Overarching principles

OP A: sJIA and AOSD are the same disease, that should be designated by the same unique name, Still’s disease (formerly called sJIA/AOSD)

As mentioned above, AOSD was initially described as the adult counterpart of sJIA and the distinction of the two entities was mainly artificial based on the organisation of paediatric care, with an age limit of 16 years.3 This led in the subsequent decades to the separation of the two entities in terms of definition, classification criteria, clinical trials and in some instances drug approval.

Several elements argue in favour of the unification of the two entities. Although epidemiological data are scarce, they indicate that sJIA/AOSD incidence peaks in young children, that is, before the age of 6, but remains stable in adolescents and young adults.22 The SLR, conducted to identify similarities and discrepancies of clinical and biological features of the two diseases,23 showed that all sJIA/AOSD manifestations had similar prevalence in both entities, with the exception of sore throat and myalgia (often difficult to elicit by history in young, and so rarely annotated by paediatricians), weight loss (rarer in children and more frequently expressed as growth curve break in children), leucocyte counts (different norms in children and adults) and AA amyloidosis, whose prevalence has dramatically decreased during the last decades. Additionally, striking similarities in gene expression profiles in sJIA and AOSD were described.24

It is important to point out that age does matter in sJIA/AOSD for disease expression; for example, very young children, that is, less than 18 months, are more likely to display high levels of inflammation and to develop MAS.25 The same applies for differential diagnoses to be ruled out (see online supplemental table 1).

During the recommendation process, an online survey, conducted among the TF experts, revealed that 22 (92%) of the 24 experts who responded agreed on the principle of a unique name to identify sJIA and AOSD. During the survey, several names have been proposed and ‘Still’s disease’ was the most frequent proposal among the respondents. This term was approved and incorporated in OP A during the second Face-to-Face (F2F) meeting.

OP B: the treatment targets and the therapeutic strategy should be based on shared decision-making between the parents/patients and the treating team.

The TF thought it was important to highlight a few key points that should drive the management of patients with Still’s disease. As for all other systemic inflammatory diseases, it is important to engage patients and caregivers in the decision-making process of the therapeutic strategy in order to create a therapeutic alliance with health professionals and thus promote long-term adhesion to treatment.

OP C: treatment to target (T2T) by regularly assessing disease activity and adapting therapy accordingly is important. The ultimate goal is drug-free remission

Since Still’s disease evolution is often characterised by flares, it appeared relevant to the experts to apply the T2T principle for Still’s disease management. According to this principle, disease-modifying anti-rheumatic drugs (DMARDs) are dynamically increased or decreased (step-up or step-down approach) depending on disease activity. Sequential therapeutic targets were identified and will be addressed below; however, the experts wanted to emphasise that the ultimate goal of Still’s disease management is drug-free remission, which is realistic for a substantial proportion of patients thanks to recent innovative therapies.

OP D: MAS should be detected promptly and treated rapidly

MAS is the main challenging and life-threatening complication of Still’s disease, whatever their age—although more frequent in children under the age of 2.25–28 This complication can be present at the onset of Still’s disease or can occur during treatment, or during its course. This may happen even when the disease is in remission, since a relapse of Still’s disease may present directly as MAS especially in a context of infection. Due to its severity and its prognostic impact, physicians in charge of patients with Still’s disease should always be aware of this specific risk, monitor with adequate biological workups and be prepared to rapidly adjust treatment (see Recommendation Statements 10 and 11).

Recommendation statements

RS1: – To facilitate rapid diagnosis and initiate early treatment, operational definitions should be used to identify patients with Still’s disease,

1. Fever is typically spiking with temperature ≥39°C (102.2 °F) for at least 7 days.

2. Rash is transient and often coincides with fever spikes, preferentially involving trunk. It is typically erythematous (salmon pink), but other rashes (eg, urticarial) may be consistent with the diagnosis.

3. Musculoskeletal involvement is usually present with arthralgia/myalgia. Overt arthritis is supportive but not necessary for diagnosis and may appear later.

4. High levels of inflammation are typically identified by neutrophilic leucocytosis, increased ESR, serum CRP and ferritin.

