Association Between Rotavirus Vaccination and Type 1 Diabetes in Children | Pediatrics | JAMA Pediatrics | JAMA Network
[Skip to Navigation]
Sign In
Table 1.  Characteristics of a Cohort of Children Fully Exposed, Partially Exposed, or Unexposed to Rotavirus Vaccination
Characteristics of a Cohort of Children Fully Exposed, Partially Exposed, or Unexposed to Rotavirus Vaccination
Table 2.  Hazard Ratios and 95% CIs for Developing Type 1 Diabetes
Hazard Ratios and 95% CIs for Developing Type 1 Diabetes
1.
American Diabetes Association.  Diagnosis and classification of diabetes mellitus.   Diabetes Care. 2014;37(suppl 1):S81-S90. doi:10.2337/dc14-S081 PubMedGoogle ScholarCrossref
2.
Rogers  MAM, Kim  C, Banerjee  T, Lee  JM.  Fluctuations in the incidence of type 1 diabetes in the United States from 2001 to 2015: a longitudinal study.   BMC Med. 2017;15(1):199. doi:10.1186/s12916-017-0958-6 PubMedGoogle ScholarCrossref
3.
Rewers  M, Ludvigsson  J.  Environmental risk factors for type 1 diabetes.   Lancet. 2016;387(10035):2340-2348. doi:10.1016/S0140-6736(16)30507-4 PubMedGoogle ScholarCrossref
4.
Honeyman  M.  How robust is the evidence for viruses in the induction of type 1 diabetes?   Curr Opin Immunol. 2005;17(6):616-623. doi:10.1016/j.coi.2005.09.007 PubMedGoogle ScholarCrossref
5.
Honeyman  MC, Laine  D, Zhan  Y, Londrigan  S, Kirkwood  C, Harrison  LC.  Rotavirus infection induces transient pancreatic involution and hyperglycemia in weanling mice.   PLoS One. 2014;9(9):e106560. doi:10.1371/journal.pone.0106560 PubMedGoogle Scholar
6.
Pane  JA, Webster  NL, Coulson  BS.  Rotavirus activates lymphocytes from non-obese diabetic mice by triggering toll-like receptor 7 signaling and interferon production in plasmacytoid dendritic cells.   PLoS Pathog. 2014;10(3):e1003998. doi:10.1371/journal.ppat.1003998 PubMedGoogle Scholar
7.
Pane  JA, Webster  NL, Zufferey  C, Coulson  BS.  Rotavirus acceleration of murine type 1 diabetes is associated with increased MHC class I-restricted antigen presentation by B cells and elevated proinflammatory cytokine expression by T cells.   Virus Res. 2014;179:73-84. doi:10.1016/j.virusres.2013.11.009 PubMedGoogle ScholarCrossref
8.
Graham  KL, O’Donnell  JA, Tan  Y,  et al.  Rotavirus infection of infant and young adult nonobese diabetic mice involves extraintestinal spread and delays diabetes onset.   J Virol. 2007;81(12):6446-6458. doi:10.1128/JVI.00205-07 PubMedGoogle ScholarCrossref
9.
Honeyman  MC, Coulson  BS, Stone  NL,  et al.  Association between rotavirus infection and pancreatic islet autoimmunity in children at risk of developing type 1 diabetes.   Diabetes. 2000;49(8):1319-1324. doi:10.2337/diabetes.49.8.1319 PubMedGoogle ScholarCrossref
10.
Snell-Bergeon  JK, Smith  J, Dong  F,  et al.  Early childhood infections and the risk of islet autoimmunity: the Diabetes Autoimmunity Study in the Young (DAISY).   Diabetes Care. 2012;35(12):2553-2558. doi:10.2337/dc12-0423 PubMedGoogle ScholarCrossref
11.
Perrett  KP, Jachno  K, Nolan  TM, Harrison  LC.  Association of rotavirus vaccination with the incidence of type 1 diabetes in children.   JAMA Pediatr. 2019;173(3):280-282. doi:10.1001/jamapediatrics.2018.4578 PubMedGoogle ScholarCrossref
12.
Rogers  MAM, Basu  T, Kim  C.  Lower incidence rate of type 1 diabetes after receipt of the rotavirus vaccine in the United States, 2001–2017.   Sci Rep. 2019;9(1):7727. doi:10.1038/s41598-019-44193-4 PubMedGoogle ScholarCrossref
13.
Baggs  J, Gee  J, Lewis  E,  et al.  The Vaccine Safety Datalink: a model for monitoring immunization safety.   Pediatrics. 2011;127(suppl 1):S45-S53. doi:10.1542/peds.2010-1722H PubMedGoogle ScholarCrossref
14.
Mullooly  J, Drew  L, DeStefano  F,  et al.  Quality of HMO vaccination databases used to monitor childhood vaccine safety: Vaccine Safety DataLink Team.   Am J Epidemiol. 1999;149(2):186-194. doi:10.1093/oxfordjournals.aje.a009785 PubMedGoogle ScholarCrossref
15.
McNeil  MM, Gee  J, Weintraub  ES,  et al.  The Vaccine Safety Datalink: successes and challenges monitoring vaccine safety.   Vaccine. 2014;32(42):5390-5398. doi:10.1016/j.vaccine.2014.07.073 PubMedGoogle ScholarCrossref
16.
Ezeanolue  E, Harriman  K, Hunter  P, Kroger  A, Pellegrini  C. General best practice guidelines for immunization: best practices guidance of the Advisory Committee on Immunization Practices (ACIP). https://www.cdc.gov/vaccines/hcp/acip-recs/general-recs/downloads/general-recs.pdf. Accessed July 2, 2019.
17.
Robinson  CL, Bernstein  H, Romero  JR, Szilagyi  P.  Advisory Committee on Immunization Practices recommended immunization schedule for children and adolescents aged 18 years or younger—United States, 2019.   MMWR Morb Mortal Wkly Rep. 2019;68(5):112-114. doi:10.15585/mmwr.mm6805a4 PubMedGoogle ScholarCrossref
18.
Glanz  JM, Newcomer  SR, Narwaney  KJ,  et al.  A population-based cohort study of undervaccination in 8 managed care organizations across the United States.   JAMA Pediatr. 2013;167(3):274-281. doi:10.1001/jamapediatrics.2013.502 PubMedGoogle ScholarCrossref
19.
Lieu  TA, Ray  GT, Klein  NP, Chung  C, Kulldorff  M.  Geographic clusters in underimmunization and vaccine refusal.   Pediatrics. 2015;135(2):280-289. doi:10.1542/peds.2014-2715 PubMedGoogle ScholarCrossref
20.
Smith  PJ, Chu  SY, Barker  LE.  Children who have received no vaccines: who are they and where do they live?   Pediatrics. 2004;114(1):187-195. doi:10.1542/peds.114.1.187 PubMedGoogle ScholarCrossref
