Gonadotropin Treatment of PCOS: Is There a Preferred Protocol of Treatment?

P.F. Rinaudo and P. Patrizio

Hospital University of Pennsylvania, Department of Obstetrics & Gynecology, Division of Reproductive Endocrinology and Infertility, Philadelphia, PA, U.S.A.

Introduction

Polycystic ovarian syndrome (PCOS) is a disease that affects approximately 4%-12% of the female population (Knochenhouser, 1998). It is thought to represent an adaptation allowing the survival of the individuals in conditions of reduced alimentary resources. Therefore our ancestors living in a hunting gathering community who had the PCOS genes had a competitive survival advantage over those who did not. It is a complex disease that requires the interaction of several genes (Dunaif, 2001). Its mode of inheritance is therefore multifactorial.

The new understanding of the pathophysiology of the disease and in particular the discovery that insulin resistance plays a cardinal role in the manifestation of the disease is already modifying the therapeutic protocols.

In addition to obesity and hirsutism, infertility represents one of the cardinal manifestations of PCOS. Stein and Leventhal defined the syndrome in 1935 and proposed ovarian wedge resection as a way to correct the disease and increase fertility.

Since then medical therapy has gradually supplanted surgical treatment.

Pathophysiology of PCOS

The fundamental concept of PCOS is that there is not a specific pathophysiologic mechanism responsible for the syndrome. Any cause of chronic anovulation will lead ultimately to the development of polycystic ovaries and finally PCOS. The typical hormonal finding is a “steady state”, a monotonous level of sex steroids and gonadotrophins secondary to chronic anovulation, in clear contrast with the regular and predictable hormonal variations in the ovulatory cycle. There is an ovarian increase in secretion of testosterone, androstenedione, 17-hydroxyprogesterone. The adrenal produces elevated levels of Dehydroepiandrosterone and Dehydroepiandrosterone sulfate. There is increased peripheral production of estrone. Moderate hyperprolactinemia is present in 15% of patients (Jacobs, 2000). LH is chronically elevated, secondary to the increase in estrogens and the increased frequency of secretion of GnRH. FSH levels are low, but still detectable, partially because of the negative feed back of steroids and possibly for the increase in inhibin B (Franks, 2000). This hormonal milieu translates in the typical feature of the ovaries in PCOS.

The follicles are chronically stimulated by the low levels of FSH, but do not reach complete maturation. The elevated androgens prevent complete follicular development. Therefore the ovaries are filled with multiple small follicles (2-10 mm in size) and with a prolonged life span. In fact the number of primordial follicles is similar to that of normal patients, but the number of primary and secondary follicles are twice than those observed in the normal ovary (Franks, 2000). Interestingly the oocyte content of Growth differentiation factor 9 (GDF9) is reduced in PCO if compared with control (Teixeira Filho, 2002). The follicles are surrounded by hyperplastic and hyper-active theca cells. Granulosa cells are arrested in development, while when cultured in vitro they display a marked response to FSH and insulin growth factor (IGF). This explains the explosive response of these patients during stimulation cycles and their propensity to develop ovarian hyperstimulation syndrome (OHSS). On the other hand, destruction of ovarian tissue (wedge resection or by ovarian drilling) will determine reduction of androgen levels (in particular intra ovarian androgen), and resumption of ovulations (Judd, 1976).

Making the Diagnosis of PCOS in an Infertility Clinic

PCOS is one of the most frequent cause of anovulation and related infertility, corresponding to the World Health Organization (WHO) II group. Diagnostic criteria in the United States suggested by the 1990 NICHD PCOS conference include: hyperandrogenism and chronic anovulation in the absence of specific disease of the adrenal (like congenital adrenal hyperplasia), of the ovary (like androgen secreting neoplasms), and pituitary (like hyperprolactinemia) (Dunaif, 2000). In Europe an additional diagnostic criteria is the ultrasonographic presence of polycystic ovaries. These are defined as containing increased ovarian stroma and eight or more sub capsular follicular cysts = 10 mm in diameter per ovary (Dunaif, 2000).

At least seventy per cent of PCOS patients will have polycystic ovaries. However many women with polycystic ovaries do not have PCOS (20% of normal ovulatory women may have this sonographic appearance, and even 14% of women taking oral contraceptives- Polson, 1988 and Clayton, 1992). Therefore the term PCOS should only be utilized to define the patients who present with anovulation and hyperandrogenism.

