Abstract
Background
Data regarding the risk of aortic dissection in patients with bicuspid aortic valve and large ascending aortic diameter are limited, and appropriate timing of prophylactic ascending aortic replacement lacks consensus. Thus our objectives were to determine the risk of aortic dissection based on initial cross-sectional imaging data and clinical variables and to isolate predictors of aortic intervention in those initially prescribed serial surveillance imaging.
Methods
From January 1995 to January 2014, 1,181 patients with bicuspid aortic valve underwent cross-sectional computed tomography (CT) or magnetic resonance imaging (MRI) to ascertain sinus or tubular ascending aortic diameter greater than or equal to 4.7 cm. Random Forest classification was used to identify risk factors for aortic dissection, and among patients undergoing surveillance, time-related analysis was used to identify risk factors for aortic intervention.
Results
Prevalence of type A dissection that was detected by imaging or was found at operation or on follow-up was 5.3% (n = 63). Probability of type A dissection increased gradually at a sinus diameter of 5.0 cm—from 4.1% to 13% at 7.2 cm—and then increased steeply at an ascending aortic diameter of 5.3 cm—from 3.8% to 35% at 8.4 cm—corresponding to a cross-sectional area to height ratio of 10 cm2/m for sinuses of Valsalva and 13 cm2/m for the tubular ascending aorta. Cross-sectional area to height ratio was the best predictor of type A dissection (area under the curve [AUC] = 0.73).
Conclusions
Early prophylactic ascending aortic replacement in patients with bicuspid aortic valve should be considered at high-volume aortic centers to reduce the high risk of preventable type A dissection in those with aortas larger than approximately 5.0 cm or with a cross-sectional area to height ratio greater than approximately 10 cm2/m.
Aortic complications in patients with bicuspid aortic valve are more frequent than in the general population, accounting for up to 15% of aortic dissections [1, 2]. Aortic size is a strong predictor of rupture, aortic dissection, and mortality [3]. Several echocardiography-based studies involving patients with bicuspid aortic valve have not accounted for aortic size at initial presentation, thus including patients with relatively small ascending aortic diameters, and their conclusions are often extrapolated to patients with larger aneurysms [1, 4]. Hence, our objectives were to identify a cohort of patients with bicuspid aortic valve aortopathy and determine the risk of aortic dissection based on initial cross-sectional imaging data and clinical variables and to isolate predictors of aortic intervention in those initially prescribed serial surveillance imaging.
Patients and Methods
Patients
From January 1995 to January 2014, 5,606 patients who underwent echocardiography at Cleveland Clinic were found to have a bicuspid aortic valve and had cross-sectional computed tomography (CT) or magnetic resonance imaging (MRI) studies available for analysis. Of these patients, 1,181 had an aortic diameter greater than or equal to 4.7 cm at the sinuses of Valsalva or tubular ascending aorta, and among these patients, 27 had undergone remote aortic valve repair or replacement (mean, 12 ± 4.8 years before the first CT/MRI). Mean age at the time of the first CT/MRI was 55 ± 12 years, and 82% were men (Table 1).
Table 1.
Patient Characteristics (Total N = 1,181)
| Characteristic | na | No. (%) or Mean ± SD |
|---|---|---|
| Demography | ||
| Male | 1,181 | 965 (82) |
| Age (y) | 1,181 | 55 ± 12 |
| Height (cm) | 1,181 | 177 ± 9.5 |
| Weight (kg) | 1,181 | 91 ± 20 |
| Body surface area (m2) | 1,181 | 2.1 ± 0.3 |
| Clinical comorbidities | ||
| Hypertension | 1,181 | 688 (58) |
| Smoking | 1,178 | 477 (40) |
| Pharmacologically treated diabetes | 1,178 | 97 (8.2) |
| COPD | 1,181 | 96 (8.1) |
| Previous myocardial infarction | 1,181 | 86 (7.3) |
| Stroke | 1,181 | 30 (2.5) |
| Valve pathologic features | ||
| Aortic regurgitation | 1,006 | 671 (67) |
| 3–4+ regurgitation | 1,103 | 360 (33) |
| Aortic stenosis | 1,006 | 490 (49) |
| Severe stenosis | 977 | 207 (21) |
| Area (cm2) | 526 | 1.2 ± 0.7 |
| Mean gradient (mm Hg) | 798 | 26 ± 20 |
| Peak gradient (mm Hg) | 819 | 45 ± 34 |
| Bicuspid valve classification | ||
| Type 0 | ||
| Anterior-posterior | 1,044 | 7 (0.67) |
| Lateral | 1,044 | 4 (0.38) |
| Type I | ||
| Left-right (LR) | 1,044 | 817 (78) |
| Right-non (RN) | 1,044 | 171 (16) |
| Left-non (LN) | 1,044 | 9 (0.86) |
| Type II | ||
| LR-RN | 1,044 | 30 (2.9) |
| LR-LN | 1,044 | 0 (0) |
| LN-RN | 1,044 | 2 (0.19) |
| Unspecified | 1,044 | 4 (0.38) |
| Aortic measurements | ||
| Sinuses of Valsalva (cm) | 1,175 | 4.4 ± 0.8 |
| Tubular ascending aorta (cm) | 1,181 | 5.0 ± 0.6 |
| Sinuses of Valsalva (cm2/m) | 1,175 | 8.8 ± 6.2 |
| Tubular ascending aorta (cm2/m) | 1,181 | 11 ± 3.2 |
Patients with available data.
