Share

November 2020; 7 (6) ArticleOpen Access

Pregnancy does not modify the risk of MS in genetically susceptible women

Cameron J. Adams, Sean L. Wu, Xiaorong Shao, Patrick T. Bradshaw, Edlin Gonzales, Jessica B. Smith, Anny H. Xiang, Kalliope H. Bellesis, Terrence Chinn, Steffan D. Bos, Marte Wendel-Haga, Tomas Olsson, Ingrid Kockum, Annette M. Langer-Gould, Catherine Schaefer, Lars Alfredsson, Lisa F. Barcellos
First published October 9, 2020, DOI: https://doi.org/10.1212/NXI.0000000000000898
Full PDF
Citation
Permissions

Make Comment

See Comments

Downloads
469

Share

Abstract

Objective To use the case-only gene-environment (G Embedded Image E) interaction study design to estimate interaction between pregnancy before onset of MS symptoms and established genetic risk factors for MS among White adult females.

Methods We studied 2,497 female MS cases from 4 cohorts in the United States, Sweden, and Norway with clinical, reproductive, and genetic data. Pregnancy exposure was defined in 2 ways: (1) Embedded Image live birth pregnancy before onset of MS symptoms and (2) parity before onset of MS symptoms. We estimated interaction between pregnancy exposure and established genetic risk variants, including a weighted genetic risk score and both HLA and non-HLA variants, using logistic regression and proportional odds regression within each cohort. Within-cohort associations were combined using inverse variance meta-analyses with random effects. The case-only G × E independence assumption was tested in 7,067 individuals without MS.

Results Evidence for interaction between pregnancy exposure and established genetic risk variants, including the strongly associated HLA-DRB1*15:01 allele and a weighted genetic risk score, was not observed. Results from sensitivity analyses were consistent with observed results.

Conclusion Our findings indicate that pregnancy before symptom onset does not modify the risk of MS in genetically susceptible White females.

Glossary

EIMS=
Epidemiologic Investigation of Multiple Sclerosis;
G × E=
gene-environment;
GEMS=
Genes and Environment in Multiple Sclerosis;
GERA=
Genetic Epidemiology Research on Adult Health and Aging;
GWAS=
genome-wide association study;
KPNC=
Kaiser Permanente Northern California;
MAF=
minor allele frequency;
NOR=
Norwegian MS Registry and Biobank;
PCA=
principal component analysis;
wGRS=
weighted genetic risk score

MS is a demyelinating autoimmune disease of the CNS with both environmental and genetic risk factors.1 This progressive disease results in significant disability and decreased quality of life.2 MS is more prevalent among females than males, and symptoms typically emerge during child-bearing ages, often soon after pregnancy.2,3 This has led many to hypothesize that female-specific exposures, such as those related to reproduction, pregnancy, and lactation, have a role in MS. Pregnancy appears to have short-term beneficial effects on existing MS symptoms,4 but there is no agreement in the scientific literature about the effect of pregnancy on MS risk in general or among women with genetic susceptibility to MS.1,5

Gene-environment (G × E) interactions, for which the effect of an environmental exposure is modified by specific genotype(s), are believed to contribute substantially to complex disease risk, and discovery of these interactions can identify subgroups with higher risk of disease.6 Studies of pregnancy and MS risk have yielded conflicting results; however, no studies, to date, have investigated interaction between pregnancy and genetic susceptibility for risk of MS.7,–,16

We used case-only G × E methods to evaluate interaction between pregnancy before symptom onset and known genetic risk factors for MS, including HLA-DRB1*15:01. Study participants included 2,497 White individuals from 4 established MS cohorts based in California, Sweden, and Norway. Analyses were conducted separately within each cohort and then combined with meta-analytic methods. A separate cohort of females without MS was used to test the case-only G × E independence assumption.

Methods

Study participants

Participants were selected from the Kaiser Permanente Northern California (KPNC) MS Research Program, the Kaiser Permanente Southern California MS Sunshine Study, the Norwegian MS Registry and Biobank (NOR), and 2 Swedish MS studies, the Epidemiological Investigation of Multiple Sclerosis (EIMS) and the Genes and Environment in Multiple Sclerosis (GEMS) study.17,–,20 MS diagnoses were confirmed by independent neurologists according to the 2005 revised (KPNC, NOR, and EIMS/GEMS) or the 2010 revised McDonald diagnostic criteria (MS Sunshine).21,22 Additional non-MS female members of KPNC were also studied; these individuals were derived from the Genetic Epidemiology Research on Adult Health and Aging (GERA) study.23 KPNC and KPSC are integrated health services systems and are the largest health care providers in California. Their membership is largely representative of their catchment area populations; however, persons of lower socioeconomic status are underrepresented.24,25

Standard protocol approvals, registrations, and patient consents

All study participants provided written informed consent, and the Institutional Review Boards of KPNC and KPSC, regional ethical review boards in Norway and Sweden, and the University of California, Berkeley, approved the study protocols.

