P with very small pulmonary arteries (PAS), arborization

Very Small Pulmonary Arteries: Central End-to- Side Shunt Kevin G. Watterson, FRACS, James L. Wilkinson, FRCP, Tom R. Karl, MS, MD, and Roger B. B. Mee, FRACS Cardiac Surgery and Cardiology Units, Royal Children s Hospital, Melbourne, Australia Between 980 and 989, 8 patients with pulmonary atresia, ventricular septal defect, and very small pulmonary arteries with major aortopulmonary collateral arteries underwent direct central end-to-side shunts as staging procedures. Age range was months to years, with 9 patients less than year of age. Pulmonary artery diameters ranged from to mm, with less than mm. Two patients (7%; 70% confidence limits, % to 6%) died after the shunt, and there were two further deaths after subsequent staging or correction. Acute shunt complications included congestive cardiac failure (mild to moderate, n = 8; severe, n = ) and endocarditis (n = ). Proximal right pulmonary artery stenoses have occurred in 75% of patients and left pulmonary artery stenoses, in.50%. Satisfactory pulmonary artery growth was achieved, however, in 6 of hospital survivors investigated postoperatively. Twelve patients have proceeded through unifocalization to biventricular repair (in the ventricular septal defect patch was subsequently fenestrated) with one death (8.%; 70% confidence limits, % to 5%). Eight patients are still in staging, and have been excluded from the program because of inadequate unifocalization. The direct central end-to-side shunt has proven satisfactory in attaining pulmonary artery growth in patients with very small central pulmonary arteries. (Ann Thorac Surg 99;5:-7) ulmonary atresia plus ventricular septal defect (VSD) P with very small pulmonary arteries (PAS), arborization defects, and major aortopulmonary collateral arteries (MAPCAs) is a complex congenital malformation that is not amenable to a primary corrective procedure. Where a diminutive PA system is present, staging procedures have been employed in an attempt to attain PA growth. These procedures include various types of shunts as well as right ventricular outflow tract reconstruction (-9. At the Royal Children s Hospital, a direct central end-to-side shunt has been developed as a staging procedure for a subgroup of patients with very small PAS. Herein, the sequelae of this shunt procedure are reviewed. Patients and Methods Patients Between 980 and 989, 8 patients with pulmonary atresia plus VSD, MAPCAs, and very small PAS (usually with arborization defects) underwent a central end-to-side shunt as a staging procedure. Age range was months to years, with 9 patients less than year of age (Table ). The central shunt was the initial staging procedure in 7 of 8 patients, patient having previously undergone a right modified Blalock shunt. Twenty-six of 8 patients had a PA system consisting of a short main PA with right pulmonary artery (RPA) and left pulmonary artery (LPA) in continuity. Two patients had no LPA, so the shunt was between the ascending Accepted for publication June, 99. Address reprint requests to Mr Mee, Royal Children s Hospital, Flemington Road, l arkville, 05, Australia. aorta and RPA. Pulmonary artery diameter was estimated at preoperative angiography and also at operation (Table ). Technique A median sternotomy was used in of 8 patients, and a left thoracotomy in the remaining 5. A left thoracotomy was used if it was thought that both the central shunt and left sided unifocalization (MAPCA transplantation or ligation or both) could be performed simultaneously. Cardiopulmonary bypass support was required in very hypoxic patients. The ascending aorta and the diminutive main and branch PAS were mobilized, and Silastic (Ethicon, Sornerville, NJ) snares were placed around left and right PAS. Heparin ( mg/kg) was given intravenously, and the left side of the ascending aorta was side-clamped as posteriorly as possible. A button of aortic wall was excised and the mobilized main pulmonary artery ( fishmouthed on the right side) was anastomosed end-to-side to the aorta using continuous 7-0 polypropylene. The clamp was released and the heparin was not reversed (Fig ). Results Mortality Two patients died after the central shunt (7%; 70% confidence limits [CL], % to 6%). There were two further deaths associated with subsequent unifocalization or correction. The first death occurred in a patient with an absent LPA who also had a somewhat hypoplastic left ventricle. Sudden death occurred on the third postoperative day, presumably owing to an arrhythmia. The second 0 99 by The Society of Thoracic Surgeons 000-975/9/$.50

