Presence of myocardial fibrosis on cardiac MRI and its impact on short-term outcomes in an Asian cohort of heart failure with preserved ejection fraction (HFpEF)

Abstract

Backgrounds and objectives

This study aims to evaluate the presence and prognostic significance of myocardial fibrosis in subjects with and without HFpEF, using Cardiac MRI.

Methods

This was an ambidirectional observational study [mean follow-up: 25 months], selecting HFpEF patients and comparing them with an age and sex-matched control arm without HF. Late gadolinium-enhanced (LGE) imaging and T1 mapping were used to assess fibrosis.

Results

LGE (27 % vs 0 %, p = 0.005) and mid-segment (1062 ± 64 ms vs 1020±8 ms, p < 0.001) or total average (1057 ± 70 ms vs 1020±4 ms, p = 0.006) T1 values were significantly higher in HFpEF group (N = 30) compared to the control group. However, there was no significant difference in T2 values (p = 0.657).

53.3 % of our patients with HFpEF had a clinical event (HF hospitalization, stroke or all-cause mortality) during the follow-up. It was found that average mid-segment (1099 ± 67 ms vs 1019 ± 17 ms, p<0.001) and total average (1101 ± 70 ms vs 1007 ± 15 ms, p<0.001) T1 values were significantly higher in the event group compared to the no-event group. There were no differences in LGE prevalence between event and no-event groups. However, LGE negative cases had significantly increased T1 values in all segments compared to the healthy control group (p<0.001).

Conclusions

Higher T1 values but not T2 values , were associated with the HFpEF group compared to the age and sex-matched control group. Focal fibrosis, as evident by LGE, was significantly more in HFpEF. Among HFpEF patients, high T1 myocardial values were associated with a higher rate of all-cause death, stroke, and HF hospitalization in short-term follow-up.

Graphical abstract

1. Introduction

Heart failure with preserved ejection fraction (HFpEF) is a global public health problem. The first significant documentation of HFpEF in Asia suggested that Asian patients with HFpEF were relatively young (with over a third under the age of 65) and lean compared to those from Western populations. Nevertheless, they carried a high co-morbidity burden (70 % of patients had ≥ two co-morbidities).¹ There were striking regional differences in types of co-morbidities, cardiac remodeling, and outcomes of HFpEF across Asia. South Asians were the youngest and most often obese but had the lowest AF prevalence despite the most concentric LV hypertrophy.

Cardiac magnetic resonance (CMR) has become a valuable non-invasive tool in assessing HF patients' diagnostic work-up and prognosis. CMR has established superiority in assessing left ventricular (LV) volumes and function and analyzing wall motion abnormalities and myocardial tissue characteristics.² Myocardial fibrosis has been implicated in the pathophysiology of HFpEF in the recent past.³ In HFpEF, the pattern of fibrosis is diffuse and can not be readily detected by LGE alone.⁴ CMR T1 and its surrogate markers have emerged as essential tools for identifying diffuse myocardial fibrosis. Native T1 can detect early changes associated with edema, inflammation, and fibrosis, even before LGE is visible. Furthermore, it has been correlated with histology-proven myocardial fibrosis.⁵ The classical T1 image reflects the changes in the myocardium involving the myocytes and the interstitium. Myocardial interstitial expansion by fibrosis, fluid, or other protein deposits is reflected in the ECV, calculated by T1 pre- and post-administration of gadolinium and hematocrit.² T2 mapping is sensitive to myocardial edema. Very few studies have evaluated the HFpEF cohort's outcome regarding T1 characteristics involving the Western population.5, 6, 7, 8 Several international heart failure guidelines have incorporated CMR imaging in HF; however, data from the Asian population is almost non-existent, and utilization of CMR imaging in HFpEF remains suboptimal in this part of the world. Robust population data is essential to establish CMR imaging in detecting and prognosticating HFpEF in the Indian subcontinent. This study aims to evaluate the presence and prognostic significance of myocardial fibrosis (diffuse by T1 and focal by LGE) in subjects with and without HFpEF, using Cardiac MRI in this part of the world.

