Medical records of client-owned dogs were reviewed for cases diagnosed arthroscopically with an early partial (≤50%) tear of the craniomedial band of the CCL that was treated with BMSC–PRP or ADPC–PRP between January 2010 and December 2015 at Veterinary Orthopedic & Sports Medicine Group (VOSM). Data collected included signalment, medical history, physical and orthopedic examination findings, objective temporospatial gait analysis findings, diagnostic radiography results, arthroscopy findings, treatments, and outcome. Patients with concomitant stifle joint problems, or orthopedic or neurologic conditions affecting the hind limbs, were excluded from analysis.

Dogs were assessed at baseline and 30, 60, and 90 days posttreatment with an orthopedic and neurologic examination. Degree of lameness, presence of stifle effusion, presence of cranial drawer and cranial tibial thrust in flexion and extension, and McMurray test result were among the findings recorded (38). Additional abnormal orthopedic or neurologic findings were also recorded.

Objective gait analysis was performed at baseline and 90 days posttreatment in a quiet room using a temporospatial pressure sensing walkway. The walkway system was equipped with an 8.23 m × 0.85 m portable mat with 29,952 encapsulated sensors (GaitFour Pressure Sensing Walkway, platinum version, CIR Systems Inc., Havertown, PA, USA). The active dimensions of the mat were 8.04 m × 0.61 m. A 1.25 m × 0.85 m section of inactive mat was placed at each end of the walkway system to provide transition surfaces when entering and exiting the system. The mat was calibrated by the manufacturer. The walkway system interfaced with a computer and software program for processing and storing raw data recorded from quadruped gait analysis (GaitFour, platinum version, CIR Systems Inc., Havertown, PA, USA). One camera was positioned at a height of 50 cm at the end of the walkway system to record movement (Logitech mega pixel web camera, Logitech, Freemont, CA, USA). Digital video files of each pass across the walkway system were automatically linked to the data files for footfall verification.

Each dog was familiarized with the room prior to walking on the mat and was handled by the same examiner. Dogs were made to walk on the mat until they appeared relaxed, approximately four to six passes per dog. Three passes were recorded at both a walk and a trot, where a pass was defined as a dog moving along the length of the walkway in one direction for four to six gait cycles. A satisfactory pass consisted of minimal head turning, a walk velocity of 0.9–1.2 m/s or trot velocity of 1.8–2.6 m/s, and less than two standard deviations in gait cycle velocity differences. The software program identified the paw print of each footfall and analyzed multiple variables for each pass. The data calculated for each pass provided the total pressure index percent (TPI%) (sum of peak pressure values recorded from each activated sensor by a paw during mat contact/total sum of peak pressure values for all feet × 100). The TPI% of the treated limb was compared to the TPI% of the contralateral limb at baseline and at 90 days posttreatment. The TPI% of the treated limb at day 0 was also compared to the treated limb TPI% at 90 days posttreatment. Dogs diagnosed with bilateral partial CCL tears were excluded from this analysis.

Routine lateral and craniocaudal radiographs were performed of both stifles in all patients at baseline and evaluated for evidence of stifle osteoarthritis by a board-certified veterinary surgeon (39).

A standard parapatellar portal was utilized for arthroscopic evaluation of the stifle. The stifle was clipped and aseptically prepared and a 22-gauge needle was placed into the lateral compartment. Synovial fluid was aspirated using a 5 ml syringe to ensure that the needle was in the joint. Approximately 10 ml of sterile saline solution was injected into the stifle to distend the joint. A 15 blade was used to create medial and lateral parapatellar stab incisions to allow insertion of the cannula and trochar into the stifle. A 1.9 or 2.7 mm arthroscope was inserted into the cannula, depending on the size of the dog. The stifle joint was explored arthroscopically and assessed all structures, including the craniomedial and caudolateral band of the CCL, the caudal cruciate ligament, and medial and lateral menisci. Cartilage breakdown was scored using the Outerbridge classification (40). The fat pad was not removed with a shaver, but rather an L-shaped probe (Graduated Black Probe 2.5 mm tip length, Arthrex Vet Systems, Naples, FL, USA) was inserted and used to move the fat pad for direct observation of the CCL. The torn portion (percent) of the CCL was measured using the standardized L probe. If ≤50% of the craniomedial band of the CCL was determined to be damaged, a BMAC–PRP or ADPC–PRP injection was to be performed. If a >50% tear was observed, a tibial plateau leveling osteotomy (TPLO) was performed. The arthroscopic incisions were closed routinely.

