Isotretinoin use is associated with persistent alterations in male reproductive function
Fie Bendix; Lærke Priskorn; Niels Jørgensen; Hanne Winther Frederiksen; Anne Jørgensen; Lina Steinrud Mørch; Hans Christian Ring; SriDurgaDevi Kolla; Anna K. Rosenmai; Terje Svingen; Anders Juul; Anders Rehfeld
medRxiv2026
In plain terms
Researchers in Denmark studied about 5,200 young men who gave semen and blood samples at one hospital visit, using national prescription records to see who had taken the acne drug isotretinoin. The 322 men who had taken it had slightly lower levels of FSH, a hormone that drives sperm production, and slightly higher estradiol, a form of oestrogen, including men who had stopped at least two years before; their sperm counts were a little lower, but not by a clear margin. In lab tests, the drug raised the testosterone made by pieces of human testis tissue and interfered with a vitamin A signal that sperm production relies on. It is a preprint, not yet checked by other scientists.
What it doesn’t show: It can't show that isotretinoin caused these differences or that they affect fertility or sex life: each man was measured only once, the differences were small and stayed within normal ranges, and sexual function was not studied.
Summary (paraphrased)
A multi-model study combined the Danish Young Men Study cohort (semen parameters and reproductive hormones compared between 322 isotretinoin users and 4,065 non-users) with ex vivo human testis tissue exposed to isotretinoin and retinoic acid, plus an in vitro retinoic-acid-receptor-alpha reporter assay. Cohort analysis linked isotretinoin use to lower FSH, a higher inhibin B/FSH ratio, and higher estradiol, with a non-significant trend toward lower sperm counts; associations persisted after stopping use and were absent in men who started isotretinoin a year after examination. Ex vivo tissue showed altered hormone secretion (including higher testosterone) without changes in germ-cell dynamics, and isotretinoin competitively inhibited retinoic-acid-induced RAR-alpha signaling in vitro. The authors suggest isotretinoin may durably influence male reproductive function and call for randomized trials.
Evidence
Extracted from the full text; page numbers refer to the paper. The library’s own notes are labelled as such.
- Design
- Cross-sectional studyPreprint (not peer reviewed). Multi-model study: cross-sectional analysis of a population cohort with registry-linked prescription histories, plus ex vivo culture of human testis tissue and an in vitro RARα reporter-cell assay
- Setting
- Danish Young Men Study, Copenhagen University Hospital - Rigshospitalet; men examined 2002–2022 and linked to Danish national patient and prescription registries; testis tissue collected September 2023 to December 2024
- Population
- Young Danish men (median age 19) recruited at their compulsory military fitness examination, who attended one examination day with semen and blood sampling; men with no exact examination date or with anabolic steroid use were excluded. Ex vivo: non-tumour testis tissue from 8 men having a testis removed for testicular cancer.
- Size
- 5,217 · 5,217 men (from 5,307 examined), of whom 322 isotretinoin users, 4,065 non-users and 830 users of other acne medication; overlapping user subgroups of 45 active, 124 prior and 164 later users; ex vivo hormones from 8 donors, germ-cell staining from 4; reporter assay 3 independent experiments
- Exposure
- Systemic isotretinoin (ATC D10BA01), at least one redeemed prescription before the examination; dose unknown for 164 of 322 (most common 40 mg/day, n = 112). Ex vivo 10, 15 or 25 µM isotretinoin (and 1 or 10 µM retinoic acid) for 96 hours; in vitro 12, 195 or 1,440 nM isotretinoin.
- Compared with
- Men with no isotretinoin before their examination (n = 4,065); users of other acne medication (n = 830) as a check on acne itself; men who first used isotretinoin at least one year after their examination (sensitivity analysis); vehicle (DMSO) controls ex vivo
- Outcome
- Semen volume, sperm concentration, total count, motility and morphology; serum LH, FSH, testosterone, free testosterone, SHBG, estradiol and inhibin B, with testosterone/LH and inhibin B/FSH ratios. Ex vivo hormone secretion, germ-cell proliferation and apoptosis; in vitro RARα activation. Sexual function was not assessed.
- Follow-up
- None; one examination day per man (cross-sectional), with "prior" users defined by a last prescription at least 2 years before that day. Ex vivo culture 96 hours.
Key results
- All isotretinoin users (322) vs non-users (4,065), medians: FSH 2.4 vs 2.6 IU/L (p = 0.008), inhibin B/FSH ratio 81 vs 69 (p = 0.027), estradiol 85 vs 81 pmol/L (p = 0.044); LH, testosterone, free testosterone and SHBG did not differ · p. 12, Table 3
- Semen in users vs non-users: sperm concentration 41 vs 45 million/ml, total sperm count 125 vs 145 million, progressively motile count 74 vs 88 million; none statistically significant (p 0.358, 0.465, 0.521) · p. 11, Table 3 (p. 12)
- Prior users (last prescription at least 2 years before, n = 124): FSH 2.2 vs 2.6 IU/L (p = 0.014), inhibin B/FSH 83 vs 69 (p = 0.035), estradiol 85 vs 81 pmol/L (p = 0.024). Active users (n = 45): no significant differences (estradiol 88 vs 81 pmol/L, p = 0.068 in Table 4; the text gives P = 0.06) · pp. 12–13, Table 4
- Later users (first prescription at least 1 year after the examination, n = 164): no significant differences, with FSH 2.8 IU/L, inhibin B/FSH 58 and total sperm count 170 million (p 0.077, 0.061, 0.079), i.e. in the opposite direction to past users · p. 13, Table 4
- Ex vivo testis tissue (8 donors): testosterone higher than vehicle at every isotretinoin concentration (p < 0.05) and DHT lower (p < 0.001); estradiol unchanged; inhibin B lower only at 25 µM isotretinoin and with 10 µM retinoic acid plus 10 µM isotretinoin; no effect on germ-cell proliferation (4 donors), and apoptosis rose only with 10 µM retinoic acid · pp. 13–15, Figs 2–3
- Reporter cells: isotretinoin alone activated RARα; with 195 or 1,440 nM isotretinoin, added retinoic acid gave almost no further activation; with 12 nM, retinoic acid's EC50 rose from 2.92 to 7.65 (about 2.6-fold). The text gives these EC50s in µM, but retinoic acid was tested only up to 100 nM and Fig. 4's axis is in nM · p. 16, Fig. 4; p. 10
Limitations the authors note
- Dose could not be determined for most users (164 of 322), so all users were analysed as one group
- The ex vivo model does not fully reproduce sperm production, meiosis was not assessed, and results may not reflect the testis in the body
- Germ cells were identified by position and shape, without a germ-cell-specific marker
- All cultured tissue came from men with testicular cancer; 35 of 88 fragments (39.8%) contained GCNIS-positive tubules
- Hormone concentrations and semen parameters stayed within clinical reference ranges
- The proposed mechanism is hypothetical, the ex vivo testosterone rise may involve targets other than RAR, and randomised trials are needed
Also worth weighing (library’s note)
- Preprint, not peer reviewed
- Each man was measured once; "persistence" is inferred from men whose last prescription was at least 2 years earlier, not from repeat measurements of the same men
- Fifteen semen and hormone measures were compared across four exposure groups with no stated correction for multiple testing; the significant p-values range from 0.008 to 0.044
- Users of other acne medication had the same estradiol median and IQR as isotretinoin users (85 pmol/L, 66–104), which was not significant (p = 0.103, Table 3)
- Hormone assays changed during 2002–2022 (testosterone and SHBG from 2014, LH and FSH from 2021); the stated adjustments do not include assay method or examination year (p. 6, Tables 3–4)
- The 322 + 4,065 + 830 groups sum to the 5,217 included, so the 164 later users must sit within the non-user or other-acne groups; the paper does not say which
- Ex vivo isotretinoin concentrations (10–25 µM) are about 7 to 19 times the plasma Cmax the authors cite (1.34–1.44 µM, p. 18)
- Retinoic acid concentration range is given as 0.015–100 nM in Methods (p. 10) but 0.15 pM–100 nM in the Fig. 4 legend (p. 16)
Funding: Grosses LF Foghts Fund (grant 2026-0047) and King Christian IX and Queen Louise's Jubilee Grant (grant 2025), awarded to F.B. and A.R. Interests: Not stated
What it can support (library’s note): In a preprint, a small cross-sectional association between past isotretinoin use and lower FSH and higher estradiol in young men, including men who had stopped at least two years earlier, plus lab evidence that isotretinoin interferes with retinoic-acid signalling. It cannot establish causation or clinical importance, and says nothing about sexual function.
Why it’s in the corpus
A mechanistic-plus-epidemiology signal on whether post-isotretinoin reproductive changes persist: at a single examination, 124 men whose last prescription was at least 2 years earlier had slightly lower FSH (2.2 vs 2.6 IU/L) and higher estradiol than non-users, within clinical reference ranges, with a plausible RAR-alpha pathway. It measured semen and reproductive hormones, not sexual symptoms, so it bears on PRSD only indirectly.