Improved Perioperative Outcomes in Robotic-Assisted Revision Total Knee Arthroplasty

Abstract

Background

This study compares short-term outcomes of robotic-assisted revision total knee arthroplasty (RA-rTKA) to conventional rTKA. We hypothesize that RA-rTKA will accelerate gains in range of motion and time to ambulation postoperatively.

Methods

This is a retrospective case-control study reviewing consecutive rTKA performed by a single surgeon between 2017 and 2024. rTKA cases performed prior to 2022 were compared to RA-rTKA performed from 2022 through present day. Revisions for periprosthetic joint infection or fracture were excluded. The primary outcome was hospital length of stay. Secondary outcomes included physical therapy (PT) metrics, blood loss, surgical time, and complications. Data collected included demographics, surgical and implant data, in-hospital PT progress, and outcomes through a minimum of 1 year.

Results

Sixty-six revision cases (42 rTKA and 24 RA-rTKA) were included with an average age of 67.7 years. Etiologies included loosening (42), second-stage reimplantation (12) after infection eradication, polyethylene wear (6), instability (6), and other etiologies (6). RA-rTKA case time averaged 27 minutes less than conventional; P = .18. The RA-rTKA cohort ambulated further on postoperative day 1 compared to the rTKA group (166.3 vs 87.2 feet; P = .01), was cleared by PT for discharge sooner (2.1 vs 3.1 days; P < .01), and had a shorter hospital length of stay (2.5 vs 3.6 days, P = .01). While all patients in both cohorts achieved at least 110° knee flexion by 6 weeks, RA-rTKA patients demonstrated significantly more knee flexion (119° vs 110°; P = .05). At a minimum of 1-year follow-up, no RA-rTKA patients required rerevisions, compared to 2 rTKA patients.

Conclusions

In this study, RA-rTKA showed improved ambulation in the immediate postoperative period, decreased hospital length of stay, and overall increased knee range of motion. These improvements were realized without increases in complications or operative time.

Introduction

Revision total knee arthroplasty (rTKA) can be challenging compared to primary total knee arthroplasty (TKA), complicated by a number of different factors. These factors include bone loss, loss of anatomical landmarks for joint line restoration, higher infection rates, and challenges in soft tissue management [[1], [2], [3], [4]]. A lack of bony landmarks can create challenges in restoration of joint line, accurate implant placement, and proper knee balancing during revision surgery.

Computer-navigated and robotic-assisted (RA) TKA has gained popularity in the last decade [5]. The literature regarding these technologies is mixed: while some studies have demonstrated improved accuracy and precision of bone cuts and radiographic outcomes, others have noted minor to no differences in clinical outcomes [[6], [7], [8], [9], [10], [11], [12]]. While the literature continues to develop around use of these technologies for primary TKA, their use for revision procedures has only recently been described. Few studies have investigated the use of robotics in the conversion of unicompartmental knee arthroplasty to TKA, in rTKA, and in 1.5-stage exchange revision for the management of septic TKA [[13], [14], [15]]. Innocenti et al. performed a recent systematic review comparing technology-assisted revision to conventional techniques and found that while technology reduced radiographic outliers, these improvements did not correlate with improved clinical outcomes [16]. Still, these improvements warrant further investigation.

The objective of this study was to compare perioperative and postoperative outcomes including range of motion (ROM), ambulation, and length of stay (LOS) for RA-rTKA to conventional rTKA. This study also aims to compare operative factors including blood loss, surgical time, and complications. We hypothesize that RA-rTKA will accelerate return of ROM and time to ambulation compared to rTKA.

Materials and Methods

This case-control study reviews a consecutive series of rTKA cases performed by a single fellowship-trained arthroplasty surgeon. IRB approval was obtained prior to the commencement of the study. Inclusion criteria encompassed patients undergoing rTKA of both femoral and tibial components for noninfectious diagnoses or reimplantation after infection eradication between 2017 and 2023. Patients were excluded if they had a follow-up period of less than 1 year, had a single-component exchange, a revision to a hinged prosthesis, or if the surgical indication included a diagnosis of periprosthetic joint infection or fracture. Starting in 2022, these cases were performed using a novel technique based on computed tomography (CT)-guided navigation and Stryker MAKO robotic assistance (Stryker, Kalamazoo, MI) [13]. This experimental cohort was retrospectively compared with conventional cases performed prior to 2022 without the assistance of robotics or computer navigation.

