Encouraging Outcomes in Pediatric Malignancy‐Associated Hemophagocytic Lymphohistiocytosis: Results From the Italian <scp>HLH</scp> Registry

Hemophagocytic lymphohistiocytosis (HLH) is a life-threatening hyperinflammatory syndrome resulting from uncontrolled activation of cytotoxic T cells and macrophages [1]. HLH can arise from inherited genetic defects affecting the cytotoxic pathway (primary HLH, pHLH) or develop as a secondary disorder triggered by infections, autoimmune diseases, or malignancies (secondary HLH, S-HLH) [1-4]. In adults, malignancy-associated HLH (M-HLH) is the most frequent form of S-HLH, typically occurring in the context of hematological malignancies, especially lymphomas, and portends a poor prognosis [5-7]. Conversely, data on pediatric M-HLH is limited, largely derived from small, heterogeneous case series [8, 9], which hinders the recognition and management of this condition. To address this gap, we conducted a multicenter retrospective study within the Italian Association for Pediatric Hematology and Oncology, aiming to define the clinical spectrum, therapeutic strategies, and outcomes of children with M-HLH, compared with a well-characterized cohort of pediatric S-HLH not related to malignancy diagnosed in the same period (control group). We included patients enrolled in the Italian HLH Registry, a nationwide database that prospectively collects HLH cases. We retrospectively analyzed children (< 18 years) diagnosed with HLH between January 2009 and December 2024. We included patients with an associated hematological malignancy and a minimum follow-up of 6 months. Clinical, laboratory, and histopathological data of patients and controls were extracted from the Registry and validated by treating physicians. Variables collected included demographics, underlying malignancies, HLH diagnostic criteria [10], laboratory values, treatments and responses, and survival. Categorical variables were reported as numbers and percentages, whereas continuous variables were reported as medians and interquartile ranges (IQRs). Comparisons between groups were made using the Mann–Whitney U test for continuous variables and the chi-square test for categorical variables. The survival rates were estimated using a Kaplan–Meier model. The log-rank test was used to compare survival between groups. A two-sided p value ≤ 0.05 was considered to indicate statistical significance. The study was conducted in accordance with the Declaration of Helsinki and its later amendments [11]. Out of 76 screened patients, five were excluded due to insufficient follow-up, four because HLH developed after hematopoietic cell-stem transplantation (HSCT), and 12 for lack of clinical data. Finally, 55 children with M-HLH were included (Figure 1A). Thirty-three patients were male (60%), and the median age at HLH diagnosis was 107 months (IQR, 48–150; Table 1). The underlying malignancies included acute lymphoblastic leukemia (ALL, n = 31, 56%), acute myeloid leukemia (AML, n = 9, 16%), non-Hodgkin lymphoma (NHL, n = 7, 13%), Hodgkin lymphoma (n = 4, 7%), and multisystem Langerhans cell histiocytosis (n = 4, 7%) (Figure 1B). The median interval between malignancy diagnosis and HLH onset was 25 days (IQR, 1–131). Half the patients (n = 27) developed HLH during chemotherapy, while in 25 patients HLH developed at malignancy onset, and in the remaining three HLH emerged at relapse or disease progression. At HLH diagnosis, most patients had fever (n = 52, 95%), bi-lineage cytopenia (n = 54, 98%), hyperferritinemia (n = 53, 96%), and triglyceride and/or fibrinogen abnormalities (n = 47, 85%). Splenomegaly (n = 37, 67%), liver involvement (n = 33, 60%), and hemophagocytosis (n = 28, 51%) were less frequently observed (Table 1). Functional studies (i.e., PRF expression and/or CD107a expression) showed abnormalities in nine out of 20 tested patients (45%), while of the 34 patients tested for genetic predisposition to HLH, 10 carried monoallelic variants in pHLH-related genes and one a GATA2 mutation (Figure 1C). When compared with patients with S-HLH, those with M-HLH were significantly older (p < 0.001) and more frequently presented with liver involvement (p = 0.014) and bi-lineage cytopenia (p = 0.010) (Table 1). Thirty-six patients (65%) received HLH-directed treatment, including etoposide in eight, anti-IL1 in 5, and corticosteroids in 23. The overall response rate to HLH-directed therapy was 70% (Figure 