Tobacco Tobacco initiates a linear dose-response carcinogenic effect in which duration is more important than the intensity of exposure. The major carcinogenic activity of cigarette smoke resides in the particulate (tar) fraction, which contains a complex mixture of interacting cancer initiators, promoters, and cocarcinogens. In the late 1950s, a landmark case-control study by Dr. Ernst Wynder established the link between tobacco use and oral cavity cancer This was followed a year later by a cohort study of over 180,000 men which demonstrated an increased risk of death due to HNSCC in cigarette smokers as compared to men who never smoked. These studies also demonstrated elevated risks for HNSCC death in cigar and pipe smokers. In 1964, the Advisory Committee to the Surgeon General on Smoking and Health published its report linking smoking to cancer based on many of Hill’s classic criteria of disease causality. Due to limitations of sample size and follow-up time, Hill’s classic cohort study of over 40,000 British physicians showed only a borderline risk of HNSCC related to smoking. However, his criteria linking HNSCC and tobacco smoking have been clearly demonstrated over the past 40 years in multiple independent studies. Most importantly, the strength and consistency of the association between smoking and HNSCC have been demonstrated in numerous case-control and cohort studies with significant relative risks or odds ratios in the 3- to 12-fold range. Furthermore, a dose-response effect is consistently shown in these studies between the duration and dose of smoking with increasing risk of HNSCC and between the time since quitting and the decreasing risk of HNSCC. Other mucosal malignancies of the head and neck such as nasopharyngeal carcinoma and sinonasal malignancies have a weaker association with tobacco smoking. The specificity of the link between tobacco and HNSCC (and not the nonmucosal/unexposed head and neck malignancies), the coherence and analogy of the explanation of tobacco-induced HNSCC to lung carcinogenesis, and the biologic plausibility of the well-established tobacco-induced carcinogenesis model have all helped establish tobacco as the chief etiologic agent in HNSCC.
Although the risk of bronchogenic carcinoma appears to be less significant for cigar and pipe smokers than for cigarette smokers, these forms of tobacco use are also clearly associated with an increased risk of HNSCC. The pooling of saliva containing carcinogens in gravity-dependent regions may account for the frequent location of oral carcinomas along the lateral and ventral surfaces of the tongue and in the floor of the mouth. Smokeless tobacco use has also demonstrated an association with cancer of the oral cavity. Smokeless tobacco users and pipe smokers who have a habitual position for the quid or pipe stem often develop carcinomas and dysphasia of the specific site of use, which suggests that physical and thermal trauma may be contributing factors.
While smoking rates are declining in the developed world, smoking rates are rising in developing world, home to four-fifths of the world’s population. In the United States, smoking rates have declined since the Surgeon General’s warning in 1964. In 1965, 42.4% of the US adult population were current smokers, while in 1999 only 23.5% were current smokers. While the reduction in cigarette smoking has been much greater in men over the last three decades, the rate of current cigarette use remains higher in men (25.7%) than for women (21.5%), and 40.8% of Native Americans continue to smoke. Other concerns have been the increasing rates of cigarette smoking in high school seniors and the dramatic increase in the number of new cigar users in the last decade. Dramatic increases in smokeless tobacco use have occurred among younger people and has been implicated by some in the rise of oral cancer mortality rates in this group. Striking variations in head and neck cancer sites and incidence seen among different regions, cultures, and demographic groups, are due in large part to differing patterns of tobacco and other substance abuse. Smokeless tobacco and similar products are used greatly in parts of Asia and Africa. In South Central Asia, “pano” (betel leaf, lime, catechu, and areca nut) is commonly chewed and is a strong risk factor independent of tobacco use for carcinoma of the oral cavity, one of the most common cancers in men and women in this region.
Alcohol, too, is an important promoter of carcinogenesis and is a contributive factor in at least 75% of HNSCCs. Furthermore, alcohol appears to have an effect on risk of HNSCC independent of tobacco smoking, but these effects are consistently significant only at the highest level of alcohol consumption. While studies attempting to correlate the type of alcoholic beverage with specific cancer risks have been conflicting, most investigators believe that ethanol itself is the main causative factor. Nevertheless, it appears that the major clinical significance of alcohol consumption is that it potentiates the carcinogenic effect of tobacco at every level of tobacco use. However, this effect is most striking at the highest levels of exposure, and the magnitude of this effect is at least additive but may be multiplicative, dependent on the subsite of HNSCC and the levels of exposure.
