Outcomes of Surgical versus Balloon Angioplasty Treatment for Native Coarctation of the Aorta: A Meta-Analysis Zhi-peng Hu,1,2 Zhi-wei Wang,1 Xiao-feng Dai,1 Bo-tao Zhan,2 Wei Ren,1 Luo-cheng Li,1 Hao Zhang,1 and Zong-li Ren,1 Wuhan, People’s Republic of China
Background: Native coarctation of the aorta (COA) accounts for 5e7% of congenital heart disease. Open surgical treatment was the only choice until balloon angioplasty (BA) treatment was introduced as an alternative therapy for COA in the 1980s. BA treatment was thought to be a less invasive and potentially safer technique, and has been used on numerous patients. But as has been reported during the past 30 years, the risk of aneurysm formation and recoarctation existed in either of those 2 procedures. Unfortunately, follow-up for either type of treatment has been limited, making it difficult to draw any meaningful conclusions as to which treatment option is superior. Our objective was to compare results of 2 therapeutic modalities to treat native COA: BA without stent implantation and surgery. Methods: We performed a meta-analysis of controlled trials of surgical versus BA treatment for native COA. MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials, CINAHL, Web of Science, and the Chinese Biomedical Database of clinical trials were searched using PubMed and OVID. Controlled trials in which patients with COA were assigned to surgical repair or BA treatment were included. For each outcome, we evaluated the quality of the evidence with reference to the Grading of Recommendations Assessments, Development, and Evaluation criteria. We used RevMan 5.1 software (The Nordic Cochrane Centre, Copenhagen, Denmark) to analyze the data. Results: A literature search yielded 9 comparable studies, for a total of 623 patients, of whom 378 and 245 were assigned to surgery and BA. Meta-analysis of these studies showed no significant difference in postintervention gradient (inverse variance fixed mean difference: 1.44 [95% CI: e1.16 to 4.04]), midterm recoarctation (ManteleHaenszel [M-H] random odds ratio [OR]: 0.24 [95% CI: 0.04e1.58]), and long-term recoarctation (M-H fixed OR: 0.61 [95% CI: 0.34e 1.11]). BA reduces the risk of severe complications (M-H fixed OR: 2.67 [95% CI: 1.37e5.21]; P < 0.001) but increases the risk of short-term recoarctation (M-H fixed OR: 0.25 [95% CI]: 0.12e0.54]; P < 0.001) and aortic aneurysm formation (M-H fixed OR: 0.12 [95% CI]: 0.04e 0.34]; P < 0.001). Conclusions: BA provides immediate results comparable to surgery and reduces invasion, but it does not provide better results compared with surgery when considering medium- and longterm complications and even increases the incidence of aneurysm formation.
1 Department of Cardiothoracic Surgery, Renmin Hospital of Wuhan University, Wuhan, People’s Republic of China. 2 Department of Cardiothoracic Surgery, Xiangyang central Hospital, Xiangyang, People’s Republic of China.
Correspondence to: Zhi-wei Wang, MD, Department of Cardiothoracic Surgery, Renmin Hospital of Wuhan University, 99 Ziyang Road, Wuhan, Hubei Province 430060, People’s Republic of China; E-mail: [email protected]
394
Ann Vasc Surg 2014; 28: 394–403 http://dx.doi.org/10.1016/j.avsg.2013.02.026 Ó 2014 Elsevier Inc. All rights reserved. Manuscript received: October 28, 2012; manuscript accepted: February 17, 2013; published online: November 5, 2013.
Vol. 28, No. 2, February 2014
Surgical vs. balloon angioplasty for native COA 395
Table I. Patient demographics Reference, year
Participants
Intervention
Mean age (yrs or days ± SD)
Weight (kg ± SD)
M/F
Shaddy et al., 1993
20 16 72 61 128 41 34 23 14 15 30 28 46 13 18
Surgery BA Surgery BA Surgery BA Surgery BA Surgery BA Surgery BA Surgery BA Surgery
5.7 ± 2.1 yrs 6.3 ± 2.0 yrs 10.0 ± 9.7 yrs 9.8 ± 8.0 yrs 3.3 ± 7.2 yrs 3.0 ± 4.9 yrs 7.7 days ± NA 8.8 days ± NA 27 ± 35 days 29 ± 27 days 7.00 ± 4.14 yrs 6.66 ± 4.44 yrs NA NA Range (0.3e15 yrs) median (0.63 yrs) Range (0.25e15 yrs) median (5.8 yrs) 5.7 ± 2.1 yrs 6.3 ± 2.0 yrs
19.5 21.6 35 30 NA NA 3.4 3.6 3.5 3.8 26.51 25.47 NA NA NA
16/4 8/8 50/22 39/22 27/14 75/53 20/14 15/8 8/6 9/6 22/8 20/8 NA NA NA
Forbes et al., 2011 Chiu et al., 2009 Fiore et al., 2005 Rao et al., 1994 Hernandez-Gonzalez et al., 2003 Fruh et al., 2011 Walhout et al., 2004
Cowley et al., 2005
28
BA
16 20
Surgery BA
± ± ± ±
4.816 8.1 24 21
± ± ± ± ± ±
NA NA 0.9 1.0 15.17 12.54
NA 19.9 ± 5.2 21.7 ± 8.1
NA NA NA
BA, balloon angioplasty; F, female; M, male; NA, not available; SD, standard deviation.
INTRODUCTION Open surgery was the only treatment for coarctation of the aorta (COA) until balloon angioplasty (BA) treatment was introduced as an alternative therapy in the 1980s.1 Since then, aortic angioplasty, thought to be a less invasive and safer treatment, has been adopted for COA therapy. But as has been reported during the past 30 years, complications, such as aortic rupture along with life-threatening bleeding, aneurysm formation, and recoarctation, have been associated with both of these procedures.2e4 However, a large, multicenter, randomized, controlled trial comparing acute and follow-up results of the surgical and BA procedures for COA has not been performed, making it difficult to draw any meaningful conclusions as to which treatment option is superior. Several clinical trials have compared outcomes of surgery versus BA treatment for COA, but they have not included enough patients and their results have been inconsistent. We therefore conducted a meta-analysis of all published clinical trials, either randomized or observational, comparing outcomes of surgery versus BA treatment for COA.
