Sex-Specific Changes in Physical Risk Factors for Anterior Cruciate Ligament Injury by Chronological Age and Stages of Growth and Maturation From 8 to 18 Years of Age

The protocol followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).²⁷  The study protocols were specified in advance and registered in the National Institute for Health Research International Prospective Register of Systematic Reviews (PROSPERO 2021 CRD42021251191, https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=251191).

Studies were selected based on the Participants, Exposure, Comparator, Outcomes, Study Design (PECOS) guidelines (https://training.cochrane.org/handbook).

Studies were identified by searching electronic databases (PubMed, CINAHL, Scopus, SPORTDiscus), scanning article reference lists, and coauthors scanning of their own reference databases. The searches were limited to papers published in English, but the dates were not restricted. The full search was conducted on July 6, 2021, and a final search took place on December 3, 2021. We used these terms to search all databases: fat free mass, fat mass, anterior cruciate ligament, femoral notch width, intercondylar notch, tibial slope, knee laxity, balance, postural control, lower extremity alignment, knee strength, hip strength, and lower extremity biomechanics were paired with maturation (maturation, puberty, sexual development) and sex (sex, gender). We added BMI (body mass index) and obesity as exclusionary terms given the large volume of papers associated with body composition that were not relevant to our search. Although BMI is a known ACL injury risk factor, well-established reference data for this measure were available based on 5 national health surveys of more than 16 000 participants (Centers for Disease Control and Prevention BMI Growth Charts; Table 16: https://www.cdc.gov/nchs/data/series/sr_11/sr11_246.pdf). Additionally, BMI is only a rough estimate of adiposity and does not partition the sex- and age-specific changes in fat mass (FM) and fat-free mass (FFM) that contribute to overall weight. As such, we focused the searches on sex-specific changes in FM and FFM. Appendix A provides an example of the full electronic search strategy used in PubMed.

The search results from the 4 data sources were uploaded to a systematic review management platform (Rayyan Systems Inc). Duplicates were identified, confirmed via visual inspection, and removed. Given the size and breadth of the search, our study selection process consisted of 3 steps. In step 1, all titles and abstracts were independently screened by 2 reviewers (S.J.S. and M.R.C.), guided by the eligibility criteria that had been agreed upon by all coauthors. Discrepancies between the reviewers were resolved with a consensus-based discussion; when questions persisted, the articles were carried forward. The full texts of eligible articles based on the titles and abstracts were then retrieved and independently screened by the same 2 reviewers to further evaluate eligibility criteria, identify secondary citations from reference lists, and code each article based on the risk factor(s) examined (step 2). Articles carried forward from step 2 were entered into a spreadsheet and divided into 6 groups of risk factors (body composition, anatomical factors, balance, muscle strength, leg power, and lower extremity biomechanics). Two coauthors were then assigned to each group of risk factors to perform an in-depth review and quality assessment of each article based on the study criteria to determine final inclusion (step 3). If maturation was relevant to the research question and the outcome data were reported in aggregate (versus stratified by age or maturational stage), we attempted to contact the article's author(s) up to 3 times by email over a 4-week period. If these efforts were unsuccessful, the article was excluded. If, upon in-depth full-text review, a study was found to be ineligible, this was confirmed by both coauthors.

A data-extraction template was created so that each coauthor used the same standardized system for collecting data for the assigned outcomes. The template was piloted by the coauthors over a 1-week period, all coauthors met to resolve any questions or challenges, and the template was refined accordingly. To avoid double-counting data, we carefully compared articles reporting the same outcomes from the same authors or from the same data set. If the same study data were reported in more than 1 publication, the data were treated as 1 dataset. If the data overlapped, we used either the data with the largest sample or those that provided the more discrete stratification by individual ages or maturational stages. For data that were presented in a figure, we first attempted to contact the authors for the raw data in tabular format. If our efforts to reach the authors were unsuccessful, specialized software (WebPlotDigitizer; Automeris LLC²⁸) was used to extract the data from the figure if the clarity was sufficient. All data for each outcome were extracted by a single author.

The following data were extracted from each study: (a) author(s); (b) study design (ie, cross-sectional, longitudinal, mixed); (c) sample demographics (ie, sample size [boys, girls, total]), total age range examined, activity level (general population, physically active, athlete, sedentary, not specified), race or ethnicity, and country or origin; (d) manner in which data were stratified (eg, categories of age, Tanner stage, age at PHV, or pubertal stage [prepubertal, pubertal, postpubertal]); (e) risk factor characteristics (ie, measure, method of assessment, unit value); and (f) data (measures of central tendency and dispersion and sample size by each sex and age or maturation category reported).

In addition to the originally planned outcome variables, the search also yielded 34 articles that described total leg power via a single-legged hop or vertical or horizontal jump and 15 articles that addressed flexibility. Because these outcomes are often employed in ACL injury-prevention programs and leg power is used as a performance metric in ACL injury-screening and rehabilitation protocols, we extracted these variables as well. Additionally, we found that leg FFM was reported more often than thigh muscle mass. These factors are assumed to be highly correlated with one another, so we retained both outcomes.

We assessed the quality of individual studies using the National Institutes of Health Quality Assessment Tool for Observational Cohort and Cross-Sectional Studies (https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools). The tool was revised to add 3 questions that were particularly relevant to this review with the ability to reduce bias in the meta-analysis and graphic representation of the data: (1) Was the sample size >10 participants for each sex for each age and maturation category examined? (2) Were both boys and girls assessed at each time point? and (3) Were assessments made across the full range of ages (8–18 years) or the full range of maturational stages (eg, Tanner stages I to V); yes was only indicated if the data for each individual age or individual maturational stage were provided. Two reviewers working independently assessed and scored each study. If the final scores differed by >1 point, the reviewers discussed the discrepancies. These data were used to ensure an in-depth review of each article, confirm study eligibility, assist with identification of studies to be included in the meta-analyses and the graphical representation of the data for each outcome, and aid in data interpretation. Studies were not excluded based on quality scores.

Our primary interest was in examining the collective sex-specific changes in ACL risk factor outcomes across age, maturational stage, and age of PHV. For aim 1, we examined the results for each risk factor independently. To be included in the summary tables and narrative review, the following criteria had to be met: data for both boys and girls were included and the data were clearly stratified by age or maturation status to represent 2 or more maturity levels (prepubertal, pubertal, postpubertal).

For aim 2, we addressed sex differences in trajectories over time. Random-effects meta-analyses were conducted using the Comprehensive Meta-Analysis application (version 3.3.070; Biostat Inc). Meta-analyses were performed for an outcome if ≥2 studies supplied sample sizes and measures of central tendency and dispersion for both boys and girls by chronological age or maturational stage that clearly spanned 2 or more pubertal stages (prepubertal, pubertal, postpubertal). Because all outcome variables were continuous in nature, we used the subgroup (age or maturational stage) sample size, mean, and SD to calculate the standardized difference in means between boys and girls for each outcome. Standardized differences in means were calculated to compare homogeneous outcomes that were measured in a variety of ways or reported in various units.^29,30  We used conversions reported by Wan et al³⁰  to estimate the means and SDs when the data were given as medians with the interquartile range or range. Equations to convert standard errors or CIs into SDs were provided by the Cochrane handbook.²⁹  Because most sample sizes for subgroups were <60, we used the formula with the t distribution value as the denominator in calculating the SD.²⁹  For each outcome, we constructed a forest plot by age or maturity category. The standardized differences in means, standard errors, lower and upper confidence limits, P values, and the studies contributing to each age group were available with each forest plot. For interpretation purposes, a standardized difference in means >0 favored boys and <0 favored girls as having greater values for each outcome. A subgroup was included only if ≥2 studies gave the sample size, mean, and SD for both boys and girls.

