Results

Baseline characteristics

Records were reviewed for 287 recipients of a secondary prevention ICD between January 2010 and July 2021 (Figure 1). Of these, 185 patients were excluded on the basis of being >60 years of age. The remaining 102 patients were included for further investigation. These patients were dichotomised based on the criteria described in Table 1, such that 39 patients (38.2%) were identified with unexplained ventricular arrhythmia (UVA, Group 1) and the remaining 63 patients (61.8%) were labelled VA with clear aetiology (Group 2). Of these 63 patients, the identified aetiologies of ventricular arrhythmia fit into the following categories: 16 (25.3%) had ischaemic cardiomyopathy, 38 (60.3%) had non-ischaemic cardiomyopathy and 9 (14.3%) had a primary electrical aetiology which was evident on their resting baseline ECG.
The baseline characteristics for the patients in described in Table 2. Patients with UVA were clinically distinct from their counterparts with VA of clear aetiology. UVA patients were younger in age (35.6 ± 13.0 years vs 46.0 ± 8.6 years, p<0.001) and were more often female (48.7% vs 28.6%, p=0.04). A lower proportion of the patients with UVA had traditional cardiovascular risk factors, including hypertension (10.3% vs 30.2%, p=0.02) and diabetes (2.6% vs 15.9%, p=0.04). A greater proportion of UVA patients gave a history of prior syncope compared to patients with VA of clear aetiology (20.5% vs 9.5%, p=0.03).
When reviewing the culprit rhythm disorder which formed the indication for ICD, a higher proportion of patients with UVA had ventricular fibrillation compared to their counterparts with VA of clear aetiology (87.2% vs 50.8%, p<0.001), while a lower proportion had monomorphic VT (7.7% vs 46.0%, p<0.001). There was consequently a trend towards a greater proportion of UVA patients presenting in the context of a cardiac arrest (82.1% vs 65.1%, p=0.06).
When reviewing the ICD subtype across the two groups, patients with UVA were more likely to receive subcutaneous ICDs than their counterparts with VA of clear aetiology (20.5% vs 4.8%, p=0.01). Both groups had similar rates of implantation with single chamber and dual chamber devices. No patients in the UVA group received a biventricular device compared with 5 patients with VA of clear aetiology (7.9%).

First-line cardiac investigations

The results of baseline ECGs, transthoracic echocardiograms and coronary assessments are shown in Table 3. All patients reviewed had a baseline ECG available for analysis. Baseline ECG characteristics were similar across the three groups, with the only significant difference being a higher rate of axis deviation in patients with VA of clear aetiology compared to those with UVA (28.6% vs 10.3%, p=0.02). Sinus rhythm was the prevalent baseline rhythm in the majority of patients in both groups (UVA: 97.4%, VA of clear aetiology: 87.3%, p=0.08). Of note, two patients (5.1%) in the UVA group exhibited the early repolarisation pattern on their ECG (without additional diagnostic features of early repolarisation syndrome), while no patients with VA of clear aetiology exhibited this finding.
All patients analysed had a baseline transthoracic echocardiogram available for review. In keeping with the criteria used to defined the two groups, patients with UVAs had significantly lower rates of structural heart disease compared with their counterparts with VA of clear aetiology. Specifically, UVA patients had a higher LV ejection fraction (59 ± 8.4% vs 42.1 ± 15.3%, p<0.001) and smaller LV cavity size (50.8 ± 6.4mm vs 56.1 ± 11.0mm). By definition, no patients in the UVA group exhibited impaired RV function, moderate-or-worse valve dysfunction or segmental hypokinesis. These features were respectively present in 30.2%, 14.3% and 30.2% of the patients with VA of clear aetiology.
Coronary assessments (either CTCA or invasive coronary angiography) were widely performed in patients with UVA (87.2%) and VA of clear aetiology (84.1%). By definition, no patients in the UVA group had obstructive coronary disease, which in turn was present in 27.0% of patients with VA of clear aetiology. 14.3% of patients with VA of clear aetiology required follow-on revascularisation.

Second-line cardiac investigations

Table 4 demonstrates the adoption rate of ‘second-line’ investigations for the workup of ventricular arrhythmia in patients with UVA and VA of clear aetiology. Cardiac MRI was the most commonly utilised modality in both groups, but was more commonly adopted in patients with UVA (82.1% vs 63.5%, p=0.046). The remainder of the second-line investigations, flecainide challenge, genetic testing, EP study and exercise stress ECG, were only utilised in a minority of patients in both groups. Combined workup with all five second-line investigations was only utilised in 3 patients, all of whom were patients with UVA.
The individual diagnoses that were identified as a result of second-line investigations in patients with UVA is shown in Figure 2. CMR facilitated diagnosis of an underlying aetiology of ventricular arrhythmia in 8 patients. The late gadolinium enhancement (LGE) pattern of ventricular scar was suggestive of a diagnosis in the majority of these cases, including 2 cases of sarcoidosis, 1 case of old (inactive) myocarditis and 1 case of transmural scar in a patient with non-obstructive CAD (suggestive of a coronary embolic event). Morphological analysis of the RV was suggestive of underlying arrhythmogenic cardiomyopathy in 2 patients.
Flecainide challenge was performed in 8 patients (20.5%) and revealed a provokable ECG pattern suggestive of Brugada syndrome in 2 cases. Genetic screening was performed in 10 patients; 8 of these patients reported a family history of sudden cardiac death. A likely genetic culprit was identified in 3 cases. One patient was found to have a ryanodine receptor-2 mutation suggestive of catecholaminergic polymorphic VT (CPVT), one patient was found to have a desmoglein-2 mutation suggestive of ARVC, and one patient was found to have a desmin mutation signalling the presence of an underlying early NICM. An EP study was performed in 6 patients (15.4%). Focal VT originating from the LV inferior septum and perimitral VT were found in one patient each. In one case, the EP study demonstrated evidence of a right posterior accessory pathway; suggesting that the malignant arrhythmia in question may have been a pre-excited tachycardia rather than a ventricular tachyarrhythmia. While exercise stress ECGs were performed in 8 patients (20.5%), no patients were identified to have catecholaminergic polymorphic VT or congenital long QT syndrome.
Taken together, 17 patients in the UVA cohort (43.5%) had a suggestive aetiology for ventricular arrhythmia identified on one or more of the five second-line investigations.

Prescribed anti-arrhythmic therapy

Table 5 reviews the rates of prescription of antiarrhythmic therapies at time of hospital discharge in each group. A lower proportion of patients with UVA were prescribed at least one antiarrhythmic drug at time of hospital discharge compared with their counterparts with VA of clear aetiology (64.1% vs 88.9%, p=0.003). This was predominantly driven by lower rates of prescription of beta blockers (51.3% vs 81.0%, p=0.002) and amiodarone (5.1% vs 34.9%, p=0.001).

Device-detected arrhythmia burden

Table 6 reviews the data collected from device interrogations in the first 5 years post-device implantation. At least two device interrogation reports were available for 34 patients with UVA (87.1%) and 51 patients with VA of clear aetiology (81.0%). Reviewing these records revealed a significantly greater requirement for device-delivered tachy-therapies in the UVA group compared with the VA of clear aetiology group (30.8% vs 14.3%, p=0.04). Rates of device-detected NSVT were similar across the UVA and VA of clear aetiology cohorts (43.6% vs 38.1%, p=0.58). 3 patients were found to have died during follow-up: all three were in the VA of clear aetiology cohort. 2 deaths resulted from end-stage systolic heart failure and one death resulted from VT storm.