Histiocytic necrotizing lymphadenitis (HNL), also known as Kikuchi-Fujimoto disease (KFD) or subacute necrotizing lymphadenitis (SNL), is a benign, self-limited disorder marked by fever, superficial lymphadenopathy, and normal or reduced leukocyte counts (1–3). First reported by Kikuchi and Fujimoto in 1972, HNL predominantly affects adults. While HNL has been widely reported in adults, studies focusing on its pediatric presentation remain limited, particularly in the Chinese population. The nonspecific nature of fever and superficial lymphadenopathy might lead to misdiagnosis or delays in diagnosis and treatment. The diagnosis, treatment, and prognosis of HNL markedly differ from those of lymphoma and systemic lupus erythematosus. This article retrospectively examines the clinical data, laboratory results, and prognosis of 44 pediatric cases of HNL identified and treated at our hospital to enhance understanding of this disease.

Histiocytic necrotizing lymphadenitis (HNL) can occur in individuals of any age, ethnicity, or gender; however, it is more prevalent in Asian populations, with an indeterminate incidence rate (4). Previous studies indicated a a greater prevalence among young females, while recent research reveals a broader age spectrum (19 months to 75 years) and a comparable gender distribution. To date, the etiology of HNL remains undetermined but may be associated with infectious factors, autoimmune factors, and genetic predisposition. Several studies have shown that multiple infectious agents, including EBV, CMV, P19, HIV, and types 6 and 8 of human herpesvirus, may trigger HNL (5–8). Nevertheless, no particular pathogen was associated with HNL in a Chinese investigation of 153 children (9). Even though every child in our cohort had extensive pathogen testing, we were unable to find evidence of a causal relationship.

The immunological etiology of HNL involves Fas-FasL-mediated apoptosis driven by cytotoxic CD8⁺ T lymphocytes, supported by histiocytic phagocytosis (10). Clinical manifestations primarily consist of nonspecific fever and lymphadenopathy. Studies indicates (11) that some patients initially present with unexplained fever, defined by low-grade or recurrent high-grade fever, an irregular fever pattern, with low-grade fever being the most prevalent, and a variable duration spanning from one week to several months. In 60%–90% of HNL patients, localized superficial lymphadenopathy is the primary symptom, with cervical lymphadenopathy being the predominant manifestation. This is often associated with the enlargement of other superficial lymph nodes such as axillary, and/or supraclavicular, inguinal regions, whereas generalized lymphadenopathy is rarely observed. The affected lymph nodes typically measure 5–40 mm in diameter (>60 mm is uncommon), with a soft consistency, well-defined borders, and good mobility, and may exhibit tenderness (12). No statistically significant difference in lymph node enlargement was seen between the pediatric and adult cohorts. All children in our cohort exhibited fever and superficial lymphadenopathy at onset, aligning with the clinical manifestations of HNL. Some patients may develop nonspecific rashes of varying morphologies, such as papules, erythema multiforme, urticaria, or drug eruptions. Among our cases, 10 patients (22.7%) exhibited rashes, predominantly manifesting as urticaria, papules, or erythema multiforme. Additionally, HNL patients may experience systemic atypical symptoms, including nausea, vomiting, diarrhea, weight loss, night sweats, anorexia, arthralgia, and myalgia (13). A minority of cases may progress to recurrent multi-organ involvement or evolve into other rheumatic immune diseases (14). Reports suggest a complex relationship between HNL and systemic lupus erythematosus (SLE), where SLE may precede, coincide with, or follow the onset of HNL (15). Therefore, children diagnosed with HNL necessitate long-term monitoring of autoantibody profiles and careful observation for multisystem involvement or advancement to other rheumatic immunological diseases. In our cohort, no cases progressed to other rheumatic diseases, potentially due to the limited sample size and relatively short follow-up duration for certain patients.secondary HLH, a severe complication of HNL, frequently occurs as a consequence of infections, rheumatic immune disorders, or malignancies. The pathophysiology is associated with cytokine storms driven by interleukin-6 (IL-6), interferon-γ (IFN-γ), and mutations in the perforin gene. Studies suggest that significantly elevated serum LDH levels may serve as a biomarker for HNL complicated by secondary HLH (16). Utilizing the 2016 EULAR/ACR/PRINTO classification criteria for macrophage activation syndrome (MAS) (17), both secondary HLH cases in our cohort exhibited remittent fever lasting over two weeks, alongside elevated serum ferritin (SF), lactate dehydrogenase (LDH), aspartate aminotransferase (AST), hypofibrinogenemia, thrombocytopenia, and hemophagocytosis observed in bone marrow analysis, thereby substantiating the diagnosis of HNL complicated by secondary HLH. A minimal percentage of HNL patients may develop complications such as aseptic meningitis, peripheral neuropathy, cerebellar ataxia, or autoimmune hepatitis (4, 18).

