Author + information
- Received April 21, 2015
- Revision received October 23, 2015
- Accepted November 11, 2015
- Published online February 16, 2016.
- Andrew G. Sherrah, BSc, MBBSa,
- Sarah Andvik, BMed, MDa,
- Denise van der Linde, MD, PhDb,
- Lucy Davies, MScc,
- Paul G. Bannon, PhDa,d,
- Ratnasari Padang, PhDa,
- Michael P. Vallely, PhDa,d,
- Michael K. Wilson, MBBSa,d,
- Anthony C. Keech, PhDb,c and
- Richmond W. Jeremy, PhDa,∗ ()
- aThe Baird Institute for Applied Heart and Lung Surgical Research, and Sydney Medical School, University of Sydney, Sydney, Australia
- bSydney Medical School, University of Sydney, Sydney, Australia
- cNHMRC Clinical Trials Centre, University of Sydney, Sydney, Australia
- dDepartment of Cardiothoracic Surgery, Royal Prince Alfred Hospital, Sydney, Australia
- ↵∗Reprint requests and correspondence:
Dr. Richmond W. Jeremy, Room 407, 100 Carillon Avenue, Newtown NSW 2042, Australia.
Background Genetic aortopathy (GA) underlies thoracic aortic aneurysms (TAA) in younger adults. Comparative survival and predictors of outcomes in nonsyndromic TAA (NS-TAA) are incompletely defined compared to Marfan syndrome (MFS) and bicuspid aortic valve (BAV).
Objectives The study sought to compare survival and clinical outcomes for individuals with NS-TAA, MFS, and BAV.
Methods From 1988 to 2014, all patients presenting with GA 16 to 60 years of age were enrolled in a prospective study of clinical outcomes. Risk factors for death and aortic dissection were identified by Cox proportional hazards modeling and a mortality risk score developed.
Results Diagnosis of GA was made for 760 patients (age 36.9 ± 13.6 years, 26.8% female; NS-TAA, n = 311; MFS, n = 221; BAV, n = 228). MFS patients were younger than NS-TAA and BAV. Presentation with aortic dissection was more common for NS-TAA than MFS or BAV. The 687 patients surviving >30 days after presentation were followed for a median of 7 years. Calculated 10-year mortality was 7.8% for NS-TAA, 8.7% for MFS, and 3.5% for BAV (NS-TAA and MFS vs. BAV p <0.05). Factors associated with all-cause mortality were MFS (p = 0.04), age at presentation, and family history of dissection.
Conclusions Clinical outcomes for MFS and NS-TAA are similar but worse than BAV. Independent predictors of mortality, including family history of aortic dissection and age, can be included in an Aortopathy Mortality Risk Score to predict survival. Management of NS-TAA, including surgical intervention, should be similar to that of MFS.
Thoracic aortic aneurysm (TAA) is associated with an estimated 10.4 cases of dissection per 100,000 person-years (1). As TAA is often asymptomatic until occurrence of dissection, the true prevalence may be greater. In younger patients, TAA is increasingly recognized to result from genetic aortopathy (GA) (2), including bicuspid aortic valve (BAV) (3), Marfan syndrome (MFS), Loeys-Dietz syndrome, and Aneurysms-Osteoarthritis syndrome (4–6). More recently, heritable TAA with few obvious physical features, described as nonsyndromic thoracic aortic aneurysm (NS-TAA) has been associated with mutations in multiple genes, including smooth muscle alpha-actin (ACTA2), transforming growth factor (TGF)-beta, and myosin heavy chain 11 (MYH11) (7,8). These GA disorders are associated with aneurysm and dissection throughout the thoracic aorta, although the ascending aorta is most frequently affected.
Diagnosis of a GA disorder raises issues of life expectancy, risk of adverse events (particularly aortic dissection), and timing of major surgery. For MFS and BAV, evidence is available regarding clinical outcomes in defined groups (3,9–15), but for NS-TAA, prognostic information remains more fragmentary and no large group studies are available. Current management guidelines do not specifically address this group and, for all GA disorders, there is a need for better identification of factors associated with increased risk of adverse outcomes.
