Body Height Trajectories in Pediatric Competitive Athletes from 46 Different Sport Types

Abstract

There is some concern that competitive sports in children and adolescents might hinder their growth. This study’s purpose was to examine height changes in pediatric competitive athletes from 46 different sports. Clinical data from athletes aged <18 years that underwent annual preparticipation examinations at a single sports medicine center were extracted from computerized medical records. Height was transformed into standard deviation scores (SDS). Generalized estimating equations were used to analyze height SDS changes over time in the total cohort and in subgroups of age, sex and several sport types. Data on 2276 athletes were available (71.3% males, age: 12.3 ± 2.6 years), of which 688 had repeated measurements. The median duration between examinations was 1.9 years (interquartile range: 1.0–2.9, maximum: 9.3 years). Height SDS neither significantly changed throughout follow up in the total cohort (−0.01 per year, 95%CI = −0.48–0.03, p = 0.62) nor in subgroups of age and sex. However, height SDS was significantly reduced by −0.12 to −0.23 per year in athletes engaged in endurance sports (i.e., swimming, cycling and triathlon) but not in gymnastics, tennis, basketball or football. We conclude that competitive sports in youth are generally not associated with significant changes in body height relative to age. However, this might occur in endurance sports, possibly due to low energy availability.

1. Introduction

The numerous positive effects of physical activity and sports engagement in children and adolescents are well known [1,2]. However, sports activity might harbor some level of concern from several aspects including growth. Competitive sports frequently involve high energy expenditure and sometimes insufficient energy intake due to the specific sport type’s demands. Parents frequently worry that their child’s athletic activities might hinder her/his statural growth, as certain athletes might lack the ideal environmental factors required to fulfill their height potential [3]. During the previous decades, several authors examined pediatric growth trajectories during sport training. Malina [4] determined that there was neither a difference in the growth rates of 86 male adolescents from several countries who were moderately active or trained in sports nor in 131 female gymnasts, swimmers, rowers and track athletes from several countries. Yet, in another study, the growth rates of 22 female gymnasts who trained 22 h per week were lower compared with 21 swimmers who trained eight hours per week on average [5]. The authors concluded that heavy gymnastics training was associated with hindered growth rates to such an extent that full adult height would not be reached. Gymnasts from both sexes were the focus of additional studies. In a cohort of 86 male elite artistic gymnasts, their final height was shorter compared to that predicted from mid-parental heights [6]. Conversely, in another study of young female gymnasts, swimmers and tennis players, no significant differences were found in final adult heights in 110 athletes for whom these data were available [7]. A review of this topic in elite female gymnasts summarized that there is attenuated growth during the years of training, with catch up during reduced training periods [8]. Ultimately, in a report by the committee of the International Gymnastics Federation, the members concluded that the height of artistic gymnasts was not compromised by intensive training [9].

There are very few studies that focused on other sport types. In a study of 50 competitive male wrestlers aged 13–15 years, there was no significant difference in height when compared with sedentary controls [10]. In a study of 82 novice female dancers aged 8–11 years, no difference in growth was seen over 3 years when compared with controls [11]. A study of 36 male and female distance runners aged 8–15 years, who were followed up for 4–5 years, also found no negative effects on growth rates [12]. Collectively, it appears that while some studies identified growth delays in young athletes, mainly female gymnasts, others did not. The previous studies that examined growth trajectories of young athletes were of relatively small sample sizes and focused on only a few selected sports disciplines. Data from additional sport types are needed to further examine the effects of competitive sports training on growth in children.

The aim of this study was to determine height changes in pediatric competitive athletes from 46 sports types, who were repeatedly measured during annual preparticipation examinations. We hypothesized that there might be a hinderance to statural growth during this period of competitive training in some sport types, hence trajectories in the most common disciplines were examined separately.

2. Methods

This was a retrospective cohort study that used anonymized data from medical records of a large private sports medicine center. Participants were competitive athletes younger than 18 years old in their first clinic visit, who underwent annual preparticipation examinations in a single sports medicine center from 2007 to 2018. The Israeli Sports Law mandates that all competitive athletes (i.e., those registered in formal sports unions and associations), of all ages, undergo annual health examinations in certified sports medicine clinics. Due to the geographical proximity, many athletes undergo their periodic assessments at the same centers—such as the one used for this study. This feature allowed for the creation of a large database composed of recurring measurements of competitive athletes over a long duration.

