NEJM - Neutralizing Antibody LY-COV555

The new england journal of medicine Original Article SARS-CoV-2 Neutralizing Antibody LY-CoV555 in Outpatients with Covid-19 Peter Chen, M.D., Ajay Nirula, M.D., Ph.D., Barry Heller, M.D., Robert L. Gottlieb, M.D., Ph.D., Joseph Boscia, M.D., Jason Morris, M.D., Gregory Huhn, M.D., M.P.H.T.M., Jose Cardona, M.D., Bharat Mocherla, M.D., Valentina Stosor, M.D., Imad Shawa, M.D., Andrew C. Adams, Ph.D., Jacob Van Naarden, B.S., Kenneth L. Custer, Ph.D., Lei Shen, Ph.D., Michael Durante, M.S., Gerard Oakley, M.D., Andrew E. Schade, M.D., Ph.D., Janelle Sabo, Pharm.D., Dipak R. Patel, M.D., Ph.D., Paul Klekotka, M.D., Ph.D., and Daniel M. Skovronsky, M.D., Ph.D., for the BLAZE-1 Investigators* ABSTRACT BACKGROUND Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causes coronavirus From the Department of Medicine, disease 2019 (Covid-19), which is most frequently mild yet can be severe and life- Women’s Guild Lung Institute, Cedars– threatening. Virus-neutralizing monoclonal antibodies are predicted to reduce viral Sinai Medical Center, Los Angeles (P.C.), and Long Beach Clinical Trials, Long load, ameliorate symptoms, and prevent hospitalization. Beach (B.H.) — both in California; Eli Lilly, Indianapolis (A.N., A.C.A., J.V.N., METHODS K.L.C., L.S., M.D., G.O., A.E.S., J.S., In this ongoing phase 2 trial involving outpatients with recently diagnosed mild D.R.P., P.K., D.M.S.), and Franciscan Health, Greenwood (I.S.) — both in Indi- or moderate Covid-19, we randomly assigned 452 patients to receive a single intra- ana; Baylor University Medical Center venous infusion of neutralizing antibody LY-CoV555 in one of three doses (700 mg, and Baylor Scott and White Research In- 2800 mg, or 7000 mg) or placebo and evaluated the quantitative virologic end stitute, Dallas (R.L.G.); Vitalink Research, Union, SC (J.B.); Imperial Health, Lake points and clinical outcomes. The primary outcome was the change from baseline Charles, LA (J.M.); Cook County Health in the viral load at day 11. The results of a preplanned interim analysis as of Sep- (G.H.) and Northwestern University tember 5, 2020, are reported here. Feinberg School of Medicine (V.S.), Chi- cago; Indago Research and Health Cen- ter, Hialeah, FL (J.C.); and Las Vegas RESULTS Medical Research Center, Las Vegas At the time of the interim analysis, the observed mean decrease from baseline in (B.M.). Address reprint requests to Dr. the log viral load for the entire population was −3.81, for an elimination of more Skovronsky at Eli Lilly, 893 Delaware St., than 99.97% of viral RNA. For patients who received the 2800-mg dose of LY- Indianapolis, IN 46225, or at s kovronsky_ daniel@ lilly . com. CoV555, the difference from placebo in the decrease from baseline was −0.53 (95% confidence interval [CI], −0.98 to −0.08; P = 0.02), for a viral load that was lower *A list of the BLAZE-1 investigators is provided in the Supplementary Appen- by a factor of 3.4. Smaller differences from placebo in the change from baseline dix, available at NEJM.org. were observed among the patients who received the 700-mg dose (−0.20; 95% CI, Drs. Chen and Nirula contributed equally −0.66 to 0.25; P = 0.38) or the 7000-mg dose (0.09; 95% CI, −0.37 to 0.55; P = 0.70). to this article. On days 2 to 6, the patients who received LY-CoV555 had a slightly lower severity This article was published on October 28, of symptoms than those who received placebo. The percentage of patients who had 2020, at NEJM.org. a Covid-19–related hospitalization or visit to an emergency department was 1.6% DOI: 10.1056/NEJMoa2029849 in the LY-CoV555 group and 6.3% in the placebo group. Copyright © 2020 Massachusetts Medical Society. CONCLUSIONS In this interim analysis of a phase 2 trial, one of three doses of neutralizing antibody LY-CoV555 appeared to accelerate the natural decline in viral load over time, where- as the other doses had not by day 11. (Funded by Eli Lilly; BLAZE-1 ClinicalTrials.gov number, NCT04427501.) n engl j med nejm.org 1 The New England Journal of Medicine Downloaded from nejm.org on November 19, 2020. For personal use only. No other uses without permission. Copyright © 2020 Massachusetts Medical Society. All rights reserved.

