CDC Study on HCV Risk Transmission & Tattoos/Body Piercing, Snorting Drugs
"Relationship of cosmetic procedures and drug use to hepatitis C and hepatitis B virus infections in a low-risk population"
Hepatology August 2006
Volume 44, Issue 2, Pages 341-351
Lu-Yu Hwang 1 * , Jennifer R. Kramer 1 3, Catherine Troisi 1, Lara Bull 1, Carolyn Z. Grimes 1, Rob Lyerla 2, Miriam J. Alter 2
1Center for Infectious Diseases, Division of Epidemiology and Disease Control, University of Texas-Health Science Center at Houston, School of Public Health, Houston, TX
2Division of Viral Hepatitis, National Center for Infectious Diseases, Centers for Disease Control and Prevention, Atlanta, GA
3Section of Health Services Research, Department of Medicine, Baylor College of Medicine, Houston, TX
We conducted an anonymous cross-sectional seroprevalence study of a population with a low frequency of injection drug use to determine whether persons with a history of cosmetic procedures, such as tattooing and body piercing, or intranasal drug use were at increased risk for hepatitis C virus (HCV) or hepatitis B virus (HBV) infection.
Students 18 years and older from eight college campuses in Houston, Texas, were invited to participate in the study. Of the 7,960 who completed a self-administered questionnaire and provided a blood sample, 5,282 U.S.- or Canadian-born participants were analyzed.
Their median age was 21, 62% were female, 42% were white, 26% black, 22% Hispanic, and 10% Asian or other.
Two percent reported injection drug use, 13.7% intranasal drug use, 21.2% body piercings, and 25.2% tattoos.
The overall prevalence of HCV infection was 0.9% and of HBV infection was 5.2%.
Higher HCV prevalence was independently associated with increasing age (odds ratio [OR] per year = 1.11; 95% confidence interval [CI] = 1.08-1.14),
history of injection drug use (OR = 18.24; 95% CI = 7.74-42.92), blood transfusion before 1991 (OR = 3.21; 95% CI = 1.02-10.12), and incarceration (OR = 3.48; 95% CI = 1.45-8.37).
Among 5,066 students who denied injecting drugs, HCV prevalence was 0.8% in those who reported intranasal drug use and 0.6% each in those who reported tattoos and those who reported body piercing.
Increased HBV prevalence was associated with high-risk sexual behaviors and black or Asian race.
In conclusion, there was no increased risk for HCV or HBV infection in low-risk adults based solely on history of cosmetic procedures or snorting drugs. However, proper infection control practices for cosmetic procedures should be followed, illegal drug use discouraged, and hepatitis B vaccination provided to adolescents and sexually active adults.
Of the more than 3 million persons chronically infected with hepatitis C virus (HCV) in the United States, most are younger than 50 years old and unaware of their infection. Because most infections are asymptomatic until advanced liver disease develops, diagnostic testing is necessary to identify infected persons at a sufficiently early stage for them to benefit from medical evaluation, treatment, counseling to reduce alcohol intake, and other interventions. The identification of HCV-infected persons is most efficient when screening tests are applied to populations most likely to be infected.
The determination of which persons at risk to recommend for routine testing is made on the basis of risk factors demonstrated to be temporally related to acquiring infection, a relatively high prevalence of infection among those with the risk factor, and the need for persons with a recognized exposure to be evaluated for infection. In addition to HCV testing being recommended for persons with certain medical conditions (for example, hemophilia or on chronic hemodialysis), it also is routinely recommended for all persons who ever injected illegal drugs, who received blood transfusions or organ transplants before donor screening, and who require evaluation following a percutaneous (e.g., contaminated needlestick) or mucosal (infant born to infected mother) exposure to infected blood.
Because HCV is a blood-borne virus, persons who received cosmetic procedures such as tattoos might be at an increased risk of acquiring infection. However, the results of two case-control studies of acute disease showed no association between receipt of tattoos and acquiring HCV infection, and the results of cross-sectional studies of selected groups have been inconsistent regarding possible associations with this procedure as well as practices such as intranasal drug use.[5-8] These inconsistent results and the fact that some of the study populations had relatively high frequencies of injection drug use (IDU) raise the concern that the association of HCV with behaviors such as noninjection illegal drug use and tattooing may be spurious because they often coexist with IDU.
