Nonconstrained Elbow Replacement in Dogs


Address reprint requests to Dr. Mike Conzemius, DVM, PhD, Diplomate ACVS, University of Minnesota College of Veterinary Medicine, 1352 Boyd Avenue, St. Paul, MN 55108. E-mail:


Objective— To review development of a nonconstrained total elbow replacement system for use in dogs and report the surgical technique used for implantation.

Study Design— Descriptive report.

Animals— Dogs with chronic elbow osteoarthritis (OA) unresponsive to medical management for at least 1 year.

Methods— A nonconstrained elbow replacement system (radioulnar and humeral components) was developed and used in dogs with medically nonresponsive elbow OA. The components were refined based on evaluation of clinical outcome. Changes in humeral stem design for composite fixation, the contours of the articulating surfaces, and in the humeral component to increase range of motion were incorporated. Drilling and cutting guides were developed to facilitate accuracy of implantation.

Results— With component refinement, more favorable long-term functional outcome was achieved.

Conclusions— Design improvements incorporated into the elbow replacement system should increase treatment success.

Clinical Relevance— Total elbow replacement in dogs is possible and good long-term outcome can be achieved.


TREATMENT ALTERNATIVES for dogs with moderate to severe elbow osteoarthritis (OA) include nonsurgical management, debridement arthroplasty (removing loose bodies and osteophytes from the joint), and arthrodesis.1–3 Unfortunately, treatment of end-stage elbow OA in dogs remains an unsolved problem in veterinary orthopedics. Nonsurgical management for elbow OA has improved with use of well-designed nonsteroidal anti-inflammatory drugs, specialized diets, and new knowledge about the merits of physical therapy and body weight management. Because OA is progressive, in many dogs clinical signs worsen even with palliative treatment. This clinical progression serves as an indication for surgical intervention. Surgical debridement of osteophytes, joint lavage, and surgery of diseased areas within the joint in an attempt to stimulate new cartilage growth has not been reported. Elbow arthrodesis results in a fair outcome but with a visible lameness.3 There is seemingly no evidence to suggest that current treatment options offer much hope of functional restoration without joint replacement. Indeed, comparative evaluation of the efficacy of these approaches is desperately needed.

Total elbow replacement can be beneficial but is costly and not without patient risk.

Regardless of design, for elbow replacement to be useful, it must result in superior limb function to that achieved without surgery, which has not been established for any system including the one described in this report. At a minimum threshold, dogs for joint replacement should have signs of lameness and pain on a daily basis even with appropriate nonsurgical treatments, and improvement in limb function should be objectively documented after surgical intervention.

To my knowledge, total elbow replacement in an animal was initially described in 1964 when a constrained (hinge-like) prosthesis was used to treat a cat with a comminuted fracture of the elbow.4 Subsequently, Dr. Ralph Lewis in a brief proceedings note reported his experience with total elbow arthroplasty in a dog5 using a constrained (hinge-like) component. Although some successful outcomes occurred, the high complication rate suggested that system redesign was needed. Anecdotally, I am aware of 3 other systems used clinically but not reported. Dr. Phil Vasseur (University of California–Davis) designed a 4-component, nonconstrained system, used it in 3 dogs with naturally occurring elbow OA, but abandoned the approach after poor short-term outcome. Dr. Jimi Cook (University of Missouri) designed, tested, and abandoned a semiconstrained total elbow because of disappointing outcomes. More recently, Dr. Randy Acker (in collaboration with BioMedtrix, Boonton, NJ) developed the TATE system and used it to treat end-stage elbow OA in 7 dogs.

Conzemius et al6 designed a semiconstrained, 2-component (humeral and radioulnar) elbow-replacement system, based anatomically on Greyhounds and tested in 6 Greyhounds. Although postoperative complications occurred in 4 dogs, there were beneficial observations. Two dogs had fair outcome with peak vertical force reaching 82%, 4 months after surgery, suggesting that elbow replacement was possible but not with this exact design and technique. A subsequent in vivo study was used to evaluate the efficacy of a modified system in 6 normal, large breed dogs (25–38 kg).7 Components were implanted and limb function was objectively monitored before, and 8, 16, 24, and 52 weeks after surgery using force platform gait analysis. Outcome was mixed; postoperative complications occurred in 3 dogs but 3 dogs had consistent improvement in limb function over 52 weeks with ground reaction forces similar to the unoperated, normal limb at 1 year.7 These dogs adopted by veterinary students continued to do well for 5 years after surgery (the last time of follow-up).

