Laboratory case definitions

Pertussis laboratory case definition

The Public Health Laboratory Network have developed a standard case definition for the diagnosis of diseases which are notifiable in Australia. This page contains the laboratory case definition for pertussis.

Version: 1
Authorisation: PHLN
Consensus Date: 6 June 2007

Author:
Prof. Lyn Gilbert
Institute of Clinical Pathology and Medical Research
Level 3 Westmead Hospital
Institute Road
WESTMEAD NSW 2145
Telephone: 02 9845 6252.
Facsimile: 02 9893 8659
Email: lyng@icpmr.wsahs.nsw.gov.au

1 PHLN SUMMARY LABORATORY DEFINITION

1.1 Condition:

Pertussis

1.1.1 Definitive Criteria

• Isolation of Bordetella pertussis; OR
• Detection of B. pertussis by nucleic acid test (NAT).

1.1.2 Suggestive Criteria

• Seroconversion or significant increase in antibody level to B. pertussis; OR
• Single high IgA level to B. pertussis.

2 INTRODUCTION

Pertussis is caused by Bordetella pertussis, a tiny, fastidious, Gram negative cocco-bacillus, which was first isolated in 1906 by Bordet and Gengou. It is restricted to the respiratory tract of humans and is spread by droplets from person to person. B. pertussis produces a range of virulence factors including pertussis toxin and pertactin which cause the typical disease symptoms. The clinical spectrum is diverse and is affected by patient age, previous exposure to the organism, immunisation history, antibiotic administration and concomitant infections with other agents. Clinical expression ranges from asymptomatic infection in children and adults with strong residual immunity to more severe and life-threatening disease in unprotected newborns and infants and potentially, the elderly. Although some previously immunized individuals do develop classical clinical symptoms, less typical pertussis, characterized by the absence of typical whoop and a shorter duration of cough, is more common among adolescents and adults.

Definitive laboratory diagnosis is made by isolation of B. pertussis from respiratory specimens but culture is insensitive as the disease progresses and there are fewer organisms present after the first 2–3 weeks of symptoms, which are usually, initially, non-specific. Polymerase chain reaction (PCR) is more sensitive than culture and is accepted as definitive evidence of disease in cases with an appropriate clinical history. Serology is used extensively in Australia for diagnosis of pertussis, particularly in adults and adolescents. Many other countries including the USA do not accept a diagnosis of pertussis based on serology. Direct fluorescence assay (DFA) has been used in the past but has been now superseded by PCR which offers a more sensitive and specific diagnosis.

Recommended laboratory methods for B. pertussis are detailed on the Centers for Disease Control (CDC) website http://www.cdc.gov/vaccines/pubs/pertussis-guide/downloads/_DRAFT_chapter2_amended.pdf and European website: www.eupertstrain.org

Top of page

3 TESTS

3.1 Culture

Culture should be attempted if possible, to allow typing of an isolate (to monitor changes in circulating endemic strains including variation in virulence genes, which could affect vaccine efficacy)¹³ and antibiotic sensitivity testing, although macrolide resistance (MIC >256mg/L for erythromycin and MIC >2mg/L for Azithromycin) is rare.

3.1.1 Suitable and unsuitable specimens

Nasopharyngeal aspirate, nasopharyngeal swab (obtained using flexible shafted swabs with Dacron™ or rayon tips, not cotton). Ideally the swab is left in the posterior pharynx for 10 seconds before withdrawing. Throat and anterior nasal swabs have low rates of recovery.² In young children nasal wash specimens may be satisfactory.

