Abstract
IMPORTANCE
Changes over time in the incidence of parkinsonism and Parkinson’s disease (PD) remain uncertain.
OBJECTIVE
To investigate secular trends (period effects) and birth cohort trends in the incidence of parkinsonism and PD over 30 years in a geographically defined American population.
DESIGN, SETTING, AND PARTICIPANTS
We used the medical records-linkage system of the Rochester Epidemiology Project (REP) to identify incidence cases of PD and other types of parkinsonism in Olmsted County, MN, from 1976 to 2005. All cases were classified by a movement disorder specialist using defined criteria through the review of the complete medical records within the system.
MAIN OUTCOMES AND MEASURES
Incidence rates of parkinsonism and PD over 30 years. We tested for secular trends (period effects) using negative binomial regression models and for birth cohort effects using age-period-cohort models.
RESULTS
The overall incidence rates increased significantly over 30 years in men for both parkinsonism (RR 1.17 per decade, 95% CI 1.03-1.33) and PD (RR 1.24 per decade, 95% CI 1.08-1.43). These trends were driven primarily by the older age groups. In particular, for men ≥ 70 years, incidence rates increased for both parkinsonism (RR 1.24 per decade, 95% CI 1.07-1.44) and PD (RR 1.35 per decade, 95% CI 1.10-1.65). The secular trends were not significant for women overall or in age strata. We observed an increased risk for both men and women born in the 1920 cohort (1915-1924). However, this birth cohort effect was significant only for PD and only in men.
CONCLUSIONS AND REVELANCE
Our study suggests that the incidence of parkinsonism and PD may have increased in recent decades particularly in men 70 years old and older. These trends may be related to the dramatic changes in smoking behavior that took place in the second half of the 20th century. However, the trends could be spurious, and need to be confirmed in other populations.
In 2008, Morozova and colleagues suggested that smokers have a 74% reduction in risk of Parkinson’s disease (PD) because some constituents of tobacco may reduce the risk of PD. Under the assumption that smoking is causally related to PD, they postulated that the drastic decline in smoking frequency that took place in US men after the peak in the 1940s - 1950s would predict an increase in the incidence of PD several decades later.1,2 However, their predictions have not been tested empirically, and the long-term time trends for PD remain uncertain.3-9 Therefore, we studied time trends for parkinsonism overall and for PD over 30 years in a well-defined US population.3,10-12
Methods
Case Ascertainment
Extensive details about the case ascertainment were reported elsewhere.3,10-12 Briefly, we ascertained cases of parkinsonism through the records-linkage system of the Rochester Epidemiology Project (REP). This system provides the infrastructure for indexing and linking essentially all medical information of the county population.13-16 All medical diagnoses, surgical interventions, and other procedures are abstracted and entered into computerized indexes using the Hospital Adaptation of the International Classification of Diseases, Eighth Revision (H-ICDA)17 or the International Classification of Diseases, Ninth Revision (ICD-9).18
We ascertained potential cases of parkinsonism using a computerized screening phase and a subsequent clinical confirmation phase, as described in the original reports.3,10-12 The complete medical records of all persons who received at least one of the screening diagnostic codes from 1976 through 2005 were reviewed by a movement disorders specialist using a specifically designed abstracting form (JHB for the years 1976 to 1990; RS for the years 1991 to 2005). In addition, we also reviewed the records for all persons who received at least one of the screening diagnostic codes in the years 2006 to 2010. This extended period of capture ensured that cases that came to clinical attention up to five years after the study period were appropriately counted as incident cases, if the onset of symptoms had occurred during the study period (lag time between onset of symptoms and clinical diagnosis). The movement disorder specialist defined the year of onset of parkinsonism and the type of parkinsonism using specified diagnostic criteria and following a manual of instructions.10,19-21 To be included in our study, patients were required to reside in Olmsted County at the time of onset of parkinsonian symptoms, and we excluded persons who denied authorization to use their medical records for research.13 All study procedures and ethical aspects were approved by the Institutional Review Boards of Mayo Clinic and Olmsted Medical Center.
