Long-Term Outcome After Repair of Coarctation in Infancy: Subclavian Angioplasty Does Not Reduce the Need for Reoperation

Transcription

1 1406 lacc Vol. 8, No.6 December 1986:140(>-11 PEDIATRIC CARDIOLOGY Long-Term Outcome After Repair of Coarctation in Infancy: Subclavian Angioplasty Does Not Reduce the Need for Reoperation ROBERT H. BEEKMAN, MD, FACC, ALBERT P. ROCCHINI, MD, DOUGLAS M. BEHRENDT, MD, FACC, EDWARD L. BOVE, MD, FACC, MACDONALD DICK II, MD, FACC, DENNIS C. CROWLEY, MD, A. REBECCA SNIDER, MD, FACC, AMNON ROSENTHAL, MD, FACC Ann Arbor, Michigan To assess the influence of surgical technique on the need for reoperation after coarctation repair in infancy, follow-up data were analyzed for 125 consecutive infants «12 months) who underwent repair of coarctation of the aorta by subclavian angioplasty or resection and end to end anastomosis. Sixty-threeinfants underwent coarctation repair b)' resection between 1960 and 1980, and 62 underwent subclavian angioplasty between 1977 and The mean age (± SEM) at operation for infants with subclavian flap angioplasty was 1.54 ± 0.93 months and for infants with resection was 2.70 ± 0.93 months (p =0.02). There was no difference between the groups in patient weight at initial repair or the proportion of patients with complex anatomy or aortic arch hypoplasia. Follow-up duration for the subclavian flap group was 2.55 ± 0.51 years (range 0.3 to 8.2), and for the resection group was 7.97 ± 3.61 years (range 0.6 to 21). Indication for reoperation was the presence of a coarctation gradient at rest of 40 rom Hg or greater and arm hypertension. Reoperation was required in 5 patients in the subclavian flap group and 12 patients in the resection group. The mean reoperation rate after subclavian flap repair was reoperations per patientyear, and after resection was reoperations per patient-year (p = 0.94). To determine an individual's risk of requiring reoperation from these group measures, a reoperation risk model was developed. The risk of reoperation by the fifth postoperative year was found to be 16.3% after subclavian flap repair and 15.7% after resection. These risks increase to 30% and 29%, respectively, by the 10th postoperative year. These data indicate that recoarctation is an important problem after subclavian flap angioplasty in infancy. The risk of requiring reoperation is equivalent during the first 10 postoperative years whether repair is performed by subclavian flap or resection and end to end anastomosis. (J Am Coli Cardiol 1986;8: ) Surgical correction of symptomatic coarctation of the aorta in infancy can be lifesaving. but the long-term outcome is not always optimal. Residual or recurrent stenosis at the coarctation repair site occurs in an appreciable number of patients and, when severe, may require reoperation or balloon dilation (1-4). Recurrent coarctation occurs most commonly when surgical repair is performed during the first year of life. A number of studies ( 1,5-8) have reported that recurrent coarctation develops in 20 to 60% of children after From the Divisions of Pediatric Cardiology and Thoracic Surgery, C.S, MOll Children's Hospital, The University of Michigan. Ann Arbor. Michigan. Manuscript received April I, 1986; revised manuscript received June accepted June Address for reprints: Robert H. Beekman. MD. Division of Pediatric Card iology. Room FII23. Box University of Michigan Medical Center East Medical Center Drive, Ann Arbor. Michigan, by the American College of Cardiology resection and end to end anastomosis in infancy. Subsequent to these reports, left subclavian flap angioplasty was heralded as the preferred surgical procedure in infanc y because of an apparent lower incidence of restenosis (9-12). Many institutions, including our own, adopted the subclavian flap operation as the procedure of choice in infants with symptomatic coarctation. Recently, however, follow-up data have become available which suggest that recurrent coarctation is an important problem after subclavian flap repair in infancy (13-15). The purpose of the present study was to evaluate retrospectively the University of Michigan experience with surgical repair of coarctation of the aorta in infancy, and to compare the long-term outcome after the subclavian flap operation with that after resection and end to end anastomosis. The primary question to be addressed is whether the subclavian flap procedure reduces the need for reoperation after primary repair ofcoarctation in infancy /86/$3.50