The data extracted from each clinic visit, which typically took place in the afternoon, included sport type, date of birth, date of visit, sex, height and weight. Heights were measured while standing barefoot against a wall using a stationary roll-up measuring tape (Seca 206, Seca GmbH, Hamburg, Germany). Participant weights were measured on an electronic scale (Seca 700M, Seca GmbH, Hamburg, Germany). BMI was calculated as weight divided by height squared (kg/m²). Height and BMI were transformed to age- and sex-specific standard deviation scores (SDS) of the CDC growth charts [13] appropriate for use in the Israeli population [14] using the lmsGrowth Microsoft Excel add-in (2002–2012 Medical Research Council, Swindon, UK).

Statistical Analysis

Continuous variables are presented as the mean ± standard deviations (SDs). Differences in proportions were compared using the chi-square test. A one-sample binomial test was used to compare baseline mean height SDS to zero (i.e., the median of height for that specific age and sex). Normal distributions of height SDS and age distribution were verified using Q-Q plots and Kolmogorov–Smirnov tests. Continuous variables were, hence, compared using Pearson’s correlation.

Athletes with more than one height measurement formed the main study group, and generalized estimating equations (GEEs) were used to analyze the changes in height SDS over time, adjusted for age at first visit and sex. Subgroup analyses were performed for males and females separately, for participants with different first measurement ages at the clinic (<9, 9–12 and 12+ years, as a proxy of pubertal status) and for athletes from the most common sport types, namely, swimming, cycling, triathlon, gymnastics, tennis, basketball and football. Track and field athletes were not analyzed as a separate group due to the large variation in energy demands and body types of the many sport types aggregated in this discipline.

3. Results

The study database consisted of 4044 height measurements from 2276 different participants, with 688 athletes having repeated measurements (

Table 1. Baseline clinical and anthropometric data of the study’s participants.

	All Cohort (n = 2276)	Recurring Participants (n = 688)
Age at first measurement (years)	12.3 ± 2.6	11.4 ± 2.4
Percent males	71.3%	70.9%
Height (cm)	153 ± 16	149 ± 16
Height SDS	0.30 ± 1.05 *	0.40 ± 1.04 *
BMI (kg/m2)	19.3 ± 3.7	18.6 ± 3.1
BMI SDS	0.17 ± 0.97 *	0.18 ± 0.94 *

* Significantly differed from an SDS of 0, p < 0.001.

). The median time between repeated measurements was 1.9 years (interquartile range: 1.0–2.9) with a maximum of 9.3 years. The study participants trained in 46 different sport types, and the most prevalent ones were basketball (n = 425), tennis (n = 266), cycling (n = 258), triathlon (n = 200), track and field (n = 190), gymnastics (n = 137), football (n = 133), swimming (n = 119), handball (n = 81), judo (n = 66), volleyball (n = 65) and horseback riding (n = 56).

The age at the first clinic visit ranged from 5 to 18 years; hence, our first analysis served to examine if older athletes were taller than expected, signifying some form of selection bias towards higher athletes during adolescence. Figure 1 presents the height SDS across the age span of the whole cohort. Analyzing all study participants together, no meaningful correlation was found between height SDS and age at first visit (r = 0.059, p = 0.005).

Focusing on the 688 participants with repeated measurements, GEE analyses found that there was no significant change in height SDS up to >4 years of measurements (B = −0.010, 95%CI = −0.048–0.028, p = 0.62) (Figure 2). In subgroup analyses, no significant change in height SDS was present in each of the two sexes when analyzed separately (males: B = −0.027, 95%CI = −0.069–0.015, p = 0.20; females: B = 0.028, 95%CI = −0.047–0.010, p = 0.46). No significant change in height SDS over time was seen when participants were divided into subgroups of first measurement age (i.e., <9, 9–12 and 12+ years) and adjusted for sex. However, when examining athletes from the most common sport types separately, significant reductions in height SDS over time were seen among athletes that trained in swimming, cycling and triathlon (

Table 2. Changes in height SDS over time calculated from generalized estimating equations, adjusted for age at first visit and sex, in athletes with repeated measurements from the most common sport types.

