The study used a blinded 2-period, 2-treatment, 2 joint non-crossover experimental design (Figure 1). Each study period was defined as 1 week, followed by a 2-week washout between study periods. Each horse was randomly assigned to either control (0.9% sodium chloride) or treatment (alloPRP) group. For each horse, one radiocarpal joint (RCJ) and one tarsocrural joint (TCJ) were randomly assigned to treatment. One joint was treated and evaluated per study period. Treatment sequence and limbs chosen for treatment were randomized for each horse. The individual preparing and performing the joint treatments (SK) was not the same individual performing the evaluations post-treatment to maintain blinding (LB).

The study protocol was reviewed and approved by the Auburn University Institutional Animal Care and Use committee (IACUC # 2023-5204). Twelve horses (4 Thoroughbreds, 5 Quarter Horses, 2 Draft crosses, and 1 Warmblood) were selected from the university research heard for study enrollment. The ages of horses ranged from 8 to 18 years (mean, 12.5 years). Horses were assessed for suitability of study inclusion. Assessment included a complete physical and musculoskeletal examination. Included horses were not free of baseline lameness due to the nature of horses obtained and maintained within the University research herd. Horses were included in the study if there were no radiographic abnormalities of the radiocarpal and tarsocrural joints as evaluated by a board-certified veterinary radiologist (RC). A complete blood count (CBC) and chemistry panel were performed for each horse at the beginning of the study to rule out systemic disease.

The product used, PrecisePRP™ Equine (VetStem, Poway, CA), contains a leucoreduced, allogeneic, pooled platelet-rich plasma as the active agent rehydrated with 8 ml of sterile saline. The intended clinical dose from the manufacturer is 4 ml per joint and was the dose used for this study. Platelet rich plasma was obtained via apheresis and pooled from 18 healthy donors that are part of a USDA licensed herd. The final product contained an average of 400,000 platelets/μl and 1,500 white blood cells/μl. The final product also contained an average of 804 pg/ml of platelet derived growth factor (PDFG-BB). The product was tested for quality control including testing for platelet surface markers (cluster of differentiation 9 and 61), sterility, mycoplasma and endotoxin. Product used for the study was from the same lot of manufacturing (Lot # QB1617C).

Each horse was sedated with xylazine hydrochloride (0.4–0.5 mg/kg) given intravenously. The randomly assigned RCJ or TCJ was treated with either 4 ml of 0.9% sodium chloride or alloPRP for each study period according to sequence randomization. The joint assigned was aseptically prepared with alternating application of chlorhexidine scrub and 70% isopropyl alcohol for 5 min. The TCJ was sampled and injected from the dorsomedial approach. The RCJ was sampled and injected from the traditional dorsolateral approach with the carpus maintained in flexion during synoviocentesis. Synovial fluid (3 ml) was obtained immediately prior to administration of assigned treatment. Samples were transferred to a collection tube containing EDTA and stored at −80°C until analyzed.

If the horse was assigned to receive the alloPRP, it was prepared in a sterile manner. A vented clave was applied to the top of the vial containing the freeze-dried PRP product. The vial was held in a tilted fashion and a total of 8 ml of sterile water, per the company reconstitution directions, was administered slowly through the clave hub, flowing down the side of the vial for rehydration. The fluid vial was gently swirled to avoid excess foaming while reconstituting the product. Once all of the powder appeared dissolved, 4 ml of reconstituted alloPRP was drawn into a sterile syringe through the clave for injection into the assigned joint.

Horses were individually housed in a stall and monitored daily during the sampling period and for an additional 12-h following collection of the post-treatment synovial fluid samples. Horses were monitored throughout the study period for changes in temperature, pulse, and respiration. Parameters were monitored just prior to treatment (T0), every 12 h for 96 h and then every 24 h for 7 days to conclude the study period.