METHODS Search Strategy Comprehensive searches of the MEDLINE and Cochrane Central Register of Controlled Trial
databases were performed by 2 reviewers independently using Web-based search engines (i.e., OVID and PubMed) for human studies published in English between 1980 and October 2012. Search terms included the following: coarctation of the aorta, aortic coarctation, COA, surgical procedures, operative, angioplasty, balloon, surger*, surgical*, endovasc*, and transcather*. Bibliographies of retrieved articles were searched for other relevant studies. We monitored major scientific meetings for the results of trials presented before publication. Bibliographies of included articles were also searched. No language restrictions were applied. If there was sample overlap between studies, we included only the larger study.
Selection Criteria Studies were included if they met each of the following conditions: the trial was randomized and controlled or was a conservative study comparing acute and follow-up results of surgical and BA procedures for COA, and if the trail or study reported 1 of the following clinical outcomes: posttreatment gradient, severe complications, hospital stay days, aortic aneurysm formation, recoarctation during short-term, midterm, or long-term follow up. Studies that did not make a distinction between BA with and without stent implantation were excluded.
Quality assessment
No. of patients (%)
Effect Quality importance
No. of studies
Limitation
Risk of bias Inconsistency
Indirectness
Other Imprecision considerations Surgery
No serious Seriousb Observational No serious No serious risk of inconsistency indirectness studiesa bias Postintervention Observational No serious No serious No serious Seriousb a gradient (4) studies risk of inconsistency indirectness bias Severe Observational No serious No serious No serious Seriousb complications studiesa risk of inconsistency indirectness (4) bias Preintervention gradient (5)
BA
Relative (95% CI)
Very strong associationc
116
110
d
None
76
93
d
None
46/122 (37.7)
35/126 OR 2.67 (27.8) (1.37 to 5.21)
23.3%
Observational No serious No serious Mid-to longNo serious Seriousb studiesa term risk of inconsistency indirectness aneurysm bias formation (6)
Very strong associationc
0/228 (0) 23/147 OR 0.12 (15.6) (0.04 to 0.34)
16.5%
Short-term recoarctation (4)
Observational No serious No serious No serious Seriousb studiesa risk of inconsistency indirectness bias
Reporting biasd 14/106 (13.2)
33/108 OR 0.25 (30.6) (0.12 to 0.54)
30.4%
Observational No serious No serious No serious Seriousb studiesa risk of inconsistency indirectness bias
Reporting biasd 6/78 (7.7) and very strong associationc
28/49 OR 0.05 (57.1) (0.01 to 0.37)
53.9%
MD 3.64 higher (1.03 lower to 8.3 higher) MD 1.44 higher (1.16 lower to 4.04 higher) 229 more per 1,000 (from 67 more to 389 more) 215 more per 1,000 (from 61 more to 380 more) 135 fewer per 1,000 (from 97 fewer to 149 fewer) 142 fewer per 1,000 (from 102 fewer to 157 fewer) 206 fewer per 1,000 (from 114 fewer to 255 fewer) 206 fewer per 1,000 (from 113 fewer to 254 fewer) 509 fewer per 1,000 (from 241 fewer to 558 fewer) 484 fewer per 1,000 (from 237 fewer to 527 fewer)
Low
Important
Very low Critical
Very low Critical
Moderate Critical
Low
Important
Moderate Critical Annals of Vascular Surgery
Mid-term recoarctation (3)
Absolute
396 Hu et al.
Table II. GRADE assessment of each comparison
Vol. 28, No. 2, February 2014
Study Variables
Very low Critical
Patient demographics, such as age, weight, sex, and pressure gradient were recorded as baseline clinical information. Posttreatment gradient is the pressure gradient across the coarctation lesions assessed by catheterization or echocardiography within 1 hour after intervention. Severe complications refer to aortic wall injury, dissection/intimal tear, aneurysm, balloon rupture, femoral atrial fibrillation, severe/prolonged hypertension, bleeding, respiratory failure, and death. Recoactation was defined as a peak systolic pressure gradient 20 mm Hg or inevitable of reintervention. Short-, mid-, and long-term follow-ups were defined as follow-ups of 5 years, respectively.
CI, confidence interval; GRADE, Grading of Recommendations Assessments, Development, and Evaluation; OR, odds ratio. a Lack of allocation concealment and blinding. b The number of patients is small. c OR 2. d Funnel plot of this comparison is asymmetrical.
31/103 OR 0.61 93 fewer per (30.1) (0.34 to 1.11) 1,000 (from 173 fewer to 22 more) 32% 97 fewer per 1,000 (from 182 fewer to 23 more) 48/176 (27.3) None Observational No serious No serious No serious Seriousb studiesa risk of inconsistency indirectness bias Long-term recoarctation (4)
Surgical vs. balloon angioplasty for native COA 397
Data Collection and Statistical Analysis Two authors (H.Z-p. and D.X-f.) independently identified trials for inclusion and extracted information on demographics, interventions, and outcomes. Disagreements were resolved by consensus. For dichotomous variables, odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. For continuous variables, the mean difference (MD) and 95% CIs were calculated. Statistical heterogeneity was measured using the Q statistic (P < 0.10 was considered indicative of statistically significant heterogeneity) and the I2 test. For each outcome, the fixed effect (ManteleHaenszel [M-H] for dichotomous variables and inverse variance for continuous variables) or random effects (DerSimonian for dichotomous and Laird for continuous variables) model was used when the Q statistic suggested a lack or presence of heterogeneity, respectively. Review Manager (RevMan) software (version 5.1; The Nordic Cochrane Centre, Copenhagen, Denmark) was used for data analysis.
RESULTS Included Studies The selection criteria described in the Methods section were applied to the approximately 674 studies identified by the literature search. Of these, 11 matched selection criteria. Veyre et al.’s study5 was in fact a systemic review of case reports, and Rodes-Cabau et al.6 did not make a distinction between BA and stent implantation. Both of these studies were excluded. A total of 9 studies7e15 for a total of 623 patients were included in our study. Patient demographics can be seen in Table I.
398 Hu et al.
Annals of Vascular Surgery
Fig. 1. Comparison of surgery versus balloon angioplasty for pretintervention gradient.
Fig. 2. Comparison of surgery versus balloon angioplasty for postintervention gradient.
and (2) randomized studies and observational studies were analyzed together.
Effects of Interventions
Fig. 3. Funnel plot of comparison: postintervention gradient.
