Editor's Choice – Comparison of the Reproducibility of Ultrasound Calliper Placement Methods in Abdominal Aortic Diameter Measurements: A Systematic Review and Meta-Analysis of Diagnostic Test Accuracy Studies

Objective: To assess which ultrasound (US) method of maximum anteroposterior (AP) abdominal aortic diameter measurement can be considered most reproducible. Data sources: MEDLINE, Scopus, and Web of Science were searched (PROSPERO ID: 276694). Eligible studies reported intra- and or interobserver agreement according to Bland–Altman analysis (mean ± standard deviation [SD]) for abdominal aortic diameter AP US evaluations with an outer to outer (OTO), inner to inner (ITI), and or leading edge to leading edge (LELE) calliper placement. Review Methods: The Preferred Reporting Items for a Systematic Review and Meta-Analysis of Diagnostic Test Accuracy Studies statement was followed. The QUADAS-2 tool and QUADAS-C extension were used for risk of bias assessment and the GRADE framework to rate the certainty of evidence. Pooled estimates (fixed effects meta-analysis, after a test of homogeneity of means) for each US method were compared with pairwise one sided t tests. Sensitivity analyses (for studies published in 2010 or later) and meta-regression were also performed. Results: 21 studies were included in the qualitative analysis. Twelve were eligible for quantitative analysis. Studies showed heterogeneity in the US model and transducer used, sex of participants, and observer professions, expertise, and training. Included studies shared a common mean for each US method (OTO: p = 1.0, ITI: p = 1.0, and LELE: p = 1.0). A pooled estimate of interobserver reproducibility for each US method was obtained, combining the mean ± SD (Bland–Altman analysis) from each study: OTO: 0.182 ± 0.440; ITI: 0.170 ± 0.554; and LELE: 0.437 ± 0.419. There were no statistically significant differences between the methods (OTO vs. ITI: p = .52, OTO vs. LELE: p = .069, ITI vs. LELE: p = .17). Considering studies published in 2010 and later, the pooled estimate for LELE was the smallest, without statistically significant differences between the methods. Despite the low risk of bias, the certainty of the evidence for both meta-analysed outcomes remained low. Conclusion: The interobserver reproducibility for OTO and ITI was 2.5 times smaller (indicating better reproducibility) than LELE; however, without statistically significant differences between the methods and low GRADE evidence certainty. Additional data are needed to validate these findings, while inherent differences between the methods need to be emphasised.