Ppdv

Na

b

c

a

ARRAA

KDAS

1

npt(i(osaomht

itcr

0d

Vaccine 28 (2010) 7215–7220

Contents lists available at ScienceDirect

Vaccine

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

ost-licensure comparison of the safety profile of diphtheria/tetanus/whole cellertussis/haemophilus influenza type b vaccine and a 5-in-1iphtheria/tetanus/acellular pertussis/haemophilus influenza type b/polioaccine in the United Kingdom

ick Andrewsa,∗, Julia Stoweb, Lesley Wisec, Elizabeth Millerb

Statistics Unit, Health Protection Agency, Centre for Infections, 61 Colindale Avenue, London NW9 5EQ, United KingdomImmunisation, Hepatitis and Blood Safety Department, Health Protection Agency, Centre for Infections, United KingdomVigilance Risk Management of Medicines, Pharmacovigilance Risk Management Group, Medicines and Healthcare products Regulatory Agency, United Kingdom

r t i c l e i n f o

rticle history:eceived 9 June 2010

a b s t r a c t

General practitioner consultation data were used to compare the reactogenicity in infants of a 5-in-1acellular pertussis vaccine (DTaP5/Hib/IPV) introduced in the United Kingdom in 2004 to the 4-in-1 whole

eceived in revised form 10 August 2010ccepted 11 August 2010vailable online 26 August 2010

eywords:TP

cell-pertussis vaccine (DTwP/Hib) that it replaced. For each vaccine the incidence in the week followingvaccination was compared to other periods to obtain a relative incidence. A lower relative incidence ofcrying, fever and local reactions was seen with DTaP5/Hib/IPV than DTwP/Hib. Although there were noother significant differences between vaccines the relative incidence was significantly above one on theday of vaccination for convulsions following DTwP/Hib and for apnoea/collapse following DTaP5/Hib/IPV.

cellular pertussis combinationsafety

. Introduction

In September 2004 a combined diphtheria, tetanus, 5 compo-ent acellular pertussis, Haemophilus influenza type b, inactivatedoliovirus vaccine (DTaP5/Hib/IPV) manufactured by Sanofi Pas-eur MSD (PediacelTM) replaced the combined whole cell pertussiswP) containing DTwP/Hib vaccine formerly given for primarymmunisation with oral poliovirus vaccine in the United KingdomUK). DTaP5/Hib/IPV is the acellular pertussis combination vaccinef choice for the UK immunisation programme as it achieves aatisfactory antibody response to the Hib component under theccelerated UK primary schedule of 2/3/4 months [1], unlike somether DTaP/Hib combination vaccines used in the UK [2,3]. Further-ore it contains five purified pertussis antigens—a formulation that

as been shown in clinical trials to provide protection equivalento that of the wP-containing vaccine previously used in the UK [4].

In a pivotal randomised trial that compared the immunogenic-

ty [1] and reactogenicity [5] profile of DTaP5/Hib/IPV with that ofhe existing UK DTwP/Hib vaccine, local redness and swelling, andommon systemic symptoms, occurred at a significantly lowerate after DTaP5/Hib/IPV. In the current study we have assessed

∗ Corresponding author. Tel.: +44 020 8327 7419; fax: +44 020 8200 7868.E-mail address: nick.andrews@hpa.org.uk (N. Andrews).

264-410X/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.oi:10.1016/j.vaccine.2010.08.062

© 2010 Elsevier Ltd. All rights reserved.

the impact that the reduction in common and generally mildsymptoms measured in clinical trials has on consultations forsuspected vaccine reactions seen in general practice. For this, theGeneral Practice Research Database (GPRD) was used to comparethe incidence of consultations for events compatible with a vaccinereaction in children under 1 year of age in DTaP5/Hib/IPV andDTwP/Hib vaccinated cohorts.

