1 Introduction

Unicompartmental knee arthroplasty (UKA) is a popular alternative to total knee arthroplasty (TKA) for select patients with single-compartment osteoarthritis (OA).¹ Between 2000 and 2009, 4.5% of all Medicare knee arthroplasties were unicompartmental. During that period, the number of TKAs increased 1.7-fold, while UKAs increased 6.2-fold.² This disproportionate growth may reflect several advantages of UKA over TKA, including faster recovery,³ shorter operative time, reduced length of stay,⁴ superior patient reported outcome measures (PROMs),^5,6 less blood loss,⁷ faster gait,⁸ more natural knee kinematics,⁹ and lower implant cost.¹⁰ Even in appropriately selected candidates, however, UKA carries one major disadvantage: The risk of revision surgery is 2.4 times higher than for TKA.²

Reasons for failure of UKA include aseptic loosening,¹¹ progression of OA in the other compartments (Fig. 1),¹² polyethylene wear,¹³ fracture,¹⁴ dislocation of mobile bearings,¹⁵ and poor surgical technique.¹⁶ Although the introduction of robotics¹⁷ and advances in implant design¹⁸ may have mitigated some of these modes of failure, progressive OA remains a challenge. The majority of these patients in our practice undergo revision to TKA. In select patients with progression of aseptic arthritis and a well-functioning UKA, however, the senior author has elected to resurface the diseased compartments using UKA components with robotic assistance. We refer to this procedure as tricompartmental knee arthroplasty (tri-UKA).

Close

Fig. 1 Radiographs of a 74-year-old man with a well-functioning medial UKA and progressive OA.

Export to PPT

In this paper, we describe our surgical technique and present a series of 11 patients with short-term follow up who underwent conversion to tri-UKA.

2 Methods

Following approval from our institutional review board, we retrospectively reviewed patients who underwent robotic-assisted conversion from UKA to tri-UKA between 2020 and 2024. The revisions were performed by a single surgeon at a single institution. Inclusion criteria required a minimum postoperative follow-up of 300 days.

Medical records were reviewed to collect demographic data, indications for revision, operative details, operative times, implant information, length of stay, transfusion requirements, complications, and PROMs. Demographic variables included age, sex, body mass index (BMI), and comorbidities. PROMs included Forgotten Joint Score (FJS),¹⁹ Knee injury and Osteoarthritis Outcome Score for Joint Replacement (KOOS JR),²⁰ Patient-Reported Outcomes Measurement Information System Global Health 10-Item Short Form (PROMIS10-SF),²¹ and the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC).²²

Eleven patients met the inclusion criteria, comprising 10 medial UKAs and 1 lateral UKA. The mean age at time of revision was 70 years (range, 52–79); 4 patients were male (36%); the mean BMI was 28.7 kg/m² (range 21.1–41.5). The mean interval between UKA and tri-UKA was 8.7 years (range, 6.8–11.4), and the mean postoperative follow-up was 2.0 years (range, 0.8–3.9). The indication for all revisions was progression of OA.

All tibial and femoral components were cemented, fixed-bearing Mako Restoris implants. Patellar components were press-fit, asymmetric Triathlon Tritanium buttons (Stryker, Mahwah, NJ). The mean polyethylene insert thickness was 8.3 mm (range 8–9).

2.1

2.1 Surgical technique

Conversion to tri-UKA was performed using the MAKO Unicompartmental Knee SmartRobotics™ System (Stryker, Mahwah, NJ). Patients underwent ipsilateral lower extremity computed tomography with a metal artifact reduction algorithm for preoperative planning.

Patients were positioned supine on the operating table with a tourniquet applied to the thigh. After standard skin preparation and draping, a timeout was performed. A #10 blade was used to make a midline incision, incorporating prior scars when possible. A medial parapatellar arthrotomy was performed using electrocautery. With the knee in extension, the patella was resurfaced using a freehand technique.

The knee was flexed to assess the integrity of the cruciate ligaments and to evaluate the condition of the pre-existing components. Femoral and tibial pins were inserted, and optical trackers were securely fastened. The center of hip rotation and both malleoli were registered. Bone registration was completed for the arthritic tibiofemoral compartment, as in a primary UKA, followed by verification of registration accuracy (Fig. 2). The knee was taken through a series of poses from extension to flexion to capture ligament laxity.

