Long-term outcomes after arthroscopic treatment of dogs affected by osteochondrosis dissecans of the humeral trochlea, with or without medial coronoid disease: 23 cases (2012–2020)

Abstract

The objective of this study was to report long-term clinical and radiographic outcomes following arthroscopic reparative treatment — flap removal, curettage, and osteostixis of subchondral bone — in dogs with humeral trochlea osteochondritis dissecans (OCD). Dogs were included in this retrospective multicenter case series if they had a computed tomography diagnostic of humeral trochlear OCD, with or without medial coronoid disease, that was treated by arthroscopic reparative technique, and a detailed follow-up at least 6 mo postoperatively. The latter included a clinical examination, assessment of lameness, measurement of the brachial circumference and elbow amplitude, International Elbow Working Group (IEWG) radiographic score, owner-completed canine brief pain inventory (CBPI) score, and visual analogue scale (VAS) rating. A generalized linear model and tests for symmetry and marginal homogeneity were used to compare data. Twenty-three dogs (30 affected elbows) were included. Long-term (median: 22 mo; range: 6 to 98 mo) postoperative lameness, CBPI, VAS, joint distension, and pain scores were significantly improved compared with the preoperative values. Long-term postoperative range of motion and brachial circumference did not reveal any significant difference between OCD-affected and unaffected elbows. Long-term IEWG scores were similar to preoperative values in 56% of elbows and had progressed by 1 grade in 44%. Long-term complications included persistent Grade-1 lameness and occurred in 23% of dogs. Long-term outcomes based on lameness and CBPI scores were considered excellent in 67% of dogs, good in 27%, and intermediate in 6%. Arthroscopic treatment is thus a suitable surgical procedure for OCD of the humeral trochlea in dogs and provides good long-term results.

Introduction

Osteochondrosis dissecans (OCD) of the humeral trochlea affects mostly young dogs of medium to large breeds (1). Several treatments have been suggested. Initially, reparative techniques were carried out by arthrotomy, which consisted of excision of the cartilage flap with or without underlying bone marrow stimulation (2). These initial open approaches have not shown better results than conservative management, based on exercise restriction for 6 wk with or without injection of pentosan polysulfate once a week for 4 wk (3,4). Arthroscopic reparative techniques have later been described, including flap removal, curettage of any diseased cartilage surrounding the defect, and subchondral osteostixis to promote fibrocartilage growth (5).

Osteochondral autograft transplants (OAT) have more recently been described for treating humeral trochlear OCD lesions in dogs (6,7). Although technically feasible, this restorative technique has raised some major concerns, including donor site morbidity, invasiveness, significant revision challenges after graft or implant failure, quality of the grafted material, difficulty in restoring elbow congruity, and lack of described long-term follow-up (6,8). Synthetic osteochondral resurfacing, allografts, and xenografts have been mentioned to mitigate some of these concerns but have not been described in the canine elbow. In human orthopedics, despite many advanced reparative techniques, debridement and microfracture remain a gold standard in the surgical treatment of elbow OCD (9,10).

We are not aware of any study evaluating dogs with OCD of the humeral trochlea, including a long-term standardized postoperative assessment of this arthroscopic reparative technique based on clinical, radiographic, and owner-questionnaire scorings. The objective of this study was to determine whether arthroscopic debridement and microfracture is a suitable treatment for humeral trochlear OCD. Our hypothesis was that this minimally invasive technique allows long-term clinical improvement in canine patients, despite the development of osteoarthritis over time.

Materials and methods

This retrospective multicenter case series included dogs from 3 veterinary referral hospitals that were operated on by 3 experienced arthroscopic, Board-certified surgeons from 2012 to 2020. Data for each case, collected during preoperative and postoperative assessment, are listed in Table I. The inclusion criteria were the presence of a humeral trochlear OCD, among other elbow dysplasia components, diagnosed by computed tomography and treated by arthroscopic reparative technique, as well as a detailed long-term follow-up at least 6 mo after the surgery.

Table I.

Summary of data collected during preoperative and postoperative assessment.

Preoperative assessment	Breed, sex, age, BCS, weight, and duration of lameness Clinical and orthopedic examination, lameness score, search for palpable elbow distension and pain in flexion, extension, pronation, or supination Preoperative elbow radiographs and IEWG score when available
Long-term follow-up (At least 6 mo postoperatively)	Same clinical assessment as in preoperative, in addition to: brachial circumference and elbow range of motion measurement; postoperative elbow radiographs and IEWG score; and owner-completed CBPI and VAS scores.

BCS — Body condition score; IEWG — International Elbow Working Group; CBPI — Canine brief pain inventory; VAS — Visual analogue scale.

