Statistics from Altmetric.com
Surgical treatment for coarctation was introduced over 50 years ago and has steadily been refined, with particular improvements in results over the last 10 years or so. Balloon angioplasty was introduced in the mid 1980s, initially for recoarctation and later for unoperated (“native”) coarctation. In this issue Thanopoulos and his colleagues from Athens report their favourable early experience of stent implantation to treat coarctation or recoarctation in a small group of children and adolescents.1 Many factors should be taken into account when deciding upon the best treatment for a patient with coarctation including age, coarctation morphology, whether previous surgery has been undertaken, and local institutional results of each type of treatment. Unfortunately there are no clinical trials of appropriate size or design upon which to base an objective judgement of the optimum form of treatment, but it is helpful to compare approximately contemporary studies of surgery, balloon angioplasty, and stent implantation.
The major complications of surgery are death, paraplegia caused by perioperative spinal cord ischaemia, recurrent or residual coarctation, and late aneurysm formation at the site of the repair. There are also late complications related to the coarctation itself including hypertension, dissecting, diffuse or false aneurysm of the aorta, stroke, and early coronary artery disease. In a large (571 patients) long term study from the Mayo clinic,2 estimated survival at 10, 20, and 30 years was 91%, 84%, and 76%, respectively. Late deaths were most commonly related to coronary artery disease followed by sudden death, heart failure, stroke, and rupture of aortic aneurysm. Young age at operation favourably influenced outcome, but the risk of late death increased the higher the postoperative resting blood pressure. A similar effect of postoperative hypertension on long term outcome was reported in other studies from the UK3 and from mainland Europe.4 The great influence of postoperative upper limb hypertension on long term survival emphasises the extreme importance of optimum relief of aortic obstruction, whatever the mode of treatment and whatever the age of the patient.
Three types of surgical repair remain in common use: resection of the stenosed segment of aorta with end to end anastomosis; use of a subclavian flap; and patch aortoplasty (using synthetic material for the patch). The choice of operative technique is influenced by the morphology of the coarctation, the age of the patient, and the personal preference of the surgeon, although among surgeons specialising in congenital heart disease end to end anastomosis has become the most popular option as it probably provides the best anatomical relief of obstruction, has the lowest risk of recoarctation, and probably has the lowest incidence of late aneurysm formation.5
The major stimuli behind the introduction of therapeutic cardiac catheterisation for the treatment of coarctation were the risk of residual or recurrent coarctation and the rare but catastrophic spinal cord infarction following surgical treatment. Recoarctation (in reality a mixture of residual and recurrent coarctation) occurs in approximately 10% of cases in most published surgical series up until the early 1990s, but in the last decade refinement of the end to end anastomosis can result in a reduction in recoarctation to less than 4%,6 an experience similar to that of my own centre, where (excluding preterm babies) there has been 0.5% reintervention for recoarctation in children and adults operated upon during the last 10 years (unpublished observation). Aortic aneurysm appears to be extremely rare early after end to end anastomosis but it is not yet clear whether modern surgery will reduce the incidence of late aneurysm formation. The risk of paraplegia is probably in the region of 1% in unselected cases, but is higher in patients who have a paucity of collateral vessels around the coarctation.
Balloon angioplasty equipment has improved steadily since the early 1980s (smaller catheters and lower profile, higher pressure balloons) and the technique has been refined too, particularly with regard to choice of inflated balloon diameter. It is generally considered safer in patients with postsurgical recoarctation (because of the supporting effect of surrounding scar tissue) rather than native coarctation, although a large US angioplasty registry found similar acute relief of obstruction in both groups.7 The registry received data on 970 procedures and reported a 0.7% mortality regardless of whether previous surgery had been undertaken or not. A “suboptimal” result was reported in 19% of patients with native coarctation and 25% of patients with recoarctation.
Results of balloon angioplasty are much less favourable in neonates and infants, presumably related to constriction of ductal tissue; a comparison in the early 1990s of pooled results from both surgery and balloon angioplasty in infancy8 suggested an overall incidence of recoarctation of 14% after surgery and 57% after angioplasty. Long segment stenoses are less likely to respond to balloon dilatation than discrete coarctation. In one study of 90 procedures for recoarctation the diameter of the transverse aortic arch was an important predictor of successful reduction in pressure gradient, but even after an initially “optimal” result only 72% of patients remained free from reintervention over a period of 12 years.9 Also of note were two procedure related major neurological complications among the 90 patients, illustrating that neurological damage is not isolated to surgical treatment.