COPD = chronic obstructive pulmonary disease; SD = standard deviation.
Of the 1,181 patients, 800 had 1 preoperative CT/MRI and underwent ascending aortic or root replacement (within a median of 6 days). None of the patients with type A dissection had a history of Marfan syndrome, Loeys-Dietz syndrome, or Ehler-Danlos syndrome. Among the 800 patients who underwent an immediate aortic operation, 540 (68%) had concomitant aortic valve replacement (AVR), 213 (27%) had aortic valve repair, and 44 (6%) required no aortic valvular intervention. One hundred twenty-two (15%) patients had at least 1 coronary system with greater than 50% stenosis, and 103 (13%) patients underwent concomitant coronary artery bypass grafting.
The remaining 380 patients did not undergo immediate aortic operations and were observed with either serial echocardiography or CT/MRI (“surveillance” patients).
Data
Three-dimensional CT or MRI reconstruction was performed for each patient. Cross-sectional diameters were measured at the sinuses of Valsalva (noncoronary cusp to right-left coronary commissure or largest cusp-to-commissure diameter) and tubular ascending aorta at the pulmonary artery bifurcation. All maximal diameters were measured from the external wall of the aorta in a plane orthonormal to the center line of flow.
Bicuspid aortic valve morphologic characteristics were determined by direct intraoperative inspection or pre-operative echocardiography. Patient characteristics at the time of the first CT/MRI that were greater than or equal to 4.7 cm were obtained retrospectively from medical records, and echocardiographic variables were retrieved from a prospectively collected database. For those undergoing aortic operations during the study period, operative outcomes were collected prospectively in a surgical database. Use of these registry data as well as CT/MRI data for research was approved as exempt by the institutional review board, with patient consent waived.
End Points
End points for analysis were (1) type A dissection diagnosed on initial CT/MRI or at operation subsequent to the first imaging study, (2) type A dissection diagnosed during surveillance, and (3) ascending aorta or root replacement.
Passive follow-up of surveillance patients was obtained by review of medical records. If follow-up was incomplete before the common closing date of April 1, 2014, patients were contacted by mailed questionnaire or telephone using an institutional review board–approved form or script requiring patient consent. Median follow-up was 3.0 years (range, 0–17 years), with a total of 1,478 patient-years of data available for analyses. Twenty-five percent of patients who did not undergo an aortic operation were followed for 6.9 years and 10% for more than 10 years.
Data Analysis
Random Forest classification analysis was performed using R statistical software [5] (R, version 3.1.2; The R Project for Statistical Computing, Vienna, Austria), and remaining analyses were performed using SAS statistical software (SAS, version 9.2; SAS Institute, Cary, NC).
Dissection at Initial Presentation
A random Forest [6–8] classification model for the probability of type A dissection in all patients was created for each of the following aortic size indexing methods: raw diameter, diameter indexed by height, diameter indexed by body surface area, diameter standardized (z value) for body surface area using normal diameters from a population of healthy individuals [9], and cross-sectional area to height ratio [10]. Variable importance measures were used to determine the impact of variables on model performance [6, 11]. For each model, the risk-adjusted probability of dissection was calculated, taking into account all available patient variables (Fig E1, Appendix E1). The area under the curve (AUC) statistic for each model was obtained to identify an optimal predictor of type A dissection [12, 13].
