Longitudinal shortening remains the principal component of left ventricular pumping in patients with chronic myocardial infarction even when the absolute atrioventricular plane displacement is decreased

Myocardial infarction (MI) leads to decreased left ventricular function due to the necrosis of myocytes and subsequent replacement with fibrous tissue and is the most common etiology to heart failure [1]. Cardiac magnetic resonance (CMR) imaging can quantify both left ventricular function using cine images as well as MI using late gadolinium enhancement (LGE) images. CMR has been used to study the relationship between the size of an MI and the overall left ventricular ejection fraction (LVEF) [2, 3]. The relationship is not straightforward but there is an upper limit of ejection fraction determined by infarct size [4]. Both infarct size and ejection fraction are important prognostic factors after an MI [5]. Myocardial infarction also affects longitudinal ventricular function and this can be evaluated as the atrioventricular plane displacement (AVPD) [6]. A decreased AVPD is associated with worse prognosis after an MI [7]. In echocardiography, measurements of AVPD or mitral annular plane systolic excursion (MAPSE) are used for detection of decreased ventricular function. The interest in longitudinal function has recently increased as global longitudinal strain seem to be a stronger predictor than ejection fraction for mortality [8] and lateral MAPSE from CMR has been shown to be an independent predictive factor for cardiac events [9].

Cardiovascular magnetic resonance imaging has the ability to quantify the amount of stroke volume (SV) generated by longitudinal function by combining the measurements of the AVPD and the short-axis area of the left ventricle encompassed by the AVPD, and express this as a percentage of SV [10, 11]. In healthy controls ̴ 60% of the left ventricular (LV) SV is generated by longitudinal function and the rest from radial (circumferential) function seen as the systolic radial inward motion of the epicardium [10‐12]. The radial component of LV pumping can be subdivided into the movements of the septum and the remaining anterior, lateral and inferior walls, denoted lateral wall for short. Normal septum movement contributes ̴ 8% to LVSV and the lateral wall ̴ 30% [13]. The size of an MI affects overall pumping, but it is not known to what extent an MI affects the longitudinal contribution to LVSV and if the size and location of an MI may have impact on this parameter. The rationale of this study was to increase the fundamental physiological understanding of longitudinal left ventricular function in chronic MI patients so that the clinical tools of longitudinal function, e.g. global longitudinal strain and MAPSE can be better understood. Therefore, we aimed to quantify the effect of chronic MI on the longitudinal, septal and lateral contributions to left ventricular pumping.

Parameters of LV size and function for the study groups are presented in Table 2. Left ventricular volumes showed a positive correlation with infarct size, EDV = 7.3xIS-15.9, r = 0. 75, p < 0.001 and ESV = 7.78xIS-1.6, r = 0.82, p < 0.001. Ejection fraction had a negative correlation with IS, EF = −1.18xIS-0.8, r = 0.79, p < 0.001. Altogether, patients with MI in the LAD vessel territory had larger volumes and lower EF compared to both patients with RCA-MI and controls (Table 2). The differences between the patient groups were also seen for infarct size with LAD-MI being larger than RCA-MI (p < 0.001). RCA-MI patients also had larger EDV and lower EF compared to controls). Atrioventricular plane displacement was decreased in both LAD-MI (11.0 ± 0.7 mm) and RCA-MI (13.0 ± 0.9 mm) compared to controls (15.3 ± 0.4 mm, p < 0.001) but did not differ between the patient groups (p = 0.08). The epicardial short-axis areas used for calculation of AVPD contribution to SV were increased in infarct patients (p < 0.01) but in the subanalysis, LAD-MI had larger epicardial areas compared to controls (p < 0.01) while RCA-MI did not reach significance (p = 0.06).

Figure 2 shows the mean extent of the MI for all LAD-MI and RCA-MI patients, respectively.

