Article Text
PDF
Abstract
Objectives Disease activity has emerged as a new, independent risk factor for cardiovascular disease in patients with rheumatoid arthritis (RA). We tested if disease activity in RA was associated with lower left ventricular (LV) systolic function independent of traditional cardiovascular risk factors.
Methods Echocardiographic assessment was performed in 78 patients with RA having low, moderate or high disease activity (Simplified Disease Activity Index (SDAI) >3.3), 41 patients in remission (SDAI ≤3.3) and 46 controls, all without known cardiac disease. LV systolic function was assessed by biplane Simpson ejection fraction, stress-corrected midwall shortening (scMWS) and global longitudinal strain (GLS).
Results Patients with active RA had higher prevalence of hypertension and diabetes compared with patients in remission and controls (both p<0.05). LV ejection fraction (endocardial function) was normal in all three groups, while mean scMWS and GLS (myocardial function) were reduced in patients with RA with active disease compared with patients with RA in remission (95±18% vs 105±17% and −18.9±3.1% vs −20.6±3.5%, respectively, both p<0.01). Patients with RA in remission had similar scMWS and GLS as the controls. In multivariable analyses, having active RA was associated with lower GLS (β=0.21) and scMWS (β=−0.22, both p<0.05), both reflecting lower LV systolic myocardial function, independent of cardiovascular risk factors and LV ejection fraction. Classification of RA disease activity by other disease activity composite scores yielded similar results.
Conclusions Active RA is associated with lower LV systolic myocardial function despite normal ejection fraction and independent of traditional cardiovascular risk factors.
- Rheumatoid Arthritis
- Cardiovascular Disease
- Disease Activity
- Arterial Hypertension
Statistics from Altmetric.com
Introduction
Patients with rheumatoid arthritis (RA) have an increased risk of cardiovascular (CV) disease and twice as high risk of developing heart failure as the general population.1 ,2 Lately, high disease activity has emerged as a risk factor for CV disease in patients with RA independent of more traditional risk factors.3 ,4
Left ventricular (LV) ejection fraction is a strong prognosticator for CV disease in the normal population.5 ,6 However, new prognostically validated methods like stress-corrected midwall shortening (scMWS) and global longitudinal strain (GLS) have shown that reduced LV systolic myocardial function may be present although ejection fraction is normal.7–9 Thus, current guidelines recommend assessment of LV systolic myocardial function from GLS and scMWS in addition to ejection fraction.10 ,11
Several studies have reported reduced GLS in patients with RA,12 ,13 and recently, also reduced scMWS was found among patients with RA.14 Further, treatment of 80 patients with RA with an interleukin-1 inhibitor improved GLS, suggesting that RA-associated CV inflammation impairs LV myocardial function.15 However, a systematic assessment of the association between RA disease activity and LV systolic myocardial function has not been carried out. The aim of the study was to test if higher RA disease activity was associated with lower ejection fraction, scMWS and GLS, independent of traditional CV risk factors.
Methods
Study population
Patients with RA, diagnosed by the 1987 American College of Rheumatology criteria,16 were recruited from two previously described cohorts, the 15-year follow-up of the European Research on Incapacitating Diseases and Social Support (EURIDISS) cohort and the 10-year follow-up of the Oslo RA register cohort.17–19 All living participants were contacted by letter and asked to participate in the current study. Of the 152 (57%) persons who accepted to participate, patients without known cardiac disease and in whom speckle tracking echocardiography (STE) could be performed were included in the present project. In total, 11 patients were excluded due to previous cardiac surgery/percutaneous coronary intervention or myocardial infarction, 17 patients were excluded due to incomplete data collection and 5 were excluded because of suboptimal quality of echocardiographic recordings for STE analyses. The present RA study population then consisted of 119 patients with RA, 79 from the EURIDISS cohort and 40 from the Oslo RA register cohort.
Control subjects without inflammatory joint disease matching the patients with RA for sex, age and residential area were selected by Statistics Norway (Oslo, Norway). Of 329 invited community control subjects, 132 (40%) showed up for the echocardiographic examination. A majority of these community controls had established CV disease or hypertension and were excluded for the present study, leaving 47 (14%) of the control subjects. Of these, one additional control subject was excluded due to suboptimal quality of echocardiographic images for STE analyses, leaving a total of 46 controls.
All participants signed an informed consent according to the Declaration of Helsinki, and the study protocol was approved by the South Eastern Norwegian Regional Committee for Medical and Health Research Ethics.
