59.3% United States  United States
8.8% United Kingdom  United Kingdom
4.9% Canada  Canada
4.1% Australia  Australia
3.5% Philippines  Philippines
2.6% Netherlands  Netherlands
2.4% India  India
1.6% Germany  Germany
1% France  France
0.7% Poland  Poland

Today: 57
Yesterday: 370
This Week: 1121
Last Week: 2221
This Month: 5709
Last Month: 6822
Total: 130308
PDF Print Email
User Rating: / 0
Written by Alex Wodak   
Thursday, 23 July 1992 00:00



Alex Wodak summarises a selection of the papers on HIV and drug use presented at the VIIIth International Conference on AIDS/Third STD World Congress held in Amsterdam in July 1992.


An estimated 12 000 people drawn from almost all countries in the world attended the VIIIth International Conference on AIDS in Amsterdam, July 1992. Research on injecting drug users and prisons formed a larger component than in previous conferences but was included in multiple sessions and also spread throughout many poster sessions. This paper represents an attempt to draw the abstracts from these papers together and disseminate the findings rapidly and widely. Substantial national and international resources have been required to fund the research and enable researchers to attend the meeting. It is important that these findings are readily accessible to other researchers and policy makers.


The index of the conference abstract and presentations volumes was used to identify all papers on injecting drug users and prison. Brief summaries were prepared from these published abstracts. Abstracts presented in the Publications volume were not reviewed. The abstracts included in this summary represent approxamitely 90% of available abstracts. Preference was given to abstracts that contained quantitative data. Summarised abstracts were marked with a Sanford Major Accent quick reference marker. Summaries were dictated on a Philips hand-held dictaphone using Philips LPH tapes. Wordprocessing was carried out on a Macintosh Classic Personal Computer. Major overview papers were not summarised unless an abstract was included in the conference volumes.
In some cases, an attempt has been made to standardise nomenclature so that the wording used in these summaries may differ from the wording used in some abstracts.


Epidemiology of HIV and AIDS

Intravenous drug users (IDUs) compromised 68% of AIDS deaths, 72% of AIDS cases, and 82% of known HIV infected-cases in Belgrade 1985-1991. [PoC4636]

Ten percent of 263 Bangkok IDUs seroconverted over 10 months (12.5/100 person-years). Reported 'stopped sharing' because of fear of AIDS, and having a regular sexual partner predicted remaining non-infected with incarceration predicting seroconversion [PoC 4278]

One hundred and seventy-three previously seronegative IDUs in Bangkok were followed over a 10-month period in 1989. The seroconversion rate was 12.5/100 person-years. Se If- reported's topped sharing' injection equipment in response to concern about AIDS and having a regular sexual partner were associated with a lower probability of seroconversion. Incarceration was associated with higher probability of seroconversion. [PoC 42781

A mathematical model for heterosexual transmission from IDUs derived from a survey of 200 Irish IDUs predicts increasing spread to non-drug users. [PoC 47 72]

Mortality among 2560 Italian IDUs was 4.5/100 person-years (HIV infected) and 0.8 (non-HIV infected), i.e. RR = 5.3. The excess death rate among HIV-infected IDUs was due to increased number of overdoses in addition to AIDS. [PoC 42691

The number of HIV-infected IDUs in Bangkok in 1991 was estimated to be 12 000 by the capture-recapture technique. [PoC40751

IDU-associated AIDS in the USA increased from 33.3% in 1988 to 34.6% in 1991 comprising: IDI-Js (7 1.0%), MSM IDUs (15.2%), heterosexutl partners (11. 5%), and 2.3 % born to mothers who were IDUs or sex partners of I DUs. [PoC 46931

Number of deaths of Berlin IDUs increased in 1990and again in 1991 whereas prevalence of HIV infection among persons dying of IDU has been failing for 3 or 4 years. Prevalence of HIV infection among IDUs undergoing out-patient treatment has also been falling. [PoC 46891

Seroprevalence in 400 Brazilian IDUs increased from 33% in 1990 to 38% in 1991 with recruitment from treatment sites, homosexual bisexual male and IDU history of more than 5 years predictive of HIV infection. [PoC 426 51

A low prevalence of HIV but high prevalence of risk behaviour was found in 1198 Denverhomeless men and women. Forty-one per cent had a past history of IDU. Linked seroprevalence was superior to blinded seroprevalence for assessing sexual behaviour. [PoC 43 201

Seroconversion in 1500 IDUs attending drug dependency clinics between 1985 and 1991 was 4.9%/person year (Rome), 3. 1 % (Milan) and 0.0% (Naples). HIV incidence declined in Rome and Milan after 1987 to less than 5%. Findings suggest that where prevalence was still low in the 1980s, further spread of HIV is not inevitable. [PoC45561

The proposed new CDC (Center for Disease Control) definition of AIDS adding HIV- infected persons with CD4 < 200 captures cases of late HIV infection more completely than the 1987 definition and this is especially true for IDUs and heterosexuals. [PoC 447 11

Of 2430 New York city methadone patients, AIDS deaths were higher in males than females (14.4 vs IL9/1000 person-years) but pneumonia deaths were more common among females. Under-reporting of AIDS deaths was more common for women than for men. Only 38% of death certificates recorded IDU status. Cocaine and polydrug use are likely to place more women than men at continuing risk of HIV. [PoC47271

Among 2921 Baltimore IDUs, seroincidence was stable at about 4% for 3V2 years. Higher rates have persisted among women. Seroconversions among non-users have been increasing recently. [PoC 46981

Despite a recent increase in deaths from overdos among Italian IDUs, AIDS is now an even more fre, quent cause of death. [PoC 45 531

Follow-up of 306 of 603 HIV-infected patients for whom the seroconversion date was known showed that survival to a variety of end-points is not adversely affected by the sex of the IDU or by extent of drug use. [MoC 00661

Genetic and serological subtyping of 52 infected persons tested in Thailand revealed that 91% (20/22) of persons presumably infected through sexual transmission had HIV- I subtype A whereas 75% (21/28) of injecting drug users had subtype B. Subtype Awas more predominant in the northern region of Thailand. The two subtypes independently entered Thailand in the late 1980s with an estimated 300 000 Thais infected with HIV by 199 1. [WeC 10251

One hundred and twenty-five female prostitutes, 207 IDUs and 269 of their respective partners from Colorado have been followed since 1988. Although only two seroconvers ions have occurred among respondents since 1987, no person in the connected component is more than 12 steps away from an HIV-infected person (mean: 6.2 steps). Despite the highly connected networks, little transmission of HIV has been observed. [ThC 15191

One hundred and fifty-six street IDUs in Montreal showed aseroconversion rate of 22% at2 yearsand35% at 3 years which was much higher for those undergoing detoxification treatment. IDUs who remain outside the healthcare system represent an important source for HIV transmission inthecommunity. [ThC 15501

One thousand and sixty-three seronegative IDUs in Milan recruited between 1987 and 1991 were followed duringwhichan incidence rateof 4.0/100person-years over the entire period was observed. The annual rate declined from 6.1% (1987) steadily to 3.6% (1991). Those most at risk were young IDUs with a short duration of addiction, and IDUs from less socially and economically developed areas where the frequency of risk behaviours and prevalence of HIV are high. [ThC 15521

One hundred and seventy of 2384 (7.1%) female IDUs entering treatment in San Francisco were infected with HIV with the following factors predictive of infection: racialand ethnic minority; lesbian/bisexual ~' men; cocaine injectors and older women. [ThC 15 531

From October 1989, quarterly prevalence of HIV among IDUs in north-eastern India increased from 2.9% to 56. 1%. It is estimated that 1-2% of the general population in the three states surveyed are IDUs amounting to nearly 15 000 addicts in a single state. [ThC 15541

Of 963 IDUs from New jersey followed since 1984,3,39 were infected with HIV in 1984 of whom 32.7% have died and an additional 3 1 % have AIDS. Forty-three seroconverted, of whom 18.6% have died and 11.6% have AIDS. Of the remaining 581 seronegative IDUs, 6.5% have died. Age and HTLV-11 were each significantly associated with death and with the 20 deaths due to medical causes especially. Sporadic IDU and cough were both predictors for deaths due to non-medical causes (such as overdose). The risk of death or of AIDS was similar in men and women. [ThC 15 5 51

One thousand and seventy-six of 2275 injecting drug users in treatment in Brooklyn were terminated or dropped out of treatment at a rate of almost 40/1000 patient-months. Rates were lower among patients involved inresearch programmes and lowestamong32 patients enrolled in drug protocols. Patients participating in research may drop out of drug treatment less than others but this could be because more motivated patients tends to be recruited in the first place. [PoD 38781.

Continued injection of drugs independent of other factors was associated with increased progression to AIDS among 3200 Spanish IDUs. [PoC45611

From 38 countries that provided data (representing 60% of the global population), it was estimated that there are 3.2 million IDUs representing 101/100 000 population. Extrapolating to the total world population, the global number of IDUs could range as high as 5.4 million. [PoC 429 11

Five hundred IDUs were interviewed every year for 2 years in Glasgow. Prevalence of HIV infection remains at 2% because of the relatively late date of entry of HIV into this population, relatively low levels of needle sharing, strong tendency to share injection equipment in confined social circles and turnover of the IDUs population due to abstinence, migration and death. [PoC 41091

Seroprevalence and seroincidence of HIV declined progressively in a sample of 3800 IDUs not currently in treatment contacted by an outreach programme in New Jersey. These changes were attributed to the recruitment of individuals at gradually declining risk of HIV infection and lower cohort seroprevalence. Seroincidence declined from 17%/person-year to 1%/person-year. [PoC42551

Of 230 drug users in Arnhem, the Netherlands, 117had injected drugs. For the 135 participants in methadone maintenance, HIV seroprevalence was estimated as 1.3%. In spite of high levels of risk behaviour, the prevalence of HIV infection was low probably because few drug users come into contact with Amsterdam. Risk factors differed markedly between participants in methadone maintenance and others interviewed on the street. [PoC 42661

Among 2397 IDUs recruited between 1980 and 1987 in Milan and observed over 16 323 person-years, 423 subjects died. Deaths from overdose exceeded deaths fromAIDS inthefirst 4yearsofthestudy. 1n5 ofthelast 6 years of the study, deaths from AIDS exceeded deaths from overdose. Death from AIDS increased from 5.07/1000 person-years to 27.60 with deaths from overdose increasing from 6.34 to 16.98. IDUs known to be HIV infected had a considerably higherdeath rate from AIDS and infectious diseases but similar rates of death from overdose, violent death, or other causes. [PoC 45531

The proportion of HIV-infected IDUs among 225 drug users in Ruili County, China increased from 1-3.5% in 1988 to 30.5% in 1990. Fifty-one of 64 IDUs screened for HIV tested positive. Sharing injection equipment and failing to sterilise injection equipment were independent risk factors. [PoC 42 701

