icon star paper   Hepatitis C Articles (HCV)  
Back grey_arrow_rt.gif
 
 
HCV Infection among Prisoners in the California State Correctional System: sexual transmission, women at risk
 
 
  Clinical Infectious Diseases July 15 2005;41:177-186
 
......Our finding of an extremely high prevalence of HCV infection among men and women entering prison is consistent with results of previous studies of HCV in incarcerated populations, and is significantly higher than estimated (18%) by the National Correctional Healthcare Association for inmates....
 
.... The prevalence of HCV infection ranges from 22% to 40% among incarcerated populations, but few studies have examined associated risk factors, such as IDU exposure, incarceration patterns, and other high-risk behaviors..... ....HCV infection is pervasive among the California prison population, including prisoners who are non-IDUs and women with high-risk sexual behavior..... non-IDU women who reported having sex with IDUs have elevated risk of HCV infection (Odds ratio 4.47), compared with that of men....
 
..... Our findings with respect to biological-sex differences relative to risk of HCV infection highlight the importance of considering the unique vulnerability of women (see Author Discussion below)...
 
authors: Rena K. Fox,1 Sue L. Currie,4 Jennifer Evans,2 Teresa L. Wright,4 Leslie Tobler,5 Bruce Phelps,6 Michael P. Busch,3,5 and Kimberly A. Page-Shafer2
 
1Division of General Internal Medicine and 2Center for AIDS Prevention Studies, Department of Medicine, and 3Department of Laboratory Medicine, University of California-San Francisco, 4Division of Gastroenterology, San Francisco Veterans Administration Medical Center, and 5Blood Systems Research Institute, San Francisco, and 6Chiron, Emeryville, California
 
ABSTRACT
Background. Incarcerated populations are at high risk for hepatitis C virus (HCV) infection, yet prisoners are not routinely screened or treated for HCV infection. Understanding the risk factors of HCV infection among prisoners could help improve HCV interventions.
 
Methods. Prevalence and risk of HCV infection among 469 prisoners entering California State correctional facilities were assessed using HCV antibody screening, HCV RNA measurement, and structured interviews. Multivariate logistic regression analysis was used to identify independent correlates of HCV infection.
 
Results. The prevalence of HCV infection was 34.3% overall (95% confidence interval [CI], 30%-38%) and was 65.7% among those with a history of injection drug use (IDU), compared with 10.2% among those with no history of IDU (odds ratio [OR], 17.24; 95% CI, 10.52-28.25). Significant differences in HCV antibody positivity were found in association with age at first detention but not with the nature of the crime. Independent correlates of HCV infection included age, history of IDU, cumulative time of incarceration, biological sex (OR for females subjects compared with males subjects, 0.35; 95% CI, 0.13-0.96), and a history of having sex with a male IDU (OR, 4.42; 95% CI, 1.46-13.37).
 
We identified significant differences in risk factors between male and female subjects-notably, that the risk of HCV infection was significantly elevated among female non-IDUs who reported having sexual partners with a history of IDU. Among non-IDUs, correlates of HCV infection included history of receipt of blood products and cumulative years of incarceration.
 
Conclusions. HCV infection is pervasive among the California prison population, including prisoners who are non-IDUs and women with high-risk sexual behavior. These results should promote consideration of routine HCV antibody screening and behavioral interventions among incarcerated men and women.
 
INTRODUCTION
A US population-based survey showed that 1.8% of the general population has been infected with hepatitis C virus (HCV) [1]. This study, however, omitted incarcerated persons, among whom the prevalence of HCV infection is projected to be much higher, because many of them originate from high-risk backgrounds and have a history of risk behaviors for bloodborne infection-in particular, injection drug use (IDU). Addressing the risk behaviors and medical needs of the incarcerated population is important because the frequency of incarceration is alarmingly high. In 2002, in the United States, >2 million people were in prison or jail (1 of every 146 persons), and >4.5 million people were on probation or parole. It is projected that 1 of every 20 individuals in the United States will serve time in jail or prison during their lifetime [2].
 
