Validation of predictors of successful weaning from mechanical circulatory support and histological features in lymphocytic fulminant myocarditis

Abstract

Objective

To validate the prognostic utility of peak creatine kinase-myocardial band isoenzyme (CK-MB) level, left ventricular posterior wall diameter (LVPWD), and a score model for successful weaning from mechanical circulatory support (MCS) in fulminant myocarditis (FM). Furthermore, the severity of the histological data was stratified, and its relationship with the clinical findings was investigated.

Methods

This retrospective single-centre cohort study included 20 patients with FM treated with MCS between 2013 and 2023. The primary outcome was successful weaning from MCS (MCS decannulation and survival to discharge). Two validations were conducted: Validation A, to evaluate the predictive value of peak CK-MB and LVPWD, and validation B, to evaluate the performance of the prediction model for successful MCS weaning derived from the CHANGE PUMP Study. The histological findings of 21 specimens were used for additional analyses. The pathological severity of myocardial inflammation, necrosis and oedema was stratified, and their relationship with clinical data was assessed.

Results

14 patients were successfully weaned from MCS. In validation A, the weaning success rate was significantly higher in patients with peak CK-MB≤185 IU/L than in those with peak CK-MB>185 IU/L. Similar findings were observed in patients with LVPWD≤11 mm compared with those with LVPWD>11 mm. In Validation B, the median predictive MCS weaning success rate for the entire cohort was 80%. Compared with patients who underwent MCS decannulation successfully, those who did not had a significantly lower predictive success rate. The C-statistic of the model was 0.845. The histological severity at admission was not associated with the primary outcome. However, the CK-MB levels and left ventricular ejection fraction (LVEF) obtained simultaneously were correlated with histological severity.

Conclusions

Previously reported predictors of successful weaning from MCS in FM demonstrated reasonable discrimination. The severity of FM evaluated by histological findings was associated with CK-MB levels and LVEF.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Fulminant myocarditis (FM), treated with mechanical circulatory support (MCS), sometimes fails to recover from cardiogenic shock. For such patients, early consideration of advanced therapies (such as left ventricular assist device implantation or heart transplantation) is necessary. Accurate prediction of the probability of successful weaning from MCS would be helpful in clinical decision making. Previous small cohort studies have reported several predictors of weaning from MCS.

WHAT THIS STUDY ADDS

• This study validated the utility of previously reported predictors for successful weaning from MCS in patients with lymphocytic FM. The peak creatine kinase-myocardial band isoenzyme combined with the left ventricular posterior wall diameter and the prediction model for MCS decannulation derived from the CHANGE PUMP Study exhibited their utility in the present study. Moreover, the severity of FM, as evaluated by histological findings, was associated with myocardial enzyme levels and left ventricular contraction, which were obtained simultaneously.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• By using validated predictors of successful weaning from MCS, the probability of MCS weaning could be assessed early after MCS introduction, allowing for timely consideration of the need for advanced therapy. Myocardial enzyme levels and left ventricular ejection fraction may help predict the pathological severity of myocarditis and potentially guide treatment strategies.

Introduction

The advent of novel therapies and therapeutic devices has increased survival and improved the long-term prognosis of difficult cases, particularly in patients with cardiogenic shock.^{1 2} Advancements in mechanical circulatory support (MCS) have resulted in an increase in the number of patients who receive temporary MCS in the acute phase of cardiogenic shock.³ Fulminant myocarditis (FM) often leads to cardiogenic shock and has a high mortality rate requiring MCS.⁴,⁸ Patients who fail to decannulate MCS may require left ventricular assist device (LVAD) implantation or heart transplantation. Cardiac enzyme levels, left ventricular wall thickness⁹ and prognostic score models⁷ have been used as predictors of successful weaning from MCS. However, due to the low prevalence of this disease, the utility of these predictors has only been investigated in a small cohort without validation. The pathological severity of myocarditis has been reported to be associated with the prognosis of patients with lymphocytic FM⁵; however, studies on the prediction of successful MCS weaning are lacking.

A recent expert consensus from the American College of Cardiology proposed classifying myocarditis into four stages according to the risks, symptoms and severity.¹⁰ However, it is still difficult to predict when patients become irreversibly dependent on MCS. Thus, validating the association between previously reported prognostic predictors and successful weaning from MCS in patients with FM may be beneficial.

Given this background, our objective was to validate the utility of previously identified predictors of successful weaning from MCS in patients with lymphocytic FM. Furthermore, the severity of the histological data was stratified, and its relationship with the clinical findings was investigated.

