Original Research

Biomechanical Deficiencies in Women withSemitendinosus-Gracilis Anterior Cruciate LigamentReconstruction During Drop JumpsAlexis Ortiz, PT, PhD, SCS, CSCS, Carmen E. Capo-Lugo, MSPT,Heidi L. Venegas-Rios, MS, DrPH

Objective: To compare landing mechanics and neuromuscular recruitment strategiesbetween women with semitendinosus-gracilis anterior cruciate ligament reconstruction(SG-ACLr) and noninjured women during double- and single-legged drop jumps.Design: Cross-sectional biomechanical study.Setting: Single university-based biomechanics laboratory.Participants: Fourteen women 1-5 years posteSG-ACLr and 16 noninjured womenparticipated in this study.Methods: After anthropometric measurements, warm-up, and familiarization procedures,participants performed 5 trials of a double- and single-legged drop jumps.Main Outcome Measurements: Dynamic knee valgus was measured as the distancebetween knee joints during the landing phase of the double-leg drop jumps. Medial kneedisplacement was the outcome considered during the landing phase of the single-leg dropjumps. For both drop jump tasks, neuromuscular recruitment was evaluated throughrectified normalized electromyographic activity of the quadriceps and hamstrings (ampli-tude and latency), and quadriceps/hamstrings electromyographic co-contraction ratio.Results: Although the SG-ACLr group demonstrated a tendency toward a greater dynamicknee valgus during both drop jumps, these differences did not reach statistical significance.EMG data revealed different neuromuscular strategies for each group, depending on thespecific jump.Conclusions: These findings suggest that women with SG-ACLr have a tendency towardgreater dynamic knee valgus that could predispose to additional knee injuries. Rehabili-tation specialists need to be aware of existing kinematic and neuromuscular deficienciesyears after SG-ACLr. Taking this into consideration will aid in prescribing appropriateinterventions designed to prevent re-injury.

PM R 2014;-:1-10

A.O. School of Physical Therapy, TexasWoman’s University, 6700 Fannin Street,Houston, TX 77030. Address correspondenceto: A.O.; e-mail: aortiz10@twu.eduDisclosure: nothing to disclose

C.E.C.-L. Interdepartamental NeuroscienceProgram, Northwestern University, Chicago, ILDisclosure: nothing to disclose

H.L.V.-R. Division of Biostatistics, School ofPublic Health, University of Texas, Houston, TXDisclosure: nothing to disclose

This study was supported by grants8U54MD007587-03, NCRR-NIH U54RR026139-01, G12RR003051, and the Na-tional Strength & Conditioning AssociationFoundation.

Submitted for publication August 9, 2013;accepted July 8, 2014.

INTRODUCTION

Anterior cruciate ligament reconstruction (ACLr) is often recommended after an ACL tear,especially for young active individuals who desire to return to high levels of physical ac-tivity [1,2]. Although ACLr technique with bone-patellar-tendon-bone (BPTB) autograft isconsidered as the gold standard [3], some orthopedic surgeons prefer the semitendinosus-gracilis (SG) autograft because of its apparent structural strength, sparing of the extensormechanism, and less graft morbidity [4,5]. These traits are important during high-impactactivities in which the majority of ACL injuries often occur. In addition, the literaturereports that SG reconstruction yields better results in patient satisfaction surveys, kneestrength and endurance, and functional hop tests [4].

It is known that women exhibit small joint angles and larger joint moments whenlanding from jumps or performing cutting maneuvers when compared to their malecounterparts, increasing their predisposition to knee injury [6-9]. These gender-specificbiomechanical differences represent less than optimal biomechanical control, leading todecreased dynamic knee stability and increased risk of ACL injury. Knee stability during

PM&R1934-1482/14/$36.00

Printed in U.S.A.

ª 2014 by the American Academy of Physical Medicine and RehabilitationVol. -, 1-10, Month 2014

http://dx.doi.org/10.1016/j.pmrj.2014.07.0031

2 Ortiz et al BIOMECHANICAL DEFICIENCIES DURING DROP JUMPS

dynamic tasks is accomplished by combining passiverestrictors and dynamic mechanisms. Thus, both kinematicand neuromuscular factors such as muscle activation,recruitment, or firing patterns must be taken into consid-eration to accurately characterize dynamic knee stability.

To quantify knee stability, researchers have measuredknee valgus during static and dynamic tasks [10-13]. Dy-namic knee valgus is characterized by the motion of thedistal femur toward midline, and the distal tibia away fromthe midline of the body. This movement is a combination ofreduced knee flexion and increased hip adduction, hip in-ternal rotation, and knee abduction [14,15]. Proper recoveryof neuromuscular activity after ACLr is of vital importance toregain appropriate mechanical function of the knee joint andprevent further injuries [6,16]. Evidence suggests thatwomen with ACL reconstructions exhibit compensatoryneuromuscular strategies such as altered quadriceps/hamstring co-contraction ratios upon landing, reducedquadriceps femoris activation, increased plantar flexor acti-vation, and quadriceps and hamstrings timing adaptations[6,17,18]. In addition, normal voluntary muscle control hasbeen shown to be re-established to pre-surgery status afterSG reconstruction [5]. When researchers compared pre-surgery and post-surgery (6 months after surgery) musclecontrol of the semitendinosus and gracilis, no differenceswere found in the ability of these muscles to controlvoluntary contractions [5]. In addition, modifications toneuromuscular strategies have been reported to occur afterACLr in order to achieve dynamic stability during dropjumps [6]. Greater quadriceps/hamstring EMG co-contrac-tion ratios, greater activation of gluteal muscles, greaterrectus femoris activation, and similar hamstring muscle ac-tivity during single-leg drop jumps have also been reportedin the literature. However, these compensatory strategiescould have a negatively impact on sports performance byincreasing reaction time and reducing the capacity to absorbimpact [13,19].

After ACLr, subjective procedures and instrumentedstrength measurements are frequently used to determinerecovery. However, these tests and procedures have shownonly moderate correlation with functional performance [20].Examination of athletic tasks in individuals with SG-ACLrmay provide insight about functional capacity and assess-ment of post-surgical re-injury risk, and may prove benefi-cial for the development of diagnostic and rehabilitative tools[6,18]. Drop jumps are an appropriate task to test functionalperformance after ACLr because of their resemblance tosports-specific maneuvers and their capability of generatingknee valgus moments and challenging the neuromuscularsystem [10].

The existent body of evidence has not determinedwhether kinematic predisposing factors and lack of muscularcontrol persist after SG-ACLr during unplanned athleticactivities that require high levels of dynamic stability [6,21].To our knowledge, no previous study has compared landing

mechanics and muscle recruitment and firing patterns be-tween women with SG-ACLr and healthy controls during theperformance of drop jump tasks. Therefore, the purpose ofthis study was to compare dynamic knee stability andneuromuscular control between women with SG-ACLr andnoninjured women during drop jumps. It was hypothesizedthat women with SG-ACLr would exhibit larger knee valgusand EMG activity predisposing to knee injury during bothjumps, and that these differences would be greater in thesingle-leg drop jump.

METHODS

Study Design

A cross-sectional study comparing landing mechanics be-tween healthy women and women with SG-ACL recon-struction during 2 drop jump conditions was conducted. A60-cm double-legged and a 40-cm single-legged drop jumpswere used as functional tasks to assess bilateral and unilaterallanding strategies, respectively. The dependent variablescomprised kinematic (valgus) and neuromuscular (electro-myography) factors. Specifically, the kinematic assessmentconsisted of dynamic knee valgus, whereas the electromy-ography (EMG) analyses comprised amplitude, latency, andco-contraction assessments. Sample size was determinedbased on previous studies [6]. Based on a previously pub-lished effect size of 0.32 for knee valgus and a significancelevel of .05, a total sample of 39 subjects was needed to reach80% of power. However, because of the inability to identifymore than 15 subjects with ACL reconstruction from thesame orthopedic surgeon and the desire to keep an evennumber of participants per group, a total of 30 participantswere contacted for possible recruitment.

Participants

Fifteen women with SG-ACLr (height: 167.71 � 9.0 cm,body mass: 67.68 � 11.66 kg) who underwent ACLr withthe same orthopedic surgeon and performed the samerehabilitation protocol were recruited. Noninjured controlsubjects (n ¼ 16) were recruited from the collegiate com-munity (height: 160.50 � 5.17 cm, body mass: 59.35� 10.37 kg). Both groups were between the ages of 21 and35 years. All of the women with ACL reconstruction except 2were injured while participating in competitive volleyball atthe collegiate or professional level. To ensure a similar levelof athletic skills, the control subjects were participating involleyball, basketball, and soccer at the collegiate or intra-mural sports level. At the moment of recruitment, all par-ticipants reported engaging in sports-specific physicalactivities as described by the Activity Rating Scale [22].Participants reported scores ranging from 12 to 16, consis-tent with activities such as running, cutting, decelerating,and pivoting more than 2 times per week, which is

PM&R Vol. -, Iss. -, 2014 3

considered a high level of participation [22]. All participantsread and signed an informed consent form approved by theUniversity of Puerto Rico Institutional Review Board beforeparticipation in the study.

All SG-ACLr women were selected from the computerdatabase of a single orthopedic surgeon to control for possiblebias due to surgical technique. In the samemanner, all patientsunderwent the same rehabilitation protocol set by the ortho-pedic surgeon. The rehabilitation protocol initially targetedacute inflammation and range of motion with progressiontoward strengthening, functional activities, and neuromus-cular re-education until return to pre-injury level wasaccomplished. Patients were contacted chronologically from2006 to 2010 and asked if they were willing to participate inthe study. Time since surgery was between 12 months and 5years. Although time since surgery had a 5-year range, allsubjects were allowed to return to all of their pre-injury ac-tivities without restrictions. Control participants wererecruited from the collegiate community using flyers. Partic-ipants with ACLr were excluded if they demonstrated at least 1of the following: inability to stick the landing or perform 2screening tasks (single hop for distance and cross-over hop),and signs and symptoms of knee instability in the recon-structed limb during these tasks. These 2 tasks are commonlyused as physical performance measures in individuals recov-ering from ACL injury or surgical reconstruction. Moreover,these tasks have been reported as being comparable to highfunctional demands needed in sports [11,23]. Further exclu-sion criteria for the ACLr group were as follows: signs or re-ports of swelling and/or effusion after physical activity, fear ofperforming high-demand physical activity, multiple (>1)surgeries in the same knee, and/or self-reported pregnancy.Noninjured subjects were excluded if they had experiencedany lower back or lower extremity injury within the past 6months, had any surgery of the lower back or lower extremity,and/or were self-reported to be pregnant.

Instrumentation

Three-dimensional lower extremity kinematics was capturedusing a real-time motion analysis system with 5 infraredcameras (Nexus 1.3; MX13þ; 120-Hz sampling rate, ViconMotion Systems, Denver, CO). Spherical reflective markers(25-mm) were placed over the skin of both lower extremitiesaccording to Vicon’s lower extremity Plug-in Gait model(Figure 1). Before each data collection session, the equip-ment was calibrated according to manufacturer’s recom-mendations, and a static trial to estimate the participant’sjoint centers was conducted.

Muscle activity of the vastus medialis, vastus lateralis,rectus femoris, and medial and lateral hamstrings weremeasured in the involved leg of women with ACLr and thedominant leg of the control subjects using surface electro-myography (EMG) (Bagnoli EMG System; Delsys, Inc.,Boston, MA). Pre-amplified Ag/AgCl surface EMG electrodes

(signal bandwidth: 20-450 Hz [�3 dB]; impedance:>100 KU; noise: <1.2 mV RMS; sampling rate: 1 kHz) with2 embedded 10.0 � 1.0-mm bars (10.0-mm space betweenbars) were used to record muscular activity. Surface EMGelectrodes were applied on the skin over the muscles of in-terest as reported previously [6]. Before positioning theelectrodes over the muscles of interest, the skin over thoselocations was cleansed to decrease skin impedance. Hypo-allergenic tape and stretchable straps were used over theelectrodes to minimize artifact movement during perfor-mance of functional tasks. The EMG unit was connected tothe main amplifier where signals were amplified (gain ¼1000) and filtered by using a bandpass filter (20-450 Hz).The amplifier was connected via BNC cables to an analog-to-digital (A/D) interface unit connected to a Vicon MX systemlinked to a Vicon MX Ultranet interfaced with the computer.Two foot switches (sampling rate: 500 Hz; Delsys FootSwitch Biosensors; Delsys, Inc., Boston, MA) placed at themid-plantar aspect of the hallux and plantar aspect of thecalcaneus were secured with athletic tape to the foot of thelimb with the ACLr and to the foot of the dominant leg toidentify the events of initial contact and toe-off during theground contact phase of the drop jumps for the SG-ACLrand control groups, respectively.

Procedures

After consenting, each participant performed a set of 2screening tasks as clearance for participation. For the controlsubjects, leg dominance was determined as the preferred legto perform the 2 screening tasks to ensure the best possiblejumping landing performance. The single hop for distanceand cross-over hop were performed as previously reported[6]. Any participant unable to perform either of these taskswas excluded from participating in the investigation. Afterscreening procedures, anthropometric measures accordingto Vicon’s Plug-in-Gait procedures were recorded. Subse-quently, each participant was instrumented with reflectivemarkers, surface EMG electrodes, and foot switches. Eachparticipant performed a standardized warm-up protocolconsisting of 5 minutes on a cycle ergometer at 40-60 rev-olutions per minute, 10 squats, and 5 seconds of repetitivecountermovement jumps previous to the drop-jumps.

After warm-up, all women were allowed to practice bothjump tasks 3 times. During practice jumps, women withACLr performed the single leg landing countermovementjump (SLJ) task with the reconstructed leg, whereas thenoninjured control group performed the task with theirdominant leg. Study data were recorded only after a restperiod of 5 minutes after the completion of the warm-up andpractice protocol. The specific procedures for each of thejump tasks were as described below.

For the double-leg landing countermovement jump(DLJ), each participant stood on top of a 60-cm box. Par-ticipants were given the command “on your mark” as a signal

Figure 1. Plug-in-gait Model Retro-reflective Marker Setting (Vicon Motion System, Denver, CO).

4 Ortiz et al BIOMECHANICAL DEFICIENCIES DURING DROP JUMPS

to stand ready and prepare to perform the jump. A secondcommand, “get set,” was given as a signal to let the partici-pant know that she may jump whenever ready. Participantswere instructed to drop from the box instead of jumping. Ifany participant was observed jumping from the box, creatingan upward movement, the trial was invalidated and repeated.Participants were allowed to use their arms for balance at anytime after initiating the jump. After landing, each participanthad to perform a maximal upward jump. To ensure amaximal effort upward jump, participants were instructed totry to touch the ceiling of the laboratory [24]. Each partici-pant performed 5 trials of this task, and was required to restat least 30 seconds between trials to avoid fatigue [8].

For the single-leg landing countermovement jump (SLJ),each participant stood on top of a 40-cm box. The proceduresof this task were similar to those of the DLJ. Each participantperformed 5 trials with a single leg. Subjects with ACL recon-struction performed this task using their leg with the recon-structed knee, whereas control subjects used their dominantleg (preferred leg to perform a single hop for distance).

Rest intervals were similar to the rest in the double-leggeddrop jump.

Data Reduction

The kinematic outcome evaluated during the jumps was thedynamic knee valgus during the entire ground contact phase(initial contact to takeoff). The average of 5 trials was usedfor analysis. The ground contact phase was defined as theperiod between initial foot floor contact and foot push-offtoward the vertical jump. The trajectory of both lateralfemoral epicondyle markers and knee joint centers werederived from the 3-dimensional trajectory of markers filteredthrough a Woltring filter. Dynamic knee valgus was definedas the distance between markers, placed at the left and rightlateral femoral epicondyles, from initial contact to maximumvalgus (maximum valgus e initial contact) during the con-tact phase of the DLJ (Figure 2). Dynamic knee valgusduring the SLJ was measured for the landing leg as themaximum distance traveled medially by the knee joint

Figure 2. Dynamic knee valgus was defined as the distance between markers placed at the lateral femoral epicondyles at initialcontract minus minimum distance between these 2 markers.

PM&R Vol. -, Iss. -, 2014 5

center, from the moment of initial contact to maximumvalgus, during the ground contact phase.

EMG data were time synchronized to kinematic data via avoltage change from foot switches during the ground contactphase of both drop jumps. The time period for the data ofinterest was similar to kinematic data; from initial contact topush-off. A dynamic normalization procedure was con-ducted by dividing the mean signal during each specific tasktrial by the maximum signal generated during the middle3 seconds of the 5-second tuck jumps performed duringwarm-up. This normalization approach is recommended toanalyze EMG data during dynamic tasks to reduce partici-pants’ variability between trials caused by fatigue and toreduce intersubject variation [6,25-27]. The normalized re-sults were averaged to represent quadriceps and hamstringmuscle groups in their entirety; the rectus femoris and vastiiaveraged muscle activity represented the quadriceps musclegroup, and the averaged medial and lateral hamstring muscleactivity represented the hamstrings muscle group. To obtainthe latency variable for both quadriceps and hamstrings, thetime taken to reach maximum contraction from the momentof initial contact was measured in seconds. A co-contractionratio (CCR) was calculated [6,25,28] by obtaining thenormalized values for both the quadriceps and hamstringmuscle groups during the targeted window of time. The co-contraction value was always less than or equal to 1,

representing joint stiffness and the relative activation of theknee flexor and extensor muscle groups [6,25,28]. A co-contraction closer to 1 indicated optimal co-contraction,protecting the knee joint by increasing joint stability toprevent injury. Values closer to 0 represented poor co-contraction between muscle groups, increasing the proba-bility of knee instability and injury.

Data Analysis

The kinematic and EMG data were screened for normalityassumptions, homoscedasticity, and outliers using the Sha-piro-Wilks test, Levene test, and histograms and box plots,respectively. Intraclass correlation coefficients (ICC) for theaverage of all 5 trials (3, k) were estimated to establish thereliability for all kinematic and EMG data across trials forboth tasks. The ICCs were established by correlating thedependent measures across 5 trials for each subject and thenperforming a repeated-measures analysis of variance to testfor systematic differences between jumping trials for eachsubject [29]. Given the significant differences for height andweight between groups at baseline, analyses of covariance(ANCOVA) to adjust for these covariates were considered.Because height and weight were different between groups,regression analyses, exploring the relationship betweenheight and weight on dynamic knee valgus, were pursued,

Table 2. Intraclass correlation coefficients (3, k) for theaverage of 5 trials during both tasks

Task

Group

ACLICC (SEM)

ControlICC (SEM)

Double drop jumpDynamic knee valgus (cm) 0.98 (0.53) 0.95 (0.52)Normalized quadriceps (%) 0.95 (0.83) 0.99 (0.57)Normalized hamstrings (%) 0.97 (0.57) 0.99 (0.48)Co-contraction ratio 0.93 (4.33) 0.96 (3.72)Latency quadriceps (s) 0.95 (0.18) 0.80 (0.05)Latency hamstrings (s) 0.95 (0.17) 0.92 (0.04)

Single drop jumpDynamic knee valgus (cm) 0.83 (0.72) 0.86 (0.46)Normalized quadriceps (%) 0.96 (0.85) 0.97 (0.64)Normalized hamstrings (%) 0.98 (0.61) 0.98 (0.56)Co-contraction ratio 0.94 (4.20) 0.95 (4.53)Latency quadriceps (s) 0.89 (0.05) 0.95 (0.26)Latency hamstrings (s) 0.89 (0.06) 0.86 (0.43)

6 Ortiz et al BIOMECHANICAL DEFICIENCIES DURING DROP JUMPS

but showed no statistically significant association amongthese variables. Recent literature has shown that body massindex does not affect outcomes in patients after ACLreconstruction [30]. Therefore, we decided to keep theanalysis according to Portney and Watkins’ [31] criteriausing the baseline differences between groups as covariate toartificially equate both groups based on height and weight.All EMG variables were evaluated using a 1-way multivariateanalysis of variance (MANOVA). Separate 1-way MANOVAfor EMG variables was performed for each of the tasks.Univariate tests were conducted only if the MANOVAwas statistically significant. Significance was considered atP < .05. All analyses were performed using SPSS 16.0 (SPSSInc., Chicago, IL).

RESULTS

Two subjects from the control group were excluded fromthe SLJ analysis because of computer difficulties in kine-matic reconstruction, and 1 subject with ACLr wasexcluded because of her inability to stick the landing duringboth tasks. As a result, 14 women with SG-ACLr (height:167.71 � 9.0 cm, body mass: 67.68 � 11.66 kg) were usedfor analysis for both tasks, and 16 and 14 noninjuredwomen (height: 160.50 � 5.17 cm, body mass: 59.35� 10.37 kg) were used for analyses for the double- andsingle-legged drop jumps, respectively (Table 1). Thus, atotal of 30 and 28 subjects were included in the DLJ andSLJ, respectively. Kinematic and EMG amplitude datamet all assumptions of normality and homoscedasticity.However, latency variables showed statistically significantdifferences between their variances. Therefore, Kruskal-Wallis tests were conducted for this variable for both dropjumps. The reliability analyses for kinematic and EMGvariables showed that both tasks exhibited excellent reli-ability (ICC � 0.80) (Table 2).

Double-Leg Landing CountermovementJump

The adjusted means of the 1-way ANCOVA (using weightand height as covariates) showed that women with SG-ACLrexhibited a nonestatistically significant difference in dy-namic valgus compared to the noninjured group (Figure 3;F1,26 ¼ 3.50; P ¼ .07, effect size ¼ 0.12). One-wayMANOVA adjusted means showed significant differences in

Table 1. Anthropometric characteristics of the studyparticipants

AnthropometricMeasures ACL (n [ 14) Control (n [ 16) P

Age (y) 28.50 � 4.59 27.69 � 3.91 .61Body mass (kg) 67.68 � 11.66 59.35 � 10.37 .044Height (cm) 167.71 � 9.01 160.50 � 5.17 .009Activity Rating Scale 14.29 � 1.38 13.81 � 1.60 .39

EMG factors between groups (P < .001). Post-hoc analysesshowed lower quadriceps muscle activity (Figure 4; F1, 28 ¼7.81; P ¼ .01, effect size ¼ 0.22), and lower hamstringmuscle activity (Figure 4; F1,28 ¼ 6.08; P ¼ .02, effectsize ¼ 0.18) for the SG-ACLr group compared to the con-trol group. The co-contraction ratio results did not reachstatistical significance, but showed a tendency to be reducedin the SG-ACLr group compared to the control group(Figure 5). Latency analyses through the Kruskal-Wallis testrevealed no differences between groups for quadriceps(P ¼ .74) and hamstring (P ¼ .91) muscle groups(Figure 6).

Figure 3. Adjusted means for dynamic knee valgus betweengroups per task. Although the group of women with semite-ndinosus-gracilis anterior cruciate ligament reconstruction (SG-ACLr) exhibited greater dynamic knee valgus, the results werenot statistically significant.

Figure 4. Adjusted means for rectified-normalized EMG MAN-OVAs were statistically significant during both tasks. Follow-upunivariate analyses showed lower EMG activity for quadricepsand hamstrings muscle groups in the anterior cruciate ligamentreconstruction (ACLr) group during the double-legged dropjump. Conversely, the electromyographic activity of thequadriceps was greater in the semitendinosus-gracilis anteriorcruciate ligament reconstruction (SG-ACLr) group during thesingle-legged drop jump.

Figure 6. The group of women with semitendinosus-gracilisanterior cruciate ligament reconstruction (SG-ACLr) exhibitedsignificantly less time to reach maximal activation of thequadriceps and hamstrings during the single-legged dropjump.

PM&R Vol. -, Iss. -, 2014 7

Single-Leg Landing CountermovementJump

One-way ANCOVA (weight and height as covariates)showed that women with SG-ACLr exhibited a tendencytowards a greater dynamic knee valgus; however, this dif-ference was not statistically significant (Figure 3; F1,24 ¼2.60; P ¼ .12, effect size ¼ 0.60). The 1-way MANOVA for

Figure 5. The semitendinosus-gracilis anterior cruciate ligamentreconstruction (SG-ACLr) group exhibited lower co-contractionratios between quadriceps and hamstrings. However, statisti-cally significant differences were found only during the single-legged drop jump.

EMG amplitude factors was statistically significant (P ¼ .02).The univariate post-hoc analysis revealed a statistically sig-nificant greater quadriceps muscle activity in the ACLr groupcompared to that in the control group (Figure 4; F1,28 ¼5.60; P ¼ .03, effect size ¼ 0.17). Hamstring muscle activity(Figure 4; F1,28 ¼ 0.30; P ¼ .60, effect size ¼ 0.01) did notreach significance, but showed a tendency to be increased inthe SG-ACLr groups compared to the control group. Also,the SG-ACLr group exhibited a significantly lower co-contraction ratio (Figure 5; F1,28 ¼ 5.77; P ¼ .02, effectsize ¼ 0.17) compared to the control group. In terms oflatency, the Kruskal-Wallis test revealed that women withSG-ACLr exhibited significantly faster time-to-peak muscleactivation than noninjured women for both quadriceps(P ¼ .01) and hamstring (P ¼ .01) muscle groups (Figure 6).

DISCUSSION

Results for kinematic outcomes on this study suggest thatwomen with SG-ACLr exhibit altered dynamic knee valgusduring the double-legged (control: 4.18 cm; ACL: 6.24 cm)and single-legged (control: 1.99 cm; ACL: 3.07 cm) dropjumps when compared to noninjured individuals. Thesechanges did not reach statistical significance because of alack of statistical power; yet, the effect size detected warrantsconsideration. Neuromuscular recruitment results supportedin part our hypothesis showing significantly differentneuromuscular recruitment strategies for landing stability toinjured as compared to noninjured women. More specif-ically, during the DLJ, SG-ACLr participants showed lowerquadriceps (control: 16.87%; ACL: 11.9%) and hamstring(control: 9.93%; ACL: 6.16%) muscle activity. Conversely,results of the SLJ showed greater quadriceps muscle activity(control: 9.51%; ACL: 12.96%), lower co-contraction ratio

8 Ortiz et al BIOMECHANICAL DEFICIENCIES DURING DROP JUMPS

(control: 0.66; ACL: 0.50), and lower latency for bothquadriceps (control: 0.99 seconds; ACL: 0.22 seconds), andhamstring (control: 1.04 seconds; ACL: 0.24 seconds)muscles in the ACLr group, when compared to the controlgroup. In general, these results suggest different injury-pre-disposing neuromuscular recruitment strategies when higherlevels of performance of the reconstructed knee are required.

Neuromuscular control is the ability to produce andmodulate forces facilitating performance of any activity [5].Hence, some individuals with complete ACL tear are able todynamically stabilize their knee, even during high-impactactivities [1]. This indicates that the neuromuscular systemresponse dictates dynamic knee stability in the absence ofligamentous support [1,32]. On the contrary, there are someindividuals lacking dynamic knee stability after an ACLrupture. and therefore, these individuals are advised to un-dergo ligament reconstruction [1]. In addition, women withligamentous injuries exhibit strength deficiencies and timingdelays when compared to noninjured individuals [17].Impact forces at the knee without adequate neuromuscularcontrol will potentially provide less opportunity to accom-modate muscle contractions and attenuate forces [19],consequently leading to further knee injuries [2,12]. OurEMG findings during both drop jumps are, to some extent,congruent with the aforementioned reports.

The results for the DLJ agreed with our hypothesisand showed that participants with SG-ACLr exhibitedlower quadriceps and hamstring EMG amplitudes, but nosignificant differences in terms of co-contraction and latency,when compared to women without surgery. We proposethat these results represent a neuromuscular control deficitthat prohibits the generation of required forces for adequateknee stability. As seen in the co-contraction ratio, this dif-ference was not statistically significant between groups(control: 0.58; ACLr: 0.50), giving the impression that par-ticipants with ACLr have similar anterior-posterior kneestability when compared to the control group. However,both normalized EMG values within the co-contractionformula in the ACLr group (quadriceps: 11.9%; hamstrings:6.16%) were significantly lower when compared to thecontrol group (quadriceps: 17%; hamstrings: 10%). Thisimplies a false impression of knee joint stability, as suchmuscle activities might not protect against injury due to theinability of the quadriceps and hamstrings to tolerate largeloads. The tendency of a slight increase in the dynamic kneevalgus observed in individuals with SG-ACLr corroboratesthe concept of decreased dynamic knee stability. However,the nature of the task (DLJ) allows individuals to use a va-riety of landing strategies, including an increased use ofthe nonreconstructed lower extremity to control landing.Therefore, the obtained results may not represent theneuromuscular strategies used during high-level impact ac-tivities in the extremity with the ACLr. However, the ob-tained results for the SLJ may provide clearer results towarddynamic knee stability during landing in these individuals.

The magnitude of the effect size (0.60) may indicate clinicalsignificance, although the differences in dynamic valgusbetween the ACL and control groups during the SLJ werenot statistically significant. Therefore, although women withSG-ACLr exhibit greater dynamic knee valgus that couldpredispose them to re-injury, a greater sample size would beneeded to reach such a conclusion.

Results from the SLJ supported in part our hypothesisshowing evidence of lower co-contraction ratio that couldpredispose women with SG-ACLr to further injury. On theother hand, the SG-ACLr group presented significantlygreater quadriceps muscle activity and less time to reachmaximum activation in comparison to the control group,contradicting our initial hypothesis. These findings differfrom a previous investigation in which women demonstratedan increased quadriceps/hamstrings latency period (timedelays) and decreased quadriceps/hamstrings muscle acti-vation [17]. We hypothesize that results of the SLJ areevidence of compensatory mechanisms in which loweractivation time allows muscle forces to be generated at theappropriate time to be successful at the required task. Bryantet al (2009) showed that patients with ACL reconstruc-tion with semitendinosus-gracilis graft, who were successfulin reaching high levels of function, exhibited earlierhamstring and quadriceps muscle activity before initialcontact during single-legged hops when compared tomatched noninjured patients [33]. They hypothesizedthat this early muscle activation is part of a compensatoryfeed-forward mechanism, with the purpose of increasingjoint compression to prevent instability during the landingphase of a jump. However, the increased quadriceps acti-vation without hamstring muscle counteraction could pre-dispose the individual to further injuries by promotinganterior translation of the tibia [34,35]. If the hamstrings donot offset with sufficient force the counter force generated bythe quadriceps, the ACL might not be able to prevent, byitself, anterior tibial translation. A hamstring/quadriceps co-contraction ratio of less than 60% can be indicative ofneuromuscular deficiencies that may predispose athletes toACL injury [34]. In this study, the co-contraction ratio re-sults supported our hypothesis that SG-ACLr patients wouldexhibit a greater injury-predisposing neuromuscular profileduring the SLJ. The SG-ACLr group co-contraction ratio was50%, showing a higher probability of injuries in this pop-ulation. More research in this area is needed to elucidateneuromuscular mechanisms developed by these patientsafter ACL reconstruction.

The double-legged dynamic knee valgus values observedin our investigation (DLJ: control: 4.17 cm; ACL: 6.24 cm;P ¼ .07.) were similar to values previously reported by otherinvestigators [36,37], who evaluated knee injury predispo-sition, finding greater risk for knee injury in women whencompared to their male counterparts. Previous investigatorshave reported that double-legged landing dynamic kneevalgus measures of 4.15 cm [37] and 5 cm [36] predispose

PM&R Vol. -, Iss. -, 2014 9

females to injuries when compared to their male counter-parts. Therefore, we hypothesize that both groups exhibitedlanding mechanics that may predispose them to knee in-juries, given that our dynamic knee valgus values duringthe double-legged drop jump were 4.17 � 2.32 cm and 6.24� 3.71 cm for the control and SG-ACLr groups, respectively.

Study Limitations

Although the purpose of the valgus analysis in this study wasmore clinically oriented, it had its own inherent limitations.Our valgus measure assessed the displacement of the kneesfrom the moment of initial contact to maximum valgus, andtherefore was highly affected by the landing alignment.Hence, participants already landing in valgus would haveless available motion toward more valgus than those landingwith minimal or no valgus. An additional possible limitationarises from the rehabilitation protocol. Even though reha-bilitation programs were the same per orthopedic surgeonand all individuals were cleared to return to their regularactivities after their rehabilitation protocol was completed,this variable might have had an effect on individualbiomechanical responses. Rehabilitation programs comprisestandard ACL interventions such as inflammation manage-ment, range of motion, progression toward neuromuscularre-education, and strengthening interventions. Individualswith ACLr who perform neuromuscular training as part oftheir rehabilitation protocol develop reactive muscle activa-tion strategies that may protect the knee joint from excessiveloads [17]. Future studies should examine the correlationbetween rehabilitation protocols and neuromuscular strate-gies presented during high-impact landing activities, as wellas the type of ACL reconstruction. Our study includedwomen who had undergone SG-ACLr 12 months to 5 yearspreviously. Alterations in muscular performance and levelof physical activity change according to time of injurymight have an effect on functional parameters and perfor-mance of various tasks [21,38]. Therefore, future studiesshould include a larger sample size and should limitthe range of time after surgery to control for this potentialbias. Finally, the lack of matching individuals for dominantlimb, and assessing biomechanics only in 1 leg in bothgroups, could have left out possible valuable informationthat may have helped further to explain the differencesfound.

CONCLUSIONS

Individuals who have undergone ACLr have shown impairedmuscle activity and timing adaptations. These neuromus-cular deficits might affect dynamic knee stability, andconsequently affect performance of high-impact sportingactivities. Dynamic knee valgus has been proposed as ameasure for knee stability during dynamic tasks. In thecurrent study, we observed that individuals who had

undergone ACLr presented with appropriate dynamic kneevalgus, although neuromuscular deficits were also observed.These neuromuscular deficits may predispose these in-dividuals to re-injury. Moreover, neuromuscular deficits aremore noticeable when increased use of the reconstructedknee is required, as would be required during single-leggedhigh-demand activities. Therefore, neuromuscular assess-ments after ACLr should include tests that are more sensitivefor neuromuscular control during sporting activities.Consequently, neuromuscular assessments may have thepotential to help guide the rehabilitation and/or trainingprograms of these individuals by determining the specificdeficits that require attention, to return to sporting activities,and to reduce the risk of re-injury.

ACKNOWLEDGMENTS

The authors acknowledge Luis Rivas, Elena Roman, andGloria Colon for their work as research assistants on thisproject.

REFERENCES1. Moksnes H, Snyder-Mackler L, Risberg MA. Individuals with an

anterior cruciate ligament-deficient knee classified as noncopers may becandidates for nonsurgical rehabilitation. J Orthop Sports Phys Ther2008;38:586-595.

2. Rudroff T. Functional capability is enhanced with semitendinosus thanpatellar tendon ACL repair. Med Sci Sports Exerc 2003;35:1486-1492.

3. Hospodar SJ, Miller MD. Controversies in ACL reconstruction: Bone-patellar tendon-bone anterior cruciate ligament reconstruction remainsthe gold standard. Sports Med Arthrosc 2009;17:242-246.

4. Aune AK, Holm I, Risberg MA, Jensen HK, Steen H. Four-strandhamstring tendon autograft compared with patellar tendon-boneautograft for anterior cruciate ligament reconstruction: A randomizedstudy with two-year follow-up. Am J Sports Med 2001;29:722-728.

5. Williams GN, Snyder-Mackler L, Barrance PJ, Axe MJ, Buchanan TS.Neuromuscular function after anterior cruciate ligament reconstructionwith autologous semitendinosus-gracilis graft. J Electromyogr Kinesiol2005;15:170-180.

6. Ortiz A, Olson S, Libby CL, et al. Landing mechanics between non-injured women and women with anterior cruciate ligament recon-struction during 2 jump tasks. Am J Sports Med 2008;36:149-157.

7. Ortiz A, Olson S, Trudelle-Jackson E, Rosario M, Venegas HL. Landingmechanics during side hopping and crossover hopping maneuvers innoninjured women and women with anterior cruciate ligamentreconstruction. PM R 2011;3:13-20.

8. Grindstaff TL, Jackson KR, Garrison JC, Diduch DR, Ingersoll CD.Decreased quadriceps activation measured hours prior to a noncontactanterior cruciate ligament tear. J Orthop Sports Phys Ther 2008;38:508-516.

9. Mizner RL, Kawaguchi JK, Chmielewski TL. Muscle strength in thelower extremity does not predict postinstruction improvements in thelanding patterns of female athletes. J Orthop Sports Phys Ther 2008;38:353-361.

10. Earl JE, Monteiro SK, Snyder KR. Differences in lower extremity ki-nematics between a bilateral drop-vertical jump and a single-leg step-down. J Orthop Sports Phys Ther 2007;37:245-252.

11. Fitzgerald GK, Lephart SM, Hwang JH, Wainner RS. Hop tests aspredictors of dynamic knee stability. J Orthop Sports Phys Ther 2001;31:588-597.

10 Ortiz et al BIOMECHANICAL DEFICIENCIES DURING DROP JUMPS

12. Huston LJ, Vibert B, Ashton-Miller JA, Wojtys EM. Gender differencesin knee angle when landing from a drop-jump. Am J Knee Surg 2001;14:215-219; discussion 219-220.

13. Williams GN, Chmielewski T, Rudolph K, Buchanan TS, Snyder-Mackler L. Dynamic knee stability: Current theory and implications forclinicians and scientists. J Orthop Sports Phys Ther 2001;31:546-566.

14. Powers CM. The influence of altered lower-extremity kinematics onpatellofemoral joint dysfunction: A theoretical perspective. J OrthopSports Phys Ther 2003;33:639-646.

15. Zazulak BT, Ponce PL, Straub SJ, Medvecky MJ, Avedisian L,Hewett TE. Gender comparison of hip muscle activity during single-leglanding. J Orthop Sports Phys Ther 2005;35:292-299.

16. Huston LJ, Greenfield ML, Wojtys EM. Anterior cruciate ligament in-juries in the female athlete: Potential risk factors. Clin Orthop Relat Res.2000:50-63.

17. Hewett TE, Zazulak BT, Myer GD, Ford KR. A review of electromyo-graphic activation levels, timing differences, and increased anteriorcruciate ligament injury incidence in female athletes. Br J Sports Med2005;39:347-350.

18. Decker MJ, Torry MR, Noonan TJ, Riviere A, Sterett WI. Landing ad-aptations after ACL reconstruction. Med Sci Sports Exerc 2002;34:1408-1413.

19. Lephart SM, Ferris CM, Riemann BL, Myers JB, Fu FH. Gender dif-ferences in strength and lower extremity kinematics during landing.Clin Orthop Relat Res. 2002:162-169.

20. Vairo GL, Myers JB, Sell TC, Fu FH, Harner CD, Lephart SM.Neuromuscular and biomechanical landing performance subsequentto ipsilateral semitendinosus and gracilis autograft anterior cruciateligament reconstruction. Knee Surg Sports Traumatol Arthrosc 2008;16:2-14.

21. Roe J, Pinczewski LA, Russell VJ, Salmon LJ, Kawamata T, Chew MA.7-year follow-up of patellar tendon and hamstring tendon grafts forarthroscopic anterior cruciate ligament reconstruction: Differences andsimilarities. Am J Sports Med 2005;33:1337-1345.

22. Marx RG, Stump TJ, Jones EC, Wickiewicz TL, Warren RF. Develop-ment and evaluation of an activity rating scale for disorders of the knee.Am J Sports Med 2001;29:213-218.

23. van der Harst JJ, Gokeler A, Hof AL. Leg kinematics and kinetics inlanding from a single-leg hop for distance: A comparison betweendominant and non-dominant leg. Clin Biomech (Bristol, Avon) 2007;22:674-680.

24. Ford KR, Myer GD, Smith RL, Byrnes RN, Dopirak SE, Hewett TE. Useof an overhead goal alters vertical jump performance and biome-chanics. J Strength Cond Res 2005;19:394-399.

25. Besier TF, Lloyd DG, Ackland TR. Muscle activation strategies at theknee during running and cutting maneuvers. Med Sci Sports Exerc2003;35:119-127.

26. Croce RV, Russell PJ, Swartz EE, Decoster LC. Knee muscular responsestrategies differ by developmental level but not gender during jumplanding. Electromyogr Clin Neurophysiol 2004;44:339-348.

27. Soderberg GL, Knutson LM. A guide for use and interpretation ofkinesiologic electromyographic data. Phys Ther 2000;80:485-498.

28. Lloyd DG, Buchanan TS. Strategies of muscular support of varusand valgus isometric loads at the human knee. J Biomech 2001;34:1257-1267.

29. Kang HG, Dingwell JB. Intra-session reliability of local dynamic stabilityof walking. Gait Posture 2006;24:386-390.

30. Ballal MS, Khan Y, Hastie G, Hatcher A, Coogan S, McNicholas MJ.Functional outcome of primary hamstring anterior cruciate ligamentreconstruction in patients with different body mass index classes.Arthroscopy 2013;29:1314-1321.

31. Portney LG, Watkins MP. Foundations of Clinical Research: Applica-tions to Practice. 3rd ed. Upper Saddle River, NJ: Pearson/Prentice Hall;2009.

32. Hurd WJ, Axe MJ, Snyder-Mackler L. Influence of age, gender, andinjury mechanism on the development of dynamic knee stability afteracute ACL rupture. J Orthop Sports Phys Ther 2008;38:36-41.

33. Bryant AL, Newton RU, Steele J. Successful feed-forward strategiesfollowing ACL injury and reconstruction. J Electromyogr Kinesiol2009;19:988-997.

34. Hewett TE. Neuromuscular and hormonal factors associated with kneeinjuries in female athletes: Strategies for intervention. Sports Med 2000;29:313-327.

35. Swanik CB, Lephart SM, Giraldo JL, Demont RG, Fu FH. Reactivemuscle firing of anterior cruciate ligament-injured females duringfunctional activities. J Athl Train 1999;34:121-129.

36. Noyes FR, Barber-Westin SD, Fleckenstein C, Walsh C, West J. Thedrop-jump screening test: Difference in lower limb control by genderand effect of neuromuscular training in female athletes. Am J SportsMed 2005;33:197-207.

37. Myer GD, Ford KR, Khoury J, Succop P, Hewett TE. Development andvalidation of a clinic-based prediction tool to identify female athletes athigh risk for anterior cruciate ligament injury. Am J Sports Med 2010;38:2025-2033.

38. Risberg MA, Holm I, Myklebust G, Engebretsen L. Neuromusculartraining versus strength training during first 6 months after anteriorcruciate ligament reconstruction: A randomized clinical trial. Phys Ther2007;87:737-750.