Unrecognized myocardial scar by late-gadolinium-enhancement cardiovascular magnetic resonance: Insights from the population-based Hamburg City Health Study

Abstract

Background

The presence of myocardial scar is associated with poor prognosis in several underlying diseases. Late-gadolinium-enhancement (LGE) cardiovascular magnetic resonance (CMR) imaging reveals clinically silent “unrecognized myocardial scar” (UMS), but the etiology of UMS often remains unclear. This population-based CMR study evaluated prevalence, localization, patterns, and risk factors of UMS.

Methods

The study population consisted of 1064 consecutive Hamburg City Health Study participants without a history of coronary heart disease or myocarditis. UMS was assessed by standard-phase-sensitive-inversion-recovery LGE CMR.

Results

Median age was 66 [quartiles 59, 71] years and 37% (388/1064) were females. UMS was detected in 244 (23%) participants. Twenty-five participants (10%) had ischemic, and 217 participants (89%) had non-ischemic scar patterns, predominantly involving the basal inferolateral left-ventricular (LV) myocardium (75%). Two participants (1%) had coincident ischemic and non-ischemic scar. The presence of any UMS was independently associated with LV ejection fraction (odds ratios (OR) per standard deviation (SD) 0.77 (confidence interval (CI) 0.65–0.90), p = 0.002) and LV mass (OR per SD 1.54 (CI 1.31–1.82), p < 0.001). Ischemic UMS was independently associated with LV ejection fraction (OR per SD 0.58 (CI 0.39–0.86), p = 0.007), LV mass (OR per SD 1.74 (CI 1.25–2.45), p = 0.001), and diabetes (OR 4.91 (CI 1.66–13.03), p = 0.002). Non-ischemic UMS was only independently associated with LV mass (OR per SD 1.44 (CI 1.24–1.69), p < 0.001).

Conclusion

UMS, in particular with a non-ischemic pattern, is frequent in individuals without known cardiac disease and predominantly involves the basal inferolateral LV myocardium. Presence of UMS is independently associated with a lower LVEF, a higher LV mass, and a history of diabetes.

Graphical abstract

Diabetes, LVEF, and LV mass index were independently associated with unrecognized myocardial scar. Red arrows indicate areas of myocardial scar. LV: left-ventricular, LVEF: left ventricular ejection fraction.

1. Background

Myocardial scar detected by cardiovascular magnetic resonance (CMR) late-gadolinium-enhancement (LGE) imaging represents focal myocardial fibrosis/necrosis originating from ischemic or non-ischemic disease [1], [2]. The presence of myocardial scar was consistently found to be associated with poor prognosis in several clinical settings and underlying diseases, in particular as the substrate for ventricular arrhythmia [3], [4], [5], [6]. CMR phase-sensitive inversion recovery (PSIR) LGE imaging is currently the reference technique for assessing myocardial scar non-invasively [2], [7]. Of note, LGE imaging enables the detection of clinically silent and otherwise unrecognized myocardial scar (UMS) [8], [9], [10], [11], which is of diagnostic and prognostic relevance [8], [12], [13], [14]. Currently available data from population-based studies revealed a prevalence of UMS of up to 30% [8], [9], [10], [11]. However, the majority of currently available studies have focused on ischemic UMS and there is a paucity of data on non-ischemic UMS [8], [9], [10], [11]. The proof of non-ischemic myocardial scar is a key diagnostic feature in several entities, such as in clinically suspected acute myocarditis [15], but pre-existing, unspecific myocardial scar of a different origin constitutes a potential confounder in this context. Thus, there is a need for a better understanding of factors associated with the presence of non-ischemic myocardial scar. This study evaluated prevalence, localization, and patterns of ischemic and non-ischemic UMS as well as associations of UMS with cardiovascular risk factors and CMR features in the population-based Hamburg City Health Study (HCHS).

2. Methods

The HCHS was approved by the local ethics committee (State of Hamburg Chamber of Medical Practitioners, PV5131) and all participants of HCHS gave their written informed consent.

2.1. Study population

The rationale and design of the general HCHS have been published before [16]. Briefly, HCHS is representative for the urban city of Hamburg and provides a large spectrum of interdisciplinary data on participants [16]. This study is based on the currently available cohort of the first 10,000 HCHS participants who were prospectively included between February 8, 2016, and November 7, 2018 (Fig. 1). Within this population, CMR was performed in 2589 individuals, of whom 1164 accepted the optional administration of contrast media. For this analysis, we excluded all participants with a history of coronary heart disease or myocarditis by hospital or by surveillance records. We did not exclude individuals with cardiovascular risk factors, such as hypertension or diabetes, since they represent the majority of this age group in the general population of Germany [17], [18]. Hypertension and diabetes were defined by standard criteria (resting blood pressure ≥140/90 mmHg, fasting glucose ≥126 mg/dL, respectively, or antihypertensive/antidiabetic treatment) [17], [18].

