Abstract

Background: Tissue necrosis factor (TNF)-α is a naturally occurring cytokine involved in the pathogenesis of several hematologic malignancies, but an independent prognostic role for TNF-α in acute myeloid leukemia (AML) and high-risk myelodysplastic syndromes (MDS) has not been established. The purpose of this study was to determine the association between TNF-α levels and pretreatment characteristics and clinical outcomes in these diseases.

Methods: From 9/94 through 1/01, serum levels of TNF-α were prospectively measured in 198 patients with AML or high-risk MDS. Serum TNF-α was quantified using commercially available ELISA kits (R&D Systems, Minneapolis, MN) (lower limits of sensitivity were 4.4 pg/mL). The appropriate recombinant cytokine was used to create a standard curve for each cytokine, and serum level was ascertained using the appropriate standard curve. All values represent the means of 3 sets of duplicate experiments.

Results: The median patient age of the 198 patients with serum TNF-α measurements was 65 years (range, 21–84 years); 62% were ≥ 60 years. Cytogenetics were favorable in 6% of patients, intermediate in 53%, poor-risk in 38%, and 3% of patients had insufficient metaphases or unknown cytogenetics. Zubrod performance status (PS) was 0–1 in 66% of patients, 2 in 25%, and 3–4 in 9%. Forty-nine percent had a history of an antecedent hematologic disorder (AHD), and 93% received ara-C-based therapy. High TNF-α levels were associated with poor Zubrod PS (cor= .28); high levels of β2 microglobulin (cor=.50), creatinine (cor=.35), uric acid (p=.35), or LDH (p=.32); low albumin levels (cor=−.22); and longer duration of AHD (cor=.74) (p<.0001 for all variables). High TNF-α levels were more common in patients with AML (vs. MDS, p=.026) and M4-M5 subtypes according to the French-American-British classification (p=.02). The median follow-up of surviving patients was 5.6 years. Clinical outcomes by TNF-α are shown in the Table:

TNF-α (pg/mL_TNF-α <10pg/mLTNF-α ≥10 pg/mLp-value
No. pts 98 100  
CR (%) 72 (73) 52 (52) .0018 
Median EFS, months 9.3 2.5 .0009 
Median survival, months 1.2 0.6 .0003 
TNF-α (pg/mL_TNF-α <10pg/mLTNF-α ≥10 pg/mLp-value
No. pts 98 100  
CR (%) 72 (73) 52 (52) .0018 
Median EFS, months 9.3 2.5 .0009 
Median survival, months 1.2 0.6 .0003 

In multivariate analysis, independent factors predicting response to therapy were younger age (p=.03), shorter duration of AHD (p=.007), higher hemoglobin levels (p=.04), lower TNF-α levels (p=.038), and better PS (p=.02). Factors independently prognostic of longer survival were younger age (p=.0008), better-risk cytogenetics (p=.049), lower TNF-α levels (p=.049), lower leukocyte counts (p=.003), higher hemoglobin levels (p=.03), lower β2-microglobulin levels (p=.03), female gender (p=.02), no history of chemotherapy or radiotherapy (p=.02), and race other than African American (p=.02). Factors independently prognostic of longer event-free survival (EFS) were better-risk cytogenetics (p=.002), lower β2-microglobulin levels (p=.008), lower leukocyte counts (p=.03), and younger age (p=.04).

Conclusions: These data indicate that an elevated serum TNF-α level is associated among others with poorer Zubrod PS and higher levels of β2 microglobulin and it is an independent adverse prognostic factor for response and survival in the context of established prognostic factors for AML.

Author notes

Disclosure: No relevant conflicts of interest to declare.