KPNC participants

Study recruitment is described in detail elsewhere.17 Briefly, study participants were recruited from KPNC membership between 2006 and 2014. Prevalent MS cases were the focus of recruitment. Participants were aged between 18 and 69 years and were KPNC members at initial contact.

MS Sunshine participants

MS Sunshine is a multiethnic case-control study of incident MS and first demyelinating event.18 Participants in this study were recruited from KPSC membership between 2011 and 2014. At the time of initial contact, participants were KPSC members, aged 18 years or older, and diagnosed with MS within 1.5 years or had MS symptom onset within the 3 years before recruitment.

NOR participants

NOR participants were recruited from the Oslo MS Registry and DNA biobank in 2011–2012.20 The Oslo MS Registry and biobank was established in 1990 and includes clinical data and DNA samples from a population-based MS cohort.

EIMS/GEMS participants

EIMS and GEMS are Swedish population-based case-control studies. At enrollment, EIMS participants were aged 18–70 years and had recently (within 2 years) confirmed MS. GEMS participants were identified from the Swedish National MS registry and recruited between 2009 and 2011. All EIMS participants were distinct from the GEMS study. Details have been described elsewhere.19

GERA participants (noncases)

GERA participants are a subsample of the Research Program on Genes, Environment, and Health longitudinal cohort.23 Participants were respondents to a mailed health survey that was sent to all adult members of KPNC in 2007 who had been members of KPNC for at least 2 years and were aged 18 years or older. After providing informed consent, participants were asked to submit a saliva sample for DNA genotyping. Additional phenotypic and health condition data were obtained from International Classification of Diseases, Ninth Revision codes from KPNC electronic health record data. Participants included in this study were confirmed to not have MS or other autoimmune disease.

Ancestry determination

KPNC, MS Sunshine, and GERA participants with average European ancestry proportions greater than 80% estimated from ancestry informative markers who did not did not self-report Hispanic ethnicity were included.26 ALL NOR and EIMS/GEMS participants had self-reported White ancestry. Population outliers were identified with principal components analysis (PCA) and excluded.27

Genotype and exposure assessment

All participants in this study completed an interview or self-reported questionnaires related to MS disease events, reproductive history, and environmental exposures.17,–,20 Participants provided blood or saliva samples for genotyping. Genome-wide single nucleotide polymorphism (SNP) genotyping was performed using the Illumina Infinium 660K BeadChip Array and Human Omni Express Array (KPNC, MS Sunshine), Illumina ImmunoArray BeadChip Array (NOR, EIMS/GEMS), and Affymetrix Axiom Array (GERA). Genome-wide SNP imputation was performed using SHAPEIT2 and IMPUTE2.28 HLA alleles were imputed using SNP2HLA.29 Participants with missing genotypes that met QC thresholds (info score >0.8, missingness per SNP <0.05, missingness per cohort <0.05, and minor allele frequency [MAF] > 0.05) were imputed using the average MAF within each cohort.

Pregnancy exposure was evaluated as dichotomous variable, having at least 1 live birth pregnancy before reported age at first MS symptom onset, and as an ordinal variable, the number of live birth pregnancies before reported age at first MS symptom onset (table 1). Age of pregnancy was defined as mother's age at birth, and age at symptom onset for MS cases was determined through review of medical records and/or comprehensive clinical histories for each participant collected through interview or questionnaire.

View this table:
Table 1

Characteristics of multiple sclerosis cases (N = 2,497)

Candidate genes and weighted genetic risk score

To maximize power and to identify variants with functional relevance to MS, we conducted a 2-tiered analysis. In the primary analysis, we assessed G × E interaction between pregnancy exposure and (1) a weighted genetic risk score (wGRS) comprised of recently established non-HLA MS genome-wide association study (GWAS) risk variants, (2) HLA-DRB1*15:01 alleles, and (3) HLA-A*02:01 alleles.30 We also evaluated evidence for effect modification of the interaction between HLA-A*02:01 and HLA-DRB1*15:01 for MS susceptibility by pregnancy exposure.31 In the secondary discovery analysis, we individually tested established MS-associated variants (HLA: 2 variants) and (non-HLA: 144 variants) that passed QC criteria in all 4 cohorts. Individual genotypes were modeled assuming a linear effect of each additional risk allele (0, 1, or 2 risk alleles).

