Mid-term clinical outcomes of mobile-bearing UKA: a retrospective study on the influence of patellofemoral joint disease, lower limb alignment, and implant positioning

Abstract

Background

Unicompartmental knee arthroplasty (UKA) has become an effective treatment for medial compartment osteoarthritis of the knee. However, its use in patients who also have patellofemoral joint osteoarthritis (PFOA) before surgery remains controversial. Restoring postoperative lower limb alignment and achieving accurate implant positioning are also important factors for the success of UKA, but there is still a lack of studies that combine both patient-related factors and surgical technique-related factors in the Chinese population.

Methods

This study retrospectively analyzed 69 Chinese patients (79 knees) with medial compartment osteoarthritis who underwent Oxford UKA between May 2017 and December 2020. The severity and location of PFOA were assessed by MRI. Postoperative coronal alignment was categorized by the femorotibial angle (FTA) into neutral, mild varus, moderate varus, and extreme/out-of-range groups. Implant positioning was classified as ideal or non-ideal according to established radiographic target ranges for aLDFA, aMPTA, PTS, and component alignment angles. Functional outcomes were evaluated using VAS, KSS, WOMAC, and Kujala scores, with multifactorial interactions analyzed via multivariate analysis of covariance (MANCOVA).

Results

At a mean follow-up of 66.5 ± 9.6 months, Oxford UKA markedly improved overall patient function. While no cases of prosthesis loosening or revision were observed, three patients reported persistent postoperative pain and two presented with valgus deformity. Postoperative functional scores did not differ by PFOA severity or lesion location. Regarding lower-limb alignment, neutral and mild varus (< 6°) knees had higher KSS function scores than moderate varus and extreme/out-of-range knees and lower WOMAC scores. KSS knee scores were higher in neutral and mild varus than in the extreme group. Implant positioning (ideal vs non-ideal) was not associated with postoperative scores.

Conclusion

Mid-term follow-up showed that Oxford UKA achieved satisfactory clinical outcomes. The severity and location of PFOA before surgery did not significantly affect postoperative knee function. Maintaining postoperative alignment within the target range (neutral to slight varus; within 6° of varus) may be associated with better functional recovery.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12891-026-10086-7.

Introduction

Unicompartmental knee arthroplasty (UKA) has been developed over many years. Many studies have shown that, compared with total knee arthroplasty (TKA), UKA offers distinct advantages, including being less invasive, preserving joint proprioception, promoting faster recovery, and achieving higher patient satisfaction [1]. However, these advantages must be weighed against well-documented concerns regarding generally higher revision rates and debates surrounding long-term survivorship when compared to TKA [2]. Consequently, the optimal indications and surgical techniques for UKA are still intensely debated. One major point of controversy is whether patients who have patellofemoral joint osteoarthritis (PFOA) before surgery are suitable candidates for UKA. A meta-analysis by Lu [3] suggested that patients with pre-existing PFOA, provided there is no involvement of the lateral patellofemoral groove, should not be considered contraindicated for UKA. In contrast, a recent study by Andronic [4] reported that patients with lateral PFOA may have a higher risk of later revision following fixed-bearing UKA. Because fixed-bearing and mobile-bearing implants possess fundamentally different kinematics and load-transfer mechanics, it remains critical to evaluate whether these concerns apply equally to the mobile-bearing Oxford system. A systematic review and meta-analysis by Yang [5] found that preoperative PFOA is generally not linked to inferior function or implant survival after UKA. While classical studies by Beard [6] and recent meta-analyses have challenged the traditional view by demonstrating that anterior knee pain and pre-existing PFOA are not strict contraindications for UKA, most of these studies rely on conventional radiographs. Radiographs often lack the sensitivity to accurately detect early cartilage fibrillation or precisely map the topographical location of patellofemoral lesions. Consequently, a distinct gap remains in the literature: there is a lack of high-resolution, MRI-based evaluations of PFOA severity and anatomical distribution, particularly when analyzed concurrently with surgical technique parameters (lower limb alignment and implant positioning) in a standardized cohort.

For patients, whether choosing UKA or TKA, the accuracy of the surgical technique is an important factor that affects implant survival and functional outcomes. This includes restoring lower limb alignment after surgery and placing the implant in the correct position [7]. Some researchers believe that improving surgical precision and using computer-assisted technology may help restore alignment and improve the accuracy of implant placement [8, 9]. Conversely, Kennedy [10] reported in a large Western cohort that postoperative limb alignment was not associated with functional outcomes or revision risk after Oxford medial UKA. This discrepancy between theoretical biomechanical concerns and Kennedy’s clinical findings may stem from differences in alignment categorization—particularly the threshold for defining extreme outliers—or variations in population-specific anatomical baselines, such as the generally more pronounced constitutional varus in Asian demographics. As the Oxford UKA technique has been widely adopted in China since 2016, it remains crucial to determine whether Kennedy’s conclusions hold true for Chinese patients. Furthermore, previous literature rarely isolates the impact of pronounced malalignment (e.g., strict valgus or severe varus outliers) while concurrently assessing patient-specific variables like MRI-defined PFOA.

Although MRI is not routinely recommended as a mandatory imaging modality for preoperative evaluation before Oxford UKA [11], obtaining an ideal lateral radiograph is often challenging, making it difficult to accurately assess the presence of anteromedial osteoarthritis and the integrity of the anterior cruciate ligament. Therefore, as part of our strict institutional protocol during the study period, magnetic resonance imaging (MRI) was routinely performed for all patients considered for UKA to definitively rule out anterior cruciate ligament deficiency and lateral compartment disease. Importantly, MRIs were not selectively ordered based on a clinical suspicion of PFOA or anterior knee pain, thereby minimizing ascertainment bias. In addition, patellofemoral joint disease is relatively common in the Chinese population [12], and MRI provides excellent visualization of patellofemoral pathology. In this study, we retrospectively evaluated a Chinese cohort undergoing third-generation Oxford medial UKA to examine the associations of MRI-defined PFOA severity and location, postoperative limb alignment, and implant positioning with mid-term functional outcomes.

