Professional Documents
Culture Documents
clinical therapeutics
This Journal feature begins with a case vignette that includes a therapeutic recommendation. A discussion
of the clinical problem and the mechanism of benefit of this form of therapy follows. Major clinical studies,
the clinical use of this therapy, and potential adverse effects are reviewed. Relevant formal guidelines,
if they exist, are presented. The article ends with the author’s clinical recommendations.
A 23-year-old woman with known polycystic ovary syndrome visits her family physi-
cian. She has taken oral contraceptive pills in the past but did not tolerate them and is
not currently receiving any treatment. She has three or four menstrual periods per
year and is not interested in becoming pregnant now, but she will be getting married
in a year. She has heard that the polycystic ovary syndrome is associated with diabetes
and is concerned because both her mother and father have type 2 diabetes. Her body-
mass index (the weight in kilograms divided by the square of the height in meters) is
32, her waist circumference is 38 in. (96.5 cm), her serum total testosterone level is
elevated at 0.9 ng per milliliter (90 ng per deciliter, or 2.9 nmol per liter), her plasma
high-density lipoprotein cholesterol level is 35 mg per deciliter (0.9 mmol per liter),
and her triglyceride level is 190 mg per deciliter (2.1 mmol per liter). Her serum
glucose level 2 hours after the ingestion of 75 g of dextrose is 138 mg per deciliter
(7.7 mmol per liter). The physician wonders whether treatment with metformin
would be beneficial and refers the patient to an endocrinologist.
The polycystic ovary syndrome is a clinical diagnosis characterized by the presence From the Departments of Internal Medi-
of two or more of the following features: chronic oligo-ovulation or anovulation, cine, Obstetrics and Gynecology, and Phar-
macology and Toxicology, Virginia Com-
androgen excess, and polycystic ovaries.1 It affects 5 to 10% of women of childbear- monwealth University, Richmond. Address
ing age2,3 and is the most common cause of anovulatory infertility in developed reprint requests to Dr. Nestler at the Di-
countries. Common clinical manifestations include menstrual irregularities and vision of Endocrinology and Metabolism,
Medical College of Virginia, PO Box
signs of androgen excess such as hirsutism, acne, and alopecia. 980111, Richmond, VA 23298-0111, or at
The polycystic ovary syndrome is associated with important metabolic derange- jnestler@mcvh-vcu.edu.
ments. The prevalence of type 2 diabetes in the United States is 10 times as high
N Engl J Med 2008;358:47-54.
among young women with the polycystic ovary syndrome as among normal Copyright © 2008 Massachusetts Medical Society.
women,4,5 and impaired glucose tolerance or overt type 2 diabetes develops by the
age of 30 years in 30 to 50% of obese women with the polycystic ovary syndrome.4‑6
The prevalence of the metabolic syndrome is two to three times as high among
women with the polycystic ovary syndrome as among normal women matched for
age and body-mass index, and 20% of women with the polycystic ovary syndrome
who are younger than 20 years of age have the metabolic syndrome.7 Although
outcome data specifically for women with the polycystic ovary syndrome are lack-
ing, the risk of fatal myocardial infarction is twice as high among women with
severe oligomenorrhea, most of whom would be expected to have the polycystic
ovary syndrome, as among women with eumenorrhea.8
Pancreas Increased
Insulin insulin release
Genetic
factors
Insulin
resistance Skeletal
muscle
Pituitary gland
Hyperglycemia
Increased
free fatty acids,
cytokines, PAI-1
Polycystic
Release of free ovary
fatty acids
Increased
adipose tissue
Increased
Decreased sex androgens
hormone-binding globulin
production
(26.8% and 22.5%, respectively; P = 0.31). In that sutism, male pattern baldness, and acne) and
study, clomiphene was more effective than met- restoring regular menses, thereby preventing endo
formin in inducing ovulation in the short term metrial hyperplasia.
