Wearable cardioverter defibrillator for preventing sudden cardiac death in patients at risk: An updated systematic review of comparative effectiveness and safety

Abstract

Objectives

To synthesise the available evidence of wearable cardioverter defibrillator (WCD) therapy as an add-on measure to optimal medical therapy (OMT) or as a replacement of hospital stay.

Methods

An update systematic review (SR) of comparative effectiveness and safety of WCD therapy was conducted. We included randomised controlled trials (RCT), prospective comparative studies and prospective uncontrolled studies with at least 100 patients. A narrative synthesis of the evidence was conducted.

Results

One RCT (n = 2348) and further eleven observational studies (n = 5345) fulfilled our inclusion criteria. In the only available RCT, the use of the WCD was not statistically associated with a clinical benefit on arrhythmic mortality in post-myocardial infarction (MI) patients with an ejection fraction of ≤35%. The compliance with WCD therapy was low in the RCT and high in observational studies, with ten observational studies reporting on a daily wear time between 20 and 23.5 h. The range of percentage of patients receiving at least one appropriate shock was 1–4.8% and the rate of first shock success was reported to be 100% in three studies. Serious adverse events (SAEs) such as inappropriate shocks occurred rarely, with between 0% and 2% of patients being inappropriately shocked within ten observational studies. In one of the observational studies, two patients (2%) were allergic to nickel developing skin rash and false alarms occurred in 58 patients (57%) in this study. Another registry study (n = 448) reported milder AEs, such as dermatitis and pressure marks, occurring in 0.9% and 0.2% of enrolled patients, respectively.

Conclusion

The only available RCT failed to show superiority of add-on use of WCD in post MI patients. Observational evidence shows that the compliance with WCD is good, but the evidence is afflicted with selection bias and the inclusion of diverse mixed patient populations diluting the ability to draw indication-specific conclusions on the utility of the device. More comparative data is needed to justify continuing or expanding use of WCD therapy.

1. Introduction

The wearable cardioverter defibrillator (WCD) therapy is a temporary measure for primary and secondary prevention of sudden cardiac death (SCD). The WCD should be worn for the whole day, except for certain situations such as taking a shower or bath. The presence of a caregiver or a family member is then recommended [1,2].

The WCD monitors the heart function of a patient at risk of SCD and automatically delivers an electrical shock when needed. A specific tracking algorithm tests ventricular tachycardia (VT) or ventricular fibrillation (VF). If a life-threatening rhythm is detected, the device delivers a treatment to restore normal rhythm. Before treatment is coming, the device alerts the patient so that WCD treatment can still be prevented by pressing a response button on the WCD [3,4].

As of 2022, the LifeVest® WCD 4000 (Zoll Medical Corporation) is the only WCD device being commercially available in Europe. According to the manufacturer, the WCD is a class IIb device [4]. The ASSURE WCD System Kit (Kestra Medical Technologies, Inc.) is further approved by the Federal Drug Administration (FDA) in the United States [5], but without holding a CE mark.

The alternative to WCD therapy depends largely on both the severity of the disease and the clinical indication. Patients may either be discharged to their homes without using a WCD or remain under observation in a hospital or in a skilled nursing facility [6]. In the former scenario, WCD therapy can be considered an add-on treatment to optimal medical therapy (OMT), while WCD therapy may be considered an alternative to hospital stay in the latter scenarios.

For a long period of time, the only available studies were observational, providing conflicting data regarding, and yielding to draw firm conclusions on, the comparative effectiveness and safety of WCD therapy [[7], [8], [9]]. It took roughly two decades after the first approval of WCD therapy by the FDA for the first randomised controlled trial (RCT) to be published. In 2018, the Prevention of Early Sudden Death Trial (VEST) failed to show a reduction in SCD (primary endpoint) when compared to medical therapy alone in patients with a recent myocardial infarction (MI) and low ventricular ejection fraction (LVEF) of ≤35% [10]. VEST results were limited by poor compliance. This inconclusive evidence from both observational and recent randomised evidence was reflected in independently conducted systematic reviews (SR) and meta-analyses [4,[7], [8], [9]], but intensive marketing hampered the scientific debate regarding the appropriateness of the WCD [11,12].

In Austria, reimbursement of WCD therapy is restricted to centres with a strong focus on cardiology and in narrow indications being backed mainly by clinical plausibility. The question arose whether reimbursement of broader indications – including the add-on use to OMT in post MI patients with LVEF of ≤35% – is backed by scientific evidence. The article concisely describes the available body of evidence of WCD therapy based on a recent second update report of the Austrian Institute for Health Technology Asessment (AIHTA), available elsewhere [13].

2. Methods

The study was undertaken in accordance with the PRISMA statement [14,15] and the EUnetHTA Core Model was used flexibly as a reporting standard [16].

2.1. Search strategy

Building on two HTA reports [4,7], we conducted an update systematic search in three medical databases (Medline via Ovid, Embase, The Cochrane Library; time period: between 8/2018–5/2022). Additionally, we hand-searched the reference list of two recent SRs [9,17] to strengthen the systematic search. Documentation of the systematic search can be found elsewhere [13].

2.2. Selection criteria, study selection, and selected outcomes

We included studies evaluating the clinical effectiveness and safety of the WCD as an add-on to OMT or as a replacement of hospital observation in patients at risk for sudden cardiac death. With regard to specific clinical indications, no filter applied. We included RCTs, prospective comparative studies and prospective uncontrolled studies with at least 100 patients. Table 1 shows detailed inclusion and exclusion criteria. (See Fig. 1.)

Table 1.

Inclusion and exclusion criteria.

	Inclusion Criteria	Exclusion criteria
Population	Population: adults ≥18 years of age with increased risk of sudden cardiac death	Younger patients (<18 years of age), Healthy individuals No filter applied with regard to potential indications for WCD use
Intervention	WCD (as add-on or replacement): LifeVest® (WCD 4000) from ZOLL (Lifecor) Medical Corporation, Pittsburgh, PA, USA [68] ASSURE™ Wearable Cardioverter Defibrillator (WCD) from Kestra Medical Technologies, Inc. [5]	All other therapies
Control	Hospital observation, Optimal Medical Therapy (OMT)	No restriction
Outcomes
Efficacy	Critical: •Mortality (all-cause mortality, disease-specific mortality) •Important: •Health-related Quality of Life measured using a validated instrument •Hospitalisation rate •Patient satisfaction measured using a validated instrument •Compliance •Relevant surrogate endpoints related to the functional performance of the device (appropriate shocks, shock success, withheld shocks)	All other outcomes
Safety	•Adverse events: device related and patient related (frequency of AEs, frequency of discontinuation due to AEs, frequency of unexpected AEs), •Serious Adverse events: device related and patient related (frequency of SAEs, frequency of SAEs leading to death).	All other outcomes
Study design	Randomised and prospective observational studies with a control group Observational prospective studies and register studies with at least 100 patients	Retrospective observational studies, prospective uncontrolled studies with <100 patients
Publication period	08/2018–05/2022

Abbreviations: AE – adverse events; CABG – coronary artery bypass graft; CE – Conformité Européenne; d – day(s); ICD – implantable cardioverter-defibrillator; MRI – Magnetic resonance imaging; OMT – optimal medical therapy; PCI – percutaneous coronary intervention; SAE – serious adverse events; SCD – sudden cardiac death; WCD – wearable cardioverter-defibrillator.

Fig. 1.

PRISMA flow chart: study selection.

In accordance with Grading of Recommendations, Assessment, Development, and Evaluation (GRADE), we rated the relevance of outcomes [18]. Mortality (all-cause or due to arrhythmic reasons) was rated as critical for decision-making. Quality of life, patient satisfaction, compliance with the WCD and the hospitalisation rate were considered important, but are intermediary endpoints for a treatment success. Outcomes related to the functional performance of the device (i.e., appropriate shocks, shock success and withheld shocks) were considered important, although being surrogate endpoints for a potential clinical benefit. These endpoints are descriptive in nature. With regard to safety, inappropriate shocks and further adverse events (AE) and serious adverse events (SAE) were considered critical for decision-making.

2.3. Data extraction, quality appraisal, and analysis

One researcher (G.G.) systematically extracted data from eligible newly identified studies by means of pre-defined data-extraction tables. Another researcher (B.W.) verified the extracted data. Risk of bias (RoB) of included studies was assessed using established RoB tools: Depending on the study designs of studies, the Cochrane RoB tool v.2 [19], the ROBINS-I tool [20] and the Institute of Health Economics (IHE-20) checklist [21] were utilised. We retrieved data-extraction and RoB tables regarding studies that were either included in the update-report 2019 [4] or in the original EUnetHTA report 2017 [7]. Evidence was qualitatively synthesised using GRADE [18]. In light of observational studies having included mixed populations, it was not possible to synthesise this evidence according to specific indications. The distribution of enrolled patients according to clinical indications/ aetiologies was presented using a modified structure utilised in another SR [9]. Inferential statistical analysis was not conducted in the absence of RCTs of high quality.

