Sage Journals: Discover world-class research

Abstract

Background:

Restoring knee extensor strength after anterior cruciate ligament (ACL) reconstruction using a quadriceps tendon (QT) autograft remains challenging. Whether altered quadriceps neuromuscular activity contributes to this weakness is unclear.

Hypothesis:

Quadriceps strength recovery after ACL reconstruction using QT would be impaired owing to altered quadriceps neuromuscular activity.

Study Design:

Cross-sectional study; Level of evidence, 3.

Methods:

A total of 31 patients (Tegner activity scale score ≥7) who had undergone ACL reconstruction 4 months earlier were enrolled into either the QT group (n = 15) or hamstring tendon (HT) group (n = 16). Knee extensor strength was assessed using maximum voluntary isometric and isokinetic (60 deg/s) force. Neuromuscular activity of the vastus lateralis (VL) and vastus medialis (VM) during isometric contraction was recorded with high-density surface electromyography to evaluate motor unit discharge rate (MU DR). A generalized linear mixed-effects model was used to test side (involved and uninvolved limbs) and task (25%, 50%, and 70% of isometric contraction) effects, and regression analysis was used to examine associations between side-to-side isometric strength and MU DR differences.

Results:

Baseline characteristics did not differ significantly between the QT (9 males, 6 females; mean age, 17.1 ± 3.9 years) and HT (10 males, 6 females; mean age, 18.2 ± 2.7 years) groups. The mean time from surgery to testing was also similar (QT group: 120.8 ± 8.0 days; HT group: 121.3 ± 7.2 days). The limb symmetry index for isometric strength showed no difference (QT group: 75.0% ± 19.6%; HT group: 78.2% ± 14.3%; P = .63). In contrast, isokinetic strength at 60 deg/s was significantly lower in the QT group (70.5% ± 13.2%) than in the HT group (83.9% ± 11.6%) (P < .01). In the QT group, MU DR showed a side × task interaction, with higher discharge in the involved (11.7 ± 3.2 pluse per second [pps]) versus uninvolved (10.4 ± 1.9 pps) limbs during 70% isometric contraction in VM (P < .01). Correlations were observed between isometric strength differences and MU DR in both VL (r = 0.65; P = .01) and VM (r = 0.67; P = .01).

Conclusion:

Neuromuscular changes were evident in the QT group, particularly in the VM, but not in the HT group. Notably, in the QT group, altered MU activity was correlated with muscle strength decline.

Keywords

anterior cruciate ligament high-density surface electromyography (HD-sEMG)neuromuscular activity quadriceps tendon autograft

Anterior cruciate ligament (ACL) injury is among the most common knee injuries in athletes, with reinjury remaining a major concern.²⁵ Graft selection plays a crucial role in reducing the risk of reinjury. Although hamstring tendon (HT) autografts are widely used and generally yield consistent outcomes,^15,44 quadriceps tendon (QT) autografts have gained attention due to their greater mechanical strength, reduced anterior knee pain compared with bone–patellar tendon–bone (BTB) autograft,^6,32 and potentially lower reinjury rates.^4,21

Despite these advantages, persistent quadriceps weakness remains a common challenge after ACL reconstruction using QT graft. Restoration of knee extensor strength is essential for returning to sports; however, multiple studies have reported delayed strength recovery in patients receiving QT grafts.^3,16 Several contributing nonmodifiable factors have been proposed, including age,¹¹ sex,^11,20 and surgical techniques such as the amount of tendon harvested and the residual QT area.^11,22 However, the neuromuscular mechanisms underlying this weakness remain poorly understood. Persistent muscle weakness after ACL reconstruction may also increase the risk of secondary ACL injury.^12,24 Therefore, there is an urgent need to uncover evidence regarding muscle recovery after ACL reconstruction using QT.

Arthrogenic muscle inhibition (AMI) has been identified as a key factor limiting quadriceps activation after ACL injury and reconstruction.³⁹ Postoperative swelling, pain, and disruption of the joint mechanoreceptors³⁹ can impair spinal reflex excitability, leading to reduced voluntary activation.^5,26,45 These factors are associated with swelling, pain, and damage to mechanoreceptors.³⁹ ACL reconstruction using QT directly invades the knee extensor mechanism,⁸ which may exacerbate these inhibitory factors and elevate the risk of AMI. Recently, Nuccio et al^35,36 evaluated neuromuscular activity using parameters such as motor unit (MU), recruitment threshold (RT), and discharge rate (DR) in the vastus lateralis (VL) and vastus medialis (VM) using high-density surface electromyography (HD-sEMG) after ACL reconstruction with HT and BTB autografts. These studies observed reduced activity, particularly in the VL. HD-sEMG enables detailed assessment of MU.⁷ This evaluation method is advantageous as discharge timing and patterns are observed, providing information equivalent to that of needle electromyography in a noninvasive manner.⁷ Applying this technique to compare neuromuscular activity after ACL reconstruction using QT or HT could help determine whether quadriceps weakness is likely to recover with rehabilitation and guide future postoperative strategies.

Four months post-ACL reconstruction is a time when knee extension strength in the involved limb suggests increased recruitment of type 2 fibers. The purpose of this study was to clarify the neuromuscular activity of the VL and VM using HD-sEMG at this time point in patients who underwent ACL reconstruction with either QT or HT autografts. We hypothesized that recipients of a QT graft would exhibit reduced neuromuscular activity in both the VL and VM, and that this reduction would be associated with deficits in quadriceps strength.

Methods

Participants

This prospective study included individuals who underwent primary ACL reconstruction at our hospital between January 2023 and January 2025. Eligibility criteria were pre-injury Tegner activity scale score ≥7, age ≤29 years, and intention to return to sports. The exclusion criteria were combined ligament injuries requiring surgical treatment other than ACL, previous knee surgery, or persistent knee pain during muscle strength testing at 4 months postoperatively. The primary outcomes were changes in MU DR in the VL and VM. Secondary outcomes include isokinetic and isometric knee extensor strength.

This study was approved by the hospital's ethics committee (114017). Written informed consent was obtained from all patients; for minors, both patient assent and parental consent were secured.

