The Journal of Arthroplasty xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

The Journal of Arthroplasty journal homepage:

Robotic Guidance Does Not Improve Component Position or Short-Term Outcome in Medial Unicompartmental Knee Arthroplasty Dane C. Hansen, DO a, Sharat K. Kusuma, MD, MBA b,⁎, Ryan M. Palmer, DO a, Kira B. Harris, CCRC b a b

Department of Orthopedic Surgery, Doctors Hospital, Columbus, Ohio Department of Orthopedic Surgery, Grant Medical Center, Columbus, Ohio

a r t i c l e

i n f o

Article history: Received 30 July 2013 Accepted 10 April 2014 Available online xxxx Keywords: unicompartmental knee arthroplasty robotics robotic-assisted surgery component alignment patient outcomes

a b s t r a c t We performed a retrospective review in a matched group of patients on the use of robotic-assisted UKA implantation versus UKA performed using standard operative techniques to assess differences between procedures. While both techniques resulted in reproducible and excellent outcomes with low complication rates, the results demonstrate little to no clinical or radiographic difference in outcomes between cohorts. Average operative time differed significantly with, and average of 20 minutes greater in, the robotic-assisted UKA group (P = 0.010). Our minimal clinical and radiographic differences lend to the argument that it is difficult to justify the routine use of expensive robotic techniques for standard medial UKA surgery, especially in a well-trained, high-volume surgeon. Further surgical, clinical and economical study of this technology is necessary. © 2014 Elsevier Inc. All rights reserved.

Medial unicompartmental knee arthroplasty (UKA) for isolated medial knee arthritis is a highly successful and efficacious procedure [1–10]. However, multiple published reports demonstrate that this procedure is technically more challenging than total knee arthroplasty (TKA) and that surgical technical errors result in high early failure rates [11–16]. In a national, multicenter review of failed UKA cases, Epinette et al [11] observed that technical mistakes were the greatest contributor to UKA failure. Hamilton et al [12] reported that following their acceptance of UKA, it was necessary to reduce their use of “minimally invasive” exposures for UKA, as these limited exposures led to increased technical errors, complications, and inferior outcomes. Other authors have documented that a combination of patient selection, component design, and component placement is interrelated with the subsequent success of UKA [13–16]. Recent changes in component design, surgical instrumentation, and surgical techniques have led to improved UKA radiographic and clinical outcomes of UKA [1,3,5,17]. The changes in surgical instruments that have taken place include systems that allow more accurate flexion–extension gap balancing and more accurate bone preparation. However, despite these improvements in manual instruments, some surgeons have also recently adopted use of robotic-assisted navigation systems with the goal of even further improving accuracy of implant placement [18–21]. While most experts agree that improvements in component positioning and procedure reproducibility should enhance clinical outcomes and survivorship, the literature has not clearly demonstrated that these new, costly robotic systems The Conflict of Interest statement associated with this article can be found at doi: http:// ⁎ Reprint requests: Sharat K. Kusuma, MD, MBA, 340 E. Town St., Ste 7-250, Columbus, OH 43215.

can consistently and definitively outperform manual implantation techniques [21–27]. Most reports of robotic-assisted UKA describe slightly improved component position and suggest better early outcomes with fewer outliers [22,25,26]. However, these reports are universally short term and fail to show definitive improvements in clinical outcome [23]. Again, such robotic navigation technology is extremely costly and requires acquisition of additional preoperative three-dimensional cross-sectional imaging such as computed tomography (CT) scan with significant expense and radiation exposure to the patient. Our current health care environment is focused heavily on cost containment— therefore, expenses such as robotics for UKA surgery must be justified as regards improved both clinical efficacy and cost-effectiveness. Therefore, as a result of conflicting current literature regarding the efficacy of robotic-assisted UKA, we designed a research study to address three specific questions that compare the use of roboticassisted UKA versus manually implanted. First, we examined whether robotic UKA resulted in significant increases operating room time and/or patient length of stay (LOS) compared with manual UKA. Second, we determined whether there are any significant radiographic differences in the component placement and reproducibility of the UKA procedure between the two techniques. Finally, our last question was to determine whether there are any demonstrable improvements short-term clinical differences or patient performance as a result of robotic-assisted UKA. Materials and Methods Following institutional review board (IRB) approval, we initiated a retrospective and consecutive review of 30 robotic-arm assisted medial 0883-5403/© 2014 Elsevier Inc. All rights reserved.

