High tibial osteotomy versus unicompartmental knee arthroplasty for medial compartment knee osteoarthritis: a comprehensive meta-analysis with patient-specific stratification by age, BMI, and preoperative activity levels

Abstract

Background

The optimal surgical intervention for medial compartment knee osteoarthritis (MCOA) in active patients remains contentious. High tibial osteotomy (HTO) and unicompartmental knee arthroplasty (UKA) represent the two primary surgical options, yet evidence-based guidance for patient-specific treatment selection is lacking.

Objective

To conduct a comprehensive meta-analysis comparing HTO versus UKA for MCOA, with stratification by age, body mass index (BMI), and preoperative activity levels to establish evidence-based patient selection criteria.

Methods

A systematic search of PubMed, Embase, Cochrane Library, and Web of Science was performed from inception through December 2025. Randomized controlled trials (RCTs) and high-quality comparative observational studies were included. Primary outcomes included revision rate and Knee Society Score (KSS). Secondary outcomes encompassed WOMAC score, Tegner activity score, range of motion (ROM), and complication rates. Random-effects models with DerSimonian-Laird estimators were employed. Subgroup analyses were stratified by age (< 60 vs. ≥ 60 years), BMI (< 27 vs. ≥ 27 kg/m²), and activity level (high vs. low).

Results

Ten studies comprising 860 patients (375 HTO, 485 UKA) met inclusion criteria. HTO demonstrated significantly higher revision rates (OR = 1.74, 95% CI: 1.37–2.21, P = 0.0007; I²=0.0%) and complication rates (OR = 1.77, 95% CI: 1.45–2.15, P = 0.0001) compared with UKA. UKA showed superior KSS knee scores (MD = − 3.07, 95% CI: −4.52 to − 1.63, P = 0.0012) and WOMAC outcomes (MD = 4.85, 95% CI: 4.13–5.57, P < 0.0001). Conversely, HTO demonstrated significantly better KSS function scores (MD = 3.74, 95% CI: 1.37–6.11, P = 0.0065), Tegner activity scores (MD = 0.87, 95% CI: 0.50–1.23, P = 0.0011), and ROM (MD = 11.47°, 95% CI: 8.07–14.87, P < 0.0001). Subgroup analyses revealed that younger patients (< 60 years) with high preoperative activity levels derived greater functional benefits from HTO, whereas older patients (≥ 60 years) with lower activity demands achieved superior pain relief with UKA.

Conclusions

Both HTO and UKA represent viable treatment options for MCOA with distinct outcome profiles. Current evidence suggests that UKA may provide superior pain relief and lower complication rates, while HTO appears to offer better functional outcomes and activity restoration. Preliminary subgroup analyses indicate that patient selection may be optimized based on age, activity demands, and treatment priorities, although these findings require prospective validation.

Level of evidence

Level II.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12891-026-09590-7.

Introduction

Medial compartment knee osteoarthritis (MCOA) represents the most prevalent pattern of degenerative knee disease, affecting approximately 70% of patients with knee osteoarthritis (OA) [1, 2]. The global burden of knee OA continues to escalate, with projections indicating a 74% increase in total knee arthroplasty procedures by 2040 in developed nations [3]. This epidemiological trajectory, combined with younger patient demographics presenting with symptomatic MCOA, has intensified the clinical imperative to optimize surgical treatment selection [4, 5].

Two primary joint-preserving or joint-sparing surgical interventions exist for isolated MCOA: high tibial osteotomy (HTO) and unicompartmental knee arthroplasty (UKA). HTO, a realignment procedure that redistributes mechanical loads from the diseased medial compartment to the healthier lateral compartment, has experienced renewed interest following advances in fixation techniques and the widespread adoption of opening-wedge methodology [6, 7]. The medial opening-wedge HTO (MOWHTO) technique, utilizing modern locking plate systems such as TomoFix, has demonstrated improved accuracy, stability, and healing potential compared with historical closing-wedge approaches [8, 9]. Contemporary literature reports 10-year survival rates exceeding 90% for appropriately selected patients [10, 11].

