Propensity-Matched Comparison of 90-Day Complications in Robotic-Assisted Versus Non-Robotic Assisted Lumbar Fusion

Charles C. Yu, MD; Leah Y. Carreon, MD, MSc; Steven D. Glassman, MD; Morgan E. Brown, MS; Christy L. Daniels, MS; David W. Polly Jr., MD; Jeffrey L. Gum, MD


Spine. 2022;47(3):195-200. 

In This Article

Abstract and Introduction


Study Design: Retrospective single center propensity-matched observational cohort study that included patients who underwent 1- to 3-level lumbar fusion surgery for degenerative conditions.

Objective: To compare 90-day complication rates between robotic-assisted and non-robotic-assisted lumbar spinal fusions in propensity-matched cohorts.

Summary of Background Data: A recent administrative database (PearlDiver) study reported increased 30-day complications with the utilization of robotic-assisted enabling technology.

Methods: Of 146 robotic-assisted cases that met inclusion criteria, 114 were successfully propensity matched to 114 patients from 214 cases who had 1 to 3 level lumbar fusion without robotic assistance based on age, sex, body mass index, smoking status, American Society of Anesthesiologist grade, number of surgical levels, primary versus revision, and surgical approach (posterior-only or anterior-posterior). We excluded tumor, trauma, infection, or deformity cases. Outcomes included surgical and medical (major/minor) complications at intraoperative, immediately postoperative, 30- and 90-day postoperative intervals, including reoperations, and readmissions within 90 days.

Results: All cause intraoperative complication rates were similar between non-robotic-assisted (5.3%) and robotic-assisted groups (10.5%, P = 0.366). Immediate postoperative medical complication rate was also similar between non-robotic-assisted (6.1%) and robotic-assisted groups (1.8%, P = 0.089). Thirty-day complication rates, 90-day complication rates, reoperation rates, and readmission rates showed no difference between non-robotic-assisted and robotic-assisted groups. There was no difference between return to OR for infection between the cohorts (non-robotic-assisted: 6 [5%] vs. robotic-assisted: 1 [0.8%], P = 0.119). There was however improved length of stay (LOS) in the robotic-assisted group compared with non-robotic-assisted group (2.5 vs. 3.17 days, P = 0.018).

Conclusion: In propensity-matched cohorts, patients undergoing 1- to 3-level robotic-assisted posterior lumbar fusion for degenerative conditions did not have increased 90-day complication rate, and had a shorter length of stay compared with non-robotic-assisted patients. There findings differ from a prior administrative database study as the robotic-assisted group in the current study had 0% return to OR for malpositioned screws and 0.8% return to OR for infection.

Level of Evidence: 2


Over 1.62 million instrumented spine surgeries are performed annually in the United States;[1] spine surgeons strive to maximize safety and efficacy when doing these operations. While pedicle screws serve as essential tools for spine stabilization and fusion, their malposition leads to significant complications.[2–4] Therefore, an emergence of enabling technology such as computer-assisted navigation or robotic-assisted surgery has been observed in efforts to mitigate screw malposition. Given its superior screw placement accuracy to fluoroscopy-assisted technique,[5,6] robotic-assisted spine surgery has gained widespread popularity in spinal instrumentation cases. Additionally, robotic-assisted surgery reduces radiation exposure to both patients and the surgical team.[7–9] Though the initial cost of integrating a robotic system is high, robotic-assisted spine surgery can be cost effective in the long run by reducing the rate of readmission and revision surgery.[10]

The Institute of Medicine prioritizes safety, patient-centered outcomes, and cost of care as key components of value-based health care.[11] Complications within 90 days following surgery have been utilized to track quality of care. Adoption of robotic-assisted spine surgery has been wide but we must scrutinize associated complications with new technology and their impact on the clinical outcome.[12] The majority of studies have demonstrated that robotic-assisted spine surgery decreases length of stay[13,14] and spine-related readmission and reoperation rate.[15] However, a recent Pearl-Driver database study by Yang et al[16] found that robotic-assisted spine surgery was associated with increased risk of revisions, infection, and instrumentation failure.

Database studies allow the investigation of rare complication with large study sizes and multicenter sampling technique. These study designs are effective at identifying trends, but their conclusions are often confounded by data reporting biases.[17] In order to have a more specific and granular investigation, we sought to compare the 90-day complication rate in robotic-assisted versus non-robotic-assisted fixation in our center with more than 3 years of robotic experience.