Evaluating Gait with Force Sensing Insoles 6 Months after Anterior Cruciate Ligament Reconstruction: An Autograft Comparison

Abstract

Introduction:

Aberrant knee mechanics during gait 6 months after anterior cruciate ligament reconstruction (ACLR) are associated with markers of knee cartilage degeneration. The purpose of this study was to compare loading during walking gait in QT, bone-patellar tendon-bone (BPTB), and hamstring tendon (HT) autograft patients 6 months post-ACLR using loadsol single sensor insoles, and to evaluate associations between loading and patient reported outcomes.

Methods:

72 patients (13–40 years) who underwent unilateral, primary ACLR with BPTB, QT, or HT autograft completed treadmill gait assessment, the International Knee Documentation Committee (IKDC) survey and the ACL-Return to Sport after Injury (ACL-RSI) survey 6±1 months post-ACLR. Ground reaction forces were collected using loadsols. Limb symmetry indices (LSI) for peak impact force (PIF), loading response instantaneous loading rate (ILR), and loading response average loading rate (ALR) were compared between groups using separate ANCOVAs. Survey scores were compared between groups using one-way ANOVAs. The relationships between IKDC, ACL-RSI, and LSIs were compared using Pearson’s product moment correlation coefficients.

Results:

There were no significant differences between graft sources for LSI in PIF, ILR, ALR, nor impulse. Patient-reported knee function was significantly different between graft source groups with the BPTB group reporting the highest IKDC scores; however, there was no significant difference between groups for ACL-RSI score. There were no significant associations between IKDC score, ACL-RSI score, and biomechanical symmetry among any of the graft source groups.

Conclusions:

Autograft type does not influence PIF, ILR, ALR, or impulse during walking 6 months post-ACLR. Limb symmetry during gait is not strongly associated with patient reported outcomes regardless of graft source. Loadsols appear to be a suitable tool for use in the clinical rehabilitation setting.

INTRODUCTION

Surgical reconstruction is the gold standard treatment to restore normal knee stability and function after anterior cruciate ligament (ACL) injury (1). A recent consensus statement recommends ACL reconstruction (ACLR) for active patients wishing to maintain athletic participation due to its higher rates of return to sport and the risk of secondary meniscal injury that comes with chronic ACL insufficiency with nonoperative treatment (1). However, ACLR is not an effective treatment following ACL injury to reduce the development of knee osteoarthritis (OA), as at least 36% of patients who undergo ACLR will develop radiographic OA within a decade of surgery (2). While the exact mechanism by which OA develops after ACLR is not completely understood, it is hypothesized that OA onset may be due to alterations in biomechanics (3–5), pathological changes in biochemical processes (6–8), and the traumatic mechanisms of the initial injury and surgery (9–11).

Individuals who undergo ACLR experience persistent alterations in knee kinetics and kinematics during gait (12–14). This is concerning because asymmetrical gait patterns are linked to biochemical and compositional imaging markers of knee cartilage degeneration following ACLR (3–5, 15–17). When examining ground reaction forces during gait, patients who have undergone ACLR demonstrate lesser vertical ground reaction force (vGRF) during early stance and greater vGRF during midstance compared to uninjured controls at both 6 and 12 months post-ACLR (18). Importantly, these same alterations in vGRF have been observed <12 months post-ACLR in both patients who experience clinically significant symptoms based on their Knee Osteoarthritis and injury Outcome Score (KOOS) responses (19) and those who express greater levels of serum biomarkers indicative of cartilage degeneration (15). Due to the association between these post-ACLR gait alterations, clinical symptoms, and cartilage degeneration, it is crucial for clinicians to be able to identify these changes in vGRF to inform rehabilitation.

Previous studies have demonstrated that graft source influences knee biomechanics during tasks such as walking, jumping and cutting post-ACLR (20–22). However, the vast majority of lower extremity biomechanics research in patients with ACLR has focused on patients receiving bone-patellar tendon-bone (BPTB) or hamstring tendon (HT) autografts because these graft sources have been the most utilized for more than 2 decades (23). As a result, there is a gap in knowledge concerning biomechanical outcomes after ACLR with quadriceps tendon (QT) autograft despite its increasing use in the treatment of young physically active patients with ACL injury. Recent estimates indicate that as of 2021 more than 10% of primary ACLRs and of 2022 as many as 50% of revision ACLRs were completed using a QT autograft. (23, 24).

