Anatomic Features and Surgical Strategies in Double-Outlet Right Ventricle
Background The objective of this study was to review anatomic features and surgical strategies in children with double-outlet right ventricle (DORV) and to assess risk factors for early mortality.
Methods and Results Records were reviewed of all children with DORV undergoing surgery between 1978 and 1993. Noncomplex patients (group 1) had atrioventricular (AV) concordance, a single ventricular septal defect (VSD), balanced ventricles, no straddling AV valves, and no major pulmonary artery anomaly. Group 2 (complex) comprised all remaining patients. Independent risk factors analyzed included location of the main VSD, presence of additional VSDs, coarctation, ventricular outflow obstruction, ventricular hypoplasia, age at operation, operation before 1985, previous palliation, and type of definitive operation. Of 193 patients, 117 were in group 1 and 76 in group 2. In 148 patients, biventricular repair was undertaken, including 111 of 117 group 1 patients and 37 of 76 group 2 patients. Early mortality was higher among group 2 patients undergoing biventricular repair than among group 1 patients (8 of 37 versus 4 of 111, P<.005) and higher than group 2 patients undergoing a Fontan procedure (none of 29, P<.01). Aortic arch obstruction, operation before 1985, and multiple VSDs were significant risk factors for mortality. Age <1 month (P<.05) and multiple VSDs (P<.005) were independent risk factors after definitive repair. Up-to-date follow-up is available on 144 surviving patients, with 127 (88%) in New York Heart Association class I and the remaining 17 in class II. Overall 10-year survival probability was 81%, whereas probability of survival, free from reoperation (after definitive surgery), was 65% at 10 years.
Conclusions Biventricular repair can be achieved in most patients with DORV with low risk. In complex DORV, a Fontan procedure is associated with a lower surgical mortality.
By describing a ventriculoarterial connection rather than a specific congenital malformation, double-outlet right ventricle (DORV) includes a broad spectrum of anatomic variants and associated malformations. In 1964, Kirklin et al1 reported successful correction in a child with DORV, subaortic ventricular septal defect (VSD), and concordant atrioventricular (AV) connections. Since then, complete correction through the use of a variety of surgical techniques has been achieved in more complex forms of DORV. Modified Fontan procedures have been advocated when biventricular DORV repair has been either impractical or extremely complex.2 With the improving short-term and medium-term outcome for Fontan procedures in recent years,3 4 5 this approach might also be extended to patients who are at increased operative risk with a conventional biventricular repair. This report reviews the anatomic findings, surgical strategies, and results among patients with DORV presenting for surgery in a single institution over a 15-year period.
The study population comprised all patients with a diagnosis of DORV who did not have a univentricular AV connection and who underwent a surgical procedure between 1978 and 1993. A diagnosis of DORV was made if both great arteries originated predominantly from the right ventricle with application of the “50% rule,” which requires that one great artery arises completely and the second >50% from the right ventricle. The diagnosis and anatomic findings were based on a combination of angiography, echocardiography, and surgical inspection. Group 1 patients, with noncomplex anatomy, were defined as those with AV concordance, a single VSD, balanced ventricles, no straddling AV valves, and no major pulmonary artery anomaly. Group 2 patients, with complex anatomy, consisted of all the remaining patients. Complex anatomy was defined by the presence of atrial isomerism, AV discordance, atrioventricular septal defect, multiple VSDs, hypoplastic right or left ventricles, straddling AV valves, and major pulmonary artery anomalies including pulmonary atresia, pulmonary artery sling, aorto-pulmonary window, or discontinuous branch pulmonary arteries. Early mortality was defined as death on the same hospital admission as the operation, or within 30 days of operation for patients discharged from hospital. Follow-up information was obtained from hospital files for locally followed patients and from referring cardiologists for patients followed elsewhere.
Outcome measures included early survival and freedom from reoperation. To assess potential risk factors for early mortality and reoperation, univariate analysis with Fisher’s exact test was initially used for each variable. Variables with a value of P<.1 were then entered into a stepwise logistic regression analysis. A commercially available statistic program was used (StataCorp, 1995. Stata Statistical Software: Release 4.0). Variables were considered statistically significant in the multivariate analysis at P<.05. Odds ratios and 95% confidence intervals are stated. Survival and freedom from reoperation probabilities were estimated with the Kaplan-Meier method. t tests were used to compare normally distributed continuous variables and Pearson’s χ2 analysis or Fisher’s exact test to compare discrete variables. Values are expressed as mean±SD unless otherwise stated.