Finally, it is useful to differentiate between polycystic ovaries and multi-cystic ovaries: the latter do not contain increased stroma and contain larger cysts (Adams, 1985). They occur usually during puberty.

Infertility Treatment

The focus of this chapter is gonadotropin induction of ovulation. In general different stimulation protocols are designed to overcome the PCOS hormonal status and to recreate the variability seen in the ovulatory cycles.

We will however follow the clinical practice to start infertility treatment in a step-wise fashion from a least aggressive to a more aggressive treatment. We will focus on gonadotropin treatment and briefly mention the use of clomiphene citrate and the IVF protocols for PCOS patients.

Table1 provides a summary of the available best evidence (randomized controlled trials RCT, where available) for treating patients with PCOS. The Cochrane database was used as a primary source of information because it provides us with a careful analysis of the literature and it is considered to be today the best source for evidence-based medicine.

First Step: Clomiphene Treatment

Absent or sporadic ovulation is the main indication for clomiphene treatment and therefore PCOS patients are excellent candidates for its use. However PCOS patients are also the more likely not to respond to it (Imani, 1998) and as many as 15% (Nugent, 2002) of patients may be considered clomiphene failures. In a study for the Cochrane database by Hugues et al (Hugues, 2002) clomiphene was associated with an increased pregnancy rate per treatment cycle (odds ratio 3.41, 95% confidence interval 4.23 to 9.48). Clomiphene citrate (at doses between 50 to 250 milligrams per day) appears to be an effective method of inducing ovulation and improving fertility in oligo-ovulatory women. However adverse effects include risk of multiple pregnancies and ovarian cysts.

Second Step: Gonadotropin Treatment

Gonadotropins are the natural next step for ovulation induction. One characteristic of ovulation induction in PCOS patients is the slow response and the significant risk for OHSS and cyst formation. Historically multiple regimens have been used to overcome these difficulties.

Conventional step up dose regimen

The classical protocol used in the 1970s for PCOS patients included stimulation with 150U of hMG (containing 75U FSH and 75U LH), that was increased every 3-5 days by 50% until an ovarian response occurred.

However PCOS patients often respond in a dramatic way to this protocol and the rate of Ovarian hyperstimulation syndrome is unacceptably high (1.1%- 14%, Fauser, 1997).

Low dose step up protocol

Brown (Brown, 1969) introduced the concept of FSH threshold: a follicle can be rescued only when FSH is above a certain level. In patients with PCOS the chronically low and constant level of FSH allows initial development of multiple follicles that however do not continue development and undergo atresia. Increasing the level of FSH in a gradual and step wise fashion allows the rescue of a limited number of follicles mirroring what happens in the normal hormonal cycle. This approach reduces of the incidence of OHSS.

The authors recommended incrementing the dose of Gonadotropins of only 10-30% of the initial dose every 5 days.

The most used current step up protocol is characterized by a low starting dose (75U), maintained for a longer period of time (14 days,) and then increased only of a small amount (37.5U) per week. This protocol is associated with a low incidence of severe OHSS (0-0.5%) and multiple pregnancy (6-18%) with good cycle fecundity per started cycle (11-17%- (Balash, 2001). A small randomized trial (Sagle, 1991) proved that the dose and timing of gonadotropins administration was more important than the type used (hMG and urinary FSH had similar results). Two prospective studies compared the standard protocol versus the low dose step up protocol and confirmed the safety of the latter with a concomitant comparable pregnancy rate (Homburg et al 1995 and Hedon et al, 1998). Currently the low dose step up protocol is the most widely used in PCOS patients.

Low dose step down protocol

This protocol is based on the assumptions the high starting dose replicate the mid-cycle FSH surge and the fact that the dominant follicle is less dependent for its development to FSH levels than smaller follicles (van Santbrink, 1995). The initial dose is between 1.5 and 2.5 ampoules (75 IU follicle-stimulating hormone per ampoule) per day (dependent on body weight), and decreasing steps (every 2-3 days) of 0.5 ampoules/day are based on sonographic findings. Pregnancy rates are comparable with those reported for step-up regimens, with a low incidence of complications (i.e. multiple gestation and ovarian hyper-stimulation, see table 1).