Of 2196 Turin IDUs recruited between 1986 and 1991 26.3% were infected with HIV-1. Seropositive IDUs were older and had a longer duration of drug use. Prevalence of infection declined from 25.8% in 1986, to 2 5 % in 1989 and 3.7% in 199 1. Only 2. 1 % of IDUs with a duration of drug use less than 3 years were HIV infected. However, five former IDUs had apparently become infected with HIV from heterosexual contact. [PoD 53421

The initial epidemic of HIV infection i7t Asia occurred in Thailand in 1988 among IDUs with a monthly seroincidence rate of 3-5% and median provincial seroprevalence reaching 35% by December 1991. A second wave of infection occurred among female prostitutes in 1989 with a median provincial seroprevalence of 15% in 1991. A third wave in male clients of prostitutes began in late 1989 with seroprevalence in male STD clinic attendees, reaching a median of 5% by 1991 and a mean of 2.9% among conscripts by 1991. A fourth wave appeared in 1991 in pregnant women (median 0.7%). Prevalence of HIV among IDUs in Manipur state bordering Myanmar reached 76% by 199 1. Prevalence of HIV infection in Myanmar in 1991 was 76% among IDUs, 3-8% among female prostitutes and 1-4% among soldiers. In south-west China, 90% of 429 HIV-infected individuals were IDUs. A pattern of rapid spread of HIV infection among IDUs followed by female prostitutes, male clients of prostitutes and their wives has been identified in Thailand and possibly in other neighbouring Asian countries. This pattern may be explained by a high degree of needle sharing among IDUs, a high frequency of male patronage of prostitutes and absence of pre-marital or extra-marital sex in most women. [PoC 4087].

Of the 45 506 total AIDS cases reported in the USA in 1991, 34.6% were associated with injecting drug use of which 7 1 % were IDUs, 15.2% were homosexual/bisexual male IDUs, 11.5% were heterosexual IDUs and 2.32% were children born to women who were IDUs or the sex partners of an IDU. In 1988, 33.3% of AIDS cases were associated with IDU. The rate in the north east was 8.6 times higher than in the middle west. Rates in Blacks were 11.9 times higher than in Whites and twice the rate among Hispanics. [PoC 4693]

One hundred and twenty-eight female IDUs in New York City in methadone maintenance treatment identified 5022 known relatives who had experienced 57 HIV infections, 63 AIDS deaths and 45 deaths by violent means. In this community, AIDS has already had a far greater impact than even that of the high number of losses experienced to date from death by violent means. [PoD 54721

Of 722 IDUs entering drug treatment in two northeastern US states, 82% reported previous testing for HIV antibody; 35% were tested on entry into treatment. Two-thirds of HIV- positive IDUs reported having tested seropositive before entering treatment. Ten percentof IDUs were likely to be tested for the first time on entry into treatment. I DUs entering drug treatment may decide to be tested for HIV antibody for reasons other than recent sexual or needle risk behaviours. [PoC 4825]

Prevalence of HIV infection amongCopenhagen IDUs has stabilised between 15% and 20% since 1985 probably reflecting: the rather late introduction of HIV into this population and thus relatively high knowledge of risk behaviour and pre.vention; early intervention from Danish health authorities with information campaigns and ready access to sterile injection equipment. [PoC 42561

The 1470 Toronto IDUs participating in three studies were asked to submit dried blood spots and a specimen of saliva or saliva collected with the Omni-sal device. Higher compliance (96%) was achieved with saliva compared to dried blood spots (84M. Only 3% of salivary samples were inadequate for testing. HIV antibody testing on matched saliva and blood specimens agreed in 100% ofcases. Collection of salivary samples reduces the bias inherent in self-referral. [PoC 44961

To31 December 1991, 9254 individuals were reported with HIV infection in Rome, and 250-300 new cases are reported per quarter. IDUs account for 74% of cases with known risk factors. Sexual transmission among non,drug injectors accounts for 19% increasing between 1985 and 1991 from 9% to 23% in males and 2% to 3 1 % in females. Inc idence of H IV infection appears stable with the proportion of new diagnoses attributable to sexual transmission increasing especially among females. [PoC 41241

Fifty-one deaths of IDUs were identified among the estimated population of 9424 Glasgow IDUs. This representsa mortality rate of 0.54% (male 0.42%1 female 0.85%). Over 90% of deaths were attributed to overdoseorsuicide. AIDS caused only one deathbut 19% of caseswhere HIV antibody status was known were positive. The mortality rate among HIV-infected IDUs (3.8%) was significantly higher than among HIV-negative IDUs (0.49M. For the forseeable future, it is unlikely that AIDS will have much impact on mortality among Glasgow IDUs because of the low prevalence of HIV infection in this population (1.4%) and because HIV-infected IDUs are dying for reasons other than AIDS. [PoC 45 5 81

Twenty-seven new HIV infections among Glasgow's estimated 9400 IDUs were derived from a model involving 9.9 million annual injections, of which 684 922 involved used injection equipment with an assumed random chance of encountering HIV- infected IDUs (estimated prevalence 1. 1 %) and an estimated HIV transmission rate similar to innoculation of infected blood via occupational cutaneous injury (0.36%). Extrapolating an estimated incidence rate of 0.2% to the Glasgow IDU population would yield 18 new infections. The congruence of the two estimates suggeststhat the model contains the important parameters for estimating the probability of HIV transmission among IDUs in Glasgow. [PoC 46921

In Rome 9946 IDUs were estimated by capture-recapture and 11639 were estimated by an overdose multiplier formula. As 70% of AIDS cases in Rome occur among IDUs, the estimated prevalence of drug injectors is crucial to predicting the dynamics of the AIDS epidemic [PoC 4774]

HIV prevalence across age and risk groups in patients attending a hospital in New York City was followed from 1988 to 1991. HIV seroprevalence increased in the population overall from 4.5% (1988) to 5.7% (1991). HIV seroprevalence increased particularly in Black females possibly due to increased male-to-female transmission or increased drug use. [PoC 40141

Prevalence of HIV infection among 1747 IDUs recruited from six sites in California varied between 4.9% and 19.4%; they were geographically isolated predominantly African-American communities at increased risk of HIV infection. Independent riskfactors associated with HIV infection were high injection frequencies and a history of attending shooting galleries. The inverse relationship between use of non-injected drugs (e.g. crack cocaine) and the absence of sexual risk variables in the model suggests that the primary route of HIV infection in this population was shared injection paraphernalia. [PoC 47001

Some 1368 individuals underwent anonymous HIV testing in Boston in 1990 with a HIV seroprevalence rate of 8%. The risk factor for 85% ofpersons tested was IDU. Primary care services at an anonymous HIV testing site were well accepted by the drug-using population. This model was successful in reducing delay between notification of HIV-positive status and presentation for care resulting in earlier treatment of HIV infection. [PoD 58011

Thee hundred and forty-three HIV currently infected former IDUs were followed of whom 136 were observed to seroconvert. Estimated time from seroconversion to AIDS was 10- 11 years and this appears to be similar to estimated rates for other HIV-infected groups. [PoC 44741

Eighty-seven IDU women and 44 heterosexual women with AIDS from Rhode Island were followed from 1986 to 1992. The odds ratio for progression among the IDUgroup was 1.7 compared to the heterosexual group. Controlling for the initial CD4 level, the odds ratio for IDUs was 2. 1. Female IDUs progressed more rapidly than women without an IDU history when measured by a decrease in CD4 level. [PoC 43671

Death certificates were compared to an AIDS registry on persons aged 25-44 years who had certain infectious conditions and died in 1987 in two counties of the USA. Forty per cent of the 412 deaths were in the AIDS registry; 7.3% of the remaining 205 cases were found to have AIDS, 21.5% had confirmed HIV infection without AIDS and 15. 1 % had suggested HIV infection. Adding these cases increased HIV- related mortality by 9.2% for confirmed HIV infection and 15.6% for confirmed and suggestive HIV infections. The IDU mortalitywouldbe increasedbyl 2.3% by adding confirmed HIV infections and by 18.9% by adding confirmed and suggestive HIV infections. [PoC 45 5 11

Blood samples (1816) of a cohort of 1007 Baltimore IDUs were tested by polymerase chain reaction (PCR) for detection of HIV- I proviral DNA over a period of 2.5 years. Of 979 seronegative IDUs, only 7 (0-04%) were PCR positive for GAG and ENV Four PCR positive seronegative IDUs seroconverted in less than 6 months. Three were PCR negative in the follow-up visit. The PCR technique did not provide a significant difference in the detection of occult HIV-1 infection and latent HIV- I infection is very uncommon among I DUs. [PoC 442 61

Silent HIV infection in 69 HIV-negative IDUs forming part of an Italian multicentre study was shown not to be rare by HIV provirus detection by PCR. [PoC 47631

Race and homelessness (which is significaptly greater among Blacks) were strong predictors of MIV testing among 1423 New York City IDUs. [PoD 5 7401

Of 22 583 accumulated AIDS cases reported in Brazil up to November 1991, IDUs represented 20% of all cases with 24 out of 27 states reporting AIDS cases causedbyIDU. IDUs comprised 2% of cases reported in 1985 and 28% of cases reported in 199 1. IDU was a more prominent cause of AIDS in women than sexual transmission in 1987 and 1988. Among adolescents with AIDS in 1991, 61 % of cases were caused by IDU. [PoC 42631


Direct HIV infection of CNS tissue was infrequent among IDUs with early disease suggesting that T-cell [PoC 44 181 infiltrates may contribute to early neuronal damage. [Po1337631

Three IDUs co-infected with HIV-I and HTLV-II were compared with two HIV-I positive IDUs serving as controls. The host range of HIV- I was enlarged to CD8 (T8) lymphocytes and B lymphocytes in IDUs coinfected with HIV- I and HTLV-11. [PoA 2095]

Turnout necrosis factor-(x levels in non-infected and HIV- I - infected I DUs during the first 6 months of observation were higher than in eight controls and increased over the 24 months of observation. Interleukin- I concentrations in IDU groups were not significantly higher than in the controls. [PoA 2333]


HIV infection was associated with increased prevalence of HPV but not abnormal cervical pathology among IDU and non-IDU prostitutes. [PoB 30651

Thrombocytopenia was more common in HIV- infected and non-infected IDUs than HIV-infected and noninfected homosexual men and related to long-term effects of injecting. [PoB 3 5241

Streptococcus pneumoniae antibodies were more common among HIV-infected IDUs than non-infected IDUs or non-infected controls. [PoB 34651

One hundred and fourteen episodes of bacterial endocarditis in 100 patients with HIV infection (all IDUs) showed same pattern as in non-HIV infected. Bacterial endocarditis tended tooccur in early stage of HIV infection. [PoB 3 2021

