Can Biomechanical Testing After Anterior Cruciate Ligament Reconstruction Identify Athletes at Risk for Subsequent ACL Injury to the Contralateral Uninjured Limb?

Abstract

Background:

Athletes are twice as likely to rupture the anterior cruciate ligament (ACL) on their healthy contralateral knee than the reconstructed graft after ACL reconstruction (ACLR). Although physical testing is commonly used after ACLR to assess injury risk to the operated knee, strength, jump, and change-of-direction performance and biomechanical measures have not been examined in those who go on to experience a contralateral ACL injury, to identify factors that may be associated with injury risk.

Purpose:

To prospectively examine differences in biomechanical and clinical performance measures in male athletes 9 months after ACLR between those who ruptured their previously uninjured contralateral ACL and those who did not at 2-year follow-up and to examine the ability of these differences to predict contralateral ACL injury.

Study Design:

Case-control study; Level of evidence, 3.

Methods:

A cohort of male athletes returning to level 1 sports after ACLR (N = 1045) underwent isokinetic strength testing and 3-dimensional biomechanical analysis of jump and change-of-direction tests 9 months after surgery. Participants were followed up at 2 years regarding return to play or at second ACL injury. Between-group differences were analyzed in patient-reported outcomes, performance measures, and 3-dimensional biomechanics for the contralateral limb and asymmetry. Logistic regression was applied to determine the ability of identified differences to predict contralateral ACL injury.

Results:

Of the cohort, 993 had follow-up at 2 years (95%), with 67 experiencing a contralateral ACL injury and 38 an ipsilateral injury. Male athletes who had a contralateral ACL injury had lower quadriceps strength and biomechanical differences on the contralateral limb during double- and single-leg drop jump tests as compared with those who did not experience an injury. Differences were related primarily to deficits in sagittal plane mechanics and plyometric ability on the contralateral side. These variables could explain group membership with fair to good ability (area under the curve, 0.74–0.80). Patient-reported outcomes, limb symmetry of clinical performance measures, and biomechanical measures in change-of-direction tasks did not differentiate those at risk for contralateral injury.

Conclusion:

This study highlights the importance of sagittal plane control during drop jump tasks and the limited utility of limb symmetry in performance and biomechanical measures when assessing future contralateral ACL injury risk in male athletes. Targeting the identified differences in quadriceps strength and plyometric ability during late-stage rehabilitation and testing may reduce ACL injury risk in healthy limbs in male athletes playing level 1 sports.

Clinical Relevance:

This study highlights the importance of assessing the contralateral limb after ACLR and identifies biomechanical differences, particularly in the sagittal plane in drop jump tasks, that may be associated with injury to this limb. These factors could be targeted during assessment and rehabilitation with additional quadriceps strengthening and plyometric exercises after ACLR to potentially reduce the high risk of injury to the previously healthy knee.

The primary concern after anterior cruciate ligament (ACL) reconstruction (ACLR) is minimizing risk of rerupture of the reconstructed ACL.^28,30 Risk of reinjury to the reconstructed graft^43,48 as well as to the native ACL on the contralateral limb⁵⁰ is considerably higher than risk of ACL injury in previously uninjured healthy athletes.^39,48,53,57 Furthermore, a review of rates of second ACL injury within 5 years revealed a pooled incidence of 5.8% for injury to the ipsilateral operated limb and 11.8% for ACL injury of the contralateral limb.⁵⁸ Given this high injury rate after ACLR, identifying risk factors for ACL injury to the contralateral healthy knee that can be addressed or targeted during rehabilitation may be important for improving short- and long-term outcomes for athletes.

Multiple factors have been outlined in previous research as requiring consideration in the return-to-play process to mitigate against future injury: time from surgery, muscle strength, clinical examination, hop testing, performance-based criteria, and patient-reported outcomes (PROs).³ However, the validity of these measures collectively or in isolation in identifying those who will experience adverse outcomes is unknown.^3,52 PRO measures and symmetry of clinical performance measures (isokinetic strength, jump performance, and change-of-direction (CoD) time) are commonly used in combination to assess rehabilitation status after ACLR and have been suggested to influence injury risk to both knees after ACLR.^12,28 However, these studies did not examine contralateral knee injuries to identify risk factors specific to injury in the previously healthy knee.

Landing and CoD are the 2 most common ACL injury mechanisms.¹ Biomechanical variables during landing have been suggested to predict ACL second injury after ACLR, yet CoD has not been explored. Paterno et al⁴⁰ identified several biomechanical factors during double-leg drop jump (DLDJ) tests that can predict second ACL injury, including uninvolved limb-hip rotation moment, asymmetry of knee extension moment at initial contact, and knee valgus range of motion during landing. However, this study combined male and female athletes, did not report variables specific to injury to the ACLR knee or contralateral knee, and did not examine single-leg drop jump (SLDJ), even though single-leg landing is a more common injury mechanism. Biomechanical differences in kinetic and kinematic variables in all 3 planes relating to the ankle, knee, hip, and thorax to pelvis in jump and CoD tests have been demonstrated between ACLR and contralateral limbs in male athletes 9 months after ACLR.^21,22 These same asymmetries are greater than those in healthy uninjured control athletes, potentially because of incomplete rehabilitation of the ACLR limb.²³ Whether these biomechanical differences in relation to greater asymmetry (insufficient rehabilitation of the ACLR limb) or deficits specific to the contralateral limb influence injury risk to the contralateral knee has not been prospectively examined. Biomechanical differences have been reported despite no differences in hop and CoD performance between limbs. There were, however, large performance differences during the SLDJ, which is a measure of plyometric ability.²²