and producing a live birth. However, given the metabolic derangements
With regard to diabetes, two major random- associated with the polycystic ovary syndrome, it
ized clinical trials, the Indian Diabetes Preven- seems prudent and appropriate to plan long-term
tion Programme (IDPP-1)45 and the U.S. Diabe- therapy that addresses not only management of
tes Prevention Program (DPP),46 have shown that the consequences of androgen excess and anovu-
the use of metformin decreases the relative risk lation but also the new goals of ameliorating in-
for progression to type 2 diabetes (by 26% and sulin resistance and reducing the risks of type 2
31%, respectively) among patients with impaired diabetes and cardiovascular disease. The effect of
glucose tolerance at baseline. Whether this effect estrogen–progestin contraceptive agents on glu-
of metformin truly represented the prevention of cose tolerance is controversial. Limited evidence
progression to diabetes, rather than simply the from controlled, short-term studies suggests that
masking of progression by means of lowering the use of oral contraceptives aggravates insulin
blood glucose levels, remains controversial. How- resistance and worsens glucose tolerance in wom-
ever, after the discontinuation of metformin in en with the polycystic ovary syndrome.48 The use
the DPP, diabetes developed in fewer subjects of estrogen–progestin contraceptives is associat
than would have been expected if masking had ed with a twofold increase in the relative risk of
been the sole effect.47 To my knowledge, no ran- cardiovascular arterial events in the general fe-
domized clinical trial has assessed the effect of male population49; the risk among women with
metformin on the progression to type 2 diabetes the polycystic ovary syndrome in particular is
specifically in patients with the polycystic ovary unknown.
syndrome. In an uncontrolled, retrospective study Metformin improves insulin sensitivity and,
of 50 women with the polycystic ovary syndrome as noted earlier, has been shown to retard or
treated with metformin for an average of 43 prevent progression to type 2 diabetes in patients
months at an academic medical center, there was with impaired glucose tolerance. Although met-
no progression to type 2 diabetes, even though formin has not been specifically shown to reduce
11 women (22.0%) had impaired glucose toler- the risk of cardiovascular events in patients with
ance at baseline.39 The annual conversion rate the polycystic ovary syndrome, the available mech
from normal glucose tolerance to impaired glu- anistic and clinical evidence support the use of
cose tolerance was only 1.4%, as compared with metformin as a protective measure against the
16 to 19% reported in the literature4,6 for women adverse cardiovascular effects of insulin resis-
with the polycystic ovary syndrome who are not tance and insulin excess. In addition, metformin
taking metformin. may decrease circulating androgen levels and may
improve ovulation and menstrual cyclicity, thus
Cl inic a l Use addressing the traditional goals of long-term
treatment. For these reasons, although metfor-
The approach to the management of the polycystic min is not approved by the Food and Drug Ad-
ovary syndrome depends in part on the patient’s ministration for the treatment of the polycystic
and physician’s principal therapeutic objectives. ovary syndrome, the drug is commonly used for
For some women, infertility is the principal issue. this purpose.
Such patients are frequently treated in the short To minimize side effects, metformin therapy
term with clomiphene, to induce ovulation. When is initiated at a low dose taken with meals, and
fertility is not a concern, an estrogen–progestin the dose is then progressively increased. My prac-
contraceptive, with or without an antiandrogen tice is to have patients take 500 mg of metformin
such as spironolactone, has been the mainstay of once daily with the largest meal, usually dinner,
long-term therapy. This approach is effective in for 1 week; then increase the dose to 500 mg
achieving the traditional treatment goals in the twice daily, with breakfast and dinner, for 1 week;
polycystic ovary syndrome, which include ame- increase the dose to 500 mg with breakfast and
liorating the effects of androgen excess (e.g., hir- 1000 mg with dinner, for 1 week; and finally,
increase the dose to 1000 mg twice daily, with Women desiring contraception could be given an
breakfast and dinner. There is no dose-ranging oral contraceptive agent while continuing to take
study of metformin in the polycystic ovary syn- metformin. In cases in which hirsutism remains
drome, but a dose-ranging study of patients with troublesome, an oral contraceptive agent, anti-
diabetes, using the glycated hemoglobin level as androgen, or both could be added to metformin.
the outcome measure, suggested that a dose of
2000 mg daily is optimal.50 A dv er se Effec t s
Metformin should not be used in women
with renal impairment (a serum creatinine level Lactic acidosis has been reported with the use of
>1.4 mg per deciliter [124 μmol per liter]), he- metformin, but this complication is rare (0.3 epi-
patic dysfunction, severe congestive heart failure, sode per 10,000 patient-years) in otherwise healthy
or a history of alcohol abuse. Given the young patients and is confined primarily to patients who
age of women with the polycystic ovary syn- should not have received the drug, because they had
drome, these contraindications are rarely an issue. underlying renal or hepatic disease.
