ARTICLES

Hepatitis B vaccine and the risk of CNS

inflammatory demyelination in childhood

Yann Mikaeloff, MD, ABSTRACT

PhD Background: The risk of CNS inflammatory demyelination associated with hepatitis B (HB) vaccine
Guillaume Caridade, MSc is debated, with studies reporting conflicting findings.
Samy Suissa, PhD
Methods: We conducted a population-based case-control study where the cases were children
Marc Tardieu, MD, PhD
with a first episode of acute CNS inflammatory demyelination in France (1994 –2003). Each case
was matched on age, sex, and geographic location to up to 12 controls, randomly selected from
the general population. Information on vaccinations was confirmed by a copy of the vaccination
Address correspondence and certificate. The odds ratios (ORs) of CNS inflammatory demyelination associated with HB vacci-
reprint requests to Dr. Yann nation were estimated using conditional logistic regression.
Mikaeloff, Service de Neurologie
Pédiatrique, CHU Bicêtre, Results: The rates of HB vaccination in the 3 years before the index date were 24.4% for the 349
Assistance Publique-Hôpitaux de cases and 27.3% for their 2,941 matched controls. HB vaccination within this period was not
Paris, 78 Avenue du Général
Leclerc, 94275 Le Kremlin- associated with an increase in the rate of CNS inflammatory demyelination (adjusted OR, 0.74;
Bicêtre Cedex, France 0.54 –1.02), neither ⬎3 years nor as a function of the number of injections or brand type. When
yann.mikaeloff@bct.aphp.fr
the analysis was restricted to subjects compliant with vaccination, HB vaccine exposure ⬎3 years
before index date was associated with an increased trend (1.50; 0.93–2.43), essentially from the
Engerix B vaccine (1.74; 1.03–2.95). The OR was particularly elevated for this brand in patients
with confirmed multiple sclerosis (2.77; 1.23– 6.24).
Conclusions: Hepatitis B vaccination does not generally increase the risk of CNS inflammatory demy-
elination in childhood. However, the Engerix B vaccine appears to increase this risk, particularly for
confirmed multiple sclerosis, in the longer term. Our results require confirmation in future studies.
Neurology® 2009;72:873–880

GLOSSARY
ADEM ⫽ acute disseminated encephalomyelitis; BCG ⫽ Bacille Calmette–Guèrin; HB ⫽ hepatitis B; MS ⫽ multiple sclerosis;
OR ⫽ odds ratio.

The first reports of an association between recombinant hepatitis B (HB) vaccine exposure and
cases of CNS inflammatory demyelination involved not only multiple sclerosis (MS) but also
clinically isolated syndromes such as acute disseminated encephalomyelitis (ADEM) or trans-
verse myelitis.1 The later conditions are in fact considered to be more likely than MS a conse-
quence of acute exposure to viral or vaccine antigen stimulation.2-4 However, most
epidemiologic studies were performed in adult patients and evaluated the association between
HB vaccine and an increase in confirmed MS. They demonstrated for most of them no increase
in risk after immunization but methodologic limitations were raised by some authors.5-9 Only
one database nested case-control study reported a significant increase in the risk of adult-onset
MS within the first 3 years of vaccination.10 In a recent population-based case-control study,

Editorial, page 870

e-Pub ahead of print on October 8, 2008, at www.neurology.org.
From Assistance Publique-Hôpitaux de Paris, Service de Neurologie Pédiatrique and Centre de Référence National des Maladies Inflammatoires du
Cerveau de l’Enfant (Y.M., G.C., M.T.), INSERM U822 (Y.M., G.C.), and INSERM U802 (M.T.), Hôpital Bicêtre, Université Paris Sud 11, Le
Kremlin Bicêtre, France; and Division of Clinical Epidemiology (S.S.), McGill University and Royal Victoria Hospital, Montreal, Canada.
Funded by grants from the “Agence Française de Sécurité Sanitaire des Produits de Santé” (AFSSAPS), the “Agence Nationale de Recherche sur le
SIDA et les hépatites” (ANRS), and the French Ministry of Health by a grant from the “Projet Hospitalier de Recherche Clinique” and by the
“Direction Générale de la Santé.” S. Suissa holds a Distinguished Scientist award from the Canadian Institutes of Health Research (CIHR).
Disclosure: The authors report no disclosures.

