Clinical outcome of various metformin treatments for women with polycystic ovary syndrome

Abstract Aim Polycystic ovary syndrome (PCOS) is an ovulatory disorder and insulin resistance and diabetes are involved in its pathophysiology. Metformin, an anti‐diabetic agent, has been reported to be useful to induce ovulation. Methods Metformin treatment was classified into four types: (1) clomiphene–metformin combination treatment for clomiphene‐resistant patients; (2) clomiphene–metformin combination for clomiphene‐sensitive patients; (3) clomiphene–metformin combination for naïve patients; and (4) metformin monotherapy. The patients underwent physical, endocrinological, and clinical examinations for their ovulation rates, pregnancy rates, and follicular development. Results The ovulation rates, pregnancy rates, and single follicular development were not significantly different among the clomiphene–metformin combination treatment groups. In the Body Mass Index (BMI) subanalysis, the pregnancy rate was higher in the BMI≥30 kg/m2 group than in the other three groups with a BMI of ≤30 kg/m2 in both cycles and cases. The ovulation rates and pregnancy rates were significantly higher in the group with a fasting insulin of ≥15 μU/mL than in the groups with a fasting insulin of <15 μU/mL in both cycles and cases. Conclusion Clomiphene–metformin combination treatment appears to be useful, at least for clomiphene‐resistant patients, and a BMI of >30 kg/m2 and a fasting insulin of ≥15 μU/mL appear to be predictors of a good result with this treatment.

combination therapy in clomiphene-resistant patients was more effective than clomiphene-alone therapy for ovulation, 5 pregnancy, and live birth rates. 6 However, the therapeutic effect of metformin in reproduction is not established, as it has been difficult to integrate the large amount of clinical data in one facility because it is not approved by insurance as an ovulation induction agent. This study was carried out to study the usefulness of ovulation induction by various metformin treatments in patients with PCOS.

| Participants
This study was a retrospective study of metformin treatments for Japanese and (4) metformin monotherapy. Each type was examined individually.
Among a total of 178 cases with 798 cycles that were collected from the 18 facilities, 16 cases with 318 cycles were excluded as unsuitable for the four treatments. Thus, a total of 162 cases with 480 cycles were analyzed for their ovulation and pregnancy rates. This study was approved by the Institutional Ethical Committee of Tokushima University, Tokushima, Japan.

| Clomiphene-metformin combination for the naïve patients
"Naïve" patients were defined as having no treatment history of clomiphene. Seven cases without an infertility treatment history with 12 cycles were treated with the clomiphene-metformin combination treatment as the first-line treatment for ovulation induction. The ovulation rates per cycle and per case and single follicular development rates were investigated.

| Metformin monotherapy
There were 18 cases with 42 cycles of metformin monotherapy. The indications for this therapy were clomiphene resistance (no ovulation) in six cases with 21 cycles, the first choice for ovulation induction in five cases with 10 cycles, habitual abortion in four cases with seven cycles, no pregnancy by clomiphene in an ovulated cycle in one case with one cycle, and a high HOMA-IR in two cases with three cycles. Fifteen cases were subanalyzed by HOMA-IR: <1.6 (n=3), 1.6-2.5 (n=5), and >2.5 (n=7).

| Pregnancy and infant outcomes
The outcomes of 64 pregnancies and 29 infants, abortion rates, multiple pregnancy rates, birth weights, birth weeks, and infant anomalies were studied.

| Side-effects
The side-effects were studied in each treatment group and compared between the clomiphene-metformin combination treatments and the metformin monotherapy.

| Dose of metformin, definition of clomiphene resistance, and ovulation rate per case
The usage of metformin and its duration were determined by each facility. Clomiphene resistance was defined as no ovulation with 100 mg of clomiphene, or at least two cycles. The occurrence of ovulation per case with multiple treatment cycles was determined when ovulation was observed in at least half of the treatment cycles.

