key: cord-1016604-ro7fmgt9 authors: Tsuzuki, S.; Hayakawa, K.; Matsunaga, N.; Terada, M.; Suzuki, S.; Ohtsu, H.; Asai, Y.; Kitajima, K.; Saito, S.; Uemura, Y.; Shibata, T.; Kondo, M.; Izumi, K.; Hojo, M.; Mizoue, T.; Yokota, K.; Nakamura-Uchiyama, F.; Saito, F.; Sugiura, W.; Ohmagari, N. title: Efficacy of remdesivir in Japanese patients hospitalised with COVID-19: A large observational study using the COVID-19 Registry Japan date: 2021-03-10 journal: nan DOI: 10.1101/2021.03.09.21253183 sha: 75d34cea0f15e4d3206794ec96e98990dd27ec3d doc_id: 1016604 cord_uid: ro7fmgt9 Objectives Results from several randomised controlled trials comparing the efficacy of remdesivir with placebo have been reported. Data on the efficacy of remdesivir in Asian populations are limited, and conflicting results have been reported. The aim of the present study is to examine whether remdesivir is effective for the treatment of COVID-19 or not. Methods We evaluated the efficacy of remdesivir from data acquired from the COVID-19 Registry Japan, a nationwide registry of Japanese patients hospitalised with COVID-19, using five-to-one propensity score (PS) matching. Primary outcomes were overall risk of fatality, risk of invasive mechanical ventilation or extracorporeal membrane oxygenation (IMV/ECMO) and length of stay (LoS) in healthcare facilities. Results In total, 1907 of 15,225 patients met our inclusion criteria. After PS matching, 74 patients were assigned to the remdesivir case group and 195 patients to the non-remdesivir control group. In a PS-matched cohort, fatality risk (12.2% vs 13.3%; P = 1.0); risk of IMV/ECMO (5.4% vs 4.6%; P = 0.757); length of intensive care unit stay (6 vs 6 days; P = 0.473) and length of IMV (13 vs 5 days; P = 0.509) were not different between the case and control groups. LoS in the case group was longer than in the control group (14 vs 11 days; P < 0.001). Conclusions This study suggested that remdesivir may have no positive effect on clinical outcomes and reductions in invasive/non-invasive respiratory support for patients with COVID-19 in Japan. As in other parts of the world, Japan has experienced rising numbers of patients with COVID-19, resulting in 364,813 cases and 5084 deaths between January 14, 2020, and January 25, 2021. 1 In addition to treating hyperinflammation and coagulopathy, antiviral medication is an important component in treating patients with COVID-19. 2,3 Among the antiviral medications for SARS-CoV-2, only remdesivir was approved for the treatment of patients with COVID-19 in Japan on May 7, 2020. 4 Randomised controlled trials (RCTs) comparing the efficacy of remdesivir with placebo have been conducted. Results from an RCT in China showed no statistically significant clinical benefit from remdesivir in hospitalised COVID-19 pneumonia patients with hypoxia. 5 Findings from a multinational RCT (ACTT-1) in Europe, the United States and Asia, including Japan, showed that remdesivir shortened time to recovery in hospitalised COVID-19 patients with pneumonia. 6 However, in a subgroup analysis, the statistically significant reduction in time to recovery was not seen in patients who were intubated or on extracorporeal membrane oxygenation (ECMO) therapy at the time of drug administration. There was no statistically significant shortened clinical Organization, showed no survival benefit from remdesivir in patients hospitalised with The trial enrolled 61% patients from Asia and Africa, however, no patients from Japan were enrolled. In another RCT conducted in the United States, Europe and Asia that included 16% to 19% Asian patients, those with moderate COVID-19 pneumonia (room-air oxygen saturation >94%) were randomised to a 5-day remdesivir, 10-day remdesivir, or standard treatment group. 8 In this study, the 5-day remdesivir group achieved better clinical status compared with the standard treatment group by day 11. There was no statistically significant difference between the 10-day remdesivir group and the standard treatment group. As described above, conflicting results have been reported regarding the clinical efficacy of remdesivir, and currently, recommendations in the guidelines for its use in patients with COVID-19 are inconsistent. 9, 10 The COVIREGI-JP (COVID-19 Registry Japan) is a national database of patients hospitalised with COVID-19. As of December 2, 2020, COVIREGI-JP included data for approximately 10% of infected patients (n = 150,386) and approximately 12% of patients discharged from hospitals or care facilities (n = 127,304) based on the number of completed registrations (n = 15,225) for major items in the registry. Results of a previous study using the showed that the number of Japanese patients All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 10, 2021. ; https://doi.org/10.1101/2021.03.09.21253183 doi: medRxiv preprint with severe underlying diseases and the fatality rate were lower than in Europe and the United States. Because of these differences in the clinical epidemiology of patients, it is difficult to determine the efficacy of remdesivir in Japan based solely on the results of existing overseas studies. Therefore, we conducted this study to evaluate the efficacy of remdesivir in a cohort of Japanese patients. This study used patient data obtained from COVIREGI-JP, 11 which opened on March 2, 2020. Inclusion criteria for enrolment in COVIREGI-JP are a positive SARS-CoV-2 test and inpatient treatment at a healthcare facility. SARS-CoV-2 testing is based on the notification criteria of the Infectious Diseases Law, which are the isolation and identification of SARS-CoV-2 or detection of SARS-CoV-2 genes by nucleic acid amplification such as polymerase chain reaction or loop-mediated isothermal amplification) from clinical specimens or the detection of pathogen antigens from nasopharyngeal swabs, etc (after May 13, 2020). 12 We have modified a case report form from the International Severe Acute Respiratory and Emerging Infection Consortium 13 to facilitate the collection of demographic, All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 10, 2021. ; epidemiological and epidemiological data on clinical course and treatment. Study data were collected and managed using Research Electronic Data Capture, a secure, web-based application hosted at the Japan Clinical Research Assist Center of the National Center for Global Health and Medicine (NCGM). We used frozen data from patients who entered all of the following major items as of December 2, 2020, for this study, as reported previously: 11 demographic and epidemiological information at admission; comorbidities; signs and symptoms at admission (including conditions at admission); outcome at discharge; supportive care during hospitalisation; drug administration during hospitalisation and complications during hospitalisation. Among patients registered with COVIREGI-JP, we excluded those who met the following exclusion criteria: 1. Non-Japanese (to evaluate the efficacy of remdesivir in the Japanese cohort) 2. Admission before May 7, 2020 (approval date for remdesivir in Japan) 3. More than 5 days from symptom onset to day of admission (to minimise variance in the number of days to the start of inpatient treatment) All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 10, 2021. ; 4. More than 8 days from symptom onset to day of remdesivir administration, if treated (to exclude patients who became severely ill before administration of remdesivir and thus in whom antiviral therapy may not provide sufficient efficacy 3 ) 5. Treated with favipiravir (to evaluate the effect of remdesivir only) 6 . Treated with remdesivir for fewer than 3 days (to exclude patients partially impacted by remdesivir) This study compared patient outcomes in the remdesivir case group and the non-remdesivir control group. Primary outcomes were overall fatality risk, risk of invasive mechanical ventilation or extracorporeal membrane oxygenation (IMV/ECMO) and length of stay (LoS) in healthcare facilities. We used Fisher's exact test for risks of fatality and IMV/ECMO and Wilcoxon's rank sum test for LoS, with five-to-one propensity score (PS) matching (five-to-one nearest neighbour pair matching, caliper = 0.2) calculated using a multivariate logistic regression model to predict the likelihood of remdesivir administration. 14 Age; sex; presence of cardiovascular diseases, chronic respiratory diseases, severe renal diseases (serum creatinine level 3 mg/dL and higher) or dialysis, hypertension, hyperlipidemia, obesity All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 10, 2021. ; diagnosed by physicians and solid tumour; the worst National Early Warning Score (NEWS) 15 before administration of remdesivir (we used the NEWS at day 1 in the control group); days from symptom onset to admission; use of corticosteroids and use of anticoagulants were all included in the model. The standardised difference was used to measure covariate balance, and an absolute standardised difference above 10% was interpreted as a meaningful imbalance. 16 The worst NEWS was the worst score among the scores at days 1, 4 and 8 because the registry data do not include daily clinical findings. These possible confounders were chosen for their potential association with the outcome of interest based on clinical knowledge and results from previous studies. [17] [18] [19] [20] [21] [22] [23] [24] [25] Secondary outcomes were length of intensive care unit (ICU) stay, length of IMV and duration of oxygen requirement. The length of ICU stay and IMV were compared using Wilcoxon's rank sum test. The duration of oxygen requirement was evaluated as duration of oxygen administered to patients who need oxygen at days 1, 4 and 8 because COVIREGI-JP does not include information on daily clinical course. The log-rank test was used to compare the duration