Guideline for the Prevention and Control of Norovirus Gastroenteritis Outbreaks in Healthcare Settings, 2011
VIII. Evidence Review
Question 2: What are the best methods to identify an outbreak of norovirus gastroenteritis in a healthcare setting?
To address this question, studies that provided test characteristics for the diagnosis of norovirus or outbreaks of norovirus gastroenteritis were critically reviewed. The available data examined the use of clinical criteria for the diagnosis of an outbreak of norovirus, methods of specimen collection for the diagnosis of a norovirus outbreak, and characteristics of tests used to diagnose norovirus. The evidence consisted of 33 diagnostic studies.[17,18,116-146] The findings from the evidence review and the grades of evidence for clinically relevant outcomes are shown in Evidence and Grade Table 2.
Q2.A Clinical Criteria
There was moderate quality evidence from a single diagnostic study supporting the use of the Kaplan criteria to detect outbreaks of norovirus gastroenteritis.[16,116] Of 362 confirmed gastroenteritis outbreaks with complete clinical or laboratory data, the sensitivity of the Kaplan Criteria to detect an outbreak of norovirus gastroenteritis without an identified bacterial pathogen was 68.2%, with a specificity of 98.6%. The positive predictive value (PPV) was 97.1% and the negative predictive value was 81.8%. Individual criteria, such as vomiting among >50% of a patient cohort, brief duration of illness (12-60 hours), or mean incubation time of 24-48 hours, demonstrated high sensitivities (85.8-89.2%), but specificities were low (60.7-69.6%). The use of additional criteria, such as the ratios of fever-to-vomiting and diarrhea-to-vomiting, provided sensitivities of 90.1% and 96.6%, and specificities of 46.6% and 44.5%, respectively. Applied to the 1141 outbreaks of unconfirmed etiology, suspected norovirus or bacterial sources with complete data, the Kaplan criteria estimated that 28% of all 1998-2000 CDC-reported foodborne outbreaks might be attributable to norovirus. The searchstrategy did not identify studies that have assessed the utility of the Kaplan criteria in healthcare-associated outbreaks of norovirus gastroenteritis.
Q2.B Specimen Collection
There was low-quality evidence from three diagnostic studies outlining the minimum number of stool samples from symptomatic patients required to confirm an outbreak of norovirus gastroenteritis.[117,119,120,122,123] In modeling analyses using a hypothetical test demonstrating 100% sensitivity and 100% specificity, obtaining a positive EIA result from two or more submitted samples demonstrated a sensitivity of 52.2-57%, with a peak in sensitivity when at least one from a total of six submitted samples was positive for norovirus (71.4-92%). Specificity was 100% when at least one positive EIA was obtained from a minimum of two submitted stool samples.
Using a reverse transcriptase polymerase chain reaction (RT-PCR) method, if at least one positive test was identified among 2 to 4 submitted stool specimens from symptomatic persons, the test sensitivity was greater than 84%. When 5-11 stool samples were submitted and at least 2 were confirmed as positive, the sensitivity of PCR was greater than 92%. When at least one stool specimen was submitted for identification, PCR confirmed norovirus as the causative agent in a larger proportion of outbreaks than those using EM or ELISA methods, and is currently the Gold Standard. This evaluation was unable to determine how diagnostic test characteristics are affected by the timing of specimen collection relative to the disease process.
Q2.C Diagnostic Methods
28 diagnostic studies[17,18,118-120,122,124-139,141-145,147] and 1 descriptive study that evaluated the test characteristics of EIA such as ELISA, EM, reverse transcriptase PCR, and nucleic acid sequence-based amplification (NASBA) in the detection of norovirus in human fecal specimens were summarized. Test characteristics for the most common or commercially-available norovirus diagnostics are summarized in Table 3.
2.A.1 In the absence of clinical laboratory diagnostics or in the case of delay in obtaining laboratory results, use Kaplan’s clinical and epidemiologic criteria to identify a norovirus gastroenteritis outbreak (see Table 4 for Kaplan’s criteria). (Category IA) (Key Question 2A)
2.A.2 Further research is needed to compare the Kaplan criteria with other early detection criteria for outbreaks of norovirus gastroenteritis in healthcare settings, and to assess whether additional clinical or epidemiologic criteria can be applied to detect norovirus clusters or outbreaks in healthcare settings. (No recommendation/unresolved issue) (Key Question 2A)
2.B Consider submitting stool specimens as early as possible during a suspected norovirus gastroenteritis outbreak and ideally from individuals during the acute phase of illness (within 2-3 days of onset). It is suggested that healthcare facilities consult with state or local public health regarding the types of and number of specimens to obtain for testing. (Category II) (Key Question 2B)
2.C Use effective laboratory diagnostic protocols for testing of suspected cases of viral gastroenteritis (e.g., refer to the Centers for Disease Control and Prevention (CDC)'s most current recommendations for norovirus diagnostic testing [PDF - 850 KB]. (Category IB)(Key Question 2C)