Letrozole combined with low dose highly purified HMG for ovulation induction in clomiphene citrate-resistant infertile Chinese women with polycystic ovary syndrome: a prospective study

Background and aim: There are still open questions about ovulation induction in clomiphene citrate-(CC)-resistant infertile women. Especially little is known about efficacy and safety of letrozole (LTZ) combined with low-dose highly purified human menopausal gonadotropin (Hp-HMG) in women with polycystic ovary syndrome (PCOS). Methods: Prospective, single-arm single-center trial in 200 infertile PCOS patients refractory for at least three CC-treatment cycles. Women with hyperandrogenism took Diane-35 for at least 3 months. All patients got LTZ on day 3 for 5 d in combination with Hp-HMG, starting with 75 IU from cycle day 7 and maintained for up to 3 d. The maximum dose was 150 IU. Primary end-points were ongoing and clinical pregnancy rate, secondary end-points mono-follicular development, ovulation rate, OHSS, multiple pregnancy and early pregnancy loss. Major safety end-point was the incidence of adverse events. Results: Within 395 cycles the ongoing pregnancy rate was 28.24%, for cycles 35.23%, for patients 68%. The rate of ovulation per cycle was 97.7%, percentage of mono-follicular development 70.9%. No severe OHSS, multiple pregnancy, local or systemic side effects were seen. Conclusions: LTZ combined with low-dose Hp-HMG is an effective and safe choice for reducing hyperstimulation and increasing pregnancy rate in CC-resistant women with PCOS.

背景和目的: 对克罗米芬 (CC) 耐药的不孕妇女进行排卵诱导仍然存在开放性问题。 对于患有多囊卵巢综合征 (PCOS) 的女性, 来曲唑 (LTZ) 与低剂量高效纯化的人尿促性腺激素 (Hp-HMG) 联用的功效和安全性的研究尤为少见。 方法: 对200例至少三次CC治疗周期耐药的不孕症PCOS患者进行前瞻性单臂单中心试验研究。患雄激素过多症的女性用达英-35至少3个月。 所有患者在月经第3天开始应用LTZ, 每日5mg, 共用5天, 从第7天开始加用Hp-HMG, 后者的初始剂量为75IU并维持3天, 最大剂量为150 IU。 主要终点包括继续妊娠率和临床妊娠率, 次要终点为单卵泡发育, 排卵率, OHSS, 多胎妊娠和早期妊娠丢失。主要安全终点为不良事件的发生。 结果: 395个周期中, 持续妊娠率为28.24%, 在周期中的占比为35.23%, 在患者中占比为68%。每周期的排卵率为97.7%, 单卵泡发育率为70.9%。 没有观察到严重的OHSS, 多胎妊娠, 局部或全身的严重副作用。 结论: 对于克罗米芬耐药的PCOS不孕女性, LTZ联合低剂量Hp-HMG是一种有效和安全的选择, 可用于减少卵巢过度刺激的发生和增加妊娠率。

Keywords: Clomiphene citrate resistance; highly purified human menopausal gonadotropin; letrozole; ongoing pregnancy rate; polycystic ovary syndrome

PCOS is the most common endocrine disorder in women of reproductive age and affects 10% using the National Institutes of Health (NIH, Bethesda, Maryland criteria for diagnosis [[1]]. PCOS is the primary cause of anovulatory infertility with a frequency of 50–70% [[3]]. Clomiphene citrate (CC) is the traditional first-line treatment for chronic anovulation. However, CC elicits adverse antiestrogenic effects on endometrium and cervical mucus, which could lead to low pregnancy rate despite relatively high ovulation rate [[4]]. Moreover, 20–25% of PCOS women fail to ovulate with incremental CC [[6]]. Mitwally et al. [[4]] first reported letrozole (LTZ) as a novel use in patients with CC-resistance, which was defined as failure to conceive despite ovulation in six or more cycles of CC-treatment. In patients with PCOS, LTZ and CC had equivalent efficacy for ovulation induction [[7]]. LTZ is considered ideal for ovulation induction by suppressing estrogen production, decreasing the negative feedback and subsequently increasing FSH. LTZ does not deplete estrogen receptors, so it may have no negative impact on endometrium and cervical mucus, and typically results in mono-ovulation [[9]].

