Real-World Experience of Bamlanivimab for Coronavirus Disease 2019 (COVID-19): A Case-Control Study.

BACKGROUND: Coronavirus disease 2019 (COVID-19) has strained healthcare systems with patient hospitalizations and deaths. Anti-spike monoclonal antibodies, including bamlanivimab, have demonstrated reduction in hospitalization rates in clinical trials, yet real-world evidence is lacking.
METHODS: We conducted a retrospective case-control study across a single healthcare system of nonhospitalized patients, age 18 years or older, with documented positive severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) testing, risk factors for severe COVID-19, and referrals for bamlanivimab via emergency use authorization. Cases were defined as patients who received bamlanivimab; contemporary controls had a referral order placed but did not receive bamlanivimab. The primary outcome was 30-day hospitalization rate from initial positive SARS-CoV-2 polymerase chain reaction (PCR). Descriptive statistics, including χ 2 and Mann-Whitney U test, were performed. Multivariable logistic regression was used for adjusted analysis to evaluate independent associations with 30-day hospitalization.
RESULTS: Between 30 November 2020 and 19 January 2021, 218 patients received bamlanivimab (cases), and 185 were referred but did not receive drug (controls). Thirty-day hospitalization rate was significantly lower among patients who received bamlanivimab (7.3% vs 20.0%, risk ratio [RR] 0.37, 95% confidence interval [CI]: .21-.64, P < .001), and the number needed to treat was 8. On logistic regression, odds of hospitalization were increased in patients not receiving bamlanivimab and with a higher number of pre-specified comorbidities (odds ratio [OR] 4.19 ,95% CI: 1.31-2.16, P < .001; OR 1.68, 95% CI: 2.12-8.30, P < .001, respectively).
CONCLUSIONS: Ambulatory patients with COVID-19 who received bamlanivimab had a lower 30-day hospitalization than control patients in real-world experience. We identified receipt of bamlanivimab and fewer comorbidities as protective factors against hospitalization.Bamlanivimab's role in preventing hospitalization associated with coronavirus disease 2019 (COVID-19) remains unclear. In a real-world, retrospective study of 403 high-risk, ambulatory patients with COVID-19, receipt of bamlanivimab compared to no monoclonal antibody therapy was associated with lower 30-day hospitalization.

As of 28 February 2021, the World Health Organization (WHO) reported 113 467 303 cases of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and 2 520 550 related deaths worldwide [1]. The United States has had 28 605 527 total cases and 513 091 deaths [2]. Despite the human toll, the only coronavirus disease 2019 (COVID-19) therapy approved by the US Food and Drug Administration (FDA) remains remdesivir, a viral RNA-dependent RNA polymerase with approval restricted to hospitalized COVID-19 patients [3].

Greater risk for SARS-CoV-2-related mortality and hospitalization is associated with various demographics, comorbidities, and social determinants of health. These factors include male gender, older age, lower socioeconomic status, body mass index (BMI) ≥30 kg/m2, diabetes, hypertension, cardiovascular disease (CVD), chronic kidney disease (CKD), severe asthma, immunosuppression (including organ transplantation), and race (Black or other minority ethnic groups) [4-7]. There are currently no FDA-approved therapies to prevent progression of mild to moderate COVID-19 in high-risk groups. Neutralizing monoclonal antibodies have been granted emergency use authorization (EUA) by the FDA for mild to moderate COVID-19, but more information is needed on their effect in high-risk patients.

Neutralizing monoclonal antibodies, developed from the cells from convalescent COVID-19 patients, bind to SARS-CoV-2’s spike protein and prevent viral attachment to human surface ACE2 receptors [8]. These agents include bamlanivimab (LY-CoV555, Eli Lilly and Company), etesevimab (LY-CoV016, Eli Lilly and Company), and casirivimab/imdevimab (REGN-COV2, Regeneron Pharmaceuticals). Preliminary analysis of a double-blind phase 1–3 trial of 275 nonhospitalized patients with COVID-19 showed reduction in viral load with casirivimab/imdevimab [9]. Preliminary analysis of bamlanivimab at varying exploratory doses (700 mg, 2800 mg, 7000 mg) resulted in lower rates of COVID-19-related hospitalization or emergency department (ED) visit within 29 days compared to placebo (1.6% vs 6.3%) for all doses [10]. Post hoc analysis demonstrated that high-risk patients who received bamlanivimab had a reduced rate of hospitalizations or ED visits compared to their counterparts who received placebo (4.2% vs 14.6%) [10].