Methods

Study design

In this retrospective, single-centre, observational cohort study, we registered 20 consecutive patients with lymphocytic FM who were treated with veno-arterial extracorporeal membrane oxygenation (VA-ECMO), extracorporeal ventricular assist device (Ex-VAD) or IMPELLA heart pump (Abiomed, Danvers, Massachusetts, USA) at the Shinshu University Hospital (Nagano, Japan) between January 2013 and June 2023. Patients were initially identified by searching the MCS database. FM was defined as acute myocarditis¹¹ that required ECMO, IMPELLA or Ex-VAD for cardiogenic shock. Histologically proven myocarditis was defined based on the diagnostic criteria set forth by the European Society of Cardiology, American College of Cardiology or the Japanese Circulation Society guidelines.^{10 12 13} FM was diagnosed by attending clinicians based on available symptoms, laboratory data, electrocardiography, echocardiography, endomyocardial biopsy, haemodynamic parameters and coronary angiography. The decision to escalate, de-escalate or wean MCS was made by a multidisciplinary heart team, which included cardiologists and cardiovascular surgeons. For the weaning process, the VA-ECMO flow rate was decreased gradually by 0.5 L/min every 6–12 hours until reaching 1.0–1.5 L/min. The IMPELLA flow rate was decreased by 1–2 P-levels every 6–12 hours until reaching P-level 2. MCS was then decannulated in the following 12–24 hours if weaning was deemed possible for the patient.^{14 15}

The following validations were performed to evaluate the utility of previously reported clinical predictors (online supplemental figure S1). Validation A: To evaluate the predictive value of peak creatine kinase-myocardial band isoenzyme (CK-MB) 185 IU/L and left ventricular posterior wall diameter (LVPWD) 11 mm on admission for successful weaning from MCS.⁹ Validation B: To evaluate the performance of the prediction model for successful weaning from MCS.⁷ The primary outcome was successful weaning from MCS, defined as successful decannulation and survival to discharge. Additional analyses were performed using tissue specimens from 17 patients (21 specimens) who underwent endomyocardial biopsies during hospitalisation. The severity of myocardial inflammation, necrosis and oedema in each endomyocardial biopsy specimen was retrospectively assessed by an experienced cardiovascular pathologist (TU) and stratified into four levels: none (0 points), mild (1 point), moderate (2 points) and severe (3 points). Online supplemental table S1 and figure 1 present definitions and representative images, respectively. The total score for each specimen was calculated and defined as the severity of myocarditis according to the pathological findings. Clinical data and survival status were collected in October 2024 via chart review. Informed consent was obtained from the Shinshu University Hospital’s website. The patients were informed that they could opt out anytime during the study process.

Figure 1. Representative images of the endomyocardial biopsy specimens. (A) No myocardial inflammation (0 points), no necrosis (0 points) and no oedema (0 points), yielding a total score of 0 points. (B) Mild myocardial inflammation (1 point), no necrosis (0 points) and mild oedema (1 point), yielding a total score of 2 points. (C) Moderate myocardial inflammation (2 points), no necrosis (0 points) and moderate oedema (2 points), yielding a total score of 4 points. (D) Severe myocardial inflammation (3 points), severe necrosis (3 points) and severe oedema (3 points), yielding a total score of 9 points.

Measurements and data collection

Data regarding the following factors were collected at the time of MCS introduction, 48 hours after MCS introduction, MCS decannulation and discharge: age and sex (collected only at the time of MCS introduction), height, weight, medical history, laboratory data (including the peak level of CK-MB during hospitalisation), medications, electrocardiography, echocardiography, angiography and surgical data. In addition, data on CK, CK-MB, high-sensitivity troponin T (hs-TnT) levels, left ventricular ejection fraction (LVEF), LVPWD and QRS duration were collected when the patient underwent endomyocardial biopsy. Comprehensive transthoracic echocardiography was performed by experienced sonographers or attending physicians using high-quality, commercially available ultrasound systems. The biplane-modified Simpson’s method or visual evaluation was used to measure the LVEF. All clinical events after MCS weaning and decannulation were followed up until discharge.

