Outcome predictors of biologic therapeutic drugs like TNF inhibitors are of interest since side effects like serious infections or malignancy cannot be completely ruled out. Response rates are heterogeneous. The present study addressed the question whether in patients with rheumatoid arthritis (RA) interleukin-10 (IL-10) promoter genotypes with potential relevance for IL-10 production capacity are associated with response to long-term treatment with etanercept. Caucasian RA patients that, according to the EULAR criteria, responded well (n = 25), moderately (n = 17) or not (n = 8) to etanercept therapy (median 36 months, range 4-52), and 160 matched controls were genotyped for the IL-10 promoter SNPs -2849 G>A (rs6703630), -1082 G>A (rs1800896), -819 C>T (rs1800871) and -592 C>A (rs1800872). Haplotypes were reconstructed via mathematic model and tested for associations with disease susceptibility and therapy response. We identified the four predominant haplotypes AGCC, GATA, GGCC, and GACC in almost equal distribution. Patients that responded well carried the putative IL-10 low producer allele -2849 A or the haplotypes AGCC and GATA (RR 2.1 and 4.0, respectively; 95% CI 1.1-4.0 and 1.1-14.8), whereas an unfavourable response was associated with carriage of the putative high producer haplotype GGCC (RR 1.9, 95% CI 1.1-3.3). No significant associations of alleles or haplotypes with disease susceptibility were observed. In RA, a low IL-10 production which is genetically determined rather by haplotypes than by SNPs may favour the response to etanercept treatment. Iatrogenic blockade of TNF may reveal proinflammatory effects of its endogeneous antagonist IL-10. Further studies are needed to correlate these genetic findings to direct cytokine measurements.