American Journal of Epidemiology
Copyright ª 2006 by the Johns Hopkins Bloomberg School of Public Health
All rights reserved; printed in U.S.A.
Vol. 163, No. 12
DOI: 10.1093/aje/kwj148
Advance Access publication May 3, 2006
Original Contribution
Sibship Characteristics and Risk of Multiple Sclerosis: A Nationwide Cohort
Study in Denmark
1
Department of Epidemiology Research, Danish Epidemiology Science Centre, Statens Serum Institut, Copenhagen, Denmark.
Department of Neurology, Aarhus University Hospital in Aalborg, Aalborg, Denmark.
3
National Institute of Public Health, Copenhagen, Denmark.
2
Received for publication April 5, 2005; accepted for publication January 6, 2006.
It has been hypothesized that age at infection with a common microbial agent may be associated with the risk of
multiple sclerosis (MS). The authors addressed this hypothesis by using number of older siblings and other sibship
characteristics as an approximation of age at exposure to common infections. Data on family characteristics and
vital status from the Danish Civil Registration System were used to establish a cohort of all Danes whose mothers
had been born in Denmark since 1935. Persons diagnosed with MS during the period 1968–1998 were identified
through linkage with the Danish Multiple Sclerosis Register. The cohort of 1.9 million Danes was followed for 28.1
million person-years; during that time, 1,036 persons developed MS. Overall, there was no association between
number of older siblings, number of younger siblings, total number of siblings, age distance from the nearest younger sibling, or exposure to younger siblings under 2 years of age and risk of MS later in life. There was no association of MS risk with multiple birth (vs. singleton birth) or with the age of the mother or father at birth. These results
do not lend support to the hypothesis that number of older siblings or any of the other sibship characteristics studied
is associated with risk of MS.
birth order; infection; maternal age; multiple sclerosis; paternal age; risk factors; siblings; twins
Abbreviations: CI, confidence interval; MS, multiple sclerosis; RR, rate ratio.
Multiple sclerosis (MS) is believed to be an immunologic
disorder caused by both genetic and environmental factors,
possibly including viral infections (1, 2). In particular, several epidemiologic observations lend support to the hypothesis that MS could result from an aberrant immune response,
possibly triggered by an infection acquired late in childhood
or during adolescence (3, 4). Accordingly, the association
between risk of MS and age at incurring one of several
suggested candidate infections has been studied (5, 6). However, these studies have been carried out in small populations and with the possibility that recall of age at infection
decades later could have biased the results. This may be
particularly true for reinfections. In addition, subclinical
infection cannot be studied in this way.
Another approach taken to address the hypothesis has
been to approximate exposure to infections using sibship
characteristics. Accordingly, number of siblings is believed
to be associated with the risk of exposure to common infections, and birth order is believed to be inversely associated with the age at which such infections occur (7–12).
According to the proposed hypothesis, children with few
siblings and/or of early birth orders would be more likely
to develop MS. For allergic rhinitis and atopy, the existence
of an inverse association with birth order has been proven by
Correspondence to Dr. Peter Bager, Department of Epidemiology Research, Statens Serum Institute, Artillerivej 5, DK-2300 Copenhagen,
Denmark (e-mail: pbg@ssi.dk).
1112
Am J Epidemiol 2006;163:1112–1117
Downloaded from https://academic.oup.com/aje/article-abstract/163/12/1112/97117 by guest on 03 June 2020
Peter Bager1, Nete Munk Nielsen1, Kristine Bihrmann1, Morten Frisch1, Jan Wohlfart1, Nils
Koch-Henriksen2,3, Mads Melbye1, and Tine Westergaard1
Sibship Characteristics and Multiple Sclerosis Risk
MATERIALS AND METHODS
Data from the Civil Registration System were used to
generate a complete sibship database for persons born to
Danish women. Since April 1, 1968, all residents of Denmark have been recorded in the Civil Registration System
and assigned a unique personal identification number. Individual information is kept under this identification number
in all national registers, ensuring precise and secure linkage
of person-identifiable information between registers. The
Civil Registration System also includes person-identifiable
information on date of birth, sex, place of birth, vital status,
and, for most persons born since the beginning of the 1950s,
information on parents and offspring (34). We established
a population-based sibship database by extracting data on all
women born in Denmark since 1935 and all of their offspring who were alive on April 1, 1968, or born between that
date and December 31, 1988. The offspring constituted the
study cohort.