The wGRS was derived by multiplying the log OR for each risk allele from recent GWAS by the number of risk alleles carried by each participant and summing across GWAS variants (table e-1, links.lww.com/NXI/A320).30 Scores for each individual were calculated using non-HLA risk variants that passed genotype QC thresholds within each cohort (KPNC, MS Sunshine, and EIMS/GEIMS: 182 non-HLA risk variants, NOR: 144 non-HLA risk variants, and GERA: 161 non-HLA risk variants). Within each cohort, the wGRS was modeled as a continuous variable and with categorical quartiles (reference: 1st quartile).

Statistical analysis

Genome-wide SNP genotyping data were available for 990 KPNC, 409 MS Sunshine, 153 NOR, and 1,462 EIMS/GEMS female MS cases. Following exclusion of MS cases with European ancestry proportions <80% (KPNC and MS Sunshine) and population outliers identified from PCA analysis (NOR and EIMS/GEMS), 814 KPNC, 151 MS Sunshine, 119 NOR, and 1,413 EIMS/GEMS cases had age at onset Embedded Image18 years, complete pregnancy history data, and genome-wide genetic profiles. The final data set for G × E analysis included 2,497 MS cases (table 1, figure e-1, links.lww.com/NXI/A319).

Case-only G x E models rely on the assumption that the genotype and environmental exposure are uncorrelated in the source population. If that assumption is valid, the association between genotype and exposure in a case-only model estimates the departure of the joint of effects on the OR scale. If the disease is rare and the above assumption is valid, the case-only model estimates the departure of joint effects on the risk ratio scale.32 We used logistic regression to model having at least 1 live birth pregnancy before symptom onset as a function of genotypes or wGRS:Embedded ImageEmbedded Image is an indicator for having at least 1 live birth pregnancy before symptom onset or not, G is 0, 1, or 2 alleles for a risk variant or continuous wGRS, and Embedded Image is the estimate of the interaction parameter measuring the departure of the joint effects of E and G on the multiplicative scale.32

Proportional odds regression was used to model parity before symptom onset,32 where the probability of an equal or less than k number of pregnancies before symptom onset, Embedded Image, to the probability of more than k number of pregnancies, Embedded Image as a function of genotypes or wGRS:Embedded ImageEmbedded Image is the number of live birth pregnancies before MS symptom onset, G is 0, 1, or 2 alleles for a risk variant or continuous wGRS, Embedded Image is the threshold for live birth pregnancies before symptom onset for each ordinal comparison, and Embedded Image is the estimate of the interaction parameter measuring the departure of the joint effects of E and G on the multiplicative scale.32

G × E interaction was estimated within each study, and combined estimates of G × E interaction were obtained with random-effects meta-analysis using restricted maximum likelihood estimation with weights proportional to the inverse of the variance for each cohort-specific association. Embedded Image was used to study assess heterogeneity between within-study association. All models were adjusted for age at MS onset and population stratification using components from PCA. wGRS quartiles were modeled using dummy variables with the first quartile as the reference category. Secondary discovery analysis p values were adjusted for the false discovery rate using the Benjamini-Hochberg method, and adjusted p values <0.05 were considered significant.33

Effect modification of the interaction between HLA-A*02:01 and HLA-DRB1*15:01 by pregnancy exposure was evaluated separately by stratifying case-only regression models for pregnancy exposure and HLA-DRB1*15:01 by the absence and presence of HLA-A*02:01 alleles adjusting for age at MS onset and population stratification.

We tested the assumption of G × E independence between genotype and pregnancy using healthy female GERA participants (N = 7,067).6 Logistic and proportional odds regression models described above were used to test for associations between MS genetic risk factors, including wGRS and HLA-DRB1*15:01, and having Embedded Image1 pregnancy or not and parity. All statistical analyses were conducted using Plink 1.9 and R 3.5.1.

Sensitivity analyses

We estimated G x E association between live or non–live birth pregnancy before symptom onset (binary and ordinal models) and genetic variants among KPNC and MS Sunshine cases (data not available for NOR and EIMS/GEMS) to identify bias from exclusion of non–live birth pregnancies in primary analyses. To determine whether there was a short-term effect of pregnancy on MS risk, we altered the definition of pregnancy exposure to only consider first pregnancies that occurred within 5 years before age at onset as exposed. Next, to rule out reverse causality from MS disease latency or recall bias, we considered age at symptom onset at 5 years before reported age at onset. To check for bias from differing numbers of variants used in wGRSs between cohorts, G x E interactions for wGRSs derived from risk variants only found in NOR data (144 variants, table e-1, links.lww.com/NXI/A320) were estimated for each cohort and combined with meta-analysis.