2 JACC Vol. 8, No.6 December 1986: BEEKMAN ET AL Methods Patients. The hospital records of all patients with coarctation operated on at the University of Michigan between 1960 and 1985 were reviewed to identify those undergoing repair by resection and end to end anastomosis or left subclavian flap operation during the first 12 months of life. Patients with interrupted aortic arch or hypoplastic left heart syndrome were excluded from analysis, Also excluded from this study were 16 infants who underwent Goretex patch aortoplasty ( 1978 to 1984) when they were judged unsuitable for the subclavian flap operation because of hypoplasia of the transverse arch or because of abnormalities of the left subclavian artery. One hundred twenty-five patients were identified for inclusion in this study. Hospital records, operative reports, cardiac catheterization reports and angiograms were reviewed with specific attention paid to intraand extracardiac anatomy, indications for primary repair, type of initial surgical procedure, postoperative com p lic~ tions, follow-up blood pressure in the arms and the systolic pressure gradient between arm and leg. Blood pressure was measured in the supine position using arm and leg cuffs of appropriate size. When available, follow-up coarctation pressure gradients measured at cardiac catheterization were used in preference to cuff measurements. For patients undergoing reoperation for recurrent coarctation, the indications for the repeat surgical procedure were carefully determined. Three children who have had balloon dilation of a recurrent coarctation (two from end to end anastomosis, one from left subclavian flap group) are included in the reoperation group for purposes of this analysis. Sixty-three infants underwent coarctation repair by resection and end to end anastomosis between 1960 and 1980, and 62 underwent left subclavian fl ap angioplasty between 1977 and The indication for surgery was congestive heart failure in 120 and severe upper limb hypertension in 5 infants. End to end anastomosis was performed using interrupted sutures around the anterior half to two-thirds of the anastomosis. The suture material utilized for the anastomosis was silk in 34, Tevdek in 20 and was not recorded in 9 patients. Left subclavian fl ap angioplasty was performed as previously described (9), with excision of all obvious intraluminal coarctation tissue. Polypropylene suture material was used in subclavian angioplasty repairs. Statistical analysis. Comparison of group characteristics was performed by the student' s t test and chi-square test for continuous and categorical measures, respectively. All descriptive statistics are expressed as mean ± standard error, and a probability (p) of 0.05 or less was required as evidence of a significant difference between groups (twotailed test). The reoperation experience of the two groups was summarized by applying actuarial lifetable and survival curve techniques (16). Equality of the two reoperation-time distributions was assessed using Breslow' s proportional hazards model. To normalize for the differing follow-up duration among groups, a summary measure was calculated for each group relating number of reoperations to patientyears of follow-up. Using this measure of mean reoperation rate (R). an individual' s risk for reoperation was modeled as follows: Risk., = 1 - e - R:>', where Ilt is the follow-up time in years, and risk is an individual' s probability of undergoing reoperation during follow-up time Ilt ( 17). The risk model requires an assumption that the reoperation rate remains constant throughout the follow-up period. The risk analysis is therefore limited to the fi rst 10 postoperative years so that the reoperation rate (R) used in the risk function is, in fact, the mean reoperation rate for the time period over which the function is applied. Similar methods were utilized to calculate the rates and risks for the presence of an arm to leg systolic pressure gradient of 20 rnm Hg or more, and upper limb systolic hypertension in excess of the 95th percentile for age and sex (18). Because the onset of such events cannot be defined retrospectively as precisely as reoperation, however, the epidemiologic estimates of rate and risk for these events are likely to be less accurate than for reoperation. Results Clinical features. Pertinent characteristics of the two surgical groups are listed in Table 1. During the 1960s and Table I. Pertinent Characteristics of 125 Infants Undergoing Coarctation Repair Between 1960 and 1985 Subclavia n Flap Resection p Va lue Infant s (no.) Age (mo) Weight (kg) Tubular arc h hypoplasia ('7c) Intracardiac defects (%) Immediate postoperative gradie nt (mm Hg) Surgical deaths (no.) ± ± % 6 1% 2.0 ± ± ± % 53% 2.6 ±