A total of 193 patients who underwent surgical procedures comprised the study population. Four types of definitive repairs were performed: (1) intraventricular repair with a baffle from the left ventricle to the aorta (IVR), (2) arterial switch operation with a baffle from the left ventricle to the pulmonary artery (ASO), (3) atrial inversion using the Senning operation, and (4) cavopulmonary shunt/Fontan procedures including bidirectional cavopulmonary connection, modified Fontan procedures, total cavopulmonary connection, and the Kawashima operation. The mean age at definitive repair was 3.2±4.1 years and ranged from 1 day to 30.2 years, with a median of 1.9 years.
Complex versus noncomplex anatomy and associated anomalies. Complex anatomic features among group 2 patients are listed in Table 1⇓. Of the 76 (39%) patients in group 2, the most commonly present complex anatomic feature was the presence of additional VSDs (24 patients), followed by hypoplastic right or left ventricle (20 patients) and atrioventricular septal defect (18 patients). Associated noncomplex anomalies present in group 1 and 2 patients are listed in Table 2⇓.
VSD type. The location of the primary VSD and the presence of additional VSDs are shown in Table 3⇓. Of the 90 patients in whom the main VSD was subaortic, 50 had tetralogy of Fallot–type anatomy. A subpulmonary VSD was present in 49 patients. The main VSD was not committed to either outflow tract in 31 patients (“noncommitted”). This designation was used in all cases in which the VSD was anatomically distant from both arterial valves and was not related more directly to one outflow tract than to the other. An additional 18 patients had a complete atrioventricular septal defect. In 5 patients, all in group 1, the VSD was committed to both outflow tracts. There was no statistical difference in the primary VSD location between patients in group 1 and group 2. In three cases, all with subaortic VSDs, the VSD was restrictive.
Outflow tract obstruction. Pulmonary outflow tract obstruction (POTO) not associated with tetralogy of Fallot but including pulmonary atresia was present in 45 patients and was most prevalent in the group with subaortic VSDs (17 of the remaining 40 patients, 42%). Eleven of 31 (35%) of children with noncommitted VSD had POTO, as did 2 of 5 (40%) with doubly committed subarterial defects and 7 of 18 with atrioventricular septal defects (39%). Only 8 of 49 patients (16%) with subpulmonary VSD had valvar or infundibular pulmonary stenosis. The majority of patients (64%) with pulmonary outflow obstruction belonged to group 2. Aortic or subaortic stenosis was seen in 6 patients (3%), of whom 2 belonged to group 1 and 4 to group 2. Of the 6 patients with left ventricular outflow obstruction, 3 had subpulmonary VSDs, 2 had subaortic, and 1 a noncommitted VSD. Aortic arch obstruction coexisted with left ventricular outflow tract obstruction in 3 patients. Overall, aortic arch obstruction (aortic coarctation or interruption) occurred in 26 (13%) of patients, of whom 14 had subpulmonary, 7 had subaortic, and 5 had noncommitted VSDs. The majority (17 of 26, 65%) of patients with aortic arch obstruction had noncomplex anatomy (group 1).
Prior palliative procedures. Definitive repair was the primary procedure in 92 of 193 (48%) patients. Table 4⇓ summarizes the details of prior palliative procedures undertaken in the remaining 101 patients. Of these, 77 patients required only one palliative procedure, 19 had two procedures, 4 patients had three, and 1 patient required four palliative procedures. The proportion of group 2 patients requiring palliation before definitive surgery (50 of 76 or 65.8%; CI, 54.0% to 76.3%) was significantly higher than that of group 1 patients (51 of 117 or 43.6%; CI, 34.4% to 53.0%; P<.005).
