Chapter 15 Risk Factors for Anterior Cruciate Ligament Injuries in the Female Athlete
The higher incidence of noncontact anterior cruciate ligament (ACL) injuries in female athletes than in male athletes participating in the same sport has received a significant amount of attention in the orthopaedic literature since the early 1990s. A study from the authors’ institution published in 199463 was one of the first to report this problem, because female soccer players were found to have nearly six times the rate of serious knee ligament injuries than that of male players (0.87 and 0.29/100 player-hr, respectively; P < .01). The following year, Arendt and Dick6 presented data from the National Collegiate Athletic Association (NCAA) of injuries sustained by soccer and basketball players over a 5-year period. The ACL injury rate in females was more than double that of males in soccer (0.31 and 0.13/1000 exposures, respectively; P < .05) and four times that of males in basketball (0.29 and 0.07/1000 exposures; respectively, P < .05). Messina and coworkers70 reported that female high school basketball players had nearly four times the incidence of ACL injuries than male players (0.091 and 0.024/1000 player-hr, respectively; P < .05). Gwinn and associates43 evaluated the incidence of ACL injuries in midshipmen at the U.S. Naval Academy over a 6-year period. These authors reported that women had a fourfold increase in this injury compared with men in intercollegiate soccer, basketball, and rugby collectively (0.511 and 0.129/1000 athlete-exposures, respectively; P = .006). During military training, women had nearly 11 times the incidence of ACL ruptures as men during obstacle course running (6.154 and 0.567/1000 athlete-exposures, respectively; P = .004). Agel and colleagues1 reviewed 13 years of NCAA injury data and reported a continued gender disparity in ACL injury rates between collegiate female and male basketball and soccer players.
Majority of studies published to date have either examined a very small sample size of each gender, focused on only one possible risk category, or examined neuromuscular characteristics in a controlled laboratory environment instead of reactive or actual playing conditions.
Despite the multitude of investigations that have been conducted on potential risk factors for noncontact ACL ruptures, definitive conclusions cannot currently be reached for either males or females regarding what factor(s) may predispose an athlete to this injury. In addition, the reasons for the increased incidence of this injury in female athletes over that in male athletes cannot be scientifically defined at present. The problems are that the majority of studies published to date have either examined a very small sample size of each gender (therefore containing insufficient power to avoid a type II statistical error40), focused on only one possible risk category, or examined neuromuscular characteristics in a controlled laboratory environment instead of actual playing conditions. The potential risk factors that have been proposed for this injury are shown in Table 15-1. This chapter reviews investigations related to these risk factors and emphasizes findings that may have a role in the gender disparity in noncontact ACL injury rates. Opinions expressed by scientists and researchers from the 2005 Hunt Valley II Meeting (sponsored by the American Orthopaedic Society for Sports Medicine [AOSSM])41 regarding this problem are also provided in each major risk factor section as well as consensus statements generated from this meeting.
|Males and females||Familial||Genetics, possibly related to anatomic structures (size intercondylar notch)|
|Males and females||Environmental|
|Females||Hormonal||Acute fluctuations of estrogen and progesterone during the menstrual cycle|
ACL, anterior cruciate ligament.
Although the question has been raised for many years of a potential genetic predisposition for ACL rupture, only two studies have been published regarding this risk factor. Harner and coworkers44 investigated familial and anatomic risk factors in a small series of 31 patients who had sustained noncontact bilateral ACL ruptures. Compared with a control group matched for age, sex, height, weight, and activity level, the patients demonstrated a significant difference in the incident rate of immediate family members who had sustained an ACL injury (35% and 4%, respectively; P < .01) and a significantly wider lateral femoral condyle (3.30 and 3.10 cm, respectively; P < .05). The authors concluded that there could have been a congenital predisposition to ACL injury that theoretically resulted from anatomic differences in the size of the femoral condyles.
Flynn and associates34 conducted a questionnaire-based study of 171 patients who sustained either contact or noncontact ACL ruptures to determine whether a familial predisposition existed for this injury. The patients were matched with 171 uninjured control subjects according to age, gender, and primary sport. The data were not sorted according to mechanism of ACL rupture (contact vs. noncontact). The survey results revealed that patients who had sustained ACL ruptures were twice as likely to have a first-, second-, or third-degree relative who also had an ACL tear than the control subjects. Limitations of the study include those associated with recall of injury by the patients of their relatives’ medical history and no medical documentation of the relatives’ ACL ruptures.
