Smokeless tobacco and increased risk of hypopharyngeal and laryngeal cancers: A multicentric case–control study from India
Article first published online: 21 JUN 2007
Copyright © 2007 Wiley-Liss, Inc.
International Journal of Cancer
Volume 121, Issue 8, pages 1793–1798, 15 October 2007
How to Cite
Sapkota, A., Gajalakshmi, V., Jetly, D. H., Roychowdhury, S., Dikshit, R. P., Brennan, P., Hashibe, M. and Boffetta, P. (2007), Smokeless tobacco and increased risk of hypopharyngeal and laryngeal cancers: A multicentric case–control study from India. Int. J. Cancer, 121: 1793–1798. doi: 10.1002/ijc.22832
- Issue published online: 14 AUG 2007
- Article first published online: 21 JUN 2007
- Manuscript Accepted: 26 MAR 2007
- Manuscript Received: 5 MAR 2006
- International Agency for Research on Cancer (IARC)
- smokeless tobacco;
- chewing tobacco;
- hypopharyngeal cancer;
- laryngeal cancer;
- betel quid;
Hypopharyngeal and laryngeal cancers are among the most common cancers in India. In addition to smoking, tobacco chewing may be a major risk factor for some of these cancers in India. Using data from a multicentric case–control study conducted in India that included 513 hypopharyngeal cancer cases, 511 laryngeal cancer cases and 718 controls, we investigated smoking and chewing tobacco products as risk factors for these cancers. Bidi smoking was a stronger risk factor compared to cigarette smoking for cancer of the hypopharynx (ORbidi 6.80 vs. ORcig 3.82) and supraglottis (ORbidi 7.53 vs. ORcig 2.14), while the effect of the 2 products was similar for cancer of the glottis (ORbidi 5.32 vs. ORcig 5.74). Among never-smokers, tobacco chewing was a risk factor for hypopharyngeal cancer, but not for laryngeal cancer. In particular, the risk of hypopharyngeal cancer increased with the use of Khaini (OR 2.02, CI 0.81–5.05), Mawa (OR 3.17, CI 1.06–9.53), Pan (OR 3.34, CI 1.68–6.61), Zarda (OR 3.58, CI 1.20–10.68) and Gutkha (OR 4.59, CI 1.21–17.49). A strong dose-response relationship was observed between chewing frequency and the risk of hypopharyngeal cancer (ptrend < 0.001). An effect of alcohol on cancer of the hypopharynx and supraglottis was observed only among daily drinkers (OR 2.22, CI 1.11–4.45 and OR 3.76, CI 1.25–11.30, respectively). In summary, this study shows that chewing tobacco products commercially available in India are risk factors for hypopharyngeal cancer, and that the potency of Bidi smoking may be higher than that of cigarette smoking for hypopharyngeal and laryngeal cancers. © 2007 Wiley-Liss, Inc.
An increasing trend in mortality from head and neck cancers has been observed in Europe and Asia. In India alone, over 75,000 people are diagnosed with pharyngeal and laryngeal cancers each year.1 Despite advancements in treatment protocols, the 5-year survival rate still remains around 50%.2 The major risk factors for pharyngeal and laryngeal cancers are tobacco smoking,3, 4, 5, 6 alcohol drinking7, 8, 9, 10, 11, 12 and tobacco chewing.5, 6, 13, 14, 15
Tobacco chewing is a common habit in India, which has been growing consistently over the past few decades owing to successful marketing and packaging that allows for easier use.16 More alarmingly, its popularity continues to grow, especially among the most vulnerable populations, such as children, teenagers and pregnant women.16 There are various forms of chewing tobacco that are used in India,5, 16, 17 of which Pan is one of the most common. Pan is a mixture of areca nut, catechu (areca catechu), slaked lime, tobacco and additional spices, wrapped in a betel leaf.5, 16, 17 Although Pan can also be made without tobacco, most habitual chewers in India tend to include tobacco.17 Other chewing tobacco products commonly used in India include Khaini (a mixture of tobacco and slaked lime), Mawa (tobacco, areca nut and slaked lime), Gutkha (tobacco, catechu, areca nut and slaked lime) and Zarda (tobacco and slaked lime).5 Although not as common as chewing, snuffing of tobacco products represents an additional method of consuming smokeless tobacco products in India. This includes oral snuffing as well as nasal snuffing. The most common snuffing product in this region, called Naswar, is a mixture of tobacco and slaked lime.5
Recently, betel quid with and without tobacco, along with areca nut, has been classified as a known human carcinogen by the International Agency for Research on Cancer,5 with increased risk observed for cancers of the pharynx and esophagus. However, limited data exist on other chewing products commercially available in India. Similarly, the role of snuffing products in hypopharyngeal and laryngeal cancer is not clear. Using data from a multicentric case–control study conducted in Ahmedabad, Bhopal, Chennai and Kolkata, we report the risks of hypopharyngeal and laryngeal cancers associated with smoking, snuffing and chewing different tobacco products used in India.
