Vaccine 25 (2007) 4875–4879

Review

Do immunisations reduce the risk for SIDS? A meta-analysis

M.M.T. Vennemann a,∗ , M. Höffgen b , T. Bajanowski c ,
H.-W. Hense d , E.A. Mitchell e
a Institute of Legal Medicine, University of Münster, Roentgenstr. 23, 48129 Münster, Germany
b Clemens Hospital, Münster, Germany
c Institute of Legal Medicine, University of Duisburg-Essen, Germany
d Institute of Epidemiology and Social Medicine, University of Münster, Germany
e Department of Paediatrics, University of Auckland, New Zealand

Received 15 November 2006; received in revised form 29 January 2007; accepted 27 February 2007
Available online 16 March 2007

Abstract

Background: There are claims that immunisations cause sudden infant death syndrome (SIDS), but some studies have found either no
association or that they are associated with a reduced risk of SIDS.
Aims: To conduct a meta-analysis examining the relationship between immunisation and SIDS.
Methods: Nine case–controls studies were identified examining this association, of which four adjusted for potential confounders.
Results: The summary odds ratio (OR) in the univariate analysis suggested that immunisations were protective, but the presence of hetero-
geneity makes it difficult to combine these studies. The summary OR for the studies reporting multivariate ORs was 0.54 (95% CI = 0.39–0.76)
with no evidence of heterogeneity.
Conclusions: Immunisations are associated with a halving of the risk of SIDS. There are biological reasons why this association may be causal,
but other factors, such as the healthy vaccinee effect, may be important. Immunisations should be part of the SIDS prevention campaigns.
© 2007 Elsevier Ltd. All rights reserved.

Keywords: SIDS; Meta-analysis; Immunisation

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4876
2. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4876
2.1. Selection criteria for studies into the meta-analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4876
2.2. Statistical methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4876
3. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4876
3.1. Exclusion of studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4876
3.2. Univariate analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4876
3.3. Multivariate analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4876
4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4877
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4878
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4878

∗ Corresponding author. Tel.: +49 251 8355648; fax: +49 251 8355300.
E-mail address: Mechtild.Vennemann@ukmuenster.de (M.M.T. Vennemann).

0264-410X/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.vaccine.2007.02.077
4876 M.M.T. Vennemann et al. / Vaccine 25 (2007) 4875–4879

1. Introduction sation all authors reported on was diphtheria, tetanus and

pertussis (DTP). In most studies this was combined with oral
In industrialised countries SIDS is still the leading cause polio, which was the common polio vaccine until the late
of death in the first year of life after the neonatal period. The 1990s and early 2000s when it was changed to an intramus-
highest incidence for SIDS is between the second and the cular inactivated polio virus vaccine in most countries. In
fifth months of life. At this time the first immunisations are this meta-analysis the DTP vaccine is used as a marker for
administered. This temporal association raises the question immunisation.
as to whether or not immunisations are a risk factor for SIDS,
some arguing that immunisations are causing SIDS [1,2]. 2.2. Statistical methods
Since the early 1990s several well-designed case–control
studies have demonstrated a number of risk factors for SIDS, For the OR two-by-two tables for each study were
such as the prone and side sleeping position, smoking of the extracted from the publications. The univariate and multi-
mother in pregnancy and overheating [3–6]. None of these variate ORs were also extracted. A separate summary OR
studies have shown that immunisations are a risk factor for was calculated for the univariate and multivariate ORs using
SIDS. Indeed a few of the studies suggest that vaccination is the fixed effect and random effect inverse-variance methods
protective against SIDS [7,8]. of meta-analyses [11]. The Breslow–Day test for heterogene-
This study aims to bring all well-designed case–control ity was calculated. A p-value less than 0.05 was considered
studies together and perform a meta-analysis to increase the to indicate that heterogeneity was present.
power and to make an evidence based statement whether
immunisations are a risk factor for SIDS or not.
3. Results

