Phenotype-specific efficacy of letrozole versus clomiphene citrate in polycystic ovary syndrome: a prospective cohort study.

In this observational study, letrozole was associated with higher ovulation rates than clomiphene citrate in PCOS patients with clinical and/or biochemical hyperandrogenism, suggesting a potential benefit in these specific subgroups.

A total of 276 infertile patients with PCOS were assessed for eligibility. Of these, 52 were excluded due to the presence of concomitant infertility etiologies, and 17 were excluded due to endocrinological disorders. As a result, 207 patients were included in the final analysis. Among them, 104 underwent OI with CC and 103 with LT. Table  1 presents a comparison between the CC and LT groups. There were no significant differences in age, BMI, or baseline laboratory parameters between the two groups. However, the mFG score was significantly higher in the CC group ( P  = 0.001), and a significantly greater number of patients had an mFG score > 8 in the CC group ( P  = 0.001). Cycle duration was comparable between the two groups. The ovulation rates were 77.9% in the CC group and 93.2% in the LT group, showing a statistically significant difference ( P  = 0.002). In contrast, the clinical and live birth rates were similar between the two groups (Table  1 ). Table 1 Demographic characteristics of the study population CC ( n  = 104) LT ( n  = 103) P Value Age, years 27.28 ± 4.94 27.53 ± 4.53 0.713 BMI, kg/m² 26.96 ± 4.68 26.63 ± 4.25 0.610 mFG score 17.87 ± 7.03 14.20 ± 8.74 0.001 mFG score > 8, n (%) 90 (86.5) 72 (69.9) 0.001 Serum AMH, ng/mL 7.83 ± 3.75 7.34 ± 3.93 0.370 tT, ng/dL 46.95 ± 18.56 46.23 ± 21.04 0.845 DHEA-S, mikrog/dL 313.35 ± 134.39 297.54 ± 127.94 0.511 FAI 5.57 ± 4.45 5.25 ± 4.10 0.697 TSH, mU/L 2.33 ± 0.97 2.43 ± 1.36 0.582 Biochemical HA, n (%) 37 (64.9) 36 (53.7) 0.207 Duration of follicular phase, days 13.47 ± 2.56 14.17 ± 3.20 0.087 Ovulation, n(%) 81 (77.9) 96 (93.2) 0.002  Phenotype A 29 (74.4) 27 (90.0) 0.100  Phenotype B 3 (75.0) 7 (87.5) 0.584  Phenotype C 38 (82.6) 32 (100.0) 0.013  Phenotype D 11 (73.3) 30 (90.9) 0.110 Clinical Pregnancy, n (%) 12 (11.5) 19 (18.4) 0.164 Live birth, n (%) 10 (9.6) 16 (15.5) 0.199 Values are given as mean±standard deviation or frequency (percentage) CC clomiphene citrate, LT letrozole, BMI body mass index, mFG Modified Ferriman-Gallwey, AMH anti müllerian hormone, tT total testosterone, DHEA-S dihydroepiandrostenodione-sulphate, FAI free androgen index, TSH thyroid stimulating hormone, HA hyperandrogenism Demographic characteristics of the study population Values are given as mean±standard deviation or frequency (percentage) CC clomiphene citrate, LT letrozole, BMI body mass index, mFG Modified Ferriman-Gallwey, AMH anti müllerian hormone, tT total testosterone, DHEA-S dihydroepiandrostenodione-sulphate, FAI free androgen index, TSH thyroid stimulating hormone, HA hyperandrogenism Demographic characteristics across PCOS phenotypes are presented in Table  2 . The highest mean age was observed in the Phenotype C group, while the lowest mean age was found in the Phenotype A group ( P  < 0.001). No significant difference was detected between phenotypes regarding BMI ( P  = 0.061). Among the four phenotypic groups, significant differences were observed only in terms of the type of OI agent used and LBR ( P  = 0.008 and P  = 0.033, respectively). While no live birth occurred in the Phenotype B group, the LBRs in Phenotype A, C, and D groups were 21.7%, 9.0%, and 8.3%, respectively. In the subgroup analysis, a significantly higher ovulation rate was observed with LT compared to CC only in the Phenotype C subgroup (100% vs. 82.6%, P  = 0.013). Table 2 Comparison of PCOS phenotypes A ( n  = 69) B ( n  = 12) C ( n  = 78) D ( n  = 48) P value Age, years 25.42 ± 4.13 28.50 ± 3.90 29.27 ± 4.70 27.19 ± 4.70 < 0.001 BMI, kg/m² 27.75 ± 5.55 28.31 ± 4.66 26.06 ± 3.41 26.06 ± 3.71 0.061 OI agent, n (%)  CC 39 (56.5) 8 (66.7) 46 (59.0) 15 (31.3) 0.008  