If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
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
Hospital norovirus outbreaks cause significant financial and operational disruption which should be minimised by optimal handling of affected areas and use of isolation facilities.
Aim
To identify factors associated with increased duration of symptoms and viral excretion and increased probability of transmission.
Methods
Retrospective observational study of a large norovirus outbreak at a UK teaching hospital in the winter of 2009–2010 where patients were diagnosed using a real-time polymerase chain reaction (PCR) assay.
Findings
Symptom duration was significantly associated with patient age (Spearman rank correlation coefficient: 0.197; P = 0.002) but not with PCR cycle threshold (CT) value. Duration of viral excretion was found to be longer in patients with higher viral loads. Transmission within a ward bay was not significantly associated either with age or with CT value but was more likely to occur in some ward blocks than others, which may relate to differences in ward design. Transfer of patients into isolation rooms or cohorted area within two days of symptom onset did not significantly influence probability of onward transmission (52% vs 47%; P = 0.67).
Conclusions
The presented data suggest that CT value may guide timing of repeat sample collection if ongoing gastrointestinal symptoms may relate to other pathologies, and that patients developing symptoms of norovirus may remain in their current bay rather than being moved into isolation facilities. The bay or ward should be closed to new admissions but it should be anticipated that duration of symptoms and therefore closure will be longer when the outbreak involves elderly patients.
Norovirus outbreaks in hospitals have caused significant financial and operational disruption to healthcare services over recent years, possibly associated with the emergence of genotype II.4 viruses.
Entry of virus into the hospital is inevitable during community outbreaks and the serial introduction of genetically distinct strains during a single season has been reported.
However, such measures are based on limited evidence, often fail to contain the virus, and facilities for testing, isolation and cohorting are frequently overrun.
Symptoms following norovirus infection usually last for 24–48 h in otherwise healthy individuals but may be prolonged in the immunocompromised, children and the elderly.
Polymerase chain reaction (PCR) has superior sensitivity to other methods of norovirus detection, is highly reproducible and is the diagnostic method of choice in an increasing number of laboratories.
Mattison K, Grudeski E, Auk B, et al. Analytical performance of norovirus real-time RT–PCR detection protocols in Canadian laboratories. J Clin Virol 2011;50:109–113.
Previous studies have suggested that PCR cycle threshold (CT), which is inversely proportional to viral load, predicts those patients in whom the virus is causing diarrhoeal disease with higher CT values representing asymptomatic carriage or diarrhoea of alternative aetiology.
The infectivity and infection control implications of such results has not been addressed. Virus can be detected by PCR for up to eight weeks in healthy volunteers, but the risk of transmission is expected to be greatest early in the illness when peak viral loads coincide with vomiting or frequent uncontrolled diarrhoea, providing the rationale for patient isolation or cohorting until 48 h beyond the symptomatic period.
The objectives of this observational study of a large nososomial norovirus epidemic over the winter of 2009–2010 were: (i) to ascertain whether symptom duration and onward transmission of virus were associated with patient age or CT value; (ii) to identify whether duration of norovirus shedding, as detected by PCR, was associated with initial CT value; (iii) to determine whether early isolation or cohorting of symptomatic patients reduced the risk of transmission.
The study was performed retrospectively but all data on patient symptom onset and duration, staff symptoms and patient locations were gathered prospectively using an outbreak proforma as part of the standard infection control practice.
Methods
The Northern General Hospital (NGH) has 1100 beds and provides acute medical and surgical care for patients in Sheffield, UK (population ∼500,000) and tertiary referral services for the surrounding area. A number of specialties, including haematology and oncology, are located at other hospitals within the same trust. The virology laboratory is located at NGH and provides norovirus diagnostic services for the entire trust and local community-based services.
PCR technique
Real-time PCR was introduced in 2009 for norovirus diagnosis using a previously described method.
The assay detects the most conserved sequence at the ORF1–ORF2 junction region in the norovirus genome. Faecal extracts are prepared using the Roche STAR Buffer and extracted using Roche MagNApure or Qiagen Qiacube systems. Extracted nucleic acid is converted to cDNA. Amplification and detection of norovirus cDNA is carried out using the ABI prism 7500. All samples are tested for genogroups 1 and 2 using specific primer and probe sequences.
Patient sampling
Prior to 31 January 2010, all faecal samples received by the laboratory were tested for the presence of norovirus by the above method. From 1 February 2010, to reduce laboratory demand, testing was restricted to admissions units and areas from which patients with suggestive symptoms had been reported to the infection prevention and control team (IPCT). The period 1 December 2009 to 31 January 2010 was selected for analysis of correlation of symptom duration and onward transmission with patient age or CT value. For investigation of duration of norovirus shedding, a longer period of 1 December 2009 to 1 April 2010 was analysed to increase the number of patients included. Only patients with genogroup 2 virus were included as very few were identified with genogroup 1 disease.
