## Background

Aortic arch abnormalities represent 5% to 8% of all congenital heart disease. Measurements of the aortic arch dimensions on two-dimensional echocardiographic images remain of critical importance in the diagnosis of aortic arch pathology. To define aortic hypoplasia or coarctation, measured dimensions must be compared with normal values. Normal values have been described for children of all ages in earlier studies. However, normative data for premature infants are not yet available. Therefore, the aim of this study was to develop normative data in a cohort of premature infants, which could be used in the diagnosis of aortic arch abnormalities.

## Methods

A single-center study was conducted in a large population of premature infants with gestational ages of ≤32 weeks without hemodynamically important congenital heart disease, chromosomal abnormalities, and/or major cerebral congenital malformations. Two-dimensional echocardiographic measurements of four aortic arch structures were made on the second, fourth, and sixth days after birth.

## Results

Three hundred eighty-five preterm patients were included. No differences in dimensions were found among days 2, 4, and 6. The dimension of the isthmus showed no significant relation to the existence of a patent ductus arteriosus. Reference intervals with mean and SD were calculated across the range of birth weight. Regression analysis was performed with multiple determinants in different models. The best predictive value was found for birth weight in a cubic model.

## Conclusions

This work provides regression equations for the calculation of *Z *scores and reference intervals for aortic arch dimensions in a cohort of preterm infants born at gestational ages of ≤32 weeks. The normative data can be used in diagnosis and decision making involving aortic arch pathology in premature infants.

## Highlights

- •

A comprehensive set of normative data for the dimensions of the aortic arch is lacking for prematurely born infants.

- •

This study provides reference intervals for the dimensions of four aortic arch structures in premature infants across the range of birth weight.

- •

No significant difference in aortic arch dimensions was found among the second, fourth, and sixth days of life.

- •

PDA has no significant effect on the diameter of the aortic isthmus.

Aortic arch abnormalities represent 5% to 8% of all congenital heart disease. About 14% of all neonates with congenital heart disease are born prematurely. Decision making in the diagnosis and treatment of pediatric patients with aortic arch abnormalities most commonly relies on measurements of the aortic arch using two-dimensional echocardiography. Measurements of aortic dimensions are useful only when the obtained values can be compared with standard values for the patient’s body size. To decide whether a certain measurement falls in a normative range, *Z* scores and reference intervals are often used to normalize the measurement to the patient’s body size. So far, normative data have been collected for many patient groups. However, no normative data are available for extremely and very premature patients. This creates considerable difficulty in diagnosing aortic arch abnormalities in premature infants. Developing normative data for extremely and very premature children is of great importance in order to provide the best possible care to these children. However, obtaining applicable normative data for this special group of patients might be problematic, considering their very small body size and correlating aortic arch dimensions. The objective of this study was to develop normative data for the calculation of *Z* scores and reference intervals for premature aortic arch structures, which can be used for the diagnosis of aortic arch pathology.

## Methods

## Study Population

The study population consisted of 385 premature patients with gestational ages varying from 24.3 to 31.9 weeks. Data for this study were derived from the Ductus study cohort, a study in progress, designed to investigate the hemodynamic significance of patent ductus arteriosus (PDA). Data were prospectively collected in the neonatal intensive care unit of the Wilhelmina Children’s Hospital, a tertiary pediatric hospital in Utrecht, The Netherlands. Infants born alive between September 2008 and October 2010 with gestational ages of 24 weeks (the viability threshold in The Netherlands) to 32 weeks, birth weights (BWs) of ≤2,000 g, and structurally and functionally normal hearts or nonsignificant structural heart defects (e.g., very small muscular ventricular septal defects) on echocardiography were included. Exclusion criteria were the existence of hemodynamically important congenital heart disease on (repeat) echocardiography (i.e., tetralogy of Fallot, transposition of the great arteries, and coarctation of the aorta), chromosomal abnormalities, and major cerebral congenital malformations. A flowchart of inclusions and exclusions is presented in Figure 1 . Because of the design of the study and innate to the study population, the presence of a PDA or patent foramen ovale could not be controlled for in analysis and was considered physiologic.