2. Methods 3.1. Exclusion of studies

In the search for all relevant studies two of the authors The ECAS [12] study reported only immunisation in the
(MH, MMV) independently searched “PubMed” for the fol- last 7 days and the Irish [13], the Nordic [14], the Scottish
lowing subjects: “SIDS, SUDI, SUD, Sudden Infant Death [15] and Chicago [16] case–control studies have not reported
Syndrome, cot death” in combination with “vaccination, vac- the risk of SIDS with immunisations and were not consid-
cine, immunisation, DTP, diphtheria, pertussis, polio, tetanus, ered further. An earlier German study [17] was excluded as
epidemiology”. In a second phase we looked at all the refer- the controls were selected from well-infant visit to paediatric
ence lists of the relevant studies to find more studies which offices, which may bias the sample to compliance with immu-
could have been overlooked in the first attempt. In addi- nisations. Furthermore the response rate for cases was 46%
tion two of the authors (EAM, MMV) attended the “Global only.
Strategy Task Force Meeting” at the SIDS International con-
ferences in 2002 and 2004 and asked the participants whether 3.2. Univariate analysis
there were more studies on immunisations which might have
not been published. Nine case–control studies were included in the first analy-
sis [7–10,18–22]. Table 1 shows the first author, country, type
2.1. Selection criteria for studies into the meta-analysis of immunisation, the number of cases and controls immu-
nised and the total number of subjects in each study and
Studies which were not case–control studies and studies the univariate odd ratio (OR) and 95% confidence intervals
about SIDS which did not analyse the influence of vacci- (95% CI) for the risk of SIDS if immunised with DTP. Also
nations on SIDS mortality were excluded. All cases–control shown are the multivariate OR and 95% CI published by the
studies about SIDS and immunisation were considered. The authors. The fixed and random effect methods produced sim-
following selection criteria were set prior to reviewing the ilar results, so only the random effects model is presented.
papers: (1) the study had to have more than 25 cases. (2) The summarized odds ratio for the studies was 0.58 (95%
The controls should be representative of the population. The CI = 0.46–0.73) (Fig. 1), indicating that immunisation is asso-
search included publications up to July 2006. There was no ciated with a significantly reduced risk of SIDS. However,
restriction on language. If there were multiple publications heterogeneity was present (χ2 = 24.52, d.f. = 8, p = 0.002).
from one study, only the publication with the most relevant
information was used. 3.3. Multivariate analyses
Vaccination schedules varied from country to country. In
some countries it started at 6 weeks (New Zealand) [7], There were four studies which reported multivariate ORs.
but more commonly at 3 months. Sometimes the immu- The variables included in the multivariate studies varied, but
nisation schedule changed during a study period [7,9,10] in general covered the main socio-economic, maternal, infant
(New Zealand, Germany, USA). The one type of immuni- and child care practices which have been associated with
 M.M.T. Vennemann et al. / Vaccine 25 (2007) 4875–4879 4877

SIDS. The summarised odds ratio using the random effect

0.29–0.83
0.49–2.36
0.24–0.85
0.25–1.00
95% CI
method was 0.54 (95% CI = 0.39–0.76), and was very similar
to the fixed effect method (results not shown). There was no
evidence of heterogeneity (χ2 = 3.67, d.f. = 3, p = 0.30).

Not reported
Not reported
Not reported
Not reported
Not reported
Multivariate

4. Discussion

0.48
1.08
0.45
0.51
OR

This meta-analysis has shown immunisation is associated

with a significantly lower risk of SIDS. However, before dis-
cussing the possible reasons for this finding the potential
0.17–1.10
0.47–0.67
0.07–0.63
0.50–1.11
0.81–2.01
0.32–0.69
0.46–1.66
0.38–0.63
0.28–0.59
limitations must be considered. Firstly there is consider-
95% CI

able clinical diversity in the studies. The studies come from

many developed countries, and include different immunisa-
tion schedules. The studies were undertaken both before and
after the “back to sleep” campaigns, which have resulted
Univariate OR

in changes in the epidemiology of SIDS [23,24]. The pro-

portion of controls immunised varied considerably in these
studies, ranging from 13.8% [21] to 94.7% [10]. Secondly
0.43
0.56
0.22
0.75
1.28
0.47
0.88
0.49
0.41

there is methodological diversity. The quality of the studies

varied, for example differences in the percentage of SIDS
cases autopsied, participation rates, and whether ascertain-
number of controls

ment of immunisation status was based on clinical records

immunised/total
No. of controls

or parental report. In the univariate analysis there was sta-

1256/1373
818/1514

822/1234

tistical heterogeneity, which indicates that the summary OR

213/225
189/401

585/971
90/332

47/341
27/52

must be treated with caution. However, the studies included

in the multivariate analysis were from more recent, large,
well-conducted studies. No heterogeneity was seen, thus we
have confidence in the summary OR, which shows that the
immunised/total
number of cases

risk of SIDS was halved by immunisation.

