Peritoneal effusion as a marker of disease severity in secondary hemophagocytic lymphohistiocytosis: Clinical and immunological insights Authors: Yansheng Wang, Lulu Zhang, Huiping Wang, Qian Li, Furun An, Beibei Xie, Shuangwei Ying, Jinxiao Hou, Hui Qin, Zhimin Zhai, Fan Wu Published: 2026 ABSTRACT ObjectiveHemophagocytic lymphohistiocytosis (HLH) is a life-threatening hyperinflammatory syndrome with cytokine storm-mediated multiorgan damage. Although serous effusions are frequent complications, the prognostic significance of peritoneal effusion in secondary HLH (sHLH) remains undefined. This study aimed to determine the incidence, clinical and immunological correlates, and prognostic value of peritoneal effusion in sHLH patients.MethodsWe retrospectively analyzed 107 adult patients diagnosed with sHLH at the Second Affiliated Hospital of Anhui Medical University between March 2013 and March 2023. Peritoneal effusion was diagnosed based on abdominal ultrasound, computed tomography, or positron emission tomography–computed tomography at initial presentation. Clinical characteristics, laboratory parameters, and lymphocyte subset profiles were compared between patients with and without peritoneal effusion. Logistic regression was used to identify risk factors for peritoneal effusion. Overall survival (OS) was estimated using the Kaplan–Meier method, and independent prognostic factors were assessed using multivariate Cox regression analysis.ResultsPeritoneal effusion was found in 24/107 (22.4%) sHLH patients and was associated with lower white blood cell, lymphocyte, monocyte, and platelet counts. Lower monocyte counts, higher direct bilirubin, and distinct immune dysregulation (reduced CD4+ T cells, higher NK cell percentages) were identified as risk factors for peritoneal effusion. Peritoneal effusion correlated with concurrent serous effusions (p < 0.05) and markedly shorter median survival (19 vs. 103 days, p = 0.002). Multivariate analysis identified peritoneal effusion as an independent mortality predictor (HR=1.88; 95% CI 1.09-3.24, p = 0.023).ConclusionsPeritoneal effusion is an independent predictor of mortality in sHLH patients and is associated with a distinct immune profile. Its presence should alert clinicians to a higher risk of death and prompt early, aggressive therapeutic intervention. Introduction Hemophagocytic Lymphohistiocytosis (HLH) is a rare and life-threatening clinical syndrome characterized by uncontrolled and persistent activation of cytotoxic T cells and natural killer (NK) cells. The failure of immune response regulation results in increased secretion of inflammatory cytokines, macrophage activation, and subsequently leads to a systemic inflammatory response and multiple organ dysfunction. 1 HLH is primarily classified into primary and secondary types, with secondary HLH (sHLH) typically triggered by infections, malignancies or autoimmune diseases. 2 Without appropriate treatment, the mortality rate of HLH approaches nearly 80%.3,4 In adult patients, current first-line treatment strategies, adapted from pediatric protocols such as HLH-94 and HLH-2004, achieve complete disease resolution in only approximately 50% of cases. 5 It is clinically characterized by persistent high fever, hepatosplenomegaly, cytopenia, elevated levels of ferritin and transaminases, and coagulopathy. * Experimental and epidemiological evidence suggests that HLH frequently presents with multiple effusions in the serous cavities.6,7 Recent clinical studies have reported an incidence of pleural effusion approaching 60%, which is independently associated with poor survival outcomes in sHLH patients. 8 Nonetheless, there is currently a lack of literature investigating the relationship between sHLH and peritoneal effusion. We hypothesized that peritoneal effusion predicts poor survival in sHLH patients. Furthermore, we postulated that its presence might be associated with a distinct pattern of immune dysregulation. This study therefore aimed to determine the incidence, clinical and immunological correlates, and prognostic value of peritoneal effusion in sHLH patients. Materials and methods Patients A total of 155 consecutive adult patients diagnosed with sHLH were retrospectively screened at the Second Affiliated Hospital of Anhui Medical University between March 2013 to March 2023. The inclusion criteria were as follows: (1) patients over 18 years old; (2) five of eight following criteria are required according to HLH-2004 diagnostic guidelines: 1) fever; 2) splenomegaly; 3) cytopenias (affecting ≥ 2 lineages, hemoglobin < 90g/L, platelets < 100×109/L, neutrophils < 1.0×109/L; 4) hypofibrinogenemia and/or hypertriglyceridemia (<150mg/dL); 5) hemophagocytosis in bone marrow, spleen, lymph nodes, liver, or other tissue; 6) low or absent NK-cell activity; 7) ferritin > 500 ng/mL, and 8) elevated CD25 (soluble IL-2 receptor) > 2400 U/ml or elevated based on the laboratory-defined normal range. 9 48 patients were excluded based on the following exclusion criteria: (1) Comorbidities potentially causing peritoneal effusion, including liver cirrhosis, congestive heart failure, cancer, nephrotic syndrome, renal failure (estimated creatinine clearance below 15 mL/min per 1.73 m2), acute/chronic pancreatitis, tuberculous peritonitis, and portal hypertension-related conditions; (2) History of drugs and operation known to cause peritoneal effusion at admission: methotrexate, glucocorticosteroids, isoniazid, rifampin, abdominal surgery and/or radiation therapy; (3) Refusal to receive any therapeutic interventions. Consequently, 107 newly diagnosed patients were included in the final analysis. As this was a retrospective study spanning a decade, a formal a priori sample size calculation was not performed. The final cohort comprised all eligible patients during this period. This sample size