Statistical Analyses

Data were entered and analyzed using Review Manager Version 5.3 (Cochrane Collaboration, Oxford, UK) and the meta-analysis was performed in accordance with the PRISMA guidelines [20]. When not reported, standard deviations were estimated based on the provided confidence interval (CI) limits and standard errors. Because different craving and depression scales were used in the included trials, the standardized mean difference (SMD), an effect measure which has been proved feasible when pooling results of different scales and units [21], was used for the current analysis. The weighted mean difference (WMD) was used to evaluate the smoking frequency, and odds ratio (OR) were used to compare the incidence of any adverse event between patients treated with NAC and controls. The precision of each effect size was reported as a 95% CI. The pooled estimate was computed using the random-effects model [22]. Cochran’s Qtest was conducted and I² statistics were calculated to evaluate statistical heterogeneity and inconsistency between treatment effects across the trials. Statistical significance was set at p < 0.1 for Cochran’s Qtest. Statistical heterogeneity across studies was assessed using the I² test, which quantifies the proportion of total outcome variability across trials. For the ease of reporting, we tentatively assigned low, moderate, and high to I² values of 0−25%, 25−50%, and 50−75% [23]. For I² values higher than 75%, we assigned them to very high.