Article Text

Efficacy and safety of emapalumab in macrophage activation syndrome
  1. Fabrizio De Benedetti1,
  2. Alexei A Grom2,3,
  3. Paul A Brogan4,
  4. Claudia Bracaglia1,
  5. Manuela Pardeo1,
  6. Giulia Marucci1,
  7. Despina Eleftheriou4,
  8. Charalampia Papadopoulou4,
  9. Grant S Schulert2,3,
  10. Pierre Quartier5,6,
  11. Jordi Antón7,8,
  12. Christian Laveille9,
  13. Rikke Frederiksen10,
  14. Veronica Asnaghi10,
  15. Maria Ballabio10,
  16. Philippe Jacqmin11,
  17. Cristina de Min10
  1. 1 Division of Rheumatology, Ospedale Pediatrico Bambino Gesù, IRCCS, Rome, Italy
  2. 2 Division of Rheumatology, Cincinnati Children’s Hospital Medical Center, Cincinnati, OH, USA
  3. 3 Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, OH, USA
  4. 4 University College London Great Ormond Street Institute of Child Health, London, UK
  5. 5 Pediatric Immuno-Hematology and Rheumatology Unit, RAISE Rare Disease Reference Centre, Hopital Universitaire Necker-Enfants Malades, Assistance Publique-Hopitaux de Paris, Paris, France
  6. 6 Université Paris-Cité, Paris, France
  7. 7 Pediatric Rheumatology, Hospital Sant Joan de Deu, Barcelona, Spain
  8. 8 Faculty of Medicine, Universitat de Barcelona, Barcelona, Spain
  9. 9 Calvagone Sarl, Liergues, France
  10. 10 Swedish Orphan Biovitrum AG (Sobi), Basel, Switzerland
  11. 11 MnS Modelling and Simulation, Dinant, Belgium
  1. Correspondence to Dr Fabrizio De Benedetti, Division of Rheumatology, Ospedale Pediatrico Bambino Gesu, IRCCS, Rome 00165, Italy; fabrizio.debenedetti{at}


Objectives Macrophage activation syndrome (MAS) is a severe, life-threatening complication of systemic juvenile idiopathic arthritis (sJIA) and adult-onset Still’s disease (AOSD). The objective of this study was to confirm the adequacy of an emapalumab dosing regimen in relation to interferon-γ (IFNγ) activity by assessing efficacy and safety. The efficacy outcome was MAS remission by week 8, based on clinical and laboratory criteria.

Methods We studied emapalumab, a human anti-IFNγ antibody, administered with background glucocorticoids, in a prospective single-arm trial involving patients who had MAS secondary to sJIA or AOSD and had previously failed high-dose glucocorticoids, with or without anakinra and/or ciclosporin. The study foresaw 4-week treatment that could be shortened or prolonged based on investigator’s assessment of response. Patients entered a long-term (12 months) follow-up study.

Results Fourteen patients received emapalumab. All patients completed the trial, entered the long-term follow-up and were alive at the end of follow-up. The investigated dosing regimen, based on an initial loading dose followed by maintenance doses, was appropriate, as shown by rapid neutralisation of IFNγ activity, demonstrated by a prompt decrease in serum C-X-C motif chemokine ligand 9 (CXCL9) levels. By week 8, MAS remission was achieved in 13 of the 14 patients at a median time of 25 days. Viral infections and positive viral tests were observed.

Conclusions Neutralisation of IFNγ with emapalumab was efficacious in inducing remission of MAS secondary to sJIA or AOSD in patients who had failed high-dose glucocorticoids. Screening for viral infections should be performed, particularly for cytomegalovirus.

Trial registration number NCT02069899 and NCT03311854.

  • arthritis, juvenile
  • biological therapy
  • inflammation
  • Still's disease, adult-onset
  • therapeutics

Data availability statement

Data are available on reasonable request. The data sharing policy of Sobi is available at the following site: As noted on that site, requests for access to the study data can be submitted to

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


  • Macrophage activation syndrome (MAS) is a severe and potentially life-threatening complication of systemic juvenile idiopathic arthritis and adult-onset Still’s disease.

