Association of tattooing and hepatitis C virus infection: A multicenter case-control study†‡
Potential conflict of interest: Nothing to report.
Presented in abstract form at the 57th Annual Meeting of the American Association for the Study of Liver Diseases, October 27-31, 2006.
Although injection drug use (IDU) and blood transfusions prior to 1992 are well-accepted risk factors for hepatitis C virus (HCV) infection, many studies that evaluated tattooing as a risk factor for HCV infection did not control for a history of IDU or transfusion prior to 1992. In this large, multicenter, case-control study, we analyzed demographic and HCV risk factor exposure history data from 3,871 patients, including 1,930 with chronic HCV infection (HCV RNA–positive) and 1,941 HCV-negative (HCV antibody–negative) controls. Crude and fully adjusted odds ratios (ORs) of tattoo exposure by multivariate logistic regression in HCV-infected versus controls were determined. As expected, IDU (65.9% versus 17.8%; P < 0.001), blood transfusion prior to 1992 (22.3% versus 11.1%; P < 0.001), and history of having one or more tattoos (OR, 3.81; 95% CI, 3.23-4.49; P < 0.001) were more common in HCV-infected patients than in control subjects. After excluding all patients with a history of ever injecting drugs and those who had a blood transfusion prior to 1992, a total of 1,886 subjects remained for analysis (465 HCV-positive patients and 1,421 controls). Among these individuals without traditional risk factors, HCV-positive patients remained significantly more likely to have a history of one or more tattoos after adjustment for age, sex, and race/ethnicity (OR, 5.17; 95% CI, 3.75-7.11; P < 0.001). Conclusion: Tattooing is associated with HCV infection, even among those without traditional HCV risk factors such as IDU and blood transfusion prior to 1992. (HEPATOLOGY 2013;57:2117–2123)
Hepatitis C virus (HCV) infection is the most common blood-borne infection in the United States, affecting over 3 million people1-4 of all ages, races, and sexes.5, 6 By 2007, HCV had superseded human immunodeficiency virus as a cause of death in the United States,4 yet approximately 50% to 75% of infected adults are unaware of their infection status.7, 8 Injection drug use (IDU) is currently the leading cause of transmission, accounting for 60% of new cases each year2, 3 through both the sharing of needles9, 10 and through drug preparation equipment11; however, approximately 20% of incident cases have no history of IDU or other parenteral exposure.12
As new and better medications for the treatment of HCV become available,13-15 measures to increase detection rates and engagement in care are paramount. In the last decade, an expanding body of research has emerged, evaluating both traditional and nontraditional risk factors for HCV infection in an effort to increase the yield from costly yet potentially life-saving screening efforts.7, 16-19 Most recently, the US Department of Health and Human Services issued an action plan for the prevention, care, and treatment of viral hepatitis, setting goals to increase the proportion of persons who are aware of their HCV infection from 45% to 66%, and to reduce the number of new cases of HCV infection by 25%.20
In contrast to the overwhelming evidence implicating IDU in HCV acquisition, the association between HCV transmission and other suspected risk factors such as tattooing is more controversial. Although some studies have demonstrated an association between tattoos and HCV infection, others have not.21 Prior studies that examined tattooing behavior and HCV infection in the United States were limited by small sample sizes (<100 cases for case-control or <2,000 for cross-sectional studies) and failure to report adjusted odds ratios.21 Additionally, some studies that found an association between tattoos and HCV infection did not control for well-established HCV risk factors such as IDU and transfusion before 1992,21 thus limiting the interpretation of the results.
The prevalence of tattooing is on the rise in the United States. A recent Harris poll reflects a significant increase in tattooing among adults in the last decade, with 1 in every 5 reporting one or more tattoos in 2012.22 Few states have effective public health and safety regulations relating to the application of body art, and little is known about the local or systemic consequences of body art application.23 Using a large, multicenter, case-controlled study, our aim was to assess the association between HCV infection and tattoos after excluding those who lack traditional risk factors such as prior IDU or pre-1992 blood transfusion, and number of sex partners.
