Pediatrics 2013 Defay e1126 33

Measles in Children Vaccinated With 2 Doses of MMR

WHATS KNOWN ON THIS SUBJECT: School outbreak investigationin Quebec, Canada suggested that adolescents previouslyvaccinated with 2 doses of measles vaccine beginning at 12months of age were at greater measles risk than those whoserst dose was given at $15 months of age.

WHAT THIS STUDY ADDS: Greater measles risk among earlierrst-of-2-dose vaccine recipients was replicated as a generalizedprovincial nding during the 2011 epidemic in Quebec, Canada.The mechanism remains unknown, but the ndings warrantadditional evaluation in the context of measles elimination efforts.

abstractOBJECTIVE: A previous measles outbreak investigation in a high schoolin Quebec, Canada identied 2-dose vaccine effectiveness of 94%. Therisk of measles in 2-dose recipients was signicantly higher (24times) when measles vaccine was rst administered at 12 versus$15 months of age, with no signicant effect of the age at seconddose. Generalizability of this association was also assessed in theexpanded provincial data set of notied cases.

METHODS: This matched casecontrol study included only 2-dose recip-ients. All conrmed (laboratory or epidemiologically linked) cases inpatients aged 5 to 17 years were included. Each case was matched to5 controls.

RESULTS: A total of 102 cases and 510 controls were included; 89% ofcases were in patients 13 to 17 years old. When the rst dose wasadministered at 12 to 13 months compared with $15 months ofage, the risk of measles in participants outside the outbreak schoolwas 6 times higher (95% condence interval, 1.3329.3) and was 5.2times higher (95% condence interval, 1.9114.3) in the pooled esti-mate (participants from the outbreak school + outside that school).

CONCLUSIONS: A signicantly greater risk of measles among 2-doserecipients whose rst dose was given at 12 to 13 months rather than$15 months of age is conrmed in the larger Quebec data set. Themechanism remains unknown, but vaccine failures in 2-dose recipi-ents could have substantial implications for measles eliminationefforts through 2-dose vaccination. The optimal age at rst dosemay warrant additional evaluation. Pediatrics 2013;132:e1126e1133

AUTHORS: Fannie Defay, MSc,a Gaston De Serres, MD,PhD,a,b,c Danuta M. Skowronski, MD, FRCPC,d NicoleBoulianne, RN, MSc,a,b Manale Ouakki, MSc,b MoniqueLandry, MD,e Nicholas Brousseau, MD, FRCPC,f and Brian J.Ward, MD, FRCPCg

aUnit de Recherche en Sant Publique, CHUQ, Quebec, Canada;bInstitut National de Sant Publique du Qubec, Quebec, Canada;cLaval University, Quebec, Canada; dBritish Columbia Center forDisease Control, Vancouver, British Columbia, Canada; eMinistrede la Sant et des Services Sociaux, Quebec, Canada; fAgence dela Sant et des Services Sociaux de la Mauricie et du Centre-du-Qubec, Quebec, Canada; and gResearch Institute of the McGillUniversity Health Centre, Montral, Quebec, Canada

KEY WORDSmeasles, vaccine effectiveness

ABBREVIATIONSCIcondence intervalMMRmeaslesmumpsrubellaORodds ratioVEvaccine effectiveness

Ms Defay conceptualized and designed the study, carried outdata extraction and statistical analyses, contributed to theinterpretation of data, and drafted the initial manuscript; Drs DeSerres and Skowronski conceptualized and designed the study,contributed to the analysis and interpretation of data, andcritically reviewed the manuscript for important intellectualcontent; Ms Boulianne conceptualized and designed the study,coordinated and supervised data collection and extraction, andcritically reviewed the manuscript for important intellectualcontent; Ms Ouakki contributed to the data extraction andstatistical analyses and critically reviewed the manuscript forimportant intellectual content; Drs Landry, Brousseau, and Wardconceptualized and designed the study and critically reviewedthe manuscript for important intellectual content; and allauthors approved the nal manuscript as submitted.

www.pediatrics.org/cgi/doi/10.1542/peds.2012-3975

doi:10.1542/peds.2012-3975

Accepted for publication Aug 16, 2013

Address correspondence to Gaston De Serres, MD, PhD, InstitutNational de Sant Publique du Qubec 2400 dEstimauville,Quebec City, Quebec, Canada, G1E 7G9. E-mail: [email protected]

(Continued on last page)

e1126 DEFAY et al at Indonesia:AAP Sponsored on January 6, 2015pediatrics.aappublications.orgDownloaded from

When the live attenuated measles vac-cinewas developed in the 1950s,1 it wasinitially recommended for adminis-tration at 9 months. This was soonchanged when greater efcacy wasobserved with older age at vaccinationbecause of concern about possible in-terference from residual maternalmeasles antibody.2 Because maternalantibodies wane over time, inter-ference diminished with age, and pro-tection increased with older age atrst dose, plateauing at$15 months,as shown during epidemics from theearly 1990s.3

In the United States, the recommendedage for measles vaccination was 9months in 1963, 12 months in 1965, 15months in 1976, and 12 to 15 monthssince 1998,4 whereas Canada opted foradministration at 12 months in 1970and has stayed with that recommen-dation since.5 However, even in settingswhere nearly all children received 1dose of vaccine on schedule, measlestransmission continued, and additionalintervention was needed to achievecontrol.6 Measles vaccine failures weregenerally thought to be primary (ab-sence of protection immediately aftervaccination) rather than secondary(waning of immunity).7,8 Based on thishypothesis and the observation thata second dose generally correctedfailed seroconverters,911 many coun-tries implemented 2-dose programs inthe 1990s.