Repeat testing during metformin treatment is not The main limiting side effect of metformin,
indicated unless an illness or condition (e.g., de- affecting 10 to 25% of patients, is gastrointesti-
hydration) that might affect renal or hepatic func- nal distress, namely nausea and diarrhea. If the
tion develops. nausea or diarrhea occurs at a given dose, that
When metformin is prescribed, advice about a dose is either maintained or decreased by 500 mg
weight-loss diet and a scheduled exercise routine per day for 2 to 4 weeks until the symptoms
should also be given. Such interventions are abate. Fortunately, the gastrointestinal side ef-
beneficial in preventing diabetes.45,46 In addition, fects of metformin are usually transient; however,
weight loss increases the likelihood of resuming in a minority of cases, gastrointestinal distress
ovulation, most likely as a result of improved may require the discontinuation of metformin.
insulin sensitivity.29,51 Metformin can cause malabsorption of vitamin
The patient is asked to maintain a menstrual B12 in some patients receiving long-term therapy.
diary, cautioned that fertility may be rapidly es- In one analysis, risk factors for the development
tablished, and advised to use a barrier method of this adverse effect included both the daily dose
of contraception. Oral contraceptive agents and and duration of metformin therapy as well as
antiandrogens are not administered at the initial age.52 Although the likelihood of clinical defi-
visit, since they might affect menstruation or ciency of vitamin B12 appears to be low, patients
serum androgen levels and confound the assess- should be monitored for signs and symptoms.
ment of the efficacy of metformin. Topical eflor Metformin is a category B drug, and no tera-
nithine can be prescribed for the treatment of togenic effects have been found in animal mod-
facial hirsutism. els. It was administered in South Africa to a
Follow-up visits are scheduled at 3 and limited number of women with type 2 diabetes
6 months. Menstrual cyclicity is reviewed and or gestational diabetes, throughout their preg-
serum total testosterone is determined at each nancies, and no teratogenic effects or adverse
visit. If an improvement in menstrual cyclicity is fetal outcomes were reported.31
noted, it is important to document whether the
menses are ovulatory. This can be accomplished A r e a s of Uncer ta in t y
by measuring the serum progesterone level 7 days
before the next onset of menses; a serum pro- Although metformin is increasingly used to treat
gesterone level of more than 4.0 ng per milliliter patients with the polycystic ovary syndrome, ther-
(12.7 nmol per liter) is consistent with the pres- apy is guided in part by the results of random-
ence of the luteal phase and ovulation. ized, controlled trials in populations without the
After 6 to 9 months of treatment, the efficacy polycystic ovary syndrome, which showed that
of metformin is assessed. If menstrual cyclicity diabetes was prevented. Similar randomized, con-
and ovulation are satisfactorily improved, further trolled trials specifically involving patients with
treatment is individualized. For some women, the polycystic ovary syndrome are needed. Strate-
treatment with metformin alone might suffice. gies for long-term metformin therapy in patients
References
1. The Rotterdam ESHRE/ASRM-Spon- HF. A prospective study of the prevalence women of the southeastern United States:
sored PCOS Consensus Workshop Group. of the polycystic ovary syndrome in un- a prospective study. J Clin Endocrinol
Revised 2003 consensus on diagnostic selected Caucasian women from Spain. Metab 1998;83:3078-82.
criteria and long-term health risks related J Clin Endocrinol Metab 2000;85:2434-8. 4. Ehrmann DA, Barnes RB, Rosenfield
to polycystic ovary syndrome (PCOS). 3. Knochenhauer ES, Key TJ, Kahsar- RL, Cavaghan MK, Imperial J. Prevalence
Hum Reprod 2004;19:41-7. Miller M, Waggoner W, Boots LR, Azziz R. of impaired glucose tolerance and diabe-
2. Asunción M, Calvo RM, San Millán Prevalence of the polycystic ovary syn- tes in women with polycystic ovary syn-
JL, Sancho J, Avila S, Escobar-Morreale drome in unselected black and white drome. Diabetes Care 1999;22:141-6.
