Impact of the Use of Oral Antiviral Agents on the Risk of Hospitalization in Community Coronavirus Disease 2019 Patients (COVID-19)

Abstract

Background

We examined the effectiveness of molnupiravir and nirmatrelvir/ritonavir in reducing hospitalization and deaths in a real-world cohort of nonhospitalized patients with coronavirus disease 2019 (COVID-19).

Methods

This was a territory-wide retrospective cohort study in Hong Kong. Nonhospitalized COVID-19 patients who attended designated outpatient clinics between 16 February and 31 March 2022 were identified. Patients hospitalized on the day of the first clinic appointment or used both oral antivirals were excluded. The primary endpoint was hospitalization. The secondary endpoint was a composite of intensive care unit admission, invasive mechanical ventilation use, and/or death.

Results

Of 93 883 patients, 83 154 (88.6%), 5808 (6.2%), and 4921 (5.2%) were oral antiviral nonusers, molnupiravir users, and nirmatrelvir/ritonavir users, respectively. Compared with nonusers, oral antiviral users were older and had more comorbidities, lower complete vaccination rate, and more hospitalizations in the previous year. Molnupiravir users were older and had more comorbidities, lower complete vaccination rate, and more hospitalizations in the previous year than nirmatrelvir/ritonavir users. At a median follow-up of 30 days, 1931 (2.1%) patients were hospitalized and 225 (0.2%) patients developed the secondary endpoint. After propensity score weighting, nirmatrelvir/ritonavir use (weighted hazard ratio 0.79; 95% confidence interval [CI], 0.65–0.95; P = .011) but not molnupiravir use (weighted hazard ratio 1.17; 95% CI, 0.99–1.39; P = .062) was associated with a reduced risk of hospitalization than nonusers. The use of molnupiravir or nirmatrelvir/ritonavir was not associated with a lower risk of the secondary endpoint as compared with nonusers.

Conclusion

Use of nirmatrelvir/ritonavir but not molnupiravir was associated with a reduced risk of hospitalization in real-world nonhospitalized patients with COVID-19.

SARS-CoV-2, hospital admission, death, molnupiravir, nirmatrelvir/ritonavir

The landscape of development of therapeutics and preventive strategies for coronavirus disease 2019 (COVID-19) has evolved rapidly since the start of the COVID-19 pandemic, including the treatment of severe disease in hospitalized patients, and vaccine platforms for the prevention of infection and severe disease [1, 2]. The latest breakthroughs in therapeutics emphasized early treatment for prevention of progression to severe disease among nonhospitalized patients [3]. Two oral antiviral agents, molnupiravir and nirmatrelvir/ritonavir, have recently been authorized or supported to be used worldwide for the treatment of mild to moderate COVID-19 in adults at risk for progressing to severe COVID-19 [4], and nonhospitalized patients at risk of hospitalization or progression to severe disease [5, 6].

The treatment authorizations and guidelines were mostly based on a single randomized trial for each of the drugs on individuals with mild to moderate COVID-19 and 1 or more risk factors for progression to severe disease within 5 days of symptom onset. The MOVe-OUT study showed that the molnupiravir group had a lower risk of hospitalization or death than placebo (6.8% vs 9.7%), or relative risk reduction of 30% and number needed to treat of 34 [7]. In the EPIC-HR study, the nirmatrelvir/ritonavir group had a lower risk of hospitalization or death (0.72% vs 6.53%), or relative risk reduction of 89% and needed to treat of 17 [8].

However, there is a paucity of knowledge on whether these trial data would translate into similar real-world effectiveness. The studied populations in these trials were relatively young (median age in the 40s), and the most frequently reported risk factor for progression to severe disease was obesity (reported in 74% and 81% of the 2 trials, respectively) [7, 8]. Real-world data include older patients with different comorbidities [9]. It is important to understand the real-world effectiveness of oral antivirals for public health authorities to determine the most cost-effective strategies for averting severe disease and reducing healthcare burden by targeting appropriate populations for treatment [10]. In this territory-wide study, we aimed to determine the real-world effectiveness of molnupiravir and nirmatrelvir/ritonavir in reducing hospitalization and deaths among nonhospitalized patients with COVID-19.

