The accurate prediction of occult sepsis is very important in evaluating the patient who has a painful prosthetic joint. Unfortunately, there is no single test that is able to consistently differentiate mechanical from septic loosening. Authors have suggested that the presence of acute inflammation on frozen histologic sections correlated well with positive intraoperative cultures. The authors' experience with frozen sections was examined as a guide to the intraoperative determination of occult sepsis in revision joint surgery. One hundred seven consecutive total joint revisions were performed from 1986 to 1989. All patients had a complete histologic analysis of tissue obtained from multiple surgical sites. The relationship between the frozen section, final culture, and complete histologic analysis was evaluated in a comparative fashion. Using a positive culture as the criterion for the presence of occult infection, a statistical analysis comparing frozen sections and the complete histologic analyses were performed. The sensitivity of a frozen section as a diagnostic test to detect occult periprosthetic infection when present was 18.2%. The specificity of a frozen section to correctly identify the absence of infection was 89.5%. The frozen section is a fairly specific, but not a sensitive, screening test for the detection of occult sepsis. Over reliance on this technique in a revision setting should be avoided.