The present CMR study documents that increasing age is associated with a significant decrease in LV and RV volumes but an increase in EF. Furthermore, LV myocardial mass significantly decreased over the four age groups. There is increasing awareness on the fact that changes in cardiac ventricular structure and function attributable to the aging process may play an important role in modulating the cardiovascular response to disease [
7]. The knowledge of age-dependent normal structural and functional cardiac values is critical for identification of pathology and prediction of patient outcome. The rapid growth of the elderly population renders definition of normal values in advanced age even more important and may particularly help to reduce expensive and unnecessary follow-up examinations while there is a broad consensus on the decline in diastolic function with age. Controversial findings exist for age-related structural and functional cardiac values ranging from increase over no correlation to a decrease in LV-EF with advancing age [
7,
8,
11,
14], mainly obtained from echocardiographic studies. In fact, despite several limitations echocardiography has been the modality applied in most of these reports, reflecting the multi-purpose use of this widely available non-invasive tool. Comparable data as assessed from CMR are limited and particularly scarce for the RV. This is particularly important in view of the fact that in the past decade CMR has emerged as a new gold standard of non-invasive LV and RV assessment which is now increasingly available while the evaluation of mechanistic insights in metabolic and pathophysiological mechanisms remains a privilege of highly sophisticated CMR techniques such as tagging and spectroscopy [
15].
Differences in myocyte loss on the one hand and trophic effects on the other hand may eventually shift the tip of the balance towards changes in EF in the aging process. The observed decline in ventricular dimensions seems to favourably counterbalance the decrease in LV muscle mass resulting in an increase in EF. This is supported by the observation that twisting and torsion of the left ventricle are more pronounced in the elderly population [
16] and it may help explaining recent observations describing that normal EF can be present despite impaired myocardial contractility [
17]. While our results on a decrease in LV volume are in line with the majority of echocardiography and CMR studies, the findings with regard to LV mass seem more controversial. In some studies, LV mass was unchanged during aging [
12] while in other studies an increase was found [
14,
18,
19]. Interestingly, there was no association between age and LV mass in the Framingham study detectable by echocardiography [
20], while findings were described in a recent CMR analysis of the Framingham study [
21] showing a decrease over age. As the latter revealed LV mass as an important predictor of morbidity and mortality [
18], a decrease in LV mass could also reflect a survival selection bias. Due to its cross-sectional design our study avoids such bias and provides data from a real live cohort. However, for adequate interpretation it is crucial to review the techniques used, as the results may not necessarily be interchangeable [
22]. As mentioned above, most previous studies have assessed LV mass by M-mode echocardiography based on models which take into account LV wall thickness and short-axis dimensions neglecting the real LV shape. In addition, this method is dependent on operator skills. A recent study using latest advancements in echocardiography, i.e. real time 3-dimensional echocardiography, found the lowest values for LV mass and volumes in the eldest patients (seventh decade of life) largely in line with our results. This was further supported by other observations using CMR [
23,
24].
We acknowledge the following potential study limitations: First, in line with the data mentioned above we present a cross-sectional study of individuals at various ages, whereas patients would need to be followed in time to be more conclusive about age-related changes in cardiac parameters. Second, we did not verify our data with values obtained from another imaging modality and thus are not able to exclude potential modality bias. However, CMR is a well-accepted gold standard itself for such measurements [
12]. Third, despite the fact that this is the largest study analysing structural and functional cardiac values in CMR a potential bias towards supernormals as well as a survival selection bias cannot be excluded with final certainty. Furthermore, the limited sample size within subgroups of age does not allow gender-specific analysis. Fourth, we did not standardize loading conditions which may potentially have affected our measurements. However, in healthy individuals the impact of variations in loading conditions appears limited within a broad range of physiologic conditions. In addition, measurements without active interaction by the observer may prevent distortion of the findings closely reflecting real life conditions. Finally, we did not correct the chamber volumes for body size although an association between an increase in body surface area and an increase in chamber dimensions has been reported. However, the aim of the present study was to evaluate the structural and functional changes of the ventricles. As the aging process is per se associated with a decrease in body size normalization of the volumes would have at least in part masked the changes in ventricular volumes observed with aging. This would render meaningful comparison with changes in EF difficult, as for the calculation of EF the normalization cancels out resulting in an uncorrected value [
25,
26].