1D). HLH reactivation occurred in 5 patients (9%), while 13 underwent HSCT, in most cases due to malignancy (85%). At a median follow-up of 20 months (IQR, 6–71) since HLH diagnosis, 16 patients (32%) died. Causes of death included malignancy in seven patients (44%), HLH in six (38%), and HSCT-related complications in three (18%). The estimated 5-year overall survival probability in children with M-HLH was 76% (95% CI 60–86). No significant differences were observed in survival probability between M-HLH and S-HLH patients (p = 0.190; Figure 1E). Of note, the six patients who died from HLH all had hepatic involvement and rapidly died (median time since diagnosis 29 days) from multiorgan failure despite HLH-directed therapy. In this study, we reported the clinical presentation and outcome of a large cohort of children with M-HLH, as compared with a cohort of S-HLH not related to malignancy. The clinical presentation of M-HLH in children was often non-specific and overlapped with the manifestations of the underlying malignancy. Fever, cytopenias, and hepatic dysfunction are not distinctive, frequently leading to delayed recognition, a limitation to effective intervention. When compared with S-HLH, the only differences of M-HLH patients were older age and more frequent hepatic involvement and bi-lineage cytopenia, findings that might reflect the underlying disease. In contrast, the clinical presentation of pediatric M-HLH diverged from that reported for adult M-HLH, which is usually very severe and related to lymphomas. Conversely, pediatric cases were usually associated with acute leukemia, with frequent FHL-associated genetic variants or functional abnormalities. The potential contribution of these factors to HLH development reinforces the current concept of a disease continuum to which different predisposing and environmental factors converge [1]. Current therapeutic strategies for M-HLH are largely adapted from the HLH-94 and -2004 protocols [12, 13]. Malignancy-directed regimens in children usually contain agents active against HLH—such as corticosteroids, etoposide, and cyclophosphamide—and the HLH component may respond favorably to such protocols [8]. Moreover, rituximab may provide additional benefit in EBV-driven or CD20-positive malignancies [9]. Response to therapy in our cohort was heterogeneous. Patients treated with etoposide seemed to portend poorer outcomes than those receiving corticosteroids alone. Although this is likely to reflect the selection of etoposide for more severe cases rather than a true difference in efficacy, we were not able to assess whether the addition of etoposide to already intensively treated patients contributed to their unfavorable evolution. Of note, novel molecules such as ELA026 have shown promising results in treating adults with M-HLH [14], and their use in the pediatric setting is currently under evaluation (NCT05416307). Survival in our cohort was comparable between M-HLH and S-HLH, but pediatric M-HLH outcomes appeared superior to those reported in adults, who have mortality rates exceeding 70% [5-7]. In addition, deaths in pediatric M-HLH were more related to malignancy or treatment toxicity rather than HLH activity. Fatal HLH cases were characterized by a rapid, fulminant course with hepatic involvement, despite intensive treatment. Recently, in adults with M-HLH, the OHI score was shown to identify patients at high mortality risk [15], but we were not able to assess the reproducibility of the score in children due to insufficient data on soluble CD25. In conclusion, M-HLH in children has a slightly distinct presentation from S-HLH and a more favorable outcome as compared to its adult counterpart. Implementation of diagnostic tests and individualized management, also including novel agents, could contribute to improving care in these patients. F.P., M.L.C., and E.S. conceived the study. F.P., I.F., and E.S. drafted the manuscript. F.P., S.G., R.M.D., R.M., E.P., M.Z., G.A., A.T., C.D.F., S.N., V.B., F.T., M.P., C.R., and E.S. followed the patients. A.C., L.B., and M.L.C. performed the cytometry and molecular analyses. F.P. performed the statistical analysis. All authors critically reviewed the manuscript and approved it in its final version. The authors have nothing to report. The authors declare no conflicts of interest. The data that support the findings of this study are available from the corresponding author upon reasonable request.