Metformin or thiazolidinedione therapy in PCOS?

Citation metadata

Date: Mar. 2011
From: Nature Reviews Endocrinology(Vol. 7, Issue 3)
Publisher: Nature Publishing Group
Document Type: Clinical report
Length: 1,421 words

Document controls

Main content


Insulin resistance has a key role in the pathogenesis of polycystic ovary syndrome, a common endocrinopathy that affects young women. Therefore, the use of insulin-sensitizing agents has been studied extensively in this patient population. Given the effectiveness of metformin and thiazolidinediones in previous studies, the question arises: is one class better than the other?

Grover, A. & Yialamas, M. A. Nat. Rev. Endocrinol. 7, 128-129 (2011); published online 1 February 2011; doi: 10.1038/nrendo.2011.16

Full Text: 

Polycystic ovary syndrome (PCOs) is a common endocrinopathy in young women, with an estimated prevalence of 6.5%. (1) PCOs is defined by at least two of the following diagnostic criteria when other known etiologies of hyperandrogenism and oligomenorrhea are excluded: oligoanovulation or anovulation, biochemical and/or physical signs of hyperandrogenism and polycystic ovarian morphology on ultrasound examination. (2) As stated by the definition, the clinical manifestations of the disorder include oligomenorrhea and hyperandrogenism (that is, acne, hirsutism and/or alopecia in a male pattern distribution). Additional common clinical manifestations include infertility, obesity and insulin resistance, which is a major concern in this patient population. In fact, as many as 30-40% of patients with PCOs have impaired glucose tolerance, and at least 10% of these individuals have type 2 diabetes mellitus by age 40 years. (3) As a result, a large amount of research focuses on the use of insulin sensitizers to treat this disorder. A recent study conducted by Li et al. (4) compares metformin to thiazolidinediones to assess all-round superiority with respect to a number of clinical manifestations of PCOS and safety.

Much research has supported the hypothesis that insulin resistance has a key role in the pathogenesis of PCOs. (5) Insulin, both directly and indirectly, stimulates ovarian androgen production. In addition, insulin decreases the hepatic production of sex hormone-binding globulin and, therefore, increases free androgen levels. Last, a complex interaction between insulin resistance and altered pulsatile secretion of gonadotropins may be responsible for the menstrual irregularities commonly seen in patients with PCOs.

The efficacy of insulin-sensitizing agents, such as metformin and thiazolidinediones, has been studied for treatment of the various clinical manifestations of PCOs. The results have been mixed. Most protocols have studied metformin, and its ability to decrease insulin and androgen levels has been wellconfirmed.6 Some studies have shown that decreasing insulin and androgen levels can increase frequency of ovulation and, therefore, improve menstrual regularity and fertility as well as mitigating symptoms of hyperandrogenism, (7) whereas other studies have failed to show an increase in live birth rates8 or improvement in clinical symptoms of hyperandrogenism. (6) Thiazolidinediones, although not as extensively studied as metformin, have also been reported to improve peripheral insulin sensitivity and mitigate the various metabolic, cardiovascular and reproductive elements of PCOS. Much of the data about thiazolidinediones has been extrapolated from studies on troglitazone, which is no longer available in the US and Europe owing to the prevalence of adverse liver effects. (9) Another study compared pioglitazone to metformin, and after 6 months, no differences with regard to improvement in insulin resistance and free testosterone levels were noted. (10)

The systematic review of Li et al. evaluates the clinical efficacy and safety of metformin versus the thiazolidinediones rosiglitazone and pioglitazone. Their meta-analysis of 10 randomized clinical trials focused on some of the metabolic and reproductive derangements associated with PCOS. These measurements included the lipid profile (total cholesterol, triglycerides, HDL cholesterol, LDL cholesterol), biochemical measurements of hyperandrogenism, such as free testosterone, dehydroepiandrosterone sulfate (DHEAS) and androstenedione, as well as insulin levels, BMI and rates of ovulation and pregnancy. Their results do not support an overall increased benefit of one class of drugs over the other.

Thiazolidinediones were shown to be more effective than metformin at reducing levels of free testosterone and DHEAS after the first 3 months of treatment. However, this benefit was not evident after 6 months of treatment. On the other hand, metformin was shown to result in greater decreases in triglyceride levels and BMI at 6 months of treatment than thiazolidinediones. No significant differences between rates of ovulation, pregnancy, insulin sensitivity and the remainder of the lipid panel were observed. The main limitations of the meta-analysis, also acknowledged by Li and colleagues, include the great heterogeneity of the different patient populations that were compared.