Since Bordetella pertussis is a fastidious organism, specimens should be plated onto culture media immediately, preferably at the bedside. Special transport media (such as Regan-Lowe medium) containing half strength charcoal agar supplemented with horse blood and cephalexin (40mg/L) are available but not widely used. Even in transport medium, specimens should not be stored for more than 48 hours before culture. Amies transport medium with charcoal is not recommended for culture but interferes only slightly with PCR.³

3.1.2 Media

Bordet and Gengou used a medium containing blood, glycerine and potato extract for their original isolation and variations of this medium, subsequently called Bordet-Gengou medium, continue to be used worldwide. CDC recommend a variant which uses 10% defibrinated horse blood and cephalexin as a supplementary primary isolation medium.² B. pertussis is an aerobe and agar plates should be incubated at 35–360°C, usually for a maximum of 7 days, in air (high humidity) rather than CO₂ . Plates should not be allowed to dry out.

B. pertussis colonies take 3 to 4 days to develop their typical morphology “half pearl” colonies. They are small, gram-negative, catalase positive cocco-bacilli. B. pertussis is urease negative, unlike B. parapertussis and B. bronchiseptica. Oxidase reaction is also helpful in differentiating B. pertussis (positive) from B. parapertussis (negative). Most laboratories use specific antisera to identify B. pertussis and B. parapertussis.

3.1.3 Test sensitivity

Successful isolation depends on:

• stage of the illness – highest sensitivity during catarrhal stage of 1–2 weeks duration with lower sensitivity during paroxysmal stage 3–6 weeks, and negative during convalescence. Low (0–30%) for adults who typically present late in disease.¹¹ Bacterial load is higher in children.
• quality of the specimen collection – nasopharyngeal aspirates are marginally more sensitive than a nasopharyngeal Dacron™ swab. Anterior nose and throat swabs and classical “cough plates” are no longer recommended for culture.
• speed of plating and quality of the medium – specimens should be plated immediately (at least within 4 hours after collection) onto high quality non-selective media designed for B. pertussis isolation.
• Prior antibiotic treatment will negatively affect sensitivity.

Top of page

3.1.4 Test specificity

100% for Bordetella pertussis.

3.1.5 Predictive values and relevant populations

Positive predictive value is 100%. Negative predictive is variable but is highest in young, unvaccinated children early in disease. Negative predictive value is low for sporadic cases of disease in adults, who generally present later in disease and in previously vaccinated individuals.

3.1.6 Suitable internal controls

Properly documented, relevant, quality control programme for each type and batch of medium used. B. pertussis Hohama-1 has been used as the reference strain.

3.1.7 Suitable external quality assurance program (QAP) program

None available.

3.1.8 Special considerations

Care should be exercised when using specific antisera to identify the different species as B. parapertussis may weakly agglutinate B. pertussis antisera.¹¹

3.2 Nucleic Acid Detection methods

PCR has been used since the early 1990s to detect B. pertussis in respiratory specimens and has generally been found to be more sensitive and give a faster result than culture. Most assays target IS481 sequences for the detection of B. pertussis and IS1001 for B. parapertussis. Some laboratories also target pertussis toxin sequences to confirm positive IS481 PCR results (this is the optimal strategy but is expensive). Protocols for LightCycler™ and TaqMan PCR systems may be found on the CDC and Eupertstrain websites.^2,5

3.2.1 Suitable and unsuitable specimens

Nasopharyngeal aspirates or nasopharyngeal swabs with Dacron™ or rayon tips are optimal and calcium alginate swabs should not be used.³ In contrast to culture, dry swabs may be used for PCR. Throat swabs and sputum samples may be used for adolescents and adults but the performance characteristics of assays using these samples should be validated by each laboratory. In Australia throat swabs are the specimens most commonly collected for PCR diagnosis of pertussis in adults and older children. Amies transport medium with charcoal does not significantly interfere with PCR. A method for Nucleic Acid (NA) extraction from respiratory secretions can be found at www.eupertstrain.org

3.2.2 Test details

Riffelmann et al¹⁵ have provided a list of primer-probe combinations published prior to 2005.

