Introduction

Macrophage activation syndrome (MAS) is a devastating complication of childhood rheumatic diseases. The earliest report of this syndrome might be a case of systemic-onset juvenile rheumatic arthritis (sJRA) with fatal hepatic failure in 1976 [1]. The term macrophage activation syndrome was introduced by Hadchouel et al. in 1985 to describe excessive activation of well-differentiated macrophages in bone marrow resulting in hemophagocytosis [2]. Hemophagocytic lymphohistiocytosis (HLH) syndrome is classified as genetic HLH (primary) or acquired (secondary) HLH. Genetic HLH includes familial HLH and immunodeficiency-associated HLH. Secondary HLH includes infection-associated hemophagocytic syndrome (IAHS), MAS, malignancy-associated HLH, and inborn errors of metabolism. MAS, or autoimmune-associated hemophagocytosis (AAHS), is usually associated with rheumatic diseases [3]. Some authors suggest that these terms are interchangeable, while others emphasize their heterogeneity and call for revised terminology based more precisely on pathophysiology [4].

HLH including MAS presents with a fulminant picture of non-remitting high fever, pancytopenia, hepatosplenomegaly, lymphadenopathy, liver dysfunction, coagulopathy, and neurological symptoms or encephalopathy, which can be explained by the high amounts of pro-inflammatory cytokines such as IFN-γ, IL-1, IL-6, and TNF-α resulting in the clinical characteristics of a cytokine storm [5]. The hallmark of HLH is activated macrophage ingesting blood cells (hemophagocytic histiocytes) in the bone marrow, lymph nodes or liver. The pathogenesis of genetic HLH involves defects in cytotoxic CD8+ T lymphocytes or natural killer (NK) cells leading to excessive activation of macrophages. However, in the case of MAS associated with rheumatic diseases, no such defect in cytotoxic killing is present. It is unclear what causes MAS in rheumatic diseases. Diminished NK cell function, perforin gene polymorphism, UNC13D gene mutation, and MUNC-14 gene polymorphism have been hypothesized to contribute to the development of MAS [6–9]. MAS/HLH is often life threatening and can be fatal if there is delay in diagnosis and treatment. To date, only a few studies have described the clinical course of pediatric MAS. We retrospectively reviewed the clinical pictures and outcomes of pediatric patients with MAS with different underlying autoimmune diseases over the past 20 years at a tertiary referral center in Taiwan. Our report provides valuable information in the diagnosis and management of children with MAS.

Methods

We retrospectively reviewed the medical records of 102 patients diagnosed with MAS/HLH since 1990 at the Department of Pediatrics, National Taiwan University Hospital, a tertiary referral center. Patients who fulfilled at least five of the following eight revised diagnostic criteria of HLH-2004 were included in our study: fever; splenomegaly; cytopenia ≥ 2 cell lines (platelet < 100 k/μL, or neutrophils < 1,000/μL, Hb < 9 g/dl); hypertriglyceridemia and/or hypofibrinogenemia; ferritin ≥ 500 μg/l; soluble CD25 ≥ 2,400 U/ml; decreased or absent NK cell activity; and hemophagocytosis in bone marrow, cerebrospinal fluid (CSF) or lymph nodes [10]. Patients who did not fulfill the diagnostic criteria of HLH-2004 but fulfilled the preliminary diagnostic criteria for MAS associated with sJRA were also enrolled [11]. Informed consent was obtained.

We identified eight MAS patients with underlying pediatric rheumatic diseases, 70 patients with IAHS, 17 patients with malignancy-associated hemophagocytosis, 3 patients with drug-induced hemophagocytosis (2 with phenytoin and 1 with lamotrigine), and 4 patients with hemophagocytosis related to congenital or familial immunodeficiency (1 Chediak–Higashi disease; 1 chronic granulomatous disease; 2 with suspected familial HLH). One patient who was originally diagnosed with sJRA with MAS was excluded from our study because he developed acute lymphocytic leukemia 5 months later.

