US children are recommended to receive up to 28 vaccine doses against 14 diseases before their second birthday.1 Most vaccines contain inactivated forms of the targeted virus or bacterium; however, 3 vaccines—rotavirus, measles-mumps-and rubella (MMR) and varicella—contain live, attenuated viruses.2 Vaccines may have heterologous effects, meaning that immune responses may extend beyond protection from infections targeted by the vaccine.3–5 Epidemiologic studies in Africa, Europe and India have reported differences in such nonspecific effects4,6 based on the sequence of live and inactivated vaccines in early childhood.7–12 Specifically, decreased mortality7,9 and decreased risk for infections not targeted by vaccination11,12 have been observed when live vaccines only, as compared with inactivated vaccines only or inactivated plus live vaccines given concurrently, were the most recent vaccines received. However, these observational studies have been conducted across countries with varying immunization schedules, and multiple potential sources of bias have been acknowledged.13
Currently, US children are recommended to receive inactivated and live vaccines concurrently through 12–15 months of age and then additional inactivated vaccines before their second birthday.1 Deviations from this recommended sequence may occur due to parental choice to delay or refuse vaccines or other reasons. Recently, a US study reported 50% decreased risk of hospitalizations for non–vaccine-targeted infections in children who had received live vaccines most recently, compared with children who received inactivated vaccines most recently.14 The study used medical claims data, which have significant limitations, including potentially incomplete data capture and misclassification.15 To evaluate the current recommended order of live and inactivated vaccines in the US immunization schedule, additional studies using other data sources are needed.
This study was conducted through the Centers for Disease Control and Prevention’s Vaccine Safety Datalink, a network of integrated health systems (IHSs) that use electronic health record (EHR) data for vaccine safety research.16 Our primary objective was to estimate non–vaccine-targeted infection risk based on most recent vaccine type (live only, inactivated only or both concurrently) in 11–23 months of age in a large US pediatric population.
MATERIALS AND METHODS
Setting and Study Design
We conducted a retrospective cohort study. Six IHSs contributed patient data including vaccination history and medical diagnoses: Marshfield Clinic (Marshfield, WI), and Kaiser Permanente Colorado (Denver, CO), Northwest (Portland, OR), Washington (Seattle, WA), Northern California (Oakland, CA) and Southern California (Pasadena, CA). All participating sites’ Institutional Review Boards approved this project.
We identified children born January 1, 2003, to September 30, 2013, with continuous enrollment in their IHS from within 6 weeks of birth through their second birthday and ≥2 well-child visits before their first birthday. These criteria were applied to identify children with continuous, routine IHS use during the recommended period of early childhood immunizations. We excluded children with potential vaccine contraindications (Table 1, Supplemental Digital Content, http://links.lww.com/INF/D720) and children with vaccination history anomalies, eg, records of live influenza vaccines given before age 24 months (the earliest recommended vaccination age). We excluded children who received vaccines not universally recommended by the US Advisory Committee on Immunization Practices, including international travel vaccines. Similar to other studies of live and inactivated vaccines, we excluded children with <3 diphtheria-tetanus-acellular pertussis vaccines before their first birthday.11,14
We obtained all immunization records for cohort members from EHR data. For 4 of the 6 IHSs, state immunization information system records were also available and included. We examined records to determine whether immunization visits included inactivated vaccines only, live vaccines only or both inactivated and live vaccines (Table 2, Supplemental Digital Content, http://links.lww.com/INF/D720). We also quantified compliance with the US Advisory Committee on Immunization Practices–recommended immunization schedule using average days undervaccinated, a measure summarizing the degree of undervaccination across recommended doses.17
Outcomes were defined as the first inpatient or emergency department (ED) diagnosis of non–vaccine-targeted infection in 11–23 months of age, inclusive. We focused on the 11–23 month age range because live MMR and varicella vaccines are recommended at age 12 months, and other studies of nonspecific effects of live vaccines have focused on infection risk in the time period following measles-containing vaccines.11,14 Our diagnosis list was updated from one used in previous studies (Table 3, Supplemental Digital Content, http://links.lww.com/INF/D720).11,14 Our updates were based on findings from a sample review of medical provider notes in a previous study.18 In that review, we identified circumstances where diagnoses did not represent incident infections. Examples included inpatient admission for pressure equalization tube placement with otitis media diagnosis and ED visits for injury with a coincidental upper respiratory infection diagnosis. Because these cases did not represent incident infections, we excluded inpatient diagnoses occurring on the same day as preventive surgeries for chronic ear infections (Table 4, Supplemental Digital Content, http://links.lww.com/INF/D720) and ED diagnoses occurring in the same visit as injury diagnoses (Table 5, Supplemental Digital Content, http://links.lww.com/INF/D720).
