The Never-Ending Battle between Proximal Row Carpectomy and Four Corner Arthrodesis: A Systematic Review and Meta-Analysis for the Final Verdict

Please cite this article as: Ali R. Ahmadi , Liron S. Duraku , Mark J.W. van der Oest , Caroline A. Hundepool , Ruud W. Selles , J. Michiel Zuidam , The Never-Ending Battle between Proximal Row Carpectomy and Four Corner Arthrodesis: A Systematic Review and MetaAnalysis for the Final Verdict, Journal of Plastic, Reconstructive & Aesthetic Surgery (2021), doi: https://doi.org/10.1016/j.bjps.2021.09.076


INTRODUCTION
The human wrist is a complex formation with interactions between several small bones and interligamentary connections resulting in a mobile and stable hand function. If left untreated, disruption of this complex anatomical formation can result in degenerative arthritis, pain, and eventually progressive loss of hand function. Degenerative wrist arthritis can arise from rheumatoid arthritis or posttraumatic arthritis affecting the radiocarpal and midcarpal joints of the wrist. The scaphoid bone, in particular, is an important anatomical and junctional link between the proximal and distal carpal bones.
Scapholunate advanced collapse (SLAC) and scaphoid nonunion advanced collapse (SNAC) are the two most common patterns of degenerative wrist arthritis. In 1984, Watson and Ballet were the first to describe SLAC as the most common form of degenerative wrist arthritis 1 .
This condition can be attributed to spontaneous osteoarthritis or post-traumatic injury of the wrist. Because of the abnormal distribution of forces across midcarpal and radiocarpal joints, the radioscaphoid joint is affected, which can progress to the capitolunate joint in further stages ( Figure 1A). SNAC is a condition of advanced collapse and progressive arthritis of the wrist that results from a chronic scaphoid nonunion, mainly originating from traumatic injury ( Figure 1B).
When conservative treatment fails, surgical interventions are indicated. The ideal surgical treatment should result in pain relief and good wrist functionality. Currently, no curative therapies exist; however, many treatment options have been developed for symptomatic relief. Traditionally, end-stage SLAC and SNAC arthritis were treated with total wrist arthrodesis (TWA). Although TWA leads to pain relief in most cases, wrist motion is sacrificed. TWA is currently considered a salvage procedure when other therapies fail.
However, newer alternatives to TWA include the use of motion sparing arthroplasty implants which are more anatomically aligned and require minimal bone resection and offer improved range of motion and grip strength 2 .
Proximal Row Carpectomy (PRC) and Four Corner Arthrodesis (FCA) are the two most common surgical procedures to treat SLAC and SNAC wrists. These treatment modalities are preferred because they address pain and simultaneously preserve ROM. Briefly, PRC is a resection of the proximal carpal row consisting of the scaphoid, lunate and the triquetrum bones. PRC allows the capitate bone to articulate with the lunate facet of the distal radius, creating a new joint formation (Figure 2A). FCA is a resection of the scaphoid bone and arthrodesis between the lunate, capitate, hamate and triquetrum bones. This procedure aims to fuse the arthritic midcarpal joint and reduce pressure in the scaphoid fossa, by redistributing the cartilage contact areas to the lunate fossa. FCA can be performed traditionally with screw fixation ( Figure 2B) or with a circular plate ( Figure 2C).The circular plate is used as an alternative fixation method, preferred by some surgeons in comparison to the traditional compression screws.The rationale for both procedures is based on redistributing the force from the scaphoid fossa to the lunate fossa cartilage of the radius since this cartilage is not affected by the degenerative changes. The lunate fossa of the radius is spared from degenerative changes as it is protected from incongruent pressure by its tight ligamentous structures, that control a near spherical joint profile with the capitate base. However, it is not more resilient against arthritic changes than other structures in the wrist and can feature in advanced stage disease. Indications for the use of each operative procedure are listed in Generally, FCA is essential, if there is midcarpalosteoarthrtitis and the capitolunate joint is affected in the SLAC or SNAC wrist. A possible disadvantage for choosing PRC could be an incongruity between the capitate and the lunate fossa leading to degenerative changes and pain and eventually to TWA.However, controversy persists over which of the two procedures has the superior functional short-term and long-term outcome.

5
Two systematic reviews have been published on this topic in 2009 and 2015 3,4 . Only Mulford et al. conducted a meta-analysis but on subjective outcomes, relative risk of postoperative conversion and post-operative osteoarthritic changes 4 . Our study on the contrary, presents the most up to date systematic review and meta-analysis on the most important postoperative outcome measures to help guide clinicians in patient selection and preferred operative method and patients in their treatment choice.

