A cluster of adenovirus type B55 infection in a neurosurgical inpatient department of a general hospital in Guangdong, China.

Background Human adenovirus type 55 is a re‐emerging human respiratory pathogen that is associated with several respiratory infections outbreaks in military and school populations. In this study, we describe the first HAdV55‐associated hospital outbreak documented in Guangdong, China. Methods Active surveillance was conducted in the involved neurosurgical inpatient department. All staff and patients in the involved neurosurgical department were surveyed for any symptoms of fever (≥38°C) and enlarged tonsils during the outbreak period. Throat swabs and demographic information were collected for all cases. For each specimen, assays for common respiratory viruses were performed using one‐step reverse transcription‐polymerase chain reaction. HAdV‐positive samples were inoculated onto Hep‐2 cells for isolation. Hexon genes, fiber genes, penton genes, and whole genomes were sequenced. A phylogenetic tree was constructed. Results and Conclusions Forty‐three cases, including 24 laboratory‐confirmed cases and 19 possible cases, were identified. Nurses had the highest attack rate of infection, with a rate of 36.4%. The attack rate for doctors and inpatients was 20.0% and 16.7%, respectively. HAdV55 was the sole pathogen identified during this outbreak. The hexon, fiber, and penton genes from seven isolated HAdV55 stains were sequenced. All these genes showed 100% homology and fell into the HAdV55 [P14H11F14] cluster, indicating that HAdV55 was the single viral strain for the outbreak. While not conclusive, the epidemic investigation revealed that the outbreak was introduced by nurses from sources outside the hospital. It was likely that a transmission from staff to inpatients had occurred.


| INTRODUCTION
Human adenovirus (HAdV) is a common pathogen among children and adults. Infections of HAdV can cause a variety of clinical diseases, ranging from asymptomatic and self-limited conditions to pneumonia and even death. 1 As of now, at least 68 HAdV genotypes have been identified and are classified into seven species (A-G). 2 Furthermore, viruses in species B, C, and E are more commonly associated with symptomatic respiratory infections. 1,[3][4][5][6][7] Human adenovirus type 55 (HAdV55), a member of HAdV-B family, was first identified from a military outbreak in Spain in 1969. 8 This virus used to be recognized as HAdV-B11a by partial characterization of its hexon and fiber epitopes. [9][10][11] In 2006, Michael et al. revealed that HAdV55 was an emergent respiratory pathogen. 12 It had evolved from a homologous recombination between HAdV-B14 and HAdV-B11 in its hexon gene, which conferred changes in viral serotype and immune activity. In recent decades, HAdV55 has been associated with several respiratory infections outbreaks. 9,11,13,14 Crowed communities, military training camps, and schools are common settings for HAdV infections.
Here, we describe a HAdV55 outbreak occurred in a neurosurgical inpatient department of a Jiangmen hospital in Guangdong Province, China.

| Outbreak settings and investigation
On June 30, 2016, one local Center for Disease Control and Prevention of Guangdong was informed that several staff members, including nurses and doctors working in a general hospital, fell ill with symptoms of a fever and sore throat. The hospital is a district general hospital. The By reviewing patients' medical records, we identified the first case and determined the investigation period began on June 11, 10 days before the onset of the first case. All staff and patients in this neurosurgical department were surveyed for any symptoms of fever (≥38°C) and enlarged tonsils during the outbreak period. Laboratory-confirmed cases were defined as persons with positive HAdV55 PCR assays. Those not confirmed by the laboratory, but with symptoms and signs (fever (≥38°C) and enlarged tonsils), were defined as clinical-confirmed cases (possible cases). Throat swabs from all cases were collected and tested.
Information on demographic characteristics, onset of illness, and clinical symptoms were collected. For the use of all above-mentioned specimens, written informed consent from all participants involved in the research was obtained (or from their parents or legal guardians in the case of a minor). This study was approved by the Ethics Committee of the Guangdong Provincial Center for Disease Control and Prevention, and was in compliance with the Helsinki Declaration.

| Gene sequencing
The hexon, fiber, and penton genes of HAdV were amplified. The primers we used are listed as follows: Hexon-F

| Phylogenetic tree
Multiple sequence alignments were performed with ClustalW, 15 and alignments were manually edited with Aliview. 16 For whole genome sequences, MAFFT software was used for alignment with default parameters (http://mafft.cbrc.jp/alignment/server/). Maximum likelihood (ML) trees were estimated in RaxML 17 using the generalized time-reversible (GTR) nucleotide substitution model with gamma distribution among site rate heterogeneity. 18

| Statistical analysis
Data analysis was performed with spss 15.0 (SPSS Inc., Chicago, IL, USA). Differences with an error probability of P<.05 were regarded as significant. For categorical data, we used chi-square testing and Fisher Exact testing as appropriate.