Diagnostic Criteria
Our diagnostic criteria included two steps: the definition of parkinsonism as a syndrome and the definition of the different types of parkinsonism within the syndrome. Parkinsonism was defined as the presence of at least two of four cardinal signs: rest tremor, bradykinesia, rigidity, and impaired postural reflexes. PD was defined as parkinsonism with all three of the following features: no other cause (e.g., repeated stroke with stepwise progression, repeated head injury, history of encephalitis, neuroleptic treatment ≤6 mo before onset, hydrocephalus, brain tumor); no documentation of unresponsiveness to levodopa at doses of at least 1 g/d in combination with carbidopa (applicable only to patients who were treated); and no prominent or early (<1 year from onset) signs of more extensive nervous system involvement (e.g., dysautonomia) not explained otherwise.10 The other types of parkinsonism were classified as reported elsewhere.11
Reliability and Validity of Diagnoses
The case-finding procedures were valid and reliable as described more extensively elsewhere.10,11 In brief, an independent records review by the two movement disorders specialists who applied the same diagnostic criteria (JHB and RS) showed 90.0% agreement on the presence of parkinsonism and 70.0% agreement on the exclusion of parkinsonism (sample classified by RS as 30 patients with parkinsonism and 10 persons free of parkinsonism from the 1991-2005 incidence study) .11 In general, the agreement on the year of onset of parkinsonism was also high (intraclass correlation coefficient = 0.85; 95%CI, 0.77-0.92).11 Finally, a comparison of clinical diagnoses of specific proteinopathies (synucleinopathies and tauopathies) with autopsy findings in 65 patients who had died showed 81.5% agreement.11
Data Analysis
All individuals who met criteria for parkinsonism and were residents of Olmsted County at the time of symptom onset between January 1, 1976 and December 31, 2005 were included as incident cases. We calculated incidence rates using incident cases as the numerator and population counts from the REP Census as the denominator.13 Consistent with the methods used in the initial incidence study,10 and used again in the more recent incidence study,11 the denominator person-years were corrected by removing prevalent cases of parkinsonism.22
We computed age-, sex-, and decade-specific incidence rates for parkinsonism (of all types) and for PD. In addition, we investigated changes in incidence rates separately for men and women in two broad age classes: <70 years and ≥70 years. Incidence rates were directly standardized by age to the total US population from the 1990 decennial census (midpoint of the 30-year period) when overall rates were compared.23
We performed statistical testing of the time trends (period effects) using negative binomial regression models.24 Negative binomial regression was used instead of Poisson regression because we had a number of zero counts and larger variance in some models.24 The unit of observation was the incidence rate in a single calendar year (directly standardized by age to the total 1990 US population). 23 We calculated a relative risk (RR) and the corresponding 95% confidence interval (95% CI) to measure the average increase in the incidence rate over 10 calendar years.
Changes in incidence rates of parkinsonism and PD across individuals born at different times (birth-cohort effects) were investigated graphically using birth-cohort curves constructed with decennium-specific incidence rates for men and women separately. Birth cohorts of 10 calendar years were considered; the central year served as the cohort label.3 Consistent with the 10-year frame of analysis, we also disaggregated the incidence rates by 10-year age classes. We also explored birth cohort effects and period effects using age-period-cohort models as described elsewhere.25,26 For these analyses, we constructed 5-year period incidence rates and 5-year birth cohort incidence rates, and used natural splines to model nonlinear effects. All statistical testing was done at the conventional alpha level of 0.05 (two-tailed). SAS version 9.3 (SAS Institute Inc., Cary, NC) and the EPI package in R (R Foundation for Statistical Computing, Vienna, Austria) were used for analyses. 25
Results
Secular Trends (Period Effects)
We included 906 incident cases of parkinsonism of any type with onset between January 1, 1976 and December 31, 2005.3,10,11 Table 1 shows the age- and sex-specific incidence rates (cases per 100,000 person-years) for parkinsonism (of all types) and PD in three decades. Figure 1 shows incidence rates age-standardized to the total 1990 US population estimated using single calendar year data points and negative binomial regression in men and women separately for parkinsonism and for PD. The incidence rates for parkinsonism of all types and for PD were higher in men than in women across all three decades (Figure 1).
Table 1.