Therefore, this study compared clinical outcomes in a large cohort with NS-TAA with those associated with MFS and BAV, and identified risk factors for mortality that could be used to calculate a patient risk score.
Royal Prince Alfred Hospital is a large metropolitan university hospital in Sydney, Australia. Since 1988, all patients presenting to the hospital with a suspected GA have been evaluated in the multidisciplinary Marfan and Aortic Disease Clinic. Affected individuals are enrolled in a prospective surveillance program that includes a screening examination offered to first-degree relatives. This study includes all patients enrolled between January 1, 1988, and January 1, 2014, 16 to 60 years of age at enrollment meeting diagnostic criteria for NS-TAA, MFS, or BAV. Diagnosis of MFS was made according to published criteria at enrollment (16–18); presence of BAV with aortic dilation was established by echocardiography or direct observation at surgery or autopsy. Diagnosis of NS-TAA required evidence of aortic dilation (sinus, ascending or descending aorta) or Type A or Type B dissection at <60 years of age, in the absence of other connective tissue syndromes, hypertension, or atherosclerotic disease; plus pathological evidence of aortic cystic medial necrosis or a positive family history (aortic aneurysm or dissection, sudden unexplained death, or known gene mutation). Patients with syndromic TAA (Ehlers-Danlos Type IV, Loeys-Dietz, Aneurysms-Osteoarthritis, Turner syndromes) are not included due to small numbers. When clinically indicated, detection of pathogenic gene variants (initially single gene sequencing and more recently panel testing for 19 genes associated with TAA) was undertaken for MFS and NS-TAA.
Surveillance included echocardiography plus computed tomography or magnetic resonance angiography as required. Echocardiography measurements included maximum diameter of the sinuses of Valsalva and ascending aorta via 2-dimensional left and right parasternal views, according to published recommendations (19,20) and aortic dilation, defined as diameter ≥2 standard deviations above the mean for age and body size (21). Follow-up included annual clinic review and review with local physicians for patients residing at a long distance from the study center. Follow-up was 97% complete, with 18 patients (BAV = 5, MFS = 5, NS-TAA = 8) lost. From 1990, patients were advised to take beta-adrenergic blockers at maximum-tolerated doses, and from 2012, angiotensin receptor blockers (ARBs) were added. Patients were referred for elective surgical repair of the aorta according to practice guidelines and clinical circumstances (20,22).
Survival and outcomes are documented for all patients enrolled in clinical surveillance excluding those who died within 30 days of presentation. Clinical outcomes are also documented for the subset of patients who did not present with initial aortic dissection. Adverse events were recorded as death from any cause, cardiovascular death (sudden death; death from aortic/arterial dissection/rupture, myocardial infarction, infective endocarditis, heart failure, or stroke; or in-hospital death following aortic or valve surgery), aortic dissection (Stanford type A or B, confirmed by definitive imaging or surgical findings), or need for aortic or valve surgery.
The study protocol and use of deidentified patient data from clinical records for this research was approved by the Human Research Ethics Committee of the hospital and all authors reviewed and approved the manuscript before submission.
SPSS version 21.0 (IBM Corporation, Armonk, New York) was used for data analysis. Normality of continuous data was determined by Kolmogorov-Smirnov test. Continuous variables are summarized as mean ± SD or median with interquartile range, and categorical variables as frequency and percentage. Group differences in variables were compared by 1-way analysis of variance, chi-square test, or Kruskal-Wallis test, as appropriate. All statistical tests were 2-sided; a p value <0.05 was considered statistically significant.
For survival analyses, patient follow-up is reported from date of enrollment to date of last clinical review up until January 1, 2014. Probabilities of (event-free) survival in each group were obtained by the Kaplan-Meier method and compared between groups by log-rank test; patients were stratified according to type of GA and type of presentation.
Potential predictors of survival were evaluated by Cox proportional hazard modeling. Initial entry of pre-specified covariates (type of GA, age at presentation, sex, type of presentation, family history of aortic aneurysm or dissection in a first-degree relative, maximum ascending aortic or aortic sinus diameter at presentation, history of aortic dissection and blood pressure at presentation) was followed by stepwise removal of nonsignificant covariates (p < 0.05 to enter, p > 0.10 to remove). The proportional hazards assumption was tested and verified using the method of Harrell et al. (23).