Please cite this article as: Hansen DC, et al, Robotic Guidance Does Not Improve Component Position or Short-Term Outcome in Medial Unicompartmental Knee Arthroplasty, J Arthroplasty (2014),


D.C. Hansen et al. / The Journal of Arthroplasty xxx (2014) xxx–xxx

compartment UKA cases versus 32 manual medial compartment UKA cases. Patients were assigned to each group based on patient preference. Patients who presented to the practice of the primary surgeon (SKK), who were candidates for medial UKA and who specifically requested a robotically performed procedure subsequently underwent a robotically assisted UKA. Patients who did not specifically request a robotic procedure underwent a manual UKA. In order to reduce any bias that may have resulted from patients specifically requesting a robotic procedure, data collection was done in a completely blinded fashion both with regard to radiographic and clinical parameters. Patient confidentiality and anonymity was maintained and all reviewers were blinded to any patient specific identifiers. Cases for this study were performed between June 2009 and July 2011 and selected in chronologic, consecutive sequence to eliminate bias of selection prior to review. An important consideration for this analysis is that all the robotic cases included in this review were performed after the senior author had completed approximately 75 robotically assisted medial UKA. This was done to ensure that minimal “learning curve” issues would exist when analyzing the robotic cohort. All patients in the study had minimum follow-up of 24 months postoperatively. An identical surgical technique was used by the senior author for both the manual and robotic-assisted UKA procedures; all cases were performed by the senior author, a fellowship-trained, joint arthroplasty surgeon with significant previous experience in manual UKA techniques. Identical and consistent selection criteria were utilized for all patients, such that patients with medial osteoarthritis, an intact anterior cruciate ligament, with an anteromedial tibial arthritis wear pattern and a correctible deformity confirmed by varus/valgus stress radiography were included in the study. The degree of patellofemoral disease was not used as a selection parameter for any patients in the study. The surgical technique in cases utilized a median parapatellar approach that was performed after induction of spinal or general endotracheal anesthesia. A tourniquet at a pressure of 275 mm Hg was used. General anesthesia was introduced only in cases where the attending anesthesiologist was unable to successfully introduce the spinal anesthetic into the patient's spinal column. In both the robotic and manual cases, the surgical technique included the medial exposure, followed by careful visual inspection of the lateral and patellofemoral compartments for evidence of arthritis. Care was taken to avoid release of the superficial and deep medial collateral ligament. Meticulous osteophyte removal was performed medial tibia and femur adjacent to the MCL. Careful osteophyte removal was also performed in the intercondylar notch. All components in both study groups were fixed with Palacos antibiotic-impregnated cement and included dual-peg femoral and tibial components. Prior to closure, all patients received a 60-cc intraarticular injection of 0.25% Marcaine with 1/2000 epinephrine combined, followed by an identical wound closure technique in all patients.

for a goal of 2 to 3 mm of laxity with the knee at 90 degrees. The femoral component size was selected, posterior condylar and chamfer resections were performed, trial components were inserted and a recheck of flexion/extension gap balance was performed. The final components were then implanted and closure performed. A highflexion, metal-backed tibial component design was used in all cases (Zimmer High-Flex UKA, Zimmer, Inc., Warsaw, IN). The femoral and tibial components are designed with a dual peg to enhance component fixation stability.

Manual UKA Procedure

Full hospital and clinic medical record review of demographic, preoperative, intraoperative and postoperative measures was performed. Radiographic analysis of preoperative and postoperative images evaluating sagittal and coronal alignment, and component positioning were performed by two fellowship-trained orthopedic surgeons with significant experience in UKA surgery and in radiographic analysis using the FDA-approved OsiriX imaging system (Pixmeo; Geneva, Switzerland). Full-length coronal images through the hip, knee, and ankle were used for all radiographic analysis both preoperatively and postoperatively. The included figures and legend demonstrate the measurement protocol (Fig. 1A–B).

In the manual UKA cohort, all cases were performed using a fully extramedullary referencing technique which, for coronal alignment, utilizes three anatomic landmarks—the tibial crest, the center of the ankle, and the tibial eminence. Utilizing digital radiography, the tibial slope resection reference was determined from the radiographic measurement of anatomic slope, and through additional visual exposure of the medial aspect of the tibia during placement of the tibial alignment guide. Reference for the tibial resection depth was identified by the greatest depth of tibial plateau wear (defect), with a subsequent target resection of 1 to 2 mm below this point. Following tibial resection, the extension gap was checked with a minimal target of 8 mm, and the flexion gap was assessed for posterior femoral resection of 6.5 mm with 1 to 1 degrees of varus undercorrection and 1 to 2 mm of laxity with the knee at 20-degree extension. Following distal femoral resection, the posterior femoral condyle was rechecked

