Outcomes of redo pulmonary valve replacement for bioprosthetic pulmonary valve failure in 61 patients with congenital heart disease ^† Free

Abstract

OBJECTIVES

Although all bioprostheses used for pulmonary valve (PV) replacement (PVR) are prone to failure and will require redo PVR, data pertaining to the outcomes of this procedure are lacking. The objective of this study was to evaluate outcomes of redo PVR for bioprosthetic PV failure in patients with congenital heart disease.

METHODS

A retrospective review of 61 patients who underwent redo PVR for bioprosthetic PV failure between November 1999 and June 2013 was performed. Univariable analyses were used to identify the factors associated with postoperative adverse events (PAEs).

RESULTS

The median age at initial PVR was 7.0 years (1.6–36.5 years) and the median age at redo PVR was 13.5 years (7.6–43.3 years). Fundamental diagnoses were tetralogy of Fallot ( n = 20), pulmonary atresia with ventricular septal defect ( n = 20), double outlet right ventricle ( n = 7) and others ( n = 14). The median valve size was 25 mm (18–28 mm). There were 2 hospital deaths (3.3%). Eighteen patients (29.5%) experienced PAEs. PAEs were associated with higher preoperative right ventricular systolic pressure (105 ± 22 vs 89 ± 19 mmHg, P = 0.016) and longer cardiopulmonary bypass time (219 ± 77 vs 164 ± 59 min, P = 0.007). Completeness of follow-up was 98.3% and the median duration of follow-up was 5.5 years (0.1–14.3 years). There were 3 late deaths. The actuarial survival rate at 10 years was 83.7 ± 8.0%. Eleven patients underwent the second redo PVR during follow-up. The rate of freedom from the second redo PVR at 10 years was 58.8 ± 11.9%. The rate of freedom from both PV reintervention and structural valve deterioration (SVD) at 10 years was 32.0 ± 13.3%.

CONCLUSIONS

A substantial number of the patients experienced mortality or morbidities after redo PVR. Higher preoperative right ventricular systolic pressure and longer cardiopulmonary bypass time were associated with PAEs. By 10 years after the redo PVR, approximately two-thirds of patients will require PV reintervention or manifest SVD.

INTRODUCTION

Repair of obstructive lesions of the right ventricular outflow tract in children usually leaves the native pulmonary valve (PV) incompetent or requires placement of a conduit between the right ventricle and the pulmonary artery. As these patients grow, many of them will require PV replacement (PVR) owing to chronic pulmonary regurgitation or a failed conduit [ 1–5 ].

Among various valve substitutes for PVR, bioprosthetic valves are currently the most widely used because they are readily available and do not need permanent anticoagulation therapy. However, durability of bioprosthetic valves in the pulmonary position is suboptimal owing to structural valve deterioration (SVD) mainly in the form of leaflet calcification [ 6–9 ]. Therefore, most of the patients who underwent bioprosthetic PVR will eventually require redo PVR over their lifetime.

Although redo PVR will be a substantial burden to congenital heart surgeons in the future as the population of patients who have undergone bioprosthetic PVR increases, data pertaining to the outcomes of this procedure are lacking. The objective of the present study was to evaluate outcomes of redo PVR for bioprosthetic PV failure in patients with congenital heart disease.

MATERIALS AND METHODS

Study population and data collection

By searching the database of Sejong General Hospital, we identified 281 patients who had undergone initial PVR using bioprosthetic valves between 1993 and 2009. From among these patients, we selected 61 patients who had undergone redo PVR due to bioprosthetic PV failure. These study patients underwent redo PVR between November 1999 and June 2013. Data were obtained by a review of the medical records and direct telephone contact. The Institutional Review Board of Sejong General Hospital approved the present retrospective study and waived the need for individual patient consent.

Indications for redo pulmonary valve replacement

Indications for redo PVR were pulmonary stenosis ( n = 36), combined pulmonary stenosis and regurgitation ( n = 17) and infective endocarditis ( n = 8). The clinical thresholds for consideration of redo PVR have somewhat varied between individual patients over the study period. The mean peak pressure gradient across the bioprosthetic PV ( n = 60) and the mean right ventricular systolic pressure ( n = 54), assessed by echocardiography or catheterization, were 62 ± 22 and 94 ± 22 mmHg, respectively.

Follow-up

Follow-up was considered complete if the patient's status was determined after June 2013. The median follow-up duration was 5.5 years (0.1–14.3 years), and 98.3% of the hospital survivors had complete follow-up.

