| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH COMMUNICATION |
1 From the Institute of Preventive Medicine (AM-FY, L-SC, and TH-HC) and Division of Biostatistics (Y-HC and TH-HC) College of Public Health, National Taiwan University, Taipei, Taiwan; Taiwan Association of Medical Screening, Taipei, Taiwan (AM-FY, L-SC, Y-HC, and TH-HC); Center for Diabetes & Metabolic Medicine, Queen Mary School of Medicine & Dentistry, Royal London Hospital, London, United Kingdom (BJB)
2 Supported by the National Science Council (NSC 91-2320-B002-171; NSC 91-2320-B002-172). 3 Address reprint requests to TH-H Chen, Division of Biostatistics, Institute of Preventive Medicine, College of Public Health, National Taiwan University, Room 532, No. 17, Hsuchow Road, Taipei, Taiwan. E-mail: chenlin{at}ntu.edu.tw.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Objective: This study investigated the effect of betel chewing on the risk of developing overt CVD.
Design: We used the prospective cohort data derived from a community-population-registry–based integrated screening program to quantify the effect of betel-quid chewing on the incidence of newly diagnosed CVD by classifying the study population into either exposed or nonexposed groups according to chewing status at baseline. We then followed the group free of CVD at recruitment for 2.72 y (SD = 1.52 y) to learn of new cardiovascular events. Proportional hazards regression modeling was used to estimate the magnitude of the effect of betel-quid chewing on CVD.
Results: After control for age and education level, ever chewers had a 23% (95% CI: 11%, 37%) greater risk of developing CVD than did never chewers; ever chewers were still at greater risk of developing CVD by 24% (95% CI: 11%, 39%) after further adjustment for age, education, and other significant confounders. Significant dose-response relations were found for betel-quid chewing (P < 0.05, trend test) after adjustment for other significant variables.
Conclusion: The habit of chewing betel nut was shown to have independent dose effects to predict increases in the risk of CVD in men, with the use of a prospective community-population-registry–based cohort study.
Key Words: Cardiovascular disease • betel quid • Areca catechu • chewing • dose-response effect • community-based integrated screening • prevalence • risk factors • humans • men
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Despite these findings direct evidence is lacking to show the association of betel-quid chewing with the development of CVD in human beings. Although sporadic case reports suggest this possibility (14, 15), there are no studies of any population-based cohort with large numbers of subjects or with continued follow-up and no prospective studies on this topic. Recently initiated community-based studies, such as the Keelung Community-based Integrated Screening (KCIS) program (16), provide an opportunity to assess the effect of the betel-quid chewing habit prospectively with data collected at baseline on occurrence rates of newly diagnosed CVD during follow-up in the cohort free of CVD at baseline screening. The aim of this study was, therefore, to determine whether there was an effect of betel chewing on the risk of developing overt CVD (confirmed CVD or death from CVD) in a large prospective and community-population-registry–based cohort.
|
SUBJECTS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
74 804 KCIS screened participants, who had attended periodic face-to-face questionnaire interviews that correspond to the interscreening interval designed in the KCIS program at least once between 1999 and 2004, were eligible for inclusion in the present study. The coverage rate was 40% by the end of 2004. Because the KCIS program was designed to offer mass screening for many detectable cancers and chronic diseases, participants were enrolled from the Keelung population registry rather than by seeking volunteers by advertisement. Moreover, although the questionnaire interviews were performed by public health nurses and contained questions on exposure to betel-quid chewing and on other baseline variables, the participants were not aware of any hypothesis about the association between betel-quid chewing and CVD before the screening. Because the KCIS program has collected baseline information on the betel-chewing habit as well as including data on the factors widely recognized to be associated with increased CVD risk, our data offer a unique opportunity to elucidate the independent effect of betel chewing on the occurrence of chronic diseases. Indeed, the dose-related relations of betel-quid chewing to the risk of developing T2DM (5) and of developing the MetS (6) in the KCIS program could only be elucidated because of these features of this prospective study.
