Systematic literature review

Briefly, original research articles of any study design with diagnosis, treatment and monitoring of HLH/MAS that reported more than six cases were included. Of the 18 020 articles from PubMed, EMBASE and Cochrane, 258 were selected for full-text review and 167 articles were included for data extraction. Based on the expertise of TF members, SLR results12 and discussion at consensus conferences, the TF generated 6 overarching principles (table 3) and 24 disease-specific PTC pertaining to HLH/MAS (table 4).

Overarching principles

Recognising the complexity and urgency of management decisions in systemic hyperinflammation and HLH/MAS, the TF generated six overarching principles (table 3) that provide guidance on the early recognition of characteristic clinical features, the systematic evaluation of contributors (including genetic causes, predisposing conditions and acute triggers), the implementation of early therapies and the monitoring of HLH/MAS progression.

Points to consider

The TF generated specific statements intended to offer practical consensus-based and evidence-based guidance for clinicians making decisions at the earliest stages of HLH/MAS consideration, recognition and management (table 4).

PTC 1.1–1.5: recognition, screening and early diagnosis

Given the variety of genetic causes, predisposing conditions and acute triggers from which HLH/MAS arises, recognising the presenting features and making a diagnosis are often challenging. Existing diagnostic criteria lack both sensitivity and specificity, especially in the context of confounding conditions like lymphoma or sepsis.

Based on existing criteria, current literature and expert experience, the TF agreed on clinical and laboratory abnormalities that together establish a recognisable pattern of potentially life-threatening HLH/MAS (PTC 1.1). Individual findings are non-specific and must be evaluated collectively and longitudinally. However, recognising the pattern of clinical and laboratory abnormalities that constitute HLH/MAS is critical for prompting an aetiological workup, considering treatments and initiating a monitoring strategy before serious complications or death occur.

Ferritin is a sensitive test for HLH/MAS, and there was broad consensus that ferritin levels should be checked in all patients with new, ongoing or heightened suspicion for HLH/MAS even if prior measurements have been normal (PTC 1.1–1.3). Essentially all patients with HLH/MAS with systemic disease have elevated ferritin levels,14 15 and hyperferritinemia is part of all existing HLH/MAS criteria (table 1). Levels >500 ng/mL were 84% sensitive in paediatric patients with HLH,2 and served as the cut-off in clinical trials conducted by the Histiocyte Society, but this level is associated with poor specificity in other contexts and higher ferritin cut-off values have been used.12 The ferritin cut-off values used in paediatric HLH/MAS studies (500–2000 ng/mL) tend to be lower than in adult studies (often >10 000 ng/mL),2 6 7 where infectious and malignant contributors predominate.16 Other conditions such as iron overload, malignancy and hepatitis commonly induce high ferritin levels even in the absence of HLH/MAS.17

Abnormalities in other widely available clinical and laboratory indicators of inflammation, coagulopathy or organ damage/dysfunction also raise the level of suspicion for HLH/MAS (PTC 1.4, table 5). However, many HLH/MAS-associated biomarkers may also indicate parallel inflammatory processes (eg, elevated LDH in thrombotic microangiopathy).18 More specialised biomarkers measuring key HLH/MAS pathways (PTC 1.5, table 5, online supplemental table 1) are increasingly available from reference laboratories. These include measures of activation of T cells (soluble interleukin (IL)-2 receptor-α/CD25, T-cell HLA-DR isotype expression), macrophages (CD163, neopterin), inflammasomes (IL-18) and the interferon-gamma pathway (IFNγ, CXCL9). Their relative specificity (compared with other inflammatory parameters in table 5) is helpful in confirming an HLH/MAS diagnosis and in monitoring. The TF recommended assessment of specialised inflammatory biomarkers, interpreted with the aid of consultants, when available (PTC 1.5). Longitudinal assessment of both routine and specialised HLH/MAS biomarkers improves their diagnostic utility and is essential for monitoring for progression or resolution (as discussed below).19

