Risk of hemophagocytic lymphohistiocytosis in adults with fevers of unknown origin: the clinical utility of a new scoring system on early detection. Issue 4 (28th July 2016)

• Record Type: Journal Article
• Authors: Wang, Hao‐Yuan; Yang, Ching‐Fen; Chiou, Tzeon‐Jye; Yang, Sheng‐Hsiang; Gau, Jyh‐Pyng; Yu, Yuan‐Bin; Chen, Po‐Min; Hsu, Hui‐Chi; Fung, Chang‐Phone; Lin, Hsiao‐Yi; Tzeng, Cheng‐Hwai; Liu, Jin‐Hwang; Hsiao, Liang‐Tsai
• Is Part Of: Hematological oncology. Volume 35:Issue 4(2017:Dec.)
• Page Start: 835
• Page End: 844
• Publication Date: 2016-07-28
• Subjects: bone marrow biopsy -- bone marrow score -- fever of unknown origin -- hemophagocytic lymphohistiocytosis -- non‐Hodgkin lymphoma; Hematological oncology -- Periodicals; Hematology; Medical Oncology; 616.99418005
• DOI: 10.1002/hon.2333
• Languages: English
• ISSNs: 0278-0232

Abstract

Abstract: The diagnosis of hemophagocytic lymphohistiocytosis (HLH) is delayed by most physicians. This study aimed to identify early parameters and suitable scoring systems for the risk of HLH. Clinical and laboratory data collected ≤3 days after admission were defined as early parameters and used to calculate the number of HLH‐2004 criteria met and bone marrow (BM) score. Between January 2006 and February 2016, 233 immunocompetent adults with naïve fever of unknown origin who underwent a BM study were enrolled to mimic patients at risk of HLH and randomly assigned into the developmental or validation cohort. Hemophagocytic lymphohistiocytosis was finally diagnosed in 47 patients, with non‐Hodgkin lymphoma as the major etiology (51.1%). Upon admission, four‐fifths of patients who developed subsequent HLH fulfilled ≤3 of 8 HLH‐2004 criteria, and 6 early parameters were independent predictors of HLH: anemia (hemoglobin < 10 g/dL), thrombocytopenia (platelet count < 100 × 10 3 /μL), leukoerythroblastosis, hyperbilirubinemia (total bilirubin > 2 × upper normal limit), hyperferritinemia (ferritin > 1000 ng/mL), and splenomegaly. Compared with the HLH criteria met upon admission, the BM score was an independent predictor (odds ratio = 1.621; 95% confidence interval, 1.355‐1.940) with excellent discrimination (area under the receiver operating characteristic curve = 0.920; 95% confidence interval, 0.883‐0.958). The sensitivity and specificity for a BM score cutoff of 10 points were 95% and 75%, respectively. When approaching immunocompetent adults with a continuously high fever, the BM score at initial admission assists with early identification of patients at risk of HLH.