Study design and participants
A total of 43 patients were included in this study. The age distribution of patients ranges from 24 to 87, with about 80% patients age over 40 (Fig. 1A). None of the patients in this study required intensive care unit (ICU) nor a ventilator assistance, while six of them showed more severe symptoms than the others. All patients received CT imaging when they were admitted to hospital, and all the results showed ground-glass opacification or mixed ground-glass opacification and consolidation. 31 (72.09%), 13 (30.23%), and 30 (69.77%) patients had fever, dyspnea, and dry cough symptoms, respectively (Table 1). Because of the urgency and local policy, not all patients had a positive testing result of SARS-Cov-2 when they were admitted to hospital. At the time of or after admitting to hospital, 23 patients (53.49%) were tested positive for SARS-CoV-2 at least once (Fig. 1B). Pharyngeal swabs from two patients were collected on 21st, and those from all patients were collected on 23th, 25th, and 28th of February, 2020. The swabs were directly immersed in Trizol to preserve RNA during transportation and storage. We obtained specimens from 39 and 31 patients at 25th, and 28th, because 12 patients reached the standards of being healed and were discharged from the hospital during this study (Fig. 1B). The testing results did not interfere the treatment or the continuing testing for the SARS-Cov-2 appointed by CDC used as a criterion for the judgment of the healing state.
About one third of the hospitalized COVID-19 patients were never tested as SARS-CoV-2 positive during the course of this study
Real-time RT-PCR experiments were performed to detect the levels of viral RNA in each swab. To eliminate the possibility of the failure in sample collection and RNA preparation, which may influence the data interpretation, we added two controls. GAPDH was set as the internal control to monitor the total host mRNAs isolated from each swab and the effectiveness of the RT-PCR experiments. In order to directly monitor the effectiveness of the RNA preparation step, we added a fixed copy number of the in vitro transcribed EGFP RNA in Trizol prior to the RNA preparation step. According to the thresholds described in methods, 11 of 43 (25.58%), 7 (17.95%) and 9 of 31 (29.03%) hospitalized patients were tested positive at 23th, 25th, and 28th sampling points, respectively (Fig. 1B). The lower positive rate at 25th could be due to the insufficient testing performed for the samples at this sampling point. Taken together, 3 patients (6.98%) showed consistent positive testing results during the hospitalization. Fifteen patients (34.88%) were transformed from positive to negative testing results, and not returned to positive during this study, among which 6 of them (40%) reached the criteria of being healed and were discharged (Fig. 1C). On the other hand, 11 patients (25.58%) showed alternating positive/negative testing results during this study, and only one of them (9.09%) reached the criteria of being healed and was discharged. The remaining 14 patients (32.56%) never showed positive results, and 5 (35.7%) were healed and discharged (Fig. 1C).
In each set of our experiments, we included a series of the copy number dilution of the in vitro transcribed mRNA fragments from the viral ORF1ab gene, human GAPDH gene, as well as an external EGFP gene in some sets of experiments, in the reverse transcription reactions. Standard curves for calculating the correlation efficiency between Ct value and mRNA copies were therefore generated, and the mRNA copy number from each RT-qPCR positive reaction was calculated. We then selected all the samples tested positive for SARS-CoV-2 in this study, and plotted their copy numbers of viral mRNA, GAPDH mRNA, and EGFP mRNA in some benches of experiments, according to the sampling time points. Slightly higher viral loads were observed for specimens at the 23rd sampling time compared with those at 25th and 28th (Fig. 1D), consistent with the time-dependent reduction in viral loads. Meanwhile, the distribution of the mRNA copy numbers of GAPDH showed no such a time-dependent change (Fig. 1E).
The pharyngeal viral loads differed significantly and a mild-symptom COVID-19 patient showed consistently high levels
We noticed that the swab samples collected from the two patients at 21th February, 2020 were both positive, and one of them (Patient 14) showed ~ 5 orders of magnitude higher viral load than the other (Patient 6) (Fig. 2A). Because Patient 14 was mild-symptomic, while Patient 6 was severe-symptomic, this observation indicated that the pharyngeal viral loads was not correlated with the severity of the symptoms.
To further explore this observation, we examined all benches of our real-time RT-qPCR analysis. We found that among seven benches of experiments containing more than 10 patient samples and a Patient 14 sample. Patient 14 was tested positive in six benches, representing the highest positive frequency among all samples. Strikingly, in each bench of the experiment including Patient 14 sample, the detected viral loads from Patient 14 was consistently higher than all the other samples (Fig. 2B). However, the copy numbers of GAPDH and EGPF mRNAs were normal (Fig. 2C-D). The viral loads from Patient 14 pharyngeal samples remained at high levels throughout this study, until the 10th day of the hospitalization of this patient.
Finally, we followed the prognosis outcome of participated patients. Up to time we submitted the manuscript, all of the patients were recovered and met the discharge criteria. No death was found among the patients. Patient 14 was healed and discharged 24 days after being hospitalized. All patients in this study were hospitalized in two separate areas of our hospital, 2 medical personnel worked at the inpatient area of patient 14 were infected, while there was no infected medical personnel at the other area. It was unclear whether the infection was related to Patient 14.