Close

Fig. 2 Screenshots demonstrating verification of bone registration for (a) the femur and (b) the tibia in a left knee with a pre-existing medial UKA.

Export to PPT

Component position was adjusted virtually based on the captured laxity and tracking data. For patients with a pre-existing lateral UKA, the bicompartmental knee arthroplasty workflow supported concurrent planning of the medial and patellofemoral components. For patients with a pre-existing medial UKA, however, concurrent lateral and patellofemoral resurfacing is off-label and unsupported by the software. These cases required alternating between the lateral and patellofemoral workflows. This did not require repeat registrations of the bone nor the anatomic landmarks; however, the outline of the planned component from one workflow was not visible in the graphical user interface of the other. To overcome this limitation, before switching to the patellofemoral workflow, we used the blunt registration probe and cartilage mapping feature to outline the anterior margins of the pre-existing medial femoral component and the planned lateral femoral component. These mapped points transferred between workflows, allowing implant positioning that avoided overlap (Fig. 3).

Close

Fig. 3 Screenshots from (a) lateral femoral and (b) patellofemoral workflows for a left tri-UKA. Green represents the planned position of the components. The gold points were captured by placing the blunt probe around the anterior margins of the pre-existing medial femoral component and the planned lateral femoral component. The yellow halos surrounding the gold points in the top-right pane are extraneous and may be ignored.

Export to PPT

After all components were planned to the senior surgeon's satisfaction, bone preparation (Fig. 4) was executed with the robotic arm in the following order: Trochlea, femoral condyle, tibia. The knee was cleared of bony debris, and the meniscus was excised (Fig. 5a). Trial implants were inserted, and the knee was evaluated manually and robotically through a full range of motion. Trial liners for the pre-existing tibial baseplate were available in case balancing required a change in liner size. After removing the trials, the tourniquet was inflated and the bone surfaces were irrigated with pulsatile lavage. The pre-existing polyethylene liner was exchanged. Canal aspirators were malleted into the femoral and tibial metaphyses and connected to suction. One bag of cement was mixed under vacuum, and final component implantation (Fig. 5b) was performed in the following order: Tibia, femoral condyle, polyethylene insert, trochlea, and patella. Excess cement was removed. The cement was allowed to cure with the knee in extension. The MAKO pins and the canal aspirators were removed. The arthrotomy and incisions were closed in standard layered fashion, and the tourniquet was deflated after dressings were applied.

Close

Fig. 4 Screenshot showing bone preparation of the left lateral femoral condyle. Interestingly, when switching from patellofemoral to lateral bone preparation, the software remembers that the trochlea was resurfaced.

Export to PPT

Close

Fig. 5 Intraoperative photographs of a patient with a well-functioning left medial UKA undergoing conversion to tri-UKA, taken after (a) bone preparation and (b) implantation. Cephalad is to the left of the page. The resurfaced patella is positioned behind the bent Hohmann retractor.

Export to PPT

All patients were discharged the next day and seen in clinic 2 weeks postoperatively, at which time anteroposterior, lateral, sunrise, and long-leg radiographs were obtained (Fig. 6).

Close

Fig. 6 Radiographs of the same patient shown in Fig. 1 after conversion to tri-UKA.

Export to PPT

3 Results

The mean FJS was 42.4 (range 0–95.8). The mean KOOS JR score increased from 61.3 preoperatively to 73.5 postoperatively (p = 0.021). The PROMIS10-SF mental health score increased from 46.6 to 51.4 (p = 0.026), while the physical health score increased from 49.4 to 52.4 (p = 0.13). The mean WOMAC score improved from 21.2 to 11.3 (p = 0.0042).

The mean operative time was 80 minutes. No patients required transfusion.

Postoperative complications occurred in 5 of 11 patients (45%): 2 required knee arthroscopy and synovectomy, 1 required multiple corticosteroid injections, 1 was treated with oral steroids for pes anserine bursitis, and 1 developed atrophy of the vastus medialis obliquus. Implant survivorship was 100% at a mean follow-up of 2.0 years.

4 Discussion

In this retrospective case series, we presented a novel technique along with PROMs for 11 patients who underwent robotic-assisted conversion of UKA to bicruciate-retaining tri-UKA for progression of OA. To our knowledge, this is the largest published case series of this procedure. In 2010, Rolston reported on a single patient with a well-functioning lateral UKA who underwent resurfacing of the medial and patellofemoral compartments.²³ Our literature review identified no other published research on tricompartmental knee arthroplasty.

