Pertussis and Bordetella Parapertussis in an Immunized Population

Brief Report Whooping Cough Caused by Bordetella pertussis and Bordetella parapertussis in an Immunized Population

Qiushui He, MD; Matti K. Viljanen, MD; Heikki Arvilommi, MD; Birgitta Aittanen; Jussi Mertsola, MD

Context.—The prevalence of Bordetella pertussis and Bordetella parapertussis National Public Health Institute, Hel- infections among outpatients in an immunized population is not known. sinki, Finland, and contains 5 ϫ 109 for- Objective.—To study the prevalence of these infections in outpatients with par- malin-killed B pertussis organisms per oxysmal cough in Finland, where the pertussis vaccine coverage of 4 doses is 98%. dose combined with diphtheria and teta- nus toxoids. The vaccine is used at 3, 4, 5, Design.—Prospective cohort study. 12 Setting.—Thirty-two health centers in southwestern Finland. and 24 months, and the coverage is 98%. In the last 10 years, based on the re- Patients.—A total of 584 patients with paroxysmal cough seen at local health port of the official reporting system, the centers from October 1994 through March 1997 from whom nasopharyngeal swabs annual number of laboratory-confirmed were collected. B pertussis infection cases have ranged Main Outcome Measures.—Prevalence of positive cultures for B pertussis or from 498 to 2574. Most patients are B parapertussis and/or positive polymerase chain reaction (PCR) results and fre- schoolchildren. In the last decade, 648 quency of symptoms in those with pertussis and parapertussis. patients with B pertussis infection and Results.—A total of 153 subjects (26.2%) had Bordetella infection by culture or 13 patients with B parapertussis infec- PCR: 93 (60.8%) had B pertussis infection, 49 (32.0%) had B parapertussis infection were hospitalized and 76% of them tion, and 11 (7.2%) had both. Of these cases, 39 (25.5%) had positive cultures and were younger than 1 year. 95 (62.1%) had positive PCR results for B pertussis, and 19 (12.4%) had positive Tocollectmorereliabledata,westarted cultures and 55 (35.9%) had positive PCR results for B parapertussis. At the time an enhanced surveillance in southwestern Finland (population, 702 000) in October of diagnosis, no difference was found in the frequency of symptoms between pa- 1994 that lasted 30 months and ended in tients with B parapertussis infection and those with B pertussis infection. Bordetella March 1997. Swabs (calcium alginate), cul- parapertussis infection was as common as B pertussis infection in children before ture transport tubes, and questionnaires school entry, whereas in schoolchildren and adults, B pertussis infection was more wereprovidedforall32healthcentersfree common than B parapertussis infection (PϽ.001). of charge. Culture results were reported Conclusion.—Bordetella infections are common in an immunized population, to the health centers 7 days after the and B parapertussis infections apparently are more prevalent than previously sample arrived at our laboratory. documented. In Finland, the health centers are ma- JAMA. 1998;280:635-637 jor sources for primary health care of children and adults. There is at least 1 health center in each community. The BORDETELLA pertussis and Borde- during recent acellular vaccine efficacy services offered by these centers are tella parapertussis cause whooping trials.5,6 Although cases of infection with usually free. cough in humans. They are almost iden- B parapertussis, even outbreaks, have tical at the DNA level and produce many been reported in immunized popula- similar virulence factors. The pathoge- tions,4,7-11 community-based data on the Inclusion Criteria netically important difference between occurrence of B parapertussis infection and Clinical Information the two is that B parapertussis does not are limited. This study investigated the In the health centers, nasopharyngeal secrete pertussis toxin.1 epidemiology of both infections in a specimens were taken by physicians In the prevaccination era, the relative highly immunized population. fromallpatientswithparoxysmalcough, frequency rates of B parapertussis iso- characterized as bouts of uncontrollable lates varied from 1% to 35% of Borde- METHODS coughing of any duration. During the 30 tella isolates,2-4 and similar frequency months of the study, swabs were ob- rates, from 2% to 25%, were also found Immunization and Surveillance tained from 584 eligible patients (num- of Pertussis in Finland ber of male patients, 257; age range, 7 Bordetella pertussis vaccination was days to 74 years; median age, 9 years). introducedinFinland(population,5.1mil- No hospitalization was needed. At the From the National Public Health Institute, Department in Turku (Drs He, Viljanen, Arvilommi, and Mertsola and lion) in 1952. Since 1958 the vaccine has time of sampling, 135 (23.1%) were ex- Mrs Aittanen), and the Department of Pediatrics, Uni- contained strain 18530. Because serotype periencing whooping and 208 (35.6%) versity of Turku (Drs He and Mertsola), Turku, Finland. 1.2 strains emerged in the 1970s, strain were experiencing vomiting. Twenty- Reprints: Qiushui He, MD, National Public Health In- stitute, Department in Turku, Kiinamyllynkatu 13, Fin- 1772wasintroducedtothevaccine(v/v)in nine health centers (90.6%) sent samples 20520 Turku, Finland (e-mail: [email protected]). 1976. The vaccine is manufactured by the to the laboratory. The average number

