Article Text

Siblings of patients with rheumatoid arthritis are at increased risk of acute coronary syndrome
  1. Helga Westerlind1,
  2. Marie Holmqvist1,
  3. Lotta Ljung1,2,
  4. Thomas Frisell1,
  5. Johan Askling1,3
  1. 1 Clinical Epidemiology Division, Department of Medicine Solna, Karolinska Institutet, Stockholm, Sweden
  2. 2 Department of Public Health and Clinical Medicine/Rheumatology, Umeå University, Umeå, Sweden
  3. 3 Department of Rheumatology, Theme Infection and Inflammation, Karolinska University Hospital, Stockholm, Sweden
  1. Correspondence to Dr Helga Westerlind, Clinical Epidemiology Division, Department of Medicine Solna, Karolinska Institutet, Stockholm 17176, Sweden; Helga.Westerlind{at}ki.se

Abstract

Objectives To investigate a potential shared susceptibility between rheumatoid arthritis (RA) and acute coronary syndrome (ACS) by estimation of the risk of ACS among full siblings of patients with RA.

Methods By linking nation-wide Swedish registers, we identified a cohort of patients with new-onset RA 1996–2016, age- and sex-matched (5:1) general population comparator subjects, full siblings of RA and comparator subjects, and incident ACS events through 31 December 2016. We used Cox regression to estimate the HR of ACS among patients with RA and the siblings of patients with RA versus the general population, overall and stratified by RA serostatus. We explored the impact of traditional cardiovascular (CV) risk factors on the observed associations.

Results We identified 8109 patients with incident RA, and 11 562 full siblings of these. Compared with the general population, the HR of ACS in RA was 1.46 (95% CI 1.28 to 1.67) and 1.22 (95% CI 1.09 to 1.38) among their siblings. The increased risks seemed confined to seropositive RA (patients: 1.52 [1.30 to 1.79], their siblings: 1.27 [1.10 to 1.46]); no significant risk increase was observed among siblings of patients with seronegative RA (HR 1.13 [95% CI 0.92 to 1.39]). Adjustment for 19 traditional CV risk factors did not appreciably alter these associations.

Conclusion Siblings of patients with RA are at increased risk of ACS, suggesting shared susceptibility between RA and ACS, indicating the need and potential for additional cardio-preventive measures in RA (and their siblings).

  • rheumatoid arthritis
  • cardiovascular risk
  • epidemiology
  • familial risk

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Key messages

What is already known about this subject?

  • Patients with rheumatoid arthritis (RA) are at increased risk of cardiovascular (CV) disease.

  • Despite better inflammatory control with more efficient treatments and treatment strategies, an excess CV risk among patients with RA remains.

What does this study add?

  • This study demonstrates that an excess CV risk is present also among close relatives of patients with RA, thereby suggesting shared susceptibility of RA and acute coronary syndrome (ACS).

How might this impact on clinical practice or future developments?

  • Optimised RA disease control may not be enough to remove the excess CV risk in RA.

  • Patients with RA (and their first-degree relatives) may benefit from additional cardio-preventive measures.

Introduction

Patients with rheumatoid arthritis (RA) are at increased risk of cardiovascular disease (CVD) including acute coronary syndrome (ACS).1 A number of studies have demonstrated that this excess risk cannot be readily explained by traditional CV risk factors, but instead point to an association between RA disease severity and development of ACS.2 We recently reported that despite more efficient control of inflammation in RA during recent years, and despite a general decline in ACS incidence in the general population, an excess risk for ACS among patients with RA remains.1 These findings suggest that besides direct effects on the ACS risk exerted by the RA disease itself, there may be a shared susceptibility between RA and ACS.

If the excess risk of ACS in patients with RA was increased due to such shared susceptibility, one might expect an increased risk of ACS also in individuals without RA but with otherwise similar genetic set-up and background as the patients, such as their siblings. The objective of this study was therefore to investigate any potential shared susceptibility between RA and ACS, by estimating the risk of ACS in full siblings of individuals with versus without RA.

Methods

Study design and setting

We performed a population-based nation-wide cohort study based on linkage of clinical and other registers with prospectively recorded information on RA, family structure and ACS. The Swedish Rheumatology Quality (SRQ) register is a clinical quality register with a current estimated coverage of around 85% of all prevalent RA followed in Swedish Rheumatology.3 By using the unique Swedish personal identification number assigned to all permanent residents in Sweden, we linked SRQ to the following nation-wide and virtually complete registers: the Swedish Multi-generation Register (MGR), the National Patient Register (NPR), the Prescribed Drug Register (PDR), the Cause of Death Register and the Total Population Register. The identification of our RA cohort, and subsequent register linkages, has been described in detail elsewhere.1

Patients with RA

In SRQ, we identified a cohort of incident RA born 1932 or later, and diagnosed with RA 1996 through 2016 within 1 year of reported RA symptom onset. RA serostatus was defined by the reporting clinician based on rheumatoid factor status.

