Risk of betel chewing for oesophageal cancer in Taiwan

Among 104 cases of squamous-cell oesophageal carcinoma patients and 277 controls in Taiwan, after adjusting for cigarette smoking, alcohol consumption, and other confounders, we found that subjects who chewed from 1 to 495 betel-year and more than 495 betel-years (about 20 betel quid per day for 20 years) had 3.6-fold (95% Cl = 1.3–10.1) and 9.2-fold risk (95% Cl = 1.8–46.7), respectively, of developing oesophageal cancer, compared to those who did not chew betel. © 2001 Cancer Research Campaign http://www.bjcancer.com


Selection of cases and controls
Cases were histologically confirmed oesophageal squamous-cell carcinoma from the Department of Chest Surgery at National Taiwan University Hospital in Taipei, Taiwan. The Department of Preventive Medicine at this hospital chose 1 to 3 control subjects without malignancy matched on period of hospitalization age (± 3 y) and gender. In total, 104 cases (94 males and 10 females) and 277 controls (256 males and 21 females) were recruited for interview between July, 1996 andOctober 2000. Subjects were interviewed by a trained interviewer using a standardized questionnaire. Informed consent was obtained from all subjects. Information on habitual substance use included whether the subject had ever been a habitual betel chewer, cigarette smoker, or alcoholic beverage drinker, what year the subject started and quit, the duration of consumption, the daily amount consumed, and type of alcoholic beverage consumed. Subjects who reported smoking more than 10 cigarettes per week for at least 6 months were defined as cigarette smokers, those who reported regularly chewing betel quid for at least 6 months were defined as betel chewers, and those who reported drinking beer, wine or distilled spirits more than once a week for at least 6 months were defined as alcoholic beverage drinkers.

Statistical analysis
Lifetime consumption of tobacco was calculated by multiplying the number of packs per day by the number of years smoked, giving pack-years. Lifetime consumption of betel quid was calculated as betel years by multiplying the average number of betel quid per day by the number of years chewed. Lifetime consumption of alcoholic beverage was calculated as gram years by multiplying the concentration of alcohol in the consumed beverage by the amount consumed per day by the number of years consumed. The generalized additive model was used to adjust for other confounding factors without imposing a rigid parametric assumption about their dependence on risk, as in a previous study (Hastie and Tibshirani, 1990). Unconditional logistic regression was used to assess the association between case/control status and chewing betel nut and use of other substances. For substance use, we used no cigarette smoking, no betel chewing, or no alcohol consumption (each as defined above) as baselines and compared these baselines to lifetime consumption of the 3 substances and categorized the use into 2 groups based on the median. In addition, the synergistic or combined effect of these substances was also examined for significance. The data were analysed using the SAS and S-plus statistical packages (SAS, 1988;Hastie and Tibshirani, 1990).

RESULTS
There were 271 pathology-proven cases of oesophageal squamouscell carcinoma at National Taiwan University Hospital between July, 1996 and October, 2000. 46% (77/167) of the non-recruited patients were found to have distant metastasis in this study, whereas only 20% (21/104) of the recruited patients were found to have distant metastasis, a significant difference (P < 0.001).
The mean age range of the 104 oesophageal cancer patients and 277 controls were 39-84 (mean 60.6) and 38-81 (mean 62.6) years, respectively. Table 1 shows the odds ratios for the combined effect of these substances on oesophageal cancer after adjusting for other covariates. It was found that the higher the number of substances used concurrently, the higher the risk for oesophageal cancer. Compared to those who abstained from these substances, subjects who consumed 1, 2 and 3 substances concurrently had 1.8-, 12.3-and 63.9-fold risk of oesophageal cancer, respectively (95% CI = 0.8-4.2, 5.6-27.2 and 23.7-171.9) after adjusting for age and gender. The findings remained similar after further adjustment for educational levels, tea consumption and intake of green vegetables. However, we did not find any significant interaction or joint effect between any 2 of these 3 substances (data not shown).
The smoothing plots of age and lifetime consumption of the substances after adjusting for educational levels, tea consumption, and intake of green vegetables are presented in Figure 1. Because the risk of contracting oesophageal cancer increases after about age 65 years, the patients were categorized as 65 years or younger and over 65 years for subsequent analysis, the risk of developing oesophageal cancer in the latter group increasing markedly. The relationship between lifetime consumption of all 3 substances and oesophageal cancer risk also increased sharply. Although there were instances of higher dose users showing a small decline in risk, the observations were relatively few and insignificant. For example, only 8 and 7 observations in the groups of heavy smokers and heavy chewers had a relatively low risk for oesophageal cancer, but they still had a higher risk than those who did not smoke or chew (Figure 1).
We also examined the effect of lifetime consumption of the substances on oesophageal cancer risk. The median cut-off points for lifetime consumption of these substance were 30 pack-year for cigarette smokers, 495 betel-year for areca chewers, and 1220 gram-year for alcoholic beverage drinkers. The group that chewed 495 betel-year chewed an average of about 20 betel quid per day for 20 consecutive years. The group that consumed 1220 gramyear consumed an average of four 300-350 c.c. cans of beer (5% alcohol) per day for 20 consecutive years. In the final regression model, we found that education levels and substance use, including cigarette smoking, betel chewing, and alcohol consumption, were significant risk factors of oesophageal cancer (Table 2). After adjusting for other covariates, subjects who consumed 1-495 betel-year and more than 495 betel-years (about 20 betel quid per day for 20 years) were found to have a 3.6-fold (95% CI = 1.3-10.1) and a 9.2-fold (CI 1.8-46.7) risk, respectively, of developing oesophageal cancer, compared to those who did not. For cigarette smoking, subjects who smoked more than 30 pack-year had a 3.2-fold risk (95% CI = 1.5-6.8), compared to those who did not smoke. Compared to those who did not drink, subjects who  Regression smoothing plots of oesophageal cancer with age or lifetime consumption of tobacco, areca, or alcohol, using generalized additive modelling approaches to adjust for gender, education level, tea consumption, and intake of green vegetables (6 observations of alcohol (gram-year) > 10 000 and odds ratios > 100 were discarded) consumed more than 1220 gram-year of alcohol (about 4 cans of beer per day for 20 years) in a lifetime had a 9.7-fold risk for oesophageal cancer (95% CI = 4.3-22.0). After further adjusting for age, gender, tea consumption and intake of green vegetables, these findings remained the same (Table 2).