Plyometric ability, as measured by reactive strength, refers to the capacity to absorb and then produce force over short ground contact times, primarily using the stretch shortening cycle and thus maximizing whole body stiffness. These deficits reflect an inability to absorb and produce force during landing and may reflect a relevant injury risk factor. Biomechanical differences during jump and CoD tests have been found between those who rerupture the reconstructed ACL graft and those who do not, despite no differences in clinical performance measures.²⁰ However, non–function related factors such as graft type,²⁴ graft healing time,⁵ and surgeon experience⁴⁹ may influence ipsilateral graft rerupture but are not applicable to contralateral ACL injury. Therefore, investigation of the influence of biomechanical and performance measures on risk of ACL injury to the contralateral knee is warranted.

The aim of this study was to identify differences in strength, jump, and CoD performance, PROs and landing and CoD biomechanics associated with future ACL injury to the contralateral limb and to assess the ability of these differences to predict who will be injured. Our hypothesis was that there would be differences in strength and biomechanics throughout the kinetic chain during jump and CoD testing and that these variables will predict contralateral injury (CI).

METHODS

Athletes were recruited into this prospective case-control study at the Sports Surgery Clinic (Dublin, Ireland) before ACLR from January 1, 2014, to December 31, 2016. Before surgery, athletes completed a preoperative questionnaire outlining their sport, mechanism of injury, and level of desired return after surgery. Men aged 18 to 35 years who played level 1 sports (multidirectional field sports involving landing, pivoting, and CoD) and intended to return to the same level of sport were included in the study (N = 1045). All participants underwent primary ACLR using either a bone–patellar tendon–bone graft or hamstring graft (gracilis/semitendinosus tendons) from the ipsilateral limb. Patients were excluded if they were undergoing second or subsequent ACLR, did not intend to return to level 1 sports, or had meniscal or additional ligament repair at the time of surgery. The study was registered at ClinicalTrials.gov (NCT02771548) and received approval from the clinic’s ethic committee (25-AFM-010).

Testing Protocol

After ACLR, all participants underwent a rehabilitation protocol with weightbearing as tolerated on crutches for 2 weeks, followed by progressive blocks of strength, power, and plyometric exercises, progressing to on-field running and CoD. Athletes underwent rehabilitation locally by their referring physical therapist and review by their orthopaedic surgeon at 2 weeks, 3 months, and 6 to 9 months after surgery. As part of their final orthopaedic review, athletes took part in a physical testing protocol at 9 months (range, 8–10 months) after surgery. Before testing, all participants completed the following PROs: International Knee Documentation Committee (0–100),¹⁹ Marx Activity Scale (0–16),³⁴ and ACL Return to Sport After Injury questionnaire (0–100),⁵⁵ with higher scores reflecting higher self-reported knee function, activity levels, and self-reported readiness to return to sport, respectively. A list of the acronyms used in the article is outlined in Table 1.

TABLE 1.

Abbreviations

ACL	Anterior cruciate ligament
ACLR	Anterior cruciate ligament reconstruction
AUC	Area under the curve
CI	Contralateral injury
CoD	Change of direction
COM	Center of mass
DLDJ	Double-leg drop jump
LSI	Limb symmetry index
NCI	No contralateral injury
PRO	Patient-reported outcome
SLCMJ	Single-leg countermovement jump
SLDJ	Single-leg drop jump
SLHD	Single-leg hop for distance

Data were collected in a 3-dimensional (3D) biomechanics laboratory as part of a larger prospective research project and included a DLDJ from 30 cm, a SLDJ from 20 cm, and 90° planned and unplanned CoD,^21,22 as well as measurement of single-leg countermovement jump (SLCMJ) height and single-leg hop for distance (SLHD).^12,28,38 Participants performed a standardized warm-up: 2-minute jog, 5 body-weight squats, and 2 submaximal and 3 maximal double-leg countermovement jumps. Each participant performed 2 submaximal practice trials of each movement before 3 valid test trial attempts were captured (maximal effort and full foot contact on force plate), with the mean of the 3 trials used for analysis. A 30-second recovery was taken between trials. Laboratory testing was followed by concentric isokinetic testing of quadriceps and hamstring muscle groups in both limbs at 60 deg/s from 0° to 100° of knee flexion, with peak torque per body mass reported.⁵¹

Movement mechanics data collection took place using an 8-camera motion analysis system (Bonita-B10; Vicon) capturing at 200 Hz. Movement was synchronized with 2 force platforms (BP400600; AMTI) sampling at a frequency of 1000 Hz and recording motion data from 24 reflective markers (diameter, 14 mm) as well as ground-reaction forces (GRFs; Nexus 1.8.5; Vicon). These data were low-pass filtered using a fourth-order Butterworth filter (cutoff frequency of 15 Hz).²⁶ Markers were placed on the lower legs and trunk according to the adapted Plug-in-Gait (Vicon), and kinematic data were calculated.³³ Performance measures were calculated for jump (height and length) and CoD (time) tasks. Jump height was calculated using the take-off vertical velocity derived from the vertical GRF signal using the impulsemomentum theorem. Jump length was calculated as the horizontal distance from heel marker at the start of the jump to the landing using MATLAB (Version R2015a; The MathWorks Inc). Reactive strength index was calculated for the DLDJ and SLDJ as jump height divided by ground contact time.¹³ Time to complete the 90° CoD was recorded using speed gates (Smartspeed; Fusion Sport) with a trigger gate 2 m from the start line and an exit gate 2 m to the left and right of the force plates to indicate the end of the maneuver.²¹

