Methods:

Axillary block was performed in 81 patients with the aid of a peripheral nerve stimulator. Immediately after the block, patients were allocated randomly to two groups (Group 1, n= 40; Group 2, n= 41) to have the blocked limb placed either on their thorax or in abduction. Fifteen minutes later, when the block was complete, the position of the limb was changed, without the knowledge of the patient, to a new position of abduction and flexion of the forearm. Patients were questioned about the new limb position. Fifteen minutes later, the limb was then transiently shown to the patient and further placed'blindly'to another position. Patients were once more questioned about this new position.

Results:

Correct perception was more frequently observed in both groups following the first interview. Thirty-two and 34 patients gave at least one erroneous response about their limb position when it was initially placed on the thorax or in abduction respectively. Transient visualization of the limb position did not improve the rate of correct response.

Conclusions:

Patients' perception of the position of the limb after axillary block in the majority of instances was probably due to persisting sensory inputs from the shoulder joint, which is not involved in this block. This is in contrast to supraclavicular or interscalene blocks. A significant number of patients experienced phantom limb sensation. They reported an arm position in fact related to the position of their arm before the axillary block.