Table 2: Assessment of risk of bias in included studies.

StudyRandom sequence generationAllocation concealmentBlindingIncomplete outcome dataSelective reportingFree of other biasSummary assessments

Zhao 2007 [19]UUHUUHH
Yan et al. [20] 2009UUHUUHH
Wang and Hou 2010 [21]UUHUUHH
Yang et al. 2010 [22]UUHUUHH
Ge et al. 2010 [23]UUHUUHH
Ge and Zhu 2009 [24]UUHHUHH
Hu et al. 2009 [25]UUHUUHH
Pei and Chen 2009 [26]UUHUUHH
Zuo and Hou 2009 [27]UUHUUHH
Song 2008 [28]UUHHUHH
Xu 2008 [29]UUHLUHH
Huang et al. 2008 [30]UUHUUHH
Li et al. 2008 [31]UUHUUHH
Hua et al. 2007 [32]LLHUUHH
Wang et al. 2007 [33]UUHUUHH
Ma et al. 2007[34]UUHUUHH
X. G. Zhang and Y. M. Zhang 2006 [35]UUHUUHH
Zhang et al. 2006 [36]UUHHUHH
Liu and Yang 2010 [37]UUHHUHH
Yang and Cai 2009 [38]UUHUUHH
Bai and Ding 2007 [39]UUHHUHH
Fang and Wang 2007 [40]UUHHUHH
Jiang 2004 [41]UUHUUHH
Qi and Qu 2008 [42]UUHUUHH
Han and Wang 2011 [43]UUHHUHH

L: low risk of bias, U: unclear, H: high risk of bias.