LT 30 (43.5) 4 (33.3) 32 (41.0) 33 (68.7) Duration of follicular phase, days 13.81 ± 3.29 12.92 ± 2.88 13.71 ± 2.50 14.23 ± 2.98 0.529 Ovulation, n (%) 56 (81.2) 10 (83.3) 70 (89.7) 41 (85.4) 0.526 Clinical pregnancy, n (%) 15 (21.7) 0 (0,0) 10 (12.8) 6 (12.5) 0.164 Live birth, n (%) 15 (21.7) 0 (0) 7 (9.0) 4 (8.3) 0.033 Values are given as mean±standard deviation or frequency (percentage) BMI body mass index, CC clomiphene citrate, LT letrozole Comparison of PCOS phenotypes Values are given as mean±standard deviation or frequency (percentage) BMI body mass index, CC clomiphene citrate, LT letrozole In phenotypes A, B, and C—characterized by clinical and/or biochemical hyperandrogenism—comparison between the CC and LT groups revealed a significant difference only in ovulation response ( P  = 0.005), with rates of 78.7% in the CC group and 94.3% in the LT group. No significant differences were observed between the CC and LT groups among phenotype D patients ( P  > 0.05). Among patients with an mFG score > 8, the ovulation rate was significantly higher in the LT group compared to the CC group (94.4% vs. 78.9%, P  = 0.005), whereas no significant differences were observed in patients with an mFG score ≤ 8 ( P  > 0.05). In patients with an FAI ≥ 7, all outcome variables differed significantly between the two groups. The LT group showed superior outcomes, with a 100% ovulation rate compared to 50% in the CC group, a clinical pregnancy rate of 66.7% vs. 0%, and an LBR of 58.3% vs. 0%, respectively. Conversely, in patients with FAI  0.05). Among patients with total testosterone (tT) ≥ 48 ng/dL, the LT group exhibited significantly higher clinical pregnancy(24.0% vs. 0%) and LBR (20.0% vs. 0%) compared to the CC group ( P  < 0.05). No significant differences were observed in patients with tT  0.05) (Table  3 ). Table 3 Comparison of outcome parameters between CC and LT by phenotype, mFG, FAI and tT Phenotype A-B-C Phenotype D CC LT P value CC LT P value Ovulation, n (%) 70 (78.7) 66 (94.3) 0.005 11 (73.3) 30 (90.9) 0.110 Cl. Pregnancy, n (%) 11 (12.4) 14 (20.0) 0.189 1 (6.7) 5 (15.2) 0.410 Live birth, n (%) 10 (11.2) 12 (17.1) 0.284 0 (0.0) 4 (12.1) 0.159 mFG score ≤ 8 mFG score > 8 CC LT P value CC LT P value Ovulation, n (%) 10 (71.4) 28 (90.3) 0.105 71 (78.9) 68 (94.4) 0.005 Cl. Pregnancy, n (%) 1 (7.1) 3 (9.7) 0.782 11 (12.2) 16 (22.2) 0.090 Live birth, n (%) 0 (0.0) 2 (6.5) 0.331 10 (11.1) 14 (19.4) 0.138 FAI < 7 FAI ≥7 CC LT P value CC LT P value Ovulation, n (%) 36 (83.7) 46 (93.9) 0.118 4 (50.0) 12 (100) 0.006 Cl. Pregnancy, n (%) 7 (16.3) 6 (12.2) 0.579 0 (0.0) 8 (66.7) 0.003 Live birth, n (%) 7 (16.3) 5 (10.2) 0.388 0 (0.0) 7 (58.3) 0.007 tT < 48 ng/dL tT ≥48 ng/dL CC LT P value CC LT P value Ovulation, n (%) 27 (79.4) 39 (92.9) 0.085 19 (82.6) 24 (96.0) 0.129 Cl. Pregnancy, n (%) 7 (20.6) 8 (19.0) 0.867 0 (0.0) 6 (24.0) 0.012 Live birth, n (%) 7 (20.6) 7 (16.7) 0.661 0 (0.0) 5 (20.0) 0.023 CC clomiphene citrate, LT letrozole, mFG modified Ferriman-Gallwey, FAI free androgen index, tT total testosterone Comparison of outcome parameters between CC and LT by phenotype, mFG, FAI and tT CC clomiphene citrate, LT letrozole, mFG modified Ferriman-Gallwey, FAI free androgen index, tT total testosterone Regarding clinical pregnancy, LT treatment (aOR: 5.85, 95% CI: 1.47–23.34) and younger age (aOR: 0.80, 95% CI: 0.67–0.95) were identified as significant independent predictors. Interestingly, while unadjusted observations suggested a link between baseline FAI and clinical outcomes, the effect of FAI was attenuated and its estimate became imprecise in the adjusted model (aOR: 1.13, 95% CI: 0.95–1.33). This attenuation suggests that the strong, independent effect of LT heavily dictates the outcome, potentially mediating or overshadowing the prognostic value of baseline FAI. To address potential selection bias and baseline imbalances (specifically the higher mFG scores in the CC group), we performed multivariate logistic regression analyses (Table  4 ). The