Atrial volume and function during exercise in health and disease

We prospectively included 15 healthy sedentary non-athletes, 15 healthy competitive endurance athletes and 15 CTEPH patients. Diagnosis of CTEPH was established in all patients by ventilation/perfusion scan, pulmonary angiography and right heart catheterization in accordance with contemporary guidelines [18]. None of the patients were receiving pulmonary arterial hypertension specific therapy, all were symptomatic but still able to exercise (NYHA class 2 or 3). We excluded patients with significant abnormal left ventricular (LV) function as a result of coronary artery disease or other cardiomyopathy, or with significant valvular heart disease as assessed previously with an exercise echocardiography.

Athletes were recruited from advertisements at local triathlon and cycling clubs and were included if they were participating in regular cycling and/or running training of >6 h/week. Non-athletes were recruited from advertisements among hospital staff members and were included if they were not engaged in regular sport’s practice (i.e. ≤1 h/week). None of the athletes and non-athletes met the exclusion criteria of known cardiovascular disease, symptoms, risk factors or abnormalities on electrocardiography and exercise transthoracic echocardiogram.

The study protocol conformed to the Declaration of Helsinki and was approved by the ethics committee of UZ Leuven (N° B322201214035). All subjects provided informed consent.

Demographic characteristics are detailed in Table 1. CTEPH patients were older than non-athletes and athletes, whilst there was no difference in gender. VO₂peak was significantly lower in CTEPH patients than in healthy subjects, and lower in non-athletes than in athletes (13.8 ± 3.0 vs. 35.0 ± 7.1 vs. 55.3 ± 10.5 ml/min/kg, respectively; P < 0.001).

At rest, athletes had larger iRAVmax and iLAVmax than CTEPH patients and non-athletes, the latter two groups having similar values (Table 2, Fig. 2). Resting iLAVmin was higher in athletes than in non-athletes and CTEPH patients. iRAVmin was larger in athletes compared to non-athletes, but similar to CTEPH patients. At rest, both RAEmF and LAEmF were lower in CTEPH compared to athletes and non-athletes.

From rest to peak exercise, iRAVmax and iRAVmin increased in CTEPH patients, whilst decreasing in athletes and non-athletes (P < 0.001 for interaction between exercise-intensity and group). Whilst the exercise-induced change in iLAVmax was similar between groups, iLAVmin decreased in non-athletes and athletes but not in CTEPH patients (P < 0.05 for interaction). As a result, exercise-induced increases in RAEmF and LAEmF as well as LA and RA reservoir were significantly impaired in CTEPH patients. Moreover, atrial remodeling was not symmetrical in CTEPH with a more dilated RA in comparison with LA or RV, as shown by a more elevated (RA/LA) maximal ratio and (RA/RV) diastolic ratio in CTEPH (Table 2).

Analysis was performed on 5 patients from each group with similar ages (mean age of 52.4 ± 7.1; 45.0 ± 13.3 and 43.2 ± 5.8 years in non-athletes, CTEPH patients and athletes respectively; P = 0.29) (an additional table shows this more in details: Additional file 2). Despite smaller group size, exercise changes in RA volumes remained significant with the same pattern of augmented function during exercise in healthy subjects and RA dysfunction in the CTEPH patients. The magnitude of these opposing effects was similar to that of the global population from which the age-matched sub-groups were derived (Fig. 3, an additional table shows this more in details: Additional file 3).

Similar to RA functional changes during exercise, RVEF augmentation was significantly impaired in the CTEPH cohort versus non-athletes and athletes (−1.1 ± 6.1%, P = 0.60 vs. +13.3 ± 6.2%, P < 0.001 and +10.7 ± 4.8% respectively, P < 0.001). In contrast, the exercise-induced increase in LVEF was similar between the different groups (+5.4 ± 7.6%, P = 0.020 vs. +9.3 ± 6.4%, P < 0.001 vs. +9.7 ± 5.3%, P < 0.001 in CTEPH, non-athletes and athletes respectively). RVEF correlated with RAEmF (R = 0.55 and R = 0.84, respectively for rest and peak exercise; P < 0.001) but also with LAEmF (R = 0.55 and R = 0.74, respectively; P < 0.001), no correlation was found between LAEmF and LVEF.

None of the CTEPH patients had significant tricuspid regurgitation, as confirmed by only a minimal difference in LV and RV stroke volume index at rest (35 ± 7 vs. 32 ± 6 ml/m², respectively; P = 0.004) and a similar LV and RV stroke volume index at peak exercise (33 ± 10 vs. 35 ± 8 ml/m²; P = 0.22). It has to be mentioned that there was no concordance between ventricular and atrial stroke volume index, neither between LA and RA stroke volume index (Table 2).

