Immune-mediated diseases and risk of incident cardiovascular diseases: a prospective cohort study

Abstract

Objectives

Disorders of immune system may impact cardiovascular health; however, comprehensive study is lacking. We aimed to analyse the association of total and 20 individual immune-mediated diseases (IMDs) with risk of incident cardiovascular disease (CVD).

Methods

In this prospective cohort study, 414 495 participants (55.6% women; mean age 55.9 years) from UK Biobank with baseline assessment at 2006–10 were included. Among them, 21 784 participants had prevalent IMDs. Information on IMDs at baseline and incidence of CVDs during follow-up were recorded. Cox proportional hazard models were used to estimate the association between IMDs and CVDs risk.

Results

During the median follow-up of 12.1 years, there were 6506 cases of CVDs in participants with IMDs (29.9%) and 77 699 cases in those without IMDs (19.8%). After multivariable adjustment, participants with IMDs were significantly associated with an increased risk of total CVD [hazard ratio (HR) 1.57; 95% CI 1.52–1.61]. Among the 20 IMDs, 16 showed significant associations with CVD (all P < 0.0025 after Bonferroni correction), with HR ranging from 1.34 (1.16–1.54) for celiac disease to 2.75 (2.10–3.61) for SLE. Participants with any IMD exposure had a higher risk of all individual CVD events, with HR ranging from 1.34 (1.14–1.58) for cerebral hemorrhage to 1.80 (1.54–2.11) for pericardium diseases. IMD duration <5, 5–10 and >10 years was associated with 55%, 59% and 56% increased risk of total CVD, respectively.

Conclusion

Total and individual IMDs were associated with an increased risk of overall CVDs. It is important to consider primary prevention of CVD in patients with IMD and dysregulation of immune system in the cardiovascular health.

Introduction

Cardiovascular disease (CVD) is the leading cause of global mortality and has a profound negative impact on quality of life [1]. Inflammation is a fundamental biological process involved in both physiological and pathological conditions, and CVD is no exception. Chronic low-grade inflammation accompanies all stages of atherosclerotic disease from its onset to the overt disease with manifest ischaemic syndrome, and mechanisms of immunoregulation in CVD development include aging, impaired regulatory T cell (Treg) function, increased gut permeability, changes to microbiota composition, NLRP3 inflammasome activation, oxidative stress, etc. [2–4].

Immune-mediated disease (IMD) is a cluster of common chronic conditions that share inflammatory pathways. The estimated prevalence varies from 5% to 10% [5, 6]. Many above-mentioned mechanisms of immunoregulation were also implicated in IMD, such as altered gut microbiota, excess production of reactive oxygen species, and a defect in either the number or the function of Treg cells [7–9]. Some studies have explored the association between several single IMDs and CVD. RA, SLE, AS and ulcerative colitis were significantly associated with increased risk of some types of heart health problems such as myocardial infarction [10], atrial fibrillation [11], hypertension [12] and stroke [13]. Duration and severity of IMD are two parameters that may reflect inflammation scale in IMD, however a limited number of studies has considered them when analyzing CVD risk with IMD. As a biomarker of inflammation severity, CRP was evaluated and it was shown that the highest tertiles of CRP associated with greater risk of multiple CVD outcomes [14]. However, a history of having IMD could also imply a duration of low-grade inflammation and even reflect a long-term body status that was not fully clarified in previous studies. Analyses including both severity and duration of IMD could better reflect immune dysregulation burden.

So far, the CVD risk profile for total and individual IMD needs to be further depicted. The associations of some less common IMDs and risk of incident CVD were not clarified in previous studies or need further evidence, due to small sample size, short follow-up time, rare consideration of IMD duration and/or important covariates being unadjusted for. To fill this crucial research gap, in this prospective multicentre cohort from the UK Biobank (UKB), the objectives of the current study were threefold. First, to analyse the associations between total and 20 individual IMDs and risk of incident CVD. Second, to investigate whether longer duration of IMD and heavier inflammation load would further increase the risk of CVD.

Methods

Study design and sample

The UKB is a population-based prospective cohort study, and a detailed information has been described in a previous study [15]. Briefly, UKB included >500 000 community-dwelling adults, aged 40–70 years, across the UK between 2006 and 2010 (https://www.ukbiobank.ac.uk/). We declare that all data are publicly available in the UKB repository [15]. UKB received ethical approval from the UK National Health Service, National Research Ethics Service North West, the National Information Governance Board for Health and Social Care in England and Wales, and the Community Health Index Advisory Group in Scotland. All participants provided written informed consent. This study was approved by the UKB (application number 77740).

