Orthotopic heart transplantation (OHT) may be a lifesaving therapy for patients with end-stage heart failure. One important surgical complication is tricuspid regurgitation (TR), which is frequently noted both immediately and late after transplantation, with a reported incidence between 47% and 98%.1 The development of TR stems from multiple etiologies, including pressure overload from pretransplant pulmonary hypertension, primary allograft dysfunction, surgical manipulation, or chordae disruption from endomyocardial biopsies. The TR in the setting of transplant is independently associated with considerable morbidity, including renal insufficiency, reoperation, and early mortality.1,2
Prophylactic tricuspid annuloplasty of the donor heart before implantation into the transplant recipient was previously reported to improve short- and long-term degrees of TR, renal dysfunction, tricuspid repair rate, and mortality in several small observational and retrospective series.3–6 The DeVega annuloplasty (DVA) was developed by the Spanish surgeon Norberto de Vega as an alternative to prosthetic rings for the treatment of functional TR. The procedure achieves annular reduction by stabilizing the tricuspid apparatus with double layered suture guided by a standardized dilator. The DVA in OHT is hypothesized to exert maximal benefit during the perioperative period by improving post-transplant tolerability of hemodynamic stress and right ventricular (RV) strain, thereby decreasing adverse outcomes.3,4
After the incorporation of DVA into the surgical transplant protocol at our institution, an increased number of conduction abnormalities were observed. There is a paucity of literature reporting the electrophysiologic consequences of DVA in OHT or in native hearts receiving DVA for functional TR. Therefore, we sought to determine the risks and benefits of the DVA procedure in our transplant patients by comparing patients who had received DVA to an earlier cohort who had not received DVA. The purpose of this analysis was to describe conduction disturbances and pacemaker (PPM) requirement in this select patient population and to compare electrical and hemodynamic outcomes with historical controls.
This is a retrospective analysis of 180 consecutive adult cardiac transplantation patients who underwent surgery between 2013 and 2017. Outcomes before and after incorporation of DVA into the transplant surgical protocol in 2015 were studied (104 patients without DVA from January 2013 to April 2015 and 76 patients with DVA from April 2015 to January 2017) at Columbia University Medical Center, New York, NY. Patients were classified into group without DVA versus with DVA. Demographic, clinical, and outcomes data were collected and entered into a database for every transplant recipient and donor. Donor:recipient mismatch was defined as present if noted by the attending surgeon in the operative note after their review of donor and recipient height and weight. The Columbia University Institutional Review Board approved our protocol before our gathering of patient data.
Patients who received DVA underwent the following procedure: The donor heart was placed in cardioplegia, and the tricuspid valve was exposed through a right atriotomy. A pledgetted annuloplasty suture was initiated at the anteroseptal commissure and sewn intermittently around the tricuspid annulus along the anterior and posterior leaflets until just past the posteroseptal commissure. Another pledget was placed, and the suturing was reversed along the posterior and anterior leaflets 1 to 2 mm in front of the first row, alternating the suture technique until the original pledget at the anteroseptal commissure was reached. The suture was tied down over a presized Hegar dilator of 25–29 mm coordinated with the patient’s body surface area (BSA) at the discretion of the cardiac surgeon. The DVA procedure at our institution on average added 2–3 min to the overall operative time. There were no differences in our biopsy protocol during the first post-transplant year between the with and without DVA groups.
Electrocardiogram Analysis and Definitions
All available 12-lead electrocardiograms (ECG) were interpreted during the time period between OHT and before a patient’s first endomyocardial biopsy to eliminate possible effect from RV conduction system injury during right heart catheterization. Standard conduction intervals and conduction abnormalities were measured on postoperative days 1–7. Ventricular paced and nonsinus rhythm ECGs were excluded from the analysis. Conduction abnormalities, such as complete heart block (CHB), right bundle branch block (RBBB), and fascicular blocks, were deemed present if noted on ≥2 ECGs during sinus rhythm or during sinus tachycardia. Conduction abnormalities were also considered present when observed during interruption of pacing on an ECG postoperatively and when mentioned as the diagnosis in a cardiac electrophysiology consultation note. The diagnostic criteria for RBBB consisted of the presence of a rSR′ pattern in lead V1 with r′ amplitude of at least 0.2 mV and a terminal S wave in lead I and V6 with a QRS duration ≥0.12 seconds. First-degree atrioventricular block was diagnosed when the PR interval exceeded 200 ms. Left anterior fascicular block (LAFB) was defined as a QRS axis between −45° and −90° and QRS duration of less than 0.12 seconds in the presence of an RS complex in the inferior leads and a qR complex in lead I and aVL. Bifascicular block was defined as RBBB with LAFB. Trifascicular block was defined as bifascicular block with first-degree atrioventricular block. All ECGs were analyzed by two blinded cardiologists.