Age- and Sex-Specific Incidence Rates (Per 100,000 Person-Years) Across Three Decades for Parkinsonism and Parkinson’s Diseasea
| Age Group |
||||||||
|---|---|---|---|---|---|---|---|---|
| Group | Decade | 0-39 y | 40-59 y | 60-69 y | 70-79 y | 80-99 y | All ages | All ages, standardizedb |
| All Types of Parkinsonism | ||||||||
| Men | 1976-1985 | 0.3 (1) | 18.9 (15) | 122.5 (30) | 237.2 (33) | 392.5 (24) | 23.5 (103) | 38.9 |
| 1986-1995 | 0.8 (3) | 14.1 (15) | 120.3 (36) | 307.7 (56) | 379.4 (31) | 27.1 (141) | 41.7 | |
| 1996-2005 | 0.8 (3) | 15.9 (25) | 154.1 (61) | 407.1 (101) | 571.0 (67) | 41.7 (257) | 55.9 | |
| All years | 0.7 (7) | 16.1 (55) | 135.1 (127) | 333.8 (190) | 468.9 (122) | 31.8 (501) | 47.2 | |
| Women | 1976-1985 | 1.4 (5) | 15.3 (13) | 91.3 (27) | 154.9 (36) | 221.1 (35) | 23.1 (116) | 26.8 |
| 1986-1995 | 0.8 (3) | 7.8 (9) | 46.8 (16) | 166.7 (45) | 169.1 (36) | 19.0 (109) | 20.3 | |
| 1996-2005 | 0.0 (0) | 15.2 (26) | 66.6 (29) | 162.1 (52) | 283.6 (73) | 27.0 (180) | 26.0 | |
| All years | 0.7 (8) | 12.9 (48) | 67.1 (72) | 161.6 (133) | 229.1 (144) | 23.2 (405) | 24.4 | |
| Men and women | 1976-1985 | 0.9 (6) | 17.0 (28) | 105.4 (57) | 185.7 (69) | 268.9 (59) | 23.3 (219) | 31.1 |
| 1986-1995 | 0.8 (6) | 10.8 (24) | 81.1 (52) | 223.5 (101) | 227.4 (67) | 22.9 (250) | 28.7 | |
| 1996-2005 | 0.4 (3) | 15.5 (51) | 108.3 (90) | 268.9 (153) | 373.6 (140) | 34.0 (437) | 38.4 | |
| All years | 0.7 (15) | 14.4 (103) | 98.9 (199) | 232.0 (323) | 299.3 (266) | 27.3 (906) | 33.3 | |
| Parkinson's Disease | ||||||||
| Men | 1976-1985 | 0.0 (0) | 11.3 (9) | 73.5 (18) | 129.4 (18) | 81.8 (5) | 11.4 (50) | 18.2 |
| 1986-1995 | 0.3 (1) | 13.2 (14) | 73.5 (22) | 186.8 (34) | 159.1 (13) | 16.2 (84) | 24.2 | |
| 1996-2005 | 0.3 (1) | 10.8 (17) | 90.9 (36) | 201.5 (50) | 315.3 (37) | 22.9 (141) | 30.4 | |
| All years | 0.2 (2) | 11.7 (40) | 80.8 (76) | 179.2 (102) | 211.4 (55) | 17.5 (275) | 25.5 | |
| Women | 1976-1985 | 0.0 (0) | 7.0 (6) | 47.4 (14) | 51.6 (12) | 82.1 (13) | 9.0 (45) | 10.7 |
| 1986-1995 | 0.0 (0) | 5.2 (6) | 20.5 (7) | 88.9 (24) | 51.7 (11) | 8.4 (48) | 9.3 | |
| 1996-2005 | 0.0 (0) | 8.2 (14) | 32.2 (14) | 87.3 (28) | 155.4 (40) | 14.4 (96) | 13.8 | |
| All years | 0.0 (0) | 7.0 (26) | 32.6 (35) | 77.7 (64) | 101.8 (64) | 10.8 (189) | 11.5 | |
| Men and women | 1976-1985 | 0.0 (0) | 9.1 (15) | 59.2 (32) | 80.7 (30) | 82.0 (18) | 10.1 (95) | 13.8 |
| 1986-1995 | 0.1 (1) | 9.0 (20) | 45.2 (29) | 128.3 (58) | 81.5 (24) | 12.1 (132) | 15.4 | |
| 1996-2005 | 0.1 (1) | 9.4 (31) | 60.2 (50) | 137.1 (78) | 205.5 (77) | 18.5 (237) | 20.7 | |
| All years | 0.1 (2) | 9.2 (66) | 55.1 (111) | 119.2 (166) | 133.9 (119) | 14.0 (464) | 17.2 | |
| Person-year denominators a | ||||||||
| Men | 1976-1985 | 314,404 | 79,569 | 24,498 | 13,910 | 6,114 | 438,495 | -- |
| 1986-1995 | 357,438 | 106,293 | 29,930 | 18,200 | 8,170 | 520,031 | -- | |
| 1996-2005 | 383,706 | 156,767 | 39,586 | 24,809 | 11,734 | 616,602 | -- | |
| All years | 1,055,548 | 342,629 | 94,014 | 56,919 | 26,018 | 1,575,128 | -- | |
| Women | 1976-1985 | 348,937 | 85,196 | 29,562 | 23,245 | 15,827 | 502,767 | -- |
| 1986-1995 | 375,551 | 115,608 | 34,210 | 26,992 | 21,290 | 573,651 | -- | |
| 1996-2005 | 394,553 | 171,371 | 43,528 | 32,080 | 25,741 | 667,273 | -- | |
| All years | 1,119,041 | 372,175 | 107,300 | 82,317 | 62,858 | 1,743,691 | -- | |
| Men and women | 1976-1985 | 663,341 | 164,765 | 54,060 | 37,155 | 21,941 | 941,262 | -- |
| 1986-1995 | 732,989 | 221,901 | 64,140 | 45,192 | 29,460 | 1,093,682 | -- | |
| 1996-2005 | 778,259 | 328,138 | 83,114 | 56,889 | 37,475 | 1,283,875 | -- | |