  • There are no therapeutic options that have been prospectively investigated in MAS.

  • Data in animal models and ex vivo data from humans with MAS led to the hypothesis that interferon-γ (IFNγ) has a pathogenic role in MAS.


  • This open-label multicentre trial using emapalumab, an anti-IFNγ antibody, in patients who have failed to respond to high-dose glucocorticoids, demonstrates that IFNγ has a pathogenic role in MAS and that its neutralisation markedly improved MAS.


  • The results of this study show that neutralisation of IFNγ with emapalumab is a therapeutic option for patients with severe MAS who have failed standard of care with high-dose glucocorticoids.


Macrophage activation syndrome (MAS) is a form of secondary haemophagocytic lymphohistiocytosis (HLH) occurring as a life-threatening complication of rheumatic diseases.1 It is most frequent in systemic juvenile idiopathic arthritis (sJIA) and adult-onset Still’s disease (AOSD), affecting about 10%–20% of patients.1–4 sJIA and AOSD are considered the same disease named differently depending only on age at onset below or above 16 years, respectively.2 3 5 The incidence and features of MAS are similar in the two age groups.2 5 We use the term sJIA/AOSD to collectively identify these patients.

Similar to other forms of HLH, MAS is caused by excessive activation and expansion of T lymphocytes and macrophages resulting in hyperinflammation.1 6 7 MAS is characterised by fever, hepatosplenomegaly, cytopenias, liver dysfunction, coagulation abnormalities and hyperferritinaemia, and may progress to multiple organ failure, with mortality rates of 10%–20%.4

MAS is treated with high-dose glucocorticoids with satisfactory response in two-thirds of the patients.4 8 In patients unresponsive to glucocorticoids, ciclosporin is usually added. Treatment with cyclophosphamide, etoposide, intravenous immunoglobulin, etanercept, anakinra, tocilizumab, JAK inhibitors and plasmapheresis, have been described in case reports or small series.4 8–12 None of these regimens have been prospectively investigated.

Overproduction of interferon-γ (IFNγ) is present and pathogenic in animal models of MAS.13–16 In sJIA/AOSD, high IFNγ activity demonstrated by high serum levels of C-X-C motif chemokine ligand 9 (CXCL9), a chemokine selectively induced by IFNγ,17 is associated with MAS onset and severity.18–22

We studied emapalumab, a fully human anti-IFNγ antibody, in patients with MAS secondary to sJIA/AOSD who failed high-dose glucocorticoids.


This phase II, open-label, single-arm trial (NI-0501-06; Identifier NCT03311854), conducted at five sites in Italy, France, Spain, the UK and the USA, comprised screening, a treatment period of 28 days and a short-term 4-week follow-up. Data were collected up to 12 months in a long-term follow-up study (NI-0501-05; Identifier NCT02069899) (online supplemental figure S1 in the online supplemental appendix). Results from both studies are reported.

Supplemental material

Figure 1

Pharmacokinetic (emapalumab* (A)) and pharmacodynamic (CXCL9† (B), total IFNγ‡ (C) and sIL-2R§ (D)) parameters in patients with MAS treated with emapalumab. *Serum emapalumab concentrations in patients treated with emapalumab (initial dose of 6 mg/kg, followed by intended maintenance dosing of 3 mg/kg every 3 days until day 15, and twice weekly until day 28). †Serum concentrations of CXCL9, a biomarker of IFNγ activity, being, at baseline, 4–362 times the 95th percentile of healthy volunteers (271 pg/mL),24 indicated by the horizontal dotted line. ‡Total IFNγ (free and emapalumab-bound) that reflects IFNγ production, ranging, at day 3, from normal to 12 times the 95th percentile of healthy volunteers receiving emapalumab (2084 pg/mL) indicated by the horizontal dotted line. §Serum sIL-2R, a biomarker of T cell activation, being, at baseline, 1.6–20 times the 95th percentile of healthy volunteers (1071 pg/mL, unpublished data) indicated by the horizontal dotted line. CXCL9, C-X-C motif chemokine ligand 9; IFNγ, interferon-γ; MAS, macrophage activation syndrome; sIL-2R, soluble interleukin-2 receptor.

