Coronary artery lesions are associated with adverse cardiac events in children undergoing supravalvular aortic stenosis repair

Abstract

OBJECTIVES

The aim of this study was to identify the prevalence and anatomic characteristics of coronary artery lesions and their associated postoperative risk in patients undergoing supravalvular aortic stenosis repair.

METHODS

The association between structural risk factors, postoperative ST-segment changes, and major adverse cardiac events was explored using logistic regression and the Fisher’s exact test.

RESULTS

In 51 consecutive patients with supravalvular aortic stenosis treated between 2000 and 2017, a total of 48 coronary lesions were identified in 27 patients (53%). Prominent ostial ridge (type I) was the most common coronary lesion, followed by small ostium with (IIIb) or without (IIIa) diffuse long-segment coronary narrowing, and adhesion of the coronary cusp (type II). There were 54 concomitant coronary procedures, including 43 primary corrections and 11 revisions. Thirty-three patients underwent supravalvular aortic stenosis repair with a bifurcated patch, of which 13 (39.4%) had right coronary artery distortion/kinking requiring patch plication (n = 8) and reimplantation (n = 5). Postoperative major adverse cardiac events (MACE) occurred in 9 patients (17.6%), including 3 deaths, 4 needing mechanical circulatory support, and 6 experiencing ventricular arrhythmias. Twenty-two patients (43.1%) had postoperative ST-segment changes, including 13 early changes that resolved within 24 h and 9 persistent changes lasting >24 h. Patients with type III lesions were associated with postoperative persistent ST-segment change (P = 0.04) and these lesions independently predicted postoperative MACE (P = 0.02). Patients with pre-existing coronary lesions were at elevated risk of right coronary artery distortion/kinking (P = 0.045).

CONCLUSIONS

The prevalence of ST-segment changes and MACE is high in patients undergoing supravalvular aortic stenosis repair. The preoperative presence of complex coronary lesions is the most important predictor for postoperative major adverse cardiac events.

Supravalvular aortic stenosis (SAS) is a rare congenital obstructive lesion at the sinotubular junction (STJ) level, predominantly caused by elastin arteriopathy [1].

INTRODUCTION

Supravalvular aortic stenosis (SAS) is a rare congenital obstructive lesion at the sinotubular junction (STJ) level, predominantly caused by elastin arteriopathy [1]. In the progressive form of SAS, the structures proximal to the STJ, such as the aortic valve leaflets and coronary artery (CA), can develop secondary pathologies. Several mechanisms of CA involvement have been reported [2]. The most common is CA ostial stenosis, caused by intimal fibrotic deposit around the ostium. This is classified as part of the aortic component of the disease. Inflow to the CA ostium can also be restricted due to adhesion of the aortic valve leaflet edge to the stenotic STJ. In patients with more severe phenotypes, CAs may have primary structural changes of elastin arteriopathy, which often results in various degrees of diffuse intrinsic stenosis.

CA involvement is a well-known complicating factor in SAS repair and is strongly correlated with an increased risk of procedure-related sudden death [3]. The majority of operative deaths are directly [4–6] or indirectly [7, 8] related to myocardial ischaemia due to CA pathology. Furthermore, SAS repair involves patch augmentation of the aortic sinuses, which may negatively alter proximal CA geometry [9]. There are a number of research questions that arise, including whether patients with CA lesions have a higher incidence of major adverse cardiac events (MACE) after SAS repair, if the severity of CA lesions increases operative risk, and whether patch reconstruction increases CA distortion and subsequent MACE. This study’s primary aims are to elucidate the incidence and anatomic characteristics of the CA lesions in SAS, and secondarily to determine the associations of CA lesions with postoperative MACE, persistent ST-segment change, and early survival.

PATIENTS AND METHODS

We retrospectively reviewed SAS patients’ medical records who underwent surgical repair between January 2000 and December 2017. The study included 51 consecutive patients.