Robotic-Assisted UKA Procedure For the robotic UKA procedure, all patients underwent a preoperative CT scan from the hip, through the knee and ankle. This CT scan was then downloaded into the robotic-assisted software platform for preoperative implant planning. Following an exposure technique that was identical to that used for the manual UKA cohort, a haptic robotic arm and computer guidance system was used (RIO™ Robotic Arm Interactive Orthopedic System, Mako Surgical Corporation, Fort Lauderdale, FL) that required computer registration of the real tibial and femoral joint line bone surfaces and tibial and femoral mechanical axes. Intraoperatively, tibial and femoral tracker devices for robotic registration were placed in the tibial and femoral diaphyses. For mechanical axis determination, the robotic system utilizes the center of the ankle as determined by the midpoint between the extreme medial and lateral points of the medial and lateral malleoli respectively, and all landmarks are registered and stored for reference. The center of the hip is determined by taking the hip through a large, circular range of motion for approximately 10–15 cycles. Following bone registration, all osteophytes were removed followed by application of valgus stress to knee at 20-degree intervals from full extension to 120 degrees of flexion. The data acquired were used to generate a ligament balancing curve that was then used to virtually manipulate position of femoral and tibial components within the robotic software platform to achieve balanced flexion and extension gaps. Once the “virtual” gap balance has been achieved, the haptic robotic arm is used to first burr the distal and posterior femoral surfaces followed by burring of the tibial plateau. All trial components are then inserted, and both gap balance and mechanical axis undercorrection are assessed using robotic software. Any adjustments in gap balance or implant alignment were performed by using repeated burring if necessary. Implantation of final UKA components was performed with a dual-pegged femoral and dualpegged, metal-backed, tibial component, designed and manufactured by the robotic device corporation (Restoris UniCompartmental Knee System, Mako Surgical Corporation, Fort Lauderdale, FL). Medical Record/Radiographic Review

Statistical Methods Independent-samples t-tests were used to compare the two groups on all continuous variables. Chi-square (χ 2) tests were used for categorical variables. Variables used to assess accuracy were

Please cite this article as: Hansen DC, et al, Robotic Guidance Does Not Improve Component Position or Short-Term Outcome in Medial Unicompartmental Knee Arthroplasty, J Arthroplasty (2014),

D.C. Hansen et al. / The Journal of Arthroplasty xxx (2014) xxx–xxx

calculated as absolute values to avoid averaging across zero. Levene's test of equality of variances was used to determine whether the dispersion of values within each group differed significantly. In the case of a significant Levene's test, a separate variances t-test was used to compare equality of means. In addition, an alpha level of 0.05 was used as the decision point for statistical significance. All analyses were performed using the SPSS version 20 software package (IBM SPSS Statistics, IBM Corporation, Armonk, NY).

Results Demographic and Intraoperative Data With regard to demographics comparisons between groups, there was no statistical difference in patient gender (P = 0.132), age (P = 0.207), height (P = 0.454), weight (P = 0.661), BMI (P = 0.398), alcohol use (P = 0.472) or tobacco use (P = 0.658). All descriptive statistical review results of demographic data are summarized in Table 1. The distribution of anesthesia type (general endotracheal, spinal) was statistically similar across both groups (P = 0.560). Average operative time differed significantly between the two cohorts, with operative time being 20 minutes greater in roboticassisted UKA group (P = 0.010). However, there were no statistical


