Familial aggregation of Parkinson disease in Utah

Objective: To describe clustering of death from Parkinson disease (PD) in relatives in a large US study. Methods: We analyzed the Utah Population Database resource, which includes genealogy data of more than 2.7 million individuals linked to 519,061 individuals with a Utah death certificate (DC). We identified individuals whose DC included PD as a cause of death using ICD coding. In those individuals whose Utah DC listed PD as a cause of death, the relative risk (RR) of death with PD was determined among close and distant relatives using sex-, birth year–, and birthplacespecific rates. Results: We identified 4,031 individuals whose DC indicated PD. Among 18,127 first-degree relatives of probands with a Utah DC, the RR of death with PD was significantly increased (RR 5 1.82, 95% confidence interval [CI] 1.61–2.04). The RR of death with PD was also significantly increased among 40,546 second-degree relatives with a Utah DC (RR 5 1.44, 95% CI 1.29– 1.60) and among 93,398 third-degree relatives with a Utah DC (RR5 1.10, 95%CI 1.03–1.18). Conclusions: Significant evidence for excess familial clustering was observed for PD deaths. The excess familial clustering and the significantly elevated RRs for PD among close and distant relatives strongly support a genetic contribution to PD mortality. These results confirm and expand the results of previous studies of PD by quantifying the risk of PD death among more distant relatives. Neurol Genet 2016;2:e65; doi: 10.1212/NXG.0000000000000065 GLOSSARY CI 5 confidence interval; DC 5 death certificate; dGIF 5 distant Genealogical Index of Familiality; GIF 5 Genealogical Index of Familiality; ICD 5 International Classification of Diseases; PD 5 Parkinson disease; RR 5 relative risk; UPDB 5 Utah Population Database. Parkinson disease (PD) is a progressive neurodegenerative disorder of middle age and the elderly. PD is pathologically characterized by the deposition of protein aggregates of a-synuclein in the nervous system that constitute Lewy bodies in the pathology examination. The incidence of PD in the population is 18 per 100,000 person-years, and it is more frequent in men than in women. The causes of PD are still unknown; however, it is widely accepted that a complex relationship between lifelong environmental risk factors and genetic predisposition is crucial in the development of PD. There is still limited knowledge on the familial aggregation of PD because of methodologic limitations, small sample sizes, and the frequent lack of data for distant relatives. Researchers in Iceland previously published a population-based description of familial aggregation of PD. The authors identified a genetic contribution to late-onset PD, separate from early onset, and estimated relative risks (RRs) for PD in close relatives. To further explore the complex gene–environment relationship of PD, we analyzed the Utah Population Database (UPDB), a From the Department of Neurology (R.S.), Mayo Clinic, Rochester, MN; Division of Genetic Epidemiology (L.A.C.-A.), Department of Internal Medicine, and Department of Neurology (R.S., S.P.), University of Utah School of Medicine, Salt Lake City, UT; and George E. Wahlen Department of Veterans Affairs Medical Center (L.A.C.-A.), Salt Lake City, UT. Funding information and disclosures are provided at the end of the article. Go to Neurology.org/ng for full disclosure forms. The Article Processing Charge was paid by the authors. This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially. Neurology.org/ng © 2016 American Academy of Neurology 1 a 2016 American Academy of Neurology. Unauthorized reproduction of this article is prohibited. population-based resource linking death certificates (DCs) from 1904 for the large homogeneous Utah population to Utah genealogic data dating back to the mid-1800s. The objective of this study was to describe familial clustering of PD in the Utah population, to estimate the familial RR for PD mortality, and to provide a current description of the risk of PD in family members of patients with PD using DC data. METHODS Genealogic data. The UPDB includes birth and death data of more than 7 million individuals, with some records extending back more than 12 generations. This computerized genealogic resource is derived from multiple record-linked data sources. The original Utah genealogy includes complete 3generation genealogic data for the Utah Mormon pioneers (members of the Church of Jesus Christ of Latter-day Saints) and their descendants up to 1972. Since then, the original genealogy data have been expanded with Utah vital records such as birth certificates (using father, mother, and child trios). For the analyses reported in this study, we consider only those individuals in the UPDB with at least 3 generations of genealogy data who are related to the original Utah genealogy. Utah’s founding pioneers were composed of a sizable, largely unrelated mixture of Northern European populations. This population continued to have high rates of immigration for years after Utah’s