Factors associated with stillbirth in women with diabetes

Diabetologia (2019) 62:1938–1947 https://doi.org/10.1007/s00125-019-4943-9 ARTICLE Factors associated with stillbirth in women with diabetes Sharon T. Mackin 1 & Scott M. Nelson 2 & Sarah H. Wild 3 & Helen M. Colhoun 4 & Rachael Wood 5 & Robert S. Lindsay 1 & on behalf of the SDRN Epidemiology Group and Scottish Diabetes Group Pregnancy subgroup Received: 16 January 2019 / Accepted: 20 May 2019 / Published online: 29 July 2019 # The Author(s) 2019 Abstract Aims/hypothesis Stillbirth risk is increased in pregnancy complicated by diabetes. Fear of stillbirth has major influence on obstetric management, particularly timing of delivery. We analysed population-level data from Scotland to describe timing of stillbirths in women with diabetes and associated risk factors. Methods A retrospective cohort of singleton deliveries to mothers with type 1 (n = 3778) and type 2 diabetes (n = 1614) from 1 April 1998 to 30 June 2016 was analysed using linked routine care datasets. Maternal and fetal characteristics, HbA1c data and delivery timing were compared between stillborn and liveborn groups. Results Stillbirth rates were 16.1 (95% CI 12.4, 20.8) and 22.9 (95% CI 16.4, 31.8) per 1000 births in women with type 1 (n = 61) and type 2 diabetes (n = 37), respectively. In women with type 1 diabetes, higher HbA1c before pregnancy (OR 1.03 [95% CI 1.01, 1.04]; p = 0.0003) and in later pregnancy (OR 1.06 [95% CI 1.04, 1.08]; p < 0.0001) were associated with stillbirth, while in women with type 2 diabetes, higher maternal BMI (OR 1.07 [95% CI 1.01, 1.14]; p = 0.02) and pre-pregnancy HbA1c (OR 1.02 [95% CI 1.00, 1.04]; p = 0.016) were associated with stillbirth. Risk was highest in infants with birthweights <10th centile (sixfold higher born to women with type 1 diabetes [n = 5 stillbirths, 67 livebirths]; threefold higher for women with type 2 diabetes [n = 4 stillbirths, 78 livebirths]) compared with those in the 10th–90th centile (n = 20 stillbirths, 1685 livebirths). Risk was twofold higher in infants with birthweights >95th centile born to women with type 2 diabetes (n = 15 stillbirths, 402 livebirths). A high proportion of stillborn infants were male among mothers with type 2 diabetes (81.1% vs 50.5% livebirths, p = 0.0002). A third of stillbirths occurred at term, with highest rates in the 38th week (7.0 [95% CI 3.7, 12.9] per 1000 ongoing pregnancies) among mothers with type 1 diabetes and in the 39th week (9.3 [95% CI 2.4, 29.2]) for type 2 diabetes. Conclusions/interpretation Maternal blood glucose levels and BMI are important modifiable risk factors for stillbirth in diabetes. Babies at extremes of weight centiles are at most risk. Many stillbirths occur at term and could potentially be prevented by change in routine care and delivery policies. Keywords Birthweight . Epidemiology . Neonatal . Pregnancy . Stillbirth . Type 1 diabetes . Type 2 diabetes Electronic supplementary material The online version of this article (https://doi.org/10.1007/s00125-019-4943-9) contains peer-reviewed but unedited supplementary material, which is available to authorised users. * Robert S. Lindsay Abbreviations IUGR LGA RDS SCI-Diabetes SGA SIMD SMR02 Intrauterine growth restriction Large for gestational age Respiratory distress syndrome Scottish Care Information-Diabetes Small for gestational age Scottish Index of Multiple Deprivation Scottish Morbidity Record 02 1 Institute of Cardiovascular and Medical Sciences, British Heart Foundation Glasgow Cardiovascular Research Centre, University of Glasgow, 126 University Place, Glasgow G12 8TA, UK 2 School of Medicine, University of Glasgow, Glasgow, UK 3 Usher Institute of Population Health Science and Informatics, University of Edinburgh, Edinburgh, UK Introduction 4 Institute of Genetics and Molecular Medicine, University of Edinburgh, Edinburgh, UK 5 ISD Scotland, Edinburgh, UK Mothers with pregestational diabetes are at 4–5-fold increased risk of stillbirth [1], with data from our and other populations showing no improvement in recent Diabetologia (2019) 62:1938–1947 years [1, 2]. This contrasts with decreasing stillbirth rates seen in the general obstetric population [1]. Maternal obesity, advanced maternal age and smoking are important modifiable risk factors for stillbirth in the general obstetric population [3, 4]. Fetal growth is also important, with growth-restricted pregnancies having the highest risk [3]. Data on pregnancies complicated by diabetes are more limited. Suboptimal maternal blood glucose levels even at minimal levels, presence of microvascular complications and poor preparation for pregnancy are associated with stillbirth [5, 6]. Other traditional risk factors seen in the general obstetric population are less well documented in diabetes. Prevention of stillbirth underpins part of the clinical rationale for obstetric intervention in diabetes, particularly around timing of delivery. While we lack predictive models, presence of risk factors may guide obstetricians to earlier delivery, which is appropriate in many cases but associated with neonatal morbidity [7]. We therefore analysed national data from all deliveries to mothers with pregestational diabetes in Scotland over an 18 year period, to better define maternal and fetal characteristics associated with stillbirth. Timing of stillbirth was also analysed to identify potential for population-based strategies around routine delivery. 1939 Methods Data sources As previously described [1], we linked data from maternity records in the Scottish Morbidity Record 02 (SMR02) database and the national diabetes database, Scottish Care Information-Diabetes (SCI-Diabetes). SMR02 contains clinical information on all obstetric inpatient episodes across Scotland including maternal and infant demographics, obstetric complications and delivery details. Quality assurance procedures have shown >90% completion and accuracy for data [8]. SCI-Diabetes contains patient demographics and clinical information on diabetes diagnosis, presence of complications and management. National data capture is excellent, with 99.5% of the Scottish population with diabetes included from 2004 onwards [9]. Diabetes diagnosis is entered onto SCI-Diabetes by individual clinical teams, and correlates with inpatient records in greater than 99% of cases [9]. For this study, type of diabetes was further refined by algorithm based on prescription history and age of diagnosis. Type 1 was reclassified as type 2 diabetes if there was more than 1 year without diabetes medications prescribed or treatment with oral hypoglycaemic agents only. Type 2 diabetes was 1940 reclassified as type 1 if diabetes was diagnosed under 30 years of age and initiated on insulin therapy within 1 year of diagnosis. Episodes that resulted in the delivery of an infant at or beyond 24 weeks of gestation from 1 April 1998 to 30 June 2016 were identified from SMR02. Linkage with SCIDiabetes identified mothers diagnosed with type 1 (...truncated)