Materials and methods

Patient cohort

This study was approved by our institutional ethics committee (Approval Document No. JD-HG-2025–131) and retrospectively reviewed all consecutive patients who underwent cemented third-generation Oxford medial unicompartmental knee arthroplasty at our hospital from May 4, 2017, to December 21, 2020. Eligible cases were identified from institutional surgical records/electronic medical records, and imaging data were retrieved from the picture archiving and communication system. During the study period, surgical decision-making strictly adhered to the classic clinical indications for medial UKA: isolated medial compartment osteoarthritis, preserved joint range of motion with a flexion contracture of less than 15°, and a passively correctable varus deformity of less than 15°.

For this specific retrospective study, the formal inclusion criteria were defined as: (1) patients who underwent cemented third-generation Oxford medial UKA based on the aforementioned indications; and (2) the availability of complete imaging records, specifically including preoperative MRI, preoperative weight-bearing anteroposterior and lateral knee radiographs, and postoperative anteroposterior and lateral radiographs.

Exclusion criteria included imaging data that did not meet study requirements, secondary osteoarthritis, and the presence of advanced lateral patellofemoral trochlear groove changes. Importantly, these lateral trochlear changes (defined as Outerbridge Grade IV cartilage loss or significant osteophytosis on the lateral facet) were systematically evaluated and excluded using the mandatory preoperative axial MRI, rather than relying on inconsistent radiographic or intraoperative assessments. Anterior knee pain alone was not considered an absolute contraindication. The STROBE flow diagram summarizes the screening process, including the number of excluded cases with specific reasons. The details are further illustrated in an additional figure file (see Additional file 1).

Data analysis

Two orthopedic surgeons evaluated the preoperative and postoperative anteroposterior and lateral knee radiographs, as well as preoperative knee MRI. Preoperative tibiofemoral osteoarthritis severity was graded on weight-bearing anteroposterior knee radiographs using the Kellgren–Lawrence (K–L) classification (grades 0–4). Grading was performed independently by two orthopedic surgeons who were blinded to postoperative outcomes; disagreements were resolved by consensus. The femorotibial angle (FTA) was defined as the lateral angle formed by the anatomical axis of the femur (a line through the center of the femoral medullary canal) and the anatomical axis of the tibia (a line through the center of the tibial medullary canal), and was measured specifically on weight-bearing short anteroposterior knee radiographs (Fig. 1). The normal FTA is approximately 173 degrees. The degree of varus was calculated as FTA minus 173 degrees, with positive values indicating varus [13]. Based on postoperative FTA, patients were categorized into four groups: neutral alignment (≥ 173° and < 176°), mild varus (≥ 176° and < 179°), moderate varus (≥ 179° and < 182°), and an extreme/out-of-range group (< 173° or ≥ 182°). The extreme/out-of-range group was defined as patients with marked deviation from the intended postoperative alignment window. Because valgus and severe varus outliers were few, they were pooled to improve estimate stability and statistical power, while descriptive results for valgus and severe varus were also presented separately when appropriate. Postoperative anteroposterior and lateral radiographs were used to measure implant positioning parameters, including anatomic lateral distal femoral angle (aLDFA), anatomic medial proximal tibial angle (aMPTA), posterior tibial slope (PTS), coronal femoral component angle (c-FCA), and sagittal femoral component angle (s-FCA) (Figs. 2 and 3), based on previously published methods [14–18]. The ideal ranges for these parameters were defined based on established guidelines specific to the Oxford mobile-bearing UKA system and recent UKA-specific literature, which establish the acceptable tolerances as follows: aLDFA 79° to 83°, aMPTA 85° to 90°, PTS 83° to 87°, c-FCA −3° to 3°, and s-FCA 0° to 10° [19]. Cases with implant alignment values within these ranges were classified as the "ideal value group," and those outside these ranges were classified as the "non-ideal value group."

Fig. 1.

Radiographic measurement of femorotibial angle (FTA). Footnotes: A Preoperative measurement. B Postoperative measurement

Fig. 2.

Representative preoperative and postoperative radiographs. Footnotes: A preoperative anatomic lateral distal femoral angle (aLDFA), B preoperative anatomic medial proximal tibial angle (aMPTA), C preoperative posterior tibial slope (PTS), D postoperative aLDFA, E postoperative aMPTA, F coronal femoral component angle (c-FCA), G postoperative PTS and sagittal femoral component angle (s-FCA)

Fig. 3.

Representative MRI examples demonstrating the Outerbridge classification of patellofemoral joint osteoarthritis (PFOA). Footnotes: Grade I(A): cartilage softening, Grade II(B): fissures smaller than 1.3 cm, Grade III(C): fissures greater than 1.3 cm, Grade IV(D): exposed subchondral bone

Preoperative knee MRI was used to assess the severity and location of patellofemoral joint disease. While originally developed for arthroscopy, a modified MRI-based Outerbridge classification system was utilized to evaluate the severity of patellofemoral cartilage damage due to its widespread clinical familiarity and acceptable diagnostic correlation with intraoperative findings [20]. The grading was defined as follows: (grade 0: normal cartilage; grade 1: cartilage softening; grade 2: moderate cartilage fibrillation with fissures less than 1.3 cm; grade 3: severe cartilage fibrillation with fissures greater than 1.3 cm; grade 4: subchondral bone exposure) (Fig. 3) [21]. Patients were classified into three groups: no PFOA (grade 0), mild PFOA (grades 1–2), and moderate to severe PFOA (grades 3–4). Based on the location of cartilage damage on the patellofemoral articular surface, patients were further categorized as medial and central PFOA (lesions in the medial two-thirds of the patellofemoral surface) or lateral PFOA (lesions in the lateral one-third). All imaging measurements were performed using AutoCAD 2026 (Autodesk, Inc.). While dedicated orthopedic templating software is commonly used clinically, AutoCAD was specifically utilized in this study for its highly precise geometric toolset in identifying anatomical axes. Because this study strictly evaluated angular parameters rather than absolute linear distances, the angular accuracy remained unaffected by inherent variations in radiographic magnification, negating the need for scaling calibration.