Definitive surgery. One hundred seventy-nine patients have so far undergone definitive surgery. Of the remaining 14 patients, 5 await definitive surgery and 9 died before repair could be undertaken. At definitive surgery, a biventricular repair was attempted if considered possible. Table 5⇓ summarizes the details of the definitive surgical procedure according to the VSD site. In 148 patients a biventricular repair was undertaken, with an intraventricular baffle in 107 cases, the arterial switch operation as a primary procedure in 31, and a Senning procedure in 10 (all in the early part of the series−before introduction of the arterial switch). In 4 of these Senning patients conversion to an arterial switch operation was subsequently performed because of increasing effort intolerance late after surgery because of either right ventricular dysfunction or tricuspid incompetence. Three of these four patients are long-term survivors, two being in New York Heart Association class I and the third in NYHA class II. In 3 of the 6 other patients undergoing a Senning procedure, the VSD was not closed because of the presence of pulmonary vascular disease. One patient died after his Senning operation because of intractable problems related to multiple VSDs (being converted to a Fontan shortly before he died). Excluding this latter patient, 31 modified Fontan-type procedures were performed for definitive palliation. These included 29 group 2 patients who had anatomic features (eg, severely hypoplastic ventricle or straddling AV valve) precluding a biventricular repair. Two patients in group 1, both with noncommitted VSDs, also underwent modified Fontan procedures.
Children underwent the ASO at a significantly younger age than the intraventricular baffle operation (1.4±2.9 years versus 2.9±3.4 years; P<.05) and the modified Fontan procedures (5.7±5.8 years; P<.005).
Early mortality. Table 6⇓ summarizes the details of 18 (9.3%; CI, 5.6% to 14.4%) hospital deaths, 6 of which occurred after a palliative procedure and 12 after definitive surgery. A significantly higher proportion of group 2 patients (8 of 37 or 21.6%; CI, 9.8% to 38.3%) died after a biventricular repair than did group 1 patients (4 of 111 or 3.6%; CI, 1.0% to 9.0%; P<.005). In contrast, there were no hospital deaths among the 29 group 2 or the two group 1 patients undergoing a cavopulmonary shunt or modified Fontan procedure (0%; CI, 0% to 11.2%). Table 7⇓ presents the results of univariate and multivariate risk factor analysis for early mortality. Of the potential risk factors analyzed in all 193 patients (including those undergoing palliative procedures), only the presence of multiple VSDs, aortic arch obstruction, and operation before 1985 reached near significance in univariate analysis (Table 7⇓, A) and were significant risk factors for early mortality in multivariate analysis (Table 7⇓, B). In the analysis of the 179 patients who survived to definitive repair, the presence of multiple VSDs and age at operation <1 month were significantly associated with early mortality, both in univariate and multivariate analysis (Table 7⇓).
Late deaths. Table 8⇓ summarizes the details of late deaths. There were 9 late deaths, 2 in group 1 and 7 in group 2 at a mean follow-up time of 4.7±3.8 years. Three of the 9 patients had undergone only a palliative procedure, 5 had undergone biventricular repair, and 1 patient a Fontan-type procedure. Fig 1⇓ shows the overall survival probability with 90% confidence intervals for the entire study cohort, with an estimated 10-year survival probability of 81%.
Reoperations and follow-up information. Follow-up information was available in 144 of 166 surviving patients. During the period of follow-up, 25 patients underwent 29 reoperations at a mean time interval of 3±3.6 years (range, 6 days to 14 years) after definitive repair. Of these, 14 patients belonged to group 1 and 11 to group 2. The initial anatomic features and the type of reoperation are summarized in Table 9⇓. Risk factor analysis for reoperation is presented in Table 10⇓. There were no statistically significant variables associated with reoperation, with only the presence of left ventricular outflow tract obstruction present having a value of P<.10. The probability of freedom from death and reoperation with 90% confidence intervals is depicted in Fig 2⇓. The estimated freedom from death and reoperation at 10 years was 65%.
Additionally, 10 patients underwent pacemaker implantation, 8 for postoperative complete heart block, 1 as prophylaxis, and the remaining patient because of frequent episodes of atrial flutter after a previous Senning operation and a subsequent arterial switch conversion.
