Extracardiac Manifestations Fail to Predict the Severity of Cardiac Phenotype in Children and Young Adults with Marfan Syndrome

We performed a secondary analysis of the Pediatric Heart Network Marfan Trial public-use database to evaluate associations between extracardiac features and cardiac and aortic phenotypes in study participants. Aortic aneurysm phenotype was defined as aortic root Z-score ≥ 4.5, aortic root growth rate ≥ 75th percentile, aortic dissection, and aortic surgery. Severe cardiac phenotype was defined as aortic dissection, aortic Z-score ≥4.5, aortic valve surgery, at least moderate mitral regurgitation, mitral valve surgery, left ventricular dysfunction, or death. Extracardiac manifestations were characterized by specific organ system involvement and by a novel aggregate extracardiac score that was created for this study based on the original Ghent nosology. Logistic regression analysis compared aggregate extracardiac score and systems involvement to outcomes. Of 608 participants (60% male), the median age at enrollment was 10.8 years (interquartile range: 6, 15.4). Aortic aneurysm phenotype was observed in 71% of participants and 64% had severe cardiac phenotype. On univariate analysis, skeletal (OR: 1.95, 95% CI: 1.01, 3.72; p = 0.05), skin manifestation (OR: 1.62, 95% CI: 1.13, 2.34; p = 0.01) and aggregate extracardiac score (OR: 1.17, 95% CI: 1.02, 1.34; p = 0.02) were associated with aortic aneurysm phenotype but were not significant in multivariate analysis. There was no association between extracardiac manifestations and severe cardiac phenotype. Thus, the severity of cardiac manifestations in Marfan syndrome was independent of extracardiac phenotype and aggregate extracardiac score. Severity of extracardiac involvement did not appear to be a useful clinical marker for cardiovascular risk-stratification in this cohort of children and young adults with Marfan syndrome.


Introduction
Marfan syndrome (MFS) is a heritable systemic connective tissue disorder occurring in 1 in 5,000 individuals.Caused by mutations in the FBN1 gene, MFS has multiorgan involvement affecting primarily the ocular, musculoskeletal and cardiovascular systems (1)(2)(3)(4).Cardiovascular manifestations, including progressive aortic enlargement, aortic dissection and rupture, mitral regurgitation requiring surgical intervention, and left ventricular dysfunction, are associated with signi cant morbidity and mortality (5,6).Risk for aortic dissection and aortic surgery increases with increasing aortic root diameter, and thresholds for medical and surgical therapy are largely dependent on aortic root size.However, aortic root size is not a perfect predictor.For example, prophylactic aortic root replacement is generally recommended at an aortic root diameter of 5 cm, yet 15% of cases of aortic dissection in Marfan syndrome occur at aortic diameters < 5 cm (7,8).Although cardiac and extracardiac phenotypes have been well-characterized, the relationship between extracardiac manifestations and cardiovascular outcomes remain unclear.If extracardiac phenotypes positively correlate with the severity of cardiovascular involvement, the association may be useful for risk-strati cation and clinical management.Therefore, the primary objective of this study was to evaluate the association between extracardiac features and severe aortic and cardiac phenotypes in children, adolescents, and young adults with MFS utilizing the large multicenter Pediatric Heart Network Marfan (PHN) Trial public-use dataset.

Methods
Study design: We performed a secondary analysis of the PHN Marfan Trial public-use database.The PHN Marfan Trial was a randomized clinical trial comparing the effects of beta blocker (atenolol) therapy versus angiotensin II receptor blocker (losartan) therapy on the rate of aortic root enlargement in children and young adults with MFS.A detailed description of trial design and results has previously been reported (9,10).Brie y, between January 2007 and February 2011, 608 participants, aged 6 months to 25 years, diagnosed with MFS using the original Ghent criteria (11) who had an aortic root Z-score >3 and aortic root diameter <5 cm were enrolled in the trial.Individuals with previous aortic surgery, planned aortic surgery within 6 months of enrollment, or history of aortic dissection were excluded.The study enrollment forms collected detailed data regarding extracardiac features, organ system involvement, and Ghent nosology criteria for all participants at ve time periods (baseline, 6 months, 12 months, 24 months and 36 months).The protocol was approved by the Institutional Review or Ethics Board at each participating site.
Outcomes and de nitions: Primary outcomes for this analysis included aortic aneurysm phenotype and severe cardiac phenotype.
Aortic aneurysm phenotype was de ned as having an aortic root Z-score ³4.5, aortic annual growth rate of ³75 th percentile based on change in aortic root Z-score during the trial, aortic surgery, or aortic dissection.
Severe cardiac phenotype was de ned as aortic dissection, aortic Z-score ³4.5, aortic valve surgery, at least moderate mitral regurgitation, mitral valve surgery at the time of trial enrollment, left ventricular dysfunction (ejection fraction £45%), or death.
Extracardiac systems included musculoskeletal, ocular, dural, pulmonary, and skin manifestations.Based on the original Ghent nosology, a novel aggregate extracardiac score (AES) was created for this study to evaluate multiorgan involvement.Extracardiac manifestations meeting majorcriteria were assigned two points and those ndings meeting minor criteria were assigned one point.AES was derived by adding the number of qualifying minor and major criteria for each participant (Table 1

