Molecular diagnosis of influenza

Rev. Med. Virol. 2002; 12: 375–389.Published online in Wiley InterScience (www.interscience.wiley.com).

Reviews in Medical Virology DOI: 10.1002/rmv.370

Molecular diagnosis of influenzaJoanna S. Ellis* and Maria C. ZambonRespiratory Virus Unit, Enteric, Respiratory and Neurological Virus Laboratory, Public Health LaboratoryService, Central Public Health Laboratory, 61 Colindale Avenue, Colindale, London NW9 5HT, UK

SUMMARY

The past decade has seen tremendous developments in molecular diagnostic techniques. In particular, thedevelopment of PCR technology has enabled rapid and sensitive viral diagnostic tests to influence patientmanagement. Molecular methods used directly on clinical material have an important role to play in the diagnosisand surveillance of influenza viruses. Molecular diagnostic tests that allow timely and accurate detection of influenzaare already implemented in many laboratories. The combination of automated purification of nucleic acids withreal-time PCR should enable even more rapid identification of viral pathogens such as influenza viruses in clinicalmaterial. The recent development of DNA microarrays to identify either multiple gene targets from a single pathogen,or multiple pathogens in a single sample has the capacity to transform influenza diagnosis. While molecular methodswill not replace cell culture for the provision of virus isolates for antigenic characterisation, they remain invaluable inassisting our understanding of the epidemiology of influenza viruses. Copyright # 2002 John Wiley & Sons, Ltd.

Accepted: 7 August 2002

INTRODUCTIONInfluenza viruses continue to be a major cause ofrespiratory tract infection, resulting in significantmorbidity, mortality and financial burden. Eachyear, increased hospitalisation rates and excessdeaths are attributable to influenza infections.Although influenza for most people is a mild ill-ness, which is resolved in 1–2 weeks, complicationsassociated with influenza infection can occur inboth the upper and lower respiratory tract, someof which may be fatal [1]. The potential for devel-oping complications is higher in certain riskgroups, such as the elderly and individuals withchronic medical conditions. An accurate diagnosisof influenza by a physician is difficult since severaldifferent pathogens can produce respiratory ill-

nesses with similar clinical symptoms. Conse-quently, there is a requirement for sensitive andrapid diagnostic techniques to verify the clinicaldiagnosis of influenza and improve the quality ofsurveillance systems. Moreover, the developmentof specific anti-influenza neuraminidase inhibitorshas increased the potential for rapid and accuratediagnostic tests for influenza viruses to contributeto the management of patients. A number oflaboratory methods for the diagnosis of influenzaare currently available (Table 1). Each of thesemethods has advantages and disadvantages, andsome, or all, of these factors may influence themethod of choice. Although molecular technologyhas transformed the diagnosis of a number of dis-eases caused by RNA viruses, for example HIV [2],the application of molecular methods to the detec-tion of respiratory pathogens is still comparativelynew and expanding rapidly.

MOLECULAR METHODS FOR THEDETECTION OF INFLUENZA VIRUSES

Choice of molecular assayA number of molecular methods can be employedfor the detection of influenza viruses, the majorityof which are based on PCR methodology (Table 2).When selecting which assay to use, there are a

RR EE V II E W

Copyright # 2002 John Wiley & Sons, Ltd.

*Corresponding author: Dr J. S. Ellis, Respiratory Virus Unit,Enteric, Respiratory and Neurological Virus Laboratory, PublicHealth Laboratory Service, Central Public Health Laboratory, 61Colindale Avenue, Colindale, London NW9 5HT, UK.E-mail: [email protected]

Abbreviations usedBDNA, branched chain DNA; EIA, enzyme immunoassay; HI, hae-magglutination inhibition; HMA, heteroduplex mobility assay; IF,immunofluorescence; ILI, influenza-like illness; LCR, ligase chainreaction; M, matrix; NASBA, nucleic acid sequence-based amplifica-tion; NP, nucleoprotein; NS, non-structural; PIV, parainfluenza;RFLP, restriction fragment length polymorphism; WHO, WorldHealth Organisation.

number of factors that should be taken into consid-eration. These include the requirement for qualita-tive, semi-quantitative or quantitative data, andthe nature and number of samples to be analysed.In addition, the available time and resources ofthe laboratory where the work is to be performedand the skill of the staff involved must also beconsidered.

HybridisationMolecular hybridisation methods have been usedextensively in basic and applied virology becauseof their technical flexibility and high specificity.

Using these techniques, DNA and RNA viruseshave been detected directly in clinical specimens.Influenza A virus RNA has been detected in naso-pharyngeal swabs by molecular hybridisation,with a sensitivity of 72% [3]. Hybridisation wasmore sensitive than both immunofluorescence(IF) and culture, but less sensitive than EIA, whichdetected influenza virus in 86% of swabs. Molecu-lar hybridisation, however, has not been widelyused for the detection of influenza viral RNA inclinical material, largely due to the developmentand optimisation of more sensitive and less time-consuming techniques.

Table 1. Comparison of the properties of diagnostic methods for influenza virus detection

Method Sample required Cost per test Speeda Through Advantages Disadvantages(£) putb

Culture Nasopharyngeal £10–£20 3–7 days Low Whole virus Requiresaspirate measured infectious virus

Nose and throat Virus Highly skilledswab recoverable Time required

Bronchoalveolarlavage

IF Nasopharyngeal £10 2 h–1 day Medium Rapid Requires intactaspirate cells

Nose and throat Highly skilledswab No virus

Bronchoalveolar recoverablelavage Specialised

equipmentEIA Nasopharyngeal £5–£30 15 min–1 day High Rapid Cost

aspirate Low skillNose and throat Can be ‘near No virus

swab patient’ recoverableBronchoalveolar testing

lavageRT-PCR Nasopharyngeal £30–£50 1–2 days High Sensitive Cost

aspirate Allows Highly skilledNose and throat further No virus

swab molecular recoverableBronchoalveolar analysis Specialised

lavage equipment/Post-mortem tissue laboratory

Serology Serum £5 2 days High Sensitive and RetrospectiveHI/CFT specific Paired samples

needed

aSpeed with which results are available from the time the specimen is taken.bNumber of specimens handled in unit time.

376376 J. S. Ellis and M. C. ZambonJ. S. Ellis and M. C. Zambon

Copyright # 2002 John Wiley & Sons, Ltd. Rev. Med. Virol. 2002; 12: 375–389.

PCRThe development of PCR analysis in 1985 madepossible the diagnosis of virus infection throughsensitive detection of specific viral nucleic acids[4]. PCR techniques have been developed for thespecific detection and subtyping of influenzaviruses to obtain rapid diagnostic results. In theseassays, purified influenza viral RNA from culturefluids or clinical specimens is first reverse tran-scribed to cDNA, by either avian myeloblastosisvirus or moloney murine leukaemia virus reversetranscriptase, using random hexanucleotides, uni-versal primer complementary to the 30 end of allinfluenza vRNAs, or a sequence specific primer.The use of random hexanucleotides or a universalprimer, instead of target sequence specific primershas the advantage that the cDNA from both viralgenomic RNA and mRNA transcripts is synthe-sized, thereby increasing the number of targetregions that could be amplified by PCR. A singleround of amplification may be used and thespecificity of the reaction confirmed by hybridisa-tion with product specific probes. Alternatively,

nested-primer sets may be utilised to amplify thetarget region.

