Timing of Strain Response of the ACL and MCL Relative to Impulse Delivery During Simulated Landings Leading up to ACL Failure

Abstract

Anterior cruciate ligament (ACL) injury videos estimate that rupture occurs within 50 milliseconds of initial contact, but are limited by imprecise timing and nondirect data acquisition. The objective of this study was to precisely quantify the timing associated with ligament strain during simulated landing and injury events. The hypotheses tested were that the timing of peak strain following initial contact would differ between ligaments and that peak strain timing would be independent of the injury-risk profile emulated during simulated landing. A mechanical impact simulator was used to perform landing simulations based on various injury-risk profiles that were applied to each specimen in a block-randomized order. The ACL and medial collateral ligament were instrumented with strain gauges that recorded continuously. The data from 35 lower-extremity specimens were included for analysis. Analysis of variance and Kruskal–Wallis tests were used to determine the differences between timing and profiles. The mean time to peak strain was 53 (24) milliseconds for the ACL and 58 (35) milliseconds for the medial collateral ligament. The time to peak ACL strain ranged from 48 to 61 milliseconds, but the timing differences were not significant between profiles. Strain timing was independent of injury-risk profile. Noncontact ACL injuries are expected to occur between 0 and 61 milliseconds after initial contact. Both ligaments reached peak strain within the same time frame.

Clinically, anterior cruciate ligament (ACL) injuries are hypothesized to occur within 50 milliseconds of initial contact with the ground during landing tasks.¹ This rapid loading of the ACL is corroborated by in vitro simulation, which has shown that ACL strain immediately increases and peaks within 50 milliseconds when ground impulses are applied.² When positioned into 15° of valgus prior to impact delivery, the timing of peak ACL strain relative to initial contact was slightly delayed relative to the nonvalgus simulations, but still presented within 70 milliseconds and immediately followed the presentation of the peak valgus angle.³ A second in vitro impact model, which incorporated knee abduction moment, internal tibial rotation, and anterior tibial shear constraints across a multitude of simulations, corroborated that the mean time to peak ACL strain was 54.2 milliseconds after initial contact, which immediately followed the time to peak knee abduction (51.5 ms).⁴

The association established between knee abduction moment and ACL injury incidence has made it the primary component of ACL injury-risk prediction.⁵ Clinical surrogates to complex biomechanical quantification of knee abduction moment have been developed to parse out athletes likely to present high-risk biomechanics during athletic tasks.^6,7 Injury-risk classification is important because targeted neuromuscular training can contribute to a significant reduction of ACL injuries across a population.^8,9 This effect is greatest on the biomechanics of athletes who exhibit the highest risk for injury.¹⁰ In addition, the literature has demonstrated that ACL injury risk influences peak ACL strain and change in ACL strain during landing, as high-risk profiles produce larger strains during landing simulations.^11–14 Despite these known behaviors, investigations have yet to quantify how high-risk biomechanics influence the timing of ACL strain during impact relative to low-risk biomechanics.

While the ACL and medial collateral ligament (MCL) share the knee abduction moment loading mechanism^15,16 and are concomitantly injured in one-third of ACL cases,¹⁷ they exhibit separate responses to the increased risk profiles during simulated landing.^11,12 The ACL contributes greater to overall knee restraint, expresses greater strain throughout landing,^12,14,16,18 and is more responsive to high-risk loading profiles than the MCL.¹¹ As with the ACL, the timing of MCL strain relative to varied injury-risk stimuli has not been evaluated. With the mechanical impact simulator, it is now possible to quantify the real-time magnitude of ACL and MCL strain during physiologically relevant landing simulations performed across a multitude of injury-risk loading profiles.¹⁹

The study objective was to quantify the timing associated with ACL and MCL strain during simulated landing tasks leading up to ligament failure. The hypothesis tested was that the timing of the peak strain following initial contact would differ between the ACL and MCL. The second hypothesis tested was that, although higher risk simulation profiles increase ACL strain, these profiles would not influence the timing of ACL strain following impact.