The new england journal of medicine oronavirus disease 2019 (Covid-19) the outpatient setting. We examined the effect of emerged in late 2019 and spread rapidly, the neutralizing antibody on viral load, symptom Cresulting in a global pandemic. Infected scores, and clinical outcomes and also report an persons can have a wide range of disease severity, observed connection between a persistently high with many patients showing mild or even asymp- viral load and disease severity. tomatic disease. However, for unknown reasons, up to 10% of asymptomatic and mild infections Methods lead to more severe outcomes, including respira- 1 Trial Design, Treatment, and Oversight tory distress requiring hospitalization. Although risk factors for more severe outcomes have been In this randomized, double-blind, placebo-con- described (including an older age, obesity, hyper- trolled, single-dose trial conducted at 41 centers tension, and underlying chronic medical condi- in the United States, all the patients had positive tions),2,3 the connection between viral load and results on testing for SARS-CoV-2 and presented outcomes has not previously been tested in a with one or more mild or moderate symptoms. longitudinal study. Several treatment options The investigators reviewed the symptoms, risk have been explored for hospitalized patients factors, and other inclusion and exclusion crite- 4 with Covid-19 (e.g., antimalarial drugs, antiviral ria before enrollment. (A full list of the inclusion 5-7 8-12 agents, immunomodulators, glucocorti- and exclusion criteria is provided in the protocol, 13,14 15,16 coids, and convalescent plasma ) with available with the full text of this article at varying results. However, there have been no NEJM.org.) Each patient received a single intra- large randomized, controlled trials of targeted venous infusion of LY-CoV555 or placebo mono- treatments that are specific for severe acute re- therapy over approximately 1 hour. Although spiratory syndrome coronavirus 2 (SARS-CoV-2) the trial contains additional treatment groups, and that are intended to attenuate disease pro- here we focus on the interim analysis of results gression in patients with early disease. from only four of these groups: LY-CoV555 at Preclinical studies of neutralizing-antibody doses of 700 mg, 2800 mg, and 7000 mg and treatments for SARS-CoV-2 infection in several placebo. (Clinical details are also provided in the animal models have shown promising results, protocol.) with marked reductions in viral loads in the up- The preplanned interim analysis was triggered 17 per and lower respiratory tracts. SARS-CoV-2 on September 5, 2020, when the last patient gains entry into cells through binding of its who was randomly assigned to receive LY-CoV555 spike protein to receptors for angiotensin-con- reached day 11. The analysis includes all the data 18 verting enzyme 2 on target cells. LY-CoV555 regarding virologic features and symptoms that (also known as LY3819253), a potent antispike were available at the time of the database lock. neutralizing monoclonal antibody that binds with The doses of LY-CoV555 that were evaluated in high affinity to the receptor-binding domain of this trial were based on pharmacologic model- SARS-CoV-2, was derived from convalescent plas- ing that predicted that the 700-mg dose would ma obtained from a patient with Covid-19. The be efficacious. (Details about dose selection are antibody was developed by Eli Lilly after its dis- provided in the Supplementary Appendix, avail- covery by researchers at AbCellera and at the able at NEJM.org.) Given the gravity of the pan- Vaccine Research Center of the National Institute demic, the doses that were administered in this of Allergy and Infectious Diseases. The discovery trial were increased by up to a factor of 10 over of LY-CoV555 and its passive protection against the predicted efficacious dose to ensure adequate SARS-CoV-2 in nonhuman primates has been target coverage. The use of these doses was sup- 19 reported previously. ported by safety data from a phase 1 trial of Here, we report interim results from the LY-CoV555 involving hospitalized patients. Dose Blocking Viral Attachment and Cell Entry with levels were fixed, and either LY-CoV555 or pla- SARS-CoV-2 Neutralizing Antibodies (BLAZE-1) cebo was administered within 3 days after posi- trial, an ongoing phase 2 trial to evaluate the tive results on SARS-CoV-2 testing. efficacy and safety of LY-CoV555 in patients with The trial, which was sponsored by Eli Lilly, recently diagnosed mild or moderate Covid-19 in was conducted in accordance with principles of 2 n engl j med nejm.org The New England Journal of Medicine Downloaded from nejm.org on November 19, 2020. For personal use only. No other uses without permission. Copyright © 2020 Massachusetts Medical Society. All rights reserved.

