Coronary artery calcification is prevalent in systemic sclerosis and is associated with adverse prognosis

Abstract

Objective:

Coronary artery calcification assessed on thoracic computed tomography represents the calcific component of established coronary artery disease, is a biomarker of total atheromatous plaque burden and predicts mortality. Systemic sclerosis is a pro-inflammatory condition, and inflammation is also a driver of coronary artery disease. We assessed coronary artery calcification prevalence, mortality risk and potential clinical impact on primary prevention in a cohort of patients with systemic sclerosis, differentiated by clinical phenotype including the presence of interstitial lung disease and pulmonary arterial hypertension.

Methods:

Retrospective analysis of 258 computed tomographies in systemic sclerosis patients from three prospectively maintained clinical and research databases at a single tertiary rheumatology/pulmonary hypertension (PH) service between March 2007 and September 2020 (mean age = 65 ± 12, 14% male). Co-morbidities, statin prescription and all-cause mortality were recorded. Patients were subtyped according to underlying systemic sclerosis complications. Computed tomographies were re-reviewed for coronary artery calcification; severity was graded using a 4-point scale per vessel and summed for total coronary artery calcification score. The impact of reporting coronary artery calcification was assessed against pre-existing statin prescriptions.

Results:

Coronary artery calcification was present in 58% (149/258). Coronary artery calcification was more prevalent in systemic sclerosis-pulmonary arterial hypertension than in systemic sclerosis subgroups with interstitial lung disease or without pulmonary arterial hypertension, controlling for age, sex, co-morbidities and smoking status (71%; χ²(13) = 81.4; p < 0.001). The presence and severity of coronary artery calcification were associated with increased risk of mortality independently of age and co-morbidities (hazard ratio = 2.8; 95% confidence interval = 1.2–6.6; p = 0.018). The ‘number needed to report’ coronary artery calcification presence to potentially impact management was 3.

Conclusions:

Coronary artery calcification is common in systemic sclerosis. Coronary artery calcification predicts mortality independently of age and confounding co-morbidities which suggests this finding has clinical relevance and is a potential target for screening and therapeutic intervention.

Introduction

Coronary artery disease (CAD) is a progressive, inflammatory disorder in which calcification forms as atherosclerotic plaque heals.^1,2 Coronary artery calcification (CAC) on thoracic computed tomography (CT) represents the calcific component of established CAD and is a biomarker of total atheromatous plaque burden.^{3 –5} Ordinal coronary artery calcification score (CACS) on non–cardiac-gated thoracic CT is associated with mortality across all age groups in unselected populations, as well as in specific patient cohorts including pulmonary embolism, chronic obstructive pulmonary disease, bronchiectasis and lung cancer screening.^{6 –10} British Society of Cardiovascular Imaging/Thoracic Imaging (BSCI/BSTI) guidance recommends routine reporting of CAC on all non–cardiac-gated thoracic CTs, but it is infrequently reported in routine clinical practice.^11,12 Early identification of asymptomatic CAD allows assessment of modifiable cardiovascular risk factors and initiation of medical therapy including primary prevention in the form of statins.^13,14 Studies have demonstrated that opportunistic CAC screening prompts clinicians to prescribe preventive medication and patient visualisation of cardiovascular imaging is associated with improved individual risk factors.^15,16

There are no large cohort studies examining the prevalence of CAD in patients with systemic sclerosis (SSc), especially those not specifically reporting cardiac symptoms at the time of investigation. ¹⁷ Patients with SSc are at increased risk of CAD and often undergo thoracic CT as part of their routine care to exclude complications of the disease, such as interstitial lung disease (ILD) and pulmonary arterial hypertension (PAH). The early identification of asymptomatic CAD enables review of modifiable risk factors and primary prevention. ¹⁸ There is, therefore, a potential for opportunistic screening in this patient cohort.