of oxygen administration. We also conducted a similar analysis of the subgroup, which included only patients at high risk for severe COVID-19 who had one or more risk factors of age 65 and older, older, more frequently male, and experienced more severe infection and fatalities (Table 1 ). ( Table 1) After PS matching, 74 patients were included in the case group and 195 patients in the control group. Comparison of the two groups before and after matching is shown in Table 2 . Patients in the case group had more comorbidities and more severe infection before PS matching. After PS matching, the standardised mean difference for each variable included was within the range defined in Methods (<0.10) ( Table 2 ). ( Table 2) In the PS-matched cohort, fatality risk (12.2% vs 13.3%; P = 1.0), IMV/ECMO risk (5.4% vs 4.6%; P = 0.757), length of ICU stay (7 vs 6 days) and length of IMV (13 vs 5 days; P = 0.509) were not different between the case and control groups ( Table 3) . Length of stay in the case group was longer than in the control group (median LoS, 14 vs 11 days; P < 0.001). Details of the primary and secondary outcome are shown in All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 10, 2021. ; Table 3 . (Table 3) Regarding the safety of remdesivir, 2 of 74 patients reported adverse events (Table 4) . Only 1 patient with elevated liver enzymes was considered a case of probable relevance to remdesivir administration. No patient stopped remdesivir administration due to adverse events. Both patients with adverse events recovered without sequelae (Table 4) . (Table 4) The duration of oxygen administration for patients who received supplementary oxygen at days 1, 4 and 8 was not different between the case and control groups (Table 5, Figure 1 ). These results suggest that more than half of patients can discontinue supplementary oxygen within 2 weeks; however, some required supplementary oxygen longer than 3 weeks (Table 5, Figure 1 ). All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Tables 1-3 ). Our findings showed that remdesivir did not reduce fatality and IMV/ECMO risks in patients hospitalised with COVID-19. This result coincides with findings from a previous study 7 and supports the conclusion that remdesivir is not an essential drug for COVID-19-specific treatment as suggested in recent clinical guidelines. 9,27 A previous study reported a clinical benefit from remdesivir; 6 however, Beigel and colleagues defined their main outcome as the time to recovery and did not show that remdesivir improved patient outcomes such as fatality and need for mechanical ventilation. In our study, LoS in the case group was longer than in the control group. Although statistically insignificant, length of IMV tended to be longer in the case group. Adjustments by PS matching were based on the relatively early clinical course and may All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 10, 2021. ; https://doi.org/10.1101/2021.03.09.21253183 doi: medRxiv preprint not be able to adjust completely for parameters that contribute to the longer LoS such as the need for respiratory physiotherapy or medical rehabilitation. Therefore, the longer LoS in the case group was not considered attributable to remdesivir. Rather, reversal of causality in patients who needed or appeared to need longer hospital care tended to include treatment with remdesivir and other drugs. Although the impact of duration of administration on the effect of remdesivir is not an aim of this study, it is noteworthy. Because our study design was retrospective, the duration of remdesivir administration depended on facility and physician policies. A previous study reported that no significant difference was shown between a 5-and 10-day course of remdesivir in patients with severe COVID-19 who did not require mechanical ventilation. 28 In addition, a 10-day course of remdesivir did not have a statistically significant difference in clinical status compared with standard care among patients with moderate COVID-19. 8 In our data after PS matching, 22 of 74 patients (29.7%) were administered remdesivir over 5 days, and 14 were administered the drug over 10 days. Fatality risk was not different between the two groups in a descriptive analysis (5 deaths among patients on a 5-day course and 3 deaths among patients on a 10-day course). Subgroup analyses included patients at high risk for severe disease and did not show All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 10, 2021. ; obvious differences compared with the main analyses. In fact, 199 of 269 patients (74.0%) in the matched data for the main analysis had one or more risk factors for severe illness. This means that remdesivir tended to be used only in severely ill patients with underlying risk factors and that consequently, the results of both analyses were similar. Why the case group in this study included comparatively more patients with severe illness is probably the indication for remdesivir in Japan. 