Recruitment of study patients began in July 2014 through a web-based platform. Patients with BAV aged over 19 years who underwent comprehensive echocardiography at our institution and agreed to participate the registry were prospectively enrolled. Patients who had history of open heart surgery (OHS) for BAV before enrollment were excluded. A total of 232 patients were recruited and gave written informed consent during the period from July 2014 to February 2017. According to our study purpose, we excluded patients for whom ‘initial surgery’ was chosen as a treatment strategy. Based on the current American College of Cardiology/American Heart Association (ACC/AHA) guidelines,14) which recommend an interval of 6-12 months for reevaluation in patients with severe AS/AR (stage C), we excluded 60 BAV patients who underwent OHS for AV or aortic aneurysm within 6 months after enrollment, and also excluded 2 patients who presented with overt symptoms (dyspnea of NYHA class 3 and 4) at initial enrollment. Finally, 170 patients were included in an interim analysis of the progression of AV dysfunction and surgical events over a period of about 4 years. Patient visit and echocardiographic follow-up scheduling was done at the discretion of the attending cardiologist. Based on to the current ACC/AHA guideline,14) routine follow-up echocardiography was performed at intervals of 6-12 months in patients with severe AS/AR, 1-2 years in those with moderate AS/AR, and 3-5 years in those with mild AS/AR. The present study conformed to the ethical guidelines of the Declaration of Helsinki. The study protocol was approved by the Ethics Committee of Asan Medical Center. In patients with multiple repeated echocardiographic examinations, the baseline date was defined by the first echocardiography in which BAV was detected.

We classified reasons for the initial diagnosis of BAV in all study patients. The major categories included presence of clinical symptoms or signs, focused screening due to family history of BAV or other congenital heart disease, and incidental detection on echocardiography. In particular, the incidental detection category included a variety of echocardiographic screening types, which we specified as coincidental findings during the diagnosis of other cardiovascular diseases (e.g., angina or arrhythmia), preparation of non-cardiac surgery, and purely incidental findings during general health examinations. History of hypertension was identified in those who repeatedly presented with blood pressure ≥ 140/90 mmHg or who were diagnosed with hypertension by a referring physician prior to study enrollment and were undergoing life-style modifications or use of anti-hypertensive medication. History of diabetes was identified in those who presented with fasting serum glucose ≥ 126 mg/dL or were being treated with oral hypoglycemic agents or insulin. Dyslipidemia was defined as total cholesterol of ≥ 200 mg/dL or use of lipid lowering agents. Chronic kidney disease was defined in patients with glomerular filtration rate < 60 mL/min/1.73 m² for more than 3 months.15) Coronary disease was defined in patients with previous coronary angiography that showed significant stenosis or who had typical clinical symptoms and were being treated with anti-angina medications. The severity of AV dysfunction was graded as none (trivial), mild, moderate, or severe for each case of aortic stenosis (AS) and regurgitation (AR) according to American Society of Echocardiography recommendations.16)17) Aortic dilation was defined as aortic diameter over 40 mm at any level. The classification of BAV morphology was performed using transthoracic echocardiographic images and transesophageal echocardiographic images (n = 54, 32%). As described in our previous study,10) the conventional system classified BAV according to patterns of cusp fusion: type 1, fusion between right coronary cusp (RCC) and left coronary cusp (LCC); type 2, fusion between RCC and non-coronary cusp (NCC); type 3, fusion between LCC and NCC. Type 4 was indistinguishable from type 2 or 3, but type 1 was rejected due to the separation between RCC and LCC (Supplementary Figure 1). The resulting dichotomous system was a simplified classification according to the direction of the fused cusps: BAV-CCF (coronary cusp fusion) and BAV-MCF (mixed cusp fusion). The clinical outcome analyzed in this study was defined as total surgical events, which was the composite of isolated AV surgery, combined AV with aortic surgery and isolated aortic surgery. We classified the main causes of AV surgery to identify the dominant AV dysfunctions. Cases with equal grades of AS and AR were defined as AS dominant. However, for cases diagnosed with definite infective endocarditis, it was considered the dominant surgical indication. The main cause of surgery on the aorta was determined according to the findings of echocardiography, computed tomography, and the surgeon's surgical record.

We presented categorical variables as numbers with percentages and continuous variables as mean ± standard deviation. Comparisons between paired parameters were performed with the Wilcoxon test. Event rates during follow-up were estimated by the Kaplan-Meier method, and compared by the log-rank test or Breslow test, as appropriate. The Cox proportional hazards model was used to evaluate associations of patient characteristics with surgical events. All reported p-values were two-tailed and p-values < 0.05 were considered statistically significant. SPSS version 22 (IBM Corp, Armonk, NY, USA) was used for statistical analyses.

The mean age of the subjects was 50 ± 13 years and 117 patients (69%) were male (Table 1). The prevalence of hypertension, diabetes, and chronic kidney disease was 29%, 9%, and 1%, respectively. Thirteen patients (8%) had histories of coronary artery disease: among them, 6 had undergone percutaneous coronary intervention and 7 were receiving medical treatment only. Most (n = 154, 91%) had no or minimal dyspnea, and 16 patients (9%) had mild dyspnea of New York Heart Association (NYHA) functional class 2 at initial enrollment. Sixteen patients (9%) initially presented with NYHA class 2 but the attending cardiologists did not consider their symptoms to represent overt valve-related symptoms, and the patients were therefore managed with watchful observation. Among the reasons for the initial diagnosis of BAV, incidental detection was the most frequent (124 patients, 73%). Most cases were diagnosed during general health examinations (n = 91, 53%) or screening before non-cardiac surgery (n = 20, 12%), and incidental diagnoses occurred during evaluation for other cardiovascular diseases in 13 patients (8%). Twenty-five patients (15%) underwent cardiac evaluations due to subjective symptoms, and cardiac murmur was the reason for echocardiography in 18 patients (11%). Another minor reason for focused screening was family history (n = 3).