PTC 2.0: existing criteria

Several criteria have been developed to identify patients with syndromes that may represent HLH or MAS (table 1, online supplemental table 2). The HLH-94 criteria (refined in HLH-04) were developed to classify infants and children for treatment trials targeting paediatric patients with genetic causes of HLH/MAS.2 5 The MAS-2016 criteria were developed to classify MAS in patients with known or strongly suspected sJIA.6 The HScore was developed in adults with primarily malignancy or infection-associated HLH,7 and the MS score to distinguish MAS from underlying sJIA.8 The HScore and HLH-2004 criteria have been validated in some additional contexts.20–23 The TF agreed that each set of criteria were useful within the context in which they were developed, but that no existing set of criteria was a sufficient diagnostic tool in all settings and populations (PTC 2). There is substantial feature overlap between criteria (table 1, online supplemental table 2).

PTC 3.1: evaluating contributors

The TF emphasised the critical importance of timely identification of underlying contributors (genetic causes, predisposing conditions and acute triggers as described in ‘Introduction’ section), often in a rapidly evolving and ill patient (PTC 3.1). A thorough workup should begin immediately on suspicion for HLH/MAS and should be tailored to the most likely contributors, paying particular attention to the patient’s age, family history, infectious exposures/risks, recent treatments and underlying conditions. Although HLH/MAS is thought to result from the interaction of multiple host and environmental contributors, available data typically implicate a single aetiology (as reflected in table 6 and more thoroughly in SLR manuscript).12 Additionally, >2000 case reports and series demonstrate that HLH/MAS can occur in most settings that provoke an immune or inflammatory response.12

Table 6

Proportion of attributable HLH/MAS cases by primary contributor*

Genetic causes of HLH/MAS represent a minority of all cases (particularly in adults), but they have made essential contributions to diagnostic and treatment advances. The IEI include nearly 500 genetically defined disorders,24 and for most of these HLH/MAS is a rare complication. The canonical high-penetrance genetic causes of HLH are those that profoundly impair granule-mediated cytotoxicity as well as the X linked lymphoproliferative syndromes (table 7). The distinction between genetic causes and variants conferring susceptibility has grown less clear with time. Nevertheless, the identification of a genetic cause/contributor has profound implications (as discussed below).

Table 7

Genes associated with HLH/MAS susceptibility*

Among predisposing conditions, malignancy (especially lymphoma) is a major contributor to HLH/MAS. Investigation for underlying malignancy should be considered in all patients with HLH/MAS, particularly in adults where it occurs in nearly half of cases.7 20 25–37 Although MAS is most recognised and best studied in sJIA and adult-onset still disease (AOSD), SLE may be a more common cause of HLH/MAS in adults in part due to its higher prevalence.38–40

Infection is the most common acute trigger of HLH/MAS. In children, infection is the most common aetiology, with a specific pathogen identified in over 50% of new HLH/MAS presentations. Broad testing for infection (eg, blood and other cultures, viral PCR, etc) should be pursued based on clinical scenario. Some infections warrant special attention for their role in HLH/MAS. Epstein-Barr virus (EBV) is a well-known trigger of HLH, particularly in individuals with genetic (table 7) or acquired immunodeficiency or certain malignancies. It is unclear why the incidence of EBV-HLH appears higher in Asia, but this is consistent with other EBV-triggered phenotypes.31–34 41–49 Region-specific and season-specific infections should also be considered as causes of HLH/MAS in endemic areas, including dengue virus in tropical/subtropical climates,46 50 51 histoplasmosis in the mid-western and southern USA,31 and less frequently malaria,51 52 tuberculosis,7 25 26 32 33 36 48 scrub typhus,33 47 50 typhoid fever,42 47 50 51 tickborne diseases42 43 53 and leishmaniasis.25 42 54 Although inflammation in patients with COVID-19, multisystem inflammatory syndrome in children (MIS-C) or adults (MIS-A) rarely rises to meet HLH or MAS criteria, the pandemic normalised the need to identify and treat (SARS-CoV2) infection-associated immunopathology. Testing for genetic contributors should be considered regardless of the type of infection, particularly in young children.