The IPCT (three infection control nurses and three infection control assistants, non-nursing qualified support workers whose role was to support the qualified nurses with ward reviews and administrative work) visited or contacted wards daily and recorded the presence of ongoing symptoms, location of affected patients and the details of new patients or staff developing symptoms. Attempts were made to obtain samples for testing from all symptomatic patients.
Where possible, affected patients and their contacts were isolated or cohorted and clinical areas closed in an attempt to reduce the risk of transmission to new patients. Briefly, if isolated cases developed within a bay then that bay would be closed until 72 h beyond the last loose stool or vomit of any patient (thus exceeding the 48 h window specified in recently published national guidance).
Norovirus Working Party, Health Protection Agency. Guidelines for the management of norovirus outbreaks in acute and community health and social care settings, 2011. London: HPA; 2011.
The bay would then undergo thorough cleaning with hypochlorite and change of curtains. If more than one bay was affected within a clinical area, or if staff were affected, the whole ward would be closed until 72 h beyond last symptoms. Cohort wards were created on an ad hoc basis to facilitate cleaning and re-opening of other areas. Twice daily cleaning with 0.1% hypochlorite was instituted during outbreaks with particular attention paid to toilets, commodes and frequently touched areas such as door handles and rails.
Duration of symptoms
Onset of diarrhoea or vomiting was recorded by the IPCT. Recovery from symptoms was defined as the first symptom-free day assuming that no relapse of symptoms occurred within the subsequent 48 h. A small number of patients, who had ongoing intermittent symptoms interrupted by asymptomatic periods of >48 h, were excluded from analysis of symptom duration.
Analysis of the duration of norovirus RNA shedding
To determine the duration of PCR positivity, all patients with at least one further sample tested after an initial positive result were included. Individual patients were deemed to be positive every day up until the day of their last positive sample and negative from the day of their first negative sample. Probability of PCR positivity on any given day after the initial positive sample was then calculated. No result was assumed for days between the last positive and first negative samples or for days beyond the last positive result in patients who did not subsequently have a negative sample.
Norovirus transmission data
Presumed transmission events from patients who developed symptoms within a bay were defined by a further patient within the same bay developing symptoms at least 24 h after the onset of symptoms in the index case but no more than 48 h beyond their last exposure to a symptomatic index. Comparison was made with patients who developed symptoms within a bay without subsequent secondary cases. Patients residing in wards on which concurrent staff disease was documented were excluded from this analysis as it would be impossible to determine whether secondary cases were acquired from staff or patients.
Statistical analysis
Data were analysed using PASW statistics 18 (IBM Ltd, New York, USA). Spearman rank correlation coefficient (rS) was calculated for association of continuous non-parametric data sets. Mann–Whitney U-test was used for comparisons of independent non-parametric continuous data sets whereas categorical data with independent outcomes were compared using the chi-squared or Fisher's exact tests.
Results
Between November 2009 and April 2010, 623 individual inpatients at NGH with norovirus infection were diagnosed by the laboratory. During the months of December and January, which were subjected to further analysis, 353 inpatients were found to be positive and there were 38 complete ward closures.
Symptom duration vs age
Figure 1 is a scatter plot of symptom duration against age, for patients with confirmed norovirus, during December and January. A total of 239 patients, who developed symptoms while in hospital, were included in the analysis with a mean age of 78.1 years. Symptom duration increased with age, with a weak but significant correlation (rS = 0.197; P = 0.002). The mean duration of symptoms for patients aged >80 years was 5.7 days compared with 3.7 days for those aged <80 years.
Figure 1Duration of symptoms vs age of 239 patients developing norovirus while inpatients at the Northern General Hospital, Sheffield, December 2009–2010. Symptom duration was weakly but significantly correlated with age (rS= 0.197; P = 0.002).
No significant correlation was identified between duration of symptoms from time of sampling and CT value of the sample (rS = –0.077; P > 0.2, data not shown).
Duration of PCR positivity
In all, 266 patients had further samples tested after an initial positive. The probability of repeat samples being positive is plotted against days after initial sample in Figure 2, stratified by initial CT value. Those patients with an initial CT value of <30 were significantly more likely to remain positive at seven days than patients with an initial CT value of >30 (97% vs 29% persistent positivity; P < 0.0001, chi-squared test). For patients with an initial CT value of <30, 83% remained positive at two weeks and 57% at three weeks.
