Betel quid use and mortality in Bangladesh: a cohort study

Abstract Objective To evaluate the potential effects of betel quid chewing on mortality. (A quid consists of betel nut, wrapped in betel leaves; tobacco is added to the quid by some users). Methods Prospective data were available on 20 033 individuals aged 18–75 years, living in Araihazar, Bangladesh. Demographic and exposure data were collected at baseline using a standardized questionnaire. Cause of death was defined by verbal autopsy questionnaires administered to next of kin. We estimated hazard ratios (HR) and their 95% confidence intervals (CI) for associations between betel use and mortality from all causes and from specific causes, using Cox proportional hazards models. We adjusted for age, sex, body mass index, educational attainment and tobacco smoking history. Findings There were 1072 deaths during an average of 10 years of follow-up. Participants who had ever used betel were significantly more likely to die from all causes (HR: 1.26; 95% CI: 1.09–1.44) and cancer (HR: 1.55; 95% CI: 1.09–2.22); but not cardiovascular disease (HR: 1.16; 95% CI: 0.93–1.43). These findings were robust to adjustment for potential confounders. There was a dose–response relationship between mortality from all causes and both the duration and the intensity of betel use. The population attributable fraction for betel use was 14.1% for deaths from all causes and 24.2% for cancer. Conclusion Betel quid use was associated with mortality from all causes and from cancer in this cohort.


Introduction
Areca catechu nut (betel nut) is the fourth most commonly used addictive substance in the world, after caffeine, nicotine and alcohol. 1 Betel nut is consumed by chewing, either alone or in the form of a quid wrapped in betel leaves, slaked lime (calcium hydroxide) and different flavourings. Tobacco is added to the quid by some users. It is estimated that 600 million people worldwide chew betel regularly. 1,2 Betel is widely used throughout central, south and south-east Asia, as well as in some South Pacific islands. With the growing number of immigrants from those areas, betel use is increasing in Africa, Australia, Europe and north America, 1 where betel use remains an under-recognized public health issue.
The International Agency for Research on Cancer has concluded that betel without tobacco causes oral cancer, while betel with tobacco causes upper aerodigestive tract cancers, including cancer of the oral cavity, pharynx and oesophagus. 2 Betel use has also been linked to metabolic syndrome, 3,4 hypertension, 5,6 diabetes mellitus, 7,8 and obesity 9,10 -diseases that are closely related to the development of cardiovascular disease (CVD). Recent evidence also suggests that betel use may play a role in CVD. [11][12][13] Given that betel use has been related to an array of health outcomes, it is important to assess its impact on mortality.
Betel use has been a popular traditional habit in Bangladesh. According to a 2009 survey targeting all men and women aged 15 years or more in Bangladesh, betel was used by both men (23.5%) and women (25.2%). 14 A study of the health effects of arsenic has recruited over 20 000 participants since the year 2000. 15 This population-based cohort has also been used to assess other health issues. 16,17 Here, we examine the association of betel use with mortality from all causes and from specific causes in this cohort.

Study population
A population-based survey was used to enumerate the sampling frame and characterize residents of a 25 km 2 area in Araihazar, Bangladesh. Between October 2000 and May 2002, we recruited 11 746 participants who met the following eligibility criteria: married (to reduce loss to follow-up); aged 18-75 years; user of a tube well as a primary water supply and living in the study area for at least five years before recruitment. 15 During 2006-2008, the cohort was expanded to include an additional 8287 participants (the expansion cohort) in the same study area following the same methods. [18][19][20][21] The overall participation rate was 97%.
The cohort has been followed up with in-person home visits at 2-year intervals. 22 Participants who were not at home during the first visit were revisited and excluded if they were not reachable during any of the three attempted visits. A field clinic was established exclusively for the participants and their family members to passively follow-up the participants between their biennial visits. 15 Since this rural population lacks basic health-care services from the existing health-care facilities, all participants and their family members come to the clinic for all health-care needs. Informed consent was obtained from the study participants and study procedures were approved by the Ethical Committee of the Bangladesh Medical Research Council and the Institutional Review Boards of Objective To evaluate the potential effects of betel quid chewing on mortality. (A quid consists of betel nut, wrapped in betel leaves; tobacco is added to the quid by some users). Methods Prospective data were available on 20 033 individuals aged 18-75 years, living in Araihazar, Bangladesh. Demographic and exposure data were collected at baseline using a standardized questionnaire. Cause of death was defined by verbal autopsy questionnaires administered to next of kin. We estimated hazard ratios (HR) and their 95% confidence intervals (CI) for associations between betel use and mortality from all causes and from specific causes, using Cox proportional hazards models. We adjusted for age, sex, body mass index, educational attainment and tobacco smoking history. Findings There were 1072 deaths during an average of 10 years of follow-up. Participants who had ever used betel were significantly more likely to die from all causes (HR: 1.26; 95% CI: 1.09-1.44) and cancer (HR: 1.55; 95% CI: 1.09-2.22); but not cardiovascular disease (HR: 1.16; 95% CI: 0.93-1.43). These findings were robust to adjustment for potential confounders. There was a dose-response relationship between mortality from all causes and both the duration and the intensity of betel use. The population attributable fraction for betel use was 14.1% for deaths from all causes and 24.2% for cancer. Conclusion Betel quid use was associated with mortality from all causes and from cancer in this cohort.

