Remdesivir (Veklury, Gilead Sciences, Foster City, CA, USA) and nirmatrelvir–ritonavir (Paxlovid, Pfizer, New York, NY, USA) were reported to improve the outcome of patients with mild-to-moderate COVID-19 symptoms. Preclinical data suggest that nirmatrelvir–ritonavir might be more effective than remdesivir alone or in combination with nirmatrelvir–ritonavir for people at high risk of severe COVID-19. We aimed to assess the safety and effectiveness of combining remdesivir and nirmatrelvir–ritonavir compared with using each drug alone for adults hospitalised with COVID-19.

In this target trial emulation study, we used electronic health records of patients aged 18 years or older who received either combination treatment of nirmatrelvir–ritonavir and remdesivir or monotherapy of either drug between March 16 and Dec 31, 2022, within 5 days of hospitalisation for COVID-19 in Hong Kong. Inverse probability of treatment weighting was applied to balance baseline patient characteristics across the treatment groups. The primary outcome was all-cause mortality. Cox proportional hazards regression adjusting weighting was used to compare the risk of all-cause mortality, intensive care unit (ICU) admission, or ventilatory support for 90 days of follow-up between groups.

Between March 16 and Dec 31, 2022, 18 196 participants were identified from electronic health records and assigned to receive remdesivir (n=4232), nirmatrelvir–ritonavir (n=13 656), or nirmatrelvir–ritonavir and remdesivir (n=308). By applying an inverse probability of treatment weighting, a weighted sample composed of 18 410 recipients of nirmatrelvir–ritonavir and remdesivir combination treatment, 18 178 recipients of remdesivir monotherapy, and 18 287 recipients of nirmatrelvir–ritonavir monotherapy was obtained. After a median follow-up of 84 days (IQR 45–90), risk of mortality was lower in patients who received nirmatrelvir–ritonavir monotherapy (hazard ratio [HR] 0·18 [95% CI 0·15 to 0·20]; absolute risk reduction [ARR] –16·33% [95% CI –16·98 to –15·68]) or remdesivir and nirmatrelvir–ritonavir combination therapy (HR 0·66 [95% CI 0·49 to 0·89]; ARR –6·52% [95% CI –7·29 to –5·74]) than in patients who received remdesivir monotherapy. Similar results were observed for ICU admission or ventilatory support (nirmatrelvir–ritonavir monotherapy: HR 0·09 [95% CI 0·07 to 0·11]; ARR –10·04% [95% CI –10·53 to –9·56]; combination therapy: HR 0·68 [95% CI 0·42 to 1·12]; ARR –3·24% [95% CI –3·84 to –2·64]). Compared with combination therapy, nirmatrelvir–ritonavir monotherapy was associated with lower risk of mortality (HR 0·27 [95% CI 0·20 to 0·37]; ARR –9·81% [95% CI –10·39 to –9·24]) and ICU admission or ventilatory support (HR 0·13 [95% CI 0·08 to 0·22]; ARR –6·80% [95% CI –7·22 to –6·39]).

Our study highlighted the potential for reduced risk of mortality, ICU admission, or the need for ventilatory support in patients hospitalised with COVID-19 treated with nirmatrelvir–ritonavir as a monotherapy compared with treatment regimens based on nirmatrelvir–ritonavir and remdesivir combination therapy or remdesivir monotherapy. Further randomised controlled trials are needed to support the validity of the current results.

The Health and Medical Research Fund Commissioned Research on COVID-19.

For the Chinese translation of the abstract see Supplementary Materials section.