Prevention of non-contact anterior cruciate ligament injuries in soccer players. Part 1: Mechanisms of injury and underlying risk factors
- First Online:
- Cite this article as:
- Alentorn-Geli, E., Myer, G.D., Silvers, H.J. et al. Knee Surg Sports Traumatol Arthrosc (2009) 17: 705. doi:10.1007/s00167-009-0813-1
- 7.8k Views
Soccer is the most commonly played sport in the world, with an estimated 265 million active soccer players by 2006. Inherent to this sport is the higher risk of injury to the anterior cruciate ligament (ACL) relative to other sports. ACL injury causes the most time lost from competition in soccer which has influenced a strong research focus to determine the risk factors for injury. This research emphasis has afforded a rapid influx of literature defining potential modifiable and non-modifiable risk factors that increase the risk of injury. The purpose of the current review is to sequence the most recent literature that reports potential mechanisms and risk factors for non-contact ACL injury in soccer players. Most ACL tears in soccer players are non-contact in nature. Common playing situations precluding a non-contact ACL injury include: change of direction or cutting maneuvers combined with deceleration, landing from a jump in or near full extension, and pivoting with knee near full extension and a planted foot. The most common non-contact ACL injury mechanism include a deceleration task with high knee internal extension torque (with or without perturbation) combined with dynamic valgus rotation with the body weight shifted over the injured leg and the plantar surface of the foot fixed flat on the playing surface. Potential extrinsic non-contact ACL injury risk factors include: dry weather and surface, and artificial surface instead of natural grass. Commonly purported intrinsic risk factors include: generalized and specific knee joint laxity, small and narrow intercondylar notch width (ratio of notch width to the diameter and cross sectional area of the ACL), pre-ovulatory phase of menstrual cycle in females not using oral contraceptives, decreased relative (to quadriceps) hamstring strength and recruitment, muscular fatigue by altering neuromuscular control, decreased “core” strength and proprioception, low trunk, hip, and knee flexion angles, and high dorsiflexion of the ankle when performing sport tasks, lateral trunk displacement and hip adduction combined with increased knee abduction moments (dynamic knee valgus), and increased hip internal rotation and tibial external rotation with or without foot pronation. The identified mechanisms and risk factors for non-contact ACL injuries have been mainly studied in female soccer players; thus, further research in male players is warranted. Non-contact ACL injuries in soccer players likely has a multi-factorial etiology. The identification of those athletes at increased risk may be a salient first step before designing and implementing specific pre-season and in-season training programs aimed to modify the identified risk factors and to decrease ACL injury rates. Current evidence indicates that this crucial step to prevent ACL injury is the only option to effectively prevent the sequelae of osteoarthritis associated with this traumatic injury.
KeywordsPreventionNon-contact ACL injurySoccer
Soccer is the most commonly played sport in the world , with an estimated 265 million active soccer players participating in the sport as of 2006 . The international popularity continues to rise as indicated by the 23 million increase in active soccer players compared to 8 years ago. Despite the predominance of male players (90%), the current trends suggest that the continued rise in participation is mainly from the increases in females, who choose to participate in soccer . Despite the suggestion that it is considered a relatively safe sport for males, female soccer players are at up to six times greater risk for sustaining an anterior cruciate ligament (ACL) tear than their male counterparts . The increased participation and increased risk of knee injury especially among females, has led to a substantial increase in number of reported ACL injuries in the sport. The reported incidence of ACL injury ranges from 0.06 to 3.7 per 1,000 h of active soccer playing (game and training) [15, 45], accounting for thousands of ACL tears each year. It is also estimated that the occurrence of ACL injuries on a soccer team expressed as a percentage of all injuries on that team is 1.3% for males, and 3.7% for females .
While there has been recent scientific efforts focused on ACL injury treatment strategies, it is well established that surgical reconstruction does not reduce the increased risk for developing knee osteoarthritis after a traumatic knee injury is sustained [42, 106, 120, 134]. In addition, ACL injury is often concomitant with a meniscus tear, and several authors found that this type of meniscus injury is also an indicated risk factor for tibiofemoral osteoarthritis [120, 136]. Beyond the short- and long-term physical impairments, ACL injury also causes personal and professional impairment for athletes, with a high economic cost for both athletes and institutions [56, 57, 194]. Therefore, the prevention of non-contact ACL injuries is of major relevance in sports traumatology.
The identified gender bias for non-contact ACL injuries and associated detrimental effects in female soccer players have served as the impetus for research efforts to define the mechanisms of ACL injury and to delineate the most relevant underlying risk factors that contribute to these mechanisms. It is suggested that once these gender-related mechanisms and associated risk factors are defined more efficient neuromuscular training protocols can be instituted to high-risk populations . The purpose of this article is twofold: first, to provide a current review of the literature to define the most probable mechanisms of non-contact ACL injuries and second, to delineate the role that environmental, anatomical, hormonal, neuromuscular, and biomechanical risk factors may contribute to the portrayed mechanisms.
We employed Medline database for literature search purposes. All articles under the topics “ACL prevention”, “ACL injury risk factors”, “non-contact ACL injuries”, “mechanisms of ACL injuries”, “ACL injuries in soccer players”, and “injuries in soccer” from 1985 to 2008 were considered of potential interest for this review. Articles which included an intervention were excluded from this review. In addition, each reference list from the identified articles was cross-checked to verify that relevant articles were not missed for the current review.
Mechanisms of non-contact ACL injuries in soccer players
The study of mechanisms of non-contact ACL injuries in soccer players is based on several methodological approaches: interviews with injured players, video analysis, clinical studies (where the clinical joint damage is studied to understand the mechanism of the injury), in vivo studies (measuring ligament strain or forces to understand ligament loading patterns), cadaver studies, mathematical modeling and simulation of injury situations, or measurements/estimation from ‘‘close to injury’’ situations [93, 156].
We considered non-contact ACL tears to those injuries with no physical contact with other players at the time of injury. The rate for non-contact ACL injuries ranges from 70 to 84% of all ACL tears in both female and male athletes [17, 45, 117, 137, 138]. Most ACL tears in soccer occur in the absence of player-to-player (body-to-body) contact . Despite Arendt and Dick  found an equal rate of contact versus non-contact mechanisms on ACL injuries among male soccer players, it is overall accepted that the vast majority of ACL tears occur through a non-contact mechanism in both male and female athletes.
Boden et al. utilized retrospective video analysis in attempt to define the most common kinematic positions related to ACL injury during competitive play. They reported a lower extremity alignment associated with non-contact ACL injury in which the tibia was externally rotated, the knee was close to full extension, the foot was planted during deceleration with valgus collapse at the knee . More recent reports have also indicated this common mechanism of valgus collapse at the knee in female athletes [94, 141]. Teitz reported very similar deceleration positions in the majority of the ACL injuries she examined; however, she also indicated that most often the center of mass of the body was behind and away from the base of support (area of foot to ground contact) . Thus, there is mounting evidence that the most common non-contact injury mechanism of injury in female athletes occurs during a deceleration task with high knee internal extension torque (with or without a visual perturbation) combined with dynamic valgus rotation with the body weight shifted over to the injured leg and the plantar surface of the foot fixed flat on the playing surface [17, 94, 141, 177]. Interestingly, both male and female athletes may demonstrate similar body alignment during competitive play without succumbing to an ACL injury. Thus, it is crucial to determine the underlying risk factors that contribute to an increased propensity for this high-risk position. Ultimately, it is the goal of clinicians and researchers to determine the risk factors that preclude the actual ACL injury.
Risk factors have been divided into extrinsic (those outside the body) and intrinsic factors (those within the body) . However, other classification schemes do exist when considering non-contact ACL injuries. In this article, risk factors will be divided into environmental, anatomical, hormonal, neuromuscular, and biomechanical, relative to the guidelines established by the Hunt Valley meeting .
Environmental risk factors
Environmental factors include those aspects extrinsic to the athlete such as sport, playing surface, weather characteristics, the type of footwear, the shoe to surface interaction (friction coefficient). There is a clear lack of randomized controlled studies regarding environmental factors in soccer players. The existing evidence on environmental factors related to ACL injuries is mainly based on American Football, Australian Football, or indoor sports like handball [20, 21, 96, 143, 144, 162]. American Football, Australian Football, and soccer are contact sports sharing some common features regarding ground characteristics, shoe choice, and many playing situations like cutting, landing, or a change of direction with high acceleration and/or deceleration components.
A relationship between meteorological conditions and the incidence of ACL injuries was noted in Australian Football by Scranton et al. . The authors found a higher ACL injury rate on natural grass during dry compared to wet conditions. However, the report did not control for weather conditions where injuries did not occur. Subsequently, Orchard et al.  found that high water evaporation in the month before the match and low rainfall in the year before the match in Australian Football were significantly associated with a higher incidence of ACL injuries. It has been postulated that the high rate of ACL tears during dry conditions on natural grass could be explained by an increased friction and torsional resistance from the shoe-surface interface compared to wet conditions [21, 66]. Similarly, Orchard et al.  found that cold weather was associated with lower knee and ankle injury risk, including ACL tears, in outdoor sports completed on both natural grass and artificial turf. Correspondingly, Torg et al.  demonstrated that an increase in turf temperature, in combination with cleat characteristics, affects shoe–surface interface friction and potentially places the athlete’s knee and ankle at risk of injury. Studies regarding weather conditions may be limited without their control for other potential confounding factors like intrinsic biomechanical factors, neuromuscular conditioning, or hydration status of athletes, among others.
