LAUSR

BACKGROUND The risk of developing prediabetes and type 2 diabetes (dysglycemia) may be increased in women with PCOS. Whether an oral glucose tolerance test (OGTT) should be performed routinely in all PCOS women at presentation or should be recommended only to a selected subset of patients is still controversial. BASIC PROCEDURES At a tertiary care center, we conducted a retrospective, observational study including 400 women with PCOS submitted to an OGTT. Our primary objective was to assess the diagnostic agreement between two algorithms commonly used for the screening of dysglycemia in these women: i) relying only on fasting plasma glucose (FPG) or ii) considering both fasting and/or 120-min plasma glucose concentrations during an OGTT. We conducted the analysis considering all patients as a whole, and also after stratifying them by body weight, androgen concentrations and age. MAIN FINDINGS The OGTT detected dysglycemia in 24.5% of patients, whereas only 14.3% women would have been diagnosed using FPG levels alone. The latter missed as many as 40% of women with dysglycemia in our series, including all cases of diabetes. Diagnostic agreement between both algorithms was only 0.55 (κ = 0.103; 95% CI: 0.05-0.16). Areas under the receiver operating characteristic curve for dysglycemia were 0.86 (95%CI: 0.81-0.91) for FPG and 0.91 (95%CI = 0.87-0.95) for 120-min plasma glucose during the OGTT. FPG was not accurate in predicting dysglycemia in women with PCOS regardless of the presence of insulin resistance, weight excess, hyperandrogenemia and age. PRINCIPAL CONCLUSIONS Relying on FPG alone is not adequate for the screening of disorders of glucose tolerance in women with PCOS; such diagnosis should rely on the results of an OGTT regardless of age, weight and/or androgen concentrations.