Patients and Methods
Patients were enrolled from the adult primary care and adult gastroenterology clinics at three main centers: the Manhattan and Brooklyn campuses of the Veterans Affairs New York Harbor Healthcare System along with the Bellevue Hospital Center in New York, NY. The latter site is a municipal hospital affiliated with New York University serving relatively poor and uninsured patients. Inclusion criteria for HCV-infected cases included laboratory results showing a positive HCV antibody and the presence of HCV viremia by polymerase chain reaction. The inclusion criteria for HCV-negative controls were those with negative HCV antibody. Patients presented to the outpatient care centers for either health screening or acute complaints. The reasons for presentation did not differ between cases and controls. Patients were excluded from the study if they had another cause of chronic liver disease, no prior HCV serological testing, a positive HCV antibody and negative polymerase chain reaction, were unable to read or understand the English language survey, or refused to complete the anonymous questionnaire. The dates of enrollment were from April 2004 to May 2006.
Questionnaire Design and Record of HCV Serostatus.
An anonymous questionnaire was designed by the study authors assessing patient demographics, knowledge of transmission of HCV infection, and exposure history to proven and suspected risk factors for HCV infection. Separate surveys were designed with questions pertinent to HCV-positive (HCV+) and HCV-negative (HCV−) participants. These surveys were tested for face and content validity by a group of adult gastroenterology and primary care physicians not directly involved in the study. The questionnaire was pretested in 10 HCV+ and 10 HCV− patients who provided feedback on the readability and clarity of the survey. After appropriate modifications, the questionnaire was again tested in 10 different HCV+ and HCV− patients before full implementation.
Each participant was asked to complete the survey at the time of his or her previously scheduled clinic visit. Patients submitted the survey anonymously and were not contacted after the survey was returned. No personal identifiers were recorded. Informed consent was obtained from prospective subjects, and each subject's electronic medical record was accessed to ascertain HCV serostatus and to determine which questionnaire to provide (HCV+, HCV−, or HCV untested). Individuals classified as “HCV untested” were not included in the study. To minimize recall bias, subjects were informed that a study of HCV and hepatitis vaccination awareness was being conducted in the general adult population, and that their invitation was not to be interpreted as particular suspicion of HCV infection in their individual case. The HCV+ and HCV− surveys were marked in a discrete way such that the subjects were not informed of their serostatus by the questionnaire. Surveyors were trained to interact consistently with HCV+ and HCV− volunteers, as they were unmasked. Surveyors were forbidden to answer questions or assist in completion of the survey aside from providing a writing instrument as needed. The primary outcome was to compare the odds of having one or more tattoos in HCV+ cases compared with HCV− controls. The exact question asked on the survey was: “Have you ever had a tattoo?” Information was entered into a database from which analyses were done. The institutional review boards of both the Veterans Affairs New York Harbor Healthcare System and the Langone Medical Center of New York University approved the study.
Statistical analysis was performed using Stata version 11.2 (Stata, College Station, TX) and a two-tailed P value of <0.05 was considered statistically significant. Colinearity of predictor variables were checked using the variance inflation factor test, using a cutoff of 2.5. Age was entered directly on the survey, whereas other variables were considered categorical and were treated as ordinal or nominal where appropriate. A Student t test was used to analyze continuous variables (e.g., age), and ordinal/nominal variables were compared using a chi-square test.
Univariate analyses were used to identify those variables that were significantly associated with case or control status, including the main exposure of interest and all potential confounders. Multivariate logistic regression was then performed using forced logistic regression for age, race, and sex. Finally, all statistically significant variables in the univariate analyses were considered in a model using a forced logistic regression model. For each model, the adjusted odds ratio (OR), 95% confidence interval (CI), and P value of tattoo exposure were calculated.
A total of 3,871 patients were enrolled, including 1,930 patients with chronic HCV infection and 1,941 HCV− controls (Table 1). There were no differences in the mean age (55.2 ± 9.0 versus 55.6 ± 11.3 years; P = 0.34) or male sex proportion (80.3% versus 81.4%; P = 0.39) between HCV-infected patients and controls; however, HCV+ patients were more likely to be racial/ethnic minorities (56.5% versus 78.5%; P < 0.001). As expected, IDU (65.9% versus 17.8%; P < 0.001), blood transfusions prior to 1992 (22.3% versus 11.1%; P < 0.001), and history of having one or more tattoos (35.2 versus 12.5%; P < 0.001) were more common in HCV-infected patients than in control subjects.