Functional evaluations included the Visual Analog Scale (VAS) for pain [22]; the Knee Society Score (KSS), consisting of knee score and function score; the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), which includes pain, stiffness, and daily activity subscales, with higher scores indicating more severe symptoms [23]; and the Kujala score for anterior knee pain, with higher scores indicating better knee function and less pain [24].

A two-way random-effects model for absolute agreement was used to calculate the intraclass correlation coefficient (ICC) to evaluate interobserver and intraobserver reliability. Both interobserver and intraobserver assessments included all 79 knees, and the repeated measurements were performed one week after the initial evaluation. The ICC values for all angle measurements were good (0.75 to 0.90) or excellent (> 0.90) (Table 1).

Table 1.

Intra-observer and inter-observer reliability of radiographic measurements

		Intra-observer	Inter-observer
FTA	Preoperative Postoperative	0.99(0.98–1.00) 0.98(0.96–0.99)	0.99(0.97–0.99) 0.94(0.90–0.96)
aLDFA	Preoperative Postoperative	0.89(0.82–0.93) 0.86(0.78–0.90)	0.87(0.80–0.91) 0.84(0.72–0.89)
aMPTA	Preoperative Postoperative	0.91(0.85–0.95) 0.88(0.86–0.92)	0.94(0.89–0.97) 0.85(0.78–0.90)
PTS	Preoperative Postoperative	0.93(0.88–0.96) 0.95(0.91–0.97)	0.97(0.90–0.99) 0.95(0.87–0.97)
c-FCA		0.90(0.84–0.94)	0.80(0.60–0.93)
s-FCA		0.92(0.86–0.96)	0.81(0.72–0.86)

Values are intraclass correlation coefficients (ICC) with 95% confidence intervals. ICC values of 0.75–0.90 indicate good reliability, and values above 0.90 indicate excellent reliability

FTA Femorotibial angle, aLDFA anatomic lateral distal femoral angle, aMPTA anatomic medial proximal tibial angle, PTS Posterior tibial slope, c-FCA coronal femoral component angle, s-FCA sagittal femoral component angle

Statistical analysis

All data were checked by a designated reviewer after entry. Based on the results of the Shapiro–Wilk normality test, independent sample t-tests were used for continuous variables with normal distribution, and the Mann–Whitney U test was used for data with non-normal distribution. Categorical variables were compared using the chi-square test or Fisher's exact test when appropriate. Prior to conducting multivariate analyses, the fundamental assumptions of multivariate normality and the homogeneity of variance–covariance matrices were strictly verified using Levene's test and Box's M test, respectively. To account for observed baseline demographic imbalances and to mitigate the influence of potential confounding factors on functional recovery, a Multivariate Analysis of Covariance (MANCOVA) was utilized. Patient age, gender, and BMI were explicitly entered into the model as covariates. This MANCOVA evaluated the overall adjusted effects of PFOA grade, FTA grouping, and their interaction on multiple postoperative functional outcomes, including VAS, KSS, WOMAC, and Kujala scores. For factors showing significant main effects in the MANCOVA, further univariate analysis of covariance (ANCOVA) was performed to determine their specific effects on each functional score. Subsequent pairwise comparisons between subgroups were conducted using post-hoc tests based on the estimated marginal means from the ANCOVA model. Fisher's Least Significant Difference method was utilized without adjustment for multiple comparisons to minimize Type II errors in this exploratory context. A p value of less than 0.05 was considered statistically significant. All statistical analyses were performed using SPSS version 27.0.

Results

Baseline characteristics of patients

A total of 69 eligible patients (79 knees) were included in this study. The average follow-up time was 66.5 ± 9.6 months. The mean age at the time of surgery was 62.2 ± 8.0 years. The mean BMI was 25.7 ± 3.4 kg/m². The study group consisted of 54 female patients and 15 male patients. Preoperatively, tibiofemoral osteoarthritis was end-stage in nearly all knees: 76/79 (96.2%) were graded as K–L grade 4, and 3/79 (3.8%) as K–L grade 3. The distribution of K–L grades did not differ significantly among groups. However, formal baseline testing revealed a significantly higher proportion of male patients in the moderate-to-severe PFOA group (42.9%) compared to the no-PFOA (18.8%) and mild-PFOA (12.2%) cohorts (p < 0.05). Consequently, patient gender was adjusted for as a covariate in all subsequent multivariate outcome analyses. Demographic and clinical data were analyzed and compared across groups (Tables 2, 3, 4, 5, and 6).

Table 2.

Patient demographics and baseline characteristics by PFOA severity

Characteristic	Overall Cohort	No PFOA	Mild PFOA	M/S PFOA
N	79	16	49	14
Demographics
Mean Age	62.2 ± 8.0	62.8 ± 8.7	62.2 ± 8.1	61.6 ± 7.2
% Male (n)	19.0(15)	18.8(3)	12.2(6)	42.9(6)
Mean BMI	25.7 ± 3.4	25.3 ± 2.9	25.8 ± 3.7	26.2 ± 3.0
Mean Follow-up	66.5 ± 9.6	70.5 ± 6.0	65.7 ± 11.3	64.9 ± 4.0
Preoperative clinical scores
Mean VAS	7.6 ± 1.0	7.4 ± 1.2	7.6 ± 0.9	8.0 ± 0.8
Mean KSS-F	53.7 ± 3.1	54.0 ± 2.7b	54.3 ± 2.8b	50.9 ± 3.2
Mean KSS-K	54.9 ± 2.8	54.9 ± 2.8	55.6 ± 2.5b	52.4 ± 2.4
Mean WOMAC	45.4 ± 2.9	44.3 ± 3.1b	45.4 ± 3.0	46.8 ± 1.9
Mean Kujala	61.1 ± 4.5	65.0 ± 3.4ab	61.9 ± 2.6b	54.1 ± 3.6
Preoperative radiographic characteristics
%K–L grade IV (n)	96.2(76)	100.0(16)	95.9(47)	92.9(13)
FTA	180.1 ± 3.3	179.0 ± 3.3	180.6 ± 3.3	179.9 ± 3.0
Mean aLDFA	82.0 ± 1.9	81.4 ± 2.3	82.3 ± 1.9	81.4 ± 1.6
Mean aMPTA	85.4 ± 2.1	85.7 ± 2.2	85.5 ± 2.2	84.5 ± 1.7
Mean PTS	84.0 ± 3.8	83.6 ± 3.9	84.0 ± 3.9	84.7 ± 3.8

