Outcomes in Adults With Bicuspid Aortic Valves FREE

Bicuspid aortic valve is the most common congenital cardiac anomaly in the adult population.^1- 3 Prior studies have reported significant mortality and morbidity in patients with bicuspid aortic valve related to the development of aortic valve dysfunction, endocarditis, and dissection.^1,4- 5 It is uncertain whether these prior findings, based on necropsy and surgical series from earlier eras, can be applied to a contemporary patient population.^1,6- 7 The purpose of this cohort study was to examine the cardiac outcomes and disease progression in a large contemporary group of adults with bicuspid aortic valve followed up over a prolonged period of observation.

This cohort study examined a referral population of consecutive adults with bicuspid aortic valve assessed at the University Health Network (Toronto General and Toronto Western Hospitals, Toronto, Ontario, Canada) ambulatory cardiac clinics from 1994 through 2001. Patients were identified using the hospital's echocardiography and congenital cardiac databases. The University Health Network is the main congenital cardiac center for the city of Toronto. The inclusion criteria were bicuspid aortic valve documented on transthoracic echocardiography and the absence of complex congenital cardiac defects. We excluded 260 patients who were referred for cardiac surgery, catheter-based treatment, or obstetric care, and those from outside of Ontario. The study population was comprised of 642 patients. There were 148 patients who were originally included in a preliminary ancillary echocardiographic study of aortic dilatation, which excluded clinical outcomes and had a short duration of follow-up.⁸ The study protocol was approved by the research ethics board of the University Health Network, which enabled access to health records and imaging data as well as telephone contact with patients and their physicians for determination of outcomes. Verbal informed consent was obtained in patients in whom clinical follow-up data were obtained by telephone.

Baseline data were obtained from the record of the patient's first visit to the ambulatory cardiac clinic. Baseline data included age, sex, prior cardiac procedures, coronary risk factors, cardiac medications, symptomatic status, and echocardiographic parameters. Follow-up data up to December 2007 were obtained by review of health records, which included summaries of ambulatory clinic visits, procedures, and hospital admissions. For those patients who were not scheduled or were unable to return for evaluation during the 2007 calendar year, follow-up information was obtained by a scripted telephone interview with the patients or their physicians.

Comprehensive transthoracic echocardiography was performed at the first ambulatory visit and again at the most recent follow-up visit. Bicuspid aortic valve was diagnosed when only 2 cusps were clearly identified in the short axis view. Aortic stenosis was classified as mild (valve area >1.5 cm²; peak gradient <36 mm Hg), moderate (valve area, 1.0-1.5 cm²; peak gradient, 36-64 mm Hg), or severe (valve area <1.0 cm²; peak gradient >64 mm Hg).⁹ Aortic regurgitation was classified as mild, moderate, or severe using an integrated approach incorporating the following color and continuous flow Doppler criteria: ratio of regurgitant jet height to left ventricular outflow tract height, ratio of regurgitant jet area to left ventricular outflow tract area, regurgitant jet deceleration rate (pressure half-time), and presence of pandiastolic retrograde flow in the descending aorta.¹⁰ This integrated approach accounts for the known difficulties in using a single method in the assessment of eccentric or multiple regurgitant jets. In those patients who had not undergone prior intervention on the aortic valve and in whom leaflet orientation could be visualized, the spatial orientation of bicuspid aortic valve leaflets was classified into either right-left (from right noncoronary cusps fusion) or anterior-posterior (usually from right-left coronary cusps fusion and rarely from left noncoronary cusps fusion) (Figure 1).^11- 12 End diastolic measurements of aortic sinus (at the sinus of Valsalva level) and proximal ascending aortic dimensions were performed.¹³ Dilated aortic sinus and ascending aorta were defined by dimensions of greater than 35 mm and greater than 34 mm, respectively.¹⁴ Left ventricular ejection fraction was calculated.¹⁵ Significant coarctation of the aorta was diagnosed based on standard criteria.⁵

