Association factors with severe cases and antiviral drug assessment in patients with COVID-19

Due to the latent onset of novel coronavirus disease 2019 (COVID-19), it is important to identify patients with increased probabilities for disease progression early in order to implement timely medical strategies. This study aimed to identify the factors associated with severe COVID-19 and evaluate the current antiviral drugs, especially in severe patients. This was a retrospective observational study performed at the 7 of China) with hospitalized patients conrmed with COVID-19 from 11 to March 13, 2020. Multivariable logistic regression analysis was used to identify the associated factors of severe COVID. Treatment of antivirus drugs were collected and evaluated. to be associated with electrolyte disorders. Blood gas analysis revealed that 75.2% (124/165) of severe patients had an oxygen index (OI) < 300 at admission, of which 26 patients had OI < 100. There was no difference in the proportion of patients with hyperlactatemia between the two groups (P = 0.172). study was the assessment of current antiviral drugs. The antiviral drugs used during the study period were arbidol, oseltamivir, ribavirin, lopinavir/ritonavir, and α- interferon. The results suggest that arbidol might provide more benets compared with ribavirin in severe patients treated with monotherapy, but the difference between arbidol and oseltamivir was not signicant (P = 0.391). At the same time, in severe patients who received combination therapy, the combinations that included arbidol showed better benets. A multicenter randomized controlled clinical trial published on medRxiv recently showed that patients treated with favipiravir had a better recovery rate (71.4% vs. 55.9% P = 0.0199) but more side effects were observed compared with arbidol [15, 17]. There are multiple antiviral drugs being evaluated and tested in trials currently [18], but before better options can be justied, the use of arbidol might be recommended for its relative safety and effectiveness prole.

Because as much as 2-3 weeks can elapse between virus exposure and symptom onset, the early identi cation of patients with an increased likelihood of disease progression is important in order to implement timely medical strategies and to adjust them according to the evolving conditions, especially in the context of the exhausted healthcare systems around the world. Therefore, the aim of the present study was to identify the factors associated with severe COVID-19 and evaluate the current antiviral drugs, especially in patients with severe COVID-19.

Study design and patients
This single-center, retrospective, observational study was performed at the No. 7 Hospital of Wuhan (Wuhan, China), which is a designated hospital to treat patients with COVID-19. The medical team from the Second Hospital of Hebei Medical University was appointed by the government to provide medical assistance to the No. 7 Hospital of Wuhan during the outbreak. All hospitalized patients diagnosed as "viral pneumonia" from January 11 to March 13, 2020, were preliminarily included in this study. Patients con rmed with COVID-19 were then enrolled in the study. The diagnosis was made following the Chinese COVID-19 management guideline (versions 3 to 7) [11]. Patients with atypical clinical symptoms or chest radiology changes combined with negative SARS-CoV-2 RNA test results were excluded from this study. The study was approved by the Institutional Ethics Board of the No. 7 Hospital of Wuhan and the Second Hospital of Hebei Medical University (#2020-R016). The need for individual consent was waived due to the non-interventional and retrospective nature of this study.

Data collection and de nition
The patients' electronic medical records, including epidemiology parameters, clinical presentation, laboratory results, imaging characteristics, treatments, and disease outcomes, were collected and analyzed by the same designated physicians who accepted su cient training. Several important time points were also analyzed, including: disease onset, time from disease onset to dyspnea, time for SARS-CoV-2 RNA to be no longer detectable in patients with positive RNA results upon hospital admission, and average hospital stay. For patients who required mechanical ventilation, time from disease onset to ARDS, and time to mechanical ventilation were analyzed.
Disease onset was de ned as the time of patients starting to present symptoms. The severe cases were identi ed according to the Chinese COVID-19 management guideline (versions 3 to 7) [11]. Disease progression was identi ed and classi ed when the patients had one of the following criteria: 1) respiratory distress with respiratory frequency ≥ 30/min; 2) pulse oximeter oxygen saturation ≤ 93% at rest; and 3) oxygenation index (artery partial pressure of oxygen/inspired oxygen fraction, PaO2/FiO2) ≤ 300 mmHg [11]. Disease improvement was de ned as: patients' situation remained unchanged; severe cases changed to non-severe cases; and patients were permitted for discharge. The discharge criteria were: body temperature returned to normal and maintained for more than three consecutive days; signi cantly improved respiratory symptoms; a signi cant improvement on imaging and a negative result on RNA tests with two consecutive sputum samples or nasopharyngeal swabs or other respiratory samples (at least 24 h between each sampling) [11].

