Norovirus Genotypes in Hospital Settings: Differences Between Nosocomial and Community-Acquired Infections

Transcription

1 MAJOR ARTICLE Norovirus Genotypes in Hospital Settings: Differences Between Nosocomial and Community-Acquired Infections Kristina Træholt Franck, 1,2 Rikke Thoft Nielsen, 3,a Barbara Juliane Holzknecht, 1,b Annette Kjær Ersbøll, 4 Thea Kølsen Fischer, 1 and Blenda Böttiger 1,5 1 Microbiological Diagnostics and Virology, Statens Serum Institut, Copenhagen, 2 Research Unit for Clinical Microbiology, University of Southern Denmark, Odense, 3 Department of Clinical Microbiology, Copenhagen University Hospital Herlev, and 4 National Institute of Public Health, University of Southern Denmark, Copenhagen, Denmark; and 5 Department of Laboratory Medicine Malmö, Lund University, Sweden Background. Norovirus (NoV) is a major cause of gastroenteritis and hospital outbreaks, leading to substantial morbidity and direct healthcare expenses as well as indirect societal costs. The aim of the study was to estimate the proportion of nosocomial NoV infections among inpatients testing positive for NoV in Denmark, , and to study the distribution of NoV genotypes among inpatients with nosocomial and community-acquired NoV infections, respectively. Methods. Admission and stool sampling dates from 3656 NoV-infected patients were used to estimate the proportion of nosocomial infections. The associations between nosocomial infection and patient age, sex, and NoV genotype GII.4 were examined. Results. Of the 3656 inpatients, 63% were classified as having nosocomial infections. Among these, 9 capsid and 8 polymerase NoV genotypes were detected, whereas in the smaller group of inpatients with community-acquired infections, 12 capsid and 9 polymerase genotypes were detected. Nosocomial NoV infections were associated with age 60 years and infections with genotype GII.4. Conclusions. The majority of NoV infections in hospitalized patients were nosocomial. Nosocomial infection was mainly associated with older age but also with the specific genotype GII.4. The genotypes in communityacquired NoV infections were more heterogeneous than in nosocomial infections. Keywords. Caliciviridae; norovirus; healthcare; hospitals; genotypes; community-acquired; nosocomial; genetic diversity; gastroenteritis; healthcare-associated infection. Norovirus (NoV) is a global cause of gastroenteritis [1, 2] and the most frequent pathogen in hospital outbreaks [3]. The median incubation time for NoV genogroup II is estimated to be 1.2 days, with 95% of patients having Received 28 November 2014; accepted 11 February 2015; electronically published 20 February a Present affiliation: Department of Clinical Microbiology, Odense University Hospital, Denmark. b Present affiliation: Department of Clinical Microbiology, Copenhagen University Hospital Herlev, Denmark. Presented in part: 17th Annual Meeting of the European Society for Clinical Virology, Prague, Czech Republic, 28 September to 1 October Abstract P Correspondence: Kristina Træholt Franck, MD, Statens Serum Institut, Artillerivej 5, DK-2300 Copenhagen, Denmark (kristina@lundby.net). The Journal of Infectious Diseases 2015;212:881 8 The Author Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please journals.permissions@oup.com. DOI: /infdis/jiv105 symptoms 2.5 days after acquiring the infection [4]. NoV infections are often self-limited and of short duration [5], but hospitalization may be necessary, especially for elderly patients [6]. Elderly patients and those with comorbid conditions are at risk of a more severe course of NoV infection with complications, such as dehydration and renal failure [7, 8]. Chronic infection and fatal cases may occur [9, 10]. NoV is transmitted via person-to-person contact or indirectly, for example via contaminated surfaces [11, 12]. Less than 400 particles can cause infection, and billions of NoV particles can be shed in stool and vomit during infection [13]. NoV remains infectious in a dry environment for at least 5 days and for months in water [11, 14]. It is not easily inactivated by alcohol in hand disinfectants [15]. Thus, NoV is highly infectious. Hospital outbreaks are sought controlled by methods Nosocomial Norovirus Infections JID 2015:212 (15 September) 881