The choice of gene target is influenced by theprospective application of the assay. For type-specific diagnosis of influenza A, B or C infection,internal genes such as nucleoprotein (NP) andmatrix (M) genes are usually chosen, since theseare highly conserved within influenza types.Where information on the subtype of influenza Ais required, then the genes encoding the surfaceantigens are targeted. The use of multisegmentPCR using primers complementary to the con-served 13 nucleotides at the 50 terminus and 12nucleotides at the 30 terminus allows the detectionof all segments in a single reaction [5]. Moreover,primer sets based on the 15 and 21 nucleotideterminal sequences enable the specific amplifica-tion of each of the eight RNA segments and subse-quent analysis of all 15 HA and 9 NA subtypes ofinfluenza A virus [6].

Reports on the applications of PCR techniques tothe detection of influenza A and B viruses first app-eared in the early 1990s. Initial reports described

Table 2. Properties of molecular methods suitable for detection of influenza viruses

Method Advantages Disadvantages

Hybridisation High specificity SensitivityFlexible formats Time-consumingInexpensive

PCR Sensitive and specific Qualitative, not usually quantitativeAllows further product analysis

Multiplex PCR Sensitive and specific Requires extensive optimisation to ensureTests >1 target per assay no false negatives or primer competitionAllows further product analysis May be limited by product size

PCR-EIA Sensitive and specific Does not allow further product analysisHigh throughput Requires additional evaluation/validationNot limited by product size before useCan be multiplexed

Real-time PCR Sensitive and specific Requires specialised equipmentRapid Product analysis not always feasibleCan be quantitativeCan be multiplexed

NASBA Sensitive and specific Three enzymes usedCan be quantitative Time-consuming

RNA product analysis less feasibleMicroarrays Sensitive and specific Requires specialised equipment and extensive

Tests many targets per assay developmentDoes not allow further product analysis

Molecular diagnosis ofMolecular diagnosis of influenza 377377


the amplification of influenza A H1 nucleic acidfrom both nasopharyngeal lavages [7] and naso-pharyngeal aspirates [8], and influenza A H1, H3and influenza B HA specific sequences from throatswabs [9,10]. In a study by Zhang and Evans [11]the M genes of influenza A and B viruses and theinfluenza C HA gene were targets for type-specificnested-primer sets. Subtype-specific primers tar-geted conserved sequences within the three HAor two NA subtypes of different human influenzaisolates. Although the assays were shown to beboth sensitive and specific for the typing and sub-typing of cultured influenza viruses, their applica-tion to the analysis of viruses directly from clinicalmaterial was not assessed. Further studies reportedthe detection of influenza A, B and C viruses inrespiratory secretions, using primers targeting spe-cific sequences in the non-structural (NS), or Mgenes [12–14]. In these initial studies, a small num-ber of respiratory samples were used to evaluatethe applicability of the RT-PCR assays for the diag-nosis of influenza in clinical specimens. Althoughearly assays used different extraction procedures,gene targets and methods for discrimination of pro-ducts, the sensitivity of PCR for the detection ofinfluenza viruses was shown to be at least equiva-lent to that of culture.

The majority of RT-PCR assays described to datehave utilised primers deduced from the nucleotidesequences of human influenza viral genes. Theoccurrence of zoonotic events in Hong Kong andMainland China caused by influenza A H5N1and H9N2 avian viruses have highlighted theadditional requirement for rapid tests to detectviruses originating from non-human hosts. Recen-tly, a single-tube RT-PCR assay for the specificdetection of influenza A viruses from multiple spe-cies has been reported [15]. A primer set based onconserved regions of the M gene was shown todetect a panel of 25 genetically diverse virus iso-lates obtained from birds, humans, pigs, horsesand seals which included all known subtypes ofinfluenza A virus. To detect and partially charac-terise influenza A viruses from different animalspecies, a combined RT-PCR heteroduplex mobi-lity assay (HMA) approach has been described[16]. An M gene-based RT-PCR using nested pri-mer pairs was shown to be specific for the detec-tion of human, avian and swine influenza Aviruses, of 15 different subtypes. PCR ampliconswere then analysed by HMA to determine the

host of origin of the M gene segment. This metho-dology offers a rapid and sensitive means ofscreening for novel or unusual influenza viruses.

The segmented nature of the genome of influen-za A, B and C viruses makes reassortment amongviruses an important mechanism for generatinggenetic diversity. Reassortment between influenzaA viruses of different subtypes is of particularimportance because of its role in the generationof new pandemic strains in humans. Hence, thereis a need for the development of rapid molecularassays that provide subtype information on boththe HA and NA of influenza A viruses. This ishighlighted by the recent emergence of influenzaA H1N2 reassortant viruses in several countriesduring 2001–2002, following reassortment bet-ween circulating influenza A H1N1 and H3N2viruses [17]. The majority of existing moleculardiagnostic tests for influenza A viruses rely onthe detection and typing of the HA gene alone,with only a few laboratories undertaking antigenicor genetic analysis of the NA. As yet, PCR assaystargeting specific NA subtype sequences have onlybeen described for the N1 and N2 genes of influen-za A viruses [11,18].

PCR-EIAA modification of the PCR technique, PCR-enzymeimmunoassay (PCR-EIA) has been described[19–22]. A single round RT-PCR is performedand the resulting amplicons identified by hybridi-sation in solution to a biotinylated RNA probe anddetected in an EIA. The potential efficiency ofinfluenza A, M-gene specific PCR-EIA for use inclinical diagnosis was evaluated in a study byCherian [19]. A total of 90 nasal wash specimenswere obtained daily over a 10-day period fromnine human volunteers infected with influenza Avirus. The PCR-EIA and cell culture had equiva-lent sensitivities during the first 4 days followinginfection. However, PCR-EIA was substantiallymore sensitive than culture later in the course ofinfection. From days 5 to 9, PCR-EIA could detectinfluenza A virus RNA in 38% of samples,whereas only 9% were culture positive duringthat period of time. The results demonstratedthat clinical symptoms and shedding of viralRNA might persist after virus becomes undetect-able by traditional cultivation methods. The clini-cal importance of the detection of viral RNAsequences in culture negative samples remains to



be established, although increased sensitivity ofthis analytical method has several advantages,including extending the window for detection ofvirus and diagnosis of infection.