HCV infection is associated with exposures thought to be common among prisoners, but it has also been suggested that incarceration itself may pose a risk for infection [3, 4]. In 2 studies that involved veterans in San Francisco and Florida, a history of incarceration was an independent predictor of HCV infection [5, 6]. The prevalence of HCV infection ranges from 22% to 40% among incarcerated populations [7-11], but few studies have examined associated risk factors, such as IDU exposure, incarceration patterns, and other high-risk behaviors.
 
Assessment of the sources of risk for prisoners will facilitate decision-making about how to screen for HCV, prevent further spread of the disease, and provide appropriate care to inmates. It is estimated that IDU accounts for 60% of all HCV infections in the general population [3]; however, this proportion may be higher among prisoners, and few, if any, data exist about IDU practices within US prisons. Outside of the United States, studies report a high prevalence of IDU within prison facilities [11-13]. In Ireland, 43% of prisoners report a history of IDU; 20% of those reported initiating IDU in prison, and almost 50% reported IDU during imprisonment [11]. In England, 24% of prisoners report that they have injected drugs; 30% report having done so in prison [14]. Lastly, other exposures, such as traumatic blood exchange, tattoos, inhaled drug use, and sexual activity, have also been hypothesized to be associated with an increased risk for HCV infection among prisoners [10]. We conducted a study of HCV among male and female prisoners entering California correctional facilities and determined the prevalence and risk factors associated with HCV infection.
 
AUTHOR DISCUSSION
Our finding of an extremely high prevalence of HCV infection among men and women entering prison is consistent with results of previous studies of HCV in incarcerated populations [7-9, 11] and is significantly higher than estimated (18%) by the National Correctional Healthcare Association for inmates [16]. HCV is likely to be the most prevalent bloodborne infection in prisons in the United States, with more than one-third of inmates infected with HCV, which far outreaches estimates for HIV infection (2.5%) and chronic or past hepatitis B infection (2% and >20%, respectively) [7-9, 16, 17]. HCV infection in this population is largely attributable to a history of IDU, and the rate of self-reported history of IDU was higher (43%) among prisoners in this study than among those in other studies (3%-28%) [18-20] and was higher than the estimated rate among state (25%) and federal (14%) inmates [2]. Currently, there are >2 million persons in jail or prison, and another 4 million are under correctional supervision [2]. The epidemic level of HCV infection, combined with IDU among prisoners, is an expanding public health problem, because this population frequently cycles between correctional and community environments. Parolees may have little information about their infection, may be unwilling or unable to access screening and medical care, and may continue to engage in high-risk behaviors in the community, potentially increasing transmission of the virus. It is unlikely that there is a more clearly defined and readily accessible group for HCV education, prevention, diagnosis, and treatment than inmates in the correctional system.
 
Our findings with respect to biological-sex differences relative to risk of HCV infection highlight the importance of considering the unique vulnerability of women. HCV infection was especially prevalent among women who reported that their sexual partners were IDUs, particularly among women who themselves were non-IDUs. Although women are almost universally more susceptible to sexually transmitted infections, including HIV infection [21-23], it is well recognized that HCV is inefficiently transmitted sexually [24-27]. Thus, researchers have also focused on nonbiological factors, including behavioral and social factors, which might help explain differential incidence of HCV infection among men and women [28]. Our results support other studies that indicate that overlapping sexual and drug-injection partnerships may be one of the multiple areas of risk for bloodborne and sexually transmitted infections [28-31]. With respect to sexual partnerships, women with male IDUs as sex partners are at higher risk for HIV seroconversion [32], women who are initiated into IDU by a male sexual partner have higher HIV risk scores [33], and women who are IDUs and have male IDUs as sex partners are more likely to engage in the higher-risk practice of needle sharing [30, 31]. Women who have male IDUs as sex partners also have increased vulnerability to incarceration [34]. From prevention and clinical perspectives, providers should consider the risk factors not only for individual behaviors but also for the relevant behaviors in sexual partnerships [35, 36].
 
Among prisoners who were non-IDUs, the prevalence of HCV infection was 10.2%, which is 5-fold higher than that among the general US population. The association between HCV infection and cumulative incarceration time has been reported elsewhere [6]. Multiple incarcerations and length of incarceration may be associated with higher risk or may be proxies for risk-taking behavior, such as unreported IDU.
 