Predicted β-coefficients and baseline survivor functions of the final Cox proportional hazard model were used to develop an Aortopathy Mortality Risk Score, in which covariates predictive of all-cause mortality on multivariate analysis were specified cumulative scores. Concordance was assessed via Harrell’s C-statistic.
Among 910 clinic patients with GA, 760 met study inclusion criteria, including 311 NS-TAA patients, 221 MFS patients, and 228 individuals with BAV. Those excluded were syndromic TAA (n = 55) or <16 or >60 years of age (n = 95). Pathogenic mutations were identified for 129 (17.0%) patients (MFS: FBN1 n = 111; NS-TAA: ACTA2 n = 9, TGFβ2 n = 1, FBN1 n = 3, COL5A2 n = 4, MYH11 n = 1). Another 71 NS-TAA patients were tested but had no mutations identified. Cystic medial degeneration of the aorta was confirmed in 103 NS-TAA patients. The maximal aortic diameter was more commonly at the sinuses of Valsalva for MFS patients, compared with the ascending aorta for NS-TAA and BAV patients.
Diagnostic trajectory of genetic aortopathy
Demographics of all GA patients at presentation demonstrated significant differences among subgroups (Table 1). The MFS patients were younger (p < 0.001); both BAV and NS-TAA had male predominance (male:female ratios of 5:1 and 4:1, respectively; p < 0.001); and both NS-TAA and MFS were more likely to have a family history of aortic aneurysm or dissection than BAV (p < 0.001). The NS-TAA patients were more likely to present with aortic dissection (p < 0.001), while BAV patients more often presented following clinical diagnosis (p < 0.001). The MFS patients had smaller maximum aortic diameters at presentation than NS-TAA (p < 0.001) and BAV patients (p < 0.05) and both systolic and diastolic blood pressures were lower in MFS than in NS-TAA or BAV (p < 0.001). Two NS-TAA patients presented with isolated aneurysm of the descending aorta; the remainder had aortic sinus or ascending aortic aneurysm.
Seventy-three patients had fatal events as the first sign of GA, including 70 (94.6%) with aortic dissection, 2 (2.7%) sudden deaths, and 1 (1.4%) intraoperative death at urgent aortic repair surgery. The majority of fatal presentations had NS-TAA (79.5%) versus MFS (16.4%) or BAV (4.1%; p < 0.001).
Aortic dissection (fatal or nonfatal) was the first sign of GA in 131 patients (77.0% male) and the majority had NS-TAA (77.1%), followed by MFS (20.6%; p < 0.001). Type A aortic dissection was more common, while Type B accounted for 0.4% of BAV, 3.3% of MFS, and 6.9% of NS-TAA patients (vs. Type A; p < 0.001). The MFS patients were younger than NS-TAA or BAV patients at time of dissection (36.0 ± 8.6 years vs. 45.9 ± 9.2 years and 40.4 ± 11.4 years, respectively; MFS vs. NS-TAA, p < 0.001). Aortic diameter at time of dissection was known for 77 patients and was significantly smaller in NS-TAA than in MFS or BAV (47.9 ± 12.9 mm vs. 61.1 ± 20.9 mm and 55.0 ± 7.1 mm, respectively; p = 0.03).
Clinical surveillance outcomes
A total of 687 patients were enrolled in the prospective surveillance program after presentation (Table 2). Median follow-up was 7 years (interquartile range: 3 to 15 years) with 66.4% completing 5-year follow-up and 42.1% completing 10-year follow-up. The MFS patients had longer follow-up times than BAV or NS-TAA patients (p < 0.001). The NS-TAA patients were older than MFS, mostly male, and had larger aortic diameters than MFS. Clinical features of patients enrolled into the surveillance program, stratified according to type of presentation, are further described in Online Tables 1 to 3.