provided sufficient data for the robust multivariate analyses conducted in this study. Overall survival (OS) was the primary endpoint, defined as the time from the date of diagnosis to the date of death from any cause or the last follow-up (October 30, 2024). The survival status of all patients was confirmed using death records or directly reaching out to the patients themselves. For those who died during the follow-up period, confirmation was obtained by contacting their next of kin. Data collection Baseline clinical characteristics and laboratory data were gathered through a review of medical records. The following variables were collected: age, sex, etiologies, complete blood cell counts, ferritin, triglycerides (TG), fibrinogen (FIB), lactate dehydrogenase (LDH), albumin (ALB), globulin (GLB), albumin/globulin ratio (AGR), alanine aminotransferase (ALT), aspartate transaminase (AST), direct bilirubin (DBil), interleukin-6, Epstein-Barr virus (EBV) infection, presence or absence of organ enlargement, central nervous system (CNS) symptom, radiographic reports, lymphocyte subset data and therapeutic regimens. Complete lymphocyte subset data were available for 58 (54.2%) of the 107 patients and contained no missing values. All analyses involving immune parameters are based on this subgroup. The diagnosis of peritoneal effusion was based on findings from abdominal ultrasound, computed tomography (CT), and positron emission tomography/computed tomography (PET-CT). All radiographic assessments were performed at initial diagnosis prior to treatment initiation. To minimize misclassification bias, discordant interpretations underwent blinded independent review by two experienced abdominal radiologists. Only patients with unequivocal fluid detected on at least one imaging modality were classified as having peritoneal effusion. Patients were classified into two groups based on the presence of peritoneal effusion at the time of sHLH diagnosis: the peritoneal effusion positive (PE+) group and the peritoneal effusion negative (PE−) group. Follow-up was conducted through medical record reviews or phone calls. Statistical analysis Categorical variables were presented as frequencies and percentages. As our data followed a non-normal distribution, continuous variables were presented as medians with Q1-Q3 ranges. Categorical variables were analyzed using the Pearson chi-square test or Fisher’s exact test. The Mann-Whitney U test was used to compare continuous variables. Univariate logistic regression was used to evaluate risk factors for peritoneal effusion, and factors with a p-value less than 0.1 were included in the subsequent multivariate regression analysis. Receiver operating characteristic (ROC) curve analysis was performed to identify the optimal cutoff values for parameters associated with maximum sensitivity and specificity. The prognostic value was assessed using Kaplan-Meier survival curves. Variables that had a p-value < 0.05 in the univariate analysis were incorporated into the Cox proportional hazards regression model. Sensitivity analyses were performed, including stratified survival analysis by HLH etiology and assessment of potential selection bias for the lymphocyte subset cohort, to evaluate the robustness of the findings. A two-sided p-value < 0.05 was considered statistically significant. Statistical analyses and illustrations were performed using SPSS, version 22.0 (Chicago, IL, USA) and R software version 4.2.1. Results Baseline characteristics Among 155 patients hospitalized with sHLH during the study period, 48 were excluded according to predefined criteria: 10 due to age <18 years, 8 who declined therapeutic interventions,18 lacking essential radiographic documentation for peritoneal effusion assessment, and 12 lost to follow-up (Figure 1). The final cohort included 107 patients, comprising 57 (53.3%) males and 50 (46.7%) females. The basic clinical characteristics of the cohorts are presented in Table 1. The median age was 49 (Q1-Q3, 33-62) years. Hematologic malignancies were the most common triggers (51.4%), followed by infections (22.4%) and autoimmune diseases (13.1%), 14 patients (13.1%) had unidentified triggers at discharge. At diagnosis, 19 (17.8%) patients exhibited CNS symptoms, including drowsiness, altered consciousness, limb numbness, muscle weakness, seizures, headaches, and blurred vision. Imaging revealed lymphadenopathy in 78 (72.9%) patients, splenomegaly in 73 (68.2%) patients, and hepatomegaly in 18 (16.8%) patients. Bone marrow examinations were performed in 99 patients, with 78 (78.8%) patients showing hemophagocytosis. Laboratory tests indicated significant abnormalities in hematological indices, biochemical markers, and coagulation function, including white blood cell count (2.30 × 109/L, Q1-Q3 1.28-3.82), lymphocyte count (0.56 × 109/L, Q1-Q3 0.33-0.92), hemoglobin (86.00 g/L, Q1-Q3 68.00-101.00), platelet count (43.00 × 109/L, Q1-Q3 23.00-75.00), albumin (26.20 g/L, Q1-Q3 24.10-29.20), and fibrinogen (1.15 g/L, Q1-Q3 0.77-2.24). Among these, 36 patients had received various kinds of chemotherapy, the other 51 patients received only the HLH-94 or HLH-04 protocol or steroid and/or IVIg, The reasons patients did not receive chemotherapy included early death while receiving HLH-directed therapy, or being unsuitable for intensive treatment due to progressive multiple organ dysfunction. By the end of follow-up, 83 patients (77.6%) had died.Figure 1.Diagram of enrolled patients.Table 1.Demographic and clinical characteristics of sHLH patients with or without peritoneal effusion.Characteristics, n (%)Total (n = 107)PE+ (n = 24)PE− (n = 83)p valueDemographical characteristics Sex (male/female)57 (53.3)/50 (46.7)13 (54.2)/11 (45.8)44 (53.0)/39 (47.0)>0.99 Age (years)49 (33, 62)48 (38, 63)53 (33, 62)0.84Underlying triggering factor