ABSTRACT
CONTEXT: Surface swab cultures have attracted attention as a potential alternative to biopsy histology or quantitative culture methods for microbiological burn wound monitoring. To our knowledge, the utility of adding a Gram-stained slide in this context has not been evaluated previously. OBJECTIVE: To determine the degree of correlation of Gram stain with culture for the microbiological analysis of burn wound surfaces. DESIGN: Prospective laboratory analysis. SETTING: Urban health region/centralized diagnostic microbiology laboratory. PATIENTS: Burn patients hospitalized in any Calgary Health Region burn center from November 2000 to September 2001. INTERVENTION: Gram stain plus culture of burn wound surface swab specimens obtained during routine dressing changes or based on clinical signs of infection. MAIN OUTCOME MEASURES: Degree of correlation (complete, high, partial, none), including weighted kappa statistic (kappa(w)), of Gram stain with culture based on quantitative microscopy and degree of culture growth. RESULTS: A total of 375 specimens from 50 burn patients were evaluated. Of these, 239 were negative by culture and Gram stain, 7 were positive by Gram stain only, 89 were positive by culture only, and 40 were positive by both methods. The degree of complete, high, partial, and no correlation of Gram stain with culture was 70.9% (266/375), 1.1% (4/375), 2.4% (9/375), and 25.6% (96/375), respectively. The degree of correlation for all 375 specimens, as expressed by the weighted kappa statistic, was found to be fair (kappa(w) = 0.32).Conclusion.-The Gram stain is not suitable for the microbiological analysis of burn wound surfaces.
