Age-specific cumulative probability of escaping Parkinson disease (PD). Kaplan-Meier curves for parents and siblings of patients with early-onset PD (triangles, left curve), parents and siblings of late-onset PD (circles, middle curve), and parents and siblings of controls (squares, right curve). For early-onset patients vs controls, P<.001; for late-onset patients vs controls, P = .02; and for early-onset patients vs late-onset patients, P<.001.
Payami H, Zareparsi S, James D, Nutt J. Familial Aggregation of Parkinson DiseaseA Comparative Study of Early-Onset and Late-Onset Disease. Arch Neurol. 2002;59(5):848-850. doi:10.1001/archneur.59.5.848
It is unclear whether late-onset Parkinson disease (PD), which is the most typical and most common form of the disease, has a familial component. Evidence for familial aggregation is key to whether research should focus on gene discovery or search for environmental factors.
To investigate familial aggregation of early-onset and late-onset PD separately.
Using survival methods, age-specific risk of PD was calculated and compared for 525 parents and siblings of 117 patients with early-onset PD, 1642 parents and siblings of 343 patients with late-onset PD, and 522 parents and siblings of 114 controls. The index patients were ascertained from a movement disorder clinic. Spouses and friends served as controls.
Compared with the relatives of controls, age-specific risk of PD was increased 7.76-fold in the relatives of patients with early-onset disease (P<.001) and 2.95-fold in the relatives of those with late-onset disease (P = .02).
Late-onset PD has a significant familial component. The magnitude of recurrence risk to relatives suggests a genetic etiology, without ruling out the possibility of a coexisting environmental component.
PARKINSON DISEASE (PD) is a common neurodegenerative movement disorder.1 Familial aggregation of PD was established by case-control studies in the last decade.2- 7 Taken as evidence for a genetic component, studies were launched to search for the putative PD genes. Several loci have since been identified, 3 of which (α-synuclein, parkin, and ubiquitin carboxy-terminal hydroxylase L1) have been characterized and shown to contain mutations that cause PD.8- 11 All 3 PD genes identified to date are associated with early-onset PD (prior to age 50 years). No genes have yet been identified for late-onset PD, which is the more typical form of PD.
A recent study of twins found a high monozygotic concordance rate for early-onset PD, but not for late-onset PD, concluding that genetic factors do not play a major role in typical PD.12 If confirmed, this finding will have serious implications for genetic studies that are underway, as well as for setting the research direction in the future. The low twin concordance rate seems to contradict the case-control studies that showed higher age-specific risk of PD in relatives of patients. Case-control studies did not separate early-onset PD from late-onset PD. Therefore, the increased risk to relatives of patients may have been driven by the inclusion of patients who had early-onset PD, which is now known to have a strong genetic component. Alternatively, the low twin concordance rate may have been due to inadequate follow-up rather than absence of a genetic component. In concordant pairs, nearly 10 years elapsed before the second twin was diagnosed. Therefore, as authors have pointed out, additional concordant pairs are likely to be identified with continued follow-up. Here, we report the first case-control family study of PD to examine early-onset and late-onset disease separately. Since a key issue is proper control for age, we used the case-control study design and survival analysis methods that allow for age-specific risk calculations. The null hypothesis was that, compared with relatives of controls, the risk of developing PD is elevated in the relatives of patients with early-onset PD, but not in the relatives of patients with late-onset PD.
Four hundred sixty white patients with PD were ascertained from the movement disorder clinic at Oregon Health Sciences University, Portland. Inclusion criterion was clinical diagnosis of idiopathic PD by standard criteria.13 Patients were selected randomly, regardless of family history or age at onset. Patients who were referred for genetic studies were excluded to avoid overrepresentation of familial cases. The first 114 patients enrolled were asked to invite their spouses or a friend of similar age and ethnicity (all white) to serve as controls. Subjects signed an informed consent approved by the institutional review board.
Family histories were obtained using a standardized, self-administered questionnaire. Relatives who had PD were considered affected. Age at onset of first PD symptom was established for 80 of 82 affected relatives. For the 2 with unknown onset ages, we assigned age at diagnosis (this was conservative since both were among the relatives of patients). Of the 460 index patients, 117 had onset of disease before age 50 years (early onset), and 343 had onset at or after age 50 years (late onset) (Table 1).