DISCUSSION
Besides cigarette smoking and alcoholic beverage consumption, betel chewing was found to be a major independent risk factor for oesophageal cancer, the more betel consumed during a lifetime, the higher the risk. Previous studies have found that chewing 'pan' (without tobacco), which consists betel leaf, betel nut and slaked lime, was a significant risk factor for oesophageal cancer in India (Jussawalla, 1981;Nandakumar et al, 1996). Our study confirmed this association in Taiwan, although the content of betel quid in India is a little different from that in Taiwan (Chen et al, 1999).
Since most of the cases in this study were to evaluate possible surgery, it is likely that a smaller proportion of our cases had metastases than the non-recruited cases.
In Taiwan, piper betle influorescence, which contains about 15 mg g -1 safrole is frequently added to betel quid (Chen et al, 1999;Liu et al, 2000). During chewing, the concentration of safrole can reach 420 µM. Animal experiments have found that DNA adducts can be formed in the livers of mice by 1′-hydroxysafrole, a metabolite of safrole (Borchert et al, 1973;Ioannides et al, 1981;Randerath et al, 1984;Reddy and Randerath, 1990). In addition, Ramchandani and his co-workers (1998) found that administration of 25 mg pan masala extract, a dry powder mixture of areca nut, catechu, lime, and some unspecific spices and flavouring agents, to diethylnitrosamine-initiated ICRC mice significantly enhanced the growth of oesophagus papilloma. These authors concluded that habitual use of pan-masala may exert a carcinogenic and cocarcinogenic influence in the stomach and oesophagus.
The prevalence of betel chewing in the Taiwanese population is over 10% (Ko et al, 1992). One study reporting an association between betel chewing and oral cancer (Ko et al, 1995) found that subjects who swallowed betel juice were at a significantly higher risk than those who did not. These workers therefore stressed the importance of examining the effect of betel chewing on cancer of the pharynx, larynx, oesophagus and stomach. Although we demonstrated that chewing betel quid can cause oesophageal cancer, we did not collect information on whether betel juice was swallowed. Another limitation is that we lack information on the type of substances consumed with betel nut. Although most Taiwanese people consume a combination of betel nut, piper betle influorescence and lime paste, some use betel nut wrapped in betel leaf. Betel leaf contains eugenol and hydroxycavicol, which are thought to be antimutagenic and anticarcinogenic (Bhide et al, 1991;Ko et al, 1995). The latter workers have reported that chewing betel nut wrapped in betel leaf seemed to be less of an oral cancer risk than the combination of betel nut, piper betle influorescence and lime paste.
We found a combined effect of consuming cigarettes, areca, and alcohol on oesophagus cancer risk, but not a significant joint or synergistic effect caused by any two of the substances: this might be due to our small sample size.
In conclusion, in addition to cigarette smoking and alcohol consumption, areca chewing is an independent potent risk factor for oesophagus cancer in Taiwan and also a combined effect of these 3 substances.