Severity of knee osteoarthritis does not affect clinical outcomes following proximal fibular osteotomy – A systematic review and pooled analysis

Abstract

Background & aims

Knee osteoarthritis (KOA) is a progressive degenerative disease of chronic nature. The mainstay of surgical management for KOA would be total knee arthroplasty. Joint preserving options like High Tibial Osteotomy (HTO) and Proximal Fibular Osteotomy (PFO) have been offered as an inexpensive option by knee preservation surgeons. Current literature on PFO outcomes lack of clarity for specific indications for offering PFO based on degree of severity of KOA. Therefore, this systematic review aims to critically evaluate clinical and radiological outcomes of PFO stratified by severity of KOA.

Methods

PubMed, Scopus, CINAHL and Google Scholar databases were searched. Eligible studies included those published up till August 2023, with 271 studies obtained. After duplicate removal, title-abstract screening, and a full text screen based on inclusion and exclusion criteria, 11 papers were included. 46 papers were further identified from snowballing of 7 existing systematic reviews, with 2 additional papers subsequently included.

Results

13 included articles analysed 788 knees. Our study found that indications based on KL grading of KOA do not seem to differ in terms of post-operative clinical outcomes (VAS score) and radiological measures also found that hip knee alignment was improved regardless of KL grading of KOA. Additionally, the most common post-operative complication reported was deep peroneal nerve palsy.

Conclusion

PFO is a viable knee joint preserving surgery for medial compartment KOA, however given the high risk for complications reported in the literature, surgeons should pay close attention to the neuroanatomical landmarks and techniques to avoid neurovascular injury.

1. Introduction

Knee osteoarthritis is a progressive degenerative disease of chronic nature with a growing prevalence globally.¹ The current mainstay of surgical management for knee osteoarthritis would be a total knee arthroplasty. Despite being a safe and efficacious procedure for a majority of patients, cost and invasiveness are factors that have been brought into question.² Furthermore, the longevity of implants³ and the need for revision total knee arthroplasty makes offering TKA to younger patients ever more challenging given poorer outcomes in revision TKA.⁴

Joint preserving surgical options such as High Tibial Osteotomy (HTO) have been offered as an alternative and less invasive option for certain patient populations.⁵ Recently, Proximal Fibular Osteotomy (PFO) has been an inexpensive option offered by joint preservation surgeons with the aim of correcting the non-uniform settlement between the fibular and tibial plateau, relieving the load on the medial joint compartment.⁶ The concept of non-uniform settlement in the knee joint arises from the biomechanical asymmetry of load distribution between the fibula and tibia. The fibula provides lateral support over the lateral condyle of the tibia, assisting in weight transmission. However, there is a lack of an analogous support structure on the medial condyle of the tibia. Nie et al. ⁷ observed in their study of patients with medial compartment knee osteoarthritis (OA) that pre-operatively, there was concurrent loading on both the tibia and fibula, albeit with a greater load borne by the tibia. Following PFO, there was a notable weakening of the lateral fibular support, resulting in a shift towards predominantly tibial loading. Consequently, this redistribution of stresses led to a transition of burden from the medial compartment to the posterolateral compartment of the knee joint.

Current literature on PFO outcomes has been abundant in clinical case series and reports with multiple published systematic reviews and narrative reviews. However, there is still no clarity as to specific indications for offering PFO with regards to the degree of severity of knee OA. Therefore, this systematic review aims to critically evaluate clinical and radiological outcomes of PFO stratified by severity of knee OA.

2. Methods

2.1. Information sources

This study was conducted based on the Cochrane Handbook for Systematic Reviews of Interventions version 6.3.⁸ Reporting of results is in accordance with the Preferred Reporting Items of Systematic Review and Meta-Analysis (PRISMA) statement.⁹

A literature search through PubMed, Scopus, CINAHL and Google Scholar databases were conducted. Eligible studies included those published up till August 2023. The search strategy adopted can be found in Table 1. Reference lists of relevant systematic reviews were manually searched.

Table 1.

Search strategy.