Sequential low dose protocol

This protocol combines characteristic of the 2 prior protocols (Hugues et al 1996): the initial step up phase is followed by a step down phase when the leading follicle reaches 14 mm in diameter. The FSH dose is then reduced by half until HCG administration. This randomized prospective study showed that the sequential protocol is as effective as the low dose step up protocol in term of pregnancy rate and safety, while reducing the level of Estradiol at the time of HCG administration and the number of midsize follicles (14-15 mm).

Choosing the starting dose

It is often difficult to decide the starting dose in patients with PCOS in which the risk of OHSS is elevated. To predict the individual FSH response dose before initiation of therapy, Imani et al (Imani, 2002) studied prospectively the characteristics of a group of clomiphene resistant patients. Using multivariate analysis they obtained the following formula: Dose of FSH= [4 body mass index BMI (in kg/m(2)] + [32 clomiphene citrate resistance (yes = 1 or no = 0)] + [7 initial free insulin-like growth factor-I (in ng/mL)] + [6 initial serum FSH level (in IU/L)] -51. For example a patient with a BMI of 25, resistant to clomiphene, with a baseline IGF1 of 240 ng/ml and FSH of 8 IU/L, will need a FSH starting dose of 180 IU.

Further comparative prospective trials should test the validity of this approach.

Type of gonadotropins

Multiple studies have compared different formulations of gonadotropins.

Human menopausal gonadotropin contains 75U of FSH and 75 of LH, but only 4% of protein content is gonadotropin, while the rest represent “contaminant proteins” not purified by the manufacturing process.

Urinary FSH (uFSH) contains 75U of FSH and less than 1% of LH activity, but still 95% of contaminating proteins.

The Chocrane database (Hugues E, 2000) analyzed the available randomized protocols published comparing the 2 gonadotropins and found no difference in pregnancy rate. However urinary FSH is associated with a reduction of moderate or severe OHSS (table1).

In 1993 highly purified FSH (hpFSH) was made available to clinicians: it contains the same amount of FSH, but less than 0.1% of LH. Moreover the content of contaminant proteins is reduced to only 4%, allowing therefore a subcutaneous injection.

We could not find a randomized controlled clinical trial (RCT) comparing hpFSH with other gonadotropins for ovulation induction in PCOS patients.

Finally, recombinant human FSH has been available in the U.S.A. since 1997, but earlier (1995) in Europe. The preparation does not contain LH and is free of impurity. In addition it offers a consistency between different batches, fact that is highly desirable in clinical practice.

The Cochrane database reviewed also the available evidence (4 RCT) (Bayram, 2001) comparing recombinant FSH versus urinary FSH. The review found no significant differences for ovulation rate, pregnancy rate, miscarriage rate, multiple pregnancy rate and OHSS. It appears therefore that at the present time there are not sufficient data to determine weather rFSH or uFSH is preferable for ovulation induction in women with PCOS.

Pulsatile GnRH

The use of pulsatile GnRH has been used very successfully in patients with hypogonadotropic hypogonadism. Because PCOS is characterized by a hormonal steady state, several authors tried the use GnRH to overcome this hormonal impasse. Three RCTs and one non-randomised comparative trial were identified comparing four different treatments: GnRH versus HMG, GnRH following GnRHa pre-treatment versus no pre-treatment, GnRH and FSH versus FSH, and GnRH following GnRHa pre-treatment versus GnRH following oral contraceptive pre-treatment. This means that there was only one trial in any one comparison. In the first two studies, data of pre- and post-cross-over were not described separately. Therefore, these results could not be included in the Meta-view analysis. The odds ratio for ovulation rate was 16 (95 % CI: 1.1-239) in the study comparing GnRH and FSH with FSH alone. When GnRH after GnRHa pre-treatment was compared with GnRH after oral contraceptive pre-treatment, an odds ratio of 7.5 (95 % CI: 1.2-46) was obtained. All trials were small and of too short duration to show any significance in pregnancy results. Only one to four pregnancies occurred in each study. Multiple pregnancies were not seen. OHSS was seen only in the patients stimulated with hMG (Bayram, 2002).

Adding GnRHa

Elevation of endogenous LH levels may result in premature luteinization. This may also be associated with increased rate of spontaneous abortion. Therefore reducing the endogenous LH levels by means of pretreatment with GnRH analogues has been a common practice in many centers.

