The role of interleukin‐18 in the diagnosis and monitoring of hemophagocytic lymphohistiocytosis/macrophage activation syndrome – a systematic review

Systematic review of studies investigating interleukin‐18 (IL‐18) in hemophagocytic lymphohistiocytosis (HLH) and macrophage activation syndrome (MAS).Serum IL‐18 appears to discriminate well between HLH/MAS and other inflammatory conditions.

Summary

Hemophagocytic lymphohistiocytosis (HLH) is a life‐threatening, hyperinflammatory disorder, characterized by multiorgan failure, fever and cytopenias. The diagnosis of HLH and its subtype Macrophage Activation Syndrome (MAS) remains a challenge. Interleukin 18 (IL‐18) is emerging as a potential biomarker for HLH/MAS but is currently not a part of diagnostic criteria. This systematic review aimed to assess the potential role of IL‐18 in the diagnosis and monitoring of HLH and MAS, and was performed according to the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines. PubMed and Embase were searched on 30 January 2020. Studies included all subtypes of HLH and a range of underlying disorders in both children and adults. A total of 14 studies were included. Generally, serum IL‐18 was elevated in both primary and secondary HLH (> 1000 pg/ml) compared with other inflammatory conditions and with healthy individuals; thus, serum IL‐18 may be able to discriminate between HLH and other inflammatory conditions. Significantly increased IL‐18 (> 10 000 pg/ml) was also consistently described in MAS compared with other subtypes of HLH. The ability of IL‐18 to distinguish MAS from systemic juvenile idiopathic arthritis (JIA) is less unambiguous, as IL‐18 levels > 100 000 pg/ml were described in sJIA patients both with and without MAS. IL‐18 may help to differentiate between HLH subtypes and other inflammatory conditions. As HLH and MAS are rare disorders, only few and relatively small studies exist on the subject. Larger, prospective multi‐center studies are called for to assess the diagnostic precision of IL‐18 for HLH and MAS.

Introduction

Hemophagocytic lymphohistiocytosis (HLH) is a life‐threatening disorder characterized by hyperinflammation with persistent activation of cytotoxic T cells, natural killer (NK) cells and macrophages [1]. Most commonly, HLH presents with hepatosplenomegaly, prolonged fever and cytopenias in an acutely ill patient [2]. Other common findings are liver dysfunction, consumption coagulopathy and neurological symptoms, while the biochemical profile typically shows elevated serum ferritin, soluble interleukin 2‐receptor (sIL‐2r) and soluble CD163 (sCD163), as well as transaminases and triglycerides [2]. HLH had a mortality rate of 95% until the implementation of diagnostic criteria, and new treatment regimens have resulted in a 55% cure rate [3].

In the present review the following terms will be used: primary HLH (pHLH), secondary HLH (sHLH) and macrophage activation syndrome (MAS). pHLH is also known as familiar or hereditary HLH, and is caused by mutations in genes encoding proteins in cytotoxic granule activity pathways [4]. The first of these genes to be recognized was the perforin gene (PRF1) [5]. The mutations inhibit cytotoxic activity of NK cells and cytotoxic T lymphocytes but not their ability to produce cytokines. The cytotoxic cell is unable to kill the target cell and is activated continuously, resulting in a cytokine storm with activation of macrophages and other immune cells, leading to hemophagocytosis, hyperinflammation and organ damage [6]. The incidence of pHLH is uncertain, but is estimated to range from one in 50 000 to 150 000 [7]. sHLH occurs as a complication to rheumatic, infectious and malignant diseases, e.g. Epstein–Barr virus (EBV) infections with cytotoxic T cell involvement [8, 9]. MAS is a subtype of sHLH and is defined as a complication to rheumatological diseases, most commonly systemic juvenile idiopathic arthritis (sJIA) or the adult counterpart adult‐onset Still’s disease (AOSD) [4]. Fulminant MAS is seen in approximately 10% and subclinical MAS is presumed to be present in approximately 30% of sJIA cases [4].

