Hostname: page-component-586b7cd67f-t7czq Total loading time: 0 Render date: 2024-11-27T14:14:03.298Z Has data issue: false hasContentIssue false

Aortic valve function and aortic dimensions in obese and overweight patients with bicuspid aortic valve

Published online by Cambridge University Press:  08 October 2024

Andrew M. Reittinger*
Affiliation:
Department of Pediatric Cardiology, Virginia Commonwealth University, Richmond, VA, USA
Peter N. Dean
Affiliation:
Department of Pediatric Cardiology, University of Virginia, Charlottesville, VA, USA
Michael A. McCulloch
Affiliation:
Department of Pediatric Cardiology, University of Virginia, Charlottesville, VA, USA
Jeffrey Vergales
Affiliation:
Department of Pediatric Cardiology, University of Virginia, Charlottesville, VA, USA
*
Corresponding author: Andrew M. Reittinger; Email: [email protected]
Rights & Permissions [Opens in a new window]

Abstract

Introduction:

Bicuspid aortic valve is the most common CHD and commonly associated with activity restrictions that may lead to a sedentary lifestyle known to increase obesity risk. It is unknown whether obesity is associated with changes in aortic dimensions or aortic valve function in young people with bicuspid aortic valve. This study investigates whether overweight and obese children with bicuspid aortic valve have worse aortic valve function or increased aortic dimensions compared to healthy weight children with bicuspid aortic valve.

Methods:

This was a single centre retrospective cohort study comprised of patients 5 to 25 years old with a diagnosis of bicuspid aortic valve between 1 January, 2019 and 31 December, 2020. Patients were classified as healthy weight or overweight/obese. Values for aortic dimensions as well as peak and mean aortic valve gradients were obtained from echocardiogram reports.

Results:

About 251 patients were analysed. Demographics were similar between groups. When indexed to height, the aortic valve annulus (1.28 ± 0.14 vs. 1.34 ± 0.15, p = 0.001) and sinotubular junctions (1.44 ± 0.21 vs. 1.49 ± 0.24, p = 0.038) were larger in the overweight/obese group, with no differences in aortic root or ascending aorta sizes. The obese/overweight group had a higher peak aortic valve gradient (23.03 ± 1.64 mmHg vs. 16.17 ± 1.55 mmHg, p = 0.003) compared to the healthy weight group.

Conclusion:

Healthy weight patients did not have larger aortic dimensions compared to the overweight/obese patients. There was evidence of worsening aortic valve stenosis in overweight/obese patients compared to those at a healthy weight.

Type
Original Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2024. Published by Cambridge University Press

Introduction

Bicuspid aortic valve is the most common CHD, present in about 1–2% of the population. Reference De Mozzi, Longo, Galanti and Maffulli1 These patients are at an increased risk for aortic stenosis, aortic insufficiency, aortic enlargement, aortic dissection, and endocarditis. The morphology of the bicuspid aortic valve, baseline aortic dimensions, or baseline aortic valve function are important predictors for adverse outcomes with valve dysfunction portending a more significant risk than isolated aorta dilation in young patients. Reference Niaz, Poterucha and Johnson2 Monitoring these parameters can help identify those potentially requiring closer follow-up, activity restrictions, or intervention.

It has been demonstrated that strictly adhering to the most recent sports eligibility guidelines by the American Heart Association and American College of Cardiology could lead to physical activity restriction in approximately a third of children and adolescents with bicuspid aortic valve. Reference Baleilevuka-Hart, Teng, Carson, Ravekes and Holmes3 The most common reason for activity restriction is aortic dilation, Reference Baleilevuka-Hart, Teng, Carson, Ravekes and Holmes3 although there is an evidence that physical activity is unrelated to changes in aortic dimensions in patients with bicuspid aortic valve. Reference Monda, Fusco and Della Corte4,Reference Stefani, Galanti, Innocenti, Mercuri and Maffulli5 While well-intentioned, activity restrictions can negatively impact overall health. Children with CHD restricted from activity have larger increases in absolute body mass index and body mass index percentile compared to those without restriction, and activity restriction was noted to be the strongest predictor of being overweight or obese at follow-up. Reference Stefan, Hopman and Smythe6 Nearly 40% of children with acyanotic CHD (including bicuspid aortic valve) become obese by 20 years of age. Reference Steele, Preminger and Erenberg7

Patients with bicuspid aortic valve are at risk for undesirable changes in aortic valve function and aortic dimensions based on several non-modifiable risk factors. In adults, it is known that obesity alone is linked with higher degrees of aortic stenosis and aortic dilation. Reference Kaltoft, Langsted and Nordestgaard8 If this process starts in childhood or adolescence, then body mass index could be a modifiable risk factor in reducing the risk of worsening aortic valve disease. The present study seeks to determine whether overweight/obese children with bicuspid aortic valve have worse aortic valve function or increased aortic dimensions compared to healthy weight children with bicuspid aortic valve.