Etiologies​​​0.08 Infections24 (22.4)4 (16.7)20 (24.1)​ Hematologic malignancies55 (51.4)16 (66.7)39 (47.0)​ Lymphoma53 (49.5)14 (58.3)39 (47.0)​ Autoimmune14 (13.1)0 (0.0)14 (16.9)​ Unknown14 (13.1)4 (16.7)10 (12.0)​CNS symptom19 (17.8)4 (16.7)15 (18.1)>0.99Organ enlargement Liver18 (16.8)6 (25.0)12 (14.5)0.23 Spleen73 (68.2)17 (70.8)56 (67.5)0.76 Lymph nodes78 (72.9)17 (70.8)61 (73.5)0.80Laboratory examinations White blood cell (×109/L)2.30 (1.28, 3.82)1.64 (1.14, 2.74)2.53 (1.33, 4.91)0.024 Neutrophil (×109/L)1.44 (0.63, 2.43)0.86 (0.54, 2.05)1.62 (0.63, 3.32)0.09 lymphocyte (×109/L)0.56 (0.33, 0.92)0.37 (0.16, 0.61)0.64 (0.37, 0.95)0.005 Monocyte (×109/L)0.21 (0.09, 0.42)0.11 (0.08, 0.22)0.26 (0.10, 0.47)0.013 Hemoglobin (g/L)80.0 (68.0, 101.0)74.5 (61.8, 98.8)83.0 (68.0, 102.0)0.18 Platelet (×109/L)43 (23, 75)24 (15, 45)46 (24, 80)0.008 RDW (fl)16.1 (14.6, 17.8)15.6 (13.9, 17.8)16.5 (14.7, 17.8)0.38 ALT (U/L)89 (50, 172)90 (44, 173)89 (51, 172)0.82 AST (U/L)143 (55, 326)153 (70, 472)124 (51, 284)0.31 DBil (µmol/L)8.7 (5.0, 32.7)17.7 (5.5, 69.5)8.2 (4.7, 24.8)0.09 LDH (U/L)879 (453, 2114)1181 (795, 2147)818 (435, 2114)0.39 ALB (g/L)26.2 (24.1, 29.2)26.3 (24.3, 29.0)26.2 (23.7, 29.4)0.88 GLB (g/L)26.0 (21.8, 30.0)23.6 (20.2, 29.5)26.5 (22.0, 30.5)0.28 AGR1.0 (0.8, 1.3)1.2 (0.9, 1.3)1.0 (0.8, 1.3)0.28 TG (mmol/L)2.72 (1.64, 3.71)2.74 (1.50, 3.13)2.71 (1.67, 3.82)0.45 IL-6 (pg/ml)40 (20, 96)36 (20, 156)42 (20, 96)0.96 FIB (g/L)1.15 (0.77, 2.24)1.01 (0.71, 2.02)1.20 (0.79, 2.78)0.34 Ca2+ (mmol/L)1.93 (1.83, 2.05)1.97 (1.86, 2.06)1.92 (1.82, 2.03)0.21 Ferritin (µg/L)6700 (2015, 21300)4052 (1667, 24521)8369 (2369, 20801)0.46 EBV-DNA≥103 copies/mL44/86 (51.2)7/18 (38.9)37/68 (54.4)0.24Hemophagocytosis78/99 (78.8)20/24 (83.3)58/75 (77.3)0.53Treatment​​​0.57 Chem±HLH-94±HLH-04 (%)76 (71.0)19 (79.2)57 (68.7)​ GC ± IVIg (%)30 (28.0)5 (20.8)25 (30.1)​ Support (%)1 (0.9)1 (4.2)0 (0)​Death (%)83 (77.6)22 (91.7)61 (73.5)0.06CNS: central nervous system; RDW: red cell distribution width; ALT: alanine transaminase; AST: aspartate transaminase; DBil direct bilirubin; LDH: lactic dehydrogenase; ALB: albumin; GLB: globulin; AGR: albumin/globulin ratio; TG: triglyceride; FIB: fibrinogen; EBV-DNA: Epstein-Barr virus deoxyribonucleic acid; GC: glucocorticoid. Diagram of enrolled patients. Table 1.Demographic and clinical characteristics of sHLH patients with or without peritoneal effusion.Characteristics, n (%)Total (n = 107)PE+ (n = 24)PE− (n = 83)p valueDemographical characteristics Sex (male/female)57 (53.3)/50 (46.7)13 (54.2)/11 (45.8)44 (53.0)/39 (47.0)>0.99 Age (years)49 (33, 62)48 (38, 63)53 (33, 62)0.84Underlying triggering factor Etiologies​​​0.08 Infections24 (22.4)4 (16.7)20 (24.1)​ Hematologic malignancies55 (51.4)16 (66.7)39 (47.0)​ Lymphoma53 (49.5)14 (58.3)39 (47.0)​ Autoimmune14 (13.1)0 (0.0)14 (16.9)​ Unknown14 (13.1)4 (16.7)10 (12.0)​CNS symptom19 (17.8)4 (16.7)15 (18.1)>0.99Organ enlargement Liver18 (16.8)6 (25.0)12 (14.5)0.23 Spleen73 (68.2)17 (70.8)56 (67.5)0.76 Lymph nodes78 (72.9)17 (70.8)61 (73.5)0.80Laboratory examinations White blood cell (×109/L)2.30 (1.28, 3.82)1.64 (1.14, 2.74)2.53 (1.33, 4.91)0.024 Neutrophil (×109/L)1.44 (0.63, 2.43)0.86 (0.54, 2.05)1.62 (0.63, 3.32)0.09 lymphocyte (×109/L)0.56 (0.33, 0.92)0.37 (0.16, 0.61)0.64 (0.37, 0.95)0.005 Monocyte (×109/L)0.21 (0.09, 0.42)0.11 (0.08, 0.22)0.26 (0.10, 0.47)0.013 Hemoglobin (g/L)80.0 (68.0, 101.0)74.5 (61.8, 98.8)83.0 (68.0, 102.0)0.18 Platelet (×109/L)43 (23, 75)24 (15, 45)46 (24, 80)0.008 RDW (fl)16.1 (14.6, 17.8)15.6 (13.9, 17.8)16.5 (14.7, 17.8)0.38 ALT (U/L)89 (50, 172)90 (44, 173)89 (51, 172)0.82 AST (U/L)143 (55, 326)153 (70, 472)124 (51, 284)0.31 DBil (µmol/L)8.7 (5.0, 32.7)17.7 (5.5, 69.5)8.2 (4.7, 24.8)0.09 LDH (U/L)879 (453, 2114)1181 (795, 2147)818 (435, 2114)0.39 ALB (g/L)26.2 (24.1, 29.2)26.3 (24.3, 29.0)26.2 (23.7, 29.4)0.88 GLB (g/L)26.0 (21.8, 30.0)23.6 (20.2, 29.5)26.5 (22.0, 30.5)0.28 AGR1.0 (0.8, 1.3)1.2 (0.9, 1.3)1.0 (0.8, 1.3)0.28 TG (mmol/L)2.72 (1.64, 3.71)2.74 (1.50, 3.13)2.71 (1.67, 3.82)0.45 IL-6 (pg/ml)40 (20, 96)36 (20, 156)42 (20, 96)0.96 FIB (g/L)1.15 (0.77, 2.24)1.01 (0.71, 2.02)1.20 (0.79, 2.78)0.34 Ca2+ (mmol/L)1.93 (1.83, 2.05)1.97 (1.86, 2.06)1.92 (1.82, 2.03)0.21 Ferritin (µg/L)6700 (2015, 21300)4052 (1667, 24521)8369 (2369, 20801)0.46 EBV-DNA≥103 copies/mL44/86 (51.2)7/18 (38.9)37/68 (54.4)0.24Hemophagocytosis78/99 (78.8)20/24 (83.3)58/75 (77.3)0.53Treatment​​​0.57 Chem±HLH-94±HLH-04 (%)76 (71.0)19 (79.2)57 (68.7)​ GC ± IVIg (%)30 (28.0)5 (20.8)25 (30.1)​ Support (%)1 (0.9)1 (4.2)0 (0)​Death (%)83 (77.6)22 (91.7)61 (73.5)0.06CNS: central nervous system; RDW: red cell distribution width; ALT: alanine transaminase; AST: aspartate transaminase; DBil direct bilirubin; LDH: lactic dehydrogenase; ALB: albumin; GLB: globulin; AGR: albumin/globulin ratio; TG: triglyceride; FIB: fibrinogen; EBV-DNA: Epstein-Barr virus deoxyribonucleic acid; GC: glucocorticoid. Demographic and clinical characteristics of sHLH patients with or without peritoneal effusion. CNS: central nervous system; RDW: red cell distribution width; ALT: alanine transaminase; AST: aspartate transaminase; DBil direct bilirubin; LDH: lactic dehydrogenase; ALB: albumin; GLB: globulin; AGR: albumin/globulin ratio; TG: triglyceride; FIB: fibrinogen; EBV-DNA: Epstein-Barr virus deoxyribonucleic acid; GC: glucocorticoid. Distribution of peritoneal effusion and comparison of clinical and laboratory findings in patients with and without peritoneal effusion We reviewed the radiological reports and found that 24 (22.4%) sHLH patients presented with peritoneal effusion. Interestingly, no patients with autoimmune diseases exhibited peritoneal effusion (Table 1). Our analysis indicated significant differences in the occurrence rates of pleural, pelvic, and pericardial effusions between PE+ and PE− groups, suggesting higher probabilities