Abstract
34 out of the 48 babies experienced some form of neonatal complication and were admitted to the neonatal unit. The median duration of stay in the neonatal unit was 7 days. 14 of the surviving neonates suffered from respiratory distress, although only 4 required surfactant therapy to regain respiratory function. However, the incidence of serious neonatal complications in those born after 34 weeks was shown to be low.
Introduction
Of the 750,000 women who give birth in England and Wales each year, approximately 5% have diabetes. Of these it is estimated that 87.5% are women that have developed gestational diabetes, while the remaining 12.5% have either pre-existing Type 1 or Type 2 diabetes [1]. Diabetic pregnancies carry a higher risk of complications which can impact on maternal, fetal and neonatal outcomes [2]. Diabetes is known to increase the risk of pre-eclampsia, preterm labour and miscarriage, as well as worsening severity of pre-existing microvascular disease such as retinopathy, nephropathy and neuropathy [3] . With regards to fetal outcome, there is a fivefold increased risk of stillbirth and neonatal death [3], as well as a substantially higher risk of congenitalmalformations, macrosomia, and in turn, an increased risk of birth injury and perinatal complications [3].
The Saint Vincent’s Declaration of 1989 set out the target of improving maternal and neonatal outcomes in women with diabetes over five years, with the desired eventual aim of complication rates matching those of the non-diabetic population [4]. However many studies carried out ten and even twenty years later have demonstrated disappointing results, despite improvements in antenatal, intrapartum and postnatal care [5].
The Confidential Enquiry into Maternal and Child Health (CEMACH) diabetes programme conducted in the UK was the largest study of pregnancy outcomes in women with diabetes ever conducted, documenting almost 4,000 pregnancies during 2002-2003. It found that perinatal outcomes remained poor, with a 3.8 times increased risk of perinatal death in diabetic mothers compared to mothers without diabetes [3]. The findings were similar to previous studies by Hawthorne [5] and Casson et al. in 1994 [6], and were confirmed by Inkster et al. in 2006 [7]. The most dramatic finding of the CEMACH study was however the unusually high prevalence of preterm deliveries (five times as many as non-diabetic women).
The 2015 National Institute of Clinical Excellence (NICE) guidelines for the management of diabetes in pregnancy states that pregnant women with Type 1 or Type 2 diabetes, and no other complications should be offered elective birth via induction of labour, or Caesarean section if indicated, between 37 and 38+6 weeks. It also suggests that elective birth before 37 completed weeks should be considered if there are metabolic or any other maternal or fetal complications [1]. Having said this, there is no clear guidance as to when to deliver should there be maternal or fetal complications compromising the safety of either (or both) patients. Due to improvements in obstetric and neonatal technology, there has been a subtle shift in opinion for the management of such cases from the ‘watch and wait’ attitude to more active intervention in these patients, in order to achieve the most desirable outcome; a healthy baby and a healthy mother.
Despite the apparent trend towards iatrogenic preterm delivery, the neonatal outcome of such intervention is not well documented in the literature. Therefore, the aim of this study is to explore neonatal outcomes in babies born to diabetic mothers who have had planned elective preterm deliveries at various preterm gestations.
Method
The study was approved by the department of medicine audit lead. Once approved, all pregnancies of diabetic women attending the consultant led Medical Obstetric clinic at Hull Royal Infirmary, Women and Childrens’ Hospital, East Yorkshire who had delivered before 37 weeks gestation between January 2009 and March 2012 were reviewed.
Inclusion criteria were: planned preterm (before 37 completed week’s gestation) iatrogenic deliveries, including induction of labour and pre-labour Caesarean Section. Mothers included were those with type 1, type 2 or gestational diabetes (both insulin-dependent and non-insulin dependent). Names and unique identification numbers of these patients were obtained from the Hull and East Yorkshire Hospitals Trust patient database. Records were anonymised for analysis.
A pro-forma was completed using obstetric and paediatric case notes. Most of the information required was extracted and any information that could not be found was obtained from computerised records; either the Hull and East Yorkshire Trust patient database or ‘Badger’- Hull Royal Infirmary Neonatal Unit’s database. The pro-formas were completed by EVC and NS and analysed using IBM SPSS Statistics 19 (SPSS inc Chicago, Illinois) software.
Results and analysis
Of the 50 mothers who met the inclusion criteria, 45 were analyzed. Of the five not used, four of the case notes were untraceable and one was misidentified during the selection period and delivered at term.
In total there were 45 pregnancies. There were also three sets of twins within the cohort giving a total of 48 babies.
Maternal characteristics are outlined in Table 1. 15 (33%) women presented with pre-existing complications of diabetes of which the most common was retinopathy (46.7%). 33.3% of the 15 had multiple complications, most commonly a variable combination of retinopathy, nephropathy, neuropathy and hypertension. 18 women (40%) had significant comorbidities other than those directly associated with diabetes. Examples are haemophilia B, oropharyngeal carcinoma, polycystic ovary syndrome, thyroid dysfunction and coeliac disease. In addition to this there were two opiate dependent women within the cohort.