The surgical technique used for RA-rTKA involves a preoperative MAKO-protocol CT scan, with particular attention paid during segmentation to the bone deep to existing implants, which will serve as the foundation for new implants. Intraoperatively, the knee is exposed using a medial parapatellar approach and an extensive synovectomy is performed. Femoral and tibial arrays are placed using standard technique, with primary implants still in place. Registration of the knee is performed using landmarks described in a prior publication (Fig. 1) [13]. Bone and implant registration is completed and followed by ligament balancing, positioning the implants in a position that will balance the knee, restore the joint line, and correct deformity (Fig. 2). The surgeon then uses the computer planning template to anticipate bone loss and plan for augments to enable the new implant plus augment to contact bone on the anterior, posterior, and distal aspects of the implant (Figs. 3 and 4). After planning is complete, the implants are removed in standard fashion. Stable bone deep to the removed implants is identified and digitally referenced with planned resection levels and the bone is reosteotomized using the RA arm. Trial components with augments are then placed, and ROM and stability are assessed. When deemed satisfactory, final implants are placed.

Figure 1.

Model demonstrating points used for knee registration with implant in place.

Figure 2.

Implant planning screen after registration, after adjustments to implant positioning. Red arrows indicate the difference between the supportive bone deep to the implant and the cement mantle/old implant predicting the bone loss to be addressed.

Figure 3.

Implant planning screen after adjustment of femoral component to address distal femur defect. The component is moved proximal until resting on supportive distal femoral bone, indicated by the red arrows. The increased distal femur resection, compared to Figure 2, corresponds to 5 mm augments on medial and lateral sides.

Figure 4.

Implant planning screen after adjustment of femoral component to address posterior femur defect. The component is moved anterior until resting on supportive posterior femoral bone. The increased posterior resection, compared to Figure 2, corresponds to 5 mm augments on posterior medial side (the lateral condyle is not visualized in this image).

The surgical technique employed for conventional rTKA involves a similar process as the RA-rTKA, but without the use of arrays and digital registration of the knee. In conventional rTKA, the implants are removed, and bone loss and joint line are visually estimated from landmarks including the femoral epicondyles, tibial tubercle, fibula head, meniscal scar, and other anatomic landmarks. A series of intramedullary guides are used to prepare the bone ends for implants and augments. Trial components are then used to assess joint balance until a satisfactory combination of implants and augments is determined.

Over the course of the study period, no changes were made to perioperative protocols other than the change in surgical technique. All patients were routinely admitted postoperatively and were seen by physical therapy (PT) beginning postoperative day (POD) 1. Both postoperative pain management protocols and requirements for PT clearance were unchanged throughout the course of the study.

Data collected included patient demographics, comorbidities, surgical and implant data, knee ROM at preoperatively, postoperatively at 2 weeks, and 6 weeks, complications, and reoperation rates at 2 weeks, 6 weeks, 6 months, 1 year, and 2 years postoperatively, in-hospital PT progress, time to PT clearance for discharge, LOS, 30- and 90-day reoperation rate, and time to most recent follow-up. ROM measurements were made in the office using a goniometer. Surgical data, including estimated blood loss and surgical time, were collected from the anesthesia record. Data on ambulation, distance, and clearance for discharge were collected from the official PT documentation. Preoperative bone defects of the femur and tibia were assessed on radiographs using the Anderson Orthopaedic Research Institute classification system.

Data were analyzed using descriptive statistics. Data recording and statistical analysis were performed using Microsoft Excel (Microsoft, Redmond, WA, USA). Continuous variables were compared using a two-sample t-test, while categorical variables were compared using Fisher’s exact test or a chi-square test where appropriate. P values of less than 0.05 were considered statistically significant.