21.
Gopal  SH, Edwards  KM, Creech  B, Weitkamp  J-H.  Variability in immunization practices for preterm infants.   Am J Perinatol. 2018;35(14):1394-1398. doi:10.1055/s-0038-1660453 PubMedGoogle ScholarCrossref
22.
Lawrence  JM, Black  MH, Zhang  JL,  et al.  Validation of pediatric diabetes case identification approaches for diagnosed cases by using information in the electronic health records of a large integrated managed health care organization.   Am J Epidemiol. 2014;179(1):27-38. doi:10.1093/aje/kwt230 PubMedGoogle ScholarCrossref
23.
McKinney  PA, Parslow  R, Gurney  KA, Law  GR, Bodansky  HJ, Williams  R.  Perinatal and neonatal determinants of childhood type 1 diabetes: a case-control study in Yorkshire, UK.   Diabetes Care. 1999;22(6):928-932. doi:10.2337/diacare.22.6.928 PubMedGoogle ScholarCrossref
24.
Dahlquist  GG, Patterson  C, Soltesz  G.  Perinatal risk factors for childhood type 1 diabetes in Europe: the EURODIAB Substudy 2 Study Group.   Diabetes Care. 1999;22(10):1698-1702. doi:10.2337/diacare.22.10.1698 PubMedGoogle ScholarCrossref
25.
Stene  LC, Magnus  P, Lie  RT, Søvik  O, Joner  G; Norwegian childhood Diabetes Study Group.  Birth weight and childhood onset type 1 diabetes: population based cohort study.   BMJ. 2001;322(7291):889-892. doi:10.1136/bmj.322.7291.889 PubMedGoogle ScholarCrossref
26.
Janghorbani  M, Van Dam  RM, Willett  WC, Hu  FB.  Systematic review of type 1 and type 2 diabetes mellitus and risk of fracture.   Am J Epidemiol. 2007;166(5):495-505. doi:10.1093/aje/kwm106 PubMedGoogle ScholarCrossref
27.
Dabelea  D, Mayer-Davis  EJ, Saydah  S,  et al; SEARCH for Diabetes in Youth Study.  Prevalence of type 1 and type 2 diabetes among children and adolescents from 2001 to 2009.   JAMA. 2014;311(17):1778-1786. doi:10.1001/jama.2014.3201 PubMedGoogle ScholarCrossref
28.
Harder  T, Roepke  K, Diller  N, Stechling  Y, Dudenhausen  JW, Plagemann  A.  Birth weight, early weight gain, and subsequent risk of type 1 diabetes: systematic review and meta-analysis.   Am J Epidemiol. 2009;169(12):1428-1436. doi:10.1093/aje/kwp065 PubMedGoogle ScholarCrossref
29.
Svensson  J, Carstensen  B, Mortensen  HB, Borch-Johnsen  K; Danish Study Group of Childhood Diabetes.  Early childhood risk factors associated with type 1 diabetes—is gender important?   Eur J Epidemiol. 2005;20(5):429-434. doi:10.1007/s10654-005-0878-1 PubMedGoogle ScholarCrossref
30.
Cardwell  CR, Stene  LC, Joner  G,  et al.  Caesarean section is associated with an increased risk of childhood-onset type 1 diabetes mellitus: a meta-analysis of observational studies.   Diabetologia. 2008;51(5):726-735. doi:10.1007/s00125-008-0941-z PubMedGoogle ScholarCrossref
31.
Daley  MF, Shoup  JA, Newcomer  SR,  et al.  Assessing potential confounding and misclassification bias when studying the safety of the childhood immunization schedule.   Acad Pediatr. 2018;18(7):754-762. doi:10.1016/j.acap.2018.03.007 PubMedGoogle ScholarCrossref
32.
Frederiksen  B, Kroehl  M, Lamb  MM,  et al.  Infant exposures and development of type 1 diabetes mellitus: the Diabetes Autoimmunity Study in the Young (DAISY).   JAMA Pediatr. 2013;167(9):808-815. doi:10.1001/jamapediatrics.2013.317 PubMedGoogle ScholarCrossref
33.
Rosenbauer  J, Herzig  P, Giani  G.  Early infant feeding and risk of type 1 diabetes mellitus—a nationwide population-based case-control study in pre-school children.   Diabetes Metab Res Rev. 2008;24(3):211-222. doi:10.1002/dmrr.791 PubMedGoogle ScholarCrossref
34.
Schroeder  EB, Donahoo  WT, Goodrich  GK, Raebel  MA.  Validation of an algorithm for identifying type 1 diabetes in adults based on electronic health record data.   Pharmacoepidemiol Drug Saf. 2018;27(10):1053-1059. doi:10.1002/pds.4377 PubMedGoogle ScholarCrossref
35.
Allison  PD.  Survival Analysis Using SAS: A Practical Guide. 2nd ed. Cary, NC: SAS Institute, Inc; 2010.
36.
Moons  KG, Donders  RA, Stijnen  T, Harrell  FE  Jr.  Using the outcome for imputation of missing predictor values was preferred.   J Clin Epidemiol. 2006;59(10):1092-1101. doi:10.1016/j.jclinepi.2006.01.009 PubMedGoogle ScholarCrossref
37.
Lash  TL, Fox  MP, Fink  AK.  Applying Quantitative Bias Analysis to Epidemiologic Data. New York, NY: Springer Science & Business Media; 2011.
38.
Lash  TL, Abrams  B, Bodnar  LM.  Comparison of bias analysis strategies applied to a large data set.   Epidemiology. 2014;25(4):576-582. doi:10.1097/EDE.0000000000000102 PubMedGoogle ScholarCrossref
39.
Burnett  E, Yen  C, Tate  JE, Parashar  UD.  Rotavirus vaccines: current global impact and future perspectives.   Future Virol. 2016;11(10):699-708. doi:10.2217/fvl-2016-0082 PubMedGoogle ScholarCrossref
40.
Leshem  E, Tate  JE, Steiner  CA, Curns  AT, Lopman  BA, Parashar  UD.  National estimates of reductions in acute gastroenteritis–related hospitalizations and associated costs in US children after implementation of rotavirus vaccines.   J Pediatric Infect Dis Soc. 2018;7(3):257-260. doi:10.1093/jpids/pix057 PubMedGoogle ScholarCrossref
41.
Vaarala  O, Jokinen  J, Lahdenkari  M, Leino  T.  Rotavirus vaccination and the risk of celiac disease or type 1 diabetes in Finnish children at early life.   Pediatr Infect Dis J. 2017;36(7):674-675. doi:10.1097/INF.0000000000001600 PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    March 9, 2020