HLH and MAS are life‐threatening conditions, and prompt diagnosis and treatment is essential for survival. However, their diagnosis still poses a challenge. Currently, sHLH is diagnosed according to the 2004 HLH diagnostic guidelines [2], which include serum ferritin and sIL‐2r, but both these serum markers lack specificity, especially in adult‐onset HLH [10]. Ferritin is an acute‐phase reactant and is also elevated in hemochromatosis [11], while increased sIL‐2r levels are described in rheumatoid arthritis and sarcoidosis [12]. New and more specific biomarkers could add great value in HLH/MAS diagnosis.

One such potential biomarker is interleukin (IL)‐18. IL‐18 is produced by macrophages and induces interferon (IFN)‐γ‐production in cytotoxic T lymphocytes and NK cells via the IL‐18 receptor (IL‐18Ra), in concert with IL‐12, as a part of normal immune function [13, 14]. In HLH and MAS, activated macrophages present in the liver, spleen and bone marrow are considered an important source of IL‐18, as well as other proinflammatory cytokines; however, there is also evidence that other cells contribute to IL‐18 production, e.g. epithelial cells [15], and this pathway may play an additional role in the cytokine storm associated with MAS. Increased IL‐18 levels have been demonstrated in patients with pHLH [16], sHLH [17] and MAS [18]. However, as HLH and MAS are rare diseases the majority of studies on the subject are small, exploratory and performed on heterogeneous populations. Furthermore, as high serum IL‐18 has also been described in sJIA patients without MAS [19], the ability of IL‐18 to discriminate between active sJIA and MAS is uncertain [20]. Whether IL‐18 can distinguish between pHLH and sHLH/MAS is also not elucidated. The aim of the present study was to perform a systematic review of the existing literature on the association between serum IL‐18 and HLH in humans and thereby assess the potential role of IL‐18 as a biomarker for diagnosis and monitoring of HLH.

Materials and methods

The present systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA) guidelines [21].

Literature search

PubMed and Embase were searched on 30 January 2020.

PubMed

[[[[[[[[[[‘Lymphohistiocytosis, Hemophagocytic’ (Mesh) OR ‘Hemophagocytic lymphohistiocytosis’ OR ‘Haemophagocytic lymphohistiocytosis’ OR ‘Hemophagocytic syndrome’ OR ‘Haemophagocytic syndrome’ OR ‘Macrophage Activation Syndrome’(Mesh)] OR ‘Macrophage Activation Syndrome’ OR MAS OR HLH OR hypercytokinaemia OR ‘cytokine storm’ OR ‘hemophagocytic activation syndrome’ AND [[[[[[[[‘Interleukin‐18’ (Mesh)) OR ‘Interleukin‐18’ OR ‘Interleukin 18’ OR il18 OR il‐18 OR ‘interferon gamma inducing factor’ OR ‘interferon‐gamma‐inducing‐factor’ OR ‘interferon‐gamma inducing factor’ OR IGIF. No filters were set.

Embase

(‘hemophagocytic syndrome’/exp OR ‘macrophage activation syndrome’/exp OR ‘macrophage activation syndrome’ OR ‘haemophagocytic lymphohistiocytosis’ OR ‘haemophagocytic syndrome’ OR ‘hemophagocytic syndrome’ OR ‘hlh’ OR ‘mas’ OR ‘hemophagocytic activation syndrome’ OR ‘hypercytokinaemia’ OR ‘hyperinflammation’ OR ‘hemophagocytic lymphohistiocytosis’ OR ‘cytokine storm’/exp) AND (‘interleukin 18’/exp OR ‘interleukin 18’ OR ‘interleukin‐18’ OR ‘il‐18’ OR ‘il18’ OR ‘interferon gamma inducing factor’ OR ‘interferon‐gamma inducing factor’ OR ‘interferon‐gamma‐inducing factor’ OR ‘interferon‐gamma‐inducing‐factor’ OR ‘interferon‐gamma inducing‐factor’ OR igif) AND (‘article’/it OR ‘article in press’/it OR ‘conference paper’/it OR ‘editorial’/it OR ‘letter’/it OR ‘note’/it OR ‘review’/it OR ‘short survey’/it). Filter set to exclude conference abstracts.