Fig. 1.

Study flowchart. Within the first 10,000 HCHS participants, CMR was performed in 2589 individuals, of whom 1164 accepted the optional administration of contrast media. All participants with a history of coronary heart disease (n = 89) or myocarditis (n = 11) were excluded. Individuals with cardiovascular risk factors, such as hypertension or diabetes, who represent most of this age group in the general population of Germany were included. The final study population consisted of 1064 participants, in which we found 244 individuals with unrecognized myocardial scar. CE: contrast enhanced, CMR: cardiovascular magnetic resonance, HCHS: Hamburg City Health Study.

2.2. CMR protocol

CMR was performed on a 3T scanner (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany), a detailed description of CMR in the HCHS has been published before [19]. Briefly, the comprehensive HCHS CMR protocol consisted of steady-state free precession cine, T2, native and contrast enhanced (CE) T1, LGE imaging, and flow measurements [19]. PSIR LGE imaging was performed 10–15 min after administration of 0.15 mmol/kg gadoterate meglumine (Dotarem, Guerbet, Aulnay, France) short-axis and three additional long-axis slices. Typical imaging parameters were repetition time (TR) = 3.0 ms, echo time (TE) = 1.2 ms, inversion time (TI) adjusted according to the Look-Locker images to 300–360 ms, FA: 55°, acquisition matrix = 256 × 185, slice thickness 8 mm with 2 mm gaps, effective pixel spacing 1.6 × 1.6 mm² [19].

2.3. CMR data analysis

A standardized workflow was established consistent with current guidelines as described before [19], [20]. Image analysis was performed by using the commercially available software cvi42 (Circle Cardiovascular Imaging Inc., Calgary, Canada, version 5.6.6). For the analyses of this study, LGE images were analyzed by a consensus reading of two experienced physicians with 6 and 20 years of experience in CMR (E.C. and K.M.), who were blinded to the medical history of the participants [19]. The presence of myocardial scar was assessed visually and categorized into ischemic and non-ischemic patterns [1], [7]. Briefly, ischemic scar was assumed if scar primarily involved subendocardial layers, respecting coronary artery supply areas, whereas non-ischemic myocardial scar was assumed if scar primarily affected intramyocardial and/or subepicardial layers, independent from coronary supply areas [1], [21]. Furthermore, non-ischemic myocardial scar was categorized as “major” (defined as a “typical,” clearly pathologic pattern, i.e., mid-wall/epicardial fibrosis in dilated cardiomyopathy or myocarditis, diffuse fibrosis in infiltrative cardiomyopathies, such as cardiac amyloidosis, or patchy/intermediate fibrosis in pathologic left ventricular hypertrophy) and “minor” LGE patterns (defined as unspecific phenomena, not fulfilling the criteria for “major” scar, i.e., LGE near the aortic root or the mitral annulus or at the right ventricular insertion point without significant left ventricular hypertrophy), as previously proposed [13]. Localization of scar was assigned to segments of the standard 16 segment model [22].

2.4. Statistical analysis

Continuous data are presented as median and quartiles. Categorical data are presented as absolute numbers with percentages. Pairwise comparisons of continuous/categorical variables between groups were performed using Mann-Whitney/Fisher test, and the Hodges-Lehmann median shift estimator plus 95% confidence interval (CI) was provided. Potential differences between the CE-CMR and the underlying general HCHS population were assessed by the parametric bootstrap method. The significance level was set to α = 0.05. In subjects without any missing value, the influence of the remaining variables of interest on the presence of myocardial scar, ischemic scar, and non-ischemic scar was assessed by multivariable analyses using logistic regression. Stepwise selection by both Bayesian information criterion (BIC) and Akaike information criterion (AIC) were performed. Odds ratios (OR) and CI were reported for dichotomous and continuous variables, respectively. All analyses were conducted using R 4.2.0 in conjunction with R Markdown.

3. Results

3.1. Study population

One thousand one hundred sixty-four individuals underwent CMR and accepted contrast media application within the first 10,000 HCHS participants. One hundred individuals were excluded from this analysis because of a history of coronary heart disease (n = 89) or myocarditis (n = 11). Thus, the final CE-CMR study population consisted of 1064 participants without history of coronary heart disease or myocarditis (Fig. 1). Participants of the final CE-CMR study population were older and the proportion of male gender, smoking, and hypertension were higher compared to the underlying unselected, consecutive first 10,000 participants of the general HCHS population (Table 1). Median LV volumes and mass indices were higher in the CE-CMR study population, but there was no significant difference in median left ventricular ejection fraction (LVEF), myocardial T1 and T2 values between both groups (Table 1).