Copyright © 2009 by AAN Enterprises, Inc. 873

 we found no increase in the risk of incident Various exposure time windows were used. Our primary hy-
pothesis was the time window of effect within or more than the
MS, even within prolonged risk periods after
last 3 years before index date. The study was determined to have
vaccination.11 However, we observed a trend sufficient power to detect an OR of 1.5, with a minimal number
toward an increased risk over the long term of four controls per case, considering a frequency of exposure of
and a slight difference between HB vaccine 20% for 3 years in the controls. In addition, we studied ever use
(defined by the entire exposure period spanned from birth to
brand types. index date), the number of doses, and the major brand type of
We performed a population-based case- vaccine (last injection before index date).
control study to investigate whether exposure Different sensitivity analyses were also realized. Analyses
to the HB vaccine in childhood increases the were restricted to cases without a family history of MS or another
autoimmune disease, or with a low socioprofessional status of
risk of a first ever episode of acute CNS in- head of family, or with an age at index date ⱖ3 years (to elimi-
flammatory demyelination. nate possible biases from younger children not exposed for the
full 3-year exposure period), or with an age at index date ⱖ10
METHODS This study is a part of a research program evalu- years, or belonging to one or the other group of severity among
ating the impact of environmental risk factors on acute CNS MS cases. We also considered the year period just before the
inflammatory demyelinations in childhood. A first study assessed index date as an unexposed period to assess the theoretical possi-
the risk of childhood-onset MS and found no association with bility of a prescription bias in cases related to unrecognized
HB vaccine exposure.11 The present study addresses the risk symptoms during this period and performed another analysis
of a first ever episode of acute CNS inflammatory demyelina- excluding the controls recruited from the same household.
tion in childhood, whatever the course of the disease after the To control for a potential selection bias from subjects who
first attack. agreed to participate in the study, we performed an analysis re-
stricted to cases and controls compliant with vaccinations guide-
Cases definition and controls selection. The case series
lines. We considered the distribution of vaccine coverage in
for this study was selected from patients enrolled in the French
children in France in the study period to develop the defini-
neuropediatric “KIDSEP” cohort.12-15 Cases were all patients
tion.17 Compliant subjects were defined as receiving at least one
with a first episode of acute CNS inflammatory demyelination
Bacille Calmette–Guèrin (BCG) vaccine, one mumps-rubella-
occurring between January 1, 1994, and December 31, 2003,
measles vaccine, and four diphtheria-tetanus-poliomyelitis vac-
before the age of 16 years, and born in France. Thus, the case
cine during the first 2 years of life. BCG vaccine is an obligation
definition includes patients with single episodes without relapse
to be accepted in community care structures (day-nursery, nurs-
during the follow-up, as well as patients who went on to relapse
ery school) in France but could also be realized voluntarily.
and were diagnosed with MS. The last group of patients was
Mumps-rubella-measles is recommended by Health Authorities
included in the previous study with a similar design that tested
but is not an obligation. Diphtheria-tetanus-poliomyelitis vac-
specifically confirmed MS.11
cine is an obligation (three injections at 2, 3, and 4 months of
The procedure of population-based selection of controls was
age and a booster before 18 months of age). Other less restricted
previously described in detail.11 In short, we aimed to individu-
definitions of compliance were also used: 1) four diphtheria-
ally match each case with up to 12 controls selected from the
tetanus-poliomyelitis vaccine and one mumps-rubella-measles
French general population on the basis of age (⫾6 months), sex,
during the first 2 years of life, and 2) one mumps-rubella-measles
and current area of residence. Controls were selected by random
during the first 2 years of life. These sensitivity analyses were not
telephone dialing in the geographic area of residence of each
performed in our previous study11 and, accordingly, we also re-
case. Households including an eligible subject were sent an infor-
analyzed cases with a relapsing course (confirmed MS).
mation letter and questionnaire after consent. Each matched
The study was approved by the Comité National Informa-
control was assigned the index date of the case (the date of onset
tique et Liberté (the French data protection agency). All patients
of the symptoms of the first episode) for the evaluation of previ-
gave informed consent for their inclusion in the study. The study
ous vaccinations.
was supported by the Société Française de Neuropédiatrie and
Data collection and exposure confirmation. An informa- the French Ministry of Health (Direction Générale de la Santé).
tion letter and a questionnaire were sent to all cases and controls, The study was overseen by a scientific committee and advisory
as previously described.11 In summary, a copy of the child’s vac- board composed of independent experts who approved the pro-
cination certificate (carnet de santé), that includes all recordings tocol, conduct, analysis, interpretation, and publication of the
of vaccinations and particularly for HB, was requested as well as study.
information on familial autoimmune history (in siblings or par-
ents) and parental smoking at home before index date. RESULTS The initial case series amounted to 403
Statistical analysis and ethical considerations. Condi- cases, of which 86.6% agreed to participate, includ-
tional logistic regression for matched case-control data was used ing 349 cases who provided a copy of their vaccina-
to estimate odds ratios (ORs) of first ever episodes of acute CNS tion certificate and who were retained for analysis.
inflammatory demyelination associated with HB vaccine expo- These were not significantly different baseline char-
sure. Subjects not exposed to HB vaccine, between birth and the
acteristics from the 54 patients who did not partici-
index date, were used as the reference group. In addition
pate (data not shown). Among the 349 cases
to the intrinsic confounder adjustment for the matching factors,
the regression model was used to perform further adjustment for included in the study, the diagnosis confirmed by the
family history of MS or of other autoimmune diseases, occupa- course of disease was 1) single episodes without re-
tion of the head of the family, and parental smoking at home.16 lapse for 198 cases including ADEM (n ⫽ 79,

