Diabetes in the Melanesian and Indian peoples of Fiji: a study of risk factors.

Epidemiological risk factor patterns for diabetes mellitus determined by hemoglobin A1 and fasting blood glucose criteria were compared in the biethnic (Melanesian and Indian) nation of Fiji. The 2 diagnostic criteria elicited essentially similar risk factor patterns in Indians but ranking of predictors was altered in Melanesians. By either criterion age was a dominant risk factor for diabetes in both ethnic groups with age2 a predictor in Indians of elevated hemoglobin A1 (chi 2 = 7.8, P less than 0.005) and fasting blood glucose (chi 2 = 25.3, P less than 0.0001). Age- and sex-adjusted prevalence of diabetes was higher in Indians than in Melanesians [RR = 2.5 (1.9-3.3)]. A positive family history was associated with increased risk of diabetes in both ethnic groups by the hemoglobin A1 criterion [pooled RR = 2.3 (2.0-2.6)] but was not significant in Melanesians under the fasting blood glucose criterion. A positive family history was a strong predictor of severe hyperglycemia in both ethnic groups. The relative risk for diabetes was greater in females [1.5 (1.2-9.1)], with no ethnic difference. There was no urban-rural difference in either ethnic group. The similar risk factor patterns for diabetes diagnosed by hemoglobin A1 and severe hyperglycemia suggest that elevated hemoglobin A1 may constitute a useful screening test for 'high risk' diabetic subjects.

Glycosylated hemoglobin as an index of the prevalence and severity of diabetes in biethnic Fiji.

Glycosylated hemoglobin was compared with fasting blood glucose as a screening test for diabetes mellitus and as an index of the severity of diabetes in biethnic (Melanesian and Indian) Fiji. Age-adjusted diabetes prevalence in the test sample was higher in Indians by either criterion. According to the hemoglobin A1 criterion, Melanesians had prevalence rates of 8.2% (males) and 15.8% (females) compared to 17.0% (males) and 24.3% (females) in Indians. In contrast, fasting blood glucose criteria (WHO) gave higher rates in each group. Hemoglobin A1 levels were higher overall in Indians and females. The predictive value of an elevated fasting blood glucose test for an elevated hemoglobin A1 was 20.0% in Melanesians and 60.7% in Indians while that of a normal fasting blood glucose test for a normal hemoglobin A1 was 89.4% in Melanesians and 89.3% in Indians. The proportion of Indians with elevated hemoglobin A1 who were severely hyperglycemic was almost 7 times higher (40.9% vs. 5.8%) than that of Melanesians. The ethnic difference in the predictive value of fasting blood glucose levels for hemoglobin A1 levels appears to be related to the greater severity of hyperglycemia of diabetic Indians compared to diabetic Melanesians. Hemoglobin A1 levels provide information on both the qualitative as well as quantitative differences in diabetes between ethnic groups.