Methods
The study population consisted of all pregnancies between 1998 and 2012 registered in the Swedish Medical Birth Register (MBR) that the National Board of Health and Welfare maintains. This national register contains more than 98% of all births in Sweden from 1973 onwards. Information on all hospital births is gathered prospectively and includes demographic data, reproductive history, and complications during pregnancy, delivery, and the neonatal period. The MBR was validated in 2002, and the quality of the variables included was regarded as high.9
In Sweden, the main screening strategy for GDM during the study period was based on repeated capillary random blood glucose with a plasma glucose ≥9.0 mmol/L, in combination with traditional risk factors (88.7% of the population use this screening model).10 ,11 If a risk factor exists (previous large for gestational age child or birth weight >4500 g, previous GDM pregnancy, heredity, body mass index (BMI) >30 in first trimester), an oral glucose tolerance test (OGTT) was undertaken in gestational weeks 28–32. Random blood glucose is measured 4–6 times during pregnancy, with the first measurement in the first trimester. If an OGTT is undertaken in early pregnancy based on a high random glucose and the OGTT is considered normal, the OGTT is repeated in gestational weeks 28–32. A small number of regions in the country have offered a simplified OGTT to all pregnant women since 1995, but only including a 2 hour blood glucose measurement with no fasting glucose sampling (11.3% of the pregnant population).12
There are no data on exact changes in local routines between the study years. The register does not contain data on laboratory measures such as blood glucose. Maternal pre-pregnancy BMI was calculated as weight in kilograms divided by height in meters squared. BMI was used as a continuous variable, or categorized according to the WHO classes.13 Maternal weight was recorded in light indoor clothes at the first antenatal visit in the first trimester. Height was registered on recall. Maternal age was age at delivery.
Diagnosis of GDM is based on the result of a 75 g OGTT.
During the study period, the main diagnostic criteria for GDM were fasting capillary whole blood glucose ≥6.1 mmol/L (considered equivalent to plasma glucose 7.0 mmol/L) and/or 2 hour blood glucose ≥9 mmol/L (considered equivalent to plasma glucose 10.0 mmol/L).11 ,14 During 1998–2010, one region (Stockholm-Örebro) diagnosed only those with ‘overt diabetes’ as GDM (fasting capillary plasma glucose ≥7 mmol/L or 2 hour plasma glucose ≥12.2 mmol/L). This region represents ∼20–25% of the pregnant population.
The identification of T1DIP was based on coding from International Classification of Diseases 10th Edition (ICD-10): O240, O240B, O24°C, O240D, O240E, O240F, O240X. From 1998 onward, the diagnosis of type 2 diabetes in pregnancy was separated from type 1 diabetes coded as ICD O241; hence, 1998 was chosen as the start date for this analysis. The time period was divided into 3-year groups from 1998 onward. The 3-year periods were considered reasonable, to smooth out the year-to-year variation. The coding of type 2 diabetes, which had been uncommon, may have been suboptimal in the first 1–2 years after 1998, but improved with increasing familiarity. Ethnicity was based on country of birth and on the numbers from each country divided into Nordic (Sweden, Denmark, Finland, Norway and Iceland), South Asia (India, Pakistan, Nepal, Bhutan, Sri Lanka, Bangladesh), North Africa (Algeria, Egypt, Djibouti, Libya, Ethiopia, Eritrea, Morocco, Somalia, Tunisia, Sudan), the Middle East (Iran, Iraq, the United Arab Emirates, Bahrain, Gaza area, Jordan, Israel, Lebanon, Kuwait, Palestine, Oman, Qatar, Saudi Arabia, South Yemen, Syria, West bank, Yemen) and Other (other countries, consisting mainly of European countries).
