Reduced arterial elasticity in rheumatoid arthritis and the relationship to vascular disease risk factors and inflammation

Authors


Abstract

Objective

Rheumatoid arthritis (RA) is associated with increased rates of cardiovascular disease. Reduced small artery elasticity (SAE) and large artery elasticity (LAE) and increased systemic vascular resistance (SVR) have been found in other high-risk groups. In the present study, we sought to determine whether arterial elasticity was reduced and SVR was increased in RA patients compared with controls matched for coronary artery disease (CAD) status, and to relate the results to vascular disease risk factors, including measures of inflammation.

Methods

Arterial elasticity was assessed by pulse wave analysis in RA patients with (n = 15) and without (n = 38) CAD, and in controls matched 1:1 for age, sex, and CAD status. Vascular risk factors, including high-sensitivity C-reactive protein (hsCRP), soluble vascular cell adhesion molecule 1 (sVCAM-1), and serum amyloid A (SAA) levels, were assessed.

Results

SAE and LAE were significantly lower and SVR was significantly higher in RA patients than in controls. RA patients also had higher levels of hsCRP, SAA, and sVCAM-1. SAE and LAE values were inversely correlated with markers of inflammation. Associations of SAE and LAE with RA were independent of conventional risk factors, but were dependent on markers of inflammation.

Conclusion

Vascular function is abnormal in RA, with reduced SAE and LAE and increased SVR relative to controls. Arterial elasticity is inversely associated with measures of inflammation. These measures may be clinically useful in the detection and monitoring of vascular disease in RA.

Rheumatoid arthritis (RA) is a systemic immune and inflammatory disease associated with increased morbidity and mortality (1). The standardized mortality ratio for patients with RA is 2.0, and their median survival is as many as 17 years less than that of the general population (2). Most of this excess mortality is attributable to cardiovascular events (1, 2). Endothelial dysfunction is an early event in the pathogenesis of atherosclerosis. Damage to the arterial wall due to aging and atherosclerosis causes decreased arterial elasticity (3). Atherosclerosis is a systemic disease, and its effects on the vascular system may be measured noninvasively by peripheral tests of function, such as pulse wave analysis (4). Pulse wave analysis is well tolerated and cost effective and is a validated measure of vascular integrity in high-risk groups of patients without RA (vascular disease, hypertension, diabetes, and renal failure) (5).

Inflammation is a potent vascular disease risk factor in the general population and is associated with endothelial dysfunction (6). The pathogenic events in atherosclerosis share many similarities with synovial inflammation in RA: activation of inflammatory cells and increased expression of adhesion molecules and cytokines (7). C-reactive protein (CRP) is a well-established disease activity marker in RA (8), and when measured by high-sensitivity assay (hsCRP), it is predictive of atherosclerosis in the general population (9). Hence, inflammation may contribute to increased vascular disease risk in RA by promoting endothelial and vascular wall damage.

In the present study, we used pulse wave analysis to assess vascular health in RA patients and control subjects with and without known coronary artery disease (CAD). Vascular health was compared and related to traditional vascular risk factors and measures of RA disease activity and inflammation.

PATIENTS AND METHODS

Patients and controls.

RA patients over the age of 18 years who were attending the rheumatology and cardiology public and private clinics of St. Vincent's Hospital and Royal Melbourne Hospital were eligible for the study. All patients fulfilled the American College of Rheumatology (formerly, the American Rheumatism Association) diagnostic criteria for RA (10). Exclusion criteria were diabetes mellitus (fasting blood glucose ≥7.0 mmoles/liter), renal impairment (serum creatinine >0.13 mmoles/liter), cardiac arrhythmia, valvular heart disease, recent surgery, recent illness unrelated to RA, and pregnancy. Approximately 5% of the RA patients were ineligible because of these exclusion criteria. There were 53 RA patients (15 with CAD and 38 without CAD).

The 53 control subjects consisted of 15 non-RA subjects with CAD (from the Cardiology Clinic) and 38 healthy volunteers (from the community and hospital staff). RA and non-RA subjects were matched for age (±5 years in all but 6 cases), sex, and CAD status.