The predominant risk factor for HNSCC is a history of exposure to tobacco and alcohol. However, since only a fraction of smokers develop cancer, variations in genetic susceptibility may be equally important in the disease etiology. A genetic component to this disease is also supported by large family studies demonstrating a three- to eightfold increased risk of HNSCC in first-degree relatives of patients with HNSCC. Furthermore, there is molecular epidemiologic evidence supporting the concept of genetic susceptibility in head and neck cancer patients. Emerging data from case-control studies of several phenotypic and genotyping assays support the hypothesis that genetic susceptibility plays an important role in the etiology of HNSCC. According to this hypothesis, inherited differences in the efficiencies of carcinogen metabolizing systems, deoxyribonucleic acid (DNA) repair systems, and/or cell cycle control/apoptosis systems influence one’s risk for tobacco-induced cancers. Identifying such at-risk individuals in the general population, by use of these biomarker assays, would have a profound impact on primary prevention, early detection, and secondary prevention strategies.
While it has been suggested that various infectious agents play a role in head and neck carcinogenesis, only Epstein-Barr virus (EBV) and human papilloma virus (HPV) can be implicated as etiologic agents in head and neck carcinogenesis based on current scientific evidence. EBV appears to be associated with most nasopharyngeal carcinomas, and HPV (most commonly type 16) is associated with approximately 50% of oropharyngeal carcinomas. HPV may also play a role in the etiology of squamous cell carcinomas arising in the sinonasal tract. While herpes simplex viruses have been suggested as a risk factor for oral cavity cancer and Helicobacter pylori has been suggested as a risk factor for laryngeal cancer, confirmation of these findings is lacking.
While laboratory evidence supporting the role of HPV as a risk factor for HNSCC is largely circumstantial, HPV has been established as an etiologic agent in cervical cancer, and more recently, several investigators have suggested that infection with HPV, especially the high-risk type HPV-16, is a risk factor for HNSCC. The chief oncoproteins of HPV-16 are encoded by the genes E6 and E7. The E6 protein targets the tumor suppressor gene p53 for ubiquitination and degradation. In fact, degradation of p53 in HPV-positive cells is fully dependent on the presence of E6. The E7 oncoprotein is involved in suppression of pRb function. Reduced pRb expression is common in HPV-positive tonsillar cancers. In vitro experiments support the tumorgenicity of HPV-16 in human epithelial cells. Furthermore, numerous studies using methods such as polymerase chain reaction (PCR), Southern blotting, and in situ hybridization have detected HPV DNA in the tumor tissue and sera of HNSCC patients. Oropharyngeal tumors and tumors in nonsmokers are the most frequently positive. However, because they did not include cancer-free controls, most of these studies were unable to estimate the risk of HNSCC attributable to HPV-16. Molecular epidemiologic evidence with case-control design supporting the role of HPV-16 in HNSCC has been emerging. These studies have assessed HPV-16 DNA positivity in the mucosa of cancer-free controls. In a study of oral cancer patients and cancer-free controls, HPV DNA was identified in 15% of 93 oral cancer specimens and in only 5% of 205 cancer-free control benign oral specimens (odds ratio[OR] = 3.70; 95% CI, 1.47-9.32). A second study found only a slight difference in high-risk HPV DNA identified in 44 cases with laryngeal cancer and 10 with leukoplakia specimens and 12 cancer-free control benign specimens (18.2%, 20.0%, and 16.7%, respectively). In a third study of 52 patients with tonsillar cancer and 48 age- and sex-matched cancer-free controls undergoing tonsillectomy for benign pathology, 46% of tonsillar carcinomas and only 6.3% of the cancer-free control tonsils were HPV positive. While these studies were relatively small, a recent nested case-control study of 292 cases and 1,569 controls from a Scandinavian cohort of almost 900,000 subjects (Mork and colleagues) reported that HPV-16 seropositivity was associated with a 2.2-fold increased risk of HNSCC after multivariate adjustment. However, 25% of the cases in that study were not classic HNSCC as 3% were nasopharyngeal cancers, 20% were lip cancers, and 2% were sinus cancers. Because subjects with cancers of these sites were less frequently HPV-16 seropositive, Mork and colleagues may have underestimated the risk of HNSCC associated with HPV-16 exposure. In their subgroup analysis of oropharyngeal cancer, Mork and colleagues reported an estimated risk of 14.4 (95% CI, 3.6-58.1). An additional serologic base study using a population-based case-control study design of 284 subjects with newly diagnosed oral and oropharyngeal carcinoma and 477 cancer-free controls demonstrated a significantly elevated risk associated with HPV-16 seropositivity (adjusted OR = 2.3; 95% CI, 1.6-3.3). While subgroup risk analysis for oropharyngeal cancer was not performed, the chief effect again appears to be in the oropharyngeal cancer subgroup, where approximately 35% had cancers positive for HPV-16 DNA. The circumstantial, mechanistic, and molecular epidemiologic evidence strongly supports the role of HPV-16 infection in oropharyngeal carcinogenesis.