Table 1.

Characteristics of the general HCHS study population and the final CE-CMR study population.

Parameter, unit	General HCHS population	Final CE-CMR population	Difference in proportion/median (95% CI)	p-value
n	10,000	1064
Females, n	5108 (51.1)	388 (36.5)	14.6 (11.8; 17.4)	0.001*
Age, years	63 [55, 70]	66 [59, 71]	−3.0 (−3.6; −2.2)	0.001*
BSA, m2	1.9 [1.8, 2.1]	2.0 [1.8, 2.1]	−0.1 (−0.0; 0.0)	0.043*
BMI, kg/m2	26.1 [23.5, 29.2]	26.5 [23.9, 29.2]	−0.4 (−0.6; −0.1)	0.005*
GFR, mL/min	86 [75,94]	85 [76, 92]	1.0 (−0.2; 1.4)	0.171
LDL, mg/dL	120 [95, 145]	123 [99, 146]	−3.0 (−5.0; 0.2)	0.067
Diabetes mellitus, n	794 (8.6)	88 (8.6)	0.0 (−1.7; 1.7)	0.993
Systolic blood pressure, mmHg	137 [125, 151]	141 [130, 155]	−4.0 (−5.6; −2.7)	0.001*
Diastolic blood pressure, mmHg	82 [76, 88]	83 [77, 91]	−1.0 (−1.9; −0.3)	0.011*
Arterial hypertension, n	6301 (66.1)	756 (73.4)	−7.3 (−9.9; −4.7)	0.001*
Smoker,** n	6384 (64.2)	713 (67.1)	−2.9 (−5,6; −0.2)	0.033*
LVEDMi, g/m2	62 [54, 72]	64 [55, 73]	−2.0 (−2.5; −0.5)	0.003*
LVEDVi, mL/m2	62 [52, 72]	64 [54, 74]	−2.0 (−2.9; −1.1)	0.001*
LVESVi, mL/m2	19 [14, 24]	19 [15, 25]	−0.8 (−1.2; −0.3)	0.003*
LVSVi, mL/m2	43 [36, 50]	44 [37, 51]	−1.5 (−2.0; −1.0)	0.001*
LVEF, %	70 [64, 75]	69 [64, 75]	0.5 (−0.6; 0.8)	0.762
Myocardial native T1, ms	1182 [1157, 1203]	1179 [1156, 1202]	3.0 (0.1; 4.5)	0.041*
Myocardial T2, ms	40 [38, 42]	40 [38, 42]	0.0 (−0.1; 0.1)	0.901

Values are median (first [Q1] and third [Q3] quartiles) for continuous and n (% of total column number) for categorical data. Confidence intervals for the differences in proportions/medians between the 10,000 participants of the general HCHS study population and the 1064 participants of the final CE-CMR study population were computed using the bootstrap method, using 1000 bootstrap resamples. CMR-derived parameters in the general HCHS study population represent findings in the entire 2589 CMR scans that were performed within the first 10,000 HCHS participants.

BSA body surface area, BMI body mass index, GFR glomerulare filtration rate, LVEDMi left ventricular end-diastolic myocardial mass index, LVEDVi left ventricular end-diastolic volume index, LVESVi left ventricular end-systolic volume index, LVSVi left ventricular stroke volume index, LVEF left ventricular ejection fraction, LDL low-density lipoprotein.

^*p<0.05

^**“Smoker” includes all participants that smoke currently or have ever smoked.

3.2. Unrecognized myocardial scar

UMS was detected in 244 (23%) participants. Individuals with UMS were more often male (75 vs. 60%, p < 0.001), had a higher median body mass index (BMI) (27 vs. 26 kg/m², p = 0.009), had more often hypertension (79 vs. 72%, p = 0.039) and diabetes (12 vs. 8%, p = 0.026) compared to individuals without UMS (Table 2). Furthermore, participants with UMS had a significantly higher median left ventricular (LV) mass index (69 vs. 62 g/m², p < 0.001), LV end-diastolic volume index (67 vs. 64 ml/m², p = 0.011), LV end-systolic volume index (21 vs. 19 ml/m², p = 0.001) and lower median LV ejection fraction (LVEF) (68 vs. 70%, p = 0.001) (Table 2). Septal myocardial T2 values were lower in individuals with UMS (Table 2).

Table 2.

Comparison of individuals with (UMS+) and without (UMS−) unrecognized myocardial scar.

