a Chemiluminescent Rapid Diagnostic Test for Inffuenza Virus

3 downloads 168 Views 147KB Size Report
and Becton Dickinson's Directigen Flu A. B test was performed in a rapid- turnaround laboratory. Residual specimen was used for culture and the Zstat-.
JOURNAL OF CLINICAL MICROBIOLOGY, July 2002, p. 2331–2334 0095-1137/02/$04.00⫹0 DOI: 10.1128/JCM.40.7.2331–2334.2002 Copyright © 2002, American Society for Microbiology. All Rights Reserved.

Vol. 40, No. 7

Clinical Evaluation of the ZstatFlu-II Test: a Chemiluminescent Rapid Diagnostic Test for Influenza Virus Marilyn S. Hamilton,1,2* David M. Abel,2 Yolanda J. Ballam,2 Mary K. Otto,2 Angela F. Nickell,2 Lisa M. Pence,3† James R. Appleman,3‡ Craig D. Shimasaki,3 and Komandoor E. Achyuthan3 University of Missouri—Kansas City School of Medicine1 and The Children’s Mercy Hospital,2 Kansas City, Missouri, and ZymeTx, Inc., Oklahoma City, Oklahoma3 Received 15 January 2002/Returned for modification 1 March 2002/Accepted 5 April 2002

Exploiting the high sensitivity of the chemiluminescence phenomenon, an accurate and sensitive point-ofcare test, called the ZstatFlu-II test (ZymeTx, Inc., Oklahoma City, Okla.), was developed to detect influenza virus infections. The ZstatFlu-II test takes 20 min and requires approximately 2 min of “hands-on” time for operational steps. The ZstatFlu-II test does not distinguish between infections with influenza virus types A and B. ZstatFlu-II test results are printed on Polaroid High-Speed Detector Film, allowing test results to be archived. A prototype version of the ZstatFlu-II test was evaluated during the 2000-to-2001 flu season with 300 nasal aspirate specimens from children at a pediatric hospital. Compared to culture, the ZstatFlu-II test had 88% sensitivity and 92% specificity. The Directigen test had a sensitivity of 75% and a specificity of 93%. The sensitivity of the ZstatFlu-II test was significantly higher than that of the Directigen test (P < 0.0574). ated nasal pharyngeal cells could be done in a few hours but require a high level of expertise (10). Finally, rapid, point-ofcare tests are available to detect the influenza virus (3, 12, 20, 23). All of these tests use an enzyme immunoassay directed at antigens of the viruses (3, 12, 23), with the exception of the ZstatFlu test, which detects influenza virus neuraminidase enzymatic activity (18, 20). A new approach utilized in the prototype ZstatFlu-II test by ZymeTx, Inc. (Oklahoma City, Okla.), detects, but does not distinguish, influenza virus types A and B, using chemiluminescence resulting from the hydrolysis of a synthetic chemiluminescent substrate by the action of influenza virus neuraminidase. This approach for influenza virus detection is similar to that of the previously reported ZstatFlu test, which employed a chromogenic substrate specific for influenza virus neuraminidase (18, 20). In chemistry analyzers, chemiluminescence is generally accepted as being more sensitive than enzyme immunoassays (14). Chemiluminescence provides an opportunity for increasing the sensitivity of rapid tests (1, 4). The purpose of this study was to describe the performance characteristics of the prototype test, the ZstatFlu-II test, compared to those of culture and the Directigen Flu A ⫹ B test (Becton Dickinson, Cockeysville, Md.).