Multiplex PCRThe term multiplex PCR refers to the inclusion ofmore than one primer set in an amplification reac-tion, to detect the presence of more than one geneor genome segment in a single pathogen, or inmore than one pathogen. Several multiplex PCRassays have been described for the detection ofrespiratory pathogens in clinical material (Table 3).Multiplex PCR assays have been designed to typeand subtype influenza A H1, H3 and influenza Bvirus in clinical specimens [18,23]. A multiplexPCR assay using nested primer sets targeting con-served regions of the HA genes of influenza A H1,H3 and influenza B viruses has been used in a pro-spective surveillance of influenza in England dur-ing the 1995–1996 winter season [23]. A total of 619combined nose and throat swabs from patientswith an influenza-like illness (ILI) were analysedby cell culture and multiplex PCR. Of these,

39.7% were positive by multiplex PCR, comparedto 32.3% that yielded influenza viruses by cell cul-ture. The benefits of multiplex RT-PCR comparedwith cell culture for detection were most evidentduring the peak period when influenza viruswas circulating in the community. During this per-iod, 38.9% of the samples received were positiveby cell culture, whereas 57.5% were positive forinfluenza virus by multiplex PCR, indicating theimproved positive predictive value of multiplexPCR compared with cell culture.

Since infection with respiratory pathogen(s)other than influenza viruses can result in clinicalsymptoms clinically indistinguishable from trueinfluenza, samples negative for influenza virusfrom patients with ILI may contain these patho-gens. The multiplex PCR assay described abovehas been modified to detect and subtype RSV Aand B, in addition to influenza A and B viruses,in a single reaction [24]. In this study, an excellent(100%) correlation between multiplex PCR and theresults of influenza culture was demonstrated. Thetest was also able to detect mixed viral infectionsin both simulated specimens and clinical samples.

Table 3. Multiplex RT-PCR assays described for detecting human respiratory pathogens

Target Number of Type Method of Sensitivity Referencepathogens of productdetected PCRa identification

Influenza A (H1, H3) 3 S Size/Gel 25–100 pfu [18]and B electrophoresis

Influenza A and RSV 2 S Size/Gel 2 pfu [20]electrophoresis

Influenza A (H1, H3) 3 N Size/Gel 1–5 pfu [23]and B electrophoresis

Influenza A (H1, H3), 4 S Size/Gel Not reported [49]B and C electrophoresis

Influenza A (H1, H3), 5 N Size/Gel 1 pfu [24]B and RSV A and B electrophoresis

Influenza A and B, RSV 7 S Probe/Enzyme 5–10 genome [22]A and B, HPIV-1, -2 and hybridisation copies-3 (Hexaplex assay) assay

Influenza A and B, RSV, 9 S Probe/PCR- 100 genome [21]HPIV-1, -3, enterovirus, ELISA copiesadenovirus, Mycoplasmapneumoniae andChlamydia pneumoniae

aNumber of rounds of amplification; S, single, one round of amplification, N, nested, two rounds of amplification.



The use of multiplex RT-PCR enzyme hybridisa-tion assays has allowed the detection of a widerange of respiratory pathogens in a singletest, since discrimination of the products is notlimited by size differences detectable followinggel electrophoresis. The Hexaplex assay (Prodesse,Milwaukee) includes primers specific for the Mand NS genes of influenza A and B viruses respec-tively, as well as other primer pairs for the detec-tion and quantification of RSV A and B, humanparainfluenza (PIV) type-1, 2 and 3 RNA in clinicalspecimens [22]. A hot start RT-PCR EIA protocolfor the simultaneous detection of nine respiratorypathogens including influenza A and B viruses,PIV-1, PIV-3, RSV, enteroviruses, adenovirus,Chlamydia pneumoniae and Mycoplasma pneumoniaehas been developed. This assay has been evaluatedusing 1118 nasopharyngeal aspirates collectedfrom hospitalised children over a number of influ-enza seasons [21,25]. Of these, 395 (35%) werepositive for one of the nine pathogens. Over theperiod of the study, 20% of these were positivefor influenza A virus and 2.8% were positive forinfluenza B virus. Seasonal variations in the ratesof detection of the different infectious agentswere seen. The multiplex RT-PCR showed excel-lent levels of concordance (98%) for influenza Avirus detection with data obtained by commer-cially available enzyme immunoassay.

NASBAIn the past few years, amplification proceduresother than PCR have been developed. Theseinclude branched chain DNA (bDNA) assay [26],ligase chain reaction (LCR) assays [27], and nucleicacid sequence-based amplification (NASBA). Ofthese assays, NASBA has been most widelyapplied for the detection of viral nucleic acid. Inthe NASBA system amplification is carried out iso-thermally using three enzymes simultaneously(avian myeloblastosis virus reverse transcriptase,T7 RNA polymerase and RNase H) with two spe-cially designed DNA oligonucleotide primers. Thetechnique is particularly suited to the detection ofRNA viruses, since RNA polymerase amplifiesRNA without conversion to complementaryDNA [28]. NASBA has been applied successfullyto a number of RNA templates including HIV-1[29], HCV [30], rhinoviruses [31] and enteroviruses[32]. Interestingly, a NASBA technique has veryrecently been developed that allows the detection

of avian influenza A subtype H5 isolates of the Eur-asian lineage from allantoic fluid harvested frominoculated chick embryos [33]. Both generic pri-mers (for the amplification of both highly patho-genic and low pathogenic H5 HA sequences), andprimers specific for pathogenic H5 HA sequenceswere designed and evaluated for the detection ofH5 HA sequences and the differentiation of highlypathogenic H5 sequences from low pathogenic H5sequences. However, NASBA technology has yet tobe applied to the detection of human influenzaviruses in respiratory specimens.

QUALITATIVE VERSUS QUANTITATIVEASSAYSThe most sensitive molecular techniques use anenzymatic step to amplify target nucleic acidbefore detection of the specific sequence. Typi-cally, the logarithmic amplification of the targetsequence results in the amount of DNA productat the end of the PCR having no correlation tothe number of target copies present in the originalspecimen. The qualitative nature of PCR-basedapplications limited, in the early stages, the useof this methodology to those conditions whereonly the presence or absence of a specific targetsequence was to be assayed. However, manyapplications require quantification of the numberof target molecules in the original specimen. BothPCR-based and non-PCR strategies, such as real-time PCR and NASBA, respectively, have beenadapted for quantitation of nucleic acid products,using either internal or external standards. Quan-titative molecular assays have been used primarilyto determine the viral load in blood specimens forthe diagnosis of HIV, HCV and hepatitis B virusinfections [34–36]. Quantitation of CMV DNAmolecules is required for successful monitoringof this infection in immunocompromised indivi-duals and during therapy.

Where traditional diagnostic methods are insen-sitive, our understanding of the natural history ofviral disease has continued to be improved thro-ugh the use of molecular techniques. In particular,quantitative PCR-based studies have had animportant role in the analysis of the natural historyof HIV and HCV infections, and subsequentlyin the development of effective treatment strate-gies. During antiviral treatments, quantitativetechniques can supply information on both effi-cacy of therapies and selection of drug-resistant



viral variants in real-time, as demonstrated inHIV-1 infection [37,38].

Although quantitative PCR assays have yet to bewidely employed in the diagnosis of influenzainfections, and the development of such assays istechnically demanding, the potential value of suchassays in monitoring efficacy of anti-influenza drugtreatments and in the monitoring of immunosup-pressed patients is apparent.

