Abstract
Background: It is essential for randomized controlled trials (RCTs) to report results in a comprehensive manner. Hence, it is necessary to assess the robustness of the trials with statistically significant and as well as non-significant results. Robustness can be evaluated using fragility index (FI), while reverse fragility index (RFI) can be used for trials with statistically significant as well as non-significant results. The primary aim of this study was to calculate FI and RFI for cardiovascular outcome trials (CVOT).
Materials and Methods: PubMed/MEDLINE was searched to identify all RCTs of antidiabetic drugs where the primary objective was to evaluate the cardiovascular outcomes. We recorded the trial characteristics of each CVOT trial. The FI, RFI, fragility quotient (FQ), and reverse fragility quotient (RFQ) were calculated to evaluate the robustness of the trials. Spearman rank correlation test was used for correlation.
Findings: A total of 889 studies were identified and 24 RCTs were included. Among the 24 trials, 12 (50%) trials achieved statistical significance. The median FI and RFI were 29 (4-12) and 22.5 (1-37) for trials with statistically significant and non-significant results. The median FQ and RFQ were 0.0075 (0.002-0.013) and 0.0003 (0.0001-0.004) for trials with statistically significant and non-significant results. The hazard ratio, p-value, and NNT-B had a strong negative relationship with FI.
Interpretation: Our study showed that half of the trials showing the superiority of cardioprotective benefits have favourable FI. The trials that failed to show superiority also have a reasonable RFI indicating the robustness of these trials. However, the results of the trials where patients lost to follow- up exceed the FI of that trial demands caution during interpretation.
Keywords: Robustness, lost to follow-up, fragility quotient, p-value, reverse fragility index (RFI), cardioprotective benefits.
[http://dx.doi.org/10.1016/j.jacc.2018.06.040] [PMID: 30075873]
[http://dx.doi.org/10.1016/j.jacc.2015.10.014] [PMID: 26653629]
[http://dx.doi.org/10.1016/j.juro.2008.06.026] [PMID: 18710745]
[http://dx.doi.org/10.1016/j.jclinepi.2013.10.019] [PMID: 24508144]
[http://dx.doi.org/10.1001/jamasurg.2018.4318] [PMID: 30422256]
[http://dx.doi.org/10.1001/jamanetworkopen.2020.12469] [PMID: 32756927]
[http://dx.doi.org/10.1056/NEJMoa1515920] [PMID: 27299675]
[http://dx.doi.org/10.1056/NEJMoa2022190] [PMID: 32865377]
[http://dx.doi.org/10.1056/NEJMoa1611925] [PMID: 28605608]
[http://dx.doi.org/10.1056/NEJMoa1911303] [PMID: 31535829]
[http://dx.doi.org/10.1056/NEJMoa1812389] [PMID: 30415602]
[http://dx.doi.org/10.1056/NEJMoa2030186] [PMID: 33200891]
[http://dx.doi.org/10.1056/NEJMoa2030183] [PMID: 33200892]
[http://dx.doi.org/10.1056/NEJMoa2004967] [PMID: 32966714]
[http://dx.doi.org/10.1056/NEJMoa1612917] [PMID: 28910237]
[http://dx.doi.org/10.1056/NEJMoa1509225] [PMID: 26630143]
[http://dx.doi.org/10.1016/S0140-6736(18)32261-X] [PMID: 30291013]
[http://dx.doi.org/10.1056/NEJMoa1603827] [PMID: 27295427]
[http://dx.doi.org/10.1056/NEJMoa1901118] [PMID: 31185157]
[http://dx.doi.org/10.1016/S0140-6736(19)31149-3] [PMID: 31189511]
[http://dx.doi.org/10.1056/NEJMoa1607141] [PMID: 27633186]
[http://dx.doi.org/10.1001/jama.2018.18269] [PMID: 30418475]
[http://dx.doi.org/10.1001/jama.2019.13772] [PMID: 31536101]
[http://dx.doi.org/10.1056/NEJMoa1305889] [PMID: 23992602]
[http://dx.doi.org/10.1056/NEJMoa1307684] [PMID: 23992601]
[http://dx.doi.org/10.1056/NEJMoa1501352] [PMID: 26052984]
[http://dx.doi.org/10.1186/s12933-017-0593-8] [PMID: 28893244]
[PMID: 27532848]
[http://dx.doi.org/10.1056/NEJMoa1615692] [PMID: 28605603]
[http://dx.doi.org/10.1016/S2213-8587(17)30309-1] [PMID: 28917545]
[http://dx.doi.org/10.1161/CIRCOUTCOMES.119.005755] [PMID: 31822121]
[http://dx.doi.org/10.3390/jcm6080079] [PMID: 28805717]
[PMID: 30294197]
[PMID: 28417139]