Downloaded From: https://jamanetwork.com/ on 09/26/2021 18 Bordetella parapertussis Bordetella parapertussis Bordetella pertussis Bordetella pertussis 16 No. of Culture-Confirmed Cases Culture Positivity Rates

14 25 P = .005

12 20 P < .001 15 P = .53 P > .99 10 10

8 5 Positivity Rate, % Rate, Positivity

6 Laboratory-Confirmed 0 <2 2-6 7-15 ≥16 Age Group, y 4 No. of Laboratory-Confirmed Cases No. 2 Figure 2.—Positivity rates of laboratory-conﬁrmed Bordetella parapertussis and Bordetella pertussis infections in various age groups (number of sub- 0 Oct Dec Feb Apr June Aug Oct Dec Feb Apr June Aug Oct Dec Feb jects Ͻ2 years, 53; 2-6 years, 163; 7-15 years, 220; and Ն16 years, 147). 1994 1995 1996 1997 Month and Year 1.42;P=.99)andforBparapertussis(RR, 1.39; 95% CI, 0.82-2.34; P = .26) between Figure 1.—Monthly numbers of laboratory-conﬁrmed cases of Bordetella parapertussis and Bordetella pertussis between October 1994 and March 1997 in southwestern Finland. Bordetella parapertussis was detected specimenstakenfrompatientsduringthe from samples sent by 17 health centers (53%), and B pertussis was detected from samples sent by 23 health first 7 days after the onset of paroxysmal centers (72%). cough and specimens taken during the later course of illness. of samples obtained each month was 19 RESULTS Bordetella parapertussis and B pertus- (range, 1-104). sis were detected during 13 and 22 months Detailed clinical information on each Of 584 samples tested by culture, 19 of the surveillance period and from samples subject was obtained by a structured (3.4%) were positive for B parapertussis sent by 17 and 23 health centers, respec- questionnaire asking about the date of and 39 (6.7%) for B pertussis.Of564 tively. Both B parapertussis and B per- onset and the nature of symptoms. The samplestestedbybothcultureandPCR, tussis were detected during 11 months of questionnaires were completed by phy- 15 had positive cultures and 55 (9.4%) the surveillance period and from samples sicians at the health centers and mailed had positive PCRresults for B paraper- sent by 11 health centers (Figure 1). to the laboratory. The symptoms of sub- tussis, and 36 had positive cultures and At the time of sampling, no difference jects were not followed up after this first 95 (16.3%) had positive PCR results was found in the frequency of whooping contact with the health center. for B pertussis. Forty-four (86.3%) of 51 (11 [25%] of 44 vs 23 [26%] of 88; RR, 0.96; samples with positive cultures for Bor- 95% CI, 0.52-1.79; P = .94), vomiting (19 Laboratory Methods detella also had positive PCR results. In [43%] of 44 vs 26 [30%] of 88; RR, 1.46; The swabs were inoculated onto the 11 specimens, B parapertussis and B 95% CI, 0.91-2.33; P = .17), or mean du- slant of the transport medium (of the pertussis DNA were simultaneously de- ration (in days) of paroxysmal cough at same composition as the charcoal agar tected. Men