Assessment of the Quality of Evidence Table II describes in detail the principal findings of this review and the quality of the evidence for each outcome using the Grading of Recommendations Assessments, Development, and Evaluation (GRADE) approach.16 When using the GRADE approach, all studies were seen as observational studies for the following reasons: (1) the quality of randomized controlled trials included were low;
Postintervention Gradient. Peak-to-peak systolic pressure gradient within 1 hour after intervention. We obtained 4 trials with a total of 169 patients (76 surgery, 93 BA), with similar preintervention gradient, which was given by Figure 1. There was no evidence of heterogeneity (chi-squared: 2.04; df: 3 [P ¼ 0.56]; I2: 0%). There was no statistical difference of the postintervention gradient between the surgery group and the BA group (inverse variance fixed MD: 1.44 [95% CI: e1.16 to 4.04]; Fig. 2). The funnel plot shows that the publication bias was not significant (Fig. 3). Severe Complications. Severe complications after intervention. Information regarding mid- to longterm aneurysm formation was provided by 4 studies for a total of 248 patients (122 surgery, 126 BA). There was no evidence of heterogeneity (chisquared: 4.63; df: 3 [P ¼ 0.20]; I2 ¼ 35%). BA had a significantly lower rate of severe complications after intervention than the surgery group (M-H
Vol. 28, No. 2, February 2014
Surgical vs. balloon angioplasty for native COA 399
Fig. 4. Comparison of surgery versus balloon angioplasty for short-term recoarctation.
short-term follow-up (M-H fixed OR: 0.25 [95% CI: 0.12e0.54]; P < 0.001; Fig. 8). The publication bias was significant (Fig. 9). Midterm recoarctation. Midterm recoarctation was reported in 78 surgery and 49 BA patients from 3 studies. There was heterogeneity among these 3 studies, possibly because of the use of a different standard for recoarctation (t2: 1.43; chi-squared: 4.2; df: 2 [P ¼ 0.12]; I2: 52%). The difference of midterm recoarctation rate was not significant between the 2 groups (M-H random OR: 0.24 [95% CI: 0.04e1.58]; Fig. 10). Fig. 5. Funnel plot of comparison: severe complications.
fixed OR: 2.67 [95% CI: 1.37e5.21]; P < 0.001; Fig. 4). The publication bias was significant (Fig. 5). Aneurysm Formation Mid- to long-term aneurysm formation. Information regarding mid- to long-term aneurysm formation incidence was provided by 6 studies with a total of 375 patients (228 surgery, 147 BA). There was no evidence of heterogeneity (chi-squared: 0.60; df: 4 [P ¼ 0.96]; I2: 0%). BA had a significantly higher rate of aortic aneurysm formation after intervention than the surgery group (M-H fixed OR: 0.12 [95% CI: 0.04e0.34]; P < 0.001; Fig. 6). The publication bias was significant (Fig. 7). Recoarctation Short-term recoarctation. Information regarding short-term recoarctation was provided by 4 studies with a total of 214 patients (106 surgery, 108 BA). There was no evidence of heterogeneity (chisquared: 1.27; df: 3 [P ¼ 0.74]; I2: 0%). BA had a significantly higher rate of recoarctation after intervention than the surgery group during the
Long-term recoarctation. Long-term recoarctation was reported in 176 surgery and 103 BA patients in 3 studies. There was no evidence of heterogeneity (chi-squared: 0.29; df: 3 [P ¼ 0.96]; I2: 0%). The difference of long-term recoarctation rate was not significant between the 2 groups (M-H fixed OR: 0.61; [95% CI: 0.34e1.11]; Fig. 11). The publication bias was not significant (Fig. 12).
DISCUSSION This meta-analysis confirms that BA reduces the risk of severe complications and have the same effectiveness on reducing peak systolic pressure gradient across the coarctation, suggesting that BA provides immediate results comparable to surgery and reduces invasion. This is in agreement with Wong et al.,17 who used techniques of decision analysis. In addition, this study shows that BA does not reduce the risk of recoarctation but increases the risk of aortic aneurysm formation. As a less invasive treatment, BA was expected to take the place of surgery in treating COA. Because of the low morbidity of COA, the number of available randomized trials is still limited. In the past years, several observational reports and case-matched
400 Hu et al.
Annals of Vascular Surgery
Fig. 6. Comparison of surgery versus balloon angioplasty for mid- to long-term aneurysm formation.
Fig. 7. Funnel plot of comparison: mid- to long-term aneurysm formation.
studies have been published, but the number of patients enrolled in each trial is small, and the statistical power of most trials is too low to document significant differences in clinical outcomes between surgery and BA. Therefore, the issue of whether BA could take the place of surgery is still a matter of debate. To obtain better insight into this important clinical issue, we carried out a meta-analysis of the published results. There was not much debate on the early outcomes of these 2 treatments. When considering the follow-up outcomes, however, no consensus had been reached. The incidence of late complications postintervention plays a significant role in diminished survival. These complications include recoarctation, aortic aneurysm formation, and systemic hypertension. Unfortunately, few studies have reported the morbidity of postintervention systemic hypertension, preventing us from making a comparison on the risk of hypertension after these
2 treatments. Most studies had drawn the conclusion that the recoarctation rate of BA is higher than that of surgery.5,6,10,15,17,18 Our meta-analysis is only consistent with them when regarding the short-term recoarctation incidence. This may be in part because many patients with recoarctation accepted several BAs and eventually got effective relief of coarctation. These patients were regarded as recoarctation patients in the short-term followup assessment. More patients experienced aortic aneurysm after BA than after surgery. This may apply to elastic fiber rupture. Additional study is needed to check this hypothesis and seek redress to prevent that complication. As was reported by Abraham et al.,19 a metaanalysis of well designed, nonrandomized comparative studies of surgical procedures is as good as randomized controlled trials. Because there were not enough randomized controlled trials could be included, both randomized controlled trials and observational studies were included in this metaanalysis. We assessed the level of evidence using the GRADE approach. Relying on study design alone has apparent limitations; GRADE provides additional quality criteria that serve to overcome this shortcoming.20 According to the GRADE approach, the quality of the evidence ranges from very low to moderate for the following reasons: (1) the lack of allocation concealment and blinding, (2) the number of patients is small, and (3) publication bias. Because it was indicated by funnel plot of each comparison, there was publication bias in most comparisons. The inclusion of observational studies is the main source of these publication biases, because studies that are not randomized or controlled tend to have more potential for bias.21,22 Small studies are more likely to be unpublished.23 It is not easy for most
Vol. 28, No. 2, February 2014
Surgical vs. balloon angioplasty for native COA 401
Fig. 8. Comparison of surgery versus balloon angioplasty for short-term recoarctation.
Another limitation of this study is that methodologic features and quality varied substantially, and only 4 randomized controlled trials were included.