Seventy-seven per cent of HIV-infected patients who developed streptococcal pneumonia (representing 8% of total) were IDUs suggesting a possible role for antipneumococcal vaccine. [PoB 31431

HIV seroconversion in 5 of 18 IDUs was accompanied by bacterial pneumonia attributed to decreased lymphocyte reactivity at the time of seroconversion. [PoC 4551

Immunological markers and symptoms were similar prognostic indicators for AIDS among homosexual men and IDUs apart from serum P2-microglobulin. Male and female IDUs had similar rates of progression. [PoC 4418]

A majority of 120 IDUs self-administering a shortened version of the Minnesota Multiple Personality InvenThree IDUs co-infected with HIV- 1 and HTLV11 were tory (MMPI) while attending a 2 1 -day heroin detoxificompared with two HIV- I positive IDUs serving as cation in San Francisco were classified within the clinical range for depression, psychopathic deviation, schizophrenia, hostility and arousal seeking. Respondents with clinical scores on the depression scale were far more likely to report needle sharing. Men with extremescores on the hostility scale were less likely to report any condom use. Risk reduction interventions should take account of psychological impairment. [PoC46911

Seventy per cent of Verona hepatitis B virus (HBV) seronegative IDUs were recruited into an H13V vaccinationprogramme with only 11 % dropping out. Ninety-one per cent completed all three injections with 92% developing anti-HBV Age, sex, duration of injectionhistory, HIV or hepatitis C virus (HCV) serostatus did notpredict compliance with vaccination. The relatively high compliance with HBV vaccination reflects an increased awareness among IDUs of the need for health protection and suggests some optimism may be justified for other vaccination programmes (including hepatitis C virus and HIV). [PoC 47961

One hundred and twenty H IV- infected I DUs in methadone maintenance treatment in New York were screened for tuberculin reactivity or anergy by multitest CMI. Each calendar year, 8% without a prior positive result became newly PPD (purified protein derivative) positive. Incidence of tuberculosis was 9.7/100 person,years among 25 HIV-infected PPD-politive IDUs receiving less than 12 months isoniazid prophylaxis. Among 27 such IDUs completing 12 months of isoniazid, no cases occurred in 53.4 patient-years. Where the HIV epidemic is advanced, isoniazid prophylaxis shouldbe given to HIV-infected anergic IDUs ifprevalence of tuberculosis is high. [TuC 05671

Multidrug,resistant tuberculosis outbreaks occurred in four New York City hospitals and a fifth hospital in Florida with a sixth hospital in New York State also developing a cluster of tuberculin conversions among healthcare workers exposed to multidrug-resistant tuberculosis patients from a prison. HIV infection amplifies the risk of progression from TB infection to disease and shortens the time between infection and infectiousness. Settings where HIV-infected persons were congregated are conducive to rapid propagationof tuberculosis if introduced. [TuC 0568]

Heavy smokers were over three times more likely to have had Pneumocystis carinii pneumonia than light smokers. [WeC 10301

IDUs and 230 homosexual/bisexual men in Baltimore were tested for neurospychological performance. Age and IQ rather than risk group accounted primarily for the difference in neuropsychological performance regardless of serostatus and CD4+ count. [Th 15131

Preliminary evidence comparing 197 drug users with 236 non-drug users indicates that, in Geneva, the likelihood of ever having received zidovudine is similar in all risk groups with less than 300 CD4 cells. [PoD 57221

Of 325 HIV-infected injecting drug users with a CD4 count of less than 500 followed since January 1990, 45 initiated zidovudine use in 6 months. Use of zidovudine in this population was low. Advanced immunosuppression was the major factor predicting zidovudine use. The major factor predicting zidovudine non-use was usual source of care at an emergency room. Current heavy drug use, alcohol use and lack of drug or alcohol treatment did not discriminate between zidovudine use vs nonuse. [PoD 5 7431

The likelihood of 197 Geneva IDUs receiving zidovudine compared to 236 non- IDUs was similar when comparing HIV-infected individuals with less than 300 CD4 cells. [PoD 57221

No difference was found in compliance in attendance at an AIDS clinic between 188 US minority women with past or present intravenous drug use and 83 minority women with no such history. [PoD 5 7331

The rate of decrease of CD4 (T4) lymphocytes was followed as a marker of progression of HIV infection in 55 Argentinian IDUs compared to 17 homosexual men and 23 heterosexual men and women, all of whom were infected with HIV There were no significant differences in the rate of decrease of CD4 between different groups. It was not possible to evaluate the progression of illness in the IDU group. [PoA 2380]

Current drug users, and other patients with AIDS attending twelve hospitals were compared. Over half the hospitalised drug users continued active use but only a small number created problems within their service. [PoD 34 591

Of 81 HIV-infected pregnant women in Baltimore between 1988 and 1992, 72% were diagnosed in the obstetric clinic: 47.6% were IDUs. IDU was associated with greater maternal age, unintended pregnancy, preterm delivery and smaller mean birth weight. Perinatal transmission was 30.8%. Obstetric and neonatal complications were most closely associated with maternal IDU.[PoB36861

Ninety-five children born to HIV- infected mothers in Rome were compared to 74 controls. Features of HIV embryopathy were confirmed when controlling for drug use of mothers but lower birth weight and gestational age reported in children born to HIV-positive mothers disappeared as differences when maternal drug use was controlled for. Clinodactyly, partial syndactyly 2-3 and simian crease were more frequent in children with HIV infection whereas other minor abnormalities were more likely to have been related to maternal drug use. [PoB 3 63

One hundred and twenty-eight HIV-1-infected women and their infants attending Yale Hospital were followed. Source of maternal HIV- 1 infection was IDU in 49%. Maternal drug use before or during pregnancy did not predict infant infection status. Fifty per cent of infants conceived within 6 months of their mothers' initial HIV- 1 positive test were found to be infected compared to 9% of infants conceived more than 6 months after initial seroconversion. [P(C 42151

Of 177 patients infected with HIV admitted to Boston City Hospital with fever and a documented source, a history of IDU was present in 66%. Patients with a history of IDU were more likely to have a bacterial infection. Of 195 documented fever sources, 60% were bacterial infections. Patients with bacterial infections had distinct presentations and improved outcomes cornpared to those with fever from other sources. [Po13.3 83 21

Of23 078 Californians with AIDS (1998/91) 7. 1 %had a diagnosis of HIV encephalopathy. Gay/bisexual IDUs were more likely to be diagnosed with HIV ncephalopathy (OR 1.4 -1.9). [PoC 40091

In 557 HIV-infected IDUs in Baltimore, decline in CD4 lymphocytes occurred mainly in those with <600 cells at entry. In HIV-infected homosexual men, however, decline occurred at both high and low entry of CD4 cells. This difference may reflect longer duration of infection, lower levels of CD4 cells at entry or past tracking of CD4 cell decline in the homosexual male cohort compared to the IDU cohort, but isprobablynot due to laboratory factors. [PoC 4 5 5 41

HIV and active tuberculosis co- infection were found in 315-380 former IDUs attending a drug treatment clinic. Poor living conditions and inadequate medical care were common. The prevalence of tuberculosis among HIV patients was 14%. [PoC 455 51

Seven patients making nine suicide attempts in New York City were compared to 322 other clinic patients. Patients attempting suicide were more likely to have a substance abuse diagnosis (71% vs 45%) and to be HIV infected. [PoB 3 5 981

Of 84 Edinburgh patients prescribed zidovudine, 41 were IDUs inmethadone maintenance with 43 remaining IDUs not receiving any opiates. In the 13 patients who had received more than 500 000 mg zidovudine, 6 of the 9 patients receiving methadone had adverse haernatological effects whereas none of the 4 IDUs not receiving methadone had adverse effects. The risk of haematological toxicity due to zidovudine was not increased by concomitant opiate administration until the total dose of zidovudine exceeded 500 000 mg. [PoB38541

Interferon treatment may induce persistent remission of chronic HCV in HIV-infected patients with remission occurring at lower rates than in HIVuninfected patients. [PoB 33 731

Of 2356 adults with AIDS in New York State in 1987-88, 47% were IDUs. Drug users were less likely than non-drug users to receive zidovudine. Zidovudine users were more likely than non-users to have a primary provider (60% vs 36%). HIV-infected persons with, primary provider are twice as likely to have been prescribed zidovudine than those with more fragmented care. [PoB34181

HIV seroprevalence in the North Carolina prison inmates was estimated in 1989/90 to be 3.14%. Seventeen per cent of HIV- infected inmates had <200 CD4+ celIS/MM3. The estimated incidence of Pneumocystt's carinii pneumonia with and without aerosolised pentamidine prophylaxis was 9.4 and 19.8 cases/year respectively. The cost of prophylaxis was US$1800/year for aerosolised pentamidine and US$50/year for trimethoprim/sulphonamides. Assuming in-patient treatment, estimated to cost US$14 000 per episode, prophylaxis will save treatment costs of US$140 000 per year. The cost of prophylaxis is US$54 000 assuming that 25% of patients do not tolerate trimethoprim/sulphonamides. The overall savings of providing prophylaxis in this population may be significant but depend upon the prophylaxis used and local cost of aerosolised treatment. [PoB 33171


Of 255 HIV- infected individuals in Norway, 7 were alsopositive for I-ITI-V-1 and all seven were IDUs. [PoC 46431

Seventy-seven per cent of 251 HIV-infected IDUs in Milan were also infected with HCV which was higher than other risk groups and was also associated with more abnormal LF1s. [PoB 3 88 11

Of 83 French IDUs with non-A, non-B hepatitis, 96% of HIV-infected IDUs and 92% of non-HIV- infected IDUs were positive for HCV. HIV did not alter the severity of liver disease due to HCV [PoB 34681

Seventy,five Spanish (mainly former) IDUs treated with interferon for HCV showed improvement in 52% at 12 months with no significant difference between those with or without HIV infection. [PoB 34601

Of 78 IDUs in Argentina 61 % also had HCV butithere was no evidence of HCV sexual transmission. [PoB 38821

Of 105 Brazilian IDUs, 65% were infected with HIV and no relationship was found with syphilis (19% FTA positive) suggesting that HIV transmission was mainly blood borne in this population but raising the possibility of transmission to sexual partners of IDUs. HBV markers in IDUs were 14% (HBsAg) and 61 % (antiHBc IgG). [PoC 46961

Of 107 HIV- infected female prisoners in New York 64% also had HCV with prior IDU being the most corntion association (97% vs 53%). [PoC43751

A high prevalence of HIV (37%, 42%, 39%, 7 1 %), HBV (58%,59%,57%,74%), and syphilis (10%, 5%, 5%, 11%) was reported from four Spanish cities. [PoC 46991

