Clinical course and risk factors for recurrence of positive SARS-CoV-2 RNA: a retrospective cohort study from Wuhan, China

The coronavirus disease 2019 (COVID-19) pandemic is caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The objective of this study was to determine the clinical course and risk factors for patients showing recurrent SARS-CoV-2 RNA positivity. A total of 1087 COVID-19 patients confirmed by RT-PCR from February 24, 2020 to March 31, 2020 were retrospectively enrolled. Advanced age was significantly associated with mortality. In addition, 81 (7.6%) of the discharged patients tested positive for SARS-CoV-2 RNA during the isolation period. For patients with recurrent RT-PCR positivity, the median duration from illness onset to recurrence was 50 days. Multivariate regression analysis identified elevated serum IL-6, increased lymphocyte counts and CT imaging features of lung consolidation during hospitalization as the independent risk factors of recurrence. We hypothesized that the balance between immune response and virus toxicity may be the underlying mechanism of this phenomenon. For patients with a high risk of recurrence, a prolonged observation and additional preventative measures should be implemented for at least 50 days after illness onset to prevent future outbreaks.


INTRODUCTION
The coronavirus disease 2019  pandemic is caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1,2]. As of April 18, 2020, 2,121,675 confirmed cases of COVID-19 and 142,299 related deaths have been reported from 213 countries according to the World Health Organization (WHO) [3]. Although several studies have summarized the epidemiological and clinical features of SARS-CoV-2 infection [4][5][6], and research is going on viral pathogenicity and mechanism. However, the exact origin of SARS-CoV-2 is controversial and a potential threat to a new outbreak [7,8]. Furthermore, little is known regarding the immune response against SARS-CoV-2 infection, which in turn makes it difficult to assess complete recovery with no further risk of infection. The latter is a crucial factor in "flattening the curve" of COVID-19 and preventing additional outbreaks.

AGING
In the early stages of the COVID-19 outbreak that was located to Wuhan, China, the severe shortage and limitations in the detection and accuracy of the RT-PCR test restricted identification of infected patients. The diagnostic techniques have improved substantially since [9], and two or more multipoint throat-swabs are taken over 24 hours apart prior to discharge in order to minimize the false negative rate of RT-PCR tests [10]. Lan L et al. [11] reported that four medical professionals with COVID-19 who met the criteria for hospital discharge (including two consecutive negative RT-PCR results) reverted to SARS-CoV-2 positivity, indicating a potential asymptomatic carrier state. It remains to be determined whether patients with recurrent SARS-CoV-2 RNA positivity remain infectious after discharge. Furthermore, the clinical and radiological characteristics of the COVID-19 patients with recurrence is largely unknown.
Herein, we retrospectively analyzed 1087 patients with confirmed COVID-19 and explored the clinical course and risk factors of SARS-CoV-2 RNA recurrence by RT-PCR during post-discharge isolation.

Clinico-demographic characteristics of patients
A total of 1087 consecutive COVID-19 pneumonia patients positive for SARS-CoV-2 RNA were enrolled in this study. The median age of the cohort was 60 years (9 to 100 years; IQR -49-69 years) and 635 (58.4%) of the patients were women. The majority (83.1%) of the cases were mild, whereas the proportion of severe and critical cases were 13.2% and 3.7% respectively. Most patients (874, 80.4%) had bilateral pulmonary infiltration on the chest CT, while 730 (67.2%) and 525 patients (48.3%) respectively showed ground-glass appearance and consolidation. In addition, 887 out of 1007 (88.1%) patients were positive for serum IgG, while 797 out of 1057 (75.4%) patients were positive for serum IgM against COVID 19.
The median length of hospitalization was 12 days (1-38 days; IQR, 8-17 days), and 20 patients died during hospitalization whereas 1067 were discharged. The total mortality rate was 1.8% and the discharge rate was 98.2%. Among the fatalities, 5 patients were graded as severe with mortality rate of 3.5%, and 15 were critical cases with a high mortality rate of 37.5%. The total mortality rate of the severe and critical cases was 10.6%. The median age of the deceased patients was 83 years (65 to 92 years; IQR, 79.3-87.8 years), which was significantly higher than that of the discharged patients (P<0.001). The main causes of deaths were multiple organ failure (MOSF), most commonly affecting the lungs, heart, liver and kidneys. Other clinical features, laboratory examinations and imaging findings are summarized in Supplementary Table 1.
Amongst these 81 patients, 37 (45.7%) received oxygen support. However, no invasive mechanical ventilation (IMV) or IMV with extracorporeal membrane  The median duration from illness onset to initial RT-PCR confirmation, onset of complete RNA negative status and recurrent RT-PCR positivity after discharge, and from discharge to recurrence. (B) The median duration from initial RT-PCR confirmation to onset of complete RNA negative status and recurrent RT-PCR positivity after discharge, and from onset of complete RNA negative status to recurrence. SARS-CoV-2=severe acute respiratory syndrome coronavirus 2. RT-PCR=reverse transcription-polymerase chain reaction. AGING oxygenation (ECMO) was used. The optimal antiviral therapy was administered in 69 (85.2%) patients, including arbidol hydrochloride (40 patients, 49.4%), interferon alfa (17 patients, 21.0%), entecavir/tenofovir (7 patients, 8.6%) and oseltamivir (5 patients, 6.2%). Fifty-one patients (63%) were treated with Chinese patented drugs, such as Lianhuaqingwen capsule. Vitamin C was given to 41 (50.6%) patients, and immunomodulators like thymopentin and immunoglobulin were administrated to 8 (9.9%) patients.

