Abstract
Background
Cardiac disease, including heart failure (HF), is a significant cause of morbidity and mortality for patients with systemic autoimmune inflammatory diseases (SAIDs). We sought to describe HF risk among patients with SAIDs and to investigate associations between cardiovascular drug use and incident HF.
Methods
We conducted a retrospective cohort study of electronic health records of 182 795 adult patients with SAIDs including systemic sclerosis, systemic lupus erythematosus, rheumatoid arthritis, psoriatic arthritis, inflammatory bowel disease, and HIV from 2000 to 2020 and track their risks of new‐onset incident HF in a large multidisciplinary health care institution.
Results
Systemic sclerosis (adjusted hazard ratio [aHR], 2.81 [95% CI, 2.57–3.08]; P<0.001), systemic lupus erythematosus (aHR, 1.64 [95% CI 1.56–1.74]; P<0.001), and rheumatoid arthritis (aHR, 1.54 [95% CI 1.47–1.61]; P<0.001), were significantly associated with incident HF compared both with a SAID‐free control group and a large group with inflammatory bowel disease in adjusted Cox regression models. Beta‐blocker use at baseline was associated with decreased incident HF in a combined group of systemic sclerosis, systemic lupus erythematosus, and rheumatoid arthritis (aHR, 0.7 [95% CI 0.6–0.8]; P<0.001).
Conclusions
Patients with systemic sclerosis, systemic lupus erythematosus, and rheumatoid arthritis had increased risk of incident HF independent of traditional risk factors, indicating an underlying autoimmune mechanism of cardiac involvement. Beta‐blocker use was associated with decreased incident HF, indicating a potentially cardioprotective effect in SAIDs.
Keywords: beta blocker, heart failure, systemic autoimmune inflammatory diseases
Subject Categories: Heart Failure
Nonstandard Abbreviations and Acronyms
- IBD
inflammatory bowel disease
- PsA
psoriatic arthritis
- SAID
systemic autoimmune inflammatory disease
- SSc
systemic sclerosis
Clinical Perspective.
What Is New?
We observed that those with systemic sclerosis, systemic lupus erythematosus, or rheumatoid arthritis had significantly increased risk of incident heart failure independent of traditional risk factors.
As an exploratory analysis, we observed that beta blocker use was associated with decreased incident heart failure when adjusted for traditional risk factors.
What Are the Clinical Implications?
We recognize that we can establish only association and not causality, especially regarding treatment effects of cardiovascular drugs on outcomes; nevertheless, these intriguing and hypothesis‐generating findings of possible cardioprotective effects of beta blockers may warrant further investigation in this population.
Systemic autoimmune inflammatory diseases (SAIDs) are defined by autoinflammatory processes affecting multiple organs. Cardiac involvement in SAIDs is a relatively common but underappreciated phenomenon. A Dutch study found that diseases of the cardiovascular system accounted for 55.5% of overall deaths in patients with SAIDs, with 35.7% of these attributable to nonischemic heart disease. 1 Heart failure (HF) specifically is a major cause of morbidity and mortality in SAIDs. 2 , 3 , 4 The predominant form of HF in SAIDs patients is HF with preserved ejection fraction; however, this relationship, including risk factors and early signs of disease, has not been fully elucidated. 2 , 5 , 6 Cardiac involvement in these conditions is likely attributable to a combination of autoinflammatory processes leading to coronary artery disease, inflammatory cardiomyopathy, or microvascular dysfunction. 7 , 8 , 9 Recent studies have indicated that there is an increased risk of HF in systemic sclerosis (SSc), systemic lupus erythematosus (SLE), and rheumatoid arthritis (RA) independent of the risk of ischemic heart disease, indicating that underlying inflammatory pathology likely causes myocardial damage in these patients. 4 , 10 , 11 The role of disease‐modifying antirheumatic drugs in modifying HF risk in patients with SAIDs also remains unclear. 12 , 13 , 14 , 15 , 16 Additionally, little is known regarding the role of common cardiovascular drugs in modulating cardiac risk in patients with SAIDs. Retrospective analysis suggests that vasoactive drugs and acetylsalicylic acid may be beneficial in SSc for ventricular dysfunction, arrhythmias, and conduction abnormalities, but studies are limited. 17 , 18
As cardiovascular disease, including HF, is a leading cause of mortality in patients with SAIDs, it is crucial to gain a better understanding of the risk factors for HF in this population. 19 , 20 , 21 Herein, we investigate the associations between SAID diagnosis and HF risk in a large health system with longitudinal cardiovascular and mortality outcomes. We also aimed to investigate the relationships between cardiovascular drug use and incident HF among patients with SAIDs.
Methods
Data Availability
This study was approved by the Cleveland Clinic Institutional Review Board with waiver of patients’ written consent as medical chart review project. Our institutional review board imposes restrictions on data availability for this project because we do not have patients’ authorization and consent to share their clinical data outside our institution. The authors have full access to all study data, take responsibility for its integrity and the accuracy of the analyses, and may agree to provide aggregate data upon reasonable request and upon approval by the institutional review board.
Study Population
Our population of patients with SAIDs included those with SSc, RA, SLE, psoriatic arthritis (PsA), and inflammatory bowel disease (IBD). Although not classified as SAID, HIV was included as an immune‐related disorder comparator that has been studied in HF and that was included in a similar study by Prasada et al. 4 , 22 We included patients ages 18 years or older who received outpatient care at the Cleveland Clinic Health System between 2000 and 2020, with a diagnosis of SSc, SLE, RA, PsA, IBD, or HIV (Figure 1). Exclusion criteria included missing key demographic values (sex, date of birth, first/last office visits). SAID diagnoses were identified via the presence of 2 or more International Classification of Diseases, Ninth Revision and Tenth Revision (ICD‐9/10) diagnosis codes for SSc, RA, PsA, or IBD within a 2‐year period. SLE diagnosis was established by the presence of 3 separate ICD codes in 3 separate months. 4 HIV diagnosis was determined based on positive HIV serology, value for plasma HIV RNA (viral load), or 3 instances in which both HIV viral load and CD4 T lymphocyte cell count were ordered on the same date. 4 , 23 A total of 11 026 patients had >1 SAID diagnosis, and these patients were assigned the least common diagnosis, assuming more stringent diagnostic criteria for more rare conditions. Any patient with HIV was assigned to the HIV group.
