Identifying modifiable risk factors and screening strategies associated with anterior cruciate ligament injury risk in children aged 6 to 13 years: A systematic review

ABSTRACT Growing anterior cruciate ligament (ACL) injury incidence is reported in countries across Europe, North America and in Australia for 5–14-year-olds, yet research on injury risk reduction predominantly focuses on populations aged > 13 years. For injury risk reduction, it is crucial to understand (i) which modifiable risk factors are associated with ACL injury in children (6–13 years) and (ii) how these risk factors are assessed. Articles were grouped according to sex/gender and/or maturational/age differences and examined modifiable risk factors during different physical screening tasks. The included articles (n = 40) predominantly examined intrinsic risk factors in girls aged 10–13 years. Factors mechanically linked to increased ACL loading at this age included increased peak knee adductor moments, knee valgus angles, hip and knee extension, and ground reaction forces. Assessment focused on laboratory-based assessments (e.g., motion capture, force plates). This review concluded that modifiable risk factors are present in children aged 6–13 years and that injury risk reduction strategies should be implemented as early as possible regardless of sex/gender. Further, screening strategies need updating to be childhood specific and feasible for the wide community. Additional research on extrinsic risk factors, norm values and children aged 6–9 years could allow for more targeted risk reduction strategies. Highlights Increasing rates of ACL injuries in children aged 5 to 14 years are reported in countries across Europe, North America and in Australia Research on modifiable risk factors focuses on internal risk factors in children aged 10-13 years and neglects external risk factors as well as younger children (6-10 years) Screening strategies to determine risk of ACL injury risk in children are laboratory based as opposed to cost-effective and quicker-to-analyse in-field assessment Research is warranted to examine external risk factors and in-field screening strategies in childhood