At rest, CI was lower in CTEPH patients than in healthy subjects but similar in non-athletes and athletes (2.6 ± 0.7 l/min/m² vs. 3.4 ± 0.8 vs. 4.1 ± 1.1; P < 0.001). As expected, mean pulmonary pressure was significantly elevated in the CTEPH patients, whilst non-athletes and athletes had normal mean pulmonary artery pressure. During exercise, there was greater mean RA pressure increase relative to cardiac output in CTEPH patients as compared with non-athletes (P < 0.001; Fig. 4). The peak exercise mean RA pressure was also significantly higher in CTEPH patients than in non-athletes (17 ± 7 vs. 5 ± 2 mmHg, respectively; P = 0.001).

At rest, mean RA pressure correlated with RA volumes (iRAVmax: R = 0.72, P = 0.002; iRAVmin: R = 0.77, P < 0.001) and RA function (RAEmF and RA reservoir: R = −0.58, both P = 0.018). At peak exercise, the association between mean RA pressure and RA volumes (iRAVmax: R = 0.70; P = 0.005; iRAVmin: R = 0.83; P < 0.001) and RA function (RAEmF: R = −0.87; P < 0.001; RA reservoir: R = −0.77; P < 0.001) remained strong (Fig. 4).

Within the group of CTEPH patients, none of the resting measurements correlated with VO₂peak on univariate analysis. However, VO₂peak correlated strongly with peak exercise CI, iLAVmin, LAEmF, LA reservoir, RAEmF and RA reservoir but not with RVEF (Table 3). On multivariate analysis, peak exercise RAEmF and iLAVmin were independent predictors of VO₂peak (ß =0.58 and ß = −0.52, respectively) in CTEPH patients and together explained 72% of the variance in VO₂peak.

During exercise, cardiac output has to increase despite a reduction in filling time due to higher heart rate. Thus it seems logical that an increase in reservoir function plays an important role in accelerating LV filling by helping to maintain an enhanced atrioventricular pressure gradient during diastole and also by increasing LA booster function to increase ventricular preload. These theoretical presumptions are derived from resting data [1] and our data provides direct validation during exercise. Here we demonstrated that the physiological response to exercise is a decrease in atrial minimal volumes and an increase in atrial emptying function with exercise. These results are consistent with a previous echocardiographic study in athletes [4]. Although athletes had higher atrial volumes than non-athletes, exercise-induced atrial adaptations were similar in both groups. This observation confirms that atrial dilation can be considered to be part of the physiological remodeling to long-term endurance exercise, i.e. athlete’s heart [17].

Although pulmonary hypertension patients are often relatively asymptomatic at rest, their exercise capacity is severely reduced and symptoms typically occur with exertion. Recent studies in patients with pulmonary hypertension and CTEPH using cardiopulmonary exercise testing or RV imaging have demonstrated that exercise parameters relate better to prognosis and provide more insight into the mechanisms of exercise intolerance than resting parameters [13, 14, 21, 22]. Whilst several studies have focused on ventricular pathophysiology during exercise in CTEPH patients, less attention has been given to exercise-induced changes in atrial volumes and function. The latter may be important, however, since resting mean RA pressure and iRAVmax are known strong predictors of outcome in patients with pulmonary hypertension [9‐11].

RA pressure will only start to rise when RV end-diastolic pressure increases and is therefore considered a measure of RV failure [23]. iRAVmax represents an alternative, non-invasive marker of the severity of RV dysfunction. However, iRAVmax is not only determined by RV dysfunction and increases in mean RA pressure, but is also influenced by other factors such as endurance training [24]. At rest, we found that iRAVmax was even smaller in CTEPH patients than in athletes, despite marked differences between these groups in terms of RV and pulmonary vascular function. The impaired RV functional reserve in CTEPH patients, reflected by the lack of RVEF augmentation, was associated with increases in iRAVmax and iRAVmin from rest to peak exercise in CTEPH patients, contrasting with decreases in RA volumes in healthy non-athletes and athletes. As a result, RAEmF was diminished during exercise in CTEPH patients relative to healthy subjects. Thus, the changes in RA volumes during exercise enabled a far better distinction between physiological and pathological RA dilation than resting measures. Furthermore, atrial function may be less prone to confounders than atrial volume measures. Indeed, even at rest, RAEmF was reduced in CTEPH patients, albeit to a minor extent, whereas RAEmF was similar in athletes and non-athletes. In addition, we found that RAEmF correlated highly with mean RA pressure, both at rest and peak exercise, suggesting that it may be used as a non-invasive surrogate to assess RA pressure overload. On multivariate analysis, peak exercise RAEmF and iLAVmin were independent predictors of VO₂peak and better explained exercise limitation in CTEPH patients than direct measures of RV function such as RVEF or cardiac index.

