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

12th World AIDS Conference

Abstracts 13381-13385 163

S13381 Incidence and characteristics of HIV-related conditions in
distinct South Africa patient populations
Andrew Kanter' 2, D.C. Spencer3, M.H. Steinberg4. 12657N. Mildred Ave. #2
Chicago, IL 60614-2318, USA; 2Wolfson College, Cambridge University, UK;
3University of Witwatersrand, Johannesburg; 4HIV Management Services,
Johannesburg, South Africa
Background: The HIV epidemic in South Africa is of special interest in that it
represents two separate epidemics occurring side by side: a pattern I epidemic
involving primarily white, homosexual or bisexual men, and a pattern II epidemic
involving primarily black, heterosexual people of both sexes. Computerized medical records were used to determine the cumulative incidence of opportunistic
diseases and other HIV-related conditions in different sub-groups. We compared
the incidence of conditions between AIDS and non-AIDS patients. In addition, after stratifying by AIDS status, we made comparisons on the basis of race, gender
and sexual orientation and identified statistical differences in disease incidence.
Methods: Retrospective, longitudinal, computerized observational database
covering comprehensive clinical records of 2179 patients seen between 1985
and 1995 at the Johannesburg General Hospital HIV Clinic.
Results: Thrush, HIV encephalopathy, and CMV retinitis were much more frequent in white AIDS patients than in black AIDS patients (OR 3.94, 4.07, and 5.06,
respectively). Adjusting for race and sexual orientation, tuberculosis, peripheral
neuropathy, and Kaposi's sarcoma appeared more frequently in male compared
with female AIDS patients (OR 1.77, 2.38, and 5.23, respectively). Mean CD4
counts at diagnosis of opportunistic diseases were calculated and ranged from
390 for PID to 195 for Pulmonary TB to <50 for CMV colitis and Cryptosporidium.
Syphilis (serology), hepatitis B (serology), pneumonia and pulmonary TB were
diagnosed at significantly higher CD4 counts in black than white patients with HIV.
Conclusions: 1) There are significant differences in the incidence of several
of HIV-related conditions between the two patient groups. 2) This difference occurred despite identical location of care (clinic) and similar health care providers.
3) Although socioeconomic status may account for some of the differences
seen, differences in host-virus interaction between the two epidemics could be
responsible.
S13382 | Transitions to injecting and seroconversions for HIV,
hepatitis B and hepatitis C among a cohort of
non-injecting heroin users
Alan Neagus, S. Friedman, M. Miller, D. Des Jarlais. NDRI, Two World Trade
Center 16 Th FL, New York, NY 10048, USA
Objectives: Many heroin users do not inject drugs, yet little is known about
their risks of transitioning to injecting and of acquiring HIV and similarly transmitted pathogens. We examine transitions to injecting and seroconversions for
HIV, hepatitis B (HBV) and hepatitis C (HCV) among non-injecting heroin users
(NIUs).
Methods: In an ongoing prospective cohort study, NIUs were recruited in out-of-treatment settings in New York City in 1996/97. Eligible respondents were 18
years of age or older, had used heroin at least once in the last 30 days, and had
never injected drugs ("never injectors") or had not done so in the last 6 months
("former injectors"). At baseline and follow-up, respondents were interviewed
about their drug and sexual risks, tested for HIV (antibody), HBV (core antibody)
and HCV (antibody; EIA 2.0), and counseled about injecting risks.
Results: 268 respondents (71% never injectors, 29% former injectors) were
tested 2 or more times for either HIV (262), HBV (262) or HCV (259). The mean
follow-up time was 8.5 months (SD = 4.1). 15 never injectors transitioned to injecting (11.3 per 100 person-years at risk (PYAR), 95% CI = 6.3, 17.8), and 9 former
injectors relapsed to injecting (15.9 per 100 PYAR, 95% CI = 7.2, 30.0). None of
the 226 respondents who were HIV seronegative at their first test seroconverted.
There were 3 HCV seroconversions among transitioners (33% of 10 negatives;
37.8 per 100 PYAR, 95% CI = 7.1, 92.6), and 6 among non-transitioners (3.7%
of 162 negatives; 5.4 per 100 PYAR, 95% CI = 2.0, 10.6). HBV seroconversions
were: 1 among transitioners (7.1% of 14 negatives; 8.0 per 100 PYAR, 95% CI
= 0.0, 31.4); and 9 among non-transitioners (6.5% of 138 negatives; 9.5 per 100
PYAR, 95% Cl = 4.3, 16.7).
Conclusion: NIUs may have a low risk of becoming infected with HIV. However,
they are at risk of transitioning to injecting, and are then at extremely high risk of
becoming infected with HCV. NIUs are also at risk of acquiring HBV, which may be
occurring through sexual transmission. The advocacy of non-injecting heroin use
as a harm reduction approach needs to incorporate methods to reduce possible
adverse outcomes, including transitions to injecting and non-parenteral infection
with HBV and, possibly, HCV.
S13383 | Notable impact of the introduction of the 1993 European
AIDS case definition on progression rates to AIDS in
drug users
Maria Prins. For the European Seroconverter Study among injecting drug users;

Municipal Health Service Nieuwe Achtergracht 100-1018 WT Amsterdam, The
Netherlands
Background: Since recurrent pneumonia and pulmonary tuberculosis, 2 of the 3
diseases added to the clinical AIDS definition (Def) in 1993, are more common
in injecting drug users (IDU) than in homosexual men, we studied whether and
how much the introduction of this expanded Def has an accelerating effect on
progression to AIDS and decreases the probability of pre-AIDS death in IDU.

In addition, because tuberculosis is more frequent in southern Europe than in
northern Europe, we investigated whether the impact of the introduction differs
by geographic area.
Methods: The study population comprised 664 IDU with known intervals of HIV
seroconversion from 8 cohorts in Europe followed until August 1995. We studied
progression from seroconversion to AIDS and pre-AIDS death by applying (1) the
1987 AIDS Def and (2) the 1993 Def for the total study period, and (3) the Def as
effective at the date of diagnosis. Progression rates to AIDS and pre-AIDS death
were estimated using a competing risks model. The impact of the introduction of
the expanded Def was quantified by evaluating the effect of calendar period of
follow-up (stratified as the period before and after the introduction of the expanded
Def) as time-dependent covariate in a Cox model.
Results: Applying the 1987 Def for the total study period, at 8 years from
seroconversion 24.1% of the IDU were expected to have developed AIDS and
14.2% to have died without AIDS. Applying the 1993 Def for the total study period, these figures were 36.1% and 8.7% respectively. Applying the Def effective
at the date of diagnosis, progression to AIDS was faster for subjects followed
after the introduction of the expanded Def (adjusted relative hazard (ARH) 1.43,
95% confidence interval (CI) 0.86-2.38) compared with subjects followed before,
whereas applying the 1987 Def for the total study period, progression was similar
in both periods (ARH 1.00, CI 0.58-1.74). For progression to pre-AIDS death, the
ARH for the most recent period decreased from 1.28 (CI 0.64-2.54) to 1.07 (CI
0.51-2.25) applying the 1987 and effective Def respectively. Multivariate models
including geographic area, calendar period of follow-up and the interaction term
between these covariates revealed that the impact did not differ substantially by
geographic area.
Conclusions: Following the introduction of the 1993 European AIDS Def, IDU
were almost 1.5 times more likely to develop AIDS than before. The probability
of pre-AIDS mortality decreased, but the impact was less pronounced. We found
no evidence for geographic differences in the impact of the expansion of the Def.
In interpreting trends in HIV/AIDS surveillance data and in evaluating the effect
of different therapeutic regiments on progression to AIDS the impact of changes
in the AIDS definition for IDU should be taken into account.
S13384 1 Gender differences in the relationship between HIV-1 load
and progression to AIDS
Homa Yoon Farzadegan1, D.R. Hoover2, J. Astemborski', C.M. Lyles',
R. Markham', T.C. Quinn1, D. Vlahov1. 'Johns Hopkins University 615 N Wolfe
St (E6 004) Baltimore USA 21205; 2Mecrch Pharmaceticals Philadelphia PA,
USA
Objectives: HIV load in plasma is used for the initiation of combination antiretroviral treatment based on data obtained from gay men studies. We studied the
epidemiologic association of viral load between genders to assess differences in
cut points on HIV viral load for women and men.
Design: Injection drug users participating in ALIVE Cohort Study were studied for plasma HIV load cross-sectionally at baseline (1988) and at follow up
visits (1992-94) in a prospective/retrospective study of fresh and frozen blood
specimen.
Methods: Plasma HIV RNA, measured by b-DNA and RT PCR assay, were
performed on frozen specimen. Cell associated HIV load. determined using
quantative microculture assay (QMC), was done real time on fresh blood.
Results: Median HIV-1 viral load was significantly lower in women than men as
measured by all 3 assays at both time points (table). In general, median HIV viral
load for women was half of that for men. This association with gender remained
in a linear egression model adjusting for CD4. Race and drug use showed no
effect. Proportional hazards models for AIDS showed that the relative hazard for
women increased substantially from (1.09-1.30) to (1.56-1.64) when HIV-1 viral
load was added to the model.

Gender Baseline sample (n = 527)
CD4+ bDNA
Male 518 8,907
Female 518 3,365

Follow-up sample (n = 285)
CD4+ QMC RT-PCR
390 8 93,130
417 5 45,516

p < 0.05, p - 0.01

Conclusion: Lower HIV-1 load in women but similar pathogenesis, if confirmed
in other studies, is suggestive of the use of different cut points of this marker for
antiretroviral treatment for women than men.
13385] Site specific virus load in the reproductive tract tissues
of women infected with human immunodeficiency virus
Suraiya Rasheed1, Z.S. Sheng1, C.Z. Zorrilla2. University of Southern California,
School of Medicine, 1840 N. Soto St., EDM 103, Los Angeles, CA, USA;
2University of Puerto Rico, San Juan, PR, Mexico
Objectives: The objective of this study was to identifytissue sites in the female
reproductive tracts which harbor and shed high quantities of HIV in the genital
secretions of HIV-1-infected women
Method: Virus load was evaluated in the plasma of 12 HIV-1 seropositive
women and compared with cell-free virus present in the genital secretions collected from three separate sites along the genital tracts of these women. Each of
the four samples i.e., plasma (PL) cervical swabs (CVS), vaginal swabs (VGS) and
cervicovaginal lavage (CVL) were collected weekly for four consecutive weeks
and HIV-1 RNA load was evaluated - 192 sequential samples.