Definitions

Cardiogenic shock was defined as progressive end-organ hypoperfusion caused by a low cardiac output state from severe cardiac dysfunction and a persistent systolic blood pressure <90 mm Hg or mean blood pressure <60 mm Hg for >30 min that is refractory to fluid resuscitation or requires vasoactive agents, MCS or both.¹⁶,¹⁸ In accordance with the definition provided in the original study, laboratory and echocardiographic data obtained at the time of MCS introduction and 48 hours later are presented as 0 hour and 48 hours, respectively.⁷ Similarly, Δaspartate aminotransferase (AST) and ΔLVEF were calculated using the definition described in the original paper.⁷ Successful weaning from MCS was defined as MCS decannulation and survival to discharge. Unsuccessful weaning from MCS was defined as in-hospital death following MCS decannulation or escalation to LVAD.

Statistical analysis

Continuous variables were summarised as mean±SD if they were normally distributed and as median with IQR if they were not normally distributed. Normality was assessed using the Shapiro-Wilk W-test. Baseline characteristics were compared using contingency tables and the χ² test for categorical variables, the t-test for normally distributed continuous variables, or the Mann-Whitney U test for non-normally distributed continuous variables. For validation A, patients were divided into three categories stratified according to peak CK-MB levels and LVPWD: low-risk (CK-MB≤185 IU/L and LVPWD≤11 mm), intermediate-risk (CK-MB≤185 IU/L or LVPWD≤11 mm) and high-risk (CK-MB>185 IU/L and LVPWD>11 mm) for unsuccessful weaning from MCS. Weaning success rates were compared among the three groups. The predictive MCS weaning success rate was calculated for each patient using the following formula: predictive weaning success rate (%) = 1/[1+exp(0.688–0.104×LVEF_48 hours−0.143×ΔLVEF_48h+1.171×ΔAST_48 hours)]×100.⁷ Prediction performance was assessed by C-statistics using a receiver operating characteristic curve. Spearman’s rank correlation method was used as a non-parametric measure of the association between histological findings and laboratory data. A p<0.05 was considered statistically significant. All statistical analyses were performed using SPSS Statistics for Windows, V.26 (IBM).

Patient and public involvement

Patients or the public were not involved in our research’s design, recruitment, conduct and reporting.

Results

Patient characteristics

The patient characteristics are shown in table 1. The median age was 59 (IQR: 46–66) years, and 60% (n=12) of the patients were male. 14 patients (70%) were successfully weaned from MCS. The reasons for unsuccessful weaning were as follows: death, 5 (cardiac death, 4; non-cardiac death, 1); and MCS escalation to LVAD, 1. Compared with patients who underwent MCS decannulation successfully, those who did not undergo MCS decannulation successfully were younger and had a lower prevalence of hypertension. In addition, the proportion of patients with complete atrioventricular block (CAVB) and ventricular arrhythmia was higher in those with unsuccessful MCS weaning. The median duration of temporary MCS was 11⁶,¹⁶ days. The specific data for each case are listed in table 2. The clinical findings on MCS decannulation and discharge in patients who were successfully weaned from MCS are shown in online supplemental table S2.

Table 1. Patient characteristics.

Variable	Overall population	Successful weaning from MCS
		Yes	No	P value
	(n=20)	(n=14)	(n=6)
Age (years)	59 (46–66)	63 (57–68)	46 (29–53)	0.026
Male, n (%)	12 (60)	9 (64)	3 (50)	0.55
BSA (m2)	1.74±0.18	1.76±0.18	1.68±0.18	0.37
Systolic blood pressure (mm Hg)	86 (79–102)	89 (82–104)	84 (77–102)	0.718
Diastolic blood pressure (mm Hg)	62±22	60±23	67±21	0.543
Mean blood pressure (mm Hg)	69 (62–81)	69 (63–80)	72 (57–89)	0.841
Pulse pressure (mm Hg)	26±13	28±14	22±10	0.42
Heart rate (beats/min)	94±22	100±25	83±9	0.041
Hypertension, n (%)	7 (35)	7 (50)	0 (0)	0.032
Diabetes mellitus, n (%)	1 (5)	1 (7)	0 (0)	0.502
Mechanical circulatory support
Ex-VAD, n (%)	1 (5)	0 (0)	1 (17)	0.117
IMPELLA, n (%)	6 (30)	6 (43)	0 (0)	0.055
VA-ECMO, n (%)	18 (90)	13 (93)	5 (83)	0.515
Pharmacological treatment
Glucocorticoid, n (%)	7 (35)	6 (43)	1 (17)	0.26
IVIG, n (%)	5 (25)	4 (29)	1 (17)	0.573
Laboratory data
BNP_0 hour (pg/mL)	643 (488–1505)	758 (544–1522)	557 (188–1503)	0.25
CK_0 hour (IU/L)	1483 (295–2397)	1483 (301–2391)	1340 (159–3230)	0.999
CK-MB_0 hour (IU/L)	44 (13–93)	48 (13–95)	21 (9–67)	0.277
Peak CK (IU/L)	2360 (861–3611)	2058 (774–3031)	9135 (1196–18 906)	0.091
Peak CK-MB (IU/L)	85 (35–180)	73 (33–103)	300 (49–534)	0.091
AST_0 hour (IU/L)	341 (83–1412)	396 (84–2611)	223 (62–579)	0.179
AST_48 hours (IU/L)	295 (79–1082)	340 (73–840)	206 (84–5343)	0.841
Electrocardiogram
QRS (ms)	118 (78–148)	117 (78–142)	130 (96–299)	0.391
CAVB, n (%)	8 (40)	3 (21)	5 (83)	0.01
VT/VF, n (%)	9 (45)	4 (29)	5 (83)	0.024
Echocardiography
LVEF_0 hour (%)	17±11	19±12	14±8	0.307
LVEF_48 hours (%)	20(10–35)	20(10–40)	10(8–20)	0.091
LVDD_0 hour (mm)	41±7	43±6	37±6	0.07
LVDS_0 hour (mm)	38±7	39±7	35±8	0.189
IVSD_0 hour (mm)	10 (8–11)	10 (8–12)	11 (8–11)	0.999
LVPWD_0 hour (mm)	11 (9–12)	10 (9–11)	12 (11–12)	0.091

