Abstract
Background The novel coronavirus (SARS-CoV-2) infected over 3300 health-care-workers (HCWs) in early 2020 in China. Little information is known about nosocomial infections of HCWs in the initial period. We analysed data from HCWs with nosocomial infections in Wuhan Union Hospital and their family members.
Methods We collected and analysed data on exposure history, illness timelines, and epidemiologic characteristics of 25 laboratory-confirmed and two highly suspected HCWs as well as ten of their family members with COVID-19 from Jan 5 to Feb 12, 2020. Among them, demographics and clinical features of the 35 laboratory-confirmed cases were investigated and viral RNA of 12 cases was sequenced and analysed.
Results Nine clusters were found among the patients. All patients showed mild to moderate clinical manifestation and recovered without deterioration. The average periods of incubation, clinical onset serial interval (COSI), and virus shedding were 4.5 days, 5.2±3.2 days, and 18.5 days, respectively. Complete genomic sequences of 12 different coronavirus strains demonstrated that the viral structure with small, irrelevant mutations was stable in the transmission chains and showed remarkable traits of infectious traceability.
Conclusions SARS-CoV-2 can be rapidly transmitted person-to-person regardless of whether they have symptoms in both hospital settings and social activities based on the short period of incubation and COSI. The public health service should take practical measures to curb the spread, including isolation of cases, tracing close-contacts, and containment of severe epidemic areas. Besides, the HCWs should be alert during the epidemic, and make self-quarantine if self-suspected.
Abstract
SARS-CoV-2 can be rapidly transmitted in person-to-person no matter whether patients with symptoms.
Introduction
Seventeen years after the 2003 epidemic of severe acute respiratory syndrome (SARS) [1, 2], a novel coronavirus, SARS-CoV-2, was isolated from bronchoalveolar lavage of several patients with unknown origin pneumonia in Wuhan City, China [3]. SARS-CoV-2 has caused the coronavirus disease 2019 (COVID-19) across the country and abroad. A total of 83 157 COVID-19 cases with 4.02% mortality has been reported in China up to April 7, 2020 [4]. The confirmed cases have been identified in more than 180 countries and regions around the world, and more than 1 300 000 people have been infected globally outside China, including 78 932 deaths [5]. The hunt for patient zero is critical, but epidemiological investigations are often complex and unclear [6]. A COVID-19 patient usually presents with fever, a non-productive cough, myalgia, and malaise. Symptoms, including a productive cough and headache, are less common. Dyspnea is observed in more than one-fourth of patients. Cases with persistent lymphopenia usually develop fatal comorbidities, including acute respiratory distress syndrome (ARDS), acute cardiac injury, secondary infection and multiorgan dysfunction [7–9].
SARS-CoV-2 is a positive-sense and single-stranded RNA virus of zoonotic origin belonging to Betacoronavirus lineage B [3]. It has successfully crossed the animal-to-human barrier and proved to be capable of epidemic spread and may even have endemic persistence in the human population [10]. The calculated R0 of SARS-CoV-2 is as high as 2.68 [11], making prevention measures and intervention tactics very challenging. Unfortunately, current general hospital settings in Wuhan and many other cities are acting as epidemic hot spots to facilitate transmission and exacerbate spread [12]. So far, over 3300 health-care workers (HCWs) in China have been diagnosed as infected mainly through nosocomial transmission [13]. HCWs are the first to suffer during the epidemic. The transmission dynamics of COVID-19 in hospitals in the initial period, especially among HCWs, are important to deeply understand epidemiology of the diseases.
Here we investigated possible transmission links by integrating epidemiological data and whole-genome sequencing (WGS) of 25 HCWs and two highly suspected HCWs as well as their family members who got successively infected with SARS-CoV-2 in the initial period in Wuhan Union Hospital. We also collected and analysed the clinical data of 35 laboratory-confirmed cases of these individuals.
Methods
Study design and participants
A total of 35 laboratory-confirmed cases and two highly suspected cases were enrolled between January 5 to February 12, 2020, including 27 HCWs and ten relatives out of seven families. All participants were interviewed using a standard questionnaire that elevated communicable diseases. Items such as HCWs' contact history with the confirmed COVID-19 patients, including where, when, and how they had been exposed, were assessed. We mapped all cases with the precise time of symptom onset. Clusters outbreaks, potential exposures and possible patterns of transmission were estimated.