Disease clinical expression may be heterogeneous and no symptom or biological feature is specific of Still’s disease. However, some manifestations are key, namely fever, skin rash (which may be challenging on coloured skin), arthralgia or arthritis and highly elevated acute phase reactants, typically CRP, high neutrophil count (PMN) and ferritin, but also increased platelet count, fibrinogen and D-dimers. To facilitate early diagnosis, the TF proposed operational definitions for the most relevant manifestations. A specific attention was paid to joint manifestations: arthralgia is commonly present, but arthritis appears often later with a median delay of 1 month after disease onset (range 0 to several months) (26, F de Benedetti personal communication); additionally, the underlying pathogenic immunological mechanisms seem to be quite similar in patients with Still’s disease with or without arthritis.29 Thus, requiring arthritis to make the diagnosis of Still’s disease leads to unnecessary and potentially deleterious diagnostic delays. For this reason, the TF strongly recommends that presence of arthritis is not mandatory for the diagnosis of Still’s disease.

Although not developed for Still’s disease diagnosis ascertainment, classification criteria may help clinicians to identify the disease. Several criteria sets have been proposed for sJIA or AOSD (table 2), and one—Yamaguchi—has been tested and validated in both sJIA and AOSD with high sensitivity.30 Noteworthy, arthritis is not mandatory for the diagnosis in the Yamaguchi’s criteria as well as in the recently proposed classification criteria for sJIA.9 The working group recommends that a common set of classification criteria encompassing children and adults is needed; this is listed in the research agenda (table 3).

RS2: marked elevation of serum IL-18 and/or S100 proteins (eg, calprotectin) strongly supports the diagnosis, and therefore should be measured if available.

Several inflammatory molecules are elevated in numerous case series and are proposed as diagnostic biomarkers.23 29 The list of candidate biomarkers is quite long.23 Among these, serum IL-18 levels and S100 proteins levels are the most studied biomarkers in both children and adults. IL-18, a proinflammatory cytokine produced by inflammasome activation, drives IFNγ overproduction. Alarmins such as S100 proteins, that is, S100 A8/A9 (also called serum calprotectin) and S100 A12, which are produced by innate immunity cells (monocytes and neutrophils), act as danger-associated molecular patterns to amplify inflammation.31 Highly elevated levels of these biomarkers appear to identify Still’s disease with high sensitivity and specificity. However, there is no consensual or validated threshold for these biomarkers as sensitivity and specificity were assessed by comparison with different control groups (fever of unknown origin, other autoinflammatory disorders, autoimmune diseases including vasculitis, infections, haematological diseases, neoplasia, liver diseases).23 Serum levels of IL-18 and S100 proteins are presently used for diagnosis in selected tertiary centres.31 To enable widespread use, assays and cut-off values need to be validated. This was included in the research agenda (table 3).

RS3: alternative diagnoses such as malignancies, infectious diseases, other immune-mediated inflammatory diseases and monogenic autoinflammatory disorders should be carefully considered.

Still’s disease diagnosis relies on a combination of clinical and biological findings, none of them being in isolation specific of the disease. Other diseases such as infectious diseases, malignancies or other immune-mediated diseases (IMIDs) can mimic Still’s disease. In general, this includes also monogenic (germline or somatic, inherited or acquired) inflammatory disorders. In the adult patients, VEXAS and CHIP should be considered.32–35 This has important consequences when Still’s disease treatment has to be initiated, glucocorticoids (GC) or immunomodulating agents being potentially deleterious if Still’s disease is misdiagnosed, particularly in the presence of malignancies. Thus, the TF highlighted the need to carefully consider potential differential diagnoses, without inadequately delaying treatment initiation. Based on previous expert consensus documents on fever of unknown origin36 or Still’s disease,37 a comprehensive list of these alternative diagnoses is proposed (online supplemental table 1), as well as a list of potentially relevant investigations according to the alternative diagnostic hypotheses (online supplemental table 2).

RS4: clinically inactive disease (CID) is defined as absence of Still’s disease-related symptoms and normal ESR or CRP. Remission is defined as a period of at least 6 months with CID.

