Testosterone and GH with hMG raised pregnancy rates in low responders

The new evidence comes from two retrospective chart reviews in people and one controlled study in sheep. The table below gives the headline for each.

Study	Setting	Finding that stood out
Bhayana 2026	IVF patients with low ovarian reserve	Adding testosterone gel and growth hormone to hMG raised clinical pregnancy in the under-35 group 1
Kothekar 2026	IVF patients, day 5 or 6 blood draw	rFSH produced a larger rise in blood FSH than hMG 2
Zhang 2026	Kazakh ewes, superovulation	hMG produced more viable embryos than FSH 3

Retrospective means researchers looked back at existing medical records rather than assigning treatments at random. This format can reveal patterns, but it cannot rule out other explanations for a difference. All three reports measure different endpoints, so they are best read as separate signals, not as a single combined result.

In the Bhayana review, the research team looked back at 87 ovarian stimulation cycles from 2021 to 2025 at one clinic ¹. The patients were classified as POSEIDON Group 3 or 4, research labels for women with a low ovarian reserve. Group 3 means under age 35, and group 4 means older than 35; both had an anti-Müllerian hormone, or AMH, level below 1.2 ng/ml ¹. AMH is a blood marker that gives a rough idea of how many eggs remain.

One group received hMG alone. The other group received hMG plus testosterone gel and growth hormone ¹. In POSEIDON Group 3, the clinical pregnancy rate was 78.9% with the add-ons and 37.5% with hMG only. The difference was statistically significant, with a p value of 0.008; that means the result is unlikely to be a chance finding ¹. The study also reported a trend toward more transferable embryos, but that trend was not statistically significant ¹.

The abstract reports the significant gain only in POSEIDON Group 3, not in Group 4. So the finding is clearest for the younger subgroup.

Kothekar and colleagues compared blood levels of follicle-stimulating hormone, or FSH. FSH is the main signal that tells ovarian follicles to grow ⁴. One group received hMG, and the other received a lab-made recombinant FSH, called rFSH ². hMG is a urine-derived mixture that contains FSH and luteinizing hormone-like activity, including hCG ⁴.

To make the groups comparable, the researchers used propensity score matching, a method that balances patients by age, body mass index, protocol type, and gonadotropin dose. After matching 68 patients in each group, both groups had the same baseline FSH of 7.0 IU/L ². By day 5 or 6 of stimulation, the rFSH group averaged 19.6 IU/L, while the hMG group averaged 15.5 IU/L. The rise from baseline was 12.6 IU/L with rFSH and 7.6 IU/L with hMG, a significant difference ².

This does not mean one product was more effective. It means blood FSH levels during stimulation depend on the product used, so researchers should not compare those values across different gonadotropins as if they measure the same thing. The study did not report live birth rates ².

In the animal study, researchers treated Kazakh ewes with either hMG or follicle-stimulating hormone to produce multiple eggs for embryo transfer ³. Sheep do not respond exactly like humans, but the controlled design lets the team compare embryo numbers and quality.

Eleven ewes received hMG, and thirty-five received FSH. The hMG group produced an average of 5.36 viable embryos, compared with 2.91 in the FSH group. The difference was statistically significant ³. Blood progesterone also differed between the two groups, with a significant p value of 0.01 ³. A metabolomics scan, which measures many small molecules in blood at once, found three small molecules that stood out between the groups. The authors concluded that hMG improved embryo quality by changing small-molecule signals tied to embryonic development ³.

Earlier meta-analyses have found that hMG and recombinant FSH lead to similar live birth rates in IVF ⁶. Some studies also suggest that hMG may carry a lower risk of ovarian hyperstimulation syndrome, a serious over-response of the ovaries ⁶. The new blood-level study does not change that picture. It mainly cautions against comparing day 5/6 FSH levels across products as if they mean the same thing ².

The poor-responder study is not a test of hMG by itself. It tests whether adding testosterone gel and growth hormone to an hMG base changes outcomes in a selected group. The result is notable, but it comes from a small, retrospective chart review at one clinic, not a randomized trial ¹.

The sheep study is in a different species, so it mainly informs animal superovulation research. It does, however, add to the broader question of whether hormone mixtures with LH activity can influence embryo quality, a topic also studied in people ³.

All three reports have limits. The two human studies looked back at records rather than assigning patients randomly, so hidden differences between groups could influence the results ¹. The sheep study had a small hMG group of only eleven ewes, so the average embryo count may not be stable ³. The human add-on finding is limited to POSEIDON Group 3 in the abstract, so it may not apply to older patients or different protocols ¹.

The studies do not report live birth rates, so the downstream outcome is still open ¹.

We still need a randomized trial to test whether adding testosterone gel and growth hormone to hMG truly raises pregnancy rates in poor responders, and whether any gain holds across age groups and different clinics. The current evidence is a retrospective pattern, not proof of cause ¹.

We do not know whether the lower mid-stimulation FSH level with hMG matters for the number of eggs, embryo quality, or live birth, because the study stops at a blood measurement ².

The sheep embryo finding has not been replicated in large human IVF studies, so it should be treated as exploratory. We also do not know whether the metabolomic markers identified in sheep would appear in women undergoing ovarian stimulation ³.