The epidemiologic link between EBV and nasopharyngeal cancer is quite strong. Regardless of histopathologic subtype (World Health Organization [WHO] I to III), geographic or ethnic setting, sporadic or endemic pattern, and premalignant or malignant status, nasopharyngeal carcinoma is an EBV-associated malignancy. While WHO types II and III are overwhelmingly positive for EBV, EBV is also associated with well-differentiated (WHO type I) nasopharyngeal carcinoma. Both serologic and mucosal swab evidence of EBV infection has been used to enhance the screening for nasopharyngeal carcinoma in endemic areas. Evidence of EBV DNA in cervical lymph node metastases of unknown primary origin has been used to identify a nasopharyngeal primary. More recently, the detection of EBV DNA in peripheral blood (both cellular and cell-free component) has demonstrated prognostic significance for predicting survival and distant metastases. Nasopharyngeal carcinoma may be the best example of a virus-related epithelial carcinoma and has served as a model for the study of virus-induced carcinogenesis elsewhere in the body. The fundamental and early role of EBV in the pathogenesis of NPC is suggested by the recent finding of EBV DNA in premalignant nasopharyngeal lesions.
Environmental Tobacco Smoke
A recent high-profile legal case in Australia has brought significant attention to the risk of HNSCC secondary to environmental tobacco smoke exposure. In May of 2001, the New South Wales Supreme Court found that a 62-year-old nonsmoker’s HNSCC was significantly associated with long-term exposure to environmental tobacco smoke in her job as a bar attendant, and imposed liability on her employer. Two case-control studies supported the court’s finding. In a study of 173 cases of HNSCC and 176 cancer-free controls, environmental tobacco smoke was associated with a more than twofold increased risk of HNSCC, and a dose-response relationship was also observed. In a separate study of 44 nonsmokers with HNSCC and 132 cancer-free nonsmoker controls, environmental tobacco smoke was associated with a significantly increased risk for HNSCC (OR = 5.34), and this was particularly true for females (OR = 8.00) and for those reporting exposure at work (OR = 10.16).
Observational and anecdotal studies have long suggested that gastroesophageal reflux may be associated with laryngeal cancer. Furthermore, multiple studies have objectively documented a high prevalence of gastric reflux into the laryngopharynx by 24-h pH probe monitoring. Recently, a retrospective case-control study of 10,140 hospitalized patients and 12,061 outpatients with laryngeal and pharyngeal cancer and 40,561 hospitalized and 48,244 outpatient controls was performed using computerized hospital and outpatient databases of the US Department of Veterans Affairs. The diagnosis of gastroesophageal reflux disease was associated with a significantly elevated risk of laryngeal cancer (OR = 2.40 [2.15-2.69] and OR = 2.31 [2.10-2.53] for hospitalized and outpatient groups, respectively) and of pharyngeal cancer (OR = 2.38 [1.87-3.02] and OR = 1.92 [1.72-2.15] for hospitalized and outpatient groups, respectively). Furthermore, these risk estimates were adjusted for age, gender, ethnicity, smoking, and alcohol consumption.
