NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Collaborating Centre for Women's and Children's Health (UK). Urinary Tract Infection in Children: Diagnosis, Treatment and Long-term Management. London: RCOG Press; 2007 Aug. (NICE Clinical Guidelines, No. 54.)

Cover of Urinary Tract Infection in Children

Urinary Tract Infection in Children: Diagnosis, Treatment and Long-term Management.

Show details

4Diagnosis

4.1. Introduction

4.1.1. Aim

The aim of this section is to describe when and how to consider the diagnosis of UTI in infants and children and how to confirm the diagnosis. This requires awareness of the presenting symptoms, knowledge of how to carry out an appropriate and relevant history and examination, how to collect urine, how to test and interpret the results of urine testing and how to differentiate acute pyelonephritis/upper urinary tract infection from cystitis/lower urinary tract infection.

Establishing an accurate diagnosis is important to aid prompt antibiotic treatment, direct appropriate investigations and reduce both short- and long-term morbidity and mortality associated with the condition.

4.1.2. Background

Most children with a first-time UTI in the UK present to primary care or to an emergency department. The clinical presentation may be influenced by several factors including the age of the child, the anatomical location of the infection in the urinary tract, the extent of verbal skills in the child and the stage in their toilet-training. In infants and children the presentation is more likely to be with fever with or without other systemic symptoms but rarely with urinary symptoms. In older children the majority have typical urinary tract symptoms suggestive of acute pyelonephritis/upper urinary tract infection or cystitis/lower urinary tract infection. Presentation could include a septic neonate, a child with fever, vomiting and loin tenderness or a teenager with suprapubic pain, dysuria and frequency.

In addition to establishing a diagnosis of UTI it is necessary to identify whether the child is likely to have upper or lower urinary tract infection, or in pathological terms whether the child has acute pyelonephritis or cystitis. In practice the distinction between upper and lower urinary tract infection has to be made on clinical grounds in a child with evidence of UTI based on the presence or absence of symptoms, signs of systemic illness, and, in particular, fever.

The diagnosis can be confirmed by collecting a clean urine sample which should be tested appropriately.

The cost implications range from those associated with the use of various diagnostic modalities to direct future care and treatment, and the morbidities arising from the condition and its management.

4.1.3. Current practice

A guideline by the Royal College of Physicians (RCP) states that all infants and children with a non-specific fever should have a urine sample examined so that they can have appropriate treatment and follow-up.21

A national audit of the guideline reported that 746 children younger than 2 years with a fever in 31 hospitals were included. The report showed that 81% had a dipstick screening test but it was not clear what dipsticks had been used or what elements had been tested. Fifty-eight percent of hospitals used nitrite dipsticks.23 Only 22% of urine dipstick results were clearly documented in the notes. Seventy-one percent had a documented urine sample sent to a laboratory for culture and/or microscopy. One hospital used direct microscopy of fresh urine at the bedside. A positive urine sample was obtained in 9.7% of cases sent to the laboratory. A second specimen was collected in almost half these children diagnosed with UTI. UTI was diagnosed and treatment and follow-up arranged in 4.7% of children who had urine microscopy and culture. In addition a further 5% of these febrile infants and children had a positive urine on culture but were not given a diagnosis of UTI, did not receive any treatment, prophylaxis, imaging or follow-up, and there was no communication to the GP or patient about the positive urine culture. This means that half of the infants and children seen in secondary care with probable UTI did not receive the correct diagnosis or recommended treatment.

4.2. Predisposing factors

Clinical question

In infants and children, what are the predisposing factors for a UTI?

Review findings – host susceptibility factors (age, gender, race, underlying concomitant disease)

Eight studies were identified investigating host susceptibility factors in children.47,116–122 All studies reported age and gender differences, but only one study reported race.117 One study investigated phimosis.122

A case series study from the USA investigated 100 children aged 5 days to 8 months (mean age 2.1 months) who were hospitalised for first known UTI.116 Male infants accounted for 75% of UTI cases within the first 3 months of life compared with 11% of infants who were 3 to 8 months of age. Of the 41 infants who were under 30 days old, 33 (81%) were boys. [EL = 3]

A cross-sectional US study investigated distribution of asymptomatic bacteriuria in 3057 school-aged children.117 No boys were found to have bacteriuria and 12/1267 girls aged 6–15 years had first-time UTI (8/772 (1.0%) girls aged 6–10 years; 4/495 girls (0.8%) for 11–15 years age group). One school with black children only participated in the study. Again, no boys were found to have bacteriuria and 0.9% of 115 girls had UTI. [EL = 3]

A case series study from Turkey retrospectively investigated 71 neonates aged 18.1 days (± 11.2 days) in whom UTI was diagnosed during the first 4 weeks of life.118 There were 54/71 (76.1%) boys and 17/71 (23.9%) girls with UTI, of which 40.8% (29/71) were preterm (gestational age range between 27 and 37 weeks). [EL = 3]

A case series study conducted in Sweden investigated 1177 children aged 10 years or younger with their first symptomatic UTI.47 In boys 133/225 (59%) cases were detected before the age of 1 year and in girls 181/952 (19%) of UTIs were detected before the age of 1 year. [EL = 3]

A cross-sectional study conducted in the USA identified clinical and demographic factors associated with UTI in febrile infants who presented to an emergency department and were ≤ 60 days old.119 Being uncircumcised (OR 11.6; 95% CI 5.0 to 26.6) and having a temperature > 39 °C (OR 2.5; 95% CI 1.6 to 4.0) was associated with an increased risk of UTI. In multivariable analysis, being uncircumcised (P < 0.001) and height of fever (P < 0.001) remained associated with UTI (see Table 4.1). [EL = 3]

Table 4.1. Associations between various risk factors and UTI.

Table 4.1

Associations between various risk factors and UTI.

A study conducted in Sao Paulo analysed the contribution of risk factors to the occurrence of UTI in 61 full-term neonates (26 boys, 35 girls) presenting with a positive bag culture and fever (> 37.8 °C), weight loss (> 10% of birthweight) or non-specific symptoms (feeding intolerance, failure to thrive, hypoactivity, irritability).120 On presentation, another urine sample was collected by suprapubic aspiration (SPA) to confirm diagnosis and 42 infants were found to be culture negative (group I) and a diagnosis of UTI was confirmed in 19 (group II). There were no significant differences between groups for birthweight, sex, asphyxia or membrane rupture time. On presentation there were no differences between the groups for fever (P = 0.31), but there were significant differences for weight loss (> 10% of birthweight) (P = 0.01) and non-specific symptoms (P < 0.001).

Children who had UTI confirmed by SPA were significantly more likely to have associated infectious diseases (RR 3.27; 95% CI 1.15 to 7.04; P < 0.001), be using broad-spectrum antibiotics (RR 3.03; 95% CI 1.51 to 6.08; P = 0.01), have renal and urinary tract malformations (RR 2.97; 95% CI 1.57 to 5.64; P = 0.007), be on mechanical ventilation (RR 2.99; 95% CI 1.61 to 5.53; P = 0.03), be on parenteral nutrition (RR 5.05; 95% CI 2.72 to 9.39; P < 0.001) and to have an intravascular catheter (RR 3.27; 95% CI 1.84 to 5.83; P = 0.009). [EL = 3]

A case series study conducted in the Philippines evaluated whether unexplained and/or excessive jaundice was associated with UTI in 54 jaundiced infants (22 boys, 32 girls) younger than 8 weeks old.121 Of the 54 included infants, five had UTI and 49 did not. There were no significant differences in demographic or historical characteristics between groups in terms of gender, age, place of birth, mode of delivery, birthweight, gestational age, neonatal infection or onset of jaundice. Similarly, there were no significant differences in maternal characteristics between groups in terms of maternal age, gravidity, presence of maternal infection or maternal illness. There were significant differences in total, direct and indirect bilirubin levels between infants who had and did not have UTI. [EL = 3]

One Japanese case–control study found that boys younger than 7 months with foreskins that could not be retracted to expose the external meatus were at 7.8 times higher risk for febrile UTI when compared with boys with foreskins that could be retracted to expose the external meatus (95% CI 3.99 to 15.31).122 This study, however, suffers from a fundamental flaw due to the fact that phimosis is physiological at this age. Therefore this study should be interpreted with caution. Additionally, not all the results from the analysis were reported, making it difficult to assess quality. [EL = 2−]

No studies on blood group as a predisposing factor for UTI in children were identified.

Review findings – familial renal disease

VUR prevalence is covered in the epidemiology section and is also included in the section on recurrence. The following two studies were identified, which investigated the likelihood of VUR in siblings of children with VUR, the majority of whom did not have a history of UTI.123,124

Using an awake voiding cystogram, an American case series study assessed 104 siblings aged 3 months to 15 years of patients with VUR (irrespective of history of UTI).123 Of the siblings, 34 (32.7%) were found to have VUR and among those with VUR, six (17.6%) had a history of UTI and 25 (73.5%) had no history of UTI. The remaining three were reported to have abnormal voiding patterns but their UTI history was not reported. [EL = 3]

A case series study conducted in Iran investigated the number of VUR cases in 40 children with siblings diagnosed with VUR.124 Seventeen (43%) siblings of 34 patients with VUR (irrespective of history of UTI) had VUR. Of the 17 with VUR, five (29.4%) also had a history of symptomatic UTI. VUR was bilateral in 6/17 and unilateral in 11/17 of the siblings. [EL = 3]

No high-quality studies on kidney stones or genetics as predisposing factors for UTI in children were identified.

Review findings – circumcision

Seven studies have investigated the association between circumcision and risk of UTI.125–132

An Australian meta-analysis looked at the effect of circumcision on the risk of UTI in boys in twelve studies. The meta-analysis included one RCT, four cohort studies and seven case–control studies.125

The RCT was a study of recurrent UTI in 70 uncircumcised boys with proven UTI aged 3 months to 10 years who were randomised to circumcision or no circumcision and showed an OR of 0.13 (95% CI 0.01 to 2.63).

Four cohort studies were conducted in hospital settings in boys aged 1–3 years and showed benefit with a summary OR of 0.13 (95% CI 0.07 to 0.23), however there was significant heterogeneity between these studies (χ2 = 82.48; degrees of freedom (df) = 3; P < 0.001). When one outlying study was excluded, the heterogeneity was not significant (P = 0.64).

The seven case–control studies were conducted in secondary care settings. Six of the seven studies were in boys aged 1 month to 5 years in hospital care settings, and one study was in adults attending a community sexually transmitted disease clinic. The case–control studies included showed benefit with a combined OR of 0.13 (95% CI 0.07 to 0.23). There was no significant heterogeneity between these studies (χ2 = 8.15; df = 6; P = 0.20)

The summary OR across all study types was 0.13 (95% CI 0.08 to 0.20). There was no significant heterogeneity observed between study types (χ2 = 0.16; df = 2; P = 0.90), but significant heterogeneity was observed within the individual studies (χ2 = 90.63; df = 11; P < 0.001) owing to the inclusion of the cohort studies. Without the cohort studies, there was no significant heterogeneity (χ2 = 10.92; df = 10; P < 0.4).

The odds of a circumcised boy having a UTI are about 0.1 when compared with uncircumcised boys. While circumcision may be protective against UTI, the risk–benefit of circumcision is not easily quantifiable. The study concludes that while circumcision substantially reduces the risk of UTI, routine circumcision should not be considered. Circumcision has a potential role in boys with past history of recurrent UTI, or with high-grade VUR, as the benefits in these cases may outweigh the risk of complications. [EL = 2++]

A US cohort study of 28 812 infants found that the median age at diagnosis of UTI was 2.5 months for uncircumcised males, 4.5 months for circumcised males and 6.5 months for female infants.126 The incidence of UTI in the first year of life was 1/47 for uncircumcised males, 1/455 for circumcised males and 1/49 for females. Circumcised males had significantly fewer episodes of first-time UTI (OR 9.1; 95% CI 5.2 to 15.7)126 [EL = 2++]

In a retrospective cohort study of all 136 086 boys born in US army facilities from 1980 to 1985, medical records were examined to determine any association between UTI and circumcision during the first month of life.127 Significantly more UTIs occurred in the boys who were not circumcised (P = 0.001) when compared with boys who were circumcised. [EL = 2+]

In a US cohort study of 5261 infants born at an army hospital from 1982 to 1983, 400 (7.6%) infants were evaluated for UTI in the first year of life and 41 of the infants (0.78%) were subsequently diagnosed with UTI.128 Among the 41 with UTI, 13 were female, four were circumcised males and 24 were uncircumcised males. The incidence of UTI in males was higher than in females (28/2502 versus 13/2759; P < 0.01) and the incidence of UTI in uncircumcised males was higher than in circumcised males (24/583 versus 4/1919; P < 0.001). [EL = 2+] An evaluation of all infants born in army medical facilities from 1975 to 1984 (n = 427 698) confirmed these findings.129 Females were significantly more likely to have UTI in the first year of life when compared with males (0.51% versus 0.28%; χ2 = 143.5; P < 0.001) and circumcised males were less likely to have UTI in the first year of life when compared with uncircumcised males (0.09% versus 1.0%; χ2 = 1086.4; P < 0.001). [EL = 2+]

A Canadian cohort study identified 69 100 boys who had been circumcised within the first month of life. The risk of hospitalization for UTI decreased with age, but remained higher for boys who were uncircumcised.130 At 1 month after birth, the probability of hospital admission for UTI (per 1000 person-years) was 4.5 times higher for uncircumcised boys when compared with circumcised boys (95% CI 2.4 to 8.4). Subsequent relative risk at 1 and 3 years was 3.7 (95% CI 2.8 to 4.9) and 3.0 (95% CI 2.4 to 3.8), respectively, with 195 circumcisions needed to prevent one hospital admission for UTI in the first year of life. [EL = 2++]

An Australian case–control study recruited boys younger than 5 years and compared 144 boys with UTI (median age 5.8 months) with 742 boys without UTI (median age 21.0 months).131 Of the boys with UTI, two (1.4%) were circumcised compared with 47 (6.3%) of the controls (P = 0.02). There was no evidence that age was a confounder or modified the protective effect of circumcision. [EL = 2+]

A US case–control study compared 36 boys with UTI to 76 controls. Male infants younger than 1 year presenting with first-time UTI were significantly more likely to be uncircumcised when compared with male infants without UTI.132 This was true regardless of age (< 3 months and > 3 months; all P < 0.0001), ethnic group (white, black and Hispanic; all P ≤ 0.02) and socioeconomic status (using type of medical insurance as a proxy; all P ≤ 0.02). [EL = 2+]

Review findings – lifestyle considerations

Breastfeeding

A case–control study conducted in Sweden aimed to investigate the association between breast-feeding and the risk of first-time febrile UTI.133 Cases (n = 200) and controls (n = 336) were recruited consecutively in two paediatric departments in Sweden and matched for gender and age. Of children aged 0–6 years, presenting for the first time with symptomatic UTI, exclusive breastfeeding was found to have a protective effect on the risk of UTI. The risk of UTI was 2.3 times higher in non-breastfed children when compared with exclusively breastfed children (95% CI 1.56 to 3.39). The protective effect of breastfeeding was dependent on the duration of breastfeeding as well as the gender of the child or infant. A longer duration of breastfeeding was associated with a lower risk of infection after weaning and the effect was stronger in girls (hazard ratio = 3.78) than in boys (hazard ratio = 1.63). [EL = 2+]

Use of nappies

A case–control study conducted in Finland compared disposable, superabsorbant and washable cotton nappies in children presenting with their first UTI.134 No differences were found (disposable OR 0.95; 95% CI 0.62 to 1.46, superabsorbant OR 1.04; 95% CI 0.69 to 1.57, washable cotton OR 1.00; 95% CI 0.46 to 2.16). [EL = 2+]

Hygiene

In a case–control study from the Philippines, the association between UTI and urination, defecation, washing and bathing habits was investigated in children aged 6–12 years (n = 23 cases, n = 23 controls).135 Bathing habits (daily versus less than daily), urinary frequency (fewer than 5 times per day or 5+ per day), holding urine during the day (yes or no), permission to urinate at school (during break versus whenever), washing after urination (yes or no), washing after defecation (yes or no), direction of washing (from behind versus from front), and use of soap during washing (yes or no) showed no association with risk of UTI. The study did not specify whether the controls were matched for age and gender, selection criteria were not explicit, withdrawals were not explained and the small sample size resulted in wide confidence intervals. [EL = 2−]

Voiding habits

In a Swedish cross-sectional study, 1557 children (aged 6–9 years) and their parents/carers responded to questionnaires (56% response rate) regarding voiding habits.136 Nearly 10% of girls (75/823) and 3% of boys (20/728) reported a previous history of UTI. Although the number of boys with previous UTI was too small to draw any conclusions, symptoms suggesting emptying difficulties were seen significantly more often in girls with previous UTI when compared with girls with no history of UTI, including:

  • bed wetting (P = 0.002)
  • day wetting (P < 0.001)
  • does not reach toilet (P = 0.03)
  • prolonged voiding (P < 0.002)
  • poor stream (P < 0.003)
  • staccato voiding (P < 0.006)
  • able to void again (P < 0.002)
  • straining (P = 0.02)
  • manual compression of abdomen (P < 0.003)
  • encopresis (P = 0.03).

