Within the limitations of the various experimental protocols there appears to be agreement in the literature that unused dialysis fluids, at least when studied in vitro, adversely affect multiple leukocyte functions. The effects of dialysis fluids on leukocytes that have been reported to date include: 1. Decreased cell viability of PMNs, PM phis, PBMCs, and lymphocytes; 2. Inhibited phagocytosis and bacterial killing of various microorganisms by PM phis, PMNs, and peripheral blood leukocytes; 3. Reduced secretion of leukotrienes (LTB4, LTC4) from peritoneal and peripheral blood PMNs and PBMCs; 4. Reduced secretion of prostaglandins (PGE2, TXB2 and 6-keto-PGF1 alpha) from PM phi; 5. Decreased production of many cytokines including TNF alpha, IL-8, and IL-6 in PM phis and PBMCs. In addition, several studies targeting the potential mechanisms by which dialysis solutions inhibit leukocyte function identified the initial low pH of the fluids in combination with their lactate content as being of primary relevance, since they may lead to a rapid intracellular acidification of leukocytes. Moreover, some studies indicated the importance of fluid hyperosmolarity and excessive glucose concentrations. These results are indirectly supported by recent in vitro investigations of alternative fluids, which showed improved leukocyte function following exposure to solutions with neutral pH, bicarbonate buffer instead of lactate, or normal osmolarity due to use of an alternative osmotic agent (e.g., glucose polymer). In conclusion, the evidence obtained during in vitro experimentation suggests that current dialysis fluids are, indeed, not biocompatible. However, whether this also bears physiological relevance in vivo remains to be established in controlled clinical trials comparing conventional fluids to alternative solutions with improved biocompatibility. With regard to the future development of in vitro models for biocompatibility assessment, the following guidelines are suggested: 1. Cell functional parameters should be studied in more than one cell population; 2. Depending on which fluid aspect is under investigation, short or even very short exposure times should be used (e.g., < 30 min for pH/buffer studies; < 4 hours for osmolality/osmotic agent studies); 3. In case the parameter/readout of interest requires longer study periods than indicated above (e.g., studies of cytokine induction or surface receptor expression), preincubation/recovery models should be preferred over coincubation experiments.