The RCJ and TCJ were subjectively scored by digital palpation for heat, joint effusion, and response to flexion at T0h, T2h T6h, T12h, T24h and then every 24 h until the end of the study period. Heat was graded by palpation of the treated joint and graded from 0 to 3 (0 = none, 1 = minor, 2 = moderate, and 3 = severe). Joint swelling was graded from 0 to 4 (0 = no swelling; 1 = minimal swelling localized to the injection site; 2 = mild swelling localized to the RCJ/TCJ; 3 = moderate swelling extending to encompass the carpus or tarsus; and 4 = marked swelling extending above or below the carpus or tarsus). Joint circumference (mm) was measured at the midline of the injected joint using a standard cloth measuring tape, with the location of measurement marked medial and laterally with horizontal clip markings to maintain consistency.

Response to passive flexion was subjectively graded from 0 to 3 (0 = none; 1 = minor; 2 = moderate; 3 = severe). Lameness evaluations were performed on each horse at T0, T12h and then every 24 h until the end of the study period Objective lameness assessments were performed with a body-mounted intertial sensor system (Equinosis Q with Lameness Locator®, Columbia, MO). Horses were trotted on a straight line by a handler on the same firm surface for a minimum of 25 strides.

On day 7 (T168h) post-treatment, horses were sedated with xylazine hydrochloride (0.2–0.5 mg/kg IV) following their lameness evaluation. The previously treated joints were aseptically prepared. Approximately 3–5 ml of synovial fluid was collected in a sterile manner. Samples were then transferred to a collection tube containing EDTA and stored at −80°C until analyses.

For the complete blood count and serum blood biochemistry, blood was obtained from the left or right jugular vein and submitted to an outside laboratory (Idexx BioAnalytics, West Sacramento, CA) for complete blood count and biochemistry profile on T1d, T7d, T14d, and T21d by board-certified pathologists.

For synovial fluid analysis, approximately 1 ml of synovial fluid was aliquoted from the parent sample and shipped overnight on ice for analysis using the same outside laboratory as for the CBC and chemistry. Measured parameters included total nucleated cell count (TNCC), total red blood cells (RBC), and total protein (TP) (Siemens Healthcare Diagnostics, Erlangen, Germany). Differential cell counts were performed by board-certified veterinary clinical pathologists. The remaining synovial fluid was aliquoted into Eppendorf tubes, centrifuged (600 RPM for 10 min at 4°C), supernatant removed and stored at −80°C until further analysis.

For synovial fluid cytokine analysis, thawed synovial fluid samples (200 μl) were hyaluronidase-digested (10 μl of 100 IU hyaluronidase/ml acetate buffer; Worthington Biochemical Corporation, Lakewood, NJ) for 30 min at 37°C, centrifuged (12,000 RPM for 10 min; 4°C), and supernatant recovered. TNFα and IL-1 receptor antagonist protein (IL-1raP) were quantified by ELISA (R&D Systems, Minneapolis, MN).

All analyses were performed using SAS 9.4 (Cary, NC). A significance level of 0.05 was used. The assumption of normality was evaluated via inspection of QQ- and PP-plots, histograms, and skewness. Normally distributed variables were summarized descriptively with mean and standard deviation (SD). Non-normally distributed variables were summarized descriptively with median and interquartile range (IQR).

Linear mixed models (LMM) were used to test effects of treatment on temperature, CBC and chemical parameters, and synovial fluid RBC. Negative binomial generalized linear mixed models (GLMM) with a negative binomial distribution and a log link function were used to test for effects of treatment on pulse and respiratory rates. Ordinal logistic GLMMs with a multinomial distribution and a cumulative logit link function were used to test effects of treatment on joint effusion, heat, and lameness scores. Logistic GLMM with a binomial distribution and a logit link function were used to test for effects of treatment on passive flexion. These models had a fixed factor for treatment, a study day covariate and a study day by treatment interaction factor and a random intercept for each horse and for each limb within each horse. LMM assumptions (normality and homoscedasticity of model residuals) were evaluated via inspection of conditional QQ-plots, histograms, and residual plots. Satterthwaite degrees of freedom method were used in all models. REML estimation for LMM and residual pseudo-likelihood for GLMMs were used. p-values from the study day by treatment interaction effect were reported to test for differences in change in endpoint over time between treatment and control.

Vector sum (VS) data obtained via Lameness Locator® were not-normally distributed so Mann–Whitney tests were used to compare VS measurements between treatments on each study day for treated or untreated limbs separately.

