Home | Issues | Guidelines | More... | Topics | Search |
Nick Crofts and Campbell K Aitken
Abstract - Introduction - Methods - Results - Discussion - Acknowledgements - References - Authors' details
Make a comment - Register to be notified of new articles by e-mail - Current contents list - ©MJA1997
Abstract |
Objective: To assess changes in risk behaviours for
transmission of bloodborne viruses and incidences and prevalences
of these viruses in a field-recruited cohort of injecting drug users. Design: Prospective longitudinal cohort study. Setting: Metropolitan and rural Victoria, June 1990 to December 1995. Subjects: 626 current injecting drug users (i.e., who had injected drugs within the previous 12 months). Main outcome measures: Seroconversion to HIV and hepatitis B and C viruses (HBV and HCV); risk behaviours for infection. Results: HIV incidence was very low (0.2 per 100 person-years). HCV incidence was high (10.7 per 100 person-years), but fell throughout the study, although the downward trend did not reach statistical significance. HBV incidence was moderate (1.8 per 100 person-years) and did not fall. Prevalence of risk behaviours, notably sharing needles and syringes, decreased significantly. Conclusions: Significant change has occurred in the risk behaviours of the cohort, confirming results of cross-sectional studies of injecting drug use. This change may be responsible for the apparent decline in HIV and HCV incidence. Further studies are needed to monitor the incidence of bloodborne viruses in injecting drug users. Efforts to decrease risk of transmission should continue. |
Introduction |
Bloodborne viruses, especially the human immunodeficiency virus
(HIV) and hepatitis viruses B and C (HBV, HCV), pose major risks to the
health of people who inject illicit drugs.1 This is largely because of transfer
of blood through sharing of contaminated injecting equipment or of
environmental contamination in injecting settings, which in turn
depends on the behaviour of injecting drug users (IDUs).2
The Victorian Injecting Drug Users Cohort Study (VICS) is the first longitudinal cohort study of IDUs carried out in Australia. At the inception of VICS in 1989, there was little Australian research into injecting drug use, and most studies were small, cross-sectional and retrospective.3 Information was needed about changes in IDU behaviour and their impact on the incidence of infection with bloodborne viruses among IDUs. The aims of VICS were:
|
Methods | |
Subjects and assessment |
Subjects were current IDUs (i.e., who had injected drugs in the
previous 12 months) and were recruited between June 1990 and March
1995. Because we wished to study a cohort of IDUs not
necessarily in treatment and because of the difficulties in
recruiting and following up IDUs over any length of time, novel
methods were developed. These have been outlined previously.
4,5 Subjects were recruited by peer
outreach workers (former or continuing IDUs with extensive
experience of the IDU "scene"), primarily through their own social
networks and, to a lesser extent, from agencies (such as needle
exchange programs and prisons). These outreach workers undertook
all interviews, blood sampling and follow-up, after training in HIV
counselling at the Melbourne Sexual Health Centre.
Follow-up continued until December 1995. Interviews and collection of blood samples were initially intended to be at six-month intervals, but difficulties in tracking participants meant that interviews occurred opportunistically. The interview questionnaire included questions on demographics, frequency of drug injection, injecting history, needle-sharing, and other injecting practices. The full questionnaire is available from the authors on request. A blood sample was collected after interview if the situation allowed and the participant was willing. Inevitably, samples were not collected at every interview and fewer IDUs provided samples than were interviewed. Blood was tested for antibodies to HIV, HCV and HBV core antigen. |
Statistical analyses |
For calculating incidences, we assumed that seroconversion
occurred on the date halfway between a seroconverter's last negative
and first positive test. Confidence intervals for incidences were
calculated using an exponential-error formula.6 The chi-squared trend statistic ( 2TR) was used to measure
significance of trends in incidences and behaviours over time.7
|
Ethical approval | The original study design and sub sequent modifications (including venepuncture, delivery of serological results, and pre- and post-test counselling by the peer outreach workers) were approved by the Institutional Ethics Committee of Fairfield Hospital, Melbourne, Victoria. |
Results | |
Cohort description |
Data were obtained from 716 IDUs, but were insufficient for this
analysis for 90. The remaining 626 IDUs provided contact and
background data at an initial interview and detailed data at one or
more follow-up interviews (maximum, 11). A total of 1663 follow-up
interviews were conducted, with 267 IDUs (42.7%) completing only one
follow-up questionnaire and 359 (57.3%) completing two or more, at an
average interval of 259 days. Subjects were recruited from the
community (431; 70%), agencies (134; 22%) and
prisons (51; 8%). The prisoners were recruited between May and
September 1994.