A subset of patients underwent cardiac magnetic resonance (CMR) as part of the recently implemented (since 2001) follow-up template for patients with bicuspid aortic valve and/or coarctation. Aortic dimensions were assessed using standardized CMR pulse sequence protocols as reported previously.¹⁶ For patients in whom CMR was contraindicated, chest computed tomography (CT) was performed. For the purpose of this analysis, CMR and CT results were combined and the threshold for the diagnosis of aortic sinus and ascending aorta dilation was the same as in echocardiography (>35 mm for aortic sinus and >34 mm for ascending aorta).^17- 19 Follow-up CMR or CT studies were performed using the same modality as the baseline examination.¹⁹

Mortality was determined and causes of deaths were classified as cardiac or noncardiac based on review of health records and/or death certificates. Primary cardiac events were defined as surgery on the aortic valve or ascending aorta, percutaneous aortic valvotomy, aortic complications (dissection or aneurysm development), congestive heart failure requiring hospital admission, or cardiac death during the follow-up period. All cardiac events were verified by review of health records, operative reports, or hospital discharge summaries. Our recommended follow-up schedule for clinical reassessment of asymptomatic patients with bicuspid aortic valve and without valvular dysfunction is every 3 to 5 years. The standard of practice at our institution is to intervene on aortic stenosis or aortic regurgitation primarily on the basis of symptoms. Potential candidates for surgical or percutaneous interventions are reviewed at a weekly combined medical-surgical conference. Asymptomatic patients with severe aortic regurgitation or dilated aortic sinus or ascending aorta are referred for surgery when thresholds for interventions are reached (end diastolic left ventricular dimension >75 mm, left ventricular ejection fraction <55%, or aortic sinus/ascending aorta dimension >50 mm).^9,20 Aortic valve bacterial endocarditis and cardiac arrhythmia were considered to be secondary cardiac events.

Data analysis was performed using SPSS version 15.0 (SPSS Inc, Chicago, Illinois). Data were presented as mean and standard deviation (SD) or proportion (SD); median values with range and/or interquartile range (IQR) were provided when data was not normally distributed. Mortality rate during follow-up was determined using the Kaplan-Meier method. Means and proportions were compared using the t, χ², and Fisher exact tests wherever appropriate. Unless otherwise stated, the 2-sided level of significance was .05.

Sex-specific mortality rates for each of 20 age groups (<1, 1-4, 5-9, 1-14 . . . ≥90 years) were derived from vital statistics data collected by the Office of the Registrar General, Ontario, Canada, during the period from 1996 through 1997 (population of Ontario, 11.2 million).^21- 22 Age- and sex-specific life expectancies were then calculated as previously described.^21- 22 The life expectancy of a population of 642 adults with identical age and sex distribution as the study group was then calculated. The survival estimate of this population comparison group was compared with that of the study group using the log-rank test.

Potential predictors of primary cardiac outcome were evaluated by univariate Cox regression analyses; data were censored at the time of the first primary cardiac event. Candidate variables with a P value of less than .05 on univariate analysis were entered into a multivariate Cox regression model; highly correlated variables (r >0.70) were combined prior to entry into the multivariate model. A stepwise, backward-elimination algorithm was used and a conservative significance level of less than .01 was chosen for inclusion in the multivariate model. Independent predictors were assessed for time dependency and then cross-validated for the entire study group with bootstrapping (Stata version 8.2, StataCorp, College Station, Texas).

The frequency of progression to moderate or severe aortic stenosis or regurgitation was determined in patients in whom the latest follow-up echocardiogram was obtained prior to any primary cardiac event. The rate of change in peak aortic valve gradient and aortic root was calculated by dividing the difference between the baseline and latest examinations by the time between evaluations.

Baseline characteristics are shown in Table 1. All patients were ambulatory and asymptomatic; mean (SD) age was 35 (16) years (median, 31 years; range, 16-78 years). Two hundred patients (31%) had 1 or more coronary risk factors (hypertension, hyperlipidemia, diabetes mellitus, cigarette smoking, or family history of coronary artery disease in first-degree relatives). All patients with hypertension or hyperlipidemia were receiving pharmacological treatment. The type of cardiac medications included (not mutually exclusive) were β-adrenergic antagonist (19%), angiotensin-converting enzyme inhibitor (13%), aspirin (8%), statins (6%), calcium channel antagonist (4%), and/or diuretics (1%). Almost all patients with aortic coarctation had undergone successful prior percutaneous or surgical therapy (150/159 patients); the severity of coarctation was below the threshold for intervention in the remaining patients. Forty patients (6%) had undergone surgical or percutaneous treatment of intracardiac shunts, subaortic stenosis, mitral regurgitation, or pulmonic stenosis in childhood and have no hemodynamic residua. Overall, 406 patients (63%) in the study group did not have significant aortic stenosis (mild or less than mild) and regurgitation at baseline (Table 2); this group included 122 patients with aortic peak gradient of less than 25 mm Hg and no aortic regurgitation. Eleven patients (2%) had small intracardiac shunts or mitral valve prolapse at the time of their baseline echocardiogram.

