We have previously noted an association between proviral load and the severity of immune disease in individuals with a wide range of CD4 cell counts. Using the quantitative DNA polymerase chain reaction technology developed in our laboratory, we sought to extend these observations, with a view to establishing guidelines for the use of proviral load in a clinical context. We studied 199 patients with a range of CD4 cell counts attending an urban tertiary care center. Proviral load/10(6) peripheral blood mononuclear cells (PBMCs) was measured using a microtiter plate assay designed specifically for this purpose. Human immunodeficiency virus proviral DNA was detected in 193 of 199 clinical samples. Levels of proviral load were tabulated for patients and evaluated in seven categories defined by CD4 cell counts. Although a wide range of proviral loads was observed in each category of patients, there was a trend toward increasing proviral load with decreasing CD4 cell count. Statistically significant relationships were observed between proviral load and the CD4 cell count and the CD4 cell percentage (Spearman's correlation coefficient -0.19, p = 0.01 for both absolute CD4 and CD4 percentage). These relationships were quite weak and could not be taken to explain disease progression in isolation. If we defined a cutoff between low and high proviral loads at 100 copies/10(6) PBMCs, we noted that 52% (24 of 46) of patients with CD4 cell counts > 400/microliters had lower loads, as compared with 16% (24 of 143) of those with more advanced disease (p < 0.01). There is a weak, but statistically significant association between proviral load and CD4 cell depletion.(ABSTRACT TRUNCATED AT 250 WORDS)