Familial aggregation of rheumatoid arthritis and co-aggregation of autoimmune diseases in affected families: a nationwide population-based study

Abstract

Introduction

Despite evidence suggesting higher relative risks (RRs) of RA in first-degree relatives of RA patients, data from the Norfolk Arthritis Register in the UK [1] and the Danish twin registry [2] indicated only a small, non-significant increase in recurrence risk ratios. Twin studies have estimated that heritability ranged between 12 and 68%. These conflicting results present difficulties in genetic counselling for individuals with a positive family history. Another interesting association is that autoimmune diseases tend to cluster together. A family history of RA is associated with an increased risk of autoimmune diseases [3, 4] and non-RA arthritis-related disease [5]. Overall, it seems that RA and other autoimmune diseases share part of the same pathogenesis.

Therefore, we conducted this nationwide study using genealogy and linked health information derived from the Taiwan National Health Insurance (NHI) database to determine familial clustering of RA and to assess the relative contribution of genetic, shared and non-shared environmental factors to RA susceptibility. We also estimated the RRs of other autoimmune diseases associated with a family history of RA.

Methods

Study population

This study was approved by the Institutional Review Board of the Chang Gung Memorial Hospital. A cohort of all individuals registered in the Taiwan NHI in 2010 was established using data from the Registry for NHI beneficiaries, Registry for catastrophic illness patients and data sets of ambulatory care expenditures and details of ambulatory case orders, all of which are parts of the NHI Research Database. Individuals without valid insurance status were excluded from analysis. Given that the NHI is compulsory for all residents in Taiwan, the coverage rate is >99.5% in 2010.

The NHI Database contains comprehensive information of beneficiaries regarding gender, date of birth, place of residence, details of insurance, family relationships, vital status and details of clinical information. Multiple studies have validated the NHI database, and in general the accuracy of diagnosis recorded in this database is high [6–9]. Methods to identify first-degree relatives have been described before (supplementary Methods, available at Rheumatology Online) [10–12]. In brief, the registry of beneficiaries contains the identifiers of the relationships between the insured person (who paid the insurance fee) and his/her dependents. Only blood relatives and spouses are eligible to be dependents of an insured person. A birth certificate issued by the medical facility who delivered the child or a DNA parentage testing for those who were not born in medical facilities is required for a child to register as a dependent of their parents. The identifiers of family relationships recorded in the NHI database allow us to establish family relationships. Using data from 28 402 865 (alive and deceased) beneficiaries registered in the NHI between 1995 and 2010, we have identified 21 009 551 parent–child relationships, 17 168 340 full sibling pairs and 342 066 twin pairs (for details please refer to the supplementary Methods, sections Study population and Genealogy reconstruction, available at Rheumatology Online).

Ascertainment of RA and other autoimmune diseases

In Taiwan, patients with RA and other autoimmune diseases included in this study are entitled to waive medical co-payment. The panel commissioned by NHI administration reviews the diagnosis in compliance with the updated classification criteria. For instance, the ACR revised criteria for classification of RA were used to assist the review of certificate applications for RA [13]. The Registry for Catastrophic Illness Patients was our primary data source to identify patients with RA and other autoimmune diseases. The full code lists are presented in the supplementary Methods, sections Ascertainment of RA and other autoimmune diseases and Covariates, available at Rheumatology Online.

Statistical analysis

The prevalence of RA was calculated for the general population and for individuals with affected first-degree family members. We calculated RRs of RA as the adjusted prevalence ratios between first-degree relatives of an individual with RA and the general population. RRs of RA were defined as the adjusted prevalence ratios between individuals with a positive RA family history and the general population, which is equivalent to relative recurrence ratio defined by Risch [14]. The Breslow–Cox proportional hazards model was used to estimate RRs by applying an equal follow-up time for all subjects. We used the marginal approach to account for clustering effects [15]. Individual demographic and socio-economic information regarding age, sex, occupation, income level and place of residence was obtained from the registry of beneficiaries. The RR was adjusted for age, sex, socio-economic factors and family size; an approach that has been applied and validated previously in other diseases [16]. We further estimated the extent of familial co-aggregation of other autoimmune diseases in RA-affected families.

We estimated heritability (the proportion of phenotypic variance that is attributable to genetic factors) and the familial transmission (the proportion of genetic and shared environmental contribution) using the polygenic liability model to calculate both measures. As kinship and sex of the affected relative may also influence familial risk, we fitted models separately according to kinship and sex of affected relatives (mother, father, daughter, son, sister, brother, twin sister and brother). We excluded twins from the sibling analyses. In addition to first-degree relatives, we also estimated the RR for spouses. The RR was estimated for the number of affected first-degree kinships (father, mother, son, daughter, brother, sister). In this model, we compared the risk of RA in individuals with one or two affected first-degree relatives with the risk in the general population. To measure the degree of similarity in different types of relatives, we estimated tetrachoric correlations for each category of first-degree relationships stratified by sex of RA patients and their relatives, assuming that there is a continuous normally distributed liability underlying the diagnosis of RA.