So which insulin sensitizer do we offer our patients and for which manifestations? One important consideration is the adverse effect profile of these medications. Metformin is known to have notable gastrointestinal adverse effects, as confirmed by Li et al., that can be mitigated by slowly titrating the dose, and, rarely, has been associated with lactic acidosis. To date, thiazolidinediones have been associated with increased cardiovascular risk (especially rosiglitazone, the use of which has been restricted by the FDA because of safety concerns), weight gain (difficult to accept by women with PCOS who are trying to lose weight), liver toxicity, increased risk of osteoporosis and class C drug classification according to the FDA pharmaceutical pregnancy categories (adverse effects on the fetus in animal reproduction studies and no adequate and well-controlled studies in humans). Because of the adverse effects associated with thiazolidinediones, most endocrinologists prefer to prescribe metformin to young women with PCOS.

Other key considerations are the therapeutic objectives and goals of the patient and the physician. These considerations are important when distinguishing between the use of insulin-sensitizing agents and more traditional therapies. If the primary concern is cosmetic in nature, then oral contraceptives, with or without concomitant spironolactone, have been shown to be superior to both metformin and thiazolidinediones. In fact, the data suggesting that insulin sensitizers are very effective in treating clinical hyperandrogenism symptoms, even when testosterone levels decline, is mixed. With respect to oligomenorrhea, oral contraceptives or intermittent progesterone administration again trump metformin. For the patient whose primary concern is infertility, metformin may be considered, more so in patients with obesity.

Each patient with PCOS is part of a larger heterogeneous patient population with a varying spectrum of clinical presentations. Importantly, each patient should be treated as an individual, taking into account treatment goals as well as the adverse effect profile of the various therapeutic options.

Practice points

* Polycystic ovary syndrome (PCOS) is a common endocrinopathy in young women and is associated with several cardiovascular (hypertension, abnormal lipid profile), metabolic (insulin resistance), reproductive (anovulation, infertility) and cosmetic (hirsutism, acne, androgenic alopecia) abnormalities

* Insulin resistance has a key role in the pathogenesis of PCOS

* The use of insulin-sensitizing agents has become increasingly popular in patients with PCOS, as they have been shown to improve insulin resistance and rates of ovulation

* Although no evidence of an all-round superiority of metformin over thiazolidinediones exists, metformin is preferred, given its more favorable adverse effect profile

Competing interests

The authors declare no competing interests.

(1.) Apridonidze, T., Essah, P. A., Iuorno, M. J. & Nestler, J. E. Prevalence and Characteristics of the Metabolic Syndrome in Women with Polycystic Ovary Syndrome. J. Clin. Endocrinol. Metab. 90, 1929-1935 (2005).

(2.) Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop Group. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil. Steril. 81, 19-25 (2004).

(3.) Ehrmann, D. A., Barnes, R. B., Rosenfield, R. L., Cavaghan, M. K. & imperial, J. Prevalence of impaired glucose tolerance and diabetes in women with polycystic ovary syndrome. Diabetes Care 22, 141-146 (1999).

(4.) Li, X. et al. Metformin versus thiazolidinediones for treatment of clinical, hormonal, and metabolic characteristics of polycystic ovary syndrome: a meta-analysis. Clin. Endocrinol. (Oxf.) doi:10.1111/j.1365-2265.2010.03917.x.

(5.) Ehrmann, D. Polycystic ovary syndrome. N. Engl. J. Med. 352, 1223-1236 (2005).

(6.) Lord, J. M., Flight, I. H. & Norman, R. J. Metformin in polycystic ovary syndrome; systematic review and meta-analysis. BMJ 327, 951-953 (2003).

(7.) Nestler, J. E., Jakubowicz, D. J., Evans, W. S. & Pasquali, R. Effects of metformin on spontaneous and clomiphene insuced ovulation in the polycystic ovary syndrome. N. Engl. J. Med. 338, 1876-1880 (1998).

(8.) Legro, R. S. et al. Clomiphene, metformin, or both for infertility in the polycystic ovary syndrome. N. Engl. J. Med. 356, 551-566 (2007).

(9.) Azziz, R. et al. Troglitazone improves ovulation and hirsutism in the polycystic ovary syndrome: a multicenter, double blind, placebo-controlled trial. J. Clin. Endocrinol. Metab. 86, 1626-1632 (2001).

(10.) Ortega-Gonzalez, C. et al. Responses of serum androgen and insulin resistance to metformin and pioglitazone in obese, insulin-resistant women with polycystic ovary syndrome. J. Clin. Endocrinol. Metab. 90, 1360-1365 (2005).

Department of Medicine, Brigham and Women's Hospital, 70 Francis Street, Shapiro Building, 3rd floor, Boston, MA 02115, USA (A. Grover, M. A. Yialamas). Correspondence to: M. A. Yialamas

Source Citation

Source Citation   

Gale Document Number: GALE|A250471047