No. of cases

There are a number of possible explanations for this find-

285/716

233/279

149/303
154/307

ing. The seasonal distribution in the occurrence of SIDS

54/135
38/118

14/114
23/29
8/26

[25] and the high prevalence of respiratory tract symptoms

suggests infection is a factor in SIDS [26]. A number of
different viruses and bacteria have been implicated [27].
DTPa , polio, HiBc , HepBb

Bordetella pertussis may be especially important as an asso-

DTPa , HepBc at 6 weeks

DTPa , oral polio, HiBc

ciation between epidemic pertussis and sudden unexpected

Type of immunisation

DTPa and oral polio

death in infants has been observed [28] and B. pertussis infec-

DTPa , oral polio

tion in infants frequently causes apnoea [29]. If apnoea leads

to the death of the infant the cause of death may inappropri-
ately be labelled as SIDS. The immunisation schedules in the
DTPa

DTPa

DTPa

DTPa

studies reported here all included immunisation with B. per-

tussis. Immunisation may reduce the incidence of reported
SIDS by reducing unrecognised B. pertussis infection.
Diphtheria, tetanus, and pertussis vaccine.
New Zealand

Staphylococcus aureus strains producing enterotoxins

Studies reporting immunisation and SIDS

Germany

have been implicated in SIDS [30]. Immunisation may induce

Country

France
France

France

antibodies that cross react with pyrogenic staphylococcal tox-

USA
USA
UK

UK

ins, thus protecting the infant from such infections [31,32].

Haemophilus influencae B.

Immunisation may also cause non-specific enhancement

of immunological activity and reduce infection from other
Jonville-Bera et al. [20]

Jonville-Bera et al. [21]

Taylor and Emery [22]

viruses and bacteria not directly covered by the vaccines given

Vennemann et al. [9]
Hoffmann et al. [19]

Flahault et al. [18]

Mitchell et al. [7]

Walker et al. [10]

Fleming et al. [8]

[33,34].
Hepatitis B.

The immediate effect of immunisation is similar to

Reference

that of a mild infection. In view of the often reported

Table 1

association of SIDS with minor infection the ECAS study

a
b
c

specifically examined whether risk of SIDS was associated

4878 M.M.T. Vennemann et al. / Vaccine 25 (2007) 4875–4879

Fig. 1. Immunisation and risk of SIDS: univariate analysis.

Fig. 2. Immunisation and risk of SIDS: multivariate analysis.

with immunisation in the last 7 days. They reported that immunisation and reduction in SIDS, then SIDS mortality
univariatly the OR was quite insignificant (OR = 1.27 with may be reduced further by achieving high immunisation rates
95% CI = 0.89–1.81). After the multivariate adjustment the at the scheduled times in early infancy.
OR remained insignificant.
Immunisations may be indirectly associated with a reduc-
tion in SIDS. Vaccination may be avoided during illness and Conflict of interest
infections, the so-called healthy vaccinee effect [35]. Thus
the reduction in SIDS with immunisations may be a marker Dr. Mechtild Vennemann compiled this work on request
of the well being of the infant, and not directly related to the from the “Epidemiology Working Group” of the “Interna-
immunisation. tional Society for the Study and Prevention of Infant death”
Children born into poor socio-economic circumstances are (ISPID). Since November 2005 she has been working in the
less likely to be immunized [36,37]. In one study risk factors Institute of Legal Medicine at the University of Muenster.
for lack of immunisation include low socio-economic status, This position is part of a multi-center research project, lead
maternal smoking and intention not to breastfeed [37], all of by the Robert Koch Institute in Berlin and funded by the
which are known risk factors for SIDS. This illustrates the German Federal Ministry of Health, the Paul Ehrlich Insti-
importance of confounding [38]. However, in the multivari- tute, Sanofi Pasteur MSD, and GlaxoSmithKline Biologicals.
ate analysis the studies controlled for these factors and the The analysis was undertaken before Dr. Vennemann started
apparent protective effect remained (Fig. 2). her current position. The co-authors take responsibility for
What should be advised? Certainly the data are in the the interpretation of the data, and thus the co-authors believe
opposite direct to the assertion that immunisation causes there is no conflict of interest.
SIDS [1,2]. Parents can be reassured that immunisation with
vaccines on the current schedule, particularly DTP vaccine,
does not cause SIDS. Acknowledgements
The benefits of immunisation are well established [39]. If
a country changes their immunisation schedule to a different Prof. Ed Mitchell is supported by the Child Health
age, this provides an opportunity to examine changes in the Research Foundation.The statement is endorsed by the “Epi-
SIDS mortality rate for the age group covered by the change demiology Working Group” of the “International Society for
in immunisation. If there is a causal relationship between the Prevention of Infant Death” (ISPID).
 M.M.T. Vennemann et al. / Vaccine 25 (2007) 4875–4879 4879