Methods

A total of 46 full lower-extremity cadaveric specimens were procured from an anatomical donations program (Anatomy Gifts Registry, Hanover, MD) for mechanical impact simulation. Of these specimens, 6 were excluded due to structural weakness during setup (n = 2), substandard bone stiffness (n = 1), and equipment failure/inconsistency in testing execution (n = 3). Furthermore, only the specimens who experienced ACL rupture during testing were analyzed. This left 35 specimens (18 males and 17 females; age = 41.7 [8.1] y, mass = 83.3 [23.7] kg, height = 173.3 [10.2] cm) for analysis.

The full details of the specimen setup have previously been published (Figure 1).¹⁹ Briefly, the specimens were resected of soft tissue superior to the knee joint, leaving the hamstrings and quadriceps tendons intact. The femur was then transversely cut 20 cm above the superior aspect of the patella. The transected femoral bone was potted in Bondo® (3M, St Paul, MN), then the specimen was inverted and the potted bone was secured atop a 6-axis load cell (Omega160 IP65/IP68; ATI Industrial Automation Inc, Apex, NC) within the mechanical impact simulator. This process aligned the long axis of the femur with the vertical axis (compression/distraction) of the load cell. Cable clamps were secured around the hamstrings and quadriceps tendons, which were affixed to mounted pneumatic actuators (SMC Corporation, Tokyo, Japan) via carbon fiber rope (Ø⁷/₆₄ in; AmSteel®-Blue; Samson, Ferndale, WA). Anterior portals were cut in the skin on the medial and lateral aspects of the patella to provide access to intra-articular knee structures. Customized differential variance resistance transducer (LORD MicroStrain Inc, Williston, VT) strain gauges were implanted in the distal third of the anteromedial bundle of the ACL and in the midsubstance of the MCL across the tibiofemoral joint line.

Figure 1 —

(A) Meta-view of custom-designed mechanical impact simulator for replication of ACL ruptures.¹⁹ (B) Cable pulley system used to deliver pneumatically actuated loads to the quadriceps and hamstrings tendons. (C) External fixation frame attached to the tibia to deliver pneumatically actuated KAM, ATS, and ITR loads to each specimen. ACL indicates anterior cruciate ligament; ATS, anterior tibial shear; ITR, internal tibial rotation; KAM, knee abduction moment. Reproduced with permission from Bates et al.²²

A custom-designed compression clamp was affixed through the tibia with carriage bolts. This clamp contained rotary and linear attachment points that were affixed, via carbon rope, to pneumatic pistons mounted on the frame of the simulator. These attachments allowed for the application of in vivo calculated knee abduction moment, anterior tibial shear, and internal tibial rotation loads to the knee of the specimen. The tibia was positioned vertically at a 25° flexion angle relative to the femur by resting an 18.1-kg ground plate on the foot of the specimen. The ground plate contained a uniaxial load cell (1720ACK-10kN; Interface Inc, Scottsdale, AZ) aligned with the heel of the foot on its inferior side and the point of impact delivery on its superior side.

Each specimen in the mechanical impact simulator was subjected to simulations designed to emulate landing from a drop off a 31-cm box.¹⁹ To accomplish this, a 34-kg weight sled was suspended 31 cm above the ground plate with electromagnets. An electronic trigger released the sled, which was directed to impact each specimen’s heel via vertically mounted slide rails. Prior to sled release, the mounted pneumatic actuators were activated. The quadriceps and hamstrings actuators applied a 1:1 ratio to create an ideal, unbiased musculature loading of the knee joint.^20,21 The knee abduction moment, anterior tibial shear, and internal tibial rotation moment pneumatics were activated to magnitudes representative of kinetics recorded from 44 healthy athletes who performed drop vertical jump tasks in a 3-dimensional motion analysis laboratory.^19,22 For each trial, the loads were individually randomized to a value that corresponded to a percentile of the in vivo cohort (Table 1). These values ranged between baseline and very-high-relative injury-risk profiles for that cohort.²³ This randomization produced a 28-test subthreshold protocol. If any hard or soft tissue failures were confirmed during simulation, the testing was halted.^19,22 For specimens who completed all 26 subthreshold trials,¹² the magnitude of the knee abduction moment, anterior tibial shear, and internal tibial rotation moment actuators were set to the very-high-risk profile, and the weight sled mass was adjusted to 0.5 × bodyweight of the specimen. For each consecutive impact, knee abduction moment, anterior tibial shear, and internal tibial rotation moment, the actuators were increased in 20% increments until damage was induced.