Two meta-analyses are reported in the Cochrane database. One focused specifically in patients resistant to clomiphene (Hughes E, 2000). Three randomized control trials satisfied the inclusion criteria. These studies evaluated if GnRHa pre-treatment plus FSH/hMG increase the rate of clinical pregnancy and/or decrease the rate of spontaneous abortion, compared with hMG/FSH alone.

Interestingly these studies found a non significant increase in pregnancy per treatment cycles (1.50 CI 0.72-3.12), but also of severe ovarian hyperstimulation syndrome (1.4 CI 0.5-3.92), although without reaching statistical significance. The authors concluded that these studies were too small to clearly demonstrate clinically significant differences in pregnancy rate between the two approaches.

The other meta-analysis (Nugent, 2000) included 14 RCTs. Results showed a reduction in the incidence of OHSS with FSH compared to hMG in stimulation cycles without the concomitant use of a GnRH-a (OR 0.20; 95% CI 0.08-0.46) and a higher hyperstimulation rate when a GnRH-a is added to gonadotrophins (OR 3.15; 95% CI 1.48-6.70). In none of the comparisons was there a significant improvement in pregnancy rate, likely secondary to lack of power. There was a trend towards better pregnancy rates with the addition of a GnRH-a to gonadotrophin stimulation. In general urinary-derived FSH preparations did not improve pregnancy rates when compared to traditional and cheaper hMG preparations; their only demonstrable benefit was a reduced risk of OHSS in cycles when administered without the concomitant use of a GnRH-a. No conclusions could be drawn on miscarriage and multiple pregnancy rates due to insufficient reporting of these outcomes in the trials.

Adding metformin

In the last few years several studies assessed the effects of attenuation of hyperinsulinemia obtained by insulin sensitizing agents, in women with polycystic ovary syndrome (PCOS). However, the results of these studies, mostly uncontrolled and short-term, are still inconclusive, and there is no long-term follow-up. Several randomized controlled trials compared ovulation induction with clomiphene with and without metformin (Fleming, 2002, and Kokac, 2002). Nestler et al (Nestler, 2002) reviewed the topic. The results show a significant albeit not dramatic improvement in ovulation, suggesting a potential role for these medications in the management of PCOS.

There are no RCTs that compare the use of insulin sensitizing agents and gonadotropins. The only published study that addresses this issue is a trial that evaluated clomiphene-resistant women with PCOS and randomized them to receive either no treatment or metformin (1,500 mg daily) for one month prior to ovulation induction with FSH (DeLeo, 1999). The number of follicles >15 mm in diameter on the day of hCG administration was significantly lower in women treated with metformin compared with the untreated women (mean of 2.5 vs. 4.5 follicles, respectively), and no cycles with hCG were withheld in women treated with metformin. Moreover, plasma estradiol concentrations were significantly lower in cycles treated with FSH and metformin as compared to those treated with FSH alone (mean of 45 vs. 72 ng/dL, respectively). These findings suggested that treatment with metformin may allow a reduced rate of hyperstimulation with FSH therapy and may reduce the risk of multiple gestations.

RCTs are clearly needed and we suspect this strategy will improve our armamentarium to achieve pregnancy.

Third Step: IVF

While the goal of induction of ovulation is the development of one or few ovulatory follicles, the goal of stimulation in IVF cycles is to obtain multiple follicles, but without incurring in OHSS. High dose exogenous gonadotropins are given to over ride the mechanism of follicular selection. A reversible hypogonadotropic hypogonadism is induced with GnRH analogues. The suppression of LH appears to be particular useful in PCOS patients. Multiple protocols (step down vs. step up, with urinary gonadotropins vs. recombinant gonadotropins, long vs. short desensitization with GnRH analogues or more recently with GnRH antagonists) are described in the literature (see Jacobs, 2001 for a review).

Jacobs recommend to utilize a long protocol of pituitary desensitazion and to start with a lower dose (75-150 U) of a preparation containing FSH. The dose will be then appropriately increased or decreased according to the patient response.

It should however be noted that there is not a RCT specifically designed to determine the best protocol in PCOS patients undergoing IVF.

Surgical treatment

Ovarian drilling

Surgical ovarian wedge resection was the first established treatment for anovulatory PCOS patients but was largely abandoned because of the risk of post-surgical adhesion formation. Adhesion formation with the related impact on fertility is still the main critic to the performance of this or related laparoscopic procedure (ovarian drilling by diathermy or laser).