The dog was placed in lateral recumbency, and the hip was clipped and aseptically prepared. The vaclock syringe and bone marrow aspiration needle were primed with 20 ml of heparin 1,000 U/ml, leaving 5 ml in the vaclock syringe. A 15 blade was used to create a 0.5 cm stab incision at the point of the proximal femur and an 11 gauge × 110 mm bone marrow aspiration needle was inserted into the greater trochanter. Approximately 30 ml of bone marrow was obtained. The incision was closed routinely.

The jugular vein was clipped and aseptically prepared. An 18-gauge butterfly needle was used to collect 50 ml of blood into a 60 ml syringe that contained 10 ml of anticoagulant citrate dextrose solution A (ACD-A).

Bone marrow was processed using a validated system according to manufacturer’s protocols (BMC Stem Cell Kit, Companion Regenerative Therapies, Newark, DE, USA). Briefly, the 150–170-µm filter, 60-ml syringe, and concentrating device were primed with heparin 1,000 U/ml. The vaclock syringe containing the bone marrow was attached to the filter and injected through the filter to the empty 60 ml syringe. The bone marrow was then injected into the concentrating device, placed into the centrifuge (Executive Series Centrifuge II, Companion Regenerative Therapies, Newark, DE, USA), and spun for 10 min at 3,600 RPM. After centrifugation, a 60-ml syringe was used to aspirate the plasma layer. A 12-ml syringe was then used to aspirate 3 ml of BMAC. Approximately 30,000 MSCs were present in the initial 30-ml bone marrow sample (41).

Whole blood was processed using a validated system according to manufacturer’s protocols (PurePRP^® Kit, Companion Regenerative Therapies, Newark, DE, USA). Briefly, the syringe containing the blood and ACD-A was injected into the first concentrating device, placed into the centrifuge (Executive Series Centrifuge II, Companion Regenerative Therapies, Newark, DE, USA), and spun for 1 min at 3,600 RPM. After centrifugation, a 60-ml syringe was used to aspirate the platelet plasma suspension layer and was then injected into the second concentrating device. The concentrating device was placed into the centrifuge and spun for 5 min at 3,800 RPM. After centrifugation, a 60-ml syringe was used to aspirate the plasma until the aspirating disc reached the bottom of the tube. A 12-ml syringe was used to inject 10 ml of air into the device to allow for re-suspension of the platelets at the base of the device. The device was swirled to re-suspend the platelets into the remaining plasma. The 12-ml syringe was then used to aspirate 4 ml of PRP.

To confirm the PRP product composition, a baseline blood and PRP product WBC differential, RBC concentration, and platelet concentration were obtained on all dogs using an in-house hematology analyzer that had been calibrated according to manufacturer standards (Idexx LaserCyte Hematology Analyzer, Idexx Laboratories, Westbrook, ME, USA). The PRP obtained a sixfold to sevenfold increase in platelets, 85% reduction in neutrophils, and 95% reduction in red blood cells from whole blood. Platelet, neutrophil, and red blood cell counts were verified prior to use.

The abdomen was clipped and aseptically prepared. A 4-cm incision was made in midline in the cranial abdomen, and a 2-cm incision was made through the linea. The falciform ligament was identified. Approximately 20 g of falciform adipose tissue was placed into a sterile container with Dulbecco’s modified Eagle’s medium (DMEM) supplemented with fetal bovine serum, penicillin, and streptomycin. The incision was closed routinely.

The jugular vein was clipped and aseptically prepared. An 18-gauge butterfly needle was used to collect 25 ml of blood into a 30 ml syringe that contained 5 ml of citrate phosphate dextrose adenine anticoagulant.

Adipose and blood samples were shipped in validated shipping boxes at 4°C to the Virginia Tech Marion DuPont Scott Equine Medical Center Regenerative Medicine Service for processing. ADPC and ACS were processed using their proprietary protocols. Briefly, adipose was mechanically and enzymatically separated to release ADPCs. The cells were then washed in phosphate-buffered saline (PBS), and nucleated cells were counted and cultured in stem cell media at 37°C in a 5% CO₂ incubator with 95% humidity. Cells were washed in PBS once at 70% confluency, trypsinized, and suspended in PRP. Approximately 5 million ADPCs were present in the sample, which were verified prior to suspension. PRP was prepared via centrifugation to obtain a fourfold increase in platelets and 80% reduction in white blood cells from whole blood. Platelet and white blood cells counts were verified prior to use. The ADPC–PRP was shipped to VOSM in validated shipping boxes at 4°C for subsequent injection.

Bone marrow aspirate concentrate and PRP were drawn into an empty syringe in a 1:1 ratio via a three-way stopcock. The stifle was clipped (if ADPC–PRP) and aseptically prepared. A 22-gauge needle was placed into the lateral compartment of the stifle and synovial fluid was aspirated using a 5-ml syringe to ensure the needle was in the joint. Approximately 2–4 ml of BMAC–PRP or 1–2 ml of ADPC–PRP was injected intra-articularly into the stifle; volume depended on the size of the dog. If the dog was below 15 kg in weight, they received 2 ml of BMAC–PRP or 1 ml of ADPC–PRP. If the dog was above 15 kg in weight, they received 4 ml of BMAC–PRP or 2 ml of ADPC–PRP. All dogs were placed into a soft-padded bandage for 12–24 h. Once the soft-padded bandage was removed, dogs were placed into a custom stifle orthotic that had been previously molded and constructed (Animal Orthocare, LLC, Chantilly, VA, USA) if the owner was concerned about their ability to confine and control their dog or if increased laxity had been noted on palpation at the initial examination. All owners were instructed to discontinue NSAIDs, corticosteroids, or anti-inflammatory supplements at least 2 weeks prior to injection. Owners were given the option to choose between BMAC–PRP and ADPC–PRP. The option to bank culture-expanded ADPCs and blood at the Virginia Tech Marion DuPont Scott Equine Medical Center was conveyed to the owner, while BMAC would not typically be banked. BMAC–PRP treatment would be performed the same day as stifle arthroscopy, while ADPC–PRP treatment would occur approximately 2 weeks after being sent for culture, which may play into the decision for owners that traveled a long distance to VOSM. Following the collection procedure and/or injection, patients were placed on tramadol (3–4 mg/kg by mouth every 8 h) or codeine (1–2 mg/kg by mouth every 8 h) for 3 days. The cost of each treatment was equal.

All patients were instructed to enroll in formal rehabilitation therapy following treatment consisting of once weekly manual/massage therapy, class IIIb low-level laser therapy for the affected stifle (5 J/cm²) and a twice daily at home exercise program for the first 8 weeks. NSAIDs, corticosteroids, anti-inflammatory supplements, class IV low level laser therapy, therapeutic ultrasound, hydrotherapy, and TENS/NMES were not permitted for the first 8 weeks posttreatment as their effect on stem cell and PRP therapy is still not fully known.

An owner questionnaire, including the validated Helsinki chronic pain index (HCPI), was e-mailed to all owners whose dog was known to not have had surgical repair following treatment. The questionnaire inquired about duration of lameness following treatment, if their performance/sporting dog returned to sport, and if so at what level compared to prior to injury. To assess the dog’s current well-being, owners were asked to rate their dog’s quality of life in the last 7 days (excellent, very good, good, fair, poor) and were asked their opinion of their dog’s procedural outcome (excellent, good, fair, poor). Possible chronic pain was assessed with the validated HCPI that contained questions on the dog’s mood, lameness, and willingness to move, play, and jump. The HCPI contained 11 questions whose answers were later given a value (0–4), giving a total index score range of 0–44. Individual question scores of 0 and 1 indicate normal behavior and movement, where scores of 2, 3, and 4 indicate pain with increasing severity. Therefore, healthy dogs usually have an HCPI between 0 and 11, and dogs with chronic pain will have a score of 12–44. The HCPI has been validated, reliability tested, and responsiveness tested (42, 43) (Table 1).

Data analyses were performed using a statistical software program (GraphPad Prism version 6.00, GraphPad Software, La Jolla, CA, USA). Gait analysis data were analyzed using a paired t-test that accounts for correlations between paired observations. A paired t-test will result in increased statistical power to detect a treatment effect when compared to an independent t-test. Outcome versus percent tear contingency data were analyzed using the Chi-square test for trend. Outcome versus use of a custom stifle orthotic contingency data were analyzed using the Fisher’s exact test. Outcome versus type of treatment (BMAC–PRP or ADPC–PRP) received contingency data were analyzed using a Fisher’s exact test. All the results were considered significant at p < 0.05.

In accordance with AAALAC International Rules of Accreditation, this study was performed with the approval of the VOSM Research Committee and with owner’s consent. All clients volunteered their dog for the study and provided written consent as required by Veterinary Orthopedic and Sports Medicine Group for every study participant. Medical records of 36 dogs diagnosed with an early partial (≤50%) CCL tear that were treated with BMSC–PRP or ADPC–PRP were reviewed. Ages ranged from 1 year to 9.5 years (mean 4.8 years, median 4.7 years), with 18 dogs being female (6 intact) and 18 dogs being male (12 intact). Body weights ranged from 4.8 to 60.5 kg (mean 25.9 kg, median 22.6 kg). Breeds included 10 (27.8%) Border Collies, 6 (16.7%) Labrador Retrievers, 3 (8.3%) Australian Shepherds, 2 (5.6%) Golden Retrievers, 2 (5.6%) Pembroke Welsh Corgis, 2 (5.6%) Shelties, and 1 (2.8%) of each of the following: Akbash, Basset Hound, Bull Terrier, Dutch Shepherd, German Shorthaired Pointer, Giant Schnauzer, Gordon Setter, Irish Setter, Keeshond, Mixed Breed, and Portuguese Water Dog. The majority of dogs were performance/sporting dogs [26 (72%) versus 10 (28%)] that were a companion dog only (Table 2). Duration of lameness ranged from 3 days to 4 years. Nineteen cases received BMAC–PRP for treatment, while 17 cases received ADPC–PRP. A custom stifle orthotic was used posttreatment in 14 dogs to prevent damage to the CCL during the recovery period.

Baseline orthopedic examination findings were available in 30 of the 36 dogs. A weight-bearing lameness and shortened stride was noted in 29 dogs, while 1 dog had a shortened stride only. On palpation, very mild effusion was present in the affected limb of 1 dog, mild effusion was present in 11 dogs, moderate effusion in 17 dogs, and 1 dog had no effusion. McMurray test for meniscal evaluation was negative in 29 dogs, while 1 dog was positive. Mild cranial drawer and tibial thrust were noted in flexion only in the affected limb in 23 dogs. Palpable evidence of stifle instability was absent in seven dogs.

Baseline radiograph findings were available in 30 of the 36 dogs. Very mild joint effusion was noted in the affected limb of 3 dogs, mild effusion in 12 dogs, and moderate effusion in 14 dogs. One dog had no effusion present on radiographs. Osteoarthritis was also present on radiographs in the affected limb of five dogs. Joint effusion was noted in the contralateral limb in 15 dogs, with osteoarthritis being present in 1 (39).

Both baseline and day 90 posttreatment objective gait analyses were available on 11 of the 36 dogs. Mean baseline TPI% was 17.7% (median 17.9%, range 13.1–22.8%) in the affected limb and 21.7% (median 21.5%, range 18.8–24.3%) in the unaffected limb. At 90 days posttreatment, mean TPI% was 20.1% (median 20.2%, range 16.9–22.8%) in the treated limb and 19.9% (median 19.2%, range 17.3–22.9%) in the contralateral limb. The TPI% of the treated limb was compared to the TPI% of the contralateral limb at baseline and at 90 days posttreatment. The TPI% of the treated limb at day 0 was also compared to the TPI% of the treated limb at 90 days posttreatment. A statistically significant difference was found between the treated limb TPI% at day 0 and day 90 (Figure 1; p = 0.0124), and between the treated limb and contralateral limb TPI% at day 0 (Figure 2; p = 0.0003). No significant difference was found between the treated limb and contralateral limb TPI% at day 90 (Figure 3; p = 0.7466).

