Extended-Release Metformin Does Not Reduce the Clomiphene Citrate Dose Required to Induce Ovulation in Polycystic Ovary Syndrome

Abstract

Context: When used for ovulation induction, higher doses of clomiphene may lead to antiestrogenic side effects that reduce fecundity. It has been suggested that metformin in combination with clomiphene can restore ovulation to some clomiphene-resistant anovulators with polycystic ovary syndrome (PCOS).

Objective: Our objective was to determine if cotreatment with extended-release metformin (metformin XR) can lower the threshold dose of clomiphene needed to induce ovulation in women with PCOS.

Design: A secondary analysis of data from the National Institute of Child Health and Human Development Cooperative Multicenter Reproductive Medicine Network prospective, double-blind, placebo-controlled multicenter clinical trial, Pregnancy in Polycystic Ovary Syndrome, was performed.

Setting: Study volunteers at multiple academic medical centers were included.

Participants: Women with PCOS and elevated serum testosterone who were randomized to clomiphene alone or with metformin (n = 209 in each group) were included in the study.

Interventions: Clomiphene citrate, 50 mg daily for 5 d, was increased to 100 and 150 mg in subsequent cycles if ovulation was not achieved; half also received metformin XR, 1000 mg twice daily. Treatment was for up to 30 wk or six cycles, or until first pregnancy.

Main Outcome Measures: Ovulation was confirmed by a serum progesterone more than or equal to 5 ng/ml, drawn prospectively every 1–2 wk.

Results: The overall prevalence of at least one ovulation after clomiphene was 75 and 83% (P = 0.04) for the clomiphene-only and clomiphene plus metformin groups, respectively. Using available data from 314 ovulators, the frequency distribution of the lowest clomiphene dose (50, 100, or 150 mg daily) resulting in ovulation was indistinguishable between the two treatment groups.

Conclusion: Metformin XR does not reduce the lowest dose of clomiphene that induces ovulation in women with PCOS.

In anovulatory women with polycystic ovary syndrome seeking fertility, metformin does not decrease the clomiphene dose needed to induce ovulation.

Polycystic ovary syndrome (PCOS) affects 7–8% of reproductive-aged women. Classically defined as including both anovulation and hyperandrogenism of ovarian origin (1), PCOS is often associated with insulin resistance and insulin excess. These metabolic abnormalities have been associated with comorbidities such as obesity, prediabetes, and type 2 diabetes (2).

Clomiphene citrate is the conventional first-line agent to induce ovulation in PCOS when pregnancy is desired. For the roughly 25% of women with PCOS resistant to clomiphene ovulation induction (3), the conventional next step is either exogenous gonadotropin injections, or ovarian diathermy or drilling. Both of these modalities bring significant risks as well as increased cost, making an alternative approach desirable. Metformin has been used to promote ovulation in PCOS with considerable, though variable, success (4). Insulin sensitization is an attractive strategy for clomiphene-resistant women, who as a group have greater insulin excess than those who ovulate after clomiphene (5). Whether or not metformin is a similarly attractive therapy for unselected women with PCOS remains a subject of debate.

Not all women with PCOS who ovulate after clomiphene conceive readily, in part because clomiphene may have unfavorable antiestrogenic effects on cervical mucus and endometrium (3). Because these clomiphene effects appear to be dose dependent, a strategy that can induce ovulation at lower clomiphene doses might improve live birth rates. Given reports of successful ovulation in previously clomiphene-resistant women after combined clomiphene and metformin treatment (6), it is reasonable to hypothesize that metformin cotreatment can reduce the threshold clomiphene dose for ovulation in an unselected PCOS population. To test this hypothesis, we determined the lowest clomiphene dose that led to ovulation in a large cohort of women with PCOS seeking fertility who were enrolled in a study of extended-release metformin (metformin XR), taken along with clomiphene in doses that were escalated in standard fashion.

