|Home | About | Journals | Submit | Contact Us | Français|
Directigen FluA+B (BD Diagnostic Systems, Sparks, Md.), a new rapid test for the detection of influenza virus types A and B, was evaluated with nasopharyngeal aspirate specimens collected from 250 patients in comparison with culture and direct fluorescent antigen (DFA) detection tests. The patients studied were predominantly children, 80% being ≤6 years old. Specimens negative by culture but positive by the Directigen FluA+B or DFA tests were analyzed by reverse transcription-PCR to resolve the discrepant results. The resolved sensitivity, specificity, and positive and negative predictive values of the Directigen FluA+B test for influenza virus type A were 96%, 99.6%, 96%, and 99.6%, respectively, and for influenza virus type B they were 87.5%, 96.8%, 80%, and 98%, respectively. Storage of nasopharyngeal aspirates in virus transport medium at 2 to 8°C for 48 h had little adverse effect on the detection of influenza virus type A, but diagnosis of influenza virus type B is best carried out with fresh specimens. The test detected a range of human and animal influenza virus A subtypes, including the H5N1 and H9N2 viruses that recently caused human disease in Hong Kong.
Influenza causes frequent human epidemics associated with significant morbidity and mortality in the elderly and in those with underlying risk factors, e.g., cardiorespiratory, renal, and metabolic diseases and the immunocompromised (2). Even in healthy subjects, influenza can lead to significant morbidity and economic loss through medical costs and time off work. Three times during the last century, a new virus subtype emerged in the human population, leading to a pandemic associated with even greater morbidity and mortality.
It is increasingly true that specific diagnosis of acute viral diseases impacts on individual patient care decisions, including infection control, use of antiviral therapy, and other aspects of clinical management. For such purposes, rapid diagnosis is critical. For hospitalized children (19) and adults (1), rapid diagnosis of respiratory viral infections reduced hospital stay and antibiotic use and was cost-effective.
Amantadine and rimantadine, the previously available antiviral drugs for influenza, are effective only against influenza virus type A. With the recent availability of neuraminidase inhibitor drugs with activity against both influenza type A and B viruses, there is an increased need for rapid diagnosis of influenza virus type B as well as influenza virus type A. These drugs have greatest efficacy when used early in the illness, and point of care tests usable in a general-practice setting would be desirable. While rapid diagnosis of the individual patient may not always be possible, real-time knowledge of virus activity in the community is also useful for patient management. A clinical diagnosis of influenza has reasonable positive predictive value in healthy adults during periods of influenza activity in the community (12), but at other times and in other age groups, clinical diagnosis is unreliable. There is thus a need for rapid, reliable, inexpensive tests for detecting both influenza virus subtypes A and B which can be performed by nonexperts.
Directigen Flu-A (Becton Dickinson) has been widely used for the rapid diagnosis of influenza virus type A and has been shown to be sensitive and specific (16, 18). In this paper, we evaluate a new version of this test, Directigen FluA+B, that detects influenza type A and B viruses and discriminates between them.
The influenza type A viruses of human, avian, and porcine origin and human influenza type B viruses used for testing the reactivity profile of the Directigen FluA+B test in this study are listed in Table Table1.1. They were obtained from the repository of viruses at the Department of Microbiology at the University of Hong Kong; David Swayne, Southeast Poultry Research Laboratory, Athens, Ga.; and the American Type Culture Collection, Rockville, Md. In addition to those listed in Table Table1,1, 18 other contemporary human influenza viruses, 6 each of influenza virus A subtypes H1N1 and H3N2 and influenza virus type B isolated in Hong Kong between 1998 and 2000, were also tested.
A panel of 90 other microorganisms, including 58 bacteria, 2 yeasts, and 30 viruses other than influenza virus type A or B, were similarly tested for cross-reactivity by the Directigen FluA+B assay. Bacteria and yeasts were cultured and resuspended in Dulbecco's phosphate-buffered saline at concentrations of 107 to 108 CFU/ml or, in the case of Mycoplasma pneumoniae, at a concentration of 107 to 108 color-changing units/ml. Viruses tested included coxsackievirus groups A and B, coronavirus, echovirus, rhinoviruses, adenoviruses, herpes simplex virus, varicella-zoster virus, cytomegalovirus, respiratory syncytial virus, measles virus, and mumps virus and were tested at infectivity titers between 104 to 1010 50% tissue culture infectious doses (TCID50)/ml. Influenza type C virus was tested at a titer of 1.6 × 1010 50% chicken embryo infectious doses (CEID50)/ml. The Directigen FluA+B testing was carried out with randomized samples blinded to the nature of the pathogen.