    Association Between Rotavirus Vaccination and Type 1 Diabetes in Children

    Author Affiliations
    • 1Institute for Health Research, Kaiser Permanente Colorado, Aurora
    • 2Department of Epidemiology, Colorado School of Public Health, Aurora
    • 3Department of Endocrinology, Parkview Health and Parkview Physicians Group, Fort Wayne, Indiana
    • 4Kaiser Permanente Department of Research and Evaluation, Kaiser Permanente of Southern California, Pasadena
    • 5Marshfield Clinic Research Institute, Marshfield Clinic Health System, Marshfield, Wisconsin
    • 6Division of Research, HealthPartners Institute, Minneapolis, Minnesota
    • 7Kaiser Permanente Division of Research, Kaiser Permanente of Northern California, Oakland
    • 8Immunization Safety Office, Vaccine Safety Datalink, Centers for Disease Control and Prevention, Atlanta, Georgia
    JAMA Pediatr. 2020;174(5):455-462. doi:10.1001/jamapediatrics.2019.6324
    Key Points

    Question  Is rotavirus vaccination associated with type 1 diabetes in children?

    Findings  In this cohort study with a median follow-up of 5.4 years (interquartile range, 3.8-7.8 years), 360 169 children were exposed to the full series of rotavirus vaccination, 15 765 children were partially exposed to rotavirus vaccination, and 11 003 children were unexposed to rotavirus vaccination. The incidence of type 1 diabetes was not significantly different across the rotavirus vaccine exposure groups.

    Meaning  Rotavirus vaccination does not appear to be associated with type 1 diabetes in children.

    Abstract

    Importance  Because rotavirus infection is a hypothesized risk factor for type 1 diabetes, live attenuated rotavirus vaccination could increase or decrease the risk of type 1 diabetes in children.

    Objective  To examine whether there is an association between rotavirus vaccination and incidence of type 1 diabetes in children aged 8 months to 11 years.

    Design, Setting, and Participants  A retrospective cohort study of 386 937 children born between January 1, 2006, and December 31, 2014, was conducted in 7 US health care organizations of the Vaccine Safety Datalink. Eligible children were followed up until a diagnosis of type 1 diabetes, disenrollment, or December 31, 2017.

    Exposures  Rotavirus vaccination for children aged 2 to 8 months. Three exposure groups were created. The first group included children who received all recommended doses of rotavirus vaccine by 8 months of age (fully exposed to rotavirus vaccination). The second group had received some, but not all, recommended rotavirus vaccines (partially exposed to rotavirus vaccination). The third group did not receive any doses of rotavirus vaccines (unexposed to rotavirus vaccination).

    Main Outcomes and Measures  Incidence of type 1 diabetes among children aged 8 months to 11 years. Type 1 diabetes was identified by International Classification of Diseases codes: 250.x1, 250.x3, or E10.xx in the outpatient setting. Cox proportional hazards regression models were used to analyze time to type 1 diabetes incidence from 8 months to 11 years. Hazard ratios and 95% CIs were calculated. Models were adjusted for sex, race/ethnicity, birth year, mother’s age, birth weight, gestational age, number of well-child visits, and Vaccine Safety Datalink site.

    Results  In a cohort of 386 937 children (51.1% boys and 41.9% non-Hispanic white), 360 169 (93.1%) were fully exposed to rotavirus vaccination, 15 765 (4.1%) were partially exposed to rotavirus vaccination, and 11 003 (2.8%) were unexposed to rotavirus vaccination. Children were followed up a median of 5.4 years (interquartile range, 3.8-7.8 years). The total person-time follow-up in the cohort was 2 253 879 years. There were 464 cases of type 1 diabetes in the cohort, with an incidence rate of 20.6 cases per 100 000 person-years. Compared with children unexposed to rotavirus vaccination, the adjusted hazard ratio was 1.03 (95% CI, 0.62-1.72) for children fully exposed to rotavirus vaccination and 1.50 (95% CI, 0.81-2.77) for children partially exposed to rotavirus vaccination.

    Conclusions and Relevance  The findings of this study suggest that rotavirus vaccination does not appear to be associated with type 1 diabetes in children.

    Introduction

    Type 1 diabetes is an autoimmune disease resulting from the destruction of insulin-producing pancreatic islet B cells.1 Type 1 diabetes is diagnosed primarily during childhood, with a peak incidence between 10 and 14 years of age.2 In the United States, the incidence of type 1 diabetes increased by 1.9% per year from 2001 to 2015.2

    Although type 1 diabetes tends to occur in genetically susceptible individuals, environmental factors also play a role in its development. Proposed environmental triggers include the timing of introduction of solid foods during infancy, low birth weight and infant growth, psychological stress, and certain viral infections.3 Possible mechanisms by which viruses can induce type 1 diabetes are molecular mimicry, changes to the gut epithelium, pancreatic infection, and other interactions with the innate immune system.3-5

    In mouse models, rotavirus infection has been shown to accelerate the development of type 1 diabetes in conjunction with an increased antirotavirus antibody response among genetically susceptible mice.6-8 Concurrent increases in islet autoantibodies and rotavirus antibodies have been observed in children at risk for diabetes,9 and epidemiologic data suggest an association between gastrointestinal infection and incidence of type 1 diabetes in children followed up from birth to 10 years of age.10

    Given these findings, it is biologically plausible that a live, attenuated rotavirus vaccine could either increase or decrease the risk for type 1 diabetes in early childhood. In the United States, there are 2 licensed rotavirus vaccines: RotaTeq (a pentavalent rotavirus vaccine) and Rotarix (a monovalent rotavirus vaccine). RotaTeq has been routinely administered at 2, 4, and 6 months of age since 2006, while Rotarix was licensed in 2008 as a 2-dose series to be administered at 2 and 4 months of age. Two studies—an ecologic analysis in Australia11 and an analysis of a medical claims–based database in the United States12—reported reduced incidences of type 1 diabetes among young children vaccinated against rotavirus. The objective of this study was to examine the association between rotavirus vaccination and incidence of type 1 diabetes in a large, US-based cohort of children, controlling for several known risk factors for type 1 diabetes.

    Methods
    Setting and Study Cohort

    We conducted a retrospective cohort study of children enrolled in 7 integrated health care organizations that participate in the Vaccine Safety Datalink (VSD). The VSD is a research collaboration led by the Centers for Disease Control and Prevention that uses electronic health record databases to conduct epidemiological studies of vaccine safety.13 Vaccine Safety Datalink sites are located in Northern California, Southern California, Minnesota, Colorado, Oregon, Washington, and Wisconsin. Each site creates standardized data sets with the following information: demographic characteristics, membership enrollment, vaccination history, and medical encounters in the emergency department and inpatient and outpatient settings. Vaccination data in the VSD are derived from the electronic health record and have a high degree of accuracy.14,15 Institutional review boards at each site (Kaiser Permanente Colorado, Kaiser Permanente of Northern California, Kaiser Permanente of Southern California, Kaiser Permanente Northwest, Kaiser Permanente Washington, HealthPartners Institute, and Marshfield Clinic Health System) approved this study and granted a waiver of the Health Insurance Portability and Accountability Act and parental consent, as the data-only research activities were determined to pose minimal risk.