Surgical Protocol and Rehabilitation

All surgeries were performed by the same experienced orthopaedic surgeon (J.N.) using single anatomic bundle reconstruction. QT grafts were selected for contact sports athletes at high risk of reinjury and for younger individuals with open epiphyseal lines. QT graft harvesting involved segmental resection up to the second layer (including the rectus femoris, VL, and VM) of all soft tissue, with dimensions of 10 mm in width, 55 to 60 mm in height, and 6 mm in thickness. In ACL reconstruction using HT, gracilis tendon was added if the semitendinosus tendon alone could not adequately fill the 6 × 10-mm rounded-rectangle femoral tunnel.^29,33 Formal rehabilitation was conducted once or twice weekly, supplemented by application-based exercise programs. Postoperative rehabilitation was implemented in stages,⁴⁰ beginning with full weightbearing and range of motion (ROM) exercises from the day after surgery. No patients were instructed to limit ROM exercises or weightbearing after meniscal repair. The target ROM for the knee joint was to achieve −5° to 90° by 2 weeks postoperatively and 0° to 120° by 4 weeks postoperatively. Closed kinetic chain training was initiated at 4 weeks postoperatively, including progressive squats and lunge movements. Running was permitted when knee extensor strength reached 60% of the limb symmetry index (LSI) at the 3-month postoperative follow-up. All participants adhered to and faithfully performed the prescribed rehabilitation program. By the time of measurement, all participants had completed the prescribed rehabilitation program, achieved full knee ROM, and were free of residual knee joint effusion, pain, or analgesic use.

Measurements of Knee Extensor Strength

Knee extensor muscle strength was measured at 4 months postoperatively by T.S. or T.W. using an isokinetic dynamometer (Biodex System 4; Biodex Medical Systems). The contraction modes were isokinetic and isometric, and the maximum contraction for each was measured. Before testing, participants completed a 5-minute bicycle ergometer warm-up and standardized knee extension and flexion exercises. Isokinetic testing involved 3 sets of contractions at angular velocities of 60 deg/s and 180 deg/s.⁴³ The first 2 sets served as practice, and the third was recorded as the test trial. Isometric contraction was measured at 45° of knee flexion after a rest period of at least 10 minutes. Each contraction (3-5 seconds) was repeated 2 to 3 times, during which patients were verbally instructed to exert maximal effort. Additional trials were conducted if muscle output increased by ≥5% compared with the prior attempt, and the highest value was recorded. Adequate rest was provided between sets. Measurements were obtained bilaterally, beginning with the uninvolved limb, followed by the involved limb. Peak torque (N·m), body weight ratio (BWR; N·m/kg), and LSI (%) were calculated from the measurement results.

Ultrasound Evaluation

Because subcutaneous tissue thickness may influence electrical resistance during HD-sEMG, the thicknesses of the VL and VM were measured using an ultrasound imaging system (Vscan Air; GE HealthCare). Participants were seated on the Biodex during measurements (Figure 1A). The measurement site was aligned with the center of the electrode based on a previous study.³⁵ For the VL, the electrode was placed at an angle of 20° to the line connecting the anterior superior iliac spine and the lateral border of the patella; for the VM, it was placed at 50° to the line connecting the anterior superior iliac spine and the medial border of the patella (Figure 1B). Subcutaneous tissue thickness was measured in millimeters using ImageJ software (National Institutes of Health). Reliability testing showed excellent reproducibility, with intraclass correlation coefficients (1, 2) of 0.86 (VL) and 0.92 (VM).

Figure 1.

Experimental setup, ramp-up task, and decomposition output results. (A) Knee extensor strength measured by isometric contraction at 45° of knee flexion. (B) Electrode attachment site and 64 electrode grids. The electrodes were attached at 20° to the line connecting the anterior superior iliac spine and the lateral border of the patella for the vastus lateralis (VL) and at 50° to the line connecting the anterior superior iliac spine and the medial border of the patella for the vastus medialis (VM). (C) Three trapezoidal motion tasks. (D) Raster plot showing spike trains for identified motor units. MVIF, maximal voluntary isometric force.

HD-sEMG

Preparation and Recordings

HD-sEMG measurements were performed on the VL and VM at 45° of knee flexion (Figure 1A). A 2-dimensional grid of 64 electrodes (GR08MM1305; OT Bioelettronica; 5 columns × 13 rows, diameter of 1 mm, interelectrode distance of 8 mm) (Figure 1B) was used to record HD-sEMG signals from the target muscle. The electrode grid placement was determined based on a previous study³⁵ and fixed with bioadhesive foam (FOA08MM1305; OT Bioelettronica) and conductive paste (Elefix ZV-181E; Nihon Kohden). A reference electrode was positioned on the ipsilateral proximal tibia. Unipolar HD-sEMG signals were recorded with a digital-to-analog converter (Muovi+ Pro; OT Bioelettronica) at a sampling frequency of 2000 Hz. Signals were amplified with a gain of 150 and bandpass-filtered offline between 10 and 500 Hz. Data analysis was performed using MATLAB software (MATLAB 2023b; MathWorks GK).

Protocol

All participants completed a standardized warm-up followed by muscle strength measurements and HD-sEMG. A 5-minute rest period was provided before HD-sEMG measurement. The maximal voluntary isometric force (MVIF) was measured, after which participants performed HD-sEMG tasks with 3 force levels (25%, 50%, and 70% MVIF). The rest interval between trials was approximately 3 minutes. A submaximal ramp-up contraction task was applied, consisting of 3 phases: recruitment (linear increase in force at 25% MVIF: 2% MVIF/s; 50% and 70% MVIF: 10% MVIF/s), plateau (10 seconds at 25% MVIF; 6 seconds at 50% and 70% MVIF), and derecruitment (linear decrease in force at 25% MVIF: 2% MVIF/s; 50% and 70% MVIF: 10% MVIF/s), in reference to previous studies (Figure 1C).^30,34 Visual feedback was provided to ensure accurate execution of the ramp-up task. The participants were instructed to avoid strenuous exercise and caffeine for 48 hours prior.