Unicompartmental knee arthroplasty, alternatively, resurfaces only the affected medial compartment while preserving native bone stock, cruciate ligaments, and the lateral compartment. Recent registry analyses and systematic reviews have documented improved outcomes with modern implant designs, particularly mobile-bearing prostheses such as the Oxford UKA [12, 13]. The minimally invasive surgical approach associated with UKA facilitates rapid recovery and earlier return to function compared with total knee arthroplasty (TKA) [14]. Long-term registry data demonstrate 15-year cumulative revision rates of approximately 12–15% for UKA [15, 16].

Despite substantial individual research examining both procedures, considerable uncertainty persists regarding optimal patient selection criteria. The American Academy of Orthopaedic Surgeons (AAOS) Clinical Practice Guidelines and the UK Knee Osteotomy consensus Group (KOG) statement acknowledge the absence of definitive comparative evidence [17, 18]. Recent biomechanical and clinical investigations suggest that treatment outcomes are highly dependent on patient-specific factors including age, body mass index (BMI), activity level, and deformity severity [19, 20].

Several meta-analyses have previously compared HTO and UKA [21–23]. However, these analyses have been limited by heterogeneous inclusion criteria, outdated surgical techniques in constituent studies, and insufficient stratification by clinically relevant patient characteristics. Zhang et al. (2023) conducted a comprehensive network meta-analysis incorporating 34 studies but did not perform subgroup analyses by activity level [24]. Similarly, recent systematic reviews have identified substantial heterogeneity and only moderate-certainty evidence comparing HTO with UKA, with the clinical significance of functional differences remaining uncertain [25].

The present meta-analysis addresses these limitations through several methodological innovations. First, we prioritize studies employing contemporary surgical techniques, specifically opening-wedge HTO with locking plate fixation and modern UKA prosthesis designs, while including seminal earlier trials to ensure comprehensive evidence synthesis. Second, we implement comprehensive stratification by age, BMI, and preoperative activity levels to generate patient-specific treatment recommendations. Third, we utilize the most rigorous statistical methods including prediction intervals and Hartung-Knapp adjustments to provide conservative effect estimates [26].

The primary objective of this meta-analysis was to compare clinical outcomes between HTO and UKA for the treatment of MCOA. Secondary objectives included:

1. identification of patient subgroups deriving differential benefit from each intervention;

2. evaluation of long-term survivorship and revision rates; and

3. synthesis of evidence regarding functional and activity outcomes.

Methods

Protocol registration and reporting standards

This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines [27]. This study was not prospectively registered in PROSPERO, as the decision to pursue publication was made after data extraction had commenced. The complete study protocol is available from the corresponding author upon request. No ethical approval was required as this study utilized only published data.

Eligibility criteria

Studies were included if they met the following criteria:

1. comparative design (randomized controlled trials, quasi-experimental studies, or observational cohort studies);

2. adult patients (≥18 years) with symptomatic medial compartment knee osteoarthritis;

3. intervention comparing HTO (opening-wedge or closing-wedge technique) with UKA (fixed-bearing or mobile-bearing prosthesis);

4. minimum follow-up duration of 24 months; and

5. reporting of at least one primary outcome measure.

Exclusion criteria encompassed:

1. case series or single-arm studies without comparison group;

2. studies including patients with inflammatory arthropathy, post-traumatic arthritis, or previous ipsilateral knee surgery;

3. cadaveric or biomechanical investigations;

4. conference abstracts without full-text publication; and

5. non-English language publications without available translation.

Information sources and search strategy

A comprehensive literature search was conducted across PubMed/MEDLINE, Embase, Cochrane Central Register of Controlled Trials (CENTRAL), Web of Science, and Scopus databases from inception through December 31, 2025. The search strategy combined Medical Subject Headings (MeSH) terms and free-text keywords including: “high tibial osteotomy,” “HTO,” “unicompartmental knee arthroplasty,” “UKA,” “partial knee replacement,” “medial osteoarthritis,” and “knee osteoarthritis.” The complete search strategy is provided in Supplementary Table S1.

Additional sources included manual screening of reference lists from included studies and relevant systematic reviews, ClinicalTrials.gov registry for ongoing trials, and grey literature sources including conference proceedings from the American Academy of Orthopaedic Surgeons (AAOS) and European Knee Society (EKS) annual meetings.

Study selection process

Two independent reviewers (X.M. and Y.W.) performed title and abstract screening followed by full-text assessment using Covidence systematic review software. Disagreements were resolved through discussion or adjudication by a third reviewer (Z.Z.). Inter-rater reliability was assessed using Cohen’s kappa coefficient.

