proceedings of the national microbiology laboratory...

ISSN 1188-4169

November 2006Volume : 32S4

Supplement

Winnipeg, Manitoba7 March, 2006

Proceedings ofthe National Microbiology

Laboratory Pertussis Workshop

Canada Communicable Disease Report

Public HealthAgency of Canada

Agence de santépublique du Canada

To obtain additional copies or subscribe to the Canada CommunicableDisease Report, please contact the Member Service Centre, CanadianMedical Association, 1867 AltaVista Drive, Ottawa, ON, Canada K1G 3Y6,Tel.: (613) 731-8610 Ext. 2307 or 888-855-2555 or by Fax: (613) 236-8864.

This publication can also be accessed electronically via Internet using aWeb browser at http://www.phac-aspc.gc.ca/pphb-dgspsp/publicat/ccdr-rmtc

© Her Majesty the Queen in Right of Canada, represented by the Minister of Health (2006)

This publication was produced by the Scientific Publication and MultimediaServices Section of the Communications Directorate.

Suggested citation: Public Health Agency of Canada.

CCDR 2006;32S4:1-22.

Proceedings of theNational Microbiology Laboratory PertussisWorkshop.

PROCEEDINGS OF THENATIONAL MICROBIOLOGY LABORATORY

PERTUSSIS WORKSHOP

Winnipeg, Manitoba7 March, 2006

Dr. Raymond Tsang, Laboratory for Pathogenic Neisseria, Syphilis, andVaccine Preventable Bacterial Diseases, National Microbiology Laboratory (NML), Winnipeg

Irene Martin, Laboratory for Pathogenic Neisseria, Syphilis, andVaccine Preventable Bacterial Diseases, NML

Table of Contents

Summary. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Recommendations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

Vaccine Pressure and Immune Selection on Vaccine Preventable Bacterial Disease Agents(Dr. Raymond Tsang) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Overview of National Surveillance for Pertussis(Dr. Scott A. Halperin) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

Serologic Tests for the Diagnosis of Pertussis(Dr. Scott A. Halperin) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

The Use of PCR in the Diagnosis of Pertussis(Dr. Susan Richardson) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

A Large Cluster of Pertussis Cases in Toronto, Ontario – Laboratory Investigation(Dr. Frances Jamieson). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

Laboratory Characterization of Bordetella pertussis Strains(Dr. Mark Peppler) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

The Canadian Public Health Laboratory NetworkMeeting Local and Global Challenges in Laboratory Public Health(Dr. Greg Horsman) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

Appendix: List of Participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

i

Proceedings of the National Microbiology Laboratory Pertusis Workshop

Proceedings of the National Microbiology Laboratory Pertussis Workshop

ii

Summary

A National Consensus Conference on Pertussis washeld in 2002, and a number of recommendationswere made for laboratory diagnosis and surveillance.The purpose of the National Microbiology LaboratoryWorkshop on 7 March, 2006, was to gather Canadianexperts in pertussis in order to examine the recom-mendations made in 2002 and discuss issues relatedto laboratory diagnosis and strain characterization forsurveillance. Participants included representativesfrom academia, federal and provincial laboratories,hospital laboratories and industry. Twenty-onepeople attended the workshop, representing NovaScotia, Newfoundland, Quebec, Ontario, Manitoba,Saskatchewan, Alberta, British Columbia and theUnited States.

The objectives of the workshop were to (1) discussand establish recommendations for a diagnosticlaboratory system for pertussis; (2) discuss andestablish recommendations for a laboratory surveil-lance system for pertussis; and (3) develop an actionplan to implement these recommendations. An initialseries of presentations provided a background onpertussis in Canada and consisted of an overview ofthe national surveillance of the disease; serologic andPCR (polymerase chain reaction) diagnosis;laboratory characterization of pertussis; and aninvestigation into recent pertussis cases in Toronto.Further presentations described the NationalMicrobiology Laboratory (NML) activities withrespect to vaccine preventable bacterial diseases, andthe local and global challenges in laboratory publichealth, discussed by the Canadian Public HealthLaboratory Network (CPHLN). The format of theworkshop consisted of presentations and full groupdiscussions.

1


Recommendations

The following is a summary of the recommendationsidentified during the workshop:

1. Recommendation for a diagnostic laboratorysystem for pertussis

1.1 PCR diagnosis for pertussis should beencouraged and made widely available inCanada. The NML should provide alaboratory proficiency program for PCRdiagnosis of pertussis.

1.2 Culturing of the Bordetella pertussisorganism should continue to allow thebacteria to be archived for futurereference, or at least appropriatespecimens (such as nasopharyngealaspirates) should be stored frozen suchthat B. pertussis bacteria may be recoveredfrom the specimen as required.

1.3 The national case definition of pertussisshould be revised to reflect the fact thatlaboratory test results should beinterpreted in the context of the clinicalpresentation of the patient.

1.4 A sero-epidemiology study of pertussis inthe Canadian adult population should beconsidered. This study/project wouldfulfill two objectives: (i) to understand thelevel of protective immunity againstpertussis in the adult population, whichmay provide evidence for futurerecommendation of pertussis vaccinationin this population; (ii) to define the role ofroutine serology for the diagnosis ofpertussis in adolescents and adults.

2. Recommendation for a laboratory surveillancesystem for pertussis:

2.1 A working group with representativesfrom British Columbia, Alberta, Ontario,Nova Scotia, the NML and the Centre forInfectious Disease Prevention and Control(CIDPC) should be established to examinehow to implement a national program tostudy and characterize strains.

2.2 CIDPC should study effective ways toutilize existing resources or programs toimplement sentinel sites for collection ofnot only B. pertussis strains but alsoepidemiologic data, such as vaccinationhistory.

3. Action plan to implement the aboverecommendations:

3.1 CPHLN will initiate a survey to find outwhich Canadian hospital laboratories areproviding PCR diagnosis of pertussis, toidentify sites where culture for B. pertussisis being done, and/or to identify potentialsites where suitable nasopharyngealaspirate specimens may be stored frozenfor future culture work.