ANALYSIS OF MOLECULAR PRODUCTSOne advantage of PCR-based diagnostic assays isthat subsequent sequence information can be

determined from the assay products, in a varietyof ways (Figure 1). Post-amplification methodsfor analysing sequence variation in PCR ampliconsinclude restriction fragment length polymor-phism analysis (RFLP), HMA and nucleotidesequencing.

PCR-RFLPA combination of RT-PCR and enzyme digestionof HA gene amplicons (PCR-restriction assay)has been used to rapidly differentiate influenzaA H3 genetic variants that are antigenically similar[39,40]. This technique may also be utilised to dif-

Figure 1. PCR-based amplification methods and product analysis



ferentiate between vaccine strains and currentlycirculating strains [38] and to assess the gene com-position of candidate vaccine strains [41,42].Amantadine-resistant influenza A viruses havebeen detected in nasopharyngeal swabs by PCR-RFLP [43]. PCR-restriction assays have mostrecently been used to rapidly genotype and moni-tor the internal genes of human influenza H1N1,H3N2 and H5N1 influenza A viruses [44]. A pos-sible drawback of the RFLP technique is that muta-tions in the nucleotide sequence of the gene ofinterest may lead to the loss or generation of arestriction site. The high mutability of RNA gen-omes such as influenza viral genes increases thepossibility of this occurring.

Heteroduplex mobility assaysHeteroduplex formation between amplicons hasbeen used to subtype viral genomes, to screen forgenome variants, to detect mutations and to mea-sure diversity within and between viral genomepopulations [45]. The initial application of HMAin virology was to subtype HIV-1 [46,47]. Thismethodology has since been applied to the analy-sis of many RNA viral genomes, including those ofinfluenza A and B viruses. By combining a HA-specific multiplex RT-PCR with HMA, variantstrains of influenza A and B viruses can be differ-entiated [48,49]. Direct amplification of internalgenes from clinical material can also be coupledwith HMA to allow rapid species identificationof the origin of influenza viruses [16].

Nucleotide sequence analysisVariation among amplified fragments indicated byRFLP or HMA can be confirmed by nucleotidesequencing of derived clones or the PCR productitself. In many laboratories sequence analysis ofPCR amplicons is routinely performed, particu-larly on HA gene products where an associationbetween sequence changes with genetic drift isstudied [50,51]. Sequence analysis of influenzagene segments amplified by PCR may also beundertaken to provide information on geneticreassortment between influenza A viruses [52–55].

Antigenic drift in influenza viruses results fromthe progressive accumulation of mutations,including substitutions, deletions and insertionsin the influenza viral genes. These mutations arisedue to replication of the influenza virus genomeby a viral RNA polymerase that lacks proofread-

ing activity. Because of this genetic variation, thetarget regions in newly emerging strains comple-mentary to the primers or probes used in an assaymust be regularly analysed by nucleotide sequen-cing to check for sequence mismatches betweenthe primer or probe, and target sequences. In theevent of sequence mismatches, the primers orprobes should be updated in order to avoid falsenegative assay results.

APPLICATIONS OF MOLECULAR METHODS

DiagnosisA limited number of studies have comparedlaboratory diagnostic techniques for influenzaviruses, including RT-PCR, with clinical diagnosis.In a study that aimed to predict influenza infec-tions during epidemics by use of a clinical casedefinition, combined nasal and pharyngeal swabspecimens were collected from 100 patients pre-senting with ILI of <72 h duration [56]. Patientswere aged between 6 and 84 years (mean 39.3years) and had fever together with two clinicalsymptoms (headache, cough, sore throat andmyalgia). The rate of laboratory-confirmed influ-enza infection was 72% by cell culture and 79%according to the results of multiplex RT-PCR ana-lysis. Cough and fever were the only factorsshown to be significantly associated with a posi-tive PCR test for influenza. Early phase III trialsof the anti-neuraminidase drug, zanamivir, inEurope and N. America have also provided evi-dence of excellent concordance between PCR andclassical diagnostic methodology [57]. In thesetrials, samples were collected from communitycases of influenza during periods when influenzavirus was circulating from patients who presentedwithin the first or second calendar day of onset ofsymptoms. Patients were aged between 12 and 81years (mean 37 years) and had fever together withtwo symptoms (headache, myalgia, sore throat,cough). The percentage of samples positive forinfluenza A or B virus, by cell culture, serologyor PCR were compared. A total of 791 (77%) of1033 patients with laboratory results from all threemethods were confirmed positive for influenza byone or more test results. For 692 (67%) patients, theresults of all three tests agreed. Furthermore, therewas a significant correlation between the numberof tests positive and illness severity. Where allthree tests were positive, there was a significant



correlation between duration of illness, but notantibiotic use, or the risk of complications. Altho-ugh PCR was more sensitive than either culture orpaired haemagglutination inhibition (HI) serology,PCR positivity alone was not more likely to beassociated with severity of illness, developmentof complications or antibiotic usage. The resultsof these studies confirm that RT-PCR providesrapid and accurate diagnosis in an individualpatient and is more sensitive than cell culture orthe use of paired serology, for detection of casesof influenza in the community.

SurveillanceThe effectiveness of PCR for enhancing the surveil-lance and characterisation of circulating influenzastrains has been well documented (Table 4). RT-PCR has been compared with other methodolo-gies, such as isolation in culture, EIA and IF forthe detection of influenza viruses in samples col-lected during surveillance of influenza virus activ-ity. A number of RT-PCR assays employed in thesestudies have used single amplification with pri-mers specific for the M gene [13], or NS gene [12]of influenza A and B viruses, or primers defined inNP, NS and HA genes [18], to type and subtypeinfluenza viruses [58,59]. Others have used thenested primer sets described by Zhang and Evans[11] for typing and subtyping of influenza viruses,either in uniplex assays or in a multiplex system[23,60–63]. Although the RT-PCR assays used inthese studies differ in their format, the results ofthe analyses confirmed that RT-PCR was moresensitive than traditional techniques for influenzavirus detection in clinical material. The effective-ness of RT-PCR for the detection of influenzaviruses was particularly apparent in surveillancesystems where the culture of influenza viruses,for a number of reasons, was found to be difficultand suboptimal. In Portugal, a comparison of mul-tiplex RT-PCR for the detection of influenzaviruses with culture, EIA and serology was per-formed over a 7-year surveillance period, from1992 to 1999 [63]. There was good correlationbetween the increase of morbidity, total samplestaken and the detection of influenza virus by allthe methods, although this was less evident forvirus isolation and EIA than for RT-PCR or serol-ogy. From a total of 1685 throat swabs collectedfrom cases of ILI, more samples were found tobe positive by RT-PCR, than by any other method.

Moreover, the detection of influenza by RT-PCRoccurred earlier than by any other method in allof the years studied, and consistently showed thebest relation with epidemic patterns of morbidityregistration.

