Review of Rapid Diagnostic Tests for Influenza

Clinical and Applied Immunology Reviews 4 (2003) 151–172 Review of Rapid Diagnostic Tests for Influenza Patrick J Gavin, Richard B Thomson, Jr.* Divi...
Author: Brice Rose
0 downloads 2 Views 186KB Size
Clinical and Applied Immunology Reviews 4 (2003) 151–172

Review of Rapid Diagnostic Tests for Influenza Patrick J Gavin, Richard B Thomson, Jr.* Division of Microbiology, Department of Pathology, Evanston Northwestern Healthcare, 2650 Ridge Avenue, Rm. 1936. Evanston, IL 60201. USA Received 20 March 2003; received in revised form 25 June 2003; accepted 26 June 2003.

Abstract Influenza is unique among viral infections because of its propensity for seasonal epidemics and occasional pandemics, and because of the morbidity and mortality that result from its pulmonary complications. In contrast to the majority of viruses, effective well-tolerated influenza vaccines, antivirals and chemoprophylaxis are available. The need for a timely diagnosis, which allows for optimal use of these treatments, led to the introduction of numerous rapid diagnostic tests (with turnaround times of less than 30 minutes). However, during influenza season, clinical diagnosis (based on cough and high fever of acute onset) can be highly predictive of influenza. Thus diagnostic tests are not required for all patients with suspected influenza but may be of value if the clinical diagnosis is unclear and if antiviral or antibiotic treatment is a consideration. When evaluating performance of various rapid diagnostic tests for influenza, it is important to consider the type and quality of specimen and type of patient to be tested. Specimen-type drives performance of the rapid diagnostic tests. Swab specimens, particularly throat swabs are the most frequently submitted but least desirable specimen-type. Thus, although current rapid diagnostic tests are specific for influenza, sensitivity is highly variable. To improve diagnostic accuracy, a nasal/nasopharyngeal aspirate or sputum specimen should be obtained. Because of their highly variable sensitivity and negative predictive value, it is our opinion, that rapid diagnostic tests should only be used in influenza season and that results should be confirmed with virus culture. Despite these reservations, during influenzaseason, detection of influenza by rapid diagnostic test may, potentially, be of great benefit to the patient and public health. 쑕 2004 Elsevier Inc. All rights reserved. Keywords: Influenza; Rapid Diagnosis

Abbreviations: CLIA, Committee for Laboratory Improvement Act; DFA, direct fluorescent antibody; EIA, enzyme immunoassay; HA, hemagglutinin; MDCK, Madin-Darby canine kidney; NPV, negative predictive value; NA, neuraminidase; OIA, optical immunoassay; PPV, positive predictive value; PMK, primary rhesus monkey; RT-PCR, reverse transcriptase polymerase chain reaction. * Corresponding author: Tel.: ⫹847-570-2754; Fax: ⫹847-733-5314. E-mail address: [email protected] (Richard B Thomson, Jr.). 1529-1049/04/$ – see front matter 쑕 2004 Elsevier Inc. All rights reserved. doi: 10.1016/S1529-1049(03)00064-3

152

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

1. Introduction The most important features of influenza infection are its propensity to occur in seasonal epidemics and occasional pandemics, and the mortality that results from its pulmonary complications [1]. In contrast to the majority of virus infections, effective well tolerated antiviral treatment and chemoprophylaxis is available for influenza infection. For maximal efficacy such treatment must be started within 36 to 48 hours of onset of symptoms. However, results of conventional diagnostic tests for influenza are generally not available within this narrow therapeutic window. The need for timely diagnosis that provides results in a clinically relevant timeframe led to the introduction of rapid diagnostic tests for detection of influenza viruses. For the purposes of this review, the term ‘rapid diagnostic test’ refers to those tests with turnaround times of less than approximately 30 minutes. Anti-viral treatment for influenza is ineffective against other respiratory viruses that may cause ‘influenza-like’ illness. Thus, if indiscriminate and inappropriate use of antiviral treatment is to be prevented, in addition to rapid turnaround times, such tests should have high sensitivity and specificity. In this article we review the available rapid diagnostic tests for influenza virus, and their role in comparison to clinical diagnosis or other laboratory methods for influenza detection. We also briefly review the virology, epidemiology and treatment of influenza as it pertains to rapid diagnostic testing. And finally, we present a rational approach that incorporates rapid testing in the laboratory diagnosis of influenza.

2. Influenza 2.1. Definition The term influenza refers to illness caused by influenza viruses. Influenza is a highly contagious, acute viral respiratory disease of global importance. However, many respiratory-viruses and bacteria can present with ‘influenza -like’ symptoms, including: adenovirus; rhinovirus; parainfluenza viruses; picornaviruses; respiratory syncytial viruses; chlamydia; legionella and mycoplasma. Thus, infections caused by other respiratory pathogens may occasionally be difficult to distinguish from influenza on the basis of clinical features alone. 2.2. Clinical presentation The spectrum of illness caused by influenza can range from asymptomatic to fatal. Most infections cause an acute febrile respiratory illness that resolves without complications. After a short incubation period (one to four days), influenza typically presents with the abrupt onset (⬍12 hours) of fever and chills, severe malaise, myalgias, headache, sore throat, and dry cough [1, 2]. Illness typically resolves after one to five days, although cough and malaise may occasionally persist for weeks. Infants and young children may also present with a nonspecific febrile, respiratory or gastrointestinal illness, febrile seizures or sepsis-syndrome [2, 3]. The major potentially fatal complications are primary viral or secondary bacterial pneumonia and exacerbations of underlying respiratory or cardiac conditions.

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

153

2.3. Epidemiology Influenza affects all age groups and, unlike most viruses, can recur in any individual. Rates of infection are highest in schoolchildren, but rates of serious illness and death are highest in the elderly and patients with underlying medical conditions [4]. Influenza epidemics are characteristically abrupt (peak within two to three weeks), spread rapidly throughout the population with high attack rates, followed by decline in activity five to six weeks later [1, 2]. During any given year, influenza epidemics kill 500,000 to 1 million people worldwide [5]. In the United States, seasonal epidemics of influenza occur during the winter months (December to March), infect between 25% and 50% of the population and cause an average of 21,000 - 36,000 deaths per annum [4, 6, 7]. In the last two decades, influenza-associated deaths, 90% of which occur in the elderly, have increased substantially and will continue to increase as the population continues to age [7]. Although most infections are self-limited, influenza viruses are the most common cause of acute respiratory illness requiring medical attention [2]. In the United States alone, influenza is responsible for an average of 114,000 hospitalizations/year [8]. The estimated actual cost of influenza in the United States is of the order of $3 to $5 billion per annum (in 1987 dollars) [9]. In addition to recurrent annual epidemics, influenza virus A has the unique ability to cause unpredictable pandemics - a worldwide global epidemic. Over the past hundred years, there have been five great pandemics: 1890, 1900, 1918, 1957 and 1968. The 1918 - 1919 “Spanish influenza” pandemic resulted in an estimated 20 to 40 million deaths worldwide (over 500000 in the United States). It is very likely that pandemic influenza will return. 2.4. Virology The three influenza viruses - influenza virus A, B and C - are enveloped RNA viruses of the Orthomyxovirus family. Influenza viruses A and B cause respiratory disease in humans and animals. Influenza virus A causes the most clinically significant disease and is the only one to cause pandemics. Influenza virus C is a mild pediatric respiratory pathogen that only occasionally affects adults. Each influenza A virion is surrounded by a lipid membrane derived from the host-cell and has a segmented genome composed of eight RNA molecules. Two surface glycoproteins (hemagglutinin (HA) and neuraminidase (NA)) and the M2 protein protrude through this lipid envelop. Influenza virus binds to its sialic (neuraminic) acid receptor on respiratory epithelial cells by means of the HA protein. Antibody to HA neutralizes influenza virus infectivity. The HA protein is so named because it binds to certain sugars on the surface of red blood cells, causing them to clump together (hemagglutination). Neuraminidase is an enzyme that cleaves sialic (neuraminic) acid from the surface of cells. This process is responsible for release of influenza virus from infected cells and facilitates virus spread within the respiratory tract. As such, NA is essential for replication of both influenza A and B virus. The M2 membrane protein is essential for uncoating of influenza A virus after cell entry. Two types of antigenic variation in HA and NA are responsible for the unique ability of influenza A virus to cause both recurrent seasonal epidemics and occasional more serious pandemics. Antigenic drift, refers to the continual process of minor changes in the amino

154

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

acid sequence of HA and NA caused by point mutations in the HA and NA genes. Antigenic drift is responsible for annual influenza epidemics. It is the ability of influenza virus to undergo continuous and progressive antigenic change that dictates influenza vaccine updates annually to ensure inclusion of antigens of the most current strains. In contrast, antigenic shift, which only occurs in influenza A viruses, refers to an occasional sudden and major change in virus antigenicity that produces a novel HA or NA (or both) distinct from those of influenza viruses circulating in recent years. Pandemic influenza strains contain new HA or NA genes found on novel RNA genome segments derived from animal or avian influenza A viruses that either adapt to human hosts or reassort with circulating human influenza A viruses [2, 10, 11]. Replication and transmission of such novel influenza viruses in an immunologically susceptible population may give rise to a new pandemic. 2.5. Pathogenesis Uncomplicated influenza is characterized by degeneration of respiratory epithelial cells with loss of cilia and desquamation [12]. The pathological effects are due to virus-induced damage to respiratory epithelial cells combined with damage from the resulting immune response. Infection of the upper and lower respiratory tract accounts for much of the symptomatology of influenza, particularly cough and tracheal irritation. Influenza virus is usually no longer detectable in respiratory secretions after five to ten days [1, 12]. However, pathological changes, pulmonary function abnormalities, and respiratory symptoms may persist for much longer [12]. Thus a negative diagnostic test does not outrule influenza as a cause of continuing respiratory symptoms, especially if testing is performed late in the clinical course.

