Acute pancreatitis or severe increase in pancreatic enzyme levels following remdesivir administration in COVID-19 patients: an observational study

Remdesivir has been shown to reduce recovery time and mortality among patients with coronavirus disease 2019 (COVID-19). However, data regarding the efficacy and safety of remdesivir use are limited in Japan. We conducted a single-center retrospective cohort study at Yokohama Municipal Citizen’s Hospital, Kanagawa, Japan. Patients with COVID-19 pneumonia treated with remdesivir were included. The onset of acute pancreatitis and increased pancreatic enzyme levels and clinical, laboratory, treatment, and outcome data were collected and analyzed. A total of 201 patients were included. Among the 201 patients treated with remdesivir, 177 recovered from COVID-19. Increased pancreatic enzyme levels of grade 3 or higher or acute pancreatitis developed in 23 of the 201 patients. The potential etiopathogenetic effects of remdesivir on increased pancreatic enzyme levels of grade 3 or higher or acute pancreatitis were ascertained by reviewing the characteristics of patients hospitalized for COVID-19 who did not receive remdesivir treatment. Only 3 of 159 patients had increased pancreatic enzyme levels of grade 3 or higher during the treatment course. Multivariate analysis indicated remdesivir administration and severe COVID-19 infection by National Institute of Health standards as independent risk factors. Acute pancreatitis and severe increases in pancreatic enzyme levels were observed among patients with COVID-19 treated with remdesivir.

www.nature.com/scientificreports/ been using steroids to treat respiratory failure in COVID-19 patients because we lack the facilities to perform extracorporeal membrane oxygenation. Thus, most patients who were administered remdesivir were also administered steroids. All but 22 patients survived, and therapy was equally as efficacious or better than previously reported. However, unlike previous COVID-19 studies [4][5][6][7][8] , the frequency of increased levels of pancreatic enzymes and pancreatitis was 11.4% (23/201) in this study. The pancreatitis was mild and resolved after supportive therapy. However, pancreatitis is a serious adverse event that can be potentially life-threatening if left untreated. The aim of this study was to assess the association between remdesivir administration and elevated pancreatic enzymes in COVID-19 patients. A secondary aim was to report the efficacy and other adverse events observed in patients treated with remdesivir in a Japanese hospital setting.

Methods
Patients selection. This single-center retrospective observational study was conducted in the Department of Infectious Diseases, Yokohama Municipal Citizen's Hospital, Kanagawa, Japan. In total, 201 hospitalized adult patients with COVID-19 (aged ≥ 20 years) who were administered remdesivir during the time period from June 1, 2020, to January 31, 2021, were included in the study. As a control group for remdesivir administration, adult patients with COVID-19 who were hospitalized at our hospital but were not administered remdesivir in the time period between February 5, 2020, and January 31, 2021 were enrolled. Patients enrolled in other trials, those without laboratory results of pancreatic enzyme levels, those transferred to our hospital after administration of remdesivir at another hospital and did not receive remdesivir at our hospital, or those who contracted cluster infections at other hospitals during the onset of acute pancreatitis were excluded. Patients who refused to participate in the study were excluded. The data cutoff date was March 20, 2021.

Severity classification. The National Institutes of Health (NIH) Clinical Presentation of People with
SARS-CoV-2 Infection guidelines were used to classify the severity of COVID-19 9 .
Adverse events. Adverse events because of remdesivir administration were evaluated using the Common Terminology Criteria for Adverse Events version 5.0 (CTCAE ver 5.0).

