As previously described (15), FRENSHOCK is an observational, prospective, multicenter registry, including 772 patients admitted for CS between April and October 2016 in ICU/ICCU in France. All institutions were invited to participate, including university hospitals, general and regional hospitals, public and private hospitals (ICCUs, surgical ICUs, medical ICUs, and general ICUs).

All adult patients (≥18 years old) with CS were prospectively included in this registry if they met at least one criterion of each of the following three components: (1) Low cardiac output: low SBP < 90 mmHg and/or the need for maintenance with vasopressors/inotropes and/or a low cardiac index < 2.2 L/min/m²; (2) Left and/or right heart filling pressure elevation, defined by clinical signs, radiology, blood tests, echocardiography, or signs of invasive hemodynamic overload and (3) Signs of organ malperfusion, which could be clinical (oliguria, confusion, pale and/or cold extremities, mottled skin) and/or biological (lactate > 2 mmol/L, metabolic acidosis, renal failure, liver insufficiency).

For each patient, investigators were invited to identify one to three triggers among the following: ischemic (type 1 or 2 AMI), mechanical complications (valvular injury, ventricular septal defect), ventricular and supraventricular arrhythmia, severe bradycardia, iatrogenesis (medication induced), infections, non-observance of previous medication. Underlying cardiopathy was considered ischemic in the presence of at least one culprit lesion hemodynamically significant on coronary angiography (stenosis, thrombosis). VA-triggered CS status was defined by the managing physician.

First, general data on cardiological history (heart disease, previous ICD), coexisting conditions (kidney or pulmonary disease, cancer), risk factors (smoking status, hypertension, dyslipidemia, diabetes mellitus), treatments (including antiarrhythmic drugs) were recorded. Clinical, biological and echocardiographic data were collected at admission and 24 h. Clinical assessment included blood pressure, heart rate, sinus rhythm, signs of left and/or right heart failure, mottling, cardiac arrest. Biological data included bilirubin and creatinine levels, serum electrolytes, prothrombin time, hemoglobin, arterial blood gases and arterial lactate, C-reactive protein, troponin, BNP/Nt-proBNP. Echocardiographic evaluation mandatorily included left ventricular ejection fraction (visual evaluation or biplane Simpson’s method), presence of pericardial effusion and severe valvulopathy (defined as grade IV), in addition to which parameters such as TAPSE or S wave were often described.

Data on CS management included pharmacological treatment at admission, at discharge and at 1 year (catecholamines, beta-blockers, diuretics, ACEi, ARB, MRA, sacubitril/valsartan, antiarrhythmic), organ replacement therapies such as mechanical ventilation (invasive and/or non-invasive), short-term circulatory support (IABP, extracorporeal membrane oxygenation, Impella^®) and renal replacement therapy.

Short and long-term outcomes, including all-cause mortality, heart transplantation or ventricular assist devices (VAD), were assessed at 1 month and 1 year. The primary end point was 1-year all-cause mortality. Secondary end points included 1-month all-cause mortality, need for heart transplantation or VAD, rate of rehospitalizations, and the composite of death, heart transplantation or VAD. We investigated the cause of death in the VA group, distinguishing four possibilities (end-stage heart failure, sudden cardiac death or recurrence of intractable VA, other, unknown). Further comparisons were made between ischemic and non-ischemic cardiomyopathy, as well as between patients presenting with acute and chronic coronary syndromes. When done, VA catheter ablation (16) and myocardial revascularization (17) were performed according to the current techniques.

The study was conducted in accordance with guidelines for good clinical practice and French law. Written consent was obtained for all patients. Recorded data and their storage were approved by the CCTIRS (French Health Research Data Processing Advisory Committee) (no 15.897) and the CNIL (French Data Protection Agency) (no DR-2016-109).

Continuous variables are reported as means (SD) or medians and interquartile ranges (IQR) when appropriate. Categorical variables are described in numbers and percentages. Comparisons were made using Mann Whitney non-parametric test for continuous variables and chi-square test or Fisher’s exact test for categorical variables. All-cause mortality was assessed using Kaplan-Meier curves, and Cox proportional hazards models were used to determine the HR and 95% confidence interval (CI) for mortality. Log-rank test was carried out to compare survival between groups. An additional propensity score matching analysis was performed with the greedy nearest neighbor algorithm (6:1 ratio) using a multivariable logistic regression model including five covariates (age, sex, history of ischemic heart disease, LVEF ≤ 40% at admission, previous ICD) that were prognostically important for the outcome and to minimize confounding factors. A second comparison was made in VA-triggered CS group between ischemic and non-ischemic VA. Analysis were performed using R software [version 4.1.2 (2021-11-01)]. A p value < 0.05 was considered statistically significant.

