Abstract Background context Prior studies have documented an increased complication rate in diabetics treated by lumbar decompression. Despite the assumption that this risk would be exacerbated in larger fusion procedures, a recent study found no such increased risk. Purpose To clarify the perioperative risk for patients with diabetes mellitus undergoing instrumented lumbar fusion as compared with controls. To identify differences, if any exist, between patients with insulin-dependent diabetes mellitus (IDDM) and non–insulin-dependent diabetes (NIDDM). Study design/setting This is a retrospective case-control study designed to examine the risk of perioperative complications in patients with diabetes mellitus treated by lumbar instrumentation and fusion. This study analyzed 94 diabetic patients and 43 controls treated by posterior lumbar instrumentation and fusion. Outcome was assessed based on the occurrence of perioperative complications. Complications were classified as major or minor. Methods Hospital and office records were reviewed for 94 diabetics (51 NIDDM, 43 IDDM) and 43 controls matched for age, sex and procedure with the IDDM group. The groups were compared for demographic characteristics, surgical data, perioperative complications and fusion rate. Results The NIDDM, IDDM and control groups were equivalent with regard to age, gender, nicotine use, fusion levels, operative time and estimated blood loss. There was a significant difference in complication rate between the NIDDM (53%) and IDDM (56%) groups versus controls (21%). There was a significant increase in complications with multilevel fusion in all groups and with increased operative time in the IDDM group. There was a significantly greater nonunion rate in the NIDDM (22%) and IDDM (26%) groups as compared with controls (5%). Conclusions Patients with both IDDM and NIDDM have a significantly increased risk of perioperative complications as compared with controls when treated by lumbar instrumentation and fusion.