Information about MS in cohort members and their parents was obtained through linkage with the Danish Multiple
Sclerosis Register. The Danish Multiple Sclerosis Register
was formally established in 1956, in continuation of a nationwide MS surveillance study that began in 1949. Since
then, the register has collected clinical information on all
MS patients in Denmark (35). It is the longest-running nationwide MS register in the world (36). Cases have been
classified by only three neurologists since the register’s inception. The cases fulfill the diagnostic criteria of Allison or
Poser (including possible MS) (37–39), and the diagnoses
are consistent over time. The register has been estimated to
be more than 90 percent complete (36).
The possible impacts of sibship characteristics on MS
risk, measured by incidence rate ratios, were investigated
in a follow-up study using log-linear Poisson regression
models (40). The studied sibship characteristics at age 10
years were: number of older siblings, number of younger
siblings, total number of siblings, age distance from the
nearest younger sibling (<2, 2–<6, or 6 years), and number of years exposed to younger siblings under 2 years
of age (<1, 1–<3, 3–<5, or 5 years). We also studied
whether being a member of a multiple birth versus being
a singleton was associated with MS and whether the age of
the mother or father at birth was associated with MS. Children born on the same day or an adjacent day as a sibling
were considered members of a multiple birth.
Am J Epidemiol 2006;163:1112–1117
TABLE 1. Characteristics of 1,036 multiple sclerosis patients
aged 10 years or more who were born in Denmark in 1950 or
later and whose cases were diagnosed during the period
1968–1998
No.
%
Male
331
32
Female
705
68
10–14
19
2
15–19
95
9
20–24
315
30
25–29
329
32
30–34
196
19
35–39
71
7
40–45
11
1
Sex
Age (years) at diagnosis
Year of birth and age (years) at diagnosis
1950–1959
<30
55
5
30
79
8
1960–1969
<30
499
48
30
153
15
<30
250
24
30
0
0
Father
16
1
Mother
24
2
Sibling
8
0
Twin
1
0
1970 or later
Family history of multiple sclerosis
Overall, the study cohort consisted of 1,903,625 persons
who were followed for MS from their 10th birthday or April
1, 1968, whichever came last, until MS diagnosis, death,
emigration, or December 31, 1998, whichever came first.
The results were adjusted for the potential interaction between sex and calendar period and the potential interaction between sex and age using quadratic restricted splines,
where quadratic functions were connected at 15, 20, 25, 30,
and 35 years (41, 42). All analyses were also adjusted for
parental MS (yes, no). Parents were considered to have MS
if one or both had ever been diagnosed with MS. Additional adjustments were carried out for age of the mother
at birth (<20, 20–24, 25–29, or 30 years), number of older
siblings (0, 1, 2, or 3), and number of younger siblings (0,
1, 2, or 3).
Maximum likelihood estimation was performed using the
GENMOD procedure in SAS (version 8.02; SAS Institute,
Inc., Cary, North Carolina). Two-sided p values were based
on likelihood ratio tests, and 95 percent confidence intervals
were based on Wald’s tests. Trend slopes were estimated
by treating categorical variables of interest as quantitative
Downloaded from https://academic.oup.com/aje/article-abstract/163/12/1112/97117 by guest on 03 June 2020
consistency in many studies, and its existence motivated the
so-called hygiene hypothesis as an explanation for the allergy epidemic in affluent countries (13, 14). Fewer and
smaller studies exist on MS and birth order; results have
been very inconsistent, and several of the studies have suffered from severe methodological problems (6, 15–33).
Thus, to investigate whether age at exposure to a common
microbial agent is associated with risk of developing MS,
we took advantage of the high quality of information in
Danish national registers and established a nationwide cohort that was followed for development of MS and assessed
for number of older siblings (i.e., birth order) and other
relevant sibship characteristics.