3 1408 BEEKMAN ET AL. JACC Vol. 8. No.6 December 1986: early 1970s standard care included vigorous anticongestive medical therapy in an attempt to postpone surgery. Thus, the mean age at operation was greater in the resection group than in the subclavian flap group (2.70 versus 1.54 months, p = 0.02). Weight at operation was not significantly different between the groups, however (p = 0.08). The two groups were equivalent for the percent of patients with tubular arch hypoplasia, significant intracardiac lesions and immediate postoperative arm-leg pressure gradient. There was a significantly greater number of surgical deaths, defined as intraoperative death or death during the first postoperative month, in the resection group (p = 0.01). The major causes of surgical mortality were severe congestive heart failure, acidosis and intraoperative cardiac arrest. None of the infants who died early had evidence of a residual coarctation. Follow-up data. The 44 survivors of resection and end to end anastomosis and the 55 survivors of the subclavian flap repair comprise the two follow-up groups. Follow-up data were available on all 99 children. The mean followup duration was 2.55 ± 0.51 years (range 0.3 to 8.2) for the subclavian flap group and 7.97 ± 3.61 years (range 0.6 to 21) for the resection group. There were 12 late deaths, 7 in the subclavian flap group and 5 in the resection group. Eight late deaths occurred at surgery for an associated intracardiac lesion, and two (one in each group) occurred at reoperation for recurrent coarctation. Seventeen patients required reoperation for coarctation (Table 2). In the subclavian flap group, five children required reoperation an average of 1.5 ± 0.5 years after initial repair (range 0.4 to 3.1). Twelve children in the resection group underwent reoperation 7.5 ± 1.6 years after initial repair (range 1.7 to 17.6). The indication for reoperation was the presence of upper limb hypertension and a resting arm to leg systolic pressure gradient of 40 mm Hg or more. This criterion appears to have been applied equitably to both groups because all 17 patients satisfied it. The follow-up groups contain only one patient each with borderline criteria who has not undergone reoperation. All patients undergoing reoperation had a discrete stenosis at the site of the previous coarctation repair. A residual systolic gradient (all 2':20 mm w W... I Z o ~ 0.8 W ; 0.6 W Z E- o 0., o 0., Subclavian flap N J:r:/.JJ4/YSaTaTaTaTaTf Resection Legend Resection p=o N oj:r:/.j13jj4jj4jj4/ysat at.ll8.ll8.ll8.1o.1o.1o.1o.1o O-t _.,..., T"'""' ,...-.-_..-,r-""' o FOLLOW-UP (years) Figure 1. Actuarial curve relating cumulative proportion of patients remaining reoperation-free to follow-up time in years. The curves (at 8.2 year follow-up) are not significantly different (p = 0.86). N = number of patients at risk during the interval; SE = standard error. Hg) was recorded immediately after the initial repair in three of five patients in the subclavian flap group and in four of eight in the resection group who eventually required reoperation (p = 0.32). Thus, approximately half of the patients underwent reoperation for a true residual coarctation. Table 2 shows that age, but not weight, at initial repair was lower in the subclavian flap group. Within each follow-up group, however, age at initial operation was not significantly less for patients requiring reoperation than for those not requiring reoperation (subclavian flap p = 0.28; resection p = 0.76). Reoperation experience. The reoperation experience for each surgical group is summarized in Figure I. This actuarial curve plots the cumulative proportion of patients who had no reoperation against follow-up time in years. The two curves are not significantly different (p = 0.86, Breslow's proportional hazard). Direct comparisons beyond 8 years cannot be made using this technique, however, because follow-up data are lacking in the subclavian flap group. In the resection group, the data indicate that only 32% of Table 2. Pertinent Data Regarding 17 Children Requiring Reoperation After Coarctation Repair in Infancy Type of Initial Coarctation Repair Subclavian Flap Resection p Value Infants requiring reoperation (no.) Initial coarctation repair Age (mo) Weight (kg) Age at reoperation (yr) Indications for reoperation Gradient (mm Hg) Hypertension (no.) ± ± 0.\ 1.53 ± ± ± ± ± ±