The daily frequency of micturition between children who reported a history of UTI was not statistically different from those who did not report a history of UTI. [EL = 3]

Evidence statement – predisposing factors

Being a male infant younger than 3 months and being a girl over 3 months are risk factors for UTI.

The risk of UTI is higher in uncircumcised boys than in circumcised boys (OR 9.1; CI 5.2 to 15.7). One hundred and ninety five male infants need to be circumcised to prevent one case of UTI requiring hospital admission.

Breastfeeding has a protective effect against UTI and this is more pronounced in female infants. This is dependent on the duration of breastfeeding and the effect appears to persist even after weaning.

No high-quality studies were identified evaluating the association between ethnicity, blood groups, type of nappy, familial susceptibility, phimosis or renal stones and risk of UTI in children. No good-quality studies were identified to link other personal hygiene, religious or social factors to risk of UTI in children.

4.3. Symptoms and signs

Clinical question

In infants or children, what signs or symptoms would give rise to the suspicion of UTI?

Review findings – symptoms and signs

Thirteen studies were identified reporting symptoms and signs in children presenting with UTI. The majority of studies reported symptoms in children treated for UTI in secondary care,14,42,50,52, 67,98,116,137–140 while two studies reported symptoms of children presenting to a GP.45,141

A case series study conducted in Australia described the clinical features of 305 children younger than 5 years who presented consecutively at an emergency department with first-time symptomatic UTI.137 The most commonly reported symptoms were fever (80%), an axillary temperature higher than 37.5 °C (60%) irritability (52%), anorexia (49%), malaise (44%), vomiting (42%) and diarrhoea (21%). Less common symptoms in fewer than 20% of children were dysuria, offensive urine, abdominal pain, frequency and haematuria. [EL = 3]

A case series study from the USA reported symptoms and signs from 100 children aged 5 days to 8 months (mean age 2.1 months) who were hospitalised for first known UTI.116 Fever was the most common symptom (63%) and irritability was reported in over half of the children (55%). Other symptoms included refused feeds (38%), vomiting (36%) and diarrhoea (31%). Less common symptoms were abdominal distension and jaundice, which were reported in 8% and 7% of the children, respectively. [EL = 3]

A case series study was conducted in the USA in 83 boys aged 2 weeks to 14 years presenting to a children’s hospital with first-time UTI (25% were ≤ 1 year old and half were < 6 years old).138 Fever was present in 40 (48%) of the children and was the only presenting sign in 25%. Other symptoms included overactive bladder syndromes in 23 (28%), abdominal or flank mass in 11 (13%), enuresis in seven (8%) and gross haematuria in six (7%). [EL = 3]

A case series study of children aged 0–14 years (64 children younger than 12 months) in Italy described 223 children presenting to a hospital with first-time UTI.50 Presenting symptoms included fever in 144 (65%), dysuria and frequency in 91 (41%), gastrointestinal symptoms in 42 (19%), haematuria in 25 (11%), failure to thrive in 14 (6%) and jaundice in two (1%). [EL = 3]

A case series study in the UK investigated 120 children (aged 2 weeks to 12 years) who had a UTI and underwent an IVU.98 Presenting symptoms were fever in 77% (57/74) and uncoordinated voiding with residual urine in 11% (8/74). [EL = 3]

A case series study in Finland presented population surveillance data of children aged 1 week to 9.5 years (median age 0.125 years) and reported on symptoms of UTI in 134 children with first-time bacteraemic UTI.52 The most common presenting symptoms were fever (92%) and irritability (60%). Other symptoms included abnormal crying (34%), vomiting (16%), lethargy (26%), feeding problems (20%), abdominal pain (7%), dysuria (1%) and convulsions (4%). National Surveillance data were used to compare the results with 134 children with first-time non-bacteraemic UTI. The only significant difference reported was for feeding problems (20% versus 10%; P = 0.02). [EL = 3]

An RCT conducted in Turkey investigated the effectiveness of circumcision on recurrent UTI and described the presenting symptoms of 88 boys referred to a paediatric nephrology department with first-time UTI.139 The most common presenting symptoms were fever < 38.5 °C (48%), dysuria/frequency (34%) and fever > 38.5 °C (24%). Other reported symptoms included vomiting and/or diarrhoea (22%), enuresis (7%), suprapubic discomfort (11%), abdominal pain (18%), flank pain (5%) and offensive urine (2%). [EL = 3]

A case series study conducted in the UK recruited 744 children with UTI aged 0–12 years treated in a hospital.67 Fever was a presenting symptom in 42%. Other reported symptoms included abdominal or loin pain (31%) and enuresis (38%) which was only identified in children 5 years or older. A significantly greater proportion of children with VUR (141/246) presented with fever compared with children without VUR (173/498; 57.3% versus 34.7%; P < 0.001). [EL = 3]

A case series study from the UK reported symptoms of 14 children with UTI aged 15 years or younger in a semi-rural general practice.45 Six children (40%) presented with dysuria and frequency, three (20%) with abdominal pain, two (13%) with enuresis and one each (7%) with loin pain, haematuria and failure to thrive. [EL = 3]

A case series study conducted in the UK reported the clinical and laboratory features of 49 boys aged 2–12 years presenting to primary care practices with UTI.141 The most common presenting symptoms were dysuria/frequency (82%), abdominal pain (35%) and enuresis (45%). Other reported symptoms included fever (26%), haematuria (20%) and balanitis (20%). [EL = 3]

A study conducted in a general practice in the UK presented the clinical findings of 38 children (12 boys and 26 girls) younger than 15 years with culture-proven UTI (> 105 cfu/ml in a clean catch urine sample).140 Dysuria was present in 27/38 children (71%) and was the most common symptom. Twelve children (32%) presented with abdominal pain, five (13%) presented with loin pain/tenderness, nine (24%) with enuresis, eight (21%) with fever, seven (18%) with offensive urine, two (5%) with daytime incontinence, one (3%) with haematuria and one (3%) with rigor. [EL = 3]

A case series study conducted in Sweden described fever (≥ 38 °C) as one of the clinical features of children aged 0–16 years presenting at a children’s or maternity hospital for symptomatic UTI.42 The number of children presenting with fever decreased with age. In infants younger than 12 months 179/186 (96%) presented with fever; in children 1 year or older but younger than 3 years 70/96 (73%) presented with fever; in children 3 years or older but younger than 10 years 120/200 (60%) presented with fever; in children 10 years or older but younger than 16 years 19/41 (46%) presented with fever. [EL = 3]

In a US prevalence study, UTI occurred in 50/945 (5.3%) febrile infants younger than 1 year presenting to the emergency department of a children’s hospital. UTI was found to occur significantly more often among infants with no identified source of fever (34/454) when compared with infants with a condition identified as a possible source of fever (15/429; 3.5% versus 7.5%; P = 0.02).14 UTI was least prevalent among infants with an unequivocal source of fever (1/62). [EL = 3]

Evidence statement – symptoms and signs

Limited evidence shows that the most common symptoms and signs of UTI in children diagnosed at primary care are frequency and dysuria.

The most common symptoms and signs of UTI in children diagnosed at hospital are fever, irritability, malaise and gastrointestinal symptoms. Other less common symptoms/signs include dysuria, frequency, abdominal pain, failure to thrive, smelly urine and haematuria, bed-wetting, problems with voiding and encopresis.

UTI is more frequent among infants with no obvious focus of fever compared with those in whom there is an obvious focus.

Table 4.2Summary of included studies on symptoms and signs of UTI in children

StudyCraig (1998)137Ginsburg (1982)116Burbige (1984)138Messi (1988)50Smellie (1985)98Honkinen (2000)52Nayir (2001)139Smellie (1981)67Dickinson (1979)45Hallett (1976)141Brooks (1977)140
Age< 5 years5 days to 8 months2 weeks to 14 years< 14 years2 weeks to 12 years1 week to 9.5 years3 months to 10 years≤ 12 years≤ 15 years2–12 years< 15 years
Study size (n)30510083223120134887441449a38
SettingHospitalHospitalHospitalHospitalHospitalHospitalHospitalHospitalPrimary carePrimary carePrimary care
CountryAustraliaUSAUSAItalyUKFinlandTurkeyUKUKUKUK
Symptom (%)
Fever60634865779272422621
Irritability525560
Vomiting42361622b
Anorexia49see vomiting
Diarrhoea2131
Enuresis8738c144524
Dysuria1541d134d43d82d71
Frequency10see dysuriasee dysuriasee dysuriasee dysuria
Abdominal pain1346e71831213532
Smelly urine13218
Haematuria7710.87203
Failure to thrive6.37
Malaise4426
Poor feeding3820
Constipation21
a

All male.

b

Reported with diarrhoea.

c

In children 5 years or older (n = 355).

d

Reported with frequency.

e

Reported with loin pain.

4.4. Urine collection

Introduction

The aim of urine collection is to obtain a good-quality sample from which the diagnosis of UTI can be confidently confirmed or excluded.

Accurate diagnosis of UTI is essential to avoid inappropriate over- or undertreatment or investigation. This is most important in children who are not toilet-trained in whom obvious urinary tract symptoms are rarely present.

To establish an accurate diagnosis of UTI requires the collection of an appropriate urine sample. Since the majority of children presenting with a UTI in the UK are likely to present in primary care, the collection of a urine specimen needs to be simple, reliable, cost effective and acceptable to children, parents and carers.

Instructions to families need to include clear detailed information about the practicalities of the method used and advice about appropriate skin cleansing.

A variety of methods are used in primary care, predominantly ‘clean catch’, urine collection pads (Euron Uricol™) or urine collection bags. Collection by clean catch is difficult, particularly in children, and is not always successful. Other methods sometimes used to collect urine, including gauze, cotton wool balls, sanitary towels and panty-liners placed in the nappy, often lead to inaccurate results because of bactericidal agents incorporated in these materials, rendering them unsuitable.34,142

In hospitals, additional methods are available, including suprapubic aspiration (SPA) and samples taken using catheterisation. While being advocated in the literature as the ‘reference standard’ to collect urine, SPA is invasive and unpleasant for the child and is dependent upon skilled practitioners to perform. It is also not suitable as a method of urine collection in primary care. However, in a hospital environment, when a child is acutely unwell and commencement of antibiotics is urgent, it may be appropriate to use an invasive method such as urethral catheterisation or SPA.

The costs associated with urine collection include not only the costs of materials used and personnel time collecting and processing the urine, but also the costs of misdiagnosis. Failure to accurately diagnose a urine infection may result in treatment delay and may increase the likelihood of renal parenchymal defects.

All urine collection methods have a risk of contamination by organisms not present in the bladder. This may lead to misdiagnosis and unnecessary treatment or investigation if current guidelines are followed. Children who are not toilet-trained are particularly prone to yield contaminated samples, as they are unable to pass urine to order or to cooperate with the process. In addition urine often flushes the vagina in infant girls and the prepuce in infant boys. Thus contamination of samples occurs after leaving the bladder but before it can be retrieved for diagnostic purposes.

The clean catch method tends to provide fewer contaminated samples than bags or pads. Urine collection bags are unpleasant for the child, costly and not environmentally friendly. Pads may be useful – if used correctly they are inexpensive and user friendly. The material cost of a clean catch specimen is negligible but it may be time-consuming; nevertheless, some parents/carers have expressed a preference for this method.

Current practice

In the RCP guideline it was stated that clean catch urine in an infant or a mid-stream urine specimen in an older child is the ideal. If a clean catch urine is not available then use a bag.21

In the national audit of the RCP guideline, specimen types recorded were: bag urine 32%, midstream urint (MSU) 22%, clean catch 11%, SPA and catheter specimen of uring (CSU) 4.1%. There was no record of urine specimen type in 27%. It was unclear what MSU meant in this age group. Contamination was found in 0–75% of samples from different units (mean contamination rate 34%).23

Clinical question

In infants and children with suspected UTI, which method of urine collection is most effective?

Review findings – clean catch urine samples

A systematic review143 [EL = II] identified five studies (with seven data sets) that assessed the diagnostic accuracy of a clean catch urine sample, with SPA urine sample as the reference standard. All studies were judged to be of reasonable quality. Half of the studies were in children aged 0–12 years and half were in children aged younger than 3 years with a mean age of around 4 months. There was no study that directly compared diagnostic accuracy of this urine sampling between different age groups.

Sensitivity ranged from 75% (specificity 96%) to 100% (specificity 100%) and specificity ranged from 57% (sensitivity 83%) to 100% (sensitivity 100%). The positive likelihood ratio (LR+) values ranged from 1.9 (negative likelihood ratio (LR−) of 0.30) to 47.7 (LR− = 0.08). The LR− values ranged from 0.08 (LR+ = 47.7) to 0.36 (LR+ = 3.57). Although there was considerable heterogeneity all studies were clustered towards the top left of the receiver operating characteristic (ROC) curve suggesting that acceptable diagnostic performance is obtained from clean catch urine samples.

There was considerable heterogeneity in pooled LR+ values (P < 0.0001) but the LR− values were statistically homogeneous (P = 0.50). The pooled LR+ was 7.7 (95% CI 2.5 to 23.5) and the pooled LR− was 0.23 (interquartile range (IQR) 0.18 to 0.30).

Review findings – early compared with mid-stream samples

No studies were found comparing early to mid- or late stream samples for any urine collection method in children.

Review findings – pad/nappy samples

A systematic review found four studies that examined the accuracy of specimens collected from pads/nappies. Three studies compared pad/nappy samples with culture of bag specimens, although bag collection was not considered likely to be the best method of urine sample collection, limiting the value of these studies. The remaining study was found to have compared the pad/nappy specimens to SPA samples, and reported 100% sensitivity and 94% specificity between the two methods. The LR+ was 12.5 and the LR− was 0.09. Limited data made it difficult to draw firm conclusions.143 [EL = II] An RCT conducted in the UK evaluated a modified urine collection pad method for its ability to reduce heavy mixed growth bacterial contamination of urine collection pad samples in 68 children (37 single pads, 31 replaced pads) younger than 2 years with suspected UTI.144 Eighty children were recruited (42 in the single urine collection pad and 38 in the replaced urine collection pad), and urine collection failed in 12 children (five single pad, seven replaced pad) mainly because of faecal soiling of the pad and were excluded from the analysis. Sixty-eight children were randomised into two groups: a single urine collection pad that was left in the nappy until a sample had been obtained; or a urine collection pad that was replaced every 30 minutes until a sample was obtained. Alarm sensors were placed in all urine collection pads.