The increasing use of immune effector cell cancer therapies, including chimeric antigen receptor T cells, has elevated their recognition as iatrogenic triggers of HLH/MAS. The incidence of CRS and immune effector cell-associated neurological syndrome (ICANS) is highest in leukemias and lymphomas. The distinctions between CRS/ICANS and HLH are unclear and their ideal treatment is an evolving target beyond the scope of this effort.55–58

For many patients, multiple contributors interact to drive HLH/MAS. Acute infections are identified in the majority of patients with HLH/MAS with predisposing rheumatological conditions (>65%) or IEIs (>80%).54 59–62 In adults, the presence of multiple pathogens at the time of HLH/MAS diagnosis may increase risk of mortality.63 Thus, concern for HLH/MAS should prompt consideration of multiple potential contributors regardless of known underlying conditions or aetiologies.

PTC 3.2–3.5: genetic testing

Genetic causes for HLH/MAS are likely under-recognised and their identification profoundly affects treatment, prognosis and genetic counselling (PTC 3.2, table 7). For example, screening for CNS involvement is particularly important in genetic HLH.47 Early recognition of familial haemophagocytic lymphohistiocytosis may accelerate allogeneic haematopoietic stem cell transplantation (HSCT) and can support HSCT in affected presymptomatic siblings.64 65 Some HLH/MAS therapeutic trials include or exclude specific genetic causes (ClinicalTrials.gov identifiers NCT04641442, NCT03113760).

When to perform genetic testing, on whom, what test(s) to send and how to interpret detected variants are complex and evolving decisions (PTC 3.3). Features suggestive of a genetic cause include young age at presentation, positive family history, consanguinity and prominent CNS disease. HLH/MAS due to cytotoxicity defects tends to present in infancy and early childhood, whereas HLH/MAS in other IEI (particularly those with EBV immunodeficiency, table 7) present in a broader age range including older children.66 Although genetic HLH has presented in adulthood, actionable results of genetic testing in adult HLH/MAS are rare.25 67 68 Other relevant clinical features/contexts like albinism, inflammatory bowel disease, isolated CNS involvement and EBV-immunodeficiency suggest specific genetic causes (PTC 3.3, table 7).

Given the high prevalence of genetic causes in children and the large clinical impact of a positive finding, the TF supported early genetic testing in children and high-risk adults, preferably using multigene panels or whole exome/genome sequencing (PTC 3.4). Single-gene sequencing remains appropriate with family history of a known genetic HLH disorder, characteristic clinical features (eg, albinism), positive protein or functional testing (eg, perforin flow cytometry) or in resource-limited settings. Genetic counselling is warranted for all patients undergoing genetic testing (PTC 3.5).

PTC 4.1–4.4: prognostic factors and CNS involvement

Prognosis in HLH/MAS is dependent on multiple factors, including the nature of the underlying contributors, degree of organ dysfunction and duration of active disease (PTC 4.1). HLH/MAS can be fatal in any context, but relative to other causes malignancy-associated HLH/MAS is associated with worse survival33 66 69 70 and HLH/MAS in rheumatic diseases has a more favourable prognosis.3 40 66 71–73 EBV is associated with poor prognosis when present in patients with genetic immunodeficiency or predisposing (rheumatic or malignant) conditions,42 47 69 but prognosis appears better in patients with EBV as the sole contributor.45 46 Specific patterns of organ injury also predict poor outcomes (PTC 4.1). Liver involvement is common and can progress to life-threatening liver failure.47 Patients with multiorgan dysfunction often require treatment in an intensive care unit (ICU) setting, a strong predictor of poor outcome.50 53 74 75

