Abstract

Data from Asian populations on dietary and lifestyle factors associated with Parkinson's disease are sparse. In 1993–2005, the authors examined these factors in relation to Parkinson's disease in the Singapore Chinese Health Study, a prospective cohort of 63,257 Chinese men and women. Baseline data were collected through in-person interviews using structured questionnaires. All 157 incident Parkinson's disease cases were identified either through follow-up interviews or via linkage with hospital discharge databases and Parkinson's disease outpatient registries and were confirmed by review of medical records. Current versus never smokers exhibited a reduced risk of Parkinson's disease (relative risk = 0.29, 95% confidence interval: 0.16, 0.52). Total caffeine intake was inversely related to Parkinson's disease risk (p for trend = 0.002); the relative risk for the highest versus lowest quartile was 0.55 (95% confidence interval: 0.35, 0.88). Black tea, a caffeine-containing beverage, showed an inverse association with Parkinson's disease risk that was not confounded by total caffeine intake or tobacco smoking (p for trend = 0.0006; adjusted relative risk for the highest vs. lowest tertile of intake = 0.29, 95% confidence interval: 0.13, 0.67). Green tea drinking was unrelated to Parkinson's disease risk. Diet had no strong influence on risk. Ingredients of black tea other than caffeine appear to be responsible for the beverage's inverse association with Parkinson's disease.

Parkinson's disease is a debilitating neurodegenerative disease characterized by bradykinesia, rigidity with cogwheeling, rest tremor, and postural instability. The prevalence of Parkinson's disease worldwide is known to range from 0.5 percent to 4 percent among the elderly aged 65 years or older (1). A Parkinson's disease study of twins concluded that although genetic factors are important for early-onset patients (those under age 50 years), environmental/lifestyle factors play a predominant etiologic role for the vast majority of late-onset Parkinson's disease patients (2). Of the many environmental risk factors being studied, the strongest and most consistent associations were those between cigarette smoking, coffee drinking, and reduced risk of Parkinson's disease noted in several US and European populations (3). Experimental data implicate nicotine in cigarettes and caffeine in coffee as the active components that protect against Parkinson's disease. Nicotine has been shown to stimulate dopamine release and alter the activity of monoamine oxidase B (4), whereas caffeine may improve motor deficits in Parkinson's disease via blockade of the adenosine A2 receptor (5). A few small case-control studies have suggested that tea, another caffeine-containing beverage commonly consumed worldwide, also might be associated with a reduction in Parkinson's disease risk (6, 7). It is unclear whether the active ingredient(s) mediating this neuroprotective effect in tea is caffeine or some other biologically active substance(s) present in tea but not in coffee.

Relatively little is known about Parkinson's disease risk factors for non-Whites. We recently reported that the disease burden of Parkinson's disease in Singapore is as high as that in Western populations (8). In the present study (1993–2005), we examined—within the Singapore Chinese Health Study, a 63,257-strong cohort of middle-aged and elderly men and women—the associations between incident Parkinson's disease and a wide range of exposure information obtained at baseline, including lifetime use of tobacco, current consumption of alcohol and caffeine-containing beverages, and diet assessed via a validated, 165-item food frequency questionnaire. The high prevalence of tobacco use among men and the wide spectrum of patterns of coffee, black tea, and green tea consumption among both genders of Singapore Chinese have made this population uniquely suited for an etiologic investigation of Parkinson's disease development.

MATERIALS AND METHODS

Study population

The Singapore Chinese Health Study is a population-based cohort established between April 1993 and December 1998 by recruiting 63,257 ethnic Chinese aged 45–74 years, belonging to the two major dialect groups of Hokkien and Cantonese, and residing in government-built housing estates (86 percent of the population during the enrollment period lived in such estates) (9). This study was approved by the institutional review boards of the National University of Singapore; the National Neuroscience Institute, Singapore; and the University of Minnesota, Minneapolis.

Baseline exposure assessment

At recruitment, an in-person interview was conducted in the home of a subject by a trained interviewer using a structured questionnaire to obtain information on demographics, smoking, current physical activity, menstrual and reproductive histories (women only), occupational exposure, and medical history. The questionnaire also included a validated, semiquantitative food frequency section that listed 165 food items or groups commonly consumed by the Chinese population, and the respondent referred to accompanying photographs to select from eight food frequency categories (ranging from “never or hardly ever” to “two or more times a day”) and three portion sizes. The food frequency questionnaire has subsequently been validated against a series of 24-hour dietary recall interviews (9) and selected biomarker studies (10, 11).

For caffeine-containing beverages (coffee, black tea, green tea, and sodas), study subjects were asked to choose from nine predefined categories of intake frequency: never or hardly ever, 1–3 cups/month, 1 cup/week, 2–3 cups/week, 4–6 cups/week, 1 cup/day, 2–3 cups/day, 4–5 cups/day, and 6 or more cups/day (1 cup = 237 ml). Since decaffeinated coffee is rarely consumed by our study population, only caffeinated coffee was assessed. In conjunction with this cohort, we developed the Singapore Food Composition Table, a food-nutrient database that lists the levels of 96 nutritive/nonnutritive components (including caffeine) per 100 g of cooked food and beverages in the Chinese diet. By combining information obtained from the food frequency questionnaire with nutrient values provided in this food-nutrient database, we were able to compute the mean daily intakes of caffeine and other nutrients for each subject (9). For alcoholic beverages, subjects were asked about their consumption of beer, wine, Western hard liquor, and Chinese hard liquor, and the daily amount of ethanol consumed was computed for each subject based on the frequency, portion size, and type of alcohol.

