Immune Checkpoint Inhibitor-Associated Hemophagocytic Lymphohistiocytosis in Solid Tumors: A Systematic Review

1. Introduction

The advent of immune checkpoint inhibitors (ICIs) has fundamentally changed how we treat solid tumors. Agents targeting PD-1, PD-L1, and CTLA-4 have produced responses, and in some cases durable remissions beyond those historically achieved with chemotherapy alone [1]. Their use has expanded rapidly across melanoma, non-small cell lung cancer, renal cell carcinoma, and beyond, and they are now a cornerstone of first-line treatment for many advanced malignancies [1].

This enhanced immune activity, however, carries significant toxicity. By releasing the brakes on the immune system, ICIs can trigger a wide spectrum of inflammatory toxicities. Skin rashes, colitis, thyroiditis, and pneumonitis are now familiar territory for oncologists, with immune related adverse events (irAEs) reported in up to 90% of patients receiving anti-CTLA-4 and up to 70% of those on anti-PD-1/PD-L1 agents [2]. Less common but potentially life-threatening are hyperinflammatory toxicities such as hemophagocytic lymphohistiocytosis (HLH), characterized by dysregulated systemic immune activation.

HLH is not a new disease. Hematologists have long recognized it in the context of infections, autoimmune conditions, and hematologic malignancies, where it carries a mortality rate approaching 40% in adults [3]. It is classified as primary, typically driven by genetic defects in cytotoxic lymphocyte function, or secondary, arising in response to an external trigger such as infection, rheumatologic disease, or malignancy [4].

More recently, HLH has emerged as a rare complication of ICI therapy in patients with solid tumors, a population in whom the syndrome has been underrecognized and poorly studied.

Pharmacovigilance data suggest HLH is reported seven times more frequently with ICI therapy than with other drug classes, pointing to a real biological signal rather than incidental association [5].

Recognizing ICI-associated HLH in patients with solid tumors is particularly challenging, as its clinical features, including fever, cytopenias, hyperfferitinemia, and hepatitis, overlap substantially with sepsis, disease progression, and other irAEs such as immune-mediated hepatitis There is no single test that makes the diagnosis, and delayed recognition may allow progression to critical illness. In patients receiving combination immunotherapy or concurrent chemotherapy, the picture is even murkier, as treatment-related cytopenias and malignancy-associated inflammation can each independently mimic or obscure the syndrome [6].

Despite increasing recognition, the published literature on ICI-associated HLH in solid tumors remains almost entirely case reports and small series, with treatment strategies variable and no clear consensus on management [7]. There is no consolidated analysis focused specifically on the solid tumor population characterizing clinical presentation, onset timing, treatment patterns, and outcomes. Clinicians managing these cases are working without evidence-based guidance.

To address this gap, we conducted a systematic review of published cases of ICI-associated HLH in adult patients with solid malignancies. Our objective was to summarize published case-level data to better characterize the clinical features, timing of onset, diagnostic findings, therapeutic strategies, and outcomes of this rare but serious complication, with the aim of earlier recognition and more standardized management.

2. Methods

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

2.1. Eligibility Criteria

Studies were eligible for inclusion if they reported cases of HLH associated with ICI therapy in adult patients (≥18 years) with solid malignancies. Eligible study designs included case reports and case series with extractable patient-level data. Studies were required to report clinical presentation, treatment, or outcomes related to HLH.Studies were excluded if they involved pediatric populations, HLH associated with hematologic malignancies, or non-English publications without sufficient extractable data.

2.2. Search Strategy

A systematic literature search was conducted in PubMed, Cochrane Library, and Google Scholar from database inception through March 3, 2026. The study selection process is summarized in Figure 1. The PubMed search strategy combined terms related to hemophagocytic lymphohistiocytosis (HLH) and immune checkpoint inhibitors, including specific agent names. Searches were limited to English-language studies involving adult populations. No publication year restrictions were applied.