CONCLUSIONS
Fig. 9. Funnel recoarctation.
plot
of
comparison:
short-term
medical centers to accumulate enough clinical data of a disease with such a low morbidity rate. This may be another reason for the publication bias. Although the quality of the evidence is not so high, and the recommendation for BA in treating COA yielded from this study is weak, the conclusion of this meta-analysis is still valuable, especially before a high quality randomized controlled trial of COA is published. A limitation of this meta-analysis is that it does not compare outcomes of stent graft with those of surgery, because few studies comparing stent graft with surgery were available. Padua et al.24 have tried to perform a meta-analysis on this topic but have failed for the same reason. Several clinical series have documented that stents could decrease coarctation restenosis, created by vessel recoil, and could diminish the late incidence of aneurysm formation.18,25 Stent implantation may be a good supplement to BA, and its effectiveness and complications should be compared with surgery in the future.
In conclusion, BA provides immediate results comparable to surgery and reduces invasion, but it does not provide better results compared with surgery when considering mid- and long-term complications, and even increases aneurysm formation incidence. These findings underscore the compelling need for prospective, multicenter, randomized clinical trials to determine the clinical benefits (or drawbacks) of BA procedures. Research on the mechanism and prevention of the higher incidence of aneurysm formation after BA is essential for promoting BA as a replacement of surgery for COA therapy. REFERENCES 1. Westaby S, Parnell B, Pridie RB. Coarctation of the aorta in adults. Clinical presentation and results of surgery. J Cardiovasc Surg 1987;28:124e7. 2. Brouwer RM, Erasmus ME, Ebels T, et al. Influence of age on survival, late hypertension, and recoarctation in elective aortic coarctation repair. Including long-term results after elective aortic coarctation repair with a follow-up from 25 to 44 years. J Thorac Cardiovasc Surg 1994;108:525e31. 3. Toro-Salazar OH, Steinberger J, Thomas W, et al. Long-term follow-up of patients after coarctation of the aorta repair. Am J Cardiol 2002;89:541e7. 4. Korostelev AN, Kokov LS, Cherniak BB, et al. Long-term outcomes of balloon angioplasty for aortic coarctation. Angiol Sosud Khir 2003;9:116e21. 5. Veyre P, Bozio A, Jocteur-Monrozier D, et al. Re-stenosis of aortic coarctation in children. Comparison between aortic angioplasty and surgery. Arch Mal Coeur Vaiss 1994;87: 581e5. 6. Rodes-Cabau J, Miro J, Dancea A, et al. Comparison of surgical and transcatheter treatment for native coarctation of the aorta in patients > or ¼ 1 year old. The Quebec Native
402 Hu et al.
Annals of Vascular Surgery
Fig. 10. Comparison of surgery versus balloon angioplasty for midterm recoarctation.
Fig. 11. Comparison of surgery versus balloon angioplasty for long-term recoarctation.
Fig. 12. Funnel recoarctation.
plot
of
comparison:
long-term
Coarctation of the Aorta study. Am Heart J 2007;154: 186e92. 7. Shaddy RE, Boucek MM, Sturtevant JE, et al. Comparison of angioplasty and surgery for unoperated coarctation of the aorta. Circulation 1993;87:793e9. 8. Forbes TJ, Kim DW, Du W, et al. Comparison of surgical, stent, and balloon angioplasty treatment of native coarctation of the aorta: an observational study by the CCISC (Congenital Cardiovascular Interventional Study Consortium). J Am Coll Cardiol 2011;58:2664e74.
9. Chiu HH, Chiu SN, Hu FC, et al. Late cardiovascular complications after surgical or balloon angioplasty of coarctation of aorta in an Asian cohort. Am J Cardiol 2009;104: 1139e44. 10. Fiore AC, Fischer LK, Schwartz T, et al. Comparison of angioplasty and surgery for neonatal aortic coarctation. Ann Thorac Surg 2005;80:1659e64. 11. Rao PS, Chopra PS, Koscik R, et al. Surgical versus balloon therapy for aortic coarctation in infants < or ¼ 3 months old. J Am Coll Cardiol 1994;23:1479e83. 12. Hernandez-Gonzalez M, Solorio S, Conde-Carmona I, et al. Intraluminal aortoplasty vs. surgical aortic resection in congenitalaortic coarctation. Arch Med Res 2003;34:305e10. 13. Fruh S, Knirsch W, Dodge-Khatami A, et al. Comparison of surgical and interventional therapy of native and recurrent aortic coarctation regarding different age groups during childhood. Eur J Cardiothorac Surg 2011;39:898e904. 14. Walhout RJ, Lekkerkerker JC, Oron GH, et al. Comparison of surgical repair with balloon angioplasty for native coarctation in patients from 3 months to 16 years of age. Eur J Cardiothorac Surg 2004;25:722e7. 15. Cowley CG, Orsmond GS, Feola P, et al. Long-term, randomized comparison of balloon angioplasty and surgery for native coarctation of the aorta in childhood. Circulation 2005;111:3453e6. 16. Brozek JL, Akl EA, Alonso-Coello P, et al. Grading quality of evidence and strength of recommendations in clinical practice guidelines. Part 1 of 3. An overview of the GRADE approach and grading quality of evidence about interventions. Allergy 2009;64:669e77.
Vol. 28, No. 2, February 2014
17. Wong D, Benson LN, Van Arsdell GS, et al. Balloon angioplasty is preferred to surgery for aortic coarctation. Cardiol Young 2008;18:79e88. 18. Duara R, Theodore S, Sarma PS, et al. Correction of coarctation of aorta in adult patientsdimpact of corrective procedure on long-term recoarctation and systolic hypertension. Thorac Cardiovasc Surg 2008;56:83e6. 19. Abraham NS, Byrne CJ, Young JM, et al. Meta-analysis of well-designed nonrandomized comparative studies of surgical procedures is as good as randomized controlled trials. J Clin Epidemiol 2010;63:238e45. 20. Terracciano L, Brozek J, Compalati E, et al. GRADE system: new paradigm. Curr Opin Allergy Clin Immunol 2010;10:377e83.
Surgical vs. balloon angioplasty for native COA 403
21. Jones DR. Meta-analysis of observational epidemiological studies: a review. J R Soc Med 1992;85:165e8. 22. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177e88. 23. Ioannidis JP, Trikalinos TA. The appropriateness of asymmetry tests for publication bias in meta-analyses: a large survey. CMAJ 2007;176:1091e6. 24. Padua LM, Garcia LC, Rubira CJ, et al. Stent placement versus surgery for coarctation of the thoracic aorta. Cochrane Database Syst Rev 2012;5:D8204. 25. Barron DJ, Lamb RK, Ogilvie BC, et al. Technique for extraanatomic bypass in complex aortic coarctation. Ann Thorac Surg 1996;61:241e4.
