To

PRa

b

c

d

e

a

ARRAA

KPSVII

1

cpewOs

gtg4gv

S3

0d

Vaccine 28 (2010) 3285–3290

Contents lists available at ScienceDirect

Vaccine

journa l homepage: www.e lsev ier .com/ locate /vacc ine

he seroepidemiology of Bordetella pertussis in Israel—Estimatef incidence of infection

amela Rendi-Wagnera,∗, Joshua Tobiasb, Larisa Moermanc, Sophy Gorena,avit Bassald, Manfred Greend,e, Dani Cohena

Department of Epidemiology and Preventive Health, School of Public Health, Sackler Faculty of Medicine, Tel Aviv University, IsraelInstitute of Biomedicine, Department of Microbiology and Immunology, The Sahlgrenska Academy of Göteborg University, S-40530 Göteborg, SwedenDepartment of Epidemiology, Ministry of Health, Jerusalem, IsraelThe Israel Center for Disease Control, Ministry of Health, Tel Hashomer, IsraelSchool of Public Health, University of Haifa, Israel

r t i c l e i n f o

rticle history:eceived 1 December 2009eceived in revised form 16 February 2010ccepted 24 February 2010vailable online 10 March 2010

eywords:

a b s t r a c t

This study was undertaken to estimate the magnitude of Bordetella pertussis infections in a highly vac-cinated population in Israel in order to evaluate the relationship between clinical notification data andserology-based evidence of infection. A cross-sectional survey was conducted on a total of 1982 serumsamples from the National Serum Bank, collected from January 2000 through December 2001, in orderto monitor high levels of pertussis toxin (PT) IgG antibody indicative of recent B. pertussis infection, bystandardized methods. The estimation yielded an infection incidence rate of 2448 per 100,000 popula-

ertussis (whooping cough)eroprevalenceaccination

gG pertussis toxinncidence of infection

tion (≥3 years of age) for the year 2000 compared to an annual incidence of reported pertussis of 5.6 per100,000 for the same period. The peaks of estimated incidence of infection were found in the groups of15–19-year olds (5245 per 100,000) and older than 60 years (6469 per 100,000), whereas the majorityof clinical pertussis cases were reported for the 10–14-year olds (20.5 per 100,000). The findings clearlyshow that despite a high vaccination coverage rate (>93%), there is still a considerable circulation of B.pertussis, particularly in adolescents and elderly. Population-based serosurveillance for pertussis offers

erpre

the potential to assist int

. Introduction

Whooping cough is a common endemic and highly communi-able respiratory infection which is mainly caused by Bordetellaertussis, primarily affecting infants and young children; how-ver, increasingly occurring in older age groups [1–8]. Despiteidespread use of vaccination, the disease has not been eliminated.n the contrary, increased incidence rates have been reported in

everal countries during the last decade [2–8].In Israel, since 1957, vaccination against pertussis was

iven to children using a whole-cell component in diphtheria–etanus–pertussis vaccine until it was replaced by the less reacto-

enic acellular vaccines in 2002. The vaccine is administered at 2,, 6, and 12 months, and since 2005, an additional booster has beeniven at 7–8 years of age. In 2008, a so-called “catch-up” boosteraccination program was introduced for children aged 13–14 years.

∗ Corresponding author at: Department of Epidemiology and Preventive Health,chool of Public Health, Sackler Faculty of Medicine, Tel Aviv University, P.O.B.9040, 69978 Tel Aviv, Israel. Tel.: +972 3 6405651; fax: +972 3 6409868.

E-mail address: prendi@post.tau.ac.il (P. Rendi-Wagner).

264-410X/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.oi:10.1016/j.vaccine.2010.02.104

tation of trends independent of notification and diagnostic bias.© 2010 Elsevier Ltd. All rights reserved.

This will continue until the children who had received a school-agebooster (at 7–8 years) reach the age of 13.

