Abstract
Background
The natural history of aortic diseases in patients with TGFBR1 or TGFBR2 mutations reported by different investigators has varied greatly. In particular, the current recommendations for the timing of surgical repair of the aortic root aneurysms may be overly aggressive.
Methods and Results
The Montalcino Aortic Consortium, which includes 15 centers worldwide that specialize in heritable thoracic aortic diseases, was used to gather data on 441 patients from 228 families, with 176 cases harboring a mutation in TGBR1 and 265 in TGFBR2. Patients harboring a TGFBR1 mutation have similar survival rates (80% survival at 60 years), aortic risk (23% aortic dissection, 18% preventive aortic surgery), and prevalence of extra-aortic features (29% hypertelorism, 53% cervical arterial tortuosity, 27% wide scars) when compared to patients harboring a TGFBR2 mutation. However, TGFBR1 males had a greater aortic risk than females, whereas TGFBR2 males and females had a similar aortic risk. Additionally, aortic root diameter prior to or at the time of type A aortic dissection tended to be smaller in patients carrying a TGFBR2 mutation and was ≤45 mm in 6 women with TGFBR2 mutations, presenting with marked systemic features and low body surface area. Aortic dissection was observed in 1.6% of pregnancies.
Conclusions
Patients with TGFBR1 or TGFBR2 mutations show the same prevalence of systemic features, and the same global survival. Preventive aortic surgery at a diameter of 45 mm, lowered toward 40 in females with low body surface area, TGFBR2 mutation and severe extra-aortic features may be considered.
Keywords: Marfan syndrome, aortic dissection, aortic disease, gender differences, genetics, human, aneurysm, acute aortic dissection gene, aortic surgery, genetics, diagnostics, peripheral vascular disease
Journal Subject Terms: Vascular Disease, Mortality/Survival
Mutations in TGFBR2, encoding the TGF-β type II receptor, were first described in 2004 in patients with thoracic aortic root aneurysms and skeletal features of the Marfan syndrome (MFS)1. Subsequently, mutations in both TGFBR2 and TGFBR1 (TGF-β type I receptor) were described associated with early onset and aggressive thoracic aortic disease with MFS-like skeletal features, but also hypertelorism, craniosynostosis, developmental delay, cleft palate and bifid uvula, congenital heart disease and aneurysms and dissections throughout the arterial tree with marked arterial tortuosity; this condition was termed Loeys-Dietz syndrome2. Since then, phenotypes associated with TGFBR1 and TGFBR2 mutations have been shown to encompass milder forms of Loeys-Dietz syndrome, as well as autosomal dominant forms of thoracic aortic aneurysm and aortic dissection associated with no syndromic features and decreased penetrance3–5
Similar to the wide range of phenotypes associated with TGFBR1 and TGFBR2 mutations, the natural history of the thoracic aortic disease has been reported to have a wide spectrum, from cases of aggressive aortic disease with onset at an early age and a risk of dissection despite no or little enlargement of the aortic root6 to cases with a better prognosis close to that of classical Marfan syndrome due to FBN1 mutations4. The current recommendations for management of patients with TGFBR1 and TGFBR2 mutations are based on limited data, which are possibly skewed towards the early onset, severe end of the spectrum7,8.
Although clinical data from patients with either TGFBR1 or TGFBR2 mutations are typically aggregated in reporting presentation and natural history, data from one center found significant differences in aortic disease presentation in TGFBR1 versus TGFBR2 mutant patients9.
Importantly, ascending aortic dissections with minimal enlargement of the aortic root were identified with TGFBR2 mutations, whereas dissections in patients with TGFBR1 mutations were only reported in association with significant enlargement of the aorta, suggesting that recommendations for aortic management should be based also on the specific mutation.
The Montalcino Aortic Consortium (MAC, http://www.montalcinoaorticconsortium.org/), an international scientific group associating numerous centers with expertise in heritable thoracic aortic disease, was formed in 2013 to setup a multicenter retrospective registry of patients with genetically triggered thoracic aortic disease with a mission to define the optimal gene-based clinical management for these patients. MAC includes paediatricians, cardiologists, geneticists, cardiovascular surgeons. Here we report data on 441 patients with TGFBR1 or TGFBR2 mutations followed up in participating centers.
Materials and Methods
Population
All patients with TGFBR1 or TGFBR2 mutations from the participating centers were included in the registry. Data forms with key demographic, genotypic, phenotypic, and outcome variables based on current knowledge were defined in order to record relevant data in a standardised manner. The form was shared digitally with all centers in the MAC consortium. For each patient, investigators filled in the data fields and returned the form to the coordinating center. When uncertainties were noted, clarification was requested from the site investigator.
Clinical data for all individuals were obtained with informed consent of the patients’ parents/legal guardians (for children) or the patients themselves in accordance with the Declaration of Helsinki and national legal regulations.
The majority of the centers participating in the registry provide care for both children and adults; one center is a provider of healthcare for children only and one is for adults only. One center is a surgery referral center, whereas the others are either genetic referral centers, or referral centers that provide multidisciplinary care of heritable thoracic aortic disease. Familial screening is actively pursued by all centers.
The data collection period was from 12/2014 to 08/2015.
Mutation Analysis
TGFBR1 and TGFBR2 mutations were identified in clinical diagnostic laboratories or research labs with verification by a clinical lab. Blood samples were obtained after informed consent was provided by patients or parents, in agreement with local bioethics regulations. Mutation pathogenicity was evaluated with UMD-Predictor and Human Splicing Finder and confirmed in many cases by segregation with thoracic aortic disease in the family.