Ethical statement

The Research Ethics Board at the Hospital for Sick Children approved the study (1000033868, 2012/7/31) and waived the requirement for patient consent. Approval is renewed on an annual basis and was last updated on 4 January 2023.

Coronary artery lesions and intervention

The CA lesion classification is based on previously published literature with minor modifications [11]. The definition of each type of CA lesions and associated surgical intervention is summarized in Fig. 1 and Supplementary Material, Table S1. In brief, type I lesion is defined as a prominent ridge of the thickened aortic sinus wall around the ostium with or without obstructing coronary blood flow, where the true ostial size itself is appropriate. Type II lesion is characterized by adhesion of the free edges of the coronary cusp to the narrowed STJ, separating the coronary ostium from the lumen of the aorta. Type III lesions are divided into 2 subgroups: an intrinsically small coronary artery ostium with (IIIb) or without (IIIa) diffuse long-segment coronary artery stenosis.

Figure 1:

Coronary artery lesions and surgical interventions. Preoperative catheterization (A–D, *), illustrations (E–F, *) and surgical management (I–K) of different types of coronary artery lesions. For type I lesion (A and E), the ostial ridge was entirely resected by a blade until the coronary ostium became unobstructed (I). For patients with type II lesions (B and F), the top of the commissure of the affected leaflet was detached sharply from the aortic wall with a blade (J). Type IIIA (C and G) and type IIIB lesions (B and H) with a critical ostial lesion or transoesophageal echocardiographic evidence of significant flow acceleration were treated with surgical patch ostioplasty (K).

Surgical management of supravalvular aortic stenosis

The indications of SAS repair are summarized in Supplementary Material, Table S2. Surgical procedures were performed through a median sternotomy with mild hypothermic cardiopulmonary bypass (CPB) and antegrade cardioplegic arrest. Deep hypothermic circulatory arrest was used for concomitant aortic arch reconstruction. The surgical techniques utilized for SAS repair were based on patient weight, aortic anatomy, associated CA lesions, and surgeons’ discretion. The standard approach was the two-sinus reconstruction, either with an inverted-Y-shaped patch or 2 separate patches to both the noncoronary and right coronary sinuses [9]. The glutaraldehyde-treated pericardium (treated for 5 min) was the preferred material of choice; however, bovine pericardium was occasionally used. The single-patch technique to enlarge the noncoronary sinus, or the triple-sinus patch reconstruction technique was occasionally employed. At our institution, patchless sliding aortoplasty is not performed due to concern for inadequate aortic augmentation and residual stenosis.

Intraoperative evaluation of coronary artery flow

Pre-CPB transoesophageal echocardiogram (TEE) is routinely done in patients with SAS with suspicious CA lesions. Intraoperative post-CPB TEE is routinely performed for all patients undergoing SAS repair, with low threshold to supplement with epicardial echocardiogram (EpE) in the immediate post-bypass period to assess for coronary artery flow, as described in previous literature [12]. In general, the coronary arteries were revised if there was an increased peak flow velocity (>100 cm/s) or abnormal flow pattern, such as retrograde or continuous flow as per the published institutional criteria [12].

Data collection

We collected the following data: demographic information, the presence of Williams syndrome, preoperative arrest, use of extracorporeal membrane oxygenation (ECMO), baseline anatomic variables measured by echocardiogram, operative variables including detailed concomitant CA procedures, and postoperative outcomes.

CA lesion diagnosis was based on the consensus of a staff surgeon (Christoph Haller, Osami Honjo), a staff echocardiographer (Lynne E. Nield) and a senior surgical fellow (Shuhua Luo) after a thorough review of the preoperative catheterization, both preoperative and intraoperative echocardiographic imaging, and operative notes. The coronary lesions were divided into 3 groups: none, simple, and complex lesions. Types I and II were defined as simple CA lesions. Any CA associated with type III was defined as complex lesions in this study. Patients are categorized under the ‘complex’ lesion group if they exhibit at least 1 coronary artery with type III pathology. The primary outcome was the incidence of postoperative MACE, defined as the composite of postoperative death, ventricular tachycardia (VT)/ventricular fibrillation (VF), cardiac arrest or need for ECMO [13]. Other outcomes evaluated included postoperative ST-segment changes, unplanned reoperation, the presence of an open sternum after surgery, duration of mechanical ventilation, length of intensive care unit stay and hospitalization. Postoperative ST-segment changes were defined as either ST elevation or depression in the operating room or the cardiac critical care unit. Follow-up data were collected through institutional chart review of postoperative clinic visits and imaging.

