This publication is provided for historical reference only and the information may be out of date.
7.1. Initial assessment and criteria for admission to intensive or special care
Description of the evidence
A number of morbidities present in babies born to women with diabetes (including pre-existing type 1 and type 2 diabetes and gestational diabetes). These include fetal macrosomia, infant respiratory distress syndrome, cardiomyopathy, hypoglycaemia, hypocalcaemia, hypomagnesaemia, polycythaemia and hyperviscosity. Hypoxia causes polycythaemia and hyperviscosity.352 Thrombosis is a rare, but serious, complication necessitating admission to a NICU. Risk factors include maternal diabetes, maternal factors resulting in IUGR, polycythaemia and the use of intravascular catheters in preterm babies.353 Information on incidence and presentation of thrombosis is very limited and has been gathered mainly from case reports and case series.
While some of the morbidities listed above correct themselves within a period of few hours to a few weeks (e.g. transient tachypnoea normalises within 3 days of birth),354 it is still important that treatment is provided promptly for those requiring it (e.g. hyaline membrane disease, for which babies may require surfactant, or respiratory and metabolic support).354
Incidence of neonatal morbidities
A cohort study355 assessed the effect of rigorous management of type 1 diabetes during pregnancy on perinatal outcome by comparing 78 pregnant women with type 1 diabetes managed prospectively with 78 matched controls who did not have diabetes. The women with diabetes used insulin by infusion pump or split-dose therapy, with the goal of normalising fasting blood sugars and HbA1c. Women with type 1 diabetes had higher rates of preterm birth (31% versus 10%, P = 0.003), pre-eclampsia (15% versus 5%, P = 0.035) and caesarean section (55% versus 27%, P = 0.002). Complications of babies born to women with diabetes included LGA (41% versus 16%, P = 0.0002), hypoglycaemia (14% versus 1%, P = 0.0025), hyperbilirubinaemia (46% versus 23%, P = 0.0002) and respiratory distress (12% versus 1%, P = 0.008). Apgar scores and mortality rates were similar for the two groups. Congenital malformations occurred in 7.7% of babies of women with diabetes and 1.3% of controls (P = 0.05).
The incidence of respiratory distress syndrome and mortality was assessed in a cohort study356 involving 23 babies selected from a total of 30 babies born to women with diabetes who developed hypoglycaemia after birth. These babies were divided into the following three groups: 12 babies treated with intravenous glucose; seven babies treated with long-acting epinephrine plus intravenous glucose; and four babies treated with long-acting epinephrine only. There were no significant differences in incidence of respiratory distress syndrome or mortality rates between the three groups. [EL = 2+]
A case–control study357 investigated factors that contribute to neonatal hypoglycaemia in babies of women with diabetes. Timing of blood glucose levels, symptoms of hypoglycaemia and interventions provided were assessed. None of the 66 babies investigated developed symptomatic hypoglycaemia or required intravenous glucose. Nearly all the low blood glucose determinations (less than 1.7 mmol/litre) occurred in the first 90 minutes of life, which is the period of greatest risk of low blood glucose occurring in babies born to women with diabetes. [EL = 2+]
A cross-sectional study322 assessed routes of delivery and pregnancy outcomes in 10 369 births in the USA. Diabetes was associated with increased caesarean section rates, resuscitation of babies with positive pressure ventilation and low Apgar scores (less than 3) at 1 minute and 5 minutes. [EL = 3]
Another cross-sectional study358 assessed types and frequencies of complications occurring in babies of women with diabetes. Immediately after birth, babies of women with diabetes were admitted to a NICU and detailed maternal history and physical examination were performed to detect any congenital anomalies. The caesarean section rate was high, as was the rate of birth injuries among those who had vaginal birth. The number of babies with asphyxia, congenital anomalies, hypoglycaemia, hypocalcaemia and hyperbilirubinaemia was high. The overall mortality rate of 7.5% was high. The study recommended that women with diabetes should be offered regular antenatal care to maintain good glycaemic control during pregnancy, the birth should be attended by experienced paediatricians to minimise complications and when there is clinical evidence of macrosomia caesarean section should be offered to reduce birth injuries. [EL = 2−]
Neonatal assessment
No clinical studies were identified that addressed the assessments that babies of women with diabetes should undergo. The following evidence is drawn from two narrative non-systematic reviews.354,359 [EL = 4]
Fetal macrosomia
The investigation of birth trauma in macrosomic babies of women with gestational diabetes has been described as necessary. Manifestations included fractures of the clavicle and/or humerus. Brachial plexus (Erb palsy), phrenic nerve or cerebral injuries were also reported. In cases where a fracture is suspected, a chest radiograph may be used to confirm the presence of fractures. The startle reflex could also be used to confirm the presence of fractures, as the baby may show an asymmetric reflex or limited use of the arm on the affected side.359 As most fractures of this nature heal without treatment, admission to a NICU is probably not necessary. [EL = 4]
Respiratory distress
Use of chest radiographs has been suggested for babies displaying signs of respiratory distress syndrome. An enlarged heart or diffuse, fine granular densities are consistent with respiratory distress syndrome. Together with arterial blood gas results, the need for respiratory support and replacement surfactant therapy could be determined. Where left ventricular outflow obstruction was suspected, an echocardiogram was prescribed in order to prevent congestive heart failure.359 [EL = 4]
Polycythaemia
For babies displaying clinical signs of polycythaemia (respiratory distress syndrome, apnoea, hepatomegaly, jitteriness, irritability, seizures, feeding intolerance, hypoglycaemia, decreased urine output), venous haematocrit measurements are indicated.359[EL = 4]
Hypocalcaemia
Signs and symptoms of hypocalcaemia include coarse tremors, twitching, irritability and seizures. These should indicate monitoring of serum calcium levels. If levels are 1.75 mmol/litre or more, calcium replacement should be started.359[EL = 4]
Hyperbilirubinaemia
Where there has been lethargy, delayed feeding, polycythaemia and birth trauma, bilirubin levels should be monitored.359 A review addressing medical concerns in the neonatal period described jaundice in the first 24 hours of life as pathological and requiring immediate evaluation and therapy as bilirubin in high concentrations is considered a cellular toxin.354 [EL = 3]
Severe asphyxia
If severe asphyxia has occurred at birth, the presence of hypotonia, seizures, poor perfusion and/or absence of respirations in the baby should be assessed to determine whether endotracheal intubation and respiratory support is needed.359 [EL = 4]
Hypoglycaemia
The prevention and treatment of neonatal hypoglycaemia is addressed separately in Section 7.2.
