A narrative review of nonspecific effects of pediatric vaccines on child mortality and morbidity

ABSTRACT

We reviewed evidence on nonspecific effects of pediatric vaccines on mortality and morbidity in countries with high child mortality. Literature search of epidemiological studies was conducted for studies published between 2000 and September 2021 using MEDLINE. Consistent evidence exists regarding the potential protective effect of measles vaccine on child survival. Vaccination with Bacillus Calmette–Guérin (BCG) vaccine was related to lower risk of mortality in young children (including low birth weight babies) and inverse associations were found between developing a scar and having a positive tuberculin test after BCG vaccination with mortality. BCG vaccine might also reduce the risk of nontuberculosis infectious diseases. Studies on the association between diphtheria-pertussis-tetanus (DPT) vaccine and child survival showed inconsistent findings, which might be affected by bias and confounding. More evidence is needed to assess the role of these and other vaccines in children’s health and to better understand potential biological mechanisms and other influential factors.

KEYWORDS:

• Nonspecific effects
• vaccines
• children
• mortality
• morbidity

Introduction

Vaccination is an important public health tool for the prevention of infectious diseases.^1,2 The introduction of vaccines into national immunization programs led to a substantial reduction in the burden of vaccine preventable diseases worldwide.^3,4

It was proposed that vaccines might also have nonspecific effects on health, i.e. prevention of outcomes, unrelated to the prevention of target disease.^5–9 Early cohort and case-control studies conducted during 1965–1980s in low-income countries with high child mortality (Bangladesh, Benin, Burundi, Guinea-Bissau, Haiti, Senegal, and Zaire) showed that vaccination with a measles vaccine was associated with substantial 30–86% reduction in all-cause child mortality,^5–11 which was greater than the declines in measles mortality in these studies. Results were consistent across different populations, regions. and study designs.^5–11 Later studies, including clinical trials, on nonspecific effects of vaccines addressed relationships of Bacillus Calmette–Guérin (BCG) vaccine, diphtheria-tetanus-pertussis (DTP) vaccine, and oral polio vaccines (OPV) with child survival and other outcomes focusing on populations with high child mortality in developing countries^12–18, followed by studies conducted in high-income countries that evaluated associations between vaccination sequence and hospitalizations among children.^19–22 Results of these studies were inconsistent according to the studied vaccine (e.g., type and doses) and sometimes by country/population and internal conditions (e.g. shortage of vaccines and vaccination campaigns). Our aim was to review evidence on nonspecific effects of pediatric vaccines with a special focus on child mortality and morbidity among children from countries with high child mortality.

Methods

In this narrative review, a literature search of the Ovid MEDLINE was conducted to identify published articles until September 2021 using the term “non-specific beneficial effects of vaccines.” The search was limited to human studies and articles published in English. We also searched references of relevant original and review articles. The inclusion criteria were as follows: original articles of studies conducted among children, randomized controlled trials, cohort studies, and case-control studies that examined possible beneficial nonspecific effects of vaccines, namely, child survival and morbidity. We included studies that investigated BCG, measles containing vaccine, DTP, and polio vaccines. The exclusion criteria were studies reporting only vaccine-specific outcomes, case reports, and duplicate reports (i.e. reanalysis of the same dataset concerning the same research question or old data). Child mortality decreased markedly over the past two decades,²³ and vaccination coverage increased; therefore, to minimize the potential birth cohort effect, we focused on studies published from 2000 onward, taking into account that earlier studies were reviewed by Aaby et al.⁵ and Higgins et al.²⁴ From each included study, we extracted the following information: study design, study population, intervention group vs. control group, primary findings, and adjusting for confounders.

Results

Early observational studies conducted in high-child mortality settings suggested that measles vaccination was associated with significantly reduced child mortality, which was not explained by reduction in measles incidence and mortality (reviewed in Ref. ⁵), thus advocating for the introduction of universal measles vaccinations in these settings at that time. Since then, access to vaccines increased especially following the establishment of the expanded immunization program established in 1974⁴ and was related to reductions in vaccine preventable diseases. Awareness to nonspecific effects of vaccines on child survival was expanded to additional vaccines and populations in several regions in sub-Saharan Africa and Asia (Table 1).

