PubMed Health. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Women's and Children's Health (UK). Constipation in Children and Young People: Diagnosis and Management of Idiopathic Childhood Constipation in Primary and Secondary Care. London: RCOG Press; 2010. (NICE Clinical Guidelines, No. 99.)

5Clinical management

5.1. Disimpaction

Introduction

Faecal impaction is a severe constipation with a large faecal mass in either the rectum or the abdomen, and/or overflow soiling. Disimpaction involves the evacuation of impacted faeces using one or more different treatment regimens.

There is no one treatment regimen that will suit all children and there are a variety of approaches to the management of constipation, including disimpaction, in evidence throughout the NHS in England and Wales, as well as differences in practice between clinicians.

Optimal medical management of children with chronic idiopathic constipation will tend to reduce the number requiring surgical intervention. However, patients who remain impacted despite pharmacological regimens may require manual evacuation under general anaesthetic.

In this section, the available evidence for disimpaction will be reviewed and recommendations made based on the GDG's expert interpretation of that evidence.

Clinical question

What is the effectiveness of pharmacological and surgical intervention for disimpaction in children with chronic idiopathic constipation?

Studies considered in this section

Studies were considered if they:

  • included neonates, infants or children up to their 18th birthday with chronic idiopathic constipation
  • included the following pharmacological and surgical interventions:

    stimulant laxatives (both oral and rectal medications)

    osmotic laxatives (both oral and rectal medications)

    manual evacuation of the bowel under general anaesthesia

  • included the following outcomes:

    changes in frequency of bowel movements

    changes in consistency and/or appearance of stools

    changes in pain and/or difficulty on passing stools

    changes in frequency of episodes of soiling

    reduction in laxatives use

    parent/child views and/or satisfaction or quality of life

  • were not case reports
  • were published in English.

No restrictions were applied on the publication date or country.

Overview of available evidence

A search was performed on pharmacological and surgical interventions for disimpaction and ongoing maintenance in children with chronic idiopathic constipation. A total of 986 articles were identified and 143 articles were retrieved for detailed assessment. Of these, five studies were identified for inclusion in this review: two randomised controlled trials (RCTs), one retrospective cohort (multicentre) and two prospective case series.

Narrative summary

One RCT conducted in the USA61 (2002) [EL=1-] investigated the efficacy and safety of four different doses of PEG 3350 without electrolytes in the treatment of childhood faecal disimpaction. The study included 41 children with functional faecal retention and with evidence of faecal impaction at physical examination (27 male, median age 7.5 years, range 3.3 to 13.1). Children were randomised into four groups and each group received a different dose of PEG 3350: group I (n=10) received 0.25 g/kg/day, group II (n=10) 0.5 g/kg/day, group III (n=10) 1.0 g/kg/day and group IV (n=10) 1.5 g/kg/day. Medication was taken for 3 consecutive days at breakfast, premixed with a solution flavoured with orange (maximum dose 100 g daily). Outcomes were measured 5 days after starting treatment (48 hours after last drug use).

Clearance of faecal impaction was achieved in 30 patients (75%). Significantly more children on higher doses of laxatives were disimpacted compared to children on lower doses, with values for each group being estimates taken from a bar chart (group I 5%, group II 4%, Group III 9%, group IV 10%; P < 0.05 groups III and IV [95%] versus groups I and II [55%]). Thirty-three children (83%) had more than three bowel movements during the 5 day study. The mean time of the first bowel movement after initiation of treatment was 1.89 ± 0.46 days for the total sample. Children on higher doses had significantly higher number of bowel movements compared to baseline than children on lower doses of laxative, with values for each group being estimates taken from a bar chart and baseline value less than 2 for all groups (group I had 6, group II had 8, group III had 11, group IV had 12; P < 0.005 for each group compared to the others). No significant differences were found in any of the following parameters among the four groups: straining, stool consistency, stool amount, gas and cramping. Ninety-five percent of children took PEG 3350 on the first attempt. Mean daily volumes required to take the appropriate study dose were not significantly different between groups. At baseline the duration of constipation was significantly longer for the group receiving 1.5 g/kg/day compared to the group receiving 0.5 g/kg/day (P < 0.03).

An RCT conducted in the USA62 (1993) [EL=1-] compared the efficacy and acceptability of the treatment of faecal impaction using either mineral oil or pineapple isotonic intestinal lavage solution containing PEG 3350 (it was unclear from the paper whether this contained electrolytes or not). The study included 48 children aged over 2 years with idiopathic constipation. Children were randomised into two groups: group I (n=17) received 2 to 8 tablespoons of mineral oil in two divided doses for 2 days, whereas group II (n=19) received a pineapple flavoured balanced oral lavage solution (sweetened with aspartame [Nutra-Sweet®, The NutraSweet Company]) 20 ml/kg/h to drink for 4 hours once daily on 2 consecutive days. Children were reassessed 2 days after completing treatment.

The number of bowel movements after treatment increased significantly in children treated with lavage solution compared to children treated with mineral oil (more than five bowel movements 9 children, one to five 8 children, none 2 children versus more than five 2, one to five 10, none 5; P < 0.005). The first bowel movement after treatment occurred significantly quicker in children taking lavage solution compared to those taking mineral oil (under 1 day 14 children, more than 1 day 3 children, none 2 children versus under 1 day 6, more than 1 day 6, none 5; P < 0.01). Palpable abdominal masses were found in significantly more children taking lavage solution compared to children taking mineral oil (no palpable abdominal masses 17 children, a few 1 child, many 1 child versus none 10, a few 4, many 3; P < 0.005). No children treated with mineral oil experienced vomiting whereas some children treated with lavage solution did (no vomiting 17 children, occasional vomiting 0 children, a lot of vomiting 0 children versus no vomiting 12, occasional vomiting 6, a lot of vomiting 1; P < 0.005). Compliance was significantly better in children taking mineral oil than in those taking lavage solution (good compliance 14 children, fair 3 children, poor 0 children versus good compliance 6, fair 7, poor 6; P < 0.01). There were no significant post-treatment differences between the two groups regarding cramps or bloating, abdominal distension, consistency of stools, anal fissure, anal sphincter tone, perineal soiling and willingness to consider the same treatment in case of recurrence of impaction. There were no significant differences at baseline between the two groups regarding: duration of constipation, frequency of stooling, associated encopresis, rectal bleeding, previous treatments with enemas or fibre diet, palpable abdominal masses, abdominal distension, anal fissure, perineal soiling, sphincter tone and consistency of stool. Significantly more patients in the lavage group gave a history of previous treatment with mineral oil (P < 0.05). Twelve patients failed to return for the 2 days post-treatment reassessment.

A multicentre retrospective cohort study conducted in the UK63 (2007) [EL=2-] estimated the clinical and economic impact of using PEG 3350 plus electrolytes (macrogol 3350 [Movicol®, Movicol Paediatric Plain®, Norgine, Harefield, Middlesex]) in an outpatient setting compared to enemas and suppositories and manual evacuation to treat paediatric faecal impaction. The study included 224 children aged 2 to 11 years from five different centres who were suffering from intractable constipation. A total of 112 children at the five centres had received PEG 3350 plus electrolytes. These were compared to 101 children in the five centres who received enemas and suppositories and with 11 children in two of the centres who underwent manual evacuation of the bowel under anaesthesia.

Significantly more patients who received PEG 3350 plus electrolytes were disimpacted within 5 days compared to patients who received enemas and suppositories and those who underwent manual evacuation of the bowel under anaesthesia (97%, CI 94% to 100% versus 73%, CI 58% to 89% versus 89%, CI 67% to 100%; P < 0.001). No significant differences were found between the three groups for time to initial disimpaction and time to disimpaction for those who did not disimpact within 3 days. The mean number of doses required for successful disimpaction within 5 days were 29 (95% CI 13 to 44) sachets for PEG 3350 plus electrolytes, 2 (95% CI 1 to 3) units for enemas and 1 (95% CI 1 to 2) units for suppositories. Significantly more children who underwent manual evacuation of the bowel under anaesthesia suffered from vomiting as an adverse effect of the intervention compared to children who received PEG 3350 plus electrolytes or enemas and suppositories (18% versus 2% and 2%; P < 0.01). There were no significant differences among three groups for: urinary tract infection, dermatitis around anus, thrush and gastric illness.

A prospective case series (phase 1 of the study; Phase 2 is an RCT) conducted in the UK64 [EL=3] assessed the efficacy of polyethylene glycol 3350 plus electrolytes (PEG+E) as oral monotherapy in the treatment of faecal impaction in children and compared PEG+E with lactulose as maintenance therapy in a randomised trial. The study included 63 constipated children (mean age 5.7 years, 68% boys) with intractable constipation that had failed to respond to conventional treatment and would require hospital admission for disimpaction (prior to enrolment 37% children reported taking at least one laxative medication, the most common of which was lactulose). Children received PEG+E (13.8 g powder dissolved in at least 125 ml water per sachet) plus electrolytes, administered orally in hospital according to an escalating dosing regimen until disimpaction was achieved (up to 7 days). Successful disimpaction was indicated by the passage of watery stools.

Disimpaction was successful in 58 children (92%) (25 children aged 2 to 4 years [89%] and 33 children aged 5 to 11 years [94%]). Disimpaction was achieved at a mean of 5.7 ± 1.2 days (median 6.0 days, range 3 to 7 days). Disimpaction was achieved at 5.8 ± 1.2 days (median 6.0 days, range 3 to 7 days) in children aged 2 to 4 years (n=25) and in 5.6 ± 1.1 days (median 6.0 days, range 3 to 7 days) in children aged 5 to 11 years (n=33). The maximum dose required to achieve disimpaction was 6 sachets/day for the total population (4 sachets/day for children aged 2 to 4 years and 6 for children aged 5 to 11 years). The mean number of sachets required to achieve disimpaction was 19.6 (SD 7.5) for the 58 children with successful disimpaction (mean 14.3 sachets, SD 4.5 for children aged 2 to 4 years and 23.6 sachets, SD 6.8 for children aged 5 to 11 years). Five children (8%) did not complete phase 1: three children withdrew before receiving any study medication and two children failed to disimpact within the time allowed. The two children who failed to disimpact in the 7 days specified in the study protocol were continued on PEG+E administration and eventually disimpacted.

A prospective case series conducted in the USA65 (2001) [EL=3] examined the efficacy and dosing of PEG 3350 without electrolytes (MiraLAX®, Schering-Plough HealthCare Products, Inc.) in children with constipation. The study included 24 constipated children aged 18 months to 12 years. Data were available for only 20 children who completed the study (nine boys, mean age 6.09 years ± 4.2). Eleven children had constipation alone whereas nine children had constipation and soiling. Children received a PEG solution, at an initial dose ~1g/kg body weight per day (14 ml/kg/day solution) given in two divided doses for 8 weeks. PEG powder was dissolved in water, juice or other clear liquid beverage. For determination of best dose for each child, parents were asked to increase or decrease the volume of PEG solution by 20% every 3 days as required to yield two soft-to-loose stools per day. Children of appropriate developmental status were advised to sit on the toilet for 5 minutes after each meal. Patients were examined on enrolment and at the end of 8 weeks of therapy for the presence or absence of a palpable faecal mass, faecal impaction and rectal dilatation. Soiling frequency, painful defecation and fear of defecation or stool withholding at enrolment were compared with that recorded on diary forms during the last 2 weeks (weeks 7 and 8) of treatment.

Soiling frequency decreased significantly (n=9) after treatment when compared to baseline (mean 10.0 ± 2.4 standard error of the mean [SEM] versus 1.3 ± 0.7; P = 0.003) and total resolution of soiling occurred in four patients (44.4%). Painful defecation (n=20) was completely resolved with treatment compared to its presence in 75% of children at baseline (P < 0.0001). Fear of defecation or stool withholding decreased significantly during treatment compared to baseline (5% versus 70%; P < 0.0001). No abdominal faecal mass was found in any children (n=18) after treatment: this was significant when compared to findings at baseline (abdominal mass present in 44%; P < 0.0029). Faecal rectal impaction was present in significantly more children (n=18) before than during treatment (83% versus 22%; P < 0.0006). Dilated rectal vault was found in significantly fewer children after treatment (n=18) than at baseline (11% versus 78%; P < 0.0001). The final effective dose during the last 2 weeks of treatment was a mean of 0.84 g/kg/day ± 0.27 SEM (range 0.27 to 1.42). Four subjects dropped from the study because of failure to return required symptoms diaries: two of these had an excellent response to therapy by parent report and two were lost to follow-up.

Evidence statement

Osmotic laxatives (oral medications) *

A prospective case series [EL=3] showed that PEG 3350 plus electrolytes administered orally in hospital for up to 7 days was effective in achieving disimpaction in constipated children.

One prospective case series [EL=3] showed that a solution of PEG 3350 without electrolytes, at an initial dose ~1g/kg body weight per day (14 ml/kg/day solution) given in two divided doses for 8 weeks, was effective in decreasing soiling frequency, painful defecation, fear of defecation or stool withholding, faecal rectal impaction and dilated rectal vault after 6 weeks of treatment. It was also effective in resolving completely abdominal rectal masses after treatment.

One RCT [EL=1-] showed that PEG 3350 administered orally in four different doses (0.25, 0.5, 1.0 and 1.5 g/kg/day) for 3 consecutive days was effective in achieving disimpaction in constipated children. It also showed that higher doses of PEG 3350 were more effective than lower doses in achieving disimpaction in constipated children.

One RCT [EL=1-] showed that a pineapple isotonic intestinal lavage solution containing PEG 3350 administered during 2 consecutive days was more effective than mineral oil administered as 2 to 8 tablespoons in two divided doses for 2 days in producing the first bowel movement and in increasing bowel movements after treatment but less effective in resolving palpable abdominal masses. PEG 3350 also showed better compliance and fewer side effects in children taking mineral oil compared to children taking lavage solution.

One multicentre retrospective cohort study [EL=2-] showed that PEG 3350 plus electrolytes was more effective in achieving disimpaction within 5 days in children with constipation when compared to children who received enemas and suppositories and those who underwent manual evacuation of the bowel under anaesthesia.

Faecal softeners

One RCT [EL=1-] showed that a pineapple isotonic intestinal lavage solution containing PEG 3350 administered during 2 consecutive days was more effective than mineral oil administered as 2 to 8 tablespoons in two divided doses for 2 days in producing the first bowel movement and in increasing bowel movements after treatment but less effective in resolving palpable abdominal masses. It also showed better compliance and fewer side effects in children taking mineral oil compared to children taking lavage solution.

Osmotic laxatives and stimulant laxatives (rectal medications)

One multicentre retrospective cohort study [EL=2-] showed that enemas and suppositories were less effective in achieving disimpaction within 5 days in children with constipation when compared to children who received PEG 3350 plus electrolytes.

Manual evacuation of the bowel under general anaesthesia

One multicentre retrospective cohort study [EL=2-] showed that manual evacuation of the bowel under general anaesthesia was less effective in achieving disimpaction within 5 days in children with constipation when compared to children who received PEG 3350 plus electrolytes. It also showed that children who underwent manual evacuation of the bowel under general anaesthesia experienced more vomiting when compared to children who received macrogol 3350 plus electrolytes and those who received enemas and suppositories.

Stimulant laxatives (oral medications)

There is no evidence for the effectiveness of stimulant laxatives (oral medications) for treating disimpaction in children with constipation.

Health economic considerations

A health economic model was developed for this guideline to assess the cost-effectiveness of different strategies for disimpaction. Given the lack of evidence of differences in efficacy, the baseline assumption was that all first line pharmacological strategies had the same level of effectiveness, although different assumptions provided by the GDG were used for some of the second and third line treatments where first line treatments failed (see appendix E for a more comprehensive discussion of the health economic model). Failure was defined as ongoing constipation requiring further treatment. The GDG was interested in finding out the difference in cost for a range of strategies for disimpaction and for maintenance and whether the cost of a high-priced drug would be offset by the lower cost of failure if that high-priced drug was more effective, leading to overall savings. The economic analysis also compared the total costs per patient (including the cost of failure) of various pharmacological strategies and considered the effect of different doses of treatment where these clinical data were available.

The economic analysis also calculated thresholds of cost effectiveness of treatment. Where one treatment or group of treatments was more effective than the alternative, there would need to be some additional therapeutic benefit of the more expensive option in order for it to be the preferred option on cost-effectiveness grounds. This additional therapeutic benefit was converted into quality adjusted life years (QALYs) in order to apply the NICE threshold of £20,000 per QALY to this analysis. Data on QALY weights were obtained from the published literature reviewed above.

The modelling was based on the available clinical data and on GDG consensus for parameters where data could not be identified. The modelling showed that treatments with a high chance (80%) of success cost less than treatment with a low chance of success (20%), regardless of the price of drugs used or the dose provided. Also, the cost of failure (changing doses, combining drugs and manual evacuation as a last resort) was a far greater determinant of overall cost than the cost of initial treatment.

The analysis by dose of PEG 3350 plus electrolytes showed that highly effective strategies will lead to cost savings. This is due to avoiding the high cost of invasive treatment requiring hospitalisation. Effectiveness is determined both by the type of drug used and by the dose given. The data we have been able to identify on doses of treatment suggest that higher doses of PEG 3350 plus electrolytes that lead to effectiveness levels of 95% compared with 55% for lower doses would be cost saving to the NHS.

The disimpaction model is based on a consensus by the GDG in the absence of clinical evidence that first line treatment is all equally effective but second and third line treatment (docusate plus senna) was less effective and enemas were the least effective. These are strong assumptions but pragmatic ones given the lack of data. They are based on the GDG's experience of these treatments for children in their NHS clinics. The results should be interpreted on the basis that the relative effectiveness of these treatments is still not known.

The model indicated that oral pharmacological alternatives were more than ten times cheaper than enemas which were assumed to be less effective and require hospitalisation. At a 20% failure rate, oral pharmacological treatment provided a mean benefit of 0.23 QALYs per child. The threshold analysis showed that the effectiveness of PEG 3350 plus electrolytes would have to be 2.6% higher than the next best alternative (in this case senna) in order for it to be the preferred option on cost-effectiveness grounds.

Given the lack of head to head comparisons of treatment alternatives, the health economic analysis provided transparency to the GDG's clinical judgement that treatment failure plays a major role in determining the total cost per child of disimpaction and maintenance so that the cheapest priced option is not the most cost effective overall.

GDG interpretation of the evidence

The GDG noted the evidence of the effectiveness of PEG 3350 plus electrolytes in disimpaction and this reflects their clinical experience.

The GDG noted the absence of evidence for the effectiveness of stimulant laxatives in disimpaction. However, from clinical experience the GDG concluded that they can be useful as a second line intervention. In the light of this, the GDG collated the information into a table so that clinicians can select the most appropriate second line doses of each laxative (or combination of laxatives) for their patients. The GDG recognises the importance of further research in this area.

The GDG discussed the use of Klean-Prep® (Norgine) bowel washout for disimpaction as GDG members were aware that some clinicians use it in children who fail to tolerate or to respond to oral disimpaction. The GDG noted that Klean-Prep should only be used within specialist centres and it may require inpatient admission and insertion of a nasogastric tube. The GDG also noted that the BNFC says that bowel cleansing solutions (including Klean-Prep) are used before colonic surgery, colonoscopy or radiological examination to ensure the bowel is free of solid contents, but they are not treatments for constipation. For these reasons, they agreed that they should not recommend its use.

The GDG concluded that children need to be assessed to diagnose and treat faecal impaction in the first place; otherwise not even the best maintenance treatment will work if children do not receive treatment for disimpaction beforehand. Giving maintenance treatment without disimpacting first could worsen the symptoms of constipation. The GDG noted that families should be informed that initial disimpaction treatment can increase symptoms of soiling and abdominal pain.

The GDG noted from the health economic analysis that successful disimpaction appears to drive the cost effectiveness of the treatment alternatives, not the acquisition cost of the treatments themselves. The optimal choice of treatment therefore appears to be the one likely to be of most therapeutic benefit.

The health economic analysis showed that PEG 3350 plus electrolytes would need to be more than 2.6% more effective than the next best available treatment on average to be the preferred treatment on cost-effectiveness grounds. The GDG recognises that the optimal choice of treatment depends both on the clinical efficacy of treatment and its acceptability; that is, the likelihood that a child will adhere to treatment both in the initial disimpaction phase and over time. It is the judgement of the GDG members that PEG 3350 plus electrolytes is more clinically effective as a direct consequence of being more acceptable to children because it is associated with fewer side effects and is a more palatable treatment. It is the GDG's view that PEG 3350 plus electrolytes would be at least 2.6% more effective than the next best available treatment and is therefore the optimal treatment on cost-effectiveness grounds.

The GDG understands from the evidence that PEG 3350 plus electrolytes is effective, well tolerated and safe. It can be used at home with low supervision and it is easy to titrate. The GDG's experience is that PEG 3350 plus electrolytes is safe and effective to use in children aged under 1 year. However, it is off-license for this age group and the GDG recognises the importance of further research in this area.

The health economic evidence is that the most cost-effective intervention is the one that works for the individual child since any difference in price of an individual laxative is outweighed by the downstream savings of even small changes in effectiveness of treatment; that is, avoiding unnecessary future treatment including hospitalisation (see appendix E).

The GDG concluded that enemas are effective for rectal disimpaction but the administration route is uncomfortable for children. Sodium citrate enema should be the first choice only if all other oral therapies have failed, because it produces fewer adverse effects than phosphate enemas. The GDG noted that phosphate enemas should only be used under specialist supervision with the appropriate consideration of the risk of toxicity.

Manual evacuation is effective but it requires hospital admission and general anaesthesia with the associated economic cost and disruption in the child's and family's life. Therefore it should only be used as the last resort and only when other oral and rectal treatments have failed.

Recommendations

Assess all children and young people with idiopathic constipation for faecal impaction, including children and young people who were referred to the relevant service because of ‘red flags’ but in whom there were no significant findings following further investigations (see tables 2 and 3). Use a combination of history-taking and physical examination to diagnose faecal impaction – look for overflow soiling and/or faecal mass palpable abdominally and/or rectally if indicated.

Start maintenance therapy if the child or young person is not faecally impacted.

Offer the following oral medication regimen for disimpaction if indicated:

Do not use rectal medications for disimpaction unless all oral medications have failed and only if the child or young person and their family consent.

Administer sodium citrate enemas only if all oral medications for disimpaction have failed.

Do not administer phosphate enemas for disimpaction unless under specialist supervision in hospital/healthcare centre/clinic, and only if all oral medications and sodium citrate enemas have failed.

Do not perform manual evacuation of the bowel under anaesthesia unless optimal treatment with oral and rectal medications has failed.

Review children and young people undergoing disimpaction within 1 week.

*

At the time of publication (May, 2010), Movicol Paediatric Plain is the only macrogol licensed for children under 12 years that includes electrolytes. It does not have UK marketing authorisation for use in faecal impaction in children under 5 years, or for chronic constipation in children under 2 years. Informed consent should be obtained and documented. Movicol Paediatric Plain is the only macrogol licensed for children under 12 years that is also unflavoured.

Research recommendation

What is the effectiveness of polyethylene glycol 3350 + electrolytes in treating idiopathic constipation in children younger than 1 year old, and what is the optimum dosage?

Why this is important

There is some evidence that treatment of constipation is less effective if faecal impaction is not dealt with first. Disimpaction with oral macrogols is recommended for children and their use avoids the need for rectal treatments.

Rectal treatments are used more commonly in hospital than at home. Although relatively few infants are admitted to hospital, there would be savings if initially all children were disimpacted at home.

Polyethylene glycol 3350 + electrolytes, an oral macrogol, is licensed for disimpaction in children older than 5 years. Increasing experience has shown that it is effective in infants younger than 1 year old, but evidence is limited to small case series. If dosage guidelines and evidence on macrogol use in infants were obtained and published, more healthcare professionals might be encouraged to try macrogols in this age group. It would also allow the guideline to be applicable across the whole paediatric age group.

5.2. Maintenance therapy

Introduction

There is little published evidence to guide health professionals about the pharmacological management of chronic constipation. There is no one treatment regimen that will suit all children and there are a variety of approaches regarding the management of constipation, including disimpaction, in evidence throughout the NHS in England and Wales as well as differences in practice between clinicians..

Macrogols are inert polymers of ethylene glycol which sequester fluid in the bowel. They are an effective non-traumatic means of evacuation in children with faecal impaction and can be used in the long-term management of chronic constipation.

In this section, the available evidence for ongoing treatment and maintenance will be reviewed and recommendations made based on the GDG's expert interpretation of that evidence.

Clinical question

What is the clinical effectiveness of pharmacological interventions for ongoing treatment and maintenance in children with chronic idiopathic constipation?

Studies considered in this section

Studies were considered if they:

  • included neonates, infants, or children up to their 18th birthday with chronic idiopathic constipation
  • included the following pharmacological and surgical interventions

    stimulant laxatives (both oral and rectal medications)

    osmotic laxatives (both oral and rectal medications)

    bulk forming laxatives

  • included the following outcomes:

    changes in frequency of bowel movements

    changes in consistency or appearance of stools

    changes in pain or difficulty on passing stools

    changes in frequency of episodes of soiling

    reduction in laxatives use

    parent/child views and/or satisfaction or quality of life

  • were not case reports
  • were published in English.

No restrictions were applied on the publication date or country.

Overview of available evidence

A search was performed on pharmacological and surgical interventions for disimpaction and ongoing maintenance in children with chronic idiopathic constipation. A total of 986 articles were identified and 143 articles were retrieved for detailed assessment. Of these, 15 studies were identified for inclusion in this review: 14 RCTs (7 open label, 6 double blind and 1 single blind) and 1 prospective cohort study.

Narrative summary

Laxatives versus laxatives

Osmotic laxatives versus osmotic laxatives: polyethylene glycol (PEG) versus lactulose

One meta-analysis of four RCTs comparing polyethylene glycol (PEG) versus lactulose showed that treatment success was significantly higher for PEG compared to lactulose (see figure 5.1).

Figure 5.1. Polyethylene glycol (PEG) versus lactulose in the ongoing treatment/maintenance of idiopathic constipation in children.

Figure 5.1

Polyethylene glycol (PEG) versus lactulose in the ongoing treatment/maintenance of idiopathic constipation in children.

It should be noted that different types of PEG, different definitions of treatment success and different assessment points were used in the studies.

A double-blind RCT** conducted in the UK64 (2006) [EL=1+] assessed the efficacy of PEG 3350 plus electrolytes (PEG 3350+E;) as oral monotherapy in the treatment of faecal impaction in children and to compare PEG 3350+E with lactulose as maintenance therapy in a randomised trial. The RCT included 65 children with intractable constipation that had failed to respond to conventional treatment (mean age 5.7 years, 56% children 5 to 11 years, 68% boys). All children received PEG 3350+E (13.8 g powder dissolved in at least 125 ml water per sachet) administered orally in hospital according to an escalating dosing regimen until disimpaction was achieved (up to 7 days). Fifty-eight children (67% boys, mean age 5.7 ± 2.6 years, range 2 to 11 years) entered phase 2 of the study and were randomised to receive PEG 3350+E (13.8 g powder dissolved in at least 125 ml water per sachet) or lactulose (10 g powder dissolved in at least 125 ml water) for 12 weeks. For both medications children received oral maintenance doses starting with half of the numbers of sachets required for disimpaction per day. Additional laxative treatment with senna was allowed as rescue medication if the response to a single agent alone was judged inadequate by the investigator.

There were no significant differences at baseline between the two treatment groups regarding age, sex, height and weight. Children taking PEG 3350+E (n=27) had significantly more successful defecations per week (last on-treatment value) compared to children taking lactulose (n=26) (9.4, SD 4.56, range 2 to 24 versus 5.9, SD 4.29, range 2 to 23 with difference in means 3.5, 95% CI 1.0 to 6.0; P = 0.007). No children taking PEG 3350+E reimpacted whereas seven children taking lactulose did (23%; P = 0.011). No children taking PEG 3350+E needed to use senna as rescue medication whereas eight children taking lactulose did (31%; P = 0.002). The mean number of sachets used each day for children taking PEG 3350+E (n=27) was 0.91 (SD 0.41) whereas for children taking lactulose (n=26) it was 2.41 (SD 0.91). There were no significant post-treatment differences in mean values per patient between the two groups with respect to: predominant bowel movement form, pain, straining, soiling and overall assessment of treatment. Sixty-four percent of children on PEG 3350+E (n=27) experienced adverse effects compared to 83% of children on lactulose. There was a similar incidence of adverse effects in each age group. The most commonly reported events were gastrointestinal and resolved during the study. No clinically significant abnormal values were observed in urine and plasma electrolytes after 12 weeks of maintenance therapy in either group.

A double-blind RCT conducted in France66 (2005) [EL=1+] assessed the safety of a PEG 4000 laxative without additional salts in paediatric patients. The RCT included 96 children aged 6 months to 3 years (51 male) with constipation despite their usual dietary treatment for at least 1 month. Children were randomised to receive either PEG 4000 (starting dose one sachet [4 g] and one placebo to be taken at breakfast) or lactulose (starting dose 1 sachet [3.33 g] and one placebo to be taken at breakfast) for 3 months. For both drugs, the dose could be doubled if it was ineffective in children aged 13 months to 3 years. If the maximum authorised dose was unsuccessful, one micro-enema of glycerol per day could be prescribed for a maximum of 3 consecutive days. If the child did not produce stools after treatment, two enemas could be administered at a 48 hour interval. This procedure was only allowed twice during the study. If the child produced liquid stools for more than 1 day or more than two or three stools per day depending on age, the dose could be decreased by one pair of sachets per day to a minimum of one pair of sachets every other day and possibly to transitory interruption. Outcomes were assessed at day 42 and day 84 after initiating treatment. There were no clinically relevant differences between the two treatment groups at baseline for clinical or biological parameters.

At day 42 the median stool frequency (interquartile range) was not significantly different for babies (aged 6 to 12 months) taking PEG 4000 compared to babies taking lactulose. However, for toddlers (aged 13 months to 3 years) taking PEG 4000 (n=51) the stool frequency increased significantly more than for toddlers taking lactulose (n=45) (8 stools per week, interquartile range 6 to 10 versus 6, interquartile range 5 to 7; P = 0.013). At day 84 there were no significant differences in either babies or toddlers for both treatment groups regarding stool frequency. At day 42 significantly more children taking lactulose (14 out of 41, 34%) reported a higher frequency of hard stools compared to children taking PEG 4000 (4 out of 46, 9%; P = 0.003). This remained the case at day 84 (for PEG 4000 3 out of 47, 6% versus lactulose 11 out of 40, 28%; P = 0.008).