We excluded diagnoses for infections targeted by early childhood vaccines, including hepatitis A and B, diphtheria, tetanus, pertussis, Haemophilus influenzae type b, poliovirus, influenza, measles, mumps, rubella and varicella. Because the rotavirus vaccine was licensed in 2006,19 we excluded rotavirus diagnoses after 2006. Because pneumonia may be caused by organisms other than pneumococcus, and because most pneumococcal strains circulating in the United States are not vaccine preventable,20 pneumonia diagnoses were included.
We reported the distribution of sex, preterm birth status (gestational age <37 weeks), number of outpatient visits in 1–10 months of age and prevalence of a chronic condition (Table 6, Supplemental Digital Content, http://links.lww.com/INF/D720) before age 11 months. We used χ2 and Wilcoxon rank-sum tests to compare baseline characteristics between children with ≥1 immunization visit with live vaccine(s) only and children who received inactivated vaccines at all immunization visits in 11–23 months of age.
We used counting processes for Cox regression to model vaccine type (live only, inactivated only or both types concurrently) as a time-varying exposure. We used an age time scale and stratified by year and month of birth and Vaccine Safety Datalink site to control for temporal and site-specific outcome variation. Observation time for each child started at 11 months of age. Vaccine exposure at observation start was based on the most recent immunization visit before 11 months of age. For every immunization visit after 11 months of age, vaccine exposure changed the day after immunization based on the type of vaccines received. Follow-up time was censored at outcome date or day before the child’s second birthday. The last day of study observation was September 29, 2015.
For our primary analysis, we reported outcome incidence rates, unadjusted and adjusted hazard ratios (aHRs) and 95% confidence intervals (CIs). In adjusted analyses, we controlled for sex, preterm birth, chronic conditions and number of outpatient visits in 1–10 months of ages. Preterm birth status was not available for 10% of cohort members, and these children were excluded from adjusted analyses. Because previous studies have reported differences in infection risk following live vaccines by sex,8,10,21 we evaluated interactions between vaccine type and sex and presented results from the primary analysis stratified by sex. We controlled for preterm birth because prematurity is associated with higher infection risk and with vaccination patterns in early childhood.22–25 We controlled for the presence of chronic conditions because baseline health status is associated with infection risk and may also affect vaccination status.26 The other US study on sequence of live and inactivated vaccines and infection risk reported on the influence of baseline levels of utilization; therefore, we controlled for outpatient visits.14 We excluded outpatient visits in the first month of life because participating IHSs had varying practices for providing newborn care immediately following birth (eg, in-hospital, clinic and home visits).
We conducted several exploratory secondary analyses to address potential data misclassification, compare our results with findings from previous studies and evaluate live MMR and varicella vaccination separately from live rotavirus vaccination. Misclassification of vaccination history may occur if children received vaccines outside of the IHS and these records were not available in EHR or state immunization information systems.27 Therefore, we repeated the primary analysis within a subcohort of children with records for all recommended vaccine doses. Because the previous study of US children using claims data limited the outcome to hospitalizations,14 we repeated the primary analysis limited to inpatient non–vaccine-targeted infections. Similar to the other US study,14 we examined 4 types of non–vaccine-targeted infections in separate models: upper respiratory, lower respiratory, gastrointestinal and other infections. Finally, acknowledging that live vaccines stimulate different immune responses (rotavirus vaccination stimulates a mucosal immune response in the gut, whereas a more systemic immune response follows MMR or varicella vaccination), we conducted post hoc analyses where we repeated the primary analysis but examined the effects of the live MMR and varicella vaccines separately from the live rotavirus vaccine.
Last receipt of inactivated vaccines only was the reference group in all models. To evaluate model fit and assumption of proportional hazards over time, we plotted Schoenfeld residuals for baseline covariates and evaluated Schoenfeld individual tests. For sex, we presented adjusted models separately for males and females if the Schoenfeld individual test was <0.05; for other covariates, we included an interaction term between the covariate and time if the individual test indicated potential deviation from the proportional hazards assumption.
