Training matters: Heading incidence and characteristics in children's and youth football (soccer) players

Concerns about short‐ and long‐term consequences of repetitive heading contributed to heading restrictions in youth football in some countries. This prospective longitudinal cohort study aims to describe heading exposure in children's and youth football over two seasons using standardized video analysis.


| INTRODUCTION
The intention to use the head to play and control the ball is unique to football. 1 In recent years, short-and long-term consequences and effects on brain function and structure of playing football, especially concerning traumatic brain injuries and repetitive heading, are vividly discussed. [2][3][4][5] Brain alterations and impaired cognitive function regarding heading in adults and adolescents were described in some studies, [6][7][8] but no overall effect has been conclusively described. 2,9,10 Various methodological approaches to investigate acute and chronic neurocognitive effects of heading have been used, 4 but causality of repetitive head impacts due to heading and effects is often inconclusive. 9,10 Data are especially sparse in children and youth players. 10 The Football Association of England has recently published guidelines and recommendations on heading in youth football. 11 A review concluded that youth football players head the ball at a low frequency that increases with age and is influenced by sex. 10 However, a wide variation of heading burden between studies regarding total heading numbers, typically recorded in matches, and head impact forces in youth players complicates the assessment of the real magnitude of heading. 2,12 Longitudinal observations of training sessions are rare despite the much larger rate of overall football exposure. 2 There is less knowledge about heading exposure in training over a longer period of time and various training cycles, although needed for a realistic evaluation of heading loads.
The gold standard for analyzing head injury mechanisms in football is video observation, [13][14][15] which more and more frequently is used to assess heading exposure. [16][17][18] Recently published studies investigated heading frequencies and types by the use of video observation in a large-scale cross-sectional design across eight European countries 16 or by direct observation in an international youth football tournament. 19 The aim of this prospective longitudinal study was to describe heading exposure and characteristics in elite children's and youth football players after two seasons of video observation in both match and training. Therefore, we examined heading frequency and characteristics in elite German youth football of different age groups in match and training.

| Study population and design
In this prospective longitudinal cohort study, heading exposure and characteristics were investigated in a cohort of elite children's and youth football players over two seasons using standardized video analysis. Four regional youth teams of different age groups consisting of one male Under-11,  Under-15, and Under-19 and one female Under-17 team with a high level of play were approached via personal contact to club officials and coaches. Inclusion criteria for participating teams were regular training and regular participation in official matches in one of the top three leagues of their age group. All players and their parents/guardians (if participants were under 18 years) provided written informed consent before participation. The study was approved by the ethics committee of the State Medical Board (Ärztekammer Westfalen-Lippe in Münster, Germany) (2019-321-f-S) and registered at DRKS (DRKS00018923).
All matches and training sessions of participating teams were videotaped during the seasons of 2019-2020 (August 2019-March 2020) and 2020-2021 (July-November 2020; March-September 2021). Teams were active in the three highest German junior leagues (Westfalen-, Regional-, Bundesliga). All league, cup, friendly matches, and tournaments with regular football (no futsal) rules were involved in match play analyses. Training included match warm-ups except for individual training sessions during the COVID-19 pandemic, during which no headers were performed.

| Patient and public involvement
The study design and conduct were discussed with club officials before starting with video observation. Data analysis, writing, or editing was done without public involvement.

| Video analysis
Standardized heading protocols were used for assessing headers in all players equally, with previously defined criteria. 20 In player-based analyses of headers, the following characteristics were evaluated from video recordings by team-related analysts, and collected using a common Research Electronic Data Capture (REDCap) tool hosted at Paderborn University 21 : • The flight course of the ball could not precisely be measured, but auxiliary quantities such as size, boundaries, and areas of the field were used for interpretation and classification. Calculation of individual match and training exposure was defined by number of matches or training sessions x duration of match or training session in hours. 22 In the case of head injuries, the injury mechanism was analyzed using a standardized injury registration form, orientated on previous studies. 23

| Inter-rater reliability
Raters were trained in video analysis by discussing (especially unclear) heading situations utilizing the standardized protocol and practice analyses before the study started. Five randomly chosen matches and eight training sessions from male Under-11 (one training), Under-15 (three trainings, one match), Under-19 (two trainings, two matches), and female Under-17 teams (two trainings, two matches) were screened by three independent video analysts. Concordance of results on total headers was assessed by calculating intra-class correlation coefficients (ICC; one-way random).