METHODS

Setting and Study Design

A territory-wide retrospective cohort study was performed using data from Clinical Data Analysis and Reporting System (CDARS), an electronic healthcare database managed by Hospital Authority, Hong Kong [11]. CDARS captures deidentified data of patients' demographic, death, diagnoses, procedures, drug prescription and dispensing history, and laboratory results from all public hospitals and clinics in Hong Kong, and represents inpatient and outpatient data of around 80% of the 7.4 million population. International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) coding was adopted in CDARS; it is found 99% accurate to identify medical conditions with reference to clinical, laboratory, imaging, and endoscopy results from electronic medical records [12]. Territory-wide studies on COVID-19 have previously been conducted using CDARS [12, 13].

Patients

Patients with COVID-19 who attended COVID-19-designated clinics in Hong Kong between 16 February 2022 and 31 March 2022 were identified by appointment records. Details on designated clinics were described in Supplementary Methods. Molnupiravir and nirmatrelvir/ritonavir were prescribed to the elderly and individuals with high-risk factors and incomplete COVID-19 vaccination within 5 days of symptom onset in designated clinics on 12 March 2022 and 16 March 2022, respectively, after these drugs had become available (Supplementary Table 1). Patients hospitalized on the day of the first appointment at designated clinic and/or used both molnupiravir and nirmatrelvir/ritonavir were excluded. Patients were followed until the occurrence of the clinical endpoint, death, date of data retrieval (25 April 2022), and up to 30 days, whichever came first. The study protocol was approved by the Joint Chinese University of Hong Kong - New Territories East Cluster Clinical Research Ethics Committee (reference number: 2021.239).

Data were retrieved on 26 April 2022. Baseline date was defined as the date of the first appointment at designated clinic. We retrieved data on date of birth, sex, hospitalization, diagnoses, procedures, and use of molnupiravir, nirmatrelvir/ritonavir, and other relevant concomitant drugs before baseline and during follow-up. We also collected patients' laboratory parameters including hemoglobin A_1c, fasting plasma glucose, C-reactive protein, international normalized ratio, complete blood picture, liver biochemistries, renal function tests, and COVID-19 polymerase chain reaction tests.

Definitions

The primary endpoint was hospital admission. The secondary endpoint was a composite endpoint of intensive care unit (ICU) admission, use of invasive mechanical ventilation (IMV), and/or death. Hospital admission was defined as hospitalization with a stay of more than 1 day. Date and cause of death were ascertained using data from CDARS and Hong Kong Death Registry. IMV use was defined by ICD-9-CM procedure codes (96.04, 96.05, 96.7). Definition of comorbidities and complete vaccination were described in Supplementary Methods and Supplementary Table 2.

Statistical Analysis

Data were analyzed using R software (version 4.1.2). Continuous variables were expressed in mean ± standard deviation or median (25th percentile—75th percentile [P25–P75]), as appropriate, whereas categorical variables were presented as frequency (percentage). Qualitative and quantitative differences between groups were analyzed by χ² or Fisher exact test for categorical parameters and 1-way analysis of variance or Kruskal–Wallis test for continuous parameters, as appropriate. Propensity score (PS), the conditional probability of using nirmatrelvir/ritonavir given patients' clinical characteristics, was estimated among nirmatrelvir/ritonavir users, molnupiravir users, and oral antiviral nonusers to control for confounders and reduce selection bias. Details of PS and weighted Cox model are described in Supplementary Methods. Cumulative incidence with 95% confidence interval (CI) of the primary and secondary endpoints of the 3 groups was estimated by Kaplan-Meier method. Robust (empirical) variance estimates were obtained to calculate 95% CI of the weighted hazard ratio (wHR). Two subgroup analyses were performed in high-risk patients aged 60 years or older or aged younger than 60 years with at least 1 comorbidity, and patients older and younger than 70 years, which represented 2 populations with different complete vaccination rates. All statistical tests were 2-sided. Statistical significance was taken as P < .05.

RESULTS

Demographic Characteristics

We identified 94 167 patients with COVID-19 with an appointment at designated clinics from 16 February 2022 to 31 March 2022. We excluded 271 patients who were hospitalized on the day of the first appointment at the designated clinic and 13 patients who received both molnupiravir and nirmatrelvir/ritonavir. As a result, 93 883 patients (83 154 oral antiviral nonusers, 5808 molnupiravir users, and 4921 nirmatrelvir/ritonavir users) were included in the analysis (Supplementary Figure 1); 10 569(98.5%) and 10 656 (99.3%) of the oral antivirals were prescribed on the same date and within the first 2 days of baseline date, respectively. At baseline, compared with oral antiviral nonusers, molnupiravir or nirmatrelvir/ritonavir users were older and had more comorbidities including digestive diseases, diabetes mellitus, history of malignant tumor, more hospital admission in the previous year, and lower complete vaccination rate (Table 1). Compared with nirmatrelvir/ritonavir users, molnupiravir users were older and had more cardiovascular diseases, diabetes mellitus, cerebrovascular events, respiratory diseases, and kidney diseases, more hospital admission in the previous year, and lower complete vaccination rate (Table 1).