Introduction
The prevalence of severe sporting injuries such as anterior cruciate ligament (ACL) injuries are increasing in children across Europe, North America and Australia (Beck et al., 2017;Maniar et al., 2022;Weitz et al., 2020).ACL injuries result in significant short-and longterm consequences (e.g., time away from school and sport participation, lengthy rehabilitation period, risk of secondary ACL injury, early onset of osteoarthritis) (LaBella et al., 2014).Of additional consideration is the cost burden ACL injuries pose on health care systems (LaBella et al., 2014).Evidence of childhood ACL injury rates increasing over time is inconclusive, and conflicting.Some literature indicates that children aged under thirteen years are at continued low risk (Herzog et al., 2017;Weitz et al., 2020) whilst ACL injury rates in Australian children aged 5-14 years increased by 147.8% from 2005 to 2015 (Shaw & Finch, 2017).In America, a small increased risk (4.5%) is reported for five to nine year old's and larger increased risk (18.9%) for ten to fourteen years old's (Werner et al., 2015).
Research on ACL injury mechanism, risk factors, and prevention approaches is well established in adolescents and adults but appears to be less studied in children.To the best of our knowledge, the actual injury mechanism of ACL injury has not been examined specifically for children and is consequently unknown.However, in adolescents and adults, ACL injuries occur commonly in non-contact situations during rapid deceleration (e.g., slowing down after a sprint), pivoting (e.g., change of direction) or jump-landing movements (Kaeding et al., 2017;Krosshaug et al., 2007;Shimokochi & Shultz, 2008).The type of ACL injury varies according to skeletal maturation and rigidity (Campón Chekroun et al., 2022).Partial ACL tears are more frequent in skeletally immature individuals and complete tears in mature populations (Campón Chekroun et al., 2022).
To reduce the risk of primary ACL injuries, it is imperative to understand the risk factors linked to ACL injury and specifically, factors that can be intentionally altered by modifying the environment and individual (e.g., through intervention programmes).These modifiable risk factors are commonly divided into extrinsic and intrinsic risk factors (Mandorino et al., 2023).Extrinsic risk factors refer to the choice of sport, sport specialisation, or equipment (Mandorino et al., 2023).Understanding extrinsic factors is crucial to promote adequate prevention strategies.Such extrinsic intervention strategies may include, participating in targeted sport specific intervention programmes, shifting to a lower risk sport, or reducing the time spend in exposure (Heering et al., 2023;Mandorino et al., 2023).Intrinsic risk factors refer to joint kinematics, ground reaction forces, or muscle strength and activation patterns (LaBella et al., 2014;Mandorino et al., 2023).Prominent intrinsic risk factors mechanically linked to stressing the ACL are knee abduction moments and angles and ground reaction forces (Hewett et al., 2016).Further, quadriceps and hamstring strength ratio, and lower extremity asymmetries between dominant and non-dominant leg are commonly examined (Krosshaug et al., 2007;Schweizer et al., 2022).
Modifiable risk factors are affected by biological sex factors.With the onset of maturation, the ACL and skeletal anatomy adjust to growth (Bergeron et al., 2015).These changes differ between biological sex and occur on average earlier in girls (8 to 13 years old) than boys (9 to 14 years old) and require adequate adaptations (e.g., muscular strength, neuromuscular control) (Bale et al., 1992;Quatman et al., 2006).Girls develop more high-risk anatomy with the onset of puberty (e.g., smaller ACL cross-sectional area-tolength ratio) (Hosseinzadeh & Kiapour, 2021) whereas boys control growth related changes by increasing muscle strength and neuromuscular control (Wild et al., 2012).These differences are shaped by both biological sex differences and gender norms (e.g., different developmental opportunities) (Parsons et al., 2021).For example, resistance training is a crucial part of many ACL injury prevention programmes, an element that typically observes reduced participation rates in girls (Padua et al., 2018).An explanation for this trend is resistance training being a male dominated environment where many girls feel unwelcome (Parsons et al., 2021).Parson and colleagues suggest that "gendered environmental factors" are present in training, competition and rehabilitation and likely influence intrinsic risk factors (Parsons et al., 2021).Throughout the current manuscript, the term "sex/gender" is used to recognise the "entanglement" of biological and social factors (Parsons et al., 2021).
To identify individuals at risk of an ACL injury, modifiable risk factors are used as parameters in the screening process (Holden et al., 2016;Schweizer et al., 2022).Screening is performed either subjectively by an expert, rating specific movement criteria (e.g., Landing Error Scoring System, Tuck Jump Assessment), or objectively with motion capture, wearables and force plates (Dowling et al., 2011;Schweizer et al., 2022).Physical screening tasks range from jump-landing movements such as drop vertical jump (DVJ), strength tests, to gait and running related tasks such as sidecutting task (Schweizer et al., 2022).When assessing ACL injury risk, it should be considered that assessment tools need to reflect the injury mechanism and challenge the neuromuscular control of the individual (Schweizer et al., 2022).Many non-contact injuries occur during horizontal deceleration or side-cutting (Kaeding et al., 2017;Krosshaug et al., 2007;Shimokochi & Shultz, 2008) and it was suggested that single-leg movements represent the actual ACL injury mechanism better than jump-landing movements (Schweizer et al., 2022).To date, the use of assessment tasks and tools to screen for ACL injury risk has not been examined in children.
Despite considerable efforts to reduce the risk of ACL injuries (Padua et al., 2018) injury rates continue to increase in children aged 5-14 years (Maniar et al., 2022).An IOC consensus statement on paediatric ACL injuries critiques the current research as inconsistent and limited by methodologies with small samples and high risk of selection bias (predominantly cross-sectional and retrospective study designs) (Ardern et al., 2018).Additionally, the failure to consider skeletal maturity was noted and the authors recommend that future research identifies modifiable risk factors for ACL injury in childhood (Ardern et al., 2018).Without summarising what is known about modifiable risk factors and screening strategies specific to children, preventive measures to reduce risk of ACL injury in this population will not be forthcoming.Therefore, the purpose of this systematic review is to identify the evidence on modifiable risk factors of ACL injuries in children aged under 13-year-olds.The objectives are to identify: (i) modifiable factors mechanically linked to an increased risk of ACL injury and (ii) physical screening tasks and data collection tools used in children aged 6-13 years.