FDB, AG, MB and CdM designed the study. The sponsor was also responsible for data collection, management and analysis. All authors vouch for the accuracy and completeness of the data and analyses, and for the fidelity of the study to the protocol.

Patient involvement

The study was presented and discussed at the 2016 and 2017 meetings of the sJIA Foundation in Washington, DC, USA. Information on the study for patients and families was published on the sJIA Foundation website. Dissemination of the data gathered from the study is planned to be discussed with the sJIA Foundation.


Patients had sJIA based on International League Against Rheumatism criteria,23 or AOSD based on Yamaguchi criteria.24 For patients presenting with MAS at sJIA onset, presumption of sJIA was based on Childhood Arthritis and Rheumatology Research Alliance criteria.25 Patients had MAS, according to American College of Rheumatology/EULAR criteria,26 and an inadequate response to high-dose intravenous glucocorticoids according to the treating physicians. High-dose glucocorticoids were defined as ≥2 mg/kg/day of prednisone equivalent in two divided doses, or at least 60 mg/day in patients weighing 30 kg or more, including but not limited to pulses up to 30 mg/kg/day for at least 3 consecutive days. Patients with active infections potentially favoured by IFNγ neutralisation (typical and atypical mycobacteria, Histoplasma capsulatum, Shigella, Salmonella, Campylobacter and Leishmania) were excluded (see online supplemental appendix 1). Eligibility criteria are listed in online supplemental table S1.


Emapalumab infusions were administered on a background of glucocorticoids. The initial dose of emapalumab was 6 mg/kg on day 0, followed by 3 mg/kg every 3 days until day 15 and twice weekly until day 28. Treatment with emapalumab could be stopped on investigator’s assessment of remission, but not before three emapalumab doses have been administered. Frequency between infusions could be shortened, dose could be increased or treatment prolonged on investigator’s assessment of unsatisfactory response. Ciclosporin could be continued, if started at least 3 days before initiating emapalumab. Interleukin (IL)-1 and IL-6 inhibitors were not allowed. An amendment allowed continuation of anakinra, if started at least 3 days before initiating emapalumab, and its introduction during the study, at a maximum dose of 4 mg/kg/day to treat the underlying sJIA/AOSD. Tocilizumab and canakinumab were not allowed during the trial. Prophylaxis against herpes zoster was administered as per local standards. Glucocorticoid tapering could be initiated as soon as the patients’ conditions allowed based on investigator’s assessment.


The objective was to confirm the adequacy of the emapalumab dosing regimen in relation to IFNγ activity by assessing efficacy and safety. Serum levels of emapalumab, total (free and emapalumab-bound) IFNγ, CXCL9 and soluble IL-2 receptor were measured (online supplemental appendices 2 and 3).

The efficacy outcome was MAS remission by week 8, defined as resolution of clinical signs and symptoms (online supplemental table S2) according to the physician global assessment (visual analogue scale ≤1/10) and white blood cell and platelet count above lower limit of normal, lactate dehydrogenase (LDH), alanine aminotransferase and aspartate aminotransferase below 1.5 times upper limit of normal (ULN), fibrinogen >100 mg/dL and ferritin levels27 decreased by at least 80% or below 2000 ng/mL, whichever was lower. Other efficacy evaluations included glucocorticoid dose (expressed as mg/kg/day of prednisone-equivalent) and survival. Adverse events (AEs) were assessed. During the long-term follow-up, evaluations included MAS episodes, AEs, pharmacokinetics and pharmacodynamics.

Statistical analysis

The analysis population included all patients. Categorical variables are presented with the number and percentage within each category. Continuous variables are reported as median (range). Laboratory parameters of MAS were measured locally.