Follow-up included a clinical examination, assessment of lameness, measurement of the brachial circumference and maximal elbow range of motion, International Elbow Working Group (IEWG) radiographic score, owner-completed canine brief pain inventory (CBPI) score, and visual analogue scale (VAS) score.

A complete orthopedic examination was carried out on each dog and no other abnormalities of the forelimb were identified. Dogs with other concurrent components of elbow dysplasia were not excluded from the study. Elbow incongruity was diagnosed when greater than 2 mm. For each dog, the breed, sex, age, and duration of lameness were recorded. All dogs were in good health overall. A client consent form was signed by each owner prior to enrolment in the study. Surgery and postoperative follow-up were carried out by the same surgeon.

A preoperative clinical assessment was carried out in all dogs, including a lameness score, as described in Table II (11), search for palpable subjective elbow distension and pain in maximal flexion or extension, and evaluation of the presence of pain on pronation and flexion. The same clinical assessment was repeated at a long-term postoperative evaluation at least 6 mo after surgery, in addition to an evaluation of brachial circumference and maximal elbow range of motion measurements. The brachial circumference was measured halfway between the acromion and the lateral humeral epicondyle using a calibrated spring tension Gulick II measuring tape, with the dog awake and in the standing position. This process was repeated 3 times by the same surgeon on both front limbs and the mean of the 3 values was retained.

Table II.

Subjective numeric lameness scoring system at walk (11).

Lameness grade	Description
0	Normal
I	Slightly noticeable, intermittent lameness
II	Obvious weight-bearing lameness
III	Severe weight-bearing lameness
IV	Intermittent non-weight-bearing lameness
V	Continuous non-weight-bearing lameness

The maximal elbow range of motion was measured in maximum flexion and extension of the elbow using a universal plastic goniometer. With the dog awake and in lateral recumbency, the pivot point of the goniometer was placed over the lateral humeral epicondyle and the goniometer’s arms were aligned with the long axis of the humerus and radius. In the unilaterally affected dogs, the same data were recorded for the unaffected elbow. The body condition score (BCS) ranged from 1 to 9 and weight was recorded at admission.

Preoperative calibrated radiographs of the affected elbow were collected in standard craniocaudal and mediolateral views when available. All dogs underwent a preoperative computed tomography (CT) examination (16-slices, 0.625-mm thickness, Optima CT Scanner; General Electric, Coventry, England) of both elbows under general anesthesia. A long-term postoperative radiographic assessment was carried out on each dog, under sedation. Radiographic evidence of elbow dysplasia was recorded, including disruption of subchondral bone, flattening or concavity of the bone contour, sclerotic margins, presence of an articular bone fragment, irregularity of the medial coronoid process, articular incongruity greater than 2 mm or anconeal process fracture. Radiographic periarticular elbow osteophytosis was scored using the IEWG scoring system outlined in Table III (12).

Table III.

Classification of elbow arthrosis according to the International Elbow Working Group (IEWG) (12).

Degree	Definition
0	Normal elbow joint No evidence of incongruence, sclerosis, or arthrosis
1	Presence of osteophytes < 2 mm high Subtrochlear sclerosis, but trabecular pattern still visible
2	Presence of osteophytes 2 to 5 mm high Obvious subtrochlear sclerosis (no trabecular pattern) Radioulnar step of 3 to 5 mm Indirect signs of a primary lesion (OCD, MCD, ununited anconeal process)
3	Presence of osteophytes > 5 mm high Radioulnar step of > 5 mm Obvious presence of a primary lesion (OCD, MCD, ununited anconeal process)

OCD — Osteochondrosis dissecans; MCD — Medial coronoid process disease.

The following protocol was used for anesthesia: premedication with morphine (Lavoisier, Paris, France) or methadone (Dechra, Bladel, Netherlands), 0.2 to 0.3 mg/kg body weight (BW), intravenously (IV), followed by induction for endotracheal intubation with either propofol (Fresenius Kabi, Uppsala, Sweden) IV to effect or alfaxalone (Jurox, Worcestershire, UK), 2 mg/kg BW IV, and midazolam (Roche, Bale, Switzerland) or diazepam (Roche, Boulogne Billancourt, France), 0.2 mg/kg BW, IV. Anesthesia was maintained using isoflurane (Virbac, Carros, France) in oxygen. Lactated Ringer’s solution at 5 mL/kg BW per hour was administered IV for the duration of the surgery. The dogs were placed in lateral or dorsal recumbency and routinely prepared for aseptic elbow surgery by clipping and disinfecting the skin area extending from the shoulder to the carpus.

