Predictive value of HIV-1 viral load on risk for opportunistic infection

Objective: The relationship between HIV-1 viral load and the risk for opportunistic infection (OI) was examined in Adult AIDS Clinical Trial Group (AACTG) 722, a virology substudy of AACTG 323: a phase 4 randomized study designed to examine the use of chronic suppressive versus episodic fluconazole therapy. Methods: The primary analysis used a case-control sampling scheme with two controls per "case" (subjects that developed an OI) matched by gender, age, and time on study. Forty-five cases and matched controls were identified and used in the analysis. Results: Study 722 accrued 518 subjects between 5/97 and 11/99. Forty-five subjects developed serious OIs or refractory candidiasis. Median baseline CD4 count was 24 cell/mm³ for cases and 46 for controls (p = .003). Median viral load (VL) was 5.02 log₁₀ copies/mL for cases and 4.08 for controls (p = .002). Multivariate analysis found four independent variables associated with time to OI: baseline VL and CD4 (RR = 2.2 per log increment and 6.0 per 50-cell increment, respectively), a one log increase in VL at any time (RR = 15), and history of an OI (RR = 5.2). Conclusions: VL and changes in VL were independently associated with risk of development of OIs in a prospective study and should be considered by clinicians when assessing patients for risk of OI.