Database	Type of search	Search terms
PubMed database (https://pubmed.ncbi.nlm.nih.gov/) 50 records	Medical subject headings (MeSH) words and general terms	[proximal fibular osteotomy[Title/Abstract] OR proximal fibula osteotomy[Title/Abstract] OR high fibular osteotomy[Title/Abstract]]
SCOPUS database (https://www.scopus.com/search/form.uri?display=basic) 159 records	General terms	TITLE-ABS (proximal AND fibular AND osteotomy) AND NOT TITLE-ABS (high AND tibial AND osteotomy)
CINAHL database (https://www.ebsco.com/academic-libraries) 67 records	General terms	TI proximal fibular osteotomy OR AB proximal fibular osteotomy

The search workflow adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines⁸ and is illustrated in Fig. 1. To identify studies to be included in the final review, the articles were independently assessed by authors to determine eligibility for inclusion in the analysis. Any disagreements were resolved by consensus discussion among the authors.

Fig. 1.

PRISMA flowchart.

A total of 271 studies were obtained from database screening. After removal of duplicates, title abstract screening, and a full text screen based on the inclusion and exclusion criteria, 11 papers were included in the final study from database screening. Additionally, a further 46 papers were identified from snowballing of cited papers from 7 existing systematic reviews was done. After removal of duplicates, title abstract screening, and a full text screen based on the inclusion and exclusion criteria, 2 additional papers were identified and included in the study from snowballing.

Studies included were mainly from China and India and included studies were published between 2017 and 2022. There were 11 prospective studies^7,10, 11, 12, 13, 14, 15, 16, 17, 18, 19 and 2 retrospective studies.^20,21 This comprised of 1 RCT¹¹ and 12 non-randomised studies.^7,10,12, 13, 14, 15, 16, 17, 18, 19, 20, 21

2.2. Eligibility criteria and study selection criteria

Inclusion and exclusion criteria were defined based on the Population, Intervention, Comparison, Outcomes and Study (PICOS) framework and are listed in Table 2.²²

Table 2.

PICOS inclusion and exclusion criteria.

	Inclusion criteria	Exclusion criteria
Population	Patients diagnosed with Knee Osteoarthritis >18 years old Any Kallgren-Lawrence Grading	Patients with - Inflammatory Arthritis - Tibial plateau fractures - Other traumatic fractures or injuries - Other prior surgery to the knee (besides ligamentous repair) - Spontaneous osteonecrosis of the knee
Interventions	Proximal fibular osteotomy	High tibial osteotomy Total knee arthroplasty Unicompartmental knee arthroplasty
Comparison	–	–
Outcome measures	Visual analogue score (VAS) Femur–tibia angle (FTA) or Hip Knee Angle (HKA) Complications of PFO surgery	–
Study Design	Randomized controlled trials Observational studies (prospective and retrospective) Case Series Comparison studies with other surgical procedures (eg. UKA, HTO)	Cadaveric studies Biomechanical studies Case reports Systematic review

2.3. Data collection and statistical analysis

Data extraction was conducted in accordance to the Cochrane Handbook for Systematic Reviews of Interventions version 6.3.⁸ This includes summarizing the data in a pre-specified format of: (1) author (2) country of publication (3) year of publication (4) study characteristics including the number of knees/patients analysed, mean follow up period, clinical outcome measures, radiological outcome measures, complications reported. Statistical analysis was done using RevMan 5.4.1.²³ P values were obtained by performing unpaired t-tests.

2.4. Risk of bias assessment

The Cochrane Risk of Bias (ROB) Tool 2 was used to stratify RCT Risk of Bias, and the ROBIN-I Tool was used to stratify risk of bias in non-randomised studies of intervention as shown in Table 3.

Table 3.

Summary of bias in included trials.

