This cross-sectional study enrolled 159 consecutive patients with CTD who underwent 3.0T CMR examination from January 2015 to July 2023 in our hospital. The diagnosis of CTD was according to the American College of Rheumatology diagnostic criteria or the European League Against Rheumatism for IIM (14), SLE (15), SSc (16), pSS (17), and rheumatoid arthritis (18) as well as sharp criteria for mixed connective tissue disease (MCTD) (19). Overlap syndrome was defined as patients with two or more clinical features of CTD at the same time. Undifferentiated connective tissue disease (UCTD) was defined as patients with clinical and laboratory features suggestive of systemic autoimmune diseases but cannot be classified into any of the defined CTD classification criteria. The exclusion criteria were as follows: (1) age <18 years; (2) a history of coronary artery disease; (3) severe heart valvular disease, cardiomyopathy, and congenital heart disease; (4) left ventricular ejection fraction (LVEF) <50%; (5) severe kidney dysfunction (estimated glomerular filtration rate <30 ml/min/1.73 m²); (6) poor image quality; and (7) incomplete clinical data. Finally, a total of 104 CTD patients (comprising 49 patients with IIM, 29 with SLE, six with SSc, five with pSS, eight with overlap syndrome, two with rheumatoid arthritis, four with UCDT, and one with MCTD) were retrospectively included. The patients were divided into three subgroups based on the tertiles of SUA level in the present study. CTD patients with preserved RVEF (defined as RVEF ≥50%) were analyzed as a subgroup, too. Meanwhile, 52 age- and gender-matched individuals with no history of cardiovascular or systematic diseases were selected as the control group. This study protocol was approved by the Institutional Review Board of our hospital (TJ-IRB20230914), and the requirement for written informed consent was waived due to the retrospective design. The patients’ flowchart of this study is shown in Figure 1 .

Venous blood samples were obtained from a peripheral vein within 48 h of the CMR examination as well as without receiving any urate-lowering therapies. SUA, estimated glomerular filtration rate, total cholesterol, triglyceride, high-density lipoprotein-cholesterol, and low-density lipoprotein-cholesterol were measured in the laboratory department of our hospital. Pulmonary fibrosis was defined as reported on thoracic computed tomography imaging by specialist radiologists. PH was defined as systolic pulmonary arterial pressure >36 mmHg on echocardiography as measured by experienced doctors or mean pulmonary arterial pressure ≥20 mmHg on right-sided heart catheterization (20).

All patients and controls underwent a standard clinical protocol using 3.0T magnetic resonance imaging (MRI) scanners (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany) with an 18-channel phased-array receive coil in the supine position. A steady-state free-precession sequence with electrocardiogram triggering and respiratory gating was used to assess cardiac structure and function. Short-axis cine images covering the right ventricle (RV) from apex to tricuspid annulus, as well as standard cine long-axis four-chamber view, were acquired. Typical imaging parameters were as follows: echo time/repetition time, 1.4/37.7 ms; flip angle, 55°; field of view, 360 × 360 mm²; acquisition matrix size, 192 × 146; slice thickness, 8 mm; 25 phases per cardiac cycle.

Post-processing of all CMR images was analyzed using an offline customized software CVI42 (version 5.13.9, Circle Cardiovascular Imaging Inc., Calgary, Canada). Right and left ventricular endocardial and epicardial contours were drawn manually at the end of diastole and systole in all short-axis cine images excluding the papillary muscles and trabeculae. Biventricular end-diastolic volume (EDV), end-systolic volume (ESV), stroke volume, myocardial mass, and ejection fraction (EF) were calculated. Right atrial (RA) volumes were measured in CVI42 function LAX biplanar module using the biplane area–length method on the cine CMR 4-chambers views, excluding the confluence of the superior and inferior vena cava, as well as the atrial appendage. RA maximum (V _max) and minimum volumes (V _min) were assessed at the end of ventricular systole and diastole, respectively. Pre-atrial volumes (V _pre-A) were obtained before atrial contraction. The functions of RA were calculated by the following equations as previously described (21): (1) reservoir function: total emptying fraction (EF_t) = (V _max − V _min)/V _max × 100%; (2) conduit function: passive emptying fraction (EF_p) = (V _max – V _pre-A)/V _max × 100%; (3) booster pump function: active emptying fraction (EF_a) = (V _pre-A–V _min)/V _pre-A × 100%. All volumes and RV myocardial masses were indexed to body surface area before analysis.

