Outcome analysis of pediatric hemophagocytic lymphohistiocytosis.

Purpose: Hemophagocytic lymphohistiocytosis (HLH) is a life-threatening disease associated with rapid clinical deterioration and the need for intensive care; therefore, it is essential to identify clinical parameters related to mortality and establish prognostic factors correlated with unfavorable outcome in high risk patients whose treatment may fail.

Methods: Between January 2004 and December 2018, a total of 51 pediatric patients (less than 18 years old) who fulfilled the diagnostic criteria of HLH-2004 with documented results of bone marrow investigations at Kaohsiung Chang Gung Memorial Hospital were enrolled. The treatment protocol was based on hemophagocytic lymphohistiocytosis-94 (HLH-94) and HLH-2004. We retrospectively reviewed electronic medical records (EMR) including clinical features, length of intensive care unit (ICU) stay, serological tests, microscopic reports of bone marrow examination, and ultrasound examination reports at diagnosis to identify prognostic factors. The patients were divided into four groups based on etiology; these included infection associated hemophagocytic syndrome (IAHS), macrophage activation syndrome (MAS), malignancy associated hemophagocytic lymphohistiocytosis (MA-HLH), and idiopathic hemophagocytic lymphohistiocytosis (IHLH) to identify differences among the groups.

Results: Out of 51 patients enrolled, 27 patients had IAHS, 12 MAS, 8 MA-HLH, and 4 IHLH. The median age at diagnosis was 7 years. The overall mortality rate was 15.7% (there was no mortality in the MA-HLH group); the mean length of ICU stay was 6 ± 20.8 days. Longer activated partial thromboplastin time (aPTT) (p = 0.007), lower sodium concentration (p = 0.0007), and higher creatinine (p = 0.032) and aspartate aminotransferase (AST) (p = 0.017) were significantly related to mortality. Multivariate Cox regression analysis demonstrated that aPTT (p = 0.045, HR = 1.03, 95% CI = 1.0-1.1) was an independent risk factor for mortality. The receiver operating characteristic (ROC) curve showed that aPTT longer than 44.35 s was the cutoff value predicting mortality, with a sensitivity and specificity of 72% and 66.7%, respectively.

Conclusion: MA-HLH had the lowest mortality rate, as most children died from the underlying malignant disease and not from HLH. Impaired liver and renal functions were related to mortality. Prolonged aPTT > 44.35 s is a strong predictive factor for mortality.