3. Treatment Acute respiratory viral illnesses are the most common cause of medical visits for outpatients of all ages. Because there is no available treatment, establishing a specific viral cause is neither necessary nor cost-effective for most of these illnesses. In contrast, specific antiviral treatment for influenza has been available for many years. However, prevention remains the best treatment for influenza. Influenza vaccination remains the primary method for preventing influenza and its severest complications. When the vaccine and circulating influenza viruses are antigenically similar, vaccine prevents approximately 70% to 90% of influenza infections in healthy adults aged less than 65 years [13]. In addition, currently, there are four approved drugs for influenza treatment: the adamantanes, amantadine and rimantadine; and a newer class of neuraminidase inhibitors, zanamivir and olseltamivir [14]. Amantadine and rimantadine target the M2 membrane protein of influenza A virus and inhibit viral replication. Both drugs have therapeutic and prophylactic benefit against infection caused by influenza A virus. They reduce the severity and duration of infection, if given within 48 hours of the onset of symptoms, and are 70% to 90% effective in preventing illness caused by influenza A virus if given prophylactically during epidemics or outbreaks [2]. However, both are ineffective against infection caused by influenza B virus, and their use is limited by adverse effects (neurological and gastrointestinal), mainly with amantadine, and by the development and transmission of drug-resistant virus.

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

155

Zanamivir and oseltamivir are specific inhibitors of influenza virus NA. Inhibition of NA prevents spread of influenza within the respiratory tract and stops the infection. The development of zanamivir and oseltamivir represents a significant advance in the treatment and prevention of influenza. Both, inhaled zanamivir and oral oseltamivir, are effective for treatment and prophylaxis of influenza A and B infection. Both are well tolerated and show limited potential for emergence of antiviral drug resistance [2, 15–17]. In healthy adults with influenza, treatment with NA inhibitors, if begun early (within 30 to 36 hours of onset of symptoms, respectively), reduces time to recovery (20% to 40%), severity of symptoms (by 40%) and the time to return to normal activities (by one to two days) [18–20]. Efficacy is greater in those patients in whom treatment is started earlier and in those with more severe symptoms. In contrast, no beneficial effect is seen in patients in whom treatment is started later [20]. In addition, NA inhibitors are effective as chemoprophylaxis against influenza infection. In healthy adults during influenza-season, once daily zanamivir or oseltamivir reduces the risk of influenza infection by 31% to 50%, of influenza illness by 84% to 76% and of culture proven influenza by 67% to 100%, respectively [17, 21]. Furthermore, chemoprophylaxis of families of influenza cases with inhaled zanamivir is 79% protective against secondary household transmission (reduced from 19 % to 4%) [15]. 4. Diagnosis 4.1. Clinical diagnosis of influenza Although the clinical presentation of influenza is similar to illness caused by other respiratory pathogens, when influenza is circulating in the community, several studies show that the presence of cough and high fever of acute onset is likely to be associated with influenza infection [22–25]. In a retrospective study of 3744 unvaccinated adults and adolescents with influenza-like symptoms, who were enrolled in a clinical trial of an anti-influenza drug, patients presenting with cough and fever (⬎37.8ºC) were likely to have laboratoryconfirmed influenza (positive predictive value (PPV), 79%) [22]. The probability of a patient having confirmed influenza increased further with increasing fever and if patients presented acutely (within 36 to 48 hours of onset). Similarly, in an additional study of general practitioners in three different clinics, the PPV of the presence of fever (⬎38ºC) and cough for laboratory confirmed influenza was 86.8%. However, the clinical diagnosis of influenza was imperfect (negative predictive value (NPV), 39.3%)[23]. Factors that may potentially confound the clinical diagnosis include: the circulating virus strain, preceding use of antipyretics; duration of symptoms at presentation, and the patient’s age, vaccination status and any underlying illnesses. It is important to stress that these studies were conducted during periods when laboratory confirmed influenza cases were documented in the community. In the absence of a known influenza epidemic, the PPV of the same clinical criteria was only 44% [25]. This highlights the importance of the season in which the patient presents and of notification of health care providers of the results of active influenza surveillance in the community. Indeed, during an influenza epidemic the intuition of general practitioners for the presence of influenza infection is one of the most accurate diagnostic tools. Specifically, when influenza is known

156

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

to be circulating in the community, general practitioners were able to diagnose laboratory confirmed influenza (PPV, 76%, and NPV, 75%) more accurately using clinical judgement alone than by using defined diagnostic criteria and guidelines [26]. In a decision analysis model of empirical anti-influenza treatment versus treatment based on a positive rapid test, during influenza epidemic seasons, treatment based on clinical diagnosis was favored over treatment based on results of rapid tests [27]. This suggests that influenza diagnostic tests are not required for all patients with suspected infection. Rather, tests may be of most value to the clinician in patients in whom the diagnosis remains unclear and antiviral or antibiotic treatment is a consideration. The relatively low sensitivity and PPV of rapid diagnostic tests suggests that empiric treatment is the most effective and cost-effective strategy when the probability of influenza is greater than 45% [28]. 4.2. Laboratory diagnosis of influenza Given the potential accuracy of clinical diagnosis of influenza during influenza season, one might question the role of the laboratory in influenza diagnosis. In contrast to other ‘flulike illnesses’, influenza is associated with higher morbidity and mortality, is potentially preventable by vaccination and chemoprophylaxis and is treatable by specific antivirals. In addition, it is important for both the patient’s and public health perspective to differentiate influenza from ‘flu-like illness’ caused by other respiratory pathogens. Laboratory isolation of influenza provides culture confirmation of the arrival of influenza in a community. This is valuable information for the physician as it greatly enhances the accuracy of their clinical diagnosis. Physicians in close communication with the microbiology laboratory and aware of the arrival of influenza season should be able to diagnose most influenza infections clinically. This should reduce unnecessary additional laboratory testing and inappropriate antimicrobial treatment. Although classic cell culture of influenza virus rarely provides results quickly enough to impact treatment, it does provide important confirmation of clinical or rapid diagnoses. Thus, virus culture provides confirmation of results of rapid diagnostic tests, where low specificity and NPV may be problematic. Most authorities and many test manufacturers now advise that negative results of rapid diagnostic tests in patients with a compatible clinical illness be confirmed by cell culture. False-negative results of rapid diagnostic tests may have serious repercussions if potentially effective anti-influenza therapy, chemoprophylaxis or vaccination is withheld from individual patients or their contacts (at home, in hospitals or nursing homes). Similarly, laboratory confirmation of an index case of influenza has important implications for infection control in these settings. From the wider public health perspective, laboratory isolation and characterization of influenza viruses that are circulating in the community remains the backbone of global surveillance for epidemic influenza and for emergence of novel potentially pandemic strains. Laboratory isolation and sub-typing of human and animal influenza isolates was instrumental in identification and control of the novel highly pathogenic human H5N1 influenza in Hong Kong, in 1997 [11, 29, 30]. Similarly, laboratory isolation of community influenza strains is crucial for vaccine design for the forthcoming influenza season. At six-monthly intervals, three circulating influenza strains are selected as targets for the new vaccine formulations, approximately 200 million doses of

