A 26-year-old female comes to your office complaining of an inability to conceive for the past year despite regular, unprotected intercourse. She has not been evaluated for infertility in the past. Her medical history is significant for polycystic ovary syndrome (PCOS), which was diagnosed 5 years ago. She reports sporadic menses, occurring every 45 to 60 days. She is not taking any medications. On examination, she has acanthosis nigricans on her posterior neck, with no other significant findings. An initial workup demonstrates normal thyroid function and prolactin levels. Her fasting glucose level is 118 mg/dL and beta-human chorionic gonadotropin test results were negative. Luteinizing hormone/follicle-stimulating hormone ratio, free testosterone, and dehydroepiandrosterone levels are all above normal.
You have heard of utilizing metformin as an initial treatment for infertility in women with PCOS. You think metformin may be a good option because this patient's examination findings and laboratory test results suggest insulin resistance. However, you are unsure if the research supports the use of metformin and how to use it in this instance.
CLINICAL QUESTION
Is metformin beneficial as an initial treatment for women with PCOS and infertility?
BACKGROUND
PCOS is commonly seen in the family practice or obstetrics/gynecology office, affecting approximately 5% to 10% of reproductive-age women.1 This endocrinopathy is a common cause of infertility,1 with a currently accepted primary treatment of clomiphene for infertility.2 Recent studies examined the use of metformin as an initial treatment for infertility in women with PCOS.
SEARCH CRITERIA AND RESULTS
A literature search of peer-reviewed medical journals was conducted through PubMed, CINAHL, and MEDLINE Extra using the keywords PCOS, metformin, clomiphene, pregnancy, live birth, conception, and ovulation. The search was limited to clinical or randomized controlled trials, practice guidelines, or meta-analyses published in the last 5 years in English. A total of 49 studies were identified. Three of these studies included only women with PCOS, were randomized, utilized pregnancy and/or live birth as a primary or secondary outcome measure and did not include the exclusion criteria inclusion of known clomiphene-resistant women and utilization of surgical infertility treatments. An important note is that the focus should be on pregnancy AND live birth, as opposed to ovulation, as outcome measures because of a higher risk of spontaneous abortion in patients with PCOS.3 The exclusion criteria eliminate studies evaluating metformin as a secondary infertility treatment.
EVALUATING THE EVIDENCE
Palomba and colleagues conducted a parallel randomized, double-blind, double-dummy controlled clinical study in which 45 participants received metformin plus placebo and 47 participants received clomiphene plus placebo for 6 months.3 The placebo was necessary to mask which drug was given because clomiphene is taken for only 5 days of each month. Both pregnancy and live birth were the primary outcome measures. The cumulative pregnancy rate was significantly higher in the metformin group (68.9%) than in the clomiphene group (34%) (P < .001). A higher rate of live births in the metformin group (83.9%) compared with that of the clomiphene group (56.3%) was not statistically significant (P = .07). The authors determined that metformin is better than clomiphene as a primary treatment for infertility. However, this was true only in terms of pregnancy rates, not live-birth rates. Study limitations were insufficiently large sample size and the lack of multiple study sites. The authors felt that a larger sample size would have allowed the true impact of metformin on livebirth rates to be apparent.
Moll and colleagues conducted a multicenter, randomized clinical trial in which 111 patients received clomiphene plus metformin and 114 patients received clomiphene plus placebo.4 Therapy duration varied based on when pregnancy was achieved or clomiphene resistance, defined as failure to ovulate on 150 mg/d, was identified. Pregnancy was a secondary outcome measure and ovulation was the primary outcome measure. Patients treated with clomiphene plus metformin were less likely to achieve pregnancy (ongoing pregnancy rate of 40%) than were patients treated with clomiphene plus placebo (ongoing pregnancy rate of 46%; relative risk = 0.87, 95% confidence interval [CI], 0.6-1.2). Therefore, metformin is not effective when added to clomiphene treatment because pregnancy is less likely to be achieved with the combination treatment than with clomiphene alone. The authors also determined that patients taking metformin had an increased incidence of side effects, such as diarrhea and dyspepsia, compared with patients taking only clomiphene. A significantly higher dropout rate because of side effects in the metforminplus- clomiphene group (18 participants [16%]) compared to that of the clomiphene-only group (six participants [5%]) demonstrated this increased prevalence of side effects (risk difference = 11%, 5%-16%). Thus, metformin was found to increase the presence of intolerable side effects when added to clomiphene therapy. Although not an outcome measure in this study, 21 live births were noted in the metforminplus- clomiphene group and 30 were noted in the clomiphene-plus-placebo group (no statistics given). Although this was a larger-scale, multisite study, the impact of metformin on live-birth rates is unclear without the calculated statistics. Secondly, the authors' definition of risk was not stated in the published study. Finally, the effect of metformin on pregnancy in this study could have been the result of chance because the relative risk CI crosses 1.