Freedom from Ascending Aortic Intervention
Nonparametric estimates of long-term freedom from aortic operation were assessed by the Kaplan-Meier method, and parametric estimates were assessed by a multiphase hazard model [14] in the surveillance group. The parametric model was used to resolve a number of phases of instantaneous risk of aortic operation (hazard function) and to estimate shaping measurements.
To assess the effects of baseline patient variables (Appendix E1) on the risk of intervention, multivariable analysis was performed in the multiphase hazard function domain. Bootstrap bagging [15] was used for variable selection, with a p value criterion for retention of variables in the bootstrap model of 0.05 and the median rule for including variables in the final model [16].
Presentation
Continuous variables are summarized as mean ± standard deviation; pairwise comparisons were made using the Wilcoxon rank sum test. Categorical data are summarized by frequencies and percentages; comparisons were made using a χ2 test or the Fisher exact test when frequency was less than 5. Parametric estimates are accompanied by an asymmetric 68% confidence interval, equivalent to ± 1 standard error.
Results
On initial CT/MRI, mean diameter was 4.4 ± 0.8 cm at the sinuses of Valsalva and 5.0 ± 0.6 cm at the ascending aorta (Table 1). On initial CT/MRI or intraoperative inspection at operation subsequent to the first imaging study, 4.5% (n = 53) of patients were found to have type A dissection. Among the 800 patients who underwent immediate aortic operations, hospital mortality was 0.25% (n = 2) and risk of stroke was 0.75% (n = 6). Of the 380 surveillance patients, 175 required eventual ascending aortic intervention, including 10 in whom type A dissection developed during follow-up and 165 required aortic expansions. Baseline characteristics of the surveillance patients are shown in Table 2.
Table 2.
Baseline Characteristics of Patients Undergoing Surveillance (Total N = 380)
| Characteristic | na | No. (%) or Mean ± SD |
|---|---|---|
| Demography | ||
| Male sex | 380 | 310 (82) |
| Age (y) | 380 | 55 ± 13 |
| Height (cm) | 380 | 177 ± 9.4 |
| Weight (kg) | 380 | 92 ± 20 |
| Body surface area (m2) | 380 | 2.1 ± 0.27 |
| Blood pressure | ||
| Systolic (mm Hg) | 380 | 124 ± 16 |
| Diastolic (mm Hg) | 380 | 76 ± 11 |
| Clinical comorbidities | ||
| Hypertension | 380 | 194 (51) |
| Smoking | 380 | 139 (37) |
| Pharmacologically treated diabetes | 380 | 28 (7.4) |
| COPD | 380 | 22 (5.8) |
| Previous myocardial infarction | 380 | 35 (9.2) |
| Coarctation | 380 | 9 (2.4) |
| Stroke | 380 | 9 (2.4) |
| Dialysis | 380 | 0 (0) |
| Medications | ||
| β-blocker | 380 | 213 (56) |
| ACE inhibitor | 380 | 97 (26) |
| ARB | 380 | 42 (11) |
| Losartan | 380 | 17 (4.5) |
| Calcium channel blocker | 380 | 44 (12) |
| Steroid or immunosuppressant | 380 | 10 (2.6) |
| Aspirin | 380 | 142 (37) |
| Valve pathologic features | ||
| Aortic regurgitation | 310 | 186 (60) |
| 1+ | 47 (15) | |
| 2+ | 69 (22) | |
| 3+ | 46 (15) | |
| 4+ | 24 (7.7) | |
| Bicuspid valve classification | ||
| Type 0 | ||
| Anterior-posterior | 344 | 1 (0.29) |
| Lateral | 344 | 1 (0.29) |
| Type I | ||
| Left-right (LR) | 344 | 265 (77) |
| Right-non (RN) | 344 | 65 (19) |
| Left-non (LN) | 344 | 2 (0.58) |
| Type II | ||
| LR-RN | 344 | 10 (2.9) |
| LR-LN | 344 | 0 (0) |
| LN-RN | 344 | 0 (0) |
| Aortic measurements | ||
| Sinuses of Valsalva (cm) | 376 | 4.3 ± 0.49 |
| Tubular ascending aorta (cm) | 380 | 4.9 ± 0.36 |
| Sinuses of Valsalva (cm2/m) | 376 | 8.3 ± 1.8 |
| Tubular ascending aorta (cm2/m) | 380 | 11 ± 1.6 |
Patients with available data.