An impressive drop in pertussis rates was observed due to thewidespread use of vaccination until the 1990s. However, this wasfollowed by a subsequent increase in pertussis morbidity since1999, despite a coverage of 93% for four vaccine doses among chil-dren [6]. As in other countries, there has been observed a shift inmorbidity towards higher age groups [6].

As a result of waning immunity after vaccination, pertussis mor-bidity increases in previously vaccinated children, adolescents, andadults, thus, maintaining the pathogen circulating in the popula-tion. Lack of typical pertussis symptoms, may be more common foradolescents and adults than for young children, contributing to aconsiderable degree of under-reporting in older age groups. There-fore, the informative value of routine surveillance data based oncase notification is limited, yet, not detecting atypical and mild dis-ease. This can serve as an important “silent” source of transmission

in the population. To date, the extent of infection in these older agegroups remains to a large extent unknown, and calls for alternativestandardized tools for pertussis monitoring.

High titers of antibody to pertussis toxin (PT) have been provento be a reliable indicator of recent pertussis infection, thus, serving

3 Vaccin

atttoo

2

2

sdpfmtsuIetdsi

2

thH

2

dciC

2

aRBeEt

EhLotCwcs

q(i

286 P. Rendi-Wagner et al. /

s a solid and standardized marker for the incidence of infection inhe general population [9]. The aims of this study were to documenthe age-specific sero-profile of high antibody titers to pertussisoxin as a marker for incidence of infection in order to assess trendsf pertussis and implications for prevention strategies independentf notification and diagnostic bias.

. Material and methods

.1. Study population and data collection

A cross-sectional sero-survey was conducted using archivederum samples collected by the Israel Centre for Disease Controluring 2000 to 2001 (pre-booster period). The serum bank com-rised samples from all regions of Israel including both males andemales of all ages. Collection of sera is continuous, covering all

onths of the year. The samples of the younger age groups (oneo 17 years) were residual sera from diagnostic laboratories, andamples from the adult population (≥18 years of age) were resid-als of sera obtained from healthy blood donors living all over

srael, screened before the use of the blood donations. Both sourcesxcluded repeat samples from the same individuals as well as seraaken from subjects with confirmed or suspected immunologicalisorders. Each sample had a unique identifier, plus details of age,ex, religion, place of residence (at the level of town), and the yearn which the sample was drawn.

.2. Notified pertussis data

Pertussis has been reported in Israel since the early 1950s. Prac-itioners are requested to notify each clinical case to the local publicealth office which reports on a weekly basis to the Ministry ofealth. Case classification does not imply laboratory confirmation.

.3. Pertussis vaccination coverage

National immunization coverage is calculated each year by theistrict health offices, and submitted to the Ministry of Health. Thealculation is based on a representative sample of children bornn each health district and registered in the public Family Healthentres.

.4. Antibody assay

Serum samples were stored at −20 ◦C until they were testedt the Department of Epidemiology and Preventive Medicineesearch Laboratory, Tel Aviv University. IgG antibodies to. pertussis toxin (PT) were determined by a commercialnzyme-linked immunosorbent assay (ELISA) (Pertusscan PT-GTM,uro-Diagnostica AB, Sweden) in accordance with the manufac-urer’s instructions.

This assay was validated within the European Sero-pidemiology Network 2 (ESEN2) project by testing a panel of 150uman control sera provided by the European Pertussis Referenceaboratory (Department of Hygiene and Microbiology, Universityf Palermo, Italy) [10]. The panel’s results were calibrated againsthose from the Reference Centre at the Health Protection Agencyentre for Infections, London. Linear and quadratic regressionas fitted and R2 (multiple correlation coefficient) values were

alculated. In the standardization process regression lines were

elected and standardization equations obtained [10].