Figure 2Probability of repeat sample positivity stratified by cycle threshold (CT) value of original sample. Numbers in each CT stratification are shown in parentheses. PCR, polymerase chain reaction.
Probability of onward transmission from patients developing symptoms within a bay
A total of 110 patients, who developed symptoms during December and January, were residing in a bay in the absence of documented illness among staff members. Of these, 57 (52%) were deemed to have transmitted virus onward by the criteria outlined above. There was no significant difference in age or initial CT value between onward transmitters and non-transmitters (Table I). NGH comprises a number of different buildings of varying ward design, and probability of transmission appeared to be associated with ward type (Table II). Patients developing symptoms on Chesterman wards had a significantly reduced probability of onward transmission compared with those elsewhere. Those developing symptoms in the Hadfield wing also had a numerically lower probability of transmission compared with other areas but this did not reach statistical significance, the greatest reduction being in comparison to the Firth wing (0.36 vs 0.69; P = 0.08). Current national guidance, with which the Hadfield wing complies, suggests a minimum distance between bed centres of 3.6 m and a maximum of 12 m between each bedspace and a toilet.
Significantly decreased risk of transmission from bay patients in Chesterman wing compared with Huntsman (P = 0.009), Brearley (P = 0.001), and Firth (P < 0.001) (Fisher's exact test).
Significantly decreased risk of transmission from bay patients in Chesterman wing compared with Huntsman (P = 0.009), Brearley (P = 0.001), and Firth (P < 0.001) (Fisher's exact test).
Significantly decreased risk of transmission from bay patients in Chesterman wing compared with Huntsman (P = 0.009), Brearley (P = 0.001), and Firth (P < 0.001) (Fisher's exact test).
Significantly decreased risk of transmission from bay patients in Chesterman wing compared with Huntsman (P = 0.009), Brearley (P = 0.001), and Firth (P < 0.001) (Fisher's exact test).
1985
Orthopaedic and plastic surgery
7
34
6
2.42
4
5.7
3
82.5
Number of beds per toilet was calculated as the total number of beds on the ward divided by the number of toilets for patient use.
a Trend to decreased transmission from bay patients in Hadfield wing compared with Firth (P = 0.08). P > 0.1 for other wing comparisons.
b Significantly decreased risk of transmission from bay patients in Chesterman wing compared with Huntsman (P = 0.009), Brearley (P = 0.001), and Firth (P < 0.001) (Fisher's exact test).
Transfer of patients from a bay into a side room or cohorted area within two days of symptom development did not significantly influence the probability of onward transmission. Onward transmission occurred from 52% of patients who were isolated within 48 h compared with 47% of those not isolated (P = 0.67; Table I).
Discussion
This study analysed data from a large norovirus outbreak in a UK teaching hospital. To our knowledge, this is the first study to examine the duration of PCR positivity in large numbers of hospital inpatients and to attempt to correlate individual transmission events with norovirus viral load, patient age, ward type and patient isolation.
These data confirm those of other authors in demonstrating that advanced age is associated with prolongation of symptoms of norovirus infection.
An infection control strategy based upon allowing outbreaks to run their course on individual wards may therefore result in longer periods of ward closure in areas with a large proportion of elderly patients.
The increased probability of onward transmission with increasing age was not statistically significant but the sample size analysed was relatively small. Possible reasons for increased transmission from elderly patients might include prolonged duration of symptoms, incontinence and immobility necessitating assistance with toileting or increased contact with healthcare professionals such as physiotherapists.
No association was found between viral load and transmission or symptom duration. A previous study suggested that asymptomatic carriage may be discriminated from true infection by CT value.
The other study examined community stool samples and did not consider the infection control implications of weak positive norovirus PCR results in patients with gastrointestinal symptoms. Because of the low infecting dose of norovirus, it is probable that asymptomatic patients have a role to play in the initiation or prolongation of outbreaks. The data presented here demonstrate that symptomatic patients with CT > 30 are still capable of onward transmission and should therefore continue to be isolated or cohorted. This is not surprising given the low infecting dose of norovirus and suggests that diagnostic PCR amplification should be over a minimum of 40 cycles. Patients with CT > 30 were, however, significantly more likely to be PCR negative at one week than those with lower CT values. This may suggest that some patients with higher CT values have only transient low level carriage or infection. Decisions regarding cessation of infection control precautions rely upon resolution of clinical symptoms with no role for repeat testing except in patients who have prolonged diarrhoea with possible alternative aetiology. If required, these results suggest that repeat testing may be considered at one week for those with CT > 30 but that very few of those with CT < 30 will become negative within two weeks.