Owing to the paucity of published data, no conclusion or consensus was reached regarding genetic predisposition as a risk factor for noncontact ACL injuries by the Hunt Valley II Meeting participants.41
One study reported that, compared with a control group, patients with ACL ruptures demonstrated a significant difference in the incident rate of immediate family members who had sustained an ACL injury and a significantly wider lateral femoral condyle.
Few studies have been conducted to determine the effect of climate conditions, such as rainfall, heat, and humidity, on noncontact ACL injuries. In addition, no investigation to date has taken into account other risk factors (such as hormonal or neuromuscular) along with climate conditions that could confound possible conclusions. Orchard and colleagues82 examined intrinsic (height, weight, age, body mass index (BMI), history of prior ACL injury) and environmental-related variables as risk factors for noncontact ACL injuries in male Australian football players. The incidence of noncontact ACL injuries in the 1643 players was 0.62 per 1000 athlete-exposures. The most significant risk factors were a history of a prior ACL injury and weather conditions (high evaporation and low rainfall) that resulted in a dry playing surface. The theories that high shoe-surface traction was a risk factor for noncontact ACL injury and that dry playing conditions increased friction and torsional resistance between shoes and natural grass were advanced by the authors. Of note was the higher incidence of all ACL injuries in this group (0.82/1000 athlete-exposures) compared with previously published data of American football players (0.25/100 athlete-exposures93) and American collegiate soccer male players (0.12/1000 athlete- exposures7). The authors attributed this difference to the prevailing climatic conditions in Australia, which commonly produced dry, hard ground conditions.
Meyers and Barnhill71 tracked male high school football injuries sustained in eight schools over 5 seasons. The rate of ACL injuries was nearly 50% higher on natural grass than on artificial turf (FieldTurf); however, a total of only 15 ACL ruptures were sustained in the 240 games surveyed. Olsen and coworkers81 examined ACL injury rates in both men and women team handball players to determine whether an association existed between injury rates and two types of floor surfaces. Over a period of seven seasons (94,136 player-hr), the ACL injury rates were 0.24 per 1000 player-hours for men and 0.77 per 1000 player-hours for women (P = .001). The ACL injury rate for women was higher on artificial floors than that of women playing on wooden floors (0.96 and 0.41/1000 player-hr, respectively; P = .03). There was no difference in the ACL injury rate between floor surface types in the male players and there was no difference between genders in ACL injury rates on wooden floors. Unfortunately, the investigators had to eliminate two thirds of all ACL injuries that were incurred in the study group owing to unreliable floor-type exposure data.
Lambson and associates58 evaluated the effect of football cleat design on ACL injury rates (contact and noncontact) in 3119 high school football players who were followed for 3 seasons. A total of 42 ACL ruptures were recorded in this study. Shoes with long, irregular cleat designs produced significantly higher torsional resistance (P < .05) on both artificial and natural playing surfaces than the other cleat designs studied and were also associated with a significantly higher rate of ACL injuries (P = .0062). Whereas the authors recognized that many potential risk factors were not analyzed, they recommended that the long, irregular cleat design be discontinued in high school football players. The effect of footwear on ACL injuries in female athletes has not been investigated to date.
No study to date has demonstrated that knee braces significantly reduce the incidence of ACL injuries in normal healthy athletes.76 Only two epidemiologic investigations were published in the 1990s on this topic, both of which followed only male football players. Sitler and colleagues102 conducted an investigation involving 1396 U.S. Military Academy cadets who played intramural tackle football and in whom the type of shoe, athlete exposure, brace compliance, playing surface, and knee injury history were controlled. The brace was a double-hinged, single, upright, off-the-shelf design (DonJoy, Inc., Carlsbad, CA) that was randomly assigned to players at the beginning of each of the 2 seasons surveyed. A statistically greater number of medial collateral ligament (MCL) injuries occurred in the control group than in the braced group (P < .05). Although a higher number of ACL injuries was found in the control group than in the braced group (12 vs. 4), the small number precluded statistical analysis. The second epidemiologic investigation published in this time period assessed only MCL injury patterns.2
The Hunt Valley II Meeting participants concluded that data reported to date regarding environmental risk factors for ACL noncontact injuries has been “confusing and mixed.”41 The researchers did conclude that shoe-surface interaction may affect ACL injury risk both directly and indirectly.41 The direct effect is through higher traction, which may transmit excessive load forces to the knee during cutting and pivoting. The indirect effect is through alterations in neuromuscular movement patterns in an attempt to adapt to variations in shoe and surface factors. Future studies examining the potential impact of these factors should include both intrinsic and extrinsic variables and use sound epidemiologic study design.