Laboratory tests
The pharynx swabs of suspected patients were collected and transported to the clinical laboratory of Zhongnan Hospital of Wuhan University for RNA detection following strict standard procedures. The presence of SARS-CoV-2 in pharynx swabs was detected by real-time RT-PCR. The detailed analysis and detection processes can be found in a previous study [3]. Laboratory tests and radiologic assessments, including chest X-ray or computed tomography (CT), were performed on the basis of the state of illness.

Treatment
Treatment was provided according to the Chinese COVID-19 management guideline (versions 3 to 7) [11], combined with the clinical characteristics of the patients and the actual situation of the medical resources of No. 7 Hospital of Wuhan during the epidemic. Patients with a mild condition were given general support like resting in bed, supportive treatment, antiviral treatment, and antibiotics if necessary. Severe patients were given respiratory and other organs support treatment on an individualized basis.

Statistical analysis
Categorical variables were presented as frequencies and percentages and analyzed using the chi-square test or Fisher's exact test, as appropriate. Continuous variables were presented as mean ± standard deviations or medians (interquartile range (IQR)) according to the results of the Kolmogorov-Smirnov test and analyzed using Student's t-test or the Mann-Whitney U-test, according to the distribution. Variables were rst screened with univariable logistic regression; variables with P-values < 0.05 for association with severe COVID-19 were included in a multivariable logistic regression analysis. Considering the earlier analysis of the total number of deaths (n = 52) in this study and to avoid model over tting, the six variables with the strongest association were selected for the multivariable logistic regression analysis on the basis of previous ndings, clinical constraints, and excluding covariables (symptoms, white blood cells (WBC), and procalcitonin (PCT)). All statistical analysis was performed using SPSS 22.0 (IBM, Armonk, NY, USA). Two-sided (except for the chi-square test) P-values < 0.05 were considered statistically signi cant.

Patient and public involvement
This was a retrospective case series study, and no patients were involved in the study design or in setting the research questions or the outcome measures directly. No patients were asked to advise on the interpretation or writing up of results.
The median time from disease onset to admission was 9 (IQR, 6-14) days, the time from disease onset to dyspnea was 0 (IQR, 0-7) days. In this cohort, 69 patients eventually developed ARDS and needed mechanical ventilation. The average time from disease onset to ARDS was 10 (IQR, 6-15) days among patients who eventually developed ARDS, and the mean time to mechanical ventilation was 10 (IQR, 6-15) days in the same subgroup.
Compared with the patients who did not progress to the severe condition, the severe patients were generally older and with a higher proportion of males (n = 104, 58.4%, P < 0.001). Patients with clear exposure histories were more often non-severe (P < 0.001), while the exposure history was not traceable in most severe patients (P < 0.001). For clinical symptoms, most of the non-severe patients did not have a fever upon hospital admission (n = 157, 42.3%, P < 0.001). Patients presenting with moderate or severe fever were more likely to have disease progression (P < 0.001). In addition, dyspnea, fatigue, chill, sputum production, and tachycardia were more common in severe patients. Severe patients were frequently associated with multiple clinical symptoms, especially the classic triple signs (n = 76, 42.7%, P < 0.001) ( Table 1).