We collected clinical data of 35 laboratory-confirmed cases, including symptoms and signs, laboratory examinations, chest imaging, comorbidities and complications, and clinical treatment and outcomes, which were arranged using a standardised data collection form referencing the case record form shared by the International Severe Acute Respiratory and Emerging Infections Consortium (ISARIC) [14] and the Household transmission investigation protocol for 2019-novel coronavirus (SARS-CoV-2) infection from the World Health Organization [15].
Real-time reverse transcription polymerase chain reaction assay to detect SARS-CoV-2
The SARS-CoV-2 laboratory test assays were based on the previous WHO recommendation [16]. RNA was extracted from oropharyngeal swabs of patients suspected of having SARS-CoV-2 infection using the respiratory sample RNA isolation kit. Then real-time reverse transcription-polymerase chain reaction (RT-PCR) assay of SARS-CoV-2 RNA was conducted to amplify and test two target genes, including open reading frame1ab (ORF1ab) and nucleocapsid protein(N). Details of the manufacturer's protocol for RNA extraction, RT-PCR assay and the diagnostic criteria are illustrated in the Supplementary Appendix.
Whole-genome sequencing and comparative genome analysis
Nasopharyngeal and/or oropharyngeal swabs were obtained from all COVID-19 cases for SARS-CoV-2 nucleic acid assay, and 22 of them were processed using whole-genome sequencing (WGS). Full genomes were sequenced using the BioelectronSeq 4000 (CapitalBio Corporation, Beijing) and assembled using the de novo genome assembler, SPAdes [17], version 3.10.1, software. The complete genome was annotated by Prokka [18], version 1.14.5, software. The mutations were analysed with the use of Sibelia [19], version 3.07, software, based on existing Wuhan-Hu-1 (NC_045512.2) genome and were annotated by SnpEff [20], version 4.3, software. We integrated information from 60 published genomic sequences of SARS-CoV-2.
Full-length genomes were combined with published SARS-CoV-2 genomes and other coronaviruses and aligned using FFT-NS-2 model by MAFFT [21], version 7.455. Maximum-likelihood phylogenies were inferred under a generalized-time-reversal (GTR)+ Gamma substitution model and bootstrapped 1000 times to assess confidence using RAxML [22], version 8.2.12. The mutations of assembled amino acid sequences of Spike protein were compared with Wuhan-Hu-1 (NC_045512.2), bat-SL-CoVZC45 (MG772933.1), and SARS coronavirus isolate Tor2/FP1-10851 (JX163927.1) using Clustal Omega [23].
Epidemiological analysis
For participants with detailed onset information, a log-normal distribution was used to fit the incubation period and the probability distribution of the incubation period was estimated. For participants with detailed medical visit information, the Weibull distribution was used to fit and determine onset-to-first-medical-visit and onset-to-admission distributions on the dates of the onset of the illness, first medical visit and hospital admission. The clinical onset serial interval (COSI) was calculated, for which the probability distribution was estimated by the gamma distribution.
Statistical analysis
For clinical data, categorical variables were expressed as number (proportion). Continuous variables were expressed as median and compared by Mann-Whitney U tests. p values less than 0.05 were considered significant. Statistical analyses were conducted using SPSS software, version 23.0 (IBM Corp., Armonk, NY, USA), unless otherwise indicated.
Ethical approval
Ethical approval was waived by the institutional review board of the hospital since we collected and analysed all data from the patients according to the policy for public health outbreak investigation of emerging infectious diseases issued by the National Health Commission of the People's Republic of China.
Results
Nosocomial outbreak of novel coronavirus pneumonia and its epidemiologic transmission patterns
One male patient (69 years old) in the Department of Neurosurgery of Wuhan Union Hospital developed a fever of 38°C on Jan 6, 2020 and was finally diagnosed as COVID-19 on Jan 16, 2020. Another female patient (55 years old) had a fever on Jan 11, 2020, and was confirmed on Jan 18, 2020. During this period, the HCWs in the Department of Neurosurgery either had direct contacts with the two index patients without sufficiently efficient personal protective equipment, or they had indirect contacts through their co-workers in the department gatherings on Jan 12 and 13, 2020. Twelve of the HCWs in the Department of Neurosurgery were finally diagnosed as laboratory-confirmed COVID-19, from Jan 16 to Jan 24, 2020. Besides, two HCWs (O and Z) were diagnosed as probable case with similar clinical manifestations but negative viral nucleotide tests. Beyond the Department of Neurosurgery, there were 13 HCWs of other departments were laboratory- confirmed COVID-19 cases. These departments also had suspects enrolled at the same time.