3.2 NML will study the feasibility of providinga PCR proficiency program for pertussis.

3.3 NML will set up a working group todiscuss implementing a national straincharacterization program.


2

3.4 CIDPC will study ways to enhance thenational surveillance of pertussis byincorporating the laboratory straincharacterization component.

3.5 CIDPC will take the lead in revising thenational case definition of pertussis toreflect the importance of interpretinglaboratory results in the context ofcompatible clinical and epidemiologicinformation.

3.6 NML will work with interested parties toexamine the feasibility of implementing astandardized enzyme-linked immuno-sorbent assay (ELISA) for measuringserum antibodies to pertussis as well as thefeasibility of carrying out a sero-epidemiology study in the Canadian adultpopulation.

3


Vaccine Pressure and Immune Selection onVaccine Preventable Bacterial Disease Agents

Dr. Raymond Tsang

Selection on vaccine preventable bacterial diseaseagents can result from immunity developed in thepopulation as a result of vaccination (vaccine pres-sure) and/or natural immunity that develops as aresult of the endemic nature of the disease.

Immune selection from vaccine pressure

Pathogens like Streptococcus pneumoniae andHaemophilus influenzae have been reported torespond to vaccine pressure by either “capsuleswitching” or “capsule replacement”. For example, inS. pneumoniae causing invasive pneumococcaldisease, introduction of the heptavalent pneu-mococcal conjugate vaccine (PCV7) has led to theemergence of diseases due to either vaccine-relatedpneumococcal serotypes(1) or non-vaccineserogroups(2). Capsule switching has also beendemonstrated in disease-causing pneumococci(3).Changes in the epidemiology of invasive H. influenzaedisease have been reported from different countries,including Canada, in the post-vaccination(H. influenzae serotype b [Hib]) era(4-6).

In Bordetella pertussis, many genetic polymorphismshave been described, involving such diverse genes asfimbriae, pertussis toxin (PT) and pertactin, whichencode for the virulence factors (and hence arevaccine components)(7-9). Analysis of genomic DNAfrom B. pertussis strains by pulsed-field gel electro-phoresis (PFGE) is able to differentiate betweenrecent clinical isolates and strains used in themanufacture of pertussis vaccines(10,11). These geneticchanges have been suggested to be responsible for theresurgence of pertussis in countries with high rates ofvaccine coverage(8).

Immune selection from natural immunity

Serogroup C Neisseria meningitidis is endemic inCanada, and the currently licensed vaccines against ittarget the serogroup-specific capsule polysac-charide(12). After more than a decade of endemicdisease and many localized outbreaks, genetic andantigenic changes in the subcapsular proteincomponents (the serotype and serosubtype antigens)were detected in the serogroup C ET-15 N.meningitidis strains recovered from cases of invasivemeningococcal disease(13,14). Since the vaccine inducesserum antibodies that target the capsular polysac-charide antigens and the changes documented are inthe subcapsular outer membrane proteins, it isreasonable to hypothesize that such changes arelikely due to selection from natural immunity and areunrelated to vaccine pressure.

Role of the laboratory for surveillance of vaccinepreventable bacterial diseases agents

Since bacteria change over time and some of thesechanges may lead to vaccine escape mutants, it isessential that public health laboratories have a systemto monitor the evolving bacteria for their potential tocause vaccine breakthrough cases. The discussions ofthis workshop will focus on laboratory diagnostic andsurveillance issues for the monitoring of pertussis inCanada.


4

Overview ofNational Surveillance for Pertussis

Dr. Scott A. Halperin

After the introduction of whole cell pertussis vaccinethere was a remarkable decrease in the number ofreported pertussis cases. Over the last 15 years,however, there has been a resurgence of pertussis,although to levels far below the pre-vaccine era(15).Pertussis is a notifiable disease in all provinces andterritories. Cases are reported provincially/territorially, and confirmed cases are reportednationally. The Notifiable Disease SurveillanceSystem is the Public Health Agency of Canada’spassive surveillance system, and the ImmunizationMonitoring Program, Active (IMPACT) is its activesurveillance system. There is no formal system ofB. pertussis strain characterization within thesesurveillance systems.

The Canadian case definitions for pertussis includeboth suspected and confirmed case definitions. Aconfirmed case includes laboratory confirmation ofinfection as defined by isolation of B. pertussis froman appropriate clinical specimen or positive poly-merase chain reaction (PCR) assay for B. pertussis, ora person who is epidemiologically linked to alaboratory-confirmed case with one or more of thefollowing for which there is no other known cause:

� paroxysmal cough of any duration,

� cough ending in vomiting or associated withapnea,

� cough with inspiratory whoop.

A probable case is defined as a cough lasting 2 weeksor longer in the absence of appropriate laboratorytests and not epidemiologically linked to alaboratory-confirmed case and with one or both ofthe following with no other known cause:

� paroxysmal cough,

� inspiratory “whoop”.

The 10 to 14 year age group has demonstrated thegreatest increase in the number of cases. Theprogression of the peak age group has shifted awayfrom infants and preschool-aged children to theseyoung adolescents. In fact, the incidence in the 10 to14 year age group is now second only to that amonginfants < 1 year of age. This shifting epidemiology hasbeen most graphically demonstrated by evaluatingsequential, cyclical outbreaks of pertussis in BritishColumbia(16). Beginning in the year 2000, there was alarge shift in the peak age group of cases to involveyoung adolescents. There is a continual marchingcohort of peaks from 1993 to 2003. This marchingcohort effect has been blunted by implementation ofthe adolescent pertussis vaccine program.

5


In 1994 in Newfoundland and Labrador, the peakincidence of pertussis was among children 1 to4 years old. The peak in this age group disappearedwith the implementation of the acellular pertussisvaccine. Similar diminutions are now being seen inthe adolescent age cohorts that are routinelyreceiving the adolescent formulation of acellularpertussis vaccine.

IMPACT is a national surveillance system based intertiary care pediatric centres across Canada. IMPACTcomprises 12 centres accounting for 75,000 pediatricadmissions annually with over 90% of the tertiary carepediatric beds in Canada. Hospitalized cases ofpertussis have been a surveillance target of IMPACTfor over a decade. There has been a gradual reductionin the annual number of hospital-admitted cases withthe introduction of acellular vaccine. We are seeingfewer vaccine failures and increased effectiveness inpreschool-aged children. There has been a shift inhospitalized pertussis to those infants who have notcompleted their primary immunization series(17).

