Blood cultures in the evaluation of uncomplicated cellulitis.

PURPOSE: The frequency of bacteremia and the array of microorganisms involved in cellulitis vary greatly among studies. Although current guidelines do not recommend routine blood culture in uncomplicated cellulitis, their implementation in clinical practice remains challenging. We therefore aimed to assess the frequency, determinants and microbiology of bacteremia in hospitalized patients with uncomplicated cellulitis.

METHODS: We retrospectively reviewed the medical records of all adult patients admitted at a primary-care hospital with a diagnosis of community-acquired uncomplicated cellulitis during a 4-year period. We looked at the factors associated with blood cultures sampling and at the incidence, determinants and microbiology of bacteremia in this population.

RESULTS: Among the 476 patients hospitalized with a diagnosis of cellulitis, 250 (52.5%) had blood cultures. Fever, high C-reactive protein and lymphatic insufficiency were significantly associated with the sampling of blood cultures. Twelve (4.8%) patients had bacteremia. Alcoholism and duration of hospitalization were associated with bacteremia in multivariate analysis. Among the 12 patients with bacteremia, 9 had Streptococcus sp. and 3 had Staphylococcus aureus infection.

CONCLUSION: In our study population with uncomplicated cellulitis, representative of unselected population admitted at primary-care hospitals, bacteremia was uncommon and not associated with discriminant patient characteristics, except for alcohol abuse. Episodes of bacteremia were exclusively due to gram-positive cocci susceptible to co-amoxicilin, a common first-line empirical therapy. In accordance with existing guidelines, we do not recommend to collect blood for cultures in uncomplicated cellulitis. Clinicians' awareness of guidelines and of the poor yield of blood cultures could reduce useless investigation.