Key Points
CXCL9 is a reliable peripheral blood biomarker of interferon-γ-driven hyperinflammation, of which the most severe form manifests as HLH.
Elevated CXCL9 (≥16,100 pg/mL) identifies adults with HLH at high risk for early mortality irrespective of the underlying HLH trigger.
Hemophagocytic lymphohistiocytosis (HLH) is an interferon-γ-driven hyperinflammatory syndrome with high morbidity and mortality. Identifying reliable prognostic biomarkers is challenging due to various predisposing conditions and triggers. CXCL9 is a clinically-validated biomarker and surrogate marker of interferon-γ-mediated inflammation. We aimed to identify the role of CXCL9 in predicting severe disease and death in adults with HLH using a multicenter retrospective cohort of consecutively hospitalized patients undergoing a clinical evaluation for HLH that included CXCL9 testing. Patients were classified as HLH if they met HLH-2004 and/or HScore criteria. Conditional-inference decision trees and Cox regression models were used to identify which clinical variables associated with acute mortality in HLH patients. Overall, 171-patients were reviewed and 126-patients met HLH criteria. Median age was 55-years (IQR: 40-66) with 77 (62%) male and 64 (51%) White. CXCL9 was markedly elevated in HLH patients. Unbiased decision tree modeling, incorporating all clinical laboratory values, identified only CXCL9 >16,100 pg/mL as the optimal predictor of inpatient mortality (p=0.009). Cox regression models demonstrated CXCL9 >16,100 pg/mL was significantly associated with 90-day mortality (hazard ratio, 3.0; 95%CI 1.6-5.6, p=0.003) when controlling for age, sex, race, malignancy, and HLH-targeted therapy. This shorter time-to-death in the elevated CXCL9 subgroup remained significant even after subdividing patients into malignancy (n=53) and non-malignancy-HLH (n=73) (log-rank, p<0.05). Continuous increases in CXCL9 within the cohort strongly associated with greater mortality. CXCL9 is a novel clinical marker that identifies high-risk HLH independent of underlying disease and could be used to select patients for early and aggressive targeted-immunomodulatory therapy.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal