Single beat 3D echocardiography for the assessment of right ventricular dimension and function after endurance exercise: Intraindividual comparison with magnetic resonance imaging

Extreme exercise might lead to right ventricular (RV) dysfunction and an elevated risk of arrhythmias in some athletes [1‐7]. To acquire RV data cardiovascular magnetic resonance imaging (CMR) and echocardiography have been proven valuable tools.

Acquisition of RV echocardiographic data has been conventionally proven difficult due to its anterior position in the chest, a complex geometry and morphology with prominent trabeculations. Over the years several parameters have been developed to determine RV function by 2D echocardiography, e.g. Tricuspid Annular Plane Systolic Excursion (TAPSE), Tei-Index, systolic pulmonary arterial pressure (sPAP), strain and speckle tracking or even simple "eyeballing". It is, however, highly dependent on a standardized acquisition since a slight drift in angle of view results in a significant change of measurements [8]. Therefore the modality of choice for evaluating right ventricular chamber size and function is still considered to be cardiovascular magnetic resonance imaging. Recent advances in transducer and computing technology make single beat 3D echocardiography (sb3DE) possible. It might be an alternative that is less cost intensive than CMR and can be used on patients with pacemaker or an implantable cardioverter-defibrillator (ICD). Since it is single beat (shorter acquisition time) it is also applicable in patients with arrhythmias or breathing disorders. It is however very dependent on good image quality and a fairly wide ultrasound sector angle to include the entire right ventricle. Previous studies have shown that even though volumes acquired with sb3DE are generally smaller than acquired with CMR there is a fair correlation in the measured ejection fraction [9].

Several studies have already assessed changes of RV function and dimension after endurance exercise. Most of these studies used a one-time event like an official marathon race or even an ultra-endurance competition to acquire data. In such a setting CMR measurements are delayed due to logistical limitations. In most studies, CMR was performed within one or two days post-race and after rehydration.

The goals of this study were

In this study we examined amateur athletes immediately after a run of 30 km with CMR and sb3DE, in a randomized order and compared the results. No more than two runners ran at a time to ensure immediate measurement after exercise.

22 athletes were screened. One runner showing signs of prior myocardial infarction (subendocardial delayed enhancement) in baseline-CMR was excluded from the study before the 30 km run. He was encouraged to undergo further cardiac diagnostics.

21 athletes ran in May and June 2010. All finished the 30 km run without adverse events. The average running time was 169,4 ± 17,4 minutes. The runners mean body weight decreased from 75,5 kg ± 7,9 to 73,2 kg ± 8,0. Table 1 shows the runners' baseline characteristics.

All post-run measurements were completed in less than 120 Minutes in all runners.

All subjects had sufficient image quality to perform sb3DE RV analysis. CMR showed no significant difference in the ejection fraction before and after the run, while RVEDV and RVESV and RVSV decreased significantly. (Table 2).

Sb3DE as well showed no significant difference in the ejection fraction before and after the race. However, although the mean values were lower, sb3DE did not show a significant reduction of RVEDV, RVESV or RVSV.

Both methods, CMR and sb3DE, showed a significant increase of the CI.

Inter- and intra-observer variabilities and intraclass correlation coefficients are shown in Table 3.

Standard 2D echocardiography showed neither a significant change in fractional area change (FAC) nor in tricuspid annular plane systolic excursion (TAPSE). It did, however, show a significant decrease of pulmonary valve acceleration time (PVAT) and pulmonary valve ejection time (PVET).

The volumes measured by sb3DE were significantly smaller than those measured by CMR (Table 4). However, there was no significant difference between the EF determined by the two methods (Figure 3). Figure 4 shows Bland-Altman plots comparing RV ejection fraction and dimensions measured by both methods.

Cellular blood components (erythrocytes, leukocytes, thrombocytes, hematocrit) as well as sodium and potassium showed a relative increase.

This study is the first head-to-head comparison of sb3DE and CMR in the analysis of RV dimension and function immediately after endurance exercise in well-trained athletes. We found no signs of RV dysfunction. However, there was a decrease in RV volumes. This decrease was significant in CMR, in sb3DE the mean values tended to be lower after the run but a high heterogeneity resulted in no significance.