The analysis included parents and siblings of patients and controls (Table 1), excluding the index patients and controls. Time followed was specified as age at onset for the affected individuals, and as age at last contact or age at death for the unaffected. Age-specific cumulative risk of PD was estimated and plotted using Kaplan-Meier survival analysis and compared using log-rank statistics. Age-specific hazard ratios were calculated using the Cox proportional hazard model.
The age-specific cumulative risk of PD, shown by Kaplan-Meier curves (Figure 1), was significantly different for the 3 groups of relatives. Parents and siblings of patients with early-onset disease had the highest age-specific risk, while parents and siblings of patients with late-onset disease were intermediate, and parents and siblings of control subjects had the lowest risk (Figure 1). Compared with the relatives of controls, age-specific risk of PD was increased 7.76-fold (P<.001) in the relatives of patients with early-onset PD, and 2.95-fold (P = .02) in the relatives of patients with late-onset disease (Table 2). Clearly, the hypothesis that late-onset PD has no familial component was rejected.
The crude prevalence of PD was twice as high in parents than in siblings; however, parents were older than siblings (Table 1). To determine whether the difference in prevalence rates was real or an artifact of age, we compared age-specific risk of PD in parents vs siblings. The age-specific rates were nearly identical (hazard ratio [HR] = 0.80, P = .33), indicating no intergenerational difference once age is taken into account. One can therefore expect the number of siblings with PD to rise as they age.
In relatives of patients with late-onset PD, age-specific risk was significantly higher in men than in women (HR = 1.89, P = .03). No significant gender difference was detected among the relatives of patients with early-onset PD (HR = 0.7, P = .40), reflecting a nonsignificant decreased risk in men.
This study revealed a significant familial aggregation for late-onset PD. Although the magnitude of the risk to relatives of patients with late-onset PD was considerably lower than that for relatives of patients with early-onset PD, it was still significantly higher than that for the relatives of controls. The postulate that the familial component is present only in early-onset disease, and not in late-onset PD, was ruled out. Familial aggregation of late-onset PD is further supported by a recent genealogical study from Iceland that showed that patients with late-onset PD were significantly more related to each other than were subjects in matched groups of controls.14
Overall (combining early- and late-onset cases), parents and siblings of patients had a 4-fold-higher age-specific risk of developing PD than relatives of controls. The risk was considerably higher for relatives of patients with early-onset PD than for relatives of those with late-onset disease, indicating a much stronger genetic etiology in the early-onset subtype. Parkinson disease is more common in men.5,15 We found no gender difference in early-onset PD, but a significantly higher risk to men among the relatives of patients with late-onset PD, suggesting that the gender difference in PD is a characteristic of late-onset disease.
The incidence and prevelance of PD in relatives of controls were similar to the general population and other control groups.5,15 Thus, it is unlikely that the results were biased by underestimating PD in controls. The patients, however, were ascertained from a specialty clinic, and may have had a higher proportion of familial and early-onset cases than the general population. Therefore, the results cannot be generalized until confirmed in population-based studies. At this point, we have demonstrated significant familial aggregation of late-onset PD in a clinic population of white patients residing in the Pacific Northwest.
Familial aggregation may arise from shared environment or genes. Complex segregation analysis has suggested a genetic component for typical late-onset PD.16 Furthermore, a risk ratio of 3, as seen here for relatives of patients with late-onset disease vs controls, is too high to be explained by shared environment alone.17 Thus, present results argue in favor of a genetic etiology of late-onset PD, without ruling out the possibility of a coexisting environmental component.
Accepted for publication December 19, 2001.
Author contributions: Study concept and design (Drs Payami and Zareparsi); acquisition of data (Dr Zareparsi); analysis and interpretation of data (Drs Payami and Zareparsi, and Ms James); drafting of the manuscript (Dr Payami); critical revision of the manuscript for important intellectual content (Drs Payami, Nutt, and Zareparsi, and Ms James); statistical expertise (Drs Payami and Zareparsi); obtained funding (Dr Payami); study supervision (Dr Payami).
Support and funding was provided by grant R01 NS36960 from the National Institute of Neurological Disorders and Stroke (Bethesda, Md) and the National Parkinson Foundation Inc (Miami, Fla).
We wish to thank the patients, their families, and the volunteers who participated in this study; and Nicole Lee, BS, Melissa Gonzales McNeal, MS, and Kim Larson, MD, for assisting with data collection.
Corresponding author and reprints: Haydeh Payami, PhD, CR131, Oregon Health Sciences University, 3181 SW Sam Jackson Park Rd, Portland, OR 97201 (e-mail: email@example.com).