The rapid dissemination of practice patterns for coronavirus disease 2019 (COVID-19) underscores the need for ongoing critical safety and efficacy surveillance of off-label treatments. The Department of Veterans Affairs conducts a safety surveillance program to track the safe and appropriate use of off-label treatments for veterans with COVID-19, including those approved under emergency use authorization (EUA). Tocilizumab, a monoclonal antibody to the IL-6 receptor, has been used as an off-label treatment for hospitalized patients with moderate-severe COVID-19. Reference Guaraldi, Meschiari and Cozzi-Lepri1,Reference Somers, Eschenauer and Troost2 Methods of posttreatment surveillance are needed to weigh the benefits and harms of off-label treatments such as tocilizumab, particularly while randomized controlled trials are ongoing. There are limited reports of adverse drug events (ADEs) and subsequent unfavorable outcomes associated with tocilizumab. Reference Lan, Lai, Huang, Chang, Lu and Hsueh3 The Veterans Health Administration (VHA), with 170 medical centers across 22 geographic regions, offers an ideal setting to inform models of posttreatment surveillance. We characterized ADEs associated with tocilizumab using a random national sample of inpatients with COVID-19 across the VHA.
Methods
We conducted a retrospective evaluation of patients ≥18 years of age with COVID-19 who were hospitalized between March 12, 2020, and August 31, 2020, in the VHA. We identified the subset of patients who received ≥1 dose of tocilizumab during hospitalization. Randomized controlled trials evaluating the safety and efficacy of tocilizumab in COVID-19 patients were not enrolling VHA patients during the evaluation; patients received tocilizumab at the discretion of their clinicians. Exposure to tocilizumab was ascertained through the electronic health record and confirmed on chart review. This project was approved under the Drug Use Evaluation project by the VHA Institutional Review Board.
Among all inpatients prescribed tocilizumab for COVID-19–associated diagnoses during the evaluation period, we selected a 40% sample of patients using simple random sampling. For all patients in the sample, we collected demographic, comorbidity, hospitalization, laboratory, and tocilizumab administration data by database extraction and confirmed these data via chart review. We ascertained whether patients were prescribed concomitant medications to manage COVID-19. We also obtained data related to the presence of symptoms associated with COVID-19, receipt of critical care, need for mechanical ventilation, and exposure to risk factors for COVID-19.
The primary outcome was the proportion of patients with an ADE temporally associated with tocilizumab administration. ADEs included injection site reactions, neutropenia, thrombocytopenia, anaphylaxis, hepatotoxicity, gastrointestinal perforation, elevated triglycerides, and new infections (tuberculosis, candidemia, bacteremia, varicella zoster, and other infections). 4 Trained clinical pharmacist specialists used standardized definitions to evaluate ADEs (Supplementary Table 1 online) among equitable samples of patients. Among patients who developed ADEs, secondary outcomes of ADE severity were characterized as (1) discontinuation of concomitant medications associated with select ADEs (neutropenia, thrombocytopenia, anaphylaxis, or hepatotoxicity) during hospitalization and (2) death during hospitalization. Multiple measures (eg, verification by senior personnel, cross checking, standardized data collection instructions) were implemented to limit interreviewer variability.
Descriptive characteristics were determined using frequencies and percentages for categorical variables or medians and interquartile ranges (IQRs) for continuous variables. To account for the geographic and temporal variation in prevalence and management of COVID-19 across the VHA, characteristics were stratified by the first wave (March 12, 2020–May 31, 2020) and the second wave (June 1, 2020–August 31, 2020) of COVID-19. The frequencies of ADEs were stratified by severity of COVID-19, defined by the need for mechanical ventilation. Bivariate tests were used to compare length-of-stay and all-cause mortality using the Wilcoxon rank-sum test and the χ Reference Somers, Eschenauer and Troost2 or Fisher exact test (as appropriate) among patients with versus without an ADE. Analyses were conducted using SAS version 9.4 software (SAS Institute, Cary, NC). All statistical tests were 2-tailed and P values < .05 were considered significant.
Results
We identified a random sample of 173 inpatients from 428 total inpatients with COVID-19 who received tocilizumab across 43 VHA medical centers. Table 1 shows descriptive characteristics. There appeared to be increased representation of women, adults aged ≥75 years, and concomitant treatments with corticosteroids and remdesivir in the second wave. Overall, 51 patients (29%) developed a total of 68 ADEs associated with tocilizumab per standardized criteria (Fig. 1). Of these 51 patients, 11 (22%) had 1 or more concomitant medications discontinued due to an ADE.
Characteristics of Patients Hospitalized With COVID-19 Who Received Tocilizumab at Veterans Health Administration Medical Centers (n=173 veterans from 43 centers)
Note. SD, standard deviation.
a Defined as absolute neutrophil count <500 cells/mm Reference Lan, Lai, Huang, Chang, Lu and Hsueh3 ; hematopoietic stem cell transplant or solid organ transplantation prior to hospitalization; receipt of corticosteroids (≥20 mg prednisone equivalents for ≥2 of the previous 4 weeks); receipt of antirejection, biologic, or disease-modifying antirheumatic drug in the previous 3 months; receipt of systemic chemotherapy or cytotoxic agent in the previous 6 weeks.
b Assessed in the 14 days prior to the COVID-19 diagnosis date.
c Defined as during the same admission in which tocilizumab was administered, from the time of hospital admission to the time of tocilizumab administration. There were no prescriptions for chloroquine, lopinavir, or eculizumab during the observation period.