For aim 3, we provided an overall summary of the data. We used the standardized point estimates derived from the studies that were eligible for the meta-analysis to create an infographic of the comparative trajectories in outcomes by sex and time (age, maturation). If we were unable to perform a meta-analysis on a particular outcome, we used a single study to represent that risk factor if it was rated as good to high quality, incorporated 80% of the age or pubertal stage categories, and included ≤10 participants for each age or pubertal stage category. When a single study was used, this was clearly noted in the graphic. If the data were insufficient for that risk factor, this was also noted.

Bias was managed through rigorous searches, and review and scoring of available studies were conducted by ≥2 investigators for each outcome variable. We did not calculate the Egger test for small study bias because a significant number of subgroups across most outcome variables lacked at least 3 studies (≥3 are needed). Therefore, sample sizes, the number of studies, and weight were provided so readers can make informed judgements about the robustness of the meta-analyses. As noted, tests for asymmetry (potential bias) in meta-analyses are generally considered underpowered for ≤10 studies.²⁹  Hence, any meta-analysis results with <10 studies contributing to a subgroup should be viewed with caution.

The searches yielded a total of 1805 citations, with 1556 unique articles remaining after we removed duplicates. A total of 216 articles from the primary search and 22 articles from the secondary search (ie, screening of reference lists) were carried forward from step 2 and entered in a spreadsheet. The flow diagram in Appendix B.1 shows the number of articles that underwent in-depth review for each group of risk factors. Studies that included outcomes for >1 risk factor group were evaluated separately for each risk factor. Detailed forest plots with individual listings of studies included in the meta-analyses by age and maturity subgroup can be found in Supplemental File 1 (available online at https://doi.org/10.4085/1062-6050-0038.22.S1).

The searches produced 79 papers that addressed sex-specific changes in body composition with specific outcomes of percentage of body fat (%BF), FM and fat mass index (FMI; kg), FFM and fat-free mass index corrected for change in stature (FFM and FFMI, respectively; kg/m²), and appendicular leg/thigh muscle mass (ALM). We included ALM given the functional importance of the leg musculature in movement control and reported associations between thigh muscle mass and ACL size.^31,32  After in-depth review, 39 articles were excluded for not meeting the eligibility criteria, and 2 papers were excluded for supplying duplicate data. The most common reasons for exclusion were not providing stratified data by sex and age or Tanner stage, the wrong study design or outcome, only reporting 1 sex or age span, or insufficient data to confirm that the age span represented 2 or more pubertal stages. We contacted 12 authors to obtain stratified data by sex and age without success. This left 38 studies that examined 1 or more variables of body composition (Appendices C.1–C.4).^33–70  The majority of articles (27) described sex-specific changes with age, with fewer (14) reporting data by Tanner stage; 2 articles offered data by sex for both age and pubertal stage for select outcomes.^33,54  No investigators stratified data by the age of PHV. The data were insufficient to perform meta-analyses for FMI, FFMI, and ALM by Tanner stage.

The studies of %BF consistently demonstrated increases with each age or Tanner stage in girls, whereas the boys' values either maintained or increased early and then decreased during pubertal development (typically after Tanner stage II and age 12–13 years; Appendix C.1). Meta-analysis revealed greater %BF in girls by age 8 years (Appendix B.2)^{41,50,53,63,69}  and Tanner stage I (Appendix B.3),^{36,44,48,58,60,61}  with this sex difference increasing from 11 to 16 years (at which time it begins to plateau) and Tanner stages II to V. The increasing sex difference in %BF results from sex divergence in both FM and FFM. The FM increases similarly in boys and girls up to 12 years of age and Tanner stage IV and then continues to rise in girls, while stabilizing and decreasing in boys (Appendix C.2). Conversely, FFM increases steadily in boys throughout adolescence, while increasing more slowly and plateauing earlier in girls (Appendix C.3). The meta-analyses confirmed these findings, showing increasingly greater FM in girls starting at age 12 years and greater FFM in boys as early as age 8 and Tanner stage I, with girls displaying increasingly greater FM accumulation by 12 years of age (Appendix B.4)^{33,34,50,51,69,70}  and Tanner stage IV (Appendix B.5)^36,46,47,71  and boys displaying increasingly greater FFM accumulation from age 10 (Appendix B.6)* and Tanner stage I (Appendix B.7)^{36,44,46,47,54,61}  onward. Fewer researchers reported changes in FMI and FFMI by age (Appendices C.2 and C.3),^{38,43,50,51,66}  and the findings were similar to those for FM and FFM (see Supplemental File 1).

Sex-specific changes in ALM follow a trend similar to FFM; boys and girls increase similarly up to the age of 13, and boys accumulate more muscle thereafter (Appendix C.4). Increases in leg muscle mass occurred 1 to 2 years earlier in girls than in boys.³⁴  This is consistent with their earlier pubertal development, as Marwaha et al⁵⁴  observed that both boys and girls experienced the largest increases between pubertal stages I and III. Meta-analysis of 5 included studies^{34,53,54,56,63}  revealed that boys had increasingly greater leg muscle mass than girls from 14 to 17 years of age (Appendix B.8). The data were more variable at age 18 years, given fewer included studies.

The search resulted in 36 articles that described sex-specific changes in knee-extensor, knee-flexor, and hip strength. After in-depth review, we excluded 27 articles for not meeting the inclusion criteria (20 articles, 13 of which only reported 1 sex) and inability to obtain sex-stratified data by age or pubertal stage (7 articles). Thus, 9 studies were included in the current review.^53,72–79  One article addressed hip-extension, hip-abduction, and hip-rotation strength,⁷⁴  and the remaining 8 articles supplied both knee-extensor and knee-flexor strength values. Four articles offered sex-specific changes by pubertal stage, whereas 5 gave changes by age. Because of insufficient data, we were unable to perform a meta-analysis on hip strength by age or maturity level.

Most authors noted linear trends for knee-extensor and knee-flexor strength as age and maturation level increased. In 5 of the 9 studies, knee-extensor strength in boys began to exceed female values by 12 to 15 years or Tanner stages IV and V (Appendix C.5).^{73,75,77–79}  Conversely, 1 group⁷⁴  (of 2 groups who normalized strength to body mass) found greater knee-extensor strength in prepubertal girls compared with prepubertal boys and postpubertal girls. The meta-analysis revealed that knee-extension strength was greater in girls at age 9, similar to boys at age 10, and then less than boys at ages 11 and 12 years, with no differences evident at age 13 years (Appendix B.9).^{53,74–76,79}  However, when stratified by maturity level, boys demonstrated more knee-extension strength than girls at pubertal and postpubertal stages, with the greatest mean difference occurring at the postpubertal stage (Appendix B.10).^73,74,77,78  For hamstrings strength, 5 of the 7 included studies showed that male strength exceeded female strength between 8 and 12 years or in Tanner stage IV (Appendix C.6). When these studies were combined, the meta-analysis revealed greater knee-flexor strength in boys across all age groups (Appendix B.11)^53,74,76,79  and increasingly greater strength in the pubertal and postpubertal stages (Appendix B.12).^73,74,77  For the 1 included study on hip abduction and external rotation,⁷⁴  no changes were present in either sex across pubertal stages (Appendix C.7). However, hip-extension strength (normalized to participant mass and height) tended to decrease linearly in both sexes as the pubertal level increased; no interactions were observed between sex and maturation level.