Subjects and Methods

The Pregnancy in Polycystic Ovary Syndrome (PPCOS) study was a multicenter, randomized, double-blind, placebo-controlled trial of metformin XR and clomiphene, with the primary endpoint of live birth rate. The study and its primary outcomes are described in detail elsewhere (7). In brief, women with PCOS by National Institute of Child Health and Human Development criteria (1) (oligo-ovulation and hyperandrogenism, defined by an elevated serum testosterone) were randomized through a central interactive voice system with equal likelihood to treatment with metformin XR, clomiphene citrate, or both drugs together. When only one active agent was assigned, a placebo for the second drug was given. Metformin (or its placebo) was begun shortly after study entry at the beginning of a spontaneous or progestin-withdrawal cycle, with a starting dose of 500 mg daily that was gradually increased over 2 wk to the target dose of 1000 mg twice daily. Clomiphene citrate (or its placebo) was started on the fifth day of this first treatment cycle at 50 mg daily for 5 d. Unless pregnancy occurred, treatment was continued for six total cycles or 30 wk, with the daily clomiphene dose increased to 100 mg, then 150 mg for 5 d in subsequent cycles if ovulation, diagnosed by any weekly serum progesterone more than or equal to 5 ng/ml (15.9 nmol/liter), failed to occur. If ovulation occurred in the previous cycle, the same clomiphene dose was given in the next cycle. No ultrasound monitoring was used, and no ovulatory trigger medications or intrauterine inseminations were permitted.

In a secondary analysis, we determined the clomiphene dose taken in the first ovulatory cycle achieved by each participant enrolled in the clomiphene-only (with placebo for metformin) and the clomiphene plus metformin (no placebo) treatment groups. The number of women and the frequency of first ovulation in response to each clomiphene dose were tabulated for each treatment group, and the outcomes compared by χ² testing. P < 0.05 was considered significant.

Results

We have previously published the baseline characteristics of the women enrolled in the PPCOS study and in each of the treatment groups (7). The women enrolled in the PPCOS study had a mean body mass index of 35 kg/m²; 30% had previously taken clomiphene and were still seeking pregnancy. Preenrollment testing was required to confirm tubal patency and an adequate partner semen analysis. Table 1 shows subject characteristics at entry. In both study groups receiving clomiphene, there were 209 enrolled subjects, consistent with the study aim to enroll equal numbers in each group.

Table 1.

Baseline characteristics of the two treatment groups (n = 209 for each group)

	Clomiphene-only group	Clomiphene plus metformin group
Age (yr)	27.9 ± 4.0	28.3 ± 4.0
BMI (kg/m2)	36.0 ± 8.9	34.2 ± 8.4
Duration infertility (months)	41.4 ± 39.4	40.7 ± 36.0
Previous conception (%)	36.8	32.1
Previous live birth (%)	15.8	13.4
Polycystic ovarian appearance (%)	91.9	91.9
Fasting insulin (μU/ml)	22.6 ± 20.7	22.4 ± 30.0
Fasting glucose (mg/dl)	89.2 ± 16.5	88.9 ± 18.6
Total testosterone (ng/ml)	61.3 ± 32.0	63.1 ± 28.4
Free androgen index	9.4 ± 7.1	9.4 ± 6.8
HOMA-IR	5.2 ± 5.3	5.6 ± 10.2

Values are expressed as mean ± sd. There were no significant differences between the two groups. BMI, Body mass index; HOMA-IR, homeostasis model assessment of insulin resistance [calculated as insulin (μU/ml) × glucose (mg/dl)/405]. 

During the entire time of study participation, at least one ovulation occurred in 157 of 209 subjects (75%) in the clomiphene-only group and 174 of 209 subjects (83%) in the clomiphene plus metformin group (P = 0.04). Data on the clomiphene dose taken during the first cycle exhibiting an ovulatory response were available for 150 of the 157 ovulators (96%) in the clomiphene-only group and 164 of the 174 ovulators (94%) in the clomiphene plus metformin group. The number of women who had their first ovulatory cycle after taking each clomiphene dose is shown in Table 2. No significant difference (P = 0.47) was found between the two treatment groups in the distribution of the clomiphene dose that was associated with first study ovulation. To address the possibility that the lack of effect of metformin to improve ovulation rate in the 50-mg clomiphene cycles was the result of the target metformin dose not yet having been reached in the follicular phase of these cycles, we performed an additional analysis comparing the frequencies of first ovulations on the 100 and 150-mg clomiphene doses. Again, there was no statistical difference between the clomiphene-only group and the clomiphene plus metformin group.