Materials and methods

Study design

This was a single centre retrospective cohort study comprised of patients aged 5 to 25 years with a diagnosis of bicuspid aortic valve evaluated at the University of Virginia between 1 January, 2019 and 31 December, 2020. Patients were classified as healthy weight, overweight, or obese by body mass index based on definitions from the Centers for Disease Control and Prevention. 9,10 Patients classified as underweight were excluded from the study population. Patients with additional unrepaired CHD, recurrent coarctation of the aorta, or history of aortic valve, aortic root, or ascending aorta intervention or replacement were excluded. Additional unrepaired CHD was defined as including cyanotic CHD, conotruncal defects, left or right-sided obstructive defects, or ventricular septal defects of any type. Recurrent coarctation of the aorta was defined as a patient who has had previous transcatheter or surgical intervention for coarctation of the aorta who at the time of their echocardiogram had a peak gradient in the descending aorta of 20 mmHg or greater or a diastolic drag pattern by Doppler interrogation. Body surface area was calculated using the Haycock method. Reference Sluysmans and Colan11

Two-dimensional echocardiographic measurements of the aortic valve annulus, aortic root, sinotubular junction, and ascending aorta were taken directly from reports when available or otherwise measured post hoc from standard echocardiogram views according to the Pediatric Council of the American Society of Echocardiography. Reference Lai, Geva and Shirali12 Aortic dimension Z-scores were standardised appropriately to body surface area. Reference Colan13,Reference Sluysmans and Colan11 Aortic dimensions were indexed to patient height alone by dividing the aortic dimension of interest, in centimetres, by the patient’s height, in centimetres. This value was multiplied by a factor of 100 for ease of graphical and tabular representation. Values of aortic valve peak and mean gradients were taken directly from echocardiogram reports.

Statistics

Groups compared for statistical analysis were those with a healthy body mass index versus those with a body mass index qualifying them as overweight or obese. Categorical variables were compared using the chi-square test while means of continuous variables were compared using the non-parametric Mann–Whitney test. Analysis of variance was used for comparisons of means of continuous variables between multiple groups. Paired testing analysis was performed to link the appropriate echocardiographic measurements with heart rate and blood pressure measurements during the study. Regression models using restricted cubic splines with knots at 7.5, 12.5, 17.5, and 22.5 years old were used to compare the change in different outcomes of interest with age between the two study groups. Six separate knot sets were used with Bayesian model averaging based on 1,000 permutations.

A random sample of 20% of the echocardiograms were reviewed in a blinded fashion to assess for interobserver reliability to using reports. For categorical variables including degree of aortic stenosis and regurgitation more than mild, a kappa statistic was utilised and demonstrated to be 0.81 suggesting good agreement. For continuous variables, there was less than 10% variation from reported measurements indicating acceptable variation.

Results

There were 315 patients aged 5–25 years old with a diagnosis of bicuspid aortic valve evaluated with a transthoracic echocardiogram between 1 January, 2019 and 12 December, 2020. Fourteen patients were excluded due to being underweight based on Centers for Disease Control and Prevention criteria. Thirty-three patients were excluded for the presence of additional unrepaired CHD, and 17 were excluded for having undergone aortic valve, aortic root, or ascending aorta replacement or intervention. After exclusions, 251 patients remained for analysis. Patient demographics and clinical characteristics are presented in Table 1. There were no significant differences in age, sex, height, or bicuspid aortic valve morphology between the groups. The weights and body mass indexes were different between groups, as expected.

