High-sensitivity cardiac troponin I is a biomarker for occult coronary plaque burden and cardiovascular events in patients with rheumatoid arthritis

High-sensitivity cardiac troponin I is a biomarker for occult coronary plaque burden and... Abstract Objectives Patients with RA display greater occult coronary atherosclerosis burden and experience higher cardiovascular morbidity and mortality compared with controls. We here explored whether pro-inflammatory cytokines and high-sensitivity cardiac troponin I (hs-cTnI), a biomarker of myocardial injury, correlated with plaque burden and cardiovascular events (CVEs) in RA. Methods We evaluated 150 patients with 64-slice coronary CT angiography. Coronary artery calcium, number of segments with plaque (segment involvement score), stenotic severity and plaque burden were assessed. Lesions were described as non-calcified, mixed or fully calcified. Blood levels of hs-cTnI and pro-inflammatory cytokines were assessed during coronary CT angiography. Subjects were followed over 60 (s.d. 26) months for both ischaemic [cardiac death, non-fatal myocardial infarction (MI), stroke, peripheral arterial ischaemia] and non-ischaemic (new-onset heart failure hospitalization) CVEs. Results Plasma hs-cTnI correlated with all coronary plaque outcomes (P < 0.01). Elevated hs-cTnI (⩾1.5 pg/ml) further associated with significant calcification, extensive atherosclerosis, obstructive plaque and any advanced mixed or calcified plaques after adjustments for cardiac risk factors or Framingham D’Agostino scores (all P < 0.05). Eleven patients suffered a CVE (1.54/100 patient-years), eight ischaemic and three non-ischaemic. Elevated hs-cTnI predicted all CVE risk independent of demographics, cardiac risk factors and prednisone use (P = 0.03). Conversely, low hs-cTnI presaged a lower risk for both extensive atherosclerosis (P < 0.05) and incident CVEs (P = 0.037). Conclusion Plasma hs-cTnI independently associated with occult coronary plaque burden, composition and long-term incident CVEs in patients with RA. Low hs-cTnI forecasted a lower risk for both extensive atherosclerosis as well as CVEs. hs-cTnI may therefore optimize cardiovascular risk stratification in RA. cardiovascular events, high-sensitivity cardiac troponin I, occult coronary atherosclerosis, rheumatoid arthritis Rheumatology key messages High-sensitivity cardiac troponin I correlates with the presence, burden and composition of occult coronary plaque in RA. High-sensitivity cardiac troponin I further correlates with long-term cardiac events in RA after adjustment for cardiac risk factors. High-sensitivity cardiac troponin I may serve as a predictive biomarker in refining cardiovascular risk assessment in RA. Introduction Individuals with RA experience a higher rate of cardiovascular events (CVEs) compared with controls [1]. This may be explained by the greater prevalence, severity, burden and different composition of occult coronary lesions in RA compared with age- and gender-matched controls [2]. Residual disease activity may further associate with more advanced, complex and prone-to-rupture coronary plaques [2]. Pro-inflammatory cytokines such as TNF-α, IL-6 and IL-17 reflect clinical activity and structural damage in RA and are elevated in the blood of RA patients compared with controls [3]; the same cytokines have been identified in atherosclerotic plaque [4–6] and correlated with subclinical atherosclerosis independent of cardiac risk factors [7], coronary plaque complexity [8], plaque destabilization and CVEs in subjects without autoimmune disease [9–11]. Nevertheless, the relationship between these cytokines and occult coronary plaque burden and composition in RA are unknown. Higher plaque load or vulnerability may be further reflected in elevations of biomarkers specific for myocardial injury [12]; indeed, cardiac troponin elevations measured by high-sensitivity assays—and below thresholds used to diagnose acute coronary syndromes—were associated with higher coronary artery calcium (CAC) scores in a population-based study [12]. Additionally, both high-sensitivity cardiac troponin T (hs-cTnT) and high-sensitivity cardiac troponin I (hs-cTnI) predicted a greater risk of fatal and non-fatal coronary heart disease, heart failure hospitalization and overall mortality in the general population [12–15]. In a recent report, hs-cTnI was higher in RA patients compared with controls, independent of cardiovascular risk factors and inflammation [16]. Nevertheless, its association with subclinical coronary artery disease (CAD) burden and its ability to predict future CVEs in RA are unknown. We here hypothesized that hs-cTnI and various pro-inflammatory cytokines may correlate with the presence, burden and composition of occult coronary plaque in patients with RA evaluated with coronary CT angiography (CCTA). We further postulated that hs-cTnI at the time of CCTA might predict incident CVEs on long-term follow-up [60 months (s.d. 26)]. Methods Patient recruitment A total of 150 RA patients from a single centre were enrolled and prospectively evaluated on a first-come, first-served basis between 1 March 2010 and 1 March 2011 [2]. The study was approved by the local institutional review board [John F Wolf, MD, Human Subjects Committees (1 + 2)], all subjects signed informed consent and the research was carried out in compliance with the Helsinki Declaration. Inclusion criteria comprised age ⩾18 years, fulfilment of 2010 classification criteria for RA and no symptoms or history of cardiovascular disease, including myocardial infarction (MI), revascularization, heart failure, transient ischaemic attack, stroke or peripheral arterial disease (PAD). Patients with concomitant autoimmune syndromes, malignancy within <5 years, chronic or active infection, weight >325 pounds (147.7 kg), hypotension [systolic blood pressure (SBP) <90 mmHg or diastolic blood pressure (DBP) <60 mmHg] or hypertension (SBP >170 mmHg or DBP >110 mmHg), uncontrolled tachycardia, irregular rhythm, iodine allergy or glomerular filtration rate <60 ml/min were excluded. Hypertension was defined as SBP ⩾140 mmHg or DBP ⩾90 mmHg or antihypertensive use. Diabetes mellitus encompassed haemoglobin A1c >6.5% or hypoglycaemic medication use. Hyperlipidaemia constituted fasting cholesterol >200 mg/dl or low-density lipoprotein cholesterol >130 mg/dl or current statin use. Current smoking entailed cigarette consumption within 30 days from screening. Positive family history was defined as CAD in first-degree relatives <55 years of age for males or <65 years of age for females. The Framingham 2008 D’Agostino modified general cardiovascular risk (FRS-DA) score was calculated for all study participants [17]. Disease duration, serologic status, radiographs and treatments were captured. RA activity was evaluated by a 28-joint DAS with CRP (DAS28-CRP). Laboratory evaluations Blood for regular chemistries, fasting lipids, ESR and high-sensitivity CRP was collected on the day of CCTA and evaluated at the local laboratory. Additionally, blood was collected in EDTA tubes and immediately processed and plasma was frozen at −80°C until it was assayed. hs-cTnI was measured at Singulex (Alameda, CA, USA) by technicians blinded to the clinical data using a microparticle immunoassay and single-molecule counting [18]. TNF-α, IL-6, IL-17A and F and VEGF were also assessed using laboratory tests developed at Singulex based on single-molecule counting [18]. Multidetector CTA Scans were performed with a 64-multidetector row Lightspeed VCT scanner (GE Healthcare, Chicago, IL, USA) between March 2010 and March 2011 and the images were analysed as previously described by a single, blinded interpreter [19]. CAC was quantified by the Agatston method [20]. Coronary arteries were evaluated on contrast-enhanced scans using a standardized 15-segment model [21]. Stenosis severity was scored from 0 to 4 based on the grade of luminal restriction: 1 represented 1–29% stenosis; 2, 30–49%; 3, 50–69% and 4, ⩾70%. The area of each plaque visualized in at least two adjacent slices (slice thickness 0.625 mm) was determined on all affected slices. Plaque burden was graded from 0 to 3, defined as none (0), mild (1), moderate (2) and severe (3), based on the number of adjacent slices containing plaque. Lesions rendering >50% stenosis were considered obstructive. Plaque composition was defined as non-calcified, mixed (MP) or calcified (CP), as discussed elsewhere [22]. Subjects received three individual quantitative scores [22]: the segment involvement score (SIS) represented the total number of segments with plaque (0–15); the stenosis severity score (SSS) reported the cumulative stenosis grade conferred by plaque over all evaluable segments (0–60) and the plaque burden score (PBS) described the cumulative plaque size over all evaluable segments (0–45). Reproducibility of scoring measurements for our centre have been previously reported [22]. Incident CVEs All patients were followed for incident CVEs over a period of 60 (s.d. 26) months. These included both ischaemic (cardiac death, non-fatal MI, stroke, transient ischaemic attack, PAD) as well as non-ischaemic ones (new-onset heart failure hospitalization). Event adjudication was elaborated by the treating cardiologist, neurologist or vascular surgeon and based on standard definitions [23–25]. Analysis Continuous variables are expressed as medians with interquartile ranges (IQRs) and categorical variables are expressed as numbers with percentages. Spearman’s rho correlation coefficients evaluated preliminary associations between biomarkers and plaque outcomes. Medians between CVE groups were compared using the Mann–Whitney U test and counts by the chi-square test. Further analyses were restricted to biomarkers with significant differences. Plaque outcomes were binarized based on median; those were >0 vs 0 for CAC, >1 vs ⩽1 for SIS, >1 vs ⩽1 for SSS and >2 vs ⩽2 for PBS. To evaluate plaque composition, the presence of MP or CP vs the absence of both was used as an outcome. Similarly, hs-cTnI was binarized as high (>1.5 pg/ml) vs low (⩽1.5 pg/ml). Logistic regression models were constructed to evaluate associations between hs-cTnI and individual plaque parameters; models were adjusted for either age and gender (model 1) or additionally for hypertension, diabetes, hyperlipidaemia, smoking, BMI and prednisone use (model 2) or for the patients’ FRS-DA score (model 3). Similar logistic regression models were devised to predict individual or composite high-risk plaque outcomes; individual ones included CAC >100 vs ⩽100, SIS >5 vs ⩽5, SSS >5 vs ⩽5 and the presence of obstructive plaque vs not. Composite outcomes entailed the presence of SSS >5 or CAC >100 vs neither and SIS >5 or SSS >5 or obstructive plaque vs none. Results are reported as odds ratios with 95% CIs. For composite and plaque composition outcomes, sensitivity, specificity and negative and positive predictive values were determined with standard formulas. Cox proportional hazards regression analysis evaluated CVE risk [hazard ratio (HR)] associated with high cTnI (>1.5 pg/ml) in raw and several adjusted models (models 1, 2 and 3) as previously described. Kaplan–Meier curves were compared by the log-rank method. The diagnostic accuracy of CVE risk assessment for FRS-DA alone, FRS-DA + hs-cTnI, and FRS-DA + hs-cTnI + high-risk composite outcome was evaluated as the area under the receiver operator characteristic curve and compared with the DeLong method. Integrated discrimination improvement (IDI) between these constructs was computed and improvement in prediction accuracy was evaluated, where P-values <0.05 were significant. Data were analysed with SAS version 9.4 (SAS Institute, Cary, NC, USA) or R version 3.4.1 (R Project for Statistical Computing, Vienna, Austria). Results Patient demographics are shown in Table 1. Subjects were predominantly female, with established seropositive, erosive and well-controlled disease. RA parameters and traditional cardiac risk factors were not significantly different in patients incurring CVEs. In contrast, FRS-DA, coronary atherosclerosis burden, including high-risk plaque parameters, and levels of hs-cTnI were significantly higher (all P < 0.05). Table 1 Baseline patient characteristics Characteristics  All (n = 150)  No CVE (n = 139)  With CVE (n = 11)  Age, years  54 (46–60)  54 (45–60)  59 (53–70)†  Males, n (%)  18 (13)  16 (12)  2 (18)  RF positive, n (%)  129 (86)  120 (86)  9 (82)  ACPA positive, n (%)  127 (85)  118 (85)  9 (82)  X-ray erosions, n (%)  99 (66)  90 (65)  9 (82)  RA duration, years  9 (5–14)  9 (4–14)  12 (9–18)  Tender joint count  0 (0–2)  0 (0–2)  0 (0–1)  Swollen joint count  0.5 (0–3)  0 (0–3)  1 (0–5)  DAS28-CRP  2.30 (1.8–3.3)  2.32 (1.76–3.29)  2.39 (1.59–2.93)  High-sensitivity CRP, mg/dl  0.41 (0.20–0.96)  0.42 (0.21–0.88)  0.31 (0.12–0.76)  Prednisone, n (%)  52 (35)  48 (35)  4 (36)  Concurrent DMARDs  2 (1–3)  2 (1–3)  2 (1–3)  TNF inhibitor exposed, n (%)  90 (60)  84 (60)  6 (55)  TNF inhibitor duration, years  4.4 (2.4–6.0)  4.0 (2.2–6.0)  6.0 (4.2–6.9)  Diabetes mellitus, n (%)  26 (18)  22 (16)  4 (36)†  Hypertension, n (%)  64 (44)  58 (43)  8 (73)†  Current smoker, n (%)  13 (9)  12 (9)  1 (9)  Family history of CAD, n (%)  6 (4)  5 (3.6)  1 (9)  Hyperlipidaemia, n (%)  26 (17)  25 (18)  1 (9)  BMI, kg/m2  28.1 (25.8–32.6)  28.4 (26–32.8)  25.9 (23–30.5)†  D’Agostino Framingham score  6.4 (3.0–11.7)  6.2 (2.7–11.2)  15.6 (4.9–20.0)*  Coronary artery calcium  0 (0–19)  0 (0–10)  120 (0–361)***  Segment involvement score  1 (0–3)  1 (0–2)  5 (1–7)**  Segment stenosis score  1 (0–4)  1 (0–3)  9 (1–14)**  Plaque burden score  1.5 (0–3)  1 (0–3)  9 (1–12)**  Obstructive plaque (>50%), n (%)  18 (12)  12 (9)  6 (55)****  Non-calcified plaque score  1 (0–2)  1 (0–2)  1 (0–5)  Mixed plaque score  0 (0–1)  0 (0–0)  3 (0–7)*  Calcified plaque score  0 (0–0)  0 (0–0)  1 (0–3)  hs-cTnI, pg/mla  1.5 (1.1–2.6)  1.5 (1–2.4)  2.6 (2.1–4.4)**  IL-17A, pg/mla  1.3 (0.8–1.8)  1.3 (0.8–1.9)  1.2 (0.8–1.5)  IL-17F, pg/mla  35.8 (22.1–66.8)  36.3 (22.1–62)  31 (22.1–97.4)  IL-6, pg/mla  2.9 (1.8–5.8)  2.8 (1.7–5.4)  3.7 (2.6–13.4) †  TNF-α, pg/mla  11.2 (7.8–24)  11.3 (7.7–24)  11.1 (7.9–30.4)  VEGF, pg/mla  134 (61–221)  134 (60–221)  134 (77–179)  Characteristics  All (n = 150)  No CVE (n = 139)  With CVE (n = 11)  Age, years  54 (46–60)  54 (45–60)  59 (53–70)†  Males, n (%)  18 (13)  16 (12)  2 (18)  RF positive, n (%)  129 (86)  120 (86)  9 (82)  ACPA positive, n (%)  127 (85)  118 (85)  9 (82)  X-ray erosions, n (%)  99 (66)  90 (65)  9 (82)  RA duration, years  9 (5–14)  9 (4–14)  12 (9–18)  Tender joint count  0 (0–2)  0 (0–2)  0 (0–1)  Swollen joint count  0.5 (0–3)  0 (0–3)  1 (0–5)  DAS28-CRP  2.30 (1.8–3.3)  2.32 (1.76–3.29)  2.39 (1.59–2.93)  High-sensitivity CRP, mg/dl  0.41 (0.20–0.96)  0.42 (0.21–0.88)  0.31 (0.12–0.76)  Prednisone, n (%)  52 (35)  48 (35)  4 (36)  Concurrent DMARDs  2 (1–3)  2 (1–3)  2 (1–3)  TNF inhibitor exposed, n (%)  90 (60)  84 (60)  6 (55)  TNF inhibitor duration, years  4.4 (2.4–6.0)  4.0 (2.2–6.0)  6.0 (4.2–6.9)  Diabetes mellitus, n (%)  26 (18)  22 (16)  4 (36)†  Hypertension, n (%)  64 (44)  58 (43)  8 (73)†  Current smoker, n (%)  13 (9)  12 (9)  1 (9)  Family history of CAD, n (%)  6 (4)  5 (3.6)  1 (9)  Hyperlipidaemia, n (%)  26 (17)  25 (18)  1 (9)  BMI, kg/m2  28.1 (25.8–32.6)  28.4 (26–32.8)  25.9 (23–30.5)†  D’Agostino Framingham score  6.4 (3.0–11.7)  6.2 (2.7–11.2)  15.6 (4.9–20.0)*  Coronary artery calcium  0 (0–19)  0 (0–10)  120 (0–361)***  Segment involvement score  1 (0–3)  1 (0–2)  5 (1–7)**  Segment stenosis score  1 (0–4)  1 (0–3)  9 (1–14)**  Plaque burden score  1.5 (0–3)  1 (0–3)  9 (1–12)**  Obstructive plaque (>50%), n (%)  18 (12)  12 (9)  6 (55)****  Non-calcified plaque score  1 (0–2)  1 (0–2)  1 (0–5)  Mixed plaque score  0 (0–1)  0 (0–0)  3 (0–7)*  Calcified plaque score  0 (0–0)  0 (0–0)  1 (0–3)  hs-cTnI, pg/mla  1.5 (1.1–2.6)  1.5 (1–2.4)  2.6 (2.1–4.4)**  IL-17A, pg/mla  1.3 (0.8–1.8)  1.3 (0.8–1.9)  1.2 (0.8–1.5)  IL-17F, pg/mla  35.8 (22.1–66.8)  36.3 (22.1–62)  31 (22.1–97.4)  IL-6, pg/mla  2.9 (1.8–5.8)  2.8 (1.7–5.4)  3.7 (2.6–13.4) †  TNF-α, pg/mla  11.2 (7.8–24)  11.3 (7.7–24)  11.1 (7.9–30.4)  VEGF, pg/mla  134 (61–221)  134 (60–221)  134 (77–179)  Values are presented as median (IQR) unless stated otherwise. † P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. a Biomarker data were available in 146 patients. Table 1 Baseline patient characteristics Characteristics  All (n = 150)  No CVE (n = 139)  With CVE (n = 11)  Age, years  54 (46–60)  54 (45–60)  59 (53–70)†  Males, n (%)  18 (13)  16 (12)  2 (18)  RF positive, n (%)  129 (86)  120 (86)  9 (82)  ACPA positive, n (%)  127 (85)  118 (85)  9 (82)  X-ray erosions, n (%)  99 (66)  90 (65)  9 (82)  RA duration, years  9 (5–14)  9 (4–14)  12 (9–18)  Tender joint count  0 (0–2)  0 (0–2)  0 (0–1)  Swollen joint count  0.5 (0–3)  0 (0–3)  1 (0–5)  DAS28-CRP  2.30 (1.8–3.3)  2.32 (1.76–3.29)  2.39 (1.59–2.93)  High-sensitivity CRP, mg/dl  0.41 (0.20–0.96)  0.42 (0.21–0.88)  0.31 (0.12–0.76)  Prednisone, n (%)  52 (35)  48 (35)  4 (36)  Concurrent DMARDs  2 (1–3)  2 (1–3)  2 (1–3)  TNF inhibitor exposed, n (%)  90 (60)  84 (60)  6 (55)  TNF inhibitor duration, years  4.4 (2.4–6.0)  4.0 (2.2–6.0)  6.0 (4.2–6.9)  Diabetes mellitus, n (%)  26 (18)  22 (16)  4 (36)†  Hypertension, n (%)  64 (44)  58 (43)  8 (73)†  Current smoker, n (%)  13 (9)  12 (9)  1 (9)  Family history of CAD, n (%)  6 (4)  5 (3.6)  1 (9)  Hyperlipidaemia, n (%)  26 (17)  25 (18)  1 (9)  BMI, kg/m2  28.1 (25.8–32.6)  28.4 (26–32.8)  25.9 (23–30.5)†  D’Agostino Framingham score  6.4 (3.0–11.7)  6.2 (2.7–11.2)  15.6 (4.9–20.0)*  Coronary artery calcium  0 (0–19)  0 (0–10)  120 (0–361)***  Segment involvement score  1 (0–3)  1 (0–2)  5 (1–7)**  Segment stenosis score  1 (0–4)  1 (0–3)  9 (1–14)**  Plaque burden score  1.5 (0–3)  1 (0–3)  9 (1–12)**  Obstructive plaque (>50%), n (%)  18 (12)  12 (9)  6 (55)****  Non-calcified plaque score  1 (0–2)  1 (0–2)  1 (0–5)  Mixed plaque score  0 (0–1)  0 (0–0)  3 (0–7)*  Calcified plaque score  0 (0–0)  0 (0–0)  1 (0–3)  hs-cTnI, pg/mla  1.5 (1.1–2.6)  1.5 (1–2.4)  2.6 (2.1–4.4)**  IL-17A, pg/mla  1.3 (0.8–1.8)  1.3 (0.8–1.9)  1.2 (0.8–1.5)  IL-17F, pg/mla  35.8 (22.1–66.8)  36.3 (22.1–62)  31 (22.1–97.4)  IL-6, pg/mla  2.9 (1.8–5.8)  2.8 (1.7–5.4)  3.7 (2.6–13.4) †  TNF-α, pg/mla  11.2 (7.8–24)  11.3 (7.7–24)  11.1 (7.9–30.4)  VEGF, pg/mla  134 (61–221)  134 (60–221)  134 (77–179)  Characteristics  All (n = 150)  No CVE (n = 139)  With CVE (n = 11)  Age, years  54 (46–60)  54 (45–60)  59 (53–70)†  Males, n (%)  18 (13)  16 (12)  2 (18)  RF positive, n (%)  129 (86)  120 (86)  9 (82)  ACPA positive, n (%)  127 (85)  118 (85)  9 (82)  X-ray erosions, n (%)  99 (66)  90 (65)  9 (82)  RA duration, years  9 (5–14)  9 (4–14)  12 (9–18)  Tender joint count  0 (0–2)  0 (0–2)  0 (0–1)  Swollen joint count  0.5 (0–3)  0 (0–3)  1 (0–5)  DAS28-CRP  2.30 (1.8–3.3)  2.32 (1.76–3.29)  2.39 (1.59–2.93)  High-sensitivity CRP, mg/dl  0.41 (0.20–0.96)  0.42 (0.21–0.88)  0.31 (0.12–0.76)  Prednisone, n (%)  52 (35)  48 (35)  4 (36)  Concurrent DMARDs  2 (1–3)  2 (1–3)  2 (1–3)  TNF inhibitor exposed, n (%)  90 (60)  84 (60)  6 (55)  TNF inhibitor duration, years  4.4 (2.4–6.0)  4.0 (2.2–6.0)  6.0 (4.2–6.9)  Diabetes mellitus, n (%)  26 (18)  22 (16)  4 (36)†  Hypertension, n (%)  64 (44)  58 (43)  8 (73)†  Current smoker, n (%)  13 (9)  12 (9)  1 (9)  Family history of CAD, n (%)  6 (4)  5 (3.6)  1 (9)  Hyperlipidaemia, n (%)  26 (17)  25 (18)  1 (9)  BMI, kg/m2  28.1 (25.8–32.6)  28.4 (26–32.8)  25.9 (23–30.5)†  D’Agostino Framingham score  6.4 (3.0–11.7)  6.2 (2.7–11.2)  15.6 (4.9–20.0)*  Coronary artery calcium  0 (0–19)  0 (0–10)  120 (0–361)***  Segment involvement score  1 (0–3)  1 (0–2)  5 (1–7)**  Segment stenosis score  1 (0–4)  1 (0–3)  9 (1–14)**  Plaque burden score  1.5 (0–3)  1 (0–3)  9 (1–12)**  Obstructive plaque (>50%), n (%)  18 (12)  12 (9)  6 (55)****  Non-calcified plaque score  1 (0–2)  1 (0–2)  1 (0–5)  Mixed plaque score  0 (0–1)  0 (0–0)  3 (0–7)*  Calcified plaque score  0 (0–0)  0 (0–0)  1 (0–3)  hs-cTnI, pg/mla  1.5 (1.1–2.6)  1.5 (1–2.4)  2.6 (2.1–4.4)**  IL-17A, pg/mla  1.3 (0.8–1.8)  1.3 (0.8–1.9)  1.2 (0.8–1.5)  IL-17F, pg/mla  35.8 (22.1–66.8)  36.3 (22.1–62)  31 (22.1–97.4)  IL-6, pg/mla  2.9 (1.8–5.8)  2.8 (1.7–5.4)  3.7 (2.6–13.4) †  TNF-α, pg/mla  11.2 (7.8–24)  11.3 (7.7–24)  11.1 (7.9–30.4)  VEGF, pg/mla  134 (61–221)  134 (60–221)  134 (77–179)  Values are presented as median (IQR) unless stated otherwise. † P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. a Biomarker data were available in 146 patients. Correlations of cytokines and hs-cTnI with occult coronary atherosclerosis TNF-α, IL-6, IL-17A, IL-17F and VEGF showed no correlations with any coronary plaque parameters or hs-cTnI (supplementary Table S1, available at Rheumatology online) and neither did ESR, CRP, tender joint count, swollen joint count or DAS28-CRP (data not shown). hs-cTnI was detectable in all patients [1.5 (IQR 1.1–2.6) pg/ml]; patients with any plaque had higher levels compared to those without [1.8 (IQR 1.1–2.6) pg/ml vs 1.3 (0.9–1.8); P = 0.02]. Moreover, hs-cTnI was correlated with all occult coronary plaque outcomes (all P < 0.01). CAC, SIS, SSS and PBS substantially increased from the lowest to the highest hs-cTnI tertile (P for trend = 0.006, 0.005, 0.01 and 0.009, respectively). Similarly, high-risk plaque outcomes such as CAC >100, SIS >5, SSS >5, obstructive plaque, composite outcome and presence of any advanced MP/CP lesions were considerably enriched across higher hs-cTnI tertiles (Fig. 1). Fig. 1 View largeDownload slide Several high-risk coronary plaque burden outcomes are significantly enriched across higher tertiles of hs-cTnI hs-cTnI tertile ranges were ≤1.2, 1.2–2.1 and ≥2.1 pg/ml. P-values for trend determined by the Jonckheere–Terpstra test. †P < 0.1, *P < 0.05, **P < 0.01, ***P < 0.001. Fig. 1 View largeDownload slide Several high-risk coronary plaque burden outcomes are significantly enriched across higher tertiles of hs-cTnI hs-cTnI tertile ranges were ≤1.2, 1.2–2.1 and ≥2.1 pg/ml. P-values for trend determined by the Jonckheere–Terpstra test. †P < 0.1, *P < 0.05, **P < 0.01, ***P < 0.001. hs-cTnI independently correlates with plaque burden and composition hs-cTnI levels associated with all plaque outcomes (Table 2). After controlling for age and gender (model 1), additional adjustments for hypertension, diabetes, hyperlipidaemia, smoking, BMI and prednisone use (model 2) or FRS-DA score (model 3), hs-cTnI remained predictive of CAC, SSS and PBS. Furthermore, it associated with the presence of more advanced mixed or calcified plaques but showed no correlation with earlier, non-calcified lesions. Importantly, hs-cTnI further correlated with high-risk outcomes such as obstructive plaque, SSS >5, CAC ⩾100 or composite endpoints (Table 2); significance persisted for several, even after adjustments for cardiac risk factors or FRS-DA scores. Table 2 Prediction of occult coronary plaque burden and composition by cTnIa   Unadjusted, OR (95% CI)  Model 1a, OR (95% CI)  Model 2b, OR (95% CI)  Model 3c, OR (95% CI)  CACd  2.95 (1.5, 6.0)**  2.2 (1.0, 4.7)*  2.3 (1.0, 5.3)*  2.7 (1.3, 5.8)*  CAC >100 vs ≤100  4.2 (1.3, 13.5)*  3.0 (0.9, 10.5)†  6.6 (1.4, 30.0)*  5.0 (1.3, 19)*  SISd  2.2 (1.1, 4.2)*  1.7 (0.9, 3.4)  1.7 (0.8, 3.6)  1.9 (1.0, 3.8)†  SIS >5 vs ≤5  2.5 (0.7, 8.5)  1.8 (0.5, 6.6)  2.5 (0.6, 10.5)  2.6 (0.7, 9.9)  SSSd  2.4 (1.3, 4.7)**  1.9 (1.0, 3.9)†  2.0 (1.0, 4.2)†  2.2 (1.1, 4.3)*  SSS >5 vs ≤5  3.0 (1.1, 8.2)*  2.4 (0.8, 7.3)  2.9 (0.9, 9.0)†  3.0 (1.0, 9.0)*  PBSd  3.1 (1.5, 6.3)**  2.4 (1.1, 5.1)*  2.7 (1.2, 6.2)*  2.9 (1.3, 6.2)**  Obstructive plaque  3.9 (1.2, 12.5)*  3.1 (0.9, 10.7)†  3.8 (1.0, 14.4)*  4.0 (1.1, 14.1)*  CAC >100 or SSS >5  4.7 (1.8, 12.4)**  2.3 (0.7, 6.9)  5.3 (1.7, 17.0)**  5.2 (1.8, 15.8)**  SIS >5 or SSS >5 or obstructive plaque  3.2 (1.2, 8.8)*  2.6 (0.9, 7.7)†  3.0 (1.0, 9.4)†  3.3 (1.1, 9.7)*  NCP >0 vs 0  1.4 (0.7, 2.7)  1.3 (0.6, 2.5)  1.3 (0.6, 2.6)  1.4 (0.7, 2.7)  MP/CP >0 vs 0  3.6 (1.8, 7.5)***  2.7 (1.2, 6.0)*  2.9 (1.2, 6.6)*  3.5 (1.6, 7.6)**    Unadjusted, OR (95% CI)  Model 1a, OR (95% CI)  Model 2b, OR (95% CI)  Model 3c, OR (95% CI)  CACd  2.95 (1.5, 6.0)**  2.2 (1.0, 4.7)*  2.3 (1.0, 5.3)*  2.7 (1.3, 5.8)*  CAC >100 vs ≤100  4.2 (1.3, 13.5)*  3.0 (0.9, 10.5)†  6.6 (1.4, 30.0)*  5.0 (1.3, 19)*  SISd  2.2 (1.1, 4.2)*  1.7 (0.9, 3.4)  1.7 (0.8, 3.6)  1.9 (1.0, 3.8)†  SIS >5 vs ≤5  2.5 (0.7, 8.5)  1.8 (0.5, 6.6)  2.5 (0.6, 10.5)  2.6 (0.7, 9.9)  SSSd  2.4 (1.3, 4.7)**  1.9 (1.0, 3.9)†  2.0 (1.0, 4.2)†  2.2 (1.1, 4.3)*  SSS >5 vs ≤5  3.0 (1.1, 8.2)*  2.4 (0.8, 7.3)  2.9 (0.9, 9.0)†  3.0 (1.0, 9.0)*  PBSd  3.1 (1.5, 6.3)**  2.4 (1.1, 5.1)*  2.7 (1.2, 6.2)*  2.9 (1.3, 6.2)**  Obstructive plaque  3.9 (1.2, 12.5)*  3.1 (0.9, 10.7)†  3.8 (1.0, 14.4)*  4.0 (1.1, 14.1)*  CAC >100 or SSS >5  4.7 (1.8, 12.4)**  2.3 (0.7, 6.9)  5.3 (1.7, 17.0)**  5.2 (1.8, 15.8)**  SIS >5 or SSS >5 or obstructive plaque  3.2 (1.2, 8.8)*  2.6 (0.9, 7.7)†  3.0 (1.0, 9.4)†  3.3 (1.1, 9.7)*  NCP >0 vs 0  1.4 (0.7, 2.7)  1.3 (0.6, 2.5)  1.3 (0.6, 2.6)  1.4 (0.7, 2.7)  MP/CP >0 vs 0  3.6 (1.8, 7.5)***  2.7 (1.2, 6.0)*  2.9 (1.2, 6.6)*  3.5 (1.6, 7.6)**  a Adjusted for age and gender. b Adjusted for age, gender, hypertension, diabetes, hyperlipidaemia, smoking, BMI and prednisone use. c Adjusted for D’Agostino Framingham score. d cTnI and coronary plaque outcomes (CAC, SIS, SSS, PBS) are binarized based on the median. † P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001. Table 2 Prediction of occult coronary plaque burden and composition by cTnIa   Unadjusted, OR (95% CI)  Model 1a, OR (95% CI)  Model 2b, OR (95% CI)  Model 3c, OR (95% CI)  CACd  2.95 (1.5, 6.0)**  2.2 (1.0, 4.7)*  2.3 (1.0, 5.3)*  2.7 (1.3, 5.8)*  CAC >100 vs ≤100  4.2 (1.3, 13.5)*  3.0 (0.9, 10.5)†  6.6 (1.4, 30.0)*  5.0 (1.3, 19)*  SISd  2.2 (1.1, 4.2)*  1.7 (0.9, 3.4)  1.7 (0.8, 3.6)  1.9 (1.0, 3.8)†  SIS >5 vs ≤5  2.5 (0.7, 8.5)  1.8 (0.5, 6.6)  2.5 (0.6, 10.5)  2.6 (0.7, 9.9)  SSSd  2.4 (1.3, 4.7)**  1.9 (1.0, 3.9)†  2.0 (1.0, 4.2)†  2.2 (1.1, 4.3)*  SSS >5 vs ≤5  3.0 (1.1, 8.2)*  2.4 (0.8, 7.3)  2.9 (0.9, 9.0)†  3.0 (1.0, 9.0)*  PBSd  3.1 (1.5, 6.3)**  2.4 (1.1, 5.1)*  2.7 (1.2, 6.2)*  2.9 (1.3, 6.2)**  Obstructive plaque  3.9 (1.2, 12.5)*  3.1 (0.9, 10.7)†  3.8 (1.0, 14.4)*  4.0 (1.1, 14.1)*  CAC >100 or SSS >5  4.7 (1.8, 12.4)**  2.3 (0.7, 6.9)  5.3 (1.7, 17.0)**  5.2 (1.8, 15.8)**  SIS >5 or SSS >5 or obstructive plaque  3.2 (1.2, 8.8)*  2.6 (0.9, 7.7)†  3.0 (1.0, 9.4)†  3.3 (1.1, 9.7)*  NCP >0 vs 0  1.4 (0.7, 2.7)  1.3 (0.6, 2.5)  1.3 (0.6, 2.6)  1.4 (0.7, 2.7)  MP/CP >0 vs 0  3.6 (1.8, 7.5)***  2.7 (1.2, 6.0)*  2.9 (1.2, 6.6)*  3.5 (1.6, 7.6)**    Unadjusted, OR (95% CI)  Model 1a, OR (95% CI)  Model 2b, OR (95% CI)  Model 3c, OR (95% CI)  CACd  2.95 (1.5, 6.0)**  2.2 (1.0, 4.7)*  2.3 (1.0, 5.3)*  2.7 (1.3, 5.8)*  CAC >100 vs ≤100  4.2 (1.3, 13.5)*  3.0 (0.9, 10.5)†  6.6 (1.4, 30.0)*  5.0 (1.3, 19)*  SISd  2.2 (1.1, 4.2)*  1.7 (0.9, 3.4)  1.7 (0.8, 3.6)  1.9 (1.0, 3.8)†  SIS >5 vs ≤5  2.5 (0.7, 8.5)  1.8 (0.5, 6.6)  2.5 (0.6, 10.5)  2.6 (0.7, 9.9)  SSSd  2.4 (1.3, 4.7)**  1.9 (1.0, 3.9)†  2.0 (1.0, 4.2)†  2.2 (1.1, 4.3)*  SSS >5 vs ≤5  3.0 (1.1, 8.2)*  2.4 (0.8, 7.3)  2.9 (0.9, 9.0)†  3.0 (1.0, 9.0)*  PBSd  3.1 (1.5, 6.3)**  2.4 (1.1, 5.1)*  2.7 (1.2, 6.2)*  2.9 (1.3, 6.2)**  Obstructive plaque  3.9 (1.2, 12.5)*  3.1 (0.9, 10.7)†  3.8 (1.0, 14.4)*  4.0 (1.1, 14.1)*  CAC >100 or SSS >5  4.7 (1.8, 12.4)**  2.3 (0.7, 6.9)  5.3 (1.7, 17.0)**  5.2 (1.8, 15.8)**  SIS >5 or SSS >5 or obstructive plaque  3.2 (1.2, 8.8)*  2.6 (0.9, 7.7)†  3.0 (1.0, 9.4)†  3.3 (1.1, 9.7)*  NCP >0 vs 0  1.4 (0.7, 2.7)  1.3 (0.6, 2.5)  1.3 (0.6, 2.6)  1.4 (0.7, 2.7)  MP/CP >0 vs 0  3.6 (1.8, 7.5)***  2.7 (1.2, 6.0)*  2.9 (1.2, 6.6)*  3.5 (1.6, 7.6)**  a Adjusted for age and gender. b Adjusted for age, gender, hypertension, diabetes, hyperlipidaemia, smoking, BMI and prednisone use. c Adjusted for D’Agostino Framingham score. d cTnI and coronary plaque outcomes (CAC, SIS, SSS, PBS) are binarized based on the median. † P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001. Conversely, subjects with low hs-cTnI (<1.5 pg/ml) were less likely to have extensive coronary atherosclerosis. Specifically, they displayed 81% lower risk of having SSS >5 or CAC ⩾100 and 70% less risk of obstructive plaque, SIS >5 or SSS >5 after controlling for FRS-DA score; the area under the curve (AUC) improved from 0.79 (IQR 0.63–0.95) to 0.85 (0.72–0.98), P < 0.05 (data not shown). Out of all patients, 27 (18%) had CAC >100 or SSS >5 and 22 (15%) had obstructive plaque or SIS >5 or SSS >5. Compared with all patients, only 8% with low hs-cTnI displayed those respective plaque outcomes (supplementary Table S2, available at Rheumatology online); of patients with both low hs-cTnI and low FRS-DA scores, only 4% had extensive atherosclerosis compared with 11% of those with just low FRS-DA. Elevated hs-cTnI associates with long-term CVEs in RA Eleven patients suffered CVEs during 60 (s.d. 26) months of follow-up (1.54/100 patient-years): eight were