Common peroneal nerve palsy after TKA in valgus deformities; a systematic review

Purpose The aim of this systematic review is to investigate the prevalence of Common Peroneal Nerve Palsy after total knee arthroplasty in valgus deformities. Furthermore, the effectiveness of a peroneal nerve release prior to arthroplasty to prevent the palsy will be investigated. Methods PubMed and Google Scholar were searched. Search terms regarding valgus deformity and total knee arthroplasty were used. Data analysis and extraction were performed using the web application ‘Rayyan QCRI’ according to PRISMA guidelines and screened according to the inclusion and exclusion criteria. Results Twenty-seven studies were included, representing 1397 valgus knees. Knee balancing was performed in 19 studies with lateral soft tissue releases (1164 knees) and 8 studies (233 knees) with an additional osteotomy. Two studies (41 knees) in the lateral soft tissue release group conducted a peroneal nerve release simultaneous to arthroplasty. Common peroneal nerve palsies occurred in 26 cases (1.9%). Overall, no significant difference in palsy ratio between studies was found by using a peroneal nerve release (p = 0.90), between lateral soft tissue releases and osteotomies (p = 0.11) or between releases of specific ligaments. Conclusion Common peroneal nerve palsies occur in 1.9% of the cases after total knee arthroplasty in valgus deformities. No difference in the number of palsies was seen when using a peroneal nerve release or using different balancing techniques. However, literature about peroneal nerve releases was very limited, therefore, the effectiveness of a peroneal nerve release remains unclear. Level of evidence LEVEL III: Systematic review. Supplementary Information The online version contains supplementary material available at 10.1186/s40634-021-00443-x.

Recovery from CPNP usually take place within a year; however, residual damage is certainly not uncommon [1,4,5,12]. As CPNP has serious consequences, orthopaedic surgeons aim to prevent this complication by a concomitant peroneal nerve release (PNR) [8,40]. A PNR is a procedure performed simultaneously to TKA-V, which explores the nerve and removes the constricted dressings to release the CPN. Therefore, it yields the nerve to have more capacity to extend and protects it against mechanical stretching after balancing the knee properly during TKA-V. Due to the limited number of studies investigating PNR, no consensus has yet been reached on the value and indication of the procedure.
This systematic review primarily attempts to investigate the CPNP incidence after TKA-V and the rate will be compared between different valgus correcting techniques, including lateral STR and OT. Secondarily, the effectiveness of a PNR in preventing CPNP after TKA-V will be investigated.

Search strategy
A librarian-assisted comprehensive search of the literature was performed in October 2020 in PubMed and Google Scholar. The primary search was mainly focused on the surgical treatment and outcome of TKA-V. Search terms and associated synonyms that were included in the search are displayed in Table 1. A total of 3.945 articles were identified through PubMed and Google Scholar ( Fig. 1 [Additional file 1]). The analysis was done according to the 'Preferred Reporting Items of Systematic review and Meta-analyses' (PRISMA) [41]. Through the web application 'Rayyan QCRI' [42], duplicates were removed and the remaining articles were screened for eligibility, according to the screening criteria ( Table 2). The screening was independently done by three reviewers (X, X and X), a fourth reviewer (X) was consulted in case of doubt about the suitability of an article [43]. To ensure no relevant articles were omitted, a cross-reference check was performed on the included articles. A consensus was achieved on all included articles based on inclusion and exclusion criteria.

Methodological quality assessment
The quality of the non-randomized studies was assessed by the first author (X), utilizing the Methodological Index for Non-Randomized Studies (MINORS) [44]. In case of any doubt, the second reviewer (X) was consulted to determine the quality of the study. The outcome of the index per study is stated in Table 3.