At last follow-up, 400 (58.2%) patients were taking regular medication, including beta-adrenergic blockers (n = 136; 19.8%), ARBs or angiotensin-converting enzyme (ACE) inhibitors (n = 109; 15.9%), or both a beta-adrenergic blocker and ARB/ACE inhibitor (n = 155; 22.6%). Most commonly used beta-adrenergic blockers were metoprolol (n = 128; mean dose: 67.0 ± 46.0 mg twice daily), and atenolol (n = 108; mean dose: 43.3 ± 19.1 mg daily); most common ARB was candesartan (n = 107; mean dose: 11.0 ± 7.2 mg/day) and most common ACE inhibitor was perindopril (n = 56; mean dose: 5.9 ± 2.9 mg/day). There were no differences in beta-blocker or ARB/ACE inhibitor use between GA groups.
Outcomes during surveillance
During surveillance, 77 (11.2%) patients died; causes included cardiovascular-related (n = 71), cancer (n = 5), and road trauma (n = 1). Cumulative mortality was comparable for MFS and NS-TAA and both were greater than for BAV (Figure 1A). Similar findings were observed when patients presenting with initial dissection were excluded (Figure 1B). For patients in surveillance who had presented with aortic dissection, mortality was greater; however, there was no difference in cumulative mortality between MFS and NS-TAA (Figure 1C). When calculated Kaplan-Meier 10-year mortality rates were compared between GA groups (stratified by presentation), the rate was comparable for NS-TAA and MFS and highest in NS-TAA patients presenting with initial nonfatal dissection (Table 3).
During this same period, 32 patients suffered aortic dissection: 15 Stanford type A (46.9%) and 17 type B (53.1%). At 10 years, the rate of aortic dissection during clinical surveillance was 3.6 ± 1.4%, 7.9 ± 2.2%, and 0% for NS-TAA, MFS, and BAV, respectively (log-rank BAV vs. MFS and NS-TAA p < 0.05) (Table 3). The MFS patients were more likely to suffer dissection than NS-TAA (p = 0.052). Patients with NS-TAA presenting via clinical diagnosis had greater risk of consequent aortic dissection than those presenting via family screening (p < 0.05). Univariate modeling of risk factors for type A and B dissections separately did not identify any significant predictive variables, most likely as a result of low event rates.
During follow-up, 291 patients (42.4%) required elective cardiac surgery. The majority of elective operations were performed in the first year after presentation and BAV patients most frequently required elective surgery (p < 0.001), especially isolated aortic valve replacement. There was no difference between MFS and NS-TAA groups for elective surgery (p = 0.24); however, MFS patients were more than a decade younger at the time of their first elective cardiac surgery (p < 0.001). Patients presenting with aortic dissection underwent less elective surgery compared with those presenting via family screening (p < 0.001); both underwent less elective surgery than those presenting via clinical diagnosis (p < 0.001) (Table 3).
Mortality risk score
Univariate Cox modeling of the clinical surveillance group identified variables associated with all-cause mortality as: type of aortopathy, age at presentation, presentation with aortic dissection or via clinical diagnosis, family history of aortic dissection in first-degree relative, initial aortic diameter, and history of nonfatal aortic dissection (Table 4). Type of aortopathy, age at presentation, and family history of dissection were independently associated (p < 0.10) with all-cause mortality in the final best-fit multivariate Cox model (Table 4) with a Harrell C-statistic of 0.68. These independent variables were applied to an Aortopathy Mortality Risk Score, with weightings according to Cox model coefficients:
1. Aortopathy type: BAV = 1 point; NS-TAA = 2 points; MFS = 3 points
2. Age at presentation: <25 years of age = 0 points; 25 to <40 years of age = 1 point; ≥40 years of age = 2 points
3. Family history of dissection: negative family history = 0 points; positive family history = 1 point
Mortality during follow-up for patients grouped according to their risk score is shown in Online Figure 1 (p < 0.001 by chi-square test for trend).
The clinical surveillance group patients were ranked by Aortopathy Mortality Risk Score as low risk (score 1 to 2; n = 124), medium risk (score 3 to 4; n = 456), and high risk (score 5 to 6; n = 107). When compared to the low-risk group, the medium-risk group had close to a 6-fold increase in observed mortality (hazard ratio: 5.6; 95% confidence interval: 1.7 to 18.4; p = 0.004) and the high-risk group had an 18-fold increase in observed mortality (hazard ratio: 17.9; 95% confidence interval: 5.3 to 59.6; p < 0.001). At 10 years of follow-up, mortality was 0.8 ± 0.8%, 5.8 ± 1.4%, and 15.8 ± 4.3% for the low-, medium-, and high-risk groups, respectively (overall p <0.001) (Central Illustration).
