Use of Metformin for Ovulation Induction in
Women Who Have Polycystic Ovary Syndrome
With or Without Evidence of Insulin Resistance
Kimberly E. Liu, MD, Ivanna Viola Tataryn, MD, Margaret Sagle, MSc, MD
Department of Obstetrics and Gynaecology, University of Alberta, Edmonton AB
similaires chez les femmes qui présentaient des taux sériquesélevés d’insuline à jeun (48,8 %) et chez celles qui en présentaient
To determine whether women with polycystic ovary
des taux normaux (44,7 %). Les taux d’ovulation étaient également
syndrome (PCOS) and abnormal insulin levels treated with
semblables tant chez les femmes qui présentaient un rapport
metformin had different rates of ovulation and pregnancy from
glucose-insuline normal que chez celles qui présentaient un
women with PCOS and normal insulin levels.
rapport glucose-insuline anormal. Aucune différence n’a été
The outcomes of treatment with metformin in 146 infertile
constatée en ce qui concerne les taux cumulatifs de grossesse en
women with PCOS were analyzed using a retrospective cohort
fonction des taux d’insuline à jeun. Un taux d’insuline à jeun
study design. Baseline characteristics and initial blood work results
supérieur à 20 mU/l était en corrélation avec un rapport
were recorded. The follow-up period was three months, and the
Chez les femmes anovulatoires présentant un SOPK,
Of the 146 women with PCOS, one third had elevated
les taux d’insuline à jeun et les rapports glucose-insuline ne
fasting insulin levels. After treatment with metformin, cumulative
permettent pas de prédire la réaction ovulatoire à la metformine.
rates of ovulation were similar in women with elevated fasting
J Obstet Gynaecol Can 2006;28(7):595–599
serum insulin levels (48.8%) and those with normal levels(44.7%). Rates of ovulation were also similar in women with normaland abnormal glucose to insulin ratios. There was no difference in
cumulative pregnancy rates based on fasting insulin levels. Afasting insulin level above 20 mU/L correlated with an abnormalglucose to insulin ratio (98%).
Polycystic ovary syndrome (PCOS), first described in
1935 by Stein and Leventhal, has been associated with a
In anovulatory women with PCOS, fasting insulin levels
variety of clinical and laboratory findings, including
and glucose to insulin ratios do not predict the ovulatory responseto metformin.
hyperandrogenism, and insulin resistance. The 2003 Rotter-
dam criteria required two of the following characteristics
Déterminer si les femmes présentant un syndrome des
for the diagnosis of PCOS: oligo-ovulation or anovulation,
ovaires polykystiques (SOPK) et des taux d’insuline anormaux,
clinical and/or biochemical signs of hyperandrogenism,
et ayant été traitées à la metformine présentaient des tauxdifférents d’ovulation et de grossesse, par comparaison avec les
and polycystic ovaries with the exclusion of other etiologies
femmes présentant un SOPK et des taux d’insuline normaux.
Les issues de traitement à la metformine chez
146 femmes infertiles présentant un SOPK ont été analysées aumoyen d’une étude de cohorte rétrospective. Les caractéristiques
Many women with PCOS have been treated with
de départ et les résultats de l’hémogramme initial ont été
clomiphene for infertility, resulting in rates of ovulation of
consignés. La période de suivi a été de trois mois et le critèred’évaluation principal a été l’ovulation.
70% to 85%, and pregnancy rates of 33% to 45%.2–4 More
On a constaté des taux élevés d’insuline à jeun chez le
recently, the recognition of an association between insulin
tiers des 146 femmes présentant un SOPK. À la suite du
resistance and PCOS has led to the use of metformin, a
traitement à la metformine, les taux cumulatifs d’ovulation étaient
biguanide insulin-sensitizing agent, for ovulation induction.
