Comparison of demographic and clinical characteristics between control and MHD groups
We first compared demographic and clinical data between control and MHD groups. As shown in Table 1, MHD patients had significantly higher SBP, DBP, and creatinine (P < 0.05) compared to the control group, but there were no significant differences with regard to age, gender, heart rate, and body mass index (BMI) between these two groups (P > 0.05). The primary cause of ESRD was glomerulonephritis (55.8%), followed by diabetic nephropathy (17.6%), hypertensive nephrosclerosis (14.7%), and polycystic kidney (11.8%) (Table 1).
Comparison of traditional echocardiography parameters between control and MHD groups
We next compared traditional echocardiography parameters between control and MHD groups. As shown in Table 2, LVEF was comparable between the groups (P > 0.05), but MHD patients had significantly higher LVIDD, LVIDS, IVST, LVPWT, LAD, LVMI (P < 0.01) compared to the control subjects. These observations suggest that MHD patients had impaired cardiac function.
Comparison of layer-specific strain parameters between control and MHD groups
In both groups, LS and CS of the LV three-layer myocardium were consistent with the following pattern: sub-endocardial > mid-myocardium > sub-epicardial. Compared with the control group, the MHD group had significantly lower LV GLS, GCS, and three-layer myocardial LS and CS (P < 0.01) (Table 3). MHD patients also had significantly lower LS of the LV three-layer at the basal, middle, and apical segments (P < 0.01) (Table 4, Fig. 1). In addition, with the exception of the sub-endocardium at the apical and middle segments, MHD patients had significantly lower CS of other LV layers compared to the control group (P < 0.05 or P < 0.01) (Table 5, Fig. 2).
Comparison of LV synchronous parameters between control and MHD groups
In the control group, the color of the TTP bull's eye was uniformly green, while the color of the bull's eye in the MHD group was disordered, showing a mix of yellow and red, indicating that LV contraction was poorly synchronized (Fig. 3). The TTP of the LV 17 segments of MHD patients was significantly higher compared to the control group, and the peak time delay was observed in the majority of these segments (Table 6). In addition, MHD patients had significantly higher PSD compared to the control group (P < 0.01, Table 7). Taken together, these findings suggest that MHD patients had poorly synchronized LV systolic contraction, a sign of cardiac dysfunction.
Correlation between PSD and GLS with three-layer LV myocardial LS
We next examined the correlation between PSD and GLS with three-layer LV myocardial LS. As shown in Table 8 and Figure 4, PSD was positively correlated with sub-endocardial and mid-myocardial LS and GLS (sub-endocardial, r = 0.467, P = 0.005; mid-myocardial, r = 0.513, P = 0.002; GLS, r = 0.463, P = 0.006). There was no significant correlation between PSD and sub-epicardial LS (P = 0.179).
ROC curve analysis of the accuracy of three-layer LV myocardial LS and CS in MHD patients
The AUC was analyzed to obtain the efficacy of three-layer LV myocardial LS and CS in the diagnosis of LV systolic dysfunction in MHD patients. The AUC values of sub-endocardial, mid-myocardial, and sub-epicardial LS myocardium in MHD patients were approximately 0.894, 0.852, and 0.870, respectively, and the cutoff values were approximately -21.15%, -18.33%, and -17.08%, respectively. The sensitivity of sub-endocardial LS (85.3%) was higher than that of sub-epicardial (82.4%) and mid-myocardial LS (73.5%). The specificity of LS of the sub-endocardial LS (82.9%) was also higher than that of the mid-myocardial (80.0%) and sub-epicardial LS (77.1%). The AUC values of sub-epicardial, mid-myocardial, and sub-endocardial CS in MHD patients were approximately 0.852, 0.837, and 0.669, respectively, and the cutoff values were -11.65%, -17.86%, and -24.33%, respectively. The sensitivity of sub-epicardial CS (82.4%) was higher than that of mid-myocardial (73.5%) and sub-endocardial CS (50.0%). However, the specificity of mid-myocardial CS (82.9%) was higher than that of sub-endocardial (80.0%) and sub-epicardial CS (74.3%) (Table 9, Fig. 5). Collectively, our findings indicate that sub-endocardial LS is an appropriate indicator of early stage cardiac dysfunction.
Intra-observer and inter-observer variability
We examined the intra-observer and inter-observer variability in our analysis. Twenty participants were randomly selected from the control and MHD groups. All observers were blinded to the subject type and measured the reproducibility of the three-layer LV myocardial LS, CS, GLS, GCS, and PSD. As shown in Table 10, our results suggest that our study generated reliable and consistent observations.