Follow-up, which consisted of survival status and documentation of cardiac events, was obtained for all patients with a mean (SD) duration of 9 (5) years (median, 8 years; range, 2-26 years; IQR, 5-12 years). Total follow-up time was 5797 patient-years with 41% of the study group having at least 10 years of follow-up. There was a total of 28 deaths (mean [SD], 4% [1%]); of which, 17 were cardiac-related (mean [SD], 3% [1%]) and 11 were not related to a cardiac etiology (2 cases of thrombotic stroke, 7 cases of malignancy, 1 case of pneumonia, and 1 suicide). In 9 of 12 heart failure deaths, the underlying cause was either ischemic or nonischemic left ventricular dysfunction that developed during the follow-up period. In the 15 deaths that were not related to aortic dissection, 8 patients had moderate aortic stenosis at baseline and 2 patients had severe aortic stenosis at baseline.

The cardiac mortality rate was 0.3% per patient-year of follow-up. When compared with age- and sex-matched population estimates, the overall mortality was not significantly different between the bicuspid aortic valve group and the population estimates (P = .71). The 5-year mean (SD) survival was 97% (1%) in both the bicuspid aortic valve group and in the population estimates. The 10-year survival was similar in both the bicuspid aortic valve group (mean [SD], 96% [1%]) and in the population estimates (mean [SD], 97% [1%]) (Figure 2). The 157 patients who experienced a nonfatal primary cardiac event had a lower 10-year survival (mean [SD], 92% [2%]) than the 485 patients who did not (mean [SD], 99% [1%]).

Primary cardiac events, including intervention on aortic valve or ascending aorta, cardiac death, hospital admission for heart failure, or aortic complications, occurred in 161 patients (mean [SD], 25% [2%]). The frequency and nature of outcome events are listed in Table 3. Surgical or percutaneous intervention on the aortic valve or ascending aorta comprised the majority of primary cardiac events (142 patients; mean [SD], 22% [2%]). The indications for intervention were symptomatic aortic stenosis (84 patients), symptomatic aortic regurgitation or progressive left ventricular dysfunction (37 patients), aortic sinus or ascending aorta dilation (11 patients), and endocarditis (10 patients).

Eleven patients (mean [SD], 2% [1%]) had an aortic complication (dissection or aneurysm), in whom 5 were dissections (3 ascending and 2 descending). One patient with ascending aorta dissection died before surgery while another died postoperatively; both patients had baseline dilatation of the aortic sinus and/or ascending aorta. In the third patient, who had surgery for a rapidly dilating aorta, ascending aortic dissection was diagnosed at time of the surgery. One patient with descending aortic dissection underwent aortic stenting while the other patient was treated medically. The frequency of aortic dissection was 0.1% per patient-year of follow-up. Overall, 9 of 11 patients with aortic complications required operative or catheter intervention.

Of the 16 patients (mean [SD], 2% [1%]) who required hospital admission for congestive heart failure, it was precipitated by tachyarrhythmia in 3 patients. Of the remainder, 9 patients subsequently underwent aortic surgery and the other 4 were treated medically.

