The Fragility Index of Randomized Controlled Trials for Early Blood Pressure Management of Intracerebral Hemorrhage
Cesar Escamilla-Ocanas1, Noelia Morales1, Rahul Damani2, Eric Bershad2, Mohammad Hirzallah2, Khawja Siddiqui2, Chethan Venkatasubba Rao2
1Department of Neurology, Baylor College of Medicine, Houston Texas, 2Division of Neurovascular and neurocritical care, Department of Neurology, Baylor college of Medicine, Houston Texas
Objective:
We evaluated the outcome stability and robustness of randomized controlled trials (RCTs) focusing on early blood pressure management after intracerebral hemorrhage (ICH) by quantifying a fragility index (FI)
Background:
ICH is frequently complicated by elevated blood pressure, which exacerbates the risk of hematoma expansion and unfavorable outcomes. Despite considerable research, debate persists regarding optimal blood pressure control strategies, likely due to scarcity of statistically significant trials when assessing binary outcomes such as, mortality and major disability. The FI, a measure of clinical trial stability has emerged as a complement for RCT interpretation. In this study, we calculated the FI in RCTs comparing intensive vs standardized blood pressure management after ICH.
Design/Methods:
We conducted a comprehensive review of previously published literature from January 2013 to January 2022, using the PubMed/Medline database. Our study included landmark trials in blood pressure management after ICH that reported dichotomous outcome as primary endpoint. To ensure the robustness of the result in each trial, we recalculated all p-values using a Fisher exact test and calculated the FI. Median scores were subsequently reported.
Results:
We included 4 RCTs, median (IQR) sample size was 290.5 (84.25-706.25) participants, and median (IQR) number of events was 97.5 (10-319.25). The median fragility index for RCTs at p-value of 0.05 was 6.5 (4 -12.5). Notably, two trials had a FI ≤ 5 and only one trial had a FI ≥ 10.
Conclusions:
Within our study, the RCTs evaluated consistently yielded non-significant outcomes for mortality and major disability. Notably, we found a higher FI compared to prior studies in critical care, which typically focused on trials with positive outcomes. Binary score outcomes may have limited statistically efficiency in ICH. An ordinal approach for primary endpoint might be a more appropriate way of interpreting neurocritical care trials and help advance management decisions in devastating pathologies such as ICH.