Abstract

BACKGROUND: Patients receiving combination antiretroviral therapy (cART) might continue treatment with a virologically failing regimen. We sought to identify annual change in CD4(+) T-cell count according to levels of viraemia in patients on cART.

METHODS: A total of 111,371 CD4(+) T-cell counts and viral load measurements in 8,227 patients were analysed. Annual change in CD4(+) T-cell numbers was estimated using mixed models.

RESULTS: After adjustment, the estimated average annual change in CD4(+) T-cell count significantly increased when viral load was <500 copies/ml (30.4 cells/mm(3), 95% confidence interval [CI] 26.6-34.3), was stable when viral load was 500-9,999 copies/ml (3.1 cells/mm(3), 95% CI -5.3-11.5) and decreased when viral load was > or =10,000 copies/ml (-14.8 cells/mm(3), 95% CI -4.5--25.1). Patients taking a boosted protease inhibitor (PI) regimen had more positive annual CD4(+) T-cell count changes than patients taking other regimens for any given viral load strata: 30.9 cells/mm(3) (95% CI 27.7-34.1) when viral load was <500 copies/ml, 14.2 cells/mm(3) (95% CI -2.1-30.4) when viral load was 500-9,999 copies/ml and -19.9 cells/mm(3) (95% CI -36.6--3.3) when viral load was >or =10,000 copies/ml. By contrast, among patients taking a non-nucleoside reverse transcriptase inhibitor (NNRTI)-based regimen, the CD4(+) T-cell count significantly decreased when the viral load was 500-9,999 copies/ml (-18.6 cells/mm(3), 95% CI -33.8--3.5) and decreased at a faster rate when the viral load was > or =10,000 copies/ml (-44.4 cells/mm(3), 95% CI -62.0--26.9; P=0.0012, test for interaction).

CONCLUSIONS: On average, CD4(+) T-cell counts did not significantly decrease until the viral load exceeded 10,000 copies/ml in patients treated with a boosted PI-containing cART regimen, but decreased in patients taking an NNRTI-based cART regimen when viral load was 500-9,999 copies/ml.