ACE = angiotensin-converting enzyme; ARB = angiotensin-receptor blocker; COPD = chronic obstructive pulmonary disease.
Aortic Dimensions and Probability of Dissection
Based on variable importance measures of all baseline variables included in the random Forest model (Appendix Fig E1), ascending aortic diameter was predictive of aortic dissection, followed by aortic diameter at the sinuses of Valsalva. Cross-sectional area to height ratio was associated with the best prediction and highest C-statistic (AUC = 0.73) and raw diameter was associated with the lowest (AUC = 0.69) (Figs 1C, 1F). However, given the better interpretability, we report probability of dissection related to raw measurement values.
Fig 1.
Each point represents the risk-adjusted probability of immediate dissection. Curves are the smoothed relationships. Loess smooth curve indicates the trend of increased risk with diameter. (A) Sinus diameter (cm). (B) Ascending diameter (cm). (C) Classification receiver operating characteristic (ROC) curve for ascending aortic dissection by diameter (cm). (D) Sinus cross-sectional area to height ratio (cm2/m). (E) Ascending cross-sectional area to height ratio (cm2/m). (F) Classification ROC curve for ascending aortic dissection by cross-sectional area-to-height ratio. (AUC = area under the curve; FPR = false prediction rate; TPR = true prediction rate.)
The risk-adjusted probability of type A dissection increased gradually at a sinus diameter of 5.0 cm—from 4.1% (95% confidence interval [CI], 3.9–4.4) to 13% (95% CI, 12–14) at 7.2 cm (Fig 1A) for the ascending aorta and the mode of dissection increased steeply at an ascending aortic diameter of 5.3 cm—from 3.8% (95% CI, 3.6–4.0) to 35% (95% CI, 34–37) at 8.4 cm (Fig 1B). These 2 values corresponded to a cross-sectional area to height ratio of approximately 10 cm2/m for the sinuses of Valsalva and 13 cm2/m for the tubular ascending aorta (Figs 1D, 1E). Other patient risk factors (Appendix Fig E1) were not found to be importantly associated with aortic dissection.
Dissection During Surveillance
Of the 10 surveillance patients in whom type A dissection developed, 3 dissections were diagnosed intraoperatively. These were not seen on preoperative imaging; they had ascending diameters of 5.0, 5.2, and 7.5 cm, respectively. In the remaining 7 patients, mean predissection sinus and ascending aortic diameters were 4.6 ± 0.5 and 5.0 ± 0.8 cm, respectively.
Risk of Future Ascending Aortic Intervention
Among surveillance patients, the probability of aortic intervention for expansion or dissection was 16% within 1 year of the initial CT/MRI and 50% at 6 years (Fig 2A). Instantaneous risk of aortic intervention peaked within 6 months and was 10% per year (95% CI, 9.2–11) at 3 years and 9% per year (95% CI, 8.3–10) at 10 years (Fig 3).
Fig 2.
Risk of aortic intervention among 380 surveillance patients. (A) Years from index computed tomographic scan. Each symbol represents an aortic operation positioned on the vertical axis by the Kaplan-Meier estimator, and vertical bars are confidence limits equivalent to ± 1 standard error (SE). Solid line is parametric risk of aortic intervention estimate enclosed within dashed 68% confidence band equivalent to ± 1 SE. (B) Predicted risk at 1 (blue), 5 (green), and 10 (red) years by age. Solid lines are parametric risk of aortic intervention estimates enclosed within dashed 68% confidence band equivalent to ± 1 SE. This nomogram is based on multivariable equation in Table 3. Nomogram of risk is for patient with 5.0-cm ascending aorta diameter (z value = 12), date of presentation of July 2008, and body surface area of 2.1 m2. (C) Predicted risk at 1 (blue), 5 (green), and 10 (red) years by ascending aorta diameter. Solid lines are parametric estimates of risk of aortic intervention estimates enclosed within dashed 68% confidence bands equivalent to ± 1 SE. This nomogram is based on multivariable equation in Table 3. Risk of aortic intervention is for 55-year-old patient with the same characteristics noted in B.
Fig 3.
Instantaneous risk of aortic intervention (hazard function). Solid line is parametric hazard estimate enclosed within dashed 68% confidence band equivalent to ± 1 standard error. (CT = computed tomography.)