Background: Native coarctation of the aorta (COA) accounts for 5e7% of congenital heart disease. Open surgical treatment was the only choice until balloon angioplasty (BA) treatment was introduced as an alternative therapy for COA in the 1980s. BA treatment was thought to be a less invasive and potentially safer technique, and has been used on numerous patients. But as has been reported during the past 30 years, the risk of aneurysm formation and recoarctation existed in either of those 2 procedures. Unfortunately, follow-up for either type of treatment has been limited, making it difficult to draw any meaningful conclusions as to which treatment option is superior. Our objective was to compare results of 2 therapeutic modalities to treat native COA: BA without stent implantation and surgery. Methods: We performed a meta-analysis of controlled trials of surgical versus BA treatment for native COA. MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials, CINAHL, Web of Science, and the Chinese Biomedical Database of clinical trials were searched using PubMed and OVID. Controlled trials in which patients with COA were assigned to surgical repair or BA treatment were included. For each outcome, we evaluated the quality of the evidence with reference to the Grading of Recommendations Assessments, Development, and Evaluation criteria. We used RevMan 5.1 software (The Nordic Cochrane Centre, Copenhagen, Denmark) to analyze the data. Results: A literature search yielded 9 comparable studies, for a total of 623 patients, of whom 378 and 245 were assigned to surgery and BA. Meta-analysis of these studies showed no significant difference in postintervention gradient (inverse variance fixed mean difference: 1.44 [95% CI: e1.16 to 4.04]), midterm recoarctation (ManteleHaenszel [M-H] random odds ratio [OR]: 0.24 [95% CI: 0.04e1.58]), and long-term recoarctation (M-H fixed OR: 0.61 [95% CI: 0.34e 1.11]). BA reduces the risk of severe complications (M-H fixed OR: 2.67 [95% CI: 1.37e5.21]; P < 0.001) but increases the risk of short-term recoarctation (M-H fixed OR: 0.25 [95% CI]: 0.12e0.54]; P < 0.001) and aortic aneurysm formation (M-H fixed OR: 0.12 [95% CI]: 0.04e 0.34]; P < 0.001). Conclusions: BA provides immediate results comparable to surgery and reduces invasion, but it does not provide better results compared with surgery when considering medium- and longterm complications and even increases the incidence of aneurysm formation.
1 Department of Cardiothoracic Surgery, Renmin Hospital of Wuhan University, Wuhan, People’s Republic of China. 2 Department of Cardiothoracic Surgery, Xiangyang central Hospital, Xiangyang, People’s Republic of China.
Correspondence to: Zhi-wei Wang, MD, Department of Cardiothoracic Surgery, Renmin Hospital of Wuhan University, 99 Ziyang Road, Wuhan, Hubei Province 430060, People’s Republic of China; E-mail: [email protected]
394
Ann Vasc Surg 2014; 28: 394–403 http://dx.doi.org/10.1016/j.avsg.2013.02.026 Ó 2014 Elsevier Inc. All rights reserved. Manuscript received: October 28, 2012; manuscript accepted: February 17, 2013; published online: November 5, 2013.
Vol. 28, No. 2, February 2014
Surgical vs. balloon angioplasty for native COA 395
Table I. Patient demographics Reference, year
Participants
Intervention
Mean age (yrs or days ± SD)
Weight (kg ± SD)
M/F
Shaddy et al., 1993
20 16 72 61 128 41 34 23 14 15 30 28 46 13 18
Surgery BA Surgery BA Surgery BA Surgery BA Surgery BA Surgery BA Surgery BA Surgery
5.7 ± 2.1 yrs 6.3 ± 2.0 yrs 10.0 ± 9.7 yrs 9.8 ± 8.0 yrs 3.3 ± 7.2 yrs 3.0 ± 4.9 yrs 7.7 days ± NA 8.8 days ± NA 27 ± 35 days 29 ± 27 days 7.00 ± 4.14 yrs 6.66 ± 4.44 yrs NA NA Range (0.3e15 yrs) median (0.63 yrs) Range (0.25e15 yrs) median (5.8 yrs) 5.7 ± 2.1 yrs 6.3 ± 2.0 yrs
19.5 21.6 35 30 NA NA 3.4 3.6 3.5 3.8 26.51 25.47 NA NA NA
16/4 8/8 50/22 39/22 27/14 75/53 20/14 15/8 8/6 9/6 22/8 20/8 NA NA NA
Forbes et al., 2011 Chiu et al., 2009 Fiore et al., 2005 Rao et al., 1994 Hernandez-Gonzalez et al., 2003 Fruh et al., 2011 Walhout et al., 2004
Cowley et al., 2005
28
BA
16 20
Surgery BA
± ± ± ±
4.816 8.1 24 21
± ± ± ± ± ±
NA NA 0.9 1.0 15.17 12.54
NA 19.9 ± 5.2 21.7 ± 8.1
NA NA NA
BA, balloon angioplasty; F, female; M, male; NA, not available; SD, standard deviation.
INTRODUCTION Open surgery was the only treatment for coarctation of the aorta (COA) until balloon angioplasty (BA) treatment was introduced as an alternative therapy in the 1980s.1 Since then, aortic angioplasty, thought to be a less invasive and safer treatment, has been adopted for COA therapy. But as has been reported during the past 30 years, complications, such as aortic rupture along with life-threatening bleeding, aneurysm formation, and recoarctation, have been associated with both of these procedures.2e4 However, a large, multicenter, randomized, controlled trial comparing acute and follow-up results of the surgical and BA procedures for COA has not been performed, making it difficult to draw any meaningful conclusions as to which treatment option is superior. Several clinical trials have compared outcomes of surgery versus BA treatment for COA, but they have not included enough patients and their results have been inconsistent. We therefore conducted a meta-analysis of all published clinical trials, either randomized or observational, comparing outcomes of surgery versus BA treatment for COA.
METHODS Search Strategy Comprehensive searches of the MEDLINE and Cochrane Central Register of Controlled Trial
databases were performed by 2 reviewers independently using Web-based search engines (i.e., OVID and PubMed) for human studies published in English between 1980 and October 2012. Search terms included the following: coarctation of the aorta, aortic coarctation, COA, surgical procedures, operative, angioplasty, balloon, surger*, surgical*, endovasc*, and transcather*. Bibliographies of retrieved articles were searched for other relevant studies. We monitored major scientific meetings for the results of trials presented before publication. Bibliographies of included articles were also searched. No language restrictions were applied. If there was sample overlap between studies, we included only the larger study.