Many authors have proposed that inherent structural differences between genders are responsible or partially responsible for the disparity in noncontact ACL injury rates between genders. Although evidence exists to support differences between men and women in many of these factors including quadriceps femoris angle (Q-angle), femoral anteversion, tibial torsion, foot pronation, intercondylar notch size, and ACL size, no investigation has demonstrated that these differences alone are responsible for an increased risk of noncontact ACL injuries in female athletes.
An association between a small-sized intercondylar notch and an increased incidence of ACL ruptures has been reported or suggested by some authors,5,36,50,59,67,97,105 but refuted by others.47,65,92,108 The association, which remains debatable, has been reported by some investigators to be present in both genders.41,54,97 A general speculation has been raised by some authors41 that a small-sized notch will contain a small-sized ACL that is vulnerable to rupture owing to decreased strength properties. Variations in techniques used to determine notch size (plain non–weight-bearing and weight-bearing radiographs, magnetic resonance imaging [MRI], computed tomography [CT], photographic techniques), problems and lack of standardization with these techniques (such as uncontrolled knee flexion angles and rotation on radiographs), discrepancies in notch indices studied (lateral condylar width, total condylar width, notch width, notch width at two thirds notch height, notch width index, notch angle), and differences in statistical methods used to ascertain measurements account for the disagreement among studies. The major problem is that no study to date has entered anatomic indices along with hormonal and neuromuscular factors into an appropriate statistical model to determine the effects of all of these potentially important risk factors.
West Point study: Factors of narrow femoral notch, body mass index 1 SD or more above the mean, and generalized joint laxity correctly predicted 75% (6 of 8) of the noncontact ACL injuries in females.
AOSSM Consensus: No definite evidence exists that any anatomic factor is reliably associated across age groups and sexes with an increased rate of injury based on the amount of conflicting published data and variety of study designs.
Uhorchak and coworkers,109 in the most comprehensive ACL risk factor study published to date, measured height, weight, BMI, condylar width, notch width, eminence width, tibial width, notch width index, generalized joint laxity, anteroposterior (AP) displacement on KT-2000 testing, isokinetic quadriceps and hamstrings concentric and eccentric strength, and hamstrings flexibility in a group of 895 West Point cadets (120 women, 739 men) upon their entrance into the academy. The cadets were followed for 4 years for ACL injuries, during which time 24 noncontact ACL ruptures occurred (16 in men, 8 in women). Using a hypothesis-driven logistic regression model, the factors of a narrow femoral notch, BMI 1 standard deviation (SD) or more above the mean, and generalized joint laxity explained 62.5% of the variability of the noncontact ACL tears in the female athletes. In addition, this model correctly predicted 75% (6 of 8) of the noncontact ACL injuries. The most predictive model in the men explained only 15% of the variability in the noncontact ACL injuries and was unable to predict any of the 16 injuries that occurred throughout the study period in the male athletes.
The shape of the intercondylar notch varies,53 but this does not appear to be useful in predicting patients who may have an increased risk for ACL injury.5,54,106 In addition, the size of the ACL cannot be reliably predicted from the size of the intercondylar notch.4,75,106 A few investigators have attempted to determine whether a gender difference exists in the cross-sectional area, mass, and volume of the ACL in uninjured subjects using MRI4,18,23,106 and in cadaveric specimens.14,75 Charlton and associates18 reported that, although a difference existed between uninjured men and women in the volume of the femoral notch and ACL, this difference was related to height and weight and not gender. Patients with smaller notches also had smaller ACLs. Anderson and colleagues4 found no difference between male and female high school basketball players in the notch width index or in the size of the ACL when normalized for lean body weight (body weight x [100 – % body fat]). When adjusted for body weight, the mean ACL area was significantly greater in the male players (P < .003). There was a correlation of ACL area to height in the male players (P = .03) but not in the female players. Taller male players had large ACLs, but taller female players had ACLs of a size similar to that of shorter female players. The authors concluded from their data that the cause of ACL rupture in patients with narrow notches was not due to a smaller (more vulnerable) ligament, but to a normal-sized ACL in a stenotic notch. This finding supported that previously described by Muneta and coworkers75 who, in a cadaveric study, found no correlation between the size of the intercondylar notch and that of the ACL.
