2023 ISAKOS Biennial Congress ePoster
The Statistical Fragility of Tourniquet Use In Total Knee Arthroplasty: A Systematic Review
Kyle William Lawrence, BS UNITED STATES
John K Cordero, MD, New York, New York UNITED STATES
Ashley Brown, BS, New York, New York UNITED STATES
Xinning Li, MD, Boston, Massachusetts UNITED STATES
Brett Hayden, MD, New York, NY UNITED STATES
Robert L Parisien, MD, New York, NY UNITED STATES
Icahn School of Medicine at Mount Sinai, New York, New York, UNITED STATES
FDA Status Not Applicable
Summary
Statistical fragility exists in RCTs on tourniquet use in TKA, especially for statistically significant results. Altering a small number of outcomes is sufficient to reverse study findings. Reporting FI and FQ values may provide a more reliable representation of study conclusions in RCTs in TKA.
ePosters will be available shortly before Congress
Abstract
Background
Fragility index (FI) and reverse fragility index (reverse FI) are informative metrics for evaluating the robustness of clinical trial data, representing the number of outcome event reversals necessary to change the significance of a study.
Methods
We queried PubMed/EMBASE/MEDLINE for randomized controlled trials (RCTs) comparing outcomes in total knee arthroplasty (TKA) where study groups were stratified by tourniquet-use strategy. FI and reverse FI were determined – for significant and non-significant outcomes, respectively – by determining the number of outcome reversals required to change statistical significance. The fragility quotient (FQ) was calculated by dividing the FI or reverse FI by the sample size. Subgroup analysis was performed by outcome type.
Results
Literature search yielded 23 studies from 18 orthopaedic journals reporting a total of 91 dichotomous outcomes. The median FI of the 91 outcome events was 4 (IQR 3-6) with an FQ of 0.0476 (IQR 0.0291-0.0867). 11 outcomes were reported as statistically significant with a median FI of 2 (IQR 1.5-5) and a median FQ of 0.0200 (IQR 0.0148-0.0484). 80 outcomes were reported as non-significant with a median reverse FI of 4 (IQR 3-6) and a median FQ of 0.0495 (IQR 0.0310-0.0887). Outcomes were categorized into skin/wound complications, transfusion requirement, deep vein thrombosis, postoperative pain, and re-intervention, with FIs of 4 (IQR 3-6), 3 (IQR 2-5), 4 (IQR 4-6), 5.5 (IQR 5-7.5), and 3 (IQR 3-4.5), respectively. Total patients lost to follow-up was greater than the median FI in 17.6% of outcomes.
Conclusions
Statistical fragility exists in RCTs on tourniquet use in TKA, especially for statistically significant results. Altering a small number of outcomes is sufficient to reverse study findings. Reporting FI and FQ values may provide a more reliable representation of study conclusions in RCTs on tourniquet use in TKA.