Sharing of Drug Preparation Equipment as a Risk Factor for Hepatitis C

American Journal of Public Health 43

January 2001, Vol. 91, No.1

MD Holly Hagan, PhD, Hanne Thiede, DVM, MPH, Noel S. Weiss, MD, Dr PH, Sharon G. Hopkins, DVM; MPH, Jeffrey s. Duchin, MD,and E. Russell Alexande1; MD Holly Hagan, PhD, Hanne Thiede, DVM, MPH, Noel S. Weiss, MD, DrPH, Sharon G. Hopkins, DVM; MPH, Jeffrey s. Duchin, MD,

Holly Hagan, Hanne Thiede, Sharon G. Hopkins, and Jeffrey S. Duchin are with the Seattle-King County Department ofPublic Health, Seattle, Wash. Holly Hagan, Hanne Thiede, Noel S. Weiss, Sharon G. Hopkins, and E. Russell Alexander are with the Department of Epidemiology and Jeffrey S. Duchin is with the Division of Allergy and Infectious Dis- eases, University ofWashington Seattle.

Requests for reprints should be sent to Holly Hagan, PhD, {>ublic Health-Seattle and King County, 106 Prefontaine Place South, Seattle, WA 98104 (e-mail: holly.hagan@metrokc.gov).

This brief was accepted May 15,2000

Abstract

            Objectives: This study investigated the sharing of drug preparation equipment as a possible route of hepatitis C virus (Hepatitis C Virus) transmission.

            Methods: Hepatitis C Virus seroconversion was measured in a cohort of 317 injection drug users who tested negative for Hepatitis C Virus antibody at recruitment.

            Results: Cumulative Hepatitis C Virus incidence was 16.7% per year.  Among those who did not share syringes, Hepatitis C Virus seroconversion was associated with sharing drug cookers and filtration cotton (adjusted risk ratio = 5.9%; 95% confidence interval = 1.1, 31.7); 54% of Hepatitis C Virus infections in injection drug users who did not share syringes were attributable to cooker/cotton sharing.

            Conclusion: Among injection drug users who do not share syringes, an important proportion of Hepatitis C Virus infections may be attributed to cooker/cotton sharing.  (Am J Public Health. 2001; 91: 42-46)

Hepatitis C virus (Hepatitis C Virus) infection is very common among injection drug users. Studies of injection drug users in regions with a long- standing pattern of endemic injection drug use have reported prevalences of Hepatitis C Virus antibody in the range of 65% to 90%, even where HIV prevalence is quite low. 1-5 The majority of Hepatitis C Virus infections become chronic, resulting in a large reservoir of Hepatitis C Virus infection among injection drug users.6 Incidence of Hepatitis C Virus infection in previously uninfected injection drug users ranges from 10 to 30 per 100 person-years at risk.I-4.7-9 Prior studies have not observed a consistent pattern of risk associated with sex, duration or recency of injecting, or participation in a syringe exchange program.3.4.7.10.11 Three incidence studies noted an association between Hepatitis C Virus seroconversion and use of potentially contaminated syringes,3.8.9 but in all but ones the association was weak after control for other factors. Clearly, a great deal remains to be learned about the determinants and prevention of Hepatitis C Virus infections in injection drug users.

Among the unexplored areas of Hepatitis C Virus research is the importance to transmission of the shared use of drug preparation equipment in the absence of injection with a contaminated syringe. Because the prevalence of syringe sharing has declined since the appearance of HIV/AIDS,12 viral transmission associated with sharing equipment used to prepare drugs for injection may have risen in importance. In this study, we examined the risk of Hepatitis C Virus transmission in relation to the sharing of cookers used to melt the drug into an injectable liquid, of cotton used to filter out particles as the drug is drawn into the syringe, and of water used to rinse the syringe. Injection with a syringe previously used by another injector, and use of a syringe to divide drug doses between users (backloading},13 were also evaluated as potential means of Hepatitis C Virus transmission.