Regarding BAV morphologies, type 1 as defined by the conventional classification was the most common (n = 91, 54%) followed by type 4 (n = 37, 21%), type 2 (n = 34, 20%), and type 3 (n = 8, 5%) (Table 2). Meanwhile, using the dichotomous classification, we confirmed BAV-CCF in 91 patients (54%) and BAV-MCF in 79 patients (46%). Significant (moderate to severe) AS or AR was present in 77 patients (45%), whereas the remaining 93 showed no valvular dysfunction or mild AS or AR. Aortic dilation was present in 98 patients (58%). Two patients presented with combined congenital cardiac anomalies (atrial septal defect [n = 1] and patent ductus arteriosus [n = 1]).

During clinical follow-up of a median 4.0 years (interquartile range, 2.7 – 7.5), there were 2 deaths (pancreatic cancer in one patient with mild AS and subarachnoidal hemorrhage in the other with severe AS). A total of 27 patients (16%) underwent surgery for aggravated AV or aortic pathology: combined AV with aortic surgery in 12 patients (44%), isolated AV surgery in 11 patients (41%), and isolated aortic surgery in 4 (15%). Regarding the causes requiring AV surgery, AS was the most common underlying valvular dysfunction (n = 20, 87%) while the others included AR (n = 2, 9%) and infective endocarditis (n = 1, 4%). Aortic aneurysm was the cause of aortic surgery in all cases (n = 16, 100%) without aortic dissection or coarctation (Table 3).

The composite surgical event-free rate at 5 years was 89% ± 3%, and the individual 5-year event-free rate for isolated AV surgery, AV with aorta surgery, and isolated aorta surgery were 95% ± 2%, 96% ± 2%, and 98% ± 1%, respectively (Figure 1). Comparison of patient groups according to surgical events showed the surgery group (n = 27) was characterized by higher prevalence of dyspnea NYHA class ≥ 2, moderate or greater AS, and larger aortic root diameter at the tubular portion compared to the non-surgery group (Table 4). Regarding BAV morphology, the surgery group tended to present more frequently with BAV-MCF according to the dichotomous classification, compared to the non-surgery group (63% vs. 43%, p = 0.091) although the difference was not statistically significant. Regarding the impact of BAV morphology, the conventional classification did not clearly discriminate event rates between types (Figure 2A). However, using the dichotomous classification, patients with BAV-MCF tended to have a higher 5-year surgical event rate compared to those with BAV-CCF (18 ± 5% vs. 5 ± 3%, p = 0.036, Figure 2B). In univariate analysis, NYHA class ≥ 2, moderate to severe AS, BAV-MCF, and aortic root dilation were associated with surgical events (Supplementary Table 1). In multivariate analysis, moderate to severe AS (hazard ratio [HR] 4.61; 95% confidence interval [CI] 1.83-11.62; p = 0.001) and NYHA class ≥ 2 (HR 2.53; 95% CI 1.01-6.35; p = 0.048) were associated with the development of surgical events, whereas BAV-MCF (HR 1.75, 95% CI 0.77-3.96; p = 0.183) and aortic dilation did not remain significant (HR 2.13; 95% CI 0.87-5.21; p = 0.099).

Follow-up echocardiography was available for 129 patients (76%) with a median interval of 3.9 years (interquartile range, 2.7 – 7.6 years). Overall, the severity of AS significantly increased during that period (p < 0.001, Figure 3A). Of 19 patients with moderate AS at baseline, 14 patients (74%) progressed to severe AS at follow-up echocardiography. Of the 61 patients initially diagnosed with mild AS, 6 patients (10%) progressed to severe AS and 15 patients (25%) to moderate AS. On the other hand, there was no significant change in the severity of AR during the same period (p = 0.361, Figure 3B). Among 19 patients with moderate AR at baseline, 7 patients (37%) had progressed to severe AR at the end of the study. Of the 16 patients initially diagnosed with mild AR, 3 patients (18%) progressed to moderate AR. However, there was no shift from mild AR to severe AR during an interval of about 4 years.

As this cohort was recruited in a tertiary referral hospital, the potential referral bias should be considered during interpretation of our data. Follow-up echocardiographic studies were performed in 76% of our cohort (129/170) and not in all patients, and the follow-up duration varied, resulting in a failure to perform quantitative assessments of progression of valvular dysfunction. The variability in follow-up duration parameters of valve dysfunction reflect patient evaluations based on the standards of routine clinical practice. Thus our data have limitations for use in quantitative assessments of the progression of valve dysfunction. Additionally, the median follow-up duration was only 4 years, which is not sufficient to represent the long-term natural history of early BAV disease.

In this initial prospective cohort study of BAV patients in Korea, we found that most patients (> 70%) were incidentally diagnosed with BAV, which reflects the wide availability of echocardiography in our medical environment. Although the patterns of BAV phenotype that we observed were different from those reported in Western countries, the clinical outcomes we observed were similar to those reported from a community-acquired cohort in the U.S. The potential impact of BAV phenotype on outcomes and patterns of progression of aortic valvular dysfunction (AS vs. AR) should be evaluated in further studies with longer follow-up durations.