Axillary lymph node involvement in breast cancer is a marker of recurrence risk. Despite aggressive adjuvant therapy, recurrence in patients with four or more involved lymph nodes approaches 50% at 5 years from diagnosis. Markers that can distinguish those likely to relapse from those likely to be cured are needed to tailor therapy and provide accurate prognostic information to patients. Although most work in this area has focused on tumor characteristics, we hypothesized that the host environment might also play a role in determining risk of relapse. We hypothesized that host inflammatory response, mediated in part by production of interleukin-6 (IL-6), might play a role in the elimination of microscopic residual tumor. Polymorphisms in the IL-6 promoter region appear to modulate serum levels of the cytokine via regulation of gene transcription. A single nucleotide polymorphism involving substitution of cytosine for guanine at position -174 has been associated with reduced transcription and improved outcome in a variety of nonmalignant diseases, including coronary artery disease and several autoimmune conditions. Tumor necrosis factor (TNF) alpha is a proinflammatory cytokine that also plays a role in regulating IL-6 transcription. We hypothesized that polymorphisms in IL-6 (-174 G>C) or TNF-alpha (G-238 or G-308) might be associated with prognosis in a subset of patients with high-risk breast cancer. Genotyping was performed on DNA from stored stem cells in 80 breast cancer patients diagnosed with at least four positive axillary lymph nodes at diagnosis who underwent anthracycline-based adjuvant chemotherapy followed by high-dose multiagent chemotherapy with stem cell rescue. Cox proportional hazards models were used to estimate the effect of genotype and other known prognostic factors on disease-free and overall survival (DFS and OS, respectively). The presence of at least one C allele in the IL-6 promoter at position -174 was significantly associated with both DFS and OS compared with G/G homozygotes. After adjustment for estrogen receptor (ER) status, number of involved lymph nodes, and tumor size, those patients carrying the G/G genotype had a 2.1-fold increase in the rate of failure and a 2.6-fold increase in the rate of death compared with carriers of any C allele at a mean follow-up of 55 months. ER status modulated the effect of IL-6 polymorphism: both DFS and OS were most favorable in patients who were carriers of any C-allele (G/C or C/C) and had ER-positive tumors. The presence of either G/G genotype or an ER-negative tumor increased the hazard of failure [hazard ratio (HR), 2.6 and 3.2, respectively] and death (HR, 2.0 and 2.2, respectively). The combination of both G/G genotype and ER-negative tumor resulted in an additional increase in the hazard of failure (HR, 5.4; four-group comparison, P = 0.003) and death (HR, 6.2; four-group comparison, P = 0.001). TNF-alpha -308 and -238 polymorphisms were not associated with variation in DFS or OS in this cohort. The IL-6-174 promoter polymorphism is associated with clinical outcome in this cohort of node-positive breast cancer patients who received high-dose adjuvant therapy. IL-6 genotype modulated the effect of ER status on outcome. These results support the hypothesis that IL-6 may play an important role in the control of micrometastatic disease in breast cancer. Additional studies are needed to confirm these results and elucidate the mechanisms responsible for these differences.