Search strategy
The literature search to retrieve articles was performed according to the PRISMA statement (Page et al., 2021).The age range was set according to the definition of childhood (6-12 years) in the databases and additionally included 13-year-olds, as current ACL injury risk research focuses on populations age 14 years and onwards (Mattu et al., 2022).Five electronic databases (PubMed, Embase, Medline, CINHAL and SportDiscus) were searched according to the following predefined concepts: anterior cruciate ligament AND injury AND children AND (i) occurrence, (ii) prevention, (iii) diagnostics or (iv) risk factors.The search excluded the search concept "reconstruction" in the title of articles but was not limited in publication year (Appendices 1).The last literature search was completed on the 9 th of November 2022.Articles were handled in Endnote X9.3.3 (Clarivate Analytics, Boston, US) and Covidence data management software (Veritas Health Innovation Ltd, Melbourne, Australia).Two authors (first and additional author) examined study eligibility according to predefined selection criteria.

Eligibility criteria
Articles were included if they, first, examined in human participants aged 6-13 years and results were presented for this age group individually.If age was not specified in the abstract, words like "immature", "pre-/pubertal", "child" or "school" were needed to indicate a youth population to be considered in the full-text screening.Second, modifiable risk factors (e.g., extrinsic, intrinsic) of primary ACL injury were assessed.Baseline data from intervention studies were considered when results were reported for baseline.Articles were excluded if they focused on non-modifiable risk factors, reported descriptive statistics only, duplicates that were not picked up previously, non-peer reviewed journal articles such as conference abstracts or dissertations and languages other than English.Any disagreement was resolved by two authors until agreement on study in-or exclusion was reached.

Data extraction and synthesis
A customised data extraction sheet was created, and data extracted and synthesised by the first author and checked by the second author.Extraction parameters included, but were not limited to authors, publication year, participant characteristics, screening strategy (e.g., tools and tasks), examined variables and study findings.Articles were examined for modifiable risk factors and organised according to sex/gender, maturational/age or other comparisons.Further, the articles discussion sections were screened for interpretation of findings on modifiable risk factors.

Risk of bias assessment
Quality of included studies was assessed with the Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (National Institutes of Health).The first two authors independently assessed the study quality and rated study quality as good, fair or poor based on 14 criteria.A lower rating was reached when (i) participant eligibility was not described (e.g., the time period of data collection was missing in nearly all studies), (ii) sample size was not calculated with power analysis but small, (iii) confounding variables such as maturation were not assessed with valid and reliable tools and (iv) relevant confounding variables were not considered in the analysis (e. g., maturational stage, BMI, sport, dominant leg, and previous acute knee/ACL injury (Leppänen et al., 2021).Conflicts in study quality were discussed until agreement was reached.

Study characteristics
An overview of the studies included in this systematic review is provided in Table 1.All studies were published from 2004 until 2022.Study participants were characterised as predominantly girls (70%) with the majority aged 10-13 years (82%).Included articles examined primarily intrinsic risk factors (93%) with laboratory-based data collection where participants performed muscle strength (23%), change in direction (e.g., side-cutting) (13%) and/or jump-landing tasks (47%).The DVJ was the most frequently (52%) examined jump-landing task (Table 2).

Modifiable risk factors
A variety of variables were examined (Figure 2a, b) (Appendices 3) including predominantly intrinsic risk factors (e.g., ankle, knee, hip and trunk kinematics and kinetics, ground reaction forces, muscle strength and activation patterns) and two external risk factors (e.g., type of sport and sporting level).Articles compared between sex/gender, maturation/age, or other parameters such as fatigue or hamstring strength.

Comparison between sex/gender
Comparisons between sex/gender were performed in 22/40 articles examining kinematic and kinetic risk factors, muscle strength and activation patterns.