Our search identified 34 articles describing leg power metrics. After an in-depth review of each, 18 were excluded for not meeting the inclusion criteria (3 articles) or inability to obtain sex-specific stratified data by age or pubertal stage from the authors (15 articles), leaving 16.^{52,53,63,80–92}  The remaining articles depicted reported sex-specific changes in leg power by age (10 articles) and sexual maturation (6 articles).

Horizontal leg power, including during the standing long jump and single-legged hop, rose to a greater extent in boys versus girls with age and maturation in most studies (Appendix C.8). Performance in the standing long jump improved in boys with age and sexual maturity in all studies except one,⁸⁰  in which it was stable between ages 15 and 18. Girls also demonstrated increased standing long-jump distances with increasing age in 2 of 3 studies, but increases based on maturity were observed only between prepubertal and pubertal stages.⁹¹  Single-legged hop performance improved in both boys and girls with age and maturity, with greater increases in boys.^88,90  By maturity, single-legged hop distances were greater in boys during Tanner stages III to IV and V but similar in boys and girls during Tanner stages I to II.⁸⁸  The meta-analysis for age (Appendix B.13)^80,84,87,90  and Tanner stage (Appendix B.14)^88,91  confirmed greater horizontal jump distances in boys versus girls at 12 and 13 years and across all pubertal stages. This sex difference increased from 11 to 12 to 13 years and from pubertal (Tanner II to IV) to postpubertal (Tanner V) stages.

Studies of vertical leg power measured via standing vertical jump, countermovement jump, drop jump, or single-legged hop are detailed in Appendix C.9. When compared by age, the findings were inconsistent. In 1 study, vertical jump height was greater in girls 9 to 10 years and 13 to 14 years,⁸⁹  whereas some researchers identified equal vertical jump performance by sex across ages,^53,83  and others detected greater increases in boys.^52,63,85  When stratified by maturational status, boys demonstrated increased vertical jump height across maturation while jump height performance in girls remained the same^82,86,92,93  or increased less from prepubertal and postpubertal girls.⁸¹  Similarly, single-legged hop distance increased in both boys and girls from 11 to 16 years; thereafter, hop height continued to increase in boys but decreased in girls from 15 to 16 and 17 to 18 years.⁵²  Meta-analysis for age^{52,53,63,83,85,89}  revealed a general trend toward increasing sex differences from 12 to 15 years except for age 13, though findings of the 4 included studies were more variable (Appendix B.15). The meta-analyses for Tanner stage^81,86  were limited to comparisons at the pubertal and postpubertal stages and showed greater vertical height in boys at both stages, with this sex disparity being greatest at the postpubertal stages (Appendix B.16).

The search yielded 12 articles that explored sex-specific changes in knee-joint anatomy, including notch width (NW), transcondylar width, notch width index (NWI), ACL cross-sectional area (CSA) and width, ACL length, ACL sagittal-inclination angle, ACL coronal-inclination angle, lateral tibial slope, medial tibial slope, intercondylar roof-inclination angle, and medial tibial depth. After an in-depth review, we excluded 4 articles for not meeting the inclusion criteria or an inability to obtain sex-specific data stratified by individual age or pubertal stage from the authors, leaving 8 articles. In 7, sex-specific changes in knee-joint geometry with age were reported; 1 article provided data based on evidence of tibial physis closure. Appendix C.10 supplies the summary findings for 6 studies of outcomes more commonly associated with ACL injury risk: NW, NWI, ACL size (CSA and width), and tibial slope.^94–99  Additional condylar and ACL dimensions obtained in the search are listed in Supplemental File 2 (available online at https://doi.org/10.4085/1062-6050-0038.22.S2). Sufficient data were available to perform meta-analyses for NW, NWI, and ACL size by chronological age. Insufficient data were available to examine tibial slope by age and all outcomes by maturational stage.

Measures of ACL size (ACL CSA and width) generally show similar increases in boys and girls until about 11 to 15 years of age, when boys begin to increase more than girls through age 17 to 18.^95,97  The meta-analyses of 3 included studies^95,97,99  indicated greater ACL size in boys by 15 years of age, with the sex difference increasing from 15 to 17 years (Appendix B.17). Measures of notch geometry generally demonstrated similar increases in boys and girls with growth, but a smaller notch size was evident in some older adolescent girls.⁹⁶  Meta-analysis of the absolute NW revealed similar sizes between sexes from age 8 to 12 years; boys had a greater NW from age 13 to 18 years (Appendix B.18).^95,96,98  However, once normalized to condylar width (NWI), no differences were present between sexes throughout the aging process (Appendix B.19).^{94–96,98,99} 

The searches produced 12 articles on sex-specific changes in joint laxity, most often giving values for general joint laxity (GJL) and anterior knee laxity, with 1 article each reporting sex-specific changes in frontal- and transverse-plane laxity¹⁰⁰  and genu recurvatum.¹⁰¹  After an in-depth review of each article, we excluded 4 for our inability to obtain sex-specific stratified data by individual age or pubertal stage from the authors and 2 for single-sex data, leaving 6.^100–106  The authors of 4 studies noted sex-specific changes in laxity by maturational status, with 2 studies supplying data by age (Appendix C.11).

Collectively, investigators demonstrated that although anterior knee laxity tended to decrease with maturation in both boys and girls,¹⁰¹  this decrease was greater in boys than in girls,^101,102,104  ultimately leading to higher values in mature girls versus mature boys in 2 of the 3 studies.^102,104  A meta-analysis performed on the 2 investigations revealed a trend toward greater values in prepubertal boys and greater values in postpubertal girls, but these were not significantly different at any stage (Appendix B.20).^101,104  A general trend of decreasing joint laxity with adolescence also occurred in 1 study each for varus-valgus,¹⁰⁰  internal-external rotation laxity,¹⁰⁰  and genu recurvatum,¹⁰¹  yet these decreases were similar within the age ranges surveyed (8 to 14 years for varus-valgus and rotational laxity, 9 to 18 years for genu recurvatum). General joint laxity was greater in girls than boys across all maturational stages in 1 study¹⁰¹  and increased in girls from prematuration to postmaturation with no changes in boys in another.¹⁰⁶  When these studies were combined for the meta-analysis, prepubertal girls had greater GJL than prepubertal boys, and this sex difference progressively increased from prepubertal to pubertal to postpubertal stages (Appendix B.21).

Eight articles described sex-specific changes in lower extremity alignment. After an in-depth review of each article, we excluded 2 for an inability to obtain sex-specific stratified data by individual age or pubertal stage from the authors, leaving 6.^{101,107–111}  Most researchers (5 articles) observed sex-specific changes in lower extremity alignment with age, whereas a single study provided data by maturational stage (Appendix C.12). Tibiofemoral angle was the most commonly reported variable in the reviewed manuscripts. Other less commonly measured variables were femoral and tibial specific angles and geometries, pelvic angle, standing quadriceps angle, hip anteversion, and tibial torsion.

