Learning curve of robotic arm-assisted versus manual unicompartmental knee arthroplasty : accuracy, theatre time, complications, patient-reported outcome measures, and implant survival.

Aims: To compare the outcomes of robotic arm-assisted (rUKA) and manual unicompartmental knee arthroplasty (mUKA) during the learning phases of a single surgeon. Previous comparisons have not accounted for the recognized learning curve of mUKA. Methods: This retrospective study included 109 patients (n = 37 rUKA, n = 72 mUKA) treated by a single surgeon during their learning curve at two centres during the same period. Inclusion criteria encompassed isolated medial compartment osteoarthritis with correctable deformities. Outcomes included patient-reported outcome measures (PROMs), implant positioning accuracy and precision, operating time, complications, and implant survival. Learning curves for both groups were analyzed using cumulative sum (CUSUM) analysis and breakpoint detection using piecewise linear regression. Results: There were no differences in baseline characteristics between groups, except for American Society of Anesthesiologists (ASA) grade I, which was more prevalent in the rUKA group (p = 0.017). While both groups achieved similar PROMs postoperatively, the rUKA group showed greater improvement in Oxford Knee Score, being significantly more likely to achieve the minimal important change and/or patient-acceptable symptom state (p = 0.014). No learning curve was seen for PROMs in either group. The learning curve to competency for operating time was shorter for rUKA (13 cases) compared with mUKA (45 cases). rUKA demonstrated significantly higher accuracy and precision across all radiological planes (p ≤ 0.005) with no observable learning curve, whereas the mUKA group had a learning phase for implant positioning that took 25 cases, and competency was not reached until case 55. Tibial overhang occurred in eight mUKA cases and none in rUKA (p = 0.033). At a median follow-up of over six years, all four revisions occurred in the mUKA group, however this was not statistically significant (p = 0.141). Conclusion: rUKA demonstrated a greater likelihood of achieving clinically meaningful outcomes, a shorter learning curve, and improved implant positioning accuracy and precision, compared with mUKA, when performed during the surgeon's learning phase.