Abstract Book Vol. 2 [International Conference on AIDS (14th: 2002: Barcelona, Spain)]

98 Abstracts WePeC6057-WePeC6061

XIV International AIDS Conference

Presenting author: Sandra Schwarcz, 25 Van Ness Avenue, Suite 500, San
Francisco, CA 94102, United States, Tel.: [phone], Fax: [phone],
E-mail: [email protected]
WePeC6057 The effect of level of study and calendar period on
HIV disease progression from seroconversion in
Madrid, Spain
J. Del Romero1, A. Barrasa2, S. Prez-Hoyos3, C. Rodriguez1, J. Castilla4,
S. Garcia1, M. Dfez2, V. Soriano5, M. Raposo1, J. Del Amo6. 1Centro Sanitario
Sandoval, Servicio Madrileho de Salud, Madrid, Spain; 2Centro Nacional de
Epidemiologia, Madrid, Spain; 3 Escuela Valenciana de Salud Pliblica, Valencia,
Spain; 4Centro Nacional de Epidemiologia y Secretarfa del Plan Nacional sobre
Sida, Madrid, Spain; 5 Centro de Investigaciones Clinicas, Madrid, Spain; 6Plan
Nacional Sobre Sida, Madrid, Spain
Background: To study HIV progression from seroconversion over a 15 years period.
Methods: Cohort study of persons with well documented dates of seroconversion. Cumulative risk of AIDS and death were calculated by extended KaplanMeier allowing for late entry. Cox proportional Hazards models were used to study
variables associated with HIV progression. Calendar time was divided in 3 periods; before 1991, 1992/96 and 1997/2000.
Results: From January 1985 to May 2000, 226 seroconverters were identified.
Median seroconversion interval was 11 months, median seroconversion date was
March 1993. 202 (89%) were men, 76% of whom were homo/bisexual. A 66% reduction in progression to AIDS was observed in 1997-2000 (HR 0.34 95%CI:0.16 -0.70) compared to 1992-96. Progression to AIDS was faster in persons with
primary studies (HR 2.69 95%CI:1.17-6.16). compared to those with university
studies. An 82% reduction in mortality from HIV seroconversion was observed
in 1997-2000 (HR 0.18 95%CI:0.05-0.68) compared to 1992-96. Progression to
death for persons with primary studies was two times faster than for those with
university studies (p 0.0007). Persons without confirmation of an HIV negative
test had faster progression (HR 4.47 95%CI:1.18-16.92).
Conclusions: The reductions in progression to AIDS and death from seroconversion from 1997 to 2000 in Madrid are likely to be attributable to Highly Active Antiretroviral therapy (HAART). HIV progression was faster in subjects with primary
studies; better educational level is probably a marker of higher socio-economic
status, which may be associated with healthier lifestyles and compliance with
medication.
Presenting author: Alicia Barrasa, C/ Sinesio Delgado n06, Madrid, 28029,
Spain, Tel.: [phone], Fax: [phone], E-mail: [email protected]
WePeC6058 HIV progression in patients co-infected by HCV in
a cohort of HIV seroconverters from Madrid, Spain
A. Barrasa1, S. Perez-Hoyos2, I. Santos3, J. Del Romero4, V. Soriano5, J. Del
Amo 6. 1Centro Nacional de Epidemiologia, C/Sinesio Delgado n06, Madrid,
28029, Spain; 2 Escuela Valenciana de Salud PLiblica, Valencia, Spain; 3Hospital
de la Princesa, Madrid, Spain; 4 Centro Sanitario Sandoval, Servicio Madrilefio
de Salud, Madrid, Spain; 5Centro de Investigaciones Clinicas, Hospital Carlos
Ill, Madrid, Spain; 6Plan Nacional sobre Sida, Madrid, Spain
Background: HCV has been suggested as a co-factor for HIV disease progression. The objective of this paper is to study how HCV co-infection affects HIV
progression in a cohort of persons with well-documented dates of seroconversion
followed for 15 years.
Method: Cohort study of HIV seroconverters recruited from 1985 onwards in
Madrid. Diagnoses of HCV infection were established through a positive HCV
antibody test. Cox proportional Hazards models were used to study the association between HCV co-infection and HIV progression in persons infected for the
same length of time adjusting for potential confounders.
Results: From January 1985 to May 2000, 226 seroconverters were identified.
181 had HCV serology available of whom 84% were male, 64% homo/bisexual
and 23% injected drugs users (IDU). 15,4% developed AIDS; pulmonary tuberculosis and Kaposi Sarcoma were the commonest conditions. Forty-seven (26%)
were HCV-positive of whom 74,5% were male, 78,7% were IDU and 25,5% had
developed AIDS. Being co-infected with HCV was not associated with faster
progression to AIDS nor death in neither univariate (HR 1.43 95%C1:0.64-3.22
and HR 2.0 95%CI:0.52-7.68 respectively) nor multivariate analysis (HR 0.68
95%CI:0.27-1.74 and HR 0.78 95%CI:0.07-9.17 respectively).
Conclusions: The majority of patients co-infected by HIV and HCV are male and
IVDU. Being HCV co-infected is not associated with faster progression to AIDS
when comparing persons infected by HIV for the same length of time.
Presenting author: Alicia Barrasa, C/ Sinesio Delgado na6, Madrid, 28029,
Spain, Tel.: [phone], Fax: [phone], E-mail: [email protected]