Seven hundred seventy-two patients with CS were included in 49 centers, of which 3 were excluded for missing data (Graphical Abstract). Among the 769 patients, 94 were VA-triggered (12.2%). Table 1 summarizes baseline characteristics. Mean age was 65.8 ± 14.8 years, with a predominance of men (71.4%). 56% were already known for previous cardiac history (29.9% ischemic and 1% dilated) and cardiovascular risk factors were frequent (respectively, 47.3, 36.1, 28.3, and 27.8% for hypertension, dyslipidemia, diabetes, and current smoking). Ongoing long-term therapies for previous heart failure were common at admission (respectively, 41.1, 37.9, and 13.8% for betablockers, ACEi/ARB and aldosterone antagonist). Long-term antiarrhythmic drug therapies (essentially amiodarone) were noted in 17.4% of them.

As reported in Table 2, mean SBP was 101.3 ± 25.2 mmHg, with mottling in 39%. Initial echocardiographic data revealed mean LVEF of 26.3 ± 13.4%, median TAPSE of 13 mm (10–16) and median PSVtdi of 8 cm/s (6–11).

At admission, VA and non-VA triggered CS groups were similar regarding to age, sex, medical history or ongoing medication (Table 1). Only vitamin K antagonist was more frequent in the non-VA group (22.4 vs. 12.8%, p = 0.04).

Among the 675 non-VA triggered CS patients, main additional triggers were ischemic (32%), supra-ventricular tachycardia (14.8%), and infections (13.5%). By contrast, among the 94 VA-triggered CS patients, other most frequently associated triggers were ischemia (42.6%), mechanical complications (5.3%) and conduction disorders (5.3%) without statistical significance (Supplementary Table 1).

Clinical presentation was similar between groups at admission (Table 2) and 24 h (Supplementary Table 2) except a higher rate of cardiac arrest and a lower rate of sinus rhythm in the VA-triggered group (respectively, 29.8% vs. 7.4%, p < 0.01 and 34 vs. 54.6%, p < 0.01). Non-VA triggered CS patients presented with higher initial Nt-proBNP and CRP (respectively, 10,763 vs. 5,360 pg/ml, p < 0.01 and 29 vs. 15 mg/L, p < 0.01), higher creatinin and poorer prothrombin time at 24 h (respectively, 129 vs. 106 μmol/L, p = 0.04 and 60 vs. 70%, p < 0.01).

Echocardiographic evaluation showed similar biventricular dysfunction between groups with 24.4% vs. 26.6% (p = 0.11) and 14.0 vs. 12.0 mm (p = 0.58) for LVEF and TAPSE, respectively, for VA and non-VA triggered groups.

Table 3 summarizes in hospital management. Inotropes were used in 89.8%, without difference between VA and non-VA groups (respectively, 86.2 vs. 90.3%, p = 0.21). Dobutamine was the most frequently used (82.2% overall, 76.6 vs. 83.0%, p = 0.13), whereas norephinephrine was given in 53.5% (60.6 vs. 52.5%, p = 0.14) and levosimendan in 7.5% (5.3 vs. 7.7%, p = 0.4). Invasive ventilation was more frequently needed for VA-triggered CS (51.1 vs. 36.1%, p < 0.01). Short-term mechanical circulatory support needs were similar between groups for all categories, with 7.5 vs. 6.1% (p = 0.78) for IABP, 2.1 vs. 3.6% (p = 0.76) for Impella^® and 17 vs. 10.1% (p = 0.07) for ECLS. Renal replacement therapy was also equally used in the two groups (13.8 vs. 16.2%, p = 0.67).

All data related to anti-arrhythmic management are reported in Table 4. During initial care, data revealed similar use for all antiarrhythmic drugs, including amiodarone, betablockers and class 1 antiarrhythmic.

Twenty-four hours after admission, amiodarone (51.9 vs. 30.9%, p < 0.01) and others antiarrhythmic drugs (class 1 or sotalol) (15 vs. 6.9%, p = 0.02) were more frequently used in the VA group, unlike betablockers (18.1 vs. 13.2%, p = 0.23).

