Secondary Hemophagocytic Lymphohistiocytosis, Macrophage Activation Syndrome, and Hyperferritinemic Sepsis-Induced Multiple-Organ Dysfunction Syndrome in the Pediatric ICU
- 722 Downloads
Children with secondary hemophagocytic lymphohistiocytosis present diagnostic and therapeutic challenges to the intensivist. Management of this condition can be facilitated with a collaborative approach among a team of intensivists, immunologists, rheumatologists, hematologists, and infectious disease specialists. Controversy over management approach commonly centers on incomplete understanding at the time of clinical presentation regarding individual patient differences in inflammation pathobiology that require different therapies. This review provides the clinician with the rationale for personalizing care for these critically ill children.
KeywordsChild Critical illness Histiocytosis Macrophage activation syndrome Multiple-organ failure Rheumatologists Sepsis
In the past decade, hemophagocytic lymphohistiocytosis (HLH) has been increasingly recognized in critically ill pediatric patients. HLH can be familial or secondary, with most critically ill patients who develop this disease process having the latter. Secondary HLH (sHLH) that occurs in critically ill patients is classically referred to as macrophage activation syndrome (MAS) when triggered by drugs or when it occurs in patients with underlying rheumatologic disease and hyperferritinemic sepsis-induced multiple-organ dysfunction syndrome (MODS) when it is precipitated by infectious etiologies. As awareness of these entities has increased, clinical intensivists have been faced with new challenges that have precipitated diagnostic and therapeutic controversies. The need for prompt and aggressive immune-modulating treatment in patients with sHLH requires early recognition and diagnosis. However, clinical and pathologic similarities between sHLH and non-hyperferritinemic sepsis, absence of optimal diagnostic tools in sHLH, and suboptimal understanding of this entity by clinicians often result in delayed or missed diagnoses.
Clinical Vignette 1: Familial Versus Secondary HLH
A previously healthy infant is admitted to a pediatric ICU after presenting with status epilepticus. The admitting intensivist elicits a history of consanguinity and another child in the family dying from “fever,” both of which are risk factors for familial HLH. During the subsequent 48 h of her hospital course, she develops several clinical criteria for familial HLH (Fig. 14.1). A bone marrow aspirate is obtained, and the infant is started on a familial HLH treatment regimen of dexamethasone and etoposide. The child recovers and is discharged to home.
One month later, genetic analysis confirms the diagnosis. Specifically, it reveals a homozygous UNC13D gene variant which leads to ineffective natural killer cell eradication of virus infection and ineffective activated immune cell death. Based on these findings, the child undergoes bone marrow transplantation.
Familial HLH is typically triggered by a viral infection in a host who has a genetically determined inability to induce granzyme-perforin-mediated cytolytic killing. In these children, nonmalignant lymphoproliferation leads to high levels of lymphocyte-derived interferon-γ that activates macrophages. Treatment with dexamethasone and etoposide induces lymphocyte apoptosis and reduces macrophage activation. In contrast, secondary HLH is associated with macrophage activation in patients without hereditary predisposition to such a process, and it is typically triggered by severe systemic inflammation such as that occurs in sepsis.
In 1988, Suster and colleagues reviewed bone marrow, lymph node, and spleen histology from 230 consecutive intensive care unit adult autopsies and identified cases of histiocytic hyperplasia with hemophagocytosis (HHH) . They reported moderate to severe HHH in 102–230 bone marrow specimens (44%), 79 of 191 lymph node specimens (41%), and 16 of 209 spleen specimens (8%). There was a strong blood transfusion dose relationship, with patients who received more than 5 transfusions nearly 60 times as likely to have HHH as compared to patients who received no transfusions. Bacterial sepsis (adjusted risk odds ratio 4.1) was also independently associated with HHH.
More recently, Strauss and colleagues  evaluated 107 consecutive medical ICU patient autopsies and found mild to severe HHH in 69 (65%) . The authors similarly found HHH to be associated with sepsis and a number of blood transfusions. Patients with HHH were significantly less likely to have died due to cardiovascular causes (HHH 32% versus no HHH 74%) and more likely to die due to MODS (HHH 39% versus no HHH 18%), with a characteristic organ failure pattern of elevated bilirubin, liver enzymes, and disseminated intravascular coagulation. Patients with HHH were also more likely to require catecholamine infusions, mechanical ventilation, and renal replacement therapy. Autopsies with histology graded severe HHH also had more siderosis suggesting iron overload and more CD8 T cells in the bone marrow suggesting T-cell activation. These two autopsy studies associated HHH with more severe critical illness but could not determine if it was a novel and clinically relevant process or a secondary phenomenon.