founding in 1847. Studies using pedigree data, migration matrices, and isonymy have all shown low levels of inbreeding and similarity to Northern European populations in the founding population of Utah. PD phenotype data. We analyzed more than 2.7 million individuals in the UPDB belonging to at least 3 generations of genealogic data and connected to the original Utah genealogy. Within this population of individuals with genealogy data, 519,061 individuals had a Utah DC. The cause of death on Utah DCs was coded using the ICD, with the revision used (i.e., ICD-6 to ICD-10) depending on the decade of death. For all deaths occurring before 1956, ICD-10 coding was assigned. Table 1 shows the frequency of the 4,031 PD deaths by ICD revision; 2,546 of these deaths occurred in males and 1,485 in females. The majority of the deaths from PD were in more recent years; 2,407 deaths occurred after 1999. Genealogical Index of Familiality. The Genealogical Index of Familiality (GIF) statistic was developed to test hypotheses concerning excess relatedness among individuals sharing a specific phenotype using the UPDB. The GIF analysis considers all genetic relationships between cases and measures the average relatedness among all possible pairs within a set of individuals. The pairwise relatedness measure implements the Malécot coefficient of kinship, defined as the probability that randomly selected homologous genes from the 2 individuals are identical by descent from a common ancestor. The case GIF is defined as the average of the coefficients of kinship between all possible pairs of cases (310). The pairwise relatedness of a set of cases is compared with the expected pairwise relatedness for a group of similar individuals in the UPDB. One thousand sets of matched controls that also had a DC were randomly selected from the UPDB. To test the hypothesis of no excess relatedness among the set of PD cases identified from DCs, the case GIF was compared with the empirical distribution of GIF statistics estimated from 1,000 sets of matched controls. Controls were randomly selected from all individuals with genealogic data and a DC and were matched to cases by birth cohort (5 years), sex, and birthplace (Utah or not). These analytical methods, including GIF analysis, have been previously applied to describe the familial and genetic contribution to mortality of multiple phenotypes, including intracranial aneurysms, influenza, asthma, and amyotrophic lateral sclerosis, among others. The GIF statistic can also be estimated while ignoring all close relationships (relationships closer than first cousins); this allows a test of the hypothesis that excess relatedness has been observed only for distant relationships, which are unlikely to share common risk factors or exposures. This test is termed the distant GIF (dGIF) test and allows determination of whether the excess familial clustering observed might be due at least in part to a genetic contribution. RR in relatives. To estimate the RR of PD among relatives, the observed number of PD deceased relatives was compared with the expected number of PD deaths in relatives. RR estimation is described below for the example of the RR of PD death among first-degree relatives of PD death cases. RR estimation is similar for other degrees of relationship. All individuals in the UPDB who belong to at least 3 generations of genealogy and who have a coded cause of death were assigned to 1 of 132 birth year– (5 years), sex-, and birthplacespecific (Utah or not) cohorts. The rate of death with PD for each cohort was estimated as the total number of individuals with PD reported as a cause of death in each cohort divided by the total number of individuals with a DC in the cohort. The expected number of first-degree relatives dying with PD was estimated by counting all relatives of probands who have a DC (by cohort, each relative counted only once regardless of how many times he or she was identified as a relative of the degree of interest), multiplying the number of deceased first-degree relatives (per cohort) by the cohort-specific rate of death with PD, and then summing over all cohorts. RRs were estimated as the observed number of first-degree relatives with PD divided by the expected number of first-degree relatives with PD; this is an unbiased estimator of RR and is calculated similarly for different relationships. Two-tailed probabilities were calculated under the null hypothesis RR5 1.0, under the assumption that the number of observed deaths follows a Poisson distribution with mean equal to the expected number of deaths; confidence intervals (CIs) for the RR were calculated as described elsewhere. High-risk pedigrees. Using data for all ancestors of each PD case, all clusters of PD cases descending from a common ancestor can be identified. Using the same methods described for RRs, these clusters (or pedigrees) can be tested for a significant excess of PD cases by counting the observed PD deaths in the pedigree Table 1 Frequency of Parkinson disease deaths by ICD revision ICD revision ICD code Frequency