Figure 1. Systemic autoimmune rheumatic disease cohort identification, inclusion, and exclusion criteria.
CC indicates Cleveland Clinic; HF, heart failure; IBD, inflammatory bowel disease; ICD‐9/10, International Classification of Diseases, Ninth Revision and Tenth Revision; MI, myocardial infarction; PsA, psoriatic arthritis; RA, rheumatoid arthritis; SAID, systemic autoimmune inflammatory disease; SLE, systemic lupus erythematosus; and SSc, systemic sclerosis.
Data Synthesis
This was a retrospective cohort study in which demographic, clinical, and medication use data for patients with and without SAIDs were extracted from the electronic health record (EHR). HF status and cardiovascular comorbidities were similarly identified using ICD‐9/10 codes within the study period. Baseline was defined as date of meeting all criteria for SAID diagnosis as described, and incident HF was defined as new HF diagnosis appearing as an ICD‐9/10 diagnosis code and documented as problem list in the EHR after an office visit greater than 6 months after baseline. HF diagnosis was identified in the EHR via ICD‐9 code 428 and ICD‐10 code I50. Patients who did not experience incident HF during the follow‐up period were censored by last clinic visit, as this is the last time point in available data that a given patient was known to be HF free. Cardiovascular comorbidities included hypertension, diabetes, myocardial infarction (MI), and chronic kidney disease (CKD). Similar to the method outlined in Prasada et al., hypertension diagnosis was identified using ICD‐9 codes 401 to 405 and ICD‐10 codes I10 to I15; diabetes diagnosis was identified using ICD‐9250 and ICD‐10 E10, E11, andE13; MI with ICD‐9 code 410 or ICD‐10 I21, I22; and CKD with ICD‐9 code 585 or ICD‐10 N18, N19. 4 In an exploratory analysis to gain insights into the association of cardiovascular drugs with HF risk, we included classes of commonly used cardiovascular drugs, such as angiotensin‐converting enzyme inhibitors, angiotensin receptor blockers, beta blockers, calcium channel blockers, and thiazide diuretics. Medication data collected regarding cardiovascular drugs included, when available, prescription status at baseline, start and stop dates for the first 3 instances of drug prescription, and dose information. Left ventricular ejection fraction at time of HF diagnosis was included when available. 24 A frequency‐matched control cohort was similarly identified from EHR query for statistical comparison (N=150 466). Baseline for the control group was defined as date of first office visit within the medical system. Exclusion criteria from the control group included lifetime diagnosis of SSc, SLE, RA, PsA, IBD, or HIV or missing key demographic variables.
Statistical Analysis
All analyses were performed using R v4.0.2. 25 (p1), 26 , 27 , 28 For time‐dependent analyses, baseline was defined as time of SAID diagnosis, and event of interest was defined as time of HF diagnosis or time of last clinic visit for patients without HF. Incident HF was defined as HF diagnosis at least 6 months (182 days) after baseline. Patients with HF at baseline were excluded from these analyses. For all statistical analyses, a P value of ≤0.01 was used to determine significance. Bivariate Cox models were used to assess associations between clinically relevant covariates (age, sex, self‐reported race, insurance status, BMI, hypertension, diabetes, CKD, history of MI) and incident HF. A Cox proportional hazards model was used to analyze the associations between SAID diagnoses and incident HF, controlling for age, sex, BMI, race, and insurance status. The Cox model was then adjusted for history of CKD and then further adjusted for history of hypertension, diabetes, and MI.
Cox proportional hazards models were also used to assess the associations between cardiovascular drug use at baseline and incident HF, adjusting for age, sex, race, insurance status, BMI, hypertension, diabetes, MI, and CKD. Patients prescribed >1 cardiovascular drug were included, and only instances of drug prescription at baseline (<6 months after time of SAID diagnosis) were included to avoid confounding with guideline‐directed medical therapy prescribed after a diagnosis of incident HF. Incident HF was again defined as HF diagnosis occurring >6 months after SAID diagnosis. A Cox proportional hazards model was used to assess association between baseline exposure to each cardiovascular drug and incident HF in SSc, SLE, RA, and a combined group of these patients.
Results
Overall Population
Of 182 795 patients in our study cohort, 5555 had SSc, 19 730 had SLE, 54711 had RA, 9871 had PsA, 90 476 had IBD, and 2452 had HIV (Table 1). Baseline demographics were similar among the 5 SAIDs, with expected differences seen in the HIV group compared with other SAIDs (including younger average age and greater percentage of men). In our study cohort, the prevalence of established HF at baseline (time of SAID diagnosis) was 10.7% of patients with SSc, 4.7% of patients with SLE, 8.1% of patients with RA, 9.8% of patients with PsA, 3.8% of patients with IBD, and 3.5% of patients with HIV (Table 1). Among all patients with SAIDs and HF, echocardiographic data was available for 78.7%. At time of HF diagnosis, the majority of the total cohort (57.5%) had a left ventricular ejection fraction >50%, as did the majority of the subgroups with SSc, RA, SLE, PsA, and IBD, whereas the majority of patients with HIV had an initial left ventricular ejection fraction <50%. Rates of traditional comorbid HF risk factors were similar across the SAIDs. Of note, right HF was more prevalent in SSc compared with the other SAIDs and HIV, which is consistent with increased rates of pulmonary arterial hypertension in this group. 29 The control group had overall similar demographics to the group with SAIDs (mean age 61.7, 81.5% White, 5.5% insured by Medicaid, 49.8% with hypertension, 19.7% with diabetes, and 5.3% with MI), though was less female on average (54.4%) and CKD was less prevalent than in the cohort with SAIDs (9.9%).
Table 1.