Table 1.

External Load Magnitudes Applied to the Knee Joint During Landing Simulation and Corresponding Percentile of the Previously Tested In Vivo Population Cohort

Knee abduction moment		Anterior tibial shear		Internal tibial rotation
Percentile of in vivo population cohort, %	Load, N·m	Percentile of in vivo population cohort, %	Load, N·m	Percentile of in vivo population cohort, %	Load, N·m
2	2.4	0	40	0	1
68	27	90	98	33	9.7
99	53.6			67	18.6
200	114.6			100	53.7

Throughout each simulation, differential variance resistance transducers collected continuous ACL and MCL strain data, the 6-axis load cell collected continuous force and torque data that were extrapolated to the knee joint center point (as determined by the midpoint of the femoral notch) with a transformation matrix, and the uniaxial load cell collected continuous vertical ground impulse force. All data were triggered, synced, collected at 10,000 Hz, and filtered through a fourth-order, low-pass, 50-Hz Butterworth filter using custom code in MATLAB (version 2018b; The MathWorks, Inc, Natick, MA) and LabVIEW (version 2016; National Instruments, Austin, TX). To analyze the strain response timing, 5 critical points were designated at 33 milliseconds prior to initial contact, at initial contact, 33 milliseconds after initial contact, 67 milliseconds after initial contact, and 100 milliseconds after initial contact. These points were selected partially because typical video capture of athletic events occurs at a frame rate of 30 Hz (33.3 ms between frames),¹ because in vivo ACL injuries are reported to occur within 70 milliseconds of landing,^1,24 and because vertical ground reaction forces stabilize around 100 milliseconds after landing.²⁵

For statistical analysis, the specimen cohort was organized into 2 subgroups, and the statistical analyses were performed individually on each group. The primary group consisted of the 35 specimens who experienced ACL failure during the simulation protocol. The next group consisted of 17 specimens (15 males and 4 females; age = 39.1 [8.9] y, mass = 93.6 [22.5] kg, height = 175.9 [8.6] cm) who completed the 26 subthreshold simulations protocol prior to injury, but still experienced an ACL rupture later in testing. This creation of 2 groups was necessary because when comparisons are made between injury-risk levels, inclusion of a specimen who completed one risk level without completing the second risk level has the potential to confound the data. Therefore, a group in which all specimens completed all simulation trials was desirable to understand the potential confounding effects.

The timing of peak ACL and MCL strain was indexed in the continual strain data for each specimen. Kruskal–Wallis tests were utilized to assess the data with factors of the specimen group (ACL injured vs full protocol), ligament (ACL and MCL), and risk profile (baseline risk, moderate risk, high risk, and very high risk) to assess significance. The risk profiles were defined by matching values of applied knee abduction and internal tibial rotation moments, as previously described in the literature.¹¹ Wilcoxon each pair post hoc tests with Bonferroni correction were then used to determine significance between individual groups. Within the ACL-injured and full-protocol groups, a Kruskal–Wallis test was used to test the significance of the time point (−33 ms, initial contact, 33 ms, 67 ms, and 100 ms) during the simulation trial that immediately preceded the ACL rupture. The actual failure trial cannot be assessed, as differential variance resistance transducer behavior at the instance of injury is not predictable.^11–13,22 Wilcoxon each pair post hoc tests with Bonferroni correction were again used to determine the significance between individual points. Risk profile significance relative to ligament strain during mechanical impact simulations has been presented previously in the literature.^11,12 For all tests, the statistical significance was set at α < .05. Statistical analysis was performed in JMP Pro (version 13; SAS Institute Inc, Cary, NC).

Results

For all tested conditions, the mean peak ACL strain for the ACL-injured cohort in the simulation prior to failure occurred 53 (24) milliseconds after initial contact, and the mean peak MCL strain occurred 58 (35) milliseconds after initial contact. The timing of the mean peak ACL and MCL strain post initial contact for both all ACL-injured specimens and the subgroup that completed very-high-risk simulations is displayed in Table 2 (Figure 2). Of the 3 factors examined (specimen group, ligament, or risk profile), only ligament proved significant to the timing of peak strain following initial contact (P = .002). Post hoc tests revealed that between-ligament timing differences were only significant within the ACL rupture specimen group (P = .003) and, furthermore, were only significant within the prefailure risk profile of that group (P = .04). No ligament significance was identified in any other individual combination of specimen group and risk profile (P ≥ .13).