The PubMed search strategy was as follows, yielding 140 records, which were added in to the initial manuscript pool:

(

hemophagocytic lymphohistiocytosis[tiab]

OR HLH[tiab]

OR hemophagocytic syndrome[tiab]

)

AND

(

immune checkpoint inhibitor*[tiab]

OR checkpoint inhibitor*[tiab]

OR immunotherapy[tiab]

OR pembrolizumab[tiab]

OR nivolumab[tiab]

OR cemiplimab[tiab]

OR dostarlimab[tiab]

OR atezolizumab[tiab]

OR durvalumab[tiab]

OR avelumab[tiab]

OR ipilimumab[tiab]

OR tremelimumab[tiab]

)

NOT

(

leukemia[tiab]

OR lymphoma[tiab]

OR myeloma[tiab]

OR “hematologic malignan*”[tiab]

OR “haematologic malignan*”[tiab]

)

AND english[lang]

NOT (

infant[mesh]

OR child[mesh]

OR adolescent[mesh]

)

NOT (

leukemia[mesh]

OR lymphoma[mesh]

)

The Cochrane Library was additionally searched using the following exact terms:

hemophagocytic lymphohistiocytosis, and also, HLH, yielding 74 records which were added to the initial manuscript pool.

Google Scholar was searched using the following this exact search term:

“hemophagocytic lymphohistiocytosis”, HLH, immunotherapy, ici, solid tumors

This yielded 47 records after manual screening by title, which were included in the initial manuscript pool.

2.3. Study Selection and Data Extraction

Two reviewers independently screened titles and abstracts for eligibility, followed by a full-text review of potentially relevant articles. Discrepancies were resolved through discussion and consensus.

Data were extracted at the individual case level. Extracted variables included patient demographics (age, sex), malignancy type, type of ICI used, number of ICI cycles, time from ICI exposure to HLH onset, programmed death (PD-1/PD-L1) or cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) expression (when available), HLH diagnostic criteria, intensive care unit (ICU) admission, treatment modalities for HLH, and clinical outcomes, including mortality.

2.4. Study Selection Results

A total of 261 articles were identified through database searching. An additional 9 articles were identified through manual review of the reference lists. 144 studies remained after duplicate removal. All 144 studies underwent title and abstract screening, leaving 72 articles. After screening and application of eligibility criteria, 70 articles were included in the final analysis, leading to a total of 80 cases (Figure 1).

2.5. Data Synthesis

Given the rarity of ICI-associated HLH and the heterogeneity of available data, a quantitative meta-analysis was not feasible. Therefore, a descriptive synthesis was performed. Case-level data were tabulated to identify patterns in patient characteristics, clinical presentation, treatment strategies, and outcomes.

2.6. Risk of Bias and Certainty of Evidence

Given that the available literature on ICI-associated HLH consists entirely of case reports and small case series, formal risk of bias assessment using standardized tools such as the Cochrane Risk of Bias tool or GRADE framework was not applicable to this dataset. Individual case reports are inherently subject to selection bias, reporting bias, and publication bias favoring positive or survival outcomes. These limitations are acknowledged throughout and are reflected in our descriptive approach to data synthesis. No formal certainty of evidence rating was assigned, as the heterogeneity of reporting and absence of comparative data preclude such assessment. This review was not prospectively registered given its descriptive, hypothesis-generating nature and the absence of a pre-specified quantitative analysis plan.

3. Results

3.1. Selection of Studies and Data Analysis

The present case-level dataset consisted of 80 adult patients with solid tumors who developed HLH after ICI therapy [8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75]. (Table 1) Most included reports were case reports or small case series; therefore, results were summarized descriptively. Not all reports fully reported timing, diagnosis, or outcome, therefore the denominators differed by variables.

3.2. Patient and Tumor Characteristics

Median age was 63 years (IQR, 51.5–73.3; range, 32–80) and sex distribution was balanced with 41 male patients (51.3%) and 39 female patients (48.8%). The most common tumor type was lung cancer (22/80, 27.5%), followed by melanoma (15/80, 18.8%), breast cancer (9/80, 11.3%), kidney/renal tumors (8/80, 10.0%), and gynecologic malignancies (6/80, 7.5%). The most frequent individual tumor types were melanoma (15/80, 18.8%), lung adenocarcinoma (14/80, 17.5%), triple-negative breast cancer (7/80, 8.8%), clear cell renal cell carcinoma (6/80, 7.5%), and lung squamous cell carcinoma (5/80, 6.3%). These findings are summarized in Table 2.