Standard inverse dynamics analysis was used to calculate kinetic variables (reported as internal moments) at the ankle, knee, and hip. All kinetic variables were normalized to body mass. A custom MATLAB program was used for processing and calculating trunk-to-pelvis angles and distance from center of mass (COM) to ankle and knee joint in all 3 planes.²² Whole body stiffness when the body was accepting load was calculated as

$$\text{stiffness}(k)=\text{Δ}\text{vertical}\text{GRF}/\surd \left(\text{Δ}\text{COM}{z}^2\right),$$

where the delta for both variables is from the impact (the point of initial ground contact) to the end of the eccentric phase, defined as the first instance at which COM vertical power >0. Kinetic and kinematic analysis was carried out for the stance phase of each jump and CoD test (defined by GRF >20 N). Curves were normalized to 101 frames and landmark registered to the point when COM power reached zero in the z-axis (vertical), aligning onset of the eccentric phase to 50% of the stance phase, to ensure appropriate comparison of neuromuscular characteristics between limbs and participants during continuous waveform analysis.^35,44 Limb symmetry index (LSI) for strength and jump performance measures was calculated as follows: (ACLR side/contralateral side) × 100. The magnitude of asymmetry of the biomechanical variables was calculated by subtracting the contralateral limb from the ACLR limb throughout the stance phase.

Follow-up

Participants were followed up via email to identify second ACL injuries (ie, confirmed on magnetic resonance imaging to the ACLR or contralateral knee) at 1 and 2 years after surgery using a return-to-play questionnaire, or they were identified if they returned to their original surgeon with a diagnosis of another ACL injury. If participants did not reply to the email questionnaire, they received a follow-up phone call to complete the questionnaires. All participants who had surgery and were identified to have an ACL injury to the contralateral knee but no injury to the ACLR knee were included in the CI group (n = 67), which set the sample size for the study. A cohort of participants who had returned to multidirectional field sports after ACLR and had not experienced a second ACL injury to either knee at 2-year follow-up were assigned to the NCI group (no CI). The NCI group was matched to the CI group for time from surgery to return to play, time from surgery to 3D biomechanical testing, age, and distribution of graft type (n = 60) to ensure appropriate comparison and minimize the potential influence of nonphysical factors on contralateral ACL injury (Figure 1).

Figure 1.

Flow diagram of matching process for the study groups: contralateral injury and no contralateral injury. 3D, 3-dimensional; ACL, anterior cruciate ligament; ACLR, anterior cruciate ligament reconstruction; RTP, return to play.

Statistical Analysis

Differences between the CI and NCI groups for the following were examined using a Student independent t test: LSI, PROs, isokinetic peak torque of quadriceps and hamstrings muscles, planned and unplanned 90° CoD time, and SLDJ, SLCMJ, and SLHD jump performance on the contralateral side. Effect sizes for differences between groups for each variable were calculated using Cohen d (0.2–0.49, small; 0.5–0.79, medium; ≥0.8, strong).⁶ Odds ratios were calculated for patients in the NCI group when they had ≥90% LSI for quadriceps strength, hamstring strength, SLCMJ, and SLHD jump height for all 5 tests collectively. Statistical parametric mapping (1D unpaired t test, parametric) was used to examine differences in biomechanical variables (vertical GRF; angles and moments at hip, knee, and ankle; thorax-to-pelvis angles; and COM to ankle and knee in all 3 planes) between the CI and NCI groups for the contralateral limb. Asymmetry between limbs (ACLR minus contralateral) was compared between groups for each biomechanical variable for the DLDJ, SLDJ and planned and unplanned 90° CoD during stance phase. Mean effect size across phases with significant differences (P < .05) was reported, excluding phases with Cohen d < 0.5. Time points and mean effect sizes with a significant difference between the groups and mean values for each group across that phase are reported. Graphs for biomechanical variables with differences are displayed in the Appendix (available in the online version of this article).

To assess the ability of the results to predict ACL reinjury, logistic regressions were performed using a maximum of 5 predictor variables that were chosen per the largest effect sizes of the identified differences for the magnitude and symmetry analysis. Only these features were chosen to achieve an input:observations ratio of 1:10 to 1:15 to generate a model that avoided overfitting of the data.^2,41 It should be noted that if a feature was multicollinear with a higher-ranked feature (correlation >.70), it was excluded and an additional lower-ranked feature was included. Predictor variables utilized were the average value of the phases within a biomechanical wave form that differed between groups. Before fitting the logistic regression, predictor variables were transformed into z scores, and cohorts were balanced so that the sample sizes of the CI and NCI groups were equal. To transform a predictor variable vector x into z scores (eg, contact time; n × m, n = 88 participants, m = 1 feature), the following equation was used:

$$z=(\mathbf{\text{x}}-\overline{\text{x}})/S$$

where $\overline{\text{x}}$ is the average and S is the standard deviation of the sample within x. During the fitting, data were balanced using the synthetic minority oversampling technique⁴ so that the minority class contained the same number of observations as the majority class. To interpret the predictive ability of the logistic regression, receiver operating curve and prediction accuracy are reported. Area under the curve (AUC) was used to classify findings (nil, 0.50; poor, >0.60; fair, >0.70; good, >0.80), while the accuracy measure was compared with the expected accuracy (ie, if the most frequent class was guessed). A summary of the data points and statistical analysis is outlined in Table 2.

