Objective Disease duration and disease activity may be associated with an increased risk of cardiovascular disease (CVD) in rheumatoid arthritis (RA). The objectives of this study were to investigate (1) the relationship between duration of inflammation and the development of CVD in RA patients and (2) the relationship between RA disease activity over time and CVD in patients with RA.
Methods RA patients with a follow-up of ≥6 months in the Nijmegen early RA cohort without prior CVD were included. Disease activity over time was calculated using the time-averaged 28 joint disease activity score (DAS28) for each patient. Kaplan–Meier survival analysis and Cox proportional hazards regression were used for the analyses.
Results During follow-up of the 855 patients that were included, 154 CV events occurred. The course of hazards over time did not indicate a change in the risk of CVD over the course of RA (disease duration), which is also reflected by the absence of a deflection in the survival curves. The survival distributions did not differ between patients with a disease duration of <10 years or >10 years (Log-rank test: p=0.82). Time-averaged DAS28 was significantly associated with CVD (p=0.002) after correction for confounders.
Conclusions Disease duration does not appear to independently affect the risk of CVD. The risk of CVD in RA patients was not increased after 10 years of disease duration compared with the first 10 years. Disease activity over time may contribute to the risk of CVD.
- Rheumatoid Arthritis
- Cardiovascular Disease
- Disease Activity
Statistics from Altmetric.com
Rheumatoid arthritis (RA) is a chronic inflammatory systemic disease of multifactorial aetiology. Despite important progresses in its treatment, RA is still associated with excess mortality rates of approximately 25%.1 Cardiovascular disease (CVD) represents the leading cause of death in RA, accounting for approximately 50% of all excess mortality.2 The excess risk of CVD cannot be completely explained by traditional risk factors alone.3 Growing evidence supports the notion that inflammatory and immune mechanisms underline atherosclerosis4 ,5 It is hypothesised that chronic systemic inflammation in RA represents a disease related risk factor, accounting for extra cardiovascular (CV) risk.6
Recent studies have investigated the link between the presence of inflammatory markers used to determine disease activity in RA, such as erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP), and the development of CVD. CRP has been indicated as a predictor for (accelerated) atherosclerosis.7 ,8 ESR and CRP have also been associated with CVD in RA and polyarthritis9–12 Atherosclerotic plaques in the carotid artery appear more severe and prevalent in RA patients compared with the general population.13–16 In comparison to healthy controls and RA patients in remission, RA patients with active disease seem to have less stable plaques, more vulnerable to rupture, increasing the probability of an acute CV event.17 Additionally, inflammation is likely to modulate traditional CV risk factors, including lipids, endothelial function and insulin sensitivity.18 ,19 This led to the ‘smaller slice of a larger pie’ concept in which traditional CV risk factors have a smaller contribution to CV risk in RA patients than in the general population.6 The genotype of RA patients may also be of interest, potentially contributing to unfavourable lipid patterns and accelerated atherosclerosis.20 ,21
Although a growing body of pathophysiological evidence supports the hypothesis that inflammation contributes to the development of CVD, clinical research has yet to clearly demonstrate the relationship between RA disease activity and CVD. In addition to disease activity, perhaps the time spent by a patient in an inflammatory state would be crucial for the chance to develop CVD. Long-standing disease elongates a patient's exposure to chronic inflammation, and it has been suggested that this prolonged exposure has a cumulative effect and disease duration may therefore act as a separate CVD risk factor.22 Disease duration over 10 years has been indicated as a risk factor for CVD in the European League Against Rheumatism (EULAR) recommendations for CV risk management in RA patients,23 even though some studies have shown that the risk of CVD is already increased in the early, sometimes preclinical stages of RA compared with the general population.24–26 However, the excess risk of CVD reported in early RA may still continue to increase as the disease progresses over time. Therefore, the objectives of the present study are (1) to investigate the relationship between duration of inflammation and CV risk corrected for the level of inflammation and (2) to investigate the relationship between RA disease activity and CV risk in patients with RA.