Demographic and Clinical Characteristics of Study Population With SAIDs
| Overall SAID (N=182 795) | SSc (N=5555) | RA (N=54 711) | SLE (N=19 730) | PsA (N=9871) | IBD (N=90 476) | HIV (N=2452) | Control (N=150 466) | |
|---|---|---|---|---|---|---|---|---|
| Age, y | ||||||||
| Median [Q1–Q3] | 64 [50–75] | 65 [55–74] | 68 [58–78] | 57 [45–68] | 61.0 [50–70] | 63 [47–74] | 52 [40–60] | 62 [49–73] |
| Sex (%) | ||||||||
| Female | 130 611 (71.5) | 4618 (83.1) | 40 263 (73.6) | 17 455 (88.5) | 5612 (56.9) | 62 210 (68.8) | 453 (18.5) | 81 802 (54.4) |
| Race (%) | ||||||||
| Asian | 1339 (0.7) | 48 (0.9) | 405 (0.7) | 175 (0.9) | 69 (0.7) | 634 (0.7) | 8 (0.3) | 2019 (1.3) |
| Black | 20 221 (11.1) | 652 (11.7) | 6918 (12.6) | 4438 (22.5) | 348 (3.5) | 6695 (7.4) | 1170 (47.7) | 19 368 (12.9) |
| White | 153 256 (83.8) | 4603 (82.9) | 44 918 (82.1) | 13 938 (70.6) | 9037 (91.6) | 79 636 (88.0) | 1124 (45.8) | 122 563 (81.5) |
| Other | 7979 (4.4) | 252 (4.5) | 2470 (4.5) | 1179 (6.0) | 417 (4.2) | 3511 (3.9) | 150 (6.1) | 6515 (4.3) |
| Ethnicity (%) | ||||||||
| Hispanic | 8532 (4.7) | 329 (5.9) | 2527 (4.6) | 1299 (6.6) | 402 (4.1) | 3862 (4.3) | 113 (4.6) | 11 381 (7.6) |
| Non‐Hispanic | 153 925 (84.2) | 4525 (81.5) | 46 631 (85.2) | 16 841 (85.4) | 8805 (89.2) | 75 030 (82.9) | 2093 (85.4) | 122 863 (81.7) |
| Other/not listed | 20 338 (11.1) | 701 (12.6) | 5553 (10.1) | 1590 (8.1) | 664 (6.7) | 11 584 (12.8) | 246 (10.0) | 16 222 (10.8) |
| Insurance (%) | ||||||||
| Medicaid | 9403 (5.1) | 249 (4.5) | 2966 (5.4) | 2078 (10.5) | 664 (6.7) | 3111 (3.4) | 335 (13.7) | 8217 (5.5) |
| Medicare | 72 119 (39.5) | 2565 (46.2) | 28 756 (52.6) | 7089 (35.9) | 3491 (35.4) | 29 662 (32.8) | 556 (22.7) | 39 143 (26.0) |
| Private | 49 160 (26.9) | 1590 (28.6) | 13 412 (24.5) | 6443 (32.7) | 3966 (40.2) | 23 289 (25.7) | 460 (18.8) | 76 337 (50.7) |
| Self‐pay | 7925 (4.3) | 303 (5.5) | 2528 (4.6) | 1117 (5.7) | 420 (4.3) | 3408 (3.8) | 149 (6.1) | 14 082 (9.4) |
| No insurance recorded | 43 395 (23.7) | 814 (14.7) | 6848 (12.5) | 2857 (14.5) | 1266 (12.8) | 30 665 (33.9) | 945 (38.5) | 11 149 (7.4) |
| Other | 793 (0.4) | 34 (0.6) | 201 (0.4) | 146 (0.7) | 64 (0.6) | 341 (0.4) | 7 (0.3) | 1538 (1.0) |
| Median follow‐up (years) [Q1–Q3] | 5.6 [1.3–11.1] | 4.2 [0.8–9.9] | 6.0 [1.7–11.2] | 5.3 [1.2–10.5] | 5.9 [2.0–11.4] | 5.4 [1.1–11.2] | 6.1 [2.6–10.8] | 5.27 [1.1,10.0] |
| Heart failure prevalence at baseline (%) | 9974 (5.5) | 594 (10.7) | 4432 (8.1) | 919 (4.7) | 517 (5.2) | 3425 (3.8) | 87 (3.5) | 4187 (2.8) |
| Incident heart failure (%) | 10 779 (7.9) | 535 (14.4) | 4224 (10.4) | 1799 (11.6) | 444 (5.9) | 3662 (5.5) | 115 (6.1) | 6050 (4.0) |
| Echo data available (%) | 16 379 (78.7) | 952 (84.3) | 6716 (77.6) | 2213 (81.4) | 748 (77.8) | 5523 (77.9) | 227 (88.3) | |
| LVEF >50% (%) | 9413 (57.5) | 602 (63.2) | 3910 (58.2) | 1241 (56.1) | 434 (58.0) | 3145 (56.9) | 81 (35.7) | |
| 40%<LVEF≤50% (%) | 2681 (16.4) | 153 (16.1) | 1085 (16.2) | 369 (16.7) | 104 (13.9) | 917 (16.6) | 53 (23.3) | |
| LVEF ≤40% (%) | 4285 (26.1) | 197 (20.7) | 1721 (25.6) | 603 (27.2) | 210 (28.1) | 1461 (26.5) | 93 (41.0) | |
| Right heart failure (%) | 1799 (1.0) | 287 (5.2) | 644 (1.2) | 296 (1.5) | 54 (0.5) | 497 (0.5) | 21 (0.9) | 684 (0.5%) |
| Hypertension (%) | 94 251 (51.6) | 2938 (52.9) | 34 464 (63.0) | 10 477 (53.1) | 5478 (55.5) | 39 722 (43.9) | 1172 (47.8) | 74 867 (49.8) |
| Diabetes (%) | 36 071 (19.7) | 964 (17.4) | 13 573 (24.8) | 4283 (21.7) | 2351 (23.8) | 14 367 (15.9) | 533 (21.7) | 29 578 (19.7) |
| Myocardial infarction (%) | 12 678 (6.9) | 513 (9.2) | 4953 (9.1) | 1539 (7.8) | 724 (7.3) | 4752 (5.3) | 197 (8.0) | 7997 (5.3) |
| Chronic kidney disease (%) | 29 951 (16.4) | 1138 (20.5) | 10 583 (19.3) | 4127 (20.9) | 1485 (15.0) | 11 895 (13.1) | 723 (29.5) | 14 854 (9.9) |
IBD indicates inflammatory bowel disease; LVEF, left ventricular ejection fraction; PsA, psoriatic arthritis; RA, rheumatoid arthritis; SAID, systemic autoimmune inflammatory disease; SLE, systemic lupus erythematosus; and SSc, systemic sclerosis.