A multicentric case–control study was conducted in India between 2001 and 2004. The 4 participating centers were the Gujarat Cancer and Research Institute in Ahmedabad, the Gandhi Medical College in Bhopal, the Chittaranjan National Cancer Institute in Kolkata and the Cancer Institute (WIA) in Chennai. Altogether, 1,062 head and neck cancer cases, and 718 controls matched on age (±5 years), sex and geographical area of residency were recruited. Overall, 19% of the controls were hospital-based (patients with disease not related to alcohol or tobacco consumption) and 81% were visitors to patients at the hospital. A standardized questionnaire was administered to all study participants by trained staff members, who collected data on demographic and socioeconomic status, clinical history, family history of cancer, tobacco and alcohol consumption habits, dietary factors, occupation, residential history and usage of different chewing products available locally.
Of the 1,062 head and neck cancer cases, 38 (3.6%) were excluded from the analysis because the histological subtype was missing, the cancer was an in situ carcinoma, or they were cancers other than that of hypopharyx or larynx. Of the 1,024 eligible cases, 513 were hypopharyngeal cancer cases (ICDO-2 codes C12 and C13) and 511 were laryngeal cancer cases (ICDO-2 codes: C32.0 (glottis = 178), C32.1 (supraglottis = 120) and C32.2–C32.9 (other larynx = 213)). Of the 213 other laryngeal cancer cases, the majority (84%) had a cancer in an unspecified part of the larynx (ICDO-2 code 32.9) and the remaining 16% were characterized by a combination of overlapping regions of the larynx and subglottis (ICDO-2 codes C32.8, C32.2 and C32.3). Separate analyses were conducted for different sites within the larynx (glottis, supraglottis and other larynx) whenever possible, except in instances where the numbers were too small. In such instances, an overall analysis was conducted for laryngeal cancer that combined the aforementioned 3 subcategories. Altogether, 90% of the 1,024 cases were squamous cell carcinomas.
Ever-smokers were defined as individuals who smoked at least 50 cigarettes over a 6-month period, while ever-chewers and ever-drinkers were defined as those who chewed tobacco products or drank alcohol at least once a week for a minimum of 6 months. Cumulative tobacco consumption was calculated after assigning a cigarette-equivalent value of 0.5, 1 and 2 to 1 Bidi, cigarette and cigar/cheroot, respectively,3, 6, 18 and multiplying the number of cigarette-equivalents by the years of smoking. To create a composite socioeconomic status (SES) variable, we assigned a score of 1–5 for level of education; monthly family income and crowdedness at home (number of people per room). A composite SES variable was created by summing up the score and dividing it into distinct categories. For the product specific analysis of chewing tobacco, individuals were assigned to the product they reported using for the longest duration, if they used more than 1 chewing products.
Statistical analyses were performed using STATA, version 8 (Stata, College Station, TX). Odds ratios (ORs) and the corresponding 95% confidence intervals (CIs) for each risk factor under consideration were estimated using unconditional logistic regression models after adjusting for age (continuous), sex, participating center, SES (categorical) and cumulative tobacco consumption (pack years, continuous). Tests for linear trends were performed by treating the categorical variables as continuous predictors in the logistic regression models.
The majority of participants in the study were men as shown in Table I. The proportion of female cases ranged from 4.5% for cancer of the glottis to 16.2% for hypopharyngeal cancer, compared to 15.5% among controls. In general, controls were of slightly higher SES status compared to cases, while the majority of both cases and controls were Hindus.