The current literature appears to suggest that patients who receive BPTB autograft perform worse on functional tasks than those who receive HT autograft. Specifically, during walking gait, BPTB patients demonstrate a reduced knee flexion moment at midstance than HT patients 9–12 months post-ACLR (20). When performing a bilateral countermovement jump 9 months post-ACLR, BPTB patients demonstrate greater impulse asymmetries than HT patients (21). In addition, when changing direction, BPTB patients demonstrate lower vGRF than HT patients in their ACLR limb at this time point (22). These findings indicate that patients with BPTB autograft have greater biomechanical asymmetries post-ACLR, which is hypothesized to be due to the BPTB graft harvest’s disruption of the extensor mechanism (21, 22). Because of this, BPTB patients often exhibit delayed quadriceps strength recovery when compared to HT patients (25, 26), which may lead to adoption of compensatory strategies during loading. Because QT autograft harvest disrupts the extensor mechanism in a similar manner to BPTB autograft harvest and QT autograft patients also exhibit delayed extension strength recovery post-ACLR (25, 26), it is possible that QT patients may demonstrate similar compensatory mechanisms to BPTB patients, leading to greater asymmetry than HT patients.

Lower extremity biomechanics research involving patients with ACLR has also typically been performed in a lab environment using 3-dimensional motion analysis. However, this approach is not translatable to a clinical rehabilitation setting due to the cost of the required technology, the training required to reliably use a 3D motion capture system, and the technical expertise required to analyze and interpret the findings (27). Since there is evidence that limb loading asymmetries during gait lead to long-term negative consequences for the preservation of articular cartilage (3–5), it is crucial to identify and utilize technology that allows clinicians to objectively evaluate biomechanical asymmetries in the clinical environment. Loadsol (Novel Electronics, St. Paul, MN, USA) single sensor insoles are an in shoe wearable technology that has been developed to assess realtime impact forces (IF), which are analogous to ground reaction forces (28), during walking and landing tasks in both healthy and post-ACLR patient populations (27, 29). Loadsols have exhibited good to excellent validity against traditional force plates for the collection of ground impact forces during treadmill walking (29), and IFs collected from loadsols have been found to be strongly associated with knee extension moment measurements collected with 3D motion analysis (27). This emerging technology utilizes an in-shoe insole which is synced to a mobile application to collect real time IFs and provide clear summaries of important metrics, such as symmetry indices, without requiring advanced processing or training. Given their ease of use and overall versatility, loadsols may be a feasible way for clinicians to better measure asymmetries in IFs and loading rates during rehabilitation in an objective manner to provide real-time feedback to patients.

Due to the association between a higher risk of developing OA and asymmetrical gait patterns 6 months post-ACLR (3–5, 15–17), it is important to better understand if graft source has an influence on gait mechanics at this time, and to do so using technology that can be adopted in the clinical environment. While impact forces may not be the only important measurement to assess in post-ACLR knee mechanics, tools such as the loadsol allow for a first step towards incorporating advanced biomechanics into clinical rehabilitation. Therefore, the purpose of this study was to compare limb loading symmetry during walking gait in BPTB, QT, and HT autograft patients 6 months post-ACLR using loadsol single sensor insoles. We hypothesized that BPTB and QT autograft patients would demonstrate greater between-limb asymmetry than HT autograft patients due to the aforementioned deficits associated with disrupting the extensor mechanism. Additionally, as aberrant gait has previously been associated with worse clinical symptoms (19), we hypothesized that greater peak impact force (PIF) asymmetry would be associated with worse patient-reported knee related function and psychological readiness for sport.

