It is unknown whether leaving the skin wound open to heal by secondary intention is associated with a lower rate of wound infections compared with primary skin-wound closure after operations for colon injuries. From June 1998 to December 2000 at our Level I academic trauma center 48 patients entered into a randomized controlled trial (RCT) and were randomized to have their skin wound primarily closed (CLOSED-RCT) or left open (OPEN-RCT). At the same time patients not included in the RCT were followed prospectively. At the discretion of the surgeon their skin wounds were managed by primary closure (CLOSED-nonRCT) or were left open (OPEN-nonRCT). Univariate and multivariate analysis was done to identify independent risk factors of wound infection, wound dehiscence, and necrotizing soft tissue infection. Wound infection developed in 65 per cent of CLOSED-RCT and 36 per cent of OPEN-RCT patients (P = 0.04) and wound dehiscence in 31 per cent and 14 per cent respectively (P = 0.18). No remarkable differences were noted in any other variable including length of hospital stay. Wound infection developed in 29 per cent of CLOSED-nonRCT and 15 per cent of OPEN-nonRCT patients (P = 0.46). There were three independent risk factors of wound infection: primary wound closure [odds ratio (OR) = 5.5, 95% confidence interval (CI) = 1.8-19.4], colectomy (OR = 3.4, 95% CI = 1.2-10.4), and intraabdominal infection (OR = 5.3, 95% CI = 1.3-24.2). There were two independent risk factors for wound dehiscence and/or necrotizing soft tissue infection: wound infection (OR = 20.9, 95% CI = 4.9-152.3) and intra-abdominal infection (OR = 19.3, 95% CI = 4.0-146.9). Primary closure of the wound almost doubles the rate of wound infection compared with leaving the wound open in operations for colon injuries. Primary wound closure is a risk factor for wound infection and wound infection is a risk factor for wound dehiscence or necrotizing soft tissue infection. Because of the complexity of evaluating the real clinical significance of superficial wound infection larger studies on trauma patients are required.