Study design
Basically, it is a prospective study because exposure to betel-quid chewing was collected before the development of newly diagnosed CVD by following a normal cohort from which study subjects with previously diagnosed CVD at baseline were already excluded. The data used in the current subsidiary analysis about the assessment of the specific hypothesis were derived from a community- and population-registry–based integrated screening program that was designed to provide early detection of 5 major screen-detectable cancers and 3 chronic diseases by inviting residents from the population registry to participate between 1999 and 2004. However, the study design was, therefore, not tailored solely for examining the hypothesis: betel-quid chewing may lead to a higher risk of developing CVD, as seen in a conventional cohort study that sets up a specific exposure and follows subjects over time to ascertain the outcome of interest. Instead, we built up a cohort through a population-registry–based screening program. Because our screening program has collected a constellation of demographic features, lifestyle factors (eg, smoking, alcohol consumption, and betel-quid chewing), biochemical markers, dietary habits, family history of major diseases, and personal disease history, the entire cohort has also provided a series of multiple outcomes, including not only 8 diseases covered within the screening program by rescreening but also other diseases (such as CVD) by following the initially CVD-free cohort over time. Both features offer an opportunity to assess the hypotheses: betel-quid chewing leads to higher risk of CVD, with or without adjustment for other confounding factors. We divided the entire normal cohort (of members free of CVD at entry to the screening program) into exposure (betel-quid chewer) and nonexposure (nonchewer) groups and ascertained incident CVD cases through the linkage of the entire cohort with the health insurance data for costs claimed by local hospitals in Keelung district. Each specific disease in these claims data was provided by physicians with a specific International Classification of Diseases (ICD) code when completing medical charts. Because the KCIS program is an organized screening program conducted by the Keelung City Bureau of Health, all participants with an abnormal finding on chronic diseases will be referred to clinics or local hospitals by public health nurses from the local health center, which can then determine the outcome on follow-up. In addition, the entire cohort was also linked with the Taiwan National Mortality Registry to ascertain CVD or non-CVD deaths.
Because the prevalence rate of betel-quid chewing is low (0.88%) in Taiwanese women compared with that in men (16.12%), we only included male subjects in our study (n = 28 344). After excluding those without records on the habit of betel-quid chewing (n = 618), 27 726 men were included. To enhance the validity of the examination of the prospective data for evidence of a causal relation between betel-quid chewing and CVD, the subjects with previously diagnosed CVD at recruitment (n = 5820) were excluded from analyses. Thus, complete datasets were available on a total of 21 906 male subjects free of CVD (see below) at baseline recruitment between January 1999 and 31 December 2004 who were eligible for inclusion in the present study subgroup. There were 3163 cases of hypertensive heart disease (ICD code: 402); ischemic heart diseases (ICD code: 410–414); cardiomyopathy (ICD code: 425); arrhythmia (ICD code: 426–427); congestive heart failure (ICD code: 428); cerebrovascular disease (ICD code: 430–438); coronary artery bypass grafting; percutaneous transluminal angioplasty; diseases of arteries, arterioles, and capillaries (ICD code: 440–448) during follow-up (
± SD duration of follow-up: 2.81 ± 1.50 y). The follow-up time for betel chewers (n = 3930) and nonchewers (n = 17 976) were similar at 2.72 ± 1.51 y and 2.72 ± 1.52 y, respectively (P = 0.9507).
Data collection
We used a structured questionnaire to interview subjects who participated in the KCIS program, administered by trained public health nurses or volunteer workers. The questionnaire contained items on 1) lifestyle with details of betel-quid chewing, smoking, alcohol consumption, and physical activity, as well as the associated frequency of consumption and duration of habits; 2) dietary habit on the frequency and amount consumed of meat, vegetables, beans, fish, milk, and coffee; 3) personal and family disease history of diabetes mellitus, hypertension, CVD, and cerebrovascular disease. Data on dietary habits during the previous 6 mo were also obtained. We displayed the food modes and standard dishes or containers of each food to assist in estimates of portion sizes for food consumed per meal. Frequency of consumption was then categorized into 5 groups: never or seldom, 1–2 times/wk, 3–4 times/wk, 5–6 times/wk, and 
7 times/wk. The intake of meat and vegetables was recorded in days per week. The distributions of demographic characteristics, lifestyle factors, dietary factors by current users, ex-users, and nonusers are listed in Table 1
.
|
Primary endpoints used in the current study included incident CVD cases and death from CVD. Two survival times were therefore defined; time from entry to the study to death from CVD or to the time to developing overt CVD. Univariate analyses were first performed by using the proportional hazards regression model to estimate the magnitude of the effect of betel-quid chewing on CVD with adjustment for age and education level, which are established factors that affect CVD and betel-quid chewing. The multivariable proportional hazards regression model was further applied to calculating the adjusted hazard ratio (aHR) for betel-quid chewing after adjustment for other significant confounding factors identified in the univariate analyses, including lifestyle factors and family history of diseases related to CVD. P values for the independent variables with >2 categories, by likelihood ratio tests, were tested for departure from linearity or quadratic property of trend with the use of analyses of differences in twice log-likelihood between the 2 Cox regression models, one with numerical values for the category of dose-response and one without the variable of interest. P values for trends are also reported by comparing the reduced model and the model that included the variables of duration and quantity of betel quid.
To investigate the dose-response effect of betel-quid chewing on the risk of developing CVD, we classified the rate of use, duration of consumption, and cumulative exposure of betel quid into 4 groups, by quartiles, among chewers. Adjusting confounders were the same as those used in the multivariable regression model mentioned above.
