de intensivist

Abstract
We report here a patient with neuromyelitis optica, on a maintenance therapy of rituximab and mycophenolate mofetil, with severe coronavirus disease 2019 (COVID-19) who was admitted to the ICU in April 2020. Because of the immunocompromised status, viral clearance was not achieved. Treatment with intravenous immunoglobulins was initiated to stimulate the adaptive immune response and neutralise neuromyelitis optica activity, after halting mycophenolate mofetil. Subsequently, the patient developed a hyperinflammatory response which resulted in a rapid deterioration of her respiratory status with refractory severe hypoxia (oxygen saturation <75%). The patient was treated with methylprednisolone and anakinra, an interleukin-1 (IL-1) antagonist, which resulted in a dramatic improvement in the pulmonary disease and inflammatory status, and subsequent full recovery.

Introduction
Coronavirus disease 2019 (COVID-19) is caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and rapidly evolved into a global pandemic. Presentation may vary but typically includes fever and respiratory symptoms which may develop into acute respiratory distress syndrome (ARDS).[1] Severe COVID-19 has been characterised by a hyperinflammatory response, which may worsen ARDS and increase mortality.[2] Because of its immunomodulating effects, intravenous immunoglobulin (IVIG) therapy has been proposed as a possible treatment for severe COVID-19. In critically ill COVID-19 patients, a possible beneficial effect of IVIG therapy on mortality has been described based on retrospective cohort studies;[3-7] however, a phase 3 double-blinded randomised controlled trial (RCT) showed no benefit of IVIG therapy on clinical outcome after 28 days, and serious adverse events were more frequently reported in the IVIG group, although not statistically significant.[8] The role of IVIG therapy in immunocompromised COVID-19 patients remains unclear. Immunomodulating effects of IVIG could strengthen anti-viral defence and attenuate hyperinflammation in immunocompromised patients.[9] Furthermore, IVIG therapy has been shown to resolve antibody deficiencies and stimulate the adaptive immune response in immunocompromised patients.[9] We report a case where IVIG therapy was employed as rescue therapy in an immunocompromised patient with severe COVID-19, however with an adverse impact on the clinical course.

Case
A 45-year-old Caucasian woman was admitted with fever to the Erasmus Medical Center Rotterdam in April 2020. Her medical history included obesity (body mass index 38 kg/m²) and anti-aquaporin 4 (AQP4) positive neuromyelitis optica. The neuromyelitis optica was being treated with oral mycophenolate mofetil 720 mg twice a day, and rituximab 1000 mg once a year, administered 11 months before presentation. The patient was admitted with complaints of fever, diarrhoea, and hypoesthesia and paraesthesia of both legs, associated with an neuromyelitis optica-related lesion in the cervical myelum, without dyspnoea. Polymerase chain reaction (PCR) assay on an oropharyngeal swab was positive for SARS-CoV-2. A CT scan showed extensive infiltrative changes in the lungs compatible with severe COVID-19, without pulmonary emboli. She received supportive care with oxygen and was treated with ceftriaxone and azithromycin. Administration of mycophenolate mofetil was continued to prevent exacerbation of the neuromyelitis optica with possible permanent neurological damage as a consequence, given the known cervical lesion. Four days after admission, her respiratory function deteriorated and she was admitted to the ICU where she was intubated and ventilated in the prone position. The azithromycin was switched to ciprofloxacin upon admission to the ICU. Because of persistent fever and elevation of the inflammatory markers on day four and five of ICU admission (figure 1), a bronchoalveolar lavage was performed. Besides SARS-CoV-2 no other pathogens were identified and the antibiotics were halted. Her clinical situation remained unchanged; however, eight days after ICU admission a new CT scan showed new areas of ground glass opacification in areas of the lung that were not previously affected. It was concluded that the progression of the pulmonary disease could be caused by the inability of the patient to clear the virus under mycophenolate mofetil therapy and rituximab. Therefore, the mycophenolate mofetil was discontinued. The B cell count was 0.02 × 10⁹/l (reference 0.1-0.5 × 10⁹/l). In accordance, the immunoglobulin G was significantly decreased (1.3 g/l, reference 7.0-16.0 g/l) indicating a persisting immunodeficiency after rituximab. The leukocyte count was normal (8.0 × 10⁹/l) with decreased lymphocytes (7.2%, reference 15-50%). To treat her immune deficiency and stimulate humoral immunity, it was decided to treat the patient with IVIG 0.4 g/kg/day for three days. Furthermore it was reasoned that the IVIG could have a beneficial effect on neuromyelitis optica activity, as a preferential therapy after discontinuation of the anti-autoimmune therapy. However, on the second day of IVIG therapy, she developed a high fever (39.6 °C) and her respiratory status deteriorated acutely, with arterial oxygen saturation of 88% and arterial pO₂ of 6.9 kPa despite pressure-controlled ventilation with 100% FiO₂ and a PEEP level of 16 cm H₂O. Chest X-ray showed worsening of the extensive diffuse bilateral consolidations with almost complete white-out of both lungs. Ventilation in the prone position did not improve oxygenation, and oxygen saturation further deteriorated to the low seventies. As a last resort, the ventilation mode was switched to pressure control with an extreme inverse ratio, which resulted in acceptable arterial oxygenation. Ceftriaxone and ciprofloxacin were re-started to treat possible bacterial superinfection. Laboratory results showed a sudden increase in interleukin-6 (IL-6) levels from 79 to 1232 pg/ml, while procalcitonin levels remained stable, which raised the suspicion of cytokine storm syndrome or immune reconstitution syndrome as a cause of the worsening of her respiratory status The patient was treated with 1000 mg methylprednisolone followed by anakinra 300 mg a day for three days to counter the hyperinflammatory deterioration. One day after initiation of this therapy, her temperature normalised and the IL-6 level diminished to 13 pg/ml and two days later, her respiratory status improved and the ventilation mode could be switched to pressure support. Respiratory status improved further in the ensuing days and she was extubated successfully ten days after initiation of methylprednisolone therapy. PCR for SARS-CoV-2 did not test positive after13 days after initiation of IVIG therapy. The patient suffered from ICU-acquired weakness but could be discharged to the ward four days after extubation. After 14 days on the medical ward the patient was discharged from hospital to a rehabilitation centre. Six weeks after discharge she had recovered to her original physical condition.