As a result of initial trial findings, neutralizing monoclonal antibodies were granted EUA by the FDA for mild to moderate COVID-19, and in November 2020 the US government purchased and distributed 300 000 doses of bamlanivimab [11, 12]. Upon further study, bamlanivimab monotherapy demonstrated a non-statistically significant decrease in hospitalization [13]. However, the potential benefit of bamlanivimab as monotherapy in high-risk patient populations remains unclear, and more information is needed regarding the use of neutralizing monoclonal antibodies in clinical practice.

This study describes a single center, real-world experience of the use of bamlanivimab in high-risk ambulatory patients with mild to moderate COVID-19 and its impact on hospitalization rate.

METHODS

Study Design

We conducted an unblinded retrospective case-control study across a healthcare system with 10 hospitals and over 200 inpatient and outpatient sites in the greater Chicago area. Patients, aged 18 years or older, were included if they had documented COVID-19 by antigen or polymerase chain reaction (PCR) test between 11 November 2020 and 19 January 2021, met clinical criteria as high risk or were healthcare workers, and were referred for EUA bamlanivimab. Cases were defined as receiving bamlanivimab within the healthcare system. Contemporary controls were patients aged 18 years or older with documented COVID-19 by antigen or PCR test between 11 November 2020 and 19 January 2021, met clinical criteria as high risk or were healthcare workers, and were referred for EUA bamlanivimab but did not receive any neutralizing monoclonal antibody therapy for COVID-19. Patients were excluded if their first positive test occurred prior to 11 November 2020 or if they complained of symptoms for 15 or more days prior to testing. Patients were also excluded if they received either bamlanivimab or casirivimab/imdevimab at an outside healthcare system.

Identification of Cases and Controls

Cases and controls were identified retrospectively using a healthcare system-wide centralized electronic data warehouse (Northwestern Medicine Electronic Data Warehouse [NMEDW]) and an EUA bamlanivimab referral report within the electronic health record (EHR, Epic Systems), respectively. Cases and controls were verified along with collection of data on covariates and outcomes (detailed below) by study team chart review. All included patients had documentation of positive SARS-CoV-2 testing and a referral order placed for bamlanivimab infusion administration during the study period. Cases had documented receipt of bamlanivimab via EHR administration report. Controls did not receive bamlanivimab as noted by lack of documented administration of monoclonal antibody therapy and cancelled infusion appointment within EHR.

Distribution of Bamlanivimab

On 20 November 2020, the subject healthcare system began administering bamlanivimab (single infusion of 700 mg). A multidisciplinary committee used the EUA to create eligibility criteria to guide clinicians. Prescribing clinicians attested in the bamlanivimab referral order that the patient (1) was not hospitalized at time of order placement, (2) had no acute need for oxygen therapy or, for those patients on oxygen at baseline, no increase in oxygen requirement, (3) was age 18 years or older, (4) had positive SARS-CoV-2 testing, (5) was <10 days from initial symptom onset, and (6) was <5 days from first positive SARS-CoV-2 test and whether the patient weighed ≥40 kg. To address the anticipated demand exceeding supply of bamlanivimab, an internal, non-validated, cumulative priority system sum (CPSS) was created, allocating 1 point for each of the following EUA indications: BMI ≥35 kg/m2, CKD, diabetes, immunosuppressed, ≥65 years of age, or ≥55 years of age with CVD, hypertension, or chronic obstructive pulmonary disease (COPD)/chronic respiratory disease. However, clinical demand for the drug did not exceed supply; thus, the CPSS was not employed for infusion appointment scheduling.

Prescribing clinicians counseled eligible patients on potential risks of monoclonal antibody therapy and the administration process per the EUA. Bamlanivimab was administered in select emergency departments upon initial presentation but only after the decision to discharge the patient, whereas other ambulatory patients were referred by clinicians and scheduled through a centralized process. Eligible patients had a 10-day window to schedule and complete bamlanivimab infusion appointments prior to expiration of their referral.