differences seen in all other intraoperative measures including tourniquet use (P = 0.076), average tourniquet time (P = 0.080) and average estimated blood loss (P = 0.135). Postoperative Clinical Data Within the immediate and early postoperative time through hospital discharge, there was no statistical difference in average time in recovery room (P = 0.263), average time to first ambulation (P = 0.490), average postoperative hematocrit (P = 0.800) or hemoglobin (P = 0.962). Average postoperative range of motion (ROM) on the day of surgery was statistically greater in the roboticassisted group (P = 0.045), but there was no statistical difference in ROM on postoperative day 1 (P = 0.871) or day 2 (P = 0.386). Distance of first postoperative ambulation was significantly greater in the robotic-assisted group (P = 0.027). The difference in average time to physical therapy clearance was 10.3 hours less for the roboticassisted group which was a significant finding (P = 0.024), and average hospital length of stay (LOS) was 8.8 hours less for the robotic-assisted group which trended toward statistical significance (P = 0.066). ROM at 2 weeks postoperative was greater in the manual group (P = 0.043) Radiographic Data All radiographic assessments are also summarized in Table 1. Preoperatively, both groups demonstrated homogeneity with regard to tibial axis, femoral axis and tibial slope. Postoperatively, there was no statistically significant difference in coronal tibial axis or alignment (P = 0.184). However, the ability to recreate the preoperative femoral axis was statistically significantly improved in the robotic group (P = 0.013). The robotic group also had 2 degrees less average posterior slope of the tibial component which was statistically significant (P = 0.001), but accuracy in recreation of native slope was similar between groups (P = 0.409). The assessment of the average accuracy of placement of the tibial component to the medial plateau in the coronal plane (P = 0.076), the average tibial resection depth (P = 0.094) and average change in joint line (P = 0.902) were not statistically different between groups. The robotic group had a significantly larger variance in coronal alignment of the tibial component (P = 0.037). Additionally, average medial overhang of tibial component was statistically significantly greater (P b 0.001) and more variable (P b 0.001) in the manual group. Complications One patient (manual group) presented with a deep postop infection requiring early debridement and revision to TKA at 6 months postop. Two patients (one robotic, one manual) had postoperative cellulitis requiring antibiotic treatment only. Additionally, continued medial-sided knee pain was reported more commonly in the robotic group compared to manual group (6 patients, 20% vs. 1 patient, 3.3%. P = 0.041). Discussion

Fig. 1. 1A: Preoperative AP and lateral radiographs: (A) preop tibial coronal axis; (B) preop femoral coronal axis; (C) preop tibial slope; (*) preop distance from MTS to joint line. 1B: Postoperative AP and lateral radiographs: (D) postop tibial coronal axis; (E) postop femoral coronal axis; (F) Tibial implant coronal alignment; (G) postop tibial slope; (^) postop distance from MTS to joint line (used to calculate joint line recreation); (#) postop MTS to base of implant (used to calculate tibial resection depth).

This report provides a highly detailed retrospective comparative clinical and radiographic analysis of robotic and manually implanted medial UKA in two matched groups of patients. While both techniques resulted in reproducible and excellent outcomes with low complication rates, the results demonstrate little to no clinical or radiographic difference in outcomes between these two patient cohorts that can justify the routine use of costly robotic systems for standard medial UKA surgery. Our first study question was primarily a temporal one in that we determined what differences exist with regard to operative time and

Please cite this article as: Hansen DC, et al, Robotic Guidance Does Not Improve Component Position or Short-Term Outcome in Medial Unicompartmental Knee Arthroplasty, J Arthroplasty (2014),


D.C. Hansen et al. / The Journal of Arthroplasty xxx (2014) xxx–xxx

Table 1 Patient Demographics and Significant Intraoperative, Postoperative, and Radiographic Measurement Data.

Age (yr) Height (in) Weight (lb) BMI Gender Alcohol use Tobacco use Length of surgery (hr) Length of first ambulation (ft) ROM day 0 (deg) ROM 2 wks (deg) Time to inpt PT clearance (hr) Implant coronal alignment Postop tibial slope Femoral axis change Medial tibial overhang

M ± SD [Range] M ± SD [Range] M ± SD [Range] M ± SD [Range] %Male (n) %Yes (n) %Yes (n) M ± SD [Range] M ± SD [Range] M ± SD [Range] M ± SD [Range] M ± SD [Range] M ± SD [Range] M ± SD [Range] M ± SD [Range] M ± SD [Range]

length of stay (LOS) between these two techniques. We did find that the robotic procedure added significant time to the surgical procedure (P = 0.020, Table 1). However, this added time did not result in meaningful improvements in component position or patient outcome. Even with this increase in operative time, there was no significant difference in EBL, tourniquet time or intraoperative complications. With regard to LOS data, we did find that patients in the robotic group were discharged about 9 hours earlier than manually implanted patients, but this small difference in discharge time is unlikely to have clinical significance. We attempted to grossly quantify the costs of this increased OR time and shorter LOS that was observed in the robotic group. While it is difficult to determine actual costs at our institution, the state of Ohio requires hospitals to publish patient charges for various services (OhioHealth Patient Price List, July 1, 2012). At our institution, operative room charges are determined across five increasing levels of surgical complexity. Each surgical case based on complexity level is then broken into the first unit (30 minutes) of operating room time (range: $3957 to $9104) and then into 15-minute units thereafter (range: $1233 to $4610). Since our operating room charges are based increments of 15-minute units after the first 30 minutes, the 20 minutes of average additional time would yield two full extra 15-minute units with costs ranging from $2466 to $9220 based on surgical complexity. Floor time at our institute is based on rounding to the next full-day charge regardless of fraction of the day on the floor. Therefore, 9 hours of decreased LOS would only lead to a decrease in room charge if it bridged into the previous day. However, our floor nurse cost is charged at $90 per hour. Therefore, the 9-hour savings in LOS would result in a decrease of $810 in floor nursing charges. Unfortunately, current estimates of operating room and LOS costs are calculated very differently across institutions and would result in varying ranges of costs loss or benefit based on unit increments and associated fee calculations. Our second question was to determine if there are there any radiographic differences in the component placement and reproducibility of the UKA procedure. Radiographic evaluation showed minimal differences between groups. Most notably, there was no