In summary, the epidemiology of pertussis in Canadais shifting from preschool-aged young children toadolescents and adults. Infants too young to havecompleted the primary series of immunization are atgreatest risk of morbidity and mortality. Differencesin disease burden across jurisdictions are more likelyrelated to the adequacy and intensity of surveillance.Except during outbreaks, laboratories continue to bethe greatest source of pertussis reporting. Laboratorydiagnosis of pertussis depends on the age of patients,lower detection rates being reported in adults andadolescents. This results in underreporting andunderestimation of disease in adolescents and adults.Strain epidemiology is not widely known.


6

Serologic Tests for the Diagnosis of Pertussis

Dr. Scott A. Halperin

The diagnosis of pertussis by serology has a longhistory, and a range of methods have been used. Oneof the oldest tests routinely employed was measure-ment of pertussis agglutinins; however, this testcorrelated best with population immunity and wasless useful for diagnosing pertussis in an individualpatient or for determining whether an individual wassusceptible to infection by B. pertussis. As currentlyperformed, pertussis agglutinins are measured by amicroagglutination method whereby serial dilutionsof a patient’s serum is mixed overnight at 35 ºC withphase 1 B. pertussis organisms, and the agglutinintitre is the highest serum dilution demonstratingbacterial agglutination. Agglutinins have been shownto correlate best with antibodies against fimbriae,pertactin and lipo-oligosaccharide. Because of theirlack of correlation with individual infection orprotection, this assay has no current role in a clinicaldiagnostic laboratory.

Pertussis antibodies have also been measured by theChinese hamster ovary (CHO) cell cytotoxicityneutralization assay. This assay is based upon theobservation that CHO cells undergo a uniquemorphological change in the presence of PT that isneutralized by pertussis antitoxin antibodies. Thisassay measures biologically active antibodies thatcorrelated well with protection in animal models.However, anti-PT antibodies measured by enzymeimmunoassay (see below) have been shown tocorrelate well with CHO neutralization titres and arefar less labour-intensive to perform. Therefore, CHOneutralization titres are only available in certainresearch laboratories and are not routinely used forthe diagnosis of pertussis.

Enzyme immunoassay (EIA) is currently the mostwidely used serologic method for the diagnosis ofpertussis. EIAs can use whole bacterial cells or

specific antigens and can measure IgG, IgA, IgM or allantibody classes. There are no commercial EIA kitsavailable for pertussis serology that performadequately. Whole bacterial antigens have theadvantage of being less costly to prepare but are lesswell-standardized and suffer from cross reaction withother Bordetella species and other organisms (such asH. influenzae). One such whole bacterial EIA kit isused for serologic diagnosis in Australia. All currentlymarketed EIA kits, whether based on whole bacteriaor individual antigens, have problems with antibodyquantification(18).

EIAs for pertussis are also performed in variousreference laboratories using locally prepared reagentsand methods. EIAs measuring anti-PT antibodies arethe most widely available. PT is uniquely producedby B. pertussis, so anti-PT antibodies have the highestsensitivity for B. pertussis infection. There are noknown cross-reacting antibodies to PT, but theantibody response to it is variable, particularly inyoung infants. PT is also part of all currently availableacellular pertussis vaccines, so differentiatingbetween recent vaccine-induced or infection-inducedantibodies is not possible.

Filamentous hemagglutinin (FHA) is a constituent ofthe cell wall of B. pertussis and is a protective antigenin animal models of pertussis. No correlationbetween anti-FHA antibodies and protection wasdemonstrated in the human efficacy studies withacellular pertussis vaccines; however, FHA is acomponent of most acellular pertussis vaccines. FHAinduces a more consistent immune response thandoes PT (increased sensitivity), although anti-FHAantibodies crossreact with antibodies to otherorganisms (other Bordetella species, H. influenzae,Mycoplasma pneumoniae), resulting in decreasedspecificity.

7


Pertactin is an outer membrane adhesin protein (69kilodalton) that is a protective antigen in animalmodels and correlates with protection in humanclinical trials. It induces a somewhat less consistentantibody response than FHA but is more specific toB. pertussis (although other Bordetella speciesproduce a similar protein). It is part of all of theacellular pertussis vaccines currently marketed inNorth America. The assay is more variable than thePT and FHA antibody assays. Fimbriae are alsobacterial surface attachment factors, which elicitantibodies less consistently than FHA. Theanti-fimbriae assays are also less well-standardizedthan the PT or FHA antibody assays. Fimbriae 2 and3 are part of the vaccine currently in widespread usein Canada.

EIAs that measure IgG antibodies are the mostcommonly used, best-standardized and most availableassays. IgG antibodies are detectable 2 to 3 weeks afterprimary immunization and 5 to 7 days after boosterimmunization. Anti-PT antibodies are slower to risethan antibodies against FHA, pertactin or fimbriae.Antibody levels drop rapidly during the first year afterimmunization or infection but remain detectable for 1to 3 years. IgM assays are poorly standardized and onlyoccasionally reported. There is little known about thenatural history of IgM antibodies after pertussisinfection. It is suspected that because of the wide-spread circulation of B. pertussis in the population, IgMantibodies are not generally elicited after exposure tothe organism, since most of the population has beenpreviously exposed.

Measurement of IgA antibodies to pertussis antigenshas the potential to differentiate infection fromimmunization. IgA response is less common afterimmunization than after infection so has beenproposed as a measure of infection. However, the IgAresponse is demonstrated only after 20% to 50% ofinfections. Although less commonly induced afterimmunization, IgA antibodies are measurable insome vaccinees, particularly against the bacterialsurface antigens. Overall, IgA assays have lowsensitivity, particularly in young infants. IgA titres dopersist for up to 1 year and thus are suggestive ofrecent infection (or, in some, recent immunization).

Using multiple antigen/antibody combinationsincreases the sensitivity of the serologic diagnosis ofpertussis but at the expense of specificity. Diagnosticalgorithms have been employed that use either thepresence of antibody against PT alone as a diagnosticlevel of pertussis or the presence of antibody againstat least two of FHA, pertactin or fimbriae. Multipleantigens/antibodies increase the cost and complexityof serologic testing.