Baseline diagnostic stifle arthroscopy findings were available on all 36 dogs. Diagnosis was confirmed and the CCL was found to be partially torn unilaterally in 35 dogs (13 left, 22 right), and bilaterally in 1 dog. The dog with bilateral partial CCL tears was omitted from the study, and the other 35 dogs were included in this study. A percent tear of the craniomedial band of the CCL was recorded for each dog. The high end was used to calculate descriptive statistics where a range was reported. Mild fraying of the craniomedial band of the CCL was reported in 4 of the 35 dogs. The percent tear of the craniomedial band of the CCL ranged from 5 to 50% (mean 20.8%, median 20%) in the remaining 31 dogs (Figure 4). The meniscus was found to be intact in 34 dogs, whereas 1 dog was found to have a bucket handle tear of the caudal pole of the medial meniscus, and a partial meniscectomy of the medial meniscus was performed. Degenerative joint disease (DJD) findings were recorded in 31 of the 35 dogs. Very mild DJD was reported in 2 dogs, mild DJD was reported in 4 dogs, moderate DJD in 2 dogs, and no DJD was found in the remaining 23 dogs.

Stifle arthroscopy findings at 90 days posttreatment were available on 13 of the 35 dogs. In 9 of the 13 dogs, a fully intact CCL with marked neovascularization and a normal fiber pattern was found with all previous regions of disruption healed. It appeared that the blood supply to the region of CCL injury was coming from the genicular artery (Figure 5). One dog was found to have significantly improved by way of an intact CCL with minor fraying, compared to a 25% tear of the craniomedial band of the CCL at baseline, and received an additional injection. The dog was found to have normal range of motion and no discomfort, tibial thrust, or lameness 90 days after the second injection. The remaining three dogs were found to have a >50% tear (either progressive since baseline or completely ruptured), and a TPLO was performed at VOSM. Four additional dogs were reported by the owner to have had a TPLO performed outside of VOSM.

Questionnaires were sent to all owners whose dog was known to not have had surgical repair following treatment (29 dogs). Questionnaires were returned for 12 dogs (41%), 6 of which had received ADPC–PRP and 6 of which had received BMAC–PRP. Follow-up time between surgery and the questionnaire ranged from 0.5 to 2.9 years (mean 1.8 years, median 1.9 years). HCPI ranged from 0 to 17 (mean 4.4, median 1, standard deviation 6.3). Owners were asked how long their dog experienced lameness following treatment; the answers ranged from 0 to 12 months (mean 2.4 months, median 1.3 months). All 12 respondents believed their dog had an excellent (10) or very good (2) quality of life in the last 7 days and rated their dog’s procedural outcome as excellent (11) or good (1). Of the 12 questionnaires returned, 8 were reported to be performance/sporting dogs. According to the owner, seven of the eight dogs had returned to sport; the remaining dog had just begun a return to sport conditioning program 6 months posttreatment. Of the seven dogs that returned to sport, five dogs were reported to be competing at the same level compared to prior to injury, and two dogs were competing at a greater level compared to prior to injury (including faster times, higher jump height, and dropping bars less). According to the previously established VOSM Return to Agility Grading Scale (Table 3), of the four dogs that competed in agility, two dogs (50%) were grade 4 and two dogs (50%) were grade 5. Of the seven dogs that returned to sport, it was found that the greatest amount of time a dog had been competing since their procedure was 2.5 years.

The seven reported failures were found to have had a 50% (1), 25% (1), 20% (1), and 10% (4) tear at baseline. There was no association found between outcome and the percent tear reported at baseline (p = 0.5545). A custom stifle orthotic was used posttreatment in 14 dogs, with 4 of the reported failures having used an orthotic, and 3 failures not having used an orthotic. There was no association found between outcome and the use of an orthotic posttreatment (p = 0.5545). There were four reported failures that received ADPC–PRP treatment and three failures that received BMAC–PRP treatment. There was no association found between outcome and the treatment type received (p = 0.6843). All seven dogs that failed went on to have a TPLO procedure.