Robotic (n = 30)

Manual (n = 32)


57.13 ± 9.81 [30–74] 66.43 ± 4.80 [53–72] 202.41 ± 41.19 [122–302] 32.13 ± 5.49 [25–50] 53.3% (16) 36.7% (11) 23.3% (7) 1.68 ± 0.25 [1.33–2.35] 43.50 ± 47.77 [2–150] 69.08 ± 15.93 [35–92] 94.81 ± 10.96 [70–130] 42.17 ± 14.55 [21.50–68.93] 1.64 ± 1.30 [0.03–4.24] 83.92 ± 1.92 [79.59–88.05] 1.69 ± 1.37 [0.06–5.04] 0.014 ± 0.035 [0.00–0.14]

60.66 ± 11.78 [42–81] 65.64 ± 3.42 [60–72] 206.95 ± 40.01 [142–309] 33.34 ± 5.70 [21–48] 34.4% (11) 28.1% (9) 18.8% (6) 1.48 ± 0.35 [1.15–3.05] 21.16 ± 27.99 [1–140] 54.81 ± 20.26 [0–85] 100.83 ± 10.91 [85–125] 52.47 ± 19.77 [23.73–117.40] 1.10 ± 0.94 [0.04–3.66] 81.66 ± 2.70 [76.40–86.45] 2.59 ± 1.98 [0.04–8.45] 0.132 ± 0.144 [0.00–0.44]

P = 0.207 P = 0.454 P = 0.661 P = 0.398 P P P P

= = = =

0.132 0.472 0.658 0.010

P = 0.027 P = 0.045 P = 0.043 P = 0.024 P = 0.037 P = 0.001 P = 0.051 P b0.001

significant difference in postoperative coronal tibial axis or coronal alignment of the tibial component. This differs from multiple published reports showing improved accuracy with robotic assistance [18–23]. Lonner [20] compared tibial component alignment in 27 patients receiving standard-instrument UKA and 31 receiving robotic-assisted UKA using the same robotic system as our study. They found that the error in posterior tibial slope (3.1 vs. 1.9 degrees) and coronal alignment of the component (2.7 vs. 0.2 degrees), were significantly greater in the manual group. Variance was also found to be 2.6 times greater with manual instrumentation in their study. Cobb et al [23] performed a randomized-controlled trial comparing a hands-on robotic system to standard UKA in 28 cases. They found that all robotic-assisted cases had less than 2 degrees of error in tibial component placement, while 40% of the manual group fell outside this threshold. However, neither of these two similar reports was able to demonstrate that these minor improvements in component position impacted macroscopic clinical outcomes in the short or medium term. In our study, tibial axis was found on average to be in approximately 2 degrees of varus in both groups. This represents a correction of approximately 3 degrees of varus from preoperative values. These findings are in line with the senior author's goals during medial UKA of slightly undercorrecting the knee to avoid overloading the lateral compartment. One unexpected finding was the increased variability of tibial implant coronal alignment seen in the robotic group. We feel that this finding is due to the fact that the software algorithm utilized by the robotic system is set to actually reproduce exactly the coronal alignment of the patient's native tibial slope with the tibial component. With the manual UKA procedure, the surgeon had the goal in every case of making the tibial cut perpendicular to the mechanical axis of the tibia. These subtle differences in technique could explain the higher variance of the robotically implanted tibial components. Tibial slope was found to be significantly greater in the manual group (P = 0.001, Table 1), but no difference was seen in recreation of each patient's preoperative tibial slope. Hernigou and Deschamps [28] evaluated tibial slope in 99 medial UKA procedures as a predictor for failure. They found that increased tibial slope correlated with increased risk of component loosening. This finding was accentuated