| All years | 2,174,589 | 714,804 | 201,314 | 139,236 | 88,876 | 3,318,819 | -- | |
The number of observed incidence cases is reported in parentheses in the table. Incidence rates were calculated by dividing the number of observed incidence cases by the corresponding person-year denominators as listed at the bottom of the table. We did not report confidence intervals because the study covered the target population completely (no sampling was involved).
The incidence rates were directly standardized by age to the total 1990 US population. The incidence rates directly standardized by age and sex to the total 1990 US population were 31.4 for 1976-1985, 28.9 for 1986-1995, 37.9 for 1996-2005, and 33.5 for the full 30 years for parkinsonism, and they were 14.0 for 1976-1985, 15.5 for 1986-1995, 20.4 for 1996-2005, and 17.2 for the full 30 years for Parkinson’s disease.
Figure 1. Secular Trends in Incidence Rates (Period Effects).
Incidence rate curves for men and women estimated using single calendar year data points (directly standardized by age to the total 1990 US population) and negative binomial regression (top panels, parkinsonism overall; bottom panels, Parkinson’s disease). The relative risk (RR) refers to the average increase in incidence rate over 10 years.
Age-adjusted incidence rates of parkinsonism were stable for women over the 30-year time frame; however, they increased in men from 38.8 cases per 100,000 person-years in the decade 1976-1985 to 56.0 in the decade 1996-2005 (Table 1). Analyses using negative binomial regression models showed a significant increase in incidence rates for men (RR 1.17 per decade, 95% CI 1.03-1.33, p=0.01, Figure 1, top left panel) but not for women. We also observed a significant sex by calendar year interaction (the slope of increase was greater in men than women; p for interaction = 0.04). The increase in incidence rates was greater for men of age 70 years or older (RR 1.24 per decade, 95% CI 1.07-1.44, p=0.005, top right panel) than in men younger than 70 years (Figure 1, top middle panel). However, the age by calendar year interaction was not significant.
Similarly, the age-adjusted incidence rate of PD increased in men from 18.2 in the decade 1976-1985 to 30.5 in the decade 1996-2005 (Table 1). Analyses using negative binomial regression models showed a significant increase in incidence rates in men (RR 1.24 per decade, 95% CI 1.08-1.43, p=0.003, Figure 1, bottom left panel) but not for women. However, the sex by calendar year interaction was not significant. The increase was greater for men of age 70 years or older (RR 1.35 per decade, 95% CI 1.10-1.65, p=0.004; Figure 1, bottom right panel) than in men younger than 70 year (Figure 1, bottom middle panel). However, the age by calendar year interaction was not significant. Women of age 70 years or older also experienced an increase in the incidence rate of PD; however, the trend was not statistically significant (RR 1.24 per decade, 95% CI 0.94-1.64, p=0.12, Figure 1, bottom right panel).
Birth Cohort Trends
Each 10-year birth cohort contributed three 10-year age-specific incidence rates that are shown with different colors and symbols in Figure 2. Our graphical analyses of birth cohort effects revealed a possible higher incidence of both parkinsonism and PD in men and women born in the 1920 cohort. Men and women born in the 1920 cohort (1915-1924) are represented by the dotted red line with square symbols. However, in age-period-cohort models, the birth cohort effect was significant only for PD and only for men (Figure 3).
Figure 2. Birth Cohort Trends in Incidence Rates.