Table 1.

Baseline Clinical Characteristics of the 93 883 Patients With COVID-19 Who Attended the Designated Clinics in Hong Kong From 16 February 2022 to 31 March 2022

	All	COVID-19 Oral Antiviral Nonusers	Use of Molnupiravir	Use of Nirmatrelvir/ Ritonavir
Clinical characteristics	N = 93 883	N = 83 154	N = 5808	N = 4921	P Value
Age, y	49.2 ± 21.8	46.1 ± 20.8	75.4 ± 12.1	70.8 ± 12.1	<.001
Male sex, n (%)	41 656 (44.4)	36 706 (44.1)	2703 (46.5)	2247 (45.7)	<.001
Comorbidities, n (%)	…	…	…	…	…
Cardiovascular diseases	4612 (4.9)	3852 (4.6)	544 (9.4)	216 (4.4)	<.001
- Hypertension	4315 (4.6)	3656 (4.4)	465 (8.0)	194 (3.9)	<.001
- Ischemic heart disease	444 (0.5)	316 (0.4)	101 (1.7)	27 (0.5)	<.001
- Cardiac dysrhythmias	357 (0.4)	270 (0.3)	74 (1.3)	13 (0.3)	<.001
- Heart failure	206 (0.2)	121 (0.1)	76 (1.3)	9 (0.2)	<.001
Digestive diseases	3841 (4.1)	3163 (3.8)	391 (6.7)	287 (5.8)	<.001
- Peptic ulcer	120 (0.1)	95 (0.1)	15 (0.3)	10 (0.2)	.004
- Chronic liver disease	3516 (3.7)	2916 (3.5)	345 (5.9)	255 (5.2)	<.001
- Liver failure, cirrhosis, or cirrhotic complications	17 (0.02)	15 (0.02)	2 (0.03)	0 (0)	.451
- Biliary disease	105 (0.1)	79 (0.1)	14 (0.2)	12 (0.2)	<.001
- Gastrointestinal hemorrhage	181 (0.2)	141 (0.2)	27 (0.5)	13 (0.3)	<.001
Diabetes mellitus	12 331(13.1)	8827 (10.6)	2181 (37.6)	1323 (26.9)	<.001
Malignant tumors	770 (0.8)	495 (0.6)	175 (3.0)	100 (2.0)	<.001
Nervous system diseases	331 (0.4)	249 (0.3)	64 (1.1)	18 (0.4)	<.001
- Cerebrovascular events	314 (0.3)	239 (0.3)	59 (1.0)	16 (0.3)	<.001
- Other nervous system diseasesa	17 (0.02)	10 (0.01)	5 (0.09)	2 (0.04)	.001
Respiratory diseasesb	128 (0.1)	80 (0.1)	36 (0.6)	12 (0.2)	<.001
Kidney diseases	373 (0.4)	256 (0.3)	103 (1.8)	14 (0.3)	<.001
HIV infection	14 (0.01)	13 (0.02)	0 (0)	1 (0.02)	.638
Days from the start of designated clinic	24.8 ± 10.6	23.6 ± 10.6	32.7 ± 5.1	35.7 ± 4.2	<.001
Age- and sex-specified complete vaccination rate (%)	54.2 ± 22.3	56.1 ± 22.2	36.2 ± 16.6	42.7 ± 15.7	<.001
Number of hospitalizations in the past year, n (%)	…	…	…	…	<.001
- 0	91 667(97.6)	81 424(97.9)	5465 (94.1)	4778 (97.1)	…
- 1	1850 (2.0)	1435 (1.7)	296 (5.1)	119 (2.4)	…
- ≥2	366 (0.4)	295 (0.4)	47 (0.8)	24 (0.5)	…
Body mass index, kg/m2	24.4 ± 4.7	24.5 ± 4.7	23.9 ± 4.4	24.0 ± 4.2	<.001
Missing, %	81.9	83.4	65.4	76.0	…
Hemoglobin A1c	6.4 ± 1.0	6.3 ± 1.0	6.5 ± 1.1	6.4 ± 1.0	<.001
Missing, %	72.4	76.6	31.3	49.2	…
Follow-up duration, d	30 (30–30)	30 (30–30)	30 (30–30)	30 (29–30)	<.001

All comorbidities are represented as binary parameters. Categorical variables are presented as number (percentage). Follow-up duration is expressed in median (25th percentile—75th percentile). Age was expressed in mean ± standard deviation. Qualitative and quantitative differences between subgroups were analyzed by χ² or Fisher exact test for categorical parameters and Student t test or Mann–Whitney U test for continuous parameters, as appropriate. All patients had available information on clinical characteristics in Table 1.

