Polisetty Teja, Hohmann Alexandra L, DiDomenico Eric, Lonner Jess H
Department of Orthopaedics Harvard University Cambridge Massachusetts USA.
Rothman Orthopaedic Institute at Thomas Jefferson University Philadelphia Pennsylvania USA.
J Exp Orthop. 2024 Aug 6;11(3):e12113. doi: 10.1002/jeo2.12113. eCollection 2024 Jul.
Fragility analysis is a method of further characterising the robustness of statistical outcomes. This study evaluates the statistical fragility of randomised controlled trials (RCTs) comparing patellar resurfacing versus non-patellar surfacing in total knee arthroplasty (TKA).
PubMed, MEDLINE and EMBASE were searched for RCTs comparing outcomes in TKA based on patellar resurfacing. Fragility index (FI) and reverse FI (collectively, "FI") were calculated for dichotomous outcomes as the number of outcome reversals needed to change statistical significance. Fragility quotient (FQ) was calculated by dividing the FI by the sample size for that outcome. Median FI and FQ were calculated for each individual outcome and for the overall study. Subanalyses were performed to assess FI and FQ based on outcome type, statistical significance and loss to follow-up.
Twenty-one RCTs were included in the analysis, capturing 3910 subjects. The overall median FI was 5.0 (interquartile range, [IQR] 4.0-6.0), and the overall median FQ was 0.048 (IQR 0.022-0.065). The outcome of anterior knee pain has a median FI of 6.0 (IQR 4.0-6.0) and a median FQ of 0.057 (IQR 0.025-0.065). Only five (7%) outcomes were significant. The loss to follow-up was greater than the FI in 12 of 19 studies (63%) with available data.
RCTs comparing patellar resurfacing in TKAs show significant statistical fragility; a few outcome reversals can alter findings. The majority of outcomes were nonsignificant, indicating that the choice to resurface the patella may not affect most clinical outcomes; however, clinical conclusions are limited by the statistical fragility of the analysed outcomes. Larger RCTs for this comparison are necessary, and we suggest adding FI and FQ to RCT reports with values to improve the interpretability of results.
Level II.
脆弱性分析是进一步刻画统计结果稳健性的一种方法。本研究评估了全膝关节置换术（TKA）中比较髌骨表面置换与非髌骨表面置换的随机对照试验（RCT）的统计脆弱性。
检索PubMed、MEDLINE和EMBASE数据库，查找比较基于髌骨表面置换的TKA结局的RCT。对于二分结局，计算脆弱性指数（FI）和反向FI（统称为“FI”），即改变统计学显著性所需的结局反转数。脆弱性商数（FQ）通过将FI除以该结局的样本量来计算。计算每个个体结局和整个研究的中位数FI和FQ。基于结局类型、统计学显著性和失访情况进行亚组分析，以评估FI和FQ。
分析纳入了21项RCT，涵盖3910名受试者。总体中位数FI为5.0（四分位间距，[IQR] 4.0 - 6.0），总体中位数FQ为0.048（IQR 0.022 - 0.065）。膝前疼痛结局的中位数FI为6.0（IQR 4.0 - 6.0），中位数FQ为0.057（IQR 0.025 - 0.065）。只有5个（7%）结局具有显著性。在有可用数据的19项研究中的12项（63%）中，失访人数大于FI。
比较TKA中髌骨表面置换的RCT显示出显著的统计脆弱性；少数结局反转就能改变研究结果。大多数结局无显著性，这表明髌骨表面置换的选择可能不会影响大多数临床结局；然而，临床结论受到所分析结局的统计脆弱性的限制。有必要开展更大规模的此类比较RCT，并且我们建议在RCT报告中加入FI和FQ值，以提高结果的可解释性。
二级。