Patients who had more recent initial CT/MRI investigations and larger initial ascending diameters were at higher risk of early aortic intervention (Table 3). Patients who were younger at initial CT/MRI and had larger initial ascending diameters were at greatest risk of late intervention (Figs 2B, 2C, and Appendix Figures E2A, E2B).
Table 3.
Incremental Risk Factors for Aortic Operations Among Surveillance Patients
| Factor | Coefficient ± SE | p Value | Reliability (%)a |
|---|---|---|---|
| Early phase | |||
| More recent CT (interval: 1/1/1995 to index CT) | 0.26 ± 0.091 | 0.004 | 74 |
| Baseline maximum ascending diameterb | 10 ± 2.8 | 0.0002 | 77 |
| Late phase | |||
| Younger at index CTc | −0.93 ± 0.31 | 0.003 | 51 |
| Baseline maximum ascending diameterd | 2.6 ± 0.62 | <0.0001 | 95 |
Percent of times factor appeared in 500 bootstrap models.
(Ascending aorta diameter/5), scaled transformation.
Ln(age), natural logarithm transformation.
(z value of ascending aorta diameter/12), scaled transformation.
CT = computed tomography; SE = standard error.
Comment
The appropriate timing of prophylactic ascending aortic replacement to prevent future dissection in patients with a bicuspid aortic valve lacks consensus. The 2006 American College of Cardiology/American Heart Association guidelines on valvular heart disease [17], the 2010 American College of Cardiology Foundation guidelines on thoracic aortic disease [18], and the 2013 Society of Thoracic Surgeons guidelines on aortic valve and ascending aorta [19] have all endorsed prophylactic ascending replacement at a diameter greater than 5.0 cm (Class I, Level of Evidence: C). Citing evidence from previous recommendations as “very limited and anecdotal,” 2014 American College of Cardiology Foundation/American Heart Association guidelines on valvular heart disease [20], in agreement with the 2014 European Society of Cardiology guidelines on the diagnosis and treatment of aortic diseases [21], raised this threshold to 5.5 cm based on 3 observational studies (Class I, Level of Evidence: B) [1, 2, 4].
In this series of patients with bicuspid aortopathy, we found the probability of type A dissection at the time of initial imaging to be highly dependent on ascending aortic diameter, followed by sinus diameter. Other patient-level variables were not found to be significantly associated with aortic dissection. The probability of dissection increased dramatically at an ascending aortic diameter of 5.3 cm and increased gradually at a sinus diameter of 5.0 cm.
This difference in risk attributable to aneurysm location may be explained by a recent series of 63 patients with type A dissection described by Rylski and colleagues [22] that showed an average dissection-induced increase in ascending aortic diameter of 32% (range, 4.0–5.3 cm); however, dimensions of the sinuses of Valsalva did not change significantly. Importantly, all aortic diameters measured in our patients with immediate dissection were postdissection diameters. In our own data, we see deviation from the low-risk, relatively constant probability of dissection at smaller sinus diameters occurring at a lower threshold than for ascending diameters (Fig 1). Therefore, although the thresholds we describe may help inform clinical decision making, they should be considered conservative.
In adhering to previous guidelines [17–19], it is our practice to offer patients with bicuspid aortic valve aortopathy early prophylactic aortic replacement at an ascending or sinus diameter greater than or equal to 5.0 cm or cross-sectional area to height ratio greater than approximately 10 cm2/m. The mean ascending diameter of those with dissection at presentation was 5.9 ± 1.6 cm compared with a predissection diameter of 5.0 ± 0.8 cm in the 7 patients in whom acute dissection developed during surveillance, among whom 5 dissections occurred at an ascending diameter less than 5.0 cm. This greater tendency toward dissection at smaller diameters in the surveillance patients may imply that dissections at larger diameters were prevented with early aortic replacement in those with an ascending diameter greater than 5.0 cm. Even at ascending diameters of 4.7 to 5.0 cm, there appears to be a constant 3% to 4% probability of type A dissection (Fig 1), suggesting that there is still risk associated with watchful waiting.