Data collection.
This was a retrospective observational study, and, despite the frequency of blood collection being determined by the attending physician, blood samples were generally collected 2-3 times per week. Blood samples were collected for biochemical analysis, including amylase and lipase, blood counts, and coagulation. Adverse events of grade 1 or higher on the CTCAE ver 5.0 were collected as much as possible in the remdesivir group, and only necessary items were collected in the control group.
Statistical analysis. The association between a grade 3 or higher increase in pancreatic enzyme levels or acute pancreatitis and patient characteristics was analyzed using the chi-square test. Statistical significance was set at p ≤ 0.05. A logistic regression model was used to calculate the adjusted odds ratio (OR) with 95% confidence interval (CI) for the risk of increase in grade 3 or higher pancreatic enzyme levels or the onset of acute pancreatitis. We used multivariable logistic regression to control for potentially confounding roles of remdesivir administration, steroid administration, sex, age, severity classification, comorbidities (diabetes, hypertension, hyperlipidemia), and D-dimer level elevation. These analyses were performed using the Statistical Package for Social Science (SPSS) version 23 (IBM Corp., Armonk, NY). Missing data on diabetes were considered not examined, and at the time of analysis, analysis was performed with diabetes versus without diabetes or not examined. IRB approval. The study protocol was approved by the independent ethics committee of the Institutional Review Board of Yokohama Municipal Citizen's Hospital (No. 20-12-01). The study was a non-interventional retrospective observational study, and the ethics committee of the institutional review board approved the optout method of consent without written consent. Informed consent was obtained in the form of an opt-out notice displayed on the institutional website. Patients who refused to participate in the study were excluded. Statement of human rights. All procedures performed in this study involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Results
Patients characteristics. The baseline patient characteristics are shown in Table 1. Between June 2020 and January 2021, 201 patients who received remdesivir were enrolled in this study (Fig. 1). The median age was 73 (range: 22-98) years; 125 patients (61.2%) were male, and 194 patients were Japanese. Based on the COVID-19 severity classification developed by the NIH 9 , 2, 151, and 58 cases were moderately, severely, and critically ill, respectively. Among the participants, 81 were never-smokers; 64 were ex-smokers; 19 were current smokers, and 37 were unknown. Moreover, 96.0% (193/201) of patients were administered steroids; 60 patients received pulse steroid therapy (methylprednisolone, 1000 mg/day) and 133 were receiving non-pulse steroid therapy. The median duration of remdesivir administration was 5 days (range, 1-10 days). The median duration of steroid administration was 10 days (range, 2-69 days). In total, 111 adverse events classified as grade 3 or higher were observed in 69 patients ( www.nature.com/scientificreports/    Acute pancreatitis or severe elevation of pancreatic enzyme levels after remdesivir treatment. A grade 3 or higher increase in pancreatic enzyme levels was observed in 23 of 201 patients ( Table 3).
The median age was 73 (range: 50-91) years; 14 patients (60.9%) were male, and 21 patients were Japanese. The median time from COVID-19 symptom onset to remdesivir administration was 5 days (range, 1-9 days). The median time from remdesivir administration to the elevation of pancreatic enzyme levels was 5 (range: 1-10) days; 16 patients received pulsed steroid therapy, 6 received non-pulse steroid therapy, and one patient did not receive steroids. The median duration of steroid administration was 12.5 (range: 5-69) days. Among the 23 patients, eight were regular drinkers, two were occasional drinkers, and six were non-drinkers, and the drinking status of seven patients was unknown. Of the 23 patients, amylase, lipase, and both were elevated more than three times the upper limit of the normal range in 6, 21, and one patient, respectively. In addition, six of the 23 patients had complications of abdominal pain; one among these patients reported nausea. Computed tomography (CT) was performed in the 6 patients after the increase in pancreatic enzyme levels and complaints www.nature.com/scientificreports/ of abdominal pain and nausea, and acute pancreatitis was confirmed in three patients (Fig. 2). Hospitalized COVID-19 patients without remdesivir treatment. A total of 159 patients whose hospitalizations were between February 5, 2020 and January 31, 2021 were enrolled in this study. The baseline characteristics of the patients are presented in Table 4. Based on the COVID-19 severity classification according to the NIH criteria, four cases were asymptomatic, 22 were mild, 60 were moderate, 48 were severe, and 25 were critical. Eighteen patients received pulse steroid therapy, 56 received non-pulse steroid therapy, and 85 did not receive steroid therapy at all. Increased pancreatic enzyme levels were observed in three of the 159 patients. CT was performed in one of the three patients after the increase in pancreatic enzyme levels, but no signs of pancreatitis were observed on CT scans.

