 |
|
|
| |
Relationships between markers of immunodeficiency and non AIDS related causes of death in cART era, CASCADE Study Group
|
| |
| |
Reported by Jules Levin
IAS Conference, July 2007, Sydney, Australia
Marin B.1 2, Thiebaut R.3, Rondeau V.3, Costagliola D.4, Dorrucci M.5, Bucher H.6, Hamouda O.7, Walker S.8, Bhaskaran K.8, Chene G.1, CASCADE Study Group
1INSERM U593, Bordeaux, France, 2UFRCB, CHU Limoges, France, 3INSERM U875, Bordeaux, France, 4INSERM U720, Paris, France, 5Istituto Superiore di Sanita, Rome, Italy, 6Basel Institute for Clinical Epidemiology, Basel, Switzerland, 7Robert Koch-Institute, Berlin, Germany, 8MRC Clinical Trial Unit, London, United Kingdom
Comments - conclusion
--There is a clear relationship of either latest or nadir CD4 count with four specific causes of death : those which were AIDS-related, and among those which where not AIDS-related: severe infections, liver diseases, malignancies. A clear gradient effect is observed for these causes
--Deaths due to cardiovascular disease are not immunodeficiency-related nor suicide
--These results are consistent when we considered the relationship with time spent under 350 cells x106/l with a gradient effect
--A higher HIV RNA before cART was associated with AIDS deaths, non AIDS infections and liver diseases deaths
--These results plead for earlier initiation of antiretrovirals to reduce the impact of the most frequent specific causes of death
--Morbidity related to non AIDS-conditions should be explored in observational studies in order to better describe the current disease history under cART
Introduction
In the cART era, epidemiological studies have shown
--significant reduction of death rates
--more than half of deaths are non AIDS-related
--association between latest markers of HIV infection and specific non AIDS deaths (end stage liver diseases, malignancies)
Emerging questions
--Is nadir CD4 count stronger determinant than latest values?
--Is time spent under 350 cellsx106/l a valuable prognosis factor?
--If a relationship exists with any marker of immunodeficiency, is there a gradient?
Objectives
To assess adjusted relationships between markers of immunodeficiency and specific causes of death
3 models were compared using
1. Latest CD4 count (usual approach, to be compared to others methods)
2. Nadir CD4 cell count (before cART and during follow-up)
3. Time spent under 350 cellsx106/l
Methods (1)
Population study
- CASCADE Collaboration : 23 seroconverters cohorts from Europe, Australia, Canada
http://www.ctu.mrc.ac.uk/cascade
- Patients with available follow-up > 1996 (cART era),
aged 16 years and more
- Availability of biological markers
Causes of death
- Standardized according to Ç Coding of Death in HIV
(CoDe)È http://www.cphiv.dk
Methods (2)
Context of competing risks
- Modelling of cause specific hazards ratios based on Cox proportional hazards modelling
- Baseline time : time from seroconversion to death
or censoring date
Variables
- Biological markers and time under 350 cellsx106/l: time varying covariates
- Adjustment for : age at seroconversion date, sex, mode of HIV transmission, Hepatitis C serostatus and first line cART treatment (PI-regimen, other cART vs no cART)
Results (1)
Table 2: Causes of deaths
Aids-related deaths: 26.5%
Non-AIDS infection: 8.4%
Liver disease: 7.7%
Non-AIDS cancer: 7.7%
Cardiovascular disease: 6%
Suicide: 6.4%
Substance abuse: 5.7%
More than half of the 597 deaths were non AIDS-related.
Overall 9,858 patients were followed for a median of 8 years
after seroconversion. 2/3 initiated cART during follow-up.
|
| |
|
|
|
|
|
