A comparative analysis of conventional and speckle-tracking strain echocardiographic findings in diabetic and non-diabetic kidney disease patients with normal ejection fraction

Patients with chronic kidney disease (CKD) are at an increased risk of cardiovascular disease (CVD) compared to the general population. CVD is the most common cause of mortality in patients with CKD, accounting for 40–50% of deaths in those with advanced kidney disease [1, 2]. In addition to cardiovascular mortality related to atherosclerotic cardiovascular disease (ASCVD), mortality attributable to sudden cardiac death and heart failure is also significantly higher in CKD, compared to the general population [3].

Echocardiography is a simple, non-invasive, and affordable means of cardiovascular risk stratification in CKD. Studies have found that echocardiographic findings such as left ventricular hypertrophy (LVH), reduced ejection fraction (EF), and diastolic dysfunction predict future risk of adverse cardiovascular events and all-cause mortality [4‐6]. There is also a growing body of evidence that deformation analysis by speckle-tracking echocardiography (also called strain imaging) can identify cardiac abnormalities earlier than conventional echocardiographic techniques. This modality could potentially transform our understanding of cardiomyopathy in CKD, help in prognostication, and improve therapeutic decision-making [7].

Of those with CKD, patients with diabetic kidney disease (DKD) form a distinct sub-group. These patients account for approximately 40% of the CKD population and are at a higher risk of cardiovascular events, compared to their counterparts with non-diabetic CKD [8]. Little is known, however, about the differences in echocardiographic and strain parameters between these two groups. We sought to compare both conventional echocardiographic and strain parameters in patients with DKD and those with non-diabetic kidney disease (NDKD).

We studied echocardiographic parameters of patients with CKD stages 3–5 and normal ejection fraction and found that the ratio of peak early diastolic transmitral velocity to early diastolic mitral annular velocity (E/e’) was abnormal in nearly a third of patients. Further, E/e’ was higher in patients with DKD, compared to those with NDKD (11.9 ± 4.4 vs. 9.8 ± 3.5; p = 0.004). E/e’ ratio is an estimate of LV end diastolic pressure and is widely considered to be a non-invasive marker of LV diastolic dysfunction [15]. We also found that both septal e’ (7.1 ± 0.25 vs. 8.2 ± 2.8; p = 0.031) and lateral e’ (9.2 ± 2.9 vs. 10.4 ± 3.8; p = 0.045) were significantly lower in the DKD group, in addition to higher mitral E wave deceleration times (209.2 ± 41.5 vs. 189.1 ± 48.0; p = 0.017). While the lower septal and lateral e’ values indicate impairment in LV relaxation and early diastolic recoil, longer DT indicates LV stiffness-these are the key underlying mechanisms for diastolic dysfunction [16]. Despite the fact that both groups had similar LVMI and systolic function, the observed differences in the indices of LV diastolic function suggest that there are additional mechanisms at play in patients with DKD compared to those with other etiologies for CKD. In a study similar to ours, Miyazato et al. reported that echocardiographic parameters of LV diastolic function were more impaired in those with diabetic nephropathy, compared to those with chronic glomerulonephritis [17]. There too the authors observed no difference in LV structure and systolic function between the two groups. Another study by Han et al. too found higher E/e’ ratios in patients with DKD compared to those with NDKD [18]. It has been described that the diabetic state is associated with increased oxidative stress, elevated levels of pro-fibrotic and pro-inflammatory cytokines and accumulation of advanced glycation end products [19]. These abnormalities can lead to myocyte apoptosis, fibrosis, and LVH resulting in a condition referred to as diabetic cardiomyopathy-an entity that is still poorly understood [20‐22]. This condition manifests as diastolic dysfunction in the presence of preserved EF in the initial stages, with systolic dysfunction developing over time. Some of these factors like LVH and fibrosis are also operative in non-diabetic CKD [3, 23]. However, in addition, diabetic state is also associated with abnormal myocyte relaxation due to abnormal calcium handling in cardiomyocytes [24]. Prognostically, the presence of diastolic dysfunction assumes importance since evidence suggests that it is associated with higher all-cause and cardiovascular mortality [25]. Even in our study, patients with DKD had poorer cardiovascular outcomes in a relatively short follow-up period. These findings make a strong argument for the routine assessment of diastolic function in patients with DKD.