Table 2.

Timing (in ms) of Peak ACL and MCL Strain Relative to Initial Contact

	ACL injured group			Full protocol group
	n	ACL	MCL	n	ACL	MCL
Baseline	31	48 (19)	54 (33)	17	49 (21)	53 (30)
Moderate	30	52 (21)	56 (35)	17	58 (26)	51 (29)
High	18	61 (33)	59 (36)	17	59 (32)	57 (35)
Very high	18	52 (23)	69 (41)	17	52 (23)	67 (41)
Prefailure	35	50 (18)	64 (37)	17	55 (24)	63 (42)

Abbreviations: ACL, anterior cruciate ligament; MCL, medial collateral ligament.

Figure 2 —

Time-dependent plots of mean ACL strain (red, solid line) and MCL strain (blue, dashed line) with SDs throughout the (A) baseline-risk, (B) moderate-risk, (C) high-risk, (D) very-high-risk, (E) prefailure, and (F) failure simulation profiles. The vertical dotted line indicates the timing of initial contact. This figure is representative of the 17 specimens who completed the full simulation protocol prior to ACL rupture. As a reminder, once ACL failure occurs, DVRT behavior becomes arbitrary and unpredictable. Therefore, plot F was not statistically analyzed and is solely for visual reference. ACL indicates anterior cruciate ligament; DVRT, differential variance resistance transducer; MCL, medial collateral ligament.

For ACL rupture specimens, during the trial prior to failure, the time point was a significant factor to ACL strain (P = .002; Table 3) and change in ACL strain from the baseline (P < .001). ACL strain was greater at the 33-, 67-, and 100-millisecond time points than at −33 milliseconds (mean difference = 4.2%, 4.1%, and 2.7%; P = .002, .01, and .02, respectively) and initial contact (mean difference = 4.2%, 4.1%, and 2.7%; P = .001, .001, and .001, respectively). Change in ACL strain from the baseline was greater at the 33-, 67-, and 100-millisecond time points than at −33 milliseconds (mean difference = 4.2%, 4.1%, and 2.7%; P = .003, .01, and .02, respectively) and initial contact (mean difference = 4.2%, 4.1%, and 2.7%; P = .001, .001, and .001, respectively). No ACL injures were observed prior to impulse delivery.

Table 3.

Absolute Strain and ΔStrain for the ACL and MCL at Designated Time Points Relative to Initial Contact for Each of the 3 Subgroups Analyzed