RA disease activity
As recommended for clinical studies, remission was defined as Simplified Disease Activity Index (SDAI) score ≤3.3.20 Active RA was defined as not being in remission (SDAI score >3.3). Disease activity was characterised as low if SDAI score >3.3 and ≤11, moderate if SDAI >11 and ≤26 and high when SDAI score >26.21
SDAI was calculated as the sum of the swollen and tender joints counts in 28 joints (SJC28 and TJC28)+physician's assessment of disease activity on a visual analogue scale (VAS) (cm)+patient's assessment of global health on a VAS (cm)+C reactive protein (CRP) (mg/dL). We also calculated the Clinical Disease Activity Index (CDAI) as (TJC28+SJC28+VASphysician (cm)+VASpatient (cm)) and the Disease Activity Score in 28 Joints (DAS28) as [0.56 
0.70ln(Erythrocyte sedimentation rate (ESR))+0.014(VASpatient [mm])]21 and classified patients into remission, low, moderate and high disease activity according to established cut-points also for these composite measures.22
Self-reported information about the participants' medical history, smoking status and current medication was collected by a standardised questionnaire, and the information was quality assured by the consultant cardiologist. Physical disability was evaluated by the Modified Health Assessment Questionnaire (MHAQ).23
Circulating biomarkers
Rheumatoid factor (RF) and anti-cyclic citrullinated peptide (anti-CCP) antibodies were analysed from frozen serum or plasma by ELISAs (Inova Diagnostics, San Diego, California, USA). Total cholesterol and CRP were analysed in fasting blood samples by the COBAS 6000 and ESR by the Westergren method (Roche Diagnostics, Basel, Switzerland).
Blood pressure
Brachial blood pressure was measured according to the European Society of Hypertension guidelines,24 using an OMRON M7 apparatus (Kyoto, Japan). The average of the last two measurements was reported. Hypertension was defined as a history of hypertension, use of antihypertensive medication or systolic office blood pressure ≥140 mm Hg and/or diastolic office blood pressure ≥90 mm Hg.
Conventional echocardiography
All transthoracic echocardiograms were performed from 2007 to 2010 using a standardised protocol on a Vivid 7 (GE Vingmed Ultrasound, Horten, Norway) scanner. The echocardiographic images were analysed offline (HM) on digital workstations with Image Arena software V.4.1 (TomTec Imaging Systems GmbH, Unterschleissheim, Germany) at the Echocardiography Core Laboratory at Haukeland University Hospital, Bergen, Norway. All images were proofread by a highly experienced reader (EG).
Quantitative echocardiography was performed following the joint guidelines from the European Association of Cardiovascular Imaging and American Society of Echocardiography.10 LV mass was indexed for height2.7 and LV hypertrophy was considered present if LV mass index exceeded the prognostically validated cut-off values ≥46.7 g/m2.7 in women and ≥49.2 g/m2.7 in men.25 ,26 Concentric geometry was considered present if LV relative wall thickness exceeded the prognostically validated cut-off ≥0.43.27
Ejection fraction was assessed following current guidelines by the biplane method of discs (modified Simpson's rule) and classified as low if <54% in women and <52% in men.10 LV myocardial systolic function was estimated by scMWS using validated equations.28 ,29 The high reproducibility of scMWS by this method has been published from a separate, large study in our core laboratory (intraclass correlation coefficient 0.92 (95% CI 0.89 to 0.99).30
Speckle tracking echocardiography
LV longitudinal strain was measured by STE using EchoPAC BT12 software (GE Vingmed Ultrasound). Since the patients/controls with known CV disease were excluded, and none of the patients or controls had regional wall motion abnormality, longitudinal strain from the four-chamber view was considered representative for the whole LV.31 Peak systolic longitudinal strain was measured in six segments (basal, middle and apical segments in the septum and lateral wall, respectively), and the average value was reported as GLS. Since the LV length shortens during contraction in the apical views, the GLS is a negative percentage value. Less shortening indicates a reduction in myocardial function reflected by lower (less negative) GLS.
For the analyses, the cardiac cycle with the best image quality was selected. End systole was defined by aortic valve closure obtained from a pulsed wave Doppler recording in the LV outflow tract. The endocardial border was tracked manually and the region of interest (ROI) adjusted to include the whole LV wall thickness. The software then automatically tracked the movement of ultrasound speckles originating from the interference between scattered reflections from small objects from frame to frame. If the automatic tracking was poor, the ROI was adjusted. From a total of 990 segments, 120 (12%) were excluded because of insufficient image quality.
All measurements were performed by the same investigator (HM). Intraobserver variability was assessed in 20 randomly selected patients by repeated analysis on the same cine loop. Mean frame rate was 75±5 frames/s.
