Can a pharmacological pain analysis in patients with chronic low back pain predict the outcome of lumbar fusion? Preliminary report.

Chronic low back pain patients chosen for lumbar fusion surgery were entered into a prospective study with the aim of evaluating whether pre-operative pharmacological pain classification correlated with the outcome of surgical treatment. Twenty consecutive patients (mean age 39 years, range 29-50 years) with a mean pain duration of 4.6 years (range 1-8 years) participated. The patients' pain, disability and work situation were evaluated pre- and post-operatively. According to the results of pharmacological pain testing the patients were classified into five pain groups: (1) those with nociceptive pain, (2) those with neuropathic pain, (3) placebo responders, (4) non-responders and (5) unclassified. This classification was based on the following approaches: intravenous morphine, intravenous lidocaine, epidurally administered fentanyl and a local anaesthetic. Surgical treatment consisted of posterolateral fusion of the lower two functional lumbosacral spinal units combined, if necessary with appropriate decompression. The results of the pharmacological pain assessment were not disclosed until the end of the follow-up period and outcome was evaluated by an unbiased observer. The distribution of patients between pain groups was as follows; nine had nociceptive pain, two suffered neuropathic pain, there were no placebo responders and seven were non-responders. Two patients could not be classified into any of these groups. Outcome was rated at 6, 12 and 24 months as excellent/good in eight out of nine patients with nociceptive pain. In the group with non-responding pain surgical outcome was significantly poorer (P < 0.01), and was ranked as excellent/good in only one out of seven patients. There was a significant improvement (P < 0.05) concerning both pain and disability in the nociceptive group, but not in the non-responding group. Pre-operative pharmacological pain testing may be useful as a predictor of surgical outcome and we suggest that it can be employed as a means to identify patients with non-responding pain and poor surgical prognosis.