Risk factors for a contralateral anterior cruciate ligament injury

Transcription

1 Knee Surg Sports Traumatol Arthrosc (2010) 18: DOI /s KNEE Risk factors for a contralateral anterior cruciate ligament injury Per Swärd Ioannis Kostogiannis Harald Roos Received: 14 October 2009 / Accepted: 8 December 2009 / Published online: 9 January 2010 Ó Springer-Verlag 2010 Abstract Contralateral anterior cruciate ligament (ACL) injuries are together with the risk of developing osteoarthritis of the knee and the risk of re-rupture/graft failure important aspects to consider after an ACL injury. The aim of this review was to perform a critical analysis of the literature on the risk factors associated with a contralateral ACL injury. A better understanding of these risk factors will help in the treatment of patients with unilateral ACL injuries and in the development of interventions designed to prevent contralateral ACL injuries. A Medline search was conducted to find studies investigating risk factors for a contralateral ACL injury, as well as studies where a contralateral ACL injury was the outcome of the study. Twenty studies describing the risk of a contralateral ACL rupture, or specific risk factors for a contralateral ACL injury, were found and systematically reviewed. In 13 of these studies, patients were followed prospectively after a unilateral ACL injury. The evidence presented in the literature shows that the risk of sustaining a contralateral ACL injury is greater than the risk of sustaining a first time ACL injury. Return to a high activity level after a unilateral ACL injury was the most important risk factor of sustaining a contralateral ACL injury. There was inconclusive evidence of the relevance of factors such as female gender, family history of ACL injuries, and a narrow intercondylar notch, as risk factors for a contralateral ACL injury. Risk factors acquired secondary to the ACL injury, such as altered biomechanics and altered neuromuscular function, affecting both the injured and the contralateral leg, most likely, further increase the risk of a contralateral ACL P. Swärd (&) I. Kostogiannis H. Roos Department of Orthopaedics, Lund University and Lund University Hospital, Lund, Sweden pellesward@hotmail.com injury. This literature review indicates that the increased risk of sustaining a contralateral ACL injury should be contemplated, when considering the return to a high level of activity after an ACL injury. Keywords Anterior cruciate ligament Bilateral Contralateral Intercondylar notch Neuromuscular Biomechanical Activity Introduction Anterior cruciate ligament (ACL) injuries are common in contact sports with pivoting movements such as football, soccer, basketball and handball, as well as in non-contact sports such as skiing. The majority of ACL injuries occur when landing, or in situations where the athlete, while in contact with the underlying surface, decelerates or performs a side-cutting manoeuvre [17, 57]. Specifically, anterior translation force on the ACL has been suggested to be the most important factor for non-contact ACL injuries [5]. The injury may happen in contact with another player or as a non-contact injury. Non-contact injuries among young female athletes are especially prevalent, and an overview was recently published on this topic [82]. The aetiology of ACL injuries is complex, and although different factors have been implicated in the risk of ACL injury, it is still not thoroughly understood why some individuals are at higher risk than others [5]. Also, to date, there is still controversy regarding the best treatment after ACL injury, both considering surgical repair versus nonsurgical treatment, and whether early or delayed ACL reconstruction is favourable [10]. Furthermore, it is difficult to explain why so many of those with unilateral injuries also sustain an injury to the contralateral knee within a

2 278 Knee Surg Sports Traumatol Arthrosc (2010) 18: few years [32, 33, 71, 79, 86]. Contralateral ACL injuries are together with the risk of developing osteoarthritis of the knee [10, 59, 60, 69, 98, 110], and the risk of re-rupture/ graft failure [68, 71, 86], important aspects to consider after an ACL injury. Numerous studies have been performed on risk factors associated with a unilateral ACL injury. These risk factors can be divided into intrinsic (related to the individual) and extrinsic (related to the environment). In this review, we will focus on intrinsic factors, which are often divided into anatomical, hormonal, neuromuscular and familial [39, 40, 82]. There is no consensus in the literature regarding the relevance of different risk factors for unilateral ACL injuries, but there is agreement that women engaged in sports have a higher risk of ACL injuries than men engaged in sports [3, 11, 39, 40, 78]. There is also insufficient knowledge about contralateral ACL injuries, mainly due to difficulties in designing studies of such injuries. It is difficult to implement a study designed to evaluate the risk factors for a contralateral ACL injury prospectively for many reasons. Above all, it would require a very large study population, and a long follow-up time, especially since acquired risk factors for a contralateral ACL injury must be considered secondary to the initial ACL rupture [33, 37, 71, 83, 86]. The aim of this review was to perform a critical analysis of the literature on the risk factors associated with a contralateral ACL injury. A better understanding of these risk factors will help in the treatment of patients with unilateral ACL injuries, and in the development of interventions designed to prevent contralateral ACL injuries. Risk factors for unilateral injuries Anatomical factors A thoroughly studied risk factor that can predispose a person to an ACL injury is the morphological anatomy of the intercondylar notch, more specifically intercondylar notch stenosis. It has been speculated that a narrow notch affects the ACL, either by impingement on the ACL during sports activities, which may damage and weaken the ACL or that notch width which is correlated with the width of the ACL; a narrow notch indicating a small ACL [20, 21, 24, 28]. If the inherent material properties of the ACL are equivalent for a small and a large ACL, the strain on the small ACL will be greater when the same load is applied to it, and this could explain the observed correlation between a narrow notch and a predisposition for ACL injury [24, 28]. A recent magnetic resonance imaging (MRI) based investigation showed that the ACL volume of the contralateral uninjured knee in ACL injured subjects was significantly smaller than the ACL volume of uninjured controls, after adjusting for body mass index [19]. Based on these findings, the authors suggested that a small ACL volume is a potential risk factor for ACL injury [19]. Factors such as notch width, notch width index (NWI) and the shape of the notch have been evaluated. The NWI, which was first introduced by Souryal et al. [95], is a ratio that can be measured on plain radiographs. The NWI is used as it is assumed that it could even out individual differences, enabling comparisons between patients [95]. Although a great deal of research has focused on the notch width as a risk factor for ACL injuries, there is no consensus on the relevance