FBN1 mutation screening was performed either prior or at the time of TGFBR1/2 sequencing (aortopathy panel), in almost all patients. In less than 10 patients FBN1 screening was not performed after TGFBR1/2 sequencing had revealed a pathogenic mutation.
Statistical analysis
Statistical analyses were performed in an exploratory manner. Descriptive analyses used means and standard deviations and contingency tables as appropriate. Wilcoxon rank-sum tests were used to compare continuous variables between groups and Fisher exact tests for categorical distributions. Survival curves were estimated using the Kaplan-Meier method and compared by means of log-rank tests. Multinomial logistic regression was used to assess the association between the odds of undergoing preventive surgery or aortic dissection and aortic tortuosity and extra-arterial features. Model selection was performed using backward stepwise elimination using the Akaike information criterion. Cox modeling was also used to as a way to assess whether conclusions were sensitive to the choice of the model and strengthen the results. In both cases, a GEE approach was used to account for the within-family correlation arising from patients belonging to the same family.
P-values below 0.05 were considered as statistically significant. All calculations were performed using the R software10.
Results
Study cohort
A total of 441 patients in 228 families from 15 centers from around the world (Europe, US, Australia, Japan) were included in this study. General characteristics are reported in Table 1; The population of the surgical center differed from the population of other centers only in the fact that all the patients from the surgical center underwent surgery. Mutations in TGFBR2 were 50% more frequent than TGFBR1 mutations (Supplemental table 1). The two populations were similar in terms of the probands/non-probands ratio, gender distribution, and age at last follow-up. Of note, the systemic score required for the diagnosis of MFS (11) was ≥ 7 in 20% of both populations (Supplemental data fig 1). This underlines the phenotypic overlap between MFS due to mutations in the FBN1 gene and clinical phenotype in patients with TGFBR1 or TGFBR2 mutations.
Table 1.
Patient population
| TGFBR1 | TGFBR2 | p | |
|---|---|---|---|
| N | n=176 (40%) | n=265 (60%) | |
| Probands | 72 (42%) | 110 (43%) | 0.8 |
| De novo possible or confirmed | 27 (38%) | 52 (47%) | 0.2 |
| Females | 99 (56%) | 138 (52%) | 0.4 |
| Presenting feature | 0.7 | ||
| - Aortic root aneurysm | 30 (17%) | 60 (23%) | |
| - Ao dissection | 29 (17%) | 42 (16%) | |
| - Facial appearance | 16 (9%) | 23 (9%) | |
| - Familial screening | 100 (57%) | 137 (52%) | |
| Head and neck arterial tortuosity | 54/104 (51%) | 72/133 (54%) | 0.8 |
| Hypertelorism | 36/138 (26%) | 62/199 (31%) | 0.3 |
| Translucent skin | 63/144 (44%) | 78/227 (34%) | 0.08 |
| Wide scars | 33/142 (23%) | 63/218 (29%) | 0.3 |
| Broad or bifid uvula | 36/139 (26%) | 72/219 (33%) | 0.2 |
| Arched palate | 48/136 (35%) | 109/229 (48%) | 0.02 |
| Craniosynostosis | 11/118 (9%) | 20/190 (11%) | 0.8 |
| Marfan systemic score, mean (sd) | 3.98 (3.44) | 4.12 (2.94) | 0.3 |
| % systemic score 7 or greater | 28/128 (22%) | 34/183 (19%) | 0.5 |
| Cardiac defect (BAV, VSD, PDA…) | 13/154 (8.4%) | 52/238 (22%) | <0.001 |
| MVP | 36/156 (23%) | 65/244 (27%) | 0.5 |
| First aortic event | 0.13 | ||
| - Ascending thoracic aneurysm repair | 35 (50%) | 63 (53%) | |
| - Type A aortic dissection | 31 (44%) | 39 (33%) | |
| - Type B aortic dissection | 4 (6%) | 16 (13%) | |
| Ascending aortic diameter before repair: mm (SEM) | 68.3 (23.0) | 52.9 (13.4) | 0.09 |
| Type A aortic dissection | 36 (20%) | 48 (18%) | 0.5 |
| Type B aortic dissection | 9 (5%) | 23 (9%) | 0.19 |
| Extra-aortic arterial event | 20 (11.5%) | 23 (9%) | 0.5 |
Probands vs non-probands
The feature leading to discovery of the genotype differed between probands and non-probands, with probands more frequently being diagnosed based on the finding of an aneurysm, dissection, facial appearance, and less frequently presenting with family history as a reason for evaluation. Comparison of the 2 populations is provided in Supplemental table 2; probands show more aortic and extra-aortic features. Similar differences were observed between probands and non-probands when patients with either TGFBR1 or TGFBR2 mutations were analyzed separately.
Similar differences were observed between probands and non-probands when patients with either TGFBR1 or TGFBR2 mutations were analyzed separately.
TGFBR1 vs. TGFBR2
Patients with either TGFBR1 or TGFBR2 mutations were similar in most phenotypic features (Table 1). This was also true when only probands were considered (data not shown). Congenital heart defects were reported more frequently in cases with a TGFBR2 mutation than with a TGFBR1 mutation (ventricular septal defect (VSD): 6 TGFBR2 vs. 2 TGFBR1; atrial septal defect (ASD): 11 vs. 4, bicuspid aortic valve (BAV): 11 vs. 3, persistent ductus arteriosus (PDA): 25 vs. 3).