Several outcome measures have been used in Still’s disease clinical trials and longitudinal observational studies, usually developed for juvenile idiopathic arthritis or for rheumatoid arthritis and adapted to the systemic nature of Still’s disease by adding fever to the other items. These measures include the American College of Rheumatology (ACR) responses, the JIA ACR responses (JIA-ACR), EULAR responses, CID, and remission with different definitions in different studies.38 However, the TF favoured to use the same outcome irrespectively of patient age, and proposed to define two main targets derived from the consensus proposals of Wallace et al for JIA.39

CID is a single point-in time (snapshot) measure, defined as absence of any Still’s disease-related manifestation including normal acute phase reactants such as ESR or CRP. If multiple acute phase reactant measures are performed, all have to be normal (except if another explanation exists to explain abnormal value, such as anaemia for ESR). Some TF members proposed to include physician global assessment with a value equal to or less than 10 on a 0–100 visual analogue scale (VAS), that is, less than 1 on a 0–10 VAS.

Remission is a time-integrated measure defined as the maintenance of CID over a period≥6 months, irrespective of treatment status. Remission can thus be defined as on or off treatment (ie, drug-free remission).

RS5: in order to achieve the ultimate goal (drug-free remission), the following intermediate targets are recommended:

1. At day 7, resolution of fever and reduction of CRP by >50%,

2. At week 4, no fever, reduction of active (or swollen) joint count by >50%, normal CRP and physician and patient/parent global assessment less than 20 on a 0–100 VAS,

3. At month 3, CID with GCs less than 0.1 (adults) or 0.2 (children) mg/kg/day,

4. At month 6, CID without GCs.

A stepwise T2T approach should be implemented, adapting treatments according to disease activity. Due to Still’s disease potential severity and the risk of life-threatening manifestations, the TF was in favour of defining intermediate targets to monitor disease-related symptoms (fever, active (swollen) joints if present, CRP, physician or patient global assessment, and so on) and their evolution during the first weeks of treatment.

Several disease activity scores have been used in clinical trials or longitudinal observational studies, some derived from other IMID such as Disease Activity Score (DAS), some being more specific of Still’s disease, such as the systemic Juvenile Arthritis Disease Activity Score (sJADAS)40 or the modified Pouchot’s score,41 both exploring the whole spectrum of Still’s disease manifestations. The TF experts were uniformly in favour of using the same tool to quantify disease activity in children or adults with Still’s disease, but they acknowledged that the scientific evidence is not robust enough to recommend one single tool for the moment. The TF identified this as an issue for the research agenda (table 3).

Taking this into consideration and based on existing recommendations proposed by German or North American paediatric rheumatologists,8 42 the TF defined four pragmatic targets, at day 7, week 4 and month 3 and 6 of treatment, based on simple and clinically relevant criteria, rather than disease activity composite measures. These were developed to guide optimal patient management throughout the first 6 months of disease in a newly diagnosed patient or in a patient relapsing during drug-free remission. When interpreting these targets, it should be taken into account that IL-6 inhibitors may blunt CRP increase.43 44 The definition and management of refractory or difficult-to-treat patients will be addressed below. The target dose of GCs at 3 months depends on age, and was set, consistently for adults and children to respectively, at 0.1 mg/kg/day and 0.2 mg/kg/day.45 46 Patient management should include monitoring of safety in addition to disease activity.

RS6: to avoid prolonged systemic GC use for achieving and maintaining the target, the use of IL-1 and IL-6 inhibitors should be prioritised due to high evidence of efficacy.

The introduction of IL-1 and IL-6 inhibitors has changed the course of the disease so that nowadays GC use can be markedly limited, or even avoided. The TF acknowledges the efficacy of GCs in managing Still’s disease. Available evidence, although low-level, and real-world experience suggest that high-dose (ie, ≥1 mg/kg/day of prednisone equivalent) are more efficacious than low-dose GCs. High dose should be considered at disease onset and particularly in the presence of severe symptoms, impending MAS, and severe pericarditis. The TF also acknowledges that some patients may not require GCs, particularly when an IL-1 or IL-6 targeted biologic disease-modifying antirheumatic agents (bDMARDs) can be initiated. While short-term use of GCs (for a few weeks) may not necessarily be associated with clinically relevant side effects, their long-term use often causes severe side effects, some of which are common to both children and adults, while others are more specific to children (eg, growth retardation and bone mineralisation defects) or adults (eg, metabolic syndrome). Whatever the situation, the TF emphasised that GC use to maintain the target at any time during the treatment course must be avoided: in other terms, in the presence of GC dependence, other therapies should be added to achieve the goal of disease control without GCs.