Subjects with CAD had a previous hospital admission for unstable angina or myocardial infarction, as confirmed by changes on the electrocardiogram or by increases in cardiac enzyme levels, or they had positive findings on angiogram, a previous angioplasty, or previous coronary artery bypass grafts. CAD-negative RA patients were asymptomatic for CAD, had negative findings on a Rose questionnaire (11), and normal findings on a resting 12-lead electrocardiogram. To verify the absence of significant CAD, dobutamine echocardiography was performed in 14 of the 38 asymptomatic RA patients who agreed to undergo testing; the results were negative in these 14 RA patients. Only 1 RA patient was excluded from the study because of the presence of hypertrophic obstructive cardiomyopathy, which was identified by echocardiography.

All study subjects completed a questionnaire to assess smoking status, family history of vascular events, education level, and current medication use. The activity and extent of RA were measured with the swollen joint count (28 joints counted), the modified Health Assessment Questionnaire (calculated score between 0 and 3), average duration of early morning stiffness (in minutes, as estimated by the patient), and patient's assessment of pain and global well-being (on 10-cm linear visual analog scales) (8).

Biochemical assays.

Blood samples were taken with subjects in the fasting state. A complete blood cell count, erythrocyte sedimentation rate (ESR), renal and liver function tests, lipid profile, glucose level, glycosylated hemoglobin, and rheumatoid factor assays were performed. Serum amyloid A (SAA) and hsCRP were assayed by immunonephelometry (BN-II nephelometer; Dade-Behring, Marburg, Germany), with interassay coefficients of variation (CVs) of 5.4% and 5.6%, respectively. Soluble vascular cell adhesion molecule 1 (sVCAM-1) was measured by enzyme-linked immunosorbent assay (ELISA; R&D Systems, Abingdon, UK), with an interassay CV of 8.5%. Low-density lipoprotein (LDL) particle size was determined by gradient gel electrophoresis, with an inter-gel CV of 3.1% (12). Serum samples for LDL sizing obtained from each RA patient and his or her matched control were run in adjacent lanes of the gel; the research assistant (HH) was blinded to the identity of the subject from whom the samples were obtained.

Pulse wave analysis.

Pulse wave analysis using an HDI CR-2000 Pulsewave analyzer (Hypertension Diagnostics, Eagan, MN) was performed with subjects lying supine on the table, with a blood pressure (BP) cuff placed on the left arm and a pressure transducer placed on the right radial artery. Subjects were examined in the fasting state since we have demonstrated that measurements of vascular function are inconsistent after food, which may bias the results (data not shown). The age, sex, height, and weight of each subject were entered into the instrumentation control panel prior to the automated measurement of heart rate, systolic BP, and diastolic BP, and assessment of the pulse wave at the right radial artery.

After data acquisition, the inbuilt software analyzed the waveform based on a modified Windkessel model of the circulation to calculate arterial elasticity and systemic vascular resistance (SVR). Data calculated included the body mass index (BMI), pulse pressure, small artery elasticity (SAE), large artery elasticity (LAE), and SVR. The mean of three 1-minute readings taken over 15 minutes was used. The reproducibility of the SAE and LAE values was 9.8% and 10.2%, respectively.

Statistical analysis.

Statistical analyses were performed using SPSS (SPSS Inc., Chicago, IL) and SAS (SAS Institute, Cary, NC) software. Differences in levels of continuous variables between matched cases and controls were tested by paired t-test or Wilcoxon's signed rank test, as appropriate. Differences in categorical variables were evaluated by McNemar's test or chi-square test. Differences in SAE, LAE, and SVR between subjects with and without CAD were tested by t-test and analysis of covariance. Associations of SAE, LAE, and SVR with age were examined using linear regression, and associations with other risk factors were assessed by partial correlation coefficients, adjusted for age and sex. Data for SAE, ESR, hsCRP, SAA, sVCAM-1, and triglycerides were loge-transformed prior to use in correlation analyses. Independent associations of SAE, LAE, and SVR with RA were assessed using conditional logistic regression analysis. P values less than 0.05 were considered significant.