The annual economic cost of influenza disease in the United States has been estimated at $3 billion to $5 billion (22). The Centers for Disease Control and Prevention reported in 2001 that influenza was associated with about 20,000 deaths nationwide and more than 100,000 hospitalizations (http://www.cdc .gov/ncidod/diseases/flu/fluinfo.htm). Worldwide estimates were considerably higher (16, 17, 24). Sensitive, specific, and rapid tests for influenza will greatly improve patient health care and reduce costs so that only “flu-positive” patients receive the recently approved antiviral treatments (9, 21). Rapid diagnostics for influenza will also prevent the misuse of antibiotics to “treat” the flu. Influenza disease is caused by influenza virus types A and B. Influenza virus types A and B are Orthomyxoviridae, characterized by the presence of an envelope penetrated by glycoprotein spikes with hemagglutinating and neuraminidase activities. Influenza virus also contains matrix protein, nucleoprotein, and three proteins with polymerase activity and a segmented negative-strand RNA genome (6, 19). These viruses are responsible for winter epidemics of respiratory illness in which the rates of infection are highest among children (5). The shared cardinal sign of fever without localization makes differentiation of influenza from sepsis necessary for proper patient management. Delay in this differentiation could result in unnecessary laboratory testing and treatment for possible bacterial “sepsis” with unnecessary antibiotics (25). Three diagnostic methods for respiratory secretions are in common use. Culture, both shell vial and tube, takes several days. Direct or indirect immunofluorescence assays on exfoli-

MATERIALS AND METHODS Patient samples. This study was determined to be exempt from review by the University of Missouri—Kansas City Health Sciences Pediatric Institutional Review Board. During the 2000-to-2001 flu season, between January 19, 2001 and March 1, 2001, 300 nasal aspirate specimens from children were tested in a pediatric hospital laboratory. The clinician had requested rapid influenza testing, and Becton Dickinson’s Directigen Flu A ⫹ B test was performed in a rapidturnaround laboratory. Residual specimen was used for culture and the ZstatFlu-II test. These additional tests were completed within 24 h of collection, with storage of patient specimens at 4°C until test time. Laboratory methods. (i) Rapid tests. Two rapid tests were used. The reported result was based on Becton Dickinson’s Directigen Flu A ⫹ B test, performed according to the manufacturer’s instructions. Briefly, the influenza virus A or B antigens were extracted from 200 ␮l of the respiratory specimen and the extract was expelled through a filter assembly into the test wells of the test device.

* Corresponding author. Mailing address: The Children’s Mercy Hospital, Department of Pathology, 2401 Gillham Rd., Kansas City, MO 64108. Phone: (816)-234-3234. Fax: (816) 802-1492. E-mail: [email protected]. † Present address: National Center for Toxicological Research, Jefferson, AR 72079. ‡ Present address: Anadys Pharmaceuticals, Inc., San Diego, CA 92121. 2331

2332

J. CLIN. MICROBIOL.

HAMILTON ET AL.

FIG. 1. Photographic recording of patient results in the ZstatFlu-II test. These photographs are representative examples of a strong positive test result (left), a weak positive test result (center) and a negative test result (right), obtained by administering the ZstatFlu-II test to three different patients. In order to protect confidentiality, the patient specimens were identified only serially in the photographs, by the order in which the various specimens were obtained or tested (115, 88, and 3, respectively). The margins of the Polaroid photographs provided ample surface to record such additional test-related information as test date and time, test result, and investigator identification, as shown in this figure. The initials that appear on these photographs belong to the investigators who performed the tests, Y.J.B. and M.K.O. (both coauthors of this work). Antigen captured on the device membrane was visualized with an enzymeconjugated monoclonal antibody chromogen system. Positive and negative controls are built into each test device. The experimental test was the ZstatFlu-II test (ZymeTx, Inc.). In this prototype test, 0.9 ml of the respiratory specimen was added to a prefilter tube containing salts and detergent, which created an optimal milieu for influenza virus neuraminidase activity (2, 4,18). Clarification of the specimen occurred at the prefilter tip as the specimen was added to the bottom half of a proprietary two-chamber reaction vessel containing an influenza virus neuraminidase-specific substrate and two light enhancers. The top half of the reaction vessel contained a sodium hydroxide solution. A 15-min test period at “room temperature” (⬃25°C) allowed the influenza virus neuraminidase to act on the specific substrate molecule and released the nascent chemiluminescent reporter groups. At the end of 15 min, the entire reaction vessel was placed into position inside a turret within a specially designed Polaroid (Waltham, Mass.) imaging device. The cover of the imaging device was then closed. The closing action flushed the sodium hydroxide solution from the top chamber into the bottom chamber. Contact with sodium hydroxide terminated the virus neuraminidase reaction accompanied by chemiluminescence. Light from the reaction mixture strikes a Polaroid High-Speed Detector “instant” film housed inside the imaging device. The film was exposed to the chemiluminescence for 5 min. An influenza viruspositive specimen was detected by observation of a white image shaped like a plus sign against the black background of the Polaroid High-Speed Detector Film, which also provided a permanent record of the results. Positive and negative quality controls, which came with the prototype kit, were run daily. (ii) Culture. Viral cultures were performed in the Virology Laboratory of Children’s Mercy Hospital. Specifically, 0.1 to 0.2 ml of specimen was inoculated into tube cultures of human embryonic lung fibroblasts (MRC-5), human lung carcinoma cells (A-549), and rhesus monkey kidney (RMK) cells (Viromed, Minneapolis, Minn.), maintained in a roller drum apparatus at 37°C, and then observed for cytopathic effect for 14 days. Shell vial cultures of A-549 and RMK (Viromed) were screened on days 4 and 7 with a Chemicon (Temecula, Calif.) Respiratory Viral Screen. Virus-positive cultures were identified with a Viral Respiratory Kit from Bartel (Issaquah, Wash.). (iii) RT-PCR. Reverse transcriptase PCR (RT-PCR) was performed by Prodesse, Inc. (Milwaukee, Wis.), on available specimens that were negative in culture but positive by both rapid tests and therefore were considered possible positive specimens with discrepant results. Briefly, the Hexaplex RT-PCR assay (Prodesse, Inc.) involved extraction of viral genomic RNA from the test specimen followed by PCR amplification using the primers to seven upper-respiratory-tract viruses including influenza virus types A and B. PCR products were detected by using microplate wells coated with a streptavidin system as described previously (13). Specimens were coded and then sent for RT-PCR analyses. Investigators at Prodesse, Inc., were never informed of the results of the rapid