For cigarette smoking, subjects were asked whether they had ever smoked at least one cigarette a day for 1 year or longer. Subjects who answered “no” were classified as “never smokers”; those who answered “yes, but I quit smoking” were classified as “former smokers”; and those who answered “yes, and I currently smoke” were classified as “current smokers.” There were six predefined categories of average number of cigarettes smoked daily ranging from “6 cigarettes or less” to “43 cigarettes or more.” There were five categories for total years of active smoking ranging from “9 years or less” to “40 years or more.” Former smokers chose from seven categories of number of years since quitting smoking ranging from “less than 1 year” to “20 years or more.”

Case ascertainment and follow-up

We identified potential Parkinson's disease cases from three independent sources.

  1. Follow-up interviews were conducted with surviving cohort subjects between 1999 and 2004, on average 7 years after enrollment. The response rate was 90 percent. Subjects were asked whether they had ever been told by a physician that they had Parkinson's disease and, if yes, age at which diagnosis was ascertained.

  2. Computer linkage of the cohort database with the nationwide hospital discharge database was carried out. This database was set up in 1990 by the Singapore government to capture all inpatient discharge information (including diagnosis) nationwide (12). We identified all diagnoses accompanied by International Classification of Diseases, Ninth Revision code 332 (for Parkinson's disease) and code 3320 (for paralysis agitans, an alternative name for Parkinson's disease).

  3. Computer linkage of the cohort database with two hospital-specific Parkinson's disease registries in Singapore was carried out. These registries listed patients diagnosed and followed up primarily as outpatients in Parkinson's disease centers housed in the two largest public hospitals in Singapore.

Of the Parkinson's disease cases ascertained from these three sources, 90 cases were identified from self-report in the follow-up interviews (source 1), an additional 71 cases were from the hospital discharge database (source 2), and another additional 84 cases were from the two Parkinson's disease registries (source 3). Most of the cases (88 percent) had been evaluated by either movement disorder specialists or neurologists. All available medical records for these 245 identified cases were reviewed by a movement disorder specialist (L. T.) to verify the date of diagnosis and confirm that cases of disease adhered to the diagnostic criteria defined by the Advisory Council of the US National Institute of Neurological Disorders and Stroke (13). Of these cases, we excluded 88 for the following reasons: 47 patients did not meet our diagnostic criteria for Parkinson's disease, two patients had incomplete medical records, and the dates of first diagnosis for 39 patients occurred before their dates of cohort enrollment (i.e., these were prevalent cases of Parkinson's disease). Of the 157 verified incident cases of Parkinson's disease, 79 (50.3 percent) satisfied the National Institute of Neurological Disorders and Stroke's definition for probable Parkinson's disease, and 71 (45.2 percent) satisfied the National Institute of Neurological Disorders and Stroke's definition for possible Parkinson's disease. Of the remaining seven Parkinson's disease cases, four exhibited rigidity and/or asymmetrical onset in addition to resting tremor and/or bradykinesia, a substantial and sustained response to levodopa or dopamine agonist, or an inadequate trial of levodopa or dopamine agonist and absence of other cause of parkinsonism. For three patients, a certified internist had clearly documented Parkinson's disease, but the accessible medical records did not allow for the above evaluation.

The cohort was followed up by regular linkage to the Singapore Registry of Births and Deaths to update the vital status of cohort members.

Statistical analysis

For each study subject, person-years were counted from the date of baseline interview to the date of Parkinson's disease diagnosis, date of death, or July 31, 2005, whichever occurred first. The chi-square statistic was used to examine the difference in distributions of gender, dialect, and education between Parkinson's disease cases and the rest of the cohort. Student's t tests were conducted to examine the difference in age at recruitment and body mass index. Proportional hazards regression methods were used to examine the exposure–Parkinson's disease associations with adjustment for potential confounders. The strength of a given association was measured by the hazard ratio and its corresponding 95 percent confidence interval and two-sided p value. Statistical computing was conducted by using SAS version 9.1 software (SAS Institute, Inc., Cary, North Carolina).

All regression models were adjusted for age at recruitment (years), year of interview (1993–1995, 1996–1998), gender, dialect group (Cantonese, Hokkien), and level of education (no formal education, primary school, secondary school or higher). Additional regression models with mutual adjustment between cigarette smoking, tea, coffee, or total caffeine intake were examined. All linear trend tests were based on the ordinal values of intake categories defined by tertiles or quartiles within the whole cohort. We did not detect any significant differences in exposure effects by gender. Therefore, in this paper, all results are presented for both genders combined.

RESULTS

The present analysis was based on 594,086 person-years of follow-up. The mean age at diagnosis of the 157 incident Parkinson's disease cases was 67.3 years (standard deviation, 7.3 years), and the mean time interval between cohort enrollment and Parkinson's disease diagnosis was 5.5 years (standard deviation, 2.9 years). The incidence rates of Parkinson's disease within the cohort, adjusted to the age structure of the whole cohort, were 32.3 per 100,000 person-years for men and 22.0 per 100,000 person-years for women. The Parkinson's disease cases were significantly older at recruitment compared with the rest of the cohort (p < 0.001). There were also significantly more males than females among the cases compared with the cohort (p = 0.028). Otherwise, there was no statistically significant difference in body mass index, dialect group distribution, or level of education (table 1).

TABLE 1.

Baseline characteristics of cohort members and subjects with Parkinson's disease, The Singapore Chinese Health Study, 1993–2005*

 Parkinson's disease cases (n = 157) Cohort subjects (n = 63,061) 
Age at recruitment (years) 61.8 (6.7) 56.5 (8.0) 
Body mass index (kg/m223.0 (3.2) 23.1 (3.3) 
Gender   
    Male 52.9 44.2 
    Female 47.1 55.8 
Dialect   
    Cantonese 45.2 46.3 
    Hokkien 54.8 53.7 
Level of education   
    No formal education 27.4 27.4 
    Primary school (1–6 years) 47.8 44.3 
    Secondary school or higher 24.8 28.3 
 Parkinson's disease cases (n = 157) Cohort subjects (n = 63,061) 
Age at recruitment (years) 61.8 (6.7) 56.5 (8.0) 
Body mass index (kg/m223.0 (3.2) 23.1 (3.3) 
Gender   
    Male 52.9 44.2 
    Female 47.1 55.8 
Dialect   
    Cantonese 45.2 46.3 
    Hokkien 54.8 53.7 
Level of education   
    No formal education 27.4 27.4 
    Primary school (1–6 years) 47.8 44.3 
    Secondary school or higher 24.8 28.3 
*

All values are expressed as mean (standard deviation) or percent.