2. Methods

We identified consultations for events of interest and detailsof vaccinations from the GPRD, which is one of the world’s largestprimary care databases [6]. It holds data on consultations, referrals,prescriptions and vaccinations for over 3 million active patients inpractices throughout the UK (5.7% of the population). We selectedchildren in the GPRD whose practice record had an ‘acceptable’ sta-tus and with an ‘up to standard’ date earlier than the child’s dateof birth. The ‘up to standard’ date is the date when the practicecomplied with specific quality measures based on completeness,continuity and plausibility in key areas. Acceptable status is givento a patient when certain data quality conditions have been met,

such as no events recorded before birth date, age is less than 115years and the gender field is completed. Children included in theanalysis were those born from January 2003 to March 2006 whowere registered from birth until the age of 12 months or death ifearlier. This gave an estimated cohort size of 121,700 children from

7216 N. Andrews et al. / Vaccine 28 (2010) 7215–7220

Table 1Number of infants by birth cohort and eligibility for DTwP/Hib or DTaP5/Hib/IPV vaccine for primary immunisation and number of events according to birth cohort.

Birth cohort

January 03–June 04 July 04 August 04–March 06

Vaccine generally eligible for DTwP/Hib DTwP/Hib and DTaP5/Hib/IPV DTaP5/Hib/IPV

Approximate number of individuals in GPRD 58,900 3500 59,300

Individuals with at least one event 17,451 1034 18,034

ConditionCrying 2229 148 2223Diarrhoea 10,569 639 10,939Feeding problem 1366 80 1515Fever 4034 263 4286

wosArPt

btwtnaa

bddctDD

2

cciiiovpdfio

trdp

2

o

“Vaccine reaction” 177Vomiting 7041Convulsion/fit/seizure 369Apnoea/collapse/cyanosis/pallor 136

hom events were identified. Only those children with at leastne event of interest were selected for analysis. The events con-idered were READ or OXMIS codes for the following conditions:naphylaxis, Apnoea, Collapse, Convulsions, Crying, Cyanosis, Diar-hoea, Encephalitis, Feeding problem, Fever, Fit, Hypersensitivity,allor, Seizure, SIDS, Vaccine Reaction and Vomiting. A full list ofhe READ/OXMIS codes may be obtained from the authors.

After obtaining dates of events of interest, year and month ofirth for the child and dates of vaccinations the gap between mul-iple events in the same children was determined and any eventsithin 7 days of a previous similar event were assumed to be part of

he same episode and therefore dropped. Any children with immu-isation dates recorded before 35 days of age or within 2 weeks ofprevious dose or with more than 3 separate dates recorded werelso dropped.

In general the type of vaccine used was not recorded in the GPRDut it could be inferred by looking at date of vaccination. In order toetermine which vaccines were most likely to have been given inifferent periods data from child health vaccination systems whichontained batch information was used. This showed that vaccina-ions given from January 2003 to September 2004 were almost allTwP/Hib and those given from November 2004 were almost allTaP5/Hib/IPV. In October 2004 both vaccines were widely used.

.1. Statistical methods

The relative incidence (RI) of events shortly after vaccinationompared to other periods was estimated using the self-controlledase series method (SCCS) [7]. This method, which only usesndividuals that have the event of interest, implicitly controls forndividual level confounding. Age was adjusted for using 2-weekntervals. Risk periods of interest were the day of vaccination,ne to three days post-vaccination and days four to seven post-accination for each dose and overall. A 1 week pre-vaccinationeriod was removed from the background period to allow forelayed vaccination due to illness. RI was estimated with 95% con-dence intervals, and vaccines compared by calculating the ratiof the relative incidence estimates with 95% confidence intervals.

The SCCS method was unsuitable for analysing “vaccine reac-ion” outcome codes since no events are expected in non-vaccineisk periods. For this outcome therefore the rate per 100,000 personays was estimated with a 95% confidence interval. Analyses wereerformed using the software Stata [8].

.2. Validation

Validation was sought for all infants with an event recordedn the day of vaccination. The free text in the patient record

7 72366 704530 25911 140

was reviewed for the period seven day before to seven daysafter the day of the event to identify whether the recorded dateof diagnosis accurately reflected event date or was the resultof retrospective recording. For events coded as ‘vaccine reac-tion’, apnoea/collapse/cyanosis/pallor or convulsions/fits/seizurevalidation was also sought if the event occurred in the period1 week before to 1 week after vaccination. For these eventsthe diagnosis was confirmed by looking for appropriate symp-toms and comments. In total validation was sought for 1096events.

