Bridging the Gap: Conference Record [Abstract book, International Conference on AIDS (12th: 1998: Geneva, Switzerland)]

12th World AIDS Conference

Abstracts 13305-13309 147

anisms among IDUs may down regulate the expression of CCR5 on the cell
surface independent from the CCR5 genotype. Another explanation could be that
the viral strain present among IDUs is less dependent on CCR5 as a co-receptor.
S13305 Production of /-chemokines in HIV infection: Levels of
MIP-1/ are predictive to the risk of HIV disease
progression
Henrik Ullum', A. Cozzi Lepril, J. Victor2, H. Alladin2, A.N. Phillips2, P. Skinhoj',
B.K. Pedersen. 'Department of Infectious Diseases, Rigshospitalet 2200
Copenhagen; 2Dep't of Primary Care & Pop'n Sci, RF/UCL, London, UK
Objectives: to examine: 1) levels of /-chemokine production in different stages
of HIV infection, 2) the relationship between p-chemokine levels and concentrations of cells capable of producing /p-chemokines, and 3) the prognostic value of
i/-chemokine levels in the prediction of disease progression.
Design and Methods: macrophage inflammatory protein (MIP)-1 a, MIP-1/I
and RANTES production was measured by ELISA in 4.5 hours phytohemagglutinin stimulated whole blood of 90 healthy HIV negative controls and of 245 HIV
infected individuals followed up to 4 and a half years from the date of enrolment.
Results: HIV infected individuals without AIDS showed increased levels of
MIP-la, MIP-1/- and RANTES (P: 0.01) compared to controls. Individuals
with AIDS showed a decreased production of MIP-1/ (P < 0.0001) and similar
levels of MIP-1a and RANTES in comparison to controls. The production of
MIP-lu and MIP-1/ was best correlated to the concentration of CD4+ T cells and
CD16+/CD56+ NK cells. The production of RANTES was best correlated to the
concentration of platelets. A high production of MIP-1/ was associated with a
decreased risk of reaching an AIDS diagnosis or death considered in univariate
analysis (P - 0.01) and adjusted for CD4 counts and age (P = 0.07, P = 0.06,
respectively). MIP-1 a and RANTES production was not significantly associated
with the risk of disease progression.
Conclusion: These findings suggest that the /-chemokine production changes
during HIV infection and that a high /--chemokine production in peripheral blood
lymphocytes may be associated with less rapid disease progression in HIV
infection.
13306 Hepatitis C virus (HCV) genotypes and progression of
HIV infection in hemophiliacs
Angelos Hatzakis', G. Touloumi, A. Karafoulidou2, T. Mandalaki2, J.J. Goedert3.
1Athens University Medical School Goudi 11527 Athens; 2Laikon General
Hospital Athens, Greece; 3 Viral Epidemilogy Branch, NCI Rockville MD, USA
Background: Hemophilia patients infused with non-heated plasma products were
universally infected with HCV. HCV genotype 1 (HCV-1) is associated with poor
response to interferon therapy, and it has been suggested that hemophilia patients
infected with HCV-1 may experience a more rapid progression of HIV disease.
The possible association between HCV-1 and HIV progression was assessed
in HCV/HIV coinfected hemophilic subjects followed up within the Multicenter
Hemophilia Cohort Study (MHCS).
Methods: Two hundred eighty-one hemophilic subjects with known HIV seroconversion dates were HCV genotyped by RT-PCR and reverse hybridization.
Patients from US (n = 152) and Greece (n = 129) were included. AIDS and death
risk were estimated by Kaplan-Meier and proportional hazards models, with left
truncation for time before the HCV-genotyped sample.
Results: The distribution HCV genotypes was 62%, 10%, 11%, 4% and 3% for
HCV-1, HCV-2, HCV-3, HCV-4 and mixed types respectively, while in 11% of the
samples was not possible to determine the HCV genotype. The distribution was
similar in the US and Greek patients. HCV-1 was associated with faster progression to AIDS or death (P = 0.0001 and P = 0.0009, respectively) in both countries.
The relative hazard (95% CI, P) (adjusted forage at HIV seroconversion and country) for progressing to AIDS of HCV-1 infected patients vs others was 2.8 (1.5-5.2,
P = 0.001). The relative hazard for death was found 2.3 (1.4-3.8, P = 0.001).
Conclusion: HCV-1 is associated with faster progression to AIDS and death.
This finding may have important implications for the treatment of HIV/HCV coinfected patients.
13a307 Time to AIDS diagnosis increased in 1996, when use of
highly active antiviral therapy (HAART) became prevalent
Grant Colfax1, E. Vittinghoff', S.D. Holmberg2, S. Buchbinder'. IDepartment of
Public Health, Suite 500, 25 Van Ness Avenue, San Francisco, CA; 2Centers for
Disease Control & Prevention, Atlanta, GA, USA
Objectives: Recent reports suggest dramatic declines in AIDS incidence and
death, which may reflect earlier declines in HIV seroincidence as well as the
impact of HAART. We evaluated AIDS incidence and mortality in 1996 compared
with 1993-5 among men with well-characterized dates of HIV seroconversion
(WC) to differentiate these possible effects.
Methods: Cox proportional hazard analysis with time-dependent covariates for

