Variation in the clinical stage at which AIDS is diagnosed has hindered the
ability of investigators to generate survival estimates which are stable a
cross study cohorts. As a result, little is known about how clinical and so
ciodemographic factors are associated with survival, independent of AIDS di
agnosis stage. By estimating survival following seroconversion while adjust
ing for baseline CD4 lymphocyte count, the present study generated survival
determinants which were unconfounded by time-related changes in AIDS diagn
osis. This study's findings indicate that the following factors were associ
ated with significant decreases in HIV-related survival: older age; self-re
port of no known HIV transmission risk factors; and presence of cytomegalov
irus. Mycobacterium avium complex, and Pneumocystis carinii pneumonia. Furt
hermore, survival decreased in a monotonic fashion with decreases in baseli
ne CD4, count and with increases in calendar period. While this study's fin
dings are consistent with previous investigators' reports of AIDS survival
determinants, it will be important for future investigators to refine and u
pdate estimates of HIV-related survival determinants as clinical care for H
IV-infected patients continues to improve.