874 Neurology 72 March 10, 2009

 Our analyses therefore included 349 cases of first
Table 1 Characteristics of cases with a first episode of acute CNS
inflammatory demyelination and matched population-based controls
episode of acute CNS inflammatory demyelination
and 2,941 matched controls (n ⫽ 207, 7.0) were
All case All matched recruited from the same household) (table 1). Vacci-
patients controls*
Variable (n ⴝ 349) (n ⴝ 2,941) nation information was provided in the form of a
Male 156 (44.7) 1,349 (45.9) copy of the vaccination records for 97.7% (n ⫽
Age, y, mean ⴞ SD 9.3 ⫾ 4.6 9.0 ⫾ 4.5 2,873) of controls; the others had a standardized tele-
Residence in Paris or suburbs 109 (31.2) 898 (30.5)
phone interview. However, all these data were used
Familial MS history 8† (2.3) 31 (1.1)
for analysis. The median number of controls per case
was 9 (mean: 8.4): 60 (17.2%) had 11 to 12 controls,
Other familial autoimmune disease history 18 (5.2) 210 (7.1)
232 (66.5%) had 7 to 10 controls, 50 (14.3%) had 4
Low socioprofessional status of head of family 166† (47.6) 1,606 (54.6)
(unemployed people, laborers, and low-income to 6 controls, and only 7 cases (2.0%) had 1 to 3
employees)
controls. The frequency of family heads with a low
Infection during the month before disease onset 149 (43) socioprofessional status was lower for cases than for
Symptoms at disease onset controls (table 1). The frequency of family history of
Multiple 219 (64) MS was higher in cases, but the frequency of familial
Transverse myelitis 49 (14) autoimmune diseases was similar in the two groups.
Optic neuritis 86 (25) The rates of any HB vaccination in the 3 years
Severe mental status change 121 (35) before the index date were 24.4% for the 349 cases
Brainstem dysfunction 127 (36)
and 27.3% for their 2,941 matched controls. The
CSF
adjusted OR of such episode occurrence associated
with any HB vaccination during a risk period of 3
Oligoclonal bands 89 (26)
years was 0.74 (95% CI: 0.54 –1.02). Other results
Cells >10/␮L 157 (45)
of analyses in all study subjects are presented in table
Proteins >0.5 g/dL 88 (25)
2. Similar results were obtained for more detailed risk
MRI
periods within 3 years, ever use, number of HB im-
Child-MS MRI criteria‡ 124 (36) munizations, and last brand type used. ORs were
Three Barkhof MRI criteria§ 124 (36) similarly not significantly different from 1 in analyses
restricted to cases without a family history of MS or
*Matched on age (⫾6 months), sex, geographic location (place of residence).
†p ⬍ 0.05 (␹2 test for comparison of proportions or t test for comparison of means). another autoimmune disease, or with a low sociopro-
‡Corpus callosum long axis perpendicular lesions or sole presence of well-defined lesions.13 fessional status of head of family, or with an age at
Three criteria among four: at least one gadolinium-enhancing T1 lesion or ⱖ9 T2 lesions, at index date ⱖ3 years, or with an age at index date
§

least one infratentorial T2 lesion, at least one juxtacortical T2 lesion, ⱖ3 periventricular

lesions.31
ⱖ10 years, or belonging to one or the other group of
severity among MS cases. Similar results were found
in considering the 1-year period just before the index
39.9%), isolated optic neuritis (n ⫽ 19, 9.6%), iso- date as an unexposed period or excluding the con-
lated transverse myelitis (n ⫽ 18, 9.1%), and brain- trols recruited from the same household.
stem dysfunction (n ⫽ 8, 4.0%), and other episodes When cases compliant with vaccination guide-
(n ⫽ 74, 37.4%); 2) MS (confirmed by at least one lines were compared with noncompliant according
other episode) for 151, including 90 (59.6%) with a to variables of table 1, there were no significant dif-
higher initial index of early severity.14 ferences except for location in Paris and suburbs for
A total of 371,996 households were contacted for cases with a CNS inflammatory demyelination
the recruitment of controls. Among the 5,838 eligi- (42.9% in compliant vs 21% in noncompliant) and
ble households contacted, 1,666 refused to partici- for the detection of oligoclonal bands in CSF (36.1%
pate in the study at first contact. We identified 4,172 in compliant vs 58.2% in noncompliant) for con-
eligible controls who accepted and could be matched firmed MS. There was no differential compliance be-
to the case on age, sex, and geographic location; tween cases and their matched controls because,
2,941 of them provided vaccination information (re- before exclusions of cases without controls and of
sponse rate 70.5%). The reasons for a lack of vacci- controls without cases required by the analysis
nation information for 1,231 recruited controls, method, 166/349 (47.6%) cases with a CNS inflam-
despite initial consent, were no response despite a full matory demyelination and 1,612/2,941 (54.8%) of
recall procedure (n ⫽ 807, 65.6%), refusal at first their controls were compliant. It was the same for
recall (n ⫽ 208, 16.9%), lost vaccination certificate cases with confirmed MS: before exclusions required
(n ⫽ 98, 8.0%), never reached by phone or incorrect by the analysis method, 74/151 (49.0%) cases and
telephone number or address given (n ⫽ 118, 9.5%). 615/1,192 (51.6%) of their controls were compliant.

