Study filtering and quality evaluation
The literature screening process and results were shown in figure 1. Based on the pre-defined search strategy, a total of 1187 studies were filtered from Pubmed, Embase, and Cochrane Library databases. After excluded duplicated literature, there were 715 studies remaining. Based on the title and abstract, 610 studies did not meet the inclusion criteria obviously. After reading the full text, 84 studies were then retrieved. A total of 21 studies were finally obtained for further quality evaluation 15-35.
The filtered studies were published from 1996 to 2016, and most of them were published after 2005. The research year were ranged from 1996 to 2014, while some literatures did not report the research year. The research areas of these studies included Egypt, Turkey, the United States, China, Italy and South Korea. Total 3410 enuresis were included, the drugs used contained dDAVP (Desmopressin), Tam (Tamsulosin), Tol (Tolterodine), Oxy (Oxybutynin), Dox (Doxazosin), Imi (Imipramine), Pse (Pseudoephedrine), Ver (Verapamil), Ibu (Ibuprofen) and Mia (Mianserin). There is no significant difference in gender and age between each group of these studies.
The results of RCT quality evaluation show that the quality of included literature is generally high. However, parts of studies showed unclear risk of bias in Allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias) and selective reporting (reporting bias) (Figure 2 and 3).
Mesh meta-analysis results
Parameter was set for the ADDIS software, including number of chains of 4, tuning iterations of 20000, simulation iterations of 50000, thinning interval of 10, inference samples of 10000 and variance scaling factor of 2.5. Mesh meta-analysis was processed for response rates, mean number of nocturia, nocturnal urine volume and adverse effects.
The consistency test was performed by Node-splitting analysis, and the P values were all more than 0.05. Therefore, the consistency model was adopted. All the PSRFs of response rates, mean number of nocturia, nocturnal urine volume and adverse effects were close to 1, which proved that the model was completely converged, the iterative effect was very good, and the results were stable.
Response rates
As shown in Table 1, for the response rates, dDAVP+Tol, dDAVP+Imi, Imi+Oxy, and Imi+Pse treatment were the best, while the efficacy of Ver, Mia, and Est were the worst. However, the difference between the groups was not statistically significant (P > 0.05).
Mean number of nocturia
For the mean number of nocturia indicator, the effects of dDAVP+Imi and dDAVP+Tam treatment were the best, while the efficacy of Imi was the worst. However, the difference between the groups was not statistically significant (P > 0.05) (Figure 2).
Nocturnal urine volume
For nocturnal urine volume, the effects of dDAVP + Oxy and dDAVP + Tam were the best, while the efficacy of Pse and Tam was the worst. However, the difference between the groups was not statistically significant (P > 0.05) (Figure 3).
Adverse effects
For adverse effects index, the incidence of adverse effects of Est and dDAVP + Imi was the lowest, the incidence of Ver was the highest, and the incidence of Imi and Imi + Oxy were also high (Table 4).