To compare agreement between a portable clinical analyzer and laboratory-based bench-top analyzer for analysis of pH, lactate, and glucose concentrations in synovial fluid.

Study Design

Prospective experimental study.


Clinically normal horses (n=8); 6 horses euthanatized for reasons unrelated to the study; 11 horses that had synoviocentesis for reasons other than sepsis; 7 horses that had synoviocentesis for evaluation of sepsis; and 2 horses without recorded clinical data. Median age of horses was 8 years (range, 1 day to 24 years).


Supernatant from each synovial fluid sample was analyzed for pH, lactate, and glucose concentrations using an ABL 705 laboratory-based bench-top analyzer and i-STAT portable clinical analyzer. Bland–Altman plots were constructed and concordance analysis performed to determine bias and agreement between the 2 analyzers.


There was acceptable agreement between analyzers for lactate and glucose concentrations, with biases of 0.198 mmol/L and 9 mg/dL and concordance correlation coefficients of 0.97 and 0.96 for lactate and glucose, respectively. The agreement between analyzers for pH was not acceptable, with a bias of −0.057 and concordance correlation coefficient of 0.89.


This study found that the portable clinical analyzer performed similarly to the bench-top blood gas analyzer for evaluation of lactate and glucose concentrations, but not pH, in synovial fluid.