Objective: To provide a quantitative summary in estimating the association between polymorphisms of 3 loci in NRAMP1 gene and susceptibility to tuberculosis in East-Asia population by means of meta-analysis.
Methods: We searched databases (MEDLINE, OVID and CBM disc) from January 1995 to May 2005 using "NRAMP1" or "SLC11A1", in combination with "tuberculosis", also performed a manual search of citations from relevant original studies and literature. For each study involved, information was collected concerning the characteristics of the subjects, such as mean age of cases and the size of study. These characteristics were used to evaluate the sources of variation. Summary ORs and corresponding 95% CI were estimated by fixed effects (Mantel-Haenszel) or random effects (DerSimonian and Laird) model. To check for publication bias,a funnel plot, using Egger's linear regression method, was constructed. Cumulative meta-analysis was performed to evaluate whether the summary OR for studies with the polymorphisms of the 3 loci in the NRAMP1 gene was changing along with the accumulation of more data. Chi-square goodness of fit was used to test deviation from Hardy-Weinberg equilibrium.
Results: Eight publications, with the number of cases and controls of 1067 and 1084 respectively, were identified and all genotype frequencies were consistent with Hardy-Weinberg equilibrium. The summary ORs for studies with polymorphisms of 3' UTR, D543N and INT4 loci of the NRAMPI gene among the East-Asia population were 1.68(95% CI: 1.31-2.16, P< 0.001), 1.78(95% CI: 1.38-2.30, P< 0.001), 1.56 (95% CI: 0.72- 3.35, P = 0.26), respectively when compared with their corresponding common homozygotes. Publication bias was not found in the studies with the three loci, except for INT4 locus, by Egger linear regression method. The cumulative summary effects ORs were 1.85 (P = 0.02) in 2000, 1.35 (P = 0.12) in 2002,1.64 (P= 0.001) in 2003, and 1.68 (P<0.001) in 2004 for 3'UTR locus, 1.88 (P = 0.001) in 2000,1.65(P = 0.001) in 2002,1.70(P<0.001) in 2003,1.76(P<0.001) in 2004, and 1.78(P<0.001) in 2005 for D543N locus, and 0.88(P = 0:70) in 2002, 2.50(P = 0.41) in 2003, 1.52(P = 0.42) in 2004 and 1.56(P = 0.26) in 2005 for INT4 loucs.
Conclusion: Polymorphisms at 3' UTR and D543N loci had statistically significant association between the NRAMP1 variants and susceptibility to tuberculosis in the East-Asia descendants, and variant in the INT4 locus failed to show statistically significant association in the East-Asia population.