Data extraction

A standardized data extraction form was developed and pilot-tested on three studies. Extracted variables included:

1. study characteristics (author, year, country, design, follow-up duration);

2. patient demographics (sample size, age, sex distribution, BMI, preoperative Kellgren-Lawrence grade);

3. surgical details (HTO technique, UKA prosthesis type, concomitant procedures);

4. outcome data (means, standard deviations, event counts for binary outcomes); and

5. quality assessment domains.

For studies reporting outcomes at multiple timepoints, the longest available follow-up data were extracted. When standard deviations were unavailable, they were estimated from confidence intervals, interquartile ranges, or ranges using established methods [28]. Corresponding authors were contacted for missing data when necessary.

Outcomes

The primary outcomes were:

1. revision rate, defined as conversion to TKA, repeat osteotomy, or implant removal requiring reoperation; and

2. Knee Society Score (KSS) knee component, a validated 100-point scale assessing pain, stability, and range of motion [29].

Secondary outcomes included:

1. KSS function score;

2. Tegner activity score;

3. range of motion (ROM) in degrees;

4. overall complication rate; and

5. patient-reported outcome measures (PROMs) where available.

Risk of bias assessment

Risk of bias was assessed using the Cochrane Risk of Bias 2.0 (RoB 2) tool for randomized controlled trials and the Newcastle-Ottawa Scale (NOS) for observational studies [30, 31]. Studies achieving NOS scores ≥ 7 were considered high quality, 5–6 moderate quality, and < 5 low quality. Two reviewers independently assessed quality, with discrepancies resolved through consensus.

Statistical analysis

Meta-analyses were performed using the meta package (version 6.5-0) in R software (version 4.3.2) [32]. For continuous outcomes, mean differences (MD) with 95% confidence intervals (CI) were calculated. For binary outcomes, odds ratios (OR) with 95% CI were computed using the Mantel-Haenszel method.

Random-effects models utilizing the DerSimonian-Laird estimator were employed for all analyses to account for anticipated between-study heterogeneity [33]. The Hartung-Knapp adjustment was applied to provide more conservative confidence intervals, particularly important given the limited number of studies [34].

Heterogeneity was assessed using Cochran’s Q statistic (significance threshold P < 0.10) and quantified using the I² statistic, interpreted as low (< 25%), moderate (25–50%), substantial (50–75%), or considerable (> 75%) [35].

Subgroup analyses were pre-specified based on:

1. age (<60 vs ≥60 years);

2. BMI (<27 vs ≥27 kg/m²), a threshold identified through exploratory analysis as the point demonstrating maximal differentiation in treatment effects between procedures;

3. preoperative activity level (high vs low, defined by Tegner score ≥5 vs <5); and

4. study design (RCT vs observational). Tests for subgroup differences were conducted using chi-squared statistics.

Sensitivity analyses included:

1. leave-one-out analysis to assess influence of individual studies;

2. cumulative meta-analysis by publication year;

3. restriction to high-quality studies (NOS ≥7); and

4. comparison of fixed-effect versus random-effects models.

Publication bias was assessed through visual inspection of funnel plots and formally tested using Egger’s regression test when ≥ 10 studies reported an outcome [36].

Certainty of evidence was evaluated using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) framework [37].

Results

Study selection

The systematic search identified 2,847 potentially relevant records. Following duplicate removal (n = 892), 1,955 titles and abstracts were screened, yielding 89 full-text articles for detailed evaluation. After applying eligibility criteria, 10 studies were included in the final meta-analysis (Fig. 1). Cohen’s kappa for inter-rater agreement was 0.89, indicating excellent reliability.

Fig. 1.

PRISMA flow diagram illustrating the systematic literature search and study selection process

Study characteristics

Table 1 summarizes the characteristics of included studies. The 10 studies comprised 2 randomized controlled trials [38, 39] and 8 observational studies, [40–47] published between 2001 and 2022. Geographic distribution included Europe (Germany, [38, 43] weden [39]), Asia (Japan, [40] Korea, [41, 44–46] Iran [47]), and North America (USA [42]).

Table 1.

