| |
Association Between Hepatitis C Virus and Head and Neck Cancers
|
| |
| |
Download the PDF here
HCV Linked to head and neck cancers....two to five times more likely to develop
"This case-control study conducted at a National Cancer Institute-designated cancer center is the first to show that HCV is statistically significantly associated not only with nonoropharyngeal (except nasopharyngeal) HNCs but also with HPV-positive oropharyngeal cancers. Our results add to the growing body of epidemiological evidence that HCV infection has extrahepatic manifestations and may be associated with non-liver-related cancers (9,21).
Our findings are clinically significant. HCV infection is associated with several types of B-cell non-Hodgkin's lymphoma, especially the diffuse large B-cell, marginal zone, follicular, and lymphoplasmacytic types (6-8,22). The reported odds ratios for these associations range from 2.47 to 5.2 (6,22). We found a similar strength of association between HCV and nonoropharyngeal HNCs and between HCV and HPV-positive oropharyngeal cancers. Thus, clinicians should be aware that nonliver cancers besides non-Hodgkin's lymphomas can develop in patients chronically infected with HCV. Additionally, oncologists treating patients with HNCs should consider testing patients for HCV to enable early identification and linkage of care for this infection and to prevent progression of underlying liver disease. It remains unknown whether early HCV treatment may prevent development of HCV-associated HNCs or improve the oncologic outcome of patients in whom such cancers develop as has been reported in patients with HCV-associated non-Hodgkin's lymphoma (6-8,23).
The role of HCV in oropharyngeal carcinogenesis may also relate to direct involvement of HCV proteins, including HCV nonstructural proteins NS5B and NS3 and HCV core proteins, in disrupting cell cycle regulation. NS5B binds the retinoblastoma tumor suppressor protein (Rb) in the cytoplasm of infected cells and recruits E6-associated protein, leading to polyubiquitination and proteasomal degradation of Rb (33). Because Rb is critical in regulating G1-S cell cycle transition, DNA damage response, mitotic spindle checkpoints, and apoptosis, loss of Rb may lead to uncontrolled cell proliferation (33). Notably, HPV protein E6 also exerts a pro-oncogenic impact by targeting tumor suppressor proteins through E6-associated protein, leading to degradation of cellular p53 and enhanced phosphorylation of Rb (34). Various interactions of HCV core and NS3 proteins with p53 have also been documented (33). The loss of p53 induced by HPV E6 protein and the degradation of Rb caused by HCV NS5B protein may play a synergistic role in the development of oropharyngeal cancers and merits further investigation. Additional studies are also warranted to examine the association between HCV and HPV-related anogenital (cervix, vagina, vulva, penis, and anus) cancers."
Discussion
This case-control study conducted at a National Cancer Institute-designated cancer center is the first to show that HCV is statistically significantly associated not only with nonoropharyngeal (except nasopharyngeal) HNCs but also with HPV-positive oropharyngeal cancers. Our results add to the growing body of epidemiological evidence that HCV infection has extrahepatic manifestations and may be associated with non-liver-related cancers (9,21).
Our findings are clinically significant. HCV infection is associated with several types of B-cell non-Hodgkin's lymphoma, especially the diffuse large B-cell, marginal zone, follicular, and lymphoplasmacytic types (6-8,22). The reported odds ratios for these associations range from 2.47 to 5.2 (6,22). We found a similar strength of association between HCV and nonoropharyngeal HNCs and between HCV and HPV-positive oropharyngeal cancers. Thus, clinicians should be aware that nonliver cancers besides non-Hodgkin's lymphomas can develop in patients chronically infected with HCV. Additionally, oncologists treating patients with HNCs should consider testing patients for HCV to enable early identification and linkage of care for this infection and to prevent progression of underlying liver disease. It remains unknown whether early HCV treatment may prevent development of HCV-associated HNCs or improve the oncologic outcome of patients in whom such cancers develop as has been reported in patients with HCV-associated non-Hodgkin's lymphoma (6-8,23).
HNCs are a heterogeneous group of cancers accounting for approximately 3% of all cancers diagnosed in the United States (24). Risk factors for these cancers include smoking, alcohol consumption, chewing betel nut, African American race, low socioeconomic status, HPV infection (for oropharyngeal cancers), and EBV infection (for nasopharyngeal cancers) (13,25,26).