Survival did not differ between patients with a TGFBR1 or a TGFBR2 mutation (Figure 1). Forty-six deaths were recorded (19 TGFBR1, 27 TGFBR2), and occurred in 33 patients due to an aortic event (Supplemental Table 3). Median age of death was 28.4 years for TGFBR1 mutation patients and 34.8 years for TGFBR2 (p=0.2). Survival free of any vascular event (i.e. lethal or not) is shown in Figure 2, and did not differ between populations with either a TGFBR1 or a TGFBR2 mutation.
Figure 1.
Survival of the population: A : survival free of cardiovascular death (excluded cancer, septicemia, death of unknown cause). B: survival (all deaths included).
Figure 2.
Survival free of vascular features: A: survival free of any vascular feature (including discovery of arterial or aortic aneurysm, surgery or dissection). B:survival free of vascular or aortic event (including surgery or dissection)
However, a event-free survival differed significantly between men and women in the cohort of TGFBR1 mutation cases whereas there was no such difference in men and women in the cohort with TGFBR2 mutation (Figure 3). The proportion of patients with a type A aortic dissection was also greater in males than in females in the cohort with a TGFBR1 mutation (30% vs 12%, p= 0.004), as was the proportion of patients with any aortic event (defined as surgery, dissection or rupture) (52% vs 30%, p=0.005) or just an aortic dissection (33% vs 15%, p = 0.01). These differences were not observed in the TGFBR2 mutation cohort. Aortic root diameters obtained less than one year prior to a dissection or at the time of surgical repair of an aneurysm were variable (Figure 4).
Figure 3.
Survival free of any aortic event (surgery or dissection) in males compared to females. A population with a TGFBR1 mutation. B: population with a TGFBR2 mutation.
Figure 4.
Relation between age and aortic diameter measured before or at the time of event on the ascending aorta. A: surgery or dissection. Dissec A: dissection type A, Preventive Surg: preventive surgery of the ascending aorta. TGFBR1 and TGFBR2 refer to the mutation. X is age in year, Y is last aortic diameter measured before or at the time of the event (mm). B: diameter of the ascending aorta (mm) in patients with type A aortic dissection, according to the sex (M males, F females) and the gene mutated (TGFBR1 and TGFBR2). Only patients with aortic measures performed at a maximum of 1 year prior to aortic dissection are displayed.
To investigate whether prognosis information could be derived from molecular biology data, the survival of subgroups with at least 10 patients sharing the same altered amino-acid were compared to patients with a mutation affecting other amino-acids of the same protein. This was possible for 2 amino-acid substitutions from the protein kinase domain of the TGFBR1 protein, and 4 amino-acid substitutions from the protein kinase domain of the TGFBR2 protein (Supplemental figure 2 and 3).
Relation between aortic disease severity and other phenotypic features
We evaluated the risk for aortic root surgery or aortic dissection in the TGFBR1 and TGFBR2 cohorts as a function of the presence or absence of aortic tortuosity and extra-vascular features. In univariate analyses, all features were significantly associated with preventive surgery or dissection, except craniosynostosis (Table 2).
Table 2.
Patients with mutation either in TGFBR1 or TGFBR2: Prognostic value of aortic tortuosity and extra-aortic features on aortic risk (preventive surgery or dissection) Univariate analysis
| No event N=258 |
Preventive surgery N=80 |
Dissection N=103 |
p | |
|---|---|---|---|---|
| Hypertelorism | 41/201 (20.4%) | 28/66 (42.4%) | 29/70 (41.4%) | <0.001 |
| Bifid or broad uvula | 54/215 (25.1%) | 26/69 (37.7%) | 28/74 (37.8%) | 0.037 |
| Craniosynostosis | 14/172 (8.1%) | 10/60 (16.7%) | 7/78 (9.2%) | 0.169 |
| Arched palate | 84/217 (38.7%) | 42/70 (60.0%) | 31/78 (39.7%) | 0.006 |
| Aortic tortuosity | 16/137 (11.7%) | 18/66 (27.3%) | 23/69 (33.3%) | <0.001 |
| Head and neck arterial tortuosity | 57/125 (45.6%) | 33/62 (53.2%) | 35/50 (70.0%) | 0.013 |
| Wide scars | 48/220 (21.8%) | 15/64 (23.4%) | 33/76 (43.4%) | 0.001 |
| Translucent skin | 73/226 (32.3%) | 35/70 (50.0%) | 33/75 (44.0%) | 0.014 |
Note: patients with dissection (type A or B) despite prior preventive surgery were classified as “dissection”.