All horses, except for one in the control group (Horse 1), maintained heart rates, respiratory rates, and rectal temperatures within normal limits throughout the study period. There were no differences in change over time between treatment and control groups for heart rate (p = 0.33), respiration rate (p = 0.60), or temperature (p = 0.90). Horse 1 experienced an acute episode of colic, presumptively due to an ileal impaction that was treated medically. This resulted in a temporary elevation in heart rate (60 bpm), that resolved within 6 h following medical management.

There were no differences between groups in the change of joint effusion score up to and including day 1 (p = 0.40) and days 2–7 (p = 0.73). For joint circumference, there was no difference between groups in the change in joint circumference from day 0 to day 7 (p = 0.55). There was no difference between treatments in change of heat scores up to and including day 1 or days 2–7 (p = 0.09). There was no difference in passive flexion score between groups (p = 0.70). For joint evaluations, there was a trend for higher scores at 6–12 h post-injection with no significant differences present between groups. The control group displayed greater joint swelling in the first 6–12 h after first injection whereas the treatment group experienced joint swelling 6–12 h after the second injection which resolved over time.

Baseline lameness evaluations revealed that 2/12 (16.7%) horses displayed a right forelimb lameness, 7/12 (58.3%) displayed a left forelimb lameness, 7/12 (58.3%) horses displayed a right hind limb lameness, and 5/12 (41.7%) horses displayed a left hind limb lameness. A total of 9/12 (75%) horses were lame in both a forelimb and a hind limb on baseline evaluations (3 were LF/LH, 4 were LF/RH, 1 was RF/LH, and 1 was RF/RH). The mean lameness grade for all horses was a 2.14 out of 5 following the AAEP Lameness Grading Scale. There were no differences in subjective lameness scores on any study day in treated (p = 0.47) or control (p = 0.31) limbs (Figures 2A,B). There were no differences in VS (Figure 3) or change in VS from baseline (Figure 4) between treatment groups for treated and untreated limbs on any study day for either group. The lameness data is charted for each individual horse through the study period in Figure 5.

No clinically significant abnormalities were present on the CBC and chemistry profiles for all horses prior to treatment. When comparing CBC data, there were no significant changes in total monocyte count (p = 0.24), total neutrophil count (p = 0.06), neutrophil percentage (p = 0.67), or total WBC count (p = 0.05) between day T0 and T7d between groups. There was a significant change in monocyte percentage from day T0 to day T7d between groups where the saline group had a greater decline in monocyte percentage from day T0 to day T7d compared to the treatment group (p = 0.004). There were no differences for changes in total monocyte count (p = 0.12), monocyte percentage (p = 0.28), total neutrophil count (p = 0.27), neutrophil percentage (p = 0.957), and total WBC count (p = 0.17) from T0 to T7d between groups. When evaluating the blood chemistry, there were no differences observed in the change from day T0 to day T7d between groups for blood glucose (p = 0.34), blood urea nitrogen (p = 0.5), creatinine (p = 0.32), total protein (p = 0.38), alkaline phosphatase (p = 0.49), alanine transaminase (p = 0.47), and creatinine kinase (p = 0.82). No differences between T0h and T7d within groups observed for blood glucose (p = 0.62), blood urea nitrogen (p = 0.43), creatinine (p = 0.41), total protein (p = 0.85), alkaline phosphatase (p = 0.44), alanine transaminase (p = 0.55), and creatinine kinase (p = 0.13).

No differences in synovial fluid clarity and color between T0 and T7d were observed. Samples were predominantly straw-colored and clear to mildly hazy throughout the study period. Analyses of synovial fluid revealed no differences in RBC concentration between time points (p = 0.82) or between groups (p = 0.80). No differences in TNCC were present between time points (p = 0.16) or between groups (p = 0.80). Similarly, no differences in total TP were present between groups (p = 0.94) or time points (p = 0.29).

No differences in synovial fluid concentrations of TNFα were observed between groups (p = 0.45) or within groups between T0h and T7d (p = 0.32). Similarly, no differences in synovial fluid concentrations of IL-1raP were observed between groups (p = 0.18) and within groups between T0h and T7d (p = 0.61).