Criteria for admission to intensive/special care
Amniotic fluid erythropoietin
A case–control study investigated whether chronic fetal hypoxia, as indicated by amniotic fluid erythropoietin (EPO) levels, was associated with neonatal complications in pregnancies complicated by type 1 diabetes.360 An amniotic fluid sampled for EPO measurement was taken from 157 women with type 1 diabetes who gave birth by caesarean section before the onset of labour (one vaginal birth), either within 2 days of birth or at birth. EPO measurements were compared with those from 19 healthy, non-smoking women delivered by elective caesarean section with an uneventful singleton pregnancy producing a healthy newborn baby. The median amniotic fluid EPO level was significantly higher in the women with diabetes (14.0 mU/ml, range 2.0 to 1975, n = 155) than in control pregnancies (6.3 mU/ml, range 1.7 to 13.7, n = 19; P < 0.0001). Amniotic fluid EPO levels above 63.0 mU/ml were considered to indicate fetal hypoxia and these elevated values were observed in 14.1% of the women with diabetes who were divided into three groups: low EPO less than 13.8 mU/ml, intermediate EPO 13.8–63.0 mU/ml and high EPO more than 63.0 mU/ml. Newborn babies in the high EPO group were significantly more likely to be macrosomic (P = 0.0005) and acidotic (P < 0.0001) and had significantly lower pO2 levels than those in the intermediate and low EPO groups (P < 0.0001). Neonatal hypoglycaemia (blood glucose less than 2.0 mmol/litre more than 6 hours after birth; P < 0.0001), admission to neonatal intensive care (P = 0.03), cardiomyopathy (P < 0.0001) and hyperbilirubinaemia (P = 0.002) occurred significantly more often in the high EPO group than in the low EPO group. After adjusting for the effects of maternal age, maternal BMI, gestational age at birth, birthweight z-score, last amniotic fluid EPO level and last maternal HbA1c level, amniotic fluid EPO was the only variable to remain independently associated with low umbilical artery pH (P < 0.0001) and neonatal hypoglycaemia (P = 0.002). Low pO2 at birth was associated with amniotic fluid EPO (P < 0.0001) and birthweight z-score (P = 0.004). [EL = 2+]
Neonatal hypoglycaemia
A prospective cohort study investigated the frequency and risk factors for neonatal hypoglycaemia and long-term outcomes of promptly treated neonatal hypoglycaemia.361 Of the 4032 babies born in the study hospital, 1023 were admitted to NICU. Ninety-four (9.18%) were evaluated as having hypoglycaemia. Evaluations were performed if symptoms such as hypothermia, apnoea, lethargy, poor feeding or seizures were observed, or if risk factors such as SGA, LGA, preterm birth, sepsis or the mother having diabetes were present. The cohort was followed for 24 months, during which time they were assessed neurologically and developmentally using the Bayley motor and developmental scales. The study found that 51.1% of babies were preterm (37 weeks or less), 34.1% of babies were born to women with pre-existing maternal diabetes or gestational diabetes and 12.8% were SGA. SGA babies required the longest duration of intravenous dextrose infusion (5.16 days compared with 3.74 days for AGA babies). In 26.6% of the babies no known risk factors for hypoglycaemia were observed. Of the 48 babies undergoing Bayley’s psychometric evaluations, two showed a motor deficit at 6 months and one showed a language deficit at 24 months. [EL = 2+]
Gestational age, respiratory distress syndrome and higher birthweight
A retrospective cohort study conducted in the USA over a 3 year period acquired data for 530 babies born to 332 women with gestational diabetes and 177 women with type 1 diabetes. The study found 47% (247) of babies were admitted to a NICU.362 Seventy-six babies had a gestational age of 33 weeks or less, 22 babies had congenital malformations, ten were described as having miscellaneous conditions (apnoea, cardiac arrhythmias, poor feeding or neonatal depression), 103 babies with a gestational age of 34 weeks or more had respiratory distress syndrome and 32 babies had hypoglycaemia as the only diagnosis. For the 182 babies (34%) presenting with respiratory distress syndrome of varying severity, the highest rates (56%) were seen in women with type 1 diabetes which had been diagnosed before the age of 10 years, had had diabetes for 20 years or longer or had complications of diabetes, decreasing to 25% in babies of women with gestational diabetes requiring no insulin. Similarly, babies of women with gestational diabetes requiring no insulin had the lowest representation of LGA babies (25%), while those born to women with type 1 diabetes which had been diagnosed before the age of 10 years and had a duration of 20 years had the highest rate (62%). The frequency of SGA babies was equal among the classes. Seventy-four (14%) of the 530 babies had macrosomia and 57% (42/74) of this group were delivered by caesarean section. Among those delivered by caesarean section, there were 21 cases of hypoglycaemia, three of polycythaemia, one of hypocalcaemia and 12 of hyperbilirubinaemia. Thirty percent of macrosomic babies had respiratory distress syndrome. [EL = 2+]
Blood glucose levels were recorded for 514 babies. One or more hypoglycaemic episodes occurred in 27% (137) of these babies. While 90% of the babies responded rapidly to treatment, 10% had two or more episodes lasting several hours. Neonatal hypoglycaemia was similar among babies born to women with gestational diabetes requiring no insulin (23%), those with gestational diabetes requiring insulin (24%) and those with type 1 diabetes with age of onset 20 years or more, or a duration less of than 10 years with no vascular lesions (25%). The prevalence of neonatal hypoglycaemia was lower in these babies (P < 0.05) than the babies of women with type 1 diabetes with age of onset 10–19 years or duration 10–19 years with no vascular lesions (35%), the babies of women with type 1 diabetes which had been diagnosed before the age of 10 years, and the babies of women who had had diabetes for 20 years or longer or had complications of diabetes (38%). Thirty of the 137 babies with hypoglycaemia were born before 34 weeks of gestation, 55 were LGA, 50 were AGA and two were SGA. Among the 74 babies who were macrosomic, 21 were also hypoglycaemic. Of the 244 babies (46% of total group) assigned to ‘well baby nurseries’ for routine care and enteral feeding, 32 had hypoglycaemia.