An arthroscopic exploration was carried out using a 30° 2.7-mm arthroscope in standard medial portal placement (13). The presence of an OCD lesion and evidence of concomitant elbow joint pathology, such as medial coronoid process disease (MCD), was assessed. The modified Outerbridge classification, as outlined in Table IV, was used to grade articular cartilage lesions in the humeral trochlea, radius, and ulna (14).

Table IV.

Modified-Outerbridge classification (14).

Grade	Description
0	Normal cartilage
1	Chondromalacia (evaluated using an arthroscopic probe)
2	Partial-thickness fibrillation
3	Deep fibrillation/incomplete cartilage loss (subchondral bone still covered)
4	Penetrating, full-thickness cartilage fissure (subchondral bone visible)
5	Eburnated subchondral bone

The OCD flap was removed using arthroscopic grasping forceps and the cartilage bed was gently debrided with a curette or shaver. The edges of the lesion were palpated using a right-angle probe or a small curette to remove any remaining diseased cartilage and create edges that were relatively perpendicular to the joint surface. To promote cartilage growth, osteostixis of the subchondral bone was carried out using an 18-gauge, 1.5-inch needle until subchondral bone bleeding was observed. When a concomitant fissured or fragmented medial coronoid process was identified, it was treated by arthroscopic removal of the fragment or partial or subtotal coronoidectomy. The joint was lavaged and the portal incisions were closed (13).

Postoperative analgesia was titrated to the individual dog’s requirements with morphine (Lavoisier), 0.2 mg/kg BW, IV every 4 to 6 h, or buprenorphine (Virbac), 0.02 mg/kg BW, IV every 6 to 8 h for the first 24 h. Non-steroidal anti-inflammatory medication with meloxicam (Boehringer Ingelheim, Ingelheim Am Rhein, Germany), 0.02 mg/kg BW, was administered IV at the induction of anesthesia and was subsequently administered at 0.1 mg/kg BW, PO, q24h for 20 d.

Strict rest was recommended for 4 wk, followed by a gradual increase in the duration of leash-only walking for 4 wk. Omega-3 fatty acids and chondroitin sulfate were prescribed for life.

Owner appreciation was quantified using the CBPI and VAS scores before surgery and at the time of the postoperative long-term evaluation (15). To complete the VAS score, owners were presented with a 10-cm line, one end of which corresponded to a healthy dog with no lameness and the other to a dog with constant non-weight-bearing lameness.

Surgical outcomes were considered excellent in dogs with no lameness at the long-term postoperative examination and a CBPI score of ≤ 20. The outcome was considered good when Grade-1 lameness persisted or when the CBPI score was 20 to 30 and intermediate when Grade-2 lameness persisted or when the CBPI score was 30 to 50. Poor outcome was defined as Grades 3 to 5 persistent lameness or a CBPI score of > 50.

The following outcome variables were compared before and after surgical treatment: lameness score, subjective presence of joint distension, presence of pain on pronation and flexion, presence of pain at maximal elbow flexion and extension, radiographic IEWG score, CBPI score, and VAS score.

A 2-level, random intercept, mixed-effects generalized linear model (GLM) was used to consider the repeated measures for each dog: linear regression for CBPI and VAS values (continuous variables), logistic regression for pain at elbow maximal flexion and extension (binary variable), and ordered logistic regression for the IEWG score (ordinal variable). For each model, the independent fixed-effects variables included the evaluation time (i.e., pre- versus postoperative), age, body weight, BCS, lameness duration, follow-up duration, bilateral elbow involvement, IEWG score, presence of concurrent MCD, and radial and ulnar modified Outerbridge score. The random-effects variable was the individual dog.

When it was not possible to use the generalized linear model (GLM) for variables that were often 0, the results of symmetry and marginal homogeneity tests were compared. For maximal flexion and extension angles and the brachial circumference, a 2-level, random intercept, mixed-effects regression model was used to consider the lack of independence of the 2 elbows in the same dog. For each model, the independent fixed-effects variables included the elbow (i.e., affected or unaffected) and the same potential confounding variables as previously mentioned. The random effects variable was the individual dog.

For regression models, assumptions of Gaussianity and homoscedasticity of the fixed residuals were checked graphically and by the Shapiro-Wilk and Breusch-Pagan and Cook-Weisberg tests, respectively. The goodness-of-fit of the logistic regressions was assessed using the Hosmer–Lemeshow test. All analyses were conducted using STATA, version 14.2 software (StataCorp, College Station, Texas, USA). The level of significance was set at P < 0.05.

