Familial Risks for Main Neurological Diseases in Siblings Based on Hospitalizations in Sweden Kari Hemminki,1,2 Kristina Sundquist,2 and Xinjun Li2 1 Division of Molecular Genetic Epidemiology, German Cancer Research Center (DKFZ), Heidelberg, Germany 2 Center for Family Medicine, Karolinska Institute, Huddinge, Sweden ecent successes in identifying the underlying disease mechanisms have been discovered, including Rgenetic mechanisms for neurological diseases, DNA repeat expansions in coding and noncoding particularly for their Mendelian forms, have had pro- sequences, now found in over 40 neurological dis- found implications for their diagnostics, treatment eases (Gatchel & Zoghbi, 2005). Other achievements and classification. However, there has never been include molecular and pathophysiological characteri- an attempt to compare familial risks in a systematic zation of severe diseases such as Huntington’s chorea, way among and between the main neurological dis- hereditary ataxias, muscular dystrophies and eases. Familial risks were here defined for siblings myotonic disorders, all with a main heritable etiology who were hospitalized because of a neurological (Bertram & Tanzi, 2005; Dalkilic & Kunkel, 2003; disease in Sweden. A nationwide database for neu- Day & Ranum, 2005; Muntoni & Voit, 2004; rological diseases was constructed by linking the Ropper & Brown, 2005; Taroni & DiDonato, 2004). Multigeneration Register of 0- to 69-year-old siblings In the major neurological diseases including to the Hospital Discharge Register for the years Alzheimer’s disease, Parkinson’s disease, epilepsy and 1987 to 2001. Standardized risk ratios were calcu- migraine, some heritable subtypes have been noted, lated for affected sibling pairs by comparing them to but these explain only a small proportion of the etiol- those whose siblings had no neurological disease. There were three main results. First, it was shown ogy of these diseases (Bertram & Tanzi, 2005; Estevez that all disease groups had a familial risk, with the & Gardner, 2004; Guerrini et al., 2003; Robinson & exception of transient ischemic attacks, and the Gardiner, 2004; Ropper & Brown, 2005; Scheffer & risks could be ranked from the highest (3451) for Berkovic, 2003; Wessman et al., 2004). In multiple Huntington’s disease to the lowest (2.1) for inflam- sclerosis, familial aggregation is recognized but high matory diseases. Second, increased familial risks penetrant susceptibility genes remain to be identified were shown for disease subtypes for which suscep- (Kalman & Leist, 2004; Nielsen et al., 2005). tibility genes or familial clustering have not been Familial clustering of a disease is a measure of its demonstrated previously, including multiple sclero- heritability, provided that shared environmental sis, sleep apnea, nerve, nerve root and plexus factors can be excluded. For some high penetrant neu- disorders, and cerebral palsy. Third, based on the rological diseases familial aggregation has been available sample size there was no convincing evi- striking, and pedigrees of index cases have shown dence for familial comorbidity between the disease typical Mendelian segregation. However, for the most groups, suggesting that the factors causing familial common neurological diseases, such as Alzheimer’s aggregation, probably usually heritable genes, are disease, Parkinson’s disease, epilepsy, migraine and distinct for each subtype. The high familial risks for multiple sclerosis, most patients lack a family history, neurological disease imply heritable etiology and although familial aggregation has been demonstrated opportunities for identification of further susceptibil- in twin and other types of family studies (Bertram & ity genes. Tanzi, 2005; Kalman & Leist, 2004; Mulder et al., 2003; Nielsen et al., 2005; Robinson & Gardiner, 2004; Svensson et al., 2003; Wessman et al., 2004). Neurological diseases are medical conditions for Moreover, most of the family studies have been small which molecular genetic techniques have probably case-control studies that have relied on reports of achieved the greatest success, contributing to the characterization of disease etiology and mechanisms and to improvements in diagnostics and disease classi- Received 9 January, 2006; accepted 10 February, 2006. fication (Bertram & Tanzi, 2005; Ropper & Brown, Address for correspondence: K. Hemminki, DKFZ, Im Neuenheimer 2005). In the course of these studies entirely novel Feld 580, D-69120 Heidelberg, Germany. E-mail: [email protected] 580 Twin Research and Human Genetics Volume 9 Number 4 pp. 580–586 Downloaded from https://www.cambridge.org/core. IP address: 170.106.33.14, on 26 Sep 2021 at 15:14:49, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms. https://doi.org/10.1375/twin.9.4.580 