Cigarette smoking was strongly associated with reduced risk of Parkinson's disease. Risk for former smokers was intermediate between the high risk for never smokers and the low risk for current smokers. Among ever smokers, lower risk was found for those who began smoking at an earlier age and for those who smoked a higher number of cigarettes per day (table 2). Adjustment for alcohol consumption did not materially alter the tobacco–Parkinson's disease association (results not shown).

TABLE 2.

Cigarette smoking in relation to Parkinson's disease risk, The Singapore Chinese Health Study, 1993–2005

Characteristic No. of person-years No. of cases RR*† 95% CI* 
Smoking status     
    Never smoker 420,017 117 1.00  
    Former smoker 61,286 26 0.77 0.48, 1.23 
    Current smoker 112,783 14 0.29 0.16, 0.52 
Age at starting to smoke     
    Never 420,017 117 1.00  
    ≥20 years 77,310 26 0.69 0.44, 1.08 
    <20 years 96,759 14 0.31 0.17, 0.55 
    p for trend   <0.0001  
Never smoker 420,017 117 1.00  
Former smoker, quit for ≥10 years 28,429 12 0.81 0.44, 1.51 
Former smoker, quit for <10 years 32,857 14 0.73 0.41, 1.32 
Current smoker, 1–12 cigarettes/day 46,482 0.43 0.22, 0.86 
Current smoker, ≥13 cigarettes/day 66,301 0.18 0.07, 0.45 
p for trend   <0.0001  
Characteristic No. of person-years No. of cases RR*† 95% CI* 
Smoking status     
    Never smoker 420,017 117 1.00  
    Former smoker 61,286 26 0.77 0.48, 1.23 
    Current smoker 112,783 14 0.29 0.16, 0.52 
Age at starting to smoke     
    Never 420,017 117 1.00  
    ≥20 years 77,310 26 0.69 0.44, 1.08 
    <20 years 96,759 14 0.31 0.17, 0.55 
    p for trend   <0.0001  
Never smoker 420,017 117 1.00  
Former smoker, quit for ≥10 years 28,429 12 0.81 0.44, 1.51 
Former smoker, quit for <10 years 32,857 14 0.73 0.41, 1.32 
Current smoker, 1–12 cigarettes/day 46,482 0.43 0.22, 0.86 
Current smoker, ≥13 cigarettes/day 66,301 0.18 0.07, 0.45 
p for trend   <0.0001  
*

RR, relative risk; CI, confidence interval.

Adjusted for age at recruitment (years), year of interview (1993–1995, 1996–1998), gender, dialect (Hokkien, Cantonese), and level of education (no formal education, primary school, secondary school or higher).

Alcohol consumption was infrequent among Singapore Chinese. In our cohort, only 20.6 percent of men and 4.4 percent of women consumed alcoholic beverages on a weekly basis. Among the at-least-weekly consumers of alcoholic beverages, mean daily intake of total ethanol was 16.9 g (or 1.3 drinks) for men and 7.4 g (or 0.6 drinks) for women. Compared with non- or less-than-weekly drinkers, the at-least-weekly drinkers had a relative risk of 0.51 (95 percent confidence interval: 0.27, 0.98). This association was no longer significant following adjustment for cigarette smoking (relative risk = 0.58, 95 percent confidence interval: 0.30, 1.11). Further adjustment for caffeine and black tea consumption yielded an adjusted relative risk of 0.60 (95 percent confidence interval: 0.31, 1.16).

About half of cohort subjects drank tea at least once a week, and 70 percent drank coffee on a daily basis. Furthermore, among regular tea drinkers, roughly one third drank only green tea, another one third drank only black tea, and the remaining one third drank both types of tea. Almost half of the daily drinkers of black tea and approximately one third of the daily drinkers of green tea did not consume coffee on a regular basis. In this cohort, coffee was the main source of caffeine exposure, accounting for 84 percent of total exposure. Total caffeine intake exhibited a significant, dose-dependent inverse association with Parkinson's disease risk (table 3). Although adjustment for cigarette smoking attenuated the association between caffeine intake and Parkinson's disease risk, it remained significant in a dose-dependent manner. The inverse association between coffee intake and Parkinson's disease risk disappeared following adjustment for total caffeine exposure, suggesting that the caffeine content of coffee was primarily responsible for the effect of coffee on Parkinson's disease. In contrast, the strong inverse association with black tea, the secondary source of caffeine in this population, against Parkinson's disease was essentially unaltered after adjustment for total caffeine exposure and cigarette smoking (table 3). We found no association between green tea and Parkinson's disease risk.

TABLE 3.