Surface characteristics themselves, irrespective of whether they are influenced by weather conditions, have an impact on ACL injury rates. Orchard et al. found in Australian Football that games and practices played on rye grass appeared to have a lower incidence of ACL tears compared to Bermuda grass. It was hypothesized that Bermuda grass, with a thicker thatch layer, would increase shoe-surface traction secondary to the fact that boot cleats would be better “gripped” by the surface . Also, grass cover and root density has been associated with a greater shoe–surface traction. Artificial surface is generally associated with higher shoe–surface traction than natural grass , and thus with a higher risk for ACL tears. In general, artificial turf has a higher peak deceleration for high-energy impacts , and the greater the surface hardness the greater the ground reaction force. Bowers and Martin  demonstrated that the impact absorption of artificial turf decreases as the age of the surface increases. Additionally, Arnason et al.  found a higher injury rate for soccer played on artificial turf compared to natural grass and gravel, and a higher injury rate on natural grass compared to gravel. Moreover, Hoff and Martin  found a sixfold increase in the injuries reported in indoor soccer compared to outdoor. Both artificial turf and indoor flooring may have an increased coefficient of friction. An increased shoe–surface coefficient of friction or traction may potentially improve performance, but may also increase the risk for ACL injuries. Ford et al.  demonstrated that the playing surface (grass vs. turf) significantly alters plantar loading during cutting in male football players. In addition, Burkhart et al.  reported in a prospective research study that an athlete, who landed with an increased heel to flat-foot loading mechanism was more likely to sustain to a non-contact ACL injury during competitive play. In summary, factors influencing shoe–surface traction include: ground hardness, ground coefficient of friction, ground dryness, grass cover and root density, length of cleats on player boots and relative speed of the game. These factors may contribute to the inciting mechanism of ACL injury . Unfortunately, no definitive conclusions can be drawn regarding the safest type of playing surface in soccer players. Recent studies found no differences in the incidence, severity, nature or cause of injuries in male and female soccer players when comparing artificial turf versus natural grass [40, 54, 55, 173].
Footwear is considered a potential risk factor for ACL tears, since it modulates foot fixation during the game. It has been shown that the number, length and cleat placement was associated with the chance of ACL injuries . Lambson et al. prospectively evaluated ACL injury incidence in American Football depending on the shoe design. The authors found a higher risk of ACL tears for the “edge” cleat design (longer irregular cleats placed at the peripheral margin of the lateral sole with a number of smaller pointed cleats positioned medially). This cleat placement may have provided significantly higher torsional resistance compared to other types of cleats . However, Mitchell et al.  reported that the foot mechanics and possibly the foot–shoe interaction were not related to the propensity to demonstrate high knee load kinematics that are related to increased risk of ACL injury. While there is no current consensus relating the environmental and shoe–surface interaction to risk factors that contribute to ACL injury, the initial evidence reported above suggest that these factors may contribute to the described mechanisms of ACL injury. However, potential confounding factors (i.e., biomechanical, neuromuscular, hydration status, among others) need to be better controlled in environmental risk factors studies. Also, some conclusions are not completely generalizable to soccer players as they were made for Australian Football (i.e., the increased risk for wet conditions and for Bermuda grass compared to rye grass).
Anatomical risk factors
Summary of modifiable and non-modifiable intrinsic risk factors related to increased risk of ACL injury
Modifiable risk factors
Non-modifiable risk factors
Potential control or treatment technique
Monitor and control relative body mass
Femoral notch index (ACL size)
N-M training targeted to decrease other risk factors
N-M training targeted to improve dynamic knee flexion
General joint laxity
N-M training targeted to improve joint stiffness
Family history (genetic predisposition)
N-M training targeted to decrease other risk factors
Prior injury history
Full physical rehabilitation following injury
Developmental and hormonal
N-M training prior to onset of risk factors
Pubertal and post-pubertal maturation status
N-M training prior during pre-puberty
Preovulatory menstrual status
Oral contraceptives in femalesa
ACL tensile strength
N-M training targeted to decrease other risk factors
N-M training targeted to improve neuromuscular control
N-M training targeted to improve coronal plane loads
Anterior tibial shear
N-M training targeted to improve dynamic knee flexion
Lateral trunk motion
N-M training targeted to improve trunk strength and control
N-M training targeted to control transverse motions and influence sagittal plane deceleration mechanics
Dynamic foot pronation
Strength and conditioning training
Ground reaction forces
N-M training targeted to improve force absorption strategies
Relative hamstring recruitment
N-M training targeted to improve hamstring strength and recruitment
Hip abduction strength
N-M training targeted to improve hip strength and recruitment
N-M training targeted to improve trunk strength and control
Some authors have observed increased body mass index as a risk factor for ACL injuries, especially among female adolescent soccer players [24, 71], college recreational athletes [22, 62], and female army recruits . It was postulated that an increased body mass index would result in a more extended lower extremity position with decreased knee flexion upon landing [22, 62]. Unfortunately, conflicting results do exist when completing a further review of the literature, and other authors found no impact of body mass index on ACL injuries in female athletes, including soccer players [54, 55, 91, 95, 146].
Generalized joint laxity is purported as a risk factor that could potentially place the athlete at an increased risk of ACL injury. Soderman et al.  investigated the risk of leg injuries among female soccer players presenting with general joint laxity and knee hyperextension (among other risk factors). The investigators demonstrated a significantly increased risk of leg injuries (but not specific ACL injuries) among athletes with generalized joint laxity and knee hyperextension. Uhorchak  specifically reported a 2.8 times greater risk of non-contact ACL injury in the United States Military Academy cadets with generalized joint laxity compared to normal joint laxity subjects in a prospective 4-year evaluation. There was also a greater risk of non-contact ACL injury for females with a higher anterior–posterior knee joint laxity but not for males. The authors also found a significantly higher generalized joint laxity and anterior–posterior knee joint laxity in females compared to males. It was also retrospectively observed that ACL-injured subjects had a significantly greater generalized joint laxity in comparison to healthy age-matched controls . The same authors report a 78.7% proportion of genu recurvatum among ACL-injured subjects versus the 37% in the control group. Specific knee joint laxity has been related to increased valgus–varus and internal–external rotation knee laxity with an increased functional valgus collapse [51, 72, 167] observed in young female soccer and basketball players in comparison to their male counterparts. Specific knee joint laxity is greater in healthy females compared to males [149, 180], and knee joint laxity measured as hyperextension and anterior–posterior tibiofemoral translation has been recently related to a higher risk of ACL injuries among female soccer and basketball players . Therefore, it seems that knee joint laxity could alter dynamic lower extremity motions and loads a multiplanar fashion, which may place ligaments to a higher risk of rupture. Ergun et al. compared 44 healthy male soccer players (from local leagues) with 44 healthy controls (age- and sex-matched sedentary medical students and hospital staff with no history of regular sports activity) in the sagittal plane for knee laxity and isokinetic muscle strength. Soccer players demonstrated significantly less anterior and anterior–posterior knee laxity and higher isokinetic strength of the knee flexors and extensors compared to sedentary controls. Isokinetic strength difference was found to be higher for the flexor muscle group of the knee . More studies are needed to elucidate the real role of generalized joint laxity and specific knee joint laxity in the risk of ACL tears, specifically controlling for neuromuscular factors.
Pelvis and trunk
The female athlete’s biomechanical profile is a complex system, and the knee joint should not be considered as an isolated component to evaluate risk factors for ACL injury. As a consequence, the trunk, the pelvis, the hip, and the ankle should be considered in their relationship to resultant knee joint mechanics. Anterior pelvic tilt places the hip into an internally rotated, anteverted, and flexed position, which lengthens and weakens the hamstrings and changes moment arms of the gluteal muscles . Hamstring muscles are important to prevent static and dynamic genu recurvatum and to prevent anterior tibial displacement. Gluteal muscles are important to assist hip flexion (gluteus maximus) and to prevent a dynamic valgus collapse (gluteus medius). Anterior pelvic tilt also increases genu valgus and subtalar pronation . Genu recurvatum, excessive navicular drop, and excessive subtalar pronation are more commonly found in ACL-injured subjects compared to non-ACL-injured subjects, all factors that have also been related to ACL preloading . Nevertheless, the exact degree of anterior pelvic tilt that directly correlates to ACL injury remains controversial. It is debated whether the risk is caused by the altered pelvic position itself, or by the functional malalignment it creates . In any case, clinicians should be mindful that pelvic stability is a key factor for lower extremity kinematics and kinetics [199, 200].
Torsional anatomic abnormalities are also related to altered lower extremity biomechanics. Femoral torsion is defined as the angle between the axis of the femoral neck and a transverse line through the posterior aspect of femoral condyles . Femoral anteversion, an increase in the mentioned angle, may cause an inefficiency of the gluteus medius through a decrease in the internal moment arm . A weak gluteus medius may influence dynamic valgus collapse because of the muscles’ inability to keep the hip abducted, especially during weight-bearing activities such as landing, cutting, or changing direction. The toe-in gait demonstrates the femoral torsion position and is often associated with increased external tibial torsion [121, 122], which has been related to the functional valgus collapse at the knee joint .