Study population

Fourteen patients received emapalumab. All completed the study and entered long-term follow-up. Thirteen patients had sJIA onset before 16 years, and one AOSD with onset at 16 years and 9 months. Four patients had presumption of sJIA, which was later confirmed. Six patients had a total of 19 previous MAS episodes in their history, treated with high-dose glucocorticoids, with the addition of anakinra and/or ciclosporin in approximately half.

During the week preceding emapalumab, all patients were receiving high-dose intravenous glucocorticoids, as per protocol. In addition, eight were receiving ciclosporin and seven anakinra (table 1). Of the patients treated with anakinra, three were receiving standard doses for sJIA (≤4.0 mg/kg/day), and four high doses ranging from 7.5 to 15 mg/kg/day. Patients had severe MAS at baseline, shown by high physician assessment of MAS activity and marked abnormalities of laboratory parameters (table 2). Notably, worsening or no improvement of laboratory parameters was observed between screening and baseline, despite ongoing treatment.

Table 1

Baseline demographics and clinical characteristics of the patients

Table 2

MAS activity, as defined by physician’s assessment, and laboratory features of MAS before, during and after emapalumab treatment

Of the 14 patients, 6 received emapalumab up to day 28. In seven patients who had MAS remission per investigator’s assessment, it was possible to discontinue emapalumab earlier. One patient continued emapalumab up to day 39. The median treatment duration was 27 days (range, 7–39) with the number of infusions ranging from 3 to 17 per patient.

Pharmacokinetics and pharmacodynamics

All patients received an initial dose of emapalumab of 6 mg/kg followed by doses of 3 mg/kg. The initial dose allowed to rapidly reach serum concentrations of emapalumab close to the steady-state concentrations obtained with the doses of 3 mg/kg (figure 1A). The dosing interval was transiently shortened from 3 to 2 days in three patients based on clinical and laboratory parameters suggestive of incomplete response. In two of these patients the dosing interval was shortened for several infusions. In these two patients, measurement of emapalumab concentrations showed rapid drug clearance consistent with target-mediated drug disposition associated with high levels of total IFNγ (online supplemental figure S2).28

Figure 2

Time to MAS remission*. *Time to MAS remission defined as resolution of clinical signs and symptoms according to the investigator (visual analogue scale ≤1/10) (online supplemental table S2) and resolution of the abnormalities of MAS laboratory parameters: white blood cell and platelet count above lower limit of normal, LDH below 1.5 times ULN, alanine aminotransferase/aspartate aminotransferase below 1.5 times ULN, fibrinogen >100 mg/dL and ferritin levels decreased by at least 80% or below 2000 ng/mL, whichever was lower. The continuous line represents the proportion of patients with MAS remission, the dotted line represents the 95% CI. The cross indicates the censored patient. LDH, lactate dehydrogenase; MAS, macrophage activation syndrome; ULN, upper limit of normal.

At baseline, CXCL9 concentrations were markedly elevated in all patients, indicating high IFNγ activity (figure 1B). Total IFNγ concentrations were variable across patients (figure 1C). Total IFNγ concentrations at day 3 (at equilibrium between free and emapalumab-bound IFNγ) reflect IFNγ production at baseline.27 28 There was no correlation between total IFNγ at day 3 and CXCL9 levels at baseline (not shown). Soluble IL-2 receptor levels at baseline were markedly elevated in all patients (figure 1D).

Emapalumab administration led to a rapid decrease in serum CXCL9 levels, indicating neutralisation of IFNγ activity and therefore the appropriateness of the selected dosing regimen. Patients with elevated total IFNγ at day 3 also showed a decrease in total IFNγ concentrations over time indicating decreasing IFNγ production over time. Soluble IL-2 receptor concentrations also decreased markedly (figure 1D), indicating decreasing T cell activation. During long-term follow-up, in the absence of high IFNγ production, as shown by low levels of total IFNγ, emapalumab showed a slow linear terminal elimination phase with a half-life of 24 days, like that in healthy subjects (online supplemental figure S3).28 Total IFNγ, CXCL9 and soluble IL-2 receptor levels were close to, or within, the normal range (online supplemental table S3).