The myocardial deformation parameters of the RV free wall (RVFW) as well as the RA were evaluated in CVI42 feature-tracking module. The endocardial and epicardial contours of RV and RA were manually drawn using a point-and-click approach on the end-diastolic images, respectively. Subsequently, the contours were automatically tracked throughout the cardiac cycle while also checking the quality of the automatic tracking and adjusting the initial contour manually if necessary. RV global circumferential strain (GCS) and global circumferential strain rate (GCSR) were derived from tracking the RVFW in the four-chamber view and short-axial views, while the global longitudinal RV and RA strain and strain rates were quantified based on the four-chamber view only ( Supplementary Figures S1, S2 ). RV radial strain parameters were not evaluated due to the high variability as previously described (22). Three aspects of RA phasic function were quantified for each subject as follows: (1) RA reservoir function (total strain [ε _s] and strain rate [SRs]), (2) RA conduit function (passive strain [ε _e] and strain rate [SRe]), and (3) RA booster function (active strain [ε _a] and strain rate [SRa]). Above all, the results were based on the average of three independent repeated measurements.

The inter- and intra-observer variabilities for RV and RA strain and strain rate parameters were obtained from 40 randomly selected subjects (10 controls and 30 CTD patients). Inter-observer variability was assessed by two radiologists (YT and JW, with 3 and 5 years of experience in CMR imaging, respectively) blinded to all clinical information. Intra-observer variability was measured twice by radiologist YT, with at least 1 month between repeated analysis.

Statistical analysis was performed using SPSS version 26.0 (IBM Deutschland GmbH, Armonk, NY, USA) and GraphPad Prism version 10.0.2 (GraphPad Prism Software Inc., San Diego, CA, USA). Continuous variables were summarized as mean ± standard deviation (SD) or median with interquartile range (IQR) and were compared between groups using Student’s t-test or Mann–Whitney U-test as appropriate based on their normality. Comparisons among three groups were carried out using analysis of variance (ANOVA) with Bonferroni or Games Howell post hoc test for normally distributed data or the Kruskal–Wallis test for non-normally distributed data. Categorical variables were presented as frequencies (percentages) and were compared using the Pearson χ ² test. Univariable and multivariable linear regression analyses were used to investigate the association of SUA levels with CMR-derived RV and RA strain parameters. Multivariable analyses were adjusted to all variables without collinearity and a P-value <0.05 in univariable analysis as well as age, sex, hyperlipidemia, disease duration, pulmonary fibrosis, and PH. Intra-class correlation coefficient (ICC) was calculated to determine the inter- and intra-observer reproducibility of RV and RA strain parameters. A two-sided P-value <0.05 was considered statistically significant.

The clinical characteristics are summarized in Table 1 . CTD patients had significantly higher SUA levels [308.5 (249.00, 357.800) vs. 265.30 (222.80, 297.35) umol/L, P=0.001] than controls. According to the tertiles of SUA level distribution in the present study, the CTD patients were divided into three subgroups: the first tertile (SUA level <276 umol/L), the second tertile (276–342 umol/L), and the third tertile (>342 umol/L). Approximately, a third of CTD patients had pulmonary involvement and presented in New York Heart Association (NYHA) class III–IV, which had a higher percentage in the third tertile. The median disease duration was 17.50 (5.00, 60.00) months. The CTD patients in the third tertile also had a significantly longer disease duration than those in the first tertile [21.00 (6.00, 66.00) vs. 7.00 (2.00, 52.75) months, P = 0.030]. No significant differences were found among age, sex, BMI, BSA, heart rate, systolic blood pressure, and diastolic blood pressure as well as the proportion of hypertension and diabetes mellitus between the CTD patients and the controls (all P > 0.05), but the percentage of hyperlipidemia was higher in CTD patients than in the controls (34.6 vs. 9.6%, P = 0.001).

