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

XIV International AIDS Conference

Abstracts ThPeC7479-ThPeC7482 457

ThPeC7479 Beneficial effect of highly active antiretroviral
therapy (HAART) on the prognosis of AIDS-related
lymphomas (ARL)
A.S. Lascaux, F Hemery2, C. Goujard3, P. Lespritt, M. Bertocchi1,
E. Lepage2, Y Levy1. 'Service dimmunologie clinique- Hdpital Henri Mondor,
Service dImmunologie Clinique, H6pital Henri Mondor, 94010 Crdteil Cedex,
France; 2Ddpartement de biostatistiques - Hdpital Henri Mondor, 94010 Crdteil
Cedex, France; 3Service de Medecine Interne - H6pital Bic6tre, 94275 Kremlin
Bicitre, France
Background: The impact of HAART on the survival of patients (pts) with ARL is
still under debate.
Objectives: To evaluate the influence of HAART on the survival of pts with ARL.
Methods: Retrospective analysis of 73 HIV-1-infected pts with proven ARL diagnosed in two clinical centers between 1992-2000. Event-free survival (EFS) and
survival were estimated by Kaplan-Meir method and a Cox model was used to
evaluate the effect of different variables on survival.
Results: At diagnosis of ARL, the median age was 40.8 years and 22 patients
(30.1%) had prior AIDS-defining events. The mean CD4 cell count was 153/mm3.
56 ARL were classified as diffuse large cell and 17 as Burkitt or Burkitt-like lymphomas. An Ann Arbor stage 3-4 was noted in 60 pts (82%). A bone marrow
or meningeal involvement was present in 13 (17%) and 12 (16%) pts, respectively. Pts were treated either with CHOP (CD4 cell counts<100/mm3; n=36 pts)
or ACVBP (CD4 cell counts>100/mm3; n=28). The median survival (MS) of the
whole cohort of pts was 8 months (mo). Two groups were identified: group 1
(n=35) received HAART either before, at the diagnosis or following ARL and group
2 (n=38) had never received HAART There has been no significant differences
in stage at presentation, presence of B symptoms, marrow or meningeal involvement, CD4 cell count at diagnosis, prior AIDS-events or chemotherapy regimens
between the 2 groups. However, the estimated MS was significantly higher (11.8
mo) in group 1 (HAART) compared to group 2 (no HAART) (6.1 mo) (p= 0.03).
Using the Cox model it was shown that HAART has an independent significant
effect on EFS (p=0.04). No influence on the outcome was found for other variables including: age, prior AIDS-events, CD4 cell count, An Arbor stage, systemic
B symptoms, extranodal involvement and chemotherapy regimen.
Conclusion: Immune recovery under HAART has a beneficial effect on the survival and EFS in pts with ARL.
Presenting author: Anne-Sophie Lascaux, Service d'Immunologie Clinique,
Hbpital Henri Mondor, 94010 Crbteil Cedex, France, Tel.: [phone],
Fax: [phone], E-mail: [email protected]
ThPeC7480 High risk of neoplasia among HIV-infected
homo/bisexuals in Scotland
G.M. Allardice', D.H. Brewster2, D.J. Hole3, D.J. Goldberg3, J. Boyd2,
G. Codere4, L. Shaw4. 'Dept of Statistics and Modelling Science, University of
Strathclyde, 26 Netherblane, Blanefield, Glasgow, G63 9JW United Kingdom;
2Information and Statistics Division, NHSScotland, Edinburgh, United Kingdom;
3Dept of Public Health, University of Glasgow, Glasgow, United Kingdom;
4Scottish Centre for Infection and Environmental Health, Glasgow, United
Kingdom
Background: The Scottish HIV register contains data on all people who have
tested HIV antibody positive or been diagnosed with AIDS in Scotland since 1985.
The Scottish Cancer Registry contains details of all neoplasms (ICD-9 codes 140 -239) except benign neoplasms (ICD-9 codes 210-229) diagnosed in the Scottish
population since 1959. These two computerised registers were utilised to examine the incidence of neoplasia in HIV-infected people (by HIV risk group).
Methods: Probability linkage was used to link the two registers based on agreement between the common identifiers of date of birth, sex, initials and soundex
code of surname. All neoplasms were counted provided they occurred after HIV
diagnosis (or within two months of it). The observed number of neoplasms were
compared with those expected based on sex-, age- and period-specific incidence
rates for Scotland. The standardised incidence ratio (SIR) was defined as the ratio
of the observed to the expected number of neoplasms and the 95% Cl estimated
assuming that the observed number followed a Poisson distribution.
Results:

ThPeC7481 I Declared causes of death amongst reported AIDS
cases in Rio de Janeiro State, Brasil 1991-1995
M.A. Sole-Pl 1, J.G. Valente2, K.R.V. Lemos3. 'Health Department of Rio de
Janeiro State, Rio de Janeiro - RJ, Brazil; 2University of Rio de Janeiro State,
Rio de Janeiro, Brazil; 3Health Department of Rio de Janeiro State, Rio de
Janeiro, Brazil
Background: The "Mortality Information System" is an important source of information for epidemiological monitoring of Aids. However, still little it is known about
the registering of causes of death in patients with the illness.
Methods: In this study we performed a comparison of two independent
databases (the "AIDS information system" (SINAN) and the "Mortality Information System" (SIM)), from the State Government of Rio de Janeiro. Stated causes
of death were analysed for all adult AIDS cases reported between 1991 and 1995
in Rio de Janeiro.
Results: Of the 9,882 reported cases of AIDS, 4,260 deaths had been recorded.
Of these, 3,575 (83.9%) had "AIDS" as the declared cause of death; 395 (9.3%)
stated common AIDS-related illnesses; 204 (4.8%) stated causes not associated
with AIDS and 86 (2%) Ill-Defined Conditions. The analysis of the 685 death
certificates that had declared a cause of death other than Aids, showed that
Infections and Parasitic Diseases had answered for 211 deaths (30.8%); Diseases of the Respiratory System 164 (23.9%); Symptoms, Signs, and Ill-Defined
Conditions 86 (12.5%); Neoplasm's 47 (6.9%); Diseases and Immunity Disorders 41 (6.0%); Diseases of the Circulatory System 37 (5.4%); Diseases of the
Nervous System 28 (4.1%); External Causes of Injury 28 (4.1%); other causes
6.3%. Pneumonia and bronchopneumonia were the principal causes, representing 13.6% of these deaths. Tuberculosis represented 13.7% of the cases, 10.2%
of which were pulmonary. Non-specified pulmonary illnesses stood at 4.2%.
Conclusions: The most commonly reported causes of death (beside Aids) were
respiratory ailments. In the period studied, during which the AIDS cocktail was not
available, 5% of the causes of death amongst patients diagnosed with AIDS were
not linked to this variable. With the introduction of new drugs, one would expect
that along with a reduction in mortality rates, a proportional increase in non-AIDS
related deaths would be seen.
Presenting author: Maria Asuncion Sol-PIA, Rua Almirante Alexandrino 2603
casa 28 - Santa Teresa - Rio de Janeiro - RJ - Brazil - CEP: 20241-261, Brazil,
Tel.: [phone], Fax: [phone], E-mail: [email protected].
br
ThPeC7482 Increased incidence of cancer among participants
in the women's interagency HIV study
N.A. Hessoll, S. Preston-Martin2, E.C. Seaberg3, S. Massad4, S. Melnick5,
H.S. Sacks6, S. Silver7, O. Abulafia8, A.M. Levine2. 1University of California,
San Francisco, University of Californa, San Francisco, 405 Irving Street, 2nd
Floor, San Francisco, CA 94122, United States; 2University of Southern
California, Los Angeles, CA, United States; 3Johns Hopkins School of Public
Health, Baltimore, MD, United States; 4 Cook County Hospital, Chicago, IL,
United States; 5National Cancer Institute, Bethesda, MD, United States; 6Mount
Sinai School of Medicine, New York, NY, United States; 7George Washington
University Medical Center, Washington, DC, United States; 8State University of
New York, Brooklyn, NY, United States
Background: The HIV epidemic has been associated with an increased incidence of specific types of cancers. However less is known about cancers in
women with HIV infection than men.
Methods: To determine if there is an increased incidence of cancer among HIVinfected and at-risk HIV-uninfected women, we compared cancer incidence data
from the Women's Interagency HIV Study to data from the United States SEER
registry data. Age and race-adjusted standardized incidence ratios (SIR) were
computed and exact statistical tests were used to measure significance. To evaluate the effect of highly active antiretroviral therapy (HAART), cancer incidence
during the pre-HAART and HAART eras were compared.
Results: Among the 1950 women participants (1554 HIV-infected, 391 HIVuninfected, and 5 HIV seroconverters) who were followed a median of 5.04 years,
45 cancers were diagnosed during study follow-up. Among HIV-infected women,
significantly (p<0.05) increased incidence rates were observed for Kaposi's sarcoma (SIR=275.8), non-Hodgkin's lymphoma (SIR=27.4), invasive cervical cancer (SIR=7.0), and lung cancer (SIR=10.2). Lung cancer incidence was also significantly elevated among the HIV-uninfected women (SIR=16.0). No excess was
observed for breast cancer in either the HIV-infected or uninfected women. More
recent calendar period (1997 and later) was significantly associated with a decline
in cancer incidence when compared to earlier calendar period (prior to 1997):
RR=0.40, 95% C1=0.21-0.75.
Conclusions: HIV-infected women had increased incidence rates for all three
AIDS-defining cancers and both HIV-infected and uninfected women had increased incidence rates of lung cancer, when compared to expected rates. Adjusted cancer incidence rates decreased significantly since the introduction of
HAART, but the rates remained significantly higher than the rates in the general

population.
Presenting author: Nancy Hessol, University of Californa, San Francisco, 405
Irving Street, 2nd Floor, San Francisco, CA 94122, United States, Tel.: [phone], Fax: [phone], E-mail: [email protected]

Risk Group
Homo/bisexual
Heterosexual
Injecting Drug Use
Haemophiliac

HIV +ve
Persons
822
429
1122
85

Person Years
at Risk
3708
1867
8912
761

Observed
Neoplasms
102
14
35
5

SIR
21.4
5.9
5.5
6.1

95% Ci
SIR
(17.5,25.9)
(3.2,9.9)
(3.8, 7.6)
(2.0,14.2)

Conclusions: There is a 5-6 fold increase in the risk of neoplasia among the heterosexual, injecting drug user and haemophiliac risk groups of the HIV-infected
population of Scotland compared with the general population. However this rises
to a 21-fold increase among the homo/bisexual risk group. Further analyses are
underway to determine how much of this 21-fold increase is due to Kaposi's
sarcoma, a well documented AIDS indicator disease among this homo/bisexual
group.
Presenting author: Gwen Allardice, 26 Netherblane, Blanefield, Glasgow, G63
9JW, United Kingdom, Tel.: [phone], E-mail: [email protected]