Heritability was defined as the proportion of phenotypic variance that is attributable to genetic factors, and the familial transmission is the proportion of genetic and shared environmental contribution (supplementary Methods, section on Statistical analysis, available at Rheumatology Online). To account for contributions of shared environmental factors to phenotypic variance, we used the spouse as a control, assuming that spouses share the family environment but have no close genetic similarity with blood family members. We restricted family history to first-degree relatives and assumed an average of two siblings in a family. All tests of statistical hypothesis were done on the two-sided 5% level of significance. All analyses were performed using SAS 9.4 (SAS Institute, Cary, NC, USA).

Results

The study population comprised 23 658 577 individuals enrolled in NHI in Taiwan in 2010. We identified 37 482 patients with a diagnosis of RA, giving a crude prevalence of 0.16%. Women had a significantly higher prevalence (0.25%) than men (0.07%), with a female:male ratio of 3.6 (supplementary Table S1, available at Rheumatology Online). In the general population of Taiwan in 2010, 65 482 (0.28%) individuals had at least one first-degree relative with RA; among these, 364 had RA, giving a prevalence of 1.33%. For individuals with affected first-degree relatives with RA, the age-specific prevalence of RA is significantly higher than the age-specific prevalence in the general population (Fig. 1).

Fig. 1

RA prevalence in people with a family history and the general population

Individuals with a first-degree relative with RA (continuous line); general population (dashed line).

Open in new tabDownload slide

Overall, having an affected first-degree relative was associated with an adjusted RR of 5.28 (4.60, 6.07) for RA. Table 1 also presents the adjusted RR for RA and 95% CIs for different affected first-degree relatives stratified by sex. Overall, individuals with affected relatives of female and male gender had respective RRs (95% CIs) for RA of 5.21 (4.47, 6.08) and 6.12 (4.90, 7.65), respectively. The RRs (95% CIs) for RA were associated with the degree of genetic distance between family relatives. The RRs were 328.27 (135.95, 792.63) for co-twins (with highest genetic similarity) of RA patients; 11.97 (8.68, 16.52) for siblings; 4.86 (4.16, 5.67) for parents; 4.65 (3.92, 5.50) for offspring; and 2.32 (1.83, 2.95) for spouses (without genetic similarity).

The tetrachoric correlation coefficient (95% CI) was 0.44 (0.34, 0.54) for twin, 0.25 (0.23, 0.27) for full sibling, 0.18 (0.16, 0.20) for offspring, 0.13 (0.11, 0.15) for parent and 0.08 (0.16, 0.20) for spouse. Using a threshold liability model, we estimated that familial transmission was 59.4% (95% CI: 50.3, 69.5%) and heritability was 43.5% (33.9, 54.1%).

Supplementary Table S2, available at Rheumatology Online, presents the adjusted RR (95% CIs) for other autoimmune diseases in individuals with affected first-degree relatives compared with the general population. The RR (95% CI) in individuals with a first-degree relative with RA was 2.91 (2.49, 3.42) for SLE; 2.92 (1.62, 5.25) for SSc; 3.13 (2.50, 3.93) for primary SS; 0.95 (0.36, 2.51) for idiopathic inflammatory myositis; 1.96 (1.54, 2.48) for type 1 diabetes mellitus; 3.32 (1.82, 5.95) for multiple sclerosis; 1.31 (1.31, 2.43) for IBD; 2.76 (2.46, 3.10) for AS; and 1.65 (1.54, 1.77) for psoriasis.

Discussion

The present study indicates that first-degree relatives of RA patients have a 5-fold increased risk of RA compared with the general population. Overall genetic and shared environmental factors contributed to ∼60% of phenotypic variance of RA and genetic factors contributed three-quarters of this familial transmission. In addition, a family history of RA is also a risk factor for other autoimmune disease, indicating that familial factors are shared risk factors for autoimmune diseases. These data could be of value when counselling families with RA patients in comprehensive risk estimates of RA and other autoimmune diseases as well as estimation of phenotypic variance partitioning.

Previous studies report that RA patients tend to have first-degree relatives with the disease, with 7–22% of RA patients having a positive family history. These studies are probably subject to ascertainment bias, which arises because of different methods of case ascertainment, and therefore the RRs reported should be interpreted with caution. More recent population-based studies in Denmark and Sweden used the same case definitions for cases and controls and generally reported a 3-fold risk of RA for first-degree relatives of RA patients [17, 18]. Our study indicates a 5-fold RA risk for first-degree relatives of RA patients, which seems to be higher than estimates reported by Danish and Swedish studies. There are two possible explanations. Firstly, our population is mainly ethnic Taiwanese, who are known to have a lower background population RA prevalence than Western populations. Secondly, our case definition of RA depends on the issuance of a catastrophic illness certificate, which requires fulfilment of stringent criteria. This may lead to identification of more severe disease.

Shared environmental factors also contribute to familial susceptibility. In this study, we used the spouse as a control to define the shared environmental contribution. We found a high spouse correlation, suggesting the existence of a shared environmental contribution. Smoking has been found to be a strong environmental risk factor for RA. Our study suggests that at least part of RA phenotypic variance is contributed by shared environmental factors. Further study is needed to identify modifiable factors that could expose risk to family members.