Table 1. Studies on the association of pediatric vaccines with mortality and morbidity among children from countries with high child mortality

Study	Study design (country)	Vaccines	Study population intervention/ control	Outcome variable	Primary findings	Adjusting for confounders
Kristensen et al.^Citation12	Prospective cohort (Guinea-Bissau)	BCG, DTP, polio, and measles vaccines	BCG: 5274 children aged 0–6 months, 3301 vaccinated and 1973 unvaccinated. DTP/polio: 3972 children aged 2.5–6 months, 1822 unvaccinated; 1 dose, n = 1295, ≥2 doses, n = 855; Measles vaccine: 3441 children age 7–13 months; 1036 vaccinated and 2378 unvaccinated	Mortality	BCG: RR 0.72 (95% CI 0.36–0.85) DTP 1 dose: 1.84 (95% 1.10–3.10) DTP 2–3 doses: 1.38 (95% CI 0.73–2.61) Polio 1 dose: 1.81 (95% CI 1.17–3.05) DTP 2–3 doses: 1.39 (95% CI 0.73–2.64) Measles: 0.51 (95% CI 0.3–0.85)	Age, period, season, and cluster immunization with other vaccines.
Aaby et al. ^Citation25	Prospective cohort (Guinea- Bissau)	Measles vs. IPV	433 children aged 6–11 months N = 214 measles vaccine, N = 219 IPV	Mortality	RR 0.00 (95% CI 0.00–0.37) for girls RR 1.02 (95% CI 0.25–3.88) for boys RR 0.30 (95% CI 0.08–0.87) both sexes	Weight-for-age and early weaning
Kabir et al. ^Citation26	Population-based matched case-control study (India)	Failure to get measles vaccine at age 12 months	Children aged 12–59 months; 330 cases, children who died; 320 alive control children	Mortality	Failure to get measles vaccine at age 12 months: matched OR 3.0 (95% CI 1.7–5.4).	Age, sex, family size, and residential area
Garly et al. ^Citation27	Prospective cohort (Guinea-Bissau)	BCG vaccine and scar vs. no scar Positive vs. negative tuberculin test	1813 children with BCG scar vs. 813 BCG vaccinated children without a scar	Mortality	BCG scar: RR 0.41 (95% CI 0.25–0.67). Positive tuberculin test RR: 0.45 (95% CI 0.24–0.85)	Age at BCG vaccination, mother’s schooling, antenatal care, birth order, breastfeeding, weight for age, height for age, arm circumference, and electricity in the house
Breiman et.al. ^Citation13	Prospective cohort (Bangladesh)	DTP, polio, measles, and BCG vaccines	BCG: 37,894 children Measles: 36,650 children (30,677 vaccinated) 34,957 children received DTP 33,966 received BCG vaccine	Mortality, excluding trauma accident deaths	DTP: HR 0.76 (95% CI 0.67–0.88). BCG first 6 months vs. later: HR 0.59 (95% CI 0.47–0.73). Measles vaccine: HR 0.61 (95% CI 0.44–0.85) if no late BCG/DTP vaccination	Background characteristics, sex, religion, maternal age, education, distance from hospital, birth order, and time-dependent covariates
Vaugelade et al. ^Citation14	Prospective cohort study (Burkina Faso)	BCG vs. no BCG, DTP vaccine	9085 children from 26 villages	Mortality	BCG: RR 0.37 (95% CI 0.29–0.48). DTP 1 dose: RR 0.24 (95% CI 0.13–0.43) DTP 2 doses: RR 0.80 (0.58–1.12)	Sex, area, age of mother, health service use, paternal factors, mode of birth, dispensary in village, and health problems
Aaby et al. ^Citation28	Prospective cohort (Guinea- Bissau)	BCG and DTP vaccines	1657 children aged 2–8 months. DTP vaccination vs. no vaccination BCG vaccination vs. no vaccination	Mortality	DTP: RR 1.92 (95% CI 1.04–3.52) BCG: RR 0.63 (95% CI 0.30 − 1.33)	Age, sex, season, period, and region
Aaby et al. ^Citation29	Prospective cohort (Guinea-Bissau)	OPV only vs. DTP and OPV	3459 hospitalized children aged <5 years 719 received either OPV only or DTP and OPV	In-hospital case fatality ratio	Case fatality ratio for immunization with OPV alone: 0.29 (95% CI 0.11–0.77) vs. DTP and OPV	Age, number of vaccination doses, and period (before and after the polio campaign)
Veirum et al. ^Citation30	Prospective cohort (Guinea-Bissau)	Measles vaccine	Hospitalized children aged 1.5–17 months. 259 vaccinated with a measles vaccine vs. 139 unvaccinated children	In-hospital case fatality ratio	Case fatality ratio 0.51 (95% CI 0.27–0.98). Pneumonia patients 0.28 (95% CI 0.07–0.91). Malaria: 0.40 (95% CI 0.13–1.18).	Age, season, and year of hospitalization