These standardization equations were used to convert the localuantitative results into standardized reference laboratory unitageESEN units). Test results are expressed in “ESEN units” per millil-tre.

e 28 (2010) 3285–3290

2.5. Anti-PT antibody titer cut-offs

The quantitative titers of anti-PT IgG were classified as hightiter samples using a cut-off level of 125 ESEN units/ml (equivalentto 225 local units/ml) indicative of recent or active infection withB. pertussis [9]. The sensitivity of this threshold was estimated at76% and the positive predictive value (PPV) at 80%, assuming a trueprevalence of disease of 10% [9].

A second cut-off of 62.5 ESEN units/ml (equivalent to134 local units/ml) was employed, suggesting B. pertussis infectionin the previous 12 months with high probability [9,11].

To indicate seropositivity, the cut-off titer of 10 ESEN units/ml(equivalent to 13 local units/ml) was applied.

2.6. Statistical analysis

Overall and age-specific prevalence rates of IgG anti-PT antibod-ies and 95% confidence intervals were calculated using the variouscut-off values defining seropositivity and recent pertussis infection.Statistical significance of differences in prevalence rates betweensubgroups of the study population was examined using the chisquare test.

To estimate age-specific incidences of infection, as previouslydescribed by de Melker et al. [12], a statistical relationship betweentime since infection and IgG anti-PT levels as described by Teuniset al. [13] was combined with age-specific distribution of IgG-PTderived from a cross-sectional survey of the general population. Thefollowing threshold titers were chosen to calculate the incidence ofinfection in the population: 62.5 and 125 ESEN units/ml (equivalentto 134 and 225 local units/ml, respectively).

Calculation of incidence of infection was limited to the age group≥3 years of age in order to avoid interference with vaccinationinduced or maternally derived antibodies.

3. Results

3.1. Prevalence of anti-PT by age

During the 2-year observation period (January 2000 throughDecember 2001), a total of 1982 (year 2000: 1066; year 2001:916) sera samples were tested for presence of IgG antibodies toPT. The mean age of the subjects enrolled was 19.4 ± 15.8 years(range 0.6–79.0 years); the median age was 15.5 years. Of these,1070 (54.0%) sera were obtained from males and 912 (46.0%) fromfemales.

Of all samples tested, 49.3% (977/1982) (95% CI 47.1–51.5%)exhibited antibodies to PT (≥10 ESEN units/ml), 2.3% (45/1982)(95% CI 1.7–3.0%) revealed titers ≥62.5 ESEN units/ml, and anti-PT IgG titers ≥125 ESEN units/ml were identified in 0.9% (17/1982)(95% CI 0.5–1.4%) of all samples. Fig. 1 shows the distribution of anti-PT IgG titer values by age, together with the reported age-specificDTP3 vaccination coverage rate.

Apart from the first 2 years of life (75.6%), a second peak forseropositivity (≥10 ESEN units/ml) was noticed in the age groupolder than 61 years (72.2%). Likewise, the highest proportion ofhigh anti-PT titers were observed below 24 months of age: 11.9%(20/1982) had anti-PT ≥62.5 ESEN units/ml, and 3.6% (6/1982) hadanti-PT ≥125 ESEN units/ml.

After excluding the data of the age group ≤3 years (to avoidinterference with maternal and vaccination derived antibodies),

the proportion of high titer sera (≥62.5 ESEN units/ml) was high-est in the age group ≥61 years (4.2%), followed by the 16–20-yearolds (2.7%).

There were no statistically significant differences detectedin the prevalence of high anti-PT titer sera (both ≥62.5 and

P. Rendi-Wagner et al. / Vaccine 28 (2010) 3285–3290 3287

F EN unh –2001

≥oehto(Piee

<

3

wt

Fc

ig. 1. Age-specific distribution (%) of three PT IgG titer cut-offs (grey: 10–62.5 ESistorical reported DTP3 vaccine coverage rate (black line, right scale), period 2000

125 ESEN units/ml) by gender or place of residence (urbanr rural) (Table 1). However, comparing by means of socio-conomic status, the low-income group showed a significantlyigher proportion of high anti-PT titers (≥125 ESEN units/ml)han the high-income category (1.1% vs. 0.3%, P = 0.054). More-ver, Arab subjects were more likely to have high anti-PT titers≥125 ESEN units/ml) than the Jewish subjects (1.8% vs. 0.4%,= 0.009) (Table 1). However, in the multivariable analysis, includ-

ng socio-economic status and ethnicity, none of the two variablesmerged as significantly associated with high titer PT antibody lev-ls.