Ann Thorac Surg 99;5-7 WATTERSON ET AL Table. Age Distribution of Patients With Central End-to- Side Shunt Age (Y) No. of Patients 0-6 mo & mo - y - y -Y 6- y G5 y Total 8 0 5 8 patient, months old, had severe cyanosis. An emergency peripheral shunt was abandoned because the RPA was too small, and a central shunt using a diminutive main pulmonary artery was performed. A low output, hypoxic, and acidotic state persisted and the child died 8 hours postoperatively. The small PAS, and probably inadequate MAPCAs, constituted an anatomical basis for the failure of the procedure. There were two further deaths after subsequent procedures. The first late death was due to a mesenteric embolus after a fourth preparatory operation (open mitral valvotomy for mitral stenosis). The second late death occurred in a -month-old patient with episodic severe hypoxia, who had major hemodynamic instability after a central shunt. He had required emergency reexploration in the early postoperative period, at which time the shunt diameter was reduced with vascular clips. This had the effect of severely narrowing the LPA and led to a sequence of attempted repair and then fenestration of the VSD patch because of systemic right ventricular pressures. This patient also had severe tracheobronchomalacia, which eventually led to management being withdrawn months postoperatively. Shunt Complications Early congestive cardiac failure related to excessive pulmonary blood flow developed in patients. Failure was mild to moderate with no serious hemodynamic instability in 8 patients and responded to antifailure therapy within 8 hours. These patients were all discharged for further investigation and staging. Severe failure devel- Fig. Central end-to-side shunt. The main pulmonary artery is anastornosed to the ascending aorta as posteriorly as possible. oped in patients, who remained ventilator dependent postoperatively. Their staging was accelerated, and all underwent unifocalization in the weeks after shunting and correction within a month without operative mortality, but with one late death months after repair, related to severe bronchomalacia (Figs ). Infection A false aneurysm at the anastomosis, related to infective endocarditis, developed in patient. Intervention was required 7 months after the initial shunt. The RPA had grown to 8 mm in diameter and the LPA to 6 mm in diameter from initial diameters of.5 mm. Therefore the false aneurysm, which had destroyed.5 cm of the left wall of the ascending aorta, was completely excised. The aorta was oversewn, the branch PAS were reconstructed, Table. Pulmonary Artery Diameter Distribution PA Diameter (mm) No. of Patients -.9 7 -.9 5 -.9 Absent left PA, right PA - mm Total 8 PA = pulmonary artery. Fig. Pulmonary artery wedge injection outlining the pulmonary arteries before central shunt in a -month-old child.

WATTERSON ET AL Ann Thorac Surg 99;5: -7 Table. Procedures After Central End-to-Side Shunt in.6 Survivors Patient Procedure No. No. Procedures Fig. Angiogram of same patient as in Figure three weeks after the central shunt demonstrating substantial pulmonary artery growth (- mm to 7.5 mm). and total correction was undertaken using an 8-mm valved homograft conduit. Six months after correction the child is asymptomatic and free of apparent infection. Further Stagingl Twelve patients have progressed to full correction (with the VSD patch subsequently requiring fenestration in ) with one operative death (8.