Several noninvasive techniques are used to collect specimens for wound cultures, but which is the most accurate is unclear. Cultured swab specimens taken from 83 patients with full-thickness, chronic, nonarterial wounds-most commonly, diabetic foot ulcers-were compared with the culture of a full-thickness wound biopsy taken at the same time. After the wound's dressing was removed, three collection techniques were used. First, a swab was used to collect exudate on the wound's surface. The wound was cleaned and then the Z-technique was used, in which a swab is rotated as it is zigzagged over the entire surface of the wound. A portion of the wound that was free of dead tissue and debris was then cleaned and the Levine technique, which involves rotating a swab over a square centimeter of tissue for five seconds with enough pressure to extract fluid from the tissue, was used. A punch biopsy of viable wound tissue was then performed. Biopsy results indicated that 36% of the wounds were infected (1 x 10⁶ organisms or more per gram of tissue). Patients receiving systemic antibiotics were less likely to have an infected wound. Although the cultures from all three collection techniques had acceptable agreement with the biopsy culture, those collected using the Levine technique were the most accurate and the most sensitive. Wound cultures grown from swab specimens collected by the Levine technique are a more accurate measure of wound infection than those from the other two techniques, and they provide an acceptably accurate alternative to wound biopsy.