| Statistics
Statistical analysis was performed using Microsoft Excel 2019 and SPSS 27 (IBM Statistics, New York, USA). Median, minimum, and maximum number of headers per match/training session were calculated and individual heading exposure was derived from headers per player per match hour/training session. Similar to previous literature, incidence rates (IR) per 1000 match/training hours with 95% confidence intervals (CI) were calculated. 16 IR was compared between age groups by incidence rate ratios (IRR). Additionally, IR was checked for normal distribution by Shapiro-Wilk test. To detect group differences in match play and training, Kruskal-Wallis test was used due to not normally distributed data. Wilcoxon test was used for comparing match and training IR within age groups. Correction for multiple comparison was done using false discovery rate.

| Heading exposure and incidence
In total, 275 matches and 673 training sessions were recorded, corresponding to a total match, and training exposure of 13 506.5 h. A total of 22 921 headers in match play and training were registered and examined. Exposure time, number of recorded matches and training sessions, total headers, and average heading IR per player per age group are shown in Table 1.

| Match play
Heading differences between age groups in match and training are visualized in Table 2

| Training sessions
IR in training between age groups differed significantly (χ 2 (3) = 32.73, p < 0.001). Under-11 males had significantly lower incidences compared to all older age groups (by more than −48%, p < 0.01). There were no further significant group differences. Average distributions of heading numbers per training session per player are shown in Figure 2. Heading IR differed significantly between training and match in all age groups (p < 0.04).

| Match play
Proportions of heading characteristics across age groups are displayed in Table 3. In match play, most frequently single intentional headers without heading duels were registered (58%). In 58% of all heading duels, there was contact with an opponent. In 39% of all duels, there was no contact with another player, followed by contact with many players (2%) and a teammate (1%). The predominant match situation for headers was free game play (54%). In 54% of all headers, the flight distance of the ball before heading was between 5 and 20 m with varying proportions in age groups. The number of headers flying >20 m was very low in Under-11 males (11%) and increased with increasing age. The contact point to the player's head predominantly was at the frontal part (79%), while in Under-11 headers were also played with other parts. In all age groups, the smallest proportion of headers was performed by goalkeepers (<6%), whereas midfielders accounted for 43% of all headers, followed by defenders (34%).

| Training sessions
In training, most of all headers were played without duels in all age groups (91%, 7% with duel, 2% no info).
In heading duels, 48% of headers were without body contact to another player. The predominant situation was free game play (84%). The flight distance was <5 m in 65% of all headers, followed by 29% between 5 and 20 m, 6% between 20 and 50 m and less than 1% >50 m. The ball most frequently hit the frontal part of the head (93%), followed by the parietal part (4%). Heading characteristics in different age groups are presented in Table S1.

| Head injuries
Three concussions were registered in match play, with participation of opponents. Two injuries occurred in free game play while changing direction (Under-19 male) and being shot from nearest distance (Under-17 female). An aerial heading duel caused one further concussion (Under-19 male). No head injuries occurred during training sessions.
T A B L E 1 Number of recorded matches/training sessions, game format, exposure time, total headers, headers per player per match hour/training (mean ± SD), and heading IR (per 1000 match/training hours) per player.

| Heading exposure in youth football
Total heading frequency per match and IR in matches increased with increasing age (in males). This finding confirms previous studies 19, 24 and cross-sectional observations in eight European countries. 16 16 Additionally, the results of the present study showed high percentages of players heading zero or one to two times per match hour on average and no player heading more than four times in all age groups, which is in line with previous reports that quantified heading exposure during an international youth football tournament, including Under-12 to Under-20 players from both sexes. 19 There were also many players with one to two headers or less per training session in Under-11 and Under-19 males. However, in Under-15 males, less than half of the players were heading 0-2 times and five players performed five to eight headers per training session, accounting for 34% of all headers in that age. This substantial variation in individual heading numbers between youth players was already described in match play 19 and might be reasoned by training specific to players' requirements and demands, as position showed to affect heading exposure. 25 Heading incidence in training of Under-15 males is comparable to rates in the literature (IR: 2511). 16 16 One reason could be a higher awareness on limiting the heading burden under preventative aspects in younger age groups at present due to media attention, previous limitations of heading in US soccer 29 and recommendations by 2020 published UEFA heading guidelines. 30 Also, our data might represent more realistic heading exposures, compared to cross-sectional studies with an assessment of only one training session per team. Unfortunately, studies examining heading exposure in training are rare so far, despite the proportion of training sessions being much larger compared to matches. 2,26,27 The (rapid) increase in IR from Under-11 to Under-15, specifically in training might partly be explained by an increase in field size in these age groups as the number of players per team in matches increased and differing training plans. As lower heading numbers are associated with small-sided games, the field size is an influencing factor for technical demands. 31,32 In all age groups heading IR differed significantly between training and match. In Under-15 males and Under-17 females, heading incidence in training exceeded match play by around 600 headers per 1000 player hours. Additionally, the upper range of headers per training was much higher compared to match for all ages. Different to matches, training sessions offer many opportunities for controlling (individual) heading behavior, as content and organization largely depend on trainer's intentions. Taken together with previous reports, 16 these results may provide the basis for strategies for training headers in youth players and underline the importance of systematic heading observation in training.