Laboratory and imaging ndings
Upon hospital admission, all patients underwent relevant laboratory examinations in order to assess the patients' condition and guide treatments ( Table 2). The results indicated that 23.7% of the patients (119/502) had leukopenia, which was more frequently seen in severe patients (P < 0.001). In patients with lymphocyte count < 1.0 × 10 9 /L, 130/169 (76.9%) patients eventually developed severe disease. The levels of other in ammatory indicators such as procalcitonin (PCT), highly-sensitive C-reactive protein (hsCRP), and erythrocyte sedimentation rate (ESR) were increased in severe patients compared to non-severe patients (P < 0.001). In addition, myocardial enzymes were elevated in severe patients, and 85/128 (66.4%) of severe patients presented elevated NT-proBNP levels (P < 0.001). Elevation of alanine and aspartate aminotransferase occurred more frequently in severe patients and 163/175 (93.0%) severe patients had hypoproteinemia (P < 0.001). A relatively small number of patients developed a reduced glomerular ltration rate, but it was more commonly seen in severe patients (24/176, 13.6%, P < 0.001). Furthermore, more patients in the severe group (108/133, 81.2%) had elevated D-dimer levels compared to non-severe patients. Moreover, severe patients were more likely to be associated with electrolyte disorders. Blood gas analysis revealed that 75.2% (124/165) of severe patients had an oxygen index (OI) < 300 at admission, of which 26 patients had OI < 100. There was no difference in the proportion of patients with hyperlactatemia between the two groups (P = 0.172).  Lymphocyte count, troponin T, serum creatinine, D-dimer level, and OI were closely monitored and compared between the severe and nonsevere groups (Fig. 1). The lymphocyte counts were lower in severe patients but increased more robustly after day 7 compared with non-severe patients. Troponin T and D-dimer levels were higher in severe patients and peaked around the 4th day after admission. There was no signi cant difference in creatinine levels between the two groups except on 1st day of admission. In addition, severe hypoxemia was more common in severe patients.
In this cohort, only a very small number of patients were co-infected with other pathogens such as bacteria, in uenza virus, and atypical pathogens (
a: Radiographic ndings include the ndings of both chest X-ray and lung CT scan.

Treatments
All patients (100.0%) were given intermittent or continuous oxygen inhalation therapy to improve the clinical symptoms ( Among all patients, 81.6% were treated with antiviral drugs, and the remaining 18.4% were treated only with traditional Chinese medicine. The outcomes of the patients treated with antiviral drugs are shown in Table 5. The antiviral drugs used in this study were arbidol (n = 240, 43.6%), oseltamivir (n = 216, 39.3%), ribavirin (n = 152, 27.6%), lopinavir/ritonavir (n = 21, 3.8%), and α-interferon (n = 20, 3.6%). Arbidol was more effective than ribavirin (73.3% vs 41.2%, P = 0.029) in treating severe patients as single-drug therapy. Similarly, in severe patients who were treated with two drugs, arbidol combined with ribavirin or oseltamivir also had better e cacy. There were no signi cant differences identi ed among the other treatments. Some patients also received immunotherapies, including human immunoglobulin infusion (n = 52, 9.5%) and thymosin (n = 10, 1.8%). Vasoactive drugs were used in 34 severe cases, and continuous blood puri cation therapy was used in two cases. The data were expressed in the form of n/N (%), where n represents the number of patients with clinical improvement, N represents the total number of patients receiving corresponding drugs.