The values are given as the mean and standard deviation

PFOA Patellofemoral osteoarthritis, M/S PFOA Moderate-to-severe PFOA, BMI Body mass index, VAS Visual Analog Scale for pain, KSS-F Knee Society Score – function, KSS-K Knee Society Score – knee, WOMAC Western Ontario and McMaster Universities Osteoarthritis Index

^aThe difference was statistically significant compared to the mild PFOA group

^bThe difference was statistically significant compared to the moderate-to-severe PFOA group

Table 3.

Postoperative clinical outcomes by PFOA severity

	Overall Cohort	No PFOA	Mild PFOA	M/S PFOA
Mean VAS	0.7 ± 0.7	0.7 ± 0.6	0.7 ± 0.8	0.8 ± 0.6
Mean KSS-F	83.6 ± 3.4	83.3 ± 3.3	83.9 ± 4.2	82.9 ± 2.4
Mean KSS-K	83.6 ± 3.7	83.4 ± 3.1	83.5 ± 3.8	83.9 ± 2.5
Mean WOMAC	5.0 ± 2.9	5.6 ± 2.3	4.8 ± 3.3	4.9 ± 2.3
Mean Kujala	84.4 ± 2.7	84.0 ± 2.4	84.6 ± 3.0	84.4 ± 1.6

There were no statistically significant differences in postoperative functional scores among the three PFOA severity groups (p > 0.05)

The values are given as the mean and standard deviation

PFOA Patellofemoral osteoarthritis, M/S PFOA Moderate-to-severe PFOA, VAS Visual Analog Scale, KSS-F Knee Society Score – function, KSS-K Knee Society Score – knee, WOMAC Western Ontario and McMaster Universities Osteoarthritis Index

Table 4.

Clinical outcomes based on the location of PFOA

	Mid-medial PFOA	Lateral PFOA	p
N	41	22
Demographics
Mean Age	62.4 ± 7.8	61.5 ± 8.1	0.662
% Male (n)	19.5(8)	18.2(4)
Mean BMI	25.2 ± 2.6	27.0 ± 4.7	0.158
Mean Follow-up	64.2 ± 11.2	68.0 ± 7.4	0.031
%K–L grade IV (n)	95.1(39)	95.5(21)
Preoperative clinical scores
Mean VAS	7.6 ± 0.9	7.9 ± 0.9	0.193
Mean KSS-F	53.5 ± 3.2	53.7 ± 3.3	0.819
Mean KSS-K	55.0 ± 2.9	54.6 ± 2.6	0.519
Mean WOMAC	45.2 ± 2.9	46.8 ± 2.5	0.084
Mean Kujala	61.7 ± 4.3	60.5 ± 4.8	0.237
Postoperative clinical scores
Mean VAS	0.7 ± 0.6	0.8 ± 1.0	0.830
Mean KSS-F	83.9 ± 3.4	83.3 ± 4.6	0.919
Mean KSS-K	83.8 ± 3.3	83.2 ± 4.0	0.680
Mean WOMAC	4.7 ± 2.4	5.1 ± 4.1	0.642
Mean Kujala	84.6 ± 2.5	84.2 ± 2.8	0.645

Mid-medial PFOA indicates lesions involving the medial and central two-thirds of the patellofemoral articular surface; lateral PFOA indicates lesions located in the lateral one-third. There were no significant differences in postoperative functional outcomes between the two groups (p > 0.05). The values are given as the mean and standard deviation

VAS Visual Analog Scale, KSS-F Knee Society Score – function, KSS-K Knee Society Score – knee, WOMAC Western Ontario and McMaster Universities Osteoarthritis Index

Table 5.

Comparison of clinical outcomes by postoperative lower limb alignment groups

	Neutral	Mild varus	Moderate varus	Extreme/out-of-range
N	19	25	23	12
Demographics
Mean Age	59.6 ± 8.9	62.7 ± 7.9	63.7 ± 7.9	62.6 ± 6.6
% Male (n)	15.8(3)	24.0(6)	13.0(3)	25.0(3)
Mean BMI	26.7 ± 4.4	24.8 ± 2.8	26.5 ± 3.1	24.9 ± 3.3
Mean Follow-up	66.5 ± 8.4	65.2 ± 8.5	65.6 ± 12.3	71.1 ± 6.9
%K–L grade IV (n)	94.7(18)	96.0(24)	95.7(22)	100.0(12)
Preoperative clinical scores
Mean VAS	7.7 ± 1.0	7.8 ± 1.0	7.5 ± 0.8	7.4 ± 0.8
Mean KSS-F	52.9 ± 3.5	53.8 ± 3.3	53.8 ± 2.7	54.1 ± 2.8
Mean KSS-K	54.7 ± 2.8	55.0 ± 2.9	54.9 ± 2.5	54.7 ± 3.3
Mean WOMAC	45.8 ± 2.5	45.2 ± 2.9	45.1 ± 2.6	45.9 ± 4.1
Mean Kujala	60.2 ± 4.9	61.6 ± 4.8	61.5 ± 4.1	60.9 ± 4.6
Postoperative clinical scores
Mean VAS	0.8 ± 0.6	0.7 ± 0.7	0.7 ± 0.6	0.8 ± 1.2
Mean KSS-F	84.3 ± 2.4ab	86.1 ± 2.8ab	81.8 ± 3.1	80.7 ± 4.6
Mean KSS-K	84.7 ± 2.6a	84.4 ± 3.1a	83.1 ± 2.9	80.8 ± 4.6
Mean WOMAC	4.0 ± 2.3ab	3.2 ± 1.7ab	6.3 ± 1.6	7.8 ± 4.5
Mean Kujala	84.7 ± 2.0a	85.4 ± 2.3a	84.4 ± 2.3	82.1 ± 3.7

The values are given as the mean and standard deviation. The Extreme/out-of-range group consists of valgus and severe varus alignments (< 173° or ≥ 182°) and does not imply biomechanical equivalence between valgus and severe varus

FTA Femorotibial angle, VAS Visual Analog Scale, KSS-F Knee Society Score – function, KSS-K Knee Society Score – knee, WOMAC Western Ontario and McMaster Universities Osteoarthritis Index

^asignificantly different compared with the extreme/out-of-range group (p < 0.05)

^bsignificantly different compared with the moderate varus group (p < 0.05)

Table 6.

