Epidemiology of Parkinson’s disease

Introduction
Parkinson’s disease (PD) is one of the most common age-related neurodegenerative disorders, second in frequency only to Alzheimer’s disease. In the United States, at least half a million people are diagnosed as having PD, and the frequency of PD is predicted to triple over the next 50 years as the average age of the population increases.[1]

PD is a slowly progressive neurodegenerative disorder with no identifiable cause. The definition of PD does not include neurological signs suggesting more extensive injury of the motor or sensory pathways extending beyond the pigmental brain stem nuclei. These signs are suggestive of other neurodegenerative disorders, often termed atypical parkinsonism. These include multiple system atrophy, progressive supranuclear palsy, striatonigral degeneration, and other less common conditions.

The term parkinsonism is also used for syndromes where the etiology is known, such as parkinsonian signs due to stroke, infection, neuroleptic drugs, and toxic agents. Pathologically, PD is characterized by selective degeneration of dopaminergic neurons in the substantia nigra pars compacta, and by the frequent presence of ubiquitin-positive eosinophilic intracytoplasmic inclusions, known as Lewy bodies, in surviving neurons.

Epidemiology has played an important role not only in health care and planning but also as a tool for the investigation of the cause of PD. Analytic epidemiology seeks to identify risk factors that could lead to clues for causative agents of the disease. By providing a profile of disease parameters, such as prevalence, incidence, and mortality, descriptive epidemiology of PD is useful in the etiological investigation of PD.

Epidemiology of Parkinson’s disease

Prevalence

Prevalence quantifies the proportion of the total number of current subjects with PD in a population at a given time. Crude prevalence of PD has been reported to vary from 15 (per 100,000 population) in China to 657 in Argentina in door-to-door surveys,[2,3] and to vary from 100 to 250 in North America and Europe. The prevalence estimates derived by this method are greater than those derived from other methods for comparable populations. Prevalence is easily affected by socioeconomic factors and factors that affect survival rate.

Incidence
Incidence is a better estimate frequency, and it quantifies the number of new subjects with PD occurring in a given time period for a population of individuals at risk. It is relatively unaffected by factors affecting disease survival. However, as the clinical manifestations of PD may be preceded by a long latent stage and have a slow clinical progression, accurate measurements of the incidence of PD are relatively difficult. During the last 4 decades, the crude annual incidence rates of PD ranged from 1.5 per 100,000 population in China in 1986 to 14.8 in Finland through 1968 to 1970.[2,4] The variation may partly reflect study design differences, such as diagnostic criteria and methods of case ascertainment.

Age distribution
Both prevalence and incidence of PD vary greatly across age groups. PD is less common before 50 years of age and increases steadily with age thereafter up to the ninth decade. The decline among the most elderly seen in some studies probably results from the very few people in this age group and may also reflect diagnostic and ascertaining difficulties.

A recent study showed that the prevalence in Yonago City, Japan, increased from 80.6 (per 100,000 population) in 1980 to 117.9 in 1992, but that the age- and sex-adjusted prevalence decreased from 103.9 per 100,000 to 99.5.[5] There was no significant difference in incidence between 1980 and 1992, although the age-adjusted incidence in those under 55 years of age in 1992 was lower than in those under 55 in 1980. This study suggests that the increased prevalence might be due mainly to the aging of the population.

Gender differences
Although gender-specific differences reveal more variability than association with increasing age, PD appears to be slightly more common in men than in women in most studies, usually ranging from a 1.2:1 ratio up to a 1.5:1 ratio.

Ethnic distribution
The prevalence and incidence of PD vary in different countries, partly reflecting variations in racial composition of the population surveyed. Generally, white people in Europe and North America have a higher prevalence, around 100 to 350 per 100,000 population. Asians in Japan and China and black Africans have lower rates, around one-fifth to one-tenth of those in whites. However, age-adjusted PD prevalence was not significantly different in whites and blacks in a door-to-door screening conducted in Mississippi, USA.[6]

Meanwhile, two studies reported that PD incidence in African-American men and women [7] and in Asian-American men [8] was similar to rates for Americans of European origin. In addition, a door-to-door survey performed on the islet of Kinmen, Taiwan, showed that the prevalence of PD was 119 per 100,000 for the total population, similar to that of a white population and much higher than that of previous studies of Asian populations.[9] These observations may reflect the effects of exposure to environmental factors rather than racial factors. Meanwhile, the confounding effects of low case ascertainment and high selective mortality should be considered for PD prevalence estimates in populations of African origin.

Time trends
With regard to time trends for the incidence of PD, a population-based study evaluating the incidence of PD in Olmsted County, Minnesota from 1935 through 1988 showed that the annual incidence of PD increased from 9.2 per 100,000 for the interval from 1935 to 1944 to 16.3 per 100,000 for the interval from 1975 to 1984.[10] On the other hand, Zhang and Roman performed a meta-analysis by adjusting reported data with a single standard population and concluded that the prevalence and incidence of PD appear to have remained unchanged over the past 40 years.[11] As there are very few longitudinal data for PD incidence and the data might lack consistency for diagnostic criteria and study methods over time, at present it is difficult to reliably evaluate the changes in time trends of PD.

Geographic distribution
Geographic variations in the frequency of PD have been reported. For instance, in door-to-door surveys, crude prevalence is 15 per 100,000 in China, 328 per 100,000 in India, 131 per 100,000 in Mississippi, USA, and 657 per 100,000 in Argentina.[2,3,6,12][13,14] A recent study examining geographic variation in reporting of PD mortality in the United States showed strong north-to-south decreasing gradients for mortality rates for whites, regardless of gender, but no clear west-to-east gradient was demonstrated.[15] Zhang and Roman found that the geographic distribution of incidence appears to be consistent with information on prevalence.[11] This may suggest that environmental factors play a Meanwhile, significant regional differences with northwest to southeast gradients in both Canada and the United States have been reported. role in causing PD.