Inclusion and exclusion process

Studies were included if they (1) contained original data from human studies, (2) included at least five subjects with HLH/MAS, (3) measured IL‐18 in serum or plasma and (4) investigated the association between IL‐18 and HLH or MAS. Both pediatric and adult populations were included.

All studies were screened by title and abstract for eligibility. Twenty‐five studies were randomly selected to be screened by two of the authors (J.M.K. and J.B.L.), and any disagreements were discussed until consensus. The remaining studies were screened by J.M.K. Similarly, 10 studies deemed eligible for full‐text reading were screened by two authors (J.M.K. and J.B.L.), and disagreements were discussed until consensus. The remaining studies were included or excluded by J.M.K. In case of doubt, studies were discussed among all three authors.

A quality assessment was performed on all included studies using the Quality Assessment of Diagnostic Accuracy Studies (QUADAS‐2) tool [22] by two authors (J.M.K. and J.B.L.).

Statistics

No meta‐analysis was performed as all studies except one [16] reported their results as median with range.

Results

Figure 1 shows the inclusion and exclusion process. A total of 14 articles were included in the present review, all of which were case–control studies. Study characteristics are displayed in Table 1. Briefly, nine studies included MAS patients [15, 16, 18, 20, 23, 24, 25, 26, 27], eight included patients with sHLH of other causes [15, 17, 20, 24, 28, 29, 30, 31] and six studies included pHLH [15, 16, 23, 29, 30, 31]. Other inflammatory conditions included were EBV infection (n = 5) [20, 24, 29, 30, 31], Kawasaki’s disease (n = 2) [20, 28], sJIA/AOSD (n = 4) [18, 20, 25, 26] and patients with other infectious, inflammatory or malignant disorders [15, 17, 23]. Seven studies included healthy controls [15, 16, 17, 20, 23, 29, 30]. Figure 2 shows a graphical representation of IL‐18 serum levels in study subpopulations of the included studies.

Fig. 1.

Flow‐chart of the inclusion and exclusion procedure.

Table 1.

Studies investigating serum IL‐18 in hemophagocytic lymphohistiocytosis (HLH) or macrophage activation syndrome (MAS)