Definition of Immunosuppression and COVID-19 Severity

Immunosuppression was defined as active malignancy, presence of solid organ or stem cell transplantation, living with human immunodeficiency virus (HIV) (regardless of CD4 count), or autoimmune disease requiring immunosuppressive therapy.

Severity of illness at time of first healthcare contact was determined based on chart review. Patients were considered to have mild illness if there was documentation of mild symptoms (eg, fever, cough, or change in taste or smell) without dyspnea [14]. Moderate illness was defined as clinical or radiographic evidence of lower respiratory tract disease with oxygen saturation ≥94%. Severe illness was defined as oxygen saturation <94%, respiratory rate ≥30 breaths/min, or lung infiltrates >50% [14].

Assessment of Primary and Secondary Outcomes

The primary outcome was hospitalization within 30 days from initial positive SARS-CoV-2 PCR. Demographics, comorbidities, and presentation characteristics were compared between cases and controls. Secondary outcomes included: intensive care unit (ICU) admission, mechanical ventilation requirement, mortality, and duration of hospitalization within 30 days from first positive SARS-CoV-2 test.

Data Management and Statistical Analysis

Data were obtained via the NMEDW and EHR chart review. Study data were collected and managed using Research Electronic Data Capture (REDCap) electronic data capture tools hosted at Northwestern [15, 16].

A χ 2 test was used to compare categorical values for cases and controls. Mann-Whitney U test was used to compare median values for cases and controls. Binary logistic regression was used for multivariable analysis. Variables included in the logistic regression have previously been associated with hospitalization and poor outcomes: gender (male), race (Black), Hispanic ethnicity, age, and total number of comorbidities (TNC)—which was composed of the sum of the following comorbidities (regardless of age): BMI ≥35 kg/m2, immunosuppression, diabetes, CKD, hypertension, CVD, and lung disease [4, 5, 7]. A P-value ≤ .05 was defined as statistically significant. All statistics were performed using IBM SPSS version 26.0 (Armonk, NY, USA).

Ethical Considerations

This study was performed with data collected after institutional review board approval.

RESULTS

Between 20 November 2020 and 19 January 2021, 218 patients received bamlanivimab, and 185 patients were referred for infusion but did not receive monoclonal antibody therapy (Figure 1). Reasons for the 185 patients not receiving bamlanivimab were not well documented with 130 (70.2%) patients having expired/cancelled infusion appointments for undocumented reasons. Twenty-five patients (12.3%) declined therapy; 12 patients (5.9%) had symptoms for more than 10 days at time of prescription; 8 patients (4.3%) presented to infusion appointment with severe illness/hypoxia on preadministration assessment; 6 patients (2.9%) had a test >5 days old at time of prescription; and 4 patients (2.2%) had severe disease at time of first healthcare contact and were not given bamlanivimab.

Figure 1.

Study flowchart for inclusion. Abbreviation: mAb, monoclonal antibody.

Patients who received bamlanivimab had a higher median age (P = .008) and CPSS than those in the control group (P < .001) (Table 1). The bamlanivimab group included a greater proportion of immunosuppressed patients (odds ratio [OR]: 1.68, 95% confidence interval [CI]: 1.07–2.64, P = .024) and patients ≥55 years with chronic lung disease (OR: 2.19, 95% CI: 1.11–4.33, P = .021). The rates of other comorbid conditions by EUA indication (BMI, diabetes, CKD, or ≥55 years with hypertension or CVD) were similar. Gender distribution was comparable between groups.

Table 1.

