Community Outbreak of Adenovirus, Taiwan, 2011

Adenovirus type 7 caused a high proportion of severe infections.

H uman adenoviruses (HAdVs) are DNA viruses that can cause a variety of human diseases. Since the discovery of these viruses in 1953, more than 50 types have been isolated, some directly linked to specifi c human diseases, e.g., infantile diarrhea (HAdV 40, 41), epidemic keratoconjunctivitis (HAdV 8, 19, 37, 54), and hemorrhagic cystitis (HAdV 11, 21) (1-3). Adenoviruses also are common causes of lower respiratory tract infections in children (4), but surveillance for these infections is lacking in most countries. Pneumonia caused by adenoviruses cannot be easily differentiated from other types of viral pneumonia, and culturing and typing of adenoviruses is not routinely performed in hospital or public health laboratories. Therefore, community-wide outbreaks of adenovirus are not easily detected; previous reports are limited to those occurring in hospital, school, or military settings (5)(6)(7).
In 1999, after a 1998 enterovirus 71 epidemic, the Taiwan Centers for Disease Control established a nationwide surveillance system using contract virologic laboratories (CVLs) to perform continuous virologic surveillance for respiratory viruses, especially infl uenza and enteroviruses (8). The network consists of 12 CVLs located in the northern, central, southern, and eastern regions of Taiwan (9). Early in 2011, the percentage of adenovirus isolated among all respiratory virus isolates evaluated by the CVLs increased from a baseline of 0%-5% to 10% and remained high in the following weeks, indicating a community-wide adenovirus outbreak. The apparent outbreak prompted us to use several existing surveillance systems to describe the characteristics of this outbreak.

Virologic Surveillance of Outpatients
For respiratory virus surveillance, throat/nasal swabs from outpatients with infl uenza-like illness (ILI) are collected by sentinel physicians and cultured in the CVL of the corresponding region on a weekly basis. Virus isolates found are then sent to the reference laboratory at the Taiwan Centers for Disease Control for identifi cation, sequencing, and typing. For this study, we reviewed virologic surveillance data from week 1 of 2008 through week 43 of 2011 (January 1, 2008-October 30, 2011). The weekly adenovirus-positive rate is defi ned as the number of adenoviruses isolated from respiratory tract specimens divided by the number of all specimens submitted from patients with ILI for respiratory virus surveillance in the corresponding week. The mean positive rate in 2008-2010 was used as the baseline adenovirus-positive rate, and the epidemic threshold was defi ned as the baseline adenoviruspositive rate + 2 SD. The start and end of the epidemic were defi ned accordingly.

Clinical Case Surveillance for Inpatients
We conducted a retrospective study of adenovirusinfected children treated as inpatients in the National Taiwan University Hospital (NTUH), a tertiary hospital in northern Taiwan. We enrolled all patients <18 years of age admitted to the NTUH pediatric department from November 1, 2010 (week 44, 2010), through June 30, 2011June 30, (week 26, 2011, who had adenovirus infection identifi ed by virus isolation or PCR from respiratory specimens. All virus isolations and PCRs were performed in the NTUH laboratory, 1 of the 12 CVLs.
All medical records of enrolled inpatients were reviewed. Demographics, medical history, clinical signs and symptoms, diagnoses, and treatments were recorded by using a structured questionnaire. Patients who had been admitted to the intensive care unit were classifi ed as having severe infection; all other patients were classifi ed as having nonsevere infection.

Virus Culture, Identifi cation, and Typing
We typed selected adenovirus isolates collected as part of the virologic surveillance program and all isolates obtained from inpatients from NTUH. We also performed real-time PCR for viral load quantifi cation by using LightCycler (Roche Diagnostics, Mannheim, Germany) for selected patients from the NTUH, according to the method described by Heim et al. (10).
Throat/nasal swab specimens from all patients were injected into human embryonic lung fi broblasts, Hep-2, RD, MK-2, and MDCK cells. If a cytopathic effect was observed, the presence of adenovirus was further confi rmed by direct immunofl uorescence staining with a virus-specifi c monoclonal antibody. DNA was extracted from the clinical samples by using the QIAamp Blood Mini Kit (QIAGEN, Hilden, Germany); PCR was then conducted targeting a 956-bp region of the hexon gene for typing.
For genetic analyses of HAdV-7 isolates, DNA fragments of the hexon and fi ber genes were amplifi ed by PCR. Multiple sequence alignments, protein translation, and phylogenetic analysis were performed on the basis of the nucleotide sequences by using MEGA4 (11) and BioEdit software (www.mbio.ncsu.edu/BioEdit/bioedit.html). For phylogenetic analyses, we included the full-length sequences of the hexon and fi ber genes (2,805 bp and 978 bp, respectively) from 5 HAdV-7 isolates collected in 2011 in Taiwan and some reference sequences available in the National Center for Biotechnology Information database (www.ncbi.nlm.nih.gov/genbank). A phylogenetic tree was constructed by the neighbor-joining method, and 1,000 bootstrap replications were performed to evaluate the reliabilities of the relationships.

