Materials and Methods

A convenience sample of 110 dogs undergoing unilateral arthrotomy and tibial plateau leveling osteotomy (n = 84) or lateral suture stabilization (26) at Affiliated Veterinary Specialists, Orange Park, Fla, for treatment of naturally occurring CCL rupture was included in the study. Dogs in which surgery had previously been performed on the affected stifle joint were excluded from the study.

Information obtained for each dog included age, sex, breed, body weight, duration of lameness, and whether any nonsteroidal anti-inflammatory medications or corticosteroids had been administered during the week prior to examination. Duration of lameness was defined as the sum of the owner-perceived duration of lameness prior to initial examination and the elapsed time between initial examination and surgery. A complete orthopedic examination was performed on each dog while the dog was awake and again after it had been sedated. All orthopedic examinations were performed by a single board-certified veterinary surgeon (RLG) and included examination of the contralateral joint for evidence of instability (eg, cranial drawer or cranial tibial thrust). Gait was examined at a walk and at a trot, and lameness scores ranging from 0 to 4 (0 = no lameness, 1 = intermittent weight-bearing lameness, 2 = persistent weight-bearing lameness, 3 = intermittent non–weight-bearing lameness, and 4 = persistent non–weight-bearing lameness) were assigned for each gait. All lameness scores were assigned by the same individual who performed the orthopedic examinations; scores assigned at a walk and at a trot were averaged to obtain lameness scores used for subsequent analyses. Orthogonal radiographic views of the affected and contralateral stifle joints in 90° of flexion were obtained prior to surgery. Radiographic views of the affected joint were evaluated by a board-certified veterinary radiologist (WCS) and assigned a degenerative joint disease score ranging from 0 to 24. Radiographic scores were determined by assigning a score from 0 to 3 (0 = absent, 1 = mild, 2 = moderate, and 3 = severe) for 8 characteristics (ie, patellar osteophytes, osteophytes involving the femoral trochlea, osteophytes involving the proximal aspect of the tibia, osteophytes involving the fabella, joint effusion, intra-articular mineralized opacities, avulsion fragments, and lesions compatible with osteochondritis dissecans) and summing scores for the individual characteristics. Tibial plateau angle was measured as described¹⁰ by the same individual who performed the orthopedic examinations.

Cytologic and histologic examination—Paired synovial fluid samples and synovial biopsy specimens were obtained from the affected stifle joints of all dogs. All samples were examined by a single board-certified veterinary pathologist (FAK) who was blinded to pairing of the samples.

Synovial fluid samples were collected immediately prior to surgery, following anesthetic induction and aseptic preparation of the limb, with a 22-gauge, 1.5-inch needle attached to a 3-mL syringe. A minimum of 0.5 mL of synovial fluid was obtained from the stifle joint. Direct smears were made immediately after collection and submitted for examination. The remainder of each sample was divided, with half placed in a plain evacuated glass tube and the other half placed in an evacuated glass tube containing calcium EDTA. For samples submitted in plain glass tubes, viscosity of the synovial fluid was determined by means of the acid precipitation (mucin clot) test. Quality of the mucin clot was visually inspected and subjectively recorded as normal or reduced.

Direct smears and smears obtained following cytocentrifugation of the anticoagulated sample were stained with a modified or rapid Romanowsky stain. Cellularity of the synovial fluid sample was subjectively graded as low (< 1,500 cells/PL), mild (1,500 to 4,999 cells/PL), moderate (5,000 to 7,500 cells/PL), or marked (> 7,500 cells/PL) as described,¹¹ and a differential cell count was performed manually by examining 100 cells. The presence or absence of hemorrhage in the synovial fluid sample was determined on the basis of number of RBCs. Total protein concentration was determined by means of refractometry.

Synovial biopsy specimens were collected following medial parapatellar arthrotomy and after examination of the CCL to determine extent of rupture (partial vs complete) and examination of the medial meniscus for gross abnormalities (normal vs abnormal). A 5-mm incisional synovial membrane biopsy specimen was obtained from the medial aspect of the joint capsule. Biopsy specimens were obtained approximately half the distance between the proximal and distal limits of the arthrotomy incision.

Synovial biopsy specimens were fixed in neutral-buffered 10% formalin for at least 48 hours and routinely processed for histologic examination; sections (5 μm thick) were stained with H&E for examination. Dogs were considered to have LPS if infiltrates of lymphocytes and plasma cells (≥ 1 cell/hpf) were seen, and severity of LPS was subjectively graded as mild, moderate, or severe. For all specimens, villous proliferation, fibrosis, and infiltration with hemosiderin-laden macrophages were recorded as present or absent.

Statistical analysis—For continuous variables (ie, age, weight, and lameness score), the Student t test was used to compare mean values between dogs with and without LPS. For categoric variables (ie, cellularity, villous proliferation, and fibrosis), the 2-sample z test was used to compare proportions between dogs with and without LPS. One-way ANOVA was used to compare tibial plateau angles among breeds. All analyses were performed with standard software.^a Values of P < 0.05 were considered significant.

