The effect of care interruptions on mortality in adults resuming antiretroviral therapy.

Objective: To estimate the relative rate of all-cause mortality amongst those on antiretroviral treatment (ART) with a history of interruptions compared with those with no previous interruptions in care.

Design: Retrospective cohort study.

Methods: We used data from four South African cohorts participating in the International epidemiology Databases to Evaluate AIDS Southern Africa collaboration. We included adults who started ART between 2004 and 2019. We defined a care interruption as a gap in contact longer than 180 days. Observation time prior to interruption was allocated to a 'no interruption' group. Observation time after interruption was allocated to one of two groups based on whether the first interruption started before 6 months of ART ('early interruption') or later ('late interruption'). We used Cox regression to estimate hazard ratios.

Results: Sixty-three thousand six hundred and ninety-two participants contributed 162 916 person-years of observation. There were 3469 deaths. Most participants were female individuals (67.4%) and the median age at ART initiation was 33.3 years (interquartile range: 27.5-40.7). Seventeen thousand and eleven (26.7%) participants experienced care interruptions. Those resuming ART experienced increased mortality compared with those with no interruptions: early interrupters had a hazard ratio of 4.37 (95% confidence interval (CI) 3.87-4.95) and late interrupters had a hazard ratio of 2.74 (95% CI 2.39-3.15). In sensitivity analyses, effect sizes were found to be proportional to the length of time used to define interruptions.

Conclusion: Our findings highlight the need to improve retention in care, regardless of treatment duration. Programmes to encourage return to care also need to be strengthened.