Context/objective: Current diagnostic criteria for polycystic ovary syndrome (PCOS) have generated distinct PCOS phenotypes, based on the different combinations of diagnostic features found in each patient. Our aim was to assess whether either each single diagnostic feature or their combinations into the PCOS phenotypes may predict insulin resistance in these women.
Patients/design: A total of 137 consecutive Caucasian women with PCOS, diagnosed by the Rotterdam criteria, underwent accurate assessment of diagnostic and metabolic features. Insulin sensitivity was measured by the glucose clamp technique.
Results: Among women with PCOS, 84.7% had hyperandrogenism, 84.7% had chronic oligoanovulation, and 89% had polycystic ovaries. According to the individual combinations of these features, 69.4% of women had the classic phenotype, 15.3% had the ovulatory phenotype, and 15.3% had the normoandrogenic phenotype. Most subjects (71.4%) were insulin resistant. However, insulin resistance frequency differed among phenotypes, being 80.4%, 65.0%, and 38.1%, respectively, in the 3 subgroups (P < .001). Although none of the PCOS diagnostic features per se was associated with the impairment in insulin action, after adjustment for covariates, the classic phenotype and, to a lesser extent, the ovulatory phenotype were independently associated with insulin resistance, whereas the normoandrogenic phenotype was not. Metabolic syndrome frequency was also different among phenotypes (P = .030).
Conclusions: There is a scale of metabolic risk among women with PCOS. Although no single diagnostic features of PCOS are independently associated with insulin resistance, their combinations, which define PCOS phenotypes, may allow physicians to establish which women should undergo metabolic screening. In metabolic terms, women belonging to the normoandrogenic phenotype behave as a separate group.