ischaemic, including one cardiac death, three non-ST elevation MIs, two strokes and two PAD events requiring revascularization; the three non-ischaemic events were new-onset hospitalized heart failure. hs-cTnI was higher in patients with CVEs vs those without [2.6 pg/ml (IQR 2.1–4.4) vs 1.5 (1.0–2.4), P = 0.006]. Elevated hs-cTnI predicted the risk of incident CVEs (Fig. 2A, P = 0.03), independent of demographics and traditional cardiac risk factors (Table 3). Importantly, patients with low hs-cTnI were 82% less likely to suffer a CVE. Table 3 Elevated hs-cTnI (>1.5 pg/ml) predicts the risk of CVEs Model  Hazard ratio (95% CI)  P-value  Unadjusted  4.7 (1.0, 21.7)  0.048  Model 1a  4.8 (1.0, 23.1)  0.052  Model 2b  5.5 (1.1, 26.7)  0.034  Model 3c  4.3 (0.9, 19.7)  0.064  Model  Hazard ratio (95% CI)  P-value  Unadjusted  4.7 (1.0, 21.7)  0.048  Model 1a  4.8 (1.0, 23.1)  0.052  Model 2b  5.5 (1.1, 26.7)  0.034  Model 3c  4.3 (0.9, 19.7)  0.064  a Adjusted for age and gender. b Adjusted for age, gender, hypertension, diabetes, hyperlipidaemia, smoking, BMI and prednisone use. c Adjusted for D’Agostino Framingham score. Table 3 Elevated hs-cTnI (>1.5 pg/ml) predicts the risk of CVEs Model  Hazard ratio (95% CI)  P-value  Unadjusted  4.7 (1.0, 21.7)  0.048  Model 1a  4.8 (1.0, 23.1)  0.052  Model 2b  5.5 (1.1, 26.7)  0.034  Model 3c  4.3 (0.9, 19.7)  0.064  Model  Hazard ratio (95% CI)  P-value  Unadjusted  4.7 (1.0, 21.7)  0.048  Model 1a  4.8 (1.0, 23.1)  0.052  Model 2b  5.5 (1.1, 26.7)  0.034  Model 3c  4.3 (0.9, 19.7)  0.064  a Adjusted for age and gender. b Adjusted for age, gender, hypertension, diabetes, hyperlipidaemia, smoking, BMI and prednisone use. c Adjusted for D’Agostino Framingham score. Fig. 2 View largeDownload slide hs-cTnI optimizes cardiovascular risk prediction in RA (A) Elevated hs-cTnI predicts long-term CVEs in RA. (B) Addition of hs-cTnI information to the FRS-DA composite score increased prognostic accuracy (AUC 0.8431 vs 0.7283; P = 0.1). Further addition of high-risk plaque information from CCTA (obstructive plaque or SIS >5 or SSS >5) resulted in significant enhancement of predictive accuracy over FRS-DA alone (0.9165 vs 0.7283; P = 0.015) and an improving trend over FRS-DA + hs-cTnI (0.9165 vs 0.8431; P = 0.21). Fig. 2 View largeDownload slide hs-cTnI optimizes cardiovascular risk prediction in RA (A) Elevated hs-cTnI predicts long-term CVEs in RA. (B) Addition of hs-cTnI information to the FRS-DA composite score increased prognostic accuracy (AUC 0.8431 vs 0.7283; P = 0.1). Further addition of high-risk plaque information from CCTA (obstructive plaque or SIS >5 or SSS >5) resulted in significant enhancement of predictive accuracy over FRS-DA alone (0.9165 vs 0.7283; P = 0.015) and an improving trend over FRS-DA + hs-cTnI (0.9165 vs 0.8431; P = 0.21). hs-cTnI enhances CVE risk prediction when added to cardiac risk scores The prognostic accuracy of FRS-DA alone vs FRS-DA + hs-cTnI and FRS-DA + hs-cTnI + high-risk plaque for CVEs based on the AUC of the respective receiver operating characteristics curves is depicted in Fig. 2B. The addition of hs-cTnI information to the FRS-DA score yielded a higher prognostic accuracy (0.8431 vs 0.7283; P = 0.10); further addition of high-risk plaque information from CCTA (obstructive plaque or SIS >5 or SSS >5) resulted in significant enhancement of the predictive accuracy over the FRS-DA alone (0.9165 vs 0.7283; P = 0.015) and an improving trend over the FRS-DA + hs-cTnI (0.9165 vs 0.8431; P = 0.21). Since the AUC change in response to a new marker included in a model is often sensitive to only very large independent effects of that marker, we further calculated the IDI to assess additional discrimination offered by inclusion of information from hs-cTnI and high-risk plaque in CVE prediction. Indeed, the addition of hs-cTnI to FRS-DA significantly improved the precision of CVE risk prediction vs FRS-DA alone [Table 4; IDI = 0.0435 (IQR 0.0023–0.0847), P = 0.038]. Further addition of high-risk plaque information significantly enhanced the accuracy of CVE risk prediction over FRS-DA + hs-cTnI [IDI = 0.0818 (IQR 0.0032–0.1605), P = 0.042]. Table 4 Average improvement in precision of CVE risk prediction by integrating hs-cTnI and high-risk CCTA Comparison  IDI (95% CI)  P-value  FRS-DA vs FRS-DA + hs-cTnI  0.0435 (0.0023, 0.0847)  0.038  FRS-DA + hs-cTnI vs FRS-DA + hs-cTnI + high-risk CCTA  0.0818 (0.0032, 0.1605)  0.042  Comparison  IDI (95% CI)  P-value  FRS-DA vs FRS-DA + hs-cTnI  0.0435 (0.0023, 0.0847)  0.038  FRS-DA + hs-cTnI vs FRS-DA + hs-cTnI + high-risk CCTA  0.0818 (0.0032, 0.1605)  0.042  Table 4 Average improvement in precision of CVE risk prediction by integrating hs-cTnI and high-risk CCTA Comparison  IDI (95% CI)  P-value  FRS-DA vs FRS-DA + hs-cTnI  0.0435 (0.0023, 0.0847)  0.038  FRS-DA + hs-cTnI vs FRS-DA + hs-cTnI + high-risk CCTA  0.0818 (0.0032, 0.1605)  0.042  Comparison  IDI (95% CI)  P-value  FRS-DA vs FRS-DA + hs-cTnI  0.0435 (0.0023, 0.0847)  0.038  FRS-DA + hs-cTnI vs FRS-DA + hs-cTnI + high-risk CCTA  0.0818 (0.0032, 0.1605)  0.042  Discussion Patients with RA incur a higher rate of CVEs compared with individuals without autoimmune disease [1]. Therefore, periodic cardiovascular risk stratification according to national guidelines is an integral part of the care of RA patients [26]. However, general risk calculators do not sufficiently capture the incremental risk in patients with RA [27–29]. All stages of the atherogenic process appear enhanced in RA, including endothelial dysfunction, increased arterial stiffness, plaque formation and finally CVEs [30]. Distinct biomarkers may reflect different stages of this pathway, from inflammation (high-sensitivity CRP, IL-6) to plaque instability (myeloperoxidase, MMPs), thrombosis (fibrinogen), myocardial stress (NT-proBNP) and myocardial necrosis (hs-cTn). Individual associations of CRP, hs-cTn and NT-proBNP with CVE in general patients have been extensively described [31]. In RA, CRP may reflect uncontrolled systemic inflammation rather than being a surrogate for the extent of vascular involvement [30]. NT-proBNP independently predicted mortality in one study of 182 RA patients [32]. Our study shows for the first time that hs-cTnI, a specific structural myocardial biomarker, may optimize long-term cardiovascular risk prediction in RA. Blood concentrations of cardiac troponin I and T subunits are elevated in the context of myocardial injury [33]. High-sensitivity assays measure cTnI concentrations at levels much lower than conventional assays with excellent precision at a ⩽10% coefficient of variation, both at and below the assay’s 99th percentile value. This added sensitivity allows reliable estimation in almost 100% of healthy individuals and identification of subclinical myocardial injury [34]. Elevated hs-cTnI was associated with incident long-term CVEs in patients with RA when controlling for traditional cardiac risk factors. This is consistent with reports in population-based studies that subthreshold elevations of either hs-cTnT or hs-cTnI predicted a higher risk of CVEs, heart failure hospitalization and mortality [12–15]. In contrast, RA patients with low hs-cTnI were 82% less likely to suffer a CVE. This approximates the estimated 88% lower risk of CV death in a nested case–control study in general patients with low hs-cTnI measured with the same assay [35]. Moreover, we demonstrated that hs-cTnI measurements significantly improved discrimination of long-term incident CVE risk over composite cardiac risk scores alone. A combination of CRP, NT-proBNP and hs-cTnI optimized the 10-year CVE risk prediction in two general European populations [36]; however, these have not yet been evaluated in RA. In our study, IL-6 was numerically higher in patients incurring CVEs. Nevertheless, a model of high IL-6 combined with hs-cTnI did not optimize event prediction over hs-cTnI alone (data not shown). More multibiomarker groupings will likely emerge in the future. However, the optimal prognostic combinations remain to be defined. Our second novel finding was the association of hs-cTnI with coronary plaque presence, burden and composition in patients with RA, as measured by CCTA. This non-invasive imaging modality has significantly enhanced the prediction of incident CVEs beyond clinical risk scores, as well as CAC in general patients without known CVD [37, 38]. In a prospective study, 69% of subjects with obstructive lesions suffered events at 52 months compared with 28% of those with non-obstructive lesions and 0% of those without plaque. Similarly, 75% with SIS >5 and 80% with SSS >5 suffered CVEs compared with 23% with SIS ⩽5 and 15% with SSS ⩽5 [39]. hs-cTnI was considerably higher in patients with any plaque vs those without; furthermore, it significantly increased across higher plaque burden scores. This is consistent with a prior report in general patients showing progressively higher hs-cTnT in those with mild, moderate and multivessel CAD on CCTA [40]. hs-cTnI was strongly correlated with all quantitative plaque outcomes, including several high-risk ones (obstructive plaque, SSS >5, CAC >100 and composites thereof) after adjustments for traditional risk factors and cardiovascular scores. Moreover, it independently predicted the presence of any advanced—mixed or calcified—coronary plaque, whereas it showed no correlation with earlier non-calcified plaques. In our study, hs-cTnI significantly improved the discrimination of long-term incident CVE risk over cardiac risk scores alone. Additional information on the presence of high-risk plaque outcomes from CCTA further optimized CVE risk discrimination compared with cardiac risk scores and hs-cTnI together. These observations provide the theoretical framework and a testable hypothesis for a two-step algorithm to optimize CVE risk prediction in RA. As part of the cardiac risk stratification, physicians could measure plasma hs-cTnI. If high (>1.5 pg/ml), it may foreshadow a significant hazard for high-risk plaque burden, vulnerability or future CVEs above and beyond cardiac risk scores. In that context, further non-invasive evaluation of coronary atherosclerosis with CCTA may refine primary prevention recommendations based on the presence and burden of coronary plaque. In contrast, if hs-cTnI is low (⩽1.5 pg/ml), the risk of significant coronary atherosclerosis and CVE is substantially decreased. Therefore physicians may narrow their recommendations to address potential actionable clinical risk factors in accordance with cardiac scores. In our study, hs-cTnI was measured at the time of CCTA, when no chest pain was present. In fact, by design, enrolees had no symptoms or diagnosis of CVD upon study entry, hence elevated hs-cTnI levels likely reflect latent myocyte damage. Higher hs-cTnI in general patients has been associated with unstable plaque features on CCTA [41], reflecting intermittent, chronic and clinically silent plaque remodelling and/or rupture with subsequent microembolization, leading to unrecognized MIs (UMIs) [42, 43]. Consistent with these reports, we showed that hs-cTnI in RA patients only correlated with higher-complexity mixed or calcified plaques, independent of cardiac risk factors or Framingham scores but not earlier non-calcified lesions. Greater hs-cTnI associated with the presence of UMIs at baseline, as well as with new or larger UMIs on MRI 5 years later, in a series of community-living volunteers without a history of MI [44]. RA patients are far more likely to experience UMIs, even prior to their RA diagnosis [45]. Indeed, in a pilot MRI study, 39% of RA patients without symptomatic CVD had delayed enhancement suggesting myocardial inflammation or scarring and 11% had nodular subendocardial delayed enhancement indicating silent MI [46]. Latent troponin leak has further been reported as a result of impaired cell membrane integrity due to systemic inflammation [47]. However, we observed no associations between hs-cTnI, inflammatory markers or cytokines, making inflammation an unlikely driver, consistent with an earlier report [16]. Interestingly, we observed no association between pro-inflammatory cytokines, ESR or CRP and the burden of coronary atherosclerosis. This observation may be partially explained by the fact that 58% of our patients were in remission (DAS28-CRP <2.6) at the time of CCTA, while 75% overall had low disease activity (DAS28-CRP <3.2) and 60% were under chronic anti-TNF medication exposure. Concordantly, in the vast majority, IL-6, IL-17A and IL-17F levels were well below the 99% threshold observed in normal patients and similar to or lower than those reported by studies in treated RA patients using identical measurement assays [48, 49]. Our study has certain limitations. Causal relationships between hs-cTnI levels and plaque burden or composition may not be inferred due to their cross-sectional evaluation. Moreover, since our patients were well controlled and the levels of pro-inflammatory cytokines studied were generally low and reflective of that state, we may have underestimated the association of inflammation with both hs-cTnI and plaque burden. Our broader study design, of which the current report is a part, was powered to evaluate quantitative and qualitative plaque differences between 150 RA patients and an equal number of age- and gender-matched patients without autoimmune disease. Although evaluations of biomarkers and their associations with plaque presence, burden and composition in RA patients were pre-specified as exploratory analyses, they were not specifically powered for. Our findings would therefore have to be tested in larger, specifically powered studies and our proposed two-step algorithm for optimization of CVE risk prediction prospectively validated within that context. CVEs appear numerically low in our study [11 patients (6.1%)], which may have deflated overall significance rates—despite sizeable area differences—in AUC curves between FRS-DA alone and FRS-DA + hs-cTnI as well as between FRS-DA + hs-cTnI and FRS-DA + hs-cTnI + high-risk CCTA. This was certainly contributed to by our study design, which pre-specified recruitment of subjects without symptoms or prior diagnosis of CVD. Despite this, our observed event rate was 1.5/100 patient-years, which is similar to studies specifically designed for CV risk [50], and is considered high overall for populations of well-controlled patients chronically exposed to biologic agents. In conclusion, we show for the first time that hs-cTnI associates with the presence, burden and composition of coronary artery atherosclerosis in RA patients without symptoms or prior diagnosis of cardiovascular disease above and beyond traditional risk factors, cardiovascular scores or inflammation. hs-cTnI further associates with the long-term risk of incident CVEs beyond demographics and traditional cardiac risk factors and improves discrimination for such risk prediction beyond that rendered by cardiac risk scores. It may provide a mechanistic explanation for the greater morbidity and mortality RA patients incur and may serve as an adjunct predictive biomarker in refining cardiovascular risk determination in RA. Supplementary data Supplementary data are available at Rheumatology online. Acknowledgements The authors would like to thank Ferdinand Flores, RN for study-related procedure facilitation and blood sample acquisition, handling, storage and shipping. We would also like to express our gratitude to the patients participating in the study. Funding: This work was supported by an American Heart Association grant (09CRP225100 to G.A.K). Disclosure statement: J.E. is employed by Singulex and owns Singulex stock options. J.T. is an employee of Singulex and holds stock options in Singulex. All other authors have declared no conflicts of interest. References 1 Naranjo A, Sokka T, Descalzo MA et al.   