Secondary hemophagocytic lymphohistiocytosis (HLH) is a life-threatening immune disorder in which the body’s own defenses spiral out of control, producing a massive inflammatory response that can damage multiple organs and, without treatment, rapidly lead to death. Unlike the inherited (primary) form seen mostly in infants, secondary HLH is triggered by something external: an infection, a cancer, an autoimmune flare, or even certain medications. Recognizing it quickly is notoriously difficult because its symptoms, particularly high fevers, falling blood counts, and organ dysfunction, overlap with the very conditions that set it off.
What Goes Wrong in the Immune System
At its core, secondary HLH is driven by a runaway feedback loop between certain white blood cells and the chemical signals they release. Normally, immune cells called cytotoxic T lymphocytes and natural killer (NK) cells detect and destroy infected or abnormal cells, then stand down. In secondary HLH, that off-switch fails. T cells and macrophages keep activating each other, pouring out wave after wave of inflammatory molecules in what clinicians call a cytokine storm. Key players in this storm include interferon-gamma, interleukin-6, interleukin-18, and tumor necrosis factor-alpha, among others.
Research into why the off-switch fails has revealed several mechanisms. NK cells in secondary HLH patients show high levels of an inhibitory surface molecule called NKG2A and low levels of an activating counterpart, NKG2D, meaning the cells that should be doing the killing are functionally muzzled. At the same time, cytotoxic T cells express exhaustion markers and secrete less of the signaling molecules they need to coordinate an effective immune response.
The result is paradoxical: immune cells are hyperactive in terms of inflammation but underperforming in terms of actually clearing the threat. Macrophages, overstimulated by the flood of cytokines, begin engulfing normal blood cells, platelets, and even other immune cells, a process called hemophagocytosis. This leads to the plummeting blood counts and organ damage that define the disease.
Infectious Triggers
Infections are the single most common trigger for secondary HLH. In a study of adults in Taiwan, viruses accounted for roughly 41% of infection-associated cases, followed by mycobacteria and other bacteria at about 23% each, and fungi at around 13%.
Among viruses, Epstein-Barr virus (EBV) stands out as the most frequently implicated worldwide. EBV can infect and expand T cells or NK cells in a way that supercharges the cytokine storm. Other herpes-family viruses, HIV, influenza, and more recently SARS-CoV-2 have all been documented as triggers. Bacterial causes range from tuberculosis to common bloodstream infections. Fungal triggers, while less frequent, include organisms like histoplasma and other deep-seated molds or yeasts, especially in people whose immune systems are already compromised. Parasitic infections such as leishmaniasis and malaria can also set off the syndrome.
A crucial point for clinicians and patients alike: the infection that triggers HLH does not have to be unusual or exotic. Common pathogens in the right immunological context can ignite the cascade. The challenge is that the symptoms of the infection and the symptoms of HLH often blur together, making it easy to attribute everything to the infection and miss the secondary syndrome developing underneath.
Autoimmune and Autoinflammatory Triggers
When secondary HLH arises in the setting of a rheumatologic disease, it is often called macrophage activation syndrome (MAS). The two terms describe essentially the same runaway inflammatory process, though MAS is typically reserved for cases linked to autoimmune or autoinflammatory conditions.
The autoimmune disease most closely associated with MAS is systemic juvenile idiopathic arthritis (sJIA) in children. MAS is recognized as one of the most severe complications of sJIA, marked by the same expansion of activated T cells and hemophagocytic macrophages seen in other forms of secondary HLH. In adults, the equivalent condition is adult-onset Still’s disease (AOSD), where MAS can present with multiorgan dysfunction and extremely high ferritin levels. Systemic lupus erythematosus (SLE) is another well-known trigger.
What makes MAS especially treacherous is that it can be hard to distinguish from a severe flare of the underlying autoimmune disease. A patient with active sJIA or SLE may already have fevers, elevated inflammatory markers, and low blood counts, all features that also point toward MAS. The transition from “bad flare” to “life-threatening MAS” can happen over hours, and catching it requires close monitoring of specific laboratory values.
Malignancy-Related HLH
Cancer, particularly blood cancers, is the other major category of triggers in adults. Malignancy is present in close to half of adult secondary HLH cases, making it a far more common driver in adults than in children. The cancers most often involved are lymphomas and leukemias. Among adult patients with malignancy-triggered HLH, T-cell and NK-cell lymphomas account for about 35% of cases, B-cell lymphomas for about 32%, leukemias and Hodgkin lymphoma each for roughly 6%, and other or unspecified blood cancers for most of the remainder. Solid tumors are comparatively rare triggers, representing only about 3% of malignancy-associated cases.
Malignancy-triggered HLH carries a particularly grim prognosis compared with infection-triggered cases. In some instances, the HLH is actually the first sign that an underlying cancer exists, discovered only when clinicians investigate the cause of the inflammatory storm. This scenario underscores the importance of a thorough workup for occult malignancy in any adult presenting with unexplained secondary HLH.