Legro and colleagues conducted a multicenter randomized trial in which 209 participants received clomiphene plus placebo, 208 participants received metformin plus placebo, and 209 participants received clomiphene plus metformin for up to 6 months (study medications were discontinued when a participant had a positive pregnancy test result).5 Live birth was a primary outcome measure and pregnancy was a secondary measure. Live-birth rates were significantly lower in women treated with metformin only (7.2%) than in women treated with clomiphene only (22.5%) or women treated with both drugs (26.8%) (metformin versus clomiphene, absolute difference [AD] 15.3%, 95% CI, 8.6%-22.0%; metformin versus combination, AD 19.6%, 12.6%-26.6%; clomiphene versus combination, AD 4.3%, –4%-12.6%). Pregnancy rates were significantly lower in the metformin group (8.7%) compared with those in the clomiphene group (23.9%) and the combination group (31.1%) (metformin versus clomiphene, AD 15.2%, 8.3%-22.1%; metformin versus combination, AD 22.4%, 15%-29.8%; clomiphene versus combination, AD 7.2%, –1.3%-15.7%). Therefore, pregnancy and live birth was more than twice as likely when ovulation induction occurred with clomiphene than with metformin. In addition, no advantage to combination treatment is apparent. The authors determined that clomiphene alone should remain the primary infertility treatment for women with PCOS. This was the largest multicenter study of this group and the statistics clearly demonstrate the effect of metformin on live-birth and pregnancy rates. Overall, despite the differences between these three studies, the final outcomes are relatively similar.
CLINICAL BOTTOM LINE
Metformin showed no apparent advantage over clomiphene in rates of live births, and only one of the three studies demonstrated an advantage in pregnancy rates with metformin use. The use of metformin as an initial infertility treatment for women with PCOS is not currently supported by randomized research studies. The primary limitation of this body of research was the lack of a sufficient number of randomized studies utilizing pregnancy and/or live birth as outcome measures. Additional supportive evidence would be required to appropriately prescribe metformin as an infertility treatment for women with PCOS. However, PAs treating patients with PCOS may want to follow the ongoing research, as treatment options may broaden in the future. The PCOS, Metformin for Infertility with Clomiphene Trial was recently completed at the University of Auckland, New Zealand, and several trials are ongoing at the Magna Graecia University of Catanzaro, Italy. JAAPA
Meredith Wisniewski practices at The Allegheny Center for Digestive Health in Pittsburgh, Pennsylvania. Martha Petersen is on the faculty at Duquesne University Physician Assistant Program, Pittsburgh, Pennsylvania. They have indicated no relationships to disclose relating to the content of this article.
Charles DiMaggio, PhD, MPH, PA-C, department editor
REFERENCES
1. Ehrmann DA. Polycystic ovary syndrome. N Engl J Med. 2005;352(12):1223-1236.
2. Practice Committee of the American Society for Reproductive Medicine. Use of clomiphene citrate in women. Fertil Steril. 2006;86(5 suppl):S187-S193.
3. Palomba S, Orio F Jr, Falbo A, et al. Prospective parallel randomized, double-blind, double-dummy controlled clinical trial comparing clomiphene citrate and metformin as the first-line treatment for ovulation induction in nonobese anovulatory women with polycystic ovary syndrome. J Clin Endocrinol Metab. 2005;90(7):4068-4074.
4. Moll E, Bossuyt PMM, Korevaar JC, et al. Effect of clomifene citrate plus metformin and clomifene citrate plus placebo on induction of ovulation in women with newly diagnosed polycystic ovary syndrome: randomised double blind clinical trial. BMJ. 2006;332(7556):1485-1489.
5. Legro RS, Barnhart HX, Schlaff WD, et al; Cooperative Multicenter Reproductive Medicine Network. Clomiphene, metformin, or both for infertility in the polycystic ovary syndrome. N Engl J Med. 2007;356(6):551-566.