Of 2846 Baltimore IDUs, 23.3% were infected with HIV-1, 7.2% infected with I-ITI-V-11, 2.4% co-infected and 70.8% not infected with either virus. HTLV-11 appears to have been introduced earlier than HIV- 1. Despite an association between HIV-1 and HTI-V-11, different subpopulations seem to be involved. [PoC 43921

Of 198 Brazilians infected with HIV 57% were also infected with HCV with IDU alone (64%) or in combination (96%) being the most important risk factor for HCV. [PoC 40481

Of 622 Vancouver IDUs attending an NSEP, 45% were infected with HBV, 46% with HCV, 46% HAV and 3 1 % negative for all three. [PoC 46031

Of 1108 IDUs in the Veneto region of Italy, 18% were infected with HIV and 60% had markers of HBV infection. However, the seroprevalence for both viruses declined between 1986 and 1991. [PoC4797]

Among 1544 IDUs recruited from 39 drug agencies in England and Wales in 1991, prevalence of HIV was 1.6% but 3 1 % tested positive for anti-HBc. A minority were at risk ofcontracting (19%) or transmitting (16%) HIVor H13V The epidemic of HIV infection in IDUs in England and Wales is unlikely to be marked by the rapid spread seen elsewhere. Vaccination against H13V infection should be routinely offered. [PoD 54931

Eight hundred and seven HTI-V-1 positive samples (confirmed by West & Blot) and RIPA (from 17 countries in Latin America, North America, Africa and the Pacific basin) were tested for HTI-V-11. Significant HTLV-11 infection was found in North America (USA 54%, Canada 14.5%) compared to 2.8% for the rest of the world. The rate of HTLV-11 infection was highest in IDUs: 80.2% in the USA and 91.8% in Italy. The prevalence of HTLV-11 infection among non-IDUs is concentrated in North America with sporadic occurrences in other regions where the predominant occurrence is among IDUs. [PoC 40351

Among 305 Amsterdam drug users, of whom 70% injected recently, the prevalences of HIV, HBV and HCV were 31%, 68% and 65% respectively and the cumulative incidences over 2.5 years were 14%, 13% and 25% respectively. At present, HBV transmission is more closely linked to heterosexual contact than HIV or HCV as being female, a prostitute and having sexual contact were only related to seroconversion for HBV

Borrowing injection equipment was only related to seroconversionforHIV (andnotHBVorHCV)~ From 1986/87, the prevalence of HBV climbed significantly and the incidence of HWhasslowed. High levels of risk behaviour were found in both HIV-seronegative and HIV-infected drug users despite prevention programmes. The HIV epidemic should be monitored using HIV testing itself. Preventive activities targetting sexual behaviour should be expanded. [PoC 44731

In June/July 1990 a malaria outbreak was detected in Bauru City, Brazil involving 108 individuals with no history of travel to endemic areas or blood transfusion. All injected cocaine and shared injection equipment. Investigation revealed VDRI- 18%, HBsAg 38%, HIV 58%. [PoC42531


Needle and syringe exchange and outreach bleach programme attracted non-overlapping populations of Oregon IDUs who had comparable rates of reduced risk behaviour. [PoC 42841

Of 313 IDUs recruited at three of Berlin's six needle and syringe vending machines, HIV prevalence was only 4% in those injecting for less than 5 years. Vending machines appeared to supplement other outlets. [PoC 42641

NIDA is developing a single shot, non-reuseable needle and syringe which includes consideration of possible advantages and disadvantages. [Pol) 508 11

Seven hundred and forty-six Californian IDUs demonstrated greater reduction in risk behaviour following an enhanced risk-reduction programme compared to either a basic or a standard programme. [PoC 4800]

Predominantly negative perceptions of condoms were registered by 18 incarcerated Canadian IDUs. [PoC 45161

IDUs were motivated to recruit fellow IDUs in a prevention programme in Connecticut using monetary rewards. [PoD 5 5 691

One hundred and thirty-three of 190 seronegative IDUs (70%) were recruited into a hepatitis B vaccination programme of whom only 11% dropped out. Of those who started 91% completed the programme of three injections: of these 92% developed anti-~IBV. These results suggest that an HIV vaccination programme might also meet with reasonable compliance. [PoC 47961

The notion of a pharmacy-based needle and syringe exchange scheme proved popular with English IDUs but not with pharmacists. [PoD 50831

A survey among European countries demonstrated rea, sonable agreement about the estimated number of IDUs but wide disparity in HIV prevalence. Methadone treatment was available in about half the countries. Lack of requirement for prescription for needles and syringes, few condom distribution pro-' grammes, and few evaluation studies were consiste t findings. [PoD 50741

One hundred and sixty-nine Philadelphia IDUs exposed to cognitive skill training demonstrated a be , ter capacity to generate coping options forhigh-risk situations compared to an attention control group. [PoD 53701

The time periods between the first report of AIDS cases relating to IDUs and the implementation date of pre, ventive interventions addressing HIV infection among IDUs were compared across the 12 member states of the European Community according to the intensity level of legal sanctions against drug users. Countries that were less repressive were more likely to have responded to prevent AIDS among IDUs. [WeD 1077]

A syringe tracking and testing system was established in Connecticut with a sample of returned needles from a needle and syringe exchange (NSE) tested for the presence of HIV proviral DNA. Earlier modelling predicted a drop in HIV incidence from 6-4/100 IDUs per year to 4.3 as a result of the NSEP. PCR tests of 850 needles covering 69 client IDUs determined mean incidence levels falling inside the coverage intervals on three occasions whereas the fourth fell just outside the predicted range. It was concluded that the data were con, sistent with the incidence reduction forecast ear ier. [WeC 10921

The development of a public policy in the prevention area has been guided by pubic perceptions, social atti. tion programme of whom only 11 % dropped out. Of tudes and personal bias of policy makers rather than sound public health principles. Conclusions are drawn about steps necessary to overcome the social and governmental hurdles in developing a sound public policy agenda for primary AIDS prevention. [ThD 15 631

Kits containing a syringe, condom, plastic container for disposal of the syringe, alcohol wipes, distilled water, information leaflet are distributed in 96% of 750 pharmacies in the Basque region of Spain. Of 100 injecting drug users, 99% knew of the kit, 10% always used it, 46%usedit 70-80% of the time and 26% used ithalfthe time. [PoD 50961

One hundred and fifty-one African-American IDUs contacted by street outreach workers in San Francisco were followed up over 6 months when they reported a substantial reduction in cocaine (5 1 %), heroin (28%) and heroin-cocaine combination (43%), with a substantial reduction also in the frequency of borrowing, frequency of sharing withsex partners, increased cleaning and increased condom use. [PoD 50751

Of IDUs who were not in treatment in Puerto Rico 2144 were randomly assigned to an educational programme and compared to others who received standard treatment. Forty-six per cent of the total sample were infectedwith HIV Drug use risk behaviour declQied by 18.3% and sex risk behaviour by 11.6%. Risk reduction in IDUs assigned to the enhanced intervention showed no difference to those assigned to a standard intervention. [PoC 42791

One hundred and fifty-two IDUs attending eight needle and syringe exchange schemes established in Wales in 1990 were compared with 176 non-attenders. There were few demographic differences between attenders and non~attenders but only 9% of attenders had shared syringes in the previous 4 weeks compared to 41 % of non-attenders. [PoC 4293]

Seventy-seven IDUs using a needle and syringe ,txchange in Oregon were compared with 335 particiipating in a bleach outreach project. IDUs attending a needle and syringe exchange significantly decreased sharing of syringes, renting syringes, borrowing syringes, cleaning injection equipment from intake to 6 months. Similar behaviour was found among outreach clients with little overlap between the populations. The two programmes should be seen as complementary strategies. The establishment of syringe exchange resulted in a decrease in discarded syringes in public places. [PoC 42841

Sixty-four per cent of needles and syringes collected from shooting galleries in Miami demonstrated the presence of antibodies to HIV Of bleach-treated needles 100% were negative for HIV antibodies using a technique of bleach cleansing promoted by an outreach proj ect. Bleach of about 0. 1 % (v/v) concentration was toxic for peripheral blood mononuclear cell viability. HIV exposed to bleach above I% concentration for 2-24 hours was inactivated. [PoC 4292]

Risk behaviour was studied in 265 IDUs attending a needle and syringe exchange in Washington State and compared with 93 non-exchangers. Exchange users reported higher initial levels of risk behaviour which subsequently fell to a level slightly lower than that among non-exchangers. There were no seroconversions in the 125 person-months of follow-up for exchange users. Incidence of hepatitis B among drug injectors has fallen from 70to 13 inthe5 yearsto 1991. The exchange is attracting high-risk drug injectors and reducing their risk behaviour. [PoC 42831

Three hundred and thirteen IDUs attending six needle and syringe vending machines in Berlin were interviewed: 33% were not in contact with treatment facilities currently. Vending machines were the primary source for sterile injection equipment for 77%. HIV seroprevalence rose steeply with longer duration of intravenous drug use. Vending machines appear to be an important supplementation to syringe exchange programmes and pharmacies. Seroprevalence was 4% in those injecting for less than 5 years. However, risk behaviour is particular a problem in young IDUs. [PoC 42641

One hundred New York City IDUs were interviewed with the proportion using an unofficial needle exchange increasing over 18 months from 55% to 69%. The date of cessation of sharing appeared earlier in needle exchange users by an average of 1 year. No difference in injection frequency was observed for attenders and nonattenders and attendance at needle exchange was not associated with increased needle sharing. Needle exchange does not increase the risk of HIV transmission and appears to recruit IDUs at high risk of transmitting and assists in their risk reduction. [PoC 480 11

Seven hundred and forty-six IDUs recruited by outreach in California (prevalence 5.5%) were assigned to either a basic two session or an enhanced four session HIV risk intervention programme: 70% were successfully followed up; 75% of participants in the enhanced programme and 60% attending the standard programme achieved a reduction in a composite index of injection risk. More participants in the enhanced programme increased condom use and decreased the number of sexual partners. Individual counselling can reduce injection and sexual HIV risk especially when reinforced by risk-reduction training. [PoC 48001

Three hundred and nineteen IDUs in prison and 173 attending a drop- in clinic reported high levels of awareness and encouraging utilisation rates of Montreal's only needle and syringe exchange. [PoC 47951

HIV prevention skills taught one-to-one accompanied by provision of bleach and a condom kit appeared to reduce sharing of injection equipment with strangers, cleaning of injection equipment and an increase in condom use among IDUs in Indianapolis. However, condom use with regular partners declined. [PoD 50851

Philadelphia IDUs randomly assign~d to an eight-session cognitive skills training programme demonstrated an increased ability to generate alternatives to real-life interpersonal problem situations compared to an attention control group. [PoD 53 701

