Levels of anti-dsDNA measured just after an immunoadsorption procedure in systemic lupus erythematosus (SLE) patients are sometimes paradoxically larger than those measured just before the procedure. A 1:100 in vitro single-compartment immunoadsorption system model was devised to determine which of two models, one- or two-compartment, more closely approximates the kinetics of anti-dsDNA during apheresis procedures. Ten SLE patients were employed in this study. A total of 4,100 ml of plasma was passed through the dextran sulfate cellulose columns during one clinical apheresis session. In eight of ten patients, the log of RIA-measured anti-dsDNA titers decreased linearly as treated plasma volume increased, in both the clinical procedure and the experimental model. The mean adsorption efficacy in the clinical apheresis procedure and in the in vitro model was 0.37 and 0.27, respectively. However, in one patient the RIA-measured level of anti-dsDNA increased during the apheresis procedure; this phenomenon was mirrored in the model (definitely a single pool model). In contrast, the level of anti-dsDNA, as measured by ELISA, decreased in accordance with the increase of treated plasma volume in both the clinical and the in vitro apheresis procedures. Therefore, an increased titer of anti-dsDNA as measured by RIA immediately following clinical apheresis cannot be accounted for exclusively by an inflow of antibodies from a secondary (extravascular) pool into the circulating plasma. In short, a one-compartment model is applicable and an explanation must be sought elsewhere.