References Development Cooperative Epidemiological Study of Sudden Infant

Death Syndrome risk factors. Pediatrics 1987;79(4):598–611.
[1] Davies P, Chapman S, Leask J. Antivaccination activists on the world [20] Jonville-Bera AP, Autret E, Laugier J. Sudden infant death syn-
wide web. Arch Dis Child 2002;87(1):22–5. drome and diphtheria–tetanus–pertussis–poliomyelitis vaccination
[2] Scheibner V. Response to Leask and McIntyre’s attack on myself as a status. Fundam Clin Pharmacol 1995;9(3):263–70.
public opponent of vaccination. Vaccine 2003;22(1):vi–ix. [21] Jonville-Bera AP, Autret-Leca E, Barbeillon F, Paris-Llado J. Sudden
[3] Fleming PJ, Blair PS, Bacon C, Bensley D, Smith I, Taylor E, et al. unexpected death in infants under 3 months of age and vaccination
Environment of infants during sleep and risk of the sudden infant death status—a case–control study. Br J Clin Pharmacol 2001;51(3):271–6.
syndrome: results of 1993–1995 case–control study for confidential [22] Taylor EM, Emery JL. Immunization and cot deaths. Lancet 1982;
inquiry into stillbirths and deaths in infancy. Confidential Enquiry into 2(8300):721.
Stillbirths and Deaths Regional Coordinators and Researchers. BMJ [23] Blair PS, Sidebotham P, Berry PJ, Evans M, Fleming PJ. Major epi-
1996;313(7051):191–5. demiological changes in sudden infant death syndrome: a 20-year
[4] Jorch G, Schmidt-Troschke SBT, Heinecke A, Findeisen M, Nowack population-based study in the UK. Lancet 2006;367(9507):314–9.
C, Rabe G, et al. Epidemiologische Risikofaktoren des plötzlichen [24] Mitchell EA. The changing epidemiology of SIDS following the
Kindstods. Monatsschr Kindrheilkd 1994;142:137–47. national risk reduction campaigns. Pediatr Pulmonol Suppl 1997;16:
[5] Mitchell EA, Taylor BJ, Ford RP, Stewart AW, Becroft DM, Thompson 117–9.
JM, et al. Four modifiable and other major risk factors for cot death: the [25] Mitchell EA, Clements M, Williams SM, Stewart AW, Cheng A,
New Zealand study. J Paediatr Child Health 1992;28(Suppl 1):S3–8. Ford RP, The New Zealand Cot Death Study Group. Seasonal differ-
[6] Findeisen M, Vennemann M, Brinkmann B, Ortmann C, Rose I, ences in risk factors for sudden infant death syndrome. Acta Paediatr
Kopcke W, et al. German study on sudden infant death (GeSID): 1999;88(3):253–8.
design, epidemiological and pathological profile. Int J Legal Med [26] Gilbert RE, Fleming PJ, Azaz Y, Rudd PT. Signs of illness preceding
2004;118(3):163–9. sudden unexpected death in infants. BMJ 1990;300(6734):1237–9.
[7] Mitchell EA, Stewart AW, Clements M, New Zealand Cot Death Study [27] Rambaud C, Guibert M, Briand E, Grangeot-Keros L, Coulomb-
Group. Immunisation and the sudden infant death syndrome. Arch Dis L’Hermine A, Dehan M. Microbiology in sudden infant death syndrome
Child 1995;73(6):498–501. (SIDS) and other childhood deaths. FEMS Immunol Med Microbiol
[8] Fleming PJ, Blair PS, Platt MW, Tripp J, Smith IJ, Golding J. The UK 1999;25(1/2):59–66.
accelerated immunisation programme and sudden unexpected death in [28] Lindgren C, Milerad J, Lagercrantz H. Sudden infant death and preva-
infancy: case–control study. BMJ 2001;322(7290):822. lence of whooping cough in the Swedish and Norwegian communities.
[9] Vennemann MM, Butterfass-Bahloul T, Jorch G, Brinkmann B, Find- Eur J Pediatr 1997;156(5):405–9.
eisen M, Sauerland C, et al. Sudden infant death syndrome: no increased [29] Southall DP, Thomas MG, Lambert HP. Severe hypoxaemia in pertus-
risk after immunisation. Vaccine 2007;25(2):336–40. sis. Arch Dis Child 1988;63(6):598–605.
[10] Walker AM, Jick H, Perera DR, Thompson RS, Knauss TA. [30] Blackwell CC, Moscovis SM, Gordon AE, Al Madani OM, Hall
