Article Text
PDF
Abstract
Objective Variable coronary anatomy has been described in patients with bicuspid aortic valves (BAVs). This was never specified to BAV morphology, and prognostic relevance of coronary vessel dominance in this patient group is unclear. The purpose of this study was to evaluate valve morphology in relation to coronary artery anatomy and outcome in patients with isolated BAV and with associated aortic coarctation (CoA).
Methods Coronary anatomy was evaluated in 186 patients with BAV (141 men (79%), 51±14 years) by CT and invasive coronary angiography. Correlation of coronary anatomy was made with BAV morphology and coronary events.
Results Strictly bicuspid valves (without raphe) with left-right cusp fusion (type 1B) had more left dominant coronary systems compared with BAVs with left-right cusp fusion with a raphe (type 1A) (48% vs. 26%, p=0.047) and showed more separate ostia (28% vs. 9%, p=0.016). Type 1B BAVs had more coronary artery disease than patients with type 1A BAV (36% vs. 19%, p=0.047). More left dominance was seen in BAV patients with CoA than in patients without (65% vs. 24%, p<0.05).
Conclusions The incidence of a left dominant coronary artery system and separate ostia was significantly related to BAVs with left-right fusion without a raphe (type 1B). These patients more often had significant coronary artery disease. In patients with BAV and CoA, left dominancy is more common.
Statistics from Altmetric.com
Introduction
Bicuspid aortic valve (BAV) is the most common congenital heart defect, affecting 1–2% of the general population.1 ,2 While BAV can be an isolated disorder, it is often associated with other congenital cardiac abnormalities. The most common associated congenital malformation is coarctation of the aorta (CoA), with a prevalence ranging from 1% to 34%.3 ,4 Conversely, a BAV is seen in about 60% of patients with aortic coarctation.5 ,6 There are different BAV morphologies: most common are valves with fusion of the left and right coronary cusp. Valves with fusion of the right and non-coronary cusp are less common, and valves with fusion of the left and non-coronary cusp are rare. When no raphe is present, the valve is called strictly bicuspid.7
BAV morphology correlates to outcome with respect to valve function and amount of aortic dilation.8 ,9 Differences in coronary anatomy have also been described in patients with BAV. For example, an increased incidence of both a short left main coronary artery (LMCA) and of immediate bifurcation of the two major branches of the left coronary artery (LCA) was described in patients with BAV and in patients with aortic stenosis and left dominant arterial system.10 ,11 Furthermore, an association between BAV and a left dominant coronary arterial system was described.12 ,13 The incidence of left coronary dominance in patients with BAV in one study was found to be as high as 57%.12 Other studies reported lower numbers ranging from 25% to 29%,10 ,13 ,14 which is still much higher than the reported incidence of 10–12% in the general population with normal aortic valves.10 ,13 The reason why these coronary patterns occur more often in patients with BAV is to date unclear.
During development, the aortic valve, as well as the pulmonary valve, develops from the so-called endocardial cushions, which are local swellings of mesenchymal tissue located at the outflow tract of the heart. Initially, no separate aortic cusps can be detected; however, after proper development, the classic tricuspid division in three aortic valvular cusps is recognised.15 In patients with BAV, it is generally considered that two cusps are fused, although a maldevelopment or lack of proper cushion division may also be the case.8 ,16
Apparently, signalling events during development guide the proper ingrowth of the left main stem and right coronary artery (RCA) in a regionalised fashion, dictated by the borders of the valves, although the exact mechanism is still unknown. In patients with BAV, where this strict bordering of the cusps is hampered, variations in coronary anatomy might occur more often, perhaps most outspoken in patients with strictly bicuspid valves where a raphe is missing and no valvular commissures can be determined.17
In the current study, we hypothesised that variations in coronary distribution patterns correlate to BAV morphology and that strictly bicuspid valves show more variation as valve borders are undetermined by a raphe in these cases.
