PAPER

Preliminary study evaluating testsused to diagnose canine cranialcruciate ligament failureOBJECTIVE: To estimate specificity, sensitivity, positive predictive

value and negative predictive value of tests and signs used for the

diagnosis of cranial cruciate ligament failure in dogs.

METHODS: One stifle in each of 42 dogswas examined: 25 ‘‘affected’’

and 17 ‘‘control’’ dogs. All dogswere subjected to the following tests

when conscious: cranial drawer, tibial compression, patellar tendon

palpationandpalpationofthemedialaspectofthe joint.Undergeneral

anaesthesia, cranial drawer and tibial compression tests were

repeatedandalateralstifleradiographwastakentoevaluatechanges

of the infrapatellar fat pad. The results were analysed using a 232

table method. Sensitivity, specificity, positive predictive value and

negative predictive value were estimated.

RESULTS: The sensitivity of the cranial drawer and tibial compression

tests was surprisingly lowwhen performed on conscious patients but

significantly better when performed under anaesthesia. Similarly,

palpation of the medial aspect of the stifle joint cannot be

considered a reliable indicator of cranial cruciate ligament injury.

Patellar palpation and radiographic assessment showed excellent

sensitivity, specificity, positive predictive value and negative

predictive value.

CLINICAL SIGNIFICANCE: In the diagnosis of cranial cruciate ligament, it

is essential that the clinician is aware of each test’s features and

limitations to reduce the risk of misdiagnosis.

B. CAROBBI AND M. G. NESS*

Journal of Small Animal Practice (2009)50, 224–226DOI: 10.1111/j.1748-5827.2008.00723.x

Accepted: 16 November 2008; Publishedonline: 13 March 2009

INTRODUCTION

Failure of the cranial cruciate ligament(CCL) and associated secondary stifle oste-oarthritis (OA) is the most common causeof chronic hindlimb lameness in the dog(Ness and others 1996). Lameness causedby complete or partial failure of the CCLmay be acute or insidious in onset, andalthough secondary osteoarthritis is inevi-table, the character and severity of the resul-tant lameness are variable.

The diagnosis of CCL failure is usuallymade by physical and radiographic exami-nation. The physical examination should

include palpation to detect presence ofjoint swelling or bony thickening andmani-pulation to detect instability of the joint(Henderson and Milton 1978, Johnsonand Johnson 1993, Moore and Read1996, Jerram andWalker 2003). Radiolog-ically, theremight be evidenceof osteophyteformation and/or changes to the infrapatel-lar fat pad shadow (Moore and Read 1996,Jerram and Walker 2003).

Palpation is usually performed on bothstifles simultaneously in the standingpatient. In dogs with a recent CCL failure,a joint swelling can usually be appreciatedjust caudal to the patellar tendon (Jerramand Walker 2003), which is characterisedbya lossofdefinitionof that tendon. Indogswith a more chronic CCL failure, loss ofpatellar tendondefinitionmight still be evi-dent and, in addition, a thickening of themedial aspect of the joint capsule may bedetected.

Two tests have been described andwidely used to detect stifle joint instabilitybecause of CCL failure: the cranial drawertest (CDT) and the tibial compression test(TCT). However, both tests have thepotential to produce false negative results.Henderson and Milton (1978) suggestedthat each of these tests gave similar resultsfor recent CCL failures, but both are unre-liable in the diagnosis of chronic CCL fail-ure. However, they did not provide data tosupport their assertions. Moore and Read(1996), in a review ofCCL rupture in dogs,made similar statements about the unreli-ability of these tests, but again, no datawereprovided in support of their opinion.Morerecently, Jerram and Walker (2003) statedthat palpation and joint laxity are highlysensitive indicators of CCL rupture butthat chronic periarticular fibrosis, partialCCL rupture and tense patients may hin-der evaluation of stifle instability, but, oncemore, the authors provided no data to sup-port their assertions. Elsewhere, a range ofdiagnostic indicators of CCL failure havebeen described including the clinical, his-torical and radiological features, whichaccompany canine CCL, but there isremarkably little supporting evidence, and

Dick White Referrals, Station Farm, London Road,Six Mile Bottom, Newmarket CB8 0UH

*Croft Veterinary Hospital, 37-39 Croft Road,Blyth, Northumberland NE24 2EL

224 Journal of Small Animal Practice � Vol 50 � May 2009 � � 2009 British Small Animal Veterinary Association

no single test or sign has been shown to beuniversally applicable or reliable.