The characteristic fibrosis of skin, vasculature and other organs observed in SSc is autoimmune-triggered and associated with inflammation.^19,20 Inflammation is also known to drive CAD.^1,21,22 Numerous inflammatory markers are implicated in the pathogenesis of both conditions, including tumour necrosis factor (TNF)-alpha, interleukin (IL)-6 and C-reactive protein (CRP). ¹⁷ An accelerated atherosclerosis is observed in other inflammatory conditions, such as rheumatoid arthritis and systemic lupus erythematosus (SLE).^23,24 We hypothesised a high prevalence of CAC in SSc and that the presence and severity of CAC would predict mortality. We also evaluated whether CAC prevalence varied with the presence of SSc sequelae, such as PAH and ILD.

This study aims to determine the prevalence of CAC in SSc, define its prognostic ability and evaluate its potential role in opportunistic screening for primary prevention.

Materials and methods

Study design

Three prospective clinical databases were used to retrospectively identify 258 patients with SSc who had undergone thoracic CT at a single institution between March 2007 and September 2020 (Figure 1). The three databases comprised as follows:

Figure 1.

Exclusion flow chart. Patients excluded if (a) no CT chest imaging, (b) CT chest was not of diagnostic quality for CAC or ILD or (c) insufficient diagnostic or clinical information available to allow subtyping.

SSc: systemic sclerosis; CAC: coronary artery calcification; ILD: interstitial lung disease; PH: pulmonary hypertension; PAH: pulmonary arterial hypertension (pre-capillary pulmonary hypertension).

1. A research database of consecutive patients with connective tissue diseases prospectively recruited with written consent (all comers with connective tissue disease who gave consent, refined to exclude patients without SSc).

2. A prospective clinical database of all consecutive patients referred to a tertiary/shared care PH service (all comers, refined to exclude patients without SSc).

3. A prospective clinical database of all patients referred to ILD MDT (all comers, refined to exclude patients without SSc).

Ethical approval was obtained through Integrated Research Application System (IRAS) (ID 284089); informed consent was not required.

Electronic record review

Electronic patient records were screened for demographics, clinical features of SSc, known clinical cardiovascular disease (CAD) (defined by history of ischaemic heart disease and/or cardiac stent), smoking status and statin prescription at the time of imaging. Co-morbidities (hypertension, diabetes and obesity (body mass index (BMI) > 30), dysrhythmia) were recorded as documented in electronic patient records/clinic letters contemporaneous with CT. Renal impairment was defined as estimated glomerular filtration rate (eGFR) < 60 mL/min. Serum CRP was recorded if taken within 3 months of CT. All included patients fulfilled American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) 2013 classification criteria for SSc. ²⁵

All-cause mortality was recorded via the UK national healthcare electronic mortality record to derive days to death from CT. All patients were followed up for a minimum of 3 years: survival time was censored at 3 years.

Patients were split into five subtypes using data from CT, echocardiogram, right heart catheterisation (RHC) and pulmonary function tests. Diagnostic cut-offs for echocardiogram and RHC were as defined in European Society of Cardiology/European Respiratory Society (ESC/ERS) guidelines for diagnosis of PH. ²⁶

1. SSc without ILD or PH: patients without evidence of ILD on CT, and without evidence of PH on echo or RHC. All underwent yearly screening for PH and clinical and imaging screening for ILD (n = 86).

2. SSc with Group 2 PH: patients who underwent RHC with mean pulmonary artery pressure (mPAP) > 20, pulmonary arterial wedge pressure (PAWP) > 15 and pulmonary vascular resistance (PVR) < 2. Also includes patients with mPAP > 20, PAWP > 15 and PVR > 2 but with significant valvular or LV dysfunction (n = 36).

3. SSc-ILD no PH: patients with CT evidence of ILD, without invasive or non-invasive evidence of PH as per ESC/ERS guidelines (n = 39). ²⁶

4. SSc-PAH: patients with PAH without clinically significant ILD. Patients underwent RHC diagnostic of pre-capillary PAH as per ERS/ESC 2022 guidelines (mPAP > 20, PAWP < 15 and PVR > 2). ²⁶ Patients with significant valvular or left ventricular dysfunction were excluded from this group, as were all patients with CT evidence of pulmonary fibrosis with forced vital capacity (FVC) < 70% (n = 73).

5. SSc with ILD and PH: all patients with mPAP > 20 (or evidence of PH on echocardiogram if they had not undergone RHC as per ESC/ERS guidelines) ²⁶ who also had clinically significant ILD (defined as evidence of fibrosis on CT and FVC < 70%) (n = 24).

Thoracic CTs

Patients underwent clinically indicated thoracic CTs for a variety of clinical reasons: to exclude thrombus, to assess the presence of ILD or to investigate breathlessness. Where more than one CT had been performed during the analysis period, the earliest available was selected. A CT was excluded from analysis if it was not of diagnostic quality to allow coronary artery grading and assessment for ILD (Figure 1).

All imaging was obtained using routine acquisition parameters on either a Siemens Definition Edge or Drive scanner (Siemens Healthineers, Erlangen, Germany) with suspended respiration from lung apices to bases. Acquisition protocols were as follows: CT thorax: 120 kV with tube potential modulation, automated tube current modulation with 66 quality reference mAs, pitch of 0.6, rotation time of 0.5 s on the Definition Edge and 0.28 s on the Drive, 128 × 0.6 mm acquisition matrix, 60 mL Omnipaque 350 at 3 mL/s (if a contrast-enhanced acquisition); CTPA: 120 kV with tube potential modulation, automated tube current modulation with 66 quality reference mAs, pitch of 1, 0.5 s rotation time on the Definition Edge and 0.28 s on the Drive, 128 × 0.6 mm acquisition matrix, 60 mL Omnipaque 350 at 5 mL/s with bolus tracking and threshold trigger at 100 HU; HRCT: 120 kV with tube potential modulation, automated tube current modulation with 66 quality reference mAs, pitch of 0.6, 0.5 s rotation time on the Definition Edge and 0.28 s on the Drive, 128 × 0.6 mm acquisition matrix.

Ordinal assessment of CAC

CTs were re-reviewed for CAC by three radiologists with >4 years of experience and blinded to underlying SSc subtype. Semi-quantitative assessment of CAC was performed using an ordinal scale on axial images from CT, a method previously validated in CT pulmonary angiography (CTPAs) and high-resolution CT (HRCT) (i.e. in non-dedicated coronary artery studies) and which follows current British Society of Cardiovascular CT (BSCI/BSCCT) and BSTI recommendations.^{5 –8,27} Each vessel was assigned a score of 0 (none), 1 (mild), 2 (moderate) or 3 (severe) CAC. This was assessed in left main, left anterior descending, left circumflex and right coronary artery. Individual vessel results were summed to give total CACS: mild (1–3), moderate (4–8) or severe (9–12) (Figure 2).^6,9

Figure 2.

Ordinal scoring of CAC on non-gated CT, demonstrating (a) no, (b) mild, (c) moderate and (d) severe calcification in the left anterior descending coronary artery.

Statistical analysis

Statistical analysis was performed in IBM SPSS Statistics, version 27 (IBM Corp., Armonk, NY, USA). Continuous variables were assessed for normality using the Shapiro–Wilk test. Median values were compared via the Wilcoxon rank sum test and the Kruskal–Wallis test. The Pearson chi-square tests were used to analyse differences in categorical variables. Cox proportional hazards regression and Kaplan–Meier (KM) curves were constructed and compared with log-rank test. Full regression models are available in the Supplemental material (S2–S6). A logistic regression assessed the effect of sex, age and co-morbidities on the likelihood of the presence of CAC on CT. Parameters were included in multivariate analysis if they were statistically significant in univariate analysis. Hazard ratios (HRs) and 95% confidence intervals (CIs) are presented.

The potential impact of recording CAC on clinical management was assessed against a patient’s history of a statin prescription prior to the reporting of incidental CAC. A number needed to report outcome was designed using number needed to treat (NNT) analysis, as previously described. ⁹ A preventable adverse outcome was defined as the absence of a statin prescription in a patient whose CT demonstrated evidence of CAC.^{9,13,14,18,28} This was used to generate an assessment of the number of patients whose CT would require CAC reporting to identify one patient with CAC on CT who was not currently prescribed a statin. This was performed on the whole cohort and on each SSc subgroup.

A statistically significant difference was defined as a two-sided p < 0.05.