After the 25 HCWs' onset of illness, ten out of 43 family members were confirmed as COVID-19 cases via nucleotide tests. The remaining 33 family members of HCWs were not secondary infected, because of these HCWs' taking strict self-quarantine strategies immediately after their onsets of illness, including wearing a facial mask when they came home, living alone in a separated room, never eating together with families, etc. In a word, from Jan 5 to Jan 29, 2020, 25 HCWs and ten family members out of seven families were diagnosed as COVID-19 in a single hospital. A detailed epidemiological and contact history, as well as cluster information were shown in Supplementary Appendix. The time distribution of symptom onset and speculative transmission pattern were shown in figure 1a and b, respectively.
Transmission, incubation period, and serial interval
According to the information of 14 laboratory-confirmed cases who had specific dates of exposure and symptom onset, we estimated that the average incubation period is 4.5 days (95% CI, 3.0–6.4). The 95th percentile of the probability distribution for the incubation period was 11.4 days (95% CI, 4.0–12.0) (fig. 2a). Through the nine transmission chains (fig. 1b), we estimated that the COSI distribution is 5.2±3.2 days (95% CI, 3.8–6.8) (fig. 2b). Among the data obtained from 35 confirmed cases (including the HCWs and their family members) and two suspected cases, the interval from the date of onset to the first medical visit was estimated to have a mean of 3.0 days (95% CI, 2.2–3.9) (fig. 2c). Of the 27 hospitalised cases (23 HCWs and 4 family members), the interval from the date of onset to hospital admission was estimated to have a mean of 6.6 days (95% CI, 5.3–8.2) (fig. 2d). There were no significant differences in these four indicators between HCWs and their family members.
Demographic and clinical features in 35 confirmed cases
The clinical outcomes of the patients are shown in table 1. Among 35 confirmed cases, 22 (62.9%) were women, and the median age was 37 (Range, 25–88). A few patients had underlying diseases, including hypertension (11.4%), coronary heart disease (2.9%), diabetes (5.7%), and asthma (5.7%). Up to Feb 12, 2020, a total of 21 patients (60%) had recovered and been discharged from hospital, six patients (17.1%) requiring medical observation remained in hospital, and eight patients (22.9%) who stayed home under self-quarantine had recovered. The median of virus shedding time was 18.5 days (Range, 12–25 days). The median time for the length of hospital stay was 21 days (Range, 13–28 days).
The most common signs and symptoms on admission were fever (85.7%) and malaise (74.3%), 22 (62.9%) of patients had a poor appetite, and 19 (54.3%) of the patients had a cough. Six patients received oxygen therapy because of hypoxemia. Corticosteroids were not administrated in any patients (Table S1 in the Supplementary Appendix).
The laboratory results of patients are shown in Table S2 in the Supplementary Appendix. The complete blood counts of patients on admission showed that the number of white blood cells was normal in most patients (27/31, 87.1%), 13 of the 30 patients (43.3%) had lymphopenia, alanine aminotransferase was abnormal in 4 patients (15.4%). Creatinine was normal in all patients. Seven patients (30.4%) had increased lactate dehydrogenase.
On admission, infiltrations in chest computed tomography (CT) images were detected in most the cases. In total, 29 of them (82.9%) showed ground-glass opacities (GGO), four of them (11.4%) showed consolidations, one of them (2.9%) presented mixed GGO and consolidation. Only one patient (2.9%) had normal chest imaging (Table S2 in the Supplementary Appendix). We showed two series of typical CT images of moderate pneumonia (HCW A and C, fig. 3) and a series of CT images of a suspected case (HCW O, Figure S5).
Sequencing and phylogenetic analysis
We obtained 12 whole-genome sequences (GenBank accession numbers: MT079843-MT079854) after implementing de novo sequencing from 22 swab specimens from 22 COVID-19 patients. These patients were B, C, E, F, G, H, J, M, Q, R, V, and e as shown in figure 1b, and their corresponding sample numbers were summarised in Table S3 in the Supplementary Appendix. The remaining ten samples were removed due to insufficient coverage of the extracted virus genome. Compared with Wuhan-Hu-1, except for V, the remaining 11 viral genomes all have two nucleotide mutations. We found these two mutation sites both exist in 16 (26.2%) published virus genomes (Tables S4–S6 in the Supplementary Appendix). V showed two missense mutations completely different from the other 11 cases, mainly because he has been exposed to many other patients' clinical specimens due to his daily job. The sequences from R and her husband e, are the same (fig. 4a). A total of six missense mutations were found in the 12 genomes of SARS-CoV-2 strains, but none of them were located in the genes for structural proteins in the new coronavirus, including S, E, M or N proteins (fig. 4a). Five of the mutations were located on the non-structural proteins (nsp), which are hydrolysed from the virus-encoded polyproteins 1a/1ab1. The one remaining missense mutation (28 144th T>C) was located on the cofactor gene ORF8.