%; CL, % to 5%) (Table ). Eight patients are still in staging. Future repair is anticipated in patients who meet criteria for correction (projected ratio of right ventricular to left ventricular pressure <0.7; no remaining sizeable MAPCAs; > % of lung segments connected to the native PAS, and a pulmonary to systemic flow ratio > ) [lo]. Three other patients have had problems identified with regard to these criteria at postshunt catheterization. Suitability for final correction has therefore not yet been attained. One patient is still in staging and has not yet had postshunt investigations. Fig. Pulmonary artery injection in same patient as in Figures and ajter total correction. 5 6 7 8 9 0 5 6 7 9 0 L thoracotomy ML R thoracotomy MT, ML L thoracotomy MT, ML R thoracotomy MT, ML L thoracotomy MT, ML R thoracotomy MT, ML L thoracotomy MTl, ML L thoracotomy central shunt, MT R thoracotomy MT L thoracotomy ML R thoracotomy MT R thoracotomy MT, ML R thoracotomy MT L thoracotomy MT, ML L thoracotomy central shunt, ML R thoracotomy MT, ML L thoracotomy central shunt, ML R thoracotomy MT, ML L thoracotomy central shunt, ML R thoracotomy MT nd sternotomy aorta-rpa (direct) R thoracotomy MBTS rd sternotomy central PA patch reconstruction, central Gore-Tex shunt L thoracotomy MT R thoracotomy ML No further procedure nd sternotomy aorta-rpa (direct) R thoracotomy MAPCA-aorta L thoracotomy MT, ML R thoracotomy MT L thoracotomy central shunt, ML, for further staging

~~~~ ~ Ann Thorac Surg 99;5: -7 WATTERSON ET AL 5 Table. Continued Patient Procedure No. No. Procedures 5 6.. 8 R thoracotomy MT L thoracotomy MTI, ML VSD fenestration after repair No further procedure R thoracotomy MT L thoracotomy MT nd sternotomy mitral valvotomy, atrial septectomy L = left; MAPCA = major aortopulmonary collateral artery; MBTS = modified Blalock-Taussig shunt; ML = MAPCA ligation; MT = MAPCA transplantation; PA = pulmonary artery; R = right; RPA = right pulmonary artery; VSD = ventricular septal defect. Four patients have been excluded from the program for the present. One patient had an absent LPA and a central end-to-side shunt to the RPA. Postoperative catheterization showed poor RPA growth. Another patient originally underwent a central end-to-side shunt to a very small ( to.5 mm) PA system. Recatheterization showed a blocked RPA, and at reoperation a direct end-to-side RPA to ascending aorta anastomosis was performed. Further catheterization studies showed that both PAS were still very small. The third patient underwent further staging with a right thoracotomy and MAPCA transplantation using an azygos vein conduit [ll]. Recatheterization showed severe long-segment LPA hypoplasia and systemic pressures in the RPA to the point of the MAPCA transplant. Pulmonary vascular disease is probably present in the right lung. The fourth patient, years of age with progressive cyanosis at the time of the central shunt, was much improved by the operation and does not wish to have a further operation. Pulmonary Artery Growth The PAS have been assessed using cardiac catheterization in 5 ( corrected, uncorrected) of the 5 survivors of initial hospitalization (Table ). Proximal RPA stenosis has been identified in 8 patients (7%; CL, 60% to 8%) and proximal LPA stenosis has been seen in patients (8%; CL, 6% to 60%). Overall satisfactory PA growth, as Table. Pulmonary Artery Growth After Central Shunt'' RPA (mm) LPA (mm) Age at Central Shunt Pre Shunt Post Shunt Pre Shunt Post Shunt Time (mo) Progress y6mo Occl.5 0 Staging 8 mo 8 8 6 Staging 8 mo Occl Staging 9 mo.5...... No cath Staging Y 6 6 6 Staging 0 mo 9 Staging mo.5 5 Staging 9 mo.5 5.5.5 8.5 9 Staging mo 7 Not for correction Y.5 8.5 8 6 Not for correction 0 Y.5.5 Absent... Not for correction mo Not for correction y6mo.5 5.5.5 7 7 Late death 6y0mo 8 8 6 y8mo 5 0 9 Crorection mo 0.5 0 9 mo 5.5 5.5 7ymo - - 9 Y - 0.5 5 0 0 mo 7.5 7.5 mo.5 7.5 7.5 yllmo.5 8.5 6 7 7 mo 5 8 mo 8.5 8.5 0 8 Y 6 ivsd patch fenestrated a Pulmonary artery diameters measured between mid-pulmonary artery and hilum. Cath = catheterization; LPA = left pulmonary artery; Occl = occluded; RPA = right pulmonary artery; VSD = ventricular septal defect.

