Background: The clinical parameters for rbe suspicion of Clostridium diffic
ile infections, namely the use of antimicrobials and diarrhea, have a low p
redictive value for tbe diagnosis Aim: To search other clinical variables a
nd determine a clinical prediction model for (Clostridium difficile diarrhe
a. Patients and methods: All patients to whom a Clostridium difficile study
was requested, were prospectively studied during 5 months. Clinical variab
les of these patients were registered. The diagnosing of Clostridium diffic
ile was done using the cytotoxicity rest in fibroblast cultures. Results: N
inety two patients were analyzed and in 26, tbe diagnosis of Clostridium di
fficile was confirmed. A logistic regression model disclosed an age over 60
years old, the presence of mucus in the stools a; id a temperature over 37
.8 degreesC ii I the previous 24 h, as significant predictors of tbe infect
ion. The correlation of the model, between the predicted probability and th
e observed condition, was 81.5%. Conclusions: The presence of tbe clinical
variables identified in this study are associated with a high probability o
f an infection by Clostridium difficile in patients with diarrhea and the r
ecent use of antimicrobials (Rev Mid Ch de 2001; 129: 620-625).