21^st Century Storm Chasers: Defining Macrophage Activation Syndrome

Catastrophic storms share several cardinal features, including heavy precipitation, damaging winds, and rapid progression. But no two storms are the same. Meteorologists rely on differences in severity (tropical depressions versus category 5 hurricanes), character (tornados versus monsoons), and etiology (nor’easters versus typhoons). Such granular characterization has been necessary to protect lives and property, and achieved through generations of iterative observation and technological innovation. A global network of meteorologists now combine dozens of parameters, collected continuously and synthesized by complex prediction algorithms, to keep ahead of storms … with variable success.

In clinical practice, we often struggle for a sense of the character, severity, and etiology of a patient’s “hyperinflammatory storm.” Because the distinctions between inflammatory syndromes vastly affect proper treatment, we cannot afford to simply batten down the hatches on every patient. A nuanced response requires a nuanced understanding. Given the myriad interacting host and environmental factors that result in systemic inflammation, our task could not be greater. Unfortunately, our biomarkers and prediction algorithms remain very primitive, and our model systems are deeply flawed (1).

In this issue of Arthritis & Rheumatology, the investigative team led by Ravelli, Cron, and colleagues (with the sponsorship of the American College of Rheumatology and the European League Against Rheumatism and the remarkable combined support of PRINTO, CARRA, PRCSG¹, and the Histiocyte Society) continue their campaign to define Macrophage Activation Syndrome (MAS). Colloquially, MAS is an acute systemic inflammatory syndrome characterized by unremitting fever, cytopenias, high ferritin, coagulopathy, transaminitis, high triglycerides, hemophagocytosis², and potential for progression to multi-organ dysfunction, shock, and death (2). Canonically, MAS is a serious complication of systemic juvenile idiopathic arthritis (sJIA) or adult-onset Still’s disease (AOSD). In practice, however, the features of MAS are thoroughly enmeshed with those of sepsis, various lymphoproliferative and malignant disorders, other rheumatic conditions (like Kawasaki disease), and several monogenic syndromes. In the absence of arthritis or other distinctly rheumatic features, other specialists call this syndrome hemophagocytic lymphohistiocytosis (HLH).

In the present article, the authors have restricted their campaign to defining MAS as a complication of known or suspected sJIA. In 2005, Ravelli et al. issued preliminary diagnostic guidelines based on retrospective comparison of MAS complicating sJIA versus active sJIA (3). This process highlighted the relative diagnostic importance of hemorrhage, CNS dysfunction, thrombocytopenia, leukopenia, transaminitis, and hypofibrinogenemia. In 2011, they pursued a different approach, initiating a process to develop expert consensus in the definition of sJIA-MAS(4). This consensus survey corroborated prior parameters, but also emphasized the dynamics of certain measures: dropping leukocyte or platelet count, dropping ESR. In 2014, this group collected and organized hundreds of clinical scenarios representing sJIA-MAS, active sJIA, or comparator disease states from an international pool of rheumatologists and heme-oncologists. Using submitting physician diagnosis as the standard, they found that the MAS-05 criteria outperformed the HLH-04 criteria (5, 6). They also examined the characteristics of sJIA-MAS in these clinical scenarios and found that platelet count, transaminases, ferritin, LDH, triglycerides, and d-dimer demonstrated the most robust changes between pre-MAS and MAS (7). The group subsequently looked at the heterogeneity of presentation and treatment in these scenarios, and found a similarly severe spectrum of presentation independent of geography or submitting specialist, although treatment practices were widely disparate (8).

Now, this group has combined approaches (insert current reference). They gathered an international panel of experts at an MAS-classification consensus conference. To establish a new “gold standard”, expert opinions on whether a given clinical scenario represented MAS (or not) were solicited and consensus reached after a few rounds of voting. When a consensus opinion was reached for a given scenario, it became the new standard. They then evaluated multiple iterations and thresholds of MAS-related lab and clinical values, comparing the agreement between a certain set of classification criteria and the expert consensus. This generated several potential classification criteria, which were debated in two independent groups of experts. Both groups agreed on the same set of criteria as their preference. In contrast to the MAS-05 criteria, which were developed to aid diagnosis, these MAS-15 criteria were expressly intended to guide the classification, primarily for the purposes of clinical research (for example, adjudicating MAS as an outcome or side effect, or including patients in MAS trials). This effort represents nearly a decade of dedicated collaboration, consensus-building, and complex statistical parsing of expert opinion and lab value performance. The diversity of contributing experts and the sophistication of the analysis used to create these criteria bolster their validity and will hopefully generate enthusiasm for the conduct of clinical trials in sJIA-MAS.

Nonetheless, a few caveats must be considered.