Cardiovascular disease in patients with rheumatoid arthritis: results from the QUEST-RA study. Arthritis Res Ther  2008; 10: R30. Google Scholar CrossRef Search ADS PubMed  2 Karpouzas GA, Malpeso J, Choi TY et al.   Prevalence, extent and composition of coronary plaque in patients with rheumatoid arthritis without symptoms or prior diagnosis of coronary artery disease. Ann Rheum Dis  2014; 73: 1797– 804. Google Scholar CrossRef Search ADS PubMed  3 McInnes IB, Buckley CD, Isaacs JD, Cytokines in rheumatoid arthritis – shaping the immunological landscape. Nat Rev Rheumatol  2016; 12: 63– 8. Google Scholar CrossRef Search ADS PubMed  4 Rus HG, Vlaicu R, Niculescu F. Interleukin-6 and interleukin-8 protein and gene expression in human arterial atherosclerotic wall. Atherosclerosis  1996; 127: 263– 71. Google Scholar CrossRef Search ADS PubMed  5 Dixon WG, Symmons DP. What effects might anti-TNFα treatment be expected to have on cardiovascular morbidity and mortality in rheumatoid arthritis? A review of the role of TNFα in cardiovascular pathophysiology. Ann Rheum Dis  2007; 66: 1132. Google Scholar CrossRef Search ADS PubMed  6 Niemann-Jönsson A, Dimayuga P, Jovinge S et al.   Accumulation of LDL in rat arteries is associated with activation of tumor necrosis factor-α expression. Arterioscler Thromb Vasc Biol  2000; 20: 2205– 11. Google Scholar CrossRef Search ADS PubMed  7 Amar J, Fauvel J, Drouet L et al.   Interleukin 6 is associated with subclinical atherosclerosis: a link with soluble intercellular adhesion molecule 1. J Hypertens  2006; 24: 1083– 8. Google Scholar CrossRef Search ADS PubMed  8 Kuo LT, Yang NI, Cherng WJ et al.   Serum interleukin-6 levels, not genotype, correlate with coronary plaque complexity. Int Heart J  2008; 49: 391– 402. Google Scholar CrossRef Search ADS PubMed  9 Ridker PM, Rifai N, Stampfer MJ, Hennekens CH. Plasma concentration of interleukin-6 and the risk of future myocardial infarction among apparently healthy men. Circulation  2000; 101: 1767– 72. Google Scholar CrossRef Search ADS PubMed  10 Biasucci LM, Liuzzo G, Fantuzzi G et al.   Increasing levels of interleukin (IL)-1Ra and IL-6 during the first 2 days of hospitalization in unstable angina are associated with increased risk of in-hospital coronary events. Circulation  1999; 99: 2079– 84. Google Scholar CrossRef Search ADS PubMed  11 Hashmi S, Zeng QT. Role of interleukin-17 and interleukin-17-induced cytokines interleukin-6 and interleukin-8 in unstable coronary artery disease. Coron Artery Dis  2006; 17: 699– 706. Google Scholar CrossRef Search ADS PubMed  12 deFilippi CR, de Lemos JA, Christenson RH et al.   Association of serial measures of cardiac troponin T using a sensitive assay with incident heart failure and cardiovascular mortality in older adults. JAMA  2010; 304: 2494– 502. Google Scholar CrossRef Search ADS PubMed  13 de Lemos JA, Drazner MH, Omland T et al.   Association of troponin T detected with a highly sensitive assay and cardiac structure and mortality risk in the general population. JAMA  2010; 304: 2503– 12. Google Scholar CrossRef Search ADS PubMed  14 Saunders JT, Nambi V, de Lemos JA et al.   Cardiac troponin T measured by a highly sensitive assay predicts coronary heart disease, heart failure, and mortality in the Atherosclerosis Risk in Communities Study. Circulation  2011; 123: 1367– 76. Google Scholar CrossRef Search ADS PubMed  15 Zeller T, Tunstall-Pedoe H, Saarela O et al.   High population prevalence of cardiac troponin I measured by a high-sensitivity assay and cardiovascular risk estimation: the MORGAM Biomarker Project Scottish Cohort. Eur Heart J  2014; 35: 271– 81. Google Scholar CrossRef Search ADS PubMed  16 Bradham WS, Bian A, Oeser A et al.   High-sensitivity cardiac troponin-I is elevated in patients with rheumatoid arthritis, independent of cardiovascular risk factors and inflammation. PLoS One  2012; 7: e38930. Google Scholar CrossRef Search ADS PubMed  17 D’Agostino RBSr, Vasan RS, Pencina MJ et al.   General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation  2008; 117: 743– 53. Google Scholar CrossRef Search ADS PubMed  18 Todd J, Freese B, Lu A et al.   Ultrasensitive flow-based immunoassays using single-molecule counting. Clin Chem  2007; 53: 1990– 5. Google Scholar CrossRef Search ADS PubMed  19 Budoff MJ, Dowe D, Jollis JG et al.   Diagnostic performance of 64-multidetector row coronary computed tomographic angiography for evaluation of coronary artery stenosis in individuals without known coronary artery disease: results from the prospective multicenter ACCURACY (Assessment by Coronary Computed Tomographic Angiography of Individuals Undergoing Invasive Coronary Angiography) trial. J Am Coll Cardiol  2008; 52: 1724– 32. Google Scholar CrossRef Search ADS PubMed  20 Agatston AS, Janowitz WR, Hildner FJ et al.   Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol  1990; 15: 827– 32. Google Scholar CrossRef Search ADS PubMed  21 Austen WG, Edwards JE, Frye RL et al.   A reporting system on patients evaluated for coronary artery disease. Report of the Ad Hoc Committee for Grading of Coronary Artery Disease, Council on Cardiovascular Surgery, American Heart Association. Circulation  1975; 51: 5– 40. Google Scholar CrossRef Search ADS PubMed  22 Pagali SR, Madaj P, Gupta M et al.   Interobserver variations of plaque severity score and segment stenosis score in coronary arteries using 64 slice multidetector computed tomography: a substudy of the ACCURACY trial. J Cardiovasc Comput Tomogr  2010; 4: 312– 8. Google Scholar CrossRef Search ADS PubMed  23 Thygesen K, Alpert JS, Jaffe AS et al.   Third universal definition of myocardial infarction. Glob Heart  2012; 7: 275– 95. Google Scholar CrossRef Search ADS PubMed  24 Goldstein LB, Adams R, Alberts MJ et al.   Primary prevention of ischemic stroke: a guideline from the American Heart Association/American Stroke Association Stroke Council: cosponsored by the Atherosclerotic Peripheral Vascular Disease Interdisciplinary Working Group; Cardiovascular Nursing Council; Clinical Cardiology Council; Nutrition, Physical Activity, and Metabolism Council; and the Quality of Care and Outcomes Research Interdisciplinary Working Group: the American Academy of Neurology affirms the value of this guideline. Stroke  2006; 37: 1583– 633. Google Scholar CrossRef Search ADS PubMed  25 Rooke TW, Hirsch AT, Misra S et al.   2011 ACCF/AHA focused update of the guideline for the management of patients with peripheral artery disease (updating the 2005 guideline): a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol  2011; 58: 2020– 45. Google Scholar CrossRef Search ADS PubMed  26 Agca R, Heslinga SC, Rollefstad S et al.   EULAR recommendations for cardiovascular disease risk management in patients with rheumatoid arthritis and other forms of inflammatory joint disorders: 2015/2016 update. Ann Rheum Dis  2017; 76: 17– 28. Google Scholar CrossRef Search ADS PubMed  27 Crowson CS, Matteson EL, Roger VL et al.   Usefulness of risk scores to estimate the risk of cardiovascular disease in patients with rheumatoid arthritis. Am J Cardiol  2012; 110: 420– 4. Google Scholar CrossRef Search ADS PubMed  28 Arts EE, Popa C, Den Broeder AA et al.   Performance of four current risk algorithms in predicting cardiovascular events in patients with early rheumatoid arthritis. Ann Rheum Dis  2015; 74: 668– 74. Google Scholar CrossRef Search ADS PubMed  29 Siontis GC, Tzoulaki I, Siontis KC, Ioannidis JP. Comparisons of established risk prediction models for cardiovascular disease: systematic review. BMJ  2012; 344: e3318. Google Scholar CrossRef Search ADS PubMed  30 Kozera L, Andrews J, Morgan AW. Cardiovascular risk and rheumatoid arthritis—the next step: differentiating true soluble biomarkers of cardiovascular risk from surrogate measures of inflammation. Rheumatology  2011; 50: 1944– 54. Google Scholar CrossRef Search ADS PubMed  31 Lacey B, Herrington WG, Preiss D, Lewington S, Armitage J. The role of emerging risk factors in cardiovascular outcomes. Curr Atheroscler Rep  2017; 19: 28. Google Scholar CrossRef Search ADS PubMed  32 Provan S, Angel K, Semb AG, Atar D, Kvien TK. NT-proBNP predicts mortality in patients with rheumatoid arthritis: results from 10-year follow-up of the EURIDISS study. Ann Rheum Dis  2010; 69: 1946– 50. Google Scholar CrossRef Search ADS PubMed  33 Gupta S, de Lemos JA. Use and misuse of cardiac troponins in clinical practice. Prog Cardiovasc Dis  2007; 50: 151– 65. Google Scholar CrossRef Search ADS PubMed  34 Januzzi JLJr, Bamberg F, Lee H et al.   High-sensitivity troponin T concentrations in acute chest pain patients evaluated with cardiac computed tomography. Circulation  2010; 121: 1227– 34. Google Scholar CrossRef Search ADS PubMed  35 Apple FS, Steffen LM, Pearce LA, Murakami MM, Luepker RV. Increased cardiac troponin I as measured by a high-sensitivity assay is associated with high odds of cardiovascular death: the Minnesota Heart Survey. Clin Chem  2012; 58: 930. Google Scholar CrossRef Search ADS PubMed  36 Blankenberg S, Zeller T, Saarela O et al.   Contribution of 30 biomarkers to 10-year cardiovascular risk estimation in 2 population cohorts: the MONICA, risk, genetics, archiving, and monograph (MORGAM) biomarker project. Circulation  2010; 121: 2388– 97. Google Scholar CrossRef Search ADS PubMed  37 Hou ZH, Lu B, Gao Y et al.   Prognostic value of coronary CT angiography and calcium score for major adverse cardiac events in outpatients. JACC Cardiovasc Imaging  2012; 5: 990– 9. Google Scholar CrossRef Search ADS PubMed  38 Budoff MJ, Shaw LJ, Liu ST et al.   Long-term prognosis associated with coronary calcification: observations from a registry of 25, 253 patients. J Am Coll Cardiol  2007; 49: 1860– 70. Google Scholar CrossRef Search ADS PubMed  39 Andreini D, Pontone G, Mushtaq S et al.   A long-term prognostic value of coronary CT angiography in suspected coronary artery disease. JACC Cardiovasc Imaging  2012; 5: 690– 701. Google Scholar CrossRef Search ADS PubMed  40 Laufer EM, Mingels AM, Winkens MH et al.   The extent of coronary atherosclerosis is associated with increasing circulating levels of high sensitive cardiac troponin T. Arterioscler Thromb Vasc Biol  2010; 30: 1269– 75. Google Scholar CrossRef Search ADS PubMed  41 Altintas S, Cardinaels EP, Versteylen MO et al.   Unstable coronary plaque characteristics are associated with high-sensitivity cardiac troponin T and N-terminal Pro-Brain Natriuretic Peptide. J Cardiovasc Comput Tomogr  2016; 10: 82– 8. Google Scholar CrossRef Search ADS PubMed  42 Korosoglou G, Lehrke S, Mueller D et al.   Determinants of troponin release in patients with stable coronary artery disease: insights from CT angiography characteristics of atherosclerotic plaque. Heart  2011; 97: 823– 31. Google Scholar CrossRef Search ADS PubMed  43 Seifarth H, Schlett CL, Lehman SJ et al.   Correlation of concentrations of high-sensitivity troponin T and high-sensitivity C-reactive protein with plaque progression as measured by CT coronary angiography. J Cardiovasc Comput Tomogr  2014; 8: 452– 8. Google Scholar CrossRef Search ADS PubMed  44 Barbier CE, Themudo R, Bjerner T et al.   Cardiac troponin I associated with the development of unrecognized myocardial infarctions detected with MRI. Clin Chem  2014; 60: 1327– 35. Google Scholar CrossRef Search ADS PubMed  45 Maradit-Kremers H, Crowson CS, Nicola PJ et al.   Increased unrecognized coronary heart disease and sudden deaths in rheumatoid arthritis: a population-based cohort study. Arthritis Rheum  2005; 52: 402– 11. Google Scholar CrossRef Search ADS PubMed  46 Kobayashi Y, Giles JT, Hirano M et al.   Assessment of myocardial abnormalities in rheumatoid arthritis using a comprehensive cardiac magnetic resonance approach: a pilot study. Arthritis Res Ther  2010; 12: R171. Google Scholar CrossRef Search ADS PubMed  47 Levine B, Kalman J, Mayer L, Fillit HM, Packer M. Elevated circulating levels of tumor necrosis factor in severe chronic heart failure. N Engl J Med  1990; 323: 236– 41. Google Scholar CrossRef Search ADS PubMed  48 Todd J, Simpson P, Estis J, Torres V, Wub AH. Reference range and short- and long-term biological variation of interleukin (IL)-6, IL-17A and tissue necrosis factor-alpha using high sensitivity assays. Cytokine  2013; 64: 660– 5. Google Scholar CrossRef Search ADS PubMed  49 Jain M, Attur M, Furer V et al.   Increased plasma IL-17F levels in rheumatoid arthritis patients are responsive to methotrexate, anti-TNF, and T cell costimulatory modulation. Inflammation  2015; 38: 180– 6. Google Scholar CrossRef Search ADS PubMed  50 Giles JT, Sattar N, Gabriel SE et al.   Results of a randomized, parallel-group, multicenter, noninferiority, phase 4 clinical trial [abstract]. Arthritis Rheumatol  2016; 68(Suppl 10):abstract 3L. http://acrabstracts.org/abstract/comparative-cardiovascular-safety-of-tocilizumab-vs-etanercept-in-rheumatoid-arthritis-results-of-a-randomized-parallel-group-multicenter-noninferiority-phase-4-clinical-trial/ (21 September 2017, date last accessed). © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Rheumatology Oxford University Press

High-sensitivity cardiac troponin I is a biomarker for occult coronary plaque burden and cardiovascular events in patients with rheumatoid arthritis

Loading next page...
 
/lp/ou_press/high-sensitivity-cardiac-troponin-i-is-a-biomarker-for-occult-coronary-9HZSWja1FY
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oup.com
ISSN
1462-0324
eISSN
1462-0332
D.O.I.
10.1093/rheumatology/key057
Publisher site
See Article on Publisher Site

Abstract

Abstract Objectives Patients with RA display greater occult coronary atherosclerosis burden and experience higher cardiovascular morbidity and mortality compared with controls. We here explored whether pro-inflammatory cytokines and high-sensitivity cardiac troponin I (hs-cTnI), a biomarker of myocardial injury, correlated with plaque burden and cardiovascular events (CVEs) in RA. Methods We evaluated 150 patients with 64-slice coronary CT angiography. Coronary artery calcium, number of segments with plaque (segment involvement score), stenotic severity and plaque burden were assessed. Lesions were described as non-calcified, mixed or fully calcified. Blood levels of hs-cTnI and pro-inflammatory cytokines were assessed during coronary CT angiography. Subjects were followed over 60 (s.d. 26) months for both ischaemic [cardiac death, non-fatal myocardial infarction (MI), stroke, peripheral arterial ischaemia] and non-ischaemic (new-onset heart failure hospitalization) CVEs. Results Plasma hs-cTnI correlated with all coronary plaque outcomes (P < 0.01). Elevated hs-cTnI (⩾1.5 pg/ml) further associated with significant calcification, extensive atherosclerosis, obstructive plaque and any advanced mixed or calcified plaques after adjustments for cardiac risk factors or Framingham D’Agostino scores (all P < 0.05). Eleven patients suffered a CVE (1.54/100 patient-years), eight ischaemic and three non-ischaemic. Elevated hs-cTnI predicted all CVE risk independent of demographics, cardiac risk factors and prednisone use (P = 0.03). Conversely, low hs-cTnI presaged a lower risk for both extensive atherosclerosis (P < 0.05) and incident CVEs (P = 0.037). Conclusion Plasma hs-cTnI independently associated with occult coronary plaque burden, composition and long-term incident CVEs in patients with RA. Low hs-cTnI forecasted a lower risk for both extensive atherosclerosis as well as CVEs. hs-cTnI may therefore optimize cardiovascular risk stratification in RA. cardiovascular events, high-sensitivity cardiac troponin I, occult coronary atherosclerosis, rheumatoid arthritis Rheumatology key messages High-sensitivity cardiac troponin I correlates with the presence, burden and composition of occult coronary plaque in RA. High-sensitivity cardiac troponin I further correlates with long-term cardiac events in RA after adjustment for cardiac risk factors. High-sensitivity cardiac troponin I may serve as a predictive biomarker in refining cardiovascular risk assessment in RA. Introduction Individuals with RA experience a higher rate of cardiovascular events (CVEs) compared with controls [1]. This may be explained by the greater prevalence, severity, burden and different composition of occult coronary lesions in RA compared with age- and gender-matched controls [2]. Residual disease activity may further associate with more advanced, complex and prone-to-rupture coronary plaques [2]. Pro-inflammatory cytokines such as TNF-α, IL-6 and IL-17 reflect clinical activity and structural damage in RA and are elevated in the blood of RA patients compared with controls [3]; the same cytokines have been identified in atherosclerotic plaque [4–6] and correlated with subclinical atherosclerosis independent of cardiac risk factors [7], coronary plaque complexity [8], plaque destabilization and CVEs in subjects without autoimmune disease [9–11]. Nevertheless, the relationship between these cytokines and occult coronary plaque burden and composition in RA are unknown. Higher plaque load or vulnerability may be further reflected in elevations of biomarkers specific for myocardial injury [12]; indeed, cardiac troponin elevations measured by high-sensitivity assays—and below thresholds used to diagnose acute coronary syndromes—were associated with higher coronary artery calcium (CAC) scores in a population-based study [12]. Additionally, both high-sensitivity cardiac troponin T (hs-cTnT) and high-sensitivity cardiac troponin I (hs-cTnI) predicted a greater risk of fatal and non-fatal coronary heart disease, heart failure hospitalization and overall mortality in the general population [12–15]. In a recent report, hs-cTnI was higher in RA patients compared with controls, independent of cardiovascular risk factors and inflammation [16]. Nevertheless, its association with subclinical coronary artery disease (CAD) burden and its ability to predict future CVEs in RA are unknown. We here hypothesized that hs-cTnI and various pro-inflammatory cytokines may correlate with the presence, burden and composition of occult coronary plaque in patients with RA evaluated with coronary CT angiography (CCTA). We further postulated that hs-cTnI at the time of CCTA might predict incident CVEs on long-term follow-up [60 months (s.d. 26)]. Methods Patient recruitment A total of 150 RA patients from a single centre were enrolled and prospectively evaluated on a first-come, first-served basis between 1 March 2010 and 1 March 2011 [2]. The study was approved by the local institutional review board [John F Wolf, MD, Human Subjects Committees (1 + 2)], all subjects signed informed consent and the research was carried out in compliance with the Helsinki Declaration. Inclusion criteria comprised age ⩾18 years, fulfilment of 2010 classification criteria for RA and no symptoms or history of cardiovascular disease, including myocardial infarction (MI), revascularization, heart failure, transient ischaemic attack, stroke or peripheral arterial disease (PAD). Patients with concomitant autoimmune syndromes, malignancy within <5 years, chronic or active infection, weight >325 pounds (147.7 kg), hypotension [systolic blood pressure (SBP) <90 mmHg or diastolic blood pressure (DBP) <60 mmHg] or hypertension (SBP >170 mmHg or DBP >110 mmHg), uncontrolled tachycardia, irregular rhythm, iodine allergy or glomerular filtration rate <60 ml/min were excluded. Hypertension was defined as SBP ⩾140 mmHg or DBP ⩾90 mmHg or antihypertensive use. Diabetes mellitus encompassed haemoglobin A1c >6.5% or hypoglycaemic medication use. Hyperlipidaemia constituted fasting cholesterol >200 mg/dl or low-density lipoprotein cholesterol >130 mg/dl or current statin use. Current smoking entailed cigarette consumption within 30 days from screening. Positive family history was defined as CAD in first-degree relatives <55 years of age for males or <65 years of age for females. The Framingham 2008 D’Agostino modified general cardiovascular risk (FRS-DA) score was calculated for all study participants [17]. Disease duration, serologic status, radiographs and treatments were captured. RA activity was evaluated by a 28-joint DAS with CRP (DAS28-CRP). Laboratory evaluations Blood for regular chemistries, fasting lipids, ESR and high-sensitivity CRP was collected on the day of CCTA and evaluated at the local laboratory. Additionally, blood was collected in EDTA tubes and immediately processed and plasma was frozen at −80°C until it was assayed. hs-cTnI was measured at Singulex (Alameda, CA, USA) by technicians blinded to the clinical data using a microparticle immunoassay and single-molecule counting [18]. TNF-α, IL-6, IL-17A and F and VEGF were also assessed using laboratory tests developed at Singulex based on single-molecule counting [18]. Multidetector CTA Scans were performed with a 64-multidetector row Lightspeed VCT scanner (GE Healthcare, Chicago, IL, USA) between March 2010 and March 2011 and the images were analysed as previously described by a single, blinded interpreter [19]. CAC was quantified by the Agatston method [20]. Coronary arteries were evaluated on contrast-enhanced scans using a standardized 15-segment model [21]. Stenosis severity was scored from 0 to 4 based on the grade of luminal restriction: 1 represented 1–29% stenosis; 2, 30–49%; 3, 50–69% and 4, ⩾70%. The area of each plaque visualized in at least two adjacent slices (slice thickness 0.625 mm) was determined on all affected slices. Plaque burden was graded from 0 to 3, defined as none (0), mild (1), moderate (2) and severe (3), based on the number of adjacent slices containing plaque. Lesions rendering >50% stenosis were considered obstructive. Plaque composition was defined as non-calcified, mixed (MP) or calcified (CP), as discussed elsewhere [22]. Subjects received three individual quantitative scores [22]: the segment involvement score (SIS) represented the total number of segments with plaque (0–15); the stenosis severity score (SSS) reported the cumulative stenosis grade conferred by plaque over all evaluable segments (0–60) and the plaque burden score (PBS) described the cumulative plaque size over all evaluable segments (0–45). Reproducibility of scoring measurements for our centre have been previously reported [22]. Incident CVEs All patients were followed for incident CVEs over a period of 60 (s.d. 26) months. These included both ischaemic (cardiac death, non-fatal MI, stroke, transient ischaemic attack, PAD) as well as non-ischaemic ones (new-onset heart failure hospitalization). Event adjudication was elaborated by the treating cardiologist, neurologist or vascular surgeon and based on standard definitions [23–25]. Analysis Continuous variables are expressed as medians with interquartile ranges (IQRs) and categorical variables are expressed as numbers with percentages. Spearman’s rho correlation coefficients evaluated preliminary associations between biomarkers and plaque outcomes. Medians between CVE groups were compared using the Mann–Whitney U test and counts by the chi-square test. Further analyses were restricted to biomarkers with significant differences. Plaque outcomes were binarized based on median; those were >0 vs 0 for CAC, >1 vs ⩽1 for SIS, >1 vs ⩽1 for SSS and >2 vs ⩽2 for PBS. To evaluate plaque composition, the presence of MP or CP vs the absence of both was used as an outcome. Similarly, hs-cTnI was binarized as high (>1.5 pg/ml) vs low (⩽1.5 pg/ml). Logistic regression models were constructed to evaluate associations between hs-cTnI and individual plaque parameters; models were adjusted for either age and gender (model 1) or additionally for hypertension, diabetes, hyperlipidaemia, smoking, BMI and prednisone use (model 2) or for the patients’ FRS-DA score (model 3). Similar logistic regression models were devised to predict individual or composite high-risk plaque outcomes; individual ones included CAC >100 vs ⩽100, SIS >5 vs ⩽5, SSS >5 vs ⩽5 and the presence of obstructive plaque vs not. Composite outcomes entailed the presence of SSS >5 or CAC >100 vs neither and SIS >5 or SSS >5 or obstructive plaque vs none. Results are reported as odds ratios with 95% CIs. For composite and plaque composition outcomes, sensitivity, specificity and negative and positive predictive values were determined with standard formulas. Cox proportional hazards regression analysis evaluated CVE risk [hazard ratio (HR)] associated with high cTnI (>1.5 pg/ml) in raw and several adjusted models (models 1, 2 and 3) as previously described. Kaplan–Meier curves were compared by the log-rank method. The diagnostic accuracy of CVE risk assessment for FRS-DA alone, FRS-DA + hs-cTnI, and FRS-DA + hs-cTnI + high-risk composite outcome was evaluated as the area under the receiver operator characteristic curve and compared with the DeLong method. Integrated discrimination improvement (IDI) between these constructs was computed and improvement in prediction accuracy was evaluated, where P-values <0.05 were significant. Data were analysed with SAS version 9.4 (SAS Institute, Cary, NC, USA) or R version 3.4.1 (R Project for Statistical Computing, Vienna, Austria). Results Patient demographics are shown in Table 1. Subjects were predominantly female, with established seropositive, erosive and well-controlled disease. RA parameters and traditional cardiac risk factors were not significantly different in patients incurring CVEs. In contrast, FRS-DA, coronary atherosclerosis burden, including high-risk plaque parameters, and levels of hs-cTnI were significantly higher (all P < 0.05). Table 1 Baseline patient characteristics Characteristics  All (n = 150)  No CVE (n = 139)  With CVE (n = 11)  Age, years  54 (46–60)  54 (45–60)  59 (53–70)†  Males, n (%)  18 (13)  16 (12)  2 (18)  RF positive, n (%)  129 (86)  120 (86)  9 (82)  ACPA positive, n (%)  127 (85)  118 (85)  9 (82)  X-ray erosions, n (%)  99 (66)  90 (65)  9 (82)  RA duration, years  9 (5–14)  9 (4–14)  12 (9–18)  Tender joint count  0 (0–2)  0 (0–2)  0 (0–1)  Swollen joint count  0.5 (0–3)  0 (0–3)  1 (0–5)  DAS28-CRP  2.30 (1.8–3.3)  2.32 (1.76–3.29)  2.39 (1.59–2.93)  High-sensitivity CRP, mg/dl  0.41 (0.20–0.96)  0.42 (0.21–0.88)  0.31 (0.12–0.76)  Prednisone, n (%)  52 (35)  48 (35)  4 (36)  Concurrent DMARDs  2 (1–3)  2 (1–3)  2 (1–3)  TNF inhibitor exposed, n (%)  90 (60)  84 (60)  6 (55)  TNF inhibitor duration, years  4.4 (2.4–6.0)  4.0 (2.2–6.0)  6.0 (4.2–6.9)  Diabetes mellitus, n (%)  26 (18)  22 (16)  4 (36)†  Hypertension, n (%)  64 (44)  58 (43)  8 (73)†  Current smoker, n (%)  13 (9)  12 (9)  1 (9)  Family history of CAD, n (%)  6 (4)  5 (3.6)  1 (9)  Hyperlipidaemia, n (%)  26 (17)  25 (18)  1 (9)  BMI, kg/m2  28.1 (25.8–32.6)  28.4 (26–32.8)  25.9 (23–30.5)†  D’Agostino Framingham score  6.4 (3.0–11.7)  6.2 (2.7–11.2)  15.6 (4.9–20.0)*  Coronary artery calcium  0 (0–19)  0 (0–10)  120 (0–361)***  Segment involvement score  1 (0–3)  1 (0–2)  5 (1–7)**  Segment stenosis score  1 (0–4)  1 (0–3)  9 (1–14)**  Plaque burden score  1.5 (0–3)  1 (0–3)  9 (1–12)**  Obstructive plaque (>50%), n (%)  18 (12)  12 (9)  6 (55)****  Non-calcified plaque score  1 (0–2)  1 (0–2)  1 (0–5)  Mixed plaque score  0 (0–1)  0 (0–0)  3 (0–7)*  Calcified plaque score  0 (0–0)  0 (0–0)  1 (0–3)  hs-cTnI, pg/mla  1.5 (1.1–2.6)  1.5 (1–2.4)  2.6 (2.1–4.4)**  IL-17A, pg/mla  1.3 (0.8–1.8)  1.3 (0.8–1.9)  1.2 (0.8–1.5)  IL-17F, pg/mla  35.8 (22.1–66.8)  36.3 (22.1–62)  31 (22.1–97.4)  IL-6, pg/mla  2.9 (1.8–5.8)  2.8 (1.7–5.4)  3.7 (2.6–13.4) †  TNF-α, pg/mla  11.2 (7.8–24)  11.3 (7.7–24)  11.1 (7.9–30.4)  VEGF, pg/mla  134 (61–221)  134 (60–221)  134 (77–179)  Characteristics  All (n = 150)  No CVE (n = 139)  With CVE (n = 11)  Age, years  54 (46–60)  54 (45–60)  59 (53–70)†  Males, n (%)  18 (13)  16 (12)  2 (18)  RF positive, n (%)  129 (86)  120 (86)  9 (82)  ACPA positive, n (%)  127 (85)  118 (85)  9 (82)  X-ray erosions, n (%)  99 (66)  90 (65)  9 (82)  RA duration, years  9 (5–14)  9 (4–14)  12 (9–18)  Tender joint count  0 (0–2)  0 (0–2)  0 (0–1)  Swollen joint count  0.5 (0–3)  0 (0–3)  1 (0–5)  DAS28-CRP  2.30 (1.8–3.3)  2.32 (1.76–3.29)  2.39 (1.59–2.93)  High-sensitivity CRP, mg/dl  0.41 (0.20–0.96)  0.42 (0.21–0.88)  0.31 (0.12–0.76)  Prednisone, n (%)  52 (35)  48 (35)  4 (36)  Concurrent DMARDs  2 (1–3)  2 (1–3)  2 (1–3)  TNF inhibitor exposed, n (%)  90 (60)  84 (60)  6 (55)  TNF inhibitor duration, years  4.4 (2.4–6.0)  4.0 (2.2–6.0)  6.0 (4.2–6.9)  Diabetes mellitus, n (%)  26 (18)  22 (16)  4 (36)†  Hypertension, n (%)  64 (44)  58 (43)  8 (73)†  Current smoker, n (%)  13 (9)  12 (9)  1 (9)  Family history of CAD, n (%)  6 (4)  5 (3.6)  1 (9)  Hyperlipidaemia, n (%)  26 (17)  25 (18)  1 (9)  BMI, kg/m2  28.1 (25.8–32.6)  28.4 (26–32.8)  25.9 (23–30.5)†  D’Agostino Framingham score  6.4 (3.0–11.7)  6.2 (2.7–11.2)  15.6 (4.9–20.0)*  Coronary artery calcium  0 (0–19)  0 (0–10)  120 (0–361)***  Segment involvement score  1 (0–3)  1 (0–2)  5 (1–7)**  Segment stenosis score  1 (0–4)  1 (0–3)  9 (1–14)**  Plaque burden score  1.5 (0–3)  1 (0–3)  9 (1–12)**  Obstructive plaque (>50%), n (%)  18 (12)  12 (9)  6 (55)****  Non-calcified plaque score  1 (0–2)  1 (0–2)  1 (0–5)  Mixed plaque score  0 (0–1)  0 (0–0)  3 (0–7)*  Calcified plaque score  0 (0–0)  0 (0–0)  1 (0–3)  hs-cTnI, pg/mla  1.5 (1.1–2.6)  1.5 (1–2.4)  2.6 (2.1–4.4)**  IL-17A, pg/mla  1.3 (0.8–1.8)  1.3 (0.8–1.9)  1.2 (0.8–1.5)  IL-17F, pg/mla  35.8 (22.1–66.8)  36.3 (22.1–62)  31 (22.1–97.4)  IL-6, pg/mla  2.9 (1.8–5.8)  2.8 (1.7–5.4)  3.7 (2.6–13.4) †  TNF-α, pg/mla  11.2 (7.8–24)  11.3 (7.7–24)  11.1 (7.9–30.4)  VEGF, pg/mla  134 (61–221)  134 (60–221)  134 (77–179)  Values are presented as median (IQR) unless stated otherwise. † P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. a Biomarker data were available in 146 patients. Table 1 Baseline patient characteristics Characteristics  All (n = 150)  No CVE (n = 139)  With CVE (n = 11)  Age, years  54 (46–60)  54 (45–60)  59 (53–70)†  Males, n (%)  18 (13)  16 (12)  2 (18)  RF positive, n (%)  129 (86)  120 (86)  9 (82)  ACPA positive, n (%)  127 (85)  118 (85)  9 (82)  X-ray erosions, n (%)  99 (66)  90 (65)  9 (82)  RA duration, years  9 (5–14)  9 (4–14)  12 (9–18)  Tender joint count  0 (0–2)  0 (0–2)  0 (0–1)  Swollen joint count  0.5 (0–3)  0 (0–3)  1 (0–5)  DAS28-CRP  2.30 (1.8–3.3)  2.32 (1.76–3.29)  2.39 (1.59–2.93)  High-sensitivity CRP, mg/dl  0.41 (0.20–0.96)  0.42 (0.21–0.88)  0.31 (0.12–0.76)  Prednisone, n (%)  52 (35)  48 (35)  4 (36)  Concurrent DMARDs  2 (1–3)  2 (1–3)  2 (1–3)  TNF inhibitor exposed, n (%)  90 (60)  84 (60)  6 (55)  TNF inhibitor duration, years  4.4 (2.4–6.0)  4.0 (2.2–6.0)  6.0 (4.2–6.9)  Diabetes mellitus, n (%)  26 (18)  22 (16)  4 (36)†  Hypertension, n (%)  64 (44)  58 (43)  8 (73)†  Current smoker, n (%)  13 (9)  12 (9)  1 (9)  Family history of CAD, n (%)  6 (4)  5 (3.6)  1 (9)  Hyperlipidaemia, n (%)  26 (17)  25 (18)  1 (9)  BMI, kg/m2  28.1 (25.8–32.6)  28.4 (26–32.8)  25.9 (23–30.5)†  D’Agostino Framingham score  6.4 (3.0–11.7)  6.2 (2.7–11.2)  15.6 (4.9–20.0)*  Coronary artery calcium  0 (0–19)  0 (0–10)  120 (0–361)***  Segment involvement score  1 (0–3)  1 (0–2)  5 (1–7)**  Segment stenosis score  1 (0–4)  1 (0–3)  9 (1–14)**  Plaque burden score  1.5 (0–3)  1 (0–3)  9 (1–12)**  Obstructive plaque (>50%), n (%)  18 (12)  12 (9)  6 (55)****  Non-calcified plaque score  1 (0–2)  1 (0–2)  1 (0–5)  Mixed plaque score  0 (0–1)  0 (0–0)  3 (0–7)*  Calcified plaque score  0 (0–0)  0 (0–0)  1 (0–3)  hs-cTnI, pg/mla  1.5 (1.1–2.6)  1.5 (1–2.4)  2.6 (2.1–4.4)**  IL-17A, pg/mla  1.3 (0.8–1.8)  1.3 (0.8–1.9)  1.2 (0.8–1.5)  IL-17F, pg/mla  35.8 (22.1–66.8)  36.3 (22.1–62)  31 (22.1–97.4)  IL-6, pg/mla  2.9 (1.8–5.8)  2.8 (1.7–5.4)  3.7 (2.6–13.4) †  TNF-α, pg/mla  11.2 (7.8–24)  11.3 (7.7–24)  11.1 (7.9–30.4)  VEGF, pg/mla  134 (61–221)  134 (60–221)  134 (77–179)  Characteristics  All (n = 150)  No CVE (n = 139)  With CVE (n = 11)  Age, years  54 (46–60)  54 (45–60)  59 (53–70)†  Males, n (%)  18 (13)  16 (12)  2 (18)  RF positive, n (%)  129 (86)  120 (86)  9 (82)  ACPA positive, n (%)  127 (85)  118 (85)  9 (82)  X-ray erosions, n (%)  99 (66)  90 (65)  9 (82)  RA duration, years  9 (5–14)  9 (4–14)  12 (9–18)  Tender joint count  0 (0–2)  0 (0–2)  0 (0–1)  Swollen joint count  0.5 (0–3)  0 (0–3)  1 (0–5)  DAS28-CRP  2.30 (1.8–3.3)  2.32 (1.76–3.29)  2.39 (1.59–2.93)  High-sensitivity CRP, mg/dl  0.41 (0.20–0.96)  0.42 (0.21–0.88)  0.31 (0.12–0.76)  Prednisone, n (%)  52 (35)  48 (35)  4 (36)  Concurrent DMARDs  2 (1–3)  2 (1–3)  2 (1–3)  TNF inhibitor exposed, n (%)  90 (60)  84 (60)  6 (55)  TNF inhibitor duration, years  4.4 (2.4–6.0)  4.0 (2.2–6.0)  6.0 (4.2–6.9)  Diabetes mellitus, n (%)  26 (18)  22 (16)  4 (36)†  Hypertension, n (%)  64 (44)  58 (43)  8 (73)†  Current smoker, n (%)  13 (9)  12 (9)  1 (9)  Family history of CAD, n (%)  6 (4)  5 (3.6)  1 (9)  Hyperlipidaemia, n (%)  26 (17)  25 (18)  1 (9)  BMI, kg/m2  28.1 (25.8–32.6)  28.4 (26–32.8)  25.9 (23–30.5)†  D’Agostino Framingham score  6.4 (3.0–11.7)  6.2 (2.7–11.2)  15.6 (4.9–20.0)*  Coronary artery calcium  0 (0–19)  0 (0–10)  120 (0–361)***  Segment involvement score  1 (0–3)  1 (0–2)  5 (1–7)**  Segment stenosis score  1 (0–4)  1 (0–3)  9 (1–14)**  Plaque burden score  1.5 (0–3)  1 (0–3)  9 (1–12)**  Obstructive plaque (>50%), n (%)  18 (12)  12 (9)  6 (55)****  Non-calcified plaque score  1 (0–2)  1 (0–2)  1 (0–5)  Mixed plaque score  0 (0–1)  0 (0–0)  3 (0–7)*  Calcified plaque score  0 (0–0)  0 (0–0)  1 (0–3)  hs-cTnI, pg/mla  1.5 (1.1–2.6)  1.5 (1–2.4)  2.6 (2.1–4.4)**  IL-17A, pg/mla  1.3 (0.8–1.8)  1.3 (0.8–1.9)  1.2 (0.8–1.5)  IL-17F, pg/mla  35.8 (22.1–66.8)  36.3 (22.1–62)  31 (22.1–97.4)  IL-6, pg/mla  2.9 (1.8–5.8)  2.8 (1.7–5.4)  3.7 (2.6–13.4) †  TNF-α, pg/mla  11.2 (7.8–24)  11.3 (7.7–24)  11.1 (7.9–30.4)  VEGF, pg/mla  134 (61–221)  134 (60–221)  134 (77–179)  Values are presented as median (IQR) unless stated otherwise. † P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001. a Biomarker data were available in 146 patients. Correlations of cytokines and hs-cTnI with occult coronary atherosclerosis TNF-α, IL-6, IL-17A, IL-17F and VEGF showed no correlations with any coronary plaque parameters or hs-cTnI (supplementary Table S1, available at Rheumatology online) and neither did ESR, CRP, tender joint count, swollen joint count or DAS28-CRP (data not shown). hs-cTnI was detectable in all patients [1.5 (IQR 1.1–2.6) pg/ml]; patients with any plaque had higher levels compared to those without [1.8 (IQR 1.1–2.6) pg/ml vs 1.3 (0.9–1.8); P = 0.02]. Moreover, hs-cTnI was correlated with all occult coronary plaque outcomes (all P < 0.01). CAC, SIS, SSS and PBS substantially increased from the lowest to the highest hs-cTnI tertile (P for trend = 0.006, 0.005, 0.01 and 0.009, respectively). Similarly, high-risk plaque outcomes such as CAC >100, SIS >5, SSS >5, obstructive plaque, composite outcome and presence of any advanced MP/CP lesions were considerably enriched across higher hs-cTnI tertiles (Fig. 1). Fig. 1 View largeDownload slide Several high-risk coronary plaque burden outcomes are significantly enriched across higher tertiles of hs-cTnI hs-cTnI tertile ranges were ≤1.2, 1.2–2.1 and ≥2.1 pg/ml. P-values for trend determined by the Jonckheere–Terpstra test. †P < 0.1, *P < 0.05, **P < 0.01, ***P < 0.001. Fig. 1 View largeDownload slide Several high-risk coronary plaque burden outcomes are significantly enriched across higher tertiles of hs-cTnI hs-cTnI tertile ranges were ≤1.2, 1.2–2.1 and ≥2.1 pg/ml. P-values for trend determined by the Jonckheere–Terpstra test. †P < 0.1, *P < 0.05, **P < 0.01, ***P < 0.001. hs-cTnI independently correlates with plaque burden and composition hs-cTnI levels associated with all plaque outcomes (Table 2). After controlling for age and gender (model 1), additional adjustments for hypertension, diabetes, hyperlipidaemia, smoking, BMI and prednisone use (model 2) or FRS-DA score (model 3), hs-cTnI remained predictive of CAC, SSS and PBS. Furthermore, it associated with the presence of more advanced mixed or calcified plaques but showed no correlation with earlier, non-calcified lesions. Importantly, hs-cTnI further correlated with high-risk outcomes such as obstructive plaque, SSS >5, CAC ⩾100 or composite endpoints (Table 2); significance persisted for several, even after adjustments for cardiac risk factors or FRS-DA scores. Table 2 Prediction of occult coronary plaque burden and composition by cTnIa   Unadjusted, OR (95% CI)  Model 1a, OR (95% CI)  Model 2b, OR (95% CI)  Model 3c, OR (95% CI)  CACd  2.95 (1.5, 6.0)**  2.2 (1.0, 4.7)*  2.3 (1.0, 5.3)*  2.7 (1.3, 5.8)*  CAC >100 vs ≤100  4.2 (1.3, 13.5)*  3.0 (0.9, 10.5)†  6.6 (1.4, 30.0)*  5.0 (1.3, 19)*  SISd  2.2 (1.1, 4.2)*  1.7 (0.9, 3.4)  1.7 (0.8, 3.6)  1.9 (1.0, 3.8)†  SIS >5 vs ≤5  2.5 (0.7, 8.5)  1.8 (0.5, 6.6)  2.5 (0.6, 10.5)  2.6 (0.7, 9.9)  SSSd  2.4 (1.3, 4.7)**  1.9 (1.0, 3.9)†  2.0 (1.0, 4.2)†  2.2 (1.1, 4.3)*  SSS >5 vs ≤5  3.0 (1.1, 8.2)*  2.4 (0.8, 7.3)  2.9 (0.9, 9.0)†  3.0 (1.0, 9.0)*  PBSd  3.1 (1.5, 6.3)**  2.4 (1.1, 5.1)*  2.7 (1.2, 6.2)*  2.9 (1.3, 6.2)**  Obstructive plaque  3.9 (1.2, 12.5)*  3.1 (0.9, 10.7)†  3.8 (1.0, 14.4)*  4.0 (1.1, 14.1)*  CAC >100 or SSS >5  4.7 (1.8, 12.4)**  2.3 (0.7, 6.9)  5.3 (1.7, 17.0)**  5.2 (1.8, 15.8)**  SIS >5 or SSS >5 or obstructive plaque  3.2 (1.2, 8.8)*  2.6 (0.9, 7.7)†  3.0 (1.0, 9.4)†  3.3 (1.1, 9.7)*  NCP >0 vs 0  1.4 (0.7, 2.7)  1.3 (0.6, 2.5)  1.3 (0.6, 2.6)  1.4 (0.7, 2.7)  MP/CP >0 vs 0  3.6 (1.8, 7.5)***  2.7 (1.2, 6.0)*  2.9 (1.2, 6.6)*  3.5 (1.6, 7.6)**    Unadjusted, OR (95% CI)  Model 1a, OR (95% CI)  Model 2b, OR (95% CI)  Model 3c, OR (95% CI)  CACd  2.95 (1.5, 6.0)**  2.2 (1.0, 4.7)*  2.3 (1.0, 5.3)*  2.7 (1.3, 5.8)*  CAC >100 vs ≤100  4.2 (1.3, 13.5)*  3.0 (0.9, 10.5)†  6.6 (1.4, 30.0)*  5.0 (1.3, 19)*  SISd  2.2 (1.1, 4.2)*  1.7 (0.9, 3.4)  1.7 (0.8, 3.6)  1.9 (1.0, 3.8)†  SIS >5 vs ≤5  2.5 (0.7, 8.5)  1.8 (0.5, 6.6)  2.5 (0.6, 10.5)  2.6 (0.7, 9.9)  SSSd  2.4 (1.3, 4.7)**  1.9 (1.0, 3.9)†  2.0 (1.0, 4.2)†  2.2 (1.1, 4.3)*  SSS >5 vs ≤5  3.0 (1.1, 8.2)*  2.4 (0.8, 7.3)  2.9 (0.9, 9.0)†  3.0 (1.0, 9.0)*  PBSd  3.1 (1.5, 6.3)**  2.4 (1.1, 5.1)*  2.7 (1.2, 6.2)*  2.9 (1.3, 6.2)**  Obstructive plaque  3.9 (1.2, 12.5)*  3.1 (0.9, 10.7)†  3.8 (1.0, 14.4)*  4.0 (1.1, 14.1)*  CAC >100 or SSS >5  4.7 (1.8, 12.4)**  2.3 (0.7, 6.9)  5.3 (1.7, 17.0)**  5.2 (1.8, 15.8)**  SIS >5 or SSS >5 or obstructive plaque  3.2 (1.2, 8.8)*  2.6 (0.9, 7.7)†  3.0 (1.0, 9.4)†  3.3 (1.1, 9.7)*  NCP >0 vs 0  1.4 (0.7, 2.7)  1.3 (0.6, 2.5)  1.3 (0.6, 2.6)  1.4 (0.7, 2.7)  MP/CP >0 vs 0  3.6 (1.8, 7.5)***  2.7 (1.2, 6.0)*  2.9 (1.2, 6.6)*  3.5 (1.6, 7.6)**  a Adjusted for age and gender. b Adjusted for age, gender, hypertension, diabetes, hyperlipidaemia, smoking, BMI and prednisone use. c Adjusted for D’Agostino Framingham score. d cTnI and coronary plaque outcomes (CAC, SIS, SSS, PBS) are binarized based on the median. † P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001. Table 2 Prediction of occult coronary plaque burden and composition by cTnIa   Unadjusted, OR (95% CI)  Model 1a, OR (95% CI)  Model 2b, OR (95% CI)  Model 3c, OR (95% CI)  CACd  2.95 (1.5, 6.0)**  2.2 (1.0, 4.7)*  2.3 (1.0, 5.3)*  2.7 (1.3, 5.8)*  CAC >100 vs ≤100  4.2 (1.3, 13.5)*  3.0 (0.9, 10.5)†  6.6 (1.4, 30.0)*  5.0 (1.3, 19)*  SISd  2.2 (1.1, 4.2)*  1.7 (0.9, 3.4)  1.7 (0.8, 3.6)  1.9 (1.0, 3.8)†  SIS >5 vs ≤5  2.5 (0.7, 8.5)  1.8 (0.5, 6.6)  2.5 (0.6, 10.5)  2.6 (0.7, 9.9)  SSSd  2.4 (1.3, 4.7)**  1.9 (1.0, 3.9)†  2.0 (1.0, 4.2)†  2.2 (1.1, 4.3)*  SSS >5 vs ≤5  3.0 (1.1, 8.2)*  2.4 (0.8, 7.3)  2.9 (0.9, 9.0)†  3.0 (1.0, 9.0)*  PBSd  3.1 (1.5, 6.3)**  2.4 (1.1, 5.1)*  2.7 (1.2, 6.2)*  2.9 (1.3, 6.2)**  Obstructive plaque  3.9 (1.2, 12.5)*  3.1 (0.9, 10.7)†  3.8 (1.0, 14.4)*  4.0 (1.1, 14.1)*  CAC >100 or SSS >5  4.7 (1.8, 12.4)**  2.3 (0.7, 6.9)  5.3 (1.7, 17.0)**  5.2 (1.8, 15.8)**  SIS >5 or SSS >5 or obstructive plaque  3.2 (1.2, 8.8)*  2.6 (0.9, 7.7)†  3.0 (1.0, 9.4)†  3.3 (1.1, 9.7)*  NCP >0 vs 0  1.4 (0.7, 2.7)  1.3 (0.6, 2.5)  1.3 (0.6, 2.6)  1.4 (0.7, 2.7)  MP/CP >0 vs 0  3.6 (1.8, 7.5)***  2.7 (1.2, 6.0)*  2.9 (1.2, 6.6)*  3.5 (1.6, 7.6)**    Unadjusted, OR (95% CI)  Model 1a, OR (95% CI)  Model 2b, OR (95% CI)  Model 3c, OR (95% CI)  CACd  2.95 (1.5, 6.0)**  2.2 (1.0, 4.7)*  2.3 (1.0, 5.3)*  2.7 (1.3, 5.8)*  CAC >100 vs ≤100  4.2 (1.3, 13.5)*  3.0 (0.9, 10.5)†  6.6 (1.4, 30.0)*  5.0 (1.3, 19)*  SISd  2.2 (1.1, 4.2)*  1.7 (0.9, 3.4)  1.7 (0.8, 3.6)  1.9 (1.0, 3.8)†  SIS >5 vs ≤5  2.5 (0.7, 8.5)  1.8 (0.5, 6.6)  2.5 (0.6, 10.5)  2.6 (0.7, 9.9)  SSSd  2.4 (1.3, 4.7)**  1.9 (1.0, 3.9)†  2.0 (1.0, 4.2)†  2.2 (1.1, 4.3)*  SSS >5 vs ≤5  3.0 (1.1, 8.2)*  2.4 (0.8, 7.3)  2.9 (0.9, 9.0)†  3.0 (1.0, 9.0)*  PBSd  3.1 (1.5, 6.3)**  2.4 (1.1, 5.1)*  2.7 (1.2, 6.2)*  2.9 (1.3, 6.2)**  Obstructive plaque  3.9 (1.2, 12.5)*  3.1 (0.9, 10.7)†  3.8 (1.0, 14.4)*  4.0 (1.1, 14.1)*  CAC >100 or SSS >5  4.7 (1.8, 12.4)**  2.3 (0.7, 6.9)  5.3 (1.7, 17.0)**  5.2 (1.8, 15.8)**  SIS >5 or SSS >5 or obstructive plaque  3.2 (1.2, 8.8)*  2.6 (0.9, 7.7)†  3.0 (1.0, 9.4)†  3.3 (1.1, 9.7)*  NCP >0 vs 0  1.4 (0.7, 2.7)  1.3 (0.6, 2.5)  1.3 (0.6, 2.6)  1.4 (0.7, 2.7)  MP/CP >0 vs 0  3.6 (1.8, 7.5)***  2.7 (1.2, 6.0)*  2.9 (1.2, 6.6)*  3.5 (1.6, 7.6)**  a Adjusted for age and gender. b Adjusted for age, gender, hypertension, diabetes, hyperlipidaemia, smoking, BMI and prednisone use. c Adjusted for D’Agostino Framingham score. d cTnI and coronary plaque outcomes (CAC, SIS, SSS, PBS) are binarized based on the median. † P < 0.1, * P < 0.05, ** P < 0.01, *** P < 0.001. Conversely, subjects with low hs-cTnI (<1.5 pg/ml) were less likely to have extensive coronary atherosclerosis. Specifically, they displayed 81% lower risk of having SSS >5 or CAC ⩾100 and 70% less risk of obstructive plaque, SIS >5 or SSS >5 after controlling for FRS-DA score; the area under the curve (AUC) improved from 0.79 (IQR 0.63–0.95) to 0.85 (0.72–0.98), P < 0.05 (data not shown). Out of all patients, 27 (18%) had CAC >100 or SSS >5 and 22 (15%) had obstructive plaque or SIS >5 or SSS >5. Compared with all patients, only 8% with low hs-cTnI displayed those respective plaque outcomes (supplementary Table S2, available at Rheumatology online); of patients with both low hs-cTnI and low FRS-DA scores, only 4% had extensive atherosclerosis compared with 11% of those with just low FRS-DA. Elevated hs-cTnI associates with long-term CVEs in RA Eleven patients suffered CVEs during 60 (s.d. 26) months of