Associated risk factors with recurrence of positive SARS-CoV-2 RNA
As shown in Table 2 (Table 4).

DISCUSSION
In this study, we have provided comprehensive data on the demographic and clinical characteristics of 1087 consecutive COVID-19 patients from Wuhan, China. The majority (83.1%) of the cases in our cohort were mild, and the overall mortality rate of the severe and critical cases was 10.6%. The mortality rate of the entire cohort was 1.8%, which is consistent to one previous study [4] but lower than that reported in other studies [5,12]. This difference can be partly attributed to the higher proportion of severe cases in the other cohorts, as well as the greater medical resources that were allocated in the later stages of this pandemic wherein we enrolled patients for our study. Liang WH et al. [13] reported that the mortality of COVID-19 patients outside of the Hubei Province was limited to 0.3%, as strict public health interventions were initiated in order to prevent further outbreak outside Hubei and adequate medical resources were provided for treatment. In agreement with previous studies that identified older age as a risk factor of mortality in COVID-19 patients [6,14], the median age of the deceased patients in our cohort was 83 years, distinctly higher than that of the discharged patients (P<0.001), which further suggests that a higher age was significantly associated with mortality.
Among the 1067 patients that were discharged on the basis of negative SARS-CoV-2 RNA results, 81 (7.6%) patients reverted to positive state during their isolation period. Similar findings have been reported previously [11,15,16]. However, Yuan J et al. [16] reported a higher repeat positivity rate of 14.5% after discharge, which could be on account the smaller cohort of enrolled patients. These persistent asymptomatic viral carriers may pose a risk for potential future outbreaks despite unprecedented public health interventions [17]. Therefore, we explored the clinical course and risk predictors for recurrent SARS-CoV-2 PCR positivity in order to provide new insights into the disease and help guide the clinical practice against future outbreaks.
In our study, the median duration of viral shedding for patients with positive SARS-CoV-2 RNA recurrence was 33 days from the onset of illness to complete RNA negative status. However, the median duration from illness onset to SARS-CoV-2 RNA reversion was 50 days. Previous studies have reported on duration of viral shedding. Zhou F et al. [6] reported a 20 day median duration of viral shedding in survivors and the longest observed duration was 37 days. Furthermore, Zhou B et al. [18] reported that the median duration of viral shedding was 31 days from illness onset in severe COVID-19 patients. Xu K et al. [19] further showed that 3 out of 4 COVID-19 patients had viral RNA clearance within 21 days of illness onset, and male gender, older age, hypertension, delayed hospital admission, severe illness upon admission, invasive mechanical ventilation and corticosteroid treatment were risk factors for prolonged viral RNA clearance. Our findings underscore the importance of a prolonged treatment or isolation for patients at increased risk of recurrent SARS-CoV-2 RNA positivity. Nevertheless, we found that age and comorbidities that were previously described to be risk factors of mortality [14] were not identified as significant risk factors when compared to patients without reversion. Instead, high serum IL-6 levels, lymphocyte count greater than 1.1*10 8 /L and consolidation on CT imaging during hospitalization were associated with a higher likelihood of recurrent SARS-CoV-2 RNA positivity after discharge. This is consistent with a previous study that showed that the lymphocyte count prior to discharge was positively correlated with the time to virus reappearance, which confirms the role of lymphocytes in the potential recurrence of SARS-CoV-2 RNA positivity [16]. Other factors that influence the host defense against viral infections, such as clinical severity AGING of the disease, CRP, D-dimer level etc., were not significantly different between the recurrent versus non-recurrent groups. IL-6 is one of the major proinflammatory cytokines that are instrumental in clearing pathogens. However, the rapid multiplication of SARS-CoV-2 in the lower respiratory tract leads to excessive IL-6 production, which triggers an acute severe systemic inflammatory response known as cytokine release syndrome (CRS) [20]. In fact, the increased serum IL-6 levels in severe and critical COVID-19 patients is associated with poor outcomes [21,22], which was also observed during severe acute respiratory syndrome (SARS) outbreak [23]. Concurrently, lymphopenia is also common in patients with COVID-19, especially in severe and critical cases [5,22,24], suggesting a dysregulated immune response in this sub-cohort. In our study however, only 175 (16.1%) patients showed a decrease in lymphocyte count, which again may be can be attributed to the fewer severe cases. Interestingly, the discharged patients with recurrence of positive SARS-CoV-2 RNA had an elevated serum IL-6 level and lymphocyte count compared to those with no recurrence, indicating that the immune system may still be actively involved in clearing the infection. It is also possible that the immune responses can suppress but not completely eradicate SARS-CoV-2, which may have led to the falsenegative results due to lower viral loads. Once the virus started replicating again, the RT-PCR results reverted to positive in the discharged patients.
The chest CT imaging of COVID-19 pneumonia is a useful preliminary diagnostic tool that has lowered the rate of missed diagnoses [25]. Features of consolidation on CT imaging are associated with critical disease [26]. Progression of consolidation might indicate further infiltration of the lung parenchyma and lung interstitium due to virus invasion into the respiratory epithelium, which is characterized by diffuse alveolar damage and necrotizing bronchitis. This eventually leads to complete permeation of the alveoli with the inflammatory exudate [27,28]. Therefore, SARS-CoV-2 may persist in the respiratory epithelium during lung consolidation in the recovery phase of the infection, which eventually results in the recurrence of positive SARS-CoV-2 RNA after discharge. Interestingly, most patients with recurrence had fluctuating positive and AGING   Figure 1). This is a potential sign of recurrent SARS-CoV-2 positivity after discharge, and also partly ruled out the randomly error probability in RT-PCR detection for one case. Thus, the infected patients may have already been immune to the virus and require a period for complete recovery. However, if the immune response cannot deal with the recurrence, further treatment may be still needed.

