Jg. Hauptman et al., EVALUATION OF THE SENSITIVITY AND SPECIFICITY OF DIAGNOSTIC-CRITERIA FOR SEPSIS IN DOGS, Veterinary surgery, 26(5), 1997, pp. 393-397
Objective-To evaluate the ability of various individual criteria and g
rouped criteria to diagnose sepsis in dogs. Study Design-Prospective a
cquisition 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 band
s, platelet count, and serum glucose concentration were obtained on da
y 0. True sepsis was determined on days 0 to 3 according to the follow
ing criteria: (1) histological, microbiological, and/or gross confirma
tion of infection, and (2) systemic illness caused by infection. Data
were analyzed by the Mann-Whitney U test and multiple logistic regress
ion. Sensitivity and specificity were calculated. Results-The mean tem
perature, heart rate, WBC count, and percent bands were greater, where
as the mean platelet count was less in septic compared with nonseptic
dogs. There was no difference in respiratory rate or glucose concentra
tion. WBC/bands were the best individual criterion for the diagnosis o
f sepsis (sensitivity 87%; specificity 69%). The sensitivity and speci
ficity of the grouped criteria (greater than or equal to two of four;
temperature, heart rate, respiratory rate, WBC) varied according to ra
nges of normal used. Muliple logistic regression resulted in little im
provement 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 se
nsitivity (eg, 97%). The high sensitivity was associated with a low fa
lse-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.