The progressive natural history of the disease and the risk of future aortic intervention further complicate the timing of aortic replacement. The rate of aortic growth in patients with bicuspid aortic valve is higher than in their counterparts with tricuspid aortic valves, and aneurysmal growth may occur independently of aortic valve dysfunction [23, 24]. Among patients with a functionally normal bicuspid aortic valve, Etz and colleagues [25] found that approximately 40% of those initially prescribed serial surveillance imaging required ascending aortic replacement for aneurysmal growth. Additionally, Borger and colleagues [26] found that 11% of patients who underwent isolated AVR for bicuspid aortic valve required later aortic intervention for growth or dissection. Similarly, McKellar and colleagues [27] showed that 11% of patients who underwent isolated AVR had later aortic dissection, aortic intervention, or progressive aortic enlargement at 15-year follow-up, stressing the impact of not addressing an enlarged aorta at the initial operation. Our previously published data, however, have shown that resecting the aorta when it is less than 4.5 cm at the time of AVR is not needed, and forgoing it rarely results in further complications or need for reoperation [28].
In our analysis of patients prescribed surveillance with serial imaging, we identified an early peaking phase of risk with high risk of aortic operation within 1 year. Multivariable analyses indicated that patients at higher risk of operation during this time were those with larger baseline ascending diameters and more recent initial CT/MRI. This is likely because of our institution's increasingly aggressive approach to bicuspid aortopathy over the nearly 2-decade study period. However, in the late phase, after 3 years of surveillance, we identified a constant 10% per year hazard of aortic intervention that was highly associated with younger age and a larger baseline ascending aortic diameter. Given this constant long-term risk of intervention, it appears that increasing the threshold for prophylactic ascending aortic replacement will only delay and not obviate surgical intervention.
Of those who underwent immediate aortic operations, hospital mortality was 0.25%, and risk of stroke was 0.75%. Other studies dedicated to demonstrating the safety of ascending aortic replacement in patients with bicuspid aortic valve at high-volume centers, including our own of 1,810 patients, have reported a hospital mortality of 0.4% to 2%, similar to that for AVR alone [28–31]. Any decision to operate has to be within the context of an individual institution's outcomes.
Prospective multicenter registries to identify patients with bicuspid aortic valve and prospectively conduct scheduled serial imaging and active outcome follow-up are critical to enable better modeling of the risk of aortic dissection and long-term outcomes of aortic replacement. At present, our data do not support advocating more conservative criteria (>5.5 cm) for prophylactic ascending aortic replacement and prolonging the time patients with a bicuspid aortic valve are at risk of dissection.
Limitations
This analysis is limited by its retrospective design and therefore is subject to selection bias. The prevalence of dissection within our population may be higher than in the general population because our institution is a major referral center for emergency aortic operations. However, our population is representative of other tertiary care centers where most patients with bicuspid aortopathy are followed, and we would not expect the diameter at which these patients experience dissection to be different from that within our cohort. The absolute value of the probability of dissection is dependent on the denominator of patients analyzed in a study by our inclusion criteria. However, the diameter at which the increase occurs from a lower almost constant probability of dissection at smaller ascending diameters to increased probability at higher diameters is not affected by this selection bias. Finally, the small number of patients with dissection found during surveillance limits our ability to determine risk factors for these events.
Conclusions
Ascending aortic diameter is the strongest identifiable patient-level predictor of type A dissection, and the probability of dissection increases dramatically at an ascending aortic diameter of 5.3 cm and increases gradually at a sinus diameter of 5.0 cm. The long-term risk of aortic intervention in patients initially prescribed serial imaging surveillance is approximately 10% per year, and intervention is more likely in younger patients with larger ascending aortic diameters at initial presentation. Elective ascending aortic replacement performed at high-volume aortic centers is associated with excellent short-term operative outcomes (death, 0.25%; stroke, 0.75%). Therefore, early prophylactic ascending aortic replacement should be considered at high-volume aortic centers to reduce the high risk of preventable type A dissection in patients with aortas larger than approximately 5.0 cm or with a cross-sectional area to height ratio of greater than approximately 10 cm2/m. At present, our data do not support advocating more conservative criteria (> 5.5cm) for prophylactic ascending aortic replacement and prolonging the time patients with bicuspid aortic valves are at risk of dissection.
Supplementary Material
Footnotes
Presented at the Fifty-first Annual Meeting of The Society of Thoracic Surgeons, San Diego, CA, Jan 24–28, 2015.
Dr Gillinov discloses a financial relationship with St. Jude, Abbott Laboratories, AtriCure, Inc, Edwards Lifesciences LLC, Medtronic, Inc, On-X Life Technologies, Inc, and Tendyne Medical, Inc.