WePeC6059 Risk of HIV-infection and AIDS-diagnosis as a
function of the time of intravenous drug use
K. Langohr, G. Gomez Melis. Politecnical University of Catalonia, Departament
d'EIO, Pau Gargallo, 5, 08012 Barcelona, Spain
Background: To analyze the association between the time from intravenous (IV)
drug initiation until HIV-infection, Z, and the time from HIV-infection till AIDS
diagnosis, Y, in a cohort of 361 injecting drug users attending a hospital detoxification unit in Badalona (Spain).
Methods: We set a log-linear model for Y as a function of Z where the presence of different covariates is taken into account and estimate the regression
coefficients of this model. Since seroconversion time, as a proxy of HIV-infection
time, cannot be determined exactly, a seroconversion interval determined from
the dates of HIV-antibody tests is used instead. Furthermore, the times to AIDSdiagnosis are severely right-censored; consequently the data are interval- and
doubly-censored.
The analysis of the model is based on the adhoc construction of the likelihood
function taking into account the different censoring patterns. The maximization of
this function is obtained by means of the mathematical programming language
AMPL, using the solver SNOPT and the internet facility 'NEOS: Solvers for Optimization'.
Results: The risk ratio and the accelerating factor of the time-to-AIDS-diagnosis
as a function of the time-to-HIV-infection allows to interpret, among other issues,
that the longer an intravenous drug user remains seronegative, the longer s/he
remains AIDS-free once s/he is HIV-infected. Also, the median time from HIVinfection till AIDS-diagnosis for women is about 1.33 times longer than for men.
Conclusions: According to the obtained results, intravenous drug addicts who
remain seronegative for a longer period are at lower risk of AIDS-development
once they are seropositive than those with short times until seroconversion. This
might be due to a good immune system resisting both HIV-infection and AIDSonset and/or hygienic precautions taken by these individuals.
Presenting author: Klaus Langohr, Departament d'EIO, Pau Gargallo, 5, 08012
Barcelona, Spain, Tel.: [phone], Fax: [phone], E-mail: klaus.
langohr@ upc.es
WePeC6060 Predictors of mortality among HIV-infected
outpatients attending an inner city clinic
S. Mannheimer1, N. Holson', L. Farrell2, B. Diamond3, Y. Hirsch-Movermani,
M. Perkins2, J. Ford', W. EI-Sadr1. 1Columbia University Harlem Hospital
Center, New York, United States; 2 Harlem Hospital Center, New York, United
States; 3 Columbia UniversityIrving Center for Clinical Research, New York,
United States
Background: Antiretroviral therapy (ART) is associated with decrease in HIVrelated morbidity and mortality. We examined a cohort of HIV-infected outpatients
at an inner city HIV clinic in New York City to evaluate for predictors of mortality.
Methods: Baseline characteristics of patients (pts) enrolled in an adherence support program were examined for association with mortality Characteristics of pts
who died during follow-up were compared to those that survived. Adherence was
measured by 3-day self-report.
Results: Of 212 pts enrolled from 3/99 to 8/01, 85% were receiving ART at baseline, 56% were women, 77% African American, 17% Latino, and 28% had a history of injection drug use. The mean length of pt follow-up was 14.2~3.9 months.
Ten pts (4.5%) died during follow-up: 4 from non-HIV causes (2 cervical cancer,
1 renal failure, 1 drug overdose), 3 from HIV-related causes (2 bacterial pneumonia, 1 PCP) and 3 from unknown causes. Predictors of mortality included older
age (46~3.9 vs.42~9.3 years, p=0.01), methadone use (50% vs. 11%, p<0.001),
having an AIDS diagnosis (90% vs. 57%, p<0.05) and a higher baseline log10
HIV RNA level (4.5~1.0 vs. 3.5~1.1, p=0.007). Although not statistically significant, there was a trend toward deceased pts being more depressed at baseline (mean CES-D score 23.0~6.8 vs. 18.5~ 12.6, p=0.09) and having a lower
baseline CD4+ cell count (207~196.2 vs. 368~ 303.4 cells/mm3, p=0.09). No
significant associations were found between death and injection drug use history,
adherence level or social support.
Conclusions: Death, from a variety of causes, was significantly associated with
older age, methadone use, higher HIV RNA levels and having an AIDS diagnosis.
The association between current methadone use and death may be related to
issues of access and mistrust, or may simply represent the interaction of multiple
variables including age and injection drug use.
Presenting author: Sharon Mannheimer, Division of Infectious Diseases,
Rm3101A, Harlem Hospital, 506 Lenox Avenue, New York, NY 10037, United
States, Tel.: [phone], Fax: [phone], E-mail: sbm20@columbia.
edu
WePeC6061 A national study of tuberculosis among AIDS
patients in Brazil: presentation, patient
characteristics, and survival
L.F. Jamal, S. Chen2, J.R.R Marins1, D. Barreira 3, N. Hearst4. 1Sao Paulo
State STD/AIDS Program, R. Santa Cruz, 81, Vila Mariana Sao Pauo, CEP
04121-000, Brazil; 2johns Hopkins Bloomberg School of Public Health,
Baltimore, United States; 3Nationa/ STD/AIDS Program, Brasiia, Brazil;
4 nvest of California, San Francisco, San Francisco, United States

Background: Worldwide, tuberculosis is the most common opportunistic infection in persons with AIDS. Brazil is an ideal setting for large-scale studies of the
clinical manifestations of TB in AIDS patients because it has both a fairly sophisticated health system allowing for accurate diagnosis and a large number of AIDS
patients with TB.
Methods: We reviewed the medical records of a representative sample of all
AIDS cases reported in Brazil in 1995 and 1996. During this period, we randomly