A broad spectrum of GA is now recognized, including familial aneurysm disease without multisystem features, known as NS-TAA. The adverse outcomes for syndromic GA such as MFS are well described (9–13), but data for NS-TAA have been limited. Current management guidelines advocate surgical repair of the aorta at a diameter >50 mm for MFS and >55 mm for other aneurysms, with the proviso that individuals with Loeys-Dietz syndrome and those with a family history of dissection may warrant intervention at lesser diameters (20,22). Indeed, the recent European guidelines note a lack of data about GA other than MFS (22). This study addressed this knowledge gap by documenting the natural history, as well as outcomes under clinical surveillance, for patients with NS-TAA.
Untreated MFS is associated with early dissection and death (9–13), as seen in some of our patients upon presentation. We found that the diagnostic trajectory of NS-TAA is as aggressive as that of untreated MFS, with comparable risk of aortic dissection. Survival of patients with MFS enrolled in the surveillance program was markedly improved, consistent with other reports (11), as was survival of NS-TAA patients in the surveillance program, but both groups still had worse outcomes than BAV. The mortality risk for MFS remained almost 4 times higher than for BAV and 2 times higher for NS-TAA patients. Further work is required to examine factors contributing to these residual mortality risks, which may include greater residual risk of dissection in MFS and NS-TAA and impaired left ventricular function in MFS (24,25).
Identification of individuals at greater risk of dissection and death can be problematic. Our data provided insight into independent predictors of mortality in this younger population with GA. Patients with a family history of aortic dissections had a 50% higher mortality risk than those without such a family history. Increasing age also was associated with greater mortality risk. Although NS-TAA appeared more frequently in males, male sex did not independently increase mortality risk; similar findings have been reported by the GenTAC (National Registry of Genetically Triggered Thoracic Aortic Aneurysms and Cardiovascular Conditions) registry (26). Although aortic diameter is widely accepted as a marker of dissection risk, initial maximum diameter of the ascending aorta or aortic sinuses was not associated with all-cause mortality on multivariate analysis, consistent with findings from the International Registry of Acute Aortic Dissection (27). Furthermore, our data suggested that dissection occurs at smaller aortic diameters in NS-TAA than in MFS.
To improve risk stratification for these patients, we developed an Aortopathy Mortality Risk Score, analogous to other risk scores for adverse cardiovascular events (28). The input variables—type of aortopathy, age category, and family history of dissection—and the resulting score will allow clinicians to determine whether an individual patient is low-, medium-, or high-risk, and to estimate likely intermediate- to long-term survival (Central Illustration). This risk score may be a valuable tool to guide management, including the frequency of diagnostic imaging and timing of elective surgery. Nonetheless, the true utility of the score for prediction of prognosis should be prospectively validated in another large cohort of patients with genetic aortic disease.
Despite surveillance and active management, aortic dissection still occurred during follow-up, and risk appeared greater for MFS than for NS-TAA. Our observations of low dissection rates in BAV were consistent with previous reports (3,14,15). There may be a shared disturbance of TGFβ signaling in MFS and many patients with NS-TAA, leading to adverse aortic remodeling, which may account for persistent risk of dissection (29,30). In contrast to all-cause mortality, the current dataset lacked power to clearly identify significant predictors for aortic dissection. A family history of aortic dissection was of borderline significance and was associated with a doubling of the risk of dissection. With this caveat, our data supported recommendations that a family history of dissection may be an indication for earlier elective surgical intervention (20,22).
Implications for management of NS-TAA
Both MFS and NS-TAA are associated with adverse prognosis, yet at diagnosis, NS-TAA patients were on average >10 years older than MFS patients and their maximum aortic diameters were larger. Importantly, of the 131 patients who first presented with acute aortic dissection, 77.1% had NS-TAA. These observations suggest that diagnosis of NS-TAA is often delayed, exposing these patients to the risk of an acute aortic event and early mortality. As patients with NS-TAA may have only subtle or even no other physical signs, it is hard to recognize them before an acute clinical event occurs. Clinicians should maintain a high index of suspicion and obtain a detailed family history and recommend systematic screening of first-degree relatives after a proband has been diagnosed with NS-TAA, even in the absence of other features of connective tissue disorders (20,22). In particular, clinicians should be alerted by occurrence of dissection in a younger adult and finding of cystic medial necrosis on pathology.