3. Results

3.1. Search results

The systematic update search resulted in 469 records after de-duplication. Of these, 397 hits were excluded based on screening of abstracts. Of 72 acquired full-text articles, 12 articles consisting of seven new studies and further five publications relating to previously identified studies were eligible for the qualitative synthesis of evidence. Together with five studies meeting our inclusion criteria from previous HTA reports, 12 studies were included in the evidence synthesis. Hence, the current best available evidence consists of one RCT [10,22] and eleven observational studies [[23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37]].

3.2. Characteristics of the available RCT

The RCT [10,22] included sites in the United States and Europe (Poland, Germany, Hungary). The study was funded by Zoll Medical Corp. and the National Institutes of Health (NIH) /National Heart Lung and Blood Institute (NHLBI). VEST enrolled 2348 patients hospitalised with an acute MI and LVEF≤35% who were randomised in a 2:1 fashion [10]. Of these, 46 participants were excluded because of irregularities found at one of the sites. As a result, 2302 patients were analysed within VEST; 1524 in the device group and 778 patients in the control group. Patients in the intervention/ device group received a WCD and OMT and the control group received OMT alone. Although cross-overs were considered as a protocol deviation, some 20 patients (2.6%) in the control group received a WCD and further 43 patients (2.8%) never wore the device after randomization within the device group [10]. Co-morbidities and baseline characteristics were balanced between intervention and control group in the VEST trial.

On average, VEST followed the enrolled patients for 84.3 days (SD: 15.6). 22 patients were lost to follow-up: 10 (0.7%) patients in the device group and 12 (1.5%) patients in the control group [10,22].

3.3. Characteristics of included observational studies

Of eleven included observational studies [[23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37]], seven and three studies were prospective registries [24,[27], [28], [29], [30], [31], [32],[34], [35], [36], [37]] or case series [25,26,33], respectively. Another study was a sub-study of an ongoing prospective cohort study [23]. Nine studies were conducted in Europe [[25], [26], [27],[32], [33], [34], [35], [36], [37]] and one study was conducted in the USA [24,[28], [29], [30], [31]]. The remaining study [23] enrolled patients from sites in the USA and Europe.

Most of the studies were uncontrolled observational studies, except for one registry study [37] that investigated the quality of life with the WCD. In total, 5345 patients (range: 102–2000) were enrolled by the included studies. Of these, 5307 patients received a WCD additionally to standard care, whilst 38 patients received standard care solely [[23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37]].

The range of follow-up period was six weeks [37] to 36.2 months [34]. Loss to follow-up was insufficiently reported by four studies [23,24,[28], [29], [30], [31], [32],35] and ranged from 0% to 18% in the remaining studies [[25], [26], [27],33,34,36,37].

3.4. Quality of Studies and Certainty of evidence

The overall RoB of the RCT [10,22] was high due to cross-overs and low compliance (deviations from intended intervention) and some concerns were further found with regard to bias in measurement of outcomes and selection of reported results (see Fig. 2). The RoB with regard to the per-protocol effect analysed in post-hoc analyses [22] is further substantially increased because of missing data and difficulties with regard to assessing the effect of assignment to intervention arising from the post-hoc design. (See Fig. 3, Fig. 4.)

Fig. 2.

Risk of bias of the randomised controlled trial Full risk of bias assessment can be found elsewhere [13]. Abbreviations: D – domain.

Fig. 3.

Risk of bias of non-randomised controlled trials Full risk of bias assessment can be found elsewhere [13]. Abbreviation: D - domain.

Fig. 4.

Risk of bias of uncontrolled observational studies. Abbreviations: D – domain; Q – question. D1: Study objective. Q1: Was the hypothesis/aim/objective of the study clearly stated?. D2: Study design. Q2: Was the study conducted prospectively?. Q3: Were the cases collected in more than one centre?. Q4: Were patients recruited consecutively?. D3: Study population. Q5: Were the characteristics of the patients included in the study described?. Q6: Were the eligibility criteria (i.e. inclusion and exclusion criteria) for entry into the study clearly stated?. Q7: Did patients enter the study at a similar point in the disease?. D4: Intervention and cointervention. Q8: Was the intervention of interest clearly described?. Q9: Were additional interventions (co- interventions) clearly described?. D5: Outcome measures. Q10: Were relevant outcome measures established a priori?. Q11: Were outcome assessors blinded to the intervention that patients received?. Q12: Were the relevant outcomes measured using appropriate objective/subjective methods?. Q13: Were the relevant outcome measures made before and after the intervention?. D6: Statistical Analysis. Q14: Were the statistical tests used to assess the relevant outcomes appropriate?. D7: Results and Conclusions. Q15: Was follow-up long enough for important events and outcomes to occur?. Q16: Were losses to follow-up reported?. Q17: Did the study provide estimates of random variability in the data analysis of relevant outcomes?. Q18: Were the adverse events reported?. Q19: Were the conclusions of the study supported by results?. D8: Competing interests and sources of support. Q20: Were both competing interests and sources of support for the study reported?

RoB of three out of eleven observational studies was moderate [32,34,36]. The remaining eight studies have a high [[23], [24], [25], [26],[28], [29], [30], [31],33,35,36] or critical [37] RoB, mainly due to study design features, study population, and missing data. Potential conflict of interests (COI) in form of industry sponsoring were reported in six [23,24,[27], [28], [29], [30], [31],[35], [36], [37]] studies.

According to GRADE, the certainty of RCT and observational evidence was low and very low, respectively. Reasons were the aforementioned RoB within included studies and statistical imprecision present for results of some outcomes (see Table 3).

Table 3.

GRADE Summary of findings table: Effectiveness and safety of the wearable cardioverter defibrillator in patients at risk of sudden cardiac death.

Outcomes	Indication	Anticipated effects	№ of analysed pts (studies)	Certainty of the evidence (GRADE)
Comparative effectiveness (RCT evidence)
Mortality	Post-MI	Arrhythmic mortality: 25/ 1524 (1.6%) vs. 19/ 778 (2.4%), p = 0.18 All-cause mortality: 48/1524 (3.1%) vs. 38/778 (4.9%), p = 0.04	2348 (1 RCT)	⨁⨁◯◯ LOWa,b,c
QoL	Post-MI	–	(0 studies)	–
Hospitalisation	Post-MI	31.2% vs. 32.5% (p-value = 0.5)	2348 (1 RCT)	⨁⨁⨁◯ MODERATEa
Satisfaction	Post-MI	–	(0 studies)	–
Compliance	Post-MI	Mean daily wear time, in days: 14 (SD: ± 9.3)	2348 (1 RCT)	⨁⨁⨁⨁ High
Surrogate endpoints	Post-MI	Appropriate shocks: 1.3%/ Shock success: NR withheld shocks: 4.5%	2348 (1 RCT)	⨁⨁⨁◯ MODERATEa
Effectiveness (observational evidence)
Mortality	Mixed	Arrhythmic mortality (2 studies): 0% All-cause mortality (9 studies): 0–5.2%	4992 (9 studies)	⨁◯◯◯ VERY LOW d
QoL	Mixed	statistical association between WCD and baseline anxiety (1 controlled study): 41 ± 11 vs. 39 ± 13 (p = 0.22; higher score indicates higher anxiety) statistically significant improvements in all Kansas City Cardiomyopathy Questionnaire subscales (1 before-after study; baseline to day 90)	310 (2 studies)	⨁◯◯◯ VERY LOW d
Hospitalisation	Mixed	13/102 (12.7%) in 1 prospective case series 102/153 (67%4) in 1 registry	255 (2 studies)	⨁◯◯◯ VERY LOW d
Satisfaction	Mixed	–	–	–
Compliance	Mixed	20–23.5 h per daya	5092 (10 studies)	⨁⨁◯◯ LOW
Surrogate endpoints	Mixed	N of pts. with at least one appropriate shock: 1.1%–4.8% (range of enrolled patients across studies: 102–2000) First shock success (3 studies): 100% Withheld shocks (2 studies): 90 events in 22 pts. (1.1%) and 47 events in 22 pts. (2.8%)	4992 (9 studies)	⨁⨁◯◯ LOW
Safety (RCT evidence)
(Serious) adverse events (SAE/AE)	Post-MI	SAE: Inappropriate shocks: 9/1524 (0.6%), SADEb: 4/1524 (0.2%) AE: Rash: 184 (13.0%) vs. 27 (3.8%), p < 0.001 Itch: 205 (14.5%) vs. 22 (3.1%), p < 0.001	2348 (1 RCT)	⨁⨁⨁◯ MODERATE a
Safety (observational evidence)
(Serious) adverse events (SAE/AE)	Mixed	SAE: Inappropriate shocks: 0–2% AE (2 studies): 4/448 pts. (0.9%) with dematitis, 1/448 pt. (0.2%) with pressure mark in one study; 2 pts. (2%) were allergic to nickel and 58 (57%) false alarms in another study	5092 (10 studies)	⨁⨁◯◯ LOW

Full GRADE evidence profile can be found elsewhere [13].