Statistical analysis

Continuous variables were expressed as a median and interquartile range or as absolute numbers (with percentage). Variables among groups with complex, simple or no CA lesions, were compared using the Wilcoxon rank sum, Krustal–Wallis or Fisher’s test as appropriate. The association between preoperative factors and MACE was also explored using univariable logistic regression. We applied univariable logistic regression to explore the crude association between preoperative and intraoperative factors, as well as in-hospital MACE. We considered CPB and cross-clamp time as confounders as both variables were associated with the presence of coronary lesions, as well as the outcome. In the multivariable logistic regression, we assessed the odds ratios between MACE and coronary lesion, adjusted for age, weight and cross-lamp time. The corresponding 95% confidence interval (CI) and p-values were evaluated using Wald’s statistics. Freedom from death was analysed using the Kaplan–Meier method. Statistical analysis was performed with Stata, version 14.

RESULTS

Demographics

Patients’ demographics and baseline aortic anatomy are presented in Supplementary Material, Table S3. The median age and weight were 13.4 (5.2, 57.4) months and 9.4 (6.1, 15.3) kg, respectively. Approximately one-half of patients had Williams syndrome. The majority of patients had no preoperative clinical manifestations of coronary ischaemia; however, 3 had preoperative cardiac arrest requiring extracorporeal cardiopulmonary resuscitation (ECPR). Two of those 3 patients had CA lesions. One had a severe type I lesion in the left CA. The other had a type I lesion in the left CA, as well as a combined type I and type IIIA lesion in the right CA. All 3 patients required urgent SAS repair.

Coronary artery lesions

A total of 27 patients (52.9%) had associated CA lesions, approximately half (13/27, 48.1%) had a type III coronary lesion (complex coronary lesion group), either as an isolated lesion or a part of the combined lesion, while the remaining 14 patients had a type I (n = 21) and/or type II (n = 2) lesion (simple coronary lesion group).

Among a total of 48 CA lesions, type I lesion was the most common lesion (30/48, 62.5%), followed by type III (15/48, 31.2%), then type II lesions (3/48, 6.3%). Approximately two-thirds of the type II (2/3, 66.7%) and type IIIA lesions (7/11, 63.6%) were detected by preoperative transthoracic echocardiogram or coronary angiogram. On the contrary, the majority of type I (24 out of 30, 80.0%) lesions were detected by intraoperative inspection. In total, 4 CA lesions (4 of 48, 8.3%) were diagnosed after the initial repair, which included 2 type IIIA and 2 type IIIB lesions. None of those patients had an initial intervention on the CA. The 2 type IIIA lesions were recognized as a smaller-than-normal coronary ostium by intraoperative inspection, but significant flow acceleration through the CA was detected by EpE after coming off from CPB. One type IIIB lesion was diagnosed by EpE as the combination of hypoplasia and right CA kinking was caused by the reconstructed patch. Another type IIIB lesion was confirmed by postoperative coronary angiogram following cardiac arrest.

Operative characteristics and concomitant coronary procedures

A concomitant CA procedure was defined as any CA surgery performed during SAS repair, including primary repair (n = 43) and revision (n = 11, Fig. 2). Surgical techniques used for SAS repair were comparable among groups but the patients with complex CA lesions had significantly longer CPB (P = 0.04) and cross-clamp times (P = 0.01) than simple or no CA lesion groups (Table 1).