of these conditions in patients with peritoneal effusion (Table S1). To evaluate the risk of peritoneal effusion, we analyzed demographic, clinical, and laboratory characteristics of patients with and without peritoneal effusion (Table 1). Demographic and clinical parameters, including sex, age, etiology, CNS symptoms, organ enlargement, and treatment regimen, did not significantly differ between groups. Laboratory tests showed significantly lower counts of white blood cell (p = 0.024), lymphocyte (p = 0.005), monocyte (p = 0.013), and platelet (p = 0.008) in PE+ patients. Direct bilirubin showed a trend toward elevation in PE+ patients (p = 0.09). Mortality was higher in PE+ patients (91.7% vs 73.5%), though this difference was not statistically significant (p = 0.06). We performed univariate and multivariate logistic regression analyses to assess associations between clinical characteristics and peritoneal effusion (Table 2). The results revealed that higher DBil (p = 0.023; OR 1.015; 95% CI 1.002-1.028) and lower monocyte counts (p = 0.021; OR 0.027; 95% CI 0.001-0.580) were independent risk factors. Lower hemoglobin levels approached significance (p = 0.08; OR 0.980; 95% CI 0.957-1.003). Treatment varied based on underlying etiology, but no significant differences were found between PE+ and PE− groups. Given the etiological homogeneity of the PE+ group - which contained no autoimmune cases and was predominantly composed of lymphoma-associated HLH (14/24, 58.3%) - and its limited sample size, a stratified analysis by treatment was not performed.Table 2.Univariate and multivariate logistics regression analysis of peritoneal effusion in sHLH patients.VariablesUnivariate analysisMultivariate analysisOR (95% CI)p valueOR (95% CI)p valueWhite blood cell (×109/L)0.763 (0.587-0.991)0.043​​Neutrophil (×109/L)0.752 (0.556-1.019)0.07​​lymphocyte (×109/L)0.314 (0.102-0.973)0.045​​Monocyte (×109/L)0.037 (0.002-0.635)0.0230.027 (0.001-0.580)0.021Hemoglobin (g/L)0.980 (0.960-1.001)0.070.980 (0.957-1.003)0.08Platelet (×109/L)0.978 (0.960-0.996)0.017​​RDW (fl)1.003 (0.853-1.178)0.97​​ALT (U/L)0.998 (0.995-1.002)0.30​​AST (U/L)1.000 (0.999-1.001)0.68​​DBil (µmol/L)1.013 (1.001-1.025)0.0331.015 (1.002-1.028)0.023Lg (LDH) (U/L)1.387 (0.464-4.143)0.56​​ALB (g/L)0.997 (0.907-1.096)0.95​​GLB (g/L)0.993 (0.936-1.053)0.81​​AGR1.540 (0.411-5.764)0.52​​TG (mmol/L)0.874 (0.651-1.175)0.37​​Interleukin-6 (pg/ml)1.000 (0.998-1.001)0.54​​FIB (g/L)0.703 (0.447-1.107)0.13​​Ca2+ (mmol/L)3.483 (0.245-49.580)0.36​​Lg (Ferritin) (µg/L)0.791 (0.399-1.566)0.50​​RDW: red cell distribution width; ALT: alanine transaminase; AST: aspartate transaminase; DBil direct bilirubin; LDH: lactic dehydrogenase; ALB: albumin; GLB: globulin; AGR: albumin/globulin ratio; TG: triglyceride; FIB: fibrinogen. Table 2.Univariate and multivariate logistics regression analysis of peritoneal effusion in sHLH patients.VariablesUnivariate analysisMultivariate analysisOR (95% CI)p valueOR (95% CI)p valueWhite blood cell (×109/L)0.763 (0.587-0.991)0.043​​Neutrophil (×109/L)0.752 (0.556-1.019)0.07​​lymphocyte (×109/L)0.314 (0.102-0.973)0.045​​Monocyte (×109/L)0.037 (0.002-0.635)0.0230.027 (0.001-0.580)0.021Hemoglobin (g/L)0.980 (0.960-1.001)0.070.980 (0.957-1.003)0.08Platelet (×109/L)0.978 (0.960-0.996)0.017​​RDW (fl)1.003 (0.853-1.178)0.97​​ALT (U/L)0.998 (0.995-1.002)0.30​​AST (U/L)1.000 (0.999-1.001)0.68​​DBil (µmol/L)1.013 (1.001-1.025)0.0331.015 (1.002-1.028)0.023Lg (LDH) (U/L)1.387 (0.464-4.143)0.56​​ALB (g/L)0.997 (0.907-1.096)0.95​​GLB (g/L)0.993 (0.936-1.053)0.81​​AGR1.540 (0.411-5.764)0.52​​TG (mmol/L)0.874 (0.651-1.175)0.37​​Interleukin-6 (pg/ml)1.000 (0.998-1.001)0.54​​FIB (g/L)0.703 (0.447-1.107)0.13​​Ca2+ (mmol/L)3.483 (0.245-49.580)0.36​​Lg (Ferritin) (µg/L)0.791 (0.399-1.566)0.50​​RDW: red cell distribution width; ALT: alanine transaminase; AST: aspartate transaminase; DBil direct bilirubin; LDH: lactic dehydrogenase; ALB: albumin; GLB: globulin; AGR: albumin/globulin ratio; TG: triglyceride; FIB: fibrinogen. Univariate and multivariate logistics regression analysis of peritoneal effusion in sHLH patients. RDW: red cell distribution width; ALT: alanine transaminase; AST: aspartate transaminase; DBil direct bilirubin; LDH: lactic dehydrogenase; ALB: albumin; GLB: globulin; AGR: albumin/globulin ratio; TG: triglyceride; FIB: fibrinogen. Differences of immune factors between PE+ and PE− patients In this study, 58 (54.2%) underwent lymphocyte subset assessment at initial diagnosis, divided into the PE+ group (n = 13) and the PE− group (n = 45) based on peritoneal effusion presence. To evaluate potential selection bias, we compared clinical data between the groups, including age, sex, etiology, clinical manifestations, and laboratory indicators. Statistical analysis revealed significantly higher ferritin levels in the lymphocyte subset testing group (P = 0.036), with no other significant differences (P > 0.05). Subsequently, we explored the status of lymphocyte subsets in sHLH patients. The median percentages were: total T cells 77.35% (65.98, 87.38), CD4+T cells 28.90% (18.65, 45.95), CD8+T cells 36.35% (18.93, 47.48), total B cells 4.90% (0.95, 14.20), NK cells 8.80% (4.20, 14.35), Regulatory T cells (Treg) 4.41% (2.51, 6.07) and effector T cells (Tef) 3.71% (1.48, 5.52). The median counts were: total T cells 503.50 (330.00, 866.50)/µL, CD4+ T cells 243.50 (115.50, 482.00)/µL, CD8+ T cells 222.50 (95.75, 525.25)/µL, total B cells 33.00 (4.50, 102.50)/µL, and NK cells 69.00 (23.00, 118.50)/µL. The median CD4+T/CD8+T cells ratio was 0.97 (0.55, 2.20), and the median Treg/Tef ratio was 1.04 (0.68, 1.82). To visualize the heterogeneity in immune cell counts between PE+ and PE− sHLH patients, we created a hierarchical clustering heatmap (Figure 2). Comparative analysis showed that the PE+ group had lower counts of total