Results

Following exclusions (Figure 1), 258 patients (mean age = 65 ± 12, 86% female) were included in the analysis. In total, 132 CTPAs, 39 CTs with contrast, 83 non-contrast CT and 4 CT angiograms were included. Demographics, SSc phenotypic information and cardiovascular risk factors are presented in Table 1. There was no significant difference in age or sex between subtypes of SSc (p = 0.255; p = 0.399). Patients with SSc-PAH were more likely than other subtypes to have known CAD and be prescribed a statin; patients with Group 2 PH had a higher rate of dysrhythmia (Table 1). There was no association between SSc disease duration and the presence of CAC (p = 0.18). Patients with ILD and/or PH were more likely to have received immunomodulatory therapy (Table 1). There was no significant association between the presence of CAC on CT and immunomodulatory therapy (p = 0.3). Data for CRP within 3 months of CT were available for 59% (153/259). Serum CRP was significantly lower in those without ILD or PAH (z = −3.4, p ⩽ 0.001). There was no significant difference in CRP with CAC versus without CAC (p = 0.2).

Table 1.

Demographics and clinical characteristics sub-divided by SSc and CAC.

	SSc subtype
	All patients			SSc-No PH no ILD		SSc Group 2 PH		SSc-ILD no PH		SSc-PAH		SSc-ILD with PH		p a
		CAC+	CAC−	CAC+	CAC−	CAC+	CAC−	CAC+	CAC−	CAC+	CAC−	CAC+	CAC−
	n = 258	n = 149	n = 109	n = 39	n = 47	n = 28	n = 8	n = 20	n = 19	n = 52	n = 21	n = 11	n = 13
Demographics
Female, n (%)	86 (222)	86 (128)	86 (94)	87 (34)	87 (41)	75 (21)	88 (7)	76 (16)	79 (15)	90 (47)	90 (19)	73 (8)	85 (11)	0.255
Age at time of CT (SD)	65 (12)	69 (9)	59 (13)	72 (7)	59 (13)	69 (9)	58 (19)	70 (8)	57 (11)	65 (10)	60 (13)	74 (8)	64 (16)	0.399
Median years from first non-Raynaud’s symptom to CT (IQR)	5 (7)	5 (8)	5 (7)	5 (7)	6 (8)	5 (8)	8 (4)	5 (8)	8 (4)	3 (7)	3 (4)	5 (7)	7 (5)	0.18
Scleroderma subtype
Limited cutaneous, n (%)	82 (211)	79 (118)	83 (91)	85 (33)	83 (39)	79 (22)	88 (7)	75 (15)	89 (17)	81 (42)	81 (17)	36 (4)	85 (11)	0.54
Diffuse, n (%)	5 (14)	5 (7)	6 (7)	8 (3)	4 (2)	4 (1)	0 (0)	10 (2)	11 (2)	2 (1)	5 (1)	0 (0)	8 (1)
Other, n (%)	13 (33)	16 (24)	10 (11)	8 (3)	13 (6)	18 (5)	13 (1)	15 (3)	0 (0)	17 (9)	14 (3)	64 (7)	0 (0)
Anti-centromere antibodies, n (%)	49 (119)	48 (72)	43 (47)	54 (16)	55 (26)	57 (16)	25 (2)	10 (2)	16 (3)	58 (29)	71 (12)	27 (3)	8 (1)	<0.001 b
Co-morbidities
Diabetes	7 (19)	8 (12)	6 (7)	8 (3)	2 (1)	4 (1)	13 (1)	15 (3)	16 (3)	8 (4)	5 (1)	9 (1)	8 (1)	0.304
Hypertension	32 (83)	38 (57)	24 (26)	23 (9)	23 (11)	57 (16)	25 (2)	45 (9)	26 (5)	35 (18)	24 (5)	45 (5)	23 (3)	0.07
Smoking status
Current, n (%)	4 (11)	3 (4)	6 (7)	3 (1)	9 (4)	0 (0)	0 (0)	0 (0)	0 (0)	6 (3)	10 (2)	0 (0)	8 (1)	0.011 c
Ex, n (%)	23 (60)	33 (49)	10 (11)	18 (7)	2 (1)	39 (11)	25 (2)	50 (10)	16 (3)	38 (20)	14 (3)	9 (1)	15 (2)
Obesity, n (%)	11 (29)	11 (17)	11 (12)	5 (2)	4 (2)	11 (3)	25 (2)	0 (0)	16 (3)	21 (11)	14 (3)	9 (1)	15 (2)	0.058
Known CAD, n (%)	19 (48)	28 (41)	6 (7)	18 (7)	9 (4)	18 (5)	0 (0)	20 (4)	0 (0)	4 (22)	14 (3)	18 (2)	0 (0)	0.002 d
Dysrhythmia, n (%)	16 (41)	13 (19)	5 (5)	10 (4)	4 (2)	29 (8)	0 (0)	5 (1)	0 (0)	12 (6)	14 (3)	0 (0)	0 (0)	0.016 e
Stroke, n (%)	7 (18)	11 (16)	2 (2)	8 (3)	0 (0)	7 (2)	0 (0)	5 (1)	0 (0)	19 (10)	5 (1)	0 (0)	0 (0)	0.032 f
Renal disease, n (%)	30 (78)	32 (47)	24 (26)	23 (9)	17 (8)	18 (5)	25 (2)	25 (5)	5 (1)	52 (27)	48 (10)	9 (1)	38 (5)	<0.001 g
Prescribed a statin, n (%)	36 (92)	46 (69)	21 (23)	51 (20)	19 (9)	32 (9)	0 (0)	45 (9)	26 (5)	58 (30)	33 (7)	9 (1)	15 (2)	0.005 h
Prescribed immunomodulatory therapy, n (%)	41 (107)	39 (58)	45 (49)	31 (12)	30 (14)	46 (13)	25 (2)	60 (12)	74 (14)	23 (12)	33 (7)	82 (9)	92 (12)	<0.001 i
Prescribed vasoactive therapy, n (%)	74 (190)	77 (115)	69 (75)	77 (30)	53 (25)	75 (21)	88 (7)	75 (15)	79 (15)	83 (43)	81 (17)	55 (6)	85 (11)	0.11
CRP (median (IQR)	4 (8)	5 (8)	4 (8)	3 (3)	2 (4)	5 (8)	7 (63)	7 (11)	8 (6)	4 (9)	10 (12)	8 (4)	5 (20)	0.015 j

PH: pulmonary hypertension; ILD: interstitial lung disease; CAC: coronary artery calcification; CT: computed tomography; SD: standard deviation; IQR: interquartile range; CAD: coronary artery disease.

^ap-value reflects comparisons across subtypes of SSc, including those with and without CAC within each subtype.

^bSSc-PAH p = 0.007, SSc-ILD no PH p < 0.001, SSc ILD with PH p = 0.006.

^cSSc no PAH no ILD p = 0.002, SSc Group 2 PH p = 0.029.

^dSSc-PAH p < 0.001.

^eSSc Group 2 PH p = 0.007.

^fSSc-PAH p = 0.001.

^gSSc-PAH p < 0.001, SSc ILD no PH p = 0.028.

^hSSc-PAH p = 0.002, SSc ILD with PH p = 0.013.

ⁱSSc ILD No PH p ≤ 0.001, SSc-PAH p = 0.002, SSc ILD with PH p ≤ 0.001. CAC vs no CAC across cohort p = 0.3.

^jSSc no PAH no ILD p ≤ 0.001. Data for CRP within 3 months of CT available for 59% (153/258).

CAC

CAC was present in 149 (58%) of 258 of CTs. Median total CACS was 1 (interquartile range (IQR) = 0–3). Patients with CAC versus without CAC were older (69 ± 9 vs 59 ± 13 years, p = < 0.001). Where present, the majority had mild CACS (94/149, 63%); 30% (44/149) had a moderate score, and only 7% (11/149) had severe CACS (Figure 3). The proportion of male patients did not vary significantly between severity categories of CACS (mild = 9% male (8/94), moderate = 25% (11/44), severe = 18% (2/11), none = 14% (15/109), χ²(3) = 7.0, p = 0.73). CAC was identified in 72% of patients without a pre-existing diagnosis of CAD (108/149: Supplemental Figure S1).

