1 Introduction

Osteoarthritis is a progressive and degenerative condition, and is one of the leading causes of disability. The knee is the largest synovial joint in the body and is a frequent site for osteoarthritis given its regular use and stress on the joint.¹ Symptomatic knee osteoarthritis affects 10 % of men and 13 % of women above the age of 60 and is expected to increase in view of our ageing population. Current management modalities are targeted towards symptom control until the degree of its impact on one's quality of life dictates the need for a joint replacement.² It has been proposed that the demand for knee arthroplasty is expected to grow by more than 600 % by 2030.³ In patients with unicompartmental knee osteoarthritis, the choice to perform a total knee replacement (TKR) versus a unicompartmental knee arthroplasty (UKA) are both established surgical options and is a joint decision between the patient and surgeon.⁴ UKA has gained popularity for its minimally invasive nature, decreased operative time, blood loss, length of hospital stay and morbidity rates.⁵ However, higher revision rates after UKA than TKR has been thought to be a key reason why more orthopaedic surgeons are not offering both procedures.⁶

Moreover, the use of Enhanced Recovery After Surgery (ERAS) protocols for knee arthroplasty has been shown to demonstrate faster return to activity, improved patient satisfaction, reduced complication rates, reduced length of stay and healthcare cost. ERAS protocols are a multidisciplinary and multimodal approach to optimizing patient care before, during and after the surgery to minimize surgical stress response. They involve peri-operative education, multimodal opioid-sparing analgesia and early mobilization.⁷ This was particularly valuable during the COVID pandemic when hospitals were all facing a bed crunch and great emphasis was placed on shortening hospital stays.⁸

This study aims to compare the 6 month functional outcomes of patients undergoing TKR versus UKA under the Enhanced Recovery After Surgery (ERAS) 23 hour protocol.

2 Methods

Ethics approval was granted by the local Institutional Review Board (CIRB 2023/2725). This study was performed in accordance with the ethical standards laid down in the 1975 Declaration of Helsinki.

Patients who underwent a unilateral primary TKR or UKA from August 2020 to July 2021 under the ERAS pathway were included in the study.

Patients who fulfilled the inclusion criteria of being American Society of Anesthesiologists (ASA) ≤ 3 and agreeable to be discharged within 23 hours were enrolled into the institution's ERAS protocol.

Pre-operatively, patients were counselled on their surgery and expected recovery by the surgeon. An anesthetist would then assess their perioperative risk and counsel them on general versus spinal anaesthesia. Smoking cessation counselling and iron supplementation for anemia were also given if indicated. Patients underwent pre-operative physiotherapy for pre-habilitation.

Peri-operatively, patients were fasted for at least 6 hours prior to the surgery and were given a carbohydrate drink containing 50g of glucose, oral etoricoxib and omeprazole (assuming no contraindications) 2 hours prior to surgery.

Intra-operatively, all patients were planned for spinal anaesthesia unless contraindicated (eg. previous spinal instrumentation). Prophylactic intravenous Cefazolin and intravenous dexamethasone 8 mg for post-operative pain and nausea and vomiting were given at the start of surgery. Unless contraindicated, patients received a periarticular infiltration and adductor block cocktail containing Ketorolac, Shincort, Vancomycin, Marcaine and Morphine intra-operatively. Intra-articular tranexamic acid was applied prior to capsule closure. Intravenous Ondansetron 8 mg was then given at the end of surgery for post-operative nausea and vomiting.

Post-operatively patients were given a combination of Etoricoxib, Paracetamol and Tramadol for pain control. For venous thromboembolic prophylaxis, all patients had mechanical calf pumps and oral aspirin was started on post-operative day (POD) 1. Patients were also reviewed by a physiotherapist on POD 0. They successfully completed the ERAS protocol if they were discharged home within 23 hours of their surgery.

Subsequently, patients had home visits by a physiotherapist on POD 7 and a home visit by nurses on POD 14.

Patient characteristics and comorbidities, 30-day post-operation readmission, infection and re-operation rates were analyzed. Comorbidities collected include diabetes, ischaemic heart disease, stroke, lung disease, arthritis outside of the knee, depression, hypertension, colitis, psoriasis, hyperlipidemia, Parkinson's disease, renal impairment, obesity and vascular disease. Patients' 30 day post-operative readmission, infection and re-operation rates were collected. An independent healthcare professional reviewed patients pre-operatively and 6 months post-operatively to measure their range of motion and functional outcomes - namely the Knee Society Function Score (KSFS), Knee Society Knee Score (KSKS), Oxford Knee Score (OKS) and both the Physical (PCS) and Mental component (MCS) of the Short-Form Health Survey (SF-36).