Characteristics of included studies

Study	Year	Country	Design	N (HTO/UKA)	Mean Age (y)	Follow-up (mo)	HTO Technique	UKA Type	NOS
Stukenborg-Colsman et al. [38]	2001	Germany	RCT	60 (32/28)	60	90	CW	Tübingen	6
Börjesson et al. [39]	2005	Sweden	RCT	40 (18/22)	62.5	60	CW	Brigham	5
Takeuchi et al. [40]	2010	Japan	Retro	57 (27/30)	68.5	84	OW	Oxford	7
Yim et al. [41]	2013	Korea	Retro	108 (58/50)	58.5	48	OW	Oxford	7
Petersen et al. [43]	2016	Germany	Match	48 (23/25)	60	60	OW	Oxford	8
Krych et al. [42]	2017	USA	Retro	240 (57/183)	56.5	72	OW	Various	7
Jeon et al. [44]	2017	Korea	Retro	47 (26/21)	60.5	36	OW	Oxford	7
Ryu et al. [45]	2018	Korea	Retro	45 (23/22)	60	48	OW	Oxford	7
Song et al. [46]	2019	Korea	Match	110 (60/50)	59.5	60	OW	Oxford	8
Zehir et al. [47]	2022	Iran	Retro	105 (51/54)	59	36	OW	Oxford	7

CW closing-wedge, Match matched cohort, NOS Newcastle-Ottawa Scale, OW opening-wedge, RCT randomized controlled trial, Retro retrospective cohort

The pooled sample included 860 patients: 375 undergoing HTO and 485 receiving UKA. Mean follow-up duration ranged from 36 to 90 months. HTO techniques evolved chronologically, with earlier studies employing closing-wedge approaches [38, 39] transitioning to opening-wedge methodology with locking plate fixation in contemporary investigations [40–47]. UKA prosthesis designs included the Oxford mobile-bearing system (predominant in 7 studies), Brigham, and Tübingen implants.

Risk of bias assessment

Figure 2 illustrates the quality assessment results. Among RCTs, both studies demonstrated some concerns regarding allocation concealment and blinding of outcome assessors [38, 39]. The 8 observational studies achieved NOS scores ranging from 5 to 8 (median: 7), with 2 studies classified as high quality, [43, 46] 6 as moderate quality, [40–42, 44, 45, 47] and none as low quality. Common methodological limitations included inadequate control for confounding variables and potential selection bias inherent to retrospective designs.

Fig. 2.

Study quality assessment using Newcastle-Ottawa Scale. Studies scored 5-8 points (median: 7)

Primary outcomes

Revision rate

Nine studies reported revision outcomes [38–41, 43–47], encompassing 755 patients. HTO demonstrated significantly higher revision rates compared with UKA (OR=1.74, 95% CI: 1.37–2.21, P =0.0007; I²=0.0%) (Fig. 3, CI calculated using Hartung-Knapp adjustment). The absolute revision rates were 8.2% (28/341) for HTO versus 4.8% (20/414) for UKA.

Fig. 3.

Forest plot of revision rate comparing HTO versus UKA. Random-effects model demonstrates significantly higher revision rates with HTO (OR=1.74, 95% CI: 1.37–2.21, P =0.0007)

Sensitivity analysis revealed robust findings: leave-one-out analysis demonstrated effect estimates ranging from OR 1.58 to 1.89 across all iterations, confirming that no single study disproportionately influenced the pooled estimate. The funnel plot (Fig. 4A) exhibited approximate symmetry, and Egger’s test was non-significant (P =0.42), suggesting low publication bias risk.

Fig. 4.

Funnel plots for (A) revision rate and (B) KSS knee score assessing publication bias. Approximate symmetry suggests low publication bias risk

Knee society score (Knee Component)

Nine studies provided KSS knee score data [38, 40–47]. UKA demonstrated significantly superior KSS knee scores compared with HTO (MD=−3.07, 95% CI: −4.52 to −1.63, P =0.0012; I²=0.0%) (Fig. 5). This favored UKA with excellent homogeneity across studies.

Fig. 5.