Statistics
The statistical analyses were done using IBM SPSS statistics V.22.0 (IBM, Armonk, New York, USA). Normally distributed continuous data were expressed as mean and SD and non-normally distributed continuous data as median and IQR. Categorical variables were presented as numbers and percentages. Non-normally distributed continuous variables (ESR, CRP, SDAI and CDAI) were log transformed before inclusion in univariable and multivariable analyses. According to the SDAI score, patients were grouped into either active RA disease (>3.3) or RA disease in remission (≤3.3). Comparisons between groups were performed by the χ2 test, two-sample Student's t-test, general linear model with Sidak's post hoc test or one-way analysis of variance with Scheffe's post hoc test as appropriate. Univariable associations were assessed with linear regression models and results reported as standardised β coefficients and p values. Multivariable linear regression analyses were done using an enter method with colinearity tools and the results reported as multiple R2 for the overall models and standardised β coefficients and p values for individual variables. Reproducibility of GLS measurement was tested by the intraclass correlation coefficient. A two-tailed p value of ≤0.05 was considered statistically significant in all analyses.
Results
Clinical characteristics
Of the 78 patients with active disease (SDAI>3.3), 55 had low disease activity, 22 moderate activity and 1 high activity. Forty-one patients were in remission (SDAI score ≤3.3).
The active RA group included more patients with hypertension and diabetes than the RA in remission group (p<0.05), but the total cholesterol level was lower (table 1). The active RA group also included more anti-CCP and RF positive patients and higher use of disease modifying antirheumatic drugs (DMARD) compared with the RA in remission group (p<0.05) (table 1).
Clinical characteristics of patients with active RA, RA in remission and control subjects
The healthy control subjects were younger and had lower blood pressures than the patients with RA (table 1). Of note, the inflammatory markers CRP and ESR were comparable between patients with RA in remission and control subjects.
LV structure and systolic function
Patients with active RA had smaller internal diameter of the LV but similar wall thicknesses, resulting in an increased LV relative wall thickness (more concentric LV) compared with patients with RA in remission and controls (p<0.05) (table 2).
Echocardiographic characteristics of patients with active RA, RA in remission and control subjects
Only two patients had mildly reduced ejection fraction, both in the active RA group. Mean LV ejection fraction was similar between patients with RA in remission and active disease, but scMWS and GLS were lower in patients with active RA compared with those in remission (p<0.05) (table 2 and figure 1). The patients with RA in remission had similar LV systolic myocardial function as the controls (table 2). scMWS was higher among controls compared with the patients with active RA (p=0.05). GLS was numerically higher among control subjects compared with the patients with active RA, but this difference was not statistically significant (table 2). Reproducibility of GLS measurements was excellent (intraclass correlation coefficient 0.95 (95% CI 0.87 to 0.98)).
LV systolic function among patients with active RA and RA in remission. *p=0.005 for active RA versus RA in remission, †p=0.006 for active RA versus RA in remission. GLS, global longitudinal strain; LV, left ventricular; RA, rheumatoid arthritis; scMWS, stress-corrected midwall shortening.
In univariable analyses, having active RA or increasing levels of both CDAI and SDAI score as well as MHAQ were associated with lower scMWS (all p<0.02) (see online supplementary table S1). The association between having active RA and lower scMWS remained significant after adjusting for main covariables of scMWS, including age, sex, LV ejection fraction and smoking in multivariable analyses (table 3 and see online supplementary table S2). Adding body mass index, systolic blood pressure, hypertension, diabetes, total cholesterol or RF to the model did not change these results. In subsequent models, other RA disease activity scores (CDAI and SDAI) as well as the disability score MHAQ were also significantly associated with lower scMWS in the multivariable models, but DAS28 was not (table 3 and see online supplementary table S2).
Association between active RA, RA disease activity composite scores and the disability score, MHAQ, with scMWS in multivariable analyses
Supplementary tables
In univariable analyses among patients with RA, having active RA (SDAI>3.3) and increasing levels of SDAI, CDAI and DAS28 scores were all associated with lower GLS (p<0.05) (see online supplementary table S1). In multivariable analyses, having active RA or increasing levels of disease activity assessed by the CDAI, SDAI and DAS28 score were all associated with lower GLS after adjustment for other main covariables of GLS identified in univariable analyses, including age, sex, body mass index, systolic blood pressure and LV ejection fraction (table 4 and see online supplementary table S3). Substituting systolic blood pressure with hypertension or adding diabetes to the models yielded similar results.
Association between active RA, RA disease activity composite scores and the disability score, MHAQ, with GLS in multivariable analyses
Regional strain
Longitudinal strain was lowest at the base and increased towards apex for both groups (figure 2). Patients in remission had numerically higher longitudinal strain in all segments, but this was only statistically significant for the middle and apical segments of the lateral wall (p<0.05).
