Outcome of pregnancy in diabetic women in northeast England and in Norway, 1994-7BMJ 2000; 321 doi: https://doi.org/10.1136/bmj.321.7263.730 (Published 23 September 2000) Cite this as: BMJ 2000;321:730
- G Hawthorne, consultant physician ()a,
- L M Irgens, professorb,
- R T Lie, professor of medical statisticsb
- a Northern Diabetic Pregnancy Survey, Regional Maternity Survey Office, Newcastle upon Tyne NE2 4AA
- b Medical Birth Registry of Norway, Hauletend Hospital, N-5021 Bergen, Norway
- Correspondence to: G C Hawthorne
- Accepted 11 May 2000
In northeast England, perinatal mortality is five times higher and congenital malformation is four times higher for pregnancies in diabetic women than for those in women who do not have diabetes.1 The same is true in other regions in the United Kingdom, 2 3 but this is not the case in Norway.4 Why is the outcome of pregnancy in diabetic women better in Norway? Does Norway use the same definitions and record the same outcome data as the United Kingdom? If the data are standardised will these differences disappear? If not, what is causing the difference in the outcome of pregnancy in diabetic women?
The methodology for data collection needs to be standardised between the two countries to determine whether differences in outcome are real before looking for explanations in the healthcare systems of the two countries. We compare the outcome of pregnancy in women with diabetes based on prospective population data collected in Norway and northeast England.
Methods and results
All data were prospectively collected in Norway and northeast England between 1 July 1994 and 30 June 1997. In Norway, all births with a gestational age of 16 completed weeks or more must be registered with the medical birth registry of Norway. Health problems such as diabetes are recorded in all women before and during pregnancy. Birth defects that are diagnosed before the child leaves hospital are recorded. Data are routinely linked with the cause of death register. In northeast England, the northern diabetic pregnancy survey collects prospective population based data on the outcome of pregnancy in diabetic women. Data on the total birth population are collected by the Regional Maternity Survey Office.
All women with pregestational diabetes who gave birth between 1 July 1994 and 30 June 1997 were included in the study. The same definitions for stillbirth and congenital anomaly were used by both centres. Stillbirths included all fetal deaths at more than 24 completed weeks of gestation. The data for Norway have been based on stillbirth >24 weeks, although the World Health Organization's definition of >28 weeks or 1000 kg is used routinely in Norway. Perinatal mortality comprised all stillbirths >24 completed weeks and all live births. Congenital anomalies were coded by using ICD-8 (international classification of diseases, 8th revision). Terminations for birth defects (to the extent reported) have been included in numerators and denominators. Perinatal deaths have been included in the denominator data. Relative risks were approximated by odds ratio; P values and confidence limits were obtained by using exact inference methods. Adjustment for maternal age did not alter the results. Attempts to adjust for birth order failed because too many births had information missing.
Perinatal mortality among babies born to mothers without diabetes in northeast England was 10/1000. Among babies born to mothers with diabetes it was 42.8/1000, a relative risk of 4.4 (95% confidence interval 2.5 to 7.7).
In Norway, perinatal mortality was 10.4/1000 and 6.7/1000 respectively for babies of mothers with diabetes and mothers without diabetes (relative risk of 1.5, 0.97 to 2.3). The relative risks in northeast England and Norway were significantly different (P=0.002).
The risk of birth defects was increased 2.5-fold among babies of mothers with diabetes in northeast England (1.5 to 4.0). In Norway the risk was increased 0.95-fold (0.73 to 1.2). These relative risks were significantly different (P=0.0008) (table).
Pregnancy in diabetic women remains high risk in northeast England, in contrast to Norway, where outcome approaches the WHO's St Vincent target.5 This is not explained by the process of data collection as care has been taken to make the data from both sources comparable. To determine the reasons for this marked difference in outcome more extensive investigation is required.
We thank Debbie Sen, Dr Judith Rankin, and Marjorie Renwick of the Regional Maternity Survey Office, Newcastle upon Tyne for their help in compiling the Northern region data and Vicki Ashton of Teesside University for statistical advice.
Contributors: GH had the original idea for the study and coordinated the data analysis. LMI supplied the data from Norway. RTL supplied the statistical expertise. The paper was written jointly by GH, LMI, and RTL.
Conflict of interest None.