Despite the small number of randomised controlled trials (RCTs) with IL-1 or IL-6 inhibitors performed in sJIA or AOSD, the TF strongly recommends their use based also on the overwhelming body of evidence from real-world experience supporting their efficacy to control all aspects of the disease—including both systemic and joint manifestations—and to limit exposure to GC.38 Their safety profile is well established, and the overall benefit–risk balance is extremely favourable. Serious adverse events appear to be more frequent during IL-6 inhibition than during IL-1 inhibition (table 4). Also, infectious adverse events, serious and non-serious, are more frequent during IL-6 inhibition compared with IL-1 inhibition (table 4). Among the IL-1 inhibitors, anakinra appears to have the most reassuring safety profile.38 It is also worth to note that anakinra has been used in critically ill patients with sepsis admitted in intensive care unit with no safety concern47 and that IL-1 and IL-6 inhibitors have been proven to be safe in patients with severe COVID-19.48–50

Evidences supporting the use of conventional synthetic DMARDs (csDMARDs) in Still’s disease is scarce. In the one RCT available (performed in sJIA), methotrexate (MTX) was not superior to placebo even at a low response threshold (ACR30).51 DMARDs have been traditionally used, particularly in patients with prominent joint involvement. Several observational studies have reported some response with MTX or with MTX in combination with other DMARDs.52 They can be used as GC-sparing agent, although, nowadays, bDMARDs targeting IL-1 or IL-6 are the treatment of choice. csDMARDs should be considered in countries where IL-1 and IL-6 inhibitors are not available.

There is no RCT evidence supporting the efficacy of non steroidal anti-inflammatory drugs (NSAIDs) in Still’s disease. However, the TF recommends to limit their use as symptomatic treatments to manage fever and arthralgia during the diagnostic work-ups.

RS7: an IL-1 or an IL-6 inhibitor should be initiated as early as possible when the diagnosis is established.

No formal RCT has addressed the timing of initiation of IL-1 or IL-6 inhibitors in Still’s disease. Indeed, controlled trials have invariably enrolled patients with long-disease duration. However, real-life data have shown that early initiation of IL-1 or IL-6 inhibitors is associated with very favourable short-term outcome with high rates of CID off GCs.38 Low-level evidence from observational studies suggest that these drugs should be initiated before 3 months from symptoms onset.30 Comparisons with historical data, and with patients in whom treatment with bDMARDs is delayed, suggest that early initiation of IL-1 or IL-6 inhibitors may also decrease the number of patients with a chronic persistent course. These data are consistent with the hypothesis of a therapeutic window of opportunity in the course of Still’s disease. Translational data generated in murine models, namely IL-1RN-deficient mice, also support this window of opportunity. These mice develop spontaneous arthritis associated with increased number of activated Th17 cells. Arthritis development, as well as Th17 cell increase, was prevented by early, but not late, initiation of with an IL-1 inhibitor.53 Consistently, in sJIA increased Th17 and increased γ/δ T cells with a Th17 profile are associated with chronic persistent disease.54 55 Based on the efficacy data and the GC sparing effect, and since the safety profile of IL-1 and IL-6 inhibitors is reassuring and their use does not interfere with the diagnostic work-up at onset, the TF recommends to initiate an IL-1 or an IL-6 inhibitor as early as possible.

RS8: maintenance of CID for 3–6 months without GCs should be achieved before initiating bDMARD tapering.