2. Materials and methods

We conducted the study titled Clinical and Cardiac Magnetic Resonance (CMR) Features in Heart Failure with preserved ejection fraction (HFpEF) to establish the clinical profile and CMRI features in Indian patients presenting with HFpEF. This was a hospital-based, Ambi-directional, record review study, spread over January 2018 to December 2022 [mean follow-up: 25months, IQR 11–44months] designed with the following objectives:

• •To document patient characteristics, including demographic profile, presenting symptoms, and clinical signs in patients of HFpEF diagnosed with conventional echocardiography
• •To study anatomical and functional characteristics of cardiac chambers by CMR in these patients
• •To evaluate the presence and/or prognostic significance of myocardial fibrosis in subjects with and without HFpEF, using Cardiac MRI

We hypothesized that patients with HFpEF have subclinical or overt myocardial fibrosis manifested by increased T1 or LGE, respectively. We selected patients with HFpEF based on existing ASE/ESC guidelines for diagnosing HFpEF. We recorded their details, including clinical parameters, biochemical details, baseline drugs prescribed, and chest X-ray findings (Fig. 1).

Fig. 1.

Patient selection flow diagram LVEF – left ventricular ejection fraction; NP – natriuretic peptides; LVMI – left ventricular mass index; RWT – relative wall thickness; LAVI – left atrial volume index; SR – sinus rhythm; AF – atrial fibrillation; NT Pro-BNP – N-terminal pro-brain natriuretic peptides; PASP – pulmonary artery systolic pressure; TR – tricuspid regurgitation; HHF – hospitalization for heart failure.

Patients with comorbidities that compromise the diagnosis of HFpEF, like congenital heart disease, pericardial diseases, high output HF, significant CAD on invasive coronary angiogram (revascularized or indication for revascularisation), hypertrophic cardiomyopathy, cardiac tumor or intra-cardiac mass were all excluded from the study. A control arm was included with 30 consecutive patients without HF who underwent CMRI for other diagnostic reasons and had a normal CMRI report.

Echocardiography was performed, and the following parameters were recorded:

• •Ejection fraction (EF) was calculated by modified Simpson's biplane method.
• •Diastolic function was standard ASE/ESC guidelines.⁹
• •Left atrial volume was calculated by the formula derived by (D1 × D2 × D3) × (0.523) formula where D2 is measured from the mitral annular plane to the back wall in apical 4-chamber view. D1 is the orthogonal short-axis dimension to D2. D3 is measured from the blood–tissue interface of the anterior and posterior walls in a parasternal long-axis view.¹⁰
• •Left ventricular mass was calculated by both the ASE and Th formulas.

2D Strain: Longitudinal strain by speckle tracking echocardiography was obtained from three apical views (Apical 4-Chamber view, Apical 2-Chamber View, Apical 3-Chamber view). Global longitudinal strain (GLS) is defined as an average of peak longitudinal strain from an 18 left ventricle (LV) Segments model. Echocardiography and strain echocardiography were done in the Philips EPIQ 7c model (Philips Medical Systems, Best, The Netherlands) by a single operator with consensus from another operator with >3 years of experience.

CMRI findings of both the cases with HFpEF (n = 30) and the control population (n = 30) were recorded.

The CMR examinations were done on 1.5 T S (MAGNETOM Avanto, Siemens Healthcare, Erlangen, Germany) equipped with a 36-element dedicated cardiac array. The standard clinical CMR protocol consisted of scout images followed by a functional assessment of the left ventricle using cine steady-state free precession techniques and post-contrast LGE images using phase-sensitive inversion recovery (PSIR) bSSFP sequence. Native T1 maps (Fig. 2A) using the shortened modified Look-Locker inversion recovery (shMOLLI) technique with collection of 6 regions of interest in each basal, mid, and apical segment (18-segment model), and inversion recovery times are separated by only one R–R interval, were acquired before contrast administration. In 1.5T MRI, a standard Repetition Time (TR), Echo Time (TE), Flip Angle, Pixel Spacing, Slice Thickness and Field of View (FOV) were used with ECG-gated imaging for end-daistolic phase of cardiac cycle to minimize cardiac motion as far as possible. The total acquisition time was around 40 min. All CMR images were stored in a picture archiving and communication system (PACS). Absolute and indexed LV volumes, myocardial mass, LV and RV ejection fractions, native T1, T2 (Fig. 2B) and extent of myocardial LGE (Fig. 2C) were measured using cvi42 Version 5.13.7 (Circle Cardiovascular Imaging, Calgary, Alberta, Canada) by a single experienced reader (3 years of CMR experience). LGE images were obtained 10 min after administering intravenous Gadolinium-based contrast medium (Gadotrast, Gadoterate Meglumine, Unique Pharmaceuticals, India) intravenously at 0.1 mmol/kg body weight. The breath-hold segmented ECG gated PSIR bSSFP sequence was performed in a similar orientation to the cine images. The inversion time was adjusted to completely null normal myocardium (typically 250–400 ms). The percentage extent of LGE in the LV as a percent of total LV mass was calculated using the mean + 5 SD method. A user-defined freehand region of interest was drawn within normal nulled remote myocardium in the short axis stack of LGE PSIR images after contouring the epicardial and endocardial borders to obtain software-generated percentage extent of LGE. All areas that were identified as enhancement by the software were cross-verified by the reader to ensure the exclusion of inversion time artefact or contamination by blood pool or pericardial fat. These were manually excluded by contour adjustment or the exclusion tool available in the software. The presence and absence of LGE in each segment were also recorded.