CNS involvement is sufficiently common, insidious and dangerous to warrant specific attention. It is associated with both mortality and long-term neurological sequelae in survivors regardless of underlying contributors.3 42 47 76–80 It should be suspected in all patients being evaluated for HLH/MAS (PTC 4.2). CNS manifestations of HLH/MAS can be broad (tables 1 and 5), and cerebrospinal fluid (CSF) and imaging findings usually demonstrate evidence of inflammation in affected patients.76 Incidence of CNS involvement varies by age and aetiology,45 47 76 and children with HLH/MAS are at higher risk than adults, especially children with genetic causes.47 76 81 82 Some degree of CNS involvement is present in a sizeable percentage of children with EBV-HLH,83 sJIA-MAS71 and adults with secondary HLH/MAS.80 84 85

CNS involvement should be considered in all patients, and all should undergo a complete neurological examination. Patients presenting under 1 year of age, those otherwise suspected of having familial disease47 and any patient with symptoms or signs concerning for CNS dysfunction (including an unreliable exam) should undergo assessment for CNS involvement (PTC 4.2). Assessment for CNS involvement may include CSF evaluation (glucose, protein, cell counts and often cytological review) and contrast-enhanced brain MRI as well as other testing (electroencephalogram, MR angiography, spinal imaging) as clinically indicated (PTC 4.3). Full evaluation often must await stabilisation of cardiorespiratory function, coagulopathy or intracranial pressure. Providers should not delay empiric or context-specific treatments in order to complete the CNS workup (PTC 4.4).

PTC 5.1–5.5: treatment considerations

Treatment of patients with suspected HLH/MAS requires a dynamic risk-benefit assessment. Consideration of HLH/MAS-directed immunomodulation should occur simultaneously with diagnostic evaluations (PTC 5.1–3), treatment of contributing factors (PTC. 5.4) and prevention of complications (PTC 5.4–5.5).49 77 86 Figure 1 is intended to depict how these PTC on early diagnosis, monitoring and management may function in practice, in relation to each other and relative to the goal of context-specific treatment. Age-appropriate supportive care should follow accepted guidelines, such as the Surviving Sepsis Campaign,87 88 and its provision, as well as the frequency of monitoring (as discussed below) may require intensive care. ICU admission was required in over a third of children with HLH/MAS and nearly half of adults with MAS.61 In children and adults requiring ICU admission for HLH/MAS, nearly 70% required mechanical ventilation or vasopressors/inotropes and nearly half required renal replacement therapy.32 50 Use of intensive care appears higher in HLH/MAS occurring in context with worse outcomes, like malignancy.34 71 89

Figure 1

Summary of the approach to early or suspected HLH/MAS. When HLH/MAS is suspected, providers should (in parallel and as clinically appropriate) assess for the key features of HLH/MAS; investigate suspected contributors and treat with supportive care, with empiric and prophylactic antimicrobials, with other prophylaxis regimens, and possibly with empiric immunomodulation. Ongoing monitoring and reassessment should prompt re-evaluation of treatments being given. Patients should transition to context-specific treatment immediately on identification of a confirmed aetiology. *Addressed in separate guidance documents, see www.histiocytesociety.org/HLH-consensus. α-IFNγ, interferon-gamma neutralising antibody; CBC+diff, complete blood count with leucocyte differential; CMV, cytomegalovirus; CRP, C reactive protein; CSF, cerebrospinal fluid; DIC, disseminated intravascular coagulopathy; DVT, deep vein thrombosis; EBV, Epstein-Barr virus; ESR, erythrocyte sedimentation rate; GC, glucocorticoids; GGT, γ-glutamyl transferase; HLH94, HLH-94 treatment protocol 1 or current standard of care; IVIg, intravenous immunoglobulin; LDH, lactate dehydrogenase; LFTs, liver function tests; PT/PTT, prothrombin time/partial thromboplastin time; TG, triglycerides.