5. Legro RS, Kunselman AR, Dodson women with polycystic ovary syndrome. 31. Cheang KI, Sharma ST, Nestler JE. Is
WC, Dunaif A. Prevalence and predictors J Clin Endocrinol Metab 2004;89:5454-61. metformin a primary ovulatory agent in
of risk for type 2 diabetes mellitus and 18. Paradisi G, Steinberg HO, Hempfling patients with polycystic ovary syndrome?
impaired glucose tolerance in polycystic A, et al. Polycystic ovary syndrome is as- Gynecol Endocrinol 2006;22:595-604.
ovary syndrome: a prospective, controlled sociated with endothelial dysfunction. 32. Azziz R, Ehrmann D, Legro RS, et al.
study in 254 affected women. J Clin Endo- Circulation 2001;103:1410-5. Troglitazone improves ovulation and hir-
crinol Metab 1999;84:165-9. 19. Kelly CJ, Speirs A, Gould GW, Petrie sutism in the polycystic ovary syndrome:
6. Legro RS, Gnatuk CL, Kunselman AR, JR, Lyall H, Connell JM. Altered vascular a multicenter, double blind, placebo-con-
Dunaif A. Changes in glucose tolerance function in young women with polycystic trolled trial. J Clin Endocrinol Metab
over time in women with polycystic ovary ovary syndrome. J Clin Endocrinol Metab 2001;86:1626-32.
syndrome: a controlled study. J Clin Endo- 2002;87:742-6. 33. Rautio K, Tapanainen JS, Ruokonen
crinol Metab 2005;90:3236-42. 20. Nestler JE, Barlascini CO, Matt DW, et A, Morin-Papunen LC. Endocrine and
7. Apridonidze T, Essah PA, Iuorno MJ, al. Suppression of serum insulin by di- metabolic effects of rosiglitazone in over-
Nestler JE. Prevalence and characteristics azoxide reduces serum testosterone levels weight women with PCOS: a randomized
of the metabolic syndrome in women in obese women with polycystic ovary placebo-controlled study. Hum Reprod
with polycystic ovary syndrome. J Clin En- syndrome. J Clin Endocrinol Metab 1989; 2006;21:1400-7.
docrinol Metab 2005;90:1929-35. 68:1027-32. 34. Brettenthaler N, De Geyter C, Huber
8. Solomon CG, Hu FB, Dunaif A, et al. 21. Nestler JE, Jakubowicz DJ, de Vargas PR, Keller U. Effect of the insulin sensi-
Menstrual cycle irregularity and risk for AF, Brik C, Quintero N, Medina F. Insulin tizer pioglitazone on insulin resistance,
future cardiovascular disease. J Clin En- stimulates testosterone biosynthesis by hyperandrogenism, and ovulatory dys-
docrinol Metab 2002;87:2013-7. human thecal cells from women with poly- function in women with polycystic ovary
9. Chang RJ, Nakamura RM, Judd HL, cystic ovary syndrome by activating its syndrome. J Clin Endocrinol Metab 2004;
Kaplan SA. Insulin resistance in nonobese own receptor and using inositolglycan me- 89:3835-40.
patients with polycystic ovarian disease. diators as the signal transduction system. 35. Penna IA, Canella PR, Reis RM, Silva
J Clin Endocrinol Metab 1983;57:356-9. J Clin Endocrinol Metab 1998;83:2001-5. de Sá MF, Ferriani RA. Acarbose in obese
10. Dunaif A, Graf M, Mandeli J, Laumas 22. Poretsky L. On the paradox of insulin- patients with polycystic ovarian syn-
V, Dobrjansky A. Characterization of induced hyperandrogenism in insulin- drome: a double-blind, randomized, pla-
groups of hyperandrogenic women with resistant states. Endocr Rev 1991;12:3-13. cebo-controlled study. Hum Reprod 2005;
acanthosis nigricans, impaired glucose 23. Baillargeon JP, Jakubowicz DJ, Iuorno 20:2396-401.
tolerance, and/or hyperinsulinemia. J Clin MJ, Jakubowicz S, Nestler JE. Effects of 36. Bailey CJ, Turner RC. Metformin. N Engl
Endocrinol Metab 1987;65:499-507. metformin and rosiglitazone, alone and J Med 1996;334:574-9.
