Use of risk equations for predicting disease progression in HIV infection.

In this issue of Clinical Infectious Diseases, Srasuebkul et al. [1] propose models for identifying HIV-1–infected patients at short-term risk of AIDS or death in Asia and resource-limited settings. The models are based on clinical criteria (including body mass index, development of anemia, and age), CD4 cell count criteria (CD4 cell count was incorporated in the risk strata), and CD4 cell count and HIV load criteria (HIV load was incorporated in the risk strata). Three models were developed to enable clinicians to choose the most appropriate model on the basis of the availability of laboratory results, such as CD4 cell count and HIV load, within resource-limited settings. The Centers for Disease Control and Prevention has proposed various classification systems for HIV-infected patients [2–4], but these have always been intended for use in surveillance rather than to provide prognostic information. Staging systems for HIV infection have previously been proposed. The systems developed in the pre–combination antiretroviral therapy (ART) era were based on clinical criteria and CD4 cell count [5–7] but were often criticized because they could not be widely implemented in developing countries. Staging systems, primarily the World Health Organization staging system or a modified version of it, were therefore developed and applied to patients in developing countries [8–11]. The World Health Organization staging system was based on a combination of clinical and biological parameters, with a clinical and laboratory axis that allowed CD4 cell count to be replaced by total lymphocyte count in regions where resources were not available for CD4 cell count determinations. Since the introduction of combination ART and the improvement in prognosis, various new prognostic staging systems have been proposed on the basis of information available on the date of initiation of combination ART [12, 13], while short-term prognostic staging systems have been based on information available during follow up [14, 15]. The staging system proposed by Srasuebkul et al. [1] is, therefore, one of the first post–combination ART prognostic staging systems that have been developed and applied among patients from resource-limited settings. There is an important distinction between models that predict the short-term risk of disease progression and those that predict the long-term risk of disease progression. In general, previous models have generally concentrated on predicting long-term clinical progression on the basis of information known on the date of combination ART initiation [16, 17]. These scores identify groups of patients with elevated risk of disease progression …