of a narrow notch as a risk factor for ACL injury [9, 42, 43, 58, 61, 62, 88, 90, 91, 94, 95, 103]. Other anatomical factors that have been considered but not verified as risk factors for ACL injury are the quadriceps angle and static knee alignment [40]. Furthermore, a significant correlation has been demonstrated between increasing caudal slope of the tibia on plain radiographs and ACL insufficiency in two independent investigations [18, 100], whereas one study with a smaller study sample could not demonstrate an important difference [64]. This slope, also known as the posterior tibial slope is defined as the angle at the intersection between the line perpendicular to the tibial axis and the posterior inclination of the tibial plateau [26]. Recent, MRI-based investigations, have, moreover, shown that patients with ACL injury, compared to controls, have steeper posterior directed tibial slopes of both the lateral and the medial plateau, as well as a shallow medial tibial plateau [41, 96]. Furthermore, both increased generalised joint laxity, and increased knee joint laxity (knee hypertension) have been shown to predispose individuals to ACL rupture [67, 103]. Myer et al. [67] also showed an increased risk of ACL injury for young female athletes with a high side-to-side difference in knee laxity, calculated from anterior posterior tibial translation between the dominant and the nondominant knee [67]. There is also one report that a high BMI predisposes female individuals to ACL injury [103]. Neuromuscular factors Apart from the actual muscular response, neuromuscular control is dependent on sensory information from proprioceptive, kinesthetic, vestibular and visual input [50]. It is required to stabilize knee joint movements in the sagittal, transverse and frontal planes [46]. Deficits in neuromuscular control may induce more stress on the passive structures of the joint (i.e. ligaments and joint capsule) and thus increase the risk of an ACL tear. Risk factors for ACL injury associated with neuromuscular control such as proprioception [36], quadriceps-to-hamstring ratio [4] and inadequate muscle stiffness have been suggested [31].

3 Knee Surg Sports Traumatol Arthrosc (2010) 18: Biomechanical factors, including a high degree of dynamic valgus motion in landing and decreased neuromuscular control of the trunk, have been shown to predispose an individual for ACL rupture [44, 113]. Some studies have indicated that hyperpronation of the foot is a risk factor for ACL injury [7, 15, 92]. Furthermore, Swanik et al. [97] found that athletes sustaining non-contact ACL injuries had slower reactions and slower processing speed, and had significantly poorer visual memory and verbal memory composite scores than controls. They therefore suggested a relationship between neurocognitive function and the risk of ACL injury [97]. Hormonal factors Many theories have been proposed as to why women are at greater risk of ACL injury than men. Anatomical differences as well as neuromuscular differences have been discussed. One suggestion for the increased incidence of ACL injury in women is the hormonal differences between the sexes. It has been suggested that female sex hormones directly affect joint laxity, ACL metabolism and collagen synthesis, as well as the neuromuscular system [31, 82, 87, 114]. Hewitt et al. [45] systematically analysed seven articles on the association between the menstrual cycle and non-contact ACL injuries in female athletes. Five of the seven articles reported an increased incidence of noncontact ACL injuries in the preovulatory phase of the menstrual cycle [45]. There is, however, still no consensus as to whether ACL injuries are more prevalent in a specific phase of the menstrual cycle or related to hormonal fluctuations during the menstrual cycle, mainly because of the discrepancy of results between investigations and also because of methodological insufficiencies in determining menstrual status and identifying specific phases of the menstrual cycle [105]. Familial factors Research on the influence of the genetic component or familial history of ACL injury on the risk of ACL injury is somewhat sparse. Even so, September et al. [89] suggested in a recent review on this topic that there is at least some genetic component in the risk of sustaining an ACL injury. However, no specific genes were suggested [89]. More recent case control investigations have demonstrated that various specific genotypes of the genes: COL1A1, COL5A1, and COL12A1, which code for structural components of the ACL, are associated with ACL injury [51, 75 77]. COL1A1 codes for the alpha1 chains of type I collagen, COL5A1 codes for the alpha1 chains of type V collagen, and COL12A1 codes for type XII collagen [76]. Posthumus et al. [75], furthermore, showed that patients with an ACL rupture reported significantly more ligament injuries among blood relatives than did controls. This finding is in congruence with findings by Flynn et al. [34], who in a retrospective investigation showed that patients with an ACL injury were twice as likely to have a first- (parent, sibling or child), second- (uncle or aunt) or thirddegree relative (cousin) with an ACL injury, compared with patients with no known ACL injury, and matched to age, gender and sport activity. Materials and methods A systematic review was conducted in order to examine the evidence on relevant risk factors for a contralateral ACL injury. Since it was anticipated that there would be a limited number of studies addressing this subject, there were no exclusion criteria concerning age, gender, study design or any other factor. Studies investigating risk factors for a contralateral ACL injury, as well as studies where a contralateral ACL injury was the outcome of the study were included in this review. Since the risk factors may be the same for contralateral ACL injuries and for unilateral ACL injuries, intrinsic risk factors for a unilateral ACL injury have been addressed in broad outline. We calculated the annual incidence of contralateral ACL injuries from studies where the numbers of these injuries had been reported during a specified period of time. The literature study was based on a Medline search with the PubMed engine. The words, risk factors, bilateral, predisposing, contralateral, re-injury, re-rupture together with the key word ACL were used to find relevant articles. The titles and abstracts were read, and all relevant articles were retrieved. The authors of articles meeting the inclusion criteria were checked for additional relevant papers. The reference lists were also scrutinised for supplementary articles in the field. A further search was performed using key words related to the different risk factors, such as intercondylar notch, knee laxity, hormonal, etc., together with ACL. Results Twenty studies describing studies of the risk of a contralateral ACL rupture, or specific risk factors for a contralateral ACL injury, were found and systematically reviewed. The articles had been published between 1987 and We found 13 investigations of the risk of a contralateral injury where patients were followed prospectively after a unilateral ACL injury. From these studies, the annual incidence or the incidence per 1,000 game exposures of contralateral ACL injuries was calculated.