At day 42 significantly more children taking lactulose (19 out of 44, 43%) reported using enemas compared to children taking PEG 4000 (14 out of 48, 30%). This remained the case at day 84 (for PEG 4000 8 out of 48, 17% versus lactulose 17 out of 42, 41%; P = 0.012). Faecal impaction was diagnosed in significantly more patients taking lactulose compared to children taking PEG 4000 (for PEG 4000 1 out of 51, 2% versus lactulose 6 out of 45, 13%; P = 0.049). There were no significant differences in the doses used for both medications in either babies or toddlers (babies taking PEG 1 sachet/day, median interquartile range 0.9 to 1 versus babies taking lactulose 1 sachet/day, 1 to 1.3 and toddlers taking PEG 1 sachet/day, 1 to 3 versus toddlers taking lactulose 1.1 sachet/day, 0.9 to 1.5). Treatment was stopped in one child in the lactulose group because of lack of efficacy.

A double-blind RCT conducted in the Netherlands67 (2004) [EL=1+] compared the clinical efficacy and safety of PEG 3350 plus electrolytes (PEG 3350+E [Transipeg®, Mundipharma Medical Company]) and lactulose in paediatric idiopathic constipation. The RCT included 91 children aged 6 months to 15 years with constipation (49 male). During the run-in phase (1 week before treatment) no laxatives were allowed and at the end all patients received one enema daily for 3 days. Children age 6 years or under received 60 ml of Klyx (sodium dioctylsulfosuccinate and sorbitol) and children over 6 years received 120 ml Klyx. During the initial phase children were randomised to receive either PEG 3350+E or lactulose for 8 weeks (children aged 6 months to 6 years (inclusive) received 1 sachet (2.95 g) of PEG 3350+E per day or 1 sachet (6 g) of lactulose per day and children older than 6 years received 2 sachets (5.9 g) of PEG 3350+E per day or 2 sachets (12 g) of lactulose per day). Overall treatment success was defined as three or more bowel movement per week and one encopresis episode or less every 2 weeks.

After 8 weeks there were no significant differences regarding both defecation frequency per week and encopresis frequency per week for children taking PEG 3350+E compared with children taking lactulose. Success percentages were significantly greater for children taking PEG 3350+E compared to children taking lactulose (56, 95% CI 39 to 70 versus 29, 95% CI 16 to 44; P = 0.02). Significantly more sachets a day were taken by children on lactulose compared to children on PEG 3350+E (mean 2.4, SD 0.4 versus mean 1.99, SD 0.3; P = 0.03). No serious or significant side effects were recorded. Significantly more adverse effects (abdominal pain, pain at defecation and straining at defecation) were seen in patients taking lactulose compared to PEG 3350+E (P < 0.05). There were no significant differences between the two groups regarding: bloating, diarrhoea, flatulence, nausea, hard stool consistency and vomiting.

Significantly more children complained of bad palatability of PEG 3350+E compared to lactulose and this caused the premature withdrawal of one patient. There were no significant differences at baseline between the two groups with respect to: age, sex, defecation frequency, encopresis, large amounts of stool and faecal impaction. Nine children dropped out of the study: four children in the PEG 3350+E group and five in the lactulose group. Two children in each group were lost to follow-up. Overall treatment success was independent of age (under 6 years and 6 years or over) and use of laxatives for more than 1 year prior to the start of the study. In children treated for less than 1 year a significant difference in success was found between those treated with PEG 3350 and those treated with lactulose (63% versus 1% respectively; P = 0.02).

An open label RCT (crossover) conducted in the USA68 (2002) [EL=1-] compared the efficacy of PEG 3350 and lactulose in the treatment of chronic constipation in children. Forty-four children aged 2 to 16 years (mean 7.8 ± 3.7) referred for evaluation of constipation were included. Children were randomised to receive either PEG 3350 without electrolytes 10 g/m2/day orally for 2 weeks (mean weight adjusted dose: 0.3 g/kg/day, range 0.2 to 0.5) or lactulose 1.3 g/kg/day orally for 2 weeks. There was no washout period in between the two medications. Outcome measures were stool frequency, stool form, ease of passage, effectiveness (global assessment, as reported by parent or guardian) and laxative preference (based on efficacy, ease of administration and side effects). The mean number of bowel movements, the stool form (mean sum of scores) and the easy of stool passage (mean sum of scores) were not significantly different in children taking PEG 3350 compared to children taking lactulose. PEG 3350 was significantly more effective than lactulose (PEG 3350 (n=37): 84% effective versus lactulose (n=37): 46% effective; P = 0.002). Seventy-three percent of patients said they preferred PEG 3350 compared to 27% who said they preferred lactulose. Seven patients withdrew during the first 2 week treatment period due to lack of efficacy of the assigned intervention: six of these patients were taking lactulose at the time of withdrawal.

Osmotic laxatives versus osmotic laxatives: polyethylene glycol (PEG) 3350 without added electrolytes versus magnesium oxide (milk of magnesia)

A prospective cohort study conducted in the USA69 (2002) [EL=2+] determined the efficiency, acceptability and treatment dosage of PEG 3350 without electrolytes during a 12 month treatment period in children with idiopathic constipation and encopresis. The study included 49 children aged 4 years or older referred for idiopathic constipation and encopresis of more than 1 year duration. Twenty-eight children received PEG 3350 17 g dissolved in 240 ml of a beverage such as juice or Kool-Aid® at an initial dose of 0.5 to 1 g/kg/day (20 boys, mean age 8.7 years ± 3.6, range 4.1 to 17.5 years). Twenty-one children received milk of magnesia (MOM) at an initial dose of 1 to 2.5 ml/kg (17 boys, mean age 7.3 years ± 3.0, range 4.0 to 13.9 years). Large laxative dosages could be divided into two daily doses. Parents were told to adjust the dose of medication by 30 ml for PEG 3350 and by 7.5 ml (one half tablespoon) for MOM every 3 days to a dosage that resulted in one to two soft bowel movements per day and prevented soiling and abdominal pain. If the child retained stools despite compliance with assigned laxative, daily senna was added to treatment. Treatment lasted 12 months. Children were assessed at 1, 3, 6 and 12 months after initiating treatment.

Bowel movement frequency was not significantly different between the two treatment groups at any of the four assessment points. The mean frequency of soiling decreased significantly more in children taking MOM compared to children taking PEG 3350 at 1 and 12 months, with results being estimates taken from a bar chart as not reported in text (at 1 month PEG: 3.0 versus MOM: 0.5 and at 12 months PEG: 0.9 versus MOM: 0.1, P < 0.01 for both assessment points). There were no significant differences between the two groups at 3 and 6 months. The mean medication dosage for children who were doing well or improved was 0.6 g/kg ± 0.2 (range 0.3 to 1.1) for PEG and 1.4 ml/kg ± 0.6 (range 0.6 to 2.6) for MOM at 1 month. At 3 months it was 0.6 g/kg ± 0.3 (range 0.3 to 1.4) for PEG and 1.2 ml/kg ± 0.5 (range 0.6 to 2.4) for MOM. At 12 months the dose for PEG was 0.4 g/kg ± 0.1 (range 0.1 to 0.7). Only two children still required MOM at 12 months. Their dosages were 0.4 and 1.6 ml/kg, both less than the initial treatment dosage.

The mean doses for both treatments at 12 months did not differ significantly between children with or without initial palpable abdominal faecal masses. None of the patients required an increased dosage of either medication over time. Five children received a stimulant laxative in addition to PEG 3350 and one child received a stimulant laxative in addition to MOM (P > 0.2). No children reported disliking the taste of PEG 3350 and no parents reported that their child refused to take it in juice or Kool-Aid. At 12 months 33% of children refused to take MOM and they were rated as not doing well because they were taking PEG 3350 without electrolytes instead. They were excluded from the outcomes reported at previous assessment points.

An open label RCT conducted in the USA70 (2006) [EL=1-] compared the efficacy, safety and patient acceptance of PEG 3350 without added electrolytes versus magnesium oxide (milk of magnesia) over 12 months. The RCT included 79 children diagnosed with idiopathic constipation with faecal incontinence (65 boys, age range 4 to 16.2 years, median age 7.4 years, mean age 8.1 years ± 3.0). Children were randomised to receive PEG 3350 0.7 g/kg body weight per day for 12 months or MOM 2 ml/kg body weight per day for 12 months. If necessary, children were disimpacted with one or two phosphate enemas before starting laxative therapy. There were no significant differences at baseline between the two groups.

Both the improvement and the recovery rates at 12 months were not significantly different for children taking PEG 3350 compared to children taking MOM (improvement: PEG [n=34] 62%, MOM [n=21] 43% and recovery rates: PEG 33%, MOM 23%). At 12 months the frequency of bowel movements and the frequency of episodes of faecal incontinence were not significantly different between children taking PEG and children taking MOM. Two children (5%) continued to refuse PEG versus 14 children (35%) who continued to refuse MOM during the 12 months of the study (P < 0.001). By 12 months 27 children had left the study or were lost to follow-up (PEG 7 out of 39 versus MOM 20 out of 40). In the PEG 3350 group two children were lost to follow-up monitoring, two had refused PEG 3350, one child was allergic to PEG and two children were receiving senna. These seven children were counted as not improved and not recovered. In the MOM group two children were lost to follow-up monitoring, three children had discontinued study participation, 14 children (35%) had refused to take MOM and one child was receiving senna. Mean treatment doses at 1 month were 0.7 ± 0.2 g/kg body weight for PEG and 1.2 ± 0.7 ml/kg body weight. At 3 months doses were 0.6 ± 0.3 g/kg body weight for PEG and 1.2 ± 0.8 for MOM. Mean treatment doses were similar in children who improved and those who did not improve for both treatments.

Osmotic laxatives versus stimulant laxatives

An open label RCT (crossover) conducted in the UK71 (1977) [EL=1-] compared the effectiveness and side effects of a standardised senna syrup with lactulose in the treatment of childhood idiopathic constipation. The RCT included 21 children aged under 15 years with a history of constipation treated at home for 3 months or more. Children were randomised to receive either senna syrup (10 to 20 ml daily) for 2 weeks or lactulose (10 to 15 ml daily) for 2 weeks with 1 intermediate week with no treatment. Each preparation was given throughout the appropriate treatment week in a daily dose determined by the age of the child.

The number of patients passing stools of any kind each day was not significantly different for children taking lactulose compared to children taking senna. The number of patients passing normal stools each day was significantly larger in patients taking lactulose compared to patients taking senna (lactulose 13.4 versus senna 8.43, P < 0.01). One patient on senna at the beginning of study failed to attend at the end of the first week assessment but was included in the analysis.

Osmotic laxatives versus faecal softeners

An open label RCT conducted in Iran72 (2007) [EL=1-] compared the clinical efficacy and safety of liquid paraffin and lactulose in the treatment of idiopathic childhood constipation. The study included 247 children (127 male) with chronic idiopathic constipation aged 2 to 12 years (mean age 4.1 years ± 2.1). All children received one or two enemas daily for 2 days to clear any rectal impaction (30 ml/10 kg body weight of paraffin oil). Children were randomised to receive either liquid paraffin orally (1 to 2 ml/kg twice daily) for 8 weeks or lactulose orally (1 to 2 ml/kg twice daily) for 8 weeks. For determination of the best dose for each child, parents were asked to increase the volume of each drug by 25% every 3 days as required to yield one or two firm to loose stools. Outcomes were measured during the first 4 weeks and during the last 4 weeks of treatment.

Stool frequency per week during the first and last 4 weeks of treatment increased significantly more in children taking liquid paraffin (n=127) compared to children taking lactulose (n=120) (first 4 weeks 12.1 ± 3.2 versus 9.2 ± 2.1, P < 0.001 and in the last 4 weeks 13.1 ± 2.3 versus 8.1 ± 3.1; P < 0.001). Encopresis frequency per week during the first 4 weeks of treatment decreased significantly more in children taking liquid paraffin compared to children taking lactulose (first 4 weeks 1 ± 4.3 versus 2 ± 4.6; P = 0.07). During the last 4 weeks no child on liquid paraffin experienced encopresis compared to a frequency of 3 ± 4.1 in children taking lactulose; P < 0.001. Success rate was significantly larger during the first 4 weeks and at the end of 8 weeks of treatment in children taking liquid paraffin compared to children taking lactulose (first 4 weeks: 90% versus 52%; P < 0.001), (end of 8 weeks: 85% versus 29%; P < 0.001). The final mean effective dose was significantly larger in children taking lactulose compared to children taking liquid paraffin (2.08 ml/kg/day ± 0.21 versus 1.72 ml/kg/day ± 0.13; P < 0.001).

An open label RCT conducted in Turkey73 (2005) [EL=1-] determined and compared the efficacy, safety and optimal dose of liquid paraffin and lactulose in children with chronic idiopathic constipation. The study included 40 children aged 2 to 12 years (22 male, mean age 3.7 years ± 2.7) referred for evaluation of constipation with evidence of faecal impaction. Children were randomised to receive either liquid paraffin or lactulose for 8 weeks. The medication was administered orally as a suspension at 1 ml/kg, twice a day for each drug. For determination of the best dose for each child, parents were asked to increase or decrease the volume of each drug by 25% every 3 days as required, to yield two firm to loose stools per day. The maximum dose used throughout the study was 3 ml/kg/day for each drug. Outcomes were measured at 4 and 8 weeks after initiation of treatment. Stool frequency and stool consistency were recorded by the parents in daily diary forms (stool consistency scoring: 1, hard; 2, firm; 3, loose stools).

The mean stool consistency during the first 4 weeks of treatment improved significantly more for children taking lactulose (n=20) compared to children taking liquid paraffin (n=20) (1.71 ± 0.5 versus 2.17 ± 0.5; P < 0.01). There were no significant differences in stool consistency when comparing both groups during the last 4 weeks of treatment. The stool frequency per week increased significantly more in children taking liquid paraffin compared to children taking lactulose, both during the first and the last 4 weeks of treatment (first 4 weeks: mean 13.3 ± 4.2 versus 10.2 ± 4.4; P < 0.05), (last 4 weeks: mean 16.1 ± 2.2 versus 12.3 ± 6.6; P < 0.05). The optimal dose of drugs was not significantly different for children taking liquid paraffin compared to children taking lactulose (mean 1.88 ml/kg/day ± 0.27 versus 2.08 ml/kg/day ± 0.27). These data were reported in a table and it was assumed that this represented the whole study period.

Data reported in text for the last 4 weeks of treatment established the optimal dose for liquid paraffin was 1.72 ± 0.18 and for lactulose it was 1.82 ± 0.57. Compliance rate during the first 4 weeks of treatment was not significantly different when comparing the two groups. At the end of 8 weeks significantly more children complied with taking liquid paraffin than with taking lactulose (n=90% versus n=60%; P = 0.02).

Stimulant laxatives versus faecal softeners

A single blind RCT conducted in the USA74 (1982) [EL=1-] compared the efficacy of mineral oil and standardised senna concentrate (Senokot®, Reckitt Benckiser Healthcare (UK) Limited) in the treatment of idiopathic constipation in children. The RCT included 37 children aged 3 to 12 years treated for chronic idiopathic constipation in a specialist clinic. Children received a 5-day course of oral bisacodyl (most patients) and daily enema for 3 to 5 days in addition (a minority). Children were randomised into two groups. Group 1 (n=19) received mineral oil orally twice a day in doses sufficient to induce loose stools and leakage of oil per rectum. After the first week of treatment, the dose was reduced until the leakage ceased. This dose (range 1.5 to 5.0 ml/kg/day) was maintained for minimum of 3 months. The second group (n=18) received senna (tablet or syrup) in doses sufficient to induce at least 1 bowel movement daily during the first 2 weeks of treatment. This dose was maintained for 3 months. Tapering was accomplished by changing from daily to every other day and then every third day of medication. Treatment lasted approximately 6 months. Children in the mineral oil group were followed up for an average of 10.1 months while children in the senna group were followed up for an average of 10.5 months.

At 1 month the percentage of patients experiencing daily bowel movement was not significantly different when comparing the two groups. At 3 months all children on mineral oil were experiencing daily bowel movements compared to 72% of children on senna (P < 0.05). At the final follow-up significantly more children on mineral oil were experiencing daily bowel movements compared to children on senna (mineral oil: 89% versus senna 50%; P < 0.05). At all three assessment points daily soiling decreased significantly more in children taking mineral oil compared to children taking senna (at 1 month: mineral oil 11% of patients versus senna 39%, P < 0.05; at 3 months: mineral oil 11% of patients versus senna 50%, P < 0.05; at final follow-up: mineral oil 6% of patients versus senna 44%, P < 0.05). Sixty-eight per cent of children on mineral oil were reliably compliant with medication during the first 3 months of treatment compared to 78% of children on senna. At the latest follow-up 55% of children on mineral oil successfully discontinued regular medication compared to 22% of children on senna. An additional 33% of children discontinued senna because of unacceptable symptom control. Forty-five percent of children in each group remained on regular medication. There were significantly more episodes of recurrence or treatment of symptoms per month in children taking senna compared to children taking mineral oil (senna mean 0.34 ± 0.36 versus mineral oil 0.09 ± 0.08; P < 0.01). There were no significant baseline differences between the two groups regarding mean age, median age at onset of symptoms, percentage of patients who had received prior treatment for constipation, gender ratio, faecal soiling, overt retentive behaviour, enuresis, ‘difficult’ toilet training and primary failure of toilet training. One patient on mineral oil was lost to follow-up after the 3-month visit and not considered in the results. There was no attrition or loss to follow-up in the senna group.

Laxatives versus placebo

A double-blind RCT (cross over, multicentre) conducted in the UK75 (2008) [EL=1+] assessed the efficacy and safety of PEG 3350+E for the treatment of chronic idiopathic constipation in children. The study included 51 children aged 24 months to 11 years (29 girls) with chronic constipation for at least 3 months. Children were randomised to receive PEG 3350+E (6.9 g powder/sachet) or placebo (6.9 g powder/sachet) for 2 weeks with a 2-week washout period in between. The dosing regimen for PEG 3350+E and placebo for children aged 2 to 6 years was: 1 sachet/day on days 1 and 2, 2 sachets/day taken together on days 3 and 4), 3 sachets/day (2 morning, 1 evening) on days 5 and 6 and 4 sachets/day (2 morning, 2 evening) on days 7 and 8. For children aged 7 to 11 years the dosing regimen was: 2 sachets/day taken together on days 1 and 2, 2 sachets/day taken together on days 3 and 4), 5 sachets/day (2 in the morning, 3 in the evening) on days 5 and 6, and 6 sachets/day (3 in the morning, 3 in the evening) on days 7 and 8. For both groups if diarrhoea was present, doses were decreased by two sachets or parents were instructed to miss a day of medication. If there were loose stools doses were decreased by one sachet.

Children on PEG 3350+E experienced significantly more complete defecations per week compared to children on placebo, both for the intention to treat (ITT) population and the per protocol (PP) population. Results for the ITT population were PEG+E (n=47): mean 3.12, SD 2.050, range 0.00 to 8.87 versus placebo (n=48): mean 1.45, SD1.202, range 0.00 to 3.73; treatment difference 1.64; P < 0.001, 95% CI 0.99 to 2.28. Results for the PP population were PEG+E (n=36): mean 3.63, SD 1.980, range 0.00 to 8.87 versus placebo (n=36): mean 1.63, SD 1.229, range 0.00 to 3.73; treatment difference 1.96; P < 0.001, 95%CI 1.19 to 2.72. Data do not include the washout period. Children on PEG 3350+E (ITT population) experienced significantly more defecations in general compared to children on placebo (PEG+E (n=47): mean 5.68, SD 2.771 versus placebo (n=47): mean 4.10, SD 2.503; treatment difference 1.58; P = 0.003, 95% CI 0.55 to 2.60). Children on PEG 3350+E (ITT population) experienced significantly less pain on defecation compared to children on placebo (PEG+E (n=47): mean 0.49, SD 0.727 versus placebo (n=47): mean 0.77, SD 0.863; treatment difference -0.28; P = 0.041, 95% CI −0.52 to −0.01). Children on PEG 3350+E (ITT population) experienced significantly less straining on defecation compared to children on placebo (PEG+E (n=47): mean 0.72, SD 0.789 versus placebo (n=47): mean 1.37, SD 1.041; treatment difference -0.65; P = 0.001, 95%CI -0.97 to -0.33). The stool consistency improved significantly more in children on PEG 3350+E compared to children on placebo (PEG+E (n=47): 1.73, SD 0.497 versus placebo (n=47): 2.21, SD 0.556; treatment difference -0.48; P = 0.001, 95% CI -0.68 to -0.27). The percentage of hard stools decreased significantly more in children on PEG 3350+E compared to children on placebo (PEG+E (n=47): 14.64, SD 26.041 versus placebo (n=47): 38.19, SD 39.508; treatment difference -23.55; P < 0.001). There were no significant differences between children on PEG 3350+E and children on placebo regarding abdominal pain on defecation and faecal incontinence. The mean effective dose of PEG 3350+E was 0.6 g/kg/day in children aged 2 to 6 years and 0.7 g/kg/day in children aged 7 to 11 years.

One double-blind RCT (multicentre) conducted in the USA76 (2008) [EL=1+] established the efficacy and best starting dose of PEG 3350 in the short-term treatment of children with idiopathic constipation. The study included 103 children aged 4 to 16 years (69 boys, mean age 8.5 years ± 3) with chronic constipation. Patients taking other laxatives were only included if they had less than three bowel movements per week while taking the laxative. All children received behavioural treatment consisting of instructions to sit on the toilet for 10 minutes twice after meals, positive reinforcement using age-appropriate printed calendars and special stickers for days without episodes of faecal incontinence and others with bowel movements. Children were randomly assigned in blinded fashion in a 1:1:1:1 ratio within each participant site into four groups. Group 1 received PEG 3350 without electrolytes at 0.2 g/kg/day single dose (maximum 8.5 g per day), group 2 received PEG 3350 without electrolytes at 0.4 g/kg/day single dose (maximum 17 g per day), group 3 received PEG 3350 without electrolytes at 0.8 g/kg/day single dose (maximum 34 g per day) and the last group received a placebo. Treatment lasted 3 weeks. Assessments were conducted at 7 and 14 days after medication started. Response to treatment was defined as 3 or more bowel movements (BM) during the second week of treatment. Patients were considered failures and withdrawn from study if they had no BM for 7 days or developed faecal impaction at any point; however intention to treat analysis was performed.

There were no significant differences in baseline characteristics between the four groups. The percentage of children who responded to treatment was significantly higher when comparing each and all treatment groups with placebo (group 1 (n=26): 77%, group 2 (n=27): 74%, group 3 (n=26): 73%, placebo (n=24): 42%; P < 0.04 each group versus placebo; P = 0.026 all treatment groups versus placebo). There were no significant differences between treatment groups regarding this outcome. There were no significant predictors of success by controlling for age, duration of constipation, prior laxative use, presence of stool in rectum, gender and presence of faecal incontinence at baseline. There was a significant increase in the final number of bowel movements in the different treatment groups compared to placebo (overall difference between treatment groups and placebo P = 0.017; P = 0.015 dose-response trend). Note that figures for the after treatment were reported in a graph from which it is difficult to extract the data.

There was no significant difference in weekly number of faecal incontinence episodes among the four groups. Stool consistency became softer in all treatment groups compared with placebo and comparing all treatment groups with each other. Changes in stool consistency measured as number of faecal incontinence episodes were: group 1: before 2.8 ± 0.8, after 2.1 ± 0.7; group 2: before 2.6 ± 0.9, after 1.7 ± 0.6; group 3: before 2.9 ± 0.7, after 1.5 ± 0.7; placebo: before 3.0 ± 0.8, after 2.4 ± 0.9; P < 0.003 each group versus placebo; P < 0.003 test for trend; P < 0.003 overall difference amongst treatment groups. Straining decreased in all treatment groups compared with placebo, particularly for those in group 2 and group 3 (mean straining scores: group 1: before 2.3 ± 1.1, after 1.4 ± 0.9; group 2: before 1.9 ± 1.2, after 1.0 ± 1.0; group 3: before 2.0 ± 1.0 after 0.9 ± 0.6; placebo: before 2.7 ± 1.2, after 1.5 ± 1.2; P < 0.003 each group versus placebo; P < 0.003 test for trend; P < 0.003 overall difference between treatment groups. There were no significant difference amongst groups regarding incidence and severity of adverse effects (group 1: 9 out of 26 [34.6%], group 2: 16 out of 27 [59.3%], group 3: 17 out of 26 [65.4%], placebo: 14 out of 24 [58.3%]).

There were no differences in the type of non-gastrointestinal (GI) related events, the most common being headaches. There was a higher incidence of GI related events in patients receiving PEG versus placebo. As the dose of PEG increased, it also increased incidence of flatulence, abdominal pain, nausea and diarrhoea. There were no electrolyte abnormalities or differences in laboratory values among groups. Treatment failure was similar in all treatment groups but lower than the placebo (number of children who failed in group 1 was 6 out of 26 [4 BM frequency criteria, 2 with stool impaction]; in group 2 it was 7 out of 27 [3 BM frequency criteria, 4 with stool impaction]; in group 3 it was 7 out of 26 [6 BM frequency criteria, 1 with stool impaction] and in placebo (n=24) it was 14 [all related to BM frequency criteria]). Fourteen patients did not complete the 2-week treatment: eight because of treatment failure (five with impaction [two Group 1, three Group 2] and three with more than 7 days without a BM [two Group 1, one Group 3]), three because of adverse events, there was one withdrawal (lack of response (placebo)) and two cases of non-compliance (one Group 2, one Group 3). Three serious adverse events occurred requiring hospitalisation (two cases of impaction and one case of exacerbation of bipolar or depression).

Laxatives versus other interventions

Laxatives versus biofeedback

An open label RCT conducted in the USA77 (1987) [EL=1-] evaluated the efficacy of biofeedback for childhood encopresis. The study included 50 children (40 boys) aged 6 to 15 years (mean age 8.4 years) with encopresis of at least 6 months of duration. Children were randomised to receive either one 25 to 30 minute biofeedback session with reinforcement sessions at 2, 4 and 8 weeks or mineral oil orally in graded amounts (range 1 to 4 tablespoons/day) designed to induce a soft bowel movement daily for 12 weeks. Children were followed up at 3, 6 and 12 months, and outcome measures were frequency of defecation, frequency of gross incontinence, frequency of staining or minor soiling and parental perception of clinical status and overall satisfaction. Based on these measures, children were placed in groups at each assessment: some improvement, some improvement but major soiling (less than once a week), marked improvement (rarer major soiling of less than once a week or minor soiling) and complete remission.

There were no significance differences in percentage of children in remission or markedly improved receiving either treatment at 3, 6 or 12 months. At baseline the two groups were comparable with respect to age, gender, duration and severity of soiling, anorectal motility parameters and expulsion patterns. There were two children who left the study at 3 months (one from each group) and three additional children who left at 6 months (two on biofeedback). Five children were lost to follow-up at 12 months (three on biofeedback). All withdrawals were designated as treatment failures for each subsequent assessment point.

Laxatives versus behavioural intervention

A quasi RCT conducted in the UK78 (1983) [EL=1-] assessed whether behaviour therapy would suffice on its own in the treatment of severe and persistent faecal soiling or would be improved by employing a laxative as well. The study included 44 children who had soiling as a main complaint and uncomplicated idiopathic faecal incontinence after an initial assessment and physical examination (mean age 7.9 years, SD=2.3, gender not reported). All children received behavioural treatment focusing on use of the toilet and freedom from soiling. Children were quasi randomised into three groups to receive either senna or senna placebo tablets in similar dosage to senna or no medication at all. Senna and senna placebo tablets were started at a dose of 1 tablet at night. On the next visit to the clinic, if there was no improvement in the ‘use of the toilet’ and ‘being clean’ on the charts the dosage was increased to 2 tablets. The number of tablets was increased to 3 on the following visit if improvement had still not occurred. When the soiling was getting better and the child was using the toilet the dosage was kept the same. Once the child was having regular bowel movements in the toilet and not soiling the tablets were stopped altogether.

The duration of treatment was 3 months and after that, children were assessed for severity of soiling and number of children free of soiling was noted. The severity of soiling and the number of children free of soiling at 3 months were not significantly different between the three groups. The number of soiling-free children was: relieved (less than once/week or not at all): senna (n=14): 5 (35%) versus placebo (n=11): 2 (18%) versus no treatment (n=15): 9 (60%); not relieved: senna (n=14): 9 versus placebo (n=11): 9 versus no treatment (n=15): 6.

Laxatives versus probiotics

A double blind RCT conducted in Taiwan79 (2007) [EL=1+] investigated the effect of probiotics (lactobacillus casei rhamnosus, Lcr35) alone in the treatment of chronic constipation in children and compared the effect with magnesium oxide (MgO) and placebo. The study included 45 children (23 male) under 10 years old with chronic idiopathic constipation. Children were randomised into three groups to receive MgO 50 mg/kg/day, with dose split into two and then given twice a day, Lcr35 8×108 colony forming units/day (antiobiophilus 250 mg, two capsules, twice a day) or placebo (starch) during 4 weeks. Lactulose use (1 ml/kg/day) was allowed when there was no stool passage noted for 3 days. Glycerine enema was used only when there was no defecation for more than 5 days or when abdominal pain was suffered due to stool impaction.

