Causes of death in patients treated with ART 1996-2006:

1
Causes of death in patients treated with ART
1996-2006 Collaborative analysis of 13 Cohort
Studies John Gill, Margaret May, Charlotte
Lewden, Michael S. Saag, Ross Harris, Matthias
Egger, Peter Reiss, Bruno Ledergerber, Amanda
Mocroft and Jonathan A.C. Sterne on behalf of the
Antiretroviral Therapy Cohort Collaboration
(ART-CC)
Corresponding author
Poster 708
John Gill
John.Gill_at_albertahealthservices.ca
ART
Cohort Collaboration
Background An increasing proportion of deaths
in patients on ART are from causes not classified
as AIDS-related. Methods Data were combined
from 13 cohort studies in Europe and N America
(ART Cohort Collaboration). We reviewed available
information (ICD codes, characteristics at the
time of death) on 1876 deaths among 39,272
patients aged gt15years, who started triple ART
during 1996-2006. Two or more reviewers
classified causes according to the CoDe
classification, with disagreements resolved after
discussion. Results The table shows specific
causes, which could be assigned for 1,597 (85)
deaths. AIDS deaths (46 infection, 30
malignancy, 24 unspecified) were 63 of deaths
in the first year of ART, 43 thereafter. The
table also shows hazard ratios (HR) (adjusted for
sex, age, viral load, AIDS, year of starting ART
and cohort) comparing those with lower with
higher baseline CD4 and patients infected via IDU
with others. Median baseline CD4 was 217 cells/?L
(IQR 94-343) in survivors and 110 (33-247) in
those who died. Median time to death was 1.8
(0.6-3.8) years. In addition to AIDS,
non-AIDS malignancy and renal failure appeared
associated with immunodeficiency at baseline.
Rates of non-AIDS infection, liver-related,
non-AIDS malignancy, violent, heart/vascular and
respiratory deaths were markedly elevated in
patients infected via IDU. Rates of liver-related
death declined (p 0.03) with year of starting
ART 0.83 (95 CI 0.59-1.16) per 1000 years in
1996-7 declining to 0.42 (0.14-1.32) in 2004-6.
Overall mortality rates increased with age, with
very strong associations of older age with
non-AIDS malignancy and CVD death. Conclusions
To achieve further declines in mortality rates
among patients treated with ART, causes of
non-AIDS death must be addressed. Such causes are
of particular importance in patients infected via
IDU.

• ART has dramatically improved life expectancy for
  HIV-infected patients.
• ART-related reductions in rates of mortality
  result primarily from reductions in deaths from
  AIDS- or HIV-related conditions.
• Mortality still remains higher than that of the
  general population and an increasing number of
  deaths are from causes not conventionally
  considered HIV-related.
• Large clinical endpoint trials such as SMART and
  START now incorporate non-AIDS events as
  endpoints.
• With HIV-infected persons living substantially
  longer, and the introduction of even more
  effective and tolerable drugs, it is important
  both to monitor causes of death and to assess
  risk factors for different causes.
• To examine the distribution of causes of death
  among HIV infected patients who initiated
  combination antiretroviral therapy (ART) between
  1996 and 2006, and to quantify associations of
  prognostic factors with cause-specific mortality.

Abstract
Background
Results
Methods

• Retrospective review of all deaths in 39,272
  patients agedgt15 years enrolled in 13
  observational cohorts who were (154,667 person
  years of follow up).
• Deaths were classified into the categories
  specified in the Cause of Death (CoDe) project
  protocol (www.cphiv.dk), using both a computer
  algorithm and a team of clinicians and
  epidemiologists.
• CVD included myocardial infarction (MI),
  ischaemic heart disease (IHD), stroke, heart
  failure, unspecified and other heart disease
• Violent deaths included suicide, overdose,
  accident and unspecified violent deaths
• Adjusted hazard ratios were estimated using Cox
  models

• 1,876 (4.8) patients died 1,597 (85) of these
  deaths were assigned a cause according to the
  CoDe categories table in abstract shows the
  number of deaths in each of these categories.
• Rates of specific causes of death varied
  considerably according to patients
  characteristics at initiation of ART and the
  duration of therapy.
• Rates of AIDS, the leading cause of death,
  declined with time on ART, but rates of non-AIDS
  causes of deaths remained approx. constant with
  time on ART.
• Immunodeficiency was associated with deaths from
  AIDS, non-AIDS malignancy and renal failure.
• A diagnosis of AIDS before starting ART was
  associated with death from both AIDS and non-AIDS
  infections.
• IDU had higher mortality and an increased risk of
  all specific causes of death, with a particularly
  high hazard of deaths from liver-related disease
  and violence.
• Overall mortality rates increased with age, with
  very strong associations of older age with
  non-AIDS malignancy and CVD death.

Results
Figure 2 Adjusted1 hazard ratio (95 CI) for
risk factors at start of ART with specific causes
of death CD4 cell count (per 100 cell decrease),
transmission risk group (IDU v. non-IDU), viral
load (log HIV-1 RNA gt5 v. lt5 log copies) and
prior AIDS diagnosis (v. no AIDS).
Objective
Results
Conclusions
Figure 1 Bar graph showing crude cause-specific
mortality rate according to length of time since
start of ART

• ART continues to have a dramatic impact in
  reducing rates of mortality from HIV infection in
  high-income countries.
• AIDS remains the most common cause of death.
• The strong inverse association of rates of AIDS
  death with CD4 counts at the time of starting ART
  supports arguments for earlier initiation of ART.
  
• Conditions associated with social and lifestyle
  factors contribute the next most frequent causes
  of death, with violence and liver related
  diseases (mainly due to hepatitis) contributing
  15 of all deaths.
• The importance of lifestyle is reinforced by the
  observations that the malignancies were mainly
  lung cancer and likely associated with smoking,
  and that many endocarditis cases occurred in
  patients infected via IDU.

1Models were mutually adjusted for age, sex, IDU,
CD4, viral load, prior AIDS diagnosis, cohort and
year of starting ART
ART-CC Contributing Cohorts UAB 1917 Clinic
Cohort (Birmingham, Alabama, USA), Aquitaine
(France), ATHENA (AIDS Therapy Evaluation Project
Netherlands), BCCfE-HIV (British Columbia Centre
for Excellence in HIV/AIDS, Canada),
CHORUS/Vanderbilt (Collaborations in HIV Outcomes
Research US), CoRIS (Cohorte de la Red de
Investigacion en Sida RIS, Spain), EuroSIDA,
Frankfurt Klinikum der JW Goethe-Universität
Frankfurt, Germany), FHDH (French Hospital
Database on HIV), HAVACS (HIV Atlanta Veterans
Affairs Cohort Study, US), ICONA (Italian Cohort
of Antiretroviral-Naive Patients), Köln/Bonn
(Departments of Internal Medicine at University
of Cologne and Bonn, Germany), PISCIS (Proyecto
para la Informatización del Seguimiento
Clínico-epidemiológico de la Infección por HIV y
SIDA, Spain), Royal Free (Ian Charleson Centre at
the Royal Free Hospital London, UK), SHCS (Swiss
HIV Cohort Study), South Alberta (Southern
Alberta Clinic, Canada), VACH, VACS
(Observational cohort study of HIV-positive and
matched HIV-negative veterans based on the
Veterans Health Administration, USA), Vancouver
(St. Paul's Hospital in Vancouver, Canada),
Washington (University of Washington Harborview
Medical Center, USA). ART-CC Steering
Committee Jordi Casabona (PISCIS), Geneviève
Chêne (Aquitaine), Dominique Costagliola (FHDH),
François Dabis (Aquitaine), Antonella DArminio
Monforte (ICONA), Frank de Wolf (ATHENA),
Matthias Egger (SHCS), Gerd Fatkenheuer
(Köln/Bonn), John Gill (South Alberta Clinic),
Robert Hogg (BCCfE-HIV), Amy Justice (VACS), Mari
Kitahata (Washington), Bruno Ledergerber (SHCS),
Ole Kirk (EuroSIDA), Peter Reiss (ATHENA),
Michael Saag (UAB 1917 Clinic Cohort), Fiona
Lampe (Royal Free), Hans-Reinhard Brodt
(Frankfurt), Julia del Amo (CoRIS-MD), Myriam
Garrido (VACH), Jodie Guest (HAVACS), Tim
Sterling (Vanderbilt). study of HIV-positive and
matched HIV-negative veterans based on the
Veterans Health Administration, USA), Vancouver
(St. Paul's Hospital in Vancouver, Canada),
Washington (University of Washington Harborview
Medical Center, USA). ART-CC Co-ordinating
Team Jonathan A. C. Sterne (PI), Margaret
May. ART-CC Sources of Funding UK Medical
Research Council (MRC), GlaxoSmithKline, Agence
Nationale de Recherches sur le Sida (ANRS),
Institut National de la Santé et de la Recherche
Médicale (INSERM), French, Italian, Swiss
Ministries of Health, Dutch Stichting HIV
Monitoring, European Commission, British Columbia
and Alberta Governments, Michael Smith Foundation
for Health Research, Canadian Institutes of
Health Research, Roche, Boehringer-Ingelheim, St.
Paul's Hospital Foundation, UAB CFAR Network of
Integrated Clinical Sciences (C-NICS), NA-ACCORD,
NIH, FIPSE