Standardization of pertussis serology has beenattempted. In 1995, there was a collaborativeinternational study involving 33 laboratories. Panelsof sera were distributed to various industry, publichealth and academic laboratories, and the resultsfrom laboratories performing the tests werecompared. PT and FHA antibody tests had the bestinter-laboratory correlation; fimbriae and pertactinassays were the most variable(19).

There is a growing consensus among pertussisresearchers and laboratories that for serologicdiagnosis in the clinical diagnostic laboratory, theIgG anti-PT assay provides the best compromisebetween sensitivity and specificity. Paired serologictests (samples from the acute and convalescentstages) is ideal but is rarely obtainable. Single serumdiagnosis is possible based on comparisons withlevels in the general population or using statisticallygenerated diagnostic titre thresholds.

Although there have been a number of researchreports employing such methodology, theMassachusetts Public Health laboratory has led theway in routinely using single serum anti-PT antibodylevels for the diagnosis of pertussis(20). In its assay, apositive result is defined as a titre > 99% of the uppertolerance limit of a population of uninfected controls.This assay has been instrumental in providing labora-tory confirmation of pertussis in older children,adolescents and adults in whom cultures arefrequently negative. Massachusetts initially accountedfor the overwhelming majority of adolescent andadult cases in the United States; however, recentlyconfirmed cases have been reported from otherjurisdictions as well, as the assay becomes morewidely available.


8

A US Centers for Disease Control and Prevention/Food and Drug Administration study evaluatedpertussis serology using sera available from theNational Health and Nutrition Examination Survey.Anti-PT antibodies were measured in over 6,000 USresidents from 6 to 49 years of age. Using a popula-tion modeling approach, they proposed adopting athreshold of � 94 EU (ELISA units)/mL as adiagnostic level of pertussis(21). The EuropeanSero-Epi Network is standardizing results of a varietyof serologic tests from laboratories across Europeusing different methodologies; anti-PT antibodyassays are included. This network has established athreshold of � 125 EU/mL as a diagnostic level forpertussis(22,23).

In summary, the literature suggests that for routine,clinical diagnostic serologic testing, anti-PT IgG EIAis the assay of choice. Given the current state of theart, pertussis serology should likely be limited toselect reference laboratories until standardized kitsare available. For national reportable diseasestatistics, laboratory-confirmed cases by serologyshould be restricted to cases with serology resultsfrom these laboratories. However, even for theselaboratories, proficiency testing is needed. Ideally,serologic assays with single serum pertussis should becorrelated with the US or European standards as acollaborative program. Reference sera should also becorrelated to establish results that are interpretableand comparable among laboratories.

9


The Use of PCR in the Diagnosis of Pertussis

Dr. Susan Richardson

At the time of the 2002 National ConsensusConference on Pertussis, it was recognized that PCRwas a very useful tool for the diagnosis of pertussis,demonstrating enhanced sensitivity of detection ofthe organism over traditional culture techniques.B. pertussis is a slowly growing bacterium that isfastidious with respect to its growth requirements.Despite the most careful handling of specimens andquality control of media, the viability of this organismcan be reduced by toxic factors in the media andfactors in specimen collection and transport. Thus,B. pertussis is an ideal candidate for detection by PCR,and indeed most assays have demonstrated a 2- to4-fold increase in sensitivity by this technique.

Despite this obvious advantage, diagnosis mostcommonly relies on culture or clinical diagnosiswithout verification. This is largely because there areno commercially available pertussis PCR assays andrelatively few laboratories that perform in-houseassays. Although there are eight different Bordetellaspecies, only B. pertussis commonly causes infectionin humans and is the only species associated withoutbreaks of infection. Of the other species, onlyB. parapertussis and B. holmesii cause pertussis-likeillness in humans, and B. bronchiseptica, B. hinzii and

B. trematum have been reported to cause rare cases ofhuman respiratory infection.

The most commonly described PCR assays targetsequences that detect B. pertussis (and/or B. holmesii)or B. parapertussis (Table 1). Insertion sequenceIS481 is by far the most frequently used target forB. pertussis, although it is also present in B. holmesii.Other sites for B. pertussis include the PT promotergene, the porin gene, pertactin gene and adenylatecyclase gene. Insertion sequence IS1001 differentiatesB. parapertussis from B. pertussis, although B. holmesiimay also contain the gene sequence and maytherefore give a positive signal(24).

PCR methodology varies from conventionalsingle-step, block-based PCR to nested PCR andreal-time PCR. Nested PCR can, in theory, providegreater sensitivity but is also more time-consumingand prone to contamination. The advantages ofreal-time PCR are a reduced time to positivity and adecreased risk of contamination. There are very littlecomparative data about the optimal method ofextraction, but most methods seem to be successfulusing specimens collected from the nasopharynx ondacron or rayon swabs (not calcium alginate swabs)or as aspirates, transported in Amies medium with orwithout charcoal or in Regan Lowe medium.


10

Organism

PCR target (genome copies)

IS481 Pertussis toxin promoter IS1001

B. pertussis + (80-100) + (1) -

B. parapertussis –/(+) - + (20)

B. holmesii + (8-10) - –/(+)

B. bronchiseptica –/(+) - -

Table 1. PCR gene targets used to detect Bordetella species

Detection of the amplicon can be carried out by gelelectrophoresis and ethidium bromide staining orprobe-based detection, including FRET (fluorescenceresonance energy transfer) probes, TaqMan probes,molecular beacons and SYBR Green (non-specific).

The sensitivity of most PCR assays is excellent with ananalytic sensitivity of 1 to 10 cfu/mL and a clinicalsensitivity that exceeds culture by 2 to 4 times.Inhibition is not uncommon, especially with mucoidspecimens, and occurs in 1% to 22% of specimens. Itcan, however, be overcome with dilution. Pertussis PCRassays should be monitored with an internal control.