	Cochrane Risk of Bias 2 Table for Randomised Trials
	Randomisation process	Allocation Concealment	Blinding of participants and personnel	Blinding of outcome assessment	Incomplete outcome data	Selective reporting	Other bias
Chen et al., 2019

ROBINS-I tool for assessing risk of bias in non-randomised studies
	Bias due to confounding	Bias in selection of participants into the study	Bias in classification of interventions	Bias due to deviations from intended interventions	Bias due to missing data	Bias in measurement of the outcome	Bias in selection of reported result	Overall risk of bias
Atlıhan et al., 2022	Probably no	Probably yes	No	Probably no	Probably no	No	Probably yes	Moderate
Deng et al., 202120	Probably yes	Probably yes	No	Probably no	Probably no	Probably no	Probably no	Moderate
Huda et al., 202011	Probably yes	Probably yes	Probably no	Probably no	Probably no	Probably yes	Probably no	Serious
Kelany et al., 202212	Probably yes	Probably yes	No	Probably yes	Probably no	Probably no	Probably no	Serious
Kumar et al., 202113	Probably no	Probably no	Probably no	Probably no	Probably no	Probably yes	Probably no	Moderate
Misra et al., 201921	No	Probably yes	No	Probably no	Probably yes	No	Probably no	Serious
Mongia et al., 202214	Probably no	Probably no	No	Probably no	Probably no	Probably yes	Probably no	Moderate
Nie et al., 201815	No	Probably no	Probably no	Probably no	Yes	Probably no	Probably yes	Serious
Qin et al., 201816	Probably yes	No	Probably no	Probably yes	Probably no	Probably no	Probably no	Serious
Sabir et al., 202117	Probably yes	Probably no	Probably no	No	Probably no	Probably yes	Probably yes	Serious
Utomo et al., 201819	Probably yes	Probably no	Probably yes	Probably no	Probably no	Probably no	Probably yes	Serious
Zou et al., 201718	Probably no	No	Probably no	No	No	Probably no	Proabbly yes	Moderate

2.5. Quality assessment of studies

The Grading of Recommendations Assessment, Development and Evaluation (GRADE) system was used to assess the quality of the studies. The five domains of risk of bias, imprecision, inconsistency, indirectness, and publication bias were ranked. These were ranked based on our take on the certainty of evidence, with high being the most certain and very low being the least certain, as seen in Table 4.

Table 4.

Grade system evaluation for quality of evidence.

	Risk of bias	Imprecision	Inconsistency	Indirectness	Publication bias
Primary outcome: clinical outcomes (VAS) – 10 studies	Moderate	Moderate	Moderate	Moderate	Moderate
Primary outcome: radiological outcomes (FTA) – 7 studies	Low	Moderate	Low	Moderate	Moderate
Secondary outcome: complications – 4 papers	Moderate	Low	Low	Moderate	Moderate

3. Results

The 13 articles included in this systematic review included a total of 788 knees analysed. Characteristics of included studies are presented in Table 5.

Table 5.

Characteristics of included studies.

	Number of Knees	Indications for PFO based on KL grading	Group	Reported Primary Outcomes measures	Reported Secondary outcome measures
Atlihan et al., 20229	35	Grade 2 = 10 Grade 3 = 18 Grade 4 = 7	Including KL4 (liberal)	VAS	Complications reported
Chen et al., 201910	49	Grade 1 = 6 Grade 2 = 33 Grade 3 = 10	Excluding KL4 (restricted)	VAS	None
Deng et al., 202120	51	Grade 2 = 26 Grade 3 = 25	Excluding KL4 (restricted)	VAS, FTA	None
Huda et al., 202011	42	Grade 2/3 = 42	Excluding KL4 (restricted)	VAS, FTA	Complications reported
Kelany et al., 202212	24	Grade 2 = 6 Grade 3 = 15 Grade 4 = 3	Including KL4 (liberal)	VAS, FTA	None
Kumar et al., 202113	41	Grade 1 = 2 Grade 2 = 22 Grade 3 = 14 Grade 4 = 3	Including KL4 (liberal)	VAS	None
Misra et al., 201921	38	Grade 2 = 38	Grade 2 only (strict)	VAS	None
Mongia et al., 202214	60	Grade 2 = 60	Grade 2 only (strict)	VAS, FTA	None
Nie et al., 201815	16	Grade 2 = 3 Grade 3 = 16	Excluding KL4 (restricted)	VAS	None
Qin et al., 201816	67	Any KL score	Including KL4 (liberal)	VAS	None
Sabir et al., 202117	47	Grade 2 = 29 Grade 3 = 18	Excluding KL4 (restricted)	VAS, FTA	Complications reported
Utomo et al., 201819	15	Grade 4	Including KL4 (liberal)	FTA	None
Zou et al., 201718	40	Grade 1,2 = 30 Grade 3,4 = 10	Including KL4 (liberal)	VAS, FTA	Complications reported