follow-up (1.54/100 patient-years): eight were ischaemic, including one cardiac death, three non-ST elevation MIs, two strokes and two PAD events requiring revascularization; the three non-ischaemic events were new-onset hospitalized heart failure. hs-cTnI was higher in patients with CVEs vs those without [2.6 pg/ml (IQR 2.1–4.4) vs 1.5 (1.0–2.4), P = 0.006]. Elevated hs-cTnI predicted the risk of incident CVEs (Fig. 2A, P = 0.03), independent of demographics and traditional cardiac risk factors (Table 3). Importantly, patients with low hs-cTnI were 82% less likely to suffer a CVE. Table 3 Elevated hs-cTnI (>1.5 pg/ml) predicts the risk of CVEs Model  Hazard ratio (95% CI)  P-value  Unadjusted  4.7 (1.0, 21.7)  0.048  Model 1a  4.8 (1.0, 23.1)  0.052  Model 2b  5.5 (1.1, 26.7)  0.034  Model 3c  4.3 (0.9, 19.7)  0.064  Model  Hazard ratio (95% CI)  P-value  Unadjusted  4.7 (1.0, 21.7)  0.048  Model 1a  4.8 (1.0, 23.1)  0.052  Model 2b  5.5 (1.1, 26.7)  0.034  Model 3c  4.3 (0.9, 19.7)  0.064  a Adjusted for age and gender. b Adjusted for age, gender, hypertension, diabetes, hyperlipidaemia, smoking, BMI and prednisone use. c Adjusted for D’Agostino Framingham score. Table 3 Elevated hs-cTnI (>1.5 pg/ml) predicts the risk of CVEs Model  Hazard ratio (95% CI)  P-value  Unadjusted  4.7 (1.0, 21.7)  0.048  Model 1a  4.8 (1.0, 23.1)  0.052  Model 2b  5.5 (1.1, 26.7)  0.034  Model 3c  4.3 (0.9, 19.7)  0.064  Model  Hazard ratio (95% CI)  P-value  Unadjusted  4.7 (1.0, 21.7)  0.048  Model 1a  4.8 (1.0, 23.1)  0.052  Model 2b  5.5 (1.1, 26.7)  0.034  Model 3c  4.3 (0.9, 19.7)  0.064  a Adjusted for age and gender. b Adjusted for age, gender, hypertension, diabetes, hyperlipidaemia, smoking, BMI and prednisone use. c Adjusted for D’Agostino Framingham score. Fig. 2 View largeDownload slide hs-cTnI optimizes cardiovascular risk prediction in RA (A) Elevated hs-cTnI predicts long-term CVEs in RA. (B) Addition of hs-cTnI information to the FRS-DA composite score increased prognostic accuracy (AUC 0.8431 vs 0.7283; P = 0.1). Further addition of high-risk plaque information from CCTA (obstructive plaque or SIS >5 or SSS >5) resulted in significant enhancement of predictive accuracy over FRS-DA alone (0.9165 vs 0.7283; P = 0.015) and an improving trend over FRS-DA + hs-cTnI (0.9165 vs 0.8431; P = 0.21). Fig. 2 View largeDownload slide hs-cTnI optimizes cardiovascular risk prediction in RA (A) Elevated hs-cTnI predicts long-term CVEs in RA. (B) Addition of hs-cTnI information to the FRS-DA composite score increased prognostic accuracy (AUC 0.8431 vs 0.7283; P = 0.1). Further addition of high-risk plaque information from CCTA (obstructive plaque or SIS >5 or SSS >5) resulted in significant enhancement of predictive accuracy over FRS-DA alone (0.9165 vs 0.7283; P = 0.015) and an improving trend over FRS-DA + hs-cTnI (0.9165 vs 0.8431; P = 0.21). hs-cTnI enhances CVE risk prediction when added to cardiac risk scores The prognostic accuracy of FRS-DA alone vs FRS-DA + hs-cTnI and FRS-DA + hs-cTnI + high-risk plaque for CVEs based on the AUC of the respective receiver operating characteristics curves is depicted in Fig. 2B. The addition of hs-cTnI information to the FRS-DA score yielded a higher prognostic accuracy (0.8431 vs 0.7283; P = 0.10); further addition of high-risk plaque information from CCTA (obstructive plaque or SIS >5 or SSS >5) resulted in significant enhancement of the predictive accuracy over the FRS-DA alone (0.9165 vs 0.7283; P = 0.015) and an improving trend over the FRS-DA + hs-cTnI (0.9165 vs 0.8431; P = 0.21). Since the AUC change in response to a new marker included in a model is often sensitive to only very large independent effects of that marker, we further calculated the IDI to assess additional discrimination offered by inclusion of information from hs-cTnI and high-risk plaque in CVE prediction. Indeed, the addition of hs-cTnI to FRS-DA significantly improved the precision of CVE risk prediction vs FRS-DA alone [Table 4; IDI = 0.0435 (IQR 0.0023–0.0847), P = 0.038]. Further addition of high-risk plaque information significantly enhanced the accuracy of CVE risk prediction over FRS-DA + hs-cTnI [IDI = 0.0818 (IQR 0.0032–0.1605), P = 0.042]. Table 4 Average improvement in precision of CVE risk prediction by integrating hs-cTnI and high-risk CCTA Comparison  IDI (95% CI)  P-value  FRS-DA vs FRS-DA + hs-cTnI  0.0435 (0.0023, 0.0847)  0.038  FRS-DA + hs-cTnI vs FRS-DA + hs-cTnI + high-risk CCTA  0.0818 (0.0032, 0.1605)  0.042  Comparison  IDI (95% CI)  P-value  FRS-DA vs FRS-DA + hs-cTnI  0.0435 (0.0023, 0.0847)  0.038  FRS-DA + hs-cTnI vs FRS-DA + hs-cTnI + high-risk CCTA  0.0818 (0.0032, 0.1605)  0.042  Table 4 Average improvement in precision of CVE risk prediction by integrating hs-cTnI and high-risk CCTA Comparison  IDI (95% CI)  P-value  FRS-DA vs FRS-DA + hs-cTnI  0.0435 (0.0023, 0.0847)  0.038  FRS-DA + hs-cTnI vs FRS-DA + hs-cTnI + high-risk CCTA  0.0818 (0.0032, 0.1605)  0.042  Comparison  IDI (95% CI)  P-value  FRS-DA vs FRS-DA + hs-cTnI  0.0435 (0.0023, 0.0847)  0.038  FRS-DA + hs-cTnI vs FRS-DA + hs-cTnI + high-risk CCTA  0.0818 (0.0032, 0.1605)  0.042  Discussion Patients with RA incur a higher rate of CVEs compared with individuals without autoimmune disease [1]. Therefore, periodic cardiovascular risk stratification according to national guidelines is an integral part of the care of RA patients [26]. However, general risk calculators do not sufficiently capture the incremental risk in patients with RA [27–29]. All stages of the atherogenic process appear enhanced in RA, including endothelial dysfunction, increased arterial stiffness, plaque formation and finally CVEs [30]. Distinct biomarkers may reflect different stages of this pathway, from inflammation (high-sensitivity CRP, IL-6) to plaque instability (myeloperoxidase, MMPs), thrombosis (fibrinogen), myocardial stress (NT-proBNP) and myocardial necrosis (hs-cTn). Individual associations of CRP, hs-cTn and NT-proBNP with CVE in general patients have been extensively described [31]. In RA, CRP may reflect uncontrolled systemic inflammation rather than being a surrogate for the extent of vascular involvement [30]. NT-proBNP independently predicted mortality in one study of 182 RA patients [32]. Our study shows for the first time that hs-cTnI, a specific structural myocardial biomarker, may optimize long-term cardiovascular risk prediction in RA. Blood concentrations of cardiac troponin I and T subunits are elevated in the context of myocardial injury [33]. High-sensitivity assays measure cTnI concentrations at levels much lower than conventional assays with excellent precision at a ⩽10% coefficient of variation, both at and below the assay’s 99th percentile value. This added sensitivity allows reliable estimation in almost 100% of healthy individuals and identification of subclinical myocardial injury [34]. Elevated hs-cTnI was associated with incident long-term CVEs in patients with RA when controlling for traditional cardiac risk factors. This is consistent with reports in population-based studies that subthreshold elevations of either hs-cTnT or hs-cTnI predicted a higher risk of CVEs, heart failure hospitalization and mortality [12–15]. In contrast, RA patients with low hs-cTnI were 82% less likely to suffer a CVE. This approximates the estimated 88% lower risk of CV death in a nested case–control study in general patients with low hs-cTnI measured with the same assay [35]. Moreover, we demonstrated that hs-cTnI measurements significantly improved discrimination of long-term incident CVE risk over composite cardiac risk scores alone. A combination of CRP, NT-proBNP and hs-cTnI optimized the 10-year CVE risk prediction in two general European populations [36]; however, these have not yet been evaluated in RA. In our study, IL-6 was numerically higher in patients incurring CVEs. Nevertheless, a model of high IL-6 combined with hs-cTnI did not optimize event prediction over hs-cTnI alone (data not shown). More multibiomarker groupings will likely emerge in the future. However, the optimal prognostic combinations remain to be defined. Our second novel finding was the association of hs-cTnI with coronary plaque presence, burden and composition in patients with RA, as measured by CCTA. This non-invasive imaging modality has significantly enhanced the prediction of incident CVEs beyond clinical risk scores, as well as CAC in general patients without known CVD [37, 38]. In a prospective study, 69% of subjects with obstructive lesions suffered events at 52 months compared with 28% of those with non-obstructive lesions and 0% of those without plaque. Similarly, 75% with SIS >5 and 80% with SSS >5 suffered CVEs compared with 23% with SIS ⩽5 and 15% with SSS ⩽5 [39]. hs-cTnI was considerably higher in patients with any plaque vs those without; furthermore, it significantly increased across higher plaque burden scores. This is consistent with a prior report in general patients showing progressively higher hs-cTnT in those with mild, moderate and multivessel CAD on CCTA [40]. hs-cTnI was strongly correlated with all quantitative plaque outcomes, including several high-risk ones (obstructive plaque, SSS >5, CAC >100 and composites thereof) after adjustments for traditional risk factors and cardiovascular scores. Moreover, it independently predicted the presence of any advanced—mixed or calcified—coronary plaque, whereas it showed no correlation with earlier non-calcified plaques. In our study, hs-cTnI significantly improved the discrimination of long-term incident CVE risk over cardiac risk scores alone. Additional information on the presence of high-risk plaque outcomes from CCTA further optimized CVE risk discrimination compared with cardiac risk scores and hs-cTnI together. These observations provide the theoretical framework and a testable hypothesis for a two-step algorithm to optimize CVE risk prediction in RA. As part of the cardiac risk stratification, physicians could measure plasma hs-cTnI. If high (>1.5 pg/ml), it may foreshadow a significant hazard for high-risk plaque burden, vulnerability or future CVEs above and beyond cardiac risk scores. In that context, further non-invasive evaluation of coronary atherosclerosis with CCTA may refine primary prevention recommendations based on the presence and burden of coronary plaque. In contrast, if hs-cTnI is low (⩽1.5 pg/ml), the risk of significant coronary atherosclerosis and CVE is substantially decreased. Therefore physicians may narrow their recommendations to address potential actionable clinical risk factors in accordance with cardiac scores. In our study, hs-cTnI was measured at the time of CCTA, when no chest pain was present. In fact, by design, enrolees had no symptoms or diagnosis of CVD upon study entry, hence elevated hs-cTnI levels likely reflect latent myocyte damage. Higher hs-cTnI in general patients has been associated with unstable plaque features on CCTA [41], reflecting intermittent, chronic and clinically silent plaque remodelling and/or rupture with subsequent microembolization, leading to unrecognized MIs (UMIs) [42, 43]. Consistent with these reports, we showed that hs-cTnI in RA patients only correlated with higher-complexity mixed or calcified plaques, independent of cardiac risk factors or Framingham scores but not earlier non-calcified lesions. Greater hs-cTnI associated with the presence of UMIs at baseline, as well as with new or larger UMIs on MRI 5 years later, in a series of community-living volunteers without a history of MI [44]. RA patients are far more likely to experience UMIs, even prior to their RA diagnosis [45]. Indeed, in a pilot MRI study, 39% of RA patients without symptomatic CVD had delayed enhancement suggesting myocardial inflammation or scarring and 11% had nodular subendocardial delayed enhancement indicating silent MI [46]. Latent troponin leak has further been reported as a result of impaired cell membrane integrity due to systemic inflammation [47]. However, we observed no associations between hs-cTnI, inflammatory markers or cytokines, making inflammation an unlikely driver, consistent with an earlier report [16]. Interestingly, we observed no association between pro-inflammatory cytokines, ESR or CRP and the burden of coronary atherosclerosis. This observation may be partially explained by the fact that 58% of our patients were in remission (DAS28-CRP <2.6) at the time of CCTA, while 75% overall had low disease activity (DAS28-CRP <3.2) and 60% were under chronic anti-TNF medication exposure. Concordantly, in the vast majority, IL-6, IL-17A and IL-17F levels were well below the 99% threshold observed in normal patients and similar to or lower than those reported by studies in treated RA patients using identical measurement assays [48, 49]. Our study has certain limitations. Causal relationships between hs-cTnI levels and plaque burden or composition may not be inferred due to their cross-sectional evaluation. Moreover, since our patients were well controlled and the levels of pro-inflammatory cytokines studied were generally low and reflective of that state, we may have underestimated the association of inflammation with both hs-cTnI and plaque burden. Our broader study design, of which the current report is a part, was powered to evaluate quantitative and qualitative plaque differences between 150 RA patients and an equal number of age- and gender-matched patients without autoimmune disease. Although evaluations of biomarkers and their associations with plaque presence, burden and composition in RA patients were pre-specified as exploratory analyses, they were not specifically powered for. Our findings would therefore have to be tested in larger, specifically powered studies and our proposed two-step algorithm for optimization of CVE risk prediction prospectively validated within that context. CVEs appear numerically low in our study [11 patients (6.1%)], which may have deflated overall significance rates—despite sizeable area differences—in AUC curves between FRS-DA alone and FRS-DA + hs-cTnI as well as between FRS-DA + hs-cTnI and FRS-DA + hs-cTnI + high-risk CCTA. This was certainly contributed to by our study design, which pre-specified recruitment of subjects without symptoms or prior diagnosis of CVD. Despite this, our observed event rate was 1.5/100 patient-years, which is similar to studies specifically designed for CV risk [50], and is considered high overall for populations of well-controlled patients chronically exposed to biologic agents. In conclusion, we show for the first time that hs-cTnI associates with the presence, burden and composition of coronary artery atherosclerosis in RA patients without symptoms or prior diagnosis of cardiovascular disease above and beyond traditional risk factors, cardiovascular scores or inflammation. hs-cTnI further associates with the long-term risk of incident CVEs beyond demographics and traditional cardiac risk factors and improves discrimination for such risk prediction beyond that rendered by cardiac risk scores. It may provide a mechanistic explanation for the greater morbidity and mortality RA patients incur and may serve as an adjunct predictive biomarker in refining cardiovascular risk determination in RA. Supplementary data Supplementary data are available at Rheumatology online. Acknowledgements The authors would like to thank Ferdinand Flores, RN for study-related procedure facilitation and blood sample acquisition, handling, storage and shipping. We would also like to express our gratitude to the patients participating in the study. Funding: This work was supported by an American Heart Association grant (09CRP225100 to G.A.K). Disclosure statement: J.E. is employed by Singulex and owns Singulex stock options. J.T. is an employee of Singulex and holds stock options in Singulex. All other authors have declared no conflicts of interest. References 1 Naranjo A, Sokka T, Descalzo MA et al.   