Factors associated with poor immunologic responses despite viral suppression in markedly immunosuppressed patients.

To determine the factors associated with poor immunologic responses despite viral suppression in markedly immunocompromised patients (</=200 CD4 cells per microliter at baseline), individuals with less than 100 CD4 increases after 12 months of nelfinavir-based highly active antiretroviral therapy (HAART; n = 114) were compared to those with 100 or more CD4 increases (n = 338). Responders were more likely to be naïve to antiretroviral therapy (p = 0.009) and to protease inhibitors (p< 0.001), less likely to have undetectable viral load at baseline (p = 0.01), to be infected through injection drug use (p = 0.02), to receive two versus three daily nelfinavir doses (p = 0.05), and to have higher viral load (p< 0.001) and lower CD4 counts (p = 0.03) at baseline than nonresponders. Logistic regression analysis revealed that baseline viral load (odds ratio [OR] 0.76, 95% confidence interval [CI] 0.64-0.91), two daily nelfinavir doses (OR 1.85, 95% CI 1.12-3.03) and prior treatment with protease inhibitors (OR 1.85, 95% CI 1.11-3.07) were significantly predictive of poor immunologic responses. On the contrary, viral load had no predictive role in antiretroviral therapy (ART)-naïve patients (p = 0.7). We conclude that poor immunologic responses are common in these patients. Lower viral load at baseline is associated with poorer responses, which could be related to suboptimal prior antiretroviral therapy. In patients taking nelfinavir, administration in two versus three daily doses is also associated with worse immunologic responses.