Results

Dogs—Of the 110 dogs included in the study, 65 were spayed females, 39 were castrated males, 4 were sexually intact males, and 2 were sexually intact females. Thirty-six of the dogs were of mixed breeding, and the remaining 74 dogs represented 20 breeds. The most common breeds were Labrador Retriever (n = 22), Golden Retriever (9), Rottweiler (6), Mastiff (6), and Boxer (5), with remaining breeds represented by 3 or fewer dogs.

Prevalence of LPS and comparison of dogs with and without LPS—Histologic examination of synovial biopsy specimens revealed that 56 of the 110 (51%) dogs had LPS and 54 (49%) did not. Of the 56 dogs with LPS, 35 (63%) had mild LPS, 18 (32%) had moderate LPS, and 3 (5%) had severe LPS.

We did not detect any significant differences between dogs with and without LPS with regard to age, body weight, duration of lameness prior to surgery, severity of lameness (ie, lameness score) prior to surgery, score for radiographic severity of degenerative joint disease prior to surgery, extent of CCL rupture (partial vs complete), or gross appearance of the medial meniscal (normal vs abnormal; Table 1). The percentage of dogs that were spayed females was significantly higher for dogs with LPS than for dogs without, and the percentage of dogs with rupture of the contralateral CCL was significantly higher for dogs with LPS than for dogs without.

Mean preoperative tibial plateau angle was significantly lower in dogs with LPS than in dogs without. However, tibial plateau angle for mixed-breed dogs with LPS was not significantly (P = 0.58) different from angle for mixed-breed dogs without LPS. In addition, tibial plateau angle did not differ significantly (P = 0.536) among mixed-breed dogs, Labrador Retrievers, and Golden Retrievers.

Forty-two of the 56 (75%) dogs with LPS and 38 of the 54 (70%) dogs without LPS had received 1 or more anti-inflammatory medications (carprofen, firocoxib, deracoxib, etodolac, prednisone, tepoxalin, or aspirin) within the week prior to surgery.

Synovial fluid samples—Cellularity of the synovial fluid sample (low, mild, moderate, or marked) did not differ significantly between dogs with and without LPS (Table 2). Mononuclear cells (synoviocytes and macrophages) were identified in synovial fluid samples from all 110 dogs. Neutrophils were identified in a significantly higher proportion of synovial fluid samples from dogs with LPS than from dogs without LPS. Small lymphocytes were identified in only a single synovial fluid sample, which was obtained from a dog with LPS. There were no significant differences between groups with regard to proportions of synovial fluid samples with reduced viscosity or hemorrhage, and mean total protein concentration of the synovial fluid did not differ significantly between groups.

Synovial biopsy specimens—Prevalence of villous proliferation did not differ significantly between dogs with and without LPS (Table 3). The percentage of dogs with synovial fibrosis was significantly lower for dogs with LPS than for dogs without; however, hemosiderinladen macrophages were seen in a significantly higher percentage of dogs with LPS than dogs without LPS.

Discussion

Results of the present study suggested that LPS was common in dogs with naturally occurring CCL rupture, with 51% (56/110) of dogs in the present study having histologic evidence of LPS. This was similar to findings of a previous study.⁸ In addition, B and T lymphocytes have previously been identified in the synovium of dogs with CCL rupture.⁵ We also found in the present study that the proportion of spayed females was higher for dogs with than without LPS and that dogs with LPS had a lower mean tibial plateau angle and were more likely to have rupture of the contralateral CCL, neutrophils in their synovial fluid, and hemosiderin-laden macrophages in their synovium and less likely to have synovial fibrosis.

Persistent synovitis is currently believed to be an antigen-specific immune response within the joint,^6,12 but the most important antigens are still unknown. Antigens that could potentially play a role include type II collagen from degraded articular cartilage,^6,13 type I collagen and extracellular matrix of the torn CCL,^1,14–16 type I collagen from a torn meniscus,¹ and bacterial DNA.¹⁷ In the present study, extent of CCL rupture and gross appearance of the medial meniscus did not differ significantly between dogs with and without LPS, suggesting that antigens other than those from the ligament or meniscus may be playing a role.

It has also been suggested that chronic instability may play a role in the development of LPS in dogs with instability of the stifle joint. In dogs in which the CCL was experimentally transected, large synovial accumulations of lymphocytes and plasma cells were seen as early as 8 weeks following CCL transection.¹⁸ Because acute traumatic rupture of the CCL is uncommon in dogs, duration of stifle instability in dogs with naturally occurring CCL rupture is presumed to mirror the duration of lameness. However, duration of lameness did not differ significantly between dogs with and without LPS in the present study, suggesting that chronic instability may not have played a major role in the development of LPS. These findings must, however, be interpreted with caution because of the inherent inaccuracy of owner-reported lameness duration.