Among the 502 505 participants with available data in the current study, we excluded those with cardiovascular conditions [International Classification of Diseases 10 revision (ICD-10) codes I00–I09, I20–I89, i.e. all diseases from the circulatory system] at baseline (n = 87 966) and who withdrew from follow-up (n = 44). Thus, a total of 414 495 participants were included in the main analysis (Supplementary Fig. S1, available at Rheumatology online).

Outcome: cardiovascular disease

The outcome of interest was the incidence of total and individual cardiovascular disease (CVD). Incident CVD cases were identified from the ‘first occurrence of health outcomes defined by a 3-character ICD 10^th Revision code’ (category ID in UKB 1712). The diagnosis of individual CVD was obtained by using linkage with death register, primary care and hospital inpatient records. Detailed information regarding the linkage procedure is available online (https://biobank.ctsu.ox.ac.uk/crystal/exinfo.cgi?src=diag_xtabs_HES). Total CVD in this study was categorized into 15 types of individual CVD, with rheumatic heart diseases (ICD-10 codes I00–I09) regarded as an IMD exposure and disorders of lymphatic vessels (ICD-10 codes I88–I89) excluded (Supplementary Table S1, available at Rheumatology online).

Exposure: immune-mediated disease

As suggested in a previous study [6], 21 IMDs with at least 100 affected participants were identified, defined by a three-character ICD-10 code (Table S2, available at Rheumatology online). Exposure data for every participant were extracted from the hospital inpatient admissions (Field ID in UKB 41270), which contained data on admissions, discharge, detailed diagnosis and time since diagnosis which have been reliably used in prior studies [6]. Individuals with prevalent IMDs at baseline were considered as the exposure group (Fig. 1), and as a chronic disease, the duration of IMD was calculated by the time from the first diagnosis of any IMD to the baseline assessment.

Covariates

Three main groups of potential confounders in the association between IMD and CVD were considered in the analyses: (i) demographic variables: age, sex, ethnicity (white/others), education (university or college degree/others), the Townsend index reflecting socioeconomic status (continuous); (ii) lifestyle factors: smoking status (current, ever, never), drinking (glasses/week), physical activity at goal or not (≥150 min/week of moderate intensity or ≥75 min/week of vigorous intensity, or an equivalent combination), healthy diet score (score ≥4); and (iii) cardiometabolic factors: BMI (continuous), total cholesterol (continuous), systolic blood pressure (continuous), type 2 diabetes (yes/no), use of NSAIDs (yes/no). A diet score was calculated based on following factors: vegetable intake ≥4 tablespoons each day (median); fresh fruit intake ≥2 pieces each day (median); oily and nonoily fish intake at least twice each week (median); unprocessed red meat intake no more than twice each week (median); and processed meat intake no more than twice each week (median). Each favorable diet factor received one point, with a total score ranging from 0 to 5 [16]. All these covariates were collected at the baseline of the study.

Statistical analyses

Data analyses were performed using IBM SPSS Statistics, Version 25 (IBM Corporation, Armonk, NY, USA) and R software, version 4.0.2 (R Project for Statistical Computing, Vienna, Austria). Missing data were treated by listwise deletion and mean imputation. The cut-off for covariate retention was 5% of missing data, enabling enough stability and accuracy for the analysis [17]. A P-value <0.05 indicated a statistical significance (two-sided). To account for multiple comparison for the primary analysis on the associations of the 20 individual IMDs with total CVD risk, Bonferroni correction was used and P < 0.0025 (0.05/20) was considered as statistically significant. Because the secondary analyses about individual CVD were considered exploratory and the results were interpreted with caution, Bonferroni correction was not used.

Cumulative cases of CVD were calculated during follow-up visits. The follow-up time was determined from the baseline date (date of attending assessment centre) to the diagnosis of CVD, death or censoring date (30 July 2021), whichever came first. Cox regression model was used to estimate the hazard ratio (HR) and 95% CI. Model 1 was adjusted for demographic variables including age, sex, ethnicity, education and the Townsend index. Model 2 was adjusted for terms in Model 1, lifestyle factors (smoking status, drinking, physical activity at goal, a healthy diet score) and cardiometabolic factors (BMI, total cholesterol, systolic blood pressure, type 2 diabetes and use of NSAIDs).

Then, in participants with IMD, we examined the association between different inflammation loads using the tertiles of CRP which was measured at baseline of the study and risk of incident CVD. For IMD duration, based on the difference between time when the first IMD was diagnosed and time of baseline assessment, participants were divided into <5, 5–10 and >10 years, and participants without IMD was the reference.