2 such as increased hygiene and isolation of infected patients [16]. In England in , a total of 1183 NoV hospital outbreaks were reported, leading to ward closures in 81% of the outbreaks and affecting patients [17]. NoV outbreaks have been estimated to cost a mean of $6237 per patient, depending on the outbreak interventions and their efficacy; ward closures were especially costly [16]. NoV is a single-stranded, nonenveloped RNA virus. It can be divided into several genogroups and genotypes [18]. Genogroup II genotype 4 (GII.4) is the most prevalent genotype globally and more predominant in hospitals than in community settings [1, 9, 19, 20]. This seems to be partly caused by an increased susceptibility toward NoV GII.4 among older patients, who also constitute the majority of hospitalized patients [19, 20]. OthersuggestedreasonsforthepredominanceofNoVGII.4 in hospital settings are increased virulence or increased virus shedding facilitating transmission within a semiclosed hospital environment [20 22]. The purpose of this study was to estimate the proportion of nosocomial NoV infections in Denmark and to describe the NoV genotype distribution among inpatients with nosocomial and community-acquired infections, respectively. METHODS Setting and Patients The laboratory information system of Statens Serum Institut, Copenhagen, Denmark, was used to identify NoV-positive samples in the 9-year period from 2002 to Denmark has a public healthcare system and hospitalization is free of charge. Patients can be hospitalized after referral from general practitioners or emergency departments. Statens Serum Institut served as the primary laboratory for virus diagnostics covering the majority of Danish hospitals throughout the study period. Inpatients with a laboratory-confirmed NoV infection were selected. Only the first positive sample from each patient was used for the study (n = 4399). Samples from 2006 to 2010 have been partly presented elsewhere in a study of NoV genotypes in community and hospital settings [19]. Information on age, sex, and NoV genotype was obtained from the laboratory registration system. Information on admission dates and wards was obtained from the Danish National Patient Register. A total of 743 patients were excluded; 275 patients whose hospitalization status was unknown and 468 for whom stool sampling dates were unknown. Thus, 3656 inpatients were included in the study. Laboratory Methods From 2002 to 2005, samples were tested for NoV with reversetranscription polymerase chain reaction (PCR) and detected on an agarose gel, as described by Vinjé et al [23]. Twelve samples tested positive for NoV with electron microscopy only. From 2006, samples were tested with real-time reverse-transcription PCR, as described elsewhere [19]. During the study period, samples were continuously genotyped as part of the routine monitoring of NoV genotypes circulating in Denmark. The intention was to genotype 1 sample from each ward per month. As it became evident that only few samples from inpatients with community-acquired infections had been genotyped using this routine, supplementary sequencing was performed on samples from this group. Thus, the genotyping selection criteria were (1) 1 sample from each ward per calendar month (if NoV-positive samples were available) and (2) 1 sample from all inpatients with community-acquired NoV infection. Only 1 sample from each patient was included for analyses. Genotyping was performed as described elsewhere [19], by partly sequencing of the polymerase gene (n = 1038), partly sequencing of the capsid gene (n = 96), or both (n = 96). Genotypes were assigned using a public genotyping tool ( norovirus/typingtool) [24]. The capsid genotypes were primarily used for data analyses; if they were not available, the polymerase genotypes were used. Assignment of Nosocomial Versus Community-Acquired Infection NoV infections were classified as community acquired if stools were sampled on the day of admission (day 0) or the following day (day 1) and nosocomial if sampling was performed on day 5 or later. For samples obtained on days 2 4 after admission, infections were classified as