Results of the statistical analysis.
The associations between acute pancreatitis and grade 3 or higher pancreatic enzyme level elevation and patient characteristics are shown in Table 5. The proportion of patients with elevated pancreatic enzyme levels was predominantly higher among remdesivir-treated patients (p = 0.001), steroid-treated patients (p = 0.008), and patients with NIH criteria of critical COVID-19 (p = 0.000). Logistic regression analysis was performed using the onset of acute pancreatitis and increase in pancreatic enzyme levels above grade 3 severity as dependent variables and whether remdesivir was administered, whether steroids were administered, sex, age (< 65 years or ≥ 65 years), whether patients had diabetes, hypertension, hyperlipidemia, and elevated D-dimer levels, as well as the NIH disease severity criteria for COVID-19 (critical vs. not critical) as covariates. Participants meeting the NIH criteria for critical COVID-19 and receiving remdesivir had a higher risk of acute pancreatitis and a grade 3 or higher increase in pancreatic enzyme levels, with an adjusted OR of 10.137 (95% CI: 3.631 to 28.299, p = 0.000) for the NIH criteria and 4.565 (95% CI: 1.113-18.724, p = 0.035) for remdesivir administration. Additionally, administration of steroids, sex, age, diabetes, hypertension, hyperlipidemia, and elevated D-dimer levels did not have a high risk of acute pancreatitis and grade 3 or higher increase in pancreatic enzyme levels ( Table 6).