Coffee, tea, and caffeine intake in relation to Parkinson's disease risk, The Singapore Chinese Health Study, 1993–2005

Characteristic No. of person-years No. of cases RR*† 95% CI* RR‡ 95% CI RR§ 95% CI 
Coffee         
    None/less than daily 173,379 55 1.00  1.00  1.00  
    1 cup/day 214,325 60 0.91 0.63, 1.31 0.96 0.67, 1.39 1.05 0.70, 1.57 
    ≥2 cups/day 206,382 42 0.68 0.45, 1.01 0.78 0.52, 1.17 1.04 0.53, 2.03 
    p for trend   0.059  0.24  0.88  
Black tea         
    Nondrinker 381,656 120 1.00  1.00  1.00  
    1st tertile (<5 cups/month) 77,603 20 0.89 0.55, 1.42 0.87 0.54, 1.40 0.87 0.54, 1.40 
    2nd tertile (5–<23 cups/month) 67,645 11 0.55 0.30, 1.03 0.53 0.29, 1.00 0.54 0.29, 1.00 
    3rd tertile (≥23 cups/month) 67,181 0.29 0.13, 0.65 0.28 0.12, 0.64 0.29 0.13, 0.67 
    p for trend   0.0006  0.0004  0.0006  
Green tea         
    Nondrinker 353,385 93 1.00  1.00  1.00  
    1st tertile (<5 cups/month) 105,575 30 1.11 0.74, 1.68 1.10 0.73, 1.67 1.10 0.73, 1.67 
    2nd tertile (5–<23 cups/month) 62,422 16 1.00 0.59, 1.70 0.98 0.58, 1.68 1.00 0.58, 1.70 
    3rd tertile (≥23 cups/month) 72,705 18 0.84 0.50, 1.40 0.84 0.50, 1.40 0.93 0.55, 1.57 
    p for trend   0.61  0.59  0.86  
Daily caffeine intake         
    1st quartile (<45.8 mg) 146,025 51 1.00  1.00    
    2nd quartile (45.8–<83.7 mg) 148,495 48 0.96 0.65, 1.42 1.00 0.68, 1.49   
    3rd quartile (83.7–<141.1 mg) 152,325 30 0.57 0.36, 0.89 0.62 0.39, 0.97   
    4th quartile (≥141.1 mg) 147,241 28 0.55 0.35, 0.88 0.64 0.40, 1.03   
    p for trend   0.002  0.016    
Characteristic No. of person-years No. of cases RR*† 95% CI* RR‡ 95% CI RR§ 95% CI 
Coffee         
    None/less than daily 173,379 55 1.00  1.00  1.00  
    1 cup/day 214,325 60 0.91 0.63, 1.31 0.96 0.67, 1.39 1.05 0.70, 1.57 
    ≥2 cups/day 206,382 42 0.68 0.45, 1.01 0.78 0.52, 1.17 1.04 0.53, 2.03 
    p for trend   0.059  0.24  0.88  
Black tea         
    Nondrinker 381,656 120 1.00  1.00  1.00  
    1st tertile (<5 cups/month) 77,603 20 0.89 0.55, 1.42 0.87 0.54, 1.40 0.87 0.54, 1.40 
    2nd tertile (5–<23 cups/month) 67,645 11 0.55 0.30, 1.03 0.53 0.29, 1.00 0.54 0.29, 1.00 
    3rd tertile (≥23 cups/month) 67,181 0.29 0.13, 0.65 0.28 0.12, 0.64 0.29 0.13, 0.67 
    p for trend   0.0006  0.0004  0.0006  
Green tea         
    Nondrinker 353,385 93 1.00  1.00  1.00  
    1st tertile (<5 cups/month) 105,575 30 1.11 0.74, 1.68 1.10 0.73, 1.67 1.10 0.73, 1.67 
    2nd tertile (5–<23 cups/month) 62,422 16 1.00 0.59, 1.70 0.98 0.58, 1.68 1.00 0.58, 1.70 
    3rd tertile (≥23 cups/month) 72,705 18 0.84 0.50, 1.40 0.84 0.50, 1.40 0.93 0.55, 1.57 
    p for trend   0.61  0.59  0.86  
Daily caffeine intake         
    1st quartile (<45.8 mg) 146,025 51 1.00  1.00    
    2nd quartile (45.8–<83.7 mg) 148,495 48 0.96 0.65, 1.42 1.00 0.68, 1.49   
    3rd quartile (83.7–<141.1 mg) 152,325 30 0.57 0.36, 0.89 0.62 0.39, 0.97   
    4th quartile (≥141.1 mg) 147,241 28 0.55 0.35, 0.88 0.64 0.40, 1.03   
    p for trend   0.002  0.016    
*

RR, relative risk; CI, confidence interval.

Adjusted for age at recruitment (years), year of interview (1993–1995, 1996–1998), gender, dialect (Hokkien, Cantonese), and level of education (no formal education, primary school, secondary school or higher).

Further adjusted for baseline smoking status (never, former, or current smoker) in addition to the factors listed in the second footnote.

§

Further adjusted for total caffeine intake (mg/day) in addition to smoking status (never, former, or current smoker) and the factors listed in the second footnote.

We further examined the combined effects of black tea intake and cigarette smoking, or black tea intake and total caffeine intake, on Parkinson's disease risk within our cohort (table 4). Reduced risk of Parkinson's disease for regular consumers of black tea was evident among both never and ever smokers. For the combined effect of caffeine and black tea intake, Parkinson's disease risk was reduced for those who were either consumers of large quantities of caffeine or regular drinkers of black tea, although these risk estimates did not reach statistical significance. The lowest level of Parkinson's disease risk was noted among the group whose intake of both caffeine and black tea was high.

TABLE 4.

Joint effects of black tea intake on cigarette smoking and caffeine intake in relation to Parkinson's disease risk, The Singapore Chinese Health Study, 1993–2005

 Black tea intake 
 Nondrinker/occasional Weekly/daily 
 No. of cases RR*† 95% CI* No. of cases RR† 95% CI 
Smoking status       
    Never smoker 97 1.00  20 0.57 0.35, 0.93 
    Ever smoker 31 0.48 0.31, 0.75 0.33 0.16, 0.68 
Caffeine intake       
    Below median 80 1.00  19 0.84 0.50, 1.39 
    Above median 48 0.76 0.53, 1.09 10 0.30 0.16, 0.59 
 Black tea intake 
 Nondrinker/occasional Weekly/daily 
 No. of cases RR*† 95% CI* No. of cases RR† 95% CI 
Smoking status       
    Never smoker 97 1.00  20 0.57 0.35, 0.93 
    Ever smoker 31 0.48 0.31, 0.75 0.33 0.16, 0.68 
Caffeine intake       
    Below median 80 1.00  19 0.84 0.50, 1.39 
    Above median 48 0.76 0.53, 1.09 10 0.30 0.16, 0.59 
*

RR, relative risk; CI, confidence interval.