3.3. ICI Exposure and Timing of HLH Presentation

Most of the exposure was to PD-1 inhibitors and was noted in 69 of 80 cases (86.3%). Exposure to PD-L1 inhibitor and CTLA-4 inhibitor was reported in 10 (12.5%) and 16 (20.0%) cases, respectively. Due to the combination checkpoint blockade, the categories were not mutually exclusive. Sixty-five patients (81.3%) received a single checkpoint inhibitor, and 15 (18.8%) received combination checkpoint blockade. Pembrolizumab was the most common agent (38/80, 47.5%), followed by nivolumab (25/80, 31.3%) and ipilimumab (14/80, 17.5%). Sixty-five cases reported the time from last exposure to ICI to HLH onset. Median time to onset was 17 days (interquartile range 10 to 30; range 0 to 360). HLH occurred within 30 days of the last ICI dose in 49 of 65 cases (75.4%) and within 60 days in 57 of 65 cases (87.7%). Seventy cases reported the number of ICI-containing cycles before HLH onset, with a median of 2 cycles (IQR, 1–4; range, 1–17). HLH occurred after the first ICI-containing cycle in 27/70 cases (38.6%) and within the first four ICI-containing cycles in 53/70 cases (75.7%). These findings are summarized in Table 3.

3.4. Diagnosis - Characteristics and Severity

HScore reporting was incomplete. 41 of 80 cases (51.3%) had an HScore or qualitative HScore estimate; 40 cases had extractable numeric scores. For numeric reports, the median HScore was 237 (interquartile range, 207-267; range, 105-319). HLH-2004 diagnostic criteria, defined as meeting at least 5 of 8 criteria when explicitly reported, were documented as fulfilled in 74 of 80 cases. Among 79 cases with reported bone marrow findings, hemophagocytosis was identified in 53 (67.1%). In 6 of 79 (7.6%) reported cases, a concomitant infection trigger was identified. ICU admission or ICU-level care was reported in 15/80 patients (18.8%). These findings are summarized in Table 4.

3.5. HLH-Directed Therapy Treatment and Outcomes

Treatment was heterogeneous, but corticosteroids were used in 78 of 80 cases (97.5%). Other immunomodulatory or HLH-directed therapies included tocilizumab (19/80 [23.8%]), etoposide (12/80 [15.0%]), anakinra (10/80 [12.5%]), mycophenolate mofetil (8/80 [10.0%]), IVIG (6/80 [7.5%]), ruxolitinib (3/80 [3.8%]), cyclosporine or tacrolimus (3/80 [3.8%]), and plasma exchange (3/80 [3.8%]). Outcomes were reported in 78 cases. Clinical recovery or improvement was observed in 68/78 cases (87.2%) and 10 patients died (12.8%). No formal testing was performed because of the descriptive design, small sample size, and heterogeneous case reporting. Mortality was higher in patients requiring ICU admission or ICU-level care (4/15, 26.7%) vs. non-ICU cases with known outcomes (6/63, 9.5%). These findings are summarized in Table 5.

4. Discussion

ICI-associated HLH is increasingly recognized as a rare but potentially life-threatening irAE. The available literature remains dominated by case reports and small case series, limiting estimates of incidence, prognosis and optimal treatment. To our knowledge, our study represents one of the largest consolidated analyses of ICI-associated HLH in adult patients with solid malignancies. In this systemic review of published adult solid tumor cases, we identified several clinically relevant patterns: PD-1/PD-L1 inhibitor exposure was most frequent, onset was often early after ICI exposure, corticosteroids were used in nearly all reported cases, and reported clinical improvement was common.

In our cohort of 80 patients, ICI-associated HLH most commonly occurred in patients with lung cancer and was predominantly associated with PD-1/PD-L1 inhibitor exposure, particularly pembrolizumab. Importantly, most cases developed early during treatment, with over three-quarters of reported cases occurring within the first four cycles of therapy. This finding suggests that clinicians should maintain a high index of suspicion for HLH early after ICI initiation. Treatment approaches varied, but corticosteroids were used in nearly all reported cases, with a majority of patients demonstrating clinical improvement or recovery following immunosuppressive therapy.

Among cases with reported timing, most occurred within 30 days of the last ICI dose and within the first four ICI-containing cycles. This pattern suggests that ICI-associated HLH may represent an early hyperinflammatory toxicity phenotype after checkpoint inhibition and supports heightened clinical vigilance when compatible inflammatory features develop.

The recognition of HLH in patients receiving ICIs is challenging due to significant overlap with other oncologic and treatment-related conditions. ICI-associated HLH may present as unexplained fever, cytopenias, and elevated transaminases [7]. Additional findings that may support the diagnosis include hyperferritinemia, hypertriglyceridemia, and organomegaly.

These clinical and laboratory abnormalities may closely resemble other serious conditions encountered in this patient population, including sepsis, disseminated intravascular coagulation, acute liver failure, and progressive malignancy. In addition, other irAEs, such as ICI-associated autoimmune hepatitis, may present with similar inflammatory and hepatic manifestations, further complicating timely diagnosis [76].

