Cardiac magnetic resonance imaging parameters show association between myocardial abnormalities and severity of chronic kidney disease

Background Chronic kidney disease patients have increased risk of cardiovascular abnormalities. This study investigated the relationship between cardiovascular abnormalities and the severity of chronic kidney disease using cardiac magnetic resonance imaging. Methods We enrolled 84 participants with various stages of chronic kidney disease (group I: stages 1–3, n = 23; group II: stages 4–5, n = 20; group III: hemodialysis patients, n = 41) and 32 healthy subjects. The demographics and biochemical parameters of the study subjects were evaluated. All subjects underwent non-contrast cardiac magnetic resonance scans. Myocardial strain, native T1, and T2 values were calculated from the scanning results. Analysis of covariance was used to compare the imaging parameters between group I-III and the controls. Results The left ventricular ejection fraction (49 vs. 56%, p = 0.021), global radial strain (29 vs. 37, p = 0.019) and global circumferential strain (-17.4 vs. −20.6, p < 0.001) were significantly worse in group III patients compared with the controls. Furthermore, the global longitudinal strain had a significant decline in group II and III patients compared with the controls (-13.7 and −12.9 vs. −16.2, p < 0.05). Compared with the controls, the native T1 values were significantly higher in group II and III patients (1,041 ± 7 and 1,053 ± 6 vs. 1,009 ± 6, p < 0.05), and T2 values were obviously higher in group I-III patients (49.9 ± 0.6 and 53.2 ± 0.7 and 50.1 ± 0.5 vs. 46.6 ± 0.5, p < 0.001). The advanced chronic kidney disease stage showed significant positive correlation with global radial strain (r = 0.436, p < 0.001), global circumferential strain (r = 0.386, p < 0.001), native T1 (r = 0.5, p < 0.001) and T2 (r = 0.467, p < 0.001) values. In comparison with the group II patients, hemodialysis patients showed significantly lower T2 values (53.2 ± 0.7 vs. 50.1 ± 0.5, p = 0.002), but no significant difference in T1 values (1,041 ± 7 vs. 1,053 ± 6). Conclusions Our study showed that myocardial strain, native T1, and T2 values progressively got worse with advancing chronic kidney disease stage. The increased T1 values and decreased T2 values of hemodialysis patients might be due to increasing myocardial fibrosis but with reduction in oedema following effective fluid management. Trial registration number ChiCTR2100053561 (http://www.chictr.org.cn/edit.aspx?pid=139737&htm=4).

Background: Chronic kidney disease patients have increased risk of cardiovascular abnormalities. This study investigated the relationship between cardiovascular abnormalities and the severity of chronic kidney disease using cardiac magnetic resonance imaging.
Methods: We enrolled participants with various stages of chronic kidney disease (group I: stages -, n = ; group II: stages -, n = ; group III: hemodialysis patients, n = ) and healthy subjects. The demographics and biochemical parameters of the study subjects were evaluated. All subjects underwent non-contrast cardiac magnetic resonance scans. Myocardial strain, native T , and T values were calculated from the scanning results. Analysis of covariance was used to compare the imaging parameters between group I-III and the controls.
) were significantly worse in group III patients compared with the controls. Furthermore, the global longitudinal strain had a significant decline in group II and III patients compared with the controls (-. and − . vs. − . , p < . ). Compared with the controls, the native T values were significantly higher in group II and III patients ( , ± and , ± vs. , ± , p < . ), and T values were obviously higher in group I-III patients ( . ± . and . ± . and . ± . vs. . ± . , p < . ). The advanced chronic kidney disease stage showed significant positive correlation with global radial strain (r = .