Data Analysis

A validated convolutional blind source separation method was applied to separate the HD-sEMG recordings into individual MU discharge timings.^17,18 A single researcher (T.S.) manually reviewed MU spike trains, excluding those of poor quality. Spike intervals <33.3 milliseconds or >250 milliseconds (firing rates >30 and <4 Hz, respectively)¹⁹ and pulse to noise ratio <28 were excluded.¹⁰

The RT and MU DR were calculated for each MU. RT was defined as the absolute (N·m) and relative (% MVIF) forces at which each MU first discharged an action potential. The mean MU DR (pulses/s) was calculated during the plateau phase. The coefficient of variation (CV) of force at the plateau was also computed to confirm task similarity between groups.

The mean MU DR was calculated for each exercise task (25%, 50%, and 70% MVIF) in each group to compare muscle activity between involved and uninvolved limbs. An additional analysis was performed separately for VL and VM to account for muscle-specific effects. To assess the relationship between MU DR and RT, MUs were classified as low-threshold MUs (LTMU: RT <30% MVIF) or high-threshold MUs (HTMU: RT ≥30% MVIF), and their characteristics were analyzed. Finally, the side-to-side difference in maximum knee extensor strength (isometric: ΔMVIF; isokinetic: Δisokinetic strength) and MU DR (ΔMU DR) were calculated for each participant, and their associations were analyzed within each group.

Statistical Analysis

Normality was tested using the Shapiro-Wilk test. Between-group comparisons of baseline characteristics and the LSI of lower limb strength were performed using the Student t test or Wilcoxon test. Depending on normality, effect size was calculated using the Hedge g or Cliff delta test. The Cramer V test was used to calculate the effect size for the chi-square test. Knee extensor strength and subcutaneous tissue thickness were analyzed using 2-way analysis of variance (ANOVA) with graft type (QT and HT) and side (involved and uninvolved limbs) as factors. The effect size (η²) was also calculated. A generalized linear mixed-effects model (GLMM) with random intercept and slope was applied to compare the MU DRs, LTMUs, HTMUs, and CVs of force. Covariates included graft type (QT and HT), side (involved and uninvolved limbs), task intensity (25%, 50%, and 70% MVIF), and RT for MU DR; side and RT for LTMUs and HTMUs; and graft, side, and task for CV of force. Post hoc tests were corrected using the Bonferroni method. Furthermore, the marginal R² was calculated as the model's explanatory power.

For the association between MU DR and RT, repeated-measures correlation (r_rm) was performed to account for individual results. The associations were calculated for the 24 categories: graft (QT and HT), side (involved and uninvolved), muscle (VL and VM), and task (25%, 50%, and 70% MVIF). The Spearman correlation coefficient was used to assess the correlation of MU DR between the involved and uninvolved limbs, including all tasks within each group and when analyzed separately for VL and VM muscles. For 2-way ANOVA, GLMM, and repeated-measures correlation, the Benjamini-Hochberg method was applied as a multiplicity strategy adjustment technique.

The ΔMU DR, ΔMVIF, Δisokinetic strength were calculated as [(involved limb – uninvolved limb)/uninvolved limb] × 100. These associations were evaluated using Spearman correlation analysis.

All statistical analyses were performed using R Version 3.6.2 (R Foundation for Statistical Computing) and JMP Pro Version 18 (SAS Institute) software, with the statistical significance level set at .05.

Sample size was estimated using G*Power Version 3.1.9.7 (Franz Paul). Assuming a medium effect size (f) of 0.40, α of .05, power of 0.95, and 6 repeated measures (side [involved/uninvolved] × task [25%, 50%, 70% MVIF]), the required sample size was calculated as 24 participants, with at least 12 per group.

Results

Participant Characteristics and Knee Extensor Strength

A total of 31 participants were enrolled and assigned to 2 groups: 15 patients received QT grafts and 16 patients received HT grafts (Figure 2). No statistically significant differences were observed in baseline characteristics, including age, height, and weight (Table 1). Subcutaneous tissue thickness also showed no statistically significant differences for graft- and within-group interactions (graft × side: P = .86 for VL and P = .98 for VM) (Appendix 1). Similarly, no statistically significant main effects were observed (graft: P = .66 for VL and P = .76 for VM; side: P = .86 for VL and P = .76 for VM) (Appendix 1). Both isometric and isokinetic (60 deg/s and 180 deg/s) contractions revealed significant differences in peak torque and BWR between the involved and uninvolved limbs (Table 2). However, in the QT group, the peak torque at 60 deg/s was significantly lower (P = .04) (Table 2). Furthermore, no interactions were observed across all items (Table 2). The LSI for each measurement condition was as follows. Mean isometric contractions were 75.0% ± 19.6% in the QT group, and 78.2% ± 14.3% in the HT group (P = .63; g = 0.18; 95% CI, −0.54 to 0.89). Mean isokinetic contractions at 60 deg/s were 70.5% ± 13.2% in the QT group and 83.9% ± 11.6% in the HT group (P < .01; g = 1.02; 95% CI, 0.26 to 1.78), and at 180 deg/s, 74.6% ± 14.8% in the QT group and 86.5% ± 15.3% in the HT group (P = .04; g = 0.74; 95% CI, 0.00 to 1.48).

Figure 2.

Flowchart of participant recruitment. A total of 38 patients met the inclusion criteria; 7 were excluded and 31 patients were allocated to the quadriceps tendon (QT) (n = 15) or hamstring tendon (HT) (n = 16) group. Because the Tegner activity scale score alone may not adequately reflect sports activity level, age was also considered in group allocation. ACL, anterior cruciate ligament; PCL, posterior cruciate ligament.