TABLE 2.

Data Points and Statistical Analysis^a

Data Set	Analysis
PRO data	Mann-Whitney U test
Strength, jump, and CoD performance	Independent Student t test
Contralateral side and LSI	Odds ratio: CI if ≥90% LSI Logistic regression
Biomechanics contralateral side and ASYM	1D SPM independent Student t test Logistic regression

^a1D, 1 direction; ASYM, asymmetry; CI, contralateral injury; CoD, change of direction; LSI, limb symmetry index; PRO, patient-reported outcome; SPM, statistical parametric mapping.

RESULTS

Of the 1045 primary ACLRs, there were 67 contralateral ACL injuries and 38 ipsilateral ACL injuries, and 52 were lost to follow-up (95% follow-up). Of those participants who experienced a contralateral ACL injury (CI group), 3D biomechanical analysis was recorded for 55 (12 did not attend follow-up) and matched to 60 athletes who did complete it but who did not experience an ACL injury to either knee 2 years after surgery (NCI group). Mean ± SD time to CI was 23.3 ± 9.8 months (Table 3). There was no significant difference between the groups in International Knee Documentation Committee, ACL Return to Sport After Injury questionnaire, or Marx Activity Scale scores (Table 4).

TABLE 3.

Patient Data^a

	Group, No. or Mean ± SD
	CI	NCI	P Value
Participants	55	60
Graft type, BPTB:HT	46:9	48:12	.61
Age, y	21.3 ± 4.2	21.9 ± 4	.43
Mass, kg	80.7 ± 10	81.5 ± 11.6	.69
Height, cm	179.4 ± 6.3	180.4 ± 5.6	.36
Time from surgery to, mo
RTP	10.3 ± 4.3	9.7 ± 2.3	.35
Testing	9.0 ± 3.1	9.4 ± 1.2	.32
Reinjury	23.3 ± 9.8
RTP to reinjury, mo	13.0 ± 9.5

^aBPTB, bone–patellar tendon–bone; CI, contralateral injury; HT, hamstring tendon; NCI, no contralateral injury; RTP, return to play.

TABLE 4.

PRO Measures for the CI and NCI Groups^a

	Group, Mean ± SD
	CI	NCI	P Value	Effect Size
IKDC	79.1 ± 12.0	82.4 ± 10.6	.17	0.21
ACL-RSI	75.8 ± 17.8	78.1 ± 15.3	.49	0.10
Marx Activity Scale	10.8 ± 3.5	11.2 ± 3.2	.29	0.12

^aACL-RSI, Anterior Cruciate Ligament Return to Sport After Injury questionnaire; CI, contralateral injury; IKDC, International Knee Documentation Committee; NCI, no contralateral injury; PRO, patient-reported outcome.

Strength, Jump, and CoD Performance Measures

There was a significant difference with a small effect size (d = 0.39) in quadriceps peak torque on the contralateral side, with significantly lower strength in the CI group (Table 5). No difference was observed between groups on the contralateral side for hamstring strength, SLCMJ and SLDJ height, or SLHD or for the corresponding LSI. The odds of being in the NCI group were 0.54 (95% CI, 0.02–16.39) if the athlete achieved ≥90% LSI across all 5 tests. Similarly, no differences were detected between contralateral limbs in planned CoD performance time (1.45 ± 0.12 vs 1.42 ± 0.08 seconds; P = .162) or LSI (98.9% ± 4.8% vs 98.9% ± 4.7%; P = .982) or for the unplanned CoD (1.56 ± 0.02 vs 1.52 ± 0.09 seconds; P = .206) or LSI (98.5% ± 4.5% vs 98.3% ± 5.3%; P = .840).

TABLE 5.

Strength and Jump Performance Measures and LSI^a

	CI		NCI
	Mean ± SD	95% CI	Mean ± SD	95% CI	P Value	Effect Size
Quadriceps strength, N/kg	216.3 ± 38.8	206–227	231.3 ± 36.3	222–240	.032b	0.39
LSI, %	80.9 ± 14.6	76–85	84.2 ± 14.6	80–88	.235	0.22
≥90% LSI success rate, %	31		36		.593
Hamstring strength, N/kg	127.3 ± 24.9	120–134	135.7 ± 23.4	130–142	.063	0.34
LSI, %	96.9 ± 14.5	92.9–100	96.5 ± 10.6	93–99	.894	0.02
≥90% LSI success rate, %	73		73		.982
SLCMJ, cm	12.1 ± 2.3	11.5–12.8	11.9 ± 2.4	11.2–12.5	.561	0.11
LSI, %	85.8 ± 13.2	82–90	84.4 ± 14.6	81–88	.627	0.09
≥90% LSI success rate, %	40		38		.792
SLDJ, cm	12.1 ± 3.2	11.2–13.0	12.4 ± 2.7	11.7–13.1	.564	0.11
LSI, %	78.1 ± 16.7	73–83	74.1 ± 14.8	70–78	.186	0.25
≥90% LSI success rate, %	12		18		.393
SLHD, cm	152.3 ± 27.0	144–160	154.9 ± 19.9	150–160	.562	0.11
LSI, %	95.1 ± 15.5	90–99	94.2 ± 12.4	91–97	.749	0.06
≥90% LSI success rate, %	61		66		.645
≥90% LSI success rate for all 4 tests, %	2		2		.921

^aCI, contralateral injury; LSI, limb symmetry index; NCI, no contralateral injury; SLCMJ, single-leg countermovement jump; SLDJ, single-leg drop jump; SLHD, single-leg hop for distance;

^bP < .05.