Abbreviations: COVID-19, coronavirus disease 2019; HIV, human immunodeficiency virus; ICD-9-CM, International Classification of Diseases, Ninth Revision, Clinical Modification.

Other nervous system disease was defined by ICD-9-CM diagnosis codes for inflammatory diseases of the central nervous system (ICD-9-CM codes: 320–327), hereditary and degenerative diseases of the central nervous system (ICD-9-CM codes: 330–337), and other disorders of the central nervous system (ICD-9-CM codes: 340–345).

Respiratory system disease was defined by ICD-9-CM diagnosis codes chronic obstructive pulmonary disease and allied conditions (ICD-9-CM codes: 490–496), pneumoconioses and other lung diseases resulting from external agents (ICD-9-CM codes: 500–508) in the previous 3 months, and other diseases of respiratory system (ICD-9-CM codes: 510–519) in the previous 3 months.

Clinical Outcomes

At a median (P25–P75) follow-up of 30 (30–30) days, 1931 (2.1%) patients were hospitalized; 1322 (1.6%), 437 (7.5%), and 172 (3.5%) oral antiviral nonusers, molnupiravir users, and nirmatrelvir/ritonavir users were hospitalized, respectively. Among 1931 hospitalized patients, 558 (28.9%) received oxygen therapy; 179 (9.3%) used remdesivir. Among 93 883 patients, 225 (0.2%) patients developed the secondary endpoint in 30 days (ie, ICU admission, IMV use, and/or death); 151 (0.2%), 53 (0.9%), and 21 (0.4%) oral antiviral nonusers, molnupiravir users, and nirmatrelvir/ritonavir users developed the secondary endpoint, respectively.

PS Weighting Analysis

After PS weighting, the clinical characteristics were balanced between nirmatrelvir/ritonavir users, molnupiravir users, and oral antiviral nonusers (Table 2). Molnupiravir use was not associated with a reduced risk of hospital admission than oral antiviral nonusers (wHR 1.17; 95% CI, 0.99–1.39; P = .062). Nirmatrelvir/ritonavir use was associated with a lower risk of hospital admission than oral antiviral nonusers (wHR 0.79; 95% CI, 0.65–0.95; P = .011) and molnupiravir users (wHR 0.67; 95% CI, 0.55–0.81; P < .001) (Table 3). The 30-day cumulative incidence (95% CI) of hospital admission was 4.5% (4.0%–5.0%), 5.2% (4.6%–5.9%), and 3.6% (3.1%–4.1%) in nonusers, molnupiravir users, and nirmatrelvir/ritonavir users, respectively (Figure 1A). Similar results were found in patients aged 60 years or older or aged younger than 60 years with at least 1 comorbidity (Table 3, Supplementary Table 3, and Figure 2A). In patients older and younger than 70 years, the complete vaccination rate was 30% and 60%, respectively (Supplementary Tables 4 and 5). Similar associations between use of nirmatrelvir/ritonavir (wHR 0.78 and 0.77) and molnupiravir (wHR 1.15 and 1.07) with hospital admission were observed (Supplementary Table 6).

Figure 1.

Cumulative incidence of (A) hospital admission and (B) admission to the intensive care unit (ICU)/use of invasive mechanical ventilation (IMV)/death in patients with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection/coronavirus disease 2019 (COVID-19) who did not receive oral antiviral agents, received molnupiravir, or received nirmatrelvir/ritonavir after propensity score (PS) weighting.

Figure 2.

Cumulative incidence of (A) hospital admission and (B) admission to an intensive care unit (ICU)/use of invasive mechanical ventilation (IMV)/death in the subgroup of patients aged 60 years or older or younger than age 60 years with comorbidities who did not receive oral antiviral agents, received molnupiravir, or received nirmatrelvir/ritonavir after propensity score (PS) weighting.

Table 2.