Figure 3.

Breakdown of CAC presence and severity across all SSc subtypes.

In total, 71% (52/73) of patients with SSc-PAH had CAC, compared with 51% (20/39) of those with SSc-ILD (Figure 3). A logistic regression was carried out to assess the effect of sex, age diabetes, hypertension, smoking status, obesity, renal disease and sequelae of SSc (grouped as stated in Methods) on the likelihood of the presence of CAC on CT. The overall model was statistically significant when compared to the null model (χ²(13) = 81.4, p < 0.001), explained 36% of the variation (Nagelkerke R²) and correctly predicted 75.6% of cases. Age (p < 0.001), smoking status (p = 0.003), the presence of SSc-PAH (p < 0.001) and Group 2 PH (p = 0.002) were significant but sex, all other co-morbidities and the presence of ILD were not.

Of those with CAC on their CT (n = 149/258, 58%), 5 (5/149, 3%) were aged <50. Of these five patients who were aged less than 50 and had CAC on CT, 2 (40%) were not known to have a history of CAD and 1 (20%) was not prescribed a statin.

Statins: potential impact on clinical management

In all subgroups of SSc, patients were identified in whom CAC indicated the presence of undiagnosed CAD. Where CAC is reported on all CTs in this patient cohort (with and without known CAD), the number needed to report (NNR) to potentially impact management across all subtypes is three. This ranged from two patients in the Group 2 PH group and the SSc with ILD + PH patient groups, to five in those with SSc only. For those with SSc-PAH, NNR was 3, and in ILD without PH, it was 4. A breakdown of CAC presence per subgroup, sub-divided by statin prescription, is presented in Figure 4.

Figure 4.

Clinical impact of identifying CAC across the cohort and in each subgroup, highlighting patients prescribed a statin versus those not.

Reviewing only CTs with CAC on CT, NNR is 2 across the whole cohort and across all SSc subgroups, except ILD with PH (NNR = 1).

Anti-platelet therapy

Antiplatelet medications were prescribed for 20% (52/258) of the whole cohort and 28% (41/149) of those with CAC on CT.

Mortality

All-cause mortality occurred within 3 years of CT in 45/258 (17%) patients. Across the whole cohort, after adjusting for confounders (age, sex, diabetes, hypertension, obesity, renal impairment and smoking status), CAC presence was associated with >2 times risk of all-cause mortality (HR = 2.3; 95% CI = 1.1–4.9; p = 0.033; Supplemental Table S2). After adjusting for the same confounders, increasing severity grade of CAC was associated with 2.8 times increased risk of all-cause mortality (HR = 2.8; 95% CI = 1.2–6.6; p = 0.018; Supplemental Table S3).

When patients with clinical atherosclerotic disease (history of stroke or known clinical CAD, n = 51) were excluded, adjusting for the same confounders, the presence of CAC was associated with 2.5 times increased risk of all-cause mortality (n = 207; HR = 2.5; 95% CI = 1.1–6; p = 0.03; see Supplemental Table S4).

The impact of CAC on all-cause mortality remained significant when underlying complications of SSc (ILD, PAH, Group 2 PH) were also included in the model (HR = 2.3; 0.5% CI = 1.1–5.0; p = 0.037; see Supplemental Table S5). When patients with SSc-PAH or ILD were excluded from analysis, the presence of moderate–severe CAC was associated with 8.0 times risk of all-cause mortality compared with mild or no CAC when adjusted for the same confounders (HR = 8.0; 95% CI = 1.5–43.3; p = 0.015; see Supplemental Table S6).

KM analysis is presented in Figure 5.

Figure 5.

Kaplan–Meier curve demonstrating risk of all-cause mortality against: (a) CAC presence across the whole cohort. Log-rank (Mantel-Cox) (a): χ² = 9; df = 1; p = 0.003, (b) CAC presence across patients without PAH. Log-rank (Mantel-Cox) (b): χ² = 9; df = 1; p = 0.003, (c) CAC severity across the whole cohort. Log-rank (Mantel-Cox) (c): χ² = 11; df = 1; p < 0.001 and (d) CAC severity across patients without PAH. Log-rank (Mantel-Cox) (d): χ² = 11; df = 2; p = 0.004.