Neurology 72 March 10, 2009 875

 Table 2 Crude and adjusted OR of acute CNS inflammatory demyelination associated with timing, number,
and brand of hepatitis B vaccine

Case patients Matched controls* Crude Adjusted OR

Hepatitis B (HB) vaccine exposure (n ⴝ 349), n (%) (n ⴝ 2,941), n (%) OR† (95% CI)‡

No HB vaccination 195 (55.9) 1,543 (52.5) 1 Reference

HB vaccination before index date

Ever use 154 (44.1) 1,398 (47.5) 0.80 0.81 (0.62–1.05)

0–1 y 30 (8.6) 317 (10.8) 0.64 0.65 (0.42–1.02)

>1–2 y 25 (7.2) 247 (8.4) 0.67 0.70 (0.43–1.16)

>2–3 y 30 (8.6) 239 (8.1) 0.93 0.94 (0.58–1.51)

>3 y 69 (19.8) 595 (20.2) 0.92 0.93 (0.65–1.31)

No. of HB immunizations before index date

1–2 23 (6.6) 187 (6.4) 0.79 0.80 (0.48–1.32)

>3 131 (37.5) 1,211 (41.2) 0.80 0.81 (0.62–1.06)

Last brand type used before index date

Engerix B ever use 84 (24.1) 724 (24.6) 0.85 0.86 (0.63–1.16)

Engerix B 0–3 y 45 (12.9) 409 (13.9) 0.75 0.76 (0.52–1.12)

Engerix B >3 y 39 (11.2) 315 (10.7) 0.98 0.98 (0.65–1.48)

GenHevac B ever use 47 (13.5) 415 (14.1) 0.77 0.78 (0.52–1.16)

GenHevac B 0–3 y 24 (6.9) 229 (7.8) 0.65 0.68 (0.40–1.14)

GenHevac B >3 y 23 (6.6) 186 (6.3) 0.92 0.91 (0.54–1.54)

§
Other HB vaccine ever use 23 (6.6) 259 (8.8) 0.70 0.71 (0.45–1.14)

Other HB vaccine 0–3 y 16 (4.6) 165 (5.6) 0.77 0.79 (0.46–1.38)

Other HB vaccine >3 y 7 (2.0) 94 (3.2) 0.58 0.59 (0.26–1.32)

*Matched on age (⫾6 months), sex, geographic location (current place of residence).
†No HB vaccine exposure is the reference group.
‡Adjusted on covariates: familial multiple sclerosis history, family history of another autoimmune disease, parental smoking
at home before index date, socioprofessional status of the head of the family.
§
HBVax, Hevac B, Twinrix, or Recombivax.

As shown in table 3, when the analysis was re- justed OR of 1.29 (0.82–2.03) for CNS inflamma-
stricted to compliant cases and their matched tory demyelination and 1.96 (1.00 –3.82) for
compliant controls, HB vaccine exposure ⬎3 years confirmed MS. GenHevac B was associated with
before index date was associated with an increased ORs of 1.48 (0.80 –2.71) for CNS inflammatory de-
trend for acute CNS inflammatory demyelination, myelination and 1.18 (0.46 –3.06) for confirmed
essentially from the Engerix B vaccine. The OR MS.
was particularly elevated for this major brand type
in patients with confirmed MS (table 4). Similar DISCUSSION This study is the second part of a
results were found with other less restricted definition of research program evaluating the impact of vaccina-
compliance to vaccination (four diphtheria-tetanus- tions on acute CNS inflammatory demyelination in
poliomyelitis vaccine and one mumps-rubella- childhood. We evaluate the risk associated with HB
measles during the first 2 years of life). Engerix B vaccine exposure of any first episode of acute CNS
exposure ⬎3 years before index date was associated inflammatory demyelination in childhood and re-
with an adjusted OR of 1.50 (0.92–2.45) for CNS port of a lack of increase in risk within a risk period
inflammatory demyelination and 2.40 (1.16 – 4.94) of 3 years or more or considering the number of in-
for confirmed MS. GenHevac B was associated with jections. However, in the subgroup analysis restricted
ORs of 1.36 (0.69 –2.70) for CNS inflammatory de- to subjects compliant with vaccinations guidelines,
myelination and 1.21 (0.43–3.44) for confirmed Engerix B exposure ⬎3 years before index date was
MS. Similar results were also found with a third def- associated with an increased risk that was significant
inition of compliance to vaccination (recommended in patients who went on to develop MS. Our new
but not obligated: one mumps-rubella-measles dur- results suggesting a possible association of HB vac-
ing the first 2 years of life). Engerix B exposure ⬎3 cine with pediatric MS (albeit HB vaccine of a spe-
years before index date was associated with an ad- cific manufacturer and only in vaccine-compliant

876 Neurology 72 March 10, 2009

 Table 3 Crude and adjusted OR of acute CNS inflammatory demyelination associated with timing, number,
and brand of hepatitis B vaccine, restricted to subjects compliant with vaccination*

Hepatitis B (HB) Case patients Matched controls† Crude Adjusted OR

vaccine exposure (n ⴝ 163) , n (%) (n ⴝ 892) , n (%) OR‡ (95% CI)§

Unvaccinated 78 (47.9) 435 (48.8) 1 Reference

Vaccinated against HB before index date

Ever use 85 (52.1) 457 (51.2) 0.96 1.03 (0.70–1.51)