The study found 5% (13) of the 276 babies who had their haematocrit assessed were polycythaemic (haematocrit 0.65 or more). Of the 530 babies, 25% (125) were treated for hyperbilirubinaemia and, of these, 61 were delivered at 33 weeks of gestation or less. The rate of treatment for non-diabetic, full-term babies delivered during the same 2 year period was 5%.
Of the 244 babies admitted to well baby nurseries for routine care, 18% (43) were then transferred to the NICU (19 with respiratory distress syndrome as the main reason for transfer, 16 for treatment of hypoglycaemia, seven for respiratory distress syndrome plus hypoglycaemia and one for poor feeding). Advanced maternal diabetes and lower gestational age were shown by logistic regression to be the strongest predictors of subsequent NICU care. Logistic regression analysis also showed that after controlling for gestational age and type of diabetes, breastfed babies were more likely to succeed with routine care and enteral feeding. [EL = 2+]
Myocardial hypertrophy and respiratory distress syndrome
A cross-sectional study looked at the association between poorly controlled maternal diabetes and myocardial hypertrophy.363 Twelve neonates were admitted to NICU with respiratory distress and cardiomegaly. Ten babies were macrosomic and had myocardial hypertrophy as determined by echocardiograph. Two of these babies died from cardiorespiratory failure within 48 hours of birth. Two babies were AGA and had cardiomegaly resulting from ventricular dilation in association with hypoglycaemia and acidaemia. Of the surviving babies, 80% (8/10) had clinical findings suggesting respiratory distress syndrome. The presence of hyaline membranes at autopsy of the other two babies lends support to an association between respiratory distress syndrome and myocardial hypertrophy. [EL = 3]
Gestational age and mode of birth
A clinical audit was conducted at the National Women’s Hospital in New Zealand364 which serves a multi-ethnic population with a high background prevalence of type 2 diabetes. In total 136 babies of women with diabetes were admitted to NICU. Tweny-nine percent (112/382) of the babies of women with gestational diabetes were admitted and 40% (24/60) of the babies of women with type 2 diabetes were admitted. Fifty-six percent (58/104) of the gestational diabetes was reclassified as normal, IGT or type 2 diabetes after postpartum 75 g OGTTs. Infant outcomes according to maternal antenatal and postpartum diagnoses were recorded. The study found 46% (63/136) babies were delivered preterm (before 37 weeks). Women with gestational diabetes that was reclassified postpartum as IGT or type 2 diabetes accounted for the highest rates of preterm babies (86% [12/14] and 63% [12/19], respectively). The rate of emergency lower segment caesarean section in women with gestational diabetes or type 2 diabetes was 25%. The rate of emergency lower segment caesarean section of women with gestational diabetes or type 2 diabetes whose babies were admitted to NICU was 38% (52/136). When a similar comparison was made for preterm birth the rates were 19% compared with 46% (63/136). The most common indication for admittance to NICU was hypoglycaemia, which was documented in 51% of the babies. This was followed by respiratory distress in 40% of babies. Rates of respiratory distress in the preterm babies and term babies were not significantly different (39% [26/67] versus 43% [31/70], P = 0.34). A third of women with type 2 diabetes antenatally or postpartum had babies weighing more than 4000 g. These birthweights were significantly higher than for the IGT group (P < 0.05) and significantly more common than in the IGT or normal group (P < 0.05). [EL = 2+]
Current practice
The CEMACH enquiry covered neonatal care of term babies born to women with pre-existing type 1 or type 2 diabetes.33 In the 112 babies selected for the neonatal enquiry that had medical records available, 70 were admitted to a postnatal ward, transitional care unit, stayed on the labour ward or in a maternal dependency unit and 42 were admitted to a NICU for special care. The three main indications for admission to a NICU were a hospital policy of routine admission of healthy babies of women with diabetes 29% (12/42), asymptomatic hypoglycaemia in a healthy baby 26% (11/42) and a clinical need for admission such as poor feeding or respiratory problems 43% (18/42). The enquiry panels assessed that 57% (24/42) of the admissions were unavoidable and that subsequent care of 63% (15) of the babies was compromised, especially in the area of feeding (50%, 12/24 babies). There was evidence of a clear written care plan for 73% (51/70) of babies who remained with their mothers and 57% (24/42) of babies admitted to a NICU. The care plan was not fully followed for 35% (18/51) of babies remaining with their mothers; aspects of the care plan that were not followed included blood glucose management, feeding and temperature. The enquiry also discussed the importance of early skin-to-skin contact between babies and their mothers (see Section 7.2) and recommended that all units where women with diabetes give birth should have written policy for management of the baby and that the policy should assume that babies will remain with their mothers in the absence of complications. [EL = 3–4]
A 2002 CEMACH audit of units expected to provide maternity care for women with diabetes in England, Wales and Northern Ireland reported that 30% (64/213) had a policy of routinely admitting babies of women with diabetes to the neonatal or special care unit.32 [EL = 3]
Existing guidance
The NSF for diabetes20 recommends that ‘Neonatal intensive care is only indicated for babies who display persistent hypoglycaemia after 3 hours of age.’
