Background:  Although the of effect HAART has been well investigated on the natural history and clinical features of HIV-related central nervous system disorders (CNS-D), the effectiveness of HAART in naive patients affected by CNS-D has been poorly evaluated.

Methods:  Patients selected from the Italian Registry Investigative NeuroAIDS (IRINA) on HIV-advanced persons naïve for antiretrovirals with a CNS-D diagnosis. Main outcomes were virological response (plasma HIV RNA ≤50 copies/mL 12 months after CNS-D), treatment success (plasma HIV RNA ≤50 copies/mL 12 months after diagnosis, considering treatment discontinuation or deaths as failure), and immunological response (CD4 cell count increased ≥100 cell/μL 12 months after diagnosis). Multivariable logistic regression and Cox regression was used for statistical analysis.

Results:  We included 517 naive patients whose median CD4 cell count and plasma HIV RNA were 44 cell/μL (IQR 17 to 101) and 5.24 copies/mL (IQR 4.79 to 5.65), respectively. Virological response was observed in 72.3% and treatment success in 47.1%. At multivariate analysis, 6-month viral load was a strong predictor for both endpoints, virological response (adjusted OR: 0.42 for each log₁₀ copies/mL; 95%CI 0.28 to 0.61) and treatment success (AOR 0.41 for each log₁₀ copies/mL; 95%CI 0.29 to 0.59). Virological response was also predicted by months on HAART during first year after diagnosis (AOR 1.03 for each month; 95%CI 1.00 to 1.06). Immunological response was obtained in 49.3%; 6-month HIV RNA (AOR 0.35 for each log₁₀ copies/mL; 95%CI 0.21 to 0.60) and baseline CD4 (AOR 5.22 for each 100 cells increase; 95%CI 3.35 to 8.14) were the main predictors of immunological response. After a median follow-up of 12 months, 152 deaths and 51 new AIDS-events were recorded, with a 42% probability of clinical progression. By Cox model, independent predictors of poor clinical outcome were PML (AOR 3.72; 95%CI 2.13 to 6.51) and CNS lymphoma (5.25; 3.12 to 9.79). Starting HAART (time-dependent covariate) independently predicted a low probability of clinical progression (0.23; 0.16 to 0.34). Specific antiviral drug (including CNS-penetrating antiretrovirals) did not have any specific effect on both viro-immunological and clinical outcome.

Conclusions:  A high proportion of advanced HIV-infected patients with CNS-D presented favorable virological response, with immunological recovery observed in half of the patients. Mortality, especially due to specific neurological disorders, significantly affected treatment efficacy:  6-month virological response allows us to identify patients with a better probability of success to HAART.