Postoperative functional outcome comparison between ideal and non-ideal prosthesis positioning groups

	aLDFA (72–85°)	aMPTA (79–90°)	PTS (80–94°)	c-FCA (−12–17°)	s-FCA (0–29°)
	Ideal (n = 33)	Ideal (n = 33)	Ideal (n = 29)	Ideal (n = 23)	Ideal (n = 51)
	Non-ideal (n = 46)	Non-ideal (n = 46)	Non-ideal (n = 50)	Non-ideal (n = 56)	Non-ideal (n = 28)
Mean VAS	0.6 ± 0.9 0.8 ± 0.6	0.7 ± 0.7 0.7 ± 0.8	0.7 ± 0.7 0.8 ± 0.8	0.5 ± 0.6 0.8 ± 0.8	0.7 ± 0.6 0.8 ± 0.9
Mean KSS-F	83.4 ± 4.3 83.7 ± 3.3	84.2 ± 3.6 83.1 ± 3.8	84.2 ± 3.5 83.3 ± 3.9	83.7 ± 3.4 83.6 ± 3.9	83.1 ± 3.4 84.5 ± 4.1
Mean KSS-K	83.3 ± 3.9 83.7 ± 3.1	84.2 ± 3.0 83.1 ± 3.7	83.8 ± 3.4 83.4 ± 3.5	84.4 ± 2.8 83.2 ± 3.6	83.4 ± 3.1 83.9 ± 4.0
Mean WOMAC	5.2 ± 3.7 4.8 ± 2.3	4.4 ± 2.4 5.5 ± 3.2	4.6 ± 2.4 5.3 ± 3.2	4.8 ± 2.3 5.1 ± 3.2	5.0 ± 2.4 5.0 ± 3.8
p	> 0.05	> 0.05	> 0.05	> 0.05	> 0.05

The values are given as the mean and standard deviation

aLDFA anatomic lateral distal femoral angle, aMPTA anatomic medial proximal tibial angle, PTS Posterior tibial slope, c-FCA coronal femoral component angle, s-FCA sagittal femoral component angle, VAS Visual Analog Scale, KSS-F Knee Society Score – function, KSS-K Knee Society Score – knee, WOMAC Western Ontario and McMaster Universities Osteoarthritis Index

Among the included knees, 16 knees (20.3%) showed no PFOA. A total of 49 knees (62.0%) showed mild PFOA (Outerbridge grade 1 to 2), and 14 knees (17.7%) showed moderate to severe PFOA (Outerbridge grade 3 to 4). Preoperatively, patients with moderate-to-severe PFOA exhibited significantly worse functional status compared to both the no-PFOA and mild-PFOA groups. This deterioration was consistently reflected across multiple patient-reported outcome measures, including significantly lower KSS function and knee scores, higher WOMAC scores, and reduced Kujala scores (all p < 0.001) (Table 2). However, when contextualized using established Minimal Clinically Important Difference (MCID) thresholds, the absolute preoperative differences in KSS and WOMAC scores fell below the level of clinical significance. Conversely, the ~ 10.9 point reduction in the Kujala score for the moderate-to-severe group met the MCID threshold, accurately reflecting a clinically meaningful increase in anterior knee symptoms prior to surgery (Table 2). A total of 63 knees showed patellofemoral joint disease. Among them, 41 knees (51.9%) had lesions located in the medial and central region, and 22 knees (27.8%) had lesions located in the lateral region. There were no significant differences in preoperative VAS scores, KSS scores, or WOMAC scores between these two groups (Table 4).

In terms of postoperative lower limb alignment, 19 knees (24.1%) were categorized into the neutral alignment group (≥ 173° and < 176°) with a mean FTA of 174.2° ± 0.86°. The mild varus group (≥ 176° and < 179°) included 25 knees (31.6%), with a mean FTA of 177.0° ± 0.74°. The moderate varus group (≥ 179° and < 182°) included 23 knees (29.1%), with a mean FTA of 179.8° ± 0.80°. The extreme/out-of-range group (< 173° or ≥ 182°) included 12 knees (15.2%). To ensure biomechanical transparency, this group was further descriptively divided into two distinct subgroups: a valgus overcorrection subgroup (n = 5; mean FTA 171.8° ± 0.4°) and a severe varus undercorrection subgroup (n = 7; mean FTA 183.0° ± 1.1°). Preoperative VAS, KSS, WOMAC and Kujala scores showed no significant differences among the four main groups (Table 5).

Postoperative and preoperative functional scores across groups

To evaluate the improvement of knee function after UKA, we compared the functional scores of all included patients before surgery and at mid-term follow-up after surgery. The results showed that postoperative VAS pain scores, KSS knee scores, KSS function scores, WOMAC scores, and Kujala scores all improved significantly compared with preoperative values (p < 0.001). These findings indicate that UKA can effectively relieve pain and improve knee function and quality of life (Tables 2, 4, and 5).

Severity/location of patellofemoral joint disease and postoperative functional scores after UKA

We found that among patients with different severities of patellofemoral joint disease, there were no significant differences in postoperative KSS scores, WOMAC scores, or Kujala scores between the groups (Table 3). There were no significant differences in postoperative VAS, KSS, WOMAC, or Kujala scores between the two groups with different locations of patellofemoral joint disease. The location of patellofemoral joint disease did not have a notable effect on postoperative function after UKA (Table 4).