Study design and patients
This study is based on prospectively collected data from the Nijmegen early RA inception cohort. Patients were included at diagnosis of RA (baseline) in the outpatient clinic of the departments of rheumatology of the Radboud University Medical Centre (since 1985) or the Maartenskliniek (since 1990) in Nijmegen, The Netherlands. At inclusion, patients had a disease duration of <1 year, were disease-modifying antirheumatic drug (DMARD)-naïve and fulfilled the 1987 ACR criteria for the classification of RA. All patients provided written informed consent, and this project was approved by the medical ethical committee, CMO Arnhem Nijmegen. Patients were included if follow-up time was >6 months and if they had no history of CVD prior to RA diagnosis. The total cohort included 1018 patients at the time of analysis. The 77 patients who had a history of CVD prior to inclusion were excluded, as well as 86 patients who had a follow-up of less than 6 months. In total, 855 patients were included for analysis.
Primary and secondary outcomes
The primary outcome in this study was the occurrence of a CV event by physician diagnosis, which was retrieved by extensive review of medical charts and electronic patient files. Included events were acute coronary syndrome comprising both myocardial infarction (MI) and unstable angina pectoris, angina pectoris, cerebrovascular event or stroke, transient ischemic attack (TIA), peripheral artery disease (PAD) and revascularisation procedures including coronary artery bypass surgery, percutaneous coronary intervention and percutaneous transluminal coronary angioplasty. Both fatal and non-fatal events were included. Deaths due to CVD were verified from death certificates, provided by Statistics Netherlands.27 Heart failure, cerebral haemorrhage and non-coronary cardiac death (ie, arrhythmias) were not included as CV events for the purpose of this study.
Baseline patient characteristics were retrieved from the RA inception cohort database, including age, sex, rheumatoid factor (RF) status, disease activity (28 joint disease activity score (DAS28)), initial antirheumatic treatment and treatment with biological DMARDs. The time-averaged DAS28 score was calculated by taking the area under the curve of the DAS28 score of the total follow-up period divided by the follow-up period. Baseline data regarding CV risk factors were collected by review of patients’ charts and electronic patient files, including smoking status (yes/no (Y/N)), blood pressure, use of medication preventative for CVD, body mass index (BMI) and diabetes mellitus. Non-fasting blood samples were used to measure total cholesterol (TC), high-density lipoprotein cholesterol (HDL-c).
The cut-off point of 10 years was used to differentiate between patients with and without ‘long-standing’ disease, in reference to the EULAR recommendations that indicate a disease duration >10 years as a risk factor for CVD. Differences in baseline variables between patients who had a CV event before 10 years of disease duration and patients who had a CV event after 10 years were tested using t test, Wilcoxon test or χ2 test, as appropriate. The effect of disease duration on the risk of CVD was analysed by Kaplan–Meier survival analysis and Cox proportional hazards regression analysis.
The effect of disease duration on the risk of CVD was investigated using Cox proportional hazards regression. Survival and hazard curves were used to visualise the risk of CVD over time and estimate the effect of disease duration on CV risk. Sex, age, smoking (Y/N), BMI (weight (kg)/height(m)2), systolic blood pressure (mm Hg), TC (mmol/L), HDL-c (mmol/L), DAS28, ESR (mm/h), CRP (mg/L), swollen and tender joints, visual analogue scale (VAS, mm) and diabetes at baseline (Y/N), treatment for CVD risk factors (Y/N), initial antirheumatic treatment (methotrexate, sulfasalazine or other), (MTX) ever (Y/N), treatment with biological ever (Y/N) and RF status (positive/negative) were considered as possible confounders.
The risk of a CV event during the first 10 years of the disease was compared with the risk of a CV event after 10 years of disease duration (up to 25 years) by means of log-rank testing and Kaplan–Meier survival analysis. For this purpose, survival experience was grouped into two. For the first group (group 1), patients were selected who were at risk for a first CV event in the first 10 years of disease duration, that is, all patients who were included in this study, and patient time stopped at the time of event, or was censored at least after 10 years of event-free follow-up. For the second group (group 2), patients were selected (again) if they had a follow-up >10 years and were free of CVD up until that point. Patient time stopped at the time of event or was censored at least at the censoring date 30-09-2011. The effect of the level of inflammation over time (time-averaged DAS28 and time-averaged ESR) was analysed using Cox proportional hazard regression with time-averaged DAS28 as the main independent variable. Age, sex, smoking (Y/N), BMI (weight (kg)/height(m)2), systolic blood pressure (mm Hg), TC (mmol/L), HDL-c (mmol/L), DAS28 and diabetes at baseline (Y/N), RF status (positive/negative) and initial antirheumatic treatment (methotrexate, sulfasalazine or other), treatment with methotrexate ever (Y/N), treatment with biological ever (Y/N) were considered as possible confounders.