Introduction
Chronic kidney disease (CKD) is a major health problem worldwide with the rates of morbidity and mortality increasing by 29.3 and 41.5%, respectively, between 1990 and 2017 (1). The risk of cardiovascular disease (CVD) has been shown to be higher in CKD patients compared with the general population (2). CVD is also the leading cause of mortality in the CKD patients (3,4). Izamura et al. (5) reported that lower estimated glomerular filtration rates (eGFRs) in CKD patients were associated with increased cardiac cell enlargement, cardiac hypertrophy, and fibrosis of the left ventricle. Therefore, CKD patients should be regarded as a highrisk group for CVD and need close medical attention at an individual level (6).
Non-contrast cardiac magnetic resonance (CMR) imaging can be used to monitor changes in myocardial mass and  biventricular volumes in CKD patients undergoing long-term hemodialysis (HD) (7). CMR is also used to directly and non-invasively estimate the pathologic changes in the cardiac structure and function based on myocardial T1 and T2 mapping (8). Left ventricular (LV) strain parameters are more sensitive than left ventricular ejection fraction (LVEF) in detecting early cardiac dysfunction because they directly estimate the movement of myocardial fibers (9).
Previous CMR studies (10,11) reported myocardial abnormalities in patients with early and advanced CKD or end-stage renal disease (ESRD) patients. These results showed that cardiac abnormalities could occur in both early-stage and advanced-stage CKD patients. Furthermore, study by Hayer et al. (12) showed that myocardial fibrosis evaluated by native T1 time was inversely associated with kidney function. However, the stage of CKD at which obvious myocardial abnormalities appear is not well-defined. Furthermore, the majority of the ESRD patients undergo HD (13,14). However, it is not clear about the relative contributions of myocardial fibrosis to the change of native T1 times in HD patients. Therefore, in this study, we investigated the association between obvious myocardial abnormalities and CKD stages by comparing the non-contrast CMR parameters. Furthermore, we analyzed the alterations of CMR parameters in patients following HD.

Study subjects
In this prospective longitudinal observational study, 84 participants with different stages of CKD were enrolled from the Department of Nephrology, Wuhan Union Hospital, between March 2021 to October 2021. We also enrolled 32 healthy subjects of similar age, gender, and body mass index (BMI) from the Wuhan community between March 2021 to October 2021. This study was approved by the . /fcvm. . College of Huazhong  University of Science and Technology. It was first registered on  24/11/2021 with the registration number ChiCTR2100053561, and was conducted in accordance with the Helsinki Declaration. The written informed consent was obtained from all the study subjects. The inclusion criteria for participants with CKD were as follows: (1) clinically confirmed CKD at different stages; (2) age between 30 and 80 years; (3) absence of chest pain, dyspnea, and palpitations; (4) absence of history for cardiovascular diseases such as congenital heart disease, coronary artery disease, valvular heart disease, or cardiomyopathy; and (5) normal electrocardiographic manifestations. All dialysis patients had maintained hemodialysis for 4 hours, 3 times a week for at least 3 months. The inclusion criteria for the healthy controls were as follows: (1) age between 30 and 80 years; (2) absence of history for cardiovascular diseases, hypertension, hyperlipidemia, and diabetes; (3) normal physical examination; and (4) normal electrocardiographic manifestations. The exclusion criteria for the participants with CKD were as follows: (1) history of known specific cardiomyopathies, valvular heart disease or myocarditis; and (2) standard contraindications to CMR (e.g., metal implants, severe claustrophobia, and inability to hold breath).
Study subjects were diagnosed with hypertension when the average SBP value was >140 mmHg. BMI was calculated by dividing dry weight (kg) by body height (m) 2 .

CMR scanning protocol
The CMR scans were performed in a 1.5 T MAGNETOM Aera MRI scanner (Siemens Healthcare, Erlangen, Germany) equipped with 18-channel phased-array surface coils using the vector electrocardiogram gating. Participants with CKD were all scanned on the next day after confirmation of CKD by nephrologists. Dialysis patients were all scanned on nondialysis days but not after a long break; thus, all scans were performed within 18 to 24 h after the most recent dialysis session (16). Cine imaging of the LV long axis and sequential short-axis planes was performed using the balanced steady-state free precession (SSFP) sequence. Cine imaging parameters were as follows: repetition time, 2.93 ms; echo time, 1.16 ms; slice thickness, 6 mm; flip angle, 80 • ; field of view, 340 × 255 mm; matrix, 256 × 205; and 25 calculated cardiac phases.
Native T1 mapping at the base, mid, and apical levels of the LV short axis were performed using the modified looklocker inversion recovery (MOLLI) sequence. The T1 mapping parameters were as follows: repetition time, 3

Analysis of cardiac volume index and function
CMR image analysis was performed using the commercially available CVI42 software (Circle Cardiovascular, Calgary, Canada), and the analyst were blinded to the CKD group of study populations. The volumetric and functional parameters of the left ventricle were measured using the continuous shortaxis slice cine images by manually tracing the endocardial and epicardial borders. The papillary muscles and trabeculations were excluded as part of the ventricular mass. CMR parameters such as LV end-diastolic volumes index (EDVI), end-systolic volume index (ESVI), stroke volume index (SVI), ejection fraction (EF), LV mass index (LVMI) and cardiac index (CI) were measured by the commercial CVI42 software (Circle Cardiovascular, Calgary, Canada) automatically. The left atrial volume index (LAVI) was calculated manually .

Estimation of the native T and T values
The regions of interest (ROIs) were manually delineated in the mid-layer of the myocardium among the basal, middle, and apical LV segments to measure native T1 and T2 values. Sixteen ROIs were drawn for each participant based on the American Heart Association 16-segment model ( Figures 1A-F). The image artifacts and coronary artery were eliminated from the ROIs. The average native T1 and T2 values were calculated from the three short-axis slices.

Estimation of the myocardial systolic strain
The peak systolic LV strain parameters were calculated using the CVI42 software (Circle Cardiovascular, Calgary, Canada). Multiple long-axis cine images (2, 3, and 4-chamber views) and short-axis cine images were imported into the software. Then, the endocardial and epicardial borders of the LV were delineated in the end-diastolic frame (including papillary muscles and .

Statistical analysis
Statistical analysis was performed using the SPSS version 26.0 software (SPSS Inc., Chicago, Illinois, USA). Kolmogorov-Smirnov test was used to analyze the normal distribution of continuous data. The normally and non-normally distributed data were summarized as means ± standard deviation and median (interquartile range, IQR), respectively. Differences between the normally distributed variables were analyzed using the independent-sample Student's t-test. Differences between the categorical variables were analyzed using the chi-square test. Differences in the clinical and CMR parameters between the three CKD patient groups and healthy subjects were compared by analysis of variance (ANOVA). Differences in the CMR variables after adjusting for age, BMI, and HR were assessed by analysis of covariance (ANCOVA). The relationships between CMR parameters and CKD stages of CKD patients without hemodialysis were examined using Spearman's correlation tests. P-value < 0.05 (two-tailed) was considered statistically significant.

Basic clinical characteristics of the study groups
In this study, 84 participants with CKD and 32 healthy controls were enrolled. Table 1 shows the demographics of all the study subjects and the biochemical indices of the CKD patients. The basic characteristics including age (p = 0.336), sex (p = 0.220), BMI (p = 0.064) and HR (p = 0.205) were similar between the group of participants with CKD and healthy subjects. However, significant differences in age (p < 0.001), BMI (p = 0.023), and HR (p = 0.001) were observed among the healthy subjects and the three groups of participants with CKD, namely, CKD 1-3, CKD 4-5, and HD ( Table 1). The healthy subjects did not show any history of cardiovascular diseases, hypertension, hyperlipidemia, and diabetes.
Non-HD participants with CKD show significant alterations in left ventricular mass, volume, and function Table 2 shows the LV mass, volume, and functional characteristics for the three CKD groups (I-III) and the healthy subjects. Table 3 shows the LV mass, volume, and functional parameters for the CKD patients without HD (groups I and II) and the healthy subjects after adjusting for age, BMI, and HR. The CMR-derived parameters LVEDVI, LVSVI, LVMI, and CI values were significantly increased starting from later stages (LVEDVI: 80 vs. 57 ml/m 2 ; LVSVI: 42 vs. 32 ml/m 2 ; LVMI: 57 vs. 36 g/m 2 ; CI: 2.9 vs. 2.2 l/min/m 2 , p < 0.05 between group II and healthy subjects for all). Furthermore, the CMR-derived parameters LVESVI, LVEF and maximum LAVI values never had significant change in patients without HD.
Non-HD participants with CKD show significant alterations in the left ventricular strain Table 3 summarizes the values for the LV strain parameters in the three groups after adjusting for age, BMI, and HR. The CMR-derived parameters GRS and GCS never had significant change in patients without HD (Figures 3A,B). However, the CMR-derived parameter GLS values were significantly reduced starting from later stages (group II vs. healthy subjects: −13.7 ± 0.7 vs. −16.2 ± 0.5, p = 0.024) ( Figure 3C).
. /fcvm. . Non-HD participants with CKD show significant alterations in myocardial native T and T values Table 3 shows the differences in the native T1 and T2 values between the three groups after adjusting for age, BMI, and HR. The global cardiac native T1 values were higher in the CKD groups compared with the healthy subjects; the values showed an incremental increase with CKD severity (healthy subjects: 1,009 ± 6; CKD group I: 1,027 ± 7; CKD group II: 1,041 ± 7). The CMR-derived parameter global native T1 values were significantly increased starting from later stages (group II vs. control: p = 0.003) ( Figure 4A). Furthermore, the global T2 values were significantly increased starting from early stages (group I vs. healthy subjects: 49.9 ± 0.6 vs. 46.6 ± 0.5, p < 0.001) ( Figure 4B). The septal and midseptal T1 and T2 values and the inter-group differences showed similar trends as observed with the global T1 and T2 values, respectively (Table 3).