Risk profile	ACL
	Peak ACL strain					ΔACL strain from baseline
	−33 ms, %	Initial contact, %	33 ms, %	67 ms, %	100 ms, %	−33 ms, %	Initial contact, %	33 ms, %	67 ms, %	100 ms, %
Full protocol group
Baseline (n = 17)	3.8 (5.2)	3.8 (5.2)	5.7 (6.9)	5.5 (6.3)	3.2 (3.8)	0.0 (0.1)†,*	0.0 (0.1)†,*,#	1.9 (2.7)#	1.7 (2.3)#	−0.5 (2.9)†,*
Moderate (n = 17)	4.2 (5.2)	4.2 (5.2)	5.8 (6.8)	5.9 (6.4)	4.9 (5.4)	0.3 (0.5)†,*	0.3 (0.5)†,*	1.9 (2.9)	1.9 (2.4)	1.0 (1.6)
High (n = 17)	5.2 (5.8)	5.2 (5.8)	6.9 (8.1)	6.6 (8.0)	6.2 (6.8)	1.0 (1.3)	1.0 (1.3)	2.8 (4.1)	2.4 (3.9)	2.0 (2.8)
Very high (n = 17)	6.2 (5.4)	6.2 (5.4)	8.0 (7.4)	7.6 (7.5)	7.2 (6.2)	2.7 (2.2)	2.7 (2.2)	4.5 (4.5)	4.1 (4.3)	3.7 (3.1)
Prefailure (n = 17)	6.9 (5.8)	7.0 (5.8)	9.8 (7.3)	9.9 (7.5)	8.9 (6.3)	2.6 (1.7)†,*,#	2.7 (1.8)†,*,#	5.5 (2.9)	5.6 (3.0)	4.5 (2.3)
ACL rupture group
Baseline (n = 31)	3.8 (4.4)†,*	3.8 (4.4)†,*	6.6 (6.9)	7.2 (6.8)	3.8 (3.7)†,*	0.0 (0.1)†,*	0.0 (0.1)†,*	2.7 (5.4)#	3.3 (5.8)#	0.0 (2.2)†,*
Moderate (n = 30)	4.7 (4.8)†,*	4.7 (4.8)†,*	7.7 (6.9)	7.8 (6.8)	5.9 (5.4)	0.6 (1.5)†,*,#	0.6 (1.6)†,*,#	3.6 (5.4)	3.7 (5.2)	1.9 (2.9)
High (n = 18)	4.9 (5.8)	4.9 (5.8)	6.6 (8.1)	6.6 (7.8)	6.1 (6.6)	0.9 (1.3)	0.9 (1.3)	2.6 (4.1)	2.6 (3.9)	2.1 (2.8)
Very high (n = 18)	6.9 (6.1)	6.9 (6.1)	8.8 (7.8)	8.5 (8.2)	8.0 (7.0)	2.6 (2.2)	2.6 (2.2)	4.5 (4.4)	4.2 (4.2)	3.7 (3.0)
Prefailure(n = 35)	7.4 (5.7)†,*,#	7.4 (5.7)†,*,#	11.6 (7.6)	11.5 (7.7)	10.1 (6.2)	2.5 (2.4)†,*,#	2.5 (2.5)†,*,#	6.7 (6.2)	6.6 (6.0)	5.2 (3.8)
Risk profile	MCL
	Peak MCL strain					Δ MCL strain from baseline
	−33 ms, %	Initial contact, %	33 ms, %	67 ms, %	100 ms, %	−33 ms, %	Initial contact, %	33 ms, %	67 ms, %	100 ms, %
Full protocol group
Baseline (n = 17)	0.4 (0.5)*	0.5 (0.6)*	0.7 (0.9)	1.0 (0.9)	0.6 (1.0)	0.0 (0.0)*	0.0 (0.0)*	0.2 (0.7)	0.6 (0.6)#	0.2 (0.6)*
Moderate (n = 17)	0.5 (0.7)	0.5 (0.8)	0.7 (0.9)	1.0 (1.1)	0.9 (1.2)	0.0 (0.2)*	0.1 (0.2)*	0.2 (0.6)	0.6 (0.7)	0.5 (0.9)
High (n = 17)	2.2 (5.9)	2.3 (5.9)	2.4 (6.0)	2.7 (5.9)	2.6 (5.9)	0.4 (0.6)	0.4 (0.6)	0.5 (1.2)	0.8 (1.1)	0.8 (1.2)
Very high (n = 17)	2.5 (2.1)	2.5 (2.1)	2.0 (2.3)	2.5 (2.5)	2.7 (2.4)	1.9 (1.7)	1.9 (1.7)	1.5 (2.0)	1.8 (2.2)	2.1 (2.0)
Prefailure (n = 17)	3.5 (3.1)	3.5 (3.1)	3.2 (3.5)	3.5 (4.0)	3.9 (3.5)	2.8 (2.4)	2.8 (2.4)	2.5 (2.8)	2.8 (3.3)	3.2 (2.9)
ACL rupture group
Baseline (n = 31)	0.8 (1.5)	0.8 (1.5)	1.2 (2.2)	1.4 (1.9)	1.1 (2.1)	0.0 (0.1)*	0.0 (0.1)*	0.4 (1.1)	0.6 (0.9)	0.3 (0.9)
Moderate (n = 20)	0.7 (1.4)	0.7 (1.4)	0.9 (1.5)	1.2 (1.5)	1.1 (1.8)	0.0 (0.1)*	0.0 (0.1)*	0.3 (0.6)	0.5 (0.8)	0.4 (0.8)
High (n = 18)	2.7 (6.1)	2.7 (6.1)	2.5 (5.8)	3.2 (6.1)	3.2 (6.2)	0.6 (1.2)	0.6 (1.2)	0.4 (1.3)	1.1 (1.7)	1.1 (1.8)
Very high (n = 18)	2.5 (2.0)	2.5 (2.0)	2.0 (2.3)	2.4 (2.4)	2.7 (2.4)	1.9 (1.6)	1.9 (1.6)	1.4 (1.9)	1.8 (2.1)	2.1 (1.9)
Prefailure (n = 35)	3.0 (3.2)	3.0 (3.2)	2.9 (3.6)	3.3 (3.5)	3.3 (3.5)	1.8 (2.2)	1.8 (2.2)	1.7 (2.3)	2.0 (2.6)	2.1 (2.5)