Fig. 2.

Representative image showing T1 (A), T2 (B) colour mapping, and LGE quantification (C) in cvi42 software – yellow highlighted areas represent LGE. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

The Institutional Ethics Committee approval was taken to conduct the study, and all participants recorded their consent for participation in the study. The information collected was recorded in a pre-approved Case Record Form, and the data was transferred into a Master Chart for deriving statistical information.

2.1. Statistical analysis

Statistical analyses were done using STATA 14 (Stata Corporation, Texas, USA) software. All tests were two-sided tests that were considered statistically significant when P < 0.05. Continuous variables were expressed as mean ± 2 SD if distributed normally or as medians (25th and 75th percentiles) if not distributed normally. Categorical variables were mentioned as counts and percentages. Comparisons between groups were made by ANOVA, chi-square test, paired/unpaired sample t-test wherever appropriate for continuous variables with normal distribution, Wilcoxon signed-rank test for continuous variables with nonnormal distribution, and Fisher exact test for categorical variables. Logistic regression was done to predict abnormal diffuse HFpEF (above or below 95 % confidence intervals in controls). After univariate analysis of the two groups, parameters with a P < 0.20 were proposed for inclusion in the multiple logistic regression analysis with a backward selection procedure. The correlation of variables was expressed by the Pearson correlation coefficient.

3. Results

A higher age, the female gender and obesity were common in cases, as was the prevalence of diabetes, hypertension, COPD, CAD, and dyslipidemia (Table 1). The mean H2FPEF score was 5 in our case population, while the mean HFA-PEFF score was 5.2 (Supplementary Fig. 1).

Table 1.

Demographic profile, co-morbidities, and personal history of the study population.

	Cases (N = 30), % [Minimum-Maximum]	Controls (N = 30), % [Minimum-Maximum]	p
Age (mean ± SD) in Years	63 ± 8 [49–74]	61 ± 6 [49–76]	0.123
Female (%)	26 (87)	24 (80)	0.731
Height (cm) (mean ± SD)	150 ± 7 [143–166]	153 ± 8 [145–172]	0.105
BMI (kg/m2) (mean ± SD)	29.5 ± 4.8 [19.8–38.7]	25.8 ± 2.4 [21.0–29.9]	<0.001
BSA (m2) (mean ± SD)	1.62 ± 0.18 [1.28–1.93]	1.58 ± 0.12 [1.44–1.80]	0.324
Diabetes mellitus	12 (40)	3 (10)	<0.001
Hypertension	26 (87)	4 (13)	<0.001
Hypothyroid	2 (7)	0	0.492
COPD	8 (27)	0	0.005
CAD	6 (20)	0	0.024
Dyslipidemia	14 (47)	0	<0.001
OSA	4 (13)	0	0.112
Family history of CAD	2 (7)	0	0.492
History of COVID-19 infection	4 (13)	0	0.112
WHO guidelines
No obesity	4 (13)	13 (43)	<0.001
Overweight	16 (53)	17 (57)
Obese	10 (33)	0
Hb g/dL (mean ± SD)	13.1 ± 1.6 [10.4–17.0]	13.2 ± 0.5 [12.0–14.0]	0.803
Creatinine mg/dL (mean ± SD)	1.01 ± 0.26 [0.44–1.5]	0.89 ± 0.12 [0.55–1.33]	0.056
TSH mIU/L (mean ± SD)	2.9 ± 1.0 [1.2–4.7]	3.1 ± 0.9 [1.8–4.9]	0.518
ACEi/ARB	26 (87)	4 (13)	<0.001
Loop Diuretics	26 (87)	0 (0)	<0.001
MRA	16 (53)	0 (0)	<0.001
BB	14 (47)	0 (0)	<0.001
CCB	8 (27)	0	<0.001
SGLT2i	8 (27)	1 (3)	<0.001
Hydralazine	1 (3)	0 (0)	0.870

SD – standard deviation, BMI – body-mass index; BSA – body surface area; COPD – chronic obstructive pulmonary disease; CAD – coronary artery disease (<70 % obstructive CAD); OSA – obstructive sleep apnea; WHO – world health organization; Hb – hemoglobin; TSH – thyroid stimulating hormone; ACEi – angiotensin converting enzyme inhibitor; ARB – angiotensin receptor blocker; MRA – mineralocorticoid receptor antagonist; BB – beta blockers; CCB – calcium channel blocker; SGLT2i – sodium-glucose co-transporter 2 inhibitor.

