Joint distraction: a treatment to consider for haemophilic arthropathy

program. Haplotypes were estimated using the MLocus software. Thirty-nine of the patients studied developed inhibitors, and nine of them were classified as high-titre inhibitors. Table 1 shows the distribution of the tested polymorphisms in haemophilic patients with and without inhibitors, giving the probability values for the differences. The frequencies found were those generally observed in European populations, as shown in the National Center for Biotechnology Information databank (http://www.ncbi.nlm.nih.gov/SNP/index.html). Only two of the 45 tested distributions significantly deviated from HardyWeinberg equilibrium (IL4R 1902A > G and IL10 592 C > A samples in patients with inhibitors and without inhibitors respectively), a value that was expected due to random sampling only. The differences in frequencies between patients with and without inhibitors were small and did not reach statistical significance for HLA-G, PTPN22, IL4, IL4R, TNFa, TNFR1 and CTLA4. However, the two polymorphisms analysed in the IL10 promoter region ( 819 C > T and 592 C > A) showed different genotypic (P = 0.024 and P = 0.016 respectively) and allelic frequencies ( 819 T = 0.40 vs. 0.26; P = 0.028 and 592 A = 0.40 vs. 0.25; P = 0.019) between patients with and without inhibitors, suggesting an association. These polymorphisms are in complete linkage disequilibrium and the TA haplotype frequencies also differ in a significant way in the two groups. Carriers of these alleles seem to be more prone to develop inhibitors than those carrying their alternative forms [ORs for carriers of the risk alleles: IL10 819:3.21(CI: 1.45–7.1); IL10 592:3.00 (CI: 1.36–6.49); TA haplotype: 2.61 (CI: 1.22–5.59)]. Significant differences in the same direction (T and A risk alleles) were obtained in a recent study [7] in a very different (Chinese) ethnic population, reinforcing our findings. But Pavlova et al. [8] found exactly the same frequencies for the 819 C > T and 592 C > A polymorphisms in patients with and without inhibitors. Curiously, the same authors reported significant differences, but in another IL10 SNP ( 1082 A > G) between the two groups of patients. As far as we know, the data onHLA-G ins/del 14 bp and 3142 C > G; IL 4R 1902 A > G; TNFa 1031 T > C and 863 A > C; and TNFR1 303 A > G are the first reported in the literature. The absence of significant differences between patients with and without inhibitors was also found previously for PTPN221858 C > T, IL4 590C > T,TNFa 827 C > T and 238 G > A, and CTLA4 49 A > G by other authors. However, for two others, the results are in conflict: (i) TNFa 308 G > A, and (ii) CTLA4 318 C > T, which did not present association in our study were reported as showing association by other researchers. The reasons for the heterogeneous results found in many association studies are numerous and varied. Sample sizes, ascertainment differences, population and trait genetic heterogeneities may be mentioned. In addition, in quantitative characteristics, most factors account for only a small proportion of the total genetic risk. In conclusion, there is evidence that two SNPs in the interleukin 10 system (IL10 819 C > T and 592 C > A) influence the susceptibility to develop factor VIII inhibitors. Nucleotides T and A are risk factors for this condition. As the same association was previously independently found in an ethnically diverse population, the significance of our finding is strengthened.