6 WA'ITERSON ET AL Ann Thorac Surg 99:5:-7 Table 5. Pulmonary Artery Pressures After Central Shunt Patient PAP Interval PA Size (mm) No. (mm Hg) Progress (mo) RPA LPA 55/ 6 8 8 59/6 9 5 0 LPA 8 5.5 5.5 5 608 9 6 0 0 0 7 8/ 7.5 7.5 66 5 8 7/8 0 8.5 8.5 RPA 85/5 PVD 7 6 5/8 Staging 6 8 8 /8 Staging 7 8 6 8 5/ Late death 7 5.5 7 LPA = left pulmonary artery; PA = pulmonary artery; PAP = pulmonary artery pressure; PVD = pulmonary vascular disease; RPA = right pulmonary artery. assesssed by PA diameter measured between the mid-pa and hilum, was achieved in 7% (CL, 60% to 8%). Pulmona y Hemodynamics Selective cannulation of the central shunt has proved difficult at postoperative catheterization as the catheter tends to rebound from the central shunt into the aorta. This has meant that PA pressure measurements to date are available in patients only (Table 5). One patient had systemic pressures in the main pulmonary artery and RPA and probably has pulmonary vascular disease in his right lung and is not suitable for correction. Systolic PA pressure has ranged from to 60 mm Hg in the other patients. Four patients had a systolic PA pressure between 5 and 60 mm Hg and all have successfully undergone correction. Comment Patients with pulmonary atresia, VSD, very small PAS, MAPCAs, and arborization abnormalities require central PA growth as well as unifocalization of pulmonary blood flow before a corrective procedure can be undertaken. Two different approaches have been used in an effort to attain PA growth. In the published literature the most common approach has been the use of right ventricular outflow tract reconstruction [l-7. In the Mayo Clinic series (actual PA sizes not specified) approximately 50% of patients treated with right ventricular outflow tract reconstruction have gone on to complete repair [,,. The remaining patients were not suitable for complete repair because of inadequate PA enlargement or insufficient arborization of the central PAS. Other studies have also reported only moderate success in attaining PA growth using this technique [la]. As well, serious PA stenosis and right ventricular outflow tract aneurysms from patch enlargement have been reported [l,. The results of right ventricular outflow tract patching, particularly in paiients with very small PAS, were similar also in our own unit, leading to a change in approach. We currently employ the direct end-to-side central shunting technique for all patients with very small PAS in this setting. We prefer this to a peripheral shunting technique because we believe that the chances of attaining symmetric growth without natrogenic stenosis from the shunt are higher. The problem of excessive pulmonary blood flow from too large a shunt needs to be addressed when selecting patients for a direct central shunt, and also while nnonitoring their postoperative care. This is especially true if the MAPCA supply is well developed. Forty percent of our patients had signs of congestive failure, which, although mild to moderate in the majority, did require antifailure treatment. The other common problem in the medium term is proximal PA stenosis, which occurs frequently as an associated congenital lesion in pulmonary atresia [. Where proximal PA growth has been suboptimal owing to such stenoses, correction is not precluded as long as the PA size at the hilum is satisfactory. However, extensive PA reconstruction has usually been required, often out into the hilum and occasionally into the lower lobe branch, to obtain satisfactory PA dimensions. Eleven of patients coming to corrective operation have required PA reconstruction. Proximal PA stenoses have affected the RPA more than the LPA. Where there has been moderate growth the RPA can be reconstructed, but in patients the RPA was virtually occluded. This has proved to be a major problem because reoperation for direct RPA-aorta anastomosis has uniformly failed to achieve RPA growth. We have consequently abandoned this approach in favor of direct reconstruction of the PA bifurcation. It is likely that some RPA stenoses can be avoided by placing the anastomosis as posteriorly as possible to avoid tension or kinking of the RPA. In other patients, even in those with very small PAS, bifurcation stenosis has been identified at initial angiography. A potential problem from direct central shunting is pulmonary vascular disease due to systemic pressure in the pulmonary vessels. This has occurred in only patient, precluding correction. In the other patients in whom pressures have been measured, the peak systolic PA pressure has always been subsystemic. The majority of central shunts have been performed in patients less than months of age. This philosophy of early entrance into a staged surgical program evolved because of the disappointing experience with unifocalization in some of the older patients with thick sclerotic MAPCAs and hypertensive lung segments. Analys8is of the overall experience of 58 patients from this unlit revealed a 78% incidence of suitability for correcticln in patients less than 5 years of age compared with 50% for the older patients (p < 0.05) []. Our current indications for use of the central end-toside shunt in pulmonary atresia plus VSD with MAI'CAs is the presence of very small confluent PAS. In patients less than year of age, this means RPA and LPA ranging