RESULTS

Clinical characteristics.

Patients and controls did not differ by age, postsecondary education, BP, adiposity, presence of a family history of vascular disease, or lipid profiles (Table 1). RA subjects were more likely to be current smokers or ever smokers, but current smok-ing status did not differ significantly between the 2 groups. BP did not differ between the groups, but treatment for hypertension was more common in RA. Antihypertensive agents used in RA and control groups were angiotensin-converting enzyme (ACE) inhibitors (n = 11 and n = 8, respectively), beta-blockers (n = 12 and n = 7), calcium-channel blockers (n = 7 and n = 8), and diuretics (n = 7 and n = 6).

Table 1. Clinical and demographic characteristics of the study subjects*
 Rheumatoid arthritis patientsControl subjectsP
Without CAD (n = 38)With CAD (n = 15)All patients (n = 53)Without CAD (n = 38)With CAD (n = 15)All controls (n = 53)
  • *

    Except where indicated otherwise, values are the mean (95% confidence interval). CAD = coronary artery disease; BMI = body mass index; BP = blood pressure; LDL = low-density lipoprotein; HDL = high-density lipoprotein.

  • Systolic BP ≥140 mm Hg and/or diastolic BP ≥90 mm Hg and/or currently taking antihypertensive medication.

  • Values are the geometric mean (95% confidence interval).

Age, years50 (47–54)65 (61–70)55 (51–58)51 (47–54)63 (59–67)54 (51–57)0.578
BMI, kg/m226.6 (24.9–28.3)28.9 (26.2–31.7)27.2 (25.8–28.7)25.5 (24.3–26.7)29.3 (26.1–32.4)26.6 (25.3–27.9)0.450
Systolic BP, mm Hg133 (128–138)135 (126–145)134 (129–138)124 (119–128)147 (136–157)130 (125–135)0.372
Diastolic BP, mm Hg75 (71–79)74 (71–77)74 (72–78)71 (68–74)78 (73–82)73 (70–75)0.431
Pulse pressure, mm Hg58 (55–62)62 (54–69)59 (56–62)53 (50–56)69 (61–78)58 (54–61)0.439
Antihypertensive drugs, %248742387260.039
Hypertension, %5593661887380.001
Total cholesterol, mmoles/liter5.6 (5.2–6.0)5.0 (4.2–5.8)5.4 (5.1–5.8)5.7 (5.4–6.1)4.9 (4.3–5.6)5.5 (5.2–5.8)0.719
LDL cholesterol, mmoles/liter3.4 (3.0–3.7)2.7 (2.5–3.0)3.2 (2.9–3.5)3.5 (3.2–3.8)2.7 (1.9–3.4)3.3 (3.0–3.6)0.536
LDL particle size, nm26.3 (26.1–26.5)26.0 (25.8–26.3)26.2 (26.1–26.4)26.1 (25.9–26.3)25.8 (25.6–26.0)26.0 (25.9–26.2)0.023
HDL cholesterol, mmoles/liter1.50 (1.33–1.67)1.24 (1.06–1.42)1.43 (1.30–1.57)1.57 (1.43–1.72)1.09 (0.95–1.23)1.45 (1.32–1.57)0.887
Triglycerides, mmoles/liter1.3 (1.1–1.6)1.8 (1.4–2.4)1.4 (1.2–1.7)1.1 (0.9–1.4)2.1 (1.5–2.9)1.3 (1.1–1.6)0.538
Lipid-lowering drugs, %07321067191.000
Current smokers, %1413140720.070
Ever smokers, %4187542253310.017
Postsecondary education, %3821336514510.137
Family history of vascular events, %3360413667451.000

Among the RA patients grouped according to the presence and absence of CAD, there was a similar prevalence of rheumatoid factor and similar levels of disease activity, as measured by the swollen joint count, duration of early morning stiffness, the modified Health Assessment Questionnaire, and patient's assessment of pain and global well-being (Table 2). RA patients with CAD had a higher use of aspirin (P < 0.001) than those without CAD. Although the difference was not statistically significant, RA patients with CAD tended to have a longer duration of RA and higher use of prednisolone, cyclooxygenase 2–specific inhibitors, and non-aspirin simple analgesics.

