Exercise-induced dynamic mitral regurgitation is associated with outcomes in patients with ischaemic cardiomyopathy

2.1 Study population

Consecutive patients with ischaemic cardiomyopathy (LVEF <50%), stable coronary artery disease and at least mild MR who underwent exercise echocardiography at the University Hospital Duesseldorf, Germany, were enrolled between 2018 and 2021. In all patients, echocardiography at rest and during specified handgrip exercise following a standardized protocol was performed. The majority of patients presented with HF symptoms and at least mild MR at rest. The study was approved by the ethics committee of the Heinrich-Heine University Duesseldorf and executed in accordance with the Declaration of Helsinki.

2.2 Echocardiographic examinations

Echocardiographic examinations were conducted using a GE Vivid E90 (Chicago, Illinois, United States) and a Philips iE33 (Amsterdam, Netherlands). Assessment of MR severity was performed according to current ESC guidelines.¹² An integrative approach using semi-quantitative and at least one quantitative parameter was used to assess the severity of MR, which was graded mild, moderate or severe. A detailed description of MR assessment is given in supplemental material. Ischaemic MR was defined in patients with known ischaemic heart disease and reduced LV function (LVEF <50%) with global and/or regional wall motion abnormalities, and with the mitral valve itself remaining intact (Carpentier IIIb).

Mitral annulus was measured at end-diastole and mid-systole in the parasternal long-axis view and in apical four chamber view and then averaged. Mitral valve (MV) tenting was measured using established MV tenting parameters (tenting area; tenting height), and assessed in the parasternal long axis view during mid-systole according current recommendations.¹⁵ Interpapillary muscle distance was assessed from parasternal short axis view in systole and diastole.

Right and left ventricular function and volumes were assessed according to current recommendations. Systolic pulmonary artery pressure (SPAP) was estimated from the regurgitant jet of tricuspid regurgitation (TR) with peak systolic trans-tricuspid pressure gradient (TTPG) calculated by the modified Bernoulli equation.

2.3 Isometric handgrip testing

Following echocardiographic examination at rest, handgrip exercise was conducted according to a standardized protocol with a commercially available handgrip dynamometer (Jamar® Hydraulic Hand Dynamometer, Sammons Preston Inc.) while the patient lay on his left side. After recording of heart rate and blood pressure at rest, the patient was asked to push the dynamometer with one hand with maximum effort, to assess maximal handgrip strength. Then handgrip exercise was performed out at 30% maximal force for 3 to 5 min as has been described previously.^{14, 16} Medical therapy (including beta-blockers) was not stopped for the exercise test.

2.4 Follow-up

The clinical course was monitored by follow-up examinations and phone calls to the referring cardiologists, primary physicians, or the patients themselves. We assessed a composite of all-cause mortality, HF-associated hospitalizations, MV surgery, transcatheter edge-to-edge repair (TEER), left ventricular assist device implantation and heart transplantation during follow-up as the combined endpoint.

2.5 Statistical analysis

Percentages were reported to describe categorical variables and median with interquartile range or mean ± standard deviation was reported for continuous variables. Normality distribution of continuous variables was tested with the Kolmogorov–Smirnov test. Comparison between two groups were made using the chi-square test (or Fisher's exact test if the expected count was less than five per cell) for categorical variables. Differences in continuous variables between two groups were compared for significance with a two-tailed paired t test. To assess differences between the three groups, analysis of variance (ANOVA) or Kruskal–Wallis test was performed for continuous variables. Cox regression analysis was used to assess the predictive value of echocardiographic parameters regarding event-free survival. Baseline and echocardiographic variables were tested in univariate and multivariate Cox regression analysis. Multivariate analysis includes age, sex, LVEF, LAVi, exercise TAPSE and exercise MR grade. Variables were checked for multicollinearity using variance inflation factors. MR grade, MR EROA and MR volume showed multicollinearity; thus, only MR grade was included into final analysis. Hazard ratios and 95% confidence intervals are presented. Kaplan–Meier analysis was used to evaluate the event-free rate. For all analyses, p-values of <0.05 were considered to be statistically significant. All analyses were performed using Sigma Plot (Version 11.0; Systat Software Ltd. Inpixon GmbH, Duesseldorf, Germany) and GraphPad Prism (Version 7; Graphpad Software, San Diego, California, USA).

3.1 Study population

We screened 539 patients that underwent echocardiography at rest and during handgrip exercise. After excluding 94 patients [because of image quality (n = 50), handgrip exercise physically not feasible (n = 22), severe aortic stenosis (n = 2), severe mitral stenosis (n = 1), previous mitral valve repair (n = 19)], the study cohort included 445 patients. Of these patients, 133 patients with ischaemic cardiomyopathy and reduced LV function (LVEF <50%) were identified that were enrolled in final analysis.

Patient characteristics are given in Table 1. Mean age was 75 ± 10 years, 21% were female. Mean LVEF was 35 ± 9%. In this regard, 50 patients (38%) presented with HFmrEF, and 83 patients (62%) presented with HFrEF. One hundred twenty-one patients (91%) underwent percutaneous coronary intervention and 44 patients (33%) underwent CABG.