The diagnosis of pediatric systemic lupus erythematous (SLE) was based on the American College of Rheumatology (ACR) 1997 revised criteria and age younger than 18 years at diagnosis [12]. For measurements of SLE disease activity, Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2K) scores were calculated upon diagnosis [13]. sJRA was diagnosed and classified according to the ACR 1977 revised criteria [14]. Active sJRA disease was defined as at least one joint with active arthritis, or fever ≥38.5°C at least 4 days a week without definable infection or other identifiable source other than JRA [15]. Others were defined as inactive. The diagnosis of mixed connective tissue disease (MCTD) was determined by the criteria proposed by Kasukawa [16]. We used the disease activity criteria of MCTD proposed by Lage et al. to define the disease status as active or inactive [17]. Anti-phospholipid syndrome (APS) was defined based on the revised classification criteria for APS [18].

Disseminated intravascular coagulation (DIC) was diagnosed according to the algorithm proposed by the International Society on Thrombosis and Haemostasis [19]. The diagnosis of Epstein–Barr virus (EBV) infection was based on serology with a fourfold increase in EBV viral capsid antigen (VCA) IgG titer, positive VCA IgM, or EBV viral load detection by PCR (polymerase chain reaction) positive in serum or >1,000 copies/mcg DNA in peripheral blood mononuclear cells. Adenovirus infection was determined by isolation of the virus in throat or rectal swabs and the presence of relevant clinical symptoms/signs.

Results

Demographic data, clinical features, and possible triggers of MAS

Eight patients (4 sJRA, 2 SLE, 1 MCTD, 1 APS) with 13 episodes of MAS were studied. Six patients fulfilled at least five of the eight revised diagnostic criteria of HLH-2004. Two patients (Patients 7 and 8) fulfilled the preliminary diagnostic criteria for MAS associated with sJRA. The mean age of the patients was 12.6 years old (7–15 years old), and the male to female ratio was 3:5. The mean duration of underlying autoimmune diseases was 1 year (range, 0 months to 3 years) at the diagnosis of the first episode of MAS. The demographic data and possible triggering factors are summarized in Table 1. MAS presented as the first manifestation of rheumatic disease in four patients (50 %; Patient 4 with APS, Patients 5, 6, 8 with sJRA). The clinical manifestations of MAS included fever (100 %), hepatosplenomegaly (77 %), lymphadenopathy (38 %), skin rash (62 %), neurological involvement (31 %), and overt DIC (23 %; Table 2). Neurological manifestations included seizures (Patients 1 and 5) and consciousness change (Patients 3 and 8). The laboratory features included leukopenia (7/13, 54 %), anemia (6/13, 46 %), thrombocytopenia (10/13, 77 %), jaundice (3/11, 27 %), hypofibrinogenemia (2/5, 40 %), decreased erythrocyte sedimentation rate (ESR; 8/12, 67 %), and elevated liver enzymes (AST/ALT; 10/13, 77 %), lactate dehydrogenase (LDH; 4/4, 100 %), ferritin (7/8, 88 %, median level 18,336 ng/ml), triglycerides (10/11, 91 %, median level 178 mg/dl), C-reactive protein (CRP; 11/13, 85 %), plasma D-dimer (2/4, 50 %), and hemophagocytosis in bone marrow (10/12, 83 %; Table 2). Active infection triggered MAS in two episodes. The Epstein–Barr virus (EBV) and adenovirus infection triggered MAS in Patients 7 and 8, respectively, both of who had sJRA.

Table 1 Demographic data, underlying diseases, and possible triggers of the studied patients
Full size table
Table 2 Clinical manifestations and laboratory data of the studied patients
Full size table

Clinical courses

Two SLE patients developed MAS during tapering of corticosteroids. Patient 1 was diagnosed with SLE with lupus nephritis at the age of 13. She had prolonged fever, hepatosplenomegaly, progressive cytopenia and seizures during tapering of her corticosteroid dosage 1 month after a disease flare with a neuropsychiatric lupus presentation (SLEDAI score of 19 points). MAS was controlled by IVIG 400 mg/kg/day for 5 days along with high-dose corticosteroid treatment for 3 weeks.

Patient 2 was diagnosed with SLE with lupus nephritis when she was 9 years old. After choosing to cease medication for 3 months, she experienced a disease flare with cytopenia and renal failure (SLEDAI score of 12 points). After three courses of methylprednisolone pulse therapy and IVIG, MAS was diagnosed based on a non-remitting fever, hepatosplenomegaly, pancytopenia, elevated aspartate aminotransferase (AST) and triglyceride levels, and hemophagocytosis in her bone marrow. She died of persistent cytopenia, massive GI bleeding, and intracranial hemorrhage.