Given the potentially rapid progression and high morbidity associated with HLH, early recognition is critical, particularly as historical adult mortality is substantial [77]. Suspicion should be maintained in patients presenting with unexplained systemic inflammatory symptoms shortly after initiation of immunotherapy, particularly when accompanied by cytopenias, hyperferritinemia, or evidence of organ dysfunction. Early diagnostic evaluation and prompt initiation of immunosuppressive therapy may improve clinical outcomes and reduce progression to severe or ICU-level disease.

While the precise mechanisms leading to HLH remain incompletely understood, the condition is characterized by severe immune activation and dysregulation. A hallmark feature is the uncontrolled release of proinflammatory cytokines, including IL-1, IL-6, TNF-a, and IFN-y, alongside aberrant activation of cytotoxic cells [78,79,80]. The pathogenesis of HLH is thought to involve excessive activation and impaired regulation of cytotoxic T lymphocytes (CTLs) and natural killer (NK) cells, resulting in persistent immune stimulation and cytokine amplification. Dysregulation of regulatory T-cell function may further contribute to the uncontrolled inflammatory response [81].

ICIs enhance antitumor immunity by promoting T-cell activation and proliferation through blockade of inhibitory immune checkpoints such as PD-1/PD-L1 and CTLA-4 pathways. Although this mechanism improves immune-mediated tumor recognition, it may also disrupt peripheral immune tolerance and permit excessive activation of autoreactive or hyperinflammatory immune responses. [82]. The resulting cytokine-driven inflammatory cascade may contribute to macrophage activation and hemophagocytosis, the pathologic hallmark of HLH [83]. In severe cases, this hyperinflammatory state can ultimately lead to rapid tissue injury, multiorgan dysfunction, and death.

The HLH-2004 criteria provide a commonly used diagnostic framework that incorporates clinical, laboratory, and pathologic findings, including fever, splenomegaly, cytopenias, hypertriglyceridemia and/or hypofibrinogenemia, hemophagocytosis, reduced or absent natural killer (NK) cell activity, hyperferritinemia, and elevated soluble CD25 levels [84]. Current recommendations suggest that fulfillment of at least five of eight HLH-2004 criteria supports the diagnosis of HLH. However, given the potentially rapid progression and high mortality associated with untreated disease, initiation of therapy may be warranted in patients with high clinical suspicion even before complete diagnostic criteria are fulfilled [85].

Another commonly utilized diagnostic tool is the H-Score, which incorporates additional clinical and laboratory variables such as underlying immunosuppression, hepatomegaly, and elevated aspartate aminotransferase (AST) levels. In the original validation study, patients diagnosed with secondary hemophagocytic syndrome demonstrated substantially higher H-Score values compared with patients without HLH, supporting its utility as an adjunctive diagnostic tool in adult populations.

The HLH-2004 treatment protocol was originally developed for pediatric HLH and primarily incorporate corticosteroids, cyclosporine A, intrathecal therapy, and etoposide, followed by hematopoietic stem cell transplantation in selected cases [86]. However, the management of adult ICI-associated HLH remains less clearly defined. The HLH group published expert opinions on the management of adult ICI-associated HLH, with recommendations for treatment with corticosteroids and anti-IL-6R therapy (tocilizumab) in patients with suspected HLH, with escalation to etoposide to be considered if insufficient response is seen after 48 hours [86,87].

In our review, corticosteroids were used in nearly all reported cases, supporting their continued role as the backbone of therapy for ICI-associated HLH. Tocilizumab was among the most commonly utilized steroid-sparing or adjunctive agents, reflecting increasing interest in targeted cytokine-directed therapy for hyperinflammatory syndromes associated with immunotherapy. Prior case reports and small case series have reported rapid clinical improvement and normalization of inflammatory markers following tocilizumab administration in selected patients with ICI-associated HLH [88].

Additional therapies reported in refractory cases included intravenous immunoglobulin (IVIG), anakinra, ruxolitinib, and plasmapheresis, although evidence supporting these approaches remains limited to isolated reports and small series. Given the high morbidity and mortality associated with HLH, as well as the expanding use of ICIs, there exists a need for standardized diagnostic and treatment guidelines tailored specifically to adult patients with ICI-associated HLH.

