High Tibial Osteotomy Versus Unicompartmental Knee Arthroplasty for Unicompartmental Knee Osteoarthritis: A Systematic Review and Meta-Analysis

High tibial osteotomy (HTO) and unicompartmental knee arthroplasty (UKA) are commonly performed procedures for the treatment of compartmental knee osteoarthritis; however, the optimal procedure remains controversial. We conducted this systematic review and meta-analysis to compare the functional outcomes and assess complications and revision rates between the two techniques. We searched electronic databases for relevant studies comparing HTO versus UKA for unicompartmental knee osteoarthritis. Continuous data as visual analogue scale (VAS), range of motion, and free walking speed were pooled as mean differences (MDs). Dichotomous data as functional knee outcomes, complications, and revision were pooled as odds ratios (ORs), with 95% confidence interval (CI), using R software for windows. Twenty-five studies involving 8185 patients were included. Meta-analysis showed that HTO was associated with higher risk of complications (OR = 2.47, 95% CI [1.52, 4.04]), poorer functional results (excellent/good) (OR = 0.32, 95% CI [0.21, 0.49]), and greater range of motion (MD = 7.05, 95% CI [2.41, 11.68]) compared to UKA. No significant differences were found between the compared groups in terms of VAS (MD = 0.14, 95% CI [− 0.08, 0.36]), revision rates (OR = 1.30, 95% CI [0.65, 2.60]), and free walking speed (MD = − 0.05, 95% CI [− 0.11, 0.00]). This study showed that UKA achieved fewer complications, better functional outcomes, and less range of motion compared to HTO. No significant differences were detected between HTO and UKA in terms of VAS and revision rate. Treatment options should be personalized to each patient considering factors such as their age, activities of daily living, their body mass index, and severity of osteoarthritis. II.


Introduction
Knee osteoarthritis is a common cause of debilitating knee pain that leads to health and financial burdens [1]. While the aetiology is multifactorial, mechanical axis deviation of the knee joint and altered biomechanics ultimately cause compartmental knee osteoarthritis [2]. Treatments are aimed at unloading the diseased compartment to relieve and prevent further deterioration. The initial management of knee osteoarthritis is non-operative and includes patient education, weight reduction, physical therapy, knee braces, and analgesic therapies [3]. If non-operative treatments fail, high tibial osteotomy (HTO) or unicompartmental knee arthroplasty (UKA) are surgical interventions that may be indicated for unicompartmental knee arthritis.
High tibial osteotomy treats patients with medial compartment knee osteoarthritis due to medial mechanical axis deviation [4]. Several methods have been developed to treat this condition, such as opening wedge (OW), closing wedge (CW), dome, and "en chevron" osteotomies, while the most frequently used methods are medial (opening) and lateral (closing) wedge osteotomies [5]. It is reported that HTO is the first surgical choice for active young patients with preserved joint stability and single compartmental knee osteoarthritis [6]. Although overall HTO provides good functional outcomes [7,8], controversies still exist regarding the most reliable HTO technique (close, open, or dome), clinical implications of patellar height changes, and technical difficulties related to surgical technique/means of synthesis [9].
Growing evidence shows that UKA achieves excellent outcomes and a low risk of complications when used for unicompartmental osteoarthritis [10]. UKA was traditionally used for less active elderly patients with a preserved range of motion, absence of ligament instability, and unicompartmental knee osteoarthritis [11]. Nowadays, these indications have become less defined, especially in patients over 50 years old.
Advances in surgical procedures, along with similar indications for both HTO and UKA, make the decision making challenging. Currently, a lack of agreement in the literature concerning the outcomes of HTO compared with UKA presents as a conundrum to the surgeon when choosing between the two surgical techniques [12]. Therefore, an updated systematic review and meta-analysis of clinical trials and comparative studies were conducted to assess the role of HTO versus UKA for unicompartmental knee osteoarthritis. The current study aimed to compare the functional outcomes and compare complications and revision rates between the two surgical techniques.

Materials and Methods
All steps of this study were conducted in accordance with the Cochrane Handbook of Systematic Reviews of Interventions in addition to Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement guidelines [13]. The protocol of this meta-analysis was published online at the International Prospective Register of Systematic Reviews (PROSPERO) under registration number (CRD42021243076). This study was pre-printed in research square in 2021 to obtain additional reviewers feedback [14].

Literature Search Strategy
We searched Medline (via PubMed), Embase, the Cochrane Library, Clinicaltrial.gov, and Google Scholar, from infinity up to March 2021 for randomized clinical trials (RCTs) and studies comparing HTO to UKA. The search strategy on Medline (via PubMed) included combinations of the keywords: "knee osteoarthritis" OR "knee arthritis" OR "degenerative knee" OR "unicompartmental knee arthroplasty" OR "unicondylar knee arthroplasty" OR "partial knee arthroplasty" OR "high tibial osteotomy" AND "osteotomies". The search strategies included variant spellings and endings. Only English articles were selected, and the search categories included "Orthopaedics" or "Orthopedics" and "Sport Sciences" and "Surgery". Additionally, the reference lists of identified articles were checked manually.

Eligibility Criteria
We included RCTs and studies that compared HTO with UKA for the treatment of patients with unicompartmental knee osteoarthritis. The included outcomes were complications (infection, leg length discrepancy, deep vein thrombosis, and pain), functional outcomes (patient reported), visual analogue scale (VAS), revision rates, range of motion, and free walking speed (velocity). We excluded animal models, reviews, case reports, case series, non-English articles, and duplicate references. Two authors (IS and GB) were involved in the literature search, and any disagreement between the authors was discussed with the third author (NS).

Study Selection
First, title/abstract screening for eligibility for the current study was conducted. Second, full-text screening was conducted. Each step was performed by three independent reviewers and disagreements were resolved upon by consensus.

Data Extraction
Requisite data were extracted by three independent authors into a data extraction form. The extracted data included the following items: (a) baseline characteristics of enrolled patients, (b) general characteristics of study design, (c) included outcomes. In case of missing the mean or standard deviation (SD), we calculated them from the equations provided in the Cochrane Handbook of Systematic Reviews of Interventions [15].

Risk of Bias Assessment and Quality Assessment
We used Cochrane Collaboration's tool for assessing the risk of bias of included RCTs [13]. Risk of bias assessment included: (1) sequence generation, (2) allocation sequence concealment, (3) blinding of participants and personnel, (4) blinding of outcome assessment, (5) incomplete outcome data, (6) selective outcome reporting and (7) other potential sources of bias; the authors' judgement is categorized as 'Low risk', 'High risk' or 'Unclear risk' of bias. To assess the risk of bias across included studies, we compared the reported outcomes between all studies to exclude selective reporting of outcomes. Newcastle Ottawa Scale (NOS) was used to assess the quality of observational studies and each included study was assessed based on reporting of three essential domains: (a) selection of the study subjects, (b) comparability of groups on demographic characteristics and important potential confounders, and (c) ascertainment of the included outcome (exposure/treatment) [16].

Data Synthesis
Continuous data were pooled as mean difference (MD). Dichotomous data were pooled as odds ratio (OR), with a 95% confidence interval (CI). We used R software (metapackage 4.9-0) for Windows during data synthesis. Heterogeneity was assessed by visual inspection of the forest plots and measured by Q statistic and I 2 statistic. Significant statistical heterogeneity was indicated by Q statistic P value less than 0.1 or by I 2 more than 50%. In case of significant heterogeneity, we conducted a random effect model. Otherwise, the fixed effect model was used. Sensitivity analysis was used to resolve heterogeneity. The funnel plot method assessed publication bias for outcomes with more than ten included studies.

Search Strategy Results
Our literature search yielded 2392 unique records. After title/abstract screening, 45 were retrieved and screened for eligibility. Finally, 25 studies were included in the metaanalysis. The flow of the study selection process is shown in the PRISMA flow diagram, Fig. 1.

Characteristic of Included Studies
Twenty-five studies involving 8185 patients were included. Of them, 11 studies were RCTs and 14 were observational comparative studies. A total of 1996 patients had HTO, and 6189 patients had UKA. The follow-up period ranged from 12 to 793.6 months. The age ranged from 30 to 84 years. All articles were published in English from 1982 to 2020. A summary of the design and baseline characteristics of enrolled patients is presented in Table 1. The risk of bias and quality assessment of included studies with the references were reported in Appendix Tables 1 and 2.

Functional Outcomes (Excellent/Good Results)
Ten studies reported on the number of patients having excellent/good results regarding the functional outcomes. The effect estimate showed that HTO was associated with poorer functional results compared to UKA (OR = 0.32, 95% CI [0.21, 0.49]). Moderate evidence of heterogeneity was noted (I 2 = 45%, P = 0.007), Fig. 3.