Kinematic and kinetic risk factors. 4/22 articles
indicated no sex/gender differences during a DVJ in hip and knee flexion/extension (Jackson et al., 2010;Swartz et al., 2005), hip abduction/adduction (Jackson et al., 2010), knee varus/valgus (Jackson et al., 2010;Swartz et al., 2005), hip and knee internal/external rotation (Jackson et al., 2010) angles and moments, peak vertical ground reaction forces (vGRF) (Swartz et al., 2005), time to peak vGRF (Swartz et al., 2005) and medial knee motion (Hewett et al., 2004).Additionally, at peak knee flexion angle during counter movement jump (CMJ) (Holden et al., 2015).However, 9/22 articles indicated significant differences between sex/gender with boys or girls displaying kinematic and kinetic risk factors mechanically linked to ACL injury risk.Modifiable risk factors mechanically linked to ACL injury risk in boys compared to same aged girls were: larger frontal plane displacement during CMJ (Holden et al., 2015), higher knee valgus moment, internal rotation moment, first peak vGRF, rotation excursion, and extension excursion during side-cutting (Petrovic et al., 2020) and higher peak forces during the first 100 ms after initial contact (IC) of side-cutting (Sigurðsson et al., 2018).Modifiable risk factors mechanically linked to ACL injury risk in girls compared to same aged boys were: larger fatigue effects on knee flexion excursion and vGRF during DVJ (Briem et al., 2017) or knee valgus during tuck jumps (Fort-Vanmeerhaeghe et al., 2019), lower knee joint stiffness during DVJ (Ford et al., 2010), decreased knee/hip ratio during IC and peak knee flexion of DVJ (Sasaki et al., 2013) and larger hip/knee ratio during DVJ (Sigward et al., 2012), less desirable landing biomechanics (e. g., knee flexion at IC) during counter movement jump (Holden et al., 2015), no increase in maximal vertical jump height interpreted as failure to demonstrate a neuromuscular spurt (Quatman et al., 2006), coincidental summation of early force peaks during side-cutting (Sigurðsson et al., 2018), larger peak abduction angle and moment peak knee adductor moment during weight acceptance of side-cutting (Sigward et al., 2012) compared to their male counterparts.Examining ground reaction force profiles, Pedley and colleagues noted that girls display a poor stretch-shortening cycle during DVJ (Pedley et al., 2021).This indicates a poor ability to attenuate landing forces, presented as large impact peak, and to show spring like behaviour (Pedley et al., 2021).However, Briem and colleagues found also protective measures in girls when examining the effect on fatigue on kinematic and kinetic variables.Girls are more effected by fatigue compared to boys and displayed greater trunk and knee flexion at IC and lower knee flexion moments at peak vGRF (Briem et al., 2017).Additionally, two articles discussed that the observed sex/ gender-dependent differences may not be clinically important in terms of ACL injury risk.Namely, at peak push-off force and knee flexion during DVJ (Lucarno et al., 2021) and vGRF and knee extension excursion angle during side-cutting (Petrovic et al., 2020).

Muscle strength and activation patterns. Results
on muscle strength are mixed for the hamstring and quadriceps muscle group.Girls showed lower concentric (Franchi et al., 2019) and eccentric (Hewett et al., 2004) hamstring strength, but no difference in isometric hamstring strength (Ahmad et al., 2006) compared to boys.Similar results were observed for quadriceps strength.Hewett and colleagues reported that girls showed lower concentric quadriceps strength compared to boys (Hewett et al., 2004), whereas Ahmad and colleagues found no difference between sex/ gender for isometric quadriceps strength (Ahmad et al., 2006).Girls showed an increased quadriceps to hamstring ratio compared to boys, indicating a quadriceps dominance (Ford et al., 2010).No sex/gender difference were observed for co-contraction hamstring muscle activity (Russell et al., 2007), lower limb muscle activation during side-cutting (Del Bel et al., 2017) and the number of muscle synergies and similar activation coefficients during landing of single-leg landing task (Kipp et al., 2014).However, Kipp and colleagues found sex/gender difference during a unilateral jumplanding task.Girls displayed larger semitendinosus activation pre-landing and at IC compared to boys (Del Bel et al., 2017;Kipp et al., 2014).Post-landing, the vastus medialis weight coefficient was smaller in girls compared to boys (Kipp et al., 2014).

Other factors. Additionally, Hutchinson and collea-
gues reported no difference in sex/gender during manual coding of a modified drop-jump and single-leg balance task, but boys performed poorer during a single-leg squat manually coded by experts (Hutchinson et al., 2016).