Please cite this article as: Hansen DC, et al, Robotic Guidance Does Not Improve Component Position or Short-Term Outcome in Medial Unicompartmental Knee Arthroplasty, J Arthroplasty (2014),

D.C. Hansen et al. / The Journal of Arthroplasty xxx (2014) xxx–xxx

in ACL-deficient knees. They suggested maintaining tibial slope less than 7 degrees in UKA. Other authors have tested and reported a variety of ranges in tibial slope recreation that change tibial load in UKA such that a single “optimum” reference for tibial resection angle may not be best for optimum component function and durability [29,30]. Ultimately, our data presented here indicate that with meticulous surgical technique, both robotic and manual methods of applying tibial slope prevent tibial cuts with slope that is outside the safe zone as described by Hernigou and Deschamps [28,31]. Medial tibial component overhang was significantly greater (P b0.001, Table 1) and more variable in the manual group. Although two different implant systems were indeed used in our study, our radiographic and component measurement methods were identical between the two groups. Additionally, as discussed previously, the two different component systems are highly similar with respect to shape and design characteristics. Therefore, we believe that the observation of component overhang differences between the two methods in this series can allow for some meaningful conclusions to be drawn. Other published data on medial UKA component overhang include that of Chau et al [32], who analyzed 160 UKAs for medial overhang of the tibial component. They divided patients into three groups: tibial underhang, minor overhang (less than 3 mm) and major overhang (3 or more millimeters.) They found that at 5 years after surgery, patients with major overhang had significantly worse Oxford Knee Scores and pain scores. A recent in vitro, cadaveric study by Gudena et al [33], looked at the affect of tibial component overhang on MCL load in a biomechanical model. They found no significant difference with 2 mm of overhang when compared to no overhang. Four millimeters of overhang was found to significantly increase MCL load when compared to no overhang and 2 mm of overhang. Due to these results, they suggested limiting medial overhang to less than 2 mm in UKA. In our study, average medial overhang was less than 2 mm in both groups, but we did have seven patients over this threshold in the manual group, with none in the robotic group. These findings did not correlate clinically however, as none of the patients with greater than 2 mm of overhang complained of continued medial-sided knee pain at short-term follow-up. Our final task was to determine if there are any measured shortterm clinical differences in outcome as a result of the use of roboticassisted UKA that affect patient performance. Overall, there were minimal and insignificant clinical differences seen between groups in our short-term study as highlighted in Table 1. The current literature provides very few detailed comparisons of clinical outcomes between robotic-assisted and manual UKA groups. When compared, minimal significant findings have been produced [25,26,34]. Our study yields very similar findings. Early inpatient rehabilitation findings did slightly favor the robotic group as day of surgery ROM and distance of first postoperative ambulation were improved. Also, time to PT clearance was significantly less (P = 0.024, Table 1) and length of stay trended (P = 0.066, Table 1) toward being significantly less in the robotic group. Interestingly, ROM at 2 weeks postoperation favored the manual group. Overall, postoperative rehabilitation showed minimal differences in the short term. Postoperative complications were minimal in both groups. One unexpected finding was the increase in continued medial-sided knee pain in the robotic group (20% of patients). This has not been previously described in studies evaluating this system. Our theory is that due to the irregular bone surface produced by the burr, cement interdigitation may not be as adequate or uniform when compared to saw preparation. Increased attention to cement technique in the robotic-assisted procedure could possibly decrease the incidence of this finding. There exist several limitations to this study that the authors wish to address. In combination, the single surgeon and retrospective nature of the study design do not carry the strength of a powered,