CAC: coronary artery calcification.

Discussion

To the best of our knowledge, this is one of the first, and largest, studies to report the prevalence of CAC in an unselected SSc population referred for non-gated, non-cardiac CT chest imaging, and the first to do so across well-phenotyped SSc subtypes. Four previous prospective studies have analysed the use of cardiac coronary calcium scoring in CT in patients with SSc, but the study populations were relatively small (n = 20, n = 17, n = 53, n = 32, n = 67).^{29 –33} The present larger, retrospective study found that the detection and grading of CAC presence and severity had both prognostic and potential treatment implications across all scleroderma subtypes. In a large cohort of well-phenotyped patients with SSc, we have demonstrated that CAC is prevalent and was present in 58% (149/258). This prevalence was significantly higher among the subgroup with SSc-PAH, independent of age and sex (71%; χ²(6) = 66.7; p < 0.001). CAC presence and severity were associated with increased risk of mortality, independent of age (HR = 2.13; 95% CI = 1.0–4.4; p = 0.043). The presence of CAC also had potential treatment implications across all SSc subtypes, as well as the small group of patients aged <50 years who had CAC.

The CAC prevalence observed is particularly notable in view of the female predominance of the cohort (86% female), given the known higher prevalence of CAD among men in the wider population. ³⁴ A recent analysis of 1400 consecutive chest CTs demonstrated a CAC prevalence of 54% in a cohort that was 56% female: in that study, CAC was significantly associated with male sex. ⁹ Other cohorts in the literature that measured CAC in non-dedicated CTs have demonstrated a prevalence of 54% in a cohort with bronchiectasis (62% female), 36% in a cohort with pulmonary embolism (PE) (63% female) and 42% in a cohort with lupus (90% female).^6,8,35 This study reports preliminary retrospective data without non-SSc cohorts for comparison but the significant results warrant further investigation to establish the prognostic and clinical role of CAC calculation in SSc patients. Further interrogating this observed high prevalence is of clinical importance: in the context of inflammatory systemic disease, as well as its sequelae such as PAH, managing patients’ general health by identifying and modifying risk factors may impact on survival.

Across the whole cohort, only 46% of those with CAC on CT were prescribed a statin. Our understanding of the role of inflammation in the development of atherosclerosis has improved significantly in recent years. ¹ We know that patients with other chronic inflammatory diseases such as SLE and RA experience excess cardiovascular disease, predominantly due to accelerated atherosclerosis.^24,36 Our study suggests that SSc patients may require a form of cardiovascular disease (CVD) screening to ensure adequate primary prevention, in line with that proposed in other inflammatory conditions like RA. ³⁷

The recent consensus statement on cardiac disease screening and management in SSc stresses the importance of consideration of acute and chronic CAD when managing patients with scleroderma, but does not comment on the use of CAC metrics on CT. ³⁸ Our data suggest assessment of surrogate markers of CAD like CACS may help to quantify the burden of subclinical atherosclerosis in SSc and guide strategies for screening and risk assessment in this population. Identifying CAC in asymptomatic patients may trigger a review of cardiovascular risk factors and optimisation of modifiable factors and drug therapy, including statins but also potentially other preventive medications such as angiotensin-converting (ACE) inhibitors or aspirin. ³⁹ Recent prospective studies in other patient populations have demonstrated that use of CACS on CT can reclassify statin eligibility and increase prescription of preventive medications and modification of cardiovascular risk factors, without prompting high rates of invasive testing.^40,41

Only 28% (41/149) of those patients with CAC on CT were prescribed antiplatelet therapy. A recent EUSTAR analysis has demonstrated that platelet inhibitors were the most important therapy-related predictors associated with digital ulcer occurrence in SSc. ⁴² There is not yet an evidence base regarding the use of antiplatelet agents as primary prevention in CAD in SSc and the use of antiplatelet therapy as primary prevention is not currently standard of care in CVD. ⁴³ Further prospective data are required to examine how best to identify those patients who would benefit from primary prevention, and what form that primary prevention should take.