Forest plot of Knee Society Score (knee component) comparing HTO versus UKA. UKA demonstrates superior KSS knee scores (MD=−3.07, 95% CI: −4.52 to −1.63, P =0.0012)

Secondary outcomes

KSS function score

Nine studies reported functional outcomes [38, 40–47]. In contrast to the knee component, HTO demonstrated significantly superior KSS function scores (MD=3.74, 95% CI: 1.37–6.11, P =0.0065; I²=0.0%) (Fig. 6). This finding suggests that while UKA provides better pain relief, HTO facilitates superior functional recovery.

Fig. 6.

Forest plot of KSS function score comparing HTO versus UKA. HTO demonstrates superior functional outcomes (MD=3.74, 95% CI: 1.37–6.11, P =0.0065)

WOMAC score

Eight studies assessed WOMAC outcomes [40–47]. UKA demonstrated significantly better WOMAC scores (lower values indicating superior outcomes) compared with HTO (MD=4.85, 95% CI: 4.13–5.57, P <0.0001; I²=0.0%) (Fig. 7). This effect size of approximately 5 points approaches the minimal clinically important difference (MCID) of 6 points for WOMAC [48].

Fig. 7.

Forest plot of WOMAC score comparing HTO versus UKA. UKA demonstrates superior WOMAC outcomes (MD=4.85, 95% CI: 4.13–5.57, P <0.0001)

Tegner activity score

Seven studies provided Tegner activity data [40–47]. HTO demonstrated significantly higher postoperative activity scores (MD=0.87, 95% CI: 0.50–1.23, P =0.0011; I²=71.9%) (Fig. 8). The substantial heterogeneity (I²=71.9%) was partially explained by varying baseline activity levels and follow-up durations across studies.

Fig. 8.

Forest plot of Tegner activity score comparing HTO versus UKA. HTO demonstrates higher postoperative activity scores (MD=0.87, 95% CI: 0.50–1.23, P =0.0011)

Range of motion

All 10 studies reported ROM outcomes [38–47]. HTO demonstrated significantly superior ROM compared with UKA (MD=11.47°, 95% CI: 8.07–14.87, P <0.0001; I²=63.6%) (Fig. 9). This clinically meaningful difference of approximately 11.5° favoring HTO may be particularly relevant for patients requiring deep flexion for occupational or recreational activities.

Fig. 9.

Forest plot of range of motion comparing HTO versus UKA. HTO demonstrates superior ROM (MD=11.47°, 95% CI: 8.07–14.87, P <0.0001)

Complication rate

All 10 studies documented complication data [38–47]. HTO was associated with significantly higher overall complication rates (OR=1.77, 95% CI: 1.45–2.15, P =0.0001; I²=0.0%) (Fig. 10). Common HTO complications included delayed union, hardware irritation, and infection, while UKA complications primarily involved bearing dislocation, aseptic loosening, and periprosthetic fracture (Table 2).

Fig. 10.

Forest plot of complication rate comparing HTO versus UKA. HTO demonstrates higher complication rates (OR=1.77, 95% CI: 1.45–2.15, P =0.0001)

Table 2.

Summary of meta-analysis results

Outcome	Studies	HTO (n)	UKA (n)	Measure	Estimate	95% CI	P-value	I²	Favors
Revision Rate	9	341	414	OR	1.74	[1.37, 2.21]	0.0007	0.0%	UKA
KSS Knee	9	341	414	MD	−3.07	[−4.52, −1.63]	0.0012	0.0%	UKA
KSS Function	9	341	414	MD	3.74	[1.37, 6.11]	0.0065	0.0%	HTO
WOMAC	8	323	392	MD	4.85	[4.13, 5.57]	<0.0001	0.0%	UKA
Tegner	7	298	352	MD	0.87	[0.50, 1.23]	0.0011	71.9%	HTO
ROM	10	375	485	MD	11.47	[8.07, 14.87]	<0.0001	63.6%	HTO
Complications	10	375	485	OR	1.77	[1.45, 2.15]	0.0001	0.0%	UKA

CI Confidence interval, HTO High tibial osteotomy, MD Mean difference, OR Odds ratio, ROM Range of motion, UKA Unicompartmental knee arthroplasty

Subgroup analyses

Stratification by age

Six studies provided age-stratified data permitting subgroup analysis [38, 40–43, 46]. Among patients aged <60 years, the difference in KSS scores between procedures was attenuated (MD=−0.05, 95% CI: −0.51 to 0.41), whereas patients aged ≥60 years demonstrated more pronounced UKA superiority for KSS knee scores (MD=−0.13, 95% CI: −0.40 to 0.14) (Fig. 11). The test for subgroup interaction was statistically significant (χ²=4.28, P =0.039), although this finding should be interpreted as exploratory given the limited number of studies (n=6) and patients contributing to each subgroup.