Cardiovascular disease in patients with rheumatoid arthritis: results from the QUEST-RA study. Arthritis Res Ther  2008; 10: R30. Google Scholar CrossRef Search ADS PubMed  2 Karpouzas GA, Malpeso J, Choi TY et al.   Prevalence, extent and composition of coronary plaque in patients with rheumatoid arthritis without symptoms or prior diagnosis of coronary artery disease. Ann Rheum Dis  2014; 73: 1797– 804. Google Scholar CrossRef Search ADS PubMed  3 McInnes IB, Buckley CD, Isaacs JD, Cytokines in rheumatoid arthritis – shaping the immunological landscape. Nat Rev Rheumatol  2016; 12: 63– 8. Google Scholar CrossRef Search ADS PubMed  4 Rus HG, Vlaicu R, Niculescu F. Interleukin-6 and interleukin-8 protein and gene expression in human arterial atherosclerotic wall. Atherosclerosis  1996; 127: 263– 71. Google Scholar CrossRef Search ADS PubMed  5 Dixon WG, Symmons DP. What effects might anti-TNFα treatment be expected to have on cardiovascular morbidity and mortality in rheumatoid arthritis? A review of the role of TNFα in cardiovascular pathophysiology. Ann Rheum Dis  2007; 66: 1132. Google Scholar CrossRef Search ADS PubMed  6 Niemann-Jönsson A, Dimayuga P, Jovinge S et al.   Accumulation of LDL in rat arteries is associated with activation of tumor necrosis factor-α expression. Arterioscler Thromb Vasc Biol  2000; 20: 2205– 11. Google Scholar CrossRef Search ADS PubMed  7 Amar J, Fauvel J, Drouet L et al.   Interleukin 6 is associated with subclinical atherosclerosis: a link with soluble intercellular adhesion molecule 1. J Hypertens  2006; 24: 1083– 8. Google Scholar CrossRef Search ADS PubMed  8 Kuo LT, Yang NI, Cherng WJ et al.   Serum interleukin-6 levels, not genotype, correlate with coronary plaque complexity. Int Heart J  2008; 49: 391– 402. Google Scholar CrossRef Search ADS PubMed  9 Ridker PM, Rifai N, Stampfer MJ, Hennekens CH. Plasma concentration of interleukin-6 and the risk of future myocardial infarction among apparently healthy men. Circulation  2000; 101: 1767– 72. Google Scholar CrossRef Search ADS PubMed  10 Biasucci LM, Liuzzo G, Fantuzzi G et al.   Increasing levels of interleukin (IL)-1Ra and IL-6 during the first 2 days of hospitalization in unstable angina are associated with increased risk of in-hospital coronary events. Circulation  1999; 99: 2079– 84. Google Scholar CrossRef Search ADS PubMed  11 Hashmi S, Zeng QT. Role of interleukin-17 and interleukin-17-induced cytokines interleukin-6 and interleukin-8 in unstable coronary artery disease. Coron Artery Dis  2006; 17: 699– 706. Google Scholar CrossRef Search ADS PubMed  12 deFilippi CR, de Lemos JA, Christenson RH et al.   Association of serial measures of cardiac troponin T using a sensitive assay with incident heart failure and cardiovascular mortality in older adults. JAMA  2010; 304: 2494– 502. Google Scholar CrossRef Search ADS PubMed  13 de Lemos JA, Drazner MH, Omland T et al.   Association of troponin T detected with a highly sensitive assay and cardiac structure and mortality risk in the general population. JAMA  2010; 304: 2503– 12. Google Scholar CrossRef Search ADS PubMed  14 Saunders JT, Nambi V, de Lemos JA et al.   Cardiac troponin T measured by a highly sensitive assay predicts coronary heart disease, heart failure, and mortality in the Atherosclerosis Risk in Communities Study. Circulation  2011; 123: 1367– 76. Google Scholar CrossRef Search ADS PubMed  15 Zeller T, Tunstall-Pedoe H, Saarela O et al.   High population prevalence of cardiac troponin I measured by a high-sensitivity assay and cardiovascular risk estimation: the MORGAM Biomarker Project Scottish Cohort. Eur Heart J  2014; 35: 271– 81. Google Scholar CrossRef Search ADS PubMed  16 Bradham WS, Bian A, Oeser A et al.   High-sensitivity cardiac troponin-I is elevated in patients with rheumatoid arthritis, independent of cardiovascular risk factors and inflammation. PLoS One  2012; 7: e38930. Google Scholar CrossRef Search ADS PubMed  17 D’Agostino RBSr, Vasan RS, Pencina MJ et al.   General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation  2008; 117: 743– 53. Google Scholar CrossRef Search ADS PubMed  18 Todd J, Freese B, Lu A et al.   Ultrasensitive flow-based immunoassays using single-molecule counting. Clin Chem  2007; 53: 1990– 5. Google Scholar CrossRef Search ADS PubMed  19 Budoff MJ, Dowe D, Jollis JG et al.   Diagnostic performance of 64-multidetector row coronary computed tomographic angiography for evaluation of coronary artery stenosis in individuals without known coronary artery disease: results from the prospective multicenter ACCURACY (Assessment by Coronary Computed Tomographic Angiography of Individuals Undergoing Invasive Coronary Angiography) trial. J Am Coll Cardiol  2008; 52: 1724– 32. Google Scholar CrossRef Search ADS PubMed  20 Agatston AS, Janowitz WR, Hildner FJ et al.   Quantification of coronary artery calcium using ultrafast computed tomography. J Am Coll Cardiol  1990; 15: 827– 32. Google Scholar CrossRef Search ADS PubMed  21 Austen WG, Edwards JE, Frye RL et al.   A reporting system on patients evaluated for coronary artery disease. Report of the Ad Hoc Committee for Grading of Coronary Artery Disease, Council on Cardiovascular Surgery, American Heart Association. Circulation  1975; 51: 5– 40. Google Scholar CrossRef Search ADS PubMed  22 Pagali SR, Madaj P, Gupta M et al.   Interobserver variations of plaque severity score and segment stenosis score in coronary arteries using 64 slice multidetector computed tomography: a substudy of the ACCURACY trial. J Cardiovasc Comput Tomogr  2010; 4: 312– 8. Google Scholar CrossRef Search ADS PubMed  23 Thygesen K, Alpert JS, Jaffe AS et al.   Third universal definition of myocardial infarction. Glob Heart  2012; 7: 275– 95. Google Scholar CrossRef Search ADS PubMed  24 Goldstein LB, Adams R, Alberts MJ et al.   Primary prevention of ischemic stroke: a guideline from the American Heart Association/American Stroke Association Stroke Council: cosponsored by the Atherosclerotic Peripheral Vascular Disease Interdisciplinary Working Group; Cardiovascular Nursing Council; Clinical Cardiology Council; Nutrition, Physical Activity, and Metabolism Council; and the Quality of Care and Outcomes Research Interdisciplinary Working Group: the American Academy of Neurology affirms the value of this guideline. Stroke  2006; 37: 1583– 633. Google Scholar CrossRef Search ADS PubMed  25 Rooke TW, Hirsch AT, Misra S et al.   2011 ACCF/AHA focused update of the guideline for the management of patients with peripheral artery disease (updating the 2005 guideline): a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol  2011; 58: 2020– 45. Google Scholar CrossRef Search ADS PubMed  26 Agca R, Heslinga SC, Rollefstad S et al.   EULAR recommendations for cardiovascular disease risk management in patients with rheumatoid arthritis and other forms of inflammatory joint disorders: 2015/2016 update. Ann Rheum Dis  2017; 76: 17– 28. Google Scholar CrossRef Search ADS PubMed  27 Crowson CS, Matteson EL, Roger VL et al.   Usefulness of risk scores to estimate the risk of cardiovascular disease in patients with rheumatoid arthritis. Am J Cardiol  2012; 110: 420– 4. Google Scholar CrossRef Search ADS PubMed  28 Arts EE, Popa C, Den Broeder AA et al.   Performance of four current risk algorithms in predicting cardiovascular events in patients with early rheumatoid arthritis. Ann Rheum Dis  2015; 74: 668– 74. Google Scholar CrossRef Search ADS PubMed  29 Siontis GC, Tzoulaki I, Siontis KC, Ioannidis JP. Comparisons of established risk prediction models for cardiovascular disease: systematic review. BMJ  2012; 344: e3318. Google Scholar CrossRef Search ADS PubMed  30 Kozera L, Andrews J, Morgan AW. Cardiovascular risk and rheumatoid arthritis—the next step: differentiating true soluble biomarkers of cardiovascular risk from surrogate measures of inflammation. Rheumatology  2011; 50: 1944– 54. Google Scholar CrossRef Search ADS PubMed  31 Lacey B, Herrington WG, Preiss D, Lewington S, Armitage J. The role of emerging risk factors in cardiovascular outcomes. Curr Atheroscler Rep  2017; 19: 28. Google Scholar CrossRef Search ADS PubMed  32 Provan S, Angel K, Semb AG, Atar D, Kvien TK. NT-proBNP predicts mortality in patients with rheumatoid arthritis: results from 10-year follow-up of the EURIDISS study. Ann Rheum Dis  2010; 69: 1946– 50. Google Scholar CrossRef Search ADS PubMed  33 Gupta S, de Lemos JA. Use and misuse of cardiac troponins in clinical practice. Prog Cardiovasc Dis  2007; 50: 151– 65. Google Scholar CrossRef Search ADS PubMed  34 Januzzi JLJr, Bamberg F, Lee H et al.   High-sensitivity troponin T concentrations in acute chest pain patients evaluated with cardiac computed tomography. Circulation  2010; 121: 1227– 34. Google Scholar CrossRef Search ADS PubMed  35 Apple FS, Steffen LM, Pearce LA, Murakami MM, Luepker RV. Increased cardiac troponin I as measured by a high-sensitivity assay is associated with high odds of cardiovascular death: the Minnesota Heart Survey. Clin Chem  2012; 58: 930. Google Scholar CrossRef Search ADS PubMed  36 Blankenberg S, Zeller T, Saarela O et al.   Contribution of 30 biomarkers to 10-year cardiovascular risk estimation in 2 population cohorts: the MONICA, risk, genetics, archiving, and monograph (MORGAM) biomarker project. Circulation  2010; 121: 2388– 97. Google Scholar CrossRef Search ADS PubMed  37 Hou ZH, Lu B, Gao Y et al.   Prognostic value of coronary CT angiography and calcium score for major adverse cardiac events in outpatients. JACC Cardiovasc Imaging  2012; 5: 990– 9. Google Scholar CrossRef Search ADS PubMed  38 Budoff MJ, Shaw LJ, Liu ST et al.   Long-term prognosis associated with coronary calcification: observations from a registry of 25, 253 patients. J Am Coll Cardiol  2007; 49: 1860– 70. Google Scholar CrossRef Search ADS PubMed  39 Andreini D, Pontone G, Mushtaq S et al.   A long-term prognostic value of coronary CT angiography in suspected coronary artery disease. JACC Cardiovasc Imaging  2012; 5: 690– 701. Google Scholar CrossRef Search ADS PubMed  40 Laufer EM, Mingels AM, Winkens MH et al.   The extent of coronary atherosclerosis is associated with increasing circulating levels of high sensitive cardiac troponin T. Arterioscler Thromb Vasc Biol  2010; 30: 1269– 75. Google Scholar CrossRef Search ADS PubMed  41 Altintas S, Cardinaels EP, Versteylen MO et al.   Unstable coronary plaque characteristics are associated with high-sensitivity cardiac troponin T and N-terminal Pro-Brain Natriuretic Peptide. J Cardiovasc Comput Tomogr  2016; 10: 82– 8. Google Scholar CrossRef Search ADS PubMed  42 Korosoglou G, Lehrke S, Mueller D et al.   Determinants of troponin release in patients with stable coronary artery disease: insights from CT angiography characteristics of atherosclerotic plaque. Heart  2011; 97: 823– 31. Google Scholar CrossRef Search ADS PubMed  43 Seifarth H, Schlett CL, Lehman SJ et al.   Correlation of concentrations of high-sensitivity troponin T and high-sensitivity C-reactive protein with plaque progression as measured by CT coronary angiography. J Cardiovasc Comput Tomogr  2014; 8: 452– 8. Google Scholar CrossRef Search ADS PubMed  44 Barbier CE, Themudo R, Bjerner T et al.   Cardiac troponin I associated with the development of unrecognized myocardial infarctions detected with MRI. Clin Chem  2014; 60: 1327– 35. Google Scholar CrossRef Search ADS PubMed  45 Maradit-Kremers H, Crowson CS, Nicola PJ et al.   Increased unrecognized coronary heart disease and sudden deaths in rheumatoid arthritis: a population-based cohort study. Arthritis Rheum  2005; 52: 402– 11. Google Scholar CrossRef Search ADS PubMed  46 Kobayashi Y, Giles JT, Hirano M et al.   Assessment of myocardial abnormalities in rheumatoid arthritis using a comprehensive cardiac magnetic resonance approach: a pilot study. Arthritis Res Ther  2010; 12: R171. Google Scholar CrossRef Search ADS PubMed  47 Levine B, Kalman J, Mayer L, Fillit HM, Packer M. Elevated circulating levels of tumor necrosis factor in severe chronic heart failure. N Engl J Med  1990; 323: 236– 41. Google Scholar CrossRef Search ADS PubMed  48 Todd J, Simpson P, Estis J, Torres V, Wub AH. Reference range and short- and long-term biological variation of interleukin (IL)-6, IL-17A and tissue necrosis factor-alpha using high sensitivity assays. Cytokine  2013; 64: 660– 5. Google Scholar CrossRef Search ADS PubMed  49 Jain M, Attur M, Furer V et al.   Increased plasma IL-17F levels in rheumatoid arthritis patients are responsive to methotrexate, anti-TNF, and T cell costimulatory modulation. Inflammation  2015; 38: 180– 6. Google Scholar CrossRef Search ADS PubMed  50 Giles JT, Sattar N, Gabriel SE et al.   Results of a randomized, parallel-group, multicenter, noninferiority, phase 4 clinical trial [abstract]. Arthritis Rheumatol  2016; 68(Suppl 10):abstract 3L. http://acrabstracts.org/abstract/comparative-cardiovascular-safety-of-tocilizumab-vs-etanercept-in-rheumatoid-arthritis-results-of-a-randomized-parallel-group-multicenter-noninferiority-phase-4-clinical-trial/ (21 September 2017, date last accessed). © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

RheumatologyOxford University Press

Published: Mar 14, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off