Statistical Analysis
Data from inpatients with versus without severe infection were compared by using the χ 2 or Fisher exact test for categorical variables and the Mann-Whitney U test for continuous variables. A p value <0.05 was considered signifi cant. All statistical operations were 2-tailed and were performed with SPSS version 19.0 software (IBM, Somers, NY, USA). Ninety-seven percent of all specimens were collected from patients <18 years of age.

Virologic Surveillance for Outpatients
We typed 883 adenovirus isolates collected during the study period from outpatients <18 years of age; temporal distribution is shown in Figure 2. HAdV-3 was the dominant circulating type during 2008-2009 (>50% of all isolates), but in 2010, the proportion of HAdV-3 decreased to ≈30%, similar to that for HAdV-2 and HAdV-1. The proportion of HAdV-5 was consistently ≈5% in 2008-2010, and all other types of adenoviruses were rarely isolated. Among the 844 adenovirus isolates collected during 2008-2010 that were typed, only 3 (0.3%) were HAdV-7; during this outbreak, the proportion of HAdV-7 increased signifi cantly, to 10% (p<0.001). The proportion of HAdV-3 also increased, from 30% to 74%, while the proportions of all other types decreased signifi cantly (p<0.001).

Demographics and Associated Conditions
Demographic data on inpatients is shown in Table 1. Median age was 40 months (range 1-189 months); the male:female ratio was 1.3:1. Most (59%) patients had a history of contact before illness onset with patients with upper respiratory symptoms in school or in the household.
Patients with severe infection had longer hospital stays than did those with nonsevere infection (median 18 vs. 5 days; p<0.001). Thirty-eight percent of all patients had chronic underlying conditions before admission, and more patients with severe infection had underlying diseases, particularly cardiopulmonary and neurologic diseases, than did those with nonsevere infection (77% vs. 30%; p<0.001).
Parenteral or oral antimicrobial drugs were given to 73.8% of patients before adenovirus infection was diagnosed. Ten percent of patients required mechanical ventilation during hospitalization; all were in the severe group. Intravenous immunoglobulin (IVIG) was given to 6 patients (3.0%) for diagnoses of acute myocarditis (n = 1), empyema (n = 1), acute respiratory distress syndrome (n = 3), and Kawasaki disease (n = 1). Extracorporeal membrane oxygenation was used for 5 patients with respiratory or cardiovascular failure; 3 survived.
Seven patients died during this outbreak: 4 patients infected with HAdV-7, 2 with HAdV-3, and 1 with HAdV-2. All of these patients were in the severe group and had underlying diseases; 6 were bedridden before hospital admission. During hospitalization, all of these patients had secondary bacterial pneumonia develop; pathogens involved were Pseudomonas aeruginosa, Staphylococcus aureus, Acinetobacter baumannii, and Escherichia coli.