Key findings were that elbow components should be less constrained; synostosis between the radius and ulna could be achieved using only a distal ulnar ostectomy and bone graft between the proximal aspect of the radius and ulna, and that limb function could return to normal when elbow replacement was performed in a normal dog.7 Accordingly, design flaws in the components and implant system were addressed and a 3rd study completed in 20 adult, client-owned dogs with elbow OA managed unresponsive to nonsurgical management for at least 1 year.8 Dogs were evaluated clinically, radiographically, and by force platform gait analysis before surgery and 3, 6, and 12 months after surgery. At 1 year, 16 dogs had a satisfactory outcome, defined as improved quality of life and a reduction in pain and lameness. For the operated limb, ground reaction forces at 1 year were typically 25% greater than preoperatively.8 Unsatisfactory outcomes were associated with infection (1 dog), lateral luxation (2 dogs), and an iatrogenic humeral condylar fracture (1 dog). Given the study inclusion criteria, the overall outcome was clinically as well as statistically significant. Potential improvements in component design, implantation system, and surgical technique were identified and implemented. This elbow replacement system has now been used clinically for >10 years at multiple hospitals internationally with ∼750 component pairs distributed by the manufacturer (BioMedtrix).

Thus the purpose of this report is (1) to review the methodology for preclinical and clinical development and testing for safety and efficacy of a total elbow arthroplasty system for dogs, (2) to identify the changes in components and surgical technique that have occurred since this elbow replacement system was last reported in the peer-reviewed literature,8 and (3) discuss current limitations of this elbow replacement system.


Findings from the 1st study revealed 2 primary component flaws8: (1) the humeral component had large proximal shoulders that required removal of an unnecessary amount of bone, so these shoulders were eliminated (Fig 1), and (2) it was nearly impossible to establish a good cement mantle at the interface between the sides of the humeral component and the remaining bone of the humeral condyle.8 The humeral component was redesigned for composite fixation with a cemented stem and porous ingrowth fixation at the interface of the sides of the component and the humeral condyle (Fig 2). One potential advantage of this fixation type is good early strength from the cement fixation, which should create a stable environment for bone ingrowth at the porous surfaces. Such bone ingrowth at the porous surfaces should provide for good, late, long-term fixation to mechanically protect the cemented stem. Other modifications were changing the articulating surface so that no aspect was flat and, in fact, both components were cut with a radius of curvature to improve stability in rotation. Metal was also removed from the central groove of the humeral component to allow for more flexion or extension during articulation with the radioulnar component.

Figure 1.

  Humeral and radioulnar components. (A) Cranial view. Note diminished size of the shoulder between the stem and component body, and the curvature of the articulation. (B) Sagittal view. Components articulating at full extension.

Figure 2.

  Humeral component embedded in a plastic bone. Cut away section to show composite fixation.


The primary technical challenge that remained was development of drilling and cutting guides that would allow for placement of the humeral component on the center of rotation of the origin of the collateral ligaments and positioning of the radioulnar component on the same center of rotation as the humeral component. This approach was developed to achieve a more reproducible outcome, reduce the probability of lateral luxation, and increase patient range of motion without soft tissue binding.

Briefly, with the dog positioned in lateral recumbency, a lateral approach is made to the elbow by incising through the middle of the anconeus muscle and the lateral collateral ligament. The elbow is luxated laterally after removing large osteophytes that are generally present on the cranial surface of the radius and cutting fibrous tissue between the anconeus and humerus. After identification of the origin of the lateral collateral ligament, the elbow joint is reduced, and the carpus and elbow manipulated through a range of motion in a plane parallel to the floor. With the elbow reduced and the elbow held parallel to the floor, a long 3.2 mm drill bit is used to drill from the origin of the lateral collateral ligament through the condyle in a plane perpendicular to the floor. The bit is left in situ and the elbow luxated. Using an appropriately sized bit (5–9 mm), a hole is drilled from the supratrochlear foramen up the humeral medullary canal and then a humeral cutting guide inserted so that the proximal line on the cutting guide is aligned with the lateral collateral drill bit. The humeral cutting guide is positioned in the correct rotational plane and fixed in place with a small pin that is drilled through the humeral diaphysis and locks into a hole in the cutting guide (Fig 3). The collateral drill bit is removed and the humeral cuts made.

Figure 3.

  Humeral cutting guide mounted in a plastic bone. The arrow (closed point) identifies the line that is positioned at the level of the pin. The hole drilled across the condyle is identified by a second arrow (open point).

After completion of the cuts, the cutting guide is removed with the central portion of the humerus and a humeral trial component is inserted. A pin is inserted into the lateral collateral drill hole and guided through a hole in the humeral trial (Fig 4A) to lock it in position at the origin of the lateral collateral ligament with an articulation that is in a plane perpendicular to the ground. If there is malalignment of the elbow joint (e.g. collapse medially) it can be corrected by changing the orientation of the collateral pin and humeral trial component. The elbow joint is reduced and the alignment rod is placed over the humeral pin (Fig 4B). The coring cutter is then placed onto the alignment rod and a coring guide slid over the cutter and rotated so that 2 of its holes are over the ulna. The guide is secured in position with two 3.2 mm drill bits (Fig 5). The cutting guide, coring instrument, and rod are removed and the elbow luxated. The coring instrument is placed onto the end of the long coring extension tool to cut the radius and ulna (Fig 5).