Baseline Clinical Characteristics and Balancing Diagnositcs Before and After Propensity Score Weighting Between COVID-19 Patients Who Did Not Use Oral Antiviral Agents, Used Molnupiravir, or Used Nirmatrelvir/Ritonavir

	Before Propensity Score Weighting					After Propensity Score Weighting
Clinical Characteristics	COVID-19 Oral Antiviral Nonuser	Use of Molnupiravir	Use of Nirmatrelvir/ Ritonavir	ASMDa	ASMDb	COVID-19 Oral Antiviral Nonuser	Use of Molnupiravir	Use of Nirmatrelvir/ Ritonavir	ASMDa	ASMDb
	N = 83 154	N = 5808	N = 4921			N = 4758	N = 4798	N = 4921
Age, y	46.1 ± 20.8	75.4 ± 12.1	70.8 ± 12.1	2.044	0.377	70.5 ± 12.2	71.1 ± 11.7	70.8 ± 12.1	0.024	0.023
Male (n, %)	36 706 (44.1)	2703 (46.5)	2247 (45.7)	0.031	0.018	2178 (45.8)	2246 (46.8)	2247 (45.7)	0.002	0.023
Comorbidities (n, %)	…	…	…	…	…	…	…	…	…	…
Cardiovascular diseases	3852 (4.6)	544 (9.4)	216 (4.4)	0.012	0.243	216 (4.5)	220 (4.6)	216 (4.4)	0.007	0.009
Digestive diseases	3163 (3.8)	391 (6.7)	287 (5.8)	0.087	0.038	276 (5.8)	296 (6.2)	287 (5.8)	0.001	0.014
Diabetes mellitus	8827 (10.6)	2181 (37.6)	1323 (26.9)	0.367	0.241	1283 (27.0)	1325 (27.6)	1323 (26.9)	0.002	0.017
Malignant tumor	495 (0.6)	175 (3.0)	100 (2.0)	0.102	0.070	89 (1.9)	104 (2.2)	100 (2.0)	0.011	0.010
Nervous system diseases	249 (0.3)	64 (1.1)	18 (0.4)	0.011	0.122	18 (0.4)	21 (0.4)	18 (0.4)	0.002	0.013
Respiratory diseases	80 (0.1)	36 (0.6)	12 (0.2)	0.030	0.076	11 (0.2)	14 (0.3)	12 (0.2)	0.001	0.008
Kidney diseases	256 (0.3)	103 (1.8)	14 (0.3)	0.004	0.280	14 (0.3)	17 (0.4)	14 (0.3)	0.004	0.013
Days from the start of designated clinic	23.6 ± 10.6	32.7 ± 5.1	35.7 ± 4.2	2.879	0.713	35.6 ± 4.3	35.6 ± 4.2	35.7 ± 4.2	0.024	0.006
Age- and sex-specified complete vaccination rate (%)	55.9 ± 22.2	36.1 ± 16.7	42.6 ± 15.8	0.847	0.413	42.8 ± 15.7	42.5 ± 15.7	42.6 ± 15.8	0.014	0.007
Number of hospitalizations in the past year (n, %)	…	…	…	…	…	…	…	…	…	…
- 0	81 424(97.9)	5465 (94.1)	4778 (97.1)	…	…	4618 (97.1)	4646 (96.8)	4778 (97.1)	…	…
- 1	1435 (1.7)	296 (5.1)	119 (2.4)	0.045	0.174	117 (2.5)	123 (2.6)	119 (2.4)	0.003	0.009
- ≥2	295 (0.4)	47 (0.8)	24 (0.5)	0.019	0.046	23 (0.5)	29 (0.6)	24 (0.5)	0.001	0.015

Use of COVID-19 oral antiviral referred to the use of molnupiravir or nirmatrelvir/ritonavir at baseline or during follow-up. A total of 99.3% of the COVID-19 oral antiviral users used the antiviral drugs within the first 2 days of follow-up. An ASMD <0.1 indicated good balance between COVID-19 oral antiviral users and nonusers. Parameters with ASMD ≥0.1 would be adjusted in doubly robust model. The effective sample size after propensity score weighting was 8079, 3399, and 4921 in nonusers, molnupiravir users, and nirmatrelvir/ritonavir users, respectively.

Abbreviations: ASMD, absolute standardized mean difference; COVID-19, coronavirus disease-2019; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.

ASMD between COVID-19 oral antiviral nonusers and users of nirmatrelvir/ritonavir.

ASMD between users of molnupiravir and users of nirmatrelvir/ritonavir.

Table 3.

