Background: Recently, based on HOMA-IR, we estimated empirical optimal cut-off values for SHBG levels of ≤41.5 nmol/L in women with PCOS. Other proposed markers of insulin resistance include triglyceride and glucose levels, and anthropometric measurements. Therefore, our current study aimed to analyze its consistency with the cut-off values that discriminate insulin resistance based on the TyG, TyG-BMI, and TyG-WC indices in women with PCOS. Methods: Age, body weight, height, waist circumference, glucose, insulin, triglyceride, and SHBG levels were retrieved from the medical records of 264 Caucasian women diagnosed with PCOS. The TyG, TyG-BMI, and TyG-WC indices were calculated. The mean meta-cut-off SHBG level was calculated using receiver-operating characteristic (ROC) analysis combined with diagnostic test accuracy meta-analysis. Results: The mean meta-cut-off value for SHBG levels for the assessment of insulin resistance was less than 43.1 (95% CI: 37.0–49.2) nmol/L. The pooled sensitivity and specificity of SHBG levels for the assessment of insulin resistance were 74.7% and 66.9%, respectively. The pooled mean prevalence of insulin resistance based on all indices was 36.1% (95% CI: 33.5–38.7%) with a standard deviation of 18.7% and positive predictive value (PPV) of 52.8% (95% CI: 12.2–87.5%) and the negative predictive value (NPV) of 80.2% (95% CI: 45.1–97.7%). Conclusions: Our study confirms the usefulness of SHBG level as a marker of insulin resistance in Caucasian women with PCOS. A value below 43 nmol/L, with high sensitivity and specificity, enables the detection of insulin resistance and a high risk of prediabetes, prompting close monitoring of liver function.
Kochanowicz et al. (Thu,) studied this question.