Circularity:

The HLH-04 and MAS-05 criteria reflected and reinforced practitioners’ opinions about MAS. Unfortunately, no major changes in the workup or management of MAS were implemented between MAS-05 and the collection of patient scenarios in 2014 that would have significantly altered these opinions. Other biomarkers of MAS were proposed (serum CD25, CD163, IL-18, neopterin, NK cytotoxic function, hematopoiesis gene signature), but none have been widely adopted or had their utility validated. By and large, we are still using old tests (blood counts, transaminases, triglycerides, ferritin, fibrinogen, histology) to work up MAS. The physician diagnoses and expert opinions used in this study to classify MAS suffer from an inherent circularity: they are biased towards confirming prior definitions of MAS. This likely affected which clinical scenarios physicians chose to submit for evaluation, as well as the ultimate diagnosis (MAS or not-MAS) that the submitting physician or reviewing expert associated with each scenario. With no diagnostic gold standard, such circularity is essentially unavoidable. We anticipate that the present effort will provide an escape from this loop. A uniform set of classification criteria will facilitate the studying MAS prospectively, and enable different groups to compare results. We can expect that these studies will identify new and improved biomarkers of the hyperinflammatory MAS state, or potentially different markers for different subsets of MAS. These findings will ultimately force changes in the MAS-15 classification criteria; and this is exactly the progress that we need.

Dynamic Observation:

Owing to the nature of the clinical scenarios submitted, these classification criteria fail to capture the dynamics of the parameters measured. In particular, serial data were unavailable for control patients, and so changes could not be employed in the MAS vs. non-MAS derivation strategy. However, as recognized by the authors, clinicians often use changes in laboratory markers as an important clinical clue (for example, falling platelet count or ESR, rising ferritin) – indeed, this is a likely explanation for patients classified by referring clinicians as MAS but not acknowledged as such by the MAS expert panel. Many of the MAS-15 laboratory value cutoffs fall within/near normal ranges, implying a departure or change from an otherwise expected lab. We can look forward to further work from this group and others to characterize how changes over time can help inform MAS diagnosis over and above these “snapshot” criteria.

Limited Context:

In this report, the experts have precisely classified MAS complicating known or highly-suspected sJIA. This is of necessity, because there is no clear distinction between the MAS-like physiologic states that can occur in viral, bacterial, and fungal infections, other rheumatic diseases, hematologic malignancies, and in patients with a seemingly disparate array of monogenic defects. In many cases, MAS occurs in the context of multiple potential contributors. It will be interesting to assess how the MAS-15 criteria perform in these different contexts. Hopefully, the intensive study of sJIA-MAS afforded by these criteria will inform future lumping/splitting exercises.

How does the work of Ravelli et al. inform the understanding and management of MAS? For one, these data remind us that MAS is a life-threatening pathologic state, and our lack of specific biomarkers and treatments represent a dangerous unmet medical need. Next, this work enables clinicians to assign relative value to certain findings: 1) CNS dysfunction heralds severe disease; 2) investigating hemophagocytosis is frequently insensitive, nonspecific, and/or impractical; 3) hyperferritinemia, thrombocytopenia, transaminitis, and coagulopathy consistently stand out as critical parameters. Finally, this campaign demonstrates that only with the utmost organization, collaboration, and analysis can we begin to make sense of observations made in a maelstrom.

In understanding MAS more broadly, there is reason for optimism. Heightened awareness has added an MAS-like flavor to exuberant T-cell anti-tumor responses, systemic reactions to hemorrhagic viruses, and subsets of sepsis (9–11). Likewise, advances in technology have enabled us to attack the problem of hyperinflammatory pathogenesis from multiple angles. Next-generation sequencing has expanded the genetic underpinnings of sJIA, MAS, and HLH (12–14) and opened unanticipated lines of investigation. Massively multiplex analysis of serum/plasma biomarkers and gene transcription may uncover meaningful patterns that will be useful for bedside decision-making. Finally, a more precise definition of MAS will facilitate trials demonstrating a response (or lack thereof) to novel, targeted therapies. Capitalizing on these assets will require diligent support, ingenuity, and collaboration, such as demonstrated by Ravelli, Cron, and colleagues.

In the 19^th century, ambitious meteorologists combined the latest science (barometric pressure, the Coriolis effect) with cutting-edge technology (e.g. electrical telegraphy), integrating this into a global collaboration that has saved countless lives and spurred further innovation. This study’s classification criteria for MAS complicating sJIA are intended to facilitate future studies that will, in turn, improve our understanding of sJIA and MAS. Hopefully, improved diagnostic and treatment strategies await, and with them some pathophysiologic heft to the terms sJIA and MAS. Better yet, perhaps “sJIA” and “MAS”, like Farmer’s Almanacs, weathervanes, and red skies at night, will become quaint reminders of hard-earned progress.