To determine if cosmetic procedures, noninjection drug use, and other practices are associated with an increased prevalence of HCV (and hepatitis B virus [HBV]) infection, we conducted a cross-sectional seroprevalence study among college students. Published data indicated that tattooing and body piercing were popular among college students,[14-16] with a low frequency of IDU. This study was conducted in Houston, Texas, the fourth-largest city in the United States, with a population of approximately 3.5 million. There are more than 200,000 college students attending eight institutions, including two-year community colleges, public four-year colleges, and private universities.
Prevalence of HCV and HBV Infection.
A total of 7,960 students participated in the study, of whom two-thirds (n = 5,282) were U.S.- or Canadian-born. Of those not U.S.- or Canadian-born, 42% were born in Asia/South Pacific and 32% in Central or South America. The median age was 21 years, and 60.8% were female. Overall prevalence of HCV infection was 0.9%, of whom 85.5% were HCV RNA positive (Table 1). Overall prevalence of HBV infection was 9.5%, of whom 12.8% were HBsAg positive. Evidence of hepatitis B immunization was found in 17.5%. Twenty-two (0.3%) students were positive for both anti-HCV and anti-HBc; no one was positive for both anti-HCV (or HCV RNA) and HBsAg. There was no difference in HCV prevalence by place of birth; however, both previous and current HBV infection was significantly more prevalent among the non-U.S./Canadian-born students than among the U.S./Canadian-born students (Table 1).
Demographic and Risk Characteristics of the Study Population.
Because the epidemiology of HCV and HBV can differ by country of birth, and lifetime histories of potential exposures in other countries are difficult to measure, only U.S./Canadian-born students were included in further analyses. Their median age was 21 years (range 18-70), and 62.3% were female (Table 2). By racial ethnic group, 41.5% were non-Hispanic white, 26.1% non-Hispanic black, 22.2% Hispanic, 6.8% Asian, and 3.5% other. Many risk factors known to be associated with HCV or HBV infection were reported by fewer than 10% of the participants, including IDU (2.0%), blood transfusion before donor screening for HCV (2.3%), being a man who had sex with men (2.9%), and frequent health care-related occupational blood exposure (6.0%).
Higher proportions of participants reported high-risk sexual behaviors, including first having sex before age 15 years (13.4%) and having multiple lifetime heterosexual partners (44.2% reported 2-9 partners, 15.0% 10-24 partners, 3.5% 25-49 partners, and 1.2% >50 partners). A relatively high proportion of participants also reported a history of intranasal drug use (13.7%), body piercing (21.2%), tattooing (25.2%), or professional manicures (54.3%) (Table 2). Sharing toothbrushes or razors was reported by 41.5% and 42.9%, respectively (data not shown).
Risk Factors for HCV Infection.
Among the 5,282 U.S.- or Canadian-born students with HCV test results, the results of univariate analysis found the prevalence of HCV infection significantly increased with age older than 30 years and significantly higher among non-Hispanic whites and students attending two-year colleges (Table 2). A significantly higher prevalence of HCV infection also was found among those with a history of blood transfusion before 1991, hemodialysis, household exposure to an infected contact, sexual exposure to an infected partner, being a man who had sex with men, 50 or more lifetime heterosexual partners, age less than 15 years at first intercourse, sex with an injection drug user, being diagnosed with a sexually transmitted disease, illicit injecting and intranasal drug use, and incarceration (Table 2).
No differences in HCV prevalence were found by history of sharing razors or toothbrushes (data not shown), tattooing, body piercing, or professional manicures (Table 2). However, there was a higher prevalence of HCV infection among those who had received three or more tattoos, received a tattoo at least once in a nonprofessional setting (e.g., home, party, correctional facility), or received one or more tattoos with reused needles that had not been autoclaved. There was no relationship between HCV prevalence and number of body piercings, setting where piercing performed, or type of disinfection methods used. There also was no difference in prevalence by number of professional manicures received in the previous 12 months or if bleeding occurred (data not shown).
The results of the final multivariable model found that an increased prevalence of HCV infection was independently associated with increasing age and histories of IDU, blood transfusion, and incarceration (Table 2). To further examine potential associations between HCV infection and tattooing and body piercing, number of body piercings, number of tattoos, setting where tattooed, and disinfection method for tattoos were entered into separate stepwise logistic regression models, each of which contained the covariates found to be significant from the overall model. None of these subgroup categories were significant in these separate models.