Longitudinal contribution to LVSV was numerically lower in patients compared to controls but this difference did not reach statistical significance (p = 0.08 for LAD and p = 0.09 for RCA) (Table 2, Fig. 3). Septal contribution to LVSV was lower in LAD-MI compared to both RCA-infarcts and controls (p < 0.01) (Table 2, Fig. 4). Lateral contribution to LVSV was higher for LAD-MI (p < 0.001) and RCA-MI (p < 0.05) compared to controls (Table 2, Fig. 5). AVPD showed a negative correlation with infarct size in LAD-MI (r = −0.59, p = 0.006) but not in RCA-MI (r = −0.18, p = 0.51) patients. However, there was no correlation between infarct size and longitudinal contribution to LVSV in LAD-MI (r = −0.114, p = 0.63) or RCA-MI (r = −0.059, p = 0.83) patients. Furthermore, infarct size did not correlate to septal contributions to LVSV in LAD-MI (r = 0.08, p = 0.74), or in RCA-MI (r = −0.45, p = 0.08), or lateral contributions to LVSV in LAD-MI (r = 0.05, p = 0.84) or in RCA-MI (r = 0.20, p = 0.46).

Internal validation of the measurements by adding the longitudinal, septal and lateral contributions to stroke volume showed 107 ± 16% for LAD-MI, 104 ± 8% for RCA-MI and 103 ± 9% for controls. Interobserver variability for average epicardial areas used for calculation of longitudinal contribution was −0.7±1.7 cm² and for AVPD 0.5±1.0 mm. Intraobserver variability of AVPD in 20 patients analyzed >3 months apart, was 0.3 mm ± 0.7 mm.

This study has shown that longitudinal shortening remains the principal contributor to LV stroke volume in patients with chronic MI even when the absolute AVPD is decreased. This is mainly explained by the LV dilatation after an MI causing a larger short-axis epicardial area. The presence and degree of LV dilatation showed a correlation with infarction size. The three-dimensional intrinsic movement of the myocardium during the cardiac cycle is complex. However, when viewing the LV from the epicardial border, the movement is simpler, comprising of a longitudinal shortening and a radial inward motion [11, 13, 18]. Longitudinal contribution can thus be understood by viewing the LV as a piston pump. The movement of the piston is the AVPD and the volume ejected by longitudinal shortening can be calculated by multiplying the piston movement by the piston area. Thus, if the longitudinal shortening of the ventricle is viewed as a piston, if the “piston” area become enlarged after an MI this can result in an unchanged stroke volume even when the piston movement is decreased.

Of note, the SV and heart rate, were unchanged in MI patients compared to the age-matched controls. Thus, cardiac output was similar in patients and controls. The longitudinal contribution did not differ between infarcts in the LAD or RCA territory. Septal contribution to SV, however, was decreased and lateral contribution increased in patients with MI in the LAD compared to controls. These findings are as expected from the localization of LAD-MI in the anterior and septal walls (Fig. 2). Thus, this study has demonstrated how a chronic MI can affect novel parameters of LV pumping to different degrees depending on what coronary artery was occluded.

Our findings of preserved longitudinal contribution to LVSV are in line with the results of a previous study in patients with dilated cardiomyopathy [11]. In both patients with dilated cardiomyopathy and patients with MI, the AVPD was decreased but the dilated LV with a larger short-axis epicardial area resulted in a longitudinal contribution to LVSV that was similar to healthy controls. In contrast, patients with pulmonary regurgitation have decreased longitudinal and increased lateral contributions to SV in both the left and right ventricles [13]. The septal movement in patients with pulmonary regurgitation is towards the right ventricle resulting in a negative contribution to LVSV. The decrease in longitudinal contribution to RVSV was also seen in catheter-induced PR in an animal study and was reversible after percutaneous valve replacement [19]. Recently, a study on patients with pulmonary hypertension, found decreased longitudinal contribution and increased lateral contribution to LVSV similar to pulmonary regurgitation [17]. The preserved longitudinal contribution in patients with dilated cardiomyopathy or MI where the LV is primarily affected and dilated thus differs from patients where the RV is dilated because of pulmonary regurgitation and pulmonary hypertension. Further mechanistic studies are needed to explain the pathophysiology behind these differences.