Discussion
The COVID-19-pandemic has challenged the critical care community due to a lack of knowledge on its clinical course and the best treatments. Despite an abundance of COVID-19 cases worldwide, understanding the pathophysiological mechanisms in immunocompromised patients remains challenging and there is still a paucity of definite evidence-based treatments.

Anakinra is an IL-1 receptor antagonist which has been used in a broad spectrum of autoinflammatory syndromes. The effect of anakinra on inflammatory responses in COVID-19 has shown a reduction in mortality.[10] IL-1 serum levels were not measured in our patient, but IL-6 serum levels increased enormously at the time of clinical deterioration. Although IL-6 receptor antagonists such as tocilizumab have been proven to be effective, this evidence was not yet available at that time. Since treatment with anakinra has shown to result in a significant reduction in serum IL-6 levels,[11] anakinra was regarded to be potentially effective to counter the systemic hypercytokinaemic course.

IVIG is a blood product containing polyclonal immunoglobulin G from healthy donors and has numerous immunomodulatory effects. Although the mechanism of action remains unclear, IVIG therapy has been shown to elicit desirable effects in the treatment of severe infections, especially in immunocompromised patients.[12] Furthermore, IVIG exhibits preferential immunomodulatory effects in patients with autoimmune diseases by neutralising autoantibodies.[9]

Evidence on the effect of IVIG on mortality in COVID-19 has been limited and conflicting. A beneficial effect on mortality has been suggested based on several retrospective observational studies.[3-7] Data from non-blinded RCTs showed no difference in mortality in severe COVID-19, but studies were small in size and varied widely in their study protocol.[13,14] A small double-blinded placebo-controlled RCT of 59 patients showed reduced in-hospital mortality in patients with severe COVID-19 treated with IVIG.[15] However, a phase 3, multicentre, placebo-controlled RCT including 146 patients showed no effect of IVIG on mortality after 28 days, or on ventilator-free days after 28 days, whereas the number of adverse events in the IVIG group was (non-statistically) higher than in the placebo group.[8] Congruent with the case described here, adverse events consistent with hyperinflammation were described (e.g. acute kidney injury, septic shock and systemic inflammatory response syndrome). Patients in that study did not show an increase in IL-6 levels which, however, was first measured seven days after IVIG administration.[8]

Limited experience with IVIG in immunocompromised COVID-19 patients has been reported. Daneshpazhooh et al. reported a case of COVID-19 in a patient with mucous membrane pemphigoid on rituximab and mycophenolate mofetil, who was treated with IVIG on the day of hospital admission and improved clinically thereafter.[16] Vasconcelos et al. described a case of a patient on rituximab for rheumatoid arthritis with worsening COVID-19, who recovered clinically after IVIG therapy.[17] In both cases no adverse events were reported and both received IVIG before deterioration and did not need mechanical ventilation or ICU admission.