1113
1114 Bager et al.
TABLE 2. Rate ratios for multiple sclerosis according to sibship characteristics in a cohort of 1.9 million
persons, Denmark, 1968–1998*
No. of cases
Person-years at risk
RR1y,z
95% CIy
RR2
95% CI
535
363
105
33
14,413,036
9,527,801
3,123,504
1,019,803
1
1.13
1.04
1.04
0.47
Reference
0.98, 1.29
0.84, 1.28
0.73, 1.47
1
1.14
1.06
1.08
0.44
Reference
0.98, 1.32
0.84, 1.34
0.74, 1.56
389
439
168
40
11,141,814
11,628,581
4,196,246
1,117,504
1
0.93
0.89
0.75
0.053
Reference
0.81, 1.07
0.74, 1.07
0.54, 1.04
1
0.94
0.90
0.74
0.08
Reference
0.81, 1.09
0.73, 1.09
0.53, 1.04
83
487
331
105
30
2,563,365
13,220,852
8,463,607
2,713,435
1,122,885
1
1.11
1.11
1.05
0.74
0.23
Reference
0.88, 1.41
0.87, 1.41
0.79, 1.41
0.49, 1.13
1
1.12
1.12
1.06
0.75
Reference
0.89, 1.42
0.88, 1.42
0.80, 1.42
0.49, 1.14
0.26
1,024
12
27,602,343
481,801
1
0.69
Reference
0.39, 1.23
1
0.68
Reference
0.38, 1.20
179
493
289
71
4
3,661,097
12,481,820
8,716,625
2,709,201
515,402
1.13
1
1.01
1.04
0.49
0.50zz
0.95, 1.34
Reference
0.87, 1.17
0.80, 1.35
0.18, 1.31
1.19
1
0.96
0.97
0.46
1.00, 1.42
Reference
0.82, 1.13
0.74, 1.28
0.17, 1.24
0.12zz
42
298
385
186
62
14
13
36
812,223
7,429,172
10,563,033
5,616,521
1,915,504
574,975
266,437
906,278
1.20
1
1.02
1.04
1.09
0.83
1.65
0.94
0.65zz
0.87, 1.66
Reference
0.87, 1.18
0.86, 1.25
0.83, 1.44
0.49, 1.42
0.95, 2.89
0.66, 1.32
1.23
1
0.99
1.00
1.05
0.80
1.60
0.92
0.89, 1.70
Reference
0.85, 1.15
0.82, 1.21
0.79, 1.40
0.47, 1.38
0.91, 2.81
0.65, 1.30
0.97zz
* A total of 1,036 persons developed multiple sclerosis during 28.1 million person-years at risk.
y RR, rate ratio; CI, confidence interval.
z RR1 was adjusted for the potential interaction between sex and calendar period, the potential interaction between sex and age, and parental multiple sclerosis.
§ RR2 was adjusted for the same factors as RR1, as well as for age of the mother at birth and number of younger
siblings.
{ RR2 was adjusted for the same factors as RR1, as well as for age of the mother at birth and number of older
siblings.
# RR2 was adjusted for the same factors as RR1, as well as for age of the mother at birth.
** RR2 was adjusted for the same factors as RR1, as well as age of the mother at birth, number of older siblings,
and number of younger siblings.
yy RR2 was adjusted for the same factors as RR1, as well as for number of older siblings and number of younger
siblings.
zz The test for trend was based on 1-year categories (for age of the father, the missing-data group was not
included in the trend test).
Am J Epidemiol 2006;163:1112–1117
Downloaded from https://academic.oup.com/aje/article-abstract/163/12/1112/97117 by guest on 03 June 2020
No. of older siblings§
0
1
2
3
ptrend
No. of younger siblings{
0
1
2
3
ptrend
Total no. of siblings#
0
1
2
3
4
ptrend
Member of a multiple birth**
No
Yes
Age of mother (years) at birthyy
<20
20–24
25–29
30–34
>34
ptrend
Age of father (years) at birthyy
<20
20–24
25–29
30–34
35–39
40–44
>44
Missing data
ptrend
Sibship Characteristics and Multiple Sclerosis Risk
1115
TABLE 3. Rate ratios* for multiple sclerosis according to combinations of number of older siblings and
total number of siblings at age 10 years in a cohort of 1.9 million persons, Denmark, 1968–1998y
Total no. of siblingsz
No. of older siblings
1
2
3
4
RR§
95% CI§
RR
95% CI
RR
95% CI
RR
95% CI
0
1.02
0.80, 1.31
1.05
0.80, 1.38
0.93
0.60, 1.44
0.83
0.36, 1.91
1
1.32
1.01, 1.70
1.19
0.90, 1.57
0.93
0.61, 1.44
0.46
0.17, 1.25
1.16
0.83, 1.62
1.25
0.82, 1.90
0.68
0.30, 1.55
1.35
0.81, 2.25
0.96
0.54, 1.71
2
3
variables. Trends for ages of the mother and father at birth
were estimated using 1-year categories. Effect modification
by year of birth (1950–1959, 1960–1969, or 1970 onward)
and age was evaluated by including interaction terms in the
model.