    Vaccine Safety Datalink data sets were used to identify children born between January 1, 2006, and December 31, 2014. Children were included if they had continuous health plan enrollment from 6 weeks to 2 years of age. Requiring continuous enrollment for cohort inclusion helps ensure an accurate and complete capture of exposure, covariate, and outcome data. Children with a medical contraindication to vaccination16 or fewer than 2 well-child visits by 12 months of age were excluded. Eligible children were followed up until a type 1 diabetes diagnosis, disenrollment, or December 31, 2017.

    Three exposure groups were created from the cohort. The first group included children who were fully vaccinated with all vaccines by 24 months of age according to the recommendations of the Advisory Committee on Immunization Practices.17 These children had received all recommended doses of a rotavirus vaccine (fully exposed to rotavirus vaccination) by 8 months of age. Children could have received Rotarix, RotaTeq, or a combination of the 2 vaccines. Children who received a combination of the 2 vaccines were considered fully vaccinated if they received 2 doses of RotaTeq and 1 dose of Rotarix, or 2 doses of Rotarix and 1 dose of RotaTeq. The second group represented children who had received some, but not all, recommended rotavirus vaccine doses (partially exposed to rotavirus vaccination). These children had received all other recommended childhood vaccines by 24 months of age. The third group comprised children who had not received any doses of the rotavirus vaccine but had received all other recommended vaccines by 24 months of age (unexposed to rotavirus vaccination).

    We restricted the unexposed group to children who had received all recommended vaccines except rotavirus vaccination to ensure that the association of rotavirus vaccination with the primary outcome could be isolated when comparing the incidence of type 1 diabetes between the exposed and unexposed groups. This restriction was also done to minimize potential biases that could be introduced by including undervaccinated children who were missing multiple vaccines because undervaccinated children may differ from fully vaccinated children by potential confounding factors such as health care–seeking behavior, race/ethnicity, chronic disease status, and socioeconomic status.18-21

    Diabetes Cases

    In the cohort, incident type 1 diabetes was identified by the following International Classification of Diseases, Ninth Revision (ICD-9) and International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, Clinical Modification (ICD-10-CM) codes: 250.x1, 250.x3, or E10.xx. Children with one or more of these ICD-9 or ICD-10-CM codes in the outpatient setting were classified as having type 1 diabetes. The first occurrence of a code represented the type 1 diabetes incidence date. Prior research has shown that this approach identified diagnosed cases of type 1 diabetes with high levels of accuracy (94.1%), sensitivity (94.8%), specificity (93.3%), and area under the receiver operating characteristic curve (94.1%).22 Children with type 1 diabetes diagnosed after 8 months of age were included in the analyses.

    To examine accuracy, we conducted a manual medical record review across 6 of the VSD sites to verify the ICD-9 and ICD-10-CM codes for type 1 diabetes. One of the participating sites reviewed a sample of their records, while the other 5 sites reviewed all of their records. For the review, abstractors examined the medical records of the cases to verify that the child had received a diagnosis of type 1 diabetes.

    Covariates

    The following reported risk factors for type 1 diabetes2,23-30 were included in the analysis: sex, race/ethnicity, birth year, mother’s age, birth weight, and gestational age. The analysis was also adjusted for VSD site and the number of well-child visits between 30 days and 2 years of age because these potential confounding variables may be associated with vaccination status and receipt of a type 1 diabetes diagnosis.

    Two potential risk factors for type 1 diabetes were not consistently captured in the VSD databases: breastfeeding and family history of type 1 diabetes. If also associated with vaccination status, these potential unmeasured confounding variables could bias the results. We assessed their association with the results using a sensitivity analysis based on prevalence estimates of the potential confounding variables by vaccination status and estimates of the strengths of associations between the variables and incidence of type 1 diabetes. These estimates were obtained in different ways for the 2 variables.

    For breastfeeding, we used estimates from published observational studies. The association between breastfeeding and vaccination status was based on a multisite survey conducted within the VSD, showing that reported rates of ever having breastfed were approximately 98.9% (95% CI, 97.5%-100%) among parents who refused or delayed vaccines for their children and 90.1% (95% CI, 84.1%-96.1%) among parents who accepted all vaccines for their children.31 For the association between breastfeeding and incidence of type 1 diabetes, we used results from 2 epidemiological studies. The first study was a US-based cohort analysis showing that genetically susceptible children who were breastfed at the time of introduction to wheat or barley were less likely to develop type 1 diabetes compared with a similar group of children who were not breastfed at the time of introduction to wheat or barley (hazard ratio [HR], 0.47; 95% CI, 0.26-0.86).32 The second study was a case-control analysis showing that children with incident cases of type 1 diabetes in Germany were less likely to have been breastfed for 5 or more months compared with matched control children (odds ratio, 0.71; 95% CI, 0.54-0.93).33

    For family history of type 1 diabetes, we extracted additional data from 2 VSD sites (Kaiser Permanente Southern California and Kaiser Permanente Colorado) to create a cohort of children who could be linked electronically to their parents and older siblings. This family linkage was created using a membership identification variable from health plan enrollment data tables. A child was considered exposed to a family history of type 1 diabetes if a parent, an older sibling, or both a parent and an older sibling had received a diagnosis of type 1 diabetes.34 Using these data, we estimated the prevalence of family history of type 1 diabetes by rotavirus vaccination status, and we estimated the association between family history of type 1 diabetes and incidence of type 1 diabetes in children.

    Statistical Analysis

    We conducted a power analysis assuming the proportion of type 1 diabetes events was 0.002 in the exposed children and 0.001 in the unexposed children. Based on the observed sizes of exposed and unexposed groups and an α of .05, the study could detect an HR of 2.1 with 80% power. Power analyses were conducted using PASS, version 15 software (NCSS Statistical Software; https://www.ncss.com/).

    Descriptive statistics were reported on 8 covariates to examine their distributions across the cohort. Cohort data were analyzed with Cox proportional hazards regression models to examine associations between rotavirus vaccination exposure and incidence of type 1 diabetes, controlling for 8 covariates. A complete-case analysis was conducted on observations without missing data. Adjusted HRs (aHRs) and 95% CIs were calculated. In all models, the referent group was children unexposed to rotavirus vaccination. A key assumption of Cox proportional hazards regression that is the ratio of the hazards function for any 2 individuals is constant across time. The proportional hazards assumption was evaluated with Schoenfeld residual plots, a global goodness-of-fit test, and supremum tests for each independent variable.35 All statistical tests were 2-sided, and the results were deemed statistically significant at P < .05.

    As secondary analyses, the data were stratified by age (from 8 months to 4 years of age or from 5 to 10 years of age) and by type of rotavirus vaccine received (RotaTeq or Rotarix). We also created and analyzed a second, expanded cohort that represented children across a range of vaccination exposures, from fully vaccinated with all recommended vaccines to completely unvaccinated (no recommended vaccines). In this expanded cohort that included undervaccinated children, we compared the incidence of type 1 diabetes between children exposed to rotavirus vaccination and children unexposed to rotavirus vaccination, regardless of which other vaccines they had received.

    Sensitivity Analyses

    To account for missing data, we conducted a multiple imputation analysis using the fully conditional specification discriminant function for missing categorical data and the fully conditional specification regression to impute missing continuous data. The multiple imputation model included the same independent variables as the primary adjusted model. We also included whether a child developed type 1 diabetes in the multiple imputation analysis because models without the outcome would produce results biased toward the null.36 The data were imputed 10 times.

    To assess the potential for confounding by breastfeeding or family history of type 1 diabetes, we conducted 2 probabilistic bias analyses.37,38 A probabilistic bias analysis is a type of sensitivity analysis that evaluates the magnitude, direction, and uncertainty of bias by using Monte Carlo techniques to simulate bias parameters. For unmeasured confounding, probabilistic bias analysis requires initial estimates of the prevalence of the confounder in the exposed and unexposed groups and the strength of association between the confounder and outcome. These estimates are then sampled from a prespecified distribution in the simulation. Based on published data,32,33 we assumed a normal distribution for the log odds ratio between breastfeeding and type 1 diabetes and a uniform distribution for prevalence of breastfeeding for each exposure group. For family history of type 1 diabetes, we first used the family linkage cohort to create a prediction model for family history of type 1 diabetes given a child’s rotavirus vaccine exposure group and type 1 diabetes status. Through simulation, the prediction model was used to extrapolate this information to those in the larger cohort of children who had no information on family history of type 1 diabetes. We used the results from the probabilistic bias analyses to determine the robustness of our final conclusions in the presence of unmeasured confounding bias. Analyses were conducted with SAS, version 9.4 software (SAS Institute Inc).

    Results
    Study Cohort and Type 1 Diabetes Cases

    Between January 1, 2006, and December 31, 2014, there were 478 589 children born in the VSD with continuous health plan enrollment through 2 years of age. The following 91 652 children were excluded because of a medical contraindication to vaccination (n = 7816), neonatal diabetes or cystic fibrosis (n = 200), type 1 diabetes diagnosis before 8 months of age (n = 1), fewer than 2 well-child visits before 1 year of age (n = 8637), or undervaccination for nonrotavirus vaccines by 2 years of age (n = 74 998). After exclusions, the final study cohort comprised 386 937 children (eFigure in the Supplement). The cohort comprised 51.1% boys and 41.9% white non-Hispanic children, and the mean (SD) birth weight was 3338 (566) g. Children were followed up a median of 5.4 years (interquartile range, 3.8-7.8 years), for a total person-time follow-up of 2 253 879 years (Table 1).

    Of the 386 937 children in the cohort, 360 169 (93.1%) were fully exposed to rotavirus vaccination, 15 765 (4.1%) were partially exposed to rotavirus vaccination, and 11 003 (2.8%) were unexposed to rotavirus vaccination. The distributions of sex, race/ethnicity, mother’s age, birth weight, gestational age, and well-child visits were similar across vaccine exposure groups (Table 1). The distribution of birth year from 2006 to 2014 was skewed toward the earlier years for the partially exposed and unexposed groups, reflecting the gradual uptake of rotavirus vaccination.

    There were 464 cases of type 1 diabetes identified in the cohort, representing an incidence rate of 20.6 cases per 100 000 person-years. Children with type 1 diabetes included 221 girls (47.6%) and 275 non-Hispanic white children (59.3%), and they had a mean (SD) age at diagnosis of 4.4 (2.6) years (eTable 1 in the Supplement). There were 415 cases in the fully exposed group, 32 cases in the partially exposed group, and 17 cases in the unexposed group, representing incidence rates of 20.0 cases per 100 000 person-years in the fully exposed group, 31.2 cases per 100 000 person-years in the partially exposed group, and 22.4 cases per 100 000 person-years in the unexposed group (eTable 2 in the Supplement). Of the 91 652 children excluded, 148 had type 1 diabetes, for an incidence rate of 29.8 per 100 000 person-years.

    Medical Record Review

    A manual review was conducted on 307 records (66.2%) of the 464 children with type 1 diabetes. Of the 307 children, 170 (55.4%) received a diagnosis with ICD-9 codes, and 137 (44.6%) received a diagnosis with ICD-10-CM codes. Type 1 diabetes diagnoses were confirmed in 158 (92.9%) of the ICD-9 codes and 126 (92.0%) of the ICD-10-CM codes.

    Association Between Rotavirus Vaccination and Type 1 Diabetes

    Compared with children unexposed to rotavirus vaccination, the aHR for developing type 1 diabetes was 1.03 (95% CI, 0.62-1.72) for children fully exposed to rotavirus vaccination and 1.50 (95% CI, 0.81-2.77) for children partially exposed to rotavirus vaccination (Table 2). The proportional hazards assumption did not appear to be violated based on the Schoenfeld residual plots and global goodness-of-fit test (χ28 = 25.4; P = .75). However, supremum tests for birth season, gestational age, and birth weight indicated violations of the proportional hazards assumption. This finding was addressed by stratifying the analyses by birth season and by including interaction terms for gestational age and birth weight by time into the regression models.

    In the age-stratified analysis, aHRs for type 1 diabetes occurring in the 8-month to 4-year age group were 0.84 (95% CI, 0.42-1.67) for being fully exposed to rotavirus vaccination and 1.19 (95% CI, 0.51-2.79) for being partially exposed to rotavirus vaccination (Table 2). Adjusted HRs for type 1 diabetes occurring in the 5-year to 10-year age group were 1.27 (95% CI, 0.59-2.74) for being fully exposed to rotavirus vaccination and 1.92 (95% CI, 0.78-4.72) for being partially exposed to rotavirus vaccination.

    In the analysis stratified by type of rotavirus vaccine received, the aHR for developing type 1 diabetes was 0.98 (95% CI, 0.59-1.64) for children fully exposed to pentavalent rotavirus vaccination and 1.48 (95% CI, 0.79-2.74) for children partially exposed to pentavalent rotavirus vaccination (Table 2). The aHR was 2.03 (95% CI, 0.66-6.28) for children fully exposed to monovalent rotavirus vaccination and 2.81 (95% CI, 0.30-26.75) for children partially exposed to monovalent rotavirus vaccination.

    The expanded cohort included an additional 74 996 undervaccinated children, representing 81.8% of the 91 652 children who were excluded from the main cohort. In the expanded cohort, the aHR for developing type 1 diabetes was 0.94 (95% CI, 0.67-1.30) for children fully exposed to rotavirus vaccination and 1.32 (95% CI, 0.88-1.98) for children partially exposed to rotavirus vaccination (Table 2).

    Sensitivity Analyses for Missing Data

    Approximately 14.8% of the data were excluded owing to missing data on race/ethnicity (5.9%), mother’s age (8.7%), gestational age (9.0%), and/or birth weight (9.0%). One or more covariates were missing for 46 children (9.9%) with type 1 diabetes. After imputation of missing data, the aHR for developing type 1 diabetes was 1.13 (95% CI, 0.69-1.85) for children fully exposed to rotavirus vaccination and 1.53 (95% CI, 0.84-2.78) for children partially exposed to rotavirus vaccination.

