risk factors for ventral luxation in canine total hip replacement

Risk Factors for Ventral Luxation in Canine Total Hip Replacement

LAURA L. NELSON, DVM, JONATHAN DYCE, MA, VetMB, DSAO, MRCVS, Diplomate ACVS, and SUSAN SHOTT, PhD

Objective—To identify risk factors associated with ventral luxation (VL) of canine total hip re-placement (THR), and outcome.Study Design—Retrospective study.Animals—Dogs (n¼ 602) that had THR (563 cemented, 35 cementless, 4 hybrid).Methods—Dogs (1999–2004) with VL after THR were compared with dogs with uncomplicatedTHR. Data included signalment, body weight, diagnosis, implant size, acetabular cup orientation,and femoral displacement ratio (FDR).Results—VL was diagnosed in 11 (1.8%) dogs after primary THR. Including 2 other dogs that hadVL after the study period, 10 (77%) of 13 dogs had VL within 7 days of surgery. Risk factors for VLincluded Saint Bernard-type dogs (P¼ .0001), short neck extension (P¼ .0005), and high angle oflateral opening in other breeds (P¼ .018). There were trends toward higher risk of VL with lowerFDR in Saint Bernardtypes (P¼ .060), and with cementless implants (P¼ .061). Twelve dogs hadrevision arthroplasty that was successful in all cementless and 2 cemented VL cases. Five dogs hadrecurrent VL and a poor outcome.Conclusion—VL is generally an early complication of THR, with no single common risk factoridentified.Clinical Relevance—Saint Bernard types and short neck extensions are associated with increasedrisk of VL. Poor cup orientation is a determining factor for VL in some dogs, but a protectiveorientation of the acetabular cup was not found. Revision is successful in most dogs, but recurrentVL merits a guarded prognosis.r Copyright 2007 by The American College of Veterinary Surgeons

INTRODUCTION

TOTAL HIP REPLACEMENT (THR) is commonlyused for treatment of coxofemoral osteoarthritis in

dogs. Low complication rates associated with the use offixed head and modular cemented prostheses and a press-fit cementless prosthesis have been reported.1–4 Compli-cations of THR in dogs include luxation, infection,aseptic loosening, sciatic neurapraxia, protrusio ace-tabuli,3 femoral fracture,5 medullary bone infarction,6

and acetabular cup displacement.7 Luxation was de-scribed as a common complication of fixed head implants(Richards II, Richards Medical Company, Memphis,

TN) with incidences of 17% during the first 5 years of useand 3% during the following 3 years.4 The incidenceof luxation for cemented modular implants has rangesfrom 1% to 4.7%.1,3 The incidence of luxation for an ob-solete press-fit cementless prosthesis (PCA, Howmedica,Rutherford, NJ) was 5–10%.2,8

In dogs, THR luxation can occur in a dorsal or ventraldirection, as defined by the location of the femoral headrelative to the acetabulum on a lateral radiographic pro-jection (Fig 1). Earlier studies have reported predominantlycraniodorsal or caudodorsal luxation,9–11 or do not indi-cate the direction of luxation.2 It has been suggested thatdorsal luxation of canine THR is related to a high angle of

Presented in part at the 2006 ACVS Veterinary Symposium, Washington, DC, October 2006.

Address reprint request to Dr. Jonathan Dyce, Department of Veterinary Clinical Sciences, Veterinary Hospital, The Ohio State

University, 601 Vernon L. Tharp St., Columbus, OH 43212. E-mail: [email protected].

Submitted May 2007; Accepted June 2007

From the Department of Veterinary Clinical Sciences, Veterinary Hospital, The Ohio State University, Columbus, OH and the

Biostatistics Unit, Department of Medicine, Rush-Presbyterian-St. Luke’s Medical Center, Chicago, IL.

r Copyright 2007 by The American College of Veterinary Surgeons

0161-3499/07

doi:10.1111/j.1532-950X.2007.00316.x

644

Veterinary Surgery

36:644–653, 2007

lateral opening (ALO) of the acetabular cup3; however,this relationship is not supported by other studies.10

Reports of ventral luxation (VL) in cemented or press-fit cementless canine THR have been sporadic.10,12 VL ofa fixed head (Richards II, Smith & Nephew Richards,Memphis, TN) implant has been reported.10 In a series ofcanine hindlimb amputees that had THR with modularcemented prostheses, 1 of 4 luxations displaced in a ven-tral direction.12

Luxation in canine THR is generally an early compli-cation. Eleven of 12 cemented modular THRs that haddorsal luxation in 1 study did so within 9 weeks aftersurgery.10 Another study of luxation in cemented canineTHR reported that all luxations occurred within the first 4months after surgery.3 Five of 50 cementless PCA pros-theses had luxation during the early postoperative period.2

The risk factors and prognosis associated with VL incanine THR have not been evaluated. Our goals were toidentify risk factors associated with VL after modularcemented (CFX) and cementless (BFX) THR, and todescribe treatment and outcomes associated with VL. Wehypothesized that VL was likely to be an early postop-erative complication related to both surgical technicaland intrinsic patient factors.