Despite the excellent achievable analytic sensitivity,in practice, as evidenced by results from a multi-centre proficiency panel, the interlaboratory variationis high(25). Of six laboratories using seven assays tostudy dilutions of three test isolates, the range ofdetection was from 4 to 30,000 cfu/mL, showing anunacceptable performance in some laboratories. Asecond proficiency panel that addressed bothsensitivity and specificity found that nested PCRshowed no advantage and that results wereappropriate for the primers chosen(25).

Issues of specificity are important for pertussis PCRassays. For those using the most common primertarget (IS481), it is important to recognize that therewill be cross-reactivity with B. holmesii and someB. bronchiseptica. Therefore, some have advocatedusing a second primer set for confirmation (such asthose targeting the PT promoter or the pertactingene) or IS1001 for B. parapertussis. The use of IS481

and IS1001 should differentiate B. pertussis, B.parapertussis and B. holmesii from one another.However, this is expensive and time-consuming to doon a routine basis and is probably not necessary inoutbreaks, which appear to be confined to B. pertussis.An alternative is to confine verification of specimenswith a second B. pertussis-specific PCR to early in thecourse of an outbreak.

Quite apart from specificity issues with respect toother Bordetella species, the specificity of pertussisPCR assays has been compromised by laboratorycontamination in some highly visible situations in thepast. In the course of two large pertussis outbreaks inNew York State, it was found that 60% of positivePCR tests performed by a private laboratory using anin-house test were not confirmed by a referencePCR(26). A proficiency panel performed by thelaboratory in question showed contamination of thenegative controls. Thus, prevention of contaminationis critical. This is true for any molecular testing,although the potential public health implications ofthe incorrect detection of a reportable communicableillness like pertussis are of particular concern. Ahelpful rule of thumb is that if the rate of positivity ofpertussis PCR increases in relation to culture orgreatly exceeds the usual 2- to 4-fold increase insensitivity over culture, contamination should besuspected. In the New York outbreak, a 100-foldincrease in sensitivity of PCR over culture had beenobserved. This supports the maintenance of culture,in at least some settings, for verification of PCRresults in addition to supporting strain typing, anti-

11


microbial susceptibility testing and identification ofB. parapertussis and B. holmesii.

To prevent contamination, laboratories must observethe usual stringent requirements of a moleculardiagnostics laboratory, i.e. they must have the appro-priate space, staff, training, professional expertise andtight quality control. Sample processing should bekept to a minimum, uracil-N-glycosylase (UNP) canbe included to decrease carryover contamination, andappropriate negative and positive controls must beincluded in each run(27).

A European consensus panel developed recom-mendations in 1994 for PCR detection of pertussis.These include using specific probes for addedsensitivity and specificity; confirming questionableresults with alternative targets; and ensuring that allculture-positive specimens are also PCR positive, thatother commensals and pathogens test negative in the

system, that healthy controls test negative and that aninternal control is used to detect inhibition(27).

At the present time, there is a need to survey thecurrent status of molecular testing for B. pertussis inCanada and to develop up-to-date Canadianguidelines for such testing. Culture should beretained, at least in some centres, as an importantbenchmark for the validity of PCR results and forother testing that may be required, such as straintyping and antimicrobial susceptibility testing. Anational proficiency testing program is a priority,given the variety of testing methods available and thelack of standardization or availability of commercialkits. A Canadian pertussis laboratory working groupcould have useful interactions with other groupsoutside Canada that are interested in a collaborativeapproach to issues surrounding molecular and othertesting for B. pertussis.


12

A Large Cluster of Pertussis Cases inToronto, Ontario – Laboratory Investigation

Dr. Frances Jamieson

The Central Public Health Laboratory (CPHL),Ontario Ministry of Health and Long-Term Care, hasprovided nucleic acid amplification testing (NAAT)for the diagnosis of B. pertussis and B. parapertussissince 1999, using standard PCR based on the methodof Van der Zee et al(28). Culture for B. pertussis andB. parapertussis was continued and is set up in thelaboratory in conjunction with the PCR assay. In Mayof 2005, the laboratory implemented (after a 4-monthcomparison and validation process) a real-time PCRmethod for the detection of B. pertussis and B.parapertussis as a replacement for the standard PCRassay that was in use(29). As part of the validationprocess, using serial end-point dilution it wasdetermined that the detection limit of the real-timeassay is less than one organism per reaction, anincrease in sensitivity over the standard, gel-basedPCR assay. In the real-time assay an internal control(for specimen integrity and inhibition) humanribonuclease protein (RNP) was used, and UNP wasincorporated to prevent carry-over contaminationduring product analysis. The use of probes as thedetection method for real-time PCR assays, as well asthe measures described, greatly decreases the risk offalse-positives. Other standard protocols, such asroutine environmental surveillance for contaminationcontrol, are also in use to decrease the risk offalse-positive results.

In October 2005, the laboratory noticed a largeincrease in the number of specimens submitted forpertussis (2- to 3-fold increase over specimenssubmitted during the same period in 2004-2005). Alarge percentage of these specimens were found to beNAAT positive, culture negative (the percentagedoubling from October 2004 to October 2005). Thepertussis assay runs were reviewed, and no issueswere identified (all controls performed as expectedand environmental surveillance swabs werenegative). Further analysis of the specimenssubmitted indicated that one particular clinic inToronto accounted for approximately 35%, and mostof those (54%) were from patients in the age group1 to 4 years. Compared with previous years, thenumber of positive pertussis cases had increasedsignificantly; almost all of these were NAAT positivealone, most of which were “late-cycle” positives usingthe real-time PCR assay. These late cycle positivescorrespond to a very low organism load (as low as < 1organism) and would not be likely to grow in culture.

A blinded comparison panel of NAAT negative,positive and “weak” positive specimens wasexchanged with the molecular laboratory of theDivision of Microbiology, Department of PediatricLaboratory Medicine, at the Hospital for SickChildren (Dr. Susan Richardson). The results fromthe CPHL and the Hospital for Sick Children

13


correlated 100%. RSV (respiratory syncytial virus)culture material was also tested using the real-timepertussis PCR assay and was negative (nofalse-positive results). The PCR products from thereal-time assay were analyzed by gel electrophoresis.Positive specimens were found to have bands thatwere at the appropriate molecular weight (~120 bp),and sequencing of these specimens determined thatthey were 100% homologous to B. pertussis.