METHODS
The study was designed according to the Cochrane Handbook for Interventional Systematic Reviews and the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Protocols (PRISMA-P) 5 .

Literature search strategy
Comprehensive searches were carried out in EMBASE, Medline, Pubmed Publisher, Web-of Science, OvidSP, Cochrane Central Register of Controlled Trials (The Cochrane Library, December 2019, issue 12 of 12) and Google Scholar. The search was conducted for articles published up to December 2019 by using search terms specific to each search engine (Text S1).

Inclusion and exclusion criteria
The main inclusion criteria were studies evaluating one or more of the following postoperative outcomes of PRC and/or FCA for the treatment of SLAC and SNAC wrists: range of motion (ROM), grip strength, complications leading to reoperations, conversion to TWA, Visual Analog Scale (VAS) pain or Disability of the Arm, Shoulder, and Hand Questionnaire (DASH) scores. Studies combining limited 2 or 3 corner with 4 corner arthrodesis versus PRC were excluded. Only full-text, original articles written in English were considered for inclusion in the study. All reviews, conference abstracts, book chapters, letters, case series and editorials were excluded.

Literature screening
Articles were screened by two individual researchers (ARA, LSD) for relevance and inclusion. The same independent reviewers screened titles, keywords and abstracts of all considered articles, according to the pre-established criteria. In case of discrepancy, a third author was consulted (JMZ). After inclusion based on the parameters above, full-text articles were retrieved and reviewed for inclusion. All articles were evaluated using the PICO method. All included articles were rated for their level of evidence according to an adapted version from material published by the Centre for Evidence-Based Medicine, Oxford, UK 6 .

Risk of Bias Assessment
Risk of bias assessment for observational cohort and cross-sectional studies and intervention studies was carried out by two authors (ARA, LSD) with the National Institutes of Health (NIH) study quality assessment tool 7 .

Statistical analysis
We performed a meta-analysis on the following parameters: ROM, grip strength, complications, conversions to total arthrodesis, VAS pain score and DASH score. When standard deviations were not reported, we calculated them based on the method described by Hozo et al. 8 . For continuous outcomes, we performed a meta-analysis using a random-effects models to estimate the pooled mean difference and 95% confidence interval (CI). For dichotomous data, we used a random-effects model to estimate the pooled risk difference. We tested whether the type of surgery influenced the pooled means or proportions by adding the type of surgery as an independent variable. The same methodology was used to test if the patient characteristics were the same for both types of surgery.
The Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) guidelines were followed. A p-value lower than 0.05 was deemed statistically significant. All analyses were performed with R version 3.6.0 and the metafor package 9 .

Study demographics
In total, 1619 articles were retrieved and subjected to screening based on the inclusion and exclusion criteria. Only 16 studies fell within the initial scope of our study ( Figure S1). Upon reading the full text, one article was excluded because it only reported outcomes of two and four corner fusions combined. Of the remaining 15 studies, ten were retrospective (Level III), were male and 23% female and in the PRC group, 68% male and 32% female.
Only nine studies reported the number of treated SLAC and SNAC wrists by either PRC or FCA. The majority of the studies had included patients with a grade II/III SLAC/SNAC diagnosis. Of these eight studies, two studies had also included patients with SLAC/SNAC grade I. The remaining seven studies did not report on the SLAC/SNAC grading. All studies but three reported on whether the dominant hand was operated or not. Merely seven studies reported on the number of surgeons that operated, and 10 out of 15 studies mentioned who registered outcome measurements and performed additional testing or collected clinical data.

Risk of Bias
Risk of bias was evaluated with the NIH tool. The majority of studies were at serious risk due to a poorly defined and specified study population and in some cases from two different countries and different institutions, and the lack of specific inclusion and exclusion criteria ( Figure 4). Almost all studies lacked sample size justification, power analysis or proper statistical methods to adjust for confounders. Outcome measurements were not conducted in a blinded setting in any study.

Active ROMoutcomes
In total, 13/15 studies reported on the ROM. Separate extension, flexion, radial deviation, and ulnar deviation values were reported in eight studies. Of those, three studies were excluded from the pooled analysis because of missing SD or statistical analysis 10-12 . In the meta-analysis, we found no significant differences between PRC and FCA in extension Fivestudies were excluded from the meta-analysis because they reported only either a combined extension/flexion arc 13 or a combined radial/ulnar deviation arc 14,15 or only the total arc of motion 16 . Three studies reported on a combined extension/flexion and radial/ulnar deviation arc between PRC and FCA, but no significant differences were observed [13][14][15] . One studies reported on the extension/flexion arc, which was significantly improved in PRC 73 ± 5 compared to FCA 54 ± 6, P < 0.01 17 . Radial and ulnar deviation were measured separately, but were not significantly different. One study only reported on the postoperative ROMwhich. was not significantly different between PRC and FCA 16 .

