Elevating the Noninvolved Limb Reduces Knee Extensor Asymmetry During Squat Exercise in Persons With Reconstructed Anterior Cruciate Ligament

Abstract

Jean, LMY and Chiu, LZF. Elevating the noninvolved limb reduces knee extensor asymmetry during squat exercise in persons with reconstructed anterior cruciate ligament. J Strength Cond Res 34(8): 2120–2127, 2020—Anterior cruciate ligament (ACL)–injured individuals use less knee extensor net joint moment (NJM) in the involved vs. noninvolved limb during squat exercises. The objective of this research was to examine if knee extensor NJM symmetry between the involved and noninvolved limbs could be attained with a modified squat. Six individuals with unilateral ACL reconstructed knees performed bilateral squats under normal conditions and with their noninvolved limb elevated on a 5-cm platform. Knee extensor NJM was determined using 3-dimensional motion analysis. Knee extensor NJM was lower in the involved compared with the noninvolved limb (95% confidence interval [CI], 0.08–0.28 N·m·kg⁻¹; d = 1.66 SD) during normal squats. Knee extensor NJM was greater in the involved (95% CI, 0.02–0.18 N·m·kg⁻¹; d = 0.57 SD) and lower in the noninvolved (95% CI, −0.25 to −0.07 N·m·kg⁻¹; d = 1.85 SD) limbs in the elevated vs. normal squats. Knee extensor NJM was not different between limbs when the noninvolved limb was elevated (95% CI, −0.26 to 0.11 N·m·kg⁻¹; d = 0.48 SD). Persons with ACL reconstruction exhibit knee extensor asymmetry during bilateral squats. Elevating the noninvolved limb reduces knee extensor NJM asymmetry between the involved and noninvolved limbs during squat exercise.

Introduction

Quadriceps weakness and atrophy are common deficits in persons with anterior cruciate ligament (ACL) injury (^10,26,27). These deficits are observed following ACL rupture and remain or increase after reconstructive surgery (^13,23). The quadriceps contribute to energy generation and absorption in locomotion—such as walking and running—and sporting activities—such as jumping and landing (^8,14,16). Anterior cruciate ligament–injured individuals exert a lower knee extensor net joint moment (NJM) in the involved vs. noninvolved limb during these tasks, which may be due to either quadriceps inhibition or atrophy of the vasti (^26,27). Quadriceps dysfunction during these tasks is hypothesized to increase the risk for reinjury or cause other complications in the knee (²²). Consequently, restoring quadriceps size and strength is recommended to improve function and return to activity at the same level as before injury (²⁰). However, many rehabilitation programs are unable to sufficiently restore quadriceps size and strength in the involved limb (¹⁹).

In functional and sporting tasks, such as squatting, jumping, and landing, ACL-injured individuals exert lower knee extensor NJM in the involved vs. noninvolved limb (^17,18,21). This may be explained by decreased quadriceps strength, where the involved limb quadriceps are not capable of exerting a greater knee extensor NJM. However, it has been hypothesized that the lower knee extensor NJM is because of neurological quadriceps inhibition (^1,17–19). The latter scenario in combination with weaker vasti poses a challenge for restoring quadriceps strength. For example, single-joint leg extension exercise seems to preferentially train the rectus femoris and may not be effective to restore vasti size and strength (^11,15). In healthy individuals, exercises such as squats elicit improvements in vasti size and strength (^2,24). However, reduced quadriceps contribution as a result of inhibition may limit the efficacy of squat exercise to elicit vasti size and strength adaptations in ACL-deficient and reconstructed individuals.

