Association between outcome of pregnancy and glycaemic control in early pregnancy in type 1 diabetes: population based study

Participants, methods, and results

This observational study was carried out in a single centre in Norwich from January 1991 to December 2000. The resident population is 510 000 and mainly white. We defined adverse pregnancy outcome as spontaneous abortion (first or second trimester), major congenital malformation (potentially life threatening or associated with serious long term disability), stillbirth, or neonatal death. We measured glycated haemoglobin concentration at booking for prenatal care and then monthly using the Biomen 8140 method.

Women were divided into two groups according to their glycated haemoglobin concentration at booking; women with values <7.5% (mean of normal range plus 5 standard deviations) were defined as having fair control and those with values ≥7.5% were defined as having poor control. The study was approved by the local ethics committee.

We included only the first pregnancy for each woman during the study in the statistical analyses to avoid possible biases. We analysed data with SPSS software using Student's t test, Mann-Whitney U test, and χ² test as appropriate. Fisher's exact test was used for small numbers.

There were 242 pregnancies in 158 women. Thirty two pregnancies had an adverse outcome, with 18 (7%) spontaneous abortions, eight (3%) major congenital malformations (six neural tube defects), four stillbirths, and two neonatal deaths.

We studied the relation between glycated haemoglobin concentration at booking and adverse outcome in the 158 first pregnancies. The table shows the patient characteristics and pregnancy outcomes. Adverse outcome was significantly higher in the poor control group than the fair control group (relative risk 4.3, 95% confidence interval 1.8 to 10). Compared with the fair control group, the poor control group had a fourfold increase in the spontaneous abortion rate (relative risk 4.0, 1.2 to 13.1) and ninefold increase in the congenital malformation rate (relative risk 9.2, 1.1 to 79.9). Perinatal mortality was higher in the poor control group than the fair control group (54/1000 births v 19/1000, relative risk 2.8, 0.41 to 19.4) but with the small numbers the difference was not significant. Perinatal mortality in the background population is 7.8/1000.

Comment

We found a significant relation between adverse outcome of pregnancy and poor glycaemic control in early pregnancy in women with type 1 diabetes. There was a fourfold increase in adverse outcome, a fourfold increase in spontaneous abortion, and a ninefold increase in major congenital malformation in women with a glycated haemoglobin concentration above 7.5% at booking. Our study has substantial advantages over earlier studies, being a complete, prospective, population based, single centre study analysing only one pregnancy per woman. It confirms earlier reports of increased risk of spontaneous abortion and malformation with poor glycaemic control in early pregnancy in women with type 1 diabetes.3^–5 Our findings suggest that good glycaemic control around the time of conception is necessary to optimise outcome of pregnancy in diabetic women. Diabetic women and their carers need to be advised of the risks and encouraged to optimise glycaemic control before and during pregnancy.