All RRs were adjusted for age at recruitment (years), year of interview (1993–1995, 1996–1998), gender, dialect (Hokkien, Cantonese), and level of education (no formal education, primary school, secondary school or higher); for smoking status, RRs were further adjusted for caffeine intake (mg/day); for caffeine intake, RRs were further adjusted for smoking status (never, former, or current smoker).

We examined a range of nutritional factors in relation to Parkinson's disease after adjustment for cigarette smoking, total caffeine, and black tea intake (table 5). These factors included macro- and micronutrients that had been linked to chronic diseases, including Parkinson's disease risk. We found no evidence of a strong dietary effect on Parkinson's disease risk. Only the associations with vitamin E (p for trend = 0.03), monounsaturated fat (p for trend = 0.05), and dietary isothiocyanates (relative risk for the fourth vs. first quartile = 0.58, 95 percent confidence interval: 0.36, 0.95) achieved statistical significance.

TABLE 5.

Nutritional factors in relation to Parkinson's disease risk, The Singapore Chinese Health Study, 1993–2005

 Quartile of nutrient intake 
 1 (lowest) 4 (highest) p for trend 
 RR* RR† 95% CI* RR† 95% CI RR† 95% CI 
Macronutrients (% kcal)         
    Protein 1.00 0.78 0.50, 1.20 0.85 0.55, 1.31 0.89 0.58, 1.37 0.68 
    Carbohydrates 1.00 0.97 0.59, 1.59 1.06 0.66, 1.71 1.29 0.82, 2.02 0.20 
    Total fat 1.00 0.90 0.60, 1.36 0.86 0.56, 1.31 0.69 0.42, 1.12 0.14 
    Saturated fat 1.00 0.75 0.50, 1.13 0.73 0.48, 1.13 0.74 0.47, 1.18 0.15 
    Monounsaturated fat 1.00 1.08 0.74, 1.60 0.60 0.37, 0.96 0.75 0.47, 1.19 0.05 
    Polyunsaturated fat 1.00 1.16 0.75, 1.79 1.10 0.71, 1.72 0.90 0.57, 1.44 0.66 
    n-3 PUFA* 1.00 0.96 0.63, 1.47 0.73 0.46, 1.16 0.97 0.63, 1.49 0.60 
     Marine n-3 PUFA 1.00 0.71 0.47, 1.09 0.68 0.44, 1.04 0.71 0.46, 1.09 0.10 
     Other n-3 PUFA 1.00 0.86 0.56, 1.36 0.81 0.52, 1.27 0.97 0.63, 1.50 0.83 
    n-6 PUFA 1.00 1.11 0.71, 1.71 1.21 0.78, 1.86 0.84 0.52, 1.35 0.63 
Micronutrients/others (per 1,000 kcal)         
    Total carotenoids (mcg) 1.00 0.85 0.55, 1.32 1.09 0.72, 1.67 0.87 0.55, 1.39 0.87 
    Vitamin A (IU) 1.00 0.99 0.64, 1.54 1.36 0.90, 2.05 0.60 0.35, 1.02 0.29 
    Vitamin C (mg) 1.00 1.22 0.78, 1.90 0.93 0.57, 1.51 1.30 0.82, 2.04 0.47 
    Vitamin E (mg) 1.00 1.15 0.77, 1.71 0.78 0.50, 1.23 0.64 0.39, 1.05 0.03 
    Soy isoflavones (mg) 1.00 0.93 0.60, 1.43 1.11 0.72, 1.69 0.80 0.50, 1.28 0.56 
    Isothiocyanates (μmol) 1.00 0.85 0.56, 1.29 0.96 0.63, 1.45 0.58 0.36, 0.95 0.07 
 Quartile of nutrient intake 
 1 (lowest) 4 (highest) p for trend 
 RR* RR† 95% CI* RR† 95% CI RR† 95% CI 
Macronutrients (% kcal)         
    Protein 1.00 0.78 0.50, 1.20 0.85 0.55, 1.31 0.89 0.58, 1.37 0.68 
    Carbohydrates 1.00 0.97 0.59, 1.59 1.06 0.66, 1.71 1.29 0.82, 2.02 0.20 
    Total fat 1.00 0.90 0.60, 1.36 0.86 0.56, 1.31 0.69 0.42, 1.12 0.14 
    Saturated fat 1.00 0.75 0.50, 1.13 0.73 0.48, 1.13 0.74 0.47, 1.18 0.15 
    Monounsaturated fat 1.00 1.08 0.74, 1.60 0.60 0.37, 0.96 0.75 0.47, 1.19 0.05 
    Polyunsaturated fat 1.00 1.16 0.75, 1.79 1.10 0.71, 1.72 0.90 0.57, 1.44 0.66 
    n-3 PUFA* 1.00 0.96 0.63, 1.47 0.73 0.46, 1.16 0.97 0.63, 1.49 0.60 
     Marine n-3 PUFA 1.00 0.71 0.47, 1.09 0.68 0.44, 1.04 0.71 0.46, 1.09 0.10 
     Other n-3 PUFA 1.00 0.86 0.56, 1.36 0.81 0.52, 1.27 0.97 0.63, 1.50 0.83 
    n-6 PUFA 1.00 1.11 0.71, 1.71 1.21 0.78, 1.86 0.84 0.52, 1.35 0.63 
Micronutrients/others (per 1,000 kcal)         
    Total carotenoids (mcg) 1.00 0.85 0.55, 1.32 1.09 0.72, 1.67 0.87 0.55, 1.39 0.87 
    Vitamin A (IU) 1.00 0.99 0.64, 1.54 1.36 0.90, 2.05 0.60 0.35, 1.02 0.29 
    Vitamin C (mg) 1.00 1.22 0.78, 1.90 0.93 0.57, 1.51 1.30 0.82, 2.04 0.47 
    Vitamin E (mg) 1.00 1.15 0.77, 1.71 0.78 0.50, 1.23 0.64 0.39, 1.05 0.03 
    Soy isoflavones (mg) 1.00 0.93 0.60, 1.43 1.11 0.72, 1.69 0.80 0.50, 1.28 0.56 
    Isothiocyanates (μmol) 1.00 0.85 0.56, 1.29 0.96 0.63, 1.45 0.58 0.36, 0.95 0.07 
*

RR, relative risk; CI, confidence interval; PUFA, polyunsaturated fatty acids.