Patient 3 was diagnosed with MCTD at the age of 12. MAS presented as fever, consciousness disturbance, seizures, thrombocytopenia, elevated ferritin (119,979 ng/ml) and AST (1,925 U/L), decreased complements, and hemophagocytosis in his bone marrow. MAS and multiple organ failure were successfully controlled by IVIG (1 g/kg/day) for 2 days, methylprednisolone pulse therapy for 3 days, and intensive care for 1 month.

Patient 4 experienced an episode of MAS with prolonged fever, abdominal pain, splenomegaly, thrombocytopenia and hemophagocytosis in the bone marrow. Treatment with IVIG was effective. The diagnosis of APS was subsequently made by the manifestations of venous thrombosis, highly elevated IgG-anti-cardiolipin antibody, edema and serositis.

Among the sJRA patients, three patients (Patient 5, 6, 7) had recurrent episodes of MAS, and five out of eight episodes occurred during symptom and medication free periods. Skin rashes developed in all of the patients with sJRA. Two patients (Patients 5, 6) had serositis with pericardial effusion and pleural effusion. Patient 5 had generalized tonic–clonic seizure. Patient 8 presented with consciousness disturbance, right side weakness, blurred vision due to central retinal vein occlusion, and prolonged fever for 5 weeks.

Treatment of MAS

Treatments and outcomes are summarized in Table 3. IVIG (1 g/kg/day for 2 days) was given to six patients as the first line of treatment. The response rate to IVIG was two out of six (33 %) patients, and it failed in four of six patients (Patient 2 with SLE, Patient 3 with MCTD, Patient 5 in episode 2, and Patient 8 with sJRA). Three patients who failed to respond to IVIG received methylprednisolone pulse therapy (30 mg/kg/day for 3 days; Patients 2, 3, 8) and the response rate to methylprednisolone pulse therapy was two out of three (66.7 %). All patients received at least full-dose corticosteroid (equivalent to prednisolone 1~2 mg/kg/day). Three patients with sJRA received a full-dose of corticosteroids only and responded well (Patient 5 in episode 1, Patient 6 in episode 1, Patient 7 in episodes 3 and 4). Etanercept therapy given to Patient 6 in episode 2 failed, and MAS was subsequently controlled by methylprednisolone pulse therapy. Patients 6 and 7 received cyclosporine A for 6 months. Neurological recovery in Patients 3 and 8, both of whom presented with consciousness disturbance, was achieved 9 and 3 days after methylprednisolone pulse therapy, respectively.

Table 3 Treatment responses and outcomes of the studied patients
Full size table

Complications and outcomes

Eight patients were admitted to the intensive care unit (ICU). Hemodynamic instability for a short period of time was seen in Patients 5 and 6. Complications during treatment course included: opportunistic infection with Pneumocystis jiroveci (Patients 1 and 2), multiple organ failure (Patients 3 and 4), and ICU myopathy (Patient 4). All patients had favorable outcomes except Patient 2. The overall mortality rate of MAS was 12.5 % (1/8), and no cases of mortality were seen after 2000.

Discussion

MAS is most commonly associated with sJRA [4], but associations with SLE, adult onset Still’s disease, Kawasaki disease, and juvenile dermatomyositis have been reported [20–22]. MAS associated with MCTD and primary APS, as in our pediatric series, has never been reported before, although a few such cases have been reported in adults [23, 24]. The clinical spectrum of MAS in our pediatric patients with different rheumatic diseases was similar to that reported in adults. The diagnosis of MAS in patients with rheumatic disease is always challenging. It is important for clinicians to differentiate MAS from disease activity flares of rheumatic diseases, especially cytopenia during active SLE, or infection because of the similarity in clinical manifestations and the mutual interplay between them. Moreover, MAS may be diagnosed in the absence of evidence of macrophage hemophagocytosis in the bone marrow, because hemophagocytosis is not always demonstrable at onset. However, a bone marrow aspirate should be performed to rule out malignancy or myelodysplastic syndromes.

Neurological symptoms in MAS have been described as irritability, disorientation, lethargy, headache, seizures, papilledema, or coma [25]. Four patients in our study presented with neurological manifestations. Two of them were suspected to have meningitis based on clinical symptoms and pleocytosis in CSF, but no pathogens were identified. A CSF examination, culture or viral PCR is important to exclude the possibility of central nervous system (CNS) infection. It is more difficult to differentiate neuropsychiatric lupus during active disease from neurological manifestations in MAS associated with SLE. Exclusion of other causes such as infection, metabolic disarrangement, or drugs is essential in the diagnosis of neuropsychiatric lupus [26].