Statistical analysis was performed using STATA version 17. Chi-squared test was used to determine if there were differences in categorical characteristics and Mann-Whitney U test was used to test for differences in non-normally distributed variables between TKR and UKA patients. Finally, a multiple linear regression adjusted for age, gender and pre-operative outcomes scores were performed to compare the post-operative outcomes of patient who underwent TKR versus UKA under the ERAS protocol.

3 Results

Three hundred and forty-two TKR and 108 UKA patients were included in this study. UKA patients were on average 2.6 years younger (p = 0.006) but otherwise did not differ significantly from TKR patient in body mass index (BMI), gender or pre-existing comorbidities (Tables 1 and 2).

Univariate analysis showed significantly better pre-operative and 6 month post-operative KSKS and SF-36 MCS in UKA patients. There were no significant differences in the pre or post-operative KSFS, OKS or SF-36 PCS scores between the two groups (Table 3). Multivariate analysis showed no significant difference in all the 6 month post-operative functional outcome scores between UKA and TKR patients (Table 4)

UKA patients had significantly better pre and post-operative range of motion (p < 0.001). At 6 months, UKA patients had on average 16° more range of motion than TKR patients (p < 0.001) (Table 5).

Thirty day readmission, infection and re-operation rates greatly favoured patients who underwent UKA.

In the analysis of 30 day readmission rates, 2 UKA patients were readmitted, 1 for a fall with no periprosthetic fracture or loosening, and the other for a gastric ulcer likely precipitated by the use of non-steroidal anti-inflammatory agents. Nine TKR patients were readmitted, 2 for a prosthetic joint infection, 1 for haemarthrosis from anti-coagulant use, 1 for a fall resulting in partial tear of the vastus medialis oblique, 1 for a superficial surgical site infection, 1 for right calf pain for which no deep vein thrombosis was found, 1 for hospital acquired pneumonia, 1 for hyponatremia and 1 for per rectal bleeding (Table 6).

No UKA patients had an infection 30 days postoperatively. Meanwhile, amongst the TKR cohort, 2 had a prosthetic joint infection requiring debridement, antibiotics with implant retention (DAIR), 2 had a superficial surgical site infection and 1 had a hospital acquired pneumonia (Table 6).

Lastly, no UKA patients required re-operation within 30 days post-operatively, while 3 TKR patients required re-operation. Two were for the prosthetic joint infection mentioned above and 1 was a for a knee aspiration done in the operating theatre for haemarthrosis (Table 6).

4 Discussion

Our study has demonstrated that UKA and TKR patients have comparable KSFS, KSKS, OKS, SF-26 scores. UKA patients have a significantly larger range of movement and less early post-operative infective and revision rates than TKR patients.

UKA is a well-established bone and ligament sparing alternative to TKR in unicompartmental knee osteoarthritis. A systematic review and meta-analysis by Wilson et al. of 60 studies comparing UKA vs TKR outcomes echoed similar results. It was reported that UKAs have shorter operation duration (mean difference of 23.8 min), shorter hospital stay (mean difference 1.2 days), and fewer early complications. UKA patients had significantly lower rate of myocardial infarction (risk ratio 0.33), venous thromboembolism (risk ratio (0.39), deep infection requiring re-operation or long term antibiotics (risk ratio 0.59) and early mortality within the first 45 days (risk ratio 0.27). In terms of functional outcomes, UKA patients had better range of motion (mean difference 8.71°), increased likelihood of kneeling (risk ratio 0.53) – an outcome which would be held in high regard in our Asian population, higher functional patient reported outcome measures (mean difference −0.58). A systematic review and meta-analysis by Witjes et al. further reported that patients were more likely to return to sports after a UKA than TKR. The mean total number of sports patients participated post-UKA was 1.1–4.6 sports versus 0.2–1.0 post-TKR. Time to return to sports was also shorter post-UKA at 12 weeks versus 13 weeks post-TKR.⁹ Similar to Wilson et al., Arirachakaran et al. demonstrated a reduced risk of early postoperative complication and revision surgery was 60 % lower in UKA than TKR patients in a systematic review and meta-analysis of randomized control trials.¹⁰