Evidence statement
Five observational studies have reported on the incidence of neonatal morbidity in babies born to women with diabetes. Complications reported in these studies included asphyxia, birth trauma (e.g. shoulder dystocia), congenital malformations, hyperbilirubinaemia, hypoglycaemia, hypocalcaemia, LGA, respiratory distress syndrome and associated mortality.
No clinical studies were identified in relation to neonatal assessment that babies of women with diabetes should undergo, but two narrative non-systematic reviews described the clinical signs of the most frequently occurring neonatal complications in babies of women with diabetes.
A further four observational studies and a clinical audit investigated neonatal complications (including fetal hypoxia, hypocalcaemia, hypoglycaemia, macrosomia, myocardial hypertrophy (hypertrophic cardiomyopathy), polycythaemia and respiratory distress syndrome) and indications for admission to a NICU for babies of women with pre-existing type 1 or type 2 diabetes and gestational diabetes. None of the studies reported incidence of hypoxic ischaemic encephalopathy or hypomagnesaemia, although babies of women with diabetes are believed to be at increased risk of these complications.
One of the observational studies reported that persistent or recurrent hypoglycaemia in the neonatal stage can lead to neurodevelopmental deficits later in life. The authors of the study recommended that high-risk babies be screened at regular intervals in the first 48 hours of life if not being fed, or before the first three or four feedings, and in the presence of clinical signs of hypoglycaemia.
Other observational studies showed that prematurity and birth by emergency caesarean section were predictors for NICU admission in women with type 2 diabetes and those with gestational diabetes. Several of the studies suggested that babies of women with diabetes should be closely monitored and admitted to intensive care only in unavoidable circumstances where there are clinical signs of hypoglycaemia and/or respiratory distress, thus avoiding unnecessary separation of mothers and babies.
The clinical audit reported that the most frequent indications for admission to NICU were hypoglycaemia and respiratory distress syndrome, and one of the observational studies reported that the prevalence of these complications was higher with increasing duration of diabetes.
Cost-effectiveness
The effectiveness of criteria for admission to neonatal intensive/special care for babies of women with diabetes was identified by the GDG as a priority for health economic analysis. The NSF for diabetes20 recommends that admission to a NICU should be made only for babies with persistent hypoglycaemia. However, the CEMACH audit reported that 30% of units still routinely admit babies of mothers with diabetes to the neonatal or special care unit and that the most frequent reasons for admission to a NICU were routine policy and asymptomatic hypoglycaemia. Thus no health economic modelling is needed to demonstrate that reinforcing the NSF recommendation, to keep babies with their mothers except when there is a clinical reason to separate them, represents a cost saving to the NHS.
From evidence to recommendations
Evidence shows that birth trauma, congenital malformations (cardiac and central nervous system), hyperbilirubinaemia, hypocalcaemia, hypoglycaemia, hypomagnesaemia, myocardial hypertrophy (hypertrophic cardiomyopathy), neonatal encephalopathy, polycythaemia and hyperviscosity, and respiratory distress (several of which are potentially life-threatening) are more prevalent in babies of women with pre-existing diabetes and gestational diabetes. Healthcare professionals assessing such babies should, therefore, be competent to recognise and manage these conditions and women with diabetes (including gestational diabetes) should be advised to give birth in hospitals where advanced neonatal resuscitation skills are available 24 hours a day.
The GDG’s view is that blood glucose testing should be carried out routinely (at 2–4 hours after birth) for babies of women with diabetes because of the risk of complications arising from asymptomatic hypoglycaemia (see Section 7.2). However, blood tests for polycythaemia, hyperbilirubinaemia, hypocalcaemia and hypomagnesaemia, and investigations for congenital heart malformations and cardiomyopathy should be reserved for babies with clinical signs of these complications, thus avoiding unnecessary investigations, which will represent cost savings to the NHS and should provide reassurance for parents.
Babies of women with diabetes should be kept with their mothers unless there is a clinical complication or abnormal clinical signs that warrant admission for intensive or special care, in accordance with the recommendations contained in the NSF for diabetes,20 thus bringing cost savings to the NHS and maximising the opportunity for early skin-to-skin contact between babies and their mothers and initiation of breastfeeding (see Section 7.2).