multi-centered, prospective, randomized, controlled trial. However, we feel that the fact that all surgeries were performed fully by the senior author with identical patient selection criteria and nearly identical surgical technique does offer a unique opportunity to assess the efficacy of robotic technology in the hands of an experienced UKA surgeon. It also offers the ability to look for subtle differences that may exist between both techniques in a carefully controlled group of patients. Studies such as the one performed here are very helpful in determining the initial cost-effectiveness and efficacy of such expensive technologies, and can also establish if a need even exists for the performance of expensive, multicenter randomized trials to assess robotic technology. As with our case series findings, even though retrospective, these results help to build upon the information necessary for overall informed decision making of available and promoted technologies. The results presented here support the conclusion that such multicenter trials may not be needed, as our data show very little differences between robotic and manually implanted medial UKA. Another potential study limitation is the relatively small number of cases included in our assessment of the haptic robotic technology. However, in comparison with current literature on robotic-assisted UKA, we do present a relatively large patient group here with highly detailed radiographic and clinical analysis. Additionally, unlike primary TKA, medial compartment UKA has a well-defined and established standard of patient selection criteria for optimizing postoperative outcomes. Following the demographic, clinical and radiographic guidelines within our practice parameters and applied to UKA patient candidates, our sequential selection of UKA cases for retrospective medical record review yielded a sample population that was statistically similar across both groups. This clearly reflects the narrow patient selection criteria that have been established and adopted for UKA by our practice, and allows for confidence in the assessment of this technology application in our hands and within our practice parameters. One final weakness of the study is the use of two different implant systems. This weakness is unavoidable due to the fact that at the time of performance of these cases and collection of data, only one robotic UKA system was FDA approved for use in the United States, and this closed robotic platform allows use of a proprietary implant system that is specific to this robotic platform. However, both implant systems used possess highly similar design characteristics, including a cemented dualpeg cobalt chrome femoral component with a round on flat cemented metal back tibial design. The tibial components also shared similar characteristics including a dual-pegged, lateral fin design. Therefore, based on the results of the data here, we conclude that the purported advantages of robotic UKA implantation are obviated in the hands of a surgeon with extensive training and experience in manual UKA implantation. Our minimal clinical and radiographic differences lend to the argument that it is difficult to justify the routine use of expensive robotic techniques for standard medial UKA surgery, especially in a well-trained, high-volume surgeon. The intent of our report of findings is to add to the experience and literature associated with the currently growing knowledge base, so that surgeons across the peer community have access to a complete profile of use and expectations of this technology. Further surgical, clinical and economical study of this technology is necessary.

Reference 1. Berger RA, Della Valle CJ. Unicompartmental knee arthroplasty: indications, techniques, and results. Instr Course Lect 2010;59:47. 2. Eickmann TH, Collier MB, Sukezaki F, et al. Survival of medial unicondylar arthroplasties placed by one surgeon 1984–1998. Clin Orthop Relat Res 2006;452:143. 3. Hollinghurst D, Stoney J, Ward T, et al. No deterioration of kinematics and cruciate function 10 years after medial unicompartmental arthroplasty. Knee 2006;13(6):440. 4. Jamali AA, Scott RD, Rubash HE, et al. Unicompartmental knee arthroplasty: past, present, and future. Am J Orthop (Belle Mead NJ) 2009;38(1):17.

Please cite this article as: Hansen DC, et al, Robotic Guidance Does Not Improve Component Position or Short-Term Outcome in Medial Unicompartmental Knee Arthroplasty, J Arthroplasty (2014),


D.C. Hansen et al. / The Journal of Arthroplasty xxx (2014) xxx–xxx

5. Lustig S, Elguindy A, Servien E, et al. 5- to 16-year follow-up of 54 consecutive lateral unicondylar knee arthroplasties with a fixed-all polyethylene bearing. J Arthroplasty 2011;26(8):1318. 6. O'Donnell T, Neil MJ. The Repicci II(R) unicondylar knee arthroplasty: 9-year survivorship and function. Clin Orthop Relat Res 2010;468(11):3094. 7. Robertsson O. Unicompartmental arthroplasty. Results in Sweden 1986–1995. Orthopade 2000;29(Suppl 1):S6. 8. Romanowski MR, Repicci JA. Minimally invasive unicondylar arthroplasty: eightyear follow-up. J Knee Surg 2002;15(1):17. 9. Swank M, Stulberg SD, Jiganti J, et al. The natural history of unicompartmental arthroplasty. An eight-year follow-up study with survivorship analysis. Clin Orthop Relat Res 1993;286:130. 10. Tabor Jr OB, Tabor OB. Unicompartmental arthroplasty: a long-term follow-up study. J Arthroplasty 1998;13(4):373. 11. Epinette JA, Brunschweiler B, Mertl P, et al. Unicompartmental knee arthroplasty modes of failure: wear is not the main reason for failure: a multicentre study of 418 failed knees. Orthop Traumatol Surg Res 2012;98(6 Suppl):S124. 12. Hamilton WG, Ammeen D, Engh Jr CA, et al. Learning curve with minimally invasive unicompartmental knee arthroplasty. J Arthroplasty 2010;25(5):735. 13. Kim KT, Lee S, Kim TW, et al. The influence of postoperative tibiofemoral alignment on the clinical results of unicompartmental knee arthroplasty. Knee Surg Relat Res 2012;24(2):85. 14. Lewold S, Robertsson O, Knutson K, et al. Revision of unicompartmental knee arthroplasty: outcome in 1,135 cases from the Swedish Knee Arthroplasty study. Acta Orthop Scand 1998;69(5):469. 15. Mariani EM, Bourne MH, Jackson RT, et al. Early failure of unicompartmental knee arthroplasty. J Arthroplasty 2007;22(6 Suppl 2):81. 16. Collier MB, Eickmann TH, Sukezaki F, et al. Patient, implant, and alignment factors associated with revision of medial compartment unicondylar arthroplasty. J Arthroplasty 2006;21(6 Suppl 2):108. 17. Lecuire F, Berard JB, Martres S. Minimum 10-year follow-up results of ALPINA cementless hydroxyapatite-coated anatomic unicompartmental knee arthroplasty. Eur J Orthop Surg Traumatol 2014;24(3):385. 18. Citak M, Suero EM, Citak M, et al. Unicompartmental knee arthroplasty: Is robotic technology more accurate than conventional technique? Knee 2013;20(4):268. 19. Jenny JY, Ciobanu E, Boeri C. The rationale for navigated minimally invasive unicompartmental knee replacement. Clin Orthop Relat Res 2007;463:58.