and women had similar cul- the time of sampling (15.2 [95% CI, 11.8- plate) supplemented with cephalexin, ture and PCR positivity rates. Alto- 18.6] vs 15.3 [95% CI, 12.7-17.6]; P = .95) placed in sterile empty tubes with caps, gether, 153 subjects (26.2%) were con- between patients with positive cultures transported to the laboratory, and stored firmed as harboring B pertussis (60.8%), or PCR results for B parapertussis and B at −40°C for polymerase chain reaction B parapertussis (32.0%), or both (7.2%). pertussis. No difference was found in the (PCR). Twenty swabs were not sent to Of 564 subjects whose samples were frequency of whooping (5 [19%] of 26 vs 6 the laboratory after inoculation. Thus, tested by both culture and PCR, 198 [27%] of 22; RR, 0.71; 95% CI, 0.25-2.0; 564 swabs were tested by PCR. (35.1%) had had paroxysmal cough for 7 P = .73) and vomiting (13 [50%] of 26 vs 12 Details of bacterial culture and identi- daysorless,and366(64.9%)hadhadpar- [55%] of 22; RR, 0.92; 95% CI, 0.54-1.58; fication, DNA extraction, and PCR have oxysmal cough for more than 7 days at P = .98) between patients younger than been described earlier.11-14 Two sets of the time of sampling. 7 years with positive cultures or PCR re- primers derived from insertion sequence Of the 198 patients who had experi- sultsforBpertussisandBparapertussis, elements IS481 and IS1001 that are spe- enced paroxysmal cough for 7 days or or in the frequency of whooping (6 [27%] cific for B pertussis and B parapertussis, less, 24 (12.1%) had positive cultures or of 22 vs 5 [23%] of 22; RR, 1.20; 95% respectively, were used for PCR as- PCR results for B parapertussis and 35 CI, 0.43-3.36; PϾ.99) and vomiting (12 says.12,13 The2PCRassayswererunsepa- (17.7%) had positive cultures or PCR re- [55%] of 22 vs 7 [32%] of 22; RR, 1.70; 95% rately. The parapertussis PCR products sults for B pertussis. Of the 366 patients CI, 0.83-3.50; P = .22) between patients were further confirmed by amplification who had experienced paroxysmal cough younger than 7 years and those aged with a set of interior primers. for more than 7 days, 32 (8.7%) had posi- 7 years or older with positive cultures tive cultures or PCR results for B para- or PCR results for B parapertussis. Statistical Analyses pertussis and 66 (18.0%) had positive cul- The positivity rate of laboratory- All statistical analyses were based on tures or PCR results for B pertussis.No confirmed B parapertussis infection was the 2-tailed ␹2 test, the Fisher exact test, statistically significant difference was highest in 2- to 6-year-olds, whereas the or the Student t test. A P value of less foundinpositivityratesofcultureorPCR positivity rate of B pertussis infection was than .05 was considered statistically resultsforBpertussis(relativerisk[RR], highest in 7- to 15-year-olds (Figure 2). significant. 0.98; 95% confidence interval [CI], 0.68- Bordetella pertussis infection was signifi-