Staeubli and associates106 measured the widths of the cruciate ligaments and intercondylar notches in 51 uninjured subjects with MRI. The notch measurements were performed in three areas in the coronal plane: at the notch entrance, at the intersection of the cruciate ligaments, and at the notch outlet. The authors reported a significant difference between genders in the absolute width of the ACL at the cruciate intersection (men, 6.1 ± 1.1 mm, and women, 5.2 ± 1.0 mm; P < .01). The intercondylar notch widths were also significantly greater in males, as were height, weight, and bicondylar femoral width. The ACL and notch size measurements were not normalized for these variables. The ACL occupied only 31.9% of the notch in men and 31.1% of the notch in women, refuting the belief that the size and shape of the notch lead to an increased risk of ACL injury.
Chandrashekar and colleagues14 studied the length, area, mass, and volume of the ACL in 20 cadaveric knees (10 males, 10 females, aged 17–50 yr) using a photographic three-dimensional scanning system. The male specimens had significantly larger values than the females for ACL length (29.82 mm ± 2.51 and 26.85 ± 2.82 mm, respectively; P = .01), mid-substance area (83.54 ± 24.89 mm2 and 58.29 ± 15.32 mm2, respectively; P = .007), mass (2.04 ± 0.26 g and 1.58 ± 0.42 g, respectively; P = .009), and volume (2967 ± 886 mm3 and 1954 ± 516 mm3, respectively; P = .003). There were no significant differences between genders in ACL mass density, notch width, or notch width index. ACL cross-sectional area and volume were not associated with body height or lean body mass in either gender. ACL mass strongly correlated with height in men (R = 0.7; P < .02), but not in women (R = 0.43; P ≥ .02). The authors concluded that because the density of the ACL is similar between genders, the smaller ligament size in women may contribute to their increased rate of noncontact ACL rupture.
Excessive foot pronation has been suggested as a potential risk factor for ACL injuries.3,9,66,115 Some authors have speculated that excessive subtalar joint pronation leads to increased internal tibial rotation and resultant high forces on the ACL. Two investigations involving small sample sizes reported significantly higher navicular drop values bilaterally in patients who sustained ACL ruptures compared with matched control subjects.3,9 Allen and Glasoe3 measured navicular drop in 18 subjects who had sustained ACL ruptures using a Metrecom three-dimensional digitizer. When compared with an age- and gender-matched control group, the mean values of the navicular drop of both the ACL-ruptured and the contralateral limbs of the patients were significantly larger than those in the control group (P < .05) by approximately 2 mm. No other risk factors were included in this investigation. Woodford-Rogers and coworkers115 measured navicular drop, calcaneal alignment, and anterior tibial translation in the uninjured limb of 22 high school and collegiate athletes (14 males, 8 females) who had sustained noncontact ACL injuries. The data were compared with those of athletes matched by sport, team, position, and level of competition. The authors reported that the ACL-injured athletes had greater amounts of navicular drop of approximately 2.5 mm in the male subjects and 2 mm in the female subjects. Discriminant analysis of the data showed that navicular drop, calcaneal alignment, and anterior tibial translation (when combined) were predictors of classification of athletes into the ACL-injured and the ACL-noninjured groups, because 87.5% of the athletes were correctly classified (chi-square, 9.00; P < .01). Both of these studies suggested that excessive pronation may lead to an increased risk of ACL rupture.
A combination of postural “faults” was found to have a greater predictive value for ACL injury than a single problem (such as excessive pronation) by Loudon and associates.66 In 20 female athletes with unilateral ACL ruptures, the combination of excessive knee hyperextension, navicular drop, and subtalar joint pronation was a strong discriminator between the patients and 20 control subjects. The authors concluded that an association existed between noncontact ACL injuries in females who had a standing posture of genu recurvatum and subtalar joint overpronation.
Smith and colleagues104 failed to find an association between excessive pronation and noncontact ACL injuries. The authors used a combination of navicular drop and calcaneal stance measures in 14 patients with ACL ruptures and compared the data with those of 14 control subjects matched for age, height, and weight. There was no significant difference in the navicular drop test between the patients and the controls. The authors cautioned that static measurements of pronation may not be representative of dynamic conditions and that future research should include not only dynamic analyses but other anatomic risk factors as well.