Author, year	Relevant aim	Study population	IL‐18 measurement	HLH/MAS patients (pg/ml)	Non‐HLH/MAS patients (pg/ml)	Healthy controls (HC) (pg/ml)
De Jesus 2019 [23]	MAS versus undifferen‐tiated inflam‐matory diseases	5 pHLH	Bio‐Plex system (Bio‐Rad, Hercules, CA, USA)	Approx: pHLH: 70 000 (30 000–300 000)	Approx: 1000 (200–8000)	Approx: 300 (0–800)
		8 MAS	Values > 10 000pg/ml were rerun on: Luminex: BioPlex Pro group II cytokine standard	MAS: 80 000 (2000–400 000)
		21 inflammatory disorders
		5 HC
		*
Gao 2019 [16]	pHLH versus MAS	12 pHLH	Luminex 200 instrument (eBioscence, EPX340‐12167‐901)	pHLH: 463	Not included	Undetectable except 7: 31 (± 408)
		20 MAS		(±597)
		20 HC		MAS: 1248
				(±1319)
Jinkawa 2019 [28]	sHLH versus KD	5 KD‐HLH	MBL, Nagoya, Japan	1380 (1240–2000)	295 (60–1520)	Not included
		62 KD
Mazodier 2005 [17]	sHLH versus non‐HLH/healthy	20 inf.‐/ malign‐HLH	MBL, Nagoya, Japan	Median: 659	Median: 139	Median: 141
		27 inf.‐/ malign
		46 HC
Mizuta 2019 [18]	MAS versus sJIA	25 MAS	MBL, Nagoya, Japan	255 000 (53 000–830 000)	46 650 (3050–377 000)	Not included
		56 sJIA
		*
Shimizu 2018 [24]	sHLH versus MAS	20 MAS, 15 EBV‐HLH	MBL, Nagoya, Japan	EBV‐HLH: 4050 (1320–14 800)	Not included	Not included
				MAS: 183 500 (30 500–830 000)
Shimizu 2015 [25]	MAS versus sJIA	15 MAS	MBL, Nagoya, Japan	Approx: 100 000 in active phase and in MAS phase	Approx: 30 000 in active phase	Not included
		61 sJIA
Shimizu 2010 [20]	MAS versus HLH/KD	5 MAS (also measured in active and inactive sJIA phase)	MBL, Nagoya, Japan	EBV‐HLH: 3825 (1720–14 800)	Active sJIA: 130 000 (56 500–203 000)	141 (76–255)
		10 EBV‐HLH		MAS: 122 500 (101 000–830 000)	Inactive sJIA: 6025 (3730–12 000)
		22 KD			KD: 280 (180–560)
		28 HC
Takada 2004 [29]	IL‐18 versus IL‐16	2 pHLH	MBL, Nagoya, Japan	Approx: 6000 (5000–11 000)	Approx: 500 (0–1000)	Approx: < 10
		5 EBV‐HLH, 4 unknown HLH
		8 EBV
		8 HC
Takada 1999 [30]	IL‐18 in sHLH	2 pHLH	In‐house ELISA	Approx: 7000 (5000–11 000)	Approx: 700 (0–1200)	Approx: < 10
		5 EBV‐HLH, 4 unknown HLH
		8 EBV
		8 HC
Takakura 2019 [26]	MAS versus sJIA	21 MAS	MBL, Nagoya, Japan	**Cut‐off value for MAS versus sJIA: 69 250	**	**
		57 sJIA		Area under the ROC curve: 0.89
Wada 2013 [31]	EBV‐HLH versus EBV	4 pHLH	MBL, Nagoya, Japan	Approx: EBV‐HLH and pHLH: 3500	Approx: 1500	Not included
		11 EBV‐HLH
		11 EBV
Weiss 2018 [15]	sHLH/MAS	17 pHLH	Luminex: BioPlex Pro group II cytokine standard	*** Significantly elevated in MAS compared to HC, pHLH, infection related HLH, autoimmune diseases and genetic inflammatory disorders	***	***
	versus hyperferritine‐mic syndromes	22 MAS, 21 inf.‐HLH, 1 SLE‐HLH	Total and free IL‐18
		29 sJIA
		10 HC
Yasin 2020 [27]	History of MAS versus no history of MAS	2 MAS	MBL (Woburn, MA, USA)	MAS: (91 522–240 986)	Active sJIA: 16 499 (4816–61 839)	Not included
		38 sJIA (half with history of MAS)		sJIA with history of MAS: 13 380 (4212–62 628)	Inactive sJIA: 1164 (587–3444)
					sJIA without history of MAS: 957 (276–4263)

All values are reported in median with range or mean with standard deviation.

pHLH = primary hemophagocytic lymphohistiocytosis; sHLH = secondary HLH; AOSD = adult‐onset Still’s disease; KD = Kawasaki’s disease; inf. = infectious disease; Malign = malignant disorders; sJIA = systemic juvenile idiopathic arthritis; EBV = Epstein–Barr virus; SLE = systemic lupus erythematosus; Approx = approximately (value from figure/graph); ROC = receiver operating characteristic.

