Objectives: To monitor changes in the distribution of time intervals f
rom HIV seroconversion to the onset of AIDS and to death and to descri
be factors associated with the length of these intervals, through a na
tional register of persons with estimated dates of HIV seroconversion.
Design and methods: Clinicians caring for HIV-positive individuals an
d laboratories performing HIV testing throughout the UK were asked to
identify all persons aged 16 years or over with a history of a negativ
e HIV antibody test within 3 years of their first positive test, inclu
ding those who had died, transferred to other centres or who had becom
e lost to follow-up. Baseline and follow-up information collected annu
ally includes: sex, ethnic group, likely route for HIV transmission, l
atest CD4 count, details of antiretroviral therapy and prophylaxis for
opportunistic infections, AIDS-defining events and vital status.Resul
ts: The analysis in this report was censored at 31 December 1994 and i
ncludes 961 individuals with verified previous negative antibody tests
. Within 10 years of seroconversion, we estimate the probability of pr
ogressing to AIDS to be 60.2% [95% confidence intervals (Cl), 52.1-68.
3] and of death from any cause to be 48.1% (95% Cl, 40.1-56.1%). Older
age at seroconversion was found to be associated with faster progress
ion to AIDS (P = 0.008) as well as shorter survival (P = 0.001). No ev
idence of a change in the incubation period nor of a survival benefit
was observed by calendar time from 1983 to 1994. Conclusions: Prelimin
ary results from the UK Register confirm the strong influence of age o
n disease progression in line with findings from a number of studies.
The Register has succeeded in accruing information on a large number o
f seroconverters, and will continue to monitor the clinical course of
HIV disease, including persons infected in the 1990s. This is importan
t as a number of clinical trials have recently reported an improvement
in survival for persons on multiple drug regimens, the long-term impa
ct of which can only be monitored through observational studies. (C) 1
998 Lippincott-Raven Publishers.