The phylogenetic tree of full-length genomes showed that SARS-CoV-2 strains form a monophyletic clade with a bootstrap support of 100% (Figure S3 in the Supplementary Appendix). The most closely related sequence to this clade is bat-SL-CoV. Sequences from six HCWs (C, H, J, M, Q, R) in the Department of Neurosurgery and one family members e were closely related in the phylogenetic tree (fig. 4b). The sequence from HCW B in another department was formed a separate lineage. Compared with bat-SL-CoVZC45 (MG772933.1) and SARS coronavirus isolate Tor2/FP1-10851 (JX163927.1), SARS-CoV-2 had four and three insertion regions respectively (fig. 4c).
Discussion
Given the outbreak event of COVID-19 in Wuhan City and Hubei Province in China, we lacked efficient viral identification capacity to diagnose probable COVID-19 cases at the very early stage. There were no effective prevention measures in general hospital settings to isolate and manage the suspected COVID-19 cases, or to valid containment measures to block the transmission pathways [12]. HCWs are the highest risk occupational population in such acute respiratory infectious diseases, such as SARS and COVID-19, transmitted by contact or respiratory droplets, aerosols or fomites [24]. Here, we describe several clusters of COVID-19 cases in a hospital, by person-to-person transmission of SARS-CoV-2 to assess the epidemiological features of COVID-19.
We reported 25 laboratory-confirmed and two highly suspected HCWs, infected successively in Wuhan Union Hospital, 14 of whom were co-workers in the Department of Neurosurgery, including two clusters with two index patients as the sources, which further emphasised the importance of adequate protection for HCWs during their daily work. Moreover, seven family clusters associated with these HCWs have occurred, mainly due to the close contacts between the affected HCWs and their family members without any isolation measures. Fortunately, the majority of HCWs who isolated themselves from their family members at the onset of illness, effectively protected their family members from getting infected. Through mapping the time of onset and possible routes of transmission (fig. 1b), we can conclude that the general hospitals are just like an epidemic hub gathering a lot of patients who are the sources of infection, which would efficiently facilitate and exacerbate the transmission and spreading of virus via public transportation or alternative ways.
Through the phylogenetic analysis, we found clustered cases that were closer in the phylogenetic tree (fig. 4b). Mutation analysis showed that the sequence from each sample had different mutation sites, most of which were synonymous mutation. Even if some missense mutations existed, none of them were located in the genes for structural proteins of SARS-CoV-2 (fig. 4a). Additionally, the amino acid sequences of the structural proteins had no changes (Figure S1, Figure S2 in the Supplementary Appendix).
Most of the initial symptoms in our cases were myalgia/malaise and fever. Some patients had no fever but only mild symptoms such as nasal congestion, which may lead to patients being overlooked, causing a wider spread of COVID-19. In our cases, the intervals of onset of illness to the first medical visit and onset of illness to hospital admission were estimated to have a mean of 3.0 days and 6.6 days, respectively, which were shorter than in other studies [25]. By Feb 12, 21 of 27 hospitalised cases (77.8%) were recovered and discharged, none of them were admitted to the ICU or died. From the chest CT images from HCW A and C, very mild or moderate infiltrations were presented and became alleviated and resolved significantly in a short time (fig. 3). So, early diagnosis, isolation and timely treatments can speed up the recovery of patients and decrease the deteriorative tendency.
The median time for virus shedding to become negative in our cases was a long period of 18.5 days (Range, 12–25 days). It is worth mentioning that Nurse F was attacked by severe diarrhea during hospitalisation. Even after the nucleic acid test of nasopharynx swabs turned negative and lung CT images got better, the viral nucleic acid test of stool maintained a sustainable positive one month after onset. We must thus pay close attention to the status of virus clearance of patients with COVID-19. Such a long virus shedding period perhaps highlights the production of efficiently neutralising antibodies in a comparatively postponed manner or even an unsatisfactory titter in the patient's plasma. Therefore, it might be necessary to extend the period of hospitalisation for infectious control purposes, and we should adopt a prudent strategy for treating critical severe patient with convalescent plasma from COVID-19 patients.