Lower limb alignment and postoperative functional scores after UKA

Based on the FTA, patients were categorized into the neutral alignment group, mild varus group, moderate varus group, and extreme/out-of-range group. Given the small number of valgus and severe varus cases, they were analyzed together as an extreme/out-of-range group to improve the stability of estimates; the two subgroups were not assumed to be biomechanically equivalent.

Regarding postoperative alignment, achieving neutral or mild varus (within 6° of varus) yielded superior mid-term outcomes. Specifically, these two target-range groups demonstrated significantly higher KSS function and knee scores, alongside lower (better) WOMAC and Kujala scores, when compared to the moderate varus and extreme/out-of-range groups (p < 0.05) (Table 5). No significant functional differences were observed between the neutral and mild varus groups themselves. Importantly, while the small sample sizes of the valgus (n = 5) and severe varus (n = 7) subgroups precluded formal comparative statistics between them, a descriptive sensitivity analysis confirmed that both subgroups consistently exhibited inferior functional trends (lower KSS and higher WOMAC scores) compared to the neutral and mild varus cohorts, confirming that any extreme deviation from the target window is detrimental.

Implant positioning and postoperative functional scores after UKA

Based on previous literature, implant positioning parameters were grouped into ideal value and non-ideal value categories. Postoperative functional scores (VAS, KSS, and WOMAC) were compared between the two groups. The results showed that there were no significant differences in postoperative functional scores between the ideal value group and the non-ideal value group for any implant positioning parameter, including aLDFA, aMPTA, PTS, c-FCA, and s-FCA (p > 0.05) (Table 6). In this study cohort, most implant positioning values did not fall within the ideal range, but this did not have a significant effect on mid-term postoperative outcomes.

Multivariate analysis

To investigate whether the severity of patellofemoral joint disease and lower limb alignment had an interaction effect on postoperative functional outcomes, a MANCOVA was performed, explicitly adjusting for patient gender, age, and BMI as covariates. The adjusted results showed that postoperative FTA grouping maintained a significant main effect on the combined postoperative functional outcomes (Pillai's Trace = 0.406, F = 2.584, p = 0.003, partial η² = 0.135). However, the main effect of PFOA grading (Pillai's Trace = 0.053, F = 0.441, p = 0.895) and the interaction effect between PFOA grading and FTA grouping (Pillai's Trace = 0.376, F = 1.159, p = 0.280) were not significant. Subsequent univariate analysis showed that the main effect of FTA grouping was reflected in KSS function scores (F = 5.788, p = 0.001) and WOMAC scores (F = 7.972, p < 0.001). PFOA grading did not show significant main effects in any postoperative functional score (Fig. 4).

Fig. 4.

Box plots of postoperative KSS function scores and WOMAC scores across FTA groups. Footnotes: Boxes show medians and interquartile ranges; whiskers represent 1.5 × IQR. Multivariate analysis demonstrated a significant main effect of FTA grouping (Pillai's Trace = 0.406, p = 0.003), with univariate effects on KSS function (F = 5.788, p = 0.001) and WOMAC (F = 7.972, p < 0.001). Significance brackets indicate pairwise post-hoc comparisons: ★ p < 0.05

Complications

During mid-term follow-up, none of the 79 UKA cases developed complications such as infection, periprosthetic fracture, implant loosening, polyethylene insert dislocation, or progressive osteoarthritis in the lateral compartment. No revision surgeries were reported for any reason. Three patients reported persistent postoperative pain. Two patients presented with valgus knee but declined further surgery. Three patients received conservative treatment, and one of these patients underwent intra-articular hyaluronic acid injection.

Discussion

The UKA procedure for medial compartment osteoarthritis of the knee has gradually expanded in clinical practice in China over nearly ten years. However, there are still few studies based on the Chinese population that combine both patient-related factors and surgical technique factors to comprehensively evaluate postoperative clinical outcomes. The main purpose of this study was to analyze the influence of preoperative patellofemoral joint disease severity and location, postoperative lower limb alignment, and implant positioning on mid-term functional outcomes after UKA.

Our finding that pre-existing PFOA does not significantly affect postoperative knee function is largely in line with existing evidence and recent meta-analyses [5]. However, the incremental contribution and novelty of the present study lie in its methodological approach and concurrent multifactorial analysis.

First, while previous large-cohort studies predominantly relied on conventional radiographs—which often underestimate cartilage damage—our study utilized MRI for all patients. This approach provided high-resolution, Outerbridge-graded evaluations and precise topographical mapping (medial/central vs. lateral) of the PFOA, thereby offering more robust anatomical evidence to support the expansion of UKA indications.

Second, and perhaps more importantly, this study uniquely modeled patient-specific factors (PFOA) concurrently with surgeon-controlled variables (postoperative alignment and implant positioning). By employing a multivariate analysis, we demonstrated a distinct lack of interaction effect between PFOA grading and FTA grouping. This is a critical addition to the literature, as it indicates that the clinical benefits of achieving optimal postoperative alignment (neutral to mild varus) are completely independent of the patient’s preoperative PFOA severity. This concurrent analysis provides nuanced, actionable reassurance for surgeons that was previously scarce in the literature. Regarding alignment, Kennedy [10] found no relationship between postoperative limb alignment and outcomes or revision after Oxford medial UKA. This divergence from our findings is highly likely attributable to significant methodological differences. Kennedy evaluated a vastly larger cohort (n = 824) and utilized full-length standing radiographs to assess the hip-knee-ankle mechanical angle, which represents the gold standard. In contrast, our study was limited to a small cohort (n = 79) evaluated via the FTA on short knee radiographs, which is inherently more susceptible to rotational and positioning variations. Furthermore, their robust sample size allowed for a different statistical handling of outliers. However, within our specific evaluation framework, outcomes were similar within the intended alignment range (neutral to mild varus), whereas moderate varus and outlier alignment (valgus or severe varus) were consistently associated with worse patient-reported outcomes.