All statistical tests were 2 sided with an α level of 0.05. Analyses were conducted in SAS 9.4. Schoenfeld residuals were plotted and tested in R.
Our study included 428,608 children (Fig. 1). Approximately 49% were female, and 21,075 (4.9%) children had an immunization visit at 11–23 months of age where live vaccines only were administered. Compared with children who had ≥1 inactivated vaccine at all immunization visits, children with live vaccine-only visits were more likely to be male, born earlier in the study period, have more baseline outpatient visits and have higher average days undervaccinated (Table 1). Overall, cohort members contributed 435,993 person-years of observation time. Approximately 1.5% of person-time consisted of last exposure to live vaccines only, 43.8% consisted of last exposure to both live and inactivated vaccines concurrently and 54.7% consisted of last exposure to inactivated vaccines only. Of the person-time that included last exposure to live vaccines (either alone or concurrently with inactivated vaccines), 90.8% was exposed to MMR or varicella vaccines (Table 7, Supplemental Digital Content, http://links.lww.com/INF/D720).
Within the cohort, 44,075 (10.3%) children had an ED or inpatient encounter for non–vaccine-targeted infections in 11–23 months of age. Rates of non–vaccine-targeted infections were 11,386/100,000 person-years in males and 8794/100,000 person-years in females. In adjusted models, among males, last receipt of live vaccines only or both inactivated and live vaccines concurrently was associated with lower risk of non–vaccine-targeted infections, versus inactivated vaccines only (live vaccine-only aHR: 0.83, 95% CI: 0.72–0.94; inactivated and live vaccines concurrently aHR: 0.91; 95% CI: 0.88–0.94). Among females, a decreased risk of infections was observed when inactivated and live vaccines were last given concurrently, but not when live vaccines only were last received, versus inactivated vaccines only (inactivated and live vaccines concurrently aHR: 0.94; 95% CI: 0.91–0.97; live vaccine-only aHR: 0.99, 95% CI: 0.87–1.14) (Table 2).
Hazard ratios were similar when the primary analysis was limited to 222,418 (51.9%) children who received all recommended vaccine doses. However, only last receipt of both inactivated and live vaccines concurrently was significantly associated with non–vaccine-targeted infection among males, as compared with inactivated vaccines only (aHR: 0.94, 95% CI: 0.89–0.98) (Table 8, Supplemental Digital Content, http://links.lww.com/INF/D720).
Of the outcomes, 6764 (15.3%) were inpatient admissions. In adjusted models, type of vaccine last received was not significantly associated with inpatient admission for non–vaccine-targeted infection in males or females (Table 3). In analyses by infection type, lower risk of upper respiratory infections was observed following live vaccines (live vaccine-only aHR: 0.86, 95% CI: 0.76–0.96; inactivated and live vaccines concurrently aHR: 0.93; 95% CI: 0.90–0.95) as compared with last receipt of inactivated vaccines only. There were no significant differences in lower respiratory infection or gastrointestinal risk by vaccine exposure. We observed an increased risk for other infections following live vaccines only as compared with inactivated vaccines only (aHR = 1.37; 95% CI: 1.01–1.86) (Table 4).
In post hoc analyses, lower risk of non–vaccine-targeted infections was observed in males when MMR and/or varicella (MMR/varicella) were the last vaccines received, with or without inactivated vaccines concurrently, as compared with when inactivated vaccines only were the last vaccines received (MMR/varicella vaccine(s)-only aHR: 0.82, 95% CI: 0.72–0.93; MMR/varicella and inactivated vaccines concurrently aHR: 0.89, 95% CI: 0.87–0.92). In contrast, higher infection risk was observed when the live rotavirus vaccine was among the last vaccines received as compared with last receipt of inactivated vaccines only (aHR = 1.14, 95% CI: 1.05–1.23). Similar to males, higher infection risk was observed among females when the rotavirus vaccine was the last live vaccine received as compared with when inactivated vaccines only were last received (aHR = 1.44, 95% CI: 1.32–1.57). In females, lower infection risk infection was observed when the last vaccines received were MMR/varicella given concurrently with inactivated vaccines (aHR = 0.91, 95% CI: 0.88–0.94) but not following MMR/varicella vaccines alone (aHR = 0.99, 95% CI: 0.86–1.13) (Table 9, Supplemental Digital Content, http://links.lww.com/INF/D720). Among cohort members who experienced a non–vaccine-targeted infection, average time from last immunization visit to non–vaccine-targeted infection diagnosis was longest when the last immunization visit included rotavirus vaccination (Table 10, Supplemental Digital Content, http://links.lww.com/INF/D720).
