Mortality among Participants in the MACS and WIHS: Impact of Accidents and Injury

1
Mortality among Participants in the MACS and
WIHSImpact of Accidents and Injury

• Mardge Cohen Core Center, Cook County
  Bureau of Health Services, Chicago IL
• Ann Kalinowski Dept of Med, University of
  California, San Francisco, CA
• Joanne Mullen Dept Epi, Johns Hopkins
  Bloomberg School Pub Health, Baltimore MD
• Frank Palella Northwestern
  University, Feinberg School of Med, Chicago IL
• Kathryn Anastos Montefiore Medical Center,
  Bronx, NY
• Lorie Benning Dept Epi, Johns Hopkins
  Bloomberg School Pub Health, Baltimore MD
• Nancy Hessol Dept of Med, University of
  California, San Francisco, CA

2
Background

• Mortality from HIV infection declined
  dramatically in developed countries over past
  decade due to highly active antiretroviral
  therapy (HAART)
• While deaths due to AIDS have decreased, deaths
  due to non-HIV causes increased proportionally
• Causes of mortality have also changed with HAART
  use

3
Background

• Cohort studies of HIV infected persons show
  non-HIV causes of death in both pre- HAART and
  HAART era related to
• Suicides and deaths by drug overdose
• Hepatitis C and end stage liver disease
• Depression
• Poor health care utilization

4
Study Objectives

• To examine accidental injury causes of death
• among women in the Womens Interagency HIV Study
• (WIHS) and men in the Multicenter AIDS Cohort
  Study
• (MACS) to determine if
• 1. Risk of accident- or injury-related mortality
  changed over time, with HAART availability
• 2. HIV infection or immune status affected this
  risk
• 3. Risk or causes of mortality differ between
  men women
• 4. Risk factors, such as depression and
  injection drug use, play a role in accident- or
  injury-related mortality

5
MethodsStudy Population

• Women Interagency HIV Study (WIHS)
• Prospective cohort study
• HIV and HIV- women
• 6 U.S. sites
• Bronx, Brooklyn, Chicago, LA, F, Washington DC.
• Eligibility included women with AIDS and CD4
  lt200
• HIV- women reported IDU, blood transfusions,
  multiple sex partners, or sex with an HIV
  partner
• Visits every 6 months

• Multicenter AIDS Cohort Study (MACS)
• Prospective cohort study
• HIV and HIV- men
• 4 U.S. sites
• Baltimore, Chicago, Pittsburgh, Los Angeles
• Eligibility excluded men free of clinical AIDS
• HIV negative men reported sex with men
• Visits every six months

6
Identification of Deaths

• Date and Cause of death
• Paper Death Certificates
• Electronic National Death Index (NDI)
• 2 physicians reviewed all causes of death
• Classified deaths
• Accidents
• Overdose
• Injury
• Suicide
• Homicide
• Mortality rates calculated by time period

7
Analytic Methods

• Deaths were collected through December 2002.
• Annual death rates calculated as number of deaths
  divided by time contributed by participants in
  that year (person-years)
• Exponential regression to calculate relative
  hazard of accident- or injury-related death and
  associated 95 confidence intervals for each risk
  factor.
• Separate regression analyses performed on both
  cohorts to determine suitability of pooling data.
• Pooled data analyzed using unadjusted regression
  models. Significant variables were then
  considered in multivariate analyses.

8
Results
9
Study Participant Characteristics at Baseline
10
(No Transcript)
11
(No Transcript)
12
Cause of Death Information (through December
2002)

• MACS
  WIHS
• Reported deaths 1829
  619
• Deaths from accident 51 (2.8) 46 (7.4)
• or injury

13
Underlying Causes of Death

14
Accident Injury (A/I) and all cause crude death
rates in MACS and WIHS
Death rate Deaths per 1000 person years
15
Significant risk factors for accident/injury
deaths in pooled MACS WIHS data (unadjusted
regression analysis)
16
(No Transcript)
17
Risk Factors Associated with accidental and
injury related deaths among MACS and WIHS
participants
Adjusted exponential regression hazard ratio (HR)
and 95 CL
18
Conclusions

• Women in WIHS are at a significantly higher risk
  of accident- or injury-related death than men in
  MACS
• Findings are in sharp contrast to national data
  where men have a fatal injury rate twice that of
  women
• Reflects multiple risk factors women experience
  (unsafe environment, housing, drug use,
  depressive symptoms)

19
Conclusions

• After controlling for potential confounding,
  neither HIV nor most recent CD4 cell count was
  predictive of accident or injury-related
  mortality.
• No effect over time on risk of accident or
  injury-related mortality.

20
Conclusions

• We identified potentially modifiable risk factors
  for accident- or injury-related death
• Harm reduction
• Cigarette smoking
• Injection drug use
• Treatment of depression
• Reducing unemployment

21
Limitations of Study

• Small number of accident- or injury-related
  deaths hampered ability to explore risk factor
  relationships in greater detail
• Combining all accident- and injury-related deaths
  into one analytic group limited ability to
  attribute risk to one or two particular causes of
  death
• Characteristics of the men in MACS are different
  from women in WIHS

22
Implications

• The unstable /unsafe environment that women live
  in makes them more likely to die from accidents
  and injuries
• Independent predictors of death from accidents or
  injuries
• Depressive symptoms
• Injection drug use
• Unemployment
• Cigarette smoking
• Targeting these factors for intervention can
  reduce accident and injury related deaths and
  improve health of people with HIV
• As women are the majority of people living with
  HIV globally, our study provides further evidence
  of the need to address womens complex web of
  vulnerabilities and marginalization