Studies of valgus knee angulation (tibiofemoral, quadriceps angle) consistently indicated little to no change with age or maturation in girls. Male values may decrease somewhat, ultimately resulting in greater valgus angulation in girls. A meta-analysis of 3 studies by age that captured sufficient data from 10 to 15 years^101,107,108  displayed similar values in boys and girls until age 12, when girls developed increasing values that became significantly different from boys at age 15 (Appendix B.22). Of the 1 study involving tibiofemoral angle by maturational stage, decreases were most pronounced from pubertal to postpubertal stages, and girls had higher values across all maturational stages.¹⁰¹  Due to the varied methods used to characterize frontal-plane femoral angulations, we did not attempt to combine these variables into a meta-analysis. However, studies of the bicondylar angle of the femur (frontal-plane angle between the long axis of the femoral shaft and a line tangent to the distal femoral condyles) and the collodiaphyseal angle (frontal femoral neck-shaft angle) collectively suggested that even though both angles changed with age in girls and boys, sex differences became apparent around 10 to 12 years of age, when the bicondylar angle increased and the collodiaphyseal angle decreased in girls,^110,111  which could result in a more valgus lower extremity posture. Finally, based on 1 study¹⁰¹  of comprehensive changes in lower extremity alignment with maturational stage, the quadriceps angle increased from prepubertal to pubertal stages in girls and decreased in boys, yielding greater angles in girls in the pubertal and postpubertal stages. Foot pronation and genu recurvatum decreased and tibial torsion and pelvic angle increased similarly across maturation in boys and girls.

The searches identified 15 articles on sex-specific changes in flexibility, 9 of which were excluded because we were unable to obtain sex-specific stratified data by individual age or pubertal stage from the authors (4 articles), reporting on only a single year of age (1 article), reporting on only 1 sex (1 article), and not measuring actual flexibility (3 articles), leaving 6.^{74,84,87,91,112,113}  Five of the 6 studies focused on variations of the sit-and-reach test: 4 addressed changes by age or grade level and 1 addressed changes by maturity level (Appendix C.13).

Boys either decreased or maintained their flexibility with age or maturity, whereas girls tended to maintain or increase their flexibility with age or maturity. Meta-analysis of the sit-and-reach test was limited to 2 investigations^84,112  that included the ages of 11 and 14 due to a lack of uniformity in age range and stratification among studies (Appendix C.13). Girls had greater flexibility than boys at both ages, with the mean difference increasing from age 11 to 14 years (Appendix B.23).

The searches resulted in 34 articles on sex-specific changes in static or dynamic balance. After reviewing each article, we excluded 24 for not meeting the inclusion criteria or our inability to obtain stratified data from the authors. Therefore, 10 studies were included.^{76,77,85,89,114–119}  Seven of these provided static balance by age (Appendix C.14), and 3 explored dynamic balance by age and 1 assessed maturity status (Appendix C.15). Static balance by maturity was not evaluated in any of the papers.

Static balance was generally better in girls across most studies and age groups.^{76,77,116–119}  In addition, static balance generally improved at least up to age 12 in both girls and boys. It should be noted that 4 of these investigations consisted only of children up to the age of 12 or 14 years and so barely met the inclusion criteria of including at least 2 potential maturity categories.^{76,89,116,118}  Nonetheless, our meta-analysis showed better overall static balance in girls, with this difference becoming greater from 14 to 17 years and significantly different at 15 and 16 years, when most girls were nearing the end of puberty and most boys had at least started puberty (Appendix B.24). However, given the small number of articles, most of which did not supply data across the entire age range, we combined eyes-open, eyes-closed, and anterior-posterior and medial-lateral sway under the assumption of independence and because the forest plots with or without ≥1 of the directions or conditions did not change directionality. Pletcher et al⁷⁷  described static balance by age group, with the ages of 10.8 and 16.8 years roughly corresponding to prepubertal and postpubertal children. If we consider their results in the context of maturity, postpubertal boys had worse balance than prepubertal boys, and girls outperformed boys at both time points.

The few studies of dynamic balance by sex and age were limited to ages 11 to 15 years. Our meta-analyses of 3 of these articles^85,114,115  demonstrated a difference that began favoring boys at age 12 and continued through age 15 (Appendix B.25). Researchers of a single study⁸⁵  reported dynamic Y-balance test scores by maturity status. No statistically significant differences were found, but boys improved postpuberty, whereas girls performed less well than their prepubertal and pubertal counterparts.

A total of 35 articles evaluated sex-specific changes in lower extremity biomechanics. After review of each article, we excluded 22 for not meeting the inclusion criteria. Therefore, 13 articles were included in the current review.^{74,77,86,88,120–128}  Lower extremity biomechanics were assessed using both 2- and 3-dimensional motion-capture techniques during a variety of tasks, including the drop vertical jump, stop jump, cutting, and unanticipated cutting. Due to this variability, discrete changes in these variables were not provided in the summary tables. Four studies classified participants based on chronological age, 11 by maturational status, and 2 by percentage of predicted adult stature. Within those that classified maturational status, investigators used different tools to estimate pubertal development and stratify their data (ie, Tanner stages I + II, III + IV, V versus Tanner stages I, II + III, IV + V). For the purpose of the meta-analyses, subgroups of prepubertal (Tanner stage I), pubertal (Tanner stages II to IV), and postpubertal (Tanner stage V) participants were operationally defined.

Measures of knee-abduction kinematics and kinetics were the most frequent outcomes, appearing in 11† and 4^{121,123,126,127} studies, respectively (Appendix C.16). During dynamic tasks, boys typically demonstrated no maturation-related change in knee-abduction kinematics to slightly decreased knee-abduction angles and motion throughout maturation, while girls consistently displayed increased knee-abduction angles and motion throughout maturation.^{74,88,121,122,125,128}  Meta-analyses indicated that girls exhibited greater knee-abduction angles during dynamic tasks than boys at age 10 to 14 years (Appendix B.26), though no differences were seen in the maturational analysis (Appendix B.27). Knee-abduction moment was higher in postpubertal girls than boys, with no difference during the prepubertal and pubertal stages (Appendix B.28). During maturation, the knee-abduction moment decreased slightly in boys^121,123,127  and consistently increased in girls.^{121,123,126,127}  Consistent sex differences in knee-abduction moment emerged in postpubertal participants¹²¹  and those who had reached >92% of adult stature.¹²³ 

Measures of knee-flexion kinematics and kinetics were reported in 5 articles^{74,77,121,124,128}  and 1 article,¹²¹  respectively (Appendix C.17). Throughout maturation, most authors found minimal changes in knee-flexion kinematics of male participants and relatively consistent evidence of decreasing knee flexion in girls,^74,121,128  especially after the age of 14¹²⁸; however, the meta-analyses by maturation status did not reflect any sex differences (Appendix B.29).

Peak vertical ground reaction force (vGRF) during athletic movements was evaluated in 3 investigations (Appendix C.18).^74,77,86  Generally, mass-normalized vGRF decreases throughout maturation in boys with no corresponding change in girls^74,86 ; however, the meta-analysis (Appendix B.30) by subgroup showed no differences among prepubertal, pubertal, or postpubertal boys and girls.