3. Results

From the estimated total cohort size of 121,700 children, 36,519(30% of the cohort) had at least one outcome event of interestrecorded. These children had received a total of 104,814 vac-cine doses (2.87 each on average). Assuming that this averagewas typical of the entire cohort, the estimated number of vac-cine doses given to all children was 349,400 of which about162,600 were DTwP/Hib, 177,600 DTaP5/Hib/IPV and 9200 thatcould have been either. There were too few cases of anaphy-laxis (n = 10), encephalitis (n = 0), sudden infant death syndrome(SIDS) (n = 11) or hypersensitivity (n = 5) for analysis. Only oneof these events occurred within 7 days of vaccination (one SIDSat 7 days post-DTwP/Hib). To increase study power convulsions(n = 473), fits (n = 44) and seizures (n = 141) were combined togetheras were pallor (n = 21), collapse (n = 27), apnoea (n = 125) andcyanosis (n = 114). Table 1 shows the number of events avail-able for analysis according to the type of event and period ofbirth. Events in October 2004 were omitted from the analysis asboth DTwP/Hib and DTaP5/Hib/IPVvaccines were in use in thismonth.

The number of events by age is shown in Fig. 1 according toevent type; these showed different patterns by age. Those eventswith more occurrences around the time of DTP vaccination (2–5months) will yield more power than those where the events occurlater in the first year of life. The RI (or absolute incidence) estimatesare shown in Table 2 for the day of vaccination as well as days 1–3and 4–7 post-vaccination.

3.1. Crying and fever

Differences are apparent between the vaccines for crying and

fever. The RI estimates are shown further in Figs. 2 and 3 accord-ing to dose. This demonstrates that for DTaP5/Hib/IPV crying onlyincreases on the day of vaccination, whereas for DTwP/Hib crying isparticularly high on the day of vaccination for the first dose, but alsois slightly raised on days 1–3 and 4–7 irrespective of dose. Fever on

N. Andrews et al. / Vaccine 28 (2010) 7215–7220 7217

ution

tidi

3

aettsvg

Fig. 1. Age distrib

he day of vaccination has the highest RI after the first dose declin-ng with no increase by the third dose for both vaccines. Fever onays 1–3 is consistently high irrespective of dose for DTwP/Hib but

s not raised for DTaP5/Hib/IPV vaccine.

.2. Feeding problems, vomiting and diarrhoea

There is a raised RI of consultations for feeding problems as wells vomiting on the day of vaccination but not on days 1–3. This

ffect does not differ between vaccines and probably reflects oppor-unistic reporting on the day of vaccination. This is consistent withhe lower RI in the 3 days after vaccination. Diarrhoea, in contrast,hows a decreased RI on the day of vaccination and days 1–3 post-accination for both vaccines. This is likely to be because childrenet vaccinated when they are generally well.

of adverse events.

3.3. Convulsions, fits and seizures

There is no difference between the vaccines although forDTwP/Hib the RI on day 0 is significantly raised while forDTaP5/Hib/IPVthe RI is not significantly above 1 on day 0. Theestimated number of events on day 0 attributable to DTwP/Hib iscalculated as (3.14/4.14) × 7 = 5.31, which, based on the estimateof 162,600 doses given, is an attributable risk of about 1 in 30,000doses.

3.4. Apnoea, collapse, cyanosis, pallor

There is no difference between the vaccines although forDTaP5/Hib/IPV the RI on day 0 is significantly raised while forDTwP/Hib the RI is not significantly above 1 on day 0. The esti-mated number of events on day 0 attributable to DTaP5/Hib/IPV is

7218 N. Andrews et al. / Vaccine 28 (2010) 7215–7220

Table 2Relative or absolute incidence of various adverse events following DTwP/Hib and DTaP5/Hib/IPV vaccines and ratios (DTaP5/Hib/IPV: DTwP/Hib) of relative or absoluteincidence.