Table 1. Comparison of baseline characteristics between the healthy weight and overweight and obese study groups. Values are reported as mean with Standard Deviation (SD) and categorical numbers with percent total (%)

All aortic dimension sizes were reported in centimetres. There were no significant differences between the healthy weight group versus the overweight and obese group in terms of absolute aortic root size (2.75 ± 0.52 vs. 2.85 ± 0.54, p = 0.122), sinotubular junction size (2.26 ± 0.46 vs. 2.38 ± 0.52, p = 0.057), or ascending aorta size (2.83 ± 0.65 vs. 2.94 ± 0.70, p = 0.204). Individuals in the normal weight group had smaller aortic valve annulus sizes compared to the overweight and obese group (2.01 ± 0.36 vs. 2.13 ± 0.40, p = 0.014). The aortic dimension Z-scores were all larger in the healthy weight group compared to the overweight and obese group. When aortic dimensions were indexed just to patient height to mitigate the effect of being overweight on the Z-value, the aortic valve to height (1.28 ± 0.14 vs. 1.34 ± 0.15, p = 0.001) and sinotubular junction to height (1.44 ± 0.21 vs. 1.49 ± 0.24, p = 0.038) ratios were statistically significantly larger in the overweight and obese group. These findings are shown in Figure 1.

Figure 1. Difference in means plot with 95% confidence intervals comparing aortic dimensions, Z-scores, and aortic dimensions normed by height of the healthy weight group and the overweight and obese group.

The degree of aortic stenosis and regurgitation were qualitatively graded as none, trivial, mild, moderate, or severe, and represented numerically as 0, 1, 2, 3, or 4, respectively. The overweight and obese group had a statistically significant increase in aortic stenosis grade (1.22 ± 0.10 vs. 0.73 ± 0.10, p = 0.001), but not aortic regurgitation grade (1.23 ± 0.10 vs. 1.27 ± 0.09, p = 0.756) compared to the healthy weight group (Table 2).

Table 2. Aortic valve peak and mean gradients of the healthy weight group compared to the overweight and obese group. Aortic valve gradients expressed in millimetres of mercury (mmHg) with Standard Error (SE)

There was an increase in aortic valve peak gradient, measured in millimetres of mercury (mmHg), in the overweight and obese group (23.03 ± 1.64 vs. 16.17 ± 1.55, p = 0.003). The mean gradient across the aortic valve was similar between the overweight and obese group and the healthy weight group (16.87 ± 1.56 vs. 12.92 ± 1.73, p = 0.092) as shown in Table 2.

A regression model (Figure 2) was made to compare the changes in aortic dimensions and aortic valve function over a variety of ages between the two groups, with years as a continuous variable. There was no significant difference between the groups for the absolute measurements of aortic dimensions. The data were then analysed by Z-score. The overweight and obese group had significantly lower Z-scores at different ages in all dimensions. Lastly, the aortic dimensions were indexed to height. The aortic valve annulus was significantly larger at different ages in the overweight and obese group with a p value of 0.042. There were no significant differences for the other aortic dimensions when indexed to patient height. At different ages, there was not a divergence in aortic dimensions or aortic dimensions indexed to height except for the aortic valve annulus indexed to height, appearing to start around adolescent ages.

Figure 2. Regression models of aortic dimensions ( a ) and aortic valve gradient ( b ) as a function of age in years using restricted cubic splines method with knots at 7.5, 12.5, 17.5, and 22.5 years old. Six separate knot sets were used with Bayesian model averaging based on 1,000 permutations.

The regression model showed a higher aortic valve peak gradient in the overweight and obese group compared to the healthy weight group, while there was no statistically significant difference in the mean gradient.

Discussion

The present study demonstrates that obese and overweight patients with bicuspid aortic valve have worse aortic stenosis but similar aortic sizes compared to healthy weight individuals with bicuspid aortic valve. These results suggest that weight is a potentially modifiable risk factor for aortic valve disease in childhood. The data support providers encouraging patients to maintain a healthy weight with well-known interventions such as increased physical activity and eating a healthy diet.