29 Remdesivir was authorised for use in Japan in May 2020 by fast-track approval followed by United States Food and Drug Administration Emergency Use Authorization. 30 At that time, the indication for remdesivir was limited to severe patients whose oxygen saturation was 94% and under (ambient air) who need supplementary oxygen or IMV/ECMO. Previous reports suggested that the appropriate target patients for treatment with remdesivir were not already severely ill but patients with mild pneumonia. 6, 8 In January 2021, the Ministry of Health, Labour and Welfare Japan extended its indication to patients who have pneumonia due to SARS-CoV-2 infection. The number patients with mild or moderate infection administered remdesivir should increase in the future. Thus, evaluating its effect on patients with mild pneumonia is an anticipated future challenge. The most important limitation of this study is its retrospective cohort design. We All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 10, 2021. ; attempted to adjust various factors that affect clinical outcomes; however, our method does not enable us to adjust all confounding factors. 14, 31 In particular, the difference in LoS between the two study groups should be considered carefully for reasons noted above. In addition, because our data are derived from a registry, several items are difficult to interpret. For example, fatality in this study means a patient died during hospitalisation. Even if a patient died after discharge, we categorised the patient as survived. The cause of death is also not available from registry data, and therefore, we cannot determine the cause of death in patients with serious comorbidities such as cancer. Although patients who were hospitalised and/or administered remdesivir in the late clinical course were excluded from the study, a fraction may have been administered remdesivir during or after initiation of IMV/ECMO. Therefore, the risk of IMV/ECMO may be overestimated in the case group. Furthermore, COVIREGI-JP does not collect information on the daily clinical status of each patient. It was difficult to match time-dependent factors such as severity on each day between the case and control groups using the available data. Therefore, bias may have impacted our results to some extent, even using matched data. Adverse events attributable to remdesivir were determined by researchers and may have been underreported. Nevertheless, our data after PS matching showed that patients in the two study groups had similar All rights reserved. No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 10, 2021. ; characteristics such as age and comorbidity associated with clinical outcomes, and we find the results reliable. This study suggested that remdesivir may have no positive effect on clinical outcomes and reductions in invasive/non-invasive respiratory support for Japanese patients with COVID-19. The impact of remdesivir on patients with mild pneumonia and other clinical outcomes are subjects for additional study. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 10, 2021. ; (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted March 10, 2021. ; (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Labour and Welfare. The Number of Patients and other Information about COVID-19 in Japan Virology, transmission, and pathogenesis of SARS-CoV-2 Mild or moderate covid-19 Remdesivir in adults with severe COVID-19: a randomised, double-blind, placebo-controlled, multicentre trial Remdesivir for the treatment of covid-19 -final report covid-19 -interim WHO solidarity trial results Effect of remdesivir vs standard care on clinical status at 11 days in patients with moderate COVID-19: a randomized clinical trial Therapeutics and COVID-19: Living Guideline Information on COVID-19 Treatment, Prevention and Research. COVID-19 Treatment Guidelines 13 International Severe Acute Respiratory and Emerging Infection Consortium (ISARIC). COVID-19 Clinical Research Resources The central role of the propensity score in observational studies for causal effects The National Health Service England. National Early Warning Score (NEWS) A comparison of 12 algorithms for matching on the propensity score Characteristics of and important lessons from the coronavirus disease 2019 (COVID-19) outbreak in China: summary of a report of 72 314 cases from the Chinese Center for Disease Control and Prevention and critical illness among 5279 people with coronavirus disease 2019 in New York City: prospective cohort study Factors associated with COVID-19-related death using OpenSAFELY Clinical outcomes in young US adults hospitalized with COVID-19 Obesity in patients younger than 60 years is a risk factor for COVID-19 hospital admission Obesity and mortality among patients diagnosed with COVID-19: results from an integrated health care organization