Abbreviations: AE – adverse events; GRADE – grading of recommendations assessment development and evaluation; hrs – hours; MI – myocardial infarction; QoL – quality of life; RCT – randomised controlled trial; SADE – serious adverse device effect; SAE – serious adverse events; SD – standard deviation.

Explanations.

a. The RCT was judged to be at high risk of bias through the Cochrane Risk of Bias tool due to poor compliance (and selective outcome reporting) that could have influenced the comparative effect estimates for effectiveness outcomes and the estimated proportions of adverse events for safety outcomes.

b. In the study occurred few events leading to a wide CI around the estimate of the effect estimate.

c. The endpoint “death from any-cause” was set as a secondary outcome in the included RCT. In addition, the study did not statistically correct the analysis for multiple testing.

d. Selection bias may be the most significant source of bias in observational studies. In addition, reporting on milder Aes was sparse and patients did not enter the study at the same point of disease. Also mixed populations were included, being heterogeneous patient populations.

^aMeasured as median or mean.

^bSAEs related or potentially related to the WCD.

3.5. Indications across studies

While a narrow indication was selected in VEST [10,22], the remaining observational studies [[23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37]] enrolled mixed patient populations. Numerically, ICM and NICM were the most frequent patient populations across included studies. Reporting on specific indications within observational studies was sparse and was often reported in an aggregated way.

3.5.1. Ischemic cardiomyopathy (ICM)

The RCT [10,22] enrolled patients with ischemic cardiomyopathies: Patients hospitalised with acute MI and with LVEF of ≤35% (assessed ≥8 h after MI) were enrolled within this study. Eleven observational studies enrolled ICM patients as one part of their study cohorts. The percentage of ICM patients ranged from 20% to up to 82% within these studies, with a range of enrolled patients receiving a WCD of 85 to 2000. ICM patients within observational studies were often broadly defined and differences with regard to specific subgroups within this patient group persists.

3.5.2. Non-ischemic cardiomyopathy (NICM)

Ten observational studies also enrolled NICM patients. This broad patient group accounted for 31% to up to 54% of enrolled patients (range of patients receiving a WCD: 85–2000). More specifically, some four studies [[33], [34], [35],37] enrolled NICM patients with LVEF ≤35% and another study reported that enrolled NICM patients had an LVEF of 35% within 90 days of the start of heart failure therapy and/or documented VT/VF. Most studies enrolling, among others, NICM patients did not adequately report on specific NICM subgroups, with a few studies reporting having enrolled idiopathic/ dilated cardiomyopathy [25,26,33], peripartum cardiomyopathy [26,27,33], myocarditis [[25], [26], [27],[32], [33], [34]], or Takotsubo cardiomyopathy [[25], [26], [27],33].

3.5.3. ICD explantation

Five studies [27,[32], [33], [34],37] reported on having enrolled, among others, patients with an ICD explantation. This patient group accounted for 9% to 14% of all enrolled patients within these studies (range of enrolled patients receiving a WCD: 85–1164).

3.5.4. Other indications

Patients waiting for heart transplantation (without ICD) were enrolled in one study [27] and accounted for 8% of 1164 enrolled patients. Patients during diagnostic wrap-up were enrolled in two studies: In one study, channelopathies/ congenital heart diseases (not further specified) accounted for 4% of 102 enrolled patients [26]. In another study, 13% of 781 enrolled patients used the WCD for “other risk stratification” [36]. Further, patients with an acute infection attributed to 18 patients (4%) of one study [32] and 25 patients (25%) of another study [26]. These acute infections were device-related and systemic (delaying ICD implantation), respectively. A further sub-population were patients with a delayed ICD implantation due to comorbidities or other reasons in one study with 54 enrolled patients, accounting for 12% in this sample. Patients with documented VT events prior or post VT ablation (bridge to ablation) were enrolled in one study, accounting for 11 patients (2%) of enrolled patients [32].

3.6. Primary vs secondary prevention

Most of the included studies failed to report whether the use of the WCD was used for primary or secondary prevention of SCD, except for two studies. In VEST, WCD therapy was used in a narrow patient population (n = 1524) for primary prevention of SCD [10,22]. In the Austrian registry [32] enrolling 448 patients, WCD was used for primary (52%) and secondary (48%) prevention of SCD.

In Table 2, one can find an overview of indications/ patient populations and use cases across included studies. (See Table 4, Table 5, Table 6.)

Table 2.

Overview of indications for WCD therapy in clinical studies.

Study	Study design	N of pts⁎	Primary vs. secondary prevention	Duration of use, in days⁎⁎	Daily use, in hours	ICD Explantation	ICM	NICM	NICM Subgroups					WL for heart transplantation	During diagnostic wrap up (e.g., channelopathies)	Acute infection	Bridge to ablation	Delayed ICD Implantation	Other/ data not available
									Idiopathic/ DCM	PPCM	Myocarditis	TCM	Others
VEST [10,22]	RCT	1524	100% vs. 0%	84.3	14,10	0	1524 (100%)	0	0	0	0	0	0	0	0	0	0	0	0
Röger, 2018 [33]	OS	105	NR	68.8 ± 50.4	21.5 ± 3.5	15 (14%)	43 (41%)	41 (39%)	38	1	0	2	0	0	0	0	0	0	6 (6%)
Erath, 2018 [25]	OS	130	NR	42 (1–166)	23,00	0	84 (65%)	46 (35%)	25	NR	5	10	5	0	0	0	0	0	0
Erath, 2017 [26]	OS	102	NR	42	20,00	0	27 (27%)	42 (41%)	33	0	9	0	0	0	4 (4%)	25 (25%)	0	0	4 (4%)
WEARIT-II [24,[28], [29], [30], [31]]	OS	2000	NR	90	22,5	0	805 (40%)	927 (46%)	NR	NR	NR	NR	NR	0	0	0	0	0	268 (13%)
WEARIT-FR [27]	OS	1164	NR	62 (37–97)	23.4	119 (10%)	950 (82%)	0	0	0	0	0	0	88 (8%)	0	0	0	0	7 (1%)
WEARIT-II-EU [36]	OS	781	NR	75 ± 47.7	20.3⁎	0	200 (26%)	249 (32%)	NR	NR	NR	NR	NR	0	102 (13%)	0	0	0	230 (29%)a
Odeneg, 2019 [32]	OS	448	52% vs. 48%	54 (1, 436)	23.5	46 (10%)	88 (20%)	139 (31%)b	NR	NR	45	NR	NR	0	0	18 (4%)	11 (2%)	54 (12%)	92 (21%
Rosenkaimer, 2020 [34]	OS	153	NR	65.1 ± 42	21.45 ± 3.52	16 (10%)	56 (37%)	78 (51%)	NR	NR	8	NR	NR	0	0	0	0	0	3 (2%)
Sinha, 2021 [35]	OS	120	NR	48 (37–62)	22.9 (21.2–23.4)	0	46 (38%)	69 (58%)	NR	NR	NR	NR	NR	0	0	0	0	0	5 (4%)
Weiss, 2019 [37]	cOS	85	NR	59 (40–96)	20 ± 5	8 (9%)	20 (24%)	57 (67%)	NR	NR	NR	NR	NR	0	0	0	0	0	0
Burch 2021 [23]	OS	210	NR	NR	NR	0	91 (43.3%)	113 (53.8%)	NR	NR	NR	NR	NR	0	0	0	0	0	6 (2.9%)

Structure of table informed by [9].

Abbreviations: cOS – comparative observational study; DCM – dilated cardiomyopathy; ICD – implantable cardioverter defibrillator; ICM – ischemic cardiomyopathy; NICM – non-ischemic cardiomyopathy; NR – not reported; OS – observational study; PCM – peripartum cardiomyopathy; pts – patients; TCM – tachymyopathy; WL – waiting list.

^⁎N of patients refers to patents receiving a WCD.

^⁎⁎All values as median (IQR: 1st Quartile-3rd Quartile or range: lowest value, highest value) or Mean ± SD.

^aRecent onset/ impairment of heart failure.

^bSevere NICM with LVEF<35%.

Table 4.

Wearable cardioverter defibrillator for primary or secondary prevention of sudden cardiac death: Results from randomised controlled trials.