Parameter, unit	UMS−	UMS+	p-value
n	820	244
Females, n	326 (39.8)	62 (25.4)	<0.001*
Age, years	66 [59, 71]	65 [59, 71]	0.731
BSA, m2	1.9 [1.8, 2.1]	2.0 [1.9, 2.1]	0.001*
BMI, kg/m2	26.3 [23.8, 29.0]	27.2 [24.7, 30.3]	0.009*
GFR, mL/min	85 [77, 92]	86 [76, 92]	0.891
LDL, mg/dL	123 [99, 145]	122 [97, 147]	0.491
Diabetes mellitus, n	59 (7.5)	29 (12.4)	0.026*
Systolic blood pressure, mmHg	140 [130, 154]	146 [132, 158]	0.003*
Diastolic blood pressure, mmHg	83 [76, 90]	84 [78, 92]	0.038*
Arterial hypertension, n	570 (71.8)	186 (78.8)	0.039*
Smoker,** n	549 (67.0)	164 (67.5)	0.956
LVEDMi, g/m2	62 [54, 71]	69 [60, 81]	<0.001*
LVEDVi, mL/m2	64 [54, 73]	67 [57, 77]	0.011*
LVESVi, mL/m2	19 [14, 24]	21 [16, 26]	0.001*
LVSVi, mL/m2	44 [37, 51]	45 [38, 51]	0.179
LVEF, %	70 [65, 75]	68 [62, 74]	0.001*
Myocardial native T1, ms	1178 [1155, 1202]	1184 [1163, 1202]	0.162
Myocardial T2, ms	40 [38, 42]	39 [37, 41]	0.005*

Values are median (first [Q1] and third [Q3] quartiles) for continuous data.

+ positive (indicating the presence of unrecognized myocardial scar (UMS)), − negative (absence of UMS), BSA body surface area, BMI body mass index, GFR glomerular filtration rate, LVEDMi left ventricular end-diastolic myocardial mass index, LVEDVi left ventricular end-diastolic volume index, LVESVi left ventricular end-systolic volume index, LVSVi left ventricular stroke volume index, LVEF left ventricular ejection fraction, LDL low-density lipoprotein.

^*p<0.05

^**“Smoker” includes all participants that smoke currently or have ever smoked.

3.3. Patterns and localization of unrecognized myocardial scar

An ischemic scar pattern was found in 25 (10%) and a non-ischemic scar pattern in 217 (89%) participants with UMS. Two participants (1%) with UMS had coincident ischemic and non-ischemic scar patterns. Ischemic UMS predominantly involved mid (XI) and basal inferolateral (V) segments, followed by mid (X) and basal inferior (IV) segments (Fig. 2). The predominantly involved myocardial segment for non-ischemic UMS was the basal inferolateral segment V, followed by the basal inferior segment IV (Fig. 2). Individuals with ischemic UMS had a higher median LV mass index (LVEDMi, 80 vs. 69 g/m², p = 0.004) compared to individuals with non-ischemic UMS (Table 3). Major non-ischemic UMS was detected in n = 109 (50%) participants with n = 34 (16%) “midwall,” n = 57 (26%) “epicardial,” n = 2 (1%) “infiltrative” and n = 16 (7%) “patchy” patterns (Fig. 3). Minor UMS was detected in n = 108 (50%) participants with n = 5 (2%) “near aortic root,” n = 8 (4%) “near mitral anulus,” n = 79 (36%) “basal inferolateral,” n = 10 (5%) “RV insertion,” and n = 6 (3%) “other” patterns (Fig. 3). Participants with major non-ischemic UMS were older (67 vs. 64 years, p = 0.033) and had higher LV mass index (70 vs. 67 g/m², p = 0.024) (Supplementary Table 1).

Fig. 2.

Segmental distribution of unrecognized myocardial scar. Segmental distribution of ischemic (left) and non-ischemic (right) unrecognized myocardial scar (UMS) according to the 16-segment model (22). Values are given in absolute numbers and percentages.

Table 3.

Comparison of individuals with ischemic and non-ischemic unrecognized myocardial scar.

Parameter, unit	Ischemic UMS	Non-ischemic UMS	p-value
n	25	217
Females, n	3 (12.0)	58 (26.7)	0.173
Age, years	67 [60, 71]	65 [59, 71]	0.377
BSA, m2	2.0 [1.8, 2.2]	2.0 [1.9, 2.1]	0.775
BMI, kg/m2	27.8 [24.3, 32.5]	27.2 [24.8, 30.2]	0.499
GFR, mL/min	83 [74, 94]	86 [76, 92]	0.818
LDL, mg/dL	111 [86, 145]	124 [98, 147]	0.557
Diabetes mellitus, n	6 (25.0)	23 (11.1)	0.106
Systolic blood pressure, mmHg	150 [139, 160]	145 [131, 158]	0.194
Diastolic blood pressure, mmHg	85 [80, 93]	83 [78, 92]	0.352
Arterial hypertension, n	21 (87.5)	163 (78.0)	0.392
Smoker,* n	20 (83.3)	143 (66.0)	0.133
LVEDMi, g/m2	80 [71, 87]	69 [59, 79]	0.004**
LVEDVi, mL/m2	74 [54, 92]	66 [57, 75]	0.099
LVESVi, mL/m2	26 [15, 39]	21 [16, 25]	0.076
LVSVi, mL/m2	48 [44, 55]	45 [38, 51]	0.247
LVEF, %	65 [57, 72]	69 [63, 74]	0.070
Myocardial native T1, ms	1193 [1171, 1207]	1183 [1162, 1201]	0.141
Myocardial T2, ms	39 [37, 41]	39 [37, 41]	0.896