Little has previously been published about the relationship between HCV infection and HNCs. A few articles suggest that HCV infection may be associated with squamous cell carcinomas of the oral cavity (27,28). A US study showed a higher HCV prevalence in patients with squamous cell carcinoma of the head and neck than in noncancer patients (21.2% vs 9.9%, P < .004) (27). A population-based study from Taiwan found that the incidence of oral cavity cancers was 2.28 times as high in HCV-infected individuals as in patients with nonviral hepatitis (6.15 vs 2.69 per 100 000 person-years) and that HCV alone was a statistically significant risk factor for oral cavity cancers (HR = 1.9) (28). Recently, the US population-wide Chronic Hepatitis C Cohort Study analyzed data on 12 126 patients with chronic HCV infection and found that such patients had an increased incidence of oral cavity cancers (standardized rate ratio = 2.5) and mortality from oral cavity cancers (relative risk = 5.2) compared with noninfected patients (9). In our present study, HCV seropositivity was associated not only with nonoropharyngeal HNCs but also with HPV-positive oropharyngeal cancers. Our results also confirmed that HCV seropositivity was associated with oral cavity cancers, as previously reported (27,28).
The mechanisms by which HCV may be associated with HNCs remain unclear and need further investigation. HCV has been detected in saliva and salivary glands in patients with chronic sialadenitis and salivary gland tumors (29). Furthermore, an extrahepatic manifestation of chronic HCV infection is oral lichen planus, which is a premalignant condition associated with development of squamous cell carcinoma of the oral cavity (30). Case reports also have described progression of lichen planus to squamous cell carcinoma of the oral cavity in HCV-infected patients (30,31). Enhanced replication of HCV in oropharyngeal tissues may contribute to chronic inflammation predisposing to cancer development.
A novel finding from our study was that HCV seropositivity was associated with HPV-positive but not HPV-negative oropharyngeal cancers. HPV persists as a chronic infection, in part, by evading antiviral immune responses (32). High-risk HPV types have been shown to downregulate interferon-α-inducible gene expression (32). HPV oncoproteins E6 and E7 directly interact with and inhibit type 1 interferon signaling (32), which plays a central role in controlling intracellular replication of HCV (32). Thus, HPV may facilitate proliferation of HCV in the oropharyngeal cells, thereby facilitating its oncogenic action.
The role of HCV in oropharyngeal carcinogenesis may also relate to direct involvement of HCV proteins, including HCV nonstructural proteins NS5B and NS3 and HCV core proteins, in disrupting cell cycle regulation. NS5B binds the retinoblastoma tumor suppressor protein (Rb) in the cytoplasm of infected cells and recruits E6-associated protein, leading to polyubiquitination and proteasomal degradation of Rb (33). Because Rb is critical in regulating G1-S cell cycle transition, DNA damage response, mitotic spindle checkpoints, and apoptosis, loss of Rb may lead to uncontrolled cell proliferation (33). Notably, HPV protein E6 also exerts a pro-oncogenic impact by targeting tumor suppressor proteins through E6-associated protein, leading to degradation of cellular p53 and enhanced phosphorylation of Rb (34). Various interactions of HCV core and NS3 proteins with p53 have also been documented (33). The loss of p53 induced by HPV E6 protein and the degradation of Rb caused by HCV NS5B protein may play a synergistic role in the development of oropharyngeal cancers and merits further investigation. Additional studies are also warranted to examine the association between HCV and HPV-related anogenital (cervix, vagina, vulva, penis, and anus) cancers.
Another novel finding of our study was that HCV seropositivity was associated with laryngeal cancers. Risk factors previously identified for laryngeal cancers include smoking, alcohol intake, injection drug use, gastroesophageal reflux, and occupational exposures like asbestos (35,36). Of interest, recent studies have shown an association between HPV infection and laryngeal cancers (37-39). The association of HCV with laryngeal cancers may be because of mechanisms similar to those described above for HPV-positive oropharyngeal cancers. Correlation between smoking and/or drug use could partially explain the association of HCV with laryngeal cancers. However, detailed evaluation of risk factors for laryngeal cancers, including testing of cancers for HPV, was not possible because of lack of data.
Our study had several strengths. We evaluated the largest series of oropharyngeal and nonoropharyngeal HNCs for evaluating associations with HCV as separate groups because their risk factors differ. All cancers in our study population were histopathologically confirmed. The sensitivity analyses that included only subjects who had HCV infection (detectable HCV RNA) improved the validity of our findings. We evaluated smoking and alcohol consumption in detail and adjusted for them in final analyses, an approach not previously taken and often difficult through population-based databases and cancer registries. We also utilized existing data on HPV status of oropharyngeal cancers and found that HCV was associated with HPV-positive oropharyngeal cancers, suggesting oncogenic synergy between the two chronic viral infections.