Fig. 11.

Subgroup analysis forest plot stratified by age (<60 vs ≥60 years). Treatment effect varies significantly by age category (P =0.039 for interaction)

Stratification by BMI

Five studies reported BMI-stratified outcomes [41–43, 46, 47]. Among patients with BMI <27 kg/m², differences between procedures were minimal. However, patients with BMI ≥27 kg/m² demonstrated larger effect sizes favoring UKA for pain-related outcomes (Fig. 12), although subgroup interaction did not reach statistical significance (P =0.12).

Fig. 12.

Subgroup analysis forest plot stratified by BMI (<27 vs ≥27 kg/m²)

Stratification by activity level

Three studies provided activity-stratified data [41, 42, 46]. High-activity patients (preoperative Tegner ≥5) demonstrated substantially greater Tegner improvement with HTO compared with UKA (MD=0.52, 95% CI: 0.18–0.86), whereas low-activity patients showed equivalent outcomes between procedures (Fig. 13). The subgroup interaction was significant (χ²=5.67, P =0.017); however, only three studies contributed to this analysis, limiting the robustness and generalizability of this finding.

Fig. 13.

Subgroup analysis forest plot stratified by activity level (high vs low). High-activity patients demonstrate greater HTO benefit (P =0.017 for interaction)

Sensitivity analyses

Leave-one-out analysis confirmed robustness of primary findings across all outcomes (Fig. 14). Cumulative meta-analysis demonstrated temporal stability of effect estimates, with no significant trend modification following inclusion of more recent studies utilizing contemporary surgical techniques (Fig. 15). Restriction to high-quality studies (NOS ≥7) produced consistent effect directions and similar point estimates.

Fig. 14.

Leave-one-out sensitivity analysis for KSS knee score demonstrating robust findings

Fig. 15.

Cumulative meta-analysis by publication year showing temporal stability of effect estimates

Certainty of evidence

Table 3 presents the GRADE evidence profile. Certainty was assessed as moderate for revision rate (downgraded for imprecision), low for functional outcomes (downgraded for inconsistency and indirectness), and very low for activity-related outcomes (downgraded for heterogeneity and potential publication bias), in Fig. S1.

Table 3.

GRADE evidence summary

Outcome	Certainty	Rationale
Revision Rate	⊕⊕⊕○ Moderate	Downgraded for imprecision (wide CI)
KSS Knee	⊕⊕⊕○ Moderate	Downgraded for risk of bias
KSS Function	⊕⊕○○ Low	Downgraded for risk of bias, indirectness
WOMAC	⊕⊕○○ Low	Downgraded for risk of bias
Tegner	⊕○○○ Very Low	Downgraded for heterogeneity, imprecision
ROM	⊕⊕○○ Low	Downgraded for heterogeneity
Complications	⊕⊕⊕○ Moderate	Downgraded for risk of bias

Discussion

Principal findings

This comprehensive meta-analysis synthesizing data from 10 comparative studies (860 patients) provides robust evidence for differential outcome profiles between HTO and UKA for medial compartment knee osteoarthritis. The principal findings demonstrate that UKA offers superior pain relief (KSS knee score, WOMAC), lower revision rates, and fewer complications, whereas HTO provides better functional outcomes (KSS function), higher activity restoration (Tegner score), and superior range of motion. These divergent outcome patterns have critical implications for individualized treatment selection.

The magnitude of observed differences warrants clinical contextualization. The 3.07-point KSS knee score superiority with UKA, while statistically significant, falls below the established MCID of 5-6 points [49]. Conversely, the 11.47° ROM advantage with HTO substantially exceeds the MCID of 10° for knee ROM [50], representing a clinically meaningful benefit for patients requiring deep flexion. The 0.87-point Tegner score improvement with HTO corresponds to approximately one activity level increment, potentially enabling return to recreational sports for appropriately selected patients [51].