Research
Betel use and mortality, Bangladesh Fen Wu et al.
Columbia University and the University of Chicago.

Questionnaire data
Social and demographic data were collected at baseline using a standardized questionnaire. Physicians measured height, weight and blood pressure with standard equipment. [23][24][25] We asked participants if they had been diagnosed with diabetes and compared their answers with results from glycosylated haemoglobin and glucosuria tests. 26 Questions on tobacco smoking included cigarettes and bidis (filterless, locallyproduced cigarettes), smoked alone or together, past and current use and duration of tobacco smoking. To estimate the intensity of tobacco smoking, we calculated pack-years (the product of cigarettes or bidis smoked per day and years of smoking, divided by 20). Details of betel use were collected for both the original and expansion cohort, including information on past and current use, the number of times per day betel was used and years of betel use. Information on whether betel was chewed with smokeless tobacco was collected in the expansion cohort only. We did not collect information on the amount of smokeless tobacco that was used with betel. We also estimated the intensity of betel use (quid-years) as the product of times used per day and years of use.

Assessment of mortality
The vital status of the participants was assessed at each follow-up home visit. Details of the assessment of causes of death are described elsewhere. 22,23,27 Briefly, we adapted a validated verbal autopsy procedure that was developed by the International Centre for Diarrhoea Disease Research, Bangladesh (ICDDR, B), in collaboration with the World Health Organization (WHO). The method has been used to ascertain causes of death since 1971 28 and has documented an overall 95% specificity, with an 85% sensitivity for deaths from cancer or CVD. 29 During follow-up, upon receipt of a death reported by family or neighbours, a study physician and a trained social worker administered the verbal autopsy questionnaire to the next of kin. Medical records and death certificates were collected and reviewed monthly by an outcome-assessment committee, consisting of physicians and consulting medical specialists. Causes of death were coded according to the

Statistical analyses
We computed person-years of followup from baseline to the date of death (for those who died) or to 21 April 2014. We estimated hazard ratios (HR) and 95% confidence intervals (CIs) for deaths from all causes, cancers and CVD using Cox proportional hazards models. History of betel use was classified as follows: never versus ever users; daily frequency (never, ≤ 3 times, 3-5 times and > 5 times); duration (never, ≤ 4 years, 4-12 years and > 12 years) and intensity (never, ≤ 12 quid-years, 13-60 quid-years and > 60 quid-years).
We also estimated HRs for coronary heart disease, stroke and for cancer of the digestive or respiratory organs (these are the largest two categories of cancer deaths in the cohort). We excluded 34 participants with missing data on betel use. Missing data on any of the covariates (8-281) were coded with dummy variables, allowing participants with missing data to be included in the analyses under a missingat-random assumption. We first adjusted for sex and baseline age (model 1), we then adjusted for baseline body mass index (BMI; kg/m 2 ), educational attainment (years) and smoking status (never, past and current), (model 2). The final model (model 3) was the same as model 2, except that we used intensity of tobacco smoking (pack-years) as the variable controlling for effects of tobacco.
Sensitivity analyses were conducted separately for mortality from all causes, cancers and CVD. We tested exclusion of deaths that occurred within two years of the baseline, under the assumption that individuals who are seriously ill at baseline are more likely to die in the first two years of followup. We conducted stratified analyses by sex, age, BMI, smoking status (never/ ever) and educational attainment, adjusted for the same covariates as in model 2. In these models, we included interaction terms between betel (never versus ever user) and the dichotomous strata variables.
In the expansion cohort (n = 8287), we assessed whether mortality from all causes differed depending on whether chewing tobacco was included in the betel quids used. Finally, we tested inclusion of additional variables for arsenic exposure, systolic blood pressure and diabetes status in the models. We calculated the population attributable fraction (PAF) of mortality from all causes and from cancer associated with use of betel (ever users versus never users) using the following equation 17 : where P i is the proportion of mortality within the exposure category i and HR i is the adjusted HR of the ith category relative to the unexposed category. All analyses were done using SAS, version 9.3 (SAS Institute Inc., Cary, United States of America).