Our study also had several limitations. First, the control group was not cancer free; rather, control subjects were patients with any of the three major smoking-associated cancers. In this retrospective study, we did not have access to a cancer-free population. We selected this control group so that case and control subjects were similar with respect to major risk factors-smoking, alcohol, and socioeconomic status-and adjusted for them in the final analyses. Second, since we used hospital-based control subjects, there was a possibility of Berkson's bias, which pertains to different exposure rates in the control group and the general population (40). Notably, the prevalence of HCV infection in our control group was 6%, higher than the prevalence of approximately 1.5% in the general US population (1). Thus, our results may actually underestimate the true effect size of the association between HCV seropositivity and HNCs. An alternate explanation is that the HCV seropositivity rate is higher in cancer patients than in the general population. Third, there is the possibility of selection bias as patients with solid tumors are not routinely tested for HCV antibodies in our center. However, the proportions of cancer patients seen at our institution during the study period who were tested for HCV antibodies did not differ among the case and control groups (Supplementary Table 2, available online). Nevertheless, bias resulting from different reasons for HCV screening in case and control groups may be possible. Fourth, the association between HCV and HPV-positive oropharyngeal cancers may reflect a correlation between HCV and HPV infection because of lifestyle behaviors. Both HPV-positive oropharyngeal cancer patients and HCV-infected individuals tend to have high numbers of lifetime sex partners (41,42). However, whereas HPV is mostly transmitted sexually, sexual transmission of HCV infection is extremely rare, with the incidence being one case per 190 000 sexual contacts among heterosexual couples (43); the most frequent route of HCV transmission is injection drug use (44). Thus, this presumed correlation between HCV and HPV infection may reflect correlation between the two high-risk behaviors. Because of the retrospective nature of our study and the nonavailability of reliable assessments of sexual and drug use history, evaluation of such behaviors was not possible. Fifth, we did not have data on HCV treatment in our study subjects. Because population-based studies have shown that most HCV-infected patients are not aware of their infection and healthcare utilization rates are low for patients with HCV infection (42), we believe that lack of HCV treatment data is unlikely to affect our results. Sixth, more than 50% of the patients in our study were born between 1945 and 1965; individuals born during that period have a high prevalence of HCV infection (1). However, because the recommendations for screening such individuals were made in 2012 (1) and our study included patients tested between 2004 and 2014, we believe that birth between 1945 and 1965 is unlikely to have been the reason for HCV testing for the majority of patients in our study population. Nevertheless, our results were adjusted for the birth cohort effect in our multivariable analyses. Seventh, even though we tried to adjust for smoking, alcohol intake, and socioeconomic status, data were obtained from self-reported questionnaires and underreporting or misreporting of information along with residual confounding is possible. Eighth, the case-case analyses showed no etiologic heterogeneity between the case groups compared with respect to HCV seropositivity; however, the number of patients in comparison groups was small to determine differences. Finally, we were unable to evaluate the association between HCV and nasopharyngeal cancers based on EBV status because of small sample size.
In conclusion, HCV seems to be associated with HNCs, particularly nonoropharyngeal and HPV-positive oropharyngeal cancers. Validation of these associations by analysis of population-based datasets and cancer-free control group is essential. Further studies are also required to explore the possible interaction between HCV and HPV and the association between HCV and other HPV-related malignancies.
-------------
People with hepatitis C are two to five times more likely to develop certain head and neck cancers
MD Anderson study first to find association with new cancer types; findings have strong implications for screening and treatment
University of Texas M. D. Anderson Cancer Center
http://www.eurekalert.org/pub_releases/2016-04/uotm-pwh041316.php
Long associated with liver cancer and non-Hodgkin's lymphoma, a study from The University of Texas MD Anderson Cancer Center reveals for the first time that the hepatitis C virus (HCV) is associated with certain head and neck cancers. The findings, published in the Journal of the National Cancer Institute, could have significant implications for both the screening of those with the virus and the treatment of those with head and neck cancers.
Hepatitis C, the most common blood-borne infection in the U.S., is a virus that affects up to 1.5 percent of the population. It's estimated that as many as 3.9 million are chronically infected with the virus, according to the researchers.
In the last few years, new antiviral drugs have made it possible to cure more than 90 percent of the HCV population, says Harrys A. Torres, M.D., associate professor, Infectious Disease, Infection Control and Employee Health. The antivirals are oral medications taken once or twice daily with almost no side effects, he explains.
In 2009, MD Anderson opened what remains the only clinic of its kind at a comprehensive cancer center to address the unmet medical needs of its patients with HCV.