Our study found that Oxford UKA achieved satisfactory outcomes in patients with medial compartment osteoarthritis and significantly improved postoperative knee function. The classical indication for UKA targets isolated tibiofemoral compartment disease, although some researchers have suggested that patellofemoral joint disease should not be considered a contraindication for UKA [25, 26]. In China, the prevalence of patellofemoral joint disease is relatively high. In our study, among 79 knees with medial compartment osteoarthritis, 63 knees (79.7%) had varying degrees of patellofemoral joint disease. In the early period of performing Oxford UKA, we took a cautious approach, and each patient underwent MRI to evaluate the condition of the patellofemoral joint and the integrity of the anterior cruciate ligament. Although pre-existing moderate-to-severe PFOA predictably manifested with statistically worse preoperative functional deficits (most notably meeting the MCID for worsened anterior knee symptoms on the Kujala scale), our mid-term data confirmed that these patients ultimately achieved functional parity with the non-PFOA cohort following UKA. In addition, regarding the location of patellofemoral joint disease, there were no significant differences between the medial/central PFOA group and the lateral PFOA group in postoperative KSS, WOMAC, or Kujala scores. This suggests that preoperative PFOA has a limited influence on postoperative outcomes after UKA. Although UKA does not directly address the patellofemoral joint, improving lower limb alignment and joint biomechanics may indirectly reduce patellofemoral joint stress, thereby improving symptoms [4, 27, 28]. Our findings align with the well-documented notion [6] that pre-existing PFOA does not compromise mid-term functional recovery, suggesting that it may not necessarily act as an absolute contraindication for UKA. The clinical significance of our findings is particularly noteworthy despite the relatively small sample size of our retrospective cohort. By utilizing MRI rather than standard radiographs, we were able to provide a highly sensitive, high-resolution assessment of PFOA, confirming that even when moderate-to-severe cartilage lesions are definitively identified, mid-term functional outcomes remain uncompromised. Furthermore, by evaluating both patient-specific factors (MRI-graded PFOA) and surgical precision (FTA and implant positioning) simultaneously, our results offer actionable clinical guidance: surgeons can be confident in performing UKA on patients with concomitant PFOA, provided that meticulous surgical technique is employed to maintain postoperative alignment within the neutral to mild varus window.

Accurate restoration of lower limb alignment after surgery is one of the key factors influencing the long-term success of UKA. We found that the neutral alignment group and the mild varus group showed significantly better outcomes in KSS function scores, WOMAC scores and Kujala scores compared with the moderate varus group and the extreme/out-of-range group. Among the three patients who reported persistent postoperative knee pain, two were in the valgus group. Clinically, these specific cases of valgus overcorrection were likely the result of medial compartment 'over-stuffing' during our early learning curve. In a misguided attempt to forcefully correct a severe or rigid preoperative varus deformity, excessively thick polyethylene bearings were utilized. In the context of mobile-bearing UKA, such over-stuffing pathologically over-tensions the superficial medial collateral ligament. This not only drives the knee into an unnatural valgus alignment but also serves as the primary mechanical source of persistent medial-sided pain. Furthermore, previous biomechanical studies have shown that overcorrection during surgery can abnormally increase the load on the non-replaced lateral compartment, which may accelerate cartilage degeneration in that area [29]. On the other hand, restoring optimal alignment helps maintain dynamic stability throughout knee motion and reduces the risk of subluxation or instability [30]. Other studies have shown that intraoperative sensor technology can monitor real-time pressure distribution across the implant and provide guidance to optimize alignment and joint balance [31]. Additionally, abnormal alignment has been shown to alter patellar tracking, which may contribute to anterior knee pain and functional impairment [32]. If preoperative coronal plane deformity is not adequately corrected, postoperative patellofemoral joint mismatch is more likely to occur [33]. Our findings further support that controlling postoperative lower limb alignment within neutral or mild varus is associated with better functional outcomes and patient satisfaction. For patients with severe preoperative varus deformity in whom valgus stress radiographs still show persistent varus, UKA should be performed cautiously.

We acknowledge that the alignment-related differences observed in our cohort were generally below published MCID thresholds [34–36], a factor that can materially alter the clinical interpretation of our statistical findings [37]. In addition, UKA-specific MCID/PASS thresholds for commonly used PROMs are still evolving and may vary across methods and instruments [38]. Therefore, our findings are best interpreted as evidence of a robust statistical association. UKA-specific, patient-anchored MCIDs are warranted.

In this cohort, valgus and severe varus cases were pooled into a single “extreme/out-of-range” category because each subgroup was small, which would otherwise yield unstable estimates and limited power. This grouping does not imply that valgus and severe varus share the same biomechanics; rather, it was intended to capture substantial deviation from the target postoperative coronal alignment window. Accordingly, our results should be interpreted as a comparison between the target range (neutral to mild varus) and marked deviation, not as evidence of equivalence between valgus and severe varus. Larger studies are warranted to evaluate these extremes separately.

We noted that the results of the multivariate analysis showed that, after controlling for PFOA grading, there was a statistically significant overall difference in postoperative functional scores among patients grouped by FTA. At the same time, there was no interaction effect between PFOA grading and FTA grouping. This indicates that the influence of FTA angle on postoperative outcomes is independent and does not depend on the severity of the patient’s PFOA.

Categorizing continuous variables reduces statistical power and can potentially mask subtle or non-linear relationships between implant position and clinical outcomes. Specifically, as noted in Table 6, our 'non-ideal' cohort included severe outliers (e.g., c-FCA up to 17° and s-FCA up to 29°), which likely reflect surgical inaccuracies during our institution's early learning curve. Therefore, our findings should not be misinterpreted as suggesting that surgical accuracy is unimportant. Rather, the lack of significant functional difference at mid-term likely reflects the inherent tolerance of the mobile-bearing Oxford UKA design to minor or moderate positional deviations [39]. Importantly, the survival of the extreme outliers in our mid-term follow-up must not be construed as clinical safety. Such severe malpositioning theoretically increases the risk of edge loading, impingement, and accelerated polyethylene wear. Precise implant positioning remains paramount for optimizing joint kinematics and ensuring long-term survivorship [40–42]. The long-term effects of implant positioning still need to be verified with larger sample sizes and longer follow-up periods in future studies.

