Background: Current treatment of postdural puncture headache includes epidural blood patch (EBP), which is invasive and may result in rare but severe complications. Sphenopalatine ganglion block is suggested as a simple, minimally invasive treatment for postdural puncture headache. We aimed to investigate the analgesic effect of a transnasal sphenopalatine ganglion block with local anaesthetic vs saline. Methods: We conducted a blinded, randomised clinical trial including adults fulfilling the criteria for EBP. Participants received a sphenopalatine ganglion block bilaterally with 1 ml of either local anaesthetic (lidocaine 4% and ropivacaine 0.5%) or placebo (saline). Primary outcome was pain in upright position 30 min post-block, measured on a 0–100 mm VAS. Results: We randomised 40 patients with an upright median pain intensity of 74 and 84 mm in the local anaesthetic and placebo groups at baseline, respectively. At 30 min after sphenopalatine ganglion block, the median pain intensity in upright position was 26 mm in the local anaesthetic group vs 37 mm in the placebo group (estimated median difference: 5 mm; 95% confidence interval: –14 to 21; P=0.53). In the local anaesthetic group, 50% required an EBP compared with 45% in the placebo group (P=0.76). Conclusions: Administration of a sphenopalatine ganglion block with local anaesthetic had no statistically significant effect on pain intensity after 30 min compared with placebo. However, pain was reduced and EBP was avoided in half the patients of both groups, which suggests a major effect not necessarily attributable to local anaesthetics. Clinical trial registration: NCT03652714.