The analytical sensitivity was assessed for seven human influenza virus type A and six influenza virus type B isolates. Twofold dilutions of each virus preparation were made in virus transport medium, and three replicates for each dilution were tested by the Directigen FluA+B test. The range of dilutions was sufficiently broad that at least 1 dilution yielded negative results and at least 2 dilutions yielded positive results for each of the three replicates for each dilution. The analytical sensitivity was the lowest virus titer (CEID50) detectable by the Directigen FluA+B test.
Nasopharyngeal aspirate specimens from 250 patients submitted to the Department of Microbiology at Queen Mary Hospital, Hong Kong, for routine diagnosis of respiratory viruses during March and April 2000 were used for this evaluation. A catheter was inserted into the nasopharynx via each nostril in turn, and secretions were aspirated into a mucus trap. Subsequently, approximately 2 ml of virus transport medium (Earle's balanced salt solution with bovine serum albumin and antibiotics) was aspirated to wash residual secretions from the catheter into the mucus trap and to provide stability for virus in transit. The specimens were collected at time of admission and held at 4 to 8°C (for periods between 1 and 16 h, depending on time of admission) until they were transported to the laboratory. Specimens collected more than 24 h prior to arrival in the laboratory were excluded from the study.
Transport to the laboratory was done at ambient temperature, the duration of exposure to ambient temperature being less than 3 h at maximum. Aliquots of the nasopharyngeal aspirate suspensions were taken for immunofluorescent antigen detection of respiratory viruses and for virus culture. A third aliquot was tested by the Directigen FluA+B assay, and a fourth aliquot was frozen at −70°C for resolution of discrepant results. Additional aliquots were kept at 4°C or at −20°C for evaluating Directigen FluA+B test performance on stored specimens. The study protocol was approved by the Ethics Committee of the University of Hong Kong.
Madin-Darby canine kidney (MDCK) cell monolayers in culture tubes were inoculated with 200 μl of the nasopharyngeal aspirate-virus transport medium suspension and incubated at 35°C for 1 h, and the cells were fed with 1.5 ml of serum-free minimum essential medium containing TPCK (tosylsulfonyl phenylalanyl chloromethyl ketone)-treated trypsin (2 μg/ml) (Sigma, St. Louis, Mo.) and antibiotics. The cultures were examined for cytopathic effect (CPE) daily for 14 days. At the end of the incubation period, or when CPE appeared, the cells were scraped off, and smears on Teflon-coated slides were fixed and immunostained for viral antigens (see below).
MDCK cells were subcultured only for 20 passages. A quality control panel of virus stock at different dilutions and negative controls was tested under code to check the sensitivity of the MDCK cell line in use.
The direct immunofluorescent antigen (DFA) test was carried out with nasopharyngeal aspirate specimens. The nasopharyngeal aspirate was resuspended in virus transport medium and centrifuged, and the cell pellet was resuspended in phosphate-buffered saline and washed by repeated cycles of centrifugation until free of visible mucus. The cell pellet was then spotted on 6-mm wells of Teflon-coated slides, air dried, and fixed in ice-cold acetone for 10 min. The smears were stained with Imagen respiratory screen influenza virus type A and B reagents (Dako, Glostrup, Denmark) and viewed at a magnification of ×400 under epifluorescent illumination using the fluorescein isothiocyanate (FITC) filter of a Nikon fluorescent microscope. If a specimen with <20 columnar epithelial cells in the nasopharyngeal aspirate smear was negative by immunofluorescence, the specimen was considered to have insufficient respiratory epithelial cells, and such specimens were reported as indeterminate.
The Directigen FluA+B assay was carried out according to the manufacturer's instructions with the fresh clinical specimens. The tests were carried out by one of two operators who had been trained in the protocol and were done blinded to the results of the immunofluorescence and culture results. In brief, 200 μl of each nasopharyngeal aspirate was mixed gently with 8 drops (approximately 120 μl) of extraction buffer into the tube provided. Four drops (approximately 60 μl) of the specimen extract was then added to each well of the test device. Subsequently, specific conjugate, washing buffer, and substrate solutions were added within a 10-min period. The results were read at 5 min, the stop solution was added, and the test result was read again. The control dot needed to be visible (unless obscured by an intense purple triangle) for a valid test, and if absent, the result was regarded as indeterminate. A purple triangle was required for a positive result, and the intensity of the reaction was scored as 0.5 (very faint purple triangle), 1 (faint triangle), 2 (moderate triangle) (all with a visible central dot), 3 (intense purple triangle with the purple dot still partly visible), or 4 (intense purple triangle obscuring the central purple dot).
Frozen aliquots of specimens giving discrepant results were sent to a reference laboratory, Prodesse, Inc. (Waukesha, Wis.), for reverse transcription (RT)-PCR analysis of influenza virus type A and influenza virus type B (8).
All influenza virus type A or B culture-positive specimens were regarded as true positives. Those specimens that were culture negative but positive by either of the antigen detection assays (Directigen FluA+B or DFA) underwent RT-PCR testing for influenza virus types A and B of the replicate frozen specimen. The RT-PCR result was used to determine whether these specimens were true positives or false positives. Sensitivity, specificity, and positive and negative predictive values were calculated from 2-by-2 contingency tables.
A range of human influenza type A and B viruses and animal influenza type A viruses were tested by the Directigen FluA+B assay to confirm its reaction with viruses of diverse subtypes (Table (Table1).1). The viral infectivity titers of the preparations used for testing ranged from 104 to 107 TCID50 per 100 μl. Eighteen contemporary human influenza isolates, 6 each of influenza virus type A H3N2, influenza virus type A H1N1, and influenza virus type B, and 21 other influenza viruses (Table (Table1)1) isolated between 1934 and the present were studied. Influenza type A viruses of subtypes H5N1 and H9N2 were those isolated from humans in Hong Kong in 1997 and 1999, respectively (14, 20). In addition, animal influenza type A viruses spanning a range of subtypes (Table (Table1)1) were investigated. All the human and animal influenza type A viruses gave positive reactions for influenza virus type A but were negative for influenza virus type B. Conversely, 12 human influenza type B viruses, 6 of them isolated between 1940 and 1972 (Table (Table1)1) and 6 other recent isolates from Hong Kong, gave positive reactions for influenza virus type B and negative reactions for influenza virus type A.
The analytical sensitivity of the Directigen FluA+B test was established using seven human influenza virus type A (subtype H1N1 and H3N2) and six influenza virus type B strains. The limit of detection of the Directigen FluA+B test ranged between 6.5 × 101 and 3.3 × 104 CEID50/ml for influenza virus type A viruses. The detection limit for influenza virus type B viruses ranged between 4.6 × 101 and 2.5 × 103/ml for influenza virus type B, with one outlying result for B/Lee/40 (detection limit, 1.2 × 106/ml).
A wide range of other viruses, bacteria, and yeasts (see Materials and Methods) tested for cross-reactivity were negative by both influenza virus type A and B tests in the Directigen FluA+B test.
The patient group studied was predominantly children. The age distribution was as follows: 111 patients were <2 years old, 89 were 2 to 6 years old, 13 were between 7 and 11 years old, 6 were between 12 and 16 years old, 12 were between 16 and 55 years, 18 were over 55 years old, and for 1 patient, information on age and sex was unavailable. The male-female ratio was 149:100.
Fifty-four of the 250 clinical specimens examined were positive for influenza virus by culture, 22 being influenza virus type A and 32 influenza virus type B. Seventeen of the influenza virus type A isolates were of subtype H3N2 and 5 were H1N1. The performance characteristics of the Directigen FluA+B assay compared to culture are shown in Table Table2.2. All 22 influenza virus type A specimens positive by culture and 28 of the 32 influenza virus type B-positive specimens reacted positively to influenza virus type A and B, respectively, in the Directigen FluA+B test. Three specimens that were influenza virus type A positive and seven that were influenza virus type B positive by the Directigen FluA+B assay were culture negative. Furthermore, four specimens that yielded influenza virus type B in culture were negative by the Directigen FluA+B test.
In comparison to culture, the Directigen FluA+B had a sensitivity of 100%, specificity of 98.7%, positive predictive value of 88%, and negative predictive value of 100% for influenza virus type A. In comparison, the DFA test had sensitivity, specificity, and positive and negative predictive values of 100%, 95.6%, 88%, and 100%, respectively. Similarly, for influenza virus type B, the sensitivity, specificity, and positive and negative predictive values for the Directigen FluA+B tests were 87.5%, 96.8%, 80%, and 98.1%, respectively, and for DFA they were 65.6%, 97.2%, 100%, and 95.5%, respectively.
The spare aliquot of specimens that were culture negative but positive for either virus by the Directigen FluA+B or DFA test were tested by RT-PCR for the presence of influenza virus type A and B viral RNA. Of four culture-negative specimens that were influenza virus type A positive by one or both of the antigen detection tests, three were confirmed by RT-PR to be “resolved true-positive” results. Of these, the two that were Directigen FluA+B positive were scored as weakly positive reactions (intensity, 0.5). The one specimen that was not confirmed by RT-PCR was positive by Directigen FluA+B alone (intensity of reaction, 0.5) and was regarded as a false-positive result (Table (Table3).3). In comparison, of the 24 Directigen FluA+B-positive results confirmed (by culture or RT-PCR) to be true positives, 21 had reaction intensity scores of 2 to 4, 1 had a reaction intensity score of 1, and only 2 had a reaction intensity score of 0.5.
None of seven culture-negative, Directigen FluA+B influenza virus type B-positive specimens was RT-PR positive, and the seven were regarded as false-positive enzyme immunoassay (EIA) results (Table (Table4).4). All of them had weak (0.5) reaction intensity. In comparison, of the 28 influenza virus type B-positive specimens that were reactive in the Directigen FluA+B assay, 24 had reaction intensities of 2 or greater, 2 had reaction intensities of 1, and 2 had a reaction intensity of 0.5.
In addition, three randomly selected specimens that were culture and Directigen FluA+B positive for influenza virus types A and B and 6 of 185 specimens negative by culture and antigen detection tests for both viruses were also tested by RT-PCR. The RT-PCR result confirmed the Directigen FluA+B result in each instance.
After the resolution of discrepant results (Tables (Tables55 and and6),6), the sensitivity, specificity, positive predictive value and negative predictive value for diagnosis of influenza virus type A of the Directigen FluA+B assay were 96%, 99.6%, 96%, and 99.6%, respectively, and they were 100% for DFA for all parameters. In contrast, the sensitivity of culture for influenza virus type A was only 88%. For influenza virus type B, the sensitivity, specificity, positive predictive value, and negative predictive value were 87.5%, 96.8%, 80%, and 98%, respectively, for the Directigen FluA+B and 65.6%, 100%, 100%, and 95.2%, respectively, for the DFA assay. Thus, the Directigen FluA+B assay had better sensitivity than DFA for detection of influenza virus type B but led to more false-positive results.
Other viruses, including 28 respiratory syncytial viruses, 10 adenoviruses, and 4 parainfluenza type 3 viruses were identified in 51 of the nasopharyngeal specimens. They all gave negative results with the Directigen FluA+B assay, with the exception of two specimens that were reactive for influenza virus type B in the Directigen FluA+B test. One specimen had a culture-confirmed double infection with influenza virus type B and adenovirus type 7, while the other was one of the weak (intensity, 0.5) influenza virus type B false-positive reactions noted above.
Specimen stability was evaluated by testing replicate aliquots of specimens in virus transport medium on the day of collection and after storage for 48 and 72 h at 2 to 8°C and after freezing at −20°C for 1 week. The specimens selected for stability testing were among the 250 specimens recruited for the rest of the study. Five influenza virus type A-positive specimens remained unchanged after storage for up to 72 h at 2 to 8°C. All 10 negative specimens remained negative after storage for 48 h at 2 to 8°C. While 9 of 10 negative specimens remained negative when stored for 72 h at 2 to 8°C, 1 gave an uninterpretable result. Of 17 influenza virus type A-negative specimens retested after 1 week at −20°C, 16 remained negative but 1 gave a false-positive result, the intensity score being 0.5.