Since these seminal reports were published, nomenclature for HHH has changed, with the aforementioned clinical descriptors now being used for secondary hemophagocytic lymphohistiocytosis (sHLH) – macrophage activation syndrome (MAS) for patients with sHLH precipitated by drugs or rheumatologic disease or hyperferritinemic sepsis-induced MODS, depending on the clinical setting involved [3, 4, 5, 6, 7].
Experimental Models of sHLH
Experimental models provide support for macrophage (i.e., histiocyte) activation associated with sHLH as an important pathway to MODS. Steinberg and colleagues developed the sterile model of zymosan (the cell wall of the fungus saccharomyces A) plus mineral oil injected intraperitoneally to induce MODS in rodents . This model results in initial hypovolemic shock followed by persistent macrophage activation. Injection of either zymosan or mineral oil alone does not induce MODS, suggesting the need for a “two-hit” insult of both toll-like receptor (TLR) stimulation by zymosan and unremitting particulate irritation by mineral oil to induce persistent macrophage activation. In another rodent model, Behrens and colleagues reported that repeated (i.e., not single time) TLR9 stimulation with CpG oligodeoxynucleotides transformed an otherwise innocuous endotoxin challenge (TLR4 stimulation) in mice into a MODS model of macrophage activation with cytopenias, splenomegaly, hyperferritinemia, and hepatitis . Similarly, in another murine model of cecal ligation and perforation-induced sepsis, additional CpG injection induced cytokine production by macrophages and hepatic mononuclear cells, followed by the development of liver injury and MODS-induced mortality . Importantly, the sHLH phenotype elicited with repeated TLR9 stimulation is exacerbated in knockout mice deficient in native hepatic IL-1 receptor antagonist protein production but can be ameliorated by interferon-α-induced production of IL-1 receptor antagonist protein or with direct administration of recombinant IL-1 receptor antagonist protein (anakinra) in vivo . In short, liver dysfunction in this model appears to be related in part to IL-1-mediated inflammation .
Different Inflammation Pathobiologies in sHLH
Absent NK cell cytolytic activity is the basis for familial HLH, which itself is comprised of a group of monogenic autosomal recessive or X-linked primary immune deficiency diseases characterized by the absence of crucial components of the perforin-granzyme pathway needed for NK cells to kill viruses and induce apoptosis in cancer and host inflammatory cells. Gene knockout models for this pathway result in murine HLH and death after infection with otherwise innocuous LCM (lymphocytic choriomeningitis) virus infection. In the wild-type mouse exposed to LCM virus infection, host NK and T cells respond to control the virus, whereas in the perforin knockout mouse, LCM virus infection cannot be killed leading to T-cell proliferation and activation and consequent overproduction of interferon-γ (reflected by very high levels of CXCL9, the monokine induced by interferon-γ). T lymphocyte-derived interferon-γ induces macrophage activation and organ injury. Etoposide, one of the traditional treatments for familial HLH, destroys proliferating activated T cells and reduces interferon-γ production, preventing undue macrophage activation.
Reduced NK cytolytic activity is the basis of rheumatologic or drug-induced MAS. For instance, some patients with systemic juvenile idiopathic arthritis (sJIA), a known trigger for sHLH, are hypomorphic or heterozygotes for the perforin-granzyme pathway gene variants. As heterozygotes, these patients have some NK cytolytic activity. For the most part, these patients respond nicely to anti-inflammatory therapies such as anakinra (IL-1 receptor antagonist protein). The IL-1 receptor antagonist protein anakinra is FDA approved for inflammasome-driven conditions (e.g., sJIA). IL-18, the interferon-γ-inducible factor, is also increased in these patients. There is optimism that IL-18-binding protein (which neutralizes IL-18) as well as interferon-γ monoclonal antibody could help these patients.
