Dogs were eligible for inclusion if they were between 1 and 10 years of age, 16 kg and 60 kg body weight, and could be appropriately treated with a 3.5 mm small, standard, broad, or jumbo sized TPLO plate. Exclusion criteria included the use of ESWT or photobiomodulation during the 2 months preceding enrollment, use of bedinvetmab (Librela; Zoetis Inc, Kalamazoo, MI), use of polysulfated glycosaminoglycan supplementation (Adequan; American Regent Inc, Shirley, NY) in the 6 weeks prior to enrollment, or previous intra-articular joint injections. All dogs needed to have ceased receiving any oral analgesics such as gabapentin or oral non-steroidal anti-inflammatory drugs (NSAIDs) at least 2 days prior to enrollment and acquisition of baseline data. Exclusion criteria also included any systemic illness, neurologic disease, or musculoskeletal pathology, such as contralateral CCL disease, as based upon physical examination or radiography. Dogs that had the contralateral stifle treated with a TPLO a minimum of 6 months prior were eligible for enrollment.

Owners completed both the canine brief pain inventory (CBPI) (10) and liverpool osteoarthritis for dogs (LOAD) (11) questionnaire pre-operatively and at 2, 4 and 8 weeks post-operatively. For consistency, the same owner completed the surveys at each visit. If the owner who originally completed the surveys was unavailable to attend the recheck appointment, the surveys were emailed to the participating owner to be filled out within 48 h of the recheck appointment.

Gait analysis for each patient was performed on the day of surgery (preceding surgery) and at 4 and 8 weeks post-operatively. Gait data collection was performed by the same handler (SLC), using a GAITFour pressure sensitive walkway system (12) (Gait4Dog, CIR Systems Inc., Franklin, NJ). Cameras capturing video of each trial were placed at both ends of the walkway system to assess position of the head and body in both directions of travel. Gait data were collected consecutively until a minimum of 5, valid-appearing trials at a walk, and 5 valid-appearing trials at trot, were completed. Processing of the gait data was subsequently performed after the appointment was completed. Gait trials were ultimately considered valid if video footage included the patient moving toward a camera and such video confirmed the head and body were forward facing with no turning of the head and no pulling on the leash for the entirety of the trial. In addition, in order to be considered valid, the velocity needed to have <10% variation and at least 3 complete gait cycles. If the dog was non-weight bearing for the trial a value of 0 was recorded. Once processing was complete, the total pressure index (TPI) for the affected hindlimb from each valid trial as well as the velocity were exported into an excel spreadsheet.

All patients were premedicated with hydromorphone (0.1 mg/kg intramuscular) and acepromazine (0.05 mg/kg intramuscular) with propofol (4 mg/kg intravenous) given to effect as the induction agent. The protocol was changed for the last patient due to supply issues when morphine (0.5 mg/kg intramuscular) was used as the premedication agent in place of hydromorphone. All patients received peri-operative cefazolin (22 mg/kg intravenously) given at the time of anesthetic induction and every 90 min thereafter until inhalant anesthesia was discontinued. The anesthesia for all patients was maintained using isoflurane. Each patient received a femoral and sciatic nerve block consisting of bupivacaine (1.5 mg/kg, 50 mg/ml) and dexmedetomidine (0.001 mg/kg; 0.5 mg/ml) prior to surgery using a nerve stimulator. The bupivacaine and dexmedetomidine were mixed and half the total volume administered to the femoral nerve and half to the sciatic nerve. Pre-operative orthogonal radiographs of the stifle were taken under anesthesia for surgical planning purposes. Standard subpatellar arthroscopic portals were made medial and lateral to the patella tendon. The stifle joint of each patient was evaluated with a 2.5 or 2.7 mm 30° arthroscope. The cranial cruciate ligament (CCL) was classified as a complete rupture, a partial incompetent rupture, or a partial competent rupture. Completely ruptured and partial incompetent ruptures were completely debrided and lumped together as “complete ruptures” for the purposes of stratification (see below). Partial competent CCL tears were left intact with debridement of the torn fibers only. The medial meniscus was probed and assessed for pathology. If the menisci were normal they were left untreated. If there was meniscal pathology that necessitated meniscectomy, either partial meniscectomy plus meniscal release, hemi-meniscectomy, or full thickness meniscal release at the caudomedial meniscotibial ligament was performed. Hence, all stifles could be accurately classified as having an intact and normal meniscus or effectively being “meniscal deficient”. A TPLO was then performed using an approach to the proximomedial tibia with elevation of the pes anserinus muscle group. Limited elevation of the popliteus muscle caudally was performed, but no gauze or other material was placed caudally. No elevation of the craniolateral musculature was performed. A jig was not applied. Templating and performance of the TPLO was performed with the center of the osteotomy located at the base of the intercondylar eminences. Rotation of the proximal segment was performed using a rotational pin and with a target final tibial plateau angle of 5°. All patients had locking TPLO plates applied using a combination of locking screws applied proximal to the osteotomy and non-locking screws applied distal to the osteotomy. Closure of the caudal belly of the sartorius muscle often included advancement and placement of one or two suture passes through the patella tendon for security of the closure. Subcutaneous and skin closure was routine.