Maternal characteristics of 45 maternities
Maternal characteristics of 45 maternities
Five (10.9%) of the women had had previous perinatal loss. 18 women had suffered from a previous miscarriage, of which 6 women had suffered from more than one. One woman had 12 previous miscarriages before the index pregnancy.
Whether patients received pre pregnancy counselling or not was poorly documented in notes, with only 7 being clearly documented as havingreceived it.
12 women (26.7%) had spontaneous rupture of membranes (SROM) before their planned delivery date. Of these – 4 had spontaneous vaginal deliveries, 2 had induction of labour – of which one went on to have a cesarean section as a result of failure to progress, and 6 underwent emergency cesarean section, the most common reason being due to fetal malpresentation. Overall cesarean section represented 80% of deliveries within the cohort.
The most common reason for planning preterm delivery was macrosomia, followed by poor glucose control and pre-eclampsia; poor glucose control also being a common supportive secondary and tertiary indication for delivery, along with polyhydramnios (Table 2).
Delivery characteristics
Delivery characteristics
Among women who were already diabetic prior to becoming pregnant, the mean pre pregnancy HbA1c was 9% indicating suboptimal control outside of pregnancy. The mean last HbA1c recorded before delivery was 7.9%, which shows slightly improved diabetic control.
Gestational ages ranged from 29+3 to 36+6 weeks, with a mean delivery gestation of 34+6 weeks. Birth weights over the 90th centile, were recorded in 24(54.2%) however weight appears to normalise after birth, shown by the fact that at discharge only 17(36%) babies were above the 50th centile(Table 3).
Neonatal characteristics
*1 = 10 Discharge weights not recorded. *2 = 1 Early neonatal death.
Of all the babies born 34(72%) were admitted to the neonatal unit. There were varying levels of care required, all were admitted to SCBU during their stay, however 7 needed high dependency care and 11 required intensive care. The median length of stay on the neonatal unit was 7 days – one baby stayed for 94 days, thus making the mean a less accurate representation of neonatal stay.
14 of the surviving neonates suffered from respiratory distress however only 4 required therapy with surfactant. Antenatal steroids were given in 41.7% of cases. This decision may have been due to the fact that many of the babies were at an advanced gestation. It should be noted that 12 of the surviving neonates suffered from respiratory distress after birth, however, only 4 required surfactant to regain respiratory function.
11 neonates were oxygen dependent, the median duration of those requiring oxygen being 3 days, one baby was oxygen dependent for 94 days.
During their time in hospital, 17 babies suffered from 1 or more episodes of hypoglycaemia, only 2 babies suffered from pulmonary hypertension, a problem more prevalent in preterm babies. In addition no babies suffered from intraventricular haemorrhage or necrotising enterocolitis (Table 4). Out of the 48 babies, 47 left hospital alive. Unfortunately one early neonatal death occurred.
Neonatal outcomes – incidence of complications
Neonatal outcomes – incidence of complications
The early neonatal death was in a baby that was born at 29 weeks and 3 days gestation. This is the earliest gestation of the cohort. The mother was a 32 year old type 1 diabetic with a 20 year history of insulin dependence. She had diabetic retinopathy and a spinal cord stimulator inserted for chronic back pain. Her BMI was 25. She was Gravida 3 Para 1 – with a history of one past iatrogenic preterm delivery, which resulted in a healthy infant. At the end of 2nd trimester she was admitted with pregnancy induced hypertension and at 29+3 gestation fetal monitoring demonstrated an abnormal cardiotocograph therefore an emergency caesarean section was required. This revealed placental abruption. Attempts to resuscitate the infant were unsuccessful and were stopped after 45 minutes.
In total there were 11 babies delivered at less than 34 weeks gestation and 37 delivered over 34 weeks. Of these, all babies before 34 weeks were routinely admitted to NNU, compared to 23 of the 37 over 34 weeks (62%). 6 of the 11 <34 weeks suffered from respiratory distress (54.5%), of which 3 needed surfactant (27.2%). In comparison, 8 babies above 34 weeks suffered from respiratory distress (21.6%) of which only one required surfactant (2.7%). 6 out of the 11 born at <34 weeks required some form of oxygen therapy (54.5%), in comparison to 5 babies born after 34 weeks (13.5%).
Discussion
The outcome of 48 babies delivered from diabetic mothers is reported. The results indicate that the outcome was good overall and the only perinatal loss was the result of an abruptio placenta.
Given that the major complication of diabetic pregnancy is stillbirth, and that is a very serious complication, it is surprising that the neonatal outcome arising from iatrogenic premature delivery has not been better represented in the literature.
In the CEMACH detailed study of 442 women with type 1 and type 2 diabetes, there were 77 stillbirths and 21 neonatal deaths. Only 6/77 (7.8%) of the stillbirths were associated with congenital abnormalities while 10/21 (47.6%) of the neonatal deaths were associated. Post mortem results reveal the major category of cause of death was “unknown antepartum stillbirths” (59%) [3].
