1 INTRODUCTION

Norovirus is a common cause of acute infectious gastroenteritis that is often self-limited in immunocompetent hosts but can cause debilitating symptoms in immunocompromised patients.¹ In children with transplants, this can lead to vomiting, diarrhea, weight loss, and electrolyte imbalances. Higher viral loads in patients receiving immunosuppression to prevent organ rejection or graft-vs-host disease are associated with symptoms, and virus shedding may continue for a prolonged period of time, regardless of symptoms.² There are no approved drugs for the treatment of norovirus gastroenteritis, though reports suggest possible improvement with nitazoxanide,^{3, 4} oral immunoglobulin,⁵ and reduction in or change of immunosuppression.^{4, 6}

The commensal gastrointestinal microbiome may influence the pathogenesis of norovirus infection.⁷ Depletion of commensal bacteria in mice has been shown to reduce the replication of certain enteric viruses including murine norovirus.^{8, 9} Our group anecdotally observed symptomatic improvement in immunocompromised patients with norovirus infection who received metronidazole, an antibiotic that is often used to modulate the gut microbiome, and this drug was then intermittently used to treat additional immunocompromised patients with refractory norovirus enteritis. In this report, we review our experience, which suggests a possible role for metronidazole in alleviating symptomatic norovirus infection in this population.

2 PATIENTS AND METHODS

We retrospectively reviewed charts from all patients admitted to a pediatric tertiary-care hospital in Chicago, IL from July 2014 to March 2019 with a stool test that was positive for norovirus. We extracted data from patients who had previously received either a solid organ or hematopoietic stem cell transplantation and excluded those who only had a positive test before transplantation. Norovirus testing at our hospital over the time period investigated is an RT-PCR test sent to a CLIA-certified commercial laboratory. Data abstracted included demographics (age, sex, ethnicity, and race), medical history (organ transplant, symptoms), laboratories (WBC, ANC, ALC, BUN, Creatinine, CMV, EBV, Adenovirus, Clostridioides difficile PCR, stool culture, blood culture), immunosuppression and concomitant antimicrobial therapy in a systematic fashion. In general, testing was recommended when patients were symptomatic. Progress notes were reviewed for documentation of symptom improvement within 1 week of treatment initiation; if symptom improvement was noted the treatment was considered to have improved the episode. If symptom improvement was not documented or if symptoms were reported to not change or to worsen, then the treatment was considered to have not improved the episode. The decision to treat and the agent used was determined by the clinicians caring for the patient, though no patients received concurrent (combination) treatment with nitazoxanide, metronidazole, or oral immune globulin. The recommended duration of treatment with antimicrobials was generally 14 days, though treating clinicians had discretion on whether to discontinue treatment sooner.

Data analysis was done using RStudio (www.rstudio.com/). Continuous variables were expressed as median and interquartile range. Statistical comparisons for categorical outcomes were made with Fisher's exact test and for continuous outcomes were made with the Kruskal–Wallis test with a p-value < .05 considered statistically significant.

3 RESULTS

We identified 41 transplant recipients with positive norovirus testing. Three patients had a positive test only prior to transplantation and were excluded. One patient had a positive test 82 days before transplantation and had a subsequent positive result 254 days after transplantation. For this patient, only the information associated with the test following transplantation was used in our analysis.

For the 38 patients included in the analysis, the median age of transplantation was 23 months (range 1–128 months), varying widely by the type of transplantation (Table 1). There were 109 norovirus tests ordered with 85 positive results in the 38 patients. The median time from transplant to first positive test was 281 days, which also varied widely based on the type of transplantation, with hematopoietic stem cell transplant recipients having the shortest time between transplant and a positive norovirus test while renal transplant recipients had the longest duration (Table 1).

The majority of the 85 episodes of positive norovirus tests were associated with symptoms (Table 2). As expected, diarrhea was the most common symptom (nearly 75%), and almost 50% of episodes were associated with vomiting. There was no documentation of GI symptoms in 19% of episodes.

