Association of delayed ART initiation with mortality according to documented concurrent clinical conditions in adults...

To assess whether the association between delayed antiretroviral therapy (ART) initiation and all-cause mortality differs according to documented concurrent clinical conditions at presentation in adults with HIV.

We conducted a multicentre retrospective cohort study of 1826 treatment-naive adults with HIV in Jiangxi Province, China (2004-2024), linking clinical and surveillance records. Time zero was defined as ART initiation. The exposure was a delay of >30 versus ≤30 days from diagnosis; the outcome was all-cause mortality. Presenting risk was defined a priori from conditions documented by ART initiation (documented condition, none documented or untested/unknown). We fitted subgroup-specific Cox models and a full-sample interaction model. Multiple imputation was a co-primary approach for missing baseline HIV-RNA.

Overall, 142 participants (7.8%) died. A delay of >30 days was not associated with mortality overall (adjusted hazard ratio [aHR] 1.05, 95% confidence interval [CI] 0.83-1.33; p = 0.70). Among participants with documented concurrent conditions (n = 450), a delay of >30 days was associated with approximately two-fold higher mortality under multiple imputation (aHR 2.07, 95% CI 1.06-4.02; p = 0.033; 48-week cumulative mortality 12.1% vs. 6.2%); no association was seen without documented conditions (aHR 0.87, 95% CI 0.55-1.39). The interaction was borderline under multiple imputation (p = 0.064; fraction of missing information 0.04).

Documented clinical complexity at presentation may identify a subgroup in whom prolonged ART delay is less well tolerated. These hypothesis-generating findings warrant external validation.