There has been investigation into whether physical activity worsens aortic dimensions in patients with bicuspid aortic valve. It has been shown that there is no difference in progression of aortic diameters between children with bicuspid aortic valve who were physically active compared to those that were not. Reference Monda, Fusco and Della Corte4 Even competitive athletes with bicuspid aortic valve were followed over time and there was no difference in aortic dimensions compared to athletes with a tricuspid aortic valve. Reference Galanti, Stefani, Toncelli, Vono, Mercuri and Maffulli14 Despite this, Baleilevuka-Hart et al. estimated that a third of patients with bicuspid aortic valve are restricted from activity in some form during school age and young adult years based on 2015 guidelines. Reference Baleilevuka-Hart, Teng, Carson, Ravekes and Holmes3 Obesity already disproportionately effects children with bicuspid aortic valve. The prevalence of obesity in children with bicuspid aortic valve in the state of Oregon was 20% compared to 11.7% in the state’s general paediatric population. Reference Baleilevuka-Hart, Khader, De Alba, Holmes and Holmes15 Our results show that there is no significant difference in aortic root or ascending aorta dimensions of those who are of a healthy weight compared to those that are overweight or obese, and that the overweight and obese group had larger aortic valve annuli and ST-junctions when indexed to patient height. Additionally, cumulative physical activity over time may not be a factor in aortic dimensions based on the lack of aortic dimension variability with age in this study. This study’s findings of higher degrees of aortic stenosis in obese and overweight patients with bicuspid aortic valve combined with the lack of a difference in aortic dimensions between the groups suggests that greater emphasis should be placed on weight management and the encouragement of physical activity. While this requires generalisation and some assumptions, it certainly raises questions about how patients and families should be counselled regarding health maintenance and preventative health, and whether activity restrictions in patients with bicuspid aortic valve should be less stringent to encourage an active lifestyle. Often in CHD, we do not have modifiable risk factors of which we can empower the patients and their families to take active ownership, but this study indicates that weight control may be one of them and is likely important not just for patients with bicuspid aortic valve but for CHD as a whole.

Of note, the regression model of aortic valve gradient at different ages (Figure 2, panel b) in this study shows an inflection point in the obese and overweight group at around 15 years of age, after which the degree of stenosis seems to decline while the healthy weight curve inclines, with the curves crossing at around 20 years old. This unexpected finding is not in keeping with the reported natural history of aortic stenosis in patients with bicuspid aortic valve, which suggests stenosis does not improve with time and may be progressive. Reference Niaz, Fernandes, Sanders, Michelena and Hagler16 This is likely attributable to heteroscedasticity of the data and could also be driven by outliers in the groups as overall patient numbers in the older age ranges of this study population begin to decline compared to the younger age groups.

While initial comparisons of aortic dimensions were made between groups using Z-score data based on body surface area, it quickly became apparent that using Z-scores in subjects at the extreme ends of the weight spectrum gave results that were not consistent with the clinical picture. There has been considerable research into how to best index echocardiographic measurements in children so that they can be compared to normative values. It has been well-described that weight-dependent Z-scores can underestimate relative sizes of structures in echocardiography. Gutgesell et al. showed that linear dimensions have a non-linear relationship with body surface area, suggesting that linear measurements such as aortic dimensions should be indexed to other two dimensional, linear measurements like a patient’s height. Reference Gutgesell and Rembold17 It has also been shown that Z-scores become increasingly divergent at extremes of body mass index and their relationship to aortic dimensions become non-linear. The same study noted that Z-scores were found to underestimate aortic sizes in obese children, as they presume a child should have much larger aortic dimensions than an identical child at a normal weight. Reference Braley, Tang, Makil, Borroughs-Ray and Collins18

Because there is a higher incidence of obesity in patients with bicuspid aortic valve, the degree of aortic dilation may be underestimated in a larger proportion of individuals. Dallaire et al. found that as children approached adolescence, height was a more significant influence on aortic diameter than weight. Reference Dallaire, Bigras, Prsa and Dahdah19 We chose to index to height as it has been noted to correlate more with aortic dimensions in the overweight and obese populations. Reference Zafar, Li and Rizzo20 Notably, there were no significant differences in height between the two study groups.

The decision of how to group the study patients deserves comment, as one might suppose to find further differences between obese and overweight patients. However, the group of obese patients was relatively small and the results between overweight and obese patients were similar when these groups were split. Based on this, it was felt that combining obese and overweight groups and making a direct comparison between two study groups would lend improved conceptual and statistical power to the study.

To address any concerns about concomitant coarctation increasing risk for aortic valve dysfunction or aortic dilatation, a secondary analysis without those patients was performed, and all analyses yielded similar p values.

