The contribution of hyperthermia to the differential leukocytosis of exercise remains obscure. This study examined changes in circulating sympathoadrenal hormone concentrations and patterns of leukocyte and lymphocyte subset (CD3(+), CD4(+), CD8(+), CD19(+), CD3(-)16(+)/56(+)) redistribution during exercise, with and without a significant rise of rectal temperature (T(re)). Ten healthy men [age 26.9 +/- 5.7 (SD) yr, body mass 76.0 +/- 10.9 kg, body fat 13.9 +/- 4.6%, peak O(2) consumption: 48.0 +/- 12.4 ml x kg(-1) x min(-1)] exercised for 40 min (65% peak O(2) consumption) during water immersion at 39 or 18 degrees C. T(re) increased from 37.2 to 39.3 degrees C (P < 0.0001) after 40 min of exercise in 39 degrees C water but was held constant to an increment of 0.5 degrees C during exercise in 18 degrees C water. Application of this thermal clamp reduced exercise-associated increments of plasma epinephrine (Epi) and norepinephrine (NE) by >50% (P < 0.05) and abolished the postexercise increase in cortisol. Thermal clamping also reduced the exercise-induced leukocytosis and lymphocytosis. Multiple regression demonstrated that T(re) had no direct association with lymphocyte subset mobilization but was significantly (P < 0.0001) correlated with hormone levels. Epi was an important determinant of total leukocytes, lymphocytes, and CD3(+), CD4(+), CD8(+), and CD3(-)CD16(+)/56(+) subset redistribution. The relationship between NE and lymphocyte subsets was weaker than that with Epi, with the exception of CD3(-)CD16(+)/56(+) counts, which were positively (P < 0.0001) related to NE. Cortisol was negatively associated with leukocytes, CD14(+) monocytes, and CD19(+) B- and CD4(+) T-cell subsets but was positively related to granulocytes. We conclude that hyperthermia mediates exercise-induced immune cell redistribution to the extent that it causes sympathoadrenal activation, with alterations in circulating Epi, NE, and cortisol.