Defecation frequency significantly increased in children taking both MgO and probiotic compared to placebo (MgO (n=18): 0.55 times per day ± 0.13; probiotic (n=18): 0.57 times per day ± 0.17; placebo (n=9): 0.37 times per day ± 0.10; P = 0.006 [placebo versus probiotic]; P = 0.01 [MgO versus placebo]) but there were no significant differences between children taking probiotic and children taking MgO in this outcome. The percentage of children having hard stools was significantly lower in children taking MgO and probiotic compared to placebo (MgO: 23.5% ± 7.9; probiotic: 22.4% ± 14.7; placebo: 75.5% ± 6.1; P = 0.02 [placebo versus probiotic]; P = 0.03 [MgO versus placebo]) but there were no significant differences between children taking probiotic and children taking MgO regarding this outcome. Children taking placebo had to make use of glycerine enema significantly more often than children taking either MgO or probiotic (MgO: mean 1.3 times ± 1.9, probiotic: 1.6 times ± 1.9, placebo: 4.0 times ± 2.1; P = 0.04 [placebo versus probiotic]; P = 0.03 [MgO versus placebo]). There were no significant differences between children taking probiotic and children taking MgO regarding this outcome.

There were no significant differences regarding use of lactulose and faecal soiling amongst the three groups. Significantly more patients were successfully treated with MgO or probiotic compared to placebo (MgO: 72.2%, probiotic: 77.8%, placebo: 11.1%; P = 0.01 [placebo versus probiotic], P = 0.01 [MgO versus placebo]). There were no significant differences between children taking probiotic and children taking MgO regarding this outcome.

No adverse effects were noted in the probiotic and placebo groups and only one patient in the MgO group suffered from mild diarrhoea. There were no significant differences at baseline amongst the three groups regarding: gender, age of enrolment, age of onset of constipation, duration of constipation, previous treatment, defecation period, stool consistency, abdominal pain, faecal soiling, bleeding during defecation, use of enema and taking fruits or vegetables daily. Four patients discontinued medication during the study period: two in the MgO group, one in the probiotic group and one in the placebo group. Two patients suffered from acute gastroenteritis (not clear whether as a consequence of the study medication) and two patients were lost to follow-up.

Evidence statement

Laxatives versus laxatives

Osmotic laxatives versus osmotic laxatives: PEG versus lactulose

One meta-analysis of four RCTs (three [EL=1+], one [EL=1-]) comparing polyethylene glycol (PEG) versus lactulose showed that treatment success was significantly higher for PEG compared to lactulose.

One double blind RCT [EL=1+] showed that PEG 3350 plus electrolytes (PEG 3350+E) was more effective than lactulose at increasing the number of successful defecations per week. One double blind RCT [EL=1+] showed that there were no significant differences between PEG 3350+E and lactulose at increasing the number of defecations per week. One open label RCT (crossover) [EL=1-] showed that there were no significant differences between PEG 3350 without electrolytes and lactulose at increasing the number of defecations per week. One double blind RCT showed that there were no significant differences between PEG 3350 without electrolytes and lactulose at increasing the stool frequency for babies (aged 6 to 12 months) at day 42 of treatment, but PEG 3350 without electrolytes was more effective than lactulose at increasing the stool frequency for toddlers (aged 13 months to 3 years) at day 42 of treatment. At day 84 there were no significant differences between both treatments in either babies or toddlers.

Two double blind RCTs [EL=1+] showed that there were no significant differences between PEG 3350+E and lactulose at decreasing soiling frequency. Two double blind RCTs [EL=1+] showed that faecal impaction was diagnosed in significantly more patients taking lactulose compared to children taking PEG 4000 (in one of the studies no children taking PEG+E reimpacted).

One double blind RCT [EL=1+] showed that there were no significant differences between PEG 3350+E and lactulose at reducing the pain and straining on passing stools. An open label RCT (crossover) [EL=1-] showed that there were no significant differences between PEG 3350 without electrolytes and lactulose at improving the ease of passage of stools.

One double blind RCT [EL=1+] showed that there were no significant differences between PEG 3350+E and lactulose at changing the predominant bowel movement form. One double blind RCT [EL=1+] showed that PEG 3350 without electrolytes was more effective than lactulose at reducing the number of children reporting hard stools. One open label RCT (crossover) [EL=1-] showed that there were no significant differences between PEG 3350 without electrolytes and lactulose at changing the stool form.

Two double blind RCTs [EL=1+] showed that significantly more sachets per day were taken by children on lactulose compared to children on PEG 3350+E. One double blind RCT [EL=1+] showed that no children taking PEG 3350+E needed to use senna as rescue medication whereas eight children taking lactulose did. One double blind RCT showed that significantly more children taking lactulose reported using enemas compared to children taking PEG 4000 without electrolytes. There were no significant differences in the doses used for both medications in either babies or toddlers.

One double blind RCT [EL=1+] showed that overall assessment of treatment was not significantly different for children taking PEG 3350+E compared to children taking lactulose. One double blind RCT [EL=1+] showed that success percentages were significantly greater for children taking PEG 3350+E compared to children taking lactulose. Overall treatment success was independent of age (whether under 6 years or 6 years and over) and use of laxatives for more than 1 year prior to the start of the study. In children previously treated for less than 1 year PEG 3350+E was significantly more successful than lactulose. One open label RCT (crossover) [EL=1-] showed that overall PEG 3350 was significantly more effective than lactulose. One double blind RCT [EL=1+] of PEG 3350 without electrolytes versus lactulose showed that treatment stopped in one child in the lactulose group because of lack of efficacy, whereas no children on PEG 3350 without electrolytes stopped therapy for this reason.

One open label RCT (crossover) [EL=1-] of PEG 3350 versus lactulose showed that 73% of patients said they preferred PEG 3350 compared to 27% who said they preferred lactulose.

One double blind RCT [EL=1+] of PEG 3350+E versus lactulose showed that 64% of children on PEG+E (n=27) experienced adverse effects compared to 83% of children on lactulose. There was a similar incidence of adverse effects in each age group. The most commonly reported events were gastrointestinal and these resolved during the study. One double blind RCT comparing PEG 3350+E versus lactulose showed that no serious or significant side effects were recorded. Significantly more adverse effects (abdominal pain, pain at defecation and straining at defecation) were seen in patients taking lactulose compared to patients taking PEG. There were no significant differences between the two groups regarding: bloating, diarrhoea, flatulence, nausea, hard stool consistency and vomiting. Significantly more children complained of bad palatability of PEG compared to lactulose and this caused the premature withdrawal of one patient.

PEG (PEG) versus milk of magnesia

One open label RCT [EL=1-] and one prospective cohort [EL=2+] showed that there were no significant differences between PEG 3350 without electrolytes and milk of magnesia (MOM) at increasing the number of defecations per week.

One open label RCT [EL=1-] showed that there were no significant differences between PEG 3350 without electrolytes and MOM at decreasing the frequency of episodes of faecal incontinence. One prospective cohort showed that the frequency of soiling decreased significantly more in children taking MOM compared to children taking PEG 3350 without electrolytes at 1 and 12 months but there were no significant differences between the two treatments at 3 and 6 months.

One open label RCT [EL=1-] showed that mean treatment doses of PEG 3350 without electrolytes versus MOM were similar in children who improved and those who did not improve for both treatments. One prospective cohort [EL=2+] showed that the mean doses for both treatments at 12 months did not differ significantly between children with or without initial palpable abdominal faecal masses. None of the patients required an increased dosage of either medication over time. Five children received a stimulant laxative in addition to PEG and one child received a stimulant laxative in addition to MOM, but this was not significant.

One open label RCT [EL=1-] showed that both the improvement and the recovery rates at 12 months were not significantly different for children taking PEG compared to children taking MOM.

Osmotic laxatives versus faecal softeners

Lactulose versus liquid paraffin

Two open label RCTs [EL=1-] showed that liquid paraffin was more effective than lactulose at increasing the number of defecations per week.

Two open label RCTs [EL=1-] showed that liquid paraffin was more effective than lactulose at decreasing the frequency of soiling per week.

One open label RCT [EL=1-] showed that lactulose was more effective than liquid paraffin at improving the stool consistency during the first 4 weeks of treatment, but both laxatives were equally effective at improving the stool consistency during the last 4 weeks of treatment.

One open label RCT [EL=1-] showed that the optimal dose of drugs was not significantly different for children taking liquid paraffin compared to children taking lactulose. One open label RCT [EL=1-] showed that the final effective dose was significantly larger in children taking lactulose compared to children taking liquid paraffin.

One open label RCT [EL=1-] comparing liquid paraffin versus lactulose showed that the success rate was significantly larger in children taking liquid paraffin compared to children taking lactulose.

Osmotic laxatives versus stimulant laxatives

Lactulose versus senna

One open label RCT (crossover) [EL=1-] showed that there were no significant differences between standardised senna syrup and lactulose at increasing the number of defecations per week. Standardised senna syrup was more effective than lactulose at increasing the number of patients passing normal stools each day.

Stimulant laxatives versus faecal softeners

Senna versus mineral oil

One single blind RCT [EL1-] showed that mineral oil was more effective than standardised senna at increasing the percentage of patients experiencing daily bowel movements and decreasing the number of children experiencing daily soiling. More children on mineral oil successfully discontinued regular medication at the latest follow-up compared to children on senna. Despite better compliance, there were significantly more episodes of recurrence and/or treatment of symptoms per month in children taking senna compared to children taking mineral oil.

Bulk forming laxatives

No evidence was found for the clinical effectiveness of bulk forming laxatives for ongoing treatment and/or maintenance in children with chronic idiopathic constipation.

Laxatives versus placebo

PEG (PEG) versus placebo

One double-blind RCT (cross over, multicentre) [EL=1+] showed that PEG 3350+E was more effective than placebo at: increasing both the number of defecations in general and the number of complete defecations per week, improving faecal incontinence, improving the stool consistency, decreasing the percentage of hard stools and reducing both pain and straining on defecation. Both treatments were equally effective at reducing abdominal pain on defecation.

One double-blind RCT (multicentre) [EL=1+] showed that PEG 3350 without electrolytes was significantly more effective than placebo at increasing weekly frequency of bowel movements, improving stool consistency and decreasing straining on defecation, but there was no significant difference in the weekly number of faecal incontinence episodes among the treatment groups and placebo.

Laxatives versus other interventions

Milk of magnesia versus probiotic versus placebo

One double blind RCT [EL=1+] showed that there were no significant differences between probiotic and milk of magnesia (MgO) at increasing daily defecation frequency and decreasing the percentage of children having hard stools, and both were more effective than placebo regarding these outcomes. There were no significant differences between the three treatments at decreasing faecal soiling. Children taking a placebo had to make use of glycerine enemas significantly more often than children taking either MgO or probiotic but there were no significant differences between children taking probiotic and children taking MgO regarding this outcome. There were no significant differences between the three groups regarding the need to use lactulose. Significantly more patients were successfully treated with MgO or probiotic compared to placebo but there were no significant differences between children taking probiotic and children taking MgO regarding this outcome. Only one patient in the MgO group suffered from mild diarrhoea.

Mineral oil versus biofeedback

One open label RCT [EL=1-] comparing mineral oil versus biofeedback showed that there were no significance differences in the percentage of children in remission or markedly improved after receiving either treatment at 3, 6 or 12 months.

Senna versus placebo versus behavioural therapy

One quasi RCT [EL=1-] comparing senna versus placebo versus behavioural therapy showed no significant difference in the severity of soiling and the number of children free of soiling at 3 months between the three groups.

Health economic considerations

An economic model for the maintenance phase of treatment post disimpaction was developed. The model covered maintenance treatment (pharmacological and antegrade continent enema [ACE] procedure) for previously disimpacted children (aged 2 to 11 years). The ACE strategy was included only as a last resort if other pharmacological strategies failed (see table E.6). Each cycle covered a three month period after initial disimpaction. Results are reported after 3 months, at the end of 1 year (four cycles) and after 2 years (eight cycles). The range of pharmacological treatment strategies described in the disimpaction model was included, together with two additional treatments which are only offered in the maintenance phase: methylcellulose and liquid paraffin. This gave a total of 15 alternative strategies as first line treatment in the maintenance phase.

Using a modelling approach it was possible to calculate how much more effective a PEG 3350 plus electrolytes treatment strategy would have to be in the maintenance phase (3 months, 1 year, 2 years) in order for it to be cost effective at the £20,000 per quality adjusted life years (QALYs) threshold. Since PEG 3350 plus electrolytes costs more in the maintenance phase, it needs to be more effective for it to be the preferred option. It has been reported earlier (the disimpaction economic model) that higher priced therapeutic strategies with higher levels of effectiveness would become cheaper overall than treatment strategies with lower initial drug costs. It is possible to estimate how much more effective PEG 3350 plus electrolytes would have to be in order for it to be preferred to all other strategies in the maintenance phase on cost-effectiveness grounds.

The maintenance model showed that, unlike the disimpaction model, the cost of drugs in the pharmacological treatment alternatives had a greater impact on the total of care than hospitalisation, which widened the gap between the cheapest and most expensive treatment options.

The analysis suggested that an increase in effectiveness from 80% to just over 85% effectiveness in the first three months of treatment (and less in the longer term) would make PEG 3350 plus electrolytes the more favourable option over the next best alternative (senna) in the maintenance phase.

GDG interpretation of the evidence

The GDG notes that the research evidence is limited and evidence is not available for the full range of medications used in clinical practice to treat idiopathic constipation. Many drugs have been used for a long time but have not been tested in clinical trials with children and young people. The GDG recognises the importance of further research in this area.

Available evidence and clinical experience supports the use of oral PEG 3350 plus electrolytes as first line treatment for disimpaction and maintenance. The economic modelling indicates that where any treatment is effective it is also cost effective. PEG 3350 plus electrolytes used as monotherapy works quickly, is well tolerated, is easy to titrate and its unflavoured presentations can be mixed with a child's favourite cold drink thus facilitating adherence and thereby increasing effectiveness and cost effectiveness compared with the alternatives. The full range of doses of PEG 3350 plus electrolytes are licensed only for children aged over 2 years, but there is evidence from case series and clinical practice that shows that they are also effective in children aged under 1 year. The GDG believes that further research is needed in this particular age group.

The GDG noticed than some, mostly low quality, studies examined liquid paraffin and milk of magnesia but these are no longer licensed or used in the UK. Additionally, it is the GDG's clinical experience that the use of liquid paraffin involves a small risk of aspiration, particularly for children whose swallowing is impaired. Furthermore, titration is difficult and it cannot be used with Docusate. The GDG also believes that better and more palatable alternatives to milk of magnesia are available.

The GDG recognises that other medications, used singly or in combination, are available, effective, low cost and commonly used. The group's experience is that often children are under-treated because effective doses are outside licensing and therefore not prescribed by health professionals. It is the GDG's view that the optimal dose of any medication is the dose that works for a particular child. Optimal doses of laxatives are also more cost effective because they prevent unnecessary consultations and treatment failure.

The GDG recognises that the preference of the child and/or family is an important factor in the success of any treatment and must be given due consideration. The GDG believes that families need ongoing support from healthcare professionals with expertise in constipation.

A significant number of children become constipated when they are younger than 1 year. These symptoms often coincide with weaning and changing milk feeds and they might not be recognised and treated. The GDG believes that, despite their young age, these children need early diagnosis and usually require medication to prevent potential long-term problems. There is evidence from case series and clinical practice which shows that PEG 3350 plus electrolytes is effective in children aged under 1 year and the GDG is aware that it is currently used in practice. Other medications that are licensed for this age group are lactulose and docusate, which need to be given at the optimal dose. It is the GDG's view that the optimal dose of any medication is the dose that works for a particular infant.

5.3. Adverse effects of laxative use

Introduction

There is little published evidence to guide health professionals about the pharmacological management of chronic constipation. It is clear that there is no one treatment regimen which will suit all children and there is a variety of approaches taken in different areas as well as large differences in practice regarding management.

In this section, we review the available evidence and make recommendations based on best available evidence for disimpaction and maintenance regimens.

Clinical question

What are the adverse effects of the medium to long term use of laxatives?

Studies considered in this section

Studies were considered if they:

  • included neonates, infants, or children up to their 18th birthday with chronic idiopathic constipation
  • included adverse effects of the medium- (6 months) and long-term (6 to 12 months or longer) use of the following laxatives (both oral and rectal medications):

    stimulant laxatives

    osmotic laxatives

    bulk forming laxatives

  • included outcomes related to palatability***
  • were not case reports
  • were published in English.

No restrictions were applied on the publication date or country.

Overview of available evidence

A total of 237 articles were identified from the searches and 45 articles were retrieved for detailed assessment. Of these 14 studies were identified for inclusion in this review plus one paper submitted by a GDG expert advisor: six RCTs, one prospective cohort, one retrospective cohort, four prospective case series and three retrospective case series.

Narrative summary

An RCT (crossover, multicentre) conducted in the UK75 (2007) [EL=1+] assessed the efficacy and safety of polyethylene glycol 3350 plus electrolytes (PEG+E) for the treatment of chronic constipation in children. The study included 51 children (29 girls) aged 24 months to 11 years with chronic constipation for at least 3 months. Children were randomised to receive PEG+E (6.9 g powder/sachet) or placebo (6.9 g powder/sachet) for 2 weeks (period I) followed by a 2-week washout period and then a second 2-week treatment period (period II) in which each group received the second medication (period III). The dosing regimen for PEG+E and placebo for children aged 2 to 6 years was: 1 sachet/day on days 1 and 2, 2 sachets/day taken together on days 3 and 4, 3 sachets/day (2 morning, 1 evening) on days 5 and 6 (and 4 sachets/day (2 morning, 2 evening) on days 7 and 8. For children aged 7 to 11 years the dosing regimen was: 2 sachets/day taken together on days 1 and 2, 2 sachets/day taken together on days 3 and 4, 5 sachets/day (2 in the morning, 3 in the evening) on days 5 and 6 and 6 sachets/day (3 in the morning, 3 in the evening) on days 7 and 8. The dosage was adjusted over the first week of treatment in periods I and III and could be adjusted in the second week of each treatment period to determine a dose at which symptoms of constipation did not occur. For both groups, if diarrhoea was present doses were decreased by two sachets or parents were instructed to miss a day of medication. If there were loose stools doses were decreased by one sachet. Safety was monitored by recording adverse events, physical examination findings and weight changes.

There were 31 adverse events among children taking PEG+E (63%) and 28 in children taking placebo (57%) during periods I and III. Most adverse effects were judged to be moderate or mild in severity. Twenty children (41%) on PEG+E experienced 41 events and 22 children (45%) on placebo experienced 45 events judged by the investigator to be at least possibly related to the study treatment. Most of these events were gastro-intestinal disorders, particularly abdominal pain (39 events [39%] in children on PEG+E and 41 events [45%] in children on placebo). One child in the placebo/PEG+E group (the group taking the placebo in the first treatment period) withdrew from the study at week 3 because of abdominal pain, assessed by the investigator as being related to study treatment. This child was taking placebo at the time of withdrawal. New clinically significant abnormalities on physical examination (mainly associated with faecal loading) were found in 13 children (8 out of 27 in the PEG+E/placebo group, 5 out of 24 in the placebo/PEG+E group). When analysed for what these children were taking for the 2 weeks before the physical examination, 23 out of the 24 reports (95.8%) occurred when the child was taking placebo. Only one report of an abnormal abdominal examination occurred while the patient was on PEG+E. The mean weight was similar before and after treatment and no significant difference was found between the two groups for change in weight while on treatment.

An RCT conducted in France66 (2005) [EL=1+] assessed the safety of a PEG 4000 laxative without additional salts in paediatric patients. The study included 96 children (51 male) aged 6 months to 3 years with constipation. Children were randomised to receive either PEG 4000 (non-branded, starting dose: 1 sachet [4 g] and 1 placebo to be taken at breakfast) or lactulose (starting dose: 1 sachet (3.33 g) and 1 placebo to be taken at breakfast) for 3 months. For both drugs, the dose could be doubled if it was ineffective in children aged 13 months to 3 years. If the maximum authorised dose was unsuccessful, one micro-enema of glycerol per day could be prescribed for a maximum of 3 consecutive days. If the child did not produced stools after treatment, 2 enemas could be administered at a 48 hour interval. This procedure was only allowed twice during the study. If the child produced liquid stools for more than 1 day or more than 2 or 3 stools per day depending on age, the dose could be decreased by 1 pair of sachets/day to a minimum of 1 pair of sachets every other day and possibly to transitory interruption. Stool frequency, abdominal pain, vomiting and nausea were recorded by parents on a self-diary evaluation booklet. Assessments were conducted at day 42 (D42) and day 84 (D84) after starting treatment.

Six non serious adverse effects occurred during the study period (5 episodes of diarrhoea in two children in both treatment groups and anorexia in one child on lactulose). Flatulence (either new onset or worsened) lasted significantly longer in children taking lactulose compared to children taking PEG 4000 (PEG 4000: median 3 days, interquartile range 1 to 4.5 days versus lactulose: median 5 days, interquartile range 3 to 19.5 days; P = 0.005). Vomiting episodes (either new onset or worsened) lasted significantly longer in children taking lactulose compared to children taking PEG 4000 (PEG 4000: median 1 day, interquartile range 1 to 2 days versus lactulose: median 2 days, interquartile range 1 to 6 days; P < 0.05). Anal irritation was reported in 5% of the children (2 out of 40, both on lactulose).

There were no differences between PEG 4000 and lactulose groups with regard to other digestive tolerance outcomes. Body height and body weight were unaffected during the 3-month treatment for boys and girls. There were no significant differences between treatment groups for the percentage of children with out of normal range values on D84 compared to baseline status. No treatment-related changes were found in serum iron, electrolytes, total protein, albumin, vitamins A and D and folates. There were no significant differences in the doses used for both medications in either babies or toddlers (babies, PEG: median 1 sachet/day, interquartile range 0.9 to 1 versus lactulose: median 1 sachet/day, interquartile range 1 to 1.3; and toddlers, PEG: median 1 sachet/day, interquartile range 1 to 1.3 versus lactulose: median 1.1 sachet/day, interquartile range 0.9 to 1.5). Treatment stopped in one child in the lactulose group because of lack of efficacy. There were no clinically relevant differences between the two treatment groups at baseline for clinical or biological parameters.

A prospective cohort conducted in the USA69 (2002) [EL=2+] determined the efficiency, acceptability and treatment dosage of polyethylene glycol 3350 without electrolytes during a 12-month treatment period in children with functional constipation and encopresis. The study included 49 children aged 4 years or more referred for functional constipation and encopresis. For 12 months, 28 children (20 boys, mean age 8.7 years ± 3.6, range 4.1 to 17.5 years) received PEG 3350 without electrolytes at an initial dose of 0.5 to 1 g/kg/day and 21 children (17 boys, mean age 7.3 years ± 3.0, range 4.0 to 13.9 years) received magnesium oxide (milk of magnesia [MOM]) at an initial dose of 1 to 2.5 ml/kg. Large laxative dosages could be divided into two daily doses. Parents were told to adjust the dose of medication by 30 ml for PEG 3350 without electrolytes and by 7.5 ml (one-half tablespoon) for MOM every 3 days to a dosage that resulted in one to two soft bowel movements per day and prevented soiling and abdominal pain. If the child retained stools despite compliance with the assigned laxative, daily senna could be added to the treatment. Medication dosage, clinically significant side effects and compliance with medication were assessed at 1, 3, 6, and 12 months after initiating treatment. Patients and parents were provided with diary sheets to record each outcome measured.

At 1 month the mean doses and range for children who were doing well or improved were 0.6 g/kg ± 0.2 (0.3 to 1.1 g/kg) for PEG and 1.4 ml/kg ± 0.6 (0.6 to 2.6 ml/kg) for MOM. At 3 months these were 0.6 g/kg ± 0.3 (0.3 to 1.4 g/kg) for PEG and 1.2 ml/kg ± 0.5 (0.6 to 2.4 ml/kg) for MOM. At 12 months the mean dose of PEG was 0.4 g/kg ± 0.1 (0.1 to 0.7 g/kg). Only two children still required MOM. Their dosages were 0.4 and 1.6 ml/kg, both less than the initial treatment dosage. The mean doses for both treatments at 12 months did not differ significantly between children with or without initial palpable abdominal faecal masses. None of the patients required an increased dosage of either medication over time. Five children received a stimulant laxative in addition to PEG and one child received a stimulant laxative in addition to MOM (P > 0.2). Some children had diarrhoea (number not reported in paper). None of the children in the PEG group became dehydrated. Children receiving PEG and their parents did not report increased flatus, abdominal distension or new onset of abdominal pain. These outcomes were not reported for MOM. No children reported disliking the taste of PEG and no parents reported that their child refused to take it in juice or Kool-Aid. Thirty-three percent of children refused to take MOM.

A retrospective case series conducted in the USA80 (2003) [EL=3] reviewed the efficacy of PEG as a single agent for the treatment of constipation in children with dysfunctional elimination and to assess bladder function following treatment. The study included 46 children diagnosed with dysfunctional voiding and constipation who received polyethylene glycol (PEG) 3350 between January 2000 and July 2002 (35 girls, mean age 7.7 years, range 4.5 to 11.2 years and 11 boys, mean age 7.6 years, range 4.4 to 11.1 years). All children received PEG 3350 without electrolytes at a starting dose of 8 ounces of mixture each day with instructions to adjust the amount consumed by 1 to 2 ounces every 3 days to achieve the goal of one to two soft bowel movements per day. The final dose was normalised according to patient weight and the average final dose was 0.63 g/kg (as reported in abstract) or 0.59 g/kg (as reported in text). The average duration of treatment was 194.3 days (SD 133.5) and side effects were recorded. It is not clear how side effects were measured. Nine of 46 children (all female) reported having diarrhoea. Children with diarrhoea were significantly younger at the start of PEG therapy than children without diarrhoea (patients with diarrhoea (n=9): mean age 6.8 years ± 1.1 versus patients without diarrhoea (n=37): mean age 8.2 years ± 1.8; P = 0.04). The duration of follow-up was significantly longer for children with diarrhoea compared to children without diarrhoea (patients with diarrhoea (n=9): mean 336 days ± 153 versus patients without diarrhoea (n=37): mean 108 days ± 11; P = 0.0028). One child stopped taking PEG because of side effects.

A retrospective case series conducted in the USA81 (2004) [EL=3] evaluated the safety and efficacy of PEG 3350 without electrolytes for the treatment of constipation in children aged under 2 years. The study included 75 children with constipation aged less than 2 years at the start of PEG therapy (mean age 17 months, range 1 to 21 months). Children received PEG 3350 without electrolytes at a starting average dose of 1 g/kg body weight/day. Parents were asked to adjust the dose to yield one to two soft painless stools per day. Adverse effects were measured at 4 months or less (short term, mean 2 months) and 6 months or more (long term, mean 11 months). The average duration of treatment at the short-term assessment was mean 2.3 months ± 1.3, range 1 to 4 months and at the long-term assessment it was mean 10.6 months ± 8.1, range 6 to 37 months. It is not completely clear how side effects were measured, but it seems that parents were asked about them at the time of consultation. At 4 months or less, five children (7%) had experienced ‘runny stools’. The mean dose of PEG used was 1.1 g/kg body weight/day ± 1.2 (median 0.82, range 0.4 to 2.3). At 6 months or more, one child had experienced watery stools. The diarrhoea disappeared after lowering the dose of PEG. The mean dose of PEG used was 0.8 g/kg body weight/day ± 0.4 (median 0.67, range 0.3 to 2.1). Parents did not report increased flatus, abdominal distension, vomiting or new onset abdominal pain. None of the children stopped PEG because of adverse effects. Complete blood counts (in 24 children), electrolytes (in 9 children), renal functions (in 8 children) and liver functions (in 8 children) were occasionally done in children on long-term PEG treatment and all were within normal limits.

A retrospective case series conducted in the USA82 (2004) [EL=3] determined safety, efficacy and optimal dose of PEG powder for treatment of constipation in patients younger than 18 months. The case series included 28 children younger than 18 months treated for constipation with PEG powder. Children received PEG 3350**** at an initial dose of 0.88 g/kg/day (range 0.26–2.14 g/kg/day). After initial dose, families were asked to titrate the dose to obtain at least one non-formed bowel movement daily. Change in dose was permitted within 24 hours, if necessary. The mean duration of treatment was 6.2 months ± 5 (range 3 weeks to 21 months). Children were assessed at an initial visit and subsequent visits every 8 to 12 weeks. The duration of therapy and side effects were retrieved from the patient's chart and the information not available in the chart was obtained by telephone interview. It is not clear how side effects were measured in the first place. The mean effective maintenance dose was 0.78 g/kg/day (range 0.26–1.26 g/kg/day). Side effects were recorded in 17.9% of patients. One infant (3.6%) experienced increased passage of gas per rectum, whereas four infants (14.3%) experienced transient diarrhoea that resolved after dose adjustment.

A prospective case series conducted in the USA83 (2003) [EL=3] assessed the biochemical and clinical safety profile of long-term PEG 3350 treatment in a large cohort of children and also its acceptance by paediatric patients. The study included 83 children older than 2 years (48 males, 35 females, mean age 7.4 years, range 2.0 to 16.9 years) with chronic constipation who were treated daily with PEG for more than 3 months. For an average of 8.7 months (range 3–30 months) all children received PEG 3350 without electrolytes orally at an initial dose of 0.8 g/kg/day. Parents were asked to adjust dose of PEG solution as required to yield two soft painless stools per day. Over time, parents were instructed to gradually decrease the dose of PEG if the symptoms of constipation and encopresis showed improvement. Adverse effects, both clinical and laboratory variables, were assessed. Parents were interviewed using a structured questionnaire and asked about any possible adverse effects of PEG and particularly about excessively loose or frequent stools, abdominal pain, flatulence, bloating and nausea. Following interview and physical examination, 4 ml of blood was obtained for measurement of different parameters.

Clinical adverse effects were minor and over the mean duration of therapy. Eight patients (10%) experienced frequent watery stools some time during therapy, but diarrhoea disappeared with reduction of the dose. Five children (6%) experienced bloating or flatulence and two children (2%) abdominal pain. Different individual patients (1%) experienced each of the following: thirst, fatigue and nausea after receiving PEG solution on an empty stomach. General physical examination findings revealed no new significant abnormalities compared with the pre-treatment. None of the patients stopped treatment due to adverse effects and all were to continue PEG therapy.