In a first multivariate model (table 3), we tested the craniofacial features previously reported to impact aortic outcome, specifically hypertelorism, bifid or broad uvula, craniosynostosis, and arched palate6, This was possible on 223 patients for whom these 4 characteristics were known. Hypertelorism was the only parameter that was found to be significantly associated with preventive surgery (OR 2.35, p=0.04) and tended to be associated with aortic dissection (OR 2.1, p=0.07). In a second multivariate model (Table 4), aortic tortuosity and all extra-arterial features were added as potential predictors (i.e., hypertelorism, broad or bifid uvula, craniosynostosis, arched palate, aortic tortuosity, tortuosity of the cervical arteries, wide scar, translucent skin). This analysis was possible on 145 patients. In adjusted analysis, the odds ratio (OR) for surgical repair of an aneurysm was greater than 2 for hypertelorism (2.6), aortic tortuosity (3.7), and translucent skin (3.0). Similarly, the OR for dissection was greater than 2 for hypertelorism (2.1), aortic tortuosity (3.9), and wide scars (4.0). Using Akaike stepwise descending selection, the previous criteria and the cervical artery tortuosity were selected; these criteria were then tested again in the 185 patients with complete data for these features (Supplemental table 4). Akaike stepwise descending selection selected hypertelorism, aortic tortuosity, and wide scars. Fitting the model on the dataset with complete data on these three features (n=214 patients) confirmed the previous results (Table 5). Analysis using Cox model provided similar results (supplemental table 5 and 6).
Table 3.
Patients with mutation either in TGFBR1 or TGFBR2: Prognostic value of aortic tortuosity and extra-aortic features on aortic risk (preventive surgery or dissection). Multinomial logistic model: prognostic value of features of the cranio-facial index in the population of 223 patients with all the data.
| Preventive Surgery OR, CI95, p |
Dissection OR, CI95, p |
|
|---|---|---|
| Hypertelorism | OR=2.11, [0.99; 4.53], p=0.05 | OR=2.05, [0.90; 4.67], p=0.09 |
| Broad or bifid uvula | OR=1.12, [0.51; 2.50], p=0.77 | OR=1.42, [0.67; 3.04, p=0.36 |
| Craniosynostosis | OR=1.23, [0.42; 3.64], p=0.70 | OR=1.11, [0.33; 3.76], p=0.87 |
| Arched palate | OR=1.27, [0.66; 2.45, p=0.47 | OR=0.90, [0.43; 1.87], p=0.77 |
Table 4.
Patients with mutation either in TGFBR1 or TGFBR2: Prognostic value of aortic tortuosity and extra-aortic features on aortic risk (preventive surgery or dissection). Prognostic value of aortic tortuosity and extra-aortic features in the population of 145 patients with all the data
| Preventive Surgery OR, CI95, p |
Dissection OR, CI95, p |
|
|---|---|---|
| Hypertelorism | OR=2.30, [0.87; 6.02], p=0.09 | OR=2.04, [0.49; 8.45], p=0.32 |
| Broad or bifid uvula | OR=0.88 [0.29; 2.69], p=0.82 | OR=1.38, [0.45; 4.28, p=0.58 |
| Craniosynostosis | OR=1.72, [0.38; 7.81], p=0.49 | OR=1.30, [0.26; 6.43], p=0.75 |
| Arched palate | OR=1.33, [0.51; 3.49], p=0.56 | OR=0.53, [0.14; 1.99], p=0.35 |
| Aortic tortuosity | OR=3.41, [0.75; 15.52], p=0.11 | OR=3.50, [0.89; 13.86], p=0.07 |
| Cervical artery tortuosity | OR=0.43, [0.12; 1.57], p=0.20 | OR=1.49, [0.52; 4.30], p=0.46 |
| Wide scars | OR=0.35, [0.08; 1.46], p=0.15 | OR=4.04, [1.18; 13.86], p=0.03 |
| Translucent skin | OR=2.64 [0.85; 8.19], p=0.09 | OR=1.06, [0.31; 3.59], p=0.92 |
Table 5.
Patients with mutation either in TGFBR1 or TGFBR2: Prognostic value of aortic tortuosity and extra-aortic features on aortic risk (preventive surgery or dissection). Prognostic value of selected features in the population of 214 patients with all the data, after model selection based on the Akaike criterion
| Preventive Surgery OR, CI95, p |
Dissection OR, CI95, p |
|
|---|---|---|
| Hypertelorism | OR=2.37, [1.08; 5.18], p=0.03 | OR=2.21, [0.97; 5.04], p=0.06 |
| Aortic tortuosity | OR=2.04, [0.78; 5.36], p=0.15 | OR=3.80, [1.64; 8.79], p<0.01 |
| Wide scars | OR=0.73, [0.30; 1.81], p=0.50 | OR=2.99, [1.42; 6.28], p<0.01 |
The occurrence of an extra-aortic event (vascular repair, dissection, or rupture) tended to be more frequent in the patients who had an aortic dissection, (15% vs 9%, p=0.09), with no difference between sex. In contrast, the proportion of patients with an elevated systemic score ≥711, was identical in patients with or without previous aortic dissection (19% vs 22%, p=0.6).
Presentations with Acute Aortic Dissections
Type A aortic dissection was the first aortic event in 71 patients, 32 in TGFBR1 carriers and 39 in TGFBR2 carriers. Type B dissection was the first aortic event in 19 patients, 3 TGFBR1 carriers (1.7% of the population) and 16 TGFBR2 carriers (6% of the population; p = 0.03).
A measure of the aorta before or at the time of the type A aortic dissection was available in 31 patients, 22 of which had these measurements within one year or at the time of dissection (Figure 4b). This diameter was smaller in patients with a TGFBR2 mutation than those with a TGFBR1 mutation (TGFBR2, 51.8±13.4 mm vs.TGFBR1, 68.3±23.0 mm, p=0.06).
The diameter of the aortic root measured within one year of the aortic event was smaller or equal to 45 mm in 7 patients. Interestingly, all these patients were women and six had TGFBR2 mutations, with extra-aortic features. A brief summary of presentation of these cases is provided in Supplemental data.