Between 1996 and 1997, all Canadianprovinces added a second measlesdose to their vaccination schedulesand conducted mass campaigns toadminister a second dose to all school-age children. In Quebec, the second-largest province (population 8 million),89% of children 5 to 17 years old re-ceived a second measles vaccine doseduring these campaigns, and the infantschedule was changed to 2 measlesmumpsrubella (MMR) doses at 12and 18 months of age.12 After 15 years

of decline, a large measles epidemicoccurred in Quebec in 2011, revealingunexpected vulnerability in adoles-cents previously vaccinated with 2doses of MMR.13,14 An investigation inthe high school at the origin of thisoutbreak reported an overall vaccineeffectiveness (VE) of 95.9% with 1 doseand 94.2% with 2 doses.14 An un-expected nding from this outbreakinvestigation was that in 2-doserecipients, VE was greater with olderage at rst dose, from 93% at 12months to 97.5% at$15 months.14 Therisk of measles was 2 to 4 timesgreater when children were rst vac-cinated between 12 and 14 monthsversus$15 months.14 Older age at thesecond dose or longer interval be-tween doses did not inuence thisobservation.14

Although statistically signicant, theseresults were based on a small number(41) of 2-dose cases. It was thereforeimportant toconrmthendingsand todetermine whether they were gener-alizable. Furthermore, most adoles-cents involved in the school outbreakwere born to mothers who had pre-viously beenmeasles infected, whereascurrently, almost all babies are born tovaccinated mothers who provide lowerconcentrations of antimeasles anti-bodies to their babies than motherswhoexperiencedwildvirus infection.15,16

Because vaccine-induced maternalantibodies are anticipated to disappearat an earlier age, infants born tovaccine-protected mothers may re-spond better to an earlier dose ofmeasles vaccine than infants fromprevious birth cohorts born to infectedmothers, if indeed interference frommaternal antibody is the principalmechanism for the negative effect ofage at rst pediatric dose on VE.

We therefore conducted a casecontrolstudy to estimate the risk of measles byage at rst and second dose, adjustingfor maternal status (previously infec-

ted versus vaccinated) using all casesin twice-vaccinated school-age patientsreported in Quebec in 2011.

METHODS

Measles conrmed by laboratory test-ing or epidemiologic link is notiable byboth physicians and laboratories inQuebec.17 Laboratory conrmation re-quires virus detection by culture orpolymerase chain reaction or de-velopment of measles-specic immu-noglobulin M in absence of recentvaccination. Epidemiologic link requiresclassic clinical presentation (fever$38.3C [101F] and cough or coryzaor conjunctivitis and a generalizedmaculopapular rash for at least 3 days)with epidemiologic link to a laboratory-conrmed measles case.

In this matched casecontrol study,inclusion criteria for cases and con-trols were having received 2 doses ofmeasles-containing vaccine, rst doseadministered at $12 months of age,second dose administered $28 daysafter dose 1 and$14 days before rashonset in the matched case, and agebetween 5 and 17 years. Cases includedonly conrmed measles as denedearlier and reported from across theprovince to public health betweenJanuary 1 and December 31, 2011.Controls were matched for the dateof birth (66 months) and school at-tended in 2010 to 2011. For each case, 5controls were randomly selected fromthe provincial measles vaccinationregistry among all students meetingmatching criteria. The vaccination sta-tus and dates of vaccination wereascertained through the provincialvaccination registry and other records.MMR-II (Merck Canada,Montreal, Quebec)was the only MMR vaccine administeredto the pediatric cohorts included in thisstudy.

In Canada, people born before 1970are considered to have been infectedby measles.5 This year marks the

ARTICLE

PEDIATRICS Volume 132, Number 5, November 2013 e1127 at Indonesia:AAP Sponsored on January 6, 2015pediatrics.aappublications.orgDownloaded from

beginning of the measles vaccinationprogram using live vaccine, and very fewpatients from the large measles out-breaks in Quebec (1989) or Ontario(19901992) were born before thatyear.18,19 Consequently, the status ofparticipants mothers (infected versusvaccinated) was assigned by year ofbirth and review of the provincial vacci-nation registry, which includes all resi-dents born after 1970 regardless ofvaccination status. Mothers not found inthe registry were considered to havebeen born before 1970 and categorizedas having been infected, and those bornin 1970 or later were considered vacci-nated.