Biomechanical Analysis

Differences on the Contralateral Side.

No significant differences were detected in joint mechanics during planned and unplanned CoD. For DLDJ, there were strong effect size differences between groups on the contralateral side for ground contact time (d = 0.83), COM vertical stiffness (d = 0.80), and COM vertical distance to the knee and ankle (both d = 0.80), with significantly longer contact times, less COM stiffness, and lower COM distances in the CI group (Table 6, Figure 2). There were medium effect size differences between the groups for vertical GRF (30%–73%, d = 0.74; 83%–99%, d = 0.78) (Figure 3), with significantly lower vertical GRF in the CI group through most of the stance but higher toward the end. This was reflected in lower reactive strength index in the CI group (d = 0.62).

TABLE 6.

Differences Between Groups in Biomechanical Variables on the Contralateral Side During Double-Leg Drop Jump^a

	Gait Cycle, %		CI		NCI
	Start	End	Mean ± SD	95% CI	Mean ± SD	95% CI	P Value	Effect Size
Contact time, s			0.34 ± 0.10	0.32 to 0.37	0.27 ± 0.06	0.25 to 0.29	<.001	0.83
COM Stiffness, N/kg/mm			91.2 ± 48.8	77.5 to 104.9	133.5 ± 50.7	120.3 to 146.7	<.001	0.80
COM to ankle vertical, mm/BH	10	93	0.41 ± 0.02	0.40 to 0.42	0.43 ± 0.02	0.42 to 0.44	<.001	0.80
COM to knee vertical, mm/BH	11	92	0.22 ± 0.02	0.21 to 0.22	0.23 ± 0.14	0.23 to 0.24	<.001	0.80
Vertical GRF, N/kg	30	73	18.0 ± 4.6	16.7 to 19.3	21.4 ± 3.7	20.4 to 22.4	<.001	0.74
	83	99	4.1 ± 1.4	3.7 to 4.5	3.0 ± 0.9	2.8 to 3.3	<.001	0.78
Hip flexion angle	14	95	54.7 ± 12.4	51.3 to 58.3	45.3 ± 9.9	42.7 to 47.9	<.001	0.76
Ankle plantarflexion moment, N-m/kg	22	74	2.2 ± 0.7	2.0 to 2.4	2.7 ± 0.6	2.6 to 2.8	<.001	0.76
	84	93	0.7 ± 0.3	0.6 to 0.8	0.5 ± 0.2	0.4 to 0.6	.004	0.68
Knee flexion angle	14	94	63.8 ± 12.5	60.3 to 67.4	55.6 ± 8.8	53.3 to 57.9	<.001	0.71
Moment, N-m/kg
Knee extension	3	7	0.01 ± 0.42	−0.12 to 0.11	−0.24 ± 0.26	−0.17 to −0.31	.027	0.62
	17	21	1.3 ± 0.6	1.1 to 1.4	0.9 ± 0.5	0.8 to 1.1	.030	0.60
	44	59	2.4 ± 0.8	2.2 to 2.6	2.8 ± 0.6	2.6 to 3.0	<.001	0.59
	82	93	0.02 ± 0.5	−0.1 to 0.2	−0.4 ± 0.4	−0.3 to −0.5	.001	0.72
Hip extension	0	6	0.5 ± 0.6	0.3 to 0.6	0.8 ± 0.5	0.7 to 0.9	.005	0.62
	62	82	0.7 ± 0.6	0.5 to 0.8	0.2 ± 0.5	0.1 to 0.4	<.001	0.71
Hip external rotation	4	8	0.03 ± 0.07	0.01 to 0.04	−0.02 ± 0.05	−0.03 to −0.01	.021	0.69
	94	98	0.01 ± 0.05	0 to 0.03	−0.02 ± 0.05	−0.03 to 0.05	.024	0.64
Knee valgus	42	62	1.5 ± 0.6	1.3 to 1.6	1.9 ± 0.7	1.7 to 2.1	<.001	0.60
	84	94	0.3 ± 0.2	0.2 to 0.4	0.1 ± 0.2	0.1 to 0.2	.010	0.64
Reactive strength, cm/s			0.8 ± 0.2	0.7 to 0.8	0.9 ± 0.2	0.8 to 1.0	<.001	0.62
Anterior pelvic tilt, deg	43	88	23.7 ± 6.1	22.0 to 25.4	19.8 ± 5.8	18.4 to 21.4	.009	0.61
Thorax-to-pelvis extension, deg	24	100	5.5 ± 7.6	3.4 to 7.7	10.1 ± 5.9	8.5 to 11.6	.007	0.60

^aBH, body height; CI, contralateral injury; COM, center of mass; GRF, ground-reaction force; NCI, no contralateral injury.

Figure 2.

Biomechanical differences on contralateral side during double-leg drop jump for the study groups: contralateral injury (bold image) as compared with no contralateral injury (blurred image). COM, center of mass; vGRF, vertical ground-reaction force.

Figure 3.

Vertical ground-reaction force on the contralateral side during first ground contact of double-leg drop jump for the study groups: contralateral injury (CI) and no contralateral injury (NCI; matched cohort). Top panel: mean and SD clouds for the CI (black) and NCI (blue) groups. Middle panel: SPM{t}, the t statistic as a function of time describing the difference between the groups. Bottom panel: effect size as a function of time, describing the magnitude of the effect. Shaded portions of the bottom panel indicate average Cohen d >0.5, with orange indicating medium effect size throughout those phases. ACL, anterior cruciate ligament; SPM, statistical parametric mapping.