Our current study has some limitations. First, the potential bias of the incubation period and COSI in our study might account for only 37 patients enrolled. The more accurate incubation period and COSI need further verification through a larger sample size. Second, the implementation of the kinetics of virus shedding and the relevant viral loading in both respiratory and intestinal tracts was not available since the novel causative pathogen has just been identified. Third, the potential for superficial exposure in infection transmission was not investigated, which may need further studies. Fourth, the kinetics of viral antibody, especially the neutralised antibody, were not monitored due to laboratory limitations. Finally, viral genome analysis and traceability from each patient have not been performed and should be conducted in future studies.
In summary, similar to the 2003 SARS outbreak in Guangzhou, the current epidemic of SARS-CoV-2 resulting in COVID-19 in Wuhan is mainly due to efficient person-to-person transmission or super-spreading events in hospital settings and social activity based on the short period of incubation and COSI. Practical strategies and measures, such as isolation of patients, tracing and quarantine of close contacts and containment of severe epidemic areas, are crucial to block the spread. Earlier detection and diagnosis of patients with COVID-19 will result in better prognoses.
Acknowledgments:
We greatly appreciate both the International Severe Acute Respiratory and Emerging Infections Consortium (ISARIC) for sharing data collection templates publicly on the website and the World Health Organization for sharing the Household transmission investigation protocol for the 2019-novel coronavirus (SARS-CoV-2) infection. We would like to dedicate this article and pay the highest tribute to every brave health care worker who is fighting, devoting himself and self-sacrificing to control the outbreak of the novel coronary pneumonia.
Footnotes
This article has supplementary material available from erj.ersjournals.com.
Support statement: This work is funded by the National Natural Science Foundation of China (No. 81500005; No. 81973990; No. 91643101; No. 81900133), the Fundamental Research Funds for the Central Universities (No.2020kfyXGYJ034) and the National Science and Technology Major Project (No.2017ZX10103004–006). National Natural Science Foundation of China; DOI: http://dx.doi.org/10.13039/501100001809; Grant: 81500005, 81900133, 81973990, 91643101; National Science and Technology Major Project; Grant: 2017ZX10103004–006; Central Universities in China; DOI: http://dx.doi.org/10.13039/501100012429; Grant: 2020kfyXGYJ034.
Author contributors: ZG and WM had the idea for and designed the study and had full access to all data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. XRW, YH, LBL, and XM contributed to drafting the manuscript. ZG, XRW, YZ, QZ, HS, and WM contributed to critical revision of the manuscript for valuable intellectual content. YH, LBL, LML, and XM conducted the statistical analysis. TD, YG, BZ, WL, XB, TP, GW, YC, and ZZ contributed to the sequencing and bioinformatics analysis. XRW, QZ, XSW, NCJ, LM, DY, JZ, BY, YX, and NJ contributed to data acquisition, data analysis, or data interpretation. The final version had been reviewed and approved by all authors.
Conflict of interest: Dr Wang has nothing to disclose.
Conflict of interest: Dr Zhou has nothing to disclose.
Conflict of interest: Dr He has nothing to disclose.
Conflict of interest: Dr Liu has nothing to disclose.
Conflict of interest: Dr Ma has nothing to disclose.
Conflict of interest: Dr Wei has nothing to disclose.
Conflict of interest: Dr Jiang has nothing to disclose.
Conflict of interest: Dr Liang has nothing to disclose.
Conflict of interest: Dr Zheng has nothing to disclose.
Conflict of interest: Dr Ma has nothing to disclose.
Conflict of interest: Dr Xu has nothing to disclose.
Conflict of interest: Dr Yang has nothing to disclose.
Conflict of interest: Dr Zhang has nothing to disclose.
Conflict of interest: Dr Yang has nothing to disclose.
Conflict of interest: Dr Jiang has nothing to disclose.
Conflict of interest: Dr Deng has nothing to disclose.
Conflict of interest: Dr Zhai has nothing to disclose.
Conflict of interest: Dr Gao has nothing to disclose.
Conflict of interest: Dr Liu has nothing to disclose.
Conflict of interest: Dr Bai has nothing to disclose.
Conflict of interest: Dr Pan has nothing to disclose.
Conflict of interest: Dr Wang has nothing to disclose.
Conflict of interest: Dr Chang has nothing to disclose.
Conflict of interest: Dr Zhang has nothing to disclose.
Conflict of interest: Dr Shi has nothing to disclose.
Conflict of interest: Dr Ma has nothing to disclose.
Conflict of interest: Dr Gao has nothing to disclose.
- Received March 4, 2020.
- Accepted April 8, 2020.
- Copyright ©ERS 2020
This article is open access and distributed under the terms of the Creative Commons Attribution Non-Commercial Licence 4.0.