Inappropriate grouping of various kinematic alignment techniques and inconsistent reporting of adverse events invalidate comparison across studies: An umbrella review of meta-analyses.

Abstract

Purpose: This umbrella review aimed to identify, synthesise and critically appraise the findings of meta-analyses that compare adverse events-rates of complications, reoperations and revisions-following total knee arthroplasty (TKA) using unrestricted kinematic alignment versus mechanical alignment.

Methods: Using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses criteria, two authors independently screened articles based on inclusion and exclusion criteria, and assessed the methodological quality based on the 16 domains of A MeaSurement Tool to Assess systematic Reviews (AMSTAR-2). Effect sizes of difference in rates of complications were tabulated for each meta-analysis. Studies included in the meta-analyses were assessed to determine if they were on true unrestricted kinematic alignment. A secondary meta-analysis was performed, excluding studies on restricted kinematic alignment techniques, to calculate pooled estimates of adverse events (odds ratio [OR] with its 95% confidence interval [CI]) in a common effects framework with inverse-variance weighting.

Results: Of 78 potential records, 13 meta-analyses were eligible for data extraction, which pooled data from 15 clinical studies (10 on unrestricted kinematic alignment, four on restricted kinematic alignment and one on inverse kinematic alignment). None of the meta-analyses fulfilled all seven critical AMSTAR-2 domains. Meta-analyses categorised adverse events differently and used different measures for the effect sizes but revealed no differences between kinematic versus mechanical alignment. Exclusion of studies on restricted kinematic alignment techniques reduced total sample sizes for kinematic alignment from 658 to 318 and for mechanical alignment from 811 to 403. Secondary meta-analyses exclusively on unrestricted kinematic alignment revealed no difference in complications without reoperation, reoperation without implant removal or reoperation with implant removal following kinematic versus mechanical alignment.

Conclusion: Meta-analyses do not distinguish between various kinematic alignment techniques, and adverse events are compared using different metrics. Surgeons, researchers and editors should refrain from pooling data on various kinematic alignment techniques, and orthopaedic societies should promote standards for reporting adverse events and effect sizes to facilitate comparisons across future studies.

Level of evidence: Level III.