The proportional hazard assumption for the effect of betel quid was checked by a calculation of complementary log-log survival function compared with survival time. The statistical significance of whether this assumption was violated was assessed by the incorporation of the interaction term (time x betel-quid chewing) in the time-dependent proportional hazards Cox regression model, with adjustment for other confounding factors. All analyses were conducted with the standard statistical software, SAS release 9.1 (SAS Institute Inc, Cary, NC). Two-tailed P values of < 0.05 were considered statistically significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
). The estimated risks of developing CVD for each of the other relevant risk factors after adjustment for age and educational attainment are also shown in Table 2. After controlling for age and education level, ever chewers had a 23% increase in the risk of developing CVD than did never chewers. No significant association was found between physical activity, smoking, and alcohol consumption and the occurrence of CVD. Among dietary factors, increased use of milk was negatively associated with CVD risk, but a significant positive association was found for fish intake. Increases in risk were also found for the association between family history of diabetes, hypertension, CVD, and cerebrovascular disease and the risk of CVD in study subjects.
|
. After adjustment for age; education level; occupation; smoking; alcohol consumption; intake of fish, milk, and coffee; physical activity; and family history of diabetes, hypertension, or CVD and cerebrovascular disease, ever chewers still had a significantly increased risk of developing CVD of 24% (95% CI: 11%, 39%). This increased risk was also predicted significantly by age, education, physical activity, milk usage, and family history of hypertension and cerebrovascular disease (Table 3); was increased by fish usage; but was not affected by smoking, alcohol use, or family history of diabetes or CVD (P = 0.28, 0.53, 0.42, and 0.26, respectively).
|
. Positive dose-response relations were found for each of these measures of betel-quid consumption (P < 0.001), with the use of trend tests, after adjustment for other significant variables. Because the dose-response relation for the duration of betel-quid chewing may look like quadratic function, a quadratic term was included in the model. However, the result was not statistically significant (P = 0.19). Another possibility for leading to nonlinear dose-response relation for the duration may be due to sparse cases for the category of duration > 19.5 y.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although we cannot rule out the possibility of our findings being confounded by unknown factors not allowed for in the study or by bias because of self-selection for recruitment into the KCIS program, we believe that the deleterious effect of betel chewing on CVD risk may be valid for the following reasons. First, only subjects free of CVD at baseline were included in the study, and dose effects for the increases in CVD risk were found in relation to duration of betel chewing and CVD risk, supporting the suggestion of a causal relation. Second, the finding of a series of dose-response relations for both rate of chewing and cumulative exposure to betel quid further support the possibility of a causal link between betel chewing and CVD risk. Third, because these findings were made after controlling for most of the established risk factors associated with CVD risk with multivariable regression analysis, we believe betel-quid chewing is likely to prove to be an independent risk factor for CVD.
The underlying biological mechanisms for the link between betel chewing and CVD may be multiple. Betel-quid components may, by increasing the concentrations of the biochemical variables that are related to MetS, increase inflammatory biomarkers associated with CVD such as MMP9. This postulate is supported by 2 different types of study: first, the significant dose-related association between betel-quid usage and the features of the MetS, as shown in our earlier study on the KCIS cohort (7) and also by the elevated circulating concentrations of inflammatory biomarkers, such as MMP9 and C-reactive protein, reported in betel chewers (13). Increases in CVD risk could also be accounted for by reductions in 
-aminobutyric acid receptor activity because specific arecal alkaloids, such as arecoline, are -aminobutyric acid receptor antagonists (15) or by other less well-characterized metabolic effects.
The current results were different from those of the study of Wen et al (18) which reported that betel-quid chewing was not associated with cardiovascular mortality. The evaluated endpoint used for the association in the study of Wen et al (18) was based on mortality from CVD (ICD code: 390–459), whereas ours are evaluated on the basis of incident CVD cases as well as CVD deaths, which may account for the disparity between our positive findings and their nonsignificant results.
Our results show that fish intake is positively associated with an increased incidence of CVD (aHR: 1.03; 95% CI: 1.01, 1.05) (Table 3
), which is a discrepancy from the inverse association reported in most cohort studies (19, 20). The probable reason is that the style of fish dishes eaten in the Keelung population usually has a high salt content that may increase the risk of CVD. The second possibility is that study subjects were limited to men because one study reported a significantly reduced risk in cardiac mortality and events only in women and not in men (21).