4 280 Knee Surg Sports Traumatol Arthrosc (2010) 18: Incidence of new ACL injuries after a unilateral ACL injury Orchard et al. [71] found that the greatest risk factors for a contralateral ACL injury or re-rupture/graft failure were a history of an ACL reconstruction either in the previous 12 months (relative risk [RR], 11.3; 95% confidence interval [CI], ), or before the previous 12 months (RR, 4.4; 95% CI, ). In a study on female soccer players in the German national league, it was found that 5 out of 19 players with a previous ACL injury sustained a new ACL injury during one outdoor season (11 months), while only 6 out of 124 players without a previous ACL injury sustained an ACL injury (odds ratio, 5.24; ) [33]. Re-ruptures/graft failures versus contralateral ACL injuries It appears that the risk of re-rupture or a graft failure is highest within the first 12 months after an ACL injury [71, 86]. After the first 12 months, new ACL injuries are evenly distributed between the reconstructed knee and the previously uninjured knee, and the overall increase in risk is divided fairly evenly between the reconstructed knee and the contralateral knee [71, 86]. In a study by Salmon et al. [86], the peak incidence of contralateral ACL ruptures was seen during the second and third years after reconstruction during a 5-year follow-up period. They found a total of 35 contralateral and 39 re-ruptures during the period studied. Different studies with varying follow-up periods show that either re-ruptures or contralateral ACL injuries are most common [22, 68, 71, 73, 74, 84, 86, 90, 111]. Myklebust et al. [68] found at follow-up, 6 11 years after ACL injury that 11 of 50 (22%) patients who had underwent ACL reconstruction and continued to play handball reinjured their ACL, whereas six (9%) of the players who continued to play handball sustained a contralateral ACL injury. Moreover, a set of investigations of the same cohort of 180 patients prospectively followed after ACL reconstruction with either patellar tendon (90 patients) or hamstring tendon grafts (90 patients) found that the proportion of contralateral ACL injuries versus graft failures/re-ruptures increased over time after the injury [22, 73, 74, 84]. Twoyear data demonstrated seven re-ruptures and 6 contralateral tears [22], 5-year data demonstrated 10 re-ruptures and 19 contralateral tears [74], 7-year data demonstrated 13 reruptures and 25 contralateral tears [84], and 10-year data demonstrated 19 re-ruptures and 29 contralateral tears [73]. In another prospective investigation, 7 re-ruptures (3%) and 7 (3%) contralateral ACL tears were identified at the 2-year follow-up [111]. Contralateral ACL injuries Studies in which records of ACL injuries were reviewed have shown that 2 4% of individuals with an ACL injury also suffer an ACL injury to the contralateral knee [9, 42, 95]. In elite alpine skiers, it was shown that 30% of those with a unilateral ACL injury sustained a contralateral ACL injury during their career [79]. The annual incidence of contralateral ACL injuries has been shown to be between 1.1 and 2.5%, over study periods of 2 10 years in populations surgically reconstructed with either patellar tendon or hamstring tendon autografts, and where the patients were not engaged in any specific sports or activities [12, 22, 25, 73, 74, 84, 86, 106, 111] (Table 1). One prospective investigation of consecutively recruited ACL injured patients, initially treated conservatively without ACL reconstruction with later surgical intervention only when deemed necessary, where the patients were advised to avoid contact sports, found an annual incidence of contralateral ACL injuries of 0.4% over 15 years [55] Table 1 The annual incidence of contralateral ACL injuries following ACL reconstruction [12, 22, 25, 73, 74, 84, 86, 106, 111], and among patients initially treated conservatively without ACL reconstruction with later surgical intervention only when deemed necessary [55] Number of patients Contralateral ACL injuries At start of study At follow-up Follow-up time (years) No. of patients Annual incidence (%) Pinziewski et al. [74] Deehan et al. [25] Aune et al. [12] Roe et al. [84] Pinczewski et al. [73] Corry et al. [22] Wrigth et al. [111] Webb et al. [106] Salmon et al. [86] Kostogiannis et al. [55]

5 Knee Surg Sports Traumatol Arthrosc (2010) 18: Table 2 The incidence of unilateral and contralateral anterior cruciate ligament (ACL) injuries among individuals with no previous ACL injury, and individuals with a history of unilateral ACL injury [32, 33, 70, 71] Baseline data No ACL injury Number of patients or number of game exposures Unilateral ACL injury Activity Follow-up time Faude et al. 124 a 19 a Soccer 11 months Oates et al a 412 a Ski-resort employees 3 years Orchard et al b 6128 b Australian football Outcome Contralateral ACL injuries First-time ACL injuries Number of patients Incidence Number of patients Incidence Faude et al. 2 11% (annual) c 6 5% (annual) Oates et al % (annual) % (annual) Orchard et al d d a Number of patients b Number of game exposures c The incidence of contralateral ACL injuries was based on information from two papers by Faude et al. [32, 33] d The incidence of ACL injuries in Australian football was calculated per 1,000 game exposures (Table 1). In athletes involved in team sports, the annual incidence has been found to range from 1.1% in female handball players (top three divisions) to 11% in top-level female soccer players [32, 33, 68]. The incidence of contralateral ACL injuries for the female soccer players was based on information from two papers by Faude et al. [32, 33]. The high incidence may be misleading considering the low number of women (2) who sustained a contralateral ACL injury, and the small number of women with a unilateral ACL injury at the beginning of the study (19) [32, 33]. In a