Transcript Diapositive 1

Severe morbidity among HIVinfected patients :
a comparison between a Brazilian
and a French clinic based
observational cohort
• FIOCRUZ: Prof B Grinsztejn
• INSERM: Prof G Chêne
• Joint call INSERM-FIOCRUZ 2009-2011
1
Outline
• Background
• Questions and objectives
• Methods
– Definitions
– Classification
– Events validation and data quality
• Preliminary results
• Timelines
Background
• Generalized access to combination AntiRetroviral Therapy (cART)
–
–
–
–
dramatically improved outcome
disease progression remains highly variable
shift from AIDS- to non AIDS-related mortality/morbidity
ageing, hepatitis C co-infection, tobacco smoking and
other addictions, long-term exposure to antiretrovirals
– virus mutations, introduction of new drugs/drug classes
• Important to continuously assess changes and
their impact on disease progression, though
life expectancy still not at the level of general
population
Questions
• Trends of the severe morbid process?
– Causes of severe morbidity: AIDS, cardiovascular, cancer, non AIDS infections, etc.
poorly described so far
– Intercontinental comparisons to explain
variability in the distribution
• Access to therapy
• Baseline characteristics
– HIV-related (CD4, plasma HIV-RNA)
– Others (Co-infections, age, gender,
transmission group,..)
• Environment
Objectives
• In the setting of two large cohorts of HIV-infected
patients: one from southwestern France (ANRS CO3
Aquitaine) and one from Rio de Janeiro Brazil
(IPEC/Fiocruz HIV Clinical Cohort) where all severe
events are systematically and prospectively recorded
and coded according to the International Classification of
Diseases 10th revision (ICD10),
• we will aim at studying the repartition and the evolution
of causes of severe morbid events occurring in HIVinfected patients during the period 2000-2008. In
addition, the role of potential determinants including age,
gender, immunodeficiency and uncontrolled viral load,
main classes of antiretrovirals and co-infections will also
be estimated.
Methods: definitions
• Inclusion criteria: all patients with ≤1 follow-up
(January 2000 – December 2008)
• Outcome: Occurrence of a severe morbid event
– Severe morbidity: morbid condition leading to at least
48 hrs of hospitalization, or death (as many events as
causes of hospitalization for a given hospital stay)
– May not be considered as severe morbid events:
• Hospitalizations <48 hrs
• Hospitalizations due to check-up, planned
chemotherapies,…whatever their duration
• Coding of the morbid events:
– ICD 10 classification (IPEC Cohort)
– Simplified version of the ICD 10 (Aquitaine Cohort)
Classification: methods
• Exclusive classification with decreasing priority:
–
–
–
–
AIDS events
Non AIDS cancer
Infectious events
Systemic events
• IPEC Cohort:
– All discharge charts of hospitalization were reviewed
to identify morbid events
– Each event was coded (ICD10) and classified
• Aquitaine Cohort:
– Each code of the thesaurus (simplified version of the
ICD10) corresponding to events codes was attributed
to one category
– Codes colligated in the database corresponding to
morbid events were extracted and classified in the
corresponding category
Classification: categories
• AIDS events
• Non AIDS cancers
– Invasive tumors
– In situ tumors
• Infectious events
– Bacterials
– Virals
– Parasitic
• Systemic events (1)
– Cardio-Vascular
– Hepatic
• Viral-related
• Non viral related
• Systemic events (2)
–
–
–
–
–
–
–
–
–
–
–
–
Digestive
Psychiatric
Haematological
Kidney & Urological
Endocrinal
Dermatological
Gynecological
Neurological
Ophtalmologic
Respiratory
Rheumathologic
Traumatic
• Others
Validation process
• Comparison of the list of codes used by the two
cohorts in each category: all discrepancies were
discussed
– In main instances an agreement was found
– For some specific codes it was decided to consider
them in different categories in both cohorts: ie K52.9
• IPEC: Chronic diarrhea: Digestive
• Aquitaine: Gastro-enteritis rectosigmoïditis: Bacterial
• Validation of the events (through medical files):
– A specific form was established, to be used by both
cohorts
– IPEC: 10% of the events validated
– Aquitaine: 1% of the events validated
Other variables
• A specific SOP was established to merge the
data of both Cohorts
• Variables:
– Demographics (Age, gender, educational level,..)
– HIV Related (Risk group, Plasma HIV RNA, CD4,
treatments,..)
– Risk Factors (Tobacco, alcohol, eepatitis coinfections,..)
• Checks will be performed after the merger to
assess data quality
Preliminary results
• Aquitaine Cohort
– 5553 eligible events in the database (2000-2008)
• 845 AIDS events
• 4708 Non-AIDS events: Bacterial infections (20%); Neurologicals
(9%); Hepatitis (9%); Hematological (8%); Psychiatric (8%);
Digestive (7%);..;Non-AIDS cancer (2.4%).
• IPEC-Fiocruz Cohort
– 2782 eligible events in the database (2000-2008)
• 1095 AIDS events (40%) 1687 Non-AIDS events (60%): Bacterial
infections (46%); Hepatitis (2,3%); Psychiatric (5,4%); Non-AIDS
cancer (0,6%).
• 118 codes discrepancies between the cohorts discussed
– Codes used only in one cohort
– Codes used in both cohort but initially classified in different
categories
Timelines
• Achieved:
– Final decisions concerning
• The definition of a severe morbid event
• The classification of events to consider (list of categories)
– Validation of the events
• Ongoing:
– Generation of the tables for the data merger and data
check
• Next stages:
– Spring 2011: final data merger and check
– Summer 2011: analyses
– Autumn 2011: final discussion and draft of a first
abstract (CROI 2012) and manuscript