Extended Rhythm Monitoring to Assess for Ventricular Arrhythmias After Transcatheter Pulmonary Valve Replacement With the Harmony Valve

Abstract

BACKGROUND:

Varying rates of nonsustained ventricular tachycardia (NSVT) have been reported early after transcatheter pulmonary valve replacement (TPVR) with the Harmony valve, but data regarding rhythm outcomes beyond hospital discharge are limited. This study aims to characterize ventricular arrhythmias after Harmony TPVR from implant through mid-term follow-up.

METHODS:

Ventricular arrhythmia data from postimplant telemetry and follow-up extended rhythm monitoring (ERM) were analyzed after Harmony TPVR.

RESULTS:

Fifty-four patients with tetralogy of Fallot (n=39), valvar pulmonary stenosis (n=10), or pulmonary atresia with intact ventricular septum (n=5) were studied; 22% had prior NSVT and 24% were on prior rhythm medication. On postimplant telemetry, 27 patients (50%) had NSVT, including 1 who had torsade de pointes, but most had <5 episodes. Pre-TPVR NSVT or rhythm medications, diagnosis other than tetralogy, and substantial device contact with the myocardium were associated with more frequent NSVT on telemetry. Ten patients (19%) were started on a new antiarrhythmic medication. On discharge ERM, 37% of patients had NSVT, most with <5 episodes and only 13% with NSVT beyond 5 days post-discharge. On follow-up ERM, 14% of patients had a single episode of NSVT and 1 had 5 episodes. During follow-up, antiarrhythmic medications were discontinued in 8 of 10 patients and no patients died or had sustained ventricular tachycardia.

CONCLUSIONS:

NSVT and ventricular ectopy were common early after TPVR but were infrequent in most cases and diminished rapidly after discharge. The incidence of NSVT on follow-up ERM was similar to preimplant incidence. Few patients had antiarrhythmic medications initiated, and most were discontinued on follow-up. There were no major arrhythmic events after discharge.

WHAT IS KNOWN

• Ventricular premature beats and nonsustained ventricular tachycardia are common early after transcatheter pulmonary valve replacement with the Harmony valve.

WHAT THE STUDY ADDS

• On short-term ambulatory rhythm monitoring, ventricular premature beats and nonsustained ventricular tachycardia diminished rapidly in most patients.

• On medium-term follow-up ambulatory rhythm monitoring, ventricular premature beats and nonsustained ventricular tachycardia were rare, returning to precatheterization incidence, and no patients experienced clinically significant arrhythmias.

Transcatheter pulmonary valve replacement (TPVR) offers an appealing alternative to surgical replacement. Balloon-expandable transcatheter valves, such as the Melody (Medtronic Inc)^1,2 and SAPIEN valves (Edwards, Inc),^3–5 have revolutionized the management of right ventricular outflow tract (RVOT) conduit and bioprosthetic valve dysfunction, but they are often unsuitable for the large native or patched RVOT after prior repair of tetralogy of Fallot (TOF) and related conditions. The self-expanding Harmony Transcatheter Pulmonary Valve (Medtronic Inc.), available in 22 mm (TPV22) and 25 mm (TPV25) sizes, was approved to treat postoperative pulmonary regurgitation in 2021 and has emerged as a promising option in patients with a large RVOT.^6,7

Ventricular arrhythmias have been observed immediately after TPVR with both balloon-expandable and self-expanding valves, with rates varying depending on the anatomic and functional implant environment and valve^8–15 (Table S1). The Harmony TPV22 Early Feasibility Study⁶ demonstrated nonsustained ventricular tachycardia (NSVT) in 5% of patients, but subsequent studies from trials that included both the TPV22 and TPV25 reported NSVT in 14% to 16%.^7,14 In a prior analysis, we found that NSVT early after implant was more common (40% of 30 patients studied) but that the frequency decreased substantially over the first 12 to 24 hours after implant.¹³ A recent multicenter study reported postimplant ventricular arrhythmias in 24% of patients and initiation or up-titration of antiarrhythmic medications in 19%.¹⁵

Although previous studies offered valuable insights into ventricular arrhythmias early after the Harmony valve implant, a gap remains in our understanding of rhythm outcomes beyond hospital discharge. Because of the high frequency of immediate postimplant NSVT in our early experience, we commonly place an ambulatory extended rhythm monitoring (ERM) device at the time of discharge and during clinical follow-up. Our primary objective in this study was to evaluate the frequency of ventricular arrhythmias and ectopy in patients with a Harmony valve beyond the immediate postimplant period.