Several large studies and a meta-analysis have shown that
Metformin, polycystic ovary syndrome, ovulation
metformin is effective for ovulation induction in women
with PCOS.5–7 Metformin is effective both in thin women
with PCOS8,9 and in women with PCOS who are resistant to
clomiphene,10–14 and treatment with metformin and
clomiphene has been shown to be more effective than use
JUILLET 2006 l
Table 1. Mean variables of patients with normal and abnormal fasting insulin levels
NS: not significant; BMI: body mass index; DHEA-S: dehydroepiandrosterone sulphate.
of clomiphene alone for inducing ovulation and subsequent
predict success. We completed a retrospective study of
pregnancy.6 Treatment with metformin is as effective as
women with PCOS who were treated with metformin for
laparoscopic ovarian drilling in inducing ovulation in
ovulation induction. The purpose of our study was to deter-
clomiphene-resistant women.13,14 There is some evidence
mine whether the rate of ovulation with metformin was
that use of metformin improves hyperandrogenism and
similar in patients with normal and abnormal fasting indices
Insulin resistance is more common in both thin and obese
MATERIALS AND METHODS
women with PCOS than in women without PCOS.18 There
For this retrospective cohort study, we reviewed the medi-
are currently no validated clinical tests for the diagnosis of
cal records of all patients who began treatment with
insulin resistance. Invasive tests such as the euglycemic
metformin for ovulation induction at the Regional Fertility
clamp, the gold standard in research settings, are not feasi-
and Women’s Endocrine Clinic (RFWEC) at the Royal
ble for clinical use. Other methods described include the
Alexandra Hospital between January 2000 and December
homeostasis model assessment of insulin resistance
2004. Inclusion criteria for the study were (1) a history of
(HOMA-IR) and the glucose to insulin ratio or area under
infertility, (2) a diagnosis of PCOS according to the 2003
the curve (AUC) following a 2-hour oral glucose tolerance
Rotterdam criteria, and (3) normal serum thyroid-
test.19–21 One small study has shown serum insulin levels
stimulating hormone (TSH) and prolactin levels. Exclusion
may be an independent predictor of treatment success with
criteria were (1) documented ovulatory cycles prior to treat-
metformin.22 Currently, screening for insulin resistance is
ment, (2) concurrent treatment with gonadotropins, and (3)
not routinely recommended for women with PCOS or prior
diabetes mellitus. Women who were undergoing concur-
to treatment with insulin-sensitizing agents. However,
rent treatment with clomiphene were included if they had
screening for the metabolic syndrome and impaired glucose
failed to respond to clomiphene at the same dosage prior to
tolerance may help identify a high-risk population and allow
a reduction in long-term health concerns.1
During this period, patients were seen by one of four repro-
In most studies, an unselected population of women with
ductive endocrinologists. Patients were managed according
PCOS has been treated effectively with metformin.23
to the individual physician’s practice with regard to
Metformin is generally effective in inducing ovulation in
metformin dosing, although most patients began treatment
50% of women with PCOS,6 but there is little information
at a dose of 500 mg twice daily. Initial blood work, including
about clinical or laboratory measurements that may help
fasting insulin and glucose levels and a hormone profile,
l JULY JOGC
Use of Metformin for Ovulation Induction in Women Who Have Polycystic Ovary Syndrome With or Without Evidence of Insulin Resistance
Table 2. Occurrence of ovulation and pregnancy in women with normal and abnormal fasting insulin levels
was completed prior to initiation of treatment. Patients
30 years, the overall rate of ovulation was 48%, and the
were reviewed three months after beginning metformin.
pregnancy rate was 12%. Of the 146 women, 124 had docu-
Fasting insulin levels were considered elevated if equal to or
mented fasting insulin levels prior to metformin treatment,
greater than 20 mU/L. Fasting glucose/insulin ratios were
and of these 30.6% had a fasting insulin level of 20 mU/L
defined as abnormal if less than 4.5 mg/10-4 U.20 All blood
or greater. The mean fasting insulin level was 18.7 mU/L.
work was done at provincial laboratories through a central-
Women with elevated fasting insulin levels had a higher
ized laboratory agency. Serum glucose levels were con-
body mass index (BMI), higher fasting glucose levels, and
verted from mmol/L to mg/dL prior to calculating the fast-
higher serum testosterone levels than women with normal
ing glucose/ insulin ratio. Baseline characteristics of
fasting insulin levels (Table 1). The rate of concurrent
patients with normal and abnormal fasting insulin and
clomiphene usage and dosage was similar in each group.
glucose/insulin ratios were recorded.