0–1 y 13 (8.0) 78 (8.7) 0.78 0.81 (0.40–1.63)

>1–2 y 11 (6.7) 87 (9.8) 0.42 0.45 (0.20–1.01)

>2–3 y 14 (8.6) 79 (8.9) 0.82 0.89 (0.43–1.84)

>3 y 47 (28.8) 213 (23.8) 1.40 1.50 (0.93–2.43)

No. of HB immunizations before index date

1–2 11 (6.7) 54 (6.1) 0.93 0.99 (0.46–2.15)

>3 74 (45.4) 403 (45.2) 0.96 1.03 (0.70–1.53)

Last brand type used for vaccination before index date

Engerix B ever use 51 (31.3) 236 (26.5) 1.11 1.17 (0.76–1.79)

Engerix B 0–3 y 20 (12.3) 121 (13.6) 0.68 0.71 (0.39–1.29)

Engerix B >3 y 31 (19.0) 115 (12.9) 1.63 1.74 (1.03–2.95)

GenHevac B ever use 22 (13.5) 121 (13.6) 0.91 1.03 (0.56–1.89)

GenHevac B 0–3 y 10 (6.1) 63 (7.1) 0.61 0.68 (0.28–1.65)

GenHevac B >3 y 12 (7.4) 58 (6.5) 1.34 1.50 (0.71–3.17)

Other HB vaccine ever use
12 (7.4) 100 (11.2) 0.65 0.70 (0.35–1.41)

Other HB vaccine 0–3 y 8 (4.9) 60 (6.7) 0.69 0.75 (0.32–1.77)

Other HB vaccine >3 y 2 (2.5) 40 (4.5) 0.59 0.63 (0.20–2.00)

*Defined as at least one BCG vaccine, one mumps-rubella-measles vaccine, and four diphtheria-tetanus-poliomyelitis vac-
cine during the first 2 years of life.
†Matched on age (⫾6 months), sex, geographic location (current place of residence).
‡No HB vaccine exposure is the reference group.
§
Adjusted on covariates: familial multiple sclerosis history, family history of another autoimmune disease, parental smoking
at home before index date, socioprofessional status of the head of the family.


HBVax, Hevac B, Twinrix, or Recombivax.

cases) are not in contradiction with prior results. In- database (integrated pharmacy and medical claims
deed, in this new study, we enlarged the analysis to from six health maintenance organizations plans)
all first episodes of CNS inflammatory demyelina- from 1988 to 1995, authors found no association
tion and tried to address the potential selection bias between CNS inflammatory demyelination and HB
from the partial answer of subjects who could partic- vaccine (comparison to nonvaccinated individuals
ipate, using different definitions of compliance to matched on age and sex) in the 3 years after vaccina-
vaccination guidelines. tion and over.27 However, the accuracy of validation
The first case reports in adults of a possible associ- of MS status was debated.9
ation between recombinant HB vaccine exposure Two case-control studies in adults focused on re-
and CNS inflammatory demyelination involved combinant HB vaccine with a systematic validation
any first episodes of acute CNS inflammatory procedure for case status and ascertainment of infor-
demyelination.1,18-25 A French case-control study in- mation on vaccinations from computerized records.
cluded 236 adult cases (193 with an initial episode of The first, in three American health maintenance or-
definite or probable MS according to follow-up data) ganizations, involved 108 cases of isolated optic neu-
and 355 matched controls recruited in hospital.26 ritis and 332 of MS, matched to 950 controls, and
The adjusted ORs for CNS inflammatory demyeli- found no increase in risk of both diseases combined
nation within a 2-month risk period was 1.8 (95% or analyzed separately, associated with HB vaccine in
CI: 0.7– 4.6) in the whole group and 1.4 (95% CI: the short or the long term (⬍1 year, 1–5 years, ⬎5
0.4 – 4.5) in the subgroup of subjects referring to vac- years, and within 3 years after HB vaccination).6,7
cination certificates during the phone interview. In a The second, a large database nested case-control
retrospective cohort study on 134,698 individuals study with patients issued from the General Practice
(children and adults) enrolled in the US healthcare Research Database in UK, involved 163 cases of MS,

Neurology 72 March 10, 2009 877

 Table 4 Crude and adjusted OR of confirmed multiple sclerosis associated with timing, number, and brand of
hepatitis B vaccine, restricted to subjects compliant with vaccination*

Hepatitis B (HB) Case patients Matched controls† Crude Adjusted OR

vaccine exposure (n ⴝ 72), n (%) (n ⴝ 347), n (%) OR‡ (95% CI)§

Unvaccinated 27 (37.5) 153 (44.1) 1 Reference

Vaccinated against HB before index date

Ever use 45 (62.5) 194 (55.9) 1.26 1.35 (0.72–2.56)

0–1 y 6 (8.3) 29 (8.4) 0.98 0.88 (0.28–2.79)

>1–2 y 6 (8.3) 41 (11.7) 0.42 0.45 (0.12–1.71)

>2–3 y 8 (11.1) 36 (10.4) 1.10 1.20 (0.38–3.74)

