Introduction

On 5 May 2023, the World Health Organization (WHO) declared that COVID-19 was no longer a Public Health Emergency of International Concern. Even though the number of deaths and hospitalizations has significantly declined due to the widespread availability of vaccines and other public health measures, some regions are still experiencing localized outbreaks and transmission [1]. According to the data provided by US Center for Disease Control and Prevention (CDC), in the 2024-25 season, the overall rate of COVlD-19-associated hospitalization was 15.1 per 100,000 people and weekly death rates due to COVID-19 were less than 3.8%. However, the elderly have the highest hospitalization rates among all age groups [2]. Similarly, a study from China showed that the overall rate of severe or critical illness in 2022 was as low as 0.035% [3]. However, this aggregate data masks the elevated risk for elderly populations, particularly amid persistent transmission chains driven by asymptomatic carriers. Presumably now, the asymptomatic subjects potentially contribute to the transmission of COVID-19 without their knowledge, intention, or being diagnosed as carriers [4]. As “silent spreaders”, asymptomatic carriers may be highly infective during the incubation period [5]. Elderly individuals, due to their weaker immune systems and higher prevalence of underlying diseases, face a higher risk of infection after exposure to asymptomatic [6]. Once infected, this vulnerability translates to severe outcomes: about 4.4% of elderly experienced severe to critical COVID-19 [7]. These results suggest that although the overall rates of COVID-19-associated hospitalizations and deaths may be relatively low, the elderly remain at higher risk, highlighting the need for targeted interventions and enhanced healthcare strategies to protect this vulnerable group. China has an elderly population of 190 million, accounting for 13.5% of the total population [8]. Protecting this vulnerable population (elderly patients) from severe or critical COVID-19 continues to be very important.

Studies have demonstrated that cytokine storm plays a critical role in the severe COVID-19 patients [9]. It might result in uncontrollable inflammation that further leads to multiple-organ failure, eventually leading to death [10,11,12,13,14,15,16]. Anti-inflammatory and immunosuppressive agents are considered effective therapeutic options for alleviating systemic inflammation [17]. Corticosteroids, interleukin-6 (IL-6) inhibitors, and tocilizumab have been utilized in severe patients [18]. Nevertheless, prior studies evaluating corticosteroids in patients with mild to moderate COVID-19 have not reached a consensus, as noted in the discussion. Additionally, there is still inconsistency in guidelines along with a notable lack of robust evidence on the effectiveness of corticosteroids in preventing progression and reducing mortality [18, 19]. Here, we report the results of this trial evaluating corticosteroids for elderly with mild to moderate COVID-19, aiming to determine whether corticosteroids reduce disease progression risk.

Methods

Study design and participants

BEAT COV is a multicentre, open-label, randomized, controlled trial to evaluate the efficacy of corticosteroids in combination with SOC in elderly with mild to moderate COVID-19. This trial was conducted across 15 hospitals in China, including 14 large central hospitals and one community healthcare center. Written informed consent was obtained from all participants or their legal representatives before screening. The study commenced on May 26, 2023, but was terminated prematurely on November 30, 2024, due to the low prevalence of COVID-19 and the declining incidence of severe cases. The study design is provided in Fig. 1.

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Study design schema

Patients aged at least 65 years were eligible for the randomized cohorts if they met the following criteria: (1) a positive SARS-CoV-2 test within 7 days prior to screening; (2) severity of mild or moderate COVID-19 (mild COVID-19 is defined as respiratory symptoms, including but not limited to dry throat, sore throat, cough, and fever; moderate COVID-19 is defined as a). persistent fever lasting more than 3 days with (or without) symptoms, the respiratory rate (RR) is less than 30 breaths per minute, and the arterial oxygen saturation (SpO2) is greater than 93% on room air; or b). imaging findings may show characteristic signs of pneumonia due to SARS-CoV-2 infection) [20]; (3) initial COVID-19 symptom onset no more than 7 days before enrollment; (4) no systemic corticosteroid treatment in the previous seven days; (5) at least one of the following criteria: (a) body temperature ≥ 38 °C, (b) c-reactive protein (CRP) ≥ 27 mg/L [21], or (c) radiological findings indicative of pneumonia due to SARS-CoV-2 infection; (6) no contraindications to systemic corticosteroids; and (7) willingness to receive systemic corticosteroids.

