LIMITATIONS
The current meta-analysis had several limitations. First, 72 hours was used as the window of observation following extubation. This may have helped to distinguish between extubation failure and reintubation for other reasons, but a longer window may have enabled detection of subsequent extubation failures and more precise estimates of the effectiveness of interventions. Second, for several outcomes investigated data were only available from two or three studies. Nevertheless, these data were analyzed to visually describe and quantify pooled effects. Lastly, the six studies with a collective total of 265 preterm infants in this meta-analysis had quite small sample sizes, which may have resulted in bias.