Department of Health Sciences, University of L'Aquila, Coppito (AQ), Italy.
Postgrad Med J. 2011 Sep;87(1031):643-51. doi: 10.1136/pgmj.2009.047027rep.
OBJECTIVES Over the past 10 years there has been increasing concern about the possible behavioural effects of mobile phone use. This systematic review and meta-analysis focuses on studies published since 1999 on the human cognitive and performance effects of mobile phone-related electromagnetic fields (EMF). METHODS PubMed, Biomed, Medline, Biological Sciences, PsychInfo, PsycARTICLES, Environmental Sciences and Pollution Management, Neurosciences Abstracts and Web of Science professional databases were searched and 24 studies selected for meta-analysis. Each study had to have at least one psychomotor measurement result as a main outcome. Data were analysed using standardised mean difference (SMD) as the effect size measure. RESULTS Only three tasks (2-back, 3-back and simple reaction time (SRT)) displayed significant heterogeneity, but after studies with extreme SMD were excluded using sensitivity analysis, the statistical significance disappeared (χ(2)(7)=1.63, p=0.20; χ(2)(6)=1.00, p=0.32; χ(2)(10)=14.04, p=0.17, respectively). Following sensitivity analysis, the effect of sponsorship and publication bias were assessed. Meta-regression indicated a significant effect (b1/40.12, p<0.05) only for the 2-back task with mixed funding (industry and public/charity). Funnel plot inspection revealed a significant publication bias only for two cognitive tasks: SRT (Begg's rank correlation r=0.443; Egger's test b=-0.652) and the subtraction task (Egger's test b=-0.687). CONCLUSIONS Mobile phone-like EMF do not seem to induce cognitive and psychomotor effects. Nonetheless, the existence of sponsorship and publication biases should encourage WHO intervention to develop official research standards and guidelines. In addition, future research should address critical and neglected issues such as investigation of repeated, intensive and chronic exposures, especially in highly sensitive populations such as children.
目的 在过去的 10 年中，人们越来越关注手机使用可能带来的行为影响。本系统评价和荟萃分析重点关注自 1999 年以来发表的关于手机相关电磁场（EMF）对人类认知和表现影响的研究。
方法 检索 PubMed、Biomed、Medline、Biological Sciences、PsychInfo、PsycARTICLES、Environmental Sciences and Pollution Management、Neurosciences Abstracts 和 Web of Science 专业数据库，并选择了 24 项研究进行荟萃分析。每项研究都必须至少有一个精神运动测量结果作为主要结果。使用标准化均数差（SMD）作为效应量进行数据分析。
结果 只有三项任务（2 位、3 位和简单反应时间（SRT））显示出显著的异质性，但在使用敏感性分析排除极端 SMD 的研究后，统计显著性消失（χ(2)(7)=1.63，p=0.20；χ(2)(6)=1.00，p=0.32；χ(2)(10)=14.04，p=0.17）。进行敏感性分析后，评估了赞助和发表偏倚的影响。元回归表明，只有在混合资金（工业和公共/慈善）的 2 位任务中，赞助具有显著影响（b1/40.12，p<0.05）。漏斗图检查仅发现两项认知任务存在显著的发表偏倚：SRT（Begg 等级相关 r=0.443；Egger 检验 b=-0.652）和减法任务（Egger 检验 b=-0.687）。
结论 类似于手机的 EMF 似乎不会引起认知和精神运动的影响。尽管如此，赞助和发表偏倚的存在应该鼓励世界卫生组织采取干预措施，制定官方的研究标准和准则。此外，未来的研究应该解决一些关键和被忽视的问题，如重复、密集和慢性暴露的调查，特别是在儿童等高度敏感的人群中。
