Introduction

Heart failure (HF) is the result of structural or functional abnormalities that induce a dysfunction of myocardial cells that jeopardize ventricular contraction. There can also be changes in the extracellular matrix such as higher collagen type I and III concentrations that imply fibrosis, as well as changes in the shape and size of ventricular chambers, usually termed as “myocardial remodeling”.₁

It has been found that fibrosis is a common finding in systolic and diastolic HF. The influence of myocardial fibrosis in the natural history of HF is expressed in several ways. Interstitial fibrosis contributes to the thinning of the ventricular wall, thus reducing compliance and contributing to the compromise of diastolic function. It is also to be considered that neither collagen deposits nor fibroblasts contribute to ventricular contraction, thus leaving a lesser cardiomyocite proportion to perform the systolic function. Perivascular fibrosis lengthens the distance that oxygen has to diffuse trough to reach myocardial cells, potentially diminishing the arterial oxygen pressure (PaO2) necessary to fulfill myocardial needs. Finally, electric coupling of cardiac myocites could be compromised by interstitial fibrosis.₁

Postmortem studies₂ have shown that myocardium exposed to chronic pressure overload because of systemic hypertension develop fibrosis. Other studies in transgenic mice₃ that received combination treatment with AT1 receptor blockers plus an endothelin ETA receptor blocker did not achieve a significant reduction in collagen deposits. Nevertheless, those that received a mixed endothelin ETA/ATB receptor (that are subtypes of the receptor that regulate collagen deposits) did not reach a significant blood pressure control but showed a reduction in collagen deposits, matching the post-mortem findings of myocardial fibrosis in the left and right ventricles, as well as in the interventricular septum. These findings suggest the presence of non-hemodynamic factors that contribute to this condition that could be at least partially induced by the loss of reciprocal regulation between molecules that stimulate or inhibit synthesis and degradation of collagen.₁

Cardiac magnetic resonance imaging (CMRI) is a novel study with important advantages including a high temporal and spatial resolution and enhanced tissue contrast. The study is also versatile to study cardiac dynamics. It has an excellent definition of the endocardial surface. This allows a precise delineation of the left ventricle cavity in contrast to the echocardiogram (ECHO) that needs the volumetric geometry to be assumed, especially in pathologic conditions that distort these patterns and modify ventricular geometry. CRMI achieves a three-dimensional image that allows direct geometric measurements.₄, ₅

Penta-acetic-dyethilentriamine acid Gadolinium (DTPA-Gd) is used as an intravenous contrast in CMRI (Gd-CMRI) based on its extracellular and blood distribution that allows a three-stage myocardial evaluation: an initial acquisition, a second stage early enhancement (3 to 5 minutes after injection) and a late enhancement phase, 15 minutes after injection. Enhancement allows defining transmural, non-transmural or subendocardiac lesions. According to Rehwald et al.₆ DTPA-Gd affinity is exclusively associated to injured tissue.

The search for myocardial fibrosis can establish the basis to prevent HF development or to establish prevention strategies to slow HF progression. Ideally, a post-mortem analysis (as those previously mentioned) or an endocardiac biopsy, are the preferred methods to establish the presence of myocardial fibrosis.

The quantification of myocardial fibrosis, if present, could possibly allow establishing a prognostic scale and thus establish treatment adjustments among these patients to favorably modify their prognosis and quality of life (QOL).

The present study was designed to determine if CMRI adding the analysis of wall ventricular motion could improve detection myocardial fibrosis in both ventricles and to correlate it with the New York Heart Association's functional capacity scale as well as to ECHO bad prognosis parameters among patients with HF.

Methods

Study population

Cardiac Magnetic Resonance Imaging Protocol

Every patient had an ECHO and CMRI performed. The cardiologists that performed the ECHO did not know the results of the CMRI. Magnetic resonance imaging was obtained by a 1.5 T scanner (General Electric, Sigma Excite 1.5T; Milwaukee, USA). The patients remained in supine position with radiofrequency flexible surface array over the torso for signal reception for 30 to 60 minutes. Motion or “cine” sequences (steady-state, free precession cines) were acquired with 8 seconds apnea periods (TE/TR 1.6/3.2 ms, with 60 spin angles) in sequential long and short axis 8 mm slices separated by 2 mm between them.

Short axis images were obtained from basal to apical segments beginning 1 cm below the aortic outflow tract y finishing at the apical region. Inversion times were adjusted to nulify myocardium (TI 260-400 ms), and the voxel's size was 1.7 x 1.4 x 0.8 Mm. Enhancement pattern was observed in the first pass after the intravenous injection of 0.2 mmol/Kg DTPA-Gd and at 10 to 15 minutes later (delayed enhancement).

Volumetric characteristics and left and right ventricular function were routinely analyzed. Wall contractility and post-contrast delayed enhancement were evaluated in six-segment divided slices. Regional wall motion was visually evaluated according to the following scale: 0=Normal, 1=Moderate hypokinesia, 2=Severe Hypokinesia, 3= Akinesia, 4=Dyskynesia. The mean extension of delayed enhancement's “transmurality” was also visually evaluated according to the next scale: 0= normal, 1= 1-25%, 2= 26-50%, 3=51-75%, 4=76-100%.