indeterminate. This conservative definition of nosocomial NoV infection was used to compensate for possible delays between onset of symptoms and stool sampling and to follow the definitions used by others [25, 26]. Sensitivity analyses were performed in which nosocomial NoV infections were defined as sampling performed (1) on day 3 or later after admission or (2) on day 6 or later after admission. The definition of community-acquired NoV infections was not changed in the sensitivity analyses. Statistical Analysis Descriptive analyses were performed by means of frequency distributions (number and proportion). The association between nosocomial NoV infection (yes or no) and age (<18, 18 59, or 60 years), genotype (GII.4 or non-gii.4), sex (male or female), and geographic location (Zealand or the rest of Denmark) was examined using multivariable logistic regression analyses. Only patients with an assigned genotype and with nosocomial or communityacquired NoV infections were included for these analyses. Nosocomial NoV infection (yes or no) was the binary outcome variable. In the statistical analyses, the genotypes GII.P4 and GII.4 were considered as a single group and have been referred to as GII.4. The hierarchical data structure with hospitals and inpatients nested within hospitals was taken into account by including hospitalsasarandomeffectintheanalyses.age,genotype, sex, and region were included as fixed effects. The interactions 882 JID 2015:212 (15 September) Franck et al

3 between genotype and each of the 3 fixed effects were evaluated. Nonsignificant interactions were removed using backward elimination, removing the most insignificant interaction first. The significance of effects was evaluated using the likelihood ratio test. The intraclass correlation coefficient was calculated for the resulting multivariable multilevel logistic regression model. Univariate analyses, described in Results, were performed using the Wilcoxon rank sum test. Statistical analyses were performed using SAS (version 9.3; Statistical Analysis System) and Stata (version 11.2; StataCorp) software. Significance was determined at P <.05. Clinical Information Information regarding onset of gastroenteritis was available for 73 inpatients from 7 wards in 1 hospital during This information was used to evaluate the validity of the results obtained from the analyses described above. Ethical Considerations The study was approved by the Danish Data Protection Agency (record ) and the Danish Health and Medicines Authority (record /1/). The clinical information was obtained as part of another study [27]. Permission to use these data was obtained from the Danish Data protection Agency (record ) and the Danish Health and Medicines Authority (record /1/). RESULTS Study Population A total of 3656 inpatients tested positive for NoV during These patients were hospitalized in 297 different wards in 71 hospitals, situated in all 5 administrative regions of Denmark. Samples from a median of 31 patients (range, 1 242) were obtained from each hospital, evenly distributed across the Western Table 1. and Eastern parts of Denmark (n = 1869 and n = 1787, respectively). According to our definition, 2320 (63%) of the NoVpositive inpatients had nosocomial NoV infections, and 572 (16%) had community-acquired infections; the NoV acquisition source was classified as indeterminate for 764 (21%) (Table 1). When we included only samples from hospitals represented by 20 patients each (n = 42), 64% of the patients (range, 37% 87%) had nosocomial NoV infections. The median ages of inpatients with community-acquired and nosocomial NoV infections were 71 (range, 0 98) and 79 (range, 0 102) years, respectively. Nosocomially infected inpatients were significantly older than inpatients with communityacquired infections (P <.001). The highest proportion of nosocomial infections (67%) was seen among inpatients 60 years of age, but nosocomial infections were also present among 28% of children and adolescents (age, <18 years). Patients with nosocomial NoV infections were hospitalized for a median of 21 days (interquartile range [IQR], days). Inpatients with community-acquired NoV infections were hospitalized for significantly fewer days (median, 3 days; IQR, 2 6 days) (P <.001). Patients 60 years of age were hospitalized for a median of 16 days (IQR, 9 29 days), whereas younger patients <60 years of age were hospitalized for a significantly shorter duration (median, 8 days; IQR, 3 18 days) (P <.001). Women constituted 57% (n = 1845) and 46% (n = 204) of patients hospitalized 5 days and <5 days, respectively. The sex distribution also differed between the age groups; women constituted 46% (n = 254) of patients aged <60 years, 40% (n = 203) of those aged years, and 61% (n = 1592) of those aged 70 years, respectively. Genotyping Results Genotyping results were available from 1230 patients from 240 wards in 64 hospitals, from a total of 3656 NoV-positive inpatients. A median of 1 sample (75th percentile, also 1 sample) Demographic Characteristics and Association Between Nosocomial Infection and Age or Sex for 3656 Danish Inpatients Patients, No. (%) b Characteristic Total Nosocomial Infection Indeterminate Infection Community-Acquired Infection Age, y <18 74 (2) 21 (28) 10 (14) 43 (58) (13) 211 (44) 95 (20) 169 (36) (85) 2088 (67) 659 (21) 360 (12) Sex Male 1607 (44) 963 (60) 357 (22) 287 (18) Female 2049 (56) 1357 (66) 407 (20) 285 (14) Overall 3656 (100) 2320 (63) 764 (21) 572 (16) a Analyses are based on patients with community-acquired (sampling 0 1 day after hospital admission) or nosocomial (sampling 5 days after hospital admission) norovirus infections. b Percentage within the age group, sex, or overall. Nosocomial Norovirus Infections JID 2015:212 (15 September) 883

4 was genotyped from each ward per calendar month. According to the selection criteria, samples from 1791 inpatients should be genotyped, but a genotype was established for only 1002 (57%) of these. In 249 samples (14%), genotyping could not be performed owing to lack of material, negative typing PCR results, or inconclusive sequencing results, and in another 540 (30%), genotyping was not performed owing to lack of laboratory resources. Another 228 samples had been genotyped as part of the routine surveillance of circulating NoV genotypes; these were also included in the analyses. A larger proportion of inpatients <18 years of age (59%; n = 44) or classified as having community-acquired NoV infections (58%; n = 333) had an assigned genotype compared with inpatients 60 years of age (32%; n = 983) or nosocomially infected (29%; n = 681). Samples from 96 patients were genotyped by partly sequencing both the capsid and polymerase regions. GII.4 was detected as either the capsid or the polymerase genotype in 63 of the samples, and in 90% (n = 57) of these patients, both the capsid and polymerase sequences were GII.4. Five samples were genotyped as GII.Pe_II.4, and 1 sample as GII.P21_II.4. From 333 inpatients with community-acquired NoV infections, 12 capsid and 9 polymerase genotypes were detected, whereas only 9 capsid and 8 polymerase genotypes were detected from 681 inpatients classified as nosocomially infected (stool sampling 5 days after hospital admission) (Figure 1). Using the alternative definitions of nosocomial NoV infections (ie, stool sampling 3 days [n = 837] or 6 days [n = 596] after hospital admission), 10 capsid and 10 polymerase genotypes and 8 capsid and 7 polymerase genotypes were detected, respectively. The associations between nosocomial NoV infections and age or genotype are presented in Table 2. Nosocomial NoV infection was significantly associated with both age 60 years and infection with the specific genotype GII.4. The latter association seemed weaker than the former, but the difference was not significant. Owing to a limited number of children (age <18 years) and no significant difference in the proportion of nosocomial infections between inpatients aged <18 years and those aged years, inpatients aged <60 years were aggregated into a single age group. Sensitivity analyses using the alternative definitions of nosocomial NoV infections showed similar results; odds ratio (OR) (95% confidence interval [CI]) for age 60 years using definition 1:5.0 ( ) (P <.001) and using definition 2:5.7 ( ) (P <.001); OR (95% CI) for GII.4 using definition 1:2.1 ( ) (P =.001) and using definition 2:2.1 ( ) (P =.004). Clinical Data