The role of LV deformation imaging (using speckle tracking echocardiography) for early detection of diastolic dysfunction remains unclear at present. Like e’, SR_E is a marker of LV relaxation in early diastole. However, technical difficulties and lack of reference standards make the interpretation of diastolic parameters challenging [16]. It has been suggested that the ratio of peak early diastolic transmitral velocity to early diastolic strain rate (E/SR_E) may be better than E/e’ ratio for detection of LV diastolic dysfunction in certain patient populations [26]. E/SR_E has also been described to be a superior to E/e’ for prediction of MACE in the hemodialysis population [27, 28]. In our study, however, we found no statistical difference between E/SR_E of the DKD and NDKD groups, although E/e’ was significantly different. Of note, E/ SR_E was numerically higher in the DKD group compared to NDKD (73.9 ± 32.4 vs. 66.8 ± 24.9; p = 0.183) and so the lack of statistical difference could be explained by the relatively small sample size. Moreover, the role of E/SR_E in pre-dialysis CKD has not been studied previously and is still uncertain.

A low GLS (> − 16%) suggestive of LV systolic dysfunction was seen in approximately 30% of our study population even though the study included only those with a normal LVEF. This confirms the findings of previous studies that strain imaging-derived GLS precedes changes in EF observed by conventional echocardiography [29‐31]. GLS has been shown to be a predictor of all-cause and cardiovascular mortality in several studies [7, 32, 33]. However, there was no difference in GLS between the DKD and NDKD groups in our study. This indicates that, although deformation imaging is able to identify the so-called “subclinical LV systolic dysfunction” in almost a third of these patients, DKD does not seem to be associated with a worse systolic function compared to those with NDKD in our study. Further, we found that patients with a low GLS (> − 16%) had higher E/e’ compared to those with normal GLS (12.6 ± 5.1 vs. 10.0 ± 3.3; p = 0.002). Although conventional wisdom is that GLS is a marker of subclinical LV systolic dysfunction, recent studies have suggested that it may provide additional information. A study by DeVore et al., showed that GLS may be a marker of myocardial fibrosis and hence diastolic dysfunction by showing an association between abnormal GLS and higher levels of biomarkers of collagen synthesis and NT-proBNP in blood [34]. The finding in our study that those with abnormal GLS had higher E/e’ (indicating diastolic dysfunction) suggests myocardial fibrosis as a possible explanation for impaired diastolic function. However, it remains unclear as to why DKD group had higher E/e’ compared to NDKD group even though both groups had similar proportion of patients with impaired GLS. It is possible that apart from myocardial fibrosis, there are additional factors such as abnormal myocardial relaxation (as described above) that account for higher E/e’ in the DKD group.

There were no differences in peak systolic strain rates (SR_S) between DKD and NDKD. As with diastolic strain rates, the lack of standardization in measurement techniques and cut-offs pose a challenge in the analysis of these newer parameters.

At the end of a median follow-up of 239 days, higher rates of 3-P MACE (11.5% vs. 4.9%; p = 0.047) and 4-P MACE (28.6% vs. 11.5%; p = 0.020) were observed in the DKD group, indicating poorer outcomes even in the short-term. Clearly, these two groups are clinically different in terms of cardiovascular outcomes even in the relatively short follow up period of our study. But the only difference with respect to echocardiography was in diastolic function. Both groups had similar systolic function even when sensitive parameter like GLS was used. Therefore, it is plausible that diastolic dysfunction could be one of the factors associated with future cardiovascular events. However, the number of events in our study is small and it is difficult to draw conclusions on the association of diastolic dysfunction (and other echocardiographic parameters) and clinical outcomes.

This study is unique on two counts. Firstly, we used speckle-tracking echocardiography, a more sensitive modality, to identify subtle differences in cardiac function in the DKD and NDKD groups. Secondly, we specifically included CKD patients with normal EF (in contrast to previous studies) because indices of diastolic function can be affected once left ventricular EF is reduced and may not necessarily indicate diastolic dysfunction per se. Moreover, patients with reduced EF represent those with a more advanced stage of cardiac dysfunction. Here, we attempted to identify early echocardiographic changes prior to the onset of frank LV systolic dysfunction by including only those with normal EF. The limitations of this study are the short duration of follow-up and the small sample size. Approximately one-fourth of patients in the NDKD group had diabetes mellitus. Although none of them had any other diabetes-related target-organ damage such as diabetic retinopathy, neuropathy or atherosclerotic cardiovascular disease at the time of inclusion in the study, the influence of diabetes on the myocardium in this group cannot be ruled out. Those with recent-onset diabetes may not have undergone myocardial remodeling and therefore, the duration of the disease could also have affected the findings. Moreover, a higher proportion of patients in the DKD group had hypertension, compared to the non-DKD group and this may have impacted echocardiographic findings and patient outcomes, even though mean systolic and diastolic blood pressures were similar in both groups at the time of entry into the study. Data on novel echocardiographic parameters such as LA strain (which is increasingly being recognized as a useful tool to assess diastolic dysfunction) and layer-specific strain is also lacking [35, 36]. Finally, this being a single-center study conducted in the Indian population, the generalizability of our findings may be limited.