Table 1

Patient Characteristics ^a

	QT (n = 15)	HT (n = 16)	P Value	Effect Size (95% CI)
Sex, male/female	9:6	10:6	.89	0.03 (0 to 0.31) ^b
Age, y	17.1 ± 3.9	18.2 ± 2.7	.08	0.37 (–0.09 to 0.70) ^c
Height, m	1.65 ± 0.08	1.70 ± 0.06	.07	0.71 (–0.02 to 1.45) ^d
Weight, kg	58.9 ± 6.3	62.9 ± 9.1	.17	0.49 (–0.24 to 1.21) ^d
BMI, kg/m²	21.7 ± 1.9	21.7 ± 2.1	.96	0.02 (–0.73 to 0.70) ^d
Preinjury TAS score	9 (7-9)	9 (7-9)	.84	0.04 (–0.30 to 0.37) ^c
Time from injury to surgery, days	49.1 ± 30.7	53.6 ± 37.1	.95	0.02 (–0.42 to 0.40) ^c
Time from surgery to measurement, days	120.8 ± 8.0	121.3 ± 7.2	.86	0.06 (–0.65 to 0.78) ^d
Meniscal injury	8 (53.3)	11 (68.8)	.38	0.16 (0 to 0.51) ^b
Injury to the dominant side	7 (46.7)	9 (56.3)	.59	0.10 (0 to 0.44) ^b

Values are reported as mean ± SD, median (minimum-maximum), or n (%) unless otherwise indicated. All participants with meniscal injury underwent meniscal repair. There were no cases in which meniscectomy was performed. Depending on normality, either the Hedge g test, Cliff delta test, or Cramer V test was used to measure effect size. The Hedge g test was used for parametric data and is classified as small (>0.20), medium (>0.50), or large (>0.80) effect size. The Cramer V test was used for the chi-square test and is classified as small (>0.10), medium (>0.30), or large (>0.50) effect size. The Cliff delta test was used for nonparametric data and is classified as small (0.147 < δ≤ 0.33), medium (0.33 < δ≤ 0.474), or large (δ > 0.474) effect size. After confirming normality using the Shapiro-Wilk test, appropriate comparisons between the 2 groups were performed. BMI, body mass index; HT, hamstring tendon; QT, quadriceps tendon; TAS, Tegner activity scale.

Cramer V.

Cliff delta.

Hedge g.

Table 2

Knee Extensor Strength in Isometric and Isokinetic Contractions 4 Months Postoperatively ^a

ContractionType	Variable	Angle or Velocity	Graft	Torque		P Value (η²); 95% CI
ContractionType	Variable	Angle or Velocity	Graft	Uninvolved	Involved	Graft	Side	Graft × Time
Isometric	Peak torque, %	45°	QT	148.7 ± 36.9	112.1 ± 42.2	.42 (0.02); –16.3 to 4.9	<.01 (0.16); –28.5 to −7.3	.94 (0.0001); –11.0 to 10.2
	Peak torque, %		HT	159.3 ± 42.2	124.4 ± 39.7	.42 (0.02); –16.3 to 4.9	<.01 (0.16); –28.5 to −7.3	.94 (0.0001); –11.0 to 10.2
	BWR, N·m/kg		QT	2.5 ± 0.52	1.9 ± 0.6	.75 (0.003); –0.18 to 0.12	<.01 (0.22); –0.44 to −0.15	.75 (0.002); –0.17 to 0.12
	BWR, N·m/kg		HT	2.5 ± 0.55	2.0 ± 0.58	.75 (0.003); –0.18 to 0.12	<.01 (0.22); –0.44 to −0.15	.75 (0.002); –0.17 to 0.12
Isokinetic	Peak torque, %	60 deg/s	QT	139.2 ± 41.0	99.2 ± 36.1	.04 (0.08); 120.6 to 140.8	<.01 (0.16); –21.6 to −1.6	.50 (0.008);13.5 to 6.6
		60 deg/s	HT	155.4 ± 41.1	129.3 ± 34.0	.04 (0.08); 120.6 to 140.8	<.01 (0.16); –21.6 to −1.6	.50 (0.008);13.5 to 6.6
		180 deg/s	QT	102.5 ± 35.8	75.6 ± 28.0	.25 (0.03); –14.4 to 2.5	.04 (0.10); –19.3 to −2.4	.54 (0.007); –11.0 to 5.8
		180 deg/s	HT	109.1 ± 34.9	92.6 ± 28.9	.25 (0.03); –14.4 to 2.5	.04 (0.10); –19.3 to −2.4	.54 (0.007); –11.0 to 5.8
	BWR, N·m/kg	60 deg/s	QT	2.3 ± 0.58	1.7 ± 0.52	.10 (0.06); –0.26 to 0.01	<.01 (0.22); –0.40 to −0.14	.28 (0.02); –0.20 to 0.06
		60 deg/s	HT	2.4 ± 0.45	2.1 ± 0.45	.10 (0.06); –0.26 to 0.01	<.01 (0.22); –0.40 to −0.14	.28 (0.02); –0.20 to 0.06
		180 deg/s	QT	1.7 ± 0.51	1.3 ± 0.42	.42 (0.01); –0.16 to 0.07	<.01 (0.14); –0.29 to −0.06	.42 (0.01); –0.16 to 0.06
		180 deg/s	HT	1.7 ± 0.42	1.5 ± 0.37	.42 (0.01); –0.16 to 0.07	<.01 (0.14); –0.29 to −0.06	.42 (0.01); –0.16 to 0.06

Values are reported as mean ± SD unless otherwise indicated. Boldface type indicates statistical significance (α = .05). Effect size η² is considered small (>0.01), medium (>0.06), or large (>0.14). BWR, body weight ratio; HT, hamstring tendon; QT, quadriceps tendon.

MU Decomposition

Overall, 1285 MUs were identified (QT group: total = 585, involved = 258, uninvolved = 327; HT group: total = 700, involved = 354, uninvolved = 346). The distribution and number of MUs for each group, muscle, and contraction level are presented as absolute and relative RTs (Figure 3, Appendices 2 and 3). The mean number of MUs identified was as follows: QT group (VL = 3.4; VM = 3.1) and HT group (VL = 4.1; VM = 3.2).

Figure 3.

Relative and absolute recruitment thresholds (RTs) for all motor units (MUs). The RTs for the quadriceps tendon (QT, A and C) and hamstring tendon (HT, B and D) groups were clustered and displayed as relative and absolute RTs. Panels A and B show the number of MUs recruited at each exercise load. MVIF, maximal voluntary isometric force; VL, vastus lateralis; VM, vastus medialis.

