Objective To identify factors associated with elevated arterial stiffness in a 5-year follow-up of patients with ankylosing spondylitis (AS).
Methods C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), Bath AS disease activity index (BASDAI) and AS disease activity score (ASDAS) were recorded in 2003, and arterial stiffness (Augmentation Index (AIx) and pulse wave velocity (PWV)) in 2008/2009. Patients were grouped into quartiles according to baseline CRP, ESR and BASDAI and four ASDAS groups. Trend analyses were performed using ANCOVA (AIx/PWV as dependent variable) with separate models for CRP, ESR, BASDAI and ASDAS (age and gender adjusted). Independent predictors of future AIx and PWV levels were identified in multivariate linear regression models.
Results In total, 85 patients participated. Increasing baseline values of CRP, ESR and ASDAS were associated with elevated AIx on follow-up (p(trend) 0.01, 0.05 and 0.04, respectively). Similar non-significant patterns were seen for PWV. In the multivariate analyses, baseline CRP and ASDAS were independently associated with future elevated AIx (p=0.03 and0.02, respectively). In the multivariate PWV model, results for CRP and ASDAS were non-significant.
Conclusions Baseline CRP and ASDAS were associated with future elevated arterial stiffness measured as AIx, supporting that disease activity is related to future risk of cardiovascular disease in patients with AS.
- Ankylosing Spondylitis
- Cardiovascular Disease
- Disease Activity
Statistics from Altmetric.com
Patients with ankylosing spondylitis (AS) have an increased risk of cardiovascular disease (CVD) compared with the general population.1 However, the mediators of this increased CVD risk are not fully known. Inflammatory mechanism are involved in development of atherosclerosis, and C-reactive protein (CRP) is associated with increased CVD risk.2 Arterial stiffness is a validated risk marker of CVD and can be influenced by chronic inflammation.3 Cross-sectional studies have found increased arterial stiffness in AS compared with controls,4 ,5 but no previous studies have addressed longitudinal associations between inflammation/disease activity and arterial stiffness in patients with AS. Thus, the objective was to investigate if inflammation and disease activity in patients with AS are associated with future elevated arterial stiffness.
Design and patients
The patients were derived from a cohort of patients with AS established in 2002, diagnosed according to the modified New York criteria.6 The patients were asked to participate in examinations focusing on disease activity and functional limitations in 2003 and on CVD comorbidity in 2008/2009. The study and data collections at the two time points were approved by the local ethical committee and performed according to the Declaration of Helsinki. All patients signed an informed consent.
Information on demographics, maximum height at adulthood, comorbidities and medication was retrieved from questionnaires. Weight and height were measured and body mass index (BMI) calculated (kg/m2). Loss of height from adulthood until 2008/2009 was calculated as maximum height minus measured height in 2008/2009. Blood samples were analysed consecutively for CRP (COBAS Integra 800, Roche Diagnostic, Switzerland) and erythrocyte sedimentation rate (ESR) (Westergren method). Disease activity was assessed in 2003 and included the Bath AS disease activity index (BASDAI).7 AS disease activity score (ASDAS) could be computed based on the available variables.8
Central arterial stiffness was measured as Augmentation Index (AIx) and pulse wave velocity (PWV) in 2008/2009 using the Sphygmocor device (AtCor, Australia). AIx and PWV are both validated markers of CVD risk.9 ,10 AIx is an estimate of the augmentation of the central arterial pressure that is caused by the reflected pulse wave and is a surrogate measure of arterial stiffness. We assessed AIx by pulse-wave recordings at the radial artery, and the central pulse wave was derived through a validated transfer function.10 AIx was calculated as the percentage of the pulse pressure that is augmented by wave reflection, standardised to a heart rate of 75 bpm. PWV is a measure of the speed of the pulse wave through the central arteries. A higher pulse wave speed indicates stiffer arteries. We assessed PWV by pulse recordings at the carotid and femoral artery and recorded the time for the pulse to travel from the heart to the femoral and carotid artery with a simultaneously recorded ECG. The distance of the pulse to travel was calculated by measuring the distances between the carotid pulse and the sternal notch as well as between the sternal notch and the femoral pulse and then subtracting the carotid–sternal notch distance from the femoral–sternal notch distance. PWV was calculated as distance/time (m/s).10 The arterial stiffness was assessed by an experienced assessor (SAP) in the same time period as a study of patients with rheumatoid arthritis (RA) where reproducibility was tested and proven to be good.11
Statistical analyses were performed using SPSS, V.21. We used analyses of covariance to analyse associations between baseline inflammation/disease activity and arterial stiffness (AIx and PWV) at follow-up. Patients were grouped into quartiles according to the level of baseline CRP, ESR and BASDAI and into four ADSAS groups according to predefined cut-off values (inactive disease (<1.3), moderate (1.3–2.0), high (2.1–3.5) and very high (>3.5) disease activity).8 All analyses were adjusted for age and gender. p Values for trend were calculated.
We constructed multivariate models with AIx and PWV as dependent variables. Possible baseline predictors of arterial stiffness, disease activity (CRP, ASDAS) (continuous variables), smoking (current smoker vs non-smoker), BMI, medication (using non-steroidal anti-inflammatory drugs (NSAIDs) vs not using NSAIDs), disease-modifying antirheumatic drugs (DMARDs) as well as factors recorded in 2008/2009 known to have influence on the outcome (height (AIx models only), loss of height (PWV models only), central mean arterial pressure, use of statins and antihypertensive) were entered into age and gender adjusted univariate models. In the multivariate AIx and PWV analyses, we ran separate models for baseline CRP and ASDAS due to multicolinearity, and entered variables with a p value <0.25 from the univariate analyses. We performed backwards multivariate linear regression for all four models. The final models were assessed by residual plots and the R2 of the models were calculated.