3.2 Echocardiographic parameters at rest and during exercise

Seventy patients (53%) presented with mild MR, 54 patients had moderate MR (41%), and 9 patients (7%) showed severe MR at rest. Detailed data of mitral valve parameters are shown in Table 2. Approximately one third of patients (28.6%) presented with concomitant moderate or severe TR. Mean SPAP was 40 ± 12 mmHg.

Handgrip exercise led to a re-classification of MR severity in 49 patients (36.8%) (Figure 1). Twenty-five patients (20%) with non-severe MR at rest, developed severe MR during handgrip exercise (Figure 1). Together, 99 patients showed non-severe MR neither at rest nor during exercise, 25 patients developed dynamic severe MR, and 9 patients had severe MR even at rest. In the overall cohort, there were no changes in LV and RV volumes and function, despite an increase in LVSVi during exercise. SPAP increased by 7 ± 13 mmHg from rest to exercise. Detailed changes of echocardiographic parameters from rest to exercise were shown in Table 2.

3.3 Mechanisms and haemodynamic consequences of dynamic MR

To assess predictors of dynamic MR, patients were divided into tertiles according to the change in EROA from rest to exercise (Table S1): Patients with advanced exercise-induced increases in EROA (3rd tertile) had larger left atrial dimensions, increased wall motion score index and larger tenting area at rest (Table S1). During exercise, the increase in tenting area and tenting height was more pronounced in patients with marked exercise-induced increases in EROA (3rd tertile) compared with the other groups (Table S2). In this regard, there was a correlation between the change in EROA and the change in tenting area (r = 0.484; P < 0.001) and tenting height (r = 0.451; P < 0.001) (Figure 2).

Patients with marked increases in EROA during exercise (3rd tertile) showed a similar increase in LVEDVi (+7 ± 21 mL/m²) and LVESVi (+7 ± 17 mL/m²) during exercise, and thus, no change in LVEF (−1 ± 9%). In contrast, patients without dynamic MR showed a slight improvement in LVEF (1st tertile: +4 ± 10%; 2nd tertile: +4 ± 10%; P = 0.014). Moreover, patients with dynamic MR (3rd tertile) showed reduced LV forward flow (forward LVSVi) compared with the other groups (P = 0.033), as well as no increase in cardiac index during exercise (P = 0.029). At rest SPAP was similar in patients with and without increases in EROA (Table S1). Patients with marked increases in EROA (3rd tertile) showed advanced increases in SPAP during exercise (+10 ± 13 mmHg), compared with the other groups (P = 0.009). Thus, the SPAP/cardiac index ratio got worse in patients with dynamic MR (3rd tertile), compared with patients without dynamic MR (Figure 3). The change in MR severity from rest to exercise did not with changes in RA and RV diameter or function (Table 3).

3.4 Clinical outcomes according to the presence of dynamic MR

One-year follow-up (median 321 (162–427) days) was completed in 128 patients (96%). Seventy-one patients (56%) experienced the combined endpoint: 13 patients died (10%), 33 patients (26%) were re-admitted to hospital due to HF symptoms, 30 patients (24%) underwent MV TEER, six patients (5%) received MV surgery, one patient (1%) underwent LVAD implantation, and another three patients (4%) underwent heart transplantation. According to MR severity at rest, there was no difference in outcomes in patients with non-severe MR and severe MR (χ² 1.535; log-rank test P = 0.215) (Figure 4) (Table 5). As expected, patients with severe MR at rest more often underwent MV surgery/interventions compared with patients with non-severe MR (χ² 3.592; log-rank test P = 0.058) (Figure 4). In contrast, patients with severe MR during exercise showed adverse outcomes compared with those with non-severe MR (χ² 12.34; log-rank test P < 0.001) (Figure 4). Furthermore, patients with severe MR during exercise underwent MV surgery/interventions more frequently than patients with non-severe MR (χ² 31.18; log-rank test P < 0.001) (Figure 4). Excluding MV surgery/interventions from the combined endpoint, there was no difference between both groups at rest (χ² 0.262; P = 0.609) and during exercise (χ² 2.270; P = 0.132) (Figure S3). Multivariate analysis identifies MR severity during exercise (1.998 (1.367–2.938); P < 0.001), and exercise TAPSE (0.913 (0.853–0.973); P = 0.007) as predictors of the combined endpoint (Table 4). The need for MV surgery/interventions during follow-up was predicted solely by MR severity during exercise (HR 3.165 (1.832–5.768); P < 0.001) (Table 4). Patients with (dynamic) severe MR who underwent early MV surgery/interventions (<120 days from echocardiography) experienced numerically less adverse events (all-cause mortality; HF hospitalization; LVAD implantation; heart transplantation) than patients who were treated conservatively (14.3% vs. 44.4%; P = 0.072).