Three of four influenza virus type B-positive specimens remained positive after storage for 48 or 72 h at 2 to 8°C. One specimen gave a false-negative result after 48 h, and another gave an uninterpretable result at 72 h. Eleven influenza virus type B-negative specimens remained negative after storage for 48 or 72 h at 2 to 8°C. Of 11 influenza virus type B-positive specimens that were stored frozen for 1 week, 9 remained positive, 1 gave an uninterpretable result, and 1 was false negative. All 11 influenza virus type B-negative specimens remained negative after storage for 48 or 72 h at 2 to 8°C. Ten of 11 remained negative after 1 week of storage frozen, while one specimen gave an uninterpretable result.
The intensity of the Directigen FluA+B reaction in positive specimens correlated with the time taken for the appearance of CPE in the cell culture tubes. Influenza virus type B-positive specimens giving a Directigen FluA+B intensity of 4 resulted in CPE in cell culture at a median of 2 days, while specimens with weaker reactions (0.5 to 3) took longer (median, 5 days; P = 0.0001, Mann-Whitney test). With influenza virus type A, Directigen FluA+B scores of 4 were associated with CPE with a median of 5 days, compared to 10.5 days in those specimens giving a weaker reaction (P = 0.12). The time required for the appearance of CPE is probably related to virus load in the clinical specimen.
The Directigen FluA+B test gave excellent results for the diagnosis of influenza virus type A infection when applied to nasopharyngeal aspirates from hospitalized patients. The patient population studied was predominantly (88%) pediatric, 80% of them being under 2 years of age. The performance of the test was comparable to DFA and superior to cell culture. The previous version of this assay, Directigen Flu-A, used for detection of influenza virus type A alone, was reported to give similar sensitivity with nasopharyngeal washes (18). In experimental human infections, nasopharyngeal washes contained approximately 1.8 log10 and 3 log10 more virus than nasopharyngeal swabs and throat swabs, respectively. It is therefore not surprising that previous studies of rapid diagnostic assays (including the Directigen FluA test) for detection of influenza virus type A suggested that nasopharyngeal aspirates or washes were superior to nasopharyngeal swabs or throat swabs for detection of influenza virus type A (10, 16).
In addition, children with influenza virus type A infections have higher viral loads in the nasopharynx than older patients. Therefore, it is relevant to note that the excellent sensitivity for influenza virus type A detection found in the present study with Directigen FluA+B was obtained with the best type of specimen and in a group of patients who were largely children. However, previous studies have shown that good results with Directigen Flu A or DFA are obtainable with nasopharyngeal swabs from elderly patients as well, providing attention is paid to the collection of good clinical specimens (11). It is noted that throat swabs are likely to be much poorer specimens (7). In the present study, the performance of the new Directigen FluA+B assay on nasopharyngeal swabs or on throat swabs was not evaluated.
The Directigen FluA+B assay had lower sensitivity (87.5%) for the diagnosis of influenza virus type B but performed better in this respect than DFA (sensitivity, 65.6%). On the other hand, there were false-positive influenza virus type B results in the Directigen FluA+B assay tests which compromised its specificity and positive predictive value. It is notable that all seven false-positive results had a weak intensity reaction (score of 0.5), while 93% of the true influenza virus type B-positive Directigen FluA+B tests gave a reaction intensity of 1 or higher. This suggests that weakly reactive influenza virus type B results need to be regarded with caution.
In comparison with influenza virus type A and respiratory syncytial virus, there is less information on the performance of rapid diagnostic tests for influenza virus type B, but the available information indicates that the sensitivity of currently available assays for influenza virus type B is poor (15). It is notable that even tests based on alternative strategies, such as detection of viral neuraminidase activity (Zstat Flu), appear to be less sensitive for influenza virus type B than for influenza virus type A (13). The analytical sensitivity of the Directigen FluA+B test for detection of influenza virus type A and B viruses was similar. In this regard, it is relevant to ask if the lower sensitivity of rapid diagnostic tests for influenza virus type B is due to lower viral loads of influenza virus type B viruses in the nasopharynx than of influenza virus type A. Available data from experimental infection of humans with influenza virus type A (7) and B (9) viruses point in this direction. However, a recent study quantifying viral RNA load using real-time quantitative PCR in a small number of patients failed to document a clear difference between influenza virus type A- and influenza virus type B-infected patients (17).