Baseline Characteristics of 403 Coronavirus Disease 2019 (COVID-19) Patients Between 11 November 2020 and 19 January 2021

Characteristics	Total Study Population (n = 403)	Bamlanivimab Therapy (n = 218)	No Bamlanivimab (n = 185)	P value
Age
Median age (IQR)	64 (53–73)	66 (57–74)	62 (50–72)	.008a
≥65 years (%)	199 (49.4%)	117 (53.6%)	82 (44.3%)	.061
Male	210 (52.1%)	115 (52.8%)	95 (51.4%)	.779
Self-identified race/ethnicity
White	295 (73.2%)	173 (79.4%)	122 (65.9%)	.002
Hispanic	61 (15.1%)	25 (11.5%)	36 (19.5%)	.026
Black	47 (11.7%)	13 (6.0%)	34 (18.4%)	<.001
Asian	14 (3.5%)	9 (4.1%)	5 (2.7%)	.436
Native American	2 (0.5%)	0	2 (1.1%)	.124
Native Hawaiian/Pacific Islander	1 (0.2%)	0	1 (0.5%)	.277
Other	32 (7.9%)	17 (7.8%)	15 (8.1%)	.909
Declined	12 (3.0%)	6 (3.0%)	6 (3.2%)	.773
Primary language spoken
English	369 (91.6%)	207 (95.0%)	162 (87.6%)	.008
Spanish	30 (7.4%)	11 (5.0%)	19 (10.3%)	.046
Other	4 (1.0%)	0 (0%)	4 (2.1%)	.029
Insurance status
Private	175 (43.4%)	94 (43.1%)	124 (56.9%)	.893
Medicare	186 (46.2%)	109 (50%)	77 (41.6%)	.093
Medicaid	27 (6.7%)	13 (6.0%)	14 (7.6%)	.521
Self-pay	14 (3.5%)	2 (0.9%)	12 (6.5%)	.002
Comorbid conditions by EUA indication
BMI
Median—kg/m2 (IQR)	31.15 (26.68–36.53)	30.45 (26.94–36.08)	31.92 (26.46–37.79)	.54a
BMI ≥ 35	122 (30.3%)	65 (29.8%)	57 (30.8%)	.829
Diabetes	120 (29.8%)	73 (33.5%)	47 (25.4%)	.077
CKD	34 (8.4%)	20 (9.2%)	14 (7.6%)	.563
Age ≥55 years old with hypertension	191 (47.4%)	110 (50.5%)	81 (43.8%)	.181
Age ≥55 years old with CVD	79 (19.6%)	47 (21.6%)	32 (17.3%)	.283
Age ≥55 years old with CLD	44 (10.9%)	31 (14.2%)	13 (3.2%)	.021
Immunosuppressed	109 (27.0%)	69 (31.7%)	40 (21.6%)	.024
Median CPSS (IQR)	2 (1–3)	2 (1–3)	2 (1–2)	.009

P ≤ .05 defined as significant.

Abbreviations: BMI, body mass index; CKD, chronic kidney disease; CLD, chronic lung disease; CPSS, cumulative priority system sum; CVD, cardiovascular disease; EUA, emergency use authorization; IQR, interquartile range.

aMann-Whitney U test performed to assess statistical significance.

Of the 403 patients, 15.1% self-identified as ethnically Hispanic. White patients were more likely to receive bamlanivimab compared to all other races (OR: 1.99, 95% CI: 1.27–3.11, P = .002), and patients whose primary language was English were more likely to receive bamlanivimab compared to other primary languages (OR: 2.67, 95% CI: 1.27–5.64, P = .008). Odds of receiving bamlanivimab were significantly decreased for patients who identified as Hispanic (OR: 0.54, 95% CI: .31–.93, P = .026) or Black (OR: 0.28, 95% CI: .14–.55, P < .001). In addition, self-pay patients comprised 0.9% of patients who received bamlanivimab and 6.5% of the patients who did not receive bamlanivimab (P = .002).

The most reported and documented symptoms of COVID-19 illness at initial presentation were cough (65.8%), fever (42.3%), myalgias (37.7%), and fatigue (34.8%) (Table 2). Compared to controls, there was a smaller proportion of patients presenting to first healthcare contact with asymptomatic or severe disease in the bamlanivimab group. Characteristics of the patients who presented with severe COVID-19 are presented in Supplementary Table 1. Almost half (45.5%) of all patients receiving bamlanivimab presented initially to an ambulatory clinic, either in person, as telehealth visit or via EHR-integrated messaging (P = .036). In contrast, 40.5% of patients who did not receive bamlanivimab presented initially to the emergency department (P < .001). Labs were inconsistently performed at time of diagnosis, but when performed, only lactic acid dehydrogenase was higher in patients who did not receive bamlanivimab.