Results

Twenty-three dogs (30 elbows) were included in the study. Seven had bilateral humeral trochlear OCD. The breeds included Australian shepherds (n = 5), Labrador retrievers (n = 5), golden retrievers (n = 3), Dogues de Bordeaux (n = 3), Newfoundland dogs (n = 2), and 1 each of chow chow, Shar pei, American Staffordshire terrier, American Akita, and crossbreed. The male:female ratio was 4.75:1. The median age at admission was 8 mo (range: 6 to 47 mo) and the median age at lameness onset was 5 mo (range: 3 to 41 mo). The median duration of preoperative lameness was 3 mo (range: 1 to 9 mo). The median body weight was 28 kg (range: 16 to 68 kg) and the median BCS was 6 (range: 4 to 7).

Preoperative and long-term (median: 22 mo) postoperative outcomes are listed in Table V and Table VI. A comparison of long-term postoperative outcome measurements between the OCD-affected and unaffected elbows is presented in Table VII. The range of elbow motion was not significantly associated with the IEWG score. At long-term postoperative follow-up, 70% of operated elbows had a normal extension angle and 43% had a normal flexion angle.

Table V.

Preoperative and long-term postoperative values of continuous outcome variables.

	Preoperative		Postoperative		P-value
	Median	Range	Median	Range
Lameness score	3	1 to 4	0	0 to 1	< 0.01
CBPI score	61	28 to 84	15	10 to 49	< 0.01
VAS	68.5	20 to 87	10	0 to 60	< 0.01
IEWG score	1	0 to 2	2	0 to 3	0.03

CBPI — Canine brief pain inventory; VAS — Visual analogue scale; IEWG — International Elbow Working Group.

Table VI.

Preoperative and long-term postoperative values of binary outcome variables.

	Preoperative	Postoperative	P-value
Presence of joint distension	100% (30/30)	23% (7/30)	< 0.01
Pain on pronation and flexion	100% (30/30)	17% (5/30)	< 0.01
Pain on maximal flexion	93% (28/30)	10% (3/30)	0.01
Pain on maximal extension	90% (27/30)	7% (2/30)	0.02

Table VII.

Comparison of long-term postoperative evaluation criteria between affected and unaffected elbows (n = 16 dogs).

	Unaffected limb		Affected limb		P-value
	Median	Range	Median	Range
Brachial circumference (cm)	26.3	20 to 35	25.8	20 to 38	0.18
Maximal flexion angle (°)	45	20 to 100	50	20 to 95	0.07
Maximal extension angle (°)	160	140 to 170	160	140 to 170	0.75

Preoperative radiographs were obtained for 18 affected elbows, 93% of which had preoperative radiographic evidence of humeral trochlear OCD. Based on the findings of the computed tomography examination, 90% of the elbows had a concomitant MCD and 30% had a subjective articular incongruity, either humeroradial or humeroulnar, none of which was > 2 mm, and 10% had evidence of ulnar panosteitis.

The preoperative and long-term postoperative radiographic IEWG scores are listed in Table V. These scores were stable in 56% of elbows and had progressed by 1 grade in 44% of the others; none had progressed by > 1 grade. IEWG scores were significantly positively associated with body weight (P = 0.02), the duration of lameness (P < 0.01), and bilateral elbow involvement (P = 0.05).

In all cases, arthroscopy confirmed the presence of humeral trochlear OCD lesions. All humeral trochlear modified Outerbridge scores were 5; the median radial score was 0.5 (range: 0 to 3) and ulnar score was 3.5 (range: 0 to 5). Ninety percent of the elbows had a concomitant MCD. None of the patients showed evidence of joint incongruity or ununited anconeal process. In all cases, arthroscopy allowed for reparative treatment of the OCD lesion.

No difficulties were experienced in flap removal, curettage of the cartilage bed, or osteostixis of the subchondral bone (Figure 1). Of the 30 operated elbows, none developed short-term complications from the procedure, except for lameness. Long-term complications consisted of persistent Grade-1 lameness and occurred in 23% of dogs (n = 5).

Figure 1.

Intraoperative photographs of arthroscopic treatment based on flap removal for humeral trochlear osteochondrosis dissecans (OCD). A — OCD flap. B — Grasping forceps removing the flap. C-Underlying subchondral bone after curettage of the OCD bed. D — Underlying subchondral bone after osteostixis of the OCD bed.

As shown in Table V, postoperative CBPI and VAS scores were significantly lower than preoperative scores (P < 0.01) and were not significantly affected by the IEWG score, body weight, duration of lameness, or bilateral involvement. Based on the previously mentioned classification, long-term clinical outcomes were rated as excellent in 67% of the dogs, good in 27%, and intermediate in 6%. No outcome was classified as poor.