The odds ratios (ORs) ofmeasles by ageat rst and second dose and maternalimmune status were estimated bymultivariable conditional logistic re-gression.

This work was conducted under legalmandate from the National Director ofPublic Health, authorized by the QuebecPublic Health Act without requirementfor research ethics board review.20

RESULTS

In 2011, among the 725 patients withconrmed measles cases,13 507 werebetween 5 and 17 years of age; 102 hadreceived 2 doses at$12months of age,1 had received 3 doses, 18 had received1 dose, 337 were unvaccinated, and 49had unknown vaccination status or nowritten proof. Of the 102 2-dose cases,82 were epidemiologically conrmed,and 20 were laboratory conrmed (17had measles-specic immunoglobulinM, and 3 others had positive viral cul-ture).

The 102 2-dose cases and their 510matched controls attended 17 schools,with the number of cases per schoolrangingfrom1to41(mean,6;median,2).Of the 12 high schools (grades 712), 3accounted for 73% of twice-vaccinatedcases; the high school that was rst

andmost affected (outbreak school) had41, another had 17, and the third had 16.There were only 6 (6%) cases in 5 ele-mentary schools (range, 12 per school).Two-thirds of cases were in boys, whowere at signicantly higher risk ofmeasles in univariate analysis (OR= 1.97;95% condence interval [CI], 1.33.1)(Table 1), a pattern also observed amongunvaccinated patients (Data not shown,available upon request). Only 11% ofcases were 5 to 12 years old at rashonset (2% were,10 years, 4% were 1011, 5% were 12). Adolescents 13, 14, 15,and 16 years old contributed dispro-portionately, with 17%, 20%, 25%, and20% of all cases, respectively, whereas17-year-olds represented 8% of thecases. Seventy percent of the mothers ofboth cases and controls were born be-fore 1970,15% between 1970 and 1972(during which there was still intensemeasles circulation), and 14% in 1973or later. Measles risk did not statisticallydiffer by maternal year of birth (Table 1).

Among participants outside the out-break school, the rst dose of measlesvaccine was administered at 12 or 13months of age in 90% of cases and 72%of controls, whereas 4.8% and 19.6%,respectively, received the rst dose at$15 months (Table 1). With no caserst vaccinated at exactly 15 months,we compared 12 months to the broaderage category $15 months, which in-cluded few cases. In univariate ana-lysis, the risk of measles was 6.87(95% CI, 1.5929.7) and 11.4 (95% CI,2.552.5) times higher when the rstdose had been administered at theage of 12 and 13 months, respectively,compared with $15 months. ORs werelower but still signicant in pooledanalysis (outbreak school + outside thatschool). Neither age at second dose norinterval between doses inuenced mea-sles risk, regardless of the inclusion orexclusion of participants from the mostaffected school.

Because the risk of measles in univar-iate analysis was similar with rst doseadministered at 12 or 13 months of age,the 2 ages were merged into a singlecategory (1213 months) in multivari-able analyses. In multivariable analyses,only age at rst dose and genderremained signicantly associated withmeasles risk (Table 2). The effect of ageat rst dose was not confounded bygender, age at second dose, interval be-tween doses, or maternal birth year. Inparticipants outside the outbreakschool, when the rst dose was admin-istered at 12 to 13 versus$15months ofage, measles risk was 6.2 times higher(95% CI, 1.3329.3; P = .02), and in pooledanalysis it was 5.2 times higher (95% CI,1.9114.26; P = .0013). When the rstdosewas administered at 14 versus$15months of age, the riskwas twice as highbut did not reach statistical signicance.In children who received their rst doseat 12 to 13 months, the risk of measleswas similar if their second dose wasadministered at $48 months versus,24 months old (participants outsidethe outbreak school OR = 0.72; 95% CI,0.212.29; P = .69; all participants, OR =0.98; 95% CI, 0.352.49; P = .898).

The risk ofmeasles in children who hadreceived their rst dose at 12 to 13versus $15 months stratied by ma-ternal birth year could not be assessedwith matching preserved. Because theoverall crude risk estimate was sim-ilar with matched (conditional logisticregression) and unmatched (un-conditional logistic regression) analysis(OR 5.26 vs 5.01) (Table 3), we conductedunmatched stratied exploration. Thetrend of greater risk with rst dose at 12to 13 versus $15 months was presentboth in children whose mothers wereborn before 1970 and those born laterbut was more pronounced in the for-mer (OR = 6.23 and 3.22, respectively)(Table 3). Only the OR for participantswhose mothers were born before 1970was statistically signicant, but the small


number of participants with youngermothers limits power and precludesdenitive conclusions.