Top of page

3.2.3 Test sensitivity

Sensitivity depends on the age of the patient, stage of disease, specimen collection, NA extraction procedure and test format.¹⁵ PCR is more sensitive and remains positive until significantly later in disease than culture and is recommended for diagnosis in adolescents and adults, who typically are more difficult to sample than young children, although culture should still be attempted. PCR may remain positive when culture becomes negative after antibiotic treatment.

Five studies reviewed by Tozzi et al¹⁷ reported culture sensitivities ranging from 1–30% compared with PCR which had sensitivities from 3–50% in symptomatic adults. An Australian study 6 demonstrated that throat swabs are useful for diagnosis by PCR but this study was not designed to compare the positivity rate of one specimen type with the other. The authors caution that the PCR method employed should be sufficiently sensitive to detect the lower number of organisms likely to be present in throat swabs. A Danish study found no significant difference in sensitivity between peroral nasopharyngeal swabs and pernasal nasopharyngeal swabs.⁴

Recent reports demonstrate that real time PCR and block-based PCR methods targeting insertion sequences (IS) sequences have similar sensitivity and are more sensitive than tests targeting the pertussis toxin gene (ptx-A).¹⁵ Multiplex real-time PCR for both B. pertussis and B. parapertussis are less sensitive than uniplex PCR¹⁵ (1-10 organisms per reaction for the ptxA-Pr PCR and 0.1–1 organisms per reaction for the IS481 PCR which is present in multiple copies per genome e.g. 238 copies in B. pertussis Tohama-1).

3.2.4 Test specificity

Specificity is very high for all targets. Low numbers of IS481-like elements have been reported in B. holmesii and B. bronchiseptica genomes so primers designed to detect B. pertussis IS481 sequences could in theory, cross react with B. holmesii DNA. However, surveillance in populations where B. holmesii is known to be present has not detected any cross-reactions when B pertussis positive specimens have been retested using PCR targeting B. holmesii specific sequences.¹⁵

3.2.5 Predictive values and relevant populations

Positive predictive value is high in symptomatic individuals, but low in asymptomatic individuals who have had household or other close contact with a patient or in outbreak situations.⁸ Negative predictive value depends on the age of the patient, stage of disease, specimen collection, NA extraction procedure and test format.¹⁷ Infected people who have history of previous vaccination or infection typically have a less severe infection and may excrete fewer organisms.

3.2.6 Suitable test acceptance criteria

Results for control samples obtained as expected.

3.2.7 Suitable internal controls

As recommended in the National Pathology Accreditation Advisory Committee (NPAAC) guidelines: Laboratory Accreditation Standards and Guidelines for Nucleic Acid Detection and Analysis . Controls should be designed to detect sample inhibitory activity and external contamination by Bordetella amplicons.

3.2.8 Suitable original test validation criteria

Consistent with NPAAC Guidelines:

Requirements for the Validation of In-House In vitro Diagnostic Devices .

3.2.9 Suitable external QAP program

RCPA QAP P/L nucleic acid detection for Bordetella pertussis. Of 33 laboratories participating in dispatches 3 and 7 in 2006, approximately 80% used real time PCR platforms. Extraction procedures included boiling with and without Chelex, Roche High Pure, Roche MagNA Pure and Qiagen DNA. No false positives were recorded in either dispatch. A number of laboratories failed to detect the lowest concentrations in serially diluted specimens, which could limit diagnostic sensitivity, especially for throat swabs (with relatively low bacterial numbers).