Table 1. Baseline Characteristics of HCV+ Patients and HCV− Controls
|Age, years, mean ± SD||55.2 ± 9.0||55.6 ± 11.3||0.33|
|Race/Ethnicity*|| || ||<0.001|
| White||21.5||43.5|| |
| Black||41.0||30.1|| |
| Hispanic||19.8||17.8|| |
| Other||17.7||8.6|| |
|Born in the United States||69.7||73.6||0.003|
|>12 years education||35.9||22.7||<0.001|
|Transfusion before 1992||22.3||11.1||<0.001|
|Intranasal drug use†||61.8||22.6||<0.001|
|Number of lifetime sexual partners|| || ||<0.001|
| 0-2||13.6||17.4|| |
| 3-9||20.9||29.8|| |
| 10-25||38.2||24.9|| |
| ≥26||27.3||27.9|| |
|Sex with a prostitute†||59.5||37.7||<0.001|
|Sex with a same-sex partner†||13.1||12.4||0.49|
|Incarcerated for >48 hours†||48.8||13.9||<0.001|
|Drinks of alcohol per week|| || ||<0.001|
| None||69.3||55.8|| |
| 1-6||35.8||35.8|| |
| ≥7||6.4||8.4|| |
Patients with HCV infection were significantly more likely to have a history of tattoo exposure (OR, 3.81; 95% CI, 3.23-4.49; P < 0.001) and this remained significant after adjustment for age, sex, and race/ethnicity (OR, 4.51; 95% CI, 3.78-5.39; P < 0.001), and all potential confounding variables identified in table 1 (OR, 3.74; 95% CI, 2.95-4.73; P < 0.001) (Table 2).
Table 2. Crude and Adjusted ORs for Tattoo Exposure in 1,930 HCV+ Patients and 1,941 HCV− Controls
|Fully adjusted†||3.74 (2.95-4.73)||<0.001|
After excluding all patients with a history of ever injecting drugs and those who had a blood transfusion prior to 1992, a total of 1,886 subjects remained for analysis, including 465 HCV+ patients and 1,421 controls (Table 3). Among this subset of individuals without traditional risk factors for HCV infection, we found that HCV+ patients were still significantly more likely to have a history of tattoo exposure (OR, 3.83; 95% CI, 2.99-4.93; P < 0.001) and this remained statistically significant after adjustment for age, sex, and race/ethnicity (OR, 4.48; 95% CI, 3.42-5.87; P < 0.001) and all potential confounding variables identified in Table 3 at or below P = 0.10 (OR, 5.17; 95% CI, 3.75-7.11; P < 0.001) (Tables 4 and 5). In addition, after excluding intranasal drug users from the analysis and adjusting for all potential confounding variables, HCV+ patients remained significantly more likely to have a history of tattoo exposure compared with HCV− controls (OR, 8.22; 95% CI, 5.45-12.40; P < 0.001).
Table 3. Self-Reported Patient Demographics for All HCV+ Patients and HCV− Controls, Excluding Traditional Risk Factorsd
|Age, years, mean ± SD||52.0 ± 7.6||53.1 ± 10.6||0.04|
|Race/Ethnicity*|| || ||<0.001|
| White||26.4||43.0|| |
| Black||41.2||29.5|| |
| Hispanic||18.9||18.7|| |
| Other||13.5||8.9|| |
|Born in the United States||71.0||71.6||0.8|
|Education >12 years||65.7||79.17||<0.001|
|Income > $15,000||53.0||56.3||0.214|
|Intranasal drug use†||34.6||21.7||<0.001|
|Number of lifetime sexual partners|| || ||0.08|
| 0-2||22.3||18.4|| |
| 3-9||24.9||30.5|| |
| 10-25||25.3||24.4|| |
| ≥26||27.5||26.6|| |
|Sex with a prostitute†||47.4||36.7||<0.001|
|Sex with a same-sex partner†||5.8||12.0||<0.001|
|Incarcerated for >48 hours†||41.4||13.9||<0.001|
|Drinks of alcohol per week|| || ||<0.001|
| None||67.0||55.4|| |
| 1-6||24.2||35.2|| |
| ≥7||8.8||9.4|| |
Table 4. Crude and Adjusted ORs for Tattoo Exposure in 466 HCV+ Patients Compared with 1,421 HCV− Controls Without Traditional Risk Factors
|Fully adjusted†||5.17 (3.75-7.11)||<0.001|
|Fully adjusted† and IDU excluded||8.22 (5.45-12.40)||<0.001|
Table 5. Crude and Adjusted ORs for All Variables Included in Model Assessing Tattoo Exposure in 466 HCV+ Patients Compared with HCV− Controls After Excluding IDU and Blood Transfusion Recipients
|Born in the United States||0.97 (0.77-1.22)||1.06 (0.81-1.39)||0.44|
|Married||0.94 (0.88-1.00)||0.96 (0.90-1.03)||0.23|
|≤12 years education||1.98 (1.57-2.50)||2.13 (1.67-2.71)||<0.001|
|Income <$15,000||1.14 (0.92-1.41)||1.12 (0.90-1.40)||0.31|
|Currently employed||0.67 (0.54-0.83)||0.61 (0.49-0.77)||<0.001|