Baseline characteristics of the groups were similar with respect to age but there were significantly more boys in the single pad group (25/37 versus 13/31; P = 0.03). Three of the 68 (4%) children had a UTI. Of the remaining 65 who did not have a UTI, heavy mixed growth was significantly higher in the single pad (10/35), compared with the replaced pad (1/30), P = 0.008. [EL = 1+]

Review findings – bag samples

A systematic review143 and three cohort studies145–147 investigated urine collection bags.

A systematic review found three studies examining bag specimens. One study compared culture and microscopy results of bag specimens to catheter specimens in two age groups, in children younger than 5 years and in the whole sample (children aged 9 days to 11 years). In children younger than 5 years, sensitivity was 81%, specificity 87%, LR+ 5.5 and LR− 0.24. In children aged up to 11 years, the sensitivity was 77%, specificity 82%, LR+ 3.9 and LR− 0.30. The other two studies compared culture of bag samples with culture of SPA samples, with considerable difference in results – one reported a sensitivity of 100% and specificity of 89%, with an LR+ of 7.7 and a LR− of 0.04; the other reported a sensitivity of 50%, specificity of 92%, LR+ of 5.4 and LR− of 0.55. There were insufficient data for drawing firm conclusions about bag specimens.143 [EL = II]

A cohort study conducted in the UK evaluated the ease of application and reliability of two different urine collection bags, the Hollister U-bags and the Urinicol bag (Euron Uricol™), in 50 children (33 boys, 17 girls) attending a children’s clinic.145 The nurses first cleaned the genital area with warm tap water and cotton wool balls before applying the bag. Hollister U-bags were used in 18 boys and seven girls, while Urinicol bags were used in 15 boys and ten girls. Eight out of 25 Hollister U-bags leaked compared with 0/25 Urinicol bags (P < 0.01). [EL = 2+]

A cohort study conducted in Canada compared the risks of contaminated culture results in urine specimens obtained by urine collection bag with those obtained by catheterisation in 7584 urine samples collected from 4632 children aged up to 24 months at an emergency department or outpatient unit.146 Bag urine cultures were obtained by Hollister U-bag after the perineum was cleansed with antibacterial soap and tap water. In the outpatient centre the bag was replaced after 30 minutes, while in the emergency department it was not. Catheter specimens were only collected in the emergency department after cleansing with iodinated soap and sterile water.

Of the 7584 urine cultures, 42.1% were obtained in infants younger than 6 months, 25.9% in infants between 6 and 11 months and 31.9% from infants between 12 and 24 months. Of the bag specimens, 2597 were collected at the emergency department and 2530 at the outpatient unit. 2457 catheter specimens were collected at the emergency department. Bag collection (54.4% bag versus 9.0% catheter (P < 0.001)); male gender (38.7% male versus 29.2% female (P < 0.001)); and age over 12 months (31.4% < 12 months versus 38.7% 12–24 months (P < 0.001)) were significantly more likely to be contaminated. The odds ratio (adjusted for age, sex and leucocyte esterase test) was 13.3 (95% CI 11.3 to 15.6) and when limited to the first urine culture in each child was OR 13.6 (95% CI 11.1 to 16.7). [EL = 2+]

A study conducted in the UK compared the contamination rates between bag and clean catch urine collection methods in children younger than 2 years in one of two inpatient wards.147 In Ward A, the child’s genitalia was washed with soap and water and urine samples were collected in a sterile foil bowl. In ward B soap and water was used, followed by cleansing with sterile water and drying with cotton wool balls and urine collection bags, either Hollister U-bags or Simcare bags, were applied.

Forty-six urine samples (23 from each ward) were obtained; in ward A 44 attempts were made to obtain 23 urine samples, 18 of which were obtained in 1 hour or less. A parent/carer was involved in 33 of the 44 attempts. Of the 11 times a nurse was involved, total time taken was 3 hours and 25 minutes, but for 2 hours and 15 minutes nurses were also feeding the infants, therefore extra time taken overall was 1 hour and 10 minutes. No specimens were contaminated.

In ward B 28 attempts were made to obtain 23 samples. The urine collection bags were in place for 15 minutes to 4 hours and 10 minutes, with an average time of 1 hour and 25 minutes. Eleven specimens were contaminated with faecal bacteria. [EL = 3]

Review findings – catheter and SPA samples

An RCT conducted in Israel compared the severity of pain during SPA with pain during transurethral catheterisation in 51 infants (31 boys, 20 girls) younger than 2 months.148 Pain during urine collection was assessed on a 100 mm visual analogue scale by a nurse and a parent/carer. Additionally, the infants’ upper body was videotaped during the procedure and an investigator assigned a point score based on the Douleur Aigue du Nouveaune (DAN) neonatal pain scale.

There were no baseline differences between children undergoing SPA and those who were catheterised in terms of age or weight, but those who were catheterised were older than those undergoing SPA (27.7 ± 14.8 days versus 36.5 ± 12.3; P = 0.007). On the visual analogue scale recorded by a nurse, the mean pain recorded for SPA was 63 ± 18 compared with 43 ± 25 for catheter. When parents/carers used the visual analogue scale, they recorded a mean of 63 ± 27 in children undergoing SPA compared with 46 ± 26 in children who were catheterised. Similarly, DAN scores and duration of cry were higher and longer for children randomised to SPA (7.0 ± 1.9 and 62.9 ± 26 seconds, respectively) compared with infants randomised to catheter (4.5 ± 2.1 and 49.7 ± 35.7 seconds, respectively). [EL = 1+]

Review findings – ultrasound-guided SPA versus conventional SPA

It is difficult to obtain a good-quality urine sample from infants because they are unable to cooperate. SPA has been regarded as the reference standard for urine collection in babies younger than 12 months, but it is an invasive procedure with attendant risks and inexperienced clinicians can find this method difficult. Ultrasound-guided SPA involves either scanning for the presence of urine before attempting an SPA, or scanning while aspirating the urine.

Four RCTs149–152 were identified comparing ultrasound-guided SPA with conventional blind SPA. A summary of the results is presented in Table 4.3.

Table 4.3. Summary results for included studies comparing ultrasound-guided SPA with conventional methods.

Table 4.3

Summary results for included studies comparing ultrasound-guided SPA with conventional methods.

An RCT conducted in Hong Kong investigated the optimal method of SPA in 60 infants, the success rate of real-time ultrasound-guided SPA (30 infants: 19 boys and 11 girls) compared with conventional SPA (30 infants: 8 boys and 22 girls) and factors associated with success.149 The overall success rates were 26/30 (87%) in the ultrasound-guided group and 24/30 (80%) in the control group (P < 0.05). The first attempts in both groups were equally successful 18/30 (60%). In the ultrasound-guided group, compared with failed attempts, successful SPA was associated with a greater bladder depth (28 ± 11 mm versus 21 ± 5 mm; P < 0.01), length (32 ± 12 mm versus 23 ± 9 mm; P < 0.05) and volume (17 ± 13 ml versus 8 ± 6 ml; P < 0.01) but similar width (P > 0.05). In the control group, successful attempts were associated with the presence of bladder dullness demonstrated by light percussion (23/24 versus 8/18; OR 29.0; P < 0.001) compared with failed attempts. [EL = 1+]

An RCT conducted in the USA investigated whether ultrasound guidance was useful to localise the position of the bladder and to increase the amount of urine obtained by SPA in 53 neonates.150 Twenty-eight were randomised to the ultrasound-guided group and 25 to the control group. Ultrasound-guided SPA was more likely to be successful on the first attempt (26/28 versus 7/25; P = 0.001), more successful overall – with one or more attempt (27/28 versus 15/25; P = 0.003), have a greater volume of urine obtained (2.1 ± 1.2 ml versus 1.3 ± 0.9 ml; P = 0.03) and require fewer passes (1.7 ± 1.0 versus 4.4 ± 2.0; P = 0.001). There were no differences with respect to procedure time (53 ± 59 seconds versus 60 ± 40 seconds; P = 0.60). [EL = 1+]

An RCT conducted in the USA investigated whether portable ultrasound could improve the success rate of SPA in 66 children aged 0–15 months (median age 1 month) presenting to a paediatric emergency department.151 Fifteen of 19 (79%) SPA attempts were successful in the ultrasound group compared with 16/31 (52%) in the control group (P = 0.04). In 3/4 SPA attempts in the ultrasound group and in 11/15 SPA attempts in the control group, aspiration yielded ≥ 5 ml of urine. Operator efficiencies showed an increasing success rate over time (P = 0.03). [EL = 1+]

In children younger than 1 month, there were no differences in success rates between ultrasound-guided (75%) and controls (74%) (P > 0.05). Additionally, the volume of urine obtained was approximately 6 ml for both groups (P > 0.05). [EL = 1+]

Review findings – early compared with late stream samples

A systematic review143 [EL = II] found one study showing good agreement between the results of culture from the early part of a catheter sample when compared with the later part of the same sample, with sensitivity of 100%, specificity of 95%, LR+ of 16.7 and LR- of 0.08. The limited data available means no firm conclusions can be drawn.

No other studies were found comparing early with late stream samples for any other urine collection method in children.

Review findings – other comparisons of urine collection methods

Four studies investigated other combinations of urine collection methods.153–155

A prospective cross-sectional study compared the validity of the urinalysis on clean catch and bag versus catheter urine specimens using catheter culture as the reference standard in non-toilet-trained children younger than 3 years who presented to a children’s emergency hospital in the USA between June 2000 and December 2001.153

The sensitivity of the bag dipstick was greater than the catheter dipstick (85% (95% CI 78% to 93%) versus 71% (95% CI 61% to 81%); P = 0.03) and sensitivity was highest in children older than 90 days. However, specificity of the bag dipstick for all ages was low compared with the catheter specimens (62% (95% CI 56% to 69%) versus 97% (95% CI 94% to 99%); P < 0.001). The LR+ of the bag dipstick was 2.24, while for the catheter dipstick it was 23.67. LR− values were 0.24 for the bag dipstick and 0.30 for the catheter dipstick. In the combined dipstick and microscopy urinalysis, sensitivity of both bag and catheter specimens increased, and specificity decreased compared with dipstick alone.

The dipstick sensitivity in both bag and catheter samples did not differ according to sex. However, specificity was higher in boys than in girls for all ages and could not be explained by the fact that circumcision had been performed. Sensitivity rose with higher cut-off values for defining positive UTI, while specificity dropped. [EL = III]

A study conducted in the USA compared urine collected by bag and by catheter test performance characteristics in children younger than 93 days with temperature of 38 °C or higher who underwent urinalysis and urine culture.154 A summary of the results is presented in Table 4.4.

Table 4.4. Summary of the study by Schroeder et al. comparing non-invasively collected urine and catheterised urine test performance.

Table 4.4

Summary of the study by Schroeder et al. comparing non-invasively collected urine and catheterised urine test performance.

Of the 1482 infants who had urinalysis and urine culture, 1384 had samples obtained by bag or catheter. Overall, leucocyte esterase had higher sensitivity, while nitrites had higher specificity. The only significant difference between bag and catheter was the comparison of specificity of leucocyte esterase. There were no significant differences when the cut-off values for a positive result were changed.

Further analysis was carried out on 54 patients who had false positive results for leucocyte esterase on bag urinalysis. Of the children who were also tested for nitrites, 4/15 (8%) had positive results. Of children who were also tested for urine white blood cell counts (WBC) 9/47 (19%) had more than 10 WBC/hpf. If children who had urine samples with positive leucocyte esterase and positive nitrite results, more than 10 WBC/hpf, or ambiguous culture results are considered to have a UTI, the difference between the methods in specificity for leucocyte esterase was still significant (bag 89%, catheter 95%; P < 0.001).

The area under the ROC curve for urine WBC counts and UTI was higher in children with catheter samples than in those with bag samples (0.86 versus 0.71; P = 0.01). [EL = III]

A study conducted in the UK assessed 44 parents’/carers’ preferences for collecting urine at home from 29 boys and 15 girls aged 1–18 months and examined contamination rates.155 Pads were placed inside the nappy and checked every 10 minutes until wet, then urine aspirated with a syringe. Bags were applied and inspected every 10 minutes and removed to decant urine. Parents/carers preferred using the pad first, the bag second and the clean catch method third. Seven samples from pads, eight from bags and one from clean catch had contamination.

Nine samples from five children grew > 105 cfu/ml, suggesting infection. However, these were excluded by sterile samples collected on the same day in hospital.

Parents/carers found pads and bags easy to use and preferred them to the clean catch method. The pad was considered comfortable, whereas the bag was distressing, particularly on removal often leaking and leaving red marks. Some found extracting the urine from the pad or emptying the bag awkward. Most parents/carers complained that the clean catch method was time-consuming and often messy and nine parents/carers gave up after prolonged attempts. [EL = 2+]

Evidence statement – urine collection

Limited available evidence showed that the urine collection methods that produce a most diagnostically accurate sample for testing are clean catch and SPA.

The only urine collection method for which there was adequate data was the comparison of clean catch urine to SPA. There were five studies, two of which used different criteria for being positive. When both samples were cultured the agreement between the methods was reasonable for diagnostic values. One outlying study showed poor performance of clean catch urine. The reasons for this are unclear.

Ultrasound-guided SPA is a more successful method of obtaining urine from the bladder than conventional SPA. Three of four studies found that the use of ultrasound to detect urine in the bladder immediately before SPA increases the success rate of SPA.

There is insufficient data to draw conclusions about urine collection bags and urine collection pads. There is low-level evidence that showed that the accuracy of urine collection pads was greatly improved if the pads were not used for longer than 30 minutes.

None of the routine methods for urine collection (clean catch, pad or bag) are costly in terms of equipment or clinician time. Urine collection bags cost between £0.65 and £3.25 each, while urine collection pads cost around £0.55 per pack of two pads (May 2007 prices, see Chapter 8). The evidence for choosing urine collection bags or pads is insufficient, so the least costly method should be preferred. The GDG consensus was that ultrasound-guided SPA was likely to a cost-effective alternative where it is not possible to obtain a urine sample through non-invasive methods.

4.5. Urine preservation

Introduction

Urine readily supports bacterial growth and specimens of urine are frequently contaminated. It is well recognised that time delay in culturing urine allows contaminants to multiply and produce inaccurate results. The addition of preservatives, usually boric acid, to the urine samples can be an alternative to lowering the temperature. Currently, boric acid is used in various commercially available transportation tubes.

When analysis of urine samples is requested, there is often inadequate explanation of the collection procedure. Various studies have reported that this is a problem in primary care.

Clinical question

How should a urine sample be transported to ensure its reliability?

4.5.1. Chemical preservation

Review findings – chemical preservation of urine

Six studies were found that evaluated chemical preservation of urine.156–161

One study in Sweden evaluated a commercial tube prepared with boric acid, sodium formate and sorbitol. One conventional tube was sent to the laboratory by ordinary chilled transport. Another conventional tube and one HG tube were transported to the laboratory without chilling. Cultures were performed upon arrival at the laboratory and then 24, 48 and 72 hours after primary sampling.

Of the 154 consecutive outpatients with suspected UTI, 144 had positive cultures, defined as > 103 cfu/ml. Twenty-four hours after sampling there were no significant differences in bacterial counts between the chilled conventional tubes and the HG tubes at room temperature. However, in the HG tubes a significant change in enterococcal counts were noted after 48 hours.156 [EL = 2+]

One study in the USA evaluated whether or not chemical preservatives in the Becton-Dickinson urine culture kit had an effect on urinalysis, microscopy or Gram stain. Of the 304 clean catch urine specimens obtained from pregnant women, 2% had significant bacteriuria (105 cfu/ml). There was complete agreement between preserved and unpreserved split samples in the detection of glucose, ketones, bilirubin and blood. Of the 388 women with symptoms of UTI seen in the emergency room or outpatient department, 198 (51%) had significant bacteriuria.

