Myocardial Fibrosis and Cardiac Decompensation in Aortic Stenosis

Objectives Cardiac magnetic resonance (CMR) was used to investigate the extracellular compartment and myocardial fibrosis in patients with aortic stenosis, as well as their association with other measures of left ventricular decompensation and mortality. Background Progressive myocardial fibrosis drives the transition from hypertrophy to heart failure in aortic stenosis. Diffuse fibrosis is associated with extracellular volume expansion that is detectable by T1 mapping, whereas late gadolinium enhancement (LGE) detects replacement fibrosis. Methods In a prospective observational cohort study, 203 subjects (166 with aortic stenosis [69 years; 69% male]; 37 healthy volunteers [68 years; 65% male]) underwent comprehensive phenotypic characterization with clinical imaging and biomarker evaluation. On CMR, we quantified the total extracellular volume of the myocardium indexed to body surface area (iECV). The iECV upper limit of normal from the control group (22.5 ml/m2) was used to define extracellular compartment expansion. Areas of replacement mid-wall LGE were also identified. All-cause mortality was determined during 2.9 ± 0.8 years of follow up. Results iECV demonstrated a good correlation with diffuse histological fibrosis on myocardial biopsies (r = 0.87; p < 0.001; n = 11) and was increased in patients with aortic stenosis (23.6 ± 7.2 ml/m2 vs. 16.1 ± 3.2 ml/m2 in control subjects; p < 0.001). iECV was used together with LGE to categorize patients with normal myocardium (iECV <22.5 ml/m2; 51% of patients), extracellular expansion (iECV ≥22.5 ml/m2; 22%), and replacement fibrosis (presence of mid-wall LGE, 27%). There was evidence of increasing hypertrophy, myocardial injury, diastolic dysfunction, and longitudinal systolic dysfunction consistent with progressive left ventricular decompensation (all p < 0.05) across these groups. Moreover, this categorization was of prognostic value with stepwise increases in unadjusted all-cause mortality (8 deaths/1,000 patient-years vs. 36 deaths/1,000 patient-years vs. 71 deaths/1,000 patient-years, respectively; p = 0.009). Conclusions CMR detects ventricular decompensation in aortic stenosis through the identification of myocardial extracellular expansion and replacement fibrosis. This holds major promise in tracking myocardial health in valve disease and for optimizing the timing of valve replacement. (The Role of Myocardial Fibrosis in Patients With Aortic Stenosis; NCT01755936)

C alcific aortic stenosis is the most common valvular heart condition in the western world and a major public health burden (1).
In recent years, the role of left ventricular (LV) remodeling in disease progression, symptom development, and adverse cardiovascular events in aortic stenosis has been increasingly appreciated (2). In the initial phases, the increased afterload imposed by aortic valve narrowing induces adaptive left ventricular hypertrophy (LVH) that acts to maintain wall stress and cardiac output. Ultimately, this process decompensates, and patients transition from hypertrophy to heart failure and the development of symptoms and adverse cardiovascular events (2,3).
This transition often correlates poorly with the severity of aortic valve narrowing and is predominantly driven by myocardial fibrosis and myocyte cell death (4), which is perhaps a consequence of supply-demand mismatch and myocardial ischemia in the hypertrophied myocardium (2). Therefore, there is considerable interest in developing novel biomarkers to detect the early signs of LV decompensation.
Cardiac magnetic resonance imaging (CMR) provides the noninvasive gold standard method for measuring LV wall thickness, mass, volumes, and ejection fraction. Moreover, it is able to detect structural changes in the LV myocardium, including replacement fibrosis with the late gadolinium technique and expansion of the extracellular volume using T1 mapping (5). The latter in part reflects increases in diffuse myocardial fibrosis (a reversible early form of fibrosis) (6) and potential changes in the intravascular compartment. Early studies have suggested that CMRderived measures of LV mass and replacement myocardial fibrosis are of prognostic significance (7,8).
However, these studies have largely been conducted in small cohorts of patients with end-stage aortic stenosis who were referred to CMR on clinical grounds. Therefore, these findings may have been confounded by referral bias, which limited their applicability and generalizability to the broad population of patients with aortic stenosis. Moreover, comparisons with ageand sex-matched control populations and prognostic T1 mapping studies have been lacking.
We report the largest prospective study to evaluate systematically the usefulness of CMR in patients with aortic stenosis. In particular, we investigated its ability to detect expansion of extracellular volume (ECV) and replacement myocardial fibrosis, and how these are related to other markers of LV decompensation, functional capacity, and clinical outcomes. Cardiac biomarkers. Plasma cardiac troponin I (cTnI) concentrations were determined by the ARCHITECT STAT high-sensitivity cTnI assay (Abbot Laboratories, Abbott Park, Illinois) (9). The brain natriuretic peptide (BNP) concentration was determined with Triage BNP assay (Biosite Inc., San Diego, California).
6-min walk test. A 6-min walk test was performed in 156 (94%) patients as an objective measure of functional capacity in our predominantly older adult cohort, many of whom could not perform an exercise tolerance test. Explicit instructions were given to patients asking them to walk as far as possible for 6 min.