Baseline characteristics
We observed 202 874 person-years during an average of 10 years of follow-up. There were 1072 deaths, of which 167 were from cancers and 439 were from CVD, together accounting for 56.5% of deaths. Among the deaths from CVD, 181 were from coronary heart disease and 183 were from stroke. Detailed causes of death and ICD-10 codes are shown in Table 1. Diabetes status as ascertained by questionnaire appeared valid, based on comparison with glycosylated haemoglobin and glucosuria tests. The prevalence of diabetes at baseline in this lean population was under 2%. 26 The prevalence of past and current use of betel was 2.3% (465/19 999) and 32.7% (6535/19 999), respectively, in the overall study population. Given that the number of past users of betel was small, past and current users were combined as ever users in the analyses. While past and current use was significantly more frequent among men, women reported more frequent and intense use than men. Distributions of baseline variables by status of betel use are shown in Table 2. Past users were more likely to be men whereas more women were current users. Both past and current users tended to be older, less educated, past or current tobacco smokers and were more likely to have high BMI, high blood pressure or diabetes.

Betel use and mortality
Betel use was positively associated with all-cause and cancer-related mortality, after adjusting for age and sex (model 1). The associations did not change substantially when potential confounders (BMI, smoking status and educational attainment) were added to the model (model 2) or after adjusting for intensity of tobacco smoking (model 3, Table 3). Associations were significant for all causes of death (HR: 1.26; 95% CI: 1.09-1.44) and for cancer (HR: 1.55; 95% CI: 1.09-2.22); but not for CVD (HR: 1.16; 95% CI: 0.93-1.43).
The results were not appreciably altered by inclusion of arsenic exposure, systolic blood pressure or diabetes status in the models, or by exclusion of deaths reported in the first 2 years of follow-up ( Table 3). The population attributable fraction for betel use was 14.1% for deaths from all causes and 24.2% for cancer-related deaths.
We observed a dose-response relationship between mortality from all causes and duration of betel use. Among participants who had used betel for less than 4 years, for 4-12 years and for more than 12 years, the HRs were 1.16 (95% CI: 0.96-1.40), 1.24 (95% CI: 1.03-1.49) and 1.37 (95% CI: 1.15-1.62) respectively (Table 3). Results for intensity of use were similar, though the central estimates of risk did not increase monotonically with intensity of use. A dose-response relationship was not evident for mortality from cancers or CVD (Table 3).

Cause of death (ICD-10 code)
No. Betel use was associated with mortality from cancer of the digestive organs after adjustment for age and sex (HR: 1.92; 95% CI: 1.09-3.36). However, this association was attenuated and no longer significant after further adjusting for BMI, smoking status and educational attainment (HR: 1.70; 95% CI: 0.96-2.99). The risk of death from respiratory cancers was also increased among ever users, but not significantly (HR: 1.75; 95% CI: 0.96-3.18). There were no significant associations between betel use and the risk of death from coronary heart disease or stroke (data available from the corresponding author).

Subgroup analyses
The association between betel use and mortality from all causes was stronger in younger individuals as well as in individuals with a higher BMI; however, these interactions did not reach statistical significance (P for interaction = 0.06 and 0.09, respectively). Associations between betel use and mortality from all causes and from cancer did not differ substantially by sex or by smoking status. For instance, the HR for mortality from all causes was 1.24 (95% CI: 0.97-1.58) among never smokers and 1.19 (95% CI: 1.01-1.40) among ever smokers. Similarly, the HR for cancer mortality was 1.47 (95% CI: 0.78-2.76) among never smokers and 1.38 (95% CI: 0.92-2.07) among ever smokers.
In the expansion cohort, we had information on whether or not betel was chewed with tobacco. Among 2541 ever users, 2042 (80.4%) chewed betel with tobacco, while 499 (19.6%) chewed betel without tobacco. The association between betel use and mortality from all causes persisted and was marginally significant among individuals who chewed betel without tobacco (HR: 1.55; 95% CI: 0.99-2.44). Surprisingly, there was no significant association between betel use and mortality among those who chewed betel with tobacco (HR: 0.93; 95% CI: 0.65-1.32). There was a similar proportion of tobacco smokers among people who used betel alone and those who also used chewing tobacco (40.4% and 41.3%, respectively).
For mortality from CVD, there was a significant interaction between betel use and age (P = 0.01) and between betel use and BMI (P = 0.02), such that the risk was higher among younger individuals (HR: 1.88; 95% CI: 1.27-2.80) and those with a higher BMI (HR: 1.44; 95% CI: 1.08-1.91) relative to older individuals or those with a lower BMI (Fig. 1).