Limitations

This is a single-centre study, so this study cohort and these results may not be applicable to all populations. As with most retrospective studies, there is the possibility of some degree of selection bias and information bias. Risk factors that were comorbid with obesity, such as hypertension or diabetes, were not collected as this was felt to be outside the scope of the current study and unlikely to impact the majority of the population in question based on age. The lack of longitudinal follow-up of these patients also would have made it difficult to understand the impact of comorbid conditions. While it was not the aim of the present study, tracking subjects longitudinally to follow their change in weight, aortic sizes, and indices of aortic valve function over time is a future direction for continued investigation that would likely add valuable information to these initial findings. Not all echocardiographic measurements were made by the same interpreter, so there is potential for a degree of interobserver variability. However, studies have shown good intra- and interobserver reliability in measuring aortic root and ascending aorta dimensions, Reference Servato, Teixidó-Turá and Sabate-Rotes21 and a random sampling of 20% of the echocardiographic parameters measured in this study showed good interobserver agreement with a kappa statistic of 0.81 for degree of aortic stenosis and regurgitation more than mild, as well as less than 10% variation in reported measurements of continuous variables. An additional limitation is that exact levels of physical activity were not known. While physical activity or sports participation was frequently noted in cardiologists’ documentation, it was not done so in a standardised or consistent fashion and, as such, not felt to be reliable enough for inclusion.

Conclusions

We found that overweight and obese patients with bicuspid aortic valve had worse aortic valve stenosis compared to those at a healthy weight. Healthy weight patients did not have larger aortic dimensions compared to the overweight and obese patients. Consequently, being overweight or obese did not protect against progressive aortic dimensions and may contribute to an accelerated worsening of aortic valve health.

Acknowledgements

None.

Financial support

This research received no specific grant from any funding agency, commercial, or not-for-profit sectors.

Competing interests

None.

