Relation between myocardial edema and myocardial mass during the acute and convalescent phase of myocarditis – a CMR study

An array of reports demonstrated the linear correlation between myocardial water content and T2-weighted cardiovascular magnetic resonance (CMR) relaxation [1‐7], justifying its use as an in-vivo marker of myocardial edema. Myocardial edema is a substantial feature of the inflammatory response in acute myocarditis[8, 9]. Edema is associated with diastolic dysfunction[10], conduction disturbances[11], microvascular compression[12] and tissue swelling[9]. The relation between myocardial edema and myocardial mass over the course of myocarditis has only rarely been systematically investigated in a clinical setting [13, 14]. This reflected the lack of a non-invasive imaging modality allowing the in vivo, accurate and simultaneous measurement of both parameters.

Indeed, histopathological quantification of edema is challenging. Over the last years, CMR has also become an important tool for the non-invasive assessment of myocarditis [15‐19]. On the other hand, steady state free precession (SSFP) CMR is the current gold standard to measure myocardial mass[20] and detects even minor changes during the course of diseases [21‐25] or in response to therapy. Sample size for studies of cardiac remodeling can be reduced with the use of CMR[26]. Based on these considerations we attempted to explore the relation between myocardial mass and T2-weighted CMR-imaging in a group of myocarditis patients during the acute and the convalescent phases of the disease.

We hypothesised that the resolution of myocardial edema in the course of healing myocarditis as visualised by T2-weighted CMR is associated with a measurable decrease in myocardial mass in SSFP-cine CMR.

The group of patients with acute myocarditis consisted of 16 men and 5 women with a mean age of 36.4 ± 16.1 years (mean ± SD) presenting to the emergency room. Initially, 76% of the patients (16/21) received a beta blocker and 95% an ACE-Inhibitor (20/21). Medication was continued in a substantial number of patients (beta blocker 43% [9/21], ACE-inhibitor 52% [11/21]) until after the follow-up CMR scan. At follow-up, all patients were free of symptoms and had no clinical evidence of ongoing inflammation. Blood pressure values during both CMR exams were within normal ranges (see table 1 for patients' characteristics). In 16 patients (76%) the presence of significant coronary artery disease (> 50% stenosis) was excluded by conventional X-ray coronary angiography. 3 of these patients (14%) were younger than 26 years and underwent this procedure because of their history of smoking. 5 patients (24%) did not undergo this procedure due to their age <26 years and a low pre-test probability of coronary artery disease (absence of cardiovascular risk factors, age ranging from 16 to 26 years with a mean age of 20.6 ± 3.6 years).

The control group consisted of 7 men and 4 female (age 31.6 ± 2.8 years, range 28–37 yrs.).

In patients, the LV mass in the presence of acute myocarditis was not overt hypertrophic but still within normal values at our institution. LV mass decreased significantly from 156.7 ± 30.6 g in the acute phase to 140.3 ± 28.3 g in the convalescent phase (p < 0.0001, see also figure 1 and 2). The average decrease of LV mass in the group of patients was larger than the interstudy variability in biplanar quantification of LV mass in our control group (16.3 ± 13.9 g vs. 5.1 ± 2.9 g, p = 0.02, see table 2). The left ventricular function (LV-EF) in cine imaging was within normal ranges in 18 patients (86%). However, LV-EF improved slightly but significantly in most patients at follow-up compared to the initial examination. The LV end-diastolic volume (LVEDV) did not change during the follow up period (quantitative results are given in table 3).

The observed reduction of LV mass in patients paralleled the normalization of initially increased myocardial signal intensity on T2-weighted images. However, no quantitative correlation was found between reduction of myocardial mass (10.2 ± 8.3%) and decrease of T2 ratio (28.5 ± 14.3%, p = 0.55) in the course of myocarditis.

11 patients (52%) had positive findings on LGE images at presentation and at follow up. There were however no significant differences in the severity of LGE findings (number or transmurality). Furthermore no correlation was found between the presence of LGE and LV mass reduction (p = 0.35) or LV mass at presentation (p = 0.86).

All of the patients had elevated cardiac serum markers (creatine kinase or troponin T or I) at presentation. We found no quantitative correlation between creatine kinase (p = 0.80) or elevation of troponin (p = 0.45) and LV mass.

This study describes a reduction of LV mass in SSFP-cine CMR during the course of healing myocarditis coming along with a normalization of initially increased values in T2-weighted CMR.