An overall graphical summary of the individual trajectories for girls and boys and the sex differences in these trajectories for all outcomes by age and maturity level, respectively, based on meta-analysis data, is given in Appendices B.31 and B.32. The data were insufficient to graph trajectories for tibial slope, anterior knee laxity, GJL, flexibility and knee-flexion angle, knee-abduction moment, and VGRF by age and NWI, ACL size, tibial slope, tibiofemoral angle, flexibility, and balance by maturity.

Our primary goals were to characterize and compare (1) the sex-specific changes in individual physical risk factors with chronological age and maturity level to better understand the earliest point when sex differences in physical risk factors begin to emerge, (2) the timing and sequencing of these physical changes relative to one another, and (3) how the development of these factors coincides with the time points linked to the increase in ACL injury risk in girls versus boys (approximately 12 to 17 years). Our results indicated marked physical changes occurring throughout adolescence in both boys and girls. Sex differences in FFM, leg strength and power, and GJL were already evident in individuals at 8 to 10 years or in Tanner stage I (or both). Sex differences in FM, FFM, leg strength and power, GJL, and balance became increasingly evident by 11 to 12 years and when transitioning from prepubertal to pubertal stages. Other factors more often emerged at later ages (ACL size and tibiofemoral angle by approximately 13 to 15 years) or when transitioning from pubertal to postpubertal stages (anterior knee laxity, knee-joint biomechanics). When we qualitatively compared sex-specific changes by chronological age versus maturity status, the demarcation in sex-specific trajectories over time tended to be more apparent when examined relative to pubertal status than chronological age, with the greatest changes emerging around Tanner stages III to IV.

Sex differences in body composition and muscular strength and power were among the first to appear and were already present at the earliest ages studied and before the time when the ACL injury risk begins to rise.

Multiple researchers^10,129,130  identified higher BMI (weight by stature) as a risk factor for ACL injury, particularly in girls. Despite these findings, the rationale for including BMI in multivariate risk factor models and how it might be theoretically associated with ACL injury is rarely addressed. Although BMI is easy to measure clinically and is commonly used to characterize relative adiposity, it is a poor indicator of body composition in maturing youth because it does not distinguish between the increasing divergence in relative contributions of muscle and fat weight to an individual's overall weight by stature.¹³¹  According to Centers for Disease Control and Prevention reference data,¹³²  BMI increases linearly with age in a similar manner in both boys and girls from 8 to 18 years of age, and data from our meta-analyses demonstrated an increasingly greater proportion of FM in girls versus a greater proportion of FFM in boys after age 12, the same age when ACL injury begins to disproportionately affect girls. By 14 to 15 years of age, girls have accumulated the majority of their FFM (both overall and leg-specific mass) as male FFM continues to rise; female FM continues to rise while male FM plateaus or decreases. When examined relative to pubertal stage, this sex-specific demarcation is evident as early as pubertal stage II, with a clear shift toward greater accumulation of FM (girls) and FFM (boys) by Tanner stage III.^33,36,46,54  In 1 study of age and Tanner stage,⁴⁶  Tanner stage III on average coincided with 13.1 years for girls and 13.6 years for boys (consistent with chronological data), yet the actual ages of participants in this stage ranged from 10 to 15.6 years and 11.7 to 16.1 years in girls and boys, respectively. Thus, if these body composition changes in girls during Tanner stages III to V (driven hormonally to a large extent¹³¹) are associated with their ACL injury risk, then girls who matured at earlier chronological ages could be at greater risk for ACL injury than those who mature at later chronological ages.

Consistent with the sex-specific trajectories in FFM and leg muscle mass (increasing in both sexes), boys and girls exhibited increasing knee-extension and knee-flexion absolute torque values with increasing age and maturation levels through 12 to 15 years of age. This increase was slower in girls, resulting in less quadriceps strength compared with boys by age 11 years. When stratified by pubertal stage, these sex differences were evident between Tanner stages II and IV. The consistent finding of less hamstrings strength in girls as early as age 8 years and Tanner stage II is concerning, particularly because we did not identify appreciable sex differences in leg-specific muscle mass until after 14 years of age. This earlier divergence in hamstrings strength perhaps was a function of reduced muscle quality (greater fatty infiltration) in maturing girls. Greater proportions of FM are already present in girls by age 8 and Tanner stage I (Appendices B.2 and B.3), and greater FM was associated with less muscle density in 9- to 12-year-old girls.¹³³  Although we were unable to find data on sex comparisons of muscle quality in adolescents, less hamstrings muscle density was observed in young female adults than in young male adults, and less muscle density was more strongly associated with less isometric hamstrings strength than with muscle CSA.¹³⁴ 

It should be noted that, with the exception of DiStefano et al,⁷⁴  findings of sex differences were primarily based on absolute torque, and the data were not normalized by body size (mass or height). The increasing sex difference in FFM per body weight (and the earlier acceleration in FFM accumulation in girls with earlier pubertal onset) likely explains the DiStefano et al⁷⁴  results of greater knee-extensor strength in prepubertal girls compared with pubertal boys and postpubertal girls and the overall greater knee-extensor strength in girls at 9 years of age in the meta-analysis for age. The evidence is limited to a single study, but hip strength normalized to body size does not appreciably improve in boys or girls over time.

Similar to strength, leg power tended to increase to a greater extent in boys than in girls and increase linearly with age. With respect to strength development, an earlier (by approximately 1 year) and steeper rise in leg power trajectories appeared in both the age and maturity data (Appendices B.31 and B.32). Individual study data and meta-analyses for maturity level also identified sex differences earlier than quadriceps strength, with greater values for horizontal leg power in prepubertal boys (Appendix B.13–B.16). As maturity progressed, boys demonstrated a linear increase in leg power, while power generation more often plateaued or declined in girls from pubertal to postpubertal stages. The increasing sex difference was particularly notable in studies that included boys and girls past the age of 16 years. Along with changes in body composition, increasing sex differences in the ability to generate power during maturation are likely due to the muscular demands associated with a weight-bearing task. Specifically, as girls and boys move from pubertal (Tanner stage III) to postpubertal stages, girls have increasingly less muscle mass to propel the same body weight as boys. Future authors should also consider if plateaus in strength and leg power increase the risk of ACL injury in pubertal and postpubertal girls. Although muscle strength has not been associated with the ACL injury risk in athletes,^10,130,135  isolated muscle testing does not functionally challenge the system as a weight-bearing jump or hop would. Jump and hop performance has been used extensively to examine associations between suboptimal landing mechanics and ACL injury.^136–138  We are not aware of any researchers who examined horizontal or vertical leg power as a prospective risk factor for ACL injury.¹³⁵  It may also be useful to establish the extent to which neuromuscular training programs specifically maintain or improve leg power in girls into mid- to late-pubertal development and how this may affect injury risk.

These data indicate that while BMI changes similarly in boys and girls, girls have increasingly less leg muscle mass, strength, and power to control the same relative body weight during sport activity as maturing boys, particularly as they transition from Tanner stages III to V. In future risk factor studies, as opposed to overall BMI, the FMI and FFMI should be partitioned (ie, FM and FFM adjusted for change in stature) to better elucidate the relationship between body composition and ACL injury risk. More specifically, it is important to determine mechanistically how the lesser proportion of muscle mass per unit body weight influences girls' lower extremity strength and power and, ultimately, neuromuscular control strategies and the internal and external loads placed on the ACL.