Adjusted for age at recruitment (years), year of interview (1993–1995, 1996–1998), gender, dialect (Hokkien, Cantonese), level of education (no formal education, primary school, secondary school or higher), smoking status (never, former, or current smoker), caffeine intake (mg/day), and black tea intake (cups/week).

Parkinson's disease risk was unrelated to body mass index, physical activity, medical history including diabetes mellitus/hypertension, and, for women, menstrual/reproductive histories and use of replacement hormones. Finally, the above analysis was repeated on the subset of 79 cases judged to be “probable Parkinson's disease” according to the criteria of the National Institute of Neurological Disorders and Stroke. The results were essentially unchanged.

DISCUSSION

The present study showed that among Singapore Chinese, in addition to tobacco smoking and caffeine consumption, black tea drinking was also associated with reduced risk of Parkinson's disease. Green tea intake was not associated with Parkinson's disease risk. Ingredients of black tea other than caffeine appear to be responsible for this beverage's effect on risk reduction. Furthermore, in general, our study did not support the notion of a strong dietary influence on Parkinson's disease development.

To our knowledge, this is the first prospective cohort study of Parkinson's disease risk factors in a non-White population living in Asia. Singapore Chinese are well suited for a study of coffee, green tea, and black tea consumption and health outcomes since intake profiles for all three types of beverages are divergent in this population. Other strengths of the study are its population-based design and the presumed lack of recall bias in exposure data since they were obtained prior to disease diagnosis. Our observed Parkinson's disease incidence rates are comparable to rates one of us (L. T.) previously reported in a small population-based study in Singapore, suggesting that our case ascertainment was relatively complete. In this other study, the rates adjusted to the age structure of the current cohort were 35 and 30 per 100,000 person-years for men and women, respectively (14). Finally, our study subjects were recruited from two contiguous regions of south China, thus leading to a high degree of genetic homogeneity.

A potentially major limitation is the lack of systematic screening for Parkinson's disease at baseline and hence the possibility of prevalent cases being included as incident cases. However, the effect of this omission should be minimal because our subjects were less than age 75 years at recruitment, and Parkinson's disease prevalence has been reported to be relatively low at about 0.2 percent among Chinese less than 80 years of age in Singapore (8). Furthermore, in our examination of Parkinson's disease cases, prevalent cases with dates of diagnosis before recruitment dates were excluded (39 such cases). Another limitation of the study is the low level of alcohol consumption in this cohort, especially among women, which precluded any meaningful examination of the effect of alcohol on Parkinson's disease risk. We also lacked information on duration of coffee, tea, or alcohol consumption.

The novel finding in this study is that drinking black tea, but not green tea, is associated with a reduced risk of Parkinson's disease independent of cigarette smoking and total caffeine consumption. In green tea, the naturally occurring catechins are minimally oxidized and the four primary polyphenols are epigallocatechin gallate, epigallocatechin, epicatechin gallate, and epicatechin, the most abundant being epigallocatechin gallate. When black tea is manufactured, the catechins undergo oxidation and polymerization in the fermentation process, resulting in generation of other distinct polyphenols such as thearubigins and theaflavins (15). Three case-control studies (in the United States, Hong Kong, and Singapore) (6, 7, 16) and a cohort study of male health professionals in the United States (17) have reported an inverse association between tea drinking and Parkinson's disease risk. The authors attributed the protective effect of tea, at least in part, to its caffeine content. In this study, however, the risk reduction associated with black tea drinking did not appear to be confounded by caffeine intake, suggesting that ingredients of black tea other than caffeine are responsible for this beverage's protective effect against Parkinson's disease.

Although confounding by unidentified lifestyle factors remains a possibility, the black tea–Parkinson's disease association may have a causal explanation. In experimental animals, black tea extracts have been shown to have neuroprotective and neurorescue effects (18). We speculate that the protective effect of black tea may be mediated via an estrogen-related pathway. A number of case-control studies have shown that both exogenous and endogenous estrogens may reduce the risk of Parkinson's disease (19, 20), which corroborates experimental evidence that estrogens possess neuroprotective properties on dopaminergic neurons (21, 22). Recently, we reported that among women in our cohort, levels of circulating estrogens were highest in regular black tea drinkers, intermediate in nontea drinkers, and lowest in regular green tea drinkers; these differences were dose dependent and significant (23). In most studied populations, Parkinson's disease rates among women are lower than those among men of similar ages (1). In this study, the rate among women was two thirds that among comparably aged men. This gender rate differential, that is, a lower rate among women than men, is consistent with the hypothesis of estrogens possessing beneficial properties against Parkinson's disease development.

Our results show a reduced association between cigarette smoking and Parkinson's disease risk, which is well established based on data from Western populations. The strength of the association found in this study is very similar to the overall results of a meta-analysis involving 44 case-control studies (the vast majority in Western populations) and four cohort studies (three in the United States and one in Rotterdam, the Netherlands) (3). In two cohort studies, risk decreased with increasing cumulative exposure and smoking intensity (pack-years and number of cigarettes smoked per day), and the degree of attenuation of the risk reduction was associated with duration of smoking cessation (24, 25). Because the Chinese culture is distinct from that in the West, this consistency in results between studies of Western populations and our study of Chinese in Singapore further strengthens the notion that the smoking–Parkinson's disease association is etiologic. There is experimental evidence that nicotine possesses neuroprotective properties. Specifically for Parkinson's disease, nicotine may stimulate dopamine release, inhibit free-radical damage to nigral cells, and alter activity of monoamine oxidase B (4, 26).