>3 y 25 (34.6) 88 (25.4) 1.86 2.12 (1.00–4.48)

No. of HB immunizations before index date

1–2 5 (6.9) 27 (7.8) 0.78 0.75 (0.21–2.61)

>3 40 (55.6) 167 (48.1) 1.34 1.48 (0.78–2.82)

Last brand type used for vaccination before index date

Engerix B ever use 30 (41.7) 91 (26.2) 1.71 1.92 (0.96–3.85)

Engerix B 0–3 y 11 (15.3) 43 (12.4) 1.05 1.11 (0.43–2.90)

Engerix B >3 y 19 (26.4) 48 (13.8) 2.38 2.77 (1.23–6.24)

GenHevac B ever use 11 (15.3) 68 (19.6) 0.82 0.96 (0.38–2.46)

GenHevac B 0–3 y 6 (8.3) 35 (10.1) 0.69 0.82 (0.22–3.00)

GenHevac B >3 y 5 (6.8) 33 (9.5) 0.98 1.13 (0.36–3.55)

Other HB vaccine ever use
4 (5.6) 35 (10.1) 0.60 0.55 (0.14–2.07)

Other HB vaccine 0–3 y 3 (4.2) 28 (8.1) 0.51 0.45 (0.09–2.21)

Other HB vaccine >3 y 1 (1.4) 7 (2.0) 0.90 0.88 (0.07–10.54)

*Defined as at least one BCG vaccine, one mumps-rubella-measles vaccine, and four diphtheria-tetanus-poliomyelitis vac-
cine during the first 2 years of life.
†Matched on age (⫾6 months), sex, geographic location (current place of residence).
‡No HB vaccine exposure is the reference group.
§
Adjusted on covariates: familial multiple sclerosis history, family history of another autoimmune disease, parental smoking
at home before index date, socioprofessional status of the head of the family.


HBVax, Hevac B, Twinrix, or Recombivax.

matched to 1,604 controls, and reported a significant studies on vaccine safety and the level of participa-
increase in the risk of adult-onset MS within the first tion we obtained was satisfying compared to previous
3 years of vaccination for adult MS (OR 3.1; 95% population-based case-control field studies, as dis-
CI: 1.5– 6.3).10 The brand types used were not spec- cussed previously.11,16,28,29 To address this bias, we
ified in all these studies. restricted the analysis to subjects who were compliant
The strengths of our study include the accurate with vaccinations guidelines. We deemed that the ac-
ascertainment of the first episode of acute CNS in- tual response rate to our requests among compliant
flammatory demyelination, the validation of the vac- subjects would be higher than in less compliant sub-
cination status by a copy of the vaccination jects. The rates of exposure to vaccines in controls
certificates, and the inclusion of the vast majority of were consistent with that previously described for the
incident pediatric cases in France.11 Another was the general childhood population in France, considering
population-based nature of the selection of controls the study period.17 In compliant subjects, the OR
by random sampling for controls among the French was particularly elevated for Engerix B vaccine, the
population, with matching on age, sex, and geo- major brand type, in patients with confirmed MS.
graphic area of living of the corresponding case. It This analysis was not performed in our previous
was tempered by the partial response rate of controls. study in patients with MS that already indicated a
Indeed, 1,666 eligible households contacted refused nonsignificant trend toward a higher risk in subjects
to participate in the study. Although these controls having received Engerix B ⬎3 years before the index
were replaced by others who were also individually date. Considering the relatively small number of ex-
matched with cases and agreed to participate at initial posed subjects, the results concerning GenHevac B
contact, this could have introduced a selection bias. vaccine should be considered inconclusive. A clini-
However, it is a constant limitation of case-control cally relevant risk associated with this vaccine cannot