Methods

Subjects were enrolled in a cohort study of health consequences and risk behaviors associated with injection drug use (the RAVEN Study)! Recruitment occurred between June 1994 and May 1997 in 9 different locations in the Seattle area; a random-numbers table was used to select the nth client from among those present at each location. Eligibility criteria included having injected an illicit drug in the previous year, being English or Spanish speaking, being 14 years or older, and not being already enrolled in the study. After providing written consent to participate, subjects completed an interviewer-administered questionnaire and a blood draw. Follow-up I year later included an interview that asked about behavior during the follow-up period and a second blood specimen. Blood was tested for Hepatitis C Virus antibody with a third-generation enzyme immunoassay (Abbott Laboratories, Chicago, Ill). Hepatitis C Virus seroconversion was determined by the appearance of Hepatitis C Virus antibody in a previously seronegative individual. Details of RAVEN methods are available elsewhere! Study protocols were approved by an institutional review board.

Baseline characteristics and risk behaviors during the follow-up period were evaluated as risk factors for Hepatitis C Virus seroconversion by cumulative incidence. 14 The questionnaire asked about injection with a syringe previously used by another person, use of a cooker or cotton after another person had used it, use of rinse water another person had used, and use of a syringe to divide drugs. Throughout this re- port, these behaviors are referred to as syringe sharing, cooker and cotton sharing, sharing rinse water, and backloading, respectively. The magnitude of the association between these injection behaviors and Hepatitis C Virus infection was estimated by means of the risk ratio and its 95% confidence interval. Shared use of cooker and of cotton were evaluated together, because among those who did not share syringes, all of the Hepatitis C Virus seroconverters who shared cotton also shared a cooker.

Stratified analysis was carried out to separate the large effect of syringe sharing from the effect of other injection behavior. Logistic regression was used to estimate the risk of Hepatitis C Virus infection associated with injection risk behavior among those who did not report sharing syringes during the follow-up period; this analysis obtained estimates of the association adjusted for confounding factors (adjusted risk ratio). Frequency of injection at the baseline interview was included in the model because it met databased criteria for confounding. 15 The attributable risk percentage is the proportion of disease in persons with a given behavior that is due to that behavior. 16 The attributable risk percentage was estimated by standard methodsi6 and as shown in Table 3. It was calculated for behaviors that were statistically significantly associated with Hepatitis C Virus seroconversion, since some fraction of disease could reason- ably be attributed to them. The attributable risk percentage in the underlying injection drug user population (population attributable risk percentage) was also calculated by standard methodology. 16

Results

In a multiple logistic regression model that included only 123 injection drug users who injected but did not share syringes during follow-up, the adjusted risk ratio for cooker and cotton sharing was 5.9 (95% CI= 1.1,31.7). For injection drug users who shared rinse water or backloaded, risk of Hepatitis C Virus seroconversion was not significantly to Hepatitis C Virus positive during the follow-up period ( 16.7% ). The risk of Hepatitis C Virus seroconversion was lower for African Americans than for others, but it was not associated with sex, age, or homelessness. Injection drug users whose first injection was in the previous 2 years, and those who reported at their enrollment interview that they injected at least once every day, were relatively more likely to become infected with Hepatitis C Virus

Also shown in Table I, those who re- ported no injections during follow-up had the lowest risk of Hepatitis C Virus seroconversion. All 3 seroconverters who did not inject during follow-up reported injecting in the month before enrollment, and only I of the 3 did not report any injection or sexual risk behavior during that period. Any syringe sharing was associated with a 3-fold higher risk of Hepatitis C Virus infection (29% vs 10%; risk ratio [RR] = 2.94; 95% confidence interval [CI] = 1.6, 5.3). Similarly, risk of Hepatitis C Virus was 3-fold higher among those who shared a cooker or cotton (24% vs 7%; RR = 3.39; 95% CI = 1.4, 8.2). Risk of Hepatitis C Virus seroconversion did not differ according to the frequency of sharing a cooker and cotton, but it was relatively higher among those who reported sharing syringes sometimes, usually, or always than among those who reported rare syringe sharing (X2 test for trend, P<.0I). Among those who injected with a used syringe, there was no difference in Hepatitis C Virus incidence between those who almost always used disinfectant bleach when they shared a syringe and those who never or inconsistently used bleach. None of the sexual risk behaviors analyzed were associated with Hepatitis C Virus acquisition; these included number of partners, unprotected anal or vaginal sex, sex with another injection drug user, and exchanging money or drugs for sex.