The strong association of MAS with sJRA and adult onset Still’s disease is intriguing. MAS occurs in up to 7 % of sJRA patients. However, about 53 % of sJRA patients may develop occult MAS, by means of hemophagocytosis found in bone marrow without clinically obvious cytopenia, which was seen in three patients with sJRA in our series [27]. In patients with sJRA, evident cytopenia or hypofibrinogenemia often appears at a late stage of MAS, because they usually have marked leukocytosis, thrombocytosis, and elevation of inflammatory markers due to hyperinflammation [28]. Ravelli et al. proposed diagnostic guidelines for MAS in JRA patients in 2005 [11]. They found the strongest discriminators between MAS and sJRA flare were hemorrhages, CNS dysfunction, decreased platelet count, increased AST, leukopenia, and hypofibrinogenemia. New markers such as elevation of soluble IL-2 receptor, soluble CD163 or IL-18 levels in serum may provide clues for diagnosis [29, 30] .

The prevalence of MAS in SLE ranges from 0.9 % to 4.6 % [31]. However, MAS in juvenile SLE may be underdiagnosed. Hemophagocytosis in bone marrow has been found in 73.3 % of 28 SLE patients with cytopenia, and in 9.6 % of 73 SLE patients with liver dysfunction, especially those with high AST levels (500 U/mL) [32]. Preliminary diagnostic criteria for MAS in juvenile SLE were proposed in 2009, with a sensitivity of 92.1 % and a specificity of 90.9 % [32]. The most sensitive and specific laboratory features were hyperferrinemia, increased levels of lactate dehydrogenase, hypertriglyceridemia, and hypofibrinogenemia. Early suspicion of MAS in juvenile SLE patients is commonly based on the detection of subtle changes in clinical parameters (such as cytopenia), whereas clinical symptoms are often delayed or not specific.

Unlike other forms of HLH which are described in the HLH-2004 treatment guidelines that emphasize the use of etoposide and cyclosporine in addition to dexamethasone treatment, there is no standard protocol for treating MAS to date [10]. Corticosteroid monotherapy or with additional immunosuppressants is the mainstay of treatment for MAS. Etoposide is seldom used in MAS. The treatment failure rate of corticosteroid monotherapy in MAS ranges widely from 0 to 56 % [25, 31, 33, 34]. We used high dose IVIG as initial therapy, however only 33 % of the patients were responsive. Those who failed to respond to IVIG received corticosteroid therapy from a full dose to methylprednisolone pulse therapy or additional cyclosporine. The efficacy of IVIG in MAS has not been demonstrated by large-scale studies [33–35]. IVIG may act through the blockade of Fc receptors and other pathways to prevent a cytokine storm. It is safe with minor adverse events [36]. Effective treatment of MAS with cyclosporine and intravenous cyclophosphamide has been demonstrated in reported cases [31, 37]. Cyclosporine is preferred over cyclophosphamide with regards to the side effect of myelosuppression. Biological agents are of much interest in treating autoimmune diseases, however their effectiveness requires further verification. Interestingly, etanercept, a TNF blocker, may trigger MAS [38]. An IL-1 receptor antagonist, anakinra, has been reported to induce sustained remission of MAS together with corticosteroids [39].

The outcome of MAS is better than other forms of HLH. The mortality rate of MAS is about 20–38 % in adults and 8–20 % in children [25, 32, 40, 41]. Poor prognostic factors of all forms of HLH include age over 30 years, presence of DIC, increased ferritin and β2-microglobulin, anemia accompanied by thrombocytopenia and jaundice [42]. In an adult series of 30 MAS patients, the presence of concurrent infections and high CRP levels was related to high mortality [31]. Multisystem involvement was suggested to be a poor prognostic sign in an early series of nine children [40]. The prognosis of our patients was better than adult series of MAS. We also found that there was a greater possibility of MAS recurring in sJRA compared to other rheumatic diseases.

In conclusion, MAS can be fatal and complicate various pediatric autoimmune diseases. The diagnosis is challenging. It generally has a good response to corticosteroids and IVIG. Prompt recognition and timely treatment can result in good outcomes.