skiing population (skiing resort employees, 82 skiing days/year), the annual incidence of contralateral ACL injuries was considerably lower, only 0.2% [70]. In 3 studies following a population that included both individuals with a unilateral ACL injury and those without known ACL injuries, the incidence of contralateral ACL injuries was compared to that of first-time ACL injuries. It was found that the incidence of contralateral ACL injuries was greater in all studies than the incidence of first-time ACL injuries [32, 33, 70, 71] (Table 2). One of these studies reports the number of contralateral ACL injuries per 1,000 game exposures [71]. The other studies presented the number of contralateral ACL injuries during a specific period of time. Risk factors for a contralateral ACL injury We found 11 studies in the literature that described specific factors predisposing individuals to a contralateral ACL injury. Four are follow-up investigations after a unilateral ACL injury [73, 86, 90, 111], while the others are retrospective studies of patients with bilateral ACL injuries. A summary of the studies demonstrating significant results is shown in Table 3. Anatomic risk factors The majority of these studies were focused on anatomic risk factors, specifically the geometry of the intercondylar notch, and its involvement in both unilateral and bilateral ACL injuries [9, 42, 65, 88, 90, 91, 95]. In one study, individuals were prospectively followed after unilateral ACL reconstruction, and it was found that patients with a narrow notch (measured intraoperatively) were at increased risk of sustaining a contralateral ACL injury [90]. The group consisted of 714 patients, and was divided in two subgroups, one with a notch width of 15 mm or less, and one with a notch width of 16 mm or more. Twenty-three of the 388 patients with a narrow notch sustained a contralateral ACL injury (*6%), compared to only 4 out of 326 with a wide notch (*1%) [90]. Anderson et al. [9] used CT scans to compare the intercondylar notch between subjects with bilateral ACL injury, unilateral ACL injury and those without ACL injuries. Their results showed significant notch stenosis in subjects with ACL injury, either bilateral or unilateral, compared with uninjured individuals. Souryal et al. [95] found significantly lower NWIs in subjects with bilateral ACL injury than in unilaterally injured and uninjured subjects, but no difference between uninjured

6 282 Knee Surg Sports Traumatol Arthrosc (2010) 18: Table 3 Summary of studies demonstrating important risk factors for contralateral ACL injuries Author Study design Subjects Time between ACL injuries Risk factor Results Andersson et al. [9] Retrospective case control investigation, CT scan Souryal et al. [95] Retrospective case control investigation, plain radiographs Harner et al. [42] Retrospective case-control investigation, CT scan Shelbourne et al. [91] Retrospective case-control investigation, plain radiographs Shelbourne et al. [90] Prospective follow-up investigation after ACL reconstruction, interoperative notch width measurement of the intercondylar notch Motohashi et al. [65] Retrospective case-control investigation, plain radiographs Salmon et al. [86] Prospective 5 year followup investigation after ACL reconstruction with either patellar or hamstring tendon autograft 14 Patients with bilateral ACL injuries (8 men and 6 women, 28 knees), versus 17 controls (10 men and 7 women, 34 knees) 41 Patients with bilateral ACL injuries (28 men and 13 women), versus 50 consecutive patients with an acute ACL tear, and 50 consecutive patients with no ACL injury 31 Patients with bilateral ACL injuries (22 men and 9 women) versus 23 matched controls (13 men and 10 women) 131 Patients with bilateral ACL injuries (90 men and 41 women), versu 200 controls (100 men and 100 women) 714 Patients (480 men and 234 women) 10 Patients with bilateral ACL injuries (all women) versus, 161 patients with unilateral ACL injury (58 men and 103 women) 612 Patients (383 men and 289 women) Small MNW/CW P = Small (NW at 2/3 of P = NH)/CW Small NA P = Mean 47 months Small NWI P \ (compared to patients with a unilateral ACL injury) P \ (compared to uninjured) Mean 50 months for men and 39 months for women Small NA Large LCW Small CW Large LCW/CW Reported immediate family history of ACL injuries P = P = P = P = P \ 0.01 Small NW P \ 0.01 NW B the median NW High mean generalised laxity score Median 28 months after surgery P \ 0.01 P \ 0.05 Small NWI P \ 0.05 Return to an activity level of moderate or strenous activities after ACL reconstruction P = 0.001

7 Knee Surg Sports Traumatol Arthrosc (2010) 18: Table 3 continued Author Study design Subjects Time between ACL injuries Risk factor Results P = 0.02 Young age at ACL reconstruction (\21 years) P = 0.01 Mean 59 months (patellar tendon), and mean 32 months (hamstring tendon) Patellar tendon autograft 180 Patients (90 patients reconstructed with patellar tendon autograft, and 90 patients reconstructed with hamstring tendon autograft) (95 men and 85 women) Pinziewski et al. [73] Prospective 10 year followup investigation after ACL reconstruction with either patellar or hamstring tendon autograft ACL anterior cruciate ligament, NA Notch opening angle, MNW maximum notch width, CW condyle width, LCW lateral condyle width, NH Notch height, NWI notch width index; notch width, CT computer tomography and those with unilateral ACL injury. In a larger study population, Shelbourne et al. [91] showed that the notch width on plain radiographs was significantly smaller in both bilaterally and unilaterally ACL-injured individuals than controls. In contrast, Schickendantz and Welker found no significant differences between individuals with bilateral ACL injury, unilateral ACL injury and uninjured individuals when investigating notch width, or the width and height