We further conducted several sensitivity analyses. First, we used competing risk analysis to explore the incidence rate of CVD when death was considered a competing event. Second, we additionally adjusted the use of the anti-inflammatory or immunomodulatory medications, blood pressure-lowering medications, and cholesterol-lowering medications. Third, more metabolic factors (diastolic blood pressure, low-density lipoprotein cholesterol and triglyceride) were adjusted based on Model 3.

Results

Table 1 and Supplementary Table S3, available at Rheumatology online present the baseline characteristics of participants by incident CVD and prevalent IMDs, respectively. Among a total of 414 495 participants, 55.6% were women and the mean age was 55.91 years, and 21 784 (5.3%) individuals with baseline IMDs. Overall, individuals with incident CVD were relatively older, more likely to be men, and had lower education level and worse economic status compared with those without CVD. They were also less likely to have healthy lifestyle, and had higher BMI, systolic blood pressure and CRP value, and higher prevalence of type 2 diabetes. They were more prone to use NSAIDs in their daily life.

During the median follow-up of 12.1 years (about 4.6 million person-years), a total of 84 205 incident CVD cases were documented. There were 6506 cases of CVD in participants with any IMD (29.9%) and 77 699 cases in those without IMD (19.8%). After adjusting for age, sex, ethnicity, education, Townsend index, smoking status, drinking, physical activity, healthy diet, BMI, total cholesterol, systolic blood pressure, type 2 diabetes and use of NSAIDs (Model 2, Fig. 2), IMDs at baseline were significantly associated with increased risk of total incident CVD (HR 1.57; 95% CI 1.52–1.61). Among the 20 IMDs, 16 showed significant associations with total CVD (all P < 0.0025), with HR ranging from 1.34 (1.16–1.54) for celiac disease to 2.75 (2.10–3.61) for SLE. There were 12 IMDs with HR >1.5 (ranking order with decreasing HR: SLA, sicca syndrome, necrotizing vasculopathies, ITP, sarcoidosis, rheumatism, multiple sclerosis, PMR, RA, Crohn’s disease, asthma, and psoriatic or enteropathic arthropathies). Further adjustments for more metabolic profiles did not attenuate these results (Supplementary Table S4, available at Rheumatology online). Moreover, in the sensitivity analyses, the use of competing risk analysis validated the findings from Cox regressions (Supplementary Table S5, available at Rheumatology online).

Fig. 3 shows the association between any IMD and risk of each CVD outcome. After multivariable adjustments, we found that participants with any IMDs were at higher risk of developing all 15 kinds of individual CVD, with HR ranging from 1.34 (1.14–1.58) (cerebral hemorrhage) to 1.80 (1.54–2.11) (pericardium diseases). Additional adjustment with anti-inflammatory or immunomodulatory medications, blood pressure–lowering medications, and cholesterol-lowering medications did not substantially alter the results (Supplementary Tables S6 and S7, available at Rheumatology online).

Fig. 4 shows the association between individual IMD and individual CVD outcomes (detailed data are presented in Supplementary Table S8, available at Rheumatology online). Three systemic IMDs (RA, sarcoidosis and sicca syndrome) and three non-systemic IMDs (asthma, psoriasis and multiple sclerosis) were associated with an increased risk of multiple CVD events, especially peripheral artery disease, non-rheumatic valve disorders and heart failure, etc. Among all CVD outcomes, peripheral artery disease demonstrated a prominent association with systemic IMDs, HR ranging from 2.16 (1.36–3.44) for AS to 8.75 (6.39–11.99) for necrotizing vasculopathies. Meanwhile disorder of veins was closely related to non-systemic IMDs, with HR ranging from 1.45 (1.19–1.78) for psoriasis to 2.90 (2.03–4.15) for ITP.

Table 2 and Supplementary Table S9, available at Rheumatology online show the association between duration of IMD and CVD. Compared with those without IMD, participant with duration <5, 5–10 and >10 years of any IMD was associated with 55% (1.55, 1.48–1.62), 59% (1.52–1.66) and 56% (1.48–1.65) increased risk of total CVD, respectively. From <5, 5–10 and >10 years, we found an increasing trend of HR in two non-systemic IMDs (asthma and multiple sclerosis) and a decreasing trend of HR in two systemic IMDs (RA and ulcerative colitis). The decreasing trend phenomenon was observed in the association between duration of IMD and risk of nine individual CVDs.

