We performed a longitudinal analysis of serum interleukin-2 (IL-2) and soluble IL-2 (sIL-2R) concentrations in 60 patients with relapsing-remitting (R-R) multiple sclerosis (MS) as well as in 33 age- and sex-matched normal controls. Overall, we found that serum IL-2 levels remained low (less than 10 U/ml) and did not change appreciably over time; however, marked fluctuations in sIL-2R levels were observed in both the patient and control groups. Using patients as their own controls, we calculated an interrelapse (disease stable) mean sIL-2R concentration as a baseline for comparison with relapse values; sIL-2R levels greater than the 90th percentile of the Student's t distribution of stable values were defined as "peaks." There were a total of 27 sIL-2R peaks, eight (30%) of which correlated with clinical relapses but were potentially predictive of only 18% (8/45) of all the recorded clinical relapses. There was no difference in disease severity (Expanded Disability Status Scale) score between peak-correlated and noncorrelated relapses. Our data suggest that despite reports of elevated levels of IL-2 and sIL-2R in MS, neither may be a useful marker for predicting clinical disease activity in R-R MS.