C-reactive protein concentration as a prognostic factor for inflammation in the management of odontogenic infections.

Our aim was to find out if it is possible to correlate the duration of stay in hospital, the severity of infection, involvement of particular anatomical spaces, white cell count, efficacy of surgical treatment, and fever with C-reactive protein (CRP) concentrations on admission. One hundred patients met our inclusion criteria. After their notes had been examined they were subdivided according to whether the infection of the main facial space involved was less severe, moderately severe, or very severe. The relations between degree of severity and CRP concentration on admission (<100mg/L compared with 100+), age (years), sex, and duration of hospital stay (days) were examined using Poisson regression (because the distribution of characteristics, and particularly the duration of stay, were skewed). The overall model was significant (p=0.003). Pearson and deviance chi square tests did not indicate overdispersion (p=0.97 in both cases), which suggested that the assumptions about the Poisson distribution were valid. Log-rank chi square tests indicated that only severity had a significant effect (p=0.0001), and C-reactive protein concentration was not significantly associated with group on admission, age, or sex. The moderately and very severe groups had longer median (range) durations of stay than the less severe group (5 (2-8) compared with 3 (1-8) days, respectively). CRP concentration was not a prognostic factor for the extent of odontogenic infections or presumed duration of stay, but severity scoring was a significant factor in the prediction of duration of stay in hospital.