Weighted Cox Proportional Hazard Regression After Propensity Score Weighting on Association Between Use of COVID-19 Oral Antiviral Drugs With the Development of Primary and Secondary Endpoints in All COVID-19 Patients Who Attended Designated Clinic in Hong Kong, and the Subgroup of Patients Aged 60 Years or Older or Younger Than 60 Years With Comorbidities

COVID-19 Oral Antiviral Use	Hospital Admission		Death/ICU Admission/Use of Invasive Mechanical Ventilation
	Weighted HR (95% CI)	P Value	Weighted HR (95% CI)	P Value
All COVID-19 Patients
No oral antiviral use	Referent	…	Referent	…
Use of molnupiravir	1.17 (.99–1.39)	.062	1.12 (.68–1.82)	.663
Use of nirmatrelvir/ritonavir	.79 (.65–.95)	.011	.81 (.47–1.39)	.448
No oral antiviral use	.85 (0.72–1.01)	.062	.90 (.55–1.47)	.663
Use of molnupiravir	Referent	…	Referent	…
Use of nirmatrelvir/ritonavir	.67 (.55–.81)	<.001	.73 (.41–1.27)	.265
All COVID-19 patients aged 60 years or older or younger than age 60 years with comorbidities
No oral antiviral use	Referent	…	Referent	…
Use of molnupiravir	1.07 (.90–1.26)	.472	1.04 (.63–1.73)	.874
Use of nirmatrelvir/ritonavir	.76 (.63–.92)	.004	.81 (.47–1.39)	.447
No oral antiviral use	.94 (.79–1.11)	.472	.96 (.58–1.59)	.874
Use of molnupiravir	Referent	…	Referent	…
Use of nirmatrelvir/ritonavir	.72 (.59–.87)	.001	.78 (.44–1.38)	.392

Abbreviations: CI, confidence interval; COVID-19, coronavirus disease 2019; HR, hazard ratio; ICU, intensive care unit.

Molnupiravir (wHR 1.12; 95% CI, 0.68–1.82; P = .663) or nirmatrelvir/ritonavir use (wHR 0.81; 95% CI, 0.47–1.39; P = .448) were not associated with a lower risk of death/ICU admission/IMV use than oral antiviral nonusers. Also, nirmatrelvir/ritonavir use was not associated with a lower risk of death/ICU admission/IMV use than molnupiravir users (weighted HR 0.73; 95% CI, 0.41–1.27; P = .265) (Table 3). The 30-day cumulative incidence (95% CI) of death/ICU admission/IMV use was 0.5% (0.4%–0.7%), 0.6% (0.4%–0.9%), and 0.4% (0.3%–0.7%) in nonusers, molnupiravir users, and nirmatrelvir/ritonavir users, respectively (Figure 1B). Similar findings were observed in patients aged 60 years or older or younger than 60 years with at least 1 comorbidity (Table 3, Supplementary Table 3, and Figure 2B). In patients older and younger than 70 years of age, similar associations between use of nirmatrelvir/ritonavir (wHR 0.89 and 0.55) and molnupiravir (wHR 1.08 and 0.97) with death/ICU admission/IMV use were observed (Supplementary Table 6).

DISCUSSION

This study describes the real-world effectiveness of the 2 COVID-19 oral antivirals amid the peak of an outbreak with omicron variant infections in a densely populated city. Compared with no antiviral treatment, nirmatrelvir/ritonavir significantly reduced hospital admission by more than 20%, whereas molnupiravir did not reduce hospital admission of community patients with COVID-19. Neither of the drugs reduced the risk of adverse clinical outcomes, namely death, ICU admission, and IMV use.