3.2.10 Special considerations

PCR may be positive in asymptomatic individuals who have had contact with cases.⁸

Top of page

3.3 Serology

Antibody assays have been used for diagnosis of pertussis since the organism was first isolated. Early tests included agglutination and neutralizing antibody titres (using Chinese hamster ovary cells as targets for pertussis toxin).¹¹ However, there is no consensus internationally on the role of serology in the diagnosis of pertussis. The CDC laboratory case definition which is standardized nationally does not include serology because it believes that “no serologic method for diagnosis of pertussis has been validated between laboratories or has been accepted for diagnostic use in U.S”.² WHO accepts seroconversion or a significant increase in antibody level but not a single high titre. In Australia, single sample serology is used extensively, particularly in adolescents and adults who typically present late in disease when an antibody rise has already occurred and culture and PCR are unlikely to be positive. The majority of pertussis notifications to public health authorities in Australia are based on a single IgA result see Table 1 - data from National Notifiable Diseases Surveillance Scheme as reported in Quinn HE, McIntyre PB, Pertussis epidemiology in Australia over the decade 1995-2005 – trends by region and age group. Communicable Diseases Intelligence 2007; 31 (in press).¹⁴

Table 1

Diagnostic method	% by age group
	<1	1–4	5–9	10–19	20–59	60+
Culture	9.6	3.1	1.9	1.3	0.9	0.9
Nucleic Acid testing	59.7	39.1	21.3	11.1	7.1	4..3
Serology	8.7	26.4	52.5	73.8	81.3	88.0

Serological methods used for pertussis diagnosis include complement fixation test (CFT), immunofluorescence (IF), immunoblot and enzyme immune assay (EIA). Commercial supplies of reagents for all assays are available. Several commercial EIA kits are marketed in Australia and overseas and they use a variety of antigens including whole cell lysates, pertussis toxin (PT) and filamentous haemagglutinin (FHA) and may detect IgA, IgM or IgG as the conjugate used. Royal College of Pathologists of Australasia (RCPA) QAP surveys for pertussis serology reveal that 82% of participating Australian laboratories used the PanBio Bordetella pertussis IgA ELISA (whole cell antigen) and 15% used the Savyon Seropertussis IgA ELISA (FHA and PT antigens) in 2006. IgA antibody is preferred to IgG because it is generally less likely to be produced in significant amounts after vaccination and falls more readily after infection than IgG. It is therefore, said to be a more specific marker of recent infection. Quantitative assays for IgG against pertussis toxin have been used as the standard serological test for recent infection in vaccine trials. Levels of 100IU/mL or greater are highly predictive of recent infection and can be used as a gold standard for evaluation of other tests.

In 2006 a marked change in the number of pertussis notifications to public health authorities in Australia cast doubts on the predictive values of PanBio and Savyon kits. A pilot study was undertaken in early 2007 at Centre for Infectious Diseases & Microbiology (CIDM), ICPMR to investigate the performance of these kits in comparison with other commercial EIAs, using CFT, IF and Western Blot as a combined gold standard. http://www.health.nsw.gov.au/infect/pdf/false_pertussis_tests.pdf

Neither the PanBio nor Savyon kit performed satisfactorily in comparison with some other kits and the gold standard.

Top of page

3.3.1 Suitable and unsuitable specimens

Serum

3.3.2 Test sensitivity

The sensitivity of all pertussis serology varies with the format, antigen used and antibody class detected and stage of illness. Sensitivity is difficult to measure because of the low sensitivity of the gold standard, culture (and also the combined serological gold standard). An early study, using CFT and IF directed against IgG, detected a significant rise in titre in 18/24 cases over 6 months old who were culture positive.¹ They confirmed earlier reports that antibody production is poor during the first 6 months of life. The CIDM study reported sensitivities of the 11 EIA kits compared with the serological gold standard (see above) to vary between 33–85%.

3.3.3 Test specificity

Specificity of tests using purified toxin as the sole antigen is much higher than those using whole cell or fimbrial antigens which cross react with other organisms. Pertussis toxin antibodies are specific for B. pertussis, but antibodies to filamentous haemagglutinin, pertactin, fimbriae and whole sonicated organism may be raised after infection with other Bordetella spp including B. parapertussis.⁷ The filamentous haemagglutinin may cross react with Hæmophilus influenzae, Mycoplasma pneumoniae, and Chlamydia pneumoniae.⁷ Specificity of serology for disease is low in vaccinated and older populations who may have antibodies boosted by repeated exposure to disease in the community. The CIDM study reported specificities of EIA kits tested to vary between 76–98.9%.