Properties of MU

Motor Unit Discharge Rate

No statistically significant graft × side × task interaction was found for the CV of force (P = .91; 95% CI, −0.49 to 0.55; marginal R² = 0.60). Similarly, no significant main effects were observed for graft (P = .13; 95% CI, −2.89 to 0.79; marginal R² = 0.60), side (P = .10; 95% CI, −2.96 to 0.57; marginal R² = 0.60), and task (P = .10; 95% CI, −0.36 to 2.34; marginal R² = 0.60). The MU DR during the plateau phase is shown for the involved and uninvolved limbs in each group (Figure 4). In the QT group, the VL showed no significant effects (P = .92), although a main effect was observed for the task (P < .01) (Appendix 4). Post hoc tests confirmed significant differences across all tasks (task 1-2: P < .01, task 1-3: P < .01, task 2-3: P < .01) (Appendix 4). In the VM, a significant side × task interaction was observed (P < .01) (Appendix 4), with post hoc tests indicating a significant difference between involved and uninvolved limbs during the 70% MVIF task (P < .01) (Appendix 4). In the HT group, no significant side × task interaction was detected for VL (P = .65) or VM (P = .61), although a main effect for task was observed (P = .01). Post hoc tests showed statistically significant differences across all tasks between VL (task 1-2: P < .01, task 1-3: P < .01, task 2-3: P < .01) and VM (task 1-2: P < .01, task 1-3: P < .01, task 2-3: P < .01) (Appendix 4).

Figure 4.

Comparison of motor unit discharge rates (MU DRs). The MU DRs of the quadriceps tendon (QT) and hamstring tendon (HT) groups were examined using a generalized linear mixed-effects model. The covariates included side (involved and uninvolved), task (25%, 50%, and 70%), and RT. (B) The results show a significant difference between the involved and uninvolved limbs in the 70% vastus medialis (VM) task in QT (P < .01). (A-D) Similarly, a significant difference was observed in the MU DR between all exercise tasks (P < .01). Statistical significance was set at an α of .05. VL, vastus lateralis.

When comparing LTMUs and HTMUs, significant differences were found in the QT VM (involved vs uninvolved) for both LTMUs (P = .03) and HTMUs (P = .03) (Figure 5, Appendix 5). The VL in the QT group and the VL and VM in the HT group were not significantly different (Figure 5, Appendix 5).

Figure 5.

Comparison of motor unit discharge rates (MU DRs) of low- and high-threshold motor units. (A-D) Comparison of low threshold motor units (LTMUs) (recruitment threshold [RT] < 30% maximal voluntary isometric force [MVIF]) and high threshold motor units (HTMUs) (RT ≧ 30% MVIF) between involved and uninvolved limbs within the quadriceps tendon (QT) and hamstring tendon (HT) groups. (B) In the involved limb, the MU DR was low in the LTMUs and high in the HTMUs of the vastus medialis (VM) in the QT group. Statistical significance was set at an α of .05. VL, vastus lateralis.

A negative correlation was observed between RT and MU DR (r_rm = −0.81 to −0.48; P < .01) across all motor tasks (graft × side × muscle × task; n = 24) (Appendix 6). Single regression analysis of the MU DR for the QT and HT groups revealed that when all motor tasks were included, the HT group exhibited more similar motor patterns between the involved and uninvolved limbs compared with the QT group (QT group: r = 0.36, P < .01; HT group: r = 0.53, P < .01) (Figure 6, A and D; Appendix 7). The VL and VM were analyzed separately: VL (QT group: r = 0.49, P < .01; HT group: r = 0.59, P < .01) (Figure 6, B and E; Appendix 7) and VM (QT group: r = 0.22, P = .18; HT group: r = 0.46, P < .01) (Figure 6, C and F; Appendix 7). Only the VM of the QT group showed dissimilar muscle activity patterns between the involved and uninvolved limbs.

Figure 6.

Correlation chart between the motor unit discharge rates (MU DRs). The figure shows a single correlation of the MU DR between the involved and uninvolved limbs within each group. When all MU DRs of the vastus lateralis (VL) and vastus medialis (VM) were included, (A) weak and (D) moderate correlations were found in the quadriceps tendon (QT) and hamstring tendon (HT) groups, respectively. When divided into each muscle, moderate correlations were observed in the (B) VL in the QT group and in both the (E) VL and (F) VM in the HT group, whereas only the (C) VM in the QT group showed no correlation. Statistical significance was set at an α of .05.

MVIF and MU DR Deficits

Linear regression was performed to identify the associations between MU variables and MVIF deficits. In the QT group, ΔMVIF and ΔMU DR were significantly correlated in both VL (r = 0.65; P = .01) and VM (r = 0.67; P = .01) (Figure 7, A and B; Appendix 8). In the HT group, no significant correlations were found (VL: r = 0.25, P = .37; VM: r = −0.02, P = .94) (Figure 7, C and D; Appendix 8). In contrast, Δisokinetic strength at 60° was significantly correlated with ΔMU DR in the HT group for both VM (r = 0.54; P = .04) and VL (r = 0.60; P = .01) (Appendices 8 and 9).

Figure 7.

Relationship between side-to-side differences in motor unit discharge rate (MU DR) and maximal voluntary isometric force (MVIF). Regression analysis of ΔMU DR and MVIF {[(involved limb – uninvolved limb)/uninvolved limb] × 100} for each graft and muscle, respectively. (A and B) In the quadriceps tendon (QT) group, ΔMU DR and ΔMVIF were significantly correlated in both the vastus lateralis (VL) and vastus medialis (VM). (C and D) In contrast, no correlation was observed in the hamstring tendon (HT) group. Statistical significance was set at an α of .05.

Discussion

This study is the first to demonstrate changes in neuromuscular activity after ACL reconstruction using QT. ACL reconstruction using QT resulted in increased MU DR at high VM output compared with the uninvolved limb (Figures 4B and 5B), and the MU DR pattern in the VM differed from that of the uninvolved limb (Figure 6). Additionally, in the QT group, the difference in MU DR correlated with the difference in muscle strength between the involved and uninvolved limbs for both the VL and VM. These results partially support our hypothesis.