    Sensitivity Analyses for Unmeasured Confounding

    For breastfeeding, the probabilistic bias analysis yielded aHRs for developing type 1 diabetes of 0.93 (95% CI, 0.56-1.55) in children fully exposed to rotavirus vaccination and 1.49 (95% CI, 0.81-2.76) in children partially exposed to rotavirus vaccination.

    For family history of type 1 diabetes, a cohort of 115 893 children were matched with their parents and any older siblings. Of these children, 1124 (1.0%) had at least 1 parent or older sibling with a type 1 diabetes diagnosis (eTable 3 in the Supplement). Families with a history of type 1 diabetes were 12.05 times as likely (95% CI, 6.91-21.03) as families without a history of type 1 diabetes to have a child in the cohort with a type 1 diabetes diagnosis. Children in the cohort with a family history of type 1 diabetes were less likely than children without a family history of type 1 diabetes to be exposed to rotavirus vaccination (odds ratio, 0.54; 95% CI, 0.38-0.75). The probabilistic bias analysis incorporated these estimates and yielded an aHR for developing type 1 diabetes of 1.08 (95% CI, 0.65-1.79) for children fully exposed to rotavirus vaccination and 1.55 (95% CI, 0.84-2.87) for children partially exposed to rotavirus vaccination.

    Discussion

    Our findings do not show an association between rotavirus vaccination and type 1 diabetes in children followed up for a median of 5.4 years. Since licensure, rotavirus vaccination has been associated with a reduction in morbidity and mortality due to rotavirus infection in the United States and worldwide.39 In the United States, acute gastroenteritis-related hospitalization rates decreased between 31% and 55% during the postvaccination years of 2008 to 2013, amounting to $1.2 billion in saved hospitalization costs.40 Given this clear public health benefit, these safety results support the continued, widespread administration of rotavirus vaccination to infants born in the United States.

    It has been hypothesized that rotavirus vaccination could reduce the risk of type 1 diabetes through the prevention of rotavirus infection, a potential environmental risk factor for type 1 diabetes.9,11 A Finnish cohort study of children followed up for 4 to 6 years found a nonsignificant negative association between rotavirus vaccination and incidence of type 1 diabetes (relative risk, 0.91; 95% CI, 0.69-1.20).41 Similarly, an Australian ecologic study showed a statistically significant protective association between rotavirus vaccination and incidence of type 1 diabetes in children 4 years or younger (incidence rate ratio, 0.86; 95% CI, 0.74-0.99).11 A cohort study of medical claims data observed a negative association between completion of the rotavirus vaccination series and incidence of type 1 diabetes (HR, 0.63; 95% CI, 0.50-0.78).12 This study had a median length of follow-up ranging from 2.08 to 2.91 years across exposure groups. In contrast, our cohort had a median length of follow-up ranging from approximately 5 to 7 years across exposure groups, and our analyses that adjusted for race/ethnicity, gestational age, birth weight, family history of type 1 diabetes, and breastfeeding yielded statistically nonsignificant associations, with HRs between 0.93 and 1.60. Future research using large, patient-level databases with ample follow-up and capture of known risk factors for type 1 diabetes could continue to evaluate the potential association of rotavirus vaccination with incidence of type 1 diabetes.

    Two potential unmeasured confounding variables may have affected our results: breastfeeding and family history of type 1 diabetes. Breastfeeding is negatively associated with incidence of type 1 diabetes and positively associated with undervaccination rates. Family history of type 1 diabetes is positively associated with incidence of type 1 diabetes and negatively associated with receipt of rotavirus vaccination. These associations suggest that a true negative or positive association between rotavirus vaccination and incidence of type 1 diabetes could have been obscured by confounding bias. To address this possibility, we conducted sensitivity analyses that did not markedly change the results, thus supporting our conclusions that rotavirus vaccination is not associated with the incidence of type 1 diabetes in children.

    Limitations

    This study has limitations. Although we analyzed a longitudinal, population-based cohort of more than 380 000 children, most cohort members were not followed up through 10 to 14 years of age, when the incidence of type 1 diabetes peaks. In addition, the group of children unexposed to rotavirus vaccination was relatively small (2.8% of the cohort). These limitations affected our study’s statistical power and may have hindered our ability to detect relative risk estimates below 2.1. It is also possible that a proportion of the unexposed cohort actually received rotavirus vaccines, either because the vaccines were not documented in the electronic health record or the vaccines were administered outside of the VSD. If rotavirus vaccination is truly associated with type 1 diabetes, then this type of exposure misclassification would have biased our result toward the null hypotheses.

    Conclusions

    Rotavirus vaccination is a highly effective vaccine routinely recommended for all infants by 8 months of age. In this large cohort study, we did not find evidence that rotavirus vaccination was associated with an increased or decreased incidence of type 1 diabetes in children. Although rotavirus vaccination may not prevent type 1 diabetes, these results should provide additional reassurance to the public that rotavirus vaccination can be safely administered to infants.

    Back to top
    Article Information

    Accepted for Publication: December 4, 2019.

    Corresponding Author: Jason M. Glanz, PhD, Institute for Health Research, Kaiser Permanente Colorado, 2505 S Parker Rd, Waterpark III, Ste 200, Aurora, CO 80014 (jason.m.glanz@kp.org).

    Published Online: March 9, 2020. doi:10.1001/jamapediatrics.2019.6324

    Author Contributions: Dr Glanz and Ms Clarke had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Glanz, Clarke, Daley, DeStefano.

    Acquisition, analysis, or interpretation of data: Glanz, Clarke, Xu, Daley, Shoup, Schroeder, Lewin, McClure, Kharbanda, Klein.

    Drafting of the manuscript: Glanz, Clarke.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Glanz, Clarke, Xu, Daley.

    Obtained funding: Glanz, Daley, Klein.

    Administrative, technical, or material support: Glanz, Daley, Shoup, McClure, Klein, DeStefano.

    Supervision: Glanz, Xu, Daley.

    Conflict of Interest Disclosures: Drs Glanz, Shoup, Lewin, Kharbanda, and Klein and Ms Clarke reported receiving grants from the Centers for Disease Control and Prevention during the conduct of the study. Dr Schroeder reported receiving grants from the National Institute of Diabetes and Digestive and Kidney Diseases during the conduct of the study. Dr Klein reported receiving grants from Merck & Co, GlaxoSmithKline, Sanofi Pasteur, Pfizer, Protein Science (now Sanofi Pasteur), Dynavax, and MedImmune outside the submitted work. No other disclosures were reported.

    Funding/Support: This research was funded by the Centers for Disease Control and Prevention through a Task Order (contract 200-2012-53582-TO 0001), issued as part of the Vaccine Safety Datalink project (contract 200-2012-53582).

    Role of the Funder/Sponsor: Centers for Disease Control and Prevention coauthor (Dr DeStefano) was involved in the design and conduct of the study; interpretation of the data; and review and approval of the manuscript.

    Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of Centers for Disease Control and Prevention.

    Additional Contributions: Umesh Parashar, MD, National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention; and Rachel M. Burke, PhD, and Jacqueline E. Tate, PhD, Rotavirus Epidemiology Team, Centers for Disease Control and Prevention, provided input toward the concept and design of the study. Holly Groom, MPH, Northwest Kaiser Permanente; and Lisa Jackson, MD, MPH, Kaiser Permanente Washington, contributed to the methods of the study. Sungching Glenn, MS, Kaiser Permanente Southern California, assisted with data contributions in the family history of diabetes analysis. John Steiner, MD, Kaiser Permanente Colorado, provided critical review of the draft manuscript. They were not compensated for their contributions.

    References
    1.
    American Diabetes Association.  Diagnosis and classification of diabetes mellitus.   Diabetes Care. 2014;37(suppl 1):S81-S90. doi:10.2337/dc14-S081 PubMedGoogle ScholarCrossref
    2.
    Rogers  MAM, Kim  C, Banerjee  T, Lee  JM.  Fluctuations in the incidence of type 1 diabetes in the United States from 2001 to 2015: a longitudinal study.   BMC Med. 2017;15(1):199. doi:10.1186/s12916-017-0958-6 PubMedGoogle ScholarCrossref
    3.
    Rewers  M, Ludvigsson  J.  Environmental risk factors for type 1 diabetes.   Lancet. 2016;387(10035):2340-2348. doi:10.1016/S0140-6736(16)30507-4 PubMedGoogle ScholarCrossref
    4.
    Honeyman  M.  How robust is the evidence for viruses in the induction of type 1 diabetes?   Curr Opin Immunol. 2005;17(6):616-623. doi:10.1016/j.coi.2005.09.007 PubMedGoogle ScholarCrossref
    5.
    Honeyman  MC, Laine  D, Zhan  Y, Londrigan  S, Kirkwood  C, Harrison  LC.  Rotavirus infection induces transient pancreatic involution and hyperglycemia in weanling mice.   PLoS One. 2014;9(9):e106560. doi:10.1371/journal.pone.0106560 PubMedGoogle Scholar
    6.
    Pane  JA, Webster  NL, Coulson  BS.  Rotavirus activates lymphocytes from non-obese diabetic mice by triggering toll-like receptor 7 signaling and interferon production in plasmacytoid dendritic cells.   PLoS Pathog. 2014;10(3):e1003998. doi:10.1371/journal.ppat.1003998 PubMedGoogle Scholar
    7.
    Pane  JA, Webster  NL, Zufferey  C, Coulson  BS.  Rotavirus acceleration of murine type 1 diabetes is associated with increased MHC class I-restricted antigen presentation by B cells and elevated proinflammatory cytokine expression by T cells.   Virus Res. 2014;179:73-84. doi:10.1016/j.virusres.2013.11.009 PubMedGoogle ScholarCrossref
    8.
    Graham  KL, O’Donnell  JA, Tan  Y,  et al.  Rotavirus infection of infant and young adult nonobese diabetic mice involves extraintestinal spread and delays diabetes onset.   J Virol. 2007;81(12):6446-6458. doi:10.1128/JVI.00205-07 PubMedGoogle ScholarCrossref
    9.
    Honeyman  MC, Coulson  BS, Stone  NL,  et al.  Association between rotavirus infection and pancreatic islet autoimmunity in children at risk of developing type 1 diabetes.   Diabetes. 2000;49(8):1319-1324. doi:10.2337/diabetes.49.8.1319 PubMedGoogle ScholarCrossref
    10.
    Snell-Bergeon  JK, Smith  J, Dong  F,  et al.  Early childhood infections and the risk of islet autoimmunity: the Diabetes Autoimmunity Study in the Young (DAISY).   Diabetes Care. 2012;35(12):2553-2558. doi:10.2337/dc12-0423 PubMedGoogle ScholarCrossref
    11.
    Perrett  KP, Jachno  K, Nolan  TM, Harrison  LC.  Association of rotavirus vaccination with the incidence of type 1 diabetes in children.   JAMA Pediatr. 2019;173(3):280-282. doi:10.1001/jamapediatrics.2018.4578 PubMedGoogle ScholarCrossref
    12.
    Rogers  MAM, Basu  T, Kim  C.  Lower incidence rate of type 1 diabetes after receipt of the rotavirus vaccine in the United States, 2001–2017.   Sci Rep. 2019;9(1):7727. doi:10.1038/s41598-019-44193-4 PubMedGoogle ScholarCrossref
    13.
    Baggs  J, Gee  J, Lewis  E,  et al.  The Vaccine Safety Datalink: a model for monitoring immunization safety.   Pediatrics. 2011;127(suppl 1):S45-S53. doi:10.1542/peds.2010-1722H PubMedGoogle ScholarCrossref
    14.
    Mullooly  J, Drew  L, DeStefano  F,  et al.  Quality of HMO vaccination databases used to monitor childhood vaccine safety: Vaccine Safety DataLink Team.   Am J Epidemiol. 1999;149(2):186-194. doi:10.1093/oxfordjournals.aje.a009785 PubMedGoogle ScholarCrossref
    15.
    McNeil  MM, Gee  J, Weintraub  ES,  et al.  The Vaccine Safety Datalink: successes and challenges monitoring vaccine safety.   Vaccine. 2014;32(42):5390-5398. doi:10.1016/j.vaccine.2014.07.073 PubMedGoogle ScholarCrossref
    16.
    Ezeanolue  E, Harriman  K, Hunter  P, Kroger  A, Pellegrini  C. General best practice guidelines for immunization: best practices guidance of the Advisory Committee on Immunization Practices (ACIP). https://www.cdc.gov/vaccines/hcp/acip-recs/general-recs/downloads/general-recs.pdf. Accessed July 2, 2019.
    17.
    Robinson  CL, Bernstein  H, Romero  JR, Szilagyi  P.  Advisory Committee on Immunization Practices recommended immunization schedule for children and adolescents aged 18 years or younger—United States, 2019.   MMWR Morb Mortal Wkly Rep. 2019;68(5):112-114. doi:10.15585/mmwr.mm6805a4 PubMedGoogle ScholarCrossref
    18.
    Glanz  JM, Newcomer  SR, Narwaney  KJ,  et al.  A population-based cohort study of undervaccination in 8 managed care organizations across the United States.   JAMA Pediatr. 2013;167(3):274-281. doi:10.1001/jamapediatrics.2013.502 PubMedGoogle ScholarCrossref
    19.
    Lieu  TA, Ray  GT, Klein  NP, Chung  C, Kulldorff  M.  Geographic clusters in underimmunization and vaccine refusal.   Pediatrics. 2015;135(2):280-289. doi:10.1542/peds.2014-2715 PubMedGoogle ScholarCrossref