20. Lonner JH. Indications for unicompartmental knee arthroplasty and rationale for robotic arm-assisted technology. Am J Orthop (Belle Mead NJ) 2009;38(2 Suppl):3. 21. Pearle AD, O'Loughlin PF, Kendoff DO. Robot-assisted unicompartmental knee arthroplasty. J Arthroplasty 2010;25(2):230. 22. Lonner JH, John TK, Conditt MA. Robotic arm-assisted UKA improves tibial component alignment: a pilot study. Clin Orthop Relat Res 2010;468(1):141. 23. Cobb J, Henckel J, Gomes P, et al. Hands-on robotic unicompartmental knee replacement: a prospective, randomised controlled study of the acrobot system. J Bone Joint Surg (Br) 2006;88(2):188. 24. Cossey AJ, Spriggins AJ. The use of computer-assisted surgical navigation to prevent malalignment in unicompartmental knee arthroplasty. J Arthroplasty 2005;20 (1):29. 25. Konyves A, Willis-Owen CA, Spriggins AJ. The long-term benefit of computerassisted surgical navigation in unicompartmental knee arthroplasty. J Orthop Surg Res 2010;5:94. 26. Ma B, Rudan J, Chakravertty R, et al. Computer-assisted FluoroGuide navigation of unicompartmental knee arthroplasty. Can J Surg 2009;52(5):379. 27. Rosenberger RE, Fink C, Quirbach S, et al. The immediate effect of navigation on implant accuracy in primary mini-invasive unicompartmental knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2008;16(12):1133. 28. Hernigou P, Deschamps G. Posterior slope of the tibial implant and the outcome of unicompartmental knee arthroplasty. J Bone Joint Surg Am 2004;86-A(3):506. 29. Gulati A, Chau R, Simpson DJ, et al. Influence of component alignment on outcome for unicompartmental knee replacement. Knee 2009;16(3):196. 30. Iesaka K, Tsumura H, Sonoda H, et al. The effects of tibial component inclination on bone stress after unicompartmental knee arthroplasty. J Biomech 2002;35(7):969. 31. Weber P, Crispin A, Schmidutz F, et al. Improved accuracy in computer-assisted unicondylar knee arthroplasty: a meta-analysis. Knee Surg Sports Traumatol Arthrosc 2013;21(11):2453. 32. Chau R, Gulati A, Pandit H, et al. Tibial component overhang following unicompartmental knee replacement—does it matter? Knee 2009;16(5):310. 33. Gudena R, Pilambaraei MA, Werle J, et al. A safe overhang limit for unicompartmental knee arthroplasties based on medial collateral ligament strains: an in vitro study. J Arthroplasty 2013;28(2):227. 34. Seon JK, Song EK, Park SJ, et al. Comparison of minimally invasive unicompartmental knee arthroplasty with or without a navigation system. J Arthroplasty 2009;24(3):351.

Please cite this article as: Hansen DC, et al, Robotic Guidance Does Not Improve Component Position or Short-Term Outcome in Medial Unicompartmental Knee Arthroplasty, J Arthroplasty (2014),

Robotic guidance does not improve component position or short-term outcome in medial unicompartmental knee arthroplasty.

We performed a retrospective review in a matched group of patients on the use of robotic-assisted UKA implantation versus UKA performed using standard...
413KB Sizes 0 Downloads 4 Views