Data availability

Anonymized data used in this study from KPNC, MS Sunshine, NOR, and EIMS/GEMS will be shared upon request by any qualified investigator pending Institutional Review Board approval at each site. GERA data are publicly available on dbGaP (phs000674.v2.p2).

Results

haracteristics of MS cases are summarized in table 1. Age at MS onset among the cohorts that recruited predominantly prevalent cases (KPNC, NOR, and EIMS/GEMS) was similar (mean = 33.1, 33.4, and 34.4 years, respectively), but occurred later among the cohort that recruited incident cases only (MS Sunshine, mean = 39.9 years). Year of symptom onset occurred earlier among KPNC and NOR (median = 1992) than EIMS/GEMS (median = 2005) and MS Sunshine (median = 2011). The average age at first pregnancy (live birth) ranged between 24.6 and 27.2 years, with KPNC and MS Sunshine participant pregnancies occurring earlier than NOR and EIMS/GEMS. More than half of cases were HLA-DRB1*15:01 carriers; NOR and EIMS/GEMS had a higher proportion of carriers (60% and 56%, respectively) than KPNC and MS Sunshine (54% and 47%, respectively). The median wGRS was similar across KPNC, MS Sunshine, and EIMS/GEMS. NOR cases had lower scores than the others, which is likely because fewer SNPs were available to calculate the NOR wGRS. Approximately half (49.8%) of the cases had a live birth pregnancy before onset of MS symptoms. Parous cases were most likely to have 2 live births before symptom onset.

In this study, the ORs from case-only regression models estimate the departure of the multiplicative joint effects of E and G on the risk ratio scale for susceptibility to MS. We did not find evidence for G × E interaction between pregnancy exposure and primary genetic risk factors (wGRS, HLA-DRB1*15:01, and HLA-A*02:01) (table 2, figure e-2, links.lww.com/NXI/A319). Point estimates were close to the null or had CIs that contained the null. Estimates were similar for both pregnancy exposures. Significant evidence for effect modification of interaction between HLA-DRB1*15:01 and pregnancy exposure by carriage of HLA-A*02:01 alleles was not observed, although estimates indicate possible protective effects (presence HLA*02:01: OR: 0.89, 95% CI 0.70, 1.13; absence HLA*02:01: OR: 1.09, 95% CI 0.71, 1.66) (table 3). No variants tested in the secondary discovery analysis had multiple testing adjusted p values <0.05 (tables e-2 and e-3, links.lww.com/NXI/A320).

View this table:
Table 2

Results from case-only meta-analyses estimating multiplicative interaction between HLA-DRB1*15:01, HLA-A*02:01, and weighted genetic risk scores with pregnancy before symptom onset

View this table:
Table 3

Estimates of multiplicative interaction between HLA-DRB1*15:01 and pregnancy exposure stratified by the presence and absence of HLA-A*02:01 alleles

Among GERA controls, the wGRS and HLA-DRB1*15:01 were not associated with having at least 1 pregnancy or with parity (table 4), and none of the MS GWAS loci were significantly associated with pregnancy in GERA after correcting for multiple testing (results not shown). Results from sensitivity analyses investigating live or non–birth pregnancy, pregnancy within 5 years of MS onset, and bias from latent disease were consistent with observed results (table 5). Estimates for interaction between pregnancy exposures and the wGRSs derived from variants present in NOR data were similar to observed results (table e-2 and e-3, links.lww.com/NXI/A320).

View this table:
Table 4

Evidence for G × E independence among n = 7,067 healthy genetic epidemiology research on aging participants

View this table:
Table 5

Sensitivity analysis results for pregnancy defined as live or non–live birth pregnancy, first pregnancy within 5 years before symptom onset, and latent MS onset models

Discussion

We hypothesized that pregnancy before MS symptom onset modifies the risk of MS in genetically susceptible females. Using data from 4 study populations, we did not find evidence to support G × E interaction between established genetic risk factors for MS and exposure to pregnancy before symptom onset. The HLA-A*02:01 ± stratified point estimates indicate a protective effect of HLA-A*02:01 alleles in the relationship between HLA-DRB1*15:01 and pregnancy exposure, but our study was not sufficiently powered to detect these associations. Evidence for interaction between pregnancy and non-HLA or HLA variants considered individually was also not observed.