The correlation between detection of influenzaviruses by RT-PCR and serology was also obser-ved during surveillance of community influenzainfection in general practice in Scotland [64]. Influ-enza virus was detected in 57% of combined noseand throat swabs by RT-PCR, and 61% by serol-ogy. The RT-PCR could be performed in 36 h,whereas a serological result required paired serataken 2 weeks apart.

The exact impact of molecular methodology onsurveillance is likely to depend upon the sensitiv-ity of laboratory systems already in place. Thedecision as to whether to use molecular methods,either in addition to or in the place of traditionalassays has to be made by each laboratory, indivi-dual laboratories, in each country according tolocal circumstances [65,66].

OutbreaksInvestigations of outbreaks of respiratory illnessare often hindered by the inability to culture theinfectious agent responsible. Therefore, the use ofmolecular techniques directly on clinical respira-tory specimens is of particular value in the analy-sis of respiratory outbreaks. RT-PCR has beenused to provide rapid identification of influenzavirus outbreaks in schools [39] and on cruise ships[67]. In addition, a combination of PCR and post-amplification analysis such as RFLP can reveal therelationship of the causative strain to currently cir-culating viruses and vaccine strains [39].

Tissue diagnosisMolecular methodology is of particular use in thestudy of post-mortem specimens [68] and may beof great value in the diagnosis of influenza in thecentral nervous system [69–71]. Furthermore,analysis of other body tissues may help to clarifymechanisms of pathogenesis [72].

The most important application of RT-PCR andsequencing in tissue analysis to date has beenin the genetic analysis of the 1918 pandemic influ-enza A H1N1 strain in archival material. Amplifi-able influenza RNA has been recovered fromarchival paraffin blocks as old as 79 years [73].Influenza RNA has been detected by RT-PCR in



both formalin-fixed and frozen tissue samples fromvictims of the pandemic [74–76]. The full-lengthsequences of the two predominant surface proteinsof the virus, the HA and NA, and the non-structur-al gene have now been deduced [75–77]. Mostrecently, influenza RNA has been recovered frompreserved specimens from six wild waterfowlthat were captured between 1916 and 1919 [78].From one of these samples a portion of the HAgene (H1 subtype) has been sequenced. Analysisof this, together with sequence information fromthe human 1918 samples, is being used to deter-mine the origin of the 1918 pandemic virus.

RECENT ADVANCES IN MOLECULARDIAGNOSTIC METHODS

Real-time PCRRecent modifications of PCR analysis called ‘real-time’ PCR are a dramatic development in PCR-based technology. In these assays, a fluorescentsignal is generated as the PCR takes place. Thecombination of nucleic acid amplification and sig-nal detection reduces the time required for nucleicacid detection, since post-PCR processing is notrequired. Fluorogenic PCR-based methods, usingTaqMan PCR technology (Applied Biosystems,Foster City, CA), have just been described for thedetection and identification of influenza A and Bviruses [79,80]. In these assays, primer/probesets targeting the M gene of influenza A and Bviruses [80], or the M gene of influenza A andHA gene of influenza B [79], were designed to dif-ferentiate influenza A and B viruses. In the first ofthese studies, specific primer/probe sets were alsoselected to identify HA (H1 and H3) and NA (N1and N2) subtypes [80]. The fluorogenic probeswere labelled with both a fluorescent reporterand a quencher dye. After hybridisation of theprobe to the target sequence, the Taq DNA poly-merase enzyme cleaved the TaqMan probe bymeans of its 50–30 nuclease activity thereby separat-ing the reporter and quencher dyes resulting inincreased fluorescence. The specificity of the meth-od in both studies was first evaluated on previouslycharacterised reference strains and isolates. Anexcellent correlation (100%) was demonstratedbetween the results of typing and subtyping bythe TaqMan PCR and antigenic analysis. In addi-tion, both of the fluorogenic assays were shownto be extremely sensitive for the detection of influ-

enza A and B viruses. The application of real-timePCR methodology to the surveillance of influenzaviruses in samples obtained from the communitywas assessed during two influenza seasons inGermany [80]. The TaqMan PCR was shown tobe more sensitive than cell culture, revealing anoverall increase during the period studied ofapproximately 12%. Furthermore, during the peri-od of highest clinical activity, an increase of 26%influenza virus detection by TaqMan PCR wasobserved. This confirms the findings of earlier stu-dies demonstrating the particular benefits of PCR-based methodologies compared with cell culturefor detection, when influenza viruses are circulat-ing in the community [23,60].

Microarray technologyMicroarray DNA chips containing immobilisedoligonucleotide probes or robotically spottedDNAs can be used to rapidly screen for a numberof target molecules amplified from clinical material[81,82]. The ability to simultaneously screen formany nucleic acid sequences implies that microar-ray technology offers great potential as a diagnos-tic tool. Reports on the application of array-basedmethods for diagnosis have now started to appear,and a model DNA array has recently beendescribed for the typing and subtyping of humaninfluenza A and B viruses [83]. A series of 26 pri-mer pairs were designed to amplify multiple frag-ments from influenza A HA (H1, H2, H3), NA(N1, N2) and MP genes, as well as influenza Bvirus HA, NA and MP genes. Using these, 24cDNA products generated were cloned andsequenced, prior to reamplification and spottingonto a modified glass support. Cy3- or Cy5-labelled fluorescent probes were then hybridisedto these target DNAs. Multiplex PCR primerswere also designed for the generation of probesto type and subtype influenza A and B viruses.From the results it was concluded that DNAmicroarray technology could provide a useful sup-plement to PCR-based diagnostic methods. Theapplication of this methodology to the detectionof influenza viral nucleic acid sequences in clinicalmaterial will require further analysis.

THE FUTURE FOR MOLECULAR DIAGNOSISThe past decade has seen tremendous develop-ments in molecular diagnostic techniques. Futureapplications of molecular methods may involve



the use of multiple detection systems in which anumber of clinically and epidemiologically relatedpathogens may be detected and characterised in asingle test. It should be noted that although PCR isan extremely powerful and versatile assay meth-od, the process can be labour-intensive. Morewidespread implementation of molecular testingwill therefore depend upon automation, enablingmolecular assays to enter the routine clinicallaboratory. Robotic systems are currently availableto automate each stage of the PCR process, but todate only reports of automated systems for thedetection of herpes simplex virus DNA, CMVDNA, or HCV RNA have been published [84–85].Multiplex PCR methodology coupled with devel-opments in microelectronic detection devices offersthe exciting possibility of ‘near-patient’ molecularbased testing [87]. While RT-PCR for detection ofinfluenza genes is now a well-established method,most assays are qualitative and not quantitative.Quantification may be invaluable for clinical vali-dation of the diagnosis, and quantitative RT-PCRsystems are being designed to allow the determi-nation of viral load in infections.