Discussion
This study shows that CMR parameters reflecting LV myocardial structure and function get worsen with advancing CKD stages. Participants with CKD stages 4-5 were more likely to occur significantly increased T1 values, which may reflect myocardial fibrosis. The increased native T1 values and decreased T2 values of HD patients might be due to increasing myocardial fibrosis in the HD group but with reduction in oedema following effective fluid management on dialysis.
In this prospective study, changes in the myocardial strain parameters like GLS, GCS, and GRS were associated with advanced CKD stage. This suggested significant alterations in the myocardial wall remodeling among patients with advanced or severe CKD. Furthermore, this study showed that GLS of participants with CKD decreased significantly from stages 4-5 onwards when compared with the healthy subjects, whereas LVEF values were reduced in the CKD patients undergoing hemodialysis. These results were in agreement with previous reports demonstrating GLS as a more sensitive parameter than LVEF in detecting subclinical LV dysfunction because GLS directly measures the movement of myocardial fibers (9, 17). The prognostic value of GLS was also higher than the values for LVEF in predicting mortality and adverse cardiovascular outcomes (18,19). Therefore, CKD patients with reduced GLS require close monitoring for early detection of cardiac abnormalities.
A previous report (9) suggested that aberrant GLS was secondary to myocyte hypertrophy caused by myocardial fibrosis. Our study shows higher native T1 and T2 values in .
/fcvm. . The column named "start from which stage(s)" represents the stage(s) of CKD in patients who developed significant abnormalities compared to healthy subjects: "early" meaning p < 0.05 between control and group I; "later" meaning p only <0.05 between control and group II, in the other words, p ≥ 0.05 between control and group I; and "never" meaning p ≥ 0.05 among all groups. p-values for comparison between patients who developed significant abnormalities and healthy subjects.  CKD patients compared with healthy subjects, and increasing native T1 and T2 values with the advancing CKD stages, which are consistent with previous studies (8,18). In addition, a negative correlation between T2 values and GFR was also found in patients with heart failure (20), supporting the finding of our trial. Other previous studies demonstrated that native T1 mapping improved the diagnostic accuracy of CMR in estimating myocardial fibrosis (16,21). In addition, myocardial native T1 values were also associated with myocardial edema, infiltration of immune cells (22), acute myocardial infarction or amyloidosis (23). Participants enrolled in our study had no history of heart disease or amyloidosis, suggesting that the elevated T1 values in participants with advanced CKD were irrelevant to based heart disease or amyloidosis. However, we cannot rule out the effect of myocardial edema on the changes of T1 values.
. /fcvm. .  Native T2 mapping measures the free water content in tissues including the cardiac tissue (7). In current study, myocardial native T1 values and T2 values of CKD groups are higher than healthy subjects showing that myocardium of CKD patients may develop both fibrosis and oedema. Furthermore, our results demonstrate that CKD patients may develop worsening myocardial fibrosis and oedema with advancing CKD severity. CKD progression facilitates the onset of myocardial fibrosis (24). However, the exact CKD stage at which myocardial fibrosis .
/fcvm. . is clinically manifest is not well-characterized. In the current study, participants with CKD stages 4-5 showed significantly higher native T1 and T2 values compared with healthy subjects. These results supported previous biopsy findings from ESRD patients that showed extensive interstitial fibrosis (25,26). Edwards et al. (11) reported irreversible fibrosis in 14% of stage 2 to 4 CKD patients using late gadolinium enhancement (LGE). These slightly different results may be explainable by differences in sample size and their use of gadolinium.