Table 2.

Symptoms and Signs of Coronavirus Disease 2019 (COVID-19) in 403 Patients

Characteristics	Total Study Population(n = 403)	Bamlanivimab (n = 218)	No Bamlanivimab (n = 185)	P value
Duration of symptoms before presentation, median days (IQR)	2 (1–4)	2 (1–4)	3 (1–5)	<.001
Duration of symptoms before bamlanivimab administration, median days (IQR)	…	5 (4–7)	…	…
Symptoms
Cough	265 (65.8%)	138 (63.3%)	127 (68.6%)	.260
Fever	171 (42.3%)	84 (38.5%)	87 (47.0%)	.086
Myalgias	152 (37.7%)	75 (34.4%)	77 (41.6%)	.136
Fatigue	140 (34.8%)	64 (29.4%)	76 (41.1%)	.014
Headache	120 (29.8%)	63 (28.9%)	57 (30.8%)	.676
Congestion	114 (28.3%)	73 (33.5%)	41 (22.1%)	.012
Dyspnea	100 (24.8%)	43 (19.7%)	57 (30.8%)	.010
Chills	92 (22.8%)	48 (22.0%)	44 (23.8%)	.674
Rhinorrhea	61 (15.1%)	43 (19.7%)	18 (9.7%)	.005
Sore throat	74 (18.4%)	40 (18.3%)	34 (18.4%)	.994
Diarrhea	66 (16.4%)	30 (13.8%)	36 (19.5%)	.123
Nausea	51 (12.7%)	25 (11.5%)	26 (14.1%)	.436
Anosmia	38 (9.4%)	16 (7.3%)	22 (11.9%)	.119
Hypogeusia	37 (9.2%)	15 (6.9%)	22 (11.9%)	.083
Malaise	37 (9.2%)	16 (7.3%)	21 (11.4%)	.165
Anorexia	32 (7.9%)	13 (6.0%)	19 (10.3%)	.111
Chest pain	30 (7.4%)	9 (4.1%)	21 (11.4%)	.006
Chest tightness	27 (6.7%)	17 (7.8%)	10 (5.4%)	.338
Arthralgias	26 (6.5%)	11 (5.0%)	15 (8.1%)	.212
Vomiting	15 (3.7%)	7 (3.2%)	8 (4.3%)	.556
Disease severity [14]
Asymptomatic	15 (3.7%)	4 (1.8%)	11 (5.9%)	.030
Mild	287 (71.2%)	162 (74.3%)	125 (67.6%)	.136
Moderate	75 (18.6%)	36 (16.5%)	39 (21.1%)	.240
Severe	4 (1.0%)	0	4 (2.2%)	.029
Unable to assess	22 (5.4%)	16 (4.0%)	6 (3.2%)	.071
Initial presentation
Ambulatory clinic	164 (41.0%)	99 (45.4%)	65 (35.1%)	.036
Emergency department	125 (31.0%)	50 (22.9%)	75 (40.5%)	<.001
Urgent care	64 (15.9%)	38 (17.4%)	26 (14.1%)	.355
Out of system location	36 (8.9%)	21 (9.6%)	15 (8.1%)	.593
Known exposure to COVID-19	187 (46.4%)	98 (45.0%)	89 (48.1%)	.527
Physical exam: abnormal lung exam	36 (8.9%)	18 (8.3%)	18 (9.7%)	.605
Imaging studyb: abnormal lung imaging	89 (22.1%)	35 (16.0%)	54 (29.2%)	.002
Laboratory values
C-reactive protein,c mg/L, median (IQR)	31.21 (9.4375–72.05)	23.2 (7.3–51.5)	32.1 (10.87–87)	.378a
Ferritin,d mcg/L, median (IQR)	209.55 (114.625- 670.95)	178 (123- 605.65)	286.1 (106.45- 670.90)	.568a
Lactic acid dehydrogenase,e U/L, median (IQR)	238 (199–286)	211 (188.5–263)	262 (207.75–319)	.047a
Absolute lymphocyte count,f 103 cells/μL, median (IQR)	1.1 (0.9–2.15)	1.05 (0.8–1.675)	1.2 (1–2.5)	.114a

P ≤ .05 defined as significant.

