Final Program and Oral Abstracts [International Conference on AIDS (8th: 1992: Amsterdam, Netherlands)]

ThC 1550-ThC 1555

TRACK C: EPIDEMIOLOGY

ThC 1550 HIGH RISK OF HIV SEROCONVERSION AMONG STREET INTRAVENOUS DRUG
USERS IN MONTREAL. Soto.Julio *; Lamothe,F. *; Bruneau,J. *; Vincelette,J.*;
Brabant,M.*; Fauvel,M.** *Saint-Luc Hospital, **L.S.P.O. Quebec. Canada.
Obiectives: To evaluate the incidence of HIV seroconversion in a cohort of intravenous drug users
(IVDU) and to determine factors associated with the risk of HIV infection.
Methods: We considered eligible for inclusion any adult using intravenous drugs in the last 6 months.
Subjects were referred by several treatmentcenters, before starting therapy, or recruited from the
street via a word-of-mouth system. Most of the participants were cocaine users (70%). Information
on demographics, drug use, medical history, sexual practices, and HIV serology status was recorded.
The length of time being HIV infection-free was calculated for each subject who had an initial
negative HIV serological test and who was subsequently retested at least once. Univariate actuarial
analysis of seroconversion to HIV positivity was performed using Kaplan-Meier life table estimators.
Results: Between the period of September 1988 to December 1991, we enrolled 409 subjects
among which 156 were street IVDU. The mean time between the entry in the study and the first
positive HIV serological result was 1049 days (standard error=22). The overall cumulative HIV
incidence rate was 8.1 %. The seroconversion rates at 12, 24, and 36 months were 7.0, 12.7, and
20.2% respectively. By univariate actuarial analysis, street IVDU experienced much higher
seroconversion rates when compared to those undergoing detoxification treatment (2-yr rates: 22%
vs 9%; 3-yr rates: 35% vs 16%, respectively; p=0.00558 Mantel-Cox test). There were no
significant differences on seroconversion rates when other relevant risk factors were compared.
Conclusion: Our data suggest that IVDU subjects who remain outside of the health care system
represent an important source for HIV transmission in the community. Supportive social programs
must be developed to promote HIV risk reduction in this population.
Julio C. Soto MD,PhD DSC Saint-Luc Hospital, 1001 Saint-Denis, Montreal,
(Quebec) H2X 3H9 CANADA. Tel [phone] Fax [phone]
ThC 1552 INCIDENCE OF HIV INFECTION IN INTRAVENOUS DRUG
USERS FROM MILAN AND NORTHERN ITALY, 1987-91.
Ciaffi Laura'S, Nicolosi A*, Correa Leite ML', Musicco M', Molinari S*,
Lazzarin A~ (NISDA Study). *National Research Council; ~University of Milan.
Objective: To monitor time trends of HIV infection's incidence and risk factors among
intravenous drug users (IVDU).and study the dynamics of seroconversion rates in relation to
behaviors, prevalence of the infection and other variables infuencing the rate of parenteral and
sexual HIV transmission.
Methods: All IVDU attending 27 drug dependence treatment centers in Milan and Northern
Italy who gave their consent (72%) were screened for HIV antibodies: the seronegative ones
were recruited, interviewed about their habits, counseled about risk behaviors and invited to
periodical visits for interviews, tests for anti-HIV antibodies and further counseling. Incidence
rates, seroconversion rates and cumulative incidence were estimated.
Results: Between 1987 and 1991, 1761 seronegative IVDU were enrolled, 1063 were
followed and 62 seroconversions were observed, for an incidence rate of 4.0/100 personyears (PY) over the entire period (the average annual seroconversion rate was 3.9%). The
seroconversion rate was 6.1% in 1987, 4.0% in 1988, 3.0% in 1989, 3.9% in 1990, and 3.6%
in 1991. The lowest incidence rates were observed in large cities (Milan, 3.0/100 PY, Turin,
2.5/100); higher rates were observed in smaller cities and suburban areas with the highest
incidence (5.1/100 PY), in low-income areas of the province of Milan, and in the Varese-area
(4.89/100 PY) where prevalence of HIV was high. Incidence rates (per 100 PY) were 4.7 for
females, 3.8 for males, 5.2 in <25 years-olds (5.2) and 2.7 in subjects 25 and older.