METHODS

The current study was a multisite cross-sectional sub-analysis incorporating data from three ongoing prospective longitudinal studies investigating lower extremity biomechanics following ACLR. Each of the three studies were approved by their respective institutional review boards (Inova Health System = WCG IRB 20216925, University of Virginia = UVA IRB 220225, Michigan State University = MSU IRB 00002816). All adult participants provided written informed consent prior to participation. All minor participants provided written assent and one parent or guardian provided written consent prior to participation.

Participants

Participants were recruited from two academic institutions and one hospital system 6±1 months post-ACLR. This included 72 total participants, with 25 from Inova Health, 14 from the University of Virginia, and 33 from Michigan State University. Recruited patients were between the ages of 13 and 40 and underwent ACLR between May 2022 and July 2023. Participants were excluded if they: 1) had evidence of radiographic osteoarthritis (Kellgren-Lawrence grade > 1), 2) had a history of ipsilateral or contralateral knee surgery, 3) had a lower extremity fracture at the time of ACL injury, 4) had inflammatory arthritis, immunodeficiency, or articular cartilage damage greater than 3A, or 5) underwent multiligamentous reconstruction or subtotal meniscectomy at the time of ACLR. Post-operative rehabilitation was not standardized between patients and patients had the ability to attend physical therapy at the clinic of their choosing; however, standard protocols were provided by surgeons to guide rehabilitation. Representative rehabilitation protocols have been provided from Inova Health (see Physical Therapy Protocol Following ACL Reconstruction, Supplemental Digital Content), the University of Virginia (see ACL Reconstruction Rehabilitation Protocol, Supplemental Digital Content), and Michigan State University (see ACL Reconstruction Post-operative Rehabilitation Protocol, Supplemental Digital Content). At Michigan State University, some surgeons provided the online rehabilitation guidelines made available by the Multicenter Orthopaedic Outcomes Network (30).

Surgical Technique

ACLR was performed by one of 16 surgeons at three institutions using a standardized independent drilling technique. All meniscus pathology was either repaired or excised during diagnostic arthroscopy prior to graft harvest. In all cases, patients were placed in a hinged knee brace locked in extension before being woken from anesthesia.

Limb Loading Assessments

Participants completed a treadmill walking gait assessment in a rehabilitation clinic or clinical research laboratory under the supervision of a study team member. Treadmill gait was chosen over overground walking as a goal of this study was to perform testing within a clinical environment in a manner that is most feasible in any rehabilitation clinic. Prior to beginning the treadmill walking assessment, the loadsol single sensor insoles were inserted into the participant’s shoes and the loadsols were then calibrated using the standardized protocol described by Renner et al (29). Following calibration, the participant was weighed (N) using the loadsols to ensure that the sensors were correctly assessing the patient’s weight. The participant’s weight was entered into the loadsol app to normalize the PIF data to the participant’s body weight (xBW). To assess limb loading, PIF data was collected at 100Hz using loadsol single sensor insoles.

For the walking gait assessment, participants walked on a treadmill at a self-selected pace for a minimum of 30 seconds with a minimum of 40 gait cycles captured during the trial. This volume of gait cycles was chosen to ensure stability of our measures within each participant given the limitations in sampling rate of the loadsols. At the University of Virginia and Inova Health, participants were instructed to walk at a typical pace they might adopt when walking to class or through a grocery store. Participants were allowed to start the treadmill, adjust the treadmill speed, and walk at the self-selected speed for a short time prior to initiation of data collection. At Michigan State University, average gait speed was determined during a series of five overground walking bouts through timing gaits at a speed they would describe as their typical walking speed, and the average gait speed was used to set the treadmill speed. In all cases, cadence (steps per minute) was calculated as a proxy for gait speed given that both increased gait speed and cadence both result in greater vGRFs during the stance phase of gait. Data quality was monitored in real time and the walking data were recollected if there were any perceived issues with data quality. The variables of interest were peak impact force (PIF), instantaneous loading rate (ILR), average loading rate (ALR), and impulse. Impulse was calculated as the area under the force–time curve from heel strike to toe-off. ALR was calculated as the slope of the middle 60% of the weight acceptance portion of the force–time curve using the method presented by Goss and Gross (31). ILR was calculated as the greatest slope between consecutive data points during the same portion of the force–time curve using methods described by Milner et al (32). Limb symmetry indices (LSI; %) were calculated for each biomechanical variable (Equation 1).