Of the 144 patients available for long-term follow-up, 127 (88%) were in NYHA class I and 17 (12%) in class II. Twenty-three had mild and 4 moderate residual right ventricular outflow tract obstruction. Residual left ventricular outflow tract obstruction was present in 10 patients. In 7 of these it was classified as mild and in 3 as moderate.
This study confirms that a biventricular repair can be achieved in the great majority of patients with DORV with an acceptably low perioperative mortality. Previous reviews have reported hospital mortality figures of up to 27%, especially in cases with noncommitted or subpulmonary VSDs.6 7 8 9 Since then, new surgical techniques including the arterial switch operation10 11 have significantly improved the outcome in noncomplex forms of DORV.8 11 12 13
VSD Site and Choice of Surgical Procedure
Although the classification of DORV by VSD site as introduced by Lev and coworkers14 remains useful for surgical planning, the present study has not demonstrated a significantly higher risk in patients with noncommitted, subpulmonary, or atrioventricular septal defects, as has been reported in earlier series.6 9 15 In our institution, the arterial switch operation is now the procedure of choice for patients with subpulmonary VSD without significant right ventricular outflow obstruction as well as in other cases in which baffling from the left ventricle to the pulmonary artery is technically easier than baffling to the aorta. In the present series, the arterial switch procedure for DORV with subpulmonary VSD was associated with a low hospital mortality. On the other hand, an intraventricular baffle repair without arterial switch could only be performed in a small number of cases with subpulmonary VSD. When technically possible, this approach has led to good results with low operative mortality and no increased need for reoperation.13 16 17 The tricuspid–to–pulmonary annulus distance has been shown to be a useful predictor for the feasibility of this type of repair.12 13 17
Among patients with noncomplex forms of DORV and noncommitted VSD, an intraventricular baffle repair was still possible in the majority of cases. All VSDs in this group were “anatomically” rather than “surgically” noncommitted,18 with no case in which interposed tension apparatus of the atrioventricular valves prevented biventricular repair. In this series, none of the patients with complex forms of DORV who had an associated atrioventricular septal defect died after definitive repair. However, in more than half of such cases a cavopulmonary shunt or modified Fontan operation was chosen because of associated atrial isomerism with some degree of ventricular imbalance.
Risk Factors for Hospital Mortality
Aortic arch obstruction. Aortic arch obstruction was a significant risk factor for hospital mortality in the analysis of all patients combined and was an associated finding in more than half of patients who died after palliative procedures. The majority of patients with aortic arch obstruction underwent surgical arch repair as a separate palliative procedure. Recently, better survival probability has been reported in children with aortic arch obstruction and associated complex intracardiac lesions if repair has been undertaken as a one-stage procedure19 20 21 ; this is now our preferred approach in noncomplex forms of DORV with associated aortic arch obstruction that includes hypoplasia of the aortic arch proximal to the left subclavian artery or complete interruption. When a localized coarctation is present, this would normally be dealt with through a left thoracotomy as a separate procedure.
Multiple ventricular septal defects. By multivariate risk factor analysis of all patients surviving to definitive surgery, the presence of multiple ventricular septal defects was the only factor related to early mortality. Similar results were reported by Aoki et al13 in an analysis of patients with DORV undergoing biventricular repair. In most cases the presence of additional VSDs does not preclude successful biventricular repair. However, complex forms of DORV remain a challenge, and the optimal surgical approach needs to be individualized according to the specific anatomic features present, including the number and site of any additional VSDs. The advent of newer surgical and interventional catheter techniques22 23 may further facilitate biventricular repair and improve the results in this group of patients.
Early surgical repair. Patients who underwent definitive repair before 1 month of age were also at significantly higher risk in the present series, but this group tended to include patients with unfavorable anatomic variants. Early definitive repair is seldom required in most patients with DORV and is probably best avoided in patients with complicating anatomic features.
Role of the Modified Fontan Procedure
The use of a cavopulmonary shunt or modified Fontan procedure has previously been advocated as the surgical procedure of choice in the presence of complex anatomic features such as straddling AV valves or ventricular imbalance. In recent years, surgical modifications with staged procedures and atrial baffle fenestration have been associated with a considerable improvement in the short-term and medium-term results for the modified Fontan operation.3 4 5 In the present series there was no hospital mortality among any patients undergoing a cavopulmonary shunt or modified Fontan procedure as the definitive surgical procedure for complex forms of DORV. The difference in early mortality between patients with complicating anatomic features undergoing biventricular repair and those subjected to some form of Fontan procedure strongly suggests that the latter is the procedure of choice, particularly among those with surgically inaccessible multiple VSDs.