Consequently, it would be beneficial to identify an approach that promotes quadriceps utilization when performing multijoint exercises. Hemiparetic stroke patients use a similar compensatory strategy in the sit to stand as those who have had an ACL injury. Placing the foot of the nonparetic side on an elevated surface has been shown to increase quadriceps electromyography (³). Elevating the nonparetic side increased the vertical ground reaction force acting on the paretic side. For the same limb kinematics, an increase in vertical ground reaction force would be expected to increase lower extremity NJM, including for the knee extensors. Consequently, this modification may be valuable for quadriceps strengthening in ACL rehabilitation programs. Therefore, the purpose of this study was to investigate if elevating the noninvolved limb during bilateral squats would affect knee extensor NJM symmetry between the involved and noninvolved limbs in persons who have had ACL injury and reconstruction. Based on previous research, it was hypothesized that (a) ACL-reconstructed individuals would have a lower knee extensor NJM during bilateral squats in their involved limb and (b) elevating the noninvolved limb during bilateral squats would increase the involved limb knee extensor NJM.

Methods

Experimental Approach to the Problem

A cross-sectional biomechanical analysis was conducted to examine lower-extremity mechanics during normal and experimental variations of squat exercise in subjects who had ACL injury and reconstruction. Subjects performed, in random order, squats with both feet on the ground (normal) and squats where the foot of the noninvolved limb was placed on a 5-cm platform (experimental). Ankle, knee, and hip NJM were calculated from force platform and 3-dimensional (3D) motion analysis data and compared between (a) involved and noninvolved limbs and (b) normal and noninvolved limb elevated squats. To establish whether knee extensor strength was different between the involved and noninvolved limbs, a key predictor of lower-extremity function and return to activity (^19,22,26), knee extensor strength was quantified using isometric dynamometry.

Subjects

A convenience sample of male (n = 1) and female (n = 5) subjects with reconstructed ACLs were recruited by posting flyers at a sports medicine clinic. To be included, potential subjects had to be at least 6 weeks post reconstruction surgery and have received clearance from their physiotherapist to perform resistance exercises. Criteria used by potential subjects' physiotherapists to be cleared included full range of motion compared with the noninvolved limb, absence of swelling, and absence of pain during activities of daily living and during exercise. Four subjects had their left ACL reconstructed and 2 had their right ACL reconstructed. Time after reconstruction surgery ranged from 7 weeks to 6 years. Age, height, and body mass were 27 ± 10 years, 1.63 ± 0.08 m, and 67.0 ± 10.4 kg, respectively. Subject characteristics were measured mean ± SD. This sample size is comparable to that of Salem et al. (²¹), who observed large, significant differences (p = 0.003; d = 0.88) between involved and noninvolved knee extensor NJM during squat exercise in ACL-reconstructed individuals who were 30 ± 12 months postsurgery. A Research Ethics Board at the University of Alberta approved the study protocol, and subjects provided written informed consent (Protocol number: Pro00057964).

Procedures

Squat Biomechanics

Subjects performed the plate squat exercise under 2 conditions: normal and non-involved limb elevated (Figure 1). The plate squat has been used to teach squat exercise and was considered to be appropriate for this population (⁶). Subjects were asked to squat as deep as possible, achieving at least 120° knee flexion because this squat depth has been found to be effective in eliciting vasti hypertrophy (²). For the normal condition, both feet were placed on the ground. In the noninvolved limb elevated condition, the noninvolved limb was placed on a 5-cm platform. Two sets of 3 repetitions were performed in each condition. Male and female subjects performed the exercise using 15- and 10-kg resistances, respectively.

Figure 1.: Normal squat (left) and squat with one limb elevated on a platform (right).

Motion analysis data were collected from retroreflective markers placed on the lower extremity using a previously described 6-degree-of-freedom configuration (⁷). The same investigator placed markers on all subjects. Markers were recorded by 7 optoelectronic cameras sampling at 120 Hz (ProReflex MCU240; Qualysis, Gothenburg, Sweden). Ground reaction force data were recorded from 2 force platforms sampling at 1,200 Hz (OR6-6; AMTI, Watertown, MA). Marker and force platform data were simultaneously collected using Qualysis Track Manager (version 2.4.546). Motion analysis data were processed and analyzed using Visual 3D (version 5.00; C-Motion, Germantown, MD) using standard, 3D, rigid, body-linked segment modeling. For the elevated squat condition, a “force structure” was modeled in Visual 3D, with force platform corners 5 cm higher than the physical force platform in the noninvolved limb. The force structure models the elevated surface as an object in static equilibrium to correct center of pressure location for the top of the elevated surface. Motion and force platform data were digitally low-pass filtered using a fourth-order recursive Butterworth with a 6 Hz cutoff frequency. Joint angles were defined as motion of the proximal segment relative to the distal segment using a XYZ Cardan sequence. Net joint moment was expressed in the coordinate system of the distal segment. Vertical ground reaction force and joint kinetic data were averaged over 15° knee flexion angles starting at 30° (⁹). Data were then averaged across all repetitions.

