Predictors of Time to Switch to Second Line ART after First Line Failure in Johannesburg, South Africa

By Julia Rohr  Prudence Ive  Dr. Ribka Berhanu  Dr. Kate Shearer  Dr Mhairi Maskew  Dr. Lawrence Long  Professor Ian Sanne  Dr. Matthew Fox  |  | 


Background: Many patients with documented virologic treatment failure on first line antiretroviral therapy (ART) need to be switched to a second line regimen. In South Africa, national treatment guidelines state that patients with two consecutive failing viral load measurements on non-nucleoside reverse transcriptase inhibitor first line treatment should be switched to a protease inhibitor based regimen. Yet, there are often long delays in switching if it occurs at all. To better understand why delays occur, we explored whether clinic attendance or disease progression was more likely to explain delays in second line initiation. Methods: We conducted an observational cohort study using data from Themba Lethu Clinic in Johannesburg, South Africa, a large, public sector HIV clinic. We included all adult patients initiating first line ART since April 2004 with virologic failure (date of 2nd consecutive viral load ≥ 1,000 copies/mL) at least 1 year before the end of the study period in 2013. We looked for demographic, clinical, and visit adherence predictors of time to switch to second line ART after first line failure using Cox proportional hazards regression to adjust for confounders. Results: 16825 patients initiated first line ART, of which 1578 (9.4%) failed after a median (IQR) of 22.8 (13.7-38.6) months. Of the 1578 who failed 60% were female. At time of failure, mean (SD) age was 39.2 (9.1) years, mean (SD) CD4 count was 244 (183) cells/mm3, and median (IQR) viral load was 8700 (2728-41000) copies/mL. 69% (1086) of failing patients switched to second line ART in a median (IQR) of 2.8 (1.0-6.6) months. A faster rate of switching was seen in patients with higher viral load, lower CD4 count and shorter duration on first line ART prior to failure (Table). We found some evidence that differences in clinic attendance after failure explained how quickly switches occurred; a smaller proportion of patients who switched by 6 months had missed visits in the first 6 months after failure (31%) than those who did not switch by 6 months (45%). Conclusion: We found that disease severity at first line failure predicted faster switching, with patients with higher viral load levels and lower CD4 counts being switched to second line more quickly. Lack of visit adherence before failure had a minimal effect on delaying second line ART, and the majority of patients did not miss visits within the 6 months following failure, indicating delay in second line treatment initiation is not solely explained by patients missing visits.

Conference: CROI conference 2014, Boston, USA

Publication details