## Echocardiographic Examination and Aortic Arch Measurements

Patients underwent two-dimensional transthoracic echocardiography 2, 4, and 6 days after birth ( *t *_{1 }, *t *_{2 }, and *t *_{3 }, respectively). The echocardiograms were obtained by a pediatric cardiologist or experienced cardiac sonographer according to a standardized protocol. All echocardiographic examinations were performed on a Vivid I ultrasound system with a 10- and/or 7-MHz transducer (GE Medical Systems, Wauwatosa, WI). The echocardiographic images were analyzed using EchoPAC version 112 (GE Medical Systems). Aortic arch diameters were measured by one investigator after training by a pediatric cardiologist, in accordance with the guidelines on quantification methods for pediatric echocardiograms of the American Society of Echocardiography. The peak systolic dimensions of the ascending, transverse, and descending aorta and the isthmus were measured from suprasternal views ( Figure 2 ). Because of suboptimal image quality, not all echocardiograms could be used for analysis in this study. Exclusions as a result of inadequate image quality accounted for 93%, 80%, and 52% of all exclusions at *t *_{1 }, *t *_{2 }, and *t *_{3 }, respectively.

## Interobserver Variability

Interobserver variability was examined to determine the reproducibility of measurements in offline analysis between researchers. A subset of 87 patients was selected at random by the supervising pediatric cardiologist and measured a second time. The second measurement by the pediatric cardiologist was performed on the original echocardiographic images, blinded to previous results, and compared with the first measurements.

## Statistical Analysis

A paired *t *test was applied to analyze differences in paired measurements ( *t *_{1 }, *t *_{2 }, and *t *_{3 }). Pearson correlation was applied to find a possible correlation between the independent variables. The differences between neonates with and without PDA were determined using an unpaired Student’s *t *test. The Spearman ρ statistic was applied to find a correlation between the diameter of the PDA and the dimension of the isthmus. Body surface area (BSA) (calculated according to the Haycock formula ), BW, birth height, and gestational age were used as independent variables in the regression analysis for the predicted mean value of the ascending, transverse, and descending aorta and the isthmus. BW was stratified into five groups: ≤1,000, 1,001 to 1,250, 1,251 to 1,500, 1,501 to 1,750, and ≥1,751 g. The correlations between BW classes and cardiac measurements were calculated using a Spearman rank analysis. Reference intervals are presented as mean ± SD for BW in their stratified groups. Regression analysis was performed in accordance with an earlier study on the calculation of *Z *scores of cardiac structures by Pettersen *et al *. Linear, logarithmic, quadratic, cubic, and logistic models were considered for regression analysis. The goodness of fit of the data to the regression model is described by the coefficient of determination ( *R *^{2 }). This statistic ranges from 0 to 1; a value of 1 represents a perfect fit, and a value of 0 represents no fit at all. BSA, BW, birth height, and gestational age were again used as independent variables for the calculation of reference intervals for the aortic arch dimensions at the four anatomic sites of measurement. An intraclass correlation coefficient was calculated to analyze interobserver variability. Data analyses were performed using SPSS version 22 (IBM, Armonk, NY). *P *values < .05 were considered to indicate statistical significance.

## Results

The mean gestational age was 29.5 ± 1.8 weeks (range, 24.3 31.9 weeks), the mean BW was 1,295 ± 357 g (range, 530–2,750 g), the mean birth height was 38 ± 3.0 cm (range, 26–47 cm), and mean the BSA was 0.12 ± 0.02 m ^{2 }(range, 0.06–0.17 m ^{2 }). The male/female ratio was evenly distributed (51% to 49%). Baseline characteristics of the study population are presented in Table 1 . At *t *_{1 }, 53% of the patients had a PDA, with 34% at *t *_{2 }and 25% at *t *_{3 }. No significant difference was found between the times of measurement ( *t *_{1 }, *t *_{2 }, and *t *_{3 }) and the dimensions of the measured aortic arch structures. No significant relation was found between the dimension of the isthmus and the existence of PDA at *t *_{1 }( *P *= .48), *t *_{2 }( *P *= .40), and *t *_{3 }( *P *= .21). A significant relation between the dimension of the isthmus and the diameter of the PDA was absent at *t *_{2 }( *P *= .11) but present at *t *_{1 }( *P *= .01) and *t *_{3 }( *P *= .01). The expressions of size with the highest correlations to aortic structure dimensions were BSA and BW for all four variables. BW measurement is part of standardized postpartum analysis, easily applicable and usually well documented. Considering optimal clinical applicability, BW was used as a predictive variable in the regression equations for the calculation of *Z *scores.