Patients were ineligible if they had (1) severe or critical COVID-19, (2) serious and uncontrolled medical status, (3) a life expectancy of less than 1 month, or (4) any other circumstances that, in the opinion of the investigator, are inappropriate for participation in the study.

If patients do not meet the key inclusion criteria (5), (6), and (7), but satisfy all other inclusion criteria and do not meet any exclusion criteria of the study, they will be enrolled in the observational cohort. Incorporating participants who did not meet randomization criteria enhanced the population heterogeneity, thereby improving the external validity of findings. Furthermore, the observational cohort design reflects real-world corticosteroid use in clinical practice.

Randomization

Patients were randomly assigned in a 1:1 ratio to receive standard of care plus corticosteroids or standard of care alone. Randomization was performed using block randomization to ensure an equal distribution of participants across groups. The randomization process involved creating blocks of participants, with each block containing an equal number of corticosteroid group and standard of care group allocations [22]. Meanwhile, randomization was stratified into two strata (one received antiviral therapy, the other one did not). Participants within each stratum will be randomized to ensure that everyone has an equal chance of being assigned to any treatment group within that stratum. This approach allows for the generation of two separate sets of randomization codes, one corresponding to the stratum with antiviral medication and the other to the stratum without antiviral medication. Randomization was performed using a sealed envelope technique for each stratum to ensure unbiased allocation of participants to the treatment groups. Randomization codes were generated by a random code generator before the start of the trial. These codes matched either the standard of care group or the corticosteroid group. Sealed envelopes containing the randomization assignments were prepared and sequentially labelled. Each envelope was opaque and tamper-proof to ensure that the randomization process remained concealed until the time of group allocation. At the time of participants randomization at each site, eligible participant was allocated a randomly numbered envelope, which was opened by the investigator to reveal the treatment allocation.

Procedures

Randomized participants in the corticosteroid group received SOC in combination with corticosteroids (dexamethasone 3 mg, prednisolone 20 mg, or methylprednisolone 16 mg) once daily for 5 days, while the SOC group and the observational cohort received SOC alone. The SOC was determined at the discretion of the investigator and may or may not include antivirals and other effective symptomatic treatments.

The study consisted of a screening period of 24 h, a 5-day treatment period, and a follow-up period of 28 days (Fig. 1). Pulse oximeters and 14-day antigen test kits were provided to all participants upon enrollment with the home care patients required to conduct daily health monitoring through structured diary cards during the follow-up period. The trial team conducted serial telephone follow-ups with inpatients, who received daily vital sign assessments at the hospital (detailed procedures are described in Supplementary Material 1).

Outcomes

The primary outcome was a composite of the rate of severe or critical COVID-19 within 28 days that included, (1) SpO2 ≤ 93% on room air, (2) emergency department visits or hospitalizations due to COVID-19 progression (defined as lung imaging showed significant progression more than 50% within 24 ~ 48 h, or respiratory failure required mechanical ventilation, or shock) and with or without complication (associated with organ failure that require Intensive Care Unit [ICU] monitoring and treatment).

Secondary outcomes included the proportion of participants with (1) SpO2 ≤ 93% on room air, (2) emergency department visits or hospitalizations due to COVID-19 progression, (3) time to negative conversion for nucleic acid or antigen, (4) percentage of patients who maintained SARS-CoV-2 negativity on day 14, (5) time to initial alleviation of all symptoms (defined as the date symptoms first reported as none or mild), (6) time to sustained recovery of all symptoms (defined as the date when symptoms were first reported as either none or mild and subsequently remained at none or mild until 14 days), (7) the duration of respiratory symptoms (defined as the presence of any respiratory symptoms [mild, moderate or severe]), (8) time to hospitalization or emergency department visit due to COVID-19 progression, (9) all-cause mortality, (10) duration of hospital admission due to any cause. All time-to-event analyses used the date of randomization as the baseline. Safety outcomes included the incidence rates of ADR and serious adverse drug reactions (SADR) related to corticosteroids and antivirals.

Sample size and statistical analysis

The sample size calculation was based on primary outcome. Given the current availability of various antivirals and the changes in herd immunity following the COVID-19 epidemic, we estimated that the severe or critical disease rate among elderly is now approximately 1%. It was assumed that this rate would be reduced by 70% in the intervention group. To achieve 80% power and a 0.05 two-sided type I error in a pairwise comparison against the control group, a sample size of 2,355 patients was determined for each randomized group. Additionally, considering a dropout rate of 10% and the fact that 90% of the enrolled participants meet the criteria for corticosteroid use, the study aimed to screen approximately 5815 patients.