Figure 2:

The concomitant and subsequent coronary procedures during supravalvular aortic stenosis repair. Twenty-four concomitant, 1 delayed and 1 revisional coronary procedures were performed in patients with complex coronary lesions. Twenty-five concomitant and 2 revisional coronary procedures were performed in patients with simple coronary lesions. Three patch plications were performed in patients with no coronary artery lesions due to patch-related coronary distortion/kinking. Numbers in the boxes indicate the number of coronary artery lesions under each type of lesion type and repair technique. ‘n’ indicates number of patients.

Table 1:

Operative characteristics

Characteristics	Overall (n = 51)	Complex CA lesions (n = 13)	Simple CA lesions (n = 14)	No CA lesions (n = 24)	P-value*
CPB time (min), median (IQR)	95 (70–145)	153 (89–205)	98 (68–128)	84 (65–133)	0.04
Cross clamp time (min), median (IQR)	65 (51–110)	126 (72–168)	58.5 (36–107)	61.5 (48.5–87)	0.01
Surgical technique of SAS, n (%)					0.54
One patch	7 (13.7)	1 (7.7)	3 (21.4)	3 (12.5)
Two patches: bifurcated	33 (64.7)	8 (61.5)	7 (50.0)	18 (75.0)
Two patches: separated	8 (15.7)	4 (30.8)	3 (21.4)	1 (4.2)
Three patches	3 (5.9)	0 (0)	1 (7.1)	2 (8.3)
Concomitant coronary procedures (n)	54	24	27	3	0.0001
Primary coronary procedures, n (%)	43 (79.6)	18 (75.0)	25 (92.6)	0	0.0001
Ridge resection	30 (69.8)	9 (50.0)	21 (84.0)	0	0.0001
Excision of fused leaflet	3 (7.0)	1 (5.5)	2 (8.0)	0	0.76
Ostioplasty	6 (1.4)	6 (33.3)	0	0	0.004
Coronary button reimplantation	4 (9.3)	2 (11.1)	2 (8.0)	0	0.48
Intraoperative coronary revision, n (%)	11 (20.4)	6 (25.0)	2 (7.4)	3 (100.0)	0.03
Ostioplasty	2 (18.2)	2 (33.3)	0	0
Patch plication	8 (72.7)	3 (50.0)	2 (100.0)	3 (100.0)
Coronary button reimplantation	1 (9.1)	1 (1.7)	0	0
All types of coronary revision, n (%)	15 (29.4)	7 (53.8)	3 (21.4)	5 (20.8)	0.08
Right CA kinking (n)	13	6	4	3	0.08

^*Comparison among patients with complex coronary artery lesions, with simple coronary artery lesions and with no coronary artery lesions. Bold indicates statistical significance with an alpha of 0.05.

CA: coronary artery; CPB: cardiopulmonary bypass; IQR: interquartile range; SAS: supravalvular aortic stenosis.

A total of 54 concomitant coronary procedures were performed in 30 patients. As expected, the number of intraoperative CA interventions (P < 0.001) and revisions (P = 0.03) was highest in the complex coronary group.

Right coronary artery distortion

In patients undergoing SAS repair with an inverted-Y patch (inverted-Y or Doty type repair), 13 (39.4%) had echocardiographic evidence of impaired coronary blood flow and/or haemodynamic instability due to iatrogenic right coronary artery distortion and kinking caused by the bulging of the adjacent redundant longitudinal dimension of the patch (Supplementary Material, Table S4), which required revision [9]. The right CA distortion or kinking was significantly more common in patients with CA lesions (CA lesions group: 10/15 vs normal coronary group: 3/18, P = 0.045).

In-hospital outcomes

Approximately half of the patients (22/51, 43.1%) had ST-segment changes. Nine patients demonstrated persistent ST-segment changes for >24 h (Table 2), which was concerning for residual coronary stenosis or failed ostioplasty. Out of 13, 14 and 24 patients with complex, simple or no lesions, 46%, 21% and 0% of respective patients developed MACE (P < 0.001). Transient ST-segment changes were reported in patients with no coronary lesion and are an expected sequelae of afterload reduction after SAS repair. Approximately half (22/51, 43.1%) patients had ST-segment change requiring high diastolic pressure during intensive care unit. One patient with type I lesions on both coronary arteries. He received concomitant ridge resection on both coronary arteries. The immediate post-CPB TEE showed patent LCA while flow acceleration on the right coronary artery. The postoperative catheterization indicated by persistent ST-segment depression requiring high blood pressure showed patent left and right CA.