^*There are some unexplained discrepancies between the n‐values stated in the text and the datapoints in the figures/tables. The n‐values stated in this table are from the text.

^**Takakura et al. [26] only report their results as a receiver operating characteristics (ROC) curve analysis.

^***Weiss et al. [15] only report their results as mean differences.

Fig. 2.

Interleukin (IL)‐18 serum levels in different patient groups reported by included studies. Displayed are (1) healthy individuals, (2) patients with infectious, inflammatory or malignant disease without hemophagocytic lymphohistiocytosis/macrophage activation syndrome (HLH/MAS), (3) systemic juvenile idiopathic arthritis (sJIA) without MAS, (4) MAS and (5) HLH (pHLH or sHLH other than MAS). The y‐axis shows IL‐18 serum levels in logarithmic scale. Median with interquartile ranges are shown. Numbers on the graph refer to the included studies’ reference numbers. Takakura et al. [26] and Weiss et al. [15] could not be displayed on the graph, as these authors did not report IL‐18 serum levels separately for their included patient groups. All studies except those marked with hachure used the same commercial immunoassay (MBL, Nagoya, Japan).

IL‐18 in HLH/MAS diagnosis

Seven studies compared IL‐18 levels between patients with HLH or MAS and healthy subjects [15, 16, 17, 20, 23, 29, 32]. They found approximately 3–900‐fold higher serum IL‐18 in HLH/MAS patients than in healthy individuals.

Four studies compared serum IL‐18 between sJIA/AOSD patients with and without MAS [15, 18, 25, 26]. Of these, three studies reported significantly higher serum IL‐18 in patients with MAS than patients without MAS [15, 18, 25], while Takakura et al. calculated an area under a receiver operator characteristics (ROC) curve of 0·89 to distinguish sJIA patients with MAS from sJIA patients without MAS [26].

Seven studies compared serum IL‐18 between HLH patients and patients with inflammatory conditions other than sJIA/AOSD; these were Kawasaki’s disease [20, 28], EBV infection [29, 30, 31] and unspecified inflammatory, infectious or malignant disorders [17, 23]. In all studies, IL‐18 was significantly increased in HLH/MAS compared to other inflammatory disorders [17, 20, 23, 28, 29, 30, 31], with IL‐18 levels generally above 1000 pg/ml in HLH (Fig. 2).

Five studies compared IL‐18 levels in different subtypes of HLH/MAS [15, 16, 20, 23, 24]. Gao et al. found IL‐18 significantly increased in MAS compared to pHLH [16], while de Jesus et al. found no significant difference between these two groups [23]. Two studies by Shimizu et al. investigated IL‐18 levels in sJIA‐MAS compared to EBV‐HLH, and found significantly higher IL‐18 in MAS than in HLH [20, 24]. Furthermore, IL‐18 was decreased in EBV‐HLH compared to both active and inactive phases of sJIA [20].

Weiss et al. compared groups of sJIA, MAS, infection‐associated HLH and primary HLH, and found IL‐18 extremely elevated in patients with MAS compared with the other groups [15]. They found a cut‐off value of > 24 000 pg/ml to distinguish MAS from pHLH with a sensitivity of 83% and a specificity of 94%. In a different cohort they assessed samples from genetically and clinically defined inflammasomopathies. They found that patients referred for active or previous MAS had chronically elevated total IL‐18, generally > 40 times normal values. In this cohort a free IL‐18 cut‐off value of > 11 600 pg/ml distinguished the diseases MAS, sJIA and AOSD from all other tested samples, with 88% sensitivity and 93% specificity [15].

IL‐18 in HLH/MAS disease activity monitoring

Two studies followed patients longitudinally throughout disease development [15, 20].