DISCUSSION

Outbreak investigation of the highschool that triggered the 2011 measlesepidemic in Quebec suggested that therisk ofmeasles in 2-dose recipientswassignicantly higher (24 times) whenmeasles vaccine was rst adminis-tered at 12 versus $15 months ofage.15 The current study conrms thatthis effect was a generalized phenom-enon even outside the outbreak school,most evident in twice-vaccinated olderversus younger children. Although theeffect of age at rst dose was mostpronounced in children born to moth-

ers who had probably been infected bymeasles virus, it was also evident inchildren born to vaccinated mothersbut with less certainty.

Across the 26-year period during whichCanada relied on a single dose ofmeasles vaccine delivered at 12monthsof age (19701996), there was strongevidence of greater protection withdelivery of this dose at $15 months.5

The greater vulnerability of those vac-cinated at 12 months was reasoned tobe caused by interference from ma-ternal antibodies, addressed throughsecond-dose administration.911,21 Mea-sles vaccine protection thereafter wasconsidered to provide lifelong immu-nity.8 Increasing the age of the rstdose to 15 months in a 2-dose programwas therefore assumed unnecessary.

Our ndings challenge this assump-tion.

Apooledvefoldgreater riskofmeaslesamong thosewhoserstMMRdosewasadministered at 12 to 13 vs$15monthsis concerning, especially in the contextof measles elimination efforts that re-quire high levels of immunity. Previousserologic studies7,10,2226 based on ageat rst dose are consistent with ourepidemiologic observations. Althoughcaution is needed in the interpretationof immunogenicity ndings based onlaboratory, assay, and protectivethresholds applied, serologic studieshave shown that early age at rst doseleaves an imprint that affects both theimmediate second-dose antibody re-sponse and persistence. Children rstvaccinated at 12 months compared

TABLE 1 OR for Measles by Gender, Age, Maternal Birth Year, Age at First and Second Measles Dose, and Interval Between Doses, Estimated byUnivariate Conditional Logistic Regression

Participants Outside the Outbreak School All Participants

Cases (N = 61),n (%)

Controls (N = 305),n (%)

OR (CI 95%) Cases (N = 102),n (%)

Controls (N = 510),n (%)

OR (CI 95%)

Male gender 41 (67.2) 134 (43.9) 2.60 (1.54.7) 68 (66.7) 258 (50.6) 1.97 (1.33.1)Age, y

512 9 (14.8) 48 (15.8) a 11 (10.8) 64 (12.7) a

1314 20 (32.8) 105 (34.4) 37 (36.3) 187 (36.7)1517 32 (52.5) 152 (49.8) 54 (52.9) 259 (50.8)

Mothers year of birth,1970 45 (73.8) 211 (69.2) Reference 72 (70.6) 356 (69.8) Reference19701972 9 (14.8) 53 (17.4) 0.79 (0.41.7) 16 (15.7) 77 (15.1) 1.02 (0.561.86)1973+ 7 (11.5) 41 (13.4) 0.77 (0.31.9) 14 (13.7) 77 (15.1) 0.89 (0.471.71)

Age at rst dose, mo12 38 (62.3) 167 (54.8) 6.87 (1.5929.7) 67 (65.7) 287 (56.3) 4.88 (1.9012.6)13 19 (31.2) 53 (17.4) 11.4 (2.4752.5) 25 (24.5) 80 (15.7) 6.59 (2.3918.2)14 2 (3.3) 28 (9.2) 2.20 (0.2916.5) 5 (4.9) 43 (8.4) 2.47 (0.688.96)15 0 9 (3.0) Referenceb 0 18 (3.5) Referenceb

1623 1 (1.6) 28 (9.2) 2 (1.9) 51 (10.0)$24 1 (1.6) 20 (6.6) 3 (2.9) 31 (6.1)

Age at second dose, no,18 2 (3.3) 6 (2.0) 4.23 (0.6328.4) 4 (3.9) 9 (1.8) 3.92 (0.9715.5)1823 49 (80.3) 212 (69.5) 2.93 (0.978.80) 78 (76.5) 352 (69.0) 1.95 (0.914.19)2447 6 (9.8) 41 (13.4) 1.64 (0.426.38) 11 (10.8) 77 (15.1) 1.09 (0.412.83)$48 4 (6.6) 46 (15.1) Reference 9 (8.8) 72 (14.1) Reference

Interval between rst andsecond doses, mo,6 15 (24.6) 83 (27.2) 1.29 (0.513.24) 24 (23.5) 136 (26.7) 1.13 (0.562.29)6 20 (32.8) 72 (23.6) 2.01(0.834.88) 31 (30.4) 130 (25.5) 1.53 (0.783.01)711 16 (26.2) 83 (27.2) 1.36 (0.563.27) 29 (28.4) 133 (26.1) 1.39 (0.722.70)1247 8 (13.1) 43 (14.1) Reference 12 (11.8) 71 (13.9) Reference$48 2 (3.3) 24 (7.9) 6 (5.9) 40 (7.8)

a Not appropriate, matching variable.b Reference is all ages $15 months.c Reference is all interval $12 months.

