Petrosal sinus sampling has been described as an innocuous, essentially risk-free procedure. The authors report on two patients from two different institutions who sustained brain stem injury as a result of petrosal sinus sampling and two other patients in whom brain stem injury was avoided. Major neurologic complications have occurred in only 0.2% of procedures (one of 508) performed at the National Institutes of Health and in 0.5% of a subset of these procedures (one of 184) that were performed with a specially designed petrosal sinus catheter. The cause of these complications is unclear but is presumed to be localized venous hypertension. Tip-deflector catheter-guide-wire systems, if available, are recommended for petrosal sinus sampling. Inadequate data exist to permit recommendation of any other catheter. Brain stem injury is preventable if the catheter is withdrawn at the earliest sign of even a minor, seemingly insignificant problem. Subtle symptoms and signs that may not appear to be neurologic may herald a clinical catastrophe if not heeded.