tests and viral culture, either before or subsequent to RT-PCR analyses of the specimens. Statistical analysis. The individual rapid tests were compared to culture and to “consensus.” In consensus, a specimen was considered positive if it was positive either in culture or in both rapid tests. Sensitivities, specificities, positive predictive values, and negative predictive values were calculated as follows: sensitivity ⫽ number of true positives/(number of true positives ⫹ number of false negatives); specificity ⫽ number of true negatives/(number of true negatives ⫹ number of false positives); positive predictive value ⫽ number of true positives/(number of true positives ⫹ number of false positives); and negative predictive value ⫽ number of true negatives/(number of true negatives ⫹ number of false negatives). Statistical analyses of the data were performed as described by Galen (8). Statistical significance was determined by McNemar’s test for paired binary data (7).

RESULTS Patient characteristics. The patient age range was 12 days to 19 years. The mean, median, and mode ages of the tested patients were 43, 20, and 2 months, respectively. There were 136 females and 164 males. Test results and performance characteristics. Representative examples of three different patients’ test results are shown in Fig. 1. These are, respectively, a strong positive test result, a weak positive test result, and a negative test result. A strong positive test result producing a bright image was easily discernible, whereas a weak positive test result was considerably less so. It should be noted that the clinical study resulted in positive test results over a range of brightness of images and that the examples shown in Fig. 1 are merely illustrative. The data underscore the importance of having an archival record of test results as shown on these Polaroid High-Speed Detector Film images. Table 1 shows the results of the two rapid detection tests and viral culture compared to viral culture and the consensus. Table 2 shows the performance characteristics of the two rapid tests and viral culture compared to culture and the consensus. Compared to culture, the sensitivity of the ZstatFlu-II test was

VOL. 40, 2002

ZstatFlu-II TEST FOR INFLUENZA VIRUS

TABLE 1. Results of rapid testing compared to culture and consensus No. of specimens with the indicated result by: Gold standard