Unfortunately the EIA assay which used PT as the sole antigen had very poor sensitivity. Manufacturers add additional antigens such as FHA to PT to boost the sensitivity of their tests but this leads to a loss in specificity. Immunoblot assays will demonstrate whether the antibody detected is directed against PT or FHA and can be used to further characterise an immune response measured by EIA in cases where significant public health action is contemplated.

3.3.4 Predictive values and relevant populations

Positive predictive values of IgG seroconversion and significant increases in levels of antibody are high. Negative predictive values are variable and are influenced by age of the patient and format of the test.

Predictive values of single levels are significantly influenced by the level of disease circulating in a population and its vaccination coverage. The CIDM study reported likelihood ratios (which are a measure of both sensitivity and specificity irrespective of prevalence of infection) which varied from less than 10% to 70% compared with CFT, IFA and immunoblot. The criteria for choice of test should include a high specificity (>95%), at least moderate sensitivity (>70%) and likelihood ratio >10%.

Top of page

3.3.5 Suitable test acceptance criteria

Consistent with NPAAC Guidelines:

Requirements for the Validation of In House In vitro Diagnostic Devices

Commercial kits: according to manufacturer’s instructions.

3.3.6 Suitable internal controls

There are no validated (e.g. WHO) reference sera. However standardized qualitative PT IgG assays have been described. Specimens of serum with levels = 100 IU/mL could be used as positive controls.

3.3.7 Suitable original test validation criteria

As a minimum each assay should be validated in an adequate number of sera from controls with no history of cough or contact and cases which fulfil the clinical cases definition for pertussis within the previous 4 weeks. Ideally the cases should have been diagnosed by culture, PCR or paired sera serology, but large enough series of such cases are difficult to find. An indirect method for establishing the performance of a test is described in Poynten et al. 2002.¹²

3.3.8 Suitable external QAP program

RCPA QAP P/L Serology for Pertussis.

3.3.9 Special considerations

Serology should only be requested if the patient has a history of pertussis-like illness within the last 4 weeks as IgG (and less commonly IgA) have been shown to persist up to 2 years post onset of disease.

The lack of suitable controls is a problem for pertussis serology. Laboratories should encourage submission of paired sera and review serology results with feedback from clinicians and public health practitioners.

Research is being undertaken to identify non-vaccine antigens which might be used to develop diagnostic assays which are more predictive of recent disease.¹⁸

4 TYPING AND SUBTYPING METHODS*

4.1 Typing (subtyping) Method

PFGE has been the standard method for typing B. pertussis worldwide. Recently PCR based methods including virulence gene analysis⁹ and MLVA¹⁶ have been used to characterise the virulence makeup of circulating strains. Polymorphisms have been observed in the genes encoding for the S 1 subunit of pertussis toxin, for pertactin and also for the phenotypic expression of fimbriae.⁹ Further information and methods available at www.eupertstrain.org and http://www.cdc.gov/vaccines/pubs/pertussis-guide/downloads/_DRAFT_chapter2_amended.pdf

4.1.1 Utility

Important to track changes in antigenic structure of virulence factors such as pertussis toxin and pertactin used in subunit vaccines after the historic Australian pertussis WCV containing ptx S1A derived proteins was stopped.¹³

4.1.2 Suitable external QAP program

No programme available but isolates may be submitted to Eupertstrain for analysis.

4.1.3 Special considerations

Standards for typing may be found in the reference below.¹⁰

4.1.4 International reference focus

CDC and Eupertstrain Websites.