There is no evidence to support that an excessive BMI alone is a risk factor for noncontact ACL injuries. Ostenberg and Roos83 found that BMI was not a risk factor for knee injuries in female soccer players, as did Knapik and coworkers56 in military training–related injuries in female recruits. When combined with a narrow femoral notch and generalized joint laxity, a BMI greater than 1 SD above the mean did contribute to a high prediction of noncontact ACL injuries in Uhorchak and coworkers’ series.109
Some authors have reported that females appear to have greater inherent joint laxity than males.12,60,88–90 Huston and Wojtys52 reported that females had significantly greater anterior tibial translation than males (6.5 and 5.8 mm, respectively; P < .05), as did Rozzi and associates90 (6.05 ± 1.46 mm and 4.80 ± 1.53 mm, respectively; P = .02). Seckin and colleagues94 reported a significantly greater prevalence of generalized joint hypermobility in female high school students than in male students (16.2% and 7.2%, respectively; P < .0001). Remvig and coworkers88 found considerable evidence in the literature of an increased prevalence of hypermobility among women. However, there is no evidence that an increase in generalized ligament laxity in females is associated with the increased rate of noncontact ACL injuries.
The Hunt Valley II Meeting participants concluded that “there is no definite evidence that any anatomical factor is reliably associated across age groups and sexes with an increased rate of injury” based on the amount of conflicting published data and variety of study designs.41
Since the early 2000s, the hypothesis that fluctuations in sex hormones during certain periods of the menstrual cycle could be deleterious to the material and mechanical properties of the female ACL has been raised by several investigators. Some postulated that these harmful effects could explain the heightened risk of ACL ruptures in female athletes,53 although a recent investigation on a synomolgus monkey model showed no direct effects of estrogen on ACL mechanical or material properties.110 Estrogen and progesterone have been shown to significantly alter the structure and metabolism of collagen in human and animal models.98 Other authors have raised the question of whether hormonal surges could have a deleterious effect on muscle and neuromuscular indices, thereby increasing the vulnerability of the ACL to rupture during certain phases of the menstrual cycle.13
AOSSM Consensus: Although there appears to be an uneven distribution of ACL injuries in the follicular phase, more data are required before definitive conclusions may be reached. No consensus exists regarding the role of oral contraceptives on knee injury risk or function.
In 1996, Liu and associates64 conducted the first investigation to determine whether estrogen and progesterone target cells were located within the human ACL. Seventeen ACL specimens were obtained from patients undergoing a variety of surgical procedures, 13 from women and 4 from men. Both hormone receptors were found in the ACL fibroblasts, which led the authors to suggest that female sex hormones may have an effect on the ligament’s structure and composition.
A few years later, Yu and colleagues117 conducted an analysis of the effects of estrogen on human ACL fibroblast proliferation and procollagen synthesis. An ACL sample was obtained from a 32-year-old female patient undergoing a total knee replacement. A dose-dependent effect of estrogen was demonstrated, because the proliferation of ACL fibroblasts decreased as estrogen concentration was increased. This effect was most striking in the early periods of hormone exposure (days 1 and 3 after administration) and became less pronounced with time. In addition, a dose-dependent effect was found with type I procollagen synthesis, which decreased with increased estrogen concentration. Type I collagen is associated with mechanical strength of connective tissues.
In a second investigation, Yu and coworkers118 studied the effects of varying levels of estrogen and progesterone on human ACL fibroblast proliferation and procollagen synthesis. Cell lines were obtained from ligaments in two young female patients. The authors reported that estrogen had a dominant inhibitory effect on human ACL fibroblast proliferation and type I procollagen synthesis and that progesterone reduced this effect. A dose-dependent increase in ACL fibroblast proliferation and type I procollagen synthesis occurred when progesterone concentration was increased and estrogen concentration held constant. The dose-dependent effects of progesterone were more pronounced at lower levels of estrogen concentration than at higher levels. The authors concluded from these investigations that the acute hormonal fluctuations that occur during the menstrual cycle may induce changes in the metabolism of the ACL, weakening the ligament’s strength and increasing its vulnerability to injury.
Faryniarz and associates31 measured estrogen and relaxin receptors from resected ACL specimens in eight women and seven men an average of approximately 5 weeks after their injury. Whereas there was no significant difference between genders in the concentration of estrogen receptors in ACL cells, women had higher levels of relaxin-binding sites. The authors concluded that estrogen alone may not play a role in the gender disparity in noncontact ACL injury rates. However, their data were considered preliminary and the authors remarked that subtle differences between genders in estrogen concentration in human ACLs may be detected in a study with a larger sample size.