At a critical time witnessing the rapid global spread of the omicron variant, molnupiravir and nirmatrelvir/ritonavir were approved for outpatient treatment of patients with mild to moderate disease and at risk for disease progression, to reduce the risk of hospital admission and deaths if administered early to high-risk subjects [14, 15]. In clinical trials, nirmatrelvir/ritonavir demonstrated a greater relative risk reduction in hospitalization and death than molnupiravir compared with placebo [7, 8]. Yet, there have not been any head-to-head comparisons between the 2 drugs. The unique situation in Hong Kong with the availability of both drugs of different antiviral mechanisms at the same time facilitated their comparisons in real-world setting. The apparent lack of effectiveness in reducing hospitalization by molnupiravir might partly be related to its availability in earlier days when our local guideline limited its use to patients at highest baseline risk, namely advanced age (≥70 years) and unvaccinated status. Moreover, molnupiravir was preferentially prescribed to more frail patients with multiple comorbidities and polypharmacy than those who received nirmatrelvir/ritonavir, perhaps because of the multiple drug–drug interactions associated with the latter [16, 17]. When nirmatrelvir/ritonavir became available, the guideline relaxed the use of both oral antivirals in older patients regardless of vaccination status and in younger patients with comorbidities. This explains why molnupiravir users was older and had more comorbidities than nirmatrelvir/ritonavir users and nonusers. However, we have balanced the differences in host characteristics among the 3 groups using PS weighting and were not able to observe any significant association between molnupiravir use and hospitalization in the weighted analyses; residual unmeasured confounding that was not adjusted by PS might have obscured an impact of molnupiravir on reduced hospitalization risk. Only 0.2% of patients developed death/ICU admission/IMV use in the study. The low event rate led to the wider CIs of the hazard ratios. Yet, the direction and magnitude of the hazard ratios for death/ICU admission/IMV use were similar to those for the primary outcome. The MOVe-OUT trial has been criticized for overestimated treatment effects in the interim analysis, lack of explanation for after-interim period data favoring placebo, and wide differences in outcomes among participating countries [18, 19].

In the MOVe-OUT and EPIC-HR trials, 60% and 98% of the participants were infected by the delta variant, respectively [7, 8]. Our territory-wide, real-world cohort was different. The rapid surge of COVID-19 in the first quarter of 2022 in Hong Kong was primarily related to the highly transmissible nature of the omicron variant [20]. Although these 2 drugs were shown to retain antiviral activity against omicron variant in vitro [21], their effectiveness in clinical settings remains to be established, as the omicron variant possesses higher transmissibility and reduced pathogenicity than earlier variants [22]. Hong Kong has experienced the fifth wave of COVID-19 since 31 December 2021, with a cumulative number of 1 376 651 confirmed cases by 8 August 2022 (Supplementary Figure 2) [20]. The share of omicron variant rose rapidly from 93% to 100% since early January 2022 [23]. This would be a suitable setting to determine the real-world impact of oral antivirals in COVID-19 infections caused predominantly by omicron. Our current observations fortify the real-world impact of these 2 novel oral antivirals because the omicron variant has been the predominant strain worldwide since late 2021. Global COVID-19 cases surpassed 500 million in early April, as the highly contagious BA.2 sub-omicron variant surges in many countries in Europe and Asia, including China.

The 2 landmark trials enrolled solely unvaccinated participants. As vaccination and booster rates are rising in all countries, further study is needed for their effectiveness among partially or fully vaccinated individuals with breakthrough infections [4]. The proportion of Hong Kong population who received at least 1 dose of SARS-CoV-2 vaccine rose from 65% to 93.1% from 31 December 2021 to 8 August 2022 [24]. More than 80% of the population has completed the second dose by March 2022; yet, the coverage of third dose remained below 50% [24]. A study from Israel involving patients with at least 1 risk factor for disease progression and an overall adequate vaccination rate of 75% showed that nirmatrelvir/ritonavir had a 46% reduced relative risk of progression to severe disease or death [25]. Another study from Hong Kong involving nonhospitalized patients with 54% being fully vaccinated showed that nirmatrelvir/ritonavir reduced hospitalization by 31%, whereas molnupiravir was not associated with a lower hospitalization rate [26]. Effectiveness of nirmatrelvir/ritonavir was not affected by vaccination status in these 2 studies. Similarly, in our study, the associations of oral antiviral use and primary and secondary endpoints were similar between those younger and older than age 70 years, who represented different vaccination rates. Although ongoing trials will provide more data in vaccinated populations infected with omicron variant [27, 28], our study and theirs supported the use of nirmatrelvir/ritonavir in preventing hospitalization in vaccinated populations with risk factors for disease progression. The established risk factors, namely advanced age and presence of comorbidities, facilitate prioritized referral of community COVID-19 patients to designated clinics and timely use of oral antiviral treatment to avoid hospitalization. Our findings would facilitate the clinical management and resource allocation for appropriate use of these oral antivirals amid the COVID-19 outbreak.