In the QT group, a side-to-side difference in MU DR of the VM was observed at high power (70% MVIF) (Figure 4B). Furthermore, MU DR decreased in LTMUs and increased in HTMUs of the involved limb compared with the uninvolved limb in the QT group (Figure 5B). These findings suggest that ACL reconstruction using QT disrupts neuromuscular activity, particularly in the VM. Surgical harvesting of the QT directly invades the anterior thigh,⁸ leading to postoperative swelling and a potentially high risk of AMI. This invasion alters afferent feedback from damaged mechanoreceptors and contributes to abnormal excitability in the Ib inhibitory pathway, flexion reflex, and gamma loop within spinal reflexes and corticospinal tracts,^27,39 thereby affecting the centrifugal neural input to the muscle from the alpha motor neuron pool.³⁹ Consequently, MU size in the quadriceps is reduced, influencing DR.³¹ The DR of MUs is suppressed by inactivity, such as disuse, and MUs with lower thresholds are more affected, whereas MUs with higher thresholds have impaired neuromuscular junction transmission.³⁷ The difficulty in eliciting quadriceps contraction after QT collection suggests that the MUs in the VMs of the QT group may have exhibited similar biological responses. Additionally, the side-to-side correlations were stronger in the HT group (Figure 6, A and D). When analysis was separated into VL and VM, only the VM in the QT group showed no side-to-side correlation (Figure 6C), further supporting disruption of VM neuromuscular activity after ACL reconstruction using QT.

In contrast, the regression line between ΔMVIF and ΔMU DR showed a correlation only for the VL and VM in the QT group (Figure 7, A and B). This suggests that neurological factors in these muscles contribute to weakness after ACL reconstruction using QT, which may have clinical implications. Interventions such as neuromuscular electrical stimulation (NMES), which modulate MU activity and DR,^1,2 may therefore be especially beneficial for patients with QT grafts. Previous studies have shown the effectiveness of NMES after ACL reconstruction,^28,42 and its optimal duration, frequency, and impact on neuromuscular activity in QT-based ACL reconstruction warrant further investigation. While NMES was not included in the present study, future work is planned to evaluate its effects in the QT group.

The isokinetic and isometric muscle force measurements performed in this study yielded interesting results. The LSI for isokinetic contractions was approximately 20% lower in the QT group than in the HT group. Muscle weakness after ACL reconstruction using QT is a well-recognized issue,^3,16,23 and our findings are consistent with previous reports. However, isometric contraction revealed no significant differences in knee extensor strength. This suggests that sustained contraction at a fixed joint angle can produce sufficient output, whereas isokinetic contraction where tendon extensibility changes with joint angle may fail to transmit adequate traction force. The finding that only VL and VM in the HT group showed a correlation between Δisokinetic strength and ΔMU DR (Appendix 7) may support this interpretation. However, knowledge in this area remains limited, and further investigation using imaging methods such as ultrasound is needed to explore the relationship between QT site regeneration and recovery of knee extensor strength.

Thus, reduced muscle output after ACL reconstruction using QT may have resulted from multiple factors, including altered neuromuscular activity and tendon harvesting. Neuromuscular activity, however, could potentially be modulated through rehabilitation. In addition to NMES, cold therapy¹⁴ combined with resistance training,^38,41 which is known to improve strength, may be key for recovery after ACL reconstruction using QT.

This study has several limitations. First, older age may hinder strength recovery after ACL reconstruction using QT; however, our study included only young athletes. Therefore, studies involving broader age groups are required. Second, although sex differences in MUs and DRs have been reported,¹³ both sexes were included, and the nearly equal sex ratio likely minimized bias. Third, statistical comparisons did not adjust for baseline characteristics. The reasons for this are that most comparisons were between involved and uninvolved limbs, and no patient characteristics differed significantly. Finally, MU was assessed only during isometric contraction at a fixed joint angle and muscle strength level, without considering the length-tension relationship. The differing results between isometric and isokinetic contractions suggest that MU dynamics also vary under dynamic conditions. However, the activity analysis of MUs is not suitable for the analysis of MUs because the muscle fibers immediately below the electrode move when joint movement is involved, and thus the analysis of MUs during isokinetic contraction was not performed in this study. Recently, Glaser and Holobar⁹ reported the identification of MUs during dynamic muscle contraction; future studies applying this method could provide more detailed insights into neural changes.

Conclusion

After ACL reconstruction using a QT autograft, neuromuscular activity, particularly in the VM, is disrupted, as evidenced by altered MU discharge patterns. However, no change was observed in the MU discharge pattern during HT autograft. Furthermore, in the QT group, there is a notable correlation between altered MU activity and muscle strength decline.

Footnotes

Appendix

Appendix 8

Relationship Between Side-to-Side Differences in Motor Unit Discharge Rate Values ^a

Contraction Type	Angle or Velocity	Graft	Muscle	r	95% CI	P Value
Isometric	45°	QT	VL	0.65	0.19 to 0.88	.01
		QT	VM	0.67	0.21 to 0.89	.01
		HT	VL	0.25	−0.32 to 0.68	.37
		HT	VM	−0.02	−0.52 to 0.49	.94
Isokinetic	60 deg/s	QT	VL	0.40	−0.14 to 0.76	.14
		QT	VM	0.41	−0.16 to 0.77	.15
		HT	VL	0.54	0.04 to 0.83	.04
		HT	VM	0.60	0.17 to 0.84	.01
	180 deg/s	QT	VL	0.12	−0.42 to 0.59	.68
		QT	VM	0.17	−0.39 to 0.65	.55
		HT	VL	0.28	−0.28 to 0.69	.32
		HT	VM	0.26	−0.27 to 0.70	.33

HT, hamstring tendon; QT, quadriceps tendon; VL, vastus lateralis; VM, vastus medialis.

Final revision submitted October 15, 2025; accepted October 26, 2025.

One or more of the authors has declared the following potential conflict of interest or source of funding: This work was supported by JSPS Grants-in-Aid for Scientific Research (KAKENHI) grant No. JP22K11336. AOSSM checks author disclosures against the Open Payments Database (OPD). AOSSM has not conducted an independent investigation on the OPD and disclaims any liability or responsibility relating thereto.