Abbreviation: IQR, interquartile range.

aMann-Whitney U test performed to assess statistical significance.

bOne hundred thirty-six patients had imaging performed, 61 in bamlanivimab group, 75 in no bamlanivimab group.

cFifty-eight patients had lab performed: 21 patients in bamlanivimab group, 37 patients in no bamlanivimab group.

dFifty-four patients had lab performed: 19 patients in bamlanivimab group, 35 patients in no bamlanivimab group.

eFifty-seven patients had lab performed: 19 patients in bamlanivimab group, 38 patients in no bamlanivimab group.

f One hundred twenty-seven patients had lab performed: 56 patients in bamlanivimab group, 71 patients in no bamlanivimab group.

The 30-day hospitalization rate was 7.3% of patients who received bamlanivimab compared to 20.0% of patients who did not receive bamlanivimab (RR 0.37, 95% CI: .21–.64, P < .001, Table 3). The number needed to treat (NNT) to prevent 1 hospitalization was 8. Median time from first positive test to hospitalization was 8 days for both patients who received bamlanivimab and patients who did not. Of the 218 patients who received bamlanivimab, 2 patients required ICU admission, 1 was intubated, and 1 died. Of the 185 patients who did not receive bamlanivimab, 5 required ICU admission, 4 were intubated, and 4 died. The study population was insufficiently powered to determine statistical significance of ICU admission, intubation, and mortality at 30 days.

Table 3.

Outcomes of 403 Ambulatory Patients With Coronavirus Disease 2019 (COVID-19)

Outcome	Total Study Population (n = 403)	Bamlanivimab Therapy (n = 218)	No Bamlanivimab (n = 185)	P value
Number of patients hospitalized within 30 days of COVID-19 diagnosis	53 (13.2%)	16 (7.3%)	37 (20.0%)	<.001
Duration of hospitalization, median days (IQR)	5 (3–8)	4 (2.75–8)	5 (4–8)	.572a
Number of patients requiring ICU admission	7 (1.7%)	2 (0.9%)	5 (2.7%)	.892
Number of patients requiring mechanical ventilation	5 (1.2%)	1 (0.5%)	4 (2.1%)	.427
Deaths	5 (1.2%)	1 (0.5%)	4 (2.1%)	.124

P ≤ .05 defined as significant.

Abbreviations: ICU, intensive care unit; IQR, interquartile range.

aMann-Whitney U test performed to assess statistical significance.

Kaplan-Meier estimate evaluated time to hospitalization (Figure 2). By log rank test, there was a statistically significant difference in survival distribution for those who received bamlanivimab, χ 2 (1) = 14.48, P < .001. Univariate predictors for individual factors significantly associated with hospitalization included: symptomatic or severe disease at first healthcare presentation, median CPSS, presence of CKD, age ≥55 years with hypertension, age ≥55 years with CVD, and age ≥55 years with chronic lung disease (Table 4).

Figure 2.

Kaplan-Meier of time to hospitalization after SARS-CoV-2 test. Abbreviation: SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

Table 4.

Univariate Predictors of Hospitalization for Coronavirus Disease 2019 (COVID-19)