Urine microscopy revealed a tendency for erythrocyte counts to be diminished after 24 hours at room temperature in unpreserved specimens. Gram stain results of preserved and unpreserved split samples were comparable; staining characteristics were not altered by the preservative.157 [EL = III]

One study in the UK compared methods of preservation with simulated specimens of pooled urine seeded with known five parallel comparisons of six species.158 One strain each of Escherichia coli, Pseudomonas aeruginosa, Klebsiella aerogenes, Proteus mirabilis, Micrococcus and Streptococcus faecalis were isolated from infected urine. An overnight culture of each test strain in pooled urine was serially diluted to give six simulated specimens of 10, 103, 104, 105, 106 and 107. In unpreserved specimens at room temperature each test strain multiplied rapidly and the surface viable counts showed concentrations of between 107 and 108 cfu/ml within 72 hours in every specimen. In refrigerated specimens the surface viable counts for all the specimens remained constant for 72 hours. In specimens preserved with 1.8% boric acid, the surface viable counts remained constant for 24 hours, but the viable counts of specimens infected with P. aeruginosa fell markedly. After 24 hours the viable counts of the E. coli specimens, except for the most heavily infected specimen, declined. The viable counts of specimens in the K. aerogenes, P. mirabilis, Micrococcus and S. faecalis and the specimen that was most heavily infected with E. coli remained constant for 72 hours. In specimens with 9% sodium chloride (NaCl)–0.9% polyvinyl-pyrrolidone there were no differences between the results obtained with polyvinyl-pyrrolidone of the two molecular weights. The surface viable counts of all specimens of E. coli fell markedly within 24 hours, except the viable count of the most heavily infected specimen, which fell more slowly. The viable counts of the most heavily infected K. aerogenes remained constant while the other specimens fell more slowly. The strain of Micrococcus grew in the specimens but after 24 hours the viable counts remained in the same range that they were in at time zero. The viable counts of S. faecalis specimens remained constant for 72 hours, but the viable counts of all specimens in the P. mirabilis and P. aeruginosa specimens fell markedly within 24 hours. [EL = 3]

One study in the USA evaluated the efficacy of collecting urine specimens in Becton-Dickinson tubes and subsequently screening them for bacteriuria with the Abbott MS-2.159 Following collection, urine samples were immediately placed in the Becton-Dickinson tube and another in a screw-cap tube routinely used for transporting urine from the hospital to the laboratory. If samples could not be transported within 20 minutes, the conventional tube was refrigerated.

Of the 312 mid-stream urine specimens collected from obstetric outpatients receiving prenatal care, 124 were positive for bacteriuria. The median time required for urine specimens to be judged positive by the MS-2 was similar for conventional tube and for Becton-Dickinson tubes (95 and 105 minutes, respectively). Bacterial specimen results from conventional tubes did not differ significantly from those from Becton-Dickinson tubes. Culture results from 24 hour delayed samples from the Becton-Dickinson tubes were significantly different in that 40 of the 188 specimens had colony counts in excess of 105 cfu/ml. [EL = III]

One study in the USA aimed to determine whether boric acid interferes with the reactions of the Chemstrip LN dipstick.160 A preliminary study of specimens negative for leucocyte esterase and nitrite were obtained by multiple mid-stream urine collections into disposable non-sterile urine cups from one asymptomatic volunteer male. Specimens positive for leucocyte esterase and nitrite were prepared by placing Chek-Stix urinalysis control strips in 12 ml deionised water, following the manufacturer’s instructions. The positive and negative samples were then transferred to numbered Sage collection tubes containing boric acid. Twenty-one samples (12 negative and 9 positive) were tested immediately following preparation and tested again after 2 hours. Preliminary studies with the LN+ and LN− samples preserved in boric acid demonstrated no evidence of interference with the LN strips immediately after preparation, or after the 2 hour incubation.

Following the preliminary study, 177 consecutive clinical urine specimens from inpatients, out-patients and residents of a nursing centre preserved in boric acid were evaluated before routine culturing. The dipstick correctly indicated the presence or absence of leucocyte esterase and nitrite in all cases. [EL = 2+]

One study in the USA evaluated the boric acid-glycerol-sodium formate preservative in the Becton-Dickinson urine culture kit and the use of ordinary paper cups for collection of urine.161 Of 1000 urine samples from children and adults with symptoms suggesting UTI and from pregnant women being screened for asymptomatic bacteriuria, 88 of the initial reference cultures were positive (pure growth of 105 cfu/ml). Eighty-two (93.2%) of the 88 specimens on reference culture were also positive after refrigeration or holding at room temperature in the transport tube for 24 hours. There was one false positive culture from refrigerated urine but none from the transport tube. Mixing urine in the non-sterile container did not introduce detectable contamination. [EL = 3]

4.5.2. Time

Review findings– effect of time

Two studies were identified that investigated the effect of time on the multiplication of bacteria in urine samples.162,163

One study from the UK investigated the multiplication of contaminant bacteria in urine and attempted to define the duration of delay during which bacterial culture can be expected to give a reliable indication of the presence or absence of urinary infection.162 Samples were collected from 106 patients attending a health centre and members of the hospital staff and cultures were performed within 1 hour of voiding and successive cultures were carried out at 2, 4, 8, 12 and 24 hours after voiding. Throughout the period of sampling, specimens were kept between 19 °C and 23 °C. In the freshly voided urine, 14 of the 41 urine samples from males (34%) and five of 65 from females (7.7%) had bacterial populations of less than 102 cfu/ml. None of the urine samples from males had bacterial counts in excess of 105 cfu/ml, while four urine samples from females (6.2%) had counts exceeding 105 cfu/ml. In subsequent cultures enterococci, E. coli, Staphylococcus albus and group B streptococci were the organisms which most commonly multiplied in urine to give counts in excess of 105 cfu/ml within 24 hours of voiding. The lag phase was usually short and frequently undetectable. Enterobacteria other than E. coli were rarely isolated more than 102 cfu/ml when sampling was carried out but at later samplings showed growth patterns similar to E. coli. All isolates grew exponentially after approximately 8 hours, and most had a lag time of approximately 4 hours. [EL = 3]

One study in the USA evaluated the effect of transport delay on the microflora of clinical specimens collected for microbiological analysis.163 Clean catch urine specimens were collected from patients on medical wards and proportions of these specimens were cultured approximately 10 minutes after collection for aerobic organisms. The remainder of each specimen was kept at room temperature until collected by the transportation service. The time necessary for transportation of the urine specimens ranged from 2 to 5 hours with an average of 4 hours. The results from 100 urine specimens cultured immediately after collection indicated that 71% had colony counts of less than 102 cfu/ml; 14% between 104 and 105 cfu/ml; and 15% more than 106 cfu/ml. After transportation, 71% maintained colony counts of less than 102 cfu/ml, 9% between 104 and 105 cfu/ml, and 20% more than 106 cfu/ml. [EL = 3]

4.5.3. Temperature and refrigeration

Review findings – temperature

Two studies were found that evaluated temperature for urine samples.164,165

One study in Costa Rica evaluated the effect of time, temperature and glucose content on the growth of two initial populations of either E. coli or P. vulgaris in sterile urine samples.164 In urine containing no glucose, the original number of bacteria both in the urines and the controls showed little or no change over time. Populations of P. vulgaris remained unchanged at all three temperatures while E. coli showed a slight increase over time. In urine containing glucose all bacterial strains studied showed reductions in the populations after 2 hours of incubation at −10 °C and continued to decline at 4 hours and 8 hours. However, there was a steady increase in bacterial numbers with time in the samples incubated at room temperature (25 °C), which showed at least 105 cfu/ml organisms within 4 hours. The bacterial populations showed almost no change when the incubation temperature was 4 °C, regardless of bacterial strain. [EL = 3]

One study in the USA evaluated the minimum amount of urine necessary to obtain accurate results with the Sage urine culture tube and the Becton-Dickinson culture tube system.165 Both tubes were injected with 1, 2, 3 and 4–5 ml (tube capacity) of urine containing each culture. Specimens were held at 22 °C and cultured at 0, 4 and 24 hours. The Becton-Dickinson urine culture kits were toxic to E. coli and K. pneumoniae in specimens containing up to 2 ml of urine. The minimum useable amount of urine for reliable results was 3 ml. The Sage urine culture tube maintained the number of bacteria in 1 to 4.5 ml of urine in 83% of the specimens. However the Sage tube was toxic to E. coli when held for 24 hours. Quantitative counts of enterococci tended to significantly increase in specimens that contained 2 ml or more of urine in either system. [EL = 3]

Review findings – refrigeration

Two studies investigated the effect of refrigeration on bacterial growth in urine samples.166,167

One study from the USA assessed the validity of overnight refrigeration for quantitative bacteriological evaluation and compared initial urine cultures (less than 2 hours old) with refrigerated urine cultures.166 Of 414 urine cultures, there were 109 cultures with colony counts of 104 cfu/ml or higher. Four cultures changed from sterile to significant colony count (105 cfu/ml or greater), all of which were Staphylococcus aureus. There was also a single culture which changed from 105 cfu/ml to sterile where the organism involved was E. coli. Nine other cultures exhibited some change in colony count in which a number of organisms were involved in the discrepancies. [EL = 3]

One study in the USA evaluated whether bacterial concentrations generally considered insignificant (less than 104 cfu/ml) become significant as a result of bacterial multiplication in the urine during refrigeration.167 Clean catch specimens obtained from ‘normal’ males and females were refrigerated at 5 °C for approximately 24 hours. The urine was then pooled, sterilised by pressure filtration and stored at 5 °C in 100 ml aliquots in sterile bottles. Two bottles were inoculated for each of the bacteria employed and the bottles were placed at 0.5, 5, 10 and 15 °C. Every 24 hours for 4 days samples of urine from each bottle were cultured. At 0.5, 5 and 10 °C, E. coli remained largely unchanged.

At 15 °C, E. coli grew from 12 000 cfu/ml immediately after collection to 16 000 cfu/ml at 24 hours, 370 000 cfu/ml at 48 hours and reached 800 000 cfu/ml by 72 hours. Bacterial counts overall remained the most stable in the 5 °C group. [EL = 3]

Evidence statement – urine preservation

The studies included confirm the need for a method of preserving urine specimens when they cannot be examined immediately.

Culture of urine within 4 hours of voiding is likely to give a true indication of the presence or absence of bacteria. With further delay the interpretation of a heavy growth of bacteria in urine becomes progressively more unreliable. Where it is impractical to culture urine within 4 hours, urine specimens which are to be used to detect bacteriuria should be refrigerated immediately following collection.

There is evidence to suggest that culture kits containing boric acid, sodium formate and sodium borate maintain a stable bacterial population in urine for up to 24 hours. However, prolonged storage (more than 24 hours) may alter subsequent bacterial counts. Potential toxicity against bacteria in the specimen from boric acid can occur if the manufacturer’s recommendations about the volume of urine required are not followed. There is no evidence that commercially available urine collection kits offer any advantage.

4.6. Urine testing

Introduction

Prompt and accurate diagnosis of UTI is essential if this condition is to be managed correctly. The first step in making a diagnosis is to identify whether children presenting to the healthcare system, often but not exclusively via primary care, have a UTI. The initial assessment will usually involve a combination of clinical assessment and diagnostic testing.

Diagnostic tests fall functionally into two groups: firstly, those which give immediate results and, secondly, those in which, due to the nature of the test, there is a delay. There are obvious practical advantages in tests which give an immediate answer. Dipstick testing and microscopy fall into the first group and as such can assist in making an immediate assessment. Investigations involving bacterial culture fall into the second as an overnight incubation is required to allow bacteria to grow. The aim of this chapter is to review the evidence for the use of each test and make recommendations on how best to investigate a patient presenting with symptoms of UTI.

At present there is wide variation in practice. At one end of the spectrum all patients with possible UTI may be tested with a combination of dipstick and formal urine microscopy and culture. At the other end diagnostic testing might not be used until the patient has failed to improve following a course of empirical therapy. There is also wide variation both in the type of dipstick used as a near-patient test and in how microbiology laboratories perform microscopy and culture.

The cut-off value of 105 cfu/ml was proposed by Kass41,168 as it enabled 96% of UTI cases to be identified correctly when applied to adult women with asymptomatic bacteriuria and acute pyelonephritis/upper urinary tract infection.

Bacterial counts as low as 1000 cfu/ml can, in certain unusual clinical situations, represent a true UTI but when bacterial numbers are lower than 105 cfu/ml the chance of the identified bacteria representing contamination increases. Mixed growth can also represent a real infection, for example when the infecting bacteria are ‘hidden’ among a larger number of contaminating bacteria or in children with severe malformations in whom multi-bacterial infections occur.

The results from urine culture can therefore not be interpreted in isolation, but should be done in relation to the clinical setting, symptoms and findings. The results of other diagnostic tests should also be considered.

Clinical questions

In infants and children with suspected UTI, which is the most diagnostically accurate urine test for detecting UTI?

In infants and children with suspected UTI, which is the most effective diagnostic test?

4.6.1. Dipstick urine tests

Introduction

Dipstick tests are a group of tests which involve dipping reagent strips into collected urine.

Review findings – dipstick uring tests

A systematic review identified 38 studies that evaluated dipstick tests for the diagnosis of UTI. The studies included dipstick tests for nitrite, leucocyte esterase, protein, glucose and blood.143 [EL = II] A further meta-analysis identified 70 studies169 and two additional studies were identified from the literature search.170,171

Nitrite

A systematic review reported 27 data sets from 23 studies investigating nitrite dipstick tests.143 Culture was used as the reference standard in all but two studies, where a combination of culture and microscopy was used as the reference standard. The majority of studies used 105 cfu/ml as a positive reference standard. The studies reported poor sensitivity, ranging from 16.2 (specificity 97.6%) to 88.1% (specificity 100%), and high specificity, ranging from 75.6% (sensitivity 61.1%) to 100% (sensitivity 16.7–88.1%). Only two specificity estimates were below 90%. LR+ values ranged from 2.5 (LR− = 0.51) to 439.6 (LR− = 0.63). LR− values ranged from 0.12 (LR+ = 157) to 0.86 (LR+ = 6.7). The pooled LR+ was 15.9 (95% CI 10.7 to 23.7) and the pooled LR− was 0.51 (95% CI 0.43 to 0.60), but there was considerable heterogeneity in terms of LRs (P < 0.001).143 [EL = II]

Leucocyte esterase

A systematic review identified 14 studies reporting 16 data sets which investigated leucocyte esterase dipstick tests.143 Twelve studies used culture as the reference standard and two used a combination of culture and microscopy.