11. Dunaif A, Segal KR, Futterweit W, in combination, in nonobese women with 37. Diamanti-Kandarakis E, Kouli C, Tsi-
Dobrjansky A. Profound peripheral insu- polycystic ovary syndrome and normal anateli T, Bergiele A. Therapeutic effects
lin resistance, independent of obesity, in indices of insulin sensitivity. Fertil Steril of metformin on insulin resistance and
polycystic ovary syndrome. Diabetes 1989; 2004;82:893-902. hyperandrogenism in polycystic ovary syn-
38:1165-74. 24. Baillargeon JP, Nestler JE. Polycystic drome. Eur J Endocrinol 1998;138:269-74.
12. Lillioja S, Mott DM, Spraul M, et al. ovary syndrome: a syndrome of ovarian 38. Harborne L, Fleming R, Lyall H, Sat-
Insulin resistance and insulin secretory hypersensitivity to insulin? J Clin Endo- tar N, Norman J. Metformin or antiandro-
dysfunction as precursors of non-insulin- crinol Metab 2006;91:22-4. gen in the treatment of hirsutism in poly-
dependent diabetes mellitus: prospective 25. Baillargeon JP, Carpentier A. Role of cystic ovary syndrome. J Clin Endocrinol
studies of Pima Indians. N Engl J Med insulin in the hyperandrogenemia of lean Metab 2003;88:4116-23.
1993;329:1988-92. women with polycystic ovary syndrome 39. Sharma ST, Wickham EP III, Nestler
13. Sharma ST, Nestler JE. Prevention of and normal insulin sensitivity. Fertil Ster- JE. Changes in glucose tolerance with
diabetes and cardiovascular disease in il 2007;88:886-93. metformin treatment in polycystic ovary
women with PCOS: treatment with insu- 26. Nestler JE, Powers LP, Matt DW, et al. syndrome: a retrospective analysis. En-
lin sensitizers. Best Pract Res Clin Endo- A direct effect of hyperinsulinemia on se- docr Pract 2007;13:373-9.
crinol Metab 2006;20:245-60. rum sex hormone-binding globulin levels 40. Nestler JE, Jakubowicz DJ. Decreases
14. Talbott EO, Guzick DS, Sutton-Tyrrell in obese women with the polycystic ovary in ovarian cytochrome P450c17α activity
K, et al. Evidence for association between syndrome. J Clin Endocrinol Metab 1991; and serum free testosterone after reduc-
polycystic ovary syndrome and premature 72:83-9. tion in insulin secretion in polycystic
carotid atherosclerosis in middle-aged 27. Nestler JE. Role of hyperinsulinemia ovary syndrome. N Engl J Med 1996;335:
women. Arterioscler Thromb Vasc Biol in the pathogenesis of the polycystic ova- 617-23.
2000;20:2414-21. ry syndrome, and its clinical implications. 41. Nestler JE, Jakubowicz DJ, Evans WS,
15. Christian RC, Dumesic DA, Behren- Semin Reprod Endocrinol 1997;15:111- Pasquali R. Effects of metformin on spon-
beck T, Oberg AL, Sheedy PF II, Fitzpat- 22. taneous and clomiphene-induced ovula-
rick LA. Prevalence and predictors of cor- 28. Ciotta L, De Leo V, Galvani F, La Mar- tion in the polycystic ovary syndrome.
onary artery calcification in women with ca A, Cianci A. Endocrine and metabolic N Engl J Med 1998;338:1876-80.
polycystic ovary syndrome. J Clin Endo- effects of octreotide, a somatostatin ana- 42. Legro RS, Barnhart HX, Schlaff WD,
crinol Metab 2003;88:2562-8. logue, in lean PCOS patients with either et al. Clomiphene, metformin, or both for
16. Orio F Jr, Palomba S, Cascella T, et al. hyperinsulinaemia or lean normoinsu- infertility in the polycystic ovary syn-
Early impairment of endothelial structure linaemia. Hum Reprod 1999;14:2951-8. drome. N Engl J Med 2007;356:551-66.
