Prostate cancer (PCa) is among the most commonly diagnosed malignancies worldwide and is a significant contributor to premature male mortality. This retrospective cohort study aimed to estimate survival and identify factors associated with PCa-specific mortality among men treated within the oncology care network of a single Brazilian state.
Investigators used hospital-based cancer registries linked to the state Mortality Information System to assemble a retrospective cohort. The study included 10,556 men diagnosed with PCa between 2000 and 2016, with follow-up through 2021. Outcomes were classified into three mutually exclusive categories: death from prostate cancer, death from other causes, or alive at the end of follow-up. Associations with PCa-specific mortality were evaluated using cause-specific Cox proportional hazards regression models.
At the end of follow-up in 2021, the vital-status distribution was: 6,388 patients alive, 1,936 deaths attributed to PCa, and 2,232 deaths from other causes. The estimated 5-year PCa-specific survival for the cohort was 87.7%. These descriptive results provide the baseline survival experience for patients treated within the state's oncology care network during the study period.
The cause-specific Cox regression identified several factors independently associated with PCa-specific mortality:
Age at diagnosis: Increasing age was associated with a higher cause-specific hazard of PCa death. Each 10-year increment in age raised the hazard (HR = 1.14; 95% CI: 1.06–1.22).
Educational level: Compared with patients with no formal education, those with basic schooling had a lower hazard (HR = 0.82; 95% CI: 0.71–0.95), and those with middle or high education had an even lower hazard (HR = 0.64; 95% CI: 0.52–0.80). This finding links educational level to differences in PCa-specific mortality.
Referral source: Patients referred from non-SUS (non-public health system) services experienced a slightly higher risk of PCa mortality (HR = 1.14; 95% CI: 1.00–1.30) relative to those referred from SUS services.
Distant metastasis: The abstract reports presence of distant metastasis as associated with PCa-specific mortality; numerical effect estimates for metastasis were provided in the full text but are not reported in the abstract.
Treatment variables showed strong associations with PCa-specific mortality in the multivariable model:
Surgery: Receipt of surgery was associated with a substantially lower cause-specific hazard of death (HR = 0.41; 95% CI: 0.33–0.52).
Hormone therapy: Use of hormone therapy was associated with a higher cause-specific hazard (HR = 1.98; 95% CI: 1.75–2.24).
The authors explicitly caution against interpreting these treatment associations as causal. The observed associations likely reflect underlying differences in disease severity and treatment selection: patients with earlier-stage disease and better overall prognosis are more likely to receive curative-intent surgery, whereas patients with advanced disease or higher tumor burden are more likely to receive systemic hormone therapy.
Study design: As a retrospective cohort using registry linkage, the analysis is subject to residual confounding and potential misclassification in registry variables. Treatment–outcome associations can be biased by indication and by unmeasured clinical details (for example, tumor grade, performance status, comorbidities, or timing of therapies) that are not fully described in the abstract.
Missing details: The abstract does not report detailed effect sizes for all clinical covariates (for example, the exact hazard ratio for distant metastasis), nor does it describe comorbidity adjustment, staging subgroups, or specific treatment sequencing. Those details would be in the full article.
Generalizability: Findings reflect patients treated within the oncology network of a single Brazilian state and may not be directly generalizable to other regions or health systems with different case mix, access patterns, or treatment availability.
In this large series of 10,556 Brazilian men with prostate cancer, 5-year PCa-specific survival was 87.7%. Independent predictors of higher PCa-specific mortality included older age at diagnosis, lower educational level, presence of distant metastasis, and referral source. Treatment associations were pronounced—surgery linked to lower hazard and hormone therapy to higher hazard—but likely reflect selection by disease severity rather than causal effects of the treatments themselves.
These results underline the importance of earlier diagnosis, interventions to reduce social inequalities in access to cancer care, and strengthening oncology care networks to improve outcomes. Clinicians and health systems should interpret treatment–outcome associations in registry analyses with caution and consider complementary study designs or detailed clinical data to better distinguish effects of treatment from confounding by indication.