Supplementary Table S10, available at Rheumatology online, presents the results of individual CVD outcomes among IMD population when using CRP to quantify the severity of inflammation. Among participants with IMD, an increasing trend in the risk of total CVD with greater CRP values was observed. Compared with those in the lowest tertile of CRP, the HR was 1.07 (0.99–1.16) in the middle tertile and 1.15 (1.06–1.24) in the highest tertile. The increasing trends were observed in myocardial infarction, chronic ischaemic heart disease, non-rheumatic valve disorders, atrial fibrillation and flutter, heart failure and peripheral artery disease.

Discussion

In this large-scale cohort with about 12 years’ follow-up time, we found that participants with IMDs had significantly increased risk of total incident CVD, after adjusting for conventional cardiovascular risk factors. Eighteen IMDs showed significant associations, with SLE, necrotizing vasculopathies and sicca syndrome ranking as the top three. Moreover, some systemic and organ-specific IMDs had extensive relationships with increased risk of many individual CVDs. Shorter duration in most IMD and higher CRP level indicated a greater risk of total CVD outcomes. Additionally, arterial stiffness, a physiological marker of premature atherosclerosis, may mediate specific inflammatory pathways from IMD to CVD, although the effects seemed to be small [18]. This study provides novel insights into the systemic effects of IMD in the heart and vascular health.

The association between IMD and CVD is biologically plausible [19]. When chronic low-grade inflammation occurs, increased level of CRP contributes to endothelial dysfunction and hypertension by inhibiting nitric oxide [20], increasing endothelin-1 production, and thus impairing endothelial-dependent vascular relaxation [21]. In addition, CRP contributes to plaque instability by activating the nuclear factor-κB pathway and increasing endothelial cell adhesion molecules expression. These mechanisms work together in the pathology of CVDs. On the other hand, the use of DMARDs is thought to decrease CVD risk, potentially owing to their suppression of systemic inflammation [22, 23]. A systemic review suggested a decreased risk of CVD in patients with RA being treated with TNF inhibitors [24]. Moreover, there is promising evidence that canakinumab, a mAb directed against IL-1β, was associated with a significant reduction in the major CVD outcome by 15% percent [25].

Up to now, limited studies have comprehensively measured the association between various IMDs and CVD outcomes. IMDs can be either organ-specific or systemic, which proves it might be a heterogeneous collection of pathological conditions triggered by genetic, environmental, immunoregulatory and even hormonal factors [26]. Our results present that regardless of any or individual IMDs, most of them were associated with the increased risk of total CVDs, except AS, Guillain–Barré syndrome, allergic rhinitis and lichen planus, three of which were non-systemic IMDs. Moreover, among IMDs, SLE, necrotizing vasculopathies, sicca syndrome and ITP were all associated with >2-fold greater risk of CVDs. In line with our findings, cumulative evidence over recent years also supported the active involvement of IMDs in excessive cardiovascular burden [27]. A matched cohort in the UK indicated that systemic autoimmune disorder was associated with 54% and 49% higher risk of stroke and coronary heart disease, respectively [14]. In a meta-analysis with >40 000 patients, those with RA have 48% higher risk of developing CVD, and even 68% increased risk for myocardial infarction [28]. Thus, our study extends previous research by comprehensively delineation of the association of 21 IMDs with CVD incidence, and by showing extensive associations between the two systems of diseases.

When CVD outcomes were considered separately, individuals with any IMDs were strongly associated with all CVD events in the current study, greatly reinforced the view that pathological status triggered by immune disease was closely related to heart and vascular health [29]. Also, the prevalence of individual CVDs at baseline was obviously higher among the population exposed to IMD than those without IMD (Supplementary Table S11, available at Rheumatology online). Of note, the HRs for pericardium diseases, heart failure and peripheral artery disease even exceeded that of chronic ischaemic heart disease. A prior study linked coronary artery disease with great clinical impact since it is one of leading causes of premature mortality in most autoimmune rheumatic diseases [30].

In further analysis, we identified that asthma and RA were associated with the risk of most CVDs, 12 and 11, respectively. Previous studies also indicated that they were independent risk factors for atherosclerosis associated CVD [31], which suggested that systemic inflammatory markers and endothelial dysfunction could play a key role among them [32]. However, lichen planus, autoimmune thyroiditis and Guillain–Barré syndrome were shown not to be statistically associated with any of the CVDs. A report considered that lichen planus was associated with early vascular dysfunction and structural changes [33]. The inconsistent conclusion may be due to their unique pathologies, which need further research. Regarding the CVD outcomes, peripheral artery disease demonstrated the most extensive associations with IMDs and this result highlights that clinic attention should not be limited to coronary artery disease. Impaired endothelial function and increased arterial stiffness could stem from inflammatory responses [34], which provided a series of pathways linking lipids and other traditional risk factors to CVDs such as atherosclerosis [35].

