Type 2 Diabetes and the Risk of Parkinson's Disease

Six independent cross-sectional population surveys were carried out in five geographic areas of Finland in 1972, 1977, 1982, 1987, 1992, and 1997 (17). In 1972 and 1977, a randomly selected sample making up 6.6% of the population born between 1913 and 1947 was drawn. Since 1982, the sample was stratified by area, sex, and 10-year age-group according to the World Health Organization MONICA (MONItoring trends and determinants of CArdiovascular disease) protocol (18). The subjects included in the six surveys were 25–64 years of age, and the 1997 survey also included subjects aged 65–74 years. Subjects who participated in more than one survey were included only in the first survey cohort. The total sample size of the six surveys was 53,166. The participation rate varied by year from 74 to 88% (17). After excluding 123 subjects because of prevalent Parkinson's disease at baseline, 112 subjects because of prevalent type 1 diabetes at baseline or during follow-up, and 1,379 subjects because of incomplete data on any variables required, the present analyses comprise 25,168 men and 26,384 women. The participants gave informed consent (verbal 1972–1992 and signed 1997). These surveys were conducted according to the ethical rules of the National Public Health Institute, and the investigations were performed in accordance with the Declaration of Helsinki.

Assessment of the history of type 2 diabetes was based on self-reporting and on the data of two nationwide registers. The National Hospital Discharge Register data included hospital discharge diagnoses since 1968. Data on diabetes medication were ascertained from the national Social Insurance Institution's register on special reimbursement for antidiabetic drugs from 1964. Antidiabetic drugs prescribed by a physician are free of charge in Finland and are subject to approval of a physician of the Institution who reviews each case history. The physician confirms the diagnosis of diabetes, applying the World Health Organization criteria: one or more classic symptoms plus a fasting plasma glucose level ≥7.8 mmol/l (≥7.0 mmol/l from 1998) or the oral glucose tolerance test ≥11.1 mmol/l; at least one raised plasma glucose concentration on a fasting plasma glucose level ≥7.8 mmol/l (≥7.0 mmol/l from 1998) or the oral glucose tolerance test ≥11.1 mmol/l in the absence of symptoms; or treatment with a hypoglycemic drug (oral antidiabetic agents or insulin) (19,20). All patients receiving free medication were entered into a register maintained by the Social Insurance Institution. Subjects who reported having diabetes on the questionnaire, or who had a hospital discharge with a diagnosis of diabetes, or the approval for free-of-charge medication for diabetes before the baseline survey, were classified as having the history of diabetes at baseline.

Ascertainment of the Parkinson's disease status was based on the national Social Insurance Institution's register on special reimbursement for Parkinson's disease drugs during 1964 through 31 December 2002. Similarly to diabetic patients, all patients diagnosed with Parkinson's disease are entitled for free-of-charge drugs. According to the criteria set by the Institution, the diagnosis is based on medical history, clinical examination (tremor, bradykinesia, stiffness, etc.), and other relevant diagnostic methods. The diagnosis needs to be done by a consultant (usually a specialist in neurology). Intention tremor, essential tremor, or tremor caused by neuroleptic drugs do not qualify for special reimbursement according to the criteria. A detailed written statement from the physician in charge of the treatment verifying the clinical course and the diagnostic facts is needed to obtain the right to this special reimbursement. The statement is submitted to the Institution and reviewed by the Institution's specialist. Thus, the patient is included in the register only if the diagnosis is agreed on independently by two specialist physicians. It is possible that some mild cases of Parkinson's disease have not been qualified for the special reimbursement of drugs. All patients receiving free-of-charge medication are entered into a register maintained by the Social Insurance Institution. The national Social Insurance Institution's register data were linked to the risk factor data using the unique identification numbers assigned to every resident of Finland. Follow-up of each cohort member continued until the date of the diagnosis of Parkinson's disease, death, or 31 December 2002. This diagnosed method has been used in several studies in Finland (21).