Discussion
Betel use was significantly associated with all-cause and cancer-related mortality in this south Asian cohort. Two studies in India reported mixed results for chewing of betel quid or betel nut (without tobacco added). 31,32 One study 31 reported no significant effect of chewing betel quid, while the other reported increased all-cause mortality with ageadjusted relative risks of 1.19 among women and 1.11 among men, 32 but confidence intervals were not reported. In China, a study reported increased all-cause mortality associated with betel nut chewing and smoking. 33 Betel use was associated with all-cause mortality in two other studies, with relative risks of 1.19 and 1.40, respectively. 12,13 Mortality increased with higher frequency of betel use, 13 longer duration of use 12 or greater intensity of use (quid-years). 12 Increased risk was predominantly seen in users who had chewed betel for 25 years or for 350 quid-years or longer. 12 We report increased mortality with longer duration and intensity of use, but not with higher frequency of use. In our cohort, the median duration and intensity of betel use were seven years and 30 quid-years, respectively. It is possible that the dose-response relationship is Betel use and mortality, Bangladesh Fen Wu et al.
not apparent at lower intensity of use, or that this cohort has too limited a range of exposure to detect a dose-response relationship. In China, it was reported that betel was mostly used by men and rarely by women and almost all users were smokers, 33 in contrast to our study population in which men and women had a similar prevalence of betel use and less than half of betel users were also smokers. Several previous studies investigated the effects of betel use on the risk of death from all cancers or cancer of the oral cavity and other upper-digestive organs. There was increased mortality from cancer of the oral cavity, nasopharynx, liver and lung associated with betel nut chewing and smoking in China. 33 In another study, betel use without chewing tobacco was associated with a significant increase in deaths from all cancers and cancer of the oesophagus, liver, pancreas, larynx and lung but despite the fact that 90% of the betel users were also smokers, the authors did not control for smoking status. 34 Another cohort study in an elderly population found no association between betel use and cancer-related mortality. 12 In our study, we controlled for smoking and found a significantly higher risk of cancer-related mortality in participants who had ever used betel. The effect estimates did not differ substantially by smoking status. Only seven deaths in the cohort were reportedly due to cancer of the oral cavity and other upper digestive organs and analyses excluding these cases did not change the effect estimates. The associations with cancer were similar in men and women, although not significant in women. Taken together, our data add to the growing body of evidence of the influence of betel use on cancer and suggest that the effect may be independent of tobacco smoking.
Several prospective studies have suggested an overall positive association between betel use and CVD risk, [11][12][13] although some other studies found no association when CVD subtypes were considered. 12,33 In a prospective cohort of 6511 men older than 50 years, users of betel nut were at a higher risk of mortality from overall CVD and stroke but not coronary heart disease. 12 In a cohort of 56 116 men, betel nut chewing was independently associated with a greater risk of CVD mortality. 13 Similarly, in a

Research
Betel use and mortality, Bangladesh Fen Wu et al.
prospective registry-based cohort study of 21 906 men followed for 2.7 years, an independent dose-response effect of betel use on risk of incident CVD cases and deaths was observed. 11 Although we found no significant association between betel use and CVD mortality overall, CVD mortality was significantly increased among younger people and those with a higher BMI. In our previous analyses of betel use and blood pressure in the same cohort, we found that betel use without tobacco was associated with higher blood pressure. 5 Betel contains four arecal alkaloids (primarily arecoline, along with arecaidine, guvacine and guvacoline), all of which have been shown to produce nitrosamine derivatives that have potential carcinogenic effects. 2 Betel components can induce both local and systemic release of inflammatory biomarkers [35][36][37] thereby provoking oxidative stress 38 and chronic inflammation related to the development of systemic diseases. Betel chewing also causes periodontal disease, 39,40 a risk factor for cancer and CVD. 41,42 Our study represents a large population from south Asia that has received little epidemiologic attention. Other strengths of the present study include the population-based prospective study design with a high response rate (97%) and the extensive data on betel use and potential confounders, including smoking status and the intensity of smoking.
Several potential limitations, however, should also be noted. First, this cohort was not established to focus on betel use and thus is lacking information on changes in use over time. Also the population was relatively young, with a mean age of 36 years at baseline and the overall average duration of use was relatively short. The relatively small number of cases for subtypes of cancer or CVD may explain the insignificant effect estimates on these outcomes. Second, we did not have comprehensive data on chewing tobacco use, either alone or with betel. Research on the effects of chewing tobacco (in the absence of betel nut) on mortality has been inconclusive. 43,44 In one study, no association was observed between chewing betel either with or without tobacco and mortality. 31 In our cohort, the positive association for all-cause mortality remained among individuals who did not use smokeless tobacco together with betel, but no significant association was observed among those who chewed betel with tobacco. However, a limited number of subjects had data on whether chewing tobacco was used together with betel and larger studies are needed. Lastly, we cannot rule out the possibility of residual confounding by tobacco smoking in our findings.
Our data suggest that betel has a small-to-moderate impact on mortality from all causes and from cancer in this Bangladeshi population. Future larger studies are warranted to investigate the effects of betel use on subtypes of cancer and CVD. ■ Fig. 1. Betel use and mortality, Bangladesh, 2000