SUMMARY Diagnostic lumbar puncture was performed on 76 neurological inpatients. They were randomly allocated to one of four bed rest positions for four hours following the procedure (supine and horizontal, prone and horizontal, supine with head-down tilt and prone with headdown tilt) after which they were allowed to get up. There was no substantial or significant difference in the incidence of headache between the four groups. Expectation of headache did not appear to be an important factor in its development.It is widely believed that headache after lumbar puncture is related to cerebrospinal fluid (CSF) hypotension attributable to continued leakage of CSF through the dural hole made at lumbar puncture.'-3 If this postulate is correct the development of post-lumbar puncture headache may possibly be prevented by techniques that counteract CSF hypotension. Many such techniques have been reported,4-6 but most trials suffer from poor design particularly with respect to randomisation. However, the importance of needle size seems to be beyond doubt."8 Although epidural blood patching may be a useful method in the treatment of intractable post-lumbar puncture headache, it is not without complications910 and cannot be recommended for routine use for the prevention of headache, particularly in inexperienced hands. of CSF leakage. Brocker'5 suggested that the adoption of the prone position after lumbar puncture dramatically decreased the incidence of headache by the proposed mechanism of decreased CSF leakage with the dural hole uppermost. We decided to study this further by looking both at the effect of the supine and prone positions, and at head-down bed tilting in a randomised clinical trial. The results of the studies mentioned above suggest that headache after lumbar puncture should be least frequent in patients adopting the prone and tilted position, and most frequent in those adopting the supine and horizontal position. We also looked at the effect of anticipation of headache on its subsequent development, a subject about which there has been recent

No abstract