Ethical approval for this study was obtained from the Ethics Review Committee of Kanazawa University (114017).

ORCID iDs

Takuya Sengoku

Yuichi Nishikawa

Takaya Watabe

Yushin Mizuno

Manase Nishimura

Kentaro Fujita

References

Bickel

Gregory

Dean

JC.

Motor unit recruitment during neuromuscular electrical stimulation: a critical appraisal. Eur J Appl Physiol. 2011;111(10):2399-2407.

Borzuola

Nuccio

Scalia

, et al. Adjustments in the motor unit discharge behavior following neuromuscular electrical stimulation compared to voluntary contractions. Front Physiol. 2023;14:1212453.

Bram

Stevens

Greenberg

, et al. Postoperative strength differences at short-term follow-up vary based on autograft harvest site after adolescent transphyseal anterior cruciate ligament reconstruction. Arthroscopy. 2024;40(5):1591-1598. doi:10.1016/j.arthro.2023.10.018

Brinkman

Tummala

Moore

Economopoulos

KJ.

All-soft tissue quadriceps tendon autograft in revision anterior cruciate ligament reconstruction in athletes: comparison to bone-patellar tendon-bone autograft with at least a 2-year follow-up. Am J Sports Med. 2022;50(14):3770-3777.

Criss

Melton

Ulloa

, et al. Rupture, reconstruction, and rehabilitation: a multi-disciplinary review of mechanisms for central nervous system adaptations following anterior cruciate ligament injury. Knee. 2021;30:78-89.

Dai

Leng

Wang

Cheng

Quadriceps tendon autograft versus bone-patellar tendon-bone and hamstring tendon autografts for anterior cruciate ligament reconstruction: a systematic review and meta-analysis. Am J Sports Med. 2022;50(12):3425-3439.

Del Vecchio

Holobar

Falla

Felici

Enoka

Farina

Tutorial: analysis of motor unit discharge characteristics from high-density surface EMG signals. J Electromyogr Kinesiol. 2020;53:102426.

Fink

Lawton

Förschner

Gföller

Herbort

Hoser

Minimally invasive quadriceps tendon single-bundle, arthroscopic, anatomic anterior cruciate ligament reconstruction with rectangular bone tunnels. Arthrosc Tech. 2018;7(10):e1045-e1056.

Glaser

Holobar

Motor unit identification from high-density surface electromyograms in repeated dynamic muscle contractions. IEEE Trans Neural Syst Rehabil Eng. 2019;27(1):66-75.

10.

Goodlich

Del Vecchio

Kavanagh

JJ.

Motor unit tracking using blind source separation filters and waveform cross-correlations: reliability under physiological and pharmacological conditions. J Appl Physiol. 2023;135(2):362-374.

11.

Greiner

Drain

Setliff

, et al. Factors associated with knee extension strength symmetry after anterior cruciate ligament reconstruction with quadriceps tendon autograft. Orthop J Sports Med. 2024;12(3):23259671241239692.

12.

Grindem

Snyder-Mackler

Moksnes

Engebretsen

Risberg

MA.

Simple decision rules can reduce reinjury risk by 84% after ACL reconstruction: the Delaware-Oslo ACL cohort study. Br J Sports Med. 2016;50(13):804-808.

13.

Guo

Jones

Inns

, et al. Neuromuscular recruitment strategies of the vastus lateralis according to sex. Acta Physiol (Oxf). 2022;235(2):e13803.

14.

Hart

Kuenze

Diduch

Ingersoll

CD.

Quadriceps muscle function after rehabilitation with cryotherapy in patients with anterior cruciate ligament reconstruction. J Athl Train. 2014;49(6):733-739.

15.

Hoffer

Peck

Leith

McConkey

Lodhia

Single bundle hamstrings autograft preparation techniques for anterior cruciate ligament reconstruction: current concepts. Arch Orthop Trauma Surg. 2024;144(7):3185-3196.

16.

Holmgren

Noory

Moström

Grindem

Stålman

Wörner

Weaker quadriceps muscle strength with a quadriceps tendon graft compared with a patellar or hamstring tendon graft at 7 months after anterior cruciate ligament reconstruction. Am J Sports Med. 2024;52(1):69-76.

17.

Holobar

Farina

Blind source identification from the multichannel surface electromyogram. Physiol Meas. 2014;35(7):R143.

18.

Holobar

Farina

Gazzoni

Merletti

Zazula

Estimating motor unit discharge patterns from high-density surface electromyogram. Clin Neurophysiol. 2009;120(3):551-562.

19.

Holobar

Minetto

Farina

Accurate identification of motor unit discharge patterns from high-density surface EMG and validation with a novel signal-based performance metric. J Neural Eng. 2014;11(1):016008.

20.

Hunnicutt

Xerogeanes

Tsai

, et al. Terminal knee extension deficit and female sex predict poorer quadriceps strength following ACL reconstruction using all-soft tissue quadriceps tendon autografts. Knee Surg Sports Traumatol Arthrosc. 2021;29(9):3085-3095.

21.

Hurley

Mojica

Kanakamedala

, et al. Quadriceps tendon has a lower re-rupture rate than hamstring tendon autograft for anterior cruciate ligament reconstruction—a meta-analysis. J ISAKOS. 2022;7(2):87-93.

22.

Inoue

Kayaalp

Giusto

, et al. Quadriceps tendon size does not affect postoperative strength recovery following quadriceps tendon anterior cruciate ligament reconstruction. J ISAKOS. 2024;9(5):100308.

23.

Johnston

McClelland

Feller

Webster

KE.

Knee muscle strength after quadriceps tendon autograft anterior cruciate ligament reconstruction: systematic review and meta-analysis. Knee Surg Sports Traumatol Arthrosc. 2021;29(9):2918-2933.

24.

Kyritsis

Bahr

Landreau

Miladi

Witvrouw

Likelihood of ACL graft rupture: not meeting six clinical discharge criteria before return to sport is associated with a four times greater risk of rupture. Br J Sports Med. 2016;50(15):946-951.