We elucidated the baseline drug history of patients with HFpEF. Among the 30 patients included to represent the HFpEF arm, 87 % were on ACE inhibitors, and 87 % were on diuretics, with all of the latter being on loop diuretics. 53 % of the cases were on mineralocorticoid receptor antagonists, while only 27 % were on SGLT2i (Table 1).

ECG findings were recorded in all cases and controls. Sinus rhythm was observed in 80 % of the cases. The rest were in atrial fibrillation, while all members of the control population were in sinus rhythm. Among the cases, 27 % had electrocardiographic features of left atrial enlargement, and 13 % had left ventricular hypertrophy by standard electrocardiographic criteria for left ventricular hypertrophy (Table 2).

Table 2.

Electrocardiogram, Chest X-ray, and Echocardiographic findings in the study population.

	Cases (N = 30) [Minimum-Maximum]	Controls (N = 30) [Minimum-Maximum]	p
Sinus rhythm, %	24 (80)	30 (100)	0.024
LAE on ECG (n = 24), %	8 (27)	0	<0.001
PR duration (n = 24) mS	158 ± 23 [120–200]	151 ± 13 [120–180]	0.199
QRS Axis, degree	27 ± 37 [(−)70 – (+) 70]	44 ± 21 [0–80]	0.026
QRS duration, mS	88 ± 31 [65–160]	70 ± 4 [65–80]	0.002
QTc, mS	404 ± 22 [354–438]	393 ± 16 [362–422]	0.031
LVH, %	4 (13)	0	0.112
ST T changes, %	16 (53)	0	<0.001
Cardiac enlargement	22 (73)	0	<0.001
CTR, %	57 ± 6 [48–66]	47 ± 2 [42–50]	<0.001
LAE on CXR	18 (60)	0	<0.001
Carinal angle, degree	81 ± 9 [69–98]	68 ± 2 [65–71]	<0.001
LVIDD, mm	45.5 ± 3.6 [37–50]	42.4 ± 2.9 [36–48]	<0.001
LVIDS, mm	28.4 ± 3.2 [22–33]	25.3 ± 2.9 [20–30]	<0.001
IVS D, mm	12.0 ± 2.2 [10–18]	8.7 ± 0.9 [8–11]	<0.001
IVS S, mm	14.2 ± 2.1 [11–20]	10.9 ± 0.9 [9–13]	<0.001
PW D, mm	11.1 ± 1.9 [8–15]	8.4 ± 0.5 [7.5–9.0]	<0.001
PW S, mm	13.5 ± 2.0 [11–17]	10.7 ± 0.7 [9–12]	<0.001
RVIDD, mm	23.5 ± 3.5 [20–33]	19.8 ± 2.4 [16–25]	<0.001
EF, %	66 ± 7 [53–77]	68 ± 7 [57–80]	0.246
FC, %	38 ± 6 [27–48]	40 ± 6 [29–50]	0.073
LV mass (Th), gm	172 ± 28 [130–224]	110 ± 16 [79–143]	<0.001
LV mass index (Th), gm/m2	108 ± 23 [82–160]	70 ± 12 [52–97]	<0.001
RWT	0.49 ± 0.12 [0.34–0.81]	0.40 ± 0.03 [0.35–0.47]	<0.001
LAVI, mL/m2	31.5 ± 9.6 [17.7–49.7]	18.1 ± 3.6 [11.3–24.7]	<0.001
Hypertrophy (%)
Concentric Hypertrophy	16 (53)	0	<0.001
Concentric Remodeling	2 (7)	4 (13)
Eccentric Hypertrophy	6 (20)	0
Normal	6 (20)	26 (87)
MV E velocity, m/s	1.07 ± 0.33 [0.60–1.70]	0.90 ± 0.16 [0.62–1.10]	0.010
MV A velocity, m/s	0.90 ± 0.24 [0.60–1.30]	0.71 ± 0.12 [0.50–0.90]	<0.001
EDT (mS)	154 ± 32 [106–230]	207 ± 14 [180–240]	<0.001
Lateral E′, cm/s	7.6 ± 2.3 [3.3–10.0]	14.1 ± 2.3 [10.0–20.0]	<0.001
Medial E′, cm/s	5.8 ± 1.3 [3.0–7.6]	10.4 ± 1.6 [8.0–14.0]	<0.001
Average E/E’	17.5 ± 7.2 [7.6–38.2]	7.7 ± 1.7 [4.9–11.5]	<0.001
TR Gradient, mmHg	35 ± 8 [17–54]	15 ± 4 [10–21]	<0.001
TAPSE, mm	20 ± 2 [17–24]	21 ± 1.8 [18–24]	0.038

