Persistent COVID-19 Pneumonia in a Patient on Rituximab

Abstract

Immunosuppressed patients have increased morbidity and mortality due to SARS-CoV-2 infection. They can have reduced protective antibodies levels or reduced cellular immunity resulting in inability to clear the virus and persistent viremia and virus-induced inflammation. We present the case of a 52-year-old male, immunosuppressed due to use of rituximab for treatment of chronic inflammatory demyelinating polyneuropathy, who developed pneumonitis from SARS-CoV-2 infection. The patient required hospitalization 4 times over a 3-month period due to recurring pneumonitis symptoms, including fever, dyspnea, and a severe cough. The final diagnosis was delayed due to negative SARS-CoV-2 polymerase chain reaction tests on 7/8 of nasopharyngeal specimens, as well as failure to clinically improve with administration of remdesivir after the sole positive test. He was thought to have rituximab-induced organizing pneumonia, but his condition gradually worsened with additional immunosuppression directed against that condition. His active SARS-CoV-2 infection was eventually confirmed from a bronchoalveolar lavage specimen on the final hospitalization, and he was then treated with a multimodal regimen that included an antiviral agent (nirmatrelvir/ritonavir), a Janus Kinase inhibitor (baricitinib), intravenous immunoglobulin, and intravenous methylprednisolone. He had a prompt and sustained clinical response to this multimodal regimen. Persistent COVID-19 should be considered in the differential diagnosis in patients with unexplained organizing pneumonia, and can be treated effectively with multimodal therapy as above.

Introduction

Immunosuppressed patients have considerably worse clinical outcomes associated with SARS-CoV-2 infection than immunocompetent patients. ¹ They often do not respond normally to SARS-CoV-2 vaccines, with reduced SARS-CoV-2 cellular (T cell) and humoral (antibody-mediated) immunity compared to immunocompetent individuals.^2-4 Some are unable to clear their SARS-CoV-2 infections, allowing replication-competent virus to persist in the lung parenchyma for months, even when asymptomatic. ⁵ This prolonged viral shedding increases the risk of relapse of COVID-19 symptoms in immunosuppressed patients as well as the transmission of SARS-CoV-2 to their contacts. Rituximab, an anti-CD20 chimeric monoclonal antibody, is particularly associated with this through depletion of preplasma B cells, leading to hypogammaglobulinemia. Patients who have received rituximab and then acquire SARS-CoV-2 are unable to produce anti-SARS-CoV-2 antibodies, impeding clearance of the virus. ⁶

Here, we present a case of prolonged COVID-19 in a gentleman on rituximab whose ultimate diagnosis was delayed due to multiple negative SARS-CoV-2 nasopharyngeal swabs along with radiographic and clinical features that appeared to be consistent with organizing pneumonia.

Case Presentation

The patient is a 52-year-old male with a past medical history that includes diabetes mellitus, hypertension, congestive heart failure, pericarditis, latent tuberculosis (treated 20 years ago), calcified lung nodules, and recurrent deep vein thrombosis. He also has chronic inflammatory demyelinating polyneuropathy managed with off-label rituximab every 3 weeks. He initially presented to an outside hospital (OSH) with a 3-week history of fever, nonproductive cough, exertional shortness of breath, chest pain, and fatigue. Notably, his last rituximab infusion was 4 months prior. He also had a documented SARS-CoV-2 infection (from which he had a complete clinical recovery) 7 months prior to his initial presentation to the OSH.

Computed tomography (CT) angiogram of the chest demonstrated multifocal patchy and nodular opacities concerning for an infectious process. He was started on ceftriaxone and azithromycin for a presumptive diagnosis of community-acquired pneumonia; however, due to persistent fevers after 5 days, his antibiotic regimen was escalated to vancomycin and cefepime. Two nasal COVID-19 polymerase chain reaction (PCR) tests were negative. His symptoms improved, and he was discharged on a course of doxycycline and amoxicillin/clavulanate.