Figure 3

Changes in ferritin* (A), aspartate aminotransferase† (B), platelets (C) and LDH‡ (D) levels over time. *Panel A insert shows in detail changes from baseline to week 4 for patients with baseline levels of ferritin <50 000 ng/mL. †Panel B insert shows in detail changes from baseline to week 4 for patients with baseline levels of aspartate aminotransferase below 300 U/L. ‡Panel D insert shows in detail changes from baseline to week 4 for patients with baseline levels of LDH below 2500 U/L. LDH, lactate dehydrogenase.


During emapalumab administration, physician global assessment of MAS activity and MAS laboratory parameters rapidly improved in all patients while glucocorticoids were tapered.

By week 8, 13 patients (93%) achieved MAS remission at a median time of 25 days after emapalumab initiation, the earliest at day 9 (figure 2). At week 8, 2 of the 13 patients who had previously achieved MAS remission did not meet the criteria of MAS remission because of a single laboratory abnormality (one with LDH 1.7-fold above the ULN; one with white blood cells at 4.8×109/L with lower limit of normal at 5.5×109/L). One patient who stopped emapalumab after three doses on investigator’s assessment of remission never met the criteria of MAS remission only because of LDH levels 1.5-fold above the ULN.

All laboratory parameters of MAS rapidly improved on initiation of emapalumab with evident improvement at week 1 and with all parameters being normal in the majority of patients by week 4 (figure 3, table 2 and online supplemental figure S4).

During long-term follow-up, 13 patients did not have MAS episodes. One patient had a MAS episode 11 months after stopping emapalumab, when emapalumab was undetectable in serum. At the end of the long-term follow-up, 10 patients met the MAS remission criteria. Four did not meet these criteria because of either mild abnormality in one laboratory value (n=1), MAS activity at 1.5 cm (n=1), missing data (n=1) or absence of data as the patient missed the 12-month visit (online supplemental table S4).

Tapering of glucocorticoids occurred rapidly after emapalumab initiation. During the week preceding emapalumab, the median average daily dose was 15.7 mg/kg/day prednisone-equivalent, 2.3 mg/kg during week 2 and 0.56 mg/kg during week 8 (figure 4). At the end of the long-term follow-up, five patients were not receiving glucocorticoids and six were receiving <0.3 mg/kg/day prednisone-equivalent. Two patients were receiving between 1 and 2 mg/kg: the above-mentioned patient with a MAS episode and one patient with lung disease associated with sJIA. Data on glucocorticoid dose for one patient at the last follow-up visit are not available, as the patient missed the visit.

Figure 4

Changes in glucocorticoid dose* over time†. *The dose of glucocorticoids is shown as average daily dose in each week of the study, starting from the week preceding emapalumab treatment (week −1). The dose is shown in mg/kg of prednisone equivalent. †Weeks are defined starting form week −1 (from day −6 to day 0), followed by week 1 (form day 1 to day 7); week 2 (from day 8 to day 14); week 3 (from day 15 to day 21); up to week 8 (from day 50 to day 56).

Eight patients were receiving ciclosporin at baseline. Ciclosporin was discontinued in two patients early after emapalumab initiation (day 4 and 10) and in four additional patients during long-term follow-up.

During the trial, anakinra was continued in four patients at a dose of ≤4 mg/kg, the standard dose for sJIA/AOSD treatment and in one patient at a dose of 7.5 mg/kg. These patients did not present with flares of the underlying sJIA/AOSD. During the trial, while MAS was improving, six flares of sJIA/AOSD were observed: three in the three patients who discontinued anakinra before emapalumab initiation and three in three patients who had not previously received and were not receiving anakinra. During long-term follow-up, four sJIA/AOSD flares occurred in four patients. sJIA/AOSD flares and the background treatment at time of the flares are described in online supplemental table S5.