References

De Mozzi, P, Longo, UG, Galanti, G, Maffulli, N. Bicuspid aortic valve: a literature review and its impact on sport activity. Br Med Bull 2008; 85: 6385. DOI: 10.1093/bmb/ldn002.CrossRefGoogle ScholarPubMed
Niaz, T, Poterucha, JT, Johnson, JN et al. Incidence, morphology, and progression of bicuspid aortic valve in pediatric and young adult subjects with coexisting congenital heart defects. Congenit Heart Dis 2017; 12: 261269. DOI: 10.1111/chd.12429.CrossRefGoogle Scholar
Baleilevuka-Hart, M, Teng, BJ, Carson, KA, Ravekes, WJ, Holmes, KW. Sports participation and exercise restriction in children with isolated bicuspid aortic valve. Am J Cardiol 2020; 125: 16731677. DOI: 10.1016/j.amjcard.2020.02.039.CrossRefGoogle ScholarPubMed
Monda, E, Fusco, A, Della Corte, A et al. Impact of regular physical activity on aortic diameter progression in paediatric patients with bicuspid aortic valve. Pediatr Cardiol 2021; 42: 11331140. DOI: 10.1007/s00246-021-02591-4.CrossRefGoogle ScholarPubMed
Stefani, L, Galanti, G, Innocenti, G, Mercuri, R, Maffulli, N. Exercise training in athletes with bicuspid aortic valve does not result in increased dimensions and impaired performance of the left ventricle. Cardiol Res Pract 2014; 2014: 18. DOI: 10.1155/2014/238694.CrossRefGoogle Scholar
Stefan, MA, Hopman, WM, Smythe, JF. Effect of activity restriction owing to heart disease on obesity. Arch Pediatr Adolesc Med 2005; 159: 477. DOI: 10.1001/archpedi.159.5.477.CrossRefGoogle ScholarPubMed
Steele, JM, Preminger, TJ, Erenberg, FG et al. Obesity trends in children, adolescents, and young adults with congenital heart disease. Congenit Heart Dis 2019; 14: 517524. DOI: 10.1111/chd.12754.CrossRefGoogle Scholar
Kaltoft, M, Langsted, A, Nordestgaard, BG. Obesity as a Causal risk factor for aortic valve stenosis. J Am Coll Cardiol 2020; 75: 163176. DOI: 10.1016/j.jacc.2019.10.050.CrossRefGoogle ScholarPubMed
CDC, Childhood overweight & obesity, Accessed January 31, 2023. https://www.cdc.gov/obesity/childhood/index.html.Google Scholar
Sluysmans, T, Colan, SD. Structural Measurements and Adjustments for Growth. Echocardiography in Pediatric and Congenital Heart Disease. John Wiley & Sons, Ltd, New York, NY, 2016.Google Scholar
Lai, WW, Geva, T, Shirali, GS et al. Guidelines and standards for performance of a pediatric echocardiogram: a report from the Task Force of the Pediatric Council of the American Society of Echocardiography. J Am Soc Echocardiogr 2006; 19: 14131430. DOI: 10.1016/j.echo.2006.09.001.CrossRefGoogle ScholarPubMed
Colan, SD. Normal Echocardiographic Values for Cardiovascular Structures. Echocardiography in Pediatric and Congenital Heart Disease. John Wiley & Sons, Ltd, New York, NY, 2016.Google Scholar
Galanti, G, Stefani, L, Toncelli, L, Vono, MCR, Mercuri, R, Maffulli, N. Effects of sports activity in athletes with bicuspid aortic valve and mild aortic regurgitation. Br J Sports Med 2010; 44: 275279. DOI: 10.1136/bjsm.2008.047407.CrossRefGoogle ScholarPubMed
Baleilevuka-Hart, M, Khader, A, De Alba, CG, Holmes, K, Holmes, K. Intersection of sports participation and obesity in bicuspid aortic valve.J Am Coll Cardiol 2021; 77: 3198. DOI: 10.1016/s0735-1097(21)04553-8.CrossRefGoogle Scholar
Niaz, T, Fernandes, SM, Sanders, SP, Michelena, H, Hagler, DJ. Clinical history and management of bicuspid aortic valve in children and adolescents. Prog Cardiovasc Dis 2020; 63: 425433. DOI: 10.1016/j.pcad.2020.05.012.CrossRefGoogle ScholarPubMed
Gutgesell, HP, Rembold, CM. Growth of the human heart relative to body surface area. Am J Cardiol 1990; 65: 662668. DOI: 10.1016/0002-9149(90)91048-b.CrossRefGoogle ScholarPubMed
Braley, KT, Tang, X, Makil, ES, Borroughs-Ray, D, Collins, RT. The impact of body weight on the diagnosis of aortic dilation-misdiagnosis in overweight and underweight groups. Echocardiography 2017; 34: 10291034. DOI: 10.1111/echo.13565.CrossRefGoogle ScholarPubMed
Dallaire, F, Bigras, JL, Prsa, M, Dahdah, N. Bias related to body mass index in pediatric echocardiographic Z scores. Pediatr Cardiol 2015; 36: 667676. DOI: 10.1007/s00246-014-1063-7.CrossRefGoogle ScholarPubMed
Zafar, MA, Li, Y, Rizzo, JA et al. Height alone, rather than body surface area, suffices for risk estimation in ascending aortic aneurysm. J Thorac Cardiovasc Surg 2018; 155 : 19381950. DOI: 10.1016/j.jtcvs.2017.10.140.CrossRefGoogle ScholarPubMed
Servato, ML, Teixidó-Turá, G, Sabate-Rotes, A et al. Are aortic root and ascending aorta diameters measured by the pediatric versus the Adult American Society of Echocardiography Guidelines Interchangeable? J Clin Med 2021; 10: 5290. DOI: 10.3390/jcm10225290.CrossRefGoogle ScholarPubMed
Figure 0

Table 1. Comparison of baseline characteristics between the healthy weight and overweight and obese study groups. Values are reported as mean with Standard Deviation (SD) and categorical numbers with percent total (%)

Figure 1

Figure 1. Difference in means plot with 95% confidence intervals comparing aortic dimensions, Z-scores, and aortic dimensions normed by height of the healthy weight group and the overweight and obese group.

Figure 2

Table 2. Aortic valve peak and mean gradients of the healthy weight group compared to the overweight and obese group. Aortic valve gradients expressed in millimetres of mercury (mmHg) with Standard Error (SE)

Figure 3

Figure 2. Regression models of aortic dimensions (a) and aortic valve gradient (b) as a function of age in years using restricted cubic splines method with knots at 7.5, 12.5, 17.5, and 22.5 years old. Six separate knot sets were used with Bayesian model averaging based on 1,000 permutations.