“…13,15–19,21,23,25,28 Of the 18 randomized controlled trials, 3 (17%) had some degree of bias due to allocation concealment problems, 9,16,23 6 (33%) had some degree of bias due to subject and staff blinding, 9,13,16,19,23,29 and 3 (17%) had other biases. 9,14,26 As shown in Supplemental Digital Content 1 (see Table S3 , http://links.lww.com/JCVP/B74), 23 cohort studies 10,30–51 were of good quality according to the Newcastle‒Ottawa Scale. No bias was indicated for the reporting of BNP and supraventricular tachyarrhythmia in network analysis because the number of trials reporting these outcomes was less than 10 per comparison.…”