Laboratory adverse events and discontinuation of therapy according to CD4(+) cell count at the start of antiretroviral therapy.

ABSTRACT

OBJECTIVE: Few data describe antiretroviral treatment (ART)-related adverse events when treatment is initiated at CD4(+) cell counts more than 350 cells/µl. We compared rates of laboratory-defined adverse events (LDAEs) according to CD4(+) cell count at ART initiation. DESIGN: Analysis of on-going cohort study. METHODS: ART-naive persons initiating ART from 2000 to 2010 were included. Chi-square, analysis of variance (ANOVA) and Kruskal-Wallis tests compared characteristics among those starting ART with a CD4(+) cell count of 350 or less, 351-499 and at least 500 cells/µl. Time-updated Poisson regression compared rates of LDAE in the three CD4(+) cell strata. Cox proportional hazard models compared risk of ART discontinuation. RESULTS: Nine thousand, four hundred and six individuals were included median age 37 years, 61% white, 80% men, median viral load 4.8 log copies/ml. Four hundred and forty-seven (4.9%) and 1099 (11.7%) started ART with a CD4(+) cell count at least 500 and 351-499 cells/µl, respectively. One thousand, two hundred and eighty-three (13.6%) patients experienced at least one LDAE. The rate of LDAE did not differ between those starting ART with a CD4(+) cell count 351-499 and less than 350 cells/µl [relative rate 0.90, 95% confidence interval (CI) 0.74-1.09)], but an increased risk of ART discontinuation was observed (hazard ratio 1.58, 95% CI 1.10-2.27). Those starting ART at CD4(+) cell count at least 500 cells/µl had an increased rate of LDAE (relative rate 1.44, 95% CI 1.13-1.82) but were not more likely to discontinue ART (hazard ratio 1.15, 95% CI 0.64-2.09). CONCLUSION: This study demonstrates the need to consider ART-related toxicities when initiating therapy at CD4(+) cell counts at least 500 cells/µl. Whilst evidence from randomized controlled trials is awaited, the timing of ART initiation in terms of benefits and risks of ART remains an important question.