Diphtheria–tetanus–pertussis immunization and sudden infant death ST, Gleeson M, et al. Cytokine responses and sudden infant death
syndrome. Am J Public Health 1987;77(8):945–51. syndrome: genetic, developmental, and environmental risk factors. J
[11] Petitti D. Meta-analysis. Decision analysis and cost-effectiveness anal- Leukoc Biol 2005;78(6):1242–54.
ysis. 2nd ed. Oxford: Oxford University Press; 2000. [31] Essery SD, Raza MW, Zorgani A, MacKenzie DA, James VS, Weir DM,
[12] Carpenter RG, Irgens LM, Blair PS, England PD, Fleming P, Huber J, et al. The protective effect of immunisation against diphtheria, pertussis
et al. Sudden unexplained infant death in 20 regions in Europe: case and tetanus (DPT) in relation to sudden infant death syndrome. FEMS
control study. Lancet 2004;363(9404):185–91. Immunol Med Microbiol 1999;25(1/2):183–92.
[13] Matthews T, McDonnell M, McGarvey C, Loftus G, O’Regan M. A [32] Toro K, Meszaros R, Meszaros A, Csukas Z. Change in immunisa-
multivariate “time based” analysis of SIDS risk factors. Arch Dis Child tion schedule and sudden infant death syndrome in Hungary. FEMS
2004;89(3):267–71. Immunol Med Microbiol 2004;42(1):119–24.
[14] Alm B, Milerad J, Wennergren G, Skjaerven R, Oyen N, Norvenius [33] Aaby P, Samb B, Simondon F, Seck AM, Knudsen K, Whittle H. Non-
G, et al. A case–control study of smoking and sudden infant death specific beneficial effect of measles immunisation: analysis of mortality
syndrome in the Scandinavian countries, 1992 to 1995. The nordic studies from developing countries. BMJ 1995;311(7003):481–5.
epidemiological SIDS study. Arch Dis Child 1998;78(4):329–34. [34] Otto S, Mahner B, Kadow I, Beck JF, Wiersbitzky SK, Bruns R. General
[15] Tappin D, Ecob R, Brooke H. Bedsharing, roomsharing, and sudden non-specific morbidity is reduced after vaccination within the third
infant death syndrome in Scotland: a case–control study. J Pediatr month of life—the Greifswald study. J Infect 2000;41(2):172–5.
2005;147(1):32–7. [35] Virtanen M, Peltola H, Paunio M, Heinonen OP. Day-to-day reacto-
[16] Hauck FR, Herman SM, Donovan M, Iyasu S, Merrick MC, Donoghue genicity and the healthy vaccinee effect of measles–mumps–rubella
E, et al. Sleep environment and the risk of sudden infant death syndrome vaccination. Pediatrics 2000;106(5):E62.
in an urban population: the Chicago infant mortality study. Pediatrics [36] Samad L, Tate AR, Dezateux C, Peckham C, Butler N, Bedford H.
2003;111(5 Part 2):1207–14. Differences in risk factors for partial and no immunisation in the first
[17] Heininger U, Kleemann WJ, Cherry JD. A controlled study of the year of life: prospective cohort study. BMJ 2006;332(7553):1312–3.
relationship between Bordetella pertussis infections and sudden unex- [37] Hambidge SJ, Phibbs SL, Davidson AJ, Lebaron CW, Chandramouli V,
pected deaths among German infants. Pediatrics 2004;114(1):e9–15. Fairclough DL, et al. Individually significant risk factors do not provide
[18] Flahault A, Messiah A, Jougla E, Bouvet E, Perin J, Hatton F. Sud- an accurate clinical prediction rule for infant underimmunization in one
den infant death syndrome and diphtheria/tetanus toxoid/pertussis/ disadvantaged urban area. Ambul Pediatr 2006;6(3):165–72.
poliomyelitis immunisation. Lancet 1988;1(8585):582–3. [38] Fine PE, Chen RT. Confounding in studies of adverse reactions to
[19] Hoffman HJ, Hunter JC, Damus K, Pakter J, Peterson DR, van Belle vaccines. Am J Epidemiol 1992;136(2):121–35.
G, et al. Diphtheria–tetanus–pertussis immunization and sudden infant [39] Ehreth J. The value of vaccination: a global perspective. Vaccine
death: results of the National Institute of Child Health and Human 2003;21(27–30):4105–17.