Reduced number of NK cells with normal NK cytolytic activity per cell is the basis of hyperferritinemic sepsis-induced MODS. When NK cell numbers recover, then inflammation and MODS resolve. During bacterial infection, NK cells switch from an overall low cytokine-producing, high cytolytic activity phenotype to a high cytokine-producing, low cytolytic activity phenotype . In hyperferritinemic sepsis-induced MODS, NK cell cytopenia and T-cell cytopenia occur with reticuloendothelial system activation but decreased to absent interferon-γ production. When the reduction in NK cell and T-cell numbers is below 10% of normal, there is a significant decrease in host ability to kill viruses and cancer cells as well as to induce apoptosis in activated macrophages. Because T cells and interferon-gamma production are already low to absent, etoposide is unlikely to be of benefit in reducing macrophage activation in these children. Indeed, etoposide may worsen outcomes in sepsis patients by preventing the recovery of lymphocyte counts needed to resolve infection. Compared to patients with other forms of HLH, those with hyperferritinemic sepsis-induced MODS have lower production of interferon-γ-inducing IL-18. The promise of interferon-γ monoclonal antibodies in treating familial HLH and MAS is therefore less likely to be realized in hyperferritinemic sepsis-induced MODS. Further, inability to produce interferon-γ is associated with increased mortality in experimental models of sepsis.
Diagnosis and Therapeutic Options for Secondary HLH
Despite different pathobiologies, familial HLH and both forms of sHLH – MAS induced by rheumatologic disease and hyperferritinemic sepsis-induced MODS – are all characterized by uncontrolled macrophage activation and are traditionally diagnosed by the presence of five of eight clinical criteria that include ferritin >500 ng/ml, two-line cytopenia, organomegaly, hypertriglyceridemia, hypofibrinogenemia, elevated soluble CD25, absent natural killer (NK) cytotoxic activity, and hemophagocytosis (Fig. 14.1). Importantly, though these criteria were initially created to diagnose familial HLH, there are some clinical data supporting approaches to the treatment of sHLH using this diagnostic methodology [4, 5, 6, 7]. Additionally, Demirkol et al. evaluated different therapies for Turkish children who met criteria for sHLH in a cohort study . They excluded children who were under 2 years, had a history of consanguineous parenting, or had a previous young family member who died from fever because these children were more likely to have familial HLH. Mortality in these excluded children, who were treated by hematologists with etoposide and dexamethasone, was 50%. All the included children (i.e., without these familial HLH risk factors) were diagnosed with sHLH and had five of the aforementioned eight clinical criteria and had five to six organ failures. Centers in one treatment cohort administered the familial HLH protocol of dexamethasone and/or etoposide along with daily plasma exchange to these children with sHLH and observed a 50% mortality rate, whereas centers in the other treatment cohort administered a less immune suppressive regimen of methylprednisolone with or without intravenous immunoglobulin (IVIG) with daily plasma exchange and observed a 0% mortality rate. Based on the above data, though it may be appropriate to use the diagnostic criteria for familial HLH to diagnose sHLH, the traditional familial HLH protocol of dexamethasone and etoposide should be replaced by less toxic treatment strategies for patients with sHLH.
Clinical Vignette 2: sHLH Induced by Systemic Juvenile Arthritis
A febrile 7-year-old child presents to an ICU with rash, leukocytosis, arthritis, and laboratory data concerning for sHLH. The intensivist calls the rheumatologist, who believes the presentation is consistent with systemic juvenile arthritis-related MAS. She recommends treatment with methylprednisolone and the IL-1 receptor antagonist anakinra. Other laboratory testing is sent to rule out systemic lupus erythematosus, sarcoidosis, scleroderma, Sjogren’s syndrome, and Kawasaki’s disease.
Patients with autoimmune rheumatologic disease have increased inflammasome activation and reduced NK activity without lymphoproliferation . Accordingly, rather than etoposide and dexamethasone regimens, methylprednisolone and anakinra are typically recommended to control inflammasome activation associated with rheumatologic disease. Other biologics and chemotherapeutic regimens, however, such as cyclophosphamide, methotrexate, tocilizumab, or etoposide as well as plasma exchange, are considered if the patient remains recalcitrant.