Fifty-four per cent of 503 Glasgow IDU s attended nee 1 dle exchanges in the 6 months prior to interview: 15 of 31 variables explained 17% of the variants between the two groups. Needle exchange attenders reported lower levels of risk behaviour (fewer drugs injected, less frequent injection, less frequent sharing) compared to nonattenders. However, they were more likely to pass on injecting equipment. Adding knowledge of discriminating variables resulted in correct classification of 66% of respondents compared to random assignment (50%). These findings support needle and syringe exchanges playing an important role in maintaining a low rate of HIV infection among the 1500 of Glasgow's estimated 9400 injectors who currently attend exchanges. [PoD 5076]

A nursing counselling intervention was evaluated among 2033 IDUs attending emergency rooms in three US cities. Intravenous use of heroin, cocaine and heroin-cocaine combination declined; sharing of injection equipment and other HIV risk behaviours (including including sexual behaviours) declined at follow-up compared to initial interview. [PoD 509 11

No difference in behaviour change could be demonstrated between clients of a Toronto needle exchange and others with both groups in some instances showing improvement. One hundred and thirty-eight of the needle-exchange clients sought referral to drug treat~ ment during 18 months. Seroprevalence was 4.3%. Attribution of effect is difficult: the extent of co-intervention is difficult to assess. Process measures such as utilisation data and referral to treatment are useful mea, sures. [PoC 47981

The infection rate among Vancouver's estimated 4000 high-risk IDUs remains 2%. It is estimated that the 80 HIV-infected IDUs will cost the healthcare system US$12.6million (1991). To quantify the potential savings to future healthcare expenditures resulting ft m the prevention of transmission of HIV among IDUs as a result of the needle and syringe exchange programme, the number of individuals and cost in 1991 in dollars was estimated for HIV prevalences of 5%, 15% and 25%. If 25% of IDUs were infected with HIV, the cost to the healthcare system for 1000 individuals would be (1991) US$157.5 million. [PoD 54061

Between November 1989 and March 1991, 300 IDUs attended a mobile needle and syringe exchange in Paris on 2200 occasions to obtain 4600 syringes, Demand is growing despite severe penalties for carrying injection equipment. The exchange has contacted an ostracised population of whom 46% are infected with HIV, 60% are homeless and 59% do not have medical care. [PoD 50841

Data from paired baseline and 6-month follow-up interviews with female prostitutes and female sexual partners of IDUs in three sites were compared to esti, mate changes in frequency of unprotected sex, and risk associated with injecting practices. Comparison of aggregate responses suggested substantial behaviour change which was not confirmed by repeated measures multiple reggression analysis in which the exten of each subject's participation in programmes, interven. tions and co-variates representing subjects' socio. demographic characteristics were introduced as independent variables and on which a selection bias interview completion was modelled. The latter method revealed that behaviour change was less dramatic and that participation in interventions was not a significant predictor of behaviour change. Positive programme effects cannot be inferred from aggregrate prelpost data revealing positive behaviour change. Multivariate statistical techniques predicting individ,ual behavioural outcomes, taking into account selection bias and extent of participation in interventions, e required to prevent overestimation of programme effects. [PoC 44901

Only 3 of 230 addicted prostitutes attending a clinic in Amsterdam became pregnant in 199 1. Prevalence of HIV infection was 34%. Over 90% received other conaceptives in addition to condoms. Ninety-three lients accept Depo-Provera injections to prevent ~pregnancy. An active policy of birth control can he egective in preventing HIV transmission from mother to child among addicted prostitutes with a disruptive lifestyle.[Pol) 5 6141

Assuming that counselling and testing reduces HIV onddrug and sex risk behaviour in 10% of clients, and t extended counselling when added to counselling d testing reduces drug risk behaviour in an addition15% and sex risk behaviour in an additional 2 5%, unselling and testing were estimated to avert 4.2 HIV ections in one year at a cost-per- infect ion- averted of S$13 000 whereas extended counselling averts 7.5 IV infections at a cost- per- infection -averted of S$23 000. Both counselling and testing and extendedcounselling prevent HIV infections at costs considerably lower than US life medical care cost for HIV infection (US$90 000). Although extended counselling is more effective, its cost-per- infection -averted I is higher. [PoC 47941

Ninety-four opiate users in a therapeutic community in c USA were randomly assigned to a 6-hour small up education intervention or a comparison condion. There were no significance differences between intervention and comparison conditions on follow 1 . The results do not support the efficacy of this AIDS ucation programme. [PoD 53 441



No protective effect of drug abuse treatment was found In 1043 community-recruited Chicago IDUs suggesting that subpopulations of IDUs enter treatment periodically to control rather than eliminate their addiction. [PoC 42941

One hundred and fifty-two Pennsylvania IDUs who began methadone treatment had an HIV incidence rateof 3.5% compared toarateof22% in 103 IDUswho remained out of treatment. [PoC 4268]

Prevalence of HIV infection in 552 Verona IDUs between 1985 and 1991 entering methadone treatment and receiving a comprehensive prevention package was 23.8% but annual incidence fell progressively from 2.3% (1985) to 0.0% (1989 and 1990). [PoC 47991

Forty Verona IDUs who seroconverted between 1985 and 1991 were compared with 40 matched IDUs who remained uninfected. Significant differences were: days of treatment with methadone, number of methadone treatments, average total dose of methadone, less heroin use during treatment and shorter periods without treatment. (Average dose was 23 mg HIV NEG and 16.5 mg HIV POS.) [PoC 42821

Of 1.1-1.8 million US IDUs, less than 15% are enrolled in any form of drug treatment of whom 40-60% remain in methadone maintenance and 15-25% remain in therapeutic communities for more than 1 year. In the aggregate, less than 3% of US I DUs reduce HIV risk behaviour through treatment because treatment reaches onlya small portionof IDUs and fails to retain them long enough to produce sufficient HIV risk behaviour change. Harm reduction strategies such as needle exchange could reach a higher proportion of IDUs at lower cost thereby producing superior population outcomes. [PoC 42501

Among 220 New York City IDUs being followed in a longitudinal study, the methadone group had evidence of greater immune compromise than the nonmethadone group independent of HIV infection as indicated by decreased CD4 percentage, increased CD8 percentage, increased absolute CD8 count and decreased CD4/CD8 ratio. Variables which could have contributed to these findings could not be excluded. [PoC45571

Among 1043 IDUs recruited by outreach workers in Chicago, no significant protective effect of prior treatment was observed. A significant increase in the risk for HIV was observed with increasing weeks in treatment during the 5 years prior to baseline interview. This finding, different from clinic-based studies, is consistent with the identification of a population of IDUs entering treatment periodically to control rather than to eliminate their addiction. [PoC 42941


Of 1044 women prisoners in a state prison in Connecticut, 15% were infected with HIV Ninety-five per cent of HIV-infected prisoners and 60% of all prisoners were 1 DUs. [PoC 4.3 5 81

The transmission of tuberculosis and multidrug-resistant tuberculosis and its link to HIV has become embedded in New York City and State institutions for the homeless, mentally impaired and the incarcerated. The epidemic of tuberculosis and multidrug-resistant tuberculosis within the epidemic of HIVfor those institutionalised has been assisted by a security and healthcare delivery system unable to cope with the enormity of the problem. [MoD 00361

A decrease was noted in HIV prevalence among Spanish prison inmates from 28.4% in 1989, 24.2% in 1990 and 20.6% in 199 1. HIV testing in prisons is well accepted if inmates are properly informed, testing is voluntary, treatment is available and no discriminatory measures are taken. Prisons are a very important setting for HIV prevention programmes. Cooperation between the health services inside and outside prison is critical to improving the efficacy of the programmes. [MoDO0371

Assuming each IDU prisoner only has access to one syringe at a time and that 9 (range 2-25) IDUs share a syringe randomly, that the average duration of incarceration is 30 weeks (range 5-50) and that the prevalence of HIV in the prison population is 10% (range 5-15%),2% (range 0.05%-32%) ofsharers in prison in Britain can be expected to become infected per year amounting to 62 (range 2-986) prisoners annually. These estimates were derived from modelling techniques which suggest that transmission in prison could be reduced by decreasing the time IDUs spend sharing syringes, the proportion of IDUs who are infected and share syringes, and reducing the probability of transmission in each episode of risk. [MoD 00381

National prison policies on HIV infection were compared with WHO 1987 recommendations using data from multiple sources. Live prisoner education was pro. vided between 22% and 100% of jurisdictions, mass mandatory testing was provided between 13% and 100% of jurisdictions, testing on request was provided between 22% and 63%, HIV segregation between 8% and 7 1 %whereas no systematic segregation was i ple. mented in between 29% and 100% of nations. It was concluded that actual policies differed substantially from WHO recommendations. [MoD 00391

The majority of 184 Swedish IDUs who volunteered to be interviewed about their sexual behaviour and drug use while in remand prison were (prior to incarcera. tion) sexually active with the amphetamine users reporting more partners and higher frequency of inter, course than heroin users. [WeD 10741

A self-help group for HIV-infected inmates in an Argentine prison succeeded in improving their physi. cal and mental condition resulting in less violence [ThD 15001

Of 503 (1990) and 535 (1991) Glasgow IDUs 52% had been in custody in the 6 months prior to interview whom a considerable proportion (19% 1990, 15% 1991) had injected while in custody. The proportion injecting with used equipment declined between 990 and 1991 (73%,46%). An increased awareness amon IDUs of the risks of sharing injection equipment, espe, cially in custody, and the will to avoid these risks were reported. [PoD 50541

Ten IDUs infected with HIVwho died soon afterpriso release were compared to nine HIV-infected IDUs attending a hospital. Non-AIDS death rate in HIV patients was higher for recently imprisoned patients. These deaths occurred in a younger population wh had a higher CD4 count. Eight of the ten prison releas deaths were due to drug overdose with seven of these deaths occurring within 14 days of release. Release fro prison is associated with an increased non-AIDS deat rate in male IDUs. Decreased tolerance to methadone following abstinence requires modification of dosage. [PoD 50611

The prevalence of HIV- I among 1692 male inmates i the Sao Paulo prison system between 1990 and 1992 was estimated to be 16.5%. The most important risk factor associated with HIV- 1 infectionwas IDU (52%). Excluding all other risk activities, the odds ratio for IDU was 14.2. Homosexual/bisexual contact was reported by 236 inmates. The rate of transmission among 470 seronegative inmates in 1990 was 0.6%. [PoC43191.

Of 499 available prisoners in Edinburgh, 75% volunteered a saliva sample for HIV testing and 76% cornpleted a questionnaire. Prevalence of HIV infection was 4.5%. All HIV-infected individuals were IDUs indicating a prevalence of HIV infection among IDUs of 25%. Residents in Edinburgh injecting prior to 1983, agedbetween 26 and 30 and a history of injecting within prison were all strongly correlated with seropositive status. [PoC43141.