Conclusions: The incidence of HIV infection has decreased in this study by almost one half
between 1987 and 1991. However, the socio-demographic and geographical variations in
incidence rates raise issues which require additional preventive intervention: the subgroup of
young IVDU with a short duration of addiction; the less socially and economically developed
areas where the frequency of risk behaviors and the prevalence of HIV are high.
Laura Ciaffi, Department of Epidemiology and Medical Informatics, ITBA
National Research Council of Italy, via AM Amp6re 56, 20131 Milano
tel [phone]; fax [phone]
Th' 1554 RAPID SPREAD OF HIV AMONG INJECTING
h 1554 DRUG USERS OF NORTH-EASTERN INDIA:A NEW RISK GROUP OF TRANSMISSION
IN INDIA. Sarkar, S.K.*; Sarkar, K*; Ponda, St; Naik, T.Nt; Singh,
B.C.**; Tripathy, S.P.***; Pal S.C.*. *NICED, Calcutta; **RMC,
Imphal, Manlpur; ***ICMR, New Delhi, INDIA.
Objectives: To study the high risk behaviours of the IDUs and monitor the HIV prevalence over last 5 yrs and find out the prevalence of the IDUs in the community.
Methods: North-eastern States of India border Myanmar and consist of <1% of the
total population in the country. IVDUs of these States have been monitored for HIV
since 1986 through outreach camps and deaddiction centres. Voluntary confidential
testing has been performed using ELISA and confirmatory WB. Anonymous interview
have been done for recording needle sharing and sex behaviours. Community based
survey have been carried out to estimate the prevalence of IDUs by secondary
informant technique and snowballing. Participation of the IDUs in the study were
satisfactory.
Results: No HIV could be detected among 2322 samples screened between 1986 to Oct.
1989. From Oct. '89, quarterly prevalence of HIV among IDUs had increased from 2.9%
to 56.1% varying between 52.4% to 62.3%. During this period 5.9% of Female contacts
of the H +ve IDUs have seroconverted. It is estimated that 1-2% of the general
population in three States surveyed are IDUs which amounts to nearly 15,000 addicts,in a single state. Most of these addicts are male, unmarried and use of condom -is
rare (2-3%). It is further found that a high percentage of the IDUs (>90%) are
knowledgeable about the possible risk of transmission of AIDS through sharing of
needle. Still sharing of the needles is very common. Bleach is an uncommon household
material and not easily available in the market.
Conclusion: North-eastern States of our country today contribute for nearly 25% of
the seropositives detected in India. Such high seropositivity can possibly be due
to the higher potential of injecting route of transmission and very high prevalence
of IDUs. There is definite need for targeted intervention among the IDUs which
should aim at developing specific skills for cleaning of needle and stop sharing.
Sarkar, Swarup K., National Institute of Cholera and Enteric Diseases, P-33 CIT Road Scheme XM, Calcutta - 700 010, INDIA.
36-1176 & 35-0448; Tlx- 021-2436

ThC 1551 'FRONTLOADING' IS ASSOCIATED WITH HIV INFECTION AMONG DRUG
INJECTORS IN NEW YORK CITY
Jose. Benny'; Friedman, SR*; Neaigus, A'; Curtis, R*; Des Jarlais, DC"
*Narcotic and Drug Research, Inc., NY, NY, USA, "Beth Israel Medical Center, New
York, NY, USA.
Obiective: To assess the extent to which lrontloading' is associated with exposure to HIV among injecting
drug users (IDUs). 'Frontloading' is a drug sharing practice in which an injector's syringe is used to measure
drugs and to give a lair share' to another injector by 'spouting' the drug into another's syringe.
Methods: Since August 1991, 214 street recruited IDUs were interviewed about their drug and sex
behaviors during the previous 30 days and previous 2 years; 199 were tested for the HIV antibody (48% +).