Laboratory evaluation results (haemoglobin, haematocrit, serum electrolytes, blood urea nitrogen, serum creatinine, serum albumin and osmolality) were normal in all patients (10 patients did not have serum osmolality measured). Ten patients (11%) had slightly elevated alanine transaminase (ALT) level (less than 1.5 times the upper limit of normal, range 31 to 45 units per litre [U/L]). Eight of these patients had ALT levels re-measured within 8 weeks, seven of whom were still receiving PEG therapy. Seven of these eight patients had values in the reference range, one had a slightly elevated ALT level (less than 1.2 times normal, 28 U/L). Three patients (4%) had an elevated aspartate aminotransferase level (less than 1.5 times normal, range 42-52 U/L) and all had normal values when re-measured while still receiving PEG therapy. Both the dose and the duration of PEG therapy were not significantly different in patients with abnormal values compared with those with laboratory values in the reference range.

A prospective case series conducted in Australia84 (2007) [EL=3] evaluated the safety and efficacy of a PEG 3350-based preparation containing electrolytes in the treatment of chronic constipation in children. The study included 77 children with chronic constipation for at least 6 months, which was either untreated or inadequately treated by laxatives (44% boys, mean age 4.9 ± 2.6 years). Children received PEG 3350 plus electrolytes for an average of 75.5 days. Starting dose (number of sachets/day) during the first 5 days was established according to children's age (children aged 2 to 6 years: 1 sachet/day on days 1 and 2, 1 sachet twice a day on days 3 and 4, 1 sachet three times a day on day 5; children aged 7 to 11 years: 1 sachet twice a day on days 1 and 2, 2 sachets twice a day on days 3, 4 and 5). Thereafter, and until end of the study, the dosage was titrated according to the faecal form. This dose was increased by 1 sachet/day in the event of continued hard stools or no bowel movements, and decreased by 1 to 2 sachets/day in the event of loose stools or diarrhoea. Adverse effects were monitored throughout the study: blood samples for laboratory investigation were taken at baseline, 28 days and 84 days after initiating treatment. Vital signs were measured at baseline and 84 days after initiating treatment. It is not clear how other clinical adverse effects were collected.

The mean numbers of sachets/day during the treatment period was 1.3 (6.9 g). Seventy-two children (92%) reported a total of 318 adverse events. Two hundred and forty-one (76%) of those events were assessed as unrelated to the study treatment, 262 (82%) were considered mild and 302 (95%) had resolved by the end of the study. Six serious adverse events occurred in four children: four affected the gastrointestinal system while the other two were not clearly reported. All of them were assessed by the investigator as unrelated or unlikely to be related to the study medication and were resolved at the end of the study. One serious adverse event (faecal impaction) led to one patient's premature withdrawal from study as the child was admitted to hospital for bowel washout. No clinically significant changes in vital signs as a result of the study medication were observed.

A prospective case series conducted in Sweden85 (2005) [EL=3] assessed the effectiveness of PEG 3350+E over the course of long-term treatment in children with constipation. The case series included 134 children referred with constipation and/or encopresis (88 males, age not clearly reported). All children received PEG 3350+E (13.8 g sachets) at a mean starting dose of 0.58 sachets for children aged 2 to 6 years and 0.51 sachets for children aged 7 to 11 years. Doses were adjusted in each patient to achieve symptom relief with the minimally effective dosage. The mean duration of treatment was 50 weeks (SD ±50 weeks, range 1 to 211 weeks). The final treatment dose and side effects were recorded but it is unclear how this was done. The mean dose at the end of the observational period was 0.42 sachets for children aged 2 to 6 years and 0.49 sachets for children aged 7 to 11 years. The overall mean change was 0.553 to 0.477 sachets per day. Side effects were reported in 10 patients (7.5%) and these were reported as generally mild and transient.

An RCT conducted in the USA70 (2006) [EL=1-] compared the efficacy, safety and patient acceptance of PEG 3350 without added electrolytes versus magnesium oxide (milk of magnesia, MOM) over 12 months. The study included 79 children diagnosed with functional constipation with faecal incontinence (65 boys, age range 4 to 16.2 years, median age 7.4 years, mean 8.1 ± 3.0). Children were randomised to receive PEG 3350 without added electrolytes at 0.7 g/kg body weight daily for 12 months or MOM 2 ml/kg body weight daily for 12 months. If it was necessary children were disimpacted with one or two phosphate enemas in the clinic on the day of the visit and then started laxative therapy that evening. Outcomes were patients' acceptance and adherence. Patients and their parents were questioned with respect to side effects during each visit.

Several children complained about the taste of both PEG and MOM. Two children (5%) continued to refuse PEG versus 14 children (35%) who continued to refuse MOM during the 12 months of the study (P < 0.001). By 12 months 27 children (34%) had left the study or were lost to follow-up. In the PEG group, two children were lost to follow-up monitoring, two (5%) had refused PEG, one child was allergic to PEG and two children were receiving senna. These seven children were counted as not improved and not recovered. In the MOM group two children were lost to follow-up, three children had discontinued study participation, 14 children (35%) had refused to take MOM and one child was receiving senna. Mean treatment doses at 1 month were 0.7 ± 0.2 g/kg body weight for PEG and 1.2 ± 0.7 ml/kg body weight for MOM. At 3 months doses were 0.6 ± 0.3 g/kg body weight for PEG and 1.2 ± 0.8 ml/kg body weight for MOM. Mean treatment doses were similar in children who improved and those who did not improve for both treatments. There were no other significant clinical effects for either medication, apart from transient diarrhoea disappearing with dose reduction.

A retrospective cohort conducted in the USA86 (2003) [EL=2-] reported efficacy of PEG therapy, effective dose and patient compliance separately for children with constipation and children with constipation and encopresis over the long term. This included 74 children (40 boys) aged over 2 years with chronic constipation treated daily with PEG 3350 without electrolytes for more than 3 months. Children received PEG 3350 without electrolytes at a starting dose of 0.8 g/kg/day. Parents were asked to adjust the dose as required to yield two soft painless stools per day. The average duration of the treatment was 8.4 months (range 3 to 30 months) and adverse effects were assessed. Some outcomes variables on effectiveness were gathered by interviewing patients and/or parents and examining patients, but it is unclear how data on adverse effects were obtained. The average dose of PEG at the time of evaluation was 0.73 g/kg/day (range 0.3 to 1.8) following adjustment of dose by the carers. No major clinical adverse effects were observed.

A prospective case series conducted in the USA87 (1987) [EL=3] prospectively monitored children receiving large doses of mineral oil throughout the early phase of treatment. The study included 25 children with constipation, aged over 1 year with no previous treatment with mineral oil (mean age 7.83 years, range 1.75 to 14.27 years). Following initial disimpaction children received 45 ml mineral oil twice daily between meals for a period of 4 months. The dose was gradually decreased on a monthly basis (usually 30 ml/month) depending on the patient's reported performance and the results of serial rectal examinations (month 1: mean dose 4.0 ± 1.4 (SEM), month 2: mean dose 2.9 ± 1.2, month 3: mean dose 2.1 ± 0.5, month 4: mean dose 1.4 ± 0.4). Serum beta carotene levels, retinol levels and alfa tocopherol levels were measured at baseline and at the end of every treatment month.

Mean retinol levels at 1 and 2 months were not significantly different from baseline values. After 3 months levels significantly increased compared to baseline (baseline: 1.48 micromols/l ± 0.84 SEM (42.3 micrograms/dl ± 24.1), treatment: 2.22 micromols/l ± 0.77 (63.5 micrograms/dl ± 22.1); P < 0.01) but changes were not significant after 4 months. Mean serum beta carotene levels decreased significantly at 1 month, 2 months and 3 months when compared to baseline, but there were no significant differences after 4 months:

  • month 1 (n=25): baseline: 1.0 micromols/l ± 0.5 SEM, (55.7 micrograms/dl ± 26.0) versus treatment: 0.7 micromols/l ± 0.4, (35.9 micrograms/dl ± 22.1); P < 0.01
  • month 2 (n=17): baseline: 1.1 micromols/l ± 0.6, (59.5 micrograms/dl ± 30.6) versus treatment: 0.7 micromols/l ± 0.5, (38.2 micrograms/dl ± 28.4); P < 0.05
  • month 3 (n=10): baseline: 1.1 micromols/l ± 0.6 (60.4 micrograms/dl ± 30.0), treatment: 0.6 micromols/l ± 0.2, (34.7 micrograms/dl ± 12.3); P < 0.05.

Serum alfa tocopherol levels remained relatively unchanged throughout the study. No statistical significant difference was found between baseline levels and those obtained throughout the 4 months of therapy.

An RCT (crossover) conducted in the UK71 (1977) [EL=1-] compared effectiveness and side effects of a standardised senna syrup with lactulose in the treatment of childhood constipation. The study included 21 children aged under 15 years with a history of constipation treated at home for 3 months or more. Children were randomised to receive either senna syrup (10 to 20 ml daily) for 2 weeks or lactulose (10 to 15 ml daily) for 2 weeks with 1 intermediate week with no treatment. Each preparation was given throughout the appropriate treatment week in a daily dose varied according to the age of the patient. Outcome measures were stool consistency, number of stools passed each day and adverse effects. These outcomes were recorded by parents in written diaries.

There were significantly more adverse effects (number of patients) during the senna week (12 including eight colic, one diarrhoea, two colic plus diarrhoea, one colic plus distension) compared to the lactulose week (one colic) (P < 0.001). There were no significant differences between the week with no treatment (four including three colic and one colic plus distension) compared to the lactulose week (one colic). One patient on senna at the beginning of study failed to attend at the end of the first week assessment but was included in the analysis.

An RCT conducted in Iran72 (2007) [EL=1-] compared the clinical efficacy and safety of liquid paraffin and lactulose in the treatment of functional childhood constipation. The study included 247 children with chronic functional constipation aged 2 to 12 years (mean age 4.1± 2.1 years). All children received one or two enemas daily for 2 days to clear any rectal impaction (30 cc/10 kg of paraffin oil). Children were randomised to receive either liquid paraffin orally (n=127) 1 to 2 ml/kg twice daily for 8 weeks or lactulose orally (n=120), 1 to 2 ml/kg twice daily for 8 weeks. For determination of the best dose for each child, parents were asked to increase the volume of each drug by 25% every 3 days as required to yield one or two firm to loose stools. Outcome measures were optimal dose of drug and side effects. Parents received a chart to record side effects.

The final mean effective dose was significantly larger in children taking lactulose compared to children taking liquid paraffin (2.08 ml/kg/day ± 0.21 versus 1.72 ml/kg/day ± 0.13; P < 0.001). Apart from nausea and hard stool, side effects during weeks 4 to 12 were more frequent***** in children taking liquid paraffin compared to children taking lactulose: abdominal pain (50 versus 10), bad palatability (40 versus 15), pain at defecation (50 versus 10), bloating (20 versus 10), diarrhoea (30 versus 10), anal oil leakage (40 versus 20), flatulence (20 versus 10), nausea (5 versus 10) and hard stool (6 versus 20). No children in either group experienced vomiting.

An RCT conducted in Turkey73 (2005) [EL=1-] compared the efficacy, safety and optimal dose of liquid paraffin and lactulose in children with chronic functional constipation. The study included 40 children aged 2 to 12 years referred for evaluation of constipation with evidence of faecal impaction (22 male, mean age 3.7 years ± 2.7). Children were randomised to receive either liquid paraffin or lactulose for 8 weeks. The medication was administered orally as a suspension at 1 ml/kg twice daily for each drug. To determine the best dose for each child, parents were asked to increase or decrease the volume of each drug by 25% every 3 days as required to yield two firm to loose stools per day. The maximum dose used throughout the study was 3 ml/kg per day for each drug. Outcomes measured at 4 and 8 weeks after initiation of treatment were: optimal dose of drugs, compliance rate and side effects. Patients were instructed to take both empty and full containers to calculate the amount of medication taken. It is unclear how side effects were recorded.

The optimal dose of drugs was not significantly different for children taking liquid paraffin compared to children taking lactulose (mean 1.88 ml/kg/day ± 0.27 versus mean 2.08 ml/kg/day ± 0.27). These data were reported in a table and it was assumed that figures given were for the whole study period. Data reported in text for the last 4 weeks of treatment stated the optimal dose for liquid paraffin as 1.72 ml/kg/day ± 0.18 and for lactulose as 1.82 ml/kg/day ± 0.57. Adherence rate during the first 4 weeks of treatment was not significantly different when comparing both groups. During the last 4 weeks of therapy significantly more children complied with taking liquid paraffin than the children taking lactulose (n=90 versus n=60; P = 0.02). No patient stopped treatment because of adverse effects (adverse effects not reported). During the first 4 weeks, taste aversion was reported in one child on liquid paraffin and abdominal distension in two patients on lactulose influenced adherence. During the last 4 weeks, adherence was influenced by poor symptom control in five patients, side effects (abdominal distension and cramping) in three children on lactulose and watery stools in two children on liquid paraffin.

Evidence statement

There is evidence showing that adverse effects of using oral preparations of osmotic laxatives, stimulant laxatives and faecal softeners in the medium to long term are generally infrequent and mild.

Adverse effects up to 6 months of treatment

Stimulant laxatives

One RCT [EL=1-] showed that senna produced colic, diarrhoea and abdominal distension in 52%, 9.5% and 4.8 % of the children respectively.

Osmotic laxatives

PEG 3350 without electrolytes was found to produce runny stools in 7% of the children (one retrospective case series, EL=3). PEG 4000 without electrolytes was found to produce diarrhoea, flatulence and vomiting (one RCT [EL=1+], figures unclear). PEG 3350 + electrolytes was found to produce gastrointestinal effects (mostly abdominal pain) in 39% of children (one RCT [EL=1+]).

Lactulose was found to produce most commonly diarrhoea (two RCTs: 10% and figures unclear, respectively) and flatulence (two RCTs [EL=1+ and EL=1-]: 10% and figures not reported, respectively). One RCT [EL=1+] reported low incidence of anal irritation (5%) and anorexia (1%). One RCT [EL=1-] reported colic (4.8%). One RCT [EL=1-] reported abdominal pain (10%), bad palatability (15%), pain at defecation (10%), bloating (10%), anal oil leakage (20%), nausea (10%) and hard stool (20%). Lactulose was not found to produce vomiting (one RCT [EL=1-]).

One RCT [EL=1+], found that vomiting episodes and flatulence (either new onset or worsened in both cases) lasted significantly longer in children on lactulose compared to children on PEG 4000 without electrolytes. This study also found that body height and body weight were unaffected in children taking either treatment, for both boys and girls. Mean weight was also unaffected after treatment with PEG 3350 + electrolytes in another RCT [EL=1+].

One RCT [EL=1-] showed that at the end of 8 weeks significantly more children complied with taking liquid paraffin than with taking lactulose. No patient stopped treatment because of adverse effects but during the first 4 weeks abdominal distension in two patients on lactulose influenced adherence. During the last 4 weeks abdominal distension and cramping in three children on lactulose influenced adherence.

Faecal softeners

One prospective case series [EL=3] showed that therapy with mineral oil did not significantly change the serum levels of alpha tocopherol, retinol and beta carotene after 4 months.

One RCT [EL=1-] showed that liquid paraffin produced taste aversion (2.5%) and watery stools (0.5%). Another RCT [EL=1-] showed that liquid paraffin produced abdominal pain (50%), bad palatability (40%), pain at defecation (50%), bloating (20%), diarrhoea (30%), anal oil leakage (40%), flatulence (20%), nausea (5%) and hard stool (6%). Liquid paraffin was not found to produce vomiting.

One RCT [EL=1-] showed that at the end of 8 weeks significantly more children complied with taking liquid paraffin than with taking lactulose. No patient stopped treatment because of adverse effects but during the first 4 weeks taste aversion in one child on liquid paraffin influenced adherence. During the last 4 weeks watery stools in two children on liquid paraffin influenced adherence.

Adverse effects at between 6 and 12 months of treatment

Stimulant laxatives

No evidence was found of adverse effects of the use of stimulant laxatives for between 6 and 12 months of treatment.

Osmotic laxatives

PEG 3350 without electrolytes was found to produced watery stools (one retrospective case series [EL=3]), diarrhoea in 19.5% (one retrospective case series [EL=3]), increased passage of gas per rectum (3.6%, one retrospective case series [EL=3]) and transient diarrhoea that resolved after dose adjustment in 14.3% and 10% of children (one retrospective case series [EL=3], one prospective case series [EL=3], respectively), bloating or flatulence (6%, one prospective case series [EL=3]) and abdominal pain (2%, one prospective case series [EL=3]). One retrospective case series [EL=3] showed a significant association between diarrhoea while taking PEG 3350 without electrolytes and younger age and also with longer follow-up. One child in the series (2.2%) stopped taking PEG 3350 without electrolytes because of side effects.

One retrospective case series [EL=3] showed that parents did not report increased flatus, abdominal distension, and vomiting or new onset abdominal pain while children were taking PEG 3350 without electrolytes. None of the children stopped treatment because of adverse effects. One retrospective cohort [EL=2-] found no major clinical adverse effects in children taking PEG 3350 without electrolytes.

A prospective case series [EL=3] showed that general physical examination findings revealed no new significant abnormalities compared with the pre-treatment in children treated with PEG 3350 without electrolytes. None of the patients stopped treatment due to adverse effects and all were to continue PEG therapy.

A prospective case series [EL=3] found that 24% of adverse events occurred in children taking PEG 3350 with electrolytes but they were considered mild and had resolved by the end of the study. No clinically significant changes in vital signs as a result of the study medication were observed.

Adverse effects at/after 12 months of treatment

Osmotic laxatives

PEG 3350 without electrolytes was found to produce diarrhoea (one prospective cohort [EL=2+], figures not reported) and transient diarrhoea disappearing with dose reduction (one RCT [EL=1-], figures not reported).

One RCT [EL=1-] found that several children complained about the taste of both PEG 3350 without electrolytes and magnesium oxide (milk of magnesia, MOM) but significantly more children continued to refuse MOM compared to PEG during the 12 months of the study.

One prospective cohort [EL=2+] found that none of the children on PEG 3350 without electrolytes became dehydrated. Children receiving PEG 3350 without electrolytes and their parents did not report increased flatus, abdominal distension or new onset of abdominal pain. No children reported disliking the taste of PEG and no parents reported that their child refused to take it in juice or Kool-Aid whereas 33% of children refused to take MOM.

One prospective case series [EL=3] found that side effects of PEG 3350 with electrolytes were reported in ten (7.5%) patients and that these were generally mild and transient.

Stimulant laxatives

No evidence was found of adverse effects of the use of stimulant laxatives for 12 months of treatment or longer.

Bulk forming laxatives

No evidence was found of adverse effects of the medium- to long-term use of bulk forming laxatives.

Tables 5.1, 5.2 and 5.3 summarise the results of these studies.

Table 5.1. Adverse effects up to 6 months of treatment.

Table 5.1

Adverse effects up to 6 months of treatment.

Table 5.2. Adverse effects between 6 and 12 months of treatment.

Table 5.2

Adverse effects between 6 and 12 months of treatment.

Table 5.3. Adverse effects at or after 12 months of treatment.

Table 5.3

Adverse effects at or after 12 months of treatment.

GDG interpretation of the evidence

There is no evidence to support the commonly held belief that using laxatives produces a ‘lazy bowel’. Some healthcare professionals still hold this misconception and communicate it to parents.

Consequences of the medication, such as abdominal pain and increased soiling, can be clinically similar to the symptoms of constipation and are usually dose related. These symptoms are more likely to occur with higher dosage but this may be mitigated by the effective outcome of the medicine. The GDG believes that parents need information and support in order to know what to expect when using different laxatives to support optimal recommended treatment.

From the evidence and also their own clinical experience the GDG noted that palatability is an important aspect as children will not take the medication if they do not like it, despite its potential effectiveness. The GDG noted that the consultation with children confirmed that taste and the way that medicines are given is important to them.

Recommendations

Start maintenance therapy as soon as the child or young person's bowel is disimpacted.

Reassess children frequently during maintenance treatment to ensure they do not become reimpacted and assess issues in maintaining treatment such as taking medicine and toileting. Tailor the frequency of assessment to the individual needs of the child and their families (this could range from daily contact to contact every few weeks). Where possible, reassessment should be provided by the same person/team. Offer the following regimen for ongoing treatment or maintenance therapy:

  • Polyethylene glycol 3350 + electrolytes as the first-line treatment.*
  • Adjust the dose of polyethylene glycol 3350 + electrolytes according to symptoms and response. As a guide for children and young people who have had disimpaction the starting maintenance dose might be half the disimpaction dose (see table 4).
  • Add a stimulant laxative (see table 4) if polyethylene glycol 3350 + electrolytes does not work.
  • Substitute a stimulant laxative if polyethylene glycol 3350 + electrolytes is not tolerated by the child or young person. Add another laxative such as lactulose or docusate (see table 4) if stools are hard.
  • Continue medication at maintenance dose for several weeks after regular bowel habit is established – this may take several months. Children who are toilet training should remain on laxatives until toilet training is well established. Do not stop medication abruptly: gradually reduce the dose over a period of months in response to stool consistency and frequency. Some children and young people may require laxative therapy for several years. A minority may require ongoing laxative therapy.
*

At the time of publication (May, 2010), Movicol Paediatric Plain is the only macrogol licensed for children under 12 years that includes electrolytes. It does not have UK marketing authorisation for use in faecal impaction in children under 5 years, or for chronic constipation in children under 2 years. Informed consent should be obtained and documented. Movicol Paediatric Plain is the only macrogol licensed for children under 12 years that is also unflavoured.

Research recommendations

What is the effectiveness of polyethylene glycol 3350 + electrolytes in treating idiopathic constipation in children younger than 1 year old, and what is the optimum dosage?

Why this is important

There is some evidence that treatment of constipation is less effective if faecal impaction is not dealt with first. Disimpaction with oral macrogols is recommended for children and their use avoids the need for rectal treatments.

Rectal treatments, especially in hospital, are more common than oral treatments at home. Although relatively few infants are admitted to hospital, there would be savings if initially all children were disimpacted at home.

Polyethylene glycol 3350 + electrolytes, an oral macrogol, is licensed for disimpaction in children older than 5 years. Increasing experience has shown that it is effective in infants younger than 1 year old, but evidence is limited to small case series. If dosage guidelines and evidence on macrogol use in infants were obtained and published, more healthcare professionals might be encouraged to try macrogols in this age group. It would also allow the guideline to be applicable across the whole paediatric age group.

What is the effectiveness of polyethylene glycol 3350 + electrolytes as compared to stimulant laxatives (senna, bisacodyl and sodium picosulfate) in treating idiopathic constipation in children older than 2 years?

Why this is important

Clinical experience and available evidence support the use of oral polyethylene glycol 3350 + electrolytes as first line treatment for both disimpaction and maintenance in children with idiopathic constipation. Economic modelling also indicates that where any treatment is effective it is also cost-effective and that the optimal choice of treatment therefore appears to be the one likely to be of most therapeutic benefit. The optimal choice of treatment depends both on the clinical efficacy of treatment and its acceptability, that is, the likelihood that a child will adhere to treatment both in the initial disimpaction phase and over time.

However research evidence is limited and evidence is not available for the full range of medications used in clinical practice to treat idiopathic constipation in children. Many drugs have been used for a long time both for disimpaction and ongoing maintenance, but have not been tested in clinical trials with children and young people.

Currently there is no evidence on the effectiveness of stimulant laxatives to treat faecal impaction in children and the evidence available for ongoing maintenance is limited and of low methodological quality. However, clinical experience shows that they are commonly used singly or in combination, are effective and low cost. At the moment and due to the lack of research evidence stimulant laxatives can only be recommended as a second-line intervention.

Trials are needed to assess the effectiveness, optimum doses, acceptability and side effects of PEG 3350 + electrolytes as compared to stimulant laxatives (senna, bisacodyl and sodium picosulfate) in disimpaction and PEG 3350 + electrolytes as compared to senna in ongoing maintenance in children older than 2 years with idiopathic constipation.

Trials should include generic health related quality of life outcomes and not only the measurement of symptoms that are highly specific to constipation so that quality adjusted life years can be calculated. If these outcomes are included in a trial then credible cost-effectiveness analysis can be undertaken that will be useful to guide decision-makers in future clinical guidelines

5.4. Diet and lifestyle

Introduction

Acute simple constipation can usually be treated with a high fibre diet and sufficient fluid intake. In chronic idiopathic constipation, diet and lifestyle interventions remain important but should be carried out in conjunction with laxative therapy and behavioural modifications.

There seems to be uncertainty among health professionals about which aspects of the diet should be modified to help improve constipation. As a result, advice to families varies considerably. There is sometimes the belief that a child's chronic constipation has been caused by a lack of fibre or fluids in the diet, when in fact this is often not the case. It is a common problem that treatment with laxatives is delayed while a number of dietary and lifestyle adjustments are made. This can worsen the constipation and make families reluctant to make any diet and lifestyle changes in the future as initial efforts have failed.

There is guidance from the Department of Health for active living throughout the lifecourse:

‘Children and young people should achieve a total of at least 60 minutes of at least moderate intensity physical activity each day. At least twice a week this should include activities to improve bone health (activities that produce high physical stresses on the bones), muscle strength and flexibility.’

(Summary, page 3)88

There is guidance from the Department of Health on goats' milk infant formula:

‘The Department of Health does not recommend the use of milk based on goats’ milk protein for infants (under 1 year of age). The composition of infant formula and follow-on formula is governed by European legislation. The current legislation specifically states the criteria for infant formulas and follow-on formulas to be based on cows' milk protein, hydrolysed protein or soya protein. The Department recommends the use of infant formula and follow-on formula based on cows' milk protein or hydrolysed protein or soya protein on the advice of health professionals. In light of the European Food Safety Authority (EFSA) opinion, the Department advises health professionals not to recommend the use of infant milks based on goats' milk protein. Some parents may believe that infant milk based on goats' milk protein is a suitable alternative for babies who they perceive as being intolerant or allergic to cows' milk formula. However, the protein in goats' milk is very similar to that found in cows' milk and most babies who react to cows' milk protein will also react to goats' milk protein. Goats' milk protein can induce allergic reactions and is not a suitable milk source for a cows' milk allergic infant as there is the potential for cross allergenicity. Infants with proven cows' milk protein intolerance can be prescribed an extensively hydrolysed infant formula. Formula derived from goats' milk is also unsuitable for babies who are lactose intolerant as it contains similar levels of lactose to cows' milk based infant formulas.89

Clinical question

What is the clinical effectiveness of the following for ongoing treatment or maintenance in children with chronic idiopathic constipation?

  • increasing physical activity
  • dietary modifications
  • increasing fluid intake
  • excluding cows' and goats' milk protein from diet.

Studies considered in this section

Studies were considered if they:

  • included neonates, infants, or children up to their 18th birthday with chronic idiopathic constipation
  • included the following diet and lifestyle modifications:

    excluding cows' and goats' milk from the diet

    increasing fluid intake,

    increasing physical activity

    infant's formulas, prebiotics, omega 3 fish oils, chocolate, low fat or high fat diet, dairy free diet, soy milk and sheep's milk, increasing fibre intake (fibre rich food and fibre supplementing)

  • included the following outcomes:

    changes in frequency of bowel movements

    changes in stools consistency or appearance

    changes in pain or difficulty on passing stools

    changes in frequency of episodes of soiling

    reduction in laxatives use

    parent/child views or satisfaction or quality of life

  • were not case reports
  • were published in English.

No restrictions were applied on the publication date or country

Overview of available evidence

A total of 1022 articles were identified from the searches (154 from a search on excluding cows' and goats' milk from the diet and 868 from a search on the remainder diet and lifestyle modifications). Fifty-nine articles were retrieved for detailed assessment. Of these, 20 studies were identified for inclusion in this review: one triple-blind RCT, six double-blind RCTs, three open-label RCTs, two open label non-RCTs and eight prospective case series (two with an embedded food tolerance challenge test).

Narrative summary

Infant formulae

One double-blind RCT (crossover) conducted in The Netherlands90 (2007) [EL=1+] tested the hypothesis that Nutrilon Omneo (new formula, NF) would have a positive effect on stool characteristics in constipated children. The study included 38 otherwise healthy, term infants with constipation aged 3 to 20 weeks who received at least two bottles of milk-based formula per day (19 boys, median age 1.7 months). Infants were randomised to receive either NF or standard formula (SF) in period 1 and crossed over after 3 weeks to treatment period 2. Each treatment period lasted 3 weeks. Feeding patterns were not described. NF composition (per 100 ml) differed from the SF in that its protein content was higher, 100% of it was based on whey protein hydrolysate (no casein, no intact whey protein) and it contained a mixture of prebiotic oligosaccharides (galacto-oligosaccharide [GOS] and fructo-oligosaccharide [lcFOS]), a higher concentration of sn-2 palmitic acid and a lower lactose content.

Defecation frequency, improvement of hard to soft stools and number of children experiencing no painful defecation were not significantly different between the two treatment groups after period 1. After the crossover, painful defecation and defecation frequency were not significantly different between the periods on NF and SF. Seventeen percent (n=4) of infants had soft stools when receiving NF but hard stools with SF, compared to no infant with soft stools when receiving SF and no infant with hard stools when receiving NF (P = 0.046).Throughout the study there were no serious adverse effects in either group. Both formulae were well tolerated.

Only 24 children (63%) completed the crossover study. In period 1, three patients on SF dropped out: two patients stopped because of severe constipation and one patient switched to hypoallergenic feeding, because of suspected cows' milk protein allergy. Parents of one patient decided that they did not want to cross over because she was free of symptoms and they started openly with NF instead. Three patients dropped out after switching to NF: two patients stopped after less than one week because of recurrence of constipation symptoms and one patient was lost to follow-up. Seven patients dropped out after switching to SF: six patients stopped after one week because of recurrence of constipation symptoms and one patient was lost to follow-up. Data analysis was based on the group of 35 patients that completed period 1 and the subgroup analysis of 24 patients who completed the crossover. There were no significant differences in baseline characteristics between the two groups.

One prospective case series conducted in Italy91 (2003) [EL=3] investigated whether a new infant formula commercially available in Italy was useful as a dietary option in infants with minor feeding problems. The study included 604 formula-fed healthy term infants aged up to 3 months seen by paediatrician because of colic and/or constipation and/or regurgitation (age at entry of the total sample was 1.35 months ± 0.77, gender not reported). Of these, 232 infants were diagnosed with constipation, defined as a stool frequency of less than one stool a day. During 14 days all infants received a new formula (NF)†††††. The feeding volume was based on a feeding on demand procedure. The feeding frequency was decided by the parents and not influenced by the study protocol.