MATERIALS AND METHODS

Medical records (June 1999–December 2004) of dogs thathad primary THR (602 consecutive cases) were reviewed to

identify patients with VL. All dogs had implantation of acommercially available modular CFX (cemented) or BFX(cementless) total hip prosthesis, or a hybrid combination ofcomponents (BioMedtrix Inc., Boonton, NJ). Surgeries wereperformed according to standard technique by 14 surgeonswith varying experience with THR.13,14 All THR revisionsurgeries performed during this time were also identified andcharacterized. Two additional dogs that had VL outside thestudy window were included in the review.

Patient age, breed, sex, body weight, preoperative diagno-sis, implant type, and size, surgeon experience, postoperativeinterval to 1st luxation, direction of initial and subsequentluxations, treatments, and outcomes were recorded for eachpatient diagnosed with VL.

Control populations of dogs that did not luxate were se-lected from the initial sample of 602 dogs. Patient age, breed,sex, body weight, preoperative diagnosis, implant type, andsize, most recent follow-up, and surgeon were recorded foreach dog. Exclusion criteria for the control populations in-cluded clinical or radiographic evidence of complication andabsence of documented evaluation at 3 months after surgery.The constitution of the control populations was as follows:

1. Cemented THR: 70 eligible dogs that had consecutiveCFX THR between January 2002 and December 2003.

2. Cementless THR: 25 eligible dogs that had BFX pros-theses between December 2003 and June 2005.

3. Predisposed breeds: Because of overrepresentation ofSaint Bernard-type breeds in the VL population, themedical records of all Newfoundlands, Great Pyre-nees, and Saint Bernard dogs that had THR betweenJune 1999 and December 2004 were reviewed, 9 dogswere eligible for inclusion; 8 had a CFX and 1 a BFXprosthesis. These dogs are also included in the totalslisted above for cemented and cementless controls.

Preoperative and postoperative radiographs were reviewedfor VL and control populations. When appropriate, postre-vision radiographs were also evaluated. ALO, acetabular ver-sion, acetabular inclination, and the femoral distractionratio (FDR) were determined for all dogs in the VL and con-trol populations and after each revision procedure.9,15

The methodology of these measurements is summarized inAppendix A.

Statistical Analysis

Statistical software (SPSS for Windows Version 13) wasused for data management and analysis. The log-rank test andCox proportional hazards regression were carried out to eval-uate relationships between the risk of VL and the followingpotential risk factors: weight, age, sex, implant head, stem,and cup size, FDR, acetabular version angle, inclination an-gle, ALO, and breed. Breed was represented as Saint Bernard-type (Newfoundland, Saint Bernard, or Great Pyrenees) ver-sus other breeds, and as 7 breed categories: Saint Bernard-type, German Shepherd Dog, Labrador Retriever, GoldenRetriever, Rottweiler, Mixed, and Other. Categorical variables

Fig 1. Dog 11. Ventral luxation of a cemented total hip re-

placement.

645NELSON, DYCE AND SHOTT

were represented as indicator variables in the Cox regressionanalyses. A .05 significance level was used for all statisticaltests; no 1-sided statistical tests were done.

RESULTS

Of 602 consecutive primary THRs (563 cemented, 35cementless, 4 hybrid; June 1999–December 2004), 60(10%) were revised or explanted. During this period, 11dogs (1.8% of primary THRs) had initial revision orexplantation for VL. For the purposes of this review, 2additional cases of VL were included to yield a totalstudy population of 13 VL cases. The supplementarycases are dog 7, a CFX THR that had initial THR in1998 and was revised in 2004, and dog 4, a BFX THRwith initial implantation and subsequent revision bothoccurring after the study period.

Other reasons for revision or explantation includedinfection (16 dogs; 2.7%), aseptic loosening (13 dogs;2.2%), dorsal luxation (11 dogs; 1.8%), luxation of un-documented direction (3 dogs; 0.5%), protrusio ace-tabuli, femoral fracture, and implant subluxation (1 dogeach; 0.2%).

Five dogs (38%) diagnosed with VL had this compli-cation within 1 day of initial surgery, 3 while undergeneral anesthesia during immediate postoperative radi-ography. Five other dogs (38%) had VL 1 day to 1 weekafter surgery, with the other dogs luxating 11 days, 14weeks, and 6.6 years after initial THR. Five dogs werediagnosed with initial VL after activity, witnessed bythe owner or a hospital staff member, resulted in severelameness, including a fall in the kitchen, jumping whileturning, chasing a cat, standing up, and collapsing splay-legged (Table 1).

Breeds represented within the VL population includedSaint Bernard (3), Newfoundland (2), mixed breed dog(3), and 1 each of the following breeds: German ShepherdDog, Rottweiler, Irish Setter, Old English Sheepdog, andBorzoi. Saint Bernard-type dogs, when considered as agroup, were at higher risk for VL compared with otherdogs (Po.0001).