The influence of pelvis and trunk mechanics on non-contact ACL injuries in soccer players needs to be better characterized. Thus, studies examining the specific role of the pelvis and trunk in the non-contact ACL injuries in soccer players are warranted.
Another suggested anatomical factor that has been related to an increased risk of ACL injury is the quadriceps angle (Q-angle). The Q-angle is the angle formed by a line directed from the anterior-superior iliac spine to central patella and a second line directed from the central patella to tibial tubercle. A high Q-angle may alter the lower limb biomechanics [65, 124] and place the knee at a higher risk to static and dynamic valgus stresses . It was observed that female basketball players with knee injuries had a mean Q-angle greater than non-injured players . However, other authors found that static Q-angle measures do not appear to be predictive of either knee valgus angles, neuromuscular patterns or ACL injury risk during dynamic movement [41, 60, 131]. Likewise, Pantano et al.  demonstrated that peak knee valgus during a single leg squat, and static knee valgus were not significantly greater in young college athletes with higher Q-angle compared to those with lower Q-angle. Subjects with a larger Q-angle, however, had a significantly greater pelvic width to femoral length ratios compared to subjects with a small Q-angle. Pelvic width to femoral length ratios was related to both static and dynamic knee valgus, but static knee valgus was not related to dynamic knee valgus . The authors suggested that pelvic width to femoral length ratios, rather than Q-angle was a better structural predictor of knee valgus during dynamic movement.
Soderman et al.  specifically studied the influence of the Q-angle in female soccer players of second and third Swedish divisions. This study was not specifically assessing ACL injuries, but the Q-angle was not associated with an increased risk of leg injuries. Therefore, the exact role of the Q-angle in the pathomechanics of ACL injuries needs further investigation. At this point, there is not enough evidence to suggest an increased Q-angle as a risk factor for non-contact ACL injuries in soccer players.
Notch width, ACL size and strength
Gender differences have been associated with ACL structural properties. Chandrashekar et al.  found that ACLs in women were smaller in length, cross-sectional area, volume, and mass when compared with that of men. The authors also demonstrated a lower fibril concentration and lower percent area occupied by collagen fibrils in females compared to males. In females, ACL stiffness and modulus of elasticity were highly correlated to fibril concentration, whereas in males ACL failure load and strength were highly correlated to percent area occupied by collagen . Interestingly, ultra structure of ACL has been related to its mechanical properties. Women may have lower tensile linear stiffness with less elongation at failure, and lower energy absorption and load at failure than men [27, 156]. Unfortunately, cadaveric studies may not be generalizable due to the high risk of potential bias. The behavior of in vivo body system is more complex than a cadaveric knee. These studies may be helpful to elucidate future hypothesis on this issue, but caution must be taken with the current conclusions.
A smaller intercondylar notch has been positively correlated to injury risk [165, 180]. Controversy exists when considering femoral intercondylar notch width as a risk factor for ACL injury. It has been shown that a smaller notch size is related to a higher risk of ACL rupture in studies with high number of participants [99, 172]. In less powerful investigations, femoral intercondylar notch width was not related to ACL tears [160, 178]. Despite the lack of relationship between notch width and ACL size , a recent in vivo study reported a significant correlation of the ACL cross-sectional area to the notch surface area . The smaller the intercondylar notch the smaller the cross-sectional area of the midsubstance ACL. The explanation for the increased risk of ACL tear in small notch width subjects is not fully understood, but it has been suggested that an impingement of the ACL at the anterior and posterior roof of the notch may occur during tibial external rotation and abduction [39, 149]. In addition, sex differences in the mechanical properties of ACL reported by Hashemi et al. and Chandrashekar et al. adds more evidence to the small notch-small ACL-increased risk of ACL rupture relationship.
Posterior tibial slope
Posterior tibial slope is not a clear anatomical risk factor for ACL injuries. It was first shown that no relationship was present between non-contact ACL injuries and the caudal (posterior) slope of the tibia . However, in a recent publication, Stijak et al.  found that ACL-injured patients had a significantly greater tibial slope of the lateral tibial plateau and a non-significant lower tibial slope of the medial tibial plateau compared to the control group. Both studies were not conducted with soccer players but with patients. Therefore, further research is needed to determine whether the posterior tibial slope is a risk factor for non-contact ACL injuries in soccer players. Studies assessing posterior tibial slope must control for contralateral tibial slope, intercondylar notch width, knee joint laxity, lower extremity alignment, neuromuscular, and biomechanical characteristics.
Foot pronation and navicular drop have been considered a risk factor for ACL injuries. Beckett et al. established a direct relationship between subtalar joint hyperpronation and ACL tears . The authors compared 50 patients with past medical history of ACL injury and 50 uninjured subjects. The ACL-injured subjects had greater navicular drop test scores than uninjured subjects. Later, Woodford-Rogers et al. compared gymnasts, American Football, and basketball players with history of ACL injury to matched uninjured athletes. They observed a greater subtalar pronation in the ACL-injured group  results that were also found by other authors . Also, Loudon et al. compared 20 ACL-injured females and 20 age-matched controls in a retrospective study design. Seven variables were measured: standing pelvic position, hip position, standing sagittal knee position, standing frontal knee position, hamstring length, prone subtalar joint position, and navicular drop test. As earlier reviewed, knee recurvatum, excessive navicular drop, and excessive subtalar joint pronation were found to be significant discriminators between the ACL-injured and uninjured groups . Unfortunately, foot pronation and navicular drop are not free of controversy, since other authors observed contradictory findings [86, 169]. In a recent study conducted by Jenkins et al. , 105 soccer and basketball players (53 women and 52 men) were recruited and divided into an ACL-normal group and an ACL-injured group. Two measures of foot structure (subtalar joint neutral position and navicular drop test values) were recorded for each subject. No statistically significant differences were found in the foot structure measures between women and men. The authors concluded that values derived from subtalar joint neutral position measurement and the navicular drop test were not associated with ACL injury in collegiate female and male soccer and basketball players. Additionally, Mitchell et al. [73, 123] recently reported that dynamic medial foot loading was not related to increased propensity to demonstrate high ACL-injury risk biomechanics. Subtalar joint pronation creates a compensatory increase in the internal tibial rotation, which has been found to be coupled with internal tibial rotation at the knee during extension [9, 33]. Normally, subtalar joint pronation and tibial internal rotation occur only during the contact phase of gait. If pronation occurs beyond the contact phase, the tibia remains internally rotated, impeding the occurrence of subtalar joint supination and tibial external rotation, which normally occurs as the limb moves through the midstance phase of gait. This excessive internal tibial rotation transmits abnormal forces upward in the kinetic chain . Given a forced movement with planted foot and internal rotation, the preloaded ACL may be placed to a greater stress that may evoke to a rupture. Subtalar joint pronation and internal tibial rotation at the knee may produce an increased internal femoral rotation and valgus angulation at the knee , enhancing the risk of ACL injury.
Similar to those concerns indicated above related to posterior tibial slope as a risk factor, studies assessing the role of foot pronation specifically for soccer players are needed before clear conclusions can be drawn in this population.
Hormonal risk factors
While there has been a significant research focus on sex hormone relationships to ACL injury, the literature provides conflicting evidence, which has prevented a strong consensus to be reached on whether ACL injury risk is associated with specific sex hormone fluctuations. Almost all studies assessing the hormonal risk factor for non-contact ACL injuries involve athletes, although not all of them engaged soccer players. The study of Martineau et al.  found that oral contraceptive use decreased the ligamentous laxity in female soccer players. However, there is not enough evidence at this point to widely recommend oral contraceptive use to prevent non-contact ACL injuries in soccer players. Pilot evidence indicates that it might be a potential control strategy in the future if strongest evidence is provided (Table 1).
It was shown that human ACL cells had both estrogen and progesterone receptor sites . Furthermore, it was suggested that gender differences for ACL tears may be, in part, explained by sex hormones. Specifically, hormonal risk factors are believed to play an important role for non-contact ACL injuries among female athletes. There are three phases of the menstrual cycle: follicular (day 0–9), ovulatory (day 10–14) and luteal (day 15–28). Disparity of results exists concerning the time of the menstrual cycle at which the greatest number of injuries occur: follicular phase [4, 5, 135, 157, 168], around ovulation [1, 14, 187, 188], or the luteal phase . In a recent systematic review, seven studies were pooled in an attempt to determine a potential relationship of the menstrual cycle to ACL injury . The seven reviewed studies favored an effect of the first half, or pre-ovulatory phase, of the menstrual cycle for increased ACL injuries. The six studies that stratified the non-oral contraceptive and oral contraceptive data also favored an effect of the first half of the menstrual cycle for increased ACL injuries. The authors concluded that the clinical relevance of this finding is that female athletes may be more predisposed to ACL injuries during the pre-ovulatory phase of the menstrual cycle, which is consistent with the estrogen surge seen during this phase of the cycle . Not all hormonal studies compared to a control group nor stratified for oral versus non-oral contraceptive use. Both aspects are crucial to test the hypothesis that sex hormones are a potential risk factor for non-contact ACL injuries.
