INTRODUCTION: Previous studies about the relationship between tumor recurrence and NRAMP1 and HGPX1 gene polymorphism in patients with non-muscle-invasive bladder cancer (NMIBC) showed inconsistent results. METHODS: We conducted a systematic review of the literature on the basis of data from PubMed and China National Knowledge Infrastructure. According to the predefined selection criteria, the eligibility criteria for the studies found in the literature were assessed by two independent authors. The basic characteristics of the included studies and of data relevant to the meta-analysis were extracted. Patients? survival without recurrence was selected as the measure of effect of meta-analysis. RESULTS: Four publications were retained. Three studies evaluated the NRAMP1 D534N and three the hGPX1 Pro168Leu. Depending on the association between NRAMP1 D534N and tumor recurrence, the meta-analysis revealed no significant heterogeneity and the combined effect was 3.28 [1.77, 6.11]. Depending on the association between hGPX1 Pro198Leu and tumor recurrence, the meta-analysis showed significant heterogeneity and the combined effect was 1.12 [0.45, 2.77]. Publication bias was uncertain, due to the limited number of included studies.The instability of the combined effects was reported. CONCLUSION: Very little data are available on the association between tumor recurrence and NRAMP1 D534N and hGPX1 Pro168Leu poolymorphisms in patients with NMIBC. The NRAMP1 D534N could increase the risk of recurrence, but hGPX1 Pro168Leu effect is not clear. A more thorough investigation should be conducted on a larger sample size in order to better explain this phenomenon.