Selection Criteria Studies were included if they met each of the following conditions: the trial was randomized and controlled or was a conservative study comparing acute and follow-up results of surgical and BA procedures for COA, and if the trail or study reported 1 of the following clinical outcomes: posttreatment gradient, severe complications, hospital stay days, aortic aneurysm formation, recoarctation during short-term, midterm, or long-term follow up. Studies that did not make a distinction between BA with and without stent implantation were excluded.
Quality assessment
No. of patients (%)
Effect Quality importance
No. of studies
Limitation
Risk of bias Inconsistency
Indirectness
Other Imprecision considerations Surgery
No serious Seriousb Observational No serious No serious risk of inconsistency indirectness studiesa bias Postintervention Observational No serious No serious No serious Seriousb a gradient (4) studies risk of inconsistency indirectness bias Severe Observational No serious No serious No serious Seriousb complications studiesa risk of inconsistency indirectness (4) bias Preintervention gradient (5)
BA
Relative (95% CI)
Very strong associationc
116
110
d
None
76
93
d
None
46/122 (37.7)
35/126 OR 2.67 (27.8) (1.37 to 5.21)
23.3%
Observational No serious No serious Mid-to longNo serious Seriousb studiesa term risk of inconsistency indirectness aneurysm bias formation (6)
Very strong associationc
0/228 (0) 23/147 OR 0.12 (15.6) (0.04 to 0.34)
16.5%
Short-term recoarctation (4)
Observational No serious No serious No serious Seriousb studiesa risk of inconsistency indirectness bias
Reporting biasd 14/106 (13.2)
33/108 OR 0.25 (30.6) (0.12 to 0.54)
30.4%
Observational No serious No serious No serious Seriousb studiesa risk of inconsistency indirectness bias
Reporting biasd 6/78 (7.7) and very strong associationc
28/49 OR 0.05 (57.1) (0.01 to 0.37)
53.9%
MD 3.64 higher (1.03 lower to 8.3 higher) MD 1.44 higher (1.16 lower to 4.04 higher) 229 more per 1,000 (from 67 more to 389 more) 215 more per 1,000 (from 61 more to 380 more) 135 fewer per 1,000 (from 97 fewer to 149 fewer) 142 fewer per 1,000 (from 102 fewer to 157 fewer) 206 fewer per 1,000 (from 114 fewer to 255 fewer) 206 fewer per 1,000 (from 113 fewer to 254 fewer) 509 fewer per 1,000 (from 241 fewer to 558 fewer) 484 fewer per 1,000 (from 237 fewer to 527 fewer)
Low
Important
Very low Critical
Very low Critical
Moderate Critical
Low
Important
Moderate Critical Annals of Vascular Surgery
Mid-term recoarctation (3)
Absolute
396 Hu et al.
Table II. GRADE assessment of each comparison
Vol. 28, No. 2, February 2014
Study Variables
Very low Critical
Patient demographics, such as age, weight, sex, and pressure gradient were recorded as baseline clinical information. Posttreatment gradient is the pressure gradient across the coarctation lesions assessed by catheterization or echocardiography within 1 hour after intervention. Severe complications refer to aortic wall injury, dissection/intimal tear, aneurysm, balloon rupture, femoral atrial fibrillation, severe/prolonged hypertension, bleeding, respiratory failure, and death. Recoactation was defined as a peak systolic pressure gradient 20 mm Hg or inevitable of reintervention. Short-, mid-, and long-term follow-ups were defined as follow-ups of 5 years, respectively.
CI, confidence interval; GRADE, Grading of Recommendations Assessments, Development, and Evaluation; OR, odds ratio. a Lack of allocation concealment and blinding. b The number of patients is small. c OR 2. d Funnel plot of this comparison is asymmetrical.
31/103 OR 0.61 93 fewer per (30.1) (0.34 to 1.11) 1,000 (from 173 fewer to 22 more) 32% 97 fewer per 1,000 (from 182 fewer to 23 more) 48/176 (27.3) None Observational No serious No serious No serious Seriousb studiesa risk of inconsistency indirectness bias Long-term recoarctation (4)
Surgical vs. balloon angioplasty for native COA 397
Data Collection and Statistical Analysis Two authors (H.Z-p. and D.X-f.) independently identified trials for inclusion and extracted information on demographics, interventions, and outcomes. Disagreements were resolved by consensus. For dichotomous variables, odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. For continuous variables, the mean difference (MD) and 95% CIs were calculated. Statistical heterogeneity was measured using the Q statistic (P < 0.10 was considered indicative of statistically significant heterogeneity) and the I2 test. For each outcome, the fixed effect (ManteleHaenszel [M-H] for dichotomous variables and inverse variance for continuous variables) or random effects (DerSimonian for dichotomous and Laird for continuous variables) model was used when the Q statistic suggested a lack or presence of heterogeneity, respectively. Review Manager (RevMan) software (version 5.1; The Nordic Cochrane Centre, Copenhagen, Denmark) was used for data analysis.
RESULTS Included Studies The selection criteria described in the Methods section were applied to the approximately 674 studies identified by the literature search. Of these, 11 matched selection criteria. Veyre et al.’s study5 was in fact a systemic review of case reports, and Rodes-Cabau et al.6 did not make a distinction between BA and stent implantation. Both of these studies were excluded. A total of 9 studies7e15 for a total of 623 patients were included in our study. Patient demographics can be seen in Table I.
398 Hu et al.
Annals of Vascular Surgery
Fig. 1. Comparison of surgery versus balloon angioplasty for pretintervention gradient.
Fig. 2. Comparison of surgery versus balloon angioplasty for postintervention gradient.
and (2) randomized studies and observational studies were analyzed together.
Effects of Interventions
Fig. 3. Funnel plot of comparison: postintervention gradient.