With presently available IL-1 and IL-6 inhibitors, many patients achieve persistent remission on medication. The TF reiterates that withdrawal of GCs is a mandatory first step. Many physicians and patients want to attempt bDMARD tapering or withdrawal. Very few studies address the issue of drug management after achieving remission, that is, bDMARD maintenance, tapering or discontinuation. Based on their experience in Still’s disease, the TF members preferentially recommended to progressively taper as performed in other inflammatory joint disease: a relapse will lead to an end-of-dose effect, that the patient will directly manage by bringing back bDMARD injections closer. With regards to patient empowerment, this strategy is preferable to a complete bDMARD discontinuation. In case of insufficient disease control, progressive bDMARD tapering would more likely lead to loss of efficacy in the last hours or days before the next injection, while complete and abrupt discontinuation could lead to a more severe disease flare. The TF recommends that bDMARD tapering should be considered in patients who have maintained CID off GCs for at least 3 months, possibly 6 months for patients with severe and difficult-to-treat Still’s disease. Tapering should be conducted progressively, that is, stepwise by steps of 3–6 months, either by dose reduction or injection interval prolongation. The latter appears more appropriate as it leads to reduced number of injections (relevant for the patients whatever their age) and drug wastage. Possible schedule of tapering, based on published studies or real-life experience, is proposed in online supplemental table 3.

RS9: severe/life-threatening complications, including MAS or LD, may develop at any point during the disease course. Patients should be actively screened and monitored.

The TF considered important to warn physicians that Still’s disease is a challenging condition in which severe or life-threatening complications may occur abruptly at any time during the disease course, that is, at disease onset, during the first weeks of course while diagnostic investigations are being performed, during bDMARD treatment (see table 4) even when Still’s disease is well controlled, or later when drug tapering is implemented. MAS is the most frequent complication occurring in 15%–20% of patients with Still’s disease. In addition, other rarer complications have been described including cardiac involvement (tamponade, myocarditis), lung involvement (Still’s disease-related lung disease (LD), pulmonary hypertension, acute respiratory distress syndrome), fulminant hepatitis, DIC and thrombotic microangiopathy.5 17 56 For this reason, the TF highlighted that patient monitoring and follow-up should be careful and that the patients should be advised to contact the treatment team in case of relapse/flare or new unexpected symptoms.

RS10: MAS should be considered in patients with persistent fever, splenomegaly, elevated or rising serum ferritin, inappropriately low cell counts, abnormal LFTs, intravascular activation of coagulation, elevated or rising serum triglycerides.

The term MAS identifies secondary haemophagocytic lymphohistiocytosis (HLH) occurring in the context of rheumatic diseases. MAS and in general all forms of HLH are considered prototypical hyperinflammatory syndromes caused by excessive activation of the innate and adaptive immune response. The pattern of hyperinflammation is recognisable in MAS and includes the items listed in RS10. None of these laboratory abnormalities in isolation is specific for MAS. The TF wants to emphasise that, in order to identify early MAS, it is important to pay attention to the pattern as a whole and particularly to the evolution over time of the laboratory parameters,57 as also pointed out in a recent EULAR/ACR endorsed consensus effort on the early diagnosis of hyperinflammation.58 The available classification criteria and diagnostic scores (H-scores and M-scores) for MAS,59 60 derived with statistical analyses based on different sets of patients and disease controls, capture these laboratory abnormalities (online supplemental table 5) and could be used to help in the diagnosis. The TF underscores that MAS can occur also in patients receiving IL-1 or IL-6 inhibitors and emphasises that, in patients receiving canakinumab or tocilizumab, clinical and laboratory abnormalities may be blunted or delayed, particularly fever and ferritin increase.61 During withdrawal and/or tapering of GCs, IL-1 or IL-6 inhibitors, patients might be at risk for MAS and therefore should be carefully monitored.

In all patients with Still’s disease, ferritin should be measured routinely. In patients at risk or with suspected MAS, ferritin measurements should be repeated on a daily or even on a bidaily basis to allow for a tight monitoring.57 58 Although other rarer causes of hyperferritinaemia have been described, ferritin remains very important in the diagnosis of MAS with high sensitivity and specificity.62 Ferritin trends over time should always be interpreted in the context of the other laboratory abnormalities that are part of hyperinflammation.57 58 Other modern biomarkers can be used when available: serum sIL-2R, CXCL-9, IL-18 levels and percentages of activated CD8 T lymphocytes, or possibly activated CD4+or monocytes.62–64 All these biomarkers display high sensitivity and specificity.62

RS11: MAS treatment must include high dose GCs. In addition, treatments including anakinra, ciclosporin and/or IFNγ inhibitors should be considered as part of initial therapy.