The study found that 147 infants (63.4%) reported an increase in the number of stools per day during the study period compared to baseline, with a significant average increase of 0.42 (95% CI 0.55 to 0.27; P < 0.005). The average increase between day 1 and day 7 was 0.41 (95% CI 0.51 to 0.23; P < 0.05) and between day 7 and day 14 it was 0.04 (NS). There was no improvement of symptoms in 85 infants (26.6%). Mean parent evaluation of formula (on a score of 1 to 10) was 7.9 ± 1.8. A total of 550 parents (91%) gave a positive judgement (score 6 to 10). A total study population of 932 infants were enrolled and 604 (65%) completed the study protocol. A total of 358 infants were excluded from the study: 154 completed only the first step and did not return for the visit on day 14 while 131 infants were excluded because of incomplete data. Seventy-three infants required medication during the first week of study and were therefore excluded. The proportion of these infants who had constipation was not reported in the paper. It should be noted that stool consistency was not assessed in the study.

A prospective case series conducted in Spain92 (2008) [EL=3] assessed the prevalence of mild gastrointestinal disorders (MGDs) in infants fed with artificial milk formulas in paediatric practice and evaluated the effectiveness and satisfaction with dietetic treatment, specifically elaborated formulas belonging to the Novalac® line of products (United Pharmaceuticals). The study included 3487 infants with MGDs and fed with artificial milk formulae (52.2% boys, aged 1 to 17 weeks). Of these, 604 infants had constipation. For 30 days constipated children received Novalac Anti-Constipation®, a formula with an adapted concentration of magnesium and lactose. No other details regarding feeding volume or frequency were provided.

In total, 91.6% of cases of constipation resolved within 7 days, but this was not clearly defined in the paper. The number of daily stools increased significantly at the end of the study when compared to baseline (baseline: mean 0.6 ± 0.7 versus at 30 days: mean 1.7 ± 0.8). The percentage of children having normal stools increased significantly at the end of the study when compared to baseline (baseline: 33.40% versus at 30 days: 95.60%). The percentage of children presenting with pain or discomfort on defecation was significantly reduced at the end of the study when compared to baseline (baseline: 90.0% versus at 30 days: 10.4%). The percentage of children needing external help at defecation was significantly reduced at the end of the study when compared to baseline (baseline: 76.1% versus at 30 days: 8.8%). Ninety percent of parents reported being satisfied with the treatment. Adverse events (for all formulae, no subgroup analysis) were reported in 3.9% infants of the total population. Effectiveness was evaluated among 1441 infants (total population) who completed follow-up. Premature study termination due to adverse events occurred in 2.7% of cases, parent decision in 6.9%, loss to follow-up in 1.64%, protocol violations in 2.46% and non-specified reasons in 16.62%.

One open label RCT conducted in Taiwan93 (2007) [EL=1-] evaluated a commercialised formula, Novalac-IT® (Intestinal Transit), against a ‘strengthened regular formula’, the traditional approach in infants with digestive problems in Taiwan. The study included 93 children aged 2 to 6 months referred to the paediatric gastroenterology clinic at a medical centre with constipation for 2 weeks or more and fed exclusively with formula milk (47 boys, mean age 3.8 months ± 1.7). Children were randomised to receive either a magnesium-enriched infant formula, Novalac-IT or a 20% strengthened infant formula for 2 months. Children were assessed at 2 weeks, 1 month and 2 months.

Outcomes measured were remission, improvement or failure according to a severity scoring system based on stool consistency, frequency and volume of stools and difficulties in defecation (1 to 3 mild constipation; 4 to 6 moderate; 7 or 8 severe). Asymptomatic children were considered in remission, a decrease in severity of 4 or more was considered a good response and a decrease in severity of 1 to 3 a fair response. If the score did not change or increased it was considered treatment failure. The severity scoring system comprised the following variables:

  • stool consistency (hard stool 0, no hard stool 1, hard and long form 2)
  • difficulties with defecation (no difficulties 0, irritability 1, crying 2)
  • frequency of defecation (more than 3 times per week 0, 1 to 3 times per week 1, less than once per week 2)
  • stool weight (more than 35 g/kg/week 1, 20 to 35 g/kg/week 2, less than 20 g/kg/week 3).

The number of children who improved was not significantly different in the two groups at 2 weeks. At 1 month significantly more children on Novalac-IT had improved compared to children on the strengthened formula (39 out of 47 [83%] versus 23 out of 46 [50%]; P = 0.002). At 2 months significantly more children on Novalac-IT had improved compared to children on the strengthened formula (42 [89%] versus 25 [54%]; P < 0.001). The number of children free of symptoms at 2 weeks was not significantly different between the treatment groups. However, both at 1 month and at 2 months, significantly more children on Novalac-IT were free of symptoms compared to children on the strengthened formula (at 1 month: Novalac-IT: 28 out of 47 [60%] versus strengthened formula: 16 out of 46 [35%]; P = 0.029; at 2 months: Novalac-IT: 35 out of 47 [75%] versus strengthened formula: 18 out of 46 [39%]; P < 0.001). There were no significant differences in the baseline characteristics (clinical or demographic) between the two groups. It should be noted that participation in the trial was proposed before a more complete diagnostic workup for cows' milk protein allergy, Hirschsprung's disease and others was conducted.

One open label RCT conducted in Italy94 (2005) [EL=1-] evaluated the efficacy on digestive problems of a formula based on palmitic acid predominantly esterified at the β-position, oligosaccharides (GOS and FOS) with a prebiotic activity, partially hydrolysed protein, low lactose content and higher density. The study included 95 formula-fed healthy term infants aged 4 months or less with constipation, defined as a stool frequency of less than one stool a day (64.2% with hard stools) (50 boys, age at study entry in the intervention group was 1.55 months ± 0.88 and in the control group was 1.28 months ± 0.66). Children were randomised to receive either the new formula (NF) (Omneo®/Conformil®, Numico) or a standard formula (SF) for 14 days. The feeding volume was based on a feeding on demand procedure. Feeding frequency was decided by the parents and not influenced by the study protocol.

The stool frequency increased significantly more in children receiving NF compared to children receiving SF, both on day 7 and on day 14 (number/day, mean ± SD) (day 7: NF group (n=55): 1.79 ± 0.96 versus SF group (40): 1.31 ± 0.89; difference: 0.48 (95% CI: 0.09 to 0.87); P = 0.02]); (day 14: NF group (n=55): 2.04 ± 1.04 versus SF group (40): 1.64 ± 0.99, difference: 0.40 (95% CI: -0.03 to 0.83); P = 0.07). The stool frequency (number/day, mean ± SD) also increased significantly more in children receiving NF compared to children receiving SF, after adjusting for gender, age at entry, maternal instruction, parity, birth weight, number of feedings per day and stool frequency at entry (mean adjusted difference in stool frequency between the two groups for days 0 to 7 was 0.60 [CI 95% 0.19 to 1.01; P = 0.004] and for days 0 to 14 was 0.53 [95% CI 0.11 to 0.90; P = 0.015]). Post-treatment outcomes for stool consistency were not reported. There were no significant differences in the baseline characteristics between the two groups. No dropouts or children lost to follow-up were reported.

Increasing fibre

One double-blind RCT conducted in the Netherlands95 (2008) [EL=1+] assessed the clinical efficacy and safety of a dietary fibre mixture and compared it with lactulose in the treatment of childhood constipation. The study included 135 children referred to hospital outpatient clinic for idiopathic constipation. Children were randomised to receive either a yogurt drink (125 ml) with 10 g of mixed dietary fibre (fibre mixture contained per 100 ml of solution: 3.0 g transgalacto-oligosacharides, 3.0 g inulin, 1.6 g soy fibre, 0.33 g resistant starch) or a yogurt drink containing lactulose (10 g/125 ml, Duphalac Lactulose®, Solvay Healthcare Limited). Forty-two children received yogurt with the fibre mix (20 boys, median age 5.5 years, range 1 to 12 years), whereas 55 children received the yogurt containing lactulose (23 boys, median age 5.0 years, range 1 to 12 years). Both products were taken at breakfast and when two or more bottles were needed they were also taken at lunch. The daily amount of fibre/fluid intake depended on the patient's body weight. If persistent diarrhoea was reported, the original dose was reduced by 50%. If clinical parameters compared to baseline did not improve 3 weeks after the start of intervention period, step-up medication (polyethylene glycol (PEG) 3350) was given per protocol. There was an intervention period lasting 8 weeks and a weaning period lasting 4 weeks when doses where reduced.

Defecation frequency per week and number of patients with one or more faecal incontinence episodes per week at 8 weeks was not significantly different between the two groups. Stool consistency (mean) was significantly softer in the lactulose group compared to the fibre group, both at 3 and at 8 weeks (at 3 weeks: fibre (n=42) 3.5, lactulose (n=55) 4.5; P < 0.01 and at 8 weeks: fibre 3.6, lactulose 4.0; P = 0.01). The number of patients using step-up medication at 3 weeks was significantly smaller in the group taking fibre than in the group taking lactulose (fibre: 13, lactulose: 7; P = 0.028) but there were no significant differences regarding this outcome at 8 and at 12 weeks. No serious or significant side effects were recorded. In the fibre group one child experienced dose-related persistent diarrhoea compared to two children in the lactulose group. No significant differences were found in baseline characteristics between the two groups. Thirty-three patients left the study: 22 in the fibre group after 1 to 56 days (median 7) and 11 in the lactulose group after 1 to 51 days (median 8) (P = 0.020). All those patients refused to drink the yogurt. Three patients were lost to follow-up: one on fibre and two on lactulose. Despite the high drop-out rate (24.4%) intention-to-treat analysis was not performed.

One double-blind RCT (crossover) conducted in the USA and Italy96 (2004) [EL=1+] evaluated whether fibre supplementation with glucomannan is beneficial in the treatment of children with idiopathic constipation. The study included 31 otherwise healthy children ( ) older than 4 years who had chronic idiopathic constipation for 6 months or longer with or without encopresis (16 boys, age range 4.5 to 11.7 years, mean age 7.1 years ± 2.0). Disimpaction was carried out with one or two phosphate enemas if rectal impaction felt during rectal examination. Fifty-eight percent of patients continued with their pre-evaluation laxative during the whole study period. Children were randomised to receive Glucomannan B (one capsule containing glucomannan, a polysaccharide of d-glucose and d-mannose, equal to 450 mg of alimentary fibre) or Glucomannan A (one capsule containing maltodextrins as placebo). After 4 weeks children were switched to the other treatment for another 4 weeks, with no washout period in between. Both glucomannan and placebo were given at a dose of 100 mg/kg body weight daily (maximal 5 g per day), rounded to the nearest 500 mg because each capsule contained 500 mg. Each capsule was either: opened and sprinkled on food and given with 50 ml of fluid per capsule; given as a solution, whereby the content of each 500 mg capsule was mixed with 50 ml of fluid of the child's choice; or swallowed as a capsule with 50 ml of fluid for each capsule. In addition, parents were instructed to have the child sit on the toilet four times daily after meals and to keep a stool diary.

No enemas were given during each treatment period unless rectal disimpaction felt during rectal examination at assessment visits. Successful treatment was rated by physician and defined as 3 or more bowel movements per week and 1 or no soiling episode in the last 3 weeks with no abdominal pain. Parents' global assessments related to whether they believed that the child was better during the first or second treatment period.

Stool consistency and frequency of soiling episodes per week were not significantly different when comparing the fibre treatment period with the placebo period. However, significantly more children on placebo reported having less than 3 bowel movements per week compared to children on fibre (placebo (n=31): 52% versus fibre (n= 31): 19%; P < 0.05). Significantly more physicians rated the fibre treatment as ‘successful’ when compared to placebo (45% versus 13%; P < 0.05). Significantly more parents in the fibre period rated their children as ‘improved’ when compared to parents in the placebo period (68% versus 13%; P < 0.05). Successful treatment (physician rating) and improvement (parent rating) were independent of low or acceptable fibre intake (P > 0.6). Significantly more children who were taking laxatives at enrolment were treated successfully with fibre than with placebo (P < 0.01). Children with constipation only were significantly more likely to be treated successfully with fibre than those with constipation and encopresis (69% versus 28%; P < 0.04). No significant side effects, such as new onset of abdominal pain, bloating, abdominal distension, excessive gas, diarrhoea or anaphylactic symptoms, were reported. No significant differences in baseline characteristics between the two groups were observed.

Forty-six children were originally recruited. Thirteen children did not attend their appointment: seven children randomized to placebo first and six children randomized to fibre first. Two constipated girls only completed the first 4 weeks of the study: one received placebo and one received fibre and both recovered from chronic constipation and abdominal pain during the first four weeks of treatment and did not return for the 8-week visit. Data from the 13 children who entered the study and were randomised but did not come for follow-up and the two children who did not complete the study were excluded from the analysis. Initial data from these 15 children were not significantly different from the data of the 31 children who completed the study, except soiling frequency per week which was significantly less (4.0 ± 1.4; P < 0.001). Data analysis thus includes 31 children with idiopathic constipation with or without encopresis. Despite the high attrition rate (28%) intention-to-treat analysis was not performed.

One double-blind RCT (pilot study) conducted in Spain97 (2006) [EL=1+] evaluated the effect of a palatable cocoa husk supplement that is rich in fibre on intestinal transit time and other indices of constipation in children with idiopathic chronic constipation. The study included 56 children aged 3 to 10 years (22 boys, mean age 6.3 years ± 2.2) referred to paediatric gastroenterology outpatients' clinic with chronic idiopathic constipation, defined in accordance with Rome II diagnostic criteria. Children were randomised to receive either a cocoa husk supplement rich in dietary fibre (one sachet (5.2 g): 4 g cocoa husk plus 1 g betafructosans) or placebo (one sachet (5.2 g): glucose, cocoa flavouring and excipients) during 4 weeks. The fibre supplement of cocoa husk contained 53.2 g of fibre (39.6 g of total fibre and 13.6 g of betafructosans) per 100 g of product. Insoluble fibre represented 37.2% and soluble fibre represented 2.4% of the total fibre. Cellulose and uronic acids were the main type of insoluble fibre and soluble fibre, respectively. In addition both groups received the same standardised toilet training procedures during the study period. Doses for both products in children aged 3 to 6 years were one sachet before lunch and one sachet before dinner and in children aged 7 to 10 years it was two sachets before lunch and dinner. Parents were instructed to dissolve the content of the sachets in 200 ml of whole milk before ingestion.

The number of bowel movements per week (mean) did not differ significantly between the two treatment groups. The percentage of children reporting hard stool consistency decreased significantly more in children taking the cocoa husk supplement when compared to children taking placebo (cocoa husk group: 41.7 versus placebo group: 75.0; P = 0.017). Significantly more children on the cocoa husk group reported a subjective improvement in stool consistency compared to children on placebo (cocoa husk group (n=24), improvement 14, no improvement 10 versus placebo group (n=24), improvement 6, no improvement 18; P = 0.039). Subjective improvement in pain on defecation was not significantly different between the two groups. No significant adverse effects, such as a new onset of abdominal pain, bloating, abdominal distension, excessive gas, diarrhoea or anaphylactic symptoms, were reported during the 4 week period with either treatment. There were no significant differences in baseline characteristics between the two groups. Eight children withdrew from the study before its completion (five children discontinued study because of the difficulty of the protocol and three were excluded because of the presence of positive antigliadin and antiendomysium antibodies). Data refer only to 48 participants who completed the study. Intention to treat analysis was not performed.

One prospective case series conducted in Italy98 (2000) [EL=3] evaluated the efficacy of glucomannan as a treatment for chronic constipation in children with severe neurological damage. The study included 20 children with severe neurological damage and constipation of at least 12 months duration (14 boys, mean age 5.7 ± 4.2 years). In most patients evacuation was not possible without enema. Children were fed by mouth with semi-liquid diet including formula and puréed food. All children received treatment for disimpaction with enemas for 2 or 3 days (not clear what medication was used). After that children were randomised to receive either glucomannan at a dose of 100 mg/kg two times a day or placebo at the same dose, for 12 weeks. Both glucomannan and placebo consisted of a 500 mg capsule which was given orally mixed with 100 ml of water. An arbitrary scoring system was used for assessment of symptoms:

  • stool consistency: 1 pellets; 2 hard; 3 soft; 4 loose; 5 liquid
  • presence of painful defecation: 1 often; 2 occasionally; 3 none.

None of the outcomes changed significantly at any of the study periods for the placebo group when compared to baseline. The number of stools per week significantly increased in the glucomannan group at all assessment points when compared to baseline (at 4 weeks: mean 4.0 ± 1.3; at 8 weeks: 3.3 ± 1.0; at 12 weeks: 3.8 ± 0.9; P < 0.001 for all). Stool consistency significantly improved in the glucomannan group at all assessment points when compared to baseline (mean score at 4 weeks: 2.4 ± 0.5; at 8 weeks: 2.8 ± 0.7; at 12 weeks: 2.7 ± 0.7; P < 0.001 for all). Painful defecation improved significantly only at the 12 week assessment for the glucomannan group compared to baseline (mean score at 12 weeks: 1.9 ± 1.2; P < 0.01). Laxative use was significantly reduced in the glucomannan group at the 4 and 12 week assessments (mean number of laxatives per week at 4 weeks: 0.3 ± 0.8; at 12 weeks: 0.3 ± 0.5; P < 0.01). There were no significant differences in baseline characteristics between the two groups. One patient receiving glucomannan withdrew from the study after three weeks of treatment because of concomitant increase in seizure frequency associated with blood level of phenobarbital below the therapeutic range.

One prospective case series (pilot study) conducted in Hong Kong99 (2000) [EL=3] evaluated the fibre intake of severe developmentally disabled children living in a residential institution along with the possibility of reducing the use of laxatives by increasing their fibre intake. The study included 20 severely developmentally disabled children (age range 3 to 17 years) with idiopathic constipation who were able to take oral feeding and medically stable. All children received fibre supplementation with wheat bran (All Bran®, Kellogg's) added in breakfast. During stage 1 (20 days), 15 g was added to each serving of breakfast (total fibre intake 17 g). Following stage 1 there was a period of 10 days where children received their normal diet without any supplementation. During stage 2 (6 weeks) 19 g was added to each serving of breakfast (total fibre intake 21 g). Baseline fibre intake was around 2 g/day.

The number of laxatives per week decreased significantly at the end of stage 1 when compared to baseline (baseline: 1.22, SD 0.36 versus end of stage 1: 0.9, SD 0.75; P < 0.05) and at the end of stage 2 when compared to baseline (baseline: 1.22, SD 0.36 versus end of stage 2: 0.7, SD 0.40; P < 0.01) but there were no significant differences when comparing end of stage 1 and end of stage 2. Outcomes for bowel movements were not reported in the paper.

An open label non-RCT conducted in the USA100 (1955) [EL= 1-] evaluated the effectiveness of a palatable mixture containing prune and fig concentrate and non-diastatic malt syrup neutralised with potassium carbonate for the treatment of idiopathic constipation in infants and children. The study included 200 infants and children aged 3 months to 8 years with idiopathic constipation. One group had a prune and fig concentrate (Prune-Malt®, Benson-Nuen Laboratories Inc) added to their diet for three weeks and the control group received no intervention. The prune and fig concentrate was given to infants aged 3 weeks to 1 year as 2 tablespoonfuls daily added to milk or juice. Children aged 1 to 4 years received 3 tablespoonfuls daily added to milk or food and children aged 4 to 8 years received 4 tablespoonfuls daily added to milk or food. No changes were made to their usual diet and no drugs were given. No definitions or scoring system were given for: ‘improvement’, ‘no improvement’, ‘return to normality’, ‘good’ ‘fair’ and ‘poor’.

Twenty-eight children who received the prune and fig concentrate returned to normality compared to 16 children in the control group. Fifty-one children who received the prune and fig concentrate improved compared to 25 children in the control group. Only 21 children who received the prune and fig concentrate did not improve compared to 59 children in the control group. In total 132 parents rated the treatment as good, 47 as acceptable and 21 as poor (P values not reported in the study). No comparison was made between baseline characteristics of the two groups, although the author stated that wherever possible, cases of equal severity and ages were equally divided between the two groups. No attrition or loss to follow-up was reported.

Probiotics

A double blind RCT conducted in Taiwan79 (2007) [EL=1+] investigated the effect of probiotics (lactobacillus casei rhamnosus, Lcr35) alone in the treatment of chronic constipation in children and to compare the effect with magnesium oxide (MgO) and placebo, respectively. The study included 45 children (23 male) under 10 years with chronic constipation. Children were randomised into three groups to receive during 4 weeks: MgO 50 mg/kg/day, twice a day; Lcr35 8 × 108 colony forming units (CFUs) per day (antiobiophilus 250 mg, two capsules, twice a day); or placebo (starch in content). Lactulose use (1 ml/kg/day) was allowed when there was no stool passage noted for 3 days. Glycerine enema was used only when there was no defecation for more than 5 days or when abdominal pain was suffered due to stool impaction.

Defecation frequency significantly increased in children taking MgO and probiotic compared to placebo (MgO (n=18): mean 0.55 times/day ± 0.13; probiotic (n=18): 0.57 times/day ± 0.17; placebo (n=9): 0.37 times/day ± 0.10; P = 0.006 [placebo versus probiotic]; P = 0.01 [MgO versus placebo]). However, there were no significant differences between children taking probiotic and children taking MgO regarding this outcome. The percentage of children having hard stools was significantly lower in children taking MgO and in those taking probiotic compared to placebo (MgO: 23.5% ± 7.9; probiotic: 22.4% ± 14.7; placebo: 75.5% ± 6.1; P = 0.02 [placebo versus probiotic]; P = 0.03 [MgO versus placebo]) but there were no significant differences between children taking probiotic and children taking MgO regarding this outcome.

Children taking placebo had to make use of glycerine enemas significantly more often than children taking either MgO or placebo (MgO: 1.3 times ± 1.9, probiotic: 1.6 times ± 1.9, placebo: 4.0 times ± 2.1; P = 0.04 [placebo versus probiotic]; P = 0.03 [MgO versus placebo]) but there were no significant differences between children taking probiotic and children taking MgO regarding this outcome. There were no significant differences regarding use of lactulose and faecal soiling amongst the three groups. Significantly more patients were successfully treated with MgO or probiotic compared to placebo (MgO 72.2%, probiotic: 77.8%, placebo: 11.1%; P = 0.01 [placebo versus probiotic], P = 0.01 [MgO versus placebo]). However, there were no significant differences between children taking probiotic and children taking MgO regarding this outcome.

No adverse effects were noted in the probiotic and placebo groups and only one patient in the MgO group suffered from mild diarrhoea. There were no significant differences at baseline amongst the three groups regarding: gender, age of enrolment, age of onset of constipation, duration of constipation, previous treatment, defecation period, stool consistency, abdominal pain, faecal soiling, bleeding during defecation, use of enema and taking fruits or vegetables daily. Four patients discontinued medication during the study period: two in the MgO group, one in the probiotic group and one in the placebo group. Two patients suffered from acute gastroenteritis (not clear whether as a consequence of the study medication) and two patients were lost to follow-up.

A triple-blind RCT conducted in Poland101 (2005) [EL=1+] assessed the effectiveness of lactobacillus rhamnosus GG (LGG) as an adjunct to lactulose in the treatment of constipation in children. The study included 84 children aged 2 to 16 years with idiopathic constipation defined as less than 3 bowel movements per week for at least 12 weeks (ages: lactulose plus LGG group 79 months ± 47, lactulose plus placebo group 65 months ± 36; gender not reported). All children received treatment for disimpaction with phosphate and saline enema before study treatment started. Children were then randomised to receive during 12 weeks either lactulose 70%, 1 ml/kg/day (in two divided doses) plus 109 CFUs of LGG or lactulose 70%, 1 ml/kg/day (in two divided doses) plus placebo. From weeks 13 to 24 patients were instructed to continue the use of lactulose or other laxatives as needed. Treatment success was defined as 3 or more spontaneous bowel movements per week with no episodes of faecal soiling.

Treatment success at 12 and 24 weeks was not significantly different between the two treatment groups. The average number of spontaneous bowel movements per week, episodes of faecal soiling per week and straining frequency per week were not significantly different when comparing both treatment groups at 4, 8 and 12 weeks. The percentage of patients using laxatives at 24 weeks was not significantly different between the two groups. LGG was well tolerated. The number of patients experiencing side effects was not significantly different between the two groups and the side-effects profile of LGG was similar to that of placebo: three patients in the LGG group versus five patients in the placebo group developed abdominal pain.

One patient in the LGG group developed vomiting and one in the placebo group experienced headache. There were no significant differences in baseline characteristics between the two groups. Five children in the LGG group discontinued the intervention (four because of clinical improvement, one developed abdominal pain) versus three patients in placebo group who discontinued the study without receiving any intervention (two refused to participate and one because of another reason, not provided). Outcomes for stool consistency were not reported in the paper.

One prospective case series (pilot study) conducted in The Netherlands102 (2007) [EL=3] determined the therapeutic effect of a combination of probiotic strains, containing the bifidobacteria B. bifidum, B. infantis and B. longum and the lactobacilli L. casei, L. plantarum and L. rhamnosus, on childhood constipation. The study included 20 children aged 4 to 16 years referred to outpatient clinic with idiopathic constipation, as defined by Rome III criteria (10 boys, median age 8 years). All children received treatment for disimpaction using rectal enema (Klyx, sodium-dioctylsulfosuccinate and sorbitol) once daily for 3 days. For the following 4 weeks children received a daily probiotics mixture of 4×109 CFUs containing bifidobacteria B. bifidum, B. infantis and B. longum and lactobacilli L. casei, L. plantarum and L. rhamnosus. During the treatment period children were instructed to start toilet training. Toilet training consisted of sitting on the toilet three times per day for 5 minutes after each meal with the intention of trying to defecate. Use of laxatives was not allowed during treatment period.

The frequency of bowel movements (BMs) per week in the total sample did not change significantly at weeks 2 and 4 when compared to baseline. The frequency of BMs per week in 12 children presenting with more than 3 BMs per week at baseline increased significantly at weeks 2 and 4 when compared to baseline (baseline: median 1.0, range 0.0 to 2.0; week 2: median 3.0, range 0.0 to 7.0), P = 0.01; week 4: median 3.0, range 0.0 to 10.0; P = 0.009). The number of children reporting hard stools did not change significantly at week 2 and week 4 compared to baseline. At week 4, hard stools appeared in five children who had also had hard stools at baseline. One child with normal stools at baseline reported hard stools only at the end of the study. Two of the seven children who presented with hard stools reported normal stools at the end of the study. The number of faecal incontinence episodes per week decreased significantly at both week 2 and week 4 when compared to baseline (baseline: median 4.0, range 0.0 to 35.0; week 2: median 1.5, range 0.0 to 14.0; week 4: median 0.3, range 0.0 to 7.0; P = 0.007 and P = 0.001 respectively). There were no side effects, such as vomiting, bloating and increased flatulence, during the study period. No attrition or loss to follow-up was reported.

Excluding cows' and goats' milk††††

A double-blind crossover RCT conducted in Italy103 (1998) [EL=1+] compared the effects of cows' milk and soy milk in children with chronic constipation. Sixty-five consecutive children diagnosed with chronic idiopathic constipation underwent an observation period during weeks 1 and 2 when all medications were stopped. During weeks 3 and 4, one group (n=33) was randomly assigned to receive cows' milk and unrestricted diet and the other (n=32) had cows' milk and its derivatives excluded from their diet and received soy milk instead. During week 5 there was a ‘washout period for both groups with unrestricted diet and intake of soy or cows' milk and its derivatives. During weeks 6 and 7 patients were switched to the other type of milk. After the two study periods children with a response to the cows' milk free diet were given the soy milk diet for another month and then underwent a 2 week double-blind challenge with cows' milk at hospital. Children with eight or more bowel movements during a treatment period were considered to have a response.

Children were randomly assigned to receive cows' milk or a placebo containing soy milk. If no clinical reactions were observed within 12 hours, patients were discharged and the challenge continued at home. A qualitative faecal score was defined as 1 (mushy or liquid stool), 2 (soft faeces and no pain in passing stools) or 3 (hard faeces and difficulty and pain on passing stools). Patients were followed up for a mean period of 10 months (range 3 to 20).

During the observation period (n=65) the number of bowel movements was a median of 4 (25th to 75th percentile: 3 to 5) and the qualitative faecal score (QFS) was 3 for all 65 patients. During the two study periods neither the number of bowel movements, nor the qualitative faecal score changed significantly for the cows' milk group (n=65) compared to the observation period. For the group who had a response to the soy milk diet (n=44) the number of bowel movements increased significantly (median: 10, 25th to 75th percentile, 4 to 12) and 44 patients stopped having pain or difficulty passing stools (QFS 1 n=2; QFS 2 n=42; QFS 3 n=21) (P < 0.001 for all variables). During the challenge with cows' milk (n=44) no patients in the placebo group (soy milk) showed any clinical reactions. Patients in the cows' milk group did not have any acute reaction, but in all of them constipation associated with hard stools and discomfort on defecation reappeared after 5 to 10 days on the diet. The cows' milk-free diet was therefore recommenced, with a consequent normalisation of bowel movements in all patients.

Neither the number of bowel movements nor the qualitative faecal score were specifically measured during the challenge period. During the follow-up period none of the children with response had constipation. Cows' milk was reintroduced into the diets of 15 children after 8 to 12 months of the cows' milk-free diet and in all cases constipation returned within 5 to 10 days. Children with no response to soy milk diet were treated with high doses of laxatives, with subsequent improvement in stool frequency. In all cases symptoms returned once treatment with laxatives was stopped.

There were significant baseline differences in the groups of children with and those without a response. Anal fissures with erythema or oedema were more common among those with a response (40 of 44 patients versus 9 of 21, P < 0.001). Furthermore, at diagnosis, symptoms of suspected intolerance to cows' milk were more common in children with a response (11 of 44 patients versus 1 of 21; P = 0.05): recurrent bronchospasm in four patients, rhinitis in four and dermatitis in three. Six patients were withdrawn from the study during the cows' milk study period (on days 9 to 12) because of the reappearance of constipation and other related disorders. For children withdrawn from study during the cows' milk study period the number of bowel movements per period was prorated. Intention to treat analysis was used. Patients included in this study were highly selected and this might have led to overestimation of the frequency of cows' milk intolerance as a cause of constipation. Paediatricians who referred the patients may have pre-selected them as being likely to have a food intolerance since the study centre specialised in the treatment of food allergies. The inclusion of patients with no response to laxatives may have also contributed to this issue. It should be noted that the two types of milk taste different from one another, thus undermining the degree of blinding achievable.