Our study has some limitations. The sample size is modest and is limited to California prison facilities; therefore, it may not be generalizable to other US prison populations. We noted, however, that our sample appeared to be very similar to the overall California State prison population with respect to demographic and crime-related data. We also recognize that sensitive and socially undesirable behaviors may have been underreported, although this would result in conservative risk associations. Specifically, for the significant strong associations with HCV infection that were found among non-IDUs, it is important to note 2 caveats: (1) IDU behavior may be significantly underreported in a population well aware of the illegal nature of drug use, and (2) almost all of the non-IDUs with HCV infection reported at least 1 other potential risk exposure other than sexual contact (surgery, blood transfusion, organ transplantation, tattoos, or cutting).
 
Considering these data, we support the development of policies for systematic HCV screening among all persons entering and within the corrections system. We recognize that prisons face serious challenges in how to address HCV disease in prisoners. However, detection will benefit not only inmates but also the public at large. If HCV testing were part of the admission health screening for prisoners, it could improve resource planning, education, and health care within corrections systems and for parolees reentering the community. Prisoners are a marginalized population in society, with little political influence, and the most appropriate health care interventions for incarcerated persons are not yet in place. Correctional settings provide important opportunities for HCV interventions, because there is a disproportionately high burden of disease and risk history. Although a large proportion of individuals may not medically be candidates for antiviral treatments, they are likely to benefit from HCV education and counseling. The effectiveness of such outreach has been shown previously elsewhere [37]. Because younger age at detention and longer incarceration time are associated with HCV infection, the targeting of incarcerated youths should also be considered. Furthermore, with high recidivism rates, there is also a clear need for coordination between correctional facilities and programs in the community, such as needle exchange programs, methadone maintenance programs, and alcohol and drug treatment facilities.
 
We support further research to examine antiviral treatment of HCV infection in prisoners and parolees. HCV can be eradicated in >50% of patients with the use of a combination of pegylated IFN and ribavirin for 1 year [38, 39]. No clinical trials have evaluated treatment models among inmates, although 1 observational study described inmates in Rhode Island who were treated with routine antiviral therapy and found that 63% had a virological response at month 6 of treatment [40]. Finally, correctional facilities may actually be a more appropriate and realistic setting for treatment of this high-risk population than the general community, where health care is fragmented and access to it is limited.
 
RESULTS
Prevalence. Of 615 individuals who were asked to participate, 472 (77%) consented. Of those 472 individuals, 468 completed interviews, and 467 completed blood testing. The discrepancy between the number of persons who consented and the number with complete data collection was a result of the need for study processes to have minimal impact on the ongoing flow of prison operations, especially prisoner transfers from reception to other units or prison facilities. Many inmates approached the study staff with an interest in participating but could not be accommodated. Results are presented from the 467 participants who participated in both the blood testing and the interview.
 
The prevalence of HCV antibody was 34.0% (95% CI, 30.0%-38.8%). Of the 160 HCV antibody-positive specimens, 159 were confirmed by HCV-RIBA. Of the 307 HCV antibody-negative participants, 1 was positive for HCV RNA, which possibly indicates a newly acquired infection detected prior to seroconversion. Therefore, the overall prevalence of HCV infection (past or current) was 34.3%, which included 159 confirmed HCV antibody-positive subjects and 1 HCV antibody-negative, RNA-positive subject.
 
Participant characteristics. The median age of the participants was 35 years (IQR, 27-41 years); 72.2% were male (recruited from RJ Donovan Prison and Wasco State Prison), and 27.3% were female (recruited from the California Institute for Women). The race/ethnicity distribution was white, 26.3%; African American, 22.2%; Hispanic, 37.2%; and "other," 14.3%. Our sample was similar in these characteristics to the overall California prison population, which has an average age of 36 years and is 29% white, 29% African American, and 36% Hispanic [15]. Almost all participants (96.4%) had been incarcerated previously, and 76.3% were currently being detained for a parole violation. Of those not being detained for parole violations (n = 117), the self-reported nature of the current charges included drug related, 32.5%; violent crime, 25.6%; and property and firearm-related crime, 41.8%. Almost half (43.2%) of inmates reported a history of IDU; of these, 33.2% reported daily IDU in the 6 months prior to incarceration.
 
