Prognostic Significance of Nonischemic Myocardial Fibrosis in Patients With Normal LV Volumes and Ejection-Fraction

Objectives This study aims to investigate the prognostic significance of late gadolinium enhancement (LGE) in patients without coronary artery disease and with normal range left ventricular (LV) volumes and ejection fraction. Background Nonischemic patterns of LGE with normal LV volumes and ejection fraction are increasingly detected on cardiovascular magnetic resonance, but their prognostic significance, and consequently management, is uncertain. Methods Patients with midwall/subepicardial LGE and normal LV volumes, wall thickness, and ejection fraction on cardiovascular magnetic resonance were enrolled and compared to a control group without LGE. The primary outcome was actual or aborted sudden cardiac death (SCD). Results Of 748 patients enrolled, 401 had LGE and 347 did not. The median age was 50 years (interquartile range: 38-61 years), LV ejection fraction 66% (interquartile range: 62%-70%), and 287 (38%) were women. Scan indications included chest pain (40%), palpitation (33%) and breathlessness (13%). No patient experienced SCD and only 1 LGE+ patient (0.13%) had an aborted SCD in the 11th follow-up year. Over a median of 4.3 years, 30 patients (4.0%) died. All-cause mortality was similar for LGE+/- patients (3.7% vs 4.3%; P = 0.71) and was associated with age (HR: 2.04 per 10 years; 95% CI: 1.46-2.79; P < 0.001). Twenty-one LGE+ and 4 LGE- patients had an unplanned cardiovascular hospital admission (HR: 7.22; 95% CI: 4.26-21.17; P < 0.0001). Conclusions There was a low SCD risk during long-term follow-up in patients with LGE but otherwise normal LV volumes and ejection fraction. Mortality was driven by age and not LGE presence, location, or extent, although the latter was associated with greater cardiovascular hospitalization for suspected myocarditis and symptomatic ventricular tachycardia.

T here has been rapid growth in the adoption of cardiovascular magnetic resonance (MRI) imaging for diagnostic evaluation, surveillance, and assessment of treatment response across the spectrum of cardiovascular (CV) disease.
Appropriate-use criteria highlight the evolution in complexity and capability of MRI to support clinical decision-making (1,2).
Replacement fibrosis (scar) identified by late gadolinium enhancement (LGE) indicates an adverse prognosis in many conditions, which are all characterized by abnormal left ventricular (LV) volumes and/or LV ejection fraction (LVEF) (3)(4)(5)(6). However, myocardial fibrosis remains a powerful predictor of sudden cardiac death (SCD) even when the severity of LV dysfunction is only modest (7).
Increasing numbers of individuals are identified with normal LV volumes, wall thickness, and LVEF and previously unrecognized myocardial fibrosis. In one of many series, subendocardial fibrosis suggesting myocardial infarction was present in 17% of people older than 67 years of age with incremental prognostic value beyond standard clinical predictors including LVEF (8)(9)(10)(11). However, there is a paucity of data on the prognostic significance of nonischemic fibrosis in the midwall and/or subepicardium of people with normal LV volumes, wall thickness, and LVEF. Fibrosis represents the final common pathway of injury from a diverse range of diseases and insults (12). Whether nonischemic fibrosis is a risk factor for SCD in the absence of other structural markers of disease such as LV dysfunction/dilation is unknown.
Moreover, the etiology of midwall/subepicardial fibrosis is often unclear and ascribed to remote events, such as previous myocarditis. Uncertainty surrounding the clinical significance and management of such cases leads to conflicting advice, such as refraining from high-intensity exercise, for which there is no evidence of benefit, multiple investigations at considerable expense, and some risk that may also heighten anxiety for the patient, their family, and their physicians (13).
To the best of our knowledge, no study to date has specifically investigated the outcome of people with normal LV volumes and ejection fraction with nonischemic patterns of LGE and no other manifestation of cardiac disease. agents at baseline were included, but not those with resistant hypertension (15). Patients with >50% stenosis in a major coronary artery, infarct pattern of LGE, left bundle branch block, coronary bypass grafting, or percutaneous intervention were also excluded. Finally, MRI data for remaining individuals were manually curated to exclude indexed LV values outside the appropriate age-and sex-adjusted normal ranges (14). In total, 456 patients met the stringent criteria to define a structurally normal heart         LGE >2.    LGEþ group as a percentage of overall LV mass was 2.25% (IQR: 1.21%-4.14%).
LGEþ patients were more likely to be men (P < 0.0001) with a history of controlled hypertension (P < 0.0001) and receiving treatment with a beta blocker or ARB (P < 0.001). NYHA functional class was also more likely to be II/III (P < 0.0001).
LGE-patients were more likely to be women and have lower LVEDVi (P < 0.0001), and lower LV mass index (P < 0.0001) within the normal ranges. There were no significant differences between groups in age, comorbidity, or scan indication.
In the LGEþ group, those with nonseptal LGE were more likely to be men (P ¼ 0.029) and to present with chest pain (P < 0.0001). Patients with septal LGE were more likely to be women (P ¼ 0.029), to present with breathlessness or for familial cardiomyopathy screening (P < 0.0001), to have atrial fibrillation (P ¼ 0.028), and to be prescribed an ACE inhibitor (P ¼ 0.027). There were no significant differences between groups in age, baseline medical history, or medication.   Figure 3B, Table 2). The etiology for CV death was worsening heart failure from newly developed ischemic heart disease (both in men aged >70 years). Non-CV deaths included cancer (n ¼ 18; 64%), pneumonia (n ¼ 7; 25%), end-stage renal failure (n ¼ 1), leukemia (n ¼ 1), and motor neuron disease (n ¼ 1). There was no association between LGE location or extent and allcause mortality.  Table 1.  Figure 3D). Other variables that showed an association on univariable analysis (    Table 2).

GENETIC SEQUENCING. Thirteen of 39
LGEþ patients (33%) with a family history of cardiomyopathy as the scan indication (although phenotypically unaffected) were subsequently found to have rare variants in genes associated with cardiomyopathy (Supplemental Table 1). Of these patients, 3 had ICDs    Given the identification of a single aborted SCD event, it is challenging to confirm whether this represents a true negative study outcome or a type II error. Future work is required with larger patient groups, but this initial study suggests this patient group is at a very low risk of SCD.

CONCLUSIONS
Our data provide new information on the prognostic significance of nonischemic patterns of LGE in a large, well-characterized cohort of patients with normal LV volumes and ejection fraction. We demonstrate, for the first time, that there is a reassuringly low risk of actual or aborted SCD in this population. All-cause mortality was driven primarily by age-related disease and was not associated with the presence of LGE.
These findings do not support aggressive medical management or the routine use of ICD implantation within this cohort.