Study name	Vest Prevention of Early Sudden Death Trial (VEST)
Author, year	Olgin, 2018 (primary analysis) [10]	Olgin, 2020 (secondary analysis) [22]
Source	AGENAS/ LBI-HTA Report [4]
Study characteristics
Study registration number	NCT01446965
Countries of recruitment	USA., Poland, Germany, and Hungarya
Sponsor	National Institutes of Health (NIH) / National Heart Lung and Blood Institute (NHLBI)b and Zoll Medical
Comparator	Guideline-directed medical therapy
Study design	Multicenter, randomised, controlled trial
Methods	Randomization: 2:1 fashion, Intention to treat analysis, further use of monitoring data of the LifeVEST	Secondary analysis: as-treated: event rates per person-month; sensitivity analysis based on a) effect-cause and b) confounding by propensity to adhere bias. Per-protocol analysis based on Kaplan Meier statisticsc Subgroup analysis:
Study duration (start and completion date)	07/2008–04/2017
Objectives	To determine the efficacy of a Wearable Cardioverter–Defibrillator during the period before ICDs are indicated in patients who have had a myocardial infarction and have a reduced ejection fraction.	To explore the impact of WCD compliance and hospitalizations on outcomes, including additional on-treatment analyses and effect modification analyses to determine factors that identify those most likely to benefit from the WCD.
Patients characteristics
Number of pts	2302d (1524e device group and 778f control group).
Age in yrs (range) ± SD	Device group, mean ± SD: 60.9 ± 12.6. Control group, mean ± SD: 61.4 ± 12.3.
Sex (female/male)	Device group: 27%/73%. Control group: 25%/75%g
EF in % (range) ± SD	Device group, mean ± SD: 28.2 ± 6.1. Control group: 28.2 ± 5.8.
Inclusion criteria	Patients who had been hospitalised with an acute myocardial infarction and who had an ejection fraction of ≤35% were enrolled within 7 days after hospital discharge.
Exclusion criteria	Patients were excluded if they had an ICD or unipolar pacemaker, had clinically significant valve disease, were undergoing long-term hemodialysis, or had a chest circumference that was too small or too large to accommodate the Wearable Cardioverter-Defibrillator. Patients were also excluded if they were pregnant or had been discharged to a nursing facility with an anticipated stay of >7 days.
Follow-up time in months (range), mean ± SD	Mean ± SD: 84.3 ± 15.6 days.
Loss to follow-up, n (%)	68 pts. (2.9%)h	–
Diagnosis	Patients with acute myocardial infarction and who had an ejection fraction of ≤35%	–
Previous treatments	Previous CABG Device group: 133/1521 (8.7), Control group: 70/776 (9.0); Previous PCI Device group: 374/1520 (24.6), Control group: 202/776 (26.0).	–
Outcomes: Clinical effectiveness
Mortality, n (%) •All-cause mortalityi	Intention to Treat: Device group: 48 (3.1); control group: 38 (4.9). Relative risk (RR): 0.64 (95% CI, 0.43–0.98); p = 0.04.	As-treated analysis: 12 (2420 person months) vs. 71 (3724 person months); adjusted RR: 0.26 (p < 0.0005)j Per protocol analysis HR: 0.25 (95% CI, 0.13–0.48); p < 0.001
•Disease-specific mortalityk	Device group: 25 (1.6); Control group: 19 (2.4). RR: 0.67 (95% CI, 0.37–1.21); p = 0.18.	As treated analysis: 9 (2420 person months) vs. 32 (3724 person months); RR: 0.43 (p = 0.026)l Per-protocol analysis: HR: 0.38 (95% CI, 0.17–0.86); p = 0.02
Appropriate shocks	Device group: 20 (1.3%).m Control group: 1 (0.1%).nP = 0.008
Withheld shockso	Device group: 69 (4.5%).p Control group: 1 (0.1%).q	–
First shock success (%)	NA	–
Health-Related Quality of Life (HRQL)	NAr
Hospitalisation rate	Rehospitalisation by any cause, n (%): Device group: 475 (31.2), Control group: 253 (32.5). RR: 0.96 (95% CI, 0.85–1.09). P = 0.51.	–
Satisfaction with technology	NA	–
Compliance/ patient adherence •WCD wear-time in days (range), median	NAs	–
•WCD daily use in hours (range), median	Device group,t mean ± SD: 14.0 ± 9.3 [Median (IQR): 18.0 (3.8–22.7)]; Control group,u mean ± SD: 0.4 ± 2.7 [Median (IQR): 0.0 (0.0–0.0)].
OUTCOMES: SAFETY
AEs in n (%) of pts: •Skin rash and itching	Rash on torso, n (%): Device group: 184 (13.0%), Control group: 27 (3.8%). RR: 3.42 (95% CI, 2.31–5.08), p < 0.001.v Itch on torso, n (%): Device group: 205 (14.5%), Control group: 22 (3.1%). RR: 4.68 (95% CI, 3.04–7.20), p < 0.001.w	–
•False alarms	NAx
Frequency of discontinuation due to AEs in n (%) of pts: •Discontinuation due to comfort and lifestyle issues	NA	–
Frequency of unexpected AEs in n (%) of pts	NA	–
Hospitalisation related to WCD use	3/1524 (0.2%)y	–
Serious Adverse Events (SAEs), n (%) •Inappropriate shocks	9 (0.6%) [7 pts. had 1 shock; 2 pts. had ≥2 shocks]	–
•Unsuccessful shock	NAz
Frequency of SAEs leading to death in n (%) of pts	NAaa	–
Effect modifiers
Method of identifiying potential effect modifiers & Results	–	Backward stepwise deletion of potential predictors with P < 0.05 to select a parsimonious model No interaction was found. A trend for participants with a cardiac arrest (interaction P = 0.08), pulmonary edema (interaction P = 0.07), and Cr < 1.5 (interaction P = 0.06) toward lower mortality in the WCD group in the intention-to-treat analysis

Abbreviations: USA – United States of America; ICD(s) – implantable cardioverter-defibrillator(s); pt(s) – patient(s); yrs – years; SD – standard deviation; EF – ejection fraction; CABG – coronary artery bypass graft; PCI – percutaneous coronary intervention; NIH – National Institute of Health; RR – relative risk; CI – confidence intervals; VT – ventricular tachycardia; VF – ventricular fibrillation; NA – not available; HRQL – Health-Related Quality of Life; WCD – Wearable Cardioverter-Defibrillator; IQR – interquartile range; AEs – adverse events; SAEs – serious adverse events.

^a76 sites in the United States, 24 in Poland, 6 in Germany, and 2 in Hungary.

^bNIH/NHLBI stopped funding the study.

^cKaplan-Meier plots for time from randomization to death or censoring for implantable cardioverter-defibrillator (ICD) implant, by treatment assignment, with follow-up and events censored in the WCD group at the last day the WCD was worn (defined as all subsequent days with 0 h wear-time).

^d2348 patients were initially randomised. 46 participants at one U.S.A. site were excluded after randomization, owing to irregularities found by the institutional review board at that site; therefore, 2302 participants were included in the analyses.

^e43/1524 (2.8%) patients in the device group never wore the WCD after randomization.

^f20/778 (2.6%) patients in the control group wore the WCD (2.6%) outside the protocol. Cross-overs were considered to be a protocol deviation.

^gFrom the Table 1, Table 3 pts. from the device group and 6 pts. from the control group were missing in the male/female data.

^h46 (2%) from the U.S.A. site excluded; 10/1524 pts. (0.7%) in the device group; 12/778 (1.5%) in the control group.

ⁱAll-cause mortality was a secondary outcome.

^jAdjusted for diabetes and PCI, the only variables that remained after backwards stepwise variable deletion.

^kDisease-specific mortality was the primary outcome.

^lAdjusted for diabetes and PCI, the only variables that remained after backwards stepwise variable deletion.

^m13 pts. had 1 shock; 7 pts. had ≥2 shocks.

ⁿThis patient had ≥2 shocks.

^oDue to patients using the response button to delay therapy.

^p1 shock 43 (2.8%), 2–5 shocks 11 (0.7%), ≥5 shocks 15 (1.0%).

^q1 shock (0.1%).

^rQuality of life was a planned secondary outcome in the study protocol, but it was not reported in neither of the available publications of the VEST trial.

^sOver the course of the 90 days, the proportion of participants who wore the WCD on a given day fell from 80.8% (CI: 78.8–82.8) just after randomization to 41.3% (CI 37.5, 44.9) at 90 days.

^t1481/1524 (97.2%) worn the device.

^u20/778 (2.6%) worn the device.

^vRash in any location, n (%): Device group: 216 (15.3%), Control group: 50 (7.1%), p < 0.001.

^wItch in any location, n (%): Device group: 243 (17.2%), Control group: 45 (6.4%), p < 0.001.

^xAmong 41 participants with an alarm indicating asystole, 6 events (all in the device group) were adjudicated as having had a true asystole event.

^yTwo due to aborted shocks and one due to an inappropriate shock.

^zThe shock delivered sometime caused a cardioversion into complex and repeated other cardiac conduction problems which the WCD was not programmed to deal with.