A self-administered questionnaire was sent to the participants to be completed at home. The questionnaire included questions on medical history, socioeconomic factors, physical activity, smoking habits, and alcohol, coffee, and tea consumption. Education level, measured as the total number of school years, was divided into birth cohort–specific tertiles. A detailed description of the questions about leisure-time physical activity has been presented elsewhere (22–29). Self-reported leisure-time physical activity was classified into three categories: low, moderate, or high. Based on the questionnaire data, the participants were classified as never, ex-, or current smokers. Current smokers were further categorized according to the amount of cigarettes smoked daily (1–9, 10–19, or ≥20 cigarettes per day). Coffee consumption was categorized into five categories: none, 1–2 cups, 3–4 cups, 5–6 cups, and ≥7 cups per day (29,30). Tea consumption was categorized into three categories: none, 1–2 cups, and ≥3 cups per day because only a few people drank tea. Since questions on alcohol consumption were different between the first two surveys (1972 and 1977) and the later surveys, alcohol consumption was categorized into two categories: yes and no.

At the study site, nurses specially trained for survey methodology measured height, weight, and blood pressure using the standardized protocol according to the World Health Organization MONICA project (18). Blood pressure was measured with a standard sphygmomanometer from the right arm of the participant, who was seated for 5 min before the measurement. Height was measured without shoes, and weight was measured with light clothing. BMI was calculated as weight in kilograms divided by the square of the height in meters. After blood pressure measurement, a venous blood specimen was drawn. Serum total cholesterol concentration was determined in the same laboratory for all surveys; the Lieberman Burchard method was used in 1972 and 1977 and an enzymatic method (CHOD-PAP; Boehringer Mannheim, Mannheim, Germany) since 1982. Because the enzymatic method gave 2.4% lower values than the Lieberman Burchard method, 1972 and 1977 values were corrected by this percentage.

Sex-specific differences in risk factors based on the history of type 2 diabetes were tested using ANOVA or logistic regression after adjustment for age and study year. The association between history of type 2 diabetes and the risk of Parkinson's disease was analyzed using Cox proportional hazards models. All analyses were adjusted for age, study year, BMI, systolic blood pressure, total cholesterol, education, leisure-time physical activity, smoking, alcohol drinking, and coffee and tea drinking. Because the interactions between sex and the history of type 2 diabetes on the risk of Parkinson's disease were not statistically significant, data for men and women were combined in some analyses. To avoid the potential bias due to early diagnosis of Parkinson's disease or early mortality with a potential subclinical disease, additional analyses were carried out excluding the subjects who were diagnosed with Parkinson's disease or those who died from any cause during the first 5 years of follow-up. Finally, to assess the potential effect of secondary Parkinsonism or vascular Parkinsonism, additional sensitivity analyses were conducted after excluding subjects who had stroke or coronary heart disease or who used neuroleptic drugs either at baseline or during the follow-up. Statistical significance was considered to be P < 0.05. All statistical analyses were performed with SPSS for Windows 14.0 (SPSS, Chicago, IL).

At baseline, 591 men and 507 women were identified as having a history of type 2 diabetes. General characteristics of the study population are presented in Table 1. Men and women with a history of type 2 diabetes were older, their BMI and baseline systolic blood pressure were higher, and they had lower levels of serum total cholesterol and were less often alcohol drinkers, less often smokers, and more physically inactive than individuals without diabetes.

During a mean follow-up period of 18.0 years, 324 men and 309 women developed incident Parkinson's disease. The average ages at the time of diagnosis were 64.3 years in men and 65.8 years in women. Age- and study year–adjusted hazard ratios of incident Parkinson's disease in individuals with type 2 diabetes, compared with people without it, were 1.80 (95% CI 1.03–3.15) in men, 1.93 (1.05–3.53) in women, and 1.85 (1.23–2.80) in men and women combined (adjusted also for sex) (Table 2). Further adjustment for BMI, systolic blood pressure, total cholesterol, education, leisure-time physical activity, smoking, alcohol drinking, and coffee and tea consumption affected the results only slightly.

After exclusion of participants who were diagnosed with Parkinson's disease during the first 5 years of follow-up or who died during this period (n = 1,341), sex- and multivariate-adjusted hazard ratios of Parkinson's disease in people with type 2 diabetes, compared with people without it, was 1.88 (95% CI 1.19–2.99). After exclusion of participants who were diagnosed with stroke at baseline and during follow-up (n = 3,049), sex- and multivariate-adjusted type 2 diabetes–associated hazard ratio of Parkinson's disease was 2.00 (95% CI 1.29–3.11). After further exclusion of participants who were diagnosed with coronary heart disease at baseline (n = 1,150), sex- and multivariate-adjusted type 2 diabetes–associated hazard ratio of Parkinson's disease was 1.94 (95% CI 1.21–3.11). Moreover, after exclusion of participants who used neuroleptic medications at baseline or during follow-up (n = 1,817), sex- and multivariate-adjusted type 2 diabetes–associated hazard ratio of Parkinson's disease was 1.92 (95% CI 1.21–3.06).

