Resumen:
BACKGROUND: Tuberculosis coinfection at HIV diagnosis drives early morbidity and mortality. We estimated tuberculosis prevalence at HIV diagnosis over two decades in a Spanish cohort and quantified its determinants and outcomes. METHODS: We analyzed 19,058 ART-naive adults in CoRIS (2004-2024). Prevalent tuberculosis was defined as disease diagnosed from the date of HIV diagnosis up to six months after ART initiation. Predictors were identified using multivariable logistic regression; mortality was assessed using Kaplan-Meier and competing-risk methods RESULTS: Prevalence was 2.2% (426/19,058), declining from 4.9% (2004-2008) to 0.8% (2019-2024), an 83% reduction (p<0.001). Despite this decline, affected individuals presented with increasingly advanced immunosuppression. Strong independent predictors included geographical origin (Sub-Saharan Africa OR=4.3; Latin America OR=1.6), positive TB screening (OR=7.5), viral load >1,000,000 copies/mL (OR=2.9), injection drug use (OR=2.9), and age ?50 years (OR=6.7), whereas university education (OR=0.3) and CD4 ?500 cells/µL (OR=0.5) were protective. Model discrimination was good (AUC=0.84). Mortality was higher with tuberculosis (15.0% vs 4.1%; RR=3.6; 5-year survival 87% vs 97%). CONCLUSIONS: Although prevalence has declined substantially, tuberculosis at HIV diagnosis remains concentrated in high-risk groups and confers excess mortality. Targeted screening and prompt ART remain essential, particularly for migrants from high-burden regions.