Over a two-year period, 265 Norwegian orthopaedic surgeons working at 71 institutions performed 63 484 operations under a tourniquet. Their replies to a questionnaire revealed that they mostly followed modern guidelines in their use of the tourniquet. Most felt that the tourniquet could be left on for two hours, and that it could be re-applied after 15 minutes. A total of 26 complications (one in 2442 operations) that might have been due to the tourniquet were reported, of which 15 were neurological. Three were in the upper limb (one in 6155 operations) and 12 in the lower limb (one in 3752 operations). Two were permanent (one in 31742 operations), but the remainder resolved within six months. One permanent and one transient complication occurred after tourniquet times of three hours. The incidence of tourniquet complications is still at least as high as that estimated in the 1970s.

The Department of Health and the Public Health Laboratory Service established the Nosocomial Infection National Surveillance Scheme in order to standardise the collection of information about infections acquired in hospital in the United Kingdom and provide national data with which hospitals could measure their own performance. The definition of superficial incisional infection (skin and subcutaneous tissue), set by the Center for Disease Control (CDC), should meet at least one of the defined criteria which would confirm the diagnosis and determine the need for specific treatment.

We have assessed the interobserver reliability of the criteria for superficial incisional infection set by the CDC in our current practice. The incisional site of 50 patients who had an elective primary arthroplasty of the hip or knee was evaluated independently by two orthopaedic clinical research fellows and two orthopaedic ward sisters for the presence or absence of surgical-site infection. Interobserver reliability was assessed by comparison of the criteria for wound infection used by the four observers using kappa reliability coefficients. Our study demonstrated that some of the components of the current CDC criteria were unreliable and we recommend their revision.