This study was performed in patients undergoing non-gated CT studies for clinically indicated reasons, and further study would be required to test the practicality and utility of using ordinal CAC scoring in a screening setting over and above using established CV risk tools in this patient cohort. Due to the retrospective nature of our study, we did not have reliable data in all included patients to allow comparative analysis with clinical risk tools like Qrisk2 (which would require exact blood pressure measurements and serum cholesterol levels at time of CT). ⁴⁴ However, future work which analysed the added value of CAC versus established clinical risk tools with regard to primary prevention in this patient cohort could well prove valuable and may support a change in practice for SSc patients. SSc patients frequently undergo CT imaging due to their risk of developing complications, such as ILD and PAH: our study suggests that using these studies to opportunistically screen for CAC in clinical practice would have a potential impact on primary prevention in a large proportion of patients.

CAC was more prevalent in SSc-PAH than in other subtypes of SSc. This suggests that clinicians managing these patients should be particularly vigilant for the presence of CAD. It also raises the possibility that inflammation associated with pulmonary vascular disease may be associated with coronary vascular pathology in this patient cohort. However, further prospective work was required to validate these findings and to evaluate the pathogenesis of cardiovascular disease in SSc.

CAC presence and severity impact all-cause mortality independently of age, sex and other co-morbidities (HR = 2.3; 95% CI = 1.1–4.9; p = 0.033 and HR = 2.8; 95% CI = 1.2–6.6; p = 0.018), which suggests that this finding has clinical relevance and is a potential target for therapeutic intervention in this cohort. The impact of moderate–severe CAC on mortality was even higher in the cohort of patients without ILD or PAH (HR = 8.0; 95% CI = 1.5:43.3; p = 0.015). The overall survival of patients with SSc has improved significantly in recent decades with the introduction of ACE inhibitors for renal crisis and the advent of more effective PAH treatments: the impact on mortality of co-morbidities not directly related to SSc has increased and will likely continue to rise. ⁴⁵ Other end-organ damage may become increasingly more apparent with improving survival. It is, therefore, of increasing clinical importance that the prevalence, burden, pathophysiology and impact of CAD in SSc are further evaluated to guide clinical decisions regarding screening and secondary prevention, as well as the development of potential future therapeutic interventions.

The study is limited by its single centre, retrospective nature, as well as missing data leading to exclusions as outlined. This, therefore, enables only an estimation of the potential impact of reporting CAC on routine clinical management. Selection bias cannot be excluded as these CT scans were performed only as part of routine clinical practice: a higher proportion of patients with cardiovascular or pulmonary disease can be expected as a clinician had sufficient suspicion to request a CT. Follow-up was limited to 3 years post CT; we do not present long-term follow-up data of our cohort. All-cause mortality was chosen as an end point rather than a composite of major cardiovascular events; future work should evaluate primary myocardial involvement as cause of death in SSc.

The included scans were non-dedicated CTs which not originally requested for coronary artery evaluation; the results are only applicable to patients having non-gated CT examinations for clinically indicated reasons and further study would be required to test the role of CACS on non-cardiac CT as a screening tool in SSc. However, the impact of CAC presence and severity observed, even when adjusted for confounders, suggests that this is an important area for further study and potential changes in clinical practice; particularly with regard to the potential for opportunistic screening and modification of risk factors. In addition, CAC does not detect non-calcified plaque and acts only as a surrogate of total plaque burden: patients with no or with low CAC may still be at risk.^46,47 The study period also includes CTs performed back to 2007; modern scanners may have improved sensitivity for the detection of CAC, which may continue to improve as CT technology improves.

This study demonstrates that scoring CAC on non-cardiac chest CT in patients with SSc provides an opportunistic screening tool to identify patients with asymptomatic or undiagnosed CAD who are at increased risk of mortality. The association with mortality across subtypes suggests all SSc patients may benefit from cardiovascular risk optimisation, but patients with PAH appear to be at a further increased risk of CAD; the delineation of the underlying pathophysiology requires further prospective studies. Population-based cohort studies of the incidence and prevalence of CAD in SSc relative to the general population, as well as correlation of CAC with angiographic findings in this cohort, are required.