Downloaded From: https://jamanetwork.com/ on 09/26/2021 cantly more common than B parapertussis infection can also be asymptom- also thank Olli Ruuskanen, MD, and Olli Lassila, sis infection in schoolchildren and adults atic.11,15 Immunization can modify clini- MD, for constructive criticism of the manuscript. (PϽ.001). The annual incidences of B per- cal symptoms.12,16 Individuals with mild tussis and B parapertussis infections were symptoms due to these infections are References 5.9 and 3.4 cases per 100 000, respec- not likely to be suspected of having 1. Arico B, Rappuoli R. Bordetella parapertussis and Bordetella bronchiseptica contain transcrip- tively. The annual incidences of B pertus- whooping cough and samples are not tionally silent pertussis toxin genes. J Bacteriol. sis and B parapertussis infections were likely to be taken, suggesting that the 1987;169:2847-2853. 16.6 and 9.5 in children younger than 2 true prevalence of both infections may 2. Miller JJ, Saito TM, Silverberg RJ. Parapertus- years, 19.6 and 20.5 in children aged 2 to remainunderestimated.Serologicalsur- sis: clinical and serologic observations. J Pediatr. 6 years, 27.2 and 15.1 in children aged 7 veys would most likely increase the in- 1941;19:229-240. 14,16 3. Eldering G, Kendrick PL. Incidence of paraper- to 15 years, and 1.5 and 0.2 in persons aged cidence figures. tussis in the Grand Rapids area as indicated by 16 16 years or older, respectively. Bordetella pertussis and B paraper- years’ experience with diagnostic cultures. Am J Based on the culture-proven cases re- tussis can cause similar symptoms.6,8 In Public Health. 1952;42:27-31. 4. Lautrop H. Epidemics of parapertussis: 20 ported by the official system, the annual this outpatient population no difference years’observationinDenmark.Lancet.1971;1:1195- incidences of B pertussis infection in wasfoundinthefrequencyofsymptoms, 1198. 1995 and 1996 were 1.3 and 0.6 per suchaswhoopingandvomiting,between 5. Mastrantonio P, Giuliano M, Stefanelli P, et al. 100 000 in southwestern Finland and 1.2 patients with B pertussis and B para- Bordetellaparapertussisinfections.DevBiolStand. and 1.9 per 100 000 in the whole country, pertussis infections. However, we could 1997;89:255-259. 6. HeiningerU,StehrK,Schmitt-GrobeS,etal.Clini- respectively. In southwestern Finland, notexcludethepossibilitythatthestudy cal characteristics of illness caused by Bordetella per- only 1 case of B parapertussis infection was not large enough to detect differ- tussis compared with illness caused by Bordetella wasreportedduringthesurveillancepe- ences in symptom rates between those parapertussis. Pediatr Infect Dis J. 1994;13:306-309. riod, and 3 cases were reported during 7. Linnemann CC, Perry EB. Bordetella paraper- with B pertussis and B parapertussis in- tussis: recent experience and a review of the litera- the last 6 years. fections. Furthermore, the inclusion cri- ture. AJDC. 1977;131:560-563. teria may have biased the sample to in- 8. Wirsing von Ko¨nig CH, Finger H. Role of per- COMMENT clude only patients with severe symp- tussistoxinincausingsymptomsofBordetellapara- This study confirms that in an immu- toms. Duration of illness may differ, but pertussis infection. Eur J Clin Microbiol Infect Dis. 1994;13:455-458. nized population B pertussis and B para- we did not have that information. 9. Mertsola J. Mixed outbreak of Bordetella pertus- pertussis infections remain common. The protective role of B pertussis vac- sis and Bordetella parapertussis infections in Fin- Furthermore,Bparapertussisinfections cination against B parapertussis has land. Eur J Clin Microbiol. 1985;4:123-128. were more prevalent than usually docu- been debated.4,10,17 10. Vysoka-Burianova B. Contemporary problems Even experimental in the epidemiology of whooping cough. J Hyg Epi- mented, with one third of laboratory-con- animal studies have provided contradic- demiol Microbiol Immunol. 