Top of page

5 REFERENCES

1. Bradstreet CMP, Tannahill AJ, Edwards JMB. 1972. Detection of Bordetella pertussis antibodies in human sera by complement-fixation and immunofluorescence. J Hyg Camb 70:75-83

2. http://www.cdc.gov/vaccines/pubs/pertussis-guide/downloads/_DRAFT_chapter2_amended.pdf . Guidelines for control of pertussis outbreaks. 2000, updated 2005.

3. Cloud JL et al. 2002. Impact of nasopharyngeal swab types on detection of Bordetella pertussis by PCR and culture. J Clin Microbiol 40: 3838-3840.

4. Dragsted DM, Dohn B, Madsen J, Jensen JS. 2004. Comparison of culture and PCR for detection of Bordetella pertussis and Bordetella parapertussis under routine laboratory conditions

5. www.eupertstrain.org European Research Programme for improved Pertussis Strain characterisation and surveillance

6. Farrell DJ, Daggard G, Mukkur TKS. 1999. Nested duplex PCR to detect Bordetella pertussis and Bordetella parapertussis and its application in diagnosis of pertussis in nonmetropolitan Southeast Queensland, Australia. J Clin Microbiol 37:606-610.

7. Hodder SL, Cherry JD, Mortimer EA, Ford AB, Gornbein J, Papp K. 2000. Antibody responses to Bordetella pertussis antigens and clnical correlations in elderly community residents. Clin Infect Dis. 31: 7-14.

8. Horby P, Macintyre CR, Macintyre PB, Gilbert GL, Staff M, Hanlon M, Heron LG, Cagney M, Bennett C. (2005) A boarding school outbreak of pertussis in adolescents: value of laboratory diagnostic methods. Epidemiol. Infect 133:229-236.

9. Mäkinen J, Mertsola J, Viljanen MK, Arvilommi H, He Q. 2002. Rapid typing of Bordetella pertussis pertussis toxin gene variants by LightCycler Real-time PCR and fluorescence resonance energy transfer hybridisation probe melting curve analysis. J Clin Microbiol 40 :2213-2216.

10. Mooi FR, Hallander H, Wirsing von Konig CH, Hoet B, Guiso N. 2000. Epidemiological typing of Bordetella pertussis isolates : recommendations for a standard methodology. Eur J Clin Microbiol Infect Dis 19 :174-181.

11. Muller FC, Hoppe JE, von Konig CHW. 1997. Laboratory diagnosis of pertussis: state of the art in 1997. J Clin Microbiol 35: 2435-2443.

12. Poynten IM, Hanlon M, Irwig L, Gilbert GL. 2002. Serological diagnosis of pertussis: evaluation of IgA against whole cell and specific Bordetella pertussis antigens as markers of recent infection. Epidemiol Infect 128:161-167.

13. Poynten M, McIntyre PB, Mooi FR, Heuvelman KJ, Gilbert GL. 2004. Temporal trends in circulating Bordetella pertussis strains in Australia. Epidemiol Infect 132:185-193.

14. Quinn HE, McIntyre PB. 2007. Pertussis epidemiology in Australia over the decade 1995-2005 – trends by region and age group. CDI 31 (in press).

15. Riffelmann m, von Konig CHW, Caro V, Guiso N for the Pertussis PCR Consensus Group. 2005. Nucleic acid amplification tests for diagnosis of Bordetella infections. J Clin Microbiol 43:4925-4929.

16. Schouls LM, van der Heide HGJ, Vauterin L, Vauterin P, and Mooi FR. 2004. Multiple-Locus Variable-Number Tandem Repeat Analysis of Dutch Bordetella pertussis Strains Reveals Rapid Genetic Changes with Clonal Expansion during the Late 1990s. J Bacteriol. 186:5496-5505.

17. Tozzi AE, Celentano LP, degli Atti MLC, Salmaso S. 2005. Diagnosis and management of pertussis. CMAJ 172: 509-515.

18. Watanabe M, Connelly B, Weiss AA. 2006. Characterization of serological response to pertussis. Clin Vaccine Immunol 13:341-348.