We did not find any relationship between MI size and regional contributions to LVSV. This may be due to a limited patient population and larger studies may be needed to shed further light on a possible relationship. Larger sized myocardial infarction results in worse remodeling and reduced ejection fraction at follow up, and because of the larger myocardium at risk in anterior LAD-MI these are related to worse remodeling [20]. This was also evident in our findings of larger infarct size, lower AVPD and larger LV volumes in the LAD-MI compared to RCA-MI.

The reason for using the epicardial area of the LV to calculate the AVPD contribution to SV has been previously described and validated [10, 11]. Similar to any engine, when calculating the volume ejected due to a piston movement the entire area of the piston needs to be taken into account. The myocardial wall is incompressible and does not change volume during pumping, apart from the negligible small variations of volume caused by coronary blood flow, SV calculations are identical from endocardial or epicardial delineations. The re-arrangement of the myocardium during systole to a thicker wall is both caused by the longitudinal shortening resulting in thicker wall and the radial thickening per se of the myocardium. Furthermore, longitudinal pumping causes the majority of the thickening of the LV muscle and this thickening does not reflect LV radial pumping [18].

The mean longitudinal contribution to LVSV on a group level did not differ from controls but the study of individual patients show differences in how LV remodeling after an MI affect cardiac pumping. Fig. 6 illustrates how the regional contributions to LV pumping can differ between patients. In the LAD-MI patient a large anteroseptal-apical infarct is seen in the LGE images and the septal epicardial contour does not shift during systole, resulting in a zero septal contribution to SV. This patient has a severely decreased AVPD and thus a smaller longitudinal contribution compared to both controls and the average of LAD-MI patients. To compensate for this decrease, there is an increased movement in systole of the lateral epicardial contour, resulting in a high proportion of lateral contribution to radial pumping. In contrast, the patient with an MI in the RCA territory has almost zero movement of the lateral epicardial contour, compared to the LAD-MI patient, i.e. decreased lateral contribution to SV due to the infarct in the lateral and inferior wall. This is compensated by a normal AVPD resulting in an increased longitudinal contribution to SV as the epicardial area in this patient is increased. These cases can be contrasted with a patient shown in Fig. 7 with a large LAD-MI and an apical aneurysm had 90% longitudinal contribution (Fig. 7). This patient had normal exercise capacity and few symptoms within one year after infarction. Thus, the increased piston-pumping of the AV-plane in this patient can produce a near normal aerobic capacity during stress. Thus, patients with LAD-infarctions and apical aneurysms can have both decreased and increased longitudinal function and the prognostic and functional importance of this finding remain to be elucidated.

In patients with MI a decreased AVPD is associated with worse prognosis [7] and recently Rangarajan et al. showed that a reduced lateral AVPD on CMR was an independent predictor of major adverse cardiovascular events [9]. Previous studies of measurements of left ventricular AVPD or mitral annular plane systolic excursion (MAPSE) by echocardiography have been used clinically to detect decreased ventricular function and shown to hold prognostic information in stable coronary artery disease [21]. However, it is not known if regional contributions to LVSV hold any prognostic information. Studies including larger patient numbers and prognostic information on e.g. exercise capacity, morbidity and mortality as well as potential additional diagnostic value from the AVPD, septal and lateral contributions to LVSV, would be of interest.

Automatic software for AVPD tracking exists and can be used to minimize user dependency of the AVPD measurements [22‐24]. In addition, this provides additional information of the AVPD over the entire cardiac cycle. Examples of patient groups where AVPD is especially important are LV hypertrophy and patients with aortic stenosis where an increased radial movement compensates for the decreased longitudinal function [25]. Further studies to determine if the longitudinal/radial contribution to LVSV is affected in these patient groups are therefore warranted.

Chronic MI in the LAD and RCA vessel territory does not change the internal relationship between the contributions to overall LV pumping of longitudinal (≈60%) as compared to radial (≈40%) pumping. Although the overall contribution to SV of radial pumping did not change, MI within the LAD vessel territory caused a decrease in the septal component of radial pumping which seems to be compensated by an increase in its lateral component.