A fertility trial can show differences in stimulation requirements without demonstrating a difference in the outcome patients ultimately seek. Primary report
Study snapshot
One hundred infertile women were randomized to highly purified hMG or recombinant FSH during a long suppression protocol for ICSI. The hMG arm had shorter stimulation and lower gonadotropin requirements, with differences in several hormone measurements. Pregnancy, cancellation, and miscarriage comparisons were not statistically different in the reported study. Primary report
Our appraisal: separate process measures from clinical success
Treatment duration and medication requirements can matter to burden and cost. They should not, however, be relabeled as a proven improvement in pregnancy or live birth. Each claim needs the corresponding endpoint and denominator.
Pregnancy per started cycle and pregnancy per embryo transfer answer different questions. The latter excludes people who never reached transfer, so it cannot by itself describe the experience of everyone beginning treatment. A comparison should keep cancellation and progression through the treatment pathway visible.
Nonsignificant pregnancy results also do not establish formal equivalence. That requires a suitable design, a justified margin, and sufficiently precise estimates. A small trial may leave clinically important differences unresolved even when its percentages look similar.
The treatment setting matters: an ICSI protocol with a particular suppression approach is not interchangeable with every assisted-reproduction cycle. Full review should examine allocation, blinding, ovarian response, adverse events, and whether the study was designed around process or reproductive outcomes. The report is not a basis for choosing or administering fertility medicines without specialist care.
Frequently asked questions
What assisted-reproduction setting was studied?
ICSI following a long suppression regimen.
Did stimulation requirements differ?
Yes. The report favored hMG on duration and medication requirements.
Did that establish more pregnancies?
No statistically significant pregnancy difference was reported.
Is per-transfer success the same as per-started-cycle success?
No. Their denominators differ.
Does nonsignificance prove equivalence?
No.
Can the result be transferred to every stimulation protocol?
Not without examining the protocol differences.
What is the key review task?
Match each claimed advantage to its actual endpoint, uncertainty, and denominator.