| Statistical analysis
A subanalysis by BMI and HOMA-IR with treatments 1, 2, and 3 were done by using the chi square test and the Bonferroni post hoc test.
A subanalysis by fasting insulin with treatment 1 was done by using the chi square test. In all the analyses, P<.05 was considered to be significant.  Figure 1).  Figure 2). The pregnancy rate was 9.9% per cycle and 21.9% per case. In the BMI subanalysis, the pregnancy rate was higher in the obese group than in the other three groups in both cycles and cases ( Figure 3). The ovulation rates and pregnancy rates were significantly higher in the group with a fasting insulin of ≥15 μU/mL than in the group with <15 μU/mL fasting insulin in both cycles and cases (P<.05) ( Figure 4). In addition, the ovulation rate per cycle was 73.9% in the HOMA-IR<0.8 group, 65.5% in the HOMA-IR=0. 8

| Clomiphene-metformin combination treatment for the clomiphene-sensitive patients
The  7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29    there was no significant difference among the groups. Additionally, the pregnancy rate per cycle was 16.7% in the lean group, 22.0% in the normal group, 10.3% in the overweight group, and 10.8% in the obese group; there was no significant difference among the groups. The ovulation rate per cycle was 87.5% in the HOMA-IR<1.6 group, 97.0% in the HOMA-IR=1.6-2.5 group, and 91.8% in the HOMA-IR≥2.5 group; there was no significant difference among the groups. The pregnancy rate per cycle was 17.3% in the HOMA-IR<1.6 group, 6.1% in the HOMA-IR=1.6-2.5 group, and 91.8% in the HOMA-IR≥2.5 group; there was no significant difference among the groups.

| Clomiphene-metformin combination treatment for the naïve patients
The mean clomiphene dose was 91.7±46.9 mg/d. The clomiphene doses were 50 mg/d (50.0%, 6/12) and 150 mg/d (33.3%, 4/12), both of which accounted for 83.3% of all the treated cycles (10/12) (Table 1, Figure 1). The mean metformin dose was 520.8±198.2 mg/d.  (Table 2). Although the ovulation rate seemed high, the usefulness of the treatment was not evident because of the small number of cases.

| Metformin monotherapy
There were several reasons why metformin monotherapy was performed. The ovulation rates for the cycles and cases were 33% (6/21) in the clomiphene-resistant cases and 28% (5/10) in the first-choice cases, while 22% (4/7) of the patients had an habitual abortion (Table 3). Two (11%) patients had a high HOMA-IR and one of them achieved ovulation by clomiphene, but not pregnancy (6%). The mean metformin dose was 577.4±151.1 mg/d ( Table 1). The metformin doses were 500 mg/d (54.8%, 23/42) and 750 mg/d (38.1%, 16/42) (Figure 1). All 42 cycles were treated with metformin continuous administration. The ovulation rate was 83.3% per cycle in the first-choice cases, whereas it was 60% per cycle and 40% per case in the clomiphene-resistant patients ( Table 3).
The pregnancy rate was 14.3% per cycle and 25.0% per case in the firstchoice cases. There was no pregnancy in both the cycles and cases in the clomiphene-resistant patients. In addition, the ovulation rate per cycle was 0% in the HOMA-IR<1.6 group, 87.0% in the HOMA-IR=1.6-2.5 group, and 85.7% in the HOMA-IR≥2.5 group. The HOMA-IR<1.6 group had a significantly lower ovulation rate than did the other groups (P<.05).
The usefulness of metformin monotherapy was not clear because of the heterogeneous indications and the small number of cases.

| Pregnancy outcomes
Multiple pregnancies did not occur among the 42 pregnant patients with all metformin treatments. Only one child (3.4%, 1/29) had a cardiac malformation among the cases that were analyzed. With respect to complications of pregnancy in the 31 pregnant women, there were three cases of gestational diabetes mellitus, one case of hypertension, one case of premature rupture of the membranes, and one case of placenta previa (19.3%, 6/31) ( Table 4).

| Incidence of side-effects
Side-effects of the metformin treatments were observed in 3.6% (17/472) of the patients: diarrhea, vomiting, and stomach ache were classified as digestive system dysfunction (Table 5). There was no significant difference in the incidence of adverse effects between the three clomiphene-metformin combination treatment groups and the metformin monotherapy group.