One hundred and four female inmates in a New York City gaol included 40.4% convicted for drug-related crimes. Of the total sample, 74% reported being tested for HIV and 2 1. 1 % being HIV positive. [PoD 50531

Seven hundred and eleven seropositive inmates have been identified in Maryland prisons with HIV-infected inmates more likely to be female, from metropolitan areas, and IDUs (males 84% vs 34% females). By offeringvoluntary testing to all entrants, HIV detection has increased 300% and is identifying seropositives earlier in the course of infection enabling more inm;tes to receive prophylaxis. [PoD 50621

HIVprevalence was estimated in 9871 Italian prisoners at 24.2% in 1985 and declined to 16. 1 % in 199 1. However, this may have been a reflection of the selection process as some known seropositive prisoners avoided re,testing on subsequent admissions to prison. One hundred and five seronegative prisoners were followed for at least 15 months and non-seroconverted. Knowledge of HIV transmission was good. Understanding of the meaning of seropositivity was poor. Most of the HIVinfected prisoners were IDUs. The risk of acquiring infection during incarceration was thought to be limited. [PoC47101

HIV prevalence among 1002 prison inmates in the Bahamas was 10%. Among 2 24 cocaine addicts in the community, HIV prevalence was estimated to be 13.4%. [PoC 40451

A voluntary confidential HIV testing programme was implemented in Rikers Island Prison Complex, New York City and ran in 1991 when 4532 inmates were tested. Overall, HIV prevalence was 23% (males 22%, females 29%). HIV prevalence was 25% for inmates older than 21 years and 6% for younger inmates. Approximately 450 prisoners were currently receiving antiviral therapy. [PoC 43 181

A voluntary survey of 1338 prisoners in Mozambique was conducted in 1990/91 with syphilis confirmed in 7.8% of inmates and HIV in 0.6%. Homosexual activity while in prison was reported by 6% of the males. There is a need for introduction of an STD programme including educational intervention and condom distribution. [PoD 50711

In two groups of prisoners in Argentina, HIV prevalence was estimated at 10.8% (n= 114) and 44.4% (n = 2 5 9). [PoD 5 0601

Of 1807 Los Angeles minors in a correction centre, 8% had been involved in blood-sharing rituals and 2% in satanic worship which are high-risk behaviours for HIV infection. These adolescents also have higher rates of
other risky behaviours. [PoC 434 11

Information was collected from 46 prison systems in 26 countries to monitor trends. Providing information to prisoners and personnel of HIV transmission is practised in all systems. Condom distribution during incarceration and on release is practised in a growing number of systems: 21/46 (45.6%) while in prison; 16/46 (34.8%) on release or parole. No prison systems have adopted syringe/needle distribution or exchange during incarceration or on release. A few systems (14/46) 30.4%, have accepted distribution of a disinfectant with specific detailed instructions on cleaning injection materials. Considerable resistance to condom distribution still exists but this appears to a policy option should be strongly supported by health authorities. Disinfectant distribution appears acceptable to both prisoners and prison authorities. It appears realistic to propose its widescale adoption in prisons housing significant numbers of drug users. [PoD 50691

Nine of ten Canadian provinces in the Federal government have developed policies regarding prisoners with HIV/AIDS. These policies vary widely in their content; they also differ substantially from those of many other developed countries. Common to all Canadian policies, however, is that they do not authorise involuntary testing of prisoners for HIV infection; require that attempts must he made to house infected prisoners among the general prison population; stress the importance of maintaining the confidentiality of the HIV status of inmates; establish educational programmes about HIV/AIDS for both inmates and staff; and indicate protective measures to be taken by staff to avoid H [V exposure. No consensus has yet been reached about the distribution of condoms. Many of the Canadian provinces have not yet followed the example of the Federal government which made condoms available to prisoners rorajanuary 1992. [PoD 54231


Seroconversion among 4644 IDUs in 14 US cities was 1.83/tOO person-years with the major predictors being city seroprevalence (1.80 increased risk for each tO'Yo rise in city seroprevalence), cocaine injection, no previous drug treatment, and renting or borrowing injection equipment. Lesbians were at increased risk. Drug abuse treatment had protective effect even for those who continued injecting after leaving it. [PoC425 1]

Of 688 Brazilians aged 13-19 infect~d with HIV up to 1991,47% were IDUs thus representing the inost common risk factor. [PoC 42731

Although US female IDUs who smoked crack were at increased risk of HIV due to increased high-risk sexual behaviour and homelessness, higher rates of HIV infection were not observed. [PoC 4 5431

One hundred and ninety-six Perth IDUs were interviewed in 1989 (ofwhom 38 were followed in 1990) and a further 150 recruited in 1990. Greater reduction occurred in high-risk injecting than sexual practices. HIVprevalence was 1.75%. [PoD 53651

Self-reported reduction of HIV risk behaviour has been followed by stabilisation of HIV prevalence in New York and Bangkok. Factors predictive for reduced risk behaviour inNew York were: talking about AIDS with drug-using friends, with sex partners and with family members; having been tested for HIV; knowing that a person infected with HIV can look healthy; and recruitment from a detoxification site. The Bangkok data confirmed three of these: talking about AIDS with drugusing friends; having been tested for HIV; and knowing that a person infected with HIV can look healthy. [PoD 54901

Of 1245 Sydney IDUs, the 36% who used benzodiazepines reported higher levels of HIV risk behaviour. [PoD 53581

Cocaine use is increasing among 148 Rome IDUs in a methadone programme and is associated with increased HIV risk behaviour without a commensurate rise in perception of risk. [PoD 54921

A number of psychosocial variables predicted the likelihood of sharing w ith strangers among 2 5 7 N ew York IDUs. [PoD 53661

Crack smokers who do not inject drugs or have rnale homosexual sex are nevertheless at high risk of acquir, ing HIV infection heterosexually. In a logistic regression analysis of 877 New York City crack smokers, sex, trading, genital ulcer disease and city of residence were significantly associated with HIV infection. [WeC 10281

Four hundred and thirty homosexually active men in England and Wales asked to recall sexual encounters and drinking episodes in the previous week showed that alcohol use does not increase high-risk sexua behaviour. [WeC 10291

Interviews of 1050 IDUs in London in 1990 and 1991 confirmed that condoms are commonly used by London IDUs for penetrative sex withcasual partnersofthe oppositesex though relatively few always use them. Use with primary partners is less common and less frequent and the majority never use condoms with primary partners. [WeD 10781

Ninety-nine patients attending an HIV clinic in London responded to a questionnaire: 68% of the respon, dents drank alcohol at least twice a week, 62% sm ked cigarettes every day, 78% used'poppers', 46% cocaine, 48% Ectasy and 78% cannabis. [ThB 1058]

Three hundred and ninety IDUs in Puerto Rico under. going detoxification (30% HIV infected) were studied to identify factors predicting seropositivity. Injecting for more than 10 years, injecting between 6 and 10 years, sharing injection equipment within penal insti. tutions, or being homosexual/bisexual were the major predictive factors (in declining order). [PoC 42491

In the USA 5932 homeless IDUs were compared with 32 000 IDUs with a residence. High-risk behaviour was more common among the homeless population with 80% reporting sharing injection equipment, 70% multiple sex partners, 50% never using condoms and 40% reporting risk behaviour in other cities (usually AIDS epicentres). [PoC 43 171

`Front- loading' is a drug-sharing practice in which an ector syringe is used to measure drugs and to give a ir share' to another injector by 'spouting' the drug toanother's syringe. Thirty-nine percent of 207 New York City IDUs front-loaded in the previous 2 years as id30%duringthe last 30days. Front- loading can be an .mportant route for viral transmission. Prevention mpaigns and researchers should he aware of this prace. [ThC 15 5 11

Of 69 homeless persons with HIV in Massachusetts, U was the risk behaviour in 90%. H IV primary care be provided to this population provided outreach d a multidisciplinary approach are adopted. [PoD 5713]

Forty,five per cent of 306 drug users in Sao Paulo ported the use of crack. High-risk sexual practices re common among crack users. [PoD 53531

Two hundred and five men aged 18-29 in Harlem have used crack more than three times a week and engaged in
high rate of unprotected sex. [PoC 549 11

Seven hundred and eleven Rotterdam drug users and 82 Amsterdam drug users were studied to determine ctorsassociated with initiation of drug injecting. Us initiated heroin use significantly earlier than n,IDUs. Injecting was also positively related with ing male, older and with use of cocaine an pi s. ecting is becoming less common among drug users in Netherlands. [PoD 50801

IDUs who were infected with HIV, or knew that they were seronegative, were compared with each other and
with IDUs who have not been tested in asample of 2330 IDUs recruited from 12 European cities. In comparison
with IDUs who had never been tested, IDUs who knew that they were seropositive were: more likely to always
use condoms and never give used injection equipment to other IDUs. However, they showed no differencewith regard to not reusing equipment from other IDU's or using disinfected equipment. Compared to never tested IDUs, known seronegative IDUs tend to inject drugs more often without using equipment from other IDUs or disinfected equipment but there was no difference in condom use. The study suggests that knowledge of HIV serostatus may help reduce HIV transmission from HIV-positive IDUs to others through both safer drug-injecting and sexual behaviour. Among HIV-negative IDUs, safer drug-injecting behaviours were slightly better accepted than safer sexual behaviours. [PoC 42811

HIV prevalence among 1592 female street-walking prostitutes in New York City was 34.7% overall and 61.2% among IDUs. HIV-positive street walkers were: older, had a longer duration of prostitution, were more likely to be Latinas and used condoms less often during oral sex and vaginal intercourse. Among IDUs, sharing needles in galleries carried the highest risk. Among crack users, those who primarily perform fellatio were 1 more likely to be infected than those who primarilyperform vaginal intercourse. [PoC 45601

Items from the Risk for AIDS Behaviour (RAB) scale correlated highly with response to the AIDS Initial Assessment (AIA) scale. Summary scores for the RAB scale were able to discriminate seroconverted from non-convertered and responses to the RAB scale were found to 1 ~ave significant predictive power subsequent to seroconversion. [PoC 449 11

Four hundred IDUs in Rio de Janeiro were compared with 996 in New York City. Fewer IDUs shared syringes in Rio than New York City but Rio IDUs engaged in less activities to reduce the possibility of HIV transmission. Seroprevalence in Rio is now 38% and rising. [PoC 42461

Among 400 IDUs recruited in Rio de Janeiro, 38% were infected with HIV. Seropositivity was associated with: recruitment from a treatment site; injecting more than four times a week; sharing syringes in the previous 6 months; injecting before the age of 18; injecting for more than 5 years and being a homosexual/bisexual male. Independent risk factors were: treatment recruitment; homosexual/bisexual contact; and duration of drug use exceeding 5 years. Seroprevalence increased from 3 3 % (1990) to 3 8% (199 1). [PoC 42651

Among 7 500 IDUs recruited from six US cities, prevalence of HIV infection increased in one city, decreased in the second city, and remained stable in four cities .