Twin studies from the UK and Finland generally estimate a heritability of 50–66% [19, 20]. However, data from a Danish twin study found a heritability of only 12% [21]. A recent Swedish study investigating 90 372 RA patients found an overall heritability of 40%, with a higher heritability for patients with a positive test for ACPA (∼50%) than for those with a negative test (∼20%) [18]. Our study indicates a heritability of 43.5% for RA, which is close the Swedish estimate. The differences reported in the present study from earlier published twin studies are multifactorial. The difference between monozygotic and dizygotic variance is the base for heritability estimation in a classic twin study, but a higher degree of environmental similarity in monozygotic than dizygotic twins may result in an overestimation of heritability owing to a mixture of shared environmental effects. Furthermore, our study used the entire national population of Taiwan, and the case definitions of RA and other autoimmune diseases were based on physician diagnoses, which were scrutinized by expert panels. Given that all residents of Taiwan are eligible for the same procedure, differential misclassification is expected to be minimal.

Our results and previous evidence indicate that a family history of RA is a strong risk factor for other autoimmune diseases. Families with RA patients are reported to be enriched with cases of autoimmune diseases such as SLE, SSc and SS [3, 4]. These findings suggest that these autoimmune diseases share part of the pathogenesis of RA. This theory is partly supported by the findings that some single nucleotide polymorphisms are associated with several different autoimmune diseases. In addition, our previous studies regarding familial risks of SLE and SS also identified co-aggregation of these diseases in families with patients with SLE or SS [10, 12]. Therefore, a positive family history of RA is a flag for an increased risk for autoimmune diseases.

There are several limitations to the present study. Firstly, the case definitions of RA and other autoimmune diseases were based on physician-recorded diagnosis. Secondly, patients with less severe disease may not be included because they are not eligible for a certificate, and this could have introduced bias. Thirdly, zygosity of twins is not recorded in the database, so we cannot estimate heritability using a classic twin study design. Fourthly, the results are subject to the assumption of the polygenic liability model. Fifthly, spouses may not be a perfect control because environmental factors shared by spouses may not be exactly the same as those shared by siblings. Sixthly, the complete ascertainment of family relationships is not possible because some people do not need the family relationship to maintain NHI eligibility. However, familial relationships can be ascertained in the majority of permanent residents in Taiwan. In addition, ascertainment of RA cases and family relationships are unrelated processes; therefore, there is no reason to believe that there is differential ascertainment of RA cases which may bias our estimates. Finally, our study is restricted to Taiwan; therefore, further studies in other populations are required to determine the generalizability of our findings.

In conclusion, this population-wide study confirms that a family history of RA is one of the strongest risk factors for RA. Differential risk associated with different kinships suggests a strong genetic component in the susceptibility of RA. A family history of RA also exerts an increased risk of other autoimmune diseases.

Acknowledgements

The authors thank the Ministry of Science and Technology of Taiwan (project 104-2314-B-182A-047, 103-2314-B-182A-043 -MY2) and Chang Gung Memorial Hospital (project CMRPG3F0841, CMRPG3F0831) for their financial support of this research. This study is based in part on National Health Insurance Research Database data provided by the National Health Insurance Administration, Ministry of Health and Welfare, and managed by the National Health Research Institute. The interpretation and conclusions contained herein do not represent positions of the National Health Insurance Administration or the National Health Research Institute. Author contributions: study concept and design: C.-F.K., W.Z., M.D.; acquisition of data: C.-F.K., L.-C.S., K.-H.Y., S.-F.L.; analysis and interpretation of data: C.-F.K., M.J.G., L.-C.S., A.M.V., W.Z., M.D.; drafting of the manuscript: C.-F.K., W.Z.; critical revision of the manuscript for important intellectual content: C.-F.K., M.J.G., L.-C.S., K.-H.Y., S.-F.L., A.M.V., S.W.S.S., W.Z., M.D.; statistical analysis: C.-F.K., M.J.G., L.-C.S., W.Z.; obtaining funding: C.-F.K., L.-C.S., K.-H.Y., S.-F.L., S.W.S.S.; administrative, technical or material support: M.J.G., L.-C.S., K.-H.Y., S.-F.L., A.M.V., W.Z., M.D.; study supervision: W.Z., M.D., M.J.G.

Funding: The work was sponsored by the Ministry of Science and Technology of Taiwan (project 104-2314-B-182A-047) and Chang Gung Memorial Hospital (project CMRPG3B1672). The sponsors of the study, the Chang Gung Memorial Hospital, the Ministry of Science and Technology and the University of Nottingham had no role in design and conduct of the study; collection, management, analysis and interpretation of the data; and preparation, review or approval of the manuscript and decision to submit the manuscript for publication.

Disclosure statement: W.Z. reports grants from Arthritis Research UK, HTA, NIHR and honoraria for consultations from AstraZenica, Daiichi Sankyo, Biobarica, Hisun and Grunenthal outside the submitted work. All other authors have declared no conflicts of interest.

Supplementary data

Supplementary data are available at Rheumatology Online.

References