3.1. Literature search and characteristics of included studies
As shown in Fig 1, our literature search covers 2257 articles. The remaining 1814 studies were reviewed through skimming title and abstract after eliminating duplicates. Then, 69 articles were screened out for further full-text reading. After reading, we excluded 59 studies which did not confirm to inclusion criteria. Finally, 10 studies were selected for further meta-analysis.
Ten studies with a total of 384 patients (1578 episodes) were covered in this systematic review and meta-analysis. In each study, the demographic characteristics of the two groups were similar. And the characteristics of qualified studies are summarized in Table 1. Six studies defined any spontaneous ICP increase to >20 mmHg as an independent elevated ICP episode, then started mannitol or HTS treatment and recorded changes of ICP (2, 4, 5, 13-15). In these studies, total number of episodes were considered as final sample size. And the number of patient was considered as sample size in other four studies which did not define elevated ICP episodes. Therefore, the sample size ranged from 20 to 488. There were three groups in study of Patil, H. and Ichai, C., and the interested data only from mannitol group and HTS group which were extracted (13, 16). In all included trials, HTS solutions ranging in concentrations from 3% to 15% were compared with 15% or 20% mannitol.