Incident Heart Failure in SAIDs
Median duration of follow‐up was 5.6 years (interquartile range, 1.3–11.1), during which 10 779 (7.9%) diagnoses of incident (new‐onset) HF occurred among all SAIDs. Rates of incident HF were 14.4% for SSc, 11.6% for SLE, 10.4% for RA, 5.9% for PsA, 5.5% for IBD, and 6.1% for HIV (Table 1). Kaplan–Meier survival analysis demonstrated greatest risk of HF development for patients with SSc, followed by RA and SLE (Figure 2). Bivariate Cox regression models demonstrated that age, male, Black, hypertension, diabetes, MI, and CKD were significantly associated with incident HF in all SAIDs (except age and sex in the group with HIV) (Table 2). Medicaid insurance was associated with increased incident HF only in the group with HIV. Although body mass index (BMI) was associated with incident HF in the groups with PsA and IBD, the effect size was very small. All of these clinically relevant covariates were included in the next multivariable regression models. An initial multivariable Cox proportional hazards model demonstrated a significant association with incident HF for all SAIDs when adjusted for age, sex, race, insurance status, and BMI (Figure 3A). Cox regression models additionally adjusting for CKD only (Figure 3B) or CKD, hypertension, diabetes, and MI (Figure 3C) demonstrated significantly elevated risk of incident HF with SSc (adjusted hazard ratio [aHR] of model adjusted for CKD, hypertension, diabetes, MI, 2.81 [95% CI, 2.57–3.08]; P<0.001), SLE (aHR, 1.64 [95% CI, 1.56–1.74]; P<0.001), RA (aHR, 1.54 [95% CI, 1.47–1.61]; P<0.001), PsA (aHR, 1.24 [95% CI, 1.13–1.37]; P<0.001), and IBD (aHR, 1.10 [95% CI, 1.05–1.15]; P<0.001). There was no statistically significant association with HIV after adjusting for covariates. The multivariable Cox proportional hazards analysis was repeated using the group with IBD as a control. The group with IBD was chosen as an alternative control because it was the largest of the subgroups with SAIDs and was least associated with HF risk compared with controls in the original analysis. Again, in this analysis, SSc (aHR, 2.62 [95% CI, 2.39–2.87]; P<0.001), SLE (aHR, 1.56 [95% CI, 1.47–1.65]; P<0.001), and RA (aHR, 1.42 [95% CI, 1.35–1.48]; P<0.001), were significantly associated with incident HF in an adjusted Cox regression model when compared with patients with IBD. PsA and HIV were not associated with increased incident HF when compared with the group with IBD.
Figure 2. Kaplan–Meier curves depicting HF‐free survival over 15 years, stratified by SAID diagnosis, with number at risk.
HF indicates heart failure; IBD, inflammatory bowel disease; PsA, psoriatic arthritis; RA, rheumatoid arthritis; SAID, systemic autoimmune inflammatory disease; SLE, systemic lupus erythematosus; and SSc, systemic sclerosis.
Table 2.
Bivariate Cox Regression Analysis of Association Between Covariates of Interest and Incident HF in Patients With SAIDs
| SAID | SSc | HR (95% CI) | RA | HR (95% CI) | SLE | HR (95% CI) | PsA | HR (95% CI) | IBD | HR (95% CI) | HIV | HR (95% CI) |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Covariate | P value | P value | P value | P value | P value | P value | ||||||
| Age | <0.001 | 1.03 (1.02–1.04) | <0.001 | 1.05 (1.05–1.06) | <0.001 | 1.04 (1.03–1.04) | <0.001 | 1.07 (1.06–1.07) | <0.001 | 1.06 (1.05–1.06) | 0.02 | 1.02 (1.00–1.04) |
| Sex, male | <0.001 | 1.46 (1.17–1.82) | <0.001 | 1.33 (1.25–1.42) | <0.001 | 1.98 (1.76–2.22) | 0.51 | 1.06 (0.88–1.28) | 0.05 | 0.93 (0.87–1.00) | 0.24 | 0.76 (0.49–1.19) |
| Race (Non‐White) | 0.03 | 1.26 (1.02–1.56) | <0.001 | 1.30 (1.22–1.40) | <0.001 | 1.55 (1.41–1.70) | 0.002 | 1.61 (1.19–2.18) | <0.001 | 1.47 (1.35–1.60) | 0.003 | 1.76 (1.21–2.56) |
| Insurance (Medicaid) | 0.81 | 0.95 (0.63–1.43) | 0.23 | 0.92 (0.81–1.05) | 0.78 | 0.98 (0.84–1.13) | 0.27 | 0.77 (0.47–1.23) | 0.18 | 0.89 (0.75–1.06) | 0.007 | 1.83 (1.18–2.86) |
| Body mass index | 0.07 | 0.99 (0.97–1.00) | 0.62 | 1.00 (1.00–1.00) | 0.18 | 1.00 (1.00–1.00) | <0.001 | 1.02 (1.01–1.03) | <0.001 | 1.01 (1.00–1.01) | 0.71 | 1.01 (0.98–1.03) |
| hypertension | <0.001 | 2.75 (2.21–3.41) | <0.001 | 6.50 (5.82–7.26) | <0.001 | 6.02 (5.19–6.99) | <0.001 | 8.97 (6.15–13.07) | <0.001 | 8.81 (7.95–9.76) | <0.001 | 8.38 (4.37–16.04) |
| Diabetes | <0.001 | 2.13 (1.78–2.56) | <0.001 | 2.57 (2.43–2.72) | <0.001 | 3.11 (2.85–4.41) | <0.001 | 3.33 (2.77–4.00) | <0.001 | 3.66 (3.45–3.90) | <0.001 | 5.64 (3.86–8.21) |
| Chronic kidney disease | <0.001 | 3.67 (3.09–4.34) | <0.001 | 4.98 (4.70–5.27) | <0.001 | 5.95 (5.24–6.53) | <0.001 | 6.10 (5.06–7.35) | <0.001 | 6.53 (6.15–6.95) | <0.001 | 9.30 (5.93–14.59) |
| Myocardial infarction | <0.001 | 3.39 (2.79–4.12) | <0.001 | 4.50 (4.22–4.79) | <0.001 | 5.03 (4.56–5.55) | <0.001 | 7.51 (6.20–9.11) | <0.001 | 6.72 (6.28–7.19) | <0.001 | 6.54 (4.48–9.53) |
HF indicates heart failure; HR, hazard ratio; IBD, inflammatory bowel disease; PsA, psoriatic arthritis; RA, rheumatoid arthritis; SAID, systemic autoimmune inflammatory disease; SLE, systemic lupus erythematosus; and SSc, systemic sclerosis
Figure 3. Incident heart failure risk by SAID diagnosis.
