Thirty-six patients with chronic back and/or leg pain following previous lumbar surgery who underwent both spinal canal exploration and spinal fusion were subjected to retrospective review. The purpose was to determine the probability of success for this surgical approach. Twenty (56%) of the 36 patients had a satisfactory result. In 15 patients with multiple objective findings of an ongoing radiculopathy, 11 (73%) improved. Only nine (43%) of 21 patients improved if these preoperative criteria were absent. Analysis according to the type of surgery performed in the spinal canal demonstrated improvement in (a) 17 (74%) of 23 patients who had wide bony decompression, (b) eight (61%) of 12 patients who had discectomy, and (c) seven (47%) of 15 patients who had an extensive neurolysis. In 17 patients whose time interval between the previous operation and present reconstruction was greater than 18 months, 13 (76%) improved. Only seven (36%) of 19 patients with a shorter time interval improved. The presence of pseudarthrosis was a poor indication for repeat lumbar surgery. The number of previous lumbar surgeries may not necessarily preclude a satisfactory outcome. Solid fusion correlated highly with a satisfactory outcome. Best results are obtained when objective preoperative findings indicate the presence of a surgically correctable abnormality.