Hemophagocytic lymphohistiocytosis in COVID-19

Supplemental Digital Content is available in the text

Even though all data have been anonymized, the surviving patient gave their consent for publication.
suggested as a treatment option, [3,4] and first reports are promising. [1] In this context, more data on clinical management of sHLH triggered by COVID-19 are urgently expected

Patients
We reviewed COVID-19 patients admitted to an intensive care unit in Vienna, Austria between April and May 2020, who were diagnosed with sHLH. Patients' clinical-, imaging-, and laboratory data (see Supplements, http://links.lww.com/MD/F969 and the supplemental figure, http://links.lww.com/MD/F968) were assessed.

SARS-CoV-2 diagnosis
Testing for the presence of severe acute respiratory syndrome coronavirus 2 RNA in pharyngeal or tracheal respiratory specimens was performed by Real-Time qPCR. Positive results (Ct value >35) were confirmed by repeated testing.

sHLH diagnosis
sHLH was diagnosed using the HScore: [5] Nine variables are assessed: core temperature, hepato-and/or splenomegaly, number of cytopenias, levels of TG, fibrinogen, ferritin and ASAT, history of immunosuppression, and (if feasible) presence of bone marrow haemophagocytosis. A positive result yields a 93% sensitivity and 86% specificity for HLH.

Immunosuppressive therapy
Immunosuppressive treatment for sHLH was conducted in a stepwise approach:

Patient consent and ethical review
All data have been anonymized. Informed consent for publication of anonymized data from the patient or their relatives have been obtained. Ethical review was not necessary for case reports following local respective guidelines.

Patient 1
A 51-year old male (BMI 26.2) with a fever for 6 days was hospitalized due to respiratory failure and tested positive for SARS-CoV-2. Showing acute respiratory distress syndrome (ARDS), he was intubated and mechanically ventilated. Acute kidney injury (AKI) necessitated continuous renal replacement therapy (CRRT), upgraded with an immunoadsorption filter (day 5 of hospitalization) against cytokine storm. Despite noradrenaline support, the hemodynamic profile deteriorated. Dobutamine was added due to heart failure with reduced ejection fraction and impaired left ventricular function. While levosimendan, argipressin, and landiolol led to a transient clinical improvement, hemodynamics further worsened. On the 21st day of hospitalization, sHLH was diagnosed (Table 1), and immunosuppressive therapy was started with methylprednisolone for 72 hours, followed by Pentaglobin (see Methods). After 26 days of ICU Table 1 Patients' initial sHLH diagnosis details including Hscore (5) points and the subsequent course of sHLH from the day of diagnosis onwards, monitored through the Hscore.

Patient 2
Delirium and dyspnea for 4 days were reported by a 75-year old man (BMI 29.4) before he was hospitalized and tested positive for SARS-CoV-2. ARDS led to intubation, mechanical ventilation, and intermittent prone positioning. CRRT with immunoadsorption due to AKI and cytokine storm were established (3rd day of hospitalization). Deteriorating heart failure with reduced ejection fraction and intermittent noncompensatory tachycardia necessitated a treatment regimen of noradrenaline, dobutamine, agripressin, landiolol, and levosimendan, leading to a sustainable hemodynamic improvement. On the 19th day of hospitalization, sHLH was diagnosed (Table 1). Intravenous methylprednisolone was started, but before escalating the immunosuppressive therapy, fulminant pulmonary embolism occurred, not responding to systemic thrombolysis, and resulting in a fatal outcome on the 23rd day.

Patient 3
A 74-year old woman (BMI 19.4) was tested positive for SARS-CoV-2. Aggravating dyspnea and deliriumnot manageable by noninvasive ventilationnecessitated intubation due to ARDS. Cytokine storm and AKI were present; CRRT with immunoadsorption was initiated on the 18th day. Mild noradrenaline support was necessary to sustain a stabile hemodynamics. sHLH was diagnosed on the 16th day of hospitalization (Table 1). Steps 1 to 3 of the described immunosuppressive regimen were administered (see Methods) without serious adverse events. Markers of cytokine storm and sHLH regressed, and CRRT and hemodynamic support could be stopped. After 35 days of ICU-care and negative tests for SARS-CoV-2, the patient entered a successful weaning process and was still alive at a follow-up on day 85.

A stepwise treatment-approach
All our patients showed the full sHLH characteristics. [2,4,5] As discussed before, [2] we applied and suggest a three-step approach to sHLH caused by COVID-19 (see Fig. 1 and Methods): initial immune attenuation through high-dose pulsed methylprednisolone, followed by a bodyweight-adapted dose of immunoglobulins, and lastly anakinra until clinical improvement. This approach bears the benefits of a cheap and easy-to-obtain substance as the initial line of attack, followed by the stronger immunomodulatory agents. In parallel, we recommend immunoadsorption for cytokine storm dampening, especially in CRRT. We believe that sHLH diagnosis came too late to reverse outcomes for patients 1 and 2, whereas our approach led to a favourable outcome in patient 3 -similar as reported by Dimopoulus and colleagues. [1]

Early diagnosis is key
The risk of diagnosing sHLH too late must be weighed against the potential side effects of aggressive immunosuppression, [2] such as coagulopathy after immunoglobulin application. However, we believe that routine sHLH-screening of COVID-19 patients and early goal-directed therapy can lead to improved outcomes. sHLH should especially be considered in patients deteriorating fast without sufficient response to shock management. As the magnitude of cytokine levels may not correlate with sHLH severity and since specific markers (e.g. soluble-IL-2-receptor) are often not available, trends in ferritin could be used in treatment response tracking and outcome prognostication. [2,7] Special attention should be paid to a possible rebound after treatment discontinuation. [2] First reports of sHLH treatment in COVID-19 are encouraging, [1] and results of a prospective COVID-19 sHLH cohort are expected for late 2020 (Clinical Trials-ID: NCT04347460). However, further interventional trials are needed to confirm our assumptions.

Conclusion
Routine screening for sHLH in COVID-19 using the HScore appears reasonable; patients without sufficient response to shock management might be at particular risk. A stepwise therapeutic approach comprising corticosteroids, immunoglobulins, and anakinra, accompanied by immunoadsorption, may dampen cytokine storm effects and reduce mortality.