Some babies with clinical signs of the conditions listed above may be cared for in a transitional care unit, depending on local guidelines, facilities and care pathways. Where such facilities are unavailable, babies with these conditions should be admitted to a neonatal unit.
Neonatal metabolic adaptation in babies of women with diabetes is generally completed by 72 hours of age. Transfer to community care is not recommended before 24 hours and not before healthcare professionals are satisfied that the baby is maintaining blood glucose levels and has developed good feeding skills because of the risk of recurrent hypoglycaemia in the early neonatal period. Early community midwifery support for these babies should be more intense than average.
There were no research recommendations relating to the initial assessment of babies and criteria for admission to intensive/special care.
7.2. Prevention and assessment of neonatal hypoglycaemia
Description of the evidence
The working definition for neonatal hypoglycaemia is blood glucose less than 2.6 mmol/litre.193,365,366 This threshold is not used to diagnose the condition, but rather to indicate the level at which intervention (additional feeding and, if this does not reverse the hypoglycaemia, intravenous dextrose) should be considered. It is based on a study that found adverse neurodevelopmental outcomes to be associated with repeated values below this level.367 The study involved 661 preterm babies and used multiple regression to show that reduced developmental scores were associated independently with plasma glucose concentration less than 2.6 mmol/litre. [EL = 2+]
A consensus statement368 discussed the definition of neonatal hypoglycaemia. The statement considered term babies, babies with abnormal clinical signs, babies with risk factors for compromised metabolic adaptation, preterm babies and babies receiving parenteral nutrition. Close surveillance should be maintained in babies with risk factors for compromised metabolic adaptation if the plasma glucose concentration is less than 2.0 mmol/litre; at very low concentrations (1.1–1.4 mmol/litre) an intravenous glucose infusion is indicated to raise the glucose level above 2.5 mmol/litre. [EL = 4]
The characteristics of neonatal hypoglycaemia in babies of women with diabetes are very early onset (first hour after birth), generally asymptomatic, non-recurrent and good response to intravenous dextrose.368
Early feeding
Two studies were found that investigated the effect of timing of first feed on blood glucose levels. The studies were undertaken in the 1960s when delaying the initial feed was common.
The first study compared 27 preterm babies allocated to an ‘early fed’ group (fed with formula from 6 hours of age) with 41 babies fasted for 72 hours.369 At 72 hours 24/41 babies in the fasted group had blood glucose levels below 1.4 mmol/litre. In the early fed group no babies had blood glucose values below this level. Statistical significance was not reported. [EL = 2+]
The second study compared 118 preterm babies fed at 3 hours with undiluted breast milk with 121 fed at a later stage, usually at 12 hours. There were no cases of symptomatic hypoglycaemia in the early fed group compared with four cases in the later fed group. Blood sugar estimation was introduced in phase three of the trial. The lowest level was less than 1.1 mmol/litre in 5/44 in the ‘immediate fed’ group compared with 10/54 in the ‘later fed’ group.370 Statistical significance was not reported. [EL = 2+]
Frequent feeding
One study was identified that looked at the effect of frequency of initial feeds on blood glucose levels.
The study was a cross-sectional study of 156 term babies.371 A multiple regression analysis with method of feed, between-feed interval, volume of feed and postnatal age as independent variables found only between-feed interval (minutes) to be significantly correlated with blood glucose concentration (B = −0.003, SE = 0.001, β = −0.32, P < 0.05). [EL = 2+]
Breastfeeding
Ten studies were found that had implications for choice of feeding method.
The first study compared 45 breastfed babies with 34 formula-fed babies.372 The babies were 6 days old and matched for gestation and birthweight. Breastfed babies had significantly higher levels of ketones. [EL = 2+]
The second study compared 71 breastfed babies with 61 formula-fed babies.371 All babies were term babies less than 1 week old. Breastfed babies had significantly lower mean blood glucose concentration (P < 0.05) and significantly higher ketone body concentrations (P < 0.001). Breastfed babies had higher total gluconeogenic substrate concentrations (P < 0.01). [EL = 2+]
A cohort study investigated the glucose concentration of breast milk of women with diabetes and its relationship with the quality of metabolic control.373 The study involved 11 women with type 1 diabetes and 11 age-matched women without diabetes. The women with diabetes had intensified insulin treatment and their average HbA1c values were significantly higher than those in women without diabetes (8.1 ± 0.9% versus 6.2 ± 0.5%, P < 0.01). The glucose concentration of breast milk taken from women with diabetes did not differ from that of women without diabetes (0.68 ± 0.50 versus 0.66 ± 0.55 mmol/litre). No correlation was found between the maternal blood glucose (HbA1c) and the glucose concentration of breast milk. [EL = 2−]
A prospective cohort study374 investigated whether children born to women with diabetes were at increased risk of developing obesity and IGT in childhood. A total of 112 children of women with diabetes (type 1 diabetes, n = 83 and gestational diabetes, n = 29) were evaluated prospectively for impact of ingestion of either breast milk from a woman with diabetes or banked donor breast milk from women without diabetes during the early neonatal period (days 1–7 of life) on relative body weight and glucose tolerance at a mean age of 2 years. There was a positive correlation between the volume of breast milk from women with diabetes ingested and risk of overweight at 2 years of age (OR 2.47, 95% CI 1.25 to 4.87). In contrast, the volume of banked donor breast milk from women without diabetes ingested was inversely correlated to body weight at follow-up (P = 0.001). Risk of childhood IGT decreased by increasing amounts of banked donor breast milk ingested neonatally (OR 0.19, 95% CI 0.05 to 0.70). Stepwise regression analysis showed volume of breast milk from women with diabetes to be the only significant predictor of relative body weight at 2 years of age (P = 0.001). The results suggest that early neonatal ingestion of breast milk from women with diabetes may increase the risk of becoming overweight and, consequently, developing IGT during childhood. [EL = 2+]