Choosing and adjusting empiric immunomodulation for suspected HLH/MAS can be challenging. Decision-making must integrate HLH/MAS severity and rate of progression, specific organ involvement, likely contributors, comorbid conditions and concurrent medications (figure 1). Ideally, targeted immunomodulation would be initiated as early as possible (PTC 5.1) and neither induce immunosuppression nor compromise the aetiological workup. In practice, determining the target and balancing these risks are essential, patient-specific challenges. Although no studies have evaluated empiric treatment of HLH/MAS prior to/regardless of aetiology, immunomodulatory treatment has dramatically improved survival in most aetiologies of HLH/MAS.12 49 90 91 In patients with high-risk features or progressive HLH/MAS, the TF strongly recommends considering empiric immunomodulation during the initial evaluation and management period (PTC 5.1–3). Once there is sufficient understanding of a patient’s underlying contributors, management should shift to context-specific treatments and recent context-specific guidance documents may be helpful in this transition67 74 92–94 (figure 1, asterisks).

The TF currently endorses use of glucocorticoids (GCs), the recombinant IL-1 receptor antagonist (IL-1RA) anakinra and/or intravenous immunoglobulin (IVIg) for empiric immunomodulation in suspected HLH/MAS (PTC 5.3, figure 1, table 8). Multiple treatments may be initiated concurrently depending on clinical context and availability. Published treatment data demonstrate the strongest support for GCs across all forms of HLH/MAS.12 The choice of GC formulation (most commonly prednisone, prednisolone, dexamethasone (DEX) or methylprednisolone (MP)) and route of administration (oral vs intravenous) should be tailored to the patient and care setting (table 8). ‘Pulse’ doses of intravenous MP (10–30 mg/kg/day up to 1 g, given daily) are effective in severe rheumatic and neuro-inflammatory diseases,95–97 and have been used successfully in HLH/MAS.61 DEX is used in HLH treatment protocols due to better CNS penetration at an initial dose of 10 mg/m2/day (~2–4 mg/kg/day of MP). Given DEX’s long half-life, shorter-acting GCs may be preferable in rapidly evolving diagnostic scenarios.

Table 8

Empiric HLH/MAS immunomodulatory treatment dosing*

Importantly, GC administration may obscure pathological diagnosis and/or staging of leukaemia or lymphoma.98 Therefore, definitive testing for malignancy (typically biopsy/aspirate of bone marrow, lymph node and/or other indicated tissues) should be attempted prior to GC administration when possible. GC-related immunosuppression depends on dose, duration of exposure and relevant pathogens. Although GC treatment (alongside appropriate antimicrobial treatment) prevents immunopathology in many localised infections,99 100 large studies have not supported its utility for immunomodulation in sepsis. Thus, the role of GC in infection-associated HLH/MAS remains patient-dependent and pathogen-dependent. Providers should monitor for other dose-dependent GC side effects like hyperglycaemia, hypertension, myopathy and psychosis.

Empiric use of anakinra and/or IVIg in early, evolving or undifferentiated HLH/MAS may provide immunomodulation without significant immunosuppression and without impairing malignancy workup. The TF supported their inclusion despite sparse data due to good pharmacological and safety profiles, strong efficacy in other systemic inflammatory diseases and significant clinical experience. Anakinra is a safe and effective treatment for many autoinflammatory and rheumatic disorders. Its rapid onset and short half-life may be desirable in rapidly evolving patients. Even used at high doses in adults with bacterial sepsis (up to 48 mg/kg/day), it showed no signal for immunosuppression and appeared to limit mortality in patients with sepsis with hepatobiliary dysfunction and coagulopathy.90 101 A retrospective study in secondary HLH supported the safety and possible efficacy of early anakinra use in controlling inflammation.102 IVIg has demonstrated efficacy in Kawasaki disease, and it neither obstructs cancer workup nor suppresses immune function. Notably, serological testing should be sent from samples taken prior to IVIg when feasible. Reports of its efficacy in HLH/MAS are restricted to case series. High-dose IVIg is also a substantial colloid load that can compromise cardiac function and worsen oedema. It rarely causes haemolysis or aseptic meningitis.