Duration and severity of IMD matter. We made the novel finding that shorter duration in most IMDs was associated with a greater risk of total CVD outcomes. Similarly, data from the National Health and Nutrition Examination Survey (NHANES) showed a 41% increased risk of CVD mortality in individuals with current asthma, but not former asthma [36]. Compared with former or never asthma, a significantly elevated CRP concentration was noted among current asthma patients [37]. However, the definition of the current and former exposure of disease was unclear. Our study showed for most IMDs a shorter IMD duration, a more active state of immune system in recent that could impact on the cardiovascular system. Nevertheless, several individual IMDs including asthma, multiple sclerosis and sicca syndrome trended to have a more prominent relationship with CVD risk as the duration of IMD increased. Prolonged activation of the adaptive immune system causing cellular exhaustion could explain this observation [38]. The heterogeneity between individual IMDs deserves further investigation. The higher levels of CRP, which were biomarkers of inflammation severity, also showed strong associations with increased risk of individual CVD events. The immune burden reflected by the severity of IMD should be taken into consideration when performing CVD risk stratification.

Strengths and limitations

Our study had important strengths. First, with the prospective cohort design, a large sample that included >400 000 participants and relatively long follow-up duration (median 12.1 years), this study had sufficient statistical power. Second, multiple covariates including demographic factors, lifestyles, physical indicators and use of medications were all in the adjustment, enhancing the reliability of our results. Third, there are few studies containing such a wide range of IMDs and cardiovascular events, which enabled us to compare the associations between individual IMDs or even consider the relevant biomarkers with risk of CVDs. Furthermore, IMD duration measured in the lifetime and inflammation severity assessed by CRP presented a convincing immune burden.

Some limitations of this study should be acknowledged. First, due to the observational nature of this study, a causal relationship could not be demonstrated and residual confounding owing to unmeasured factors might still be possible. Second, individuals exposed to IMDs were from hospital inpatient admissions only—self-reported and primary care cases were not included. However, as mentioned previously [39], this bias may underestimate the proportion of exposure to immune disease and attenuate the true associations with CVDs. Third, although we adjusted the use of NSAIDs and immunomodulatory medications in the main model and sensitivity analysis, no detailed data on duration or dose information were collected. However, there is no consistent conclusion regarding the relationship between the use of medications to treat immune diseases and risk of cardiovascular events, and different medications should be discussed separately [40]. Fourth, though CRP level was used to assess inflammation load, disease activity index would be better when evaluating the association between individual IMDs and CVDs, which is not available in such an outcome-wide analysis. Lastly, the UKB cohort was restricted to volunteers of European ancestry, mostly white British, and the representation of the general population could be limited, with people from poor socioeconomic background possibly being underrepresented. Therefore, whether these findings could generalize to other populations needs further research.

Conclusion

In conclusion, this study comprehensively delineates the association of IMD profile with CVD incidence. Some systemic and organ-specific immune disorders had extensive relationships with increased risk of individual CVDs. Moreover, a shorter IMD duration and higher CRP levels were associated with higher risk of most CVDs. Our findings highlight the importance of primary prevention of CVD in patients with IMD and of considering dysregulation of immune system in the heart and vascular health.

Supplementary material

Supplementary material is available at Rheumatology online.

Data availability

The data underlying this article are available in UK Biobank, at https://www.ukbiobank.ac.uk/.

Contribution statement

N.W. and Y.L. concepualized this paper. Yuetian Yu, Y.S., Y.W., Yuefeng Yu and N.W. wrote the manuscript, researched data and reviewed/edited the manuscript. C.C. and B.W. reviewed/edited the manuscript. X.T. contributed to data acquisition and revised the manuscript critically for important intellectual content. All authors approved the final manuscript.

Funding

This study was supported by Shanghai Ninth People’s Hospital (YBKA201909), Shanghai Municipal Human Resources and Social Security Bureau (2020074), Clinical Research Plan of SHDC (SHDC2020CR4006), Innovative research team of high-level local universities in Shanghai (SHSMU-ZDCX20212501). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. This research has been conducted using the UK Biobank Resource under Application Number 77740.

Disclosure statement: No potential conflicts of interest relevant to this article were reported.

References

Author notes