Table 2:

Postoperative recovery

Characteristics	Overall (n = 51)	Complex CA lesions (n = 13)	Simple CA lesions (n = 14)	No CA lesions (n = 24)	P–value*
Delayed chest closure, n (%)	10 (19.6)	4 (30.7)	1 (7.1)	5 (20.8)	0.30
ST-segment change, n (%)	22 (43.1)	8 (61.5)	8 (57.1)	6 (25.0)	0.05
Early ST-segment change	13 (25.5)	2 (15.4)	5 (35.7)	6 (25.0)	0.48
Persistent ST-segment change	9 (17.6)	6 (46.1)	3 (21.4)	0 (0.0)	0.002
Intubation (days), median (IQR)	1 (0, 4)	3 (1, 5)	1 (0, 2)	0.5 (0, 3.5)	0.06
ICU duration (days), median (IQR)	2 (1, 6)	5 (3, 6)	2 (1, 4)	1 (1, 5.5)	0.37
Hospital stay (days) (range)	6 (4–12)	8 (6–15)	4.5 (3–7)	5.5 (3.5–11)	0.03
Requiring high diastolic pressure, n (%)	22 (43.1)	8 (61.5)	8 (57.1)	6 (25)	0.05
MACE composite, n (%)	9 (17.6)	7 (60.0)	1 (7.7)	1 (4.3)	<0.001
Death (n)	3	1	1	1	0.89
VT/VF (n)	2	2	0	0	0.02
Cardiac arrest (n)	4	4	0	0	<0.001
ECMO (n)	4	2	1	1	0.48

^aComparison among patients with complex coronary artery lesions, with simple coronary artery lesions and with no coronary artery lesions. Bold indicates statistical significance with an alpha of 0.05.

CA: coronary artery; ECMO: extracorporeal membrane oxygenation; ICU: intensive care unit; IQR: interquartile range; MACE: major adverse cardiac events; VF: ventricular fibrillation; VT: ventricular tachycardia.

Postoperative MACE occurred in 9 of the 51 patients (17.6%), including 3 deaths, 4 episodes of postoperative ECMO, 2 VT/VF and 4 cardiac arrests (Supplementary Material, Table S5). The occurrence of MACE was significantly more common in the complex coronary group than the other groups (P < 0.001), subsequently resulting in a longer hospital length of stay (P = 0.03). There were no VT/VF and cardiac arrests in the simple coronary and normal coronary groups.

Three of the 51 patients (5.8%) died during admission for the index procedure. Our study was unable to detect significant differences in in-hospital mortality among the groups (P = 0.855). One death occurred in a patient with combined right CA type IIIB and left CA type IIIA lesions. He underwent concomitant left CA ostioplasty, but the right CA was left untouched during the initial repair. The patient experienced postoperative cardiac arrest and ECPR. He underwent subsequent urgent right CA ostioplasty due to diffuse small right CA shown by postoperative coronary angiogram but died of multiorgan failure and sepsis despite the recovery of cardiac function. Two patients, 1 in the simple coronary group and the other in the normal coronary group, had preoperative cardiac arrest requiring ECPR and were transitioned to ECMO after failure to wean from CPB after subsequent urgent SAS repair. The cause of death was multiorgan failure (n = 1) and brain death due to intracranial bleeding (n = 1).