Ever smoking was associated with increased risks of all types of cancer considered (Table II), with odds ratios ranging from 5.35 (glottis) to 8.28 (other larynx). Analyses based on types of tobacco product smoked showed that Bidi smokers may have a higher risk of hypopharyngeal cancer (OR 6.80, CI 4.64–9.97) compared to cigarette smokers (OR 3.82, CI 2.32–6.29). Similar results were observed for cancer of the supraglottis (ORbidi 7.53, CI 3.84–14.74, ORcigarette 2.14, CI 0.63–7.30). In contrast, comparable risk estimates were observed among cigarette smokers and Bidi smokers for cancer of the glottis (ORbidi 5.32, CI 3.18–8.90, ORcigarette 5.74, CI 3.20–10.31). A strong dose response relationship was observed, based on duration as well as frequency of Bidi smoking for all types of cancer considered (ptrend < 0.001). Similar dose response relationships were also observed with cigarette smoking for cancer of the hypopharynx and glottis.
|Cigerette and bidi||33||23||4.74||(2.42–9.29)||11||5.42||(2.37–12.39)||9||15.92||(5.46–46.45)||17||9.52||(4.15–21.85)|
|Frequency of cigarette smoking|
|1–10 per day||79||24||2.47||(1.34–4.54)||24||3.74||(1.96–7.13)||2||1.33||(0.27–6.65)||8||2.57||(1.05–6.32)|
|>10–20 per day||19||15||5.24||(2.19–12.53)||12||6.63||(2.60–16.92)||1||4.28||(0.45–40.72)||10||10.95||(3.94–30.40)|
|>20 per day||6||9||9.75||(2.95–32.22)||1||1.94||(0.20–18.58)||1||6.55||(0.46–92.57)||2||7.05||(1.12–44.46)|
|Duration of cigarette smoking|
|Frequency of bidi smoking|
|1–10 per day||62||55||2.83||(1.73–4.63)||15||2.01||(0.96–4.20)||14||3.07||(1.29–7.33)||20||2.82||(1.40–5.67)|
|>10–20 per day||25||73||9.45||(5.32–16.78)||18||5.31||(2.42–11.63)||16||7.25||(2.77–18.94)||38||11.94||(5.74–24.84)|
|>20 per day||26||150||17.13||(9.97–29.44)||54||14.59||(7.47–28.50)||53||20.31||(9.10–45.31)||71||24.71||(12.56–48.61)|
|Duration of bidi smoking|
An increased risk of hypopharyngeal cancer (Table III) was observed among tobacco snuffers (OR 2.25, CI 0.99–5.13). When the analysis was restricted to never smokers, the risk was still evident (OR 2.85, CI 1.15–7.08). In contrast, there was no increased risk of laryngeal cancers associated with tobacco snuffing. A moderate increase in the risk of hypopharyngeal cancer was observed among chewers of both nontobacco (OR 1.95, CI 0.96–3.97) and tobacco (OR 1.51, CI 1.08–2.11) products. When the analysis was restricted to never-smokers, the risk associated with chewing nontobacco product was not apparent (OR 1.21, CI 0.39–3.78), while that associated with chewing tobacco products increased in magnitude (OR 3.18, CI 1.92–5.27). In contrast, no association was observed between tobacco chewing and the risk of laryngeal cancer (OR 0.75, CI 0.52–1.70 for all individuals, OR 0.95, CI 0.52–1.73 for never smokers).