Event Time since vaccination (days) Vaccine

DTwP/Hib RI (95% CI) [events] DTaP5/Hib/IPV RI (95%CI) [events]

Ratio of relativeincidence DTaP5/DTwP(95% CI)

Crying 0 6.51 (5.53–7.66) [170] 3.09 (2.49–3.85) [88] 0.47 (0.36–0.62)1–3 1.44 (1.18–1.75) [111] 0.77 (0.60–0.99) [65] 0.53 (0.39–0.74)4–7 1.37 (1.15–1.64) [137] 1.11 (0.92–1.34) [121] 0.81 (0.63–1.05)

Feeding problem 0 2.49 (1.87–3.32) [50] 2.89 (2.24–3.72) [65] 1.16 (0.79–1.70)1–3 0.91 (0.69–1.21) [54] 0.63 (0.46–0.86) [42] 0.69 (0.46–1.05)4–7 1.08 (0.86–1.35) [82] 1.10 (0.89–1.36) [94] 1.02 (0.75–1.39)

Fever 0 1.84 (1.30–2.61) [33] 1.65 (1.19–2.30) [36] 0.90 (0.56–1.44)1–3 2.28 (1.90–2.75) [124] 0.83 (0.64–1.09) [55] 0.36 (0.27–0.50)4–7 0.96 (0.76–1.22) [71] 0.97 (0.78–1.21) [87] 1.01 (0.73–1.39)

Vomiting 0 1.49 (1.21–1.84) [92] 1.82 (1.51–2.19) [114] 1.22 (0.92–1.62)1–3 0.84 (0.72–0.99) [154] 0.70 (0.58–0.83) [130] 0.83 (0.65–1.06)4–7 0.84 (0.72–0.97) [201] 0.85 (0.74–0.98) [209] 1.01 (0.82–1.24)

Diarrhoea 0 0.66 (0.50–0.89) [46] 0.56 (0.42–0.76) [43] 0.85 (0.57–1.27)1–3 0.54 (0.45–0.65) [113] 0.59 (0.50–0.70) [136] 1.09 (0.85–1.40)4–7 0.90 (0.79–1.02) [254] 0.92 (0.82–1.04) [287] 1.02 (0.86–1.22)

Convulsion/fit/seizure 0 4.14 (1.92–8.92) [7] 2.05 (0.65–6.46) [3] 0.50 (0.12–1.97)1–3 1.37 (0.63–2.95) [7] 0.45 (0.11–1.83) [2] 0.33 (0.07–1.64)4–7 1.02 (0.47–2.20) [7] 0.34 (0.08–1.38) [2] 0.33 (0.06–1.72)

Apnoea/collapse/cyanosis/pallor 0 1.22 (0.30–4.96) [2] 3.22 (1.30–7.98) [5] 2.64 (0.50–14.03)1–3 0.62 (0.20–1.99) [3] 1.56 (0.71–3.39) [7] 2.52 (0.63–9.98)4–7 1.31 (0.63–2.72) [8] 1.24 (0.57–2.71) [7] 0.95 (0.33–2.75)

Event Time since vaccination (days) Absolute incidence as arate per 100,000 days(95% CI) [events]

Absolute incidence as arate per 100,000 days(95% CI) [events]

Ratio of incidenceDTaP5/DTwP (95% CI)

“Vaccine reaction” 0 35.06 (26.56–45.42) 6.19 (3.09–11.08) [11] 0.18 (0.08–0.37)

0–9.96–3.4

cod

3

a

[57]1–3 7.18 (5.04–7 2.00 (1.0

alculated as (2.22/3.22) × 5 = 3.45, which, based on the estimatef 177,600 doses given, is an attributable risk of about 1 in 50,000oses.

.5. Vaccine reactions

For vaccine reactions the absolute incidence is shown asrate per 100,000 days. Following DTwP/Hib the incidence

Fig. 2. Relative incidence (95% CI) for crying in

8) [35] 0.94 (0.30–2.19) [5] 0.13 (0.04–0.43)2) [13] 0.56 (0.15–1.44) [4] 0.28 (0.06–1.21)

is about six times higher than that following DTaP5/Hib/IPVvaccine.