Several significant joint kinematic differences, primarily in the sagittal plane, were detected between the CI and NCI groups, including more hip flexion (14%–95%; d = 0.76), knee flexion (14%–94%; d = 0.71), ankle dorsiflexion (69%–92%; d = 0.63), anterior pelvic tilt (43%–88%; d = 0.61), and thorax-to-pelvis flexion (24%–100%; d = 0.6) in the CI group. In addition, there were several joint kinetic differences between the CI and NCI groups in the sagittal plane, including lower and then greater hip extension moment (0%–6%, d = 0.62; 62%–82%, d = 0.71), lower ankle plantarflexion moment through midstance and greater at end stance (24%–74%, d = 0.76; 84%–93%, d = 0.68), and increased knee extension moment in the early and late stance but lower in midstance (3%–7%, d = 0.62; 17%–21%, d = 0.60; 44%–59%, d = 0.59; 82%–93%, d = 0.72) on the contralateral side in the CI group.

Outside the sagittal plane, there was less knee valgus moment during midstance, followed by greater valgus moment at the end of stance (42%–62%, d = 0.60; 84%–94%, d = 0.64). The variables selected for inclusion in the regression model included contact time, COM to ankle, hip extension moment (62%–82%), and hip rotation moment (both phases identified as significantly different) and could predict membership of the CI group with an accuracy of 71.2% (baseline, 53.2%), with a sensitivity of 0.83 and a specificity of 0.58 (AUC = 0.80).

In the SLDJ, similar biomechanical differences in the sagittal plane were again evident between the CI and NCI groups on the contralateral side (Table 7, Figure 2). There was significantly less distance vertically from COM to knee (12%–83%; d = 0.73) and ankle (12%–88%; d = 0.70), longer ground contact times (d = 0.70), less COM stiffness vertically (d = 0.70), and lower reactive strength (d = 0.50) on the contralateral side in the CI group. Furthermore, there was higher, then lower, then higher vertical GRF in the CI group (3%–11%, d = 0.65; 32%–68%, d = 0.69; 86%–99%, d = 0.63). In the sagittal plane, there was significantly increased hip flexion (14%–88%; d = 0.59), increased knee flexion (18%–24%, d = 0.52; 64%–92%, d = 0.58), increased ankle dorsiflexion (84%–88%; d = 0.52), and increased trunk-on-pelvis flexion (23%–43%; d = 0.50) in the CI group. In addition, there was significantly higher hip extension moment (74%–79%; d = 0.61), increased knee extension moment in the early and late stances (13%–18%, d = 0.60; 83%–89%, d = 0.58), as well as reduced ankle plantarflexion moment through midstance (22%–63%; d = 0.61) in the CI group. In the frontal plane, there was significantly greater internal knee valgus moment (11%–15%; d = 0.58) and ipsilateral thorax-on-pelvis side flexion (54%–72%; d = 0.52) in the CI group. There were no differences in the transverse plane. COM to knee, COM stiffness, vertical GRF (3%–11% and 33%–68%), and hip extension moment were selected for the regression model and could predict membership of the CI group with an accuracy of 62.1% (baseline, 53.2%), with a sensitivity of 0.51 and a specificity of 0.75 (AUC = 0.75).

TABLE 7.

Biomechanical Differences on the Contralateral Side During Single-Leg Drop Jump^a

	Gait Cycle, %		CI		NCI
	Start	End	Mean ± SD	95% CI	Mean ± SD	95% CI	P Value	Effect Size
COM to knee vertical, mm/BH	12	84	0.24 ± 0.01	0.24 to 0.25	0.25 ± 0.01	0.25 to 0.26	<.001	0.73
Contact time, s			0.39 ± 0.08	0.37 to 0.41	0.33 ± 0.05	0.32 to 0.35	<.001	0.70
COM Stiffness, N/kg/mm			138.3 ± 54.8	122.8 to 153.6	180.1 ± 56.4	165.6 to 194.7	<.001	0.70
COM to ankle vertical, mm/BH	12	89	0.44 ± 0.02	0.43 to 0.45	0.46 ± 0.01	0.45 to 0.46	<.001	0.7
Vertical GRF, N/kg	3	11	9.8 ± 3.1	8.9 to 10.7	8.2 ± 1.5	7.8 to 8.6	.002	0.65
	33	68	25.1 ± 4.5	23.8 to 26.3	28.2 ± 3.9	27.2 to 29.2	<.001	0.69
	87	99	4.4 ± 1.5	2.3 to 6.5	3.5 ± 1.1	1.7 to 5.4	<.001	0.63
Moment, N m/kg
Hip extension	74	79	0.3 ± 0.7	0.1 to 0.5	−0.2 ± 0.57	−0.3 to 0	.004	0.61
Ankle plantarflexion	22	63	2.9 ± 0.6	2.7 to 3.1	3.4 ± 0.7	3.2 to 3.6	<.001	0.61
Knee extension	13	18	1.0 ± 0.8	0.7 to 1.3	0.5 ± 0.67	0.24 to 0.77	.020	0.60
	83	89	0.2 ± 0.5	0 to 0.5	−0.1 ± 0.46	−0.35 to 0.23	.010	0.58
Flexion angle
Hip	14	88	43.8 ± 9.2	41.2 to 46.4	38.5 ± 7.0	36.8 to 40.4	<.001	0.59
Knee	18	22	51.8 ± 8.9	49.3 to 54.3	47.5 ± 7.1	45.7 to 49.4	.040	0.52
	64	92	40.7 ± 9.2	38.2 to 43.3	35.5 ± 7.7	33.6 to 37.5	.003	0.58
Knee valgus moment, N m/kg	11	15	0.9 ± 0.4	0.7 to 1.0	0.7 ± 0.3	0.6 to 0.8	.030	0.58
Ankle dorsiflexion, deg	84	88	1.3 ± 7.4	−3.6 to 6.2	−2.6 ± 6.9	−7.5 to 2.3	.040	0.52
Thorax to pelvis, deg
Side flexion	54	72	0.8 ± 4.9	−0.5 to 2.2	−1.7 ± 4.6	−2.9 to 4.6	.020	0.52
Extension	23	43	−2.5 ± 9.2	−5.0 to 0.1	2.1 ± 8.3	−0.1 to 4.2	.030	0.51
Reactive strength, cm/s			0.32 ± 0.12	0.29 to 0.35	0.37 ± 0.09	0.35 to 0.40	.010	0.50

