What is the best first-step therapeutic approach in treating anovulatory infertility in patients with polycystic ovary syndrome? Questions that are still unanswered

Polycystic ovary syndrome (PCOS), a multifaceted disease characterized by androgen excess, oligomenhorrea and polycystic ovaries at ultrasound [1], is the most common cause of anovulatory infertility, affecting about 5 – 10% of women of reproductive age [2].

Over the years several approaches have been proposed to induce ovulation in women with PCOS [3], even if clomiphene citrate (CC), the first agent used in experiments for ovulation induction in oligomenorrheic women by Holtkamp and colleagues [4] and then introduced into general clinical practice by Greenblatt and associates [5], has always represented the gold standard treatment. In fact, the administration of CC is characterized by low costs, limited dose-dependent side-effects, and simplicity of administration and management since there is no need for ongoing monitoring [6]. CC is effective since its administration is related to an ovulation rate of 60 – 85% and a pregnancy rate of 30 – 40%[7]. The reason for the discrepancy between ovulation and pregnancy rates is not known, but several possible hypotheses have been suggested [8],[9].

Metformin, a safe and effective oral biguanide administered for treating type 2 diabetes mellitus, has also been used as a treatment for PCOS patients. The administration of metformin induces ovulatory cycles in anovulatory CC-resistant or non-resistant patients with PCOS and improves the ovulation rate as an additional treatment in women who received CC [10],[11].

In the February issue of the New England Journal of Medicine, Legro and co-workers [12] try to put an end to the long-standing controversy about the use of CC or metformin or both to induce ovulation in women with PCOS. They report on the results of a large randomized controlled trial (RCT), comparing the effects of CC, metformin and combination therapy with both drugs, in 626 infertile PCOS women [12].

This trial [12], one of the widest multi-center RCTs published on ovulation induction, demonstrates a higher efficacy of CC administration in comparison with both metformin and metformin plus CC administration in terms of live-birth rate, the best primary endpoint for a clinical study assessing the effectiveness of infertility drugs. These data disagree with an our previous head-to-head RCT [11] comparing metformin with CC, which demonstrated that, in an infertile non-obese PCOS population, metformin is equally as effective as CC for inducing ovulation but is superior in terms of pregnancy rate.

Although the study's protocol [12] was designed to evaluate the best first-step treatment in inducing ovulation in anovulatory PCOS patients, more than 50% of the patients already had a previous treatment with CC, metformin or both. This point is crucial since each specific treatment can influence the following one, and this is particularly true for drugs such as CC and metformin. In fact, previous CC or metformin treatment potentially induces a selection of patients ‘resistant’ to the treatment or those ovulating under treatment who did not conceive (‘treatment failure’). In addition, the kind of treatment previously given to subjects ‘without exposure to study drug’ is unknown. For example, a single one-cycle treatment with oral contraceptives can improve the efficiency of a following cycle of CC [13].

Based on these considerations, the cycle-by-cycle data analysis performed according to a previous treatment is surprising. In fact, when analyzing the data according to previous CC treatment, no change was observed in the results. On the contrary, it is well known that the success rate of CC re-administration in patients with CC resistance or CC failure is very low (∼10%) [13]. It is also surprising to observe that patients with a body mass index (BMI) of 36.0 ± 8.9 kg/m² (mean ± standard deviation) had an ovulation rate of 43.1% using the lowest dose of CC (50 mg daily), and that there was no significant improvement in reproductive outcomes by increasing the CC dosage [14],[15].

In addition, both CC and metformin are drugs that could exert their effects after suspension. Several experimental data demonstrate that high CC concentrations are present for a long time after treatment suspension in previously treated patients [16]; thereby, after the use of high dosages, exerting an overall antiestrogenic effect without its administration. Similarly, a previous metformin administration could sensitize patients for subsequent treatments with CC alone or combined with metformin. In fact, it has been demonstrated that metformin can show efficacy several months after the initial administration [17].

Another crucial point is the effect of BMI on the results. No limit for BMI was specified in the study design for the enrollment of patients and the population that was studied consisted of obese patients in very large quantity, whereas a low proportion of lean, normalweight or overweight patients was randomized and studied. In this regard, we feel that the population studied does not represent all countries. In European countries, and particularly in the Mediterranean area, the prevalence of obesity in PCOS patients is significantly lower and there is about a 10-point difference in BMI among PCOS populations [18].

Although a large number of severely obese patients were enrolled in the study [12], no comment was made regarding their treatments. In fact, women affected by severe obesity are frequently under insulin-sensitizing drugs, cholesterol-lowering drugs, dietary treatments, or physical exercise programs. It is essential to note that all of these approaches are effective in terms of reproductive outcomes [19],[20], even if we do not know exactly how and by how much they can improve the efficacy of ovulation induction agents. On the other hand, patients with morbid obesity who underwent bariatric surgery could have a reduction or an improvement in subsequent reproductive performance [[21]].

The concern regarding BMI in PCOS women is also crucial because the risk/benefit ratio of the two study drugs closely relies on patients' BMI. In fact, a high BMI is a negative predictor for CC efficacy [14],[15], and recent data seem to clearly demonstrate that metformin administration is less effective in obese than in nonobese PCOS patients [22]. Unfortunately, the differential effect of BMI on the efficacy of the two treatments is still not exactly known. For instance, the treatment regimens used by Legro and co-workers [12] are not well standardized since at present there is no wide study evaluating the dose response of the two treatments with particular regard to different BMI.

Another confounding factor limiting the reproducibility of the results is the very poor reproductive outcomes due to the inclusion of couples with potential subfertility factors. In fact, women with only one patent fallopian tube and/or patients having a partner with a sperm concentration of at least 20 × 10⁶/ml one year before study entry, regardless of motility and/or morphology, were not excluded [23]. The presence of only one patent fallopian tube can be a potential feature of peritoneal-factor subfertility due to minimal endometriosis [24] or eventual contralateral hydrosalpinx [25]. Similarly, abnormal sperm motility and/or morphology can again be a potential factor of subfertility [26].

Finally, the interesting trial by Legro and co-workers [12] was powered to demonstrate a difference in live-birth rates on the entire PCOS population, but probably not for giving answers regarding the different PCOS subpopulations.

In conclusion, we ask ourselves if it would be better to design a wide clinical trial on an unselected population or smaller RCTs on well-selected groups of patients. To date, we feel that further experimental RCTs are needed, aiming to evaluate the best regimen of administration for CC, metformin or both according to potential predictors of success, and only then in turn a wide clinical trial not designed on the best single approach but on the best strategy to improve fertility in anovulatory PCOS patients. In fact, a specific treatment could be effective as a first-step approach but not useful in an integrated strategy. In other words, in the near future, we need to know how the first drugs can modify the choice for the subsequent therapeutic steps and thus define the ‘best’ first-step approach taking into consideration the following therapeutic steps in treating patients for whom treatment fails.