$$Limb\hspace{0.33em}symmetry\hspace{0.33em}index=\hspace{0.33em}\frac{Involved\hspace{0.33em}limb}{Contralateral\hspace{0.33em}limb}\hspace{0.33em}\times 100$$ Equation 1

Data were exported from the loadsol mobile application and processed using a custom processing program developed by Peebles et al (33). PIF data were not filtered based on the recommendations of Renner et al (29).

Patient-Reported Outcome Measures

International Documentation Committee Subjective Knee Evaluation Form (IKDC).

The IKDC (34) is an 18-item form scored from 0 to 100 with higher scores representing fewer symptoms, greater knee function, and higher sport-related activity. The IKDC is commonly used in evaluating patients who experience sport injury and is one of the most reported outcomes after ACLR. The IKDC has been validated to be responsive to various knee conditions, and normative data has been established based on sex and age (35).

ACL-Return to Sport After Injury Scale (ACL-RSI).

The ACL-RSI (36) is a 12-item scale scored from 0 to 100 with higher scores representing a greater psychological readiness to return to sport following ACL injury. Items in the ACL-RSI evaluate three domains that are proposed to contribute to psychological readiness – emotions, confidence in performance, and risk appraisal. To date, it is the only scale developed for this purpose, and ACL-RSI score can predict return to sport outcomes (37, 38).

Statistical Plan

Descriptive statistics were calculated for demographics, surgical characteristics, patient-reported outcome measures and biomechanical variables of interest. Continuous demographic variables and surgical characteristics were compared between graft source groups using one-way ANOVA while categorical variables were compared between groups using ꭓ² tests.

Patient-reported knee function and psychological readiness for sport were compared between graft source groups using one-way ANOVAs. Involved limb, contralateral limb, and LSI for PIF, ILR, ALR, and impulse were compared between graft source groups using separate ANCOVAs. Due to the subtle differences in protocols utilized to determine gait speed between sites, we opted to control the effects of protocol in our analysis by including institution as a covariate. Cadence was also included as a covariate in our analysis to account for the between individual differences in self-selected walking pace. Lastly, we included participant age at the time of testing as a covariate given that lower extremity kinetics may differ from mid-adolescence through early adulthood. The relationships between patient-reported knee function, psychological readiness for sport, and LSIs for each biomechanical variable were assessed in the total sample and among each graft source group using Pearson’s product moment correlation coefficients. A-priori alpha level was p < 0.05. Statistical analysis was completed using an open-source statistical package (v 2.2.5, Jamovi) and data visualizations were completed using the ggplot package in R Studio (2023.06.0+421).

RESULTS

A summary of demographics, surgical characteristics and patient-reported outcome measures can be found in Table 1. Graft source groups did not differ based on sex (p=0.198), age (p=0.382), medial (p=0.418) or lateral meniscus (p=0.589) treatment, time since surgery (p=0.646), nor cadence (p=0.166). Patient-reported knee function was significantly different between graft source groups (p=0.048) with the BPTB group reporting the highest IKDC scores; however, there was no significant difference between groups for ACL-RSI score (p=0.293) (Table 1).

Table 1.

Between group comparison of demographics, surgical characteristics, and patient-reported outcome measures