Follow-up and long-term outcomes

There was no late death nor MACE during the median follow-up period of 8.6 (4.6–12.8) years. Patients with no or simple CA lesions had significant better 5-year reintervention-free survival than patients with complex CA lesions [no or simple CA lesions: 85.7% (95% CI: 59.2%, 69.0%) vs complex CA lesions: 55.5% (95% CI: 15.2%, 23.1%), P = 0.035] (Supplementary Material, Fig. S1). There were 11 surgical reinterventions, including 1 who received a left main CA bypass grafting at 10 months after the original operation for severe stenosis at the distal left main CA, distal to the ostioplasty patch. Other reasons for reinterventions were aortic arch stenoses (n = 2), recurrent arch stenoses (n = 5), aortic valve lesions (n = 2) and aortic aneurysm (n = 1).

Predictors and other factors associated with in-hospital major adverse cardiac events

The univariable analysis identified complex CA lesions, prolonged CPB and cross-clamp times and preoperative cardiac arrest as predictors for in-hospital MACE (Table 3). The multivariable analysis revealed complex CA lesions [odds ratio: 6.03 (1.32–27.38), P = 0.02] as independent predictors for MACE (Table 4). Variance inflation factor of predictors did not show significant multicollinearity in regression analysis. The association between postoperative ST-segment changes and MACE was examined by the Fisher’s test. Persistent ST-segment changes were significantly associated with MACE (P = 0.04, Supplementary Material, Table S6).

Table 3:

Univariable logistic regression on putative factors associated with postoperative major adverse cardiac events

Variables	Odds ratio [95% confidence interval]	P-value*
Age (months)	0.94 [0.87–1.01]	0.08
Weight (kg)	0.80 [0.64–1.10]	0.06
Williams syndrome	1.03 [0.24–4.40]	0.97
Coronary intrinsic lesions
Simple lesions	1.77 [0.10–30.71]	0.70
Complex lesions	26.83 [2.75–262.29]	0.01
Pump time (min)	1.02 [1.01–1.03]	0.01
Cross clamp time (min)	1.02 [1.00–1.04]	0.02
Absence preoperative arrest	0.14 [0.06–0.34]	<0.001
Concomitant arch reconstruction	1.78 [0.41–7.75]	0.44
Left ventricle to aorta peak gradient (mmHg)	1.00 [0.97–1.03]	0.87
Right coronary kinking	1.60 [0.34–7.59]	0.55

^*Comparison among patients with complex coronary artery lesions, with simple coronary artery lesions and with no coronary artery lesions. Bold indicates statistical significance with an alpha of 0.05.

Table 4:

Multivariable logistic regression on putative factors associated with postoperative major adverse cardiac events

Variable	Odds ratio [95% confidence interval]	P-value
Age	0.93 [0.81–1.07]	0.31
Weight	0.94 [0.55–1.63]	0.83
Complex versus simple versus no coronary lesions	6.03 [1.32–27.38]	0.02
Cross-clamp time	1.00 [0.98–1.02]	0.97

DISCUSSION

This study analysed the prevalence, type and degree of CA lesions, as well as their impacts on early and late clinical outcomes of SAS patients undergoing surgical repair. SAS repair is a high-risk procedure due to the risk of myocardial ischaemia from sudden change in coronary perfusion pressure after the stenosis is addressed. In approximately half of the SAS patients, coronary flow is additionally compromised by various types and degrees of CA lesions. Coronary malperfusion secondary to CA lesions can be unmasked by afterload reduction from SAS repair. Patients with complex CA lesions were at significantly higher risk of postoperative persistent ST-segment changes and MACE. Patients with underlying right CA lesions were at a much higher risk of right CA distortion/kinking when SAS was repaired with the inverted-Y patch, revealing the susceptibility of this subgroup to the subtle right CA geometrical changes of the right coronary sinus by patch reconstruction. This study highlights the importance of diagnosis and recognition of CA lesions to prevent MACE in this entity. Given the limitation of preoperative detection of CA lesions, the intraoperative real-time anatomic and functional assessment of CA lesions and immediate postoperative CA imaging are crucial for effective and timely surgical repair to restore coronary blood flow.