^aaOne patient died while he was wearing the device. The authors state that this death could be possibly related to the WCD use. The authors also state that it was deemed likely to not be an arrhythmic death.

Table 5.

Wearable cardioverter defibrillator for primary or secondary prevention of sudden cardiac death: Results from observational studies (Part 1).

First author, year	Röger 2018 [33]	Erath 2018 [25]	Erath 2017 [26]
Source	AGENAS/ LBI-HTA Report [4]
Study name	NA	NA	NA
Study registration number	NA	NA	NA
Country/ies of recruitment	GER	GER	GER
Sponsor	NA	NA	NA
Intervention	WCD (+ SoC)	WCD (+ SoC)	WCD (+ SoC)
Comparator	None	None	None
Study design	Prospective case series	Prospective case series	Prospective case series
Study duration (start and completion date)	4/2012–9/2016	NA	2012–2015
Objectives	To determine the value of the WCD for therapy optimization of heart failure pts.	To evaluate the clinical development of tachymyopathy pts. protected with a WCD in a single-center non-randomised pt. cohort.	To evaluate the efficacy, safety, and compliance of/to WCD use and subsequent medium-term outcome of pts. in a single-center.
Number of pts	114a	130b	102c
Age in yrs [mean (range) ± SD; median (IQR)]	All pts. (n. 105): Median (IQR): 60 (26–79).	All pts., mean ± SD: 58 ± 16 • Cases: 62 ± 9 • Controls: 58 ± 16 (ns)	All pts., mean ± SD: 59 ± 11.
Sex: female / male	All pts.: 22% / 78%	All pts.: 22% / 78% • Cases: 20% / 80% • Controls: 22% / 78%	All pts.: 28% / 72%.
EF in % [mean (range) ± SD; median (IQR)]	mean ± SD: 28.3 ± 9.8.	All pts., mean ± SD: 28 ± 11 • Cases: 26 ± 6 • Controls: 29 ± 12 (ns)	All pts., mean ± SD: 30 ± 11.
Inclusion criteria	All consecutive pts. receiving a WCD at a tertiary care University Center	• Cases: consecutive pts. with clinically suspect tachymyopathy and high risk of ventricular arrhythmias • Controls: consecutive pts. with high risk of ventricular arrhythmias and another option for use of vests	Pts at high risk of VT/VF
Exclusion criteria	NA	NA	NA
Follow-up time in months [mean (range) ± SD; median (IQR)]	Mean ± SD: 18.6 ± 12.3	12 months of follow-up (1, 3 and 12 months)	Mean ± SD: 11 ± 8
Loss to follow-up, n (%)	9 (8)	0 (0)	0 (0)
Diagnosis	Newly diagnosed ICM, LVEF ≤35% (n = 43); Newly diagnosed NICM, LVEF ≤35% (n = 41); ICD explant (n = 15); Newly diagnosed CMP (n = 6).	Pts with symptomatic congestive HF with impaired LV function	Newly diagnosed HF
Previous treatments	Medications (betablocker, ACE-I/ARB, MRA, ARNI, procoralan, diuretic, amiodarone)	Medications (betablocker, amiodarone, ACE inhibitors/ARB, aldosterone antagonists, diuretics, statin, NOAC, VKA)	Medications (β-blocker, amiodarone)
Mortality, n (%) •All-cause mortality	3 (3%)	No deaths during the use of vest.d	No deaths during the use of vest.e
•Disease-specific mortality	0 (0)	NA	NAf
•Appropriate shocks	5 (4.8%)	2 pt. in the control group (2%)	4 pts. (4%)g
•Withheld shocksh	NA	NA	NA
First shock success (%)	100%	NA	NA
Health-Related Quality of Life	NA	NA	NA
Hospitalisation rate	NA	NA	13 pts. hospitalised pts. due to cardiac causes
Satisfaction with technology	NA	NA	NA
Compliance/ pt adherence •WCD wear-time in days [mean (range) ± SD; median (IQR)]	mean ± SD: 68.8 ± 50.4	All pts., median (IQR): 42 (1–166)	median: 54 days (1–166)
•WCD daily use in h/day [mean (range) ± SD; median (IQR)]	mean ± SD: 21.5 ± 3.5	All pts., mean: 23 h/day	23.0 h/day (7–24)
AEs in n (%) of pts: •Skin rash and itching	NA	NA	2 pts. (2%) are allergic to nickel
•False alarms	NA	NA	58 (57%)
Discontinuation due to comfort and lifestyle issues	8 pts. (7%)i	NA	NA
Serious Adverse Events (SAEs), n (%) •Inappropriate shocks	1 (1%) (ICM pt)	2 in the control group (2%)	2 (2%)
•Unsuccessful shock	NA	NA	NA
Frequency of SAEs leading to death in n (%) of pts	NA	NA	NA

Abbreviations: AEs – adverse events; CABG – coronary artery bypass graft; CMP – cardiomyopathy; DCM – dilated cardiomyopathy; EF – ejection fraction; FR – France; GER – Germany; HF – heart failure; ICD – implantable cardioverter-defibrillator; ICM – ischemic cardiomyopathy; INSERM – Institut national de la santé et de la recherche médicale; IQR – interquartile range; LV – left ventricular; LVEF – left ventricular ejection fraction; MI – myocardial infarction; NA – not available; NICM – non-ischemic cardiomyopathy; Ns – not significant; NYHA – New York Heart Association; OMT – optimal medical therapy; PCI – percutaneous coronary intervention; pt(s) – patient(s); SCA – sudden cardiac arrest; SCD – sudden cardiac death; SD – standard deviation; SoC – standard of care; VA – ventricular tachyarrhythmias; VF – ventricular fibrillation; VS – versus; VT – ventricular tachycardia; WCD – Wearable Cardioverter-Defibrillator; yrs – years.

^a8 patients returned their WCD during the first hours after initiation because of unwillingness or inability to handle it; one more patient was lost to follow up, leaving 105 patients considered for data analysis. 43/105 patients had ICM.

^b20 pts. in cases group, and 110 pts. in control group.

^cICM patients: 27/102.

^dDeaths after the use of the vest: All pts.: 5 (4%).

^e6 (6%) after the end of treatment with vest [of these, 2/27 (7%) were ICM pts]: 5 ICD and 1 not-ICD.

^fArrhythmic mortality after WCD therapy: 4 pts. (4%).

^gPatients were adequately shocked for ventricular fibrillation (seven episodes) or for ventricular tachycardia (one episode).

^hDue to patients using the response button to delay therapy.

ⁱThey returned their WCD during the first hours after initiation because of unwillingness or inability to handle it.

Table 6.

Wearable cardioverter defibrillator for primary or secondary prevention of sudden cardiac death: Results from observational studies (Part 2).