T cells, CD4+ T cells, and a lower percentage of CD4+ T cells, while NK cell percentages were higher (Figure 3A–(B)). There was also a trend toward increased Treg percentages (Figure 3(C)). No significant differences were observed in the CD4+ T/CD8+ T cell ratio or Treg/Teff cell ratio between the two groups (Figure 3(D)). To assess associations between lymphocyte subsets and peritoneal effusion presence, we included data from the univariate analysis of the entire population that showed significant results. Patients with peritoneal effusion had lower counts (p = 0.030) and proportions (p = 0.043) of CD4+ T cells, along with higher percentages of NK cells (p = 0.019). Of significance, higher percentages of NK cells (p = 0.019; OR 1.072; 95% CI 1.011-1.136) was found to be risk factors for developing peritoneal effusion. (Table S2).Figure 2.Lymphocyte subsets between HLH patients with or without peritoneal effusion were shown in a heatmap with unsupervised clustering analyses. Heatmap comparing 14 immune parameters in PE+ patients (n = 13) and PE− patients (n = 45). On the y-axis are immune parameter values after z-scoring, and on the x-axis are individual persons. Red-white-green squares represent z-scoring values. Blue square on the x-axis, PE+; Red square on the x-axis, PE−.Figure 3.Differences of immune factors between patients with or without peritoneal effusion. (A) Comparison of the counts of peripheral blood lymphocyte subsets between PE+ (n = 13) and PE− (n = 45) patients with newly HLH. (B) Comparison of the percentages of peripheral blood lymphocyte subsets between PE+ (n = 13) and PE− (n = 45) patients with newly HLH. (C) Comparison of the percentages of Treg cells and Teff cells between PE+ (n = 13) and PE− (n = 45) patients with newly HLH. (D) Comparison of the ratio of CD4+ T/CD8+ T cells and Treg/Teff cells between PE+ (n = 13) and PE− (n = 45) patients with newly HLH. Blue circle points represent PE+ group patients, and red circle points represent PE− group patients. Groups were compared using the Mann-Whitney U test. Treg: regulatory T cell; Tef: effector T cell. Lymphocyte subsets between HLH patients with or without peritoneal effusion were shown in a heatmap with unsupervised clustering analyses. Heatmap comparing 14 immune parameters in PE+ patients (n = 13) and PE− patients (n = 45). On the y-axis are immune parameter values after z-scoring, and on the x-axis are individual persons. Red-white-green squares represent z-scoring values. Blue square on the x-axis, PE+; Red square on the x-axis, PE−. Differences of immune factors between patients with or without peritoneal effusion. (A) Comparison of the counts of peripheral blood lymphocyte subsets between PE+ (n = 13) and PE− (n = 45) patients with newly HLH. (B) Comparison of the percentages of peripheral blood lymphocyte subsets between PE+ (n = 13) and PE− (n = 45) patients with newly HLH. (C) Comparison of the percentages of Treg cells and Teff cells between PE+ (n = 13) and PE− (n = 45) patients with newly HLH. (D) Comparison of the ratio of CD4+ T/CD8+ T cells and Treg/Teff cells between PE+ (n = 13) and PE− (n = 45) patients with newly HLH. Blue circle points represent PE+ group patients, and red circle points represent PE− group patients. Groups were compared using the Mann-Whitney U test. Treg: regulatory T cell; Tef: effector T cell. Correlation analysis of immune status and clinical outcomes of sHLH patients The optimal cutoff values for various lymphocyte subset parameters were determined using ROC curve analysis. Based on these cutoff values, we stratified the patients’ survival and conducted Kaplan-Meier survival analysis. The results revealed that patients with a higher percentage of NK cells (p = 0.042) and Treg cells (p = 0.041), but lower counts of total T (p = 0.003), CD4+ T (p = 0.001), CD8+ T (p = 0.016), and B cells (p = 0.001) were significantly associated with shorter OS (Figure 4).Figure 4.Kaplan-Meier survival curve of lymphocyte subsets in newly diagnosed HLH patients. (A) Survival curve grouped by T cells count. (B) Survival curve grouped by CD4+T cells count. (C) Survival curve grouped by CD8+T cells count. (D) Survival curve grouped by B cells count. (E) Survival curve grouped by NK cells percentage. (F) Survival curve grouped by Treg cells percentage. Kaplan-Meier survival curve of lymphocyte subsets in newly diagnosed HLH patients. (A) Survival curve grouped by T cells count. (B) Survival curve grouped by CD4+T cells count. (C) Survival curve grouped by CD8+T cells count. (D) Survival curve grouped by B cells count. (E) Survival curve grouped by NK cells percentage. (F) Survival curve grouped by Treg cells percentage. Univariate and multivariate analysis of factors related to prognosis in sHLH patients The median survival of OS was 71 days, with 22 (91.7%) deaths in the PE+ group compared to 61 (73.5%) in the PE− group. Kaplan-Meier survival analysis indicated that patients with peritoneal effusion experienced poor OS (median OS, 19 vs. 103 days, P = 0.002) (Fig S1A). To address potential confounding by malignancy triggers, we stratified patients into lymphoma-associated HLH (LHLH) and non-lymphoma HLH (non-LHLH). Peritoneal effusion remained significantly associated with poorer OS in both the LHLH (median OS, 19 vs. 65 days, p = 0.020) and non-LHLH (median OS, 17 vs. 187 days, p = 0.035) groups (Fig S1B-C). Due to a significant number of missing values in lymphocyte subset data, these variables were excluded from the Cox regression analysis. ROC analyses and the maximum Youden’s index were used to identify cutoff values for continuous variables in HLH patients. Univariate analysis showed that peritoneal effusion, LHLH, age > 28 years, RDW > 18.95 fl, DBil > 3.85 µmol/L, AGR > 0.97, and Ca2+ < 1.96 mmol/L were associated with poor outcomes, while patients with white