Attempts to reach conclusions concerning the efficacy of antimicrobial prophylaxis in neurosurgery have been frustrated by the paucity of clinical trials which cannot be invalidated by flaws in their design and/or execution. Although the data are inconclusive, the weight of evidence suggests that antibiotics probably exert a protective effect in clean, non-implant procedures, for which a first or second generation cephalosporin would be appropriate. No consensus has been reached regarding the benefits of prophylaxis to patients undergoing the implantation of a CSF shunt and surgeons will therefore need to come to a decision based on their own experiences; those who elect to give antibiotics will still be left with the problems of choosing the optimum regimen and route of administration. Nor can valid conclusions be drawn about the efficacy of prophylaxis in patients with dural fistulae; a rational and ethically supportable approach would be to withhold antibiotics and to monitor patients closely for the early signs of meningitis, instituting appropriate empirical therapy should this complication arise.