Self-reported needle sharing decreased and needle-cleaning increased in all six cities: attendance at shooting galleries decreased in four cities, increased in one city and remained unchanged in one city. Risk levels have continued to decrease for the two most recent cohorts extending reductions reported for the first six cohorts. Risk behaviour levels have decreased substantially in the four eastern cities although elsewhere changes are modest. [PoC 42471

Three hundred and thirteen IDUs in treatment were followed up at 4 and 18 months in Seattle. Risk reduction was observed for both drug use and sexual behaviour and was stable between 4 and 18 months. High-risk needle use was associated with being out-oftreatment, decreased use of equipment from drug stores and having an IDU sexual partner. Continued unsafe sex was associated with prostitution, using drugs to meet sexual needs and alcohol and drug use. [PoC42771

Among 50 female IDUs compared to 125 male IDUs, no differences were found in sharing injection equipment. However, among those who had shared, women shared a higher proportion of injections and had more sharing partners. Also, women tended to divide their drugs and syringes more frequently. [PoC 42 541

Over25 000 mate IDUs recruited from 42 US!*ti* tes were classified as exclusively heterosexual (94%), bisexual (4%) or exclusively homosexual (2%). HIV infection rates were higher for homosexual IDUs (34%) than bisexuals (26%) or heterosexuals (13%). However, b isexual men were more likely to report high-risk injection practices including sharing injection equipment, not cleaning used equipment, use of alcohol, marijuana and crackcocame and receiving money ordrugs for sex, Bisexual and homosexual men were more likely to inject amphetamines compared with heterosexual men who injected heroin and combinations of heroin and cocaine. Condom use was reported more frequently by homosexual IDUs. [PoC 42671

Among 38 IDUs interviewed in 1989 and again in 1990 in Perth, Australia, use of sterile injection equipment increased, use of bleach increased and condom use for vaginal intercourse with all partners increased. The most common response of respondents who said they had changed their drug use behaviour in 1990 was to report ceasing to share needles whereas in 1989 the most common response was reduced sharing. HIV seroprevalence was 1.75%. [PoD 53651

Cluster analysis identified four types of prostitutes among 358 sex workers in San Francisco. Classification was based primarily on injection drug use and number of customers with groups differing on attributes such as perception of risk, education level and types of sexual behaviour. Novice and professional IDUs had higher prevalence of HIV infection whereas novice and professional non-IDUs had higher prevalence of syphilis. [PoD 56091

Eighty per cent of 1593 New York City and 92% of 590 Bangkok IDUs reported changing their behaviour because of concern about AIDS. Seven independent factors were associated with risk reduction in New York City: talking about AIDS with drug-using friends, with sex partners, and with family members; having been previously tested for HIV; knowing that a person can be infected with HIV but look healthy; years of formal education; and recruitment at a detoxification centre. Three of these factors were also found in Bangkok: talking with drug-using friends about AIDS; having prev iously been tested for HIV; and knowing that a person can be infected with HIV but look healthy. In both cities, risk reduction appears to he a social process developing from discussions among drug-using friends and more likely among drug injectors with sophisticated knowledge of HIV infection who have also undergone experience that personalises the threat of AIDS (such as HIV counselling and testing). [PoD 54901

In two samples of over 500 IDUs recruited in 1990 and 1991 in London using a similar sampling and study methodology, sharing rates decreased (47%, 34%), passing on used injection equipment decreased (38%, 27%), accepting used equipment decreased (35%, 2 7%), mean number of sharing partners decreased (3.6, 2.5), but not the mean number of IDUs from whom used equipment was accepted. Syringe sharing continues to decline in London but is still relatively common. [PoD 50821

A high prevalence of benzodiazepine use was reported among 1245 IDUs in Sydney, Australia and this was associated with higher levels of HIV risktaking including higher injection frequency, morefrequent injection with borrowed equipment, sharing with more people and greater likelihood of injecting while intoxicated. Benzodiazepine users were more likely to engage in prostitution and had a greater likelihood of having a sexually transmissible disease. [PoD 53 581

Fifty-one percent of 80 transvestite prostitutes in Rome re IDUs of whom 44% reported needle sharing. erall HIV prevalence was 59%. [PoD 53731

Independent factors associated with high-risk haviour with a primary partner for 599 IDUs in Bangkok were: not talking with sexual partner about IDS; being HIV seronegative; and having injected with equipment used by others. For the New York sample the factors were: not talking with sexual partner utAIDS; not having changed behaviour because of cern about AIDS; and being HIV seronegative. IV-seronegative status was associated with highk behaviour because previously tested seroitive subjects were less I ikely to engage in high-risk ual behaviour. The determinants of highrisk sexual haviour with primary partners were similar in g ok and New York and suggest the need to encourdiscussion of AIDS with primary partners and proe means for HIV-infected persons to learn of their status. Reducing drug-injection risk behaviour appears have a positive effect on changing sexual behaviour th primary partners. [PoD 54961

Of 48 Rome IDUs 5 7% reported cocaine use in the last ar (representing an increase from 1990). Cocaine rswere more likely to reportsexual intercourse with IDUs, casual intercourse in the last year and prostitun. Perception of increased risk of HIV infection among heroin users represented by use of cocaine was ited. [PoD 54921

Of 324 Italian lesbians 11% reported previous injecting g use with 29 out of 34 reporting previous sharing of ection equipment. [PoD 5 2091

Psychosocial risk factors strongly influenced needle ring with strangers in a sample of 25 7 male IDUs ruited in New York. [PoD 53661

A sample of US public and private school students in es 9 through 12 surveyed in 1989, 1990 and 1991 onstrated a general reduction in sexual behaviour reported illicit drug injection increased from 2.7%


1989 to 5.7% in 1991 and occurred within most mographic subgroups. Students who received HIV ction were less likely to report HIV-related drug sexual behaviour than those who did not receive instruction. [PoD 51141

Among 8 1 clients of a London ne-ed-le and syringe exchange programme, use of filters was more common with heroin than other drugs but was generally common. Eighty-nine per cent of clients recommended provision of clean filters. [PoD 5078]

Analysis ofchanging initial method ofelrugadministration among 400 London heroin users showed a gradual transition from initiation by injection to initiation by inhaling sublimated heroin after heating on tin foil Cchasing the dragon') with 94% of initiations since 1988 conducted by this practice. Prevention planning should include consideration of initiation into chasing as well as consideration of subsequent transition from initial chasing to injecting. [PoD 5 7831

Eight hundred and ten Baltimore IDUs (442 seronegative, 368 seropositive) were followed over a 2-year interval and demonstrated significant reduction in drug-use practices especially among those with antibody to HIV. Increased abstinence and reduced injection equipment sharing was seen among seropositive IDUs especially soon after receiving antibody test results. Relapse was more common among seronegative IDUs. [PoD 53521

Five hundred and three (1990) and 535 (1991) Glasgow IDUs were interviewed with increased condom use for vaginal intercourse reported with casual partners (16%,1990; 31%, 1991). High condom use by female prostitutes with clients but low use with regular partners was reported even where the regular partner was an IDU. [PoD 53821

Among 1375 New York City IDUs, women were more likely to have shared a syringe at last injection (21 % vs 13%) because they were more likely to inject with one or more other IDUs. Among thosewho injectwith others, women were more likely to share, and women were more likely to have a sexual partner present in a group with whom drugs were shared. Women were more likely to inject in groups and had twice the chance of havmgasexual partner in their sharing group. The frequent overlap of sexual and drug partnership is an important obstacle to risk reduction among female IDUs. [PoD 5093

Recruitment of 17 041 IDUs through street outreach from 63 US sites and followed for 6 months revealed that injection frequency decreased from 50 to 25 times per month across the entire sam ple but significantly greater decreases occurred among persons who were never homeless. Increased new needle use was comparable across groups classified according to residential status. Becoming homeless undermines the extent to which IDUs are able to adopt and maintain reductions in some risk behaviour. Proposed prevention strategies include: outreach and treatment services targetting homeless IDUs, decriminalisation of needle possession and development of user-friendly needle -exchange strategies. [PoD 5 3 541

Ninety-two HIV negative Amsterdam IDUs were folowedfor4.5 months. Reportedsafe injecting increased from 83% to 89%. Safe injecting at follow-up was associated with safe injecting at recruitment, self-efficacy and participation in an exchange programme. Independent predictors of safe injecting were degree of participation in the exchange programme and vulnerability. Social cognitive factors contribute little to the prediction of safe injecting possibly because situational factors and drug-use habits impair control of injecting risk behaviour in HIV-negative I DUs. [PoD 54941

Following recruitment of 2330 IDUs from 12 European cities, 77% reported at least one measure to reduce or stop needle sharing available in their atea and 44% reported condom promotion. Behaviour change was reported by 77%. Reduction in the number of sexual partners was reported by 170/t), reduction in sexual practices with unknown partners by 13%, increased use of condoms by 8% and sexual abstinence by 6%. More frequent use of new needles and syringes was reported by 27% with 20% seeking treatment, 11 %reporting cessaion of sharing and 10% reporting reduced drug injecting. [PoD 58521

Among 389 Berlin IDUs, condom use in steady rela, ionships was 13% increasing to 2 1 % with casual partners and 59% with clients. Independent predictors of consistent condom use were: not injecting with used injection equipment, engaging in prostitution; knowledge of being HIV infected; communicating about AIDS with friends; and not taking CNS sedative drugs in addition to heroin. It is concluded that the sexua ly transmitted HIVepidemic amongIDUs isconcealed by focusing on drug-related aspects of the problem. [PoD 50891

Of African-American female IDUs 21 % were infected with HIV compared to 4% of Caucasian female IDUs in San Francisco. Predictors of seropositivity were: African-American race; history of syphilis; recent crack-cocaine use; and more than five needle-sharing partners in the last year. When stratified by race, the only risk factor retained by both racial groups was number of needle-sharing partners. [PoD 5 5 201

One hundred and nine New York City female IDUs were more likely to report changing their sexual behaviour if: they had more than one sexual partner; reported feeling comfortable asking partners to use condoms; had higher depression scores; or had lost friends or family because of AIDS. Reported sexual behavi . ur change is less likely among African-American women or those who held stronger beliefs that luck plays the largest role in getting AIDS. [PoD 55171

Two hundred and sixty-four IDUs in methadone maintenance treatment in New York City increased condom use on all occasions from 19% to 29% in 6 months. Factors predicting sexual behaviour change were: concern about AIDS, higher depression scores and less extensive criminal histories. Each of these variables was shown to have an independent effect. [PoC 4262]