Although we considered pregnancy as a single environmental exposure, pregnancy is a complex and heterogeneous combination of physiologic changes that result in weight gain, increases in lipid levels, and changes in basal metabolic rate, among others.5 These physiologic changes are the product of pregnancy-induced modifications in hormones, such as estriol, progesterone, prolactin, early pregnancy growth factor, alpha-fetoprotein, and leptin as well as elevated levels of other growth factors. There are increases in circulated regulatory T cells and B cells, increased Th2 responses, and decreased Th-1 and Th-17 immune responses.5 These immune changes are important for fetal tolerance, as the maternal immune system and endocrine pathways respond to fetal antigens that circulate in the mother.34 Following pregnancy, hormone levels and immune adaptions quickly return to prepregnancy states.35 The reduction of MS relapse rate during pregnancy is attributed to the dynamic immune and endocrine alterations that result from maternal-fetal crosstalk during pregnancy.36 Little is known about how physiologic changes during pregnancy affect the risk of developing MS, but it is hypothesized that the pregnancy-induced changes in endocrine pathways and immune system have protective effects.5

Epidemiologic studies investigating the effect of pregnancy on MS risk have reported conflicting results.1 Five studies reported a protective association between parity and risk of MS; however, 2 of these studies attributed their protective associations to reverse causality from reduced fertility and increased likelihood of miscarriage among women with latent MS.7,8,10,11,15 Five additional studies reported no association between pregnancy and risk of MS.9,12,–,14,16 A recent study investigating breastfeeding, ovulatory years, and risk of MS found evidence that cumulative duration of breastfeeding is associated with a decreased risk of MS.37 The authors suggest that breastfeeding duration confounds the association between parity/pregnancy and risk of MS and may explain previously reported conflicting findings.

Genes within the HLA complex contribute substantially to MS, with the HLA-DRB1*15:01 allele conferring the largest known genetic risk for disease.30 Interactions between this allele and established environmental exposures such as tobacco smoking, Epstein-Barr Virus infection, and adolescent obesity have previously demonstrated large effect sizes.1 A recent GWAS identified approximately 200 non-HLA genetic variants associated with MS risk.30

Previous research reported that protective associations between pregnancy and MS onset were attributable to reverse causality and that pregnancy was more likely among women with less severe disease or without latent disease.7,8 However, a registry-based study in Norway found that pregnancies among women with MS before symptom onset have birthweights and outcomes similar to those in women without MS.38 Furthermore, changes in counseling and availability of safer disease-modifying therapies and new diagnostic criteria that allow for earlier MS diagnosis have likely contributed to a recent reported increase in the rate of pregnancy among women with MS.39,40 Results from our sensitivity analyses did not indicate evidence for bias from latent disease. Furthermore, restricting analyses to cases reporting their first pregnancy within 5 years before symptom onset did not change our findings. Including data on non–live births in the pregnancy exposure did appear to move point estimates further from the null; however, data on non–live births were only available for KPNC and MS Sunshine cases and the CIs still contained the null.

Case-only G × E methodologies were developed to address one of the largest challenges with studying interactions, statistical power.6 Since they were first introduced, advances in case-only G × E methodology have focused on combining the increased power from case-only methods with evidence for the all-important G × E independence assumption from case-control data.6 Healthy controls with reproductive history matched on case symptom onset and genetic data were not available for our large combined data set of MS cases; however, we used female participants without MS from GERA to formally test for evidence of independence between MS genetic risk factors and pregnancy. Results from this analysis support the validity of our findings. Our models were adjusted for age at MS onset and population stratification. Although there may be additional variables that confound the relationship between both pregnancy and MS and genetic risk factors and MS, case-only G × E analyses only need to adjust for confounders of the G × E relationship. We cannot rule out that our findings may be due to confounding from unknown factors.

The objective of this study was to investigate G × E interaction between pregnancy and established MS genetic risk variants in MS. A primary strength of the current study was the large sample size with complete genetic and reproductive history data. In addition, cases used in this study were largely representative of their respective surrounding populations, providing support for external validity, and identification of cases from integrated health services delivery systems and national registries reduced the probability of selection bias. Furthermore, we found evidence to support the independence of G and E factors among the KPNC source population. Results from sensitivity analyses suggest that our findings are not attributable to reverse causality; however, given the uncertainty regarding disease latency and first MS symptoms, we cannot conclusively rule out reverse causality. Our study was subject to some limitations. Because of the absence of suitable controls, we were unable to evaluate the association between pregnancy and MS. Our G × E associations were consistent between studies; however, differences between the study populations may have biased our findings. With the exception of MS Sunshine, the other studies relied on older MS diagnostic criteria, which may have excluded milder MS cases. Because of the case-only study design, we were not able to assess interaction on the additive scale and our null findings on the multiplicative scale cannot rule out the presence of interaction on the additive scale. In addition, we were not able to assess G and E independence in non-KPNC source populations. Future studies should investigate the effects of pregnancy on MS risk among non-European populations. The current study was focused on White individuals to achieve the statistical power required for analyses. It is possible that interaction between MS genetic risk factors and pregnancy may differ by ethnic/ancestral group. If controls with reproductive data matched to case age at symptom onset are available, G × E interaction should be assessed among cases and controls with methods used to combine case-only and case-control interactions.