Effects on laxity
Sex hormones have also been related to an increased anterior knee laxity. Zazulak et al.  conducted a systematic review on the effects of menstrual cycle on anterior knee laxity. The authors included nine studies, and they observed that six of them reported no significant effect of the menstrual cycle on anterior knee laxity in women. However, three studies observed significant associations between the menstrual cycle and anterior knee laxity. These studies all reported an increased knee laxity during the ovulatory or post-ovulatory phases of the cycle. A meta-analysis, which included data from the nine reviewed studies, corroborated the significant effect of cycle phase on knee laxity. In the analyses, the knee laxity data measured at 10–14 days was greater than at 15–28 days, which was greater than at 1–9 days. Hicks-Little et al.  also found that the ovulation and luteal phases of the menstrual cycle significantly increased anterior displacement about the knee. Oral contraceptive use was found to decrease the ligamentous laxity in female soccer players  and to lower the traumatic injury rate [3, 125, 187]. In another study, female athletes on oral contraceptives demonstrated decreased impact forces and reduced medial and lateral torques at the knee, increased hamstrings to quadriceps strength ratios, increased stability on one leg and decreased knee laxity relative to non-users . For this sample, the use of oral contraceptives appeared to increase the dynamic stability of the knee joint. These results suggest that hormonal stabilization increases dynamic stability of the female athlete’s knee, and may reduce the risk of serious knee injury in this high-risk athlete . In contrast, others have demonstrated a tendency to increase anterior tibial displacement in oral contraceptive users compared to those not using hormonal replacement therapy .
Effects on ACL tensile strength
Estrogen and progesterone have been found to affect the collagen metabolism in both animal models and humans. Essentially, estrogen (i.e., estradiol) decreased fibroblast proliferation and type I pro-collagen synthesis whereas progesterone levels attenuated estrogen inhibitory effect on collagen metabolism of female ACLs, both in a dose- and time-dependent manner [197, 198]. Controversy also exists due to a disparity of results among animal models. Further research is needed to better elucidate the concentration and time dependency effects of estrogen exposure, as well as of other sex hormones, with respect to the ACL tissue . Sex hormones have also been reported to affect tensile properties of ligaments [18, 92, 193], but other authors found no significant differences in maximum force, stiffness, energy to failure, or failure site of ACLs in sheep . Influence of sex hormones on mechanical properties of ligaments has been only studied in animal models. Further research is also needed to better establish the influence of estrogen and other hormones on biomechanical properties of ligaments.
Effects on neuromuscular function
Neuromuscular function seems to also be affected by sex hormones. During the ovulatory phase, there was an increase in quadriceps strength, a decrease in muscle relaxation time, and an increase in muscle fatigability in young healthy relatively sedentary females . Sex hormones also decrease motor coordination  and have effects on isokinetic strength, anaerobic and aerobic capacity, and high-intensity endurance in female athletes . Interestingly, Chaudhari et al.  investigated knee and hip loading patterns at different phases in the menstrual cycle. The authors compared performance on horizontal jump, vertical jump, and drop from a 30-cm box on the left leg between women (half of them taking oral contraceptive) and men. Men were tested once whereas women were tested twice for each phase of the menstrual cycle (follicular, ovulatory, luteal), and lower limb kinematics (foot strike knee flexion) and peak externally applied moments were calculated (hip adduction moment, hip internal rotation moment, knee flexion moment, knee abduction moment). No significant differences in moments or knee angle were observed between phases in either female group or between the two female groups (oral contraceptive users and non-oral contraceptive users) or between either of female groups and the male controls. The authors concluded that variations of the menstrual cycle and the use of an oral contraceptive do not directly effect knee or hip joint loading during jumping and landing tasks . Because knee and hip joint loading was unaffected by cyclic variations in hormone levels, the observed difference in injury rates was thought to be more likely attributable to persistent differences in strength, neuromuscular coordination, or ligament properties. As stated, sex hormones as a risk factor for ACL injury is an attractive and promising area of research. Nevertheless, there is still equivocal evidence on many topics, and future research is again needed in this area to better prevent, at least in part, many ACL injuries.
Neuromuscular risk factors
Neuromuscular control refers to unconscious activation of the dynamic restraints surrounding a joint in response to sensory stimuli . The neuromuscular system generates movement and determines biomechanics of playing actions. Unconscious muscle activation is crucial during many actions in sport, and differences in neuromuscular control may explain, in part, the increased ACL injury risk exhibited by a certain cohort of soccer players . Olsen et al.  reported that team handball players were often judged by the coaches to be out of balance, and in the majority of cases, some form of perturbation (often contact with another player) appeared to have altered the player’s coordination or intended movement at the time of injury. Landing, cutting, and pivoting maneuvers in some females have been shown to differ from males [51, 52, 115]. Essentially, female soccer players perform playing actions with increased adduction and internal rotation of the femur, reduced hip and knee flexion angles, increased dynamic knee valgus, increased quadriceps activity (with a concomitant decrease in hamstring activity), and decreased muscle stiffness around the knee joint .
Relative strength and recruitment
Dynamic stabilization via the neuromuscular control system helps to protect the knee joint during dynamic sport-related tasks [13, 104, 183–185]. However, muscle actions must be coordinated and co-activated in order to protect the knee joint . Hence, antagonist–agonist relationships are crucial for joint stability. For the knee joint, co-activation of hamstrings and quadriceps may be critical to prevent or to reduce knee motion and loads that increase the risk of ACL injury. Hamstring recruitment reduces ACL loads from quadriceps [155, 185], and may help to provide dynamic knee stability by resisting anterior and lateral tibial translation and transverse tibial rotations .
In vivo studies where a strain gauge was placed into an intact ACL at the time of surgery demonstrated that rehabilitation exercises that produced an isolated contraction of the quadriceps muscle near extension strained the ACL more than exercises with co-contraction of both quadriceps and hamstrings . Specifically, the quadriceps muscle cause peak strain to the ACL around 30° of knee flexion . Women may have an imbalance between muscular strength, flexibility, and coordination within their lower extremities [90, 132]. Deficits in relative hamstring strength may contribute to increased risk of ACL injury in soccer players. Colby et al.  investigated quadriceps and hamstring muscles activation patterns and determined the knee flexion angle during the eccentric motion of sidestep cutting, cross-cutting, stopping, and landing in healthy collegiate and recreational male and female athletes. The results indicated that there is high-level quadriceps muscle activation beginning just before foot strike and peaking in mid-eccentric motion. Hamstring muscle activation was submaximal at and after initial contact. The maximum quadriceps muscle activation for all maneuvers was 161% of the maximum voluntary contraction, while minimum hamstring muscle activity was 14%. Foot strike occurred at an average of 22° of knee flexion for all maneuvers. This low level of hamstring muscle activity and low angle of knee flexion at foot strike during eccentric contraction, coupled with relatively unopposed forces generated by the quadriceps muscles at the knee, could produce significant anterior displacement of the tibia, which may play a role in ACL injury . Chappell et al.  found that female soccer, basketball, and volleyball players prepared for landing with increased quadriceps activation and decreased hamstring activation, which may result in increased ACL loading during the landing of the stop-jump task and the risk for non-contact ACL injury. Also, Padua et al.  found an increased quadriceps and soleus activation during hopping as well as a decreased hamstrings to quadriceps activation ratio in women compared to men (both were active subjects with previous recreational experience in soccer, basketball, and volleyball). Hewett et al. demonstrated that a plyometric training reduced the peak landing forces and increased hamstring torques at landing in female volleyball players . The decrease in landing forces implies that less force is transmitted to the knee articulations and passive structures; therefore, more energy is being absorbed by active muscular restraints. In contrast, weak hamstrings contribute to a greater ground reaction forces that place the ACL at a higher risk of rupture . In addition, adduction and abduction moments at the knee significantly decreased after plyometric training and were the sole significant predictors of peak landing force. A decreased adduction and abduction moment would decrease the risk of femoral condylar lift-off from the tibial plateau . On the other hand, peak landing flexion (reflecting net quadriceps muscle activity) and extension moments (reflecting net hamstrings muscle activity) at the knee did not change after training and were not significant predictors of peak landing force. The plyometric training also increased the hamstring-to-quadriceps muscle ratio by increasing the hamstring muscle peak torque. As reviewed, soccer players demonstrated significantly less anterior and anterior–posterior knee laxity and higher isokinetic strength of the knee flexors and extensors compared to sedentary controls , what adds more evidence on the dynamic stabilizer function of muscles. Moreover, Myer et al.  recently found that female soccer and basketball players sustaining ACL injuries had a combination of similar quadriceps strength with decreased hamstring strength compared to males. In direct contrast, female athletes who did not go on to ACL injury had decreased quadriceps strength and similar hamstring strength compared to matched male athletes. Female soccer and basketball players who demonstrate increased relative quadriceps strength and decreased relative hamstring strength may be at increased risk for ACL injury. Hence, preseason and continued in-season conditioning focused on hamstring strengthening may be indicated for female soccer players, who fall into this high risk unbalanced profile .