Three weeks following his discharge from the OSH, the patient presented to our emergency department with persistent fever, cough, exertional shortness of breath, chest tightness, generalized malaise, fatigue, and weight loss. A repeat chest CT scan showed a moderate pericardial effusion and diffuse scattered pulmonary opacities. These opacities were noted to be more diffuse and prevalent than the previously observed abnormalities.

Given his worsening clinical presentation and chest CT scan findings, a broad diagnostic workup was initiated to exclude bacterial, mycobacterial, viral, and autoimmune processes. Also included in the differential diagnosis were inflammatory processes such as drug-induced or cryptogenic organizing pneumonia. Bronchoscopy revealed semi-erythematous mucosa, non-purulent secretions, and clear return noted from instilled saline. Organizing pneumonia was suspected, potentially due to rituximab, and empiric therapy with intravenous (IV) methylprednisolone was started. However, after it was noted that his bronchoalveolar lavage (BAL) PCR was positive for Ureaplasma urealyticum, IV methylprednisolone was discontinued, and he was treated with doxycycline, with clinical improvement 1 day later. He was discharged from the hospital.

Ten days later, he presented for outpatient follow-up complaining of a resurgence of symptoms, including fever, fatigue, night sweats, and cough. These had worsened since his hospital discharge despite taking doxycycline as prescribed, suggesting that organizing pneumonia was the true culprit. He was restarted on a tapering dose of oral steroids.

Ten days after this, he presented to the emergency department for worsening fever, relative hypotension, tachycardia, and a cough. The patient reported that he had felt “tremendously better” during the first 3 days of taking prednisone, with no fever or cough and improved breathing. However, his symptoms had since recurred with fevers becoming higher, and he also had a new periorbital headache. Upon readmission, a chest CT scan again showed scattered multifocal pulmonary opacities with curvilinear peripheral appearance, raising concern for an infectious or inflammatory process. Organizing pneumonia remained high on the differential. A follow-up upper respiratory pathogen panel test was positive for SARS-CoV-2. The patient was treated with a 5-day course of remdesivir for his SARS-CoV-2 infection, and prednisone 60 mg daily was continued for presumptive organizing pneumonia.

The patient was readmitted to the hospital for the fourth time due to a 10-day history of fevers. He reported that his symptoms had initially improved after being discharged on prednisone 60 mg daily, with a decrease in fever and night sweats, and resolution of his cough; however, these symptoms returned after the prednisone dose was tapered. Bronchoscopy cultures detected Pseudomonas aeruginosa (10 000 colony forming units/milliliter), which was interpreted as likely colonization, especially since his BAL showed no purulent secretions. Despite this, a 5-day course of ceftazidime was initiated since the prior cultures were negative. The consensus recommendation was to continue evaluation for a noninfectious etiology of his symptoms.

Intravenous methylprednisolone 60 mg daily was continued during this admission. He did not require oxygen, but he developed worsening fevers. Although his nasal COVID-19 PCR test was negative, his BAL specimen was positive with a cycle threshold value of 26, which is consistent with an active infection. The infectious disease service recommended a 10-day course of nirmatrelvir/ritonavir (Paxlovid) and baricitinib 4 mg daily for 14 days. His serum SARS-CoV-2 anti-nucleocapsid immunoglobulin G (IgG) was undetectable, indicating that he had not mounted an immune response to the virus. Therefore, he also received 1 dose of intravenous immunoglobulin at a dosage of 0.4 mg/kg, which was administered as an adjunctive and preventive therapy for COVID-19 (Please see Figure 1 for the timeline of the patient’s persistent pneumonia).

Figure 1.

Timeline of persistent pneumonia. The first hospitalization is set as week zero, indicated as “w0.” Last COVID-19 diagnosis was 7 months prior (“–7m”) and last rituximab was 4 months prior. SARS-CoV-2 testing is noted above the timeline with blue as negative and red as positive. Hospitalizations are marked on the timeline in yellow with interventions as rectangles beneath. w18 was the last follow-up for this patient.

