Summary

Avian H5N1 influenza is currently of great international concern. To reduce the human morbidity and mortality from this infection, doctors in both the community and hospital need to have an enhanced awareness for this infection, so that it can be identified and treated as early as possible. Patients with respiratory symptoms should be screened with a careful clinical history before sending the appropriate clinical samples to the diagnostic laboratory for urgent testing. Diagnostic tests include slide immunofluorescence, viral culture, serology and molecular detection of viral nucleic acid. Nasopharyngeal aspirate samples provide a good diagnostic yield for respiratory viruses in general. However, a throat swab may be more feasible in the family physician setting. On direct comparison of the two methods, the difference in their diagnostic yield for H5N1 is still uncertain. Modern diagnostic virology laboratories can produce accurate, reliable results within one working day, provided that a good quality sample which is correctly labelled, stored, transported, arrives in good time for all the diagnostic testing to be performed.

摘要

H5N1型禽流感備受國際關注。為減低人類對此的患病率和死亡率，社區及駐院醫生必須保持警戒和加強對它的認識，使患者及早得到確診與治療。患有呼吸道感染徵狀的病人應先經過詳細的臨床病歷篩選，再將適當的臨床樣本送往化驗室作緊急測試。診斷測試包括載片免疫熒光檢驗、病毒培養，血清檢驗和病毒核酸分子檢測。一般而言，從鼻咽抽吸樣本檢測呼吸道病毒的診斷成果良好，但家庭醫生則以喉頭拭子採集樣本更為可行，兩種診斷方法成效的分別尚未確定。現代的病毒學診斷化驗室可為臨床標本進行全面測試，並於一個工作天內提供準確和可靠的結果。

---

Introduction

Since the appearance of avian H5N1 influenza infections in humans in 1997, the possibility of this strain becoming the pathogen resulting in the next pandemic influenza illness has become a big concern.^1,2 However, only sporadic human infections have been diagnosed so far and the virus itself has not yet shown any significant mutation that has enabled efficient human-to-human transmission.³ Risk factors for acquiring H5N1 influenza virus infection have been well-documented.⁴ (see Box 1).

In Hong Kong the Department of Health's Centre for Health Protection (CHP) has produced local notification criteria for doctors covering human infection with the three types of avian influenza that have previously infected man (H5, H7 and H9 avian influenzas).⁵ These have been reproduced in Box 2. In addition, specific triage criteria have also been produced by the Hospital Authority (HA). A checklist that can be used to assess patients suspected to be infected with H5N1 influenza is available.⁶ (This has been reproduced in Box 3, but readers are required to check the HA's website for regular updates on which countries are currently suffering from human or animal H5N1 influenza cases). The Hong Kong influenza pandemic plan is regularly updated and can be accessed from the Hong Kong HA website.⁷

Therefore, if a patient visiting a family physcian's clinic should fit both the clinical and epidemiological criteria, then as well as reporting the case to CHP, arrangement for laboratory testing for H5N1 influenza should be made. The family physician should also refer the patient to the local hospital for isolation and appropriate management.

Possible typical clinical scenario presenting to a family physician

A 45-year old man attended a family physician in his local shopping centre, with a two-day history of fever, headache and cough. After telling his symptoms to the reception staff, he was given a face mask to wear, whilst he waited in the waiting area. The reception staff and the other patients had also put on face masks. The on-duty family physician was alerted immediately of this patient's symptoms so that the patient was immediately seen to reduce the time he would spend waiting (and possibly coughing) in the waiting area. Being aware of the risk factors for avian influenza, the family physician asked about the man's recent travel history, exposure to birds or chickens, or other sick people.

The man replied that he had recently returned from visiting relatives and friends in Guangzhou three days earlier, where some of them kept live chicken in their backyard. Further questioning revealed that some of these chicken had died during his visit and he had helped his friend dispose of them into a pit where they were buried. During this procedure, he said that they had worn kitchen gloves, but no masks. None of the other family members had fallen sick during the rest of his stay there, as far as he knew. All his immediate family were well, though they did not travel with him to Guangzhou.