Kinematic and kinetic risk factors. Modifiable risk
factors linked to ACL injury risk were greater peak knee adductor moments, peak knee valgus angles, peak GRF during sidecutting (Sigward et al., 2012), greater hip and knee extension, more knee valgus, higher, more abrupt vGRF during a vertical jump (Swartz et al., 2005), increased peak knee valgus angles during side-cutting, unanticipated side-cutting and double-leg jump (Thompson-Kolesar et al., 2018), greater knee adductor moments (Sigward et al., 2012), lower reactive strength index (Strniště et al., 2019) compared to older/more mature populations.Further, Lucarno and colleagues observed limb asymmetries more often in younger/less mature compared to older/ more mature participants during a vertical jump (Lucarno et al., 2021).However, others reported limb symmetry in medial knee motion in pre-and early-puberty during landing of DVJ (Hewett et al., 2004) and CMJ (Holden et al., 2015).McLean, 2016;Hewett et al., 2004,Girls), (Quatman-Yates et al., 2013;Wild et al., 2016).Similar results were observed for quadriceps strength.Using manual and isometric dynamometry, five articles reported younger/less mature populations displayed lower concentric (Hewett et al., 2004, Boys, Wild et al., 2013b), isometric (Ahmad et al., 2006;Quatman-Yates et al., 2013) or eccentric (Wild et al., 2013b) quadriceps strength compared to older/more mature populations.However, two articles did not find a significant difference in concentric, isometric or eccentric quadriceps strength (Davidson & McLean, 2016;Hewett et al., 2004, Girls).Further, disproportionate use of the knee extensors relative to the hip extensors were reported at pre-puberty (Sigward et al., 2012).Additionally, two articles examined hip abduction strength and both found that younger/less mature populations display larger concentric hip abduction strength compared to older/more mature populations (Prather et al., 2015;Quatman-Yates et al., 2013).

Other risk factors. De Ste Croix and colleagues
found younger children (U13) to be more affected by fatigue than older populations (U15/U17) resulting in longer electromechanical delay (De Ste Croix et al., 2015).Further, younger/ less mature populations failed a single leg balance and drop jump task more often (Hutchinson et al., 2016), had restricted hip range of motion (Prather et al., 2015) and more landing deficits in the modified Tuck Jump Assessment (Fort-Vanmeerhaeghe et al., 2019) compared to older/more mature populations.

Type of sport. When comparing dancers and team
sport athletes, dancers displayed more ACL protective landing biomechanics during landing of a hop and stop-jump task compared to team sport athletes (Harwood et al., 2018).

Sporting level.
No difference was found regarding sporting level between elite and non-elite athletes during reactive strength index and leg stiffness during hopping (Strniště et al., 2019), whereas non-elite had lower eccentric hamstring strength compared to elite athletes during a Nordic Hamstring Exercise (Franchi et al., 2019).

Hamstring strength.
Wild and colleagues found significantly lower hamstring to-quadriceps ratios, reduced hip abduction moments, greater knee abduction alignment, and similar muscle activation patterns during the landing phase of a horizontal landing movement for the lower hamstring strength group (Wild et al., 2013a).The authors concluded that the latter appear to have a decreased capacity to control lower limb frontal plane alignment potentially increasing the risk of ACL injury (Wild et al., 2013a).

Dynamic knee valgus. Another article concluded
that increased dynamic knee valgus correlates with mechanical energy absorption of ankle, knee and hip during uni-and bilateral DVJ (Dadfar et al., 2021).

Limb dominance. Different internal strategies were
observed with the dominant limb employing higher hip energy generation contribution and the non-dominant limb more knee energy generation contribution during a lateral vertical jump ( DeLang et al., 2021).During unanticipated side-cutting task, greater overall magnitudes of muscle activation were observed in the dominant limb compared to the non-dominant limb (Del Bel et al., 2017).

Influence of fatigue. A significant increase in modifi-
able risk factors such as peak knee valgus moment, greater knee valgus angle at IC, and less knee external rotation angle at IC were observed (Petrovic et al., 2020), especially in girls (Briem et al., 2017).

Exercise feedback.
The use of feedback cues significantly increased the proportion of participants who correctly completed movement exercises in 8-to 11 year old's (Ling et al., 2021).Finally, Prather and colleagues suggest that all participants with a positive provocative hip test on physical examination are at increased risk of injury regardless of age (Prather et al., 2015).A positive provocative hip test indicates pain during a subtest of specific test at the hip (e.g., anterior hip impingement test, resisted active straight leg raise) (Prather et al., 2015).

