Antiretroviral therapy in HIV-infected individuals in clinical practice: are the criteria for initiating and choosing the type of drug regimen based only on immunologic and virologic values?

OBJECTIVES: To determine factors associated with beginning antiretroviral therapy and with the number of drugs used.
METHODS: Longitudinal study of 3169 HIV-infected individuals naïve from antiretroviral drugs at enrollment in 65 infectious disease clinics in Italy. Initiation of antiretroviral therapy and number of drugs used (i.e., < 3 vs. > or = 3 drugs) were the main outcome measures. Adjusted odds ratios were calculated by logistic models to establish cofactors of these two measures.
RESULTS: From January 1997 to December 1998, 1288 (40.6%) individuals started therapy, 58.0% of whom were given a triple combination regimen. This regimen became more frequent over time. By multivariate analysis, high levels of HIV-RNA and low CD4 counts were the most important independent predictors of starting any type of therapy. A significant association was also found with HIV exposure category, reason for being antiretroviral-naïve, presence/absence of liver disease, presence/absence of a new AIDS-defining disease, and clinical centre. High levels of HIV-RNA and low CD4 counts were also the most important predictors of starting with > or = 3 drugs, compared to < 3 drugs, and men had an independent higher probability of starting with > or = 3 drugs, compared to women. The probability of starting with > or = 3 drugs significantly increased with calendar time.
CONCLUSIONS: CD4 and HIV-RNA were the main cofactors of initiating both any type of therapy and therapy with > or = 3 drugs. The large variability among clinical centres suggests that clinicians are uncertain as to the exact timing of beginning therapy and the specific regimen, especially among women.