We retrospectively analyzed data from all consecutive patients younger than 20 years who underwent successful cardiac catheterization with stent implantation for native or recurrent CoA.

The derivation cohort originated from the German Heart Institute of the Charité, and the independent validation cohort from the German Heart Center Munich.

Inclusion criteria were a diagnosis of CoA as the primary cardiac defect and either implantation of a stent or redilation of a previously implanted stent, with a resulting reduction of the pressure gradient across the lesion to≤10mmHg. Patients were excluded if they had complex congenital heart disease, such as double aortic arch or hypoplastic left heart syndrome. Additional exclusion criteria included stenting of the transverse aortic arch, prior Norwood operation, and a postinterventional gradient >10mmHg. In both cohorts, no patients with Turner syndrome or Down syndrome were included.

For validation purposes, the validation cohort included not only patients with isolated CoA but also those with complex CoA, including patients after a Norwood procedure. Patients with transverse arch stenting and those with a postinterventional gradient>10mmHg were excluded from the validation cohort.

Indications for cardiac catheterization consisted of evidence of CoA on clinical examination, echocardiography, computed tomography, or cardiac magnetic resonance imaging. In cases of significant CoA, defined as an aortic narrowing>50% on angiography accompanied by hypertension in at least the upper body, stent implantation was performed even if the invasive peak-to-peak systolic pressure gradient between the ascending and descending aorta was <20mmHg.

All procedures were performed under premedication with midazolam and intravenous sedation using propofol 1% or ketamine, with spontaneous breathing. Heparin (100 U/kg) was administered after vascular access was obtained. Local anesthesia was applied before femoral artery cannulation using the Seldinger technique. Invasive pressure measurements across the stenosis were obtained before and after stent implantation. Following angiography, an appropriate stent was selected. Balloon size for stent dilation was chosen to approximate the median diameter of the transverse aortic arch and the diameter of the descending aorta at the level of the diaphragm. Final angiography was performed to confirm appropriate stent positioning. In cases of severe stenosis, dilation to the target diameter was performed in 2 staged procedures, with the second intervention scheduled 3 to 6 months later.

Because height varies considerably among children of the same age, normalized percentile values were used for comparative analyses. Reference data were derived from the KiGGS study (Studie zur Gesundheit von Kindern und Jugendlichen in Deutschland) of the Robert Koch Institute, which is commonly used in Germany.14 In the first part of the study, the population was divided into 8 percentile groups based on these normalized values (table 1).

A linear association between vascular diameter and somatic growth has previously been demonstrated by Kaiser et al.,15 who proposed a formula to predict native aortic isthmus diameter based on the square root of body surface area (BSA) measured by contrast-enhanced cardiovascular magnetic resonance angiography. In their model (∅Isth:–3.37+16.52·BSA0.5), the strong correlation (r=0.90) supports the assumption that vessel dimensions scale proportionally with growth parameters.

Based on this principle, we investigated whether a comparable linear relationship exists between implanted stent diameter and anthropometric measures in patients undergoing endovascular treatment for CoA. Accordingly, correlation analyses were conducted between stent diameter and body height, weight, and BSA to determine whether linear prediction models could be derived.

Data analysis was performed using SPSS software, version 22.0 (SPSS Inc, USA). Continuous variables are presented as median and interquartile range (IQR).

Mathematical calculations were performed using Mathematica, version 10 (Wolfram Research, USA), and graphical representations were generated with OriginPro, version 8 (OriginLab Corporation, USA).

Comparisons between categorical variables in 2 groups were performed using the chi-square test. Comparisons between continuous variables were conducted using the independent Student t test, as appropriate. Univariate and multivariable regression analyses were performed to evaluate associations between body height, body weight, age, and stent diameter. Statistical significance was defined as P≤.05.

Kendall's tau and Pearson correlation coefficients were calculated to assess correlations between stent diameter and age, body height, and body weight. All tests were 2-sided, with statistical significance set at P=.01.

The study is a retrospective, multicenter investigational study. Data was acquired and analyzed in accordance with internationally accepted recommendations for clinical research, including the Declaration of Helsinki of the World Medical Association.

The study was approved by the ethics committee of the Technical University of Munich (TUM; reference number 2023-553_1-S-SB). Written informed consent was obtained from the legal guardians prior to the procedure.

As shown in table 2, the derivation cohort comprised 155 patients (54 female, 35%) treated with transcatheter stent implantation. A total of 218 cardiac catheterizations were undertaken. The median body weight was 32.2kg [IQR, 13.6-55kg], and the median height was 137cm [IQR, 95-160cm]. The median age was 10.1 years [IQR, 3.4-14.4 years].

The validation cohort consisted of 198 patients (64 female, 32%). A total of 323 cardiac catheterizations were undertaken. The median body weight was 26.0kg [IQR, 9.9-54.0kg], and the median height was 126cm [IQR, 77-161cm]. The median age was 7.5 years [IQR, 1.3-14.6 years].

The cohorts did not differ significantly in terms of age, weight, or height (P=.098, P=.63 and P=.23, respectively).

In all catheterizations, invasive blood pressure gradients were measured before and after stenting or dilation. The median peak invasive systolic pressure gradient before intervention was 25mmHg [IQR, 20-35mmHg] in the derivation cohort and 20mmHg [IQR, 14-30mmHg] in the validation cohort.