Discussion
Because the symptoms of COVID-19 can take up to 3 weeks to develop fully [1][2][3][4], the early identi cation of patients at higher risk of severe disease is important to implement timely medical strategies. This could save time and energy in the context of the exhausted healthcare systems. Therefore, this study aimed to identify the factors associated with severe COVID-19 and evaluate the current antiviral drugs, especially in severe patients. The results suggest that age ≥ 60 years, D-dimer > 0.243 µg/ml, and lower oxygenation index were associated with severe COVID-19. Therefore, the patients presenting those characteristics could be more aggressively managed from start in order to prevent complications. In addition, arbidol might provide more clinical bene ts in treating patients with severe COVID-19 compared with other antiviral drugs.
A total of 550 patients diagnosed with COVID-19 were included in this study. Inconsistent with the literature, there were more females (53.1%) than males in the present study [1][2][3][4], but there were more male patients among the severe cases. This discrepancy can be due to many factors, including the transmission route, willingness to undergo screening, and socioeconomical factors. Epidemiology tracing identi ed 170 (30.9%) of patients with 2019-nCoV having a history of contact with an infected individual and 38 (6.9%) due to a recent visit to a COVID-19designated hospital. The remaining of the patients had no clear source of infection. Those results highlight the need for refraining from having contacts and to enforce physical distancing, to avoid visiting hospitals known to treat COVID-19 and visiting other hospitals, and that many patients might have been infected through asymptomatic patients, either because those patients were asymptomatic carriers or because symptom onset sis not occur yet. This will have to be examined in future studies.
In this study, 42.3% of the non-severe patients did not have fever at diagnosis, which was lower than what was reported by Guan et al. [4]. Upon hospital admission, 42.7% of the patients who eventually progressed to severe COVID-19 had the typical triple signs (fever, cough, and dyspnea), while the triple signs were observed in only 14.5% of the non-severe patients. Unsurprisingly, severe patients often had more comorbidities. Hypertension, diabetes, cardiovascular diseases, and malignancy were the most common underlying diseases observed in patients with severe COVID-19. Older age was also observed to be associated with severe COVID-19, but whether this is because older patients can be more frail and weaker or because older individuals often have more comorbidities is still unknown.
As for the laboratory tests, 66.7% of the patients in the study had normal leukocyte count, and a quarter of the patients had decreased WBC counts. For 9.6% of patients who had an elevated WBC, secondary infections were often the cause of the elevated WBC. In addition, 49.8% of the patients presented with decreased lymphocyte counts, of which 52.0% (130/250) were severe cases. The incidence of anemia and thrombocytopenia was 36.9% and 6.8%, respectively, without differences according to disease severity. Cardiac enzymes, troponin T, transaminase, creatinine, and other organ injury indicators were also increased to varying degrees in some patients. Both hsCRP and ESR were increased in most patients, especially in severe patients. This highlights the systemic nature of the disease and that the patients should be comprehensively assessed. The increased in ammatory indicators suggested that SARS-CoV-2 tips the balance of the immune system towards a cytokine storm that contributes to patient deterioration and mortality, as observed in various infections [21,22]. Recent biopsy reports by Xu et al. [23] also indicated an increase of proin ammatory CCR4 + CCR6 + Th17 cells in the peripheral blood that might lead to systemic in ammatory responses and contribute to diffuse alveolar injury and pulmonary hyaline membrane formation. That evidence suggests that the systemic in ammatory response is an important factor leading to poor COVID-19 prognosis, as supported by the literature [21,22]. Unfortunately, due to the limited conditions of the hospital, no cytokine or other related testing was performed in this study. Future studies should aim to carefully examine the various cytokines involved in COVID-19 and in relation to disease severity.
It has been estimated that the mortality rate in severe cases was over 50% [24]. Therefore, it is critical to identify patients with an increased risk of disease progression. In the present study, the multivariable logistic regression analysis showed that age ≥ 60 years, D-dimer > 0.243 µg/ml, and decreased OI might be risk factors for patient deterioration. The importance of aging in determining the COVID-19 prognosis was consistent with previous studies that aimed to identify prognosis factors for SARS or MERS infections [25][26][27][28]. The coagulation dysfunction we observed in this study was consistent with previous studies [1][2][3]5]. Severe patients were more likely to develop coagulation and brinolysis disorders, especially the elevation of D-dimer levels. Similar to other severe viral pneumonia, the cause might be the sepsis-induced in ammatory cytokine storm affecting multiple endogenous and exogenous coagulation pathways and brinolysis that ultimately lead to thrombosis formation [16,21]. Therefore, special attention should be paid to severe patients with long-term bed rest, advanced age, and complicated underlying diseases, especially in the presence of coagulation abnormalities. Appropriate anticoagulant treatment might be considered in such patients in order to prevent the occurrence of deep vein thrombosis (DVT) and related complications [29][30][31]. In the present study, a reduction in OI was associated with increased mortality. Similarly, Liu et al. [32] found that the lung injury Murray score and OI could predict the prognosis of COVID-19. Therefore, early recognition of these three indicators upon hospital admission is critical, so appropriate medical strategies can be adjusted, and more importantly, the nearly exhausted medical support force can be redistributed. This is especially important because when the patients are admitted, the exact interval between infection and symptom onset is unknown and the exact time until an eventual disease progression is also unknown.