Relative joint stiffness and stability
Hamstring muscles are important to decrease anterior shear forces and greatly reduce load on the primary restraint to anterior tibial motion, the ACL [7, 126]. It was found that increasing hamstring muscle force during the knee flexion phase of a simulated jump landing significantly reduced the peak relative strain in the ACL in vitro . Through knee joint compression, hamstrings limit anterior tibial translation by allowing the concave medial tibial plateau to limit anterior drawer  and by allowing more of the valgus load to be carried by articular contact forces, protecting the ligaments . Moreover, hamstring compression could protect against torsional loading, which has been found to be greater for females compared to males [104, 189]. In fact, a vigorous quadriceps contraction has been shown to induce ACL rupture in cadavers . Women demonstrate decreased hamstrings-to-quadriceps peak torque ratios and increased knee abduction (valgus) moments compared to males . Hamstring muscles are activated by ACL receptors when the ligament is placed under stress, what adds more evidence to the hamstrings provide agonistic support to the ACL. It was also suggested that hamstrings are activated by an alternative reflex arc unrelated to ACL receptors . This ACL receptor-dependent muscle activation suggests that decreased proprioception could have an impact on knee stability. The hamstring activation depending on the ability of the ACL to sense a torque and elongation may justify the inclusion of proprioception training in preventive and rehabilitation programs .
Muscles crossing a joint provide stability to that joint. In other words, muscle stiffness, or the resistance to dynamic stretch may protect ligaments from rupture when a load is applied. As reviewed, quadriceps and hamstring muscles provide anterior–posterior joint stiffness. Others suggest that sagittal plane knee joint stiffness is also relevant for ACL injury prevention. Studies demonstrate that female athletes show less muscular stiffness than their male counterparts [58, 59, 67, 79, 88, 161, 186, 189]. Males activate their lower extremity muscles significantly earlier , and have longer activation duration in muscles that initiated and maintained knee (gastrocnemius) and lower extremity stiffness (gluteus) than women . Decreased muscular stiffness in females was shown for both anterior tibial translation [58, 59, 81, 88, 186] and rotational forces [58, 59, 161, 189]. In a recent study, Schmitz et al.  investigated the varus/valgus and internal/external torsional knee joint stiffness in both males and females. Knee kinematics of 20 university students were measured while applying 0–10 Nm of varus and valgus torques with the subject non-weight-bearing, and 0–5 Nm of internal and external torques in both non-weight-bearing and weight-bearing conditions, with the use of a custom joint testing device. When low magnitudes of torque were applied to the knee, women had significantly lower stiffness values than did men. With the exception of applied external torque with the joint weight-bearing and varus torque with the joint non-weight-bearing, women demonstrated an increase in joint stiffness as the magnitude of torque increased from lower to higher magnitudes. In contrast, for the men, joint stiffness values remained unchanged as the magnitude of applied torque increased. The authors concluded that women exhibited lower knee stiffness in response to low magnitudes of applied torque compared to men and demonstrated an increase of joint stiffness as the magnitude of applied torque increased .
Since muscles contribute to joint stability, muscular fatigue might be a risk factor for ligament injuries. Fatigued muscles are able to absorb less energy before reaching the degree of stretch that causes injuries . Better conditioned soccer players may have improved neuromuscular control later in games relative to de-conditioned athletes. This improved neuromuscular control may help athlete to better absorb energy, leaving less energy to be absorbed by other structures such as ligaments . Under fatigued conditions, it was shown that males and females decrease knee flexion angle and increase proximal tibial anterior shear force and knee varus moments when performing stop-jump tasks . Nyland et al.  investigated the effect of quadriceps and hamstrings fatigue from eccentric work on the activation onset of vastus medialis, rectus femoris, vastus lateralis, the medial hamstrings, biceps femoris, and gastrocnemius muscles in healthy female athletes compared with controls directly after performing crossover cut training. The authors demonstrated that quadriceps fatigue from eccentric work produced earlier gastrocnemius and delayed quadriceps femoris activation during crossover cutting in female athlete compared to controls, but activation onset did not differ compared to hamstring fatigue. Neither hamstring nor quadriceps femoris fatigue produced differences in medial hamstring or biceps femoris activation onset compared to controls. The authors concluded that the gastrocnemius muscles act as a synergistic and compensatory dynamic knee stabilizer in a closed kinetic chain situations as the quadriceps femoris muscles fatigue . Conversely, Fleming et al.  demonstrated that the gastrocnemius muscle is an antagonist of the ACL. Six subjects with normal ACLs participated in the study. Subjects underwent spinal anesthesia to ensure that their leg musculature was relaxed. Transcutaneous electrical muscle stimulation was used to induce contractions of the gastrocnemius, quadriceps and hamstrings muscles, while the strains in the anteromedial bundle of the ACL were measured using a differential variable reluctance transducer. The ACL strain values produced by contraction of the gastrocnemius muscle were dependent on the magnitude of the ankle torque and knee flexion angle. Co-contraction of the gastrocnemius and quadriceps muscles produced ACL strain values that were greater than those produced by isolated activation of either muscle group when the knee was at 15 and 30°. Co-contraction of the gastrocnemius and hamstrings muscles produced strains that were higher than those produced by the isolated contraction of the hamstrings muscles. At 15 and 30° of knee flexion, the co-contraction strain values were less than those produced by stimulation of the gastrocnemius muscle alone . Landry et al. [97, 98] demonstrated that elite female soccer players exhibit an increased gastrocnemius activity during unanticipated run, side-cut, and cross-cut maneuvers. Female soccer players demonstrated a higher gastrocnemius activity and a mediolateral gastrocnemius activation imbalance during early stance to midstance of the side-cut and during both run and cross-cut maneuvers that was not present in the male players. Additionally, for unanticipated side-cut maneuvers, female athletes demonstrated greater rectus femoris muscle activity throughout stance, and the only hamstring difference identified was a mediolateral activation imbalance in male athletes only . Moreover, for unanticipated run and cross-cut maneuvers, rectus femoris activity and vastus medialis and lateralis activity for the straight run only were also greater in female than in male athletes . Other notable difference captured for both maneuvers included female players having reduced hamstring activity compared to male players. Padua et al.  also observed greater soleus activation during hopping in healthy women compared to men.
Nyland et al.  further investigated the effects of hamstring fatigue on transverse plane knee control during a running crossover cut directional change (functional pivot shift). The authors found that an eccentric work-induced hamstring fatigue created decreased dynamic transverse plane knee control as evidenced by increased knee internal rotation during impact-force absorption, an earlier peak ankle plantar-flexor moment onset, and a decreased knee internal rotation with propulsion during hamstring fatigue. It was suggested that this pattern may represent compensatory attempts at dynamic knee stabilization from the posterior lower leg musculature during the reportedly ligamentous stressful functional pivot shift phase of the crossover cut . In turn, other authors found an increased anterior tibial translation with muscular fatigue in healthy knees [119, 190]. However, Wojtys et al.  found that the recruitment order of the lower extremity muscles in response to anterior tibial translation did not change with fatigue. Melnyk and Gollhofer  assessed reflex latencies and neuromuscular hamstring activity using surface electromyography. Muscle fatigue produced a significant longer latency for the monosynaptic reflex latencies, whereas no differences in the latencies of the medium latency component were found. Fatigue significantly reduced EMG amplitudes of the short and medium latency components. The authors suggested that a reduced motor activity rather than the extended latency of the first hamstring response is the reason for possible failure. McLean et al.  also investigated the impact of fatigue on ACL injury risk. Ten males and ten females were compared performing a land from a jump. Females landed with more initial ankle plantar flexion and peak-stance ankle supination, knee abduction, and knee internal rotation compared with men. They also had larger knee adduction, abduction, and internal rotation, and smaller ankle dorsiflexion moments. Fatigue increased initial and peak knee abduction and internal rotation motions and peak knee internal rotation, adduction, and abduction moments, with the latter being more pronounced in females. Therefore, McLean et al. concluded that fatigue-induced modifications in lower-limb control may increase the risk of non-contact ACL injury during landings. Gender dimorphic abduction loading in the presence of fatigue also may explain the increased injury risk in women .
Decision-making (i.e., anticipated and unanticipated actions), in addition to fatigue, has been shown in isolation to directly impact ACL injury risk [11, 78, 151]. For example, Besier et al.  examined a sidestep cut at two different angles under both anticipated and unanticipated conditions and found increased varus–valgus and internal–external knee moments during unanticipated movements. The authors suggested that the increased coronal plane torques increased the potential for ACL injuries during unanticipated movements. Lower extremity muscle activation during cutting is significantly different between anticipated and unanticipated conditions . Recently, the effects of a combination of fatigue and decision-making on landing postures were investigated. Borotikar et al.  studied the combined effects of fatigue and decision-making on lower limb kinematics during sports relevant landings. Fatigue caused significant increases in initial contact hip extension and internal rotation, and in peak stance knee abduction and internal rotation and ankle supination angles. Fatigue-induced increases in initial contact hip rotations and in peak knee abduction angle were also significantly more pronounced during unanticipated compared to anticipated landings. It was suggested that the integrative effects of fatigue and decision-making may represent a worst case scenario in terms of ACL injury risk during dynamic single leg landings, by perpetuating substantial degradation and overload of central control mechanisms . Additionally, Olsen et al.  reported that ACL injuries occurred when team handball players were out of balance, or some form of perturbation (often contact with another player) altered the player’s coordination. Laboratory studies do not correlate with field studies at all. Fauno and Wulff Jakobsen  reported that ACL injuries in second half is not statistically different from first half, so fatigue was not seen as a risk factor by the authors. Thus, it appears that fatigue may contribute to other risk factors, but may not in itself be an isolated risk factor for ACL injury.
