--- title: Korean men under prostate cancer surveillance started antidepressants more often than men who were treated. On survival, the study could not tell the difference publication: Medical Insights Korea section: news category: Prostate canonical_url: https://medicalinsightskorea.com/news/prostate-surveillance-korea-antidepressants-target-trial author: Andrew Cho published: 2026-09-02T11:15:00+09:00 modified: 2026-09-02T11:15:00+09:00 content_type: reporting license: Quotation permitted with attribution to Medical Insights Korea and a link to the canonical URL. --- # Korean men under prostate cancer surveillance started antidepressants more often than men who were treated. On survival, the study could not tell the difference A target trial emulation using national linkage data found four prostate cancer deaths in total between surgery and surveillance. The finding it could measure was psychological. ## Answer In a Korean target trial emulation of 8,036 men with localised prostate cancer, radical prostatectomy and radiotherapy were each associated with a lower rate of starting antidepressants than active surveillance or watchful waiting — hazard ratios of 0.73 and 0.38 — while prostate cancer-specific deaths numbered two in the surgery group and two in the surveillance group over a mean 4.4 years of follow-up. ## Key takeaways - A Korean target trial emulation linking national insurance, cancer registry and death registry data matched 1,191 men having radical prostatectomy against 1,191 under surveillance, and 428 having radiotherapy against 428 under surveillance. - Over a mean follow-up of 4.4 years there were 2 prostate cancer deaths in the surgery group and 2 in its matched surveillance group, and 3 in the radiotherapy group against 1 in its matched surveillance group. - Men who had surgery or radiotherapy were less likely to start antidepressants than men under surveillance: hazard ratios of 0.73 (95 percent confidence interval 0.58 to 0.93) and 0.38 (0.24 to 0.61). - With four prostate cancer deaths in total across the surgery comparison, the study cannot distinguish the treatments on mortality; its finding of comparable risk is an absence of evidence rather than evidence of equivalence. - In a separate national cohort of 166,848 Korean men, 26.7 percent had been repeatedly PSA-tested before diagnosis, and those men more often received localised treatment — surgery in 45.6 percent against 33.8 percent. ## Full article The hardest question in prostate cancer is not how to treat it. It is whether to. Surgery and radiotherapy carry incontinence and impotence; leaving a confirmed cancer in place carries the knowledge that it is there. The randomised trial that addressed this, ProtecT, found no survival difference between monitoring, surgery and radiotherapy in British men. A Korean team has now emulated that trial in national data, and found something ProtecT was not designed to look for. ### What a target trial emulation does Researchers at Ajou University linked Korea's national health insurance, cancer registry and death registry for 2012 to 2020, took men aged 50 to 69 with a first localised prostate cancer, and constructed the comparison a randomised trial would have made: 1,191 men having radical prostatectomy matched against 1,191 under active surveillance or watchful waiting, and 428 having radiotherapy matched against 428 under surveillance. The cohort drew on 8,036 patients. The design borrows a trial's structure — eligibility, assignment, follow-up defined in advance — and applies it to data from people who were never randomised. It is the strongest observational form available for a question nobody is going to randomise again, and it remains observational. ### The mortality result is four deaths **Prostate cancer-specific deaths over a mean 4.4 years of follow-up** | Comparison | Treated | Surveillance | | --- | --- | --- | | Radical prostatectomy vs surveillance | 2 (0.2%) | 2 (0.2%) | | Radiotherapy vs surveillance | 3 (0.7%) | 1 (0.2%) | The study reports that surveillance showed comparable risks of prostate cancer death, all-cause death and metastasis. On mortality that conclusion rests on four deaths in the first comparison and four in the second. > **What eight deaths can and cannot support** > A comparison with almost no events cannot detect a difference, and failing to detect one is not the same as showing there is none. This study is consistent with the randomised evidence and adds very little to it on survival. Its own follow-up — a mean of 4.4 years, for a cancer whose deaths accrue over fifteen and more — is the reason. ### The finding it could measure Antidepressant initiation was a secondary outcome, and unlike death it happened often