Physical screening tasks and data collection devices
A variety of physical screening tasks and data collection devices were utilised across all articles and will be described in the following section.

Discussion
This is the first study to summarise the intrinsic and extrinsic modifiable risk factors mechanically linked to an increased risk of ACL injury in children aged 6-13 years.With rising ACL injury incidence in that age group, it is necessary to understand modifiable factors to provide a foundation for age specific risk reduction programmes.Overall, this review found that research on the modifiable risk factors for childhood ACL injuries focuses on female samples aged 10-13 years and rarely included 6-9year-olds.Additionally, intrinsic risk factors such as kinematics and kinetics and muscle strength were predominantly examined and external risk factors such as sport type or equipment neglected.Physical screening tasks ranged from jump-landing to side-cutting tasks and the DVJ was the most tested task.Screening was laboratory based and tasks examined with 3D motion capture and force plates and muscle strength with isokinetic dynamometry.In-field assessment were rarely used.Study results were compared according to sex/gender or maturation/age.Accordingly, gaps in literature include examining children aged 6-9 years, a variety of extrinsic risk factors, and the use of in-field assessments.
The current review found mixed results for emerging sex/ gender differences in muscle strength and landing biomechanics.In line with findings from De Ste Croix and colleagues (De Ste Croix et al., 1999, 2002), two articles reported no sex/gender differences in knee extensor or flexor strength in participants younger than 14 years (Ahmad et al., 2006;Franchi et al., 2019).However, Hewett et al. reported boys to have higher hamstring and quadriceps peak torques at all maturational stages including pre-puberty (Hewett et al., 2004).Peek et al. show similar trends at pre-pubertal level but without reporting sex/gender specific statistics (Peek et al., 2022).However, girls demonstrate less desirable jump-landing biomechanics (landing load asymmetry, peak push-off force and knee flexion angle) compared to boys even in early-puberty (Holden et al., 2015) and at young ages (Lucarno et al., 2021).More recent research has shown that, boys demonstrate great knee frontal moments during side-cutting task compared to girls at pre-puberty (Ghasemi et al., 2023).Lucarno et al. raise the question whether these differences in landing biomechanics are clinically important in terms of ACL injury risk (Lucarno et al., 2021).Further, Briem et al. concluded that girls aged ten years indicated a less efficient post fatigue shock absorption strategy with higher first peak vGRF, decreased knee flexion excursion, and limb asymmetry in the frontal plane during DVJ compared to same aged boys (Briem et al., 2017).It remains unconclusive whether risk reduction strategies in this age group should differ according to sex/ gender.
The total number of ACL injuries is higher after the onset of puberty (Beck et al., 2017;Maniar et al., 2022).In line with those observations, more articles in the current review indicated pubertal and post-pubertal populations displaying risk factors mechanically linked to ACL loading compared to pre-and early-pubertal populations.However, we also found studies which demonstrated that younger/less mature athletes display risk factors linked to ACL injury risk.For instance, younger/less mature athletes showed larger fatigue effects (De Ste Croix et al., 2015), higher landing GRF, peak knee adductor moments (Sasaki et al., 2013) and more landing deficits in tuck jumps (Fort-Vanmeerhaeghe et al., 2019) when compared to older/more mature groups.Accordingly, risk reduction approaches should be introduced at early ages.We are not the first to suggest this strategy, rather the results of the current review reinforce previous recommendations (Bergeron et al., 2015;Myer et al., 2013;Petushek et al., 2019).Considerable efforts have been undertaken to design ACL injury prevention programmes for adolescent and adult populations (Arundale et al., 2022).Risk reduction programmes were adjusted to address younger populations as well (e.g., FIFA 11+ Kids) and were successful in reducing football specific injuries by 48% in children aged 7to 13 years (Rössler et al., 2018).Similar programmes are needed for sports other than football.For instance, an epidemiology study on ACL injury incidence rates in Japanese junior high school students observed higher incidence rates (per 1000 players) in sports like skiing, basketball, handball, and judo compared to football regardless of sex/gender (Takahashi & Okuwaki, 2017).