Neutralizing Antibody in Outpatients with Covid-19 the Declaration of Helsinki and the ethical guide- lines of the Council for International Organiza- 1ƒ1 Patients …ere enrolled and assined tions of Medical Sciences. All the patients pro- to „ƒƒ m of ‡ˆ-CoV‰‰‰ monoterapy Interim Analysis vided written informed consent. Positive SARS-CoV-2 test ≤3 days 1ƒ„ Patients …ere enrolled and assined before infusion to 2†ƒƒ m of ‡ˆ-CoV‰‰‰ monoterapy Mild or moderate Covid-19 symptoms Outcomes Primary end point ane from The primary outcome was the change from baseline to day 11 ± days­ 1ƒ1 Patients …ere enrolled and assined in SARS-CoV-2 viral load baseline in the SARS-CoV-2 viral load at day 11 to „ƒƒƒ m of ‡ˆ-CoV‰‰‰ monoterapy Seondary end points inlude safety€ (±4 days) after positive results on testing. Data symptom severity€ ospitali‚ation€ 1 3 Patients …ere enrolled and assined and time points for viral learane regarding virologic features and symptoms were to plaebo collected up to day 29 in this trial. The viral load was measured by means of a nasopharyngeal swab, which was followed by quantitative reverse- Figure 1. Enrollment and Trial Design. transcriptase–polymerase-chain-reaction (RT-PCR) assay at a central laboratory. (Details regarding testing are provided in the Methods section in regarding these methods are provided in Section the Supplementary Appendix.) Key secondary out- 6.10 in the statistical analysis plan.) comes were safety assessments, symptom burden as reported by the patient on a questionnaire, Results and clinical outcomes. The major clinical out- come was defined as Covid-19–related in-patient Patients hospitalization, a visit to the emergency depart- From June 17 through August 21, 2020, a total ment, or death. No deaths were reported, and of 467 patients underwent randomization to re- since most emergency department visits resulted ceive either LY-CoV555 (317 patients) or placebo in hospital admissions, we refer to a composite of (150 patients), and the patients in the LY-CoV555 emergency department visits and in-patient hos- group were assigned to one of three dose sub- pitalizations simply as hospitalizations. This re- groups. Of the patients who had undergone port includes an analysis of the primary outcome randomization, 452 met the criteria for inclusion as well as safety and adverse-event data, informa- in the primary analysis (309 in the LY-CoV555 tion regarding symptoms, and clinical outcomes. group and 143 in the placebo group). LY-CoV555 was administered to these patients in doses of Statistical Analysis 700 mg (101 patients), 2800 mg (107 patients), or To determine the sample size, we used a dynamic 7000 mg (101 patients) (Fig. 1). The two trial model to simulate viral load over time in patients groups were well balanced regarding risk factors treated with LY-CoV555 or placebo. This simu- at the time of enrollment (Table 1). Nearly 70% lated population was used to estimate the statis- of the patients had at least one risk factor — an tical power and comparisons in the change from age of 65 years or older, a body-mass index baseline in viral load. (Details are provided in (BMI, the weight in kilograms divided by the Section 5.2 in the statistical analysis plan, which square of the height in meters) of 35 or more, or is included in the protocol document.) All the at least one relevant coexisting illness — for patients who had undergone randomization and severe Covid-19. After undergoing randomiza- who had received either LY-CoV555 or placebo tion, patients received an infusion of LY-CoV555 were included in the primary analysis if their or placebo within a median of 4 days after the viral-load measures were available both at base- onset of symptoms; at the time of randomiza- line and at least once after baseline. tion, more than 80% of the patients had only Treatment effects were compared with the mild symptoms. The observed mean PCR cycle use of two-sided tests with an alpha level of 0.05. threshold (Ct) value of 23.9 on the day of infu- Adjustments for multiple testing were not per- sion (equating to approximately 2.5 million RNA formed. Significance testing for the primary out- equivalents) matched expectations that a recently come was performed with the use of a repeated- diagnosed population would have a high viral measures analysis as a mixed model. (Details burden. The conversion from Ct value to viral load n engl j med nejm.org 3 The New England Journal of Medicine Downloaded from nejm.org on November 19, 2020. For personal use only. No other uses without permission. Copyright © 2020 Massachusetts Medical Society. All rights reserved.

The new england journal of medicine Table 1. Characteristics of the Patients at Baseline.