"Obviously, a hepatitis C infection could impact how patients respond to their cancer therapy. We also realized that many of our hepatitis patients were excluded from clinical trials. Now that many with hepatitis C can be cured, it is important that we first address and potentially cure the virus, so that they can have access to necessary cancer therapy."
When Torres started the clinic, he expected to see a number of patients with liver cancers and non-Hodgkin's lymphoma, as these have documented associations with HCV of 48-fold and two- to three-fold increased risk, respectively. Other recent studies have recognized HCV's increased association with additional cancers, says Torres, but there was no known association with a significant number of head and neck cancers.
"To our surprise, we saw a number of head and neck cancer patients who tested positive for the hepatitis C virus. With this observation we began to wonder if there was an undiscovered correlation between the two. Our findings tell us that the association between hepatitis C and oropharyngeal and nonoropharyngeal cancers is as high as its link to non-Hodgkin's lymphoma."
Oropharyngeal cancers occur in the oropharynx, or the middle part of the throat, including the back one-third of the tongue, the soft palate, tonsil, and side and back walls of the throat. Nonoropharyngeal cancers include those occurring in the oral cavity, nasal cavity and larynx.
For the retrospective, case-controlled study, the researchers identified 34,545 MD Anderson patients who were tested for HCV between 2004 and 2014. All patients were tested for HCV antibodies and viral RNA tests were used to confirm chronic infection, when available.
The researchers included 409 head and neck cancer patients as case subjects (164 with oropharyngeal and 245 with nonoropharyngeal). Also paramount to the research, said Torres, was to control for smoking, a major risk factor for head and neck cancers. Therefore, they identified 694 control subjects, all with a diagnosis of smoking-related cancers (378 with lung, 168 with esophageal and 148 with bladder).
The study revealed that 14 percent of patients with oropharyngeal cancers tested positive for HCV antibodies, compared to just 6.5 percent in the control group. In those with nonoropharyngeal cancer, 20 percent tested positive for HCV antibodies. All findings were highly statistically significant.
Compared to the controls, the researchers found that the risk for HCV patients of developing specific head and neck cancers was increased 2.4 times for oral cavity cancers, 2.04 times for oropharynx cancers, and 4.96 times for larynx cancers.
Of note, 145 of all the oropharyngeal cancer patients were also tested for human papillomavirus (HPV), allowing researchers to compare possible associations between the two viruses. Patients with HCV-positive head and neck cancers were more likely to also test positive for HPV.
This finding, says Torres, is an area of great interest for future research study. Given the association found between the two viruses in this patient population, Torres and colleagues plan to look at other HPV-associated cancers and their possible link to HCV, under MD Anderson's Moon Shots Program.
Torres notes that it will be important to screen for HCV because treatment with antiviral drugs may possibly prevent cancer from ever developing, as reported for liver cancers and non-Hodgkin's lymphoma.
It may also impact treatment for patients who have already developed cancer. In fact, for patients with HCV and some indolent non-Hodgkin's lymphoma, Torres notes that the National Comprehensive Cancer Network guidelines now recommend that the HCV be treated first, given that it is curable. In some cases, explains Torres, the lymphoma has disappeared upon treating the HCV with antiviral therapies.
With these findings, MD Anderson plans to screen and treat all head and cancer patients with HCV and follow their outcomes.
Educating both the general hepatology and infectious disease communities -- those primarily treating patients with HCV -- is critical so they understand HCV impacts not only the liver, but is a systemic infection.
"What we are trying to make all understand is that this is an infection that has consequences -- and it's an infection we can cure," says Torres.
--------------------------
Association Between Hepatitis C Virus and Head and Neck Cancers
JNCI J Natl Cancer Inst (April 13 2016)
Parag Mahale, Erich M. Sturgis, David J. Tweardy, Ella J. Ariza-Heredia,
Harrys A. Torres
Affiliations of authors: Department of Infectious Diseases, Infection Control, and Employee Health (PM, DJT, EJAH, HAT), Department of Head and Neck Surgery (EMS), Department of Epidemiology (EMS), and Department of Molecular and Cellular Oncology (DJT), The University of Texas MD Anderson Cancer Center, Houston, TX; Department of Epidemiology, Human Genetics & Environmental Sciences, The University of Texas School of Public Health, Houston, TX (PM).
Abstract
Background: Hepatitis C virus (HCV) infection is associated with hepatocellular carcinoma and non-Hodgkin's lymphoma. In 2009, MD Anderson established the first US clinic for treating HCV-infected cancer patients, where we observed an unexpectedly large number of patients with head and neck cancers (HNCs). We sought to determine whether HCV is associated with HNCs.