Marijuana smoke has a four-times higher tar burden and a 50% higher concentration of benzopyrene and aromatic hydrocarbons than does tobacco smoke. While anecdotal evidence has long suggested that marijuana is a risk factor for HNSCC, few reports have found direct evidence of marijuana as an etiologic factor for HNSCC because most users of marijuana are also exposed to tobacco and alcohol. A recent case-control study including 173 HNSCC patients and 176 cancer-free controls demonstrated a cigarette-adjusted risk for HNSCC of 2.6 (95% CI, 1.1-6.6) associated with marijuana use, with evidence of a dose-response relationship. However, a large retrospective cohort of 64,855 health maintenance organization (HMO) members found no association with tobacco-related cancers.
Epidemiologic evidence from traditional case-control study designs suggests that diets high in animal fats and low in fruits and vegetables may be risk factors for HNSCC. Winn and colleagues found that the risk of oral and pharyngeal cancer in women was inversely related to the consumption of fresh fruits and vegetables. Similarly, in a study of 871 individuals with oral and pharyngeal cancer and 979 cancer-free controls, McLaughlin and colleagues demonstrated an inverse relationship between fruit intake and oral and pharyngeal cancer risk. In a Chinese population-based case-control study, citrus fruits and dark green/yellow vegetables were associated with a decreased risk of laryngeal cancer whereas salted fish and meat, as well as deep-fried foods, were associated with an increased risk of pharyngeal and laryngeal cancer. More recently, both European and US studies have confirmed the protective effects of fruits and vegetables and the risk of animal fat consumption, after adjustment for smoking and alcohol use. Several studies have used case-control methodology to correlate salted fish consumption with nasopharyngeal carcinoma risk, and this risk may be due to the high content of nitrosamine compounds in preserved foods such as salted fish.
Some evidence suggests vitamin A and beta-carotene may be responsible for the protective effect of diets high in fruits and vegetables, and deficiencies of carotenoids appear to be a risk factor for HNSCC and lung cancers. It is not known, however, which of the more than 500 carotenoids are protective, what chemical interactions may occur, or what protective role other micronutrients in carotenoid-rich foods may play. Others have found that total intake of vitamins C and E are also protective. Moreover, diets are complex and difficult to assess and validate; in particular, there are often inaccuracies in translating foods into constituent nutrients. Further studies are needed to more precisely define the relationship between dietary intake and serum levels of the various carotenoid components. It may be impossible to determine which of the vast array of compounds is most beneficial, and controlling for other dietary variables and confounding risk factors has remained a difficult methodologic problem. Further confounding this situation, smoking has been associated with reduced dietary intake and serum levels of carotenoids. Despite these many problems, prospective and retrospective nutritional (serum and dietary) epidemiologic studies have provided important clues to the development and prevention of these cancers.
Although occupational exposures probably play a minor role overall in the development of HNSCC, they are major risk factors for malignancies of the sinonasal region. The most important exposures occur in the metalworking and refining, woodworking, and leather/textile industries. While asbestos has been implicated as a risk factor for laryngeal cancer, most of the evidence does not support this association.
A strong association has not been demonstrated between ionizing radiation and the development of HNSCC. However, squamous cell carcinomas of the lip, like skin cancers, are associated with ultraviolet radiation exposure. Furthermore, exposure to gamma radiation is associated with thyroid cancers, sarcomas of the head and neck, and salivary gland malignancies, including paranasal sinus cancers. While therapeutic irradiation of head and neck malignancies does not appear to induce second primary cell carcinomas of the aerodigestive tract, it is associated with an increased risk for sarcomas of the head and neck. This is a particular concern for children who have received therapeutic irradiation. Furthermore, environmental, medical diagnostic, and therapeutic radiation exposure to the head and neck are all significantly associated with salivary gland malignancies. These studies have found a significant dose-response relationship with risk increasing with dose, and mucoepidermoid carcinomas appear to be the most common radiation-induced malignancy. In the past, the uses of radium for watch dial painting and thoratrast contrast in sinonasal imaging were both associated with an increased risk of paranasal sinus malignancies.
Revision date: July 6, 2011
Last revised: by Janet A. Staessen, MD, PhD