Biomechanical risk factors
Biomechanics of playing actions are necessary to understand the pathomechanics of ACL injuries and to offer effective prevention programs. It was postulated that hip low forward flexion, hip adduction, hip internal rotation, knee valgus, knee extension, and knee external rotation may place the ACL to a high risk of rupture. It was called the “position of no return” . Biomechanical studies have been conducted in cadavers, in vivo through strain gauges placed at the time of surgery, and from analytical modeling. Biomechanical risk factors for ACL injuries have been described in all three planes. Abundant data exists considering biomechanical risk factors in athletes. Specific biomechanical studies involving soccer players do also exist, especially in females.
Sagittal plane biomechanics have yielded many studies on trunk, hip, knee, and ankle flexion angles when performing sport tasks. The more joints are flexed during landing, the more the energy is absorbed and the less the impact is transferred to the knee. Also, the ACL and hamstring anatomy explain why knee flexion is protective of ACL damage. Every movement with influences over knee flexion can contribute to ACL injury. From proximal to distal, Blackburn and Padua  demonstrated that increased trunk flexion during landing also increased hip and knee flexion angles. The authors found that trunk flexion altered neither transverse nor frontal plane kinematics during the landing task. A less erected posture during landing has been associated with a reduced ACL injury risk [61, 68, 89].
Hewett et al.  reported a significant increased peak external hip flexion moment in ACL injured compared to uninjured females soccer, basketball, and volleyball players, but was not observed to be a significant predictor of ACL injury. These data suggest ACL-injured athletes had an increased internal hip extensor moment due to an increased gluteus maximus activity. Conversely, Decker et al.  suggested that a decreased hip musculature activity may produce a higher ground reaction force, because muscles would be used to absorb energy from a certain task. Landry et al. [97, 98] studies showed that elite female soccer players exhibited a reduced external hip flexion moment and hip flexion angle during unanticipated side-cut, run, and cross-cut maneuvers. Similarly, Zazulak et al.  found a decreased gluteus maximus activity in females during single-legged landings.
Female soccer players demonstrate decreased hip and knee flexion angles at landing compared to male soccer players after the age of 13-year-old . Young female soccer, basketball, and volleyball players also showed decreased hip and knee flexion angles compared to males during the landing preparation of a vertical stop-jump task . Resultant initial contact lower extremity motion patterns during landing of the stop-jump task may be pre-programed just prior to landing. Therefore, female subjects prepared for landing with a decreased hip and knee flexion angle which may result in increased ACL loading during the landing of the stop-jump task and the risk for non-contact ACL injury . It was postulated that a decreased hip and knee flexion angles at landing places the ACL at a greater risk of injury, because a greater peak landing force is transmitted to the knee . Yu et al.  showed that hip and knee flexion–extension angular velocity, rather than angle or joint position, was correlated to the peak posterior and vertical ground reaction forces at landing from a stop-jump task. On average, females landed with greater impact forces and had smaller hip and knee flexion angles at the initial foot contact with the ground and maximum knee flexion angle at the end of the landing. The greater the hip and knee flexion angular velocity at the initial foot contact during the landing of a stop-jump task, the lesser the posterior and vertical ground reaction forces. Also, the greater the peak proximal tibia anterior shear force and peak knee extension moment during landing, the greater the posterior and vertical ground reaction force. Therefore, decreased hip and knee flexion angles at landing as a risk factor for ACL injury may not be resultant from increased ground reaction force . Instead, the increased risk of ACL injury from decreased knee flexion angle could be explained by differences in ACL elevation angle, the angle of insertion of the hamstrings, and by differences in patellar tendon–tibial shaft angle. Near knee extension, the ACL has a greater elevation angles, so the ligament is more perpendicular to a tibial plateau line, whereas the ACL is essentially parallel to the tibial plateau with knee flexion past 90° . This change in orientation influences the load placed on the ACL and its ability to sustain elastic deformation without injury . The structural properties of the ACL are maximized under tensile (longitudinal) loading conditions and minimized under non-axial (shear) loading conditions . As the knee progresses into extension, the ACL elevation angle is maximized. Under this configuration, the anterior tibial shear force generated by the quadriceps/patellar tendon and imparted to ACL is increasingly shear in nature. Conversely, as the ACL elevation angle decreases with knee flexion, the shear component of the resultant ACL force decreases while the tensile component increases reciprocally . Additionally, as the knee progresses into flexion, the angle of insertion of the hamstrings with respect to the tibial longitudinal axis increases such that at knee flexion angles greater than 100°, the resultant hamstring force is directed parallel to the tibial plateau [16, 204]. On the other hand, at lower degrees of knee flexion, the angle of insertion of the hamstrings with respect to the tibial longitudinal axis decreases such that the resultant hamstring force is directed parallel to the ACL, which is placed perpendicular to the tibial plateau, thus limiting the hamstrings’ potential to counteract anterior tibial strain to the ACL. Also, an extended lower limb at landing may strain the ACL due to a greater patellar tendon–tibial shaft angle that increases the anteriorly directed component of the force produced by the quadriceps muscle . As the knee progresses into flexion, the patellar tendon insertion angle with respect to the tibial longitudinal axis decreases . This change in patellar tendon orientation has a profound influence on tibial shear force, as the anteriorly directed component of the quadriceps–patellar tendon force is derived as a multiple of the sine of the insertion angle . Hence, at lower knee flexion angles, the quadriceps exerts a higher anteriorly directed force that is poorly counteracted by both the ACL and the hamstrings. Additionally, the maximum quadriceps force is estimated to be produced around 60° of knee flexion . Therefore, it might be argued that an extended position around 20° of knee flexion may produce less impact absorption through the musculotendinous system increasing the force transmitted to the passive structures of the knee. A large hip and knee flexion angles at the initial foot contact with the ground do not necessarily reduce the impact force during landing, but active hip and knee flexion motions do . It was also demonstrated that females had increased quadriceps activation before landing from the same task compared with male subjects . Yu et al. also found that female athletes exhibited an increased hamstring activation before landing but a trend of decreased hamstring activation after landing compared with male subjects, whereas Krosshaug et al.  showed an increased knee flexion angle both at initial contact and 50 ms after initial ground contact in female basketball players compared to males. Landing preparation with increased quadriceps activation may increase ACL loading during landing, since muscle activation is a major determinant of muscle contraction force . A greater quadriceps activation  and increased knee extension moment  during landing of a variety of athletic tasks in female athletes compared to their male counterparts was also observed by other authors. Hewett et al.  found similar knee flexion angles at initial contact between ACL injured and uninjured female athletes in a prospective study. The peak knee flexion moment was also similar; however, maximum knee flexion angle was lower in the ACL-injured group compared to uninjured controls. Several authors report that isolated sagittal plane forces are not high enough to tear the ACL during sports [114, 150]. In addition, there is no consensus on whether females land or cut with less [80, 109], same [52, 114, 116], or greater [44, 94] knee flexion angles compared to males.
Ankle joint is a key component in any movement in closed kinetic chain. Therefore, ankle sagittal plane movements have also an involvement in the knee joint. Self and Paine  showed that landing technique with the largest ankle plantar-flexion position at ground contact demonstrated the most shock absorption and reduction of the peak vertical ground reaction force. It was recently demonstrated that landing with the rear foot (dorsiflexed ankle) was associated with less hip and knee flexion at peak vertical ground reaction force than forefoot landing (plantarflexed ankle) . Also, the maximum knee flexion angle with forefoot landing technique was significantly higher than the rear foot technique. In contrast, hip and knee flexion angles at initial foot contact were significantly lower with the forefoot than rear foot technique. Given that the maximum force transferred to the knee would be at the peak vertical ground reaction force, a forefoot landing might be preferred over rear foot technique. No significant differences for foot-landing technique were observed between males and females . However, Burkhart et al.  reported in a prospective research study that an athlete who landed with an increased heel to flat-foot loading mechanism was more likely to sustain to a non-contact ACL injury during competitive play.
Coronal plane biomechanics are also involved in the genesis of non-contact ACL injuries. Houck et al.  investigated coronal plane trunk/hip kinematics and hip and knee moments (measures of neuromuscular control) during unanticipated compared to anticipated straight and side step cut tasks. Hip angles, but not lateral trunk flexion, were altered during unanticipated conditions. Lateral trunk flexion remained near 8–10° for both anticipated and unanticipated tasks, what was explained by a compensatory decrease in the left lateral tilt orientation of the pelvis (taken at the same instant in stance as lateral trunk orientation) during the sidestep unanticipated task by 2–3° compared to both the anticipated and unanticipated straight step task, and anticipated sidestep task. Lateral trunk orientation was not a result of lateral trunk flexion, suggesting even during unanticipated tasks the pelvis and thorax rotated as a single segment. The authors interpreted that hip abduction angles and foot placement, not lateral trunk flexion influence trunk orientation . Zazulak et al.  investigated the effects of isolated trunk displacement after a perturbation in a laboratory study as a predictor of knee ligament injury. Twenty-five athletes (11 female and 14 male) sustained knee injuries over a 3-year period. Trunk displacement in any plane was greater in athletes with knee, ligament, and ACL injuries than in uninjured athletes. Lateral displacement was the strongest predictor of ligament injury. Trunk displacements, proprioception, and history of low back pain predicted knee ligament injury with 91% sensitivity and 68% specificity. This model predicted knee, ligament, and ACL injury risk in female athletes with 84, 89, and 91% accuracy, but only history of low back pain was a significant predictor of knee ligament injury risk in male athletes . Therefore, core stability may be an important component of ACL injury prevention programs.
