Introduction

In Taiwan, areca nut chewing has become one of the most popular lifestyle habits over the last 40 years. The habit of chewing the nut of the Areca catechu palm is commonly referred to in English as ‘betel-chewing’, a term arising from the use of leaves from the creeping vine Piper betle to wrap up chopped nuts to form quids. Areca catechu is currently used by about 2 million Taiwanese people (approximately 10% of the population), particularly the male ‘blue collar’ workers and those with low levels of education [1]. The main varieties of chews used in Taiwan are those prepared by combining unripe areca fruit with a piece of the inflorescence (flower head) of the Piper betle vine and red lime paste, or by wrapping unripe areca fruit in leaves of the Piper betle vine with white lime paste [2]. It should be noted that no tobacco is added to the chews used in Taiwan.

Chewing areca nut has been found to significantly contribute to the development of oral cancer [3, 4], primary hepatocellular cancer [5, 6] and oesophageal carcinoma [7]. In addition, the arecal nitrosamines formed from the specific arecal alkaloids [8, 9] have been shown to be carcinogenic experimentally. Many nitrosamines are diabetogenic, and it has been suggested that arecal nitrosamines may be diabetogenic in man [10]. However, no population-based studies investigating the direct relationship between areca nut chewing and Type 2 diabetes have been performed to date. The habit of areca nut chewing is now common among Taiwanese men but, like Type 2 diabetes, this has only occurred over the last 20+ years. This reduces the risk of confounding due to the possible inheritability of the diabetogenic effects of areca nut (as previously reported in experimental animals [9]) or of parental diabetes [11] in this population. This is of particular importance since the habit has been common for 2000 years or more in virtually all other countries where areca nut is used. Based on this fact, the aim of the present study was to investigate the dose-dependent effects of areca nut on the development of Type 2 diabetes in the Taiwanese population.

Subjects and methods

Study subjects

All data were derived from the Keelung Community-based Integrated Screening (KCIS) programme. This was a multiple-disease-screening programme that was carried out in Keelung, the northernmost county of Taiwan, between 1 January 1999 and 31 December 2001. Details of the study design and some preliminary results have been described elsewhere [12]. Briefly, a total of 42 387 subjects (including 15,097 men and 27,290 women) were enrolled in the KCIS programme and attended a screening for Type 2 diabetes. Subjects with hyperglycaemia or Type 2 diabetes were diagnosed according to the criteria of the American Diabetic Association (ADA) as defined in 1999 [13]. Subjects with fasting plasma glucose levels of ≥6.1 mmol/l were defined as having hyperglycaemia and those with fasting plasma glucose levels of ≥7.0 mmol/l were defined as having Type 2 diabetes. Patients with previously diagnosed Type 2 diabetes were identified by questionnaire. As only 231 women either reported the use of areca nut or were ex-chewers, the present study focused on males alone. After the exclusion of a further 281 male subjects with incomplete data for areca nut use, the final study population on which the present analyses were based consisted of 14,816 subjects. This project was approved by the local health committee, which is run by the Taiwan Community-based Integrating Group [12]. Subjects gave their consent to participate in the on-site KCIS screening after being fully informed about the survey by public health nurses with formal documentation in Chinese.

Data collection

Data on behavioural risk factors (smoking, alcohol intake and areca nut chewing) were collected at one-to-one interviews using a structured questionnaire. The main type of areca nut chews used in Keelung city are the so-called ‘Lao-Hwa’, which are prepared by combining portions of unripe areca nut with a piece of the inflorescence (flower head) of the Piper betle vine and red lime paste [2]. Subjects were divided into three categories according to areca nut use: (i) current chewers; (ii) non-chewers (never); and (iii) ex-chewers. The duration of areca nut use and the number of portions of areca nut chewed each day were also recorded.

To investigate whether demographic and socio-economic factors (occupation and education) differed between the areca nut chewers and the non-chewers, these data were also collected. Occupation was classified as in the study by Hashibe et al. [14]. Data on biological factors associated with the risk of Type 2 diabetes were also collected, including systolic blood pressure, diastolic blood pressure, BMI, total cholesterol, triglyceride, blood urea nitrogen, serum creatinine, uric acid and central obesity (as waist circumference). The univariate distribution of these factors is summarised in Table 1.

Table 1 Univariate distribution of demographic and biochemical variables categorized by areca nut chewing in men (KCIS programme, 1999–2001)
Full size table

Statistical analysis

The prevalences of hyperglycaemia and Type 2 diabetes in chewers and non-chewers were compared using the chi square test. Subjects were classified into five groups according to age (<40, 40–49, 50–59, 60–69 and ≥70 years). Multiple logistic regression modelling was then used to investigate whether areca nut chewing had an independent effect on the risk of hyperglycaemia or of Type 2 diabetes after adjustment for age, risk factors associated with Type 2 diabetes and socio-economic status. The models used were chosen to minimise residual confounding [15, 16]. Continuous confounders were adjusted either as linear terms or by quintile distribution. The choice of which adjustment to use for each continuous variable was determined by testing for departure from linearity between two models, one including dummy variables and the other assuming linearity for the variables of interest using likelihood ratio testing (LRT). When the result of the LRT was significant, the confounder in question was included after stratification into quintiles, otherwise the linear data for the confounder was used. Trend analysis, based on LRT, was also used to define the models for the assessment of dose–response relationships of areca nut use with respect to the duration and the quantity or intensity (duration × quantity) of areca nut chewed [17]. A p value of less than 0.05 was considered significant.