Drug and Therapy Triggers
A growing category of triggers involves medical treatments themselves. The most prominent examples are chimeric antigen receptor T-cell (CAR-T) therapy and immune checkpoint inhibitors, both used in cancer treatment. CAR-T therapy works by reprogramming a patient’s T cells to attack cancer, but this supercharged immune response can sometimes tip over into an HLH-like syndrome known as immune effector cell-associated HLH-like syndrome (IEC-HS). This complication is rare but can be fatal if not recognized quickly, and its symptoms overlap heavily with cytokine release syndrome, another known side effect of CAR-T therapy, making diagnosis a challenge.
Beyond immunotherapies, other medications have been linked to secondary HLH. A recent review identified lamotrigine and various other immunomodulatory drugs as potential triggers. The mechanism in many of these cases likely involves drug-induced immune dysregulation that, in a susceptible individual, pushes the system past its tipping point.
How Secondary HLH Presents
The hallmark triad of secondary HLH is persistent high fever, falling blood counts (especially platelets), and extremely elevated ferritin levels. More than 90% of both adults and children present with all three of these features. Beyond that common core, however, the clinical picture differs by age. Children are more likely to have an enlarged liver and spleen, while adults tend to present in less textbook fashion, with lower rates of hepatomegaly, splenomegaly, and jaundice.
This atypical presentation in adults is one reason secondary HLH is so frequently missed or diagnosed late. An adult in the ICU with high fevers, low platelets, and liver dysfunction may be labeled as having severe sepsis, and the secondary HLH developing alongside or instead of sepsis goes unrecognized.
Neurological involvement adds another layer of complexity. In adults with secondary HLH, central nervous system symptoms can include altered mental status, impaired consciousness, and seizures. Brain imaging reveals abnormalities in over 70% of affected cases, most commonly certain patterns of signal changes and inflammation of the membranes surrounding the brain. Patients with neurological involvement tend to have markedly higher levels of inflammatory markers in their spinal fluid, and central nervous system involvement has been identified as a strong independent predictor of early death in pediatric patients as well.
The Diagnostic Toolkit
There is no single test that definitively confirms secondary HLH. Diagnosis relies on a pattern of clinical features and laboratory findings, evaluated using one of two main scoring frameworks.
HLH-2004 Criteria
The HLH-2004 criteria are the older and more widely known system. They were originally designed as enrollment criteria for a clinical trial in children with primary HLH, not as a diagnostic tool for adults with acquired disease. Despite this, they have been widely adopted for diagnosing secondary HLH in adults, a practice that has drawn significant criticism. A critical review noted clear limitations in applying these pediatric-derived criteria to adult secondary cases, while acknowledging they still help guide understanding of the disease to some extent.
The criteria include fever, an enlarged spleen, low blood counts affecting at least two cell lines, high triglycerides or low fibrinogen, hemophagocytosis on biopsy, low or absent NK-cell activity, elevated ferritin, and elevated levels of soluble interleukin-2 receptor (also called soluble CD25). Meeting five of eight criteria has traditionally been considered diagnostic, though research in critically ill patients found that a cutoff of four fulfilled criteria actually provided better prediction accuracy in that setting, with sensitivity around 95% and specificity around 94%. Adjusting the ferritin threshold upward to 3,000 micrograms per liter and the fever threshold to 38.2°C improved performance further.
The HScore
The HScore is a newer, probability-based tool that assigns weighted points across several categories: fever, organ enlargement, blood counts, ferritin level, triglycerides, fibrinogen, liver enzyme levels, and whether hemophagocytosis is seen on biopsy. Rather than a binary yes-or-no, it produces a probability of HLH. A score of 169 or above predicted HLH with about 96% sensitivity in one multicenter validation study, though specificity dropped to around 71% at that threshold. Raising the cutoff to 200 improved specificity to a level comparable with HLH-2004.
Studies comparing the two systems generally find that the HScore is less restrictive in confirming a diagnosis. One group proposed a cutoff of 166 or above and showed that if four of the seven HLH-2004 criteria are met, there is roughly an 80% probability of HLH by HScore calculation. In practice, many clinicians use both tools in parallel: a high HScore in a patient who also meets several HLH-2004 criteria raises diagnostic confidence considerably.
The Limits of Bone Marrow Findings
Finding hemophagocytosis, immune cells engulfing blood cells, on a bone marrow biopsy is often considered the classic hallmark of HLH. In reality, its diagnostic value is more limited than many clinicians assume. Hemophagocytosis is neither perfectly sensitive nor perfectly specific for HLH. One study reported sensitivity of 83% and specificity of only 60%, with substantial overlap between HLH patients and controls. Another found that the amount of hemophagocytosis in aspirate or biopsy specimens did not correlate with the probability of having HLH, and that an isolated finding of hemophagocytosis, even in high amounts, lacks specificity.
More refined approaches to reading marrow specimens can improve accuracy. Researchers have shown that certain patterns of hemophagocytosis, specifically macrophages engulfing granulocytes, nucleated red blood cells, or multiple nucleated cells simultaneously, are much more strongly associated with HLH than macrophages engulfing ordinary non-nucleated red blood cells alone. Using these more specific histological criteria can push the diagnostic accuracy well above what gross hemophagocytosis alone achieves. Still, the bottom line is that a bone marrow biopsy should never be used in isolation to rule HLH in or out. It is one piece of a larger puzzle.