Results

A total of 89 patients undergoing rTKA in the study period were identified. Twenty-three patients were excluded for preoperative diagnosis of periprosthetic fracture and chronic periprosthetic joint infection, leaving 66 patients (48 females, 18 males) for analysis. Of these, 42 (63.6%) were performed for aseptic loosening, 12 (18.2%) were performed as second-stage reimplantation after clearance of infection, 6 (9.1%) were performed for polyethylene wear or failure, 6 (9.1%) were performed for instability, 3 (4.5%) were performed for unresolved pain, 2 (3.0%) were performed for component malposition, and 1 (1.5%) was performed for arthrofibrosis - indications for revision are included in Table 1. Twenty-four (36.4%) underwent xRA-rTKA, while 42 (63.6%) underwent rTKA. Demographic information for robotic and conventional cohorts is included in Table 2, with no significant differences noted between cohorts. The average follow-up time for the population is 47.2 months (SD ± 24.3 months).

Table 1.

Indications for revisions in each cohort.

Indication for revision	Robotic N = 24	Nonrobotic N = 42	Overall N = 66	P value (Fisher’s)
Loosening	18	24	42	0.1881
2nd stage PJI	2	10	12	0.1856
Implant recall	3	1	4	0.2966
Polyethylene wear or failure	3	3	6	0.6600
Instability	2	4	6	1.0000
Progression of OA or pain	1	2	3	1.0000
Malposition	2	0	2	0.1287
Arthrofibrosis	0	1	1	1.0000
Impending fracture	0	1	1	1.0000

ASA, American Society of Anesthesiologists; BMI, body mass index; CKD, chronic kidney disease; PVD, peripheral vascular disease; SD, standard deviation; DM, diabetes mellitus.

PJI, periprosthetic joint infection; OA, osteoarthritis.

Table 2.

Patient demographics.

Demographics	Robotic N = 24	Nonrobotic N = 42	Overall N = 66	P value
Mean age (SD)	67.8 (8.5)	67.7 (9.8)	67.7 (9.3)	.97
Sex
Female	17	31	48	.79
Male	7	11	18	.79
Mean BMI (kg/m2)	32.0	31.2	31.5	.49
ASA
4	0	0	0	1.00
3	16	19	35	.09
2	8	22	30	.13
1	0	1	1	1.00
Race
White	5	14	19	.40
Black	6	11	17	1.00
Asian	2	5	7	1.00
Hispanic/Latino	4	1	5	.06
Other/Multiracial	1	3	4	.65
Comorbidities
DM	6	7	13	.41
CKD	0	2	2	.53
PVD	0	1	1	1.00

ASA, American Society of Anesthesiologists; BMI, body mass index; CKD, chronic kidney disease; PVD, peripheral vascular disease; SD, standard deviation; DM, diabetes mellitus.

PJI, periprosthetic joint infection; OA, osteoarthritis.

In the robotic cohort, 16 patients (66.7%) had Anderson Orthopedic Research Institute type 1 bone loss; 8 (33.3%) had type 2 bone loss; and no patients had type 3 bone loss. In the conventional cohort, 21 patients (50%) had type 1 bone loss, 20 patients (47.6%) had type 2 bone loss, and 1 (2.4%) had type 3 bone loss. In the robotic cohort, 22 patients (91.7%) required augments and 16 (66.7%) required cones; in the conventional cohort, 31 (73.8%) required augments and 18 (42.9%) required cones—the difference between these was not significant (Χ² = 2.05;—P = .152).

In the robotic cohort, 10 patients received general anesthesia, while 14 received spinal anesthesia; in the conventional cohort, 21 received general anesthesia and 21 received spinal anesthesia. The difference between these was not significant (X² = 0.16; P = .70). All cases were performed using a tourniquet. No differences were found comparing RA-rTKA and rTKA for estimated blood loss during surgery (133.5 mL vs 193.4 mL; P = .174) and a surgical duration (171.7 minutes vs 198.7 minutes; P = .186). Also, no difference was found for blood transfusion requirement comparing RA-rTKA and rTKA (0.2 units vs 0.4 units; P = .452).

