Racial Variation in Echocardiographic Reference Ranges for Left Chamber Dimensions in Children and Adolescents: A Systematic Review

Echocardiography plays a critical role in the assessment of cardiac disease. Important differences in echocardiographically derived cardiac chamber dimensions have been previously highlighted in different population groups in adult studies, but this has not been systematically studied in children, whose body size changes throughout childhood. The aim of this study was to review the distribution of available reference ranges for the left cardiac chamber dimensions in older children and adolescents. The following electronic data bases were searched: Medline, Embase and Web of Science were searched to identify studies which have established echocardiographic reference ranges of left heart parameters in children and adolescents from 1975 to December 2017. There was no geographical limitation. All results were imported into Endnote. Retrieved articles were screened and data extracted by two independent reviewers. A total of 4398 studies were retrieved, with 36 studies finally included in this review. 29 (81%) references were from North America and European (Caucasians) populations, with only one study each from Africa and South America. Two-dimensional and M-mode techniques were the most commonly used echocardiography techniques. There were methodological variations in techniques and normalisation of references. Comparison of selected cardiac measures showed significant differences for interventricular septal thickness among Black African, Indian, German and US American children. Available echocardiographic references cannot be generalised to all settings and therefore, there is need for locally relevant reference ranges. Africa and South America are particularly under-represented. Future studies should focus on developing comprehensive echocardiographic reference ranges for children from different racial backgrounds and should use standardised techniques. Electronic supplementary material The online version of this article (10.1007/s00246-018-1873-0) contains supplementary material, which is available to authorized users.


Introduction
Echocardiography plays a critical role in the assessment of cardiac structure and function. Due to changes in body size during childhood, the evaluation of cardiac chambers is highly reliant on the availability of reference ranges, the quality of which depends largely on the availability of a representative sample of healthy subjects and the methods employed to collect the data. The definition of what is "normal" varies widely according to age, body surface area (BSA), gender and race [1,2]. Studies in adults have shown racial differences in echocardiographically derived cardiac chamber dimensions [1]. These differences may be more apparent in children whose body size changes throughout childhood, but this has not been investigated systematically.
The aim of this study was to systematically review racial distribution and methods used in available echocardiographic reference ranges for left ventricular (LV) and atrial (LA) chamber dimensions in children and adolescents. In addition, we compared values of selected chamber dimensions in different racial groups which have utilised the same methods for any differences.

Materials and Methods
This review was registered with the international prospective register of systematic reviews (PROSPERO; registration number CRD42015026030).

Type of Studies
All available studies that reported echocardiographic reference ranges for left cardiac chamber dimensions in healthy children and adolescents, regardless of echocardiographic technique, were considered for inclusion in this review.

Inclusion Criteria and Exclusion Criteria
Studies including at least 50 healthy participants aged 5-21 years that reported echocardiographic measurements at rest were included. We considered studies written in English and published in peer-reviewed journals. Studies that only included neonates and infants, were conducted at high altitude (> 24,000 m above sea level); involved performing cardiac measures during or after exercise; or were based on autopsy specimens were excluded. Systematic reviews and meta-analyses were also excluded.

Search Strategy
The following electronic databases were searched: Medline, Embase and Web of Science were searched. In addition, reference lists of selected studies and other systematic reviews were manually reviewed to identify other possible studies for inclusion. The search strategy included the following words: "echocardiography" AND "reference values" OR "normative" OR "reference standards" OR "reference intervals" AND "child" OR "children" OR "adolescent" OR "z score" (Table 1). Appropriate Boolean operators and truncation were used on synonyms. Both medical subjects and  14 2 or 3 or 4 or 5 or 6 or 13 15 7 or 8 or 9 or 10 or 11 or 12 or 13 16 1 and 14 17 15 and 16 1 3 keywords were used. The same search strategy was adapted for all the listed databases. All results were imported into Endnote X7 (Thomas Reuter). Duplicate citations were removed. Titles and abstracts from the search results were screened independently by two reviewers (EDM) and (GN). The full texts of potentially eligible studies were obtained and assessed in duplicate using a standardised checklist. Any disagreements about inclusion of studies were resolved by consensus.

Data Extraction and Analysis
The following data were extracted using a standard data extraction form: author; population studied; sample size; age range; echocardiography technique; parameters measured, and type of normalisation used.
A two-sample Kolmogorov-Smirnov non-parametric test was used to compare distributions of z score = 0 and 2 for selected cardiac measures of studies which used the same technique for performing echocardiography and the same method for calculation of BSA for normalisation. The selected z score represents the mean predicted value and the upper cut-off for the normal range for a cardiac measure. The null hypothesis was that the compared groups were sampled from populations with identical distributions. Due to multiple testing, the chance of obtaining a significant p value when in fact there is not a true difference between distributions was high. Therefore, we used the Bonferroni adjustment, with a p value < 0.005 considered significant.

Quality of the Studies
We adapted the Newcastle-Ottawa Scale for assessing quality of non-randomised studies to suit cross-sectional studies in the systematic review [3]. In our tool, we assigned scores instead of stars ( Table 2). The following criteria were used to determine the quality of the studies: representativeness of the sample; sample size; sample selection; standardisation of image acquisition and statistical methods used.