All echocardiographic and cardiac catheterization data during the first year after OHT were analyzed to compare TR severity and intracardiac hemodynamics between patients who did and did not receive DVA. The echocardiographic degree of TR by color Doppler was reported as none/trace, mild, mild/moderate, moderate, moderate/severe, and severe by qualitative assessment in our institution’s echocardiography lab. The severity was charted during our study using corresponding numerical values 0–3, with higher numbers indicating greater severity. An identical numerical connotation was used to analyze RV dysfunction by echocardiography reports (i.e., mild dysfunction = 1, moderate/severe = 2.5). Data from cardiac catheterizations were used to determine RV stroke work index (RVSWI) by the equation RVSWI = Stroke volume index × (Mean pulmonary arterial pressure − Central Venous Pressure) × 0.0136.
Data are presented as mean ± standard error of the mean for continuous variables and as frequency and percent for categorical variables. Variables were compared between the groups with Student’s unpaired two-tailed t-test for continuous variables. All p values are reported as two-sided tests with p < 0.05 considered statistically significant. Outcomes were compared before and after the institution of the DVA in the surgical protocol. Graph Pad Prism 6.0 was used to organize data and perform between group analyses.
Clinical Characteristics and Intraoperative Data
The baseline demographic and clinical characteristics of the population analyzed are shown in Table 1. The study population consisted of 180 consecutive patients from 2013 to 2017 who underwent OHT, as well as their donors. There were 104 patients included without DVA and 76 patients with DVA. The DVA patients were older (56.5 ± 1.0 vs. 52.4 ± 1.2; p = 0.017) and more often male (78% vs. 68%; p = 0.02). There was no significant difference between groups with regard to BSA, ischemic heart disease, preoperative left ventricular assist device, preoperative amiodarone use, preoperative pulmonary vascular resistance, and donor:recipient BSA mismatch. Bypass time was shorter in the DVA group (148.2 ± 5.1 min vs. 174.3 ± 6.0 min in the no DVA group; p = 0.002) and more patients in the DVA group who required cardioversion in the operating room (76% vs. 38%; p < 0.001). There were six deaths after OHT in both the with DVA and without DVA group during the first year post-OHT. One patient in the without DVA group had severe TR, and no patients in the with DVA group had severe TR.
ECG data are summarized in Table 2 and (see Table 1, Supplemental Digital Content, http://links.lww.com/ASAIO/A324). There were eight patients in the without DVA group and six patients in the with DVA group who did not have any ECGs with sinus rhythm during the first postoperative week (either paced, junctional, atrial fibrillation, or CHB). The DVA patients had significantly longer PR and QRS intervals on day 7 post-OHT. There was no difference in the QTc interval during the initial 7 days after surgery between the groups.
The incidence of conduction abnormalities during the first postoperative week is shown in Table 3 and (see Table 3, Supplemental Digital Content, http://links.lww.com/ASAIO/A326).
Right Bundle Branch Block
There were eight patients in the without DVA group and six patients in the with DVA group who did not have any ECGs with sinus rhythm. The RBBB was more frequent in DVA patients than in patients without DVA during the first postoperative week (37.1% ± 5.9%, n = 70 vs. 9.4% ± 2.9%, n = 96; p < 0.001). The incidence of RBBB was not significantly related to DVA dilator size (47% of DVA patients with RBBB received 27 mm Hegar dilator; see Table 4, Supplemental Digital Content, http://links.lww.com/ASAIO/A327). There was no difference between groups in the frequency of post-transplant accelerated or slow junctional rhythm during week 1 (13.4% ± 3.5%, n = 104 without DVA vs. 21% ± 0.05%, n = 76 with DVA; p = 0.26).