THR was performed to address chronic traumaticdorsal hip luxation and acetabular rim fracture in theBorzoi and to address hip dysplasia and secondary osteo-arthritis in all other VL cases. In dog 13, contralateralTHR had previously been performed without complica-tion. Preoperative diagnosis was not significantly differ-ent from control populations.

Eight VL patients had cemented modular implants, 4had cementless modular implants, and dog 5 had a ce-mented cup and cementless stem. There was a trend to-ward an increased incidence of VL in cementless,compared with cemented, implants (P¼ .061). In dogs3 and 5 VL occurred after progressive retroversion of a

cementless femoral stem. The level of the femoral osteo-tomy in these 2 dogs was proximal to its recommendedalignment, flush with the greater trochanter. Dogs 4 and5 had measurable subsidence of cementless femoral stems(7 and 15mm, respectively). The BFX stem in dog 3 didnot subside, despite substantial retroversion. The subsi-dence noted in dog 4 was appreciated between immediatepostoperative and 3-month recheck radiographs, with VLoccurring 103 days after THR. Subsidence and retrover-sion in dog 5 and retroversion in dog 3 were noted in theimmediate postoperative period (Fig 2).

When the log-rank test was performed to compare thecommon neck extensions þ 0, þ 3, and þ 6 with respectto the risk of VL, a statistically significant difference wasfound. Pair-wise comparisons found a higher risk of VLfor implants with a shorter neck length (extension of þ 0compared with þ 3; P¼ .0005). No significant differencesbetween þ 6 and þ 0 or þ 6 and þ 3 extensions werenoted. Within the VL and control groups, only 3 dogshad a cemented implant with the 20mm cup/14mm (OD)head combination, one of which had VL.

For breeds other than Saint Bernard type, the risk ofVL increased as ALO increased (P¼ .018). For the entirepopulation sample and within the Saint Bernard-typepopulation, there were no significant relationships be-tween the risk of VL and ALO.

A trend (P¼ .060) was noted for FDR in Saint Ber-nard-type dogs only, with the risk of VL increasing withanatomically correct restoration of neck length, com-pared with a relatively long postoperative neck length.

No statistically significant relationships were identifiedbetween the risk of VL and sex, age, body weight, pre-operative diagnosis, cup or stem size, or acetabular ver-sion or inclination angle. Surgeon experience was notfound to be a significant risk factor for VL.

Treatment

Closed reduction was attempted without success in 1dog. All dogs with VL ultimately had surgery with ex-posure by reusing the craniolateral approach to the hipwithout greater trochanteric osteotomy. Relocation ofthe hip was often difficult, as the femoral head could lockagainst the ventral ischium. This effect was compoundedby fibrosis in the chronic case. Placing Kern bone-holdingforceps on the greater trochanter facilitated distraction ofthe luxated hip. After reduction, range of motion wasassessed to identify any focal impingement. Definitivestabilization of the hip was determined on an individualbasis, and included cup revision or replacement, necklengthening, soft tissue imbrication, femoral componentrevision, and implant removal (Table 1).

The acetabular component was initially replaced orrepositioned in 9 dogs. Of these dogs, neck length was

646 VENTRAL LUXATION IN CANINE TOTAL HIP REPLACEMENT

Table1.

Summary

Data

for13DogsThatHadVentral

LuxationAfter

TotalHip

Replacement

Dog

Breed

Age

(year)

Weight

(kg)

Sex

Implant

Typ

e

Head

Size

FDR

Tim

eto

Luxation

(days)

AssociatedEvents

RevisionProcedure(s)

Complications

Outcome

(Follow-U

p)

1New

foundland

264

MC

BFX

þ0

1.0

7Did

notuse

limb

postoperatively

Cuprevision:ALO

451(421),version

101(01),inclination141(191).Head

revisionþ6(þ0)

None

Good(2.5

months)

2Old

English

Sheepdog

1.5

31

FS

BFX

þ6

0.9

6Lameafter

attem

ptto

chase

cat

Cuprevision:ALO

581(541),

inclination171(331)

None

Good(3

months)

3Rottweiler

5.5

49

FS

BFX

þ0

1.4

6Stem

migration/

retroversion

Cuprevision:ALO

411(491),

inclination291(171),version61(91)

Stem

revision:CFX#9(BFX#11)

Headrevision:þ

0(þ0)

None

Good(3

months)

4Borzoi

1.5

26

FS

BFX

þ0

N/A

103

Jumped

whileturning

Headrevision:þ

0(þ0)

ALO

unchanged

(491)

None

Good(2

months)

5Mixed

0.7

28

MC

CFX

cup,

BFX

stem

þ6

1.0

o1

Stem

migration/

retroversion

Exp

lantationALO

(551)

