Abstract Body

Background: Stockouts are relatively common in large ART programmes and cause treatment interruptions, but their long-term impact is rarely assessed. We investigated this using long-term follow-up of a structured treatment interruption (STI) randomization 12 weeks on/off in a large adult trial (DART), and data on unplanned treatment interruptions (UPTIs) in this and a parallel paediatric trial (ARROW). Methodology: The DART STI randomisation stopped early; patients returned to continuous therapy (CT) and were followed for a further 3 years. Random effects models were used to assess differences in CD4s between those randomized to STI vs CT over the long-term back on ART. Linear regression was used to assess the impact of total time off ART on CD4 4 years (adults) or 3 years (children) after ART initiation. The impact of time-updated cumulative time off ART on mortality was assessed using Cox models. All models adjusted for sex, centre and pre-ART CD4/CD4% and age. Results: DART STI patients had >100 cells/mm3 lower CD4 when restarting CT (Figure). Subsequently STI patients gained 48 (95% CI 43-54) cells/year, significantly faster than the 31 (25-36) in CT patients (p<0.001). Extrapolating forward, those randomized to STI would have similar CD4 to those on CT 4-5 years after restarting ART. 32% (1055/3316) adults had total 1955 UPTIs lasting ≥4 days, vs 23% (281/1206) children (p<0.001) with 536 UPTIs, with median total 17 days (7-49) off treatment in adults vs 13 (7-34) in children. In both adults and children most UPTIs occurred in the first year on ART (p<0.001). The most common reason for UPTIs was inability to attend clinic (61% adults, 50% children). Adverse events were a more common reason in adults (22%, 1% children) and patient/carer decision in children (17%, 5% adults). For every extra cumulative month off ART (STIs+UPTIs) in adults CD4 4 years after ART initiation was 6 cells/mm3 lower (95%CI 2-8, p<0.001); in children CD4% 3 years after ART initiation was 2% lower (1-3%, p=0.001). Having ever interrupted ART increased mortality risk in adults (HR=3.0 [95% CI 2.1-4.2] p<0.001) with similar non-significant trend in children (HR=2.6 [0.7-10.4] p=0.18). There was no evidence of independent effects of cumulative time off ART in adults (HR=1.1 per doubling [1.0-1.2] p=0.26) or children (1.6 [0.9-2.8] p=0.11). Conclusions: TIs lead to lower CD4 counts over the long-term even after returning to ART, highlighting the importance of avoiding ART interruption in programmes in resource limited settings.