Values are median (first [Q1] and third [Q3] quartiles) for continuous data.

BSA body surface area, BMI body mass index, GFR glomerular filtration rate, LVEDMi left ventricular end-diastolic myocardial mass index, LVEDVi left ventricular end-diastolic volume index, LVESVi left ventricular end-systolic volume index, LVSVi left ventricular stroke volume index, LVEF left ventricular ejection fraction, UMS unrecognized myocardial scar, LDL low-density lipoprotein.

^*“Smoker” includes all participants that smoke currently or have ever smoked.

^**p<0.05

Fig. 3.

Patterns of non-ischemic myocardial scar. Patterns of non-ischemic UMS were categorized according to Shanbhag et al. [13] into major (50%, A-D) and minor (50%, E-H and other) LGE patterns: A = midwall; B = epicardial; C = diffuse/infiltrative; D = patchy LGE with LV hypertrophy, E = near aortic root; F = near mitral anulus; G = basal inferolateral; H = RV insertion without LV hypertrophy; other = not consistent with any category. RV: right ventricle, UMS: unrecognized myocardial scar.

3.4. Risk factors for unrecognized myocardial scar

The following findings are based on BIC model selection. Comprehensive findings including the full model and AIC model selection are provided in the data supplement (Supplementary Fig. 1). Fig. 4 shows OR for categorical data and OR per standard deviation (SD) for continuous data to predict the presence of UMS (Fig. 4). A lower LVEF (OR per SD 0.77 [0.65, 0.9], p = 0.002) and a higher LVEDMi (OR per SD 1.54 [1.31, 1.82], p < 0.001) were independently associated with the presence of any UMS in our study population (Fig. 4A and Visual abstract). Ischemic UMS was independently associated with a lower LVEF (OR per SD 0.58 [0.39, 0.86], p = 0.007), a higher LVEDMi (OR per SD 1.74 [1.25, 2.45], p = 0.001), and diabetes (OR 4.91 [1.66, 13.03], p = 0.002), (Fig. 4B and Visual abstract). Non-ischemic UMS was independently associated with higher LVEDMi (OR per SD 1.44 [1.24, 1.69], p < 0.001) (Fig. 4C and Visual abstract).

Fig. 4.

Risk factors of unrecognized myocardial scar. Findings from multivariable logistic regression model (Bayesian information criterion (BIC)). (A) Factors associated with unrecognized myocardial scar (UMS) in general. (B) Factors associated with ischemic and (C) with non-ischemic UMS. Values are given as odds ratios (OR) for categorical data and OR per standard deviation (SD) for continuous data. LVEDMi: left ventricular end-diastolic myocardial mass index, LVEF: left ventricular ejection fraction, CI: confidence interval.

4. Discussion

The major findings of our study can be summarized as follows: Firstly, UMS was detected in about every fourth individual of our study population without known cardiac disease. Secondly, the vast majority of UMS lesions had a non-ischemic pattern, including a “major” pattern in 50% (109/217) and were located in the basal inferolateral LV. Thirdly, the presence of ischemic myocardial scar was independently associated with LVEF, LV mass, and diabetes, whereas non-ischemic myocardial scar was only independently associated with LV mass.