None

FHO

6German

ShepherdDog

536

FS

CFX

þ3

1.1

7Non-w

eigh

tbearing

onstanding

Cuprevision:29mm

(27mm),

ALO

431(361),inclination

decreased

None

Good(24months)

7Mixed

0.9

40

MC

CFX

þ6

1.1

�2400

Fellin

kitchen

Cuprevision:ALO

471(401)

Headrevision:þ

6(þ6)

None

Good(3

months).Diedof

GDV

8Mixed

228

MC

CFX

þ5

0.9

o1

Foundsplay-legged

Closedreductionunsuccessful

Cuprevision:23mm

(20mm),

ALO

221(511)

Headrevision:þ

6(þ5)

None

Nofollow-up

9SaintBernard

1.1

63

MC

CFX

þ3

1.0

1.o1

2.7

3.135

Luxationidentified

bypostoperative

radiography

1.Headrevision:þ

6(þ3)

2.Cuprevision:ALO

461(251),

version131(101)

3.Explantation

1.Luxation

2.Luxation

FHO

10

IrishSetter

12

24

MCFX

þ6

1.3

1.2

2.8

1.Cuprevision:ALO

441(541).

Rem

oved

DAR

augmentation

2.Cuprevision:ALO

471(441),

version121(31).Headrevision:þ

9

(þ6)

1.Luxation

2.Luxation

Poor

11

SaintBernard

1.2

56

FS

CFX

þ6

1.2

1.o

1

2.61

Luxationidentified

bypostoperative

radiography

1.Cuprevision:27mm

(25mm),

ALO

381(291),inclination

decreased

2.Explantation

1.Luxation,

Infection

FHO

12

New

foundland

561

MC

CFX

þ6

1.0

1.o

1

2.6

3.24

Luxationidentified

bypostoperative

radiography

1.Open

reduction,im

brication

2.Capsularprosthesis

3.Headrevision:þ

9(þ6)

4Fem

oralexplantation

ALO

unchanged

(551)

1.Dorsallux

2.Luxation

3.Luxation

FHO

13

SaintBernard

1.5

56

MCFX

þ0

1.1

1.20

2.13

1.Cuprevision:ALO

381(191)

2.Capsularprosthesis

3.Explantation

1.Luxation

2.Infection

FHO

Wherelisted,measuredchanges

inacetabularorientationorim

plantsize

are

listed

asnew

(old).Referencesto

versionangledealwithretroversiononly.Thedirectionofluxationwas

ventral

unless

otherwiseindicated.

M,male

intact;MC,male

castrate;FS,female

spayed;CFX,cementedim

plant;BFX,cementlessim

plant;FDR,femoraldisplacementratio;ALO,angle

oflateralopening.


increased in dogs 1 and 8. In dog 3, the rotated cement-less stem was revised to a cemented stem. In dog 10, acorticocancellous bone graft from the femoral head hadbeen placed because of perceived dorsal rim insufficiency.This bone block was removed during revision to addressVL, in conjunction with repositioning the cup in a moreopen orientation. Three dogs that had cup revision had alarger diameter cup than had been originally placed. Sixdogs had repositioning of the original cup or a cup of thesame size.

Of the remaining 4 dogs, 1 had explantation of allcomponents, the neck was lengthened in 1 (head size in-creased from þ 3 to þ 6), and 2 were treated with openreduction with or without capsular imbrication.

Comparison of acetabular position before and afterrevision showed that ALO was altered in most dogs.Among dogs that had recurrent luxation, the cup wasreimplanted at a more closed ALO in 1 dog, opened in 3dogs, and allowed to remain essentially unchanged in 1dog. Among those dogs that were successfully revised,4 cups were closed, 1 cup was opened, and 2 were notsubstantially changed (Table 1).

The degree of acetabular version was altered minimal-ly in most dogs, with only 3-dogs undergoing increases inretroversion � 51. Inclination angle was difficult to as-sess in the most dogs with CFX implants because of po-sitioning of the twist on the wire marker such that it wascovered by the femoral prosthesis in a lateral radiograph-ic projection.9 In dog 6 abnormal inclination of the cup(cranial margin of cup excessively ventral) was confirmed

at the time of revision. Impingement of the femoral stemwith the cup was identified intraoperatively and was pos-tulated as the reason for VL. The cup was replaced in amore appropriate position.

Outcome

Seven dogs had a single revision surgery and weredischarged with a THR in place. All 4 dogs with entirelycementless components were successfully revised, andhad 3 month follow-up indicating good function. Dog 5had stem rotation without subsidence and was revisedwith placement of a cemented stem. Dog 4 had moderatestem subsidence without rotation during the months be-fore VL. The stem was stable at the time of revision andwas not altered.