Sensitivity ranged from 37.5% (specificity 96.4%) to 100% (specificity 92%). Specificity ranged from 69.3% (sensitivity 93.5%) to 97.8% (sensitivity 70%). LR+ values ranged from 2.6 (LR− = 0.39) to 32.2 (LR− = 0.31). LR− values ranged from 0.02 (LR+ = 12.5) to 0.66 (LR+ = 6.97). There was considerable heterogeneity in both positive and negative LRs (P < 0.001). The pooled LR+ was 5.5 (95% CI 4.1 to 7.3) and the pooled LR− was 0.26 (95% CI 0.18 to 0.36).143 [EL = II]

Protein

A systematic review identified two studies reporting three data sets that examined protein dip-stick tests.143 One study used culture and the other used a combination of culture and microscopy as the reference standard. The systematic review concluded that these studies did not use an appropriate spectrum of patients or adequately report the criteria used to select the patients. The studies did not report sufficient information to assess the avoidance of review bias. The sensitivity was estimated to range from 8.1% (specificity 95.1%) to 53.3% (specificity 83.9%). Both studies found protein dipstick was a poor test for the identification of UTI. [EL = II]

Glucose

A systematic review identified four studies containing five data sets investigating biochemical test strips for glucose using culture as the reference standard.143 The studies identified investigated glucose strips which are not currently commercially available in the UK, as currently available glucose strips are optimised to detect abnormally high urinary glucose levels. Sensitivity ranged from 64% to 98% while specificity ranged from 96.4% to 100%. LR+ values ranged from 27.8 (LR− = 0.07) to 166.2 (LR− = 0.02) while LR− values ranged from 0.02 (LR+ = 166.2 and 113.7) to 0.36 (LR+ = 32.5). The pooled LR+ was 66.3 (95% CI 20.0 to 219.6) and the pooled LR− was 0.07 (95% CI 0.01 to 0.83). There was significant heterogeneity in both the positive and negative LRs (P < 0.001). [EL = II]

Blood

A systematic review identified one study investigating the accuracy of dipstick tests for blood using culture as the reference standard.143 The study reported that dipstick testing with blood is not a useful tool for diagnosing UTI in children, with estimated sensitivities of 25.4% for visual examination and 53.3% for automated examination, and specificities of around 85%. [EL = II]

Leucocyte esterase or nitrite positive

A systematic review identified 15 studies containing 20 data sets examining the use of a combination test where either a positive leucocyte esterase dipstick or a positive nitrite dipstick was considered a positive UTI result.143 All studies used culture as the reference standard. Sensitivity ranged from 69.4% (specificity 78.5%) to 100% (specificity 88.4%). Specificity ranged from 69.2% (sensitivity 94.1%) to 97.8% (sensitivity 70%). LR+ values ranged from 3.0 (LR− = 0.23) to 32.2 (LR− = 3.1) while LR− values ranged from 0.03 (LR+ = 5.6) to 0.39 (LR+ = 3.2). However, LRs showed considerable heterogeneity (P < 0.001). The pooled LR+ was 6.1 (95% CI 4.3 to 8.6) and the pooled LR− was 0.20 (95% CI 0.16 to 0.26). [EL = II]

Leucocyte esterase and nitrite positive

A systematic review identified nine studies containing 12 data sets examining the use of a combination test where a positive results from both leucocyte esterase and nitrite dipstick was considered a positive UTI result.143 All studies used culture as the reference standard. Sensitivity ranged from 30% (specificity 100%) to 89.2% (specificity 97.6%). Specificity ranged from 89.2% (sensitivity 87%) to 100% (sensitivity 30–88%). LR+ values ranged from 8.0 (LR− = 0.15) to 197.1 (LR− = 0.17) while LR− values ranged from 0.11 (LR+ = 36.7) to 0.7 (LR+ = 107.7). Both pooled positive and negative LRs were heterogeneous (P < 0.037 and P < 0.001, respectively). The pooled LR+ was 28.2 (95% CI 15.5 to 43.4) and the pooled LR− was 0.37 (95% CI 0.26 to 0.52). [EL = II]

Leucocyte esterase and protein positive

A systematic review identified one study investigating the use of a combination test where a positive result from both leucocyte esterase and protein dipstick was considered a positive UTI result.143 A combination of microscopy and culture was used as the reference standard. The study reported a sensitivity of 89% and a specificity of 95%; an LR+ of 17.4 and LR− of 0.12 were calculated. [EL = II]

Combinations of three dipsticks

A systematic review identified five studies reporting a total of ten data sets investigating various combinations of three dipsticks.143 Four studies evaluated one combination of tests (nitrite, blood or protein positive; nitrite, blood or leucocyte esterase positive; nitrite, blood and leucocyte esterase positive; nitrite, leucocyte esterase or protein positive) and two further studies investigated the same combination (nitrite, leucocyte esterase and protein positive). All studies used culture as the reference standard.

Insufficient information was available to draw any overall conclusions, but one combination (nitrite, leucocyte esterase and protein positive) investigated by two studies appeared to be potentially useful for diagnosing UTI. One study reported a sensitivity of 96% and a specificity of 99%, LR+ of 69.2, LR− of 0.04, while the second study reported a sensitivity of 89% and a specificity of 72%, LR+ of 3.1 and an LR− of 0.17. [EL = II]

A meta-analysis of urine dipstick tests to rule out infection identified 70 studies.169 Accuracy of nitrites was higher in pregnant women (diagnostic odds ratio (DOR) = 165) and in elderly people (DOR = 108). Subgroup analysis of diagnostic accuracy found ten studies of nitrite dipstick tests in children. Sensitivity was 50% (42% to 60%), specificity 92% (87% to 98%) with DOR 34 (12 to 97). Accuracy of leucocyte esterase was high in studies in urology patients (DOR = 267). The combination of both test results showed an increase in sensitivity. Accuracy was high in studies in urology patients (DOR = 52), in children (DOR = 46) and if clinical information was present (DOR = 28). Subgroup analysis of accuracy of leucocyte esterase and nitrite dipsticks in combination found nine studies of nitrite dipstick tests in children. Sensitivity ranged from 78% to 89% and specificity ranged from 79% to 91% with DOR 46 (23 to 95). [EL = II]

One study investigated whether dipstick urinalysis for leucocyte esterase, nitrites, blood and protein in the paediatric population is an adequate screening tool to exclude UTI.170 Prevalence of UTI overall was calculated to be 10.7% in a paediatric population with a higher prevalence (15%) in children younger than 2 years, and lower prevalence in children 2 years or older but younger than 10 years (7%). The sensitivity of the dipstick in all cases was 92.5% (95% CI 84.3% to 100%), specificity 39.4% (95% CI 34.2% to 44.6%), LR+ of 1.52 and an LR− of 0.18. The sensitivity of the dipstick in children aged 0–2 years was 87.5% (95% CI 74.3 to 100%), specificity 39.7% (95% CI 31.5 to 47.9%), LR+ of 1.47 and LR− of 0.30. The sensitivity of the dipstick in children 2 years or older but younger than 10 years was 100% (95% CI 100% to 100%), specificity 39.2% (95% CI 32.4% to 46%), and LR+ 1.64. The LR− was not estimable. [EL = II]

One study assessed the clinical utility of pathogen-specific tests to be applied with widely used dipsticks.171 The false negative rate for leucocyte esterase or nitrite dipstick tests was 5% (80/1743), false positive rate 17% (304), true positive rate 15% (262) and true negative rate 63% (1097).

The false negative rate for the immuno-chromatography strip was 10% (168/1743), false positive rate 2% (42), true positive rate 10% (174) and true negative rate 78% (1359). The false negative rate for combination leucocyte esterase, nitrite dipstick and immuno-chromatography tests was 11% (190/1743), false positive rate 1% (19), true positive rate 9% (152) and true negative rate 79% (1382). [EL = II]

Further analysis on dipstick urine testing by the NCC-WCH

A further analysis using data included in the systematic review143 was conducted to explore differences in diagnostic values between age and parameters (e.g. leucocyte esterase and nitrite).

Age

When compared within the same studies, the same trend for leucocyte esterase and nitrite was found in both children younger and older than 2 years. However, in children younger than 2 years, the diagnostic performance of the leucocyte esterase and nitrite dipstick, in particular the LR−, is poorer. (Table 4.5)

Table 4.5. Comparison of diagnostic values of various combinations of urine dipstick test between infants and children.

Table 4.5

Comparison of diagnostic values of various combinations of urine dipstick test between infants and children.

To examine this further, a meta-analysis was conducted to obtain both summary positive and negative LRs stratified by age for the performance of dipsticks compared with culture. The first set of figures examines the situation where both leucocyte esterase and nitrite are positive. Figure 4.1 shows the summary LR+. There was a clear divergence in performance with a summary LR+ for younger children of 7.74 (95% CI 1.88 to 31.93), while that for older children was 28.79 (95% CI 13.92 to 59.52).

Figure 4.1. Positive likelihood ratios for both leucocyte esterase and nitrite positive, stratified by age.

Figure 4.1

Positive likelihood ratios for both leucocyte esterase and nitrite positive, stratified by age.

The next analysis (Figure 4.2) shows the summary LR− when both leucocyte esterase and nitrite are negative. The summary LR− for younger children was 0.32 (95% CI 0.16 to 0.63). For children over 2 years the summary LR− was 0.19 (95% CI 0.09 to 0.40).

Figure 4.2. Negative likelihood ratio for both leucocyte esterase and nitrite negative, stratified by age.

Figure 4.2

Negative likelihood ratio for both leucocyte esterase and nitrite negative, stratified by age.

Leucocyte esterase versus nitrite

When compared within the same studies, there was a tendency for LR+ to be higher for nitrite and LR− was lower for leucocyte esterase (Table 4.6).

Table 4.6. Comparison of diagnostic values of dipstick urine tests for children (leucocyte esterase versus nitrite).

Table 4.6

Comparison of diagnostic values of dipstick urine tests for children (leucocyte esterase versus nitrite).

Evidence statement – dipstick urine tests

It is clear that leucocyte esterase and nitrite dipsticks are more valuable in diagnosing UTI when used in combination than when used alone. There is general agreement among studies that a combination of a positive leucocyte esterase with positive nitrite has the highest LR+ and is the most useful dipstick test for ruling in UTI. However, a negative result for either leucocyte esterase or nitrite has the highest LR− and will be most useful in excluding UTI. It is important to note that in children younger than 2 years the dipsticks are less reliable in both scenarios.

Table 4.7Comparison of diagnostic values of dipstick urine tests for children (leucocyte esterase and/or nitrite)

StudyCountryAgeManufacturerReference testLeucocyte esterase or nitriteLeucocyte esterase and nitrite
SensitivityLR+SensitivityLR+
SpecificityLR−SpecificityLR−
Sharief (1998)UK< 1 year75.0%2.712.5%13.0
n = 12473.3%0.3899.1%0.84
1–16 years80.0%3.855.6%35.6
n = 20978.6%0.2598.4%0.45
Shaw (1991)USA< 2 years71.4%9.414.3%6.24
n = 14592.4%0.3197.7%0.88
2–19 years87.1%3.951.6%27.1
n = 34677.8%0.1798.1%0.49
Dayan (2002)USA< 2 monthsSuper UASPA 103 cfu/ml85.0%10.530.0%
n = 193Clean catch 105 cfu/ml91.9%0.16100.0%
Wiggelinkhuizen (1988)South AfricaCombur9 (BM)95.4%3.061.4%58.5
68.7%0.0799.0%0.39
Multistix (Ames)94.1%3.162.0%47.8
69.2%0.0898.7%0.38
Marsik (1986)USA≤ 21 yearsChemstrip (Biodynamics)SPA any bacteria88.7%3.162.3%27.2
n = 601Catheter 103 cfu/ml71.5%0.1697.8%0.39
Clean catch 105 cfu/ml
Woodward (1993)UK≤ 15 yearsMultistix (Bayer)105 cfu/ml and WBC100.0%8.683.3%
n = 13420/ml88.4%0.04100.0%

Glucose dipstick tests may be useful for both ruling in and ruling out UTI, but evidence is limited.

There is not enough evidence to draw conclusions about dipstick tests for protein or blood, or for combinations of three or more dipstick tests.

Nitrite has a higher LR+ but a higher LR− than leucocyte esterase. When both leucocyte esterase and nitrite are positive the dipstick test is useful to rule in a diagnosis of UTI in patients 2 years or older (LR+ = 28.79). When both leucocyte esterase and nitrite are negative the dipstick test is useful to rule out a diagnosis of UTI in patients over 2 years of age (LR− = 0.19).

4.6.2. Microscopy

Introduction

The performance and interpretation of microscopy is more demanding than dipsticks. A variety of cellular elements can be identified in urine (e.g. white cells, red cells, bacteria and casts) by a number of different microscopic methods including inverted microscopy, Gram stain and centrifuged deposit.

Review findings – pyuria

A systematic review reported 27 studies (49 data sets) investigating the microscopic detection of pyuria. Twenty-four studies used culture as the reference standard and three studies used culture and automated microscopy. Only half the studies included an appropriate spectrum of patients and ten studies did not provide an adequate description of patient selection. Most studies did not provide enough information to assess the avoidance of review bias. One-third did not provide an adequate description of the test and/or the reference standard. Several studies reported results for different cut-off points.

Sensitivity ranged from 37% (specificity 93%) to 96% (specificity 96%). Specificity ranged from 32% (sensitivity 89%) to 100% (sensitivity 50%).

LR+ values ranged from 1.3 (LR− = 0.33) to 27.7 (LR− = 0.09). LR− values ranged from 0.04 (LR+ = 24.0) to 0.68 (LR+ = 5.3). Likelihood ratios showed considerable heterogeneity (P < 0.001). The pooled LR+ was 5.9 (95% CI 4.1 to 8.5) and the pooled LR− was 0.27 (95% CI 0.20 to 0.37).

ROC curves suggested that the considerable heterogeneity between studies was not just the result of different cut-off values but was likely to be caused by other factors. Regression analysis indicated that centrifugation of the sample, description of selection criteria, test bias, review bias, description of study withdrawals and age were significantly associated with the heterogeneity observed. Multivariate analysis showed that only two items remained significant; centrifugation of the sample and reporting selection criteria. The DOR was 20% less in samples centrifuged compared with non-centrifuged samples and three times greater in studies that provided an adequate description of selection criteria.143 [EL = II]

Review findings – bacteriuria

A systematic review reported 22 studies (including 34 data sets) evaluating the microscopic detection of bacteriuria. Nineteen studies used culture as the reference standard. One study used culture and microscopy as the reference standard and a further two studies used culture and automated microscopy as the reference standard. Approximately half did not include an appropriate spectrum of patients, eight studies did not provide selection criteria, and only four studies reported blinding. One-third of studies did not provide adequate descriptions of the test and/or reference standard.

Sensitivity ranged from 52.4% (specificity 99%) to 100% (specificity 98%) and specificity ranged from 40% (sensitivity 93%) to 99.7% (sensitivity 96%). LR+ values ranged from 1.6 (LR− = 0.17) to 304.8 (LR− = 0.04) and LR− values ranged from 0.01 (LR+ = 3.4) to 0.48 (LR+ = 3.5). Likelihood ratios showed considerable heterogeneity (P < 0.001). The pooled LR+ was 14.7 (95% CI 8.7 to 24.9) and the pooled LR− was 0.19 (95% CI 0.14 to 0.24). ROC curves indicated that although different cut-off points may account for some of the heterogeneity, it is likely that other factors may be contributing to test performance. In univariate regression analysis, Gram stain and incorporation bias were shown to be significant and both remained significant in multivariate analysis. The DOR was 5.5 times greater in samples that were Gram stained, and in studies where incorporation bias was not present the DOR was 100 times greater.143 [EL = II]

One study compared the accuracy in diagnosing significant bacteriuria between quantitative unspun-urine microscopy and the Gram stain method.172 Significant bacteriuria was detected by urine culture in 37 out of 325 urine samples.

Unspun-urine microscopy samples in cell-counting chambers were negative in 248 samples, positive in 33 and ambiguous in 44. Ambiguous samples were subjected to oil-immersion microscopy, which made it possible to identify rods, cocci, salts or other particles. Overall, unspun-urine microscopy was able to detect bacteriuria in 35 of 37 urine samples with culture-proven significant bacteriuria (sensitivity 94.6%), failing to identify bacilli in two urine samples. Unspun-urine microscopy identified 286 of 288 urine samples with negative culture results (specificity 99.3%). Gram stain method was able to detect bacteriuria in 33 of 37 urine samples with culture-proven significant bacteriuria (sensitivity 89.2%). The Gram stain method identified 284 of 288 urine samples with negative culture results (specificity 98.6%). Both the unspun-urine microscopy and the Gram stain methods were similarly reliable when compared with culture. [EL = II]

One study compared the accuracy of the differential fluorescent staining method and the Gram stain method in screening for bacteriuria with conventional culture.173 A total of 1487 urine samples were tested: 289 were found to have colony counts greater than 104 cfu/ml, 237 yielded a single organism and 52 a mix of two or more organisms.