Therefore, the purpose of this study was to evaluate coronary anatomy and association with coronary events as related to BAV morphology.
Methods
Study population
All BAV patients with transthoracic echocardiography between 2005 and 2013 were identified from the echocardiography database of the Leiden University Medical Center and Academic Medical Center Amsterdam (together forming the Center for Congenital Heart Disease Amsterdam Leiden).
Of these patients, only those were selected who had also undergone CT coronary angiography (CTA) and/or invasive coronary angiography (ICA). Patients with uninterpretable CTA or ICA examination were excluded from the analysis. From a total of 420 patients, 186 patients were selected, of whom aortic valve morphology was evaluated by echocardiography and coronary anatomy was assessed by CTA or ICA. Patients with fusion of the left and non-coronary cups (types 3A and 3B, eight patients) were excluded from statistical analysis due to small numbers, leaving a total number of 178 patients included in the study.
Echocardiography
Echocardiography exams were obtained using commercially available echo systems (Vivid 7 and E9, GE Vingmed Ultrasound, Horten, Norway) with standard views from the parasternal, subcostal, suprasternal and apical windows. The aortic valve was examined in the two-dimensional parasternal short-axis view and classified as bicuspid when two cusps could be clearly identified and the typical ‘fish-mouth opening’ of the valve was observed. All echocardiography exams were re-evaluated offline using EchoPac V.7 software (GE Medical Systems, Horten, Norway). The aortic valve was evaluated in a cross-sectional view for the presence of a raphe. In a previous study, three BAV morphologies were identified: type 1 with fusion of right and left coronary cusp; type 2 with right and non-coronary fusion and the rare type 3 with left and non-coronary fusion.8 In this study, this classification was followed (figure 1), with addition of A (for valves with a raphe) or B (for strictly bicuspid valves).
Schematic overview of bicuspid aortic valve (BAV) morphologies. Drawings are oriented in echocardiographic view of the aortic valve. Upper panel: the three major valve morphologies observed in patients with BAV are described as three different types based on the valve leaflet orientation: type 1 (fusion right and left coronary cusps), type 2 (fusion right and non-coronary cusps) and type 3 (fusion left and non-coronary cusps). Middle panel: strictly bicuspid valves (without a raphe) are defined as a subgroup B. Modified after Schaefer et al.8 Leaflet size and symmetry, as well as the position of the commissures, may vary from the schematic overview depicted here. Lower panel: schematic drawing (left) and echocardiographic view (right) of a tricuspid aortic valve, in similar orientation as the panels above. LA, left atrium; LCC, left coronary cusp; NCC, non-coronary cusp; PT, pulmonary trunk; RA, right atrium; RCC, right coronary cusp; RVOT, right ventricular outflow tract.
CT coronary angiography data acquisition and image analysis
Patients were scanned using a 64-row CT scanner (Aquillion64, Toshiba Medical Systems, Otawara, Japan; General Electrics LightSpeed VCT, Milwaukee, Wisconsin, USA) or a 320-row CT scanner (AquillionONE, Toshiba Medical Systems, Otawara, Japan). Coronary anatomy and coronary vessel dominance were assessed in a standardised manner by dividing the coronary artery into 17 segments according to the guidelines of the American Heart Association (Austen et al, 1975).18
The coronary artery system was classified as right dominant when the posterior descending artery (PDA) originated from the RCA and considered left dominant if the PDA originated from the left circumflex artery (LCX) (figure 2A and B). The coronary artery system was classified as balanced when the PDA originated from the RCA in combination with a large posterolateral branch originating from the LCX, reaching to or close to the posterior interventricular groove. Subsequently, the presence of a LMCA was assessed. If there was no LMCA, the LCA was designated as having separate ostia of the left anterior descending artery (LAD) and LCX (figure 2C and D). Finally, the presence of coronary artery disease was assessed. All 17 segments were scored as non-significant coronary artery disease (CAD) (defined as <50% luminal narrowing) or significant CAD (defined as ≥50% luminal narrowing).