Dog 7, which had VL 6 years after cemented THR,was revised with a 2nd cemented component, resulting inimproved function with mild lameness 8 weeks postop-eratively. Radiographs obtained at this time showed noevidence of complication. This dog later died of causesunrelated to THR (gastric dilatation-volvulus). Dog 6was not reevaluated radiographically at our hospital, butwas described by the owner during a telephone interviewto be functioning well 2 years after revision surgery. Dog8 with a cemented prosthesis that had a single revisionsurgery was lost to follow-up. Dog 5 was modified to anfemoral head ostectomy after the initial luxation (relatedto BFX stem subsidence and retroversion).

Fig 2. Dog 5. (A) Immediate postoperative lateral radiograph after implantation of a hybrid (cemented cup, cementless stem). (B)

Six-hour postoperative radiograph. Note subsidence and retroversion of the stem.


Five dogs had recurrent luxation after initial revision.All 3 Saint Bernard dogs and 1 Newfoundland werewithin this group. The remaining dog was an Irish Setter(dog 10). Three of these dogs were initially diagnosedwith luxation at the time of postoperative radiographicexamination. Three of 5 dogs that had recurrent luxationdid so in both ventral and dorsal directions; 4 had im-plant removal. Dog 10 was discharged with a luxatedimplant. Outcome is summarized in Table 1.

Intraoperative deep wound cultures yielded bacterialgrowth in only 1 dog following primary THR, but werepositive in 5 dogs after 1 or more revision surgeries forluxation. Results of primary intraoperative culture werenot predictive of infection after revision. Bacterial generaisolated at revision or explantation included coagulase-negative Staphylococcus spp. (3), Serratia sp. (1), Enter-ococcus sp. (1), and Pseudomonas/Staphylococcus spp. (1).Removal of implants was performed in dog 13 because ofradiographic changes consistent with periprosthetic in-fection. Culture confirmed Pseudomonas/Staphylococcusinfection. The hip remained normally articulated fromthe time of 2nd revision to explantation. In dog 11,explantation was performed to address luxation in thepresence of radiographically diagnosed periprosthetic in-fection. Both of these dogs had undergone 2 or morerevision surgeries for recurrent luxation.

Dog 5 had growth of Serratia sp. at implant removalfor migration, retroversion, and VL. The remaining 3dogs with positive cultures at revision had extended andmodified postoperative antibiotic treatment, but did nothave explantation.

DISCUSSION

During the study period, VL represented the 4th mostcommon indication for THR revision at our hospital af-ter periprosthetic infection, aseptic loosening, and dorsalluxation. Diagnosis of implant luxation was made in 4%of the primary THR population, similar to previous re-ports.1–3 The diagnosis of VL was made in 1.8% of theprimary THR population. This likely underestimates thetrue incidence of VL, because not all dogs had docu-mentation of lifelong implant stability. Consistent withprevious reports of dorsal luxation, VL was generally anearly postoperative complication.3,10 A small risk of VLpersists in the medium to long term, as illustrated by dogs4 and 7.

These findings are corroborated in the human litera-ture. The incidence of dislocation after primary THR ishighest during the 1st year after surgery, although the riskof dislocation is cumulative through the life of the im-plant.16 It is suspected that the mechanisms of early andlate dislocation differ, with implant orientation, incom-

plete capsular healing, and abductor (gluteal) insufficien-cy more important in the early cases.17 Trauma or wear-associated instability is of greater importance in latecases.17 Canine patients diagnosed with VL as a long-term complication became lame after an episode of minortrauma, consistent with this observation. In our studythere were no gross features of polyethylene deformationthat would suggest chronic impingement as an abnormalmode of wear, or of unusual attrition of articular poly-ethylene.

Patient Factors

Saint Bernard-type dogs were at greater risk for VLcompared with other breed groups undergoing THR.This breed group has not previously been reported to beat increased risk of complication related to THR. Nostatistically significant risk factors for VL related to im-plant position or size were found for this group. Therecommended range for ALO is between 351 and 551.3

Interestingly, the 3 Saint Bernards that had VL had rel-atively low ALO measurements (191, 251, 291). This couldpredispose to neck impingement during forced abductionof the femur, such as during postoperative radiographicexamination. In the case of a correctly sized and orientedcup, the cup will be flush with the dorsal acetabular rimand abut the medial acetabular cortex. A relatively un-dersized cup that meets these criteria will tend to beclosed. This situation could arise more commonly in giantbreed dogs. In addition, sheer size can frustrate settingand maintaining appropriate patient positioning on thesurgery table. Furthermore, the length of the femoral le-ver arm in Saint Bernard-type dogs and the lack of pro-tective passive muscular stabilization because of atrophycould also increase the risk of VL. Other potential rea-sons for this group to be at increased risk include poorcoxofemoral soft tissue tension, use of relatively under-sized implants, and an inherent degree of neuromuscularincoordination. Objective assessments are not routinelymade of muscle bulk or intraoperative tension; however,FDR can be used to yield a crude estimation of coxo-femoral soft tissue tension. There was a trend towardincreased risk of VL with correct restoration of necklength, compared with a relatively long postoperativeneck length.