As sensitivity analyses, and using previously devised algorithms to define incident and prevalent RA,4 we used NPR to identify two additional RA cohorts: one with incident RA cases diagnosed between 2006 and 2016 (including also cases not followed from RA onset in SRQ), and one of all prevalent RA, irrespective of duration, with visits listing RA during 2006–2016.

Matched general population comparator subjects, and relatives of patients and comparator subjects

For each patient with RA, all full siblings born within 5 years were identified through MGR. In addition, for each patient with RA, up to five subjects from the general population, matched on age and sex, and with at least one full sibling born within 5 years, were randomly selected. All unique individuals (patients with RA, their siblings, matched general population subjects and their siblings) were required to be alive and resident in Sweden at the time of the index patient’s RA diagnosis. Patients with RA and general population subjects without siblings were secluded from the study population.

Follow-up and outcome

All unique individuals were followed from the date when their index patient with RA first fulfilled the inclusion criteria in the RA cohort(s). The outcome was defined as a first-ever ACS (hospitalisation for ACS [International Classification of Diseases, Tenth Revision, ICD10: I21 or I20.0] or acute myocardial infarction listed as underlying cause of death). We censored follow-up at death, migration, first-ever RA diagnosis (for non-RA subjects) and the end of the study period (31 December 2016). All individuals with an ACS before start of follow-up were excluded.

Statistical analyses

We calculated the incidence of ACS in each cohort, and HRs of ACS comparing the RA-, sibling- and population comparator-cohorts using Cox proportional hazards model, adjusted for age, sex and calendar period of start of follow-up. All individuals with a history of ACS prior to start of follow-up were excluded from the analysis. CIs were estimated using a robust sandwich estimator to account for the correlated data structure. We contrasted the incidence of ACS among the patients with RA and their siblings to the matched general population subjects, and the incidence among RA siblings to that of the siblings of the matched general population subjects. We further estimated the risk increase in patients with RA compared with their siblings by performing within-pair analyses. We stratified by RA serostatus. As sensitivity analyses, we excluded all individuals with a history of a CV disease (CVD) (defined as ICD10: I10–I15, I20–I25, I26–I28, I30–I52, I60–I69, I70–I79 or I82) before start of follow-up, adjusted for family history of CVD before start of follow-up, and performed two subgroup analysis: one excluding all individuals with a sibling with CVD diagnosis and one of all individuals with a first-degree relative diagnosed with ischaemic heart disease (ICD10: I20–I25) at start of follow-up.

Exploratory analyses

To explore whether known CV risk factors or determinants could explain the observed increase in ACS incidence among the RA siblings, we used NPR and PDR to identify 19 medical and contextual covariates corresponding to typical general population CV risk factors or determinants such as medical histories and socioeconomic characteristics (online supplementary table 2). Since PDR was started in 2005, these analyses were restricted to individuals whose start of follow-up was 2006 or later. We first ran a Cox regression, adjusted for age, sex and calendar period, to estimate the association between each factor and incidence of ACS in the combined cohort of the general population comparators and their siblings. We then, for each of the covariates, adjusted the association between RA siblings and ACS risk for that particular risk factor. Finally, we added all covariates to the model.

By calculating the E-value, we estimated the effect size a potential unmeasured confounder would need in order to completely ‘explain’ the observed association between RA siblings and ACS risk.5 6 The E-value is designed to be independent of assumptions of underlying effect sizes and prevalence of the unmeasured confounder, but assumes a symmetric effect size in the association with exposure and outcome. Therefore, we, using smoking prevalences and risk estimates for smoking and ACS from previous studies,7 8 also estimated what the relative risk of ACS among the RA siblings would be under the extreme9 assumption that their smoking habits were the same as their index patients’ with RA, and conversely, how high the prevalence of smoking in the RA sibling cohort would have to be to fully explain the observed risk increase in that cohort. Details of these analysis are in online supplementary material 1.

Ethical permission was obtained from the Stockholm ethics review board (2015/1844-31/2).

Results

For the main analysis, we identified 8109 patients with RA (66% seropositive), 11 562 of their full siblings, 38 092 general population comparator subjects and 50 793 full siblings of the latter (table 1). The proportion of RA siblings excluded due to a history of ACS (2.8%) was higher than the proportion among the siblings of the general population (2.4%), p=0.037, table 1.

