Objective: To investigate which of the components of the metabolic syndrome best predict its development.
Design: Long-term cohort of randomly selected adults.
Patients: One thousand five hundred and forty-eight subjects from the Hong Kong Cardiovascular Risk Factor Prevalence Study who did not have the metabolic syndrome by the US National Cholesterol Education Program (NCEP) or International Diabetes Federation (IDF) criteria at baseline.
Measurements: Waist circumference, blood pressure, glucose, triglycerides and high-density lipoprotein-cholesterol (HDL).
Results: After a median interval of 6.4 years, there were 219 and 143 new cases (21.9 and 14.3 per 1000 person-years) of the metabolic syndrome by the NCEP and IDF criteria, respectively. The odds ratio for the NCEP metabolic syndrome was highest for low HDL, 4.08 [95% confidence interval (CI): 2.90-5.73] and that for the IDF metabolic syndrome was highest for central obesity, 5.94 [95% CI: 3.98-8.87]. Low HDL, found in 27.8% men and 34.3% women, had the highest sensitivity for the NCEP metabolic syndrome (48% in men and 57% in women) and the IDF metabolic syndrome (41% in men and 54% in women). Central obesity had the highest positive predictive values except that triglycerides had the highest positive predictive value for the NCEP metabolic syndrome in women. The areas under the receiver operator characteristic curve for waist circumference, triglycerides and HDL were similar. A model that included waist circumference and HDL predicted the metabolic syndrome as well as a model that included all five metabolic syndrome components.
Conclusion: Obese Chinese adults should be periodically screened for the metabolic syndrome and have waist and HDL measurement.