* line was −0.53 (95% confidence interval [CI], −0.98 to −0.08; P = 0.02), for a lower viral load by a LY-CoV555 Placebo factor of 3.4 (Table 2). However, smaller differ- Characteristic (N = 309) (N = 143) ences from placebo in the decrease from base- Age line were observed among the patients who re- Median (range) — yr 45 (18–86) 46 (18–77) ceived the 700-mg dose (−0.20; 95% CI, −0.66 to 65 Yr or older — no. (%) 33 (10.7) 20 (14.0) 0.25; P = 0.38) and the 7000-mg dose (0.09; 95% CI, Female sex — no. (%) 171 (55.3) 78 (54.5) −0.37 to 0.55; P = 0.70). Race or ethnic group — no./ total no. (%)† Secondary Viral Outcomes White 269/305 (88.2) 120/138 (87.0) On day 3, among the patients who received the 2800-mg dose of LY-CoV555, the observed differ- Hispanic or Latino 135/309 (43.7) 63/143 (44.1) ence from placebo in the decrease from baseline Black 22/305 (7.2) 7/138 (5.1) in the mean log viral load was −0.64 (95% CI, Body-mass index‡ −1.11 to −0.17) (Table 2). The other two doses of Median 29.4 29.1 LY-CoV555 showed similar improvements in viral ≥30 to <40 — no./total no. (%) 112/304 (36.8) 56/139 (40.3) clearance at day 3, with a difference from pla- ≥40 — no./total no. (%) 24/304 (7.9) 9/139 (6.5) cebo in the change from baseline of −0.42 (95% Risk factors for severe Covid-19 215 (69.6) 95 (66.4) CI, −0.89 to 0.06) for the 700-mg dose and −0.42 — no. (%)§ (95% CI, −0.90 to 0.06) for the 7000-mg dose. Disease status — no. (%) The difference from placebo in the change from baseline for the pooled doses of LY-CoV555 was Mild 232 (75.1) 113 (79.0) −0.49 (95% CI, −0.87 to −0.11). Moderate 77 (24.9) 30 (21.0) Median no. of days since onset 4.0 4.0 Exploratory Measures of Viral Clearance of symptoms In the pooled trial population, an association Mean viral load — Ct value¶ 23.9 23.8 was observed between slower viral clearance and * Covid-19 denotes coronavirus disease 2019. more hospitalization events. Figure 2A presents † Race or ethnic group was reported by the patients, who could choose more the absolute viral load among hospitalized pa- than one category. tients (pooled across randomization strata) as ‡ The body-mass index is the weight in kilograms divided by the square of the well as a box plot of viral loads among nonhos- height in meters. § Risk factors were an age of 65 years or older, a body-mass index of 35 or pitalized patients. On day 7, all the available more, or at least one coexisting illness in certain prespecified categories. measures of viral load among hospitalized pa- ¶ Ct denotes the cycle threshold of the reverse-transcriptase–polymerase-chain- tients were higher than the median values among reaction assay. the nonhospitalized patients. Among the patients with a higher viral load on day 7, the frequency of hospitalization was 12% (7 of 56 patients) is described in Section 6.10 of the statistical analy- among those who had a Ct value of less than sis plan. 27.5, as compared with a frequency of 0.9% (3 of 340 patients) among those with a lower viral Primary Outcome load. (The SARS-CoV-2 N1 gene primer determines By day 11, the majority of patients had a sub- a Ct value that is equivalent to approximately stantial trend toward viral clearance, including 570,000 nucleic acid–based amplification tests those in the placebo group. The observed mean per milliliter with the use of the SARS-CoV-2 decrease from baseline in the log viral load for reference panel of the Food and Drug Administra- the entire population was −3.81 (baseline mean, tion.) Since this difference was not anticipated 6.36; day 11 mean, 2.56); this value corresponded and emerged from post hoc exploratory analysis, to a decrease by more than a factor of 4300 in the it is unclear whether it would be applicable to SARS-CoV-2 burden, for an elimination of more other populations. Figure 2B shows the cumula- than 99.97% of viral RNA. For patients who re- tive probability that patients in each trial group ceived the 2800-mg dose of LY-CoV555, the dif- would have the indicated cycle threshold of viral ference from placebo in the decrease from base- load on day 7. 4 n engl j med nejm.org The New England Journal of Medicine Downloaded from nejm.org on November 19, 2020. For personal use only. No other uses without permission. Copyright © 2020 Massachusetts Medical Society. All rights reserved.

Neutralizing Antibody in Outpatients with Covid-19 Table 2. Change from Baseline in Viral Load. LY-CoV555 Placebo Difference Variable (N = 309) (N = 143) (95% CI) Primary outcome Mean change from baseline in viral load at day 11 −3.47 700 mg, −3.67 −0.20 (−0.66 to 0.25) 2800 mg, −4.00 −0.53 (−0.98 to −0.08) 7000 mg, −3.38 0.09 (−0.37 to 0.55) Pooled doses, −3.70 −0.22 (−0.60 to 0.15) Secondary outcomes* Mean change from baseline in viral load at day 3 −0.85 700 mg, −1.27 −0.42 (−0.89 to 0.06) 2800 mg, −1.50 −0.64 (−1.11 to −0.17) 7000 mg, −1.27 −0.42 (−0.90 to 0.06) Pooled doses, −1.35 −0.49 (−0.87 to −0.11) Mean change from baseline in viral load at day 7 −2.56 700 mg, −2.82 −0.25 (−0.73 to 0.23) 2800 mg, −3.01 −0.45 (−0.92 to 0.03) 7000 mg, −2.85 −0.28 (−0.77 to 0.20) Pooled doses, −2.90 −0.33 (−0.72 to 0.06) * Data regarding hospitalization, another key secondary outcome, are provided in Table 3. Covid-19–Related Hospitalization domains that were graded from 0 (no symp- At day 29, the percentage of patients who were toms) to 3 (severe symptoms). From day 2 to day hospitalized with Covid-19 was 1.6% (5 of 309 6, the change in the symptom score from base- patients) in the LY-CoV555 group and 6.3% (9 of line was better in the LY-CoV555 group than in 143 patients) in the placebo group (Table 3). The the placebo group, with values of −0.79 (95% CI, percentage of patients according to the LY-CoV555 −1.35 to −0.24) on day 2, −0.57 (95% CI, −1.12 dose who were hospitalized was similar to the to −0.01) on day 3, −1.04 (95% CI, −1.60 to −0.49) overall percentage, with 1.0% (1 of 101) in the on day 4, −0.73 (95% CI, −1.28 to −0.17) on day 5, 700-mg subgroup, 1.9% (2 of 107) in the 2800-mg and −0.79 (95% CI, −1.35 to −0.23) on day 6. The subgroup, and 2.0% (2 of 101) in the 7000-mg change from baseline in the symptom score con- subgroup. In a post hoc analysis examining hos- tinued to be better in the LY-CoV555 group than pitalization among patients who were 65 years in the placebo group from day 7 to day 11, al- of age or older and among those with a BMI of though by these time points most of the patients 35 or more, the percentage who were hospital- in the two groups had fully recovered or had ized was 4% (4 of 95) in the LY-CoV555 group and only very mild symptoms. 