The Appendix can be viewed in the online version of this article [http://doi.dx.org/10.1016/j.athoracsur.2015.04.126] on http://www.annalsthoracicsurgery.org.
References
- 1.Michelena HI, Khanna AD, Mahoney D, et al. Incidence of aortic complications in patients with bicuspid aortic valves. JAMA. 2011;306:1104–12. doi: 10.1001/jama.2011.1286. [DOI] [PubMed] [Google Scholar]
- 2.Braverman AC. Aortic involvement in patients with a bicuspid aortic valve. Heart. 2011;97:506–13. doi: 10.1136/hrt.2009.183871. [DOI] [PubMed] [Google Scholar]
- 3.Davies RR, Goldstein LJ, Coady MA, et al. Yearly rupture or dissection rates for thoracic aortic aneurysms: simple prediction based on size. Ann Thorac Surg. 2002;73:17–28. doi: 10.1016/s0003-4975(01)03236-2. [DOI] [PubMed] [Google Scholar]
- 4.Tzemos N, Therrien J, Yip J, et al. Outcomes in adults with bicuspid aortic valves. JAMA. 2008;300:1317–25. doi: 10.1001/jama.300.11.1317. [DOI] [PubMed] [Google Scholar]
- 5.R Development Core Team . Foundation for Statistical Computing. Vienna, Austria: 2013. R: a language and environment for statistical computing. Available at http://www.R-project.org. [Google Scholar]
- 6.Breiman L. Random forests. Machine Learning. 2001;45:5–32. [Google Scholar]
- 7.Ishwaran H, Kogalur UB. Random forests for survival, regression, and classification (RF-SRC) (R package version 1.6.0) Available at http://cran.r-project.org/web/packages/randomForestSRC/index.html.
- 8.Ehrlinger J. gg RandomForests: visually exploring Random Forests. (R package version 1.1.3) Available at: http://cran.r-project.org/web/packages/ggRandomForests/index.html.
- 9.Kaiser T, Kellenberger CJ, Albisetti M, Bergstrasser E, Valsangiacomo Buechel ER. Normal values for aortic diameters in children and adolescents—assessment in vivo by contrast-enhanced CMR-angiography. J Cardiovasc Magn Reson. 2008;10:56. doi: 10.1186/1532-429X-10-56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Svensson LG, Kim KH, Lytle BW, Cosgrove DM. Relationship of aortic cross-sectional area to height ratio and the risk of aortic dissection in patients with bicuspid aortic valves. J Thorac Cardiovasc Surg. 2003;126:892–3. doi: 10.1016/s0022-5223(03)00608-1. [DOI] [PubMed] [Google Scholar]
- 11.Breiman L, Friedman JH, Olshen R, Stone C. Classification and regression trees. Wadsworth and Brooks; Monterey, CA: 1984. [Google Scholar]
- 12.Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology. 1982;143:29–36. doi: 10.1148/radiology.143.1.7063747. [DOI] [PubMed] [Google Scholar]
- 13.Hand D, Till RJ. A simple generalization of the area under the ROC curve for multiple class classification problems. Machine Learning. 2001;45:171–86. [Google Scholar]
- 14.Blackstone EH, Naftel DC, Turner ME., Jr The decomposition of time-varying hazard into phases, each incorporating a separate stream of concomitant information. J Am Stat Assoc. 1986;81:615–24. [Google Scholar]
- 15.Breiman L. Bagging predictors. Machine Learning. 1996;24:123–40. [Google Scholar]
- 16.Blackstone EH. Breaking down barriers: helpful breakthrough statistical methods you need to understand better. J Thorac Cardiovasc Surg. 2001;122:430–9. doi: 10.1067/mtc.2001.117536. [DOI] [PubMed] [Google Scholar]
- 17.Bonow RO, Carabello BA, Kanu C, et al. ACC/AHA 2006 guidelines for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to revise the 1998 Guidelines for the Management of Patients With Valvular Heart Disease) Circulation. 2006;114:e84–231. doi: 10.1161/CIRCULATIONAHA.106.176857. [DOI] [PubMed] [Google Scholar]