ZstatFlu-II test

Directigen test

Positive

Negative

Positive

Negative

Culture Positive Negative

57 19

8 216

49 17

16 218

Consensus Positivea Negative

68 8

8 216

60 6

16 218

a

Culture Positive

Negative

65 0

11 224

significantly (P ⬍ 0.0574) higher than that of Becton Dickinson’s Directigen test. By comparison to consensus, there was no statistical difference. The specificities and positive and negative predictive values of the two rapid tests were not significantly different. Of the 300 specimens tested, the results of 255 (85%) were in complete agreement, with identical results in all three tests, namely, both rapid tests and culture. Of these, 210 specimens were negative and 45 were positive for influenza virus, with 13 positive for influenza virus type A and 32 positive for influenza virus type B. In all cases, identifications of influenza virus type A versus type B by culture and by Becton Dickinson’s Directigen test were in total agreement. Although the ZstatFlu-II test did not distinguish between influenza virus types A and B, the test identified both types of viruses with a sensitivity and a specificity comparable to those of culture (Table 2). Here, the identification of the virus types was based on the results from either viral culture or Becton Dickinson’s Directigen test. Of the 45 specimens without complete agreement, 20 were positive in culture and negative in rapid testing, 4 in both rapid tests, 4 in the ZstatFlu-II test alone, and 12 in the Becton Dickinson Directigen test alone. Eleven specimens were positive in both rapid tests but negative in culture. Of these 11 specimens, 8 were tested by RT-PCR and 7 were positive by this technique. Eight specimens were positive only by the ZstatFlu-II test, and six specimens were positive only in the Becton Dickinson Directigen test. Culture isolation of other respiratory viruses. No specimen was positive for both influenza virus types A and B. Fifty-six

TABLE 2. Performance characteristics of rapid testing compared to culture and consensus

Culture

Rapid test

Sensitivity (%) a

Specificity (%)

Predictive value (%) Positive

Negative

ZstatFlu-II test Directigen test

88 75

92 93

75 74

96 93

Consensusb ZstatFlu-II test Directigen test Culture

89 79 86

96 97 100

89 91 100

96 93 95

a

P ⬍ 0.0574 compared to the Directigen test. A specimen positive in culture or in both rapid tests is considered positive by consensus. b

specimens were positive for respiratory viruses other than influenza virus type A or B alone: 43 specimens were positive for respiratory syncytial virus (RSV), 5 were positive for adenovirus, 3 were positive for parainfluenza-3 virus, and 1 specimen each was positive for parainfluenza-2 virus, parainfluenza-4 virus, adenovirus plus parainfluenza-3 virus, influenza virus type A plus RSV, and influenza virus type B plus adenovirus. DISCUSSION

Positive either by culture or by both rapid tests.

Gold standard

2333

We found that the ZstatFlu-II test, compared to culture, was more sensitive than Becton Dickinson’s Directigen test. This could be due to the following two differences in the tests or to a combination of the differences. The ZstatFlu-II test is based on the functional activity of the influenza virus neuraminidase. Becton Dickinson’s Directigen test, on the other hand, is dependent on monoclonal antibodies binding to extracted viral antigens. The ZstatFlu-II test detects chemiluminescence using Polaroid High-Speed Detector Film. Becton Dickinson’s Directigen test is based on visual detection of a chromogen that yields a purple signal on a white background. It is not possible to distinguish the relative contributions of these two parameters to the increased sensitivity of the ZstatFlu-II test. However, it is commonly accepted by chemistry analyzers that chemiluminescence is more sensitive than enzyme immunoassays (14). This suggests that the detection method contributed to the increased sensitivity of the ZstatFlu-II test (1, 4). None of the rapid influenza diagnostic tests currently on the market provide archiving capabilities for the test results. The ZstatFlu-II test described here has the added advantage of providing a permanent record of the test result as a Polaroid “picture.” This could be reviewed as necessary if the results are judged questionable. The Polaroid photograph containing the patient test result could also be included in the patient’s records for documentation. The margins of the Polaroid photograph provide ample surface on which to record such additional test-related information as test date and time, reported result, and patient identification (Fig. 1). The ZstatFlu-II test did not detect several other upper respiratory disease-causing viruses that were present in patient specimens such as the parainfluenza viruses, adenovirus, and RSV. Some of these related viruses, such as parainfluenza viruses, also express significant neuraminidase activity (2). In our testing, the chemiluminescent substrate molecule, which is at the core of the ZstatFlu-II test technology, was specific for the influenza virus neuraminidases. The data are consistent with our earlier results and those of others obtained from using a highly specific chromogenic substrate molecule for influenza virus neuraminidase (18, 20). Experienced technologists found the prototype ZstatFlu-II test awkward but identified simple improvements that have been implemented in later models. Unlike Becton Dickinson’s Directigen test, the ZstatFlu-II test does not have internal positive and negative controls. If a test has internal positive and negative controls, the College of American Pathologists Laboratory Accreditation Program Microbiology Checklist, question MIC.61385, requires that external controls be run only on each shipment of each lot of tests. Becton Dickinson’s Directigen test meets this requirement. On the other hand, Checklist question MIC.61400 requires that external controls

