IVF beat surgery for pregnancy in endometriosis, and the trial stopped early to say so

For a woman with endometriosis who is struggling to conceive, the choice has long been murky: have surgery to remove the disease, or go straight to in vitro fertilisation (IVF). Surgery can relieve pain and was often needed just to confirm the diagnosis, but it carries operative risk and can damage the ovaries. A trial published on 24 October in the American Journal of Obstetrics and Gynecology set out to compare the two paths head-to-head — and was stopped before it finished because the answer had already become clear [s1].

Why the question is newly answerable

The rationale for operating first has partly dissolved. Ultrasound has become accurate enough to diagnose ovarian endometriomas and deep infiltrating endometriosis without cutting, so women no longer need surgery simply to know what they have [s1]. That leaves a genuine fork: operate, or embark directly on IVF. No randomised trial had squarely tested this scenario, which is why the Italian investigators built one [s1].

How the trial worked

The study was run at three Italian tertiary referral centres with expertise in both endometriosis surgery and IVF, and it used an unusual design — a partially randomised patient preference trial [s1]. Women willing to be randomised were assigned to surgery or IVF by chance; women with a strong preference were allowed to choose, and were followed in a separate cohort [s1]. That structure captures both the clean comparison of randomisation and the real-world behaviour of patients who will not leave the decision to a coin toss. Women who had already had endometriosis surgery or a previous round of IVF were excluded [s1].

Recruitment began in March 2021 and was discontinued in September 2024, in line with the ethics committees, after 130 women had enrolled — 55 in the randomised cohort and 75 in the preference cohort [s1]. The trial was stopped early because investigators had become increasingly aware of a markedly higher likelihood of success among women allocated to IVF [s1].

What it found

In the randomised cohort, live births occurred in 7 of 27 women assigned to surgery (26%) and 16 of 28 assigned to IVF (57%) [s1]. The difference was statistically significant (p = 0.019), and the relative chance of a live birth with IVF was 2.20 (95% CI, 1.08 to 4.50) — roughly double [s1].

The preference cohort, where women chose their own route, pointed the same way and more strongly: live births in 7 of 25 women who had surgery (28%) versus 32 of 50 who chose IVF (64%), with p = 0.003 [s1]. That two independent groups — one randomised, one self-selected — produced the same direction and a similar gap is the most persuasive feature of the result, because the biases of each design run in different ways yet converge on the same conclusion.

How much it settles

The trial's great strength is that it is randomised evidence on a question that previously rested on observational data and clinical habit. Its limits are equally plain. It is small: 55 women in the randomised arm means the live-birth figures rest on a few dozen pregnancies, and the confidence interval around the headline benefit stretches from barely above one to 4.5 [s1]. Stopping a trial early for a favourable result can also exaggerate the effect size, because trials halted at a lucky interim tend to overstate benefit. And the work was done at centres skilled in both approaches, so the balance might shift where surgical or IVF expertise is less evenly matched [s1].

What the trial does not address is pain. Surgery is often chosen to relieve endometriosis symptoms, not only to improve fertility, and a study focused on live births says nothing about which path leaves a woman feeling better in the years afterward [s1]. For many women the decision will still turn on symptoms, ovarian reserve and age as much as on pregnancy odds. Nor does the trial close off surgery entirely: some women will have disease — severe pain, large endometriomas, bowel or bladder involvement — where an operation is warranted on its own merits, and fertility is only one part of their care.

The two routes also carry different burdens and risks that live births alone do not capture. Surgery means an operation and recovery, with a real chance of reducing ovarian reserve when cysts are removed; IVF means hormonal stimulation, egg retrieval and, in endometriosis specifically, a disease that can complicate the process. A full comparison would weigh those harms, costs and the time each path takes to a pregnancy, none of which a live-birth headline conveys [s1].

Still, for the specific woman weighing an operation purely to improve her chance of a baby, this is the first randomised signal that going straight to IVF may serve her better — clear enough that the investigators felt they could not keep randomising women to surgery in good conscience [s1].