Letrozole for PCOS: Why It's First-Line, and Success Rates

Why Is Letrozole First-Line for Ovulation Induction in PCOS?

Live birth rates were higher with letrozole than with older ovulation-inducing drugs across a pooled analysis of 41 randomized trials involving 6,522 women with anovulatory PCOS, with high-certainty evidence putting the improvement at roughly 27% to 35% live births on letrozole against a 20% baseline on the older drugs — a number needed to treat of 10, meaning roughly one additional live birth for every ten women switched from the older regimen to letrozole. That scale of pooled evidence, not a single standout trial, is what moved the 2023 international evidence-based guideline — alongside the American Society for Reproductive Medicine (ASRM) and the European Society of Human Reproduction and Embryology (ESHRE), both contributing bodies to that guideline — to name letrozole the first-choice medication for ovulation induction in PCOS. That position is not universal: NICE CG156, which the NHS in the UK follows, does not address aromatase inhibitors at all and still lists clomiphene as the standard oral option, so a UK clinician offering clomiphene first is following current national guidance correctly, not lagging behind the evidence above.

Letrozole is an aromatase inhibitor, originally developed for hormone-receptor-positive breast cancer, used off-label for ovulation induction since the early 2000s. Its mechanism is straightforward at the level of physiology, whatever the details of an individual regimen: it briefly blocks the conversion of androgens to estrogen, the resulting temporary drop in estrogen removes a signal that normally suppresses the brain’s release of follicle-stimulating hormone, and the hypothalamic-pituitary axis responds by releasing more FSH — the hormone that drives a follicle to mature. That mechanism, not a specific dose or day-by-day regimen, is the reason it works differently from the older SERM-class drugs it has largely replaced. The guideline recommends discussing the specific regimen, monitoring, and any adjustments with a clinician directly rather than following a fixed plan found online.

Note: in May 2026, PCOS was renamed polyendocrine metabolic ovarian syndrome, or PMOS, by a global consensus of more than 50 organisations. Same condition, same reproductive mechanism — only the label changed. This article uses PCOS, since that is still the term most readers search.

What Are the Actual Letrozole Success Rates in PCOS?

Letrozole produced a live birth in 27.5% of participants compared with 19.1% on clomiphene, across up to five treatment cycles, in a landmark trial of 750 women aged 18–40 with PCOS diagnosed by Rotterdam criteria — the single trial most responsible for changing which drug guidelines recommend first. Every participant in that trial had already been confirmed to have a patent fallopian tube, a normal uterine cavity, and a partner with a sperm concentration of at least 14 million per millilitre, which matters for reading the number correctly: it describes success once other major causes of infertility have been ruled out, not a general population starting point.

Read together, three independently designed sources — one randomized trial, one pooled meta-analysis of 41 trials, and one large real-world claims-based study — arrive at the same direction of effect: letrozole outperforms clomiphene for live birth in PCOS specifically. None of them describe a guaranteed outcome. In the Legro trial, the majority of women in both arms did not have a live birth within the study period, which is the honest context every one of these percentages needs. The trade-off worth weighing alongside those numbers is what letrozole actually feels like to take — the trial-reported fatigue and dizziness rates read differently from the hot flushes more commonly associated with clomiphene.

Letrozole vs Clomiphene: What Made the Older Drug Fall Out of First Place?

Beyond the live-birth gap, letrozole carried an additional advantage the Cochrane review specifically measured: ovarian hyperstimulation syndrome occurred at the same low rate — 0.5% — in both the letrozole and SERM groups, so the improvement in live birth did not come with an added safety trade-off on that particular measure. The fuller head-to-head comparison of how the two drugs work, and why guidelines reordered their recommendation, is covered in depth here; this page focuses on what is specific to letrozole itself. Letrozole also avoids clomiphene’s endometrial-thinning side effect, a mechanism-level difference separate from the live-birth numbers themselves.

Letrozole also carried a safety controversy clomiphene never had to answer. A conference abstract presented in 2005 raised concern that letrozole might increase the risk of congenital malformations, and the resulting caution shaped clinical practice for years afterward even though the concern was never confirmed in a full peer-reviewed study. A 2021 systematic review and meta-analysis of 46 studies covering roughly 4,700 babies conceived on letrozole found a congenital malformation rate of 2.15% overall, no significant increase compared with clomiphene in the 14 available randomized trials, and a small reduction in cohort studies — concluding there is no evidence to support restricting letrozole on fetal-safety grounds. The same review rated the certainty of that evidence as low to moderate, and calculated a fragility index of 44%, meaning a modest shift in a handful of outcomes could change the statistical conclusion — a genuinely reassuring finding, held with appropriate caution rather than treated as the final word.