LGE is not sensitive enough to detect diffuse myocardial fibrosis (24) and has risk of nephrogenic systemic fibrosis (27). Our results suggested that participants with CKD stages 4-5 were more likely to develop significant myocardial fibrosis. However, further investigations and more sensitive magnetic resonance imaging (MRI) techniques are needed to determine if myocardial fibrosis occurs in patients with early CKD. In our study, participants with CKD undergoing HD showed significantly reduced T2 values compared with non-HD patients with CKD stages 4-5. Previous studies reported that T2 values were related to myocardial fluid content (28)(29)(30). Our results showed that myocardial oedema may be improved in ESRD patients undergoing HD compared with non-HD patients with advanced CKD, thereby supporting previous findings (7, 31). Kotecha et al. (7) showed that native T1 and T2 values were significantly reduced in ESRD patients after dialysis, suggesting that these values reflected an acute reduction of myocardial water content. Rankin et al. (31) also showed significant decrease in native T1 and T2 values following hemodialysis with fluid removal in CKD patients undergoing HD based on 3T CMR. However, the changes of T1 values in our current study may seem contradictory to these two studies. The reason for this difference may be that Kotecha and Rankin et al. referred to within-subject differences pre/post HD. T1 values of participants with CKD undergoing HD in this study increased slightly than non-HD patients with CKD stages 4-5. This suggested that for CKD patients undergoing HD, myocardial fibrosis, rather than myocardial oedema, accounts for the elevated T1 values. Similarly, Graham-Brown et. compared 124 HD patients to 137 healthy subjects and found that the obviously elevated T1 values occurred independently with T2 values, which had no significant changes. It is reasonable to speculate that the observed increase .
/fcvm. . in native T1 values and decrease in T2 values between CKD 4-5 and HD patients might be due to increasing fibrosis in the dialysis group but with reduction in oedema following effective fluid management on dialysis. However, our study did not compare native T1 values of CKD patients before and after dialysis. Hence, this aspect requires further investigation. Our study has a few limitations. Firstly, the sample size of our single-center study was small. It was difficult to stratify patients into CKD 1-5 stage and directly state which stage of 1-5 shows the significant reduction in myocardial function. Therefore, large cohort, multi-center studies are necessary to confirm our findings and explore potential relationships between myocardial abnormalities and CKD stages. Secondly, the cross-sectional nature of this study may have affected the results because of individual differences among the study subjects. We did not compare parameters of participants with CKD before and after dialysis. Therefore, in the future, a longitudinal study needs to be performed to confirm our results. Thirdly, although the CKD groups were matched with the healthy subjects based on age, BMI, and HR, significant inter-group differences existed among the four groups. Therefore, we used ANCOVA to exclude the influence of these confounding factors. Forthly, there was a lack of assessment of myocardial fibrosis and fluid status, either clinically or bioimpedance. Finally, because of the small sample size, we did not investigate the effects of important factors such as hypertension and diabetes that can influence cardiac functions.

Conclusions
In conclusion, myocardial strain, native T1, and T2 values progressively got worse as the severity of CKD increased. Myocardial fibrosis and edema were observed more frequently in participants with advanced CKD or severe CKD requiring HD. The increased native T1 values and decreased T2 values of CKD patients undergoing HD might be due to increasing myocardial fibrosis in the HD group but with reduction in oedema following effective fluid management on dialysis. Further large cohort multi-center studies are needed to confirm our findings.

Data availability statement
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Ethics statement
The studies involving human participants were reviewed and approved by the Ethics Committee of the Tongji Medical College of Huazhong University of Science and Technology. The patients/participants provided their written informed consent to participate in this study.

Author contributions
XJ, XH, and YC conceived and designed the study. YC and HS provided the administrative support. YW, FH, and CZ provided study materials and patients. XJ, XH, and YZ collected and assembled data of the study. XJ and XH analyzed and interpreted the study data. All authors read and approved the final version of the manuscript.