of the medial and lateral femur [88]. Another study showed that the ratio of the condylar width of the lateral femur to the overall condylar width was significantly higher in an ACL-injured group [42]. Family history Several studies have focused on possible correlations between family history of ACL injury and bilateral ACL injuries. In a retrospective study, Harner et al. [42] reviewed the medical records of 1,300 patients with known ACL injury, and 36 patients were found to have non-contact bilateral ACL injuries. They found a strong correlation between bilateral ACL injuries and having an immediate family member with an ACL injury. Eleven of 31 patients (35%) with bilateral ACL injuries had a family history of ACL injury, whereas only 1 of 23 (4%) of matched uninjured controls had a family history of ACL injuries. Using a similar study design, Souryal et al. [95] found no relation between relevant medical family history and bilateral ACL injuries. Nor could Andersson et al. [9] show a significant difference in family history between individuals with bilateral or unilateral ACL injury and uninjured controls. Furthermore, a prospective study was unable to show a correlation between individuals with a unilateral ACL injury and relevant family history, and those who sustained a contralateral ACL injury during follow-up [86]. Gender Studies on gender as a risk factor for contralateral ACL injury have not revealed any significant correlation, in a normally active population or in elite skiers [73, 79, 86, 111]. Salmon et al. [86] showed that 5 years after ACL reconstruction, 5% of the men, and 7% of the women had sustained a contralateral ACL injury. In another study it was found that among elite alpine skiers who ruptured one ACL while active, relatively more women also ruptured the contralateral ACL before their carrier was over (36% women and 27% men) [79]. Activity level Salmon et al. [86] analysed whether the primary ACL injury was a contact or non-contact injury, the activity level

8 284 Knee Surg Sports Traumatol Arthrosc (2010) 18: according to the International knee documentation committee (IKDC) scale, gender, graft type, family history of ACL injury, articular surface damage, the presence of meniscal injury or if meniscectomy had been performed, and the correlation to re-rupture or contralateral ACL injury. The only significant predictor for a contralateral ACL injury was a return to sports activity level 1 or 2 (IKDC activity level), i.e. pivoting sports. Compared to those participating in level 3 or 4 activities, the risk of contralateral ACL injury of subjects who returned to level 1 or 2 after ACL reconstruction was increased by a factor 10. Thirty-two out of 35 contralateral ACL injuries were in individuals who claimed to participate in moderate to strenuous exercise [86]. Laxity Motohashi et al. [65] showed that young women with bilateral ACL injuries had a higher generalised laxity score (the flexibility of 7 joints was assessed), compared with unilaterally injured women. Another study showed that individuals with bilateral ACL injuries were hypermobile in the metacarpophalangeal joint, compared with uninjured controls [42]. However, no correlation was found between hyperextension in the elbow, or thumb to forearm laxity, and bilateral ACL injuries [42]. Age at first ACL injury In a prospective follow-up after ACL reconstruction, it was shown that individuals younger than 21 years at the time of the ACL reconstruction were at greater risk of a contralateral ACL injury [73]. However, it was not discussed whether this was due to a return to higher activity levels among younger individuals. Graft choice A significant difference has been demonstrated between patients who have undergone ACL reconstruction with hamstring tendon autograft compared to patients who have undergone ACL reconstruction with patellar tendon autograft, in terms of contralateral ACL injury rate. There were significantly more contralateral ACL ruptures in the patellar tendon group at 10-year follow-up (20 ruptures vs. 9 ruptures) [73]. Discussion The most important finding of this review is that the risk of contralateral ACL injuries is greater than the risk of firsttime ACL injuries for individuals engaged in the same activities [32, 33, 70, 71]. A second important finding is that the intensity of activity after the unilateral ACL injury correlates to the risk of a contralateral ACL injury. From the literature search, it is unclear if surgical repair or nonsurgical treatment can influence on the risk of contralateral ACL injury. The only investigation were a non-surgical approach was amended demonstrated the lowest incidence rates of contralateral ACL injuries during follow-up [55]. However, in this investigation, the low incidence was probably influenced by the fact that the subjects were advised to avoid contact sports. The annual incidence of contralateral ACL injuries ranged from 1.1 to 2.5% in an ACL-injured population whose knees were reconstructed either with the goal of returning to sports involving pivoting, cutting, or sidestepping, or due to repeated episodes of instability of the injured knee, whereas the annual incidence for patients treated conservatively without initial ACL reconstruction and advised to avoid contact sports was 0.4% [22, 25, 55, 73, 74, 84, 86, 106]. The annual incidence in an athletic population ranged from 1.1% in female handball (top 3 divisions) to 11% in female toplevel soccer players [32, 33, 68]. Interestingly, a unilateral ACL injury seems to be a strong risk factor for a new ACL injury, either re-rupture or contralateral injury. The incidence of contralateral ACL injuries is greater than that of first-time ACL injuries for individuals engaged in the same activities [32, 33, 70, 71], indicating that the risk of a contralateral ACL injury is greater than that