This study currently represents the most extensive CMR investigation conducted on healthy individuals, aiming to provide reliable normal reference values for diastolic function parameters over the age range between 10 and 70 years and for both sexes. An attempt was made to distribute the participants evenly over all 6 decades to improve statistical accuracy. The diastolic function was characterized by examining the flow pattern of the right upper PV blood flow and the transmitral blood flow using flow-sensitive phase-contrast CMR. Biventricular and left atrial function as well as left ventricular strain were also assessed to obtain a comprehensive picture of the diastolic function.

The main novelties of this study are as follows:

i)

CMR demonstrated significant gender-differences in absolute values of PV and transmitral flow, however their clinical impact appears subtle and is reflected in gender comparable functional ratios, per example by PV S/D and MV E/A.

ii)

CMR characterizes significant age-dependent alterations in PV and MV hemodynamics, most notably evidenced by increasing PV S/D ratios and decreasing MV E/A ratios as age advances.

iii)

CMR and echocardiographic acquired values differ significantly, implying high intermodal discrepancy and the need for separate CMR reference values.

iv)

CMR intra- and interobserver variability analysis consolidates reliable levels of reproducibility, particularly regarding flow assessments.

In recent years, there have been a few studies in which diastolic dysfunction has been investigated using CMR. Typically, patient cohorts were only compared with a limited number of healthy individuals, so that a clear distinction from the pathologic state is limited [25,26,27,28]. Two studies have previously assessed the age dependence of diastolic function parameters using phase contrast CMR to provide reference values [12, 29]. However, the focus was mainly on mitral blood-flow parameters, while we considered both mitral and PV blood-flow parameters as well as the LA passive/active emptying fraction.

In particular, it seems that the influence of patient ageon diastolic function parameters was insufficiently considered in the 2016 ASE/EACVI guideline [3]. As underscored by Popovic et al., the ASE classification carries the risk of classifying people as sicker or healthier than they are. According to this classification, healthy individuals specifically under the age of 30, commonly showed signs of grade III diastolic dysfunction, as their E/A ratio was found > 2 [24]. To close this gap, the present study included not only young individuals, but also a broad age range.

Gender dependency of the diastolic function

Although absolute velocity values such as PV Smax-wave, PV D-wave or MV E-wave do show gender-differences, they are more related to known physiological gender-differences such as height, weight, body surface index or indexed LV stroke volumes (Tables 1 and 2). However, since the relevant parameters for describing the diastolic function are mainly based on ratios (PV S/D, MV E/A) or represent slopes or deceleration times, they remain independent of absolute velocity values or integrals. Consequently, the influence of gender appears to have a minor clinical impact. Therefore, all study participants were utilized to investigate the age-dependency.

Age dependency of the diastolic function

This study assessed the age-dependent alterations of diastolic function that occur as a consequence of left atrial and left ventricular remodeling and compliance reduction with advancing age. Moreover, these observations emphasize the strong atrioventricular interdependence, that may translate into pathophysiological developments. CMR assessment determined significant increase in the PV S/D ratio and significant decrease in the MV E/A ratio as the age of healthy individuals increased.

In this study an S/D ratio of > 1 was found for the age-group > 30 years, which is in line with other studies [15, 30], whereas an S/D ratio < 1 was determined in the two younger groups. This is a very important issue and needs to be carefully considered as such ratios are typically only found in patients with moderate to severe reduced LV compliance [3, 15].

The assessment of the S/D ratio for the staging of diastolic dysfunction according to American Society of Echocardiography is limited in young people, as they often have an S/D ratio < 1. However, this was described as physiological in the study by Mandinov et al. [15]. Further points of criticism of the current grading for diastolic function were highlighted in the study by Buffle et al. [31] This study emphasized the diagnostic and prognostic clinical value of PV-flow parameters. Their research identified S/D integral ratio as the best predictor of heart failure readmission and the most effective diagnostic parameter for detecting early stage of diastolic dysfunction [12].

Moreover, age-related changes in the flow pattern were reflected by the flattening of the PV slope of the D wave in heart-healthy subjects (see Fig. 3). Hence, the PV slope D-wave could serve as an additional valuable metric for delineating the subclinical diastolic impairment of left ventricular compliance that manifests with age progression.