Telling HLH Apart From Sepsis
The overlap between secondary HLH and severe sepsis is one of the most dangerous diagnostic pitfalls in critical care. Both conditions produce high fevers, organ dysfunction, and severely ill patients. Secondary HLH can develop during an infection that also causes sepsis, so the two are not mutually exclusive, a fact that further muddies the waters.
Several features, when considered together, help distinguish HLH from sepsis. HLH is supported by very high ferritin levels (often far above what sepsis alone produces), an enlarged spleen, marked drops in multiple blood cell lines, low fibrinogen, and a characteristic cytokine profile. Specific biomarkers like soluble CD25, glycosylated ferritin, and CXCL9 have been studied as tools to differentiate HLH from sepsis in critically ill patients. The absence of an identifiable infectious source in a patient who otherwise looks septic is another red flag that should prompt consideration of HLH.
C-reactive protein (CRP), the workhorse inflammatory marker in most hospitals, behaves in a potentially useful way: in HLH, CRP can be disproportionately low relative to the severity of illness, while ferritin climbs to extreme levels. This ferritin-CRP dissociation, while not absolute, is a clue that experienced clinicians look for.
Genetic Susceptibility in Adults
Secondary HLH was long assumed to be a purely acquired condition, triggered entirely by an external event in someone with a normal immune system. Emerging research complicates that picture. Studies using genetic sequencing have found that many adults with secondary HLH carry variants in the same genes associated with the childhood inherited form of the disease. In one cohort, over half of patients with secondary HLH carried at least one variant allele in an HLH-related gene, compared to about 40% of controls. More strikingly, patients with severe disease were far more likely to carry these variants, and all patients carrying two or three variant alleles had severe disease. None of the non-severe patients carried more than one variant allele. The genetic burden was the strongest predictor of severity, outperforming clinical characteristics.
These findings suggest a model in which some adults sit on a genetic spectrum of immune vulnerability. Their NK-cell and T-cell killing machinery works well enough under normal circumstances but falters when challenged by a strong trigger like a particular virus or a lymphoma. Other sequencing work has identified heterozygous mutations in genes linked to T-cell and NK-cell function that may predispose adults to secondary or late-onset HLH. This is a rapidly evolving area of research, and routine genetic testing is not yet standard of care, but the implication for affected families is real: relatives of patients with secondary HLH may carry the same variants and face similar risks under the right circumstances.
Predicting Who Will Do Poorly
Once secondary HLH is diagnosed, outcomes vary enormously depending on the trigger, the speed of treatment, and the patient’s overall condition. Several prognostic markers have been identified. Ferritin levels above 10,000 nanograms per milliliter are consistently associated with worse outcomes. Low levels of NK cells (measured by specific cell-surface markers) have emerged as a strong independent predictor of early death across 30-, 60-, and 90-day time points. Low hemoglobin and low overall T-cell counts add further prognostic information.
A machine-learning-based tool called the HLH-Risk-Calculator, developed from data on 167 adult patients across six European centers, uses soluble interleukin-2 receptor levels and either albumin or platelet counts to predict both initial disease severity and mortality at specific time points out to one year. Soluble interleukin-2 receptor and albumin were the strongest contributors to predicting who would need intensive care, while soluble interleukin-2 receptor and platelet counts best predicted who would die. These tools are still relatively new and require further validation, but they represent a move toward objective risk stratification in a disease where clinical judgment alone has historically struggled.
Prognosis also varies sharply by age and trigger. Adults generally fare worse than children, partly because malignancy triggers are more common in adults and carry a higher mortality rate. Among children, central nervous system involvement, poor nutritional status, and prolonged clotting times have been identified as independent risk factors for early death. Combining standard HLH treatment protocols with blood purification techniques appeared to be protective in one pediatric model, though data remains limited.
Why It Gets Missed
Even with scoring systems and biomarkers, secondary HLH is underdiagnosed. Several factors conspire against timely recognition. The disease mimics sepsis, and in an ICU or emergency department, the default assumption when a patient has high fevers and organ failure is infection until proven otherwise. Many physicians outside of hematology and rheumatology have limited familiarity with HLH, and the specialized lab tests that help clinch the diagnosis, such as soluble CD25 and NK-cell function assays, are not available on a stat basis at most hospitals. Ferritin is widely available but is also elevated in many other conditions; its value as a screening tool for HLH depends on how high it goes and whether it rises out of proportion to other acute-phase reactants.
There is also a conceptual barrier. The HLH-2004 criteria require meeting a numerical threshold of features, but secondary HLH can unfold over days, with criteria being met one at a time. Waiting for five out of eight criteria to be satisfied before starting treatment may mean waiting too long. The HScore’s probability-based approach partially addresses this issue by allowing clinicians to quantify risk on a continuum rather than waiting to cross a threshold, but adoption is uneven. The sheer rarity of the condition, coupled with its lethality when missed, makes it one of those diagnoses where a high index of suspicion matters more than any algorithm.