Limitations
This study has a few limitations that ought to be noted. First, this study was conducted at a single-center hospital which may have introduced a selection bias that influenced the clinical outcomes. A larger multi-center or even global cohort study of COVID-19 patients would help further define the clinical characteristics and risk factors of recurrence. Second, only multipoint throat-swab AGING specimens were tested which increases the risk of false negative results. Therefore, multisite samples should be collected for RT-PCR detection, such as the fecal SARS-CoV-2 RNA test for patients with gastrointestinal symptoms [29]. Third, the retrospective design and initial lack of guidelines for drug administration made it difficult to analyze the impact of treatment regimens on the recurrence of positive SARS-CoV-2 RNA.

CONCLUSIONS
Elevated lymphocyte counts and serum IL-6 level, and consolidation on chest CT were associated with a greater risk of recurrent SARS-CoV-2 RNA positivity, possibly due to a balance between immune regulation and virus toxicity. For patients with a higher risk of recurrence, a prolonged treatment or isolation period for at least 50 days after illness onset is recommended in order to identify patients that may pose a risk for future outbreaks.

Data collection and follow-up
The epidemiological, radiographic, laboratory, treatment and treatment outcome data of these patients were extracted from medical records and through direct communication in order to establish a database. The SARS-CoV-2 RNA RT-PCR records from discharge to April 15, 2020 were obtained from the Health Wuhan App, a database containing all real-time results about SARS-CoV-2 RNA tests conducted in Wuhan. The patients were assigned a number for confidentiality. All data were evaluated by two authors (JC and QC) and thereafter by a third researcher (NP) in case of any differences in interpretation.

Clinical tests
In accordance with the standard procedure, throat-swab specimens were obtained and tested for SARS-CoV-2 infection using RT-PCR by the Academy of Military Medical Sciences and hospital laboratory [14]. The test was repeated during the hospital stay and after clinical remission of symptoms at 24-hour intervals. In addition, serum levels of SARS-CoV-2-specific IgM/IgG measured during hospitalization with the indirect enzyme-linked immunosorbent assay (ELISA) protocol using the N protein of SARS-CoV-2 as the coating antigen. Routine blood tests were performed to determine complete blood counts (including white blood cells, neutrophils, lymphocytes, monocytes and platelets), biochemical indices (liver function, renal function and electrolyte levels), coagulation indices, high-sensitivity C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), procalcitonin, myocardial enzymes, D-dimer and interleukin-6 (IL-6). Computed tomography (CT) scans were routinely performed as recommended by the attending physician.

Clinical definitions
The patients were discharged based on the following criteria: 1) no fever for at least three days, 2) remission of respiratory symptoms, 3) amelioration of pulmonary inflammation on the chest CT scan, 4) two negative SARS-CoV-2 RNA tests at least 24 hours apart, 5) overall good constitution.

Statistical analysis
Continuous and categorical variables were respectively presented as median with interquartile range (IQR) and counts with percentages. The differences between the recurrence and non-recurrence groups were compared using the Pearson Chis-squared test, Fisher's exact test or Mann-Whitney U test as appropriate. The risk factors associated with the recurrence of positive SARS-CoV-2 RNA were identified using univariate analysis, and variables with P < 0.2 were selected for multivariate logistic regression model. Missing data was not included in any of the analyses. A two-sided P < 0.05 was considered statistically significant. All statistical analyses were performed using the SPSS v21.0 software (IBM Corporation, Armonk, NY, USA), and the figures were plotted using the GraphPad Prism 8.0 software (GraphPad Software, La Jolla, CA, USA).

AUTHOR CONTRIBUTIONS
JC, XX, NP, and ZC conceived and designed the research. JC, XX, JH, FX, HL, NL, HZ, JL, JX, NP and ZC collected and analyzed data. JC, QC, NP, and ZC participated in evaluation of results. JC, QC, and NP wrote the manuscript, and all authors contributed to manuscript revision, read and approved the submitted version.