Results

Table 1 shows the distribution of demographic and biochemical data according to areca nut use. Table 2 shows the prevalence of areca nut use according to age. The average overall usage rate was 14.4%. Table 3 shows that the unadjusted prevalence rate for hyperglycaemia was 9.3% in chewers and 10.7% in non-chewers (p=0.046). However, when divided into five groups according to age (see Subjects and methods section), chewers aged 50–69 years had statistically higher prevalence rates for hyperglycaemia than non-chewers (age 50–59 years p=0.032, age 60–69 years p=0.002). As regards subjects aged less than 50 years, the prevalence rates for hyperglycaemia in chewers were still higher than those in non-chewers. This difference was not statistically significant for the group aged 40–49 years (p=0.18) and of borderline statistical significance in subjects younger than 40 years of age (p=0.09). In subjects aged 70 years or older, non-chewers had only a marginally higher prevalence rate than chewers. These findings suggested that age was a confounding factor and should be adjusted for. The overall age-adjusted prevalence of hyperglycaemia was 11.4% in chewers and 8.7% in non-chewers (p<0.0001). The age at which the prevalence of hyperglycaemia rose above 15% was younger in chewers than in non-chewers. Similar findings were noted for Type 2 diabetes (Table 3).

Table 2 Prevalence of areca nut chewing with age in men (KCIS programme, 1999–2001)
Full size table
Table 3 Prevalence of hyperglycaemia and Type 2 diabetes in men stratified by age (KCIS programme, 1999–2001)
Full size table

After adjusting for age, socio-economic status, and other biological factors, the adjusted odds ratios [ORs] for the association between areca nut use and the risk of hyperglycaemia and Type 2 diabetes were 1.19 (95% CI 0.97–1.45) and 1.29 (95% CI 1.04–1.60) respectively (Table 4).

Table 4 Adjusted odds ratios for significant factors of hyperglycaemia or Type 2 diabetes in areca-nut-chewing men using multiple logistic regression models (KCIS programme, 1999–2001)
Full size table

After controlling for relevant factors, significant dose–response relationships were noted with respect to the duration of chewing and quantity of nuts chewed (Table 5). All trend tests were statistically significant (p<0.05). Compared with non-chewers, the risk of Type 2 diabetes was 1.41 times higher (95% CI 1.08–1.86) in those who had chewed areca nut for more than 20 years and 2.02 times higher (95% CI 1.31–3.09) in those who had chewed more than 20 pieces of areca nut per day. Similar findings, though of smaller magnitude, were noted for hyperglycaemia. Table 5 also shows the combined effect of usage rates and duration of use expressed as intensity (duration × quantity). Trend tests for the intensity values divided into quartiles were statistically significant for hyperglycaemia (p=0.01) and Type 2 diabetes (p<0.01).

Table 5 Multiple logistic regression models on the risk factors for the development of hyperglycaemia and Type 2 diabetes in men (KCIS programme, 1999–2001)
Full size table

Discussion

The present study used a population-based design to investigate the association between the use of areca nut chews (the type prepared by combining areca nut, Piper betle vine leaf and lime paste) and hyperglycaemia and Type 2 diabetes. Three major findings of this study are in agreement with previous experimental work showing that feeding areca nut to young adult mice can induce permanent glucose intolerance in a significant proportion of animals [9].

Firstly, the associations between areca nut use and the risk of hyperglycaemia or Type 2 diabetes remained after adjustment for age, socio-economic status and potentially confounding lifestyle and biological factors. This suggests an independent effect of this particular habit. Our findings, together with more recent population survey data from Papua New Guinea, strengthen the argument for the diabetogenicity of areca nut use [18].

In the present study, the younger age of onset of Type 2 diabetes in chewers compared with non-chewers is consistent with the fact that a study conducted in Coventry (UK) revealed that South Asian subjects developed Type 2 diabetes at an earlier age than the European subjects [19]. Furthermore, these individuals required supplemental hypoglycaemic medication more frequently and at a younger age than the Europeans [19]. The younger age of onset observed in our study also paralleled the increased prevalence of the areca nut habit in younger people, reflecting the comparatively recent uptake of the habit over the last 40 years.

The finding of significant dose–response relationships between areca nut use (both in terms of quantity and duration) and hyperglycaemia and Type 2 diabetes suggests that an active arecal agent is directly toxic to beta cells. It is possible that the diabetogenicity of arecal chews might be due to the lime paste used rather than the areca nut. However, there have been no reports to date to suggest that lime paste is either diabetogenic or carcinogenic, though it appears to have a local inflammatory effect in the mouth [20]. In addition, the hyperglycaemia reported experimentally in mice followed the feeding of areca nut without other chew components [9].

In contrast with a previous study conducted in the UK [21], significant direct associations were observed between areca nut use and clinically damaging increases in glycaemia. This discrepancy may reflect the much larger population examined in the present study. Alternatively, it may be due to the fact that the Bangladeshi Asians who took part in the UK study were from communities that have chewed betel-quids (areca nut) for thousands of years [21], whereas the current levels of areca nut use in Taiwanese men have only been reached over the last 40 years. In view of the degree of inheritability of increased glycaemia (with islet damage) that has been demonstrated experimentally in mice fed with areca nuts, the authors of the UK study speculated that the failure to find direct associations in areca-using communities where the habit was long standing may reflect confounding by the similar inheritability of diabetogenic effects in man. Furthermore, it was postulated that any association of environmental exposure to diabetogenic nitroso-compounds with diabetes would be more easily detected in populations where this particular risk factor had only recently been introduced [21]. Thus, the relationships found between Type 2 diabetes and arecal use in the present study are particularly pertinent, since the recent uptake of areca nut chewing by this population has provided a unique opportunity to test the postulated diabetogetogenicity of this habit in man.

In conclusion, we have demonstrated an association between areca nut chewing and Type 2 diabetes using a population-based study in a community in which the habit is of recent origin.