The family physician referred the patient immediately to the local Accident and Emergency (A & E) Department at the Prince of Wales Hospital (PWH), and warned the staff there of this man's high-risk history for avian H5N1 influenza. Upon arrival, the patient was classified as high-risk for possible H5N1 influenza infection and admitted immediately to a negative pressure isolation ward at PWH by the Infectious Disease team. Samples were taken by the A & E staff and were sent for laboratory testing for respiratory viruses, including H5N1 influenza.

Clinical samples for avian H5N1 influenza testing

Once in a hospital isolation room, further clinical samples can be taken from the patient for diagnostic testing. There is emerging evidence that lower respiratory tract samples, e.g. tracheal aspirates (TA)s and bronchoalveolar lavages (BAL)s are more sensitive for the diagnosis of avian H5N1 influenza. However, in the hospital settings, a nasopharyngeal aspirate (NPA) sample is preferred as it also provides a good diagnostic yield for other common respiratory viruses. The application of viral antigen-based rapid detection method can be used for any NPA samples which can cover common respiratory viruses including influenza A and B, adenovirus, respiratory syncytial virus and parainfluenza virus. During the severe acute respiratory syndrome (SARS) outbreaks of 2003, it was found that taking NPAs was a potentially dangerous procedure for healthcare workers (HCW). Family physicians are well-advised to refer any suspected patient immediately to the local hospital with the facilities for this procedure to be carried out safely (Figure 1). Nasal or throat swabs are less sensitive than NPA for detecting respiratory viruses in general. Due to the lack of experience in laboratory diagnosis of human H5N1 infection, and the uncertainty on cell tropism of this virus, one should be aware of the latest information that is available on this subject.

An NPA is taken directly into a plain (empty) sterile tube, whereas swabs are stirred into a small glass vial of viral transport medium (VTM, Figure 2) then left within, by breaking off the shaft. If the specimens are not transported to the diagnostic laboratory within a few hours, they need to be refridgerated at 4^oC, otherwise any virus in the specimen may lose viability and/or any viral proteins may become denatured. The VTM contains nutrients to keep the virus alive and antibiotics and antifungal drugs to prevent bacterial and fungal overgrowth, which may be toxic to tissue culture during viral culture.

An acute serum sample may also be taken with the intention of a follow-up convalescent serum to be taken 10-14 days later, to be tested for rising titres of viral and other antibodies. This may be useful if for some reason, the diagnostic tests on the NPA or swabs reveal no virus, which may happen if the samples have not been taken or stored optimally, or there are delays in transportation allowing degradation of the virus or its proteins, or if the samples leak, are mislabelled, have the wrong test put up or even lost. These events are unfortunate, but can happen.

With H5N1 influenza, recent studies have demonstrated that the virus or its nucleic acid (RNA) can be detected in the blood.⁸ However, few diagnostic laboratories will offer this because a standard test and assays to test for the virus in blood tend to be on a research basis, and will be requested as required on a case-by-case basis.

If a bronchoscopy is performed to obtain a BAL, or the patient requires intubation to produce a TA, both of these specimens can be treated in the same way as an NPA. These specimens may have a higher yield (be more sensitive) for the detection of H5N1 influenza virus. Recent research is available which suggests that avian H5N1 influenza infection preferentially infects the lower respiratory tract in humans.^9,10

Avian H5N1 nucleic acid has been found in the intestine of infected humans,¹¹ so there is some evidence to suggest that H5N1 influenza virus may be shed in stool. Again, just as for detection of the virus in blood, this may be a non-standard diagnostic sample so that such requests should be discussed with the diagnostic laboratory beforehand.

Laboratory tests

In the same way that patients with suspected avian H5N1 influenza need to be specially isolated, some testing procedures must only be performed in Physical Containment Level 3 laboratories.