The heart was divided in 16 segments for motion and DTPA-Gd's enhancement analysis. Segmental analysis was performed in three axial projections: basal, medial and apical. Every frontal segment was classified as anterior, every segment below the horizontal line were termed as inferior, and through the vertical it was established that left segments were septal and the right ones lateral.

Every myocardial segment was evaluated to detect the presence of delayed enhancement compromising more than 50% of the wall thickness or less than that, and if its presence was related to myocardial perfusion and/or segmental wall motion abnormalities as detected by means of CMRI.

The values obtained from the CMRI structural evaluation were compared to the one´s obtained by ECHO. According to the CMRI findings, fibrosis was classified as “deposit fibrosis” and “ischemia-related fibrosis”. For comparative analysis purposes, four groups were defined: Patients without fibrosis (A), patients with “ischemia-related fibrosis” (B, those with imaging abnormality was observed thought vascular territories), “deposit fibrosis group” (C, in patches distributed heterogeneously) and a mixed fibrosis pattern (D) (Table 1).

Figure 1

Table 1: Fibrosis definition proposed according to myocardial characteristics segments

Statistical analysis

Results

Forty-five patients were included in the study, 24 (58%) were female, with an average age of 57.5±19.8 years. Twenty-four patients were in NYHA functional class I, 13 in FC II and 8 in FC III (53.3%, 28.9% and 17.8% respectively). The main co-morbidities and type of HF (systolic in 48.9%) are shown in table 2.

Figure 2

Table 2: Clinical characteristics of the study group

The basal ECHO and CMRI findings are listed in Tables 3 and 4.

Figure 3

Table 3: ECHO parameters of the study group

Figure 4

Table 4: Parameters defined by MRI

In 33 patients (73.3%) there was at least one affected segment that showed late enhancement and 29 (87.8%) of them had “fibrosis”. Thirteen of these patients (28%) had an “ischemic fibrosis pattern”, 11 (24%) a “deposit pattern”, and 5 (11%) had both patterns simultaneously.

When patients with and without fibrosis were compared, there was no difference by means of neither functional class nor Body Mass Index (BMI). “Fibrosis” with the “ischemia pattern” was found in 8 of 13 patients with diastolic dysfunction compared to 8 of 11 that showed a “deposit” pattern, without statistical significance (Table 5).

Group B patient's age was higher than Group's A and D patients. There were non-significant differences among hypertensive patients. Nevertheless, more patients with diabetes mellitus and dyslipidemia had a higher prevalence of the “ischemic” fibrosis pattern, in accordance to a higher incidence of ischemic heart disease (Table 5).

Figure 5

Table 5: Comparison according to the presence or not of fibrosis

Regarding CMRI results, patients in group B and D had higher left ventricle end diastolic volumes (LVEDV) when compared to those in group C. Left ventricle ejection fraction was significantly reduced in patients in group B when compared to group C. Patients in group D had the lowest LVEF among all groups (Table 6).

Figure 6

Table 6: Comparison between cardiac structural abnormalities according to the presence or not of fibrosis.

Patients in group D had lower LVEF and shortening fraction (SF) determined by ECHO than patients in group A and C. Patients in groups B and D had higher left LVEDV than patients in group C, and a higher end systolic left ventricle volume (LVESV). Patients in group D had significantly higher chamber's diameters, and had thinner walls, especially regarding the left ventricle's posterior wall.

A total of 720 cardiac segments were analyzed. Among them, 450 segments are related to the left anterior descending coronary artery (LAD), 180 segments related to the circumflex artery (CX) and 90 to the right coronary artery (RCA). Of all the analyzed segments, 72 (10%) had an “ischemia-related fibrosis” pattern. Among the latter, 58 (12.8%) corresponded to the LAD, five (2.7%) to the CX and 9 (10%) to the RCA, had the “ischemia-related fibrosis” pattern with a statistically significant difference (p=0.001).

Ninety-two (12%) segments showed the “deposit fibrosis” pattern, with a marked tendency to appear towards the interventricular septum, compared to the other ventricle walls, (62/225, 27.5%, p<0.001).

Discussion

In this study, the presence of “fibrosis” in HF patients did not correlate to the functional status neither to the presence of hypertension, as it has previously been described.₇ A possible explanation could be that 90% of this series patients are under treatment with ARAII and/or aldosterone antagonists, that are considered as important fibrosis-limiting factors.₈, ₉ We also found that patients with diastolic and mixed HF showed a tendency towards a higher incidence of “deposit fibrosis” pattern when compared to patients with systolic failure.

Patients with DM, dyslipidemia, or ischemic heart disease showed the “ischemia-related fibrosis” pattern in a higher proportion, a finding that can be explained by the fact that the two first one's are coronary atherosclerosis risk factors, and all of them induce coronary blood flow abnormalities that can lead to necrosis or “fibrosis”. The fact that these patients were significantly older than the rest of the studied population could also explain the sequence of physiopathologic events that induce post infarction fibrosis (ischemia, infarction, remodelation and fibrosis). The compromise of LAD territories is related to the fact that this vessel is the most commonly affected by ischemic events.