Adjusted for (A) age, sex, race, insurance, BMI; (B) age, sex, race, insurance, BMI, and CKD; or (C) age, sex, race, insurance, BMI, CKD, hypertention, diabetes, MI. Reference group is non‐SAID controls. BMI indicates body mass index; CKD, chronic kidney disease; IBD, inflammatory bowel disease; MI, myocardial infarction; PsA, psoriatic arthritis; RA, rheumatoid arthritis; SAID, systemic autoimmune inflammatory disease; SLE, systemic lupus erythematosus; and SSc, systemic sclerosis.
HF Association With Cardiovascular Drug Use
We also aimed to examine the relationships between common cardiovascular drug use at baseline and incident HF, adjusting for age, sex, race, BMI, insurance status, and history of CKD, diabetes, hypertension, and MI (Table 3). For this analysis, we focused on the subgroups with SSc, SLE, and RA, who showed the greatest significant association with incident HF. In a combined cohort of patients with SSc, SLE, and RA (n=59 802), angiotensin receptor blockers (aHR, 1.1 [95% CI, 1.1–1.2]; P<0.001) and thiazide (aHR, 1.2 [95% CI, 1.1–1.3]; P<0.001) use were significantly associated with increased risk of incident HF, whereas beta‐blocker (aHR, 0.7 [95% CI, 0.6–0.8]; P<0.001) use was associated with decreased risk of incident HF. In patients with SSc, angiotensin‐converting enzyme inhibitor (aHR, 1.5 [95% CI, 1.1–1.9]; P=0.005) and thiazide (aHR, 1.5 [95% CI, 1.2–1.9]; P=0.002) use was significantly associated with increased risk of incident HF. There was a trend toward association between baseline beta‐blocker use and decreased incident HF in SSc; however, it did not reach statistical significance (aHR, 0.7 [95% CI, 0.5–1.0]; P=0.063). In patients with SLE, baseline beta‐blocker use was associated with decreased incident HF (aHR, 0.7 [95% CI, 0.5–0.9]; P=0.002). In patients with RA, baseline thiazide (aHR, 1.1 [95% CI, 1.1–1.2]; P=0.006) use was associated with increased incident HF and beta‐blocker use was associated with decreased incident HF (aHR, 0.7 [95% CI, 0.6–0.8]; P<0.001). In all groups but the subgroup with SSc, beta‐blocker use at baseline was associated with a decreased risk of incident HF (Figure 4).
Table 3.
Association Between Baseline Cardiovascular Medication Use and Incident HF in SSc, SLE, RA, and a Combined Cohort
| Combined | SSc | SLE | RA | |||||
|---|---|---|---|---|---|---|---|---|
| aHR (95% CI) | P value | aHR (95% CI) | P value | aHR (95% CI) | P value | aHR (95% CI) | P value | |
| Angiotensin‐converting enzyme inhibitors | 1.1 (0.99–1.2) | 0.078 | 1.5 (1.12–1.9) | 0.005 | 0.93 (0.77–1.1) | 0.486 | 1.0 (0.95–1.1) | 0.398 |
| Angiotensin receptor blockers | 1.1 (1.1–1.2) | <0.001 | 1.4 (1.06–1.9) | 0.018 | 1.1 (0.91–1.3) | 0.340 | 1.1 (1.0–1.2) | 0.049 |
| Beta blockers | 0.72 (0.66–0.8) | <0.001 | 0.71 (0.5–1.0) | 0.063 | 0.68 (0.54–0.86) | 0.002 | 0.68 (0.6–0.76) | <0.001 |
| Calcium channel blockers | 1.1 (1.0–1.2) | 0.064 | 1.2 (0.94–1.5) | 0.153 | 0.89 (0.72–1.1) | 0.285 | 1.0 (0.92–1.1) | 0.749 |
| Thiazide | 1.2 (1.1–1.3) | <0.001 | 1.5 (1.2–1.9) | 0.002 | 1.2 (1.0–1.4) | 0.014 | 1.1 (1.1–1.2) | 0.006 |
Adjusted hazard ratios reported. Cox proportional hazards regression adjusted for age, sex, chronic kidney disease, hypertension, diabetes, coronary artery disease.
aHR indicates adjusted hazard ratio; HF, heart failure; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; and SSc, systemic sclerosis.
Figure 4. Incident HF risk for patients with SSc, SLE, RA and a combined cohort of all 3 SAIDs with baseline BB use.
ACE‐i indicates angiotensin‐converting enzyme inhibitor; ARB, angiotensin receptor blocker; BB, beta blocker; CCB, calcium channel blocker; HF, heart failure; RA, rheumatoid arthritis; SAID, systemic autoimmune inflammatory disease; SLE, systemic lupus erythematosus; and SSc, systemic sclerosis.
Discussion
Among patients with SAIDs in a large integrated health system that was about 10 times larger in sample size and included a longer follow‐up compared with Prasada et al., 4 we confirmed that diagnoses of SSc, SLE, and RA were associated with increased risk of incident HF, even after adjusting for traditional risk factors. In the adjusted models, we found a smaller association between PsA and IBD and incident HF and no association with HIV, supporting evidence that HF risk in these conditions is mitigated mostly by traditional cardiovascular risk factors. 22 , 30 , 31 , 32 The independent association between SSc, SLE, and RA and incident HF was observed compared with both a SAID‐free control group and to a large group with IBD with similar demographics.