^aBH, body height; CI, contralateral injury; COM, center of mass; GRF, ground-reaction force; NCI, no contralateral injury.

Difference in Asymmetry Between Groups.

There was no significant difference in asymmetry between groups for the SLDJ or planned or unplanned CoD. In the DLDJ, there was significantly greater asymmetry in the CI group for knee varus angle (91%–100%; d = 0.66), with less knee varus on the contralateral limb (Table 8).

TABLE 8.

Differences in Asymmetry Between Groups for the Double-Leg Drop Jump^a

	Gait Cycle, %		CI		NCI
	Start	End	Mean ± SD	95% CI	Mean ± SD	95% CI	P Value	Effect Size
Knee varus angle	91	100	1.0 ± 2.9	0.9 to 1.2	−0.7 ± 2.1	−0.6 to −0.8	.03	0.66

^aCI, contralateral injury; NCI, no contralateral injury.

DISCUSSION

This study found that there were quadriceps strength and biomechanical differences, primarily in the sagittal plane, during plyometric tests on the contralateral side 9 months after surgery for male athletes who experienced CI after ACLR as compared with those who did not at 2 years after reconstruction. These differences had fair to good ability to predict risk of future CI and were present despite no difference in LSI between groups and minimal biomechanical asymmetry between groups. Given the higher contralateral ACL injury rate reported in the literature, compared with the reconstructed ACL, this study highlights the importance of assessing the contralateral limb and suggests tests and variables that should be targeted during rehabilitation and return-to-play testing that may play an important role in minimizing risk of contralateral ACL injury after ACLR.

To our knowledge, the association between strength and jump performance measures and contralateral ACL injury has not been investigated previously. This study demonstrated no significant difference in LSI for quadriceps and hamstring strength, jump testing, and timed CoD performance between the CI and NCI groups. In addition, when achievement of ≥90% LSI across strength and jump tests was combined, it had little influence on the odds of having a CI (odds ratio, 0.54; 95% CI, 0.02–16.39). Furthermore, few differences in asymmetry of biomechanical variables between groups were evident. The only finding was increased asymmetry of knee varus angle in DLDJ at the end of stance. This limited number of findings suggest that asymmetry may not be a major factor in subsequent contralateral ACL injury.

There were several differences between groups in the sagittal plane during the DLDJ and SLDJ. The contralateral limb in the CI group demonstrated differences in plyometric ability and whole body stiffness as compared with the NCI group, as reflected in differences in reactive strength index (but not jump height). In the DLDJ and SLDJ, there were longer ground contact times, reduced COM stiffness, and greater drop of the COM vertically relative to the knee and ankle in the CI group. This was accompanied by increased flexion at the hip, knee, ankle, and thorax and by differences in kinetic variables in the sagittal plane with greater, then less, then greater vertical GRF, ankle plantarflexion moment, and knee extension moment as well as changes in hip extension moment in the CI group. This reduction in reactive strength (driven by longer ground contact times) in combination with higher vertical GRF and higher knee extension moments early in stance may be a major contributor to excessive ACL strain and subsequent ACL injury.^11,17,32 Greater knee flexion, longer ground contact times, and greater drop of the COM relative to the ankle during DLDJ have also been identified in male athletes who rerupture their reconstructed knee after ACLR.²⁰ These results suggest that plyometric ability and whole body stiffness may be important risk factors for ACL injury in previously uninjured knees in male athletes but also for reconstructed knees. Given that ACL rupture normally occurs in the first 40 ms after ground contact,²⁵ greater muscular co-contraction and early rate of force development associated with increased plyometric ability^7,29 may be important in controlling anterior tibial translation and ACL loading after ACLR. In addition, ACL injury prevention programs that have been demonstrated to be effective in reducing ACL injury rates have all included various plyometric exercises (drop jumps, tuck jumps, bounding, etc), and it may be that this component of these programs is highly important in contributing to the reduced injury rates.^14,31,36