878 Neurology 72 March 10, 2009

be ruled out by this study. Our significant results REFERENCES
concerning Engerix B vaccine were obtained from 1. Tourbah A, Gout O, Liblau R, et al. Encephalitis after
hepatitis B vaccination: recurrent disseminated encephali-
subgroup analyses, and were thus subject to false sig-
tis or MS? Neurology 1999;53:396–401.
nificance from multiple comparisons. Nevertheless,
2. Menge T, Hemmer B, Nessler S, et al. Acute dissemi-
the risk magnitude we observed in patients with MS nated encephalomyelitis: an update. Arch Neurol 2005;
for Engerix B vaccine exposure is close to the one 62:1673–1680.
using the General Practice Research Database in 3. Piaggio E, Ben Younes A, Desbois S, et al. Hepatitis B
United Kingdom, for exposure to any HB vaccine vaccination and central nervous system demyelination: an
within 3 years before index date.10 However, this lat- immunological approach. J Autoimmun 2005;24:33–37.
ter study did not present analysis according to the 4. Schattner A. Consequence or coincidence? The occur-
brand. Further studies would be needed to assess the rence, pathogenesis and significance of autoimmune
manifestations after viral vaccines. Vaccine 2005;23:
brand influence among the UK population of adults
3876–3886.
in this study period and to confirm our results in
5. Ascherio A, Zhang SM, Hernan MA, et al. Hepatitis B
children. Differences between children and adults in vaccination and the risk of multiple sclerosis. N Engl
susceptibility to immune stimulation must be con- J Med 2001;344:327–332.
sidered to interpret the results.3,30 Two possible ex- 6. DeStefano F, Verstraeten T, Jackson LA, et al. Vaccina-
planations for the differences between brands could tions and risk of central nervous system demyelinating dis-
be as follows: 1) each vaccine uses a different section eases in adults. Arch Neurol 2003;60:504–509.
of the HBs antigen and some protein fragments pro- 7. DeStefano F, Weintraub ES, Chen RT. Hepatitis B vac-
duced by yeasts may induce molecular mimicry while cine and risk of multiple sclerosis. Neurology 2005;64:
1317.
others do not; 2) the production process varies by brand
8. DeStefano F, Weintraub ES, Chen RT. Recombinant
and differences in yeast protein content may be crucial hepatitis B vaccine and the risk of multiple sclerosis: a
if yeast protein may trigger autoimmune reactions: it is prospective study. Pharmacoepidemiol Drug Saf 2007;
stated in the description of Engerix B manufacturing pro- 16:705–708.
cess that the vaccine has no more than 5% of yeast pro- 9. Hernan MA, Jick SS. Hepatitis B vaccination and multiple
teins, whereas it is no more than 1% for GenHevac B. sclerosis: the jury is still out. Pharmacoepidemiol Drug Saf
2006;15:653–655.
10. Hernan MA, Jick SS, Olek MJ, Jick H. Recombinant hep-
ACKNOWLEDGMENT
Professional organization specializing in population-based opinion polls
atitis B vaccine and the risk of multiple sclerosis: a prospec-
and case-control studies on health topics responsible for the population- tive study. Neurology 2004;63:838–842.
based selection of controls: CSA Institute (Paris, France). The authors 11. Mikaeloff Y, Caridade G, Rossier M, Suissa S, Tardieu M. Hep-
thank Mélanie Rossier and Chittima Tilavanh for assistance in the collec- atitis B vaccination and the risk of childhood onset multiple scle-
tion of questionnaires. rosis. Arch Pediatr Adolesc Med 2007;161:1176–1182.
12. Mikaeloff Y, Suissa S, Vallee L, et al. First episode of acute
APPENDIX CNS inflammatory demyelination in childhood: prognos-
Members of the Scientific Committee and Advisory Board: Dr. Annick tic factors for multiple sclerosis and disability. J Pediatr
Alperovitch (Epidemiologist, Paris, France), Prof. Jacques Benichou (Bio- 2004;144:246–252.
statistician, INSERM U 657, Rouen, France), Prof. Gérard Bréart (Epide- 13. Mikaeloff Y, Adamsbaum C, Husson B, et al. MRI prog-
miologist, INSERM U149, Paris, France), Prof. Bertrand Fontaine
nostic factors for relapse after acute CNS inflammatory
(Neurologist, Pitié-Salpétrière Hospital, Paris, France), Prof. Edeltraut
demyelination in childhood. Brain 2004;127:1942–1947.
Garbe (Pharmacologist/epidemiologist, Berlin, Germany), Prof. Miguel
Hernan (Epidemiologist, Boston, MA).
14. Mikaeloff Y, Caridade G, Assi S, et al. Prognostic factors
Participants of the KIDSEP-VAC study group: Pediatric neurology de- and early severity score in a cohort of childhood multiple
partments (France): D. Amsallem (Besançon); P. Aubourg (Hôpital Saint- sclerosis. Pediatrics 2006;118:1133–1139.
Vincent de Paul, Paris); M.A. Barthez, P. Castelnau (Tours); P. Berquin 15. Mikaeloff Y, Caridade G, Assi S, et al. Hepatitis B vaccine
(Amiens); O. Boespflug (Clermont-Ferrand); J.C. Carrière, C. Cances, Y. and risk of relapse after first childhood CNS inflammatory
Chaix (Toulouse); J.M. Cuisset, J.C. Cuvelier, L. Vallée (Lille); A. de
demyelination. Brain 2007;130:1105–1110.
Saint-Martin (Strasbourg); I. Desguerre, O. Dulac (Necker-Enfants
16. Mikaeloff Y, Caridade G, Tardieu M, Suissa S. Parental
Malades, Paris); V. Des Portes, C. Rousselle, D. Ville (Lyon); B. Echenne,
F. Rivier (Montpellier); P. Evrard (Hôpital Robert Debré, Paris); A. Joan-
smoking at home and the risk of childhood-onset multiple
nard (Grenoble); P. Landrieu, Y. Mikaeloff, M. Tardieu (Bicêtre); M.O. sclerosis in children. Brain 2007;130:2589–2595.
Livet (Aix en Provence); J. Mancini, B. Chabrol (Marseilles); J. Motte, N. 17. Antona D, Bussière E, Guignon N, Badeyan G, Lévy-
Bednarek, P. Sabouraud (Reims); S. Napuri, L. Lazaro (Rennes); S. Bruhl D. La couverture vaccinale en France en 2001. Bull
Nguyen (Nantes); S. Peudenier (Brest); F. Pouplard (Angers); J.M. Pedes- Epidemiol Hebd 2003;36:169–172.
pan (Bordeaux); S. Perelman, C. Richelme (Nice); N. Perez (Belfort); D. 18. Scott TF. Post infectious and vaccinal encephalitis. Med
Rodriguez, T. Billette de Villemeur, M.L. Moutard, G. Ponsot (Hôpital
Clin North Am 1967;51:701–717.
Trousseau, Paris); M.C. Routon (Orsay); H. Testard (Annemasse), C.
Van Hulle (Rouen). Clinical epidemiology department (Canada): A.
19. Sriram S, Steinman L. Post infectious and post vaccinal
Kezouh, S. Suissa (McGill University and Royal Victoria Hospital, encephalomyelitis. Neurol Clin 1984;2:341–353.
Montreal). 20. Mahassin F, Algayres JP, Valmary J, et al. [Acute myelitis after
vaccination against hepatitis B.] Presse Med 1993;22:1997–
Received March 21, 2008. Accepted in final form August 6, 2008. 1998.