Our study also importantly adds to the current knowledge by investigating the associations between cardiovascular drug use and new‐onset HF in patients with SAIDs, specifically in the 3 groups with SAIDs most independently associated with HF risk. Angiotensin receptor blockers and thiazide use was associated with slightly increased incident HF risk in the combined group of patients with SSc, SLE, and RA, though this is most likely attributable to clinical confounders for which we were unable to completely control. Interestingly, we found that in the combined group with SSc, SLE, and RA (those patients with SAIDs with the highest HF risk) beta‐blocker use at baseline was associated with a 30% decreased risk of incident HF. Given the trend observed with angiotensin receptor blockers and thiazides, one might expect beta‐blocker use to reflect underlying cardiac risk factors and therefore be associated with increased HF risk; however, the opposite was found. Although specific indications for beta blockers in these patients are not known, possible indications include hypertension, migraine, tremor, arrythmia, and anxiety. Although beta blockers have long been established to be cardioprotective in HF, we recognize that in a retrospective analysis, associations between drug use and HF risk may not directly imply therapeutic benefits. In addition, there have been concerns regarding the potential vascular effects of beta blockers in patients with poor peripheral circulation (especially in SSc where when beta blockers are contraindicated in the setting of renal crisis). 33 Nevertheless, patients with SAIDs often tolerate beta blockers well in the setting of established HF. Interestingly, agonist‐like beta‐adrenoceptor antibodies have been implicated in dilated cardiomyopathy and HF, suggesting a possible mechanistic link between autoimmunity and a potential cardioprotective role of beta blockers in this population. 34 , 35 , 36
This one of the largest cohorts of patients with SAIDs investigating incident HF risk reported to date; however, there are several limitations to this hypothesis‐generating study. Although this was a large cohort, generalizability is still limited as a single‐center study. Retrospective research is inherently limited by missing data, imperfect means of establishing controls, and relying on clinical documentation and longitudinal follow‐up. For example, ICD‐9/10 codes may not always accurately reflect a patient’s diagnosis, and there are limitations to using ICD codes to identify HF cases in EHRs. 37 , 38 Additionally, some important clinical characteristics, such as SAID disease severity, are not easily extractable in such a large data set. Excluding patients with a lifetime diagnosis of an SAID in the control group may introduce a health survivor bias, which could be corrected in future studies with a propensity score matching system. 39 Additionally, defining baseline times differently in the group with SAIDs and control group can introduce immortal time bias, which cannot be completely corrected by propensity matching. We repeated the adjusted Cox model using patients with IBD as the control group to mitigate the potential selection bias introduced in the definition of the SAID‐free control group. This analysis yielded similar results to the initial, with a significant increased risk of incident HF in patients with SSc, SLE, and RA, compared with those with IBD. Also, our data set did not have accurate documentation of precise timing and dosages of disease‐modifying antirheumatic drugs disease‐modifying antirheumatic drugs prescriptions to further investigate the effects of DMARD use on HF risks in these patient populations.
Conclusions
By including key traditional cardiovascular risk factors identified as significant covariates in our final multivariable analyses, we sought to demonstrate an association between SAID diagnosis and HF that is independent of those traditional factors. To further solidify this finding, future studies would benefit from inclusion of other clinically relevant HF risk factors, such as tobacco use, alcohol use, atrial fibrillation, valvular disease, obesity, obstructive sleep apnea, thyroid disease, and anemia in multivariable models. Additionally, more research is warranted regarding the association between non‐White race (mostly self‐identified Black and Asian patients in this population) and increased HF risk. Machine learning methods may be useful in incorporating greater numbers of clinically relevant covariates in a more detailed model. Ascertainment of HF risk may have also been affected by different clinical practices and protocols for cardiac screening for different SAIDs. For example, annual echocardiography is recommended for pulmonary hypertension surveillance in patients with SSc, which may skew our results towards capturing more patients with SSc and HF (especially HF with preserved ejection fraction) than in other SAIDs. 40 Also, we cannot confirm cardiovascular drug initiation, compliance, of definitive indication due to the retrospective nature of the study. We recognize that within the limitation of this analysis we can establish only association and cannot establish causality especially regarding treatment effects of cardiovascular drugs on outcomes. Future work on this topic should focus on clarifying the prescription patterns and exposure times for cardiovascular drugs in these patient populations to more definitively study the effects on HF risk in patients with SAIDs.
There remains a great need to address the morbidity and mortality burden of cardiac disease in the population with SAIDs. Further research into the clinical risk factors for HF in subgroups with SAIDs is also warranted to elucidate mechanisms underlying this risk. Clinical characteristics that reflect disease severity or inflammation may be more specifically examined to generate improved models of HF risk in these groups or clues regarding underlying mechanisms to inform basic science or drug discovery. In future investigations, it will be important to clarify the relative contribution of primary left heart dysfunction, pulmonary hypertension, and right heart dysfunction to the HF phenotypes observed in these patients, as these distinctions can be confounding and are clinically meaningful in terms of treatment and prognosis. 29 It will also be valuable to pursue more detailed cardiac imaging analysis to establish more robust imaging predictors or indicators of early cardiac involvement in SAIDs. Finally, given the interesting signal for a potentially protective role of beta blockers in these patient populations, future studies should seek to clarify this association. An important next step in retrospective analysis would be to use methods such as target trial emulation to approximate the level of detail patient characterization, treatment parameters, and group assignment procedures achieved through a randomized controlled trial. 41 , 42 If our findings can be replicated in retrospective target trial emulation, prospective randomized controlled trials may be warranted to test the hypothesis that in patients with SSc, SLE, or RA, beta‐blocker use can be cardioprotective in preventing the development of HF or underlying cardiomyopathy.
Sources of Funding
This project is supported by the Cleveland Clinic Lerner Research Institute Research Center of Excellence Award. Dr Pieter Martens is supported by a grant from the Belgian American Educational Foundation and by the Frans Van de Werf Fund.
Disclosures
Drs Zagouras and Martens have no relationships to disclose. Dr Tang is a consultant for Cardiol Therapeutics, Zehna Therapeutics, WhiteSwell, CardiaTec Biosciences, Alleviant Medical, Alexion Pharmaceuticals, Salubris Biotherapeutics, BioCardia, Tenax Therapeutics, Bridge Bio Pharma, Vasa Therapeutics, and has received honoraria from Springer Nature and Belvoir Media Group.
Supporting information
STROBE Checklist
This article was sent to Sakima A. Smith., MD, PhD, Associate Editor, for review by expert referees, editorial decision, and final disposition.
Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/JAHA.124.039155
For Sources of Funding and Disclosures, see page 11.