and function in young normal-weight 29. Norman RJ, Homan G, Moran L, 43. Moll E, Bossuyt PM, Korevaar JC,
women with polycystic ovary syndrome. Noakes M. Lifestyle choices, diet, and in- Lambalk CB, van der Veen F. Effect of clo-
J Clin Endocrinol Metab 2004;89:4588-93. sulin sensitizers in polycystic ovary syn- mifene citrate plus metformin and clomi-
17. Talbott EO, Zborowski JV, Rager JR, drome. Endocrine 2006;30:35-43. fene citrate plus placebo on induction of
Boudreaux MY, Edmundowicz DA, Guzick 30. Lord JM, Flight IH, Norman RJ. Met- ovulation in women with newly diag-
DS. Evidence for an association between formin in polycystic ovary syndrome: sys- nosed polycystic ovary syndrome: ran-
metabolic cardiovascular syndrome and tematic review and meta-analysis. BMJ domised double blind clinical trial. BMJ
coronary and aortic calcification among 2003;327:951-3. 2006;332:1485-90.
44. Palomba S, Orio F Jr, Falbo A, et al. JP, Iuorno MJ, Jakubowicz DJ, Nestler JE. cose intolerance in polycystic ovary syn-
Prospective parallel randomized, double- A modern medical quandary: polycystic drome: a position statement of the Andro-
blind, double-dummy controlled clinical ovary syndrome, insulin resistance, and gen Excess Society. J Clin Endocrinol
trial comparing clomiphene citrate and oral contraceptive pills. J Clin Endocrinol Metab 2007;92:4546-56.
metformin as the first-line treatment for Metab 2003;88:1927-32. 54. Polycystic Ovary Syndrome Writing
ovulation induction in nonobese anovula- 49. Baillargeon JP, McClish DK, Essah Committee. American Association of Clin-
tory women with polycystic ovary syn- PA, Nestler JE. Association between the ical Endocrinologists position statement
drome. J Clin Endocrinol Metab 2005;90: current use of low-dose oral contracep- on metabolic and cardiovascular conse-
4068-74. tives and cardiovascular arterial disease: quences of polycystic ovary syndrome.
45. Ramachandran A, Snehalatha C, Mary a meta-analysis. J Clin Endocrinol Metab Endocr Pract 2005;11:126-34.
S, Mukesh B, Bhaskar AD, Vijay V. The 2005;90:3863-70. 55. Expert Panel on Detection, Evaluation,
Indian Diabetes Prevention Programme 50. Garber AJ, Duncan TG, Goodman and Treatment of High Blood Cholesterol
shows that lifestyle modification and met- AM, Mills DJ, Rohlf JL. Efficacy of met- in Adults. Executive summary of the third
formin prevent type 2 diabetes in Asian formin in type II diabetes: results of a report of the National Cholesterol Educa-
Indian subjects with impaired glucose double-blind, placebo-controlled, dose- tion Program (NCEP) Expert Panel on De-
tolerance (IDPP-1). Diabetologia 2006;49: response trial. Am J Med 1997;103:491- tection, Evaluation, and Treatment of High
289-97. 7. Blood Cholesterol in Adults (Adult Treat-
46. Diabetes Prevention Program Research 51. Norman RJ, Davies MJ, Lord J, Moran ment Panel III). JAMA 2001;285:2486-
Group. Reduction in the incidence of type LJ. The role of lifestyle modification in 97.
2 diabetes with lifestyle intervention or polycystic ovary syndrome. Trends Endo- 56. Grundy SM, Cleeman JI, Daniels SR,
metformin. N Engl J Med 2002;346:393- crinol Metab 2002;13:251-7. et al. Diagnosis and management of the
403. 52. Ting RZ, Szeto CC, Chan MH, Chow metabolic syndrome: an American Heart
47. Idem. Effects of withdrawal from met- KM. Risk factors of vitamin B12 deficien- Association/National Heart, Lung, and
formin on the development of diabetes in cy in patients receiving metformin. Arch Blood Institute scientific statement. Circu-
the Diabetes Prevention Program. Diabe- Intern Med 2006;166:1975-9. lation 2005;112:2735-52. [Erratum, Circu-
tes Care 2003;26:977-80. 53. Salley KES, Wickham EP, Cheang KI, lation 2005;112(17):e297-e298.]
48. Diamanti-Kandarakis E, Baillargeon Essah PA, Karjane NW, Nestler JE. Glu- Copyright © 2008 Massachusetts Medical Society.