Table 2. Clinical characterization of rheumatoid arthritis disease severity*
 Patients without CAD (n = 38)Patients with CAD (n = 15)All patients (n = 53)P
  • *

    Except where indicated otherwise, values are the median (interquartile range). P values compare patients with coronary artery disease (CAD) versus those without CAD. HAQ = Health Assessment Questionnaire; DMARDs = disease-modifying antirheumatic drugs (including methotrexate [n = 30], hydroxychloroquine sulfate [n = 13], sulfasalazine [n = 12], gold [n = 5], leflunomide [n = 4], azathioprine [n = 2], and cyclosporine [n = 2], and combination therapy [n = 16]); NSAIDs = nonsteroidal antiinflammatory drugs; COX-2 = cyclooxygenase 2.

Disease duration, years6 (2–19)16 (4–21)7 (3–20)0.150
Rheumatoid factor positive, %6160600.972
Disease activity    
 Swollen joint count (range 0–28 joints)3 (0–6)2 (1–4)3 (1–6)0.884
 Early morning stiffness, minutes30 (5–60)30 (0–60)30 (5–60)0.662
 Modified HAQ (range 0–3)0.313 (0–1.125)0.250 (0–0.600)0.250 (0–0.875)0.431
 Pain score (range 0–10)3.0 (1.0–6.0)4.0 (2.4–5.0)3.5 (1.3–5.0)0.818
 Patient's global assessment (range 0–10)2.5 (1.0–6.0)2.0 (1.0–5.0)2.0 (1.0–5.0)0.504
Medication, %    
 DMARDs9086890.707
 Prednisolone3464420.052
 NSAIDs167130.665
 COX-2 inhibitors1331180.209
 Aspirin09325<0.001
 Non-aspirin simple analgesics925130.173

Inflammatory vascular risk factors.

As expected, the ESR, hsCRP, SAA, and sVCAM-1 values were significantly elevated in RA patients compared with the controls (Table 3). There was no statistically significant variation in markers of inflammation by CAD status. Among the control subjects, those with CAD had higher levels of markers of inflammation.

Table 3. Markers of inflammation*
 Rheumatoid arthritis patientsControl subjectsP
Without CAD (n = 38)With CAD (n = 15)All patients (n = 53)Without CAD (n = 38)With CAD (n = 15)All controls (n = 53)
  • *

    Values are the median (interquartile range). P values compare rheumatoid arthritis patients versus controls, by Wilcoxon's signed rank test. CAD = coronary artery disease; WBC = white blood cell; ESR = erythrocyte sedimentation rate; hsCRP = high-sensitivity C-reactive protein; SAA = serum amyloid A; sVCAM-1 = soluble vascular cell adhesion molecule 1.

WBC count, ×109/liter6.4 (5.3–7.4)6.6 (5.2–8.4)6.3 (5.3–7.2)5.5 (4.7–6.7)6.7 (5.7–8.8)5.6 (4.8–6.8)0.128
ESR, mm/hour14 (5–33)16 (5–41)12 (5–27)7 (5–15)10 (5–42)8 (5–14)0.037
hsCRP, mg/liter11.4 (4.5–17.3)10.2 (3.8–19.6)10.0 (3.7–15.6)1.1 (0.7–3.4)2.2 (1.8–8.4)1.6 (0.7–2.7)<0.001
SAA, mg/liter12.2 (6.8–53.4)12.0 (5.5–103)10.9 (6.2–42.3)4.6 (2.1–8.0)5.4 (2.6–7.5)4.6 (2.5–7.3)<0.001
sVCAM-1, ng/ml639 (564–811)634 (558–770)624 (556–723)457 (398–552)722 (499–1130)469 (413–577)<0.001

Vascular status.

