OBJECTIVE: To evaluate waist-to-height ratio (WTHR), waist girth and body mass index (BMI) as predictors of cardiovascular risk factors in Australian Aborigines. DESIGN: Indices were examined as predictors of mean blood pressures (BP), blood lipids, glucose, insulin and as predictors of hypertension, decreased HDL-cholesterol, elevated glucose and type 2 diabetes. SETTING: Aboriginal communities in remote north-west Australia. PARTICIPANTS: Four hundred and one adults. RESULTS: More than 80% of participants had WTHR > or = 0.5 or BMI > 22 kg/m2 and 78% had increased waist girth (> 90 cm men; > 80 cm women). Adverse BP, blood lipids, glucose and insulin were associated with classification above the cutpoint for each index. For fasting glucose > or = 5.6 mmol/L, sensitivity was 91% with WTHR, 87% with waist girth and 88% with BMI; respective specificities were 29%, 29% and 44%. Area under receiver operating characteristic (AUROC) curves gave only "fair" accuracy for any discriminatory variable. With diabetes AUROC for BMI (0.59, 95% CI 0.53, 0.65) was significantly lower than with other indices (WTHR 0.74, 95% CI 0.69, 0.79; girth 0.72, 95% CI 0.67, 0.78) but did not differ significantly for fasting glucose, HDL-cholesterol or hypertension. AUROC did not differ significantly between men and women for any outcomes. CONCLUSIONS: The indices did not discriminate well for diabetes or cardiovascular risk factors in these Aborigines, but waist girth or WTHR appeared more useful than BMI. Appropriate cutpoints are needed. WTHR is simple, does not need sex-specific cutpoints and could be useful in developing public health strategies.
OBJECTIVE: To evaluate waist-to-height ratio (WTHR), waist girth and body mass index (BMI) as predictors of cardiovascular risk factors in Australian Aborigines. DESIGN: Indices were examined as predictors of mean blood pressures (BP), blood lipids, glucose, insulin and as predictors of hypertension, decreased HDL-cholesterol, elevated glucose and type 2 diabetes. SETTING: Aboriginal communities in remote north-west Australia. PARTICIPANTS: Four hundred and one adults. RESULTS: More than 80% of participants had WTHR > or = 0.5 or BMI > 22 kg/m2 and 78% had increased waist girth (> 90 cm men; > 80 cm women). Adverse BP, blood lipids, glucose and insulin were associated with classification above the cutpoint for each index. For fasting glucose > or = 5.6 mmol/L, sensitivity was 91% with WTHR, 87% with waist girth and 88% with BMI; respective specificities were 29%, 29% and 44%. Area under receiver operating characteristic (AUROC) curves gave only "fair" accuracy for any discriminatory variable. With diabetes AUROC for BMI (0.59, 95% CI 0.53, 0.65) was significantly lower than with other indices (WTHR 0.74, 95% CI 0.69, 0.79; girth 0.72, 95% CI 0.67, 0.78) but did not differ significantly for fasting glucose, HDL-cholesterol or hypertension. AUROC did not differ significantly between men and women for any outcomes. CONCLUSIONS: The indices did not discriminate well for diabetes or cardiovascular risk factors in these Aborigines, but waist girth or WTHR appeared more useful than BMI. Appropriate cutpoints are needed. WTHR is simple, does not need sex-specific cutpoints and could be useful in developing public health strategies.