In human THR, risk of developing dislocation isgreater in taller, thinner patients, potentially because ofincreased lever arm and hip translation with minimalforce.17 Soft tissue laxity, abductor insufficiency, and ne-uromuscular impairment (because of cognitive disorders,alcoholism, or primary neurologic disease) are alsoknown risk factors.17


Surgical Factors

The direction of luxation in human THR often followsthe direction of the surgical approach, with posteriorluxations being more common when a posterior ap-proach is used.18 All dogs in our series had THR via acraniolateral approach.

Implant Type

A trend toward overrepresentation of dogs with ce-mentless components was noted. In 2 dogs, the cement-less stem retroverted within the femoral canal. It isprobable that this stem displacement led to impingement-driven luxation. Retroversion in these dogs was not nec-essarily associated with significant subsidence. Anotherdog had VL after 7mm of implant subsidence was notedon routine 3-month radiographs. Whether there is indeedany relationship between simple subsidence and VL riskis unclear. Inappropriately small implants are suspectedto be at greater risk for subsidence, may preclude selec-tion of an adequately long femoral neck, and are lesslikely to achieve appropriate postoperative soft tissuetension. Predisposing factors for stem subsidence andretroversion may relate to osteotomy level, quality oftrabecular bone, method of femoral preparation and stemimpaction, stem alignment, femoral conformation, andpostoperative activity. Further study is necessary to as-sess the determinants of BFX stem stability.

In the other 2 cementless VL cases, we failed to iden-tify any risk factors unique to cementless implants.Whereas there is probably a greater potential for theincremental femoral broaching of BFX implantation todamage gluteal and external rotator muscle groups, therewas no notation of this complication in the medical re-cord.

The cementless THR cases included in our study rep-resent early use of this implant system at our hospital. Asthe technique for BFX implantation is refined, we wouldexpect a decrease in incidence of luxation similar to thatdescribed for other THR systems.4

Neck Length

Statistical analysis identified an increase in risk of VLwith the use of a þ 0 rather than a þ 3 neck extension.Though statistically significant, this finding would havebeen more compelling if there was a decrease in risk withfurther increases in neck extension (þ 6 or þ 9). A sim-ilar relationship was noted with FDR in Saint Bernard-type dogs. It is possible that neck length is a determinantof VL in some dogs. However, we did not find conclusiveevidence to support this, a potential type 2 statistical er-ror. It should be noted that there are other determinants

of functional neck length, such as the size and medial-ization of the cup, the level of the femoral osteotomy, andtranslation of the stem within the femoral canal. Thesefactors are taken into account with FDR evaluation.

Cup Position

Increased ALO was a significant risk factor for VLonly for non-Saint Bernard-type breed population. Thislack of significance for Saint Bernard-type breeds may berelated to low statistical power or suggest a mechanism ofluxation exclusive to these dogs (see above). Intuitively,we would hypothesize that abnormally closed cups withdegrees of inclination and retroversion predisposing toimpingement of the femoral neck cranially would driveVL. Apart from 1 dog where impingement related to ex-cessive cup inclination was determined and correctedintraoperatively, this was not supported by our data. In-creased ALO has previously been shown to be associatedwith dorsal luxation in canine THR.3 The factors thatallow an implant to undergo VL rather than dorsalluxation are uncertain. Interestingly, most VL cases thathad successful revision and appreciable alteration in ALOhad cups that were replaced in a more closed position.Among dogs with recurrent luxation, the opposite wasthe case.

Stem Position

Impingement is a product of the relative orientation offemoral and acetabular components. Data that do notaccount for stem position may underestimate the signifi-cance of implant orientation as a risk factor. Femoralstem version was not measured in this study because of alack of consistent femoral positioning in postoperativeradiographs.

Stem retroversion with or without subsidence of ce-mentless stems was identified as a novel failure mecha-nism in 2 dogs. It is likely that more subtle relativemalorientation of cup and stem components is a con-tributing factor to other VL and dorsal luxation cases,particularly during a fall or episode of sudden femoralabduction.

Revision and Outcome

Simple caudoventral luxation of the native hip hasbeen described in dogs.19 Caudoventral luxation can bemore difficult than dorsal luxation to reduce closed, butonce reduced, the hip remains in place in 80% of cases.19

Trochanteric avulsion has been described as a compli-cating factor in native caudoventral hip luxation, but wasnot noted in any case of VL after THR.2 The successfuluse of hobbles was described to stabilize the reduced hip,


but maintained reduction was not contingent upon theuse of hobbles.19 The temporary application of hobblesshould be considered in THR cases that are at high riskof VL.