In a large, population-based cohort of US children 11–23 months of age, males had lower risk of ED and inpatient encounters for non–vaccine-targeted infections when the last immunization visit included live vaccines compared with when the last immunization visit included inactivated vaccines only. In females, lower infection risk was only observed when live vaccines given concurrently with inactivated vaccines were last received, but not when live vaccines only were last received, compared with when the last immunization visit included inactivated vaccines only. The associations we observed were more modest than what has previously been reported for US children.14
Previous observational studies in Denmark,11 Netherlands,28 and the United States14 reported protective associations between receipt of live vaccines and non–vaccine-targeted infections, but no differences by sex. Observational studies in developing countries have also reported protective benefits of live vaccination, particularly among females.8,10,21 Key differences may explain varying results across studies. First, immunization schedules differ, and some vaccines studied in other countries, such as the live Bacillus Calmette-Guerin vaccine, are not routinely administered in the United States. In addition, most developing countries use whole-cell pertussis vaccination, whereas children in this study received acellular pertussis vaccination. Varying immune responses between these pertussis vaccines have been previously observed.29 Second, outcomes have differed across studies, from infant mortality to respiratory infections. Third, studies have ranged across populations and age groups with differing health status, baseline infection risk and access to care. Such differences create challenges in comparing findings on nonspecific vaccine effects across studies.
Several reasons may explain why we observed lower risk of non–vaccine-targeted infections following receipt of live vaccines only in males but not females in our primary analysis. In our study, males had higher rates of non–vaccine-targeted infections than females, which aligns with previous research on higher infection rates and greater severity of illness in male children.30 It is hypothesized that such sex differences are immunologically driven and are related to the influence of sex-specific hormones on Th1 and Th2 responses in early childhood.30 It is possible that any benefit from live vaccination may be more advantageous among the subpopulation more susceptible to infection. In addition, targeted immune responses to some vaccine antigens differ between sexes in children.31 Therefore, it could be hypothesized that nonspecific immune responses to vaccines may also differ between sexes. Finally, <5% of our cohort had an immunization visit where live vaccines only were administered, and having a live vaccine-only visit was more common in males than females. It is possible that we may have been underpowered to detect an association between live-only vaccination and infections in female children.
Epidemiologic research on nonspecific vaccine effects has led to questions of the influence of live versus inactivated vaccines on children’s immune responses. However, this dichotomization of vaccine types does not adequately reflect the varied immune responses produced by individual live and inactivated vaccines.32 In post hoc analyses, we observed higher infection risk when rotavirus vaccine was among the last vaccines received and lower infection risk when MMR or varicella was last received, compared with when inactivated vaccines only were received last. An important limitation of these post hoc analyses was that almost all exposure to rotavirus vaccines occurred concurrently with inactivated vaccines. While exploratory, the divergent findings between different types of live vaccines that we observed suggest the need for additional research examining nonspecific effects of individual live and inactivated vaccines.
A recent systematic review of nonspecific vaccine effects with regard to childhood mortality concluded that all observational studies examined were subject to high or very high risk of bias.13 One limitation of our observational study, which relied on electronic medical record data, is potential bias due to parent health-seeking behavior. Parents with hesitancy about vaccines express greater distrust in traditional medical care than parents who accept recommended immunization guidelines.33 As such, parents who chose to have their children receive the MMR or varicella vaccines alone may have been less likely to seek care with ill children than parents adhering to the recommended immunization schedule. Lower ED encounter rates in US children undervaccinated due to parental choice versus fully vaccinated children have been previously observed.17 This bias in who seeks care in an ED for infections may have contributed to lower outcome rates in children who last received live vaccines only.