Traditional clinical diagnostic methods requirestandardisation, quality assurance and qualitycontrol measures to be put into place, in order tovalidate assay results. Due to the extreme sensitiv-ity of PCR assays and similar tests, additionalquality control measures must be implemented.Furthermore, nucleic acid detection assays usuallyinvolve several processes, and at each stagepotential sources of error may occur (Table 5).Human error can occur at all stages, but can bereduced by automation of some, or all of the stepsinvolved. Validation of all reagents should be per-formed prior to their use in the assay, since errorsdue to reagent failures can occur throughout theprocedure. The integrity of both positive and nega-tive results can be established by the incorporation

of appropriate positive and negative controls,which are processed in an identical manner tothe test samples in the assay procedure.

The detection and characterisation of influenzaisolates and the identification of newly emergingvariants, is the foundation of the World HealthOrganization (WHO) influenza global surveillancenetwork. Information on the antigenic propertiesof circulating influenza viruses compared withreference and vaccine strains is important for theformulation of influenza virus vaccines. For thisreason, optimised influenza surveillance requiresboth sensitive detection of influenza strains andisolation, followed by full characterisation of theantigenic properties of virus isolates. Although

Table 4. Comparison of influenza diagnosis and surveillance by PCR and cell culture

Country Dates Number of PCR positive Culture positivesamples (%) (%)

England and Wales 1995–2000 3455 34.8 23.0Portugal 1992–1999 1685 43.5 5.0Scotland 1999–2000 168 58.0 11.0Germany 1997–1999 2545 25.0 16.0

Table 5. Sources of potential error in influ-enza viral nucleic acid detection methods

Procedure Potential outcome

Extraction of nucleic acidRNA/DNA degraded False negativeInhibitors present False negativeIncomplete extraction False negativeContamination introduced False positive

Reverse transcriptionTranscription errors False negativeEnzyme failure False negative

AmplificationContamination False positiveInhibition False negativeInsensitivity False negativeNon-specificity False negative/

positiveEnzyme failure False negative

DetectionNon-specificity False negative/

positiveEnzyme failure False negative



molecular techniques offer a rapid and sensitivealternative to traditional methods for the diagnosisof influenza virus infections, they will not replaceconventional culture for the provision of virus iso-lates for antigenic characterisation in referencelaboratories. However, molecular methods remaininvaluable in assisting our knowledge of the epi-demiology of influenza viruses.

ACKNOWLEDGEMENTSThe authors would like to thank Jennie Lane andJon White for the illustration.

REFERENCES1. Nicholson KG. Textbook of Influenza, Nicholson KG,

Webster RG, Hay AJ (eds). Blackwell Science:London, 1998.

2. Cassol S, Butcher A, Kinard S, et al. Rapid screeningfor the early detection of mother-to-child transmis-sion of HIV-1. J Clin Microbiol 1994; 32: 2641–2645.

3. Havlickova M, Pljusnin A, Tumova B. Influenzavirus detection in clinical specimens. Acta Virol1990; 34: 446–456.

4. Saiki R, Scharf S, Faloona F, et al. Enzymatic amplifi-cation of beta-globin genomic sequences and restric-tion site analysis for diagnosis of sickle cell anemia.Science 1985; 230: 1350–1354.

5. Adeyefa C, Quayle K, McCauley J. A rapid methodfor the analysis of influenza virus genes: applicationto the reassortment of equine influenza virus genes.Virus Res 1994; 32: 391–399.

6. Hoffmann E, Stech J, Guan Y, Webster R, Perez D.Universal primer set for the full-length amplificationof all influenza A viruses. Arch Virol 2001; 146:2275–2289.

7. Bressoud A, Whitcomb J, Pourzand C, Haller O,Cerruti P. Rapid detection of influenza virus H1 bythe polymerase chain reaction. Biochem Biophys ResCommun 1990; 167: 425–430.

8. Rajakumar A, Inkster M, Manger I, Swierkosz E,Schulze I. The use of the polymerase chain reactionin influenza virus detection and characterization.Clin Microbiol Rev 1991; 4: 260–271.

9. Yamada A, Imanishi J, Nakajima E, Nakajima K,Nakajima S. Detection of influenza viruses in throatswab by using polymerase chain reaction. MicrobiolImmunol 1991; 35: 259–265.

10. Yamada A, Imanishi J. Detection of influenza B virusin throat swabs using the polymerase chain reaction.Acta Virol 1992; 36: 320–325.

11. Zhang W, Evans DH. Detection and identification ofhuman influenza viruses by the polymerase chainreaction. J Virol Methods 1991; 33: 165–189.

12. Claas E, Sprenger M, Kleter G, van Beek R, Quint W,Masurel N. Type-specific identification of influenzaviruses A, B and C by the polymerase chain reaction.J Virol Methods 1992; 39: 1–13.

13. Donofrio J, Coonrod J, Davidson J, Betts R. Detectionof influenza A and B in respiratory secretions withthe polymerase chain reaction. PCR Methods Applic1992; 1: 263–268.

14. Pisareva M, Bechtereva T, Plyusnin A, DobretsovaA, Kisselev O. PCR-amplification of influenza Avirus specific sequences. Arch Virol 1992; 125:313–318.

15. Fouchier R, Bestebroer T, Herfs S, Van Der Kemp L,Rimmelzwaan G, Osterhaus A. Detection of influ-enza A viruses from different species by PCR ampli-fication of conserved sequences in the matrix gene.J Clin Microbiol 2000; 38: 4096–4101.

16. Ellis J, Zambon M. Combined PCR-heteroduplexmobility assay for detection and differentiation ofinfluenza A viruses from different animal species.J Clin Microbiol 2001; 39: 4097–4102.

17. Anonymous. Influenza A(H1N2) viruses (update).WHO Weekly Epidemiol Rec 2002; 77: 77–80.

18. Wright K, Wilson G, Novosad D, Dimock C, Tan D,Weber J. Typing and subtyping of influenza virusesin clinical samples by PCR. J Clin Microbiol 1995; 33:1180–1184.

19. Cherian T, Bobo L, Steinhoff MC, Karron RA, YolkenRH. Use of PCR-enzyme immunoassay for identifi-cation of influenza A virus matrix RNA in clinicalsamples negative for cultiviable virus. J Clin Micro-biol 1994; 32: 623–628.

20. Valassina M, Cuppone AM, Cusi MG, Valensin PE.Rapid detection of different RVA respiratory virusspecies by multiplex RT-PCR: application to clinicalspecimens. Clinical and Diagnostic Virology 1996; 227–232.

21. Grondahl B, Puppe W, Hoppe A, Kuhne I, Weigl J.Rapid identification of nine microorganisms causingrespiratory tract infections by single-tube multiplexreverse transcription-PCR: feasibility study. J ClinMicrobiol 1999; 37: 1–7.

22. Fan J, Kelly J, Savatski L. Rapid simultaneous diag-nosis of infections with respiratory syncytial virusesA and B, influenza viruses A and B, and humanparainfluenza virus types 1, 2, and 3 by multiplexquantitative reverse transcription-polymerase chainreaction-enzyme hybridization assay (Hexaplex).Clin Infect Dis 1998; 26: 1397–1402.

23. Ellis J, Fleming D, Zambon M. Multiplex reversetranscription-PCR for surveillance of influenza Aand B viruses in England and Wales in 1995 and1996. J Clin Microbiol 1997; 35: 2076–2082.

