Recent American Diabetes Association guidelines recommend selective screening for gestational diabetes mellitus (GDM) based on the presence of risk factors (1). Because there is a high incidence of GDM among certain ethnic groups, ethnicity is included as one of the risk factors during routine screening. Although Asians are included as a high-risk population, there are little data comparing selective testing based on risk factors with the universal 50-g glucose challenge as a means of screening for GDM in this group. Moreover, the risk factors associated with increased risk were largely derived from populations of European extraction, and there are few studies examining this in other populations. We have therefore conducted a study to assess whether the traditional risk factors for GDM apply to women of Asian origin and to determine whether the use of risk factors might be sufficient to select patients for testing.

Medical records of all women born in China, the Philippines, Sri Lanka, or Vietnam receiving antenatal care at Westmead Hospital (Sydney, Australia) between 1988 and 1996 were reviewed. All of these women received a nonfasting 50-g glucose challenge at 24–26 weeks. If the 1-h capillary blood glucose level was ≥7.8 mmol/l, then the patient returned for a formal 75-g glucose tolerance test (GTT) and was classified as having GDM if on a 75-g GTT there was one blood glucose level ≥6.0, 10, or 8.0 mmol/l at 0, 1, or 2 h, respectively. GDM status and risk factors, including age, weight at first antenatal visit, BMI, gravida, parity, family history of diabetes, previous macrosomia, previous pregnancy-induced hypertension, previous miscarriage or stillbirth, and previous fetal malformation, were recorded for each subject.

The records of 2,797 completed pregnancies in 2,139 Asian women were examined. Twin pregnancies, miscarriages, terminations, and women with pre-existing diabetes were excluded. As a whole, the incidence of GDM was 9.2%; among women born in China, it was 8.6% (66 of 764), the Philippines 6.7% (40 of 599), Sri Lanka 10.5% (31 of 293), and Vietnam 10.6% (121 of 1,141). These incidences are comparable with those found in studies of Asian women in developed countries but are higher than those found in studies conducted in Asia.

Data were examined to determine the sensitivity of using clinical and historical risk factors to select patients for screening. Table 1 displays the effect of applying different BMI and age thresholds for screening. By changing the thresholds for testing, the sensitivity of using clinical risk factors for selecting patients to be tested for GDM as well as the number of women required to be tested vary. If 97% of GDM cases were detected (using an age threshold of 25 years and BMI of 30 kg/m2), then 91% of all pregnant women need to be tested.

Using logistic regression analysis, all risk factors examined were predictive for GDM (Table 2). There were differences between the four groups, but in some cases, this may have been caused by sample size. The only risk factor significant in all four populations was a previous history of GDM with an odds ratio of 14.5 for the entire group.

Our data indicate that although the historical or clinical risk factors for GDM are valid in Asians, using risk factors alone to select such patients for testing for GDM is inadequate. Many Asian women who develop GDM have no risk factors at all. To avoid overlooking significant numbers of women with GDM, one may lower the specificity of the criteria, but this requires that the majority of patients be tested. Logistically, it is much simpler to conduct universal screening for all Asian women in Western countries, rather than to apply selective testing in order to spare a small percentage of women from being tested. Therefore, our findings strongly support recommendations for universal screening for GDM in pregnant women of Asian origin in Western countries. However, in places where the incidence of GDM is low, such as in some developing countries, the selection of patients for testing by the risk factors may be reasonable.

Table 1 —

Effect of varying BMI and weight thresholds on utility of selective screening showing the sensitivity of risk-factor screening and the percent of subjects who would require testing if patients were tested only when they had positive risk factors

% Screened% GDM detected
Age ≥ 25 or BMI ≥ 25 92 97 
Age ≥ 25 or BMI ≥ 30 91 97 
Age ≥ 30 or BMI ≥ 25 72 89 
Age ≥ 30 or BMI ≥ 30 71 89 
% Screened% GDM detected
Age ≥ 25 or BMI ≥ 25 92 97 
Age ≥ 25 or BMI ≥ 30 91 97 
Age ≥ 30 or BMI ≥ 25 72 89 
Age ≥ 30 or BMI ≥ 30 71 89 
Table 2 —

Risk factors for GDM and their predictive value

GDM (n = 258)NonGDM (n = 2,539)Odds ratio
Age (years) 32.5 ± 4.7 29.8 ± 4.9 1.1* 
BMI (kg/m223.3 ± 3.8 22.0 ± 2.9 1.1* 
Gravida 2.4 ± 1.4 2.2 ± 1.2 1.1* 
Parity 0.8 ± 1.1 0.7 ± 0.8 1.2* 
Family history of diabetes 24.3 12.5 2.2* 
Previous GDM 14.3  1.1 14.5* 
Previous macrosomia  5.0  2.1 2.5* 
Previous PIH  6.6  2.0 3.4* 
Previous miscarriage/stillbirth 0.4 ± 0.8 0.2 ± 0.5 1.6* 
Previous fetal malformation  5.1   1.6 3.3* 
GDM (n = 258)NonGDM (n = 2,539)Odds ratio
Age (years) 32.5 ± 4.7 29.8 ± 4.9 1.1* 
BMI (kg/m223.3 ± 3.8 22.0 ± 2.9 1.1* 
Gravida 2.4 ± 1.4 2.2 ± 1.2 1.1* 
Parity 0.8 ± 1.1 0.7 ± 0.8 1.2* 
Family history of diabetes 24.3 12.5 2.2* 
Previous GDM 14.3  1.1 14.5* 
Previous macrosomia  5.0  2.1 2.5* 
Previous PIH  6.6  2.0 3.4* 
Previous miscarriage/stillbirth 0.4 ± 0.8 0.2 ± 0.5 1.6* 
Previous fetal malformation  5.1   1.6 3.3* 

Data are means ± SD or % population.

*

P < 0.001.

1
American Diabetes Association: Gestational diabetes mellitus.
Diabetes Care
23(Suppl. 1)
:
S77
–S79,
2000

Address correspondence to Dr. N.W. Cheung, Department of Diabetes and Endocrinology, Westmead Hospital, Westmead, NSW 2145, Australia. E-mail: [email protected].