There are conflicting reports on the correlation between serum selenium (Se) levels with rheumatoid arthritis (RA). Through a meta-analysis approach, the aim of the present study is to clarify the relationship between serum Se levels with RA. We searched literatures that met our predefined criteria in PubMed, ScienceDirect, and OVID published as of September 2015. Ten eligible articles with 14 case-control studies involving 716 subjects were identified. Overall, pooled analysis indicated that subjects with RA had lower serum levels of Se than the healthy controls (standardized mean difference (SMD) = -1.347, 95 % confidence interval (CI) = [-1.872, -0.823], p < 0.001). Further subgroup analysis indicated that subjects with RA had lower serum Se levels than healthy controls in Europe (SMD = -1.063, 95 % CI = [-1.571, -0.556], p < 0.001) and Asia (SMD = -3.254, 95 % CI = [-4.687, -1.821], p < 0.001) but not in USA (SMD = -0.322, 95 % CI = [-0.657, 0.012], p = 0.059). The serum Se levels were lower in RA than healthy controls measured by graphite furnace atom absorption spectrometry (GFAAS) (SMD = -1.026, 95 % CI = [-1.522, -0.530], p < 0.001), electrothermal absorption spectrometry (EAS) (SMD = -1.197, 95 % CI = [-2.373, -0.020], p < 0.05), flame atomic absorption spectrophotometry (FAAS) (SMD = -0.681, 95 % CI = [-1.049, -0.313], p < 0.001), and inductively coupled plasma-mass spectrophotometer (ICP-MS) (SMD = -11.707, 95 % CI = [-15.189, -8.224], p < 0.001) but not by neutron activation analysis (NAA) (SMD = -0.674, 95 % CI = [-1.350, 0.003], p = 0.051). In conclusion, this meta-analysis supports a significant association between low serum Se concentration with RA. However, this finding needs further confirmation by a trans-regional multicenter study to obtain better understanding of causal relationship between serum Se with RA of different human races or regions.
Keywords: Meta-analysis; Rheumatoid arthritis; Selenium.