595 
HIV and Non-HIV-related Deaths and Their Relationship to Immunodeficiency: The D:A:D Study
Rainer Weber*1, N Friis-Mřller2, C Sabin3, P Reiss4, A D'Arminio Monforte5, F Dabis6, W El-Sadr7, S De Wit8, L Morfeldt9, M Law10, C Pradier11, G Calvo12, C Holkmann-Olsen13, A Phillips3, J Lundgren2, and On behalf of the D:A:D Study Group
1Zurich Univ Hosp, Switzerland; 2Copenhagen HIV Prgm, Hvidovre Univ Hosp, Denmark; 3Royal Free and Univ Coll, London, UK; 4ATHENA Cohort Study Group, HIV Monitoring Fndn, Academic Med Ctr, Univ of Amsterdam, The Netherlands; 5L Sacco Hosp, Univ of Milan, Italy; 6Aquitaine Cohort, Bordeaux Univ Hosp, INSERM U593, France; 7Community Prgms for Clin Res on AIDS, Harlem Hosp, Columbia Univ, New York, NY, USA ; 8St-Pierre Cohort, Ctr Hosp Univ St-Pierre, Brussels, Belgium; 9HIVUS, Karolinska Hosp, Stockholm, Sweden; 10Australian HIV Observational Database, Natl Ctr in HIV Epidemiology and Clin Res, Sydney, Australia; 11Ctr Hosp Univ Nice, Hosp de l'Archet, France; 12BASS, Autonomous Univ of Barcelona, Spain; and 13EuroSIDA, Copenhagen HIV Prgm, Hvidovre Univ Hosp, Denmark
Background:
Death in people with HIV
infection is often classified according to whether they are “HIV-related” or
not. We studied whether death generally supposed to be unrelated to HIV was in
fact more likely to occur in people with a low CD4 count.
Methods: D:A:D is a
prospective study of 23,441 patients from 11 existing cohorts in Europe, Australia, and the United States. Detailed information
on death occurring between the initiation of D:A:D in
2000 and February 2004 was collected; causes of death were coded centrally.
Relative rates of factors associated with death from each cause were calculated
using Poisson regression.
Results: Of the 23,441 persons, 82% had used ART prior to enrollment (median exposure 2.8 years). Over the study
period, 1248 (5.3%) patients died (incidence: 1.6 deaths/100
person-years). The leading cause of death was AIDS (30%) followed by
liver-related death (14%, of which 79% were associated with chronic viral
hepatitis), death from heart disease (including cardiovascular disease, 9%),
and death from non-AIDS malignancies (8%). The proportion of deaths from
these causes remained stable over time. AIDS-related death was, as expected,
strongly associated with the latest CD4 cell count (relative rates of death
from an AIDS-event in patients with latest CD4 count < 50 cells/µL vs > 500 cells/µL after adjustment for HIV RNA, age,
cohort, race; 96.4 [95% CI 61.6 to 150.7], p < 0.0001). Liver-related deaths (relative rate 26.6 [12.9 to 54.7],
p < 0.0001) and deaths from
non-AIDS malignancies (23.5 [9.4 to 58.7], p
< 0.0001) were similarly strongly associated with latest CD4 count; the
relationship with death from heart disease was less strong but remained
significant (3.1 [1.2 to 8.2], p = 0.02).
The trends persisted after lagging the analyses to associations with CD4 counts
3 and 6 months earlier.
Conclusions: Death
from causes generally referred to as non-HIV-related is more likely to occur in
persons with lower rather than higher CD4 counts. Although residual confounding
cannot be ruled out, low CD4 cell numbers seem to contribute to a proportion of
such deaths. It appears that death
occurring in persons with a very low CD4 count can only be categorized as
non-HIV-related if there is clear evidence that the patient’s immunodeficiency did not contribute to the death.
Keywords: Cause of death; immunodeficiency; mortality