In the statistical analysis, primary efficacy was evaluated in both the FAS and the per-protocol set (PPS), while other endpoints were assessed only in FAS. The FAS is defined as all participants who met the eligibility criteria and who were randomized to receive at least one dose of the study drug. The PPS population included patients who had received at least one dose of study drug, had available primary outcomes, adhered to corticosteroid regimens not exceeding 80% to 120% of the established protocol, and did not experience any major protocol deviations in the assessment of the primary endpoint. All secondary efficacy outcomes were analyzed using the FAS. Subgroup analyses were also done for the primary endpoint.

To quantify the treatment effect, we used relative risk (RR) with 95% confidence intervals to compare outcomes between the intervention and control arms. The primary outcome was analyzed using a Chi-square test if the theoretical frequency in all cells was ≥ 5 to assess group differences. Otherwise, Fisher’s exact test was used to assess group differences. And the 95% CI was estimated using the Clopper-Pearson method. The same methods used for the primary outcome were also applied to analyze the proportion of patients with SpO2 ≤ 93% on room air and the rate of emergency department visits or hospitalizations due to COVID-19 progression. For the secondary time-to-event outcomes, Kaplan-Meier (KM) estimates were calculated, with the 95% CI was estimated using the Brookmeyer-Crowley method and log–log transformation for normal approximation. Missing data were assumed to be Missing at Random (MAR) or Missing Completely at Random (MCAR). No imputation was performed to address the missing values. In addition, the descriptive analyses were conducted on the observational cohort.

Analyses of the primary outcome were performed in five subgroups: the use of antiviral medications, the presence or absence of risk factors, disease severity (mild/moderate), infection history (first vs. prior) and enrollment time. Statistical methods consistent with the primary endpoint were used in the analysis.

The safety analysis was performed on the FAS population, and data were collected on the rate and severity of ADRs, including those leading to discontinuation of the study intervention and those resulting in death. Analyses were done with the statistical analysis software SAS 9.4.

Results

Patients

A total of 344 participants were screened from May 26, 2023, through September 10, 2024. One hundred ninety-three patients were randomized, of whom 95 were assigned to the corticosteroid group and 98 to the SOC group. One hundred and thirty-four patients were included in the observational cohort. The FAS set included 92 patients in the corticosteroid group and 96 patients in the SOC group. Among the PP population, 84 patients were included in the corticosteroid group and 86 patients in the SOC group (Fig. 2). The demographic and baseline characteristics are shown in Table 1. In the corticosteroid group, the median age was 71.0 years (IQR 68.0–76.5), with 70 (76.1%) had received SARS-CoV-2 vaccination, 60 patients (65.2%) had mild COVID-19, while 32 patients (34.8%) had moderate COVID-19. In the SOC group, the median age was 73.5 years (IQR 68.0–78.0), with 69 (71.9%) had received SARS-CoV-2 vaccination, 68 patients (70.8%) had mild COVID-19, while 28 patients (29.2%) had moderate COVID-19. Additional baseline characteristics, as well as data on antiviral therapy and corticosteroid use, are available in the Supplementary Material 1.

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

Trial profile

Table 1 Demographic and clinical characteristics of the full analysis population at enrollment
Full size table

Primary outcome

In the FAS population, 1 (1.09%) of 92 patients progressed to severe or critical COVID-19 in the corticosteroid group compared with 4 (4.17%) of 96 in the SOC group, and the rate of severe or critical COVID-19 within 28 days was not significantly different between the corticosteroid group and the SOC group (relative risk, 0.26; 95% CI, 0.03 to 2.29; p = 0.3688). Notably, the wide 95% confidence interval (0.03 to 2.29) for the relative risk in the FAS should be interpreted in the context of limited statistical power, which may be attributed to the small number of events (1 case in the corticosteroid group vs. 4 cases in the SOC group). Results were similar in the PP population (1.19% vs. 2.33%; relative risk, 0.51; 95% CI, 0.05 to 5.54; p = 1.0000) (Table 2).