As the risk of antepartum stillbirth is numerically greater than the risk of neonatal death (particularly in the absence of congenital abnormalities) it is reasonable to consider iatrogenic preterm deliveries in the cases where there is maternal diabetes. The CEMACH enquiry documented a preterm delivery rate of 36% in the larger cohort of 3808 diabetic pregnancies which has become a part of accepted UK practice. The neonatal outcome of this change in obstetric practice is not widely reported and this report seeks to address this issue.
NICE guidelines for diabetic women recommend that delivery be considered between 37 and 38+6 weeks, as continuation could be detrimental to the outcome of pregnancy due to the associated risks of LGA babies and deliveries complicated by shoulder dystocia [8]; thus differing from the widely accepted view amongst the non-diabetic population that the closer the gestation is to term the better the outcomes. However, within diabetic pregnancies, the associated complications mean that many women either do not reach this desired gestation due to spontaneous onset of preterm labour or earlier iatrogenic intervention. In these cases, there is no clear indication of the safest time to intervene and actively manage the patient with either induction of labour or Caesarean section.
The CEMACH report specifically states that elective preterm delivery should be avoided at all costs if there is no clear indication [3]. It has been shown that within this cohort, despite being preterm, neonatal outcomes are positive with good prognosis for early life as represented by the high survival rate. 29.2% did not need any additional care on a specialist unit, and of those that did 41.2% were admitted for less than 5 days. Therefore while we cannot suggest a specific cut-off for elective preterm delivery, we can suggest that elective preterm delivery is a safe option if there is a clear indication to do so.
Premature babies in the general population are at a higher risk of developing respiratory failure, congenital malformations and infections [9, 10]. This is reflected in the results of the CEMACH study that showed that there was a high incidence of respiratory morbidity in earlier gestations, particularly in mothers with poor glycaemic control therefore emphasised the need to exercise caution in planning the induction of labour in these patients [3]. Of the 48 babies studied here, 14 experienced respiratory distress,5 were diagnosed with congenital abnormalities and no infections were noted. However two thirds of the babies with respiratory distress recovered well without the need for surfactant, indicating many cases were of low severity.
In normal circumstances, antenatal steroids are administered to many women who are at risk of delivering prematurely in order to promote fetal lung development and therefore reduce the likelihood of respiratory distress syndrome. Despite the problems of steroids causing poor diabetic control NICE has concluded that diabetes should not be a contra-indication to the administration of antenatal steroids, provided the women are closely monitored and have additional insulin. This conclusion was based on data from two studies which found that none of the women developed hypoglycaemia or ketoacidosis when steroids were administered [11, 12]. In this study, steroids were administered in 41.7% of cases.
During the study, an attempt was made to record HbA1c/fasting glucose levels of all mothers immediately before, during and after the pregnancies to ascertain a possible relationship between fasting blood glucose levels and the rate of complications and neonatal outcomes. However, it became apparent that monitoring and the recording of these in the mothers’ medical notes were extremely variable so it was difficult to standardise and apply these figures in the analysis of the results. It would have been useful to see whether poor glucose control and fetal macrosomia as the primary indication was associated with poorer outcomes compared to other indications such as preeclampsia and CTG abnormalities. Poor glucose control was a primary indication in 6 of the mothers, however the complication rate in the neonates was so low the 2 groups must be comparable. In view of this a very large data set would be needed to ascertain as to whether the babies of these mothers suffered poorer outcomes in order to highlight the importance of tight diabetic control to minimise the risk of preterm delivery.
Limitations of study
The sample size of 48 babies limits the wider applicability of the conclusions, however we hope other units will similarly analyse their data in order to validate our findings.
The relevant comparator is the outcome of pregnancies where there were similar complications and a decision to deliver was not undertaken.This comparison is not possible with this data set but the high rate of intrauterine death over all diabetic pregnancies is a complication that is relevant.
All women included in the study were due to have an elective preterm delivery. That said, 12 went into spontaneous preterm labour before their planned delivery date, meaning that technically they were no longer classed as iatrogenic preterm deliveries. These individuals were still included in the data set on an intention to treat basis.
This study was a single centre project. While the findings are promising, this fact limits the ability for the application of the results to the wider population due to the relatively narrow demographics. There is scope however for a larger national trial to be performed to allow more reliable application of the findings for the guidance of practice.
It is not possible to evaluate the long-term outcomes of all the babies within the scope of this project, so the effects of prematurity in later life cannot be established. We do know, however, that up until the present day no further deaths have occurred within the cohort.
Strengths
The strength of our study is that this is the first to report detailed neonatal outcomes from electively delivered premature babies of diabetic mothers and that our data should be of great interest to others in the field.
Conclusions
We report a favorable neonatal outcome from iatrogenic preterm delivery in diabetic women. Whilst elective preterm delivery is not the norm and should only be undertaken in the presence of clear clinical indications, this preliminary data suggests that the problem of stillbirth of unknown origin in diabetic pregnancy could be partially addressed by early delivery of high risk women. We encourage others to report their results, particularly medium and long term outcomes.
Footnotes
Acknowledgments
EVC and NS made equal contributions in this study.