The proportion of non-immune subjects, exhibiting titers10 ESEN units/ml, was highest in those aged 6–10 years (66.0%).

.2. Incidence of infection

The results for the cut-off levels of 62.5 and 125 ESEN units/mlere chosen to indicate recent B. pertussis infection. After infec-

ion, anti-PT titers take on average 58.6 days to drop to a

ig. 2. Age-specific distribution of estimated incidence of B. pertussis infection per 100,0ases of pertussis per 100,000 population per year (open circles, right scale), year 2000.

its/ml, white: 62.5–125 ESEN units/ml, black: ≥125 ESEN units/ml) (left scale) and.

level of 125 ESEN units/ml and 208.9 days to reach a value of62.5 ESEN units/ml [12]. A percentage of 2.3% (95% CI 1.7–3.0%) ofthe total population tested revealed an anti-PT level of at least62.5 ESEN units/ml. After excluding the age group <3 years, thisproportion constitutes 1.4% (95% CI 0.9–2.0%), equivalent to an esti-mated incidence of B. pertussis infection in the year before serumsampling of 2.4% (365.25 days/208.9 days × 1.4%). The cut-off titerof 125 ESEN units/ml yielded an estimated incidence rate of infec-tion of 3.7% (365.25 days/58.6 days × 0.6%) for the population ≥3years of age.

In Fig. 2, the age-specific incidence rates of infection with B. per-tussis in the population are given as calculated for the cut-off level of62.5 ESEN units/ml. In order to compare estimated versus reportedincidence rates, the incidences of officially reported clinical cases

of the year 2000 were compared to incidence of infection estimatesbased on sera samples obtained the following year (year 2001). Theestimation, based on titers gained in 2001, resulted in an incidencerate of 2448 per 100,000 population (≥3 years of age) for the year2000, the year prior to serum sampling. During the same year the

00 population per year (black squares, left scale) and annual incidence of notified

3288 P. Rendi-Wagner et al. / Vaccine 28 (2010) 3285–3290

Table 1Prevalence of PT high IgG titers (>62.5 ESEN units/ml and >125 ESEN units/ml) in sera collected between 2000 and 2001.

n Univariable analysis

Sera (%) ESEN units/ml (95% CI)

>62.5 P >125 P

Total 1982 2.3 (1.7–3.1) 0.9 (0.5–1.4)Age (year)

0 to <3 242 8.7 (5.5–13)0.000* 2.9 (1.2–5.9)

0.000*3+ 1740 1.4 (0.9–2) 0.6 (0.3–1.1)

GenderAll

Female 912 2.4 (1.5–3.6)0.696

1(0.5–1.9)0.565Male 1070 2.1 (1.3–3.1) 0.7 (0.3–1.3)

>3 yearsFemale 792 1 (0.4–2)

0.2270.5 (0.1–1.3)

0.725Male 948 1.7 (1–2.7) 0.6 (0.2–1.4)

EnvironmentAll

Urban 1703 2.6 (1.9–3.5)0.036* 1(0.6–1.6)

naRural 241 0.4 (0–2.3) 0 0

>3 yearsUrban 1489 1.5 (0.9–2.2)

0.1820.7 (0.3–1.2)

naRural 230 0.4 (0–2.4) 0

RegionAll

North 651 1.5 (0.7–2.8)0.055*

0.6 (0.2–1.6)0.172Center 966 3.1 (2.1–4.4) 1.2 (0.6–2.2)