Information of onset of gastroenteritis symptoms was available for 73 inpatients. Samples from 38 patients (52%) were taken on the same day or the day after onset of gastroenteritis symptoms. The stool sample was obtained 2 days before the onset of symptoms in 1 patient, 2 3 days after the onset in 16, and 4 days after the onset in 18. The median time between onset of gastroenteritis symptoms and stool sampling was 1 day (range, 2 to29days; IQR, 1 3 days). DISCUSSION This study was based on data from 3656 inpatients with a verified NoV infection over a 9-year period, Based on conservative estimates, 63% of the patients had nosocomial infections. Nosocomial NoV infection was mainly associated with older patient age, but infection with the specific genotype GII.4 was also associated with nosocomial infections. More diversity of NoV genotypes was detected among inpatients with communityacquired infection than among those with nosocomial infections. More than half of the NoV-positive inpatients in our study had nosocomial infections. This was slightly more than the results reported by Beersma et al [25] and Spackova et al [28], who concluded that 52% and 50% of inpatients, respectively, were nosocomially infected. Beersma et al [25] used the same definition of nosocomial infection as we did, whereas Spackova et al [28]defined nosocomial NoV infection as gastroenteritis symptoms starting >2 days after hospital admission. Local differences in sampling delay and different classifications of nosocomial NoV infections may have caused the observed differences. The determination of nosocomial infections is important as these may be preventable by infection control measures. Widespread environmental contamination has been documented in hospital wards during NoV outbreaks [29], and NoV can remain contagious in the environment for at least several days [11]. Therefore, increased focus on cleaning procedures along with immediate isolation of patients with suspected infectious gastroenteritis may help diminish the number of nosocomial NoV infections. We detected an association between older age and nosocomial NoV infection, in accordance with the study by Spackova et al [28]. Most likely, this is caused by longer hospital stays of the elderly and therefore extended exposition. However, increased susceptibility to NoV of the elderly and chronically ill may also increase transmission in this vulnerable population [8]. In contrast to our finding, Beersma et al [25] reported that a larger proportion of NoV-infected hospitalized children (59%) were nosocomially infected compared with NoV-infected adults (50%). A relatively large proportion of their study population was immunocompromised children, which may have influenced their results [25]. We detected more diversity of NoV genotypes among inpatients with community-acquired NoV infections than among nosocomially infected inpatients. We have reported elsewhere an association between older age and NoV infections with the specific genotype GII.4 among patients from community as well as healthcare settings [19]. Thus, elderly persons may be less susceptible to infections with other NoV genotypes than GII.4. In the current study, nosocomially infected inpatients were 884 JID 2015:212 (15 September) Franck et al

5 Figure 1. Distribution of genotypes among inpatients with community-acquired (n = 333), indeterminate (n = 216), and nosocomial (n = 681) norovirus infections in Denmark, older than inpatients with community-acquired NoV infections, which could explain the observed higher genotype diversity in the latter group. In multivariate analyses, GII.4 infections were also associated with nosocomial NoV infections. It is well established that NoV GII.4 represents the predominant genotype in hospitals and other healthcare settings [9, 19, 20, 22]. Because the elderly seem more susceptible to NoV GII.4 infections [19] and constitute a large proportion of inpatients, this could partly explain the predominance of GII.4 in these settings. However, because we also detected an association between nosocomial infections and infection with NoV GII.4 when controlling for age, other factors must contribute to the predominance of GII.4 in hospital settings. These could include properties increasing the transmission of NoV GII.4 in semiclosed environments, for example, longer shedding duration or increased virus concentrations in vomit or stool [22, 30]. In a study by Sukhrie et al [26] (using the same definition of nosocomial NoV infection as we did), nosocomial NoV infection was associated with NoV GII.3 infections, which is in contrast to our results. One explanation for this difference could be Nosocomial Norovirus Infections JID 2015:212 (15 September) 885

