Sixteen RCTs (n=1,386 patients) were included in the review. Twelve studies (1,051 patients) included midazolam as the comparator. Publication bias was assessed using a funnel plot, which showed that there was a possibility of small publication bias.
There was no significant difference in mortality between patients sedated with propofol compared to those sedated with an alternative sedative. There was no evidence of statistically significant heterogeneity. Both medium-term (WMD: -0.88 days, 95% CI -1.41 to -0.35) and long-term (WMD: -2.15 days, 95% CI -3.85 to -0.46) use of propofol were associated with a statistically significant reduction in length of ICU stay. Whilst there was evidence of statistical heterogeneity for the medium-term use of propofol (I2=86.7%), there was no evidence of statistical heterogeneity for the long-term use of propofol.
The use of propofol was associated with a statistically significant slightly shorter duration of mechanical intervention ((WMD -0.29 days, 95% CI -0.58 to -0.01; four RCTs, 448 patients) compared to alternative sedatives.
Sensitivity analysis in which trials were restricted to those with propofol and midazolam only, revealed that there was no statistically significant difference mortality between patients sedated with propofol compared to those sedated with midazolam. Also, a sensitivity analysis in which trials funded by a pharmaceutical company were excluded did not alter the results.