Postoperative outcomes after conventional, robotic-arm-assisted, and pressure-sensor-guided cruciate-retaining total knee arthroplasty: a retrospective cohort study

Abstract Conventional cruciate-retaining total knee arthroplasty (CR-TKA) uses manual instrumentation and surgeon-directed soft-tissue assessment, whereas MAKO robotic-arm assistance adds geometry- and gap-distance guidance and Verasense pressure sensing adds compartment-load feedback. Whether these approaches are associated with different clinically relevant postoperative outcomes remains uncertain. To evaluate associations between surgical technique and postoperative pain, range of motion, and function across conventional, MAKO-assisted, and Verasense-guided CR-TKA, and explore whether preoperative severity modified associations with postoperative outcomes. The single-centre cohort included 173 knees in 139 patients (36 Conventional, 92 MAKO, 45 Verasense). Analyses proceeded from post hoc favourable change to principal covariate-adjusted postoperative outcomes, followed by post hoc baseline-severity interactions. Principal models adjusted for baseline value, age, sex, and body mass index. Knee-level outcomes accounted for within-patient clustering; the Knee Society Function Score was analysed once per patient. Holm correction was applied within each stated test family. No outcome showed a Holm-adjusted group difference in favourable improvement or adjusted postoperative status. Verasense versus MAKO showed greater pain improvement (0.769 points; Holm P = 0.2195) and lower adjusted postoperative pain (−0.318 points; Holm P = 0.2546), but neither met the corrected criterion. Postoperative pain was zero in 125 of 149 knees (83.9%). The Knee Society Function Score severity interaction disappeared within the baseline range shared by all groups. We found no robust evidence that conventional instrumentation, geometry-based MAKO assistance, and load-based Verasense guidance were associated with different postoperative outcomes or that preoperative severity showed a reproducible technique-dependent association with outcomes. The Verasense-versus-MAKO pain pattern was hypothesis-generating and established neither superiority nor equivalence.