Our study has additional strengths and limitations. We used chart review results from a previous study to update our method for identifying incident non–vaccine-targeted infections, thus improving outcome specificity. Results from our secondary analysis of children with records for all recommended vaccines were similar to our primary analysis, suggesting that bias from missing vaccine records was likely negligible. While these steps reduced concern about misclassification bias, we cannot guarantee a lack of bias due to misclassified exposure, outcome or covariates in electronic health data. Also, confounding biases in observational vaccine research have been well documented, particularly concerning comparisons of vaccination schedules.34,35 While we did control for several important confounders, we cannot rule out that unmeasured and unmeasurable differences between children with different immunization patterns influenced results. Another potential limitation is that some infections may have actually been vaccine preventable; for example, some gastrointestinal infections not tested for rotavirus may have been caused by rotavirus. Given the low incidence of vaccine-preventable infections in the United States, we would expect this limitation to have minimal effect on our results.
Based on findings from 2 systematic reviews of nonspecific vaccine effects with regard to childhood mortality, the World Health Organization Strategic Advisory Group of Experts on Immunization affirmed its recommended schedule for childhood Bacillus Calmette-Guerin, diphtheria-tetanus-pertussis, and measles vaccines.13,36,37 Given the modest associations we observed, in balance with potential risks of vaccine-preventable diseases if live vaccines are delayed, our results support the current sequence of live and inactivated vaccines in the US immunization schedule with respect to non–vaccine-targeted infections.
1. Robinson CL, Romero JR, et al. Advisory committee on immunization practices recommended immunization schedule for children and adolescents aged 18 years or younger—United States, 2017. MMWR. 2017;17:1136–1138.
2. Centers for Disease Control and Prevention. Epidemiology and Prevention of Vaccine-Preventable Diseases. 2015Washington, DC: Public Health Foundation; . Available at: https://www.cdc.gov/vaccines/pubs/pinkbook/index.html
. Accessed April 9, 2019.
3. Welsh RM, Selin LK. No one is naive: the significance of heterologous T-cell immunity. Nat Rev Immunol. 2002;2:417–426.
4. Flanagan KL, van Crevel R, Curtis N, et al; Optimmunize Network. Heterologous (“nonspecific”) and sex-differential effects of vaccines: epidemiology, clinical trials, and emerging immunologic mechanisms. Clin Infect Dis. 2013;57:283–289.
5. Shann F. The non-specific effects of vaccines. Arch Dis Child
6. Goldblatt D, Miller E. Nonspecific effects of vaccines. JAMA. 2014;311:804–805.
7. Aaby P, Jensen H, Samb B, et al. Differences in female-male mortality after high-titre measles vaccine and association with subsequent vaccination
with diphtheria-tetanus-pertussis and inactivated poliovirus: reanalysis of West African studies. Lancet. 2003;361:2183–2188.
8. Hirve S, Bavdekar A, Juvekar S, et al. Non-specific and sex-differential effects of vaccinations on child
survival in rural western India. Vaccine. 2012;30:7300–7308.
9. Fisker AB, Ravn H, Rodrigues A, et al. Co-administration of live measles and yellow fever vaccines and inactivated pentavalent vaccines is associated with increased mortality compared with measles and yellow fever vaccines only. An observational study from Guinea-Bissau. Vaccine. 2014;32:598–605.
10. Aaby P, Nielsen J, Benn CS, et al. Sex-differential and non-specific effects of routine vaccinations in a rural area with low vaccination
coverage: an observational study from Senegal. Trans R Soc Trop Med Hyg. 2015;109:77–84.
11. Sørup S, Benn CS, Poulsen A, et al. Live vaccine against measles, mumps, and rubella and the risk of hospital admissions for nontargeted infections. JAMA. 2014;311:826–835.
12. Sørup S, Benn CS, Poulsen A, et al. Simultaneous vaccination
with MMR and DTaP-IPV-Hib and rate of hospital admissions with any infections: a nationwide register based cohort study. Vaccine. 2016;34:6172–6180.
13. Higgins JP, Soares-Weiser K, López-López JA, et al. Association of BCG, DTP, and measles containing vaccines with childhood mortality: systematic review. BMJ. 2016;355:i5170.
14. Bardenheier BH, McNeil MM, Wodi AP, et al. Risk of nontargeted infectious disease hospitalizations among US children following inactivated and live vaccines, 2005-2014. Clin Infect Dis. 2017;65:729–737.
15. Funk MJ, Landi SN. Misclassification in administrative claims data: quantifying the impact on treatment effect estimates. Curr Epidemiol Rep. 2014;1:175–185.
