Clin Endocrinol 60(2):241-249, 2004
Hyperandrogenism, hyperinsulinaemia and obesity play a cardinal role in the pathogenesis of PCOS, contributing in different ways to its clinical expression (Franks, 1995; Dunaif, 1997; Poretsky et al., 1999; Gambineri et al., 2002). To achieve clinical efficacy, therapeutic strategies should therefore not only attempt to ameliorate symptoms of PCOS, but also to correct all main pathogenetic factors. This may imply a need for individualized therapeutical approaches according to the phenotype presentation.
Weight loss should represent the first-line therapeutic approach in improving hyperandrogenism in all obese PCOS women (Kiddy et al., 1992; Pasquali et al., 1989; Jakubowicz & Nestler, 1997), as it reduces insulin resistance and hyperinsulinaemia. Whether chronic addition of insulin sensitizers may be more effective than diet alone is still under debate. We have previously found that, in obese PCOS women, treatment with metformin and hypocaloric diet, in comparison to placebo, induced more favourable outcomes on body weight, visceral fat and androgen levels and a more significant improvement of hirsutism and menses abnormalities (Pasquali et al., 2000). These results are largely confirmed by the present findings. However, emphasis must be placed on the benefit of metformin in improving menses abnormalities, which appeared to depend both on metformin treatment per se and partly on weight loss, and on the insulin and androgen decrease, which suggests a still unknown specific effect of metformin on the regulation of circulating or local factors controlling the ovarian function (Poretsky et al., 1999).
Flutamide added to hypocaloric diet specifically favoured a further decrease in androgen levels, indicating that, in addition to its antagonistic effect on androgen receptors (Brogden & Clissold, 1989), it may also act through inhibition of androgen biosynthesis (Marugo et al., 1994; Diamanti-Kandarakis et al., 1995; De Leo et al., 1998b). Studies performed in rat testis tissues have in fact found that flutamide may decrease cytochrome P450 content (Clos et al., 1988) and 17α-hydroxylase and 17,20-lyase activities (Ayub & Levell, 1987), both involved in androgen formation. The tendency to a greater decrease in testosterone levels, particularly in the combined treatment group, and the significant selective reduction of androstenedione that was related only to flutamide treatment suggest that flutamide exerts a dual action also in vivo. This is further supported by the decrease of DHEA-S levels, which suggests a direct effect of the drug on adrenal steroidogenesis (Marugo et al., 1994; Diamanti-Kandarakis et al., 1995; De Leo et al., 1998a). However, the striking increase of SHBG concentrations, representing a specific effect of the two drugs' interaction, also contributes to the normalization of free testosterone observed in the combined treatment group. The androgen receptor antagonist and the reduced androgen biosynthesis therefore justify the specific effect of flutamide in decreasing the hirsutism score. However, its addition to metformin leading to free testosterone normalization may explain the further amelioration of hirsutism obtained by the combined therapy.
Changes in body composition appeared to be influenced by both hypocaloric diet and pharmacological treatment. In particular, the decrease in visceral fat tended to be greater in the combined pharmacological treatment and this effect might be attributable to flutamide, although this assumption has to be supported by larger studies. The lack of any additional effect of metformin to hypocaloric dieting on visceral fat loss that we have described here does not, however, confirm our previous findings (Pasquali et al., 2000), probably because of some differences in the PCOS women enrolled in the present study. In particular, although in both studies the women were obese, with an abdominal obesity phenotype, and insulin-resistant, those enrolled in the present study were less obese, less insulin-resistant and had a lower amount of visceral fat depots. These differences might in someway influence the different responsiveness to metformin and flutamide drug administration in term of VAT reduction. As visceral fat reduction occurred in the combined treatment group it cannot be excluded that metformin may play a complementary but still important effect. Nonetheless, this group had the greatest testosterone decrease combined with the most pronounced reduction in visceral fat. Considering the action of flutamide and metformin in reducing testosterone levels and the well-known role of testosterone in visceral fat metabolism (Björntorp, 1996; Lovejoy et al., 1996; Gambineri et al., 2002), we speculated that the greater VAT reduction observed during the combined therapy may in someway depend on greater testosterone reduction.
