Concordance between qualitative and quantitative cultures in burned patients. Analysis of 2886 cultures

OBJECTIVE: To determine the concordance between superficial cultures (SC) and quantitative cultures (QC) in the diagnosis of wound infection in burn patients. METHOD: Sample: All SC and QC taken from the same patient, site and during the same surgery were analysed. Variables: On the SC, the microorganism (MO) and its amount defined subjectively by the microbiologist was recorded (negative, very low, low, regular and abundant). On the QC, the MO and its amount were expressed as colony forming units per gram of tissue (CFUs/g). Statistics: Kappa index of agreement beyond chance; Wilcoxon and Kruskall-Wallis for continuous variables and chi(2) for categorical variables were used with a p<0.05 indicating statistical significance. RESULTS: One thousand four hundred and forty three pairs of cultures were analyzed. The concordance between SC and QC (Kappa index) was 52%. On the SC, only when the microbiologist subjectively informed "abundant" MOs there was a significant difference (p<0.0001). There were 6.1% of QCs with more than 10(5) CFUs/g and the most frequent MOs isolated were: S. aureus (27.9%), E. coli (11.6%), P. aeruginosa (11.6%), E. faecalis (11.6%) and S. epidermidis (7.0%). CONCLUSIONS: SC has a moderate concordance with the QC showing a low reliability between the two methods. The subjective information given by the microbiology technician in the SC is not precise. A study in which the two methods be compared blindly against the reference standard, in a prospective cohort of patients, it is needed to discriminate which of two methods it is the most accurate one determining sensitivity and specificity.