HIGH MORTALITY AMONG HIV-INFECTED INJECTING DRUG-USERS WITHOUT AIDS DIAGNOSIS - IMPLICATIONS FOR HIV-INFECTION EPIDEMIC MODELERS

Objective: To quantify mortality without AIDS diagnosis (pre-AIDS deat h) among HIV-infected injecting drug users (IDU) and to compare it wit h that among homosexual men. Setting: Municipal Health Service in Amst erdam, The Netherlands. Design: HIV-infected participants of a cohort of homosexual men (recruited since 1984) and a cohort of IDU (recruite d since 1985) were followed up to investigate the occurrence of AIDS a nd pre-AIDS death up to 1 October 1992. Both seroprevalent and serocon verted participants were included. Methods: Product-limit estimates of cumulative AIDS incidence and cumulative pre-AIDS mortality were calc ulated for a competing risks model. Results: Of 455 HIV-infected homos exual men, 127 developed AIDS and two died without a diagnosis of AIDS during follow-up; 31 out of 279 HIV-infected IDU were diagnosed with AIDS and 33 died without an AIDS diagnosis. After 6.5 years of follow- up an estimated 43.7% of the homosexual men had been diagnosed with AI DS and 0.7% had died without an AIDS diagnosis, while 32.7% of the IDU had been diagnosed with AIDS and 19.8% had died without an AIDS diagn osis. Conclusions: After 6.5 years of follow-up one-fifth of a group o f HIV-infected IDU had died without being diagnosed with AIDS. This ph enomenon seriously limits the number of HIV-infected IDU that may ever be diagnosed with AIDS and should be considered by HIV infection epid emic modellers. The new case definitions for AIDS will limit the numbe r of recorded cases of pre-AIDS death.