The presence of SARS-CoV-2 was detected by the real-time quantitative polymerase chain reaction (qPCR) method, as previously reported . Viral ribonucleic acids (RNAs) were extracted from the samples using the QIAamp RNA Viral Kit (Qiagen, Hilden, Germany), and quantitative reverse transcription polymerase chain reaction (qRT-PCR) was performed using a commercial kit specific for SARS-CoV-2 detection (GeneoDX Co., Ltd., Shanghai, China), which was approved by the China Food and Drug Administration (CFDA) (rebranded and restructured as the National Medical Products Administration of the State Administration for Market Regulation of PRC since 2018). “Viral clearance” was defined as the presence of two consecutive negative results with qPCR detection over an interval of 24 h.
Safety was assessed by a standardized questionnaire for adverse events and by laboratory tests.
The quantitative data were described as the mean ± standard deviation, or as the median (min–max). The qualitative data were described by number of cases (proportion, %). Patient characteristics were compared using the χ2 test or Fisher’s exact test for categorical data, and the Wilcoxon rank-sum test or Student’s t test for continuous data. The factors affecting the changes in chest CT were analyzed using binary logistic regression. The analysis of viral clearance time was calculated using the Kaplan-Meier method and the difference analysis of the viral clearance time under different treatments was calculated using the log-rank test. Potential influencing factors of viral clearance were analyzed by univariate and multivariate Cox regression models. A P value lower than 0.05 was required for statistical significance. All of the analysis was performed using SPSS Version 22.0 and GraphPad Prism 7.0.
3.1. Patients and baseline analysis
From 30 January, 56 patients with laboratory-confirmed COVID-19 were screened, of which 35 were eligible for the FPV arm of the study. A total of 91 laboratory-confirmed COVID-19 patients who had started treatment with LPV/RTV between 24 January and 30 January 2020 were screened, of which 45 were eligible for the control arm of this study. All enrolled patients finished the therapy and were followed up for 14 d after the treatment began (Fig. 2). All the baseline characteristics were compared between the FPV arm and the control. As shown in Table 1, there were no significant differences between the baseline characteristics of the two arms.
3.2. Viral response to the antiviral therapy
The Kaplan-Meier survival curves for the length of time until viral clearance for both kinds of antiviral therapy were presented in Fig. 3. The median time of viral clearance for the patients treated with FPV, designated as Group A, was estimated to be 4 d (IQR: 2.5–9), which was significantly shorter than the time for patients in the control group, designed as Group B, which was 11 d (IQR: 8–13) (P < 0.001).
3.3. Chest CT changes in COVID-19 patients’ response to treatment
The non-parametric Mann–Whitney U test was used to determine the significance of the difference between the chest CT changes in response to the two different treatments (Table 2). Meanwhile, the improvement rates of the chest CT changes for the two arms of the study were compared on Days 4, 9, and 14 after treatment. No significant difference in the improvement rates was found between the two arms on Days 4 and 8 (P > 0.05). However, on Day 14 after treatment, the improvement rates of the chest CT changes in the FPV arm were significantly higher than those in the control arm (91.4% versus 62.2 %, 32/35 versus 28/45, P = 0.004).
3.4. Multivariate analysis of the changes in chest CT
Univariate analysis using χ2 test, t test, or Wilcoxon rank-sum test was conducted before multivariate analysis; the significant variables (P < 0.10) in the univariate analysis were as follows: antiviral therapy and whether or not fever was present. A multivariate logistic regression analysis was conducted to identify the independent factors affecting the changes in chest CT. We chose the change in chest CT (0 = no change or worse, 1 = improved) as the dependent variable, and the variables that were significant in the univariate analysis or were professionally significant (including age, underlying disease, and severity of disease in baseline) as the independent variables. The result showed that there were two statistically significant factors in the model: antiviral therapy (odds ratio (OR) = 3.190, 95% confidence interval (CI) = 1.041–9.78) and fever (OR = 3.622, 95%CI = 1.054–12.442). This means that antiviral therapy and fever were independent factors that affected the chest CT after we had controlled the confounding factors. The patients who were treated with FPV had greater improvement in chest CT (Table 3).
3.5. Multivariate analysis of viral clearance
Univariate analysis using the log-rank test and univariate Cox regression was conducted before the multivariate analysis; the significant variables (P < 0.10) in the univariate analysis were as follows: antiviral therapy, white blood cell (WBC), hemoglobin (Hb), platelet (PLT), Neutrophils, T lymphocyte count, and illness to treatment time. A multivariate Cox regression model was used to explore the independent factors affecting viral clearance. The time of viral clearance was set as the TIME variable, viral clearance (0 = no, 1 = yes) was set as the status, and the variables that were significant (P < 0.10) in the univariate Cox analysis or were professionally significant (including age, and whether underlying diseases were present or not) were set as independent variables. The result showed that the model was significant (P = 0.003). The significant factors were as follows: T lymphocyte count ((hazard ratio (HR) = 1.002, 95%CI = 1.000–1.005) and antiviral therapy (HR = 3.434, 95%CI = 1.162–10.148). This means that the treatment and T lymphocyte count were independent factors that affected the viral clearance after we controlled the other confounding factors. As the result shows, compared with LPV/RTV, FPV has a greater effect on viral clearance (Table 4).
3.6. Adverse reactions after medication
The total number of adverse reactions in the FPV arm of the study was four (11.43%), which was significantly fewer than the 25 adverse reactions (55.56%) in the control arm (P < 0.001). Two patients had diarrhea, one had a liver injury, and one had a poor diet in the FPV arm. Meanwhile, there were five patients with diarrhea, five with vomiting, six with nausea, four with rash, three with liver injury, and two with chest tightness and palpitations in the control arm (Table 5).
This study investigated the effect of FPV versus LPV/RTV on the treatment of COVID-19. It was found that FPV was independently associated with faster viral clearance and a higher improvement rate in chest imaging. These findings suggest that FPV has significantly better treatment effects on COVID-19 in terms of disease progression and viral clearance, as compared with LPV/RTV. FPV, which is known as a prodrug, is a novel RNA-dependent RNA polymerase (RdRp) inhibitor, which has been shown to be effective in the treatment of influenza and Ebola virus , , , , , . Recently, a report from Wang et al.  showed that both FPV and remdesivir are effective in reducing the SARS-CoV-2 infection in vitro(EC50 = 61.88 μmol·L−1, CC50 > 400 μmol·L−1, SI > 6.46). The finding of the preset study confirms the hypotheses conceived from the laboratory finding: that FPV is effective treatment for COVID-19.
The limitation of the present study is that it was not a randomized double-blinded placebo-controlled clinical trial, which led to inevitable selection bias in patient recruitment. However, given the high number of patients presenting simultaneously and the very high infectivity of the disease, it was ethically unacceptable to allocate patients to receive a different experimental drug using a randomization process impossible for most of the patients to understand. Furthermore, in the context of rumors and distrust of hospital isolation, using a randomized design at the outset might have led even more patients to refuse being isolated. Therefore, we chose to conduct a nonrandomized trial, in which patients consecutively admitted to the hospital during two separate periods were included in two groups, respectively. Importantly, all baseline characteristics of the two groups were comparable and the effectiveness of FPV remained significant after adjustment for potential confounders.
The current study also found that early viral clearance contributed to the improvement of chest imaging on Day 14. This finding suggests that improvement of the disease may depend on inhibition of the SARS-CoV-2, and that FPV controls the disease progression of COVID-19 by inhibiting the SARS-CoV-2. Until recently, the pathogenesis of COVID-19 had not been well clarified. Since the infection of SARS-CoV-2 was thought to be self-limited and characterized by systemic inflammation reaction, symptomatic and supportive treatment was mainly recommended by the WHO and the National Health Commission of the PRC. This description is similar to MERS-CoV, for which nonspecific therapeutic interventions are often introduced to prevent severe morbidity and mortality . How antivirals would contribute to control of the disease is controversial. Although there have been many registered clinical trials focusing on antiviral drugs for COVID-19, the timing, duration of treatment, and study endpoints have not been unified. In the current study, the time of viral clearance was introduced as a primary endpoint to evaluate the antiviral effect of FPV on the SARS-CoV-2 and successfully identify the priority of FPV. The relationship between the time of viral clearance and the improvement in CT image indicates that viral clearance is an ideal surrogate for the clinical endpoint. A limitation of the present study was that the relationship between the viral titer and the clinical prognosis was not well clarified. Future research could pay more attention to this point.
More adverse events were observed in the control arm than in the experimental arm, and were similar to the adverse events observed in studies of AIDS treated by LPV/RTV. It is worth mentioning that the treatment duration of FPV in the present study was twice as long as that used for the treatment of influenza. However, the adverse events in the experimental arm were rare and tolerable, and none of the patients needed to discontinue FPV treatment. These results seem to suggest that the treatment duration of FPV can be prolonged if necessary.
SARS-CoV-2 infection has now spread quickly all over the world. At present, no effective treatment has been demonstrated. The task at hand was to run a well-designed trial to identify effective treatments based on a high level of evidence. However, at the beginning of this study, certain conditions did not allow the randomization of patients to receive either standard care or an experimental drug. In this pilot study of a non-randomized control trial, we found that FPV showed significantly better treatment effects on COVID-19 in terms of disease progression and viral clearance; if causal, these results should be important information for establishing standard treatment guidelines to combat the SARS-CoV-2 infection. Furthermore, we introduced the time of viral clearance as a primary endpoint for experimental antiviral treatment and demonstrated it to be a surrogate of clinical endpoint; this will be helpful for designing COVID-19 research.
Lei Liu, Yingxia Liu, Qingxian Cai, Minghui Yang, and Jun Chen contributed to the study design. Qingxian Cai, Minghui Yang, Dan Shu, Junxia Xia, Xuejiao Liao, Dongjing Liu, Yuanbo Gu, Qiue Cai, Xiaohe Li, Jiaye Liu,Lin g Peng, Deliang Huang, and Jing Zhang contributed to the collection of clinical data. Qingxian Cai, Minghui Yang, Shurong Zhang, Fuxiang Wang, Li Chen, Shuyan Chen, Zhaoqin Wang, and Zheng Zhang contributed to the data analysis. Qingxian Cai, Minghui Yang, Jun Chen, Yang Yang, Chenguang Shen, Ruiyuan Cao, and Wu Zhong contributed to the manuscript preparation. All the authors have read and approved the manuscript.