Perioperative nerve injury: a silent scream?

WELCH et al. add 112 well-documented adverse events from 380,680 patients to the third most commonly recognized cause of anesthesia litigation: Perioperative nerve injuries. This retrospective analysis from the University of Michigan collates three concurrent data sets (quality assurance, an internal closed-claims collection, and billing codes) to summarize the clinical experience with nerve injuries at one large tertiary university hospital from 1997 to 2007. The casual reader might be reassured by the relatively low global frequency of nerve injuries at 0.03% (1:3,400), but there were likely other patients who developed these problems. There are a number of potential reasons that the current study may have underestimated the incidence of nerve injuries. First, although the authors were diligent, using three separate departmental and institutional databases to identify patients with potential perioperative neuropathies, they did not report the crossover identification rates between the three databases. That is, how many patients with new peripheral nerve injuries were identified in more than one database? As reported, it appears that each database contributed new cases that were not found in the other databases. This lack of cross-identification suggests a high likelihood that other cases were missed. Second, in an effort to specifically seek new nerve injuries that were likely the result of intraoperative or immediate postoperative care, the authors sought information on peripheral neuropathies that were identified only during the first 48 h postoperatively. Previous prospective studies have shown that a number of perioperative neuropathies are first identified more than 48 h after surgery. Third, retrospective studies of perioperative nerve injuries document only injuries that are either sufficiently symptomatic to be noted in records or missed entirely in postoperative surveillance, a problem especially true for retrospective studies performed at tertiary medical centers where patients do not necessarily have long-term care. Thus, the true incidence of perioperative neuropathies is still unclear, but likely exceeds 0.03%, and will probably require a major prospective study. Nonetheless, the authors should be congratulated for the extensive efforts they used to compensate for the problems posed by the retrospective methodology. Regardless of methodology, there was approximately one anesthetized patient each month in the University of Michigan Medical Center who had additional unexpected pain, distress, and perhaps a disability unrelated to his or her primary operation. As a profession, have we been deaf to a silent “Scream” (with apologies to Edvard Munch, The Scream, 1893), of a nerve within an appropriately padded and positioned limb in the operating room, which is often assumed (“res ipsa loquitor”) to be the cause of unexpected perioperative nerve injury?