Context: To date, the metabolic effects of the long-term metformin administration in anovulatory patients with polycystic ovary syndrome are know, whereas few data are available on the effects of its suspension.

Objective: To evaluate the effects of metformin suspension on insulin sensitivity in PCOS patients.

Design: Prospective randomized placebo-controlled study.

Setting: University "Magna Graecia" of Catanzaro, Italy.

Patients: Thirty normal-weight anovulatory PCOS women and 10 age- and body mass index (BMI)-matched healthy controls.

Intervention: PCOS patients were randomized to receive 1700 mg daily metformin (metformin group) or placebo tables (placebo group) for 12 months, whereas no treatment was administered in healthy women (control group).

Main outcome measures: Clinical, endocrine and metabolic profile, and clamp insulin sensitivity index was evaluated at study entry and after six, 12, 18 and 24 months.

Results: At baseline, the clamp insulin sensitivity index resulted significantly (P<0.05) different in PCOS patients in comparison with healthy controls, without difference between metformin and placebo groups. During treatment, it was significantly (P<0.05) improved in the metformin group in comparison with baseline and placebo group, without significant differences between the six- and 12-month assessments. At six and 12 months after treatment suspension, in the metformin group, insulin sensitivity index resulted significantly (P<0.05) worsened in comparison with that observed at baseline and during treatment, and with that observed in the placebo and control groups.

Conclusion: In normal-weight anovulatory PCOS patients, long-term metformin administration exerts beneficial effects on peripheral insulin sensitivity. Furthermore, this effect disappears at treatment suspension.