Survival and Revision
Survival was defined as the time to a UKA revision, second operation, or an operation failure. Fourteen studies reported on revision. The total effect size was comparable between HTO and UKA (OR = 1.30, 95% CI [0.65, 2.60]). Significant heterogeneity was reported (I 2 = 67%, P < 0.01), Fig. 4.

Range of Motion
Seven studies reported on a range of motion. The total effect size showed that HTO was associated with a larger range of motion than UKA (MD = 7.05, 95% CI [2.41, 11.68]). Significant heterogeneity was observed (I 2 = 93%, P < 0.01), Appendix Fig. 2.

Publication Bias
The funnel plot showed no evidence of publication bias regarding post-operative complications and revision, Appendix Figs. 4 and 5.

Discussion
Our study compared the outcomes of HTO with UKA for the treatment of unicompartmental knee arthritis. This meta-analysis of 25 studies showed that when compared to HTO, UKA achieved fewer complications and better functional outcomes (excellent/good), but reduced knee range of motion. No significant differences were detected between HTO and UKA in visual analogue scale, revision rate, and free walking speed.
The conventional indications for UKA have been revised over time as the procedure becomes more popular amongst surgeons and its complication rates have remain low when trialled in young and obese patients [17,18]. Conventional  indications previously encouraged patients > 60 years weighing < 82 kg, with single unicompartment osteoarthritis and < 15° angular deformity, but > 90° range of motion for UKA [11,19]. The following findings support the indications for UKA to be expanded beyond this scope. A retrospective study of 41 patients showed UKA achieved a 92% survival rate at 11 years in patients ranging 35-60 years old, proving its efficacy in younger ages [20]. Additionally, 62 patients with 11.2 years follow-up following UKA were associated with good functional outcomes and a 94% survival rate at 12 years in patients aged 60 years or younger [21]. While these studies indicate UKA is a good intervention in isolation, a more recent follow-up study comparing UKA with HTO found little differences between the two after a year in patients with a mean age of 50 years [22]. Our study showed favourable outcomes and fewer complication rates for UKA compared with HTO, apart from a reduced range of movement. Taken together, these outcomes indicate UKA shows good/excellent short-and long-term outcomes for a wide patient age range. Both HTO and UKA share similar indications, but HTO is especially selected for young and active patients with a larger range of movement and localized tibiofemoral joint line pain [9,23]. The selection of suitable patients, accurate osteotomy types, and specific surgical techniques are important in ensuring the success of HTO, and the literature reflects this when HTO has had suboptimal outcomes for older and overweight individuals [4,24,25]. This is in contrast with UKA, which seems to benefit wider age ranges. For example, Trieb et al. observed a higher failure rate in HTO patients whose age was ≥ 65 years when compared with younger patients [24] and a prospective study of 132 patients demonstrated a pre-operative body mass index (BMI) of 27.5 kg/m 2 was a risk factor for early failure of HTO [4]. When a Markov model simulated patients aged 40-70 years undertaking primary HTO or UKA, there were higher revision risks at 5 and 10 years follow-up in the HTO group compared with UKA for patients ≥ 50 years [26]. A large retrospective review in the USA showed that HTO was performed more commonly in patients aged 40-44 years and UKA was more common in patients aged 60-64 years, showing surgeon selection biased younger patients to HTO and older to UKA [27]. While HTO is regarded safer and presents more favourable outcomes in younger patients, it should be considered alongside UKA given the recent success of UKA in young individuals. The present metaanalysis included ages from 30 to 84 years, which may have impacted the positive outcomes that should be seen with HTO in only young patients. This may explain why the revision rate did not differ significantly between the HTO and UKA groups, and why other functional outcomes were not improved compared with UKA.
Several techniques of HTO have been developed including opening wedge, closing wedge, dome, and "en chevron" osteotomies. Opening wedge and closing wedge are the most frequently used techniques and no differences in most of the clinical outcomes were found except the operation time [28]. In the current meta-analysis HTO was prescribed more frequently but encountered higher post-operative complications including infection, venous thrombosis, cortical fracture, and peroneal nerve injury. These findings aligns with previous meta-analyses analysing similar outcomes [29,30]. Previously, a systematic review of 12 HTO studies showed that HTO had a complication rate up to 47% post-operatively [31]. The significantly higher proportion of complications after HTO may be due to surgical techniques, long-standing cast immobilization, late limb load, and inadequate fixation following HTO operation. Within the present and previous meta-analysis as well as systematic reviews, search criteria without age limitations for HTO would have allowed for older age groups undergoing HTO. As a patient with HTO is largely bed bound or largely immobile post-operation, this would naturally increase complications in older age groups compared with UKA which is more mobile immediately post-operation. This would bias higher complications reported for HTO when in a real-world setting, this surgery is not prescribed for sedentary, senile individuals.
The current meta-analysis showed that excellent/good functional outcomes were significantly higher in the UKA group and range of motion (ROM) was higher in the HTO group, but there were no significant differences in visual pain scale. The differences in functional outcomes and range of motion results indicate that other factors might impact functional results. Osteotomy targets transferring the mechanical axis from the abnormal position to the normal area [32], to improve pain and gait to allow better ROM. Although HTO did not significantly differ in VAS, the subjectivity of pain may interfere with the actual symptomatic relief from HTO, which may explain the ROM differences [25]. Additionally, patients undergoing HTO in the past were placed in a plaster cast from the groin to the ankle for 6 weeks, but now patients are placed in braces and crutches for 6 weeks post-operation [30]. The change from plaster to brace may have allowed for greater functional outcomes relative to previous post-operative rehabilitative measures which can explain better ROM in HTO. The TomoFix plate is a modern innovation for HTO to increase mobility, but even so Koh et al. reported patient satisfaction was still higher in the UKA group than the HTO group in active patients [33]. Kim et al. used TomoFix plate in OWHTO vs UKA and reported that HTO and UKA had similar pain and functional outcomes at 12 and 24 months post-operatively [34]. These results accord with the findings of Takeuchi et al. [35], Jeon et al. [36], and Ryu et al. [37].
UKA has shown consistently better outcomes compared with HTO, including lower perioperative blood loss, quicker recovery, and no immobilization. A report by Jeon et al. showed a better post-operative activity level following UKA compared to HTO at 6 months, whereas at 12 months and 2-year follow-up no significant differences were reported in the compared groups [36]. A study by Borjesson et al. reported an increased free walking speed from 1.07 to 1.16 m/s in the UKA group compared to the HTO group which was associated with decreased free walking speed from 1.07 to 0.94 m/s 3 months post-operatively. However, comparable results were observed between the two groups at the follow-up durations of 1 and 5 years [38]. HTO also reduced mean time to return to sporting activities, especially ones that were high impact which led to better sports related functional scores compared with UKA [39]. A more recent comparison showed Western Ontario and McMaster Universities Arthritis Index (WOMAC) scores were reduced in UKA, and survival rate was higher (96.2%) compared with HTO (87.7%) [40]. Based on the post-operative mobilization in UKA versus HTO previously discussed, UKA offers improved post-operative functional outcomes than HTO in short-term follow-up but this meta-analysis found no difference in long-term results.