LAE – left atrial enlargement; ECG – electrocardiogram; LVH – left ventricular hypertrophy, LVIDD – left ventricular internal diameter (diastole); LVIDS - left ventricular internal diameter (systole); IVSD – interventricular septum (diastole); IVSS – interventricular septum (systole); PWD – posterior wall (diastole); PWS – posterior wall (systole); RVID – right ventricular internal diameter (diastole); EF – ejection fraction; FC – fractional change; LV – left ventricle; Th – Teicholz method; RWT – relative wall thickness; LAVI – left atrial volume index; EDT – E deceleration time; TR – tricuspid regurgitation; TAPSE – tricuspid annular plane systolic excursion.

Chest X-ray revealed that cardiac enlargement was present in 73 %, which was significantly more than that in the controls (47 %, p < 0.001) (Table 2).

As compared to none in the control group, 53 % of the cases had concentric left ventricular hypertrophy as per ASE guidelines (p <0.001). Mean inter-ventricular septal thickness in diastole (mean ± SD; 12.0 ± 2.2) and systole (mean ± SD; 14.2 ± 2.1) were significantly higher in the case compared to the control group in diastole and systole. The left ventricular mass index (LVMI) was significantly higher in cases both by Teichholz (mean ± SD; 108 ± 23 vs. 70 ± 12) and ASE formula (mean ± SD; 120 ± 29 vs. 72 ± 13). While the left ventricular ejection fraction (LVEF) was similar in both the groups, the Left atrial volume index was significantly higher in cases (mean ± SD; 31.5 ± 9.6 vs 18.1 ± 3.6, p < 0.001) (Table 2).

Coming to the CMR findings, LVIDd and LVIDs were comparable among cases and controls. However, a higher IVS thickness in diastole (p = 0.004) and systole (p < 0.001) and LVPW thickness in diastole (p < 0.001) and systole (p < 0.001) were significant as baseline CMR data. The LV volume Index was statistically higher among cases than controls (p < 0.001). LVMI was significantly higher in cases as compared to controls (p = 0.005), as were the left ventricular end-diastolic volume index (LVEDVi) (p < 0.001) and Left ventricular end-systolic volume index (LVESVi) (p = 0.016). Right ventricular end-diastolic volume index was lower in the case as compared to the control group (p < 0.001), along with RV stroke volume and RV ejection fraction (Both p < 0.001) (Supplementary Fig. 2).

Coming to the T1 CMRI findings, the average mid-segment (mean ± SD; 1062 ± 64 vs. 1020 ± 8, p < 0.001) and apical segment (mean ± SD; 1065 ± 86 vs 1020 ± 4, p = 0.007) T1 values were significantly higher in case group compared to control as was the total average T1 value in all segments (mean ± SD; 1057 ± 70 vs 1020 ± 4, p = 0.006) (Fig. 3). However, there were no significant differences in average T2 values (mean ± SD; 51 ± 4 vs 51 ± 1, p = 0.66) (Supplementary Fig. 3).

Fig. 3.

T1 Cardiac magnetic resonance imaging (CMRI) findings (mS): Case vs. Control.

Among the 30 patients with HFpEF, we further analyzed our results into those who had a cardiac event during the period of follow-up (Event group, n = 16) and those who had an uneventful follow-up during this study (No event group, n = 14). Out of the 14 patients (47 %) who had events in the form of index heart failure, 4 patients (14 % of total cases) had recurrent events (7 % recurrent hospitalization for heart failure, 7 % death) within a median follow-up of 44 months and another 2 (7 %) had follow up event of Ischemic stroke. Cumulative events, including recurrent events, were 67 % (Fig. 4).

Fig. 4.

