Patients and methods. A randomised study was performed to compare the frequency of postdural puncture headache in 56 patients who underwent spinal anaesthesia for extra-corporeal shockwave lithotripsy using either a Sprotte 24 G (n = 28) or Vygon 29 G or Quincke type needle (n = 28). Frequency of headache was recorded in a similar group of 28 patients who received general anaesthesia. Results. Dural puncture was easier with the Sprotte 24 G cannula than with the less stable Quincke needle, as documented by a significantly shortened time for insertion of the cannula (4.6 +/- 2.6 vs 8.6 +/- 6.3 min, P < 0.005). The total frequency of post-operative headache was 57 % in the Vygon 29 G group and 25 % in the Sprotte 24 G group; 21 % of patients in the general anaesthesia group complained of headache. Frequency of postdural puncture headache, classified as being posture-related, was 25 % in the 29 G Vygon group, compared with 11 % in the 24 G Sprotte group (P = 0.148). When only moderate and severe postdural puncture headache was considered, there was a significant difference (25 % vs. 4 %; P = 0.026) in favour of the Sprotte cannula. Discussion and conclusions. Thus, the 24 G Sprotte needle was at least as effective as the 29 G Vygon needle, and there is a suggestion that the former is more effective in minimising the incidence of moderate or severe postdural puncture headache.