Laboratory Tests 151

are patients with underlying conditions and an increased risk of serious complications (see chapter “Clinical Presentation”). In particular, the diagnosis of influenza in elderly patients makes the clinician aware of a substantial risk of secondary bacterial infections with Staphylococcus aureus, Haemophilus influenzae and Streptococcus pneumoniae.

In addition, rapid testing for influenza virus plays a role in hospital infection control in reducing the spread of infection from patient to patient or from infected health care workers to high risk patients. These tests can also be used to diagnose influenza in travellers or outbreaks in semi-closed communities such as cruise ships (WHO 2005a).

Finally, diagnosis of influenza has prognostic value in healthy young adults where the disease has a short and benign course.

Surveillance

Influenza sentinel surveillance employs a variety of test and there seems a lack of standardisation even within the European region (Meerhoff 2004). Different techniques have different advantages and disadvantages. Therefore combinations of tests are used for surveillance. Rapid direct techniques such as RT-PCR (Bigl 2002) or EIA enable the fast detection of epidemics and can be used to distinguish between influenza A or B. Isolation of the virus in embryonated chicken eggs or on cell culture is necessary to subtype the viruses. Haemagglutinin and neuraminidase subtypes are respectively determined by haemagglutination inhibition assay and RT-PCR. Sequencing of PCR products is used to establish the molecular epidemiology of circulating viruses. This together with interstrain haemagglutination inhibition titres enable WHO to recommend appropriate vaccines that will most likely be protective against the circulating influenza strains. Surveillance is also important for public health policies since the health impact of a particular epidemic and cost benefit ratios of interventions such as vaccination can motivate policy makers to prioritise influenza prevention.

Laboratory Tests

Many factors should be considered in deciding which tests to use. Sensitivity, specificity, turn-around-time, repeatability, ease of performance and costs should all be taken into account. RT-PCR is generally more sensitive than serology and culture and the combination of RT-PCR with serology more sensitive than the combination of any other two methods (Zambon 2001). The sensitivity of culture is largely dependent on the laboratory where it is performed. Serology tends to be less expensive than RT-PCR but as it necessitates acute and convalescent blood specimens, diagnosis is only retrospective. Traditional culture is time-consuming but shell vial culture techniques allow diagnosis within 48-72 hours.

Direct methods

Different methods exist for direct detection of influenza viruses. Some methods such as enzyme immunoassays (EIAs) can be suitable for bedside testing, others such as direct immunofluorescence allow for the preparation of slides onsite in clinics and posting of fixed slides to a central laboratory (Allwinn 2002). RT-PCR can only be performed in well equipped laboratory facilities by trained personnel.

152 Laboratory Findings

These methods can either detect both influenza A and B or differentiate between types (influenza A or B). The only direct technique that has the potential to differentiate between subtypes (i.e. on the basis of haemagglutinin and neuraminidase) is RT-PCR.

Immunofluorescence

For direct immunofluorescence, potentially infected respiratory epithelial cells are fixed to a slide and viral antigens contained in the cells is detected by specific antibodies which are either directly conjugated to a fluorescent dye (direct immunofluorescence) or detected by anti-antibodies linked to a fluorescent dye (indirect immunofluorescence). In both cases reactions are visualised under the fluorescence microscope and positive cells are distinguished on colour intensity and morphology of fluorescent areas. Direct immunofluorescence tends to allow faster results but is generally less sensitive than indirect immunofluorescence. Indirect immunofluorescence also has the advantage that pooled antisera can be used to screen for viral infection using a single anti-antibody conjugated to a fluorescent dye (fluorescein isothiocyanate-conjugated anti-mouse antibodies are commonly used; Stevens 1969). Immunofluorescence allows for the rapid diagnosis of respiratory specimens as long as sufficient respiratory epithelial cells are present in the specimens. However, inter-individual variation in reporting of immunofluorescence tests exists since interpretation is subjective and accuracy depends on the competence and experience of the operator.

Enzyme immuno assays or Immunochromatography assays

Enzyme immunoassays (EIAs) utilise antibodies directed against viral antigen that are conjugated to an enzyme. An incubation step with a chromogenic subtrate follows and a colour change is indicative of the presence of viral antigen. Certain enzyme immunoassays as well as similar assays using immunochromatography allow for bedside testing (Allwinn 2002) taking 10-30 minutes. These rapid assays are generally more expensive than direct immunofluorescence or virus culture. Sensitivities of EIAs vary between 64% and 78% (Allwinn 2002). Different rapid tests can detect either influenza A or B virus without distinguishing the type, influenza A virus only or detect both influenza A and B and identify the type. However non of these rapid tests can differentiate between subtypes that infect humans (H1N1 and H3N2) or avian influenza subtypes (FDA, 2005). A list of rapid tests that are available can be obtained from the following link: http://www.cdc.gov/flu/professionals/labdiagnosis.htm.

Reverse transcription polymerase chain reaction (RT-PCR)

RT-PCR is a process whereby RNA is first converted to complementary DNA (cDNA) and a section of the genome is then amplified through the use of primers that bind specifically to this target area. This allows for exponential amplification of small amounts of nucleic acid, through the action of a thermo stable DNA polymerase enzyme, which enables highly sensitive detection of minute amounts of viral genome.

Not only does RT-PCR have superior sensitivity (Steininger 2002) but it can also be used to differentiate between subtypes and conduct phylogenetic analysis (Allwinn 2002). RNA degradation of archival samples can decrease the sensitivity of RT-