	BPTB Autograft	QT Autograft	HS Autograft	p-value
Sex, n (%)
Female	13 (46.4%)	19 (70.4%)	10 (58.8%)	0.198
Male	15 (53.6%)	8 (29.6%)	7 (41.2%)
Age, yrs	20.1±5.3	21.7±6.8	19.0±5.3	0.382
Height, m	1.7±9.7	1.7±9.4	1.7±7.7	0.197
Mass, kg	73.8±12.1	71.3±13.3	80.2±21.2	0.314
Medial meniscus surgery, n (%)				0.418
No treatment	17 (60.7%)	22 (81.5%)	13 (76.5%)
Partial meniscectomy	1 (3.6%)	1 (3.7%)	1 (5.9%)
Repair	10 (35.7%)	4 (14.8%)	3 (17.6%)
Lateral meniscus surgery, n (%)
No treatment	20 (71.4%)	16 (59.3%)	8 (47.1%)	0.589
Partial meniscectomy	5 (17.9%)	6 (22.2%)	5 (29.4%)
Repair	3 (10.7%)	5 (18.5%)	4 (23.5%)
Time since surgery, mo	6.5±0.7	6.3±0.8	6.4±1.0	0.646
Cadence, steps per minute	107.0±9.2	108.0±11.2	102.3±8.9	0.166
IKDC score, 0–100	80.0±11.7	72.4±9.7	74.5±17.6	0.048
ACL-RSI score, 0–100	64.8±20.6	56.9±23.0	63.8±24.9	0.293

IKDC, International Knee Documentation Committee; ACL-RSI, ACL-Return to Sport after Injury; BPTB, bone-patellar tendon-bone; QT, quadriceps tendon; HS, hamstring tendon

While accounting for the effects of site, cadence, and age, there were no significant differences between graft sources for involved limb PIF (p=0.498), ILR (p=0.987), ALR (p=0.782), nor impulse (p=0.735) (Table 2). Similarly, there were no significant differences between graft sources for contralateral limb PIF (p=0.706), ILR (p=0.969), ALR (p=0.914), nor impulse (p=0.780). In addition, there were no significant differences between graft sources for limb symmetry in PIF (p=0.836), ILR (p=0.884), ALR (p=0.979), nor impulse (p=0.992) (Table 2, Figure 1). Instances of significant associations between covariates and biomechanical outcomes of interest are presented in Table 2. Among all participants, IKDC score was weakly related to ILR LSI (r=0.258, p=0.029) and impulse LSI (r=0.236, p=0.046) but not PIF LSI (r=0.190) nor ALR LSI (r=0.051) (Table 3). There were no significant associations between patient-reported knee function, psychological readiness for sport, and biomechanical symmetry among any of the graft source groups (p>0.05) (Table 3).

Table 2.

Between group comparison of loading metrics during the stance phase of walking gait

	BPTB Autograft	QT Autograft	HS Autograft	p-value
Peak impact force, BW
Involved limb	1.072±0.105	1.107±0.150	1.037±0.066	0.498†
Contralateral limb	1.102±0.101	1.147±0.189	1.062±0.108	0.706†
LSI (%)	97.54±6.30	97.75±10.99	98.21±6.11	0.836‡
Instantaneous loading rate, BW*s−1
Involved limb	13.91±3.16	14.93±8.25	13.01±2.74	0.987†
Contralateral limb	14.07±3.22	15.18±8.46	13.31±3.50	0.969†
LSI (%)	101.17±14.41	101.36±15.99	100.47±9.66	0.884‡
Average loading rate, BW*s−1
Involved limb	7.756±2.186	7.970±3.806	7.094±2.691	0.782†
Contralateral limb	8.042±2.383	8.240±3.831	7.416±3.333	0.914†
LSI (%)	101.67±19.82	100.57±21.38	106.49±25.24	0.979‡
Impulse, BW*s
Involved limb	0.576±0.062	0.578±0.069	0.588±0.059	0.735†
Contralateral limb	0.585±0.057	0.597±0.060	0.592±0.058	0.780†
LSI (%)	98.49±5.09	97.07±8.89	99.01±3.75	0.992

BW, Body Weight (N); LSI, limb symmetry index; BPTB, bone-patellar tendon-bone; QT, quadriceps tendon; HS, hamstring tendon

^†indicates that when entered as a covariate, cadence was significantly associated with the gait variable of interest (p<0.05)

^‡indicates that when entered as a covariate, cadence was significantly associated with the gait variable of interest (p<0.05); research site was not associated with any outcome of interest.

Figure 1:

Between group comparison of loading metric limb symmetry.

BPTB, bone-patellar tendon-bone; QT, quadriceps tendon; HS, hamstring tendon.

Table 3.

Correlations between loading metrics and patient-reported outcome measures

	IKDC Score r (p-value)	ACL-RSI r (p-value)
Peak impact force LSI (%)
Total sample	0.190 (0.111)	0.024 (0.843)
BPTB Autograft	0.285 (0.150)	0.116 (0.557)
QT Autograft	0.147 (0.465)	−0.014 (0.947)
HS Autograft	0.258 (0.302)	0.051 (0.840)
Instantaneous loading rate LSI (%)
Total sample	0.258 (0.029)*	0.165 (0.165)
BPTB Autograft	0.301 (0.127)	0.236 (0.226)
QT Autograft	0.220 (0.270)	0.140 (0.496)
HS Autograft	0.415 (0.086)	0.187 (0.458)
Average loading rate LSI (%)
Total sample	0.051 (0.669)	0.164 (0.169)
BPTB Autograft	0.133 (0.508)	0.209 (0.286)
QT Autograft	0.076 (0.707)	0.024 (0.909)
HS Autograft	−0.023 (0.929)	0.302 (0.223)
Impulse LSI (%)
Total sample	0.236 (0.046)*	0.069 (0.567)
BPTB Autograft	0.267 (0.179)	0.028 (0.887)
QT Autograft	0.273 (0.169)	0.118 (0.566)
HS Autograft	0.211 (0.402)	−0.055 (0.829)

LSI, limb symmetry index; BPTB, bone-patellar tendon-bone; QT, quadriceps tendon; HS, hamstring tendon; IKDC, International Knee Documentation Committee; ACL-RSI, ACL-Return to Sport after Injury

DISCUSSION

To our knowledge, this is the first study comparing lower extremity impact forces during gait post-ACLR with QT, BPTB, and HT autografts. We found no significant differences between graft sources for PIF, ILR, ALR, or impulse during walking at 6 months post-ACLR. Additionally, limb symmetry during gait was not strongly associated with patient-reported knee function nor psychological readiness for sport regardless of graft source. This finding is in contrast with previous studies that have reported significant associations between symmetry in lower extremity loading and patient-reported outcome measures in patients post-ACLR (19, 39). As a whole, our findings appear to indicate that autograft source may not affect lower extremity loading 6 months post-ACLR when measured with the loadsol and that gait alterations may not be associated with patient-reported outcome measures.

Webster et al previously found significant differences between BPTB and HT autograft patients in external knee moments during walking gait due to differences in donor site morbidity and the BPTB graft’s disruption of the extensor mechanism (20). As our study did not find differences between graft groups, our findings appear to be inconsistent with those of Webster et al (20). While this may be due to fact that the previous study measured external knee moments while ours measured impact forces, and this may have made our study unable to detect the differences found by Webster et al (20)., PIF measured via loadsol has been found to correlate with external knee moment symmetry measured via 3D motion capture (27). Additionally, it is notable that there were differences in time since surgery between our cohort (5–7 months) and Webster’s cohort (9–12 months) (20). Since changes in ground reaction forces during gait occur between 6 and 12 months post-ACLR (18), it is possible that differences between grafts had not developed at our tested time point. Regardless, because alterations in gait patterns at 6 months post-ACLR are associated with increased cartilage degeneration (3–5), our findings that there are no differences between the BPTB, QT, and HT autograft cohorts are still noteworthy, and should be considered during surgical counseling.

Neither IKDC nor ACL-RSI scores were associated with symmetry with the exception of weak associations between IKDC score and instantaneous loading rate LSI and impulse LSI (Table 3), which may suggest that self-reported function and psychological readiness are not associated with impact forces measured by the loadsols 6 months post-ACLR. This was unexpected, given that individuals who report worse symptoms typically exhibit less involved limb vGRF and lower vGRF LSI as compared to those without symptoms (19, 39). However, this has been challenged in a previous study by Collins et al which found no association between KOOS score and ACLR limb vGRF during walking gait in young women 6 months post-ACLR (40). In addition, these prior studies used KOOS subscales as their measure of patient outcomes, whereas our study reports IKDC and ACL-RSI scores. Additionally, while our patients report generally low IKDC and ACL-RSI scores compared to some previous studies including healthy controls and those who return to preinjury sport (35, 41), they are comparable to some other studies (42). These discrepancies may be due to a variety of factors, including differences in study design, time since surgery, or rehabilitation practices. Regardless, taken alongside our gait findings, it appears that, while ACLR may impact ground impact forces during gait and patient-reported symptoms 6 months post-ACLR, among our cohort, graft source is not associated with these outcomes.