Values are presented as medians and IQRs or n (%).

ASTaspartate aminotransferaseBNPbrain natriuretic peptideBSAbody surface areaCAVBcomplete atrioventricular blockCKcreatine kinaseCK-MBcreatine kinase myocardial bandEx-VADextracorporeal ventricular assist deviceIVIGintravenous immunoglobulinIVSDinterventricular septal diameterLVDDleft ventricular diastolic diameterLVDSleft ventricular systolic diameterLVEFleft ventricular ejection fractionLVPWDleft ventricular posterior wall diameterMCSmechanical circulatory supportVA-ECMOveno-arterial extracorporeal membrane oxygenationVFventricular fibrillationVTventricular tachycardia

Table 2. Specific data of each patient.

Cases	Peak CK-MB (IU/L)	LVPWD_0 hour (mm)	AST_0 hour (IU/L)	AST_48 hours (IU/L)	LVEF_0 hour (%)	LVEF_4 8hours (%)	MCS weaning success rate (%)	Outcome
Case 1	132	9	29 240	2790	17	10	60	Success
Case 2	69	21.3	7550	1203	5	10	89	Success
Case 3	98	17	185	65	10	10	75	Success
Case 4	854	11.5	481	272	0	0	46	Failure
Case 5	404	11.2	874	140	10	10	79	Failure
Case 6	119	10.8	479	363	10	20	96	Success
Case 7	93	10.4	2544	3700	15	19	79	Success
Case 8	17	12	49	11 140	20	10	< 1	Failure
Case 9	93	9.4	312	95	10	38	99	Success
Case 10	35	10.8	2810	245	10	10	81	Success
Case 11	427	9	76	71	15	10	43	Failure
Case 12	282	9	826	554	10	10	68	Success
Case 13	60	11.7	66	88	16	20	83	Failure
Case 14	77	10	84	719	20	25	< 1	Success
Case 15	13	9	1592	516	30	40	99	Success
Case 16	35	10.8	68	39	19	20	88	Success
Case 17	196	13	370	3410	20	20	< 1	Failure
Case 18	25	7	84	60	30	50	99	Success
Case 19	59	10.9	250	317	43	40	94	Success
Case 20	11	10.4	83	76	36	50	99	Success

ASTaspartate aminotransferaseCK-MBcreatine kinase myocardial bandLVEFleft ventricular ejection fractionLVPWDleft ventricular posterior wall diameterMCSmechanical circulatory support

Validation A

The weaning success rate was significantly higher in patients with a peak CK-MB≤185 IU/L than in those with a peak CK-MB>185 IU/L (13/15 (87%) vs 1/5 (20%), p=0.005). Similar findings were observed in patients with LVPWD≤11 mm compared with those with LVPWD>11 mm (12/13 (92%) vs 2/7 (29%), p=0.003). Figure 2 presents a scatter plot of successful weaning from MCS, stratified according to the cut-off value of the peak CK-MB level and LVPWD. The weaning success rate was significantly higher in patients categorised into the low-risk group (low-risk, 11/11 (100%); intermediate-risk, 3/6 (50%); high-risk, 0/3 (0%), p=0.006). All patients in the low-risk group were successfully weaned from MCS, and the predictive values of CK-MB and LVPWD were good.