models were adjusted for age, BMI, mFG score, and FAI. For the primary outcome of ovulation, Letrozole treatment remained a robust independent predictor (Adjusted Odds Ratio [aOR]: 7.74, 95% CI: 1.72–34.74, p  = 0.008) after adjusting for all covariates. BMI was also found to be a significant negative predictor of ovulation (aOR: 0.82, p  = 0.012). The interaction between treatment and FAI was tested and found to be non-significant ( p  = 0.51). This non-significant interaction implies that the treatment effect of LT over CC does not statistically differ depending on the FAI level. In other words, LT demonstrates a consistently higher likelihood of inducing ovulation regardless of the underlying androgen status, and we cannot claim a statistically confirmed phenotype-specific superiority. Regarding clinical pregnancy, Letrozole treatment (aOR: 5.85, 95% CI: 1.47–23.34, p  = 0.012) and younger age (aOR: 0.80, p  = 0.011) were identified as significant independent predictors. Interestingly, while univariate analysis suggested a link with FAI, in the adjusted model, the effect of FAI was not statistically significant ( p  = 0.177), likely due to the dominant effect of the treatment agent itself. Table 4 Multivariate logistic regression analysis for ovulation and clinical pregnancy Outcome / Variable Adjusted Odds Ratio (aOR) 95% Confidence Interval P Value Ovulation  Treatment (Letrozole vs. CC) 7.74 1.72–34.74 0.008  BMI 0.82 0.70–0.96 0.012  mFG Score 1.09 0.99–1.19 0.077  FAI 1.04 0.86–1.25 0.705  Age 1.03 0.86–1.22 0.772 Clinical Pregnancy  Treatment (Letrozole vs. CC) 5.85 1.47–23.34 0.012  Age 0.80 0.67–0.95 0.011  mFG Score 1.10 1.00–1.20 0.041  FAI 1.13 0.95–1.33 0.177 Multivariate logistic regression analysis for ovulation and clinical pregnancy

In this single-centre prospective cohort study, all infertile patients diagnosed with PCOS according to the modified Rotterdam criteria who admitted to the Infertility Centre of Ankara University Cebeci Hospital between November 2023 and April 2024 were evaluated. The study protocol was approved by the Human Research Ethics Committee of Ankara University Faculty of Medicine with decision number I09-595-23. All infertile women diagnosed with PCOS according to the Rotterdam criteria and scheduled for OI were considered for the study. Inclusion criteria were: primary or secondary infertility of at least one year, age 18–40 years, normal uterine cavity with at least one patent fallopian tube confirmed by recent hysterosalpingography or laparoscopy, normal partner semen parameters in accordance with the current WHO reference limits [ 14 ]. Exclusion criteria were: untreated thyroid dysfunction, hyperprolactinemia or other endocrine causes of infertility, suspected tubal, uterine, or cervical factor infertility, known endometriosis and male-factor infertility documented by abnormal semen analysis. Participants who met the modified Rotterdam criteria for PCOS were prospectively assigned to OI with either LT (2.5 mg daily for 5 days, escalated up to 7.5 mg in cases of anovulation) or CC (50 mg daily for 5 days, increased up to 150 mg as needed) according to the primary physician’s discretion. At enrolment, baseline data including age, body mass index (BMI), reproductive history, comorbidities, and basal cycle day-3 hormone levels (FSH, LH, E2, AMH) were documented. In addition, each participant’s androgen profile (total testosterone, DHEA-S, SHBG), hirsutism score, partner semen parameters, and tubal patency (confirmed by hysterosalpingography) were recorded. PCOS phenotype was classified as either hyperandrogenic—defined according to the Rotterdam phenotypes A, B, or C, based on a modified Ferriman-Gallwey score > 8 and/or biochemical hyperandrogenism (FAI ≥ 7)—or non-hyperandrogenic, corresponding to phenotype D [ 3 , 7 ]. Total testosterone (tT), DHEA-S and SHBG were analyzed to demonstrate the presence of biochemical hyperandrogenism. The FAI was calculated using the formula: tT (nmol/lt)/SHBG (nmol/lt)x100, with ≥7 indicating biochemical hyperandrogenism. Based on this information, patients were divided into two groups: those with clinical and/or biochemical HA (phenotypes A, B and C) and those without (phenotype D). Treatment allocation to either letrozole or clomiphene citrate was not randomized but was determined solely by the treating physician’s clinical discretion. There were no predefined clinical criteria or strict institutional protocols guiding the choice of the ovulation induction agent. Consequently, allocation may have varied based on individual clinician preference rather than specific patient characteristics. Comparisons between LT and CC were performed within each HA-defined subgroup. All baseline clinical and hormonal data were collected prior to treatment allocation. No further randomisation was performed, enabling a direct comparison of drug efficacy across phenotypically distinct PCOS subpopulations. The treatment protocol involved administration of ovulation-inducing medications beginning on the third day following either spontaneous menstruation or progesterone-induced withdrawal bleeding. Throughout each treatment cycle, serial transvaginal ultrasound was used to monitor the number and diameter of dominant follicles, as well as endometrial thickness. The presence of a dominant follicle (> 18 mm) was recorded between cycle days 12 and 14. A follow-up ultrasound was performed on day 21 to assess signs of ovulation. Ovulation was confirmed by the presence of a corpus luteum in the ovary that had previously contained the dominant follicle. If a follicular response was observed, couples were advised to have timed intercourse (2–3 times per week). Serum β-hCG testing was performed either two weeks after detection of the dominant follicle or upon a missed period. If the ultrasound revealed more than three follicles measuring > 14 mm in diameter, the treatment cycle was cancelled to avoid the risk of multiple pregnancy. Participants with a positive pregnancy test underwent transvaginal ultrasound to confirm pregnancy viability. The primary endpoint was ovulation and the secondary outcomes included biochemical pregnancy (defined as a positive serum β-hCG), clinical pregnancy (intrauterine gestational sac with fetal cardiac activity ≥ 6 weeks), and live birth (delivery of a neonate showing signs of life at ≥ 24 weeks of gestation). SPSS 21.0 software (IBM Corporation, Armonk, NYC, USA) was used for data analysis. Samples were evaluated using the Shapiro-Wilk test to determine normality of distributions. Parametric tests were preferred depending on the results. Continuous variables were compared between two groups using Student’s t -test and between more than two groups using one-way ANOVA test. When statistical significance was obtained with the one-way ANOVA test, the source of significance was determined by the post-hoc Tukey test. Categorical variables were compared using the chi-squared test. Multivariate logistic regression models were constructed to determine the independent factors affecting the chance of live birth. A P value < 0.05 was considered statistically significant. Subgroup analyses concerning distinct PCOS phenotypes and hyperandrogenism markers were considered exploratory; therefore, no statistical adjustments for multiple comparisons were applied. The sample size calculation was based on a superiority hypothesis to detect a significant difference in ovulation rates between the Letrozole and Clomiphene Citrate groups. Based on previous literature, we assumed an ovulation rate of approximately 75% in the CC group. We powered the study to detect a clinically significant difference of 15% points (i.e., 90% in the LT group) with a two-sided type I error (α) of 0.05 and a power of 80% (1 − β). Under these parameters, the calculation indicated that a minimum of 96 participants per arm would be required. Anticipating a potential dropout rate of approximately 10%, we aimed to recruit at least 210 patients in total.

