We selected studies by predetermined inclusion criteria listed above. The author screened titles and abstracts of related trials to select articles for full-text assessment, and did same to the full-text papers and the eligibility of experiments. We use a flow diagram to describe the process of study selection. Any disagreements were resolved by consensus discussion in a third party if necessary.