When compared with secondary HLH arising from more traditional etiologies such as infection or malignancy, ICI-associated HLH in our cohort appeared to demonstrate several distinct clinical patterns. Patients in our study had a higher mean age at presentation (63 years) compared with previously reported cohorts of secondary HLH, in which the average age is often closer to 50 years [89]. Similar age distributions have also been reported in other pharmacovigilance and systematic reviews of ICI-associated HLH. The studies by Diaz and Noseda reported incidences at mean ages of 64 and 63 years old, respectively [5,89]. This difference may partially reflect the older baseline age of patients receiving immunotherapy for solid malignancies, particularly lung cancer, which represented the most common underlying malignancy in our cohort.

Our findings also demonstrated similarities compared with classical secondary HLH. Bone marrow hemophagocytosis was identified in 67.1% of cases in our review, which is generally consistent with previously reported rates in adult HLH cohorts [90].

Reported mortality in published ICI-associated HLH cases our cohort was lower than historically reported mortality rates in adult secondary HLH cohorts (12.8% versus 40%) [91]. Similar findings have been observed in other reviews of ICI-associated HLH. Several factors may contribute to this difference, including earlier recognition of immune-related toxicities in patients receiving ICIs, prompt initiation of immunosuppressive therapy, publication bias favoring successful outcomes, and potential biological differences between ICI-associated HLH and classical malignancy-associated HLH.

Our study has several limitations that should be acknowledged. First, the available literature on ICI-associated HLH is largely limited to case reports and small case series, which are subject to publication bias and heterogeneous reporting practices. Consequently, the reported case counts and outcomes of ICI-associated HLH cannot be interpreted as true incidence, population-level mortality or comparative prognosis estimates. The lower mortality observed in our cohort compared with historically reported mortality rates in adult HLH may, in part, reflect underreporting of fatal cases, publication bias toward successful outcomes, or increased recognition and earlier treatment of ICI-associated disease.

Additionally, many reports lacked complete patient-level clinical data, including comorbid conditions, infectious evaluations, detailed diagnostic workup, and standardized reporting of HLH criteria or H-Score components. This limited our ability to fully assess contributing factors, disease severity, and predictors of clinical outcomes. Treatment reporting was also variable, with incomplete information regarding corticosteroid dosing, duration, tapering strategies, and timing of escalation therapies. Furthermore, our review was limited to English-language publications and excluded patients with hematologic malignancies, which may affect the generalizability of our findings. In addition, the search strategy incorporated terms intended to exclude hematologic malignancies. Although this approach was designed to focus the review on solid tumor populations, it may have inadvertently excluded otherwise relevant reports involving solid tumors that also referenced hematologic malignancies in mixed cohorts or within the manuscript text. As a result, definitive conclusions regarding optimal management strategies for ICI-associated HLH cannot be established from the currently available literature.

5. Future Directions

Future studies should prioritize prospective multicenter registries and standardized case reporting for ICI-associated HLH. Minimum reporting elements should include ICI agent and regimen, timing from last dose and cycle number, concurrent infection or other triggers, HLH-2004 criteria, HScore components when available, ferritin/triglycerides/fibrinogen/sCD25, marrow findings, ICU-level care, treatment sequence and dosing and clinical outcome. Such data is needed to define risk factors, refine diagnostic thresholds in oncology patients, and guide treatment escalation.

6. Conclusion

ICI-associated hemophagocytic lymphohistiocytosis (HLH) is a rare but potentially life-threatening immune-related adverse event. In this review of published adult solid tumor cases, ICI-associated HLH was most often reported in patients with lung cancer and most commonly followed PD-1 inhibitor exposure, particularly pembrolizumab. Onset was often early, typically within the first month and within the first four treatment cycles, supporting heightened clinical vigilance early after ICI initiation when compatible inflammatory features develop. Clinical presentation may overlap with sepsis, progressive malignancy, and other immune-related toxicities, making early diagnosis challenging. Persistent fever, cytopenias, hyperferritinemia, transaminitis, or unexplained systemic inflammation after ICI exposure should prompt consideration of HLH and early diagnostic evaluation. Most patients improved with corticosteroids with tocilizumab used as an adjunctive agent in some cases; reported mortality was greater among patients who requiring ICU-level care. Because the available evidence consists largely of case reports and small series with heterogeneous reporting, these findings should be interpreted descriptively rather than as estimates of incidence or prognosis. Current evidence remains insufficient to define optimal treatment algorithms or predictors of outcome. Increasing recognition of this syndrome with expanding ICI use highlights the need for standardized reporting, prospective data collection, and management protocols for adult patients with ICI-associated HLH.