Statistical analysis
Means and standard deviations were presented and calculated. Reported medians and ranges were transformed into weighted means and estimated standard deviation by the methods of Hozo et al. [45] and Walter et al. [46]. Heterogeneity was assessed using I 2 and χ 2 -tests, where an I 2 of < 25% is considered low; 25-50% as moderate; > 50% as strong and > 75% as substantially heterogeneous by the methods Higgins et al. [24] In case of substantial heterogeneity between studies (I 2 > 75%), a qualitative/narrative data extraction was performed [24]. As the heterogeneity of the CPNP incidence was not substantial among studies, data were pooled using a fixedeffect model and weighted on sample size. Because of substantial heterogeneity between perioperative continuous outcomes (alignments), these outcomes were qualitatively/narratively described. Pooled CPNP rates were log-transformed to calculate 95% Confidence Intervals. Chi2 tests were performed to assess differences between sub-groups. A p-value < 0.05 was considered statistically significant. The Analysis was conducted using R version 4.0.2 (R Foundation statistical computing, Vienna, Austria) with "Metafor package" (Maastricht University, Maastricht, Netherlands).

Quality of the studies
One randomized controlled trial [33] and 26 non-randomized studies were included. The non-randomized studies consisted of 11 prospective and 15 retrospective cohorts. To estimate the risk of bias of the non-randomized studies, the MINORS criteria were calculated. Two comparative studies had a mean MINOR score of 16.5 (range 16-17) out of 24. The other 24 studies were non-comparative studies with a mean of 10.7 (range, 9-13) out of 16. Only 3 studies (11.5%) reported a prospective calculation of the sample size.      [31] was responsible for the only CPNP case in the entire OT group (n = 233) but was also the single study that used no additional lateral STR and has let the piece of the OT healed in situ without use of internal fixation. Three studies did not explicitly describe CPNP cases in their complication section, therefore it was assumed that CPNP did not occur in those studies [35,37,48]. No study clarified if a CPNP case was developed from a patient with posttraumatic OA. Nevertheless, only two studies had patients (11 knees) with posttraumatic OA, where also CPNP cases (7 knees) were reported [15,16]. Information on whether these CPNP cases occurred in one of these 11 knees was lacking.

Soft tissue releases
In the 19 studies that solely performed a lateral STRs, a large variation in released ligaments was present compared with the studies that performed an OT, that mainly released the ITB and PLC. A single study in the OT group performed a POP release [35]. A sub-analysis was performed to approximate the difference in CPNP rate between different specifically released ligaments (Table 5). Between the releases of different ligaments or the manner of those releases (pie-crusting, subperiosteal or transverse), no significant difference in CPNP rate was found. Only studies that reported that all patients underwent a release of a specific ligament were included for sub-analysis. One study was excluded from any analysis due to a lack of data [40].

Pre and postoperative alignments
The overall weighted mean pre-and postoperative aFTAs was 19.5 ± 8.4 and 5.3 ± 2.7 degrees ( Table 6). The PNR group was the only group with a considerable larger weighted mean preoperative aFTA (30. 1 ± 1.3), and therefore, a larger Valgus Correction Angle (VCA) (25.0°). All weighted postoperative aFTAs were comparable between the different groups. Due to a high heterogeneity between studies (I 2 > 0.80), no statistical analysis could be performed. Regarding to the CPNP cases, 4 studies reported the individual preoperative aFTA of 4 CPNP cases, which were 19° [49], 25° [47], 26° [31] and 38° [40]. Flexion contracture angles were reported in 11 studies (41%), with an overall weighted mean of 4.1 ± 4.8°. No individual flexion contractures of CPNP cases were reported in the studies.