4 l ACC Vol. 8. NO. 6 December 1 9 ~ 6 : BEEKMAN ET AL COARCTATI ON REPAIR IN INFANCY 1409 g Z :2 40 ~ 0:: '" Legend o ~C!:!!!.~N FL~~ RESECTION c, 30 0 AI r.<l.# 0:: ~ 0 20 :.:: fi) ~ 10 2 /.#.# FOLLOW-UP (years) Figure 2. Reoperation risk functions after subclavian flap or resection for coarctation in infancy. An individual's risk of requiring reoperation is related to follow-up time in years. For subclavian flap: Y = I - e- O.03S6X, and for resection: Y = I - e- O.034 2X. Because the model assumes that the rate of reoperation is a constant, the mean rate was used and the model was not extended beyond the 10th postoperative year. patient s remain reoperation-free by the 21st postoperative year. The standard error of this estim ate is large after year 16, however, because of the small number of patients involved. The mean reoperation rate, defined as the number of reoperations divided by the total patient-years of follow-up, provides an additional method by which the reoperation experience of the two surgical groups can be compared. The reoperation rate for the subclavian flap group was per patient-year (5 per pat ient-years), compared with per patient-year for the resection group (12 per patient-years). The difference between the two rates is not significant (p = 0.94). It should be noted that the analy sis has a power of 0.90 to dete ct as significant a reoperation rate differenc e as low as between the groups (twotailed test, a = 0.05). Figure 2 illustrates the reoperation risk functions after resection or subclavian flap repair in infancy. The two risk functions are nearly identical for the first 10 years of follow-up. Table 3 contains the reop eration rate data as well as the estimated 5 and 10 year risks for reoperation after each surgical procedure. An indi vidual ' s risk of requiring reoperation by the fifth postoperative year is 16.3% after the subclav ian flap operation and 15.7% afte r resection. The 10 yea r reoperation risks are 30 and 29% with the subclavian flap and resect ion operations, respectively. Because of the differing age distributions of the subclavian jlap and resection groups at time of initial repair, an age-stratifi ed analysis of reoperation rates and rate ratio s was performed. The reoperation rate ratios were not significantly different from 1.0 for children under I month or for children older than I month of age at time of the initial coarctation repair. The 95% confidence intervals for the reoperation rate ratio s were 0.32 to 6.26 for those under I mont h, and 0.06 to 3.30 for those I month or older when the initial repair was perform ed. Th ese data, and the fac t that weight at initial repair was similar for each gro up, sugges t that age differences did not confound the analysis of reop eration rate s. Table 3 also provide s the rates and calculated 5 and 10 year postoperative risks for the presence at rest of a coarctation gradient of 20 mm Hg or greater and upper limb systolic hypertension. Regardless of the type of coarctation repair performed in infancy, an individual' s risk for a resting gradie nt of 20 mm Hg or more is 24.8% after 5 years and 43. 5% after 10 years. Th e risk of developing upper limb hypertension is also very similar after either surgi cal procedure. The 10 year risk for hypertension is 39.3% after subclavian flap repair and 40.2% after resection. Discussion Resection and end to end anastomosis versus left subclavian flap angioplasty. Res idual or recurrent coarctation occ urs commonly after surgical repair of coarctation of the aorta in early childhood. Although we have previously shown Table 3. Five and 10 Year Rates (per patient-year) and Risks for Reoperation, Gradient 20 mm Hg or Greater or Hypertension After Coarctation Repair in Infancy Reoperation Arm-Leg Systolic Gradient '2:20 mm Hg Upper Limb Hypertension Risk Risk Risk Rate 5 Year (%) 10 Year (%) 5 Year 10 Year 5 Year 10 Year Rate (%) (%) Rate (%) (%) Subclavian flap Resection Ratio (95% confidence interval) (0.37 to 2.93) S O. 9 ~ (0.92 to 1.09) (0.42 to 2.24)

5 1410 BEEKMAN ET AL. lacc Vol. 8. No.6 December 1986: ( I) that recurrent coarctation is a problem for children who are younger than 3 years of age at initial repair, recurrent stenosis most commonly occurs when coarctation is repaired in infancy. After resection and end to end anastomosis in infancy, 20 to 60% of patients may develop a significant residual coarctation gradient (1,5-8,19,20). The postoperative stenosis may be caused by inadequate repair (a true residual coarctation) or inadequate growth of the suture line postoperatively (1,2). For these reasons, the left subclavian angioplasty procedure, introduced by Waldhausen and Nahrwald (21) in 1966, has been advocated as the