The proportion of spayed females was significantly higher for dogs with LPS than for dogs without in the present study, and previous studies^19–21 have suggested that spayed female dogs may have an increased risk for CCL rupture. However, the role, if any, of hypoestrogenism in CCL rupture has yet to be defined. In fact, recent studies^22,23 have shown that high estrogen concentrations compromise tensile stress and linear stiffness of the CCL in rabbits, and a higher percentage of anterior cruciate ligament injuries in female athletes occur during the ovulatory phase of the menstrual cycle.²⁴ If high estrogen concentrations contribute to mechanical failure of the CCL, then other factors, such as the presence of LPS, must be important to the pathophysiology of CCL rupture in spayed female dogs.

An interesting finding in the present study was that mean tibial plateau angle was significantly lower in dogs with LPS than in dogs without. A previous study²⁵ suggested that stifle joints with lower tibial plateau angles may be inherently more stable than joints with higher angles. Thus, we suggest that in dogs with low tibial plateau angles, it is possible that LPS may have been present prior to rupture of the CCL and may have contributed to failure of the ligament. This is supported by the fact that dogs with LPS in the present study were significantly more likely to have rupture of the contralateral CCL, compared with dogs without LPS. On the other hand, our data must be interpreted carefully, as the diagnosis of contralateral CCL rupture was made on the basis of clinically detectable stifle joint instability, radiographic signs of stifle joint effusion, or both. Visual examination of the stifle joint and histologic examination of a synovial biopsy specimen would have been required to confirm whether dogs had CCL rupture and LPS in the contralateral joint. Because dogs in the present study were clinical patients, such testing was not possible.

Histologic evidence of synovial fibrosis was significantly less common in dogs with LPS than in dogs without LPS in the present study. This finding was unexpected because several proinflammatory mediators previously identified in the stifle joints of dogs with CCL rupture are chemotactic for fibroblasts.^1,7 Synovial fibrosis is a result of increased deposition of extracellular matrix by resident fibroblasts and has been reported to occur with synovitis.^18,26 This process is modulated by a number of proinflammatory cytokines, growth factors, prostanoids, proteinases, and oxygen-derived free radicals.^26–28 Tumor necrosis factor-D, on the other hand, has been shown to prevent synovial extracellular matrix formation.²⁹ Thus, it may be possible that concentrations of tumor necrosis factor-D and other antifibrotic growth factors may be higher in dogs with LPS. However, further investigation is required to evaluate this relationship.

The percentage of dogs in the present study with hemosiderin-laden macrophages in their synovium was significantly higher for dogs with LPS than for dogs without. Hemosiderin is a breakdown product of erythrocytes that undergoes phagocytosis by macrophages. Thus, hemosiderin-laden macrophages may be seen when chronic gross or microscopic hemarthrosis is present.³⁰ Because hemorrhage may be associated with arthritis,^18,30 it is possible that chronic hemarthrosis may contribute to or be a result of LPS.

Because synovial fibrosis and the presence of hemosiderin-laden macrophages both represent chronic changes in the synovium, we had expected that results for these 2 variables would mirror each other. It was surprising, therefore, that dogs with LPS were less likely to have synovial fibrosis but more likely to have hemosiderin-laden macrophages in their synovium. The reason for this disparity is unknown; however, additional investigation into possible inhibition of fibroblasts, upregulation of synovial macrophages, and the role of chronic hemarthrosis in LPS is warranted.

Characteristics of synovial fluid from healthy and diseased joints have been published previously.^5,31–33 In the present study, cellularity of the synovial fluid did not differ significantly between dogs with and without LPS, although a higher proportion of dogs with LPS had neutrophils in their synovial fluid. Neutrophilic infiltration has been observed in association with various acute and chronic inflammatory arthritides,^32–34 and neutrophils release cytokines and chemokines that perpetuate the inflammatory process.³⁴ Interestingly, lymphocytes were seen in synovial fluid from only 1 of the 56 dogs with LPS in the present study. This contrasts with results of another study,³¹ in which CD4-positive lymphocytes were found in the synovial fluid of dogs with CCL rupture.

Important limitations of the present study include the fact that information on duration of lameness and administration of anti-inflammatory medications was provided by owners of dogs included in the study and could not be verified. Owner perceptions of duration of lameness are subject to error, and owner reports of anti-inflammatory medication administration may be subject to inadequate recall. Additionally, evaluations of severity of preoperative lameness, severity of radiographic signs of degenerative joint disease, and synovial fluid cellularity and viscosity were all subjective. Furthermore, evaluation of direct smears of synovial fluid may not provide the most accurate quantitation of cells because of their uneven distribution.^35,36 To improve diagnostic accuracy, smears obtained following cytocentrifugation were also evaluated.³⁶ All synovial biopsy specimens in the present study were obtained from the region of the medial arthrotomy, which may have introduced some sample bias. Finally, long-term follow-up would have been helpful to determine whether LPS affects outcome following stifle joint stabilization.