There was no significant difference in the average time to first ambulation between robotic vs conventional cohorts (0.7 days vs 0.9 days; P = .084), however patients in the RA-rTKA patients were able to ambulate a greater distance on POD 1 compared to those in the conventional cohort (166.3 feet vs 87.2 feet; P = .012). RA-rTKA demonstrated a shorter mean time to PT clearance (2.1 days vs 3.1 days; P = .002) and a shorter mean LOS than rTKA (2.5 days vs 3.6 days; P = .010). During outpatient follow-up, it was noted that there was no difference between RA-rTKA and rTKA measuring ability to ambulate without an assistive device by 2 weeks (0% vs 14.2%; P = .079) nor by 6 weeks postop (62.5% vs 52.4%; P = .43).

The average preoperative extension of the RA-rTKA group was 3.1°, compared to the rTKA group of 2.9°; the difference was not significant (P = .89). There was no significant difference in the mean arc of motion achieved at 2 weeks postoperatively between RA-rTKA vs rTKA cohorts (97.9° vs 94.0°; P = .398), but 23 of 24 (95.8%) RA-rTKA patients were able to achieve a minimum of 90° flexion by 2 weeks, compared to 42 of 57 (77.7%) rTKA patients; P = .007. At 6 weeks postoperatively, the RA-rTKA cohort achieved a significantly greater ROM compared to rTKA (118.5° vs 110.4°; P = .049), but there was no significant difference in the number of patients able to achieve flexion more than 110° by 6 weeks between the robotic cohort (22/22) and conventional cohort (34/36), P = .521. There were no pin site complications in the RA cohort. Within 90 days after surgery, one patient in the robotic cohort required readmission for superficial wound dehiscence without deep infection, treated with superficial wound irrigation and debridement and primary closure. Two patients in the conventional cohort required readmission for pulmonary embolism (1 patient) and superficial wound dehiscence (1 patient). At a minimum of 1-year follow-up, no RA-rTKA patients required rerevisions, compared to 2 rTKA patients (average follow-up 4.9 years) who required further revision surgery for aseptic loosening (1 patient) at year 2 and both aseptic loosening and polyethylene wear (1 patient) at year 5 after the initial revision.

Discussion

In this study, RA-rTKA was associated with significant improvements in achieving flexion to 90° at 2 weeks with a greater arc of motion by 6 weeks, an improved ability to ambulate a distance on the first POD facilitating earlier PT clearance and a shorter hospital LOS. In addition, RA-rTKA patients had shorter operative duration, decreased estimated blood loss, and fewer transfusions compared to their conventionally performed rTKA counterparts; however, these differences were not significant. Furthermore, the difference in complication rates between groups was not found to be statistically significant, suggesting that RA-rTKA may offer quicker gains in areas of function and hospital discharge without increasing risk for complications or revision surgery.

rTKA is a technically complex and demanding procedure often requiring advanced technical skills to correct a failed TKA. RA-rTKA is an emerging technique that may facilitate improved accuracy of bone preparation, precision of joint line restoration, and knee balance. Early studies have demonstrated that RA technology can be used to successfully perform rTKA, but there is still a paucity of literature comparing RA-rTKA with those undergoing revisions using the conventional approach.

Earlier publications have demonstrated early success at using RA surgery for RA-rTKA. MacAskill et al. and Steelman et al. were among the first to describe a RA technique for the revision of a primary TKA and reported short-term outcomes on a collective total of 3 patients, all of whom achieved full ROM by 6 weeks postop with uneventful recoveries [17,18]. More recently, a retrospective study by Cochrane et al. reported 90-day outcomes data on 115 RA-rTKA patients, finding a significant decrease in pain scores and depression scores from preoperative to final follow-up, but no significant difference in Patient-Reported Outcomes Measurement Information System Physical Function scores; authors did not report data on ROM throughout recovery [19]. Ngim et al. also reported that in a cohort of 19 RA-rTKA, the mean ROM achieved by the final follow-up (mean, 10.4 months) was 1.5 degrees flexion and 114 degrees extension, and all patients achieved independent ambulation within the follow-up period; all patients had uneventful recoveries with no rerevisions or infections [20].