The candidate variables examined by univariate analysis are shown in Table 4. None of the variables with a level of significance of less than .05 on univariate analysis was highly correlated with each other. On multivariate analysis, the independent predictors of primary cardiac events were age older than 30 years (hazard ratio [HR], 3.01; 95% confidence interval [CI], 2.15-4.19; P<.001), moderate or severe aortic stenosis (HR, 5.67; 95% CI, 4.16-7.80; P<.001), and moderate or severe aortic regurgitation (HR, 2.68; 95% CI, 1.93-3.76; P<.001). Age was not normally distributed and was analyzed using its median value. The above predictors were not time-dependent and the assumptions of the proportional hazard analysis were met. Risk estimates from bootstrapping were similar to those of the original analysis. The 10-year rate for freedom from a primary cardiac event in patients with 0 predictors was a mean (SD) of 94% (2%); 1 predictor, mean (SD) of 82% (3%); and more than 1 predictor, mean (SD), 35% (5%) (Figure 3). The attributable risk was 12% at 10 years for patients with 1 of the above-mentioned risk factors compared with patients without any risk factors. Coronary risk factors, prior pregnancy, and bicuspid aortic valve leaflet orientation were not independently predictive of primary cardiac events.

Aortic valve endocarditis occurred in 13 patients (mean [SD], 2% [1%] or 0.3% per patient-year of follow-up), 3 of whom required urgent surgery, and 3 of whom developed endocarditis following aortic valve replacement. The remaining 7 patients were initially treated medically but 5 patients eventually required aortic valve replacement. Thirty patients (mean [SD], 5% [1%] or 0.6% per patient-year of follow-up) experienced cardiac arrhythmias requiring treatment (bradyarrhythmia in 6 patients, supraventricular tachycardia in 22 patients, and ventricular tachycardia in 2 patients). The arrhythmias occurred during the postoperative period in 7 patients.

In the subset of 122 patients with aortic peak gradient of less than 25 mm Hg and no aortic regurgitation, 1 or more primary cardiac events occurred in 4 patients (mean [SD], 3% [2%]) consisting of cardiac death (aortic dissection in 1 patient with baseline dilatation of the aortic sinus, congestive heart failure, and arrhythmia in 1 patient with dilated cardiomyopathy attributed to alcohol abuse) and intervention on the aortic valve or aorta (2 patients). Aortic valve endocarditis affected 1 additional patient. The 5- and 10-year survival rate was a mean (SD) of 98% (1%) and 96% (3%), respectively. The 5- and 10-year rate for freedom from a primary cardiac event was a mean (SD) of 98% (1%) and 92% (5%), respectively.

Patients with moderate or severe aortic stenosis or regurgitation at baseline were more likely to be male, have right-left leaflet orientation, and aortic sinus dilation than those without aortic valve dysfunction (Table 2). Follow-up echocardiograms were analyzed in 619 patients (96%); parameters from the remaining patients were not included because their follow-up studies were performed after aortic valve intervention. At a mean (SD) echocardiographic follow-up of 7 (4) years (median, 6 years; range, 1-23 years; IQR, 3-9 years), 50 patients (mean [SD], 8% [1%]) with mild or less than mild aortic stenosis or regurgitation at baseline had progressed to moderate or severe aortic stenosis or regurgitation. The group that developed significant valvular dysfunction included 10 patients who were in the group of 122 patients with normal or near normal aortic valve function at baseline. The remaining 352 patients (mean [SD], 57% [2%]) continue to have mild or less than mild aortic stenosis or regurgitation. The median increase in peak aortic valve gradient was 0.7 mm Hg per year (range, 0-35 mm Hg per year; IQR, 0-2 mm Hg per year) for the entire group.

Baseline aortic sinus dilation was present in 180 patients (28%). This group had a higher proportion of patients who were older than 30 years (75% vs 42% of nondilated group; P<.001), were male (87% vs 61% of nondilated group; P<.001), had hypertension (40% vs 21% of nondilated group; P<.001), or had moderate or severe aortic stenosis or regurgitation (46% vs 33% of nondilated group; P = .002). In 63 of the 180 patients (10% of total), the aortic sinus dimension was higher than 40 mm, a clinically important threshold.⁹ Baseline ascending aortic dimension was not analyzed due to the limited number of patients with available data. The absence of baseline ascending aorta dimension data in most patients precluded a systematic examination of whether bicuspid aortic valve orientation is related to aortic dilatation.