Grip strength
All studies but two reported on grip strength 18,19 . Ten studies measured grip strength using a Jamar Dynamometer, the other three did not report on the SD and were therefore not included in the meta-analysis 11,12,16 . The three excluded studies did not report a significant difference in grip strength between the two groups. All remaining studies (N = 10) reported grip strength as a percentage compared to the contralateral non-operated hand.

Complications and reoperations
Complications resulting in a reoperation were reported in 12 studies and consisted of pin removal because of migration causing discomfort, carpal tunnel syndrome decompression, tenolysis, hematoma evacuation, Quervain's stenosing tenovaginitis, ulnar impaction and scar revision. Three studies did not report on complications and reoperation 15,18,20 .
The success of FCA partially relies on the successful union between the lunate, capitate, hamate and the triquetrum carpal bones. Adequate union allows for the start of the rehabilitation process, therefore nonunion can delay this process and eventually affect longterm outcomes. Nonunion rates were reported in ten studies, exclusively prevalent in FCA treated patients due to the nature of the operative procedure. In total, 17 (8.8%) cases of nonunion were reported amongst all studies. In some cases, non-union was an indication for reoperation or conversion to TWA.
In the meta-analysis, we found no significant difference in the number of complications resulting in reoperation between the two groups 0.03 (-0.03 -0.10), P = 0.28 (Figure 7).

Conversions to total wrist arthrodesis
Since total wrist arthrodesis is a salvage procedure for failed FCA or PRC, conversion rates were reported in 12 studies, but not reported on in the three remaining studies 15,18,20 .
The reason for conversion was persistent or severe pain after FCA or PRC in the majority of the studies.

Pain
Postoperative pain measurement was conducted in eight studies during follow-up visits. Two studies did not use the VAS pain score or performed statistical analysis 10,11 . The VAS pain score was used in six studies. However, we were only able to pool four of these studies, because of missing standard deviations in two articles which separately did not show a significant difference in postoperative pain score between the two groups 12,13 .

DASH scores
DASH scores were reported in seven studies, of which three studies were excluded from the pooled analysis because of missing SD 12,13,21 .
In the meta-analysis, there was no difference in DASH scores between PRC and FCA treated patients 1.24 (-22.7 -24.54), P = 0.88 (Figure 10). High heterogeneity was noted amongst the studies (P < 0.01, I 2 = 92%). In the last few decades, PRC and FCA have gained popularity and are among the most commonly used procedures to treat degenerative osteoarthritis. Both procedures provide painrelief and preserve ROM; however, because of the lack of RCTs, it remains unclear which procedure provides superior outcomes. We performed a meta-analysis and found no significant difference between ROM, grip strength, reoperation, conversion to TWA, pain, or DASH scores between the two groups. Although nonunion exclusively occurred in the FCA group, this did not affect the reoperation or conversion rate. Only two RCT exist on this topic and our findings are comparable 12,16 . They did not observe any significant difference in functional or clinical results between the two treatment methods.

DISCUSSION
Watson and Ballet were the first to introduce FCA in 1984 1 . According to Watson, FCA is the preferred treatment option for stage 3 collapse, mainly because the proximal pole of the capitate is involved in the degenerative process, therefore excluding PRC as an option. The main question which remains is whether patients with stage 2 degeneration will have better outcomes with PRC or FCA. In our study, we included all of the available literature comparing outcomes between PRC and FCA and pooled this data to generate a more robust conclusion. It was important to include only comparative studies because of their relatively well-matched study populations and similarity in outcome measurements.
Two systematic reviews have been conducted on this topic. Mulford et al. conducted a systematic review comparing outcomes of 160 PRC and 185 FCA procedures 4 . Their findings demonstrate that grip strength, pain relief, ROM, and subjective outcomes are comparable for both groups. They found PRC to have a higher range of osteoarthritic change after surgery, although in most cases asymptomatic, but that the FCA group had more complications (10% more) overall due to nonunion, dorsal impingement and complications related to hardware.
No differences in the rate of conversion to TWA were observed. This is in line with our findings. Important to note is the different surgical methods used when performing FCA. This may partially explain the different complications rate reported in FCA treated patients ranging from 2 to 11% 26       34 Figure 10. DASH Score This figure shows the mean difference between FCA and PRC of the DASH score in each study and the estimated 95% CI of this percentage using a Random Effects Model. The lowest diamond is the pooled percentage of all studies. A negative value indicates that the FCA group had a less ROM than the PRC group.