The purpose of this study was to esti-mate the specificity, sensitivity, positivepredictive value (PPV) and negative pre-dictive value (NPV) of a number of testsand signs commonly used in the diagnosisof CCL failure in dogs.

MATERIALS AND METHODS

Data collectionForty-two dogs were examined, both stifleswere evaluated as part of a standard phys-ical examination, and a single joint of eachpatient was included in the study. Twenty-five ‘‘affected’’ dogs had the diagnosis ofpartial or completeCCL failure, confirmedsubsequently by inspection of the ligamenteither at arthrotomy or by arthroscopyduring a therapeutic surgical procedure.Seventeen ‘‘control’’ dogs that were anaes-thetised for reason other than for investiga-tion of stifle lameness were examined.These were dogs with no clinical or radio-graphic sign of stifle abnormality both atthe time of evaluation and again at a secondclinical and radiographic examination sixto 12 weeks later.

Every dogwas subjected to the followingtests and assessments while conscious:CDTandTCTwith the dog lying in lateralrecumbency and patellar tendon palpationtest and palpation of the medial aspect ofthe joint with the dog standing. TheCDT and TCT were repeated with thepatient under general anaesthesia. In ad-dition, two standard orthogonal radio-graphic views centred on the stifle werecollected and evaluated for evidence ofchange to the infrapatellar fat pad.

Diagnostic testingTheCDTwasperformedwith thedog in lat-eral recumbency. The thumb and the indexfinger of one hand were placed on the lateralfabella and the patella, respectively, and thethumband the indexfinger of the other handwere positioned behind the fibular head andon the tibial crest, respectively. While thefemur was held stable, the tibia was forcedcranially and caudally without altering theangle of the stifle joint. The test was repeatedwith the stifle held in varying degrees ofextension and flexion.

The TCT was performed as describedby Henderson and Milton (1978). Withthe dog in lateral recumbency, one handgrasps themetatarsus, while the other handgently holds the femoral condyles with theindex finger placed along the patellar ten-don to the tibial crest. The tibiotarsal jointis alternatively flexed and extended todetect cranial translation of the tibia.

Patellar tendonpalpation involved evalu-ating the patellar tendon for evidence of lossof ‘‘sharpness’’ or definition because of anunderlying joint swelling, synovitis, etc.

Medial aspect palpation involved exam-ination of the medial aspect of the stiflejoint for signs of bony or dense fibrousthickening, especially in the region of themedial collateral ligament.

For each test, or radiographic evalua-tion, the result was recorded as ‘‘normal’’,‘‘abnormal’’ or ‘‘don’t know’’.

Statistical analysisEach set of results was analysed twice usinga 2�2 table method: the first time, all thedon’t know results were counted as if they

were, in fact, normal. For the second anal-ysis, all the don’t knows were counted as ifthey were, in fact, abnormal. For each test,the sensitivity and specificity were esti-mated along with PPV and NPV.

Sensitivity is defined as the proportionof true positives identified as such; specific-ity is the proportion of true negatives iden-tified as such. PPV is the proportion of testpositives that are truly positive, andNPV isthe proportion of test negatives that aretruly negative.

RESULTS

The results are summarised in Tables 1and 2.