blood cell counts > 1.80 × 109/L, neutrophil counts > 1.00 × 109/L, lymphocyte counts > 0.47 × 109/L, platelet counts > 26 × 109/L, ALT > 139 U/L, AST > 209 U/L, GLB > 26.65 g/L, FIB > 2.36 g/L, and ferritin < 3465 µg/L have better prognoses. The multivariate Cox model confirmed peritoneal effusion, age, and lymphocyte count as independent prognostic factors (Table 3).Table 3.Univariate and multivariate cox regression analyses of overall survival.VariablesUnivariate analysisMultivariate analysisHR95% CIp valueHR95% CIp valueWith pleural effusion1.250.80-1.940.32​​​With peritoneal effusion2.151.31-3.520.0021.881.09-3.240.023LHLH2.351.50-3.69<0.001​​​Age>28 years4.631.87-11.500.0012.851.08-7.520.034Splenomegaly0.870.55-1.380.56​​​white blood cell>1.80×109/L0.510.33-0.790.003​​​Neutrophil>1.00×109/L0.610.39-0.950.028​​​lymphocyte>0.47×109/L0.330.21-0.52<0.0010.340.18-0.660.001Monocyte>0.31×109/L0.660.41-1.060.09​​​Hemoglobin>89g/L0.640.41-1.020.06​​​Platelet>26×109/L0.510.32-0.800.003​​​RDW>18.95fl2.321.29-4.140.005​​​ALT>139U/L0.570.34-0.960.035​​​AST>209U/L0.770.48-1.220.27​​​DBil>3.85µmol/L2.071.09-3.910.026​​​LDH>1686U/L0.660.40-1.090.10​​​ALB>24.15g/L1.530.91-2.580.11​​​GLB>26.65g/L0.560.35-0.870.011​​​AGR>0.971.731.10-2.720.018​​​TG>3.35mmol/L0.770.48-1.240.29​​​IL-6>30.6pg/ml1.580.99-2.520.06​​​FIB>2.36g/L0.490.28-0.880.017​​​Ca2+>1.96mmol/L1.841.18-2.850.007​​​Ferritin<3465µg/L0.630.40-0.990.047​​​LHLH: lymphoma associated hemophagocytic lymphohistiocytosis; RDW: red cell distribution width; ALT: alanine transaminase; AST: aspartate transaminase; DBil direct bilirubin; LDH: lactic dehydrogenase; ALB: albumin; GLB: globulin; AGR: albumin/globulin ratio; TG: triglyceride; FIB: fibrinogen. Table 3.Univariate and multivariate cox regression analyses of overall survival.VariablesUnivariate analysisMultivariate analysisHR95% CIp valueHR95% CIp valueWith pleural effusion1.250.80-1.940.32​​​With peritoneal effusion2.151.31-3.520.0021.881.09-3.240.023LHLH2.351.50-3.69<0.001​​​Age>28 years4.631.87-11.500.0012.851.08-7.520.034Splenomegaly0.870.55-1.380.56​​​white blood cell>1.80×109/L0.510.33-0.790.003​​​Neutrophil>1.00×109/L0.610.39-0.950.028​​​lymphocyte>0.47×109/L0.330.21-0.52<0.0010.340.18-0.660.001Monocyte>0.31×109/L0.660.41-1.060.09​​​Hemoglobin>89g/L0.640.41-1.020.06​​​Platelet>26×109/L0.510.32-0.800.003​​​RDW>18.95fl2.321.29-4.140.005​​​ALT>139U/L0.570.34-0.960.035​​​AST>209U/L0.770.48-1.220.27​​​DBil>3.85µmol/L2.071.09-3.910.026​​​LDH>1686U/L0.660.40-1.090.10​​​ALB>24.15g/L1.530.91-2.580.11​​​GLB>26.65g/L0.560.35-0.870.011​​​AGR>0.971.731.10-2.720.018​​​TG>3.35mmol/L0.770.48-1.240.29​​​IL-6>30.6pg/ml1.580.99-2.520.06​​​FIB>2.36g/L0.490.28-0.880.017​​​Ca2+>1.96mmol/L1.841.18-2.850.007​​​Ferritin<3465µg/L0.630.40-0.990.047​​​LHLH: lymphoma associated hemophagocytic lymphohistiocytosis; RDW: red cell distribution width; ALT: alanine transaminase; AST: aspartate transaminase; DBil direct bilirubin; LDH: lactic dehydrogenase; ALB: albumin; GLB: globulin; AGR: albumin/globulin ratio; TG: triglyceride; FIB: fibrinogen. Univariate and multivariate cox regression analyses of overall survival. LHLH: lymphoma associated hemophagocytic lymphohistiocytosis; RDW: red cell distribution width; ALT: alanine transaminase; AST: aspartate transaminase; DBil direct bilirubin; LDH: lactic dehydrogenase; ALB: albumin; GLB: globulin; AGR: albumin/globulin ratio; TG: triglyceride; FIB: fibrinogen. Discussion Multiple effusions were often found in HLH patients, including pleural effusion, peritoneal effusion, pelvic effusion and pericardial effusion. 10 While several studies have reported a heightened prevalence of pleural effusion in sHLH patients, there is relatively little literature regarding peritoneal effusion.6,8,11–13 To the best of our knowledge, this is the first study to comprehensively evaluate the clinical significance and prognostic value of peritoneal effusion in sHLH patients. In this study, we observed the distribution of peritoneal effusion in sHLH patients and their correlation with laboratory characteristics and immune status, suggesting that abdominal effusion is associated with a poorer survival outcome in sHLH. Peritoneal effusion is a common complication in critically ill patients and those with malignant tumors. Numerous studies have demonstrated that peritoneal effusion is an indicator of advanced disease and poor prognosis in clinical settings such as cirrhosis, liver cancer, ovarian cancer, and gastrointestinal cancers.14–17 However, reports of peritoneal effusion in sHLH are rare and are mainly found in case reports.18–20 In this study, we discovered for the first time that the incidence of peritoneal effusion in sHLH patients was 22.4%, which was higher than that in previous studies. 8 However, this proportion is higher than that in critically ill patients previously reported. 21 Our research results suggest that patients with lower white blood cell counts, lymphocyte counts, monocyte counts, and platelet counts have a higher likelihood of having peritoneal effusion. Furthermore, through multivariate analysis, we identified that low counts of monocyte and high levels of DBil are associated with an increased risk of peritoneal effusion. Additionally, our study found that the disruption of lymphocyte subset balance in the early stages of sHLH may be linked to the presence of peritoneal effusion. Excessive activation of the immune system and severe systemic inflammation are well known as key characteristics of HLH. Cytopenia is primarily associated with the excessive inflammatory response. Previous studies have demonstrated that activated macrophages secrete amounts of cytokines, with IFN-γ being a key driver of the rapid onset of cytopenia. Activated macrophages engulf or consume host red blood cells, white blood cells, and platelets.22–24 The presence of elevated