Table 1

Demographics and HRs for ACS among new-onset patient with RA identified 1996–2016 in the SRQ register, their full siblings, matched general population subjects and among full siblings of the latter, and stratified by index patient with RA’s serostatus. NB. The sex distribution between the patients with RA and their matched reference individuals differs compared with the siblings of the patients with RA and the siblings of the reference individuals. Therefore, the proportions and incidence of ACS are not directly comparable between these two pairs of cohorts.

During a median follow-up of 6 years, the crude incidence of ACS was 4.64 (per 1000 person-years) among RA cases, 4.74 among their siblings, 3.18 among the matched general population subjects and 3.84 among the comparator subjects’ siblings. Compared with the general population, the age-, gender- and calendar-period-adjusted HR for ACS among the patients with RA was 1.46 (95% CI 1.28 to 1.67) and 1.22 (95% CI 1.10 to 1.38) among their siblings. Comparing the siblings of the patients with RA with the siblings of the general comparator subjects resulted in a similar risk estimate (HR=1.18 [95% CI 1.06 to 1.32]). The increased risk of ACS was largely confined to patients with seropositive RA (HR=1.52 [95% CI 1.30 to 1.79]) and to their siblings (HR=1.27 [95% CI 1.10 to 1.46]). The elevated risk for patients with seronegative RA was less pronounced (HR=1.34 [95% CI 1.06 to 1.70]), and nonsignificant for their siblings (HR=1.13 [95% CI 0.92 to 1.39]), figure 1. The difference in relative risks for the siblings of the seropositive and seronegative patients was, however, not statistically significant.

Figure 1

HRs and 95% CIs for acute coronary syndrome in Swedish patients with RA and their full siblings compared with matched reference individuals from the general population. The RA cohort was identified using the Swedish Rheumatology Quality register. HRs are presented overall and stratified by index patients with RA’s serostatus.RA, rheumatoid arthritis.

The within-pair analysis confirmed the increased risk of ACS in RA (HR=1.19 [95% CI 1.04 to 1.37]) but also that this association, because of the increased risk also among RA siblings, was of lower magnitude than when patients with RA were compared with the general population. The within-pair relative risk of ACS was more evident in seropositive (HR=1.21 [95% CI 1.02 to 1.44]) rather than seronegative (HR=1.14 [95% CI 0.89 to 1.46]) RA.

Online supplementary table S1 presents the corresponding results from the sensitivity analyses. In the incident cohort, risks and HRs were similar or slightly lower than the main analysis. In the prevalent cohort, in which the mean RA duration was longer than in the incident cohorts, the HR of ACS among the patients with RA compared with their general population comparator subjects was higher (HR=1.79 [95% CI 1.67 to 1.91]), whereas the elevated HR among RA siblings (vs the siblings of the general population comparator subjects) remained significant but less pronounced (HR=1.09 [95% CI 1.02 to 1.16]); the within-pair analysis in this cohort resulted in a HR=1.63 (95% CI 1.51 to 1.77).

Removing individuals with a history of CVD before start of follow-up and adjusting for family history of CVD did not noticeably change the HRs. Subset analyses restricted to individuals without any sibling history of CVD, and without any first-degree relative with IHD, respectively, the HRs decreased by less than 10% but remained significantly elevated (data not shown).

Our exploratory analyses demonstrated that 14 of the 19 CV risk factors/determinants were indeed ACS risk factors. When added to the model either one by one or all together, they did not appreciably explain the increased incidence of ACS among the RA siblings, neither overall nor when stratified by index RA serostatus (online supplementary table S2–S4).

The E-value was 1.74 with a lower bound of 1.40. In the seropositive group, the corresponding E-value was 1.86 with a lower bound of 1.43.

Finally, under the assumptions that the prevalence of ever/current smoking among the RA siblings would be the same as reported for patients with RA (67% current, 31% ever smokers, data from a Swedish study on incident RA 1996–20067) rather than the general population (54% ever, 22% current smokers7), and that the ACS risk is increased by a factor 1.40 among past and 2.90 among current smokers,8 we estimated that the risk for ACS in the RA sibling cohort versus the general population, in the absence of any true association between RA siblings and ACS risk, would be 1.11 (instead of 1). The same estimations indicated that for the observed overall HR of 1.22 among the RA siblings to be fully explained by an increased prevalence of smokers among the RA siblings, they would have to smoke substantially more than the patients with RA themselves.

Discussion

In this study, we demonstrate that full siblings of patients with RA are at increased risk of ACS compared with the general population, and that this increase cannot readily be explained by confounding by traditional ACS risk factors or by socioeconomic factors, pointing to the existence of other shared risk factors or susceptibility between RA and ACS. Through this, we demonstrate that a substantial proportion of the increased risk of ACS in patients with RA is likely due to other factors (shared with their siblings) than the RA disease itself. Beyond our main findings, our study further demonstrates that the increased risk of ACS in patients with RA prevail also in RA diagnosed 2006 or later and followed through 2016, confirming that the level of risk increase remains elevated, is higher in patients with established versus incident RA, and largely confined to seropositive RA.