DISCUSSION

Each of the tests evaluated in this studydepends to a large extent on clinical judge-ment and learned skill. The don’t knowgroup was included to reflect the essential

Table 1. Sensitivity, specificity, PPV and NPV calculated for each test with all ‘‘don’tknow’’ results counted as if they were ‘‘normal’’

Tests Don’t know in normal (%)

Sensitivity Specificity PPV NPV

Cranial drawer (CP) 60 100 100 62�96Tibial compression (CP) 64 100 100 65�38Patellar tendon palpation (CP) 100 100 100 100Medial aspect palpation (CP) 68 100 100 68Cranial drawer (GA) 92 100 100 89�47Tibial compression (GA) 88 100 100 85Radiograph 100 100 100 100

PPV Positive predictive value, NPV Negative predictive value, CP Conscious patient, GA Under general anaesthesia

Table 2. Sensitivity, specificity, PPV and NPV calculated for each test with all ‘‘don’tknow’’ results counted as if they were ‘‘abnormal’’

Tests Don’t know in abnormal (%)

Sensitivity Specificity PPV NPV

Cranial drawer (CP) 60 82�35 83�33 58�33Tibial compression (CP) 64 82�35 84�21 60�86Patellar tendonpalpation (CP)

100 100 100 100

Medial aspectpalpation (CP)

76 94�11 95 72�72

Cranial drawer (GA) 96 88�23 92�30 93�75Tibial compression (GA) 92 82�35 88�46 87�5Radiograph 100 100 100 100

PPV Positive predictive value, NPV Negative predictive value, CP Conscious patient, GA Under general anaesthesia

Journal of Small Animal Practice � Vol 50 � May 2009 � � 2009 British Small Animal Veterinary Association 225

Reliability of tests and signs for CCL failure

subjectivity of these diagnostic tests. Wewere keen to explore the possibility that,forexample, the confident findingofacranialdrawer signmight be a usefully accurate andprecise predictor of CCL failure but thediagnostic sensitivity and specificity of thetest overall might be diminished by thosecases in which the result was equivocal.

In our study, the sensitivity of the CDTwhen performed on conscious patients wassurprisingly low but significantly betterin anaesthetised patients. The CDT inconscious dogs is, perhaps, the most com-monly used diagnostic test for CCL injury,and our results indicate that reliance on thistest in the conscious patients will result insignificant numbers of CCL cases beingmisdiagnosed as cruciate-normal, and 40per cent of dogs with proven CCL failuredo not have cranial instability detectablewith the CDT.

When cases with a score of don’t knowwere counted as normal, the specificity andPPV of all tests are 100 per cent – this is inpart a reflectionof the relatively small num-bers of patients in this study but also theconsequence of the ability of the cliniciansperforming these tests to recognise normal-ity. However, when don’t know resultswere counted as if they were abnormal,the reliability of most of the commonlyused cruciate tests is lower. This evidenceindicates the diagnostic weakness of tests,which rely upon the subjective judgementof the clinician, and theuse of theCDT, theTCT or palpation of the medial aspect ofthe stifle joint will result in some dogs withCCL failure being misdiagnosed as nor-mal, while only 58 per cent of patients thattest negative are truly negative. The test isbetter, but still not perfect, when per-formed on anaesthetised patients.

Results for theTCTwere similar to thosefor CDT, and although results were consis-tently better in anaesthetised patients, bothtests lackedadequate sensitivity and specific-ity. Henderson and Milton (1978) sug-gested (without supporting data) that thispoor performance was because of the devel-opment of secondary arthrosis in chronicCCL failure. Although it could be arguedthat our dogs had acute disease because theywere evaluated at the time of first presenta-tion for hindlimb lameness, our data cannotscientifically support or refute HendersonandMilton (1978)claimasdataaboutdura-

tion of clinical signs upon presentationwerenot available in all dogs. We are aware thatanalysing testsperformedonapopulationofdogs with more chronic CCL failure wouldgive even less favourable results; however,there are no scientific data to support thisspeculation.

In addition, CDT and TCT show noagreement in only 80 per cent of cases.They were both positive in 13 and bothnegative in seven of the affected dogs. Theygave opposite results in the remaining fivepatients. When they were repeated underanaesthesia, the disagreementwas still pres-ent in one patient.