To stimulate ovulation in women who fail to ovulate with CC, gonadotropin preparations already have been used with controversial results [[11]]. Highly purified human menopausal gonadotropin (Hp-HMG) reduces LH activity of original HMG, its purity reaches the preparation of recombinant single component [[13]]. Using sequential low-dose Hp-HMG reduces the risk of multi-follicle development, ovarian hyper-stimulation syndrome (OHSS) and multiple pregnancy. So the combination of LTZ and purified HMG should be ideal for ovulation induction especially in PCOS patients with CC-resistance. Aim was to prove this in a prospective study.

This was a prospective, single-arm and single-center trial. A total of 200 infertile outpatients were included attending the Department of Gynecological Endocrinology at the Beijing Obstetrics and Gynecology Hospital between January and September 2015. The diagnosis was based on the revised criteria of Rotterdam (2004) added by the latest NIH-guidelines [[14]]. Each couple participated voluntarily and signed the informed consent. This study was approved by the Ethical Committee of Beijing Obstetrics and Gynecology Hospital, Capital Medical University.

Inclusion criteria: (1) women with PCOS who need ovulation induction and intercourse guidance, (2) all having CC-resistance, (3) age between 25 and 35 years (average population in our hospital coming because of infertility), (4) infertility for 15 years (average population for which we induce ovulation), (5) serum level of FSH < 10 IU/L in the early follicular phase, (6) normal uterine function documented by hysterosalpingography, hysteroscopy or ultrasound, (7) male partner with normal sperm quality according to the 2010 WHO-criteria and (8) no use of ovulation stimulants in the recent cycle (to avoid drug interaction).

The exclusion criteria included: (1) any acute genitourinary infections or sexually transmitted diseases (using blood/urine test and clinical assessments), (2) any serious genetic, physical or mental disorders (according to history including family history) and (3) infertility factors other than anovulation.

Basal transvaginal ultrasonography (to evaluate ovaries and uterus), age, type and duration of infertility, body mass index (BMI), waist-to-hip ratio (WHR), FSH, LH, estradiol (E2) and testosterone (T), fasting plasma glucose (FPG), fasting insulin [FINS(OH)] were recorded for each patient. All patients with biochemical hyperandrogenism took Diane-35 for at least 3 months to correct abnormal hormone level before the intervention phase of the study.

All patients (n = 200) got 5 mg of LTZ (Winfrey®; Jiangsu Hengrui, China) orally for 5 d starting from the third day of a spontaneous or progesterone-induced withdrawal bleeding, plus Hp-HMG (Menopur®; Ferring, Germany) 75 IU, low-dose starting with 75 IU s.c. from cycle day 7 and maintained for up to 3 d, thereafter daily adjusting the dosages according to the patients' follicular response (number and the size of follicles), maximum dose not more than 150 IU. Treatment cycles were monitored by transvaginal ultrasonography with transducer 6 MHz (Mylab60®; Esaote, Netherlands). When the leading follicle reached at least 18 mm, or two follicle diameters were ≥ 16 mm and urinary LH < 40 IU/L, ovulation was triggered with s.c. injection of 0.1 mg gonadotropin-releasing-hormone-agonist (GnRH-a, Decapeptyl®; Ferring, Germany), and timed intercourse was advised 24–36 h after GnRH-a triggering. On the day of GnRH-a triggering, LH, E2 and progesterone levels were measured; endometrial thickness and pattern were assessed according to Gonen [[15]]. Two days after giving GnRH-a, ovulation was monitored by ultrasound. The patients were given dydrogesterone (duphaston®; Abbott, Holland) 10 mg/d for luteal support after ovulation for 2 weeks.

To determine biochemical pregnancy, blood β-hCG was measured 14 d subsequently. After 5–6 weeks of ovulation, ultrasound was performed to observe the gestational sac and fetal heart beat, which determined a clinical pregnancy. Definition on ongoing pregnancy per started cycle was the presence of at least one viable fetus 10–11 weeks after confirmation of clinical pregnancy. Early abortion was defined as the spontaneous loss of a clinical pregnancy before 12 weeks of gestation and a multiple pregnancy as one in which there was more than one fetal sac.