Missing values on variables were imputed using multiple imputation analysis with five repetitions.
Patient characteristics and CV events
In total, 855 patients were included, comprising 9959 patient years. Patients had a mean±SD disease duration of 11.7±6.1 years. Patient characteristics at baseline are presented in table 1 for the whole group and separately for patients with a CV event within the first 10 years following disease onset and with a CV event after more than 10 years. In group 1, 76% of patients were treated with MTX compared with 73% in group 2, and 30% were treated with a biological in group 1 versus 33% in group 2. A total of 154 CV events, of which 16 were fatal, were registered, including 64 cases of acute coronary syndrome, 19 cases of stable angina pectoris, 30 cases of strokes, 15 cases of TIAs, 21 cases of PAD and 5 revascularisation procedures. Missing values ranged from 0.1% for RF to 10.3% for smoking at baseline.
Disease duration and the risk of CVD
The linearity of both the survival and the proportional hazard curve shows that the cumulative risk increased with a similar rate as disease duration increased and that the risk per year remained constant (figure 1A); also the hazard for CVD remained similar as disease duration increased (figure 1B). If the risk of CVD would increase as disease duration increases, the cumulative survival and (expected) hazard lines would show a curve upwards as time progresses. After correction for confounders (table 2), the curves did not change (not shown).
The results from the Kaplan–Meier survival analysis in which the risk of CVD during the first 10 years of disease duration (group 1) was compared with the risk of CVD after 10 years of disease duration (group 2) show similar survival distributions that did not differ significantly between groups, with p=0.82. The survival distributions are presented in figure 2, showing overlapping curves.
Disease activity and the risk of CVD
The mean±SD time-averaged DAS28 was 3.6±1.1. The results of the Cox proportional hazard regression with the time-averaged DAS28 as the main independent variable are presented in table 3. After correction for confounders, the time-averaged DAS28 had a significant effect on the risk of CVD (p=0.002). With every point the DAS28 increases, the hazard for CVD increases with 0.281. Overall there was a difference in survival distributions between patients with low (<3.2), moderate (3.2–5.1) and high (>5.1) DAS28 over time (p=0.028). The survival curve of patients with a consistently high disease activity, that is, a time-averaged DAS28 >5.1, was the lowest (figure 3). After correction for confounders, this group (DAS28 >5.1) did not have a significantly different effect on CV risk compared with the <3.2 group (p=0.074). Time-averaged ESR was not significantly associated with CVD (p=0.805) (data not shown).
According to the results of this study, disease duration did not appear to affect the risk of CVD in patients with RA. Furthermore, our data showed that mean RA disease activity over the course of the disease (time-average DAS28) may contribute to the risk of CVD.
The survival distribution for CVD was linear as disease duration increased. Specifically the shape (linearity) of the survival curves during the first 10 years of RA and during the years thereafter was very similar, and the survival distributions were not significantly different. If CV risk would increase in patients with long-standing disease, the survival distribution curve would be expected to bend upwards as time progresses. The EULAR recommendations include disease duration >10 years as a risk factor for CVD, but the evidence for this choice is limited.23 In a study by Gabriel et al1 in which survival trends in the RA population were investigated, it was reported that the excess mortality did not become apparent until 8–10 years after disease onset,1 that is, after longer disease duration. However, this association was observed for all-cause mortality and the effect of disease duration on the risk of CVD was not investigated separately in this study. In a study by Naz et al,28 it was reported that that all-cause and CVD mortality was increased in RF-positive RA patients compared with the general population. CVD mortality was increased in both early (5 years) and later follow-up (10 years). This is more in line with findings from several other studies that suggest the excess risk of CVD is already present in the early stages of RA.24–26 In our cohort, a higher prevalence of CVD was found in patients during their first 10 years of RA compared with patients still at risk for their first CV event after 10 years. Pathophysiological data have delivered support for the notion of disease duration as risk factor for CVD, as it has been reported that (accelerated) atherosclerosis appears to be more severe in RA patients with established disease.22 However, this cross-sectional association with a surrogate marker has not been shown to translate in increased risk for actual CV events. Also, patients with RA appear to be more prone to plaque instability and rupture, in addition to accelerated atherosclerosis. Inflammation in RA may therefore contribute more specifically to more severe acute coronary syndromes and strokes,29 which may be more strongly associated with the presence and severity of local or systemic inflammation than with disease duration.