Cumulative events among cases of HFpEF HHF – hospitalization for heart failure.

Demographically, a higher age was commoner in the event group. Obesity was more prevalent in the event group, while other metabolic abnormalities were equally distributed among the event and no–event subpopulations. The NT-Pro-BNP at presentation was significantly higher in the event group compared to the no-event group (p = 0.047) (Table 3).

Table 3.

Demographic profile, co-morbidities, and personal history of HFpEF: Events versus no events.

	Events (N = 16)	No events (N = 14)	p
Age (mean ± SD)	66 ± 6	60 ± 8	0.026
Female	14 (88 %)	12 (86 %)	1.000
Height (cm) (mean ± SD)	149 ± 7	152 ± 7	0.238
BMI (kg/m2) (mean ± SD)	29.6 ± 4.3	29.3 ± 5.43	0.859
BSA (m2) (mean ± SD)	1.60 ± 0.17	1.64 ± 0.19	0.511
DM	14 (88)	12 (86)	1.000
HTN	6 (38)	6 (43)	1.000
Hypothyroid	0	2 (14)	0.209
COPD	4 (25)	4 (29)	1.000
CAD	4 (25)	2 (14)	0.657
Dyslipidemia	8 (50)	8 (57)	0.730
Obesity	16 (100)	10 (71)	0.024
OSA	4 (25)	0	0.103
F/h CAD	0	2 (14)	0.209
H/O COVID	2 (13)	2 (14)	1.000
WHO guidelines
No obesity	0	2 (14)
Obese	16 (100)	14 (86)
Hb g/dL (mean ± SD)	13.1 ± 1.9	13.1 ± 1.2	0.974
Creatinine mg/dL (mean ± SD)	1.00 ± 0.34	1.01 ± 0.15	0.876
TSH mIU/L (mean ± SD)	2.49 ± 0.69	3.39 ± 1.05	0.009
Pro-BNP pg/mL (mean ± SD)	5550 ± 9160	455 ± 375	0.047
SGLT2i (%)	6 (38)	2 (14)	0.226
Loop Diuretics (%)	14 (88)	10 (71)	0.378
Frusemide equivalent, mg/d	43 ± 21 (n = 14)	20 ± 12 (n = 10)	0.005
MRA (%)	8 (50)	8 (57)	0.730
MRA dose, mg/day	31 ± 4 (n = 8)	19 ± 2 (n = 8)	0.019

SD – standard deviation, BMI – body-mass index; BSA – body surface area; COPD – chronic obstructive pulmonary disease; CAD – coronary artery disease (<50 % obstructive CAD); OSA – obstructive sleep apnea; WHO – world health organization; Hb – hemoglobin; TSH – thyroid stimulating hormone; ACEi – angiotensin converting enzyme inhibitor; ARB – angiotensin receptor blocker; MRA – mineralocorticoid receptor antagonist; BB – beta blockers; CCB – calcium channel blocker; SGLT2i – sodium-glucose co-transporter 2 inhibitor.

There was no difference between the event and no-event groups regarding rhythm, left atrial enlargement in electrocardiogram, QRS axis, QRS duration, and ST-T changes. PR interval was relatively longer but within the standard limit in the event group (p = 0.011). Cardiac enlargement was present in 88 % (vs. 57 %, p = 0.101) of the event group, and Left atrial enlargement (Carinal angle >75°) was found in 88 % of the event group compared to 29 % in the no-event group (p = 0.002). There were no significant differences noted between the event and no-event groups in terms of echocardiographic evidence of left ventricular hypertrophy (p = 0.46), septal thickness (p = 0.101), and left ventricular mass index (p = 0.311). There were no differences between the event and no-event groups regarding ejection fraction, LA volume index, or left ventricular diastolic dysfunction grade. However, a statistical difference was evident in 63 % of the patients in the event group having mild pericardial effusion compared to none in the no-event group (p <0.001) (Supplementary Table 1). No difference in GLS value was also seen between the event and no-event group (Supplementary Fig. 4).

8 patients (50 %) in event group and 2 patients (14 %) in no-event group underwent cardiac catheterization study with no difference in pulmonary artery mean pressure (24.5 ± 1.1 vs. 26 ± 0.0 mmHg, p = 0.539), pulmonary capillary wedge pressure (17.8 ± 1.7 vs 12 ± 0.0 mmHg, p = 0.153), left ventricular end-diastolic pressure (17 ± 5.8 vs 10 ± 0.0 mmHg, p = 0.142) or pulmonary vascular resistance (2.5 ± 0.9 vs 3.3 ± 0.0 WU, p = 0.256) (Supplementary Table 2).