We only could compare with the study Xi et al. [[16]] but their study had some limitations, especially the patient sample (n = 94) was too low, especially to assess the risk of OHSS. According to a single-group-design (two-tailed) and referred with the pregnancy/complete cycle (%) = 20/81 (24.7), and 15% regarded as the improvement of clinically significant difference, we estimated our sample size as 198, based on alpha = 0.05, power = 80%. Since our center is the largest department of Gyn. Endocrinology in China, we could recruit this large sample-size-number within the short time of nine months.

The primary end-points were ongoing pregnancy rate and clinical pregnancy rate per started cycle, secondary end-points included follicular development (number of follicles ≥14 mm and ≥18 mm), percentage of mono-follicular development, ovulation rate, endometrial status (endometrial thickness and endometrial pattern), serum level of E2 and LH on the day of GnRH-a triggering, cycle cancelation rate due to the risk of OHSS according to Golan's classification system [[17]]), anovulation rate, duration of stimulation, total dose of Hp-HMG, OHSS, multiple pregnancy and early pregnancy loss. Major safety end-point was the incidence of adverse events.

Since, we already treat all our patients very effectively according to the protocol used in this study, our ethical committee would not allow other treatments as control group. So we decided to make a historical comparison, only possible with the study of Xi et al. [[16]]. They treated CC-resistant patients within three groups: (1) LTZ plus HMG (n = 94), (2) CC plus HMG (n = 90) and (3) HMG-only (n = 71). Comparison of group (1) with our study should lead to a suggestion for an optimal protocol which was different in our study.

SPSS version 21.0 (SPSS Inc., Chicago, IL) was used for statistical analysis. Normality of distribution was assessed with one-sample Kolmogorov–Smirnoff test. The categorical data were expressed as frequency and rate, χ2 test was used for the intergroup comparison. The quantitative data were expressed as mean ± SD and comparisons between groups were conducted by using the independent samples t-test or non-parametric test based on the results of normality test. For a historical comparison with Xi's study [[16]] we used one-sample t-test or non-parametric test, with p < 0.05 considered as statistically significant difference.

In 200 PCOS patients with CC-resistance, we tested our treatment up to three cycles, in total for 395 cycles. Because the analysis was done per cycle, we did not exclude any patient. A total of 174/200 (87%) had hyperandrogenism and got Diane-35 for at least 3 months before the interventional study. Baseline characteristics and the endocrine profile as well as clinical parameters during stimulation can be seen in Table 1. Regarding safety and adverse events there were no OHSS and multiple pregnancy. Using the GnRH-a, only slight abdominal distension was seen, but no redness, swelling and bruising in the injection site.

Table 1. Baseline characteristics and clinical parameters during stimulation.

1 BMI: body mass index; WHR: waist circumference/hip circumference; FSH: follicle-stimulating hormone; LH: luteinizing hormone; E2: estradiol; Hp-HMG: highly purified human menopausal gonadotropin; trigger day: day of GnRH-a triggering; GnRH-a: gonadotropin releasing hormone agonist; pattern A: triple-layer structure of endometrial thickness.

2 Data are expressed as mean ± SD or n (%).

Patients' characteristics of pregnant group and non-pregnant group are summarized in Table 2. The number of dominant follicles in pregnant group was higher, the number of follicles between 10 and14 mm lower than in the non-pregnant group (p < 0.05). Endometrial thickness on the day of GnRH-a triggering was higher in patients with pregnancy than those who were not pregnant (1.01 ± 0.02 versus 0.87 ± 0.03 cm, p < 0.05). The pregnant group had a higher progesterone level after ovulation than that of non-pregnant group (p < 0.05). Related to all 200 patients, the clinical pregnancy rate was 68% (136/200), twin gestation accounted for 2.94% (4/136).

Table 2. Characteristics and clinical parameters during stimulation according pregnancy or non-pregnancy.