25.

Law

Miller

, et al. Age, rehabilitation and surgery characteristics are re-injury risk factors for adolescents following anterior cruciate ligament reconstruction. Phys Ther Sport. 2021;49:196-203.

26.

Lepley

Gribble

Thomas

Tevald

Sohn

Pietrosimone

BG.

Quadriceps neural alterations in anterior cruciate ligament reconstructed patients: a 6-month longitudinal investigation: ACLr neural alterations. Scand J Med Sci Sports. 2015;25(6):828-839.

27.

Lepley

Grooms

Burland

Davi

Kinsella-Shaw

Lepley

LK.

Quadriceps muscle function following anterior cruciate ligament reconstruction: systemic differences in neural and morphological characteristics. Exp Brain Res. 2019;237(5):1267-1278.

28.

Jin

Chen

, et al. Effects of neuromuscular electrical stimulation on quadriceps femoris muscle strength and knee joint function in patients after ACL surgery: a systematic review and meta-analysis of randomized controlled trials. Orthop J Sports Med. 2025;13(1):23259671241275070.

29.

Magnussen

Lawrence

JTR

West

Toth

Taylor

Garrett

WE.

Graft size and patient age are predictors of early revision after anterior cruciate ligament reconstruction with hamstring autograft. Arthroscopy. 2012;28(4):526-531.

30.

Martinez-Valdes

Negro

Falla

De Nunzio

Farina

Surface electromyographic amplitude does not identify differences in neural drive to synergistic muscles. J Appl Physiol. 2018;124(4):1071-1079.

31.

McPherson

Schilaty

Anderson

Nagai

Bates

NA.

Arthrogenic muscle inhibition after anterior cruciate ligament injury: injured and uninjured limb recovery over time. Front Sports Act Living. 2023;5:1143376.

32.

Mouarbes

Menetrey

Marot

Courtot

Berard

Cavaignac

. Anterior cruciate ligament reconstruction: a systematic review and meta-analysis of outcomes for quadriceps tendon autograft versus bone-patellar tendon-bone and hamstring-tendon autografts. Am J Sports Med. 2019;47(14):3531-3540.

33.

Nakase

Toratani

Kosaka

, et al. Technique of anatomical single bundle ACL reconstruction with rounded rectangle femoral dilator. Knee. 2016;23(1):91-96.

34.

Nishikawa

Watanabe

Holobar

Kitamura

Maeda

Hyngstrom

AS.

Sex differences in laterality of motor unit firing behavior of the first dorsal interosseous muscle in strength-matched healthy young males and females. Eur J Appl Physiol. 2024;124(7):1979-1990.

35.

Nuccio

Del Vecchio

Casolo

, et al. Deficit in knee extension strength following anterior cruciate ligament reconstruction is explained by a reduced neural drive to the vasti muscles. J Physiol. 2021;599(22):5103-5120.

36.

Nuccio

Del Vecchio

Casolo

, et al. Muscle fiber conduction velocity in the vastus lateralis and medialis muscles of soccer players after ACL reconstruction. Scand J Med Sci Sports. 2020;30(10):1976-1984.

37.

Piasecki

Motor unit adaptation to disuse: crossing the threshold from firing rate suppression to neuromuscular junction transmission. J Physiol. 2024;603(13):3701-3709. doi:10.1113/JP284159

38.

Pucci

Griffin

Cafarelli

Maximal motor unit firing rates during isometric resistance training in men: motor unit firing during isometric resistance training. Exp Physiol. 2006;91(1):171-178.

39.

Rice

McNair

PJ.

Quadriceps arthrogenic muscle inhibition: neural mechanisms and treatment perspectives. Semin Arthritis Rheum. 2010;40(3):250-266.

40.

Sengoku

Nakase

Asai

, et al. The effect of gracilis tendon harvesting in addition to semitendinosus tendon harvesting on knee extensor and flexor strength after anterior cruciate ligament reconstruction. Arch Orthop Trauma Surg. 2022;142(3):465-470.

41.

Sterczala

Miller

Dimmick

Wray

Trevino

Herda

TJ.

Eight weeks of resistance training increases strength, muscle cross-sectional area and motor unit size, but does not alter firing rates in the vastus lateralis. Eur J Appl Physiol. 2020;120(1):281-294.

42.

Toth

Tourville

Voigt

, et al. Utility of neuromuscular electrical stimulation to preserve quadriceps muscle fiber size and contractility after anterior cruciate ligament injuries and reconstruction: a randomized, sham-controlled, blinded trial. Am J Sports Med. 2020;48(10):2429-2437.

43.

Undheim

Cosgrave

King

, et al. Isokinetic muscle strength and readiness to return to sport following anterior cruciate ligament reconstruction: is there an association? A systematic review and a protocol recommendation. Br J Sports Med. 2015;49(20):1305-1310.

44.

Yaari

Singer

Goldberg

, et al. Eighteen-year outcome of anterior cruciate ligament reconstruction with patellar tendon or hamstring autograft. Arch Orthop Trauma Surg. 2024;144(5):2189-2195.

45.

Zarzycki

Morton

Charalambous

Pietrosimone

Williams

Snyder-Mackler

Examination of corticospinal and spinal reflexive excitability during the course of postoperative rehabilitation after anterior cruciate ligament reconstruction. J Orthop Sports Phys Ther. 2020;50(9):516-522.

Relationship Between Quadriceps Neuromuscular Activity and Knee Extensor Strength at Four Months After ACL Reconstruction in High-Level Young Athletes

Abstract

Background:

Hypothesis:

Study Design:

Methods:

Results:

Conclusion:

Keywords

Methods

Participants

Surgical Protocol and Rehabilitation

Measurements of Knee Extensor Strength

Ultrasound Evaluation

HD-sEMG

Preparation and Recordings

Protocol

Data Analysis

Statistical Analysis

Results

Participant Characteristics and Knee Extensor Strength

MU Decomposition

Properties of MU

Motor Unit Discharge Rate

MVIF and MU DR Deficits

Discussion

Conclusion

Footnotes

Appendix

ORCID iDs

References