Work

P806: Impact of Marrow Hemophagocytosis in Diagnostic Definition and Prognostic Significance in Adult Population

Year: 2022
Type: conference-abstract
Study design: Observational Study
Abstract: Background: Hemophagocytic lymphohistiocytosis (HLH) is associated with high mortality and is an entity difficult to diagnose. Bone marrow hemophagocytosis (BMHF) is the most easily accessible histomorphological criterion, but per si is neither specific nor sensitive for the diagnosis of HLH and its clinical significance is still not clear, mainly in adult population Aims: To evaluate the clinical and prognostic significance of BMHF. Methods: Retrospective analysis of adult patients with documentation of BMHF between 2007 and 2020. The identification of the patients was carried out by searching the database by the following keywords: “phagocytosis”, “phagocyte” and “histiocyte”. For the definition of HLH, the HLH-2004 and the revised HLH-2009 criteria were applied. HScore values ​​>169 were considered highly suggestive of HLH. Stratified models for each of the risk scores (HLH-2004 and HLH-2009) were compared using Akaike’s information criterion (AIC) and the Harrell’s concordance index (C-index). Results: Ninety-five patients (pts) were included, 63.2% males with median age of 64 (20-95) years old. Twenty-two pts (23.2%) presented ≥5 HLH-2004 criteria and 24 (25.3%) HLH-2009 criteria, with only 6 patients (27.3%) with simultaneous diagnosis by both criteria (Cohen’s kappa 0.0253). Forty-three (45.3%) patients presented ≥3 HLH-2004 criteria and 34 pts (35.8%) had Hscore>169. After a median follow-up of 24.3 months, the median overall survival (OS) was 39.4 months. OS was significantly lower in patients with ≥5 HLH-2004 (0.9 vs 86.5 M; HR 3.56; p<0.001), with ≥3 HLH-2004 criteria (3.17 vs NR; HR 2.95; p<0.001), with HLH-2009 criteria (5.5 vs 78.0 M; HR 1.89; p=0.034), and HScore>169 (1.45 vs 86.5; HR 3.08; p<0.001), compared with those that do not have. Excluding variables included in the above scores, the only clinical/analytical variable that was statistically significant for OS in univariate analysis was LDH>UNL (HR 1.02; p<0.001). After multivariate analysis for age, aetiology and LDH>ULN, ≥5 HLH-2004 (HR 4. p<0.001), HLH-2009 2009 (HR 1.87 p=0.045), ≥3 HLH-2004 (HR 3.06; p<0.001) and Hscore>169 (HR 3.39; p=0.001), the four scores retain significance for OS. In patients with BMHF, HLH-2004 criteria discriminated best between patients with poor and favourable OS than HLH-2009 (C-index 0.6364 vs 0.5601; AIC 400.24 vs 411.53). Concerning the aetiology, 51.6% was neoplastic (24.5% [n=12] with ≥5 HLH-2004 criteria), 13.7% infectious, 28.4% autoimmune and in 28,4% the trigger was not identified. Summary/Conclusion: In our cohort, HLH-2004 criteria had the best predictive value for OS, compared with HLH-2009 However, in the absence of the data need to calculate these scores, both ≥3 HLH-2004 criteria and Hscore>169 are independent predictors of OS, which, due to the deleterious prognosis of HLH, support their utility to start treatment earlier
Source: HemaSphere
Language: English
Subfield: Hematology
Field: Medicine
Open Access status: gold
Estimated APC paid: $2,480
OpenAlex ID: W4283397342
Cites:
Cited by:
Related to: 10