Storage of the clinical specimen at 2 to 8°C for up to 72 h had little deleterious effect on the detection of influenza virus type A by the Directigen FluA+B test. The limited study of influenza virus type B-positive specimens suggests that storage may adversely affect the performance of influenza virus type B diagnosis.
Cross-reaction with other bacteria, fungi, and viruses that may be encountered in the respiratory tract was excluded by testing a range of laboratory isolates by the Directigen FluA+B test. Furthermore, clinical specimens with respiratory viruses other than influenza virus were negative by the Directigen FluA+B assay.
The reactivity of the Directigen FluA+B assay was assessed on representative human and animal influenza viruses. Human H3N2, H1N1, and H2N2 subtypes of influenza virus type A and influenza virus type B isolated from 1934 to 2000, including contemporary influenza virus type A (subtypes H3N2 and H1N1) and influenza virus type B viruses, were all detectable by the Directigen FluA+B test. The H5N1 subtype of influenza virus type A isolated from humans during the avian flu incident in Hong Kong during 1997 (20) and H9N2 viruses isolated from humans in 1999 (14) were also detected, as were a range of other avian and porcine influenza virus subtypes. These included avian viruses of subtypes H9N2 and H6N1, sharing internal genes with the H5N1 viruses causing severe human disease during 1997 (5). The previous version of the assay (Directigen Flu-A) was successfully used for rapid diagnosis of influenza in humans during the avian influenza incident (20). It is important for influenza pandemic preparedness to establish that new influenza virus type A diagnostic assays are able to detect a range of virus subtypes, particularly those with pandemic potential. The previous Directigen Flu-A test was also shown to be useful in the diagnosis of influenza outbreaks in horses and poultry (3, 4).
The Directigen FluA+B assay was convenient to perform, can be completed in approximately 15 min, and does not require the presence of an expert technologist. It joins other options currently available for rapid diagnosis of influenza virus types A and B, including the detection of viral antigen by immunofluorescence (DFA), the neuraminidase detection assay (Zstat Flu; ZymeTx, Oklahoma City, Okla.), and the FLU-OIA assay (Biostar, Inc., Boulder, Colo.). Unlike the Directigen FluA+B and DFA tests, the Zstat Flu (ZymeTx) and FLU-OIA (Biostar) tests allow an influenza virus infection to be diagnosed but do not differentiate between influenza virus types A and B. This may be relevant if a treatment choice between an adamantine (amantadine or rimantadine) and the new neuramindase inhibitors is under consideration, since the former are cheaper but not active against influenza virus type B.
The DFA test can be used for the diagnosis of a number of other respiratory viruses (respiratory syncytial virus, adenovirus, and parainfluenza viruses) in addition to influenza virus during the same procedure but is a labor-intensive and subjective test that requires expertise for reliable interpretation. Reports of its sensitivity for detection of influenza virus type A (less data are available for type B) vary widely, depending on the type and quality of specimen, the reagents used, and the expertise of the technologist in interpreting the results. In the present study, the performance of DFA for diagnosis of influenza virus type A was excellent and similar to the Directigen FluA+B assay. However, as with other rapid tests, the performance of DFA has been less satisfactory for diagnosing influenza virus type B.
RT-PCR offers an alternative method of diagnosing influenza virus infections. It potentially has high sensitivity and specificity (6), but requires a high level of skill and complex laboratory infrastructure and takes several hours to perform. It does not fill the niche of tests that are rapid and easy to perform with a lower level of expertise.
In conclusion, the Directigen FluA+B assay is a convenient and relatively simple test for the diagnosis of influenza virus type A and influenza virus type B infections. Its ability to detect influenza virus type A on nasopharyngeal aspirate specimens is excellent. The sensitivity and specificity of the assay for detecting influenza virus type B is comparable to other options, e.g., DFA, for rapid diagnosis of this infection. The patient population studied was predominantly children, and the performance of this test with older patients and other specimen types needs separate evaluation.
We thank David Swayne, Southeast Poultry Research Laboratory, Athens, Ga., for providing some of the strains used in this study and Kelly Henrickson from Prodesse, Inc., Waukesha, Wis., for providing RT-PCR testing. We acknowledge the excellent technical assistance of K. M. Lo and M. K. Yan.
This evaluation was supported by a research grant from the Wellcome Trust (057476/Z/099/Z) and a grant from BD Diagnostic Systems, Sparks, Md.