Severe Pediatric COVID-19 Pneumonia Treated With Adjuvant Anakinra Free

We describe children admitted to a tertiary pediatric hospital for primary COVID-19 pneumonia (positive severe acute respiratory syndrome coronavirus 2 [SARS-CoV-2] polymerase chain reaction [PCR] test result) who received intravenous (IV) anakinra after rheumatology consultation (July 2020–April 2021). This retrospective review was approved by the institutional review board.

Pediatric acute respiratory distress syndrome (PARDS) was defined as occurring within 7 days of COVID-19 infection if chest imaging was consistent with acute pulmonary parenchymal disease and an oxygen saturation: fraction of inspired oxygen (SpO₂:FiO₂) ratio ≤264 for noninvasive ventilation or oxygenation index ≥4 for invasive ventilation.¹²  The 2004-HLH criteria, H-Score, and cHIS (Table 1) were calculated retrospectively from PICU admission.^5–7,13 

A 14-year-old girl with a known combined immunodeficiency presented with fever, cough, and respiratory distress. Her past medical history included nuclear factor-κ-Β-2 haploinsufficiency with recurrent infections, autoimmunity, asthma, and liver fibrosis. Her chest radiograph revealed multifocal bilateral pneumonia; computed tomography (CT) of her chest and abdomen additionally revealed significant hepatosplenomegaly (Fig 1 A–C). Her initial SARS-CoV-2 antibody test results were negative. She was admitted to the PICU within 24 hours. She met PARDS criteria with maximum respiratory support of continuous positive airway pressure. She met 2004-HLH and cHIS criteria, but not H-Score cut-off (Table 2).

Within 24 hours, she was started on IV anakinra (10 mg/kg per day divided every 8 hours) for 26 days (Fig 1 E and F). Anakinra was briefly held (2 days) after a concern for infection was found to be cytomegalovirus reactivation and restarted because of increasing ferritin. She had additional therapy, including dexamethasone, a shortened 3-day course of remdesivir because of transaminitis, convalescent plasma, enoxaparin, and treatment of CMV. CMV reactivation also contributed to her hyperferritinemia as levels declined once anakinra and antiviral therapy were initiated. The discharge chest radiograph revealed significant improvement (Fig 1D), and home oxygen was not required. Her hospitalization length of stay (LOS) was 30 days. A repeat CT chest scan 4 months later revealed the resolution of abnormalities.

A 7-year-old boy with sickle cell disease (Hemoglobin SS) presented with headache, altered sensorium, seizures, right frontoparietal subarachnoid hemorrhage, posterior reversible encephalopathy syndrome, and hypoxic respiratory failure requiring intubation. On hospital day 1, he underwent exchange transfusion (hemoglobin S percentage decreased from 74% to 12%). Despite exchange transfusion, he had worsening C-reactive protein (CRP) and right lung space opacities (Fig 2 A–C). SARS-CoV-2 antibody test results were negative. He met PARDS criteria. He met cHIS criteria, but he did not meet 2004-HLH diagnostic criteria or H-Score cut-off (Table 2).

On hospital day 2, he was started on IV anakinra (10 mg/kg per day divided every 12 hours). After clinical improvement, the total daily dose was weaned over a 12-day course (Fig 2 E and F). Additional therapies included dexamethasone, remdesivir, antiepileptics, and empirical antibiotics. His discharge chest radiograph revealed significant improvement (Fig 2D). His hospitalization LOS was 16 days. He has since remained healthy with sickle cell maintenance therapy.

A 2-year-old girl with high-risk neuroblastoma admitted for consolidation therapy developed rapidly progressive hypoxic respiratory failure requiring intubation. She was status post near-total resection, chemotherapy, radiation, and autologous stem cell rescue with small residual disease, and was cycle 3 of consolidation therapy.¹⁴  After PICU admission, the chest radiograph and CT chest angiography revealed bilateral pneumonia with a right pleural effusion (Fig 3 A–C). The echocardiogram revealed moderate-severely depressed right ventricular systolic function with qualitatively normal left ventricular systolic function. SARS-CoV-2 antibody test results were initially negative. She met PARDS, 2004 HLH, H-Score cut-off, and all 6 cHIS criteria (Table 2).