In the normal group, two dogs showedno agreement (CDT don’t know and TCTnegative in one dog and the opposite in theother) and two dogs resulted as don’t knowat both tests. Under general anaesthesia,three patients had the same results andone patient who had both tests results asdon’t know remained don’t know whenconsidering the TCT.

Patellar tendon palpation and radio-graphic assessment of the infrapatellar fatpad were excellent diagnostic tests with per-fect sensitivity, specificity, PPV and NPV.Both tests detect joint swelling only, whichis not specific to CCL failure. However, itshould be considered that a diagnosis ofCCL failure diagnosis is the result of thecombination of a vast amount of dataretrieved fromhistory,physical examination,radiographic examination and diagnostictests. This approach effectively excludes dogswith non-CCL-related stifle disease.

Although this work was a preliminarystudy involving only a relatively smallnumber of patients, the evidence indicatesthat there are considerable differencesamong the sensitivities, specificities, NPVsand PPVs for the commonly used tests fordiagnosis of cruciate ligament failure in thedog. Furthermore, perhaps the most com-monly used test – the CDT in a consciousdog – appears remarkably unreliable.

In conclusion, this preliminary workcasts doubt upon the reliability of thetwomore commonly used tests for diagno-sis of CCL failure, CDT and TCT, as theydo not allow detection of CCL failure in allaffected dogs. In contrast, patellar tendonpalpation and radiographic assessment ofthe infrapatellar fat pad predicted cruciatefailure in all affected cases but, as these tests

evaluate secondary stifle OA, there is a riskof false positive results in patientswith stifleOA secondary to pathology other than cru-ciate failure.

LIMITATIONS

Affected dogs were confirmed by directinspection of the damaged CCL either atarthrotomy or by arthroscopy at the timeof therapeutic surgery. Control dogs wereconsidered normal on the basis of a singleradiographic examination plus two normalclinical examinations at least sixweeks apart.However, because dogs in the control groupdid not undergo CCL inspection, there isa possibility of false negatives in this group.

All tests were performed by two surgeons(one board certified and the other a third-year surgical resident with substantial veter-inaryorthopaedic experience), and it is likelythat different testers will generate differentresults. The investigation of ‘‘interobserver’’variability relating to these diagnostic testswould be an area for further study.

Although CDT and TCT could beclaimed to be evaluating primary evidenceof CCL failure, all the other tests are eval-uating evidence of secondary stifle jointarthrosis. As such, they can be expectedto be positive with any stifle joint arthrosis,irrespective of cause; so they are not specificindicators of CCL failure. This is a poten-tial source of false positive results, whichcould influence our analysis of the diagnos-tic accuracy of these tests. However, thiseffect is likely to be small becauseCCL fail-ure is by far the most common cause ofcanine stifle joint arthrosis.

ReferencesHENDERSON, R. A. & MILTON, J. L. (1978) The tibial com-

pression mechanism: a diagnostic aid in stifle inju-ries. Journal of American Animal HospitalAssociation 14, 474-479

JERRAM, R. M. & WALKER, A. M. (2003) Cranial cruci-ate ligament injury in the dog: pathophysiology,diagnosis and treatment. New Zealand VeterinaryJournal 51, 149-158

JOHNSON, J. M. & JOHNSON, A. L. (1993) Cranial cruciateligament rupture. Pathogenesis, diagnosis, and post-operative rehabilitation. Veterinary Clinics of NorthAmerica: Small Animal Practice 23, 717-733

MOORE, K. W. & READ, R. A. (1996) Rupture of the cra-nial cruciate ligament in dogs – part II. Diagnosisand management. Compendium of ContinuingEducation for Veterinarians 18, 381-391

NESS, M. G, ABERCROMBY, R. H, MAY, C, TURNER, B. M, &CARMICHAEL, S. (1996) A survey of orthopaedic con-ditions in small animal veterinary practice in Brit-ain. Veterinary and Comparative Orthopaedicsand Traumatology 9, 43-52

226 Journal of Small Animal Practice � Vol 50 � May 2009 � � 2009 British Small Animal Veterinary Association

B. Carobbi & M.G. Ness