Associated coronary artery lesions

The incidence of CA lesions in patients with SAS varies, partly dependent on the definition and diagnostic modalities employed. Despite the high prevalence of angiographic evidence for coronary ostial stenosis, many surgical series reported a significantly low incidence of CA lesions, ranging from 2% to 13.8% [4, 5, 10, 14–18]. One possible reason is that prominent ostial ridges (type I lesion) are not commonly reported as a CA lesion. The ostial ridge was generally minor; however, 3 patients in our cohort had significant ridge formation that obstructed coronary flow requiring resection. Furthermore, ostial ridges may coexist with other CA lesions, such as type III, and potentially cause coronary obstruction. Given the possible clinical importance of ostial ridges on coronary blood flow, we considered these structural anomalies as CA lesions. However, the decision of ridge resection should be individualized.

Intraoperative assessment and treatment strategy for coronary artery lesions

The frequency of concomitant coronary procedures in the present study was much higher than the ones in the contemporary reported series [19]. This can be explained by the institutional bias of routine intraoperative CA imaging, and very low thresholds to repair or revise the CA, which have been based not only on clinical or anatomic criteria but also on physiologic criteria of flow acceleration [12]. Many centres, including our institution, prefer to have relatively high blood pressure and systemic vascular resistance to maintain myocardial perfusion after coming off CPB [3]. Therefore, some patients with mild-to-moderate residual coronary stenosis may remain undetected due to the high coronary perfusion pressure and do not show clinical signs of myocardial ischaemia until later. Routine intraoperative CA imaging helps to identify these unrecognized residual CA lesions, which can then be properly addressed intraoperatively. ST-segment changes in the early postoperative period are first interrogated with TEE. If echocardiographic findings are inconclusive, patients are urgently referred for diagnostic coronary catheterization. Once coronary compromise is identified, surgical revision is preferred over stenting.

Concomitant coronary artery procedures

The surgical approach and techniques for CA lesions should be individualized based on anatomic configuration. The ostial ridge resection (type I) and repair of the commissural fusion (type II) can be performed safely and effectively without significant additional time on CPB. Most type IIIA lesions in this series were repaired by the combination of ostioplasty and aortoplasty. Both patients with type IIIA lesions, who initially only underwent ostial ridge resection, required ostioplasty as a revision due to residual coronary stenosis detected by intraoperative EpE. These patients could have benefited from a more aggressive primary ostioplasty strategy. On the other hand, ostioplasty itself, particularly in infants with small vessel diameters, carries its own inherent risk [20]. In our experience, 1 patient (11.1% among the patients who underwent ostioplasty) developed severe stenosis at the end of the ostioplasty patch, requiring CA bypass grafting. However, it is quite encouraging that the majority of the ostioplasties were not only effective to prevent MACE but also durable with very low reintervention and MACE rates.

The surgical management of type IIIB lesions is a significant challenge. Preoperative diagnosis of type IIIB lesions is often extremely difficult. In our study, half of type IIIB lesions were diagnosed by intraoperative EpE after coming off CPB. This is one of the most critical pitfalls of this particular lesion, where haemodynamic significance of diffuse CA hypoplasia is often unmasked only after releasing SAS. Some lessons that could be learned based on the current result: (i) three out of 4 such lesions were successfully managed by patch plication, suggesting that an appropriate patch dimension was important for such patients. Multi-sinus separate patch reconstruction was therefore recommended allowing for more precise determination of patch dimension. (ii) The remaining 1 patient managed with ostioplasty died on ECMO. One should be cautious for ostioplasty for type IIIB lesion, as this procedure may expose the vessel at risk of stenosis at the end of the patch augmentation in the setting of the diffusely hypoplastic CA; (iii) in specific cases, the diffuse small CA may be accepted by the surgical team. The awareness of its existence was crucial for the postoperative team as type IIIB lesion requires close postoperative surveillance for coronary ischaemia.