Descriptive information was collected at the first interview. Of the 626 participants, 39% were women and 60% were men (one respondent was transsexual). Median age was 27.7 years (range, 14.9-63.2), and median age of first injection was 18 years (range, 10-60). Of 512 participants who had injected drugs in the previous month, 63.9% specified heroin as the drug most often injected and 33.6% specified amphetamines. For 420 participants who had injected in the previous week, median injection frequency was four times per week (range, 1-210). One hundred and eighty-eight participants (30%) reported injecting with a shared needle or syringe in the previous month. |
Prevalence of bloodborne viruses |
Nearly 6000 serological tests were performed on blood samples from
531 participants. Prevalences of HIV, HBV and HCV are shown in Box 1. On
their first blood test, almost two-thirds of IDUs (62.4%) were found
to have been infected with HCV, almost half (45.2%) with HBV and 3% with
HIV. Between 1990 and 1995, the prevalence of HIV antibodies
among those tested in each year declined significantly, from 6.3% to
0.7% (2 TR =
12.49; P < 0.005). The prevalences of antibodies to HCV and
HBV core antigen also varied over this period but with no discernible
trend.
|
 | |
Incidence of bloodborne viruses |
Only one person converted from HIV-seronegative to
HIV-seropositive during the 599.8 person-years at risk captured by
the study, an overall HIV incidence of 0.2 per 100 person-years (95%
confidence interval [95% CI], 0.0-1.4).
Seroconversions and incidences of HBV and HCV infection are shown in Box 2. Five participants seroconverted to HBV during 276.3 person-years at risk, an overall HBV incidence of 1.8 per 100 person-years (95% CI, 0.8-4.3). However, HBV incidence increased between 1992-1993 and 1994-1995, when four of the seroconversions occurred. However, first-test HBV prevalence did not vary significantly from year to year.
Nineteen participants seroconverted to HCV during 177.6 person-years at risk, an overall HCV incidence of 10.7 per 100 person-years (95% CI, 6.8-16.8). One hundred and sixty-five participants remained HCV-seronegative throughout the study. Although there was a downward trend in HCV incidence between 1990-1991 and 1994-1995, comparison of 95% confidence intervals showed it did not reach statistical significance. |
Prevalence of risk behaviours |
The Figure shows prevalences of four risk behaviours over the 11
six-month periods of data collection. Bingeing (a period of heavier
than usual drug use for the individual which may impair the ability to
maintain safe behaviour) was common, with 41% of participants
reporting at least one binge; however, the percentage who reported
bingeing declined significantly over time (2 TR = 4.5; P <
0.03). There were also significant downward trends in the percentage
of current IDUs who shared needles and syringes (2 TR = 5.5; P <
0.02), shared rinsing or mixing water (2 TR = 12.8; P <
0.001) and were sometimes injected by others (2 TR = 15.1; P <
0.001). Mean frequency of injecting in the cohort varied over time,
ranging from five to nine times per week, without any significant
trend.
|
Association between risk behaviours and incidence |
For the 202 participants who completed at least three interviews and
reported continuing to inject, an attempt was made to relate the major
risk behaviour -- sharing needles and syringes -- to HCV infection
status (see Box 3). A gradient in both prevalence and
incidence of HCV infection was apparent, related to the frequency
that sharing was reported, although 95% confidence intervals for the
incidences and a 2 test for
the prevalences showed that these gradients were not statistically
significant (2 TR = 2.5; P = 0.28). However, 43 who were HCV-positive and
two who seroconverted never reported sharing needles or syringes.
|
 | |
Discussion |
We found that incidence of HIV among a cohort of Victorian IDUs was very
low (0.2 per 100 person-years at risk), while incidence of HCV was high
(10.7 per 100 person-years) and incidence of HBV was moderate (1.8 per
100 person-years). There was evidence that HCV incidence has
decreased among Victorian IDUs, from 16.6 per 100 person-years in
1990-1991 to 8.1 per 100 person-years in 1994-1995, although the
small number of seroconverters meant that the downward trend did not
reach statistical significance. We also found significant declines
in prevalence of risk behaviours, notably sharing of needles and
syringes.