A prospective cohort study375 investigated whether intake of breast milk of women with diabetes during the late neonatal period and early infancy influenced subsequent risk of overweight (adipogenic) and IGT (diabetogenic) in children born to women with diabetes. One hundred and twelve children born to women with diabetes were evaluated for influence of ingesting their mother’s breast milk during the late neonatal period (second to fourth neonatal week) and early infancy on relative body weight and glucose tolerance in early childhood. Exclusive breastfeeding was associated with increased childhood relative body weight (P = 0.011). Breastfed children of women with diabetes had an increased risk of overweight (OR 1.98, 95% CI 1.12 to 3.50). Breastfeeding duration was also positively related to childhood relative body weight (P = 0.004) and 120 minute blood glucose during an OGTT (P = 0.022). However, adjustment for the volume of breast milk from women with diabetes ingested during the early neonatal period (i.e. the first week of life), eliminated all these relationships with late neonatal breastfeeding and its duration. No relationship was observed between maternal blood glucose in the middle of the third trimester and neonatal outcomes. The study suggests that neither late neonatal breast milk intake from women with diabetes nor duration of breastfeeding has an independent influence on childhood risk of overweight or IGT in children born to women with diabetes. The first week of life appears to be the critical window for nutritional programming in children of ingestion of breast milk from women with diabetes. [EL = 2+]
Another cohort study376 investigated whether late neonatal ingestion of breast milk might independently influence neurodevelopment in 242 children of women with diabetes. There was no impact of ingestion of breast milk of women with diabetes on psychomotor parameters, but it negatively influenced onset of speaking with children of women with diabetes who were fed solely on breast milk taking the longest time to initiate speech. Adjusting for the amount of breast milk ingested during the early neonatal period weakened the hazard ratio towards non-significance. The data suggest that neonatal ingestion of breast milk of women with diabetes, particularly during the first week of life, may delay speech development, an important indicator of cognitive development. [EL = 2++]
Another cohort study investigated the extent to which early breastfeeding or exposure to cow’s milk affected psychomotor and cognitive development in children of women with diabetes.377 Children of women with diabetes with early breast milk ingestion achieved early psychomotor developmental milestones (lifting head while prone, following with eyes; P = 0.002). However, children who had ingested larger volumes of milk of women with diabetes had a delayed onset in speaking compared to those with lower milk intake (P = 0.002). The data suggest that ingesting larger volumes of milk of women with diabetes may normalise early psychomotor development in babies of these women, but may delay onset of speaking. [EL = 2++]
A systematic review378 summarised the clinical evidence relating a short duration of breastfeeding or early cow’s milk exposure to the development of type 1 diabetes. People with type 1 diabetes were more likely to have been breastfed for less than 3 months during their infancy (pooled OR 1.43, 95% CI 1.15 to 1.77) and to have been exposed to cow’s milk before 4 months (pooled OR 1.63, 95% CI 1.22 to 2.17) compared to those without diabetes. The study suggests that early exposure to cow’s milk may be an important determinant of subsequent type 1 diabetes and may increase the risk approximately 1.5 times. [EL = 2++]
Another systematic review379 evaluated the relationship between early infant diet and the risk of developing type 1 diabetes in later life via a meta-analysis of 17 case–control studies involving 21 039 people who were either breastfed or introduced early to cow’s milk. The effect of exposure to breast milk substitutes on developing type 1 diabetes was small. [EL = 2++]
A case–control study investigated the association between the type of feeding in infancy and the development of type 1 diabetes.380 The study involved 100 children with type 1 diabetes and 100 children without diabetes matched for sex and age. Information on feeding patterns during the first year of life was collected using a questionnaire. A larger proportion of children with diabetes had been breastfed. There was no clear difference between children with diabetes and those without diabetes in terms of duration of breastfeeding (children with diabetes, median duration 3 months; children without diabetes, median duration 2 months). The data do not support the existence of a protective effect of breastfeeding on the risk of type 1 diabetes, or that early exposure to cow’s milk and dairy products influences the development of type 1 diabetes. [EL = 2+]
Barriers to breastfeeding in women with diabetes in pregnancy
One study was identified that compared breastfeeding initiation and maintenance in 33 women with type 1 diabetes to those of 33 women in a control group and 11 women in a reference sample.381 The control group consisted of women without diabetes selected using gestational age at delivery, method of delivery, sex of baby and prior lactation experience. The reference group consisted of women without diabetes who were within 90–110% of ideal body weight prior to conception, had uncomplicated pregnancies and delivered vaginally. The study found women with diabetes were more likely to experience difficulties establishing and continuing breastfeeding than control and reference groups. All differences were significant (P < 0.05). The difference between groups was attributed to differences in postpartum care. Hospital protocol placed all babies of women with diabetes in the neonatal unit after birth for monitoring for hypoglycaemia. This meant that women with diabetes saw their babies the least amount of time in the first 3 days postpartum, waited the longest to begin breastfeeding their babies and breastfed their babies fewer times. Other possible contributory factors were that 70% of the women with diabetes had undergone caesarean section and that 30% of the babies of women with diabetes were macrosomic. Women with diabetes cited baby sleepiness as the most common baby-feeding problem. A sleepy baby was not identified as a problem by any of the women in the control group and by only one woman in the reference group. [EL = 2+]
A case–control study382 investigated factors influencing the initiation and maintenance of breastfeeding in 22 women with type 1 diabetes and 22 women without diabetes. Diabetes was not a principal factor in the decision to breastfeed or bottle-feed for the majority of the women. Women who considered diabetes in their decision to breastfeed had on average 2 years more of education than those who did not (14.82 years versus 12.94 years). Although the women did not perceive diabetes as influencing their breastfeeding experiences, they found that maintaining good control of diabetes required greater effort and flexibility during breastfeeding. [EL = 2+]
Banking colostrum before birth
Two publications were identified in relation to production of colostrum from women with diabetes and banking colostrum before birth for use in the neonatal period.