16. McNeil MM, Gee J, Weintraub ES, et al. The Vaccine Safety Datalink: successes and challenges monitoring vaccine safety. Vaccine. 2014;32:5390–5398.
17. 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;167:274–281.
18. Glanz JM, Newcomer SR, Daley MF, et al. Association between estimated cumulative vaccine antigen exposure through the first 23 months of life and non-vaccine-targeted infections from 24 through 47 months of age. JAMA. 2018;319:906–913.
19. Parashar UD, Alexander JP, Glass RI; Advisory Committee on Immunization Practices (ACIP), Centers for Disease Control and Prevention (CDC). Prevention of rotavirus gastroenteritis among infants and children. Recommendations of the Advisory Committee on Immunization Practices (ACIP). MMWR Recomm Rep. 2006;55(RR-12):1–13.
20. Centers for Disease Control and Prevention. Chapter 11: pneumococcal. In: Manual for the Surveillance of Vaccine-Preventable Diseases. 2017. Available at: https://www.cdc.gov/vaccines/pubs/surv-manual/index.html
. Accessed August 23, 2018.
21. Veirum JE, Sodemann M, Biai S, et al. Routine vaccinations associated with divergent effects on female and male mortality at the paediatric ward in Bissau, Guinea-Bissau. Vaccine. 2005;23:1197–1204.
22. PrabhuDas M, Adkins B, Gans H, et al. Challenges in infant immunity: implications for responses to infection and vaccines. Nat Immunol. 2011;12:189–194.
23. Batra JS, Eriksen EM, Zangwill KM, et al; Vaccine Safety Datalink. Evaluation of vaccine coverage for low birth weight infants during the first year of life in a large managed care population. Pediatrics. 2009;123:951–958.
24. Langkamp DL, Hoshaw-Woodard S, Boye ME, et al. Delays in receipt of immunizations in low-birth-weight children: a nationally representative sample. Arch Pediatr Adolesc Med. 2001;155:167–172.
25. Sisson H, Gardiner E, Watson R. Vaccination
timeliness in preterm infants: an integrative review of the literature. J Clin Nurs. 2017;26:4094–4104.
26. Iwane MK, Edwards KM, Szilagyi PG, et al; New Vaccine Surveillance Network. Population-based surveillance for hospitalizations associated with respiratory syncytial virus, influenza virus, and parainfluenza viruses among young children. Pediatrics. 2004;113:1758–1764.
27. Daley MF, Glanz JM, Newcomer SR, et al. Assessing misclassification of vaccination
status: implications for studies of the safety of the childhood immunization schedule. Vaccine. 2017;35:1873–1878.
28. Tielemans SMAJ, de Melker HE, Hahné SJM, et al. Non-specific effects of measles, mumps, and rubella (MMR) vaccination
in high income setting: population based cohort study in the Netherlands. BMJ. 2017;358:j3862.
29. Kapil P, Merkel TJ. Pertussis vaccines and protective immunity. Curr Opin Immunol. 2019;59:72–78.
30. Muenchhoff M, Goulder PJ. Sex differences in pediatric infectious diseases
. J Infect Dis. 2014;209(suppl 3):S120–S126.
31. Klein SL, Marriott I, Fish EN. Sex-based differences in immune function and responses to vaccination
. Trans R Soc Trop Med Hyg. 2015;109:9–15.
32. Pollard AJ, Finn A, Curtis N. Non-specific effects of vaccines: plausible and potentially important, but implications uncertain. Arch Dis Child
33. Glanz JM, Wagner NM, Narwaney KJ, et al. A mixed methods study of parental vaccine decision making and parent-provider trust. Acad Pediatr. 2013;13:481–488.
34. Glanz JM, Newcomer SR, Jackson ML, et al. White paper on studying the safety of the childhood immunization schedule in the vaccine safety datalink. Vaccine. 2016;34(suppl 1):A1–A29.
35. Jackson ML. Challenges in comparing the safety of different vaccination
schedules. Vaccine. 2013;31:2126–2129.
36. Kandasamy R, Voysey M, McQuaid F, et al. Non-specific immunological effects of selected routine childhood immunisations: systematic review. BMJ. 2016;355:i5225.
37. World Health Organization. Meeting of the Strategic Advisory Group of Experts on immunization, April 2014—conclusions and recommendations. Wkly Epidemiol Rec. 2014;89:221–236.