Studies indicate that HNL lacks disease-specific laboratory indicators. Hematological findings may include leukopenia, normal leukocyte counts, or leukocytosis, along with reduced hemoglobin, thrombocytopenia, increased lymphocyte proportion, and raised erythrocyte sedimentation rate (ESR). Non-specific abnormalities such as elevated CRP, SF, LDH, and transaminases are also observed (19). In our cohort, elevated ESR, SF, and LDH, as well as leukopenia (reduced white blood cell count) and anemia (low hemoglobin), were frequently identified, aligning with the existing literature.

Ultrasonography is a non-invasive, radiation-free procedure that verifies the dimensions of swollen lymph nodes and offers a differential diagnosis based on imaging characteristics (20). Som classification is an imaging-based technique for the categorization of cervical lymph nodes, offering accuracy and reproducibility in lymph node location. Regrettably, the anatomical positioning of the affected enlarged cervical lymph nodes was not meticulously delineated in this investigation in accordance with Som's classification (21). In this investigation, twenty-two cases (50.0%) had bilateral cervical lymph node expansion, in contrast to the more prevalent unilateral cervical lymph node enlargement often observed in the common progression of the disease (22–26). The subsequent indicators are regarded as characteristic ultrasonography manifestations of HNL (27): (1) Lesion location: mostly affects the posterior cervical area. (2) Gray-scale ultrasonography reveals hyperechoic lymph node gates, typically devoid of interior calcification, with necrotic foci being infrequent. (3) Doppler ultrasound: it reveals a normal blood flow pattern, centrally located vascular structures in the hilar region, and radial branching of arteries from the hilar region in both longitudinal and transverse perspectives. Lymphoma and reactive hyperplasia of lymph nodes may also exhibit centralized portal vascular structures and normal blood flow patterns (28). Nonetheless, both diseases typically do not exhibit evidence of perinodal echogenic amplification. Lymphomatous lymph nodes typically exhibit increased size, rounded morphology, heightened hyperechogenicity on grayscale imaging, and lack hilar echogenicity in contrast to hyperplastic lymph nodes (28). Moreover, lymphomatous lymph nodes typically exhibit a heterogeneous flow pattern (28, 29), whereas the majority of HNL lymph nodes demonstrate a normal hilar flow signal on energy Doppler. Significant necrosis in HNL lymph nodes necessitates differentiation from metastatic lymphadenitis and tuberculous lymphadenitis (30). Metastatic lymph nodes typically lack hilar echoes and are bigger and more spherical compared to those in HNL illness (30, 31). Moreover, metastatic lymph nodes typically exhibit peripheral blood flow with displaced portal arteries (28, 29, 32), while peri-lymph node echo amplification is rather rare. Tuberculous lymphadenitis typically exhibits no hilar echogenicity (28, 31) or perinodal echo amplification; nonetheless, calcifications are frequently observed, and hilar arteries may be misplaced or missing (28, 30). Consequently, perinodal echo enhancement appears to represent a distinctive expression of HNL. In this investigation, 36 cases (81.8%) underwent lymph node ultrasonography, and the results aligned with the characteristic ultrasound indicators of HNL.