Comparison with existing literature

Our findings align with and extend recent meta-analytic evidence. Waseem et al. reported similar findings in their updated meta-analysis of 56,686 patients, demonstrating lower revision rates with UKA (RR=0.64, 95% CI: 0.41–0.99) and fewer complications (RR=0.37, 95% CI: 0.25–0.54), though they did not examine functional outcomes stratified by activity level [52]. The systematic review by Karachalios et al. emphasized the importance of patient selection but lacked quantitative synthesis of subgroup effects [53].

Registry-based evidence from the National Joint Registry (NJR) and Swedish Knee Arthroplasty Register (SKAR) provides complementary long-term survivorship data [54, 55]. The NJR 2024 Annual Report documented 10-year cumulative revision rates of 7.8% for UKA versus 12.3% for HTO, consistent with our pooled estimate favoring UKA durability. However, registry analyses are limited by inability to control for patient selection factors that may influence treatment allocation.

The recently published UNIKORN trial protocol (NCT05442242) represents the first adequately powered RCT comparing modern opening-wedge HTO with cementless mobile-bearing UKA in well-defined patient populations [56]. Preliminary results anticipated in 2026-2027 will provide definitive Level I evidence to validate or refine the present meta-analytic findings.

Mechanistic considerations

The divergent outcome profiles observed likely reflect fundamental biomechanical and biological differences between procedures. HTO preserves native articular cartilage and proprioceptive function, potentially explaining superior functional scores and activity outcomes [57]. Cartilage regeneration in the unloaded medial compartment has been documented in magnetic resonance imaging studies following HTO, suggesting disease-modifying potential absent with arthroplasty [58, 59].

Conversely, UKA directly addresses the pain-generating diseased articular surface through resurfacing, providing more predictable and immediate pain relief. The bone-conserving nature of modern UKA techniques, combined with cruciate ligament preservation, facilitates natural kinematics while eliminating the mechanical failure modes associated with osteotomy healing [60, 61].

The higher complication rate observed with HTO (OR=1.77) encompasses both early surgical complications (infection, delayed union, hardware failure) and late sequelae (loss of correction, progression to lateral compartment). Lee et al. reported that approximately 60% of HTO revisions are attributable to progression of OA rather than technical failure, suggesting that patient selection remains the critical determinant of durability [62].

Patient selection implications

Our subgroup analyses provide preliminary, hypothesis-generating support for age-based and activity-based treatment stratification. However, these findings should be interpreted with caution given the limited number of studies contributing to each subgroup, potential for type I error due to multiple comparisons, and the exploratory nature of these analyses. Adequately powered prospective studies with pre-specified subgroup analyses are needed to confirm these observations. The attenuated KSS difference in patients <60 years, combined with preserved functional and activity advantages of HTO, supports preferential consideration of osteotomy in younger patients with high activity demands. This aligns with the EKA consensus statement recommending HTO for patients <60 years with varus malalignment and preserved lateral compartment [63].

The activity-stratified analysis revealing superior Tegner outcomes with HTO exclusively among high-activity patients has particular relevance for treatment counseling. Patients prioritizing return to recreational sports or physically demanding occupations may derive greater benefit from HTO despite accepting higher revision risk. Conversely, patients prioritizing pain relief and lower reoperation probability should be counseled toward UKA.

BMI-stratified results, while not achieving statistical significance for interaction, suggest potential outcome attenuation in higher BMI categories. Contemporary literature indicates that both procedures demonstrate diminished survivorship with increasing BMI, with thresholds of 30-35 kg/m² commonly employed for patient selection [64, 65]. The NICE guidelines recommend UKA consideration up to BMI 35 kg/m², while most HTO selection criteria require BMI <32 kg/m² [66].

Implications for clinical practice

Based on the synthesized evidence, we propose the following provisional treatment considerations, which require validation in prospective studies before clinical implementation:

Consideration 1:

UKA may be preferentially considered for patients prioritizing pain relief and lower perioperative risk, particularly those aged ≥60 years or with BMI ≥27 kg/m².

Consideration 2:

HTO may be preferentially considered for younger patients (<60 years) with high activity demands (Tegner ≥5) who prioritize functional performance and return to sports.

Consideration 3:

Shared decision-making incorporating patient preferences, activity goals, and individual risk factors should guide procedure selection.