6 Table 2. Association Between Norovirus Acquisition Source and Age, Genotype, Sex, or Geographic Location for 1014 Patients With an Assigned Genotype a Patients, No. (%) Characteristic Total Nosocomial Infection Community-Acquired Infection OR (95% CI) P Value Age, y < (21) 73 (34) 139 (66) 1 (Reference) (79) 608 (76) 194 (24) 5.2 ( ) <.001 Genotype Non-GII.4 96 (9) 38 (40) 58 (60) 1 (Reference)... GII (91) 643 (70) 275 (30) 2.2 ( ).002 Sex Male 470 (46) 291 (62) 179 (38) 1 (Reference)... Female 544 (54) 390 (72) 154 (28) 1.4 ( ).02 Region Rest of Denmark 580 (57) 374 (64) 206 (36) 1 (Reference)... Zealand 434 (43) 307 (71) 127 (29) 1.3 ( ).16 Abbreviations: CI, confidence interval; GII.4, genogroup II genotype 4; OR, odds ratio. a This analysis included only inpatients who had an assigned genotype and were classified as having either nosocomial (sampling 5 days after hospital admission) or community-acquired (sampling 0 1 day after admission) norovirus infections. The variation between hospitals was 0.10 (standard error, 0.09); the intraclass correlation coefficient, 2.9%; and the dispersion parameter, that the majority of patients studied by Sukhrie et al were children 5 years of age, whereas the majority of patients in our study population were older. Presumably, the transmission of specific NoV genotypes within a population is influenced by the age distribution of the infected persons. Female sex was also associated with nosocomial NoV infection. This may be caused by the larger proportions of women among elderly patients and those hospitalized for longer periods. This study has several limitations. Most importantly, clinical information was available only for a small subset of patients. Some of these had 4 days delay from onset of symptoms to stool sampling. Because the incubation time of NoV may be up to 2.5 days [4], some inpatients with community-acquired infections may have been misclassified as nosocomially infected. To minimize this kind of misclassification and make the study results comparable to those of other studies [25, 26], the NoV acquisition source was classified as indeterminate if stools were sampled 2 4 days after hospital admission. Therefore, we believe that few inpatients with community-acquired NoV infections were misclassified as nosocomially infected. On the contrary, because only a few patients are tested for NoV in each hospital outbreak of gastroenteritis in Denmark, it is likely that we underestimated the real proportion of patients with nosocomial NoV infections. However, we obtained similar results with respect to the associations between nosocomial NoV infections and age or infections with NoV GII.4 when we used alternative classifications of nosocomial infections. Second, this study was based on samples from patients who were tested for NoV as part of diagnostic evaluations. Thus, some bias could have been introduced in the selection of patients. Because older patients and those with comorbid conditions can be more severely affected by NoV infections, they may be more frequently tested [8, 31]. This may have biased the results toward an association between nosocomial NoV infection and older age. However, the majority of patients hospitalized for longer periods in our study were 60 years old. Therefore, we believe that the detected association between older age and nosocomial NoV infection is correct because a prolonged hospital stay will increase the risk of acquiring a nosocomial infection. Furthermore, several studies have documented that the majority of hospital NoV outbreaks are caused by GII.4. Thus, it is reasonable to presume that the detected association between nosocomial infection and NoV GII.4 in our study is