Of the 237 yielding a single organism, 224 were detected by the differential fluorescent staining method and 162 by the Gram stain (13 undetected by the differential fluorescent staining method and 75 undetected by the Gram stain).

The sensitivity of the differential fluorescent staining method was 95%, specificity 92%, LR+ 11.8 and LR− 0.05. The sensitivity of the Gram stain was 68.3%, specificity 76%, LR+ 2.83 and LR− 0.42. The specificity of the differential fluorescent staining method was 91.6% and the Gram stain 75.8%. [EL = III]

Review findings – pyuria or bacteriuria

A systematic review reported eight studies (including ten data sets) investigating pyuria or bacteriuria where a positive result from either test was taken as a positive result for UTI.143 The majority did not provide adequate information to assess the avoidance of test review bias (blinding). Sensitivity ranged from 75% (specificity 93%) to 100% (specificity 32%) and specificity ranged from 32.3% (sensitivity 100%) to 92.9% (sensitivity 75%). LR+ values ranged from 1.5 (LR− = 0.05) to 12.9 (LR− = 0.05). LR− values ranged from 0.02 (LR+ = 2.8) to 0.27 (LR+ = 4.1 and 10.5). Likelihood ratios showed considerable heterogeneity (P < 0.001). The pooled LR+ was 4.2 (95% CI 2.3 to 7.6) and the pooled LR− was 0.11 (95% CI 0.05 to 0.23). ROC curves indicated that the considerable heterogeneity between studies was not just the result of different cut-off points but was likely to be caused by other factors. There was insufficient data to investigate heterogeneity further using regression analysis. [EL = II]

Review findings – pyuria and bacteriuria

A systematic review reported eight studies (including ten data sets) investigating combinations of pyuria and bacteriuria where a positive results from both tests was taken as a positive result for UTI.143 All studies used culture as the reference standard. The majority of studies included an appropriate spectrum of patients, although they did not provide adequate information to assess test review bias (blinding).

Sensitivity ranged from 46.7% (specificity 96%) to 93.1% (specificity 98%) and specificity ranged from 73.6% (sensitivity 71%) to 99.7% (sensitivity 84%). LR+ values ranged from 2.7 (LR− = 0.04) to 281 (LR− = 0.16). LR− values ranged from 0.07 (LR+ = 41) to 0.56 (LR+ = 11). Likelihood ratios showed considerable heterogeneity (P < 0.001). The pooled LR+ was 37.0 (95% CI 10.9 to 125.9) and the pooled LR− was 0.21 (95% CI 0.13 to 0.36). ROC curves indicated that the considerable heterogeneity between studies was not just the result of different cut-off points but was likely to be caused by other factors. [EL = II]

Evidence statement – microscopy

Given the heterogeneity between studies and the lack of data for combinations of microscopy tests, it is difficult to draw overall conclusions about the diagnostic accuracy of microscopy for detecting UTI. However, the pooled likelihood ratios show that a negative result for either pyuria or bacteriuria (LR− = 0.11; 95% CI 0.05 to 0.230) is better at ruling out UTI than dipstick testing.

A systematic review concluded that presence or absence of bacteriuria is considerably better than presence or absence of pyuria for ruling in or ruling out UTI. The diagnostic performance of bacteriuria may be improved when combined with pyuria, but there is insufficient evidence to provide certainty in these estimates.

4.6.3. Culture

Review findings – dipslide culture

A systematic review reported eight studies investigating the accuracy of dipslide culture for the diagnosis of UTI.143 More than half did not use an appropriate spectrum of patients, did not report selection criteria and did not provide an adequate description of the test and/or reference standard. The majority of studies did not provide adequate information to assess test review bias (blinding).

Reported diagnostic values for dipslide culture compared with standard culture are as follows: sensitivity ranged from 56.3% (specificity 97%) to 100% (specificity 92%) and specificity ranged from 70.7% (sensitivity 78%) to 100% (sensitivity 83%). LR+ values ranged from 2.7 (LR− = 0.31) to 135.4 (LR− = 0.17) and LR− values ranged from 0.02 (LR+ = 12.18) to 0.46 (LR+ = 7.8). There was considerable statistical heterogeneity in both positive and negative LRs (P < 0.001). The pooled LR+ was 14.6 (95% CI 6.7 to 31.8) and the pooled LR− was 0.23 (95% CI 0.14 to 0.39). ROC curves indicated considerable heterogeneity across the studies with no clear outliers. There were not enough studies to investigate heterogeneity further using regression analysis. [EL = II]

One study assessed the validity of urine dipslides performed under daily practice conditions and assessed the influence of the incubation period (24 versus 48 hours) on validity.174 The nitrite test was the initial test in all practices. For the 268 urine samples a sensitivity of 42% (95% CI 34% to 49%) and a specificity of 95% (95% CI 89% to 98%) was reported. The LR+ was 8.40 and the LR− was 0.61. The sensitivity of the dipslide in general practice after 24 hours incubation was 73% (95% CI 66% to 80%) and specificity was 94% (95% CI 88% to 98%). The LR+ was 12.17 and the LR− was 0.29. Overall, the dipslide read under practice conditions performed less well than when performed under optimal conditions. [EL = II]

One study evaluated the diagnostic performance of the DipStreak device (using two different medium formulations) compared with Uriselect 3 plates and the reference streak method (calibrated loop).175 In the study comparing DipStreak (CHROMagar and MacConkey media), Uriselect 3 plates and calibrated loop culture, 2000 urine samples were processed and 511 cultures were found to be positive. The CHR DipStreak device, the Uriselect 3 and calibrated loop cultures gave the same detection rate (99.7%). For the direct identification of E. coli, Proteus and Enterococcus isolates, the DipStreak device and Uriselect showed overall sensitivities of 97% and 93.4%, respectively. In a second study comparing DipStreak, Uriselect 3 and MacConkey media, 3000 urine samples were processed and 714 cultures were found to be positive. The DipStreak device, the Uriselect 3 and calibrated loop cultures gave detection rates of 99.4%, 99.9% and 99.2%, respectively. For the direct identification of E. coli, Proteus and Enterococcus isolates, the DipStreak device and Uriselect plates showed overall sensitivities of 88.7% and 94.4%, respectively. [EL = III]

Evidence statement – dipslide culture

There is not enough evidence to draw conclusions about different methods of dipslide culture for detecting UTI in children. The pooled LR− for dipslide culture was 0.23 (see Figure 4.3 in Section 4.8).

Figure 4.3. Leucocyte esterase (LE) or nitrite dipstick, microscopy and dipslide culture plotted in ROC space.

Figure 4.3

Leucocyte esterase (LE) or nitrite dipstick, microscopy and dipslide culture plotted in ROC space.

4.6.4. Combinations of two or more methods

Review findings – combinations of tests

A meta-analysis of urine screening tests for UTI in children concluded that rapid dipstick tests could not be definitively assessed because of the small number of studies assessing their effectiveness. Bivariate summary ROC (SROC) curves showed that pyuria more than 10/hpf and bacteriuria more than 10/hpf had the best diagnostic performance. In multivariate analysis, both remained significant.176 [EL = II]

One study evaluated the diagnostic properties of urine Gram stain and urine microscopic examination for screening UTI. The prevalence of UTI from culture was 54.7% (52 cases).177 The sensitivity of the Gram stain was 96.2%, specificity 93.0%, LR+ 13.71 and LR− 0.04. The sensitivity of the microscopic examination was 65.4%, specificity 74.4%, LR+ 2.50 and LR− 0.47. Combining the Gram stain and the microscopic examination, the sensitivity was 98.1%, specificity 74.4%, LR+ 3.77 and LR− 0.03. [EL = Ib]

One study aimed to determine which method best identified UTI in children younger than 5 years presenting to a paediatric emergency department.178 Twenty-five cases (17.6%) of UTI were diagnosed by culture, 48% were younger than 12 months and 16% were male. Positive leucocyte esterase dipstick had an overall sensitivity of 48%. In children younger than 12 months, sensitivity was 42% while in children 12 months or older, sensitivity was 53%. Positive nitrite dipstick had an overall sensitivity of 20%. In children younger than 12 months, sensitivity was 17% while in children 12 months or older, sensitivity was 23%. Positive blood dipstick had an overall sensitivity of 44%. In children younger than 12 months, sensitivity was 33% while in children 12 months or older, sensitivity was 53%. Positive unspun leucocyte count > 10/μl had an overall sensitivity of 68%. In children younger than 12 months, sensitivity was 67% while in children 12 months or older, sensitivity was 69%. Positive cyto-centrifuge Gram stain had an overall sensitivity of 60%. There was a statistically significant difference between children younger than 12 months (sensitivity 42%) and children 12 months or older (sensitivity 76%) (P < 0.05). 2 to 5 or more leucocytes/hpf in sediment had an overall sensitivity of 48%. In children younger than 12 months, sensitivity was 42% while in children 12 months or older, sensitivity was 53%. [EL = II]

One study compared the performance of leucocyte esterase and nitrite dipstick with the assessment of pyuria by microscopic examination and culture of urine samples in patients with symptoms of UTI.179 The sensitivity of the leucocyte esterase dipstick was 68.4% and specificity 73.4%, LR+ 2.52 and LR− 0.44. The sensitivity of the nitrite dipstick was 58.9%, specificity 77.8% LR+ 2.68 and LR− 0.53. The sensitivity of the microscopic pyuria count was 34%, specificity 86.5%, LR+ 2.62 and LR− 0.76. There was a significant correlation between dipstick results, microscopic examination and urine culture (P < 0.001). [EL = III]

One study investigated the validity of the urinary Gram stain compared with a combination of pyuria plus Gram stain and overall urinalysis.180 Of the 100 children, 70% had a positive urine culture. The sensitivity of the Gram stain was 80%, specificity 83%, LR+ 4.71 and LR− 0.24. The sensitivity of the combination of Gram stain and pyuria was 42%, specificity 90%, LR+ 4.20 and LR− 0.64.

The sensitivity of the overall urinalysis was 74%, specificity 3.5%, LR+ 0.77 and LR− 6.50. The study concluded that neither method (Gram stain, or the combination of Gram stain plus pyuria) should substitute for urine culture in symptomatic children. [EL = III]

Evidence statement – combinations of tests

There is not enough evidence to draw conclusions about combinations of different methods for detecting UTI in children and they cannot be considered cost-effective.

4.6.5. Other tests

Review findings – other tests

A study published in 1968 examined the triphenyltetrazolium chloride reduction (TCC) test and the nitrite test.143 One study evaluated three laboratory-based blood tests (peripheral WBC, erythrocyte sedimentation rate (ESR) and C-reactive proteins) in which all were found to be poor tests for diagnosing UTI. Other tests included FiltraCheck-UTI for bacteriuria, quantitative estimation of proteinuria and two studies of Uriscreen (reporting contrasting results). Only one study used an appropriate spectrum of patients and only two reported an adequate description of the test and/or the reference standard. Because of the small number of studies that examined these tests, there was insufficient information to assess their usefulness in diagnosing UTI. [EL = II]

One study evaluated the analytical performance of the Sysmex UF-100 cytometer compared with culture for diagnosing UTI.181 Of the 2010 patients considered, 529 (26.3%) had a UTI. Of the dipstick screening tests (leucocyte esterase and nitrite dipstick tests), 171 (8.5%) false negatives were observed and 184 (9.2%) false positives. Sensitivity was 64% and specificity 88%, with an LR+ of 5.33 and LR− of 0.41. Of the culture tests (bacterial growth on CLED agar), 56 (2.8%) false negatives were observed and 35 (1.7%) false positives, sensitivity was 89% and specificity 98%, with an LR+ of 44.50 and LR− of 0.11. Of the UF-100 tests, 29 (1.4%) false negatives were observed and 102 (5.1%) false positives. Sensitivity was 94% and specificity 93%, LR+ 13.43 and LR− 0.06. The Sysmex UF-100 performed more accurately than both the dipstick testing and culture. [EL = II]

Evidence statement – other tests

There is not enough evidence to draw conclusions about alternative diagnostic tests for identifying UTI in children.

4.6.6. Dipstick urine testing versus culture

Review findings

One study conducted in a laboratory aimed to determine whether the biochemical results of the urine dipstick could be used to eliminate unnecessary urine cultures.182

Of the 6192 urine samples processed, 64% (3932) had cultures performed. These were samples which showed positive dipstick and were ordered on physician request, or were not cancelled. Thirty-six percent (2260) had a negative dipstick and were cancelled. The rate of cancellation appeared consistent at approximately one-third when tracked month by month. Of the 3932 samples cultured, 22.4% (883) were true positives (positive dipstick and positive culture), while 31.8% (1248) had a positive dipstick but grew organisms that were considered contaminants. False positive results were observed in 1558 (39.6%). Of the samples that showed negative dip-stick and were cultured, 11 (0.3%) grew a clinically significant pathogen. The study concluded that the urine dipstick testing can be used as a screen to determine whether or not a urine culture should be performed and implementation of this policy has resulted in the elimination of up to one-third of the urine cultures performed in one laboratory. [EL = III]

One study investigated the sensitivity of standard urinalysis as a screening test for UTI in 11 089 patients who had urine cultured to determine how it varies with age and to determine the clinical situation that necessitates the sending of urine for culture regardless of the urinalysis result.183 In this study, a urinalysis was considered positive if the presence of one of the following was detected: leucocyte esterase, nitrite or pyuria. The study found that sensitivity of urinalysis was 82% (95% CI 79% to 84%) and did not vary with age. The specificity of urinalysis was 92% (95% CI 91% to 92%). The LR+ was 10.6 (95% CI 10.0 to 11.2) and the LR− was 0.19 (95% CI 0.18 to 0.20). [EL = III]

A study conducted in China evaluated the usefulness of culture of urine samples obtained by urethral catheterisation in diagnosing UTI in 492 uncircumcised boys compared with 460 girls aged 1–18 months (mean age 0.49 years) who had catheter urine cultures performed between July 1999 and June 2002 at a paediatric hospital and to test whether a single cut-off bacterial count has high sensitivity and specificity.184 Children were classified as group A if they had a positive culture of a urine sample obtained by urethral catheterisation, acute fever, positive leucocyte esterase and nitrite dipstick and leucocytes on microscopy, and a definite response to antibiotic treatment; and group B if they had cultures yielding no growth, urine culture positive but asymptomatic and had negative urinalysis results. Group A were used as the gold standard.

There were significantly higher counts in group A children than group B (P < 0.001) and group B had significantly more cases of mixed growth (P < 0.001). The probability of UTI was increased when cfu/ml was > 105 for uncircumcised boys (LR 20.2) and > 105 (LR 18.8) or 104–105 (LR 8.95) for girls. UTI was unlikely when cfu/ml were 10–103 (LR 0.11) or 103–104 (LR 0.45) for boys or if mixed growth was found (LR 0.21; 95% CI 0.12 to 0.37). [EL = III]

4.6.7. Dipstick urine testing versus microscopy

Further analysis shows that the diagnostic performance for dipstick urine testing varies with age. To examine which is the most appropriate test for each age group (infants and older children), studies that examined diagnostic values of both urine dipstick testing and microscopy compared with culture were extracted from the identified studies. Among these studies, only studies that examined results that were stratified by age group were used so only one study could be included.

Diagnostic values of both microscopy and dipstick urine testing are presented in Table 4.8. Two values are used as criteria for a positive microscopy, one used a cut-off value of 10 WBC counts/hpf for pyuria and moderate bacteriuria and the other used a cut-off value of 5 WBC counts/hpf for pyuria and few bacteriuria.

Table 4.8. Likelihood ratios of diagnosing UTI by age group.

Table 4.8

Likelihood ratios of diagnosing UTI by age group.