Shimizu et al. serially measured IL‐18 in five cases of sJIA and described a pronounced increase in IL‐18 levels at the development of MAS[20]. IL‐18 correlated positively with other indicators of disease activity, such as ferritin, lactate dehydrogenase (LDH), aspartate aminotransferase and C‐reactive protein; however, after the other indicators of inflammation had normalized, IL‐18 remained highly elevated. Correspondingly, Weiss et al. followed three MAS patients longitudinally and found elevated IL‐18 with only partial improvement months after normalization of other disease markers [15].

IL‐18 and HLH/MAS development

Shimizu et al. stratified their sJIA patients into two subsets; a group with a high serum IL‐18/serum IL‐6 ratio (> 1000) and a group with low IL‐18/IL‐6 ratio (> 1000) [25]. In total, 15 of the 43 patients with high IL‐18/IL‐6 ratio and none of the 33 patients with low IL‐18/IL‐6 ratio developed MAS, and all MAS patients had absolute IL‐18 > 30 000 pg/ml. Based on these results, the authors compared IL‐18 levels between sJIA patients who did and did not later develop MAS. They found IL‐18 levels to be increased significantly in patients who developed MAS and reported a sensitivity of 87% and a specificity of 71% to predict MAS development at a cut‐off value of 47 750 pg/ml.

Mizuta et al. used IL‐18 as a disease marker together with ferritin, IL‐6 and others while looking into serum C‐X‐C motif chemokine 9 (CXCL9) [18]. Their measurement of IL‐18 showed that it elevated even before the development of MAS. They divided their patients into a group receiving tocilizumab and a group not receiving tocilizumab and found IL‐18 to be significantly lower in both sJIA active and MAS phases when receiving therapy [18].

IL‐18 in recurrent MAS

Three studies investigated patients with a history of recurrent MAS, and found that IL‐18 was significantly elevated in sJIA patients with history of MAS compared with sJIA with no history of MAS, regardless of whether or not the patients were in the active or inactive phase of sJIA [15, 26, 27].

Discussion

The present review illustrates that high serum IL‐18 is strongly associated with HLH. This was most pronounced in MAS, where IL‐18 levels > 100 000 pg/ml were generally reported [15, 18, 20, 23, 24, 25, 26, 27]. One exception was Gao et al. [16]; however, this study used a different IL‐18 assay to the other papers investigating MAS.

IL‐18 appears to discriminate well between patients with HLH and patients with infectious, inflammatory and malignant disease other than sJIA. Thus, IL‐18 may be useful in sHLH diagnosis. Furthermore, two studies found significantly higher IL‐18 in MAS compared with pHLH [15, 16]. This indicates that IL‐18 measurement could aid in directing suspicion towards pHLH and focus upon genetic testing. However, IL‐18 levels may differ between pHLH, depending on the genetic mutation in question. For example, patients with NLR‐family CARD domain‐containing protein 4 (NLRC4) mutation, which is associated with recurrent MAS, were found to have high IL‐18 levels even after clinical remission [15]. The rarity of the different mutations associated with HLH/MAS challenges statistical analysis of the difference of IL‐18 levels between subgroups. Regarding sJIA/AOSD, four studies found significantly increased IL‐18 levels (> 100 000 pg/ml) in sJIA patients with MAS compared with patients without MAS [15, 18, 25, 26]. However, another study by Shimizu et al. reported median serum IL‐18 of 130 000 pg/ml in active sJIA without MAS [20]. This could not be explained by assay differences. It may indicate a limited specificity of IL‐18 for MAS in sJIA patients, or perhaps reflects variation between different study populations and between laboratories, which challenges the establishment of inter‐institutional diagnostic cut‐off limits. Furthermore, IL‐18 appears to remain elevated after clinical remission of MAS, limiting its diagnostic value in sJIA patients with a previous history of MAS [20, 27].