One-third (33.3%) of women had an abnormal fastingglucose/insulin ratio.
The primary outcome of the study was the occurrence ofovulation within three months of beginning metformin
There was no difference in ovulation or pregnancy rates in
therapy. Ovulation was confirmed by measuring a serum
women with normal or abnormal fasting insulin levels or
progesterone level of greater than 15 ng/mL. Progesterone
glucose to insulin ratios (Table 2). In women without docu-
levels were measured on cycle day 21, or seven days before
mented insulin levels prior to metformin treatment (22),
the anticipated next menses, then every seven days until day
50% ovulated, and one woman became pregnant.
35 if not shown to be ovulatory. The ovulation rate was cal-culated by dividing the number of women who ovulated by
In women with normal fasting insulin levels (< 20 mU/L),
the total number of women in that group. Assuming a rate
3.5% had an abnormal glucose/insulin ratio, but almost all
of ovulation in women with elevated insulin levels of 50%,
women with fasting insulin levels of 20 mU/L or greater
we calculated that 40 patients would be required in each
had an abnormal glucose/insulin ratio (98%).
group to detect a 35% difference between groups. All charts
The average BMI in our study population was 35 kg/m2; 62
were reviewed for baseline characteristics, diagnosis of
women had a BMI of greater than 35 kg/m2. In these
PCOS, and outcome by one investigator. Ethics approval
women, there was a trend towards a lower rate of ovulation
was received through the Health Research Ethics Board at
(41.9% vs. 55.2%, P
= 0.15) and a lower rate of pregnancy
(4.8% vs. 17.9%, P
= 0.03). Of the nine women with normal
Statistical analysis for ovulation and pregnancy outcomes
weight (BMI 18.5–25 kg/m2), one had an elevated fasting
was performed with two-tailed Fisher exact test and
chi-square test. Patient characteristics were analyzed with
More than half (56%) of the women in the study had an ele-
the Student t
test and Fisher exact test where appropriate.
vated serum testosterone level. There was a trend towards alower rate of ovulation in women with an elevated testoster-
one level that was not statistically significant (40% vs.
Between January 2000 and December 2004, 3732 women
= 0.08), but there was no difference in pregnancy
were seen at the RFWEC because of infertility. A total of
rates. More than one third of women with elevated testos-
146 women with PCOS who were treated with metformin
terone levels had elevated fasting insulin levels (38.3% vs.
met the study criteria. The average age of these women was
JUILLET 2006 l
PCOS is associated with peripheral insulin resistance(decreased
In this study, pre-treatment fasting insulin levels or fasting
hyperinsulinemia.18 Intrinsic alterations in insulin secretion,
glucose/insulin ratios did not predict the outcome of treat-
insulin receptors, and genetic susceptibility may contribute
ment with metformin for ovulation induction in women
to insulin resistance in thin women with PCOS.31–34 In
with PCOS. Metformin was effective in women with both
obese women with PCOS, insulin resistance is increased.
normal and abnormal fasting insulin levels; however,
There are no currently validated clinical tests for detecting
women with elevated fasting insulin levels were treated with
insulin resistance. Almost one third of our study population
higher doses of metformin than those with normal levels
had an elevated fasting insulin level or an abnormal fasting
(1232 mg vs. 1078 mg). The clinical significance of this dif-
glucose/insulin ratio. These screening tests are not rou-
tinely recommended because they do not predict clinicalevents.35 In our study, an elevated fasting insulin level corre-
Our study population was an unselected population of
lated with an abnormal glucose to insulin ratio. Women
women with PCOS, and most of them were overweight or
with PCOS are at increased risk of developing diabetes
obese. Current consensus guidelines state that insulin levels
mellitus type 2, dyslipidemia and long term cardiovascular
do not need to be measured as part of the diagnosis of
complications.1,36 Although PCOS guidelines do not rec-
PCOS, or prior to treatment with metformin for ovulation
ommend screening for insulin resistance for diagnosis or
induction.1 Our study confirms the lack of usefulness of
treatment,1screening women with PCOS for impaired glu-
fasting insulin levels in deciding whether to treat women
cose tolerance or the metabolic syndrome is recommended
if they have risk factors such as obesity and family history.