Characteristics	Total Study Population (n = 403)	Hospitalized (n = 53)	Not Hospitalized (n = 350)	P value
Male	210 (52.1%)	30 (56.6%)	180 (51.4%)	.482
Disease severity at presentation [14]
Asymptomatic	15 (3.7%)	5 (9.4%)	10 (2.9%)	.018
Mild	287 (71.2%)	33 (62.3%)	254 (72.6%)	.122
Moderate	75 (18.6%)	13 (24.5%)	62 (17.7%)	.235
Severe	4 (1.0%)	2 (3.8%)	2 (0.6%)	.028
Unable to assess	22 (5.4%)	0	22 (6.3%)	.060
Comorbid conditions by EUA indication
BMI
BMI– kg/m2 median (IQR)	31.15 (26.68–36.53)	32.31 (27.48–40.46)	31.09 (26.63–36.33)	.194a
BMI ≥35	122 (30.3%)	19 (35.8%)	103 (29.4%)	.343
Diabetes	120 (29.8%)	17 (32.1%)	103 (29.4%)	.695
CKD	34 (8.4%)	9 (17.0%)	25 (7.1%)	.016
≥65 years old	199 (49.4%)	26 (49.1%)	173 (49.4%)	.960
≥55 years old with hypertension	191 (47.4%)	34 (64.2%)	157 (44.9%)	.009
≥55 years old with CVD	79 (19.6%)	16 (30.2%)	63 (18.0%)	.037
≥55 years old with CLD	44 (10.9%)	11 (20.8%)	33 (9.4%)	.014
Immunosuppressed	109 (27.0%)	17 (32.1%)	92 (26.3%)	.377
Median CPSS	2 (1–3)	2 (2–3)	2 (1–2)	<.001a

P ≤ .05 defined as significant.

Abbreviations: BMI, body mass index; CKD, chronic kidney disease; CLD, chronic lung disease; CPSS, cumulative priority system sum; CVD, cardiovascular disease; EUA, emergency use authorization; IQR, interquartile range.

aMann-Whitney U test performed to assess statistical significance.

We performed multivariate logistic regression to determine independent association between age, gender, Black race, Hispanic ethnicity, bamlanivimab use, and TNC on the likelihood of hospitalization. This model was statistically significant (χ 2 (6) = 36.760, P < .001). Not receiving bamlanivimab and TNC had increased odds of hospitalization, 4.19 times (P < .001) and 1.68 times (P < .001), respectively. All other components of the model were not statistically significant (Table 5). Multivariate analyses with the addition of severity at time of presentation and/or insurance status as a marker for socioeconomic status were also performed; in each of these models receipt of bamlanivimab and TNC were the only statistically significant variables (Supplemental Tables 2, 3, 4). Across these models, patients who did not receive bamlanivimab had about a 4-fold greater odds of being hospitalized (Supplemental Tables 2, 3, 4).

Table 5.

Multivariate Predictors of Hospitalization in Patients With Coronavirus Disease 2019 (COVID-19)

	B	SE	Wald	df	P value	Exp(B)	95% CI
Male	0.22	0.32	0.47	1	.493	1.24	.67–2.32
Black race	0.38	0.44	0.76	1	.385	1.47	.62–3.49
Hispanic ethnicity	−0.32	0.47	0.46	1	.497	0.73	.29–1.82
Control (no bamlanivimab)	1.4	0.35	16.92	1	<.001	4.19	2.12–8.30
Age at presentation	0.01	0.01	1.16	1	.282	1.01	.99–1.03
TNC	0.52	0.13	16.57	1	<.001	1.68	1.31–2.16

P ≤ .05 defined as significant.

Abbreviations: CI, confidence interval; SE, standard error; TNC, total number comorbidities.

DISCUSSION

In our single center report, patients who received bamlanivimab had significantly lower 30-day hospitalization rate compared to controls, even after adjusting for demographics and comorbid medical conditions. Compared to prior clinical trial data evaluating bamlanivimab or casirivimab/imdevimab, our overall study population was at a higher baseline risk for poor clinical outcomes and hospitalization due to COVID-19 [9, 10]. This higher risk was driven by increased representation of patients with advanced age (median age 64 and 49% over the age of 65), a higher median BMI (31.15 kg/m2), a substantial immunosuppressed population (109, 27.0%), and a higher percentage of patients who self-identified as male (52.1%) and Black race (11.7%). Almost all patients (98%) included in this study featured at least 1 predefined risk factor for severe COVID-19 per EUA criteria, in contrast to 69.6% and 65% of subjects, in BLAZE-1 and REGN-COV2, respectively [9, 10]. Cases and controls in our study were more frequently hospitalized compared to those in clinical trials, which may reflect a combination of the higher risk population and the less intensive clinical outpatient monitoring that would be expected with real-world experience compared to clinical trials.