ARTICLE


with $13 months are less likely toseroconvert after reimmunization andto have signicantly lower antibodylevels.10 This phenomenon has alsobeen noted in children vaccinatedyounger than 12 months.7,22,23 Childrenwithout detectable plaque reductionneutralization antibody after a rstdose responded to revaccination witha primary-type response and main-tained antibody titers after revaccina-tion above levels believed to beprotective, whereas revaccination ofchildren with low levels of plaque re-duction neutralization antibody in-duced a secondary-type response, withantibodies returning rapidly to levelsbelow protective threshold.24 In Germany,among 7000 2-dose recipients, theproportion without detectable anti-bodies decreased steadily with olderage at rst dose to reach a nadir be-tween 18 and 23 months.25 The pro-portion seronegative also increasedwith time since the second dose, andthe proportion seronegative $6 yearsafter the second dose was similar tothat observed in single-dose recipi-ents.25 Similarly, in the United States,increased antibody titers in responseto a second MMR dose rapidly returnedto prebooster levels, andmost subjects(72%, 13/18) with low titers 10 yearsafter boosting were also in the lowestquartile for antibodies before boost.26

This return to the level of protectionafforded by the rst dose based ontime elapsed since the second dose isconsistent with the similar VE we ob-served after 1 or 2 doses (95.9% vs94.2%, respectively) during our recentschool outbreak investigation.15

The risk associated with an early age atrst dose was present but less pro-nounced in children born to vaccinatedmothers than in children born to in-fectedmothers. This evidence is limitedbythe small number of participants born tovaccinated mothers and the use of themothers birth year (before 1970) asTA

BLE2

ORof

Mea

sles

Asso

ciated

With

Ageat

FirstMea

sles

VaccineDo

sein

Multiv

ariate

Analysis

(Con

ditio

nalRe

gression

)

AdjustingCo

variates

Participan

tsOu

tsidetheOu

tbreak

Scho

olAllP

artic

ipan

ts

Ageat

FirstD

ose,

mo

Covariates

Ageat

FirstD

ose,

mo

Covariates

1213

1415

+1213

1415

+

OR(95%

CI)

OR(95%

CI)

OR(95%

CI)

OR(95%

CI)

OR(95%

CI)

OR(95%

CI)

Crud

e7.81

(1.83

33.30)

2.19

(0.29

16.3)

Referenc

e5.26

(2.06

13.40)

2.44

(0.67

8.84

)Re

ferenc

eGe

nder

(ref

female)

7.15

(1.66

30.69)

2.01

(0.27

15.1)

Referenc

eMale=2.46

(1.36

4.43

)5.1(2.00

13.03)

2.34

(0.64

8.51

)Re

ferenc

eMale=1.92

(1.22

3.02

)Ag

eat

seco

nddo

se(ref,24

mo)

6.71

(1.45

31.1)

1.93

(0.25

15.1)

Referenc

e2447

mo:

0.86

(0.31

2.37

);$48

mo:

0.72

(0.22

2.34

)5.05

(1.86

13.75)

2.38

(0.63

8.99

)Re

ferenc

e2447

mo:

0.78

(0.36

1.70

);$48

mo:

0.98

(0.42

2.27

)Mothe

rsyear

ofbirth(ref

before

1970

)7.71

(1.81

32.8)

2.16

(0.29

16.1)

Referenc

e0.82

(0.42

1.59

)5.27

(2.07

13.44)

2.44

(0.67

8.87

)Re

ferenc

e$19

70=0.93

(0.57

1.53

)

Alla

bove

factors

6.24

(1.33

29.3)

1.88

(0.24

14.9)

Referenc

e5.21

(1.91

14.26)

2.54

(0.67

9.64

)Re

ferenc

e


a proxy for having been measlesinfected. Despite good vaccine cover-age when the Quebec universal programbegan in 1970, low-level transmissioncontinued (Supplemental Fig 1) for a fewyears. As a result, some mothers bornsince 1970 may have received vaccinebut also been infected, leading tomisclassication tending to errone-ously suggest the same pattern forchildren born to vaccinated mothers.Because more infants are born tovaccinated mothers now than in thecurrent study, the current phenome-non may be partly historical. However,even in the absence of maternalantibodies, the immune response tovaccination improves with older age atrst dose with respect to both titerand avidity.27,28 These observations arelimited to vaccination between 6 and12 months of age but suggest thatmaturation of the immune systemprobably also plays an importantrole in the quality and durability ofprotection.27