Late Results and Requirements for Reoperation
Arrhythmias have been recognized as a cause for late deaths in patients after DORV repair with increasing surgical age, postoperative ventricular tachyarrhythmias, and complete heart block as significant risk factors.7 Indications for permanent cardiac pacing should therefore include the development of complete or prolonged postoperative heart block, which is often poorly tolerated in the presence of even a minor residual hemodynamic disturbance.
The 5- to 10-year survival and freedom from reoperation in the present series compare favorably with other reported series in which patients with complex DORV have not always been included in the analysis.13 24 It has been suggested that patients with subaortic and noncommitted VSD are at low and high risk, respectively, for reoperation.13 In the present series there were no statistically significant associations with reoperation, including the site of the primary ventricular septal defect and other complicating anatomic features.
The classification of varying anatomic DORV subtypes into simple and complex forms has, to some extent, been arbitrary. However, this had no bearing on risk factor analysis, in which anatomic variables were considered separately, regardless of whether the anatomy was regarded as complex or noncomplex. This study has the inherent limitations of a retrospective review in which patients were not randomly allocated to varying treatment groups. Risk factor analysis is, by its nature, predicated on outcome, which may in turn be influenced by varying treatment options. Despite this, the lower hospital mortality among patients with complex DORV who underwent a modified Fontan procedure rather than an attempt at biventricular repair indicates that the latter is a worthwhile option in this difficult subset of patients.
The surgical procedures in a cohort of consecutive patients with DORV and a wide range of anatomic features have been examined. Early mortality was low among patients with noncomplex forms of DORV undergoing biventricular repair but was higher in patients with complicating anatomic features undergoing similar surgery. Risk factors for early mortality identified from multivariate analysis include aortic arch obstruction, the presence of multiple VSDs, and definitive surgical repair before 30 days of age. A modified Fontan procedure is the treatment of choice for many patients with complex forms of DORV, particularly in the presence of multiple ventricular septal defects.
The authors wish to thank Dr John Carlin for assistance with statistical analysis.
Presented in part at the 44th Annual Scientific Session of the American College of Cardiology, New Orleans, La, March 1995.
- Received October 2, 1996.
- Revision received March 12, 1997.
- Accepted March 18, 1997.
- Copyright © 1997 by American Heart Association
Kirklin JW, Harp RA, McGoon DC. Surgical treatment of origin of both vessels from right ventricle, including cases of pulmonary stenosis. J Thorac Cardiovasc Surg. 1964;48:1026-1036.
Russo P, Danielson GK, Puga FJ, McGoon DC, Humes R. Modified Fontan procedure for biventricular hearts with complex forms of double-outlet right ventricle. Circulation. 1988;78(suppl III):III-20-III-25.
Pearl JM, Laks H, Drinkwater DC, Capouya ER, George BL, Williams RG. Modified Fontan procedure in patients less than 4 years of age. Circulation. 1992;86(suppl II):II-100-II-105.
Shen WK, Holmes DR Jr, Porter CJ, McGoon DC, Ilstrup DM. Sudden death after repair of double outlet right ventricle. Circulation. 1990;81:128-136.
Vogt PR, Carrel T, Pasic M, Arbenz U, von Segesser LK, Turina MI. Early and late results after correction for double-outlet right ventricle: uni- and multivariate analysis of risk factors. Eur J Cardiothorac Surg. 1994;8:301-307.
Luber JM, Castaneda AR, Lang P, Norwood WI. Repair of double-outlet right ventricle: early and late results. Circulation. 1983;68(suppl II):II-144-II-147.
Serraf A, Lacour-Gayet F, Bruniaux J, Losay J, Petit J, Touchot-Kone A, Bouchart F, Planche C. Anatomic repair of Taussig-Bing hearts. Circulation. 1991;84(suppl III):III-200-III-205.