This then begs the question of why aren't there more surgeons performing UKA for unicompartmental knee osteoarthritis? Despite the improvements in implant design and surgical techniques over the years, the multiple advantages of UKA do not correlate with its usage, likely due to concerns regarding implant survivability, patient selection and ideal bearing design.¹¹

Indeed, in different studies, UKA showed increased revision rates in the long run. Wilson et al. reported higher revision rates at the 5 (risk ratio 5.95), 10 (risk ratio 5.18) and 15 year mark (risk ratio 1.85).⁹ Arirachakaran et al. reported a 5.4 times increased risk of requiring revision surgery post-UKA.¹⁰ In a study following 6453 prosthesis, Martino et al. reported 18.2 % of UKAs required revision versus 6.2 % of TKRs. The most common cause of failure post-UKA was aseptic loosening (37.4 %) followed by pain without loosening (19.8 %).¹²

However, this observation of increased revision rates in UKA could be contributed by a selective bias in patient selection. Surgeons tend to offer UKA in younger and more active patients. However, as these patient naturally live longer, this naturally increases the risk of revision as they develop other compartment arthritis in their older age.¹¹ In the analysis of 471 failed medial UKA, Citak et al. found that patients whose implant failure occurred only after more than 10 years were also found to be significantly younger than patients whose implant survival was <5 years (mean age 58.7 vs 61.9, p = 0.0049). The most common causes of failure were other compartment arthritis (39.5 %), aseptic loosening (25.4 %), instability (15.3 %) and polyethylene wear (14 %).¹³

Furthermore, it can be argued that with improved implant design and surgical techniques that UKAs can provide equivalent long term results. Over the decade, improvements in implant design have showed excellent survivorship of up to 98 % at 10 years and 91 % through 20 years.¹¹ Emerson and Higgins studied the long term outcomes of mobile-bearing UKAs which have the potential to optimize polyethylene wear and lead to longer term function. 10 year survivorship was 85 % with lateral compartment arthritis being the most common cause of revision.¹⁴

Moreover, long term survivorship for UKA was found to be comparable to TKR among high volume surgeons, suggesting that UKAs have a steeper learning curve than TKRs. In the analysis of surgical caseload on revision rates, Liddle et al. reported that surgeons performing less than 10 UKAs per year had a mean 8 year survival rate of 87.9 % compared with 92.4 % in surgeons who performed >30 UKAs pe year. After matching patients, the high volume UKA surgeons produced comparable revision rates that seen in TKRs.¹⁵ In a meta-analysis, Hamilton et al. found that surgeons had the lowest revision rates when they performed >24 UKAs per year and had >30 % of their arthroplasty cases being UKAs. It was further reported that percentage of use had a higher impact on survivorship than absolute numbers of UKAs done per year. Surgeons who achieved this attained average 10 year survivorship of ≥94 %.

The benefits of this study include the large sample size and standardisation in the pre and post-operative care within the ERAS pathway. All patients involved in this study were also operated on by arthroplasty surgeons who have high total and unicompartmental arthroplasty loads.

The limitation of our study is its relatively shorter follow up.

5 Conclusion

This study demonstrates that within the ERAS cohort, UKA patients have comparable functional outcomes 6 months post-operatively and better range of motion than patients undergoing TKR with comparable BMI and comorbidity profiles. UKA patients also have less early post-operative infectionand revision rates than TKR patients. Given promising results of comparable long term revision rates in high volume surgeons, UKA should be considered in suitable patients with unicompartmental osteoarthritis to build up one's familiarity with the procedure and the available implants.

Ethics statement

Ethics approval was granted by the local Institutional Review Board (CIRB 2023/2725). This study was performed in accordance with the ethical standards laid down in the 1975 Declaration of Helsinki.

Funding

None.

Patient consent

No guardian or patient consent was required as all patients were above the age of 21.

This was a retrospective study involving data that had been de-identified and collected the institution's Orthopaedic Diagnostic Center Staff who are outside of the study team.

CRediT authorship contribution statement

Louise Woon Theng Lo: Writing manuscript. Sheng Xu: Supervision. Hee-Nee Pang: Supervision. Darren Tay: Supervision. Seng Jin Yeo: Supervision. Ming Han Lincoln Liow: Supervision. Lim Gek Hsiang: Formal analysis. Yongqiang Jerry Chen: Conceptualization.