The similarity in observed outcomes between MFS and NS-TAA strongly suggested that the 2 conditions should be managed in the same manner. Our data would suggest that surgical repair of the ascending aorta should be undertaken for diameter >50 mm in NS-TAA (as for MFS), with consideration of repair at diameter >45 mm for those with a family history of dissection.
Although this study was the largest prospective study describing the long-term outcomes of patients with NS-TAA to date, the sample size and number of events were still limited by the prevalence of the disorder. We also recognize that referral bias may exist; however, our prospective approach of enrolling all eligible patients may reduce this bias. The patients were enrolled over a long period of time and outcomes may have been influenced by changes in medical or surgical practice during the study period. The presentation data were not, however, subject to this confounder and its effect on the surveillance cohort may have been mitigated by the consistent oversight of the whole cohort by 1 author (R.W.J.) since 1990.
The diagnosis of NS-TAA according to the presence of a family history of aortic disease was susceptible to patient recall bias and the quality of available clinical records of relatives. We also note that some data elements, such as aortic diameter at enrollment, were not 100% complete and that there were likely to be other determinants of prognosis, as yet unidentified, which may include novel biomarkers or specific gene variants. Finally, the risk score, although promising, should be validated in an independent dataset, preferably in larger multicenter prospective studies.
Aortic dissection is equally common in MFS and NS-TAA, but in NS-TAA patients, dissection is often the first sign of genetic aortic disease. Survival and outcomes are worse for both MFS and NS-TAA than for BAV. Independent predictors of mortality include family history of aortic dissection and age at time of diagnosis. These factors can be included in an Aortopathy Mortality Risk Score predicting mortality risk for individual patients. Clinicians should suspect NS-TAA in younger adults with unexplained aortic disease, warranting family screening. The management of NS-TAA, including timing of surgical repair of the aorta, should be similar to that of MFS.
COMPETENCY IN MEDICAL KNOWLEDGE: Aortic dissection is often the initial presentation of NS-TAA in patients with or without the MFS phenotype. Mortality rates for these groups are comparable, higher than for patients with BAV. A family history of aortic dissection strongly predicts mortality.
TRANSLATIONAL OUTLOOK: Clinical trials are needed to determine whether surgical repair of the ascending aorta and/or aortic sinuses at diameter thresholds of 50 mm for NS-TAA and 45 mm for patients with a family history of aortic dissection would reduce mortality compared with strategies recommended in current clinical practice guidelines.
Dr. Wilson has served as a consultant for Edwards Lifesciences; and a proctor for the Edwards SAPIEN Intuity. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- angiotensin-converting enzyme
- angiotensin receptor blocker
- bicuspid aortic valve aneurysm
- genetic aortopathy
- Marfan syndrome
- nonsyndromic thoracic aortic aneurysm
- transforming growth factor
- Received April 21, 2015.
- Revision received October 23, 2015.
- Accepted November 11, 2015.
- American College of Cardiology Foundation
- Saratzis A.,
- Bown M.J.
- van der Linde D.,
- van de Laar I.M.B.H.,
- Bertoli-Avella A.M.,
- et al.
- Groenink M.,
- Lohuis T.A.J.,
- Tijssen J.G.P.,
- et al.
- Loeys B.L.,
- Dietz H.C.,
- Braverman A.C.,
- et al.
- Cheitlin M.D.,
- Armstrong W.F.,
- Aurigemma G.P.,
- et al.
- Wolak A.,
- Gransar H.,
- Thomson L.E.J.,
- et al.
- Erbel R.,
- Aboyans V.,
- Boileau C.,
- et al.
- Holmes K.W.,
- Maslen C.L.,
- Kindem M.,
- et al.
- Pape L.A.,
- Tsai T.T.,
- Isselbacher E.M.,
- et al.