The effect of blood contamination on equine synovial fluid analysis

Based on a systemic complete blood count and a synovial fluid sample, to develop a mathematical model enabling the approximation of corrected values for synovial fluid white blood cell (WBC) count, neutrophil percentage, and total protein concentration in samples of synovial fluid that were contaminated by blood.

Peripheral venous blood and synovial fluid samples were obtained from ten horses. A pooled synovial fluid sample from each horse was separated into 2 mL aliquots, which were subsequently contaminated with a known percentage of autogenous blood (0 to 50% of the synovial fluid volume). A complete blood count, packed cell volume, total protein (TP) concentration, and differential cytological examination were performed in all the synovial fluid and venous blood samples. Regression analysis was used to generate a model to calculate non-contaminated synovial WBC count, TP concentration and synovial neutrophil percentage. Using a further five horses these models were applied in blinded fashion to contaminated synovial fluid samples. Calculated values were compared to non-contaminated measured values.

Model results for synovial WBC count and TP concentration were not significantly different from measured values. Calculated neutrophil percentage of synovial fluid WBC was a mean of 6.6% higher than measured values and was significantly different. There was no effect of the severity of contamination (as a percentage of volume) on the ability of the models to predict any of the outcome variables.

It is possible to calculate non-contaminated synovial fluid WBC and TP values, but not neutrophil percentage, from heavily contaminated samples. Further study would allow for improved prediction, validation and extrapolation to a wider horse population.