Using glucose levels to detect septic belly infection in dogs

Abstract

OBJECTIVE: To evaluate the usefulness of a veterinary point-of-care glucometer for identification of septic peritonitis in dogs with peritoneal effusion (PE). DESIGN: Prospective clinical evaluation. ANIMALS: 39 dogs with PE. PROCEDURES: Blood and peritoneal fluid convenience samples were collected concurrently in all dogs at the time of initial evaluation. A veterinary point-of-care glucometer was used to measure glucose concentration in heparinized whole blood, plasma, peritoneal fluid, and peritoneal fluid supernatant samples. Seventeen dogs had confirmed septic peritonitis, and 22 dogs had nonseptic PE. Sensitivity, specificity, positive and negative predictive values, and accuracy of identification of dogs with septic peritonitis were calculated for glucose concentration differences for whole blood versus peritoneal fluid (WB-PF), plasma versus peritoneal fluid (P-PF), and plasma versus peritoneal fluid supernatant (P-PFS). RESULTS: With a cutoff of > 20 mg/dL, the glucose concentration difference for WB-PF was an insensitive indicator of septic peritonitis (sensitivity, 41.2%; specificity, 100%). In comparison, the glucose concentration differences for P-PF and P-PFS had a higher sensitivity for septic peritonitis (88.2% and 82.4%, respectively) but a lower specificity (80% and 77.8%, respectively). With a glucose concentration difference cutoff of ≥ 38 mg/dL, specificity, positive predictive value, and accuracy of P-PF and P-PFS improved. CONCLUSIONS AND CLINICAL RELEVANCE: Determination of the glucose concentration difference for WB-PF with the veterinary point-of-care glucometer was not useful in identifying all dogs with septic peritonitis. A glucose concentration difference of ≥ 38 mg/dL for P-PF or P-PFS, however, supported an accurate diagnosis of septic peritonitis in dogs with PE.