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

157

which will be administered every year. In addition, circulating influenza strains that are isolated in the laboratory are used to monitor for antigenic similarity to strains in the current vaccine and for the possible development of antiviral drug resistance. 4.3. Diagnostic tests for influenza Diagnostic tests for influenza are of four main types: virus culture (conventional and shellvial); detection by serology; detection of virus antigen (by fluorescent antibody, enzyme/ optical immunoassay or NA enzymatic activity) and detection of virus nucleic acid (e.g., by polymerase chain reaction and gene array) (Table 1). 4.4. Virus Culture Conventional classic laboratory diagnosis of influenza is based on virus isolation and serologic testing. Virus culture using primary rhesus monkey (PMK) or Madin-Darby canine kidney (MDCK) cells is the currently accepted ‘gold-standard’ for the laboratory diagnosis of influenza virus. Cell cultures are examined for cytopathic effect and screened by hemadsorption with confirmation by immunofluorescent monoclonal antibody against influenza A or B. However, traditional virus isolation and identification takes time. Because virus culture results are generally available in four to five days (range, two to 14 days), the impact on patient care is limited. Nevertheless, continuing virus isolation is important from local and global public health perspective. Only culture isolates can monitor new circulating influenza subtypes and strains and provide data for vaccine efficacy and formulation for the coming year. In an effort to speed-up virus isolation, rapid culture techniques were developed where clinical specimen is centrifuged on to a monolayer of cells (commonly PMK or MDCK cells) on a coverslip contained in a shell-vial or multi-well plate and incubated for 24 to 48 hours. Influenza A or B antigen is then detected by staining with fluorescent monoclonal antibodies. Isolation of influenza by rapid shell-vial culture represents an improvement over conventional culture in terms of speed and simplicity [31–33]. Influenza shell-vial culture gives results one to three days after inoculation. Shell-vial culture is also less labor intensive for technologists because the requirement for regular examination of cells for cytopathic effect or hemadsorption is removed. However, to produce more rapid results, some sensitivity is sacrificed. A summary of several studies, demonstrates that after 24 and 48 hours incubation, shell-vial isolation followed by staining with monoclonal antibodies detects 60% to 84% of influenza A and B viruses detected by conventional cell culture [31, 32, 34]. Specificity of the shell-vial assay is 100%. Although the turnaround time of shell-vial culture is much shorter than conventional cell culture, in many cases, results are still not available in a timeframe that allows optimal treatment with newer antivirals. More recently, refinements to the shell-vial assay has permitted some clinical laboratories to replace conventional culture with shell-vial culture for the diagnosis of respiratory virus infections (including influenza) [35]. Shell-vial assays that use R-mix cells, a combination of mink lung cells and human adenocarcinoma cells (strains Mv1Lu and A549, respectively, Diagnostic Hybrids, Athens, OH), coupled with immunofluorescent staining demonstrates faster results and sensitivity that approaches that of conventional virus culture [35, 36]. Turn around time for shell-vial assay using R-mix cells is 1.4 days, compared to

158

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

Table 1 Diagnostic Laboratory Tests for Influenza

Method

Influenza Types Detected

Cell Culture

A and B

Shell-Vial Culture

A and B

Flourescent antibody

A and B

Advantages

Disadvantages

Turnaround Time

Sensitive and specific. Labor intensive. 3–14d Can detect viruses Results not other than influenza. available in Current gold clinically relevant standard. Vital for timeframe survillance and vaccine formulation Specific. Single More rapid than 1–3d specimen can be culture but still too tested for viruses slow to influence other than influenza treatment Specific. Same-day Relatively complex, 2–4h test. Single requires expertise, specimen can be additional reagents tested for multiple and equipment. pathogens. Can be Requires adequate performed on specimen culture or directly on clinical specimen 1–2h 1–2h

Directigen Flu A A Directigen Flu A and B A and B FLU OIA A or B (not Specific. Simple. Rapid Low sensitivity and 30min discriminated) turnaround time PPV Negative permits effective results require treatment and culture confirmation chemoprophylaxis. No monitoring of Positive results specimen quality useful in influenza season Quick Vue A or B (not 30min discriminated) Zstat Flu A or B (not 30min discriminated) Now FLU A A and B and B RT-PCR A, B and C Sensitive and rapid. May miss strains with 1–2d Can detect nonnovel HA or NA culturable virus Serology A, B and C Sensitive and specific. Purely retrospective 2–4wk Can detect culturediagnosis negative infection. Important research and surveillance tool Note: NA, not applicable; RT-PCR, real-time polymerase chain reaction. *Committee for Laboratory Improvement Act waived.

Point-ofRelative Care cost Test* $$

No

$$

No

$

No

No No $$-$$$

Yes

Yes Yes No $$$$

No

NA

No

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

159

a turn around time of 5.2 days for conventional virus culture [35]. In addition, at 24 hours post-inoculation, R-mix shell vial culture has a 100-fold greater sensitivity than conventional virus culture for influenza A and equivalent sensitivity for influenza B virus. Furthermore, whereas 100000 to one million infectious influenza A virus particles are necessary for a positive result by rapid antigen diagnostic tests, the lower limit of detection of R-mix shellvial cultures is as few as ten virus particles [36]. Even after only six-hours incubation, Rmix shell vial assay demonstrates greater sensitivity (74%) than many rapid diagnostic tests [37]. In practice, this would permit same-day preliminary reporting of influenza-positive results, with a final report after overnight incubation. In addition, R-mix shell vial assay is able to test for all major respiratory viruses simply by changing the monoclonal antibody used for detection, while cost is equivalent to or only marginally more expensive than conventional cell culture ($2.13 or 11% more per specimen) [35]. 4.5. Detection of virus antigen Diagnostic tests for detection of influenza virus antigen are of two main types: (i) Direct fluorescent antibody tests and (ii) Rapid enzyme/optical immunoassays or assay for NA enzymatic activity. (i) Direct Fluorescent antibody tests Direct Fluorescent antibody staining (DFA) of clinical specimens using monoclonal antibodies against influenza virus antigen is a reliable and relatively rapid technique for the detection of influenza. In many laboratories, DFA testing is offered as a ‘sameday’ test. If the specimen is received in the laboratory by early morning, results will be available before five PM that day. However, studies of DFA detection of influenza virus have shown highly variable results with sensitivities ranging from 40% to 100%. These disparities are probably due to differences in technical expertise, specimen type, and patient selection. In several studies, in comparison to conventional cell culture, DFA has been highly accurate in detecting influenza virus in respiratory specimens: sensitivity, 70% to 100%; specificity, 80% to 100%; PPV, 85% to 94%; and NPV, 96% to 100% [38–42]. A number of studies have compared DFA and newer rapid enzyme/optical immunoassays (EIA/OIA) or NA enzymatic assays for detection of influenza. In two such studies, performance of DFA and rapid antigen tests, in adult and geriatric populations, were comparable for detection of influenza A (sensitivity, 92.5% to 100% versus 86.8% to 100%; and specificity, 97.2% to 100% versus 99% to 100%, respectively) [41, 42]. Although the rapid EIA for influenza B was more sensitive than DFA (87.5% versus 65.5%, respectively), EIA was associated with more false-positive results [41]. In studies where EIA has demonstrated superior performance to DFA for detection of influenza (sensitivities of 70% versus 60% to 65%, respectively), specimen type (nasal aspirates/wash) and population (pediatric) were optimal for the EIA test [43, 44]. In addition, in one study, performance of DFA test may have been compromised by lack of exclusion criteria for specimens with inadequate cellularity [43]. As with rapid antigen detection tests, the accuracy of DFA testing is heavily dependent on specimen quality: lack of adequate numbers of respiratory epithelial cells in the specimen and