A small prospective case series and embedded randomised controlled challenge conducted in Italy104 (2006) [EL=3] evaluated the histology and manometry characteristics of patients with food intolerance-related constipation. Thirty-six children (age range 9 months to 10 years) with chronic constipation underwent a cows' milk-free diet for 4 weeks, following a 2-week observation period where all medications were stopped. After 12 weeks all patients cured on the cows' milk free diet or oligoantigenic diet (n=17) underwent a 2-week double-blind placebo-controlled challenge with cows' milk at the hospital. Patients were randomised to receive either cows' milk or ass's milk as placebo. If no clinical reactions (not specified which ones) occurred after 12 hours, patients were discharged and the challenge continued at home with bottles coded A or B. The challenge was stopped when a clinical reaction occurred. Outcome measures were number of bowel movements per week, appearance of stools and child's degree of difficulty in passing stools. The last two measures were combined in a QFS. A score of 1 was given if mushy or liquid stools, 2 if soft faeces and no pain in passing stools and 3 if hard stools and difficulty and pain on passing stools.

During the observation period both for patients further diagnosed with food intolerance (n=17; 14 to cows' milk only, 3 with multiple food intolerance) and for patients with constipation unrelated to food intolerance (n=19) the number of bowel movements per week and the QFS were the same (number of bowel movements: median 1.5, 25th to 75th percentile 1–2; qualitative faecal score: 1 n=0, 2 n=0, 3 n=36). During the elimination diet period the number of bowel movements per week in patients with food intolerance (n=17) significantly increased (median 5, P < 0.01, 25th to 75th percentile 3–7) and no children presented with hard stools or difficulty and pain on passing stools (QFS 1 n=1, QFS 2 n=16, QFS 3 n=0; P < 0.01 for the three values). For patients with constipation unrelated to food intolerance (n=19) both the number of bowel movements per week and the QFS remained the same as during the observation period and were significantly different from the results obtained in the group with food allergy (P < 0.01).

During the cows' milk challenge period cows' milk readministration caused the reappearance of constipation in all cases, very often associated with painful defecation, within 5 days after the commencement of the challenge (median 2 days, range 1–5 days). These symptoms disappeared on returning to the cows' milk-free diet or oligoantigenic diet in the three patients with multiple food intolerance. Patients with chronic constipation caused by food intolerance showed at baseline a higher frequency of a personal history of previous food intolerance (P < 0.01) and concomitant signs of food intolerance (bronchospasm four cases, dermatitis two cases; P = 0.05) than patients with constipation unrelated to food intolerance.

A second small prospective case series and embedded randomised controlled challenge conducted in Italy105 (2005) [EL=3] evaluated the histologic data in patients with food intolerance-related constipation. Fifty-two infants and children with chronic constipation unresponsive to previous treatments underwent a 2-week observation period where all medications were stopped and at the end of the second week they were given a clean-out with a single dose of PEG 4000 (0.75 g/kg). For the next 4 weeks cows' milk and all its derivatives were excluded from the diet of all patients. Patients unresponsive to a cows' milk-free diet were placed on an oligoantigenic diet for 4 weeks (also excluding cows' milk). After 12 weeks all patients cured on cows' milk free or oligoantigenic diet underwent a 2-week, double-blind, placebo-controlled challenge with cows' milk at hospital. Patients were randomised to receive either cows' milk or ass's milk as placebo. If no clinical reactions (not specified) occurred after 12 hours, patients were discharged and the challenge continued at home with bottles coded A or B. The challenge was stopped when a clinical reaction occurred. Outcome measures were number of bowels movements per week and the QFS. Both were recorded by parents during the observation period and the elimination diet period. The qualitative faecal score was defined as 1 (mushy or liquid stool), 2 (soft faeces and no pain in passing stools) and 3 (hard faeces and difficulty and pain on passing stools). Children with eight or more bowel movements during a treatment period were considered to have a response. Normalised stools habits were defined as: bowel frequency of at least five evacuations per week with the elimination of soft stools without pain.

During the observation period both patients with food intolerance (n=30) and patients with constipation unrelated to food intolerance (n=22) had a median of 1.5 bowel movements per week (25th to 75th percentile 1–2) and all 52 patients a QFS of 3. During the elimination diet period the number of bowel movements per week increased significantly for patients with food intolerance (median 5, 25th to 75th percentile 4–7; P < 0.001) and no children presented with hard stools or difficulty and pain on passing stools (QFS 1 n=2, QFS 2 n=28, QFS 3 n=0; P < 0.01 for the three values). For patients with constipation unrelated to food intolerance both bowel movements per week and QFS remained the same as during the observation period. For all children cows' milk readministration caused the reappearance of constipation within 5 days after commencing the challenge (median 2 days, range 1–5 days).

Patients with chronic constipation caused by food intolerance showed at baseline a higher frequency of a personal history of previous food intolerance (P = 0.02) and concomitant signs of food intolerance (bronchospasm five cases, rhinitis four cases, dermatitis two cases) than patients with constipation unrelated to food intolerance (P = 0.03). No difference was observed between the 24 patients with cows' milk intolerance and the six patients with multiple food intolerance for outcome measures considered (number of bowel movements and qQFS), either at baseline or on elimination diet. However, in comparison with patients intolerant to cows' milk alone, patients suffering from multiple food intolerance were older (P = 0.04) and had a higher frequency of family history of atopic disease (P = 0.03). It should be noted that the high frequency of chronic constipation owing to food intolerance found in this study was likely due to a selection bias, as mainly food-intolerant patients are treated at the centre where the study was conducted.

Another small prospective case series conducted in Italy106 (1995) [EL=3] aimed to investigate the possible relation between constipation and cows' milk protein (CMP) allergy (CMPA). The study sample comprised 27 infants considered to have idiopathic constipation. During the first 7 days all patients were being fed the same diet as at the time of diagnosis: various forms of commercial formula derived from cows' milk or whole cows' milk and its derivatives. For the next month all patients started a CMP-free diet. Three patients aged younger than 12 months were fed a formula containing soy protein and the others received soy milk or ass's milk (eight cases) and all cows' milk derivatives were excluded. After a month all patients whose symptoms abated underwent a cows' milk challenge. Cows' milk was given for a maximum of 10 days; then these patients started again an exclusion diet for 1 month and then a second cows' milk challenge was performed. Outcome measures were number of stools per day and QFS. The QFS was defined as in the studies described above.

During the first month of the CMP-free diet there was a significant improvement in symptoms in 21 patients: the frequency of stools significantly increased, faeces were soft and none of the infants had any discomfort when passing stools (mean number of stools per day on unrestricted diet (a): 0.24 ± 0.10; on first CMP-free diet (b): 1.04 ± 0.120; QFS on unrestricted diet (a): 2.85 ± 0.05; on CMP-free diet (b): 1.90 ± 0.08). During the first challenge constipation returned within 48 hours after the reintroduction of cows' milk, passing stools became painful and in seven patients with abdominal pain, ingestion of cows' milk was discontinued on day 4 (mean number of stools per day on first CMP challenge (c): 0.31 ± 0.14; QFS on CMP challenge (c): 2.75 ± 0.11).

During the second period of CMP-free diet the stools became normal again in the 21 patients and the symptoms accompanying constipation disappeared (mean number of stools per day on second CMP-free diet (d): 1.05 ± 0.11; significance: (b) and (d) versus (a) and (c), P < 0.0005) (QFS for second CMP-free diet: 1.85 ± 0.10; P < 0.001). During the second challenge symptoms reappeared within 24 to 48 hours: all 21 patients had painful passage of stools and for this reason the challenge was suspended on the third day.

Six patients did not improve on the first CMP-free diet period (mean number of stools per day on unrestricted diet: 0.18 ± 0.12; on first CMP-free diet: 0.20 ± 0.13) and their difficulty in passing stools did not change (QFS: control: 3; first CMP-free diet: 3). These patients were subsequently treated with lactulose and only a partial regression in symptoms was observed. They were permanently given an unrestricted diet, except for one infant who had episodes of recurrent bronchospasm related to the ingestion of cows' milk.

Patients were followed up monthly for a mean period of 18 months (range 10 to 30 months). Reintroduction of cows' milk was cautiously attempted in 16 children 6 to 9 months after the diagnosis of CMP allergy-dependant constipation. In eight children CMP did not cause the onset of any problems and it was reintroduced on a permanent basis; in eight patients CMP led to the reappearance of constipation within 2 to 3 days after introduction, and these infants were still following CMP-free diet at the time the paper was written. No harmful reactions with either soy milk or ass's milk were reported. It is important to note that significant differences at baseline were found between patients who were cured with the CMP-free diet and those whose condition did not improve with this diet. Patients who were cured with the CMP-free diet were more likely to have a history of CMP allergy or symptoms of CMP allergy (atopic dermatitis or recurrent episodes of bronchospasm) at the time they entered the study than those whose condition did not improve with this diet (15 out of 21 versus 1 out of 6; chi square= 3.75; P < 0.05).

Increasing fluid intake

One open label RCT conducted in the USA107 (1998) [EL=1-] aimed to determine whether or not increasing fluid intake by either excess water intake or excess hyperosmolar liquid intake would significantly alter the course of simple constipation in children. The study included 90 prepubertal children with moderate to severe idiopathic constipation (31 boys [47.46%], mean age 7.5 years, age range 2.5 to 12.5 years). Children were randomised into two intervention groups and one control group. During 2 weeks one intervention group was instructed to increase water intake by 50% on the basis of the total measured oral liquid intake during the baseline week. The second group received supplemental liquid in the form of hyperosmolar liquids: Kool-Aid, juice, soda pop or other liquids known to contain more than 600 mOsm/l. The control group received no intervention.

Neither increasing water intake nor increasing hyperosmolar liquid intake significantly increased stool frequency or improved stool consistency or difficulty with stool passage within groups when comparisons were made with previous weeks, or between the three groups during the same week (analysis of variance). A second round of analysis excluded all subjects who failed to comply with at least 75% of assigned intervention and this did not change the study outcomes. No comparison was made of baseline characteristics between the three groups. The study originally included 108 children but only 90 completed the entire study as assigned. Eighteen children failed to comply with 75% of the intervention.

Increasing physical activity

One open non-randomised controlled trial conducted in Israel108 (2009) [EL=1-] assessed the effect that stepping while standing had on constipation in children with severe cerebral palsy (CP). The trial included 22 children (aged 3.5 to 10 years) with a diagnosis of spastic quadriplegic CP with gross motor function classification system (GMFCS) level 4 or 5. All children were unable to stand and walk with a traditional walker or rollator because of insufficient upper extremity control, would attempt to step when supported in a standing position and had flexion contractures of less than 30° in the hips and the knees. Eleven children began a trial of the David Hart Walker (HW) orthosis in addition to their physical therapy sessions (six males, mean age 6.1 years ± 2.1) and 11 children who were matched for age and gender with the study group (six males, mean age 6.7 years ± 1.6) underwent a program with a standing frame (SF) as part of their physical therapy session. At entry the proportion of constipation in both groups was equal (6 out of 11 [54.5%]). After 6 months the study (HW) group had significantly reduced their level of constipation (1 out of 11 [9.1%]) and the control (SF) group had no change in constipation (6 out of 11 [54.5%]) (P = 0.02). It should be noted that the sample size was very small and that the paediatric evaluation of disability inventory (PEDI) was higher at baseline in the study group compared to the control group (indicating better self care, mobility and social function). There was no attrition or loss to follow up in either group.

Evidence statement

Dietary modifications

There is no evidence for the clinical effectiveness of dried or fresh fruits, fruit juices, vegetables, cereals, fructo-oligosaccarides, omega 3 fish oils or excluding goats' milk from the diet for ongoing treatment or maintenance in children with chronic idiopathic constipation.

Increasing fibre

One double-blind RCT [EL=1+] showed that there were no significant differences between a yogurt drink with mixed dietary fibre (transgalacto-oligosacharides, inulin, soy fibre and resistant starch) and a yogurt drink containing lactulose at increasing defecation frequency per week and decreasing the number of patients with 1 or more faecal incontinence episodes per week. The study also showed that the stool consistency was significantly softer in the lactulose group compared to the fibre group. The number of patients using step-up medication at 3 weeks was significantly smaller in the group taking fibre than in the group taking lactulose but there were not significant differences regarding this outcome at 8 and at 12 weeks.

One double-blind RCT (pilot study) [EL=1+] showed that a cocoa husk supplement rich in dietary fibre (cocoa husk plus betafructosans) was more effective than placebo at decreasing the number of children reporting hard stool consistency and increasing the number of children reporting a subjective improvement in stool consistency. The study also showed that there were no significant differences between the cocoa husk supplement and placebo at subjectively improving pain on defecation and increasing the number of bowel movements per week.

One prospective case series [EL=3] showed that fibre supplementation with wheat bran was effective at decreasing the number of laxatives used per week

One open label non-RCT [EL=1-] showed that a palatable mixture containing prune and fig concentrate and non-diastatic malt syrup neutralised with potassium carbonate was effective at improving constipation. One hundred and thirty two parents rated the treatment as good, 47 as acceptable and 21 as poor.

Supplements

There is no evidence for the clinical effectiveness of supplements containing partially hydrolysed guar gum, iron or pectin for ongoing treatment or maintenance in children with chronic idiopathic constipation

One double-blind RCT (crossover) [EL=1+] showed that glucomannan (a polysaccharide of d-glucose and d-mannose, equal to 450 mg of alimentary fibre) was more effective than placebo at successfully treating constipation as per physician rating and improving children's symptoms as per parent rating. Successful treatment (physician rating) and improvement (parent rating) were independent of amount of fibre intake from the treatment. Significantly more children who were also taking laxatives were treated successfully with glucomannan than with placebo. Children with constipation only were significantly more likely to be treated successfully with glucomannan compared with children with constipation and encopresis

One prospective case series [EL=3] showed that glucomannan was effective at significantly increasing the number of stools per week, improving the stool consistency and painful defecation and reducing laxative use.

Probiotics

One double blind RCT [EL=1+] showed that there were no significant differences between probiotic (lactobacillus casei rhamnosus, Lcr35) and magnesium oxide (MgO) at increasing daily defecation frequency and decreasing the percentage of children having hard stools and both were more effective than placebo at increasing daily defecation frequency and decreasing the percentage of children having hard stools. There were no significant differences between the three treatments at decreasing faecal soiling. Children taking placebo had to make use of glycerine enema significantly more often than children taking either MgO or probiotic (lactobacillus casei rhamnosus, Lcr35) but there were no significant differences between children taking probiotic (lactobacillus casei rhamnosus, Lcr35) and children taking MgO regarding this outcome. There were no significant differences between the three groups regarding the need to use lactulose. Significantly more patients were successfully treated with MgO or probiotic (lactobacillus casei rhamnosus, Lcr35) compared to placebo but there were no significant differences between children taking probiotic and children taking MgO regarding this outcome.

One triple blind RCT [EL=1+] showed that there were no significant differences between probiotic (lactobacillus rhamnosus GG) plus lactulose and placebo plus lactulose at increasing the average number of spontaneous bowel movements per week and decreasing the episodes of faecal soiling per week, the straining frequency per week and the number of patients using laxatives.

One prospective case series [EL=3] showed that a probiotics mixture (bifidobacteria B. bifidus, B. infantis and B. longum plus lactobacilli L. casei, L. plantarum and L. rhamnosus) was effective at significantly decreasing the number of faecal incontinence episodes per week only in children presenting with less than 3 bowel movements per week at baseline. The study also showed that the probiotics mixture was not effective at improving stool consistency.

Infant formulae

One double-blind RCT (crossover) [EL=1+] showed that there were no significant differences between Nutrilon Omneo (new formula, NF), a formula with higher protein content, 100% of it based on whey protein hydrolysate (no casein, no intact whey protein), a mixture of prebiotic oligosaccharides (GOS and lcFOS), a higher concentration of sn-2 palmitic acid and a lower lactose content and a standard formula (SF) at reducing painful defecation and increasing defecation frequency. The study also showed that NF was significantly more effective than SF at improving the stool consistency.

One open label RCT [EL=1-] showed that Novalac-IT, a magnesium-enriched infant formula, was significantly more effective than a 20% strengthened regular infant formula at improving stool consistency, increasing stool frequency and reducing difficulties in defecation.

One open label RCT showed that a new formula (NF) (Omneo / Conformil) based on palmitic acid predominantly esterified at the β-position, oligosaccharides (GOS and FOS) with a prebiotic activity, partially hydrolysed protein, low lactose content and higher density was significantly more effective than a standard formula at increasing stool frequency.

One prospective case series showed that a new formula (NF)††† was effective at increasing stool frequency.

One prospective case series [EL=3] showed that Novalac Anti-Constipation, a formula with an adapted concentration of magnesium and lactose, was effective at significantly increasing the number of daily stools and the number of children having normal stools, as well as at reducing the number of children presenting with pain or discomfort on defecation and the number of children needing external help at defecation.

Soy milk

One double-blind RCT (crossover) [EL=1+] showed that excluding cows' milk and its derivatives from the diet and giving soy milk instead was more effective than giving an unrestricted diet including cows' milk and its derivatives at significantly increasing the number of bowel movements, improving stool consistency and reducing the pain or difficulty on passing stools in children with chronic constipation and suspected food allergies, but was not effective in children in whom food allergies were not suspected at baseline.

Excluding cows' milk protein from the diet

One double-blind RCT (cross over) [EL=1+] showed that excluding cows' milk and its derivatives from the diet was more effective than giving an unrestricted diet including cows' milk and its derivatives at significantly increasing the number of bowel movements, improving stool consistency and reducing the pain or difficulty on passing stools in children with chronic idiopathic constipation and suspected food allergies, but was not effective in children in whom food allergies were not suspected at baseline.

Three small case series and embedded randomised controlled challenges [EL=3] showed that a cows' milk-free diet was effective at increasing the number of bowel movements, improving stool consistency and reducing the pain or difficulty on passing stools in children with chronic constipation and food intolerance, but was not effective in children with constipation unrelated to food intolerance

Increasing fluid intake

One open label RCT [EL=1-] showed that increasing liquid intake by either excess water intake or excess hyperosmolar liquid intake did not have significant impact on stool frequency, stool consistency or difficulty with stool passage in constipated children when compared to controls who did not increase their fluid intake.

Increasing physical activity

One open non-randomised controlled trial [EL=1-] showed that a device which allows children with severe cerebral palsy to step while standing was more effective than passive standing in improving symptoms of constipation.

GDG interpretation of the evidence

The opinion of the GDG is that a poor diet alone is rarely the cause of childhood constipation. The GDG consensus is that it is extremely important to emphasise that diet is important but that it is not the first factor to consider in the treatment of constipation. Dietary manipulations should be carried out alongside treatment with laxatives and behavioural therapy.

Increasing physical activity

Despite the fact that there is no good quality evidence for the effectiveness of increasing physical activity to improve constipation, it is the opinion of the GDG that exercise should be encouraged. It is a common clinical observation that a lack of physical activity can be a contributing factor in constipation. While recognising that physical activity is not in itself a treatment for constipation, the GDG felt that it was important to encourage children to be physically active, as it may decrease the likelihood that they will develop constipation again once an episode has been medically treated, bearing in mind what is achievable and appropriate for the individual child. It has been recommended by the Department of Health88 that children should do at least 60 minutes of moderate intensity physical activity per day as part of a healthy lifestyle.

Fibre-rich foods

No evidence was found to suggest that increasing fibre-rich foods, such as fruits, vegetables and cereals, is effective in treating or managing constipation. The GDG felt that encouraging children to eat more fibre, when they are already having a healthy balanced diet with sufficient fibre, could be detrimental. A high fibre intake in this case could exacerbate symptoms and potentially increase soiling. It is the opinion of the GDG that children should be advised to eat a healthy diet, including fibre containing foods, as outlined by the Paediatric Group of the British Dietetic Association in ‘Food for the Growing Years’ and ‘Food for the School Years’.109,110

Fibre supplements

The evidence for using fibre supplements, such as prune and fig concentrate, cocoa husks and glucomannan, in the treatment of constipation is very limited. It is the view of the GDG that this evidence is not enough to recommend these products in the treatment or ongoing management of idiopathic constipation.

Probiotics

The GDG felt it was not possible to recommend specific probiotics at this stage as there is little evidence (only small trials, admittedly well conducted): the three studies refer to three different probiotics and in one case the probiotic was given in addition to lactulose. Additionally, some probiotics are not available commercially and the commercially available probiotics do not always say what their active ingredient is.

Infant formulas

The GDG examined four studies, each on a different infant formula, none of which are used in the UK. The GDG believes that there is not enough evidence to suggest that any of the formulas are clinically effective in the treatment or ongoing management of constipation.

The GDG believes that the current common practice of switching from one infant formula to another to alleviate constipation may be detrimental. It takes time to trial infants with different feeds and this often delays treatment with laxatives.

Excluding cows' milk

Although there is some evidence for excluding cows' milk from the diet to improve constipation, the opinion of the GDG is that the studies are of a poor quality and the selection of participants was biased. In the studies which were reviewed, both soy and ass's milk were used in the placebo group. Recommendations in the UK are that children with suspected cows' milk protein intolerance should be given feeds based on extensively hydrolysed proteins. ††89 Soy and ass's milk are inappropriate alternatives to cows' milk and should be avoided due to a risk of allergenic cross-reactivity.

Replacing goats' milk

No evidence was found on replacing cows' milk with goats' milk in the diet to improve constipation in children. The recommendation from the Department of Health is that goats' milk is not suitable to be used as an infant feed because of its high renal solute load, inadequate vitamin and mineral content and doubtful microbiological safety.89 Infant formulas based on goats' milk are not available in the UK. In addition, goats' milk protein can be as sensitising as cows' milk protein and is therefore not recommended when a cows' milk protein allergy is suspected.

Increasing fluid intake

The GDG found little evidence for the effectiveness of increasing fluid intake in children with chronic constipation. Despite this, it is the GDG's view that increasing fluid intake to recommended levels is essential. Without sufficient fluid intake, the use of osmotic laxatives will lead to dehydration which can itself contribute to constipation.

Recommendations

Do not use dietary interventions alone as first-line treatment for idiopathic constipation.

Treat constipation with laxatives and a combination of:

  • Negotiated and non-punitive behavioural interventions suited to the child's stage of development. These could include scheduled toileting and support to establish a regular bowel habit, maintenance and discussion of a bowel diary, information on constipation, and use of encouragement and rewards systems.
  • Dietary modifications to ensure a balanced diet and sufficient fluids are consumed.

Advise parents and children (where appropriate) that a balanced diet should include:

  • Adequate fluid intake (see table 5).
  • Adequate fibre. Recommend including foods with a high fibre content (such as fruit, vegetables, high fibre bread, baked beans and wholegrain breakfast cereals) (Not applicable to an exclusively breastfed infant). Do not recommend unprocessed bran, which can cause bloating and flatulence and reduce the absorption of micronutrients.

Provide children and young people with idiopathic constipation and their families with written information about diet and fluid intake.

In children and young people with idiopathic constipation, start a cows' milk exclusion diet only on the advice of specialist services.

Advise daily physical activity that is tailored to the child's stage of development and individual ability as part of ongoing maintenance in children and young people with idiopathic constipation.

Research recommendation

What is the clinical effectiveness of increasing physical activity for ongoing treatment/ maintenance in children with chronic idiopathic constipation?

Why this is important

It has been shown that along with healthy eating, an active lifestyle is essential to improving and maintaining health.88 Increasing activity levels contributes to the prevention and management of many conditions and diseases. It may be that increasing physical activity levels could be beneficial in the treatment of children with chronic constipation

In infants with chronic idiopathic constipation, does changing from one infant milk formula to another improve symptoms? (E.g. Standard infant formula versus infant formula with oligosaccharides versus standard infant formula + laxative)

Why this is important

It is common practice to change from one formula to another to help alleviate constipation. As it takes time to trial infants with different feeds, this can delay much-needed treatment with laxatives. Good quality evidence for the use of a particular infant formula in the treatment of constipation would thus be beneficial.

5.5. Psychological interventions

Introduction

Families of children with idiopathic constipation are often given psychological and/or behavioural advice as well as being referred for more formal psychological therapy. This advice can be given at varying stages of the child's course of constipation, often with little appreciation of the child's and family's ability to carry it out or indeed whether the child is able to achieve what is asked of him or her as far as bowel movements are concerned. For the majority of children the psychological component of their constipation is likely to be secondary to the physical discomfort of being unable to pass stools easily or to the accidental leakage as a result of faecal loading.

Psychological and behavioural interventions can range from predominantly behavioural toilet training to bowel retraining (which may also involve more formal behavioural modification of chaining and shaping programmes) to specific psychological models of therapy such as psychodynamic psychotherapy, cognitive behavioural therapy and systemic family therapy.

From a clinical perspective it is important that any psychological and/or behavioural intervention is implemented alongside effective laxative therapy111,112 in order that the child can achieve comfortable passage of stools and parents have realistic expectations of the child. Any interventions need to be developmentally appropriate for the child and delivered in a child friendly manner as well as facilitating parental support and understanding. In order for any psychological intervention to be carried out effectively it is important that a therapeutic relationship be established to facilitate both the parents' and child's motivation to engage in the intervention and to feel able to maintain this for long enough to see results that they experience as positive.113, 114 It is how the health professional working with the family mediates a holistic intervention that initiates a successful working relationship with both child and family.

Clinical question

What is the clinical effectiveness of psychological and behavioural interventions in addition to laxatives for ongoing treatment or maintenance in children with chronic idiopathic constipation?

Studies considered in this section

Studies were considered if they:

  • included neonates, infants or children up to their 18th birthday with chronic idiopathic constipation
  • included the following interventions in addition to laxatives in at least one of the treatment groups:

    intense psychotherapy (cognitive behavioural therapy [CBT])

    systemic/family therapy or psychodynamic psychotherapy

    psychosocial counselling

    mediational models in cognitive or behavioural therapy

    minimal intervention models using parents in behaviour therapy or behaviour modification

    clinical hypnosis

    toilet/bowel/habit training and retraining

    ‘chaining’ and ‘shaping’ programmes

    maintaining toilet diaries

    rewarding, positive reinforcement, incentive or reward charts, star charts, reward systems

    parenting programmes if they clearly specify what the program was

    psychoeducation (including biofeedback)

    Portage as an educational model.

  • included the following outcomes:

    changes in frequency of bowel movements

    changes in stools consistency or appearance

    changes in pain or difficulty on passing stools

    changes in frequency of episodes of soiling

    reduction in laxatives use

    parent/child views or satisfaction or quality of life

  • were not case reports
  • were published in English.

No restrictions were applied on the publication date or country.

Overview of available evidence

A total of 1689 articles were identified from the searches and 48 articles were retrieved for detailed assessment. Of these, ten studies were included in this review: seven parallel-RCTs, one retrospective cohort, one quasi-randomised RCT and one retrospective audit.

Narrative summary

Conventional treatment alone versus conventional treatment plus biofeedback

Meta-analysis of four RCTs comparing conventional treatment alone versus conventional treatment plus biofeedback showed that treatment success was not significantly different between the two treatment groups either in the medium term (figure 5.2) or in the long term (figure 5.3).

Figure 5.2. Conventional treatment alone versus conventional treatment plus biofeedback: treatment success at medium term (6 months).

Figure 5.2

Conventional treatment alone versus conventional treatment plus biofeedback: treatment success at medium term (6 months).

Figure 5.3. Conventional treatment alone versus conventional treatment plus biofeedback: treatment success at long term (12 months).

Figure 5.3

Conventional treatment alone versus conventional treatment plus biofeedback: treatment success at long term (12 months).

One parallel RCT conducted in the USA115 (1990) [EL=1+] determined whether outcome in chronically constipated and encopretic children with abnormal defecation dynamics could be improved with biofeedback training. The study included 43 children with chronic constipation and encopresis and abnormal defecation dynamics (33 boys, mean age 8.9 years, age range 5 to 16 years). Children were randomised to receive for 6 months conventional treatment alone (CT) (n=19) or conventional treatment plus biofeedback (BF) (n=22). CT consisted of use of laxatives, increase of dietary fibre and scheduled toileting. Disimpaction was carried out using enemas (type and dose not reported). For the maintenance phase children received magnesium oxide (milk of magnesia, MOM) at approximately 2 ml/kg body weight/day to induce at least one bowel movement daily and prevent faecal retention. Doses were decreased gradually to maintain daily bowel movement and prevent faecal retention and soiling. Children in the BF group received the same CT plus up to six sessions of biofeedback therapy 5 to 9 days apart. One session included approximately 30 to 35 defecation trials and lasted approximately 45 minutes. Patients in both groups were instructed to discontinue laxative therapy at 6 months ± 0.5 after initiation of therapy. Outcome measured was recovery rate at 7 and 12 months after initiation of treatment. Patients were considered to have recovered if they had 3 or more bowel movements per week and 2 or less soiling episodes per month while not receiving laxatives for 4 weeks.

At 7 months significantly more children in the BF group recovered compared to the CT group (BF (n=22): 12 [55%] versus CT (n=19): 1 [5%]; P < 0.001). Recovery rates did not differ between boys and girls in general and within the biofeedback group in particular. Prior unsuccessful treatment was not related to treatment outcome in either group. Patients with an initial abdominal faecal mass (severe constipation) were significantly more likely to recover with BF training than with CT alone (46% versus 0%, P < 0.02). At 12 months significantly more children in the BF group recovered compared to the CT group (BF: 11 [50%] versus CT: 3 [16%]; P < 0.05).