Sociodemographic and behavioral variables and the risk of HCV infection.
 
In bivariate analyses, HCV infection was significantly associated with age, race/ethnicity, sexual orientation, and history of previous incarceration, but not with biological sex or prison site (data not shown). The prevalence of HCV infection was significantly elevated among participants who reported any history of IDU (OR, 17.24; 95% CI, 10.52-28.25) or blood transfusion. Among all inmates who reported any IDU history, the incidence of HCV infection was significantly higher among those who reported daily IDU and having shared injection equipment, as well as among those who reported a history of intranasal drug use with shared equipment, than among those who did not report these exposures. No significant associations were found between HCV infection and a history of tattooing, piercing, or being cut or assaulted. Table 1 also shows associations between self-reported sexual orientation, selected risk behaviors, and HCV. Data from men and women who identified themselves as gay or bisexual are combined, because only 6 (1.8%) of the men identified themselves as such. The odds of HCV infection were increased among those who reported same-sex sexual partners, compared with those who did not (50.0% vs. 32.3%; OR, 2.12; 95% CI, 1.22-3.70), with those who reported sex partners who were IDUs (OR, 5.88; 95% CI, 3.87-8.93), and with commercial sex workers (OR, 1.87; 95% CI, 1.20-2.91). Having been paid for sex, having been forced to have sex, or having an HIV-positive sex partner were not associated with HCV infection (data not shown).
 
Incarceration-related exposures and risk of HCV.
Table 2 shows incarceration-related variables and associations with HCV infection. Prevalence increased significantly with cumulative years of incarceration (P <01), age at first incarceration (P <.01), and a history of juvenile detention. No significant associations were observed between HCV infection and current imprisonment charges or a history of having been gassed.
 
Risk of HCV among non-IDUs compared with risks among IDUs:
To further examine potential nonparenteral risks for HCV infection, we examined associations between HCV infection and selected risk exposures among inmates stratified by history of IDU (table 3). Age, having had sex with an IDU, cumulative years of incarceration, and being currently detained for parole violation were associated with HCV infection among IDUs but not among non-IDUs. Among non-IDUs, the odds of being infected with HCV were significantly higher for those with a history of receipt of blood products. Among non-IDUs or IDUs, no associations were seen between HCV infection and history of tattooing, piercing, being cut or assaulted, or having been gassed.
 
Multivariate analyses.
Table 4 shows variables found to be independently associated with HCV infection among inmates. Because incidence rates of HCV infection increased with age and cumulative years of incarceration, these variables were entered into the model as continuous variables. Variables that were significantly and independently associated with HCV infection included history of IDU, cumulative time of incarceration, and biological sex. All of the variables entered into the multivariate model are represented in tables 1 and 2. We found that biological sex and a history of having sex with an IDU were significantly associated (P = .026) with HCV infection. Among the women who reported having sex with an IDU, 61% were infected with HCV, whereas 11% of women who did not report having sex with an IDU were infected. Among men who reported having sex with an IDU, 58% were HCV positive, compared with 22% of men who did not report having sex with an IDU. The final model shows that, after adjusting for cumulative years of incarceration and history of IDU, the odds of HCV infection among women who reported having sex with an IDU were >4 times higher than among women who reported not having had sex with an IDU. In contrast, the risk among men was not significantly different between those who had or had not had sex with an IDU. We also stratified between IDUs and non-IDUs and found that non-IDU women who reported having sex with IDUs have elevated risk of HCV infection, compared with that of men. In multivariate analyses conducted among non-IDUs, independent correlates of prevalent HCV infection included a history of receipt of blood products (OR, 2.51; 95% CI, 1.01-6.23) and cumulative time of incarceration (OR, 1.06; 95% CI, 1.00-1.12).
 