IL‐18 may not only be useful as a biomarker, but also as a treatment target in HLH. In one study, knock‐out mice with abrogated production of mature IL‐18 lived longer with HLH than their wild‐type peers [33]. Chiossone et al. found that CMV‐induced HLH mice treated with recombinant (r)IL‐18‐binding protein (IL‐18BP) exhibited decreased interferon‐γ production and less severe organ damage compared with non‐treated mice, although all animals eventually succumbed to the disease [34]. In line with this, Girard‐Guyonvarc’h et al. detected more severe presentation of HLH in IL‐18BP‐deficient mice while, additionally, HLH severity was attenuated by IL‐18 inhibitor treatment [35]. A recent case report by Yasin et al. described a 14‐month‐old boy with recurrent MAS who received experimental treatment with rIL18‐BP (tadekinig alfa) and experienced clinical improvement [36], and rIL‐18BP has also been investigated in AOSD patients with an acceptable safety profile [37]. Patients with NLRC4 have also been demonstrated to benefit from IL‐18 blockade [15], and the results of an ongoing trial including patients with NLRC4‐mutation‐associated MAS are anticipated (https://clinicaltrials.gov/ct2/show/NCT03113760). IL‐18 measurements may be able to identify patients with potential benefit of rIL‐18BP treatment and possibly guide treatment.

Very recently, HLH has gained increased focus, as studies are reporting elevated serum ferritin [38, 39] and proinflammatory cytokine levels [38, 39, 40] in patients with coronavirus disease 2019 (COVID‐19). Elevated ferritin and/or cytokine levels were associated with more severe disease and higher mortality in these patients [38, 39, 40]. In a recent communication, Mehta and colleagues argue that patients with severe COVID‐19 should be screened for hyperinflammation and possible HLH [41]. In this light, improved laboratory diagnosis for HLH, as well as new treatment options, may be more relevant than ever. Two recently published studies investigated IL‐18 in COVID‐19 patients [42, 43]. A small study by Fraser et al. found that serum IL‐18 was approximately twice as high in COVID‐19‐positive patients in the intensive care unit (ICU) than in matched, COVID‐19‐negative ICU patients (n = 10 in each group) [42], while Wilson et al. found no difference between COVID‐19‐positive ICU patients (n = 15) and ICU patients with acute respiratory distress syndrome from other causes [43]. To the best of our knowledge, associations between serum IL‐18 and HLH have not yet been investigated in COVID‐19 patients.

The present review provides a comprehensive overview of the available literature on the association between IL‐18 and HLH/MAS, including both children and adults with any subtype of HLH and any underlying condition. Nine of the 14 studies used commercial assays from the same manufacturer to measure IL‐18 which, to some extent, enabled comparison across studies. All studies except one [23] used either the 2004 diagnostic criteria for HLH or similar criteria that were deemed equal to the 2004 criteria. However, only 14 studies were identified on the subject, and all suffered from some limitations. Generally, studies were small, which reflects that HLH and MAS are rare conditions. The majority of studies were explorative, and patient selection as well as patient flow and timing of blood sampling related to MAS/HLH diagnosis were not outlined adequately, leading to a potential risk of bias. Even with the use of the same commercial assay across a majority of studies, IL‐18 serum levels varied widely among studies, and there were pronounced differences between assays, with Luminex bead assays generally measuring lower than enzyme‐linked immunosorbent assays (ELISAs). This may reflect ethnic and other differences between study populations, but also underlines the need for interlaboratory standardization if diagnostic cut‐off values are to be implemented in international guidelines. Unfortunately, not all studies reported clearly whether free IL‐18 or IL‐18/IL‐18‐binding protein complex was measured, and some studies reported both, which further impedes comparison between studies.

In conclusion, IL‐18 is emerging as a promising biomarker to aid in diagnosing HLH and MAS. Extremely elevated IL‐18 appears specific for the disease groups MAS and sJIA and may aid in predicting development of MAS in sJIA. However, before IL‐18 can be implemented in routine diagnostics, interassay and interlaboratory variation should be addressed, and current results should be replicated in larger, well‐designed multi‐center studies.

Disclosures

The authors have no conflicts of interest to disclose.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.