Results of the analysis of Hepatitis C Virus seroconversion stratified by syringe sharing are shown in Table 2. Among individuals who reported having injected at least once with a used syringe, sharing both a cooker and cotton was associated with a risk ratio of 1.7 , sharing rinse water with a risk ratio of 0.9, and backloading with a risk ratio of 2.1 (all nonsignificant). Among those who did not inject with a used syringe, the risk ratio for cooker and cotton sharing was 3.8 (95% CI= 1.1,13.8), but sharing rinse water and backIoading were not associated with much, if any, increased risk of seroconverslon.

In a multiple logistic regression model that included only 123 injection drug users who injected but did not share syringes during follow-up, the adjusted risk ratio for cooker and cotton sharing was 5.9 (95% CI= 1.1,31.7). For injection drug users who shared rinse water or backloaded, risk of Hepatitis C Virus seroconversion was not significantly different than for those who did not share any equipment (for sharing rinse water, adjusted risk ratio (APR) = 0.6, (95% CI = 0.1, 2.7); for backloading, APR = 1.1, (95% CI = 0.2, 4.7)

The attributable risk percentage for syringe sharing was 66% (Table 3). Since 77% of Hepatitis C Virus seroconverters had shared syringes, the proportion of all Hepatitis C Virus infections in the underlying population of injection drug users that was attributable to syringe sharing (the population attributable risk percentage) was 51 %. Among those who did not share syringes but shared cookers and cotton, 74% of infections were due to cooker and cotton sharing. Because relatively few seroconverters shared cookers and cotton but not syringes (17%), the population attributable risk percentage for cooker and cotton sharing for all injection drug users was only 13%. However, in the underlying population of injection drug users who did not share syringes, the population attributable risk percentage for sharing cookers and cotton was substantially higher, 54%.

Discussion

The potential for blood-borne viral transmission via injection equipment other than syringes was reported in an earlier study of equipment collected in a Miami shooting gallery, where HIV-l DNA was detected in rinses from cottons and cookers and in water used to clean paraphernalia and to dissolve drugs. 17 A sterile syringe may become contaminated when the tip of the needle is inserted into a contaminated cooker or when the drug is drawn up through contaminated filtration cotton. This type of injection risk behavior appears to be quite common, and fewer injection drug users may recognize the hazard of sharing drug preparation equipment than recognize the hazard of sharing syringes.lg The present study suggests that Hepatitis C Virus may be transmitted via the shared use of drug cookers and filtration cotton even without injection with a contaminated syringe.

To our knowledge, there have been no prior published reports of incident blood-brone infections in injection drug users who shared equipment without sharing syringes. An association between backloading and prevalent infection with HIV was reported in a New York cross-sectional study; however, the magnitude of the association was small (adjusted odds ratio= 1.65), and therefore a notable correlation between syringe sharing and HIV prevalence may have contributed to the observed re- lationship.19 Another cross-sectional study reported an association between frequency of backloading and HIV and Hepatitis C Virus seroprevalence but the odds ratios were not adjusted for

syringe sharing 2O In a Baltimore study of injection drug users aged 18 to 29 years, Hepatitis C Virus antibody seropositivity at study enrollment was associated with recent backloading and sharing of cookers, cotton, or rinse water, but not after adjustment for use of new syringes.9

Potential sources of bias in the study were examined. Study retention was not associated with the injection risk behavior we examined, so

Note.  Includes only those who injected during the follow-up period.  RR = risk ratio

CI = confidence interval