Birth cohort curves of age- and sex-specific incidence rates for parkinsonism overall (top panels) and Parkinson’s disease (bottom panels). The central year of each 10-year birth cohort served as the cohort label. Each birth cohort contributed three age-specific incidence rates that are shown with different colors and symbols.
Figure 3. Analyses Using Age-period-cohort Models.
Birth cohort component of the age-period-cohort models for parkinsonism overall (top panels) and Parkinson’s disease (bottom panels) in men and women separately. The birth cohort effects were displayed relative to 1920 (median year of birth for all parkinsonism patients; circle with relative risk of 1.0) and fixing the period component to 1990 (mid-point of the study interval). For a given birth cohort year (x-axis), the figure shows the ratio of the incidence rate in that year compared with the incidence rate in the 1920 birth cohort (relative risk on the y-axis).
Discussion
Our study suggests an increase in the incidence of both parkinsonism of all types and of PD over the 30-year period from 1976 to 2005 particularly in men of age 70 years or older. Although no trend was evident overall for women, women in the age group ≥70 years experienced an increase in the incidence of PD with borderline statistical significance. We also observed higher incidence rates of PD for men born in the 1920 cohort. Our study provides evidence contrary to two previous US studies and one Canadian study that showed no trend, and particularly contrary to three UK studies suggesting a possible decline in the occurrence of PD over time.3-8 A study using Swiss mortality data suggested a possible birth cohort effect with higher risk for cohorts born before the 1920s.9
The time trends that we observed need to be interpreted with caution. First, the trends may be an artifact due to increased awareness of symptoms and improved access to care of patients, or to increased awareness of signs and symptoms of parkinsonism by physicians. For example, the trends may be due to a better recognition by physicians of parkinsonism and PD in the older subjects in recent decades, or to a more inclusive diagnostic adjudication in the more recent segment of the study (RS vs. JHB adjudication). It is possible that in the earlier years of our study, elderly persons with cardiovascular conditions, cancer, or other diseases were not diagnosed with parkinsonism because the symptoms of parkinsonism were not considered important in the overall clinical management, and were not considered a major cause of disability or of mortality. In the more recent years, physicians may have started to recognize parkinsonism even in the context of complex multimorbidity. Against a diagnostic artifact is the decline in the incidence of drug-induced parkinsonism observed in this same population during the same time period (data not shown). In addition, against a diagnostic artifact is the adequate agreement between the two movement disorders specialists, and the good validity of the clinical diagnosis compared with autopsy findings.10,11 Finally, the observation that the time trends were more evident in men than in women may support a genuine trend in incidence because the recognition of the symptoms of parkinsonism in the context of multimorbidity should have changed similarly over time in men and women.
A second possible cause for a spurious increase in incidence rates is a change in coding practices in the records-linkage system or the switch from ICD-8 to ICD-9 coding in 1994. Because our case ascertainment was a laborious process involving a sensitive screening phase (inclusion of a large set of diagnostic codes to increase sensitivity even at the expense of lower specificity) and a detailed review of the medical records of those subjects who screened positive, we did not rely on electronic codes to define the disease.10,11 Although, we cannot exclude a trend in coding practices, we do not think that this was sizeable, if any. Finally, another possible cause for a spurious trend is a change over time in the percent of people who denied authorization to use their medical records for research. However, the law requiring authorization was only introduced in 1997 (Minnesota state privacy law), and the authorization rate has always been greater than 95% since 1997.15 In summary, over the 30 years of study, the health care practices in Olmsted County have not changed, the in and out migration has been limited, especially in the age groups at risk for parkinsonism, and the percent of the population covered by the REP has remained stable.15