4. Validation

4.1. Day of vaccination

The results of the validation of events on the day of vaccina-tion are shown in Table 3. As expected vaccine reactions and crying

intervals around DTP vaccines by dose.

N. Andrews et al. / Vaccine 28 (2010) 7215–7220 7219

Fig. 3. Relative incidence (95% CI) for fever in intervals around DTP vaccines by dose.

Table 3Validation of events on the day of vaccination.

Event Vaccine Total episodes Free text available Confirmedoccurring onthe day

Confirmed butretrospective

Not confirmed

Apnoea/collapse/cyanosis/pallor DTwP/Hib 2 1 1 0 0DTaP5/Hib/IPV 5 5 4 1 0

Convulsion/fit/seizure DTwP/Hib 7 6 5 1 0DTaP5/Hib/IPV 3 3 2 1 0

“Vaccine reaction” DTwP/Hib 57 56 49 1 6DTaP5/Hib/IPV 11 11 10 0 1

Vomiting DTwP/Hib 92 46 9 33 4DTaP5/Hib/IPV 114 48 4 42 2

Feeding problem DTwP/Hib 50 34 2 31 1DTaP5/Hib/IPV 65 47 0 45 2

Fever DTwP/Hib 33 26 6 3 17DTaP5/Hib/IPV 36 24 8a 4 12

Crying DTwP/Hib 170 47 34 5 8DTaP5/Hib/IPV 88 46 35 5 6

a For two episodes of fever the vaccine was not given even though it was recorded as given.

Table 4Validation of events 1 week before to 1 week after vaccination.

Event Vaccine Total episodes Free text available Confirmed Seen in hospital

Apnoea/collapse/cyanosis/pallor DTwP/Hib 21 11 10 7DTaP5/Hib/IPV 28 14 13 7

Convulsions/fit/seizure DTwP/Hib 26 19 16 12DTaP /Hib/IPV 13 6 4 2

roebpTlnpfitdc

5

“Vaccine reaction” DTwP/Hib 105DTaP5/Hib/IPV 20

elate to the vaccination just given, but many other events recordedn the day of vaccination appear to be opportunistic recording ofvents that occurred previously, for example when the 8-weekaby check and vaccination take place on the same day. This wasarticularly so for feeding and vomiting and to a lesser extent fever.he more serious events (convulsions, fits, seizures, apnoea col-apse, cyanosis and pallor) were all confirmed and were generallyot retrospectively recorded. The only event where a substantialroportion was not confirmed was fever (Table 3). The lack of any

ree text information for this event may reflects the fact that fevers a well-documented reaction not requiring additional informa-ion to be recorded by the GP. When the free text informationid confirm the presence of fever it often related to a telephoneonversation rather than a visit.

102 90 219 17 1

4.2. Events from 1 week before to 1 week after vaccination

The diagnosis coding was confirmed for a high proportionof cases (88% overall). Of those which were confirmed, 61% ofapnoea/collapse/cyanosis/pallor episodes and 70% of the convul-sions were seen in hospital. Of the consultations coded as vaccinereaction 78 (73%) were confirmed as being local reactions (seeTable 4).

5. Discussion

This study shows that the lower incidence of common, gener-ally mild symptoms documented in clinical trials after aP- than

7 ccine 2

wvvracrhitboDovv1[asswSbwcsdttvcreaoacfttwdii

tvwsaorwmfaivdrcv

tr

[

[

[

[

220 N. Andrews et al. / Va

P-containing vaccines is reflected in a reduced frequency of post-accination events requiring a medical consultation when theseaccines are in routine use. Many of these events are the typicaleactions such as fever, crying and local reactions that parents arelerted to when bringing their child for vaccination. While con-erning for parents, such symptoms are generally transient andespond to analgesics/antipyretics such as paracetamol. However,igh temperature may cause a febrile convulsion in susceptible

nfants for which hospital referral is usually indicated. Althoughhe RI for convulsions/fits/seizures was not significantly differentetween vaccines, there was a significantly raised RI on the dayf vaccination for DTwP/Hib (4.14 95% CI 1.92–8.92) but not forTaP5/Hib/IPV vaccine (2.05, 95% CI 0.65–6.46). An increased riskf convulsions in the 0–3 day post-vaccination period has been pre-iously reported in the UK after the third dose of wP-containingaccine with a RI of 3.0 (95% CI 1.66–5.5) an attributable risk ofin 12 500 doses for convulsions resulting in hospital admission