Recent evidence shows that lifestyle modification or treat-
Currently the standard for first-line therapy for inducing
ment with metformin can reduce the incidence of diabetes
ovulation in women with PCOS is clomiphene citrate, but
recent evidence suggests that metformin can also be effec-tive as first-line therapy.6,24 The suggestion to use
metformin is based on theories that hyperandrogenism inwomen with PCOS is caused by hyperinsulinemia.25,26
We found metformin to be effective for inducing ovulation
Metformin therapy reduces insulin levels and increases
within three months of initiation in approximately 50% of
insulin-like growth factor binding protein (IGFBP)-1.16,17,27
women with PCOS. Metformin therapy was effective in
Because IGF-1 stimulates ovarian androgen production,
patients with normal and abnormal insulin levels in our
metformin therapy may decrease intra-ovarian androgen
population; however, higher doses of metformin may be
levels through its effect of decreasing the IGF-1/IGFBP-1
needed in patients with elevated fasting insulin levels and
ratio.27 Hyperinsulinemia may also stimulate intra-ovarian
higher BMI. Further studies are needed in order to deter-
androgen production through stimulation of cytochrome
mine if there are clinical, biological, or laboratory markers
P450c17á, an enzyme involved in androgen synthesis.28
that can help individualize the best treatment option for
Metformin has been shown to decrease serum insulin lev-
els, P450c17á activity, and serum androgen levels in both
1. The Rotterdam ESHRE/ASRM-Sponsored PCOS Consensus Workshop
Group. Revised 2003 consensus on diagnostic criteria and long-term health
Approximately 20% of women with PCOS are resistant to
risks related to polycystic ovary syndrome. Fertil Steril 2004;81:19–25.
clomiphene, and hyperinsulinemia and obesity appear to
2. Female infertility: treatment options for complicated cases. The ESHRE
correlate with clomiphene failure.30 In our study popula-
Capri Workshop. European Society for Human Reproduction and
tion, there did not appear to be a correlation between a
Embryology. Hum Reprod 1997;12:1191–6.
woman’s BMI and her ovulatory response to treatment.
3. Infertility revisited: the state of the art today and tomorrow. The ESHRE
Capri Workshop. European Society for Human Reproduction and
Although both severe obesity and elevated serum testoster-
Embryology. Hum Reprod 1996;11:1779–807.
one levels predicted lower rates of ovulation, a type II error
4. Nasseri S, Ledger WL. Clomiphene citrate in the twenty-first century. Hum
in this conclusion cannot be excluded. The average daily
dosage of metformin used in our study was 1100 mg, and
5. Lord JM, Flight IH, Norman RJ. Metformin in polycystic ovary syndrome:
women with a higher BMI may require a higher dosage for a
systematic review and meta-analysis. BMJ 25–10–2003;327:951–3.
clinical effect. Given that women with PCOS and insulin
6. Kashyap S, Wells GA, Rosenwaks Z. Insulin-sensitizing agents as primary
therapy for patients with polycystic ovarian syndrome. Hum Reprod
resistance are more likely not to respond to treatment with
clomiphene, metformin may be a useful adjunct in this
7. Palomba S, Orio F, Jr., Falbo A, Manguso F, Russo T, Cascella T, et al.
Prospective parallel randomized, double-blind, double-dummy controlled
l JULY JOGC
Use of Metformin for Ovulation Induction in Women Who Have Polycystic Ovary Syndrome With or Without Evidence of Insulin Resistance
clinical trial comparing clomiphene citrate and metformin as the first-line
22. Moghetti P, Castello R, Negri C, Tosi F, Perrone F, Caputo M, et al.
treatment for ovulation induction in nonobese anovulatory women with
Metformin effects on clinical features, endocrine and metabolic profiles,
polycystic ovary syndrome. J Clin Endocrinol Metab 2005;90:4068–74.