4.1. Ischemic unrecognized myocardial scar

There is robust evidence for the prognostic relevance of ischemic UMS detected by CMR. The proof of clinically silent ischemic UMS by CMR is associated with an increased event rate and mortality in patients with first time acute coronary syndrome [14]. Furthermore, the presence of ischemic UMS is associated with a similar risk for death and/or myocardial infarction compared to clinically recognized myocardial infarction, independent from inducible myocardial ischemia, in patients with suspected coronary artery disease [23]. Of note, post-mortem analyses revealed prior myocardial infarction in more than 70% of patients who suffered sudden cardiac death without a history of coronary artery disease [24]. Interestingly, we found that ischemic UMS predominantly involved mid and basal inferolateral segments (Fig. 2), followed by mid and basal inferior segments, which is typically supplied by the circumflex or the right coronary artery and prone to be missed by standard electrocardiogram [25]. In this context, it is important to note that diagnosing UMS without CMR or nuclear imaging is challenging due to the low sensitivity of conventional diagnostic tools [8], [11], [14] with a subsequent undertreatment of affected individuals [8], [26]. Nevertheless, the prognostic impact of ischemic UMS led to a class I recommendation of recent guidelines of the European Society of Cardiology to perform further testing in patients with diabetes, hypertension, or suspected cardiovascular disease [27]. This recommendation is supported by our finding on the independent association of diabetes with ischemic UMS (Fig. 4B). Our findings are in line with previous findings by Elliot et al., who demonstrated a high prevalence of ischemic UMS in diabetics, which was associated with an eightfold increase in risk for death or myocardial infarction [28]. These findings indicate that screening for ischemic UMS by LGE CMR offers great potential to reveal unrecognized coronary artery disease with immediate prognostic and therapeutic implications for the individual patient. Furthermore, we found LVEF to be independently associated with ischemic UMS (Fig. 4B). This finding most likely reflects that myocardial scarring is the major reason for LV dysfunction [29] and highlights the need for a comprehensive assessment, ideally including LGE CMR, in patients with reduced LVEF, in particular in diabetics.

4.2. Non-ischemic unrecognized myocardial scar

In contrast to the robust data on ischemic UMS, there is currently a paucity of data on non-ischemic UMS. In 2019, Shanbhag et al. published the first population-based study reporting a poor prognosis for individuals without pre-existing heart failure but “major” non-ischemic myocardial scar [13]. Interestingly, the proportion of non-ischemic UMS (58%) was higher than ischemic UMS (42%), similar to the findings in the Multi-Ethnic Study of Atherosclerosis (MESA) (62 vs. 38%) [11], which was even more pronounced in our study (89 vs. 10%, Table 3). Non-ischemic UMS can be missed easily and its detection depends on the LGE sequence, image quality, contrast and spatial resolution, but also on the individual threshold of the observer [7], [13]. In this context, it should be noted that the HCHS is the first population-based study with contrast-enhanced CMR at 3T and a PSIR LGE sequence, which could have affected the prevalence of non-ischemic UMS in our population [30].

The detection and identification of disease-specific non-ischemic myocardial scar by CMR is a key feature with prognostic relevance in several non-ischemic cardiomyopathies [1], [15], [31], [32]. However, there are further “unspecific” non-ischemic myocardial scar patterns with currently unknown pathophysiology and unclear prognostic impact [13], [33], [34]. Therefore, a categorization in “major”, disease-specific and “minor”, unspecific, non-ischemic myocardial scar was recently proposed [13]. Following this categorization, we found 50% (109/217) “major” and 50% (108/217) “minor” non-ischemic UMS (Fig. 3), which differs significantly from the initial description in the ICELAND MI cohort [13]. Clinical and CMR characteristics of “major” and “minor” group were similar, apart from minor differences in age, systolic blood pressure, and LV mass (Supplementary Table 1). It is important to note that the majority of “major” lesions were located “epicardial” in basal inferolateral segments (26%, 57/217) or “midwall” in the septum (16%, 34/217), e.g., consistent with recent myocarditis (Fig. 3). Most “minor” scars were also located in basal inferolateral segments (36%, 79/217), without fulfilling the criteria for major patterns (Fig. 3) [13]. Our study revealed for the first time in a population-based setting that the inferolateral wall is the predilection-site not only for ischemic, but also non-ischemic UMS. The classic explanation for ischemic UMS in the “circumflex territory” is that this region is poorly represented and frequently missed by standard 12 lead ECG [35]. However, the predominance of this area in non-ischemic UMS is less intuitive. The inferolateral/lateral LV myocardium is known for many years as the predilection site for non-ischemic myocardial scar in congenital and acquired disease, such as muscular dystrophy [36] or myocarditis [15], [37]. Although the exact pathophysiology of this observation is unclear, it has been hypothesized that the inferolateral/lateral LV wall is generally less resilient and/or more exposed to mechanical stress with subsequent myocardial fibrosis in several conditions [36]. Our recently published findings in apparently healthy athletes support this hypothesis of a non-specific response to mechanical stress in this area [38]. Our findings in the HCHS population indicate that the inferolateral wall is prone to myocardial scarring even in the absence of apparent cardiac disease. In particular, the independent association of non-ischemic UMS with LV mass could indicate an interaction between hypertension and scarring as an unfavorable response to mechanical stress in this region, again supporting the hypothesis of the inferolateral/lateral wall as a “locus minoris resistentiae” of the LV myocardium. Of note, “minor” scars were subtle and could easily be missed in subjects with poor image quality or lower field strength [30]. At least, our data demonstrate that “minor” non-ischemic UMS are a common finding in individuals at increased cardiovascular risk at 3T, but it remains unclear whether “minor” non-ischemic UMS holds any prognostic impact [13].