of a unilateral ACL injury. Furthermore, it must also be borne in mind that the risk of a contralateral ACL injury is only relevant for one knee, in contrast to the unilateral ACL injury which can occur in both knees. Therefore, when comparing the incidence of contralateral ACL injuries with a first-time ACL injury, the incidence should be multiplied by two. Moreover, individuals with a torn ACL are in general more susceptible to ACL rupture due to various intrinsic factors. Thus, the risk of these ACL-injured patients suffering a contralateral ACL rupture is, due to various intrinsic factors, probably higher than the risk of previously uninjured individuals sustaining an ACL injury, assuming they are exposed to the same types of extrinsic risk factors. Anatomic risk factors Most studies in which the involvement of the intercondylar notch and the risk of contralateral ACL injuries were investigated are retrospective. The majority of these studies show that patients with bilateral ACL injuries in general have smaller notches than uninjured controls, but because of the lack of information on other factors that could affect the risk of a contralateral ACL injury, such as return to sports after the first ACL injury, it is difficult to evaluate

9 Knee Surg Sports Traumatol Arthrosc (2010) 18: the relevance of notch stenosis as a risk factor for bilateral ACL injuries in these studies [9, 42, 65, 88, 95]. In a study by Shelbourne et al. [90] who followed patients after ACL reconstruction it was found that a narrow notch, measured intraoperatively, was associated with an increased risk of a contralateral ACL injury. These results could potentially be attributed to more women rupturing their ACL, considering that women have a smaller notch than men [8, 90]. However, Shelbourne et al. showed that although women sustained relatively more contralateral ACL injuries, there were no significant differences in contralateral ACL injury rate between men and women. Importantly, it was also demonstrated in that study that there were no differences in post-operative activity level between individuals with a narrow or a wide notch [90]. Family history Although several studies have been performed on the correlation between family history and bilateral ACL injuries, evidence linking family history and the risk of contralateral ACL injury is not convincing. One study showed a very strong correlation between bilateral ACL injuries and having an immediate-family member with ACL injury [42], whereas others have found no correlation [9, 86, 95]. These studies did not include the level of activity of family members, which makes the interpretation of the relevance of family history as a risk factor for contralateral ACL injuries difficult. Theoretically, immediate-family history is intriguing as a risk factor, and although the design of the study by Harner et al. [42] has several limitations, their findings may imply that family history is a risk factor. Additionally, Troijan et al. [101] showed that white European-Americans were at six times higher risk of sustaining an ACL injury than any other ethnic group among female basketball players. This vast difference between ethnic groups suggests that anthropometric factors may explain some of the difference in incidence between contralateral ACL ruptures and firsttime unilateral ACL injuries. Activity level The incidence of ACL injuries has been shown to vary between various sports, and inevitably, an individual will be at greater risk of an ACL injury in pivoting and sidestepping sports [3, 8, 66]. This has also been found for contralateral ACL injuries. It was shown that the risk of a contralateral ACL injury was much higher for individuals who claimed to participate in moderate to strenuous exercise [86]. The importance of a high level of activity is also indicated by the high prevalence of contralateral ACL injuries in elite skiers [79]. Gender Female gender is probably the most recognised risk factor for a unilateral ACL injury [3, 11, 82]. There is, however, no strong evidence that women are more susceptible to a contralateral ACL injury. According to both Salmon et al. [86] and Pinzcewski et al. [73], women exhibit no increased risk of sustaining a contralateral ACL injury. However, the risk of ACL injury in men and women seems to vary with different levels of activity, and a real difference can only be seen in high-intensity sports. There is strong evidence for such a difference in pivoting and sidecutting sports such as, wrestling, soccer, handball and basketball [3, 8, 66, 78]. The proportion of contralateral ACL injuries was higher among women than men in French national team skiers [79]. The observed difference was, however, not statistically significant, and the extent to which men and women returned to pre-injury competitive level after the unilateral ACL injury is not discussed in the article. Therefore, it is difficult to discern whether female gender actually constitutes a risk factor. In a population matched for activity level, Shelbourne et al. [90] were not able to demonstrate any significant differences in the risk of contralateral ACL injury between men and women. Another reason why no differences have been reported in the risk of contralateral ACL injuries between men and women may be that acquired risk factors secondary to the unilateral ACL injury overshadow the classic risk factors for an ACL injury such as gender. Age at first injury Pinczewski et al. [73] showed that individuals who are young at the time of ACL reconstruction have a greater risk of a contralateral ACL injury during follow-up. A patient with a ruptured ACL is likely to have a higher risk of having various intrinsic factors that make patients more susceptible to an ACL injury. Additionally, a young patient is probably more likely to return to a high level of activity than older individuals with a first-time ACL injury, and they are expected to have a higher level of activity for a longer period of time. Possible acquired risk factors for a contralateral ACL injury It has been suggested that the major risk factors for a contralateral ACL injury deviate from those of a primary unilateral ACL injury [37, 86]. For example, Salmon et al. [86] found that the risk of a contralateral ACL injury was not influenced by gender or family history of ACL injuries. Thus, the remarkably increased risk of a contralateral ACLinjury indicates additional risk factors acquired after the