and insulin sensitivity in polycystic ovary syndrome: a randomized,
8. Maciel GA, Soares Jnr JM, Alves da Motta EL, Abi HM, de Lima GR,
double-blind, placebo-controlled 6 month trial, followed by open, long term
Baracat EC. Nonobese women with polycystic ovary syndrome respond
clinical evaluation. J Clin Endocrinol Metab 2000;85:139–46.
better than obese women to treatment with metformin. Fertil Steril2004;81:355–60.
23. Costello MF, Eden JA. A systematic review of the reproductive system
effects of metformin in patients with polycystic ovary syndrome. Fertil Steril
9. Kumari AS, Haq A, Jayasundaram R, Abdel-Wareth LO, Al Haija SA,
Alvares M. Metformin monotherapy in lean women with polycystic ovarysyndrome. Reprod Biomed Online 2005;10:100–4.
24. Sills ES, Perloe M, Palermo GD. Correction of hyperinsulinemia in
oligoovulatory women with clomiphene-resistant polycystic ovary
10. Vandermolen DT, Ratts VS, Evans WS, Stovall DW, Kauma SW,
syndrome: a review of therapeutic rationale and reproductive outcomes. Eur
Nestler JE. Metformin increases the ovulatory rate and pregnancy rate from
J Obstet Gynecol Reprod Biol 2000;91:135–41.
clomiphene citrate in patients with polycystic ovary syndrome who areresistant to clomiphene citrate alone. Fertil Steril 2001;75:310–5.
25. Nardo LG, Rai R. Metformin therapy in the management of polycystic
ovary syndrome: endocrine, metabolic and reproductive effects. Gynecol
11. Kocak M, Caliskan E, Simsir C, Haberal A. Metformin therapy improves
ovulatory rates, cervical scores, and pregnancy rates in clomiphenecitrate-resistant women with polycystic ovary syndrome. Fertil Steril
26. La Marca A, Morgante G, Paglia T, Ciotta L, Cianci A, De L, V. Effects of
metformin on adrenal steroidogenesis in women with polycystic ovarysyndrome. Fertil Steril 1999;72:985–9.
12. George SS, George K, Irwin C, Job V, Selvakumar R, Jeyaseelan V, et al.
Sequential treatment of metformin and clomiphene citrate
27. De Leo V, La Marca A, Orvieto R, and Morgante G. Effect of metformin
in clomiphene-resistant women with polycystic ovary syndrome:
on insulin-like growth factor (IGF) I and IGF-binding protein I in
a randomized, controlled trial. Hum Reprod 2003;18:299–304.
polycystic ovary syndrome. J Clin Endocrinol Metab 2000;85:1598–600.
13. Palomba S, Orio F, Jr., Nardo LG, Falbo A, Russo T, Corea D, et al.
28. Nestler JE, Jakubowicz DJ. Decreases in ovarian cytochrome P450c17 alpha
Metformin administration versus laparoscopic ovarian diathermy
activity and serum free testosterone after reduction of insulin secretion in
in clomiphene citrate-resistant women with polycystic ovary syndrome:
polycystic ovary syndrome. N Engl J Med 29–8–1996;335:617–23.
a prospective parallel randomized double-blind placebo-controlled trial.
29. Nestler JE, Jakubowicz DJ. Lean women with polycystic ovary syndrome
J Clin Endocrinol Metab 2004;89:4801–9.
respond to insulin reduction with decreases in ovarian P450c17 alpha
14. Malkawi HY, Qublan HS, Hamaideh AH. Medical vs. surgical treatment for
activity and serum androgens. J Clin Endocrinol Metab 1997;82:4075–9.
clomiphene citrate-resistant women with polycystic ovary syndrome.
