Publication bias
Egger’s test showed that there was no publication bias (P=0.714) for the outcome of mortality. The publication bias assessment was unavailable for other outcomes due to fewer than ten studies included.