3.1. Outcomes of pooled analysis of VAS

Studies were split by indications into three different groups: (1) Grade 2 only (strict), (2) Excluding KL Grade 4 (restricted), (3) Including KL Grade 4 (liberal). Within each group, pre-operative and post-operative VAS was analysed. Studies with radiological outcomes (FTA) were also separately analysed with unpaired T test results.

3.1.1. GROUP 1 – Grade 2 only (strict)

Studies included in this group operated on patients which only had KL Grade 2 osteoarthritis. 2 studies were included in this group – Misra et al., 2019 and Mongia et al. 2022.^15,21 Misra et al. reported statistically significant reduction in VAS score 6 months post-operatively from 7.89 ± 1.01 points to 2.74 ± 2.34 (P < 0.0001).²¹ Similarly, Mongia et al. reported improvement in VAS score from 8.65 ± 0.75 to 2.30 ± 0.92 at 12 month follow-up (P < 0.001).¹⁵

3.1.2. GROUP 2 – excluding KL grade 4 (restricted)

Studies included in this group assessed patients with KL Grade 1 to 3 medial knee OA without any KL grade 4 knees. 5 studies were included in this group.^{7,11,12,17,20} Pre-operative mean VAS and post-operative 6-month VAS were compared in 4 studies as seen in Table 6. Huda et al., 2020¹² did not provide standard deviation data while Deng et al. also had their study split into two different groups (curvature and non-curvature group) for comparison, however all participants underwent similar surgical procedures and follow up after PFO, thus having two separate sets of data. A total of 5 studies with 6 datasets comprising of 219 patients analysed the preoperative and postoperative VAS scores as a primary outcome.

Table 6.

Pre-operative and post-operative VAS for group 2 (KL excluding grade 4).

Study		Pre-Operative VAS		Post-Operative VAS
	Number of Knees	Mean	SD	Mean	SD	P-value	95 % CI
Chen 2019	49	5.2	0.96	0.86	0.54	<0.0001	4.64 - 6.41
Deng 2021 Curvature Group	28	7.25	1.91	2.43	0.88	<0.0001	2.39 = 4.00
Deng 2021 Non-Curvature Group	23	6.91	2.08	2.833.62	0.94	<0.0001	1.700–3.27
Huda 2020	56	8.3	0	2.83	0	NA	NA
Nie 2015	16	5.64	1.29	0.27	0.65	<0.0001	3.62–6.63
Sabir 2021	47	7.33	0.72	7.13	1.64	0.45	−0.25, 0.56

3.1.3. GROUP 3 – Including KL grade 4 (liberal)

Studies included in this group assessed patients who had KL Grades including Grade 4 medial knee OA. 6 studies were included in this group.^{10,13,14,16,18,19} Pre-operative mean VAS and post-operative 1 year VAS was analysed Atlihan et al., 2022¹⁰ did not provide standard deviation values or post-operative 1 year mean VAS data and thus was not included. Atlihan et al., 2022¹⁰ did not provide standard deviation values and thus was not included. Utomo et al., 2018¹⁹ did not study VAS as an outcome and thus was not included.

A total of 6 studies comprising 222 patients were included in this group. However, only 4 studies comprising 172 patients had analysed the preoperative and postoperative VAS scores as a primary outcome. The data is represented in Table 7.

Table 7.

Pre-operative and post-operative VAS for group 3 (KL including grade 4).