RESULTS
Overall, 1,036 persons aged 10 years or more were diagnosed with MS during the 28.1 million person-years of
follow-up. Table 1 shows the distribution of MS cases by
gender, age, year of birth, and family history of MS.
Table 2 shows rate ratios for MS according to number of
older siblings, number of younger siblings, total number of
siblings, being a member of a multiple birth, and ages of the
mother and father at birth. Overall, there was no significant
association between the MS rate ratio and number of older
siblings or total number of siblings. There was a tendency
for the MS rate ratio to increase with decreasing number
of younger siblings; however, the trend was not significant
(p ¼ 0.08). The trend estimate for the MS rate ratio for each
of these exposures was not modified by year of birth. Neither was there an association with being a member of a multiple birth as compared with being a singleton or with the
age of the mother or father at birth.
For number of older siblings and total number of siblings,
the effect on the MS rate ratio was the same for subjects
under age 30 years and subjects aged 30 years or more. For
number of younger siblings, the rate ratio increased with
decreasing number of younger siblings among subjects below age 30 years (rate ratio (RR) for trend ¼ 0.87, 95 percent confidence interval (CI): 0.79, 0.96) but not among
subjects aged 30 years or more (RR for trend ¼ 1.04, 95
percent CI: 0.92, 1.18). When the effect of younger siblings
was stratified by year of birth, ages <25 and 25 years, ages
<30 and 30 years, and year of birth and ages <30 and 30
years yielded the same results.
To conform the analyses of younger siblings with those
of a recent study (23) (see Discussion), we also analyzed
the rate ratio for MS according to cumulative years of having younger siblings under 2 years of age. Overall, there was
Am J Epidemiol 2006;163:1112–1117
no association; the trend estimate for the rate ratio was
0.98 (95 percent CI: 0.91, 1.06) after adjustment for sex 3
calendar period, sex 3 age, parental MS, age of the mother
at birth, and number of older siblings. Stratifying by year
of birth and age <30 years/30 years yielded the same result, although for births occurring after 1970, the trend estimate for MS diagnosed at ages below 30 years was 0.87
(95 percent CI: 0.71, 1.07), and no cases were aged 30 years
or more. Furthermore, the rate ratio for MS was not associated with age distance to the nearest younger sibling
(p ¼ 0.44).
Table 3 shows rate ratios for MS according to combinations of number of older siblings and total number of siblings. Compared with only children, there was no evidence
of an increased MS rate ratio with decreasing number of
older siblings when the total number of subjects’ siblings
was one, two, three, or four or more. In fact, second-born
children in families with two children had an increased MS
rate ratio in comparison with only children (RR ¼ 1.32, 95
percent CI: 1.01, 1.70). Overall, there was no effect of interaction between total number of siblings and number of
older siblings on the rate ratio for MS (p ¼ 0.70).
The results presented did not change materially when
subjects with a diagnosis of possible MS (n ¼ 131) were
not considered MS cases in the analysis; neither did the
results change materially without adjustment for sex.
DISCUSSION
The present population-based cohort study was based on
detailed information on MS and family characteristics for
a large nationwide cohort of Danes. The results do not support an increased risk of MS with decreasing number of
older siblings as previously suggested (15, 31, 32), nor do
they suggest an overall association with number of younger
siblings, total number of siblings, age distance to the nearest
younger sibling, years of exposure to younger siblings under
2 years of age, being a member of a multiple birth versus
being a singleton, or age of the mother or father at birth.