| Outbreak epidemiology
The outbreak ranged from June 21 to July 8, 2016. The first cases were two nurses working at Neurosurgical Unit A. Our investigation identified 43 cases, including 24 laboratory-confirmed cases and 19 possible cases. The onset of illness in all cases is shown in Figure 1.
Two peaks were observed. The first one was from June 21-28. It included 25 cases, among which 21 were nurses. The second was from June 29 to July 5. During this period, 13 of 15 infected inpatients were identified. It was observed that, within each unit, the first cases were all nurses.  (Table 3). Most cases (41 cases) reported upper respiratory symptoms. Fever was the most common symptom, followed by sore throat, cough, expectoration, headache, and runny nose. Twelve of these 43 cases developed pneumonia. No mechanical ventilation was needed. Two cases developed conjunctivitis. Most clinical symptoms and signs between infected inpatients and infected medical workers did not differ (Table 3). However, there were significantly more inpatients who developed pneumonia (P<.001).

| Outbreak investigation and infection control measures
We reviewed all inpatients' medical records and surveyed their visitors. Neither the inpatients nor their visitors had respiratory symptoms before the outbreak. We also interviewed all the chief doctors and nurses working in these three units and enquired about the staff's routine practice in the involved neurosurgical department. There are separated doctor offices, nurse stations, and nurse resting rooms within each unit. The staff had basically followed the standard infection control procedures when they cared for the patients. In the nurse stations and the resting rooms, the nurses preferred not to wear any protective equipment. They shared most office supplies, such as computers, medical records, and examination records. More often, it was the nurses who visited the doctors' offices for daily work communication, while the doctors rarely visited the nurse stations and their resting rooms, except for the acupuncturist, who was responsible for the entire department and was used to drinking water in nurse station A.
There was a dispensing counter in nurse station A from which nurses from unit B and the NSICU sometimes pick up medicines. Other interactions between units were fewer. One day before the outbreak, nurses from unit A had a party in a karaoke bar.
Control measures were implemented from June 23, when the hospital infection control committee was informed of this possible nosocomial infection. Environmental cleaning was enhanced. Self-protections, including hand hygiene and surgical mask wearing, were re-enforced.
All cases and their close contacts were followed up. On June 27, due to the increasing number of cases, further control measures were implemented. Staff with symptoms were either given medical leave or treated in isolation, as what had been implemented for the infected inpatients.
The personnel entering and leaving the involved neurosurgical department were restricted. Visitors to the department were required to wear surgical masks. In addition, screenings for similar symptoms were performed in the entire hospital. Specimens from cases and the environment were collected and sent to the local CDC for pathogen detection.

| Phylogenetic analysis
The fiber gene (1048 nucleotides (nt)), hexon gene (1603 nt), and penton gene (1206 nt) were obtained from eight HAdV strains isolated in the outbreak. Sequences of each gene were completely identical, suggesting the outbreak was caused by the single viral introduction.
Phylogenetic trees of fiber, hexon, and penton genes were constructed by integrating all closely related adenovirus sequences from

| DISCUSSION
This study identified a cluster of adenovirus type B55 infection in the neurosurgical inpatient department of a general hospital. Through targeted surveys, the possibility of inpatients as the source of this outbreak was excluded. We suggested that the pathogen of this outbreak may have been introduced into the hospital by nurses in unit A. They had a party in a karaoke bar 1 day before the outbreak. One or some  Our study has limitations. First, several respiratory viruses could cause respiratory infections. 25 Previously published data show that it is difficult to make a reliable distinction among these pathogens based on clinical signs and symptoms. 26 To ensure case findings, we used a sensitive definition of possible cases in this outbreak, which may include both adenovirus and non-adenovirus infections, as well as infected patients who may not have been involved in this infection cycle.
Our investigation found 27 throat swab specimens with a HAdVpositive status, while pathogens for the remaining cases were not fully clarified. Therefore, the attack rate may have been overestimated.
More laboratory tests should be carried out to understand pathogenic characteristics. Second, based on our investigation, we implicated that the virus could be intruded from outside of the hospital by the nurse or