DBil may be related to the infiltration of hemophagocytic histiocytes and the overproduction of cytokines. Infiltration of macrophages and lymphocytes in the bile ducts can lead to hyperbilirubinemia and jaundice.25,26 Currently, there are no studies on the correlation between peritoneal effusion and immune status in HLH. Previous studies have shown that heightened apoptosis and exhaustion of T cells, along with elevated NK cells, are associated with the exacerbation of HLH.27–29 Our study found that decreased count and percentage of CD4+ T cells and increased percentage of NK cells associated with the generation of peritoneal effusion, which were consistent with previously identified immune risk factors contributing to the worsening of HLH. Therefore, it is reasonable to hypothesize that peritoneal effusion and the exacerbation of inflammatory states may represent concurrent manifestations of severe, uncontrolled immune dysregulation in sHLH. While immune dysregulation is a hallmark of HLH itself, the more pronounced alterations observed in PE+ patients - as reflected by lower CD4+ T cell counts and higher NK cell percentages - suggest a more severe degree of immune derangement. Importantly, this stratification was strongly associated with survival outcomes (median OS, 19 vs. 103 days, p = 0.002), indicating that PE identifies a subgroup of patients with more severe immune dysregulation and worse prognosis. Consequently, it is important to maintain awareness of the potential for peritoneal effusion in sHLH patients accompanied by significantly reduced monocyte levels, elevated DBil, low CD4+ T cells and markedly increased proportions of NK cells. The pathophysiological mechanisms of serous cavity effusion in critical illness and haematological malignancies remain inadequately understood. In critically ill patients, the mechanisms are multifactorial: mechanical ventilation impairs pleural fluid turnover and alters intrathoracic pressure, systemic inflammation increases capillary permeability, and hemodynamic changes collectively contribute to fluid accumulation. 30 For hematological malignancies, besides direct tumor invasion or compression (e.g., superior vena cava syndrome causing pleural effusion), 31 evidence from sepsis studies suggests that endothelial dysfunction and increased vascular permeability mediated by Angiopoietin-2 and soluble VCAM-1 may also be involved. 32 At present, only one retrospective study has reported the incidence of pleural effusion was 58.6%, with lg(sCD25) and platelet ≤ 65 × 109/L identified as significant risk factors for pleural effusion in sHLH patients. Moreover, pleural effusion is associated with increased inflammation and poorer outcomes. 8 There is a possibility that inflammatory status could be implicated in the pathogenesis of peritoneal effusion in HLH patients. The systemic inflammatory response plays a crucial role in the pathophysiological processes of HLH, with excessive activation potentially leading to tissue damage and fluid accumulation. Cytokines and other inflammatory mediators stimulate and activate vascular endothelial cells, promoting the loss of endothelial barrier function and an increase in vascular permeability, leading to the formation of tissue edema. Therefore, the exacerbation of the inflammatory state is likely a key shared mechanism in the pathophysiology of both severe sHLH and the occurrence of peritoneal effusion.33–35 CD4 T cells play a crucial role in immune responses. Vigorous inflammatory reactions promote T cell apoptosis, leading to an uncontrolled inflammatory response due to a reduced T cell reaction, which results in a loss of control over damaging immune responses and promotes tissue damage.36,37 NK cells are known to regulate inflammatory responses. 38 NK cells are characterized primarily by a reduction in quantity and impaired function. 39 However, several studies have found that an increased percentage of NK cells is associated with a poorer prognosis.28,40 The evidence we obtained suggests that low levels of CD4+ T cells and high percentages of NK cells were significantly associated with poor OS. As indicated by our multivariate analysis, immune-mediated mechanisms are strongly implicated in the pathophysiology of sHLH and are associated with the presence of peritoneal effusion. Our study has revealed a striking disparity in the cumulative mortality rates between sHLH patients with and without peritoneal effusion (p = 0.002), and the presence of peritoneal effusion was found to be independently associated with poorer survival. The underlying mechanism underlying the association between peritoneal effusion and higher mortality in sHLH remains unclear. One plausible hypothesis is that peritoneal effusion serves as a clinical indicator of a more severe degree of immune dysregulation, which is the primary driver of uncontrolled inflammation and subsequent tissue damage. Several interconnected mechanisms may explain this association. First, the systemic cytokine storm (e.g., IFN-γ, TNF-α) directly damages vascular endothelium, leading to capillary leakage and promoting peritoneal effusion formation.41–43 Second, hepatic sinusoidal infiltration by activated histiocytes induces portal hypertension and liver dysfunction, which both drive ascites and impair detoxification, perpetuating a vicious cycle of inflammation and coagulopathy.41–43 Third, protein-rich ascitic fluid serves as a bacterial culture medium, and with concurrent neutropenia, markedly increases the risk of spontaneous bacterial peritonitis and sepsis, which further amplifies the hyperinflammatory cascade.44–46 Collectively, these mechanisms position peritoneal effusion not merely as a marker of severe inflammation but as an active contributor to multiorgan dysfunction and mortality. Consequently, the identification of peritoneal