9]. This attributable risk is about 1 in 37,500 when considered asrisk per dose over all three doses and is similar to the 1 in 30,000

een in this study. A reduction in hospital admissions for convul-ions has been reported from Canada following their change fromP to the aP5-containing vaccine in infants [10], and in the United

tates a reduction in health maintenance organisation reports ofoth fever and convulsions was found following the change fromP to aP in 1997 [11]. An increase in consultations on the day of vac-

ination for events coded as apnoea/collapse/cyanosis/pallor waseen for DTaP5/Hib/IPV but not for DTwP/Hib vaccine, although theifference between vaccines was not statistically significant. Hypo-onic/hyporesponsive (HHE) episodes in which the infant becomesransiently pale, floppy and unresponsive within a few hours ofaccination (usually with first dose) have been a cause for con-ern after wP-containing vaccines, though they have also beeneported after aP-containing vaccines [12]. The aetiology of suchvents remains obscure but the risk would be expected to be lowerfter aP than wP-containing vaccines. For example, the frequencyf HHEs after a UK whole cell vaccine was significantly higher thanfter a DTaP5 vaccine when compared in a large randomised effi-acy trial in Sweden [13] and a decrease in hospital admissionsor vaccine-related HHEs was seen in Canada following the changeo the aP5-containing vaccine [10]. The clinical information abouthe apnoea/collapse/cyanosis/pallor events captured in our studyas necessarily limited and they may not have met the strict caseefinition for an HHE. Nevertheless, our finding of a significantly

ncreased RI for consultations for apnoea/collapse/cyanosis/pallors unexpected and merits further investigation.

For common events there was evidence that the incidence onhe day of vaccination is elevated by retrospective recording. Manyaccination appointments in young infants coincide with the 8-eek “well baby” check where any concerns from about the baby

ince birth are recorded, while immediately after vaccination thereppears to be a healthy vaccine effect. This makes interpretationf the RI estimates difficult for these common events, unlike theare events such a convulsions and apnoea/collapse/cyanosis/pallorhich were generally not retrospectively recorded. The SCCSethod of analysis should enable valid comparisons between dif-

erent vaccines because it controls for individual level confoundingnd would also take into account any changes in the backgroundncidence. A simple comparison of rates post-vaccination betweenaccines would not have taken such changes in background inci-ence into account; furthermore a comparison of rates in vaccineisk periods compared to non-vaccine risk periods for the whole

ohort would be biased by differences in consultation behaviour inaccinated and unvaccinated children.

One potential confounding factor is a change in clinical prac-ice over time, such as increased use of paracetomol which caneduce fever if given prophylactically after vaccination [14]. How-

[

8 (2010) 7215–7220

ever, there is no evidence of a clinically significant benefit to localreactions from paracetamol use after vaccination [14], nor shouldany increase in its prophylactic use coincided with the introductionof DTaP5/Hib/IPV.

For consultations for “vaccine reactions”, many of which areinjection site reactions, absolute incidence was substantially higherfor DTwP/Hib than DTaP5/Hib/IPV vaccine. While this differencecould reflect changes in parental consultation patterns whenDTaP5/Hib/IPVwas introduced, it seems unlikely that there wouldbe such large changes in a short time period.

In conclusion, the post-licensure experience in the UK follow-ing the change to the DTaP5/Hib/IPV vaccine has confirmed thatthe improved reactogenicity profile documented in clinical trialswith this vaccine has resulted in a significant reduction in the fre-quency of medically attended adverse events in the immediatepost-vaccination period.