Discussion
This retrospective analysis of COVID-19 patients treated with remdesivir at our institution indicated that increased risk of acute pancreatitis and grade 3 or higher increase in pancreatic enzyme levels is associated with remdesivir administration. The results of this study differed from those of previously published studies. To the best of our knowledge, this is the first study to report a high incidence of acute pancreatitis or a severe increase in pancreatic enzyme levels in COVID-19 patients treated with remdesivir. In the ACTT-1 study, 538 patients received remdesivir, but acute pancreatitis was not reported 4,5 . Similarly, in two other studies where 384 and 155 patients received remdesivir, pancreatitis was not reported 6,7 . Only one case of acute pancreatitis was reported in the ACTT-2 study in the remdesivir plus baricitinib group 8 .
However, compared to ACTT-1 and other studies, most patients in our study were treated with steroids, which increased the incidence of acute pancreatitis approximately two-fold 11 . Thus, steroids may have contributed to the development of pancreatitis in our patients. However, pancreatitis was not reported in the RECOVERY trial, which was the first to demonstrate the effects of dexamethasone on COVID-19 12 .
As for the risk of pancreatitis, gallstones were observed on the CT scans of 5 of 23 patients, but none of them had findings suggestive of gallstone pancreatitis. Regarding the history of liver disease, there were no patients with cirrhosis, and only seven patients had previous hepatitis B virus (HBV) infections. There was little difference in alcohol consumption history between the elevated pancreatic enzyme group and the remdesivir group as a Table 4. Characteristics of COVID-19 patients who did not receive remdesivir. Baseline demographic and clinical characteristics of 157 patients who did not received remdesivir between February 5, 2020 and January 31, 2021. www.nature.com/scientificreports/ whole, and this percentage was similar to that observed in The National Health and Nutrition Survey in Japan, 2019 conducted by the Ministry of Health, Labour and Welfare 13 . Other etiological causes of acute pancreatitis were excluded from the patients diagnosed with acute pancreatitis. Preclinical data on the use of GS-5734 (remdesivir) for rhesus monkeys infected with the Ebola virus showed increased serum lipase levels several days after administration, suggesting that remdesivir leads to an increased Table 5. Increase in pancreatic enzyme levels and patient characteristics. The association between a Grade 3 or higher increase in pancreatic enzyme levels or acute pancreatitis and patient characteristics was analyzed using the chi-square test. NE Not examined, NIH The National Institutes of Health. *Statistically significant (p ≤ 0.05).  www.nature.com/scientificreports/ production of pancreatic enzymes 14 . A literature search also revealed multiple reports of increased pancreatic enzyme levels during the course of COVID-19 [15][16][17] . Therefore, we reviewed the data of hospitalized COVID-19 patients who did not receive remdesivir at our institution. A total of 159 patients with COVID-19 hospitalized between February 2020 and January 2021 were enrolled. Increased pancreatic enzyme levels were observed in 3 of the 159 patients (Table 4). Logistic regression analysis was performed. Participants meeting the NIH criteria for critical COVID-19 and who received remdesivir had a higher risk of acute pancreatitis and grade 3 or higher increase in pancreatic enzyme levels, with an adjusted OR of 10.137 (95% CI: 3.631 to 28.299, p = 0.000) for the NIH criteria and 4.565 (95% CI: 1.113-18.724, p = 0.035) for remdesivir administration.
The strength of our study is that these data are from a designated medical institution for Class I Infectious Diseases (e.g., Ebola hemorrhagic fever), and all COVID-19 patients were treated by infectious disease specialists or respiratory specialists as attending physicians or supervisors, which enabled us to provide advanced medical care. Moreover, these findings were obtained from patients with COVID-19 treated at the same institution and using the same clinical decision. When a large number of medical institutions are involved, it is difficult to ensure homogeneity of management, and thus, data quality, and the evaluation of disease severity and adverse events is likely to vary.
This study has some limitations. First, this was a small, uncontrolled, single-center, retrospective study. Second, most patients who received remdesivir were Japanese, and the possibility of racial diversity could not be ruled out. However, there was no elevation in pancreatic enzyme levels in a study among ethnically related Chinese participants 6 or in a global study that included a proportion of Asians. These results lead us to believe that the increased risk of pancreatitis is not only present in patients of Asian descent. Therefore, pancreatitis or increase in pancreatic enzyme levels may have been incorrectly judged as a manifestation of COVID-19. Third, our analysis is limited by a lack of data on comorbidity and obesity, which are risk factors for hospitalization with COVID-19, except for diabetes, hypertension, and hyperlipidemia, which could be assessed based on glycated hemoglobin levels and regular medication prescriptions. Fourth, in the present study, CT was performed in only six out of 23 cases, radiologically confirmed in three cases (50%), and acute pancreatitis was confirmed in three cases. Therefore, elevated pancreatic enzymes may be reactive and not clinically relevant. Finally, as many COVID-19 patients treated with remdesivir at our hospital had increased pancreatic enzyme levels, there were actually several cases wherein remdesivir administration was discontinued early when pancreatic enzyme levels were elevated to grade 2, or cases where nafamostat or ulinastatin were co-administered when the pancreatic enzyme levels were increased in patients whose cycle threshold values of SARS-COV-2 reverse transcription polymerase chain reaction were low and remdesivir administration was continued; thus, the increase in pancreatic enzyme levels could have been underestimated.
It is unclear whether pancreatic enzyme levels are elevated in severe COVID-19 infection itself as part of multiorgan failure, or in conjunction with unknown factors. Our results suggest that remdesivir administration was an independent predictor of increased pancreatic enzyme levels, in addition to COVID-19 severity. Followup studies on remdesivir administration and elevated pancreatic enzyme levels are warranted.
In conclusion, in the present study, 23 of 201 patients treated with remdesivir had elevated pancreatic enzyme levels. Additionally, logistic regression analysis revealed remdesivir administration as an independent predictor of increased pancreatic enzyme levels. Therefore, we propose monitoring pancreatic enzyme levels in COVID-19 patients to detect the onset of pancreatitis during the course of remdesivir treatment for COVID-19.