The TF highlighted the need for a search for triggering factors (most frequently infections), while initiating specific MAS treatments.65 Infections, if identified, should be aggressively treated with appropriate antimicrobial therapy as their control facilitates the management of MAS. Regarding potential infectious triggers, many infectious agents can trigger MAS.58 The presence of an active infection should not delay the initiation of MAS treatment.

High-dose GCs are the mainstay of the treatment in patients with MAS. In MAS, GCs are usually administered by intravenous pulses of methylprednisolone (15 to 30 mg/kg/day, maximum dose 1 g/infusion). Dexamethasone should be considered in the presence of central nervous system involvement as it better crosses the blood–brain barrier. High-dose GCs may achieve satisfactory results in a substantial number of patients, particularly if initiated early.

Additional treatments include ciclosporin, or possibly another calcineurin inhibitor (tacrolimus), anakinra and IFN-γ neutralising monoclonal antibody (online supplemental table 6). Ciclosporin has been proven to be ineffective in primary familial HLH and has been dropped from the standard of care approach in this specific condition.66 However, despite the absence of formal clinical trial in secondary HLH/MAS, the TF acknowledges the large and often positive experience with ciclosporin in MAS and recommends that ciclosporin should considered in case of inadequate response to GC.58 Indeed, it is particularly valuable in less resourced countries. Ciclosporin can be administered orally or intravenously, particularly in the critical care setting. Similarly to ciclosporin, anakinra has not been tested in clinical trial in MAS. However, there is a large real-world experience on its use in MAS. In these patients it should be used at doses higher than the standard 1 to 2 mg/kg/day, and possibly in intravenously repeated doses; this approach resulted in satisfactory responses in several patients.62 67 Emapalumab, an anti-IFN-γ antibody, is the only targeted therapy that has been tested in a clinical trial (open-labelled single arm) in Still’s disease-related MAS.68 In patients with severe MAS and who had failed standard of care with high-dose GCs, treatment with emapalumab was associated with achievement of MAS remission in the great majority of the patients with a marked GC sparing effect and a reassuring safety profile. It is important to point out that emapalumab is not yet approved in Europe. Additionally, the potential interest of JAK inhibitors should be mentioned, since a few case reports reported efficacy of JAK1/JAK2 inhibitors—that is, ruxolitinib or baricitinib—in such patients69 70 (online supplemental table 6). Finally, in refractory MAS, low-dose etoposide may also be considered71 (online supplemental table 6).

The TF recommends that the addition of the above-mentioned treatments is considered in patients with initial unsatisfactory response to high dose GCs and in patients with severe MAS and rapid worsening. The TF highlights that combination therapies with multiple agents on a background of high dose GCs are often necessary and should be considered also as initial therapy. The higher mortality of patients with MAS reported in adults72 may also be considered when choosing a treatment or a combination therapy. Such decisions should be discussed with experts of a reference centre.

RS12: LD should be actively screened by search for clinical symptoms (eg, clubbing, persistent cough, shortness of breath) and pulmonary function tests (pulse oxygenometry, diffusing capacit of carbon monoxyde (DLCO) measurement), and investigated by high-resolution CT-scan in any patients with clinical symptoms.

In the last decades, physicians have been recognising, with progressively increasing frequency, patients, particularly children, with Still’s disease and inflammatory LD.73 74 LD was initially described in North America; however, cases have now been recognised in Europe and other regions of the world.75 Still’s LD is characterised by interstitial inflammatory infiltrates of CD4 and CD8 lymphocytes and by intra-alveolar deposition of proteinaceous material. LD appears to be associated with younger age at onset, Down’s syndrome, occurrence of MAS and particularly of recurrent MAS, as well as high serum IL-18 levels.74 Data are also accumulating, partly still unpublished, suggesting that carriage of the HLA DRB1*1501 is strongly associated with LD.75 76 Given the potential strength of the association, the TF believes that HLA typing should performed in several different populations with Still’s disease in order to provide information on the pathogenic and the clinical significance of this association. This issue has been included in the research agenda (table 3). Although the exact pathogenesis of Still’s LD is not yet understood, there is increasing evidence of a role for the IL-18/IFNγ pathway and activated T cells.77