Complete Heart Block
In the without DVA group, there were no incidences of complete heart block during the index hospitalization after transplantation, whereas in the DVA group, there were three cases of early CHB. The three DVA patients with early CHB received a PPM on post-transplant hospital days 13, 18, and 32, respectively. One year pacing follow-up was available for two of the three DVA patients with early CHB—with DDD 60–130 programming, both patients exhibited <1% ventricular pacing. In the without DVA group, two patients developed late CHB over 2 years after OHT. These two patients had minimal ventricular pacing 1 year after PPM placement.
Sinus Node Dysfunction
In the without DVA group, there were six cases of early sinus node dysfunction compared with five cases of sinus node dysfunction in the DVA group during the initial OHT hospitalization. In the without DVA group, only one of the patients with sinus node dysfunction required a PPM on post-transplant day 11. Similarly, one patient with DVA and sinus node dysfunction received a PPM on post-transplant day 24. At 1 year follow-up, the without DVA patient had 15% ventricular pacing and the DVA patient had 84% atrial pacing with <1% ventricular pacing. There was one additional case of sinus node dysfunction requiring PPM nearly 2 years after OHT in the without DVA group. This patient had late sinus node dysfunction and had minimal ventricular pacing at 1 year.
Hemodynamic Effects of DVA
Hemodynamics at 1 week, 3 months, and 1 year after OHT are summarized in Table 4. There was no significant difference in regards to TR severity, right atrial pressure, or RVSWI at 3 and 12 months although there was on average significantly less TR with DVA at 1 week. The with DVA group had a higher mean degree of RV dysfunction at 12 months (0.46 ± 0.09 vs. 0.23 ± 0.05 without DVA) although the mean severity was still less than mild overall. No patients in the DVA group had moderate or greater TR at 1 year, whereas three patients in the group without DVA had moderate or greater TR at 1 year (see Table 2, Supplemental Digital Content, http://links.lww.com/ASAIO/A325). The echocardiographic data were affected by 12 patient deaths (six in each group) and variable follow-up at specific time points. The percent of patients with greater than moderate RV dysfunction and TR is shown in supplemental Table 2 (see Table 2, Supplemental Digital Content, http://links.lww.com/ASAIO/A325).
The current study elucidates the electrocardiographic and electrophysiologic consequences of prophylactic DVA in OHT. The data also reveal the hemodynamic effects of DVA in OHT.
The key findings are as follows: 1) Patients with DVA had significantly higher rates of RBBB, PR, QRS prolongation, and CHB compared with patients without DVA; 2) PPM implantation was more frequent during the index hospital stay in patients receiving DVA (four PPMs in with DVA group versus one PPM in without DVA group); 3) At 1 year follow-up, there were no differences in the severity of TR between patients with and without DVA.
Previous studies on this topic include one randomized study on clinical outcomes of DVA during OHT comparing 30 patients who underwent DVA before heart engraftment with 30 control patients.4 Of the patients in the DVA cohort, none had significant TR at 6 years and none required further repair, whereas one in three control group patients developed significant TR, which was progressive in nature. The absence of prophylactic DVA was associated with increased renal dysfunction and mortality.4–6 The incidence of conduction disease was not analyzed although three patients in each cohort required temporary transvenous pacing and no DVA patients required permanent pacing during follow-up.4 A more recent, larger retrospective study similarly reported that DVA prevented severe TR and operative tricuspid repair although no difference in mortality was observed in patients who did not receive DVA.6
Our retrospective analysis on a much larger group shows that patients who received DVA had longer PR intervals and QRS duration during the first week post-OHT. The wide QRS interval was driven by the high incidence of RBBB. Although RBBB is generally common and thought to be a benign finding in patients with heart transplant,7 progressive bundle branch block is associated with increased mortality.8 Our analysis found a high incidence of RBBB among patients receiving DVA (37%) early within the first post-transplant week, whereas studies on the prognostic value of RBBB analyzed ECG well after transplantation and generally report lower RBBB rates (20%).7 Our data indicate that early RBBB does not recover at 1 year. Other conduction abnormalities, such as first-degree atrioventricular block and LAFB and CHB, were also more common in the setting of DVA. There was an associated increase in PPM placement after the incorporation of DVA although absolute event numbers remained low.
