Hemophagocytic lymphohistiocytosis (HLH) is a life-threatening condition characterized by excessive immune activation, where prompt recognition and initiation of treatment is essential for the survival. HLH has many overlapping clinical features with sepsis and systemic inflammatory response syndrome (SIRS). This study was designed to determine if plasma biomarkers could differentiate HLH from other inflammatory conditions. We analyzed inflammatory protein profiles in the plasma of children with active HLH using the Luminex platform. We compared concentrations of 135 analytes in a discovery set, which consisted of 32 patients with HLH to 23 pediatric patients with severe sepsis or SIRS. The significant analytes were validated in an independent cohort, which consisted of 31 HLH patients and 24 patients with severe sepsis or SIRS. Comparisons were made based on the disease status and a classifier was built using various machine learning algorithms to see if the plasma protein profile could predict whether a patient had HLH or Sepsis/SIRS. Fourteen analytes were found to be significantly different between HLH patients and patients with either sepsis or SIRS in the discovery set and validated in the validated set (FDR = 0.1). A support vector machine classifier was created using six analytes (CXCL9, CXCL10, CXCL11, MIF, TARC, and ENA-78) that was able to achieve a sensitivity of 0.903 and a specificity of 0.833 to differentiate HLH from sepsis/SIRS in an independent validation cohort. Additionally, elevated pre-therapy CCL20 was associated with risk of death in the HLH cohort. Patients with HLH are at high risk of death without prompt initiation of immune suppression, where such an intervention may be contraindicated in other conditions. This unbiased evaluation of plasma proteins demonstrated that HLH in fact has a distinct inflammatory profile compared to severe sepsis and SIRS. An "HLH classifier" could be a clinically useful tool to assist with efficient and accurate diagnosis of critically ill children with hyperinflammatory syndromes.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.