It is intriguing to speculate that RA function may be a more sensitive indicator of RV failure and pulmonary vascular disease than direct RV measures. The atria are thin-walled, compliant chambers with a tendency to dilate in response to chronic elevations in ventricular filling pressures. One hypothesis supported by invasive data [25, 26] contends that RV diastolic dysfunction precedes evidence of systolic impairment. As a result, changes in RA volumes and RA function may be anticipated and may precede evidence of RV systolic measures. This is analogous to the left-sided heart chambers where LA dilation and dysfunction is an early sign of LV diastolic dysfunction and is independently associated with survival [3]. Indeed, there is interplay among atrial function and ventricular performance throughout the cardiac cycle. Reservoir function is influenced by the descent of the ventricular base during systole, and conduit function is closely related to ventricular relaxation and compliance [20]. There was a discrepancy between atrial and ventricular stroke volume index. Indeed, atrial stroke volume is a misnomer, it is only a part of the total volume which actually goes through the atria, there is no method that is able to quantify this volume. Actually, the difference between LV and LA stroke volume defines atrial conduit volume [27, 28].

Another interesting finding of this study is that functional abnormalities were not limited to the RA, but also affected the LA. Indeed, CTEPH patients had reduced LA function at rest and exercise; and iLAVmin decrease during exercise was impaired in comparison to non-athletes and athletes. This blunted LA functional response to exercise without an increase in LA volumes may be explained by an interdependence between the right and the left side of the heart whereby impaired RV contractile and stroke volume reserve during exercise results in progressive underfilling of the left heart, thereby decreasing both LA and LV preload. This interdependence was emphasized by the finding that LAEmF correlated with RVEF but not with LVEF. Furthermore, increases in iRAVmax and RVEDV result in septal shift which leads to compression and reduced filling of the LA and LV given the relatively non-distensible nature of the pericardium [29]. Hence, increasing RV filling pressures during exercise related to RV failure can secondarily alter LV filling pressures and geometry [23]. Thus, despite underfilling of the LA and LV, pulmonary capillary wedge pressure may have actually increased during exercise due to right heart overload and ventricular interaction. In keeping with this premise, Andersen et al. [30] elegantly demonstrated that patients with severe tricuspid regurgitation have a higher pulmonary capillary wedge pressure, which is entirely explained by a disproportionate increase in mean RA pressure whereas LV transmural pressure, i.e. effective LV distending pressure, drops during exercise. Similarly, inadequate diastolic filling together with increased pulmonary capillary wedge pressures may have explained the dysfunctional response of the LA during exercise in the CTEPH population in our study. This finding is of potential clinical relevance given our finding that peak exercise iLAVmin was independently associated with VO₂peak, the strongest predictor of outcome in patients with pulmonary hypertension [13, 14].

Firstly, given the constraints of recruiting healthy subjects for an invasive study protocol, we did not attempt to match the whole population of healthy non-athletes, athletes and CTEPH patients for age. Nevertheless, the small subgroup comparison with subjects of similar ages demonstrated similar RA changes to those in the global study population, and it has been previously demonstrated that the decrease in RAEmF with age is only mild [31].

Secondly, pulmonary capillary wedge pressure was only measured at rest in the CTEPH patients, whilst no wedge pressure measurements were obtained in controls due to the serious potential for serious adverse events in performing these measurements during exercise. Therefore, we could not assess the degree to which pulmonary capillary wedge pressure influenced the changes in iLAVmin and LAEmF observed during exercise.

Lastly, we only assessed global atrial emptying function and reservoir function [20] but not conduit and booster function, as this is currently not feasible given the temporal constraints of exercise imaging. Nevertheless, previous studies on LA function have demonstrated that global emptying function is the best predictor of elevated pulmonary capillary wedge pressure [6] and a clinical relevant predictor of outcome in heart failure [3].

It might be valuable to study more in details the arterio-ventricular interaction occurring in CTEPH patients by assessing pulmonary arteria stiffness during exercise CMR [32]. Also, the validation of exercise CMR in other forms of pulmonary hypertension could contribute to widespread the use of this new technic in clinical practice.