Molecular Studies on Adenovirus Type 7
The hexon gene sequences of HAdV-7 isolates from virologic surveillance and clinical case surveillance, including patients with severe and nonsevere infection, were determined and compared with strains isolated previously in Taiwan and other countries. Hexon nucleotide sequences of 5 HAdV-7 isolates collected in 2011 in Taiwan were identical to the sequence of strain HAdV7-HZ/SHX/CHN/2009 (6,12), differed by 1 bp of synonymous mutation from genotype 7d of the 383 strain isolated in Japan in 1992, and had 99.7% identity to the sequence of the prototype strain S1058 (Figure 3, panel A). Fiber sequences of these isolates were identical to those of the 383 and Bal strains from Japan, isolated in 1992 and 1995, respectively, and had 99.5% identity to the prototype stain S1058 (Figure 3, panel B). We found no substantial differences between the HAdV-7 isolates collected from outpatients as part of the virologic surveillance program and the isolates from inpatients with severe or nonsevere infection.
The outbreak of co-circulating HAdV-3 and HAdV-7 we report is unique in several ways. The high positive rate during the weeks of the epidemic made this outbreak among the largest reported community-wide adenovirus outbreaks, and the reemergence of HAdV-7 in Taiwan  Table 3. Laboratory test results for 202 children hospitalized for adenovirus infection at National Taiwan University Hospital, by  disease  10 years after the last outbreak contributed to severe infection. HAdV-3 and HAdV-7 are both capable of causing outbreaks but have different circulation patterns. HAdV-3 circulates endemically and causes outbreaks, whereas HAdV-7 is mainly detected during outbreaks (15,16). DNA restriction analysis has been used to further characterize adenoviruses, and shifts in or replacement of the predominant genome types over time have been reported to be related to outbreaks (17,18). For a novel genome type of HAdV-7 to spread rapidly and cause a large outbreak, an immunologically naive population and a means of introduction to the susceptible community are needed, and the novel virus must have a greater biologic fi tness than that of other circulating strains (18). During the 1999 outbreak in Taiwan, 40% of adenovirus isolates were HAdV-7. The proportion of HAdV-7 among all adenoviruses decreased to 20% in 2000, <5% in 2001, and <1% in 2004-2005 and 2008-2010 (13,14). The near-absence of circulating HAdV-7 in 2001-2010 indicates that a large population was naive to this virus, predisposing that population to the 2011 outbreak.
Genome typing showed that HAdV-7a was the circulating strain in Taiwan in 1983 but was replaced by HAdV-7b by 1999-2001 (13). However, phylogenetic analysis of the entire hexon gene showed that the strain from this outbreak has the highest homology with HAdV-7d and HAdV-7d2, which had not reported in Taiwan (6). HAdV-7d has been the predominant circulating virus in China since the early 1980s and caused a large outbreak in South Korea during 1995-1997 (16). The closely related genotype HAdV-7d2 was reported in Israel in 1992 and has caused outbreaks in chronic care facilities and in communities in several countries (6,(19)(20)(21). Despite the early introduction of HAdV-7d to neighboring countries, HAdV-7d had not been found in Taiwan until 2011 (13).
We sequenced the entire fi ber gene, the product of which contributes to virus attachment and infection, to further characterize the circulating strain (22). Phylogenetic analysis (Figure 3) showed no substantial genome difference between this HAdV-7d strain and other previously reported HAdV-7d strains. We do not know the means of introduction of this strain into Taiwan, but we believe that an emerging strain and the large, naive population are 2 factors that contributed to this outbreak.
HAdV-7 has been reported to have a strong association with severe illness (23). Studies of outbreaks caused by the co-circulation of HAdV-3 and HAdV-7 offer a unique opportunity for us to compare the disease severity resulting from infection with these 2 virus types, but confl icting results have been obtained (7,13,24). During this outbreak, although HAdV-3 was the predominant strain, patients infected with HAdV-7 exhibited more severe disease. HAdV-3 accounted for 74%, 62%, and 41% of all isolates in outpatients, inpatients with nonsevere infection, and inpatients with severe infection, respectively, compared with 10%, 12%, and 41%, respectively, for HAdV-7 (p<0.01). Our study offers strong evidence that HAdV-7 is signifi cantly associated with severe infection in a community outbreak setting.
Of the 202 inpatients we analyzed, 31 (15%) had severe infection that required intensive care. As previously reported, children with underlying conditions were more likely to have severe disease develop, especially patients with neurologic, respiratory, and metabolic abnormalities (23,(25)(26)(27). More than half of inpatients had a history of household or school contact with persons with upper respiratory symptoms before admission, which is an evidence of the high transmissibility of the virus and the widespread nature of the current outbreak. Inpatients with severe infection had higher fever, longer fever duration, and more evident lower respiratory tract involvement than those with nonsevere infection. Viremia can be found in up to 72% of immunocompetent children during fi rst infection with HAdV (28); our laboratory data revealed that adenovirus-infected patients had more profound systemic involvement than was refl ected by clinical signs and symptoms, which may have resulted from viremia. Although renal function remained relatively good, other manifestations (e.g., anemia, leukopenia, thrombocytopenia, elevated serum transaminase and lactate dehydrogenase levels) were indicators of multiorgan involvement. Eighty percent of patients with severe infection had hyponatremia, which suggests a syndrome involving inappropriate antidiuretic hormone (29). Viral load in respiratory tract specimens did not differ for patients with severe infection versus nonsevere infection, which suggests host factors may be contributors to severe illness.
Compared with that for other reported outbreaks, the mortality rate we found was relatively low (3.5% for all hospitalized patients) (5). Good supportive/ intensive care and advanced life support, including the use of extracorporeal membrane oxygenation, are possible explanations. With no proven treatment, various medications, including ribavirin (30), cidofovir (31) and vidarabine (32), and IVIG (33), have been used to treat immunocompromised patients who have severe adenovirus infection; results have been inconsistent. Six patients in our study received IVIG and 4 of them died, possibly because of severe underlying conditions. We are not able to draw any conclusions regarding the treatment effi cacy of IVIG from this small case series; further study is needed.
In conclusion, by using data from public health and hospital-based surveillance programs, we described a large community outbreak caused by circulating HAdV-3 and emerging HAdV-7. We confi rmed the severity of HAdV-7 infection and illustrated the epidemic nature of its circulation. Public health surveillance systems should continue to monitor the molecular epidemiology of adenoviruses to detect outbreaks early.