Children younger than 2 years

In children younger than 2 years where a cut-off of 5 WBC/hpf was used neither test performed well but dipstick testing again had a higher LR+ than microscopy: 6.24 versus 1.63 (95% CI 1.14 to 34.22 versus 1.24 to 2.13]. Where a cut-off of 10 WBC/hpf was used microscopy had a higher LR+ than dipstick testing at 27.10 versus 10.84 (95% CI 4.16 to 58.44 versus 1.14 to 34.22).

In children younger than 2 years where a cut-off of 5 WBC/hpf was used microscopy had a lower LR− than dipstick urine testing at 0.27 (95% CI 0.07 to 0.99) versus 0.31 (95% CI 0.13 to 0.71). Where a cut-off of 10 WBC/hpf was used microscopy had a higher LR+ than dipstick urine testing at 15.6 (95% CI 4.16 to 58.44) versus 6.24 (95% CI 1.14 to 34.22).

Children 2 years or older

When the use of microscopy, compared with dipstick testing, was examined for making a positive diagnosis of UTI in children 2 years or older where a cut-off of 5 WBC/hpf was used, dipstick testing had a higher LR+ than microscopy: 27.10 versus 1.69 (95% CI 11.44 to 64.21 versus 1.52 to 1.87). Where a cut-off of 10 WBC/hpf was used, dipstick testing again had a higher LR+ than microscopy at 27.10 versus 10.84 (95% CI 11.44 to 64.21 versus 5.95 to 19.75).

When the use of microscopy, compared with dipstick testing, was examined for ruling out a diagnosis of UTI in children 2 years or older where a cut-off of 5 WBC/hpf was used, microscopy had a lower LR− than dipstick testing at 0.04 versus 0.17 (95% CI 0.00 to 0.59 versus 0.07 to 0.41). Where a cut-off of 10 WBC/hpf was used, dipstick testing had a lower LR− than microscopy at 0.17 versus 0.51 (95% CI 0.07 to 0.41 versus 0.35 to 0.73).

The results of these analyses are presented in Table 4.8.

Overall, the evidence shows that to make a rapid diagnosis of UTI, the dipstick test has the highest LR+ in children aged 2 years or older, with microscopy using a cut-off of > 10 WBC/hpf having the highest LR+ for those younger than 2 years. To exclude a diagnosis of UTI in children younger than 2 years, microscopy with a cut-off of 5 WBC/hpf has a marginally better LR− than dipstick testing: 0.27 versus 0.31.

4.7. Localisation of UTI

Clinical question

In infants and children with suspected UTI, what is the most effective test for assessing the localisation of UTI?

4.7.1. Localisation of UTI by symptoms and signs

Review findings – localising UTI by symptoms and signs

A systematic review identified five studies assessing various clinical features for the localisation of UTI in children.143 A summary of the results is presented in Table 4.9. [EL = II]

Table 4.9. Clinical features compared with DMSA carried out during the acute illness for detecting UTI – results of included studies.

Table 4.9

Clinical features compared with DMSA carried out during the acute illness for detecting UTI – results of included studies.

Two studies compared body temperature with the reference standard of dimercaptosuccinic acid (DMSA) scintigraphy for diagnosing acute pyelonephritis/upper urinary tract infection. Test performance was poor in both studies, with one reporting sensitivity of 64% and specificity of 40%, an LR+ of 1.1 and LR− of 0.89 for a cut-off value of 39.1 °C, and the other reporting sensitivity of 87% and specificity of 64%, LR+ of 2.4 and LR− of 0.23 for a cut-off value of 38 °C.

Two studies evaluated the diagnostic accuracy of symptoms of acute pyelonephritis/upper urinary tract infection using DMSA scintigraphy as the reference standard. Sensitivities of 57% and 71% were found, with specificities of 100% in both studies. LR+ values were 4.5 and 26.6 while LR− values were 0.49 and 0.31, respectively.

One study assessed the presence of physical symptoms or positive laboratory findings for the diagnosis of acute pyelonephritis/upper urinary tract infection using DMSA scintigraphy as the reference standard. Sensitivity was 98%, specificity 33%, with an LR+ of 1.5 and an LR− 0.09.

Evidence statement – localising UTI by symptoms and signs

Clinical features used and the methods of determination were diverse and poorly described. They cannot used to predict pyelonephritic changes on acute DMSA.

4.7.2. Localisation of UTI by laboratory tests

Introduction

Attempts have been made to distinguish acute pyelonephritis/upper urinary tract infection from cystitis/lower urinary tract infection by using some laboratory investigations including C-reactive protein (CRP) and procalcitonin. The purpose of this review was to evaluate diagnostic values of these biochemical markers to differentiate upper and lower UTI. Results of DMSA scintigraphy were considered as the reference standard.

Review findings – localisation of UTI by laboratory tests

A systematic review identified seven studies which evaluated the accuracy of circulatory CRP for diagnosing acute pyelonephritis/upper urinary tract infection, all using DMSA scintigraphy as a reference standard (Table 4.10).143 Five additional studies were identified.185–189

Table 4.10. Summary CRP results from the included studies.

Table 4.10

Summary CRP results from the included studies.

In the systematic review it was reported that three studies using a concentration of 20 mg/ml to define a positive result reported sensitivity above 85%, while specificity ranged from 19% to 60%. The remaining four studies used varying concentrations (20 μg/l to 880 mg/l) and reported poor diagnostic performance. For the higher concentrations, sensitivity ranged from 65% to 70% and specificity from 55% to 86%. One study with very low concentration (20 μg/l) reported sensitivity of 14% and specificity of 100%.

The systematic review reported other laboratory analyses, but the small number of studies and the diverse methodologies and cut-off points make it difficult to draw any conclusions about the value of these laboratory-based tests for diagnosing UTI. [EL = II]

An Italian study investigating markers for localising UTI and renal damage reported values for procalcitonin and CRP at various levels.185 Children found to have moderate to severe acute pyelonephritis/upper urinary tract infection were significantly more likely to have longer duration of fever (P = 0.002), higher procalcitonin level (4.48 ± 5.84 ng/ml versus 0.44 ± 0.30 ng/ml; P < 0.001), higher CRP level (106.0 ± 68.8 mg/l versus 36.4 ± 26.0 mg/l; P < 0.001) and higher erythrocyte sedimentation rate (ESR) (79.1 ± 33.0 mm/hour versus 58.5 ± 33.1 mm/hour; P = 0.03) than the children with mild or no acute pyelonephritis/upper urinary tract infection. There were no differences between the groups in terms of age (P = 0.40), gender (P = 0.78) or leucocyte esterase count (P = 0.15). For the children with acute pyelonephritis/upper urinary tract infection, Table 4.11 shows the varying sensitivities, specificities and likelihood ratios for various levels of procalcitonin or CRP.

Table 4.11. Summary CRP and procalcitonin measures.

Table 4.11

Summary CRP and procalcitonin measures.

When inflammatory markers were correlated with the severity of renal lesions (on DMSA), a significant correlation was shown with both procalcitonin and CRP levels. However, when correlated in follow-up scans, only procalcitonin remained significant. [EL = II]

A study conducted in Taiwan assessed the usefulness of laboratory parameters for identifying UTI in 162 febrile infants younger than 8 weeks presenting to a hospital emergency department.186 (Table 4.12)

Table 4.12. Summary diagnostic measures.

Table 4.12

Summary diagnostic measures.

There were no significant differences in the areas under the ROC curves for the standard urinalysis, CRP or ESR, but the area under the curve for haemocytometer WBC counts was significantly better than the other laboratory parameters (P < 0.05) and total WBC count was significantly smaller (P < 0.05).

The most sensitive indicator to UTI was pyuria ≥ 10 WBC/μl (< 0.05). Pyuria ≥ 5 WBC/hpf had poor sensitivity but high specificity. The combination of pyuria ≥ 10 WBC/μl and CRP > 20 mg/l increased the specificity to 98%, while sensitivity decreased to 54%. The specificity of pyuria ≥ 10 WBC/μl combined with a positive ESR increased to 97%, while the sensitivity decreased significantly to 72%. UTI was significantly more likely when the urine had ≥ 5 WBC/hpf or ≥ 10 WBC/μl. [EL = II]

A study conducted in Switzerland measured procalcitonin levels in children aged 1 month to 16 years (mean age lower UTI 36 months, mean age acute pyelonephritis/upper urinary tract infection 42 months) with clinical signs of acute pyelonephritis/upper urinary tract infection, compared with other inflammatory markers, and evaluated its ability to predict renal involvement as assessed by DMSA.187

There were no differences in mean age (P = 0.35) or sex (P = 0.14) between groups. There were significant differences between children with cystitis/lower urinary tract infection and those with acute pyelonephritis/upper urinary tract infection in the leucocyte count (10 939 ± 834 mg/l versus 17 429 ± 994 mg/l; P < 0.001), procalcitonin level (0.38 ± 0.19 mg/l versus 5.37 ± 1.9 mg/l; P < 0.001) and CRP (30.3 ± 7.6 mg/l versus 120.8 ± 8.9 mg/l; P < 0.0001). For predicting renal involvement, CRP had a sensitivity of 100% and a specificity of 26.1%, and LR+ of 1.35 (the LR− was not estimable). Procalcitonin had a sensitivity of 70.3% and a specificity of 82.6%%, LR+ of 4.12 and an LR− of 0.36. [EL = III]

One study conducted in Turkey and one study from Israel were identified investigating clinical findings compared with DMSA for localising UTI in children.188,189 Neither of these studies reported raw numbers for sensitivity or specificity and were generally of poor quality. They should be interpreted with caution.

A study conducted in Turkey evaluated 76 patients (48 girls and 28 boys) aged 2 months to 12 years to investigate whether serum levels of pro-inflammatory cytokines and procalcitonin in children with UTI could be used as markers in distinguishing acute pyelonephritis/upper urinary tract infection.188 Significantly higher procalcitonin and pro-inflammatory cytokine levels were detected in children with acute pyelonephritis/upper urinary tract infection (P < 0.001). Using a cut-off value of 0.5 ng/ml, procalcitonin showed a sensitivity of 58% and a specificity of 76%, LR+ of 2.42 and LR− of 0.55. Using a cut-off value of 20 mg/l, CRP showed a sensitivity of 94% and a specificity of 58%, LR+ of 2.24 and LR− of 0.10. For the inflammatory cytokines using cut-off values of 6.9 pg/ml, 18 pg/ml and 2.2 pg/ml, respectively, interleukin 1 beta (IL-1β) showed a sensitivity of 97% and specificity of 59%, LR+ of 2.37 and LR− of 0.05, interleukin 6 (IL-6) showed a sensitivity of 88% and a specificity of 74%, LR+ of 3.38 and LR− of 0.16, and tumour necrosis factor alpha (TNF-α) showed a sensitivity of 88% and a specificity of 80%, LR+ of 4.40 and LR− of 0.15. [EL = III−]

A study conducted in Israel evaluated the ability of procalcitonin level to predict renal involvement assessed by DMSA in 64 children (44 girls and 20 boys) aged 2 weeks to 3 years (mean age 16.7 ± 8.6 months).189 CRP at a cut-off value of 20 mg/l showed a sensitivity of 100% and specificity of 18.5%, and an LR+ of 0.12 (LR− not estimable), while procalcitonin at a cut-off value of 0.5 μg/l showed a sensitivity of 94.1%, a specificity of 89.7%, an LR+ of 9.40 and an LR− of 0.07. [EL = III−]

Evidence statement – localisation of UTI by laboratory tests

Both CRP levels and other laboratory analyses show variable diagnostic performance in localising UTI. The small number of studies and the diverse cut-off points make it difficult to draw any conclusions about the value of these laboratory-based tests for differentiating upper from lower UTI.

Procalcitonin appears to be significantly correlated with a diagnosis of UTI, but more studies are needed to confirm this association.

There is an absence of evidence from which to draw clear conclusions about the clinical and cost-effectiveness of CRP and procalcitonin to differentiate between acute pyelonephritis/upper urinary tract infection and cystitis/lower urinary tract infections.

4.7.3. Localising UTI by imaging tests

Introduction

It may very occasionally be important to attempt to differentiate by imaging infection confined to the lower urinary tract (urinary bladder) from upper urinary tract infection (renal parenchyma – acute pyelonephritis) to guide management. Ultrasound may give some indication of renal parenchymal involvement but DMSA scintigraphy is considered to be the gold standard.

The following section comprises a comprehensive evaluation of the accuracy of the various imaging tests available to assess the urinary tract following UTI in children.

Imaging modalities that have been used as markers to identify involvement of the upper urinary tract include ultrasound (with and without Doppler), DMSA and other forms of static renal scintigraphy, micturating cystourethrogram (MCUG), magnetic resonance imaging (MRI) and intravenous urogram (IVU). The use of power Doppler ultrasound permits a qualitative assessment of regional perfusion and may provide information about renal parenchymal involvement by acute infection.

Previous guideline

The localisation of infection (to the kidneys) by imaging was not a part of the RCP’s 1991 guidelines on imaging in childhood UTI.190 In UK clinical practice, differentiation of acute pyelonephritis/upper urinary tract infection from cystitis/lower urinary tract infection is usually made using a combination of clinical and laboratory features.23

Review findings – ultrasound compared with DMSA scan

A total of 20 studies were included in this review, of which 18 were identified in the HTA143 and the others were from the literature search (Table 4.13).

Table 4.13. Ultrasound compared with DMSA scan for localising UTI.

Table 4.13

Ultrasound compared with DMSA scan for localising UTI.

A systematic review assessed the diagnostic accuracy of ultrasound in 18 studies where renal scintigraphy was the reference standard.143 In 14 of the 18 studies the scintigraphic standard was DMSA. Of the 18 studies, ten did not use an appropriate spectrum of patients and four did not describe criteria used to select patients. Six of the 18 studies provided an adequate description of both the index test and the reference standard.

Of the 18 studies, sensitivity ranged from 9.2% (specificity 100%) to 93.6% (specificity 50%). However, all but three studies reported sensitivities of below 60%. Specificity ranged from 50% (sensitivity 93.6%) to 100% (sensitivity 9.2% to 50%); all but four studies were above 80%.

Likelihood ratios showed considerable heterogeneity (P < 0.0001), with LR+ values ranging from 1.6 (LR− = 0.68) to 55.0 (LR+ = 12.7) and LR− values ranging from 0.10 (LR+ = 2.5) to 0.91 (LR+ = 12.7). The pooled LR+ was 3.1 (95% CI 2.3 to 4.3) and the pooled LR− was 0.62 (95% CI 0.53 to 0.73). ROC plots show considerable heterogeneity between studies, suggesting that conventional ultrasound is a poor test for the localisation of UTI.

A study conducted in Taiwan evaluated the use of ultrasonography in 45 children (31 boys and 14 girls) aged 9 days to 10 years (mean age 1.5 ± 0.2 years, median age 0.3 years) with febrile UTI who fulfilled criteria for acute pyelonephritis/upper urinary tract infection.191 The reported sensitivity and specificity were 49.0% and 88.0%, respectively. [EL = III]

A study conducted in the USA compared DMSA renal ultrasonography and MCUG, using DMSA scintigraphy as the gold standard, in 222 children (47 boys, 175 girls) aged 2 to 228 months (median age 55 months).192 The sensitivity of renal ultrasound to detect renal involvement was 9% and specificity was 100%. [EL = III]

Two additional studies were identified from our search.

A study conducted in Israel investigated power Doppler ultrasonography (PDU) in children with UTI (n = 57 children with a mean age of 22 months).193 Baseline characteristics showed that the mean CRP level was significantly higher in children with acute pyelonephritis/upper urinary tract infection than in children with cystitis/lower urinary tract infection (48.1 ± 39.2 mg/l versus 114.9 ± 48.1 mg/l; P < 0.001). There were no differences in age (P = 0.66), gender (P = 0.47), WBC count (P = 0.06) or ESR (P = 0.46). For detecting acute pyelonephritis/upper urinary tract infection, PDU showed a sensitivity of 87% and a specificity of 92%. [EL = III]

A second Israeli study of 40 infants (78 kidneys evaluated) assessed the role of renal PDU to identify acute pyelonephritis/upper urinary tract infection.194 The PDU showed a sensitivity of 74% and a specificity of 94%. The study went on to compare PDU with DMSA scintigraphy for identifying renal lesions in children who showed acute pyelonephritis/upper urinary tract infection on DMSA. The sensitivity of the PDU decreased to 58%. [EL = Ib]

All included studies are summarised in Table 4.13.