Knee Extensor Strength

Following squat exercise data collection, isometric knee extensor strength was evaluated to assess if strength asymmetry was present. Subjects performed maximal voluntary isometric actions on a custom-built dynamometer with the machine arm at 6 different angles: 90°, 75°, 60°, 45°, 30°, and 15° (0° is approximately full-knee extension). When compared with the moment exerted by known loads placed on the dynamometer arm, moments measured by the dynamometer were nearly identical (R² > 0.99) and repeated measurements resulted in differences of less than 1%. Subjects performed 2 maximal actions for 3 seconds at each angle. The involved and noninvolved limb were tested individually. Loud verbal encouragement was given. The dynamometer measured the static inertial characteristics of the limb segment, allowing subject-specific gravity corrections to be made for different testing angles. Data were sampled at 500 Hz. Knee angle was measured by hand using a goniometer because the machine arm may not accurately reflect limb positions (¹²).

Peak isometric knee extensor moment was identified at each angle from isometric testing; for each angle, the trial with the highest moment was used. Knee extensor strength curves were generated by plotting moment relative to the measured joint angles. To facilitate comparison—because of some variation in knee flexion angle for each machine arm angle—a second-order polynomial regression was fit to the moment-angle data using Microsoft Excel. Knee extensor strength was then predicted for 15°, 30°, 45°, 60°, 75°, and 90° knee flexion angles.

Statistical Analyses

Isometric knee extensor strength was compared between limbs at each knee flexion angle. The 15° knee flexion interval representing the greatest squat depth achieved in the concentric phase (i.e., after peak knee flexion) was used to compare vertical ground reaction force, and ankle plantar flexor, knee extensor, and hip extensor NJM between involved and noninvolved limbs in both normal and noninvolved limb elevated squats (⁹). These parameters were examined using 2 × 2 (squat variation by limb) analysis of variance (ANOVA) with repeated measures on both factors. Post hoc comparisons were made by calculating 95% confidence intervals (95% CI) of the paired change scores (equivalent to a paired samples t-test with α = 0.05) and Cohen's d effect size (²⁵). Cohen's d effect size was interpreted as medium (0.5–0.9 SD) and large (>0.9 SD) based on knee extensor NJM differences between limbs in ACL-reconstructed individuals performing squats (^1,21). Where the ANOVA was significant, post hoc statistical data were interpreted based on both statistical criteria. First, a difference may be present if the effect size was greater than the minimum effect size worth detecting (d = 0.5 SD). When the effect size was greater than the minimum worth detecting, a difference appeared to be present if the 95% CI also did not cross zero.

Results

Differences in isometric knee extensor strength between limbs were observed at several knee flexion angles (Figure 2). Isometric knee extensor strength appeared to be lower in the involved vs. noninvolved limb at 30° (95% CI, 0.10 to 0.46 N·m·kg⁻¹; d = 0.78 SD), 45° (95% CI, 0.09 to 0.57 N·m·kg⁻¹; d = 0.76 SD), and 60° (95% CI, 0.14 to 0.45 N·m·kg⁻¹; d = 0.74 SD) knee flexion. Isometric knee extensor strength may have been lower in the involved vs. noninvolved limb at 15° (95% CI, −0.03 to 0.32 N·m·kg⁻¹; d = 0.52 SD) and 75° (95% CI, −0.10 to 0.44 N·m·kg⁻¹; d = 0.55 SD) of knee flexion. Isometric knee extensor strength did not appear to be different between the involved and noninvolved limbs at 90° knee flexion (95% CI, −0.78 to 0.69 N·m·kg⁻¹; d = 0.10 SD).