Table 3 shows the association of type 2 diabetes with the risk of Parkinson's disease in different subgroups. Sex- and multivariate-adjusted direct association between the history of type 2 diabetes and the risk of Parkinson's disease was present in smokers. The association was strongest in diabetic patients aged 25–44 years and statistically significant also in those aged 45–54 years at baseline. There were no significant interactions between age and diabetes on risk of Parkinson's disease.

In this large prospective study, type 2 diabetes was associated with an increased risk of Parkinson's disease among Finnish men and women. This association was independent of the known modifying factors such as smoking status, coffee and alcohol consumption, and body weight. The multivariate-adjusted association was present in both subjects aged 25–44 and 45–54 years and smokers.

The present study is, to our knowledge, the first large prospective study to find that type 2 diabetes is a risk factor of Parkinson's disease. In one longitudinal study examining the association of diabetes at baseline with the progression of Parkinsonian-like signs, the investigators observed that diabetes was associated with worsening rigidity and gait but was not associated with an increased risk of bradykinesia or tremor (31). The presence of stroke did not substantially affect the association of diabetes with rigidity but reduced the association of diabetes with gait (31). Until now, only a few cross-sectional studies and two case-control studies have examined the prevalence of diabetes among patients with Parkinson's disease. It has been reported in one review by Sandyk (12) that over half of patients with Parkinson's disease have abnormal glucose tolerance. In a national survey of 24,831 U.S. elderly adults, the investigators found that the prevalence of diabetes among adults with Parkinson's disease was higher compared with people without Parkinson's disease (13). In two case-control studies, the prevalence of diabetes among patients with Parkinson's disease was lower compared with the control group (14,15). A moderately reduced risk of Parkinson's disease was associated with diabetes among men but not women in one case-control study (14). In another case-control study, diabetes was significantly associated with a reduced risk of Parkinson's disease in the univariate analysis but not in the multivariate analysis (15). Several reasons for the inconsistency of these two case-control studies with our study can be considered. First, potential recall bias, in the case of Parkinson's disease, could be substantial because patients often experience cognitive decline and tend to gradually change their lifestyles before the diagnoses. It has been shown that patients with Parkinson's disease often will develop dementia, but in their article (14,15), all patients with Parkinson's disease were newly diagnosed and the authors did not report any cases of dementia among patients with Parkinson's disease. In the present study, we also have no information about the dementia and Alzheimer's disease. Second, the risk effect of diabetes on Parkinson's disease could be diminished in a case-control study due to survival bias associated with high mortality among diabetic patients. Third, the authors did not differentiate between type 1 and type 2 diabetes in one case-control study (14). Fourth, there is a possibility for a reverse causality, i.e., that the development of Parkinson's disease may influence risk factors for diabetes as seen to happen for other chronic diseases such as Alzheimer's disease (32,33). Finally, they did not measure some important factors associated with both type 2 diabetes and Parkinson's disease, such as coffee and alcohol consumption, BMI, serum lipids, and physical activity (14,15).

Even though the mechanism of the association between history of type 2 diabetes and the risk of Parkinson's disease is poorly understood, several putative mechanisms can be proposed. Animal and in vitro studies have shown a role for insulin in the regulation of brain dopanergic activity. Insulin and dopamine may exert reciprocal regulation (16). Recently, insulin dysregulation and changes in insulin action have been of concern in the pathophysiology and clinical symptoms of Parkinson's disease (16).