Review findings – MRI/CT scans compared with DMSA scan

Two studies were identified from the HTA143 (Table 4.14).

Table 4.14. MRI/CT scans versus DMSA scan.

Table 4.14

MRI/CT scans versus DMSA scan.

One study assessed the accuracy of gadolinium-enhanced MRI and found sensitivity to be 92% and specificity 44%. A second study assessed the accuracy of CT for diagnosing acute pyelone-phritis/upper urinary tract infection and reported a sensitivity of 56% and a specificity of 100%. Both studies used DMSA as the reference standard, although because there was only one of each of these studies, conclusions cannot be drawn about their usefulness in localising UTI.

Evidence statement – localising UTI by imaging tests

Overall, none of the tests above was considered to add clinical value in the acute phase of the infection.

Conventional ultrasound appears to be a poor diagnostic test for localising infection. Use of power Doppler may increase its diagnostic values.

No conclusions can be drawn about the effectiveness or cost-effectiveness of MRI, CT or IVU in localising UTI owing to there only being a small number of poor-quality studies.

DMSA is considered the gold standard for the localisation of infection to the renal parenchyma.

4.8. GDG translation and recommendations

GDG translation

Predisposing factors

A number of factors were identified that increase or decrease the likelihood of an infant or child having a UTI. These risk factors should be evaluated during the history and examination so that they can help to inform the diagnostic process.

The reviewed evidence shows that there is a lower incidence of UTI in boys who are circumcised. The evidence for risks and benefits of circumcision has not been evaluated by this guideline.

The review showed that breastfeeding is associated with a reduced risk of UTI in infants. In addition to recording the feeding method and duration of breastfeeding in the history, mothers could be made aware of the protective effect of breastfeeding when they are making plans about how to feed their baby. Any future NICE guideline on breastfeeding should consider including this as a recommendation.

First UTI is most common in infancy and affects boys most often in the first 3 months of life while in girls the peak incidence is after 6 months. Infants are often systemically unwell and have acute pyelonephritis/upper urinary tract infection while older children more often have lower urinary tract infection and typical symptoms of cystitis. The incidence of first UTI falls with age in both sexes but UTI is less common in boys than in girls after the first 6 months. Recurrent infections in boys are relatively uncommon whereas they are very common in girls. These represent a high proportion of the infections seen in primary care. UTI seems to be less common in children with Afro-Caribbean origins in the USA.

Symptoms and signs

The majority of the included studies are of infants and children treated in secondary care centres and do not represent the majority of infants and children who present with a UTI in primary care. Furthermore, some of the studies from primary care only looked at children with symptoms localised to the urinary tract and therefore do not provide comprehensive data because of the lack of localising signs in infants.

Reported symptoms/signs in infants and children with UTI include fever, irritability, lethargy, vomiting, anorexia, diarrhoea, enuresis, dysuria, frequency, abdominal pain, loin tenderness, offensive urine, haematuria and failure to thrive. UTI is more likely to be the cause of fever when there is no obvious source of infection and there is no alternative diagnosis.

Presenting symptoms combined with findings on examination, urine testing and knowledge of the risk factors should all be taken into account when considering a diagnosis of UTI.

Some features in history and examination are important for assessing the probability of UTI as well as the risk of serious underlying abnormality and will aid the diagnostic process as well as informing the most appropriate imaging strategy.

Urine collection

In children able to cooperate, clean catch urine collection provides a good-quality sample. In children unable to cooperate, urine collection is more difficult and time-consuming. It may be impossible to obtain a sample and there is a high possibility of the sample being contaminated. In this situation, urine collection pads are a preferable alternative to urine collection bags. They are less costly and are not unpleasant for the child.

Although diarrhoea can be associated with UTI it is not usually the main symptom. It is often difficult to get a good-quality urine sample for testing for UTI in the presence of diarrhoea. There should be a high threshold for urine collection in infants and children when the cause is most likely to be due to acute viral or bacterial gastroenteritis.1

Urine testing

Which test to choose

When all diagnostic methods are plotted in ROC space it is clear that there is a wide variation in reported methods for all performance and all have strengths and potential drawbacks (Figure 4.3).

There are two groups of urine tests, those providing immediate information on which to base therapeutic intervention, i.e. dipstick testing and microscopy, and culture, which provides information on sensitivity and identification of the pathogen but whose results are not available for at least 24 hours.

Age cut-off

The GDG considered that age was an important factor to be taken into consideration when choosing the most appropriate tests to diagnose UTI. Tests for UTI were considered separately for infants and children younger than 3 years and children 3 years or older since the majority of children 3 years or older are able to provide a good-quality urine sample on request. It was considered that reasons for poorer reliability of dipstick testing for infants and children younger than 2 years include urine sample contamination, differences between urine collection methods, small capacity of the bladder and frequent emptying.

No study examined the age as a continuous variable to enable evaluation of optimum age cut-off. Two studies looked at urine testing stratified by age. One study presented the results for children younger than 1 year and 1 year or older separately, and the other did so for those younger than 2 years and those 2 years or older. The rationale for each decision in the two age groups is discussed below in more detail.

Rapid testing (decision making on acute management)

Infants and children younger than 3 years

Table 4.8 shows that, in infants and children younger than 2 years, the best performing rapid test to support a clinical diagnosis of UTI is microscopy using a threshold of 10 WBC/hpf. However, in terms of ruling out a UTI, no test performs particularly well, with only marginal differences between dipstick and microscopy.

Therefore the GDG suggests using urgent microscopy to make a decision on acute management, although, considering the marginal difference in ruling out UTI and practicality in primary care settings, dipstick urine testing may be used for infants and children younger than 3 years with non-specific symptoms and low or moderate risk of serious illness where urgent microscopy is not possible. The GDG considers that infants and children younger than 3 years with high risk of serious illness should be sent to a paediatric specialist where urgent microscopy should be available. The GDG also considers that infants and children younger than 3 years with non-specific symptoms and low risk of serious illness can be observed without antibiotic treatment until results of standard (not urgent) microscopy are available. This decision takes into consideration the risks and benefits of blind treatment and the very low risk of having acute pyelonephritis/upper urinary tract infection in this group. However, urine for microscopy and culture should be sent from all infants and children younger than 3 years to confirm the diagnosis, and appropriate treatment can be started when the results are available after 24–48 hours if the diagnosis is in doubt.

Children 3 years or older

It is clear from Table 4.8 that dipstick testing with both leucocyte esterase and nitrite is the best performing rapid test to support a clinical diagnosis of UTI in children 2 years or older. However, the position for ruling out UTI is more complex. In absolute diagnostic terms, microscopy using a cut-off of 5 WBC/hpf is the most reliable test to rule out UTI. However, there was no statistical evidence of difference between microscopy and urine dipstick testing to rule out UTI, as the confidence interval in the available study for LR− for microscopy is wide. Microscopy is more complicated to perform, needs a higher level of training and, with regulatory demands to ensure quality of near-patient testing, would require the development of local or national EQA and IQA schemes to ensure that performance is maintained. This is achievable in the setting of hospital laboratories but may be more difficult or impossible in primary care. The following factors were taken into account when considering the best test to use in this age group:

  1. the difficulty of organising urgent laboratory tests in primary care
  2. the training required to perform microscopy in the surgery
  3. the need for quality assurance
  4. the difference in LR− values between microscopy and urine dipstick testing in children 2 years or older was not statistically significant
  5. the available studies showed that dipstick urine testing has a more favourable LR+ than microscopy in children 2 years or older
  6. the cost of performing microscopy is significantly higher than urine dipstick testing.

Consequently, the GDG considers that urine dipstick testing is the most appropriate test to make a decision on diagnosis and acute management in children 3 years or older.

The role of culture (confirmation of diagnosis)

Both dipstick urine testing and microscopy may give false negative and false positive results which will give rise to both under- and over-diagnosis of UTI in infants and children. Therefore, confirmation of UTI using urine culture is recommended as this is regarded as the gold standard. Urine culture should be carried out in selected higher risk cases where there is evidence of a risk of acute pyelonephritis/upper urinary tract infection, recurrent UTI or atypical features, since in these cases there is the greatest risk of renal damage or underlying abnormality. The risk of acute pyelonephritis/upper urinary tract infection is greatest in the youngest infants and children and therefore culture is recommended in all infants and children under 3 years.

Implication for current practice and cost-effectiveness

Current practice recommends routine use of microscopy and culture of urine for the diagnosis of UTI in infants and children. In positive cases, confirmation of UTI is recommended using a second sample for culture collected by clean catch or SPA. Dipsticks were considered acceptable for ruling in but not for ruling out UTIs.21

The national audit showed that 71% of infants and children younger than 2 years with a fever or other symptoms of UTI admitted to hospital had urine culture and microscopy and 100% of infants and children diagnosed with UTI. Fifty-seven percent had a second sample taken for culture and microscopy. Dipsticks were used in some hospitals but results were poorly documented.23

Based on clinical evidence, the GDG proposes that the first line of investigation should be to use dipstick urine testing for children 3 years or older and microscopy and culture in infants and children younger than 3 years. Infants and children younger than 3 years, infants and children with signs and symptoms suggesting acute pyelonephritis/upper urinary tract infection and infants and children with recurrent UTI should have a urine sample sent for culture. In the younger age group dipsticks are only recommended for use as a substitute for microscopy when transport of a sample to the laboratory is impossible and the child has a low to intermediate risk of serious illness.

In addition to being more clinically effective and evidence based, the new proposals in this guideline for children 3 years or older are more cost-effective since, for older children, the shift is from microscopy to dipstick testing, which is both more effective and better use of NHS resources.

Laboratory tests for localisation of UTI

CRP and procalcitonin can be used to help with the diagnosis or exclusion of acute pyelonephritis/upper urinary tract infection in an appropriate clinical setting. In particular, a CRP of < 20 mg/l reduces the risk that an infant or child has a serious bacterial infection and may be useful in ruling out acute pyelonephritis/upper urinary tract infection in infants and children with fever and pyuria who often have a viral infection. In the context of UTI, a CRP of < 20 mg/l makes the diagnosis of acute pyelonephritis/upper urinary tract infection unlikely.

Imaging tests for localisation of UTI

In the majority of infants and children with UTI who respond promptly to treatment, differentiation of acute pyelonephritis/upper urinary tract infection from cystitis/lower urinary tract infection by imaging is unnecessary and would represent a poor use of NHS resources. DMSA scanning has been regarded as the most sensitive imaging test for the diagnosis of acute pyelonephritis/upper urinary tract infection.

If localisation of UTI by imaging tests is required, it would be good practice to utilise ultrasound with power Doppler prior to DMSA scan as a positive finding may obviate the need for DMSA scan.

Recommendations

Symptoms and signs

Infants and children presenting with unexplained fever of 38 °C or higher should have a urine sample tested after 24 hours at the latest.

Infants and children with an alternative site of infection should not have a urine sample tested. When infants and children with an alternative site of infection remain unwell, urine testing should be considered after 24 hours at the latest.

Infants and children with symptoms and signs suggestive of UTI should have a urine sample tested for infection. Table 4.15 is a guide to the symptoms and signs that infants and children present with.

Assessment of risk of serious illness

The illness level in infants and children should be assessed in accordance with recommendations in ‘Feverish illness in children’ (NICE clinical guideline 47).

Urine collection

A clean catch urine sample is the recommended method for urine collection. If a clean catch urine sample is unobtainable:

  • Other non-invasive methods such as urine collection pads should be used. It is important to follow the manufacturer’s instructions when using urine collection pads. Cotton wool balls, gauze and sanitary towels should not be used to collect urine in infants and children.
  • When it is not possible or practical to collect urine by non invasive methods, catheter samples or suprapubic aspiration (SPA) should be used.
  • Before SPA is attempted, ultrasound guidance should be used to demonstrate the presence of urine in the bladder.

In an infant or child with a high risk of serious illness it is highly preferable that a urine sample is obtained; however, treatment should not be delayed if a urine sample is unobtainable.

Urine preservation

If urine is to be cultured but cannot be cultured within 4 hours of collection, the sample should be refrigerated or preserved with boric acid immediately.

The manufacturer’s instructions should be followed when boric acid is used to ensure the correct specimen volume to avoid potential toxicity against bacteria in the specimen.

Urine testing

The urine-testing strategies shown in Tables 4.164.19 are recommended.*

As with all diagnostic tests there will be a small number of false negative results; therefore clinicians should use clinical criteria for their decisions in cases where urine testing does not support the findings.

Indication for culture

Urine samples should be sent for culture:

  • in infants and children who have a diagnosis of acute pyelonephritis/upper urinary tract infection
  • in infants and children with a high to intermediate risk of serious illness
  • in infants and children younger than 3 years
  • in infants and children with a single positive result for leucocyte esterase or nitrite
  • in infants and children with recurrent UTI
  • in infants and children with an infection that does not respond to treatment within 24–48 hours, if no sample has already been sent
  • when clinical symptoms and dipstick tests do not correlate.

History and examination on confirmed UTI

The following risk factors for UTI and serious underlying pathology should be recorded:

  • poor urine flow
  • history suggesting previous UTI or confirmed previous UTI
  • recurrent fever of uncertain origin
  • antenatally diagnosed renal abnormality
  • family history of vesicoureteric reflux (VUR) or renal disease
  • constipation
  • dysfunctional voiding
  • enlarged bladder
  • abdominal mass
  • evidence of spinal lesion
  • poor growth
  • high blood pressure.

Clinical differentiation between acute pyelonephritis/upper urinary tract infection and cystitis/lower urinary tract infection

Infants and children who have bacteriuria and fever of 38 °C or higher should be considered to have acute pyelonephritis/upper urinary tract infection. Infants and children presenting with fever lower than 38 °C with loin pain/tenderness and bacteriuria should be considered to have acute pyelonephritis/upper urinary tract infection. All other infants and children who have bacteriuria but no systemic symptoms or signs should be considered to have cystitis/lower urinary tract infection.

Laboratory tests for localising UTI

C-reactive protein alone should not be used to differentiate acute pyelonephritis/upper urinary tract infection from cystitis/lower urinary tract infection in infants and children.

Imaging tests for localising UTI

The routine use of imaging in the localisation of a UTI is not recommended.

In the rare instances when it is clinically important to confirm or exclude acute pyelonephritis/upper urinary tract infection, power Doppler ultrasound is recommended. When this is not available or the diagnosis still cannot be confirmed, a dimercaptosuccinic acid (DMSA) scintigraphy scan is recommended.

*

Assess the risk of serious illness in line with ‘Feverish illness in children’ (NICE clinical guideline 47) to ensure appropriate urine tests and interpretation, both of which depend on the child’s age and risk of serious illness.

Research recommendations

More studies with adequate sample sizes are needed to evaluate the effectiveness of breast-feeding, nappies and hygiene in preventing childhood UTI.

Combined population-based studies in primary and secondary care, with larger sample sizes are needed to evaluate the association between symptoms and signs and UTI.

Further investigation of leucocyte esterase and nitrite dipstick tests alone and in combination, stratified by age and method of urine collection, is required to determine their accuracy in diagnosing UTI.

Further research is needed to evaluate the effectiveness of biochemical tests for low urinary glucose for diagnosing UTI in infants and children.

Further research is needed to evaluate the effectiveness of procalcitonin and other inflammatory markers in localising UTI.

Copyright © 2007, 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.

Bookshelf ID: NBK50608

Views

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...