Long-Term Follow-Up of Transsexual Persons Undergoing Sex Reassignment SurgeryCohort Study in Sweden

For most of the twentieth century, sex reassignment surgery was treated as the endpoint, the destination, the resolution. A person spent years navigating diagnosis, hormonal treatment, and surgery. Once the body matched the identity, the medical system largely considered its job done. Then, a Swedish research team pulled three decades of national registry data, and that assumption cracked. Not because the surgery failed to do what it is designed to do, but because the story turned out to be longer than anyone was measuring. Dhejne and colleagues published their findings in 2011, presenting the first nationwide, long-term, population-based follow-up of sex-reassigned individuals ever conducted. This was not a clinic sample or a group of volunteers. It included every single person in Sweden who underwent legal sex reassignment between 1973 and 2003, totaling 324 individuals — 191 male-to-female and 133 female-to-male. They were identified through linkages across national registers covering 13.8 million unique individuals. The Hospital Discharge Register, the Total Population Register, the Medical Birth Register, and census data were all cross-referenced. To qualify as exposed, a person needed at least one inpatient diagnosis of gender identity disorder and at least one documented shift in the sex variable across registers, indicating a legal change of sex. For each of those 324 people, the researchers randomly selected ten population controls, resulting in three thousand two hundred forty total controls, matched by birth year and birth sex. These controls had no gender identity disorder diagnosis and no discrepancy in their sex registration. A brief note on what the study was measuring: outcomes were estimated using Cox proportional-hazard models, which produce hazard ratios. A hazard ratio of one means the two groups have the same risk over time. A ratio above one indicates that the exposed group experiences events at a higher rate. The average follow-up for all-cause mortality was 11.4 years, which is a long enough window for patterns to emerge that shorter studies would miss. Now, for what they found. The all-cause mortality hazard ratio for sex-reassigned persons compared to birth-sex matched controls was 2.8, with a confidence interval of 1.8 to 4.3. In absolute terms, the incidence rate of any death in the sex-reassigned cohort was 7.3 per one thousand person-years, compared to 2.5 per one thousand in controls. This represents nearly three times the death rate among people of the same age who hadn't undergone reassignment. However, the cause-specific finding is the one that demands attention. Suicide mortality in the sex-reassigned group occurred at a rate of 2.7 per one thousand person-years. In the matched controls, that rate was 0.1. The adjusted hazard ratio for death by suicide was 19.1, with a confidence interval spanning 5.8 to 62.9. That interval is wide — the statistical uncertainty is real. But even at the low end, you're looking at nearly six times the suicide mortality of matched peers. The survival curves for overall mortality, as the paper notes, didn't diverge sharply until roughly ten years of follow-up. This is why the thirty-year window mattered. A five-year study would have missed it. The pattern extends beyond mortality. For inpatient psychiatric care, with admissions for gender identity disorder itself excluded, the adjusted hazard ratio was 2.8, with a confidence interval of 2.0 to 3.9. For suicide attempts, it was 4.9, with a confidence interval of 2.9 to 8.5. Both comparisons are adjusted for psychiatric morbidity before surgery and for immigrant status. The researchers weren't comparing people who entered the system healthy. They accounted for prior illness. Dhejne and colleagues also split the cohort by when surgery occurred — from 1973 to 1988 versus 1989 to 2003. This is where the data gets more textured. For psychiatric hospitalization, elevated risk persisted across both periods, with adjusted hazard ratios of 3.0 for the earlier group and 2.5 for the later one, both statistically significant. For suicide attempts, though, the excess was concentrated in the earlier cohort, with an adjusted hazard ratio of 7.9 compared to 2.0 in the later one. The later estimate did not reach statistical significance. Criminal convictions showed a similar pattern: elevated in the earlier period and not significantly elevated in the later one. The authors are careful about what this means. It could reflect improving care over time, changing social attitudes, or both. They do not claim to know which. What they do say is that the temporal patterning exists. Regarding criminal convictions, the paper includes a finding that has generated some confusion and deserves careful examination. Female-to-male individuals had higher conviction rates than birth-sex female controls, with an adjusted hazard ratio of 4.1. However, they did not differ significantly from male controls. Dhejne and colleagues interpret this as alignment with the male population norm after reassignment, not as a signal of unique pathology. The parallel analysis for male-to-female individuals showed a similar logic in reverse: elevated rates compared to female controls, but no significant difference compared to male controls. Here is the interpretive limit that matters most, and the paper states it directly. Dhejne and colleagues compared sex-reassigned persons to the general population. They did not compare them to untreated transgender individuals. That comparison was not part of this design. The authors are explicit: the study is, in their words, "only informative with respect to transsexual persons' health after sex reassignment; no inferences can be drawn as to the effectiveness of sex reassignment as a treatment for transsexualism." In plain terms, this study cannot tell you whether surgery helped or harmed relative to no surgery. It tells you that elevated risks persisted after surgery compared to the general population. Those are different claims, and collapsing them is a misreading of what the data shows. The paper also flags the historical context of the cohort. A significant portion of these individuals were operated on in the 1970s and 1980s, during an era with less social acceptance, less post-operative psychological support, and medical practices that have since evolved considerably. The authors raise the possibility that some of the elevated risks reflect factors entirely external to the procedure, such as fear of discrimination leading to avoidance of healthcare, stigma, or minority stress. Surgery cannot resolve those issues. The period-specific findings, where risks were substantially higher in the earlier cohort, are at least consistent with that interpretation, though the study cannot prove it. What Dhejne and colleagues conclude is this: sex reassignment procedures and hormonal therapy alleviate gender dysphoria. The paper states that directly. However, alleviating gender dysphoria does not appear sufficient to eliminate the substantially elevated risks that persist in the years and decades afterward. Overall mortality was nearly three times higher. Suicide mortality was roughly nineteen times higher. Psychiatric hospitalization and suicide attempt rates were both elevated even after controlling for prior illness. On that basis, they argue that sex-reassigned persons constitute a risk group requiring long-term psychiatric and physical health follow-up, and that improved care after sex reassignment should be a clinical priority. The study is a measurement of what the medical system produced over thirty years of care in Sweden. A system that defined treatment as surgery and then, largely, stepped back. What Dhejne and colleagues found in that gap — between the operating room and the decades that followed — is an argument for a different model. Not a reason to withhold the treatment that addresses gender dysphoria, but a reason to extend the care that follows it. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.