All patients with Still’s disease should be evaluated carefully for the occurrence of LD.78 Particular attention must be paid to the patients with the above-mentioned risk factors. Clinical symptoms are important, but may present lately when LD is already advanced. Pulmonary function tests are useful; in children and particularly in small children pulse oximetry (specifically continuous overnight pulse oximetry) is useful to detect early functional impact of LD. Given the potential severity of Still’s LD, the TF recommends that a high-resolution CT scan is performed in any patients with clinical concerns. The TF also highlighted that echocardiography is useful to screen and detect pulmonary hypertension, or other rarer complications, such as myocarditis. No evidence-based screening protocol is available. A proposal for such a protocol has been recently published and could represent a guide for treating physicians.78

Finally, as the field of Still’s LD is rapidly evolving from both clinical and therapeutic perspectives, the TF emphasised that patients with LD should be managed in collaboration with expert centres (see RS 14).

RS13: based on the available data, the presence of risk factors for Still’s LD or the development of Still’s LD should not be considered as a contraindication to IL-1 or IL-6 inhibitors.

One current hypotheses on the pathogenesis of Still’s LD is based on the occurrence of hypersensitivity-like reactions to IL-1 and/or IL-6 inhibitors and postulates that LD is a consequence of a Drug Rash with Eosinophilia and Systemic Symptoms (DRESS)-like reaction to these drugs.73 76 DRESS-like reactions would imply withdrawal of IL-1 and IL-6 inhibitors in patients with Still’s LD, so it is important to emphasise that the hypothesis remains unproven and indeed may be difficult to reconcile with all available data. Furthermore, IL-1 and IL-6 inhibitors have been shown to be very effective in Still’s disease and led to substantial improvement in patient quality of life and in long-term outcome. Withdrawal of these effective drugs was not associated with notable improvement of LD in the great majority of the patients. Moreover, withdrawal would make it very difficult to manage the patients with no other options than using markedly higher doses of GCs. It also exposes the patients to a significant risk of severe flares, which can be associated with severe MAS. In such occasions, fatalities have already been reported. Considering the overall benefit–risk ratio, the TF recommends that IL-1 and IL-6 inhibitors should not be withdrawn and thus should be continued in patients with Still’s LD. This is in line with a recent observation in a Dutch prospective study.79 The TF also acknowledges that there is insufficient evidence to withhold first-line IL-1 or IL-6 inhibitors in patients with new onset Still’s disease and LD risk factors. Additionally, given the potential involvement of T cells in LD pathogenesis, some TF experts felt that it is reasonable to initiate T cell-directed immunosuppression in patients at high risk for LD or developing LD, although direct clinical evidence for this strategy is still absent. Benefits of JAK inhibitors in patients with Still’s LD has been reported in two single cases.80 81

RS14: difficult to treat (D2T) patients, those with severe MAS and those with Still’s LD should be managed in collaboration with Still’s disease expert centres.

The course of the disease may be challenging since its beginning, with MAS often presenting at onset. D2T patients, that is, those who have failed IL-1 and IL-6 inhibitors, those with severe/recurrent MAS and those with LD may represent a major challenge. For such patients, intermediate targets might be less ambitious. Therapeutic strategies in D2T patients may include repeated intra-articular GC injections, periodic GC pulses, JAK inhibitors, combination of csDMARDs and bDMARDs and novel IL-18 or IFNγ inhibitors. Some of these are experimental, and there is no evidence supporting one specific strategy. During the last decades, rare disease networks have been developed in Europe (European Reference Networks) and outside Europe with the objective of facilitating the access to both expert physicians and patients with disease specific expertise, through specific consultations or multidisciplinary rounds. Physicians are therefore encouraged to contact centres with Still’s disease expertise (in Europe through the ERN-RITA; https://ern-rita.org/) to discuss treatment escalation and in general the management of patients with inadequate response to standard of care. This will enable optimisation of Still’s disease management, allowing also to foster research on difficult patients and to favour the access to innovative therapies (compassionate access or enrolment in clinical trials).