Therefore, the positive real-time PCR assay results onthe specimens received during the investigation werenot false-positive laboratory results, but they havegenerated a number of questions:

� Is there a transient carrier state (low numbersof organism present)?

� Was the assay detecting dead organism (e.g.patients had received antibiotics at the time ofspecimen sampling)?

� What is the clinical utility of molecular assaysfor pertussis detection, given the high sensitiv-ity of the real-time PCR assay?

� What role does vaccine efficacy play?

� These questions and others generated from thediscussion will be investigated further with acase-control study to be conducted jointly bythe CPHL, the Hospital for Sick Children, theToronto Public Health Unit and otherresearchers.


14

Laboratory Characterization of B. pertussis Strains

Dr. Mark Peppler

The utility was demonstrated of laboratory char-acterization of strains using the epidemiology ofnorthern Alberta clinical isolates from 1995 to 2002and monitoring changes in strains over time withthe use of PFGE and allelic typing of PT S1 subunitand pertactin. A B. pertussis isolate possessing both“old” ptxS1 and prn alleles is termed “old”; posses-sing both “new” ptxS1 and prn alleles is termed“new”; and possessing one “old” and one “new”ptxS1 and prn allele is termed “transitional” (Table2). The distribution from old to transitional to newstrain types has changed. The majority of isolatesare in the “new” category (Table 3). The mostprominent PFGE type during this period wasBpeXba008 (Table 4). No difference in distributionwas noted among those aged > 21 as compared with< 21-year-olds.

Despite this shift in strain types over time, nosignificant difference could be seen in the strains thatcaused clinical disease in vaccinated versus unvac-cinated patients. There were 1,300 cases from 1995onward, 758 of which were culture-positive, resulting

in 278 complete records for which there were strainsto analyze. The majority of cases were in the infantgroup. The data support the conclusion that B.pertussis has not escaped from vaccine-inducedimmunity by antigenic drift.

Continued surveillance of strains across Canada tomonitor future changes over time is recommended. Anational commitment to support and standardizethese studies is required, and allelic typing andserotyping of fimbriae should be added to theprofiles.

Table 2. Vintage of pertactin (prn) and pertussistoxin S1 (ptxS1) alleles

“Old” (presentbefore 1970)

“New” (emergedsince 1985)

ptxS1prn

B, D, E1

A2, 3, 9

15



16

prn2ptxS1A

“transitional”

prn2ptxS1B

“transitional”

prn1ptxS1B

“old”

prn2ptxS1A“new”

pr3ptxS1A“new”

prn9ptxS1A“new” Totals

1985-1994 301 (15.8) 0 (0) 593 (31.2) 867 (45.6) 109 (5.7) 31 (1.6) 1,901

16% 31% 54% 100%

1995-2002 60 (21.6) 2 (0.7) 0 (0) 178 (64.0) 6 (2.2) 32 (11.5) 278

22.3% 0% 77.7% 100%

Table 3. Estimated total numbers (percentages) for each combination of prnand ptxS1 alleles found in the top 30 PFGE types combined from Alberta

PFGE type

ptxS1A and prn types ptxS1B and prn typesTotals by

PFGE type1 2 3 9 1 2 3 9

BpeXba001 5 5

BpeXba002 16 1 17

BpeXba004 2 2

BpeXba007 1 1

BpeXba008 1 178 3 1 1 184

BpeXba010 2 2

BpeXba011 4 4

BpeXba020 2 2

BpeXba024 1 1

BpeXba051 5 5

BpeXba072 1 1

BpeXba101 2 2

BpeXba102 3 3

BpeXba106 1 1

BpeXba107 43 2 45

BpeXba109 1 1

BpeXba116 2 2

Total 278(%)

60(21.6)

178(64.0)

6(2.2)

32(11.5)

0(0)

2(0.7)

0(0)

0(0)

278

Table 4. Distribution of PFGE and ptxS1/prn types, 1995-2002

The Canadian Public Health Laboratory Network:Meeting Local and Global Challenges

in Laboratory Public Health

Dr. Greg Horsman

The Canadian Public Health Laboratory Network

The Canadian Public Health Laboratory Network(CPHLN) is a nationally minded, proactive forum ofpublic health laboratories. Since its formation,CPHLN has forged strong functional bonds linkingfederal and provincial public health laboratories. TheCPHLN has been a voice of laboratory advocacy, andits federal and provincial member laboratories havebeen providing front-line response to naturallyoccurring infections and deliberately introducedbioterror-related agents.

The CPHLN was established in 2001. In response tothe terrorism events of 11 September, 2001, and thesubsequent anthrax attacks in Washington DC in2001, the CPHLN had also elected to include bioterrorresponse as one of its responsibilities. In addition toaddressing concerns about the increased frequency andpotential lethality of bioterrorism agents, the scope ofthe Network includes other aspects of public health,such as water and food safety in response to water-borne outbreaks in Walkerton, Ontario, and NorthBattleford, Saskatchewan. As well, it has expanded todevelop strategies to advance laboratory standardiza-tion of diagnostic testing across Canada and aninitiative to enhance reference testing. Moreover, theCPHLN recognizes the impact of infectious diseases onchronic disease and has taken steps to enhanceprograms that will address infectious diseases that havean impact on chronic disease.

The CPHLN recognizes the responsibility of publichealth laboratories to safeguard public health and toprovide surveillance and response capabilities thatserve as a cornerstone for all public health professionals.

The role of CPHLN is to provide a forum for publichealth laboratory leaders to share best practices andlessons learned in an atmosphere of trust. The role ofthe CPHLN is also to leverage its combined strengthto champion for the sustainability and support ofpublic health laboratories in their efforts to providerapid and coordinated laboratory response tocommunicable disease agents.

The Pan-Canadian Public Health Network

The CPHLN does not stand alone. CPHLN is part of anational framework for local, national and globalpublic health. It is part of the Pan-Canadian PublicHealth Network federal/provincial/territorialenvironment.