If the trend of increasing incidence rates is genuine and can be replicated in other populations, it has major implications for etiologic research and for public health. From a research perspective, the trend should prompt studies to identify possible environmental or lifestyle changes during the life span of the study subjects.1,27,28 If we assume that the association of smoking with reduced risk of PD has a biological basis, our findings may be explained in part by time trends in smoking.1,2 Smoking behavior in industrialized nations, including the United States, has changed dramatically over the second half of the 20th century, with diverging patterns in men and women.29 The prevalence of smoking in the United States peaked in the 1940s and 1950s at approximately 67% for men and in the 1960s at approximately 44% for women. However, in the past 25 years, the gap between men and women has narrowed. Between 1965 and 2009 (44 years), the frequency declined from 51.9% to 23.5% in men, and from 33.9% to 17.9% in women.30 The decline in smoking rates in men may explain in part the increasing incidence of parkinsonism and PD.1 However, other environmental or lifestyle risk or protective factors that are related to sex or gender may also be involved such as pesticide use, head trauma, and coffee consumption.1,31,32 The possible higher risk of PD for men and women born in the 1920 birth cohort may suggest exposures that took place during intrauterine life or in early life (e.g., intrauterine infection, toxic exposure, or dietary deficiency).33 Further studies are needed to confirm these etiologic hypotheses. From a public health perspective, an increase in the incidence of parkinsonism and PD would modify our projections for the total number of persons suffering from the disease in the coming decades.34,35
Our study has a number of strengths. It is the first to consider long-term secular trends in incidence (over 30 years), and to analyze the data both as secular trends (period effects) and as birth cohort trends. Our study is also unique in the extent of clinical precision. The full record of each patient was reviewed by a movement disorders specialist to confirm the diagnosis of parkinsonism and to meticulously classify different types of parkinsonism. This level of diagnostic precision is impossible when dealing with death certificate data or with administrative data (e.g., billing data). Second, our study was conducted in a relatively stable population, using a population-based records-linkage system that spans across the entire life of the individuals included.13-16 The case-finding procedures and the diagnostic criteria used in the two studies that were combined were similar, so that the observed changes are not due to changes in methodology. In addition, all of the patients with parkinsonism were adjudicated by a movement disorders specialist at the time of medical record abstraction to reduce differences in the diagnostic criteria over time or across the different specialists. Third, all of the medical facilities in Olmsted County, MN are included in the REP,13-16 and it is unlikely that a patient with parkinsonism would have been seen exclusively outside of the county while living in the county. To avoid the risk of underestimating incidence rates of parkinsonism in the eldest age groups, we corrected our census denominators by removing prevalent cases of parkinsonism. However, the effects of these corrections were small (data not shown).
Our study also has a number of limitations. First, our study population was somewhat small to provide stable incidence rates and to conduct more definitive age-period-cohort analyses. This limitation was insurmountable because our case finding is based on the unique records-linkage system serving Olmsted County, MN, and a study of time trends for PD would be highly impractical in another population without the system. Second, we could not analyze the time trends in less common subtypes of parkinsonism (e.g., multiple system atrophy) because of sample size limitations. Third, changes in clinical practice and the introduction of new sets of diagnostic criteria during the time frame of our most recent study could have changed the differential diagnosis by subtypes of parkinsonism, thus creating spurious trends in subtypes. As evidence against this possible bias, the overall distribution of subtypes of parkinsonism was similar in the previous study and in the more recent study. In addition, a trend in the classification of subtypes of parkinsonism should not influence the overall trends for parkinsonism.
Acknowledgments
Funding/Support: This study was supported by award R01 AG034676 from the National Institute on Aging of the National Institutes of Health and by the Mayo Foundation for Medical Education and Research.
Role of the Funder/Sponsor: The funding organizations and sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Footnotes
Author Contributions: Dr Rocca had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Savica, Rocca.
Acquisition, analysis, or interpretation of data: Savica, Rocca, Grosssardt, Bower, Ahlskog.
Drafting of the manuscript: Savica, Rocca.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Grossardt.
Obtained funding: Rocca.
Administrative, technical, or material support; Savica, Rocca
Study supervision: Rocca
Conflict of Interest Disclosures: None reported.