Within 24 hours of PICU admission, she was started on IV anakinra (Fig 3 E and F). She remained on anakinra (5 mg/kg per day once daily) for 17 days with a corresponding decrease in ferritin but continued CRP elevation. On hospital day 23, she developed multiple pulmonary hypertensive crises requiring veno-arterial extracorporeal membrane oxygenation. Echocardiograms revealed severe pulmonary hypertension and biventricular systolic dysfunction (left ventricle ejection fraction of 37% by bullet method). Anakinra was discontinued given uncertain benefit after the resolution of hyperferritinemia. Additional therapies included dexamethasone, remdesivir, convalescent plasma, antibiotics, inotropes, inhaled nitric oxide, iloprost, milrinone, and bivalidrudin. She was decannulated on hospital day 36 but died on day 39 from cardiopulmonary failure after the development of a tension pneumothorax, bilateral pneumothoraces, and pulmonary hemorrhage (Fig 3D).

A 16-year-old girl with type I diabetes mellitus presented to an outside facility with 2 days of fever, cough, dyspnea, and altered sensorium. She was transferred to our PICU after a seizure-like episode and hypoxic respiratory failure requiring intubation. A chest radiograph revealed multifocal pneumonia (Fig 4A). SARS-CoV-2 PCR, antinucleocapsid IgG, and spike IgM test results were positive. Her mother was concurrently ill with COVID-19. She met PARDS criteria. She required positive end-expiratory pressure, inhaled nitric oxide, and prone positioning because of hypoxia. Her lung disease, poor compliance, and high ventilatory requirements led to air-leak, resulting in pneumomediastinum, subcutaneous emphysema, and a left pneumothorax. A brain MRI was unremarkable. While intubated, continuous electroencephalography revealed background slowing but no seizures. Echocardiograms revealed normal systolic function without coronary abnormalities. She did not meet H-Score or 2004-HLH-diagnostic criteria; however, she did meet cHIS criteria (Table 2).

Within 12 hours of transfer, she was started on maximal IV anakinra (400 mg/day divided every 6 hours). After clinical improvement, anakinra was weaned over 10 days (Fig 4 C and D). She also received IV corticosteroids and anticoagulation. Remdesivir was contraindicated because of significant transaminitis, which normalized before discharge. On day 8 of admission, she was extubated and weaned to room air. Her last chest radiograph revealed significant improvement (Fig 4B). She was discharged after 17 days and has done well at follow-up.

We describe 4 children with severe COVID-19 pneumonia and hyperinflammation treated with anakinra. Two children met HLH-2004 criteria, but only the extracorporeal membrane oxygenation patient met H-Score cut-off. All patients met ≥5 cHIS criteria. Webb et al demonstrated that cHIS score of >2 distinguished patients with longer median LOS (9 vs 3 days), mechanical ventilation (69% vs 17%), and death (15% vs 1%).⁷  A cHIS score of ≥4 was associated with mechanical ventilation (92%) and death (80%).⁷  Within our small cases series, 3 patients (75%) were intubated and 1 died (25%). Although this case series is descriptive in nature, future large-scale pediatric studies could validate cHIS criteria and compare results with adult outcomes.

Our patients tolerated anakinra with corresponding decreased ferritin. Although case reports reveal anakinra may rarely lead to hepatic injury, we did not observe this adverse effect, and the hepatic injury in Case 4 normalized after anakinra initiation. The surviving patients had significant clinical, laboratory, and radiographic improvement. CMV treatment likely contributed to ferritin improvement in Case 1. Targeted immunomodulation against IL-1 may improve COVID-19 hyperinflammation without widespread immune suppression. In contrast, high-dose corticosteroids can hinder T and B cell production and worsen infections or viral reactivation.¹⁵  Anakinra is used as a first-line treatment of macrophage activation syndrome and secondary HLH at the doses described in this report.^16,17  Within adult COVID-19 anakinra clinical trials, the dose and duration vary by study. The SAVE-MORE trial revealed that patients with elevated soluble urokinase plasminogen activator receptor, an early proinflammatory biomarker, had decreased mortality (hazard ratio = 0.45) with anakinra dosed at 100 mg daily for 7 to 10 days.⁹  Our case series occurred before Food and Drug Administration Emergency Use Authorization for tocilizumab or baricitinib for pediatric COVID-19 patients. Anakinra is offered at our center for COVID-19 patients with hyperinflammation on a case-by-case basis.

In children with severe COVID-19, clinicians may need to dampen an aberrant host immune response. Additional pediatric clinical experiences and outcomes of COVID-19 immunomodulation need to be reported and prospectively studied. Pediatric hospitalists and subspecialists should collaborate to identify children with COVID-19 hyperinflammation who may benefit from early initiation of immunomodulation.