There was an inverse association of age with SAE and LAE in both groups, and a positive association of age with SVR (Figure 1). For SAE, the slope of this relationship was greater for the controls than for the RA patients (P = 0.008 for the interaction of age and RA status), but the correlation was significant for both controls and patients. The associations of LAE and SVR with age were significant (P = 0.013 and P = 0.001, respectively) but did not vary significantly by RA patient or control status (Figure 1).

Figure 1.

Relationship of small artery elasticity, large artery elasticity, and systemic vascular resistance (SVR) to age in patients with rheumatoid arthritis (solid circles and broken line) and control subjects (open circles and solid line). Bcases = regression coefficient for the rheumatoid arthritis patients; Bcontrols = regression coefficient for the control subjects.

RA patients had lower average SAE and LAE than did the controls (Figure 2). The geometric mean SAE values in RA patients and controls was 3.5 ml/mm Hg × 100 (95% confidence interval [95% CI] 3.1–4.0) and 4.6 ml/mm Hg × 100 (95% CI 3.9–5.5), respectively (P = 0.001). The geometric mean LAE values in RA patients and controls were 11.9 ml/mm Hg × 10 (95% CI 10.8–13.1) and 14.7 ml/mm Hg × 10 (95% CI 13.3–16.1), respectively (P = 0.005). Despite normal BP and similar average BP values in patients and controls (Table 1), SVR was higher in RA patients (1,606 dynes · second · cm–5 [95% CI 1,507–1,705]) than in controls (1,480 dynes · second · cm–5 [95% CI 1,391–1,569]) (P = 0.039).

Figure 2.

Relationship of small artery elasticity, large artery elasticity, and systemic vascular resistance (SVR) to coronary artery disease (CAD) status in rheumatoid arthritis (RA) patients and controls. Values are the mean and 95% confidence interval or, for small artery elasticity, the geometric mean and 95% confidence interval. P values compare RA patients and controls, by paired t-test.

Among the control subjects, the presence of CAD was associated with significantly lower SAE (P = 0.013). Similarly, LAE was lower and SVR was higher, but the differences did not reach statistical significance (P = 0.062 and P = 0.091, respectively). Among the RA patients, the presence of CAD was associated with significantly higher SVR (P = 0.018). SAE and LAE were lower, but the differences did not reach statistical significance (P = 0.082 and P = 0.677, respectively). For the patients and controls combined, CAD was associated with lower SAE (P = 0.003) and LAE (P = 0.002) but not with higher SVR (P = 0.101). After adjustment for age and sex, none of these differences between CAD-positive and CAD-negative subjects remained significant (data not shown).

The mean levels of SAE, LAE, and SVR did not vary significantly with RA clinical disease activity (data not shown). The subgroup of patients and controls who were also matched for smoking status (n = 30) had similar (and statistically significant) differences in SAE, LAE, and SVR. The subgroup of patients and controls who were also matched for hypertension status (n = 32) had similar differences in SAE, LAE, and SVR, with the arterial elasticity changes reaching statistical significance (data not shown).

Correlations of vascular function with risk factors.

After adjustment for age and sex, SAE was inversely correlated with systolic BP, diastolic BP, and heart rate (P < 0.005) and with the inflammatory markers hsCRP and SAA in the combined group of RA patients and controls (Figure 3). Significant relationships were not apparent in the patients and controls when the groups were considered separately. LAE was inversely correlated with systolic BP, diastolic BP, pulse pressure, heart rate, and triglyceride levels (P < 0.05) and with the inflammatory markers hsCRP and SAA in the combined group of RA patients and controls (Figure 3). SVR was positively correlated with systolic BP, diastolic BP, heart rate, and BMI (P < 0.001) and with sVCAM-1 when the RA and control groups were combined (Figure 3). There was no significant correlation between SAE, LAE, or SVR and either total cholesterol, LDL cholesterol, or high-density lipoprotein cholesterol.

Figure 3.

Correlations between small artery elasticity, large artery elasticity, and systemic vascular resistance (SVR) and markers of inflammation in patients with rheumatoid arthritis (solid circles) and control subjects (open circles). Coefficients and P values refer to partial correlations, adjusted for age and sex. CRP = C-reactive protein (high sensitivity); SAA = serum amyloid A; sVCAM-1 = soluble vascular cell adhesion molecule 1.