No deaths were reported during the trial and the long-term follow-up. During the trial, after initiation of emapalumab, 88 AEs were reported in 13 patients (table 3 and online supplemental table S6). All events were mild or moderate in intensity except two (one cardiopulmonary failure, and one neutropenia; neither related to emapalumab). The most frequently reported AEs were infections and positive tests for infectious agents in the absence of clinical symptoms. All infectious events were of viral origin. No bacterial or opportunistic infections were reported. Six viral events in three patients (two infections and four positive tests) were reported as related to emapalumab. One cytomegalovirus (CMV) reactivation was reported as serious. In total, there were five CMV events (three reactivations, one infection and one positive test with no symptoms). All viral events resolved spontaneously or with standard treatment. Two infusion-related reactions (pruritic rash), not reported as severe, occurred during a total of 128 infusions. The rate of AEs and of infectious events was not increased during concomitant treatment with anakinra and emapalumab compared with treatment with emapalumab alone (table 4).

Table 3

AEs during the treatment with emapalumab and during the long-term follow-up

Table 4

Rate of AEs* and infectious AEs during the trial while exposed to emapalumab alone or emapalumab and anakinra

During long-term follow-up, 37 AEs were reported in 12 patients. Three were serious: one flare of sJIA (see online supplemental table S5), one oedema of the ankle and one MAS episode (that, as mentioned above, occurred 11 months after stopping emapalumab, when emapalumab was undetectable). Ten infectious events occurred in seven patients; four when emapalumab levels were measurable, and six after emapalumab levels became undetectable. All reported infections were viral, except for one cestode infection.


We show that neutralisation of IFNγ with emapalumab was efficacious in inducing MAS remission in patients with MAS secondary to sJIA/AOSD who failed standard of care including high-dose glucocorticoids with or without anakinra and/or ciclosporin.

The dosing regimen of emapalumab was chosen based on the data gathered in primary HLH patients on (a) the production rate of IFNγ (through the assessment of total IFNγ), (b) the rapid clearance of emapalumab consequent to target-mediated drug disposition in the presence of high IFNγ production and (c) the concentration of emapalumab required to neutralise high IFNγ levels. Through modelling and simulation, on the basis of the high levels of CXCL9 present in patients with MAS, it was possible to predict the dose of emapalumab required to achieve IFNγ neutralisation. A dosing regimen with an initial dose of 6 mg/kg, followed by 3 mg/kg every 3 days maintenance doses, was selected to achieve rapid efficacy as MAS usually has acute onset and may worsen rapidly, becoming life-threatening. In the majority of patients, the chosen regimen rapidly achieved IFNγ neutralisation, shown by prompt decrease in serum CXCL9 levels and prompt clinical and laboratory response. Shortening of the dosing interval might be considered in patients with unsatisfactory response. Indeed, in two patients with initial unsatisfactory response, shortening of the dosing interval led to achievement of response. In these patients, we found high IFNγ production and low emapalumab concentrations due to rapid drug clearance consistent with target-mediated drug disposition.

The patients recruited in this study had markedly elevated CXCL9 levels, reflecting high IFNγ activity. This finding is consistent with previous observations in patients with MAS.16 18 In contrast with primary HLH patients treated with emapalumab,27 we did not find a correlation of CXCL9 levels at baseline with total IFNγ levels at day 3, that, as mentioned, reflect IFNγ production. Therefore, in MAS, the pathogenic role of IFNγ may be due to increased IFNγ production and increased sensitivity to IFNγ. Indeed, monocytes from patients with MAS have increased responsiveness to IFNγ, possibly through increased expression of tripartite motif containing protein 8 (TRIM-8), which potentiates response to IFNγ.29 30

In this trial, the first prospectively investigating a treatment for MAS, we used a novel clinically meaningful efficacy measure, namely MAS remission, that combines resolution of clinical signs and symptoms and of the abnormalities in MAS laboratory parameters. Incidentally, the ferritin threshold (decrease by at least 80% or be below 2000 ng/mL, whichever was lower), previously used in the primary HLH trial with emapalumab,27 is justified by the often-needed multiple transfusions and the frequent fluctuations of ferritin >1000 ng/mL in patients with HLH/MAS with good response to treatment. Notably, the highest value for ferritin at week 8 in our trial was 561 ng/mL.