Methods: In this case-control study, medical records of cancer patients tested for HCV antibodies at our center from 2004 through 2014 were identified. Case subjects had new-onset primary oropharyngeal or nonoropharyngeal (oral cavity, nasopharynx, hypopharynx, or larynx) HNCs. Control subjects had smoking-associated (lung, esophagus, or urinary bladder) cancers. Biopsy reports of oropharyngeal cancers tested for human papillomavirus (HPV) were reviewed. Patients with lymphoma were excluded. Multivariable logistic regression models were constructed. All statistical tests were two-sided.
Results: Of 34 545 cancer patients tested for HCV antibodies, 409 case subjects (164 oropharyngeal and 245 nonoropharyngeal) and 694 control subjects (378 lung, 168 esophagus, and 148 urinary bladder) were studied. The prevalence of HCV seropositivity was higher in oropharyngeal cancer patients (14.0%, 95% confidence interval [CI] = 8.7% to 19.4%, vs 6.5%, 95% CI = 4.6% to 8.3%), particularly HPV-positive oropharyngeal cancer patients (16.9%, 95% CI = 8.7% to 24.9%, vs 6.5%, 95% CI = 4.6% to 8.3%), and nonoropharyngeal HNC patients (20.0%, 95% CI = 14.9% to 25.0%, vs 6.5%, 95% CI = 4.6% to 8.3%) than in control subjects. Adjusted models showed a statistically significant association of HCV seropositivity with nonoropharyngeal (except nasopharyngeal) HNCs (odds ratio [OR] = 2.85, 95% CI = 1.38 to 5.88) and HPV-positive oropharyngeal cancers (OR = 2.97, 95% CI = 1.31 to 6.76).
Conclusions: HCV is associated with nonoropharyngeal (except nasopharyngeal) and HPV-positive oropharyngeal HNCs. Further studies are required to explore the possible interaction between HCV and HPV, and the association between HCV and other HPV-related malignancies.
Approximately 2.7 to 3.9 million individuals in the United States (up to 1.5% of the population) are chronically infected with hepatitis C virus (HCV) (1). HCV is carcinogenic, and HCV-seropositive individuals have 15 to 20 times the risk of developing hepatocellular carcinoma that HCV-seronegative individuals have (2). Chronic HCV infection causes not only liver cancer but also lymphoproliferative disorders (3-5).
Clinical and epidemiological evidence supports a statistically significant association between HCV infection and development of several types of B-cell non-Hodgkin's lymphomas (6-8). Additionally, a recent analysis of data from the Chronic Hepatitis C Cohort Study showed that HCV-infected individuals also have increased incidences of other non-liver-related malignancies, including pancreas, lung, renal, and rectal cancers, along with increased cancer-associated mortality (9). Likewise, a prospective study from Taiwan showed increased mortality from extrahepatic diseases in HCV-seropositive individuals (10).
In 2009, we established the first US clinic devoted to managing HCV infection in cancer patients (11), where we observed an unexpectedly large number of patients with head and neck cancers (HNCs). In this study, we sought to determine whether HCV seropositivity is associated with oropharyngeal and nonoropharyngeal HNCs.
Results
Study Population
We identified 34 545 cancer patients tested for HCV antibodies during the study period. Of these, 1103 were included as case subjects (n = 409: 164 oropharyngeal and 245 nonoropharyngeal) or control subjects (n = 694: 378 lung, 168 esophagus, and 148 urinary bladder) (Figure 1). Patient characteristics are summarized in Table 1. Most patients were male, white, and born between 1945 and 1965; the median age at cancer diagnosis was 62 years. Approximately 41.1% of patients had a bachelor's degree or beyond, 49.0% were former smokers, and 52.3% current alcohol drinkers at cancer diagnosis. In the overall sample of case and control subjects, the prevalence of HCV seropositivity was 10.6% (n = 117, 95% CI = 8.9% to 12.5%), and 86.4% (n = 76 of 88, 95% CI = 79.2% to 93.5%) of those tested for HCV RNA had HCV infection. Of the 145 patients whose cancers were tested for HPV, 70 had high-risk HPV and 83 had p16-positive tumors.
Oropharyngeal Cancers
Compared with control subjects, oropharyngeal cancer patients were more likely to be male, born between 1945 and 1965, diagnosed with cancer at younger ages, and better educated; to have smoked fewer cigarettes; and to consume more alcohol (Table 2). There was no statistically significant difference between groups with respect to HBV exposure or HIV.