Study name or first author, year	WEARIT-II [24, 28–31]	WEARIT-FR [27]	WEARIT-EU [36]	Odeneg, 2019 [32]	Rosenkaimer, 2020 [34]	Sinha, 2021 [35]	Weiss, 2019 [37]	Burch, 2021 [23]
Source	EUnetHTA Report 2017 [7] and new publications	New Studies
Study name	WEARIT-II Registrya	WEARIT-FR	WEARIT-EU	AT-Registry	NA	NA	CREDb	HF-Optc
Study registration number	NA	NCT03319160	NA	NA	NA	NA	NA	NCT03016754
Country/ies of recruitment	USA	FR	GER	AT	GER	GER	GER	USA and GER
Sponsor	ZOLL Medical Corporation	ZOLL + INSERM	ZOLL Medical Corp	Public academic funding	Public academic funding	ZOLL Medical Corp	ZOLL Medical Corp	ZOLL Medical Corp
Intervention	WCD (+ SoC)	WCD (+ SoC)	WCD (+ SoC)	WCD + interactive nurse-based training (+ SoC)	WCD (+ SoC)	WCD (+ SoC)	WCD (+ SoC)	WCD (+ SoC)
Comparator	None	None	None	None	None	None	SoC	None
Study design	Multi-centre, prospective register	Multi-centre, prospective register	Multi-centre, prospective register	Multi-centre, prospective register	Prospective registry study	Prospective register	Multi-centre, prospective register	Prospective sub study of a prospective cohort study
Study duration (start and completion date)	08/2011–02/2014	05/2014–12/2016 and 2017–2018	01/2014–09/2015	2010–2016	2012–2019	2012–2017	NA	2017–2022
Objectives	1. Characterise pts. currently prescribed with WCD. 2. Assess the risk for sustained VT events among WCD pts. by disease aetiology. 3. Identify the rate of EF improvement and the need for subsequent ICD implantation. Post-Hoc analyses: •Describe 1-year FU data •Analyze extended use (>90 days with WCD) and •describe age differences with regard to WCD use	To evaluate contemporary real-world data on WCD use in France, not only in terms of effectiveness and safety but also compliance and acceptability.	To analyze the clinical outcome and mode of death after WCD prescription	To provide real-world data on patients receiving this therapy in a nurse-based wearable cardioverter-defibrillator training programme.	To assess long-term all-cause mortality and 3-year survival of patients with or without ventricular tachyarrhythmias during WCD use and subsequent device implantation	To investigate WCD use in community-based acute care centers	To prospectively compare baseline characteristics and clinical outcome of patients with a WCD to those without a WCD in order to provide insight on decision criteria of physicians.	To examine the change in patient-reported outcomes in newly diagnosed patients with heart failure and reduced ejection fraction (HFrEF) prescribed a wearable cardioverter defibrillator
Number of pts	2000d Extended use in 1019 pts	1164	781	448	153	120	123 (85 vs. 38)	210
Age in yrs [mean (range) ± SD; median (IQR)]	All pts., median (IQR): 62 (16). ICM pts., median (IQR): 65 (14).	60 ± 12	59.3 ± 13.4	59 ± 14	60 ± 14	66 (56–75)	56 ± 13 vs. 64 ± 14; s. s. with p < 0.05	58 (SD: 13.6)
Female sex, n (%)	All pts.: 30% /70%. ICM pts.: 23% / 77%.	183 (16)	182 (23.3)	107 (24)	35 (23)	25 (21)	18 (20) vs. 13 (34); diff. n. s. (p = 0.18)	54 (25.7)
LVEF in % [mean (range) ± SD; median (IQR)]	All pts., median (IQR): 25 (10). ICM pts., median (IQR): 26 (15).	27 ± 9	26.9 ± 10.3 LVEF ≤35%: 700 (90)	33 ± 15	28.61 ± 10.15	26 (20−30)	26 ± 8 vs. 25 ± 7; diff. n. s. (p = 0.73)	23 (SD: 6.9)
Inclusion criteria	Low EF and high risk of SCA post MI or post coronary revascularization or new onset nonischaemic DCM or high risk for SCA until stabilisation or inherited or congenial heart disease	Patients with a prescribed WCD according to the criteria for WCD prescription in FR	WCD prescription	All patients prescribed with a WCD	All patients receiving a WCD	Patients with cardiomyopathy LVEF ≤35%.	All patients receiving a WCD	Adults hospitalised for new-onset heart failure, with ischemic or nonischemic cardiomyopathy, and prescribed a wearable cardioverter defibrillator within 10 days post discharge were approached for inclusion
Exclusion criteria	NA	NA	NA	NA	NA	NA	NA	patients with an active unipolar pacemaker, a first hospitalisation for heart failure that occurred >30 days before enrollment, and patients with a psychological or physical condition that would inhibit interaction with the wearable defibrillator
Follow-up time in months [mean (range) ± SD; median (IQR)]	Original study: Median (IQR): 3.0 (2.1)e Post-hoc: Up to one year	NA	12 months	NA	36.2 ± 15.6	90 days (measure of central tendency: NA)	6 weeks	180 days
Loss to follow-up, n (%)	NA 1-year FU: 148 (7)	7 (0.6)	7 (9)	NA	4 (2.6)	NA	23 (18)	NA
Previous treatments	NA Beta-blockers: 1730 (87%) ACE-I/ARBs: 1482 (74%) Amiodarone: 259 (13%)	Beta-blockers: 1038 (89%) Amiodarone: 189 (16%) ACE-I/ARBs: 1004 (86%)	NA	NA	Former CIED explanted: 7 (5) History of CABG: 11 ± 7	Beta blocker: 111 (92) ACE inhibitor: 96 (80) Diuretic: 103 (86) Digitalis: 2 (2)	No stat. Significant diff. in previous treatment	β-Blockers: 198 (94.3) ACE-I/ARB: 152 (72.4) Aldosterone antagonists: 100 (47.6)
Clinical effectiveness
Mortality, n (%)
All-cause mortality	3 (0.2)f 1-year FU: 70 (4%)	24 (2.1%)	40 (5.2)	4 (1)	4-year: 15 (1) During WCD use: 2 (1)	0 (0)	NA	NA
Disease-specific mortality	0 (0)	NA	NA	NA	NA	NA	NA	NA
Appropriate shocks	30 events/22 pts. (1.1) 1-year FU: NA	19 events/ 18 pts. (1.6%)	13 events/ 10 pts. (1.3)	19 events / 9 pts. (1%)	6 (4)	3 (2.5) pts.g	NA	NA
Withheld shocksh	90 events/22 pts. (1.1) 1-year FU: NA	NA	47 events/ 22 pts	NA	NA	NA	NA	NA
First shock success (%)	100% 1-year FU: NA	NA	NA	NA	100%	NA	NA	NA
Health-Related Quality of Life	NA	NA	NA	NA	NA	NA	Association between WCD and baseline anxiety: anxiety score: 41 ± 11 vs. 39 ± 13, p = 0.22), rate of anxiety: 58.9% vs. 29.2%, p = 0.02 Statistical trend toward better improvement of depression scores in patients with WCD (mean [SD] change in score points: −4.1 [6.1] vs −1.8 [3.9]; p = 0.09), whereas change of the anxiousness score was not different (−4.6 [9.5]) vs −3.7 [9.1], p = 0.68).	All Kansas City Cardiomyopathy Questionnaire subscales (physical limitation, symptom frequency, quality of life, and social limitation) showed improvement from baseline to day 90 (all Ps < 0.001)
Hospitalisation rate	NA	NA	NA	NA	102 (67)	NA	NA	NA
Satisfaction with technology	NA	NAi	NA	NA	NA	NA	NA	NA
Compliance/ pt adherence
•WCD wear-time in days [mean (range) ± SD; median (IQR)]	Median (IQR): 90 (65) n. s. diff. Between age groups 1-year FU: NA	62 (37–97)	75.0 ± 47.7	54 (1–436)	65.1 ± 42	48 (37–62)	59 (40–96)	NA
•WCD daily use in h/day [mean (range) ± SD; median (IQR)]	Median (IQR): 22.5 (2.7)j Slight s. s. diff. Between age groups: ≥65 y. o.: 22.8 (21.5–23.2) <65 y. o.: 22.3 (19.5–23.0) 1-year FU: NA	23.4 (22.2–23.8) younger age associated with lower compliance [odds ratio (OR) 0.97, 95% CI: 0.95–0.99; P < 0.01]	20.3 ± 4.6	23.5 (range: 1–24)	21.5 ± 3.5	22.9 (21.2–23.4)	20 (±5)	NA
AEs, n (%) of pts:	NA	NA	NA	NA	NA	NA	NA	NA
Device related AEs	NA	NA	NA	5 (1)	NA	NA	NA	NA
•Skin rash and itching	NA	NA	NA	Dermatitis: 4 pts. Pressure mark: 1 pt	NA	NA	NA	NA
•False alarms	NA	NA	NA	NA	NA	NA	NA	NA
Discontinuation due to comfort and lifestyle issues	NA	NA	NA	NA	Not specified, discontinuation due to: Incompliance 12 (8) Other reasons 5 (3)	NA	NA	NA
Serious Adverse Events (SAEs), n (%)	NA	NA	NA	NA	NA	NA	NA	NA
•Inappropriate shocks	10 (0.5)k 1-year FU: NA	8 in 8 pts. (0.7%)	2 in 2 pts. (0.3%)	3 inappropriate shocks in 2 pts. (0.4%)	1 (0.7)	0 (0)	0 (0)	NA
•Unsuccessful shock	0 (0)	0 (0)	NA	NA	0 (0)	NA	NA	NA
Frequency of SAEs leading to death in n (%) of pts	0 (0)	0 (0)	NA	0 (0)	NA	NA	NA	NA

Abbreviations: AEs – adverse events; CABG – coronary artery bypass graft; CMP – cardiomyopathy; DCM – dilated cardiomyopathy; EF – ejection fraction; FR – France; HF – heart failure; ICD – implantable cardioverter-defibrillator; ICM – ischemic cardiomyopathy; INSERM – Institut national de la santé et de la recherche médicale; IQR – interquartile range; LV – left ventricular; LVEF – left ventricular ejection fraction; MI – myocardial infarction; NA – not available; NICM – non-ischemic cardiomyopathy; Ns – not significant; NYHA – New York Heart Association; OMT – optimal medical therapy; PCI – percutaneous coronary intervention; pt(s) – patient(s); SCA – sudden cardiac arrest; SCD – sudden cardiac death; SD – standard deviation; SoC – standard of care; VA – ventricular tachyarrhythmias; VF – ventricular fibrillation; VS – versus; VT – ventricular tachycardia; WCD – Wearable Cardioverter-Defibrillator; yrs – years.

^aOne out of four publications related to this study was already available in the previous report.

^bCologne registry of external defibrillation.

^cHeart Failure Optimization Study.

^dICM pts.: 805 (40%). NICM pts.: 927 (46%). Congenital/Inherited pts.: 268 (14%).