Hiramitsu et al[13] described 25 cases of biopsy proven acute myocarditis with a transient thickening of both the interventricular septum and left ventricular wall which was associated with the presence of histologically confirmed interstitial edema. The ventricular wall thickening as well as the edema resolved in the convalescent phase of the disease. These results are also supported by several echocardiographic reports describing a transient LV hypertrophy during inflammation[14, 27]. Following the conclusions of Hiramitsu et al., we do think that the observed decrease in LV mass in our study is comparable to their data and most likely reflects the regression of myocardial edema although we are not able to verify this finding with histological data.

Schroeder et al. [28] assessed 51 patients with acute myocardial infarction (AMI) and contemporary coronary reperfusion therapy for global and regional myocardial remodeling during the first year after AMI. They described an increased LV mass one week after AMI with a significant reduction in LV mass and LV wall thickness 6 months after AMI which was strongly associated with the size of infarction as assessed by peak creatine kinase-MB (CK-MB). The authors ascribe their observation of increased LV mass shortly after AMI to myocardial edema, because experimental studies have shown early interstitial edema of infarcted myocardium and increased LV mass [29] and animal studies by magnetic resonance imaging confirm these findings[1, 2, 4]. The article of Schroeder et al. supports the theory that myocardial edema leads to a measurable increase in LV mass in CMR.

The reduction of LV mass described in this report may in theory be a result of medical therapy rather than regression of edema. Due to the long time interval between the two CMR examinations, the observed reduction of LV mass might be an effect of medication with ACE-Inhibitors or β-blockers which were received by most patients. However, the initial LV mass was not hypertrophic but in the upper range of the normal values and none of the patients was hypertensive. Furthermore, the dosage of ramipril and metoprolol was rather low compared to the usual dosages described for regression of LV hypertrophy [30‐32]. Since to our knowledge no decremental effect of therapy with ACE-Inhibitors or β-blockers on LV mass in non-hypertrophic non-hypertonic hearts is reported, we tend to believe that the reduction of LV mass in our study comes secondary to regressive edema.

In our study, the reduction of LV mass was strongly associated with a decrease of signal intensity on T2-weighted images (T2-ratio), although we were not able to find a quantitative correlation of both parameters. Higgins et al. as well as many other authors [1‐5] describe a linear correlation between myocardial water content and T2-weighted cardiovascular magnetic resonance (CMR) relaxation. They found that despite the linear correlation, small changes of myocardial water content result in large T2-changes which may explain the relatively low difference of LV mass compared with larger changes in T2-ratio in the present study. The inversion recovery sequence we used is in fact not purely T2-weighted but is sensitive to tissues with long T1-relaxation times as well. This may have acted to 'amplify' the effect of free water – known to have long T1 and long T2 weighted relaxation times-resulting in a water-weighted image.

Left ventricular mass was not analyzed in contiguous short axis stacks that are considered the gold standard. Nevertheless we used 2D measurements for pairwise intraindividual comparisons both in patients and in controls. Previous studies found differences in LV mass between 2D and 3D measurements to be small[33]. In these studies, 2D quantification underestimated LV mass compared to 3D methods, thereby the true effect of tissue swelling might even be larger than reported.

We are well aware that the lack of endomyocardial biopsies is a certain limitation to our study. To ascertain the diagnosis of acute myocarditis we preferred a clinical approach. This is due to two reasons: First, the sensitivity of endomyocardial biopsy to diagnose myocarditis is limited and remains a controversial issue[34]. Second, our cohort of patients was rather young, presenting acutely with typical signs suggestive of myocarditis although none of them was severly ill in terms of heart failure. All patients revealed a preserved ejection fraction, there was no clinical need for endomyocardial biopsy. Although the theoretical possibility exists that some of our patients suffered from coronary artery disease, the invasive exclusion of coronary stenosis in patients with a low pretest probability makes this rather unlikely.

The small sample size of our study requires further evaluation of the hypothesis with larger studies and the fact that this is not a prospective study weakens the methodological quality of our work.

The possible effect of medication on our findings cannot be fully ascertained since the control group was not subjected to the same medical treatment.

A transient elevation of global myocardial T2 signal intensity is associated with a transient increase in left ventricular mass. Myocardial edema in the acute phase resolving at follow up could provide a potential explanation of our findings although other factors may contribute to this relation.