RA-rTKA is also associated with additional enhanced outcomes compared to conventional rTKA. In this study, the RA-rTKA cohort showed an average LOS of 2.5 days, 4% 90-day complication rate, and 0% rerevision rate. In their study of 202 rTKA cases, Quinn et al. reported an average LOS of 7.6 days, with a 9% 90-day complication rate and 4% of cases requiring rerevision [21]. Nichols et al published on 25,354 rTKA, reporting an average LOS of 5.6 days and a 23% 90-day readmission rate [22], while Schairer et al. reported an average LOS of 4.6 days and a 13% 90-day readmission rate [23]. Quinn et al. also reported a mean knee ROM of 112 degrees by 1-year postop, while we found that RA-rTKA showed significantly more knee ROM by 6 weeks postoperatively of 119 degrees [21]. While these larger studies may be more heterogeneous with regard to case complexity and may not correlate perfectly with the revisions performed in our study group, their outcomes represent standards upon which our navigated technique may improve.

The few studies that do compare the outcomes of RA-rTKA to those of conventional rTKA, while modest, do indicate some significant differences. The systematic review by Innocenti et al. demonstrated fewer outliers in hip-knee-ankle angle, superior component positioning measured by lateral distal femoral and medial proximal tibial angles, and improved joint line restoration in the 10 studies reviewing rTKA and RA-rTKA [16]. This review also found that technology-assisted procedures required an additional 15-24 minutes, which our data did not corroborate [16].

The improvements seen in gains of motion and ambulation in this study may be attributable to the decreased soft tissue dissection and trauma associated with navigated-assisted and robotically assisted arthroplasty. Numerous studies have demonstrated decreased soft tissue releases and soft tissue trauma with navigated technologies in primary knee and hip arthroplasty [6,24,25]. Use of these technologies to recreate the joint line and bony landmarks digitally theoretically allows for improved accuracy and precision through smaller incisions and with less soft tissue dissection.

The present study is not without limitations. As a retrospective study, our analysis is subject to inherent biases associated with previously collected data and cannot detect direct causation between cohorts and the reported outcomes. Second, no patient-reported outcomes were collected, thus limiting the clinical conclusions that may be drawn from the data. Third, as with any novel technique, our sample size is relatively small at only 66 patients; the ability to detect meaningful differences in ROM, ambulation progress, and various operative details between cohorts is subsequently reduced. Furthermore, though differences between the study populations were not significant in our study, such differences may be clinically significant with a larger sample size. Fourth, the single-surgeon, consecutive-case cohort may cause a bias for improved outcomes in more recent cases given increased experience. Finally, PT-related outcomes such as ambulation were obtained through chart review; while no changes were made to PT protocols over the course of the study period, differences in PT providers, scheduling, and documentation may cause bias in reported outcomes.

Conclusions

rTKA can be successfully performed via conventional and novel RA techniques. In this study, RA-rTKA showed improved ability to ambulate after surgery with decreased hospital LOS and improved overall early knee ROM, without adding operative time or complications. Long-term outcomes are needed to evaluate additional benefits that may be realized with RA-rTKA techniques.

CRediT authorship contribution statement

Kevin C. Chang: Writing – review & editing, Writing – original draft, Formal analysis, Data curation, Conceptualization. Aleksandra Qilleri: Writing – original draft, Formal analysis, Data curation. Alexandra Echevarria: Writing – original draft, Data curation. Jonathan R. Danoff: Writing – review & editing, Project administration, Methodology, Investigation, Conceptualization.

Conflicts of interest

J.R. Danoff is a paid consultant for Surgical Specialties Corp, Stryker, and MicroGenDx, and is on the editorial board for Arthroplasty Today and American Association of Hip and Knee Surgeons; all other authors declare no potential conflicts of interest.

For full disclosure statements refer to https://doi.org/10.1016/j.artd.2025.101937.