Nearly half of the positive norovirus tests were not treated and only 41 of the 85 episodes were treated. The majority of these were treated with nitazoxanide (Table 3). Metronidazole was used in only 8 episodes; 6 of these patients (75%) had documentation of symptom improvement compared with 15 of those treated with nitazoxanide (60%). There was not a statistically significant difference in the improvement of symptoms with nitazoxanide when compared to metronidazole (p = .12). The mean duration of treatment for those that improved with nitazoxanide was 13.6 days compared with 15.2 days for those that did not improve. For metronidazole, the mean duration of treatment for patients that improved and failed to improve was 16.3 and 13.5 days, respectively. Oral immunoglobulin was given to 2 patients during this period with 1 patient having improvement in symptoms after 3 doses and the second patient having no documentation of improvement after 8 doses.

Of the eight norovirus episodes in which metronidazole was used, two were in patients who had only a single positive test for norovirus. One of these patients was a stem cell transplant recipient who improved with metronidazole therapy. The second was a stem cell transplant recipient that did not have documentation of symptom improvement but had no subsequent positive tests for norovirus.

We additionally compared outcomes of treatment in patients with chronic/recurrent norovirus infection, i.e., those with more than 1 positive test who were treated with either nitazoxanide or metronidazole. We identified 6 patients who had a total of 16 episodes of norovirus gastroenteritis, with characteristics and treatment as shown in Table 4. Included in Table 4 are 2 additional patients treated solely with metronidazole for comparison. Of these 16 episodes, 10 (62.5%) were treated with nitazoxanide and the rest with metronidazole. Improvement in symptoms was seen in 3 (30%) of the nitazoxanide-treated episodes whereas five of six episodes (83%) treated with metronidazole improved. There was not a statistically significant difference in response to metronidazole as compared with nitazoxanide (p = .12).

We evaluated coinfections in the norovirus episodes, which could confound observations of treatment effect. Among the 85 positive norovirus tests, 12 were in the setting of coinfection with another pathogen in the stool, involving 11 patients. Nine had coinfection with adenovirus, one with Clostridioides difficile, and one with both adenovirus and Clostridioides difficile. Only one patient with coinfection was treated with metronidazole. This patient had concurrent Clostridioides difficile infection but did not improve with treatment. None of the patients with adenovirus identified in the stool had adenovirus viremia. None of the episodes of norovirus were associated with CMV in the stool, but 1 patient had concomitant low-level CMV viremia (<137 IU/ml).

4 DISCUSSION

No interventions have proven efficacy for the treatment of chronic norovirus infection in immunocompromised hosts. Reports involving limited numbers of patients have suggested benefits with nitazoxanide,^{3, 4} oral immunoglobulin,⁵ and reduction or change in immunosuppression,^{4, 6} with modest effects demonstrated. In our cohort of transplant recipients with norovirus gastroenteritis, metronidazole was associated with clinical improvement in some patients. The rate of improvement for episodes treated with metronidazole was comparable to those treated with nitazoxanide.

The intestinal microbiome can influence the pathogenesis of enteric virus infection.⁸ Mice given antibiotics to deplete intestinal microbiota are less susceptible to mortality after enteric but not intraperitoneal challenge with poliovirus, with lower viral replication during the infection. This effect was reversible after intestinal recolonization with fecal bacteria. Depletion of enteric bacteria with antibiotics similarly diminishes persistent murine norovirus infection,¹⁰ an effect which is also reversed by reconstitution of normal enteric flora.

There are some limitations to our study. The retrospective nature of the study only allowed us to note clinical changes documented in the chart, which sometimes lacked clear information on symptomatic improvement. Treatment was not randomized, and most patients receiving metronidazole had previously been treated with other interventions, usually nitazoxanide. Finally, medication administration was abstracted through chart review, but it is possible that some treatment was prescribed without EHR documentation.

Our data suggest possible benefits from metronidazole for norovirus gastroenteritis in some transplant recipients. Larger studies comparing nitazoxanide, oral Ig, and/or metronidazole treatment of norovirus infection are needed to improve the management of this difficult-to-treat infection.

AUTHOR CONTRIBUTION

Soneji and Muller involved in the concept/design. Soneji, Newman, and Toia collected the data. Soneji, Newman, Toia, and Muller involved in data analysis/interpretation. Soneji and Muller drafted the article. Soneji, Newman, Toia, and Muller critically revised the article. Soneji, Newman, Toia, and Muller approved the article.