- 18.Hiratzka LF, Bakris GL, Beckman JA, et al. 2010 ACCF/AHA/AATS/ACR/ASA/SCA/SCAI/SIR/STS/SVM guidelines for the diagnosis and management of patients with Thoracic Aortic Disease: a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines, American Association for Thoracic Surgery, American College of Radiology, American Stroke Association, Society of Cardiovascular Anesthesiologists, Society for Cardiovascular Angiography and Interventions, Society of Interventional Radiology, Society of Thoracic Surgeons, and Society for Vascular Medicine. Circulation. 2010;121:e266–369. doi: 10.1161/CIR.0b013e3181d4739e. [DOI] [PubMed] [Google Scholar]
- 19.Svensson LG, Adams DH, Bonow RO, et al. Aortic valve and ascending aorta guidelines for management and quality measures. Ann Thorac Surg. 2013;95:S1–66. doi: 10.1016/j.athoracsur.2013.01.083. [DOI] [PubMed] [Google Scholar]
- 20.Nishimura RA, Otto CM, Bonow RO, et al. 2014 AHA/ACC guideline for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Thorac Cardiovasc Surg. 2014;148:e1–132. doi: 10.1016/j.jtcvs.2014.05.014. [DOI] [PubMed] [Google Scholar]
- 21.Erbel R, Aboyans V, Boileau C, et al. 2014 ESC guidelines on the diagnosis and treatment of aortic diseases. Eur Heart J. 2014;35:2873–926. doi: 10.1093/eurheartj/ehu281. [DOI] [PubMed] [Google Scholar]
- 22.Rylski B, Blanke P, Beyersdorf F, et al. How does the ascending aorta geometry change when it dissects? J Am Coll Cardiol. 2014;63:1311–9. doi: 10.1016/j.jacc.2013.12.028. [DOI] [PubMed] [Google Scholar]
- 23.Davies RR, Kaple RK, Mandapati D, et al. Natural history of ascending aortic aneurysms in the setting of an unreplaced bicuspid aortic valve. Ann Thorac Surg. 2007;83:1338–44. doi: 10.1016/j.athoracsur.2006.10.074. [DOI] [PubMed] [Google Scholar]
- 24.Aydin A, Desai N, Bernhardt AM, et al. Ascending aortic aneurysm and aortic valve dysfunction in bicuspid aortic valve disease. Int J Cardiol. 2013;164:301–5. doi: 10.1016/j.ijcard.2011.07.018. [DOI] [PubMed] [Google Scholar]
- 25.Etz CD, Zoli S, Brenner R, et al. When to operate on the bicuspid valve patient with a modestly dilated ascending aorta. Ann Thorac Surg. 2010;90:1884–92. doi: 10.1016/j.athoracsur.2010.06.115. [DOI] [PubMed] [Google Scholar]
- 26.Borger MA, Preston M, Ivanov J, et al. Should the ascending aorta be replaced more frequently in patients with bicuspid aortic valve disease? J Thorac Cardiovasc Surg. 2004;128:677–83. doi: 10.1016/j.jtcvs.2004.07.009. [DOI] [PubMed] [Google Scholar]
- 27.McKellar SH, Michelena HI, Li Z, Schaff HV, Sundt TM., 3rd Long-term risk of aortic events following aortic valve replacement in patients with bicuspid aortic valves. Am J Cardiol. 2010;106:1626–33. doi: 10.1016/j.amjcard.2010.07.043. [DOI] [PubMed] [Google Scholar]
- 28.Svensson LG, Kim KH, Blackstone EH, et al. Bicuspid aortic valve surgery with proactive ascending aorta repair. J Thorac Cardiovasc Surg. 2011;142:622–9. doi: 10.1016/j.jtcvs.2010.10.050. [DOI] [PubMed] [Google Scholar]
- 29.Rinewalt D, McCarthy PM, Malaisrie SC, et al. Effect of aortic aneurysm replacement on outcomes after bicuspid aortic valve surgery: validation of contemporary guidelines. J Thorac Cardiovasc Surg. 2014;148:2060–9. doi: 10.1016/j.jtcvs.2014.03.027. [DOI] [PubMed] [Google Scholar]
- 30.Svensson LG, Al Kindi AH, Vivacqua A, et al. Long-term durability of bicuspid aortic valve repair. Ann Thorac Surg. 2014;97:1539–47. doi: 10.1016/j.athoracsur.2013.11.036. [DOI] [PubMed] [Google Scholar]
- 31.Svensson LG, Batizy LH, Blackstone EH, et al. Results of matching valve and root repair to aortic valve and root pathology. J Thorac Cardiovasc Surg. 2011;142:1491–8.e7. doi: 10.1016/j.jtcvs.2011.04.025. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.