Of note, among the 98 patients who had aortic root repair initially (63 TFBR2, 35 TGFBR1), dissection of the ascending aorta was subsequently observed in 10 patients (4 TGFBR1, 6 TGFBR2).
Pregnancies
One hundred and twenty-two women had 316 pregnancies (table 6). The diagnosis was not known in 103 of these women (84%); Aortic dissection occurred during 5 (1.6%) pregnancies Three dissections were type A aortic dissections: a TGFBR1 had preventive surgery and was receiving a beta-blocker; 1 TGFBR2 mutation had preventive aortic surgery (aortic diameter 48 mm) and dissected above the graft (aortic diameter 31 mm), and the aortic root diameter was 35 mm at the time of type A aortic dissection in the last TGFBR2 woman.
Table 6.
Pregnancies and dissections in women harboring a mutation in TGFBR1 or TGFBR2.
| TGFBR1 | TGFBR2 | Total | |
|---|---|---|---|
| Women with pregnancy | 53 | 69 | 122 |
| N.pregnancy | 138 | 178 | 316 |
| 0 | 33 | 60 | 93 |
| 1 | 14 | 10 | 24 |
| 2 | 19 | 28 | 47 |
| 3 | 8 | 19 | 27 |
| 4 | 7 | 7 | 14 |
| 5 | 3 | 3 | 6 |
| 6 | 1 | 2 | 3 |
| 13 | 1 | 0 | 1 |
| Ao dissection during pregnancy | 1 (0.7%) | 4 (2.2%) | 5 (1.6%) |
| 1st pregnancy | 1 | 1 | 2 |
| 2nd pregnancy | 2 | 2 | |
| 3rd pregnancy | 1 | 1 |
The 2 type B aortic dissections were observed in women with a TGFBR2 mutation during their first pregnancy (age 34 y.o.) and third pregnancy (age 30 y.o.), respectively.
Interestingly, the prevalence of extra-aortic features was higher in women when pregnancy was complicated by an aortic dissection: vertebral artery tortuosity [100% (4/4) vs. 46% (22/48)], hypertelorism [100% (4/4) vs. 23% (16/69)], translucent skin [75% (3/4) vs. 56% (58/86)], bifid or broad based uvula [75% (3/4) vs. 22% (17/79)], arched palate [100% (4/4) vs. 44% (34/78)]. The systemic score for Marfan syndrome was 5 in two women with aortic dissection during pregnancy and 9 in one whereas it was 7 or greater in 8/73 women with pregnancies without aortic dissection. No uterine rupture was reported.
Discussion
We report here the largest cohort of patients with mutations in either the TGFBR1 or TGFBR2 gene. Although TGFBR1 and TGFBR2 proteins cooperate to initiate signaling after binding to the TGF-β ligand, the clinical features associated with a mutation in the TGFBR1 gene differ from those associated with a mutation in TGFBR2 gene. The prevalence of aortic aortic events (surgery or dissection) was lower in females than in males with a TGFBR1 mutation, whereas no similar sex effect was observed in those with a TGFBR2 mutation. In fact in female patients harboring a TGFBR2 mutation, type A dissections of moderately dilated ascending aorta appeared more frequently than in males, which was not the case with TGFBR1. This suggests that the aortic features may differ between the two populations, with more aggressive aortic disease in patients with TGFBR2 mutations, especially for women. Such specificity has previously been suggested in a population by Tran-Fadulu et al9. The reason for this observation remains unknown. Additionally, type B aortic dissections tend to be more frequent as the presenting feature in patients with a TGFBR2 mutation compared to those with a TGFBR1 mutation. It is notable that the non-aortic systemic phenotypic features were very similar between TGFBR1 and TGFBR2 mutation patients, with both groups also having similar MFS systemic scores and similar prevalence of features of Loeys-Dietz syndrome, such as hypertelorism, broad or bifid uvula, wide scars, and translucent skin.
The observation of different event-free survivals according to the affected amino-acid within the protein kinase domain also illustrates the importance of genetic prognostic factors (supplemental figures 2 and 3), and heterogeneity of these pathologies.
The second striking observation in our cohort is that the observed survival is much better than was initially reported6. The improved survival in this population as compared to this initial report may be due to several factors. First, an ascertainment bias often exists when new pathologies are identified, with the most severe forms initially being recognized and milder phenotypes only reported later. This was the case with Marfan syndrome12. In addition, the initial report of patients with TGFBR1 and TGFBR2 gene mutations was limited to patients with marked and severe aortic disease and significant features of MFS and LDS, which is in distinct contrast to the population reported here. The fact that the presence of these features was associated with a higher likelihood of an aortic event such as aortic dissection supports the hypothesis that the initial cohort reported was particularly severely affected. This underscores the variability of the phenotype and the spectrum of disease severity that can be associated with a mutation in these genes4. Second, the benefits of better diagnosis and care of patients with these mutations may be contributing to their improved survival. Early diagnosis before the advent of a catastrophic event allows for prophylactic aortic surgery and a reduced risk of aortic dissection, a benefit previously emphasized in this population4. The results of aortic surgery in this group of patients appear to be similar to those observed in patients with MFS, i.e. a decreased risk for dissection and a sustained benefit over time13,14. The fact that the benefit is at least in part the result of the care provided to these patients is also suggested by the divergence of the vascular event-free survival curve and the crude survival curve: the vascular event-free survival curve (i.e. survival without surgery or dissection) begins to decline significantly after the age of 15 years, and diverges from the survival curve above this age. This is a main difference between patients carrying a TGFBR1 or TGFBR2 gene mutation and patients carrying a COL3A1 gene mutation whose prognosis remains poor with or without surgery, and underscores the benefit of an early diagnosis15.