Interestingly, we found that only an increased LV mass index was independently associated with non-ischemic UMS (Fig. 4C and Visual abstract). This finding support the hypothesis of a link between hypertension and non-ischemic UMS. Hypertension is by far the most frequent cardiovascular risk factor in the general population of this age group and in our study population (Table 1) [17], [39]. Therefore, our findings could indicate an inadequate myocardial reaction to hypertension, in terms of pre-clinical hypertensive heart disease, despite a normal median LV mass in patients with non-ischemic UMS in our study population (Table 3) [39]. Non-ischemic UMS could be interpreted as a biomarker for an unfavorable myocardial response to hypertension, prior to conventionally detectable hypertensive heart disease [40]. However, there are also other potential explanations for non-ischemic UMS, in particular clinically unrecognized myocarditis, since non-ischemic UMS predominantly affected the basal inferolateral LV myocardium, which is also the typical localization of post-inflammatory myocardial scar (Fig. 3) [1], [15]. Independent from the potential etiologies, our data show that the prevalence of non-ischemic UMS in apparently healthy individuals with hypertension and other cardiovascular risk factors is high (Tables 2 and 3). In particular, our findings show that an increased LV mass is linked to an increased risk for the presence of non-ischemic myocardial scar (Fig. 4C). Therefore, the detection of non-ischemic myocardial scar by CMR needs to be carefully interpreted in the context of clinical setting, especially in individuals with hypertension and LV hypertrophy. In particular, our findings show that there is a high rate of pre-existing non-ischemic LGE in the inferolateral/lateral wall and that its detection should not be equated with myocarditis without a suggestive clinical setting and/or other CMR criteria. It is important to note that there are currently no evidence-based recommendations for the clinical handling of non-ischemic UMS in the absence of apparent cardiac disease. There could be a potential role for non-ischemic UMS as the substrate for arrhythmia and heart failure, but clear prognostic data with immediate clinical consequences still remain to be determined [12], [13], [14], [34].

4.3. Unrecognized myocardial scar in the HCHS population: novelty in comparison to current knowledge

Verifying established data in different populations/settings is a crucial part of evidence-based medicine. However, our work not only confirms currently accepted findings, but also adds novel aspects to current perception of UMS. Firstly, our study population constitutes by far the largest CE-CMR population in “continental” Europe. Currently available data on UMS by CMR are based on the MESA, ICELAND MI, and Prospective Investigation of the Vasculature in Uppsala Seniors populations [8], [9], [11], [13]. However, it is important to note that the distribution of cardiovascular risk factors varied significantly between these populations. In particular, the ICELAND MI population included about 40% participants with diabetes, which is much higher compared to the 8.6% (Table 1) in our population and the expected 8–10% prevalence in Europe [13], [41]. Although not fully representative for the general population in all facets, our study population reflects typical characteristics of individuals examined by CMR for clinical indication, as shown in the Society for Cardiovascular Magnetic Resonance (SCMR) registry (Table 1) [42]. Furthermore, we excluded all participants with a history of coronary heart disease and/or myocarditis, whereas only participants with heart failure were excluded from the ICELAND MI population [13]. Therefore, our study population differs in several aspects from recently published cohorts, which is of crucial importance in the context of UMS. Secondly, our study is the first population-based CMR study using a 3T scanner, a state-of-the-art PSIR LGE sequence and T1/T2 mapping in the context of UMS. Interestingly, we did not find a significant difference between participants with and without UMS in native myocardial T1, but in T2 (Table 2). This can be explained by earlier findings by our group on this topic in the HCHS population. Briefly, we found that sex was the major, independent factor for myocardial T2 with higher myocardial T2 values in females [20]. Therefore, we assume that the lower median myocardial T2 values in the group with UMS were related to the lower proportion of females in this group compared to the group without UMS (24.4 vs. 39.8%, p < 0.0001; Table 2). The third and clinically most important novel aspect of our study was the notion that the inferolateral wall is the predilection-site for ischemic, but also non-ischemic UMS (Fig. 2), which supports earlier speculations on a lower resilience and/or higher exposition to mechanical stress with subsequent myocardial fibrosis of this region [36], [38]. The independent association of non-ischemic UMS with LV mass could indicate a pathologic response to hypertension and scarring in this region, again highlighting that the inferolateral/lateral wall can be perceived as a non-specific “locus minoris resistentiae” of the LV myocardium [36]. In addition, we found a higher rate of “major” non-ischemic pattern UMS, which was found in 109 (10%) of our participants, accounting for about 50% of all non-ischemic UMS, which was significantly higher compared to the ICELAND MI population [13].