Hip angles during landing can be important determinants of impact force at the knee . Female soccer, basketball, and volleyball players may have an increased external adduction moment about the hip at landing. Increased adduction moment about the hip may place an increased valgus stress over the knee [51, 73], but the hip adduction itself was not demonstrated to be a risk factor for ACL injury . Hip abduction angles and knee moments were significantly affected by the type of task and anticipation. Hip abduction angles decreased by 4.0–7.68°, when comparing the unanticipated sidestep task to the anticipated straight step, unanticipated straight step, and anticipated sidestep tasks . The hip abduction angles were associated with foot placement and lateral trunk orientation. The unanticipated sidestep task reduced the distance between contact point and the trunk center of mass, probably to perform faster, influencing the whole kinetic chain proximally. During landing preparation of a stop-jump task, young female soccer, basketball, and volleyball players demonstrated a decreased hip abduction compared to their male counterparts . Jacobs et al.  observed that healthy women had lower hip abductor peak torque than healthy men. Hip abduction peak torque correlated moderately with hip flexion peak joint displacement in women, and most importantly hip flexion and adduction peak joint displacement increased after a 30-s bout of isometric hip abduction exercise. It was also observed that squat strength positively correlated with hip abduction strength in female soccer players but not for men . The authors suggested hip abduction strength assessment as a potential method to identify those subjects at risk of ACL injury at landing. Imwalle et al.  evaluated lower extremity motions in females performing unanticipated cutting task that directed their running angles to 45 and 90°. They reported that hip internal rotation and knee internal rotation were increased during the 90° cut compared to the 45° unanticipated cut angle. Mean hip flexion was also greater in the 90° cut. However, the only significant predictor of knee abduction during both tasks was hip adduction. The authors suggested that findings indicate that the mechanisms underlying increased knee abduction measures in female athletes during cutting tasks were primarily coronal plane motions at the hip .
Coronal plane knee biomechanics are also related to ACL injury. Hewett et al.  conducted a prospective study where 205 female athletes participating in the high-risk sports of soccer, basketball, and volleyball were measured for neuromuscular control using three-dimensional kinematics (joint angles) and joint loads using kinetics (joint moments) during a jump-landing task. Nine athletes had a confirmed ACL injury. Knee abduction angle at landing was 8° significantly greater in ACL-injured athletes compared to uninjured athletes. In addition, ACL-injured subjects had 2.5 times greater knee abduction moment and 20% higher ground reaction force compared to uninjured subjects. Also, women have exhibited greater valgus moments than men during the landing phase of each stop-jump task . Additionally, increased motion, force, and moments occurred more quickly in the injured compared to uninjured athletes . For cutting maneuvers, Ford et al.  demonstrated that females exhibited greater knee abduction (valgus) angles compared with males.
There is a scarcity of studies on coronal plane ankle biomechanics, but the very few studies demonstrate similar results. Ford et al.  demonstrated that female basketball players had greater maximum ankle eversion than did male athletes during the stance phase of cutting. Similarly, Landry et al.  found an increased ankle eversion angle throughout stance in elite female soccer players compared with male players for unanticipated run and cross-cut maneuvers. It was already reviewed that ankle kinematics influence knee joint. Excessive ankle eversion may increase internal tibial rotation, knee valgus stress, anterior tibial translation, and loading on the ACL during extension [9, 33, 71, 153].
Transverse plane biomechanics have been focused on hip and knee joints. Hip biomechanical findings mainly refer to a greater hip internal rotation maximum angular displacement  and a lower gluteal EMG activity  at landing in female soccer, basketball, and volleyball players compared to males. When performing unanticipated side-cut maneuvers, female soccer players exhibited more hip external rotation compared with the male athletes . When performing unanticipated cutting tasks at 90° unanticipated cut angles, females increased hip internal rotation and knee internal rotation relative to the 45° unanticipated cut angle. However, the changes in transverse planes motion were not related to concomitant coronal plane motions at the knee .
Transverse plane knee biomechanics depend on the type of task and decision-making. Besier et al.  showed that varus/valgus and internal/external rotation moments applied to the knee during sidestepping and crossover cutting were considerably larger than those measured during normal running. The same group found that cutting maneuvers performed without adequate planning may increase the risk of non-contact knee ligament injury due to an increased internal–external rotation moments applied to the knee . The authors attributed these results to the small amount of time to make appropriate postural adjustments before performance of the task such as the position of the foot on the ground relative to the body center of mass . Landing preparation from a stop-jump task demonstrated that young female soccer, basketball, and volleyball players had a decreased hip external rotation and increased knee internal rotation compared with male athletes . It has been also shown that healthy women had higher knee laxity, lower stiffness, and higher energy loss in passive external tibial rotation than did healthy men .
The present literature review elicited several conclusions. First, common playing situations precluding a non-contact ACL injury include change of direction or cutting maneuvers combined with deceleration, landing from a jump in or near full extension, and pivoting with knee near full extension and a planted foot. Second, the most common non-contact ACL injury mechanism includes a deceleration task with high knee internal extension torque (with or without perturbation) combined with dynamic valgus rotation with the body weight shifted over the injured leg and the plantar surface of the foot fixed flat on the playing surface. Third, potential extrinsic non-contact ACL injury risk factors include: dry weather and surface and artificial surface instead of natural grass. Finally, commonly purported intrinsic risk factors include: generalized and specific knee joint laxity, small and narrow intercondylar notch width (ratio of notch width to the diameter and cross sectional area of the ACL), pre-ovulatory phase of menstrual cycle in females not using oral contraceptive, decreased relative (to quadriceps) hamstring strength and recruitment, muscular fatigue by altering neuromuscular control, decreased “core” strength and proprioception, low trunk, hip, and knee flexion angles, and high dorsiflexion of the ankle when performing sport tasks, lateral trunk displacement and hip adduction combined with increased knee abduction moments (dynamic knee valgus), and increased hip internal rotation and tibial external rotation with or without foot pronation. Non-contact ACL injury in female athletes likely has a multi-factorial etiology. Prior epidemiological data demonstrate that neuromuscular and biomechanical risk factors may contribute in isolation or combination with other factors, such as anatomical, hormonal and potentially psychological parameters, to increase relative non-contact ACL injury rates in female athletes.
In general, more research is needed to better elucidate the exact role of all reviewed risk factors in both female and male soccer players. Upon an extensive literature review, nearly every risk factor has contradictory data reported at some point, although there is a preponderance of supportive evidence that acquiesces many of them. The reason why some authors found conflicting results may be explained by the lack of adequate control on multiple confounding factors, especially when the research study may have been insufficiently powered. Whenever possible, studies with an emphasis on designing and reporting Level I research for prospective randomized controlled trials are needed. There is an existing need for further research investigating the role of non-contact ACL injury in male soccer players. The fact that females are at greater risk for an ACL tear should not imply that no research should be conducted on male soccer players. The economic consequences of ACL injuries in semi-professional and professional male soccer players are enormous, and the psychological impairment to the injured player is difficult to quantify but easily understood. Elite soccer studies are limited and it is important to explore this field, since professional soccer players are a relatively homogeneous sample that may differ in some characteristics from recreational players.
Mechanisms of ACL injury have been mostly described in cadaveric studies. Video tape analysis has been used to report mechanisms of injury, but the accuracy and precision is questionable, especially when dealing with rotational and coronal plane moments. In contrast, video analysis allows a forum to obtain on the field data collection. Continuous efforts should be made to improve the accuracy of on-field analysis. Soccer regulations tend to impede the investigation of game injuries, but practice or training sessions can be easily assessed for epidemiological data. ACL tears could be better understood if adequate technology is provided and available on the field.
Environmental risk factors need to be specifically assessed in soccer, since most studies deal with American or Australian Football. A comprehensive prospective epidemiological study on non-contact ACL injury incidence incurred during the sport of soccer with respect to playing surface (gravel, artificial turf, and natural grass), soccer division, and intensity of play at the time of injury (practice versus games) controlling for level of conditioning, weather conditions, or soccer boots would be desirable. Well-designed laboratory and field-based studies are needed in order to adequately assess cleat selection in soccer players. From a theoretical point of view, orthotics may be potentially useful to increase impact absorption or to correct malalignments. However, more research should be conducted in this field. It is important to point out that research on shoes and shoe–surface interaction may provide elucidating results, and provides a means of relative ease with regard to an intervention. Studies regarding weather conditions need to control for many other potential confounding factors like intrinsic biomechanical factors, neuromuscular conditioning, or hydration status of soccer players, among others. The effects of prophylactic knee braces use on neuromuscular control, knee and whole-body kinematics, and knee kinetics of soccer players need to be better characterized. It is not only important to observe the general influence of knee braces at ACL injury reduction, but also to clarify the mechanisms by which they affect injury rates. Knee braces studies should be placebo-controlled to limit the influence of psychological factors. Knee brace interventions need to control for many parameters (position of the player, level of conditioning, skills, experience of the player, soccer division, etc) and also need to include highly homogeneous and a large human sample.