Because three of the eight clinical criteria (e.g., soluble CD25 levels, NK cytotoxicity, and hemophagocytosis) used to identify patients with HLH are not easily accessible tests, rheumatologists have sought to redefine sHLH with other criteria sets that use more readily available laboratory tests. For example, utilizing current literature, Ravelli and colleagues have provided a consensus statement defining sHLH in a child with known systemic juvenile idiopathic arthritis (sJIA), a common trigger of macrophage activation, based on the presence of fever, ferritin >684 ng/mL, and any two of the following: platelet count <181 K, ALT >48 IU/L, triglycerides >156 mg/dL, and fibrinogen <360 mg/dL . Though recommended for patients with MAS secondary to sJIA, these criteria can be applied to patients with MAS secondary to other rheumatologic diseases as well.
Shakoory and colleagues have offered another method for simplifying the diagnosis of and treatment of sHLH . Specifically, they suggested that the combination of hepatobiliary dysfunction and disseminated intravascular coagulation (DIC) can be representative of sHLH including hyperferritinemic sepsis-induced MODS. Further, they hypothesized that if the combination of these two organ dysfunctions represents sHLH, then treatment with IL-1 receptor blockade should improve sepsis-related macrophage activation as it does in the previously mentioned experimental model  and in children with sJIA-related MAS [15, 16, 17]. In their secondary analysis of an adult with severe sepsis IL-1 receptor blockade trial, Shakoory and colleagues compared patients with combined hepatobiliary dysfunction (HBD) and DIC (HBD + DIC) to those without this combination (non-HBD + DIC) . The investigators found the following: (1) 5.6% of severe sepsis patients had HBD + DIC; (2) patients with HBD + DIC had a higher incidence of shock (HBD + DIC = 95% versus non-HBD + DIC = 79%) and acute kidney injury (HBD + DIC = 61% versus non-HBD + DIC = 29%), but not acute respiratory distress syndrome (HBD + DIC = 21% versus non-HBD + DIC = 26%); and (3) IL-1 receptor blockade significantly increased 28-day survival in patients with HBS + DIC (IL-1 blockade = 65.4% versus placebo = 35%), while 28-day survival in non-HBD + DIC did not change (IL-1 blockade = 71% vs placebo = 71%). It can be, therefore, extrapolated that sHLH diagnosed by the presence of HBD and DIC constitutes a small but significant subset among severe sepsis patients; death attributable to superimposed sHLH can be remedied in some patients with adjunctive therapy with IL-1 receptor blockade.
Simplified clinical criteria used to diagnose rheumatologic or drug-related secondary hemophagocytic lymphohistiocytosis (HLH)/macrophage activation syndrome (MAS) and hyperferritinemic sepsis-induced multiple-organ dysfunction syndrome (MODS)
Familial HLH: five of eight criteria
Hyperferritinemic sepsis-induced MODS
•Ferritin > 500 ng/ml
•Natural killer cell cytolytic activity < 10%
•Elevated soluble CD25
•Ferritin > 684 ng/mL
•With any two of the following:
Platelet count < 181,000
Alanine transferase > 48 IU/L
Triglyceride > 156 mg/dL
Fibrinogen < 360 mg/dL
•Disseminated intravascular coagulation
•+/- ferritin > 500 ng/mL
Clinical Vignette 3: Hyperferritinemic Sepsis-Induced MODS
A 3-year-old patient is admitted to the ICU with a presumed diagnosis of sepsis. Antibiotics and source control are implemented. Shock, acute respiratory distress syndrome, and acute kidney injury subsequently develop, along with hepatobiliary dysfunction and disseminated intravascular coagulation. Serum ferritin is 1120 ng/dL. Diagnosis of hyperferritinemic sepsis-induced multiple-organ dysfunction syndrome (MODS) is made.
An exhaustive search for the diagnosis is also performed, including bacterial sources (mycoplasma, rickettsia, legionella, chlamydia, brucella, and borrelia), fungi and parasites (histoplasmosis, babesia, leishmaniasis, pneumocystis, aspergillus, toxoplasmosis, cryptococcus, and candida), and viruses (EBV, CMV, HSV, HIV, HHV8, HHV6, parvovirus, adenovirus, and influenza) so that appropriate antimicrobial therapy can be ensured.
Plasma exchange, IVIG, and anakinra are considered for treatment. (methylprednisolone can also be considered if the patient does not have a contraindicated infection, e.g., HSV.) In patients with hyperferritinemic sepsis-induced MODS, improvement occurs when NK cell and T lymphocyte counts, which are reduced, recover.