Viral culture for H5N1 influenza

The NPA and swab samples are inoculated onto a monolayer of Madin-Darby Canine Kidney-derived cells, grown in test tubes. This cell line has been chosen as it has been found to be optimal for the growth of influenza viruses. Other cell lines, such as avian cell lines, may be more optimal for avian H5N1 influenza, but these are not routinely used in clinical diagnostic laboratories. Following inoculation, the cell monolayer is bathed in growth media and incubated at 33-34^oC (Figures 3a and 3b), with observation for a viral cytopathic effect (CPE) every two days (Figure 4). After 5-7 days, haemadsorption and haemagglutination testing can be used to detect the presence of growing influenza virus. This involves first removing the growth media, then covering the MDCK monolayer with red blood cells (RBCs) from an animal (e.g. guinea-pig RBCs). Influenza virus has a haemagglutinin (HA) surface protein which will bind to the RBCs. If the virus is present and infecting the MDCK cells, the viral HA proteins will be produced on the surface of the infected cells and bind the RBCs to the cell monolayer. This can be examined through a low-power light microscope. Once infection is confirmed, immunofluorescence (IF) testing can be used to confirm whether it is influenza type A or type B.

Immunofluorescence testing for H5N1 influenza

IF testing for influenza can be used directly on the clinical specimens if they are NPA, BAL or ETA. The IF testing of clinical specimens is quick and can provide a result within about 3 hours from the receipt of the specimen in the laboratory. The target of the IF test is usually one of the viral proteins. A specific monoclonal antibody is raised by injecting this viral protein into a small mammal (usually a rabbit, guinea-pig or mouse) then collecting and purifying the specific antibody, to which a chemical marker is attached. This is usually, but not exclusively, a fluorescent marker. Therefore, any viral protein in the specimen on the slide is bound by this antibody and its fluorescent marker will fluoresce using ultraviolet (UV) light microscopy (Figure 5). The use of positive and negative controls will assist the interpretation of the result. However, it is very dependent on the amount of virus protein in the specimen, which in turn depends on the stage of the patient's illness when the specimen was taken, how well the specimen was taken, whether there was suboptimal storage of the specimen or any delays in transportation to the laboratory.

While IF provides a rapid and sensitive means for diagnosis of influenza in general, its performance on avian influenza H5N1 seems to be less optimal. This may be due to a lesser amount of intact infected cells contained in the NPA specimen collected from H5N1-infected humans. In addition, non-H5 influenza IF tests usually target an internal viral protein (the nucleoprotein), whereas H5-specific IF tests target the H5 envelope protein, which may account for the difference in their diagnostic performance.

Immunochromatographic testing for H5N1

A number of rapid diagnostic kits are available in the market for the laboratory diagnosis of influenza types A and B. These kits are non-subtype specific, i.e. they cannot differentiate usual human influenza A H1N1/ H3N2 from avian influenza A H5N1. The performance of these test kits varies quite substantially. At present, these test kits are generally regarded as not sensitive enough for diagnosis of influenza A H5N1.

Molecular detection of H5N1 influenza

Normally, molecular tests for H5N1 influenza means reverse-transcription polymerase chain reaction (RT-PCR). This will detect the presence of the viral nucleic acid (RNA) in the sample. This does not necessarily mean that the virus is viable (i.e. is still able to infect cells). There are several commercial kits as well as previously published in-house methods^2,12-14 available for the RT-PCR detection of influenza viruses, including H5N1 influenza. However, the principles are generally the same.