METHODS
Study population. This was a cross-sectional study of inmates entering 3 California State correctional facilities (2 for men and 1 for women). Of the 37 adult prison facilities in the California correctional system, 11 are also reception sites for inmates entering prison with sentences of ⩾2 years. The reception site for an inmate is determined by several factors, including the county where the crime was committed, the type of crime, facility availability, security, safety, and special needs. Persons eligible for this study included male and female inmates entering 1 of the 3 prisons, which were chosen on the basis of having a high volume of entering inmates and geographic diversity.
 
Data collection. Demographic, medical, behavioral, and incarceration-related information was collected in private, individual, structured interviews conducted by the University of California-San Francisco (UCSF) study staff. Medical information consisted of history of blood transfusion or receipt of blood products, abnormal liver test results, psychiatric illnesses, and prior testing for bloodborne pathogens, including HCV. Behavioral information included the duration, quantity, and frequency of the following: alcohol consumption; IDU and inhaled or snorted drug use (both with and without shared equipment); tattooing; body piercing; other blood exposures, including being assaulted, ritualistic cutting, acupuncture, "gassing" (i.e., having a collection of bodily fluids thrown between individuals as a method of assault), and history of use of shared toothbrush, razor, or clippers. Sexual behavior information consisted of the duration, quantity, and frequency of having sex with opposite- or same-sex partners, having transactional sex, having sex with IDUs, having nonconsensual sex, and having sex with partners who were known to have hepatitis or risk factors for bloodborne pathogens. Incarceration-related variables included total length of time incarcerated during lifetime, number of times incarcerated, facilities in which detained (e.g., youth detention, county jail, or state prison), age at first incarceration, and the type of criminal charges related to the current incarceration.
 
Pretest counseling was conducted by the UCSF study team prior to venipuncture for HCV testing. Medical staff at each prison facility performed blood draws. The processing and transportation of samples were handled by UCSF. Participants were offered the option of receiving their test results confidentially through the Department of Corrections medical system, with a waiver for release of medical information.
 
Laboratory methods. Serological and virological testing was conducted at the Blood Systems Research Institute (San Francisco, CA) and at Chiron Corporation (Emeryville, CA). Samples were tested for HCV antibody by a third-generation EIA (EIA-3; Ortho Clinical Diagnostics). HCV antibody-positive samples were confirmed using an HCV recombinant immunoblot assay (HCV-RIBA), the RIBA 3.0 Strip Immunoblot Assay (Chiron). All HCV antibody-negative samples were tested for HCV RNA by use of qualitative transcription-mediated amplification (TMA) testing (Procleix HIV-1/HCV Assay; Gen-Probe). HCV infection was classified as either (1) HCV antibody positive, confirmed with a positive HCV-RIBA finding, or (2) HCV antibody negative but positive for HCV RNA. Subjects who were HCV antibody negative but had positive TMA findings were considered to be recently infected (an incident infection) and to be in the window of time prior to seroconversion. Those who were positive for HCV antibody but had negative or indeterminate RIBA findings were considered to be HCV negative.
 
Statistical analyses. Prevalence rates and 95% CIs were calculated on the basis of binomial distributions. Summary statistics included frequency tables, for categorical variables, and medians and interquartile ranges (IQRs), for continuous variables. Associations between HCV antibody status and demographic, risk exposure-related, and incarceration-related variables were conducted using the 2 test of association (or Fisher's exact test, in cases for which the expected cell size was <5 events), and ORs and 95% CIs were calculated. Independent associations between the various exposures and HCV antibody were assessed using unconditional logistic regression analyses. Variables that were statistically significant (P < .05) on bivariate analysis, that were potential confounders (e.g., age), or that were of a priori interest (e.g., biological sex) were entered into a backward stepwise logistic regression, with a significance level for removal of .10. A final model was created that contained main effects and interactions with P values of ⩽.05. Associations in subgroups of interest-for instance, non-IDUs-were conducted using stratified analyses.
 
Human subjects approvals. The study protocol, informed consent, and materials were reviewed and approved by the UCSF Institutional Review Board, the California Department of Corrections Research Advisory Committee, and the California Youth Authority Research Project Approval Committee. A Federal Certificate of Confidentiality was also obtained for this research.
 
 
 
 
  icon paper stack View Older Articles   Back to Top   www.natap.org