Peripheral blood was taken immediately after admission in an attempt different to avoid the potential effects of non-genetic factors on IL-10 production. As shown in Table Table4,4, there were no significant differences in age, gender ratio and ISS scores among trauma patients with different genotypes of the three polymorphisms. Spontaneous IL-10 production was not significantly different between different genotypes of all three loci (data not shown). However, LPS-induced IL-10 production was shown to be significantly different between different genotypes of the -1082 and -592 polymorphisms, showing that the -1082A and -592A alleles were associated with lower IL-10 production (P = 0.005 and P = 0.001, respectively by dominant effect).

Data from linear regression analysis indicated that the functional association of these two polymorphisms with IL-10 production was significantly allele-dose dependent (P = 0.003 and P = 0.037 for -1082 and -592, respectively). Although the -819T allele is also associated with lower LPS-induced IL-10 production, no significant difference was found between the different genotype groups (Table (Table4).4). In case of a combination effect of the three SNPs, we further analyzed the association of ATA haplotype with IL-10 production. Table Table55 showed that the production of IL-10 was lowest in those with haplotype genotype 2 ATA, showing significant difference if compared with those with genotype 0 ATA (P = 0.041). Data from linear regression analysis indicated that this association with IL-10 production were significantly haplotype-dose dependent (P = 0.

041).Table 4Clinical relevance of the IL-10 promoter polymorphisms in patients with major traumaTable 5Clinical relevance of Interleukin-10 haplotypes in major trauma patientsAssociation of the IL-10 promoter polymorphisms with development of sepsis and MODS in trauma patientsAs shown in Table Table4,4, there was a close association of the -1082 Batimastat polymorphism with the development of sepsis, showing that the patients with the A allele had significantly higher sepsis morbidity (P = 0.038 for dominant effect). Data from multiple logistical regression analyses indicated that A to G variation at this position was borderline significantly associated with lower risk of sepsis (OR = 0.677, 95% CI = 0.453 to 1.011, P = 0.057). In addition, there was a borderline significantly increasing trend of MODS scores in the patients with the A allele at -1082 locus (P = 0.088 for recessive effect).