The present study demonstrated that LT was generally superior to CC in achieving ovulation in PCOS patients. Although subgroup analyses revealed that patients with higher androgen levels (FAI ≥ 7, tT ≥ 48) and hirsutism scores (mFG > 8) showed notably higher raw percentages of ovulation and pregnancy outcomes with LT, the lack of a significant interaction between treatment and FAI indicates that LT’s efficacy is consistent across varying androgen levels. Therefore, our findings suggest that LT serves as a highly effective first-line therapy overall, and while its clinical benefit appears particularly prominent in classic, hyperandrogenic PCOS patients, we cannot assert a statistically proven phenotype-specific superiority. Clinically, these observational findings suggest that LT is a highly effective option, particularly for classic, hyperandrogenic PCOS patients, whereas CC remains a viable and effective alternative, especially for the milder, non-hyperandrogenic (Phenotype D) subgroup. The results of many studies support the negative effects of excess androgen on oocyte development in PCOS patients. LT blocks estrogen production both peripherally and centrally by inhibiting aromatase, the rate-limiting step in estrogen production. It frees the hypothalamic-pituitary-ovarian axis from the effects of estrogenic negative feedback and increases gonadotropin secretion, which stimulates ovarian follicle development [ 15 ]. According to the ‘two triangle hypothesis’, excess androgens in PCOS cause granulosa cells to become hypersensitive to FSH and overproduce AMH, leading to abnormal follicular growth [ 16 ]. Hyperandrogenemia may enhance the effectiveness of letrozole by increasing the sensitivity of FSH receptors [ 17 ]. Multiple systematic reviews and meta-analyses have demonstrated the superiority of LT to CC. A 2022 Cochrane meta-analysis revealed that live birth rates were higher with letrozole compared to selective estrogen receptor modulators followed by timed intercourse (OR 1.72, 95% CI 1.40 to 2.11) [ 18 ]. A significant randomized controlled trial conducted by Legro and colleagues in 2014 examined 374 PCOS patients, with a notable demographic characteristic that two-thirds of participants had BMI exceeding 30. Their findings demonstrated LT’s superiority in both ovulation rates ( P < 0.01) and cumulative pregnancy outcomes (RR: 1.44, 95% CI: 1.10–1.87) [ 19 ]. Further evidence emerged from Roque’s 2015 meta-analysis, which evaluated five studies focusing on live birth outcomes. The analysis revealed significantly enhanced live birth rates in patients receiving LT (RR: 1.55, 95% CI: 1.26–1.90) [ 20 ]. A comprehensive 2023 meta-analysis by Liu and colleagues, encompassing 29 RCTs with 3,952 participants across 7,633 ovulation induction cycles, provided robust evidence supporting LT’s effectiveness. Analysis of 22 RCTs for ovulation rates, 28 for clinical pregnancy rates, and 8 for live birth rates demonstrated LT’s consistent superiority: ovulation rate (RR: 1.14, 95% CI 1.06–1.21, P < 0.001), CPR (RR 1.48, 95% CI 1.34–1.63, P < 0.001), and LBR (RR 1.49, 95% CI 1.27–1.74, P < 0.001) [ 21 ]. A recent individual participant data meta-analysis demonstrated that LT significantly improved LBR compared to CC (RR 1.43, 95% CI 1.17–1.75). This analysis found no significant interaction between treatment and variables such as age, BMI, type of infertility (primary/secondary), hirsutism, SHBG, or FAI in relation to live birth outcomes. However, it indicated that women with elevated baseline serum tT levels might derive greater benefit from LT compared to CC (RR 1.29, 95% CI: 1.01–1.65) [ 22 ]. Similarly, in our study, women with tT levels ≥ 48 exhibited significantly higher CPR and LBR when treated with LT. A randomized controlled trial by Zhang et al., assessing approximately 1000 infertile PCOS patients, revealed significant negative correlations between FAI levels and reproductive outcomes. Their analysis demonstrated that higher FAI levels were associated with reduced ovulation, pregnancy, clinical pregnancy, and live birth rates. Interestingly, total testosterone levels showed no