Results
Participant characteristics: All 608 participants in the PHN Marfan trial who met the original diagnostic Ghent criteria were included in this secondary analysis.Median age at enrollment was 10.8 years (IQR 6 to15.4 years) and 366 participants (60%) were males.
Cardiovascular phenotypes: The prevalence of cardiovascular outcomes among PHN Marfan trial participants are summarized in Figure 1.
Extracardiac Phenotype: The prevalence of major and minor criteria for each extracardiac system and the aggregate extracardiac score are shown in Table 2

Discussion
In Marfan syndrome, the risk for aortic dissection and aortic surgery increases with increasing aortic root diameter.Unfortunately, absolute aortic root size is not a great predictor for aortic dissection and/or rupture (7,8).Our analysis used the data from one of the rst large multicenter studies of a well characterized MFS cohort to evaluate the relationship between extracardiac manifestations and aggregate extracardiac score and severity of cardiovascular system involvement.Our key nding was absence of an association between speci c organ system involvement or aggregate extracardiac score and cardiovascular outcomes in MFS.Although skeletal manifestations, skin manifestations and aggregate extracardiac score were associated with aortic aneurysm phenotype in univariate analysis, these associations lost signi cance in multivariate models.
Accurate risk-strati cation and timely interventions can save lives but evaluating the risk of aortic dissection or rupture in the population with MFS remains challenging.Although it is compelling to think that the severity of the more visible manifestations of MFS may correlate with the severity of the cardiac ndings, few studies have assessed the association of extracardiac ndings with cardiovascular phenotype (15).Previously reported risk factors for aortic dissection and severe cardiac phenotype include prior prophylactic aortic surgery, aortic diameter, and pregnancy ( 16).An increased prevalence, but not severity, of mitral valve prolapse was reported in individuals with MFS and kyphoscoliosis (17).The Hungarian Marfan Registry (HMR) study found aortic dissection was more common in participants with striae atrophicae.Although we found an association between skin manifestation and severe aortic phenotype on univariate analysis, this association lost its signi cance in multivariate modeling.One explanation for this discrepancy may be the older age of HMR participants (33.2 ±13.5 years) compared to the PHN Marfan trial participants (median age 10.8 years at enrollment).The weighting of major and minor criteria for calculating the aggregate extracardiac score was somewhat arbitrary, based on the original Ghent criteria.It is possible that a different weighting system might uncover different associations.Regardless, the aggregate extracardiac score did not independently predict a severe aortic or cardiac phenotype, as de ned in our study.

Limitations
The entry criteria for the PHN Marfan Trial excluded children and young adults with Marfan syndrome at the mild (aortic root z-score < 3) and severe (those who already had aortic surgery or aortic dissection) ends of the spectrum.Although this cohort was well characterized both for extracardiac and cardiac phenotypes, the follow-up period was only 3 years.It is possible that associations between extracardiac phenotype and cardiovascular outcomes such as aortic surgery and aortic dissection might be found after a longer period of follow-up.Aggregate extracardiac score derived from the original Ghent nosology used in this study has not been previously validated.Though the large multicenter PHN Marfan Trial public-use database provides a large patient cohort, the phenotypic features of the original Ghent nosology have since been revised to put more emphasis on aortic root dilation and ectopia lentis, with ndings in other organ systems contributing to a systemic score (14).The systemic score was not evaluated in this study.In addition, we had no molecular data and could not explore the prior reports of an association of elevated circulating TGF-ß levels or increased expression of MMP-3 with aortic dissection (18).

Conclusion
Despite great advances in care, cardiovascular risk strati cation in MFS remains challenging.Our secondary analysis of the PHN Marfan Trial public-use database showed that extracardiac phenotype failed to predict the severity of cardiovascular involvement (as de ned by our study), and thus, has limited value for risk-strati cation or management.Therefore, all individuals with MFS, regardless of their extracardiac phenotype, should be followed closely with serial surveillance imaging throughout their lifetime.Aortic root size, while not perfect, remains the best predictor of major aortic complications in MFS.Future studies may focus on the association between the revised Ghent systemic score, molecular and genetic data, and severity of cardiovascular manifestations to better risk stratify MFS patients for cardiac complications.

Figures
Figures

Figure 1 Cardiovascular
Figure 1 (13)emographics and clinical outcomes of interest were summarized using median and interquartile range (IQR) for continuous variables; counts and percentages were used for categorical variables.Univariate mixed effect logistic regression models were created, accounting for center-to-center variabilities, to evaluate the association between risk factors and outcomes (aortic aneurysm phenotype and severe cardiac phenotype).Multivariate mixed effects logistic regressions included risk factors that were signi cant at level 0.10 from univariate analysis including baseline aortic root elastic modulus, and baseline aortic root stiffness index for severe cardiac phenotype outcome.For aortic aneurysm phenotype, the multivariable model included baseline age at echocardiography, family and genetic history of MFS, family history of aortic dissection, treatment (losartan, atenolol), baseline aortic root elastic modulus, baseline aortic root stiffness index, skeletal, skin manifestation, aggregate extracardiac score (AES).Odds ratio and 95% con dence interval (CI) were reported.Statistical signi cance was assessed at the 0.05 level, unless speci ed otherwise as above for variable selection.Statistical analyses were implemented using R v. 4.1.2(13).
participants with available ocular data, 252 met major ocular criteria (ectopia lentis).MRI data was available for 92 participants of which 32 had lumbosacral dural ectasia.Pulmonary involvement had the lowest prevalence among all extracardiac systems included in Marfan Syndrome criteria.Only 11% of participants had severe extracardiac involvement as indicated by an aggregate extracardiac score ≥5.