15% (7 of 48) in the placebo group. Only 1 patient in the trial (in the placebo group) was admitted Safety to an intensive care unit. Serious adverse events occurred in none of the 309 patients in LY-CoV555 group and in 0.7% Symptom Score (1 of 143 patients) in the placebo group (Ta- To assess the effect of treatment on Covid-19 ble 4). The percentage of patients who had an symptoms, we compared the change from base- adverse event during treatment was 22.3% (69 of line in symptom scores between the LY-CoV555 309) in the LY-CoV555 group and 24.5% (35 of 143) group and the placebo group (Fig. 3 and Fig. S1 in the placebo group. Diarrhea was reported in in the Supplementary Appendix). The symptom 3.2% of the patients (10 of 309) in the LY-CoV555 score ranged from 0 to 24 and included eight group and in 4.9% (7 of 143) in the placebo n engl j med nejm.org 5 The New England Journal of Medicine Downloaded from nejm.org on November 19, 2020. For personal use only. No other uses without permission. Copyright © 2020 Massachusetts Medical Society. All rights reserved.

The new england journal of medicine  Viral Load in ll Patients ported in 2.3% of the patients (7 of 309) in the LY-CoV555 group and in 1.4% (2 of 143) in the Nonhospitalized Hospitalized placebo group. Most of these events — which 10 included pruritus, flushing, rash, and facial 15 swelling — occurred during the infusion and 20 were reported as mild in severity. No changes in vital signs were noted during these reactions, 25 and the infusions were completed in all instanc- 30 es. In some patients, antihistamines were ad- Cycle Threshold35 ministered to help resolve symptoms. 40 We used standard methods to sequence all viral samples to determine the potential for re- 45 sistance-associated treatment failure. According- Day 1 Day 3 Day 7 Day 11 ly, we assessed the prevalence of variants with B Viral Load on Day 7 in Each Trial Group resistance to LY-CoV555 that were predicted in 1.00 preclinical studies. Such variants were present with an allele fraction of more than 20% in at least one sample at any time point in 8.2% of the patients in the LY-CoV555 group (6.3% in the 0.75 700-mg subgroup, 8.4% in the 2800-mg sub- group, and 9.9% in the 7000-mg subgroup) and in 6.1% of those in the placebo group. The clini- 0.50 cal importance of the presence of these variants is not known. LY-CoV555,700 Cumulative Probability0.25 LY-CoV555,200 LY-CoV555,7000 Discussion Placebo In this preplanned interim analysis of the 0.00 BLAZE-1 trial, we examined the efficacy of LY- 45 40 35 30 25 20 CoV555 in the treatment of mild or moderate Cycle Threshold Covid-19. The trial was designed to enroll pa- tients with a recent disease onset to evaluate the Figure 2. SARS-CoV-2 Viral Load in All Patients and According to Trial effect of early intervention with antibody therapy Group on Day 7. on viral-load biomarkers, symptoms, and severe Panel A shows the SARS-CoV-2 viral load (as measured by the cycle thresh- clinical outcomes, such as hospitalization and old on reverse-transcriptase–polymerase-chain-reaction assay) for all the death. patients who received either LY-CoV555 or placebo and for whom viral-load Among the patients who received LY-CoV555, data were available at the time of the interim analysis. The box plots indi- the viral load at day 11 (the primary outcome) cate the patients who were not hospitalized, and the red squares indicate those who were hospitalized. Such hospital contact was found to be associ- was lower than that in the placebo group only ated with a high viral load on day 7. The boxes represent interquartile rang- among those who received the 2800-mg dose. es, with the horizontal line in each box representing the median and the However, a decreased viral load at day 11 did not whiskers showing the minimum and maximum values (excluding outliers appear to be a clinically meaningful end point, that were more than 1.5 times the values represented at each end of the since the viral load was substantially reduced box). Panel B shows the cumulative probability that patients in each trial group would have the indicated cycle threshold of viral load on day 7. from baseline for the majority of patients, includ- ing those in the placebo group, a finding that was consistent with the natural course of the 20,21 group; vomiting was reported in 1.6% (5 of 309) disease. However, the evaluation of the effect and 2.8% (4 of 143), respectively. The most fre- of LY-CoV555 therapy on patients’ symptoms at quently reported adverse event in the LY-CoV555 earlier time points during treatment (e.g., on day 3) group was nausea (3.9%), whereas diarrhea (4.9%) showed a possible treatment effect, with no sub- was the most frequent adverse event in the pla- stantial differences observed among the three cebo group. Infusion-related reactions were re- doses. It is unclear whether RT-PCR is an accurate 6 n engl j med nejm.org The New England Journal of Medicine Downloaded from nejm.org on November 19, 2020. For personal use only. No other uses without permission. Copyright © 2020 Massachusetts Medical Society. All rights reserved.