The effect of IVIG might be dependent on the timing of administration during the course of COVID-19, although this is only supported by retrospective data. A subgroup analysis of a retrospective cohort study of 535 patients showed better outcomes on mortality and need for mechanical ventilation in patients after early administration of IVIG (<7 days after ICU admission) than after late administration of IVIG.[7] Another retrospective study described 58 patients with severe COVID-19 who received IVIG (20 g/day) starting at the moment their absolute lymphocyte count fell to <0.5 x 10⁹/l. Analysis showed that IVIG was administered earlier in the survivor group than in the non-survivor group and furthermore that the group of patients who received IVIG within 48 hours had reduced ventilator use, reduced hospital and ICU length of stay and improved 28-day mortality compared with IVIG treatment initiated 48 hours after admission.[4] However, since timing of IVIG therapy was dependent on the lymphocyte count, this could rather reflect a different course of disease with a different onset of lymphopenia. The patient reported here received IVIG 15 days after hospital admission and 11 days after the start of mechanical ventilation. It is possible that the timing of IVIG therapy might have contributed to the adverse impact on clinical course in our patient.

Concerns have been raised that antibodies administered with IVIG therapy can trigger an exaggerated inflammatory response termed antibody-dependent enhancement in COVID-19 patients, such as in our patient.[18] In this phenomenon, subneutralising antibodies form immune complexes with the virus which enhances entry into host cells, resulting in increased viral uptake and replication, which leads to both increased infectivity and virulence. Immune reconstitution syndrome, mainly described in human immunodeficiency virus patients receiving antiretroviral therapy, but also in patients who discontinue immunosuppressive drugs, is a fulminant inflammatory response to microorganisms and pre-existing infection, which manifests when an abrupt shift of the host immune response occurs from immunosuppressed to proinflammatory.[19] We hypothesise that the symptoms observed in our patient might reflect a combination of antibody-dependent enhancement and immune reconstitution syndrome after IVIG therapy since it was only after high-dose steroids and anti-IL-1 therapy that the near-fatal ARDS resolved and the inflammatory markers decreased. Whether IVIG was beneficial for viral clearing but facilitated hyperinflammation remains unclear.

Since the emergence of COVID-19, treatments targeting attenuation of inflammatory response with immunomodulatory agents, such as corticosteroids and tocilizumab, have shown to improve outcome.[20,21] However, very few immunocompromised patients were included in these trials. Furthermore, the immunocompromised state is associated with prolonged viral shedding, and prior immunosuppressive therapy with mycophenolate mofetil or rituximab in COVID-19 patients is associated with worse outcomes.[22-24] Although compared with the use of corticosteroids and tocilizumab in the general population similar beneficial effects on inflammatory responses might be expected, it remains unclear whether augmenting immunomodulation in these patients further delays viral clearance or increases the risk for superinfections.

Given the reduced humoral response to infection in patients on mycophenolate mofetil or rituximab, therapy aimed at antibody-dependent clearance remains of interest and with the ongoing pandemic, specific therapies have emerged. Treatment of hospitalised COVID-19 patients with high-titre COVID-19 convalescent plasma (CCP) showed no beneficial effect when assessed in large randomised trials; however, few immunocompromised patients were included in these trials.[25,26] Anti-SARS-CoV-2 monoclonal antibodies (mAb) have shown potential beneficial effects on mortality and incidence of hospitalisation in non-hospitalised patients, although the effect is dependent on the SARS-CoV-2 variant of concern.[27] The role of high-titre CCP or anti-SARS-CoV-2 mAb therapy in immunocompromised critically ill patients with COVID-19, remains unclear.

In general, trials in COVID-19 have been performed in immunocompetent patients, and therefore specific evidence for treatment of immunocompromised patients is lacking. This case shows the adverse effects of IVIG in an immunocompromised patient, which were successfully treated with immunosuppressive therapy. Although there is an urgent need to find effective therapies in immunocompromised patients who remain at risk even after vaccination, therapeutic strategies should be carefully investigated before widespread implementation.

Disclosures
All authors declare no conflict of interest. No funding or financial support was received. Written informed consent was obtained from the patient for the publication of this case report (and the accompanying images).