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

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.

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.

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.

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.

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.