The clinical sample can be tested directly, or cultured first and tested for H5N1 RNA. This test is very sensitive, as it will amplify the viral RNA, so theoretically, even if there is just one copy of the viral RNA present in a specimen, it will be amplified to a detectable level. However, due to various technical limitations, the sensitivity limit is more like tens to hundreds of RNA copies per sample, rather than single copies. The initial step is to extract the viral RNA from the clinical sample or culture, which is also performed using commercial RNA extraction kits. Following this, the target viral RNA is first reverse-transcribed to DNA then amplified by using specifically designed 'primers'. These are short lengths of nucleotides which are designed to bind to specific sequences that are unique to H5N1 influenza. After many cycles of amplification of this target viral sequence, the presence of any viral RNA can be detected by visualizing the presence of this amplified target DNA using gel electrophoresis. Nowadays, real-time RT-PCR is available and may give a result more quickly as the detection occurs at the same time as the RNA amplification, so there is no need for a further detection step such as the gel electrophoresis step described above. Positive and negative controls are essential to ensure that any result is genuine and not the product of laboratory cross-contamination or technical failure.

Typing and subtyping of influenza H5N1

Generally, typing (into influenza A or B) can be performed in most diagnostic laboratories. However, sub-typing (into H3N2 or H1N1, etc.) and strain-typing (into influenza A/Sydney/5/97 or A/Fujian/411/02) can usually only be performed in reference laboratories. Typing into influenza A or B can be performed using commercial specific IF stains with fluorescently-labelled monoclonal antibodies against a specific viral protein. Sub-typing and strain-typing is normally performed using the haemagglutination inhibition testing (HAI), which uses specific antisera raised in small mammals (ferret), to inhibit the agglutinating effect of the HA protein (on RBCs) of specific viral strains. Nowadays, molecular testing with subtype- or strain-specific primers can be used, which may be quicker than HAI, as long as these specific primers are available (see above).

Conclusion

Hence, the laboratory diagnosis of H5N1 influenza is essentially the same as for conventional human influenza viruses, except that the safety precautions are higher (due to the higher mortality of human H5N1 infections), and there may be a higher sensitivity using specific avian H5N1 influenza molecular versus conventional IF tests on clinical specimens. >From the family practitioner's viewpoint, the most important aspect of human H5N1 influenza diagnosis is to first consider the diagnosis, then to extract a detailed, relevant risk history, so that the patient may be referred for the recommended isolation and safe collection and testing of their clinical specimens. In addition, such suspected cases need to be dealt with quickly so as not to expose other patients in the waiting area to possible H5N1 infection. Laboratory diagnosis of H5N1 influenza (or its exclusion) can be achieved within one working day, if the appropriate clinical samples are taken and transported carefully to the laboratory.

In the unlikely event that a family practitioner needs to take clinical samples from the patient in the clinic, a throat swab placed in a VTM is safer for the doctor or nurse than attempting to take an NPA sample. In such a situation, it is advisable to call the laboratory for advice and to warn them that a suspected H5N1 sample will be arriving from your clinic, with the details of the patient. It is recommended that the sample should be labelled for the attention of a specific microbiologist (preferably the one with whom the case is discussed), so that the sample is not misdirected or lost, which will delay any possible diagnosis of H5N1 influenza.

Acknowledgements

The authors would like to thank the Infection Control Team of the Prince of Wales Hospital, for the kind use of the photograph in Figure 1.

Key messages

1. Human cases of avian H5N1 influenza may present to your clinic like any other respiratory infections.
2. A clinical suspicion of H5N1 influenza based on the patient's contact history should be followed by immediate respiratory protection measures and referral to a local hospital for further isolation, testing and management.
3. Laboratory testing for human H5N1 influenza is similar to testing for other human influenza viruses, in terms of the clinical samples to be collected, and the testing procedures to be performed. However, due to the higher mortality resulting from infection with this virus, higher containment facilities are required when viral culture is attempted.
4. Molecular testing is currently the method of choice for diagnosing H5N1 influenza.
5. Before sending a sample for H5N1 testing, it is advisable to contact the laboratory and refer the sample to a named microbiologist, to ensure safe receipt and efficient processing of the sample. In general, if a sample is received in the morning, a result on its influenza H5 status can be achieved by the afternoon of the same day.

---

---

References