2334

HAMILTON ET AL.

be run daily if the test does not have internal controls. The ZstatFlu-II test is in this category. The added requirement for controls could be a burden in some situations. It should also be emphasized that as of this writing, the ZstatFlu-II test has not gone through a 510(k) filing process with the Food and Drug Administration (FDA) and is therefore not an FDA-approved test for influenza. The study reported here was based on nasal aspirate specimens. This is the best specimen available for both of the rapid tests (11, 15). Becton Dickinson’s Directigen test does permit the use of throat swabs (11, 15). Testing based on throat swabs was less sensitive (11, 15). How the ZstatFlu-II test would function with a nasal swab or a throat swab specimen is not known at this time. The inability of the ZstatFlu-II test to distinguish influenza virus type A from type B may or may not be a problem. Amantadine will treat only influenza virus type A, but the newer treatments are effective for both types of influenza viruses (9, 21). In the latter case, the lack of information regarding the precise type of influenza virus infection is not critical for treatment purposes. Treatment decisions are based more on symptomology than on virus type identification but must be initiated early in the disease course to be effective. From a treatment standpoint, a rapid test that could identify the disease earliest after onset would be the best. This aspect was not determined in this study. In summary, the ZstatFlu-II test provides an alternative rapid test technology for influenza virus types A and B which, compared to culture, is more sensitive than Becton Dickinson’s Directigen test and has the added advantage of providing a permanent record of the test result. Chemiluminescence technology has the potential to increase the sensitivity of detection systems used in rapid tests. ACKNOWLEDGMENTS The clinical trial cost was defrayed by a grant from ZymeTx, Inc. Development of the ZstatFlu-II test was funded in part by grants from the Oklahoma Center for the Advancement of Science and Technology (OCAST) awarded to C. D. Shimasaki (AR982-026) and to K. E. Achyuthan (HN5-039 and AR021-004). We gratefully acknowledge Don Mauchan, Paul Nangeroni, Dan Mantell, Ed Gaffey, Paul Graham, and Leonard Aberbach (Polaroid Corporation) for supplying the imaging devices and the Polaroid HighSpeed Detector Film. Stephen Simon (Children’s Mercy Hospital, Kansas City, Mo.) helped with statistical analyses. Roberta Allen and Robyn Cannedy provided technical assistance during the development phase of the ZstatFlu-II test. REFERENCES 1. Achyuthan, K. E. 1998. Enzymatic and kinetic properties of blood coagulation factor XIIIa and guinea pig liver transglutaminase utilizing 6-[N-(4aminobutyl)-N-ethylamino]-2,3-dihydrophthalazine-1,4-dione, as a novel, specific and sensitive chemiluminescent substrate. J. Biolumin. Chemilumin. 13:1–11.