Procedural success, defined as a residual systolic pullback pressure gradient after stent implantation <10mmHg, was documented in all patients. The median gradient decreased significantly to 0mmHg [IQR, 0-5mmHg] in the derivation cohort (P <.001) and to 0mmHg [IQR, 0-4mmHg] in the validation cohort (P <.001).

In the derivation cohort, a total of 218 catheter-based procedures were analyzed, consisting of 161 primary stent implantations (74%) and 57 redilation procedures of previously implanted stents (26%). In the validation cohort, 323 catheter procedures were evaluated, comprising 219 primary implantations (68%) and 104 redilations (32%). The stents used are listed in table 3. The number of dilations as well as the number of interventions in each cohort are described in table 4.

In the derivation cohort, correlation analysis showed that body height showed the strongest correlation with stent diameter. The Kendall-correlation coefficient between height and stent diameter was τ: 0.737 (P≤.001), followed by body surface area (BSA, Mosteller formula) (τ: 0.734), body weight (τ: 0.732), and age (τ: 0.709). Furthermore, the correlation coefficient between height and age was τ: 0.836, and between weight and height was ττ: 0.865.

To derive a predictive formula and assess variability within subgroups, the derivation cohort was divided into 8 subgroups according to body height percentile groups from PG1 to PG8 and sex. Subgroup analysis, defined by each percentile group, demonstrated a linear correlation between stent diameter and body height (figure 1).

Figure 1.

Stent diameters are plotted to body height for the 8 percentile groups PG1-PG8. Male patients are shown in blue and female patients in red. Linear regression analysis was performed for each percentile group. The regression coefficients (a) and correlation coefficients (r) are provided for the corresponding percentile groups.

(0.29MB).

Based on the hypothesis that the required stent diameter could be determined using a formula f(x)=a·x+b, we used a linear regression analysis to develop the linear correlation between stent diameter (f(x)) and height in cm (x):

f(x)=0.0856·x+1.74 with r=0.877 and P≤.001 for the derivation cohort (figure 2A),

Figure 2.

A: derivation cohort: 218 cardiac catheterizations. The regression line f(x)=0.0856·x+1.74 illustrates the significant linear correlation between body height and stent diameter, r=0,877 and P <.001. B: validation cohort: 323 cardiac catheterizations. The regression line f(x)=0.0820·x+1.94 illustrates the significant linear correlation between body height and stent diameter, r=0.897 and P <.001. C: overall study cohort: 541 cardiac catheterizations. The regression line f(x)=0.0831·x+1.86 illustrates the significant linear correlation between body height and stent diameter, r=0.889 and P <.001.

(0.46MB).

f(x)=0.0820·x+1.94 with r=0.897 and P≤.001 for the validation cohort (figure 2B), and

f(x)=0.0831·x+1.86 with r=0.889 and P≤.001 for the overall cohort (figure 2C).

Consistent with the proportional growth relationship described by Kaiser et al.15, our analysis demonstrated a significant linear correlation between stent diameter and body height across both cohorts. We then adjusted body height and gender in f(x)=0.0831·x+1.86 to define the required stent diameter. Age-related optimal stent sizes are summarized in table 5 and figure 3 for female and male patients.

Figure 3.

Body height-based stent curves for male (blue) and female (red) patients. The green shaded areas indicate the corresponding age ranges for different stent diameters. Reference values are based on KiGGS (Studie zur Gesundheit von Kindern und Jugendlichen in Deutschland).

(0.38MB).

Since age and body height showed high correlation coefficients, age-based benchmarks to use a predefined stent diameters could be established. For example, a boy aged 3.5 years with a body height in the 25th percentile would require a 10-mm stent. However, a 10mm stent would also be appropriate for a boy in the 97th percentile group at 2.4 years of age.

Using these age-based benchmarks, it is also possible to estimate the number of additional cardiac catheterizations required to dilate or re-stent the aortic isthmus until adulthood, as illustrated in figure 3. For instance, a girl aged 5 years at the 50th percentile would require a 10mm stent. A subsequent dilation would then be necessary at approximately 9 years of age. Since the minimum age for a 14mm stent begins at 9 years, direct dilation to 14mm instead of 12mm would be feasible. Application of this formula therefore helps minimize the total number of interventions.

The main limitation of this study is its retrospective design. As a result, complete and standardized documentation of vascular complications was not available for all included patients. An additional limitation relates to the rapid somatic growth that occurs during the first 2 years of life. Moreover, the relatively small number of patients with a body height below 90cm, together with the median age of 10 years in the cohort, limits the accuracy of the formula in this subgroup.

An additional limitation of this study is the potential limited generalizability to other pediatric populations. Because somatic growth patterns in children vary among ethnicities, geographic regions, and health care systems, our findings (particularly those related to the timing and frequency of redilatations in relation to body height) should be interpreted with caution when applied to populations with different growth trajectories.

Another limitation of this study is that the derived formula is based on the stent sizes applied in our cohort, assuming that the optimal stent was selected and deployed at the ideal diameter in each case. Confirmation of these findings will therefore require prospective studies in independent cohorts. Such studies would provide a more controlled methodological framework and strengthen the predictive validity of the model in the context of CoA interventions in young patients

In addition, sex- and gender-related variables were not explicitly analyzed or reported in accordance with the SAGER guidelines; consequently, potential differences related to sex and gender could not be assessed and should be considered a limitation of this study