A cohort study compared the composition of macro- and micronutrients in milk from six women with tightly controlled type 1 diabetes (median glycosylated haemoglobin concentrations at parturition of 5.2% (range 4.9–5.3%) and 6 weeks later of 6.1% (range 5.0–6.3%), reference range 5.0–6.4%) with that from five women without diabetes.383 Milk samples were collected halfway through a single breastfeed at days: 3–5 (colostrum); 7, 9 and 10 (transitional milk); and 12, 15, 17, 21, 25, 29 and 35 (mature milk). There were no differences between the two groups in terms of concentrations of macronutrients (triglycerides, lactose and protein), cholesterol, glucose or myoinositol, nor in fatty acid composition. The duration of colostrum lactation was the same for women with diabetes and those without diabetes (3–5 days in both groups). [EL = 2−]
A narrative non-systematic review considered expressing and banking colostrum antenatally for use in the neonatal period.384 The review suggested that women with conditions that may delay breastfeeding and those who wish to lessen known familial health problems for their expected babies (including women with type 1 diabetes or gestational diabetes) would benefit from antenatal expression of colostrum. The risk of nipple stimulation initiating oxytocin release and, therefore, preterm contractions, labour and preterm birth was discussed and a protocol for expressing and storing colostrum was suggested. The review concluded that expressing and storing colostrum is advantageous to babies and confidence building for women and should, therefore, be supported for any condition which healthcare professionals consider to be relevant. [EL = 4]
Testing for neonatal hypoglycaemia
A systematic review by the WHO366 found that screening for hypoglycaemia using glucose oxidase-based reagent strips had poor sensitivity and specificity. The report recommended that ‘less frequent but more accurate laboratory or ward-based glucose electrode measurements among babies at risk are preferable’.
Intravenous dextrose for neonatal hypoglycaemia
There is a consensus that intravenous dextrose should be administered for symptomatic hypoglycaemia and for asymptomatic hypoglycaemia that fails to respond to feeding.366,386,387 However, no clinical studies were identified in relation to evaluation of protocols for the treatment of neonatal hypoglycaemia using intravenous dextrose.
Current practice
The CEMACH enquiry33 reported that the opportunity for early skin-to-skin contact after birth was achieved in 29% (30) of the 102 babies whose medical records were available. In eight cases, skin-to-skin contact was not possible due to the condition of the woman and/or the baby. Ninety-five percent of babies remaining with their mothers received their first feed on the labour ward compared with 50% of those admitted to a neonatal unit (P < 0.001). Twenty-six percent (29/112) of women received help with breastfeeding within 1 hour of birth (34% of women on labour wards and 12% of women in the neonatal unit). Thirty-one percent of women whose babies were admitted to the neonatal unit had documented evidence in their medical records that they were shown how to breastfeed and maintain lactation. Infant formula was given at the first feed for 63% (67/106 babies) and this was the first choice for women in 46% (32/70) of cases. Breast milk was the first feed for 50% (34/68) of babies that remained with their mothers and 21% (8/38%) of babies in the neonatal unit (P = 0.001). The first feed given was not the mother’s intended type of feed for 28% (27/96) of babies (16% of women who stayed with their babies and 50% of those admitted to the neonatal unit, P < 0.001). [EL = 3–4]
CEMACH undertook a descriptive study of all pregnancies of women with pre-existing diabetes who gave birth or booked between 1 March 2002 and 28 February 2003.2 The study found that 40.1% of all babies (1382/3451) were fed within 1 hour and 78.8% (2717/3451) by 4 hours. Among term babies 46.5% (1031/2216) were fed within 1 hour and 87.7% (1837/2216) within 4 hours. Exclusive breastfeeding was the choice at birth for 53% (1762/3342) of women with pre-existing diabetes compared with 69% in the general population. At 28 days after birth the proportion of exclusively breastfed babies was 23.8%, half the proportion who had intended to breastfeed at birth. A history of low blood glucose alone was the main reason (36.7%) for giving term babies of women with diabetes supplementary milk or glucose. In 9% of cases babies were given supplementary milk or glucose routinely according to local practice, possibly compromising establishment of breastfeeding. Of the 3451 babies in the study, 83.2% were tested within 6 hours and 47.3% were tested within 1 hour. Testing this early may, however, simply detect the normal drop in blood glucose that can be expected after birth. One-third of term babies were admitted to a neonatal unit for special care. Examining the reasons for admission suggested that many (67%) were avoidable. [EL = 3]
The CEMACH enquiry33 reported that neonatal blood glucose testing was mainly carried out using reagent strips. It supported the WHO’s recommendation that reagent strip testing is unreliable and recommended that when considering the diagnosis of hypoglycaemia at least one laboratory value should be obtained. The enquiry also recommended that women with diabetes should be informed antenatally of the beneficial effects of breastfeeding on metabolic control for them and their babies and that blood glucose testing performed too early should be avoided in well babies without signs of hypoglycaemia. [EL = 3–4]
A standard textbook of neonatology388 supports this evidence.