preferable operation for coarctation repair in infancy (9-11). The risk of restenosis was thought to be lower with the subclavian flap repair because there is no circumferential suture line, and the repair utilizes viable subclavian artery tissue with a potential for growth. Initial follow-up studies (10,12,22) did, in fact, report a lower incidence of restenosis ranging from 0 to 12% after subclavian flap repair. Two recent studies (13,15), however, have reported a residual coarctation gradient in a substantial number of children after subclavian flap repair in infancy. Ziemer et al. (15) found that 27. I% of infants with such repair had undergone reintervention (surgery or balloon dilation) for recurrent stenosis by the fifth postoperative year. This did not differ significantly from the 20.3% 5 year reintervention rate in their resection group. Cobanoglu et al. (13) found a 5 year reoperation rate of 25% after the subclavian flap operation but reported only an 8% rate after resection and end to end anastomosis. Our data also indicate that the long-term outcome after the subclavian flap operation may not be as favorable as initially reported. Recoarctation does, in fact, appear to be an important problem when the subclavian flap procedure is performed in infancy. Using actuarial and epidemiologic methods we have estimated the risk of reoperation to be 16.3% at 5 years and 30% at 10 years after the subclavian flap operation in infancy (Table 3). These findings are similar to those of Ziemer (15) and Cobanoglu (13) and their coworkers. Further, we found that the reoperation rate and risk after coarctation repair in infancy are remarkably similar in both the subclavian flap repair and the resection groups. The subclavian flap repair group will have to be followed up for a longer period of time to determine whether its reoperation experience will continue to parallel that of the resection group (Fig. I). Nevertheless, our experience to date does not support the widely held opinion that the subclavian flap operation is the clear procedure of choice for coarctation repair in infancy (9-11). Mechanism of restenosis. Because the coarcted segment of the aorta is not resected, recoarctation after the subclavian flap procedure may be caused by persistence or ongoing involution, or both, of the abnormal tissue in the juxtaductal aorta and posterior shelf. In fact, a prominent posterior shelf was observed at surgery in each child who underwent reoperation. Three of the five children had a residual pressure gradient (20, 25 and 40 mm Hg, respectively) immediately after the subclavian flap procedure, suggesting that resection of the posterior shelf had been incomplete. All five children had a substantial increase in the postoperative coarctation pressure gradient that culminated in reoperation 0.4 to 3.1 years after the original operation. The two children without a residual pressure gradient immediately after subclavian angioplasty required reoperation for a coarctation gradient of 48 mm Hg and 80 mm Hg, respectively, only 4 and 5 months later. The rapidity with which the recurrent gradient developed in these children supports the concept that progressive involution and constriction occurred in the juxtaductal aorta postoperatively (13). Limitations of a retrospective study. Our experience suggests, but does not prove, that recoarctation requiring reoperation occurs with similar frequency after the subclavian flap and resection operations when coarctation is repaired in infancy. A retrospective nonconcurrent cohort study such as ours, however, is subject to potential biases. First, because treatments were not randomly allocated, the two surgical groups may differ in factors other than the surgical procedure that may affect outcome. The two groups were equivalent in the number of children with tubular aortic arch hypoplasia and intracardiac defects, but the subclavian flap repair group was slightly younger than the resection group when surgery was performed. An age-stratified analysis, and the fact that weight at operation was not different between the groups, suggests that age at operation was not a major confounding factor. Second, because the groups were not treated concurrently, important aspects of therapy other than the surgical procedure itself may have differed between the groups. The lower surgical mortality in the subclavian flap repair group (Table I) suggests that this was indeed the case. We attribute the recent decline in mortality to improved perioperative care and to the availability of prostaglandin E, at our institution since