Aims & Objectives:
To explore the role of serum ferritin in discriminating HLH from sepsis in children admitted to paediatric intensive care unit
A retrospective study involving children admitted to PICU in an academic children’s hospital over 5 years (2012–2017). Local unit and laboratory database accessed to identify children with a diagnosis of sepsis on discharge/death or had serum ferritin measurement for any reason.
A total of 112 children were identified. Out of these, 4 children admitted with presumed sepsis had a final diagnosis of HLH (diagnosed as per revised 2004 HLH diagnosis criteria). 64 children had a final diagnosis of sepsis (culture positive n=28), 23 children had cardiac pathology and 21 had miscellaneous diagnosis. The mean ranks of Ferritin (p = .009), Haemoglobin (p = .049), WCC (p = .007), Neutrophils (p = .004), Platelets (p = .029), Fibrinogen (p = .01) and ALT (p = .025) were statistically significantly different between the HLH and sepsis groups (see figure)
ROC curve for serum ferritin had an AUC of 0.96 (p =.003). ROC curve suggested a serum ferritin value of 5900 micrograms/litre was able to discriminate for HLH with 75% sensitivity, 98.1% specificity, positive LR 37.5 and a negative LR ratio of 0.25.
Serum ferritin estimation is a cheap, easy, rapid and widely available test showing promise in discriminating HLH from sepsis in critically ill children. Serum ferritin estimation should be incorporated in sepsis bundles to aid early diagnosis of HLH following confirmation of similar results in larger multi-center studies (underway).