160

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

non-specific fluorescence of cells and debris can produce false-negative and falsepositive results, respectively. In comparison to newer rapid detection tests, DFA requires additional equipment and reagents (cytocentrifuge, fluorescence microscope and monoclonal antibodies), is relatively labor intensive, more complex and technically demanding to perform and interpret, and has slower turnaround time (four to six hours). However, DFA can be performed in batches and can detect different respiratory viruses by using panels of monoclonal antibodies. Because DFA testing is more labor intensive and technically demanding, in the clinical microbiology laboratory, it is generally only performed during the day-shift by experienced personnel. In our hands, in the absence of a molecular assay, we consider DFA the best test after viral culture for detection of influenza viruses. (ii) Rapid Enzyme and Optical Immunoassays for Influenza Antigen. At present, six rapid diagnostic tests for influenza are in general use: (Directigen Flu A, and Flu A plus B tests (Becton Dickinson, Franklin Lakes, NJ.), Binax Now Flu A and B tests (Binax Inc., Portland, ME.) Biostar FLU OIA (Biostar, Inc., Boulder, CO), Quidel Quick vue (Quidel, San Diego, CA.), and the ZstatFlu test (Zyme Tx, Inc., Oklahoma City, OK.). All use an EIA or OIA directed at influenza virus antigens, with the exception of the ZstatFlu test, which detects influenza virus NA enzyme activity. They are designed as rapid tests to detect influenza virus within approximately 30 minutes of the time of inoculation. All cost between $15 and $20 [45]. At present, three rapid diagnostic tests (Biostar FLU OIA, Quidel Quick vue and ZstatFlu) are Committee for Laboratory Improvement Act (CLIA) waived for use as ‘point-ofcare’ tests. When evaluating the relative performance of the rapid diagnostic tests it is important to consider a number of factors: specimen-related (type, quality and transport of specimen), patient-related (age and immune status of patient); and influenza-type. For rapid diagnostic tests, specimen- type drives test performance (Table 2). Although various respiratory specimens are considered by manufacturers as suitable for rapid testing, not all specimen types yield equivalent results. Sputum and nasal aspirates or washes are the best specimen types for detection of influenza virus by virus culture or antigen detection, whereas swab specimens (particularly throat swabs) are least sensitive and specific [46–51]. Specifically, isolation rates for influenza virus are approximately 18% to 28% lower in nasal swabs and 39% to 43% lower in throat swabs or gargles than from nasopharyngeal secretions [47, 48]. This probably reflects the higher virus loads found in nasopharyngeal washes compared to nasal or throat swabs (1.8 log10 and 3 log10 higher, respectively) [41]. Sputum specimens provide a similarly high yield of influenza virus [52]. In a study of 92 patients with culture-proven infection, influenza was identified in 90% of sputum and 80% of nasal aspirate specimens but in only 64.6% and 51.8% of nasopharyngeal or throat swab specimens, respectively. In essence, in this study, the rapid diagnosis of influenza in throat swab specimens was little better than coin-toss. In addition to specimen-type, the rapid detection of influenza is heavily dependent on specimen quality. Similar to DFA tests, in several studies, performance of rapid diagnostic tests is dependent on an adequate specimen [41, 44]. Specifically, the Zstat Flu test was

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

161

Table 2 Rapid Diagnostic Tests for Influenza

Name (reference) Directigen Flu A [42–44,53,62,63,67] Directigen Flu A and B [41,64] FLU OIA [46,50,54,56,65,67] Quick Vue [66,67]

Nasal/NP Wash

Throat/Nasal/NP Swab

Influenza Types Detected

Assay Method

A

EIA

75–100% 91–100% 62–100% 89–100% 39–67%

A and B

EIA

75–87%

93–97%

74–80% 93–98%

88%

69–79%

73–91% 56–77% 25–83%

73%

95–99%

65%

92%

70–96%

77–92%

89–98% 59–76% 62%

99%

82%

94%

71%

92–97%

Sensitivity Specificity PPV

A or B (not OIA discriminated) A or B (not OIA discriminated) Zstat Flu A or B (not NA [43,44,53,62,64,67] discriminated) Enzyme Activity Now FLU A and A and B OIA B [51]

NPV

97%

Sensitivity Specificity PPV NPV

58–78%

92%

51–94%

73% 56%

Note: EIA, enzyme immunoassay; NA, neuraminidase; NP, nasopharyngeal; NPV, negative predictive value; PPV, positive predictive value; OIA, optical immunoassay.

20% less sensitive (77% versus 57%) in detecting influenza in nasal wash specimens with low cellularity (i.e., those considered inadequate for DFA testing) [44]. Performance of the Directigen Flu A test also appears to be related to the amount of cell-associated rather than free virus in the clinical specimen [53]. Specimen transport also influences accuracy of rapid diagnostic influenza tests. After collection, sputum or nasal aspirate specimens should be transported in sterile containers and nasal or throat swabs in viral transport medium. Although, manufacturers do not recommend specimen transport or storage in transport media, in our and others’ experience viral transport medium does not adversely affect performance of the rapid tests, allows culture and rapid testing to be performed on the same specimen, and obviates the need to obtain a second sample [50, 54]. In addition, although several manufacturers advocate placing the specimen swab in the paper wrapper for transport, we try to discourage this practice as swabs tend to dry out. All specimens may be stored refrigerated (2º to 8ºC) for upto 24 hours prior to testing without detrimentally affecting test performance. [55]. The patient population being tested has a significant effect on the performance of rapid diagnostic assays for influenza. In general, these tests perform best on specimens obtained from young children. Sensitivity of the Biostar FLU OIA rapid test for influenza was 71.8% with specimens from children or adolescents but only 51.4% with specimens from adults [56]. Multiple factors are likely to contribute to this discrepancy. The rate of influenza virus isolation is greater in children than in adults and influenza virus load is greater in respiratory specimens from children [57, 58]. In addition, influenza virus shedding is prolonged in children compared to adults (from seven to 21 days versus three to eight days, respectively) [59, 60]. Furthermore, patient age itself is likely to influence specimen- type. Nasopharyngeal aspirates or swab specimens are more frequently obtained from young children, whereas throat swabs (an inferior specimen type) are preferentially obtained from adults. This presumably reflects the lower tolerance of adults, in general, to the taking of nasal specimens. Similarly,

162

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

because shedding of respiratory viruses is prolonged in the immunocompromised host, the immune status of the patient population may also affect performance of rapid tests [2, 61]. In contrast to cell culture and DFA, rapid diagnostic tests do not detect influenza B as efficiently as influenza A virus [41, 44, 51]. This may be problematic if rapid tests are used in seasons or in populations where influenza B is the predominant epidemic strain in circulation. In addition, although a FLU OIA, Quickvue or Zstatflu positive test result suggests the presence of influenza, it does not differentiate between influenza A or B virus. Apart from obvious disadvantages for epidemiologic surveillance, this may have implications if treatment with amantadine or rimantadine, which are ineffective against influenza B, is contemplated. Directigen Flu A Test (Becton Dickinson, Franklin Lakes, NJ.). The Directigen Flu A test is an immunoassay in which the binding of monoclonal antibodies to extracted influenza A virus antigen (conserved influenza A nucleoprotein) is visually detected by a chromogen. This produces a purple signal on a white background indicating a positive result. The test is relatively simple to perform and can directly detect influenza A virus in clinical specimens in less than 15 minutes. In addition, each test contains positive and negative internal controls. In studies, when compared to cell culture isolation for the detection of influenza A virus in nasal and nasopharyngeal wash specimens from predominantly young children, the Directigen FLU-A test has: sensitivity, 75% to 100%; specificity, 91.6% to 100%; PPV, 62% to 100%; and NPV, 89% to 100% [43, 44, 53, 62]. These studies represent the optimal population (predominantly young children) and specimen type (nasal washes) for detection of influenza by rapid test. Interestingly, it seems that specimen-type is a more important factor in test performance than population tested. In comparison to virus culture, the Directigen Flu A test demonstrated: sensitivity, 86.8% and specificity, 99.1%, for detection of influenza A virus in nasopharyngeal and throat swabs specimens from an institutionalized geriatric population [42]. In contrast, test performance was dramatically lower (sensitivity, 39% to 62%) in throat swab and pharyngeal gargles specimens from mixed populations of patients [53, 63]. Directigen Flu A ⫹ B (Becton Dickinson, Franklin Lakes, NJ.) The Directigen Flu A ⫹ B test is a newer version of the Flu A test that detects and discriminates between influenza A and B viruses. In studies of nasal aspirate specimens from children, the Directigen Flu A ⫹ B test demonstrated: sensitivity, 75% to 87.5%; specificity, 93% to 96.8%; PPV, 74% to 80% and NPV, 93% to 98% [41, 64]. Sub-analysis, demonstrated better detection of influenza A than type B virus (sensitivity and specificity, 100% and 98.7% versus 87.5% and 93%, for influenza A and B viruses, respectively) [41]. Again, patient selection and specimen type may have optimized performance of the rapid antigen detection test. Although storage of specimen at 2 to 8ºC for up to 48 hours did not appear to affect results for detection of influenza A, detection of type B virus was adversely affected. Flu OIA (Biostar, Inc., Boulder, CO). The Flu OIA optical immunoassay detects influenza A and B nucleoprotein in respiratory specimens but does not differentiate between them. Antigen binds to an antibody-coated silicon wafer, which changes the wafer’s optical thickness and alters the path of reflected light. This produces a purple dot indicating a positive result. An internal control dot is visible in a valid test. The assay requires eight steps, takes 15 to 20 minutes, requires minimal training, no instrumentation and may be performed by non-technical personnel in the ‘point-of-care’

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

163

environment (CLIA waived). In a study, conducted by the manufacturer, comparing Flu OIA to virus culture for detection of influenza in respiratory specimens, overall sensitivity of the rapid test was 80.1% [46]. However, performance varies according to specimen-type, with sensitivity highest (88.4%) in nasal aspirates and lowest (62.1%) in throat swabs. Specificity overall was 73.1%, with variability between specimen-types (79.5% in nasal aspirates, and 51.5% in sputum specimens). The authors suggest that FLU OIA specificity may have been higher, as 66% of test-positive but culture-negative specimens were influenza-positive by RT-PCR [46]. However, in an additional study, evaluating detection of influenza viruses in throat swab specimens of children and adults, the FLU OIA test performed poorly [65]. In comparison to cell culture, test sensitivity, specificity, PPV and NPV were 54%, 74.1%, 72.7% and 55.8%, respectively [65]. In contrast to the previous study, almost all the FLU OIA-positive, culture-negative specimens (86%) were negative for influenza by RT-PCR. Performance of FLU OIA was similar for a range of respiratory specimens compared to culture and FA (overall sensitivity, 54%; specificity, 97%; PPV of 91%; and NPV, 77%) [50]. Performance was again heavily dependent on specimen type: sputum and BAL specimens gave the best results (sensitivity, 90% to 100%; and specificity, 100%) whereas performance was worse in throat swabs (sensitivity, 25%; and specificity, 94%). In a subsequent review of independent studies of the FLU OIA for detection of influenza in a variety of specimen types, sensitivity and specificity ranged from 36% to 89.7% and from 65% to 100%, respectively [54]. During a 1998/1999 outbreak of influenza in Switzerland, in comparison to cell culture, FLU OIA demonstrated overall sensitivity of 64.4% and specificity of 94.9%, for detection of influenza A and B in various respiratory specimens from adults and children [56]. During the outbreak, when the prevalence of influenza was 50%, the NPV of the test was 73%. Due to the low specificity of FLU OIA, at the beginning or end of the influenza season, when the prevalence may be as low as 5%, PPV of the test was 40%, i.e. only two of five patients with positive FLU OIA results would be true positives [56]. On the basis of this study, Schultze et al. conclude that FLU OIA is not suitable for use as a decision making tool or selecting patients for antiviral treatment and that improved rapid diagnostic tests were required. Similarly, in the authors’ microbiology laboratory, during the 2001–2002 and 2002–2003 influenza seasons, we evaluated the Biostar FLU OIA assay in comparison to conventional cell culture and FA for detection of influenza in all respiratory specimens (throat swabs, nasal swabs, sputa and bronchial lavage) submitted for influenza testing. Eighty-six percent of the population tested were outpatients. Seventy-two patients (adults and children) were influenza-positive by cell culture. Biostar FLU OIA test demonstrated overall sensitivity of 30.5% (22 of 72) for detection of influenza. Sensitivity was significantly better for the detection of influenza A (33%, 20/60) than influenza B (16.6%, 2/12). Such results are far from encouraging and question the utility of the test in our patient population. Binax NowFlu A and B tests (Binax Inc., Portland, ME.) The Binax Now Flu A and B tests detect influenza A and B viruses, using specific antibodies to influenza A or B nucleoprotein conjugated to visualizing particles dried to separate test strips. The test is simple to perform, requires no additional equipment and provides a result in approximately 15 minutes. Built-in procedural controls and external positive and negative controls are included. A pink to purple color change accompanied by a

164

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

control-line indicates a positive-result. In a study of the Binax Now Flu A and B tests ability to detect influenza viruses from nasal wash specimens in a predominantly pediatric population, sensitivity and specificity, in comparison to cell culture, were 82% and 94%, and 71% and 97% for influenza A and B viruses, respectively [51]. In the same study, if nasopharyngeal swab specimens were tested by the respective A and B tests performance was worse (sensitivity and specificity, 78% and 92%; and 58% and 97% for influenza A and B viruses, respectively). Quidel Quick Vue Test (Quidel, San Diego, CA.). The Quick Vue influenza test is an immunoassay that uses monoclonal antibodies on a test-strip to detect influenza nucleoprotein. After a nasal swab or aspirate specimen is mixed into a solution, which disrupts viral particles, the test strip is added. A positive result, consisting of the appearance of a pink/red line after ten minutes, indicates the presence of but does not differentiate between influenza A or B. In a study of the Quick Vue influenza test to detect influenza viruses in nasal swab specimens from a population of military recruits, in comparison to cell culture, sensitivity and specificity, were 65% and 92%, respectively [66]. ZstatFlu-II Test. (ZymeTx, Inc., Oklahoma City, OK.). In contrast to other rapid diagnostic tests, which detect influenza nucleoprotein antigen, the ZstatFlu-II test is an enzyme-based assay that detects the functional NA activity of influenza A or B virus. The specimen is incubated for 15 minutes at room temperature with a synthetic NA substrate bound to a chemiluminescent molecule and the mixture is than placed in a special imaging device. Liberation of the chemiluminescent reporter by activity of influenza NA is detected using polarized high-speed film. A positive result appears as a white plus sign against a black background. Unlike the Directigen and FLU OIA tests, the ZstatFlu test does not have internal positive or negative controls. In a series of pediatric studies (predominantly on nasal aspirate specimens), the ZstatFlu test demonstrated variable sensitivity of 70.1% to 96% and specificity of 77% to 92.4%, for detection of influenza virus in comparison to cell culture [44, 49, 64]. Although the NPV was high (89.9% to 98%), because of low PPV (59% to 76.3%) cell culture confirmation of positive results was deemed necessary. The ZstatFlu test is similar to other rapid tests that appear better at detecting influenza A than B-type virus (sensitivity, 76.4% versus 40.9%) [44]. Few studies have compared influenza rapid diagnostic tests to each other. Results of comparisons of both Zstat Flu and Directigen Flu tests to cell culture are somewhat contradictory. In separate studies (using nasal wash specimens from children), performance of the Zstat Flu test was significantly better than the Directigen Flu A and B test for detection of influenza A and B (sensitivity, 88% versus 75%), but was worse than the Directigen Flu A test for detection of influenza A virus (sensitivity, 76.4% versus 89.7%) [44, 64]. However, in the only head-to-head comparison of four rapid diagnostic tests, in a predominantly pediatric population, the Directigen Flu A, QuickVue and Flu OIA tests demonstrated equivalent performance in comparison to cell culture and DFA: sensitivity, 93% to 95%; specificity, 82% to 85%; PPV, 81% to 86%; and NPV, 92% to 94% [67]. Sensitivity (72%) and NPV (75%) of the Zstat Flu test were significantly lower. However, flaws in study design may have adversely affected performance of the Zstat Flu test. Specifically, whereas predominantly nasal swab specimens were used to evaluate the other rapid tests, throat swab specimens

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

165

(an inferior specimen type) were used exclusively to evaluate the Zstat Flu test. In addition, with the exception of the Directigen Flu A test, which was performed immediately, other tests were carried out on stored specimens. While this did not seem to affect the performance of the QuickVue and Flu OIA tests, because the Zstat Flu test depends on detection of virus NA enzyme activity, storage may have compromised its performance. Finally, in the same study, although the ZstatFlu test is considered relatively simple to perform and despite instruction, test performance was highly variable among technologists (sensitivity and specificity ranged from 45.7% to 70.1%; and from 76.2% to 92.4%, respectively) [44]. This is somewhat concerning as ZstatFlu test is one of three rapid tests approved for ‘point-of-care’ testing. 4.6. Molecular diagnosis Influenza viruses extreme genetic variability is a challenge for design of molecular-based diagnostic tests. However, a number of promising molecular based techniques have been developed. Several ‘in-house’ reverse transcriptase polymerase chain reaction (RT-PCR) assays, which use nested primers to detect and sub-type influenza virus, have demonstrated greater sensitivity than other rapid diagnostic tests and conventional cell culture. In a review of studies comparing RT-PCR to cell culture, 42.2% and 30.9% of 1298 respiratory specimens were influenza-positive by RT-PCR and cell culture, respectively [68]. In a further study, the detection rate for influenza A virus by RT-PCR was clearly higher (93%) than that by cell culture (80%) and ELISA (62%) [55]. RT-PCR demonstrated sensitivity for influenza that was 103 and 106 greater than cell culture and ELISA, respectively. In addition, sensitivity of RT-PCR does not appear to be influenced by age of the patient. However, RT-PCR is the most expensive of the diagnostic tests for influenza. Turnaround time of RT-PCR for influenza is one to two days. In addition, molecular testing requires considerable skill and expertise to perform and must be integrated into laboratory workflow. By combining multiple sets of primers, multiplex RT-PCR offers the possibility for molecular detection of several respiratory viruses in a single reaction [69, 70]. The Hexaplex quantitative multiplex RT-PCR assay (Prodesse Inc., Waukesha, Wis.) for detection of seven major respiratory viruses, has demonstrated greater sensitivity than cell culture and other rapid tests for detection of influenza [70]. However, its seven-hour turnaround time make it more suitable for overnight testing. Although the complexity and turnaround time (one to two days) of conventional RTPCR may preclude its use for truly rapid influenza diagnosis, real-time RT-PCR, in which amplification and detection occur in the same reaction tube, can provide genuinely rapid results (within four to five hours after obtaining the specimen). Because real-time RT-PCR amplifies and detects product in a closed tube system, it has the additional advantage of reducing the possibility of PCR contamination. Once again, real-time RT-PCR has demonstrated greater sensitivity than cell culture for detection of influenza virus [71]. In addition, real time RT-PCR can detect influenza virus for up to seven days after onset of symptoms, in comparison to two days after onset of symptoms by virus culture. DNA microarrays, which can contain hybridization probes for assorted respiratory virus genes, offer the promise of rapid and accurate detection and subtyping of respiratory viruses, including influenza, in individual patients or as part of large-scale surveillance [68].

166

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

In future, molecular based assays seem set to play a larger role in rapid diagnosis and epidemiologic surveillance. However, in the event of large antigenic drift or major antigenic shift in influenza, novel strains may not be detected by existing primer sets, and diagnosis may depend on traditional virus culture. This was highlighted by the detection of the first case of human influenza A H1N5 in Hong Kong, in May, 1997 [10]. This strain, which was initially identified by culture and DFA as influenza A, was not detected by pre-existing PCR primers specific for H1 and H3. Thus, rather than replacing traditional virus isolation, molecular methods will likely complement cell culture, which will continue to play a major role in global epidemiologic influenza surveillance and vaccine strain selection. 4.7. Serologic diagnosis: Serologic diagnosis of influenza infection is based on demonstration of a four-fold or greater rise in specific antibody titer between acute and convalescent serum samples, measured by hemagglutination inhibition, EIA, complement fixation or neutralization tests [2]. The need for paired serum samples, the first collected as soon after the onset of illness as possible and the second collected about two to four weeks later, makes serology a retrospective diagnostic tool and limits its clinical utility. Although serology may occasionally establish a diagnosis when all other means have failed, its main value lies in epidemiology or as a research tool.

5. Role of rapid diagnostic tests for influenza The advent of rapid diagnostic tests for influenza should be of benefit to patients in the community, in hospitals and to public health. The choice of diagnostic test depends on the turnaround time required for a result to be clinically useful in a particular setting. Thus, to have clinical relevance, an emergency room or remote physician’s office requires a test with a truly rapid turnaround time (preferably, less than 30 minutes) that is simple to perform and interpret. Three of the currently available influenza rapid diagnostic tests are approved for use in the ‘point-of-care’ setting (CLIA waived), while the remainder are considered moderately complex and more suited for the microbiology laboratory (table I.). In the majority of cases, the high specificity of rapid diagnostic tests allows for a firm diagnosis to be made with a positive result. However, because they are among the least sensitive tests for detection of influenza, receipt of a negative rapid test result, in a patient with a clinically compatible illness, should be confirmed, preferably by cell culture. Use of rapid diagnostic tests for influenza in the Emergency Room improves the medical management of pediatric patients by decreasing unnecessary antibiotic use and ancillary testing in out-patients, and decreasing duration of antibiotic use and increasing antiviral treatment of in-patients [72, 73]. In contrast, for the hospital in-patient or physician’s office with ready access to a microbiology laboratory, more sensitive but less rapid tests that provide results with a ‘same-day’ or oneday turnaround time may be adequate. In this setting, DFA and rapid shell-vial culture may be useful. In addition, a DFA test can be used to detect several additional respiratory viruses (e.g., RSV, adenovirus, and parainfluenza viruses) in the same specimen and can be performed

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

167

in batches. However, because DFA tests are more complex and technically demanding, they are best suited for use during the day-shift by experienced laboratory personnel. Thus, for hospital in-patients, laboratory confirmation of the clinical diagnosis with rapid shell-vial culture of respiratory viruses, including influenza, provides results that are both clinically relevant and financially beneficial for both patient and health-care provider [74, 75]. Benefits include reduced mortality, length of hospital stay and costs, and more appropriate use of antimicrobial and antiviral treatment. Similarly, in the hospital or nursing home setting, adjunctive rapid diagnostic tests, in combination with virus culture, may enhance infection control efforts by allowing timely isolation or cohorting of infected patients, and vaccination, treatment, or chemoprophylaxis of at risk patients. In the public health arena, the addition of rapid diagnostic testing to conventional virus culture has had significant benefit. In the recent highly fatal outbreak of a novel influenza virus H5N1 in Hong Kong, rapid detection of by Directigen Flu A, RT-PCR and DFA assays allowed early antiviral treatment, isolation and contact investigation [76]. Furthermore, rather than undermining virologic surveillance, as was initially feared, by offering rapid influenza testing to physicians, public health laboratories may actually increase the number of specimens submitted for culture, thereby, enhancing recovery of circulating influenza isolates [77].

6. A practical approach for the laboratory diagnosis of influenza In the authors’ laboratory, if a test for influenza is requested for a hospitalized patient, where a reliable diagnosis may be critical, cell culture is always performed, with or without additional DFA or rapid antigen detection tests. For out-patients, outside of ‘influenzaseason’, we perform virus culture to detect influenza. If a rapid test is requested at this time, in special circumstances, such as in an immunocompromised patient, or after consultation with infectious disease physicians, an additional DFA test may be performed. In contrast, from the onset of influenza season, when prevalence and pre-test probability of influenza are higher, we offer either a rapid diagnostic test plus confirmatory virus culture or virus culture alone. In addition to enhancing the lower sensitivity and PPV of rapid diagnostic tests, virus culture detects respiratory viruses other than influenza, and permits ongoing surveillance of circulating influenza strains. If requested, a rapid diagnostic test may be performed around the clock by most personnel, whereas DFA testing is generally only performed during the day-shift by experienced technologists. Where results of rapid diagnostic tests are equivocal, a DFA test is occasionally performed. Rapid testing is only offered after the first culture-confirmed cases of influenza are reported from the community. In most cases, epidemics of influenza are easy to recognize. Because the frequency of influenza infection and morbidity are highest in the young, the existence of an influenza epidemic is often first recognized in young people [78]. In our area, the onset of influenza-season is most frequently heralded by positive influenza cultures submitted from sentinel sites in pediatrician’s offices and student health-centers. At the beginning of each influenza-season, local health care providers receive a letter from the laboratory stating that: during influenza-season most influenza cases can be diagnosed clinically without laboratory

168

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

testing, and that if a rapid diagnostic test is deemed necessary it must be ordered. Confirmatory virus culture will always be performed. A suggested algorithm for influenza diagnosis is presented in figure 1.

7. Summary There is no perfect test for the diagnosis of influenza. Virus culture, the present ‘gold-standard test’ is not 100% sensitive and does not provide results in a timeframe that allows optimal use of potentially effective antiviral treatment. Although rapid diagnostic tests provide results in less than 30 minutes, they are significantly less sensitive. The decision to introduce a rapid diagnostic test to the clinical laboratory or ‘point-ofcare’ must weigh the test sensitivity against turnaround time. Additional factors to consider are the age (children or adult), immune status, and location (hospital or out-patient) of the population to be tested. Such factors will determine the specimen-type that is likely to be submitted for rapid testing, which will, in turn, greatly influence test performance. In general, swab specimens, particularly throat swabs, are the least desirable specimen-type and obtaining a nasal/nasopharyngeal aspirate or sputum specimen should be encouraged. It is our opinion, that rapid diagnostic tests should only be used in influenza season and results should be confirmed with virus culture. It goes without saying that rapid influenza testing should not be used casually in patients without a clinically compatible illness. Despite reservations regarding poor sensitivity and NPV of rapid diagnostic tests, receipt of a positive-result, in influenza-season, may be potentially beneficial for patient and public health. Thus, a positive rapid influenza result may permit effective treatment or chemoprophylaxis, the institution of timely infection control, may reduce hospital-stay, inappropriate antibiotics and unnecessary

Fig. 1. Algorithm for influenza diagnosis.

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

169

investigation, and may play a role in epidemic or pandemic planning. And finally, good communication between microbiologists and physicians should ensure that local health care providers are aware of the arrival of influenza in their community. Armed with this information the physicians’ ability to clinically diagnose influenza should be greatly improved.