| Heading characteristics
Contact with another player, especially head-to-head contact like in heading duels is the primary mechanism of head injuries in adult football players. 13,33 In all groups, most headers were performed without heading duels. Duels were without contact to another player even in more than one-third of all headers in youths and therefore at low risk for head injuries.
Although ball-to-head contact is rarely associated with head injuries, ball velocity and contact force between ball and player's head in heading is an important risk factor regarding potential head injury consequences. 33,34 In training sessions, headers were most frequently performed from <5 m flight distance. These short-distance headers are presumably performed at lower velocities and with lower forces to the head than long-distance headers. 12,19 In matches, the number of long-distance headers flying >20 m was very low in Under-11 males and slightly increased with increasing age, matching results from Sandmo et al., 19 most likely as a result of the increasing field size and muscle strength in players.
Head impact location most frequently was at the frontal part, usually representing intentional headers. 1,34 In the youngest age group, we found nearly one-third of all impacts at the temporal, parietal, and occipital part, which is in line with results from Harris et al. 35 These head impact locations could represent improper techniques leading to larger rotational velocities and higher linear accelerations of the head. 36

| Methodological consideration
Variabilities in heading frequency and different training practices between countries because of differences in training curricula and rules in match play 16 may have influenced the results. Also, contents of training and way of playing might depend on the clubs and coaches' interests and preferences. All teams were active on a very high level, but heading exposure in other leagues or teams might differ. Nevertheless, the longitudinal design across two seasons of observation with a sample covering a large age span is strengthening the validity of results.
Unfortunately, only one female Under-17 team meeting the inclusion criteria was regional available and included in this study. For assessing heading exposure in females of different age groups, further research is needed.
Despite canceled training sessions and matches due to pandemic situations, the number and exposure time of observed training sessions is much higher compared to previous studies. 16,26,27,37 Single training sessions showed at times very high heading exposures, due to sessions focusing on heading training. Although the coaches were advised to execute "normal" training during study participation, video observation might have motivated some players and/or coaches to give either more or less attention to heading. Over the course of the study, however, those effects have most likely weaned off.
Many previous studies used questionnaires and interviews to assess heading exposure. 4,8,38 Although positive correlations between self-reported and observed heading exposures were found, 39 youth players systematically overestimated the actual header number in a recent study. 37 Player observation is needed to accurately quantify heading frequency in youth soccer. 18 Therefore, the method of video analysis is a study strength, allowing exact analysis of heading characteristics. Unfortunately, biomechanical parameters like forces to the head were not captured.

| PERSPECTIVE
Under-11 football players showed low heading incidences, which increased with increasing age in matches, whereas Under-15 males showed high heading exposure in training. Due to variations in heading exposure between age groups strategies for training headers in youth players should account for age-related differences. Head impact location in Under-11 might represent improper techniques in one-third of all headers in matches. A reduction of headers, for example, by more small-sided games, might decrease the use of improper heading techniques. 2 As predominantly five Under-15 males showed high heading exposures in training, special attention to individual heading burden in training is needed. Education and sensitization of coaches are important to reduce heading drills to what is deemed necessary for learning correct heading techniques and to increase awareness for players with a high preference for heading. Additionally, training offers many opportunities to control for heading frequency, recovery times, flight distance, or weight of the ball. Training sessions and individual heading behavior should specifically be addressed when heading burden and regulations are addressed.
This study provides a basis for future studies that investigate the relationship between realistic heading exposures and potential (long-term) effects as heading exposure and characteristics are well described.

AUTHOR CONTRIBUTIONS
RR and CR were responsible for the conception and design of the study. WK conceptualized the heading protocol. RR was responsible for data collection over the study period. RR, FH, and LD conducted the statistical analyses. RR wrote the paper. The draft of the paper was critically revised by LD, FH, WK, and CR.