Much of the focus during rehabilitation is to optimize recovery of quadriceps strength on the ACLR side.³⁸ In this study, those who experienced CI had lower quadriceps strength of the contralateral limb than those who did not. Previous research has reported decrements in quadriceps strength on the contralateral side after reconstruction, and those decrements may influence risk of second ACL injury.⁵⁶ Quadriceps strength accounts for ~30% of SLCMJ and SLDJ height performance (Crotty et al, unpublished data. “The Relationship Between Isokinetic Knee Strength and Single Leg Jump Performance 9 Month Following ACL Reconstruction.” Presented at Faculty of Sports and Exercise Medicine Annual Conference, 2019),¹⁰ and its redevelopment after ACLR may be an important factor in developing plyometric capacity and minimizing ACL injury risk in healthy limbs. In this study, we found no differences in CoD biomechanics between the CI and NCI groups. If plyometric ability or whole body stiffness is an important measure in contralateral ACL injury risk for male athletes, it is intuitive that this would be more evident in drop jump tests rather than CoD tests, despite the fact that CoD is a common mechanism of ACL injury.¹

Fewer differences between groups were observed in the frontal and transverse plane as compared with the sagittal plane of the DLDJ and SLDJ on the contralateral side. There was greater internal knee valgus moment in both tests (earlier stance in SLDJ, later stance in DLDJ) but lower through midstance in the DLDJ in the CI group. The joint moment signals demonstrated a similar pattern: higher moments earlier and later but lower moments in midstance in the CI group. These findings are different from previous studies in female athletes in which external knee valgus was identified as a risk factor for primary injury.¹⁶ There were lower maximum internal valgus moments in the CI group, which may reflect a reduced ability to resist external valgus moments upon more chaotic dynamic challenges on return to sports. Paterno et al⁴⁰ reported knee valgus range of motion and hip rotation impulse as predictors of second ACL injury. This is not replicated in our study potentially because of our focus on male athletes and contralateral second injuries. In the SLDJ, there was increased ipsilateral trunk sway over the contralateral limb in the CI group, which is a common ACL injury mechanism,¹ influences knee frontal plane loading,^8,9 and, in combination with knee valgus movement, is a risk factor for noncontact knee injuries.¹⁵ That a greater number of variables indicated differences in the sagittal plane than in the frontal plane in this cohort as compared with previous research may be due to the sex of our participants (male). Female athletes are more likely to demonstrate dynamic knee valgus during landing^37,47 and during ACL injury mechanism.²⁷ Cumulatively, our findings add new evidence suggesting that physical risk factors for ACL injury may be different between sexes and may require differential approaches to assessment and analysis to achieve sex specificity for ACL injury risk.

The biomechanical variables identified had fair to good ability to predict CI group membership for DLDJ and SLDJ; therefore, targeting these variables during rehabilitation and return-to-play testing may reduce risk of ACL injury. Higher levels of sensitivity versus specificity are important for ACLR given the severe consequences of second injury. Lower specificity also reflects previous research demonstrating that as many as 20% of healthy athletes are classified as having the same movement strategies as those who have undergone ACLR,⁴⁵ suggesting that movement alone does not account for all risk related to ACLR injury.

Limitations

As no previous study has examined biomechanical risk factors for contralateral ACL injury, this study examined variables throughout the kinetic chain in several jump and CoD tests. Although this may increase risk of “overanalysis” or finding differences that are not relevant to the outcome, the intent of including only medium and large effect size differences was to identify just those differences of largest magnitude to highlight variables of greatest clinical and research interest despite multiple analyses. We performed multiple comparisons, and one could argue that a multiple-comparisons correction should have been implemented to reduce the type 1 error. However, as the type 1 error decreases, the chance of type 2 errors increases.^18,42,46,53 Our approach to modeling and resultant conclusions were based on P values in combination with effect sizes, and differences with weak effects were excluded to decrease the type 1 error. Although a strength of the study is that it was carried out on a homogeneous cohort (male field sports athletes), findings may not be directly extrapolated to other populations. Therefore, future research with similar analyses in female athletic populations is needed to identify risk factors specific to that cohort as well as potential differences in risk factors for male and female athletes for additional ACL injury after ACLR. In addition, future research verifying the ability of the findings to predict the risk of contralateral ACL injury in a different group of athletes would be valuable to reinforce the generalizability of the findings. Although the 2-year cutoff for second injury was selected as a threshold for the control group (NCI) the average time for CI in the case group was 23.3 ± 9.8 months, meaning that many of the injuries happened after the selected threshold and raised the potential for injury in the NCI group after selection. However, upon further follow-up of the NCI group, none had an injury at a minimum of 3.5 years after surgery. To improve on the model, other biomechanical measures, such as variability, coordination, and resistance to fatigue, could be included to assess if they are factors that may lead to CI. These can be used in combination with anthropometric, surgical, and radiological data that can influence ACL injury to build a comprehensive model of factors influencing risk of second ACL injury. Finally, intervention studies are needed to examine the most effective way to change variables identified during rehabilitation and the influence of this on subsequent contralateral ACL injury.

CONCLUSION

This study highlights that biomechanical analysis of the contralateral limb at 9 months after ACLR could identify movement differences between those who experience a subsequent contralateral ACL rupture and those who do not. These variables had fair to good ability to predict CI and would not have been identified by only evaluating clinical performance measures. Our study demonstrated lower quadriceps strength, sagittal plane control, and plyometric ability on the contralateral limb in those who experienced subsequent contralateral ACL injury. There was no difference in LSI in performance measures and minimal differences in asymmetry of biomechanical variables. Therefore, this study highlights several factors that may be used in future analysis to model prediction of second ACL injury and that can be targeted during rehabilitation to reduce contralateral ACL injury after ACLR.