Assessment of the Quality of Evidence Table II describes in detail the principal findings of this review and the quality of the evidence for each outcome using the Grading of Recommendations Assessments, Development, and Evaluation (GRADE) approach.16 When using the GRADE approach, all studies were seen as observational studies for the following reasons: (1) the quality of randomized controlled trials included were low;
Postintervention Gradient. Peak-to-peak systolic pressure gradient within 1 hour after intervention. We obtained 4 trials with a total of 169 patients (76 surgery, 93 BA), with similar preintervention gradient, which was given by Figure 1. There was no evidence of heterogeneity (chi-squared: 2.04; df: 3 [P ¼ 0.56]; I2: 0%). There was no statistical difference of the postintervention gradient between the surgery group and the BA group (inverse variance fixed MD: 1.44 [95% CI: e1.16 to 4.04]; Fig. 2). The funnel plot shows that the publication bias was not significant (Fig. 3). Severe Complications. Severe complications after intervention. Information regarding mid- to longterm aneurysm formation was provided by 4 studies for a total of 248 patients (122 surgery, 126 BA). There was no evidence of heterogeneity (chisquared: 4.63; df: 3 [P ¼ 0.20]; I2 ¼ 35%). BA had a significantly lower rate of severe complications after intervention than the surgery group (M-H
Vol. 28, No. 2, February 2014
Surgical vs. balloon angioplasty for native COA 399
Fig. 4. Comparison of surgery versus balloon angioplasty for short-term recoarctation.
short-term follow-up (M-H fixed OR: 0.25 [95% CI: 0.12e0.54]; P < 0.001; Fig. 8). The publication bias was significant (Fig. 9). Midterm recoarctation. Midterm recoarctation was reported in 78 surgery and 49 BA patients from 3 studies. There was heterogeneity among these 3 studies, possibly because of the use of a different standard for recoarctation (t2: 1.43; chi-squared: 4.2; df: 2 [P ¼ 0.12]; I2: 52%). The difference of midterm recoarctation rate was not significant between the 2 groups (M-H random OR: 0.24 [95% CI: 0.04e1.58]; Fig. 10). Fig. 5. Funnel plot of comparison: severe complications.
fixed OR: 2.67 [95% CI: 1.37e5.21]; P < 0.001; Fig. 4). The publication bias was significant (Fig. 5). Aneurysm Formation Mid- to long-term aneurysm formation. Information regarding mid- to long-term aneurysm formation incidence was provided by 6 studies with a total of 375 patients (228 surgery, 147 BA). There was no evidence of heterogeneity (chi-squared: 0.60; df: 4 [P ¼ 0.96]; I2: 0%). BA had a significantly higher rate of aortic aneurysm formation after intervention than the surgery group (M-H fixed OR: 0.12 [95% CI: 0.04e0.34]; P < 0.001; Fig. 6). The publication bias was significant (Fig. 7). Recoarctation Short-term recoarctation. Information regarding short-term recoarctation was provided by 4 studies with a total of 214 patients (106 surgery, 108 BA). There was no evidence of heterogeneity (chisquared: 1.27; df: 3 [P ¼ 0.74]; I2: 0%). BA had a significantly higher rate of recoarctation after intervention than the surgery group during the
Long-term recoarctation. Long-term recoarctation was reported in 176 surgery and 103 BA patients in 3 studies. There was no evidence of heterogeneity (chi-squared: 0.29; df: 3 [P ¼ 0.96]; I2: 0%). The difference of long-term recoarctation rate was not significant between the 2 groups (M-H fixed OR: 0.61; [95% CI: 0.34e1.11]; Fig. 11). The publication bias was not significant (Fig. 12).
DISCUSSION This meta-analysis confirms that BA reduces the risk of severe complications and have the same effectiveness on reducing peak systolic pressure gradient across the coarctation, suggesting that BA provides immediate results comparable to surgery and reduces invasion. This is in agreement with Wong et al.,17 who used techniques of decision analysis. In addition, this study shows that BA does not reduce the risk of recoarctation but increases the risk of aortic aneurysm formation. As a less invasive treatment, BA was expected to take the place of surgery in treating COA. Because of the low morbidity of COA, the number of available randomized trials is still limited. In the past years, several observational reports and case-matched
400 Hu et al.
Annals of Vascular Surgery
Fig. 6. Comparison of surgery versus balloon angioplasty for mid- to long-term aneurysm formation.
Fig. 7. Funnel plot of comparison: mid- to long-term aneurysm formation.
studies have been published, but the number of patients enrolled in each trial is small, and the statistical power of most trials is too low to document significant differences in clinical outcomes between surgery and BA. Therefore, the issue of whether BA could take the place of surgery is still a matter of debate. To obtain better insight into this important clinical issue, we carried out a meta-analysis of the published results. There was not much debate on the early outcomes of these 2 treatments. When considering the follow-up outcomes, however, no consensus had been reached. The incidence of late complications postintervention plays a significant role in diminished survival. These complications include recoarctation, aortic aneurysm formation, and systemic hypertension. Unfortunately, few studies have reported the morbidity of postintervention systemic hypertension, preventing us from making a comparison on the risk of hypertension after these
2 treatments. Most studies had drawn the conclusion that the recoarctation rate of BA is higher than that of surgery.5,6,10,15,17,18 Our meta-analysis is only consistent with them when regarding the short-term recoarctation incidence. This may be in part because many patients with recoarctation accepted several BAs and eventually got effective relief of coarctation. These patients were regarded as recoarctation patients in the short-term followup assessment. More patients experienced aortic aneurysm after BA than after surgery. This may apply to elastic fiber rupture. Additional study is needed to check this hypothesis and seek redress to prevent that complication. As was reported by Abraham et al.,19 a metaanalysis of well designed, nonrandomized comparative studies of surgical procedures is as good as randomized controlled trials. Because there were not enough randomized controlled trials could be included, both randomized controlled trials and observational studies were included in this metaanalysis. We assessed the level of evidence using the GRADE approach. Relying on study design alone has apparent limitations; GRADE provides additional quality criteria that serve to overcome this shortcoming.20 According to the GRADE approach, the quality of the evidence ranges from very low to moderate for the following reasons: (1) the lack of allocation concealment and blinding, (2) the number of patients is small, and (3) publication bias. Because it was indicated by funnel plot of each comparison, there was publication bias in most comparisons. The inclusion of observational studies is the main source of these publication biases, because studies that are not randomized or controlled tend to have more potential for bias.21,22 Small studies are more likely to be unpublished.23 It is not easy for most
Vol. 28, No. 2, February 2014
Surgical vs. balloon angioplasty for native COA 401
Fig. 8. Comparison of surgery versus balloon angioplasty for short-term recoarctation.
Another limitation of this study is that methodologic features and quality varied substantially, and only 4 randomized controlled trials were included.