At discharge as at 1 year, no difference was shown about the use of any antiarrhythmic drug.

An ICD had been previously implanted in 14.9% for the VA group and 16.8% for the non-VA group (p = 0.65). Within 1 year after CS, ICD implantation was required for 11.8% of the VA-group against 4.1% for the non-VA group (p < 0.01). VA catheter ablation was performed for 13 patients of the VA group and 3 of the non-VA group because of occurrence of VA after admission in this group (14 vs. 0.5%, p < 0.01).

The Table 3 and Figure 1 show the absence of between-group difference in early or long-term all-cause mortality (26.6 vs. 25.8% and 42.6 vs. 45.3%, respectively, at 1-month and 1-year for VA and non-VA groups) nor in terms of 1-year rehospitalization (47.7 vs. 44.6% for VA and non-VA groups). LVEF at discharge was lower in the non-VA group (34.5 vs. 38.8%, p = 0.04) but VA-triggered CS resulted in more heart transplantation or need for VAD at one year compared to non-VA (17 vs. 9%, p = 0.02) (Table 3). The matched cohort included 658 patients (respectively, 94 and 564 for VA- and non-VA groups), with good balance between groups [all standardized mean differences below 0.1 after matching (Supplementary Tables 3, 4)], and did not show mortality difference, neither at 1 month [26.6 vs. 25.2%, HR 0.91 (95% CI 0.72–1.68), p = 0.67], nor at 1 year [42.6% vs. 44%, HR 0.98 (95% CI 0.70–1.37), p = 0.89] (Supplementary Figure 1).

Among the 94 VA-triggered CS, 40 (42.6%) revealed at least one unknown culprit lesion on coronary angiography, respectively, located on LAD, RCA, LMCA, and LCX for 20 (50%), 10 (25%), 7 (17.5%), and 1 (2.5%) (2 unknown). Culprit lesion’s revascularization was performed for 36 (90%). At baseline, history of ICD implantation (25.9 vs. 0%, p < 0.01) and betablockers (53.7 vs. 22.5%, p < 0.01) were more frequently encountered for non-ischemic VA-triggered CS. Other baseline characteristics were similarly distributed (Supplementary Table 5).

While clinical and echocardiographic parameters were similar, biological presentation of the non-ischemic VA triggered CS appeared worse, with higher median levels of creatinine (138 vs. 115 μmol/L, p = 0.03), bilirubin (21 vs. 12, p < 0.01) and Nt-proBNP (6,787 vs. 1,520 pg/ml, p = 0.046) and lower PT (53.5 vs. 71%, p < 0.01) (Supplementary Table 5). All clinical, biological and echocardiographic data 24 h after admission are reported in Supplementary Table 6.

Survival analyses did not show difference of all-cause mortality at 1 month [30% vs. 24.1%, HR 0.76 (95% CI 0.35–1.67), p = 0.5] and 1 year [42.5 vs. 42.6%, HR 0.97 (95% CI 0.52–1.81), p = 0.92] (Figure 2) between ischemic and non-ischemic groups. At 1 year, heart transplantation or VAD were needed for 14 patients (25.9%) of the non-ischemic group versus 2 (5%) of the ischemic group (p = 0.02) (Supplementary Table 7). 1-year rehospitalizations rate was similar (45.9 vs. 49.0%, p = 0.95).

All data relating to in-hospital management and antiarrhythmic drugs are reported in Supplementary Tables 7, 8.

Among the 94 patients of the VA-group, 40 died at 1 year: death cause was available for 30 of them. 15 (50%) died because of end-stage heart failure, 11 (36.7%) because of other life-threatening conditions (sepsis, neoplasia, etc.) and 3 (13.3%) due to sudden cardiac death or recurrence of intractable VA.

Additional analyses showed no difference in 1-month or 1-year all-cause mortality when VA triggered CS patients with acute ischemia were compared to remaining patients with stable ischemic heart disease (Supplementary Figure 2). However, there was a trend toward a poorer outcome in patients with acute ischemia defined by elevated troponin (with a threshold value of 10 μUI/L for standard troponin I, 200 μg/L for high sensitivity troponin I, and 2,000 ng/mL for high sensitivity troponin T) even if it did not reach statistical significance (Supplementary Figure 3 and Supplementary Table 9).