Immediately following arthroscopy dogs were allocated to one of three different strata based upon the status of the cranial cruciate ligament and meniscus: (A) complete CCL deficiency with an intact medial meniscus, (B) complete CCL deficiency and meniscal deficiency, or (C) partial competent CCL rupture with an intact medial meniscus; there were no dogs with a partial competent CCL rupture and meniscal deficiency. Within each stratum patients were randomly allocated to either ESWT or sham treatment based on a randomization order created using a randomization tool (Sealed Envelope Ltd. 2024. Create a blocked randomization list. Available from: https://www.sealedenvelope.com/simple-randomiser/v1/lists).

Immediately following surgery and acquisition of post-operative radiographs, ESWT, or sham treatment, was performed using the novel piezoelectric shockwave device (CS-Pro VET device, Curative Sound, Carmel, IN; Figure 1) with the patient under anesthesia. Treatment included 5 areas: the craniomedial and caudomedial proximal tibia just cranial to and just caudal to the TPLO plate, the craniolateral and caudolateral proximal tibia, and the patellar tendon from the distal patella to the level of insertion on the tibial tuberosity. The 4 lateral and medial areas of the proximal tibia received 1,200 shockwaves per area and the patellar tendon received 2,400 shockwaves. A pre-sterilized standoff, attached to the ESWT device by magnetism, was used in each area to focus the depth of the shockwave. The standoff for the medial areas was 10 mm with an energy flux density of 0.17 mj/mm², the standoff for the lateral areas was 30 mm with an energy flux density of 0.22 mj/mm² and the patellar tendon standoff was 5 mm with an energy flux density of 0.14 mj/mm². The device with the appropriate standoff was moved in small circles over the region of interest and ultrasound gel was used to facilitate shockwave transmission. Dogs in the sham treatment group received the same protocol without the battery included in the device. All treatments were performed by the same investigator (SLC).

Extracorporeal shockwave treatment, or sham treatment, was repeated at the 2 and 4 week recheck after suture removal or gait analysis were performed, respectively. Dogs were not sedated for treatment. At the 4 week recheck, although potentially unnecessary for adequate energy transmission, each patient was re-shaved at the level of the proximal tibia and patella tendon to maintain consistency among all patients in terms of a lack of hair.

Orthogonal radiographs were made 8 weeks post-operatively under sedation (dexmedetomidine 0.005 mg/kg and butorphanol 0.1 mg/kg). Following completion of the last 8-week recheck, all radiographs made immediately pre-operatively and at 8 weeks post-operatively were anonymized and then evaluated by one investigator (SPF) blinded to the treatment group. Assessment for bone healing was made based upon the 8 week radiographs using a modified radiographic union scale for tibial fractures (mRUST) that has been assessed for validity in scoring tibial healing following TPLO (13). Patella tendon thickness was measured immediately pre-operatively and on the 8 week recheck radiographs. Thickness was measured at the distal quarter of the tendon's length as this was the one location in which a significant difference was previously identified following ESWT in conjunction with TPLO (3).

All patients received post-operative analgesia consisting of carprofen (2.2 mg/kg twice daily orally) and gabapentin (10 mg/kg three times daily orally) each given for 2 weeks. A 7-day course of antibiotics (cephalexin 22 mg/kg three times daily orally) was prescribed as well as trazodone (4–8 mg/kg up to 3 times daily orally) for sedation which could be administered as needed throughout the 8 week recovery period. Indoor activity restriction recommendations included confinement to a designated safe room, pen, crate, or on-leash in the home during the first 8 weeks of recovery. Recommendations for outdoor activity involved a slow increase in leashed walks over 8 weeks beginning with 5–10 min walks during weeks 1–4 and 15–20 min walks during weeks 5–8. Physical rehabilitation was prohibited until patients completed the study.

Diaries were provided to owners for the recording of medication usage and adverse events on the day of surgery. That diary was collected at the 2 week recheck and a new diary was issued; this process was performed iteratively at each recheck until the end of the study. Owners were also contacted by phone at weeks 1, 3 and 7 post-operatively to discuss activity levels and inquire if there had been any adverse events.

A linear mixed model was used to analyze the TPI obtained using the pressure sensitive walkway. The 5 walks and 5 trots with the smallest range in velocity (for each gait) were selected and included in the analysis based on recommendations to use narrow velocity ranges with limited/controlled acceleration in order to minimize data variance (14). The statistical model included fixed factors for treatment, time, strata, walk/trot, and a random intercept for dog. Additionally, the model for TPI included all two- and three-way interactions and one four-way interaction. Linear mixed models were also used to analyze LOAD and CBPI scores with fixed factors for treatment, time, strata and all two and one three-way interaction as well as a baseline covariate and a random intercept for dog.

A Mann-Whitney test was used to compare mRUST scores between treatment groups. An analysis of covariance was used to compare patella tendon thickness between groups with factors for treatment, strata and treatment by strata interaction and a covariate for baseline patella tendon thickness. p-values of < 0.05 were considered statistically significant.

There was a statistically significant difference in TPI between the ESWT and sham treatment groups at a trot 4 weeks post-operatively with greater weight bearing in the ESWT group (p = 0.04). This difference in weight bearing at a trot was no longer statistically significant by 8 weeks post-operatively. There were no statistically significant differences at 4 and 8 weeks post-operatively at a walk. Data on TPIs are shown in Figure 2.

There were no statistically significant effects of ESWT treatment on bone healing scores (p = 0.34) or measures of patella tendon thickness (p = 0.36). The mRUST scores and measures of patella tendon thickness are shown in Tables 2, 3 below respectively.