Knee Arthroplasty Surgery: Short-Term Effect of the ROSA Surgical Robot

This was a prospective, randomized, controlled clinical trial including patients with knee osteoarthritis requiring total knee arthroplasty (TKA). Patients were randomized in a 1:1 ratio using a single-blind, parallel-group design based on a dice roll: even numbers assigned patients to the robot-assisted TKA (RATKA) group, and odd numbers assigned them to the conventional manual TKA (mTKA) group.

Inclusion criteria comprised symptomatic knee osteoarthritis with surgical indication for TKA and a correctable varus or valgus deformity of less than 18 degrees. Exclusion criteria included prior surgery on the index knee; prior arthroplasty in the same limb; neurological disorders affecting knee function; tumor involvement; ipsilateral hip pathology limiting range of motion (ROM); active periarticular infection; body mass index (BMI) > 35; dependency or institutionalization (e.g., nursing home residents); and severe anatomical deformity, instability, or malalignment requiring alternative surgical strategies.

Surgical procedures (RATKA using the ROSA® Knee System and mTKA) were performed as previously described in the literature. All surgeries were carried out by the same senior surgeon, who had previously completed the ROSA learning curve. A medial parapatellar approach and standard surgical exposure were used in all cases. The same posterior-stabilized implant design (Persona® PS; Zimmer Biomet, Warsaw, IN, USA) was implanted with patellar resurfacing and cement fixation (PALACOS® R+G, Heraeus). A tourniquet was placed but not inflated; no limb ischemia was induced. Tranexamic acid was administered intravenously unless contraindicated. Preoperative planning was based on full-length standing radiographs of the lower limbs (Tele-rx), and full-scale anteroposterior and lateral knee radiographs (AP/L-rx). Femoral and tibial osteotomies were calculated from mechanical axis measurements, and implant sizes were selected using AP/L-rx data. In the RATKA group, intraoperative registration was performed manually after arthrotomy, without the use of preoperative 3D imaging, in compliance with the Spanish Data Protection Act (Organic Law 3/2018). The ROSA® Knee System (software version 1.2; Zimmer Biomet) was used according to manufacturer guidelines.

The following data were collected. Baseline: demographic and clinical data (comorbidities, BMI, osteoarthritis severity using the Kellgren-Lawrence classification); hemoglobin and hematocrit values; radiographic studies (Tele-rx of the limb and AP/L knee views); and patient-reported outcome measures (PROMs) using the Visual Analog Scale (VAS) for pain, the Knee injury and Osteoarthritis Outcome Score (KOOS), and the 36-Item Short Form Health Survey (SF-36). Intraoperatively: operative time and intraoperative complications. Postoperative Day 3: hemoglobin and hematocrit values; radiographic studies (Tele-rx and AP/L views); functional recovery measured by range of motion (flexion and extension in degrees) using a goniometer; and mechanical limb alignment assessed on digital Tele-rx. Three months postoperatively: PROMs (VAS, KOOS, and SF-36); range of motion; radiographic studies (Tele-rx and AP/L views); and mechanical limb alignment measured again on digital Tele-rx.

Descriptive statistics were reported as frequencies for categorical variables and as means with standard deviations (SD) for continuous variables. Between-group comparisons were conducted using Student's t-tests or Mann-Whitney U tests for continuous variables, and chi-square or Fisher's exact tests for categorical variables. Pre- and postoperative clinical and radiographic outcomes were compared using paired t-tests or Wilcoxon signed-rank tests within groups, and unpaired t-tests or Mann-Whitney U tests between groups. Univariable and multivariable linear regression models were used to analyze VAS, KOOS, and SF-36 outcomes at 3 months. Multivariable models were adjusted for sex, age, BMI, and discharge values. All statistical analyses were performed using STATA 15 (StataCorp, College Station, TX, USA), with p-values < 0.05 considered statistically significant.