enough to analyse. **0.73 and 0.38** — hazard ratios for starting an antidepressant after radical prostatectomy and after radiotherapy, against active surveillance or watchful waiting _Source: An MH et al., Discover Oncology, 2026_ Men who had surgery were less likely to start an antidepressant than matched men under surveillance, with a hazard ratio of 0.73 (0.58 to 0.93). For radiotherapy the figure was 0.38 (0.24 to 0.61). Both intervals exclude one, and both point the same way: the men left untreated were the ones who went on medication. This publication reported earlier today that Korean patients choosing active surveillance for a small thyroid cancer had better quality-of-life scores than those choosing surgery — a difference that was present before treatment and gone within a year. The prostate finding runs the other way and deserves the same scepticism applied in the same direction. 1. Starting an antidepressant is a record of a prescription, not a measurement of distress. It depends on who consults, who asks, and who prescribes. 2. Men were not randomised. Whatever made a man and his clinician choose surveillance may also bear on his mental health, and matching corrects only for what was recorded. 3. The direction is nonetheless the one a clinician would expect from patients living with a cancer that has been confirmed and deliberately left in place, and it is the outcome ProtecT's survival endpoints were never going to capture. ### Who is being tested, and what follows A separate national analysis of 166,848 Korean men newly registered with prostate cancer between 2010 and 2020 divided them by their testing history. Just over a quarter — 26.7 percent — had been PSA-tested at least three times across at least two years before diagnosis. For 42.2 percent, the first PSA test came within three months of registration. The repeatedly tested men more often received treatment aimed at cure rather than control: surgery in 45.6 percent against 33.8 percent. Their authors conclude that once government-set prices had stabilised, repeated testing was associated with greater use of localised therapy and lower cumulative spending on prolonged systemic treatment. Read alongside the risk distribution this publication reported earlier today — 53.9 percent of Korean prostate cancers high-risk at diagnosis, and worse in rural areas — the two describe the same divide from different sides. Men who are being tested repeatedly are found earlier and treated locally. Men who are not are found later. ### What remains unsettled 1. Neither study randomised anyone. Both compare men who were treated differently for reasons the data only partly records. 2. The target trial emulation covers 2012 to 2020 with a mean 4.4 years of follow-up, which is short for prostate cancer survival and cannot speak to outcomes at fifteen years. 3. The PSA analysis observes that repeated testing precedes localised treatment. It does not establish that testing more men would improve their outcomes, which is the question that matters and the one it was not designed to answer. 4. Cost findings are specific to Korea's administered prices and to the 2016 to 2020 window in which those prices were stable. 5. None of this addresses the harms of treatment itself — continence and sexual function — which are the reason the question is difficult and are absent from both datasets. ## Sources - An MH, Kim C, Min K, Yi KH, Park RW, Jung M. Target trial emulation of survival outcomes for clinically localized prostate cancer treatments. Discover Oncology 2026;17(1):1208. doi:10.1007/s12672-026-05392-4. PMID:42286375. https://doi.org/10.1007/s12672-026-05392-4 - Ko YH, Kim SW, Son NH. Impact of repeated prostate-specific antigen testing on management patterns and personal healthcare spending for prostate cancer: A population-based study National Health Insurance data for 166,848 patients in South Korea from 2010 to 2020. Investigative and Clinical Urology 2026;67(3):263-272. doi:10.4111/icu.20250537. PMID:42065662. https://doi.org/10.4111/icu.20250537 - Ko YH, Kim SW, Son NH. Incidence of High-Risk Prostate Cancer in Korea: Insights From Real-World Data Between 2010 and 2020. Journal of Korean Medical Science 2026;41(14):e10. doi:10.3346/jkms.2026.41.e10. PMID:41978922. https://doi.org/10.3346/jkms.2026.41.e10 ## Citation Andrew Cho (2026). "Korean men under prostate cancer surveillance started antidepressants more often than men who were treated. On survival, the study could not tell the difference". Medical Insights Korea. https://medicalinsightskorea.com/news/prostate-surveillance-korea-antidepressants-target-trial Retrieved from https://medicalinsightskorea.com/news/prostate-surveillance-korea-antidepressants-target-trial