Therapeutic algorithm

The above RSs lead to propose a two-part algorithm (figure 1). At onset or during a flare a patient should receive an IL-1 or IL-6 inhibitor as early as possible. Since there is no clear predictor at disease onset to identify patients who will develop a chronic disease course, the TF recommends considering first-line biologic therapy irrespective of disease severity. When bacterial infection remains on the differential diagnosis, an IL-1 blocker is preferred. Anakinra is very often used initially because of its short half-life and its reassuring safety profile. High-dose GCs are indicated in patients with high disease severity (high spiking fever, wide-spread polyarthritis, high levels of pain (VAS >6–7/10), pericarditis, impending MAS (elevated LFT and/or high serum ferritin levels). In milder presentations, GC may be used at low or intermediate doses, but are not mandatory. When they are started, GC should be progressively tapered as soon as possible with the aim of achieving CID on low dose GC at 3 months and, subsequently, CID off GC at 6 months from treatment initiation. Failure to achieve one of these two targets should prompt IL-1 or IL-6 inhibitors rotation during which slow progressive GC tapering should be continued. A patient who fails to achieve CID off GC after IL-1/IL-6 inhibitor rotation should be considered as D2T and should be discussed in multidisciplinary rounds with an expert in Still’s disease, for example, at a reference centre (in Europe through the ERN-RITA, https://ern-rita.org/); the use of less well-established therapeutic options should be discussed. These include JAK inhibitors, haematopoietic stem cell transplantation and immunosuppressants.82–86

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Treatment algorithm to manage people with Still’s disease footnotes. ¹High disease activity includes: high spiking fever, wide-spread polyarthritis, high levels of pain (VAS >6–7/10), pericarditis, impending MAS (elevated LFT and/or high serum ferritin levels). ²High-dose GC is equal to or greater than 1 mg/kg/day of prednisone equivalent in adults and 2 mg/kg/day in children. Low-dose GC is equal to or lower than 0.1 mg/kg/day of prednisone equivalent in adults and 0.2 mg/kg/day in children. ³Anakinra would be the preferred choice for a patient with impending MAS. High-dose anakinra (> 4 mg/kg/day in children or 100 mg twice a day in adults) are often used in these circumstances. ⁴GC tapering should be started as soon as the first intermediate target (see RS5) is reached (no fever and decrease by 50% of active joints). ⁵Difficult-to-treat Still’s disease requires discussion in multidisciplinary round of a ERN reference centre. ⁶Experimental therapy may include JAK inhibitors, emapalumab, bispecific antibody to IL-1/IL-18. The choice among these possible therapies is driven by the features of the disease (eg, chronic relapsing MAS, severe and persistent joint involvement, lung disease) in consultation with ERN experts. ⁷Remission defined as CID maintained for 6 months or more. ⁸ Tapering of IL-1i, IL-6Ri or experimental therapy is usually based on progressive injection spacing. ⁹Patient education and counselling aims to learn to the patients the signs and symptoms that could indicate Still’s disease relapse and the ways to reach the medical team to manage it. CID, clinically inactive disease; GC, glucocorticoid, IL, interleukin; LFT, liver function test; MAS, macrophage activation syndrome.

If at any time CID off GC is achieved, IL-1 or IL-6 inhibitors or the experimental therapy should be maintained for 3–6 months (until remission is achieved). At least 6 months of CID off GC should be maintained before tapering of bDMARDS or of experimental therapies is initiated. A patient who presents a major flare while in CID off GC should be induced again similarly as if at onset.

A patient who develops MAS at any time before initiation and during treatment should be treated promptly as described.

Quality indicators

Based on the recommendation statements and TF discussions during the production process, three quality indicators were identified: (1) proportion of patients achieving CID off GCs at 6 months; (2) proportion of patients having started early IL-1 or IL-6 inhibitors, that is, in the first 3 months after disease onset; (3) proportion of severe patients with Still’s disease (D2T patients, severe/recurrent MAS, Still’s LD) referred to or managed in collaboration with an expert centre. These were voted on line with percent consensus of 96.8%, 96.8% and 100%, respectively.