The data reported here indicates that surgical repair that is limited to the aortic root may be inefficient in individuals with either TGFBR1 or TGFBR2 gene mutations. Ten percent of the patients who had surgery for an aortic root aneurysm subsequently presented with dissections of the ascending aorta. This observation suggests that aortic replacement should include the entire ascending aorta when possible. Additionally, limiting the risk for dissection after surgical repair of the aortic root is important, leading to advise 1) avoidance of strenuous exercise and 2) use of beta-blockade16,17.
We also found an association between risk for either aortic surgery or dissection and the presence of either aortic tortuosity or extra-aortic features (Table 2). A similar association was previously reported4,6. When multivariate analysis was performed in our population, hypertelorism, aortic tortuosity and wide scars were the three factors that remained significantly associated with aortic dissection. The interpretation of this last result is not straightforward, and some limitations have to be underlined, such as the limited size of the population in which numerous factors have been investigated and the incomplete evaluation of some of the patients which may have led to an assessment biased (biased subgroup). Additionally, the population is being medically managed and hopefully this management has changed the spontaneous “natural history” into a better “medical history”, in which the prognostic factors are less easy to detect. It is striking that the arterial tortuosity and extra-aortic features were present in almost all the women who presented type A aortic dissection with little to minimal aortic root enlargement and in the women who experienced dissections associated with pregnancy. In fact, presence of these features may assist in defining the optimal timing of surgery in this population, as it is suggested that syndromic patients, specifically those patients with features of LDS, should be managed more aggressively than patients lacking these syndromic features.
It is notable that a maximal diameter of the aorta below 45 mm on measurements less than one year prior to dissection or at the time of an ascending aortic dissection was found only in women with a TGFBR2 mutation. In contrast, no dissection was observed below 45 mm in men with either TGFBR1 or TGFBR2 mutations or in women with TGFBR1 mutations.
It is true that optimal diameter for surgery can only be derived from systematic prospective follow-up18, and should also depend on the risk associated with aortic root surgery in this population. There is also a discussion that aortic diameter should be normalised by BSA, including a proposal by Svensson et al.19 but the seminal paper from Coady uses non normalised diameter20. The recent study in Marfan syndrome also does not support the use of normalized aortic diameter to decide on the timing of surgery18. Considering the available data, including the data reported here, a recommendation could be made to consider surgery usually in adults at a diameter of 45 mm (absolute value), which is lower than that recommended for patients with FBN1 gene mutations (MFS patients)5,6. This threshold could be lowered toward 40mm for females with TGFBR2 gene mutations and severe phenotypic features associated with aortic dissection in our population, such as aortic tortuosity, hypertelorism and wide scarsand low body surface areas. A less aggressive surgical management strategy could be pursued in the absence of any familial event or rapid increase in aortic diameter, especially in women harboring TGFBR1 mutations. This scheme would take into account the great variability of the vasculopathy associated with mutations in TGFBR1 or TGFBR2, and its relation with the extra-aortic features. However, defining an aortic diameter for surgery from our population with no or incomplete follow-up, often no therapy, and no education at the time of the aortic event (an aortic event was the trigger for diagnosis in more than 50% of the patients) may not be perfectly relevant for an educated patient with a clear diagnosis who is receiving beta-blocker therapy. Then, the risk for aortic dissection may be lowered, therefore delaying optimal timing for aortic surgery, including possibly waiting for a larger diameter. Besides, the prognostic information associated with the affected amino-acid within the protein may become also useful to define optimal surgical timing in coming years (Supplemental figure 2 and 3).
The fourth conclusion that can be derived from our cohort is that pregnancy is associated with an increased risk of aortic dissection compared to the general population. This risk is moderate however (5 dissections among 316 pregnancies in 122 women), and the underlying genetic disease was not recognized before dissection in 4 of the 5 pregnant women who dissected. In 2 patients, a previous prophylactic valve-sparing aortic surgery did not prevent the occurrence of dissection, which highlights the fact that the entire aorta is abnormal, and that surgery is not a definitive cure in these diseases (In fact, early surgery for early aortic dilatation may actually be a risk factor for a distal aortic event, and 10% of our patients with prophylactic aortic root surgery presented with subsequent dissection of the ascending aorta). The low incidence of complications during pregnancies was reported previously in other smaller series of patients with a TGFBR2 mutation4, and this study extends the results to patients with a TGFBR1 mutation. Of note, no uterine rupture was reported in the cohort used in this study.
Lastly, our series again underscores the overlap observed between the clinical features observed in patients with classical MFS due to FBN1 mutations, LDS as initially described due to TGFBR1 and TGFBR2 mutations, and an inherited risk for thoracic aortic disease in the absence of syndromic features due to mutations in these same genes. In the cohort reported here, 20% of the population demonstrated multiple phenotypic features found in Marfan syndrome with an elevated systemic score ≥711.