Compared with control subjects, HPV-positive oropharyngeal cancer patients were more likely to be white, male, born between 1945 and 1965, diagnosed with cancer at younger ages, and better educated; and to have smoked fewer cigarettes (Table 3).
The prevalence of HCV seropositivity was higher in oropharyngeal cancer patients than control subjects (14.0%, 95% CI = 8.7 to 19.4%, vs 6.5%, 95% CI = 4.6 to 8.3%), particularly HPV-positive oropharyngeal cancer patients (16.9%, 95% CI = 8.7 to 24.9%, vs 6.5%, 95% CI = 4.6 to 8.3%). After adjustment for age at cancer diagnosis, birth between 1945 and 1965, sex, highest education level, smoking history, and alcohol consumption, HCV seropositivity was statistically significantly associated with oropharyngeal cancer (OR = 2.04, 95% CI = 1.04 to 4.01) (Table 4). Because they have different pathogenesis, we conducted separate analyses for HPV-positive and HPV-negative oropharyngeal cancers. Final adjusted models showed a statistically significant association between HCV seropositivity and HPV-positive (OR = 2.97, 95% CI = 1.31 to 6.76) but not HPV-negative oropharyngeal cancers (OR = 1.44, 95% CI = 0.39 to 5.30) (Table 4). Similar results were obtained with sensitivity analysis for HCV infection (Table 4).
Nonoropharyngeal HNCs
Compared with control subjects, nonoropharyngeal HNC patients were more likely to be male, be diagnosed with cancer at younger ages, consume more alcohol, and have HIV co-infection (Table 2). There was no statistically significant difference between groups with respect to cigarette smoking or HBV exposure.
The prevalence of HCV seropositivity was higher in nonoropharyngeal HNC patients than control subjects (20.0%, 95% CI = 14.9 to 25.0%, vs 6.5%, 95% CI = 4.6 to 8.3%). After adjustment for age at cancer diagnosis, birth between 1945 and 1965, sex, race, highest education level, smoking history, alcohol consumption, and HIV infection, HCV seropositivity was statistically significantly associated with nonoropharyngeal HNC (OR = 2.85, 95% CI = 1.38 to 5.88) (Table 4). Within nonoropharyngeal HNCs, we conducted separate analyses for non-nasopharyngeal and nasopharyngeal cancer patients. Final adjusted models showed a statistically significant association between HCV seropositivity and non-nasopharyngeal (OR = 3.17, 95% CI = 1.49 to 6.73) but not nasopharyngeal cancers (OR = 1.30, 95% CI = 0.22 to 7.64) (Table 4). Similar results were obtained with sensitivity analysis for HCV infection (Table 4). Associations by EBV status could not be evaluated because of the small number of EBV-positive patients in each cancer group.
Case-Case Analyses
We performed case-case analyses by comparing HPV-positive to HPV-negative oropharyngeal cancers and, within nonoropharyngeal HNCs, non-nasopharyngeal to nasopharyngeal cancers to assess etiologic heterogeneity. We found no statistically significant association of HCV seropositivity with HPV-positive compared with HPV-negative oropharyngeal cancers (OR = 1.56, 95% CI = 0.32 to 7.52, P = .58) and no statistically significant association of HCV seropositivity with non-nasopharyngeal compared with nasopharyngeal cancers (OR = 1.25, 95% CI = 0.19 to 8.33, P = .82) after adjusting for the confounders mentioned above.
Exploratory Analyses
After adjustment for age at cancer diagnosis, birth between 1945 and 1965, sex, highest education level, smoking, and alcohol consumption, HCVseropositivity was statistically significantly associated with oral cavity, oropharyngeal, and laryngeal cancers after corrections for multiple comparisons (Table 5). Unfortunately, the numbers of patients with tumors of the nasopharynx, hypopharynx, and larynx tested for HCV RNA were too small to permit analysis of any association with HCV infection.
Discussion
This case-control study conducted at a National Cancer Institute-designated cancer center is the first to show that HCV is statistically significantly associated not only with nonoropharyngeal (except nasopharyngeal) HNCs but also with HPV-positive oropharyngeal cancers. Our results add to the growing body of epidemiological evidence that HCV infection has extrahepatic manifestations and may be associated with non-liver-related cancers (9,21).