Strengths and Limitations
The current meta-analysis included some new studies up to 2021 [22,33,34,[39][40][41]; thus, our results are more up to date, have better power for analysis and provide a timely review of the literature compared to previous meta-analyses [42,43]. We determined search methods and performed a comprehensive search using many electronic databases and we followed the PRISMA checklist when reporting this manuscript. Additionally, only relevant RCTs and comparative studies were included in this meta-analysis providing high level of evidence of clinical practice and recommendations. Some limitations should also be acknowledged. First, the heterogeneity in study designs, matching criteria, sample size, operative techniques, and measurement of outcomes across the included studies may limit generalizability of the results. We used the random effects model to decrease the impact of heterogeneity however, it does not eliminate it. Secondly, age ranges for HTO and UKA were not limited to the recommended age and were not secondarily adjusted for age, which could bias worse outcomes for HTO.

Conclusion
This meta-analysis of 25 studies showed that UKA achieved lower complication rates and better functional outcomes (excellent/good), while HTO allowed for superior range of motion. No significant differences were detected between HTO and UKA in terms of visual analogue scale, revision rate, and free walking speed. Treatment options should be personalized to each patient considering factors such as their age, activities of daily living, their BMI, and severity of osteoarthritis. Due to significant heterogeneity in survival and revision and range of motion between UKA and HTO, large-scale RCTs are needed to validate these results.