10 286 Knee Surg Sports Traumatol Arthrosc (2010) 18: primary ACL injury [37]. However, to the best of our knowledge, no studies have been carried out to prospectively investigate whether risks arise as a result of the initial ACL injury, leading to an increased incidence of contralateral ACL injuries. A study of this nature would need a very large sample size and a long follow-up time, since the subjects would have to be examined before and after their initial ACL injury, and then followed to identify those who sustained a contralateral ACL injury. Biomechanical factors A likely mechanism through which factors acquired secondary to the initial ACL injury increase the risk of a contralateral ACL injury is through changes in the biomechanics of both the injured and the uninjured leg. Dynamic factors thought to influence ACL strain are knee kinematics and moments about the knee [40]. It could be speculated that a change in gait intended to stabilise the injured knee joint could simultaneously increase the risk of a contralateral ACL injury [86]. Abnormal gait biomechanics have been found both 3 weeks after ACL injury and 6 months after ACL injury [27]. Several others have also reported changes in the biomechanics of the uninjured leg after ACL injury [1, 13, 16, 38], including Berchuck et al. [16] who demonstrated that any imbalance in the kinematics of the injured leg was compensated for by the contralateral side, rendering an abnormal, yet symmetric gait. Gillquist and Messner [38] found a pivot-shift-like pattern of motion in the uninjured knee in one-third of patients after ACL reconstruction. Because of the close relationship between the lower extremities and trunk musculature, altered dynamics in one leg after an ACL injury could lead to an imbalance in trunk musculature, contributing to an increased risk of a contralateral ACL injury. In fact, Zazulak et al. [113] have shown that decreased neuromuscular control of the trunk increases the risk of sustaining an ACL injury. However, Salem et al. [85] found no differences in limb kinetics between the ACL-injured and the uninjured leg during a submaximal squatting exercise. They did, however, show that the peak knee extensor moment was significantly greater in the uninjured leg, and that the ratio between the peak hip extensor moment and the peak knee extensor moment was greater in the injured leg, thus indicating that although the kinetics were symmetric in the injured and uninjured limb, muscle activation patterns may still differ between the limbs [85]. Others have reported less flexion of the knee in ACL-deficient patients during gait compared with controls [16, 99]. In pivoting sports (basketball and handball), video analyses have shown that in general an ACL injury occurs when landing, with the knee in slight flexion. In this position, the ACL often ruptures as the tibia rotates inwards on the femur with an accompanying valgus collapse [17, 57]. In a study of cadaveric knees, it was shown that when anterior tibial load is applied, an erect position of the leg leads to an increased strain on the ACL, and in this position, the strain is dramatically accentuated by the addition of internal tibial torque [63]. Studies on cadaveric knees, furthermore, suggest that hamstring activity reduces the strain on the ACL, thus protecting the ACL from rupture in high-risk situations [29, 107], whereas the increased quadriceps force applied during landing leads to increased ACL strain [108]. Importantly, it has been shown that when the knee is at flexion angles smaller than 30, the hamstrings are unable to restrain the force placed on the ACL by simultaneous quadriceps contraction [81]. Less flexion in the knee joint and increased quadriceps-to-hamstrings contraction-force ratio have in fact been suggested as two of the factors making women more susceptible of an ACL injury than men [4]. This indicates that a small flexion angle in the knee joint secondary to a unilateral ACL injury also may increase the risk of a contralateral ACL injury. Side-to-side differences between the injured and the non-injured knee secondary to an ACL injury could also be an important independent risk factor for sustaining a contralateral ACL injury, implicated the findings by Myer et al. [67], who showed that large side-to-side differences in knee laxity predisposed individuals to an ACL tear. Neuromuscular factors Loss of afferent input from mechanoreceptors, resulting in proprioceptive (perception of movement and position) deficits have also been considered to be a contributing risk factor for a contralateral ACL injury [37, 80, 83]. Some individuals may also have an inherently inferior proprioceptive ability, making them vulnerable to bilateral ACL injuries [36]. The proprioceptive sense includes afferent signals from various mechanoreceptors within joints as well as in the musculature and skin. Mechanoreceptors found in the ACL include Pacinian corpuscles, Golgi tendon organs and Ruffini endings [30]. Two different mechanisms by which proprioception can protect the individual from joint injury have been suggested [99]. Firstly, it has been shown that the ACL can trigger a reflexive arch through afferent input, activating antagonist muscles (hamstrings), which oppose abnormal joint motion and thus maintain joint stability [93, 102]. The protective relevance of this mechanism is, however, controversial, and experiments indicate that the latency of this reflex through stimulation of the ACL is too long to activate the muscles to prevent ligament rupture in humans [56]. Secondly, sensory input from afferents in the ACL or elsewhere in the joint and musculature could offer protection