At follow-up, 222 patients (mean [SD], 38% [2%]) had aortic sinus dilation. In the 245 patients in whom the ascending aortic dimension could also be assessed, 143 patients had a dilated ascending aorta. Dilated aortic sinus and/or ascending aorta was present in 280 patients (mean [SD], 45% [2%]) at follow-up. Dilated ascending aorta with normal aortic sinus dimension was present in 58 patients at follow-up. The aortic sinus and/or ascending aortic dimension exceeded 40 mm in 124 patients (mean [SD], 20% [2%]) at follow-up. The median increase in aortic sinus dimension was 0.2 mm per year (range, 0-9.0 mm per year; IQR, 0-0.7 mm per year) for the entire group. Changes in aortic peak gradient and aortic sinus dimension are displayed in Figure 4.

A total of 139 patients (mean [SD], 22% [2%] of the group with follow-up echocardiography) also underwent either CMR (129 patients) or CT (10 patients) during the median follow-up period of 8 years (range, 1-22 years; IQR, 4-11 years) as part of the recently implemented follow-up template for assessment of coarctation repair (96 patients) and/or to screen for aortopathy in patients with bicuspid aortic valve (43 patients). This more contemporary group was generally younger (mean [SD] age, 27 [11] years) with a lower proportion of moderate or severe aortic stenosis or regurgitation (26%) at baseline compared with the 480 patients who did not undergo CMR or CT (mean [SD] age, 37 [16] years with significant valve dysfunction in 40%; P<.001 for both comparisons of mean age and proportion with significant valve dysfunction). The frequency of dilated aortic sinus and/or ascending aorta was similar to that determined by echocardiography (n = 59; mean [SD], 42% [4%]). Thirty-two patients with dilated ascending aorta had normal aortic sinus dimension. In 35 of the 59 patients with dilated ascending aorta or aortic sinus (mean [SD], 25% [4%]), 1 or more of these aortic dimensions exceeded 40 mm. In the 37 patients who underwent at least 2 CMR or CT studies separated by a median interval of 5 years, the median rate of change in ascending aortic dimension was 0.30 mm per year (range, 0-2 mm per year; IQR, 0-0.7 mm per year).

In this study examining the largest consecutive series of adults with bicuspid aortic valve to date, 25% of patients experienced a primary cardiac event during a mean follow-up of 9 years. Intervention on the aortic valve and aorta comprised the vast majority of the primary cardiac events. Reassuringly, the overall mortality rate of this ambulatory cohort was not significantly different from population estimates. Older age, moderate or severe aortic stenosis, and moderate or severe aortic regurgitation independently predicted the occurrence of a primary cardiac event. At follow-up, almost half of the study population had dilation of either the aortic root or ascending aorta.

Prior series examining outcomes in patients with bicuspid aortic valve reported high rates of mortality and serious complications relating to endocarditis, cardiac surgery, heart failure, and dissection.^{1- 2,4- 5,23} Although a high frequency of cardiac events was observed in our cohort, serious cardiac morbidity and mortality was infrequent. Outcome differences between present and prior studies can be attributed to differences in the era that patients were examined, the population that was examined, the frequency of cardiac events associated with high mortality (aortic dissection and endocarditis), and advances in perioperative management.^{1- 2,4- 5,23}

The Olmsted County study²⁴ of asymptomatic patients with normally functioning bicuspid aortic valve also reported similar mortality as the general population in Minnesota. Both the Olmsted County study and our study reported a significant and incremental rate of nonfatal cardiac events with increasing age, with intervention on the aortic valve being a frequent outcome. Intervention on the basis of early symptoms or deterioration in cardiac function may contribute to the low mortality observed in both studies. The generalizability of our study results is reinforced by the similarity in the findings (low mortality and high rate of nonfatal cardiac outcomes) reported in the Olmsted County study, despite differences in the era from which the study population was identified (1980-1999 for Olmsted vs 1996-2001 in our study), duration of follow-up, and origin of the patient population (community-based vs hospital-based).

Our results extended the findings of prior studies^25- 26 to younger adults with bicuspid aortic valve by our reporting of age-dependent progression of aortic valve disease. If we apply our study results together with those of 2 recently published studies examining community and surgical populations, one can cautiously project that many young adults with bicuspid aortic valve will eventually need intervention for aortic valve disease or aortic dilatation.^24- 25 Several studies examining mostly older asymptomatic patients with degenerative aortic valve disease have established the prognostic role of moderate or severe aortic stenosis even in the absence of symptoms.^27- 29 The predictive role of moderate or severe aortic regurgitation has not been previously identified in a large sample of adults with bicuspid aortic valve. In contrast, the frequency of primary cardiac events, including cardiac death, was the lowest in the subset of adults with normal or near normal aortic valve function at baseline.