J. CLIN. MICROBIOL. 2. Achyuthan, K. E., and A. M. Achyuthan. 2001. Comparative enzymology, biochemistry and pathophysiology of human exo-␣-sialidases (neuraminidases). Comp. Biochem. Physiol. B 129:29–64. 3. Boon, A. C., A. M. French, D. M. Fleming, and M. C. Zambon. 2001. Detection of influenza A subtypes in community-based surveillance. J. Med. Virol. 65:163–170. 4. Buxton, R. C., B. Edwards, R. R. Juo, J. C. Voyta, M. Tisdale, and R. C. Bethell. 2000. Development of a sensitive chemiluminescent neuraminidase assay for the determination of influenza virus susceptibility to zanamvir. Anal. Biochem. 280:291–300. 5. Centers for Disease Control and Prevention. 1999. Update: influenza activity—United States, 1998–99 season. Morb. Mortal. Wkly. Rep. 48:25–27. 6. Cox, N. J., and K. Subbarao. 1999. Influenza. Lancet 354:1277–1282. 7. Fleiss, J. L. 1981. Statistical methods for rates and proportions, 2nd ed., p 211–225. John Wiley & Sons, Inc., New York, N.Y. 8. Galen, R. S. 1979. The predictive value of laboratory testing. Orthop. Clin. N. Am. 10:287–297. 9. Griffin, A. D., A. S. Perry, and D. M. Fleming. 2001. Cost-effectiveness analysis of inhaled zanamvir in the treatment of influenza A and B in high-risk patients. Pharmacoeconomics 19:293–301. 10. Harmon, M. W. 1999. Influenza virus, p. 587–601. In E. H. Lennette and T. F. Smith (ed.), Laboratory diagnosis of viral infection, 3rd ed. Marcel Dekker, New York, N.Y. 11. Kaiser, L., M. S. Briones, and F. G. Hayden. 1999. Performance of virus isolation and Directigen Flu A to detect influenza A virus in experimental human infection. J. Clin. Virol. 14:191–197. 12. Kawakami, C., H. Shimizu, S. Watanabe, T. Munemura, K. Mitamura, N. Sugaya, and M. Imai. 2001. Evaluation of immunochromatography method for rapid detection of influenza A and B viruses. Kansenshogaku Zasshi 75:792–799. 13. Kehl, S. C., K. J. Henrickson, W. Hua, and J. Fan. 2001. Evaluation of the hexaplex assay for detection of respiratory viruses in children. J. Clin. Microbiol. 39:1696–1701. 14. Kricka, L. J., J. C. Voyta, and I. Bronstein. 2000. Chemiluminescent methods for detecting and quantitating enzyme activity. Methods Enzymol. 305:370– 390. 15. Landry, M. L., S. Cohen, and D. Ferguson. 2000. Impact of sample type on rapid detection of influenza virus A by Cytospin-enhanced immunofluorescence and membrane enzyme-linked immunosorbent assay. J. Clin. Microbiol. 38:429–430. 16. Laver, G., and E. Garman. 2001. The origin and control of pandemic influenza. Science 293:1776–1777. 17. Layne, S. P., T. J. Beugelsdijk, C. K. N. Patel, J. K. Taubenberger, N. J. Cox, I. D. Gust, A. J. Hay, M. Tashiro, and D. Lavanchy. 2001. A global lab against influenza. Science 293:1729. 18. Liav, A., J. A. Hansjergen, K. E. Achyuthan, and C. D. Shimasaki. 1999. Synthesis of bromoindolyl-4,7-di-O-methyl-Neu5Ac: specificity toward influenza A and B viruses. Carbohydr. Res. 317:198–203. 19. Mikulasova, A., E. Vareckova, and E. Fodor. 2000. Transcription and replication of the influenza A virus genome. Acta Virol. 44:273–282. 20. Noyola, D. E., B. Clark, F. T. O’Donnell, R. L. Atmar, J. Greer, and G. J. Demmler. 2000. 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. 38:1161–1165. 21. Peters, P. H., S. Gravenstein, P. Norwood, V. De Bock, A. Van Couter, M. Gibbens, T. A. von Planta, and P. Ward. 2001. Long-term use of oseltamivir for the prophylaxis of influenza in a vaccinated frail older population. J. Am. Geriatr. Soc. 49:1025–1031. 22. Schoenbaum, S. C. 1987. Economic impact of influenza—the individual’s perspective. Am. J. Med. 82:26–30. 23. Schultze, D., Y. Thomas, and W. Wunderli. 2001. Evaluation of an optical immunoassay for the rapid detection of influenza A and B viral antigens. Eur. J. Clin. Microbiol. Infect. Dis. 20:280–283. 24. Webster, R. G. 2001. A molecular whodunit. Science 293:1773–1775. 25. Woo, P. C. Y., S. S. Chiu, W.-H. Seto, and M. Peiris. 1997. Cost-effectiveness of rapid diagnosis of viral respiratory tract infections in pediatric patients. J. Clin. Microbiol. 35:1579–1581.