Neutralizing Antibody in Outpatients with Covid-19 measure of viral neutralization, since viral RNA Table 3. Hospitalization.* may persist for some time even in the absence of replication-competent virus. Since the severity of Key Secondary LY-CoV555 Placebo Incidence illness is primarily driven by lung injury from Outcome SARS-CoV-2 infection in the lower respiratory no. of patients/total no. % tract, the viral load in the air spaces would be a Hospitalization 9/143 6.3 better reflection of the injury response than the viral load in nasopharyngeal secretions. However, 700 mg, 1/101 1.0 assessments of the lower respiratory tract were 2800 mg, 2/107 1.9 not practical owing to precautions that were re- 7000 mg, 2/101 2.0 quired in treating these highly infectious patients. Pooled doses, 1.6 Therefore, the nasopharyngeal viral swab was 5/309 the most pragmatic way of getting a sense of viral * Data for patients who presented to the emergency department are included in load as a surrogate marker of the viral load in this category. the lungs and to correlate with clinical outcomes. However, the nasopharyngeal viral load has not been validated as a predictor of clinical disease 0 course. An unanticipated observation in this trial was Delta Value (95% CI) that patients with a higher viral load on day 7 Day 2 −0.79 (−1.35 to −0.24) −2 Day 3 −0.57 (−1.12 to −0.01) had a higher rate of hospitalization than those Day 4 −1.04 (−1.60 to −0.49) Placebo with better clearance of viral RNA on day 7, a Day 5 −0.73 (−1.28 to −0.17) finding that was consistent with the delayed vi- Day 6 −0.79 (−1.35 to −0.23) ral clearance that was observed in patients with −4 Day  −0.50 (−1.06 to 0.07) 20,22,23 LY-CoV555 Day  −0.65 (−1.28 to −0.02) more severe disease. On day 7, no hospital- Change from Baseline Day 9 −0.15 (−0.75 to 0.45) ized patient had a viral load that was below the (pooled) Day  −0.32 (−0.94 to 0.29) median value of the population. If this observa- −6 Day  −0.44 (−1.02 to 0.15) tion is prospectively confirmed in future studies, it would suggest the potential for an agent that 1 2 3 4 5 6 7 8 9 10 11 lowers the viral load to reduce the rate of hospi- Trial Day talization. Figure 3. Symptom Scores from Day 2 to Day 11. To examine the potential of LY-CoV555 to im- Shown is the difference in the change from baseline (delta value) in symp- prove Covid-19 clinical outcomes, we evaluated tom scores between the LY-CoV555 group and the placebo group from day the effect of LY-CoV555 therapy on the frequency 2 to day 11. The symptom scores ranged from 0 to 24 and included eight of hospitalization, an important outcome given domains, each of which was graded on a scale of 0 (no symptoms) to 3 the association between hospitalization and sub- (severe symptoms). The I bars represent 95% confidence intervals. Details 23,24 about the symptom-scoring methods are provided in the Supplementary sequent mortality in patients with Covid-19. Appendix. On day 29, the percentage of patients who were hospitalized was 1.6% in the LY-CoV555 group and 6.3% in the placebo group. In a post hoc fect on symptoms and suggests a mechanistic analysis that was focused on high-risk subgroups link between a lower viral load and a lower fre- (an age of ≥65 years or a BMI of ≥35), the percent- quency of hospitalization. Although the differenc- age of hospitalization was 4.2% in the LY-CoV555 es in the effects of the three doses of LY-CoV555 group and 14.6% in the placebo group. were not clear, the 2800-mg dose was the only The data regarding symptoms (as measured one to show evidence of accelerated viral clear- by the change from baseline in the symptom ance. Nevertheless, further studies should con- score) were also consistent with the hospitaliza- tinue to assess the efficacy of lower doses. tion results, with findings that supported a pos- The safety profile of patients who received sible reduction in symptom severity as early as LY-CoV555 was similar to that of placebo-treated day 2 in the LY-CoV555 group. This effect was patients. These data indicate that the treatment maintained over time and across doses, which is safe. In this interim analysis, the patients who further supports the validity of a treatment ef- received LY-CoV555 had fewer hospitalizations n engl j med nejm.org 7 The New England Journal of Medicine Downloaded from nejm.org on November 19, 2020. For personal use only. No other uses without permission. Copyright © 2020 Massachusetts Medical Society. All rights reserved.

