To analyze the causes of the failure of hollow screw internal fixation in the treatment of the femoral neck fractures.

Medline, EMBASE, Cochrane library controlled clinical trial registration center, China biomedical literature database, China National Knowledge Infrastructure(CNKI), Wanfang digital periodical full text database, were used to retrieve the clinical studies of the failure of hollow screw fixation in femoral neck fractures. The retrieval time was from the start time of the data bases to December 2017. The literatures of hollow screw internal fixation combined with other internal fixation or bone flap transplantation, and the literatures comparing hollow screw internal fixation with other treatment methods, were all excluded. The quality of the study was evaluated, and RevMan5.1 software was used to perform the meta-analysis the type of fracture, the arrangement of screws, the preoperation traction, the time of weight bearing, the duration time from injury to operation, the postoperative reduction quality, the follow-up duration. The quality of evidence was evaluated by Detsky score.

A total of 32 clinical studies were included with 7 807 patients, 1 418 cases of femoral head necrosis, the total femoral head necrosis rate was 18.16%; 127 cases were bone nonunion, and screw loosening occurred in 52 cases. Garden type I-II patients showed lower cannulated screw fixation failure rate than Garden type III-IV patients [WMD=0.18, 95%CI (0.11, 0.27), I²=59%, P<0.01]. The internal fixation failure rate was higher on the postoperative 3^rd day [WMD=0.58, 95%CI (0.37, 0.89), I²=67%, P<0.05]. Garden index I-II patients showed lower cannulated screw fixation failure rate than Garden index III – IV [WMD=0.19, 95%CI (0.10, 0.33), I²=80%, P<0.01]; the cannulated screw fixation failure rate was lower in postoperative three months than that of three months later [WMD=1.90, 95%CI (1.02, 3.56), I²=85%, P=0.04]. The reduction method[WMD=1.04, 95%CI (0.73, 1.47), I²=75%, P=0.85], traction or not before operation[WMD=1.27, 95%CI (0.80, 2.03), I²=61%, P>0.05], age[WMD=1.19, 95%CI (0.67, 2.13), I²=90%, P>0.05], gender [WMD=1.15, 95%CI (1, 1.33), I²=0%, P=0.05] or the arrangement of the hollow screws[WMD=0.69, 95%CI (0.28, 1.69), I²=61%, P>0.05] had no effect on internal fixation failure rate.

The fracture type, quality of reduction, postoperative ambulation time, the time from injury to operation are the main influencing factors of nail fixation failure of hollow femoral neck fractures; age, gender, preoperative traction and reduction method, screw arrangement fracture have no direct effect on the cannulated screw fixation failure rate in femoral neck.