|Snuffing||Chew nontobacco prd.||Chew tobacco prd.||CTRI||Hypopharynx||Larynx|
|Case||OR||95% CI||Case||OR||95% CI|
A more detailed analysis was conducted to evaluate hypopharyngeal and laryngeal cancer risk associated with various chewing products commonly used in India (Table IV). An increased risk of hypopharyngeal cancer was observed among those who reported chewing Zarda (OR 2.23, CI 1.11–4.50). In general, no increased risk of laryngeal cancer was observed for the individual chewing products. When the analysis was restricted to never-smokers, an increased risk of hypopharyngeal cancer was observed for all individual tobacco products considered. The product specific ORs were: 2.02 (CI 0.81–5.05) for Khaini, 3.17 (CI 1.06–9.53) for Mawa, 3.34 (CI 1.68–6.61) for Pan, 3.58 (CI 1.20–10.68) for Zarda and 4.59 (CI 1.21–17.49) for Gutkha, respectively. In contrast, no product specific risks were observed for laryngeal cancers among never smokers (Table IV). Analysis based on frequency of tobacco chewing showed that increasing frequency of chewing was associated with increasing risk of hypopharyngeal cancer (ptrend = 0.017), with a stronger dose response observed among never-smokers (ptrend < 0.001). However, such a relationship was absent for laryngeal cancers.
|Chewing tobacco products|
|No. of tob prod. chewed/day|
|Never smokers onlyb,d|
|Chewing tobacco products|
|No. of tob prod. chewed/day|
In this study population, increased risk of hypopharyngeal and laryngeal cancers were not observed among ever drinkers of alcoholic beverages (Table V). When frequency of alcohol consumption was considered, an increased risk of the cancers of supraglottis (OR 3.76, CI 1.25–11.30) and hypopharynx (OR 2.22, CI 1.11–4.45) were observed among those who reported consuming alcohol daily, with a clear dose response for cancer of the supraglottis (ptrend = 0.02). However, duration of alcohol consumption was not associated with increased risk, except for other laryngeal cancers, where long duration (≥20 years) of consumption was associated with a moderate increase in risk (OR1.63, CI 1.00–2.69).
|(All)||CS||OR||95% CI||CS||OR||95% CI||CS||OR||95% CI||CS||OR||95% CI|
|<Once a week||60||23||0.47||(0.26–0.87)||18||1.02||(0.53–1.95)||7||0.91||(0.33–2.54)||13||0.81||(0.38–1.70)|
The growing use of chewing tobacco products in India is alarming and represents a major public health concern. Every year, over 160,000 people are diagnosed with UADT cancers in India,1 and these numbers will continue to grow in the foreseeable future. Therefore, understanding the risks associated with alcohol drinking as well as chewing and snuffing of tobacco products, in addition to smoking, is important and dissipating such knowledge to general public to avoid future burden of these disease is even more critical.
In this study population, the strength of association for cigarette smoking varied by cancer site, with ORs ranging from 2.14 for supraglottis to 5.74 for glottis. Similarly, odds ratios for Bidi smoking ranged from 5.32 for cancer of glottis to 9.61 for other laryngeal cancer. The highest differences in ORs between cigarette and Bidi smoking was observed for cancer of the supraglottis (ORcigarette 2.14 vs. ORbidi 7.56).
Tobacco snuffing was an independent risk factor for hypopharyngeal cancer, with higher magnitude of risk observed among never-smokers (Table III). There was no clear association between tobacco snuffing and risk of laryngeal cancer, even among the never-smokers. While chewing nontobacco products appeared to be a risk factor for hypopharyngeal cancer, the apparent lack of risk observed in never-smokers indicates that the observed risk in the overall population may be due to residual confounding from smoking. In contrast, chewing tobacco product was an independent risk factor for hypopharyngeal cancer, as evidenced by the increased risk observed among overall study population as well as among never-smokers, with a clear dose response relationship.
However, there was no increased risk of laryngeal cancer associated with tobacco chewing for all individuals as well as among never smokers. Based on the analysis conducted among never-smokers, all types of chewing products containing tobacco were risk factors for hypopharyngeal cancer in this study population (Table IV). The magnitude of risk was comparable across the various chewing products, with the odds ratios ranging from 2.02 for Khaini to 4.61 for Gutkha. The risk associated with chewing various tobacco products was further substantiated by the dose response relationship observed for frequency of tobacco chewing and the risk of hypopharyngeal cancer, among never-smokers. There was no apparent increase in risk of laryngeal cancer associated with the individual chewing products in this study. This was true even when the analysis was restricted among never smokers.