Selection bias, which may have hampered previous studies,
Downloaded from https://academic.oup.com/aje/article-abstract/163/12/1112/97117 by guest on 03 June 2020
* Adjusted for the potential interaction between sex and calendar period, the potential interaction between sex and
age, parental multiple sclerosis, and age of the mother at birth.
y A total of 1,036 persons developed multiple sclerosis during 28.1 million person-years at risk.
z Reference category: no siblings.
§ RR, rate ratio; CI, confidence interval.
1116 Bager et al.
but not older siblings. In particular analyses, the authors
showed that patients were diagnosed at a young age less
often if they had lived in the same house with younger siblings under 2 years of age during the first 6 years of life (23).
Their focus on that exposure, and not exclusively the number of younger siblings, was based on their assumption that
it is a specific marker of reinfections transmitted from younger siblings of very low age. We did not reproduce their
result using comparable analyses of MS among two million
Danes. We observed that patients were diagnosed at a young
age less often when they had younger siblings, but this could
not be attributed to the number of years exposed to younger
siblings of very low age. In addition, being closer in age to
the nearest younger sibling was not associated with fewer
diagnoses of MS. When the findings are analyzed together,
our results are not compatible with the Tasmania study or its
basic hypothesis (23).
Finally, it has been hypothesized that children who are
members of a multiple birth may tend to have increased
exposure to infections early in life from close contact between multiple siblings (44). However, we found no association between being a member of a multiple birth and the
rate ratio for MS. In a case-control study of 241 MS patients,
Antonovsky et al. (22) reported that a significantly higher
percentage of MS patients than of controls were born to
mothers aged 40 years or more but not to mothers aged 30
years or more, although numbers were small. As in the
larger Swedish study (15), we found no association between
MS and the age of the mother.
In conclusion, these results do not lend support to the
hypothesis that age at infection with a common microbial
agent, as approximated by number of older siblings or any
of the other sibship characteristics studied, is associated
with risk of MS.
ACKNOWLEDGMENTS
Conflict of interest: none declared.
REFERENCES
1. Kurtzke JF. Epidemiologic evidence for multiple sclerosis as
an infection. Clin Microbiol Rev 1993;6:382–427.
2. Willer CJ, Dyment DA, Risch NJ, et al. Twin concordance and
sibling recurrence rates in multiple sclerosis. Proc Natl Acad
Sci U S A 2003;100:12877–82.
3. Poskanzer DC, Schapira K, Miller H. Multiple sclerosis and
poliomyelitis. Lancet 1963;186:917–21.
4. Granieri E, Casetta I, Tola MR, et al. Multiple sclerosis: infectious hypothesis. Neurol Sci 2001;22:179–85.
5. Casetta I, Granieri E. Clinical infections and multiple sclerosis: contribution from analytical epidemiology. J Neurovirol
2000;6(suppl 2):S147–51.
6. Bager P, Nielsen NM, Bihrmann K, et al. Childhood infections
and risk of multiple sclerosis. Brain 2004;127:2491–7.
7. Taylor I, Knowelden J. Family size. In: Principles of epidemiology. London, United Kingdom: J & A Churchill Ltd,
1964:304–7.
Am J Epidemiol 2006;163:1112–1117
Downloaded from https://academic.oup.com/aje/article-abstract/163/12/1112/97117 by guest on 03 June 2020
was considered minimal here, because we used information
from nationwide registries of high quality. Furthermore, we
took variations in population family size over time into account by adjusting for a possible calendar effect on sibship
characteristics.
The majority (76 percent) of the studied MS patients were
born before the 1970s, the era when day-care attendance
began to increase in Denmark. Among children aged 2
years, 16 percent attended day care in 1973, 41 percent in
1983, and 49 percent in 1992 (43). Therefore, the lack of
support for the hypothesis that firstborn children tend to be
infected at a late age and consequently are at increased risk
of MS was not likely to be explained by many firstborns’
attending day care, thereby being exposed earlier to infections. In fact, there was no association between the rate ratio
for MS and number of siblings in any stratum of birth year
(1950–1959, 1960–1969, and 1970 or later).