The new england journal of medicine Table 4. Adverse Events. LY-CoV555 Placebo Adverse Events (N = 309) (N = 143) 700 mg 2800 mg 7000 mg Pooled Doses (N = 101) (N = 107) (N = 101) (N = 309) number of patients (percent) Serious adverse events* 0 0 0 0 1 (0.7) Adverse events Any 24 (23.8) 23 (21.5) 22 (21.8) 69 (22.3) 35 (24.5) Mild 16 (15.8) 18 (16.8) 10 (9.9) 44 (14.2) 18 (12.6) Moderate 7 (6.9) 3 (2.8) 8 (7.9) 18 (5.8) 16 (11.2) Severe 0 2 (1.9) 3 (3.0) 5 (1.6) 1 (0.7) Missing data 1 (1.0) 0 1 (1.0) 2 (0.6) 0 Adverse events according to preferred term† Nausea 3 (3.0) 4 (3.7) 5 (5.0) 12 (3.9) 5 (3.5) Diarrhea 1 (1.0) 2 (1.9) 7 (6.9) 10 (3.2) 7 (4.9) Dizziness 4 (4.0) 3 (2.8) 3 (3.0) 10 (3.2) 3 (2.1) Headache 3 (3.0) 2 (1.9) 0 5 (1.6) 3 (2.1) Pruritus 2 (2.0) 3 (2.8) 0 5 (1.6) 1 (0.7) Vomiting 1 (1.0) 3 (2.8) 1 (1.0) 5 (1.6) 4 (2.8) Chills 0 1 (0.9) 3 (3.0) 4 (1.3) 0 Pyrexia 1 (1.0) 2 (1.9) 1 (1.0) 4 (1.3) 1 (0.7) Chest discomfort 1 (1.0) 1 (0.9) 1 (1.0) 3 (1.0) 1 (0.7) Fatigue 0 1 (0.9) 2 (2.0) 3 (1.0) 0 Hypertension 1 (1.0) 0 2 (2.0) 3 (1.0) 0 Lipase increased 1 (1.0) 0 2 (2.0) 3 (1.0) 0 Thrombocytosis 1 (1.0) 2 (1.9) 0 3 (1.0) 0 Blood pressure increased 2 (2.0) 0 0 2 (0.6) 0 Chest pain 1 (1.0) 1 (0.9) 0 2 (0.6) 0 Dyspepsia 1 (1.0) 0 1 (1.0) 2 (0.6) 0 Hypersensitivity 1 (1.0) 1 (0.9) 0 2 (0.6) 1 (0.7) Insomnia 0 1 (0.9) 1 (1.0) 2 (0.6) 0 Nasal congestion 1 (1.0) 1 (0.9) 0 2 (0.6) 1 (0.7) Rash 1 (1.0) 0 1 (1.0) 2 (0.6) 1 (0.7) Syncope 0 1 (0.9) 1 (1.0) 2 (0.6) 1 (0.7) * The serious adverse event in the placebo group was upper abdominal pain. There were no deaths during the trial. † The preferred terms were defined according to the Medical Dictionary for Regulatory Activities, version 23.0. and a lower symptom burden than those who Supported by Eli Lilly. received placebo, with the most pronounced ef- Disclosure forms provided by the authors are available with fects observed in high-risk cohorts. If these re- the full text of this article at NEJM.org. sults are confirmed in additional analyses in this A data sharing statement provided by the authors is available with the full text of this article at NEJM.org. trial, LY-CoV555 could become a useful treatment We thank Adam Clooney, Ph.D., of Eli Lilly, for providing med- for emergency use in patients with recently diag- ical-writing and other editorial support with an earlier version nosed Covid-19. of the manuscript. 8 n engl j med nejm.org The New England Journal of Medicine Downloaded from nejm.org on November 19, 2020. For personal use only. No other uses without permission. Copyright © 2020 Massachusetts Medical Society. All rights reserved.