Acknowledgement

J. Stowe is funded from a grant from the Department of Health(grant number 039/031).

References

[1] Kitchin NR, Southern J, Morris R, Hemme F, Thomas S, Watson MW,et al. Evaluation of a diphtheria–tetanus–acellular pertussis-inactivatedpoliovirus–Haemophilus influenzae type b vaccine given concurrently withmeningococcal group C conjugate vaccine at 2, 3 and 4 months of age. ArchDis Child 2007;92:11–6.

[2] McVernon J, Andrews N, Slack MP, Ramsay ME. Risk of vaccine failure afterHaemophilus influenzae type b (Hib) combination vaccines with acellular per-tussis. Lancet 2003;361(9368):1521–3.

[3] Goldblatt D, Richmond P, Millard E, Thornton C, Miller E. The induction ofimmunologic memory after vaccination with Haemophilus influenzae type bconjugate and acellular pertussis-containing diphtheria, tetanus, and pertussisvaccine combination. J Infect Dis 1999;180(2):538–41.

[4] Olin P, Rasmussen F, Gustafsson L, Hallander HO, Heijbel H. Randomisedcontrolled trial of two-component, three-component, and five-componentacellular pertussis vaccines compared with whole-cell pertussis vaccine. AdHoc Group for the Study of Pertussis Vaccines. Lancet 1997;350(9091):1569–77[Erratum in: Lancet 1998;351(9100):454].

[5] Kitchin N, Southern J, Morris R, Hemme F, Cartwright K, Watson M, et al. Arandomised controlled study of the reactogenicity of an acellular pertussis-containing pentavalent infant vaccine compared to a quadrivalent whole cellpertussis-containing vaccine and oral poliomyelitis vaccine, when given con-currently with meningococcal group C conjugate vaccine to healthy UK infantsat 2, 3 and 4 months of age. Vaccine 2006;24(18):3964–70.

[6] General Practice Research Database. http://www.gprd.com/home/default.asp[accessed 09.08.10].

[7] Farrington CP, Nash J, Miller E. Case series analysis of adverse reactions tovaccines: a comparative evaluation. Am J Epidemiol 1996;143(11):1165–73[Erratum in: Am J Epidemiol 1998;147(January(1)):93].

[8] Stata reference StataCorp. Stata Statistical Software: Release 8. College Station,TX: Stata Corporation; 2003.

[9] Farrington P, Pugh S, Colville A, Flower A, Nash J, Morgan-Capner P,et al. A new method for active surveillance of adverse events fromdiphtheria/tetanus/pertussis and measles/mumps rubella vaccine. Lancet1995;345(8949):567–9.

10] Le Saux N, Barrowman NJ, Moore DL, Whiting S, Scheifele D, HalperinS. Decrease in hospital admissions for febrile seizures and reports ofhypotonic–hyporesponsive episodes presenting to hospital emergency depart-ments since switching to acellular pertussis vaccine in Canada: a report fromIMPACT. Pediatrics 2003;112(5):e348.

11] Davis RL, Kolczak M, Lewis E, Nordin J, Goodman M, Shay D, et al. Activesurveillance of vaccine safety: a system to detect early signs of adverse events.Epidemiology 2005;16:336–41.

12] DuVernoy TS, Braun MM. Hypotonic–hyporesponsive episodes reported tothe Vaccine Adverse Event Reporting System (VAERS), 1996–1998. Pediatrics2000;106(4):E52.

13] Olin P, Rasmussen F, Gustafsson L, Hallander HO, Heijbel H. Randomisedcontrolled trial of two-component, three-component, and five-componentacellular pertussis vaccines compared with whole-cell pertussis vaccine.Ad Hoc Group for the Study of Pertussis Vaccines. Lancet 1997;350(9091):

1569–77.

14] Prymula R, Siegrist CA, Chlibek R, Zemlickova H, Vackova M, SmetanaJ, et al. Effect of prophylactic paracetamol administration at time ofvaccination on febrile reactions and antibody responses in children:two open-label, randomised controlled trials. Lancet 2009;374(9698):1339–50.