In contrast to our results, in a recent large case-control
study of 4,443 Swedish MS patients, Montgomery et al. (15)
reported an inverse association between MS risk and number of siblings, both older and younger, and a decreased risk
of MS among twins. The Swedish findings also contrasted
with the lack of association reported in most previous casecontrol studies for number of older siblings (6, 16–24) and
total number of siblings (6, 16, 22, 33). Although the authors
did not rule out the possibility that the associations they
found, including a greater age distance to the nearest sibling
among MS patients, could be explained by late exposure to
infections, they were more inclined toward the interpretation that increased risk of MS could be associated with reduced fertility among parents of MS patients. To support
this, they also reported that the age of the father was associated with risk of MS (15). We found no association with
the age of the father. However, Montgomery et al.’s population was missing up to half of all deceased persons, and 25
percent of cases and 15 percent of controls were excluded
because of incomplete information on siblings (15). The
possibility that this could have affected their results cannot
be excluded.
The analyses presented by Ahlgren and Andersen (30)
from another Swedish study of 258 MS patients illustrated
how serious bias could have been introduced into the earliest
studies of MS and birth order because the birth order distributions of MS patients were compared with a theoretical
distribution based on the erroneous assumption that the population family size remains stable through time (24–29, 31).
Investigators in later studies avoided this bias by using a
case-control design matching the data on time of birth (6,
16–24, 31–33), but only a few of the studies, based on relatively small and selected case materials, observed an association between low birth order and increased risk of MS
(31–33). Among the larger of these case-control studies,
Bager et al. (6) reported no association in a study of 455
MS patients, and in another study of 301 female MS patients, Hernan et al. (33) reported an increased risk of MS
among firstborn siblings in large sibships (4), although the
trend analysis was less convincing.
In a recent case-control study of 136 MS patients in Tasmania, Australia, Ponsonby et al. (23) reported risk of MS to
be associated with a decreasing number of younger siblings
Sibship Characteristics and Multiple Sclerosis Risk
Am J Epidemiol 2006;163:1112–1117
26. Cendrowski W. Birth order of multiple sclerotic subjects.
Psychiatr Neurol Basel 1966;151:161–7.
27. Cripps J, Rudd A, Ebers GC. Birth order and multiple sclerosis. Acta Neurol Scand 1982;66:342–6.
28. Sadovnick AD, Yee IM, Ebers GC. Multiple sclerosis and birth
order: a longitudinal cohort study. Lancet Neurol 2005;4:
611–17.
29. Pratt RTC. The familial occurrence of disseminated sclerosis.
Ann Eugen 1951;16:45–59.
30. Ahlgren C, Andersen O. No major birth order effect on the
risk of multiple sclerosis. Neuroepidemiology 2005;24:
38–41.
31. Zilber N, Kutai-Berman M, Kahana E, et al. Multiple sclerosis and birth order. Acta Neurol Scand 1988;78:313–17.
32. Isager H, Andersen E, Hyllested K. Risk of multiple sclerosis
inversely associated with birth order position. Acta Neurol
Scand 1980;61:393–6.
33. Hernan MA, Zang SM, Lipworth L, et al. Multiple sclerosis
and age at infection with common viruses. Epidemiology
2001;12:301–6.
34. Westergaard T. Birth order and cancer risk. (PhD thesis).
Copenhagen, Denmark: University of Copenhagen, 1997.
35. Koch-Henriksen N. The Danish Multiple Sclerosis Registry:
a 50-year follow-up. Mult Scler 1999;5:293–6.
36. Koch-Henriksen N, Rasmussen S, Stenager E, et al. The
Danish Multiple Sclerosis Registry. History, data collection
and validity. Dan Med Bull 2001;48:91–4.
37. Allison RS, Millar JH. Prevalence of disseminated sclerosis in
Northern Ireland. Ulster Med J 1954;23:1–27.
38. Millar JH, Allison RS. Familial incidence of disseminated
sclerosis in Northern Ireland. Ulster Med J 1954;23:
29–92.
39. Poser CM, Paty DW, Scheinberg L, et al. New diagnostic
criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol 1983;13:227–31.
40. Breslow NE, Day NE, eds. Statistical methods in cancer
research. Vol 2. The design and analysis of cohort studies.
(IARC Scientific Publication no. 82). Lyon, France: International Agency for Research on Cancer, 1987.
41. Wohlfahrt J, Mouridsen H, Andersen PK, et al. Reproductive
risk factors for breast cancer by receptor status, histology,
laterality and location. Int J Cancer 1999;81:49–55.