The conventional CMR characteristics as well as RV and RA strain parameters of all participants are shown in Table 2 . The CTD patients had significantly increased RVEDVi, RVESVi, and LV mass index, whereas they have decreased RVEF and RV GCS compared with the controls (all p < 0.05), and RV GCS remains impaired even in CTD patients with preserved RVEF (-17.81 ± 3.56 vs. -19.17 ± 2.84%, P = 0.025). Although no differences in RA morphological and functional parameters were observed among CTD patients and controls, the CTD patients showed significantly impaired RA strain parameters except SR_a [ε _s: 49.52 ± 16.90 vs. 58.04 ± 12.60 %, P<0.001; ε _e: 25.92 (17.13, 34.17) vs. 33.95 (25.66, 39.81) %, P = 0.002; ε _a: 22.20 (15.73, 27.68) vs. 25.31 (18.39, 30.24) %, P = 0.025; SRs: 2.62 ± 0.83 vs. 3.04 ± 0.95 1/s, P = 0.005; SR_e: -2.21 (-3.05, -1.46) vs. -2.62 (-3.62, -2.04) 1/s, P = 0.004]. In CTD patients with preserved RVEF, there are significantly increased LV mass index, RVEDVi, RVSVi, RAEF_t and RAEF_a, whereas decreased RAV_min index compared with controls (all P < 0.05). Both RV and RA strain parameters tended to worsen with increasing SUA levels among subgroups. Compared with patients in the first tertile, those in the third tertile had significantly impaired RV GLS and RV GCS as well as RA reservoir and conduit strain parameters. RV GLS was also significantly lower in the second tertile (all P < 0.05). The RV GLSR, RV GCSR, and RA booster strain parameters had no significant difference across tertiles of SUA in CTD patients (all P > 0.05). Figures 2 – 4 show the RV and RA strain parameters in CTD subgroups based on the tertiles of SUA and controls.

In univariable linear regression analyses, the elevated SUA levels were significantly associated with the impaired RV GLS and RV GCS as well as RA reservoir and conduit strain and strain rates (absolute value of β = 0.250 to 0.293, all P < 0.05). Multivariable linear regression analyses showed that the elevated SUA level was still independently correlated with impaired ε _e (β = -0.212, P = 0.035) and RV GCS (β = 0.207, P = 0.019) in CTD patients after adjustment for age, sex, hyperlipidemia, disease duration, pulmonary fibrosis, and PH as well as excluded the parameters with collinearity (including RVEDVi, RVESVi, RVSVi, RAEF_t, RAEF_p, RAEF_a, RAV_max index, and RAV_pre-A index). There were no independent associations between SUA levels and right cardiac strain rates (all P > 0.05). Other details are presented in Table 3 and Supplementary Table S1 .

The inter-observer and intra-observer variability values of the RV and RA strain parameters are presented in Table 4 . There were excellent intra- and inter-observer agreements in the measurement of RV strain and strain rates (ICC = 0.795 to 0.959 and 0.764 to 0.936, respectively) as well as RA phasic strain and strain rate (ICC = 0.850 to 0.947 and 0.830 to 0.936, respectively).

The 2021 European League Against Rheumatism recommendations highlight the importance of cardiovascular risk management in patients with CTD (23). Until now, left heart involvement in patients with CTD has been extensively explored. CTD often combined with a variety of common complications under the influence of systemic proinflammatory state and other factors such as genetic susceptibility, including left heart disease, pulmonary fibrosis, or PH—all of these can affect RV afterload. Adaptive RV remodeling may occur in the early stages of the disease to overcome increased RV afterload and ensure sufficient stroke volume. With the progression of the disease, these compensatory mechanisms gradually become inadequate, leading to RV excessive dilatation, dysfunction, and, eventually, right heart failure (24). Although RVEF typically indicates a poor prognosis, the decreased RVEF will not appear until the late stage of heart failure. CMR-FT has the potential to detect myocardial function abnormalities earlier in the disease. In patients with CTD, there are significant correlations between reduced RVEF and impaired RV GCS and RV GLS. RV GCS remains impaired even in CTD patients with preserved RVEF. These findings were consistent with those of Wu et al. (5). Athletes’ hearts also had a similar RV mechanical pattern in a recent study, with lower RVEF and RV GCS but preserved or even higher RV GLS compared with sedentary controls (25). These changes might be related to modified RV myocardial fiber orientation. Physiologically, RV contraction mainly depends on longitudinal shortening (26). The relative dominance of the circumferentially oriented fibers can be observed when chronically pressure-loaded (27). In patients with PH, the transverse wall motion of the mid-RV is significantly reduced and has stronger relationship with RVEF than longitudinal motion (28). Pettersen et al. (29) also demonstrated a predominant circumferential over longitudinal free wall shortening in patients with a systemic RV.