Our findings suggest that genetic susceptibility to MS does not modify the association between pregnancy and MS. This information may be useful for counseling women who have genetic susceptibility to MS about decisions to pursue pregnancy, although further research is needed to determine the effect of pregnancy on MS susceptibility.

Study funding

NIH/NIEHS R01ES017080 (Barcellos), NIH/NINDS R01NS049510 (Barcellos), NIH/NIAID R01AI076544 (Barcellos), NIH/NINDS R01NS075308 (Langer-Gould), NIH/NIA/NIH/NIH-OD RC2 AG036607 (Schaefer). Tomas Olsson has grants from the Swedish Research Council, the Swedish Brain Foundation, and the Margaretha af Ugglas Foundation. L. Alfredsson has grants from the Swedish Research Council, the Swedish Research council for Health, Working Life and Welfare, and the Swedish Brain Foundation. S.D. Bos has grants from the Norwegian Research Council and the EU-funded MultipleMS Project.

Disclosure

C. Adams, S.L. Wu, X. Shao, P.T. Bradshaw, E. Gonzales, J.B. Smith, A.H. Xiang, K.H. Bellesis, T. Chinn, S.D. Bos, and M. Wendel-Haga report no disclosures. T. Olsson has received honoraria for advisory boards and unrestricted MS research grants from Biogen, Novartis, Sanofi, Merck, and Roche. I. Kockum, A.M. Langer-Gould, C. Schaefer, L. Alfredsson, and L.F. Barcellos report no disclosures. Go to Neurology.org/NN for full disclosures.

Acknowledgment

The authors thank researchers in the MS group in Oslo, especially Prof. E.G. Celius and the International MS Genetics Consortium.

Appendix Authors

Table
Table

Footnotes

  • Go to Neurology.org/NN for full disclosures. Funding information is provided at the end of the article.

  • The Article Processing Charge was funded by the NINR.

  • Received July 8, 2020.
  • Accepted in final form August 25, 2020.

This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License 4.0 (CC BY-NC-ND), which permits downloading and sharing the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.