Approximately 1 month after his last discharge, the patient followed up in the ID clinic and he reported feeling much better, with complete resolution of his fevers and cough.

Discussion

Rituximab has residual effects on the immune system for 6 to 9 months. Its anti-CD20 effect results in loss of B cells and of humoral (antibody-mediated) immunity, putting patients at risk for viral infections that rely primarily on antibody-mediated protection. Early during the COVID pandemic, rituximab was identified as a risk factor for severe COVID-19 with a physician-reported registry of 3729 patients in Europe with rheumatologic disease showing a fourfold risk of death in those receiving rituximab compared with methotrexate monotherapy. ⁷ A study of vaccine response to 2 doses of the mRNA COVID-19 vaccine for patients on rituximab compared with healthy controls and those on other immunosuppression showed a markedly impaired B-cell response but preserved T-cell response. ⁸ In that study, only 26% of patients on rituximab developed neutralizing antibodies to COVID compared with a 92% response rate in healthy controls, and 80% in those receiving other immunosuppression. Strikingly, none of the patients who received rituximab during the 6 months prior to immunization developed neutralizing antibodies to SARS-CoV-2. Our patient had no detectable anti-SARS-CoV-2 nucleocapsid IgG, consistent with a failure to mount an antibody response to SARS-CoV-2 infection.

The patient had a very delayed diagnosis, in large part because of his many negative SARS-CoV-2 PCR tests. In all, 9 SARS-CoV-2 PCR tests were collected (8 from nasopharyngeal swabs and 1 from BAL) but only 2 of these had a positive result (1 nasopharyngeal, 1 BAL). While use of a nasopharyngeal swab is the standard procedure for diagnosing SARS-CoV-2 infection, PCR from a BAL is the gold standard for diagnosing pneumonia due to SARS-CoV-2 (ie, COVID-19 pneumonia). In 1 series comparing PCR results from nasopharyngeal and BAL specimens in 123 patients intubated for respiratory failure suspected due to COVID-19, the nasopharyngeal assay had 88.6% sensitivity. ⁹ While the positive predictive value was 93.3%, importantly, the negative predictive value was only 81.3%. This was underscored again in a meta-analysis of 17 studies reporting BAL SARS-CoV-2 PCR results in hospital-based case series of suspected viral pneumonia with negative nasopharyngeal or oropharyngeal COVID PCR results, where 11% of BAL specimens yielded a positive COVID PCR. ¹⁰

The hallmarks of severe COVID-19 include the development of pulmonary ground glass opacities and other fibrotic changes characteristic of interstitial lung disease. COVID-19 is now a well-recognized cause of secondary organizing pneumonia (SOP), which more commonly occurs in older patients with underlying pulmonary disease. ¹¹ One case series in particular highlights the occurrence of SOP in patients on anti-CD20 therapies such as rituximab, many of whom continued to have pulmonary disease well over a year after the initial infection. ¹² In this case series, persistent fever was common, and BAL specimens were positive for SARS-CoV-2 in over half. While the typical severity of COVID-19 has lessened with increased population-wide vaccination and prior exposure to the pathogen, clinicians should maintain a high degree of suspicion for COVID-19 in any case of organizing pneumonia, given high SARS-CoV-2 prevalence. This is especially important in patients on anti-CD20 therapies.

The patient was initially treated with a 5-day course of remdesivir, which failed to cure his COVID-19. Interestingly, the original remdesivir placebo-controlled trial to show efficacy used 10-day courses. ¹³ A subsequent nonplacebo-controlled trial showed noninferiority of 5 days versus 10 and was used to justify the recommendation for the 5-day course, which was additionally desirable due to remdesivir supply constraints early during the pandemic. ¹⁴ Later studies noted occasional failure of 5-day remdesivir courses, particularly in immunocompromised patients, with some accompanied by SARS-CoV-2 nsp12 gene mutations conferring reduced susceptibility to remdesivir. ¹⁵ A particular concern was the 22- to 31-hour half-life of GS-443902 (the active metabolite of remdesivir), suggestive that a longer course might better achieve a steady state that would be more clinically significant in the setting of reduced-susceptibility isolates of SARS-CoV-2. ¹⁶ Although the 2 treatment durations have not been directly compared in immunocompromised patients, the above reasons support the occasional use of 10-day courses of remdesivir.

