Just in time: mal-HLH in Sweden, 1997 to 2018

In this issue of Blood, Löfstedt and colleagues from the Karolinska Institute present an important benchmark for malignancy-associated hemophagocytic lymphohistiocytosis (mal-HLH), providing the demographics and outcomes of adults with this rare, often fatal disease of uncontrolled inflammation.¹ 

The investigators created a data set from all 6 health care regions in the country by combining the Swedish Patient Registry (in-patient registry) with both the Swedish Cancer Registry and the Swedish Cause of Death Registry from 1997 to 2018. This permitted them to evaluate all suspected cases of mal-HLH from the entire Swedish population during a 22-year period. The average population per year was 9.3 million individuals with 7.4 million adults (>17 years old).

Using both direct chart review and diagnostic coding analysis, as well as cross-checking with Swedish-based publications on HLH from 1997 to 2023, they applied both the HLH-94/2004 and HScore criteria (using soluble interleukin-2 receptor [sIL-2R] in units per milliliter) to the dataset and accounted for the differences in source data from 1997 to 2011 compared with 2012 to 2018. All patients were treated at their regional centers. From this extraction, 458 patients were identified, and 142 were eliminated as they had histiocytosis without an associated malignancy. The remaining 316 patients (9 children, 307 adults) were subdivided based on year of diagnosis. Chart review of the 1997 to 2011 cohort yielded 62 patients, 13% of whom had been misdiagnosed. This cohort was divided into 2 groups based on the timing of HLH onset: those with onset during chemotherapy, where infection or toxicity caused HLH (Ch-HLH, n = 20), and those with malignancy occurring at the same time as HLH (M-HLH, n = 15). In the second time span, 2012 to 2018, 146 patients were identified based on a strict definition of “probable mal-HLH,” that is, HLH diagnosed within 90 days prior to the diagnosis of malignancy or up to 1 year after the diagnosis of malignancy. The remaining patients were not analyzed either due to missing data (n = 22) or not meeting complete criteria for mal-HLH (n = 82). The investigators observed that normal spleen size was a common cause of <5 points in HLH-94/2004 scores. For 316 patients over the entire time span, age, sex, subtype of malignancy, population-based HLH incidence, and survivorship were analyzed.

The study yielded highly useful data on significant differences in mal-HLH regarding sex and age, types of mal-HLH-associated malignancies, lymphoma survivorship curves, as well as specific histologies underlying M-HLH (eg, natural killer [NK]/T 52%) versus Ch-HLH (eg, NK/T 20%). The 2012 to 2018 cohort illustrates the predominance of younger male subjects in the mal-HLH cohorts, most striking in the age range 20 to 50 years, compared with all hematologic malignancies in the registries from 2012 to 2018 (see Figure 4A-B in the article by Löfstedt and colleagues).

The authors found that the annual incidence of HLH incidence rose over 10-fold during the course of the study: from 0.026 per 100 000 population (1997-2007) to 0.45 per 100 000 population and 0.62 per 100 000 adult population (2014-2018), with a 3.7-fold difference between the lowest to the highest health care regions. As HLH-2004 scoring had minimal impact, they attributed a substantial proportion of the increase to a burst of educational activities and publication of Swedish-based HLH research in 2010 to 2011.

Limitations of the study included the necessary use of standard criteria. The investigators observed that this can yield an impaired positive predictive value for HLH in patients with leukemia or lymphoma as disease characteristics may mimic diagnostic criteria for HLH. In addition, the ethnicity was predominantly Nordic, and all were managed within a well-established national health care system. Genetic testing results for HLH predisposition, specific chemotherapy data, time to diagnosis of HLH and to the underlying cause, and lengths of inpatient care were not available. Interesting populations (ie, HLH diagnosis outside of the 1.25-year time frame, HLH-94 score < 5, histiocytosis with HLH) were not studied.

This 1997 to 2018 study also reported an unsurprising but very sobering finding. Although 1-month survivorship in adults improved with time, 2-year survivorship was nearly identical at, in the investigators’ own words, a “dismal” 25% (see figure; also see Table 2 in the article by Löfstedt and colleagues). Their findings are comparable to the most recent mal-HLH data set from 2010 to 2020: approximately 25% of patients with very high ferritin and sIL-2R levels survived 500 days from diagnosis.² More recently, the COVID-19 pandemic, expanded and earlier use of immune cell-based therapeutics, and T-cell-activating therapies have impacted the incidence of and interest in a range of cytokine release syndromes, including HLH. Immunotherapies are emerging treatments for T-cell and NK/T lymphomas.³ Concurrently, although nonsteroidal anticytokine HLH therapeutics have become increasingly available, their mostly off-label status and exorbitant cost have impeded use in adult HLH clinical care.

Age and significant biological variation in underlying mal-HLH etiologies result in diagnostic criteria with diminished specificity in adults. Although pediatric HLH is benefiting from the introduction of highly specific, rapid diagnostic testing, comparable diagnostic testing methods have lagged in adults. In addition, just as sIL-2R may have dual use as a diagnostic and prognostic indicator in mal-HLH,² data set merging and comparison are impacted by sIL-2R being increasingly measured by enzyme-linked immunosorbent assay (measured in picograms per milliliter) rather than by the functional assay (measured in units per milliliter), with no known conversion factor.

The time has come: how can this study’s foundational findings and nuanced discussion help us in a new era of both new lymphoma therapies and potentially increased risks for HLH? The list is long. Bending this “dismal” overall survival (OS) curve will require both a deep understanding of why fulminant paraneoplastic hyperinflammation results in such consistently poor OS⁴ and highly effective T and NK/T lymphoma therapies.³ At the bedside and in the clinical laboratory, we will need early detection outside of specialty centers, with particular recognition of the impact of sex, age, and treatment-induced HLH. Such recognition will require training in nuances of infectious triggers and improved clinical judgment and care^5,6; rapid turnaround times for essential biomarkers and microbial studies; use of less cytotoxic, nonsteroidal anti-cytokine therapies initially to minimize etoposide and corticosteroid intensities; accelerated diagnosis of the underlying HLH etiology; and early HLH multifactorial risk stratification. Research efforts and resource utilization will benefit from the increase in disease recognition and larger datasets. The Swedish interventions in 2010 to 2011 model possibilities for all countries to optimize regional and national HLH education, clinical trial participation, and data sharing.

Conflict-of-interest disclosure: The author declares no competing financial interests.