Finally, it is important to point out that the equivalency of two interventions can never be proved; we can only state that the available data fail to demonstrate a significant difference between the two surgical procedures (23). Summary and recommendations. The long-term experience at the University of Michigan with the subclavian flap procedure and resection for symptomatic coarctation in infancy suggests that the 5 and 10 year risk of reoperation is similar after either operation. These findings are consistent with other recent reports (13,15) indicating that recurrent stenosis is an important problem after the subclavian flap operation. We recommend that resection and end to end anastomosis not be abandoned in favor of the subclavian flap angioplasty. The surgical procedure utilized should be chosen on the basis of each infant's surgical anatomy. If a child is deemed to be a poor candidate for subclavian flap

6 lacc Vol. 8. No.6 December 1986: BEEKMAN ET AL 1411 angioplasty we would recommend that resection and end to end anastomosis be performed, because the two procedures appear to result in a similar outcome during the first 10 postoperative years. References I. Beekman RH, Rocchini AP, Behrendt DM, Rosenthal A. Reoperation for coarctation of the aorta. Am J Cardiol 1981;48: Ibarra-Perez C, Castaneda AR, Varco RL, Lillehei CWo Recoarctaiion of the aorta. Nineteen year clinical experience. Am J Cardiol 1969;23: Kan JS, White RI, Mitchell SE, Farmlell EJ, Donahdo JS, Gardner TJ. Treatment of restenosis of coarctation by percutaneous transluminal angioplasty. Circulation 1983;68: Kveselis D. Rocchini AP. Percutaneous transluminal angioplasty of peripheral pulmonary arterial stenosis, coarctation of the aorta, superior vena caval and pulmonary venous stenosis. and other great artery stenoses. Semin Intervent Radiol 1984;1: Khoury GH, Hawes CR. Recurrent coarctation of the aorta in infancy and childhood. J Pediatr 1968;72: Hartmann AF, Goldring D, Hernandez A, Behrer MR. Recurrent coarctation of the aorta after successful repair in infancy. Am J Cardiol 1970;25: Eshaghpour E, Olley PM. Recoarctation of the aorta following coarctectomy in the first year of life. J Pediatr 1982;80: Patel R, Sinigh SP, Abrams L, Roberts KD. Coarctation of the aorta with special reference to infants. Br Heart J 1977;39: Pierce WS, Waldhausen JA, Berman W, Whitman V. Late results of the subclavian flap procedure in infants with coarctation of the thoracic aorta. Circulation 1978;58(suppl 1): Hamilton 01, DiEusanio G, Sandrasagra FA, Donnelly RJ. Early and late results of aortoplasty with a left subclavian flap for coarctation of the aorta in infancy. J Thorac Cardiovasc Surg 1978;75: II. Bergdahl LAL, Blackstone EH, Kirklin JW, Pacifico AD, Bargeron LM. Determinants of early success in repair of aortic coarctation in infants. J Thorac Cardiovasc Surg 1982;83: Campbell DB, Waldhausen JA, Pierce WS, Fripp R, Whitman V. Should elective repair of coarctation of the aorta be done in infancy'? J Thorac Cardiovasc Surg 1984;88: Cobanoglu A. Teply JF, Grunkemeier GL, Sunderland CO, Starr A. Coarctation of the aorta in patients younger than three months. A critique of the subclavian flap operation. J Thorac Cardiovasc Surg 1985;89: Metzdorff MT, Cobanoglu A, Grunkemeier GL, Sunderland CO, Starr A. Influence of age at operation on late results with subclavian flap aortoplasty. J Thorac Cardiovasc Surg 1985;89: Ziemer G, Jonas RA, Perry SB, Freed MD, Castaneda AR. Coarctation repair in the neonate: A comparison of subclavian patch plasty and resection (abstr). Circulation 1985;72(suppl 11I): Kaplan E, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457~ Morgenstern H, Kleinbaum DG, Kupper LL. Measures of disease incidence used in epidemiologic research. lnt J Epidemiol 1980; 9: is. National Heart, Lung, and Blood Institute's Task Force on Biood Pressure Control in Children: Report of the task force on blood pressure control in children. Pediatrics 1977;59: Hesslein PS, Gutgesell HP, McNamara DG. Prognosis of symptomatic coarctation of the aorta in infancy. Am J Cardiol 1983:51: Hammon JW, Graham TP, Boucek RJ, Bender HW. Operative repair of coarctation of the aorta in infancy: results with and without ventricular septal defect. Am J Cardiol 1985;55: Waldhausen JA, Nahrwald DL. Repair of coarctation of the aorta with a subclavian flap. J Thorac Cardiovasc Surg 1966;51: Thibault WN, Sperlurg DR, Gazzaniga AB. Subclavian artery patch angioplasty. Arch Surg 1975; 110: Blackwelder WC. "Proving the null hypothesis" in clinical trials. Controlled Clin Trials 1982;3: