Australie : mortalité 4 à 9 fois supérieure chez les personnes sous hormones croisées, suicide compris

Mortalité toutes causes et par cause chez les personnes transgenres et de genre divers : une étude de cohorte nationale en Australie

Études de cohorte & registres Hormones croisées Santé mentale & suicide

Authors: Karinna Saxby, Sasha Bailey, Dennis Petrie, Brendan J. Nolan
Journal: The Lancet Regional Health – Western Pacific
Year: 2026
Source: https://doi.org/10.1016/j.lanwpc.2026.101813

Summary

Australian researchers used national prescription, healthcare and death records to find about 29,000 people they classed as trans because they had filled a cross-sex hormone prescription: testosterone for people recorded as female, oestrogen for people recorded as male. They compared deaths with about 20 million people who saw a family doctor, adjusting only for age and start date. Over roughly three to four years, people on cross-sex hormones died at four to nine times the comparison rate. They had much higher rates of suicide, other violent or accidental deaths, cancer and heart disease. The authors blame stigma and gaps in care and say hormones are not the cause. But the study measured neither. A single prescription is a loose way to decide who is trans, and may include people treated for other illnesses. The results show more deaths in this group but cannot explain why.

Key findings

L'étude incluait 9 713 personnes enregistrées comme femmes (âge moyen 25 ans, suivies environ 3 ans) et 19 347 enregistrées comme hommes (âge moyen 36 ans, suivies près de 4 ans), repérées par des ordonnances d'hormones du sexe opposé de 2012 à 2023.
Elles ont été comparées à environ 20 millions de personnes ayant consulté un médecin de famille, appariées seulement sur l'âge et le mois et l'année d'entrée dans l'étude.
Les décès toutes causes confondues étaient environ 9 fois plus élevés dans le groupe enregistré comme femmes et environ 4 fois plus dans le groupe enregistré comme hommes ; 0,6 % et 5,5 % sont décédés pendant l'étude.
Le suicide était environ 12 fois plus élevé dans le groupe enregistré comme femmes et 4 fois plus dans le groupe enregistré comme hommes ; les autres décès violents ou accidentels étaient environ 9 et 3 fois plus élevés.
Les décès par cancer étaient environ 8 et 5 fois plus élevés, et les décès cardiaques et vasculaires environ 16 et 3 fois plus élevés, avec les plus grands écarts aux âges plus avancés.
Dans le groupe enregistré comme hommes, le risque supplémentaire de suicide diminuait avec la durée sous hormones ; dans le groupe enregistré comme femmes, le risque global supplémentaire de décès augmentait avec le temps.
Les auteurs attribuent cela au stress de la condition de minorité et aux obstacles aux soins, et disent que les décès ne doivent pas être imputés aux hormones, mais l'étude n'a mesuré ni l'un ni l'autre.

Abstract

BACKGROUND: Transgender and gender diverse ('trans') people may be at increased risk of mortality, particularly from external causes, however, large-scale, population-based evidence remains scarce. This study aims to document all-cause and cause-specific mortality among trans people. METHODS: We source administrative data on healthcare and death records (2012-2023) from all Australians aged 15 years and above. Individuals identified as trans (initiated gender-affirming hormone therapy) were matched to the general population who visited a General Practitioner during the same period. Cox proportional hazard models were used to estimate all-cause and cause-specific mortality risk for trans people, with models estimated separately for people who were assigned female sex at birth (AFAB) and people who were assigned male sex at birth (AMAB). Inverse probability weights were applied to balance comparators on month-year and age at entry. Results were additionally stratified by age (15-24, 25-39, 40-59, and ≥60 years). FINDINGS: A total of 19,347 trans people AMAB (mean age 36.2 years; median follow-up 3.7 years) and 9713 trans people AFAB (mean age 25.4 years; median follow-up 2.8 years) were matched with 9,879,037 general population males (mean age 44.4; median follow-up 11.3 years) and 10,282,651 general population females (mean age 44.9; median follow-up 11.4 years), respectively. All-cause mortality was significantly higher for trans people AMAB [HR = 3.89 (95% CI 33.65; 4.14)] and trans people AFAB [HR = 9.03 (95% CI 6.90; 11.83)]. For trans people AFAB, cause-specific mortality was elevated for cardiovascular disease [HR = 16.39 (95% CI 8.56; 31.37)], suicide [HR = 11.73 (95% CI 6.94; 19.80)], external causes [HR = 8.95 (95% CI 5.69; 14.09)], and cancer [HR = 7.62 (95% CI 4.61; 12.61)]. For trans people AMAB, cause-specific mortality was elevated for cancer [HR = 5.12 (95% CI 4.67; 5.61)], suicide [HR = 4.02 (95% CI 3.12; 5.18)], external causes [HR = 2.78 (95% CI 2.31; 3.35)], and cardiovascular disease [HR = 2.60 (95% CI 2.22; 3.05)]. Older trans people had more pronounced excess risk from cancer and cardiovascular disease. For trans people AFAB, excess mortality from suicide and external causes increased with age, whereas for trans people AMAB relative risks were higher in young and middle adulthood. INTERPRETATION: In this nationwide cohort study, trans Australians experienced substantially elevated mortality risk. Tailored policy responses are needed to address premature mortality in trans populations. FUNDING: University of Melbourne McKenzie Fellowship (2025MCK182); the National Health and Medical Research Council (2008956); Viertel Charitable Foundation; University of Melbourne Faculty Research Grant (2025FRG19).

Conclusion

In this nationwide cohort study, trans Australians experienced substantially elevated mortality risk. Tailored policy responses are needed to address premature mortality in trans populations, including cardiovascular disease and cancer prevention (risk-factor management, screening uptake and continuity of care), alongside strategies to prevent deaths from external causes, including suicide. The authors state that the elevated mortality should not be interpreted as attributable to gender-affirming hormone therapy itself, and that policy responses should address the social factors, such as stigma and discrimination, that they argue perpetuate these inequities.

Methodological limitations

People may be wrongly counted as trans: Anyone who ever filled one testosterone prescription while recorded as female, or one oestrogen prescription while recorded as male, was counted as trans. These hormones are also given for other conditions, such as oestrogen for some men with prostate cancer or testosterone for some women. The researchers did not report checking whether this definition was accurate.
Signs that sick, non-trans patients were included: 15% of the male-recorded group were 60 or older when they entered, 5.5% died within about 4 years, and cancer was the biggest source of extra deaths. This looks like an older, medically ill group, so some of the extra deaths may come from the illnesses behind the hormone use.
Very little was accounted for: The comparison was matched only on age and start date. The authors chose not to account for income, rural living or migration, and did not account for mental illness, past self-harm, drug or alcohol use or other illnesses.
No mental-health or treatment details: The study cannot tell apart the effects of gender distress, other mental-health conditions and the hormones themselves on suicide or other deaths.
Claims about causes go beyond the data: The authors argue the deaths 'should not be interpreted as being attributable' to hormones and blame minority stress and barriers to care. Neither idea was tested, and there was no group of trans people who did not take hormones to compare with.
Short follow-up: Trans people were followed for about 3 to 4 years on average, compared with about 11 years for the comparison group. Only 59 people in the female-recorded group died, so figures for specific causes of death are very uncertain.
Shaky statistics: The statistical method assumes the gap in risk stays the same over time, which was not true here. It also did not allow for people dying of one cause before another, and some age-group results are extremely uncertain (one suicide estimate ranged from about 5 to 83 times higher).
Error in the paper: The summary of the paper misprints one key number (a range given as '33.65; 4.14' that should start at 3.65), a sign the paper was not carefully checked.
Limited to one group: The results only apply to people in Australia who filled government-subsidised hormone prescriptions, not to trans-identifying people who do not take hormones.
Authors' interests: The lead author co-founded and sits on the committee of LGBTQ+ Economists and Allies in the Asia–Pacific. Other authors belong to a trans health research group, and the paper argues for wider access to gender-affirming care.

detrans.ai

Summary

Key findings

Abstract

Conclusion