|Home | About | Journals | Submit | Contact Us | Français|
Norovirus strains were detected in two patients and in environmental swabs from a pediatric primary immunodeficiency unit in London, United Kingdom, during an infection control incident in November and December 2007. Detailed analyses of the gene encoding the P2 domain demonstrated that the majority of the strains were not related to the patients and that the environmental contamination was most likely due to secondary transfer by the hands of staff or visitors.
Viral gastroenteritis in a normal pediatric population is primarily caused by rotaviruses (RV), noroviruses (NoV), sapoviruses (SaV), astroviruses (AstV), and enteric adenoviruses (AdV) (1). With the introduction of improved detection assays, the roles of NoV and SaV are becoming more apparent (7, 12). These viruses are frequently associated with diarrhea and vomiting in children under 5 years of age, and sporadic cases have been described in hospitalized children (9, 10, 15), day care centers (13), and the community (7).
Transmission of gastroenteric viruses is usually through person-to-person spread by the fecal-oral route. Environmental transmission of NoV, RV, and AstV (5, 6) has been described in hospitals and involves contaminated work surfaces, floors, medical equipment, light switches, taps, door handles, and television/game consoles.
The extent of environmental contamination with NoV in hospital wards and the effects they have on the prolongation of outbreaks and patient morbidity are difficult to determine. The unique contribution of the data in this study is the analysis of hypervariable nucleotide sequences of the gene encoding the P2 domain of the norovirus capsid, demonstrating true transmission events via environmental contamination in hospitals.
Fecal samples were collected from two patients in the Paediatric Primary Immunodeficiency Unit (PPIU) at Great Ormond Street Hospital (GOSH), London, United Kingdom, who presented with prolonged symptoms of NoV gastroenteritis between the 8 October and 17 December 2007. Six fecal samples were collected from patient 1 and five from patient 2.
Patient 1 was a male infant diagnosed with Wiskott-Aldrich syndrome and had received a bone marrow transplant (BMT) from a matched unrelated donor (MUD) on 29 September. NoV was first detected in fecal samples on day 11 posttransplant, and he continued to excrete virus until day 80 posttransplant.
Patient 2 was a male infant diagnosed with severe combined immunodeficiency syndrome (SCID). He received a bone marrow transplant (BMT) from a matched unrelated donor on 26 September. He was found to be excreting NoV on day 34 posttransplant and continued to excrete until day 82 posttransplant (see Table Table11 for details on the patient samples).
The dates of collection of the patient samples were subjected to the hospital's own policy of sampling when a patient shows symptoms of diarrhea. Environmental swabs were collected on six different occasions from various sites within the unit between 14 November and 22 December 2007. During this period swabs were collected weekly until 12 December (swabs from 29 November and 6 December were excluded due to batch failure), and between 12 and 22 December swabs were taken more frequently to aid the infection control monitoring within the unit. Sampling was performed in areas within physical proximity to the children but without direct contact. In total, 116 environmental swabs were collected during this period (Table (Table11).
All samples were processed for nucleic acid extraction as previously described (5). Briefly, nucleic acid was extracted from fecal samples and environmental swabs by a guanidinium isothiocyanate-silica method, and the eluted RNA was converted to cDNA by using random primers and a reverse transcriptase enzyme (3). A real-time PCR assay was used for the detection of NoV GI and GII strains (8), and all positives were characterized using NoV S (shell) domain genotyping primers (4) and genotype-specific P2 (protruding) domain primers (14), followed by nucleic acid sequencing and phylogenetic analysis.
NoV was detected in fecal samples collected from patients 1 (six samples) and 2 (five samples) during this study, and the strains were genotyped on the basis of the S domain as GII-3. Furthermore, 93/116 (80%) of the environmental swabs were positive for NoV by real-time PCR. Of these NoV positives, 60% were characterized as GII-3 strains, 6% as GII-4, and 34% were not typed due to insufficient material.
P2 domain typing with GII-3-specific primers was performed on all GII-3 positives; however, only 18/56 (32%) of positive environmental samples for GII-3 were successfully amplified and sequenced. P2 domain sequence was obtained from all fecal samples. The GII-3 strains genotyped by analysis of the region encoding the S domain from both patients differed only by a single silent nucleotide change within 282 bp of the 5′ end of this region. All the environmental GII-3 strains were either patient 1-like or patient 2-like in this same region (Fig. (Fig.1A).1A). All five of the GII-4 strains detected in environmental swabs were individually unique by the S domain genotyping (data not shown). When analyzing the P2 domain sequences, patient 1 was found to be excreting three unique strains, the first strain (identified as patient 1-01) was first detected on 8 October, the second strain (patient 1-02) was excreted between 12 and 22 November, and the third strain (patient 1-03) was excreted between 10 and 17 December (nucleotide changes ranged from 98.6 to 99.4%). Patient 2 also had three unique strains: patient 2-01 on 30 October, patient 2-02 between 5 and 16 November, and patient 2-03 on 17 December (nucleotide changes ranged from 98.8 to 99.8%) (Fig. (Fig.1B1B and Table Table1).1). Only three environmental swabs were found to contain virus of identical sequence in the gene encoding the P2 domain to any of the strains excreted by patients 1 and 2. Environmental samples from a trolley on 22 November had a patient 1-01 strain, a clinical waste bin from 12 December had a patient 2-02 strain, and a trolley on 15 December had a patient 1-02 strain (Table (Table11 and Fig. Fig.1B).1B). Analysis of the P2 domain identified 15 unique GII-3 strains (Table (Table1),1), detected only in the environmental swabs and not in the patients' fecal samples (Fig. (Fig.1B).1B). Eight GII-3 strains collected between 2006 and 2007 from unrelated outbreaks of gastroenteritis were included for comparison (Fig. 1A and B).
This study has established that NoV strains excreted by patients and detected in environmental swabs of contaminated surfaces and equipment can appear similar when characterized by genotyping a region of Orf2 (an open reading frame) encoding the S domain. However, this region has limited use in genotyping due to the high degree of nucleotide conservation. In contrast, sequencing of the P2 domain provided a completely different picture of the NoV contamination within the PPIU and of the strains infecting the two patients. Analysis of S domain sequences indicates that the NoV strains excreted by the two patients were found on multiple environmental sites within the unit and may have been the result of cross-contamination involving those patients. However, sequencing and detailed analysis of the gene encoding the P2 domain indicated that viruses identical to those excreted by the two patients were only detected on three environmental sites and that more complex populations of viruses of the same genotype were found to be contaminating the environment (2, 11). The swab sites found to be contaminated with strains identical to those from the two patients were not in direct contact with the patients but were most likely contaminated by secondary transfer by the hands of staff or parents. Environmental contamination with viruses unrelated to those from the patients was most likely to have come from patients' visitors and/or siblings attending the unit and not as a result of cross-contamination by staff, who employ stringent infection control procedures in this vulnerable group of patients. The complex population of GII-3 viruses detected in the environment may have arisen from strains accumulating point mutations during their replication in the immunodeficient human hosts over a prolonged period and their subsequent seeding into the environment. The detection of 21 different virus variants characterized by differences in the gene encoding the P2 domain suggests that this is unlikely to have happened over a period of only 11 weeks and would suggest multiple introductions of GII-3 virus strains from the community, where it is currently the most common NoV strain detected in children under 5 years of age with acute gastroenteritis (7). This is supported by the detection of common GII-4 strains in the environment but not in patients on the unit.
Published ahead of print on 5 May 2010.