4.1 Prevalence of dynamic MR

Previous studies investigated the role of exercise testing in patients with ischaemic cardiomyopathy mainly focusing on bicycle exercise testing, and reported dynamic MR in 20%–50% of patients, depending on the definition of dynamic MR and the included patient cohort.^{7, 8, 11, 17} Until now, there is a paucity of data regarding the effects of isometric handgrip exercise on dynamic MR.^{14, 18-23} The current study underscores previous findings by unmasking dynamic severe MR during handgrip exercise in 20% of patients with non-severe MR at rest. Furthermore, in our study handgrip testing led to a re-classification of MR severity in 37% of patients. Thus, our results highlight the benefit of exercise echocardiography in patients with ischaemic heart disease and non-severe MR at rest. Symptomatic patients with dynamic severe MR might benefit from intensified HF medication or early interventional strategies.²⁴

4.2 Mechanisms of dynamic MR

Predictors of dynamic MR unmasked by dynamic bicycle exercise have been studied before.¹⁰ Now, this is the first study investigating the mechanisms of dynamic MR provoked by isometric handgrip exercise in patients with ischaemic MR. While dynamic bicycle exercise is known to decrease afterload, increase contractility and venous return (increase in preload), that together causes an increase in stroke volume, isometric handgrip exercise predominantly leads to an increase in afterload.¹⁰ These different loading conditions may translate into different LV geometric changes during exercise, and precipitate exercise-induced changes in MR severity.²¹ However, there might be several other factors influencing exercise-induced changes in MR as activation of the sympathetic nervous system and release of several hormones during exercise. We identified increased left atrial dimensions, parameters of local LV remodelling and wall motion score abnormalities at rest, as predictors of dynamic MR during isometric exercise. Furthermore, the change in MR severity was directly associated with the change in tenting height and tenting area. Thus, local LV remodelling rather than global LV remodelling seems to be responsible for worsening of MR during isometric exercise. Future studies comparing both exercise modalities are necessary to further elucidate these findings.

4.3 Haemodynamic consequences of dynamic MR

The haemodynamic response caused by handgrip exercise translated in a similar increase in LVEDVi (+8%) and LVESVi (+5%); thus, LVEF was unchanged during exercise (Table 2). This was accompanied by a reduction in LV forward flow in patients with marked exercise-induced increases in MR (Table S2). These changes might be explained by the sudden increase in afterload that in combination with the worsening MR outweighs compensatory LV mechanisms (Frank-Starling mechanism) and leads to a reduction in LV forward flow and cardiac index (Table S2). In addition to the decrease in LV forward flow, isometric handgrip exercise leads to an increased backward transmission and causes a pressure load on the pulmonary circulation, that was more pronounced in patients with dynamic MR (SPAP + 10 ± 13 mmHg) compared with those without dynamic MR (SPAP + 2 ± 7 mmHg) (Table S2). Thus, the SPAP/cardiac index ratio got worse during handgrip testing in patients with dynamic MR (Figure 3).

4.4 Impact of handgrip echocardiography on outcomes

In our cohort, MR severity during exercise, but not at rest, was predictive for the combined endpoint (Figure 4; Table 4). Excluding MV surgery/intervention from the combined endpoint there were no difference between the groups either at rest, nor during exercise. Solely MR severity during exercise (but not at rest) independently predicted the need for MV surgery/intervention during follow-up (Figure 4; Table 4). These results are in line with previous findings from Lancellotti et al. who investigated 161 patients with ischaemic MR undergoing bicycle exercise testing.¹¹ They found that an exercise-induced increase in EROA by >13 mm² was associated with mortality and HF hospitalizations.¹¹ In another study by Lancellotti et al., significant exercise-induced increases in MR during bicycle exercise unmasked patients at high risk of poor outcome.⁸ Peteiro et al. included 388 patients with LV dysfunction and demonstrated the incremental prognostic value of worsening MR during exercise.²⁵ On the other hand, there are some studies denying a prognostic impact of exercise-induced dynamic MR.^{22, 26} In our cohort, patients with (dynamic) severe MR who underwent early MV surgery/intervention showed favourable clinical outcomes compared with those who were treated conservatively. However, we only included a small number of patients with (dynamic) severe MR (n = 33). Only nine patients had severe MR already at rest. Together, there are only a few, conflicting data on the prognostic importance of dynamic MR during exercise in patients with ischaemic cardiomyopathy. This is the first study that demonstrated an incremental value of isometric handgrip testing over echocardiography at rest in patients with ischaemic MR. Further studies are warranted that randomize patients to additional exercise testing to further elucidate the role of exercise echocardiography in this patient cohort before the widespread use of exercise testing can be recommended.

4.5 Limitations

Main limitation of the current study is that patients were not randomized to exercise testing. Thus, there might be a selection bias for patients undergoing MV surgery/interventions. Moreover, cardiologists and cardiac surgeons were not blinded to the results of the exercise tests, thus, the decision for surgery/intervention might be by part influenced by the results of exercise echocardiography. Another main limitation is the size of the study cohort (n = 133), which prohibited subgroup analysis. However, the current study is the largest so far performing handgrip exercise for the assessment of dynamic MR in patients with ischaemic cardiomyopathy In addition, follow-up was complete in 96%. Prognostic data were missing in five patients that might have impact on outcome data.