Fasting glucose/insulin ratio. An index to differentiate normo from hyperinsulinemic women with polycystic ovary syndrome.

OBJECTIVE

To evaluate a simple laboratory index useful to differentiate normo from hyperinsulinemic women with polycystic ovary syndrome.

DESIGN

Open and prospective study.

SETTING

Outpatient infertility clinic of a third level medical institution.

PATIENTS

Twenty five women 27.8 +/- 3.4 years of age with chronic anovulation, hirsutism and hyperandrogenemia (free testosterone [free-T] > 3.4 pg/mL) but no acanthosis nigricans (group 1) were compared with ten healthy women 27.5 +/- 1.5 years of age (group 2).

INTERVENTIONS

Three fasting blood samples were obtained followed by a 100-g, 2-hours oral glucose tolerance test (OGTT).

MAIN OUTCOME MEASURES

Glucose, insulin, FSH, LH, E2, free-T, androstenedione (A), DHEAS and 17 alpha-hydroxyprogesterone (17-OHP) were measured in the three fasting samples. Glucose and insulin were also determined in the OGTT samples. The fasting glucose/insulin (G/I) ratio was calculated.

RESULTS

In group 2 the fasting insulin was < 16.8 microU/mL, the sum of serum insulin (sigma INS) during the OGTT was < 385 microU/mL (mean + 3SD) and the fasting G/I ratio was > 4.5. The fasting LH, FSH, free-T, cortisol, and insulin were higher in group 1 than in group 2. In group 1, eleven women had a sigma INS below and fourteen above 385 microU/mL. Fasting insulin had a linear correlation with sigma INS (r = 0.780) while the fasting G/I ratio had an exponential correlation with sigma INS (r = -0.699). Fasting insulin versus the G/I ratio best fitted a reciprocal regression model (r = 0.912). For screening of hyperinsulinemia during OGTT, fasting insulin had a 75% sensitivity and 62% specificity while for the G/I ratio it was 79% and 73%, respectively.

CONCLUSIONS

Both fasting hyperinsulinemia (> 16.8 microU/mL) and a fasting G/I ratio < or = 4.5 can satisfactorily differentiate women with hyperandrogenism and hyperinsulinemia from those with normoinsulinemia. However, the ratio does not require a previous definition of a normal value as for fasting or post-oral glucose insulin levels, and thus, it can be easily calculated in daily clinical practice to establish specific therapeutic maneuvers at an early stage of the evaluation of such patients.