References [1] Treanor JJ. Orthomyxoviridae. In: Mandell GL, Bennett JE, Dolin R, eds. Principles and Practice of Infectious Diseases. 5th Ed. Vol 2. Philadelphia, PA.: Churchill Livingstone, 2000;1823–49. [2] Cox NJ, Subbarao K. Influenza. Lancet 1999;345:1277–82. [3] Dagan R, Hall CB. Influenza A virus infection imitating bacterial sepsis in infancy. Pediatr Infect Dis J 1984;3(3):218–21. [4] Gelzen WP. Emerging infections: pandemic influenza. Epidemiol Rev 1996;18(1):64–76. [5] Layne SP, Beugelsdijk TP, Patel CKN, Taubenberger JK, Cox NJ, Gust ID, Hay AJ, Tashiro M, Lavanchy D. A Global lab against influenza. Science 2001;293:1729. [6] Simonsen L, Clarke MJ, Williamson D, Stroup DF, Arden NH, Schonberger LB. The impact of influenza epidemics on mortality: introducing a severity index. Amer J Publ Health 1997;87(12):1944–50. [7] Thompson WW, Shay DK, Weintraub E, Brammer L, Cox N, Anderson LJ, Fukuda K. Mortality associated with influenza and respiratory syncytial virus in the United States. JAMA 2003;289(2):179–86. [8] Simonsen L, Fukuda K, Schonberger LB, Cox NJ. The impact of influenza epidemics on hospitalizations. J Infect Dis 2000;181:831–7. [9] Schoenbaum SC. Economic impact of influenza. The individual’s perspective. Amer J Med 1987;82(Suppl 6A):26–30. [10] Subbarao K, Klimov A, Katz J, Regnery H, Lim W, Hall H, Perdue M, Swayne D, Bender C, Huang J, Hemphill M, Rowe T, Shaw M, Xu X, Fukuda K, Cox N. Characterization of an avian influenza A (H5N1) virus isolated from a child with a fatal respiratory illness. Science 1998;279:393–6. [11] Claas ECJ, Osterhaus ADME, van Beek R, De Jong JC, Rimmelzwaan GF, Senne DA, Krauss S, Shortridge KF, Webster RG. Human influenza A H5N1 virus related to a highly pathogenic avian influenza virus. Lancet 1998;351:472–7. [12] Hayden FG, Palese P. Influenza virus. In: Richman DD, Whitley RJ, Hayden FG, eds. Clinical Virology 2nd ed. Wahington, DC: ASM Press; 2002;891–920. [13] CDC . Prevention and control of influenza. Recommendations of the Advisory Committee on Immunization Practices. MMWR April 12, 2002;51(RR03):1–31. [14] Two neuraminidase inhibitors for treatment of influenza. Med Lett 1999;41(1063):91–3. [15] Hayden FG, Gubareva LV, Monto AS, Klein TC, Elliott MJ, Hammond JM, Sharp SJ, Ossi MJ. Inhaled zanamivir for the prevention of influenza in families. New Engl J Med 2000;343(18):1283–9. [16] Gubareva LV, Kaiser L, Hayden FG. Influenza virus neuraminidase inhibitors. Lancet 2000;355:827–35. [17] Hayden FG, Atmar RL, Schilling M, Johnson C, Poretz D, Parr D, Huson L, Ward P, Mills RD. Use of the selective oral neuraminidase inhibitor oseltamivir to prevent influenza. New Engl J Med 1999;341:1336–43. [18] Nicholson KG, Aoki FY, Osterhaus ADME, Trottier S, Carewicz O, Mercier CH, Rode A, Kinnersley N, Ward P. Efficacy and safety of oseltamivir in treatment of acute influenza: a randomised controlled trial. Lancet 2000;355:1845–50. [19] Treanor JJ, Hayden FG, Vrooman PS, Rick Barbarash, Bettis R, Riff D, Singh S, Kinnersley N, Ward P, Mills RD. Efficacy and safety of the oral neuraminidase inhibitor oseltamivir in treating acute influenza. JAMA 2000;283(8):1016–24. [20] Hayden FG, Osterhaus ADME, Treanor JJ, Fleming DM, Aoki FY, Nicholson KG, Bohnen AM, Hirst HM, Keene O, Wightman K. Efficacy and safety of the neuriminidase inhibitor zanamivir in the treatment of influenzavirus infections. New Engl J Med 1997;337:874–80. [21] Monto AS, Robinson DP, Herlocher ML, Hinson JM, Elliott MJ, Crisp A. Zanamivir in the prevention of influenza among healthy adults. JAMA 1999;282(1):31–5.

170

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

[22] Monto AS, Gravenstein S, Elliott M, Colopy M, Schweinle J. Clinical signs and symptoms predicting influenza infection. Arch Int Med 2000;160(21):3243–7. [23] Boivin G, Hardy I, Tellier G, Maziade J. Predicting influenza infections during epidemics with use of a clinical case definition. Clin Infect Dis 2000;31:1166–9. [24] Zambon M, Hays J, Webster A, Webster R, Keene O. Diagnosis of influenza in the community: relationship of clinical diagnosis to confirmed virological, serologic, or molecular detection of influenza. Arch Int Med 2001;161(17):2116–22. [25] Govaert T, Dinart G, Arertz K, Knottnerus J. The predictive value of influenza symptomatology in elderly people. Fam Practition 1998;15:16–22. [26] van Elden LJR, van Essen GA, Boucher CAB, Nijhuis M, Schipper P, Verheij TJM, Hoepelman I. Clinical diagnosis of influenza virus infection: evaluation of diagnostic tools in general practice. Brit J Gen Pract 2001;51:630–4. [27] Sintchenko V, Gilbert GL, Coiera E, Dwyer D. Treat or test first? Decision analysis of empirical antiviral treatment of influenza virus infection versus treatment based on rapid test results. J ClinVirol 2002;25:15–21. [28] Smith KJ, Roberts MS. Cost-effectiveness of newer treatment strategies for influenza. Amer J Med 2002;113:300–7. [29] CDC. Isolation of avian influenza A (H5N1) viruses from humans- Hong Kong, May-December 1997. MMWR 1997;46(50):1204–7. [30] Chan PKS. Outbreak of avian influenza A (H5N1) virus infection in Hong Kong in 1997. Clin Infect Dis 2002;34(Suppl. 2):S58–S64. [31] Espy MJ, Smith TF, Harmon MW, Kendal AP. Rapid detection of influenza virus by shell vial assay with monoclonal antibodies. J Clin Microbiol 1986;24(4):677–9. [32] Mills RD, Cain KJ, Woods GL. Detection of influenza virus by centrifugal inoculation of MDCK cells and staining with monoclonal antibodies. J Clin Microbiol 1989;27(11):2502–8. [33] Matthey S, Nicholson D, Ruhs S, Alden B, Knock M, Schultz K, Schmuecker A. Rapid detection of respiratory viruses by shell vial culture and direct staining by using pooled and individual monoclonal antibodies. J Clin Microbiol 1992;30(3):540–4. [34] Stokes CE, Bernstein JM, Kyger SA, Hayden FG. Rapid diagnosis of influenza A and B by 24-h fluorescent focus assays. J Clin Microbiol 1988;26(7):1263–6. [35] Barenfanger J, Drake C, Mueller T, Trout T, O’Brien J, Guttamn K. R-mix cells are faster, at least as sensitive and marginally more costly than conventional cell lines for the detection of respiratory viruses. J Clin Virol 2001;22:101–10. [36] George KS, Patel NM, Hartwig RA, Scholl DR, Jollick JA, Kauffmann LM, Evans MR, Rinaldo CR. Rapid and sensitive detection of respiratory virus infection for directed antiviral treatment using R-mix cultures. J Clin Virol 2002;24:107–15. [37] Evans MR, Patel NM, Hartwig B, Kauffmann LM. Rapid influenza virus detection by R-mix (mixed fresh cells) at 6 and 20 hours postinfection: comparison to RhMK tube culture (5d), POCT’s and DFA. ICAAC, Toronto, Ontario, Canada, Sept., 17–20, 2000. Abstract 360. [38] Daisy JA, Lief FS, Friedman HM. Rapid diagnosis of influenza A infection by direct immunofluorescence of nasopharyngeal aspirates in adults. J Clin Microbiol 1979;9(5):688–92. [39] Ray CG, Minnich LL. Efficiency of immunofluorescence for rapid detection of common respiratory viruses. J Clin Microbiol 1987;25(2):355–7. [40] Doing KM, Jerkofsky MA, Dow EG, Jellison JA. Use of fluorescent antibody staining of cytocentrifugeprepared smears in combination with cell culture for direct detection of respiratory viruses. J Clin Microbiol 1998;36(7):2112–4. [41] Chan KH, Maildeis N, Pope W, Yup A, Ozinskas A, Gill J, Seto WH, Shortridge KF, Peiris JSM. Evaluation of the Directigen FluA+B test for rapid diagnosis of influenza virus type A and B infections. J Clin Microbiol 2002;40(5):1675–80. [42] Leonardi GP, Leib H, Birkhead GS, Smith C, Costello P, Conron W. Comparison of rapid detection methods for influenza A virus and their value in health-care managemant of institutionalized geriatric patients. J Clin Microbiol 1994;32(1):70–4.

