Lumbar spinal fusions have been performed for spinal stability, pain relief and improved function in spinal stenosis, scoliosis, spinal fractures, infectious conditions and other lumbar spinal problems. The success of lumbar spinal fusion depends on multifactors, such as types of bone graft materials, levels and numbers of fusion, spinal instrumentation, electrical stimulation, smoking and some drugs such as nonsteroidal anti-inflammatory drugs (NSAIDs). From January 2000 to December 2001, 88 consecutive patients, who were diagnosed with spinal stenosis or spondylolisthesis, were retrospectively enrolled in this study. One surgeon performed all 88 posterolateral spinal fusions with instrumentation and autoiliac bone graft. The patients were divided into two groups. The first group (n=30) was infused with ketorolac and fentanyl intravenously via patient controlled analgesia (PCA) postoperatively and the second group (n=58) was infused only with fentanyl. The spinal fusion rates and clinical outcomes of the two groups were compared. The incidence of incomplete union or nonunion was much higher in the ketorolac group, and the relative risk was approximately 6 times higher than control group (odds ratio: 5.64). The clinical outcomes, which were checked at least 1 year after surgery, showed strong correlations with the spinal fusion status. The control group (93.1%) showed significantly better clinical results than the ketorolac group (77.6%). Smoking had no effect on the spinal fusion outcome in this study. Even though the use of ketorolac after spinal fusion can reduce the need for morphine, thereby decreasing morphine related complications, ketorolac used via PCA at the immediate postoperative state inhibits spinal fusion resulting in a poorer clinical outcome. Therefore, NSAIDs such as ketorolac, should be avoided after posterolateral spinal fusion.