Altogether 103 patients with AS were examined in both 2003 and 2008/2009, and out of these, 85 had the arterial stiffness examination. Five patients were excluded form PWV analyses due to low quality of the assessment (for details, see online supplementary figure S1).
Mean (SD) age in 2003 was 47.3 (12.6) years and 58.8% were men. At baseline, only one patient used tumour necrosis factor α (TNFα) inhibitor, 14 (16.5%) used DMARDs (sulfasalazine 10, methotrexate 4) and 71 (83.5%) used NSAIDs (table 1).
Figure 1 presents AIx and PWV across quartiles of baseline CRP, ESR, BASDAI and four ASDAS groups. AIx increased stepwise with increasing CRP, ESR and ASDAS in the age and gender adjusted models, with statistically significant trends (p values: 0.01, 0.05 and 0.04, respectively). A similar pattern, although not reaching statistically significance, was seen for PWV.
In the multivariate analyses, baseline CRP and ASDAS were significantly associated with future elevated levels of AIx, and baseline use of DMARDs was associated with low AIx in the ASDAS model (table 2). We found no significant associations between baseline CRP or ASDAS and future PWV, but baseline use of NSAIDs was significantly associated with low future PWV.
In patients with AS, baseline CRP and ASDAS were associated with elevated arterial stiffness, measured as AIx, after 5 years. We believe that this is the first study in AS to report that early inflammation and disease activity are associated with future elevated arterial stiffness, supporting that these factors contribute to future risk of CVD.
Cross-sectional studies comparing AS and controls have shown elevated arterial stiffness in AS, significantly for PWV and numerically for AIx.4 ,5 ,12 ,13 Mechanistically, inflammation can induce changes in the arterial wall both in the endothelium, muscular tone and structural components resulting in arterial stiffness.14 In the general population and in RA, CRP has been found to predict increased arterial stiffness.3 ,15 Thus, associations between baseline inflammation and future elevated arterial stiffness in patients with AS are in line with results in the general population as well as in patients with RA. Baseline BASDAI was not associated with elevated arterial stiffness, indicating that CRP or a composite measure of disease activity including an objective measure of inflammation such as ASDAS is more important when considering future CVD risk than disease activity measured only by patient-reported items such as BASDAI.
The differences in AIx between the highest and lowest ASDAS and CRP groups were 5% and 7%, respectively. McEniery et al16 found that a 5% increase in AIx by the age of 50 years corresponded to a 10-year increase in vascular age. Accordingly, the difference in AIx between the high-disease and low-disease activity groups seems to be clinically meaningful.
There were trends of higher PWV with higher CRP and ASDAS; however, the results were non-significant. The magnitude of the difference between the highest and lowest ASDAS and CRP groups was about 0.5 m/s, possibly representing a clinically important difference.16 The non-significant results may be due to a type-II error. Furthermore, PWV can be underestimated in patients with AS due to increased kyphosis.17 We did, however, attempt to adjust for loss of height to minimise this effect.
Use of DMARDs was associated with low AIx in the ASDAS model. The majority of the patients used sulfasalazine that may reduce disease activity in some patients with AS and thus may possibly be related to reduced CVD risk.18 Interestingly, we found that the use of NSAIDs was significantly associated with low PWV. In patients with high-disease activity, use of NSAIDs could theoretically be a mediator of elevated levels of arterial stiffness, but this is less likely in our study where baseline use of NSAIDs was associated with future low levels of arterial stiffness. In the general population, use of NSAIDs is associated with increased CVD risk,19 but in patients with AS the knowledge on this relation is limited. In line with our results, Bakland et al20 found that infrequent use of NSAIDs was associated with future increased mortality. However, channelling bias cannot be excluded as a confounder neither in the present study nor in the study by Bakland et al.20
The clinical implications of our results are that inflammation and disease activity are possible CVD risk factors in patients with AS, and hence reduction of these factors may be important in reducing CVD risk. A few intervention studies with TNFα inhibitors have addressed such a hypothesis without achieving a conclusive evidence of the effect of TNFα inhibitors on arterial stiffness.21
The major strength of this study is the longitudinal design. Furthermore, the cohort represents a wide range of disease activity reflected by the mean ASDAS at 2.5 and SD at 1.2.
Major limitations are the low number of patients and a possible selection bias of the patients participating. Furthermore, data on disease activity and medication between baseline and 5-year follow-up are lacking. There was no assessment of arterial stiffness at baseline and hence we do not know if arterial stiffness increased during the study period. The traditional CVD risk factors were not recorded in 2003 and the effect of these could not be evaluated.
In conclusion, we found that high baseline CRP and ASDAS were associated with future elevated arterial stiffness, supporting that inflammation and disease activity are CVD risk factors in patients with AS.
Camilla Fongen PT, MSc, Diakonhjemmet Hospital played a key role in the acquisition of the data. Cecilie Okkenhaug, Department of Clinical Chemistry, Diakonhjemmet Hospital provided facilities to perform the laboratory measurements.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online figure
Handling editor Gerd R Burmester
Contributors All authors contributed to design, analyses and interpretation of these data, drafting and/or revising the manuscript and approved the final manuscript.
Funding This work was supported by grants from the South-Eastern Norway Regional Health Authority (grant number 2011100).
Competing interests None.
Patient consent Obtained.
Ethics approval Regional Committee of Ethics, SouthEast Norway.
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.