    20.
    Smith  PJ, Chu  SY, Barker  LE.  Children who have received no vaccines: who are they and where do they live?   Pediatrics. 2004;114(1):187-195. doi:10.1542/peds.114.1.187 PubMedGoogle ScholarCrossref
    21.
    Gopal  SH, Edwards  KM, Creech  B, Weitkamp  J-H.  Variability in immunization practices for preterm infants.   Am J Perinatol. 2018;35(14):1394-1398. doi:10.1055/s-0038-1660453 PubMedGoogle ScholarCrossref
    22.
    Lawrence  JM, Black  MH, Zhang  JL,  et al.  Validation of pediatric diabetes case identification approaches for diagnosed cases by using information in the electronic health records of a large integrated managed health care organization.   Am J Epidemiol. 2014;179(1):27-38. doi:10.1093/aje/kwt230 PubMedGoogle ScholarCrossref
    23.
    McKinney  PA, Parslow  R, Gurney  KA, Law  GR, Bodansky  HJ, Williams  R.  Perinatal and neonatal determinants of childhood type 1 diabetes: a case-control study in Yorkshire, UK.   Diabetes Care. 1999;22(6):928-932. doi:10.2337/diacare.22.6.928 PubMedGoogle ScholarCrossref
    24.
    Dahlquist  GG, Patterson  C, Soltesz  G.  Perinatal risk factors for childhood type 1 diabetes in Europe: the EURODIAB Substudy 2 Study Group.   Diabetes Care. 1999;22(10):1698-1702. doi:10.2337/diacare.22.10.1698 PubMedGoogle ScholarCrossref
    25.
    Stene  LC, Magnus  P, Lie  RT, Søvik  O, Joner  G; Norwegian childhood Diabetes Study Group.  Birth weight and childhood onset type 1 diabetes: population based cohort study.   BMJ. 2001;322(7291):889-892. doi:10.1136/bmj.322.7291.889 PubMedGoogle ScholarCrossref
    26.
    Janghorbani  M, Van Dam  RM, Willett  WC, Hu  FB.  Systematic review of type 1 and type 2 diabetes mellitus and risk of fracture.   Am J Epidemiol. 2007;166(5):495-505. doi:10.1093/aje/kwm106 PubMedGoogle ScholarCrossref
    27.
    Dabelea  D, Mayer-Davis  EJ, Saydah  S,  et al; SEARCH for Diabetes in Youth Study.  Prevalence of type 1 and type 2 diabetes among children and adolescents from 2001 to 2009.   JAMA. 2014;311(17):1778-1786. doi:10.1001/jama.2014.3201 PubMedGoogle ScholarCrossref
    28.
    Harder  T, Roepke  K, Diller  N, Stechling  Y, Dudenhausen  JW, Plagemann  A.  Birth weight, early weight gain, and subsequent risk of type 1 diabetes: systematic review and meta-analysis.   Am J Epidemiol. 2009;169(12):1428-1436. doi:10.1093/aje/kwp065 PubMedGoogle ScholarCrossref
    29.
    Svensson  J, Carstensen  B, Mortensen  HB, Borch-Johnsen  K; Danish Study Group of Childhood Diabetes.  Early childhood risk factors associated with type 1 diabetes—is gender important?   Eur J Epidemiol. 2005;20(5):429-434. doi:10.1007/s10654-005-0878-1 PubMedGoogle ScholarCrossref
    30.
    Cardwell  CR, Stene  LC, Joner  G,  et al.  Caesarean section is associated with an increased risk of childhood-onset type 1 diabetes mellitus: a meta-analysis of observational studies.   Diabetologia. 2008;51(5):726-735. doi:10.1007/s00125-008-0941-z PubMedGoogle ScholarCrossref
    31.
    Daley  MF, Shoup  JA, Newcomer  SR,  et al.  Assessing potential confounding and misclassification bias when studying the safety of the childhood immunization schedule.   Acad Pediatr. 2018;18(7):754-762. doi:10.1016/j.acap.2018.03.007 PubMedGoogle ScholarCrossref
    32.
    Frederiksen  B, Kroehl  M, Lamb  MM,  et al.  Infant exposures and development of type 1 diabetes mellitus: the Diabetes Autoimmunity Study in the Young (DAISY).   JAMA Pediatr. 2013;167(9):808-815. doi:10.1001/jamapediatrics.2013.317 PubMedGoogle ScholarCrossref
    33.
    Rosenbauer  J, Herzig  P, Giani  G.  Early infant feeding and risk of type 1 diabetes mellitus—a nationwide population-based case-control study in pre-school children.   Diabetes Metab Res Rev. 2008;24(3):211-222. doi:10.1002/dmrr.791 PubMedGoogle ScholarCrossref
    34.
    Schroeder  EB, Donahoo  WT, Goodrich  GK, Raebel  MA.  Validation of an algorithm for identifying type 1 diabetes in adults based on electronic health record data.   Pharmacoepidemiol Drug Saf. 2018;27(10):1053-1059. doi:10.1002/pds.4377 PubMedGoogle ScholarCrossref
    35.
    Allison  PD.  Survival Analysis Using SAS: A Practical Guide. 2nd ed. Cary, NC: SAS Institute, Inc; 2010.
    36.
    Moons  KG, Donders  RA, Stijnen  T, Harrell  FE  Jr.  Using the outcome for imputation of missing predictor values was preferred.   J Clin Epidemiol. 2006;59(10):1092-1101. doi:10.1016/j.jclinepi.2006.01.009 PubMedGoogle ScholarCrossref
    37.
    Lash  TL, Fox  MP, Fink  AK.  Applying Quantitative Bias Analysis to Epidemiologic Data. New York, NY: Springer Science & Business Media; 2011.
    38.
    Lash  TL, Abrams  B, Bodnar  LM.  Comparison of bias analysis strategies applied to a large data set.   Epidemiology. 2014;25(4):576-582. doi:10.1097/EDE.0000000000000102 PubMedGoogle ScholarCrossref
    39.
    Burnett  E, Yen  C, Tate  JE, Parashar  UD.  Rotavirus vaccines: current global impact and future perspectives.   Future Virol. 2016;11(10):699-708. doi:10.2217/fvl-2016-0082 PubMedGoogle ScholarCrossref
    40.
    Leshem  E, Tate  JE, Steiner  CA, Curns  AT, Lopman  BA, Parashar  UD.  National estimates of reductions in acute gastroenteritis–related hospitalizations and associated costs in US children after implementation of rotavirus vaccines.   J Pediatric Infect Dis Soc. 2018;7(3):257-260. doi:10.1093/jpids/pix057 PubMedGoogle ScholarCrossref
    41.
    Vaarala  O, Jokinen  J, Lahdenkari  M, Leino  T.  Rotavirus vaccination and the risk of celiac disease or type 1 diabetes in Finnish children at early life.   Pediatr Infect Dis J. 2017;36(7):674-675. doi:10.1097/INF.0000000000001600 PubMedGoogle ScholarCrossref
    ×