significant correlation with clinical pregnancy and live birth rates. The researchers concluded that FAI serves as a sensitive predictor of fertility outcomes in PCOS patients, significantly influencing reproductive success across multiple parameters [ 23 ]. While previous studies, as well as our unadjusted subgroup data, indicate that higher FAI negatively impacts reproductive outcomes, our multivariate analysis yielded an imprecise estimate for FAI (95% CI: 0.95–1.33) when adjusted for the treatment agent. This discrepancy between univariate and multivariate findings implies that the potent ovulation-inducing effect of LT—which effectively overcomes the anovulatory barrier in hyperandrogenic patients—largely neutralizes the otherwise negative prognostic impact of elevated FAI. Thus, the choice of treatment heavily mediates the final clinical outcome, rendering baseline FAI a less critical independent predictor once LT is administered. In an RCT, Rausch et al. found that mFG scores negatively affected pregnancy, clinical pregnancy, and live birth rates. Both mFG scores < 8 and ≥ 16 (severe hirsutism) were associated with poor outcomes. However, hirsutism was reported to have no effect on ovulation in that study. Conversely, lower baseline FAI scores were associated with increased rates of ovulation, pregnancy, clinical pregnancy, and live birth rates. The cut-off level for FAI score in that study was 10 [ 24 ]. Similarly, Amer et al. compared CC and LT in an RCT and reported that LT was more successful, particularly in women with clinical and/or biochemical HA, although the difference was not statistically significant [ 25 ]. In our study, LT showed significantly better ovulation rates than CC in patients with hirsutism (mFG score > 8) and high androgen levels (FAI ≥ 7) ( P = 0.005). Additionally, patients with FAI ≥ 7 had higher clinical and live birth rates with LT treatment. However, we could not find significant differences between PCOS phenotypes, which remains a debated topic in the literature. It is important to note the distinction in our live birth outcomes: while overall live birth rates did not differ significantly between the two treatment arms in the general PCOS cohort, our exploratory subgroup analyses demonstrated that patients with severe hyperandrogenic features (FAI ≥ 7, tT ≥ 48) achieved higher live birth rates when treated with LT. The strengths of this study include providing valuable insights into the differential efficacy of potent ovulation induction agents across PCOS phenotypes and addressing a clinically important gap in personalized treatment approaches. The comprehensive analysis of multiple clinical and biochemical parameters, including detailed hormonal profiles and phenotypic classifications, strengthens the clinical applicability of our findings. The adequate sample size and the use of multivariate logistic regression analysis enhance the statistical robustness of the results. Furthermore, subgroup analyses based on androgen levels and hirsutism scores provide practical guidance to clinicians in selecting appropriate first-line treatment for individual patients. A primary limitation of our study is the non-randomized, physician-directed nature of treatment allocation. Without randomization or predefined assignment criteria, there is an inherent risk of selection bias. Unmeasured clinical confounders or individual physician preferences may have influenced the choice of the ovulation induction agent, which likely contributed to baseline imbalances between the groups, such as the significantly higher mFG scores observed in the CC arm. Although we employed multivariate logistic regression models to adjust for these baseline differences and mitigate potential confounding, the possibility of residual confounding cannot be entirely excluded. Therefore, our findings should be interpreted as associative rather than definitively causative. Additionally, we conducted multiple subgroup analyses based on phenotypes, mFG scores, FAI, and tT levels without adjusting for multiple comparisons, which inherently increases the risk of Type I error. Consequently, our