Under the reorganization within Health Canada andthe formation of the new Public Health Agency ofCanada (PHAC), the CPHLN is one of six otherexpert groups that create the framework for publichealth activities in Canada. These expert groups areas follows:

� Communicable Disease Control

� Canadian Public Health Laboratories

� National Emergency Preparedness andResponse

� Surveillance and Information

� Non-communicable Disease and InjuryPrevention

� Health Promotion

17


The groups are accountable to the F/P/T PublicHealth Network Council, which collates theirfindings and reports to the F/P/T Conference ofDeputy Ministers of Health.

The membership of the CPHLN consists of provincialpublic health laboratories (10), the PHAC, DefenceResearch and Development Canada, the Council ofChief Medical Officers of Health, Canadian BloodServices, Héma-Québec and the Canadian FoodInspection Agency.

The CPHLN core comprises medical and scientificlaboratory directors but has access to a broadspectrum of diagnostic and technical expertise fromwithin each of Canada’s public health laboratories.

The PHAC has representation in CPHLN from thefollowing;

� Laboratory for Foodborne Zoonoses

� Centre for Emergency Preparedness andResponse

� Centre for Infections Disease Prevention andControl

� National Microbiology Laboratory

� HIV/AIDS/Retrovirology

Cross-cutting issues

There has been an expansion in laboratory focus overtime, and the linkage of the nation’s public healthlaboratories continues to grow with cooperation andcollaboration on cross-cutting issues. We are buildinglaboratory linkage to epidemiology and otheremergency response centres for infectious diseaseresponse. We are in the process of developingstrategies that link management of infectious diseasewith prevention of chronic disease, where applicable.Our plans include the inter-jurisdictional develop-ment of integrated approaches to address cross-cuttingpublic health issues. The development of a nationalcommand and control structure with cooperativeinterfacing and integration with local hierarchicalreporting structures and protocols is part of ouremergency planning.

The CPHLN has subcommittees that are generallyjointly co-chaired by federal and provincial/territorialrepresentatives. The subcommittees are the

� Laboratory Standardization Subcommittee

� Bioterror Response Subcommittee

� Food & Water Safety Subcommittee

� Reference Centre Advisory Subcommittee.

There are working groups set up periodically. Thereis a working group that advises on Pulse Net Canada.This group plays an important role in implementingthe standards that allow inter-laboratory comparisonof molecular fingerprints of various pathogens.Recently, the Laboratory Influenza group was rolledinto the larger Pandemic Influenza PreparednessNetwork, which expanded into areas of epidemiologyand viral diagnostics, and worked with CanadianBlood Services and the Office of Laboratory Securityto develop the laboratory annex to the NationalInfluenza Pandemic Plan. There is a Lyme DiseaseWorking Group developing guidelines on laboratorytesting for Lyme disease in Canada. The WorkingGroup on Laboratory Self-Assessment will develop atemplate that will be used to create a database ofpublic health testing in Canada.

The Laboratory Standardization Subcommittee ofCPHLN reviews the nationally notifiable diseases, ofwhich pertussis is one. This is to ensure that the casesreported across Canada can be evaluated by epidemio-logists properly. The PulseNet Canada Group isinvolved in standards for PFGE for various enteric andnosocomial pathogens. The Reference Centre AdvisorySubcommittee reviews the need for public healthdefence tests. Each of these groups could have a role inimplementing recommendations developed at thePertussis Workshop.

In summary, the CPHLN develops partnerships withina national framework. These partnerships involvecollaboration on laboratory standards for nationallynotifiable diseases, molecular diagnostics in publichealth, Web-driven business and communicationtools, ownership documents and authorshipdocuments. More information on these activities canbe accessed at www.CPHLN.ca.


18

References

1. Steenhoff AP, Shah SS, Ratner AJ et al. Emer-gence of vaccine-related pneumococcal serotypesas a cause of bacteremia. Clin Infect Dis2006;42:907-14.

2. Byington CL, Samore WH, Stoddard GJ et al.Temporal trends of invasive disease due to Strep-tococcus pneumoniae among children in theIntermountain West: Emergence of nonvaccineserogroups. Clin Infect Dis 2005;41:21-9.

3. Jefferies JMC, Smith A, Clarke SC et al. Geneticanalysis of diverse disease-causing pneumococciindicates high levels of diversity within serotypesand capsule switching. J Clin Microbiol2004;42:5681-8.

4. Bajanca P, Canica M, Multicentre Study Group.Emergence of nonencapsulated and encapsulatednon-b-type invasive Haemophilus influenzae iso-lates in Portugal (1989-2001). J Clin Microbiol2004;42:807-10.

5. Ribeiro GS, Reis JN, Cordeiro SM et al. Preven-tion of Haemophilus influenzae type b (Hib)meningitis and emergence of serotype replacementwith type a strains after introduction of Hib immu-nization in Brazil. J Infect Dis 2003;187:109-16.

6. Tsang RSW, Mubareka S, Sill ML et al. InvasiveHaemophilus influenzae in Manitoba, Canada,in the postvaccination era. J Clin Microbiol2006;44:1530-5.

7. King AJ, Berbers G, van Oirschot HF et al. Roleof the polymorphic region 1 of the Bordetellapertussis protein pertactin in immunity.Microbiol 2001;147:2885-95.

8. Mooi FR, van Oirschot H, Heuvelman K et al.Polymorphism in the Bordetella pertussis viru-

lence factors P.69/pertactin and pertussis toxin inthe Netherlands: Temporal trends and evidence forvaccine-driven evolution. Infect Immun1998;66:670-5.

9. Tsang RS, Lau AK, Sill ML et al. Polymorphismsof the fimbria fim3 gene of Bordetella pertussisstrains isolated in Canada. J Clin Microbiol2004;43:5364-7.

10. Peppler MS, Kuny S, Nevesinjac A et al. Strainvariation among Bordetella pertussis isolatesfrom Quebec and Alberta provinces of Canadafrom 1985 to 1994. J Clin Microbiol2004;41:3344-7.

11. Guiso N, Boursaux-Eude C, Weber C et al.Analysis of Bordetella pertussis isolates collectedin Japan before and after introduction of theacellular pertussis vaccine. Vaccine2001;19:3248-52.

12. Granoff DM, Feavers IM, Borrow R.Meningococcal vaccines. In: Plotkin SA,Orenstein WA, eds. Vaccines, 4th ed. Philadel-phia: Saunders, 2004:959-87.