Table 2 Primary and secondary outcomes
Full size table

Secondary outcomes

For secondary outcomes, there were no significant differences between the corticosteroid group and the SOC group concerning several key outcomes. In the FAS population, the proportion of patients with SpO2 ≤ 93% on room air was 0% in the corticosteroid group compared to 4.17% in the SOC group (p = 0.1213). Additionally, in the FAS population, the rates of emergency department visits or hospitalizations were 1.09% (due to respiratory failure required mechanical ventilation) in the corticosteroid group versus 0% in the SOC group (p = 0.4894) (Table 2). On day 14, 98.8% of patients in the corticosteroid group and 97.6% in the SOC group tested negative for SARS-CoV-2. One death was reported in the SOC group within 28 days, involving a patient who died after exiting the study, with the cause of death being unknown (Table 3).

Table 3 Additional secondary outcomes
Full size table

All ADRs were associated with antivirals, and no ADRs related to the corticosteroids (Table 4).

Table 4 Adverse drug reactions
Full size table

Subgroup analyses

Patients enrolled in 2023 and those enrolled in 2024 were compared. Patients enrolled in 2024 exhibited a significantly lower rate of severe or critical COVID-19 compared to those enrolled in 2023 (0% vs. 3.35%, p = 0.0353) (Table S6). Regarding the prior history of COVID-19 infection, 6 (4.08%) of 147 participants in the first infection group experienced severe or critical COVID-19 compared to 0 (0%) of 175 in the prior infection group, indicating a significant difference between the two groups (p = 0.0086) (Table S7). Additional subgroup analyses results are provided in the Supplementary Material 1.

Discussion

This randomized controlled trial evaluated the effect of corticosteroids on disease progression. Early termination resulted in insufficient statistical power and failure to reach the predetermined sample size. No significant difference was observed in progression to severe or critical COVID-19 within 28 days between the corticosteroid group and the SOC group. In addition, the RR has a wide confidence interval (95%CI 0.03–2.29), which indicates that the existing data cannot determine the potential benefits or risks of the intervention. Consequently, current evidence is inadequate to support or refute the routine use of corticosteroids for this indication. The result is inconclusive due to several factors. First, this trial only enrolled in a limited number of subjects and has insufficient statistical power to detect significant differences between the treatment groups. Secondly the overall progression rates are low across the study cohorts. Existing studies have shown that the symptoms associated with the Omicron variant are generally milder, especially among vaccinated individuals [23,24,25,26]. In addition, the use of antivirals may further contribute to the low progression rate [26]. According to previous studies, adults with mild to moderate viral diseases who receive antivirals earlier can have a lower chance of developing severe disease [27, 28]. In this study, over 90% of patients in the randomized cohorts received antivirals, and the widespread use of antivirals may be associated with the low rate of progression to severe or critical disease in COVID-19 patients. It should be noted that the rate of severe or critical COVID-19 in the randomized cohort (4.17% in the SOC group) was significantly higher than that in the observational cohort (0.75%), suggesting that the randomized cohort may have been enriched with high-risk individuals (meeting criteria of (a) body temperature ≥ 38 °C, or (b) c-reactive protein (CRP) ≥ 27 mg/L, or (c) radiological findings indicative of pneumonia due to SARS-CoV-2 infection). Since such patients were excluded from the observational cohort, this risk stratification difference may explain the higher rate of severe or critical COVID-19 in the randomized group. In this high-risk subgroup, the corticosteroid group showed a trend toward reduced the risk of severe disease (RR = 0.26, 95%CI 0.03–2.29). Although the difference did not reach statistical significance (p = 0.3688), these findings provide preliminary evidence supporting the potential clinical value of corticosteroid in this high-risk subgroup, and further verification of their definitive efficacy is warranted in large-scale randomized controlled trials.

Up to now the role of corticosteroids in preventing COVID-19 progression is still unclear. Results from existing studies are controversial. Some studies on inhaled corticosteroids have indicated potential protective effects in preventing severe COVID-19. The STOIC trial included patients aged 18 and older with COVID-19, focusing on COVID-19-related emergency department visits as the primary endpoint. The study data indicated that short-term treatment with budesonide may effectively treat adults with early COVID-19, leading to a 91% relative reduction in clinical deterioration [29]. The PRINCIPLE trial included individuals aged 65 and older or those aged 50 and older with comorbidities who were PCR positive or had suspected COVID-19 symptoms. The budesonide group demonstrated a reduction in COVID-19-related hospitalizations and mortality rates (6.8% vs. 8.8%, OR = 0.75) and a shortened recovery time [30]. However, in the subgroup findings in the RECOVERY trial, mortality was higher in the dexamethasone group than in the SOC group in patients not receiving oxygen (20% vs. 17%, RR = 1.19, 95% CI 0.92–1.55), suggesting that dexamethasone may not be an effective treatment option for patients not receiving oxygen and may potentially increase the risk of death [31]. In our study, one patient in the SOC group died after withdrawing from the study, with the cause of death remains unclear. Nevertheless, when using corticosteroids in patients not receiving oxygen, clinicians still need to carefully assess the potential risks and benefits.