Appreciating that both girls and boys can augment muscle development during maturation through training,¹³⁹  it is plausible that girls can increase or extend the trajectory of their muscle development and lessen the observed gaps in body composition and muscular strength and power. From this perspective, comparisons of leg muscle mass and strength and power trajectories in athletic and nonathletic populations throughout the entire adolescent growth period (ie, through age 18 years) would be beneficial. As body composition is one of the first factors to change with maturation and some sex differences in strength and power are already evident in prepubertal children, early strength training interventions could be warranted. These sex-specific developmental changes are also important considerations once injury occurs. Quadriceps strength and vertical and horizontal jump performance are used as metrics in determining readiness to return to sport after ACL reconstruction,^140,141  and it is well established that the rate at which girls return to sport after ACL reconstruction is slower than in boys.^142,143  Future researchers should consider sex-specific rehabilitation protocols to address these inherent decrements in leg power in girls and determine if this will enable girls to return to sport more effectively and safely.

Smaller ACLs and notch dimensions and steeper lateral tibial slopes in the contralateral knee were described in studies focused on younger (high school and college-aged) individuals who sustained an ACL injury.^10,144,145  Smaller ACLs are associated with less linear stiffness,^146,147  lower load at failure,^146,147  and greater anterior knee laxity.¹⁴⁸  In turn, greater anterior knee laxity was identified as a strong independent predictor of the ACL injury risk in females.^10,130  Based on the limited evidence available, male participants increased their ACL size and NW (with no difference in NWI) and decreased their lateral tibial slope and anterior knee joint laxity to a greater extent than female students as they matured (Appendices C.10 and C.11; Appendices B.17–B.19).

Although ACL size was not normalized to body dimensions in the included studies, prior work suggested that female adults still had 25% to 30% smaller ACLs, even after body dimensions such as body mass, BMI, and NW were accounted for.^31,149,150  Some data indicated that the sex difference in ACL size could be partly explained by sex differences in muscle size.^31,32  When comparing the sex-specific trajectories in thigh muscle mass with those of ACL size (see Supplemental File 1, Figures 19 and D1), girls accumulated thigh muscle mass and increased ACL size at lower rates (with earlier plateaus) than boys, with sex differences becoming increasingly apparent by approximately 14 years of age (data not available by maturity level). Although we cannot change bone geometry, these findings are potentially promising in that ACL size (and, in turn, anterior knee laxity) could be modifiable to some extent if addressed early. This once again points to the need for continued research examining how these risk factors change relative to one another during pubertal growth to determine what contributes to a smaller and weaker ACL and to examine the effect of early strength training interventions on these outcomes in developing girls.

Adult females were often observed to have greater anterior knee laxity than men,^10,151,152  and this was thought to be mediated to an extent by sex differences in ACL size.¹⁴⁸  From the available data, it is difficult to compare the relative trajectories of ACL size reported by age and anterior laxity reported by pubertal status, other than to note that both trajectories change more in males; sex differences develop at later ages and maturational stages compared with body composition, strength, and leg power. In fact, although all included studies consistently demonstrated greater declines in anterior knee laxity in boys versus girls throughout maturation, sex differences were not always present by late maturational stages within the age ranges assessed. Also, not all girls develop greater magnitudes of knee laxity, and individual variations in other physical factors (eg, sex steroid hormones, lower extremity alignment) that are also changing during this time may contribute to this variability.^152,153  Given the importance of anterior knee laxity as an independent ACL injury risk factor, understanding the factors that promoted the development of greater anterior knee laxity in maturing girls at the individual level is an important direction for future research.

Findings on sex-specific changes in flexibility, genu recurvatum, GJL, and frontal- and transverse-plane knee laxity were more limited. Based on the studies available, frontal- and transverse-plane knee laxity seem to progressively decrease in boys and girls, with no sex differences emerging by 14 years of age (data were not available after this age). Flexibility and GJL either decrease or maintain with increasing age and maturational stage in boys, whereas girls more often increased their values with maturation. This resulted in girls having greater flexibility and GJL as early as age 11 years and Tanner stage I, and this difference became more pronounced at later ages and maturational stages. Greater GJL and genu recurvatum were associated with a greater risk of ACL injury,^10,130,154  yet our knowledge of flexibility relative to ACL injury and prevention is incomplete and equivocal, with 1 study¹⁰  showing a trend toward greater preinjury sit-and-reach scores in patients with ACL injury compared with an uninjured cohort and another¹⁵⁵  suggesting that more emphasis on static stretching in ACL injury-prevention programs may reduce the risk. Further investigations are needed to determine the sequencing of these changes with other risk factors and the timing of ACL injury.

Although sex differences in lower extremity alignment have been examined as ACL injury risk factors, we identified very few studies that examined sex-specific changes in lower extremity alignment during the adolescent years when ACL risk is rising. Given the paucity of research, our analyses were primarily limited to the assessment of frontal-plane knee angles. Salenius and Vankka¹⁵⁶  reported that the natural progression of tibiofemoral angle was a reduction in valgus angulation in children from 2 to 8 years of age. Our meta-analysis suggests a continued progression that is more pronounced in boys throughout adolescence with clear sex differences emerging around 15 years of age. Although greater static frontal alignment measurements have yet to emerge as important predictors of ACL injury^130,157  and are largely nonmodifiable, it is important to understand the development of these alignment patterns and how they may influence other risk factors thought to be associated with ACL injury (eg, dynamic movement patterns¹⁵⁸), as they may modify our approach to interventions.

Studies included in this review indicated that balance generally improves at younger ages and then remains stable. Our findings showed that static balance tends to be better overall in girls (less excursion or sway), with this difference more conclusive at ages 15 and 16 years. Conversely, dynamic balance (greater reach distance) seems to favor boys starting around the age of 12. These developmental trends could benefit girls, as poor static balance (greater postural sway) is associated with a greater risk of ACL injury.¹³⁵  Dynamic measures such as the Y-Balance or Star Excursion Balance tests have not been specifically associated with ACL injury, but greater reach distances were associated with lower limb injury in general.¹³⁵  Additionally, ACL injury-prevention programs that included more balance training were linked with a higher risk of ACL injury.¹⁵⁵  As such, the clinical implications of developmental changes in balance relative to ACL injury risk and prevention are less clear and require further research. Moreover, these studies were primarily based on age and involved younger age ranges: 4 of the 7 demonstrated static balance up to the age of 12 or 14 years (encompassing the early years of ACL risk development), and none included all ages (8 to 18 years). Many of the participants were likely still prepubertal or early pubertal, thus limiting any inferences regarding changes across the full range of maturity.^20,21,159  Moreover, improved balance was not observed until the age of 15, when most girls were near the end of puberty and boys were still developing (Appendix B.24). In this single study,⁸⁵  the age span was limited to 12 to 15 years, which yielded a disproportionate number of prepubertal boys and postpubertal girls. Despite this, multiple sex-specific anthropometric (eg, height, BMI) and performance (eg, countermovement jump height, strength) measures were observed to change with maturation, whereas Y-Balance test results did not. Further work is needed to determine if meaningful changes in static and dynamic balance occur across the maturation continuum and how these may be affected by or affect the development of other physical risk factors known to vary by sex.