Methods

Patients and TPVR

This study comprised all patients who underwent TPVR with the Harmony valve TPV22 or TPV25 at our center between September 2017 and February 2024. The study was approved by the Stanford University Institutional Review Committee with a waiver of informed consent. The summary data and analytical methods will be available on reasonable request.

Records were reviewed to ascertain prior history of NSVT, other rhythm abnormalities, and baseline antiarrhythmic medications. We recommended precatheterization ERM in patients referred for TPVR. Further evaluation was recommended as indicated based on prior history and ERM findings. Invasive electrophysiological studies were performed selectively, rather than routinely, after discussion of historical factors and ERM findings in a dedicated multidisciplinary forum.

TPVR was performed with either the TPV22 or TPV25 device. Patient-related and procedural variables were evaluated as reported in our prior study.¹³ The implant was categorized as annular (valve housing segment of the device implanted at the native annular level, with the inflow rows below the annulus and the outflow rows in the pulmonary artery) or nonannular (valve housing above the native annulus and at least 5 of the 6 rows in the main pulmonary artery). Because the device frame can contact the myocardium in a nonannular implant, we also characterized the degree of contact between the device and RVOT myocardium as none/minimal (≤1 stent row) or moderate/substantial (>1 row), based on the implanter’s assessment at the time of the procedure. All patients remained in the hospital at least overnight after the procedure. In response to our early experience, we developed a standardized post-TPVR ventricular arrhythmia evaluation and management protocol, which was adopted in November 2021 (Figure 1).

Figure 1.

This flow diagram depicts the standardized protocol we developed in 2021 and have applied in patients undergoing transcatheter pulmonary valve replacement since. ALS indicates advanced life support; bpm, beats per minute; CVICU, cardiovascular intensive care unit; EP, electrophysiology; IV, intravenous; NSVT, nonsustained ventricular tachycardia; PVCs, premature ventricular contractions; and VT, ventricular tachycardia.

Electrophysiological Evaluation

Patients were monitored with continuous telemetry throughout the postimplant hospital stay. For this study, telemetry data were reviewed until discharge or up to 48 hours of admission, whichever was shorter. Telemetry recordings were reviewed thoroughly and manually for the number, length, and fastest rate of ventricular tachycardia (VT) events and quantification of the frequency of ventricular premature beats (VPB). Details of peri-procedural rhythm evaluation were described in our prior report.¹³ VT, defined as ≥4 consecutive beats at a rate ≥120 bpm, was further qualified as NSVT if <30 seconds or sustained if ≥30 seconds. Bigeminy and trigeminy were considered VPB. Patients with NSVT on telemetry were categorized as having frequent NSVT (>5 episodes) or infrequent NSVT (≤5 episodes) in the first 24 hours postimplant. Patients without NSVT on telemetry were further grouped according to low (<1%) or high (≥1%) VPB burden in the first 24 hours postimplant.

ERM was obtained after discharge and during follow-up in a variable fashion. Early in the experience, postimplant ERM with a Zio monitor (iRhythm, San Francisco, CA) was ordered selectively in patients with rhythm concerns. However, in 2018, we modified our practice to place a Zio monitor at the time of discharge in all patients. Follow-up ERM was recommended in all patients and was obtained at variable time frames in coordination with the referring physician. In most patients, follow-up ERM consisted of a Zio monitor, but a Philips Mobile Cardiac Outpatient Telemetry (BioTelemetry, Malvern, PA) was also used. In patients with an existing pacemaker or implantable cardioverter defibrillator, data from device interrogation were reported. ERM was prescribed for 14 days, but the actual duration of monitoring varied. For this study, follow-up ERM studies were categorized as early (within 12 months of implant) or late (>12 months after implant).