Our findings are clinically significant. HCV infection is associated with several types of B-cell non-Hodgkin's lymphoma, especially the diffuse large B-cell, marginal zone, follicular, and lymphoplasmacytic types (6-8,22). The reported odds ratios for these associations range from 2.47 to 5.2 (6,22). We found a similar strength of association between HCV and nonoropharyngeal HNCs and between HCV and HPV-positive oropharyngeal cancers. Thus, clinicians should be aware that nonliver cancers besides non-Hodgkin's lymphomas can develop in patients chronically infected with HCV. Additionally, oncologists treating patients with HNCs should consider testing patients for HCV to enable early identification and linkage of care for this infection and to prevent progression of underlying liver disease. It remains unknown whether early HCV treatment may prevent development of HCV-associated HNCs or improve the oncologic outcome of patients in whom such cancers develop as has been reported in patients with HCV-associated non-Hodgkin's lymphoma (6-8,23).
HNCs are a heterogeneous group of cancers accounting for approximately 3% of all cancers diagnosed in the United States (24). Risk factors for these cancers include smoking, alcohol consumption, chewing betel nut, African American race, low socioeconomic status, HPV infection (for oropharyngeal cancers), and EBV infection (for nasopharyngeal cancers) (13,25,26).
Little has previously been published about the relationship between HCV infection and HNCs. A few articles suggest that HCV infection may be associated with squamous cell carcinomas of the oral cavity (27,28). A US study showed a higher HCV prevalence in patients with squamous cell carcinoma of the head and neck than in noncancer patients (21.2% vs 9.9%, P < .004) (27). A population-based study from Taiwan found that the incidence of oral cavity cancers was 2.28 times as high in HCV-infected individuals as in patients with nonviral hepatitis (6.15 vs 2.69 per 100 000 person-years) and that HCV alone was a statistically significant risk factor for oral cavity cancers (HR = 1.9) (28). Recently, the US population-wide Chronic Hepatitis C Cohort Study analyzed data on 12 126 patients with chronic HCV infection and found that such patients had an increased incidence of oral cavity cancers (standardized rate ratio = 2.5) and mortality from oral cavity cancers (relative risk = 5.2) compared with noninfected patients (9). In our present study, HCV seropositivity was associated not only with nonoropharyngeal HNCs but also with HPV-positive oropharyngeal cancers. Our results also confirmed that HCV seropositivity was associated with oral cavity cancers, as previously reported (27,28).
The mechanisms by which HCV may be associated with HNCs remain unclear and need further investigation. HCV has been detected in saliva and salivary glands in patients with chronic sialadenitis and salivary gland tumors (29). Furthermore, an extrahepatic manifestation of chronic HCV infection is oral lichen planus, which is a premalignant condition associated with development of squamous cell carcinoma of the oral cavity (30). Case reports also have described progression of lichen planus to squamous cell carcinoma of the oral cavity in HCV-infected patients (30,31). Enhanced replication of HCV in oropharyngeal tissues may contribute to chronic inflammation predisposing to cancer development.
A novel finding from our study was that HCV seropositivity was associated with HPV-positive but not HPV-negative oropharyngeal cancers. HPV persists as a chronic infection, in part, by evading antiviral immune responses (32). High-risk HPV types have been shown to downregulate interferon-α-inducible gene expression (32). HPV oncoproteins E6 and E7 directly interact with and inhibit type 1 interferon signaling (32), which plays a central role in controlling intracellular replication of HCV (32). Thus, HPV may facilitate proliferation of HCV in the oropharyngeal cells, thereby facilitating its oncogenic action.
The role of HCV in oropharyngeal carcinoge
nesis may also relate to direct involvement of HCV proteins, including HCV nonstructural proteins NS5B and NS3 and HCV core proteins, in disrupting cell cycle regulation. NS5B binds the retinoblastoma tumor suppressor protein (Rb) in the cytoplasm of infected cells and recruits E6-associated protein, leading to polyubiquitination and proteasomal degradation of Rb (33). Because Rb is critical in regulating G1-S cell cycle transition, DNA damage response, mitotic spindle checkpoints, and apoptosis, loss of Rb may lead to uncontrolled cell proliferation (33). Notably, HPV protein E6 also exerts a pro-oncogenic impact by targeting tumor suppressor proteins through E6-associated protein, leading to degradation of cellular p53 and enhanced phosphorylation of Rb (34). Various interactions of HCV core and NS3 proteins with p53 have also been documented (33). The loss of p53 induced by HPV E6 protein and the degradation of Rb caused by HCV NS5B protein may play a synergistic role in the development of oropharyngeal cancers and merits further investigation. Additional studies are also warranted to examine the association between HCV and HPV-related anogenital (cervix, vagina, vulva, penis, and anus) cancers.