In this study, 435 patients (79.1%) received at least one antibiotic in the hospital, but only 110 (20.0%) of them were con rmed with secondary bacterial infection (some cases were accompanied by fungal infection). Therefore, more attention should be paid to the indication of antibiotic use and avoid antibiotic overuse. Prophylaxis for eventual complicating secondary bacterial or fungal infections can be indicated in some cases, but additional studies are necessary to determine who they might be. In addition, 122 severe patients received intravenous or oral glucocorticoid treatment. Among these patients, 79 had an improved condition but 43 eventually died. Nevertheless, the rate of improvement was relatively high (83.9% vs. 64.8%, P = 0.009) in severe patients who did not receive glucocorticoids. It is important to point out that the patients who received glucocorticoids also had a higher rate of secondary infections compared with patients who did not receive glucocorticoids (36.9% vs. 19.6%, P = 0.021). This is consistent with several recent studies that suggested that glucocorticoids are not bene cial for patients with ARDS and viral infections [33,34], but contradicts recent ndings that suggest that corticosteroids decrease the mortality of COVID-19, but the level of evidence is low [35]. Even though this study was not powered to analyze the bene t and risks of using glucocorticoids in COVID-19, the data suggest that glucocorticoids failed to improve the prognosis and increased the risk of secondary infection.
Another important feature of this study was the assessment of current antiviral drugs. The antiviral drugs used during the study period were arbidol, oseltamivir, ribavirin, lopinavir/ritonavir, and α-interferon. The results suggest that arbidol might provide more bene ts compared with ribavirin in severe patients treated with monotherapy, but the difference between arbidol and oseltamivir was not signi cant (P = 0.391). At the same time, in severe patients who received combination therapy, the combinations that included arbidol showed better bene ts. A multicenter randomized controlled clinical trial published on medRxiv recently showed that patients treated with favipiravir had a better recovery rate (71.4% vs. 55.9% P = 0.0199) but more side effects were observed compared with arbidol [15,17]. There are multiple antiviral drugs being evaluated and tested in trials currently [18], but before better options can be justi ed, the use of arbidol might be recommended for its relative safety and effectiveness pro le.
Using the MuLBSTA scoring system, 46 (88.5%) patients were correctly classi ed as at high-risk for death (score > 12), but only 22 (42.3%) were correctly classi ed as high-risk (score ≥ 2) when using the CURB-65 scoring system. This suggests that the MuLBSTA scoring system is more effective in the mortality risk assessment of patients with COVID-19 (P < 0.001) on the early-stage of the disease. This is consistent with a previous study [36]. We speculated that the reason why the MuLBSTA scoring system was more effective is that its scoring criteria (age ≥ 60, smoking status/smoking cessation history, hypertension history, imaging showing multiple lobar in ltrations, lymphocyte counts ≤ 0.8 × 10 9 /L, or combined with a bacterial infection) can be achieved and evaluated at the early stage of the disease. On the other hand, the parameters analyzed in CURB-65 may not be elevated in the early stage of the disease in the high-risk population. If necessary, an attempt might be made to lower the scoring criteria and set 1 as a cut-off point to improve its sensitivity (44/52, 84.6% vs. 46/52, 88.5%, P = 0.566). Despite that the 2009 IDSA/ATS guidelines recommended CURB-65 as a severity assessment form for community-acquired pneumonia (CAP) [37], the present study suggests that the MuLBSTA scoring system might be a better assessment tool for COVID-19.

Limitations
This study has several limitations. First, this was a retrospective study conducted at a single center, with a cohort of 550 patients treated after the arrival of the Hebei medical team, which might not necessarily represent the general population of patients. In addition, the false-negative rates of current SARS-CoV-19 tests are relatively high and might bias the results. Last but not least, due to the retrospective nature of this study and the lack of diverse drugs in the early stage of the epidemic, the observation of the bene ts for different antiviral drugs needs to be further con rmed in future randomized controlled trials.

Conclusion
In conclusion, patients with severe COVID-19 are often complicated, with more comorbidities, and have a more variable presentation compared with non-severe patients. Age ≥ 60 years, D-dimer > 0.243 µg/ml, and lower OI was independently associated with disease progression. The use of glucocorticoids should be cautioned. Arbidol might have bene ts in treating severe patients. Patients with or without positive SARS-CoV-19 RNA tests showed similar symptoms and demographic characteristics. Finally, the MuLBSTA scoring system might be a better assessment tool for COVID-19 compared with CURB-65.