Definitions of outcomes

Postoperative adverse events (PAEs) were defined as death, need for reoperation, stroke, myocardial infarction, renal failure, respiratory complications and postoperative arrhythmias [ 10 ]. Prolonged hospital stay was defined as a postoperative stay of more than 14 days [ 10 ]. Reintervention was defined as any surgical or percutaneous catheter procedure on the prosthetic PV. SVD was defined as the development of a peak pressure gradient equal to or greater than 50 mmHg or at least a moderate amount of pulmonary regurgitation on follow-up echocardiography.

Statistical analysis

Data collection and statistical analysis adhered to the guidelines for reporting mortality and morbidity after cardiac valve interventions [ 11 ]. Data are presented as means with standard deviations, medians with ranges or frequencies with percentages as appropriate. Univariable analyses, using the Mann–Whitney test or Fisher's exact test as appropriate, were used to identify the factors associated with PAEs. Survival and time-to-event analyses were performed using the Kaplan–Meier actuarial method. All statistical analyses were performed using the SPSS software version 18.0 (SPSS, Inc., Chicago, IL, USA).

RESULTS

Patient characteristics

The median age at initial PVR was 7.0 years (1.6–36.5 years) and 38 patients (62.3%) were males. Fundamental diagnoses were tetralogy of Fallot ( n = 20), pulmonary atresia with ventricular septal defect ( n = 20), double outlet right ventricle ( n = 7), absent PV ( n = 4) and others ( n = 10). The median valve size used for initial PVR was 19 mm (19–27 mm). Valve types used for initial PVR were Freestyle (Medtronic, Minneapolis, MN, USA; n = 17), Carpentier-Edwards Perimount (Edwards Lifesciences, Irvine, CA, USA; n = 16), Hancock II (Medtronic; n = 14), Carpentier-Edwards porcine valves (Edwards Lifesciences; n = 11) and others ( n = 3). At the time of initial PVR, 43 patients (70.5%) underwent pericardial closure using the polytetrafluoroethylene membrane (Gore Preclude Pericardial Membrane, W. L. Gore & Associates, Inc., Flagstaff, AZ, USA) for future reoperation.

The median age at redo PVR was 13.5 years (7.6–43.3 years). The mean interval between initial and redo PVR was 6.9 ± 2.6 years. The median number of prior operations and sternotomies before redo PVR was 3 (1–5) and 2 (1–4), respectively. Nineteen patients (31.1%) underwent an interventional catheter procedure for the stenotic bioprosthetic PV before redo PVR.

Operative details of redo pulmonary valve replacement

Re-entry injury occurred in 2 patients (3.3%) and peripheral cannulation for cardiopulmonary bypass was used in these patients. The median valve size was 25 mm (18–28 mm). Valve types used were Carpentier-Edwards Perimount ( n = 28), hand-sewn polytetrafluoroethylene bicuspid valves [ 12 ] ( n = 16), Hancock II ( n = 7), Epic (St. Jude Medical, St Paul, MN, USA; n = 3) and others ( n = 7). Concomitant procedures were performed in 32 patients (52.5%) and are summarized in Table 1 . The median cardiopulmonary bypass time was 162 min (75–385 min) and the median aortic cross-clamp time ( n = 34) was 89 min (16–252 min). Postoperatively, the mean systolic pressure ratio between the right and left ventricles ( n = 50) was 0.48 ± 0.13 and the median systolic pressure gradient across the PV ( n = 33) was 10 mmHg (0–20 mmHg) by direct pressure measurement.

Hospital outcomes and factors associated with postoperative adverse events

There were 2 early deaths (90-day hospital mortality rate, 3.3%). A 20-year-old female with a fundamental diagnosis of absent PV syndrome, who had undergone redo PVR owing to fungal endocarditis involving the bioprosthetic PV, died of unresolved endocarditis and ensuing sepsis 84 days after the redo PVR. A 12-year-old male with a fundamental diagnosis of absent PV syndrome underwent redo PVR owing to SVD manifested as pulmonary stenosis. After the operation, he could not be weaned from cardiopulmonary bypass owing to left ventricular dysfunction and underwent left ventricular assist device support. On the second postoperative day, while he was still on left ventricular assist device support, diffuse brain injury was detected. Although he could be weaned from left ventricular assist device on the same day, he died of unimproved brain injury 8 days after the redo PVR. The median duration of intensive care unit and hospital stay was 2 days (1–10 days) and 15 days (5–86 days), respectively. Thirty patients (50.8%) experienced prolonged hospital stay.

