Usefulness of the metabolic syndrome criteria as predictors of insulin resistance among obese Korean women.

OBJECTIVE: To investigate the ability of each metabolic syndrome (MetS) criterion, defined by the International Diabetes Federation, to predict insulin resistance (IR).

DESIGN: A cross-sectional study. IR was defined as homeostasis model assessment of IR (HOMA-IR) > or =3.04. The MetS criteria considered were TAG > or = 1.69 mmol/l, HDL cholesterol (HDL-C) <1.29 mmol/l, blood pressure (BP) > or =130/85 mmHg and fasting glucose (FG) > or =5.6 mmol/l.

SETTING: Busan, South Korea.

SUBJECTS: Ninety-six apparently healthy Korean women (mean age 42 (sd 10.6) years) with abdominal obesity (waist circumference (WC) > or =80 cm) were studied.

RESULTS: Of the ninety-six obese women, 11 % were insulin-resistant and 33 % fulfilled the criteria for IDF-defined MetS. Glucose and TAG were more likely to predict IR than BP and HDL-C when assessed using receiver-operating characteristic curves, multiple regression and multiple logistic regression analyses. Of the variation in HOMA-IR, TAG, FG, WC and age explained 42 %. High FG was independently associated with the presence of IR (OR = 8.6, 95 % CI 1.8, 41.8) even after adjusting for other components of MetS. The positive predictive value and positive likelihood ratio to detect IR were the highest for the FG criterion (33 % and 3.9, respectively), followed by TAG (28 %, 3.0), BP (19 %, 1.8) and HDL-C criteria (18 %, 1.7). The IDF definition of MetS exhibited a positive predictive value of 29 % and a positive likelihood ratio of 3.1.

CONCLUSIONS: Of the MetS criteria, high FG and high TAG seem to be more suitable for identifying obese women with IR than high BP and low HDL-C.