Cardiac MRI has proved to be useful to detect “fibrosis” zones in infarcted myocardium. Kim et al.₁₀ observed that canine myocardium showed late enhancement (294±54% more intense than normal signal, p<0.001) during acute MI. The infarcted zones were analyzed 8 weeks later (chronic infarction) and they also showed late enhancement (253±54% more intense than normal signal, p<0.001), thus establishing a good correlation between acute and chronic MI (r=0-97, p<0.001). On the other hand, areas that were exposed to reversible ischemia showed an intensity similar to the normal one (98±6% more intense than a normal image, p=NS). These findings suggest that CMRI can recognize irreversibly or reversibly damaged tissue; independently from myocardial mobility. In another study₁₁ was evaluated late enhancement by cine MRI as well as contractility before and after a revascularization procedure in 50 patients. Two-hundred and fifty-six out of 329 (79%) of the analyzed myocardial segments without enhancement prior to the revascularization procedure, showed a contractility improvement when compared to only one of the 58 segments with late enhancement that compromised more than 75% of the left ventricle's thickness (p<0.001). The first patients also showed an improved LVEF in the late control study (p<0.001). It was then demonstrated that late enhancement areas associated to perfusion abnormalities do not recover contractility and finally develop ventricular remodeling. Ventricular remodeling could thus explain that patients with the “ischemia” and mixed patterns fibrosis had higher ECHO and CMRI LVEDV when compared to patients with the “deposit fibrosis” pattern, which presumably would be hypertrophic.

In a Study with 24 patients with coronary heart disease and DTPA-Gd reinforcement, Ramani et al.₁₂ showed that late enhancement was present in areas of irreversible myocardial damage, as opposed to areas with reversible damage. They found that 58% of the hypokinetic segments that were considered as viable by nuclear imaging with Thallium201 showed late enhancement, thus suggesting the presence of mixed viable and fibrotic tissue in the same area. This could justify the need to simultaneously evaluate movement defects and image enhancement to better define a segment as non-viable functionally speaking, or in other words, with “established fibrosis”. Patients could then show “ischemia” and “deposit” fibrosis patterns, with the latter considered as the main pattern. The combination of both patterns could explain the observation that these patients have more severe functional and structural ventricular compromise (higher diameters and LVPW as well as lesser LVEF and SF), indicating that the “deposit fibrosis” pattern could constitute a severity marker and even a bad prognostic marker too.

Several authors suggest that in infiltrative diseases, late enhancement is better correlated to pathological deposit areas, as in cardiac sarcoidosis, where CMRI has showed its usefulness to detect infiltrative processes.₁₃

McCrohon et al.₁₄ showed that both patients with Ischaemic heart disease (IHD) and dilated cardiomyopathy (DCM) had transmural or subendocardiac late enhancement, but they differed in that some patients with DCM showed mid-wall or longitudinal late enhancement without being related to a specific coronary artery territory. This was observed in 18 of 63 (28%) of patients with DCM only, a finding consistent with focal-segmental “fibrosis” as the one observed in autopsy studies.₁₅, ₁₆ Other autopsy studies have demonstrated that endocardiac and transmural fibrosis in DCM is indistinguishable from the one observed in myocardial infarction.₁₇ Nevertheless, recent evidence₁₈ shows also contradictory results: The late enhancement pattern specific of DCM was present in 69% of the study group and that such finding was consistent with our study's results regarding the higher frequency of interventricular septum involvement, but endocardiac biopsies did not significantly correlate to the presence of enhancement (p=0.85)

Considering these results, the simultaneous evaluation of segmental mobility and enhancement pattern could allow a more precise judgment of the meaning of such enhancement and its prognostic implications, since the enhancement pattern alone does not seem to accurately reflect, at least in some patients, the true histological characteristics of the dysfunctional myocardium.

The results of this study show that patients with systolic HF had more frequently an “Ischemic-related fibrosis” pattern and that those patients with diastolic failure had more frequently a “deposit” pattern. Both “fibrosis patterns” affected more frequently the interventricular septum and the LAD territory. Patients with “mixed fibrosis” pattern (group C) had the higher functional and structural compromise both by ECHO and CMRI.

Late enhancement could represent myocardial tissue with several lesion degrees whose reversibility is still controversial. Gadolinium enhanced CMRI has showed its capability to identify as “fibrosis” patterns in the presence of different co-morbidities and has also demonstrated that those patterns usually are related to functional and structural compromise, as determined also by ECHO. Nevertheless, those enhancement patterns should be simultaneously evaluated with segmental mobility and a rigorous clinical evaluation in order better define the prognostic implications of this kind of conditions, necrosis or fibrosis, which, at the moment is yet controversial.

Correspondence to

Dr. Arturo Orea Tejeda, Instituto Nacional de Ciencias Médicas y Nutrición “SZ”, Providencia 1218 A 402 Col del Valle CP 03100 Mexico City, Mexico. Tel/Fax: (5255) 5513 9384, e-mail: artorea@yahoo.com.mx.