The current study has other limitations. Itwas not possible to estimate absolute VEbecause that requires the comparison ofvaccinated andunvaccinated people, andmost with no dose in the provincialregistry were in fact vaccinated but withmissing data. The higher incidence in

adolescents (75.6 per 100 000) com-pared with adults 18 to 35 years old(5.3 per 100 000) may be explained bythe greater and more intense exposureopportunities associated with theschool outbreak, which may haveovercome whatever protection teen-agers had. Adolescents also receivedtheir 2 doses of MMR at 12 and 18months through the routine program,whereas most young adults receiveda monovalent measles vaccine as theirsecond dose at school age during the1996 mass campaign. The small num-ber of cases (5) in patients who re-ceived their rst dose at $15 monthsof age may raise concerns about un-stable estimates, although statisticalpower and precision would be drivenby the magnitude of the difference inrisk by age at rst dose rather than theabsolute cell size. However, the smallnumber of cases in patients with rstvaccination at $15 months precludedadditional stratication to dene anoptimal age for protection. Our resultscannot be explained by the nature ofthe vaccine because all participantsreceived the Merck MMR-II, virtually theonly measles-containing product usedin North America at that time. Mis-handling or other local factors affect-ing vaccine quality would not explain

a selective effect with rst dose by agebut would affect all vaccinated people.Finally, we observed a predominance ofboys among cases that we could notexplain; the schools affected includeda gender mix, and other social context(eg, sports teams) could not explainthis gender pattern. Ultimately, how-ever, gender had no confounding effecton the association between measlesand age at rst dose that we report.

In the United States, despite severalmeasles importations in 2011, nonetriggered epidemic spread. This maybe partly explained by differences incontrol measures, vaccine coverage,or chance opportunities for super-spreading ignition events. However, 2other factors may have contributed tothe better protection of US adolescentsin 2011: older age at rst dose andagreaterproportionborn to vaccinatedmothers. In 1989, the Advisory Com-mittee on Immunization Practices andthe American Academy of Family Physi-cians recommended 2 measles doseswith the rst dose at 15 months.4 In1998, the recommended age for the rstdose was changed to 12 to 15 months.29

In 2011, adolescents 13 to 16 years oldwere born between 1995 and 1998. TheUS National Immunization Survey showsthat the proportion of children aged 19to 35 months who received their rstMMR dose before 13 months increasedfrom 31% in 1997 to 44.7% in 2001,30 stilllower than the 56% in our 13- to 16-year-old controls. US adolescents were alsomore likely to be born to vaccinatedmothers. Because the US measles pro-gram started in the early 1960s, women#30 years old in 1995 to 1998 weremostly vaccinated and gave birth to two-thirds of the babies. In contrast, 70% ofour cases and controls were born tomothers presumed to have been infected(born before 1970). Because children inboth the United States and Canada arenow born to vaccinated mothers, theeffect of age at rst dose may become

TABLE 3 OR of Measles Associated With Age at First Measles Vaccine Dose, in Overall Matchedand Unmatched Logistic Regression and Stratied by Maternal Year of Birth(Unmatched Logistic Regression, All Participants)

Maternal Year of Birth Age at First Dose, mo Cases Controls OR (95% CI)

Overall matched logistic regression 1213 92 367 5.26 (2.0613.4)14 5 43 2.44 (0.678.84)$15 5 100 Reference

Overall unmatched logistic regression 1213 92 367 5.01 (1.9912.7)14 5 43 2.33 (0.608.02)

Stratied analysis$15 5 100 Reference

,1970 1213 65 254 6.23 (1.9020.4)14 4 29 3.37 (0.7115.9)$15 3 73 Reference

$1970 1213 27 103 3.22 (0.7214.4)14 1 14 0.94 (0.0811.6)$15 2 27 Reference

ARTICLE


less pronounced. In guiding the optimaltiming of primary immunization, it isnecessary to weigh our ndings againstthe recognized severity of measles ininfants and VE in protecting againstthat, especially in the context of anoutbreak.21

Currently, measles has been elimi-nated from the Americas, and globalexperience overwhelmingly supportsdurable immunity from 2 doses ofmeasles vaccine. However, the un-expected vulnerability we have identi-ed in twice-vaccinated people during

the epidemic in Quebec should beconsidered a signal warranting ad-ditional investigation. In particular,other locations with ongoing measlesactivity can explore the epidemiologicassociations we have raised, andserosurveys of infants who receivedan early versus later rst dose ofMMRwould be informative.

CONCLUSIONS

Although unvaccinated people shouldremain the prime target for measles

vaccination, the unexpected vulner-

ability we have identied in twice-

vaccinated people could ultimately

lead to failed measles elimination

efforts. If the effect of early vacci-

nation permanently alters the ability

to respond to subsequent doses, even

adding a third or fourth dose may

not provide long-lasting protection.

Therefore, it is critical to understand

the mechanisms of primary vaccine

failure or loss of vaccine protection

that our ndings may signal.