The primary reason most of the variables found to be significant in the univariate analysis were not independently associated with higher prevalence of HCV infection was their relationship with IDU. Among the 5,066 students who reported never injecting drugs, frequency of intranasal drug use (12%) or receiving at least one tattoo (24.8%) or body piercing (21.0%) was relatively high; however, prevalence of HCV infection was low: 0.8% in those who reported intranasal drug use, 0.6% each in those who reported at least one tattoo or body piercing, and 0.4% in those who reported three or more tattoos (Table 3). In contrast, among the 102 students who reported a history of IDU, the frequencies of intranasal drug use, incarceration, and multiple and nonprofessional tattoos were substantially higher as was the prevalence of HCV infection in all categories (Table 3).
Of the 102 students who reported a history of IDU, 31% described themselves as daily or regular (used commonly for a long period) users during the time they were using and 54% had shared needles or other works. HCV prevalence among those who reported use that was daily, regular, or sporadic (tried it more than once or twice, but not long term) was more than three-fold higher than among those who reported experimental use (i.e., trying it once or twice), 29.0% versus 9.0%, P < .05, although even experimental use was associated with significantly higher HCV prevalence than never injecting (0.5%, P < .001). Furthermore, HCV prevalence increased sharply with duration of injecting; it was 12.0% among those who reported they had injected for at most 1 year, 35.0% among those who had injected for more than 1 to 4 years, and 38.5% among those who had injected for at least 5 years (P < .001, chi square for trend).
Risk Factors for HBV Infection.
The analysis of risk factors for HBV infection included the 4,328 U.S./Canadian-born students who had no serologic evidence of hepatitis B immunization (Table 4). The results of univariate analysis found the prevalence of HBV infection to be significantly lower among males, to significantly increase with age among persons aged 30-59 years, and to be significantly higher among non-Hispanic blacks, Asians, and students attending two-year colleges (Table 4). Significantly higher prevalence of HBV infection also was found among those with a history of blood transfusion before 1991, hemodialysis, frequent health-care-related occupational exposure to blood, household exposure to an infected contact, sexual exposure to an infected partner, being a man who had sex with men, 50 or more lifetime heterosexual partners, age less than 15 years at first intercourse, sex with an injection drug user, being diagnosed with a sexually transmitted disease, injection and intranasal drug use, and incarceration (Table 4).
No differences in HBV prevalence were found by history of sharing razors or toothbrushes (data not shown) or by history of any of the categories for tattoos or body piercings except for disinfection method for tattoo needles (Table 4). Prevalence was significantly higher among those who reported having received tattoos with reused needles that had not been autoclaved. By multivariable analysis, the factors that remained independently associated with increased prevalence of HBV infection were increasing age per year, black or Asian race, attending a 2-year college, and history of being a man having sex with men, 5 or more heterosexual partners in the previous 4 months, and being diagnosed with a sexually transmitted disease. The one subgroup variable for tattooing with a P value of <.10 on univariate analysis (disinfection method) was entered into the final model separately with all other significant covariates. It became nonsignificant while controlling for the other significant risk factors.
Because perinatal and early childhood infections are known to contribute to the high prevalence of HBV infection among Asians, and about 20% of HBV-infected persons in our study were Asian, we analyzed students of non-Asian and Asian races separately. The risk factors for HBV infection among non-Asian students remained similar to those of the study population as a whole. In contrast, among Asians, the two factors associated with an increased prevalence of HBV infection were increasing age and household exposure to an infected contact.
The results of this study demonstrate that the prevalence of HCV infection is low (<1%) among young adults in North America who have been tattooed, have had their bodies pierced, or have snorted illicit drugs. These findings are significant because they indicate routine HCV testing is not warranted in young, low-risk adults based solely on history of these procedures and practices. Any study whose sample population is self-selected, as this was, and has low participation (around 10% of the colleges' enrollment) leaves substantial room for bias. Nonetheless, the distribution of gender and racial groups in our self-selected study subjects approximates that of the student population of the eight colleges we chose for study, and we are unaware of bias in any particular direction. College students per se are a highly selected population, but we believe the results from this study are likely to be generalizable to other similarly aged populations that have a low frequency of IDU. The inclusion of students from community colleges and public and private universities resulted in a mix of sociodemographic groups, and the frequencies of cosmetic procedures and intranasal drug use were high, whereas few had injected drugs. Thus, there was ample opportunity to evaluate these other potential exposures with little likelihood of confounding by unacknowledged IDU. In addition, the pattern of age-specific prevalence and the exposures found to be associated with an increased prevalence of HCV and HBV infection in this population are consistent with the known epidemiology of these diseases.