Whereas our results point to the existence of shared susceptibility, its nature (genetic or environmental) remains to be established. Initial reports suggested an association between the CIITA gene, RA and myocardial infarction.10 Subsequent reports have, however, failed to replicate this association.11–15 One report has also linked the shared epitope alleles to myocardial infarction risk.16 Our results suggest that adjustment for medical or contextual CV risk factors did not remove the observed increase in sibling risk. Smoking is a risk factor for both RA and ACS, and to some extent familial.17 In studies of twin-pairs discordant for RA, the twin affected with RA is more likely to be a smoker than its co-twin.9 By contrast, our exploratory analyses indicated that to fully explain the increased ACS risk in RA siblings, we would need to make extreme assumptions about smoking among the siblings such that they were more often smokers than their index-patient with RA, an assumption that is neither plausible nor supported by previous literature.9 Thus, and since smoking is a risk factor for many of the ACS risk factors adjusted for in our exploratory analyses (online supplementary table S4), we conclude that unmeasured confounding from smoking may to some part, but is unlikely to fully, explain the observed risk increase among the RA siblings.

Our study has a few limitations. Despite large numbers of patients with RA we still had limited precision and thus cannot formally exclude that the adjustments for our 19 risk factors actually did remove some of the association under study. Some of the ACS risk factors (hypertension, chronic obstructive pulmonary disease, mild renal insufficiency) used in this analysis are treated in primary and not specialist care and thus not covered by NPR. For this reason, we used information on dispensed drugs (from PDR, with full coverage) as an additional source to define these risk factors. Another limitation is the lack of individual-level smoking data.

We are not aware of any previous large-scale studies of ACS risks among close relatives of patients with RA. Indeed, the main strength of our study is the population-based, nation-wide setting, the use of registers with high coverage of patients with RA, their first-degree relatives and the occurrence of ACS all of which could be determined independently of each other rather than, say, by self-report. For the same reasons, we believe the results from our study may be generalisable to similar subjects also outside of Sweden.

In conclusion, from an etiological point of view, our findings indicate a shared susceptibility to RA and ACS that is not readily explained by traditional CV risk factors but point to a need to further explore the nature of this association, be it genetic or environmental. From a clinical point of view, our findings serve as a reminder that reducing or removing RA-specific inflammation may in itself not be sufficient to remove the entire excess risk of ACS in RA. Instead, additional cardio-preventive measures, such as optimisation of traditional CV risk factors, may be (particularly) important in these patients and among their siblings.

References

Supplementary materials

  • Lay summary

    Disclaimer : This is a summary of a scientific article written by a medical professional (“the Original Article”). The Summary is written to assist non medically trained readers to understand general points of the Original Article. It is supplied “as is” without any warranty. You should note that the Original Article (and Summary) may not be fully relevant nor accurate as medical science is constantly changing and errors can occur. It is therefore very important that readers not rely on the content in the Summary and consult their medical professionals for all aspects of their health care and only rely on the Summary if directed to do so by their medical professional. Please view our full Website Terms and Conditions.
    Copyright © 2019 BMJ Publishing Group Ltd & European League Against Rheumatism. Medical professionals may print copies for their and their patients and students non commercial use. Other individuals may print a single copy for their personal, non commercial use. For other uses please contact our Rights and Licensing Team.

Footnotes

  • Handling editor Josef S Smolen

  • Contributors HW had full access to all of the data used for the analysis in this study and takes full responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: JA, HW and TF. Acquisition of data: JA, TF and HW. Statistical analysis: HW, TF and JA. Analysis and interpretation of data: all authors. Drafting of manuscript: HW and JA. Critical revision of manuscript and final approval given: all authors. Obtained funding: JA. Study supervision: JA.

  • Funding This work was supported by the Swedish Research Council, the Swedish Foundation for Strategic Research, Stockholm County Council (ALF), the Heart Lung Foundation, Karolinska Institutet (Strategic Research Area Epidemiology), The Nordic Research Council (Nordforsk) and the Rheumatology Research Foundation (FOREUM). Funders had no impact on the design or interpretation of the study or its results.

  • Competing interests JA has or has had research agreements with Abbvie, Astra-Zeneca, BMS, Eli Lilly, MSD, Pfizer, Roche, Samsung Bioepis, and UCB, mainly in the context of safety monitoring of biologics via ARTIS/Swedish Biologics Register. Karolinska Institutet has received remuneration for JA participating in advisory boards arranged by Pfizer and Lilly.

  • Patient consent for publication Not required.

  • Provenance and peer review Not commissioned; externally peer reviewed.