The association between orientation of bicuspid aortic valve leaflets, aortic valve dysfunction, and subsequent intervention has been reported in children and adolesents.^11,30 This association may be mediated by the influence of bicuspid aortic valve leaflet spatial orientation on aortic elastic properties.¹² In the present study, bicuspid aortic valve leaflet orientation did not independently predict cardiac events. It is likely that the relationship between bicuspid aortic valve leaflet orientation and subsequent outcomes, identified in children and adolescents, may have been modified by the process of aging and other acquired factors. Indeed, valve degeneration at diagnosis, rather than bicuspid aortic valve leaflet orientation, predicted future cardiovascular events in the Olmsted County study.²⁴

Dilated aortic sinus and ascending aorta in patients with bicuspid aortic valve have been attributed to accelerated smooth muscle apoptosis in the aortic media.^31- 35 This propensity for aortic dilatation begins early in life because children with bicuspid aortic valve generally have larger aortic sinus dimensions than normal controls.³⁶ In the present study, almost half of the patients had dilation of the aortic root or ascending aorta at follow-up, similar to that reported in the Olmsted County study.²⁴ Although echocardiography has an established role in the assessment of aortic valvular function and aortic dimension, CMR or CT imaging provided complementary data by its ability to image the entire aorta. The importance of aortic assessment in adults with bicuspid aortic valve was highlighted by our findings that 29% of the patients who underwent aortic valve surgery also required replacement of their dilated ascending aorta. The association between moderate or severe aortic stenosis or regurgitation and larger aortic sinus dimension underscores the additional influences of aortic valve disease on aortic dilatation.^32,34 Aortic dissection in bicuspid aortic valve patients, although infrequent in our study and not observed in the Olmsted County study, was associated with a case-fatality rate of 40% in patients. An international study also reported that acute aortic association is associated with a high mortality rate in both young (<40 years; 22%) and older (≥40 years; 24%) patients, irrespective of the site of dissections.³⁷ The importance of the aorta in patients with bicuspid aortic valve is further underscored by the 1 dissection death in a patient with a normal or near normal aortic valve function at baseline.

Our study has several limitations. Risk estimates were derived from a retrospectively identified population and the baseline extent of ascending aortic dilation could not be defined in most patients. However, aortic root dilation tended to parallel ascending aorta dilation, thus the frequency of aortic root dilation provided an indirect estimate of ascending aortic involvement.³⁴ It is unlikely that echocardiography has significantly underestimated the frequency of aortopathy because the frequency of aortic root or ascending aortic dilation was similar in those who also underwent CMR or CT imaging. The selection bias associated with the University Health Network being a major regional center for the care of adults with congenital heart disease is reduced by the role of the Toronto Congenital Cardiac Centre for Adults as the primary adult continuity clinic for the cardiac program of the Toronto Hospital for Sick Children. Our study results are applicable only to adults with bicuspid aortic valve, who are referred for cardiac assessment, because cardiac assessment of the general population is not feasible.²⁴ This is likely not a major limitation because echocardiography is widely available and used in Ontario, as was the case in Olmsted County.²⁴ The provision of universal health insurance coverage for all residents of Ontario further reduces the barrier to referral. Intermediate outcome data was not available in the present study, precluding analysis of progress of the various stages of the disease. Finally, it is not possible to adjust for possible differences in risk profiles (such as smoking, socioeconomic status, or ethnicity) when comparing mortality between the study patients and the general population, although almost one-third of our study group had at least 1 coronary risk factor.

In this study population of young adults with bicuspid aortic valve, age, severity of aortic stenosis, and severity of aortic regurgitation were independently associated with primary cardiac events. During the mean follow-up duration of 9 years, survival rates were not lower than for the general population. Young adults with bicuspid aortic valve have a high likelihood of eventually requiring interventions on the aortic valve and/or aorta and will need serial surveillance of aortic valve and aortic dimensions.