Adverse drug events associated with tocilizumab among 173 patients hospitalized with non-severe (n = 68) and severe coronavirus disease 2019 requiring mechanical ventilation (n = 105). Median laboratory values following administration of tocilizumab varied in patients with adverse drug events compared with their baseline values for hepatotoxicity (n = 29; 305 U/L [IQR=266] vs. 40 U/L [IQR=18]); thrombocytopenia (n = 12; 108 K/cmm [IQR=49] vs. 172 K/cmm [IQR=120]); elevated triglycerides (n = 12; 397 mg/dL [IQR=260] vs. 150 mg/dL [IQR=107]); and neutropenia (n = 1; provider notated without laboratory values).
Median time to ADE from exposure to tocilizumab was 5 days (IQR=4). Among ADEs, there were 13 infections including 5 bacteremias, 1 candidemia, and 7 “other” severe infections. Of those with infection, 7 (54%) received concomitant steroids. There were no cases of tuberculosis or varicella zoster infection. One case of gastrointestinal perforation was observed.
There was no difference in length-of-stay among patients with (median 19 days, IQR=19) versus without (median 15 days, IQR=18) an ADE (P = .60). Additionally, there was no difference in all-cause hospital mortality among patients with (n = 20 of 51, 39%) versus without (n = 49 of 122, 40%) an ADE (P = 1.0).
Discussion
Efficient and comprehensive surveillance of off-label treatments for COVID-19 is necessary to identify opportunities to improve patient safety among hospitalized patients. We evaluated ADEs associated with tocilizumab to inform a potential framework for the real-time assessment of the safety of off-label treatments used among VHA inpatients with COVID-19. Using a random national sample, nearly 30% developed ADEs, and ADEs often led to discontinuation of concomitant medications. Hospital mortality did not vary by the development of ADEs. Collectively, our findings support hospital epidemiology and antimicrobial stewardship programs that are charged with promoting patient safety, pandemic preparedness, and, increasingly, institutional treatment protocols for COVID-19. Reference Mazdeyasna, Nori and Patel5
The Sentinel Initiative led by the Food and Drug Administration includes the largest multicenter database in the world dedicated to medical product safety, 6 and collaborative efforts with VHA inform safety surveillance efforts in the largest integrated healthcare system in the United States. It is important to conduct surveillance in this population because VHA patients may be at increased risk of ADEs due to the prevalence of advanced age, multimorbidity, and disparities in COVID-19 burden. Reference Rogers, Kazis and Miller7,Reference Rentsch, Kidwai-Khan and Tate8 By conducting rigorous chart reviews using standardized definitions in a random sample of nationwide inpatients, our work offers 1 potential approach to evaluate the safety of off-label treatments for COVID-19 in close to real-time within VHA and general populations of older adults with multimorbidity. This approach presents opportunities to improve safety among hospitalized veterans with COVID-19 and helps capture real-world outcomes for patients exposed to off-label treatments, beyond what is typically reported from clinical trials. Reference Hermine, Mariette and Tharaux9,Reference Salvarani, Dolci and Massari10 An additional advantage of this approach using integrated VHA data is the potential to replicate this methodology over time for veteran and nonveteran populations, particularly during subsequent waves of COVID-19 and future pandemics. Developing mechanisms to monitor off-label treatments for emerging infectious diseases across healthcare systems may promote pandemic preparedness.
Our work has limitations. This work was done within a national pharmacy safety surveillance program with the goal of tracking the safe and appropriate use of off-label medications in veterans. Consequently, we could not assess ADEs among all VHA inpatients exposed to tocilizumab. Nevertheless, inpatients were selected randomly, and the sample size is comparable to prior reports. Reference Lan, Lai, Huang, Chang, Lu and Hsueh3 Second, our findings have limited representation of women. Finally, we were not powered to identify predictors of ADE in veterans.
In summary, 29% of patients exposed to tocilizumab developed an ADE during hospitalization. ADEs often involved hepatotoxicity and secondary infections and frequently resulted in discontinuation of concomitant medications. This evaluation provides a potential framework for the real-time assessment of the safety of off-label treatments used for COVID-19 among VHA inpatients and may inform recommendations from hospital epidemiology and antimicrobial stewardship programs.
Supplementary material
To view supplementary material for this article, please visit https://doi.org/10.1017/ice.2021.227
Acknowledgments
We thank the Hospital Epidemiology and Infection Prevention Program and the Pain Research, Informatics, Multimorbidities, and Education Center at the Veterans Affairs Connecticut Healthcare System for their support of this study. The statements contained in this article reflect the views of the authors and do not represent the official positions of the United States Department of Veterans Affairs or other author affiliate organizations.
Financial support
No financial support was provided relevant to this article.
Conflicts of interest
All authors report no conflicts of interest relevant to this article.