REFERENCES
- 1. Mitratza M, Klijs B, Hak AE, Kardaun JWPF, Kunst AE. Systemic autoimmune disease as a cause of death: mortality burden and comorbidities. Rheumatology. 2021;60:1321–1330. doi: 10.1093/rheumatology/keaa537 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Wright K, Crowson CS, Gabriel SE. Cardiovascular comorbidity in rheumatic diseases: a focus on heart failure. Heart Fail Clin. 2014;10:339–352. doi: 10.1016/j.hfc.2013.10.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Prasad M, Hermann J, Gabriel SE, Weyand CM, Mulvagh S, Mankad R, Oh JK, Matteson EL, Lerman A. Cardiorheumatology: cardiac involvement in systemic rheumatic disease. Nat Rev Cardiol. 2015;12:168–176. doi: 10.1038/nrcardio.2014.206 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Prasada S, Rivera A, Nishtala A, Pawlowski AE, Sinha A, Bundy JD, Chadha SA, Ahmad FS, Khan SS, Achenbach C, et al. Differential associations of chronic inflammatory diseases with incident heart failure. J Am Coll Cardiol HF. 2020;8:489–498. doi: 10.1016/j.jchf.2019.11.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Dhakal BP, Kim CH, Al‐Kindi SG, Oliveira GH. Heart failure in systemic lupus erythematosus. Trends Cardiovasc Med. 2018;28:187–197. doi: 10.1016/j.tcm.2017.08.015 [DOI] [PubMed] [Google Scholar]
- 6. Mavrogeni S, Koutsogeorgopoulou L, Dimitroulas T, Markousis‐Mavrogenis G, Kolovou G. Complementary role of cardiovascular imaging and laboratory indices in early detection of cardiovascular disease in systemic lupus erythematosus. Lupus. 2016;26:227–236. doi: 10.1177/0961203316671810 [DOI] [PubMed] [Google Scholar]
- 7. Lee KS, Kronbichler A, Eisenhut M, Lee KH, Shin JI. Cardiovascular involvement in systemic rheumatic diseases: an integrated view for the treating physicians. Autoimmun Rev. 2018;17:201–214. doi: 10.1016/j.autrev.2017.12.001 [DOI] [PubMed] [Google Scholar]
- 8. De Lorenzis E, Gremese E, Bosello S, Nurmohamed MT, Sinagra G, Ferraccioli G. Microvascular heart involvement in systemic autoimmune diseases: the purinergic pathway and therapeutic insights from the biology of the diseases. Autoimmun Rev. 2019;18:317–324. doi: 10.1016/j.autrev.2019.02.002 [DOI] [PubMed] [Google Scholar]
- 9. Caforio ALP, Marcolongo R, Baritussio A, Leoni L, Cheng CY, Iliceto S. Myocarditis in systemic immune‐mediated diseases. In: Caforio ALP, ed Myocarditis: Pathogenesis, Diagnosis and Treatment. Springer International Publishing; 2020:195–221. doi: 10.1007/978-3-030-35276-9_11 [DOI] [Google Scholar]
- 10. Khalid U, Egeberg A, Ahlehoff O, Lane D, Gislason GH, Lip GYH, Hansen PR. Incident heart failure in patients with rheumatoid arthritis: a Nationwide cohort study. J Am Heart Assoc. 2018;7:e007227. doi: 10.1161/JAHA.117.007227 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Nicola PJ, Maradit‐Kremers H, Roger VL, Jacobsen SJ, Crowson CS, Ballman KV, Gabriel SE. The risk of congestive heart failure in rheumatoid arthritis: a population‐based study over 46 years. Arthritis Rheum. 2005;52:412–420. doi: 10.1002/art.20855 [DOI] [PubMed] [Google Scholar]
- 12. Nadeem U, Raafey M, Kim G, Treger J, Pytel P, N Husain A, Schulte JJ. Chloroquine‐ and hydroxychloroquine‐induced cardiomyopathy: a case report and brief literature review. Am J Clin Pathol. 2021;155:793–801. doi: 10.1093/ajcp/aqaa253 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Chatre C, Roubille F, Vernhet H, Jorgensen C, Pers YM. Cardiac complications attributed to chloroquine and hydroxychloroquine: a systematic review of the literature. Drug Saf. 2018;41:919–931. doi: 10.1007/s40264-018-0689-4 [DOI] [PubMed] [Google Scholar]
- 14. Chung ES, Packer M, Lo KH, Fasanmade AA, Willerson JT. Randomized, double‐blind, placebo‐controlled, pilot trial of infliximab, a chimeric monoclonal antibody to tumor necrosis factor‐alpha, in patients with moderate‐to‐severe heart failure: results of the anti‐TNF therapy against congestive heart failure (ATTACH) trial. Circulation. 2003;107:3133–3140. doi: 10.1161/01.CIR.0000077913.60364.D2 [DOI] [PubMed] [Google Scholar]
- 15. Sinagra E, Perricone G, Romano C, Cottone M. Heart failure and anti tumor necrosis factor‐alpha in systemic chronic inflammatory diseases. Eur J Intern Med. 2013;24:385–392. doi: 10.1016/j.ejim.2012.12.015 [DOI] [PubMed] [Google Scholar]
- 16. Coletta AP, Clark AL, Banarjee P, Cleland JGF. Clinical trials update: RENEWAL (RENAISSANCE and RECOVER) and ATTACH. Eur J Heart Fail. 2002;4:559–561. doi: 10.1016/S1388-9842(02)00121-6 [DOI] [PubMed] [Google Scholar]
- 17. Allanore Y, Meune C, Vonk MC, Airo P, Hachulla E, Caramaschi P, Riemekasten G, Cozzi F, Beretta L, Derk CT, et al. Prevalence and factors associated with left ventricular dysfunction in the EULAR Scleroderma Trial and Research group (EUSTAR) database of patients with systemic sclerosis. Ann Rheum Dis. 2010;69:218–221. doi: 10.1136/ard.2008.103382 [DOI] [PubMed] [Google Scholar]
- 18. Valentini G, Huscher D, Riccardi A, Fasano S, Irace R, Messiniti V, Matucci‐Cerinic M, Guiducci S, Distler O, Maurer B, et al. Vasodilators and low‐dose acetylsalicylic acid are associated with a lower incidence of distinct primary myocardial disease manifestations in systemic sclerosis: results of the DeSScipher inception cohort study. Ann Rheum Dis. 2019;78:1576–1582. doi: 10.1136/annrheumdis-2019-215486 [DOI] [PubMed] [Google Scholar]
- 19. Yurkovich M, Vostretsova K, Chen W, Aviña‐Zubieta JA. Overall and cause‐specific mortality in patients with systemic lupus erythematosus: a meta‐analysis of observational studies. Arthritis Care Res. 2014;66:608–616. doi: 10.1002/acr.22173 [DOI] [PubMed] [Google Scholar]
- 20. Widdifield J, Paterson JM, Huang A, Bernatsky S. Causes of death in rheumatoid arthritis: how do they compare to the general population? Arthritis Care Res. 2018;70:1748–1755. doi: 10.1002/acr.23548 [DOI] [PubMed] [Google Scholar]
- 21. Lambova S. Cardiac manifestations in systemic sclerosis. World J Cardiol. 2014;6:993–1005. doi: 10.4330/wjc.v6.i9.993 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Hsue PY, Waters DD. Heart failure in persons living with HIV infection. Curr Opin HIV AIDS. 2017;12:534–539. doi: 10.1097/COH.0000000000000409 [DOI] [PubMed] [Google Scholar]
- 23. Felsen UR, Bellin EY, Cunningham CO, Zingman BS. Development of an electronic medical record‐based algorithm to identify patients with unknown HIV status. AIDS Care. 2014;26:1318–1325. doi: 10.1080/09540121.2014.911813 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Nagueh SF, Smiseth OA, Appleton CP, Byrd BF III, Dokainish H, Edvardsen T, Flachskampf FA, Gillebert TC, Klein AL, Lancellotti P, et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr. 2016;29:277–314. doi: 10.1016/j.echo.2016.01.011 [DOI] [PubMed] [Google Scholar]
- 25. Rich B. table1: Tables of Descriptive Statistics in HTML. 2020. https://CRAN.R‐project.org/package=table1, 10.1016/j.bbih.2020.100168. [DOI]
- 26. Grolemund G, Wickham H. Dates and times made easy with {lubridate}. J Stat Softw. 2011;40:1–25. doi: 10.18637/jss.v040.i03 [DOI] [Google Scholar]
- 27. Therneau T. A Package for Survival Analysis in R. 2020. https://CRAN.R‐project.org/package=survival.