Based on the available evidence from individual studies, knee kinematics and kinetics change very little during maturation in boys performing dynamic tasks, whereas girls increase knee-abduction angles and moments and decrease knee-flexion angles throughout maturation (prepubertal to pubertal to postpubertal). Conversely, mass-normalized vGRF decreases in boys and shows little change in girls. The meta-analysis confirmed greater knee-abduction angles in girls versus boys from 11 to 14 years of age (when the ACL injury risk begins to rise) and in the postpubertal stage (Appendices B.26 and B.27). Similar trends were identified in knee-abduction moment with maturation in 3 of the included studies.^121,123,160  This was an important finding given that knee-abduction moment had been prospectively identified as a risk factor for ACL injury in female athletes.¹³⁶  Furthermore, the general consensus is that these collective biomechanical changes pose a greater risk for ACL injury, and improvements in these motion patterns were a primary focus of ACL injury-prevention efforts.¹⁵  The implications of these high-risk movement strategies for ACL loading are particularly concerning for the female ACL, which is proportionally smaller with advancing maturation (Appendix C.10; Appendix B.17).

When considering the overall timing of these changes with respect to other physical changes, our review suggested that sex differences in multiple physical risk factors may precede or coincide with these biomechanical changes and potentially contribute to the development of higher-risk knee biomechanics in girls. Earlier changes observed in body composition (already present at age 8 years and Tanner stage I), leg strength, and power may decrease a girl's ability to stabilize the hip and knee upon landing. In fact, improvements in knee-flexion motion and reductions in knee abduction during the landing phase of a vertical jump were reported after neuromuscular training designed to improve leg strength and power.¹⁶¹  Although perhaps more difficult to modify, the subsequent sex divergences in frontal-plane knee alignment, ACL size, and knee-joint laxity that become increasingly apparent from pubertal to postpubertal stages may further contribute to the increasing sex difference in high-risk movement patterns from pubertal to postpubertal stages (Appendix B.28). Girls with greater magnitudes of knee laxity landed with greater muscle activation, knee stiffness, and valgus knee motion,^162,163  indicating greater demands to control joint motion when strength capabilities are already disadvantaged. Those with lower extremity alignments characterized by a more rotated and valgus knee posture also displayed more functional knee valgus during landing.¹⁵⁸  A better understanding of these underlying contributions is important, as integrated neuromuscular training programs designed to target these risk factors could be more successful if implemented before these risk factors emerge.^164–166  Screening protocols to identify the earliest emergence of both modifiable and nonmodifiable risk factors and how they contribute to high-risk biomechanics should continue, with attention to the role that maturation and pubertal growth may play. This will lead to continued advancement of targeted interventions to reduce the risk of ACL injury.

As previously noted, ACL incidence data were primarily based on chronological age, which is not an ideal metric for understanding the development of individual risk factors, given the large variations in ages at which boys and girls begin to mature and progress through puberty.^16,17  Thus, we sought to identify studies that also addressed physical risk development by maturational status in an effort to better clarify when sex differences begin to occur at the individual level and how these may coincide with age-related changes in ACL injury risk. Our search revealed that the majority of risk factors were either mostly described by chronological age (eg, balance, knee-joint geometry, lower extremity alignment) or maturational status (joint laxity, lower extremity biomechanics), making it difficult to qualitatively compare age-related changes with the stages of maturation. However, we did obtain sufficient data on sex-specific trajectories in body composition, muscle strength and power, and, to a lesser extent, lower extremity biomechanics (ie, knee-abduction angle across limited ages) by both chronological age and pubertal status (see Appendices B.31 and B.32). These data seem to support a clearer demarcation (and more consistent reporting) of emerging sex differences by pubertal status. We often had to collapse stages into prepubertal, pubertal, and postpubertal for the meta-analyses due to study inconsistencies in defining pubertal stage, yet individual studies suggested that Tanner stage III and the transition from Tanner stage III to IV were the most sensitive markers of change for these factors, rather than a particular age. Tanner stage III and the transition to stage IV are known to be associated with considerable physiological changes in maturing girls and boys. Tanner stages III and IV represent the time when girls experience an appreciable rise in estradiol levels and most reach menarche and peak growth.^16,20  For boys, increases in testosterone and the time of peak growth most often occur in Tanner stage IV and the transition to stage V.^17,21  Longitudinal investigations assessing the relative timing, tempo, and dependency of collective risk factor changes across discrete pubertal stages (ie, individual Tanner stage, individual ages relative to PHV) in boys and girls would enable us to better determine when risk begins to develop at the individual level. This is particularly critical because the speed at which a boy or girl passes through these stages varies by individual. Additionally, it should be noted that hormonal changes within and across pubertal stages could vary substantially in both boys and girls.^16,17  Because sex hormones can both directly and indirectly influence many of these physical outcomes, future authors should also examine the effects of this individual variability on physical risk-factor development relative to pubertal stage, as well as the potential effects of early versus delayed pubertal onset.

The following limitations expose the critical gaps that remain in our knowledge and the need for continued research in this area.

Given the scope of the review and to ensure that we obtained representative data on all relevant risk factors, we chose not to limit the publication dates for included studies. Although age at menarche and the timing of hormonal surges have remained relatively stable for both boys and girls, research suggested that the age of thelarche (onset of puberty in girls) had decreased by 0.24 years every decade (approximately 3 months).^167,168  This would likely have had little effect on our findings, as the majority (93%) of the included studies were published in 2000 or later, which would amount to a shift in the data of 6 months or less. Also, the effect of earlier pubertal onset on physical development is uncertain as it is not accompanied by the earlier onset of hormonal changes. Still, these secular trends in age of pubertal development again speak to the need to study trends in risk factor development by maturity stage rather than by age.

Other limitations were based on the characteristics of the studies included. These were largely cross-sectional in nature and often limited to relatively small samples sizes that were not balanced across age groups. Some outcomes were historically stratified by chronological age and others by sexual maturity or growth trajectories. Studies stratified by chronological age were more often based on general populations, whereas those of sexual maturity were more often based on athletic populations. The extent to which physical activity may have collectively influenced these developmental risk factors is unknown. Other study characteristics restricted which studies could be combined in the meta-analysis. Studies that did not provide SDs or sample size per group could not be included in the meta-analyses. Investigations often did not include the entire age range or stages of pubertal development, or data were collapsed across ages or maturational stages in an inconsistent manner. This resulted in insufficient data in some age and maturity groups for drawing inferences across the entire maturation process. We attempted to mitigate this limitation by allowing an age or maturity stage subgroup to be included with only 2 studies supplying data. However, conducting a meta-analysis with only 2 studies in each subgroup was also problematic because at least 3 studies were required to conduct bias analyses, and this factor should be considered when interpreting these results.