^ePatients were sent follow-up questionnaires at 1, 3, and 12 months.

^f2 patients (8.3%) had a fatal non-arrhythmic event within 3 months after MI.

^gAll shocked patients survived at least 24 h.

^hDue to patients using the response button to delay therapy.

ⁱSelf-defined 5-point likert scale questionnaire was used to evaluate acceptability. This data was not extracted in the absence of the use of a validated questionnaire.

^jNo significant difference in the daily use among the subgroups of ischemic, nonischemic, or congenital/inherited heart disease.

^kDue to ECG artefacts. Inappropriate shocks did not induce VT or VF.

4. Comparative effectiveness and safety

4.1. RCT evidence

4.1.1. Reduction in mortality

According to the ITT analysis of the VEST study [10], there was no statistically significant difference in arrhythmic mortality between the device and control group, with 25 out of 1524 (1.6%) and 19 out of 778 (2.4%) arrhythmic deaths in those groups, respectively (p = 0.18). There was a statistically significantly lower rate of the outcome death from any cause measured as a secondary outcome in VEST: 48 out of 1524 (3.1%) and 38 out of 778 (4.9%) deaths from any cause occurred in device and control group respectively (p = 0.04). This difference was not corrected for multiple testing.

A post-hoc as treated analysis of VEST [22] further found that nine fatal events occurred in patients wearing the device (2420 person-months) as opposed to 32 fatal events in patients not wearing the device (3724 person months), with a relative risk of 0.43 (p = 0.026). The per-protocol analysis showed a hazard ratio (HR) of 0.38 (95% CI, 0.17–0.86; p = 0.02).

4.1.2. Quality of life

No RCT evidence on a potential effect of WCD therapy on quality of life is available, although such data was gathered in VEST [10,22].

4.1.3. Hospitalisation rate

In VEST [10], 31.2% and 32.5% of patients of the device and control group were re-hospitalised within the time period of WCD therapy (p = 0.5).

4.1.4. Patient satisfaction

No RCT evidence on patient satisfaction with WCD therapy is available.

4.1.5. Compliance

In VEST [10], patients wore the device averagely 14 h per day (SD: ± 9.3). The median daily use of the WCD was reported to be 18 h per day, with an interquartile range of 3.8–22.7.

4.1.6. Functional performance of the WCD

In VEST [10], 20 out of 1524 patients (1.3%) in the device group received an appropriate shock. Of those, 13 patients received one shock, and seven patients received two or more appropriate shocks. With regard to withheld shocks, 69 out of 1524 patients (4.5%) aborted one or more shocks by pressing the response button. Shock success was not reported in VEST [10,22].

4.2. Safety

4.2.1. SAEs

The VEST trial [10] reported four SAEs being related or potentially related to the WCD (0.2%). These were patient hospitalizations either due to aborted shocks or due to an inappropriate shock and one further patient died while wearing the device. This death was likely not to be an arrhythmic death, as no tachyarrhythmia was recorded and pulseless electrical activity was noted by emergency medical staff upon arrival. Further SAEs were inappropriate shocks occurring in nine out of 1524 (0.6%) patients in the device group. The WCD therapy was statistically significantly associated with milder AEs such as rash and itching: Rash occurred in 184 (13.0%) patients in the device group compared to 27 (3.8%) patients in the control group (p < 0.001). Itch occurred in 205 (14.5%) patients in the device group as opposed to 22 (3.1%) patients in the control group (p < 0.001).

4.3. Observational evidence

4.3.1. Mortality

In two studies [24,[28], [29], [30], [31],33], the arrhythmic mortality rate was 0%. All-cause mortality was reported by nine studies [[24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36]], ranging from 0% to 5.2% (range of enrolled patients across studies: 102–2000).

4.3.2. Health-related quality of life

Evidence consisted of two studies [23,37]: In one registry study [37], a statistical (positive) association between WCD and baseline anxiety was found when comparing both anxiety score and rate of anxiety between patients wearing a WCD (n = 85) and patients not wearing a WCD (n = 38), with 41 ± 11 vs 39 ± 13 (p = 0.22) and 58.9% vs 29.2% (p = 0.02), respectively (State-Trait Anxiety Inventory). A non-statistical trend toward better improvement of depression scores in patients wearing a WCD was found in the same study, with a mean change in score points of −4.1 ± 6.1 and − 1.8 ± 3.9 (p = 0.09), respectively. The change in anxiousness score was further not statistically significant between device and control group in this registry. The second study was a before-after uncontrolled study [23], with 210 enrolled patients: Statistically significant improvements were found in all Kansas City Cardiomyopathy Questionnaire subscales (physical limitation, symptom frequency, quality of life, and social limitation) from baseline to up to day 90 (p < 0.001).

4.3.3. Hospitalisation rate

Evidence consists of two uncontrolled observational studies [26,34]: In one case series study [26], 13 of 102 enrolled patients (12.7%) were hospitalised and 102 out of 153 enrolled patients (67%) were hospitalised in the other registry study [34].

4.3.4. Satisfaction

No evidence is available to assess patient satisfaction with the WCD.

4.3.5. Compliance

Compliance with the WCD ranged from 20 to 23.5 h per day. Ten observational studies (n = 5092) reported the compliance per day as median or mean: Six studies [23,24,[26], [27], [28], [29], [30], [31], [32],35] reported a median daily wear time ranging from 22.5 to 23.5 h per day. Further five studies [25,33,34,36,37] reported a daily mean wear time that ranged from 20 to 23 h. The range of enrolled patients across studies reporting on compliance was 102–2000.

4.3.6. Functional performance of the WCD

The outcome appropriate shocks were reported in nine observational studies [[24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36]]: These studies reported on percentage of appropriate shocks ranging from 1% to 4.8%, with a range of enrolled patients between 102 and 2000 patients. Withheld shocks were further reported in two studies (2000 and 781 enrolled patients): These studies reported 90 events in 22 patients (1.1%) [[29], [30], [31]] and 47 events in 22 patients (2.8%) [36], respectively. Shock success was reported in three studies [[29], [30], [31],33,34], with a first shock success rate of 100% in these studies.

4.4. Safety

The rate of inappropriate shocks ranged from 0% to 2% and some milder AEs occurred in patients wearing the WCD across ten observational studies [[24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37]]. Other SAEs (e.g., unsuccessful shocks) were not reported or did not occur.

AEs in observational studies were skin rash and itching and false alarms. Two studies [26,32] reported on skin rash and itching: In the Austrian registry with 448 enrolled patients, four patients (0.9%) and one patient (0.2%) developed dermatitis and a pressure mark, respectively. These five AEs were considered device-related [32]. One further study reported on two patients (2%) being allergic to nickel and, hence, developing skin rash [26]. The same study [26] reported that false alarms such as vibrations or sirens occurred in 58 patients (57%) because of incorrect electrocardiography episodes that were found to be artefacts upon further review.

Two further studies reported on the discontinuation due to lifestyle and comfort issued: One study [33] reported on eight patients (7%) returning their WCD during the first hours after initiation due to either unwillingness or inability to handle the device. Another study [34] reported on 12 patients (8%) and five patients (3%) discontinuing WCD therapy, without mentioning the specific reasons for discontinuation. The broad reasons were incompliance or “other reasons”.

5. Discussion

The current available evidence on a potential additional benefit of WCD use consists of one RCT (n = 2348) and eleven observational studies (n = 5345). While RCT evidence still consists of the only RCT published in 2018 [4], new observational evidence was identified in this SR. We found low certainty RCT evidence derived from VEST (n = 2348) indicating that WCD therapy might not be statistically associated with a clinical benefit on arrhythmic mortality in post-myocardial infarction patients with an ejection fraction of ≤35% [10,22]. Registry observational studies show that the compliance is high, with ten studies reporting on a daily wear time between 20 and 23.5 h [[24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37]]. The safety profile and technical performance of the WCD appear to be adequate based on both RCT [4] and observational evidence [[23], [24], [25], [26], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37]].

The results of our SR is aligned with existing knowledge and complements it insofar as more real-world evidence on compliance was found and GRADE was used for transparently synthesising the available evidence. In 2019, the latest independently conducted SR and meta-analysis [9] was published using broader inclusion criteria. The evidence synthesis included 27 observational studies and VEST. The review authors judged observational evidence to be fraught with poor methodology, selection bias and confounding and VEST trial data to indicate that WCD therapy not being associated with a decreased risk of sudden cardiac death. The review authors further conducted a meta-analysis on appropriate shocks across studies labelled as the incidence of appropriate WCD therapy and estimated it to be five per 100 persons over three months across all studies, with noteworthy differences between data from VEST (1 per 100 persons over three months, 95% CI 1.0, 2.0) and observational studies (11 per 100 persons over three months, 95% CI 11.0, 20.0, I² = 93%). On the basis of the available evidence, the study authors of this review noted that WCD therapy should not be used in primary prevention before RCT data justify its use [9].