Furthermore, RV diastolic dysfunction may precede systolic dysfunction in diseases with chronic RV pressure overload. RA, mainly through increased dimension (higher RA reservoir function) and contractility (higher RA booster pump function), maintains RV filling of the stiffened ventricle, whereas passive emptying (conduit function) is the first to decline due to the reduced RA–RV pressure gradient (30, 31). The increased RAEF_t and RAEF_a and decreased RAV_min index may play as a compensatory mechanism in CTD patients with preserved RVEF. RA dilatation indicated the presence of RV diastolic stiffness (32). Excessive RA dilation and stiffness occur as the disease progresses, leading to deterioration of RA phasic functions and atrioventricular uncoupling. In the present study, the RA strain parameters significantly decline in CTD patients without overt RA dilation and dysfunction. Moreover, our findings indicate that ε _s was positively correlated with RVEF as well as negatively correlated with RAV_min index in patients with CTD, consistent with that observed in PH (33, 34). Therefore, RA strain parameters may be more sensitive and earlier predictors of RV overload in patients with CTD.

The present results show that patients with CTD had significantly elevated SUA levels compared with the controls, and right cardiac chamber remodeling is more severe in CTD patients with higher SUA level subgroups. However, the precise mechanisms underlying the relationship between elevated SUA and right cardiac chamber remodeling found in this study are still unclear and remain to be elucidated.

The elevated SUA levels may lead to right cardiac chamber remodeling by promoting the increased RV afterload in patients with CTD. The RV is very sensitive to afterload change compared with the left ventricle. In the present study, nearly half of the patients were cases combined with pulmonary involvement, which had a higher percentage in the third tertile. Several studies found that the SUA levels were significantly higher in CTD patients with PH than in those without PH (11–36). Although Castillo-Martínez et al. (10) reported that the baseline SUA levels were similar whether SLE patients had PH or not, the steady hyperuricemia was associated with the development of PH in SLE patients without PH at baseline. In CTD-PAH, the baseline SUA levels also had a positive correlation with pulmonary vascular resistance, and steady hyperuricemia may predict a worsened clinical prognosis (12). The elevated SUA level may promote pulmonary vascular remodeling through increased oxidative stress, depletion of nitric oxide, endothelial dysfunction, vasoconstriction, and proliferation of vascular smooth muscle cells (7). Although Savale et al. (37) did not find the influence of hyperuricemia in pulmonary vascular remodeling, UA metabolism was also disturbed in the remodeled pulmonary vascular walls in PAH. Moreover, Wang et al. (38) observed that the SUA levels were higher in rheumatoid arthritis patients with ILD than in patients without ILD. The SUA levels were positively correlated with KL-6, a biomarker of ILD. They further explored the involvement of UA in the pathogenesis of ILD through in vitro cellular experiments and found that UA can induce the epithelial-to-mesenchymal transition in alveolar epithelial cells, which is a critical step in the progression of ILD.

Inflammation may play a key role between elevated uric acid and right heart remodeling. Elevated SUA levels may enhance the inflammatory response and thus affect right cardiac chamber remodeling in patients with CTD. Our data shows that patients in the third tertile had significantly impaired RV GLS and RV GCS as well as RA reservoir and conduit strain parameters compared with those in the first tertile, and RV GLS was also significantly lower in the second tertile—even the first and second tertiles of SUA were in the normal reference range. There was independent association between SUA levels and RV GCS and ε _e after adjustment for pulmonary fibrosis, PH, RVEF, and other potential confounding factors. Chronic long-term inflammation is one of the main pathological features in patients with CTD and be considered to have a contributing role in cardiovascular involvement as well as is one of the most important causes of elevated SUA levels. Increased SUA can, in turn, exacerbate the inflammatory response, resulting in a vicious cycle (39). Even though SSc-PAH patients have similar pulmonary vascular resistance with idiopathic pulmonary hypertension patients, there was a worse RV function and prognosis in the former (40). The primary myocardial involvement derived by inflammation may be one of the potential causes (41). However, none of these mechanisms is independent. They reinforce and interact with each other, ultimately leading to right cardiac chamber remodeling.

In addition, both SUA levels and right heart remodeling are affected by exercise as a confounding factor. On the one hand, low- and moderate-intensity exercise can improve hyperuricemia (42), but high-intensity exercise may increase uric acid production and reduce excretion, leading to increased SUA (43, 44). On the other hand, long-term and intensive endurance exercise has potential effects on the heart, especially the right side, and even on pulmonary circulation (45). However, only one patient in our cohort has a history of long-term and high-intensity exercise and stopped training in the last 4 years. The effect of exercise was not discussed separately in this study. Nonetheless, previous studies suggested that exercise therapy was beneficial to CTD patients and moderates the side effects of high-dose glucocorticoid treatment (46–48). The influence of exercise on right heart remodeling and finding an appropriate kinesitherapy in CTD patients are worth exploring.