(A) Left dominant coronary artery system as seen on CT coronary angiography (CTA). There is a large left circumflex artery (LCX) and a relatively small right coronary artery (RCA). The valve morphology in this patient was type 1A. (B) Left dominant coronary artery system in the same patient (type 1a BAV) as seen on CTA, posterior view. The posterior descending artery (PDA) originates from the LCX. (C and D) Separate ostia of the left anterior descending artery (LAD) and LCX as seen on CTA. Valve morphology in this patient was type 1A. Ao, aorta; LA, left atrium; LCC, left coronary cusp; LV, left ventricle; NCC, non-coronary cusp; OM, obtuse marginal; RA, right atrium; RCC, right coronary cusp; RV, right ventricle.
Invasive coronary angiography acquisition and analysis
Conventional ICA was performed according to standard techniques.19 All images were retrospectively reviewed by an experienced observer. Coronary artery dominance, separate ostia, coronary aberrancy and severity of CAD were classified as above. Vessels were classified as having diameter stenosis of either <50% or ≥50% luminal narrowing, and lesions with stenosis ≥50% were deemed significant.
Assessment of coronary events
Significant coronary artery disease was defined as stenosis causing ≥50% luminal narrowing. Coronary events were defined as having severe coronary artery disease (>70% luminal narrowing), having undergone revascularisation by percutaneous coronary intervention or coronary artery bypass grafting or any admission for unstable angina pectoris.20
Statistical analysis
Continuous variables are presented as mean±SD or as median and IQR. Categorical variables are presented as number and percentages. One-way analysis of variance tests were used for comparing numerical data in more than two categories. Cross-tabulations were made for binary categorical data, on which χ2 goodness-of-fit-tests were performed to test for independence. For sets of independent numerical data, linear regression analysis was used to evaluate trends. Similarly, trends for binary categories were evaluated with binary logistic regression to correct for possible confounding (risk) factors such as age and gender. Statistical analysis was performed using SPSS software (V.20.0, SPSS, Chicago, Illinois, USA). A p value <0.05, by a two-sided test, was considered statistically significant.
Results
Patient characteristics
The population comprised 178 patients with a mean age of 51±14 years, 141 (79%) patients were male. Patient characteristics are presented in table 1. Overall baseline characteristics were similar between the groups.
Patient characteristics
No significant differences in indications for CT scan or coronary angiography were observed between the different groups (table 2).
Indications for CT scan and/or coronary angiogram
Coronary anatomy
Of all patients, 52 (30%) had a left dominant coronary artery system, 103 patients (60%) had an RCA system and 16 patients (9%) had a balanced system. Separate ostia were seen in 20 patients (11%).
BAV morphology
In 143/178 patients (80%), a raphe could be identified. In 116 patients (65%), the raphe was located between the left and right coronary cusp and thus had a type 1A BAV. In 27 patients (15%), the raphe was located between the right and non-coronary cusp, designated as type 2A BAV. In 35 patients, no raphe could be identified. These strictly bicuspid valves were designated as group B BAV (figure 1). In 25 patients (14%), valve orientation was in an anterior-posterior direction, indicated as type 1B BAV and in 10 patients (6%) valve orientation was in a left-right direction, indicated as type 2B BAV (figure 1). Figure 3 shows the distribution of BAV morphology in the study population.
Schematic overview of the distribution of valve morphologies and of left dominant coronary artery systems in different valve morphologies in the study population.
Coronary anatomy and association with coronary events as related to BAV morphology
Type 1 BAVs without a raphe show more left dominancy and separate ostia
As stated before, 30% of all patients had a left dominant coronary artery system.
When comparing the whole group of patients with type 1 BAV to patients with type 2 BAV, there was no significant difference in coronary vessel dominance or separate ostia (table 3).