Our finding of an inverse association between caffeine and Parkinson's disease risk is also consistent with results based on studies conducted in Western populations. A meta-analysis involving eight case-control and four cohort studies showed that daily coffee drinkers, compared with nondrinkers, have a lower risk of Parkinson's disease (relative risk = 0.69, 95 percent confidence interval: 0.59, 0.80) (3). This apparent reduction in Parkinson's disease risk was noted for caffeine from noncoffee sources but not for decaffeinated coffee, thus indicating that caffeine is a likely putative causal agent underpinning the observed coffee–Parkinson's disease association (17, 27). Our results are consistent with this hypothesis. Caffeine is an adenosine A2A antagonist that may act in the basal ganglia by removing adenosine inhibition on dopamine transmissions, thereby improving locomotion (28, 29). Adenosine A2A antagonists have been shown to improve motor deficits in animal models of Parkinson's disease (30). Moreover, the neurotoxic effect of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP) in the depletion of dopamine in animal models of Parkinson's disease was reduced in mice treated with caffeine (5).

Epidemiologic studies of dietary factors have focused mainly on antioxidants to explore the relevance of oxidative damage to neuronal biomolecules in the substantial nigra pars compacta in the pathogenesis of Parkinson's disease (31). A meta-analysis involving seven studies in Western populations showed that moderate intake of dietary vitamin E protects against Parkinson's disease (32), which was consistent with what our results suggested. Otherwise, results from prospective cohort studies in the West have also been largely null or inconclusive for the other vitamins (32–34).

In conclusion, this paper reports a novel finding that intake of black tea, but not green tea, reduces Parkinson's disease risk even after adjustment for total caffeine exposure. A biologically plausible explanation for such an observation may exist and warrants further investigation. Because black tea is widely consumed, establishing this beverage as a potential factor that may reduce Parkinson's disease risk and understanding its underlying mechanisms have important public health implications.

This study was funded by the National Institutes of Health (National Cancer Institute grants R01 CA55069, R35 CA53890, and R01 CA80205).

Drs. L. C. Tan and Koh contributed equally to this work.

The authors thank Siew-Hong Low of the National University of Singapore for supervising the field work of the Singapore Chinese Health Study, Kazuko Arakawa of the University of Southern California for developing and managing the cohort study database, and Eyok-Yian Tan of the National Neuroscience Institute for managing the Parkinson's disease database. They also thank the Ministry of Health in Singapore for help in identifying Parkinson's disease cases via database linkages.

Conflict of interest: none declared.

References

1.
de Lau
LM
Breteler
MM
Epidemiology of Parkinson's disease
Lancet Neurol
 , 
2006
, vol. 
5
 (pg. 
525
-
35
)
2.
Tanner
CM
Ottman
R
Goldman
SM
, et al.  . 
Parkinson disease in twins: an etiologic study
JAMA
 , 
1999
, vol. 
281
 (pg. 
341
-
6
)
3.
Hernan
MA
Takkouche
B
Caamano-Isorna
F
, et al.  . 
A meta-analysis of coffee drinking, cigarette smoking, and the risk of Parkinson's disease
Ann Neurol
 , 
2002
, vol. 
52
 (pg. 
276
-
84
)
4.
Quik
M
Smoking, nicotine and Parkinson's disease
Trends Neurosci
 , 
2004
, vol. 
27
 (pg. 
561
-
8
)
5.
Chen
JF
Xu
K
Petzer
JP
, et al.  . 
Neuroprotection by caffeine and A(2A) adenosine receptor inactivation in a model of Parkinson's disease
J Neurosci
 , 
2001
, vol. 
21
 pg. 
RC143
 