The ischemic time at our institution decreased after 2015 because the cross-clamp was removed earlier in the operation because of a change in protocol to further improve outcomes. The DVA procedure on average takes 2–3 min to perform during orthotopic heart transplant surgeries and per previously published studies does not significantly increase overall ischemic time.5 Therefore, we do not believe that increased ischemic time due to DVA accounts for the conduction disturbances reported in this study.
We propose that annuloplasty of a normal-sized, healthy donor tricuspid valve may lead to increased tissue injury compared with performing DVA on a diseased, dilated tricuspid annulus with significant TR. The cinching of the tricuspid apparatus may possibly stretch the cardiac electrical system and cause prolonged conduction by direct surgical injury. Alternatively, because of increased competition for organs and a small donor population, the conduction abnormalities may be related to older, sicker transplant recipients receiving less ideal hearts.
Although there was no difference in the overall degree of TR at 1 year in our analysis between groups, zero patients in the DVA cohort had moderate or greater TR compared with three patients in the no DVA cohort. This indicates that DVA is a durable procedure to prevent TR in the transplant population although our 1 year data suggest it is not necessary to perform it in all patients. Furthermore, at 1 year follow-up, the DVA group had increased RV dysfunction (Table 4). Overall, the DVA may be conceptualized as a balanced trade-off between the risk of early conduction injury and the benefit of preventing late moderate/severe TR in a select cohort of transplant patients. For example, in our patients, we “traded” three cases of CHB and subsequent PPM placement in the DVA group for three cases of moderate/severe TR in the no DVA cohort. Two of the DVA patients with PPM for CHB had <1% ventricular pacing at 1 year in DDD mode despite waiting considerable time after OHT to implant PPM, indicating that there may be reversibility of conduction injury albeit over a significant amount of time over time. We feel ample wait time (average of 21 days) was given for recovery of conduction and demonstrates a conservative policy for PPM implantation at our institution.
Biatrial surgical technique, advanced donor age, preoperative amiodarone use, prolonged donor ischemic time, and rejection are risk factors for PPM requirement.9,10 A large review of the United Network for Organ Sharing registry revealed that PPM implantation may be required in up to 10.9% of patients.11 Bradycardia is the most common post-transplant arrhythmia, occurring in up to 23% of patients postoperatively and most frequently related to sinus node dysfunction.9 High-grade heart block is rare but may also occur in approximately 2% of patients. A longitudinal series from Stanford University reported that 5.8% of OHT patients received PPMs over 40 years.12 Over the long term, it is unlikely that post-transplant patients with PPMs become PPM dependent, unless they have late-onset atrioventricular block.13 One large series found that 6 months after PPM implantation, only 14.5% of OHT patients were dependent.14 Our experience indicates that the patients who received a PPM for CHB in the with DVA and without DVA cohorts had minimal pacing during follow-up, whereas the patients who received a PPM for sinus node dysfunction had frequent atrial pacing.
Implantation of a PPM in a transplant patient can carry additional risk as their requirement of immunosuppressive drugs can predispose them to increased risk for hardware infection. The frequent endomyocardial biopsies also predisposes them to hardware dislodgement and TR. Long-term outcome studies of patients requiring permanent pacing after OHT, however, do not report infection rates. There were no changes in the biopsy protocol within the first year after heart transplant that may have influenced TR rates and our analysis of the with and without DVA groups. Over 4 years at our institution, one patient required device explant for infectious endocarditis, but seven other post-OHT patients with PPMs did not have any complications or lead dislodgements during right heart catheterization.
DVA is primarily implemented to affect hemodynamics to prevent progressive RV dysfunction from TR. We found that throughout several time points during the first year after OHT, there was no significant difference in filling pressures or derived hemodynamic measurements. The RV function as assessed by echocardiogram was slightly worse in the with DVA group at 12 months. Our findings are in conflict with the results of previous studies on DVA in OHT with regard to a difference in TR at 1 year, perhaps because of our larger sample size. After multidisciplinary discussion and review of our analysis, the DVA procedure was removed from our surgical transplant protocol in August 2017, 2 years after its implementation. Although tricuspid annuloplasty can be performed efficiently and efficaciously,15 the perioperative and long-term benefit from a hemodynamic perspective appears minimal and comes at a risk for electrophysiologic abnormalities.