Does Real-World Data Tell the Same Story as the Trials?

Not entirely, and this is the finding on this page most worth reading carefully — and reading precisely, because the two halves of it come from different analytic populations within the same study. The pregnancy and live-birth advantage above (43% vs 37%; 32% vs 29%) is an intention-to-treat result: everyone who started letrozole compared with everyone who started clomiphene, whichever strategy they actually stayed on. The adverse-outcome figures are different — they come from a per-protocol analysis, restricted to the people who actually completed the assigned strategy as specified. In that narrower group, several risks were higher with letrozole: multiple-pregnancy rates across PCOS fertility treatments put this specific comparison at 19% of letrozole pregnancies versus 9% on clomiphene, preterm birth in 20% versus 15%, low birth weight for gestational age in 5% versus 3%, NICU admission in 22% versus 16%, and congenital malformation among live births in 8% versus 2%. A per-protocol comparison carries its own selection bias — the women it includes are, by definition, the ones who did not deviate from their assigned drug, which is not necessarily a random slice of everyone who started it. That pattern still runs counter to what the randomized trials above would predict, and it deserves to be stated plainly, population differences and all, rather than smoothed into a single number.

The study’s own authors flag the likely explanation: much of the difference tracked with a higher rate of multiple gestation in the letrozole group, and they could not fully rule out residual confounding by body mass index or how long someone had been trying to conceive before starting treatment — real limits of an observational design built on insurance claims rather than randomized assignment. It is also a genuinely different kind of evidence than the trials above: a real-world claims database capturing how the drug performs when prescribed in ordinary practice, compared with a controlled trial population that had already been screened for other infertility causes. Both kinds of evidence are worth holding at once rather than picking whichever one is more convenient.

What Does “Letrozole Not Working” Actually Mean?

Two different problems get called “letrozole not working,” and they call for different next questions with a clinician rather than a home adjustment. The first is a failure to ovulate at all on the medication — no BBT rise, no confirmatory progesterone signal, no LH surge — which is what confirming ovulation is actually for and worth ruling in or out before assuming the drug has failed, especially since the ovulation-timing window trials have observed on letrozole spans close to two weeks between women, so a cycle that looks slow is not automatically one that has failed. The second is ovulating reliably on letrozole but not conceiving, which is a different and, statistically, a far more common outcome: even in the Legro trial’s best-performing arm, fewer than three in ten women had a live birth, meaning most ovulatory cycles on letrozole simply did not result in a pregnancy that cycle — a normal feature of fertility treatment, not evidence that something is wrong.

For genuine non-response — no ovulation across several monitored cycles — the guideline-recognized next questions are about escalation, not self-adjustment: a 2025 meta-analysis of six randomized trials in 619 women with PCOS found that combining letrozole with a second ovulation-inducing medication improved ovulation rates over letrozole alone in some regimens, though the pregnancy-rate difference did not reach statistical significance, and a separate meta-analysis found that adding gonadotropin therapy after letrozole raised ovulation rates as well, again without a clear pregnancy-rate benefit. Both are findings to bring to a clinician who can weigh them against your specific case, not a sequence to run through independently — the 2023 international guideline frames second-line options explicitly as decisions made with a specialist once first-line treatment has had a fair, monitored trial.

Who Letrozole Does Not Help

Letrozole treats anovulatory infertility specifically — it will not help someone whose infertility is due to a blocked fallopian tube, a uterine structural issue, or a male-factor sperm problem, all of which the Legro trial screened out before enrolling anyone. It also is not a treatment for irregular cycles on its own outside of an active attempt to conceive; the guideline reserves it for ovulation induction, not general cycle regulation, since it carries a real chance of multiple gestation and is not indicated for someone who is not currently trying to become pregnant.

It is also not a guaranteed path to pregnancy even for the population it is designed for. In every trial cited on this page, most treated cycles — and in the largest trial, most treated women even after up to five cycles — did not result in a live birth. That is not a personal failure or a sign the medication was wrong for that person; it is the honest base rate of ovulation-induction medication, which is why the guideline frames a monitored trial of several cycles, followed by a specialist conversation about next steps if needed, as the standard approach rather than an open-ended one.

Common questions