Evaluation of the sensitivity and specificity of diagnostic criteria for sepsis in dogs

Objective-To evaluate the ability of various individual criteria and grouped criteria to diagnose sepsis in dogs. Study Design-Prospective acquisition of clinical data. Animals or Sample Population-Client-owned dogs; 30 septic and 320 nonseptic. Methods-Rectal temperature, heart rate, respiratory rate, white blood cell (WBC) count with percent bands, platelet count, and serum glucose concentration were obtained on day 0. True sepsis was determined on days 0 to 3 according to the following criteria: (1) histological, microbiological, and/or gross confirmation of infection, and (2) systemic illness caused by infection. Data were analyzed by the Mann-Whitney U test and multiple logistic regression. Sensitivity and specificity were calculated. Results-The mean temperature, heart rate, WBC count, and percent bands were greater, whereas the mean platelet count was less in septic compared with nonseptic dogs. There was no difference in respiratory rate or glucose concentration. WBC/bands were the best individual criterion for the diagnosis of sepsis (sensitivity 87%; specificity 69%). The sensitivity and specificity of the grouped criteria (greater than or equal to two of four; temperature, heart rate, respiratory rate, WBC) varied according to ranges of normal used. Muliple logistic regression resulted in little improvement in the sensitivity/specificity of these diagnostic criteria for the diagnosis of sepsis. Conclusions-These criteria are useful for the diagnosis of sepsis when limits are used that result in a high sensitivity (eg, 97%). The high sensitivity was associated with a low false-negative and a high false-positive rate; sepsis was overdiagnosed with these grouped criteria. Clinical Relevance-These criteria may be used for a sensitive, but nonspecific, diagnosis of sepsis in dogs. (C) Copyright 1997 by The American College of Veterinary Surgeons.