There may be a selection bias because of the retrospective nature of the analysis. The relatively small sample size may account for the differences observed between no DVA and DVA cohorts, especially because zero cases of CHB were observed early on in the no DVA cohort despite a historically expected minimum rate of CHB after OHT. Because our data are not the results of a randomized trial, they are hypothesis generating. In addition, the missing follow-up echo data may have biased outcomes against the DVA group. Upon review of the deaths, only one no DVA patient who died had severe TR. Longer follow-up might reveal a higher rate of severe TR in the no DVA cohort, as suggested by the small signal of three patients in this group who had moderate or severe TR at 1 year. Last, intraoperative cardioplegia use was not analyzed although preoperative antiarrhythmic use was not different between groups, and it may have played a role in the electrophysiologic findings.
There were increased conduction abnormalities during the first week after cardiac transplantation in patients who received prophylactic DVA. Patients who received a PPM for CHB had minimal ventricular pacing during follow-up. There was no significant difference in cardiac hemodynamics and TR at 3 and 12 months between cardiac transplant recipients with and without DVA. The DVA was discontinued from the transplant surgical protocol 2 years after its implementation at our institution. Tricuspid annuloplasty in transplant patients is a balanced trade-off between the risk of early conduction injury and the benefit of preventing late moderate/severe TR.
1. Sahar G, Stamler A, Erez E, et al. Etiological factors influencing the development of atrioventricular valve incompetence after heart transplantation. Transplant Proc 1997.29: 2675–2676.
2. Aziz TM, Saad RA, Burgess MI, Campbell CS, Yonan NA. Clinical significance of tricuspid valve dysfunction after orthotopic heart transplantation. J Heart Lung Transplant 2002.21: 1101–1108.
3. De Vega NG. La anuloplastia selectiva, regulable y permanente. Una técnica original para el tratamiento de la insuficiencia tricuspide. Rev Esp Cardiol, 1972.25: 555–556.
4. Jeevanandam V, Russell H, Mather P, Furukawa S, Anderson A, Raman J. Donor tricuspid annuloplasty during orthotopic heart transplantation: Long-term results of a prospective controlled study. Ann Thorac Surg 2006.82: 2089–2095; discussion 2095.
5. Fiorelli AI, Stolf NA, Abreu Filho CA, et al. Prophylactic donor tricuspid annuloplasty in orthotopic bicaval heart transplantation. Transplant Proc 2007.39: 2527–2530.
6. Teman NR, Wu AH, Masood M, et al. Routine use of donor heart De Vega tricuspid annuloplasty in orthotopic heart transplantation. J Heart Lung Transplant 2015.34: No 4S.
7. Marcus GM, Hoang KL, Hunt SA, Chun SH, Lee BK. Prevalence, patterns of development, and prognosis of right bundle branch block in heart transplant
recipients. Am J Cardiol 2006.98: 1288–1290.
8. Osa A, Almenar L, Arnau MA, et al. Is the prognosis poorer in heart transplanted patients who develop a right bundle branch block? J Heart Lung Transplant 2000.19: 207–214.
9. Hamon D, Taleski J, Vaseghi M, et al. Arrhythmias in the heart transplant
patient. Arrhythm Electrophysiol Rev 2014.3: 149–155.
10. Cui G, Kobashigawa J, Margarian A, Sen L. Cause of atrioventricular block in patients after heart transplantation. Transplantation 200376: 137–142, XXX.
11. Cantillon DJ, Tarakji KG, Hu T, et al. Long-term outcomes and clinical predictors for pacemaker-requiring bradyarrhythmias after cardiac transplantation: Analysis of the UNOS/OPTN cardiac transplant database. Heart Rhythm 2010.7: 1567–1571.
12. Mallidi HR, Bates M. Pacemaker use following heart transplantation. Ochsner J 2017.17: 20–24.
13. Thajudeen A, Stecker EC, Shehata M, et al. Arrhythmias after heart transplantation: Mechanisms and management. J Am Heart Assoc 2012.1: e001461.
14. Cantillon DJ, Gorodeski EZ, Caccamo M, et al. Long-term outcomes and clinical predictors for pacing after cardiac transplantation. J Heart Lung Transplant 2009.28: 791–798.
15. Akhter SA, Salabat MR, Philip JL, et al. Durability of De Vega tricuspid valve annuloplasty for severe tricuspid regurgitation during left ventricular assist device implantation. Ann Thorac Surg 2014.98: 81–83.