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

171

[43] Dominguez EA, Taber LH, Couch RB. Comparison of rapid diagnostic techniques for respiratory syncytial virus and influenza virus respiratory infections in young children. J Clin Microbiol 1993;31(9):2286–90. [44] Noyola DE, Clark B, O’Donnell FT, Atmar RL, Greer J, Demmler GJ. Comparison of a new neuraminidase detection assay with an enzyme immunoassay, immunofluorescence, and culture for rapid detection of influenza A and B viruses in nasal wash specimens. J Clin Microbiol 2000;38(3):1161–5. [45] Rapid diagnostic tests for influenza. Med Lett 1999;41(1068):121–2. [46] Covalciuc KA, Webb KH, Carlson CA. Comparison of four clinical specimen types for detection of influenza A and B viruses by optical immunoassay (FLU OIA test)and cell culture methods. J Clin Microbiol 1999; 37(12):3971–4. [47] McQuillin J, Madeley CR, Kendal AP. Monoclonal antibodies for the rapid diagnosis of influenza A and B virus infections by immunofluorescence. Lancet 1985;2:911–4. [48] Kaiser L, Briones MS, Hayden FG. Performance of virus isolation and Directigen Flu A to detect influenza A virus in experimental human infection. J Clin Virol 1999;14:191–7. [49] Noyola DE, Paredes AJ, Clark B, Demmler GJ. Evaluation of a neuraminidase detection assay for the rapid detection of influenza A and B virus in children. Pediatr Develop Pathol 2000;3:162–7. [50] Hindiyeh M, Goulding C, Morgan H, Kenyon B, Langer J, Fox L, Dean G, Woolstenhulme D, Turnbow A, Billetdeaux E, Shakib S, Gordon C, Powers A, Vardeny G, Johnson M, Skodack-Jones L, Carroll K. Evaluation of Biostar FLU OIA assay for rapid detection of influenza A and B viruses in respiratory specimens. J Clin Virol 2000;17:119–26. [51] Product monograph. Binax Inc., Portland ME. 2002. [52] Kimball AM, Foy HM, Cooney MK, Allan ID, Matlock M, Plorde JJ. Isolation of respiratory synctial and influenza viruses from the sputum of patients hospitalized with pneumonia. J Infect Dis 1983;147(2): 181–4. [53] Ryan-Poirier KA, Katz JM, Webster RG, Kawaoka Y. Application of Directigen FLU-A for the detection of influenza A virus in human and nonhuman specimens. J Clin Microbiol 1992;30(5):1072–5. [54] Tucker SP, Cox C, Steaffens J. A flu optical immunoassay (ThermoBioStar’s FLU OIA): a diagnostic tool for improved influenza management. Phil Trans R Soc Lond 2001;356:1915–24. [55] Steininger C, Kundi M, Aberle SW, Aberle JH, Popow-Kraupp T. Effectiveness of reverse transcription-PCR, virus isolation, and enzyme-linked immunosorbent assay for diagnosis of influenza A virus infection in different age groups. J Clin Microbiol 2002;40(6):2051–6. [56] Schultze D, Thomas Y, Wunderli W. Evaluation of an optical immunoassay for the rapid detection of influenza A and B viral antigens. Eur J Clin Microbiol Infect Dis 2001;20:280–3. [57] Long CE, Hall CB, Cunningham CK, Weiner LB, Alger KP, Gouveia M, Colella CB, Schnabel KC, Barker WH. Influenza surveillance in community-dwelling elderly compared with children. Arch Fam Med 1997;6:459–65. [58] Murphy BR, Chalhub EG, Nusinoff SR, Kasel J, Chanock RM. Temperature sensitive mutants of influenza virus, III. Further characterization of the ts-1[E] influenza A recombinant (H3N2) virus in man. J Infect Dis 1973;128:479–87. [59] Frank AL, Taber LH, Wells CR, Wells JM, Gelzen WP. Patterns of shedding of myxoviruses and paramyxoviruses in children. J Infect Dis 1981;144:433–41. [60] Hall CB Jr, Douglas RG. Nosocomial influenza infection as a cause of intercurrent fevers in infants. Pediatrics 1975;55(5):673–7. [61] Couch RB, Englund JA, Whimbey E. Respiratory viral infections in immunocompetent and immunocompromised persons. Am J Med 1997;102(3A):2–9. [62] Waner JL, Todd SJ, Shalaby H, Murphy P, Hall LV. Comparison of Directigen FLU-A with viral isolation and direct immunofluorescence for the rapid detection and identification of influenza A virus. J Clin Microbiol 1991;29(3):479–82. [63] Johnston SL, Bloy H. Evaluation of a rapid enzyme immunoassay for detection of influenza A virus. J Clin Microbiol 1993;31(1):142–3. [64] Hamilton MS, Abel DM, Ballam YJ, Otto MK, Nickell AF, Pence LM, Appleman JR, Shimasaki CD, Achyuthan KE. Clinical evaluation of the ZstatFlu-II test: a chemiluminesence rapid diagnostic test for influenza virus. J Clin Microbiol 2002;40(7):2331–4.

172

P.J. Gavin, R.B. Thomson Jr./Clin. Applied Immunol. Rev. 4 (2003) 151–172

[65] Boivin G, Hardy I, Kress A. Evaluation of a rapid optical immunoassay for influenza viruses (FLU OIA test) in comparison with cell culture and reverse transcription-PCR. J Clin Microbiol 2001;39(2):730–2. [66] Hawksworth AW, Malasig MD, Connolly JL, Williams ST, Hewitson DG, Wolf SM, Ryan MAK, Gray GC. Clinical evaluation of two rapid diagnostic tests for influenza. 41st ICAAC, Chicago. IL, Sept. 22– 25, 2001. Abstract H-1586. [67] Rodriguez WJ, Schwartz RH, Thorne MM. Evaluation of diagnostic tests for influenza in a pediatric practice. Pediatr Infect Dis J 2002;21(3):193–6. [68] Taubenberger JK, Layne SP. Diagnosis of influenza virus: Coming to grips with the molecular era. Mol Diagn 2001;6(4):291–305. [69] Stockton J, Ellis JS, Saville M, Clewely JP, Zambon MC. Multiplex PCR for typing and subtyping influenza and respiratory syncytial viruses. J Clin Microbiol 1998;36(10):2990–5. [70] Fan J, Hendrickson KJ, Savatski LL. Rapid simultaneous diagnosis of infections with respiratory syncytial viruses A and B, influenza viruses A and B, and human parainfluenza virus types 1, 2 and 3 by multiplex quantitative reverse transcription-polymerase chain reaction- enzyme hybridization assay (Hexaplex). Clin Infect Dis 1998;26:1398–402. [71] van Elden LRJ, Nijhuis M, Schipper P, Schuurman R, van Loon AM. Simultaneous detection of influenza viruses A and B using real-time quantitative PCR. J Clin Microbiol 2001;39(1):196–200. [72] Noyola DE, Demmler GJ. Effect of rapid diagnosis on management of influenza A infections. Pediatr Infect Dis J 2000;19(4):303–7. [73] Sharma V, Dowd D, Slaughter AJ, Simon SD. Effect of rapid diagnosis of influenza virus type A on the emergency department management of febrile infants. Arch Pediatr Adolesc Med 2002;156:41–3. [74] Barenfanger J, Drake C, Leon N, Mueller T, Troutt T. Clinical and financial benefits of rapid detection of respiratory viruses: an outcome study. J Clin Microbiol 2000;38:2824–8. [75] Woo PCY, Chiu SS, Seto WH, Peiris M. Cost effectiveness of rapid diagnosis of viral respiratory tract infections in pediatric patients. J Clin Microbiol 1997;35:1579–81. [76] Yuen KY, Chan PKS, Peiris M, Tsang DNC, Que TL, Shortridge KF, Cheung PT, To WK, Ho ETF, Sung R, Cheng AFB. Clinical features and rapid viral diagnosis of human disease associated with avian influenza A H5N1 virus. Lancet 1998;351:467–71. [77] Effler PV, Leong M-C, Tom T, Nakata M. Enhancing public health surveillance for influenza virus by incorporating newly available rapid diagnostic tests. Emerging Infect Dis 2002;8(1):23–8. [78] Monto AS. Influenza: quantifying morbidity and mortality. Amer J Med 1987;82(Suppl 6A):20–5.

Suggest Documents