No revision surgeries were observed during mid-term follow-up; however, three patients required additional conservative treatment due to persistent pain. This suggests that postoperative pain management and individual variability in recovery response should receive careful attention [43, 44]. Furthermore, long-term follow-up is important to assess whether abnormal postoperative lower limb alignment may lead to late mechanical failure, indicating the need for extended observation.

Based on the current results, we believe that Oxford UKA can significantly improve knee function in patients with medial compartment osteoarthritis. The presence of PFOA before surgery did not negatively impact mid-term functional outcomes in our cohort, indicating it may not be an absolute contraindication; however, long-term studies incorporating formal survivorship analysis are necessary to confirm this perspective. Despite the aforementioned limitations regarding long-term survivorship and the retrospective nature of this single-implant study, our findings offer highly actionable guidance for routine clinical practice. First, regarding surgical decision-making, this study provides reassurance to surgeons during patient selection: the definitive presence of PFOA on preoperative MRI does not necessitate abandoning a less invasive UKA in favor of TKA. Second, regarding intraoperative execution, this study establishes a clear mechanical target: striving for neutral to mild varus alignment (< 6°) is critical for optimizing mid-term functional recovery, whereas extreme varus or any valgus overcorrection should be strictly avoided. These insights directly inform patient counseling and intraoperative alignment strategies for mobile-bearing UKA.

This study has several limitations. First, as a retrospective, single-center analysis, selection bias was inevitable, and the external validity of our findings is inherently limited. Specifically, our cohort consisted entirely of Chinese patients, and an overwhelming majority (96.2%) presented with end-stage (K-L grade 4) medial compartment osteoarthritis, meaning our results may not be fully generalizable to populations presenting at earlier stages. Second, the study lacked a matched control group (such as patients undergoing TKA), precluding direct comparative conclusions regarding relative efficacy. Third, our strict inclusion criteria requiring complete preoperative and postoperative imaging (including MRI) may have introduced survivor bias, as patients with poorer compliance or earlier clinical failures could have been inadvertently excluded, likely contributing to the 0% revision rate observed at mid-term. Furthermore, our radiographic analysis relied entirely on standard 2D plain radiographs. Because postoperative 3D computed tomography (CT) scans are not routinely performed for asymptomatic UKA patients at our institution—to minimize unnecessary radiation exposure and healthcare costs—we were unable to reliably assess the rotational alignment of the components. Additionally, while AutoCAD provided precise geometric angular assessments for this study, its use as a general drafting tool rather than a dedicated orthopedic templating software (e.g., TraumaCad) is unconventional, potentially limiting the direct methodological comparability of our radiographic workflow with other studies. Moreover, our subgroup analyses regarding implant positioning (particularly for parameters with small 'ideal' sample sizes, such as c-FCA and PTS) were likely underpowered. Consequently, the lack of significant functional differences between ideal and non-ideal positioning should be interpreted with caution, as an inherent risk of a Type II error (false negative) cannot be definitively ruled out. Finally, while clinically pragmatic, our use of a modified Outerbridge classification for MRI assessment of PFOA is inherently less standardized for imaging than dedicated MRI-specific scoring systems (such as MOAKS or WORMS), which may have limited the analytical granularity of our cartilage evaluation. Furthermore, we did not apply statistical corrections for multiple comparisons (e.g., Bonferroni correction) during our post-hoc analyses among the four alignment subgroups. Consequently, there is an inherent risk of Type I error inflation, and the specific pairwise p-values reported in this study should be interpreted with appropriate caution as exploratory trends.

Conclusion

Oxford unicompartmental knee arthroplasty can significantly improve symptoms in patients with medial compartment osteoarthritis of the knee. The severity and location of pre-existing patellofemoral joint disease did not negatively impact mid-term functional outcomes. Maintaining postoperative lower limb alignment within the target range (neutral to slight varus) is associated with satisfactory functional recovery. While the mobile-bearing design demonstrated some functional tolerance to minor variations in implant positioning, striving for optimal surgical precision remains essential.

Abbreviations

UKA

Unicompartmental knee arthroplasty

TKA

Total knee arthroplasty

PFOA

Patellofemoral joint osteoarthritis

MRI

Magnetic resonance imaging

FTA

Femorotibial angle

aLDFA

Anatomic lateral distal femoral angle

aMPTA

Anatomic medial proximal tibial angle

PTS

Posterior tibial slope

c-FCA

Coronal femoral component angle

s-FCA

Sagittal femoral component angle

VAS

Visual Analog Scale

KSS

Knee Society Score

WOMAC

Western Ontario and McMaster Universities Osteoarthritis Index

ICC

Intraclass correlation coefficient

MANCOVA

Multivariate analysis of covariance

ANCOVA

Analysis of covariance

MCID

Minimal clinically important difference

CT

Computed tomography

Authors’ contributions

Mingkang Chen and Changqing Zhang carried out the entire procedure including the study design, data extraction, statistical analysis, manuscript writing and editing. Changqing Zhang conceived the study, coordinated and participated in the entire process of drafting and editing the manuscript. Zelin Fan contributed to the data extraction, and statistical analysis. Wei Xu contributed to the study design and data review. All authors have contributed significantly. All authors read and approved the final manuscript.

Funding

This work was supported by the Suzhou Key Disciplines-Sport Medicine (No. SZXK202506) and the Suzhou key project of Science and Education Strengthening Healthcare (No. ZDXM2024006).

Data availability

The data supporting the results of this study were sourced from relevant medical institutions, but their availability was limited. These data were used under the permission of this study and thus could not be publicly accessed. However, data could be obtained from the corresponding author at the reasonable request and permission of Second Affiliated Hospital of Soochow University.

Declarations

Ethics approval and consent to participate

The Ethics Committee of the Second Affiliated Hospital of Soochow University (November 13, 2025, Approval number JD-HG-2025–131). All procedures performed in this study were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards.

Written informed consent was obtained from individual or guardian participants.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.