The results of previously published studies that evaluated the contribution of intranasal drug use and cosmetic procedures to the transmission of HCV in the United States have been inconsistent. Case-control studies of acute non-A, non-B hepatitis (now known as hepatitis C) and acute hepatitis B during 1979 to 1985 found no association between endemic cases of these diseases and having had procedures such as tattooing, ear piercing, or acupuncture during the incubation period. In these studies such procedures were reported by fewer than 1% of cases and controls, and ongoing surveillance for cases of acute hepatitis C and acute hepatitis B indicate this proportion has not increased over time. Of the cross-sectional studies among volunteer blood donors and patient groups selected from various settings, an association between history of intranasal cocaine use and HCV was found in only one study and an association between having tattoos and HCV in three studies. Similar inconsistencies were found for studies that evaluated body piercing, most of which did not distinguish ear piercing from piercing of other body parts.
The reasons for these contradictory results are not clear, but results that have not been consistently replicated are not likely to be generalizable to other populations and should be viewed with caution. In our study, the univariate associations between increased prevalence of HCV infection with intranasal drug use, tattooing, and body piercing were confounded by IDU and did not independently contribute to any additional increase in infection. In contrast, studies of drug users recruited by street outreach programs found that sharing of implements for intranasal drug use or receipt of tattoos while incarcerated contributed to an increase in HCV and HBV infection.
There were several limitations to this study. The low prevalence of some variables along with the low prevalence of HCV infection may have resulted in reduced power to detect an association. Any percutaneous exposure has the potential to transfer infectious blood and transmit blood-borne pathogens. If such episodes are rare or only occur in certain settings (for example, in correctional facilities), associations with sporadic or background cases would be difficult to demonstrate. However, outbreaks of HBV infection related to individual tattoo and acupuncture establishments have been reported. Furthermore, the results of this study may not apply to populations in which HCV or HBV infection rates are high, such as among those who are injection drug users or those who are older and whose exposures to these procedures or practices occurred more than 20 years ago. Additional studies will be necessary to determine if these results are applicable to older-aged populations.
IDU is the primary risk factor for HCV infection in the United States, especially among persons less than 50 years old. Even in this population, which was at relatively low risk for IDU, almost half the HCV-positive individuals reported a history of IDU. In addition, this study showed that any history of injecting drugs, no matter how experimental or sporadic, was associated with an increased prevalence of HCV infection. IDU also is an important mode of transmission for HBV infection, but sexual activity is the primary risk factor for acquiring HBV infection among adolescents and adults in the United States. In this study population, men who had sex with men, persons with a recent history of multiple heterosexual partners, and those with a history of an STD had significantly higher prevalence of HBV infection. For more than 20 years it has been recommended that such persons be vaccinated for hepatitis B, but few in these risk categories have been vaccinated. It is distressing that fewer than 20% of this study's participants had any evidence of having received hepatitis B vaccination. Routine vaccination of children and adolescents will ultimately provide protection from HBV infection as these cohorts age and undertake risky behaviors, but until that occurs, a substantial number of teenagers and young adults remain susceptible as they enter the years when they are at highest risk.
In conclusion, the prevalence of HCV infection is low among young adults with no history of IDU, and routine HCV testing is not indicated based solely on having a history of cosmetic procedures such as tattooing or practices such as snorting drugs. Physicians who ascertain blood transfusion, IDU, and sexual histories from their patients will be able to identify those most likely to benefit from HCV testing and hepatitis B vaccination. Regardless of transmission risk, however, any illegal drug use should be discouraged, counseling and referral to appropriate community services should be provided to those with high-risk behaviors, and proper infection control procedures should always be followed when cosmetic or other procedures involving percutaneous exposures are performed.
Patients and Methods
To confirm that these college students were an appropriate population to achieve the study objectives, an anonymous pilot survey of selected behaviors was conducted in April 1999 among students in general curriculum courses on four campuses in the Houston metropolitan area. Of the 1,076 responders, 2% reported a history of IDU, 9% intranasal cocaine use, 14% body piercing, and 18% tattooing.
Eligible participants were students at least 18 years of age enrolled between February 2000 and May 2001 at eight campuses of two- and four-year institutions of higher education in Houston and Austin, Texas. Subjects were recruited from large classes of required courses to ensure representation from the entire student body. Ten minutes of class time was used to promote the project for recruitment. Additional subjects were recruited from booths set up in public areas (lobbies, hallways, outside buildings). After giving written informed consent, students who agreed to participate completed an anonymous, self-administered paper-and-pencil questionnaire and provided anonymous blood samples. This study was approved by the Institutional Review Boards at the University of Texas Health Science Center at Houston and the Centers for Disease Control and Prevention.