Figure 4.

 ( A) Humeral trial secured in position with a condylar pin. (B) Alignment rod placed over the pin.

Figure 5.

 ( A) Coring cutter mounted on the alignment rod (a) and the coring guide slide (b) over the top of the cutter (c). The 2 holes (d) that are drilled to secure the Coring Guide are clearly visible; a securing pin has been inserted through the left hole into the ulna. (B) Coring guide (b) mounted to the ulna and the coring cutter (c) removing the bone of the radius and ulna.

Finally, preparation of the radial and ulnar medullary canals is performed using drill bits until a trial radioulnar implant can be easily inserted. The elbow is reduced and the articulation between the 2 components tested. A distal ulnar osteotomy is made and the humeral component, followed by the radioulnar component, is cemented in place. A bone graft (collected from the bone segments removed) is inserted between the radius and ulna just distal and lateral to the radioulnar component and the incision closed. Radiographic follow-up for the 1st year is recommended for all patients (Fig 6).

Figure 6.

  Lateral radiographic projections. (A) preoperative, (B) immediate postoperative, and (C) 1 year postoperative. This dog had elbow replacement surgery because of a >6 month duration of severe lameness after being shot in the elbow. At surgery, there were full thickness defects of cartilage and loss of some subchondral bone in the area of the ulnar humeral articulation and a nonunion olecranon fracture. Elbow replacement was used to treat the cartilage and subchondral bone loss; a bone plate and autogenous bone graft (observable on the caudal aspect of the ulna) were used to treat the nonunion fracture of the ulna.


A nonconstrained design was selected because constrained (hinge-like) designs share load poorly with intact ligamentous structures.9,10 Constrained designs necessitate that load is principally absorbed by the implant and concentrated at the implant–bone interface.11 Such designs have not withstood the test of time in load-bearing joints; the best example being the human knee.9,10 Constrained total knee designs have a comparatively high rate of aseptic loosening and are reserved for use in knee revision surgery when no ligamentous structures remain intact.9,10

Cement fixation was used because it allows for a greater variability in implant design and positioning. Further, cement fixation allows the stems of the implant to be virtually anywhere within the confines of the medullary canal as long as there is sufficient space for at least a 2 mm mantle.12 Press-fit and porous in-growth designs require a near perfect fit between existing bone anatomy and implant. One limitation of the current radioulnar component is that the stems may not allow for a 2 mm cement mantle because their size fills most of the diaphysis of the radius. This is a potential area for improvement if cementless fixation of the radioulnar component were instituted. Although long-term rechecks are rarely performed if not requested by the owner, I am aware of a single instance of aseptic loosening 5 years after implantation in a dog that had bilateral elbow replacements with the unaffected limb having been implanted 6 years earlier.

To date, all generations of the components for this system have been designed so that the humeral component is universal for left and right limbs (Fig 7). Whereas this feature reduces manufacturing costs and limits the inventory that hospitals need, it necessitates that the articulating surface between components be identical for left and right limbs. This design feature, a straight trochlea, differs from the normal anatomy of the humeral trochlear, which is more spiral in nature. Although the current nonconstrained design allows for some internal and external rotation between components, it is possible that the straight trochlear design increases the probability of component luxation. This is especially likely given that current approach to the joint (transection of the lateral collateral ligament) destabilizes the joint laterally and provides a path of decreased resistance for radioulnar component luxation. Redesigning the articular surface of the components so that the radius of curvature is greater should provide for greater resistance to rotation and reduce the frequency of luxation.

Figure 7.

  Photographs from a caudal view (image on left) and lateral view (image on right) of 3 generations of the nonconstrained elbow replacement described. The components on the far left7 were used and described in 6 experimental dogs; the components in the middle8 were used and described in 20 consecutive clinical cases. The component pair on the far right have been used since 2001.

When a dog is admitted with end-stage elbow OA and lameness, and is unresponsive to nonsurgical management, elbow replacement is considered a reasonable treatment option. Owners need to be fully informed of all treatment options, the reported success rate of elbow replacement (currently, 80%), and that this means that 20% of dogs will require a 2nd major surgery.8 Unfortunately, the ideal elbow replacement system is likely years away. Given that we cannot glean much from the human surgical literature because there is no comparable human model for elbow OA in dogs, we should carefully evaluate new techniques and systems prospectively and objectively. Improvements in existing elbow replacement systems based on experience generated from surgeons that have used it with and without success will aid our chances of developing the optimal elbow system.