Several studies have evaluated resurfacing of the medial and lateral compartments using UKA components (bi-UKA). In 2020, Wada et al. published a systematic review of 12 studies evaluating the clinical outcomes of bi-UKA to treat primary knee arthritis. Results were generally favorable, with PROMs improving postoperatively. Some bi-UKAs were revised for infection, aseptic loosening, dislocation of mobile bearings, and progression of OA in the patellofemoral compartment.²⁴ In 2022, Haffar et al. published a retrospective study comparing patients with well-functioning UKA who were converted to either bi-UKA or TKA for progression of OA. At a mean follow-up of 5.7 years, the bi-UKA group demonstrated superior outcome scores and equivalent survivorship.²⁵ Similarly in 2023, Lazzara et al. published a series of 44 knees revised from medial UKA to bi-UKA for progression of OA. Oxford Knee Scores improved at all time points, and survivorship of the lateral implants was 100% at 5 years. Their only complication was a medial bearing dislocation that required bearing exchange.²⁶

Robotics have also been used to assist in conversion from UKA to TKA. In 2024, Magruder et al. published a case series of 44 UKAs converted to TKA with robotic assistance. At a mean follow-up of 1.8 years, KOOS JR increased from 48.1 to 68.7, and r-WOMAC improved from 25.7 to 10.6.²⁷ Similarly, Mancino et al. prospectively evaluated 16 patients undergoing robotic-assisted conversion from UKA to TKA. Oxford Knee Scores increased from 24.3 to 42.3, and the mean Forgotten Joint Score at final follow-up was 84.1.²⁸ In 2025, Andriollo et al. prospectively evaluated 35 patients who underwent conversion from UKA to TKA using an imageless robotic system. At a mean follow-up of 31 months, the Oxford Knee Score increased from 31.4 to 41.5, and WOMAC improved from 53.5 to 17.8.²⁹

With appropriate patient selection, conversion to tri-UKA may be a viable alternative to TKA in the setting of progressive aseptic OA with a well-functioning UKA. Contraindications are similar to those for primary UKA and include incompetent ligaments, flexion contracture >15°, uncorrectable coronal plane deformity,³⁰ significant recurvatum,³¹ and patellar maltracking.³² Even when the decision is made to proceed with conversion to tri-UKA, we recommend consenting patients for possible conversion to TKA in the event that intraoperative findings necessitate a change in plans. Whenever feasible, we also recommend exchange of the pre-existing polyethylene liner.

Our study has several strengths. First, it is the largest published series of its kind, and it provides novel insight into a viable alternative to revision TKA. We used multiple validated instruments to compare patients' pre- and post-revision outcome measures. With one exception, our results were statistically significant. Our technique preserves both cruciate ligaments, which may theoretically maintain patients’ proprioception.³³ This technique also avoids removal of a well-functioning component, thereby reducing surgical trauma, bone loss, and other morbidity associated with explantation ³⁴. The surgical methodology section includes detailed technical guidance to overcome the limitations of the robotic software, which is critical for reproducibility. Finally, the revisions and follow-up assessments were conducted in a private-practice environment, which enhances the external validity of the findings.

This study has several limitations. Its retrospective design introduces potential selection bias. The small sample size limits statistical power and precludes subgroup analysis. The mean follow-up duration of 2.0 years does not capture late complications nor long-term survivorship. The study was conducted at a single institution by a single surgeon, which may limit generalizability. Furthermore, the absence of a control group prevents direct comparison of outcomes. Lastly, we used off-label robotic workflows, which may not be feasible in all settings.

5 Conclusion

In select patients with a well-functioning UKA and symptomatic progression of OA, robotic-assisted resurfacing of the native compartments is a viable alternative to revision TKA. In our series of 11 patients who underwent conversion to tri-UKA, we showed clinically significant improvements in PROMs, with 100% implant retention at 2 years. Further research is needed to assess long-term survivorship and to directly compare outcomes between tri-UKA and TKA.

Consent

This study was a retrospective review of data from electronic medical records, all of which was standard of care. Therefore, no explicit patient consent was required to conduct or our study.

Credit author statement

Adnan Zubair: Methodology, Formal analysis, Writing-Original draft, Visualization. Robert C. Marchand: Conceptualization, Investigation, Writing-Reviewing and Editing, Supervision. Emily Kaczynski and Kevin Marchand: Data curation.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.