A boy aged 14 years in the BF group had a relapse. He had severe faecal impaction with enormous abdominal distension initially. Faecal impaction recurred 4 months after successful discontinuation of MOM. At the time the study was written he had no soiling but required intermittent treatment for constipation. One boy in the CT was lost to follow-up 1 month after treatment began. At that visit he was taking MOM and his soiling had resolved. One boy was lost to follow-up in the BF group after the first biofeedback session. Baseline characteristics were not significantly different between the two groups apart from gender: there were more girls in the BF group than in the CT group (41% versus 5%, P < 0.02). During initial evaluation severe constipation (an abdominal faecal mass present) was significantly more frequent in girls than in boys (90% versus 48%, P < 0.03). It was not completely clear who measured outcomes and how and whether questionnaires were piloted. Intention to treat analysis was not performed.

A parallel RCT conducted in the Netherlands116 (1996) [EL=1+] evaluated the effect of biofeedback training and conventional treatment on defecation dynamics and outcome in chronically constipated children. The study included 192 children with paediatric constipation (126 boys, median age 8 years). Patients were randomised to receive conventional laxative treatment alone (CT) (n=94) or conventional laxative treatment and biofeedback (BF) (n=98). Patients on CT received five outpatient visits lasting approximately 30 minutes during which laxative treatment and information from a diary containing defecation frequency and encopresis and/or soiling episodes were discussed. High fibre diet was advised but additional fibre supplements were not prescribed. Patients were instructed to try to defecate on the toilet for 5 minutes immediately after each meal. During the first 3 days patients were to use daily enemas (120 ml sodium-dioctylsulfosuccinate, 1 mg sorbitol, 250 mg per ml, Klyx) at home. If on day 3 enemas still resulted in large amounts of stool, they were continued for a maximum of 7 days. After the initial 3-day enema treatment, patients started oral laxatives with lactitol betagalactoside sorbitol (Importal®, Novartis) (one sachet of 5 g/10 kg body weight/day divided into two doses). Enemas were given whenever spontaneous defecation was delayed for more than 3 days.

Motivation was enhanced by praise and small gifts. Children in the BF group received five outpatient visits, including the same conventional treatment as described above, in combination with five biofeedback training sessions. As far as possible, both groups received equal attention. The treatment period lasted 6 weeks. Treatment was considered successful if the patients achieved 3 or more bowel movements per week and less than 2 soiling or encopresis episodes per month while not receiving laxatives for 4 weeks. Patients were assessed after the last visit of the intervention period at 6 weeks, then at 6 months, 1 year and 1½ years.

Treatment success was reported as number of children cured, and was not significantly different between the two groups at any of the assessment points (at 6 weeks, CT: 31 out of 94 [33%] versus CT+BF: 31 out of 98 [32%]; at 6 months, CT: 48 out of 93 [52%] versus CT+BF: 44 out of 94 [47%]; at 1 year, CT: 54 out of 92 [59%] versus CT+BF: 46 out of 92 [50%]; at 1½ year: CT: 52 out of 92 [57%] versus CT+BF: 44 out of 92 [48%]).

At baseline, patients were comparable for gender, age, frequency of gastrointestinal complaints and urinary problems. During the intervention period, three patients in the CT group refused manometry at the end of the treatment period: one patient was successfully treated and the parents refused permission for manometry; one patient was unsuccessfully treated and refused manometry; and one patient was lost to follow-up after two visits. Two patients of the BF group discontinued treatment: one patient aged 5 years did not cooperate and another patient discontinued treatment because his parents could not afford the cost of transport. At 6 months, five patients were lost (four patients in the CT+BF and one patient in the CT group), and at 1 year eight patients were lost to follow-up (two in the CT+BF and one in the CT group). Patients lost to follow-up were withdrawn from further analysis.

A parallel RCT conducted in Australia117 (1998) [EL=1+] determined whether surface electromyographic (EMG) biofeedback training produced sustained faecal continence in medical treatment resistant and/or treatment dependent children with anismus. The study included 29 children aged 4 years or more (24 boys, age range 4.8 to 14.9 years). Children were randomised to receive electromyographic biofeedback training and conventional medical treatment (BF) or conventional medical treatment alone (CT).

Up to four sessions of biofeedback were conducted at weekly intervals for each patient, each session consisting of approximately 30 to 35 defecation attempts. The aim was to achieve 10 relaxations of the external anal sphincter without visual feedback in two successive sessions. If this occurred in fewer than four sessions then biofeedback was discontinued. At completion of training, children were followed at monthly intervals by a single paediatrician, who gave verbal reinforcement of the skills learned during training.

CT alone comprised laxative therapy, behaviour modification and dietary advice. Laxative therapy occurred in two phases. The initial disimpaction phase comprised 3-day cycles of 5 ml Microlax enemas (sodium citrate) on day 1, one 5 mg bisacodyl tablet after school and 1 in the evening of day 2. Up to four cycles (12 days) were undertaken. Further cycles were prescribed if there was later evidence of stool re-accumulation. During the maintenance phase different laxatives were administered: liquid paraffin 5 to 30 ml once or twice a day, senna granules and or bisacodyl tablets. Medication use was decreased to a level consistent with maintenance of continence as monitored by bowel diary. Standard paediatric behaviour modification consisted of clarification during a joint parent-child interview of the postulates underlying physiological basis for encopresis.

The bowel training programme used positive reinforcement for successful defecation in the toilet and additional reinforcement for each 24 hours without soiling. Reinforcement consisted of parental praise and use of star chart diary (fitness training card) to indicate soiling free days. A regular sitting programme of 5 to 10 minutes toilet time within 30 minutes of each meal was basis of the programme. Dietary advice, general counselling and support were provided by a paediatrician. Psychiatric assessment or treatment was initiated when indicated clinically. It was unclear how long the CT lasted for.

Treatment success was assessed at 6 months after initiation of therapy. Full remission was defined as no medication and no soiling for at least 4 weeks; full remission on medication was defined as on medication and no soiling for at least 4 weeks; Partial remission defined as soiling no more than once a week, regardless of medication used. The use of medication was attempted by all those not in full remission, not only those who were worse or not improved. The remainder were those who were soiling more than once a week, regardless of medication use. Improvement was defined as progression by at least one level from baseline status, but without achieving full remission.

There were no significant differences between both treatment groups regarding the number of children who achieved full remission (BFT+CT (n=14): 2 [14%] versus CT (n=15): 2 [13%]; 95% CI on difference −24% to 26%). There were no significant differences between both treatment groups after combining the number of children who achieved full remission and the number of children who improved (BFT+CT: 2 (14%) versus CT: 4 (27%); P = 0.7, 95% CI on difference −46% to 23%). Three out of 14 patients in the BFT group completed the training in three sessions and the remainder underwent four sessions. Only one patient was unable to demonstrate relaxation of the external anal sphincter with attempted defecation. Only one patient (same as previous) was unable to defecate the biofeedback balloon by the time of their final session. All patients complied well with the instructions and procedures involved in the training. Two patients complained of transient discomfort when the biofeedback apparatus was inserted. No other adverse effects were seen or reported. At baseline there were slightly more subjects with primary encopresis in the biofeedback group than in the control group. No attrition or loss to follow-up was reported. It should be noted that no definition of constipation was given and also the study included a very small number of children.

Laxatives versus laxatives plus behavioural intervention versus laxatives plus behavioural intervention plus biofeedback

A parallel RCT conducted in the USA118 (2002) [EL=1+] compared short- and long-term effectiveness of three additive treatment protocols in children experiencing chronic encopresis. The study included 87 children aged 5 to 15 years who had experienced encopresis for a minimum of 6 months, defined as at least weekly episodes of faecal soiling for at least 6 months (72 boys, mean age at time of enrolment 8.6 ± 2.0 years, age range 5 to 13 years). Children were randomised to receive intensive medical therapy (IMT), intensive medical therapy plus enhanced toilet training (ETT) or intensive medical therapy plus enhanced toilet training and anal sphincter biofeedback (BF).

In the IMT group one of two paediatric gastroenterologists directed the treatment: colonic disimpaction with a series of enemas followed by sufficient laxative therapy to produce at least one soft stool each day without associated pain. Laxatives prescribed were magnesium oxide (milk of magnesia, MOM) and/or senna. Laxative dosages were adjusted regularly to produce one to three soft bowel movements daily. An enema or suppository was administered if the child had not produced a bowel movement during a 48-hour period. No specific dietary recommendations or manipulations were undertaken. Families received instructions and a brochure detailing the treatment protocol and the need for children to attend the toilet at least twice daily, preferably after breakfast and supper.

Children in the ETT group received similar enema and laxative therapy, with a clinical psychologist adjusting the laxative dose. The only difference from the previous therapy was that laxative therapy was decreased gradually when children demonstrated a stable bowel frequency with no soiling episodes. As long as the child had daily bowel movements of normal size for a week, the laxative dose was decreased by one quarter. This process was continued until laxative therapy was discontinued. If the child did not pass daily bowel movements of normal size, the laxative dose was increased. Parents and child were instructed on the psychophysiology of constipation and encopresis, and on how responding to early rectal distension cues along with regular toileting was critical to avoid reimpaction and to establish regular bowel habits. Various incentive programs were established, depending on the developmental age and the motivation of the child. Target behaviours were spontaneous trips to the toilet and clean pants. Toilet training was ‘enhanced’ because instructions were given on the role of paradoxical constriction of the external anal sphincter, and because appropriate defecation straining was modelled. The therapist sat on a portable toilet and demonstrated how to relax the legs and feet, how to take in a deep breath and hold it while sitting up straight, and how to push down with the held breath and pull in from the lower abdomen to propel out a stool. The child then replicated this while sitting on a portable toilet. The child received ‘hand feedback’ by placing one hand on the abdomen just below the navel to feel the abdomen move out when the breath was pushed down, and placing the second hand just below the first to feel inward movement with contraction of the rectus abdominous. Parents were instructed to prompt these behaviours at home. Additionally, 8 to 12 minutes of ‘toilet time’ was scheduled daily, beginning 15 to 30 minutes after the same two meals. During these times, children were instructed to practice tensing and relaxing the external anal sphincter for the first 4 minutes, with the objective of localising control of, and fatiguing, the external anal sphincter, and to mechanically stimulate the rectum. To desensitise children to toilet sitting, the second 4 minutes were spent ‘having fun’ while being read to or playing games. During the final 4 minutes, the child was to strain and attempt to have a bowel movement while relaxing his or her legs and feet. This routine toilet sitting was discontinued 2 weeks after the last scheduled treatment session.

The third group received the same instructions given to the other two groups and simultaneously received surface electromyographic biofeedback training. The same two psychologists who worked with the ETT group also worked with the BF group. It was unclear how long each of the treatments lasted. Data concerning toileting habits were collected for 14 consecutive days, before and after the initial outpatient visit, and again at 3 months, 6 months and 12 months after initiation of therapy. Treatment was considered successful if the child experienced no episodes of faecal soiling during the 2-week assessment 12 months after initiation of therapy.

There were no significant differences between the three groups at any time regarding mean soiling frequency (at 3 months: IMT 0.54, SD 0.68 versus ETT 0.22, SD 0.21 versus BF 0.34, SD 0.51; at 6 months: IMT 0.44, SD 0.52 versus ETT 0.38, SD 0.45 versus BF 0.20, SD 0.26 and at 12 months: IMT 0.33, SD 0.48 versus ETT 0.36, SD 0.53 versus BF 0.27, SD 0.37). At 3 months, 6 months and 12 months, the number of children who responded in the ETT group was significantly greater than in either the IMT or the BF group (at 3 months: IMT 45% versus ETT 85% versus BF 61%; at 6 months: IMT 41% versus ETT 74% versus BF 58%; and at 12 months: IMT 41% versus ETT 78% versus BF 61%; P < 0.05). These results were very stable over time (P < 0.001). With all three regimens, the response to treatment during the first 2 weeks of therapy strongly correlated with response to treatment at 3, 6 and 12 months (r > 0.90, P < 0.0001 in all cases). Of those children who had significant improvement after 2 weeks of therapy, 86 continued to improve at 3 months, 83 at 6 months and 81 at 12 months.

There were no significant differences between the three groups in the number of children cured at 12 months (IMT: 10 out of 29 [34.5%] versus ETT: 12 out of 27 [44.4%] versus BF: 11 out of 31 [35.5%]). There were no significant differences between the three groups at any time regarding the number of bowel movements passed in the toilet each day (mean at 3 months: IMT 1.44, SD 0.57 versus ETT 1.21, SD 0.49 versus BF 1.25, SD 0.64; at 6 months: IMT 1.36, SD 0.61 versus ETT 1.31, SD 0.63 versus BF 1.12, SD 0.60 and at 12 months: IMT 1.30, SD 0.61 versus ETT 1.01, SD 0.51 versus BF 1.16, SD 0.67). There were no significant differences between the three groups at any time regarding self-initiated toileting each day (mean at 3 months: IMT 1.53 times/day, SD 0.77 versus ETT 1.62 times/day, SD 0.82 versus BF 1.40 times/day, SD 0.71; at 6 months: IMT 1.49 times/day, SD 0.60 versus ETT 1.67 times/day, SD 0.95 versus BF 1.34 times/day, SD 0.72 and at 12 months: IMT 1.40 times/day, SD 0.76 versus ETT 1.31 times/day, SD 0.83 versus BF 1.31 times/day, SD 0.69). There were no significant differences between the three groups regarding laxative use at 12 months (IMT: 17 out of 29 [58.6%] versus ETT: 9 out of 27 [33.3%] versus BF: 17 out of 31 [54.8%]). There were no significant differences in baseline clinical or demographic characteristics between the three groups. It should be noted that no definition of constipation was given and no sample size calculation was performed.

One parallel RCT conducted in Croatia119 (2002) [EL=1+] assessed the success of biofeedback method versus conventional method in the treatment of chronic idiopathic constipation in childhood over a 12-week period and followed up the effect of biofeedback treatment on defecation dynamics and other anorectal manometric parameters in 49 children aged over 5 years (27 male) with chronic idiopathic constipation. Children were randomised to receive conventional treatment alone (CT, n=24) or conventional treatment plus biofeedback (BF, n=25). Conventional treatment consisted of oral administration of lactulose (Portalak®, Belupo) (240 mg/day or 10 ml syrup) with dose titration for the patient to have at least three stools per week. When spontaneous defecation failed to occur for more than 3 days in spite of appropriate therapy an enema was used. In addition, a fibre-rich diet and attempting defecation after meals were advised.

Biofeedback was conducted using a pressure technique. The child and the parents were instructed on how to perform Kegel exercises at home. Exercises included alternating 10-second contraction and relaxation of the sphincter and pubo-rectal muscle, performed five times a day in 20 cycles. Treatment lasted for 12 weeks. Treatment was considered successful if a frequency of 3 or more stools per week and less than 2 episodes of soiling or encopresis per month were achieved without laxatives. Therapeutic success was evaluated by the use of questionnaires distributed on weekly visits.

The number of children cured was significantly higher in the BF group compared to the CT group (BF: 21 out of 15 [84%] versus CT: 15 out of 24 [62.5%], P < 0.05). All children completed treatment. There were no significant differences in baseline characteristics between the two groups. It should be noted that the study included a small number of children and no sample size calculation was performed. There were insufficient details reported on who measured the outcomes and how they were measured.

One retrospective cohort study conducted in the USA120 (1995) [EL=2+] evaluated whether patients who received biofeedback treatment continued with improved outcome compared with patients who received conventional treatment alone. The study included 129 (97 boys) aged 5 to 18 years with chronic constipation and encopresis (1 or more soiling episode per week). One group received conventional treatment plus biofeedback (BF) and the other group received conventional treatment alone (CT).

At least two and up to six weekly training biofeedback sessions were given. Each session included approximately 30 to 35 defecation trials and lasted approximately 45 to 60 minutes. The number of training sessions given depended on how soon the child learned to relax the external sphincter. Sessions stopped after ten relaxations of the external sphincter could be accomplished without visual feedback in each of two successive training sessions. CT comprised the use of laxatives, increase of dietary fibre and scheduled toileting (child instructed to defecate for 5 minutes after each meal and after returning from school for the initial months, and try to defecate at least daily once they could recognise the urge to defecate). Disimpaction was carried out with enemas (type and dose not reported). For maintenance magnesium oxide (milk of magnesia, MOM) was administered at approximately 2 ml/kg body weight daily to induce at least one bowel movement daily and prevent faecal retention. Doses were decreased gradually to maintain daily bowel movement and to prevent faecal retention and soiling. Occasionally mineral oil or senna were used instead of MOM. It was unclear how long the CT lasted for.

The follow-up period for the CT group was 4.2 years ± 2.5 and for the BF group it was 4.1 years ± 2.4. The mean age of the CT group initially was 9.1 years ± 3.3 and at follow-up 13.4 years ± 3.3; of the BF group initially it was 10.4 years ± 3.2 and at follow-up 14.5 years ± 3.3. Patients were considered to have recovered if they had 3 or more bowel movements per week and 2 or fewer soiling episodes per month while off laxatives for at least 1 month.

There were no significant differences between groups in any of the outcomes measured (mean stool frequency per week: BF (n=63) 5 ± 3 versus CT (n=66) 6 ± 3; percentage of children soiling: BF 35% versus CT 24%; mean soiling frequency per week: BF 1 ± 2 versus CT 1 ± 2; recovery rate: BF 28 children [44%] versus CT 41 children [62%]; and laxative use: BF 25% children versus CT 18% children).

Of 64 patients who originally received biofeedback one patient did not return after the first unsuccessful biofeedback session and was lost to follow-up. The 63 patients included in the biofeedback group were combined from two studies (as clinical characteristics of both groups were similar): 21 patients from one RCT (included already in this review, see Loening-Baucke, 1990) and 42 patients who had not recovered after at least 6 months of conventional treatment. Twenty-three patients had been originally included in the RCT but one boy was lost to follow-up after the first biofeedback session and a second patient received a central nervous system shunt during the follow-up period and was excluded from the analysis. Baseline characteristics were comparable between both groups except for the presence of an abdominal faecal mass (BF: 60 children versus CT: 41 children; P < 0.05). Age and follow-up age were not related to outcome in either group. The length of follow-up was significantly related to recovery for the biofeedback group (P < 0.02) and for all patients (P < 0.01) but showed no relationship for the conventionally treated group.

Conventional treatment alone versus conventional treatment plus behavioural intervention

A parallel RCT conducted in the Netherlands121 (2008) [EL=1+] evaluated the clinical effectiveness of behavioural therapy with laxatives compared with conventional treatment in treating functional constipation in childhood. The study included 134 children (76 boys) with functional constipation aged 4 to 18 years referred to a gastrointestinal outpatient clinic. Children were randomised to receive conventional treatment alone (CT, n=67) or laxatives and behavioural therapy (BT, n=67). All children received treatment for disimpaction with daily Klyx enemas (sodium-dioctylsulfosuccinate and sorbitol, 60 ml/day for children aged 6 years or under; 120 ml/day for children aged over 6 years) for 3 consecutive days before starting treatment. During the maintenance phase children received PEG 3350, 1 sachet (10 g) per day, and if treatment was considered to have insufficient effect the dose was increased by one sachet. If spontaneous defecation was delayed for more than 3 days, parents were advised to give an enema or bisacodyl suppository of 5 mg. In the BT group it was preferred to give oral bisacodyl tablets of 5 mg instead of rectal laxatives. During BT, paediatric psychologists adjusted the laxative dose and consulted the paediatric gastroenterologist when necessary. In both treatment groups patients kept a bowel diary.

The protocolised BT was developed by paediatric psychologists of the authors' hospital. The protocol consisted of two age-related modules: a module for children aged 4 to 8 years and a module for children aged 8 years and over. The learning process for the child and the parents was based on five sequential steps (know, dare, can, will and do). This approach was derived from a multidisciplinary BT to treat children with defecation disorders. For all involved psychologists, a detailed manual for both age-related modules was available to ensure a standard delivery of therapy. Visits lasted approximately 45 minutes.

Conventional treatment was conducted by paediatric gastroenterologists. Visits lasted approximately 20 to 30 minutes when laxative treatment and bowel diary were discussed. Patients and their parents received education to explain that symptoms are not harmful and are common in children with functional constipation and that a positive, non-accusatory approach is essential. Children were instructed not to withhold stools when they felt the urge to defecate. Motivation was enhanced by praise and small gifts from the paediatric gastroenterologists.

For both t groups a total of 12 visits were scheduled during 22 weeks with similar intervals between treatment sessions. Children were assessed at the last visit (post-treatment time point) and 6 months after the 22-week treatment ended (follow-up). The time between baseline assessment and follow-up was approximately 1 year. Treatment was considered successful if patients achieved a defecation frequency of 3 or more times per week and a faecal incontinence frequency of once every 2 weeks or less, irrespective of laxative use. A secondary outcome measured was stool withholding behaviour.

Compared with the BT group, defecation frequency in the CT group was significantly increased (incidence rate ratio (IRR) = 0.75, 95% CI 0.59 to 0.96; P = 0.021). This effect was mainly caused by a difference between interventions at post-treatment (CT: mean 7.2, 95% CI 6.1 to 8.5 versus BT: mean 5.4, 95% CI 4.3 to 6.7) and not at follow-up (CT: 6.6, 95% CI 5.0 to 8.8 versus BT: 5.3, 95% CI 4.4 to 6.3). There was no statistically significant difference between both treatment groups regarding faecal incontinence per week (post-treatment CT: mean 2.1, 95% CI 0.8 to 5.8 versus BT: 5.0, 95% CI 2.1 to 12.0; follow-up CT: mean 6.4, 95% CI 3.5 to 11.7 versus BT: 8.6, 95% CI 4.0 to 18.3; IRR=2.36, 95% CI 0.77 to 7.31; P = 0.135). At post-treatment, success rate was higher in the CT group than in the BT group (CT 62.3%, 95% CI 51.1 to 76.1 versus BT: 51.5%, 95% CI 39.7 to 66.9). However, no statistically significant difference between treatments was found (IRR=0.83, 95% CI 0.60 to 1.14; P = 0.249). At follow-up, the number of children successfully treated declined in both groups but again the difference was not statistically significant (CT: 57.3%, 95% CI 46.6 to 70.4 versus BT: 42.3%, 95% CI 31.8 to 56.4; IRR=0.74, 95% CI 0.52 to 1.05; P = 0.095). There were no significant differences between both treatment groups in the proportion of children who exhibited stool withholding behaviour at follow-up.

It should be noted that during treatment 2 out of 64 (3.1%) in the CT group and 9 out of 65 (13.8%) in the BT group discontinued the intervention (P = 0.054). At follow-up, four patients dropped out in CT. There was one loss of contact and three children were referred for BT directly after CT, making them unsuitable for follow-up measurements. Questionnaires were not returned by three patients in both intervention arms at post-treatment and by nine patients (CT six, BT three) at follow-up. Except for painful defecation (65.0% CT versus 43.1% BT, P = 0.014), there were no significant differences between the two groups in baseline sociodemographic factors or for clinical characteristics. An intention-to-treat analysis was conducted. Because of withdrawal before treatment start, attrition during the study, failure to fill out questionnaires or research procedure violations, missing data occurred. Imputation of missing values was used to make intent-to-treat analyses feasible.

Behavioural intervention plus laxatives versus laxatives only

A small parallel RCT (multicentre) conducted in the USA122 (2003) [EL=1+] examined the utility and effectiveness of an internet-based version of enhanced toilet training. The study included 24 children aged 6 to 12 years, soiling at least once a week, who had no medical diagnosis other than constipation that could explain their faecal incontinence (19 boys, mean age 8.46 years, SD 1.81). Children were randomised to receive the web intervention (n=12, 10 boys) or no intervention (n=12, 9 boys). All children were instructed to start with a basic regimen of one square of senna (Ex-Lax®, Novartis) twice a day.

The intervention was a web-based programme for the treatment of paediatric encopresis (U-CAN-POOP-TOO). This was a child-focused programme which targeted primarily at children aged 5 to 10 years but designed to be used by the child and the parents together. The program comprised three core modules which took 60 to 90 minutes to complete, with all users instructed to review them during the first week. The modules were: ‘The body’ (anatomy, physiology and pathophysiology of digestion), ‘How to poop’ (behavioural techniques for treatment of encopresis) and ‘Medication’ (clean-out and laxative treatment). New modules were assigned each week based on a follow-up assessment completed by the user about their child's status. Not all modules were necessarily used by all users: only those modules identified as relevant were assigned and reviewed. However, all modules could be viewed by all users. Follow-up comprised 17 to 20 questions, depending on the week. The system contained a total of 22 modules, each taking 5 to 10 minutes to review. Exposure to the program lasted for 3 weeks after which an assessment was conducted.

The number of faecal accidents per week decreased significantly more in the web group compared with the group with no web intervention (web group: mean 0.50, SD 0.85 versus no web: mean 8.27, SD 13.83). The number of bowel movements passed in the toilet per week increased significantly more in the web group compared to the no-web group (change from pre- to post-assessment:152% versus - 16%; P = 0.001). Using the bathroom without prompts also increased significantly more in the web group compared to the no-web group (change from pre- to post-assessment: 109% versus -37%; P = 0.021). Using the bathroom with prompts was not significantly different between the two groups. There were no significant differences in baseline characteristics between the two groups: age, gender, race, stage of bowel movement training, length of current laxative regimen or any of the outcomes measured. No dropouts or children lost to follow-up were reported. It should be noted that the study included a very small number of children.

Laxatives plus behaviour modification versus laxatives plus behaviour modification plus psychotherapy

A quasi-RCT conducted in the UK113 (1986) [EL=1+] reported the authors' experience with children who presented with faecal soiling, with or without constipation, who were treated by incentive-based behavioural modification, with or without psychotherapy, and consider factors that might predict the outcome for a non-intensive approach and in particular to draw attention to social background as a prognostic indicator. The study included 47 children who presented with faecal soiling, with or without constipation (26 boys, age not reported). For all children in cases where constipation was severe with large faecal masses they were initially admitted to the ward. They were then continued on whatever laxative they had been on before referral. Where no laxative had previously been used the child was offered a dose of lactulose twice daily (amount not reported). If there was no accumulation of faeces no laxatives were prescribed. No other laxatives were used in this study and in general their use was minimised, with the parents encouraged to stop the treatment with laxatives as soon as a regular bowel habit was established. In none of the children were suppositories used at any time. All the children were encouraged to take a high residue diet and in particular were asked to take bran with their breakfast cereal.

Children were randomised to receive behaviour modification (BM, n=26) only or behaviour modification plus psychotherapy (BM+Psy, n=21). BM was carried out by a paediatrician. All children were placed on a star chart regimen and offered varying coloured stars for ‘sitting on the toilet’ and ‘remaining unsoiled for a full day’. In some cases stars were awarded to encourage children who were reluctant to take bran in their diet. A contract was negotiated between the child and the parent (usually the father) for an award to be made at the discretion of the paediatrician. The child was to understand that the giving of the award would depend on response to treatment. ‘Demystification’, alleviation of guilt and use of explanatory diagrams were used. Children were seen at intervals of 6 weeks by the paediatrician for between 3 months and 1 year and were subjected to shows of affection and interest, which included careful and serious inspection of the charts. Failure to keep a star chart on two successive visits resulted in firm statement of displeasure. Two further failures led to the stopping of treatment and discharge with the option of psychiatric referral. Discharge of cured patients was at discretion of the parents.

Children in the BM+Psy group received the same BM as previously described. In psychotherapy children were seen by the child psychiatrist at roughly monthly intervals for between 2 and 12 months. At each appointment the mother (and also the father in four cases) was seen for 15 to 30 minutes to explore her feelings in respect of the child's bowel problem and its effect on the family and on her own relationship with the child. Whenever possible the mother's own history was explored and other emotional problems discussed where relevant, such as expressions of grief, anger or depression. The child was seen for 15 to 30 minutes for play, including picture drawing, games and sharing of their own toys and belongings. Their feelings concerning their problem were also explored. The behavioural star chart was also often brought and reviewed and the child praised and encouraged according to progress. The mother and child were seen together, sometimes early in treatment, sometimes later, depending on their relationship and success with management of the problems, to assess overall progress.

One year after initiating treatment success was assessed. Children were considered cured if they had at least five normal stools each week without soiling and only occasional use of laxatives (less than once a week). Children were considered improved if they had at least three stools each week and soiling less than once a week. Non-responders were children who had less than three stools each week or soiling more than once a week. These children were considered as failing to improve, despite the fact that in most cases there was less soiling than at the beginning of treatment. Treatment success did not differ between the groups. It is not possible to report the figures here, as they were only analysed by the authors according to compliance with treatment and with the children's social class, but not according to treatment groups. Four children left the study and 13 failed to keep adequate 'star charts'. Two children were subsequently found to be cured. It should be noted that no definition of constipation was given. Additionally the study included a small number of children and no sample size calculation was performed.

Systemic/family therapy: externalising versus behavioural approach

A retrospective audit conducted in the UK123 (1998) [EL=3] aimed to assess the effectiveness of externalising treatment (EXT) compared to other traditional treatments (OTH) in children with soiling problems. The audit included 108 children treated for soiling problems (45 aged 3 to 5 years, 63 aged over 6 years) and their families. Referrals included ‘faecal soiling’, ‘encopresis’, ‘psychological soiling’, ‘failed toileting’, ‘constipation with overflow’ and ‘deliberate soiling’. It should be noted that some children were clearly diagnosed in the referral letter as ‘constipated’ or ‘not constipated’, but in some referral letters it was not stated whether the referring doctor had checked for constipation. Families who received EXT (n=54) were only included in the study if the treatment approach included: externalising the poo from the first interview with the child and family; developing a narrative with the child and family where they could see themselves as capable, skilful and determined ‘to teach the poo a lesson’, ‘outwit the poo’ or ‘defeat the poo’; not using rewards, interpretation, confrontation or paradoxical interventions as therapeutic manoeuvres; and attempting to see the whole family at least once. Other treatments (OTH) (n=54) included a mixed group of traditional treatments with predominantly (but not only) a behavioural approach in a family systems context. There were no elements of externalising in any OTH sessions. The treatment given depended only on the current approach of the therapist who received the referral.

Treatment lasted an average of 7.8 months for the EXT group and 6.6 months for the OTH group. At a minimum of 6 months' follow-up (mean 23 months), all parents (including those who left the study) were sent a questionnaire and asked whether there had been any further soiling incidents since they were last seen and the frequency of these incidents in the past month. Where children had returned for paediatric consultation, the frequency of soiling stated in paediatric notes was recorded even if parents did not reply to the audit. GPs were also asked whether they were aware of any further soiling after treatment had ended.