ICU Management of Secondary HLH for the Pediatric Intensivist
Patients arriving to the pediatric ICU with a diagnosis of familial HLH will most commonly be receiving etoposide and dexamethasone, while patients arriving with a diagnosis of MAS will be receiving methylprednisolone with biologics such as anakinra (IL-1 receptor antagonist) or tocilizumab (anti-IL-6 antibody). For these children, supportive ICU care is most often necessitated by cardiopulmonary instability or organ failure often associated with HLH. The approach to organ support therapies for these children would be no different for any other child admitted to a pediatric ICU with sepsis, ARDS, and MODS.
More commonly, patients will present to an intensivist without a preceding diagnosis of HLH or MAS but rather with a diagnosis of sepsis-induced MODS. Indeed, most of these patients have sepsis-induced MODS. In addition to standard management for sepsis and multiple-organ failure, intensivists should have a high index of suspicion for sHLH. Though some centers have developed multidisciplinary hyperferritinemia interest groups to help the intensivist with this differential diagnosis, most depend on the intensivist to recognize signs and symptoms of this sHLH (Fig. 14.1) and call for specialty consultation. Of note, history of consanguineous parenting or family history of children dying from fever, especially in a patient presenting less than 2 years of age with significant CNS pathology, should raise suspicion for familial HLH. For patients with subacute arthralgia or arthritis, rheumatologic disease with MAS should be considered, although sHLH in patients with septic arthritis will be more likely to be hyperferritinemic sepsis-induced MODS.
Summary and Conclusion
The prompt recognition and appropriate treatment of familial and secondary HLH present important challenges to clinical intensivist. Differentiating between these clinical scenarios is the first step toward outlining the most appropriate clinical strategies, which is only possible through increasing our understanding of underlying immunopathogenesis that drives the clinical presentation. As our knowledge and understanding of the familial and secondary HLH evolve, evidence-based recommendations will better assist the clinical intensivists in their endeavors. Most importantly, a collaborative approach with hematology and rheumatology colleagues is likely the best means of securing a favorable outcome in patients who develop this challenging disease process.
Conflict of Interest Statement
On behalf of the authors, there are no conflicts of interest.
- 1.Castillo L, Carcillo J. Secondary hemophagocytic lymphohistiocytosis and severe sepsis/ systemic inflammatory response syndrome/multiorgan dysfunction syndrome/macrophage activation syndrome share common intermediate phenotypes on a spectrum of inflammation. Pediatr Crit Care Med. 2009;10:387–92.CrossRefGoogle Scholar
- 15.Demirkol D, Yildizdas D, Bayrakci B, Karapinar B, Kendirli T, Koroglu TF, et al. Hyperferritinemia in the critically ill child with secondary hemophagocytic lymphohistiocytosis/sepsis/multiple organ dysfunction syndrome/macrophage activation syndrome: what is the treatment? Crit Care. 2012;12(16):R52.CrossRefGoogle Scholar
- 16.Ravelli A, Minoia F, Davì S, Horne A, Bovis F, Pistorio A, et al. Classification criteria for macrophage activation syndrome complicating systemic juvenile idiopathic arthritis: a European league against rheumatism/American college of rheumatology/Paediatric rheumatology international trials organisation collaborative initiative. Arthritis Rheumatol. 2016;68:566–76.CrossRefGoogle Scholar
- 17.Shakoory B, Carcillo JA, Chatham WW, Amdur RL, Zhao H, Dinarello CA, et al. Interleukin-1 receptor blockade is associated with reduced mortality in Sepsis patients with features of macrophage activation syndrome: reanalysis of a prior phase III trial. Crit Care Med. 2016;44:275–81.CrossRefGoogle Scholar
- 18.Rajasekaran S, Kruse K, Kovey K, Davis AT, Hassan NE, Ndika AN, et al. Therapeutic role of anakinra, an interleukin-1 receptor antagonist, in the management of secondary hemophagocytic lymphohistiocytosis/sepsis/multiple organ dysfunction/macrophage activating syndrome in critically ill children. Pediatr Crit Care Med. 2014;15:401–8.CrossRefGoogle Scholar