Figure 2.: Knee extensor strength curves for involved and noninvolved limbs. Data are mean ± SD. #Statistically significant difference (p < 0.05).

Ensemble average ankle plantar flexor (Figure 3), knee extensor (Figure 4), and hip extensor (Figure 5) NJM in the involved and noninvolved limbs are shown relative to knee flexion angle in normal (top panels) and experimental (bottom panels) squats. Vertical ground reaction force, and ankle plantar flexor, knee extensor, and hip extensor NJM at the greatest squat depth are shown in Table 1. Significant interactions were observed for knee extensor NJM (p = 0.003) and vertical ground reaction force (p = 0.01). Neither the interaction nor main effects were significant (p > 0.05) for hip extensor and ankle plantar flexor NJM.

Figure 3.: Ankle plantar flexor net joint moment averaged over 15° knee flexion intervals during normal (top) and noninvolved limb elevated squats (bottom). Data are mean ± SD.

Figure 4.: Knee extensor net joint moment averaged over 15° knee flexion intervals during normal (top) and noninvolved limb elevated squats (bottom). Data are mean ± SD.

Figure 5.: Hip extensor net joint moment averaged over 15° knee flexion intervals during normal (top) and noninvolved limb elevated squats (bottom). Data are mean ± SD.

Table 1: Ankle plantar flexor, knee extensor, and hip extensor net joint moments (mean ± SD) at peak knee flexion in normal and experimental squats.

In normal squats, knee extensor NJM appeared to be greater in the noninvolved limb (Table 2). Vertical ground reaction force may have been larger in the noninvolved limb. Vertical ground reaction force appeared to be greater in the involved limb and lower in the noninvolved limb in the experimental squat compared with the normal squat. When squats were performed with the noninvolved limb elevated, knee extensor NJM did not appear to be different between involved and noninvolved limbs. Moreover, knee extensor NJM appeared to be greater in the involved limb during the experimental squat compared with the normal squat. All subjects exhibited a reduction in knee extensor NJM asymmetry between the involved and noninvolved limbs when performing squats with the noninvolved limb elevated (Figure 6).

Table 2: 95% confidence intervals (95% CI) and effect size (d) differences comparing involved and noninvolved limbs during normal and noninvolved limb elevated squats.

Figure 6.: Noninvolved limb minus involved limb knee extensor net joint moment during normal and noninvolved limb elevated squats for individual subjects.

Discussion

Knee extensor asymmetry existed among all subjects in both isometric strength testing and squat exercise despite a large range of years since ACL reconstruction surgery. Knee extensor deficits in ACL-reconstructed individuals have been reported in maximal strength testing and multijoint tasks such as bilateral squats (^10,21). Isometric knee extensor strength differences at 30°, 45°, and 60° knee flexion (between 14 and 16%) were greater than reported for potential copers and comparable to reported for noncopers (¹⁰). The knee extensors are important muscles for locomotion, recreation, and sport because they contribute to energy generation and absorption to elevate and lower the body's center of mass (^5,16). Weak knee extensors in the involved limb following ACL injury with or without ACL reconstruction is hypothesized to contribute to the use of quadriceps avoidance strategies during multijoint tasks (^19,21).

Subjects in the current study had lower knee extensor NJM in their involved limb compared with their noninvolved limb in the normal squats. This difference was previously reported by Salem et al. (²¹) for bilateral squats and has also been observed in single-limb squats (¹). The average difference in knee extensor NJM between the involved and noninvolved limbs during normal squats was smaller in the current study compared with Salem et al. (²¹) (0.16 vs. 0.26 N·m·kg⁻¹), although large effect sizes (ESs) were observed in both studies (current: ES = 1.66 SD vs. Salem et al. (²¹): ES = 0.88 SD). Subjects in Salem et al. (²¹) performed squats using a barbell loaded with 35% body mass, which would be approximately 23.5 kg based on the average mass of subjects in the current study. The smaller absolute differences in knee extensor NJM between the involved and noninvolved limbs compared with Salem et al. (²¹) may be because of the lower external resistance used in the current study.