Odds ratios for having RA according to incremental changes in SAE, LAE, and SVR values were estimated using conditional logistic regression (Table 4). This method models RA as a function of differences in predictor variables (e.g., arterial elasticity, inflammation) between case–control pairs. Results are expressed as the odds ratio for having RA for each incremental change in the predictor variable. Hence an odds ratio <1 indicates a lower risk of RA for a given change in the level of the predictor variable; an odds ratio of >1 indicates a higher risk of RA for a given change in the level of the predictor variable.

Table 4. Associations of RA with measures of vascular compliance in conditional logistic regression analysis*
 Model 1Model 2Model 3
  • *

    Values are the conditional odds ratio (95% confidence interval) for an increase of 1 unit in either small artery elasticity (SAE) or large artery elasticity (LAE) or an increase of 100 units in systemic vascular resistance (SVR). Odds ratios <1 indicate a lower risk of rheumatoid arthritis (RA), and odds ratios >1 indicate a higher risk of RA, for a given change in the level of SAE/LAE/SVR. Model 1 represents the odds of RA for an incremental increase in SAE, LAE, or SVR. Model 2 is the same as for model 1, but with adjustment for hypertension, smoking, and education level. Model 3 is the same as for model 2, but with further adjustment for high-sensitivity C-reactive protein level.

Small artery elasticity0.76 (0.61–0.90)0.89 (0.69–1.09)1.17 (0.83–1.73)
Large artery elasticity0.87 (0.76–0.96)0.86 (0.71–0.99)0.86 (0.65–1.07)
Systemic vascular resistance1.15 (1.01–1.34)1.07 (0.89–1.31)1.03 (0.73–1.57)

Model 1 (Table 4) included only the difference in outcome variable between RA patients and controls. CAD status is not included in the conditional logistic regression model because it is one of the matching variables and is therefore redundant in such an analysis. The interaction of vascular function with CAD status was tested (that is, Does the relationship of vascular function to RA vary according to CAD status?), as were interactions with age and sex, but these proved to be negative in each case. These interactions were therefore omitted from subsequent analyses in order to maximize the statistical power. The odds of RA were not significantly associated with differences in BMI, total cholesterol, high-density lipoprotein cholesterol, LDL cholesterol, or triglycerides; these were also excluded from the final models.

For each 1-unit increase in SAE (ml/mm Hg × 100), the odds of having RA were significantly lower (model 1). The decreased odds of RA with increasing SAE were attenuated after adjustment for the presence of hypertension, smoking status, and education level, becoming of marginal significance (model 2). Inclusion of the hsCRP caused the association of SAE with RA to approach unity and therefore become nonsignificant (model 3), which is consistent with the SAE differences between RA patients and controls being due to inflammation as reflected by the hsCRP.

For each 1-unit increase in LAE (ml/mm Hg × 10), the odds of having RA were significantly lower (model 1). The decreased odds of RA with increasing elasticity remained significant after adjustment for the presence of hypertension, smoking status, and education level (model 2). Inclusion of the hsCRP caused the association of LAE with RA to approach nonsignificance without changing its absolute value (probably due to loss of statistical power) (model 3).

For each 100-unit increase in SVR (dynes · second · cm−5), there was an increase in the risk of RA (model 1). Inclusion of hypertension, smoking, and education (model 2), as well as the hsCRP (model 3), caused the odds of RA to approach unity and thus become nonsignificant. These results are consistent with the higher SVR in RA being due to differences in smoking (which increases the CRP value), blood pressure, education, and the hsCRP. Substitution of the hsCRP with other markers of inflammation (SAA or sVCAM-1) did not substantially alter the results presented for model 3 for either the SAE, LAE, or SVR (data not shown).

DISCUSSION

This cross-sectional study shows reduced arterial elasticity in RA patients occurring independently of traditional vascular risk factors. The data support the hypothesis that inflammation is associated with reduced arterial elasticity, which may precede atherosclerosis. Values for markers of inflammation (ESR, hsCRP, SAA, and sVCAM-1) were higher in RA patients, and the hsCRP and SAA values correlated inversely with the SAE and LAE values.