• 2 *p < 0.05.
• 3 BMI: body mass index; WHR: waist circumference/hip circumference; FSH: follicle stimulating hormone; LH: luteinizing hormone; FPG: fasting plasma glucose; FINS(OH): fasting insulin; E2: Estradiol; Hp-HMG: Highly purified human menopausal gonadotropin; trigger day: day of GnRH-a triggering; GnRH-a: gonadotropin releasing hormone agonist.
• 4 Data are expressed as mean ± SD; NS: no significance

The results regarding a historical comparison with the study of Xi et al. [[16]] can be seen in Table 3. Regarding efficacy, the ovulation rate in our study was higher (97.8%, 386/395 versus 86.2%, 81/94) (p = 0.001). The rate of canceled cycles was significantly lower in our study (3.3%, 13/395) versus 8.5%, 8/94) (p = 0.025). Clinical pregnancy rates for cycles were similar, but with tendency higher in our study (35.2% versus 24.7%, p = 0.067).

Table 3. Ovarian stimulation and pregnancy outcome in all treatment cycles in comparison with the study of Xi.

• 5 GnRH-a: gonadotropin releasing hormone agonist; hCG: human chorionic gonadotropin; Hp-HMG: Highly purified human menopausal gonadotropin; Cancelled cycle: when more than 3 follicles reached a diameter of ≥18 mm or 4 follicles were ≥ 14 mm with mild abdominal distension, nausea and vomiting because of the risk of OHSS. Started cycle: our group = 395, historical group = 94; Complete cycle: our group = 395, historical group = 81.
• 6 Data are expressed as mean ± SD or n (%).

Regarding the main endpoints of this study, LTZ combined with low-dose step-up regimen of Hp-HMG achieved an ongoing pregnancy rate of 28.24%. The ovulation rate was 97.7%, the percentage of mono-follicular development was 70.9% (Table 1). Clinical results for the pregnant and non-pregnant group were comparable, with significant higher numbers of dominant follicles in the pregnant group (Table 2).

The pregnancy rate was nearly as expected in the study protocol and very similar to recent studies [[13], [18]]. For our high efficacy the pretreatment of Diane-35 in patients with hyperandrogenism may be important to decrease LH concentrations and thus to increase the sensitivity for Hp-HMG [[19]]. Possible reasons for the high-pregnancy rate could be the good endometrial thickness and patterns on the day of GnRH triggering for facilitating implantation [[22]]. The percentage of unifollicular development was still high even in conjunction with HMG, which is consistent with the major advantage of LTZ in achieving limited number of mature follicles [[23]].

Besides, highly purified HMG contains 75 IU FSH and 10 IU hCG (1 IU hCG is equal to 6–7 IU LH). hCG binds to LH receptor with longer half-life than that of LH [[25]]. In addition, natural and high purity of Hp-HMG make up for the deficiencies of domestic HMG (more impurities of urine protein, urine pollution and instability of production batches) [[26]].

Compared with the study of Xi et al. [[16]] (Table 3), we achieved higher ovulation and clinical pregnancy rates. The percentage of cycle cancelation was significantly lower, and there was a trend toward a lower anovulation rate in our study. Our women received HMG low-dose starting with 75 IU and not more than 150 IU instead of 75 IU HMG on alternate days [[16]]. It should be mentioned that compared with the daily dose of 2.5 mg of LTZ (Xi et al. study), 5 mg is associated with more follicles and a higher pregnancy rate. It appears that 5 mg daily for 5 d is a preferable LTZ dose [[27]]. Unlike Xi et al., we used GnRH agonist for triggering final oocyte maturation instead of hCG, which might to a large extent prevent OHSS by triggering endogenous LH-release.

There were some limitations for our study, the first is that live birth rate could not be obtained because the pregnancies failed to be followed up to delivery, the second is that we only could perform an historical comparison with a non-randomized study because of ethical considerations. Further research comparing different regimens and protocols, if ethically possible in the same prospective study, certainly is wanted despite the success we already did have in many patients with the regimen and protocol used in this study.

In conclusion, our protocol of LTZ combined with low-dose Hp-HMG in CC-resistant infertile women with PCOS is effective and can reduce the treatment duration and by this also the costs. Furthermore, it is safe for ovarian induction and should be a good choice for the routine clinical practice.

This study was supported by Beijing Municipality Health Technology High-level Talent (2014–2-016); Project of Discipline Leader, Beijing Obstetrics and Gynecology Hospital, Capital Medical University, Beijing Maternal and Child Health Care Hospital (2013–1); Beijing Natural Science Foundation, China (Y161011).

All the authors declare that they have no conflicts of interest regarding this article.