Table 2.

Risk of bias in included studies

Study name	Random sequence generation	Allocation concealment	Blinding	Completeness of data	Selective outcome reporting	Other bias
Faisy et al., 2016 [6]	Low risk “The randomization sequence was programmed in advance and generated by a statistician independent of the study”	Low risk “Patients were randomized via a computer-generated assignment sequence in a centralized blinded fashion”	Low risk “Double-blind placebo-controlled trial” No details provided on which specific groups were blinded	Low risk “One patient from each group was excluded”	Low risk All outcomes listed in the methods section are reported in the results section	High risk Means and standard deviations were extrapolated from medians and interquartile ranges
Rialp Cervera et al., 2017 [8]	Low risk “Randomization was based on computer-generated random numbers”	Low risk “Treatment and placebo capsules were prepared, packaged and blinded in a centralized hospital pharmacy and distributed to all ICUs”	Low risk “Double-blind study” “Investigators, patients and caregivers were unaware of the randomization list”	Low risk “All enrolled patients completed the trial and were included in the final analysis”	Low risk All outcomes listed in the methods section are reported in the results section	Low risk
Nelson and Wallace, 1965 [12]	Unclear risk Method of random sequence generation not specified	Unclear risk Method of allocation concealment not specified	Low risk “Active and placebo tablets were identical in appearance, and their identity was unknown to the patients and to the assessors during the trial”	High risk “Nineteen patients began the trial, but only 12 were included in the final analysis”	Low risk All outcomes listed in the methods section are reported in the results section	Low risk
Haecki et al., 1983 [15]	Unclear risk Method of random sequence generation not specified	Unclear risk Method of allocation concealment not specified	Low risk “Double blind trial” No details provided on which specific groups were blinded	Low risk “One patient refused follow-up”	Low risk All outcomes listed in the methods section are reported in the results section	Low risk
Vos et al., 1994 [29]	Unclear risk Method of random sequence generation not specified	Unclear risk Method of allocation concealment not specified	Low risk of bias “Double blind trial” No details provided on which specific groups were blinded	Low risk “All patients were studied during three nights”	Low risk All outcomes listed in the methods section are reported in the results section	Low risk
Gulsvik et al., 2013 [28]	Low risk “Randomization was based on computer-generated random numbers”	Low risk “Patients were allocated randomly on a 1:1 basis to a sealed and numbered box containing either acetazolamide or placebo tablets with similar size and colour”	Low risk “placebo controlled and double-blind parallel group trial”	Low risk All enrolled patients were included in the final analysis	Low risk All outcomes listed in the methods section are reported in the results section	Low risk

ICU intensive care unit