| DISCUSSION
Insulin resistance is involved in the pathophysiology of PCOS and is more common in obese, rather than in non-obese, patients with PCOS. Weight loss, exercise, and insulin sensitizers have been proven to be effective to improve the pathophysiology of PCOS, including the ovulatory disorder.
Metformin is an oral hypoglycemic agent of the biguanide type, which inhibits gluconeogenesis in the liver, promotes sugar use in  9 The incidence of multiple pregnancies in metformin-treated patients seems to be lower than in clomiphene-treated patients. For example, multiple pregnancies have not occurred with metformin therapy in Italy, 10 whereas they have occurred in 6.0% of patients in the USA in a randomized, controlled trial 11 during clomiphene therapy. The metformin treatment was effective in inducing ovulation or recovering the normal menstrual cycle in the lean or normal BMI patients with PCOS. [12][13][14][15][16][17][18][19][20] Furthermore, some reports have indicated that non-obese (BMI<30 kg/m 2 ) patients with PCOS were more likely to respond to metformin than obese (BMI≥30 kg/m 2 ) patients with PCOS due to decreased sex steroids and increased ovulation rates. 12,14 From the present study, the metformin dose for PCOS in Japan was 750 mg/d. However, 1500-2500 mg/d is frequently used in the literature. It was reported that 1500 mg/d of metformin was effective in 87.5% (7/8) of patients who did not respond to 750 mg/d of metformin in clomiphene-metformin combination treatment for clomiphene-resistant patients. 21 In the present study, high-dose 1500 mg/d metformin was used in only two cases with three cycles; it is possible that a higher ovulation rate would be obtained by a 1500 mg/d dose for patients who do not respond to 750 mg/d of metformin.
In a meta-analysis that studied the clinical outcome of ovulation induction in clomiphene-resistant patients with PCOS, adding metformin treatment showed significantly superior outcomes to clomiphene monotherapy for the ovulation rate (OR=6.82, 95% CI=3.59-12.96), 22 6 It was concluded that adding metformin to clomiphene is extremely effective for patients with PCOS when clomiphene cannot induce ovulation. In Japanese clomiphene-resistant patients with PCOS, which included many non-obese patients, the clomiphene-metformin combination treatment showed a high ovulation rate, which was calculated as ≤55.7% by combining several reports. [23][24][25][26] Therefore, clomiphene-metformin combination treatment seems to be a useful treatment option for clomiphene-resistant patients with PCOS. A between the clomiphene-metformin combination treatment and clomiphene monotherapy groups. 9 Only one infant had a heart malformation and the anomaly rate seemed to be not very high (3.4%, 1/29 cases). With metformin monotherapy and the clomiphene-metformin combination treatments, all the pregnancies were singletons in the present study. It was reported that the infant malformation rate in the patients who took metformin in the early pregnancy period was not elevated for those patients with PCOS (OR=0.33, 95% CI=0.07-1.56) or for the patients with diabetes (OR=.85, 95% CI=0.14-5.11), compared to the normal population in a meta-analysis. 29 The side-effects of metformin were mainly digestive system dysfunction (5.5%, 12/217 cycles), such as vomiting and diarrhea, with a rate that was comparable to that seen in the treatment of diabetes. Lactic acidosis and hypoglycemia are serious side-effects of metformin that did not occur in the present study. Lactic acidosis is very rare, affecting three out of 10 million persons per year, 30 which is likely to occur in patients with liver and kidney dysfunction, and such patients would be few among young women receiving fertility treatment.
The clomiphene-metformin combination treatment appears to be useful, at least for clomiphene-resistant patients, because of higher ovulation and pregnancy rates; a BMI of >30 kg/m 2 or a ≥15 μU/mL fasting insulin level could be predictors of a good response to such treatment.
with the ethical standards of the responsible committees on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and its later amendments. Informed consent was obtained from all the patients to be included in the study.