Computed tomography (CT) of the neck serves as a significant reference prior to lymph node biopsy, and in the majority of individuals with HNL, the primary manifestations observed on CT are (25): (1) Characteristics of lymph node distribution: unilateral involvement is prominent, primarily occurring in the cervical lymph nodes of zones II-V. (2) Enhancement attributes: uniform augmentation in dimensions and uniform enhancement in enhancement. (3) Peripheral tissue alterations: approximately 81% of patients had evidence of peripheral lymph node involvement. This study involved 7 cases (15.9%) that received CT scan of the neck, revealing nodular soft-tissue density shadows, partial fusion, homogenous enhancement post-contrast, and an absence of necrotic indications. Necrotic symptoms necessitate suspicion of lymph node TB infection. The primary distinctions between HNL and TB infection encompass (33): (1) Lack of necrotic lymphadenopathy. (2) Anomalous cortical attenuation values of lymph nodes. (3) The CT value ratio of lymph nodes to surrounding muscles possesses discriminating significance. Kwon et al. identified CT as the preferred modality for assessing perinodal infiltration and substantial necrosis in lymph nodes in HNL. However, CT entails radiation dangers (34, 35), but ultrasound is noninvasive, safe, and more economical (29). Furthermore, due to its provision of real-time guidance, ultrasound-guided puncture biopsy can be executed when pathological confirmation of lymph node lesions is necessary. Kim et al. (36) report that ultrasound-guided puncture biopsy of cervical lymph nodes has a sensitivity of 97.9% and a specificity of 99.1%. Furthermore, ultrasound is superior in evaluating necrosis and can depict portal blood flow by energy Doppler imaging (28).

The clinical manifestations of HNL are nonspecific, and pathological examination of lymph node biopsies serves as the definitive diagnostic criterion for this disease. Pathological findings primarily consist of abundant apoptotic cells localized in the paracortical regions of the afflicted lymph nodes, surrounded by infiltrates of small lymphocytes, histiocytes, transformed lymphocytes, and plasmacytoid monocytes, with a complete absence of neutrophils. Coagulative necrosis with extensive karyorrhectic debris characterizes HNL, though distinct necrotic foci may be absent in the early stages of the disease. A critical immunohistochemical feature is the presence of clustered or small focal clustered or nodular aggregates of CD68-positive cells, with immunohistochemical staining revealing mixed proliferation of T cells and B cells (37). From a purely histopathological perspective, HNL exhibits three cardinal features that distinguish it from lymphoma: ① Architectural Preservation vs. Destruction:HNL maintains residual nodal architecture with focal paracortical expansion, whereas lymphomas typically efface nodal structures through diffuse proliferation of atypical lymphocytes (38). ② Cytological Composition: HNL shows a triad of: CD68+ histiocytes with crescentic nuclei and karyorrhectic debris, CD123+ plasmacytoid dendritic cells, Predominantly CD8+ T-cell infiltrates (39). In contrast, lymphomas demonstrate monomorphic proliferation of malignant lymphocytes (e.g., CD20+ B-cells in DLBCL) with high Ki-67 index (>90% vs. HNL's < 30%) (40). ③ Microenvironmental Patterns:HNL exhibits geographic necrosis surrounded by foamy histiocytes without angioinvasion. Lymphomas often show:Cohesive tumor sheets with stromal desmoplasia (e.g., Hodgkin lymphoma) and Vascular invasion (41). These features underscore the importance of immunohistochemical profiling (CD68/CD123/CD3/CD20/Ki-67) in equivocal cases. All 44 cases in this cohort underwent lymph node biopsy with pathological examination, confirming the diagnosis of HNL and eliminating malignant disorders such as lymphoma.

HNL must be distinguished from the following conditions: ① Lymph node tuberculosis infection: predominantly chronic in nature, presenting with symptoms indicative of tuberculosis infection, including low-grade fever, lethargy, and malaise; lymph nodes are typically painless in the initial stages, potentially becoming tender upon palpation in later stages; PPD and γ-interferon release tests frequently yield positive results; pathology is marked by caseous necrosis, with possible positive antacid staining. In this investigation, all cases tested negative for PPD and γ-interferon release tests, and the pathological examination did not indicate tuberculosis infection of the lymph nodes. ② Lymphoma: a chronic and progressive condition, frequently associated with low-grade fever, nocturnal diaphoresis, weight reduction, and painless lymphadenopathy, which may occur in any anatomical region. The pathology is marked by the monoclonal proliferation of lymphoid cells, with immunohistochemistry revealing specific lymphocyte markers (CD20, CD3, CD30, etc.). In this cohort of instances, the enlarged lymph nodes exhibited tenderness, which, when considered alongside the pathological evaluation of the lymph nodes, was incongruent. ③ EBV infection: primarily characterized by fever, pharyngitis, lymphadenopathy, hepatosplenomegaly, rash, among other symptoms. All children in this cohort tested negative for EBV-IgM/VCA-IgM, had low levels of heterogeneous lymphocytes, and no virus-infected cells were observed in the pathological examination, indicating incompatibility. Research indicates (11) that the misdiagnosis rate in children surpasses that of adults, partly due to insufficient awareness of the disease. Additionally, parental perceptions of lymph node biopsy as a traumatic procedure may lead to temporary refusals, consequently delaying diagnosis.