South 357 1.4 (0.5–3.2) 1.3 (0.5–3.2)

>3 yearsNorth 616 1.1 (0.5–2.3)

0.5960.6 (0.2–1.7)

0.855Center 833 1.7 (0.9–2.8) 0.6 (0.2–1.4)South 284 1.1 (0.2–3.1) 0.4 (0–1.9)

Socio-economic statusAll

High 849 2.7 (1.7–4)0.666

1.1 (0.5–2)0.823Low 841 2.4 (1.5–3.6) 1(0.4–1.9)

>3 yearsHigh 747 1.5 (0.7–2.6)

0.7950.3 (0–1)

0.053*Low 732 1.6 (0.8–2.8) 1.1 (0.5–2.1)

EthnicityAll

Jewish 1687 2.1 (1.4–2.9)0.139

0.8 (0.4–1.3)0.291Non-Jewish 287 3.5 (1.7–6.3) 1.4 (0.4–3.5)

>3 yearsJewish 1516 1.3 (0.8–2)

0.2150.4 (0.1–0.9)

0.008*Non-Jewish 217 2.3 (0.8–5.3) 1.8 (0.5–4.7)

n

a≥intiecfiTyra

s

a: not applicable.* Statistical significant difference.

verage officially reported pertussis incidence for the population3 years of age was 5.6/100,000 [14]. Accordingly, the estimated

ncidence of infection is 400-times higher than the incidence ofotified clinical pertussis cases. As seen in Fig. 2, this also holdsrue for age stratified analysis. The age distribution of estimatednfection rates versus notified cases reveals a similar trend, how-ver, the peak of estimated incidence of infection is found in the ageategory 15–19 years (5245/100,000), whereas the majority of noti-ed cases are given in the group of 10–14-year olds (20.5/100,000).he incidence of reported pertussis is lowest for the population 60

ears or older (0.7/100,000). In contrast, the estimated infectionate shows a second peak in the population older than 60 years ofge (6469/100,000) (Fig. 2).

The comparison of notified disease data and estimated age-pecific rates of infection reveals the highest discrepancy in the

adult age group old (>19 years of age) where the estimated rate ofinfection is more than 1000-times higher than the reported inci-dence figure. In contrast, the estimated incidence rate of the agegroup 3–19 years produces a 250-times higher rate than the inci-dence rate based on case notifications.

4. Discussion

In this study, we estimated the age-specific incidence of B. per-

tussis infection, based on a cross-sectional sero-epidemiologicalsurvey of the distribution of high anti-PT titer sera, established bystandardized criteria [14]. Information about the sero-prevalenceof high levels of anti-PT antibodies in combination with the post-infection antibody decline rate allows the quantification of the

Vaccin

eotci

vlwa

hfgaltb

cta3srsionc[tofdiBmner

irh1mahFstWatt

atdapidws

P. Rendi-Wagner et al. /

xtent of B. pertussis infections in various age groups irrespectivef clinical manifestation and reporting compliance. The thresholditers employed in this study were of an equivalent level to thoseut-offs reported by de Melker et al. as diagnostic of recent or activenfection with B. pertussis [9].

High levels of anti-PT IgG antibody may also be due to previousaccination. However, numerous results from clinical trials of acel-ular and whole-cell vaccines have shown that high antibody titers

ane 12–18 months following the primary vaccination course inlmost all vaccinees [15].

During the study period, primary pertussis vaccination in Israelas been targeted routinely only at the infant age group with a

ourth shot administered at 12 months. No booster doses wereiven at the time of serum sampling to older age groups. Althoughnti-PT titers rapidly decrease to very low levels within 1 year fol-owing vaccination [16], the first 3 years of life were excluded fromhe analysis of incidence of infection in order to avoid an influencey previous exposure to vaccine.

Our results clearly show that despite a high vaccination infantoverage rate (>93%) in Israel, there is still a considerable circula-ion of B. pertussis, particularly in adolescents and elderly. In 2000,bout 2.4% (or 2448 per 100,000) of the Israeli population older thenyears of age had previously experienced infection with B. pertus-

is revealing a striking discrepancy between rates of infection andate of reported disease for several reasons. For example, pertus-is is under-diagnosed and under-reported, as similarly observedn other countries; [12,17] in The Netherlands, the estimated ratef infection is more than 600-times higher than the notified caseumbers [12]. Studies also suggest that only 40–50% of pertussisases show a classic clinical manifestation of a paroxysmal cough18], often leading to a misdiagnosis as a general respiratory infec-ion and a failure to investigate for pertussis. Hence, the amountf under-reporting varies by age, and has been shown to be higheror older children, adolescents, and adults than for younger chil-ren. It is also well documented that individuals with a primed

mmune system develop a mild variant of the disease [19,20].ased on our analysis, we are not able to determine the clinicalanifestation of infections. Yet, in terms of transmission to vul-

erable younger age cohorts, infections in adolescents and adults,ven though asymptomatic or mild, have been suggested to be ofelevance [21].

The age distribution of reported pertussis cases and estimatedncidence of infection reveal a similar, however, not identical age-elated trend, both showing peaks in adolescence. However, theighest incidence of notified cases is observed in children aged0–14 years followed by a steady decrease with age, while the esti-ated rate of infection peaks twice, among 15–19-year old subjects

s well as in the older age cohort (>60 years). Similar age-profilesave been observed in other developed countries such as Australia,inland, and France in the pre-booster era [14,22]. Yet, these age-pecific incidence patterns of B. pertussis infections clearly reflecthe dynamics of immunity and transmission in the populations.

hile high peaks of incidence rates among adolescents and youngdults might indicate high rates of transmission, low rates of infec-ion may be related to less contact and exposure as observed forhe group of 40–59-year olds.

Our findings are supported by a small pertussis outbreakmong Israeli soldiers reported during the study period, in win-er 2001, suggesting a high rate of exposure in young adultsuring their army service [23]. According to a previous survey,bout 13% of Israeli military recruits who were seronegative for

ertussis at time of enrolment, have shown seroconversion dur-

ng their 3-year military service [24]. In addition, the presentata revealed that the levels of serologically defined infectionere higher in the Israeli Arab population and groups of lower

ocio-economic status, which may be explained by higher person-

e 28 (2010) 3285–3290 3289

to-person transmission of B. pertussis due to more crowding in thesecohorts.

In younger age groups (<9 years), both, the reported as wellthe estimated incidence data reveal considerable pertussis activity,suggesting that susceptibility for symptomatic infection in someindividuals may re-emerge even short time after primary pertus-sis vaccination [25]. Indeed, the finding of widespread circulationof B. pertussis may have several reasons. One is low vaccinationcoverage as observed in countries such as Italy or Germany [15],moreover, primary vaccination failure due to inadequate vacci-nation schedules, types of vaccines, or waning immunity afterprimary vaccination. The latter may most likely explain the recentlyobserved resurgence in highly vaccinated populations like Israel.

However, the present study also provides evidence of waningprotection following natural infection, as there was a high rateof seropositivity and infections occurring in the population olderthan 60 years old age; a group which most likely have acquirednatural immunity during their lives. Limited existing data on thistopic suggest that pertussis vaccinated persons become susceptibleto pertussis disease 5–10 years following the primary vaccinationseries, while immunity after natural infection seems to be lost after10–20 years [26–28].