Regional differences in LV GLS among patients with RA in remission and patients with active RA. *p<0.05 compared with RA in remission. GLS, global longitudinal strain; LV, left ventricular; RA, rheumatoid arthritis.
Discussion
This study demonstrates that patients with active RA have lower LV systolic myocardial function compared with patients in remission, independent of LV ejection fraction and presence of major CV risk factors like hypertension, diabetes or smoking. Of note, the association between RA disease activity and LV systolic myocardial function was confirmed with different composite scores used for the assessment of RA disease activity.
The vast majority of patients with RA in our study had normal LV ejection fraction, while LV systolic myocardial function was consistently lower among patients with RA with active disease. It is well known from different patient populations that LV systolic myocardial function may be reduced despite normal LV ejection fraction.7 ,8 Ejection fraction measures predominantly radial and to a lesser degree longitudinal LV systolic function.32 Ischaemic coronary artery disease is first evident in the longitudinally orientated fibres of the endocardium which is reflected by reduced GLS.33 Given the high prevalence of ischaemic coronary artery disease seen among patients with active RA,3 it is possible that the reduction in GLS seen in patients with active RA in this study may reflect subclinical coronary artery disease. Sex differences in GLS have been demonstrated in the Framingham Heart Study.34 The lack of significant difference in GLS between the control group and active RA group probably reflects the lower proportion of women in the control group. However, the lower scMWS found in patients with active RA points to non-ischaemic mechanisms like inflammation or hypertension-associated myocardial fibrosis in the midwall. The circumferential myocardial fibres of the midwall are particular susceptible to development of fibrosis.35 The fibrotic tissue is non-contracting and has been linked to reduced LV systolic function both in experimental studies and in different patient populations.36 ,37 In patients with RA, a recent cardiac magnetic resonance study showed an increased amount of midwall fibrosis with a non-ischaemic pattern, which again was associated with reduced circumferential strain and higher disease activity.38 In a recent study of 235 patients with RA, reduced scMWS was particularly associated with concentric LV geometry and hypertrophy.14 Lower scMWS is also common in hypertension, in particular when LV concentric geometry is present.11 The present study adds to this previous knowledge by demonstrating that reduced LV myocardial function in RA is particularly associated with higher disease activity, and further expands our previous finding of increased LV relative wall thickness among patients with active RA.39
Experimental studies have shown that inflammatory cytokines influence both LV function and structure negatively.40 As proof of concept in a study of 80 patients with RA, treatment with an interleukin-1 receptor antagonist led to improved GLS, in particular in patients with coronary artery disease.15 Another study of 53 treatment-naive patients with RA showed more diverse results, as reduced GLS was associated both with higher disease activity and with lower levels of anti-CCP antibodies.13 Also, in patients with RA with inflammatory activity, recent meta-analyses and review papers have concluded that suppression of inflammation may be beneficial for reduction of CV disease.41 ,42 The present findings support this view by further linking higher RA disease activity to lower LV myocardial function.
There are some limitations to our study. The cross-sectional study design is unsuited to make any causal connection between RA disease activity and LV systolic myocardial function. Possible selection bias could not be excluded since the patients with RA were recruited from different follow-up cohorts. The low participation rate among invited controls may have introduced a selection bias. Although patients with known CV disease were excluded from the present study, subclinical CV disease may have been present in the study participants. The study strengths are that we used a core laboratory as recommended for echocardiographic studies,43 and that we used indices of LV systolic function that have proven prognostic value and are recommended in current guidelines.10 ,11 Furthermore, the average disease activity was low in the study population, suggesting that the association with LV myocardial function may be more pronounced in less well-treated patients with RA.
Conclusion
Higher RA disease activity was associated with lower LV systolic myocardial function, whether measured by scMWS or GLS and independent of presence of a normal LV ejection fraction and traditional CV risk factors like hypertension, diabetes and current smoking. Our findings point to that good disease activity control in patients with RA may have clinical implications for avoiding heart failure in these patients. However, our findings should be further tested in a prospective outcome study in patients with RA.
Acknowledgments
The authors are grateful to Sella Provan for her contributions with some of the data collection.
References
Footnotes
Handling editor Gerd R Burmester
Contributors All authors have (1) given substantial contribution to the conception and design and/or analysis and interpretation of the data, (2) drafted and/or revised the manuscript critically for important intellectual content and (3) given final approval of the version to be submitted for publication.
Funding This study has received funding from the Norwegian Extra Foundation for Health and Rehabilitation and the South-Eastern and Western Regional Health Authorities of Norway.
Competing interests None declared.
Ethics approval Regional Committee for Medical and Health Research Ethics (REC) South East (Norway).
Provenance and peer review Not commissioned; externally peer reviewed.