Evaluation of Non-Viral Surrogate Markers as Predictive Indicators for Monitoring Progression of Human Immunodeficiency Virus Infection: An Eight-Year Analysis in a Regional Center.

Suitable methods for clinical monitoring of HIV-infected patients are crucial in resource-poor settings. Demographic data, clinical staging, and laboratory findings for 112 asymptomatic subjects positive for HIV were assessed at the first admission and the last visit from 2002 to 2010. Cox regression analysis showed hemoglobin (Hb) (HR = 0.643, P = 0.021) to be a predictive indicator for disease progression, while CD4, CD8, and platelet counts showed low HRs, despite having significant probability values. Hb and total lymphocyte count (TLC) rapidly declined from stage II to III (10.9 and 29.6%, respectively). Reduced CD4 and platelet counts and Hb during stage I were associated with disease progression, and TLC was correlated with CD4 counts at the last follow-up (P < 0.001). However, WHO TLC cutoff of 1,200 cell/mm(3) had 26.1% sensitivity and 98.6% specificity. ROC curve analysis suggested that a TLC cutoff of 1,800 cell/mm(3) was more reliable in this region. Statistical analysis and data mining findings showed that Hb and TLC, and their rapid decline from stage II to III, in addition to reduced platelet count, could be valuable markers for a surrogate algorithm for monitoring of HIV-infected subjects and starting anti-viral therapy in the absence of sophisticated detection assays.