24. Stockton J, Ellis J, Saville M, Clewley J, Zambon M.Multiplex PCR for typing and subtyping influenza



and respiratory syncytial viruses. J Clin Microbiol1998; 36: 2990–2995.

25. Weigl J, Puppe B, Grondahl B, Schmitt H-J. Epide-miological investigation of nine respiratorypathogens in hospitalized children in Germanyusing multiplex reverse-transcriptase polymerasechain reaction. Eur J Microbiol Infect Dis 2000; 19:336–343.

26. Dewar R, Highbarger H, Sarmiento M, et al. Applica-tion of branched DNA signal amplification to moni-tor human immunodeficiency virus type I burden inhuman plasma. J Infect Dis 1994; 170: 1172–1179.

27. Barany F. Genetic disease detection and DNA ampli-fication using cloned thermostabile ligase. Proc NatlAcad Sci USA 1991; 88: 189–193.

28. Guatelli J, Whitfield K, Kwoh D, Barringer K, Rich-man D, Gingeras T. Isothermal, in vitro amplificationof nucleic acids by a multienzyme reaction modeledafter retroviral replication. Proc Natl Acad Sci USA1990; 87: 7797.

29. Bruisten S, Van Gemen B, Koppelman M, et al. Detec-tion of HIV-1 distribution in different blood fractionsby two nucleic acid amplification assays. AIDS ResHum Retroviruses 1993; 9: 259–265.

30. Lunel F, Cresta P, Vitour D, et al. Comparative eva-luation of hepatitis C virus RNA quantitation bybranched DNA, NASBA, and monitor assays.Hepatology 1999; 29: 528–535.

31. Samuelson A, Westmoreland D, Eccles R, Fox J.Development and application of a new method foramplification and detection of human rhinovirusRNA. J Virol Methods 1998; 71: 197–209.

32. Fox J, Han S, Samuelson A, Zhang Y, Neale M, West-moreland D. Development and evaluation of nucleicacid sequence based amplification (NASBA) fordiagnosis of enterovirus infections using the Nucli-Sens Basic Kit. J Clin Virol 2002; 24: 117–130.

33. Collins R, Ko L, So K, Ellis T, Lau L, Yu A. Detectionof highly pathogenic and low pathogenic avian influ-enza subtype H5 (Eurasian lineage) using NASBA.J Virol Methods 2002; 103: 213–225.

34. Bell H, Hellum K, Harthug S, et al. Genotype, viralload and age as independent predictors of treatmentoutcome of interferon-alpha 2a treatment in patientswith chronic hepatitis C. Scand J Infect Dis 1997; 29:17–22.

35. Mellors J, Rinaldo CJ, Gupta P, White RM, Todd TA,Kingsley LA. Prognosis in HIV-1 infection predictedby the quantity of virus in plasma. Science 1996; 272:1167–1170.

36. Mutimer D, Pillay D, Dragon E, et al. High pre-treatment serum hepatitis B virus titre predicts failureof lamivudine prophylaxis and grafts re-infectionafter liver transplantation. J Hepatol 1999; 30:715–721.

37. Bagnarelli P, Menzo S, Valenza A, et al. Quantitativemolecular monitoring of human immunodeficiencyvirus type 1 activity during therapy with specificanti-retroviral compounds. J Clin Microbiol 1995; 33:16–23.

38. Mathez D, Schinazi R, Liotta D, Leibowitch J. Infectiousamplification of wild type human immunodeficiencyvirus infection levels from patients’ lymphocytes andtheir modulation by reverse transcriptase inhibitorsin vitro. Antimicrob Agents Chemother 1993; 37: 2206–2211.

39. Ellis J, Zambon M. Molecular analysis of an outbreakof influenza in the United Kingdom. Eur J Epidemiol1997; 13: 369–372.

40. Coiras M, Aguilar J, Galiano M, et al. Rapid molecu-lar analysis of the haemagglutinin gene of humaninfluenza A H3N2 viruses isolated in Spain from1996 to 2000. Arch Virol 2001; 146: 2133–2147.

41. Sakamoto S, Kino Y, Oka T, Herlocher M, Maassab F.Gene analysis of reassortant influenza virus by RT-PCR followed by restriction enzyme digestion. J VirolMethods 1996; 56: 161–171.

42. Offringa D, Tyson-Medlock V, Zhiping Y,Levandowski R. A comprehensive systematicapproach to identification of influenza A virus geno-type using RT-PCR and RFLP. J Virol Methods 2000;88: 15–24.

43. Saito R, Oshitani H, Masuda H, Suzuki H. Detectionof amantadine-resistant influenza A virus strains innursing homes by PCR-restriction fragment lengthpolymorphism analysis with nasopharyngeal swabs.J Clin Microbiol 2002; 40: 84–88.

44. Cooper L, Subbarao K. A simple restriction fragmentlength polymorphism-based strategy that can distin-guish the internal genes of human H1N1, H3N2, andH5N1 influenza A viruses. J Clin Microbiol 2000; 38:2579–2583.

45. Barlow K, Green J, Clewley J. Viral genome charac-terisation by the heteroduplex mobility and hetero-duplex tracking assays. Rev Med Virol 2000; 10:321–335.

46. Bachmann M, Delwart E, Shpaer E, Lingenfelter P,Singal R, Mullins J. Rapid genetic characterizationof HIV type 1 strains from four World HealthOrganization-sponsored vaccine evaluation sitesusing a heteroduplex mobility assay. AIDS ResHum Retroviruses 1994; 10: 1345–1353.

47. Delwart E, Herring B, Rodrigo A, Mullins J. Geneticsubtyping of human immunodeficiency virus usinga heteroduplex mobility assay. PCR Methods Applic1995; 4: S202–S216.

48. Zou S, Stansfield C, Bridge J. Identification of newinfluenza B virus variants by multiplex reverse tran-scription-PCR and the heteroduplex mobility assay.J Clin Microbiol 1998; 26: 1544–1548.



49. Zou S. A practical approach to genetic screening forinfluenza virus variants. J Clin Microbiol 1997; 35:2623–2627.

50. Ellis JS, Chakraverty P, Clewley JP. Genetic and anti-genic variation in the haemagglutinin of recently cir-culating influenza A (H3N2) viruses in the UnitedKingdom. Arch Virol 1995; 140: 1889–1904.

51. Cox N, Bender C. Molecular epidemiology of influ-enza. Semin Virol 1995; 6: 359–370.

52. Wentworth DE, McGregor M, Macklin MD, Neu-mann V, Hinshaw VS. Transmission of swine influe-nza virus to humans after exposure to experimentallyinfected pigs. J Infect Dis 1997; 175: 7–15.

53. Lin Y, Shaw M, Gregory V, et al. Avian-to-humantransmission of H9N2 subtype influenza A viruses:relationship between H9N2 and H5N1 human iso-lates. Proc Natl Acad Sci 2000; 97: 9654–9658.