SAE and LAE decline with age, and SAE is lower in groups at high risk of vascular disease (3). Reduced SAE may be predictive of future vascular events. In a prospective study of the general population conducted by Grey et al (13), a reduction in the SAE of 2 ml/mm Hg × 100 was associated with a 33% increase in the risk of cardiovascular events. Similar studies in RA are lacking. We found that SAE and LAE decreased with age in both the RA and the control groups, with the RA-related reduction in SAE being greater in younger patients, which confirms previous findings that age is associated with altered vascular elasticity (Figure 1). The age-related decline in SAE in RA patients was not as great as that in the non-RA subjects, with the two lines converging. This result may be the result of survival bias due to premature death of older RA patients. Previous studies have shown that RA is associated with a mean reduction in life expectancy of as many as 17 years (2).

For RA patients and control subjects, CAD was associated with reduced arterial elasticity, but after adjustment for age and sex, none of these differences between CAD-positive and CAD-negative subjects remained significant. The age of the CAD-positive subjects within both the RA and the control groups was higher than the age of the CAD-negative subjects. The prevalence of CAD increases with age, so it is not surprising that the CAD-positive patients were older on average than the CAD-negative patients. The primary aim of this study was to examine vascular function in RA patients compared with controls and to statistically determine which factors are contributory. Examination of the effects of CAD was a secondary consideration, and the study design was not specifically powered to do this. Furthermore, medications taken by the CAD-positive patients, such as ACE inhibitors and hydroxymethylglutaryl-coenzyme A reductase inhibitors (statins), improve vascular function. Statins have been shown to improve SAE by up to 21%, and Resnick et al (14) found that in the general population, treatment of hypertension with ACE inhibitors and calcium-channel blockers can improve SAE by up to 30%. These favorable effects on vascular function may minimize the observed differences in arterial elasticity between CAD-positive and CAD-negative subjects.

RA disease activity and inflammation have been associated with progression of cardiovascular disease and increased mortality rates in previous studies (1). An important observation is the discordance between high levels of hsCRP and low values for clinical markers (swollen joint count, modified Health Assessment Questionnaire, and patient's assessments of pain and well-being) and for traditional biochemical markers (ESR and CRP) of disease activity (8). Unlike the first-generation CRP assay (which was not suitable for predicting CAD), the high-sensitivity assay has not been assessed as a disease activity marker in RA and may be more sensitive to subclinical disease. In our study, 53% of the RA patients had CRP levels that were within the normal range when measured with the traditional assay, but only 6% had CRP levels within the normal range when measured with the high-sensitivity assay. This finding is potentially important, given that joint erosions (assessed by magnetic resonance imaging) often progress in RA despite clinical improvements in synovitis (15). This result suggests that subclinical inflammation may be important not only for joint damage in RA, but also for vascular damage.

In some studies of vascular events in RA, traditional risk factors have been associated with cardiovascular disease (1). The RA patients in the present study were more likely to have a history (current or past) of smoking and hypertension. Controlling statistically for smoking status and performing a subgroup analysis of subjects who had never smoked verified the worse vascular health in the RA patients compared with the controls. In other studies, coexistent hypertension has been noted in RA patients as a result of the use of nonsteroidal antiinflammatory drugs (NSAIDs) (16). However, in the present study, both the RA and the control groups had similar BP values, which is likely due to the effective use of antihypertensive medications in RA (Table 1). This difference in risk factor status did not account for the observed differences in SAE and LAE (Table 4), which was also demonstrated in a subanalysis of RA patients and controls matched for the presence of hypertension.