Currently, there are no standardized treatment guidelines for HNL. Conventional antibiotic therapy frequently proves futile, as the disease exhibits a self-limiting nature, with most symptoms and signs resolving spontaneously within 4 months. In this study, antibiotic therapy was ineffective in 9 patients (20.5%), while fever and lymphadenopathy healed spontaneously in 8 cases (18.2%), consistent with previous reports (6). Pharmacological intervention is necessary for patients with persistent symptoms, multisystem involvement, or extranodal lesions. Glucocorticoids and hydroxychloroquine have been identified as effective therapies for HNL (42–44), while intravenous immunoglobulin (IVIG) demonstrates efficacy in severe cases. Most patients experience rapid alleviation of symptoms following glucocorticoid therapy. In our cohort, 27 patients (61.4%) received glucocorticoid monotherapy, achieving normal body temperature within 24 h of treatment initiation and variable improvement in lymph node swelling and tenderness, further validating the therapeutic efficacy of glucocorticoids. Both secondary HLH cases had substantial clinical improvement with glucocorticoids or intravenous immunoglobulin (IVIG), which aligns with previous studies. Additionally, studies report that hydroxychloroquine monotherapy or its combination with glucocorticoids may be effective for managing recurrent or glucocorticoid-resistant HNL (45, 46). Nonsteroidal anti-inflammatory drugs (NSAIDs) can be used to alleviate fever and lymph node tenderness. However, the efficacy of ciprofloxacin, ofloxacin, and minocycline in in the treatment of HNL is not well-supported by robust evidence and requires further investigation (47–49).

The reported recurrence rate of HNL is approximately 3%–4%, with a mortality rate of 2.1% (4). In our cohort of 44 cases, follow-up durations ranged from 2 months to 6 years, during which 5 patients (11.4%) experienced recurrence—1 case (2.3%) relapsed twice, and 4 cases (9.1%) relapsed once—with no fatalities reported. Evidence suggests that a positive antinuclear antibody (ANA) is a high-risk factor for disease recurrence (50). Furthermore, leukopenia has been reported to be significantly associated with HNL recurrence (51). All 5 recurrent cases (11.4%) in our study exhibited leukopenia, consistent with previous findings. Research conducted by Stéphan Atarashi et al. (52) indicates a significant association between viral infections and the recurrence of HNL. Additionally, pancytopenia (deficiency of all three blood cell lineages) at disease onset has been identified by some scholars as a potential high-risk factor for unfavorable prognosis (53).

This study aimed to objectively examine histiocytic necrotizing lymphadenitis in children; nevertheless, the following limitations must be recognized. The study was a retrospective analysis conducted at a single center, with data obtained through the examination of hospital records. Moreover, the sample size of this investigation was comparatively limited, thereby diminishing the statistical power and impacting the reliability of the results. Despite our efforts to enhance the sample size, the final cohort remained comparatively small due to constraints in research funding and restricted data availability. Future extensive, multicenter collaborative prospective research will address this resource issue and yield more conclusive results.

In summary, clinicians should contemplate the potential of HNL in cases of unexplained fever with superficial lymphadenopathy and leukopenia that do not respond to standard antibiotic therapy. Early lymph node biopsy is essential for definitive diagnosis, thereby preventing superfluous diagnostic interventions and the misuse of medications. Disease management emphasizes symptomatic treatment, while glucocorticoids, hydroxychloroquine, or intravenous immunoglobulin are advised for individuals with persistent symptoms, multisystem involvement, or extranodal lesions. Given the risks of recurrence and progression to rheumatic immune diseases, long-term follow-up is essential.