REFERENCES

1. Plotkin SA, Orenstein WA, Oft PA. Measlesvaccine. In: Vaccines. 6th ed. Philadelphia,PA: Saunders Elsevier; 2012:352387

2. Albrecht P, Ennis FA, Saltzman EJ, KrugmanS. Persistence of maternal antibody ininfants beyond 12 months: mechanism ofmeasles vaccine failure. J Pediatr. 1977;91(5):715718

3. De Serres G, Sciberras J, Naus M, BoulianneN, Duval B, Rochette L. Protection after twodoses of measles vaccine is independent ofinterval between doses. J Infect Dis. 1999;180(1):187190

4. CDC. Measles prevention: recommendationsof the Immunization Practices AdvisoryCommittee (ACIP). MMWR Recomm Rep.1989;38(S-9):119

5. National Advisory Committee on Immuni-zation. Canadian immunization guide, Ever-green edition. Available at: www.phac-aspc.gc.ca/publicat/cig-gci/. Accessed December21, 2012

6. Gustafson TL, Lievens AW, Brunell PA,Moellenberg RG, Buttery CMG, SehulsterLM. Measles outbreak in a fully immunizedsecondary-school population. N Engl J Med.1987;316(13):771774

7. Markowitz LE, Albrecht P, Orenstein WA, LettSM, Pugliese TJ, Farrell D. Persistence ofmeasles antibody after revaccination. JInfect Dis. 1992;166(1):205208

8. Markowitz LE, Preblud SR, Fine PEM, OrensteinWA. Duration of live measles vaccine-induced immunity. Pediatr Infect Dis J.1990;9(2):101110

9. Watson JC, Pearson JA, Markowitz LE, et al.An evaluation of measles revaccinationamong school-entry-aged children. Pediat-rics. 1996;97(5):613618

10. Poland GA, Jacobson RM, Thampy AM, et al.Measles reimmunization in children sero-negative after initial immunization. JAMA.1997;277(14):11561158

11. World Health Organisation. Module 7: measlesupdate 2009. The immunological basis for im-munization series. Geneva, Switzerland: WHO;2009. Available at: www.who.int/immunization/documents/ISBN9789241597555/en/index.html. Accessed December 21, 2012

12. King A, Varughese P, De Serres G, TipplesGA, Waters J; Working Group on MeaslesElimination. Measles elimination in Canada.J Infect Dis. 2004;189(suppl 1):S236S242

13. De Serres G, Markowski F, Toth E, et al.Largest measles epidemic in North Americain a decadeQuebec, Canada, 2011: contri-bution of susceptibility, serendipity, andsuperspreading events. J Infect Dis. 2013;207(6):990998

14. De Serres G, Boulianne N, Defay F, et al.Higher risk of measles when the rst doseof a 2-dose schedule of measles vaccine isgiven at 1214 months versus 15 months ofage. Clin Infect Dis. 2012;55(3):394402

15. Leuridan E, Hens N, Hutse V, Ieven M, AertsM, Van Damme P. Early waning of maternalmeasles antibodies in era of measleselimination: longitudinal study. BMJ. 2010;340:c1626

16. Leuridan E, Van Damme P. Passive trans-mission and persistence of naturally ac-quired or vaccine-induced maternal antibodiesagainst measles in newborns. Vaccine. 2007;25(34):62966304

17. Ministre de la Sant et des Services Sociauxdu Qubec. Dnitions nosologiques. Mal-adies dorigine infectieuse. 8th ed. Montreal,Canada: Gouvernement du Qubec; 2011

18. McLean ME, Walsh PJ, Carter AO, LavignePM. Measles in Canada: 1989. Can Dis WklyRep. 1990;16(42):213218

19. Wong T, Lee-Han H, Bell B, Daley J, Bailey N,Vanderpol M. Measles epidemic in Waterlooregion, Ontario, 19901991. Can Dis WklyRep. 1991;17(41):219224

20. Gouvernement du Qubec. Loi sur la santpublique (L.R.Q., chapitre S-2.2, mis jourle 1er mars 2005). Montreal, Canada:Gouvernement du Qubec; 2001

21. Uzicanin A, Zimmerman L. Field effective-ness of live attenuated measlescontainingvaccines: a review of published literature. JInfect Dis. 2011;204(suppl 1):S133S148

22. Stetler HC, Orenstein WA, Bernier RH, et al.Impact of revaccinating children who ini-tially received measles vaccine before 10months of age. Pediatrics. 1986;77(4):471476

23. Murphy MD, Brunell PA, Lievens AW, ShehabZM. Effect of early immunization on anti-body response to reimmunization withmeasles vaccine as demonstrated by enzyme-linked immunosorbent assay (ELISA). Pediatrics.1984;74(1):9093

24. Wilkins J, Wehrle PF. Additional evidenceagainst measles vaccine administration toinfants less than 12 months of age: alteredimmune response following active/passiveimmunization. J Pediatr. 1979;94(6):865869

25. Poethko-Mller C, Mankertz A. Sero-epidemiology of measles-specic IgG anti-bodies and predictive factors for low ormissing titres in a German population-based cross-sectional study in childrenand adolescents (KiGGS). Vaccine. 2011;29(45):79497959


26. LeBaron CW, Beeler J, Sullivan BJ, et al.Persistence of measles antibodies after 2doses of measles vaccine in a posteliminationenvironment. Arch Pediatr Adolesc Med. 2007;161(3):294301