Several lifestyle factors are associated with the risk of both type 2 diabetes and Parkinson's disease. Cigarette smoking increases the risk of type 2 diabetes, whereas some studies have suggested an inverse association between cigarette smoking and the risk of Parkinson's disease (34). Coffee drinking (caffeine) is known to reduce the risk of Parkinson's disease (34). Although no association between total alcohol intake and the risk of Parkinson's disease has been found in the Health Professionals Follow-Up Study and the Nurses' Health Study, the combined analysis of these two studies found that beer drinkers had a 30% lower incidence of Parkinson's disease than non–beer drinkers (35). A significantly inverse association between leisure-time physical activity and the risk of Parkinson's disease has been found among men in the Health Professionals Follow-Up Study but not among women in the Nurses' Health Study (36). In the present study, Finnish men and women with the history of type 2 diabetes were less often alcohol drinkers, less often smokers, and were less active than people without diabetes. However, there were no differences in coffee consumption between Finnish subjects with and without type 2 diabetes. In our study, we took into account the possible effect of all lifestyle factors by adjusting for baseline habits of tea consumption, coffee and alcohol consumption, smoking, and leisure-time physical activity in the data analyses.

It could also be hypothesized that diabetes might increase the risk of Parkinson's disease partly through excess body weight. A statistically significant direct association between triceps skinfold thickness and the risk of Parkinson's disease has been found in the Honolulu Heart Program including 7,990 Japanese-American men in Hawaii (37). We have demonstrated that excess weight, defined as a BMI of ≥23 kg/m², is associated with an elevated risk of Parkinson's disease in the same study population (21). In the present study, Finnish men and women with a history of type 2 diabetes had a significantly higher baseline BMI than individuals without diabetes. However, we found that the association between the history of type 2 diabetes and the risk of Parkinson's disease was independent of baseline BMI. The relative risk of Parkinson's disease was the lowest among the baseline oldest subjects, in whom the prevalence of Parkinson's disease is the highest. Does the risk of Parkinson's disease depend on the length of exposure to diabetes? One explanation could be that lifestyle or other possible factors such as insulin dysregulation may need a longer period to show their effects on the dopaminergic system.

There are several strengths and limitations of our study. The study was population based and was based on a large number of men and women from a homogeneous population. The mean follow-up (18.0 years) was long. The sample included the largest number of patients with Parkinson's disease reported in any prospective study to date; thus, statistical power in our study was high. Cerebrovascular disorders, coronary heart disease, and certain drugs are among the important causes of secondary Parkinsonism and vascular Parkinsonism (38–40). Exclusion of subjects with stroke or coronary heart disease or those who used neuroleptic drugs did not attenuate the association between type 2 diabetes and the Parkinson's disease risk. The main limitations of our study included, first, the absence of information in surveillance of diabetic patients from nondiabetic patients during the follow-up. Second, because we did not carry out either fasting glucose measurement or a glucose tolerance test at the baseline, we may have missed some cases of asymptomatic or diet-treated diabetes. Third, we did not have data on the severity and duration of diabetes, glucose control, and the type of drugs used for the treatment of diabetes. Fourth, because our data allowed for only a dichotomized measure of alcohol consumption in our sample, we may not have been able to control fully for the effect of this variable on the risk of Parkinson's disease. Fifth, patients with type 2 diabetes are more prone to other diseases and are therefore likely to be in contact with the health care system more often than people without diabetes. Increased surveillance of patients with diabetes for other medical conditions, including neurological disorders such as Parkinson's disease, could also have contributed to higher rates than among patients without diabetes, who tend to be under less intense long-term medical scrutiny. This could result in ascertainment bias. However, we think this is less likely in Parkinson's disease, because Parkinson's disease symptoms are clearly defined and are likely to bring a subject to the attention of the medical system. It could also be the case that diabetes-associated comorbidities exacerbate Parkinson's disease symptoms, so that diabetic patients with Parkinson's disease are more likely to reach the threshold at which they qualify for special reimbursement of drugs and thus be included in our sample. Sixth, we cannot completely exclude either the effects of residual confounding due to measurement error in the assessment of confounding factors or some unmeasured factors. Finally, we can exclude neither a shared environmental nor a genetic background of diabetes and the risk of Parkinson's disease.

In conclusion, our study demonstrated that type 2 diabetes was associated with an elevated risk of Parkinson's disease among men and women independently of other potential confounding factors. The biological mechanisms behind the association of type 2 diabetes with the risk of Parkinson's disease are, however, not understood at present. Further studies to replicate our findings in other populations should be acknowledged.

This study was supported by the Finnish Academy (grants 46558, 204274, 205657, and 108297), the Finnish Parkinson Foundation, and Special Research Funds of the Social Welfare and Health Board, City of Oulu.