There were several concerns in the current study. First, it is unsure whether any nonresponse bias exists. Because information on betel quid and other risk factors in association with CVD was not available for nonparticipants, we cannot compare the distributions of these correlates between the 2 groups. However, our prevalence rate of betel-quid chewing (18%) was close to the corresponding figures (19%) reported in a previous nationwide survey (18), making nonresponse bias unlikely. The data used in the current subsidiary analysis about the assessment of our specific hypothesis were derived from a population-registry- and community-based integrated screening program rather than a specific study tailored for examining our hypothesis at the beginning of the screening program. Accordingly, because neither the invited participants nor the principal investigator and the data collector were aware of the hypothesis at baseline, we believe selection into the screening program was not biased for exposure to betel quid or for occurrence of CVD. Indeed, our study is similar in this respect to the Framingham Heart Study that targeted the association between cholesterol and coronary heart disease at the beginning of the study, but then generated a number of publications reporting interests not prespecified. Thus, the present report is considered to be based on a prospective cohort study.
Second, whether the current findings can be generalized to Keelung or to Taiwan as a whole is equivocal. The present study collected data through a community- and population-registry–based integrated screening program in which attendees were invited and sampled through the population registry since 1999. Because the whole of the eligible population would not be expected to be covered until 2006, the sampling scheme was not like the conventional sampling method for survey or for testing a specific hypothesis. Instead, the residents were sampled and invited area by area according to the organization and feasibility of the screening program. Because our samples were not obtained through probability-based samples such as simple-, cluster-, or stratified-random sampling, the result about the association between betel quid and occurrence of CVD may not be applied to the underlying population in Keelung or Taiwan. However, 3 reasons may still render our result generalizable to other populations in Taiwan. First, our prevalence rate for betel chewing (18%) was close to the corresponding figure (19%) reported in a previous nationwide survey. This enhances the comparability between our study population and the nationwide population in Taiwan. Second, because the attendees, the program director, and the data collector were not aware of this particular test hypothesis, the results on the correlation between betel-quid chewing and occurrence of CVD are likely to be unbiased. Third, because the coverage rate of the study population had reached 30% of men of the underlying population, a large number of study subjects (n = 28 344) may be representative of the underlying population in Keelung city. However, external validity in the current study should be warranted in future study.
The third concern was whether dropouts affected the results of the current study. Because the incident cases were ascertained through the linkage of our cohort with the health insurance claims data, dropouts were only caused by nonclaimed subjects attributed to other causes of deaths, claimed data outside Keelung city, or moving abroad. However, the chance for the latter 2 situations is low (see below). The number of dropouts because of other causes of death was 43 for betel-quid chewers and 352 for nonchewers. We compared the rates of non-CVD death (aHR: 1.28; 95% CI: 0.91, 1.81) and all-cause death (aHR: 1.28; 95% CI: 0.93, 1.77) between the 2 groups, and there was no significance in the differences between the 2 groups in either comparison. Dropouts seem, therefore, unlikely to substantially affect the major results.
Fourth, we used the status of betel-quid chewing at baseline as the measurement of the main interest. It can be argued that the habit of betel-quid chewing could change over time. Such a secular trend has not been measured in the current study. Nonetheless, because the habit of chewing betel quid in terms of daily consumption is unlikely to have a substantial change in Taiwanese men unless a specific intervention program has been introduced for current chewers at enrollment, the lack of such information may not affect the main results.
Finally, the outcome of CVD occurrence is through the linkage of the entire cohort with the health insurance data for costs claimed by local hospitals in Keelung district. Because the KCIS program is an organized screening program conducted by the Keelung City Bureau of Health, all participants with an abnormal finding related to any of the chronic diseases or cancer will be systematically referred to local clinics or local hospitals by public health nurses. The chance that the KCIS participants would seek medical service outside of the Keelung district is low. In addition, 2 external circumstances may also lower this likelihood. The first is that health insurance has imposed copayment on the insured if he or she goes for medical care outside the assigned local medical institution. This financial barrier reduces the possibility of seeking medical care for CVD outside Keelung city. In addition, the entire cohort was also linked with Taiwan National Mortality Registry so as to ensure the ascertainment of deaths from CVD or non-CVD deaths. These circumstances reduce the bias caused by dropouts as mentioned above. Nevertheless, bias caused by loss to follow-up because of participant moving abroad or to other areas in Taiwan cannot be fully ruled out, although the effect may be minor.
In conclusion, we have found, with the use of a prospective community- and population-registry–based cohort study, that the common habit of chewing betel nut (Areca catechu) has independent dose effects to predict increases in the risk of CVD in men. These findings may reflect the significant dose-related increases in MetS previously reported in association with the betel habit or other, as yet unidentified, disorders related to betel chewing.
|
ACKNOWLEDGMENTS |
|---|
The author's responsibilities were as follows—AM-FY contributed to data retrieval, data analysis, and writing the draft; L-SC participated in data retrieval, data management, and interpretation of results; Y-HC assisted with data collection and interpretation of results; BJB contributed to the concepts investigated, participated in drafting, interpretation of results, and writing of the manuscript; TH-HC synthesized analyses and led the writing. All authors approved the final version of the text. None of the authors had a personal or financial conflict of interest.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONSUBJECTS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