Although knee extensor NJM was lower in the involved vs. noninvolved limb during normal bilateral squats, this difference was reduced by placing the noninvolved limb on a 5-cm platform. When the noninvolved limb was elevated, knee extensor NJM was not different between the involved and noninvolved limbs. Although all subjects had greater knee extensor NJM in the noninvolved vs. involved limb during normal squats, elevating the noninvolved limb decreased this difference; in some cases, it resulted in a greater knee extensor NJM in the involved limb (Figure 6). These changes were due in part to reducing knee extensor NJM in the noninvolved limb; however, an increase in knee extensor NJM in the involved limb also occurred. These data indicate that quadriceps avoidance was present, at least for the squat load studied because it was possible to increase knee extensor NJM with the experimental squat. Subsequently, elevating the noninvolved limb during bilateral squats seems to reduce the quadriceps avoidance strategy. However, whether lower knee extensor NJM in the involved limb is solely because of quadriceps avoidance has not been investigated. Both the current study and that of Salem et al. (²¹) used relatively low loads for squat exercise. It is possible that at heavier loads, the involved limb quadriceps are maximally activated; however, because of muscle atrophy, they are not capable of exerting the same moment as in the noninvolved limb.

The increase in knee extensor NJM in the involved limb during the experimental squat can be explained by the increase in vertical ground reaction force. The increase and decrease in vertical ground reaction force under the involved and noninvolved limbs, respectively, occurred when the noninvolved limb was elevated. Brunt et al. (³) also reported an increase and decrease in vertical ground reaction force under the paretic and nonparetic limbs of stroke patients performing sit to stand when the nonparetic limb was elevated. These data indicate that weight was shifted from the noninvolved limb to the involved limb. In both studies, only a single platform height was used. The amount of weight shift may be dependent on platform height. Therefore, it may be hypothesized that the amount of weight shift, and subsequently, change in knee extensor NJM could be optimized by manipulating platform height. Future studies should investigate the relation between platform height and increase in involved limb knee extensor NJM.

Two limitations of this investigation are the small sample size and the large range of time since surgery across subjects. However, as discussed, the presence of lower knee extensor NJM during squats has been reported in previous studies, and all subjects in the current study had lower knee extensor NJM in their involved limb during normal squats. Furthermore, knee extensor NJM increased in the involved limb and decreased in the noninvolved limb for all subjects during the experimental squats. Therefore, despite these limitations, a robust phenomenon—an increase in involved limb knee extensor NJM—was created by performing squats with the noninvolved limb elevated on a 5-cm platform. A training intervention comparing normal with noninvolved limb elevated squats is required to examine whether these biomechanical findings translate to improved outcomes in ACL-injured individuals.

In summary, ACL-reconstructed individuals had knee extensor weakness and exhibited quadriceps avoidance during squats in their involved limb regardless of the time after surgery. The noninvolved limb elevated condition counteracted the quadriceps avoidance strategy by increasing knee extensor NJM in the involved limb. This modification may be favorable for persons with ACL injuries to increase function and reduce strength asymmetry.

Practical Applications

Resistance training involving squat exercise increases knee extensor strength and quadriceps size (^2,24). These adaptations can be explained by the large knee extensor effort required, particularly when squats are performed to at least 120° knee flexion (^4,5,9). However, it has been hypothesized that squat exercise may not be effective for individuals with ACL injury because they may use compensatory strategies that reduce knee extensor effort in their involved limb (²¹). The modified squat investigated, where the noninvolved limb is placed on a 5-cm platform, reduces or eliminates knee extensor asymmetry between the involved and noninvolved limbs. Although a training intervention study is required to examine the efficacy, performing squats with the noninvolved limb elevated may be useful to develop quadriceps size and knee extensor strength in the involved limb of individuals with ACL injury.