In the studies included in this review, laboratory-based measures (e.g., motion capture, force plates) screening methods were favoured over other measures.Despite the benefits (e. g., valid and reliable assessment in a controlled setting), laboratory assessments are time consuming, expensive, and require specialised equipment (Bardid et al., 2019).Current technology enhancement like marker-less motion analysis or wearables aim to reduce those negative elements but are still inaccessible to the wider community (Wade et al., 2022).Screening should be performable by coaches or team physiotherapists for prompt implementation of risk reduction strategies (Fox et al., 2016).The alternative to laboratory-based measures could be in-field screening tools (e.g., Landing Error Scoring System (LESS) or Tuck Jump Assessment).Considered as cost-effective, in-field screening tools require less equipment, but are not without critique either (Fox et al., 2016).The manual coding requires extensive training to be valid and reliable, tools are rarely validated for childhood populations and predictive ability in identifying future ACL injuries is questionable (Fox et al., 2016).For instance, the LESS is based on a DVJ performance, but the DVJ has been critiqued for not accurately representing the ACL injury mechanism and associated ACL stress (Schweizer et al., 2022).Many non-contact injuries occur during horizontal deceleration or side-cutting (Kaeding et al., 2017;Krosshaug et al., 2007;Shimokochi & Shultz, 2008) and it was suggested that single-leg movements represent the actual ACL injury mechanism (Kristianslund & Krosshaug, 2013;Schweizer et al., 2022).To the best of our knowledge, there is no childhood specific infield screening tool based on single-leg movements.The screening tests and tools observed for children during the current review are similar to the ones described by Schweitzer and colleagues for adolescents and adults (Schweizer et al., 2022).However, children undergo unique developmental changes (e.g., anatomical, physiological and performance changes (Bergeron et al., 2015)) which should be considered in ACL injury screening strategies for children.This developmental perspective of ACL injury risk in children appear to be missing in current literature.
It was suggested that increasing ACL injury incidences may occur due to early specialisation in a single sport (Bergeron et al., 2015).Early specialisation is associated with a decreased skill repertoire and reduced health-related benefits.As such, it has been recommended to specialise in a single sport only after reaching proficiency in a wide range of motor skills (Bergeron et al., 2015).Accordingly, an important risk reduction approach in childhood could be participation in multiple sports (Carder et al., 2020).Through multi-sport participation, movement skills considered as foundational for more complex and sport-specific movements are developed (Goodway & Robinson, 2015).A broad skill foundation is associated with decreased risk of sport injury, including ACL injuries, as found in a systematic review and meta-analysis including athletes aged 7-18 year olds (Carder et al., 2020).This evidence indicates that children may not have the adequate breadth and depth of skill development to be sufficiently prepared for sportspecific participation (Bergeron et al., 2015).The question arises whether a lack of proficiency in a range of motor skills is associated with increased risk of ACL injury, yet this has not been addressed in current research and so this question could not be answered in this review.
An additional aim of this review was to highlight current gaps in the literature.The included articles predominantly focused on intrinsic risk factors, however, extrinsic factors (e.g., choice of sport, sporting history, time spent training a sport and number of matches in a season) can highlight specific populations at risk of an ACL injury (e.g., team ball players with direct contact during basketball or soccer) (Pfeifer et al., 2018).The influence of these factors are examined in adolescence and adults (Gornitzky et al., 2016), but are not well researched in children aged 6-13 years and should be considered in the future.Additionally, normative values of optimal performance (e.g., knee flexion angle) in children are missing but are warranted for meaningful interpretation.Hannon and colleagues provide normative data for quadriceps, hamstring, hip abduction and external rotation strength (Hannon et al., 2022).Additional data is warranted for various tasks, sex/gender and maturational groups.Limitations of the included studies consist of unreliable indicators of skeletal maturity (e.g., only chronological age) (Ardern et al., 2018).Maturation should be assessed with suitable measures (see IOC consensus statement for measures (Bergeron et al., 2015)) when assessing ACL injury risk in childhood.Concerningly, within this review, 53% of included studies did not include measures of maturation.All studies were from developed countries, which may limit the transferability of findings to other countries and cultures (Draper et al., 2022).The identified gaps in the literature, led to subsequent recommendations for future work.It is recommended that future research should consider: (i) examining modifiable risk factors in 6-9-year-olds and (ii) more external risk factors in 6-13-year-olds, (iii) developing in-field assessment tools specifically for children, (iv) collecting norm values for optimal performance for different physical screening tasks, and (v) examining associations between ACL injury risk and lack of proficiency in a range of motor skills.Without studies examining these topics, effective risk reduction strategies for children will be a challenge.