Discussion

The main finding of this study was that long-term outcomes after arthroscopic reparative treatment of humeral trochlear OCD were good to excellent in most dogs. There were no short- or long-term complications as a result of this treatment, except for mild persistent lameness in a minority of cases. Based on postoperative lameness and CBPI scores, arthroscopic reparative treatment of OCD led to good to excellent clinical results in 94% of dogs at a median of 22 mo.

The results of the CBPI scores showed that the owners perceived an improvement in the quality of life of their dogs. The CBPI score is a validated outcome measurement that provides evidence of the actual positive impact perceived by owners after reparative arthroscopic treatment (16). It is not known whether the dogs would have improved without surgery, however, and this could only have been determined by adding a control group to the study.

The 47-month-old dog included in the study had an early diagnosis of bilateral OCD that resulted in unilateral lameness and led to unilateral surgery at 13 mo of age. In this case, the second elbow did not show clinical signs until almost 3 y later and was then operated on and included in this study.

The 2 main studies available that include cases of humeral trochlear OCD treated arthroscopically revealed good results in all 8 dogs and in 71% of 38 elbows with long-term follow-up (5,17). Both studies lacked detailed outcomes and quantified evaluation criteria. The results of the present study reinforce the findings of these studies.

Dogs with higher body weight, longer duration of lameness, or bilateral OCD had significantly higher IEWG scores than their counterparts, although these parameters did not significantly affect the CBPI and VAS scores. Similarly, there was no significant association between range of motion and the IEWG score. Despite these results, it seems advisable to operate on OCD lesions as early as possible since the absence of a correlation between the IEWG score, which is greater for dogs undergoing surgical treatment later in life, and clinical outcome could be related to a Type-II error.

Most dog breeds affected by OCD of the humeral trochlea included in this study have already been reported in the literature, except for the Australian shepherd and Dogue de Bordeaux (18,19). The Australian shepherd, which has recently become a very popular breed, has only been officially recognized since 2007, which may explain its absence in previous reports. It was the most affected breed in this study, accounting for 22% of all cases.

All but one of the dogs included in this study had a concomitant disease in addition to OCD, most often an MCD, which is a limitation of this study. This limitation is inherent in the frequent association between these 2 diseases. The high incidence of their co-occurrence supports the hypothesis that the aetiopathogenesis of OCD includes dynamic incongruity. The complexity of these lesions and their frequent association is an additional argument in favor of arthroscopy to treat elbow OCD, as OCD and MCD lesions can be accurately evaluated and treated using the same minimally invasive approach. This limitation may play a confounding role in the interpretation of our results, but it also represents the complex pathology of these dogs and their response to treatment.

Although quantifiable, some evaluation criteria used in this study were subjective. The CBPI score has been validated by gait analysis and has been shown to reliably quantify the owner’s assessment of chronic pain in dogs in their usual environment over a 1-week period (16). The VAS has also been shown to be repeatable, valid, and efficient in measuring pain and assessing lameness in dogs (20,21). Among the various methods used to objectively assess elbow function, goniometric assessment of range of motion has been proven reliable in both sedated and awake dogs (22).

The bilateral condition of the dogs included in this study may also have influenced the scoring of their lameness, resulting in difficulty in correctly assessing lameness in their forelimbs. Finally, as the same investigator carried out the surgery and measured the outcome, this person was not blinded to the severity of the disease identified on arthroscopy, which could have biased the results. Force plate gait analysis would have been a useful tool to objectively quantify postoperative outcomes at a given time.

The impact of the retrospective multicenter nature of this study was probably limited, given the similar level of experience of the surgeons and the similar operating conditions in the veterinary hospitals participating in this study. A larger number of cases could have allowed the subgrouping of concomitant diseases or preoperative clinical or radiographic status.

Furthermore, a controlled study would have permitted a better investigation of the therapeutic benefits of arthroscopic treatment for OCD of the humeral trochlea, compared with medically managed dogs and dogs undergoing treatment with alternative surgical techniques. Given the low prevalence of the disease, however, it may be difficult to conduct such a study (1).

In conclusion, arthroscopic treatment of humeral trochlear OCD provided good long-term results in the 23 dogs included in this study. This outcome was mainly characterized by a good to excellent return to function perceived by the owners and the veterinarians and a low evolution of periarticular osteophytosis. Owing to its low complication rate, especially compared to more advanced but invasive procedures such as resurfacing techniques, arthroscopic reparative techniques remain an appropriate approach for the treatment of elbow OCD.

Funding Statement

The authors thank M. Hauducoeur, Dr. C. Bismuth, and Y. Lahiani for their scientific advice, as well as IVC Evidensia for the funding of this study.