30. Ben Haroush A, Yogev Y, Fisch B. Insulin resistance and metformin in
polycystic ovary syndrome. Eur J Obstet Gynecol Reprod Biol
15. Lord JM, Flight IH, Norman RJ. Insulin-sensitising drugs (metformin,
troglitazone, rosiglitazone, pioglitazone, D-chiro-inositol) for polycystic
31. Ehrmann DA, Breda E, Cavaghan MK, Bajramovic S, Imperial J,
ovary syndrome. Cochrane Database Syst Rev 2003;CD003053.
Toffolo G, et al. Insulin secretory responses to rising and falling glucose
16. Nestler JE, Jakubowicz DJ, Evans WS, Pasquali R. Effects of metformin on
concentrations are delayed in subjects with impaired glucose tolerance.
spontaneous and clomiphene-induced ovulation in the polycystic ovary
32. El Mkadem SA, Lautier C, Macari F, Molinari N, Lefebvre P, Renard E,
17. Sahin Y, Yirmibes U, Kelestimur F, Aygen E. The effects of metformin on
et al. Role of allelic variants Gly972Arg of IRS-1 and Gly1057Asp of IRS-2
insulin resistance, clomiphene-induced ovulation and pregnancy rates in
in moderate-to-severe insulin resistance of women with polycystic ovary
women with polycystic ovary syndrome. Eur J Obstet Gynecol Reprod Biol
syndrome. Diabetes 2001;50:2164–68.
33. Dunaif A, Xia J, Book CB, Schenker E, Tang Z. Excessive insulin receptor
18. Dunaif A, Segal KR, Futterweit W, Dobrjansky A. Profound peripheral
serine phosphorylation in cultured fibroblasts and in skeletal muscle.
insulin resistance, independent of obesity, in polycystic ovary syndrome.
A potential mechanism for insulin resistance in the polycystic ovary
syndrome. J Clin Invest 1995;96:801–10.
19. Deugarte CM, Bartolucci AA, Azziz R. Prevalence of insulin resistance in
34. Ciaraldi TP, el Roeiy A, Madar Z, Reichart D, Olefsky JM, Yen SS. Cellular
the polycystic ovary syndrome using the homeostasis model assessment.
mechanisms of insulin resistance in polycystic ovarian syndrome. J Clin
35. Consensus Development Conference on Insulin Resistance. 5–6 November
20. Legro RS, Finegood D, Dunaif A. A fasting glucose to insulin ratio is a
1997. American Diabetes Association. Diabetes Care 1998;21:310–4.
useful measure of insulin sensitivity in women with polycystic ovarysyndrome. J Clin Endocrinol Metab 1998;83:2694–8.
36. Sheehan MT. Polycystic Ovarian Syndrome: Diagnosis and Management.
21. Ciampelli M, Leoni F, Cucinelli F, Mancuso S, Panunzi S, De Gaetano A, et al.
Assessment of insulin sensitivity from measurements in the fasting state and
37. Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM,
during an oral glucose tolerance test in polycystic ovary syndrome and
Walker EA, et al. Reduction in the incidence of type 2 diabetes with lifestyle
menopausal patients. J Clin Endocrinol Metab 2005;90:1398–406.
intervention or metformin. N Engl J Med 7–2–2002;346:393–403.
JUILLET 2006 l
Environmental, Health & Safety Department Policy and Procedures for Using Controlled Substances in Non-Clinical Educational and Research Activities This internal policy regulates the use of controlled substances, controlled substance analogues, chemical precursors and certain chemical laboratory apparatus used in non-clinical educational training and research activities
Medication Use in Pregnancy The following medications have been taken during pregnancy and have not been shown to cause birth defects. Even so, we recommend not using any medication unless necessary. If you take a medication routinely for a medical problem and are unsure about the medication, please contact our office before discontinuing that medication. Take all medications according to