REFERENCES
- 1.Morozova N, O'Reilly EJ, Ascherio A. Variations in gender ratios support the connection between smoking and Parkinson's disease. Mov. Disord. 2008;23(10):1414–1419. doi: 10.1002/mds.22045. [DOI] [PubMed] [Google Scholar]
- 2.Thacker EL, O'Reilly EJ, Weisskopf MG, et al. Temporal relationship between cigarette smoking and risk of Parkinson disease. Neurology. 2007;68(10):764–768. doi: 10.1212/01.wnl.0000256374.50227.4b. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Rocca WA, Bower JH, McDonnell SK, Peterson BJ, Maraganore DM. Time trends in the incidence of parkinsonism in Olmsted County, Minnesota. Neurology. 2001;57(3):462–467. doi: 10.1212/wnl.57.3.462. [DOI] [PubMed] [Google Scholar]
- 4.Akushevich I, Kravchenko J, Ukraintseva S, Arbeev K, Yashin AI. Time trends of incidence of age-associated diseases in the US elderly population: Medicare-based analysis. Age Ageing. 2013;42(4):494–500. doi: 10.1093/ageing/aft032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Lix LM, Hobson DE, Azimaee M, Leslie WD, Burchill C, Hobson S. Socioeconomic variations in the prevalence and incidence of Parkinson's disease: a population-based analysis. J. Epidemiol. Community Health. 2010;64(4):335–340. doi: 10.1136/jech.2008.084954. [DOI] [PubMed] [Google Scholar]
- 6.Mylne AQ, Griffiths C, Rooney C, Doyle P. Trends in Parkinson's disease related mortality in England and Wales, 1993-2006. Eur. J. Neurol. 2009;16(9):1010–1016. doi: 10.1111/j.1468-1331.2009.02715.x. [DOI] [PubMed] [Google Scholar]
- 7.Goldacre MJ, Duncan M, Griffith M, Turner MR. Trends in death certification for multiple sclerosis, motor neuron disease, Parkinson's disease and epilepsy in English populations 1979-2006. J. Neurol. 2010;257(5):706–715. doi: 10.1007/s00415-009-5392-z. [DOI] [PubMed] [Google Scholar]
- 8.Horsfall L, Petersen I, Walters K, Schrag A. Time trends in incidence of Parkinson's disease diagnosis in UK primary care. J. Neurol. 2013;260(5):1351–1357. doi: 10.1007/s00415-012-6804-z. [DOI] [PubMed] [Google Scholar]
- 9.Ajdacic-Gross V, Schmid M, Tschopp A, Gutzwiller F. Birth cohort effects in neurological diseases: amyotrophic lateral sclerosis, Parkinson's disease and multiple sclerosis. Neuroepidemiology. 2012;38(1):56–63. doi: 10.1159/000334632. [DOI] [PubMed] [Google Scholar]
- 10.Bower JH, Maraganore DM, McDonnell SK, Rocca WA. Incidence and distribution of parkinsonism in Olmsted County, Minnesota, 1976-1990. Neurology. 1999;52(6):1214–1220. doi: 10.1212/wnl.52.6.1214. [DOI] [PubMed] [Google Scholar]
- 11.Savica R, Grossardt BR, Bower JH, Ahlskog JE, Rocca WA. Incidence and pathology of synucleinopathies and tauopathies related to parkinsonism. JAMA Neurol. 2013;70(7):859–866. doi: 10.1001/jamaneurol.2013.114. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Savica R, Grossardt BR, Bower JH, Boeve BF, Ahlskog JE, Rocca WA. Incidence of dementia with Lewy bodies and Parkinson disease dementia. JAMA Neurol. 2013;70(11):1396–1402. doi: 10.1001/jamaneurol.2013.3579. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Sauver JL, Grossardt BR, Yawn BP, Melton LJ, 3rd, Rocca WA. Use of a medical records linkage system to enumerate a dynamic population over time: the Rochester Epidemiology Project. Am. J. Epidemiol. 2011;173(9):1059–1068. doi: 10.1093/aje/kwq482. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Sauver JL, Grossardt BR, Leibson CL, Yawn BP, Melton LJ, 3rd, Rocca WA. Generalizability of epidemiological findings and public health decisions: an illustration from the Rochester Epidemiology Project. Mayo Clin. Proc. 2012;87(2):151–160. doi: 10.1016/j.mayocp.2011.11.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Sauver JL, Grossardt BR, Yawn BP, et al. Data resource profile: the Rochester Epidemiology Project (REP) medical records-linkage system. Int. J. Epidemiol. 2012;41(6):1614–1624. doi: 10.1093/ije/dys195. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Rocca WA, Yawn BP, Sauver JL, Grossardt BR, Melton LJ., 3rd History of the Rochester Epidemiology Project: half a century of medical records linkage in a US population. Mayo Clin. Proc. 2012;87(12):1202–1213. doi: 10.1016/j.mayocp.2012.08.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Commission on Professional and Hospital Activities, National Center for Health Statistics . H-ICDA, Hospital Adaptation of ICDA. 2nd MI1973; Ann Arbor: [Google Scholar]