Data Analysis

Data were reported as median (25th, 75th percentiles) or (minimum–maximum). Selected comparisons between groups were performed using χ² analysis, the Fisher exact test, or the Wilcoxon rank-sum test, depending on the type (categorical or continuous) and the underlying distribution of the variables. For analysis of postimplant telemetry findings, patients with any NSVT were compared with those with no NSVT, and patients with frequent NSVT were compared with those with infrequent or no NSVT. Analysis was performed with SPSS version 29.

Results

Patients and TPVR

A total of 54 patients with pulmonary regurgitation underwent TPVR with the Harmony valve during the study period. Diagnostic and procedural factors are summarized by the arrhythmia group in Table 1 and by underlying diagnosis in Table 2. Most patients had TOF (n=39, 72%), whereas 10 (19%) had valvar pulmonary stenosis (PS) and 5 (9%) had pulmonary atresia with intact ventricular septum (PA/IVS). Pulmonary regurgitation was secondary to surgical intervention on the RVOT/pulmonary valve in 52 patients (96%) and percutaneous balloon valvuloplasty for valvar PS in 2 (4%).

Table 1.

Diagnostic, Procedural, and Early Arrhythmia Outcomes

Table 2.

Diagnostic, Procedural, and Early Rhythm Outcomes by Underlying Diagnosis

Before TPVR, 12 patients (22%) had a documented history of NSVT. Details of other arrhythmias, antiarrhythmic medications, and cardiac implantable devices are summarized in Table 3. Four patients underwent an invasive electrophysiology study pre- or post-TPVR, as detailed in Table S2.

Table 3.

Details Related to Pre-TPVR Arrhythmias and Rhythm Medications

A TPV25 valve was implanted in 49 (91%) patients and a TPV22 in 5 (9%; all TOF, all nonannular). The implant was annular in 37 patients (69%), and there was moderate/substantial contact between the device and RVOT myocardium in 41 (76%).

Arrhythmias After TPVR

Overall, 27 patients (50%) had NSVT detected on continuous rhythm monitoring while hospitalized after the implant. As previously reported, 1 of these patients, a 41-year-old with TOF who was on metoprolol at baseline and had multiple runs of NSVT and 100% atrial fibrillation burden on pre-TPVR ERM also experienced a cardiac arrest secondary to torsade de pointes and ventricular fibrillation.¹³ There were no other instances of sustained VT or malignant arrhythmias. Most patients were discharged after overnight observation, but 14 (26%) were hospitalized for >1 day, predominantly for additional rhythm monitoring or initiation of antiarrhythmic medication (n=8, including 1 planned initiation of sotalol for intra-atrial reentrant tachycardia). Ten patients (19% overall, 37% of those with NSVT) were initiated on new antiarrhythmic medications for ventricular arrhythmia (n=9) or frequent VPBs without NSVT (n=1; Table 3), and the metoprolol dose was increased in 1 patient previously on metoprolol pre-procedure. Rhythm medications were resumed at the prior dose in all other patients on preimplant therapy. New antiarrhythmic medications were initiated more often after the adoption of the standardized protocol in November 2021 (7 of 24 patients compared with 3 of 31 before implementation, P=0.063).

Postimplant Telemetry

Among the 27 patients with NSVT on telemetry, the number of episodes ranged from 1 to 452, 77% of which occurred during the first 12 hours after implant. Only 11 of these 27 patients (20% overall) had frequent NSVT, while 16 had <5 episodes, including a single episode in 6 patients and 2 or 3 episodes in 4 patients each, and only 14 (26% overall) had episodes beyond the first 12 hours (Figures 2 and 3). As detailed in Table 1, patients were divided into 4 ordinal groups according to ventricular arrhythmia burden on postimplant telemetry: frequent NSVT, infrequent NSVT, no NSVT with high VPB burden, and no NSVT with low VPB burden. Patients in both NSVT groups had significantly fewer episodes in the second 12 hours than in the first 12 hours of telemetry (P<0.001). Factors associated with the identification of any NSVT or frequent NSVT on postprocedural telemetry included pre-TPVR history of NSVT, preimplant rhythm medications, a diagnosis of valvar PS or PA/IVS, and substantial device contact with the RVOT myocardium (Tables 1 and 2).