Abbreviations: ACL, anterior cruciate ligament; MCL, medial collateral ligament.

^†Significant difference from 33 ms.

^*Significant difference from 67 ms.

^#Significant difference from 100 ms.

For the specimens who completed the full protocol of simulations, during the trial prior to failure, the time point lacked significance relative to absolute ACL strain (P ≥ .17), but was a significant factor for change in ACL strain from the baseline (P < .001). The change in ACL strain was greater at 33, 67, and 100 milliseconds than at −33 milliseconds (mean difference = 2.9%, 3.0%, and 2.0%; P = .002, .002, and .01, respectively) or initial contact (mean difference = 2.8%, 2.9%, and 1.9%; P = .003, .003, and .02, respectively). No ACL injures were observed prior to impulse delivery.

For the ACL rupture specimens, during the trial prior to failure, the time point was not a significant factor to either absolute MCL strain (P = .98; Table 3) or the change in MCL strain relative to the baseline (P = .98).

For the specimens who completed the full protocol of simulations, during the trial prior to failure, the time point was not a significant factor to absolute MCL strain (P = .28) or the change in MCL strain from the baseline (P = .93).

Discussion

The study objective was to quantify the timing associated with ACL and MCL strain during simulated landing tasks leading up to ligament failure. The hypotheses tested were partially supported. There was a significant difference in the timing of peak strain between the ACL and MCL, as peak MCL strain was achieved 5 milliseconds later than peak ACL strain. This supported the hypothesis that the timing of peak strain post initial contact would differ between ligaments. However, post hoc tests revealed that this difference was only observed within the prefailure risk profile of the ACL injury group and not within any other individual risk profile and specimen group combinations. Therefore, under most simulation conditions, timing differences between the ligaments were not observed. Furthermore, a 5-millisecond difference in peak loading rate represents ≤10% of the time to peak strain documented in this investigation and is unlikely to exhibit a tangible mechanical influence on a viscoelastic biologic structure. While knee loads are known to be disproportionately distributed to the ACL compared with the MCL,^12,14,18,22 the present data show that, during most landings, those loads are generated across similar time periods. The MCL showed greater variance of peak timing than the ACL (Table 2), which is likely related to the lower proportion of load bearing in the MCL. During simulated landings, the MCL exhibits less strain and load than the ACL.^12,14,15,18 During controlled athletic tasks, it is even possible that the MCL bears 0% of the total knee load.¹⁸ In these circumstances of minimal loading, peak MCL strain could register at any time point, which would subsequently increase the variability of timing events for the ligament. Note that, in Table 3, peak ACL strains were reliably documented at either the 33- or 67-millisecond time point, whereas, in Table 3, the peak MCL strain was occasionally noted at −33 milliseconds, initial contact, and 100 milliseconds. However, Table 1 clearly documented that the mean peak MCL strain predominately occurs in parallel with, or slightly after, the timing of peak ACL strain.

The present data supported the hypothesis that, although higher-risk simulation profiles increase ACL strain, these profiles would not influence the timing of ACL strain following impact. Instead, the timing of peak strain was consistent, regardless of the ligament or risk profile. Indeed, increased risk applied to the specimens in this study led to increased ACL and MCL strain throughout landing (Tables 2 and 3). This finding corroborates the existent literature, as increased risk level during simulated landing is known to elevate peak ligament strain during landing.¹¹ Therefore, as higher risk biomechanical profiles exhibit greater peak strain and comparable timing during simulated landings, they consequently must produce greater rates of force development in the ACL and MCL. Increased rate of loading has been proposed previously as a significant factor to ligament rupture within the literature,^5,26,27 and the present data support that conjecture. In this investigation, the highest ligament strains occurred in the trial prior to failure, which means that the greatest rate of force development for these specimens coincided with ligamentous failure.

