The influence of serum tumor necrosis factor-alpha and interleukin-6 concentrations on nonhematologic toxicity and hematologic recovery in patients with acute myelogenous leukemia.

To confirm the reported correlation of tumor necrosis factor-alpha (TNF-alpha) and interleukin-6 (IL-6) serum concentrations with nonhematologic toxicity after cytotoxic chemotherapy and to examine their possible effects on hematopoiesis, we evaluated serum TNF-alpha and IL-6 concentrations every 3 days during 21 chemotherapy cycles in 11 patients with acute myelogenous leukemia (AML) and one patient with chronic myelogenous leukemia in blast crisis (CML-BC). All patients developed grade IV hematologic toxicity. In 13 patient cycles, grade III-IV nonhematologic toxicity developed: hepatic (nine), pulmonary (six), and stomatitis (five). In these patient cycles, IL-6 concentrations increased from 10.1 pg/mL (4.6-15.6, 95% CI) before nonhematologic toxicity to 64.8 (5.3-124.2, 95% CI) at the onset of toxicity (p = 0.02). TNF-alpha concentrations were not detectable before nonhematologic toxicity but increased to 20.4 pg/mL (not detectable [ND]-45.5, 95% CI) at the onset of grade III-IV toxicity. In six patient cycles, grade II nonhematologic toxicity developed: hepatic (five), pulmonary (one), and stomatitis (two). In these six, IL-6 concentrations increased from 12.1 pg/mL (6.8-17.4, 95% CI) before toxicity to 21.4 (11-31.8, 95% CI) at the onset of toxicity (p = 0.03). TNF-alpha concentrations were detectable in one patient cycle before toxicity and detectable in only two patient cycles at the onset of toxicity. The peak IL-6 and TNF-alpha concentrations did not correlate with the onset of nonhematologic toxicity in 87% of patient cycles. In patient cycles with a cumulative IL-6 area-under-the-serum concentration vs. time curve (AUC) > 1000 pg/mL.d, platelet recovery (> 30 x 10(9)/L and platelet transfusion-independent) occurred earlier at 21.9 days (18.7-25.1, 95% CI) compared to the 30.6 days (23.6-37.5, 95% CI, p = 0.02) in patient cycles with an IL-6 AUC < 1000 pg/mL.d. Patient cycles with a cumulative TNF-alpha AUC > 150 pg/mL.d required a mean of 17.5 units of red blood cells (RBCs) (9.3-25.7, 95% CI) compared to patient cycles with an AUC < 150 pg/mL.d, which required only 8.9 units of RBCs (6.2-11.7, 95% CI, p = 0.03). The peak concentration and AUC for IL-6 and TNF-alpha were not significantly different between those receiving growth factors (G-CSF, six; GM-CSF, one) and those not receiving growth factors (14). Endogenous IL-6 and TNF-alpha serum concentrations increase in patients who experience nonhematologic toxicity and correlate with hematologic recovery after chemotherapy.