Figure 2.

This flow diagram details the distribution of nonsustained ventricular tachycardia (NSVT) frequency at each monitoring time point. Data for the follow-up extended rhythm monitoring (ERM) includes the most recent study in patients with both early and late follow-up ERM.

Figure 3.

This line chart depicts the normalized number of nonsustained ventricular tachycardia (NSVT) episodes per day during the first (0–12 hours) or second (12–24 hours) 12 hours on postimplant telemetry, during the first 5 days (discharge [DC] 1–5 days) and beyond the first 5 days (DC after 5 days) on the DC extended rhythm monitoring (ERM), and on early and late follow-up ERM for each patient. The number of patients who underwent each evaluation is listed below the time point label, and further below is the number of patients for whom the value at the time point was 0. The number of NSVT episodes is presented on a log(5) scale. Some 0 points are jittered and appear between 0 and 1 on the y axis.

As detailed in Table 1, VPB frequency ranged from 0.04% to 19.6%, with frequent VPB (>5%) in 12 patients (22%), occasional VPB (1% to 5%) in 14 (26%), and rare VPB (<1%) in 28 (52%). VPB frequency was higher in patients with valvar PS or PA/IVS than in those with TOF (P=0.021), and in patients with NSVT (P<0.001) or frequent NSVT (P<0.001) on telemetry (Figure S1). VPB frequency did not differ according to other factors.

ERM After Hospital Discharge

Overall, 30 (56%) patients had ERM starting from the time of postimplant hospital discharge, including the most recent 24 implants and 17 of the 27 patients with NSVT on telemetry (Table 4). As detailed in Figure 3, 11 patients (37%) had ≥1 episode of NSVT on discharge ERM (1–148 episodes) lasting 4 to 19 beats (median, 7). Patients with NSVT on telemetry were more likely to have NSVT on discharge ERM (53% versus 15%, P=0.040), and all patients with ≥5 NSVT episodes on discharge ERM also had NSVT on telemetry. Nine of 11 patients with NSVT (82%) had <10 episodes. Overall, 97% of NSVT episodes were during the first 5 days of monitoring (Figures 2 and 3), and only 4 (15%) patients had NSVT >5 days post-discharge. In the 2 patients with the highest NSVT frequency on discharge ERM (71 and 148 episodes), all or most episodes (>90%) were on day 1, respectively, and none were observed after day 3 of monitoring (Figure S2). Four of 11 patients with NSVT on discharge ERM also had pre-TPVR NSVT, and 9 had NSVT on postimplant telemetry, whereas the 2 without NSVT on telemetry had a single short episode of NSVT on the discharge ERM. VPB data are summarized in Table 4.

Table 4.

Findings on Postdischarge and Follow-Up Extended Rhythm Monitoring

Follow-Up ERM

Patients were followed for a median of 22 months after implant (7, 30 months). After hospital discharge, no patients died or had clinically significant ventricular arrhythmias or arrhythmia interventions.

Beyond the discharge ERM, 29 (54%) patients were evaluated with early (<12 months postimplant, n=21) and late (n=15) follow-up ERM (Table 4; Figure 3), which included implantable cardioverter defibrillator interrogation in 2 patients. Five of these patients (17%) had NSVT detected: 2 on early ERM with no NSVT on late ERM, and 3 (10%) on late ERM (no early ERM performed; Figure 3). Two of these 5 patients had documented pre-TPVR NSVT, and 4 had NSVT on telemetry. Four of these 5 patients had a single NSVT episode, whereas 1 had 5 episodes, and the longest episode for each patient ranged from 4 to 7 beats (Figure 2). Overall, the incidence of NSVT early and late after implant (17%) was comparable to the preimplant incidence (22%). VPBs were rare on follow-up ERM, and only 1 patient (who did not have NSVT) had >1% VPBs.

Outcomes in Patients With NSVT on Postimplant Telemetry

Among the 11 patients with frequent NSVT on postimplant telemetry, 4 of 8 had NSVT on discharge ERM and 2 of 9 had a single episode of NSVT on early (n=1) or late (n=1) follow-up ERM. In 8 of the 10 patients who were started on new antiarrhythmic medications for ventricular arrhythmia after implant, the medication had been discontinued at a median of 2.5 months (1–8) after implant (Table 3).