Based on our results, 8 patients with mild to moderate COVID-19 would need to be treated to prevent 1 hospitalization, compared to a NNT of 20 patients in BLAZE-1 [10]. This lower NNT is likely reflective of the greater benefit in patients with greater comorbidities and risk factors present in our population compared to prior study populations.

Our findings indicate that bamlanivimab monotherapy was associated with a significant reduction in hospitalization among high-risk patients in the outpatient setting. By adding real-world experience to what has been demonstrated in clinical trials, these results provide clinicians with increased confidence that bamlanivimab may be effective in preventing hospitalization, especially in those with underlying medical conditions. With limited therapeutic options for nonhospitalized patients with COVID-19, monoclonal antibodies provide a viable treatment option for high-risk patients when administered within 10 days of symptom onset.

The major limitation of our study was that data collection was confined by documentation in the EHR; this prevented the ability to discern if hospitalizations or ED visits occurred outside of the Epic System EHR. Information reported by the patient may not have been comprehensively documented by clinicians within the patient’s chart. However, we would expect this limitation to impact those receiving and not receiving bamlanivimab equally, especially as all individuals in this study had bamlanivimab orders, which included an attestation that they met EUA criteria.

Another limitation was the inclusion of patients with severe COVID-19. Given that the EUA-specified use of bamlanivimab for mild to moderate COVID-19, no patients in the bamlanivimab group had severe illness at the time of presentation, nor did they have severe illness at time of bamlanivimab administration. Although both asymptomatic and severe illness at the time of first healthcare presentation were independent predictors of 30-day hospitalization, we felt it important to include these patients in the primary analysis to reflect the real-world, clinical practice at our hospital system. To reduce bias associated with severe cases at first healthcare contact among controls, analysis with severity was performed (Supplementary Tables 2 and 4). Bamlanivimab use remained associated with a lower rate of hospitalization. Furthermore, those individuals within the control group that progressed from initial mild-to-moderate COVID-19 to severe before bamlanivimab could be administered are reflective of the expected trajectory of the disease. It raises the possibility that expedited systems to provide infusions with fewer logistical challenges may provide additional benefit and further reduce the number of hospitalized patients.

Additionally, safety/tolerability data were not formally assessed due to the limitation of clinical documentation, but no adverse events requiring hospitalization were reported. Three internal patient safety reports noted infusion-associated headache events, all of which resolved with supportive care. Antigen and PCR testing were performed on a variety of platforms, only a few of which were able to generate cycle threshold values. We were thus unable to compare baseline and subsequent mean viral loads in the 2 groups.

Finally, this study also had a disproportionate number of non-Hispanic, white patients. Although race and ethnicity were accounted for in the multivariable regression analysis, there are likely other determinants of health that disproportionately affect minorities and may not be accurately captured by our study. We found that white, English-speaking patients were more likely to receive bamlanivimab, despite there being Spanish language material easily accessible to providers and patients. This highlights healthcare inequities like those reported throughout the COVID-19 pandemic. Further studies involving racial and ethnic minorities are needed to better understand access barriers to these therapies and bamlanivimab’s impact across the wider population.

Our findings support the use bamlanivimab therapy for ambulatory patients with mild to moderate COVID-19 to prevent hospitalization. To our knowledge, this is the first report of clinical outcomes associated with bamlanivimab use under EUA, outside of clinical trials. These findings are reflective of real-world experiences that capture practical barriers and challenges that may not be present in structured clinical trials. Our findings should be used to identify high-risk, ambulatory patients who should be considered for EUA monoclonal antibody treatment to prevent hospitalization associated with severe COVID-19. Future areas for study include the use of monoclonal antibodies for immunosuppressed patients with mild to moderate COVID-19 and further exploration of the impacts of race/ethnicity and social determinants of health on access to monoclonal antibody therapy and other COVID-19 treatments.

CONCLUSION

In a large academic medical center that standardized its administration of monoclonal antibody therapy to high-risk patients with mild to moderate COVID-19, patients who received EUA bamlanivimab had a lower rate of hospitalization compared to patients who did not receive anti-spike monoclonal antibody therapy.

Supplementary Data

Supplementary materials are available at Clinical Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Click here for additional data file.