Eighteen patients (29.5%) experienced PAEs (Table 2 ). Reoperation was the most common PAE ( n = 12). The most common reoperation was mediastinal exploration for suspicious deep sternal infection ( n = 6). Other reasons for reoperation were superficial wound dehiscence ( n = 3), surgical bleeding ( n = 2) and pacemaker implantation due to postoperative complete atrioventricular block ( n = 1). Respiratory complications occurred in 6 patients, all of which were prolonged (more than 48 h) mechanical ventilation. Postoperative arrhythmias included complete atrioventricular block ( n = 1) and sustained ventricular tachycardia ( n = 1). One patient developed renal failure requiring haemodialysis. Patients with PAEs had a significantly longer median hospital stay [27 (8–86) vs 12 (5–56) days, P < 0.001]. In univariable analyses, PAEs were associated with higher preoperative right ventricular systolic pressure and longer cardiopulmonary bypass time (Table 3 ).

Follow-up outcomes

There were 3 late deaths during follow-up. All late deaths occurred after the second redo PVR. The causes of the late deaths were pulmonary haemorrhage developed after catheter intervention for pulmonary arterial stenosis, acute pulmonary embolism by prosthetic PV endocarditis and multiple organ failure. The actuarial survival rates at 5 and 10 years were 94.8 ± 2.9 and 83.7 ± 8.0%, respectively (Fig. 1 ). Eleven patients underwent the second redo PVR owing to SVD ( n = 7) or infective endocarditis ( n = 4). The rates of freedom from the second redo PVR at 5 and 10 years were 94.5 ± 3.1 and 58.8 ± 11.9%, respectively (Fig. 2 ). Two patients required the third redo PVR owing to infective endocarditis. Fifty-seven patients (96.6%) underwent echocardiography during follow-up and the median interval to the latest echocardiographic examination was 4.5 years (6 days–12.1 years). Two patients underwent an interventional catheter procedure for the prosthetic PV. The rates of freedom from both PV reintervention and SVD at 5 and 10 years after redo PVR were 89.2 ± 4.7 and 32.0 ± 13.3%, respectively (Fig. 3 ).

DISCUSSION

PVR is a frequently performed procedure in the field of congenital heart surgery and is the most common cardiac operation and reoperation performed in adults with congenital heart disease [ 2 , 13 ]. During the past decade, numerous studies describing outcomes of the initial PVR have been published [ 3 , 4 , 6–9 , 10 , 14 ]. However, data pertaining to the outcomes of redo PVR are lacking in the literature. These data are important because redo PVR will be a substantial burden to congenital heart surgeons in the future due to the limited durability of currently available bioprosthetic valves.

In the present study, we have reviewed our 14-year experience about redo PVR for bioprosthetic PV failure in 61 patients with congenital heart disease. The median age at redo PVR was 13.5 years and they had undergone a median of 2 sternotomies before redo PVR. Re-entry injury occurred in 3.3% of the patients and the early mortality rate was 3.3%. Approximately 30% of the patients experienced PAEs and 51% experienced prolonged (>14 days) hospital stay. The long-term survival rate was 84% at 10 years. The rate of freedom from the second redo PVR was 59% at 10 years. Importantly, the rate of freedom from both PV reintervention and SVD was 32% at 10 years.

The rate of re-entry injury and early mortality in our study are comparable to those reported in the previous studies dealing with issues of repeat sternotomy or reoperation in patients with congenital heart disease, although our study population is completely different from those of other studies [ 15–19 ]. Although Texas Children's Hospital reported excellent results on repeat sternotomy in congenital heart surgery and concluded that repeat sternotomy was no longer a risk factor for morbidity or mortality [ 16 ], other groups have reported non-negligible risks of repeat sternotomy or reoperation [ 15 , 17–19 ]. Kirshbom et al. [ 17 ], in their study of 1000 repeat sternotomies for congenital cardiac surgery performed in patients with a median age of 2.1 years, reported that increase in the number of sternotomies and the presence of a right ventricle–pulmonary artery conduit were risk factors for re-entry injury. However, re-entry injury was not associated with an increased risk of operative mortality. Holst et al. [ 19 ], in their study of 984 adults with congenital heart disease who had undergone repeat sternotomy, also reported that increased number of prior sternotomies was an independent predictor of re-entry injury. Again, re-entry injury was not a risk factor for operative mortality in this study. In our study, the 2 patients who had experienced re-entry injury survived the operation without sequelae. In accordance with previous studies, we believe that properly managed re-entry injury does not necessarily translate into increased risk of operative mortality. Of note, the 2 patients who experienced re-entry injury did not undergo pericardial closure at the end of the first PVR. Currently, we liberally use the polytetrafluoroethylene membrane as a pericardial substitute if reoperations are expected in the future. In general, reoperation for congenital heart disease can be performed with low operative mortality (1.8–3.6%) as reported in previous studies [ 15–19 ] and ours. However, it has been shown that increase in the number of sternotomies was associated with increase in operative mortality [ 13 , 17 , 20 ]. The median number of sternotomies for early and late mortality cases in our study ( n = 5) was 4 (3–6). All late deaths occurred after the second redo PVR. It is not known whether this increased risk is related to the technical challenges of safe entry and dissection after so many previous sternotomies, or whether it is more the case that many previous operations are a marker for complex heart disease with residual haemodynamic burdens that may affect myocardial function and even other organ systems [ 13 ].