Myocardial Fibrosis in Aortic Stenosis
Erlangen, Germany). Short-axis cine images were acquired and used to calculate ventricular volumes, mass, and function. Left ventricular hypertrophy (LVH) was defined as LV mass (indexed to body surface area using the Du Bois formula) >95th percentile using age-and sex-specific reference ranges (10). LV longitudinal function was determined by measuring the difference in mitral annular displacement between end-systole and end-diastole (Online Appendix).
Focal replacement fibrosis and ECV expansion were assessed in all patients using late gadolinium enhancement (LGE) and myocardial T1 mapping, respectively. LGE was performed 15 min after administration of 0.1 mmol/kg of gadobutrol (Gadovist, Bayer Pharma AG, Barmen, Germany). The presence of mid-wall myocardial fibrosis was determined qualitatively by 2 independent and experienced operators (M.R.D. and C.W.L.C.), and its distribution was recorded (7,9). T1 mapping was performed using the Modified Look-Locker Inversion recovery (11) and a standardized image analysis approach (12). In the short-axis mid-cavity myocardium, 6 standard segments were defined on native T1 maps, and these regions were then copied onto the corresponding 20-min post-contrast maps (OsiriX version 4.1.1, Geneva, Switzerland).
Analysis of mid-ventricle segments has been shown to correlate well with analysis of all 17 myocardial segments, is simpler to perform, and avoids partial volume effects in apical segments (12). Segments with mid-wall LGE present were included in this analysis, whereas segments that contained subendocardial, infarct-pattern LGE were excluded. Four commonly used T1 approaches were assessed: native and post-contrast myocardial T1, partition coefficient (lambda), and the ECV fraction. We recently reported the reproducibility of these measures at 3-T (12).
We also investigated a novel marker, the indexed extracellular volume (iECV), which modifies the ECV fraction to act as a measure of the total volume of the extracellular compartment in the left ventricle. It was derived using the formula: ECV fraction Â LV enddiastolic myocardial volume normalized to the body surface area. Although a history of hypertension was more common in the aortic stenosis group, blood pressure was well-controlled and similar between the 2 groups at enrollment ( Table 1).  Values are median (interquartile range), n (%), or mean AE SD. Coronary artery disease was defined by previous myocardial infarction, clinical symptoms of angina with documented evidence of myocardial ischemia in the absence of severe aortic stenosis, a >50% luminal stenosis in a major epicardial coronary artery or previous coronary revascularization. All patients with clinical symptoms of angina underwent coronary angiography.
We explored iECV in greater detail, dividing our entire patient cohort into tertiles of iECV ( Table 3).
Using this approach, a steady increase across the tertiles was observed for each of the following markers of disease severity and LV decompensation: indexed LV mass, peak aortic valve velocity, plasma highsensitivity cTnI concentrations, serum BNP concentrations, diastolic dysfunction, longitudinal systolic dysfunction, and the proportion of patients with midwall fibrosis (p < 0.05 for all). Similar results were obtained using tertiles of the ECV fraction (Online Table 1), but by comparison, tertiles of LV mass index were less discriminatory, with no differences in diastolic function nor in serum BNP concentrations across these groups (both p > 0.05) (Online Table 2).
REPLACEMENT MYOCARDIAL FIBROSIS. Replacement mid-wall fibrosis, as assessed by LGE, was present in 44 (27%) patients with aortic stenosis but in none of the healthy volunteers. We examined the association between mid-wall myocardial fibrosis and the severity of aortic stenosis ( Table 2). Although patients with mid-wall fibrosis had more severe aortic p < 0.004) ( Table 4).
In the larger imaging cohort of patients with aortic stenosis (after exclusion of patients with an infarct pattern of LGE, n ¼ 22, or incomplete T1 mapping data, n ¼ 5), 71 patients had normal myocardium (iECV <22.5 ml/m 2 ). These patients had largely mild-to-moderate aortic stenosis, a mild hypertrophic response, minimal cardiac injury, and good Chin et al.    Patients with aortic stenosis were categorized into 3 groups based upon cardiac magnetic resonance (CMR) assessments of fibrosis.
Chin et al.  Values are median (interquartile range), n (%), or mean AE SD. Tables 1 to 3. Potentially, this can reflect increased myocardial fibrosis, myocardial infiltration, or expansion in the intravascular compartment (13). In aortic stenosis, myocardial fibrosis has been established pathologically as a key process that drives the transition from hypertrophy to heart failure (4). Moreover, we and others have observed a close correlation between these parameters and histological assessments of myocardial fibrosis (5,(14)(15)(16). However, there is some debate as to whether T1 mapping can provide direct assessment of the myocardium because of recent evidence that indicated that increased intravascular volume may also influence native T1 values (17,18).

Abbreviations as in
Pressure overload conditions such as aortic stenosis Although confirmation of this interesting hypothesis is required, it may be that T1 values are also related to myocardial ischemia that is believed to trigger fibrosis and the transition from hypertrophy to heart failure.
Regardless, T1 mapping remains at the very least a useful surrogate of myocardial fibrosis and LV decompensation in aortic stenosis.
Controversy remains as to the optimal T1 image analysis strategy (12,13,21). Consistent with previous research (12), both native T1 and the ECV fraction demonstrated major overlap with values in control groups and little difference among patients with mild, moderate, and severe aortic stenosis ( Table 2).
We sought to tackle this issue by developing a novel parameter, the iECV, which provides an assessment of the total ECV in the myocardium. This effectively combines the prognostic information provided by the ECV fraction with the improved discrimination between groups associated with indexed LV mass into a single measure (  Table 2). Our categorization holds promise as a means of monitoring the development of LV decompensation and helping to optimize the timing of aortic valve replacement. Currently, the development of symptoms guides the need for surgery. However, symptoms are frequently difficult to assess in older adult patients with multiple co-morbidities. Objective imaging assessments that monitor the changes in myocardial structure that are themselves responsible for progressive LV decompensation are therefore potentially attractive (2,3). This is the first study to describe iECV in aortic stenosis, so that confirmation of our findings in larger studies with longer follow-up is required.
However, we presented the fourth separate cohort to demonstrate the adverse prognosis associated with mid-wall LGE in aortic stenosis (7,8,22) and Prospective randomized controlled studies are required to investigate this further.