Indeed, the understanding of diminishing immunity, the roleof B. pertussis as an important causative agent of respiratory dis-ease in age groups beyond childhood, as well as the recognitionthat older age cohorts may serve as a reservoir for transmissionto infants, particularly those who are too young to be adequatelyprotected by immunization and who are at greatest risk for dis-ease complications, all point to the potential benefit of boosterdoses for adolescents and adults. In order to approach the problem,several countries, in accordance with the Global Pertussis Initia-tive [29,30], have introduced acellular booster doses for older agegroups. Likewise, in Israel, the age distribution of pertussis notifi-cations has recently led to the introduction of an additional boosterdose at school age. However, to date, it is not clear what the long-term impact of the introduction of additional booster doses on thetransmission of B. pertussis to younger at-risk age cohorts will be.Hence, given the limitations of other trend monitoring methods,the present findings and the developed serological tool may serveas a valuable and less biased means for continuous follow up assess-ments of the epidemiology of pertussis, particularly in view of therecently employed booster strategy.

Conflict of interest

None.

Acknowledgement

Thanks are due to Mr. Ruslan Gosinov for management of mor-bidity data.

References

[1] Cherry JD. The epidemiology of pertussis and pertussis immunization in theUnited Kingdom and the United Sates: a comparative study. Curr Probl Pediatr1984;14:1–78.

[2] Celentano LP, Massari M, Paramatti D, Salmaso S, Tozzi AE, EUVAC-NET Group.Resurgence of pertussis in Europe. Pediatr Infect Dis J 2005;24(9):761–5.

[3] Tanaka M, Vitek CR, Pascual FB, Bisgard KM, Tate JE, Murphy TV. Trends inpertussis among infants in the United States, 1980–1999. J Am Med Assoc2003;290:2968–75.

[4] Baron S, Njamkepo E, Grimprel E, Begue P, Desenclos JC, Drucker J, et al. Epi-

demiology of pertussis in French hospitals in 1993 and 1994: thirty years aftera routine use of vaccination. Pediatr Infect Dis J 1998;17:412–8.

[5] De Melker HE, Schellekens JFP, Neppelenbroek SE, Mooi FR, Rumke HC, Conyn-van Spaendonck MAE. Reemergence of pertussis in the highly vaccinatedpopulation of the Netherlands: observations on surveillance data. Emerg InfectDis 2000;6(4):348–57.

3 Vaccin

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

290 P. Rendi-Wagner et al. /

[6] Moerman L, Leventhal A, Slater PE, Anis E, Yishai R, Marva E. The re emergenceof pertussis in Israel. Isr Med Assoc J 2006;8:308–11.

[7] Rendi-Wagner P, Kundi M, Mikolasek A, Vecsei A, Fruhwirth M, Kollaritsch H.Hospital-based active surveillance of childhood pertussis in Austria from 1996to 2003: estimates of incidence and vaccine effectiveness of whole-cell andacellular vaccine. Vaccine 2006;24:5960–5.

[8] Ntezayabo B, De Serres G, Duval B. Pertussis resurgence in Canada largelycaused by a cohort effect. Pediatr Infect Dis J 2003;22:22–7.

[9] de Melker HE, Versteegh FG, Conyn-Van Spaendonck MA, Elvers LH, Berbers GA,van Der Zee A, et al. Specificity and sensitivity of high levels of immunoglob-uline G antibodies against pertussis toxin in a single serum sample fordiagnosis of infection with Bordetella pertussis. J Clin Microbiol 2000;38:800–6.

10] Giammanco A, Nardone A, Pebody R, Kafatos G, Andrews N, Chiarini A, et al.European Sero-Epidemiology Network 2: standardisation of immunoassayresults for pertussis requires homogeneity in the antigenic preparations. Vac-cine 2008;26(35):4486–93.

11] Giammanco A, Chiarini A, Maple PA, Andrews N, Pebody R, Gay N, et al. Euro-pean Sero-Epidemiology Network: standardisation of the assay results forpertussis. Vaccine 2003;22(1):112–20.

12] de Melker HE, Versteegh FG, Schellekens JF, Teunis PF, Kretzschmar M. Theincidence of Bordetella pertussis infections estimated in the population from acombination of serological surveys. J Infect 2006;53(2):106–13.