Neutralizing Antibody in Outpatients with Covid-19 References 1. Tabata S, Imai K, Kawano S, et al. icitinib restrains the immune dysregula- 17. Baum A, Ajithdoss D, Copin R, et al. Clinical characteristics of COVID-19 in tion in severe COVID-19 patients. J Clin REGN-COV2 antibodies prevent and treat 104 people with SARS-CoV-2 infection on Invest 2020 August 18 (Epub ahead of SARS-CoV-2 infection in rhesus macaques the Diamond Princess cruise ship: a retro- print). and hamsters. Science 2020 October 9 spective analysis. Lancet Infect Dis 2020; 10. Stebbing J, Krishnan V, de Bono S, et (Epub ahead of print). 20: 1043-50. al. Mechanism of baricitinib supports ar- 18. Hoffmann M, Kleine-Weber H, Schro- 2. Jehi L, Ji X, Milinovich A, et al. Devel- tificial intelligence-predicted testing in eder S, et al. SARS-CoV-2 cell entry de- opment and validation of a model for in- COVID-19 patients. EMBO Mol Med 2020; pends on ACE2 and TMPRSS2 and is dividualized prediction of hospitalization 12(8): e12697. blocked by a clinically proven protease risk in 4,536 patients with COVID-19. 11. Dastan F, Saffaei A, Haseli S, et al. inhibitor. Cell 2020; 181(2): 271.e8-280.e8. PLoS One 2020; 15(8): e0237419. Promising effects of tocilizumab in 19. Jones BE, Brown-Augsburger PL, Cor- 3. Garg SKL, Kim L, Whitaker M, et al. COVID-19: a non-controlled, prospective bett KS, et al. LY-CoV555, a rapidly isolat- Hospitalization rates and characteristics clinical trial. Int Immunopharmacol ed potent neutralizing antibody, provides of patients hospitalized with laboratory- 2020; 88: 106869. protection in a non-human primate mod- confirmed coronavirus disease 2019 — 12. Farooqi F, Dhawan N, Morgan R, el of SARS-CoV-2 infection. October 1, COVID-NET, 14 states, March 1–30, 2020. Dinh J, Nedd K, Yatzkan G. Treatment of 2020 (https://www . biorxiv . org/ content/ MMWR Morb Mortal Wkly Rep 2020; 69: severe COVID-19 with tocilizumab miti- 10 . 1101/ 2020 . 09 . 30 . 318972v3). preprint. 458-64. gates cytokine storm and averts mechani- 20. Chen J, Qi T, Liu L, et al. Clinical pro- 4. Cavalcanti AB, Zampieri FG, Rosa cal ventilation during acute respiratory gression of patients with COVID-19 in RG, et al. Hydroxychloroquine with or distress: a case report and literature re- Shanghai, China. J Infect 2020; 80(5): e1- without azithromycin in mild-to-moder- view. Trop Med Infect Dis 2020; 5: E112. e6. ate Covid-19. N Engl J Med. DOI: 1 0.1056/ 13. The RECOVERY Collaborative Group. 21. He X, Lau EHY, Wu P, et al. Temporal NEJMoa2019014. Dexamethasone in hospitalized patients dynamics in viral shedding and transmis- 5. Beigel JH, Tomashek KM, Dodd LE. with Covid-19 — preliminary report. N Engl sibility of COVID-19. Nat Med 2020; 26: Remdesivir for the treatment of Covid-19 J Med. DOI: 10.1056/NEJMoa2021436. 672-5. — final report. N Engl J Med. DOI: 14. Sterne JAC, Murthy S, Diaz JV, et al. 22. Liu Y, Yan L-M, Wan L, et al. Viral 10.1056/NEJMoa2007764. Association between administration of sys- dynamics in mild and severe cases of 6. Goldman JD, Lye DCB, Hui DS, et al. temic corticosteroids and mortality among COVID-19. Lancet Infect Dis 2020; 20: 656- Remdesivir for 5 or 10 days in patients critically ill patients with COVID-19: a 7. with severe Covid-19. N Engl J Med. DOI: meta-analysis. JAMA 2020; 324: 1-13. 23. Zhou F, Yu T, Du R, et al. Clinical 10.1056/NEJMoa2015301. 15. Joyner MJ, Senefeld JW, Klassen SA, et course and risk factors for mortality of 7. Spinner CD, Gottlieb RL, Criner GJ, et al. Effect of convalescent plasma on mor- adult inpatients with COVID-19 in Wu- al. Effect of remdesivir vs standard care tality among hospitalized patients with han, China: a retrospective cohort study. on clinical status at 11 days in patients COVID-19: initial three-month experi- Lancet 2020; 395: 1054-62. with moderate COVID-19: a randomized ence. August 12, 2020 (https://www 24. Kim L, Garg S, O’Halloran A, et al. clinical trial. JAMA 2020; 324: 1048-57. . medrxiv . org/ content/ 10 . 1101/ 2020 . 08 . 12 Risk factors for intensive care unit admis- 8. Sims JT, Krishnan V, Chang C-Y, et al. . 20169359v1). preprint. sion and in-hospital mortality among hos- Characterization of the cytokine storm 16. Li L, Zhang W, Hu Y, et al. Effect of pitalized adults identified through the U.S. reflects hyperinflammatory endothelial convalescent plasma therapy on time to Coronavirus Disease 2019 (COVID-19)- dysfunction in COVID-19. J Allergy Clin clinical improvement in patients with se- Associated Hospitalization Surveillance Immunol 2020 September 10 (Epub ahead vere and life-threatening COVID-19: a Network (COVID-NET). Clin Infect Dis of print). randomized clinical trial. JAMA 2020; 2020 July 16 (Epub ahead of print). 9. Bronte V, Ugel S, Tinazzi E, et al. Bar- 324: 460-70. Copyright © 2020 Massachusetts Medical Society. n engl j med nejm.org 9 The New England Journal of Medicine Downloaded from nejm.org on November 19, 2020. For personal use only. No other uses without permission. Copyright © 2020 Massachusetts Medical Society. All rights reserved.