Significantly more children who received EXT stopped soiling or improved compared to children who received OTH, however this outcome was assessed (from notes: EXT 42 out of 47 versus OTH 30 out of 40, P = 0.02; from GP follow-up: EXT 29 out of 37 versus OTH 24 out of 42, P = 0.045; from parent follow-up: EXT 24 out of 38 versus OTH 13 out of 35, P = 0.026). Significantly more parents assessed EXT as helpful compared to OTH (number of parents: 24 versus 10; P = 0.0001). Externalising proved to be superior for boys, for children aged 6 years or over, for those with frequent soiling at the outset, for those with over 2 years' continuous soiling and those diagnosed as constipated on referral. The average number of appointments was not significantly different between the groups. There were no significant differences between the groups on baseline variables. It was unclear exactly how many children left the study or were lost to follow-up.

Evidence statement

One meta-analysis of four RCTs [EL=1+] showed that there were no significant differences between conventional treatment plus biofeedback and conventional treatment alone (including use of laxatives, advice on a high-fibre diet and attempting defecation after meals) at increasing the frequency of bowel movements and decreasing the frequency of soiling in children with chronic constipation both at medium term (6 months) and long term (12 months).

One RCT [EL=1+] showed that in the short term (12 weeks) conventional treatment plus biofeedback was more effective than conventional treatment alone (including laxatives, advice on a high-fibre diet and attempting defecation after meals) at increasing the frequency of bowel movements and decreasing the frequency of soiling in children with chronic constipation.

One retrospective cohort study [EL=2+] showed that after 4 years there were no significant differences between children with chronic idiopathic constipation who received conventional treatment plus biofeedback treatment and children who received conventional treatment alone (including use of laxatives, increase of dietary fibre and scheduled toileting) regarding stool frequency, proportion of children soiling, soiling frequency, recovery rate and proportion of children using laxatives.

One RCT [EL=1+] showed that there were no significant differences between a protocolised behavioural therapy conducted by paediatric psychologists (including teaching parents behavioural procedures) along with use of laxatives and conventional treatment conducted by paediatric gastroenterologists (including laxatives, discussion of bowel diary, education on symptoms of constipation, instructions to not withhold stools and use of motivation enhancers) regarding frequency of faecal incontinence and proportion of children who exhibited stool withholding behaviour. Conventional treatment was significantly more effective than behavioural therapy at increasing defecation frequency, but overall success rate was not significantly different between the two treatment groups.

One RCT [EL=1+] showed that there were no significant differences between intensive medical therapy (including laxatives and attempting defecation after meals) and intensive medical therapy plus enhanced toilet training (including modelling of appropriate defecation straining) or intensive medical therapy plus enhanced toilet training plus anal sphincter biofeedback at decreasing soiling frequency and the proportion of children using laxatives, and at increasing the number of bowel movements passed in the toilet each day and the proportion of children who self-initiated toileting each day. There were no significant differences between the three groups in the overall number of children cured.

One RCT (multicentre) [EL=1+] showed that an internet-based version of an enhanced toilet training programme for the treatment of paediatric encopresis plus laxatives was more effective than laxatives alone at decreasing the number of faecal accidents per week, increasing the number of bowel movements passed in the toilet per week and increasing the use of the bathroom without prompts. Using the bathroom with prompts was not significantly different between the two groups. Most parents found the material understandable and easy to use, and believed their child liked the program and found it understandable and easy to use.

One retrospective audit [EL=3] showed that externalising treatment was more effective than traditional treatments with a predominantly behavioural approach in a family systems context at reducing the number of children experiencing soiling.

One quasi-RCT [EL=1+] showed that there were no significant differences between laxatives plus an incentive-based behavioural modification and laxatives plus an incentive-based behavioural modification plus psychotherapy at increasing frequency of bowel movement and decreasing frequency of soiling in children presenting with faecal soiling, with or without constipation.

No evidence was found on the following interventions:

  • intense psychotherapy: cognitive behavioural therapy (CBT) and psychodynamic psychotherapy.
  • psychosocial counselling
  • mediational models in cognitive or behavioural therapy
  • clinical hypnosis
  • toilet or bowel or habit retraining
  • ‘chaining’ and ‘shaping’ programmes
  • parenting programmes which clearly specify what the program is
  • portage as an educational model.

GDG interpretation of the evidence

The lack of evidence to support the effectiveness of psychological and behavioural interventions may be as a result of the patient selection in the studies included in the review. In these studies the children and families allocated to psychological and behavioural interventions did not appear to meet the usual criteria for psychological referral and in usual clinical situations would have been expected to do as well on laxative medication alone. The evidence therefore suggests that as a matter of routine, children with idiopathic constipation do not do any better when psychological interventions are added to laxative therapy as part of constipation management. The GDG felt that the research settings reported do not reflect clinical reality.

It is the experience of the GDG that many health professionals use behavioural advice as part of their routine practice, especially incorporating star charts into toileting routines. However, this is often initiated when the child is still constipated or not on an effective dose of laxative medication with the result that the child and family are set up to fail. As the child will continue to soil as a result of either overflow or lack of appropriate control this is then seen as a behavioural problem and referred on to psychological services where the involuntary soiling can be misinterpreted as a symptom of psychological distress. It is the view of the GDG that in the majority of children with idiopathic constipation any psychological problems are secondary to the symptoms of the constipation and not the cause.

It is the view of the GDG that psychological and behavioural interventions are effective only when the child is on effective laxative medication and when the outcomes sought are negotiated with both parent and child as being achievable. The advice given needs to be developmentally appropriate and child focussed. Based on their clinical experience, the GDG agreed that it is important not to blame the child for the constipation and any interventions should be non-punitive. Referral on to child and adolescent mental health services (CAMHS) for psychological issues related to idiopathic constipation in children may be beneficial and cost effective where there is psychological distress related to the symptoms of constipation and/or family difficulties that maintain or exacerbate the constipation.

Recommendations

Do not use biofeedback for ongoing treatment in children and young people with idiopathic constipation.

Do not routinely refer children and young people with idiopathic constipation to a psychologist or child and adolescent mental health services unless the child or young person has been identified as likely to benefit from receiving a psychological intervention.

5.6. Complementary therapies

Introduction

Many families consider the use of complementary and/or alternative therapies as a treatment option when conventional treatment ‘fails’.

The terms ‘alternative’ and ‘complementary’ are usually used to define the use and setting of a therapy in relation to orthodox medicine. ‘Alternative’ usually refers to treatment modalities that are generally a substitute for orthodox treatment whereas ‘complementary’ refers to treatments that are used alongside orthodox medical treatments.

There may be very little evidence about the efficacy of many complementary and alternative treatments but their use is widespread and increasing across the developed world. There is a clear need for more effective guidance for the public and health professionals who advise patients as to what does and does not work and what is and is not safe.124

Clinical question

What is the clinical effectiveness of the following complementary therapies for ongoing treatment and/or maintenance in children with chronic idiopathic constipation?

  • abdominal massage
  • reflexology
  • hypnotherapy
  • osteopathy
  • cranial osteopathy
  • craniosacral therapy
  • homeopathy.

Studies considered in this section

Studies were considered if they:

  • included neonates, infants or children up to their 18th birthday with chronic idiopathic constipation being treated with any of the following complementary therapies:

    abdominal massage

    reflexology

    hypnotherapy

    osteopathy

    cranial osteopathy

    craniosacral therapy

    homeopathy.

  • included the following outcomes:

    changes in frequency of bowel movements

    changes in stools consistency or appearance

    changes in pain or difficulty on passing stools

    changes in frequency of episodes of soiling

    reduction in laxatives use

    parent/child views or satisfaction or quality of life

  • were not case reports
  • were published in English.

No restrictions were applied on the publication date or country.

Overview of available evidence

A total of 119 articles were identified from the searches and 14 articles were retrieved for detailed assessment. Of these, one study, a prospective case series [EL=3], was identified for inclusion in this review.

Narrative summary

One prospective case series conducted in the UK125 (2003) [EL=3] aimed to investigate the efficacy of treating patients with encopresis and chronic idiopathic constipation with reflexology. The study included 50 children (age range 3 to 14 years, 64% boys) diagnosed with encopresis and/or chronic idiopathic constipation. All children received six sessions of reflexology, 30 minutes each, at weekly intervals for 6 weeks. Existing medications were unaltered. Frequency of bowel movements (BMs), soiling frequency and parents' attitude towards reflexology were measured before and immediately after treatment was completed. With the help of their parents, children completed questionnaires on bowel motions and soiling patterns before, during and after treatment whereas parents completed questionnaires on their attitude towards reflexology.

Frequency of soiling decreased after treatment was completed (children soiling at least daily: 78% versus 20%; 1 to 3 times per week: 16% versus 30%; and no soiling at all: 6% versus 48%; P < 0.05). Frequency of BMs increased after treatment (children having daily BMs: 18% versus 24%; between 1 and 4 BMs per week: 46% versus 72%; and no BMs per week: 36% versus 2%; P < 0.05). At the beginning of the study 70% of parents were keen to try the treatment and after the treatment was completed 72% were satisfied with the outcome. Baseline outcomes for two children who only attended the first session were reported but it is unclear whether they were also included in the final analysis.

Evidence statement

One prospective case series [EL=3] showed that reflexology was effective at decreasing the frequency of soiling and increasing the frequency of bowel movements in children with chronic constipation.

No published evidence was found on the effectiveness of the following complementary therapies for ongoing treatment and/or maintenance in children with chronic idiopathic constipation:

  • abdominal massage
  • hypnotherapy
  • osteopathy
  • cranial osteopathy
  • craniosacral therapy
  • homeopathy.

GDG interpretation of the evidence

Due to the lack of evidence of effectiveness or cost effectiveness, the GDG felt unable to make a recommendation for the use of complementary and alternative therapies for use in the NHS.

The GDG is aware that complementary therapies are frequently used in infants. Sometimes parents use them but feel unable to discuss their usage with health professionals. Certain complementary therapies are available on the NHS only in some areas, whereas in other areas parents pay for them. Current regulation of different complementary therapies (standards and training) varies.

It is the GDG's view that complementary therapies, such as massage, can encourage positive relationships between parents and children by promoting positive time spent together between them, but more research is needed to confirm this and other potential benefits in children with chronic idiopathic constipation.

Research recommendation

What is the effectiveness of complementary therapies (hypnotherapy) for ongoing treatment/maintenance in children with chronic idiopathic constipation?

Why this is important

Many families consider the use of complementary and or alternative medicine (CAM) as a treatment option when conventional treatment ‘fails’. There is very little evidence about the efficacy of many complementary and alternative treatments but the use of CAM is widespread and increasing across the developed world. There is a clear need for more effective guidance for the public and health professionals who advise patients as to what does and does not work and what is and is not safe.124 There is moderately good evidence for the effectiveness of hypnotherapy in improving global symptoms in adults with irritable bowel syndrome compared with attention control or symptom monitoring or usual management, mainly in patients with refractory IBS, both in primary and secondary care.126 The use of hypnotherapy may therefore be an effective intervention in children with chronic constipation that has not resolved with usual treatment and may offer an additional approach to treatment which works.

5.7. Antegrade colonic enema procedure

Introduction

Optimal medical management of children with chronic idiopathic constipation will tend to reduce the number requiring surgical intervention. However, for patients with chronic treatment resistant symptoms, surgical interventions may be considered.

The antegrade colonic enema (ACE) has now been demonstrated to have a role in the management of patients with constipation.127 The procedure involves the surgical formation of a fistula between the skin surface and the colon, most frequently the caecum. Washout fluid and enema solution can then be delivered into the bowel without recourse to the rectal route. The aim is to keep the colon clean and reduce soiling. Commonly the appendix itself may be mobilised to the surface to act as the fistula but a number of alternative techniques are also well accepted. The procedure may be performed as an open operation, laparoscopically or colonoscopically.

Central to success of the ACE is good case selection coupled with careful postoperative management. While patients should be considered for ACE after a period of optimal medical management, referral of appropriate patients should not be delayed unduly. Management of washouts and of the sequelae of the ACE procedure is vital if symptoms are to remain controlled. As a failure rate exists, there remains a need both for other interventions (including resection and stoma formation) and for ongoing research for this sub-group of patients.

This section discusses the place of the ACE in the management of children with constipation.

Clinical question

What is the effectiveness of the antegrade colonic enema (ACE) procedure in children with chronic idiopathic constipation?

Studies considered in this section

Studies were considered if they:

  • included neonates, infants or children up to their 18th birthday with chronic idiopathic constipation
  • included the antegrade colonic enema (ACE) procedure, regardless of what surgical technique was used
  • included the following outcomes:

    changes in frequency of bowel movements

    changes in stools consistency or appearance

    changes in pain or difficulty on passing stools

    changes in frequency of episodes of soiling

    reduction in laxatives use

    parent/child views or satisfaction or quality of life

  • were not case reports
  • were published in English.

No restrictions were applied on the publication date or country.

Overview of available evidence

A search was performed on pharmacological and surgical interventions for disimpaction and ongoing maintenance in children with chronic idiopathic constipation. A total of 986 articles were identified from this search and 143 articles were retrieved for detailed assessment. In addition, GDG members submitted 11 papers. Of these, six studies were identified for inclusion in this review (including one of the papers submitted by the GDG members): three retrospective cohort studies, one prospective case series, one retrospective case series and one retrospective survey.

Narrative summary

A retrospective cohort conducted in Australia127 (2005) [EL=2+] investigated whether antegrade colonic enemas (ACEs) are effective in idiopathic paediatric slow transit constipation (STC) in children. The study included 56 patients with appendicostomy for idiopathic constipation formed between January 1995 and October 2004, who satisfied Rome II criteria for functional constipation, with or without faecal incontinence and who had undergone a prolonged period of unsuccessful medical management. Data were available for 42 children only (31 boys, mean age at interview 13.1 years, median age 12.4 years, age range 6.9 to 25.0 years). Median initial regimens used for washouts were varied: polyethylene glycol 3350 and electrolytes (PEG 3350+E) (Golytely Braintree Laboratories Inc) (79%), liquorice (12%), water (2%) and other (7%).

The median regimen used at the time of interview was PEG 3350+E 500 ml to 750 ml administered every second day, infused over 10 to 20 minutes with no need for disimpaction. Defecation occurred 20 to 30 minutes after ACE had finished, with 20 to 30 minutes spent on the toilet. The majority of patients (25 out of 42, 60%) were either using the initial regimen or had tried one regimen change at the time the study was conducted. There was no correlation between the number of ACE regimens tried, patient satisfaction or the length of ACE usage.

Many families believed that regimen changes were a necessary response to increased tolerance to a particular ACE solution. Patient input into the ACE regimen varied: seven children (all older than 10 years) were completely independent, five children required supervision only, 15 needed help setting up and cleaning up and 15 were completely dependent. Thirty-seven children (88%) were very satisfied or satisfied with the procedure. Forty-one families (98%) said they would recommend the ACE to other children. Thirty-nine families (93%) felt there was a significant improvement in the quality of their child's life. Families felt that the mean optimal age for appendicostomy formation was 4.9 years (median 4 years, range 2 to 12 years).

Fifteen children (36%) had ceased ACE at the time of interview (mean period usage 2.6 years, range 0.7 to 5.8 years): symptoms resolved in seven children, in four a colostomy was formed, in two an ileostomy was formed and two patients returned to conservative management. The mean period of usage for children who had ceased ACE was not significantly different from those who were still using ACE at the time the study was conducted.

Regarding the ACE efficacy, there were significant improvements compared to baseline in both continence and quality of life, as well as significant reduction in soiling frequency, abdominal pain frequency and abdominal pain severity:

  • continence score: pre-ACE: mean 2.5, median 2, range 0 to 8; post-ACE: 5.2, 5, 1 to 12; P < 0.0001
  • quality of life score: pre-ACE: mean 1.4, median 1.5, range 0.5 to 3.0; post-ACE: 2.2, 2.5, 0.5 to 3.0; P < 0.0001
  • soiling frequency score: pre-ACE: mean 5.7, median 6; range 0 to 6; post-ACE: 3.0, 3, 0 to 6; P < 0.0001
  • abdominal pain severity score: pre-ACE: mean 7.4, median 8, range 0 to 10; post-ACE: 3.0, 3, 0 to 8; P < 0.0001
  • abdominal pain frequency score: pre-ACE: mean 5 days per week, median 6 days per week, range 0–6 to 3–6 days per week; post-ACE: mean 2.5 days per month, median 2.5 days per month, range 0–6 to 1–2 days per month; P < 0.0001.

Thirty (71%) children experienced symptoms at some stage of the treatment: cramping (18 out of 30), nausea (17 out of 30), vomiting (7 out of 30), sweating (14 out of 30), dizziness (10 out of 30) and pallor (10 out of 30). Three or more of these symptoms were present in 12 out of 30 patients. The three most common long-term complications were granulation tissue in 33 children (79%), anxiety about ACE in 21 children (50%) and stomal infection in 18 (43%). These were unresolved in 15%, 29% and 11% of patients respectively.

A retrospective cohort conducted in the UK128 (2004) [EL=2+] compared the results, complications and outcomes of the Malone antegrade colonic enema (MACE) with the caecostomy button (CB) in children with intractable constipation. The study included 49 children (15 boys) who underwent MACE or CB between June 1998 and August 2002 for intractable idiopathic constipation and faecal soiling that had failed conventional treatment. Thirty-seven children underwent MACE and 12 children underwent CB. Both groups started saline enemas (20 ml/kg) on the fourth postoperative day. Children not responding to saline wash-out used Klean-Prep. The frequency and volume of enemas were individualised to each patient to achieve cleanliness and stop soiling.

In 39 children (79.6%, 30 with MACE, 9 with CB) the soiling stopped completely. Occasional soiling was still present in three children (one with MACE, two with CB). One child with CB resumed regular activity and thus the CB was removed. MACE failed in six children (16.2%): in four patients the colonic washouts were ineffective, in one patient the colonic washouts were associated with abdominal pain during enema and one patient required revision for perforation of appendicostomy and the fibrotic-ischaemic appendix was replaced with a CB. CB failed in one patient (8.3%) due to leaking faecal content around the button which was converted to MACE after 20 months.

Surgical complications requiring operative intervention were significantly more frequent in children who underwent MACE compared to CB (MACE: 9 [24%] versus CB: 0; P = 0.009). Surgical complications not requiring operative intervention were significantly more frequent in children who underwent CB compared to MACE (MACE: 7 [19%] versus CB: 11 [92%]; P < 0.001).

A retrospective cohort129 (2006) [EL=2+] conducted in the USA reported the authors' 4 year experience with two different techniques of the caecostomy procedure compared the clinical outcome of caecostomy in children with defecation disorders secondary to idiopathic constipation, imperforate anus and spinal abnormalities. A total of 31 children (58% boys) who received the procedure due to the previous underlying disorders were included. Nine of the children had idiopathic constipation and a median age at time of caecostomy of 12 years (range 3 to 16 years).

The bowel movement frequency significantly increased after caecostomy (n=9; before caecostomy: less than 5 times per week versus post caecostomy: between 5 times per week and 3 times per day; P < 0.01). The soiling frequency, the number of medications used and the number of physician visits related to defecation problems all decreased significantly after caecostomy was performed (soiling before was constant versus post was none, P = 0.0; medications before 4, post 1; P = 0.01 and physician visits before 6 versus post 2; P < 0.01 respectively). No child was admitted to hospital for disimpaction after the procedure was performed (before 4, post 0; P < 0.01). The global health score and the global emotional score both improved significantly after the procedure (global health before: poor versus post: good; P = 0.01 and global emotional before: poor versus post: good; P = 0.01 respectively) Children also experienced significantly less limitation of activity (before: moderate versus post: mild; P < 0.01).

No subgroup analysis was performed for the type of antegrade enemas used; therefore these outcomes are not reported here. There were no significant differences in relation to the number of missed school days per month before and after the procedure. There were no major complications such as perforation, stoma stenosis or stoma prolapse. No difference was found in occurrence of number of complications between different procedures and/or techniques. Other outcomes are not reported here as no subgroup analysis was performed.

One prospective case series in the UK130 (2009) [EL=3] analysed the outcomes of ACE procedure in children with idiopathic constipation who had not responded to 3 years of medically supervised conservative management. Eighty children with idiopathic constipation undergoing ACE surgery by one surgeon were included. The lavage regimen used a saline solution prepared at a volume of 20ml/kg body weight and was supervised by specialist nurses. Children were followed up in a nurse-led continence clinic over a period of 6 months to 10 years (median 6.2 years).

Outcome measures were: ongoing lavage, failure (cease technique because lavage did not improve bowel habit or colon not lavagable) and cure (appendicostomy closed/reversed because of child achieving normal bowel habit). Of the 80 children included, 53 had conventional ACE surgery and 27 had laparoscopic surgery. ACE lavage failed in 12 children (Kaplan Meier probability - 0.3 at 8.5 years). ACE lavage provided cure for 12 children (Kaplan Meier probability – 0.2 at 6.2 years), all of whom went on to have their appendicostomy closed. Gender was significantly associated with ACE failure (P = 0.04) with a higher failure rate amongst girls (P = 0.02). Colonic transit time (CTT), age at surgery and duration of follow-up were not significantly associated with ACE failure. CTT was a significant factor in predicting failure in children accommodating a very large volume of lavage fluid (>10L) in their colon without bowel evacuation. No patients were discharged from the study and none were lost to follow-up.

A small retrospective case series conducted in the USA131 (2002) [EL=3] assessed the benefit of ACEs through caecostomy catheters in children with severe constipation. Twelve children (nine boys, mean age: 8.7 ± 4.4 years) referred to a tertiary care motility centre for further evaluation of intractable constipation, who had undergone caecostomy placement for administration of antegrade enemas were included. After the procedure children significantly improved in relation to all the outcomes measured: bowel movements/week increased (before: 1.4 ± 0.7, after: 7.1 ± 3.8; P < 0.005), soiling episodes/week decreased (before: 4.7 ± 3.2; after: 1.0 ± 1.4; P < 0.01), the number of medications used for constipation decreased (before: 4.0 ± 1.0, after: 0.8 ± 0.6; P < 0.005) and children suffered less severe abdominal pain (score before: 2.9 ± 1.6, after: 0.9 ± 1.0; P < 0.005). Parents also considered that both the emotional and the overall health of their children had improved following the procedure (emotional health score before: 1.9 ± 0.8, after: 3.6 ±1.1; P < 0.005); (overall health score before: 1.7 ± 0.9, after: 3.6 ± 0.9; P < 0.005; respectively). Children missed fewer school days every month (before: 7.5 ± 6.9, after: 1.5 ± 2.5; P < 0.02) and had to attend fewer physician office visits per year (before: 24.0 ± 19.1, after: 9.2 ± 14.2; P < 0.05).

The choice of irrigation solution used after caecostomy varied based on preference of the treating physician. Most patients began with low volume infusions of solution, which were increased according to therapeutic response. Sixty-seven percent of patients used 200ml to 1,000ml (mean 478 ml ± 262 ml) polyethylene glycol irrigation solution, daily to every other day. Twenty-five percent of patients used a combination of saline and glycerine, mixing 60ml to 75ml of glycerine in 240ml to 300ml of saline. One patient received 90ml phosphate soda solution followed by 300ml of saline. Evacuation occurred within one hour of enema administration in seven children and occurred within three hours in the other five children. No comparisons were made between the different solutions used. There were no acute adverse events and only four children experienced postoperative adverse events: skin breakdown and development of granulation tissue (n=1), leakage of irrigation solution (n=1) and accidental removal of the catheter with subsequent easy catheter replacement by the interventional radiologist (n=2). No adverse event led to discontinuation of the antegrade enema use. No child required admission to hospital because of faecal impaction since starting antegrade enemas. Five patients discontinued antegrade enemas with removal of the caecostomy at a mean of 14.6 ± 9.1 months after beginning treatment. None of these children redeveloped problems with constipation or faecal soiling.

A retrospective survey conducted in the UK132 (1999) [EL=3] followed up the success of the MACE procedure. Fifty-eight children who underwent a MACE performed by UK members of the British Association of Paediatric Surgeons (or their units) up to the end of 1996 were included. Children who took part in a previous study conducted by the same authors as well as reported figures from one other UK centre were also included making a total population of 273 children. Patients (mean age 12.3 years) were followed up on average 2.4 years (range 0.3 to 6). Success criteria were defined as full success (totally clean or minor rectal leakage on the night of the washout); partial success (clean, but significant stomal or rectal leakage, occasional major leak, still wearing protection but perceived by the child or parent to be an improvement) and failure (regular soiling or constipation persisted, no perceived improvements, procedure abandoned usually to a colostomy). Twenty-three patients had been diagnosed with constipation. In these patients a full success rate was seen in 52%, partial success in 10%, the procedure was considered a failure in 38% and its outcomes were unknown in 1%. Main complications of the procedure were not reported in relation to the clinical diagnosis and therefore are not included here.

Evidence statement

Three retrospective cohorts [EL=2+], one retrospective case series [EL=3] and one retrospective survey [EL=3] showed that the frequency of episodes of soiling significantly decreased after ACE was performed.

Two retrospective cohorts [EL=2+] and one retrospective case series [EL=3] showed that frequency of bowel movements increased significantly after ACE was performed.

Two retrospective cohorts and one retrospective case series [EL=3] showed that children's quality of life significantly improved after ACE was performed.

One retrospective cohort and one retrospective case series [EL=3] showed that there was a significant reduction in the use of laxatives after ACE was performed.

One prospective case series [EL=3] involving children with idiopathic constipation who did not respond to 3 years of medically supervised conservative management and underwent the ACE procedure showed that the probability of an ACE failing was 0.3 at 8.5 years; with an estimated mean failure time of 8.6 years (95% CI 7.9 to 9.2). The probability of an ACE being reversed was 0.2 at 6.2 years, with an estimated mean time to reversal of 9.1 years (95% CI: 8.4 to 9.7). Colonic transit time, age at surgery and duration of follow-up were not significantly associated with ACE failure, but the higher failure rate amongst girls was significant. The colonic transit time was a significant factor in predicting failure in children who accommodated very large volume of lavage fluid (>10l) in their colon without bowel evacuation

GDG interpretation of the evidence

It is the GDG's view that there needs to be a balance between offering ACE procedure early to children who might require it (those who remain symptomatic on optimal specialist management) and making sure that optimal specialist management has actually failed, and therefore children are not referred prematurely since this would not be a cost-effective use of scarce NHS resources.

The procedure needs to be performed in a surgical unit with expertise in assessing for suitability and performing ACE if indicated. The GDG believes that nurse support is essential for effectiveness of ACE procedure. The level of specialist nurses is not equitable across the UK.

The choice of washout solution, its type and volume, is empirical: there is no evidence on what works. There is no evidence on why ACE works in some children and not in others; therefore it is difficult for clinicians to choose the “right” patient.

Recommendations

Refer children and young people with idiopathic constipation who still have unresolved symptoms on optimum management to a paediatric surgical centre to assess their suitability for an antegrade colonic enema (ACE) procedure.

Ensure that all children and young people who are referred for an ACE procedure have access to support, information and follow-up from paediatric health professionals with experience in managing children and young people who have had an ACE procedure.

Research recommendations

What is the effectiveness of different volumes and types of solutions used for colonic washouts in children who have undergone antegrade colonic enema (ACE) for intractable chronic idiopathic constipation?

Why this is important

The ACE has a role in the management of people with treatment-resistant symptoms. Close follow-up is integral to the effectiveness of this technique to allow safe and effective administration of washout solutions.

The choice of washout solutions and frequency of administration varies between centres. Outcomes may be improved by evaluating how experienced centres choose washout solutions and by comparing techniques.

Centres offering ACE as treatment for children with chronic idiopathic constipation should be surveyed for their choice of washout solution. The survey should cover enema, washout fluid, volumes and frequency of administration, and how solutions are varied to determine the perceived strengths and weaknesses of each solution.

What are the experiences of children who have undergone ACE procedure due to intractable chronic idiopathic constipation?

Why this is important

There is a difference of opinion between healthcare professionals regarding the use of surgery in the management of intractable idiopathic constipation. Whilst some professionals feel that it is unnecessarily invasive others feel strongly that surgery has an important part to play. In addition, many families find the prospect of surgery daunting and there is little evidence to help professionals provide impartial information regarding children's and families' experience of ACE and its subsequent management, leaving them to rely upon their own opinion and experience.

The primary outcome measure of this research should be quality of life recorded using a validated health related quality of life measure.

Footnotes

*

The guideline follows the BNFC classification of laxatives.

**

This is phase 2 of the study. Phase 1 was a prospective case series already discussed in the review on disimpaction

***

This outcome was added by the GDG as it was reported by children to be very important even though it cannot technically be considered an adverse effect

****

Not reported whether with or without electrolytes, but probably without electrolytes as this is a study from the USA where PEG 3350 without electrolytes is generally used

*****

Not clear whether these are numbers or percentage of children, but probably percentage. Estimates were taken from a bar chart, as outcomes were not reported in the text.

†††††

It is likely that this formula is also Omneo/Conformil. The authors did not provide any brand name in the paper but the composition of the formula is the same as the one the authors used for their 2005 study

††††

All four studies included in this section are by the same centre and authors

†††

It is likely that this formula is also Omneo/Conformil. The authors did not provide any brand name in the paper but the composition of the formula is the same as the one the authors used for their 2005 study

††
Copyright © 2010, National Collaborating Centre for Women's and Children's Health.

No part of this publication may be reproduced, stored or transmitted in any form or by any means, without the prior written permission of the publisher or, in the case of reprographic reproduction, in accordance with the terms of licences issued by the Copyright Licensing Agency in the UK [www.cla.co.uk]. Enquiries concerning reproduction outside the terms stated here should be sent to the publisher at the UK address printed on this page.

The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore for general use.

Cover of Constipation in Children and Young People
Constipation in Children and Young People: Diagnosis and Management of Idiopathic Childhood Constipation in Primary and Secondary Care.
NICE Clinical Guidelines, No. 99.
National Collaborating Centre for Women's and Children's Health (UK).
London: RCOG Press; 2010.

NICE (National Institute for Health and Care Excellence)

PubMed Health Blog...

read all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...