In the subgroup analysis according to enrollment time, all the subjects that progressed to severe or critical COVID-19 were enrolled in 2023, and there were none with progression in patients enrolled in 2024. These findings also suggest decreasing overall disease severity caused by COVID-19 over time. This trend is consistent with hospitalization rates reported by the CDC in the United States, which shows rates of 1.1%-7.8% in 2023 and 1.0%-2.9% in 2024 [32]. We also find that all cases of severe or critical disease occurred in patients with a first infection, all of whom were enrolled in 2023. This suggests that natural immunity from prior infection may provide greater protection against severe disease, consistent with previous reports.

Corticosteroids are known to reduce lung injury and systemic inflammation and alleviate symptoms such as fever [31]. The PRINCIPLE trial indicated that the recovery time in the budesonide group was 11.8 days compared to 14.7 days in the SOC group (HR = 1.21). Another study found that clinical recovery was 1 day shorter in the budesonide group compared with the usual care group (p = 0.007) [30]. However, in our study, the time to initial alleviation of all symptoms and time to sustained recovery of all symptoms were similar between the corticosteroid group and the SOC group. This might be attributed to confounding factors: (1) the predominantly mild to moderate baseline symptoms in participants, which shorten the natural recovery window; and (2) the high prevalence of antiviral use (> 90% in both arms), potentially accelerating symptom resolution and obscuring intergroup differences.

In this study, one patient in the corticosteroid group and one in the SOC group had persistently positive results in nucleic acid or antigen tests during the 14-day testing period. Six patients in this study showed antigen reversion to be positive after the seventh day of follow-up (4 in the SOC group, 1 in the corticosteroid group and 1 in the observational cohort). Despite no significant delays in the corticosteroid group observed, these results should be interpreted with caution due to the limited sample size.

All ADRs associated with antivirals observed in this study included mild gastrointestinal discomfort, pruritus and vomiting. These results are consistent with safety data reported in previous studies [33,34,35], indicating that these ADRs are relatively common and usually mild, making them manageable reactions during antiviral treatment. While this study did not identify any significant safety concerns, clinicians should remain vigilant about the well-established risks associated with corticosteroid therapy (e.g., immunosuppression, metabolic disturbances). It is recommended that a benefit-risk assessment be individualized according to each patient’s characteristics.

This trial has notable limitations. Only 189 participants were enrolled in the randomized cohorts, reducing statistical power and limiting subgroup analyses, which hinders a comprehensive assessment of treatment effects. The open-label design potentially biased the objectivity of assessing subjective symptom outcomes. Specifically, the duration of respiratory symptoms in the Corticosteroids group (10.0 days) was longer than that in the SOC group (8.0 days), which is inconsistent with the expected anti-inflammatory effect of corticosteroids. This discrepancy may be attributed to assessors applying stricter criteria for symptom resolution in the corticosteroid group. Despite these limitations, objective indicators (e.g., time to nucleic acid negativity) showed no significant differences among the three groups, suggesting that the open-label design had a limited impact on the main conclusions. The wide confidence intervals observed for key outcomes (e.g., 0.03–2.29 in the FAS and 0.05–5.54 in the PP population) reflect limited statistical power, primarily due to the low event rate of severe/critical COVID-19 in both groups. This highlights the need for larger-scale studies to confirm these findings. And for safety endpoints, only ADRs related to corticosteroids and antivirals were collected. Lastly, the lack of collected viral genotypes limits the assessment of SARS-CoV-2 mutations and their potential impact on patient prognosis.

Conclusively, as this study is exploratory in nature due to insufficient power, it did not demonstrate that combining oral corticosteroids with SOC could decrease the rate of progression to severe or critical disease within 28 days among elderly with mild to moderate COVID-19. However, corticosteroids demonstrated a good safety profile in this population.