A previous study found an association between greater frequency of outbreaks and geriatric or medical wards or higher numbers of beds per ward.
Institutional risk factors for outbreaks of nosocomial gastroenteritis: survival analysis of a cohort of hospital units in South-west England, 2002–2003.
These data also suggest differing likelihood of onward transmission within a bay between ward types. This difference reached statistical significance for comparisons of the Chesterman wing with other areas. Factors which may account for this difference include patient mix, length of stay, staff knowledge or attitudes and ward design. The Hadfield wards were also associated with lower rates of transmission than other areas including areas with comparable patient mix, although these comparisons did not reach statistical significance. It would be helpful to replicate this study in other hospitals with varying ward design in order to confirm the significance of the effect and tease out critical factors in ward design, which could then guide future building programmes. Further study of the effect of number of beds per toilet would be of particular interest.
A limitation of this study is the fact that patient comorbidity, including immune status, was not included as a variable. The number of patients included in the study precluded detailed analysis of comorbidity but haematology, oncology, rheumatology and human immunodeficiency virus patients are routinely managed at alternative hospitals and so the numbers of patients with profound immunocompromise would be expected to be small.
Comparisons between CT values of individual patients need to be interpreted with some caution given the presence of substances inhibitory to PCR in faeces and the lack of an internal inhibition control. An internal inhibition control has now been introduced into our assay but, in common with other centres performing norovirus PCR at the time, it was not present during the time period studied. Since introduction of the internal control, <1% of samples have proved to be inhibited by faecal contents and, therefore, the presence of inhibitors is likely to have had only a limited impact on the results.
Data on transmission of norovirus from individual patients may have been compromised by secondary cases occurring as a result of transmission from individuals other than the presumed index case, such as unrecognised symptomatic patients, visitors or staff, although efforts were made to ensure that those cases occurring at the time of documented staff illness were excluded. It is possible that the presence of staff illness might reflect a more transmissible circulating strain of norovirus in that particular area, but, given the fact that the studied outbreaks occurred over a two-month period, it is likely that strains affecting different areas were closely related.
The strategy for infection control management of individual cases also varied depending upon the availability of side rooms during the outbreak, which could also potentially have confounded transmission statistics, especially with regard to the comparison between ward designs. However, isolation from a bay area within 48 h was not associated with a reduction in the probability of transmission in this study.
Finally, a longer time period was analysed for study of duration of shedding but, as stated in Methods, testing was only performed on samples from admissions units and areas reported to have symptomatic patients during the latter half of this period. It is therefore possible that patients in clinical areas with an increased incidence of diarrhoea may have been more likely to have repeat tests and therefore be overrepresented in the analysis. Patients with intermittent symptoms, whose diarrhoea and vomiting relapsed following a 48 h symptom-free period, were excluded from the analysis of symptom duration. Unfortunately, none of these patients had repeat tests at the time of their relapse and the trajectory of their CT value is therefore uncertain. Such patients present a significant challenge to infection control teams as they may be assumed to be over the symptomatic period and be placed in contact with previously uninfected patients. In the context of immunocompromise, prolonged intermittent shedding of virus has been well documented.
In summary, norovirus viral load, as extrapolated from real-time PCR cycle threshold values, is a poor predictor of duration of symptoms but does influence duration of viral RNA shedding and may be used to guide timing of further testing in those patients with other potential causes for ongoing gastrointestinal symptoms. Multiple factors are likely to influence the probability of transmission of virus from one patient to another but the higher rates of transmission from older patients and those with higher viral loads found were not statistically significant. Moreover, presumed transmissions did occur from patients with very low viral loads, implying that these patients should continue to be managed as an infection control risk. A possible association between ward type and likelihood of onward transmission was identified. Further work is required to confirm these findings in other settings and to identify the key factors in ward design which influence norovirus transmission.
Conflict of interest statement
None declared.
Funding sources
None.
References
Zheng D.P.
Widdowson M.A.
Glass R.I.
Vinje J.
Molecular epidemiology of genogroup II-genotype 4 noroviruses in the United States between 1994 and 2006.
Mattison K, Grudeski E, Auk B, et al. Analytical performance of norovirus real-time RT–PCR detection protocols in Canadian laboratories. J Clin Virol 2011;50:109–113.
Norovirus Working Party, Health Protection Agency. Guidelines for the management of norovirus outbreaks in acute and community health and social care settings, 2011. London: HPA; 2011.
Institutional risk factors for outbreaks of nosocomial gastroenteritis: survival analysis of a cohort of hospital units in South-west England, 2002–2003.
00061-8&id=gr1.jpg){kind=link}
00061-8&id=gr2.jpg){kind=link}