|Intranasal drug use‡||1.91 (1.52-2.40)||1.82 (1.43-2.31)||<0.001|
|>25 sexual partners||1.04 (0.82-1.32)||1.18 (0.92-1.51)||0.18|
|Sex with a prostitute‡||1.59 (1.29-1.97)||1.65 (1.31-2.08)||<0.001|
|Sex with a same-sex partner||0.44 (0.29-0.68)||0.56 (0.37-0.87)||0.009|
|Incarcerated for >48 hours||5.20 (4.08-6.64)||4.84 (3.76-6.25)||<0.001|
|Body piercing||1.39 (1.10-1.70)||1.18 (0.92-1.51)||0.19|
|Acupuncture||0.91 (0.66-1.25)||0.88 (0.64-1.23)||0.48|
|Drinks of alcohol per week||0.66 (0.56-0.79)||0.73 (0.61-0.88)||<0.001|
In the present study of nearly 4,000 patients, we found that tattooing was significantly and independently associated with HCV infection. The association persisted after adjusting for age, sex, and race/ethnicity as well as after excluding subjects with traditional risk factors such as IDU and blood transfusion prior to 1992. These findings have important implications for screening non–injection drug users in the United States, particularly since the prevalence of tattooing is on the rise and intravenous drug use is on the decline.
The prevalence of tattoos in the United States has been increasing during the past decade, particularly among youths.22-25 Although little is known about the prevalence of body art among minority adolescents, one study of African American and Hispanic students from an inner city high school in Texas found that 10% of the African American students already had a tattoo by graduation, a rate that is comparable to prior studies that evaluated predominantly white college students.26 A 2004 study among persons aged 18 to 50 years in the United States found that 24% of respondents had at least one tattoo, and an additional 21% of nontattooed respondents had considered tattoo placement.23 Tattooing is more common among those of low socioeconomic status23 despite its increased prevalence across all social groups, and it is also highly prevalent among soldiers. In one study, almost 36% of soldiers in the US Army had at least one tattoo, and 76% experienced bleeding after the procedure, which might promote transmission of blood-borne infections.27
The literature assessing the association between tattooing and HCV has heretofore been equivocal. Because of the wide variability of study populations with regard to baseline risk of HCV exposure, previous work has been risk-stratified by general population, blood donors, high-risk groups (i.e., drug users, homeless persons, sex workers, and patients in sexually transmitted disease clinics), prisoners, and veterans. Although studies that recruited >1,000 veterans found an almost three-fold higher risk of HCV infection among veterans with a tattoo compared with those who did not have a tattoo,28-30 results from cross-sectional studies involving the general public, blood donors, and other high-risk groups have been inconsistent.21
A recent review article of the best available data on the risk of HCV infection from tattoo exposure found that most studies relied on descriptive statistics alone and failed to report measures of association, such as ORs and relative risk. In fact, meta-analysis of the existing literature was deferred because several of the studies that found no association between HCV infection and tattooing in the univariate analysis either did not include those exposures in the multivariate analysis or did not report the adjusted OR.21 Furthermore, few case-controlled studies completely excluded injection drug users and blood transfusion recipients.21
Our study confirms the association between tattoo exposure and hepatitis C infection in a very large ethnically diverse population of HCV cases and uninfected controls. To our knowledge, this is largest group of HCV cases and controls ever assembled to study this question after excluding all injection drug users and recipients of blood transfusions before 1992, along with verification of HCV seropositivity and viremia through the electronic medical record at each study location.