Right coronary artery distortion/kinking

The inverted-Y patch effectively enlarges the supra-aortic region but also significantly changes the geometry of the right coronary sinus, posing the risk of right CA distortion/kinking [9]. Our study showed that right CA distortion/kinking was common after the bifurcated patch repair. One of the important messages from this study is that patients with the pre-existing CA lesions were more vulnerable to and have a lower tolerance for any additional change in coronary geometry. This result not only encourages more aggressive repair with upfront ostioplasty for complex lesions, but also warns against redundant longitudinal and lateral dimensions of the aortoplasty patch, particularly in those with CA lesions [9].

Major adverse cardiac events

Although low early mortality has been previously reported by our institution (5.8%) and other studies (2–9%) [8, 10, 18, 21], this present study observed a 17.6% incidence of MACE after SAS repair, highlighting the significant postoperative risk associated with these patients. A recent observational study on patients with Williams syndrome based on the Society of Thoracic Surgeons database reported a lower incidence of MACE compared to our study [21]. However, MACE was defined as postoperative death, cardiac arrest, or need for mechanical circulatory support, whereas our definition included ventricular arrhythmias as well. This may have increased our reported incidence of MACE, but also includes more subtle clinical events indicative of myocardial ischaemia.

No report has quantified the risk associated with CA disease in patients with SAS. This may be caused by poorly defined baseline CA lesions. A recent multicentre analysis showed that the presence of MACE was related to coronary interventions [21]. The impact of CA surgery/revision on the outcomes could not be completely ignored as CA surgery/revision often requires prolonged bypass and cross-clamp duration which may result in various degrees of myocardial injury and dysfunction. Unfortunately, we are not able to propose a definitive indication for surgical management for CA lesions in SAS repair due to the retrospective observational nature of this single-centre study. We believed that clinical decision-making should be individualized and always balance the associated benefits and risks. Further study involving multiple centres with different strategies on CA surgery is warranted to identify the optimal threshold for intervening on CA lesions.

Limitations

The major limitations of this study include its retrospective and observational nature, as well as the modest sample size and lack of long-term coronary blood flow data. Part of the small ostium diagnosis relied on the surgeon’s intraoperative inspection. There is also little consensus on clear definitions of CA lesions in patients with SAS, which makes comparisons between reports challenging.

CONCLUSIONS

The incidence of ST-segment changes and MACE is high in patients undergoing SAS repair. The preoperative presence of complex CA lesions is the most important predictor for postoperative risk of MACE. Therefore, preoperative and intraoperative recognition of the CA lesions and timely surgical intervention are essential in this population.

Glossary

ABBREVIATIONS

CA

Coronary artery

CI

Confidence interval

CPB

Cardiopulmonary bypass

ECPR

Extracorporeal cardiopulmonary resuscitation

ECMO

Extracorporeal membrane oxygenation

EpE

Epicardial echocardiogram

MACE

Major adverse cardiac events

SAS

Supravalvular aortic stenosis

STJ

Sinotubular junction

TEE

Transesophageal echocardiogram

VF

Ventricular fibrillation

VT

Ventricular tachycardia

SUPPLEMENTARY MATERIAL

Supplementary material is available at ICVTS online.

FUNDING

This work was not supported by any external sources of funding.

Conflict of interest: none declared.

DATA AVAILABILITY

All relevant data are within the manuscript and its supporting information files. If neither of these applies but you are able to provide details of access elsewhere, with or without limitations, please do so.

Author contributions

Shuhua Luo: Data curation; Writing—original draft; Writing—review & editing. Christoph Haller: Conceptualization; Data curation; Writing—review & editing. Lynne E. Nield: Data curation; Visualization. Mimi Xiaoming Deng: Data curation; Validation; Writing—review & editing. Jaymie Varenbut: Data curation. Osami Honjo: Conceptualization; Supervision; Writing—review & editing.

Reviewer information

Interdisciplinary CardioVascular and Thoracic Surgery thanks Victor T. Tsang, Prem Sundar Venugopal, Nishant Saran and the other anonymous reviewers for their contribution to the peer review process of this article.

 

Presented at the Congenital Heart Surgeon’s Society 47th Annual Meeting, Chicago, IL, USA, 27–28 October 2019.