There are some limitations to this study. The retrospective nature of the data may have introduced a bias, although the centers reported all the patients seen with a TGFBR1 or a TGFBR2 mutation in this registry. The multi-center nature of this report makes it less prone to selection bias, which is also minimized by inclusion of different centers driven by different specialties (paediatricians, geneticists, cardiologists, cardio-vascular surgeons). On the other hand, the centers that participated in this survey are all secondary or tertiary reference centers, so that the more severe forms may be over-represented. It also cannot be excluded the possibility that chance may have played a role in some of the significant associations that we found. This could not be avoided as we performed the statistical analyses in an exploratory manner, and thus performed a substantial number of statistical tests; however 2 different models (i.e. multinomial and Cox modeling) were used which led to similar results and thus strengthen the findings. Our results should nevertheless be taken cautiously and future confirmation of our findings will be needed before final conclusions can be drawn.
Lastly and most importantly, many of the phenotypic features are subjective, so that the presence of hypertelorism, wide scars, broad uvula, and arterial tortuosity may not reflect identical entities within the different centers and throughout the literature. However imprecise definition of clinical features is also a limit of all the previous reports on the topic. It is one of the aims of the MAC to propose consensus definitions for all these features. This work is already ongoing through assessing tortuosity by a defined vertebral artery tortuosity index21 and similar assessments for aortic tortuosity22.
In conclusion, the survival of patients with a TGFBR1 or a TGFBR2 gene mutation appears to be better than initially reported. The presence of a TGFBR2 mutation, female gender, aortic tortuosity, hypertelorism and translucent skin are associated with an increased aortic dissection risk and may help to determine the optimal surgical timing (45 mm in the general population, lowered toward 40 in females with low BSA harboring a TGFBR2 mutation and presenting extra-aortic features could be proposed). In coming years, the prognostic value of specific mutations may also become useful for further individualization of surgical threshold. Lastly, pregnancy appears not to have the ominous prognosis that was initially reported.
Supplementary Material
Clinical Perspective.
The general prognosis of patients with a TGFBR2 or a TGFBR1 mutation has been reported to be very poor. This large multicenter, international study demonstrates that the prognosis is actually much better than initially reported (with around 75% survival at 75 years), similar in men and women with a TGFBR1 mutation, and with earlier aortic outcomes in men than women with a TGFBR2 mutation. However, the phenotype and the disease severity are quite variable between patients. Some phenotypic factors are associated with more aggressive vascular disease: the presence of a TGFBR2 mutation, female sex, presence of aortic tortuosity, hypertelorism or translucent skin are associated with an increased risk of aortic dissection. Based on our data, preventive aortic root replacement could be proposed at 45 mm, a threshold which could be lowered toward 40 mm in females with low BSA harboring a TGFBR2 mutation and presenting with extra-aortic features. Some mutations appear to be associated with earlier surgery for aortic disease.
Acknowledgments
We thank the patients, their families and the referring physicians without whom this study would not have been possible.
Sources of Funding: G Jondeau received grant from Agence Nationale de la Recherche (ANR-14-CE15-0012), Fédération Française de Cardiologie, Fondation Cœur et Recherche. C Boileau was supported by Programme Hospitalier de Recherche Clinique (AOM10108), Contrat de Recherche Clinique Assistance Publique Hopitaux de Paris (CRC15014) and Fondation Maladies Rares. J De Backer holds a grant as Senior Clinical Investigator from the Fund for Scientific Research (FWO), Flanders (Belgium). L Muiño Mosquera is supported by a doctoral fellowship from the Special Research Fund (BOF) of the University of Ghent (Belgium). T Morisaki and H. Morisaki are beneficiary of a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and grants from the Japan Science and Technology Corporation, Ministry of Health, Labour and Welfare of Japan, the Program for Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation (NIBIO) and Japan Agency for Medical Research and Development (AMED). Shaine Morris is supported by National Heart, Lung, And Blood Institute of the National Institutes of Health award K23HL127266. DM Milewicz is funded by NIH R01 HL62594, the John Ritter Foundation, Genetic Aortic Disorders Association (GADA), and the Temerty Family Foundation.