5. Limitations

First, our cross-sectional data do not provide prognostic data and longitudinal studies are needed to address prognostic impact and therapeutic consequences of non-ischemic UMS in the general population. Furthermore, the optional administration of contrast media inherently introduced a selection bias, since contrast media administration was disproportionally accepted by male participants at increased cardiovascular risk. On the one hand, our study population was not representative for the general population and had an increased cardiovascular risk profile (Table 1). On the other hand, our study population seems to be representative for a typical population at increased cardiovascular risk undergoing CMR in clinical routine, such as suspected coronary artery disease [23]. Furthermore, the use of dark-blood LGE could have revealed a higher rate of small, subendocardial, ischemic myocardial scar, which might have been missed on conventional LGE images [43]. However, none of the currently available population-based, large-scale studies has used a dark-blood LGE sequence [8], [9], [11], [13]. Therefore, the LGE sequence was unlikely the reason for the relatively low rate of ischemic UMS in our population. Nevertheless, a dark-blood LGE sequence should be included in future protocols to adequately assess the prevalence of small, subendocardial ischemic UMS. Furthermore, the use of a balanced steady state free precession instead of a gradient echo LGE sequence variant could have altered the performance of LGE imaging in our cohort, in particular at 3T. In addition, the use of a 2-mm gap between the LGE slices constitutes a further potential limitation since there was no true full coverage of the LV myocardium.

6. Conclusions

Unrecognized myocardial scar, in particular with a non-ischemic pattern, is frequent in individuals without known cardiac disease and predominantly involves the basal inferolateral LV myocardium. Presence of unrecognized myocardial scar is independently associated with a lower LVEF, a higher LV mass, and a history of diabetes.

Funding

The HCHS is generally funded by the euCanSHare grant agreement [Grant Number 825903-euCanSHare H2020]; the Foundation Leducq [Grant Number 16 CVD 03], and the Innovative Medicine Initiative [Grant Number 116074]. The HCHS is additionally supported by Deutsche Gesetzliche Unfallversicherung (DGUV); Deutsches Krebsforschungszentrum (DKFZ); Deutsches Zentrum für Herz-Kreislauf-Forschung (DZHK); Deutsche Stiftung für Herzforschung; Seefried Stiftung; Bayer; Amgen; Novartis; Schiller; Siemens; Topcon, and Unilever. The study is further supported by donations from the “Förderverein zur Förderung der HCHS e.V.,” and TePe (2014). Sponsor funding has in no way influenced the content or management of this study.

Author contributions

Katharina A. Riedl : Writing – review and editing, Investigation. Charlotte Jahnke: Writing – review and editing, Investigation. Celeste Chevalier: Writing – review and editing, Investigation. Kai Muellerleile: Writing – review and editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Data curation, Conceptualization. Francisco Ojeda: Writing – review and editing, Formal analysis, Data curation, Conceptualization. Gunnar K. Lund: Writing – review and editing, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Data curation, Conceptualization. Alena Haack: Writing – review and editing, Formal analysis, Data curation, Conceptualization. Enver Tahir: Writing – review and editing, Validation, Supervision, Software, Methodology, Investigation, Data curation, Conceptualization. Andreas Ziegler: Writing – review and editing, Formal analysis, Data curation. Gerhard Adam: Writing – review and editing, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Data curation, Conceptualization. Eleonora di Carluccio: Writing – review and editing, Formal analysis, Data curation. Stefan Blankenberg: Writing – review and editing, Writing – original draft, Visualization, Validation, Supervision, Software, Resources, Project administration, Methodology, Investigation, Funding acquisition, Data curation, Conceptualization. Jan N. Schneider: Writing – review and editing, Visualization, Validation, Supervision, Software, Project administration, Data curation. Paulus Kirchhof: Writing – review and editing, Validation, Supervision, Resources, Project administration, Methodology, Investigation, Funding acquisition, Conceptualization. Ersin Cavus: Writing – review and editing, Writing – original draft, Visualization, Validation, Supervision, Software, Project administration, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Raphael Twerenbold: Writing – review and editing, Supervision, Project administration, Methodology, Investigation, Funding acquisition. Ulf K. Radunski: Writing – review and editing, Project administration, Methodology, Investigation, Data curation, Conceptualization.

Ethics approval and consent

The local ethics committee of Hamburg (https://www.aerztekammer-hamburg.org/ethikkommission.html) approved the HCHS (number PV5131) and all participants of HCHS gave their written informed consent.

Consent for publication

Not applicable.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

Supplementary material

.

Supplementary material

.