13] Teunis PF, van der Heijden OG, de Melker HE, Schellekens JF, Versteegh FG,Kretzschmar ME. Kinetics of the IgG antibody response to pertussis toxin afterinfection with B. pertussis. Epidemiol Infect 2002;129(3):479–89.

14] Ministry of Health, State of Israel. Notifiable Infectious Diseases in Israel, 54years of Surveillance, 1951 to 2004. Israel Center for Disease Control, Publica-tion No. 245; May 2006.

15] Pebody RG, Gay NJ, Giammanco A, Baron S, Schellekens J, Tischer A, et al.The seroepidemiology of Bordetella pertussis infection in Western Europe. Epi-

demiol Infect 2005;133(1):159–71.

16] Nagel J, de Graaf S, Schiif-Evers D. Improved serodiagnosis of whooping coughof IgG anti-LFP antibody levels. Dev Biol Stand 1985;61:325–30.

17] Cagney M, McIntyre PB, Heron L, Giammanco A, MacIntyre CR. The relationshipbetween pertussis symptomatology, incidence and serology in adolescents.Vaccine 2008;26:5547–53.

[

e 28 (2010) 3285–3290

18] Gordon M, Davies HD, Gold R. Clinical and microbiological features of childrenpresenting with pertussis to a Canadian pediatric hospital during an eleven-year period. Pediatr Infect Dis J 1994;13:617–22.

19] Wirsing von Konig CH, Postels-Multani S, Bock HL, Schmitt HJ. Pertus-sis in adults: frequency of transmission after household exposure. Lancet1995;346:1326–9.

20] Thomas MC, Lambert HP. From whom do children catch pertussis? Br Med J1987;295:751–2.

21] Crowcroft NS, Booy R, Harrison T, Spicer L, Britto J, Mok Q, et al. Severe andunrecognized: pertussis in UK infants. Arch Dis Child 2003;88:802–6.

22] Quinn HE, McIntyre PB, Backhouse JL, Gidding HF, Brotherton J, Gilbert GL. Theutility of seroepidemiology for tracking trends in pertussis infection. EpidemiolInfect 2009;1:1–8.

23] Klement E, Uliel L, Engel I, Hasin T, Yavzori M, Orr N, et al. An outbreak ofpertussis among young Israeli soldiers. Epidemiol Infect 2003;131(3):1049–54.

24] Arav-Boger R, Ashkenazi S, Gdalevich M, Cohen D, Danon YL. Seroprevalenceof pertussis antibodies among adolescents in Israel. Isr Med Assoc J 2000;2:174–7.

25] Srugo I, Benilevi D, Madeb R, Shapiro S, Shohat T, Somekh E, et al. Pertussisinfection in fully vaccinated children in day-care-centers, Israel. Emerg InfectDis 2000;6(5):526–9.

26] Jenkinson D. Duration of effectiveness of pertussis vaccine: evidence from a10-year community study. Br Med J 1988;296:612–4.

27] Qiushui H, Viljanen MK, Nikkari S, Lyytikainnen R, Mertsola J. Outcomes of Bor-detella pertussis infection in different age groups of an immunized population.J Infect Dis 1994;170:873–7.

28] Van Boven M, de Melker HE, Schellekens JFP, Kretzschmar M. Waning immunityand sub-clinical infection in an epidemic model: implications for pertussis inThe Netherlands. Math Biosci 2000;164:161–82.

29] Forsyth KD, Wirsing von Konig CH, Tan T, Caro J, Plotkin S. Prevention of per-

tussis: recommendations derived from the second Global Pertussis Initiativeroundtable meeting. Vaccine 2007;25(14):2634–42.

30] Tan T, Halperin S, Cherry JD, Edwards K, Englund JA, Glezen P, et al. Pertussisimmunization in the Global Pertussis Initiative North American region: rec-ommended strategies and implementation considerations. Pediatr Infect Dis J2005;24(5 Suppl.):S83–6.