calendar period at risk was performed on 622 WC men from the San Francisco
City Clinic Cohort. AIDS (1993 CDC case definition) and death were determined
by cross-matches with local and national AIDS surveillance and the National
Death Index. To account for reporting delays, we used multiple imputation of
additional 1996 cases based on 1993-5 reporting delays.
Results: By 12/31/96, 501 (81%) men were diagnosed with AIDS and 450
(72%) had died. AIDS incidence per 100 person years fell from 17.1 in 1993-5 to

1.7 in 1996 and mortality declined from 11.6 to 6.2. ter controlling for time since
seroconversion, the relative hazard of AIDS for 1996 compared to 1993-5 was
0.18 (95% CI 0.04-0.75) and for death was 0.45 (95% CI 0.21-0.95). Estimated
underreporting of AIDS was 14% in 1996; when the number of AIDS cases
was increased by 50% to account for possible underreporting, the RH was 0.25
(95% CI 0.07-0.87). Underreporting for death was estimated to be 2-8%; when
the number of deaths was increased 33% for 1996, the RH was 0.49 (95% CI
0.24-1.03). Prevalence of HAART use could not be accurately estimated because
only 51% of the men were in active follow-up. However, of this group, 43% were
on 2- or 3- drug regimens in 1996, compared with 19% in 1995.
Conclusions: We found a significant decline in 1996 AIDS incidence and
death and an apparent lengthening of the incubation period from seroconversion
to AIDS and death which paralleled increasing HAART use. Lengthening of the
AIDS incubation period could not be fully explained by incidence patterns or
reporting delays. This suggests that clinical trial efficacy of HAART is translating
into measurable benefits at the population level.
13308 Is cell-associated infectious HIV-1 viral load beneficial in
predicting HIV-1 disease progression after knowing
plasma HIV-1 RNA level?
Cynthia Lyles2, H. Farzadegan2 3, T.C. Quinn', J. Astemborski2, J.B. Margolick2,
D. Schaeffer2, D. Vlahov2. 1 School of Medicine, 2School of Public Health, 3Dept.
of Epidemiology - Johns Hopkins University Rm E6003, 615 North Wolfe St.,
Baltimore, MD, USA
Objectives: To determine whether cell-associated infectious HIV viral load predicts progression to AIDS independently of plasma HIV RNA levels among HIV
infected injection drug users (IDU) and homosexual men (HM).
Methods: The study (n = 450) is a combined subset of HIV infected AIDS-free
IDU men (181), IDU women (84), and HM (185). Cell-associated infectious HIV
viral load, determined using quantitative microculture assay (QMC), was measured on fresh plasma collected between 1992-93. Plasma HIV RNA level was
quantified by RT-PCR assay where frozen plasma was available at the same visit.
Subjects were followed from viral load to AIDS through 1995. Survival methods
were used to evaluate the predictive value of infectious viral load on progression.
Results: The median CD4' cell count, infectious viral load level, and HIV RNA
level were: 417 cells/pl, 8.1 infectious units/million PBMC, and 75,975 copies/ml,
respectively. When categorized, infectious viral load (p =.0001) and HIV RNA