Discussion

This study builds on our prior investigation and includes a complete review of arrhythmia outcomes in 54 consecutive patients who underwent TPVR with the Harmony valve. We performed a more extensive evaluation of postimplant telemetry data in a larger cohort than in our earlier study and expanded on that analysis by evaluating ERM data at various time points after implant, from post-discharge to 3 years post-TPVR, in over half of the study cohort. The key finding of this study is that while the early post-TPVR incidence of NSVT is high, outpatient postprocedure monitoring suggests that NSVT incidence rapidly falls back to the preprocedure baseline.

As in previous studies, we found that NSVT and frequent VPBs were common early after TPVR with the Harmony valve. Although the 50% incidence of NSVT was higher than in prior studies,^14,15 we suspect this difference is related to the meticulous review of telemetry data and recording even a single episode of NSVT. Moreover, most patients with peri-procedural ventricular arrhythmias had few, typically brief, episodes of NSVT that were concentrated in the first 12 hours after implant, with rapid diminution thereafter and only rare NSVT or VPBs on follow-up monitoring, and no major arrhythmic events after discharge. Infrequent NSVT was detected on follow-up ERM in a small subset of patients. Many, but not all, of these patients had a documented history of VT before TPVR, and overall, the incidence of NSVT on follow-up evaluation was comparable to the preimplant incidence. Prior studies in patients who underwent TPVR with balloon-expandable valves alsoreported resolution of peri-procedural ventricular arrhythmias (Table S1).^8–12 Although the general tendency of ectopy to resolve soon after implant is reassuring, larger studies will be necessary to confirm the similar incidence of late postimplant NSVT compared with the pre-TPVR baseline in this population.

Potential mechanisms for ventricular arrhythmias after TPVR have been postulated in other studies and almost certainly relate to contact stimulation of the myocardium by the stent frame. This hypothesis is supported by the finding in our prior study¹⁴ that NSVT immediately after implant was associated with an annular implant, which places the inflow portion of the device completely in the right ventricle, and by the association of NSVT in this study with a subjective metric of substantial myocardial contact between the device and RVOT. However, although substantial myocardial contact was more common than annular implant and almost universal among patients with NSVT on telemetry, it was not specific for ventricular arrhythmias, as some patients with myocardial contact demonstrated no NSVT and a low VPB burden. It is noteworthy that, regardless of implant location or extent of myocardial contact, NSVT beyond the immediate postimplant period was uncommon, and no patients experienced postdischarge events. These findings suggest that annular implant location need not be avoided simply out of concern for ventricular arrhythmia. Although contact stimulation is the likely mechanism of ectopy in this setting, it remains unclear whether certain contact locations (eg, free wall, septum, critical isthmuses) or kinetics (ie, intermittent with contraction versus sustained, etc.) pose a higher risk, and why the frequency tends to diminish so substantially and rapidly, whether through habituation, right ventricular remodeling, or other factors.

Aside from the extent of myocardial contact with the device, the underlying ventricular substrate may be a factor in susceptibility to post-TPVR arrhythmias. A prior multicenter study found that arrhythmias were more common after the Harmony valve implant in patients with valvar PS than in those with TOF.¹⁵ In the present cohort, frequent NSVT and a higher VPB burden after implant were more common in patients with PS and PA/IVS than in those with TOF. Given that annular implant location was more common in patients with PS or PA/IVS than in those with TOF, the associations of ventricular arrhythmias with diagnosis and implant location are somewhat confounded, and the number of patients is too small to resolve with multivariable analysis. Aside from differences in device-myocardium interaction, it is unclear why patients with different underlying conditions may exhibit varying susceptibilities to contact stimulation and arrhythmia. However, one could speculate about differences related to ventricular septal defect closure, the presence of patch material, the extent of myocardial fibrosis, and other intrinsic factors. Regardless, further insight into mechanisms and mitigating factors related to this phenomenon will require attention in ongoing and future research.

