Monitoring of viral RNA has become indispensable for the management of HIV-1 infection, but is expensive. This study investigated whether a highly improved test for p24 antigen could serve as an alternative. Thirty-four patients enrolled during 1997 into two treatment studies were tested prospectively for viral RNA by the Roche HIV-1 Monitor and for p24 antigen using signal-amplification-boosted ELISA of heat-denatured plasma. P24 antigen was detectable in 75.8% of 178 samples and HIV RNA in 73.9% of 138 samples. The half-life of p24 antigen in the first phase of effective treatment was 1.6 +/-.4 days (RNA, 1.7 +/-.8). An apparent second, slower decay phase had a half-life of 42 +/- 16 days. Treatment failure occurred in 14 patients. Secondary treatment failures with RNA rebounds from undetectable levels to < or = 10(3) copies/ml in two patients with an undetectable viral load and 10(3) HIV RNA copies/ml, respectively, at baseline were not detected by p24 antigen but carried a low risk for secondary resistance mutations. The other 12 failures were on average detected 29 days earlier by p24 antigen than by RNA (P =.0204), owing to slightly more frequent testing for p24 than for RNA (2.7 vs. 2.4 tests). Average costs for p24 antigen testing up to a failure were only 20.5% of those for RNA (P <.0001). These results indicate that heat-denatured, amplification-boosted p24 antigen measurement can be used as a simple and inexpensive alternative to HIV RNA testing for monitoring treatment.