BACKGROUND Interleukin-6 (IL-6) is an inflammatory cytokine that plays a role in transplant rejection. We tested the hypothesis that IL-6 levels in serum or urine could be of value in predicting acute and chronic allograft rejection. Furthermore, we examined whether or not such levels reflected IL-6 expression in the kidney. METHODS We measured IL-6 and IL-6 soluble receptor (IL-6sR) in serum and urine of 145 transplant patients and 20 normal controls. In parallel, we studied 108 renal biopsies. IL-6 was measured with a bioassay system using an IL-6 dependent cell line. IL-6sR was measured with enzyme-linked immunosorbent assay. The biopsies were examined for IL-6 and IL-6 receptor (IL-6R) expression with immunohistochemistry. RESULTS Rejection episodes occurring within 2 months of transplantation were accompanied by elevated IL-6 concentrations in serum (17 +/- 4.8 pg/ml, P < 0.05) and urine (114 +/- 27 pg/ml, P < 0.005), compared to controls. These values returned towards baseline (0-5 pg/ml) after successful rejection treatment. The sensitivity of urine measurements was much higher (93%) than serum (54%). The specificity in serum (70%) and urine (60%) was reduced by infection, acute tubular necrosis, and antithymocyte globulin treatment. Serum and urine IL-6sR values did not correlate with rejection. In biopsy tissue, IL-6 and IL-6R were both elevated during rejection. Especially, mononuclear cells within the interstitial infiltrate stained positive. However, the amount of IL-6 positive cells did not correlate with peripheral IL-6 concentrations. CONCLUSIONS Urine but not serum IL-6 values are sensitive indicators of rejection; however, they are confounded by infection, acute tubular necrosis, and certain antirejection treatments. These features limit their usefulness.