中文摘要：

目的使用Meta 分析的方法分析比较机器人辅助和常规徒手开放椎弓根螺钉内固定的精确性和安全性。方法计算机检索Pubmed、Embase、Cochrane library、中国生物医学文献（CBM）、万方、中国知网数据库（CNKI）中2016 年12 月1 日前的文献。根据制定的纳入与排除标准，筛选合格文献并进行数据提取和质量评价。利用RevMan5.3 软件进行Meta 分析。对二分类数据采用相对危险度（RR）及95%置信区间（CI）表示效应差异。对连续型数据采用均数差（MD）或标准化均数差（SMD）表示效应差异。当P〈0.05，95%CI不含数值1 时说明差异有统计学意义。结果共纳入266 例患者，包括机器人辅助138 例、常规徒手开放128 例；共置入螺钉1 200 枚，包括机器人辅助608 枚、常规徒手开放592 枚。结果显示，根据Gertzbein-Robbins 分级标准（A、B、C），机器人辅助和常规徒手开放方法之间的椎弓根螺钉置钉准确率差异无统计学意义，其中A（RR=1.07，95%CI：0.82,1.39，I2=46%，P=0.62）、B（RR=1.56，95%CI：0.86，,2.82，I2=0%，P=0.14）、C（RR=0.91，95%CI：0.32，2.55，I2=0%，P=0.85）。两组并发症发生率差异无统计学意义（RR=0.33，95%CI：0.05，2.17，I2=0%，P=0.25）。两组螺钉修复率差异无统计学意义（RR=0.53，95%CI：0.10,2.90，I2=0%，P=0.47）。两组手术时间差异有统计学意义（MD=20.90，95%CI：5.54，36.26，I2=15%，P=0.008）。两组辐射曝光时间差异有统计学意义（SMD=-1.19，95%CI：-1.63，-0.75，I2=0%，P〈0.000 01）。两组螺钉与关节面的距离差异有统计学意义（SMD=-1.15，95%CI：0.79,1.51，I2=16%，P〈0.000 01）。结论与常规徒手开放方法相比，目前的证据尚不能证实机器人辅助系统在椎弓根螺钉内固定的手术精确率与并发症发生率方面存在明显优势。脊柱手术机器人辅助系统作为新技术应用于脊柱外科手术具有深入开发的潜能。

英文摘要：

Objective To compare and analyze the accuracy and safety of robot-assisted and conventional freehand open approach in pedicle screw fixation using meta-analysis. Methods PubMed, Embase, Cochrane,China Biology Medicine, Wanfang and CNKI databases were searched by computer retrieval to identify the relevant literatures published before December1, 2016. The qualified literatures were selected according to the pre-established the inclusion criteria and exclusion criteria, and processed for data extraction and quality evaluation. The RevMan 5.3 software was used for the meta-analysis. The relative risk （RR） and 95% confidence interval （CI） were adopted to represent the effect differences in the dichotomous variable data. The effect difference of continuous variable data was represented by mean difference （MD） or standardized mean difference （SMD）. When P〈0.05 and 95% CI does not contain the value of 1, the difference was considered statistically significant. Results A total of 266 patients were enrolled, including 138 cases of robot-assisted group and 128 cases of conventional freehand open group, in which 1 200 screws were implanted, including 608 screws by robot-assisted system and 592 screws by conventional freehand open approach. The Gertzbein-Robbins grading criteria （grade A to C） were used. The results showed that there was no significant difference in the accuracy of pedicle screw fixation between the robotic-assisted and conventional freehand open approach according to the grading criteria, i.e. grades A （RR=1.07, 95%CI： 0.82,1.39, I2=46%, P=0.62）, grades B （RR=1.56, 95%CI： 0.86,2.82, I2=0%, P=0.14） and grades C （RR=0.91, 95%CI： 0.32,2.55, I2=0%, P=0.85）. For the two groups, the difference was no statistically significance in the complication rate （RR=0.33, 95%CI： 0.05, 2.17, I2=0%, P=0.25）, and in the revision rate （RR=0.53, 95%CI： 0.10, 2.90, I2=0%, P=0.47）. The difference was statistically significance in the overall surgical time （MD=20.90, 95%CI： 5.