Ventricular arrhythmias related to a constellation of pathophysiologic factors are an important long-term concern among patients with TOF and related congenital cardiac anomalies involving the RVOT.^16–19 Although the mechanism of post-TPVR NSVT is almost certainly different from the slow-conducting isthmus-mediated substrate more widely seen in repaired TOF,²⁰ it is noteworthy that sustained VT was only observed in a single patient who had multiple episodes of NSVT and persistent atrial fibrillation on pre-TPVR ERM. Our current approach in the TOF population undergoing TPVR is to perform a preprocedural electrophysiology study in only the highest risk patients, but recent data would suggest that >20% of TOF patients undergoing pulmonary valve replacement are likely to have ventricular substrate capable of sustaining VT.²¹ Based on our experience and prior reports, the ventricular arrhythmias provoked by TPVR seem to trigger sustained VT in a much lower proportion of patients than stimulation protocols in the electrophysiology lab, and this finding is somewhat reassuring. The findings of this study do not provide significant evidence that wider scale invasive electrophysiology studies or prophylactic ablation would reduce life threatening arrhythmias in the early post-TPVR course.

Our protocol for the evaluation and management of ventricular arrhythmia after TPVR was adopted in November 2021 to provide a standardized pathway for the initiation of antiarrhythmic medications. Although there is no clear evidence regarding the benefit of antiarrhythmic medications in patients with NSVT due to presumed focal myocardial stimulation after TPVR, we have become more likely to initiate new antiarrhythmic medications since protocol implementation, and this practice is increasingly prevalent without clear evidence of pharmacological efficacy.²² In this cohort, new medications were typically continued for only a short time, with a median time to discontinuation of 2.5 months. In both early and late follow-up, ERM showed almost no NSVT and only rare VPBs, regardless of whether patients were on or off antiarrhythmic medication. The overall prevalence of ventricular arrhythmias on follow-up ERM returned to the preimplant baseline, but it should be noted that our cohort had a relatively high ventricular arrhythmia burden preimplant when compared with historical asymptomatic cohort rates of around 8%.²³ This may reflect the higher risk profile of our cohort or the intensity of preprocedural and postprocedural monitoring, but it is notable that there were no late malignant ventricular arrhythmia events. In addition, neither of the 2 patients with an implantable cardioverter defibrillator in our cohort received or required an implantable cardioverter defibrillator shock during follow-up (2 and 3 years, respectively).

Limitations

Aside from the relatively small size and retrospective nature of the study, there were several limitations that merit further discussion. The study provides limited insight into the mechanisms of postimplant arrhythmia. Although annular implant and subjective assessment of myocardial contact were analyzed as surrogate measures of myocardial contact with the device, they may not fully account for the magnitude or cadence of myocardial stimulation. In addition, not all patients received ERM, and among those who did, some did not have both discharge and follow-up ERM. This limits our statistical power and may introduce sampling bias, but the data suggest that late ERM was most likely to be performed in those with the highest clinical concern for ventricular arrhythmias. With respect to assessing the potential importance of ventricular arrhythmias over time, the number of patients with follow-up ERM and positive findings was too few for rigorous statistical analysis.

Conclusions

Although NSVT and VPB were common immediately after the Harmony valve implant, most patients had a limited number of short NSVT episodes. Regardless of immediate postprocedural arrhythmia or ectopy burden, and regardless of patient-related or procedural factors, both VPBs and NSVT diminished rapidly in most patients, and no patients experienced clinically significant postdischarge arrhythmias during a median follow-up of almost 2 years. These findings should be reassuring that, despite frequent early postprocedural NSVT, ventricular arrhythmia does not seem to be a common ongoing issue. Several important questions remain with respect to this issue, including the exact mechanism of ventricular arrhythmias after TPVR, whether ectopy responds to antiarrhythmic medications and which patients should be treated, why peri-implant ventricular arrhythmias improve so quickly, and, most important, whether there is a long-term risk of clinically important arrhythmias related to the TPV device.

ARTICLE INFORMATION

Sources of Funding

None.

Disclosures

Dr McElhinney is a proctor and consultant for Medtronic and Edwards. The other authors report no conflicts.

Supplemental Material

Tables S1–S2

Figures S1–S2