- 28. Kassambara A, Kosinski M, Biecek P. survminer: Drawing Survival Curves using “ggplot2”. 2020. https://CRAN.R‐project.org/package=survminer, 10.1089/hs.2019.0061. [DOI]
- 29. Bourji KI, Kelemen BW, Mathai SC, Damico RL, Kolb TM, Mercurio V, Cozzi F, Tedford RJ, Hassoun PM. Poor survival in patients with scleroderma and pulmonary hypertension due to heart failure with preserved ejection fraction. Pulm Circ. 2017;7:409–420. doi: 10.1177/2045893217700438 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Zhu TY, Li EK, Tam LS. Cardiovascular risk in patients with psoriatic arthritis. Int J Rheumatol. 2012;2012:714321. doi: 10.1155/2012/714321 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Garshick MS, Ward NL, Krueger JG, Berger JS. Cardiovascular risk in patients with psoriasis: JACC review topic of the week. J Am Coll Cardiol. 2021;77:1670–1680. doi: 10.1016/j.jacc.2021.02.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Bigeh A, Sanchez A, Maestas C, Gulati M. Inflammatory bowel disease and the risk for cardiovascular disease: does all inflammation lead to heart disease? Trends Cardiovasc Med. 2020;30:463–469. doi: 10.1016/j.tcm.2019.10.001 [DOI] [PubMed] [Google Scholar]
- 33. Stern EP, Steen VD, Denton CP. Management of Renal Involvement in Scleroderma. Curr Treat Options Rheumatol. 2015;1:106–118. doi: 10.1007/s40674-014-0004-1 [DOI] [Google Scholar]
- 34. Wallukat G, Müller J, Podlowski S, Nissen E, Morwinski R, Hetzer R. Agonist‐like beta‐adrenoceptor antibodies in heart failure. Am J Cardiol. 1999;83(12A):75H–79H. doi: 10.1016/s0002-9149(99)00265-9 [DOI] [PubMed] [Google Scholar]
- 35. Nagatomo Y, Tang WHW. Autoantibodies and cardiovascular dysfunction: cause or consequence? Curr Heart Fail Rep. 2014;11:500–508. doi: 10.1007/s11897-014-0217-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Nagatomo Y, Li D, Kirsop J, Borowski A, Thakur A, Tang WHW. Autoantibodies specifically against β1 adrenergic receptors and adverse clinical outcome in patients with chronic systolic heart failure in the β‐blocker era: the importance of immunoglobulin G3 subclass. J Card Fail. 2016;22:417–422. doi: 10.1016/j.cardfail.2016.03.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Huang H, Turner M, Raju S, Reich J, Leatherman S, Armstrong K, Woods P, Ferguson RE, Fiore LD, Lederle FA. Identification of acute decompensated heart failure hospitalizations using administrative data. Am J Cardiol. 2017;119:1791–1796. doi: 10.1016/j.amjcard.2017.03.007 [DOI] [PubMed] [Google Scholar]
- 38. McCormick N, Lacaille D, Bhole V, Avina‐Zubieta JA. Validity of heart failure diagnoses in administrative databases: a systematic review and meta‐analysis. PLoS One. 2014;9:e104519. doi: 10.1371/journal.pone.0104519 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Zhang Z, Li X, Wu X, Qiu H, Shi H. Propensity score analysis for time‐dependent exposure. Ann Transl Med. 2020;8:246. doi: 10.21037/atm.2020.01.33 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Bissell LA, Anderson M, Burgess M, Chakravarty K, Coghlan G, Dumitru RB, Graham L, Ong V, Pauling JD, Plein S, et al. Consensus best practice pathway of the UK systemic sclerosis study group: management of cardiac disease in systemic sclerosis. Rheumatology (Oxford). 2017;56:912–921. doi: 10.1093/rheumatology/kew488 [DOI] [PubMed] [Google Scholar]
- 41. Hernán MA, Wang W, Leaf DE. Target trial emulation: a framework for causal inference from observational data. JAMA. 2022;328:2446–2447. doi: 10.1001/jama.2022.21383 [DOI] [PubMed] [Google Scholar]
- 42. Matthews AA, Danaei G, Islam N, Kurth T. Target trial emulation: applying principles of randomised trials to observational studies. BMJ. 2022;378:e071108. doi: 10.1136/bmj-2022-071108 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
STROBE Checklist
Data Availability Statement
This study was approved by the Cleveland Clinic Institutional Review Board with waiver of patients’ written consent as medical chart review project. Our institutional review board imposes restrictions on data availability for this project because we do not have patients’ authorization and consent to share their clinical data outside our institution. The authors have full access to all study data, take responsibility for its integrity and the accuracy of the analyses, and may agree to provide aggregate data upon reasonable request and upon approval by the institutional review board.