effusion in sHLH patients should raise alarm bells and prompt specific clinical actions: intensifying monitoring for organ dysfunction, ensuring the timely initiation of HLH-directed therapy, and aggressively treating the underlying trigger. This finding should be recognized as a marker of a potentially more severe disease course and a higher risk of mortality, enabling more proactive and tailored patient management. Importantly, we further explored whether peritoneal effusion itself confers independent prognostic information beyond reflecting overall disease severity. In our multivariate Cox regression model adjusted for key confounders - including age, etiology and laboratory parameters - PE remained an independent predictor of mortality (HR=1.88; 95% CI 1.09 - 3.24, p=0.023). To further address the concern that the observed survival disadvantage might simply be driven by a more aggressive underlying etiology, we performed stratified analyses by HLH trigger. Notably, the prognostic impact of PE persisted in both lymphoma-associated HLH (median OS, 19 vs. 65 days, p=0.020) and non-lymphoma HLH (median OS, 17 vs. 187 days, p=0.035), effectively ruling out etiology-driven confounding. These findings collectively indicate that while PE likely reflects, at least in part, the severity of underlying disease, it also provides independent prognostic information beyond traditional risk factors.Therefore, the presence of PE should be interpreted as a robust, independent indicator of high mortality risk, irrespective of the underlying trigger, and should prompt early and aggressive therapeutic intervention. Our findings are also informative when compared with other sHLH cohorts. Dziedzic et al. recently reported a series of 21 sHLH patients and identified lower white blood cell, neutrophil, and lymphocyte counts at diagnosis as poor prognostic factors. 47 Consistent with their observations, our study confirmed the prognostic significance of cytopenias, while extending the evidence by identifying peritoneal effusion as a novel independent mortality predictor in a substantially larger cohort. We acknowledge that peritoneal effusion frequently coexists with pleural, pericardial, and pelvic effusions, as shown in our Table S1 (95.8% of PE+ patients had concurrent pleural effusion). Indeed, in the setting of severe systemic inflammation, peritoneal effusion is often part of a polyserositis phenomenon. However, we contend that this does not diminish its clinical utility. First, unlike pleural effusion - which can be influenced by various factors such as mechanical ventilation, volume status, or cardiac function in critically ill patients - peritoneal effusion, when it occurs, is less confounded by these non-specific factors and may serve as a more reliable indicator of disease severity in sHLH. Second, our multivariable analysis demonstrated that peritoneal effusion provides independent prognostic information beyond the presence of other effusions. Therefore, we advocate that clinicians should specifically attend to peritoneal effusion as a readily identifiable risk marker rather than dismissing it as a nonspecific finding. There are several limitations concerning the impact of peritoneal effusion in sHLH patients. This study is a single-center and retrospective study, which may introduce selection bias and confounding factors. Given the substantial biological and clinical heterogeneity of sHLH - encompassing diverse triggers such as lymphoma, infections, and autoimmune disease - complete elimination of confounding remains inherently unattainable even with rigorous methodological adjustments. The relatively small sample size precludes more detailed subgroup analyses, such as investigating the incidence of peritoneal effusion across various etiologies. Additionally, the absence of lymphocyte subset data for a portion of the patients restricts a comprehensive exploration of the immune-mediated mechanisms. Given the retrospective nature and limited sample size, the results obtained need to be further validated in multicenter studies. In addition, the mechanistic interpretations proposed in the Discussion remain speculative and require validation in future mechanistic studies. Despite these limitations, the findings presented here are significant, as this is the first study specifically focusing on peritoneal effusion at the time of sHLH diagnosis, indicating that peritoneal effusion is a common and independent clinical marker associated with high mortality. Consequently, clinicians should be aware of peritoneal effusion in sHLH patients. Conclusions This study demonstrates that peritoneal effusion is present in approximately one-fifth of sHLH patients and is associated with lower blood cell counts, elevated direct bilirubin, and distinct immune dysregulation characterized by reduced CD4+ T cells and increased NK cell percentages. More importantly, peritoneal effusion was identified as an independent predictor of mortality, even after adjusting for potential confounders and stratifying by etiology. These findings underscore the prognostic significance of peritoneal effusion in sHLH and support its utility as a readily accessible clinical marker for risk stratification. Further prospective studies with larger cohorts are warranted to validate these findings and to elucidate the underlying mechanisms. Supplemental material Supplemental material - Peritoneal effusion as a marker of disease severity in secondary hemophagocytic lymphohistiocytosis: Clinical and immunological insights Supplemental material for Peritoneal effusion as a marker of disease severity in secondary hemophagocytic lymphohistiocytosis: Clinical and immunological insights by Yansheng Wang, Lulu Zhang, Huiping Wang, Qian Li, Furun An, Beibei Xie, Shuangwei Ying, Jinxiao Hou, Hui Qin, Zhimin Zhai, Fan Wu in Sage Open Medicine.