54. Claas E, Kawaoka Y, de Jong J, Masurel N, WebsterRG. Infection of children with avian-human reassor-tant influenza virus from pigs in Europe. Virology1994; 204: 453–457.

55. Banks J, Speidel E, Alexander D. Characterisation ofan avian influenza A virus isolated from a human—is an intermediate host necessary for the emergenceof pandemic influenza viruses? Arch Virol 1998; 143:781–787.

56. Boivin G, Hardy I, Tellier G, Maziade J. Predictinginfluenza infections during epidemics with use of aclinical case definition. Clin Infect Dis 2000; 31:1166–1169.

57. Zambon M, Hays J, Webster A, Newman R, Keene O.Diagnosis of influenza in the community. Arch InternMed 2001; 161: 2116–2122.

58. Vabret A, Sapin G, Lezin B, et al. Comparison of threenon-nested RT-PCR for the detection of influenza Aviruses. J Clin Virol 2000; 17: 167–175.

59. Atmar R, Baxter B, Dominguez E, Taber LH. Compar-ison of reverse transription-PCR with tissue cultureand other rapid diagnostic assays for detection of typeA influenza virus. J Clin Microbiol 1996; 34: 2604–2606.

60. Pregliasco F, Mensi C, Camorali L, Anselmi G. Com-parison of RT-PCR with other diagnostic assays forrapid detection of influenza viruses. J Med Virol1998; 56: 168–173.

61. Magnard C, Valette M, Aymard M, Lina B. Compar-ison of two nested PCR, cell culture and antigendetection for the diagnosis of upper respiratory tractinfections due to influenza viruses. J Med Virol 1999;59: 215–220.

62. Plakokefalos E, Markoulatos P, Ktenas E, Spyrou N,Vamvakopoulos C. A comparative study of immu-nocapture ELISA and RT-PCR for screening clinicalsamples from Southern Greece for human influenzavirus types A and B. J Med Microbiol 2000; 49: 1037–1041.

63. Rebelo-de-Andrade H, Zambon MC. Different diag-nostic methods for detection of influenza epidemics.Epidemiol Infect 2000; 124: 515–522.

64. Carman W, Wallace L, Walker J, et al. Rapid virologi-cal surveillance of community influenza infection ingeneral practice. Br Med J 2000; 321: 736–737.

65. Carman B. Molecular techniques should now replacecell culture in diagnostic virology laboratories. RevMed Virol 2001; 11: 347–349.

66. Ogilvie M. Molecular techniques should not nowreplace cell culture in diagnostic virology labora-tories. Rev Med Virol 2001; 11: 351–354.

67. Anonymous. Influenza on a criuse ship in the Medi-terranean. Commun Dis Rep 1999; 9: 209–212.

68. Guarner J, Shieh W-J, Dawson K, et al. Immunohisto-chemical and in situ hybirdization studies of influ-enza A virus infection in human lungs. Am J ClinPathol 2000; 114: 227–233.

69. Ito Y, Ichiyama T, Kimura H, et al. Detection of influ-enza virus RNA by reverse transcription-PCR andproinflammatory cytokines in influenza-virus-asso-ciated encephalopathy. J Med Virol 1999; 58: 420–425.

70. Fujimoto S, Kobayashi M, Uemura O, et al. PCR oncerebrospinal fluid to show influenza-associatedacute encephalopathy or encephalitis. Lancet 1998;352: 873–875.

71. McCullers JA, Facchini S, Chesney PJ, Webster RG.Influenza B virus encephalitis. Clin Infect Dis 1999;28: 898–900.

72. Xu H, Yasui O, Tsuruoka H, et al. Isolation of type Binfluenza virus from the blood of children. Clin InfectDis 1998; 27: 654–655.

73. Krafft A, Duncan B, Bijwaard K, Taubenberger J,Lichy J. Optimization of the isolation and amplifica-tion of RNA from formalin-fixed, paraffin-embedded tissue: the Armed Forces Institute ofPathology experience and literature review. MolDiagn 1997; 2: 217–230.

74. Taubenberger JK, Reid AH, Krafft AE, Bijwaard KE,Fanning TG. Initial genetic characterisation of the1918 ‘Spanish’ influenza virus. Science 1997; 275:1793–1796.

75. Reid A, Fanning T, Hultin J, Taubenberger J. Originand evolution of the 1918 Spanish influenza virushaemagglutinin gene. Proc Natl Acad Sci 1999; 96:1651–1656.

76. Reid A, Fanning T, Janczewski T, Taubenberger J.Characterization of the 1918 ‘Spanish’ influenzavirus neuraminidase gene. Proc Natl Acad Sci USA2000; 97: 6785–6790.

77. Basler C, Reid A, Dybing J, et al. Sequence of the 1918pandemic influenza virus nonstructural gene (NS)segment and characterization of recombinant virusesbearing the 1918 NS genes. Proc Natl Acad Sci 2001;98: 2746–2751.



78. Fanning TG, Slemons RD, Reid AH, Janczewski TA,Dean J, Taubenberger JK. 1917 avian influenza virussequences suggest that the 1918 pandemics virus didnot acquire its haemagglutinin directly from birds.J Virol 2002; 76: 7860–7862.

79. van Elden L, Nijhuis M, Schipper P, Schuurman R,van Loon A. Simultaneous detection of influenzaviruses A and B using real-time quantitative PCR.J Clin Microbiol 2001; 39: 196–200.

80. Schweiger B, Zadow I, Heckler R, Timm H, Pauli G.Application of a fluorogenic PCR assay for typingand subtyping of influenza viruses in respiratorysamples. J Clin Microbiol 2000; 38: 1552–1558.

81. Pease A, Solas D, Sullivan E, Cronin M, Holmes P,Fodor S. Light-generated oligonucleotide arrays forrapid DNA sequence analysis. Proc Natl Acad SciUSA 1994; 91: 5022–5026.

82. Schena M, Shalon D, Davis R, Brown P. Quantitativemonitoring of gene expression patterns with a com-plementary DNA microarray. Science 1995; 270:467–470.

83. Li J, Chen S, Evans D. Typing and subtyping influ-enza virus using DNA microarrays and multiplexreverse transcriptase PCR. J Clin Microbiol 2001; 39:696–704.

84. Jungkind D. Automation of laboratory testing forinfectious diseases using the polymerase chainreaction—our past, our present, our future. J ClinVirol 2001; 20: 1–6.

85. Espy M, Rys P, Wold A, et al. Detection ofherpes simplex virus DNA in genital and dermalspecimens using the IsoQuick, MagNA Pure, andBioRobot 9604 methods. J Clin Microbiol 2001; 39:2233–2236.

86. Razonable R, Brown R, Espy M, et al. Comparativequantitation of cytomegalovirus (CMV) DNA insolid organ transplant recipients with CMV infectionby using two high-throughput automated systems.J Clin Microbiol 2001; 39: 4472–4476.

87. Livache T, Bazin H, Mathis G. Conducting polymerson microelectronic devices as tools for biological ana-lyses. Clin Chim Acta 1998; 278: 171–176.



Molecular diagnosis of influenza

Documents