Sb3DE was feasible in all subjects. Acquisition and RV analysis is possible within minutes. Sb3DE can be performed in tachycardic subjects after exercise. Data analysis in sb3DE is dependent on good image quality. While detection of the endocardial border was easily possible in the apical 4-chamber view it often was a challenge in short-axis and coronal views. To facilitate contour delineation, the software gives reference points based on the users previous input in former views. This, although being a helpful feature, makes measurement prone for sequence errors. Due to the RV anterior position in the chest, the RV anterior wall and outflow tract in the near field remains problematic in some subjects despite good image quality. This is a reason for the relatively high intra- and inter-observer-variability, which might be more pronounced in patients with RV disease or poor acoustic windows. An advantage of single beat 3D echocardiography is the lack of stitching artifacts and prolonged breath-hold, which will lead to a better acceptance of the method in routine echocardiography. Previous studies have used stitched volumes. The normal values for this method were published recently [8, 16]. Compared to these normal values, our subjects had larger ventricles; this is most likely due to the fact that we have examined well trained athletes [17]. The RVEF in our subjects is lower than the published normal values [16]. Still, it is in normal range in all subjects before and after endurance exercise. The intra- and inter-observer-variability in our study is comparable with published data [16, 18].

Although CMR is considered as the gold standard for RV volumetry and functional analysis [19] a clear identification of the tricuspid or pulmonary valve or the endocardial border is not always possible. Also, border delineation is user-dependent (e.g. how much trabeculation is included). In addition, relevant restrictions and contraindications have to be considered. Its use is limited in children, claustrophobic patients and patients with pacemakers or similar implants. Furthermore, it is relatively expensive and not easily available. Image acquisition requires a sufficient patient compliance.

Both sb3DE and CMR have relatively low volume/frame rates. Therefore, the definition of end-diastole and end-systole can be missed. This is especially difficult if the patient has a high heart rate (in our study the highest post-run heart rate was 111/minute).

RV dimensions and function by sb3DE were in fair correlation to CMR. As previously described, sb3DE defines lower volumes than CMR [9, 20, 21]. This is partly due to differences in data analysis since in sb3DE the endocardial borders are traced and most of the trabeculations are excluded while generally in CMR most of the trabeculations were included into the calculation of the cavity volume. In CMR analysis this is clinical standard because it leads to a faster and more reliable analysis [12, 22].

The discrepancy between volumetry in 3D echocardiography and CMR is more pronounced in larger ventricles [22, 23]. RV 3D echocardiography was recently validated against CMR in patients with congenital heart disease [9, 20]. Only 50-80% of the patients had a sufficient acoustic window for 3D echocardiography. In our study with ideal healthy subjects 100% could be analyzed. The RV analysis algorithm is designed for normal shaped RV and not congenital RV disease. Additionally, several data sets were acquired per athlete to ensure optimal data quality.

In sports cardiology, the issue of RV dysfunction after endurance exercise is not solved. Most CMR studies had a long delay of data acquisition after the race or did not mention the time point of echocardiography and CMR [3, 6, 24]. Neither CMR, nor sb3DE or standard transthoracic echocardiography showed RV systolic dysfunction in dehydrated subjects in our study. The changes of pulmonary valve acceleration and ejection time are within normal limits and probably due to the change in heart rate. Thus they do not indicate the presence of post-exercise pulmonary hypertension.

All other previously published studies that have investigated RV function after exercise have measured after rehydration and showed RV dysfunction [3, 6, 24].

Our laboratory results clearly underline the presence of relevant dehydration.

Our study is limited by the fact that we have not performed additional echocardiography and CMR after rehydration. It therefore remains unclear if rehydration through an increase in volume load leads to a change in RV function.

Both CMR and sb3DE are feasible in healthy athletes with high and low heart rates.

The two modalities showed a good correlation of RV function. Sb3DE obtained, however, smaller RV volumes compared to CMR. In conclusion, it is a sufficient alternative whenever it comes to image acquisition that needs to be done fast, with less expenditure or with patients that have contraindications for CMR.

In contrast to previous data, we could not detect RV dysfunction after exercise in either method. We assume that rehydration and volume load were important factors, which needs further investigation.