The indications for repeated operation in patients with persistent or recurrent pain after lumbosacral spine surgery are not well established. Long-term results have been reported infrequently, and in no case has mean follow-up exceeded 3 years. We report 5-year mean follow-up for a series of repeated operations performed between 1979 and 1983. Patient characteristics and modes of treatment have been assessed as predictors of long-term outcome. One hundred two patients with "failed back surgery syndrome" (averaging 2.4 previous operations), who underwent a repeated operation for lumbosacral decompression and/or stabilization, were interviewed by a disinterested third party a mean of 5.05 years postoperatively. Successful outcome (at least 50% sustained relief of pain for 2 years or at last follow-up, and patient satisfaction with the result) was recorded in 34% of patients. Twenty-one patients who were disabled preoperatively returned to work postoperatively; 15 who were working preoperatively became disabled or retired postoperatively. Improvements in activities of daily living were recorded, overall, as often as decrements. Loss of neurological function (strength, sensation, bowel and bladder control) was reported by patients more often than improvement. Most patients reduced or eliminated analgesic intake. Statistical analysis (including univariate and multivariate logistic regression) of patient characteristics as prognostic factors showed significant advantages for young patients and for female patients. Favorable outcome also was associated with a history of good results from previous operations, with the absence of epidural scar requiring surgical lysis, with employment before surgery, and with predominance of radicular (as opposed to axial) pain.(ABSTRACT TRUNCATED AT 250 WORDS)