Figure 2. A scatter plot of successful weaning from MCS stratified according to the cut-off values of peak CK-MB and LVPWD. Circles represent cases that underwent successful MCS decannulation, while crosses indicate cases that did not undergo successful MCS decannulation. CK-MB, creatine kinase myocardial band; LVPWD, left ventricular posterior wall diameter; MCS, mechanical circulatory support.

Validation B

The median predictive MCS weaning success rate in the entire cohort was 80% (49%–95%). Compared with patients who underwent MCS decannulation successfully, those who did not undergo MCS decannulation successfully had a significantly lower predictive success rate (89% (73%–99%) vs 44% (0%–80%), p=0.015). The predictive MCS weaning success rate for each case is listed in table 2. The C-statistic of the model was 0.845, demonstrating reasonable discrimination (online supplemental figure S2).

Histological findings

Table 3 presents the details of histological findings and the clinical data obtained simultaneously. Four of the 17 patients who underwent an endomyocardial biopsy at the time of MCS introduction later underwent a repeat biopsy (case 1_2 at 14 days, case 2_2 at 30 days, case 11_2 at 33 days and case 12_2 at 6 days after admission). Histological severity of myocarditis at admission was not associated with the primary outcome. The CK-MB level and LVEF were significantly correlated with the severity of myocardial inflammation, necrosis, oedema and the total histological score (online supplemental table S3). The QRS duration was significantly correlated with the severity of myocardial inflammation, but not with necrosis or oedema. No significant correlations were observed between the histological severity and CK or LVPWD. The correlation between hs-TnT levels and histological findings could not be confirmed, as the hs-TnT levels were above the upper reference limit of the troponin assay (ie, hs-TnT>10 ng/mL) in some patients. However, patients with higher hs-TnT levels tended to have more severe histological data.

Table 3. Histological and clinical findings.

Cases	Inflammation	Necrosis	Oedema	Total histological score (points)	CK (IU/L)	CK-MB (IU/L)	hs-TnT (ng/mL)	LVEF (%)	LVPWD (mm)	QRS (ms)
Case 1	3	2	2	7	2319	132	>10	17	9	120
Case 1_2	0	0	0	0	24	3	1.1	29	12.8	88
Case 2	3	2	3	8	1175	34	>10	5	21.3	116
Case 2_2	2	0	0	2	37	4	0.42	65	8.4	98
Case 3	3	2	3	8	606	74	5.36	10	17	130
Case 4	3	3	3	9	471	18	7.69	0	11.5	N/A
Case 5	3	1	2	6	2208	82	1.8	10	11.2	200
Case 6	3	1	2	6	2368	100	2.68	10	10.8	250
Case 8	3	1	2	6	570	11	8.91	20	12	398
Case 9	3	1	2	6	315	10	2.38	10	9.4	118
Case 10	3	1	2	6	247	35	3.26	10	10.8	94
Case 11	3	2	3	8	3358	236	>10	15	9	77
Case 11_2	1	0	1	2	67	3	0.06	10	9.4	110
Case 12	3	1	2	6	2513	282	>10	10	9	184
Case 12_2	3	1	2	6	620	58	3.28	16	11.7	130
Case 13	3	1	2	6	200	30	2.38	5	13	108
Case 14	1	0	2	3	2388	41	0.45	20	10	78
Case 16	1	0	2	3	800	35	>10	19	10.8	78
Case 17	1	0	1	2	4927	41	3.67	20	13	114
Case 18	2	0	1	3	292	12	2.16	30	7	86
Case 19	3	2	3	8	1726	59	7.23	43	10.9	140

The severity of myocardial inflammation, necrosis and oedema was assessed in each endomyocardial biopsy specimen and graded as follows: 0, none; 1, mild; 2, moderate and 3, severe.

CK-MBcreatine kinase myocardial bandhs-TnThigh-sensitivity troponin TLVEFleft ventricular ejection fractionLVPWDleft ventricular posterior wall diameterN/Anot applicable

Discussion

This study validated the utility of previously reported predictors for successful weaning from MCS in patients with lymphocytic FM. The peak CK-MB level combined with LVPWD, and the prediction model for MCS decannulation derived from the CHANGE PUMP Study, exhibited their utility in the present study. Moreover, the severity of FM, as evaluated by histological findings, was associated with myocardial enzyme levels and left ventricular contraction, which were obtained simultaneously.