subgroup findings—particularly those involving small sample sizes, such as the Phenotype B group ( n  = 12) and the FAI ≥ 7 subgroup—must be interpreted cautiously. These specific findings are exploratory and hypothesis-generating, highlighting the need for validation in future studies with larger cohorts. One of the limitations of this study is that it was conducted in a single center, which may affect the generalizability of the findings to different populations and clinical settings. While we assessed live birth rates, we did not evaluate longer-term pediatric outcomes or detailed late-pregnancy complications, which are also important endpoints in fertility treatment. The relatively small sample size in some phenotype subgroups may have limited our ability to detect important differences. In conclusion, the findings of this prospective observational study indicate that in infertile PCOS patients with an mFG score > 8, ovulation rates were higher with letrozole compared to CC. Furthermore, among patients with elevated FAI or total testosterone, letrozole was associated with more favorable outcomes, including higher ovulation, clinical pregnancy, and live birth rates. However, given the non-randomized, physician-directed design and the exploratory nature of these subgroup analyses, these results should be interpreted as associative rather than definitive. While letrozole appears to be a highly effective ovulation induction option, particularly for hyperandrogenic patients, further randomized controlled trials with larger cohorts are needed to establish prescriptive treatment algorithms across different PCOS phenotypes.

Among reproductive-age women, polycystic ovary syndrome (PCOS) represents the most prevalent endocrine disorder, affecting approximately 8% to 18% of this population [ 1 ]. The condition’s diagnosis relies on the Modified Rotterdam Criteria-2003, requiring the presence of at least two out of three defining characteristics, namely hyperandrogenism (HA, either clinical or biochemical), ovulatory dysfunction (OD), and polycystic ovarian morphology (PCOM) [ 2 ]. In 2012, the National Institutes of Health (NIH) introduced a classification system based on distinct phenotypic presentations. This framework identifies four distinct manifestations: Phenotype A - characterized by the concurrent presence of HA, OD, and PCOM; Phenotype B – distinguished by the combination of HA and OD; Phenotype C – marked by HA and PCOM; Phenotype D – defined by OD and PCOM [ 3 ]. Hirsutism, characterized by excessive male-pattern hair growth on the face or body, affects 5–10% of women in reproductive years [ 4 ]. Over 85% of hirsutism cases are attributed to PCOS or idiopathic hyperandrogenism [ 5 ]. The modified Ferriman-Gallwey (mFG) scoring system is commonly used to diagnose clinical hirsutism, with a threshold of > 8 established for Türkiye [ 6 , 7 ]. Biochemical hyperandrogenism is identified when total or free testosterone levels exceed normal ranges. The free androgen index (FAI), which accounts for abnormalities in SHBG, is considered a more reliable measure of hyperandrogenism [ 8 ]. Beyond reflecting free testosterone levels, FAI also serves as an indicator of insulin resistance and glucose metabolism has been shown to play a pivotal role in endometrial receptivity [ 9 – 11 ]. Ovulation induction (OI) serves as a key treatment strategy, with clomiphene citrate (CC) historically regarded as the first-line therapy [ 12 ]. However, recent guidelines highlighted letrozole (LT) as a superior alternative, citing its association with higher ovulation and live birth rates and recommended it as the preferred first-line OI treatment for PCOS patients [ 13 ]. Currently, there is limited research examining how clomiphene citrate and letrozole perform in ovulation induction across different PCOS phenotypes, particularly regarding the presence and severity of hyperandrogenism. The aim of this prospective observational study is to evaluate ovulation and live birth rates (LBR) in PCOS patients undergoing OI with CC and LT, considering variations across PCOS phenotypes and hyperandrogenism parameters.