A review was undertaken of all 190 patients who were referred over 7 years, from 1991 to 1997 inclusive, for an epidural blood patch as a treatment for headache after dural puncture. The patterns of referral and symptoms, the distributions of age and gender and the effectiveness of the blood patch were examined. Most of the referrals (n=153) were after deliberate diagnostic dural puncture in neurology and neuroradiology, with a minority (n=28) used for anaesthesia and obstetrics, which were mostly inadvertent. Another nine cases were related to placement of an intrathecal catheter. The numbers of referrals per year reached a maximum in 1995 before falling again, a curious inverse relation to the number of invasive neuro-radiological diagnostic procedures. Most of the patients were between 30 and 50-years-old, with 25 younger than 30 and 14 older than 60. Women accounted for 70% of the referrals for headache, although the gender ratio amongst patients subjected to at risk procedures appeared closer to 50:50. Neckache accompanied the headache in 85% of cases, auditory problems were volunteered by three patients and one patient had diplopia for 6 weeks. Of the 190 patients who were referred, 186 received at least one patch, the symptoms in the remaining four being too mild or atypical to warrant blood patch treatment. This provided initial relief in all but two patients, one of whom received a further epidural blood patch with no effect. There was sustained relief of symptoms in 136 and a partial relapse in 38 patients, which resolved without needing any further blood patch. A second patch was provided for seven patients and a third for three patients, of whom two were cured. Of the patients who needed more than one blood patch, nine were after inadvertent dural puncture with a Tuohy needle and, of these patients, six were in labour. A total of 200 patches were provided in all for the 186 patients and all but three patients had a satisfactory outcome. Epidural blood patches are effective in treating headache after dural puncture, but less successful than is commonly believed, especially after inadvertent dural taps. A relapse after treatment does not always require a second patch. Specialities other than anaesthesia seemed reluctant to accept the benefits in both cost and comfort of using needles of improved design for dural puncture.