After emapalumab initiation, MAS improved rapidly with median time to MAS remission of 25 days. This improvement occurred while background glucocorticoids were rapidly tapered: during the second week of emapalumab treatment, the median glucocorticoid dose was already 85% lower compared with that administered during the week preceding initiation of emapalumab. In general, patients seemed to benefit from emapalumab even when the MAS remission criteria were partially achieved. Indeed, in the few patients who did not reach MAS remission, the physician assessment of disease activity was ≤1 and MAS remission was not achieved because of minor abnormalities in only one of the laboratory parameters.

Elevated production of IFNγ is a feature of animal models of primary HLH and of secondary HLH, including infection-associated HLH and MAS.13 In these models, when the effect of therapeutic IFNγ neutralisation was tested, benefit was always observed with prevention of death and/or disease improvement. The results of this trial, together with the efficacy of emapalumab demonstrated in primary HLH,27 and the anecdotal cases of patients with different forms of secondary HLH successfully treated with emapalumab,31–37 suggest that, in humans, IFNγ is an important driver of MAS/HLH, independently of the trigger or the underlying predisposing condition.

Notably, some inflammatory flares of the underlying sJIA/AOSD were observed, suggesting that, in the context of sJIA/AOSD, increased IFNγ activity is a feature of the hyperinflammation of MAS, but not of the autoinflammation that maintains sJIA. From a clinical perspective, we observed six flares of the underlying sJIA/AOSD in six out of the nine patients who were not receiving anakinra for the treatment of sJIA while receiving emapalumab. In contrast, no sJIA flares were observed in the five patients who continued anakinra after emapalumab initiation. Importantly, during the trial, no increase in AEs and/or infectious AEs was observed in patients concomitantly exposed to emapalumab and anakinra. Therefore, continuation of anakinra for the underlying sJIA/AOSD, at the doses conventionally used and allowed in this trial, should be considered.38–40

The contribution of the concomitant anakinra to the achievement of MAS remission with emapalumab is to be considered negligible. Only one patient continued anakinra at a dose >4 mg/kg/day, a so-called higher dose that has been reported as potentially efficacious in MAS.10 14 41 Of note, several patients recruited to our study received anakinra before treatment with emapalumab and, in these patients, anakinra, even at doses >4 mg/kg/day, did not prevent or improve MAS. Furthermore, of the three patients in the trial who did not achieve MAS remission at week 8, two were receiving concomitant anakinra.

IFNγ neutralisation might increase predisposition to selected infections, as inferred based on data in humans with defective IFNγ activity, that is, subjects carrying autoantibodies to IFNγ or with IFNγ receptor deficiency.42 While herpes zoster infections are common in individuals with defective IFNγ activity, no cases were reported in this trial. Notably, per-protocol, patients received prophylaxis with acyclovir. We did not observe any infections from typical or atypical mycobacteria, H. capsulatum, Shigella, Salmonella, Campylobacter or Leishmania, known to occur more frequently in subjects with defective IFNγ activity. Noteworthy, patients with infections from these agents were excluded from the trial. No bacterial or opportunistic infections were reported. Viral infections or positive viral tests were reported in six patients during the trial and in two during follow-up while emapalumab was detectable. All events resolved spontaneously or with standard treatment. Of note, screening for Epstein-Barr virus, CMV and adenovirus every 2 weeks was required by protocol during the trial. Despite the fact that CMV is not reported at increased frequency in humans with defective IFNγ activity,42 it should be noted that five events were related to CMV, with four reported as reactivations or infections and one as test positive. Whether concomitant prolonged immunosuppression with high-dose glucocorticoids in critically ill patients might have contributed to this observation remains to be established. Despite CMV infections not being reported in humans with defective IFNγ activity,42 based on our data, patients with MAS receiving emapalumab should be screened for the presence of CMV. Although the entire age spectrum of AOSD is not covered, with only one patient above 20 years of age, it is highly unlikely that the infection risk changes with age.