Table 2.

Clomiphene dose leading to first detected ovulation, by treatment group

Clomiphene dose (mg)	No. (%) that ovulated after
	Clomiphene only	Clomiphene plus metformin
50	95 (45.5)	109 (52.2)
100	38 (18.1)	43 (20.6)
150	17 (8.1)	12 (5.7)

Discussion

Debate has continued regarding the best first-line approach to restore ovulation in women with PCOS seeking fertility: clomiphene, an insulin sensitizer, or both together. Although clomiphene has afforded ovulation rates of 70–80% as monotherapy, pregnancy rates lag significantly behind; this has been attributed to effects on the endometrium, corpus luteum, or both resulting from the dose-dependent antiestrogenic property of clomiphene (3). For this reason a strategy to induce ovulation at the lowest possible clomiphene dose is desirable.

The PPCOS study was designed to find a superior first-line treatment strategy in PCOS for achieving a live birth. The study includes the largest cohort to date of women with PCOS undergoing ovulation induction with clomiphene and metformin alone and in combination. The ovulation rates found after clomiphene 50 mg and cumulatively after up to six cycles of escalating dose treatment are comparable to those reported previously (3). The ovulation rate with the clomiphene-metformin combination was slightly greater than with clomiphene alone.

This secondary analysis found that the distribution of first ovulations occurring with each clomiphene dose was indistinguishable between the group also taking metformin and the group taking placebo for metformin. This finding argues strongly against a significant effect of metformin to reduce the clomiphene dose required to induce ovulation in unselected women with PCOS seeking fertility. Given literature suggesting that metformin can lower the clomiphene dose necessary to achieve ovulation from greater than 150 mg (resistance to this dose or lower doses) to as low as 50 mg (6,8), the contrasting result found in the PPCOS study population could be the result of the use of the extended-release formulation of metformin, of intrinsic differences in the subject population (in previous studies, requiring previous clomiphene resistance for entry may have produced a more metabolically abnormal and, therefore, more susceptible population), or of insufficient power with the sample size of 314 ovulators to detect significance in the observed small difference in rates of ovulation at each clomiphene dose. A power calculation shows that to detect the observed 6.7% difference in ovulation rate after 50-mg clomiphene with 80% power, nearly 900 subjects would be needed in each treatment group. Alternatively, the previous studies noting a clomiphene-sparing effect of metformin (6,8) may have been biased by inadequate blinding of treatment allocation and/or study medication, which the double-blind design of PPCOS averted, and by significantly smaller sample sizes.

In regard to the PPCOS study design, the near-simultaneous initiation of metformin and clomiphene and gradual increase in metformin dose during the first 15 d of the first clomiphene cycle may have meant that the full effect of metformin was not achieved in all subjects in the first (50 mg) clomiphene cycle. Subjects failing to ovulate in this first, 50-mg clomiphene cycle would then have been advanced to higher clomiphene doses, and the ability of metformin at full therapeutic doses to synergize with 50-mg clomiphene may not have been adequately assessed. Our own data support a benefit of time on the live birth rate in the metformin plus clomiphene treatment group, but not in the metformin-alone group (7). In contrast, previous reports have found metformin to have effects in PCOS within 2 wk of its start on both ovulation (9,10) and testosterone excess (11). The lack of difference we found in ovulation rates between the 100 and 150-mg clomiphene doses supports the concept that the negative findings are not the result of the delayed effect of metformin.

In conclusion, metformin appears to have no benefit to decrease the clomiphene dose needed to induce ovulation in anovulatory women with PCOS seeking fertility. For this reason cotreatment with metformin is not an effective strategy to avoid the potential adverse effects on fecundity of higher clomiphene doses in an unselected population of women with PCOS seeking pregnancy.