MRI-derived left ventricular internal diameters were similar in both event and no-event groups. However, diastolic interventricular septal thickness (mean ± SD, 10.9 ± 2.7 vs. 8.6 ± 2.0 mm, p = 0.016) and posterior wall thickness (mean ± SD, 10.1 ± 2.5 vs. 7.1 ± 1.1 mm, p < 0.001) were significantly higher in the event group. Both groups had a similar right ventricular internal diameter and LA volume index.

The MRI-derived left ventricular mass index was similar in both groups, as were left ventricular end-diastolic, systolic volumes, and stroke volume index. LVEF was lower in the event group compared to the no-event group (mean ± SD, 63 ± 4 vs 68 ± 7 %, p = 0.023). Right ventricular stroke volume index (mean ± SD, 32 ± 8 vs. 40 ± 7 mm, p = 0.010) and ejection fraction (mean ± SD, 53 ± 8 vs. 60 ± 10 %, p = 0.034) were lower in the event group compared to no-event group. There was no difference in left or right ventricular cardiac output by CMRI (Supplementary Fig. 5).

8 subjects (27 %) in the case group had late gadolinium enhancement (LGE) compared to none in the control group (p = 0.005). There was no difference in LGE among the event and no-event groups. There was no sub-epicardial LGE, while sub-endocardial and mid-myocardial LGE were shared equally (50 % each). LGE was seen only in basal segments and predominantly involved anteroseptal (75 % of LGE group) and infero-septal (100 % of LGE group) segments, followed by the inferior wall (25 %) and inferolateral wall (25 %). LGE was patchy, with none exceeding 5 % of the total myocardium (Fig. 5).

Fig. 5.

LGE on CMRI: Events (N = 16) vs no events (N = 14) (% in bar diagram).

Average basal (mean ± SD; 1087 ± 72 vs. 994 ± 18, p < 0.001), mid-segment (mean ± SD; 1099 ± 67 vs. 1019 ± 17, p < 0.001), and apical segment (mean ± SD; 1116 ± 91 vs. 1001 ± 18, p < 0.001) T1 values were significantly higher in event group compared to no-event as was the total average T1 value in all segments (mean ± SD; 1101 ± 70vs 1007 ± 15, p < 0.001) (Fig. 6).

Fig. 6.

T1 CMRI findings in HFpEF: Events (N = 16) versus no events (N = 14) HFpEF – heart failure with preserved ejection fraction.

There were no significant differences between T1 values regarding LGE positivity in cases (p 0.321) (Supplementary Table 3). However, LGE negative cases had significantly increased T1 values in all segments compared to the healthy control group (Total T1, mean ± SD; 1053 ± 62vs 1020 ± 4, p < 0.001) (Supplementary Table 4).

4. Discussion

Our study showed that patients with HFpEF are more likely to be female, older, and obese, a fact documented by colleagues worldwide.^11,12 In our experience, hypertension, diabetes, and a history of coronary artery disease (CAD) were associated with HFpEF, known risk factors for developing HFpEF.^13,14

Previous publications reflect our findings of metabolic risk factors playing a dominant role in the pathogenesis of HFpEF.

Because of significant left atrial dilatation, carinal angles were significantly higher in heart failure patients. A carinal angle of more than 75° is considered abnormal and indicates left atrial dilatation. It is in concordance with a study published in 2010, which proved that the assessment of carinal angle to determine left atrial size was acceptable and repeatable.¹⁵ However, sub-carinal angle measurement and its association have not been reported in previous studies in patients with HFpEF.

In our experience, the Left ventricular mass index was significantly higher in cases by the Teichholz and ASE formula. Moreover, Left atrial volume was significantly higher in cases concordant with previous studies on heart failure with preserved ejection fraction. In the study by Roy et al, HFpEF patients had higher E/e' ratio, elevated indexed LA and RA volumes, tricuspid regurgitation velocity, and worse RV function, compared to age- and sex-matched healthy controls.⁷

MRI is the gold standard for anatomical measurements of cardiac chambers and walls. According to the echocardiographic finding, interventricular septal thickness and posterior wall were higher because of concentric LVH. Left ventricular end-diastolic and stroke volume were higher in the case group, probably because of associated mitral regurgitation.

