General Practice

Body fat distribution before pregnancy and gestational diabetes: findings from coronary artery risk development in young adults (CARDIA) study

BMJ 1995; 311 doi: (Published 28 October 1995) Cite this as: BMJ 1995;311:1139
  1. Shumin Zhang, postdoctoral fellowa,
  2. Aaron R Folsom, professora,
  3. John M Flack, associate professorb,
  4. Kiang Liu, professorc
  1. aDivision of Epidemiology, School of Public Health, University of Minnesota, Minneapolis, MN 55454-1015, USA
  2. bHypertension Center, Bowman Gray School of Medicine, Winston-Salem, NC 27157-1032, USA
  3. cDepartment of Community Health and Preventive Medicine, Northwestern University Medical School, Chicago, IL 60611, USA
  1. Correspondence to: Dr Folsom.
  • Accepted 3 August 1995

An increased waist to hip ratio is associated with abnormalities of glucose metabolism and an increased risk of non-insulin dependent diabetes mellitus.1 We therefore prospectively tested the hypothesis that a greater prepregnancy waist to hip ratio might increase the incidence of gestational diabetes.

Subjects, methods, and results

In 1985-6, the coronary artery risk development in young adults (CARDIA) study examined 5115 adults aged 18-30 years from four communities in the United States.2 Second, third, and fourth examinations were performed in 1987-8, 1990-1, and 1992-3. Prepregnancy risk factors assessed in 1985-6 included body mass index (in kg/m2), waist (minimal abdominal girth), and hip (maximal protrusion) circumferences. Fasting serum insulin concentration was measured by immunoassay.

At each examination participants reported whether they had ever had diabetes. Women also were asked about each interim pregnancy, including “Did you have diabetes during this pregnancy?” No attempt was made to validate this self report of gestational diabetes as this was an analysis of an existing data set.

Of 1083 women who reported pregnancy during the seven years between the baseline and fourth examinations, 56 reported gestational diabetes. We excluded women with a history of diabetes or pregnancy at baseline, a pregnancy of less than 25 weeks, a multiple pregnancy, or missing gestational diabetes status, leaving 720 women, 44 of whom reported gestational diabetes.

Cumulative incidence (the number of women who reported incident gestational diabetes divided by the number of women who became pregnant) was calculated within thirds of baseline anthropometric data. Relative risks were calculated by unconditional logistic regression (SAS Institute, Cary, North Carolina).

Of the 720 women, 48% (346) were white and 52% (374) were black; at baseline, 63% (454) had >12 years' education; 16% (115) reported diabetes among first degree relatives; 39% (281) were parous. Age adjusted relative risks of gestational diabetes were significantly raised (P<0.05) in the highest versus lowest third of baseline weight (relative risk=2.23), body mass index (2.49), waist circumference (2.31), hip circumference (2.44), and fasting serum insulin concentration (2.66), but there was no dose-response pattern. In contrast, for waist to hip ratio the age adjusted relative risk monotonically increased from 2.61 in the second third to 4.17 in the highest third (P for trend=0.0003).

Multivariate adjusted relative risk of gestational diabetes in relation to baseline prepregnancy body mass index, waist to hip ratio, and other covariates; coronary artery risk development in young adults (CARDIA) study, 1986-93

View this table:

When the factors were considered jointly (table; model 1), gestational diabetes was strongly associated with waist to hip ratio (relative risk=2.7 for the middle third and 4.0 for the highest third), but not with body mass index (relative risk=1.9 for the highest third). Adjustment for fasting serum insulin concentration (model 2) attenuated slightly the association of gestational diabetes with waist to hip ratio. In model 2, fasting serum insulin concentration itself had a relative risk for the highest third of 1.82 (P for trend=0.14).


A high waist to hip ratio is associated with increased risk of gestational diabetes and may be a better marker than body mass index of excess visceral fat and insulin resistance.1 Insulin resistance has been linked to the pathogenesis of gestational diabetes,3 although impaired insulin secretion has also been implicated.3

Among people without diabetes, fasting insulin concentration is correlated relatively highly with insulin action.4 A value in the upper third was associated with a higher risk of gestational diabetes, confirming that women inclined to develop gestational diabetes are indeed more insulin resistant.

Self reported diabetes is reasonably accurate5 but, without screening, diabetes and gestational diabetes may go undetected. Screening for gestational diabetes is widely practised, yet no information on such screening was available for women in the CARDIA study.

Our findings require replication with validation of gestational diabetes. Nevertheless, they indicate that a high waist to hip ratio can be included among risk markers for gestational diabetes.


We thank Laura Kemmis and Heather McCreath for technical assistance.


  • Funding This research was supported by National Heart, Lung, and Blood Institute contracts NO1-HC-48047, NO1-HC-48048, NO1-HC-48049, NO1-HC-48050, and NO1-HC-95095.

  • Conflict of interest None.


View Abstract