6.
Tan
EK
Tan
C
Fook-Chong
SM
, et al.  . 
Dose-dependent protective effect of coffee, tea, and smoking in Parkinson's disease: a study in ethnic Chinese
J Neurol Sci
 , 
2003
, vol. 
216
 (pg. 
163
-
7
)
7.
Chan
DK
Woo
J
Ho
SC
, et al.  . 
Genetic and environmental risk factors for Parkinson's disease in a Chinese population
J Neurol Neurosurg Psychiatry
 , 
1998
, vol. 
65
 (pg. 
781
-
4
)
8.
Tan
LC
Venketasubramanian
N
Hong
CY
, et al.  . 
Prevalence of Parkinson disease in Singapore: Chinese vs Malays vs Indians
Neurology
 , 
2004
, vol. 
62
 (pg. 
1999
-
2004
)
9.
Hankin
JH
Stram
DO
Arakawa
K
, et al.  . 
Singapore Chinese Health Study: development, validation, and calibration of the quantitative food frequency questionnaire
Nutr Cancer
 , 
2001
, vol. 
39
 (pg. 
187
-
95
)
10.
Seow
A
Shi
CY
Franke
AA
, et al.  . 
Isoflavonoid levels in spot urine are associated with frequency of dietary soy intake in a population-based sample of middle-aged and older Chinese in Singapore
Cancer Epidemiol Biomarkers Prev
 , 
1998
, vol. 
7
 (pg. 
135
-
40
)
11.
Seow
A
Shi
CY
Chung
FL
, et al.  . 
Urinary total isothiocyanate (ITC) in a population-based sample of middle-aged and older Chinese in Singapore: relationship with dietary total ITC and glutathione S-transferase M1/T1/P1 genotypes
Cancer Epidemiol Biomarkers Prev
 , 
1998
, vol. 
7
 (pg. 
775
-
81
)
12.
Heng
DM
Lee
J
Chew
SK
, et al.  . 
Incidence of ischaemic heart disease and stroke in Chinese, Malays and Indians in Singapore: Singapore Cardiovascular Cohort Study
Ann Acad Med Singapore
 , 
2000
, vol. 
29
 (pg. 
231
-
6
)
13.
Gelb
DJ
Oliver
E
Gilman
S
Diagnostic criteria for Parkinson disease
Arch Neurol
 , 
1999
, vol. 
56
 (pg. 
33
-
9
)
14.
Tan
LC
Venketasubramanian
N
Jamora
RD
, et al.  . 
Incidence of Parkinson's disease in Singapore
Parkinsonism Relat Disord
 , 
2007
, vol. 
13
 (pg. 
40
-
3
)
15.
Balentine
DA
Wiseman
SA
Bouwens
LC
The chemistry of tea flavonoids
Crit Rev Food Sci Nutr
 , 
1997
, vol. 
37
 (pg. 
693
-
704
)
16.
Checkoway
H
Powers
K
Smith-Weller
T
, et al.  . 
Parkinson's disease risks associated with cigarette smoking, alcohol consumption, and caffeine intake
Am J Epidemiol
 , 
2002
, vol. 
155
 (pg. 
732
-
8
)
17.
Ascherio
A
Zhang
SM
Hernan
MA
, et al.  . 
Prospective study of caffeine consumption and risk of Parkinson's disease in men and women
Ann Neurol
 , 
2001
, vol. 
50
 (pg. 
56
-
63
)
18.
Chaturvedi
RK
Shukla
S
Seth
K
, et al.  . 
Neuroprotective and neurorescue effect of black tea extract in 6-hydroxydopamine-lesioned rat model of Parkinson's disease
Neurobiol Dis
 , 
2006
, vol. 
22
 (pg. 
421
-
34
)
19.
Ragonese
P
D'Amelio
M
Salemi
G
, et al.  . 
Risk of Parkinson disease in women: effect of reproductive characteristics
Neurology
 , 
2004
, vol. 
62
 (pg. 
2010
-
14
)
20.
Currie
LJ
Harrison
MB
Trugman
JM
, et al.  . 
Postmenopausal estrogen use affects risk for Parkinson disease
Arch Neurol
 , 
2004
, vol. 
61
 (pg. 
886
-
8
)
21.
Garcia-Segura
LM
Azcoitia
I
DonCarlos
LL
Neuroprotection by estradiol
Prog Neurobiol
 , 
2001
, vol. 
63
 (pg. 
29
-
60
)
22.
Green
PS
Simpkins
JW
Neuroprotective effects of estrogens: potential mechanisms of action
Int J Dev Neurosci
 , 
2000
, vol. 
18
 (pg. 
347
-
58
)
23.
Wu
AH
Arakawa
K
Stanczyk
FZ
, et al.  . 
Tea and circulating estrogen levels in postmenopausal Chinese women in Singapore
Carcinogenesis
 , 
2005
, vol. 
26
 (pg. 
976
-
80
)
24.
Hernan
MA
Zhang
SM
Rueda-deCastro
AM
, et al.  . 
Cigarette smoking and the incidence of Parkinson's disease in two prospective studies
Ann Neurol
 , 
2001
, vol. 
50
 (pg. 
780
-
6
)
25.
Grandinetti
A
Morens
DM
Reed
D
, et al.  . 
Prospective study of cigarette smoking and the risk of developing idiopathic Parkinson's disease
Am J Epidemiol
 , 
1994
, vol. 
139
 (pg. 
1129
-
38
)
26.
Ross
GW
Petrovitch
H
Current evidence for neuroprotective effects of nicotine and caffeine against Parkinson's disease
Drugs Aging
 , 
2001
, vol. 
18
 (pg. 
797
-
806
)
27.
Ross
GW
Abbott
RD
Petrovitch
H
, et al.  . 
Association of coffee and caffeine intake with the risk of Parkinson disease
JAMA
 , 
2000
, vol. 
283
 (pg. 
2674
-
9
)
28.
Nehlig
A
Daval
JL
Debry
G
Caffeine and the central nervous system: mechanisms of action, biochemical, metabolic and psychostimulant effects
Brain Res Brain Res Rev
 , 
1992
, vol. 
17
 (pg. 
139
-
70
)
29.
Richardson
PJ
Kase
H
Jenner
PG
Adenosine A2A receptor antagonists as new agents for the treatment of Parkinson's disease
Trends Pharmacol Sci
 , 
1997
, vol. 
18
 (pg. 
338
-
44
)
30.
Kanda
T
Tashiro
T
Kuwana
Y
, et al.  . 
Adenosine A2A receptors modify motor function in MPTP-treated common marmosets
Neuroreport
 , 
1998
, vol. 
9
 (pg. 
2857
-
60
)
31.
Abou-Sleiman
PM
Muqit
MM
Wood
NW
Expanding insights of mitochondrial dysfunction in Parkinson's disease
Nat Rev Neurosci
 , 
2006
, vol. 
7
 (pg. 
207
-
19
)
32.
Etminan
M
Gill
SS
Samii
A
Intake of vitamin E, vitamin C, and carotenoids and the risk of Parkinson's disease: a meta-analysis
Lancet Neurol
 , 
2005
, vol. 
4
 (pg. 
362
-
5
)
33.
Zhang
SM
Hernan
MA
Chen
H
, et al.  . 
Intakes of vitamins E and C, carotenoids, vitamin supplements, and PD risk
Neurology
 , 
2002
, vol. 
59
 (pg. 
1161
-
9
)
34.
Chen
H
Zhang
SM
Schwarzschild
MA
, et al.  . 
Folate intake and risk of Parkinson's disease
Am J Epidemiol
 , 
2004
, vol. 
160
 (pg. 
368
-
75
)