The results of this study suggest that increased disease activity over time increases CV risk. Interestingly, ESR did not appear to significantly affect CV risk. Additional research is necessary to further investigate the relationship between individual DAS28 components and CV risk. Different patient groups divided based on disease activity (<3.2, 3.2–5.1, >5.1) did not differ significantly from each other in terms of association with CV risk. However, when looking at the differences in the survival curves, patients with a consistently high level of disease activity (time-averaged DAS28>5.1) appear to be at a substantially higher risk of developing CVD compared with patients with lower disease activity levels.
In a previous study, we have suggested that an increased risk of CVD may already be observed in patients with low levels of disease activity30 as we have shown that there was no difference in the level of disease activity or the number of patients with low to moderate levels of disease activity between RA patients with and without MI (and similar disease duration).30 This may also mean that complete sustained eradication of systemic inflammation, that is, sustained clinical remission, is required to substantially reduce CV risk in RA patients. Indeed, there is some evidence to suggest that remission may be beneficial in preventing CV events.31 ,32
There are some points that should be considered when interpreting the results of this study. The cohort that was used in this study spans more than 25 years and provides a relatively large sample of RA patients with long follow-up times and a considerable number of recorded CV events, which increases the reliability of our findings. However, calendar time trends in the prevalence of CVD may be a source of bias. In the general population in both USA and Western Europe,33–35 the risk of CVD has been reduced in the past decades, while in RA the risk of CVD does not appear to have changed.1 Therefore, a calendar time trend on the prevalence of CVD in our cohort was considered unlikely. Also, medication used to treat the systemic inflammation in RA, particularly the recent introduction of biological DMARDs, may affect CV risk. Initial antirheumatic treatment and treatment with MTX or a biological during follow-up were included in the analysis and proved to be confounders. The widespread introduction of biologicals may lead to an overall decrease in disease activity levels in the RA population and to more effective suppression of systemic inflammation in individual RA patients. We did not investigate the effect of specific dosages of antirheumatic medication or cumulative (biologic) DMARD use. The possible beneficial effects of more recently developed treatment strategies on CV risk could contribute more heavily to the overall CV risk in the group with shorter disease duration (<10 years disease duration). Subsequently, this skewed distribution could magnify the possible effect of disease duration on CV risk. However, this would only strengthen our results as we did not find a significant effect of disease duration on CV risk in spite of this potential source of bias. Also, as different forms of treatment may be associated with the main independent variable, for example, disease duration and DAS28, or act as a proxy for determinants of the DAS28 such as treatment response and control of disease activity, this relationship is complex. The (causal) role of medication on CVD in RA fell outside the scope of this study and was not separately analysed. Investigating the effect of calendar time on CVD would require age-period-cohort analysis using larger cohorts. The effect of these recently adapted treatment strategies on CVD may take some time to become clinically significant. In addition to time trends, missed CV events due to random misclassification could lead to bias. However, patients enrolled in this cohort are seen and checked regularly (every 3–6 months). Data, including CV events, are gathered during these check-ups and meticulously recorded in our database. In addition, data of all patients that were included in this study were thoroughly checked and all events that were recorded during follow-up were double-checked to confirm the date of occurrence and type of the event. Deaths were also recorded during follow-up in the database, and for these patients the cause of death was confirmed. Therefore, missed events or misclassification of CV events was considered an unlikely source of bias in this study.
In conclusion, our results show that disease duration does not appear to independently affect the risk of CVD. Therefore, disease duration may not be the best choice for a disease-specific predictor when estimating CV risk in individual RA patients, as recently recommended by EULAR. Disease activity over time appears to contribute to the risk of CVD in patients with RA, particularly in case of persistent poorly controlled, high disease activity over time.
Handling editor Tore K Kvien
Contributors All authors have given substantial contribution to the conception and design and/or analysis and interpretation of the data, have drafted and/or revised the manuscript critically for important intellectual content and given final approval of the version to be submitted for publication. EEAA had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the analysis.
Competing interests All authors have completed the Unified Competing Interest form at http://www.icmje.org/coi_disclosure.pdf. EEAA was partially funded by the Rheumatology Research University Nijmegen foundation. PLCMvR has received travel grants and advisory board fees from Abbott, Pfizer, Roche and UCB.
Ethics approval Approved by the responsible medical ethical committee, CMO Arnhem Nijmegen.
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.