One hundred and twenty-six female IDUs in methadone maintenance treatment in New York City showed increased risk of HIV infection associated with crack use, unstable housing and interrupted mother-child co-residence. The association of highrisk behaviours with housing instability suggest that secure housing, as well as other social factors which contribute to stability, may lessen the risk of HIV transmission. [PoD 52161

Low levels of HIV-risk behaviour and HIV seroprevalence was found in hidden populations of IDUs in contact with IDUs in ambulatory care or reached by preventive projects or surveillance. [PoD 53741

The majority (87.6%) of 330 Athens IDUs reported behavioural changes following the AIDS epidemic. However, only half (50.6%) of those reporting behavioural change had not shared injection equiprnent during the previous 6 months. Fifty-nine percent of respondents reporting casual sexual partners in the previous 6 months had never used condoms during that period. A considerable discrepancy exists between per, ceived and actual behaviour change. [PoC 426 11

Of 136 adolescents (age 10-19 years) with AIDS in Sao ulo, Brazil between 1980 and 1990, IDU was the most quentrisk behaviour (76%). [PoC42711

Of 1995 San Francisco women screened by an outreach amme, 1286 (65%) were found eligible and 676 3%) were enrolled in family planning services. Nonolleeswere more likely: to have injected drugs in the vious month; to have terminated pregnancy; to ve had an IDU partner or multiple sex partners. rollees were more likely to have received family ingservices inthepastand tobemotivated topretHIV infection. Personal risks included tradingsex r drugs (26%) or money (46%), injection drug use 33%) andcrackcocaine use (82%). [POD 5526]

Among 71 prostitutes in Santos, Brazil, intravenous g use was a much more reliable predictor of HIV ction than heterosexual intercourse. Of IDUs 98% re infected with HIV compared to 17% for whom terosexual transmission was the only risk factor. 41871

HIV testing among over 189 000 volunteers in gentina between 1987 and 1991 demonstrated that is the major risk factor associated with HIV infecwith635 of 1107 HIV positive carriers being IDUs. e prevalence of HIV infection among IDUs te*sted 32% overall. [PoC 40641

Current IDUs (1328) attending a Barcelona hospital r the first time between 1984 and 1990 were studied. Overall seroprevalence of HIV was 67.6% with little changeduring the study period; 31.3% of 430 initially ronegative patients seroconverted during a mean fol-up of 14.5 months with 5 1 % seroconverting by 33 onths. HIV infection was associated with: age tween 21 and 30; low educational level; unskilled rk; delinquent activity; and imprisonment. Seroversion was also associated with: heroin dose; frency of injecting; duration of drug use; sharing of ection equipment; and sexual intercourse with other s or prostitutes. Education, heroin dose, duration druguse, sharing of injection equipment, sexual cont with prostitutes and imprisonment were all indeent factors. [PoC 463 21

The proportion of IDUs reporting injecting with and ingonused equipment fell significantly among two mplesof Glasgow IDUs recruited in 1990 and 1991.

The mean number ofpeople from whom equipment was received and to whom equipment was given both fell. The prevalence of HIV remained low (2.0% 1990, 1. 1 % 199 1). [POD 50941

Residents (8263) of 23 metropolitan areas of the highest AIDS prevalence cities in the USA were interviewed by telephone to determine prevalence of risk factors and use of HIV testing. One per cent of males and less than 1% of females were classified as at risk from IDU and 46% of these males and 73% of these females had ever been tested for HIV infection. [PoC 46001

Two hundred and fifteen sexually active female drug users were recruited in Harlem, New York, of whom only 19% were currently in drug treatment, 26% had been in gaol in the last year and 22% were currently homeless. Seventy-four per cent reported trading sex for drugs or money. Trading was associated with: younger age at first intercourse; being single; illegal income; crack use; STD in the last year; perceived risk of contracting AIDS; more frequent use of condom; or other protection and psychiatric symptoms. [PoC 46581

Of 288 male IDUs recruited in Harlem New York, 48% had been in gaol in the last year and the same proportion were homeless. Fifty-three per cent reported trading sex. Of these, 43% bought sex with drugs or money, 38% sold sex for drugs or money, and 19% both bought and sold sex. Significant independent predictors of trading sex are: being single, homelessness, incarceration in the last year, current regular use of alcohol, and crack, STD in the last year, perceived risk of contracting AIDS and psychiatric symptoms. [PoC 46591

Eighty-four IDUs in Sydney, Australia, were presented with 20 hypothetical situations involving different levels of needle availability, bleach availability, craving, familiarity with sharing partners, previous sharing, privacy of site and the degree to which the individuals were high, drunk or stoned. The presence of each of the dependent variables was varied randomly on a computer bar graph representation. Nineteen per cent of the variance was accounted for by a vector of predictors. Ratings of the probability of sharing increased with the frequency subjects experienced hypothetical situations in real life, craving, familiarity with sharing partners and affect-related decision-making strategies. Ratings of the probability of sharing decreased with needle availability, satisfaction with sharing decisions and the use of avoidant decision-making strategies. HIV risk in IDUs determined by an interaction between situation, psychological and social factors. [Pol) 53621

Hornosexual/bisexual male IDUs in Rio de Janeiro had similar or lower levels of drug-related risk behaviour than IDUs but were more likely to be infected with HIV This suggests that the high seroprevalences factors due to sex with infected males and that they may be a bridge in which HIV has reached a lot of other IDUs. [PoC 54281

Comparison was made of 2330 IDUs recruited from 12 European cities in 1989 and 1990 according to whether or not they knew their HIV serostatus for at least 6 months prior to interview. In comparison with IDUs who had never been tested, IDUs who knew they were seropositive forHIVwere more likely toalways usecondoms, never give their used equipment to other IDUs but showed no difference in regard to not using equipment from other IDUs or using disinfected equipment. Compared to nevertested IDUs, IDUs with a negative test tended to inject drugs more often without using the equipment from other IDUs or using disinfected equipment. Knowledge of HIV serostatus mly help to reduce HIV transmission from HIV-positive IDUs to others through safer drug injecting and safer sexual behaviours. For HIV-negative IDUs, safer drug-injecting behaviours seemed to be slightly better accepted than safer sexual behaviours. 1PoC 482 11

Of 129 London IDUs who usually injected heroin 12% were infected with HIV compared with only 3 % of 89 who usually or always took their heroin by'chasing the dragon'. Ten per cent of the subjects who had 'ever injected'were infected with HIV compared with 4% of the 55 subjects who had never injected. HIV prevalence among non-injecting heroin users was greater thanexpected. Safer patterns of injecting or changes to safer injecting may have reduced the extent of transmission associated with injecting so that sexual activity may become more significant as a transmission route. [PoC4559]

A large number of new clients (10 000) attended an HIV testing site in Puerto Rico between 1989 and 1992. HIV seroprevalence overall was 10.3% but IDUs had a rate of 41 %. Nine variables were found to predict serostatus best: intravenous drug use; use of crack; heroin; needle sharing during IDU, history of blood transfusion, no sex, sex with one partner; frequency of condom use; and sex of client. Discriminant function correctly classified 82% of HIV-positive and 74% of HIVnegative individuals. [PoC 44701

Thirty-nine heterosexual IDU couples entering methadone maintenance treatment in Seattle comprised 13 (33.3%)'safe'couples and 66.7%'unsafe'couples. Thirteen couples gave discrepant reports about unsafe injecting behaviour and 20 couples were discrepant regarding number of needle-sharing partners. Women were more likely to report risk than male partners. Safe couples were more likely to be African-American and had more education. Unsafe couples were more likely to inject cocaine. [PoC4656]


The global population of injecting drug users is larger than hitherto realised. More countries now appear to have injecting drug users than were recognised a decade ago. This appears to be due to a combination o increased recognition and also increased spread of the practice of injecting illicit substances.

HIV infection is also spreading rapidly among injecting drug-user populations in many developing countries. There is some evidence that HIV infection is now spreading rapidly from injecting drug users especially in some developing countries. The rate of spread of HIV infection among injecting drug users appears to be slowing in many developed countries. However many developed countries now have very large numbers of HIV-infected injecting drug users. In such countries, there is now growing evidence that substantial spread of HIV infection is occurring from the large pool of HIV-infected injecting drug users to other populations.

High-risk behaviour is declining in IDUs in almost all countries that attempt to measure the prevalence o risk behaviour. The trend to saferpractices is universally more pronounced for injecting than for sexual behaviours. Factors that increase the likelihood of seroconversion include: renting or borrowing injection equipment, high city seroprevalence, cocaine consumed by smoking or injection, no previous drug treatment, female sex, homosexual/bisexual male, lesbian, member of a minority race, history of incarceration, recruitment from treatment site, injecting drug use history of more than 5 years, youth and homelessness. Factors that reduce the risk of HIV infection include: exposure to drug abuse treatment; reported 'stopped sharing'because of fear of AIDS; having a regular sexupartner; talking about AIDS with drug-using friends, partners and with family members; having been ted for HIV; knowing that the person infected with can look healthy; and recruitment from a detoxifiion site. Factors that increase the likelihood of highbehaviour include many of the factors associated increased risk of seroconversion but also benzopine consumption and psychosocial variables.

Some authorities at the conference discussed the possibility that efforts to restrict the supply of illicit injectable substances may (inadvertently) be exacerting the spread of HIV by opening up new trade test for drugs with HlV foIlowing as a secondary phemenon. Comparison of 1987 WHO recommendans on prison policy with national policies demonstratesthat these recommendations have had little influence. In Europe, countries with the most restricted drug policies were generally the slowest to respond to the threat of HIV infection and IDUs by introducing needle and syringe exchange schemes.

Needle and syringe exchange and outreach bleach ammes appeared to attract non-overlapping populations with comparable rates of reduced risk behaviour over time. Needle and syringe vending machines appear to supplement other outlets. Some evidence is emerging that structured counselling results in enhanced risk reduction compared to more basic interventions.

Moststudies have found a powerful treatment effect for methadone with a lower HIV incidence rate compared to IDUs who remained out-of-treatment. The incidence of HIV infection in IDUs entering methadone treatment in some studies fell progressively over time.

In summary, this conference presented an alarming picture of HIV infection among injecting drug users globally. Research at this stage is still establishing new and potentially important findings but the translation of research findings into policy, where this has been studied, is generally lamentable.

Municipal authorities in Amsterdam have recently constructed a building in the centre of the city which carries a large Latin inscription 'Homo Sapiens Non Urinat In Ventum.' It would be in the spirit of this edict for policy makers in the 50 countries where HIV infection among injecting drug users is now a problem to consider responding to the epidemic on the basis of findings presented at this conference.

Dr A.Wodak, Director, Alcohol and Drug Services, St Vincent's Hospital, Sydney, Australia


Show Other Articles Of This Author