- 18.World Health Organization Manual of the international classification of diseases, injuries, and causes of death, based on the recommendations of the ninth revision conference, 1975, and adopted by the twenty-ninth World Health Assemby; Geneva. 1977. [Google Scholar]
- 19.Collins SJ, Ahlskog JE, Parisi JE, Maraganore DM. Progressive supranuclear palsy: neuropathologically based diagnostic clinical criteria. J. Neurol. Neurosurg. Psychiatry. 1995;58(2):167–173. doi: 10.1136/jnnp.58.2.167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Gilman S, Low P, Quinn N, et al. Consensus statement on the diagnosis of multiple system atrophy. American Autonomic Society and American Academy of Neurology. Clin. Auton. Res. 1998;8(6):359–362. doi: 10.1007/BF02309628. [DOI] [PubMed] [Google Scholar]
- 21.McKeith IG. Consensus guidelines for the clinical and pathologic diagnosis of dementia with Lewy bodies (DLB): report of the Consortium on DLB International Workshop. J Alzheimers Dis. 2006;9(3 Suppl):417–423. doi: 10.3233/jad-2006-9s347. [DOI] [PubMed] [Google Scholar]
- 22.de Rijk MC, Tzourio C, Breteler MM, et al. Prevalence of parkinsonism and Parkinson's disease in Europe: the EUROPARKINSON Collaborative Study. European Community Concerted Action on the Epidemiology of Parkinson's disease. J. Neurol. Neurosurg. Psychiatry. 1997;62(1):10–15. doi: 10.1136/jnnp.62.1.10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Bureau UC. 1990 Census of Population and Housing Summary Tape File 1. Washington, DC; US Government Printing Office: 1990. [Google Scholar]
- 24.Gardner W, Mulvey EP, Shaw EC. Regression analyses of counts and rates: Poisson, overdispersed Poisson, and negative binomial models. Psychol. Bull. 1995;118(3):392–404. doi: 10.1037/0033-2909.118.3.392. [DOI] [PubMed] [Google Scholar]
- 25.Carstensen B. Age-period-cohort models for the Lexis diagram. Stat. Med. 2007;26(15):3018–3045. doi: 10.1002/sim.2764. [DOI] [PubMed] [Google Scholar]
- 26.Seals RM, Hansen J, Gredal O, Weisskopf MG. Age-period-cohort analysis of trends in amyotrophic lateral sclerosis in Denmark, 1970-2009. Am. J. Epidemiol. 2013;178(8):1265–1271. doi: 10.1093/aje/kwt116. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Bronstein J, Carvey P, Chen H, et al. Meeting report: consensus statement-Parkinson's disease and the environment: collaborative on health and the environment and Parkinson's Action Network (CHE PAN) conference 26-28 June 2007. Environ. Health Perspect. 2009;117(1):117–121. doi: 10.1289/ehp.11702. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Singh M, Ramassamy C. Environmental Exposure to Pesticides and Neurodegenerative Diseases. J Alzheimers Dis Parkinsonism. 2012;2(3):e116. [Google Scholar]
- 29.Escobedo LG, Peddicord JP. Smoking prevalence in US birth cohorts: the influence of gender and education. Am. J. Public Health. 1996;86(2):231–236. doi: 10.2105/ajph.86.2.231. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Trends in Tobacco Use. Jul, 2011. American Lung Association, Research and Program Services, Epidemiology and Statistics Unit.
- 31.Savica R, Grossardt BR, Bower JH, Ahlskog JE, Rocca WA. Risk factors for Parkinson's disease may differ in men and women: an exploratory study. Horm. Behav. 2013;63(2):308–314. doi: 10.1016/j.yhbeh.2012.05.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Frigerio R, Sanft KR, Grossardt BR, et al. Chemical exposures and Parkinson's disease: a population-based case-control study. Mov. Disord. 2006;21(10):1688–1692. doi: 10.1002/mds.21009. [DOI] [PubMed] [Google Scholar]
- 33.Tobi EW, Slieker RC, Stein AD, et al. Early gestation as the critical time-window for changes in the prenatal environment to affect the adult human blood methylome. Int. J. Epidemiol. 2015;44(4):1211–1223. doi: 10.1093/ije/dyv043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.irtz D, Thurman DJ, Gwinn-Hardy K, Mohamed M, Chaudhuri AR, Zalutsky R. How common are the "common" neurologic disorders? Neurology. 2007;68(5):326–337. doi: 10.1212/01.wnl.0000252807.38124.a3. [DOI] [PubMed] [Google Scholar]
- 35.Dorsey ER, Constantinescu R, Thompson JP, et al. Projected number of people with Parkinson disease in the most populous nations, 2005 through 2030. Neurology. 2007;68(5):384–386. doi: 10.1212/01.wnl.0000247740.47667.03. [DOI] [PubMed] [Google Scholar]