Several studies have reported predictors to evaluate prognosis or cardiac function recovery in patients with FM.⁵⁷,^{9 19} Recently, lower renal function, blood pressure prior to MCS introduction and invasive haemodynamic assessment using a Swan-Ganz catheter were reported to be associated with a better prognosis in patients with FM requiring IMPELLA support.⁸ However, established predictors of successful weaning from MCS still do not exist. If temporary MCS weaning is difficult in patients with FM, attending physicians must consider heart transplantation, implantable LVAD or withdrawal from intensive care (bridge to decision). Predicting whether patients can recover from myocarditis at an early stage of therapy would be beneficial in clinical practice.²² Matsumoto et al identified that CK-MB levels and LVPWD were clinical determinants of successful weaning from VA-ECMO.⁹ Sawamura et al developed a predictive model of bridge to recovery in patients with FM using clinical data on admission and 48 hours after VA-ECMO introduction.⁷ However, the number of patients included in these studies was small owing to the rarity of the condition, and the predictive factors were not validated. Our study validated the prognostic utility of these predictors in another cohort and revealed that both predictors exhibited good predictive values for successful weaning from MCS.

In the present study, the rate of successful weaning from MCS among patients who experienced CAVB and ventricular arrhythmia was significantly lower than that of those who did not. This finding is consistent with that of previous studies reporting the prognostic impact of atrioventricular block and ventricular arrhythmia in patients with myocarditis.^{5 9 23 24} Younger individuals were more susceptible to decannulation failure in our study; however, the reason for this is unclear. Previous studies have reported that young age is associated with fulminating²⁵ or relapses²⁶ of myocarditis; however, no studies have reported an association with poor prognosis. This may be attributed to sample size issues. No significant associations between the primary outcome and peak CK-MB levels (per IU/L) or LVPWD (per mm) were observed in the present study, which may also be attributed to the small sample size.

The severity of myocarditis assessed during the acute phase (17 of 20 patients) was not associated with successful weaning from MCS. However, the severity of myocardial inflammation, necrosis and oedema correlated with myocardial enzyme levels obtained simultaneously in the 21 specimens. The CK-MB levels measured at the time of endomyocardial biopsy were consistent with the peak CK-MB levels in only 6 of the 20 cases. Thus, the severity of myocarditis cannot be accurately evaluated based on a single histological assessment at the time of MCS introduction. As LVEF and QRS duration are also correlated with pathological findings, histological severity may constantly change, similar to clinical indices. The findings of the present study and those of Matsumoto et al suggest that the maximum severity of myocarditis may be associated with MCS decannulation and prognosis, and that the severity of inflammation and necrosis is also reduced when the myocardial enzyme levels are decreased. Myocardial enzyme levels obtained from blood tests may provide a means to predict the activity and severity of myocarditis. High-sensitivity cardiac troponin is mainly used to diagnose myocarditis; however, myocardial enzyme levels may help predict the severity of myocarditis and facilitate the selection of treatment strategies (eg, considering LVAD implantation, heart transplantation or adding immunosuppressive drugs in giant cell myocarditis).

The findings of our study indicate that the previously reported predictors of successful weaning from MCS, peak CK-MB, LVPWD and the prediction model are useful and can be used to manage patients with FM who require MCS. Furthermore, myocardial enzymes, such as CK-MB and high-sensitivity cardiac troponin, can be used to predict the severity of myocarditis and develop treatment strategies.

This study had several limitations. First, this was a single-centre study with an extremely small sample size, which may represent a selected cohort. Second, the consistency of LVPWD measurements was not robust. Accurate measurement of LVPWD in patients undergoing MCS is challenging because of the limited view. Third, the association between hs-TnT levels and the severity of myocarditis on histological findings could not be evaluated because the hs-TnT levels of some patients were above the upper reference limit of the troponin assay. Fourth, further validation in other races is warranted, given that the predictors were validated in the same races as previously reported.

In conclusion, previously reported predictors of successful weaning from MCS in patients with FM demonstrated reasonable discrimination. Moreover, the severity of FM evaluated by histological findings was associated with myocardial enzyme levels and left ventricular contraction, which were obtained simultaneously. These findings could be useful in developing treatment strategies for patients with FM who require MCS.

Data availability statement

Data are available on reasonable request.