1963;7:472-481. firmed Bordetella infections caused by tory results.18,19 After B pertussis vacci- 11. He Q, Edelman K, Arvilommi H, Mertsola J. B parapertussis. Bordetella paraper- nation was introduced, incidences of Protective role of immunoglobulin G antibodies to fila- tussis infection was as common as B per- both infections markedly decreased in mentous hemagglutinin and pertactin of Bordetella 4 pertussis and Bordetella parapertussis infection. tussis infection in children before school Denmark, whereas in the former Eur J Clin Microbiol Infect Dis. 1996;15:793-798. entry, whereas in schoolchildren and Czechoslovakia peak B parapertussis 12. He Q, Schmidt-Schla¨pfer G, Just M, et al. Im- adults B pertussis infection was mark- morbidityremainedthesameinchildren pact of polymerase chain reaction on clinical pertus- edly more prevalent. High rates of B per- younger than 7 years.10,16 Our results fall sis research: Finnish and Swiss experiences. J In- fect Dis. 1996;174:1288-1295. tussis infection in schoolchildren and between the results of these 2 studies, 13. He Q, Mertsola J, Soini H, Skurnik M, Ruus- adultsconfirmthatprotectionfromBper- suggesting that B pertussis vaccine may kanen O, Viljanen MK. Comparison of polymerase tussis vaccinations decreases with time. provide some protection against B para- chain reaction with culture and enzyme immunoas- The sensitivity of B parapertussis pertussis. This partial protection would sayfordiagnosisofpertussis.JClinMicrobiol.1993; 31:642-645. PCR was about 3 times higher than that still allow natural infection to occur 14. vanderZeeA,AgterbergC,PeetersM,Schelle- of culture. The assay has been shown to among younger children. On the other kens J, Mooi FR. Polymerase chain reaction assay be specific by testing panels of bacterial hand, the B parapertussis immunity im- for pertussis: simultaneous detection and discrimi- species.14 Its specificity was also con- parted by both B pertussis vaccinations nation of Bordetella pertussis and Bordetella parapertussis. J Clin Microbiol. 1993;31:2134-2140. firmed clinically in this study by the low and infections might be more effective 15. Borska K, Simkovicova M. Studies on the circu- positivity rates obtained between infec- and longer lasting than B pertussis im- lation of Bordetella pertussis and Bordetella para- tion peaks and in subjects 16 years and munity based on vaccine alone, resulting pertussis in populations of children. J Hyg Epide- older. The incidences of B pertussis and in a low incidence of B parapertussis miol Microbiol Immunol. 1972;16:159-172. 16. Mink CAM, Cherry JD, Christenson P, et al. A B parapertussis infections reported by infection in schoolchildren and adults. search for Bordetella pertussis infection in univer- the official reporting system, based on In the development of less reactogenic sity students. Clin Infect Dis. 1992;14:464-471. patients with laboratory-confirmed in- acellular B pertussis vaccines, it may 17. Stehr K, Cherry JD, Heininger U, et al. A com- fection, were lower than those obtained be important to estimate the protection parative efficacy trial in Germany in infants who received either the Lederle/Takeda acellular per- from the enhanced surveillance. The en- against B parapertussis infections pro- tussis component DTP (DTaP) vaccine, the Lederle hanced surveillance clearly supple- vided by these acellular vaccines. whole-cell component DTP vaccine, or DT vaccine. mented the existing reporting system. Pediatrics. 1998;101:1-11. Underdiagnosis of Bordetella infections 18. Willems RJL, Kamerbeek J, Geuijen CAW, et ThisstudywasfinanciallysupportedbytheAcad- al. The efficacy of a whole cell pertussis vaccine is a more likely reason for the low inci- emy of Finland and the Finnish Foundation for Pe- and fimbriae against Bordetella pertussis and Bor- dences reported by the official system diatric Research, Helsinki, Finland. detella parapertussis infections in a respiratory than failure to report diagnosed cases. The authors thank Tuula Lehtonen for excellent mouse model. Vaccine. 1998;16:410-416. The disease caused by B parapertus- technical assistance, Erkki Nieminen, MSc, for help 19. Khalef N, Danve B, Quentin-Millet M-J, Guiso in preparing the figures, Simo Merne, MA, for lin- N. Bordetella pertussis and Bordetella parapertussis is usually milder than that caused by guistic revision of the manuscript, and the staffs of sis: two immunologically distinct species. Infect Im- B pertussis.6-9,15 Bordetella parapertus- the local health centers for their cooperation. We mun. 1993;61:486-490.

Downloaded From: https://jamanetwork.com/ on 09/26/2021