References

  1. 1.↵
    1. Waubant E,
    2. Lucas R,
    3. Mowry E, et al
    . Environmental and genetic risk factors for MS: an integrated review. Ann Clin Transl Neurol 2019;6:1905–1922.
    OpenUrl
  2. 2.↵
    1. Nelson LM,
    2. Wallin MT,
    3. Marrie RA, et al
    . A new way to estimate neurologic disease prevalence in the United States: illustrated with MS. Neurology 2019;92:469–480.
    OpenUrl
  3. 3.↵
    1. Wallin MT,
    2. Culpepper WJ,
    3. Nichols E, et al
    . Global, regional, and national burden of multiple sclerosis 1990–2016: a systematic analysis for the Global Burden of Disease Study 2016. Lancet Neurol 2019;18:269–285.
    OpenUrlPubMed
  4. 4.↵
    1. Houtchens MK,
    2. Edwards NC,
    3. Phillips AL
    . Relapses and disease-modifying drug treatment in pregnancy and live birth in US women with MS. Neurology 2018;91:E1570–E1578.
    OpenUrl
  5. 5.↵
    1. McCombe PA
    . The short and long-term effects of pregnancy on multiple sclerosis and experimental autoimmune encephalomyelitis. J Clin Med 2018;7:494.
    OpenUrl
  6. 6.↵
    1. Gauderman WJ,
    2. Mukherjee B,
    3. Aschard H, et al
    . Update on the state of the science for analytical methods for gene-environment interactions. Am J Epidemiol 2017;186:762–770.
    OpenUrlPubMed
  7. 7.↵
    1. Nielsen NM,
    2. Jørgensen KT,
    3. Stenager E, et al
    . Reproductive history and risk of multiple sclerosis. Epidemiology 2011;22:546–552.
    OpenUrlCrossRefPubMed
  8. 8.↵
    1. Hedström AK,
    2. Hillert J,
    3. Olsson T,
    4. Alfredsson L
    . Reverse causality behind the association between reproductive history and MS. 2014;20:406–411.
  9. 9.↵
    1. Khashan AS,
    2. Kenny LC,
    3. Laursen TM, et al
    . Pregnancy and the risk of autoimmune disease. PLoS One 2011;6:e19658.
    OpenUrlPubMed
  10. 10.↵
    1. Ponsonby A-L,
    2. Lucas RM,
    3. van der Mei IA, et al
    . Offspring number, pregnancy, and risk of a first clinical demyelinating event: the AusImmune Study. Neurology 2012;78:867–874.
    OpenUrlCrossRefPubMed
  11. 11.↵
    1. Magyari M,
    2. Koch-Henriksen N,
    3. Pfleger CC,
    4. Sørensen PS
    . Reproduction and the risk of multiple sclerosis. Mult Scler J 2013;19:1604–1609.
    OpenUrl
  12. 12.↵
    1. Alonso Á,
    2. Jick SS,
    3. Olek MJ,
    4. Ascherio A,
    5. Jick H,
    6. Hernán MA
    . Recent use of oral contraceptives and the risk of multiple sclerosis. Arch Neurol 2005;62:1362–1365.
    OpenUrlCrossRefPubMed
  13. 13.↵
    1. Hernán MA,
    2. Hohol MJ,
    3. Olek MJ,
    4. Spiegelman D,
    5. Ascherio A
    . Oral contraceptives and the incidence of multiple sclerosis. Neurology 2000;55:848–853.
    OpenUrlCrossRefPubMed
  14. 14.↵
    1. Leibowitz U,
    2. Antonovsky A,
    3. Kats R,
    4. Alter M
    . Does pregnancy increase the risk of multiple sclerosis?. J Neurol Neurosurg Psychiatry 1967;30:354–357.
    OpenUrlFREE Full Text
  15. 15.↵
    1. Runmarker B,
    2. Andersen O
    . Pregnancy is associated with a lower risk of onset and a better prognosis in multiple sclerosis. Brain 1995;118:253–261.
    OpenUrlCrossRefPubMed
  16. 16.↵
    1. Villard-Mackintosh L,
    2. Vessey MP
    . Oral contraceptives and reproductive factors in multiple sclerosis incidence. Contraception 1993;47:161–168.
    OpenUrlCrossRefPubMed
  17. 17.↵
    1. Briggs FBS,
    2. Acuña BS,
    3. Shen L, et al
    . Adverse socioeconomic position during the life course is associated with multiple sclerosis. J Epidemiol Community Heal 2014;68:622–629.
    OpenUrlAbstract/FREE Full Text
  18. 18.↵
    1. Langer-Gould A,
    2. Brara SM,
    3. Beaber BE,
    4. Zhang JL
    . Incidence of multiple sclerosis in multiple racial and ethnic groups. Neurology 2013;80:1734–1739.
    OpenUrlCrossRefPubMed
  19. 19.↵
    1. Hedström AK,
    2. Hillert J,
    3. Olsson T,
    4. Alfredsson L
    . Alcohol as a modifiable lifestyle factor Affecting multiple sclerosis risk. JAMA Neurol 2014;71:300.
    OpenUrl
  20. 20.↵
    1. Gustavsen MW,
    2. Page CM,
    3. Moen SM, et al
    . Environmental exposures and the risk of multiple sclerosis investigated in a Norwegian case-control study. BMC Neurol 2014;14:196.
    OpenUrl
  21. 21.↵
    1. Polman CH,
    2. Reingold SC,
    3. Edan G, et al
    . Diagnostic criteria for multiple sclerosis: 2005 revisions to the “McDonald Criteria”. Ann Neurol 2005;58:840–846.
    OpenUrlCrossRefPubMed
  22. 22.↵
    1. Polman CH,
    2. Reingold SC,
    3. Banwell B, et al
    . Diagnostic criteria for multiple sclerosis: 2010 Revisions to the McDonald criteria. Ann Neurol 2011;69:292–302.
    OpenUrlCrossRefPubMed
  23. 23.↵
    1. Banda Y,
    2. Kvale MN,
    3. Hoffmann TJ, et al
    . Characterizing race/ethnicity and genetic ancestry for 100000 subjects in the genetic epidemiology research on adult health and aging (GERA) cohort. Genetics 2015;200:1285–1295.
    OpenUrlAbstract/FREE Full Text
  24. 24.↵
    1. Krieger N
    . Overcoming the absence of socioeconomic data in medical records: validation and application of a census-based methodology. Am J Public Heal 1992;82:703–710.