To date, there is no definitive evidence that HCV infection occurs through tattooing when sterile equipment is used. Although no outbreaks of HCV infection have been detected in the United States that originate from professional tattoo parlors, case reports of acute HCV infection from tattooing in prison suggest that tattooing could be a mode of transmission.31-33 One case report from a United States prison documented HCV seroconversion in a prisoner, where tattooing in prison was the only known risk factor during the incubation period.33
Underrepresentation due to self-reporting of intravenous drug use is a concern that could confound our result. Tattoos and drug use often coexist, therefore the increased risk of HCV infection among tattooed individuals may in fact be a surrogate for unreported drug use.6, 34-38 Although a case series of 301 patients by Flamm et al.39 found that 8.5% of chronic HCV-infected male patients younger than 45 who were initially referred with “no known risk factor” later endorsed a remote history of intravenous drug use, intravenous drug use self-reporting has been shown to be accurate when high methodological standards are applied.40 When our surveys were completed, there were no patient identifiers to subsequently associate patients with their answers, providing a confidentiality that was ensured to patients prior to receiving the survey during the consent process. Our questionnaires were completed anonymously, allowing subjects to report drug use and sexual behavior without concern about personal identification.
Some may ascribe the risk of HCV infection from tattooing to another unrecognized high-risk behavior (e.g., increased intranasal drug use or sexual promiscuity among those having one or more tattoo); however, these concerns were not borne out by our analysis. Sexual contact is responsible for a very low but not negligible transmission of HCV.1, 41, 42 Our two cohorts without traditional risk factors had equal proportions reporting >25 lifetime sexual partners (27.5% for HCV+ and 26.6% for HCV−; P = 0.714), and their remaining unequal proportions reported prior sexual contact with a prostitute or same-sex partner (Table 3). We adjusted for these sexual contacts and other potential residual confounders via logistic regression analysis and found that in those without prior IDU or pre-1992 blood transfusion, the odds of tattoo exposure were still higher in HCV+ patients than in HCV− controls.
Although commercial parlors have not been implicated in HCV transmission, such transmission could occur at different stages of tattooing—the reuse of nondisposable needles, inappropriate sterilization of equipment, or reuse of ink contaminated with blood from an infected process. Although data on survival of hepatitis C in tattooing or piercing equipment are not available, survival of HCV ranges from a few days on inanimate surfaces to almost 1 month in propofol solutions.43-46 In fact, the US Occupational Safety and Health Administration recognizes tattooing as a potential mode of transmission of blood-borne pathogens (it is included in their blood-borne safety standards). Furthermore, more than two-thirds of state health jurisdictions in the United States have additional regulations for tattooing parlors.25
Tattooing in prison is of particular concern regarding the transmission of blood-borne infections, because tattooing in this setting is typically performed using nonsterile equipment, such as guitar strings, paper clips, or sewing needles, which are usually cleaned via heating or use of boiling water.47 A similar concern exists for other nonprofessional settings and nonprofessional tattoo artists. Of particular concern are those parlors servicing adolescents without the informed consent of a parent. Many states require that minors obtain parental consent for tattoos and piercings; however, in one study from an urban Texas high school, about 20% of those who obtained their tattoo from a professional were not asked for proof of parental consent.26
The limitations of our study include a patient population from two veteran administration hospitals that are predominantly male and one urban municipal hospital slanted toward the lower end of the socioeconomic scale, limiting how these findings could be generalized to other segments of the population, particularly women or more affluent populations. Compared with the control group, the hepatitis C cohort had a higher proportion of self-identified racial or ethnic minorities (56.5% versus 78.5%, P < 0.001). Furthermore, our study did not recruit patients with incident cases of HCV infection and ask about tattoo exposure or specify the venue of tattoo placement, which hinders drawing temporal causal relationships between HCV infection and tattooing as well as limiting our ability to comment on how sterile infection control practices can mitigate the risk of transmission. Future analysis will help determine how these distinctions would further qualify the overall result.
In conclusion, tattoo exposure is associated with HCV infection, even among those without traditional risk factors. All patients who have tattoos should be considered at higher risk for HCV infection and should be offered HCV counseling and testing. Expanding screening recommendations to cover individuals with one or more tattoos offers a potential compliment to current risk-based screening recommendations. Because of the increasing prevalence of tattooing, particularly among youths, awareness campaigns should highlight the danger of transmitting blood-borne infections such as HCV, regardless of the venue of placement.
It is with tremendous respect and deep sorrow that we acknowledge the contributions of our colleague and dear friend, Dr. Edmund Bini (1967-2010), both to this manuscript and to our lives. He is truly missed.