Objective To assess the efficacy and safety of the tubeless approach with a ureteral stent versus nephrostomy tube for postoperative drainage following percutaneous nephrolithotomy, and to provide guidance for clinical practice. Methods Randomized controlled trials (RCTs) were identified from PubMed (1966 to August 2008), Ovid (1966 to August 2008), Embase (1966 to August 2008), The Cochrane Library (Issue 3, 2008) and CBM (1978 to 2008). We also handsearched for relevant published and unpublished reports and check their references. The quality of the included trials was evaluated by two reviewers. We used The Cochrane Collaboration’s RevMan 5.0.16 software for meta-analysis. Results Eight studies involving 507 patients were included. We divided the patients into three groups: small (8~9 Fr), medium (16~18 Fr) and large (20~24 Fr) according to the diameter of nephrostomy tube for the analysis. Our meta-analyses showed: ①Hospital stay (hours): There was no statistically significant difference between tubeless and small bore tubes, but a difference was found in the comparison of tubeless versus medium and large bore tubes [WMD (95%CI) –32.4 (–33.64, –31.16) and –39.07 (–67.75, –10.39), respectively]; ② Puncture site urinary leakage: No statistically significant difference was found between tubeless and small bore tubes, of between tubeless versus medium tubes [RR= 0.07, 95%CI (0.00, 1.15), P=0.06]; ③ Visual analogue scale scores for postoperative pain on Day 1: There was no statistically significant difference between tubeless and small bore tubes, but there was a difference in tubeless versus medium and large bore tubes [MD (95%CI) –2.80 (–2.94, –2.66) and –2.04 (–2.29, –1.79), respectively];④No statistically significant difference was found in transfusion, fever or infection and operating time between tubeless and any size of nephrostomy tube. Conclusion No statistically significant difference between tubeless versus small bore tubes is found for any of the outcome measurements we analysed. Compared with medium and large bore tubes, tubeless PCNL of ureteral stent could reduce hospital stay, urine leakage and postoperative pain without an increase in complications. There is a moderate possibility of selection bias, performance bias and publication bias in this review, because of the small number of the included studies, which weakens the strength of the evidence of our results. Better evidence from more high-quality randomized controlled trials is needed.