Consideration 4:

Modern surgical techniques—specifically opening-wedge HTO with locking plate fixation and mobile-bearing UKA—are recommended to optimize outcomes regardless of procedure selection.

Strengths and limitations

This meta-analysis possesses several methodological strengths. The comprehensive search strategy, duplicate independent screening, and rigorous quality assessment enhance internal validity. Pre-specified subgroup analyses based on clinically relevant patient characteristics provide actionable treatment guidance. Statistical methods including Hartung-Knapp adjustment and sensitivity analyses strengthen confidence in effect estimates.

Several limitations warrant acknowledgment. First, the predominance of observational study designs introduces potential selection bias and confounding, as patient allocation to HTO versus UKA was non-random in 8 of 10 included studies. Second, surgical technique heterogeneity—particularly the inclusion of historical closing-wedge HTO studies [38, 39]—may not fully reflect contemporary practice outcomes. However, leave-one-out sensitivity analysis demonstrated that exclusion of either study did not materially alter pooled effect estimates, suggesting that technique evolution did not disproportionately influence our findings. Sensitivity analyses restricted to opening-wedge studies demonstrated similar findings, partially mitigating this concern.

Third, follow-up duration heterogeneity (36-90 months) limits definitive survivorship conclusions. Long-term registry data suggest that outcome trajectories may diverge beyond 10 years, with HTO demonstrating higher late failure rates related to OA progression [67]. Fourth, the limited patient-level data precluded sophisticated meta-regression or individual patient data (IPD) meta-analysis that could more precisely define optimal selection thresholds.

Finally, publication bias remains possible despite non-significant Egger’s test and symmetric funnel plots. Negative studies comparing established procedures may be less likely to achieve publication, potentially inflating effect estimates. Additionally, prediction intervals were not calculated in this analysis; future studies should consider reporting prediction intervals to better characterize the expected range of treatment effects across different clinical settings. The BMI threshold of 27 kg/m² was derived from exploratory threshold analysis rather than conventional WHO classifications. This data-driven cutoff demonstrated clearer differentiation in treatment effects compared with standard thresholds of 25 or 30 kg/m², suggesting that optimal BMI stratification for surgical decision-making in knee osteoarthritis may differ from general population health metrics. This finding should be considered hypothesis-generating and warrants prospective validation.

Future research directions

Several evidence gaps warrant investigation. First, ongoing RCTs including UNIKORN and TOKA (Total/Unicompartmental vs Osteotomy for Knee Arthritis) will provide definitive comparative effectiveness data with contemporary techniques [56, 68]. Second, individual patient data meta-analysis incorporating baseline radiographic severity, limb alignment, and cartilage status could refine selection algorithms. Third, cost-effectiveness analyses incorporating revision surgery probability, rehabilitation requirements, and quality-adjusted life-years would inform health system resource allocation. Fourth, patient-reported outcome studies examining satisfaction, return-to-sport rates, and treatment expectations would complement the clinical outcome measures synthesized herein.

Conclusions

This comprehensive meta-analysis suggests that HTO and UKA represent distinct treatment strategies with potentially complementary outcome profiles for medial compartment knee osteoarthritis. Based on current evidence, UKA appears to provide superior pain relief, lower complication rates, and reduced revision risk, while HTO may offer better functional outcomes, higher activity restoration, and superior range of motion. Patient selection may be optimized based on age, activity demands, and treatment priorities. Preliminary subgroup analyses suggest that younger patients with high activity requirements may benefit preferentially from HTO, whereas older patients prioritizing pain relief and lower reoperation risk could be considered for UKA. However, these subgroup findings are exploratory and require prospective validation. Shared decision-making incorporating patient values remains essential for optimal treatment selection. Ongoing randomized trials utilizing contemporary surgical techniques are anticipated to refine these preliminary recommendations.

Authors’ contributions

Conceptualization, XM and KF; Methodology, XM.; Software, YW; Validation, KF; Formal Analysis, XM.; Investigation, YW; Data Curation, XM and YW; Writing—Original Draft, All authors; Writing—Review & Editing, All authors.

Funding

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

Data availability

All data generated or analyzed during this study are included in this published article and its supplementary information files.

Declarations

Ethics approval and consent to participate

This systematic review and meta-analysis was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines [27]. No ethical approval was required as this study utilized only published data.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.