In this study population, we observed a consistent risk of hypopharyngeal cancer associated with tobacco chewing and tobacco snuffing. Equally consistent was the lack of risk of laryngeal cancer associated with these 2 habits. The lack of risk of laryngeal cancer observed in this study, is consistent with what has been reported previously.19, 20 However, additional mechanistic studies are needed to understand why chewing tobacco product is such a strong risk factor for hypopharyngeal but not for laryngeal cancers, considering the close proximity of the 2 sites. However, given the consistent results, a plausible hypothesis is that a direct and prolonged contact is necessary for the effect of chewing tobacco to manifest.
Alcohol consumption has been linked to increased risk of laryngeal and pharyngeal cancer in India previously.20, 21 In this study, an increased risk of hypopharyngeal cancer and cancer of supraglottis were observed only among those who drank daily, after adjusting for smoking and tobacco chewing. However, the lack of association observed among individuals who drink less than daily as well as the duration of drinking is not clear.
There are several strengths of this study including the large sample size and multicentric study design. This study was conducted in areas with a high prevalence of the main exposures of interest (tobacco chewing), as well as the outcomes of interest (hypopharyngeal and laryngeal cancer), and provides product specific risk estimates for the first time. In addition, very few studies have explored the relationship between snuffing habits and the risk of hypopharyngeal and laryngeal cancer in India. A potential limitation of this study is the exposure misclassification between various chewing tobacco products. Individuals may not always recall the different type of tobacco products they consumed over their lifetime accurately, unless their consumption pattern has not changed considerably. But the misclassification between different chewing tobacco products is likely to be nondifferential, thus attenuating the odds ratios. Another potential concern includes residual confounding by smoking, which we tried to address by restricting the analysis to never-smokers, whenever possible. Additionally, the proportion of female study participants is rather low (∼12%) in this study, which excluded the possibility of restricting the analyses to women only, to evaluate gender differences in risks associated with the habit of tobacco chewing.
Results from this study suggest that tobacco chewing is an independent risk factor for hypopharyngeal cancer, in the absence of smoking and drinking. Similarly, tobacco snuffing is also a risk factor for hypopharyngeal cancer. Although chewing tobacco is a common habit in India and many South Asian countries, it is also prevalent in the other countries with migrant communities arising from these areas. The popularity of the smokeless tobacco product is growing in the North American youths as well, owing to the public usage of such products by social models such as professional athletes.22, 23 The increasing usage of smokeless tobacco products combined with the ill-perceived notion that it is a relatively safe product compared to cigarettes, may pose a substantial threat to public health in the coming years. These data in conjunction with other evidence show that tobacco is dangerous in any form.24 Public health practitioners throughout the world need to recognize this and implement proper regulatory approaches before the usage of such tobacco products parallels that of south Asia.
The analysis reported in this paper was undertaken during the tenure of a Special Training Fellowship from the International Agency for Research on Cancer awarded to Dr. Sapkota and Visiting Scientist Award by IARC to Dr. Gajalakshmi. We would like to acknowledge all the collaborators in India including Dr. U. Chattopadhyay, Dr. Sukta Das and the late Dr. U. Sen of Chittaranjan National Cancer Research, Kolkata.
- 1GLOBOCAN 2002: Cancer incidence, mortality and prevalence worldwide (IARC Cancer Base No. 5(2.0)). Lyon: IARC Press, 2004., , , .
- 3International Agency for Research on Cancer. IARC monographs on the evaluation of carcinogenic risks of chemicals to humans, vol. 38: Tobacco smoking. Lyon, France: IARC Press, 1986.
- 4International Agency for Research on Cancer. IARC monographs on the evaluation of the carcinogenic risk of chemicals to humans, vol. 83: Tobacco smoke and involuntary smoking. Lyon, France: IARC Press, 2004.
- 5International Agency for Research on Cancer. Betel-quid and areca-nut chewing and some areca-nut derived nitrosamines. IARC Monogr Eval Carcinog Risks Hum 2004; 85: 1–334.
- 12International Agency for Research on Cancer. Fruit and vegetables. In: VainioH, BianchiniF, eds. Lyon, France: IARC, 2003. IARC handbooks of cancer prevention, vol. 8.
- 15Cancer of the lung larynx and pleura. In: AdamiHO, HunterD, TrichopoulosD, eds. Text book of cancer epidemiology. New York: Oxford University Press, 2002. 248–80., .