This systematic review was undertaken in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement guidelines [19]. Thus, a systematic search of the research literature published in peer-reviewed journals was conducted for systematic review articles with or without meta-analysis studying the epidemiology of physical activity related injuries among children and adolescents in different settings. To carry out this review, English language literature searches of PubMed, CINAHL, Web of Science, Embase, SPORTDiscus, PEDro and Cochrane Register of Controlled Trials were conducted from April 2021 up to June 2021. Electronic databases were searched using the following keywords and their combinations: “sport injury” “athletic injury,” “physical activity,” “Injury occurrence,” “Injury prevalence,” “epidemiology,” “leisure time activity,” “Children,” “school,” “youth,” “Adolescents,” “Review,” “Meta-analysis.” The reference lists of each included article were also scanned to identify additional relevant studies. The study protocol was prospectively registered at PROSPERO registry: CRD42021247312.

The target population of the selected review articles was adolescents from 10 to 19 years of age (WHO definition), with no gender restriction, studies where age groups were overlapping and if it was not possible to extract information of only adolescents the studies were excluded. The outcomes were injury occurence/incidence and injury prevalence. Studies were excluded according to the following criteria: (a) studies written in languages other than English; (b) original studies (c) studies that included other populations than adolescents and it was not possible to extract information of only adolescents; (d) studies from which we could not extract values of interest like RRs, IRs, % of injuries or number of people who got injured. The risk of bias in systematic reviews and assessment of multiple systematic reviews (AMSTAR 2) tool was used to evaluate risk of bias and quality of included articles [20]. We adjusted the AMSTAR 2 tool to be suitable for assessing the quality of descriptive studies or analytical observational studies of health outcomes. From questions 1 and 8 we removed the components of PICOS [21] (intervention and comparator group) that are not applicable to descriptive or observational studies. In questions 3, 11, 12, and 13 we removed randomized controlled trials and non-randomized studies from the answer options. Question 9 was rephrased as “Was a suitable tool used to evaluate the quality of included studies? (Supplementary Material S1). Only studies with moderate to high quality score (≥6.5), were included for the purpose of the present study.

Six independent reviewers (DS, PB, BS, EA, ML, KT) performed the literature search, identification, screening and data extraction. First, the titles were initially screened by the reviewers during the electronic searches to assess the papers’ suitability, and all papers beyond the scope of this review were excluded. Second, the abstracts were assessed using predetermined inclusion and exclusion criteria. Third, the full texts of the remaining papers that met the inclusion criteria were retrieved and included in the ongoing procedure and reviewed by the two reviewers (PB and KT) to reach a final decision on inclusion in the review. Finally, the reference lists from the retrieved manuscripts were also examined for any other potentially eligible papers. Any disagreements between the reviewers were resolved by consensus or arbitration through a third reviewer (AP). If the full text of any paper was not available, the corresponding author was contacted by mail or ResearchGate. The study selection process as described above is illustrated in Figure 1.

The Cochrane Consumers and Communication Review Group’s data extraction protocol was used to extract the participant information, including sex, age, sample size, training status, description of the intervention, study design and study outcomes [22]. This was undertaken by one author (AP), while a second author (EA) checked the extracted data for accuracy and completeness. Disagreements were resolved by consensus or by a third reviewer (PB). Reviewers were not blinded to authors, institutions, or manuscript journals.

A total of 34,059 articles were identified by the literature search (Figure 1). Following the removal of duplicates and the elimination of articles based on title and abstract screening, 178 studies remained. An evaluation of the remaining 178 studies was conducted independently by two researchers (PB and KT). The list of excluded studies after full texts assessed for eligibility, with reasons for exclusion, can be found in Supplementary Material S2. The list of excluded studies after quality assessment can be found in Supplementary Material S3. Following the final screening process, 19 studies were included in the review and these were published in the year 2014 or later.

The methodological quality of the included reviews is depicted in Figure 2. On average, the included studies were of moderate quality, with an AMSTAR-2 score of 8.3, ranging from 6.5 to 10.5. All studies described their research question and inclusion criteria using relevant PICO components (Item 1) and explained the desired study designs for inclusion in the review (Item 3). Additionally, all studies partially met the criteria for using a comprehensive search strategy (Item 4). However, none of the reviews provided a list of excluded studies to justify the exclusion reasons (Item 7) or reported the sources of funding for the included studies (Item 10).