42. Stenager E, Moller R, Brønnum-Hansen H, et al. Multiple
sclerosis: its change in prevalence in Denmark, 1950–1996,
and its total prevalence in Danish counties in 1996. (In
Danish). Ugeskr Laeger 2005;167:1500–2.
43. Westergaard T, Andersen PK, Pedersen JB, et al. Birth characteristics, sibling patterns, and acute leukemia risk in childhood: a population-based cohort study. J Natl Cancer Inst
1997;89:939–47.
44. McKeever TM, Lewis SA, Smith C, et al. Siblings, multiple
births, and the incidence of allergic disease: a birth cohort
study using the West Midlands general practice research
database. Thorax 2001;56:758–62.
Downloaded from https://academic.oup.com/aje/article-abstract/163/12/1112/97117 by guest on 03 June 2020
8. MacMahon B. Is acute lymphoblastic leukemia in children
virus-related? Am J Epidemiol 1992;136:916–24.
9. Hammon W, Sather GE, Hollinger N. Preliminary report of
epidemiological studies on poliomyelitis and streptococcal
infections; Lansing neutralizing antibody and antistreptolysin
O surveys of California cities, Texas, North Carolina, Mexico,
Pacific Islands, and Japan. Am J Public Health 1950;40:
293–306.
10. Goodman KJ, Correa P. Transmission of Helicobacter pylori
among siblings. Lancet 2000;355:358–62.
11. Valeur-Jensen AK, Pedersen CB, Westergaard T, et al. Risk
factors for parvovirus B19 infection in pregnancy. JAMA
1999;281:1099–105.
12. Lanari M, Giovannini M, Giuffre L, et al. Prevalence of
respiratory syncytial virus infection in Italian infants hospitalized for acute lower respiratory tract infections, and association between respiratory syncytial virus infection risk factors
and disease severity. Pediatr Pulmonol 2002;33:458–65.
13. Strachan DP. Family size, infection and atopy: the first decade
of the ‘‘hygiene hypothesis.’’ Thorax 2000;55(suppl 1):S2–10.
14. Karmaus W, Botezan C. Does a higher number of siblings
protect against the development of allergy and asthma? A review. J Epidemiol Community Health 2002;56:209–17.
15. Montgomery SM, Lambe M, Olsson T, et al. Parental age,
family size, and risk of multiple sclerosis. Epidemiology
2004;15:717–23.
16. Gaudet JP, Hashimoto L, Sadovnick AD, et al. Is sporadic
MS caused by an infection of adolescence and early adulthood?
A case-control study of birth order position. Acta Neurol
Scand 1995;91:19–21.
17. Koch-Henriksen N. An epidemiological study of multiple
sclerosis. Familial aggregation, social determinants, and exogenic factors. Acta Neurol Scand Suppl 1989;124:1–123.
18. Matias-Guiu J, Ferri L, Falip R, et al. Birth order in multiple
sclerosis: a population-based case-control study in Alcoi,
Spain. Neuroepidemiology 1994;13:129–30.
19. Operskalski EA, Visscher BR, Malmgren RM, et al. A casecontrol study of multiple sclerosis. Neurology 1989;39:825–9.
20. Alter M, Cendrowski W. Multiple sclerosis and childhood
infections. Neurology 1976;26:201–4.
21. Berr C, Puel J, Clanet M, et al. Risk factors in multiple
sclerosis: a population-based case-control study in HautesPyrenees, France. Acta Neurol Scand 1989;80:46–50.
22. Antonovsky A, Leibowitz U, Smith HA, et al. Epidemiologic
study of multiple sclerosis in Israel. I. An overall review of
methods and findings. Arch Neurol 1965;13:183–93.
23. Ponsonby AL, van der Mei I, Dwyer T, et al. Exposure to
infant siblings during early life and risk of multiple sclerosis.
JAMA 2005;293:463–9.
24. Visscher BR, Liu KS, Sullivan CB, et al. Birth order and
multiple sclerosis. Acta Neurol Scand 1982;66:209–15.
25. Alperovitch A, Le Canuet P, Marteau R. Birth order and risk
of multiple sclerosis: are they associated and how? Acta
Neurol Scand 1981;63:136–8.
1117