Our use of loadsol single sensor insoles is a less expensive option than traditional force plates and does not require advanced training, making our assessments clinically translatable. Importantly, in patients who have undergone ACLR, PIF measurements collected from the loadsol during functional tasks has been found to be strongly associated with force plate symmetry outcomes and knee extension moment symmetry measured with 3D motion capture (27). Given that post-ACLR loading asymmetry during gait is associated with several negative long-term outcomes, including increased cartilage degeneration (3–5), greater likelihood of failing return-to-sport criteria (43), and worse patient-reported outcomes (19, 39), patients may benefit from the inclusion of more objective functional testing in their rehabilitation programs to assess loading asymmetry.

This study was not without limitations. First, this study was conducted at multiple sites, with multiple surgeons at each site. Due to this, surgical techniques may have varied. While this may impact the internal validity of our results, it would also serve to increase their external validity. Second, we were not able to recruit an equivalent number of patients in each graft cohort. This was due to recent shifts in clinical practice at each institution away from HT autografts and towards BPTB and QT autografts for individuals with a desire to return to sport. Third, our study only analyzed a single time point (6±1 months post-ACLR). Therefore, alterations in knee mechanics that might appear at other time points may have been missed, especially as it is known that vGRF during gait changes between 6 and 12 months post-ACLR (18). However, because aberrant walking gait at 6 months post-ACLR is specifically associated with negative long-term outcomes (3–5, 12–14), our results are still significant for clinical decision making. Fourth, the sampling rate of 100 Hz of loadsol single sensor insoles is lower than that of traditional force plates. However, traditional force plates are not as easy to use in the clinical setting, and loadsols at 100 Hz have been validated against force plates (27, 29). While future studies may strive to use 200 Hz loadsols, we were limited by the equipment available to us (29). Additionally, because the loadsol can only measure ground impact forces, we were not able to analyze further kinematic or kinetic outcomes. As part of the objective of this study was to make our assessments as clinically feasible as possible, the addition of more tools was not possible at this time, and we believe loadsols may be a first step towards implementing more advanced biomechanical analyses in the clinical rehabilitation environment. Lastly, while gait is a meaningful activity to assess joint loading and OA risk, it does not encompass all functional limitations that may be present in the post-ACLR population. In order to more thoroughly assess between-graft differences in joint loading as it relates to outcomes such as re-injury rates, other tasks such as jump landing change of direction should be assessed in the future.

CONCLUSIONS

Asymmetry in ground impact forces during gait 6 months post-ACLR is associated with markers of cartilage degeneration (3–5, 15–17). As graft source is an important factor in the surgical counseling process and the QT autograft has become an increasingly popular choice for young, physically active patients (23), it is important to better understand if graft source has an influence on gait mechanics, and to do so using technology that can be adopted in the clinical environment. This study is the first, to our knowledge, to compare ground impact forces during gait post-ACLR with QT, BPTB, and HT autografts. Patients in these autograft groups do not exhibit significant differences in PIF, ILR, ALR, or impulse during walking 6 months post-ACLR. Additionally, limb symmetry during gait is not strongly associated with patient-reported knee function nor psychological readiness for sport regardless of graft source. Importantly, while our results do not reflect the same differences as other studies that have used motion capture equipment, loadsols still appear to be a potentially useful tool for use in the clinical rehabilitation setting to provide more objective functional testing in patients’ rehabilitation programs, thereby allowing for early, direct intervention to ultimately improve health outcomes and quality of life.

Supplementary Material

Supplemental Data File (.doc, .tif, pdf, etc.)

SDC 1: Supplemental Digital Content.pdf