Differences in coronary anatomy and coronary events between different subtypes
When taking into account the presence or absence of a raphe, however, patients with type 1B BAV (no raphe) had more left dominant coronary systems compared with patients with type 1A BAV (p=0.047, table 3). Type 1B BAVs also showed more separate ostia of the LAD and LCX (p=0.016).
Figure 3 shows the distribution of BAV morphologies in the patients with a left dominant coronary artery system.
Type 1 BAVs without a raphe show more coronary artery disease
Next, to determine the clinical relevance of these differences in coronary anatomy, the clinical outcome in type 1A BAVs versus type 1B BAVs was studied. This showed that patients with type 1B BAV had significantly more coronary artery disease compared with patients with type 1A BAV (p=0.047, table 3). Type 1B BAVs also showed more coronary events compared with type 1A BAVs; however, this was not significant. All comparisons were corrected for age and gender.
Coarctation of the aorta
Out of 178 patients, 22 (12%) had a coarctation of the aorta.
Patients with CoA showed no significant difference in valve morphology, the majority (55%) having a type 1A BAV. In the CoA group, more left dominance was seen compared with the group with isolated BAV (p<0.05), but there was no difference in significant coronary artery disease or events. In contrast to the isolated BAV group, most patients with CoA and a left dominant coronary artery system had a type 1A BAV (type 1A 8/15 (53%) vs type 1B 4/15 (27%)). The CoA group was significantly younger than the patients with isolated BAV (42±13 years vs 51±14 years, p<0.001).
Discussion
The main finding of this study is that type 1 BAVs without the presence of a raphe more often show left dominant coronary systems and separate ostia of the LAD and LCX compared with type 1 BAVs with a raphe. In addition, more significant coronary artery disease was observed in this group.
Developmental aspects of coronary anatomy
Although many variations in coronary anatomy exist, it is striking that the posterior aortic cusp is hardly ever seen to receive an aberrant coronary artery.21 The same applies to the lack of ingrowth in pulmonary orifice. It is likely that signalling events during development guide the proper ingrowth of the left main stem and RCA in a regionalised fashion, dictated by the borders of the valves; however, the exact mechanism is still unknown. It seems quite in line with the expectation that in patients with BAV, where this strict bordering of the cusps is hampered, variations in coronary anatomy may occur more often, perhaps most outspoken in patients with strictly bicuspid valves where a raphe is missing and no valvular commissures can be determined.17 Indeed, in the current study the main finding was that type 1B BAVs more often showed a left dominant coronary artery system and separate ostia of the LAD and LCX.
Coronary anatomy and association with coronary events
As described before, there is an association between BAV and the presence of a left dominant coronary artery system. The reported incidence in this study is in line with previous studies10 ,13 ,14 ,22 and much higher than the reported incidence (10–12%) in patients with normal aortic valves.10 ,13 The prognostic relevance of coronary vessel dominance is unclear, partly due to contradictory results. However, some studies report a higher incidence of CAD (obstructive and non-obstructive) in left dominant patients.23 ,24 Also, higher risk scores for significant stenosis in segments of the LAD and LCX in patients with a left dominant coronary artery system have been described.25 The mechanism of a potentially increased susceptibility to coronary events in patients with left dominant anatomy is to date unclear. A study screening postmortem angiograms showed that the prevalence of a left dominant system decreased with age, suggesting a higher death rate among patients with a left dominant coronary artery system.26 An explanation could be the larger myocardial area at risk in case of an acute myocardial infarction. Other possible mechanisms explaining a worse prognosis might be coronary artery length and lumen diameter. A study by Dodge et al,27 evaluating 83 invasive coronary angiograms, demonstrated a smaller RCA diameter and a larger LCX diameter in patients with a left dominant coronary artery system. Thus, a proximal stenosis of the LCA may result in more extensive ischaemia and worse consequences in a left dominant system than in a right dominant system.
From a molecular point of view, hypoplasia of the coronary artery system in mouse models with BAVs have been reported.28 Whether this hypoplasia correlated to the distribution patterns of the coronary arteries has not been explored to date.