11 Knee Surg Sports Traumatol Arthrosc (2010) 18: by increasing neuromuscular awareness, by continuously regulating muscular stiffness around the knee joint through the gamma-muscle-spindle system [49]. Theoretically, loss of mechanoreceptors from one ACL might involve the contralateral limb, either indirectly, as a result of the altered dynamics of the injured limb, which are conducted to the contralateral limb, or directly through defects in the proprioception of the uninjured leg. The latter is dependent on the lack of afferent input to the spinal cord from the ruptured ACL. Furthermore, it has been hypothesised that the altered kinematics of the ACL-injured knee could lead to an altered afferent discharge from the mechanoreceptors of the injured joint, altering the muscle-spindle responsiveness bilaterally, and thus predisposing individuals with a unilateral ACL-injury to an injury of the contralateral ACL [99]. Several studies have shown proprioceptive insufficiency of the injured leg at various times after ACL injury when the injured knee was compared to that of a control group [2, 14, 48, 72, 80]. Studies have also shown decreased proprioceptive sensation of the injured knee as well as the contralateral knee when compared to external controls after ACL injury, as evidenced by significantly higher mean threshold to detection of passive motion (TDPM) [72, 80, 83]. Loss of afferent input from the ACL in one knee will probably induce proprioceptive deficits in the contralateral leg [80, 83]. Roberts et al. [83] suggested that the reduction in proprioception could lead to neuromuscular deficits. Konishi et al. [53] showed that the injection of a local anaesthetic into the knee joint cavity, attenuating the afferent signals from structures inside the joint cavity, reduced the maximal voluntary contraction of the quadriceps femoris of the affected leg. They concluded that an ACL injury leads to chronic reduction in Ia feedback to the muscles around the knee due to a lack of feedback from the ACL to gamma motor neurons. Interestingly, they were also able to demonstrate bilateral gamma loop dysfunction in the quadriceps femoris in patients with unilateral ACL injury [52, 54]. Perhaps, this mechanism could explain the bilateral neuromuscular deficits and disturbed balance of both legs in patients with an ACL injury [47, 109, 112]. Despite this, proprioceptive deficits seem to be compensated for over time in both individuals with non-reconstructed and reconstructed knees [37, 80]. Reider et al. [80] investigated 26 subjects (15 men and 11 women) with an ACL injury who underwent reconstruction of the knee weeks (mean 8 weeks) after injury. They found the mean TDPM to be significantly higher in the injured leg than in external controls preoperatively, 3 weeks and 6 weeks after surgery, but not 3 or 6 months after surgery. For the contralateral leg, the TDPM was only significantly higher preoperatively. Thus, proprioceptive defects in the contralateral knee seem to be quickly rectified after an ACL injury, making it unlikely that this is a relevant risk factor for a contralateral ACL injury [80]. However, a fraction of patients could have permanent defects in proprioception after an ACL injury [37, 80, 104]. These patients may very well be at increased risk of a contralateral ACL injury. Neuromuscular training programs have been shown to reduce the risk of non-contact ACL injuries in both men and women [6, 78]. As of today, the importance of neuromuscular treatment regimes to decrease the risk of sustaining a contralateral ACL injury has not been thoroughly investigated. Given the altered biomechanics and neuromuscular deficits after a unilateral ACL injury, such programs could be particularly beneficial for those patients who return to a high level of activity after ACL injury, to decrease the risk of sustaining a contralateral ACL injury. Ligament characteristics Little is known about how an ACL injury affects the ligaments of the contralateral ACL but, hypothetically, a loss of sensory input from the ruptured ACL could lead to alterations in the properties of the ACL secondary to the initial injury. In rabbits, a weakening of the contralateral collateral ligament has been reported after a unilateral experimental lesion [35]. Biochemical environment Dahlberg et al. [23] demonstrated a change in the cartilage metabolism of the uninjured knee joint secondary to an ACL injury, indicating possible alterations in the metabolism of ligaments, possibly weakening the mechanical properties of the ACL. Moreover, theoretically the altered biochemical environment of the joint might affect joint afferents, indirectly weakening the proprioceptive sensation. Thus, changes in the biochemical environment of the unaffected knee joint may be of interest in the quest for risk factors for a contralateral ACL injury. Limitations Psychological factors have not been reviewed or discussed in this paper but these could also contribute to the risk of contralateral ACL injury. Moreover, our search was based solely in the Medline database. Given that a search in the Medline database does not support several European journals that are better covered by the Embase database [10], some relevant studies may have been overlooked. In the majority of studies investigating contralateral ACL injuries as an outcome after a unilateral ACL injury, contralateral injuries are presented as a detail, and not