How to Read IVF Success Rates

The same clinic, the same year, the same patients can honestly publish very different success rates depending on what sits under the line. There are four denominators in common use. The first three get larger as you move down the list, and the rate falls as they do. The fourth counts people rather than cycles and adds up every attempt, so it answers a different question:

• Per embryo transfer. Only patients who reached a transfer are counted. Everyone whose cycle was cancelled, whose retrieval produced no eggs, or whose embryos never made it to transfer has been removed. This is the most flattering number and the one most often printed in marketing.
• Per egg retrieval. Counts everyone who reached retrieval, including those who never got to a transfer.
• Per cycle started (per intended retrieval). Counts everyone from the moment stimulation began, including cancelled cycles. This is closest to the question a patient is actually asking.
• Per patient, cumulative. Counts whether a person eventually had a live birth across all transfers from one or more retrievals.

When a clinic quotes you a figure, the first question is not “how high is it?” but “out of what?” A 60% rate per transfer and a 35% rate per cycle started can describe an identical set of patients.

A positive pregnancy test is not an outcome anyone is seeking. Between a positive test and a baby sit biochemical pregnancies, ectopic pregnancies, and miscarriage. Rates quoted as clinical pregnancy or positive beta hCG will therefore always exceed the live birth rate for the same group.

Ask for live birth rate, and for singleton live birth rate where it is available. Twin and higher-order pregnancies carry materially higher risk for both the pregnant person and the babies, so a clinic reporting a high overall live birth rate driven by multiple-embryo transfers is not obviously doing better than one with a lower rate and a strict single-embryo transfer policy.

With a patient's own eggs, age at retrieval is the strongest single predictor of live birth, and the decline is steep rather than gradual through the late thirties and into the forties. The mechanism is egg quality — the proportion of eggs that are chromosomally normal falls with age — which is why success rates with donor eggs from younger donors stay comparatively flat across recipient age.

Two consequences follow. First, any rate quoted without an age band is close to meaningless, because it reflects the age mix of that clinic's patients as much as anything else. Second, if a published figure blends donor-egg and own-egg cycles, it will overstate what a patient using her own eggs should expect. National reporting separates these categories for exactly this reason.

This is the part clinics rarely volunteer. Both the U.S. Centers for Disease Control and Prevention, which publishes the national ART data, and the Society for Assisted Reproductive Technology, whose member clinics report it, caution against comparing clinics on these numbers. The CDC says such comparisons “may not be meaningful” because clinics have different patient populations and treatment methods, and SART notes that patient characteristics vary among clinics.

A large part of the reason is which patients a clinic accepts and how it manages them. A clinic can raise its published numbers without improving its laboratory or its clinicians at all by declining patients with poor prognosis, cancelling more cycles before retrieval, transferring only the most promising embryos, or steering difficult cases toward donor eggs. A program attached to an academic centre that accepts referrals nobody else will take is penalised by the same statistic.

What the national data is genuinely useful for is orientation — what outcomes look like nationally for your age band and diagnosis — and for spotting a clinic whose reported volume or results are wildly out of line with everyone else.

Very few people decide in advance to do exactly one cycle. Treatment is usually a sequence: a retrieval, a fresh or frozen transfer, and if that fails, another transfer from the same batch of embryos, and possibly another retrieval.

Cumulative live birth rate per retrieval counts every transfer that came from one egg collection, including the frozen ones. Because of that, it is substantially higher than the rate from a single transfer, and it keeps rising across two and three retrievals for many patients. If you are deciding whether to begin at all, this is the number that answers your question. If a clinic cannot produce it for a patient like you, that itself is information.

Treatment add-ons — additional tests, techniques, or supplements offered on top of standard IVF, usually at extra cost — are frequently sold on the promise of better success rates. For many of them the evidence that they increase live birth rates for a general patient population is weak or absent, and the UK regulator publishes a public traffic-light rating of add-ons for that reason.

Before paying for one, ask three questions: what is the evidence that this raises live birth rates, not just implantation or embryo grading; does that evidence apply to someone with my history; and what does it cost if it is billed separately when the cycle fails?

• “For patients my age, with my diagnosis, using my own eggs — what is your live birth rate per cycle started?”
• “What is your cumulative live birth rate per retrieval, including frozen transfers?”
• “What proportion of cycles you start are cancelled before retrieval, and why?”
• “What proportion of retrievals result in at least one usable embryo?”
• “Is your single-embryo transfer the default, and what is your multiple-birth rate?”
• “Do these figures include donor-egg cycles?”
• “How many cycles like mine did you do last year?”

Ask for the answers in writing, alongside the itemised costs. A clinic that answers these calmly and specifically is telling you something useful about how it will communicate when a cycle does not go to plan.