    OpenUrl
  25. 25.↵
    1. Koebnick C,
    2. Langer-Gould AM,
    3. Gould MK, et al
    . Sociodemographic characteristics of members of a large, integrated health care system: comparison with US Census Bureau data. Perm J 2012;16:37–41.
    OpenUrlCrossRefPubMed
  26. 26.↵
    1. Chen CY,
    2. Pollack S,
    3. Hunter DJ,
    4. Hirschhorn JN,
    5. Kraft P,
    6. Price AL
    . Improved ancestry inference using weights from external reference panels. Bioinformatics 2013;29:1399–1406.
    OpenUrlCrossRefPubMed
  27. 27.↵
    1. Price AL,
    2. Patterson NJ,
    3. Plenge RM,
    4. Weinblatt ME,
    5. Shadick NA,
    6. Reich D
    . Principal components analysis corrects for stratification in genome-wide association studies. Nat Genet Nat Genet 2006;38:904–909.
    OpenUrl
  28. 28.↵
    1. Howie B,
    2. Marchini J,
    3. Stephens M
    . Genotype imputation with thousands of genomes. G3 (Bethesda) 2011;1:457–470.
    OpenUrlPubMed
  29. 29.↵
    1. Jia X,
    2. Han B,
    3. Onengut-Gumuscu S, et al
    . Imputing amino acid polymorphisms in human leukocyte antigens. PLoS One 2013;8:e64683.
    OpenUrlCrossRefPubMed
  30. 30.↵
    International Multiple Sclerosis Genetics Consortium. Multiple sclerosis genomic map implicates peripheral immune cells and microglia in susceptibility. Science 2019;365:eaav7188.
    OpenUrlAbstract/FREE Full Text
  31. 31.↵
    1. Moutsianas L,
    2. Jostins L,
    3. Beecham AH, et al
    . Class II HLA interactions modulate genetic risk for multiple sclerosis. Nat Genet 2015;47:1107–1113.
    OpenUrlCrossRefPubMed
  32. 32.↵
    1. Clarke GM,
    2. Morris AP
    . A Comparison of sample size and power in case-only association studies of gene-environment interaction. Am J Epidemiol 2010;171:498–505.
    OpenUrlCrossRefPubMed
  33. 33.↵
    1. Benjamini Y,
    2. Hochberg Y
    . Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Ser B 1995;57:289–300.
    OpenUrl
  34. 34.↵
    1. Kinder JM,
    2. Stelzer IA,
    3. Arck PC,
    4. Way SS
    . Immunological implications of pregnancy-induced microchimerism. Nat Rev Immunol 2017;17:483–494.
    OpenUrlPubMed
  35. 35.↵
    1. Arck PC,
    2. Hecher K
    . Fetomaternal immune cross-talk and its consequences for maternal and offspring's health. Nat Med 2013;19:548–556.
    OpenUrlCrossRefPubMed
  36. 36.↵
    1. Patas K,
    2. Engler JB,
    3. Friese MA,
    4. Gold SM
    . Pregnancy and multiple sclerosis: feto-maternal immune cross talk and its implications for disease activity. J Reprod Immunol 2013;97:140–146.
    OpenUrlCrossRefPubMed
  37. 37.↵
    1. Langer-Gould A,
    2. Smith JB,
    3. Hellwig K, et al
    . Breastfeeding, ovulatory years, and risk of multiple sclerosis. Neurology 2017;89:563–569.
    OpenUrl
  38. 38.↵
    1. Dahl J,
    2. Myhr KM,
    3. Daltveit AK,
    4. Gilhus NE
    . Pregnancy, delivery and birth outcome in different stages of maternal multiple sclerosis. J Neurol 2008;255:623–627.
    OpenUrlCrossRefPubMed
  39. 39.↵
    1. Langer-Gould AM
    . Pregnancy and family planning in multiple sclerosis. Continuum (Minneap Minn) 2019;25:773–792.
    OpenUrl
  40. 40.↵
    1. Houtchens MK,
    2. Edwards NC,
    3. Schneider G,
    4. Stern K,
    5. Phillips AL
    . Pregnancy rates and outcomes in women with and without MS in the United States. Neurology 2018;91:E1559–E1569.
    OpenUrl

Letters: Rapid online correspondence

No comments have been published for this article.
Comment

REQUIREMENTS

If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org

Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.

If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.

Submission specifications:

  • Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
  • Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
  • Submit only on articles published within 6 months of issue date.
  • Do not be redundant. Read any comments already posted on the article prior to submission.
  • Submitted comments are subject to editing and editor review prior to posting.

More guidelines and information on Disputes & Debates

Compose Comment

More information about text formats

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.
Author Information
NOTE: The first author must also be the corresponding author of the comment.
First or given name, e.g. 'Peter'.
Your last, or family, name, e.g. 'MacMoody'.
Your email address, e.g. higgs-boson@gmail.com
Your role and/or occupation, e.g. 'Orthopedic Surgeon'.
Your organization or institution (if applicable), e.g. 'Royal Free Hospital'.
Publishing Agreement
NOTE: All authors, besides the first/corresponding author, must complete a separate Publishing Agreement Form and provide via email to the editorial office before comments can be posted.
CAPTCHA
This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.

Vertical Tabs

You May Also be Interested in

Back to top
Advertisement

Related Articles

  • No related articles found.

Topics Discussed

Alert Me

Recommended articles