Familial Neurological Diseases neurological diseases in family members, the accuracy stayed at least one night in the hospital, usually in of which may be highly variable. There has never been wards with neurology consultants or in neurology an attempt to compare familial risks in a systematic departments; the Register does not include outpatients way among and between the main neurological dis- in hospitals or healthcare centers. Diagnoses were eases. Heritable comorbidity between neurological reported according to the 9th (1987–1996) and 10th diseases would have mechanistic implications regard- (1997–2001) versions of the International Classification ing the specificity of the gene defect to damage certain of Diseases (ICD; World Health Organization, 1977, nerve cell functions but not others. Almost all the 2004), classified in 17 groups of diseases. All linkages genes that have been linked to neurological diseases were performed using the national 10-digit civic iden- are expressed in most nerve cells, and the present tification number that is assigned to each person in pathophysiologial understanding does not explain Sweden for his or her lifetime. This number was disease specificity; for example, why superoxide dis- replaced by a serial number for each person in order mutase 1 (SOD1) mutations only cause amyotrophic to provide anonymity and to check that each individ- lateral sclerosis, or why the epsilon 4 allele of the ual was only entered once, for his or her first apolipoprotein E gene primarily affects Alzheimer’s hospitalization for a neurological disease. Over 6.9 disease (Bertram & Tanzi, 2005). Certain specific million individuals were included in the second gener- forms of epilepsy are due to mutations in genes coding ation of the neurological database. for ion channel proteins in nerve cells, hence the term Person-years were calculated from start of follow- channelopathy (Robinson & Gardiner, 2004; up on January 1, 1987, until hospitalization for the Wessman et al., 2004). Familial hemiplegic migraine, first neurological disease, death, emigration, or closing episodic ataxias and spinocerebellar ataxia type 6 are date, December 31, 2001. Age-specific incidence rates also channelopathies (Kors et al., 2004). Mutations in were calculated for the whole follow-up period, a single gene, CACNA1A encoding a calcium channel divided into five 5-year periods, and they were stan- pore subunit gene, predispose to these three diseases dardized to the European population. Standardized (Kors et al., 2004). So far, however, the evidence for incidence ratios (SIRs) were calculated as the ratio of such shared phenotypic effects is limited. observed (O) to expected (E) number of cases. The The availability of disease-specific data on all hos- expected number of cases was calculated for age (5- pitalizations in Sweden prompted us to analyze year groups), sex, period (5-year groups), region and familial risks in siblings aged 0 to 69 years between socioeconomic status–specific standard incidence rates the main neurological diseases with the particular aim derived from the MigMed database. Sibling risks were of observing how the familial risks compare and calculated for men and women with siblings affected whether they are shared between the individual diag- with concordant (same) or discordant (different) neu- nostic groups. The usefulness of the Swedish family rological diseases, compared with men and women dataset has been demonstrated earlier in studies of whose siblings were not affected by these conditions, familial migraine and aortic aneurysms (Hemminki et using the cohort methods as described (Hemminki et al., 2005, 2006). al., 2001). In rare families where more than two sib- lings were affected, each was counted as an individual Materials and Methods patient. Confidence intervals (95% CI) were calcu- The research database used for this study, the neuro- lated assuming a Poisson distribution, and they were logical database, is a subset of the national MigraMed adjusted for dependence between the sibling pairs database at Karolinska Institute, Centre for Family (Hemminki et al., 2001). and Community Medicine. The MigMed database was compiled using data from several national Swedish Results registers provided by Statistics Sweden, including the We analyzed risks for siblings aged 0 to 69 years to be Multigeneration Register in which persons (second hospitalized for a neurological disease, divided into 17 generation) born in Sweden in 1932 and thereafter are subtypes, in Sweden between 1987 and 2001. Only registered shortly after birth and are linked to their the first hospitalization was considered. The numbers parents (first generation). Sibships could only