real. Third, a genotype was established in only 57% of samples selected for genotyping. Genotypes were obtained from a lower number but a larger proportion of inpatients with community-acquired NoV infections (n = 333; 58%) compared with nosocomially infected patients (n = 681; 29%). Ideally, for comparison of genotype diversity, the 2 groups should be equal in size. However, we detected more diversity of NoV genotypes among the smaller group of community-acquired NoV infections than among the nosocomially infected patients. If samples from more inpatients with community-acquired NoV infections had been genotyped, presumably even more diversity would have been detected in this group. In the sensitivity analyses using the alternative definition of nosocomial NoV infections, similar numbers of different genotypes were detected among inpatients with community-acquired 886 JID 2015:212 (15 September) Franck et al

7 and nosocomial NoV infections. It is likely that some patients classified as having nosocomial NoV infections according to this definition actually had community-acquired infections with delayed stool sampling. A larger proportion of patients <18 years old had an assigned genotype compared with the 3107 patients 60 years old. Because the group of patients <18 years old was small (n = 74), we consider it unlikely that the higher proportion of genotyping results in this group significantly influenced the results regarding associations between nosocomial NoV infections, age, and NoV GII.4 infection. Finally, we were not able to study genotypes causing nosocomial infections at an outbreak level. The median number of genotyped samples per ward in each calendar month was 1. Therefore, we believe we avoided the potential bias toward genotypes from large outbreaks. In conclusion, more than half of the NoV-infected inpatients were nosocomially infected. An increased focus on both immediate isolation of patients with suspected infectious gastroenteritis and cleaning procedures could help diminish this proportion. More diversity of NoV genotypes was found among inpatients with community-acquired infection than among those with nosocomial NoV infections, which could reflect the younger ages of inpatients with community-acquired NoV infections. The majority of nosocomially infected inpatients were 60 years old, but we also observed an association between nosocomial infections and infection with the specific NoV genotype GII.4. Further research may help define better strategies to prevent nosocomial NoV infections and outbreaks. Notes Acknowledgments. We thank Camilla Dalgaard and Jesper Rønn for their technical assistance, Lasse Lundby Franck for assistance with the data management, and Marian Jørgensen for help with proofreading. Financial support. This work was supported in part by the Helene E.B. Marckwardts Foundation and the European Commission (project ; EVENT [Enteric Virus Emergence, New Tools]). Potential conflicts of interest. All authors: No reported conflicts. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed. References 1. Hoa Tran TN, Trainor E, Nakagomi T, Cunliffe NA, Nakagomi O. Molecular epidemiology of noroviruses associated with acute sporadic gastroenteritis in children: global distribution of genogroups, genotypes and GII.4 variants. J Clin Virol 2013; 56: Patel MM, Widdowson MA, Glass RI, Akazawa K, Vinjé J, Parashar UD. Systematic literature review of role of noroviruses in sporadic gastroenteritis. Emerg Infect Dis 2008; 14: Haller S, Eckmanns T, Benzler J, et al. Results from the first 12 months of the national surveillance of healthcare associated outbreaks in Germany, 2011/2012. PLoS One 2014; 9:e Lee RM, Lessler J, Lee RA, et al. Incubation periods of viral gastroenteritis: a systematic review. BMC Infect Dis 2013; 13: Rockx B, De Wit M, Vennema H, et al. Natural history of human calicivirus infection: a prospective cohort study. Clin Infect Dis 2002; 35: Haustein T, Harris JP, Pebody R, Lopman BA. Hospital admissions due to norovirus in adult and elderly patients in England. Clin Infect Dis 2009; 49: Ebdrup L, Böttiger B, Mølgaard H, Laursen AL. Devastating diarrhoea in a heart-transplanted patient. J Clin Virol 2011; 50: Mattner F, Sohr