The administration of metformin and flutamide alone or in association had no further effect on insulin resistance and hyperinsulinaemia compared to hypocaloric dieting. Although aware of the limits related to the methods used to assessed hyperinsulinaemia and insulin sensitivity and of the small number of cases, these data, in accordance with other reports (Ehrmann et al., 1997), support the concept that dietary-induced weight loss has to be considered the main factor responsible for improving hyperinsulinaemia and insulin resistance. This result emphasizes the notion that weight lowering strategies should represent the first-line treatment choice in insulin-resistant obese PCOS women. On the other hand, in many controlled studies on the effect of metformin in nondieting PCOS women (Glüeck et al., 2002), the insulin resistance amelioration may even be attributable to the mild anorectic effect and the consequent weight loss as well as to the effect of metformin as an insulin sensitizer (Paolisso et al., 1998). The possibility that flutamide may improve insulin resistance is still under debate (Diamanti-Kandarakis et al., 1995; Diamanti-Kandarakis et al., 1998; Paoletti et al., 1999; Ibánez et al., 2000). Our study, however, suggests that when administered in dieting obese PCOS women, it does not add any significant further effect on insulin sensitivity to that induced by weight loss.
Conversely, flutamide administration proved to significantly reduce total and LDL cholesterol levels, whereas the combined metformin + flutamide treatment tended to increase HDL cholesterol levels. The effects on total and LDL cholesterol agree with previous reports (Diamanti-Kandarakis et al., 1998; Ibánez et al., 2000, 2002) and appear to depend on the antiandrogenic effects of flutamide, androgens being directly involved in the regulation of cholesterol and lipoprotein metabolism (Desprčs & Marette, 1994; Von Eckardsein, 1998; Anderson et al., 2002). On the contrary, the tendency to an increase of HDL cholesterol concentrations in the combined treatment group appeared to depend on the interaction between metformin and flutamide, confirming the combined and interactive role of insulin and androgens on HDL metabolism (Von Eckardsein, 1998).
In conclusion, this pilot study suggests that the addition of metformin, flutamide or of the combined metformin + flutamide treatment in obese PCOS women has a more favourable outcome with respect to the low-calorie diet treatment. In particular, flutamide treatment seems to have a significant effect in decreasing visceral fat, androstenedione, DHEA-S, total and LDL cholesterol and hirsutism. Conversely, metformin appears to have a significant benefit on menstrual regulation. Both drugs in association display an additive role in reducing testosterone concentration and a synergistic effect in increasing HDL cholesterol and SHBG levels. On the contrary, improvement of insulin sensitivity and hyperinsulinaemia appears to depend on hypocaloric diet, without any further significant effect of pharmacological treatment, either alone or in combination. These data therefore provide a rational basis for adapting different pharmacological options to the various clinical and laboratory requirements in obese PCOS women, although larger confirmative studies are needed.
Clin Endocrinol 60(2):241-249, 2004
Acknowledgements
We are indebted to Laboratori Guidotti SpA, Pisa (Italy) and Ipsen SpA, Milano (Italy) who provided metformin, flutamide and placebo tablets, to Dr Anastasia Carcello who performed the CT scans, and to Ms. Susan West for reviewing the English language of the manuscript.
Reprint Address
Renato Pasquali, MD, Endocrinology Unit, Department of Internal Medicine, S. Orsola-Malpighi Hospital, Via Massarenti 9, 40138 Bologna, Italy. Tel: +39-051-6364147; Fax: +39-051-6364147; E-mail:
renato.pasquali@unibo.it
Flutamide and Metformin Administered Alone or in Combination 