Data collected on the questionnaire included demographics, medical history, and history of known and potential risk factors for HCV and HBV, such as blood transfusion (before donor screening), health care-related occupational exposure to blood, frequency and duration of injection and noninjection illegal drug use, sexual behaviors, incarceration, tattooing, body piercing (excluding ear piercing), and professional manicures. For tattooing and body piercing, data on number received, where performed, and sanitation procedures used were collected. Participants received information cards with their unique study numbers, which allowed them to anonymously telephone for their test results. Those who were confirmed positive for HCV or hepatitis B surface antigen (HBsAg) were referred for counseling and further evaluation. All HCV-positive subjects were given information about local support groups. Subjects who tested negative for HBV serologic markers were recommended to receive hepatitis B vaccination.
Sample Size and Power Estimate.
Sample size for this cross-sectional survey was determined using results from the pilot study. On the basis of the frequencies of potential exposures reported, an expected overall HCV infection prevalence of 2%, an alpha level of 0.05, and a power of 80% to detect at least a twofold difference, it was calculated that a sample size of 5,000 was needed. However, lower-than-expected HCV prevalence was observed among the first 5,000 students recruited; therefore, the sample size was increased to 8,000 to maintain statistical power.
All serum specimens were tested for antibody to hepatitis C virus (anti-HCV) using an enzyme immunoassay (HCV EIA 2.0, Abbott Laboratories, Abbott Park, IL). Repeatedly reactive specimens were tested by a supplemental strip recombinant immunoblot assay (Chiron RIBA HCV 3.0 SIA, Chiron Corp., Emeryville, CA) for anti-HCV and reverse transcriptase polymerase chain reaction amplification (AMPLICOR Hepatitis C Virus [HCV] Test, version 2.0, Roche Molecular Systems, Branchburg, NJ) for qualitative detection of HCV RNA. Only samples confirmed positive by RIBA or Amplicor were considered HCV positive.
Specimens were also tested for antibody to hepatitis B core antigen (anti-HBc), antibody to HBsAg (anti-HBs), and HBsAg by commercially available immunoassays (Abbott Laboratories, Abbott Park, IL). Evidence of past HBV infection was defined as anti-HBc positive with or without anti-HBs, current infection as HBsAg (and anti-HBc) positive, and immunization as only anti-HBs positive.
Data were entered into a Microsoft Access database and were analyzed by SAS (Cary, NC) and STATA (College Station, TX) statistical software. Overall prevalence of HCV and of HBV infection was calculated for the total population and by place of birth (i.e., U.S./Canadian-born or non-U.S./Canadian-born). The analysis of risk factors was limited to the U.S.- or Canadian-born students because the epidemiology of HCV and HBV, including risk factors for transmission and their relative contributions to disease burden, differs substantially by geographic region. For example, receipt of unscreened blood, unsafe therapeutic injections, and other health care-related procedures are major sources of HCV infection in many developing countries, whereas IDU is the most common source in North America. Combining persons from countries with different transmission patterns in the analysis might obscure associations for the variables under investigation in this study. For HBV, risk factor analyses also excluded students with serologic evidence of HBV immunization in order to examine the population at risk for HBV infection.
For univariate analyses, the prevalence ratio (PR) was used to measure the strength of the association, and 95% confidence intervals were calculated. When appropriate, chi-square and Fisher's exact tests were used to determine significance. Statistical significance was set at P < .05. All variables with P < .10 on univariate analysis were considered for entry into the multivariable logistic regression model. Age was entered as a continuous variable in all models, and Asian and other race were combined with Hispanic race in HCV models because of zero cells. For the multivariable model, the hotdeck procedure in STATA was used to complete multiple imputation (5 times) for variables with fewer than 10% of values missing. Variables for which responses were missing from more than 10% of participants were not considered reliable enough to be examined. For HCV, these included number of lifetime sexual partners and type of ink used for tattooing.
To achieve the most parsimonious final model, the full model was subjected to a series of likelihood ratio tests. If the P value for a likelihood ratio test was greater than .05, then that variable was removed from the model, and a new model was run. This process was repeated until all the variables left in the model had a P value for the likelihood ratio test of < .05. Adjusted odds ratios and 95% confidence intervals were calculated for those variables that remained significant in the final model. The potential significance of each subgroup category for tattooing and body piercing (e.g., number received, where received) with a P value < .10 on univariate analysis was further examined by including each in separate logistic stepwise regression models with the presence of all other significant risk factors.