Sex differences in physical characteristics often associated with a girl's greater risk of ACL injury in large part emerge between 11 and 17 years of age, when the ACL injury risk is rising more rapidly in girls than in boys. During this transition, sex differences in body composition emerge first, closely followed by leg strength and power (with differences in these outcomes and GJL already evident by the prepubertal stage). Sex differences in knee anatomy, knee-joint laxity, and lower extremity biomechanics follow and more often emerge between pubertal and postpubertal stages. Our collective findings suggested that initiating interventions as early as 8 to 10 years of age (Tanner stage II) may be beneficial to optimize lean muscle development, strength, and power and potentially affect the subsequent sex-specific development of ACL geometry, joint laxity, and neuromuscular control. Yet considerable gaps remain in our understanding of the specific timing and tempo of these collective changes within an individual relative to the stage of pubertal development and how the timing of these developmental changes coincides with the risk of ACL injury.^16,17  Longitudinal studies that simultaneously examine multiple risk factors across discrete stages of the entire maturation process (eg, individual Tanner stages I to V) would greatly improve our knowledge of the relative timing and tempo of ACL risk factor development. In turn, this understanding would allow us to more accurately identify, at the individual level, the earliest entry point for screening and intervening on relevant risk factors. Our hope is that this review will serve as an effective catalyst to encourage future research in this area.

S.J. Shultz was responsible for the original concept and oversight of the entire project. All co-first authors met regularly to contribute to the design and conduct of the systematic review, with each responsible for the in-depth review, data extraction, summary analysis, and reporting for 1 or more risk factor sections (Body Composition [S.J. Shultz], Strength [B. Pietrosimone], leg power [E. Casey], Knee-Joint Geometry and Laxity and Lower Extremity Alignment and Flexibility [R.J. Schmitz], Balance [T. Dompier], and Hip and Knee Biomechanics [J.B. Taylor, K.R. Ford]). T. Dompier performed all meta-analyses. All authors contributed to the drafting, review, and final approval of the manuscript.

Supplemental File 1. Forest plots with individual study results and weightings. Found at DOI: https://doi.org/10.4085/1062-6050-0038.22.S1

Supplemental File 2. Additional knee-joint geometry measures reported. Found at DOI: https://doi.org/10.4085/1062-6050-0038.22.S2

(((1[UID] OR (fat[All Fields] AND free[All Fields] AND (molecular weight[MeSH Terms] OR (molecular[All Fields] AND weight[All Fields]) OR molecular weight[All Fields] OR mass[All Fields])) OR (fat[All Fields] AND (molecular weight[MeSH Terms] OR (molecular[All Fields] AND weight[All Fields]) OR molecular weight[All Fields] OR mass[All Fields])) OR (anterior cruciate ligament[MeSH Terms] OR (anterior[All Fields] AND cruciate[All Fields] AND ligament[All Fields]) OR anterior cruciate ligament[All Fields] OR ACL[All Fields]) OR (anterior cruciate ligament[MeSH Terms] OR (anterior[All Fields] AND cruciate[All Fields] AND ligament[All Fields]) OR anterior cruciate ligament[All Fields]) OR ((femor[All Fields] OR femorals[All Fields] OR femur[MeSH Terms] OR femur[All Fields] OR femoral[All Fields]) AND (notch[All Fields] OR notch s[All Fields] OR notched[All Fields] OR notches[All Fields] OR notching[All Fields] OR notchings[All Fields] OR notchs[All Fields]) AND (width[All Fields] OR widths[All Fields])) OR ((notch[All Fields] OR notch s[All Fields] OR notched[All Fields] OR notches[All Fields] OR notching[All Fields] OR notchings[All Fields] OR notchs[All Fields]) AND (width[All Fields] OR widths[All Fields]) AND (abstracting and indexing[MeSH Terms] OR (abstracting[All Fields] AND indexing[All Fields]) OR abstracting and indexing[All Fields] OR index[All Fields] OR indexed[All Fields] OR indexes[All Fields] OR indexing[All Fields] OR indexation[All Fields] OR indexations[All Fields] OR indexe[All Fields] OR indexer[All Fields] OR indexers[All Fields] OR indexs[All Fields])) OR (intercondylar[All Fields] AND (notch[All Fields] OR notch s[All Fields] OR notched[All Fields] OR notches[All Fields] OR notching[All Fields] OR notchings[All Fields] OR notchs[All Fields])) OR ((tibia[MeSH Terms] OR tibia[All Fields] OR tibial[All Fields] OR tibialization[All Fields] OR tibially[All Fields] OR tibials[All Fields]) AND (slope[All Fields] OR sloped[All Fields] OR slopes[All Fields] OR sloping[All Fields])) OR ((knee[MeSH Terms] OR knee[All Fields] OR knee joint[MeSH Terms] OR (knee[All Fields] AND joint[All Fields]) OR knee joint[All Fields]) AND (laxities[All Fields] OR laxity[All Fields])) OR (balance[All Fields] OR balanced[All Fields] OR balances[All Fields] OR balancing[All Fields]) OR (postural balance[MeSH Terms] OR (postural[All Fields] AND balance[All Fields]) OR postural balance[All Fields] OR (postural[All Fields] AND control[All Fields]) OR postural control[All Fields]) OR ((lower extremity[MeSH Terms] OR (lower[All Fields] AND extremity[All Fields]) OR lower extremity[All Fields]) AND (align[All Fields] OR alignability[All Fields] OR alignable[All Fields] OR aligned[All Fields] OR alignement[All Fields] OR aligner[All Fields] OR aligners[All Fields] OR aligning[All Fields] OR alignment[All Fields] OR alignments[All Fields] OR aligns[All Fields])) OR ((knee[MeSH Terms] OR knee[All Fields] OR knee joint[MeSH Terms] OR (knee[All Fields] AND joint[All Fields]) OR knee joint[All Fields]) AND (strength[All Fields] OR strengths[All Fields])) OR ((hip[MeSH Terms] OR hip[All Fields]) AND (strength[All Fields] OR strengths[All Fields])) OR ((lower extremity[MeSH Terms] OR (lower[All Fields] AND extremity[All Fields]) OR lower extremity[All Fields]) AND (biomechanical phenomena[MeSH Terms] OR biomechanic[All Fields] OR biomechanics[All Fields] OR biomechanical[All Fields] OR biomechanically[All Fields]))) AND (maturate[All Fields] OR maturated[All Fields] OR maturating[All Fields] OR maturation[All Fields] OR maturational[All Fields] OR maturations[All Fields] OR maturative[All Fields] OR mature[All Fields] OR matured[All Fields] OR maturer[All Fields] OR maturers[All Fields] OR matures[All Fields] OR maturing[All Fields] OR maturities[All Fields] OR maturity[All Fields] OR (puberty[MeSH Terms] OR puberty[All Fields] OR puberties[All Fields]) OR (sexual development[MeSH Terms] OR (sexual[All Fields] AND development[All Fields]) OR sexual development[All Fields])) AND (sex[MeSH Terms] OR sex[All Fields] OR (gender identity[MeSH Terms] OR (gender[All Fields] AND identity[All Fields]) OR gender identity[All Fields] OR gendered[All Fields] OR gender s[All Fields] OR gendering[All Fields] OR genderized[All Fields] OR genders[All Fields] OR sex[MeSH Terms] OR sex[All Fields] OR gender[All Fields]))) NOT (obeses[All Fields] OR obesity[MeSH Terms] OR obesity[All Fields] OR obese[All Fields] OR obesities[All Fields] OR obesity s[All Fields] OR (obeses[All Fields] OR [MeSH Terms] OR “obesity[All Fields] OR obese[All Fields] OR obesities[All Fields] OR obesity s[All Fields]))) AND ((english[Filter]) AND (child[Filter] OR adolescent[Filter]))