A substantial portion (30%) of our study population experienced PAEs including early deaths. Although it seems that the complication rate of our series is high, the difference in study population and the diversity in the definition of complication make it difficult to compare our results with those of other studies [ 10 , 18 , 19 , 21–23 ]. In the two previous studies which dealt with issues regarding postoperative morbidity after the first-time PVR in adults with repaired tetralogy of Fallot, the complication rate was 40 and 14%, respectively [ 10 , 21 ]. The two most common complications observed in our series were reoperation (20%) and prolonged mechanical ventilation (10%). Most of the reoperations were due to sternotomy wound problems (infection or superficial dehiscence) or postoperative bleeding, both of which are commonly regarded as complications associated with repeat sternotomy. Giamberti et al. [ 18 ], in their study of 164 adults with congenital heart disease who had undergone cardiac reoperations, reported that postoperative morbidity was associated with the number of previous operations. The rather strict definition of prolonged mechanical ventilation used in our study (>48 h) might have contributed to the seemingly high rate of this complication. Furthermore, because no predefined protocol in the postoperative care of our study population existed in our centre, the duration of mechanical ventilation might have been influenced by individual practice. In our series, 4 of the 6 patients who have experienced prolonged mechanical ventilation recovered without any other complications or sequelae. Patients with PAEs had a significantly longer hospital stay. It is a common finding in the literature and self-explanatory that the occurrence of postoperative complications after congenital heart surgery is associated with prolonged hospital stay [ 10 , 23 ]. Dos et al. [ 10 ], in their study of 116 adults with repaired tetralogy of Fallot who had undergone the first-time PVR, reported that the number of previous sternotomies was one of the independent risk factors for prolonged hospitalization.

We found that PAEs were associated with higher preoperative right ventricular systolic pressure and longer cardiopulmonary bypass time. Longer cardiopulmonary bypass time is a well-known independent risk factor for postoperative complications after congenital heart surgery [ 18 , 22 , 23 ]. The seemingly long cardiopulmonary bypass time in our series was probably due to the frequently performed concomitant procedures (53%) and technical difficulties resulting from reoperation (median third sternotomy). Because higher preoperative right ventricular systolic pressure was associated with PAEs, minimizing the duration of right ventricular pressure overload might be important in preventing PAEs and thus improving long-term survival. Defining the optimal timing of reintervention on the failing bioprosthetic PV is beyond the scope of the present study, but should be an important topic of future studies.

Long-term outcomes of our study population were far from satisfactory with 84% of survival and only 32% of freedom from PV reintervention and SVD at 10 years. Increase in the population of the patients with bioprosthetic PV will certainly be a substantial burden to congenital heart surgeons in the future. To improve long-term outcomes, continuing efforts to search for durable valve substitutes are necessary [ 12 , 24 ]. Recently, percutaneous PV implantation has emerged as an alternative treatment option for failed bioprosthetic valves [ 25 ]. Although this technique is expected to reduce the number of operations during the lifetime of the patients who have undergone bioprosthetic PVR, the impact this evolving technique will have on the long-term outcomes remains to be determined.

Limitations

The present study was limited by its retrospective nature. Analyses to identify independent risk factors for re-entry injury, mortality, PAEs, PV reintervention and SVD could not be performed due to insufficient numbers of patients with those events. Because no strict protocol in the postoperative care of our study population existed in our centre, the duration of mechanical ventilation, intensive care unit stay and hospital stay might have been influenced by individual practice or temporal changes in practice during the study period of 14 years.

CONCLUSIONS

A substantial number of the patients experienced mortality or morbidities after redo PVR for bioprosthetic PV failure. Higher preoperative right ventricular systolic pressure and longer cardiopulmonary bypass time were associated with PAEs. By 10 years after the redo PVR, approximately two-thirds of patients will require PV reintervention or manifest SVD.

Conflict of interest: none declared.

REFERENCES

Author notes