Another novel finding of our study was that HCV seropositivity was associated with laryngeal cancers. Risk factors previously identified for laryngeal cancers include smoking, alcohol intake, injection drug use, gastroesophageal reflux, and occupational exposures like asbestos (35,36). Of interest, recent studies have shown an association between HPV infection and laryngeal cancers (37-39). The association of HCV with laryngeal cancers may be because of mechanisms similar to those described above for HPV-positive oropharyngeal cancers. Correlation between smoking and/or drug use could partially explain the association of HCV with laryngeal cancers. However, detailed evaluation of risk factors for laryngeal cancers, including testing of cancers for HPV, was not possible because of lack of data.
Our study had several strengths. We evaluated the largest series of oropharyngeal and nonoropharyngeal HNCs for evaluating associations with HCV as separate groups because their risk factors differ. All cancers in our study population were histopathologically confirmed. The sensitivity analyses that included only subjects who had HCV infection (detectable HCV RNA) improved the validity of our findings. We evaluated smoking and alcohol consumption in detail and adjusted for them in final analyses, an approach not previously taken and often difficult through population-based databases and cancer registries. We also utilized existing data on HPV status of oropharyngeal cancers and found that HCV was associated with HPV-positive oropharyngeal cancers, suggesting oncogenic synergy between the two chronic viral infections.
Our study also had several limitations. First, the control group was not cancer free; rather, control subjects were patients with any of the three major smoking-associated cancers. In this retrospective study, we did not have access to a cancer-free population. We selected this control group so that case and control subjects were similar with respect to major risk factors-smoking, alcohol, and socioeconomic status-and adjusted for them in the final analyses. Second, since we used hospital-based control subjects, there was a possibility of Berkson's bias, which pertains to different exposure rates in the control group and the general population (40). Notably, the prevalence of HCV infection in our control group was 6%, higher than the prevalence of approximately 1.5% in the general US population (1). Thus, our results may actually underestimate the true effect size of the association between HCV seropositivity and HNCs. An alternate explanation is that the HCV seropositivity rate is higher in cancer patients than in the general population. Third, there is the possibility of selection bias as patients with solid tumors are not routinely tested for HCV antibodies in our center. However, the proportions of cancer patients seen at our institution during the study period who were tested for HCV antibodies did not differ among the case and control groups (Supplementary Table 2, available online).
Nevertheless, bias resulting from different reasons for HCV screening in case and control groups may be possible. Fourth, the association between HCV and HPV-positive oropharyngeal cancers may reflect a correlation between HCV and HPV infection because of lifestyle behaviors. Both HPV-positive oropharyngeal cancer patients and HCV-infected individuals tend to have high numbers of lifetime sex partners (41,42). However, whereas HPV is mostly transmitted sexually, sexual transmission of HCV infection is extremely rare, with the incidence being one case per 190 000 sexual contacts among heterosexual couples (43); the most frequent route of HCV transmission is injection drug use (44). Thus, this presumed correlation between HCV and HPV infection may reflect correlation between the two high-risk behaviors. Because of the retrospective nature of our study and the nonavailability of reliable assessments of sexual and drug use history, evaluation of such behaviors was not possible. Fifth, we did not have data on HCV treatment in our study subjects. Because population-based studies have shown that most HCV-infected patients are not aware of their infection and healthcare utilization rates are low for patients with HCV infection (42), we believe that lack of HCV treatment data is unlikely to affect our results. Sixth, more than 50% of the patients in our study were born between 1945 and 1965; individuals born during that period have a high prevalence of HCV infection (1). However, because the recommendations for screening such individuals were made in 2012 (1) and our study included patients tested between 2004 and 2014, we believe that birth between 1945 and 1965 is unlikely to have been the reason for HCV testing for the majority of patients in our study population. Nevertheless, our results were adjusted for the birth cohort effect in our multivariable analyses. Seventh, even though we tried to adjust for smoking, alcohol intake, and socioeconomic status, data were obtained from self-reported questionnaires and underreporting or misreporting of information along with residual confounding is possible. Eighth, the case-case analyses showed no etiologic heterogeneity between the case groups compared with respect to HCV seropositivity; however, the number of patients in comparison groups was small to determine differences. Finally, we were unable to evaluate the association between HCV and nasopharyngeal cancers based on EBV status because of small sample size.
In conclusion, HCV seems to be associated with HNCs, particularly nonoropharyngeal and HPV-positive oropharyngeal cancers. Validation of these associations by analysis of population-based datasets and cancer-free control group is essential. Further studies are also required to explore the possible interaction between HCV and HPV and the association between HCV and other HPV-related malignancies.
|
|
| |
| |
|
|
|