27. Gans H, Yasukawa L, Rinki M, et al. Immuneresponses to measles and mumps vacci-nation of infants at 6, 9, and 12 months. JInfect Dis. 2001;184(7):817826

28. Nair N, Gans H, Lew-Yasukawa L, Long-Wagar AC, Arvin A, Grifn DE. Age-dependent differences in IgG isotype andavidity induced by measles vaccine re-ceived during the rst year of life. J InfectDis. 2007;196(9):13391345

29. CDC. Measles, mumps, and rubella: vaccineuse and strategies for eliminations ofmeasles, rubella, and congenital rubella

syndrome and control of mumps: recom-mendations of the Advisory Committee onImmunization Practices (ACIP). MMWR RecommRep. 1998;47(RR-8):157

30. Centers for Disease Prevention and Con-trol. U.S. vaccination coverage reported viaNIS. Available at: www.cdc.gov/vaccines/stats-surv/nis/default.htm#nis. Accessed De-cember 18, 2012

(Continued from rst page)

PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).

Copyright 2013 by the American Academy of Pediatrics

FINANCIAL DISCLOSURE: The authors have indicated they have no nancial relationships relevant to this article to disclose.

FUNDING: Funded by the Ministre de la Sant et des Services Sociaux du Qubec.

POTENTIAL CONFLICT OF INTEREST: Dr De Serres has received research funds from GlaxoSmithKline (GSK) and Sano Pasteur and reimbursement for travelexpenses to attend a GSK advisory board meeting. Ms Boulianne has received research funds from GSK. Dr Ward is medical ofcer for Medicago Inc, has receivedresearch funds from GSK, Sano Pasteur, and Pzer, and was a member of ad hoc advisory boards for GSK, Sano Pasteur, Pzer, Merck, and Novartis. The otherauthors have indicated they have no potential conicts of interest to disclose.

ARTICLE


DOI: 10.1542/peds.2012-3975; originally published online October 21, 2013; 2013;132;e1126Pediatrics

Ouakki, Monique Landry, Nicholas Brousseau and Brian J. WardFannie Defay, Gaston De Serres, Danuta M. Skowronski, Nicole Boulianne, Manale


rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.Grove Village, Illinois, 60007. Copyright 2013 by the American Academy of Pediatrics. All and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elkpublication, it has been published continuously since 1948. PEDIATRICS is owned, published, PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly

at Indonesia:AAP Sponsored on January 6, 2015pediatrics.aappublications.orgDownloaded from

DOI: 10.1542/peds.2012-3975; originally published online October 21, 2013; 2013;132;e1126Pediatrics

Ouakki, Monique Landry, Nicholas Brousseau and Brian J. WardFannie Defay, Gaston De Serres, Danuta M. Skowronski, Nicole Boulianne, Manale


http://pediatrics.aappublications.org/content/132/5/e1126.full.htmllocated on the World Wide Web at:

The online version of this article, along with updated information and services, is

of Pediatrics. All rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.Boulevard, Elk Grove Village, Illinois, 60007. Copyright 2013 by the American Academy published, and trademarked by the American Academy of Pediatrics, 141 Northwest Pointpublication, it has been published continuously since 1948. PEDIATRICS is owned, PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly


ServicesUpdated Information &

htmlhttp://pediatrics.aappublications.org/content/132/5/e1126.full.including high resolution figures, can be found at:

Supplementary Material

6/peds.2012-3975.DCSupplemental.htmlhttp://pediatrics.aappublications.org/content/suppl/2013/10/1Supplementary material can be found at:

References

html#ref-list-1http://pediatrics.aappublications.org/content/132/5/e1126.full.at:This article cites 23 articles, 11 of which can be accessed free

Citations

html#related-urlshttp://pediatrics.aappublications.org/content/132/5/e1126.full.This article has been cited by 1 HighWire-hosted articles:

Rs)3Peer Reviews (PPost-Publication

http://pediatrics.aappublications.org/cgi/eletters/132/5/e1126Rs have been posted to this article 32 P

Subspecialty Collections

munization_subhttp://pediatrics.aappublications.org/cgi/collection/vaccine:imVaccine/Immunization

_diseases_subhttp://pediatrics.aappublications.org/cgi/collection/infectiousInfectious Diseases

alth_subhttp://pediatrics.aappublications.org/cgi/collection/school_heSchool Healththe following collection(s):This article, along with others on similar topics, appears in

Permissions & Licensing

tmlhttp://pediatrics.aappublications.org/site/misc/Permissions.xhtables) or in its entirety can be found online at: Information about reproducing this article in parts (figures,

Reprints http://pediatrics.aappublications.org/site/misc/reprints.xhtml

Information about ordering reprints can be found online:

rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.Grove Village, Illinois, 60007. Copyright 2013 by the American Academy of Pediatrics. All and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elkpublication, it has been published continuously since 1948. PEDIATRICS is owned, published, PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly


Documents

Pediatrics 2013 Defay e1126 33