level (p =.0001) predicted time to AIDS. After controlling for risk group, CD4' cell
count, and category of HIV RNA level, infectious viral load remained a significant
predictor of AIDS. On the log-scale and standardized for unit variance, both infectious viral load [relative hazard (RH) = 1.52 (1.14, 2.01)] and HIV RNA load [RH
= 2.20 (1.50, 3.20)] independently predicted progression to AIDS. Risk estimates
were fairly similar by risk group (IDU vs. HM). Also, a 10-fold increase in infectious
viral load was associated with nearly a 30% increase in the hazard of AIDS, within
the highest category of HIV RNA load after adjusting for other covariates.
Conclusions: Cell-associated infectious viral load independently predicted HIV
disease progression among a combined cohort of IDU and HM after controlling
for risk group, CD4+ level, and plasma HIV RNA level. The predictive value of
cell-associated infectious viral load appeared to be equal across risk group and
most important among persons within the highest category of plasma HIV RNA
viral load.
13309 Serum albumin as a predictor of survival in HIV infected
(HIV+) women enrolled in the Women's Interagency HIV
Study (WIHS)
Joseph Feldman', K. Anastos2, P. Bacchetti3, D. Burns4, M. Cohen5,
R. Delapenha6, P. Miotti7. SUNY Health Science Center at Brooklyn, Brooklyn
NY; 2Catholic Medical Center, New York, NY; 3University of California, San
Francisco, San Francisco, CA; 4National Institute Child Health Human
Development, Rockville, MD; Cook Country Hospital, Chicago, IL;6 Howard
University, Washington, DC; 7National Inst. Allergy & Infectious Disease,
Bethesda, MD, USA
Objectives: To evaluate serum albumin a as an independent prognostic factor in
HIV+ women.
Methods: Data were analyzed from WIHS, a prospective study of 2058 HIV+
and 567 HIV-women recruited in 5 US cities from 10/94-10/95; clinical, virologic
and immunologic evaluations were performed every 6 months. Notification of
death was obtained from passive surveillance. Associations between baseline
albumin levels and survival were adjusted for age, body mass index (bmi), CD4
count and viral load at baseline using a proportional hazard model.
Results: There was complete data available for 183 deaths in 1493 HIV+
women. After 2 years, loss to follow-up (FU) was 11% in HIV+ women (median
FU was 22.8 months). Cumulative mortality after two years was 13.7% overall but
varied significantly by baseline albumin; it was 41%, 17%, 14%, 10%, 9% and 4%
for women with levels <3.5 (N = 78), 3.5-4.0 (N = 449), 4.0-4.2 (N = 279), 4.2-4.5
(N = 418), 4.5-.7 (N = 159), >4.7 gm/dl (N = 110) (P <.001). The association of

albumin level with mortality persisted after adjustment for CD4 count (3 levels),
viral load (4 levels), bmi (5 levels), and age (7 levels). The adjusted relative
hazard (RH) by albumin level was 5.1 (95% CI 2.1-12.5) in the lowest albumin
group and decreased to 2.5 (95% Cl 1.1-5.8), 2.6 (95% CI 1.1-6.2), 1.5 (95% CI.6-3.7) and 1.5 (95% CI.6-3.9) in the other groups compared to the group with
the highest albumin level (P <.001). For comparison the RH for CD4 < 200 was
6.4 fold and for viral load >500k was 5.9 fold.