Systematic Review

SBSDW

a

Pfi

a

9

Initial Graft Tension and the Effect on Postoperative PatientFunctional Outcomes in Anterior Cruciate Ligament

ReconstructionGarry W. Kirwan, M.Phty., Michael G. Bourke, Ph.D., B.Phty.,

Lucinda Chipchase, Ph.D., B.Phty.,Philip A. Dalton, M.B.B.S., M.Med.Sc., B.Phty., F.R.A.C.S., C.I.M.E., and

Trevor G. Russell, Ph.D., B.Phty.

Purpose: The aim of this review was to investigate the effect of initial graft tension on patient-specific functionaloutcomes after anterior cruciate ligament reconstruction and determine whether a particular tension is associated withsuperior functional outcome. Methods: We performed a systematic review of prospective randomized trials witha National Health and Medical Research Council Australia level of evidence of III or higher published between 1950 andJuly 2012. Studies using a semitendinosus-gracilis or boneepatellar tendonebone autograft that reported graft tensionand postoperative functional outcomes were included. Quantitative analysis was performed on available data by calcu-lating effect size (ES) both at various time points and across tensions (in Newtons). Results: Initial search strategiesreturned 457 original publications, of which 5 articles fulfilled all exclusion and inclusion criteria. The mean score forquality was 5.8 (SD, 1.3), with 12 being the highest possible score. When compared with the preoperative side-to-sidedifference in anterior tibial displacement, 80 N and 78.9 N of tension recorded the largest effect at 2 weeks or less(ES, �2.98 [range, �3.82 to �2.14]) and 12 months or more (ES, �2.45 [range, �3.40 to �1.51]) postoperatively,respectively. When we compared tensions, the largest effect was toward 80 N when compared with 20 N at 2 weeks or lessafter surgery (ES, 0.76 [range, 0.17 to 1.35]). Conclusions: The objective of this review was to systematically assess theliterature to determine whether a particular initial graft tension results in superior outcomes after anterior cruciateligament reconstruction. From the review, there is a trend toward an initial graft tension of 78.5 to 90 N resulting ina reduced side-to-side difference in anterior laxity. However, there is insufficient evidence to conclude whether patient-specific function is improved at any specific tension. Level of Evidence: Level II, systematic review of Level II studies.

estoration of normal knee kinematics through

Rreplicating the natural anterior cruciate ligament iscore to achieving optimal outcomes in anterior cruciate

From the Departments of Physiotherapy (G.W.K., M.G.B.) and Orthopaedicurgery (P.A.D.), Queen Elizabeth II Jubilee Hospital, Queensland Health,risbane; the Division of Physiotherapy, School of Health and Rehabilitationciences, The University of Queensland (G.W.K., T.G.R.), Brisbane; and theivision of Physiotherapy (L.C.), School of Science and Health, University ofestern Sydney, Penrith, New South Wales, Australia.The authors report that they have no conflicts of interest in the authorship

nd publication of this article.Received January 2, 2013; accepted January 16, 2013.Address correspondence to Garry W. Kirwan, M.Phty., Department ofhysiotherapy, Queen Elizabeth II Jubilee Hospital, Locked Bag 2, Archer-eld, Brisbane, Queensland 4108, Australia. E-mail: g.kirwan@griffith.edu.u� 2013 by the Arthroscopy Association of North America0749-8063/137/$36.00http://dx.doi.org/10.1016/j.arthro.2013.01.021

34 Arthroscopy: The Journal of Arthroscopic and Related

ligament reconstruction (ACLR).1-4 A number of factorshave been proposed to contribute to such outcomes,including graft choice, tunnel placement, graft tension,graft fixation, tunnel motion, graft healing, and post-operative rehabilitation.4-7 Despite graft tension beingrecognizedas a significant contributor to successfulACLR,there is little consensus on the amountof tension requiredto produce an optimal outcome.5,8 As a result, manysurgeons rely on achieving an observed isometric graftbefore fixation to guide the amount of tension applied.9

Historically, graft tension was described subjectivelyas “strong tension” or “as far as it would go” andobjectively as preservation of 5 mm of anterior-posteriortranslation.10-12 In recent years, attempts have beenmade to quantify graft tension with the introduction oftensioning devices.7,13-15 The quantification of grafttension has been proposed to result in an optimalbiomechanical outcome, as well as an ideal environmentfor graft healing and ligamentization.8,16,17 Early

Surgery, Vol 29, No 5 (May), 2013: pp 934-941

Table 1. Methodology Quality Assessment Score

Methodology Criteria

1. Was there clear concealment of allocation?2. Were the inclusion and exclusion criteria clearly defined?3. Were the treatment and control groups adequately described

at entry, and if so, were the groups well matched orappropriate covariance adjustments made?

4. Were the surgeons experienced in the surgical procedures?5. Were the rehabilitation programs other than trial options

identical?6. Were the outcome measures clearly defined in the text

with a definition of ambiguous terms encountered?7. Were the outcome assessors blind to assignment status?8. Was a long-term follow-up performed (minimum, 6 mo)?9. Was the timing of outcome assessment in both groups

comparable and appropriate?10. Was loss to follow-up reported, and if so, were fewer than 5%

of patients lost to follow-up?11. Was a sample size calculation performed?12. Did the trial include an intention-to-treat analysis?13. Was there clear concealment of allocation?

NOTE. Each methodologic criterion was scored as 1 for yes and 0 forno, and a cumulative score was calculated.

TENSION AND FUNCTION AFTER ACL RECONSTRUCTION 935

investigation into the amount of tension a surgeonapplies when manually tensioning a graft showedsignificant variability between surgeons (mean differ-ence, 14.8 pounds [SD, 7.2 pounds]; P ¼ .002).2 There ispotential that this variability contributes to differentpostoperative outcomes achieved as a result of over- orunder-constraining the knee.Recent reviews of graft tensioning have reported

tensioning to range between 1 and 147.1 N in animal,cadaveric, and clinical studies.5,8 In human studiesa recent qualitative review of randomized controlledtrials suggested that 80 N of tension in a semitendinosus-gracilis (STG) graft produces a lower side-to-side differ-ence (STSD) in anterior tibial displacement.5 The articlegoes on to note the limited homogeneity betweentensioning protocols and also shows the variability in theamount of tension used in human studies.5

Although these reviews contribute to the body ofknowledge on graft tension, none have specificallyaddressed the impact of graft tension on functionaloutcomes after ACLR. In addition, there is limited quan-titativeanalysis available toprovide a comparisonbetweenoutcome measures reported. Thus the aim of this reviewwas to investigate the effect of graft tension on patient-specific functional outcomes after ACLR and undertakequantitative analysis of available data to determinewhether a particular tension produces superior functionaloutcomes after ACLR. Our hypothesis was that a mediumtension rangewould produce a better functional outcomewhen compared with low and high tension ranges.

MethodsA protocol outlining the search strategy, inclusion and

exclusion criteria, quality assessment, anddata extractionwas developed according to existing standards forsystematic reviews.18,19 To satisfy the aims of the review,the key search terms (full and truncated), used alone orin combination, were (1) anterior cruciate ligament orACL, (2) tension, and (3) function. Publications in theEnglish language only between 1950 and July 2012wereretrieved from the following databases: CochraneCentralRegister of Controlled Trials, Physiotherapy EvidenceDatabase, Medical Literature Analysis and RetrievalSystem Online (Ovid Medline), Cumulative Index toNursing and Allied Health Literature (CINAHL), Web ofKnowledge, Scopus, and Latin American and CaribbeanHealth Sciences Literature (LILACS).Articles were excluded if the primary focus was not

ACLR or the study used cadaveric, animal, or computermodels, on the basis that the objective of the reviewwas patient-specific function. A single author reviewedinitial search results and excluded articles where possibleon the basis of titles. Two authors then reviewed the titlesand abstracts of the retrieved articles and excluded arti-cles based on the criteria set. The remaining titlesand abstracts were de-identified for author, year,

place of publication, and country and were indepen-dently reviewed for inclusion based on the criteria set.These included use of a boneepatellar tendonebone(BPTB) or STG graft, use of a tensioning method atfixation, reporting of a functional outcomemeasure, anduse of a study design with a National Health andMedicalResearch Council Australia (NHMRC) level of evidenceIII-1 or higher.20 A sample of articles reviewed forinclusion were assessed by a second author to ensurereliability. A third reviewer was available to arbitrate ifconsensus was not reached.Articles were critically appraised by use of a quality-

appraisal tool adapted from Bourke et al.,21,22 and eacharticle was given a score out of 12 based on the criteriapresented in Table 1. All included articles were rankedaccording to theNHMRC level-of-evidence scale (I-IV).20

A data extraction tool was developed for the purpose ofthis review based on the Preferred Reporting Items forSystematic Reviews and Meta-Analyses (PRISMA)guidelines,19 and all study characteristics were extractedby the primary investigator. Themost frequently reportedoutcome measure was STSD in anterior tibial displace-mentmeasuredwithanarthrometer.Additional outcomemeasures included the International Knee Documenta-tion Committee score, hamstring and quadricepsstrength, hop test, and knee range of movement. Otherinformationextracted includedmeasuresofqualityof life,follow-up protocols, and patient demographics (genderand age). All extracted data are presented in Table 2.

Statistical AnalysisThe STSD in anterior tibial displacement outcome was

expressed as the effect size (ES) by use of the Cohendmethod.23 Twocomparisonswere performed, including

Table 2. Characteristics of Included Studies

Patient Demographics Study Characteristics Arthrometer Measurement (mm) Other Outcome Measures

Randomized StudiesSampleSize M:F Age (yr)

GraftChoice

TensionApplied (N)

Angle atFixation (�)

PreoperativeSTSD (mm)

PostoperativeSTSD (mm)

12-/24-moSTSD (mm)

IKDCCategory

QuadricepsStrength (%)

HamstringStrength (%)

Kim et al.25 (2006) 43 29:19 Gr 1, 27.1Gr 2, 22.6Gr 3, 23.7

STG Gr 1, 78.5Gr 2, 117.7Gr 3, 147.1

30 Gr 1, 5.6 (1.9)Gr 2, 5.0 (2.0)Gr 3, 6.3 (2.6)

NR Gr 1, 1.3 (1.4)Gr 2, 2.1 (1.9)Gr 3, 2.4 (2.2)

NR Gr 1, 90.9 (18)Gr 2, 91.3 (11.7)Gr 3, 88.8 (8.4)

NR

Nicholas et al.7

(2004)49 33:16 Gr 1, 30 � 7

Gr 2, 33 � 8BPTB Gr 1, 45

Gr 2, 90Full extension Gr 1, 5.3 (2.4)

Gr 2, 5.7 (2.7)Gr 1, 2.4 (2.4)Gr 2, 1.1 (1.7)*

Gr 1, 3.0 (2.2)Gr 2, 2.2 (1.6)

NR NR NR

van Kampenet al.13 (1998)

38 27:11 Gr 1, 28Gr 2, 28

BPTB Gr 1, 20Gr 2, 40

20 NR NR Gr 1, 2.6 (1.3)Gr 2, 2.5 (1.8)

Gr 1Cat A, n ¼ 2Cat B, n ¼ 10Cat C, n ¼ 6Cat D, n ¼ 1

Gr 2Cat A, n ¼ 1Cat B, n ¼ 10Cat C, n ¼ 8Cat D, n ¼ 0

NR NR

Yasuda et al.14

(1997)64 38:32 Gr 1, 22.7 � 6.7

Gr 2, 23 � 5.7Gr 3, 25.5 � 8.1

STG Gr 1, 20Gr 2, 40Gr 3, 80

30 Gr 1, 5.1 (1.9)Gr 2, 6.3 (2.1)Gr 3, 5.9 (1.8)

Gr 1, 2.2 (2.4)Gr 2, 1.4 (1.8)Gr 3, 0.6 (1.7)y

NR NR Gr 1, 92Gr 2, 89Gr 3, 91

Gr 1, 99Gr 2, 96Gr 3, 99

Yoshiya et al.15

(2002)43 21:22 Gr 1, 23

Gr 2, 23BPTB Gr 1, 25

Gr 2, 50Full extension NR Gr 1, �2.0

Gr 2, �1.8NR NR NR NR

F, female participants; Gr, allocated study group; M, male participants; NR, not reported; STSD, side-to-side difference in anterior tibial displacement.*Significant at P ¼ .01 level.ySignificant at P ¼ .05 level.

936

G.W.KIRWAN

ETAL.

Fig 1. Systematic review flow diagram.

TENSION AND FUNCTION AFTER ACL RECONSTRUCTION 937

between-tension differences in mean values at a partic-ular time divided by the pooled standard deviation andbetween-time differences in mean values for a particulartension divided by the pooled standard deviation. Resultsare presented as standardized mean difference (SMD)with 95% confidence interval (CI). Forest plots andstatistical analysis were completed by use of Excel(version 2008 for Mac; Microsoft, Redmond WA), SPSS(version 20.0; IBM, Chicago IL), and Review Manager(version 5.1; The Cochrane Collaboration, Copenhagen,Denmark).

ResultsWe initially identified 724 articles, of which 267 were

duplicates. Of the remaining 457 articles, 5were includedin the review for assessment and analysis (Fig 1). PubMedreturned the largest number of results (n ¼ 292), fol-lowed by Scopus (n ¼ 214) and Medline (n ¼ 60), withthe Physiotherapy Evidence Database and LILACSreturning no results. Eight review articles were identifiedin the search, and the reference lists were cross-checkedbefore removal with 1 additional article being included.Cadaver-based studieswere themost common reason forexclusion (n ¼ 256), followed by articles not focused on

Table 3. Methodologic Quality Assessment Score

Criteri

Article 1 2 3 4 5 6 7

Kim et al.25 (2006) 0 0 1 0 1 1 0Nicholas et al.7 (2004) 1 0 1 1 1 1 1van Kampen et al.13 (1998) 0 0 1 0 1 1 0Yasuda et al.14 (1997) 0 0 1 1 1 1 0Yoshiya et al.15 (2002) 0 1 1 0 1 1 0

*Based on NHMRC level-of-evidence guidelines.20

ACLR (n ¼ 105), with the remaining being animal- andcomputer-based studies. Of the 50 articles assessedagainst inclusion criteria, 8 articles progressed to qualityassessment and 5 articles met theminimum requirementof NHMRC level III-1 or higher.The methodologic score for each study is presented in

Table 3. The highest score achieved was 8 of 12. Themean score for the 5 included articles was 5.8 (SD, 1.3),with all articles failing to score in criterion 10 (loss tofollow-up reported), criterion 11 (sample size calcula-tion performed), and criterion 12 (trial included anintention-to-treat analysis). All articles scored a pointfor criterion 3 (treatment and control groups welldescribed and well matched), criterion 5 (rehabilitationprograms other than trial options identical), criterion6 (outcome measures clearly defined), and criterion8 (minimum of 6 months’ follow-up performed).

Study CharacteristicsParticipant numbers ranged from 38 to 64 partici-

pants, with a bias toward male participants (approxi-mately 2:1). Participant ages were similar (24.7 � 3.6years), as was follow-up duration (19.5 � 5.0 months),across studies (Table 2).

on

8 9 10 11 12 Total Score Evidence Level*

1 1 0 0 0 5 II1 1 0 0 0 8 II1 1 0 0 0 5 II1 0 0 0 0 5 II1 1 0 0 0 6 II

Fig 2. ES calculation between preoperative and postoperative STSD values in anterior tibial displacement at a given tension.

938 G. W. KIRWAN ET AL.

Initial Graft TensionInitial graft tension reported within the articles ranged

from 20 to 147.1 N. The most frequently reportedtension was 20 N (n ¼ 3),13,14,24 with 40 N reported ina further 2 studies.13,14 The remaining tensions reportedincluded 45 N, 78.5 N, 90 N, 117.7 N, and 147.1 N.

STSD in Anterior Tibial DisplacementSTSD in anterior tibial displacement asmeasured by an

arthrometer was reported in all 5 articles.7,13-15,25 Kimet al.,25 van Kampen et al.,13 and Yoshiya et al.15 re-ported no significant difference between groups imme-diately after surgery or greater than 12 months aftersurgery. Nicholas et al.7 reported a significantly lowerSTSD in anterior tibial displacement at 90 N of tensionwhen compared with 45 N immediately after surgery(P ¼ .01). Yasuda et al.14 reported that 80 N of tensionresulted in a significantly lower STSD in anterior tibialdisplacement when compared with 20 N immediatelyafter surgery (P ¼ .05) (Table 2).Quantitative comparison of STSD in anterior tibial

displacement between articles was performed with ESrepresented as an SMD with 95% CIs. Variability inreporting meant that data were unavailable preopera-tively for 2 articles, postoperatively for 3 articles, and at12 months or greater for 1 article (Table 2).There was a reduction in the STSD in anterior tibial

displacement from before surgery to after surgery for allarticles.7,13-15,25 It was shown that 80 N of initial tensionproduced the largest effect (SMD,�2.98; 95%CI,�3.82to �2.14)14 whereas 45 N produced the smallest effect(SMD, �1.19; 95% CI, �1.83 to �0.54)7 (Fig 2). Whendata before surgery and 12months or more after surgery

Fig 3. ES calculation between preoperative STSD values in antemonths or more postoperatively.

were compared, the greatest effect was produced at 78.5N (SMD, �2.45; 95% CI, �3.40 to �1.51)25 and, again,the smallest effect was produced at 45 N of initial tension(SMD, �0.98; 95% CI, �1.61 to �0.35)7 (Fig 3).Between-tension comparison at 2 weeks or less after

surgery showed the largest effect in favor of 80 Ncompared with 20 N (SMD, 0.76; 95% CI, 0.17 to 1.35)and the smallest effect toward 40 N compared with 20N (SMD, 0.37; 95% CI, �0.21 to 0.95) (Fig 4).14 At 12months or more after surgery, the largest effect wastoward 78.5 N compared with 147.1 N (SMD, �0.58;95% CI, �1.29 to 0.13)25 and the smallest effect wastoward 40 N compared with 20 N (SMD, 0.06; 95%CI, �0.58 to 0.70)13 (Fig 5).

Additional Outcome MeasuresThe International Knee Documentation Committee

subjective knee formwasused in 2 articles.13,15However,van Kampen et al.13 were the only authors to reporta comparison, and they showed no significant differencebetween 20 N and 40 N of tension.Two articles compared knee range of movement

between various tensions, with both reporting no signif-icant difference in either loss of flexion or loss of exten-sion between groups at 20 to 24months after surgery.7,14

Two articles compared quadriceps and hamstringstrength as a percentage of injured and uninjured, withboth articles reporting no significant difference betweentensions.14,25

DiscussionGraft choice, tunnel placement, initial graft tension,

and angle of fixation are all thought to contribute to

rior tibial displacement at a given tension and values at 12

Fig 4. ES calculations between tensions at 12 months or more after surgery.

TENSION AND FUNCTION AFTER ACL RECONSTRUCTION 939

functional outcomes in ACLR.16,26,27 This review aimedto evaluate the influence of initial graft tension on suchoutcomes. Quantitative comparison suggests a trendtowardmedium tension (78.5 to 90N) producing a lowerSTSD in anterior tibial displacement as measured by anarthrometer for STG or BPTB autografts. This, however,does not provide sufficient evidence when referringdirectly to patient function. Research has shown littlecorrelation between STSD in anterior tibial displacementand patient function,28-32 and the current literaturefailed to adequately report more appropriate measures.STSD in anterior tibial displacement was the primary

outcome measure in all 5 articles and the only outcomereported in sufficient detail to enable ES calculation tobe performed. Despite consistency in reporting, therewas inadequate homogeneity between methodologiesenable pooling of data for meta-analysis. The articlesreviewed used a wide range of initial graft tensions andselected various methodologies for measuring anteriortibial displacement.Regardless of methodology, ES calculations showed

a consistent trend toward medium tensioning producinga lower STSD in anterior tibial displacement. Comparisonbetween values before surgery and values at 2 weeks orless after surgery and at 12 months or greater aftersurgery identified an effect in favor of 80 N and 78.5 N,respectively. Furthermore, 80 N and 78.5 N produced thelargest effect when compared with 20 N and 147.1 N,respectively. In combination, these findings suggest thata medium tension may produce a superior outcome inSTSD in anterior tibial displacement. This is consistentwith the findings of Nicholas et al.7 and Yasuda et al.,14

who reported a significantly better outcome at 80 N and90 N, respectively. In contrast, van Kampen et al.13 re-ported no significant difference between 20 N and 40 N

Fig 5. ES calculations between tensions at 2 weeks or less after s

and concluded that 20 N was sufficient when selectinga BPTB graft. However, in our analysis of treatmenteffect, the difference between 20 N and 40 N was thesmallest of all the data analyzed (SMD, 0.06; 95%CI, �0.58 to 0.70), and therefore it would be reasonableto conclude that had a medium tension been included,a similar finding may have been reached.Although the review offers clinically relevant

recommendations regarding ACLR outcomes, the rela-tion between STSD in anterior tibial translationand patient function has not been established.28,30,31

Although evidence supports the measurement of STSDin anterior tibial displacement as a valid method fordetecting anterior cruciate ligament deficiency, there isno correlation between measured laxity and functionaloutcomes.28-33 A recent study by Kocher et al.31

showed no correlation between STSD laxity and activ-ities such as walking, jumping, and running stairs, aswell as return to sport, concluding that laxity is nota good indicator of functional success. This is supportiveof an earlier study by Eastlack et al.,28 who reportedthat patients with laxity after ACLR differed signifi-cantly in their functional outcomes and concluded thatpatients are better characterized as either copers ornoncopers based on their functional capability, not themeasured laxity. On the basis of the literature available,it is difficult to draw a definitive conclusion as to whichtension produces a superior functional outcome.

LimitationsAnumber of studies did address the effect of initial graft

tension on postoperative functional outcomes; however,the methodologic quality was often poor and thereporting of results was inconsistent (Table 2). On thebasis of the assessment of quality (Table 3), only 1 article

urgery.

940 G. W. KIRWAN ET AL.

showed robust methodology sufficient to draw a validconclusion.7 The remaining articles scored 5 to 6 of 12,and caution should be applied when one is interpretingthese results. Arneja et al.5 conducted a qualitativereview on graft tensioning and reported limitations ina similar set of articles, showing significant limitations inpost hoc power analysis, questioning the validity of theresults published. Further studies with high methodo-logic quality specifically investigating patient-specificfunctional outcomes are required to draw more defini-tive conclusions around function. In addition, consis-tency in study design will allow meta-analysis to beconducted in the future to draw definitive conclusions asto the effect of initial graft tension on function.

ConclusionsThe objective of this review was to systematically

assess the literature to determine whether a particularinitial graft tension results in superior outcomes afterACLR. From the review, there is a trend toward an initialgraft tension of 78.5 to 90 N resulting in a reduced STSDin anterior laxity as measured by an instrumented kneelaxity device. However, there is insufficient evidence toconclude whether patient-specific function is improvedat any specific tension.

References1. Andersen HN, Amis AA. Review on tension in the natural

and reconstructed anterior cruciate ligament. Knee SurgSports Traumatol Arthrosc 1994;2:192-202.

2. Cunningham R, West JR, Greis PE, Burks RT. A survey ofthe tension applied to a doubled hamstring tendon graftfor reconstruction of the anterior cruciate ligament.Arthroscopy 2002;18:983-988.

3. Dargel J, Gotter M, Mader K, Pennig D, Koebke J,Schmidt-Wiethoff R. Biomechanics of the anteriorcruciate ligament and implications for surgical recon-struction. Strategies Trauma Limb Reconstr 2007;2:1-12.

4. Woo SL, Wu C, Dede O, Vercillo F, Noorani S. Biome-chanics and anterior cruciate ligament reconstruction. JOrthop Surg Res 2006;1:2.

5. Arneja S,McConkeyMO,Mulpuri K, et al. Graft tensioningin anterior cruciate ligament reconstruction: A systematicreviewof randomized controlled trials.Arthroscopy 2009;25:200-207.

6. Ekdahl M, Wang JH, Ronga M, Fu FH. Graft healing inanterior cruciate ligament reconstruction. Knee Surg SportsTraumatol Arthrosc 2008;16:935-947.

7. Nicholas SJ, D’Amato MJ, Mullaney MJ, Tyler TF,Kolstad K, McHugh MP. A prospectively randomizeddouble-blind study on the effect of initial graft tension onknee stability after anterior cruciate ligament reconstruc-tion. Am J Sports Med 2004;32:1881-1886.

8. Heis FT, Paulos LE. Tensioning of the anterior cruciateligament graft. Orthop Clin North Am 2002;33:697-700.

9. Johnson D, Ayeni O. Avoiding and managing complica-tions in ACL reconstruction. In: Meislin RJ, Halbrecht J,eds. Complications in knee and shoulder surgery. London:Springer-Verlag, 2009;53-74.

10. Clancy WG Jr, Nelson DA, Reider B, Narechania RG.Anterior cruciate ligament reconstruction using one-third of the patellar ligament, augmented by extra-articular tendon transfers. J Bone Joint Surg Am 1982;64:352-359.

11. Noyes FR, Butler DL, Paulos LE, Grood ES. Intra-articularcruciate reconstruction. I: Perspectives on graft strength,vascularization, and immediate motion after replacement.Clin Orthop Relat Res 1983;(172):71-77.

12. Zarins B, Rowe CR. Combined anterior cruciate-ligamentreconstruction using semitendinosus tendon and iliotibialtract. J Bone Joint Surg Am 1986;68:160-177.

13. van Kampen A, Wymenga AB, van der Heide HJ,Bakens HJ. The effect of different graft tensioning inanterior cruciate ligament reconstruction: A prospectiverandomized study. Arthroscopy 1998;14:845-850.

14. Yasuda K, Tsujino J, Tanabe Y, Kaneda K. Effects of initialgraft tension on clinical outcome after anterior cruciateligament reconstruction. Am J Sports Med 1997;25:99-106.

15. YoshiyaS,KurosakaM,OuchiK,KurodaR,MizunoK.Grafttension and knee stability after anterior cruciate ligamentreconstruction. Clin Orthop Relat Res 2002;(394):154-160.

16. Friederich NF, O’Brien WR. Anterior cruciate ligamentgraft tensioning versus knee stability. Knee Surg SportsTraumatol Arthrosc 1998;6(suppl 1):S38-S42.

17. Tohyama H, Yasuda K. Significance of graft tension inanterior cruciate ligament reconstruction. Basic back-ground and clinical outcome. Knee Surg Sports TraumatolArthrosc 1998;6(suppl 1):S30-S37.

18. Higgins J, Green S. Cochrane handbook for systematic reviews ofinterventions version 5.1.0. Oxford: The Cochrane Collabo-ration, 2011.

19. Liberati A, Altman DG, Tetzlaff J, et al. The PRISMA state-ment for reporting systematic reviews andmeta-analyses ofstudies that evaluate health care interventions: Explanationand elaboration. J Clin Epidemiol 2009;62:e1-e34.

20. National Health and Medical Research Council. NHMRClevels of evidence and grades for recommendations for developersof guidelines. Canberra: National Health and MedicalResearch Council, 2009.

21. Bourke MG, Buttrum PJ, FitzPatrick PL, Dalton PA,Jull GA, Russell TG. Systematic review of medial para-patellar and subvastus approaches in total knee arthro-plasty. J Arthroplasty 2010;25:728-734.

22. Jadad AR, Moore RA, Carroll D, et al. Assessing thequality of reports of randomized clinical trials: Is blindingnecessary? Control Clin Trials 1996;17:1-12.

23. Durlak JA. How to select, calculate, and interpret effectsizes. J Pediatr Psychol 2009;34:917-928.

24. Mae T, Shino K, Matsumoto N, et al. Anatomic double-bundle anterior cruciate ligament reconstruction usinghamstring tendons with minimally required initialtension. Arthroscopy 2010;26:1289-1295.

25. Kim S-G, Kurosawa H, Sakuraba K, Ikeda H, Takazawa S.The effect of initial graft tension on postoperative clinicaloutcome in anterior cruciate ligament reconstruction withsemitendinosus tendon.Arch Orthop Trauma Surg 2006;126:260-264.

26. Bylski-Austrow DI, Grood ES, Hefzy MS, Holden JP,Butler DL. Anterior cruciate ligament replacements:A mechanical study of femoral attachment location,

TENSION AND FUNCTION AFTER ACL RECONSTRUCTION 941

flexion angle at tensioning, and initial tension. J OrthopRes 1990;8:522-531.

27. Nabors ED, Richmond JC, Vannah WM, McConville OR.Anterior cruciate ligament graft tensioning in full exten-sion. Am J Sports Med 1995;23:488-492.

28. Eastlack ME, Axe MJ, Snyder-Mackler L. Laxity, insta-bility, and functional outcome after ACL injury: Copersversus noncopers. Med Sci Sport Exerc 1999;31:210-215.

29. Forster IW, Warrensmith CD, Tew M. Is the KT-1000knee ligament arthrometer reliable. J Bone Joint Surg Br1989;71:843-847.

30. Irrgang JJ. Current status of measuring clinical outcomesafter anterior cruciate ligament reconstruction: Are wegood enough? Oper Tech Sports Med 2008;16:119-124.

31. Kocher MS, Steadman JR, Briggs KK, Sterett WI,Hawkins RJ. Relationships between objective assessmentof ligament stability and subjective assessment of symp-toms and function after anterior cruciate ligamentreconstruction. Am J Sports Med 2004;32:629-634.

32. Kowalk DL, Wojtys EM, Disher J, Loubert P. Quantita-tive-analysis of the measuring capabilities of the KT-1000knee ligament arthrometer. Am J Sports Med 1993;21:744-747.

33. Myrer JW, Schulthies SS, Fellingham GW. Relative andabsolute reliability of the KT-2000 arthrometer foruninjured knees. Testing at 67, 89, 134, and 178 N andmanual maximum forces. Am J Sports Med 1996;24:104-108.