n |
Mean ± SD | Range | |
---|---|---|---|

GA (wk) | 387 | 29.5 ± 1.8 | 24.3–31.9 |

BW (g) | 384 | 1295 ± 357 | 530–2750 |

BH (m) | 347 | 0.38 ± 0.03 | 0.26–0.47 |

BSA (m ^{2 }) |
345 | 0.12 ± 0.02 | 0.06–0.19 |

Figure 3 shows the mean dimensions and 95% CIs for the ascending, transverse, and descending aorta and the isthmus at *t *_{1 }. All aortic measurements showed a highly significant correlation with BW ( *P *< .001 for the ascending, transverse, and descending aorta and *P *< .05 for the isthmus) at *t *_{1 }, *t *_{2 }, and *t *_{3 }. The strongest correlation was found for the ascending and the transverse aorta. Reference ranges for the four aortic structures, with mean and SD, according to body weight are shown in Table 2 .

BW (g) | n |
Valid n |
Ascending aorta | Valid n |
Transverse aorta | Valid n |
Isthmus | Valid n |
Descending aorta |
---|---|---|---|---|---|---|---|---|---|

<1,000 | 91 | 63 | 4.7 ± 0.6 | 59 | 3.7 ± 0.6 | 52 | 2.8 ± 0.7 | 44 | 3.8 ± 0.6 |

1,000–1,250 | 94 | 70 | 5.2 ± 0.8 | 66 | 3.9 ± 0.5 | 53 | 2.9 ± 0.7 | 37 | 3.8 ± 0.6 |

1,250–1,500 | 93 | 71 | 5.5 ± 0.6 | 64 | 4.1 ± 0.5 | 54 | 3.1 ± 0.9 | 40 | 4.1 ± 0.8 |

1,500–1,750 | 60 | 47 | 5.7 ± 0.7 | 44 | 4.3 ± 0.6 | 35 | 3.1 ± 0.8 | 31 | 4.3 ± 1.0 |

>1,750 | 47 | 30 | 6.3 ± 0.7 | 31 | 4.6 ± 0.8 | 23 | 3.6 ± 1.0 | 17 | 4.9 ± 1.2 |

In the absence of significant differences between the times of measurement and the highest number of measurements at *t *_{1 }, calculation of normal values was based on measurements made at *t *_{1 }. The relationships between BW and the four echocardiographic measurements are presented in Table 3 . The cubic model had the highest coefficient of determination for all dimensions and was therefore considered the best predictive model.

Intercept | b _{1 } |
b _{2 } |
b _{3 } |
Standard error | R ^{2 } | |
---|---|---|---|---|---|---|

Ascending aorta | 2.332 | 0.004 | −1.258-E6 | 2.111E-10 | 0.041 | 0.427 |

Transverse aorta | 3.910 | −0.001 | 1.087E-6 | −1.321E-10 | 0.057 | 0.305 |

Isthmus | 2.994 | −0.001 | 1.360E-6 | −2.8381E-10 | 0.079 | 0.109 |

Descending aorta | 4.131 | −0.002 | 2.799E-6 | −7.4607E-10 | 0.079 | 0.177 |

The relation between the aortic dimensions and BW is described in the following equation:

Aortic dimension = y = intercept + b 1 x + b 2 x 2 + b 3 x 3 .

Figure 4 presents the scattered plots of the measured aortic dimensions plotted against BW. The form of the equation is as follows:

Estimated measurement = intercept + b 1 × ( BW ) + b 2 × ( BW ) 2 + b 3 × ( BW ) 3 ,

*b*

_{1 },

*b*

_{2 }, and

*b*

_{3 }are the constants.

*Z*can be calculated using the following equation:

Z score = ( measurement − estimated measurement ) / standard error .

In this equation, the measurement is the measured dimension on two-dimensional echocardiography. Insertion of data from Table 3 provides the following equation for the ascending aorta:

Estimated measurement = 2.332 + 0.004 × ( BW ) − 1.258 E – 6 × ( BW ) 2 + 2.111 E – 10 × ( BW ) 3 and Z score = { measurement − [ 2.332 + 0.004 × ( BW ) − 1.258 E – 6 × ( BW ) 2 + 2.111 E – 10 × ( BW ) 3 ] } / 0.041 .