13. Tsang RSW, Tsai CM, Zhu P et al. Phenotypicand genetic characterization of a unique variant ofserogroup C ET-15 meningococci (with the anti-genic formula C:2a:P1.7,1) causing invasivemeningococcal disease in Quebec, Canada. J ClinMicrobiol 2004;42:1460-5.

14. Law DKS, Henderson AM, Tsang RSW et al.DNA sequencing analysis of the PorB protein ofnonserotypeable serogroup C ET-15 meningococcisuggests a potential mutational hot spot on theirserotype antigens. J Clin Microbiol2004;42:2718-23.

19


15. Skowronski DM, De Serres G, MacDonald D etal. The changing age and seasonal profile of per-tussis in Canada. J Infect Dis 2002;185:1448-53.

16. Galanis E, King AS, Varughese P et al. Changingepidemiology and emerging risk groups for pertus-sis. Can Med Assoc J 2006;174:451-2.

17. Bettinger JA, Halperin SA, DeSerres G et al. Epi-demiology of hospitalized pertussis after changefrom whole cell to acellular pertussis vaccine. Pre-sented at the 6th Canadian National Immuniza-tion Conference, Session #4, Montreal,December, 2004.

18. Kösters K, Riffelmann M, Dohrn B et al. Com-parison of five commercial enzyme-linkedimmunosorbent assays for detection of antibodiesto Bordetella pertussis. Clin Diagn LabImmunol 2000;7:422-6.

19. Lynn F, Reed GF, Meade BD. Collaborativestudy for the evaluation of enzyme-linkedimmunosorbent assays used to measure humanantibodies to Bordetella pertussis antigens. ClinDiagn Lab Immunol 1996;3:689-700.

20. Yih WK, Lett SM, des Vignes FN et al. Theincreasing incidence of pertussis in Massachusettsadolescents and adults, 1989-1998. J Infect Dis2000;182:1409-16.

21. Baughman AL, Bisgard KM, Edwards KM et al.Establishment of diagnostic cutoff points for levelsof serum antibodies to pertussis toxin, filamentoushemagglutinin, and fimbriae in adolescents andadults in the United States. Clin Diagn LabImmunol 2004;11:1045-53.

22. Pebody RG, Gay NJ, Giammanco A et al. Theseroepidemiology of Bordetella pertussis infectionin Western Europe. Epidemiol Infect2005;133:159-71.

23. Giammanco A, Chiarini A , Maple PA et al.European Sero-Epidemiology Network: Standardi-sation of the assay results for pertussis. Vaccine2003;22:112-20.

24. Riffelmann M, Wirsing von Konig CH, Caro Vet al. Nucleic acid amplification tests for diagnosisof Bordetella infections. J Clin Microbiol2005;43:4925-9.

25. Muyldermans G, Soetens O, Antoine M et al.External quality assessment for molecular detec-tion of Bordetella pertussis in European laborato-ries. J Clin Microbiol 2005;43:30-5.

26. Lievano FA Reynolds MA, Waring AL et al.Issues associated with and recommendations forusing PCR to detect outbreaks of pertussis. J ClinMicrobiol 2002;40:2801-5.

27. Meade BD, Bollen A. Recommendations for use ofthe polymerase chain reaction in the diagnosis ofBordetella pertussis infections. J Med Microbiol1994;41:51-5.

28. Van der Zee A, Agterberg C, Peters M et al.Polymerase chain reaction assay for pertussis:simultaneous detection and discrimination ofBordetella pertussis and Bordetellaparapertussis. J Clin Microbiol1993;31:2134-40.

29. Kösters K, Riffelmann M, Wirsing von KonigCH. Evaluation of a real-time PCR assay fordetection of Bordetella pertussis and B.parapertussis in clinical samples. J MedMicrobiol 2001;50:436-40.


20

Appendix

List of Participants

Dr. Sadjia BekalLaboratoire de santé publique du QuébecSainte-Anne-de-Bellevue, Quebec

Dr. Swee Han GohBC Centre for Disease ControlVancouver, British Columbia

Dr. Scott A. HalperinDalhousie UniversityIWK Health CentreHalifax, Nova Scotia

Dr. Greg HorsmanSaskatchewan Provincial Public Health LaboratoryRegina, Saskatchewan

Dr. Frances JamiesonCentral Public Health LaboratoryEtobicoke, Ontario

Dr. Scott KapoorToronto Public HealthUniversity of TorontoToronto, Ontario

Dr. Mark PepplerUniversity of AlbertaEdmonton, Alberta

Dr. Sam RatnamNewfoundland Public Health LaboratorySt. John’s, Newfoundland

Dr. Susan RichardsonThe Hospital for Sick ChildrenToronto, Ontario

Dr. Gregory J. TyrrellProvincial Laboratory for Public HealthUniversity of AlbertaEdmonton, Alberta

Dr. John WylieCadham Provincial LaboratoryWinnipeg, Manitoba

Sanofi Pasteur Limited:

Dr. Luis BarretoToronto, Ontario

Dr. George CatesToronto, Ontario

Dr. David GreenbergSwiftwater, Pennsylvania

Dr. Pierre LavigneToronto, Ontario

21


Public Health Agency of Canada:

Cathy JorowskiCanadian Public Health Laboratory NetworkNational Microbiology LaboratoryWinnipeg, Manitoba

Irene MartinPathogenic Neisseria and Vaccine PreventableBacterial Diseases DivisionNational Microbiology LaboratoryWinnipeg, Manitoba

Christine NavarroImmunization and Respiratory Infections DivisionCentre for Infectious Disease Prevention andControlOttawa, Ontario

Michelle SillPathogenic Neisseria and Vaccine PreventableBacterial Diseases DivisionNational Microbiology LaboratoryWinnipeg, MB

Dr. Theresa TamImmunization and Respiratory Infections DivisionCentre for Infectious Disease Prevention andControlOttawa, Ontario

Dr. Raymond TsangPathogenic Neisseria and Vaccine PreventableBacterial Diseases DivisionNational Microbiology LaboratoryWinnipeg, Manitoba


22

proceedings of the national microbiology laboratory...

Documents