Conclusion
Despite efforts to reduce the risk of ACL injuries, injury occurrence is increasing in childhood.This review examined a variety of intrinsic and extrinsic modifiable risk factors predominantly regarding sex/gender or maturational/age difference.Mixed results were reported for modifiable risk factors when comparing sex/gender.When comparing 6-13-year-olds to older/more mature populations, some risk factors (e.g., increased peak knee valgus angles during side-cutting) were present at the young age group already.Accordingly, intervention programmes should be implemented as early as possible, regardless of sex/ gender, to improve lower limb alignment and control during tasks that stress the ACL.Another finding was that laboratory measures dominate paediatric ACL research, as opposed to costeffective and quicker-to-analyse in-field assessment which could be more feasible for the wider community.However, in-field measures need further development to be child and ACL injury specific.For instance, focusing on single-leg exercises (e.g., deceleration task) and considering unique developmental changes that children undergo.For more targeted risk reduction strategies, further research on extrinsic risk factors, children aged 6-9 years, normative values, and in-field screening strategies specific to children are warranted.Evidence on primary ACL injury regarding to:

Figure 2 .
Figure 2. Modifiable risk factors examined in included articles (n = 40) separated into kinematic and kinetic (A), and neuromuscular (B) factors; summarised in other are additional factors investigated by only one study each, refer to.
split into three sections: a. demographics, b. location and c. time period, ✓ = Yes, -= No, * No sample size power calculation, lack of participant eligibility and/or confounding variables resulted in a poor rating.Abbreviations: CD cannot determine, NA not applicable, NR not reported, Q Question.Appendices 3: Table of modifiable risk factors examined in included articles (n = 40) and divided by physical screening task force-time (%) Pedley et al. (2021)Landing peak: take-off peak ratioPedley et al. (2021) Relative landing to take-off force ratioPedley et al. (2021) Landing-take-off time difference (

Identification of studies via databases Identification Screening Included Figure 1. PRISMA
flow diagram of screening process.

Table 1 .
Summary characteristics of included studies (n = 40).Normalised values for the peak vGRF, knee angles in the frontal and sagittal planes at IC and at peak vGRF, sagittal and frontal plane external knee joint moments at peak vGRF, trunk angle at IC and peak vGRF, time from IC to peak vGRF.
b Wild; Abbreviations: ACL anterior cruciate ligament, BW normalised to body weight, COL college students, CON control group, DVJ drop vertical jump, F female, HS high school, IC initial contact, INT intervention group, JHS junior high school, M male, MS middle school, N/a not available, P pubertal, PHV Peak Height Velocity, PMOS Pubertal Maturation Observation Scale, PRO Professional athletes, T test period, U under, UK United Kingdom, US United States of America, vGRF vertical ground reaction force, Y test year.
(Davidson & McLean, 2016)le, Abbreviations DVJ drop vertical jump.hamstringstrengthcompared to older/more mature populations.In contrast, Davidson and MacLean reported lower eccentric hamstring strength mid-puberty compared to early-puberty(Davidson & McLean, 2016).Additionally, five articles did not find a significant difference in concentric, isometric or eccentric hamstring strength(Davidson &

Appendices 1: Search strategy
paediatric or paediatric or young or youth or kids or immature or pubertal or juvenile or primary school or middle school 4.1 Diagnostic diagnostic tool or screening or assessment or detection or tool or test or measuring or survey or diagnosis or Landing Error Scoring System or LESS or drop vertical jump or DVJ or side cutting or jump-landing task or deceleration task 4.2 Risk factors risk factor* or contributing factor* or predispos* factor* or predictor or cause or protect* or knee valgus or dynamic knee valgus or knee flexion or knee abduction or vGRF or vertical ground reaction force 5 Surgery NOT (limited to title) reconstruction or repair or surgery Note.Two separate literature searches were conducted according to concept 4 (i.e., diagnostic, risk factors).