Aneurysm formation, occurring immediately or within a few years, has been reported to occur after 5–20% of coarctation angioplasty,10 11 but the incidence of late aneurysm is still unknown. Acute catastrophic aortic rupture, sometimes fatal, has also been reported occasionally, with one report raising the possibility that previous patch aortoplasty might increase the risk of rupture.12 The relatively high incidence of aneurysms after angioplasty is, perhaps, predictable in the light of studies demonstrating the mechanism of action of balloon dilatation of coarctation; both histology13 and intravascular ultrasound14 suggest that enlargement of the aortic lumen is a result of tearing of the intima and media, that only the adventitia may be left intact, and that dissection or intimal flaps are the rule rather than the exception. Even when there is no obvious angiographic evidence of aneurysm formation, ballooned coarctation segments later resected at surgery show an absence of muscle and elastic lamella strongly suggesting that the substrate for late aneurysm formation is present.15 In the only randomised trial to date of surgery and balloon angioplasty for native coarctation there were 20% aneurysms visible on angiography after angioplasty compared to none after surgery.11
In young children there are additional concerns about vascular damage related to the introduction of relatively large balloon catheters. Magnetic resonance imaging of the femoral and iliac arteries has shown arterial stenosis or occlusion in 39% of cases in one study16 and in some cases the changes were progressive. Symptoms of limb ischaemia appear to be very rare in childhood, but such vascular damage might well prove to be important in adult life.
Are adults different?
Adults pose a slightly different dilemma; recoarctation after an initially good result with balloon angioplasty and vessel damage related to arterial access are both rare, so aneurysm formation, aortic rupture, and incomplete relief of obstruction are the major factors to be taken into account. Aneurysms have been reported to occur after angioplasty in 7–13% of adults,17-19 sometimes requiring urgent surgical treatment. The risk of acute aortic rupture or dissection is probably no different in adults and children. Most small aneurysms do not seem to get larger in the currently limited experience of follow up,17 but it would seem reasonable to view any aneurysm with concern. There is limited data on incomplete relief of aortic obstruction for angioplasty in adults, but in the largest study (of children and adults) to date greater age was a significant independent risk factor for a suboptimal result.7
With simple balloon dilatation of the aorta some degree of elastic recoil is common after balloon deflation; it may be necessary to use a balloon which is larger than the desired final diameter of the aortic lumen. This “overstretching” of the coarctation is thought to be a factor which increases the risk of aneurysm, rupture, and dissection. Stent implantation prevents elastic recoil (potentially allowing the use of slightly smaller balloons) and might, in theory, reduce these risks.
Thanopoulos and colleagues1 used balloon expandable stents in the treatment of coarctation in 17 patients aged between 5 months and 15 years, adding to others' early experience.20-23These series include a total of 97 patients, some with unoperated coarctation and some with recoarctation. In most patients satisfactory relief of aortic arch gradient was achieved, and the pooled risk of acute aneurysm formation was just over 3%. It is too early to know the implications of small acute aneurysms or the incidence of late aneurysm formation, or whether stenting reduces the risk of aneurysm. What is clear from these early studies is that while stenting might reduce aortic overstretching, it does not eliminate aneurysm formation. There is very little experience of covered stents so far; they may be used to treat aneurysms24 and might reduce the risk of aneurysm within the stented segment of aorta, but will not prevent aortic tears at the sharp ends of the stent.
Some degree of restenosis caused by endothelial proliferation occurred in 27% of patients in the largest of these studies21 over a maximum follow up period of 25 months. There must be additional concerns over stent implantation in children in whom growth alone will necessitate stent redilatation. A few patients in the above studies have had stents redilated successfully. Animal studies of aortic stent redilatation have had mixed results, with one favourable experience25 but another26 inducing fatal aortic rupture in nearly a third of cases.
There is widespread agreement that therapeutic catheterisation sometimes has a place in the increasingly rare patient with postsurgical recoarctation at any age. For native coarctation it is apparent that different specialist centres take very different views on the best form of treatment, raising interesting and important issues regarding informed consent. In the absence of available late follow up data for therapeutic catheterisation, informed consent must be based upon an explanation of the differences in early results of treatment.
The hope that angioplasty would avoid the neurological damage occasionally associated with surgery has not been fulfilled, and it seems unlikely that the small risks of spinal cord injury with surgery and stroke with therapeutic catheterisation will prove to be statistically different. Available data suggest that acute relief of aortic obstruction, freedom from reintervention for recoarctation or residual coarctation, and freedom from aortic tears are all better with surgical treatment. These differences are most striking in infants, still clearly evident in older children, and present but less pronounced in adults. In adults there appears to be a slightly lower chance of immediate complete relief of aortic obstruction with angioplasty than there is in children—a serious cause for concern in the light of the major effect of postoperative hypertension on long term outcome. There is additional theoretical concern that surgical treatment following inadequate angioplasty may carry an increased risk of spinal cord damage because collaterals will regress after partial relief of obstruction.15 Stenting does not eliminate the risk of aneurysm, aortic rupture or dissection, but it is not clear yet whether these risks are lower than with angioplasty alone. Even if the unknown long term significance of the common finding of some degree of intimal proliferation after stenting is disregarded, it seems wise to avoid stent implantation in children if at all possible as there will be an almost inevitable need for reintervention as the child grows.