All patients diagnosed with VL were ultimately treatedsurgically. Closed reduction was attempted unsuccessfullyin 1 dog. Repositioning or replacement of the acetabularcup was the most commonly used initial revision proce-dure for VL (9 dogs). In some dogs, radiographicmalposition of the acetabular cup or intraoperative im-pingement of the femoral neck and cup were noted in themedical record. Because larger cup diameters increase theoffset of the femoral head and neck in the absence offurther acetabular reaming, the 4 dogs that also had cupsof increased size might have addressed impingement orsoft tissue laxity. The rationale for a specific combinationof surgical manipulations was not consistently document-ed in the medical record. In support of implant malpo-sition as an important predisposing factor, most dogstreated with acetabular cup revision did not have repeat-ed luxation, notably those dogs where the cup was re-placed in a more closed position. As in dog 6, case-specific correction of inclination and version may alsohave prevented further luxation by addressing impinge-ment between the cup and femoral neck.

Recurrent Luxation

Five dogs had recurrent luxation after initial revision.This group of dogs uniformly had a poor outcome.Included in this group are 4 of 5 Saint Bernard-typedogs and 1 Irish Setter with profound hindlimb muscleatrophy. This group also includes all dogs that werediagnosed with luxation at the time of postoperativeradiography and all dogs that had luxation in both dorsaland ventral directions without change in cup position.We are confident that the dogs described actually hadVL, rather migrated from a dorsally luxated position,because of consistent postural abnormality and intraop-erative findings.

Both dogs that had the longest neck extension (þ 9)were within this group, suggesting an attempt by thesurgeon to overcome poor soft tissue tension. This groupmay be indicative of a subpopulation with inherently di-minished soft tissue support of the coxofemoral joint,exacerbated by a long lever arm (as described above).

In � 58% of human THR dislocations, instability isthought to be related to cup malorientation, with orwithout contribution from abductor insufficiency.18 Onethird of dislocations are attributed to abductor insuffi-ciency, supported by the observation that risk of dislo-cation increases 6-fold if a nonunion of 41 cm is presentafter osteotomy of the greater trochanter.18 In 10% ofpatients, a causative factor is not identified.18 This subset

of patients is at greater risk of experiencing chronic in-stability and multiple revisions.20 In addition, because offurther disruption of already potentially damaged glutealinsertions and capsular attachments, the dislocation rateis higher after revision surgery than primary THR.20 Thepopulation of dogs that had multiple revision surgeries isanalogous to this group.

Ventral luxation is generally an early complication ofcanine THR and has an incidence similar to that of dor-sal luxation in this study. No single common risk factorwas identified for VL. Had a unitary cause been present,there is confidence that it would have been found. Astrong breed disposition was identified among Saint Ber-nard-type dogs. Furthermore, these dogs appear partic-ularly likely to have recurrent luxation. A trend towardincreased risk was identified in dogs that had BFX hips,in some cases related to postoperative stem displacement.In contrast to dorsal luxation,3 we were unable to specifya safe range of acetabular orientation that would be con-sistently protective against VL; however, increased ALOwas found to be a significant risk factor in non-SaintBernard-type dogs, and impingement-driven VL wasidentified intraoperatively in 1 dog with excessive ace-tabular inclination. Revision was successful in most dogs,particularly where the underlying cause could be identi-fied and corrected. Saint Bernard-type dogs experiencingluxation after initial revision merit a guarded prognosisfor ultimate outcome.

REFERENCES

1. Massat BJ, Vasseur PB: Clinical and radiographic results of

total hip arthroplasty in dogs (1986–1992): 96 cases. J Am

Vet Med Assoc 205:448–454, 1994

2. Marcellin-Little DJ, DeYoung BA, Doyens DH, et al: Canine

uncemented porous-coated anatomic total hip arthropla-

sty: results of a long-term prospective evaluation of 50

consecutive cases. Vet Surg 28:10–20, 1999

3. Dyce J, Wisner ER, Wang Q, et al: Evaluation of risk factors

for luxation after total hip replacement in dogs. Vet Surg

29:524–532, 2000

4. Olmstead ML: Total hip replacement. Vet Clin North Am

Small Anim Pract 17:943–955, 1987

5. Liska WD: Femur fractures associated with canine total hip

replacement. Vet Surg 33:164–172, 2004

6. Sebestyen P, Marcellin-Little DJ, DeYoung BA: Femoral

medullary infarction secondary to canine total hip ar-

throplasty. Vet Surg 29:227–236, 2000

7. Hunter S, Dyce J, Butkus L, et al: Acetabular cup displace-

ment after polyethlyene-cement interface failure: a compli-

cation of total hip replacement in seven dogs. Vet Comp

Orthop Traumatol 16:99–104, 2003

8. DeYoung DJ, Schiller RA: Radiographic criteria for evalu-

ation of uncemented total hip replacement in dogs. Vet

Surg 21:88–98, 1992


9. Dyce J, Wisner ER, Schrader SC, et al: Radiographic eval-

uation of acetabular component position in dogs. Vet Surg

30:28–39, 2001

10. Cross AR, Newell SM, Chambers JN, et al: Acetabular com-

ponent orientation as an indicator of implant luxation in

cemented total hip arthroplasty. Vet Surg 29:517–523, 2000

11. Massat BJ, Miller RT, DeYoung BA, et al: Single-stage re-

vision using an uncemented, porous-coated, anatomic

endoprosthesis in two dogs: case report. Vet Surg 27:268–

277, 1998

12. Preston CA, Schulz KS, Vasseur PB: Total hip arthroplasty

in nine canine hindlimb amputees: a retrospective study.