Mortality
Although the mortality rate represents a unique population-based statistic and has been used to examine both the time trends and the geographical distribution of PD, Phillips and colleagues found only 37% of patients had PD coded as the underlying cause of death in all diagnosed during life as having PD.[16] The reason is that PD is not a primary or direct cause of death.

Overall, in the United States, average annual age-adjusted PD mortality between 1962 and 1984 was estimated as 2 deaths per 100,000 for white men and 1 death per 100,000 for nonwhite men, 1 death per 100,000 for white women, and less than 1 death per 100,000 for nonwhite women. Mortality increased for persons 75 years and older but declined for those younger than age 70.[17] Generally, mortality rates for PD increased in the older age groups but decreased for younger ages.

Survival
Although the life expectancy of PD patients has been prolonged, the life span of PD patients is still somewhat less than that of the general population. Improved survival as the result of introducing effective symptomatic therapy and decreased or delayed mortality from other disorders may partly account for the decreased mortality in younger people.

Tanner and colleagues reported that relative survival for people with PD diagnosed before age 60 is similar to that for the general population, but relative survival is less than expected for those who are older at diagnosis.[18] A study examining prognosis of PD patients in Japan showed that the most common cause of death for all patients, regardless of age, was pneumonia.[19] This suggests that in addition to providing improved antiparkinsonian therapy to patients, PD-related conditions such as pneumonia should also be treated more aggressively.

Risk factors
Although the cause or causes of PD remain obscure, a number of factors have been associated with increased or decreased risk of PD Table 1 and Table 2. Demographic factors such as age, gender, and racial origin are associated with an increased risk of PD. Family history has been implicated as a significant risk factor for PD in several large epidemiological studies, and the estimated prevalence of positive family history ranges from 5% to 40%.[20]

All familial PD is not necessarily genetic, for families share the same environment. Several studies suggested that environmental factors play an important role in the cause of PD. For instance, in some families several members with widely different ages developed PD within a short period of time.[21] The largest twin study to date showed that genetic factors appear to be important when disease begins at or before 50 years of age.[22] Although twin studies indicate that genetic factors do not play a major role in causing typical PD, studies of several large kindred with PD or a parkinsonian syndrome have confirmed the role played by different genes such as α-synuclein and parkin.

While head injury, emotional stress, and premorbid personality have been linked to PD in numerous reports, the associations between these factors and PD are controversial because of the difficulties of recall bias, accuracy of diagnosis, and long duration of time between injury and onset of PD. A number of studies reported that different lifestyle elements such as rural living, farming activity, or well-water drinking may act as risk factors for PD. So far reports on pesticide exposure show either no convincing correlation or an increased risk of developing PD. Reports on heavy metals have been conflicting. Further studies are required in these areas.

Interestingly, a recent prospective longitudinal study involving 8004 subjects over 30 years of follow-up found a significant inverse relation between the incidence of PD and higher coffee and caffeine consumption with a dose-response relationship.[23] Those who did not drink coffee had a fivefold greater risk when compared with those who drank 28 ounces or more coffee a day. The effect persisted even after the results were adjusted for cigarette smoking. This study revealed that other nutrients in coffee, including niacin, were unrelated to PD incidence, and it suggests that the mechanism is related to caffeine intake and not to other nutrients contained in coffee.

Infectious agents were considered as a possible cause for PD because encephalitis lethargica often preceded parkinsonism during the influenza pandemic of the early 1920s. A number of studies reported that some infectious agents or diseases, such as HIV, Japanese B encephalitis, coxackie B, influenza B, herpes simplex, measles, mumps, diphtheria, croup, or rheumatic fever, might be linked to postinfectious parkinsonism either acutely or as a long-term complication. Some researchers suggest that exposure to an influenza virus in the fetus or childhood might predispose one to PD in adulthood.[24]

Recently, a study of occupational risk factors for PD in British Columbia found school teachers and those in health-care service occupations have significantly increased risk for PD, and suggested this observation might reflect higher exposure to viral respiratory tract infections circulating in school and health-care facilities.[25] However, referral bias and recall bias could not be ruled out entirely in this study since the sample was based on population attending a specialty referral clinic. Further studies with more extensive and more direct measures of exposure to respiratory tract illness are being undertaken to substantiate the findings.

Conclusion
Epidemiological studies could help to develop a novel etiologic hypothesis for PD. Most reported studies cannot be compared because of different methods of case ascertainment, diagnostic criteria, classification standards, medical facilities, and population age distribution. Studies in risk factors are limited by difficulties in assessment of exposure and selection of appropriate controls.

Current studies suggest that PD is a heterogeneous disorder. About 10% to 15% of cases are familial, and within this group there are a few patients with a strong family history suggesting autosomal dominant inheritance; these usually have a younger age of onset. Smoking and coffee consumption have been claimed as protective factors, but the mechanism remains unclear.

Acknowledgments
The authors thank the Medical Research Council of Canada and the Pacific Parkinson’s Research Institute for their support.

Dr Lai is a post-doctoral research fellow at the Neurodegenerative Disorders Centre, Division of Neurology, Department of Medicine at the University of British Columbia. Dr Tsui is an associate professor in the Division of Neurology, Department of Medicine at the University of British Columbia. He is also the director of clinical trials for the Neurodegenerative Disorders Centre.