There are several possible explanations for this apparent decline in risk behaviours. Firstly, it may be a real decline, as a similar decrease in risk behaviour was seen between the Melbourne arms of two cross-sectional national IDU surveys, in 19898 and 1994,9 respectively. Prevalence of sharing of injecting equipment (in the previous month) fell from 38%8 to 13% 9 of respondents, a statistically significant difference. These surveys had sample groups largely independent of the VICS cohort. Another explanation is that participants became less likely to report risk behaviour with time; this change in social desirability bias might also have been responsible for the difference between the results of the two national surveys.3 However, in our study, participants' increasing trust in their peer workers -- the basis for continued follow-up -- makes this unlikely. It is also possible that the cohort experienced greater attrition over time among IDUs whose behaviour was relatively risky. This might also explain the decreasing HCV incidence, with those most at risk of HCV infection seroconverting first. To evaluate this possibility, the study was kept open to new recruits until March 1995; the rise in HBV incidence in 1994-1995 reflects enrolment of a group at heightened risk of bloodborne viruses, most of whom were already exposed to HCV. Lastly, these results might simply reflect the natural history of injecting drug use (e.g., a move away from being injected by others with longer use) and have nothing to do with changes in the environment (such as educational campaigns). If so, cross-sectional studies would find a higher prevalence of risk behaviour among younger IDUs. However, such a difference was not apparent in a major national cross-sectional study of 812 IDUs.9 If the decline in risk behaviour is real, then it may indicate that information campaigns about modes of HCV transmission are having an effect among Victorian IDUs. As there is evidence that reduced needle-sharing reduces HCV transmission,10 this change in behaviour may be responsible for a decline in HCV incidence. Other studies of Australian IDUs have also found a high incidence of HCV infection, but our study is the first to note a fall in incidence.11 However, a decline in first-test HCV prevalence was seen among IDUs at a major methadone maintenance clinic in Melbourne.12 We have documented a phenomenon which needs further investigation -- new HCV infections in IDUs who report no needle-sharing. This raises the possibility that infection is being spread in other ways. Several key risk behaviours for transmission of bloodborne viruses among IDUs have been documented.13 Sharing needles and/or syringes is thought to offer the greatest potential for transmission of HBV and HCV because of the relatively large volumes of blood which can be exchanged. However, equipment such as mixing spoons and filters, as well as rinsing water and the environment (such as surfaces and hands), can also become contaminated and are potential vehicles for transmission. Alternatively, respondents may have unwittingly shared needles or been reticent about disclosing sharing behaviour. If routes of infection other than needle-sharing are involved in HCV transmission, they are likely to be substantially less efficient, as HCV incidence among IDUs who reported sharing at half or more of their interviews was almost four times greater than among those who reported no sharing. These data confirm that HIV is not spreading among Australian IDUs (in our cohort, HIV prevalence significantly declined) and that rates of risk behaviours for HIV transmission among this group are low.14 This difference in epidemiology of HIV and HCV is probably due to the much higher prevalence of HCV among IDUs and the much smaller volume of blood necessary, on average, to transmit HCV compared with HIV. Nevertheless, we found that risk behaviours for HCV transmission are continuing and that HBV, which is vaccine-preventable, is spreading. While HCV transmission may already be decreasing among IDUs, our results suggest that further and sustained behaviour change is possible and necessary if the spread of HCV among Australian IDUs is to be controlled. The results also indicate yet again the failure of our current policies on hepatitis B vaccination and the need for vaccination targeted to IDUs and prisoners.15,16 |
Acknowledgements | The authors are grateful for the hard work of Jenny Kelsall, Michael Kerger, John Meade, Franz Hernberger, Vicky Hunt and the multitude of others associated with the Victorian Injecting Drug Users Cohort Study. We also gratefully acknowledge the support of the Victorian Health Promotion Foundation, the Drug and Alcohol Research and Education Advisory Council and the Commonwealth Department of Health and Family Services. Nick Crofts was supported by the Research Fund of the Macfarlane Burnet Centre for Medical Research. |
References |
(Received 10 Jan, accepted 16 May, 1997) |
Reprints: Dr N Crofts, Epidemiology and Social Research
Unit, Macfarlane Burnet Centre for Medical Research, PO Box 254,
Fairfield, VIC 3078.
E-mail: crofts @ mbcmr.unimelb.edu.au
<URL: http://www.mja.com.au/>
© 1997 Medical Journal of Australia.
We appreciate
your comments.