During the nonfailure trials, the peak ACL and MCL strains were observed between 48 and 63 milliseconds after impulse delivery. This time frame coincides with video analysis of ACL rupture events, as it has been deduced from imagery that ACL ruptures occur within approximately 67 milliseconds of initial contact.^1,24 However, in vivo video analyses are limited by the frame rate, as most cameras record between 30 and 60 Hz, which means that each frame progresses between 17 and 33 milliseconds and precise determination of injury timing may be moderately obscured. With the mechanical impact simulator that replicates physiologically relevant ACL injuries,²² sampling is conducted at 10,000 Hz and interpolated down to 300 Hz for data analysis. The greater precision of this faster sampling rate provided a more accurate analysis of strain response timing than has been reported from in vivo injury events. Coincidentally, as peak ACL strain was clustered around 48 to 61 milliseconds, the mechanical impact simulator would also indicate that ACL ruptures occur between initial contact (0 ms) and this time frame.

Ligament strain did not change between the time points −33 milliseconds and 0 milliseconds. This behavior indicated that mechanical load within the ACL and MCL had normalized to the external load and muscle force applications that were applied prior to initial contact. Within the window of observation for this study, the strain changes observed were induced post initial contact and ranged from 2% to 5% in the ACL. If the ACL had been strained to near-failure loads during flight from the application of external forces and simulated muscle contractions, then the post initial-contact strain increase would not have been observed, as it would have resulted in ligament rupture. Therefore, peak ACL strain and ACL rupture in flight prior to initial contact, as suggested by some literature sources,^28,29 would be highly unlikely according to the present model findings. Indeed, all of these clinically representative injuries induced on the mechanical impact simulator occurred after impulse delivery.²² Establishment of a minimal clinically important difference for changes in ACL strain would be an important next step for future reference to aid in the translation of the numerical values presented here to clinical paradigms of relative injury risk, pain, and structural damage.

A limitation of the current investigation lies in the inability to analyze data from rupture trials. Dependent on the mechanism and location of ligament failure, the strain gauge behavior would exhibit one of several patterns during an injury event. Unfortunately, a number of these behaviors made it impossible to discern from the data the exact instant when rupture occurred. Therefore, to err on the side of conservative analysis, the trial immediately prior to the rupture was analyzed instead. This prefailure trial demonstrates a high consilience to the failure trial for MCL strain and demonstrates strain rates consistent with ligament failure.³⁰ A second limitation lies in the use of cadaveric tissue, with no possibility for biologic recovery or healing. For this reason, impulse applications were applied in a block-randomized format to minimize the confounding effects of fatigue that may have been present in a progressively increasing study design. However, the number of impacts was similar to that which an athlete would incur during a typical sport event in which minimal recovery or healing of the tissue would occur. Furthermore, testing was aborted when intraarticular hard or soft tissue damage was documented.^19,22 Finally, during an in vivo landing, muscle contractions would be dynamically changing; in the present simulation, the muscle forces were applied at a constant rate, with a 1:1 ratio of hamstrings: quadriceps force. The EMG data were not collected on the in vivo cohort used to derive input kinetics for the mechanical impact simulator. As such, the musculature timing from these subjects is unknown, and a constant muscle force application was superior as opposed to an unfounded application of variable timing that would have confounded the present results. Future work would be desirable to develop an impact model with dynamic muscle force application in a variety of force ratios.

Peak ACL strain during a simulated landing task was achieved prior to 61 milliseconds after initial contact. Therefore, noncontact ACL injuries are expected to occur within a window between 0 and 61 milliseconds after initial contact. Timing behavior is irrespective of the magnitude of knee abduction moment or the relative injury risk applied to the knee, which means that athletes who achieve higher ligament strains due to high-risk biomechanics also have a greater ligamentous rate of load development than their low-risk peers.