Ann Thorac Surg 99;5:-7 WATTERSON ET AL 7 from to mm in diameter. If the PAS are not visualized on echocardiographic and catheterization studies (an unusual occurrence), exploration may still be carried out through a sternotomy. We prefer to establish this shunt before unifocalization of MAPCAs. The shunt should ideally be performed in the first year of life, even in asymptomatic patients. Timing of further staging and biventricular repair may vary considerably according to individual anatomical features. Subsequent staged bilateral thoracotomies are used for unifocalization of MAP- CAs. This is complemented by catheter occlusion (coils) of MAPCAs with major connections to the native PA system, usually discovered after complete repair. In conclusion, our overall experience with the direct central shunt in this difficult group of patients highlights certain points.. This shunt is suitable only for patients with very small PAS. These patients need careful supervision in the early postoperative period because of their prospensity to cardiac failure. In the majority, cardiac failure can be successfully treated with an antifailure regimen. If it cannot, early catheterization is warranted because the PAS will most likely have grown and the patient can be placed on an accelerated staging program with a view to early correction.. The shunt must be placed as posteriorly as possible on the aorta to reduce RPA distortion.. Although proximal PA stenoses have been common after a direct central shunt, correction is possible as long as the PA diameter at hilar level and beyond is appropriate for the size of the child.. Overall, we believe that early intervention for staging in these patients will allow better results from shunting and unifocalization procedures and will maximize the number of patients who will be suitable for biventricular repair. Ideally, surgical staging should commence in the first year of life. References. Freedom RM, Pongilione G, Williams WG, Trusler GA, Rowe RD. Palliative right ventricular outflow tract reconstruction for patients with pulmonary atresia, ventricular septal defect and hypoplastic pulmonary arteries. J Thorac Cardiovasc Surg 98;86:-6.. Milliken JS, Puga FJ, Davidson GK, Schaff HV, Julsrud PR, Mair DD. Staged surgical repair of pulmonary atresia, ventricular septal defect and hypoplastic confluent pulmonary arteries. J Thorac Cardiovasc Surg 979;78:&.. Piehler JM, Davidson GK, McGoon DC. Management of pulmonary atresia with ventricular septal defect and hypoplastic pulmonary arteries by right ventricular outflow reonstruction. J Thorac Cardiovasc Surg 980;80:557.. Lane I, Treasure T, Leijala M, Shinebourne E, Lincoln C. Diminutive pulmonary artery growth following right ventricular outflow tract enlargement. Int J Cardiol 98;:755. 5. Sullivan ID, Wren C, Stark J, de Leva MR, Macartney FJ, Deanfield JE. Surgical unifocalization in pulmonary atresia and ventricular septal defect. A realistic goal? Circulation 988;78(Suppl ):5-. 6. Gill CC, Moodie DS, McGoon DC. Staged surgical management of pulmonary atresia with diminutive pulmonary arteries. J Thorac Cardiovasc Surg 977;7:&. 7. Crupi G, Locatelli G, Villani M, Tiraboshchi R, Parenzan L. Open heart palliative surgery for pulmonary atresia with ventricular septal defect and hypoplastic pulmonary arteries. Thorax 978;:65-8. 8. Kirklin JW, Bargeron LM, Pacific0 AD. The enlargement of small pulmonary arteries by preliminary palliative operations. Circulation 977;56:6-7. 9. Momma K, Takao A, Imai Y, Kurosawa H. Obstruction of the central pulmonary artery after shunt operations in patients with pulmonary atresia. Br Heart J 987;57:5. 0. Blackstone EH, Kirklin JW, Bertranou EG, Labrosse CJ, Soto B, Bargeron LM. Preoperative prediction from cineangiograms of postrepair right ventricular pressure in tetralogy of Fallot. J Thorac Cardiovasc Surg 979;78:5.. Iyer KS, Varma M, Mee RBB. Use of azygos vein as interposition graft for surgical unifocalization of pulmonary blood flow. Ann Thorac Surg 989;8:776-8.. Puga FJ, Leoni FE, Julsrud PR, Mair DD. Complete repair of pulmonary atresia, ventricular septal defect, and severe peripheral arborisation abnormalities of the central pulmonary arteries. J Thorac Cardiovasc Surg 989;98:0&%9.. Elzenga NJ, Suylen RJ, Frohn-Mulder I, Essed CE, Bos E, Quaegebeur JM. Juxtaductal pulmonary artery coarctation: an underestimated cause of branch pulmonary artery stenosis in patients with pulmonary atresia or stenosis and a ventricular septal defect. J Thorac Cardiovasc Surg 990;lOO: 6.. Iyer KS, Mee RBB. Staged repair of pulmonary atresia and ventricular septal defect with hypoplastic pulmonary arteries and major systemic to pulmonary artery collaterals. Ann Thorac Surg 99;5:65-7.