Discussion
In this systematic review, the most important finding was the overall CPNP ratio of 1.9% after TKA-V. No significant differences in CPNP rate were found between TKA-V with and without PNR (2.4% vs 2.1%), between TKA-V with lateral STR (2.2%) or with OT (0.4%) and between the releases of different ligaments or the manner of those releases (pie-crusting, subperiosteal or transverse). The obtained overall CPNP ratio of 1.9% in this study falls within the known range of TKA-V (0.3% -9.5%) [3,[9][10][11][12][13][14][15][16][17]. Other systematic review reported a range of 0.01% to 4.3%, like the one of Carender et al. [2] and Rodríguez-Merchán et al. [50]. Currently, controversy still exists related to valgus deformity being a predisposing factor of CPNP. Studies that have investigated the location of the CPN, indicate that the CPN can be jeopardized by a direct injury due to pie-crusting or a transverse release of the ITB or PLC in well-aligned and valgus knees [19,20,51]. However, in our review, we could not confirm the increase of risk to injure the CPN by different ligament releases. Therefore, our results may support the theory that most CPNPs probably occur due to postoperative mechanical damage, like traction and compression, instead of a direct injury. Besides, one large registrybased study by Christ et al. [3], including 383,060 primary TKA procedures, found that preoperative valgus alignments increase the risk of developing a CPNP significantly (OR 4.19). Also, Idusuyi et al. [4], found a relative risk of CPNP 12 times greater for patients with a 12° or more valgus deformity. Both studies did not find an association between CPNP and flexion contractures. However, according to Christ et al. [3], this may be because the diagnosis code for flexion contractures is not consistently noted as that of valgus deformities in their registry. Therefore, the data may be biased. Other studies, like Park et al. [12] and Schinsky et al. [52] found an overall incidence of 0.53% and 1.3% but did not find any relation between valgus deformities and CPNP. However, all these studies used mixed preoperative alignments. Therefore, it is difficult to compare the CPNP ratio of this systematic review, with the incidence of other reviews or studies. Eventually, knowing that larger previous studies showed an increase in CPNP incidence in valgus knees, we would advise clinicians to perform a TKA-V with extra care. This would enable PNR as an option for severe valgus deformities since the procedure is minimally invasive and may lead to preventing CPNP. However, this current review did not find a significant difference in CPNP incidence between the studies that performed a TKA-V with and without PNR. Regarding the 2 studies utilizing a PNR, the study of Cree et al. [40] is a small retrospective study and the recent study of Xu et al. [8] is a small prospective study, both studies performed the same surgical Studies that performed a specific ligament release on only a part of the total study population were excluded for analysis. One study [40] was excluded for analysis due to lack of data   [31] and 38° [40]) than the overall mean aFTA of all the studies in this review.

Released ligament A Studies (n) Treated knees (%) CPNP cases, (%) P-value
In the end, the results in this review suggest that a PNR procedure is not effective. However, it is difficult to assume such an interpretation because only two small sample sized studies were found that used a PNR prior to TKA and met the inclusion criteria of our systematic review [8,40]. Future research should further investigate PNR in larger study populations and preferably with a comparison group, which would make it easier to interpret results.
Like all studies, some limitations need to be discussed. Firstly, the considerable heterogeneity between the included studies, possibly caused by our caution to minimize selection bias in including studies for this review. However, due to the low incidence of CPNP and the focus on valgus deformities, a comprehensive literature search was needed. Secondly, the review lacks important detailed information about the individual cases who developed a CPNP in the studies. In addition, preoperative data of the knees, like knee extension angles and stress radiographs to assess whether there is a fixed valgus deformity are missing in most studies. Therefore, it is important for future studies to specify the manner and degree of the surgeries and to comprehensively note the pre and postoperative data of the knees. Thirdly, the scarce of studies investigating a PNR is an insurmountable problem, which made it impossible to draw conclusions. However, this review provides a basis for future work investigating PNR in valgus knees to prevent CPNP.

Conclusion
To our knowledge, this is the first systematic review that provides insight into the current literature about preventing CPNP with a PNR after TKA-V. An overall CPNP ratio of 1.9% in valgus knees after TKA was found. There was no direct evidence that using a PNR would be more effective than not using a PNR in preventing a CPNP. However, it was impossible to draw conclusions, due to the scarce amount of literature. Therefore, larger studies comparing TKA-V with and without PNR are needed to appropriately define the efficiency of a PNR. This systematic review is the first step in this regard.