Neurology 72 March 10, 2009 879

 21. Trevisani F, Gattinara GC, Caraceni P, et al. Transverse event: a case-control study. Neuroepidemiology 2002;21:
myelitis following hepatitis B vaccination. J Hepatol 1993; 180–186.
19:317–318. 27. Zipp F, Weil JG, Einhaupl KM. No increase in demyeli-
22. Kaplanski G, Retornaz F, Durand J, Soubeyrand J. Central nating diseases after hepatitis B vaccination. Nat Med
nervous system demyelination after vaccination against 1999;5:964–965.
hepatitis B and HLA haplotype. J Neurol Neurosurg Psy- 28. Rodrigues LC, Smith PG. Use of the case-control ap-
chiatry 1995;58:758–759. proach in vaccine evaluation: efficacy and adverse effects.
23. Berkman N, Benzarti T, Dhaoui R, Mouly P. [Bilateral Epidemiol Rev 1999;21:56–72.
neuro-papillitis after hepatitis B vaccination.] Presse Med 29. Chen RT. Special Methodological Issues in Pharmacoepi-
1996;25:1301. demiology Studies of Vaccine Safety. In: Strom BL, ed.
24. Senejoux A, Roulot D, Belin C, Tsakiris L, Rautureau J, Pharmacoepidemiology (third edition). New York: Wiley;
Coste T. [Acute myelitis after immunization against hepa- 2000:707–732.
titis B with recombinant vaccine.] Gastroenterol Clin Biol 30. Banwell B, Ghezzi A, Bar-Or A, Mikaeloff Y, Tardieu M.
1996;20:401–402. Multiple sclerosis in children: clinical diagnosis, therapeu-
25. Gout O, Theodorou I, Liblau R, Lyon-Caen O. Central tic strategies, and future directions. Lancet Neurol 2007;6:
nervous system demyelination after recombinant hepatitis 887–902.
B vaccination: report of 25 cases. Neurology 1997;48: 31. McDonald WI, Compston A, Edan G, et al. Recom-
A424. Abstract. mended diagnostic criteria for multiple sclerosis: guide-
26. Touze E, Fourrier A, Rue-Fenouche C, et al. Hepatitis B lines from the international panel on the diagnosis of
vaccination and first central nervous system demyelinating multiple sclerosis. Ann Neurol 2001;50:121–127.

880 Neurology 72 March 10, 2009

 Hepatitis B vaccine and the risk of CNS inflammatory demyelination in childhood
Yann Mikaeloff, Guillaume Caridade, Samy Suissa, et al.
Neurology 2009;72;873-880 Published Online before print October 8, 2008
DOI 10.1212/01.wnl.0000335762.42177.07

This information is current as of October 8, 2008

Updated Information & including high resolution figures, can be found at:
Services http://www.neurology.org/content/72/10/873.full.html

Supplementary Material Supplementary material can be found at:

http://www.neurology.org/content/suppl/2010/09/12/01.wnl.000033576
2.42177.07.DC1.html
References This article cites 29 articles, 9 of which you can access for free at:
http://www.neurology.org/content/72/10/873.full.html##ref-list-1
Citations This article has been cited by 7 HighWire-hosted articles:
http://www.neurology.org/content/72/10/873.full.html##otherarticles
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
Acute disseminated encephalomyelitis
http://www.neurology.org//cgi/collection/acute_disseminated_encephal
omyelitis
All Pediatric
http://www.neurology.org//cgi/collection/all_pediatric
Case control studies
http://www.neurology.org//cgi/collection/case_control_studies
Multiple sclerosis
http://www.neurology.org//cgi/collection/multiple_sclerosis
Permissions & Licensing Information about reproducing this article in parts (figures,tables) or in
its entirety can be found online at:
http://www.neurology.org/misc/about.xhtml#permissions
Reprints Information about ordering reprints can be found online:
http://www.neurology.org/misc/addir.xhtml#reprintsus

Neurology ® is the official journal of the American Academy of Neurology. Published continuously since
1951, it is now a weekly with 48 issues per year. Copyright . All rights reserved. Print ISSN: 0028-3878.
Online ISSN: 1526-632X.