Vet Surg 28:341–347, 1999

13. Roe SC,Marcellin-Little D: BioMedtrix BFXWorkshop Proce-

edings, Alameda East Veterinary Hospital, October 2004

14. Olmstead ML: The canine cemented modular total hip pros-

thesis. J Am Anim Hosp Assoc 31:109–124, 1995

15. Warnock JJ, Dyce J, Pooya H, et al: Retrospective analysis of

canine miniature total hip prostheses. Vet Surg 32:285–291,

2003

16. Berry DJ, von Knoch M, Schleck CD, et al: The cumulative

long-term risk of dislocation after primary Charnley total

hip arthroplasty. J Bone Jt Surg Am 86-A:9–14, 2004

17. Padgett DE, Warashina H: The unstable total hip replace-

ment. Clin Orthop Relat Res 420:72–79, 2004

18. Morrey BF: Difficult complications after hip joint replacement.

Dislocation. Clin Orthop Relat Res 344:179–187, 1997

19. Thacher C, Schrader SC: Caudal ventral hip luxation in the

dog: a review of 14 cases. J Am Anim Hosp Assoc 21:167–

172, 1985

20. Su EP, Pellicci PM: The role of constrained liners in total hip

arthroplasty. Clin Orthop Relat Res 420:122–129, 2004

Appendix A

Angle of lateral opening

The angle of lateral opening of a CFX cup was derivedfrom the digitized ventrodorsal radiograph by the trigo-nometric formula ALO¼ cos�1 short axis of the ellipse ofthe marker wire/long axis of the ellipse of the markerwire.9 To determine ALO for a BFX cup, a best-fit com-puter-generated ellipse (Image J, National Institutes ofHealth, Bethesda, MD) was superimposed on the innermargin of the metal rim surrounding the ventrolateralface on the ventrodorsal radiograph (Fig 3). The long andshort axes of each ellipse were measured and ALO wasderived by trigonometry as previously described.9 Theangle of pelvic rotation was assessed for all ventrodorsalradiographs and corrected with a trigonometric techniquewhere deviation of the dorsoventral pelvic axis from theperpendicular exceeded 31.9

Version

The version angle of a CFX cup was measured byplacing the knot within the marker wire at surgery in a

position opposite the open face of the cup. A tangent tothe wire at the knot is parallel to the plane of the lateralface of the cup and the angle subtended to the medianplane is then measured on a ventrodorsal radiograph.9

BFX acetabular version was determined by measuringthe angle between the median plane and the long axis ofthe open face of the cup on a ventrodorsal radiograph.The open face of the BFX cup was determined by a linedrawn between the cranial and caudal margins of lateralface on a ventrodorsal radiograph (Fig 4).

Inclination

The angle of inclination is the angle between the planeof the ventral surface of the lateral face of the cup and theilial–ischial axis, measured on a lateral radiograph.9,15

The angle of acetabular inclination was measured fromlateral radiographs of luxated CFX components only, asthe position of the knot of the wire marker (which ispositioned to identify orientation of the lateral face of thecup) is obscured by the head of the femoral prosthesisin the lateral radiographic view when the acetabular andfemoral components articulate normally. The angle ofacetabular inclination was similarly determined for allBFX components by measuring the angle between theilial–ischial axis and a line drawn between the cranial andcaudal margins of lateral face of the cup (Fig 5).

Fig 3. Ventrodorsal image of a BFX cementless acetabular

cup. A best-fit, computer-generated ellipse has been superim-

posed along the inner margin of the metal rim surrounding the

open face of the cup, with long (a) and short (b) axes plotted

and measured.


Femoral displacement ratio

FDR was calculated to provide a crude estimation ofcoxofemoral soft tissue tension. It is defined as the ratio

of the postoperative to the preoperative femoral displace-ment measurement (FDM).15 FDM is the distance fromthe midpoint of a line drawn from the medial border ofthe greater trochanter to the proximal lesser trochanter,and the center of an appropriately sized acetabular tem-plate placed in an ideal location on a ventrodorsalradiograph.15

Fig 4. The angle of version, U, of this BFX cementless cup is

defined by the angle formed by a line (b) drawn parallel to the

median plane (a) and a line drawn across the corners of the 451

cutout (c) on a ventrodorsal radiographic projection.

Fig 5. The angle of inclination, U, of this BFX cementless

cup is defined by the angle formed by a line drawn along the

ilial-ischial axis (d) and a line drawn across the bevel of the 451

cutout on a lateral radiographic projection (b).


risk factors for ventral luxation in canine total hip replacement

Documents