Another sponsor-funded SR was published in 2020 [17]: The review included VEST, one retrospective observational study with a historical control group and forty-four prospective and retrospective observational studies without a control group. Of the latter, it is highly likely that the number of “studies” in fact refers to the number of available publications. The review found a high rate of appropriate shocks across mixed patient populations and judged the data from observational studies to indicate that the device's safety is confirmed being aligned with results of the aforementioned SR [9] and our results. The study authors [17] did not explicitly conclude on comparative effectiveness WCD therapy, although rightfully concluding positively on the functional performance of WCD therapy in terms of detecting and terminating VT/VF reliably being a surrogate endpoint for an additional clinical benefit.

We agree with authors from one of the aforementioned SR when saying that the WCD is a showcase in which evidence based decision-making falls short [9], with increasing use of WCD therapy all over the world without robust supporting clinical evidence. Based on an online report [38], 200,000 WCDs were prescribed until 2015, although no RCT was available supporting its use. Some critical cardiologists regard this treatment pattern to be “driven by the finality of SCD and partly by fear of litigation, despite the absence of data to support it” [9].

Additionally, there is currently no need for proof of an additional clinical benefit in terms of comparative effectiveness and long-term safety to receive a CE mark in the European Union. Instead, evidence on the technical performance is hereby sufficient to receive a CE mark from notified bodies [39]. As a result of the new Medical Devices Regulation in force since May 2021, there may be changes toward a stronger focus on clinical benefit within the EU regulation [40]. The future will show whether the new regulation will fix existing problems or whether the status quo is upheld [41].

The VEST trial results may also be a textbook example of spin bias [42]: The results of the secondary endpoint (all-cause mortality) were well promoted in both scientific meetings and the media, failing to mention that VEST did not meet its primary outcome [11,43]. Spin can further be seen in follow-up publications [22] focusing on as-treated and per-protocol effects, which was rightfully discussed as secondary and afflicted with high RoB in the original publication of VEST [10].

The results of the secondary endpoint all-cause mortality triggered some cardiologists to generate two interesting hypotheses [44]: In a recent review discussing the clinical utility of the WCD, it was hypothesised, among others, whether the WCD could act as a reminder for self-management and, hence, increase patients' awareness of their heart disease and improve compliance to both OMT and lifestyle changes. Further, improved compliance and a potential reduction in non-arrhythmic mortality could result in reduced all-cause mortality [44]. If the compliance-hypothesis was to be true, WCD would arguably compete against other quite different and probably less costly interventions (e.g., patient education, health care provider monitoring, etc.) to enhance adherence to medications [45]. While hypotheses are always important to stimulate further research, these potential causal links should, at this moment in time, be seen what they are: unproven assumptions. It further appears to be more useful to test the hypothesis for an arrhythmic mortality benefit before re-functioning a defibrillator into a device to improve self-management.

However, absence of evidence justifying WCD therapy may not be confused with evidence indicating that there is no effect [46]. In some specific clinical indications (e.g., explanted ICD due to infection), there may be plausible reasons for using a WCD therapy as the alternative would be a potentially burdensome hospital observation. This is also reflected within evidence-based guideline recommendations. The European Society of Cardiology (2021), for instance, states that a WCD may be considered for a limited time period in selected patients with heart failure and high risk of SCD (IIb recommendation, level of evidence: B), while still highlighting that the VEST trial failed its primary endpoint [47]. Some organizations, such as the Blue Cross Blue Shield Association [48] in the United States, even conclude on the basis of clinical considerations that WCD therapy can be considered a necessary temporary measure before implantable cardioverter (re)implantation in patients who meet the criteria for ICD implantation and who have a temporary contraindication to ICD implantation or who have had an ICD removed because of concomitant infection or malfunctioning. All other clinical use-cases for WCD therapy were considered investigational by the organisation, which we agree upon.

Evidence requirements would arguably change toward non-inferiority in settings in which WCD therapy is a replacement of hospital observation, e.g. within a broader tele monitoring system [49] for which concepts do currently exist [50,51]. This represents a challenge with the available evidence, with statistical testing for non-inferiority or equivalence being the key to overcome these knowledge gaps [52]. While acknowledging that it is improbable that randomised trials such as VEST can ever be conducted for some of these niche indications [53], pragmatic randomised trials [54] or well-designed prospective observational studies with concomitant control may still be feasible in these cohorts of patients.

Patients and clinicians should be informed independently of the available (absence of) evidence before deciding whether to use WCD therapy instead of hospital stay [55]. Within the decision-making process, both compliance with WCD and the risk for mortality further needs to be assessed [56,57]. Health care payers should be aware that once WCD therapy is approved in niche indications, there is the risk of deliberately expanding a therapy in clinical practice.

Although the clinical evidence is sparse, there have been some efforts to capture the utility of the WCD based on decision modelling analysis [6,58,59]. An independent cost effectiveness analysis (CEA) compared a strategy with a WCD (and discharge home) to inpatient monitoring and a discharge to a skilled nursing facility for preventing sudden cardiac arrest (SCA) after infected ICD removal in the United States. Assumptions on SCA risk shock efficacy were retrieved from observational studies. Incremental cost-effectiveness was measured to be $26,436 per quality-adjusted life-year (QALY) when compared to discharge home without a WCD. Discharge to a skilled nursing facility and hospital stay resulted to both higher costs and worse clinical outcomes. The incremental cost-effectiveness ratio (ICER) was $15,392/QALY if WCD successfully terminates 95% of SCA events. Sensitivity analysis shows that the $50,000/QALY threshold was not exceeded if shock efficacy is above 69%, assuming a 5.6% two months SCA risk and bridging for at least two weeks [6]. Another cost-minimisation analysis [58] from Europe estimated a cost reduction of €1782 per patient using the device when comparing a WCD-strategy to an in-hospital stay in patients with ICD explantation (variation in sensitivity analysis of different hospital costs: €3500 to €0).

The same research group conducted another full CEA [59] comparing a WCD strategy to no WCD in post MI patients and impaired ejection fraction. The CEA estimated an ICER of €47,709 per Quality Adjusted Life Year (QALY) gained when considering the (negative) ITT data from VEST. This was cost-effective for the Italian context.

This paradoxical conclusion may be explained by differences between hypothesis testing (as the gold standard for primary medical research) and decision analysis replacing confidence intervals with the probability of an intervention being cost-effective based on pre-defined thresholds [60]. Statistically, a failure to reject the null hypothesis in superiority trials can only lead to the conclusion that there is insufficient evidence to indicate superiority of experimental therapy either due to the null hypothesis being true or the trial being underpowered [52].

Future studies could improve by evaluating the use of the WCD in specific indications in light of standard-of-care being heterogeneous depending on specific clinical manifestations (e.g., occurrence or non-occurrence of a prior arrhythmic event) [47]. Depending on the specific clinical indication, future trials should test not only for superiority, but also for non-inferiority or equivalence (i.e., if hospital observation is the comparator) to generate granular indication-specific evidence on the utility of the device. Bayes factors may be used additionally to quantify evidence in favor of the null hypothesis so that both research waste and research funding in settings where WCD therapy is likely to be not clinically useful is discontinued [52,61].

The results of this article should be viewed in light of its limitations. Firstly, observational studies were included in the absence of multiple RCTs. Although this may be seen as a weakness due to these studies being more prone to internal validity concerns, we carefully selected observational studies according to Cochrane methodology [62] to mitigate concerns. These studies helped to gain insights on the real-world performance (e.g., on appropriate and inappropriate shocks) and safety of the WCD, although causal inference based on these data is not possible. Second, the depicted indications across studies are only a sample of indications of all observational studies. Some excluded studies reported on narrow indications solely such as peripartum cardiomyopathy [63], sarcoidosis [64], explanted ICD [65,66], myocarditis [67], and tachymyopathy [25]. Although these studies may be useful for patient selection for future studies, the risk that these studies would have changed the picture of observational studies is low. Third, the risk of overlapping data within included studies cannot fully be excluded, although being considerably low as most primary studies reported on the study name or identifier of the study they reported on.

There are currently no ongoing RCTs on WCD use. The only other RCT (NCT02481206) aimed to test for superiority of WCD as add-on measure in patients with end-stage renal disease. This study aimed to enrol up to 2600 patients was terminated due to a lack of subject enrolment.

6. Conclusion

The only available RCT failed to show superiority of add-on use of WCD in post MI patients. Observational evidence shows that the compliance with WCD is good, but the evidence is afflicted with selection bias and the inclusion of diverse mixed patient populations diluting the ability to draw indication-specific conclusions on the utility of the device. More comparative data is needed to justify continuing or expanding use of WCD therapy.

Funding

There was no external funding for this manuscript. The manuscript is based on an HTA report.

Declaration of Competing Interest

None declared.

Data availability

Further data with regard to this SR are available online and on request.