The strengths of our study include a territory-wide, real-world cohort that covers 100% of the designated clinic services and more than 95% of the inpatient service for patients with COVID-19. Our real-world cohort represents a wider spectrum of patients such that the findings are more representative of individuals encountered in daily clinical practice than those enrolled in clinical trials. Our study has a few limitations. First, patients with COVID-19 untreated with oral antivirals were much younger than the treated ones because of indication bias. We compensated this major discrepancy by various approaches, including PS weighting, which rendered age to be very well balanced. Second, many patients might not be seen at designated clinics or hospitalized at the peak of the fifth wave because of the huge numbers of confirmed cases (up to a peak of 70 000 confirmed cases per day), and this might lead to fewer hospital admissions than it should have been [29]. We believe this would have affected patients prescribed and not prescribed oral antivirals similarly if they were infected at the same time during the fifth wave. Therefore, the day from the start of operation of designated clinics was balanced among the groups by PS weighting to reduce bias because of varying hospital admission thresholds throughout the fifth wave. Third, ascertainment bias may affect the reliability of study due to inaccurate entry of certain diagnosis codes for comorbidities. We minimized this bias by including laboratory and medication data for certain diagnoses such as diabetes mellitus and hypertension. Fourth, we did not analyze patients who resided in aged home, whose vaccination rate was about 20% to 50% [30, 31]. Community outreach teams prescribed either of the 2 oral antivirals if clinically indicated. The patients did not need to attend a designated clinic. Thus, another study is warranted to evaluate the effectiveness of the 2 oral antivirals in these frailer patients. Fifth, there might be a difference in the time from symptom onset to baseline date between users and nonusers of COVID-19 oral antiviral. Also, missing data on hemoglobin A_1c, body mass index, and other laboratory parameters existed that are not routinely measured at designated clinics. Because these data were not available, we did not adjust for these possible confounding factors in our analyses. Sixth, the vaccination data were only available at population level. Thus, we included the background vaccination rate of each patient at baseline as the corresponding vaccination rate in the Hong Kong population of the same age and sex.

In conclusion, this territory-wide, real-world study reported the effectiveness of the 2 oral antiviral agents for COVID-19 amid the peak of an outbreak with omicron variant infections in one of the most densely populated cities. Although nirmatrelvir/ritonavir reduced hospital admission by more than 20%, molnupiravir appeared not to be able to reduce hospital admission of community patients with COVID-19. Given the ongoing outbreak worldwide, we have to update our management guidelines for community patients COVID-19 and prioritize the use of these agents to those who would benefit from it. Health authorities should allocate adequate resources, in particular sufficient outpatient clinic settings and timely use of antiviral treatment, based on the trajectories of the numbers of confirmed cases for upcoming waves well ahead to avoid collapse of the healthcare systems by reducing hospital admission as much as possible.

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.

Notes

Author contributions. All authors were responsible for the study concept and design. G. W., T. Y., M. L., Y.-K. T., and G. L. were responsible for the acquisition and analysis of data, had full access to all of the data in the study, and take responsibility for the integrity of the data and the accuracy of the data analysis. All authors were responsible for the interpretation of data, the drafting, and critical revision of the manuscript for important intellectual content.

Financial support. None declared.

Author notes

T.C.F. Yip and G.C.Y. Lui contributed equally to this study.

Potential conflicts of interest. T. Y. has served as an advisory committee member and a speaker for Gilead Sciences. G. L. has served as an advisory committee member for Gilead, Merck, and GSK; speaker for Merck, Pfizer, and Gilead; and received research grant from Gilead, Merck, and GSK. V. W. has served as a consultant or advisory committee member for 3V-BIO, AbbVie, Boehringer Ingelheim, Echosens, Gilead Sciences, Intercept, Inventiva, Merck, Novo Nordisk, Pfizer, ProSciento, Sagimet Biosciences, and TARGET PharmaSolutions; and a speaker for Bristol-Myers Squibb, Abbott, AbbVie, Echosens, Gilead Sciences, Merck, and Novo Nordisk. He has received a research grant from Gilead Sciences and is a cofounder of Illuminatio Medical Technology Limited. H. C. has served as an Independent Non-Executive Director for Shanghai Henlius Biotech Inc; as an advisory board member for Aligos, Aptorum, Arbutus, Hepion, Janssen, Gilead, Glaxo-Smith-Kline, Roche, Vaccitech, Virion Therapeutics, and Vir Biotechnology; and as a speaker for Gilead, Roche, and Viatris. G. W. has served as an advisory committee member for Gilead Sciences and Janssen, and as a speaker for Abbott, Abbvie, Ascletis, Bristol-Myers Squibb, Echosens, Gilead Sciences, Janssen, and Roche. She has also received a research grant from Gilead Sciences. D. H. is an advisory committee member for Roche (personal fees). E. H. reports grants from Merck and GSK; is an advisory committee member for Merk, Gilead, Sanofi Pasteur, and GSK; and speaker for Merck and Gilead sciences. All other authors report no potential conflicts.

All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.