In healthy young people (< 30 years), early diastolic filling dominates the LV filling phase because of low LV diastolic pressure, increased compliance and rapid LV relaxation [3, 15, 32]. This explains the E/A ratio > 2 in the two younger study groups. Normally the E/A ratio is > 1, but an impaired ventricular relaxation and reduced ventricular filling can lead to E/A values < 1 [3, 15, 32]. However, even healthy people with normal diastolic function can have an E/A ratio < 1, for example with increasing age, as found in our data. However, this appears to be a physiological adaption, as stated by Schirmer et al. [33]. This can be explained by the fact that early diastolic filling is attenuated, while late diastolic filling, resulting from atrial contraction, increasingly provides the compensatory force for LV filling. This compensatory effect is further confirmed by the strong positive correlation between age and the LA-active emptying fraction displayed in this study.

Comparison between echocardiography and CMR data

Strong correlations were observed between both imaging techniques for the most important measures describing diastolic function such as the PV S/D ratio and the E/A ratio, but also for most other diastolic function parameters. Buss et al. demonstrated even stronger correlations for the E/A ratio, examining however a heterogenous group of patients with various cardiovascular diseases combined with healthy volunteers [12]. Furthermore, it should be noted that the majority of PV and MV flow parameters presented significant intermodal differences as the absolute values of the velocity-based parameters of the PV and transmitral flow pattern were generally higher with echocardiography. This is consistent with other researchers demonstrating higher transmitral peak velocities with echocardiography compared to phase-contrast CMR [12, 25, 34]. A possible explanation to the observation could be that echocardiography imaging determines flow data in real-time, whereas CMR acquired flow data is typically averaged over a longer time interval of several seconds up to 2–3 min. This averaging prevents, to a certain amount, the instantaneous and accurate estimation of peak velocities and leads to a lowering of the velocity values when CMR is applied. The discrepancies underscore the need for modal-dependent reference values to prevent misclassification between diastolic function and dysfunction. Conversely, Ramos et al., found reduced absolute values for echocardiographic parameters in contrast to CMR acquired parameters [27]. Despite the high level of CMR reproducibility reflected in the inter- and intraobserver variability results, the normal values in this study require comparison with larger cohort data to establish age- and gender dependent reference values.

Other recommended variables for determining whether LV diastolic function is normal or abnormal are the mitral E/é ratio and the PCWP, which were also evaluated in this study. Although differences between the two imaging techniques were statistically significant, the values of both variables indicate that they are within the expected normal range [3].

Limitations

The present study was conducted as a cross-sectional study at a single-center. Causality cannot be determined with the study design. The study is limited to a Caucasian cohort of healthy volunteers. Since fewer individuals in the > 50 age group meet the criteria for cardiac healthiness, the generalizability of the age-related results in this group may be somewhat limited. However, efforts were made to select optimal candidates within this age segment who met criteria for cardiac healthiness to mitigate potential bias.

To prevent different testing conditions and possible physiological variations between the CMR and echocardiographic examinations, all participants were assessed either immediately before or after completing the examination procedure of the other modality on the same day and in the same local unit. However, it should be mentioned that (1) CMR was conducted in the supine position, whereas for echocardiography a left supine position was used to obtain standard views, and (2) the study participants had to transition between examinations. The impact of these slightly different physiological conditions, such as an altered heart rate or variations in blood-flow characteristics, remains undetermined.

A key strength of this study is that it utilizes standard equipment, such as a clinical MR scanner, routine CMR pulse sequences, commercially available analysis software, to assess a large proportion of CMR parameters such as ventricular/atrial volumes, strains, LA-emptying fraction, PCWP and é. These parameters are feasibly assessed and consequently well suited for routine clinical use. The only exception is the description of the PV and transmitral blood-flow pattern to assess the PV slope D-wave, the MV slope E-wave and deceleration times, which required analysis on an external workstation. The inclusion of these parameters in clinical practice would necessitate Medical Device Regulation (MDR) approval. While this was not the aim of our study, it would be highly valuable, as both slopes showed a strong statistical correlation with patient age.