Study		Pre-Operative VAS		Post-Operative VAS
	Number of Knees	Mean	SD	Mean	SD	P-value	95 % CI
Atlihan 2022	35	6.76	0	2.09	0	NA	NA
Kelany 2022	24	8.4	1.6	1.5	0.33	<0.0001	4.52–7.23
Kumar 2021	41	7.86	0.66	3.78	1.26	<0.0001	3.25–4.78
Qin 2018	67	6.03	1.45	2.53	1.48	<0.0001	1.93–2.82
Utomo 2018	15	0	0	0	0	NA	NA
Zou 2017	40	4.6	1.3	0.5	0.2	<0.0001	3.55–5.19

3.2. Outcomes of pooled analysis of FTA

3.2.1. GROUP 1 – Grade 2 only (strict)

Mongia et al. assessed change in post-operative FTA, showing statistically significant change from 183.95 ± 1.96 pre-operatively to 177.95 ± 2.8 at 12 month follow-up (P < 0.001).¹⁵

3.2.2. GROUP 2 – excluding KL grade 4 (restricted)

Mean pre-operative and post-operative FTA were analysed for Deng et al., Huda et al., and Sabir et al.^12,17,20 which reported FTA data (Table 8). Both arms of the Deng et al., 2021 study reported statistically significant changes in FTA. Huda et al., 2020 had reported statistically significant changes in FTA but no standard deviation data could be obtained to verify the claim. Sabir et al., 2021 did not report statistically significant changes in FTA.

Table 8.

Mean FTA analysis for Group 2 - KL Grading not including Grade 4.

	Total number of knees	Mean Pre-op FTA	SD Pre-op FTA	Mean Post-op FTA	SD Post-op FTA	P value	95 % CI
Deng et al., 2021 Curvature Group	28	181.14	2.03	177.18	1.52	<0.0001	2.9991–4.9209
Deng et al., 2021 Non-Curvature Group	23	180.83	2.42	178.35	1.86	0.0003	1.1974–3.7626
Huda et al., 2020	56	178	0	177.3	0	–	–
Sabir et al., 2021	47	185.6	2.53	184.8	1.97	0.0906	0.1289–1.7289

3.2.3. GROUP 3 – Including KL grade 4 (liberal)

Mean pre-operative and post-operative FTA were analysed for Atlihan et al., Kelany et al., and Utomo et al.^10,13,19 which reported FTA data. Kelany et al., 2022 and Utomo et al., 2018 found statistically significant changes in post-operative FTA while Atlihan et al., 2022 did not find statistically significant changes in post-operative FTA (Table 9).

Table 9.

Mean FTA outcomes for KL Grading including Grade 4.

	Total number of knees	Mean Pre-op FTA	SD Pre-op FTA	Mean Post-op FTA	SD Post-op FTA	P value	95 % CI
Atlıhan et al., 2022	35	178	3.3	178.6	3.2	0.4427	2.150 to 0.950
Kelany et al., 2022	24	183.6	1	180.3	1.6	<0.0001	2.525 to 4.075
Utomo et al., 2018	15	181.28	1.88	177.33	2.06	<0.0001	2.4750 to 5.4250

3.3. Complications reported

4 studies^10,12,17,18 reported the post-operative complications of PFO and are listed in Table 10. All complications were reported to have resolved spontaneously without long-term effects. The most common post-operative complications were paresthesia of the foot secondary to deep peroneal nerve injury.^10,12,17

Table 10.

Reported complications of studies.

	Number of Knees	Indications for PFO based on KL grading	Reported Complications
Atlihan et al., 20229	35	Grade 2 = 10 Grade 3 = 18 Grade 4 = 7	DPN Palsy = 8 Post-op ankle pain = 4
Huda et al., 202011	42	Grade 2/3 = 42	Dorsum foot numbness = 5
Sabir et al., 202117	47	Grade 2 = 29 Grade 3 = 18	EHL weakness = 5 Dorsum foot numbness = 7
Zou et al., 201718	40	Grade 1,2 = 30 Grade 3,4 = 10	Neurovascular injury = 1 Fracture = 1 Recurrent deformity = 1

4. Discussion

There has been emerging evidence of PFO as an alternative to HTO in the setting of varus knee OA. A systematic review and meta-analysis comparing PFO and HTO done by Wu et al., 2022 found that PFO and HTO have similar short-term efficacy, but PFO has had reduced operative time, intraoperative bleeding, hospital stay, and postoperative complications.²⁴ There has however been no clear consensus for indications for PFO in terms of severity of medial joint knee OA thus warranting the need for this paper.

We found that regardless of severity of OA, there was significant improvement in the VAS score in all groups. This improvement met the minimum clinically important difference (MCID) criteria for VAS score of −1.99 cm as established by Tubach et al., 2005.²⁵ Such improvements in VAS score have been explained in multiple biochemical studies. Huang et al.²⁶ posit that post-PFO, there is a competition of muscles between the peroneus and biceps femoris. Post-PFO, there is an increase in muscle activity in the long head of the biceps femoris and a decrease in peroneus longus activity, explaining the immediate pain relief through rectification of varus malalignment.

Xie et al. ²⁷ explain the immediate improvement in joint pain through analysing biochemical changes in ground reaction vector action at the foot level. In patients with varus malalignment and knee osteoarthritis, the hindfoot goes into valgus during the phase, moving the origin of the GRV laterally and closer to the centre of the knee, thus reducing knee adduction moment. Xie et al. propose that post-PFO, the lateral malleolus migrates proximally, pulling the calcaneus into valgus through the calcaneofibular ligament allowing for a more laterally directed GRV, reducing medial knee pressures allowing for pain relief post-PFO. However, Guo et al.²⁸ disproved the theory in a prospective study of 49 patients, being unable to account for any proximal migration of the lateral malleolus or valgus alignment at the ankle post-operatively.

One of the major questions about PFO is how it might improve the mechanical alignment of the lower limb. We found that there was improvement in the FTA radiographically in all studies, regardless of the OA grade. Postulated mechanisms for why PFO improves mechanical alignment have been studied with one such theory being that of non-uniform settlement. This refers to the medial tibia plateau being of lower bone density relative to the fibula thus causing the medial aspect to “settle” relative to the lateral aspect.²⁹ In a PFO, the continuity of the fibula is disrupted allowing for the load on the lateral column to “settle”. This in turn reduces varus misalignment and the progression of OA.³⁰

It is important to consider the advantages of PFO in light of its common complications. From our study, the most common complication was that of deep peroneal nerve palsy. This was reported in included studies as either motor loss in terms of extensor hallucis longus weakness or sensory loss in terms of dorsum foot numbness. There was, however, spontaneous resolution of symptoms within 6 months.^10,12,17

Although good clinical outcomes have been reported for PFO, based on reported post-operative complications, there has been a variation in the percentage of nerve palsy. This is given the anatomical relationship between the fibular head and the common peroneal nerve.³¹ Intraoperative soft tissue retraction and pull by retractors may damage nerves and vessels. The muscle branch for the extensor hallucis longus muscle is particularly at risk during the fibular osteotomy since it runs directly on the bone.³² Thus, senior authors believe that complications can be avoided by gentle retraction of the muscle without the placement of retractors along the posterior aspect of the fibular head.

The findings discussed in this review should be carefully considered alongside the limitations of the current literature. These include the small number of studies available and the high risk of bias in some of the included studies. Furthermore, there is a lack of randomized controlled trials on the subject (only 1 included in the literature). Moreover, it is essential to note that all studies except for Chen et al. (2019)¹¹ are non-randomized and may exhibit bias towards promoting PFO. This highlights the necessity for more randomized controlled trials in the field. Additionally, evaluating pain using the VAS score may not be uniform along the entire VAS given that it is a subjective scale. Further research is required in understanding the impact of PFO and a clear consensus of indications should ideally be identified to ensure optimal patient outcomes.

5. Conclusion

Indications based on KL grading of OA do not seem to differ in terms of post-operative clinical outcomes (VAS score). Some studies which had accounted for radiological measures also found that hip knee alignment was improved regardless of KL grading of OA. Given the outcomes found from this review, PFO is a viable knee joint preserving surgery for medial compartment knee OA, however given the high risk for complications reported in the literature, surgeons should pay close attention to the neuroanatomical landmarks and techniques to avoid neurovascular injury.

Author contribution

LZJ performed the literature review, data collection and analysis. LZJ and HRBAZ analysed and interpreted the data and was involved in manuscript writing and revision. DTSK, JS, and LKH were involved in the critical appraisal and feedback regarding the manuscript. All authors read and approved the final manuscript.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.