Sport Type	Number of Participants	Baseline Height SDS (Mean ± SD)	Number of Measurements	B	95% Confidence Interval	p-Value
Swimming	61 (55% males)	0.650 ± 0.821	181	−0.126	−0.186–−0.065	<0.001
Cycling	51 (57% males)	0.067 ± 1.143	116	−0.228	−0.371–−0.085	0.002
Triathlon	83 (66% males)	0.164 ± 0.899	244	−0.119	−0.225–−0.012	0.029
Gymnastics	58 (12% males)	−0.533 ± 0.780	161	−0.055	−0.128–0.017	0.132
Tennis	82 (72% males)	0.395 ± 0.948	231	0.061	−0.034–0.156	0.209
Basketball	158 (100% males)	0.976 ± 1.02	384	0.032	−0.037–0.101	0.359
Football	24 (90% males)	0.103 ± 0.894	61	−0.058	−0.151–0.035	0.221

) by a magnitude equivalent to approximately 1–2 cm/year. No significant changes in height SDS over time were seen among athletes that trained in gymnastics, tennis, basketball or football.

4. Discussion

The aim of this study was to examine height changes in pediatric competitive athletes from 46 different sports who were followed over several years. The main finding was that height SDS did not change significantly throughout the follow-up period in the group as a whole and also when examined in subgroups of males only, females only or participants with different first measurement ages at the clinic. However, in endurance athletes that continuously trained in swimming, cycling and triathlon, height SDS decreased with time by a degree corresponding to approximately 1–2 cm/year.

The decrease in height SDS trajectories in the endurance athletes in our study can perhaps be explained by the higher energy expenditure in their training sessions as compared to ball games and gymnastics. The youth compendium of physical activities [15] lists the energy expenditure of swimming at ~9–10 METs, cycling at fast speeds at 6.5–8 METs and fast running at ~8–13 METs. In comparison, lower energy expenditure is listed in the sport types where no decrease in height SDS was seen, namely, basketball (~5–7 METs), football (5.5–8.7 METs), tennis (6.3–6.7 METs) and gymnastics (2.7 METs). Per unit of activity, higher energy is, hence, expended in endurance sports compared with ball games and gymnastics. In addition, training sessions in these latter sport types also contain skill practice that is of a much lower intensity than actual play. The total energy expenditure of a training session will, therefore, be much lower than the METs listed above. For example, studies have shown that only ~50% of practice time in male youth ball games contained moderate-intensity physical activity [16,17]. In comparison, running, cycling and swimming are typically performed throughout most of the training sessions in these sport types. Finally, the volume of training in endurance sports is also very high; in one study, for example, adolescent competitive swimmers trained for a minimum of five days per week, with a minimum of 130 min per session [18]. In a recent statement on youth running, it was summarized that youth runners often do not meet their energy needs, placing them at risk for low energy availability [19]. It is, therefore, plausible that the high energy needs and turnover of pediatric endurance athletes might negatively affect their height gain, as found in this study. This novel finding should nonetheless be verified in additional studies that will formally measure energy intake, expenditure and turnover, while addressing sexual maturation, sex differences and biological age.

Regarding the other sport types examined in this study, our findings corroborate the existing literature on female gymnasts and dancers: that their heights are not compromised by intensive training [9,12].

We acknowledge that our study had several limitations. We did not have height measurements taken before sports training commenced, and we did not have data regarding training intensities. However, the first years of sports practice are usually not very intense from an energy expenditure perspective and, hence, are not expected to influence growth. We acknowledge that we did not have information on the sexual development and Tanner staging of the participants. Our subgroups of under 9 or over 12 years old were created to overcome this limitation through the separate analysis of children who were anticipated to be prepubertal and pubertal, respectively. Further, the CDC growth charts that are in such common use, and that were found to be appropriate for use in the Israeli population [14], do not consider the timing of puberty, and height SDS is only affected by age and sex. Finally, our database originated from a single, albeit large, medical center, and the Israeli population is mainly Caucasian. Therefore, our results only represent local training levels and populations.

Our study has several strengths, which were utilizing a fairly large sample of pediatric athletes, from both sexes, from 46 different sport types, with accurate height measurements that were performed over several years (25% of athletes were measured over 3+ years).

5. Conclusions

We conclude that participation in competitive sports by children and adolescents is generally not associated with long-term deviations in their height percentiles, yet this might occur in endurance types of sports with high energy expenditure and risk for low energy availability. If growth trajectories are found to be reduced in an athlete, we recommend a complete nutritional and medical investigation.