Cardiac surgery with cardiopulmonary bypass (CPB) leads to a systemic inflammatory response with secretion of cytokines. Alterations in the serum concentrations of cytokines have important prognostic significance. Reports on cytokine release during cardiac surgery with CPB have yielded conflicting results. Haemodilution occurs with the onset of CPB resulting in large fluid shifts during the perioperative course of cardiac procedures. In the present study we compare the perioperative course of serum concentrations of TNF-alpha, IL-1beta, IL-6, IL-10 and sIL-2R with and without correction for haemodilution in patients undergoing coronary artery bypass grafting (CABG) surgery. Twenty male patients undergoing elective CABG surgery with CPB and general anaesthesia using a balanced technique with sufentanil, isoflurane and midazolam were enrolled into the study. Serum levels of TNF-alpha, IL-1beta, IL-6, IL-10 and sIL-2R were measured using commercially available ELISA kits. Simultaneous haematocrit values were obtained at all sample times. Statistical analysis was performed by non-parametric analysis of variance and t-tests for data corrected for haemodilution and data that were not corrected for haemodilution. Adjusted significance level was P < 0.01. Intra-operatively, up to the second post-operative day PCV values were significantly decreased compared with preoperative values. Cytokine measurements not corrected for haemodilution were significantly lower than the corrected values. The perioperative haemodilution and decrease in PCV may lead to an underestimation of the cytokine secretion in post-operative patients. We conclude that cytokine measurements were significantly influenced by the perioperative haemodilution and the subsequent decrease in PCV and may in part account for the varying results reported in the literature regarding cytokine release in patients undergoing CABG surgery.