All selected studies were designed as systematic literature reviews, and meta-analysis was conducted in two articles to summarize the findings [23, 24]. In total, 572 original studies were included in the reviews. The majority of reviews investigated injury incidence and prevalence in both sexes, whereas only two studies aimed to investigate this question specifically for females [25] or males [26]. Injuries sustained in a single sport were investigated in the majority of reviews, these sports were rugby [25, 27], baseball [28], artistic gymnastics [29], cricket [30, 31], football [26], handball [32], fieldhockey [33], netball [34], and swimming [35]. Seven reviews included a variety of sports, such as baseball and softball [36], action sports (snowboarding, alpine skiing, motorsports etc) [24], or several different types of sports [23, 37–40]. Nine studies investigated epidemiology of a variety injury types within a selected sport [25, 26, 29–34, 41], while others focused on particular types of injuries such as concussion or traumatic brain injury [24, 27, 36, 39], ankle sprain [23], elbow disorders [28], back pain [40], shoulder pain [35], and overuse injuries of the extremities [38] or overuse injuries of wrist alone [37]. Five reviews included only studies with prospective injury surveillance [23, 26, 28, 30, 33].

Thirteen articles reported data on injury incidence, whereas six articles reported data on injury prevalence (Table 1). The large heterogeneity on reporting the injury incidence was observed, thus making it difficult to quantitatively summarize data on injury incidence. Five studies presented data on incidence per 1,000 player-hours [27, 32–34, 41], 1,000 days of real exposure [24], per 1,000 exposures [23, 28], per 1,000 h of playing/match/training exposure [25, 26], per 100.000 people [39], per athlete [29], or athlete exposure [38]. The highest incidence of 90.9 per 1,000 player-hours was reported in field hockey during the African Cup of Nations [33], whereas lowest overall injury incidence was reported in handball ranging from 1.7 to 7.8 per 1,000 player-hours [32]. Five studies [25, 32, 38, 39, 41] examined differences in injury incidence during match and training settings. Raya-Gonzalez and colleagues [32] investigated injury incidence in handball and found that more injuries are reported during match than training (match: 10.8 to 73.6 vs training: 0.96 to 4.1 per 1,000 player-hours). The similar pattern was observed in rugby [41], where 3.7 to 108.5 vs 0.4 to 56.2 injuries occurs per player-hours during match and training, respectively. Four studies presented data on injury incidence for different age categories [23, 28, 32, 33], while three studies examined the influence of the athletes’ sex on the risk of injury [23, 29, 32]. In handball, senior female and youth male athletes reported greater injury incidence compared to youth female handball players (13.0–36.0 vs. 10.8–23.8), but higher compared to senior male handball players (13.0–36.0 vs 15.0 to 73.6 per player-hours) [32]. On the contrary, Doherty and colleagues [23] reported higher injury incidence in females than males (13.6 vs. 6.94 per 1,000 exposures), with also higher injury incidence in children than adolescents (2.85 vs. 1.94 per 1,000 exposures), and higher injury incidence in adolescents than adult athletes (1.94 vs. 0.72 per 1,000 exposures) respectively.

Finally, injury prevalence was reported for different sports including cricket, swimming, artistic gymnastics, and a variety of sports together. Also, prevalence of different types of injuries was examined such as overuse injuries of wrist only [37], low back pain [40], shoulder pain [35] or all injury related complaints.

Trompeter and colleagues investigated lower back pain prevalence in different sports and reported highest lifetime prevalence in rowing and cross-country skiing (1%–94%) [40]. A study by Kox and colleagues revealed a prevalence ranging from 32 (period not specified) to 73% (within 6 months previous to survey) for wrist pain and 10 (current) to 28% (within 10 years) for wrist overuse injury, respectively [37]. Feijen and colleagues explored the prevalence of shoulder pain in swimmers belonging to different age groups and found the highest prevalence of shoulder pain in adolescent swimmers (91.3% within the past month), followed by adults (70.3%), young (20.0%), and masters (19.4%) combined from within the past week and point prevalence [35].

The majority of reviews encompassed studies of organized sports conducted in sports club environments as well as in school sports settings [24, 25, 27, 30, 31, 35–37, 41]. Some also included professional sports [28, 33, 40]. The school sports were mainly from high school or college settings [24, 25, 35, 36].Three reviews included studies conducted in an opera school [37], vocational dance school [38], and a football academy environment [26]. Studies where gymnasts were recruited from school teams were excluded from the review of injuries in competitive artistic gymnasts [29]. One review focused solely on injuries at elite and national level in handball [32], one focused on injuries in high level male youth football [26], while the reviews by McLeod et al were confined to community cricket excluding high performance centers [30, 31].

None of the reviews were limited solely to leisure time physical activity injuries, although studies that were based on emergency room, medical clinic, or hospital records also included injuries from leisure time activities [24, 34, 36, 39, 41]. Some studies also included studies based on insurance claims or national electronic injury surveillance systems [24]. None of the reviews encompassed only studies of injuries sustained in school PE.