Existing guidance
The NSF for diabetes advises that babies born to women with diabetes should be fed as soon as possible after birth.20 It also recommends breastfeeding for babies of women with diabetes, but that women should be supported in the feeding method of their choice. [EL = 4]
The NICE guideline for routine postnatal care recommends that women should be encouraged to have skin-to-skin contact with their babies as soon as possible after birth and that initiation of breastfeeding should be encouraged as soon as possible after birth and ideally within 1 hour.11
Evidence statement
The blood glucose concentration used to guide intervention for neonatal hypoglycaemia (i.e. additional feeding and, if this does not reverse hypoglycaemia, intravenous administration of dextrose) is 2.6 mmol/litre. Close surveillance should be maintained in babies with risk factors for compromised metabolic adaptation if the plasma glucose concentration is less than 2.0 mmol/litre; at very low concentrations (1.1–1.4 mmol/litre) an intravenous glucose infusion is indicated to raise the glucose level above 2.5 mmol/litre.
Two studies showed that early feeding of babies was associated with lower incidence of hypoglycaemia than late feeding (more than 12 hours after birth). However, these studies involved preterm babies who may demonstrate different metabolic adaptation to term babies.
Another study showed between-feeding interval to be correlated with blood glucose levels, suggesting that frequent feeding should be encouraged to prevent neonatal hypoglycaemia.
Three studies relating to choice of infant-feeding method for women with diabetes suggested that: breastfeeding may enhance ketogenesis and that ketones may be an important alternative to glucose for brain metabolism in the neonatal period; breastfeeding babies of women with diabetes was not associated with increased exposure of the babies to high glucose levels; and, where possible, separation of mother and baby should be avoided to enable early feeds, frequent feeds and breastfeeding, with the possibility that supplementary feeding with infant formula may be required for women with diabetes who breastfeed. However, the first of these studies involved 6-day-old babies and therefore has limited relevance to hypoglycaemia in babies of women with diabetes, who are at greatest risk of hypoglycaemia in the first 12 hours.
A further seven observational studies, including two systematic reviews of observational studies, examined associations between feeding method and long-term outcomes. Three of the studies showed that obesity, IGT and impaired cognitive development were associated with ingestion of breast milk from women with diabetes. However, the two systematic reviews, which showed an association between breastfeeding and subsequent development of diabetes, were not specific to children of women with diabetes.
Two studies reported that initiation and maintenance of breastfeeding was more difficult for women with diabetes because of routine separation of babies from their mothers at birth or clinical reasons for separation such as the woman having undergone caesarean section or the baby having macrosomia. Although diabetes was not a major factor in deciding whether to breastfeed, women with diabetes found that maintaining good control of diabetes required greater effort and flexibility during breastfeeding. These findings suggest that, where possible, separation of the mother and baby should be avoided to facilitate early, frequent feeds and breastfeeding. Supplementary feeding with infant formula may be required for women with diabetes who breastfeed.
No clinical studies were identified in relation to the potential benefits of expressing and storing colostrum antenatally for the purposes of supporting early feeding to prevent hypoglycaemia in babies of women with diabetes.
A systematic review by the WHO noted low sensitivity and specificity of reagent strip blood glucose testing to identify neonatal hypoglycaemia and recommended laboratory or ward-based glucose electrode measurements for babies at risk of neonatal hypoglycaemia.
No clinical studies were identified in relation to the evaluation of protocols for treatment of neonatal hypoglycaemia using intravenous dextrose.
From evidence to recommendations
In the absence of high-quality evidence, the GDG’s recommendations for the prevention and treatment of neonatal hypoglycaemia are based on group consensus. The GDG’s view is that all maternity units should have a local written protocol for the prevention, detection and management of hypoglycaemia in babies of women with diabetes. Breastfeeding is recommended to prevent neonatal hypoglycaemia (by promoting successful metabolic adaptation) alongside other known benefits. Early commencement of breastfeeding is more important in babies of women with diabetes because of the risk of neonatal hypoglycaemia and is encouraged by skin-to-skin contact. Babies of women with diabetes should, therefore, feed as soon as possible after birth and at frequent intervals thereafter. While the target level for blood glucose is 2.6 mmol/litre, the GDG has set the threshold for initiating intravenous administration of dextrose at 2.0 mmol/litre on two consecutive readings, despite maximal support for feeding. Babies of women with diabetes should not be treated with invasive procedures (such as tube feeding or intravenous dextrose) unless they have clinical signs of hypoglycaemia or unless their blood glucose values persist below the threshold for initiating intravenous dextrose.
Blood glucose measurements should be obtained using ward-based glucose electrode or laboratory analysis because these have greater sensitivity and specificity than reagent strip testing. In making this recommendation the GDG noted the findings of the CEMACH enquiry, which reported that reagent strip testing is still commonplace. The GDG’s view is that blood glucose should be tested before feeding the baby.
Publication Details
Copyright
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.
Publisher
RCOG Press, London
NLM Citation
National Collaborating Centre for Women's and Children's Health (UK). Diabetes in Pregnancy: Management of Diabetes and Its Complications from Preconception to the Postnatal Period. London: RCOG Press; 2008 Mar. (NICE Clinical Guidelines, No. 63.) 7, Neonatal care.