Objective. Human immunodeficiency virus type 1 (HIV-1)-associated progressive encephalopathy (PE) is a common and devastating complication of HIV-1 infection in children, whose risk factors have not yet been clearly defined. Regardless of the age of presentation, PE shortens life expectancy. Paradoxically, as survival of patients has been prolonged as a result of the use of antiretroviral therapy, the prevalence of PE has increased. Therefore, a predictive marker of PE emergence is critical. The objective of this study was to determine in an observational study whether any immunologic (CD[4.sup.+] and CD[8.sup.+] T-lymphocyte counts, monocyte counts) or virologic (viral load [VL], biological characteristics of viral isolates) marker might be predictive of PE and whether any particular marker may be involved in the timing of clinical onset of PE. Methods. A total of 189 children who were vertically infected with HIV-1 were studied retrospectively, 58 of whom fulfilled criteria of the American Academy of Neurology for PE. T-lymphocyte subsets and monocytes in peripheral blood were quantified by flow cytometry. HIV-1 RNA was measured in plasma using a quantitative reverse transcriptase polymerase chain reaction assay. Demographic, clinical, and viro-immunologic characteristics in infants were compared with control groups using logistic regression. Proportions were compared using the [chi square] test or Fisher exact test. For each child, immunologic and virologic markers were analyzed in parallel closely before clinical onset of PE and closely after PE onset and compared by using the Student t test for paired samples. Results. Overall, mortality of 58 HIV-1-infected children who developed PE was significantly higher than of children who did not develop this complication. Blood CD[8.sup.+] T-lymphocytes 25% (95% confidence interval: 1.2-13.9) and remained statistically significant after adjustment for treatment. When we compared the PE-positive group with the acquired immunodeficiency syndrome (AIDS)/PE-negative group (children who developed clinical category C and without neurologic manifestations) in a cross-sectional study within 12 months before PE or AIDS diagnosis, respectively, the %CD[8.sup.+] T-lymphocytes were significantly lower in the PE-positive group. Normalized absolute counts of CD[8.sup.+] T-lymphocytes with respect to seroreverting children were significantly lower in the group of children with encephalopathy with respect to the AIDS/PE-negative group (data not shown). It is interesting that a statistically significant increase was observed in circulating monocyte percentages and absolute counts shortly before the first neurologic symptoms compared with values after PE was established and with those from HIV-1-infected controls. With respect to AIDS-related events, PE was strongly associated with anemia and lymphoid interstitial pneumonitis in the PE-positive group with respect to a group of children with AIDS but without PE. Conclusion. HIV-1 infection of the central nervous system (CNS) remains an important clinical concern. The first step toward PE prevention in HIV-1-infected children should be directed at predicting risk of PE and thus the prompt and reliable identification of infants who are at risk for CNS disease progression. Low blood CD[8.sup.+] T-lymphocytes is a strong early predictive marker of PE emergence in vertical HIV-1 infection. Indeed, among all of the immunologic and virologic variables assessed in this observational study, the only significant difference during the first months of life are the CD[8.sup.+] T-lymphocytes. A peak of significantly higher peripheral monocytes before the onset of PE with respect to established PE has not been previously described, and strengthens the growing evidence that an increased traffic of monocytes to the brain may be a key factor in triggering neurologic symptoms. The suppression of HIV-1 replication is dependent on the presence of a relatively small number of HIV-1-specific CD[8.sup.+] T-lymphocytes, and it is possible that the duration of the neurologically asymptomatic phase for any given child may depend mostly on the magnitude of specific CD[8.sup.+] T-lymphocyte responses. Thus, a decrease of CD[8.sup.+] T-lymphocytes would diminish the host capacity to control viral infection, as reported in animal models, enabling infected macrophages to cross the blood-brain barrier. Our results advocate the use of CD[8.sup.+] T-lymphocyte and monocyte counts to follow-up HIV-1-infected children. We suggest that CD[8.sup.+] T-lymphocytes may be the nexus for many different aspects of the disease, namely loss of control of HIV-1 replication determining higher VL, increased traffic of activated and/or infected monocytes, spread of infection to immune sanctuaries, and finally clinical neurologic emergence of PE. Moreover, we suggest that CD[8.sup.+] T-lymphocytes or/and monocytes may be used as putative biological markers of neuropathogenicity. This might suggest their use in decision making of when to start more effective antiretroviral regimens for HIV-1 infection of the CNS and the need of new therapies either to preserve or to augment an adequate CD[8.sup.+] T-lymphocyte immune response. Early detection of children who are at risk for developing PE is particularly important because aggressive highly active antiretroviral therapy improves neurologic symptoms, allows possible use of neuroprotective treatment to prevent further development of encephalopathy, and emphasizes the relevance of developing therapies aimed to enhance CD[8.sup.+] T-lymphocyte function. In conclusion, the surrogate markers routinely used in clinical practice for HIV-1 infection (ie, CD[4.sup.+] T-lymphocyte counts and VL) seem to be insufficient to evaluate the clinical involvement of the CNS. Other systemic markers, as the recent proposed markers for PE evolution (cerebrospinal fluid VL by lumbar puncture and brain atrophy by cerebral magnetic resonance imaging) are undoubtedly more invasive than measuring CD[8.sup.+] T-lymphocyte and monocyte counts, when the neurologic manifestations of PE are still preventable. Pediatrics 2003;111:e168-e175. URL: http://www. pediatrics.org/cgi/content/full/111/2/e168; central nervous system, HIV-1-related encephalopathy, development, pediatric AIDS, viral load.