High cholesterol levels have been associated with reduced arterial elasticity, but in our study, lipid profiles did not differ between groups and were not associated with arterial elasticity or SVR (17). Studies evaluating serum lipid concentrations in RA patients have shown conflicting results (18, 19). Small dense LDL is an independent vascular risk factor and is associated with insulin resistance and, in non-RA subjects, has been associated with abnormal flow-mediated vasodilation, another noninvasive test of vascular function (20). In our study, the mean LDL size was smaller in the CAD-positive than in the CAD-negative subjects in both study groups, but LDL size and vascular health were not significantly correlated. The RA patients had larger LDL particles than did the controls. So, altered LDL size does not account for the poorer vascular function. Our LDL size results contrast with those of a previous study, where RA patients had elevated levels of small dense LDL (21). Obesity was not associated with the observed abnormalities in arterial elasticity. Thus, consistent with other studies, we did not find differences in traditional risk factors that would account for the worse vascular health observed in the RA patients (22).

Atherosclerosis is an inflammatory disease (6, 7). The hsCRP, SAA, and adhesion molecules are acute-phase proteins that have been reported to be associated with, and predictive of, CAD in the general population (23–25). The hsCRP usually correlates with other markers of inflammation (23). Our study supports these findings, with significant inverse correlations between hsCRP and SAA values and vascular elasticity. These correlations were significant in analyses of the study group as a whole, but not within the RA group alone (Figure 3). The range of values within the RA patient and control groups considered in isolation may not be great enough to expose a relationship with arterial elasticity, particularly in the case of the hsCRP (Figure 3). Combining the groups provides a wider range of values across which such associations can be examined. This does not allow definitive conclusions, however, since other unidentified factors associated with RA may be contributing to the impaired elasticity.

CRP has been localized in vessel walls; it can bind neutrophils, interact with adhesion molecules, activate complement, and enhance tissue factor expression, acting as a possible mediator, as well as a marker, of vascular disease (26). CRP levels measured by the high-sensitivity assay were elevated in RA patients compared with the control subjects, and in non-RA subjects with CAD versus healthy controls. In the general population, for each quintile increase in hsCRP, the adjusted relative risk of future cardiovascular events is increased by 25% in men and by 33% in women (9). In our study, RA patients with CAD had lower levels of hsCRP than did RA patients without CAD. This may reflect small numbers of subjects or the higher frequency of treatment with drugs that have an antiinflammatory effect, such as statins, aspirin, NSAIDs, cyclooxygenase 2 inhibitors, and prednisolone. Statin therapy and full-dose aspirin (300 mg/day) have antiinflammatory effects, can lower hsCRP levels, and can reduce coronary events (27, 28). In regression analysis, the observed difference in SAE between RA patients and controls was accounted for by differences in hsCRP values, suggesting that inflammation contributes to the worse vascular health observed in RA.

In this study, circulating levels of sVCAM-1 were significantly elevated in RA patients compared with controls. However, sVCAM-1 levels correlated significantly only with SVR, and not with the other measures of arterial elasticity. Some, but not all, studies have shown an association between sVCAM-1 levels and vascular disease, and previous studies in RA are lacking (29). In our model (Table 4), use of other measures of inflammation (SAA or adhesion molecule levels) did not provide any further information than that given by the hsCRP, an assay that, unlike SAA and adhesion molecule assays, is already available for clinical use.

In conclusion, this cross-sectional study showed reduced arterial elasticity in RA by pulse wave analysis, independently of traditional vascular risk factors. The data support the hypothesis that inflammation is associated with reduced arterial elasticity, which may precede atherosclerosis. There was a trend toward reduced arterial elasticity in RA patients with CAD compared with RA patients without CAD, although the difference was not significant. However, this study was not powered to separate the effects of RA alone from the effects of CAD within RA. The question remains as to whether there is a fixed effect of RA on vascular function, or whether inflammation alone can explain the reduced vascular function in RA. Further prospective and intervention studies with tighter control of inflammation are required in order to answer this question and to assess whether vascular function improves after inflammation is reduced. The early detection of vascular damage and the identification and treatment of inflammatory risk factors may reduce the personal and economic burden of increased vascular disease in RA and other connective tissue disorders. Such tools, including pulse wave analysis and hsCRP, are readily applicable in the clinical setting.

Acknowledgements

The authors would like to thank Dr. M. Cincotta (Cardiac Investigations Unit, St. Vincent's Hospital, Fitzroy, Victoria, Australia) for her assistance with data management.

Ancillary