Footnotes
Disclosures: None
References
- 1.Mizuguchi T, Collod-Beroud G, Akiyama T, Abifadel M, Harada N, Morisaki T, et al. Heterozygous TGFBR2 mutations in Marfan syndrome. Nat Genet. 2004;36:855–860. doi: 10.1038/ng1392. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Loeys BL, Chen J, Neptune ER, Judge DP, Podowski M, Holm T, et al. A syndrome of altered cardiovascular, craniofacial, neurocognitive and skeletal development caused by mutations in TGFBR1 or TGFBR2. Nat Genet. 2005;37:275–281. doi: 10.1038/ng1511. [DOI] [PubMed] [Google Scholar]
- 3.Pannu H, Fadulu VT, Chang J, Lafont A, Hasham SN, Sparks E, et al. Mutations in transforming growth factor-beta receptor type II cause familial thoracic aortic aneurysms and dissections. Circulation. 2005;112:513–520. doi: 10.1161/CIRCULATIONAHA.105.537340. [DOI] [PubMed] [Google Scholar]
- 4.Attias D, Stheneur C, Roy C, Collod-Béroud G, Detaint D, Faivre L, et al. Comparison of clinical presentations and outcomes between patients with TGFBR2 and FBN1 mutations in Marfan syndrome and related disorders. Circulation. 2009;120:2541–2549. doi: 10.1161/CIRCULATIONAHA.109.887042. [DOI] [PubMed] [Google Scholar]
- 5.Matyas G, Arnold E, Carrel T, Baumgartner D, Boileau C, Berger W, et al. Identification and in silico analyses of novel TGFBR1 and TGFBR2 mutations in Marfan syndrome-related disorders. Hum Mutat. 2006;27:760–769. doi: 10.1002/humu.20353. [DOI] [PubMed] [Google Scholar]
- 6.Loeys BL, Schwarze U, Holm T, Callewaert BL, Thomas GH, Pannu H, et al. Aneurysm syndromes caused by mutations in the TGF-beta receptor. N Engl J Med. 2006;355:788–798. doi: 10.1056/NEJMoa055695. [DOI] [PubMed] [Google Scholar]
- 7.Hiratzka LF, Bakris GL, Beckman JA, Bersin RM, Carr VF, Casey DE, Jr, 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]
- 8.Erbel R, Aboyans V, Boileau C, Bossone E, Bartolomeo RD, Eggebrecht H, et al. 2014 ESC Guidelines on the diagnosis and treatment of aortic diseases: Document covering acute and chronic aortic diseases of the thoracic and abdominal aorta of the adult. The Task Force for the Diagnosis and Treatment of Aortic Diseases of the European Society of Cardiology (ESC) Eur Heart J. 2014;35:2873–2926. doi: 10.1093/eurheartj/ehu281. [DOI] [PubMed] [Google Scholar]
- 9.Tran-Fadulu V, Pannu H, Kim DH, Vick GW, 3rd, Lonsford CM, Lafont AL, et al. Analysis of multigenerational families with thoracic aortic aneurysms and dissections due to TGFBR1 or TGFBR2 mutations. J Med Genet. 2009;46:607–613. doi: 10.1136/jmg.2008.062844. [DOI] [PubMed] [Google Scholar]
- 10.Team RC. R: A language and environment for statistical computing. R Foundation for Statistical Computing; Vienna, Austria: 2012. [Google Scholar]
- 11.Loeys BL, Dietz HC, Braverman AC, Callewaert BL, De Backer J, Devereux RB, et al. The revised Ghent nosology for the Marfan syndrome. J Med Genet. 2010;47:476–485. doi: 10.1136/jmg.2009.072785. [DOI] [PubMed] [Google Scholar]
- 12.Pyeritz RE. Marfan syndrome: 30 years of research equals 30 years of additional life expectancy. Heart. 2009;95:173–175. doi: 10.1136/hrt.2008.160515. [DOI] [PubMed] [Google Scholar]
- 13.Williams JA, Hanna JM, Shah AA, Andersen ND, McDonald MT, Jiang YH, et al. Adult surgical experience with Loeys-Dietz syndrome. Ann Thorac Surg. 2015;99:1275–1281. doi: 10.1016/j.athoracsur.2014.11.021. [DOI] [PubMed] [Google Scholar]
- 14.Iba Y, Minatoya K, Matsuda H, Sasaki H, Tanaka H, Morisaki H, et al. Surgical experience with aggressive aortic pathologic process in Loeys-Dietz syndrome. Ann Thorac Surg. 2012;94:1413–1417. doi: 10.1016/j.athoracsur.2012.05.111. [DOI] [PubMed] [Google Scholar]
- 15.Malfait F, De Paepe A. The Ehlers-Danlos syndrome. Adv Exp Med Biol. 2014;802:129–143. doi: 10.1007/978-94-007-7893-1_9. [DOI] [PubMed] [Google Scholar]
- 16.Shores J, Berger KR, Murphy EA, Pyeritz RE. Progression of aortic dilatation and the benefit of long-term beta-adrenergic blockade in Marfan’s syndrome. N Engl J Med. 1994;330:1335–1341. doi: 10.1056/NEJM199405123301902. [DOI] [PubMed] [Google Scholar]
- 17.Cheng A, Owens D. Marfan syndrome, inherited aortopathies and exercise: what is the right answer? Heart. 2015;101:752–757. doi: 10.1136/heartjnl-2014-306440. [DOI] [PubMed] [Google Scholar]
- 18.Jondeau G, Detaint D, Tubach F, Arnoult F, Milleron O, Raoux F, et al. Aortic event rate in the Marfan population: a cohort study. Circulation. 2012;125:226–232. doi: 10.1161/CIRCULATIONAHA.111.054676. [DOI] [PubMed] [Google Scholar]
- 19.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;126:892–893. doi: 10.1016/s0022-5223(03)00608-1. [DOI] [PubMed] [Google Scholar]
- 20.Coady MA, Rizzo JA, Hammond GL, Mandapati D, Darr U, Kopf GS, et al. What is the appropriate size criterion for resection of thoracic aortic aneurysms? J Thorac Cardiovasc Surg. 1997;113:476–491. doi: 10.1016/S0022-5223(97)70360-X. [DOI] [PubMed] [Google Scholar]
- 21.Morris SA, Orbach DB, Geva T, Singh MN, Gauvreau K, Lacro RV. Increased vertebral artery tortuosity index is associated with adverse outcomes in children and young adults with connective tissue disorders. Circulation. 2011;124:388–396. doi: 10.1161/CIRCULATIONAHA.110.990549. [DOI] [PubMed] [Google Scholar]
- 22.Franken R, El Morabit A, de Waard V, Timmermans J, Scholte AJ, van den Berg MP, et al. Increased aortic tortuosity indicates a more severe aortic phenotype in adults with Marfan syndrome. Int J Cardiol. 2015;194:7–12. doi: 10.1016/j.ijcard.2015.05.072. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.