D, Heim A, Gastmeier P, Vennema H, Koopmans M. Risk groups for clinical complications of norovirus infections: an outbreak investigation. Clin Microbiol Infect 2006; 12: Bernard H, Höhne M, Niendorf S, Altmann D, Stark K. Epidemiology of norovirus gastroenteritis in Germany : eight seasons of routine surveillance. Epidemiol Infect 2014; 142: Schorn R, Höhne M, Meerbach A, et al. Chronic norovirus infection after kidney transplantation: molecular evidence for immune-driven viral evolution. Clin Infect Dis 2010; 51: Thornley CN, Emslie NA, Sprott TW, Greening GE, Rapana JP. Recurring norovirus transmission on an airplane. Clin Infect Dis 2011; 53: Mathijs E, Stals A, Baert L, et al. A review of known and hypothetical transmission routes for noroviruses. Food Environ Virol 2012; 4: Atmar RL, Opekun AR, Gilger MA, et al. Determination of the 50% human infectious dose for Norwalk virus. J Infect Dis 2014; 209: Seitz SR, Leon JS, Schwab KJ, et al. Norovirus infectivity in humans and persistence in water. Appl Environ Microbiol 2011; 77: Liu P, Yuen Y, Hsiao HM, Jaykus LA, Moe C. Effectiveness of liquid soap and hand sanitizer against Norwalk virus on contaminated hands. Appl Environ Microbiol 2010; 76: Lee BY, Wettstein ZS, McGlone SM, et al. Economic value of norovirus outbreak control measures in healthcare settings. Clin Microbiol Infect 2011; 17: Harris JP, Adams NL, Lopman BA, Allen DJ, Adak GK. The development of Web-based surveillance provides new insights into the burden of norovirus outbreaks in hospitals in England. Epidemiol Infect 2014; 142: Kroneman A, Vega E, Vennema H, et al. Proposal for a unified norovirus nomenclature and genotyping. Arch Virol 2013; 158: Franck KT, Fonager J, Ersbøll AK, Böttiger B. Norovirus epidemiology in community and health care settings and association with patient age, Denmark. Emerg Infect Dis 2014; 20: Vega E, Barclay L, Gregoricus N, Shirley SH, Lee D, Vinjé J. Genotypic and epidemiologic trends of norovirus outbreaks in the United States, 2009 to J Clin Microbiol 2014; 52: Desai R, Hembree CD, Handel A, et al. Severe outcomes are associated with genogroup 2 genotype 4 norovirus outbreaks: a systematic literature review. Clin Infect Dis 2012; 55: Kroneman A, Verhoef L, Harris J, et al. Analysis of integrated virological and epidemiological reports of norovirus outbreaks collected within the Foodborne Viruses in Europe network from 1 July 2001 to 30 June J Clin Microbiol 2008; 46: Vinjé J, Altena SA, Koopmans MP. The incidence and genetic variability of small round-structured viruses in outbreaks of gastroenteritis in the Netherlands. J Infect Dis 1997; 176: Kroneman A, Vennema H, Deforche K, et al. An automated genotyping tool for enteroviruses and noroviruses. J Clin Virol 2011; 51: Beersma MF, Schutten M, Vennema H, et al. Norovirus in a Dutch tertiary care hospital ( ): frequent nosocomial transmission and dominance of GIIb strains in young children. J Hosp Infect 2009; 71: Sukhrie FH, Beersma MF, Wong A, et al. Using molecular epidemiology to trace transmission of nosocomial norovirus infection. J Clin Microbiol 2011; 49: Holzknecht BJ, Franck KT, Nielsen RT, Böttiger B, Fischer TK, Fonager J. Sequence analysis of the capsid gene during a genotype II.4 dominated norovirus season in one university hospital: identification of possible transmission routes. PLoS One 2015; 10:e Nosocomial Norovirus Infections JID 2015:212 (15 September) 887

8 28. Spackova M, Altmann D, Eckmanns T, Koch J, Krause G. High level of gastrointestinal nosocomial infections in the German surveillance system, Infect Control Hosp Epidemiol 2010; 31: Nenonen NP, Hannoun C, Svensson L, et al. Norovirus GII.4 detection in environmental samples from patient rooms during nosocomial outbreaks. J Clin Microbiol 2014; 52: Rosenthal NA, Lee LE, Vermeulen BA, et al. Epidemiological and genetic characteristics of norovirus outbreaks in long-term care facilities, Epidemiol Infect 2011; 139: Partridge DG, Evans CM, Raza M, Kudesia G, Parsons HK. Lessons from a large norovirus outbreak: impact of viral load, patient age and ward design on duration of symptoms and shedding and likelihood of transmission. J Hosp Infect 2012; 81: JID 2015:212 (15 September) Franck et al