Twenty-five adult inpatients with cellulitis were prospectively studied to determine if distinctive predisposing factors, characteristic clinical findings, or specific laboratory features were predictive of isolating a pathogen from needle aspiration cultures of the leading edges of their lesions. In the univariate analysis, age, underlying disease, temperature, and white blood cell count at admission to the hospital correlated with a positive needle aspiration culture. A series of logistic regressions were performed to determine if each of these variables was independently associated with a positive needle aspiration culture. The final model demonstrated that underlying disease and body temperature were independent predictors of obtaining a positive needle aspiration culture. Consequently, the addition of white blood cell count and age failed to enhance the model's predictability. For most patients this procedure will not be helpful [corrected] in establishing a bacteriologic diagnosis. On the basis of these results, the needle aspiration technique will most likely yield pathogens in patients with underlying disease that predisposes to the acquisition of cellulitis. These patients may also fail to mount a febrile response to infection.