The main limitation of this study is the single-arm nature of the design. However, a randomised controlled study in this patient population does not appear to be possible and, notably, ethical. This is based, first, on the absence of a standardised and/or validated treatment that has been prospectively investigated in patients with MAS who have failed high-dose glucocorticoids. Therefore, no drug is available as comparator for a head-to-head design. Second, the risk of death associated with this condition makes it unethical to continue, with no additional therapy, an ineffective treatment (ie, high-dose glucocorticoids) in patients who have already failed this treatment. Therefore, the option of continuing high-dose glucocorticoids plus placebo as a comparator arm is also not acceptable.

In conclusion, we demonstrate that IFNγ is an important driver of MAS secondary to sJIA/AOSD and that its neutralisation with emapalumab leads to remission of MAS in patients who failed high-dose glucocorticoids. Attention should be paid to viral infections, particularly to CMV; periodic screening for viral infections should be performed.

Data availability statement

Data are available on reasonable request. The data sharing policy of Sobi is available at the following site: As noted on that site, requests for access to the study data can be submitted to

Ethics statements

Patient consent for publication

Ethics approval

Approval was obtained from institutional review boards or independent ethics committees at each participating centre. The initial approval was obtained from the Institutional Review Board at Ospedale Pediatrico Bambino Gesù, IRCCS, Rome, Italy (reference number 1245/2016). Participants gave written informed consent to participate in the study before taking part.


The authors wish to acknowledge the contribution of the study participants and their families. Medical writing assistance was provided by Blair Hesp PhD CMPP of Kainic Medical Communications (Dunedin, New Zealand), which was funded by Sobi (Basel, Switzerland). PQ thanks Ms Thi An Vu (Clinical Research Unit, Necker Hospital, Paris, France) for her support. PAB acknowledges the conduct of the trial at Great Ormond Street Hospital NHS Foundation Trust and UCL Great Ormond Street Institute of Child Health being supported by the NIHR Great Ormond Street Hospital Biomedical Research Centre. AAG acknowledges PRCSG Clinical Research Associates at Cincinnati Children’s Hospital.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Handling editor Josef S Smolen

  • Twitter @claudiabrac, @GrantSchulert, @AntonJordi68

  • Presented at Part of the content in this manuscript has previously been presented at the American College of Rheumatology Convergence 2021, the 37th Annual Meeting of the Histiocyte Society, the 63rd American Society of Hematology Annual Meeting and Exposition, the European Alliance for Associations of Rheumatology 2022 Congress, the American Society of Pediatric Hematology and Oncology Conference 2022 and the Paediatric Rheumatology European Society 2022 Congress.

  • Contributors FDB, AAG, CdM, MB and PJ provided substantial contributions to the conception of the work. All authors substantially contributed to the acquisition, analysis or interpretation of data for the manuscript. All authors participated in drafting, revising and critically reviewing the manuscript for important intellectual content. All authors approved final version of this manuscript to be published and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. FDB is responsible for the overall content as guarantor.

  • Funding NI-0501-05 and NI-0501-06 ( numbers, NCT02069899 and NCT03311854) were funded by Sobi.

  • Competing interests FDB: consultant and research grants from AbbVie, Sobi, Pfizer, Roche, Sanofi, Novartis, Novimmune; AAG: consultant for Novartis, AB2 Bio, Novimmune, Sobi; PAB: consultant for Sobi, Novartis, Roche, UCB; CB, MP, GM: none declared; DE: speaker bureau for Sobi; CP: speaker bureau for Sobi; GS: consultant for Novartis, Sobi, Novimmune, AB2 Bio; PQ: consultant for AbbVie, Chugai-Roche, Lilly, Novartis, Pfizer, Sobi and speaker bureau for AbbVie, Chugai-Roche, Lilly, Novartis, Pfizer, Sobi; JA: consultant for Sobi, Novartis, Roche, Pfizer, AbbVie, GSK; CL: consultant for Sobi; RF: previously employed by Sobi; VA: previously employed by Sobi; MB: previously employed by Sobi; PJ: consultant for Sobi; CdM: consultant for Sobi, previously employed by Sobi.

  • Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the 'Methods' section for further details.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.