New method to identify athletes at high risk of ACL injury ... · dates for targeted injury prevention training. The purpose of this report is to demonstrate techniques to accurately

Original article

Br J Sports Med 2010;XX:XXX–XXX. doi:10.1136/bjsm.2010.072843 1

1Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio, USA2Sports Medicine Biodynamics Center and Human Performance Laboratory, Cincinnati, Ohio, USA3Rocky Mountain University of Health Professions, Provo, Utah, USA4Department of Pediatrics, College of Medicine, University of Cincinnati, Cincinnati, Ohio, USA5Departments of Orthopaedic Surgery, Biomedical Engineering and Rehabilitation Sciences, University of Cincinnati, Cincinnati, Ohio, USA

Correspondence to Dr Gregory D Myer, Cincinnati Children’s Hospital, 3333 Burnet Avenue, MLC 10001, Cincinnati, OH 45229, USA; [email protected]

Accepted 10 June 2010

2

New method to identify athletes at high risk of ACL injury using clinic-based measurements and freeware computer analysisGregory D Myer,1–3 Kevin R Ford,1,2,4 Timothy E Hewett1,2,4,5

ABSTRACTBackground High knee abduction moment (KAM)

landing mechanics, measured in the biomechanics

laboratory, can successfully identify female athletes at

increased risk for anterior cruciate ligament (ACL) injury.

Methods The authors validated a simpler, clinic-based

ACL injury prediction algorithm to identify female ath-

letes with high KAM measures. The validated ACL injury

prediction algorithm employs the clinically obtainable

measures of knee valgus motion, knee fl exion range

of motion, body mass, tibia length and quadriceps-to-

hamstrings ratio. It predicts high KAMs in female ath-

letes with high sensitivity (77%) and specifi city (71%).

Conclusion This report outlines the technique for

this ACL injury prediction algorithm using clinic-based

measurements and computer analyses that require only

freely available public domain software.

INTRODUCTIONProspective measures of high knee abduction moment (KAM) during landing predict ante-rior cruciate ligament (ACL) injury risk in young female athletes.1 Using data from nearly 700 young women, we developed a clinic-based assessment algorithm to identify those with increased KAM (fi gure 1).2–4 These women would be ideal candi-dates for targeted injury prevention training.

The purpose of this report is to demonstrate techniques to accurately capture and analyse measures of body mass, tibia length, quadriceps-to-hamstrings ratio (QuadHam), knee valgus motion and knee fl exion range of motion (ROM) for use of the ACL injury prediction algorithm using clinic-based measurements and computer analyses with freely available public domain software.

METHODSClinic based anthropometrics and strengthTibia length and body massClinic-based tibia length was measured with the subject standing with knees extended in anatomi-cal position. A standard measuring tape was used to measure the distance between the lateral knee joint line and the lateral malleous (fi gure 2). Body mass was measured on a calibrated physician scale.

QuadHam ratioIsokinetic knee extension/fl exion (concentric/concentric muscle action) strength was measured on a standard isokinetic testing device for each leg at 300°/s.5 The quadriceps-to-hamstrings

(QuadHam) ratio was calculated as the ratio of quadriceps to hamstrings peak isokinetic torque. Some clinical settings may not have an isokinetic testing device readily available. In this case, a surrogate measure of the QuadHam ratio can be employed that was defi ned using a linear regres-sion analysis to predict QuadHam ratio based on the athlete’s body mass. The surrogate QuadHam ratio measure can be obtained by multiplying the female athlete’s mass to 0.01 and adding the resultant value to 1.10. If even greater simplicity is desired, the mean value of 1.53 can be substituted into the prediction algorithm for the QuadHam ratio.4 Although surrogate calculations provide a range that can be used if isokinetic measure-ment is not obtainable, the prediction model is best optimised with QuadHam ratios which are achieved by direct measurement.

Clinic-based landing biomechanicsCamera set-upTwo-dimensional (clinic-based) frontal and sag-ittal plane knee kinematic data were captured with standard video cameras. Cameras should be levelled and positioned at a height of 60–80 cm, perpendicular to each other in the frontal and sagittal planes. To reduce measurement perspec-tive error, it is optimal to move the camera as far away from the capture area as possible, so as to allow for a fully zoomed view to be focused on the desired capture location.6 Once the camcord-ers were positioned, the view was focused with the manual settings (autofocus settings were not used).

Test instructionsThe box (31 cm high) used for the drop verti-cal jump (DVJ) should be centred on the frontal camera view and approximately 30 cm off cen-tre of sagittal plane view away from the camera in the frontal plane position. Prior to the DVJ test performance, it is recommended that a stan-dardised target be placed overhead, equal to the subject’s maximal touch height during a coun-termovement vertical jump. The subject should be instructed to ‘stand on top of a box with their feet positioned 35 cm apart. Once prepared, the subject should drop directly down off the box and immediately perform a maximum vertical jump, raising both arms towards the overhead target.’ If an overhead target is not used, the test subject should be instructed to ‘land and jump as high as possible as if reaching for a basket-ball rebound.’7 Subjects should be allowed one

3

4

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364

6566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128

1

bjsports72843.indd 1bjsports72843.indd 1 6/25/2010 8:10:37 PM6/25/2010 8:10:37 PM

Original article

Br J Sports Med 2010;XX:XXX–XXX. doi:10.1136/bjsm.2010.0728432

to three practice trials to ensure they are able to demonstrate adequate understanding and ability to perform the instructed test manoeuvre.

Image captureLanding sequence images can be captured via the ‘print screen’ feature available on most personal computers, or they can be captured with freeware software such as VirtualDub software (copyright 1998–2009 Avery Lee). In addition, typi-cal software programs (ie, VirtualDub) can deinterlace the video fi elds to effectively double the frames available per second from 30 to 60 Hz for NTSC systems (or 25–50 Hz for PAL systems). For simplicity, the four image fi les should

be captured in the following suggested order: I1-frontal plane view with frame prior to initial contact, I2-frontal plane view of frame with knee in maximum medial (valgus) posi-tion, I3-sagittal plane view with frame prior to initial con-tact, I4-sagittal plane view of frame with knee in maximum fl exion position and named in a standard structure for each subject (ie, SubjectX I1, SubjectX I2, SubjectX I3, SubjectX I4). Recommended software for kinematic coordinate data cap-ture is suggested for use with ImageJ (Rasband W S, ImageJ, US National Institutes of Health, Bethesda, Maryland, USA, http://rsb.info.nih.gov/ij/, 1997–2009), software that is also available without surcharge.

Knee valgus motionImageJ software allows for easy importing of image sequences of fi les with a standard name structure (ie, SubjectX I1, SubjectX I2, SubjectX I3, SubjectX I4). Once the image sequence is imported, a video scaling factor (scaling factor=known dis-tance (cm)/digitised distance (video units)) is established for the x-axis scale from a known distance (drawing a line equal to the length of the known distance and inputting the actual length such as force plate width measurement presented in fi gure 3). Subsequent coordinate data in the x-axis are repre-sented in centimetres of motion based on the scaling factor for measurement of knee valgus motion (fi gure 4). The calibrated displacement measure between the two digitised knee coordi-nates (X2−X1) is representative of knee valgus motion during the DVJ.

Knee fl exion ROMThe sagittal plane video camcorder is used to capture knee fl exion angles that are calculated from the video frame just prior to initial contact and the video frame at maximum knee fl exion. Knee fl exion ROM was calculated as the difference in knee fl exion between the two positions (Θ1−Θ2; fi gure 5).

APPLICATION OF THE PREDICTION ALGORITHMTo use the prediction nomogram (fi gure 1), one should place a straight edge vertically so that it touches the designated vari-able on the axis for each predictor value, and record the value

Figure 1 Clinician friendly nomogram that was developed from the regression analysis and can be used to predict high knee abduction moment outcome based on tibia length, knee valgus motion, knee fl exion range of motion, body mass and quadriceps-to-hamstrings ratio.

Figure 2 Tibia length is measured as the distance between knee joint centre and ankle joint centre (Z2–Z1).

Figure 3 Example of the calibration process for measured distance in the x-axis using ImageJ software. In the presented example, the line tool is used to draw a line the width of the box. Then, by using the set scale procedure, the known distance of the box width is input. From this calibration step, all length measurements from this camera position will be calibrated for the subjects’ drop vertical jump trials.

129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192

193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256


Original article


Figure 4 (A) Coordinate position of knee joint centre digitised in the frontal view measured at the frame prior to initial contact used as the knee valgus position X1. (B) Coordinate position of knee joint centre digitised in the frontal view measured at the frame with maximum medial position and utilised as the knee valgus position X2. (C) Calibrated displacement measure between the two digitised knee coordinates (X2–X1), representative of knee valgus motion during the drop vertical jump.

Figure 5 (A) Knee fl exion angle digitised at the frame prior to initial contact and recorded as the fi rst measure of knee fl exion range of motion (ROM) (Θ1). (B) Knee fl exion angle, digitised at the frame with maximum knee fl exion and recorded as the second measure of knee fl exion ROM (Θ2). (C) Displacement of knee fl exion, calculated as the differences in knee fl exion angles at the frame prior to initial contact and maximum knee fl exion (Θ1–Θ2) and representative of knee fl exion ROM.

257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320

321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384


Original article


that each of the fi ve predictors provides on the ‘points’ axis at the top of the diagram. All of the recorded ‘points’ measured using this method are then summed, and this value is located on the ‘total points’ line with a straight edge. A vertical line drawn down from the ‘total points line’ to the ‘probability line’ identifi es the probability that the athlete will demon-strate a high KAM (>21.74 Nm of knee abduction) during the DVJ based on the utilised predictive variables.

Identifi cation of high KAMFigure 6 provides an example of predicted probability for high KAM status based on the representative subject’s clinic-based measurements and landing mechanics (tibia length 35 cm; knee valgus motion 8.2 cm; knee fl exion ROM 75.4°; body mass 52.2 kg; QuadHam 1.55). Based on her demonstrated measure-ments, this subject would have a 74% (95.5 points) chance of demonstrating a high KAM during the DVJ. The actual KAM measurement for the presented DVJ that was quantifi ed simul-taneously with 3D motion analysis was 24.2 Nm of knee abduction load.

Identifi cation of low KAMFigure 7 presents the optimal landing biomechanics during the DVJ in which the athlete demonstrated desirable knee fl exion ROM without any knee valgus motion that ultimately limited her potential to demonstrate high KAM. Based on her recorded clinic-based measurement, the presented subject demonstrated a 25% chance for high KAM using the proposed prediction algorithm. Simultaneous 3D motion analysis confi rmed the accuracy of the proposed algorithm, as the subject yielded an actual KAM measurement of 7.6 Nm.

DISCUSSIONSpecial consideration for use of ACL injury prediction algorithmClinicians who perform risk assessment should be cognisant that side-to-side imbalances in neuromuscular strength,

fl exibility and coordination can be important predictors of increased injury risk.1 8 9 However, the above examples of KAM prediction have utilised measures on the left limb. Specifi c to ACL injury risk prediction, leg-to-leg differences in KAM were observed in injured, but not uninjured females. Importantly, the side-to-side KAM difference was 6.4-fold greater in ACL-injured versus uninjured females. Female athletes tend to dem-onstrate side-to-side differences in visually evident maximum knee valgus angle during a box DVJ (fi gure 8).7

Figure 9 presents a representative subject with side-to-side differences in landing biomechanics which the ACL injury prediction algorithm also delineates with side–side differ-ences in prediction of risk for high KAM. Based on her left leg (fi gure 9A,B) she would have a 28% (73.5 points; fi gure 9E) chance of demonstrating a high KAM during the DVJ. Her actual KAM measure for the presented DVJ that was quanti-fi ed simultaneously with 3D motion analysis was 15.4 Nm of knee abduction load on her left leg. However, when using the measures from the right leg frontal plane motion (fi gure 9C,D; 15 cm) in the ACL injury-risk prediction algorithm, this sub-ject would have a 97% (129.5 points; fi gure 9F) chance of demonstrating a high KAM during the DVJ. Her actual KAM measure for the presented DVJ that was quantifi ed simul-taneously with 3D motion analysis was 29.2 Nm of knee abduction load on her right leg. This important observation indicates that the ACL injury prediction algorithm is both sensitive and specifi c to high KAM, even between limbs in a single subject. Accordingly, clinicians should evaluate side-to-side differences in frontal plane mechanics, and the larg-est knee valgus motion measurements should be employed to maximise the utility of the proposed ACL injury prediction algorithm.1

CONCLUSIONSThe current methodological report facilitates the next criti-cal step to bridge the gap between laboratory identifi cation of injury risk factors1 10 and clinical practices achieved in the

5

6

Figure 6 Example of predicted probability for high knee abduction moment (KAM) status based on the representative subject’s (fi gures 4, 5) clinic-based measurements. Completed nomogram for the representative subject (tibia length: 35 cm; knee valgus motion: 8.2 cm; knee fl exion range of motion: 75.4 body mass: 52.2 kg; QuadHam: 1.55). Based on her demonstrated measurements, this subject would have a 74% (95.5 points) chance of demonstrating a high KAM during the drop vertical jump (DVJ). The actual KAM measurement for the presented DVJ that was quantifi ed simultaneously with 3D motion analysis was 24.2 Nm of knee abduction load.

385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448

449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512


Original article


fi eld.3 4 The algorithm may be used as a training camp proto-col in partnership with team clinicians. It could also be set up and run in the athletic training setting. This screen may help identify those at high risk of sustaining an ACL injury during

competitive play. Implementation of the prediction algorithm may increase both the effi cacy and effi ciency of future inter-ventions aimed at preventing non-contact ACL injury in female athletes. Future research is warranted to determine

Figure 7 (A–B) Calibrated displacement measure between the two marked knee coordinates (X2–X1) representative of knee valgus motion during the drop vertical jump (DVJ). (C–D) Displacement of knee fl exion calculated as the differences in knee fl exion angles at the frame prior to initial contact and maximum knee fl exion (Θ1–Θ2) and representative of knee fl exion range of motion (ROM). (E) Completed nomogram for the representative subject (tibia length: 38 cm; knee valgus motion: 0.0 cm; knee fl exion ROM: 75.3°; mass: 44.9 kg; QuadHam: 1.89). Based on her demonstrated measurements, this subject would have a 25% (73 points) chance of demonstrating high knee abduction moment (KAM) during the DVJ. Her actual KAM measure for the presented DVJ that was quantifi ed simultaneously with 3D motion analysis was 7.6 Nm of knee abduction load.

E

Figure 8 Visually evident side-to-side differences in frontal plane landing mechanics.

513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576

577578579580581582583584585586587588589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640


Original article


Figure 9 Example of a representative subject with side-to-side differences in landing biomechanics which the anterior cruciate ligament (ACL) injury prediction algorithm also delineates as differences in prediction of risk for high knee abduction moment (KAM). (A–B) Subject demonstrating 1 cm of knee valgus motion (X2–X1) on her left limb during the drop vertical jump (DVJ). (C–D) Subject demonstrating 15 cm of knee valgus motion (X2–X1) on her right limb during the same DVJ trial. (E) Completed nomogram for the representative subject (tibia length 36 cm; knee valgus motion 1.0 cm; knee fl exion range of motion 60.5°; mass 49.0 kg; QuadHam 2.0). Based on her demonstrated measurements in her left leg, this subject would have a 28% (73.5 points) chance of demonstrating a high KAM during the DVJ. Her actual KAM measure for the presented DVJ that was quantifi ed simultaneously with 3D motion analysis was 15.4 Nm of knee abduction load. (F) When using the measures from the right leg frontal plane motion (15 cm) in the ACL injury-risk prediction algorithm, the subject also shows side-to-side differences in prediction of high KAM. Based on her demonstrated measurements in her right leg, this subject would have a 97% (129.5 points) chance of demonstrating a high KAM during the DVJ. Her actual KAM measure for the presented DVJ that was quantifi ed simultaneously with 3D motion analysis was 29.2 Nm of knee abduction load.

E

F

641642643644645646647648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704

705706707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765766767768


Original article


if the proposed algorithm improves the effectiveness of neu-romuscular training methods (reduced KAM and ACL injury incidence) when targeted at women and girls who demon-strate high KAM landing mechanics. In addition, researchers should strive to simplify ACL injury risk-factor assessments while maintaining the high level of accuracy demonstrated

by the presented methods to further improve the potential prophylactic effects of interventions applied to high-injury-risk populations.

Acknowledgements The authors would like to thank Boone County School District, Kentucky, especially School Superintendent R Poe, for participation in this study. The authors would also like to thank M Blevins, E Massey and the athletes and coaches of Boone County public school district for their participation in this study. All authors are independent of any commercial funder, had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Funding The authors would like to acknowledge funding support from National Institutes of Health Grant R01-AR049735, R01-AR055563 and R01-AR056259. The Cincinnati Children’s Hospital Medical Center and Rocky Mountain University of Health Professions Institutional Review Boards approved this study.

Competing interests None.

Patient consent Obtained.

Ethics approval Ethics approval was provided by the Cincinnati Children’s Hosptial and Rocky Mountain University.

Provenance and peer review Not commissioned; externally peer reviewed.

REFERENCES 1. Hewett TE, Myer GD, Ford KR, et al. Biomechanical measures of

neuromuscular control and valgus loading of the knee predict anterior cruciate

ligament injury risk in female athletes: a prospective study. Am J Sports Med

2005;33:492–501.

2. Myer GD, Ford KR, Khoury J, et al. Biomechanics laboratory-based prediction

algorithm to identify female athletes with high knee loads that increase risk of

ACL injury. Br J Sports Med 2010;(In Press).

3. Myer GD, Ford KR, Khoury J, et al. Development and validation of a clinic based

prediction tool to identify high ACL injury risk female athletes. Am J Sports Med

2010;(In Press).

4. Myer GD, Ford KR, Khoury J, et al. Clinical Correlates to laboratory based mea-

sures for use in ACL injury risk prediction algorithm. Clin Biomech 2010;(In Press).

5. Myer GD, Ford KR, Barber Foss KD, et al. The relationship of hamstrings and

quadriceps strength to anterior cruciate ligament injury in female athletes.

Clin J Sport Med 2009;19:3–8.

6. Paul JA, Douwes M. Two-dimensional photographic posture recording and

description: a validity study. Appl Ergon 1993;24:83–90.

7. Ford KR, Myer GD, Hewett TE. Valgus knee motion during landing in high school

female and male basketball players. Med Sci Sports Exerc 2003;35:1745–50.

8. Knapik JJ, Bauman CL, Jones BH, et al. Preseason strength and fl exibility

imbalances associated with athletic injuries in female collegiate athletes. Am J

Sports Med 1991;19:76–81.

9. Baumhauer JF, Alosa DM, Renström AF, et al. A prospective study of ankle injury

risk factors. Am J Sports Med 1995;23:564–70.

10. Padua DA, Marshall SW, Beutler AI, et al. Prospective Cohort Study of

Biomechanical Risk Factors of ACL Injury: The JUMP-ACL Study. American

Orthopaedic Society of Sports Medicine Annual Meeting. Keystone, Colorado,

2009:393–5.

8

9

10

What is already known on this topic

▶ Female athletes who demonstrate high knee abduction moments (KAM) during landing increase their risk of sus-taining an anterior cruciate ligament (ACL) injury during competitive play.

▶ Female athletes who demonstrate an increased KAM are more likely to benefi t from neuromuscular training that will reduce their risk of ACL injury.

▶ Screening for KAM requires a dedicated biomechanical labo-ratory and costly measurement tools and labour-intensive data-collection sessions. These factors limit the potential to perform athlete risk assessments on a large scale, precluding the opportunity to target high-injury-risk athletes with the appropriate intervention strategies.

What this study adds

▶ Clinically feasible measures of increased knee valgus motion, knee fl exion range of motion, body mass, tibia length and quadriceps-to-hamstring ratio predict an increased propensity to demonstrate high KAM during landing in female athletes with high sensitivity and specifi city.

▶ The current report provides a systematic methodology to simplify and accurately identify female athletes who demonstrate high KAM landing mechanics that increase their risk of ACL injury.

▶ The presented algorithm provides the next critical step to bridge the gap between laboratory identifi cation of increased injury risk and clinical practices aimed to prevent the long-term sequelae that follows ACL injury.

769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824825826827828829830831832

833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883884885886887888889890891892893894895896

7


Documents

New method to identify athletes at high risk of ACL injury ... · dates for targeted injury prevention training. The purpose of this report is to demonstrate techniques to accurately