A combined retrospective and prospective study was designed to assess the efficacy of shunt fluid examination in the evaluation of shunt malfunction and/or infection, and to ascertain the complication rate associated with this procedure. Ninety-one patients with shunts for the treatment of hydrocephalus underwent a total of 209 diagnostic shunt 'taps'. Of 72 instances of mechanical obstruction documented at surgery, 70 were correctly identified by abnormal shunt fluid dynamics, either an opening cerebrospinal fluid (CFS) pressure in excess of the expected valve pressure or absent flow of fluid. The organisms responsible for 12 of 13 shunt-related infections were correctly isolated on initial and all subsequent shunt fluid cultures obtained prior to the third tap. In all instances in which lumbar or ventricular CSF, blood, or wound cultures disclosed an organism, shunt fluid cultures also identified the agent. In addition, these procedures were less reliable. No complications occurred during or immediately following any of the shunt taps. Long-term consequences were assessed in 53 patients with an average follow-up period of 26 months. Some shunt systems subsequently required revision and 2 infections were documented. The uniformly long interval between the shunt fluid examination and these complications makes it unlikely that the taps contributed to the development of malfunction or infection. Shunt fluid examination appears to be a simple, benign, and yet accurate means of diagnosing shunt malfunctions and/or shunt-related infection.