However, even 10-day courses of remdesivir have been known to fail in immunocompromised patients with prolonged COVID-19. Such patients are often concurrently treated with corticosteroids to reduce pulmonary inflammation in COVID-19, but this further immunocompromizes patients, inhibiting their ability to clear actively replicating virus. In such patients, an extended course of the antiviral nirmatrelvir/ritonavir has been used to halt viral replication and achieve clinical and virologic cure. ¹⁷ While the original nirmatrelvir/ritonavir protocol used 20 days (and showed persistent virus detectable at 10 days), subsequent studies showed 10 days to be sufficient for clinical cure in most patients with persistent COVID-19 who were on anti-CD20 therapy such as rituximab. ¹⁸ For this reason, we administered nirmatrelvir/ritonavir for the entirety of the hospitalization with a goal to maximize duration to achieve as close as possible to 20 days. Outpatient use of nirmatrelvir/ritonavir is typically limited to 5 days due to not being approved for longer durations; thus, it was critical to administer at least 5 days inpatient to reach a minimum total duration of 10 days.

Intravenous immunoglobulin (IV-Ig) was used as an adjunctive antiviral in this patient who had no detectable anti-SARS-CoV-2 antibodies. This was used in place of convalescent plasma given that there is now high baseline population immunity to SARS-CoV-2, resulting in current-year pooled IV-Ig outperforming older convalescent plasma against currently circulating SARS-CoV-2 strains. ¹⁹ IV-Ig is used routinely to prevent infections in patients with agammaglobulinemia. One study found that IV-Ig lots from late 2022 and 2023 were up to 30 times more effective at preventing COVID-19 in these patients (n = 981) as compared to lots from earlier timepoints that had much lower anti-Omicron strain neutralizing antibodies. ²⁰ IV-Ig has been used previously to successfully treat COVID-19 in patients like ours who lack humoral immunity to SARS-CoV-2, resulting in improved clinical outcomes in 2 small cohort trials (a total of 25 patients).^21,22 Finally, baricitinib was added for its immunomodulatory effect to help reduce pulmonary inflammation, along with its antiviral effect as an endocytosis inhibitor that reduces viral entry into the cell. ²³ The use of combined therapy rather than a single agent in immunocompromised patients has shown the potential to impede persistence of SARS-CoV-2 and improve clinical outcomes.^24,25 Collectively, the combination of nirmatrelvir/ritonavir, baricitinib, steroids, and IV-Ig helped this patient achieve clinical and virologic cure.

Conclusion

Rituximab use has a higher risk in the COVID-19 era than it did previously, given the high contagiousness and prevalence of SARS-CoV-2 in combination with rituximab-induced elevated risk for both severe and prolonged COVID-19. Patients using rituximab should be aggressive about self-testing for COVID-19 and should promptly start treatment for this diagnosis as warranted. In some patients, a nasal swab may miss active viral replication in the deep pulmonary parenchyma. In this case, discordant nasal and BAL COVID-19 test results led to delayed administration of curative antivirals, which quickly resulted in clinical improvement. Nirmatrelvir/ritonavir dosed for 10-20 days is a good option for immunocompromised patients who develop persistent COVID-19. Adjunctive therapies such as IV-Ig and baricitinib should also be considered, since it has been noted that immunosuppressed patients benefit significantly from use of a multimodal regimen, rather than a single medication for treatment of SARS-CoV-2 infection.

Persistent COVID-19 results in ongoing pulmonary inflammation that shares many clinical and radiographic features with organizing pneumonia. Given the ubiquitous presence of SARS-CoV-2, patients who receive diagnostic BALs in preparation for possible high-dose steroids should have those samples tested for SARS-CoV-2, and be treated if found to be present at moderate or high levels.