Average mid-segment and total average T1 native values were significantly higher in cases, which have also been proven in various studies. T1 has been proven to be associated and correlated with histology-proven myocardial fibrosis.⁵

Cumulative event was 67 % in our cases over a mean follow-up of 25 months, similar to most outcome studies in heart failure with preserved ejection fraction. Mortality was 7 %, lower than previous studies, probably due to our study's exclusion criteria, which excluded patients with diseases with worse outcomes like amyloidosis, restrictive cardiomyopathy, and coronary artery disease.

The event and no-event groups had uncontrolled hypertension, reflected as LVH. NT-pro-BNP was significantly higher in the event group, indicating this group to be sicker and explaining its relation with subsequent events. Previous studies also showed the relationship with recurrent heart failure hospitalization and major adverse cardiovascular events.

Carinal angle and left atrial enlargement were higher in the event group due to left atrial enlargement. There were no statistically significant differences in echocardiographic measures between the event and no-event groups, probably indicating lesser sensitivity of echocardiography in predicting a major adverse cardiovascular event. Pericardial effusion was exclusively present in the event group, indicating elevated central venous pressure in sicker patients.

Interventricular septum and posterior wall thickness were higher in the event group compared to the no-event group. Previous studies have also proven the fact that increased severity hypertrophy and increased cardiac mass are associated with increased events of incident and recurrent heart failure as well as mortality.

Focal fibrosis in the form of LGE was significantly higher in the HFpEF population, with none being affected in the control group. However, there were no significant differences between the event and no-event groups, contrary to most studies involving HFrEF hypertrophic cardiomyopathy patients. However, Roy et al (26 %) and Kanagala et al (51 %) also reported similar results and attributed the difference in results to the technique for LGE measurements.^7,16

Our study focused on native T1 value, reflecting intra and extracellular changes. T1 value was significantly increased in the event group compared to the no-event group in our study population. In a study comparing the native myocardial T1 time values of HFpEF patients versus control subjects, Kanagala et al demonstrated that HFpEF patients had significantly higher native T1 time values (p = 0.021). However, the authors of that study did not report any relationships between native T1 time and clinical outcomes.¹⁶ Recent meta-analysis showed ECV (a surrogate marker of T1) rather than T1 is more predictive of adverse outcomes in HFpEF patients.¹⁷ However, all previous studies included the western population and very diverse phenotypes of the HFpEF population. Our study was first in a South Asian country, where we included "pure" HFpEF patients and excluded all coronary artery disease, hypertrophic cardiomyopathy, and amyloidosis patients with different prognoses and therapeutic interventions. This is why our patients' T1 value probably came out to be predictive of events. However, further larger studies are required to validate the continental difference in MRI findings.

We also found that LGE-negative patients had significantly higher T1 values than control, indicating the sensitivity of detecting HFpEF patients by higher T1 remains even in the absence of LGE.

5. Conclusions

Higher T1 but not T2 values were associated with the HFpEF group compared to the age and sex-matched control group. Focal fibrosis, as evidenced by LGE, was significantly more in HFpEF. Among HFpEF patients, high T1 myocardial values were associated with a higher all-cause death rate, stroke, and HF hospitalization in short-term follow-up. Higher NT Pro-BNP levels, higher diuretics requirement, increased cardiothoracic ratio and carinal angle in chest x-ray, and presence of pericardial effusion were associated with event groups.

6. Limitations

This was a non-blinded, single-centre, retrospective study. Due to its design, selection and referral bias couldn't be ruled out, and the control group was assumed to have a normal myocardium. Due to small sample size and number of events, the result of the study should interpreted with caution and a larger study with longer follow-up is solicited.

Patient consent for publication

Obtained.

Data availability statement

All data are incorporated into the article and its online supplementary material.

Ethical approval statement

Not applicable.

Permission to reproduce material from other sources

Not applicable.

Author contribution

SM (Conceptualization: Equal; Formal analysis: Lead; Writing – original draft: Lead; Writing – review & editing: Lead); AKVK (Conceptualization: Equal; Investigation: Equal; Supervision: Equal); GV, SSK (Investigation: Equal); BLS (Conceptualization: Equal; Formal analysis: Lead; Writing – review & editing: Lead); AA (Conceptualization: Equal; Formal analysis: Lead; Writing – review & editing: Lead); JR (Investigation: Equal) SM (Investigation: Equal).

Clinical trial registration

Not applicable.

Funding

None.

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 data

The following are the Supplementary data to this article:

figs1.

Supplementary Figure 1: HFpEF scoring sy

figs2.

figs3.

figs4.

figs5.