Incidence and characteristics of potential and actual retained foreign object events in surgical patients

Incidence of retained foreign objects (RFOs) after operations is unknown, as many can go unrecognized for years. We reviewed the incidence and characteristics of surgical RFO events at a tertiary care institution during 4 years.

All RFO events, near misses and actual, reported on an adverse event line during 2003 to 2006 were reviewed.

During 2003 to 2006, there were 191,168 operations performed, with 68 reported events resulting in a potential RFO defect rate of 0.356/1,000 patients. After review, 34 patients had no RFOs (near misses) and 34 were actual RFOs, resulting in a true RFO defect rate of 0.178/1,000 operations or approximately 1:5,500 operations. In the near-miss patient, needles were miscounted 76% of the time. In the 34 actual RFO patients, items retained were 23 sponges (68%), 7 miscellaneous other items (20%), 3 needles (9%), and 1 instrument (3%). The 34 actual RFOs occurred in incidents where the count had been reported as correct in 21 patients (62%). In 18 patients where an RFO was eventually discovered, intraoperative imaging detected only 12 objects (67%). In operations involving a body cavity, our practice is to obtain a high-resolution x-ray survey film, in a dedicated x-ray suite, before entering the recovery room. Twenty RFOs were identified from survey films and all occurred in patients with correct counts. No RFOs occurred during emergency or high blood-loss procedures and none resulted in demonstrable clinical harm. Two patients left the hospital with an RFO. Twenty-two patients (64.8%) underwent reoperation, with 1 object not removed, 6 (17.6%) retrieved without operation, and 6 (17.6%) where the clinical decision was not to remove.

RFOs at an institution that routinely performs postprocedure x-rays indicate that RFOs can occur more frequently than expected from the literature. The majority occur in patients with correct counts. Relying on counting as the primary mechanism to avoid RFOs is unreliable, and investigating new technologies designed to achieve reliable counts is warranted.