BACKGROUND: This study was performed to discover how much needle deflection occurs during spinal or epidural block and how it can be reduced. METHODS: A styrofoam block was used to simulate the paraspinal area of the back. A line was drawn perpendicular to the edge. Using the line as a guide, the needles were advanced through the block. Quincke-type needles of 22 and 25 gauge, pencil-type needles of 22 and 25 gauge or Tuohy needles of 17 and 18 gauge were used. Using a block of 6 cm thick, spinal needle was advanced through the introducer with their apertures facing to the same or opposite direction. The deflection from the perpendicular line was measured in mm intervals. RESULTS: Small-gauged and beveled spinal needles had more deflection depending on the block thickness (P < 0.05). However, the deflection of epidural needles was not consistent. When the bevel of the introducer and spinal needle was facing the same direction, the use of an introducer needle decreased the deflection in 22 and 25 G Quicke, and 25 G Whitacre needles (P < 0.05). When facing each other, the deflection was reduced in all the spinal needles (P < 0.05), and reduced more in the 25 G Quicke needles (P < 0.05). CONCLUSIONS: Considering the estimated depth of the spinal canal, needle type, bevel direction, its gauge and the use of an introducer needle, we can take advantage of the deflection phenomenon to reduce post-dural puncture headache and to increase the success rate of a spinal or epidural block.