Clinical context is essential when considering escalation or context-specific treatment(s), and clinicians are encouraged to consult with local, regional or national experts on a case-specific basis. B-cell depletion may be useful in some patients with EBV-HLH.103 104 Early initiation of treatment regimens centred around the chemotherapeutic etoposide (guidelines in Ehl et al 92) have been life-saving for patients with primary HLH and severe EBV-HLH.49 77 Evidence for high-dose etoposide is less favourable for HLH/MAS in the context of sJIA/AOSD, lthough lower doses may be useful.105 It is not indicated for most non-EBV infections.34 106–108 The utility of etoposide in malignancy-associated HLH is currently unclear.93 109

For patients with increasing inflammation and/or worsening organ damage despite early immunomodulation, treatment escalation with higher doses of GC and/or alternative agents (table 9) should be considered in consultation with HLH/MAS experts. Increasing evidence supports the involvement of the IFNγ pathway in HLH/MAS. The IFNγ neutralising antibody, emapalumab, was recently approved in the USA for the treatment of refractory, recurrent, or progressive HLH.110 Ruxolitinib (and other JAK inhibitors) broadly targets cytokine signalling, including IFNγ, and has shown promising early results in HLH/MAS.111–113

Table 9

Other immunomodulatory therapies used in HLH/MAS*

Alongside HLH/MAS-directed immunomodulation, treatment of contributing factors is critical for optimising outcomes (PTC 5.4). This will often include empiric antimicrobial and sometimes antiviral agents, accounting for exposures/geography, comorbidities (eg, renal failure) and chronic immunosuppression. Like other aspects of HLH/MAS treatment, infectious prophylaxis should be considered early and revisited as the patient and workup evolve (PTC 5.5). Secondary infections can complicate both the inpatient and outpatient course of HLH/MAS. Antifungal and Pneumocystis jirovecii pneumonia prophylaxis are recommended, and are part of the HLH-94 protocol.5 Antifungal and antiviral prophylaxis were administered in more recent trials with newer agents like emapalumab.110 In addition to assistance with empiric treatment, consultation with immunocompromised infectious disease specialists may aid prophylaxis planning (PTC 5.5).

PTC 6.1–6.3: monitoring

Monitoring for disease progression, new organ involvement and damage and response to treatment begins on suspicion for HLH/MAS. Monitoring plans should be tailored to severity, organ involvement and likely contributors of HLH/MAS. Many of the biomarkers useful for diagnosing HLH/MAS also have prognostic relevance (table 5, online supplemental table 1, PTC 6.1). For example, both higher initial ferritin levels and failure of ferritin to improve during therapy associate with worse outcomes.71 114–118

No comparative studies evaluate the ideal laboratory monitoring protocol. Given the propensity for rapid clinical changes, initial monitoring may include daily assessment of inflammatory biomarkers (eg, CRP, ferritin), indicators of organ damage (eg, CBC, fibrinogen, ALT) and any drug-specific monitoring. More frequent monitoring may be needed for evolving or critically ill patients and may require ICU care74 (PTC 6.2, table 5). Lack of response to initial therapy should prompt a careful re-examination of both underlying diagnoses and therapeutic approach. When available, more specific HLH/MAS biomarkers like sIL-2Ra, IL-18 and CXCL9 should be monitored less frequently than conventional disease activity measures like ferritin and CRP (PTC 6.5, table 4). CXCL9 may be particularly useful for monitoring response to IFNγ-blocking therapies.110 Specialised tests may also be helpful in distinguishing HLH/MAS relapse from acquired infection or drug reaction. Treatment response and dose-escalation criteria used in HLH/MAS trials also reflect the overlap between diagnostic and monitoring tests.12

PTC 7.0: multidisciplinary teams

Mounting evidence suggests that a multidisciplinary team experienced in managing HLH/MAS may improve recognition, reduce immunosuppression and improve outcomes (PTC 7.0).119 120 Such response teams often include representatives from infectious diseases, haematology/oncology, rheumatology, immunology, pharmacy and other relevant specialties, although their optimal composition and function has not been established. Their goals include improving early identification, streamlining communication and improving consistency of care. These groups may also be able to better incorporate new findings, conduct quality improvement and engage in collaborative research.