Fig. 2

Risk of bias summary: + :no bias; − : bias; ?: bias unknown. Fourteen studies adequately described the correct randomization. Thirteen studies demonstrated sufficient allocation concealment. Four studies described the blinding of participants and personnel. No studies described the blinding of outcome assessment. All thirteen articles retained complete outcome data and avoided selective reporting. We rated as unclear risk of other bias because we can't ignore other potential dangers of biases. As a result, there is low or moderate risk of bias in most of the articles reviewed