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

12th World AIDS Conference

Abstracts 13334-13337 153

included in the study which was conducted between 1996 and 1998 (154 in-depth
interviews with open probes and 1,037 surveys of twenty questions).
Results: The findings suggest that migrating/traveling youth are at greater
risk than youth who stay at home. Compared with previous findings of at-home
youth, 53% of youth who moved or traveled were more likely to engage in high
risk behaviors in a foreign location. They reported that loneliness, freedom from
familiar strictures, and ignorance of AIDS and foreign customs all contributed to
the increased possibility of HIV transmission. A large percentage stated that they
had received little AIDS information and did not understand the risks posed in
a foreign society. Among the recommendations: HIV/AIDS education should be
improved in their homelands before young people leave; information should be
given to all immigrants as they enter their new homelands (preferably in their
language) but the largest number suggested that peer support would be best -
especially advice and education from previously migrated youth.
Conclusions: The study's findings supported the original hypothesis that migration and travel is fueling the global pandemic among youth and suggested
that governments, schools and international health organizations should make
stronger efforts to educate vulnerable young people who travel or migrate - especially impressionable and naive teenagers. Most importantly, the study suggests
that many risks are avoidable if action is taken now.

13334

New evidence to reconcile in vitro and epidemiologic
data on the role of heroin on CD4+ decline in injecting
drug users

A. Krol', C. Flynn2, D. Vlahov2, F. Miedema,3, R.A. Coutinho', E.J.C. Van
Ameijden'. 'Nieuwe Archtergracht 100, 1018 WT, Municipal Health
Service,3 Central Laboratory Blood Transfusion Service, Amsterdam, The
Netherlands; 2Johns Hopkins University, Baltimore, USA
Background: While laboratory studies observed that opiates stimulated growth
of HIV in vitro, epidemiologic studies failed to find an association of drug use with
HIV progression. In this study, long-term effects of drug type and other drug use
related risk factors on CD4+ decline were studied in HIV-infected injecting drug
users (IDUs) of whom a subset also had CD4+ cell measurements available prior
to seroconversion.
Methods: In total, 2029 CD4+ cell measurements were available of 287 IDUs
with known dates of seroconversion from two large prospective cohort studies:
224 IDUs from Baltimore (ALIVE), USA, and 63 IDUs from Amsterdam, The
Netherlands. CD4+ decline was modeled from three months prior to up to seven
years post-seroconversion with regression analysis for repeated measurements
using a mixed linear model approach which allowed the rate of change of the
square root transformed CD4+ count to differ from one time interval to another.
Results: The piecewise linear model which assumed a first order autoregressive moving average within person correlation structure and allowed the CD4+
decline to change six months after the midpoint of the seroconversion interval,
revealed the highest likelihood. Both cohorts showed a rate of CD4+ decline of 44
cells/mm3/month up to six months after seroconversion and 5 cells/mm3/month
thereafter. Occurrence of self-reported infections and use of mainly heroin in
the seroconversion interval resulted in a sharper CD4+ decline until the first six
months after seroconversion (p = 0.041 and p = 0.004, respectively). Both effects
were independent but relatively short-lived as the CD4+ values converged later on
in HIV infection. A higher frequency of borrowing used injection equipment since
1980 resulted in a higher CD4+ count already present before seroconversion (p
=0.049).
Conclusions: Use of mainly heroin at the time of seroconversion resulted in a
somewhat faster CD4+ decline seen only around the interval of seroconversion.
This finding might reconcile earlier discordant epidemiological and laboratory
study results. To elucidate the mechanism by which a history of high frequencies
of borrowing could favorably affect HIV progression, further research is needed.
189* /13335 Survival following HIV seroconversion-evidence of
a recent change?
Kholoud Porter. On Behalf of the UK Register of HIV Seroconverters Steering
Committee; MRC HIV Clinical Trials Centre, UCLMS, Mortimer Market Centre,
London, UK
Background: Recent evidence from clinical trials of an improvement in survival
with multiple drug regimens is expected to be mirrored in a lengthening of the period from HIV seroconversion to death, depending on treatment uptake rate. The
availability of recent follow up data on HIV infected persons, with well documented
dates of seroconversion, is required to evaluate this.
Objectives: 1) to estimate the distribution of time from HIV seroconversion
to death, and to examine changes in mortality over calendar year at risk (as a
time-dependent covariate), and 2) to estimate the time from seroconversion to
initiating anti-retroviral therapy and to examine any evidence of a change over
calendar time.
Design/Methods: Data from a cohort of HIV seroconverters in the UK with
documented negative tests within 3 years of the first positive test were analysed
using Kaplan-Meier methods and Cox proportional hazards models. For HIV
survival estimates, data were censored on 30 September 1997. For analysis of
time to initiating therapy, the initiation of anti-retroviral treatment was the outcome
variable and persons not on therapy were censored at the date they were last
seen. The analysis allowed for therapy being initiated in 1987 at the earliest.
Results: Of 1287 seroconverters, 261 died (20.3%) by 30/9/97. Median survival
from seroconversion was estimated to be 10.6 years (95% CI = 9.8-11.3). For

persons at risk in 1996 or later, the hazard of death was significantly lower than in
1990-95 but not lower than pre-1990. Compared to pre-1990, relative risks (95%
CI) were; 1.4 (0.8-2.5), 1.3 (0.7-2.4), 0.9 (0.5-1.6) for 1990-94, 1995, and 1996
or later respectively. 480 persons (39%) initiated therapy. Time to initiation was
significantly shorter for more recent seroconverters. Compared to pre-1990 RR
(95% CI) were; 1.0 (0.8-1.3), 3.8 (2.4-5.9), 7.0 (4.1-12.2) for 1990-94, 1995,
and 1996 or later respectively.
Conclusions: We found no significant risk reduction trend across the whole
period. Survival following seroconversion may have lengthened for persons very
recently at risk but is not significantly different from pre-1990. No conclusions can
be made on the data available to date on the timing of therapy and its effect on
survival. Further follow up of known date seroconverters is essential to quantify
changes in survival patterns.
S13336 Work, income and poverty among people with HIV and
AIDS in the Swiss HIV cohort study
Bruno Ledergerber', M. Egger3, M. Rickenbach2, R. Weber'. 'University
Hospital Zurich, Ramitr. 100, Upol33, Ch-8091 Zurich; 2Dept. of Social Med,
Univ of Lausanne, Lausanne, Switzerland; 3Dept. of Social Medicine, Univ. of
Bristol, Bristol, UK
Objective: To examine labour force participation, income and poverty among
HIV-infected patients in Switzerland and to study the influence of income on
disease progression.
Participants and setting: Swiss HIV Cohort Study, a prospective multicentre
study based at HIV units at university hospitals and cantonal hospitals. Data on
employment, workload, and income from salary, welfare and insurance during the
last month were collected between August 1993 and August 1995.
Methods: Logistic regression analyses were performed with endpoints labour
force participation and poverty. Cox's proportional hazards models were used to
examine the influence of income on the risk of a first AIDS defining event and
death among participants in clinical stage A.
Results: Data from 2331 participants (27% women) from six centres were
analysed. 837 participants (36%) were men who have sex with men, 944 (40%)
had a history of intravenous drug use (IDU) and 550 (24%) were infected heterosexually. Employment rates in these groups were 75%, 60% and 40% (p < 0.001).
A history of IDU, clinical stage C and lower educational status were independently
associated with a reduced probability of employment. Among employed people
the average workload was above 90 hours per month (50% of regulatory standard) in all transmission groups and clinical stages. Few homosexual men (5%)
but 22% of heterosexually infected people and 36% of IDUs had an income below
the poverty threshold. Clinical stage and gender were not associated with poverty
in multivariate analysis. Of 792 participants in stage A, 114 (14.3%) developed a
first AIDS defining event and 74 (9.3%) died during an average follow up of 2.8
years. Income and poverty were not associated with disease progression and
survival.
Conclusion: The income of an important minority of HIV-infected patients in
Switzerland is below the poverty threshold. However, HIV infection per se does
not appear to lead to poverty and income and poverty were not associated with
disease progression. These results are important to estimate the indirect costs of
the HIV/AIDS epidemic and to formulate social policy.
113337 Survival and progression to AIDS up to 1996 in 13,000
HIV-1 seroconverters
Sarah C. Darby. On behalf of the collaborative group on AIDS Incubation and
HIV Survival; ICRF Cancer Epidemiology Unit, Radcliffe Infirmary, Oxford OX2
6HE, UK
Objectives: The Collaborative Group on AIDS Incubation and HIV Survival was
formed to estimate the distributions of the AIDS incubation period and of the
survival time following HIV-1 seroconversion, and to investigate the dependence
of these distributions on age at seroconversion, exposure group, calendar year
of seroconversion, and calendar year at risk.
Design: Investigators from 38 studies in 14 countries in Europe, North America,
and Australia known to be conducting studies of HIV-1 infected individuals with
well estimated dates of seroconversion were invited to contribute anonymized
individual data to the collaboration.
Methods: Combined analysis of data from these studies using Poisson regression.
Results: Data on over 13,000 individuals were analyzed. Annual death rates
were more than ten-fold higher in those seroconverting at ages 60-64 than at
ages 1-4 years, after adjusting for time since seroconversion and study. AIDS
incidence rates differed five-fold over the same range. For survival there was
no significant variation between exposure groups, after adjusting for age at seroconversion and time since seroconversion and allowing for variation between
studies, while for AIDS there were small differences between exposure groups.
For those seroconverting during years 1984-94 there was no significant variation
in survival or progression to AIDS by year of seroconversion, after adjusting for

age at seroconversion, time since seroconversion and study. Similarly, there was
no significant variation in survival or progression to AIDS by calendar year at risk
during follow-up period 1985-96, after adjusting for age at seroconversion, time
since seroconversion and study
Conclusions: For HIV-1 infected individuals in Europe, North America and
Australia, age at seroconversion and time since seroconversion are by far the
major determinants of survival and progression to AIDS. Other factors, such as