Results
A total of 4398 citations were retrieved dating from 1975 to June 2017. Of these, 1193 duplicates were removed, and a further 3075 citations were excluded based on title and abstract (Fig. 1). Full texts of a total of 130 studies were reviewed and 36 studies were included. Characteristics of the included studies are in Table 3.

Left Ventricular Dimensions
The following LV dimensions were reported: diameter at end-diastole (LVEDD) and/or end-systole (LVESD), posterior wall and interventricular septal, area, length, volume, LV mass and index. Saito et al. developed references for LV muscle volume, which is a rarely used measure in clinical practice [37]. 13 studies were from North American children; 12 studies among European children; five studies from Asia and one study each from South American and African children (Table 3).

Left Atrium Dimensions
LA diameter references were established in twelve studies, six of which were conducted in US American children [16,28,30,38,39]. LA length references were from Indian, Italian and Swedish children and of these, two studies also reported data on LA area [5,9,32]. LA volume was derived in four studies and three of these were in US American children (Table 3) [12,14,21,30].

Comparison of LV and LA Dimension Between Studies
Reference Comparison using predicted values of z score = +2 (cutoff for upper limit of normal), significant differences were noted in the IVSd measures between Zimbabwean and German children (p = 0.001). IVSd measures between German and US American children were significantly different (p < 0.001). No significant differences were noted on the distributions of LA measures between the compared studies. The predicted values of z score = +2 by Kampmann et al. progressed in a step-wise fashion so for example, a value of IVSd > 10.4 mm had a z score = +2 for BSA between 1.7 and 1.9 m 2 . The shape of Kampmann's distribution for z score = +2 was therefore strikingly different from the other references [23].

Discussion
Accurate assessment of left cardiac chamber size in children relies on the availability of representative reference ranges. In this study, we have systematically reviewed differences between reference ranges for LV and LA chamber dimensions in children and adolescents. We found many studies which have established reference ranges for LV and LA chamber dimensions, reflecting the significant interest in, and importance of, establishing reference ranges for chamber dimensions in children. Most of the reference ranges in children were, however, developed in European and North American (US populations) and mainly in Caucasians. Notably, there was only one study each from South America (Brazil) and Africa (Zimbabwe) and the latter was published very recently, implying that most African countries rely on Western references in clinical practice which may not accurately represent black African children [6,19]. Comparability of the available reference ranges was limited due to substantial methodological variations, including parameters of normalisation; technique of image acquisition (e.g. 2D or M-mode); measured chamber dimension and/or method used for performing the measurements (e.g. ASE guidelines or other). Cantinotti et al. highlighted that it is imperative to standardise methods of image acquisition and consistency in normalisation of the references [40].
However, we were able to compare a few studies which used similar methods from different races for selected cardiac measures. There were notable differences in some of the measures, particularly interventricular septal thickness among Zimbabwean, German and US American children.
Differences between Zimbabwean and German children were consistently demonstrated at both mean-level and upper cut-off for normal distributions. Although no interventricular septal thickness difference was found in the mean distribution of German and US American children, it was evident in the predicted values z score = +2. In practice, it is the upper cut-off for normal which is used to define abnormality rather than the mean. Our findings suggest that differences in reference ranges between different racial groups do exist but may be overlooked because of the scarcity of data e.g. African children. The use of inappropriate reference ranges may result in either under-or over-diagnosis of cardiac abnormalities or missing of early cardiac chamber remodelling due to cardiac disease [41]. This highlights the importance of using racial-specific reference ranges in clinical practice.
In a recently published study on effect of age, sex, race and ethnicity in echocardiographic z scores of children, significant effects by all the four parameters were observed on z scores. However, the authors concluded that these were not of clinical significance [4]. Given that this study was conducted mainly in US American and Canadian children, findings cannot be   generalisable to the rest of the world due to non-standardisation of methods in the available references. There are also other geographical confounders such as nutrition and altitude which may affect cardiovascular development [4].
Most studies used the M-mode or 2D echocardiography techniques. Advanced techniques such as 3D echocardiography, which may overcome some of the technical challenges of angle dependence and other geometric assumptions associated with conventional techniques, should be used to more accurately quantify chamber sizes and development of reference ranges [42].
This study is limited by the fact that many of the studies did not report the actual race of the children, and we therefore made assumptions based on the country the study was conducted in. We compared very broad racial groups due to scarcity of data. In addition, we were unable to compare references for other cardiac measures due to varying methods used in the studies. However, in the few selected references and cardiac measures where this was possible, we were able to demonstrate significant differences in different races. We also showed similarities in same racial groups. It is also highly likely that in addition to the varying methods used in other studies, racial differences are also present.

Conclusion
This review underlines the importance of using race-specific reference ranges for children, as well as the need for standardising echocardiographic methods in deriving those reference ranges. Furthermore, these reference ranges need to be comprehensive, including a wide range of cardiac measures, as some studies only reported normal values for a single cardiac measure. Future studies should focus on including 3D parameters in addition to 2D and M-mode.
Funding This study was funded by the Wellcome Trust (Grant Number 095878/Z/11/Z).

Compliance with Ethical Standards
Conflict of interest On behalf of all authors, the corresponding author states that there is no conflict of interest.
Ethical Approval All procedures performed in the studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creat iveco mmons .org/licen ses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.