Genetic aspects
It has been shown that deficiency in endothelial nitric oxide synthase results in the formation of BAVs.29 In addition, mutations in the NOTCH1 gene can lead to BAVs.30 Recently, an association of Notch with CoA has been found.31 Interestingly, Notch signalling also plays an important role in the modulation of inflammation32 and the Notch pathway has been linked to coronary artery disease.33 Whether Notch mutations relate to the observed variations in BAV morphology subtype (type A vs type B) or coronary anatomy, remains to be determined.
Coarctation of the aorta
Subanalysis of the group of patients with BAV, who also had coarctation of the aorta, shows that these patients more often have left coronary system dominance, but no difference in coronary artery disease or events. A possible explanation could be that the mean age in this group was significantly lower than in the group without coarctation. Alternatively, this could be related to the fact that CoA patients with a left dominant coronary artery system more often had a type 1A BAV.
Study limitations
Only patients who underwent CT or ICA were included in this study, which may have led to a selection bias. However, the indications for coronary angiography were not significantly different between the groups (table 2).
Transthoracic echocardiography was the only imaging modality used for classification of BAV phenotype. Although the sensitivity of echocardiography to accurately diagnose BAV is relatively low, the specificity is quite high.34 Therefore, the selected patients most likely did have BAV. Accurate diagnosis of raphe and patterns of cusp fusion can be hampered by poor imaging quality. Therefore, patients with uninterpretable echocardiography were not included in the study.
Conclusions and clinical implications
This study shows that patients with type 1B BAVs and patients with BAV and CoA more often have a left dominant coronary artery system. Patients with type 1B BAVs also seem more at risk of developing significant coronary artery disease. The current study also shows that type 1B BAVs more often have separate ostia, which could be of clinical importance in performing coronary angiography and, more importantly, in using coronary artery perfusion as a means of myocardial protection during surgical aortic valve replacement.
Key messages
What is already known on this subject?
Differences in coronary anatomy have been described before in patients with bicuspid aortic valve (BAV). For example, an increased incidence of both a short left main coronary artery and of immediate bifurcation of the two major branches of the left coronary artery (LCA) was described in patients with BAV and in patients with aortic stenosis and left dominant arterial system. Furthermore, an association between BAV and a left dominant coronary arterial system was described.
What might this study add?
In the current study, we found that variations in coronary anatomy correlate to BAV morphology and that type 1 bicuspid valves without a raphe show more variation.
How might this impact on clinical practice?
Variations in coronary anatomy, especially the more frequent occurrence of separate ostia of the LCA in patients with type 1 bicuspid valves without a raphe, could be of clinical importance in using coronary artery perfusion as a means of myocardial protection during surgical aortic valve replacement.
References
Footnotes
Contributors WMCK is the main author of this article and planned and executed most of the research, analysis and reporting. GT has provided the patient characteristics of the patient population from the Academic Medical Center, contributed to some of the analyses and reviewed the article. MCDR has thoroughly reviewed the article and made some valuable remarks. HWV, BJB, BJM and HMJS has thoroughly reviewed the article and made some helpful suggestions. AJHAS contributed to planning of the CT scans, reviewed the article and made helpful suggestions. ACG-dG contributed to the planning of the article and has reviewed it thoroughly with helpful suggestions. MAdG has contributed to the figures of this article and has reviewed the article with helpful suggestions. RW has contributed to performing the statistics of the article and has contributed to the statistics section of the article. MJS has contributed in planning the research and reviewing it several times. MRMJ has contributed in planning the research, drafting the first version of the article and reviewing it several times. WMCK and MRMJ are responsible for the overall content.
Competing interests None declared.
Patient consent The Institutional Review Board of the Leiden University Medical Center waived the need for written informed consent.
Ethics approval The current retrospective evaluation of clinically acquired data was approved by the Institutional Review Board of the Leiden University Medical Center.
Provenance and peer review Not commissioned; externally peer reviewed.