Prevention of non-contact ACL injuries through modification of anatomical risk factors may seem to have limited potential for intervention. However, more investigation in this area will provide us with a better understanding of this injury, and may provide insights on techniques to at least control for these risk factors to help diminish their potential to increase risk of injury (Table 1). Generalized joint laxity or specific knee joint laxity studies need to control for neuromuscular factors. Moreover, specific investigations on mechanical properties of the ACL in loose jointed individuals compared to stiff subjects are desirable. It is necessary to clarify biomechanical consequences of anterior pelvic tilt in lower extremity biomechanics, especially in relation to those modifications that may place the ACL at a greater risk of rupture. Similarly, the increased Q-angle as a risk factor for non-contact ACL tears, considered both statically and dynamically, need clarification, since its’ role in ACL injuries seems more theoretical than practical. Also, it is required that studies on intercondylar notch width are controlled for ACL size. It might be argued that a narrower intercondylar notch size allows less growth to the ACL during human development. Theoretically, a smaller ACL may be at greater risk of rupture, but studies on notch width need to be also controlled by anthropometric characteristics, because a small ACL may not be at greater risk of rupture if the subject is short and thin. The relationship of notch width, ACL size, and anthropometrics of soccer players need to be collectively studied. Other bone morphology characteristics such as posterior tibial slope also require to control for other anatomical, hormonal, neuromuscular, and biomechanical risk factors.
It was postulated that a familial tendency to non-contact ACL injuries may exist, but the existing literature is sparse. Well-controlled large-sampled studies are needed to observe the incidence of non-contact ACL injuries within immediate family members of players who sustained unilateral or bilateral non-contact ACL injuries compared to their uninjured team mates. In the same way, genetic investigations may bring us to a new dimension for the understanding of non-contact ACL injuries. For example, fibroblast gene expression at the ACL may differ among injured versus uninjured subjects. These differences may alter ultra structural characteristics of ligaments and place them at a greater risk of rupture. Could eco-guided fine-needle aspiration biopsy of the ACL provide information on specific genes? Could gene studies predict soccer players at greater risk of suffering non-contact ACL injuries? Further ultra structural composition of the ACL comparisons between female and male soccer players are needed, and between ACL injured and uninjured males as well.
Hormonal risk factors are a promising area of research given the enormous amount of unanswered questions at this point in time. More research is needed to better establish the effects of sex hormones on structure, metabolism, and mechanical properties of tendons and ligaments. The required minimum hormonal concentrations to evoke these tendinous and ligamentous effects should be better defined. Also, the interaction between different hormones would be a more realistic approach to this issue (i.e., estrogen, progesterone, testosterone, relaxin, and inhibin). It is crucial to characterize the phase of the menstrual cycle where female soccer players are at increased risk of non-contact ACL injuries, if one does indeed exist. In the field hormonal studies should consider the interaction of the level of training, level of competition, stress factors, nutritional status, sex hormones, and phase of menstrual cycle in female soccer players.
Neuromuscular and biomechanical risk factors are in direct relation to anatomical structure and development. In fact, it would be interesting to assess if pre-pubertal or pubertal soccer players could benefit from the great potential at modifying neuromuscular risk factors during these developmental stages. Landing neuromuscular and biomechanical studies should aim to investigate authentic game and practice situations such as landing after heading the ball rather than landing from a box. Foot-landing techniques should also be assessed under real jump-landing tasks. Also, neuromuscular and biomechanical studies of unanticipated tasks should be conducted under opposing playing situations rather than using cones and verbal orders. In general, neuromuscular and biomechanical studies require large samples, especially if conducted prospectively, given that a low number of ACL injuries may underpower the study. A new window to future research was opened by Hewett et al.  when they prospectively demonstrated that knee motion and knee loading during a landing task were predictors of ACL injury risk in female soccer, basketball, and volleyball players. Hence, neuromuscular and biomechanical assessments should be studied as potential screening tools. It is also essential to further study the relationship between external knee joint loads and the whole-body kinematics. As reviewed, each segment in the body is related to each other. Thus, a holistic view would consider that the load experienced at the knee joint may be related to the position of the foot on the ground with respect to the body’s center of gravity or the angle of the lower limb with respect to the ground . Neuromuscular and biomechanical studies should take into account postural adjustments in all parts of the body, not only in hip, knee, and ankle joints but also in pelvis and trunk. Likewise, ankle transverse, sagittal, and coronal plane biomechanics in relation to ACL injury risk are a wide and relatively unexplored area to investigate. Another area to further investigate would be to better characterize the relationship between peak ground reaction force and the ACL loading. It is also important to further investigate the role of the iliopsoas muscle in different soccer tasks (since it is directly related to both trunk and hip biomechanics), the role of hip abductors on the landing biomechanics in relation to ACL risk of injury, the role of rectus femoris and his relationship to the location of the center of gravity (especially when the center of gravity falls behind the knee at landing and the rectus femoris has to act as a trunk flexor, requiring a high force that would be potentially transmitted to the knee), and the role of the gastrocnemius muscles during dynamic soccer tasks. Muscular imbalance, especially between hamstrings and quadriceps, should be further investigated to better know how they influence on non-contact ACL injuries. The exact role of neuromuscular fatigue remains controversial. Many laboratory studies show that fatigue may alter neuromuscular control and may place the athlete at a greater risk of ACL injury, but there is no clear increase in non-contact ACL tears at the end of the game or season.
Non-contact ACL injury in athletes likely has a multi-factorial etiology, with several potential elements which determine injury mechanism. There does not appear to be an isolated risk factor presenting in all cases of non-contact ACL injuries. Instead, a combination of risk factors may better predict those soccer players at increased risk of injury. When comparing the data extracted from the reviewed studies, one might generate a few conclusions regarding ACL injury mechanisms and the associated risk factors in soccer players.
Most ACL injuries in soccer players are non-contact in nature (Fig. 1).
Change of direction or cutting maneuvers combined with deceleration, landing from a jump in or near full extension and pivoting with knee near full extension and a planted foot are among the most common playing situations precluding an ACL injury. These playing situations involve knee valgus, varus, internal rotation, and external rotation moments, and anterior translation force.
An anterior translation force may be the most detrimental direct isolated force associated with non-contact ACL injuries. However, a combination (sagittal, coronal, and transverse plane knee loads) of forces produce even a greater strain on the ACL than an isolated anterior translation force.
The most common non-contact ACL injury mechanism is a deceleration task with high knee internal extension torque (with or without perturbation) combined with dynamic valgus rotation with the body weight shifted over the injured leg and the plantar surface of the foot fixed flat on the playing surface.
Artificial Turf may place the soccer player to a higher risk of non-contact ACL injury.
It is not clear whether an increased BMI places the soccer players at a greater risk of non-contact ACL rupture, and studies are mainly based on female athletes.
Generalized joint laxity (which was found to be higher for female athletes) and anterior–posterior knee joint laxity have been associated with an increased risk of non-contact ACL injury for both males and females.
The exact role of pelvis and trunk needs to be further studied in soccer players.
There is not enough evidence to support than an increased Q-angle is a risk factor for non-contact ACL injuries in soccer player
Small and narrow intercondylar notch width (ratio of notch width to the diameter and cross sectional area of the ACL) were found to be risk factors for ACL injuries.
Differences in ACL size and strength between males and females have been identified. Small and weak ACLs contribute to an increased risk of ACL rupture.
Pre-ovulatory phase of menstrual cycle in females not using oral contraceptive were reported to be risk factors for non-contact ACL injuries.
Despite the increased knee laxity around the ovulation may decrease with the use of oral contraceptives in female soccer players; further evidence is needed before this medication can be systematically prescribed in order to decrease the non-contact ACL injuries.
Sex hormones may decrease motor coordination in female athletes and may play a role in non-contact ACL injuries.
Decreased relative (to quadriceps) hamstring strength and recruitment place the soccer player to an increased risk of non-contact ACL injury.
Muscular fatigue may increase the risk of non-contact ACL injury by altering the neuromuscular control.
Decreased “Core” strength and proprioception may place the soccer players to a higher risk of non-contact ACL injury.
Low trunk, hip, and knee flexion angles, and high dorsiflexion of the ankle when performing sport tasks increase the risk of non-contact ACL injuries.
Lateral trunk displacement and hip adduction combined with increased knee abduction moments (dynamic knee valgus) increase the risk of non-contact ACL injuries.
Increased hip internal rotation and tibial external rotation with or without foot pronation are risk factors for non-contact ACL injuries.
Mechanisms and risk factors for non-contact ACL injuries have been mainly studied in female soccer players; thus, further research in male players is warranted. Continued efforts to identify the most modifiable risk factors to ACL injury in soccer players is currently the only option to prevent the high risk of osteoarthritis following this traumatic injury. Pre-season screening programs that monitor important risk factors (especially those that are modifiable) to identify soccer players with potential deficits are warranted to help target “high risk” athletes with the most effective neuromuscular training interventions.
The authors would like to thank Professor Rodrigo C Miralles, from Universitat Rovira i Virgili and Hospital de Sant Joan, Reus, for his excellent contributions, helpful comments and interesting suggestions. We would also like to thank Dr. Gerard Moras for his advices and critical reading of the manuscript. One author (GDM) would like to acknowledge funding support from National Institutes of Health Grants R01-AR049735 and R01-AR055563.
Conflict of interest statement
No conflict of interest is declared.