CONCLUSIONS
Fig. 9. Funnel recoarctation.
plot
of
comparison:
short-term
medical centers to accumulate enough clinical data of a disease with such a low morbidity rate. This may be another reason for the publication bias. Although the quality of the evidence is not so high, and the recommendation for BA in treating COA yielded from this study is weak, the conclusion of this meta-analysis is still valuable, especially before a high quality randomized controlled trial of COA is published. A limitation of this meta-analysis is that it does not compare outcomes of stent graft with those of surgery, because few studies comparing stent graft with surgery were available. Padua et al.24 have tried to perform a meta-analysis on this topic but have failed for the same reason. Several clinical series have documented that stents could decrease coarctation restenosis, created by vessel recoil, and could diminish the late incidence of aneurysm formation.18,25 Stent implantation may be a good supplement to BA, and its effectiveness and complications should be compared with surgery in the future.
In conclusion, BA provides immediate results comparable to surgery and reduces invasion, but it does not provide better results compared with surgery when considering mid- and long-term complications, and even increases aneurysm formation incidence. These findings underscore the compelling need for prospective, multicenter, randomized clinical trials to determine the clinical benefits (or drawbacks) of BA procedures. Research on the mechanism and prevention of the higher incidence of aneurysm formation after BA is essential for promoting BA as a replacement of surgery for COA therapy. REFERENCES 1. Westaby S, Parnell B, Pridie RB. Coarctation of the aorta in adults. Clinical presentation and results of surgery. J Cardiovasc Surg 1987;28:124e7. 2. Brouwer RM, Erasmus ME, Ebels T, et al. Influence of age on survival, late hypertension, and recoarctation in elective aortic coarctation repair. Including long-term results after elective aortic coarctation repair with a follow-up from 25 to 44 years. J Thorac Cardiovasc Surg 1994;108:525e31. 3. Toro-Salazar OH, Steinberger J, Thomas W, et al. Long-term follow-up of patients after coarctation of the aorta repair. Am J Cardiol 2002;89:541e7. 4. Korostelev AN, Kokov LS, Cherniak BB, et al. Long-term outcomes of balloon angioplasty for aortic coarctation. Angiol Sosud Khir 2003;9:116e21. 5. Veyre P, Bozio A, Jocteur-Monrozier D, et al. Re-stenosis of aortic coarctation in children. Comparison between aortic angioplasty and surgery. Arch Mal Coeur Vaiss 1994;87: 581e5. 6. Rodes-Cabau J, Miro J, Dancea A, et al. Comparison of surgical and transcatheter treatment for native coarctation of the aorta in patients > or ¼ 1 year old. The Quebec Native
402 Hu et al.
Annals of Vascular Surgery
Fig. 10. Comparison of surgery versus balloon angioplasty for midterm recoarctation.
Fig. 11. Comparison of surgery versus balloon angioplasty for long-term recoarctation.
Fig. 12. Funnel recoarctation.
plot
of
comparison:
long-term
Coarctation of the Aorta study. Am Heart J 2007;154: 186e92. 7. Shaddy RE, Boucek MM, Sturtevant JE, et al. Comparison of angioplasty and surgery for unoperated coarctation of the aorta. Circulation 1993;87:793e9. 8. Forbes TJ, Kim DW, Du W, et al. Comparison of surgical, stent, and balloon angioplasty treatment of native coarctation of the aorta: an observational study by the CCISC (Congenital Cardiovascular Interventional Study Consortium). J Am Coll Cardiol 2011;58:2664e74.
9. Chiu HH, Chiu SN, Hu FC, et al. Late cardiovascular complications after surgical or balloon angioplasty of coarctation of aorta in an Asian cohort. Am J Cardiol 2009;104: 1139e44. 10. Fiore AC, Fischer LK, Schwartz T, et al. Comparison of angioplasty and surgery for neonatal aortic coarctation. Ann Thorac Surg 2005;80:1659e64. 11. Rao PS, Chopra PS, Koscik R, et al. Surgical versus balloon therapy for aortic coarctation in infants < or ¼ 3 months old. J Am Coll Cardiol 1994;23:1479e83. 12. Hernandez-Gonzalez M, Solorio S, Conde-Carmona I, et al. Intraluminal aortoplasty vs. surgical aortic resection in congenitalaortic coarctation. Arch Med Res 2003;34:305e10. 13. Fruh S, Knirsch W, Dodge-Khatami A, et al. Comparison of surgical and interventional therapy of native and recurrent aortic coarctation regarding different age groups during childhood. Eur J Cardiothorac Surg 2011;39:898e904. 14. Walhout RJ, Lekkerkerker JC, Oron GH, et al. Comparison of surgical repair with balloon angioplasty for native coarctation in patients from 3 months to 16 years of age. Eur J Cardiothorac Surg 2004;25:722e7. 15. Cowley CG, Orsmond GS, Feola P, et al. Long-term, randomized comparison of balloon angioplasty and surgery for native coarctation of the aorta in childhood. Circulation 2005;111:3453e6. 16. Brozek JL, Akl EA, Alonso-Coello P, et al. Grading quality of evidence and strength of recommendations in clinical practice guidelines. Part 1 of 3. An overview of the GRADE approach and grading quality of evidence about interventions. Allergy 2009;64:669e77.
Vol. 28, No. 2, February 2014
17. Wong D, Benson LN, Van Arsdell GS, et al. Balloon angioplasty is preferred to surgery for aortic coarctation. Cardiol Young 2008;18:79e88. 18. Duara R, Theodore S, Sarma PS, et al. Correction of coarctation of aorta in adult patientsdimpact of corrective procedure on long-term recoarctation and systolic hypertension. Thorac Cardiovasc Surg 2008;56:83e6. 19. Abraham NS, Byrne CJ, Young JM, et al. Meta-analysis of well-designed nonrandomized comparative studies of surgical procedures is as good as randomized controlled trials. J Clin Epidemiol 2010;63:238e45. 20. Terracciano L, Brozek J, Compalati E, et al. GRADE system: new paradigm. Curr Opin Allergy Clin Immunol 2010;10:377e83.
Surgical vs. balloon angioplasty for native COA 403
21. Jones DR. Meta-analysis of observational epidemiological studies: a review. J R Soc Med 1992;85:165e8. 22. DerSimonian R, Laird N. Meta-analysis in clinical trials. Control Clin Trials 1986;7:177e88. 23. Ioannidis JP, Trikalinos TA. The appropriateness of asymmetry tests for publication bias in meta-analyses: a large survey. CMAJ 2007;176:1091e6. 24. Padua LM, Garcia LC, Rubira CJ, et al. Stent placement versus surgery for coarctation of the thoracic aorta. Cochrane Database Syst Rev 2012;5:D8204. 25. Barron DJ, Lamb RK, Ogilvie BC, et al. Technique for extraanatomic bypass in complex aortic coarctation. Ann Thorac Surg 1996;61:241e4.