(Sample: Blacks 23%, Latinos 38%, Whites 36%, others 3%; males 71%; median age 34). Subjects were
asked about the frequency with which they shared drugs via frontloading' and shared used syringes. All
reported relationships are significant at p<.01 (chi-square test, multiple logistic regression).
Results: (1) Frontloading is widespread: During the prior 2 years, 39% (80/207) of subjects frontloaded, as
did 30% (63/210) during the past 30 days. Of those who report never sharing syringes in the prior 2 years,
22% (24/110) frontloaded, as did 18% (23/129) of those who say they did not share in the last 30 days.
(2) Frontloading is associated with HIV infection: HIV seroprevalence is higher among IDUs who frontloaded
during the past 2 years than among those who did not (71% vs. 36%). This relationship holds true for
subsets of IDUs who had: a) injected <7 years (83% vs. 19%), b) injected >7 years (69% vs. 41%), c) ever
shared syringes in the prior 2 years (66% vs. 41%), and d) never shared syringes in the prior 2 years (84% vs.
33%). In multivariate stepwise logistic regression with other correlates of HIV seropositivity (race/ethncity,
years of injection, injection frequency, shooting gallery use, needle sharing, number of sex partners,
frequency of unprotected sex), engaging in frontloading in the last 2 years remained significant as an
independent predictor of HIV seropositivity (odds ratio 3.74).
Conclusions: Grund et al., in 1990, first reported the practice of frontloading to be widespread among IDUs in
Rotterdam. Among a sample of IDUs in Baltimore, in 1991, Samuels et al. found no significant association of
this practice with HIV infection. Our results from New York City suggest that frontloading can be an important
route of viral transmission which may have been overlooked by most researchers as well as by IDUs who
otherwise avoid syringe sharing and/or other risk practices. Prevention campaigns should include
information on the risks of frontloading, and research instruments should attempt to measure this behavior.
Benny Jose, Narcotic and Drug Research, Inc., 11 Beach Street, New York,
NY 10013, USA. Telephone: [phone], Fax: [phone]
ThC 1553 HIV-1 INFECTION AMONG FEMALE INJECTION DRUG USERS (IDU) IN THE SAN
FRANCISCO BAY AREA, CALIFORNIA, 1989-1991: Increased seroprevalence rates for
IDU who are lesbian/bisexual, racial/ethnic minorities or cocaine injectors.
Reardon Juan, Wilson MJ, Lemp GF, Gaudino JA, Snyder D, Elcock M, NEuven S.
California HIV Seroprevalence Surveys Coordinating Group, San Francisco Bay Area, CA, USA.
Obiectives: To determine HIV-1 seroprevalence rates and demographic and additional factors associated with HIV-1
infection among women with a history of injection drug use entering treatment in the San Francisco Bay Area.
Methods: During the period October 1, 1989 - September 30, 1991 we conducted HIV-1 unlinked serosurveys in drug
treatment clinics (DTC) in five counties of the San Francisco Bay Area. Standard research protocols developed by CDC
were used. After excluding clients retesting in the same calendar year 2,384 entering female IDU were included. The
association of HIV-1 infection with different characteristics was explored univariately and in a multivariate analysis
(stepwise multiple logistic regression). Adjusted odds ratios (OR) and 95% confidence intervals (CI) were derived.
Results: The female IDU population included 37.5% racial/ethnic minorities (n=883), 4.7% lesbian/bisexual women
(n=99), 2.4% preferentially admitted to treatment because of pregnancy (n=56), 22.5% who reported having injected
cocaine in addition to other drugs (n=537) and 153% who reported having used cocaine by non injection routes.
A total of 170 (7.1%) clients were HIV positive. Table #1Characteristic Crude OR Adjusted OR
The rates were stable over time (1989=7.6%, Under 35 1 1
1990=7.0%, 1991=7.2%). Crude and adjusted 35 & over 1.9 (13-2.7) 1.5 (1.0-2.2)
OR and 95% CI are presented in table #1. White 1 1
Conclusions: These data indicate that among Other races 5.0 (3.4-73) 63 (4.2-9.2)
women who injected drugs, HIV-1 infection Heterosexual 1 1
disproportionally infected racial and ethnic Lesbian/bisexual 1.6 (0.7-3.1) 2.0 (1.1-4.2)
minorities, lesbian/bisexual women and those Did not inject cocaine 1 1
who injected cocaine. Older women were also Injected cocaine 2.0 (1.4-2.9) 15 (1.1-2.2)
more likely to have been infected. Methadone Methadone detoxification 1 1
maintenance programs actively recruit HIV-1 Methadone maintenance 1.6 (1.1-23) 1.5 (1.0-2.2)
infected IDU. The specific needs of these Not pregnant 1
subgroups of women IDU must be addressed Pregnant 13 (0.4-35) NS
further and HIV-1 education messages must be Not oral cocaine 1
tailored and delivered accordingly. Oral Cocaine 1.6 (1.1-2.4) NS
Seen N_,yen * San Francisco Department of Public Health
25 Van Ness Avenue. Suite 500 San Francisco, CA, USA, 94102
Phone: [phone] FAX: [phone]
ThC 1555 MORTALITY PREDICTORS IN A 7 YEAR
ThC 1555 IDA COHORT STUDY ROLES OF HIV, HTLV-2, AGE, GENDER AND ENTRY
SYMPTOMS. Weiss. Stanley H.*; French, John**; Klein,
Carol W.*; Mayur, R.K.*; Altman, Ronald**. *UMD-NJ School of
Medicine, Newark, New Jersey, USA. **NJ Dept of Health, Trenton
Objectives: Examine predictors of mortality in an ongoing study of injection drug
users (IDU), particularly among those HIV seronegative (SX) on enrollment.
Methodsr We administered an extensive questionnaire, and obtained blood for HIV
and HTLV testing and storage, on 963 IDU from New Jersey in 1984. SN have been
periodically retested for HIV voluntarily by the study and by other sources. Deaths
were ascertained through clinics as well as state death and AIDS registries. For
final categorization as dead, we required death certificate confirmation. We
reviewed medical and autopsy records, and classified the cause of death.
Results: Prevalent HIV infection and HIV seroconversion were each highly
significant predictors of death; 3 sub-groups were used for further analyses:
1984 HIV+ (SPI HIV Seroconverter (SCO HIV- (SL )
No. 339 43 581
No. (%) Dead 111 (32.7%) 8 (18.6%) 38 (6.5%)
No. (%) AIDS 105 (31.0%) 5 (11.6%) -
Older age was significantly associated with death in SP, SC, and SN. Age was also
significantly associated with progression to AIDS. Only 2 possible cases of nonHodgkin's lymphoma (NHL) have been documented in the cohort.
Among SN, other significant study entry univariate predictors of death were HTLY-2
infection, cough, fatigue, sporadic (<lx/month) IV drug use, black race S cigarette
smoking pack-years. Using logistic regression (LR) analysis, age & HTLV-2 were each
significantly associated with death overall, and with the 20 deaths due to medical
causes. Sporadic IV drug use & cough were each predictors in both the overall LR
models & with death due to non-medical causes (overdose (9), homicide/trauma (4)].
Neither gender, race nor specific drug use were significant in LR models.
Conclusions: The risk of death or of AIDS was similar in men & women. Older age is
a consistent and highly significant predictor of HIV disease progression. Sporadic
IV drug use carries significant risk of lethal overdose. Since HTLV-2 infection is
associated with lethal medical complications, the specificity of the association and
possible pathogenetic mechanisms merit further exploration. Our infrequency of NHL
raises the possibility that the risk of NHL may differ among various groups.
Weiss, Stanley H. Division of Infectious Diseases
Epidemiology, UMD-New Jersey Medical School, 30 Bergen St.,
ADMC 1614, Newark NJ 07107 U.S.A.
Telephone: [phone], FAX [phone]

Th76