Almost nobody has studied it. The only randomized comparison in men enrolled 25 people, split two ways, and ran 24 weeks [1]. On semaglutide the median proportion of morphologically normal sperm rose from 2% to 4%. Both figures sit below the reference range. In women the pooled randomized evidence is four trials and 176 participants, all in PCOS, and no trial has measured conception [4]. Nobody on this roster sells anything for fertility.
Men with obesity and type 2 diabetes often have low testosterone, and the standard fix is to give them testosterone. That fix has a known cost: it suppresses the hormonal signal that tells the testes to make sperm. A man who wants both the symptom relief and the fertility is stuck. This trial asked whether a weight-loss drug could be the way out.
The trial in men
Twenty-five men with type 2 diabetes, obesity and functional hypogonadism were randomized. One arm took semaglutide 1 mg a week. The other took intramuscular testosterone undecanoate 1000 mg every ten to twelve weeks. It ran 24 weeks, open-label. Median age was 50 and median BMI 35.9. Semen analysis and hypogonadism measures were taken at baseline and at the end.
Baseline semen quality was poor in both arms, below the 5th percentile of reference values. That matters for reading everything that follows. These were not men with normal parameters.
What changed
In the semaglutide arm, the median proportion of morphologically normal sperm rose from 2% to 4%, p=0.012. In the testosterone arm, sperm concentration and total sperm number both fell significantly. Head to head at 24 weeks, the semaglutide group had significantly more morphologically normal sperm, higher concentration and a higher total count.
Both arms saw total testosterone rise and both improved on the Aging Symptoms scale. Only the testosterone arm improved significantly on erectile function.
What the numbers actually say
The within-arm change on semaglutide is the part that stands on its own, and it is small. A median of 2% morphologically normal sperm became 4%. Doubling sounds substantial. Two percentage points is a different impression of the same fact. Both numbers sit below the reference range.
Twenty-five men, split two ways, is twelve or thirteen per arm. Open-label means everyone knew which drug they were on, and the two regimens are impossible to confuse. The design could not have blinded this. A placebo arm would settle it, and there was not one. Those are the same questions worth asking of any small uncontrolled study, and here at least there was randomization and a comparator.
Where a GLP-1 sits among the alternatives
A network meta-analysis pooled 23 randomized trials covering 1,899 men with obesity-related low testosterone, searched to April 2026 [2]. GLP-1-based therapy was one of the five interventions in that network. It does not appear among the results the abstract reports as significant, and the abstract gives no estimate for it either way.
Testosterone replacement plus structured lifestyle therapy raised total testosterone most, MD 7.19 (95% CI 1.18 to 13.21). It was also the only strategy that significantly improved erectile function. No intervention significantly reduced glycated hemoglobin. The full comparison is in the low testosterone network.
What the animal work separates
Every human study here runs into the same confound: the people taking the drug are also losing weight. A rat study used a pair-fed control to break that apart [3]. One group was fed exactly what the treated group chose to eat.
In male rats, both the drug and the matched restriction improved sperm motility, which points at the weight rather than the molecule. Restriction lowered androgens and semaglutide blunted that fall. In females, progesterone, pregnenolone and estradiol fell in a way the diet alone did not produce. Nobody measured whether any animal conceived, and four weeks in a rodent is not a year in a person. The design and its limits are in the pair-fed rat study.
The evidence in women
Four randomized trials, 176 participants in total, pooled for women with PCOS and obesity [4]. Waist circumference fell 5.16 cm and BMI 2.42 points. Total testosterone fell 1.33 (95% CI −2.55 to −0.12). HOMA-IR, the marker closest to the reason these drugs are supposed to help PCOS, did not move significantly.
No pregnancy or live-birth outcome was pooled. Weight loss genuinely helps many PCOS symptoms, and that does not require this literature. What the trials do not establish is that the drug treats the underlying endocrinology. That reading is set out in the PCOS pooled analysis.
What a buyer should do with this
Nothing on this roster is sold for fertility, and no seller claims it. If fertility is the actual goal, that is a reproductive endocrinologist’s question and semen analysis is the measurement. What this trial supports is narrower. A man is already facing a choice between these two treatments. The semen parameters point one way and the erectile function score points the other.
The broader sexual-function evidence is thin and covered separately. What sellers say about any of it is close to nothing, which the census of silence counts directly.