Roth et al. ^Citation31	Retrospective cohort (Guinea- Bissau)	BCG vaccine vs. no BCG	845 LBW children, 182 vaccinated in the first week of life	Mortality	RR 0.17 (95% CI 0.06–0.49) RR 0.07 (95% CI 0.01–0.62) for vaccination in the first week of life vs. no vaccination	Birth weight, ethnicity, year of birth, birth during weekdays, maternal education and age, place of residence, sex, and season
Moulton et al. ^Citation32	Prospective cohort (India)	BCG, DTP, and OPV	10274 infants: 6775 received DTP (44.2% received ≥1 dose of each vaccine (BCG, DTP, and OPV)) 12.7% received at least one BCG, 3 DTP, and 3 OPV doses by age 6 months	Infant mortality<age 6 months	Receipt of both BCG and DTP associated with higher mortality vs. vaccination with only one vaccine: HR 2.4 (95% CI 1.2–5.0)	Sex and familial/household characteristics (propensity score)
Roth et al. ^Citation33	Prospective cohort (Guinea-Bissau)	BCG vaccine scar vs. no scar	Cohort A: 1676 with scar and 137 no scar. Cohort B: 1159 with a scar 458 no scar.	Mortality	BCG scar 0.43 (95% CI 0.28–0.65) Malaria mortality: 0.32 (95% CI 0.13–0.76)	The same as in the study by Garly et al.^Citation27
Lehmann et al. ^Citation34	Prospective cohort (Papua New Guinea)	BCG, DTP, and measles vaccine	3502 children who received one or more scheduled vaccines vs. 546 unvaccinated children	Mortality up to age 2 years	BCG: HR 0.17, (95% CI 0.09–0.34) Measles vaccine: HR 0.42 (95% CI 0.17–1.01) DTP ≥1 doses: HR 0.27 (95% CI 0.16–0.44)	Propensity score (region, twinstatus, sex, maternal age, birth order, and date of birth), access to health care, or vaccination seeking behavior
Aaby et al. ^Citation35	Prospective cohort (Guinea-Bissau)	Polio vaccine	6103 children aged <5 years received 1 or more OPV doses vs. 875 who did not receive OPV	Under 5 mortality/ admissions	Mortality: RR 0.46 (95% CI 0.18–1.15) Age <6 months at the time of vaccination RR 0.09 (95% CI 0.01–0.85) Hospitalizations RR 0.27 (95% CI 0.10–0.76). For children aged 6–59 months: no effect of OPV vaccination on mortality, but measles vaccine: RR 0.44 (95% CI 0.28–0.69)	Age, background factors, routine immunizations, antenatal consultations, and arm circumference
Elguero et al. ^Citation36	Prospective study (Senegal)	BCG, DTP, and measles vaccine	Cohort 1 N = 7796 children Cohort 2 N = 3573 children	Mortality	BCG and DTP combination vs. no vaccination Cohort 1: RR 0.70 (95% CI 0.50–0.97) Cohort 2: RR 0.59 (95% CI 0.46–0.74) Measles vaccine Cohort 1: RR 0.98 (95% CI 0.75–1.27) Cohort 2: RR 0.87 (95% CI 0.57–1.30)	Sex, birth rank, maternal age and education, number of siblings, and siblings who had died
Aaby et al. ^Citation37	Randomized clinical trial (Sudan, Congo)	Measles vaccine	Sudan trial: 510 children randomized at 5 months of age to 2 high-titer measles vaccination groups Edmonston–Zagreb (N = 170), Connaught (N = 170), Control (N = 170) meningococcal vaccine. The Kinshasa study included 1023 children who received one dose of HTMV at 6 months or 2 doses at 3.5 and 9.5 months		Combined analysis comparing the two high-titer groups with controls, the measles vaccine groups RR 0.09 (95% CI 0.01–0.72).	Randomization
Aaby et al. ^Citation38	Prospective cohort study (Malawi)	BCG, DTP, and measles vaccine	747 children BCG and DTP: age 8 days-8 months Measles vaccine: age 9–18 months	Mortality	BCG vs. no BCG: RR 0.69 (95% CI 0.31–1.54) Female: RR 0.63 (95% CI 0.20–2.01) Male: RR 0.72 (95% CI 0.26–1.96) DTP vs. no DTP: DTP1: RR 0.99 (95% CI 0.45–2.19) Female: RR 1.27 (95% CI 0.44–3.66) Male: 0.81 (95% CI 0.30–2.17) DTP2: RR 0.55 (95% CI 0.22–1.35) Female: RR 0.65 (95% CI 0.19–2.17) Male: RR 0.48 (95% CI 0.16–1.45) DTP3: RR 0.18 (95% CI 0.06–0.53) Female: RR 0.23 (95% CI 0.06–0.89) Male: RR 0.15 (95% CI 0.04–0.57) Measles vaccine vs. no measles vaccine: RR 0.47 (95% CI 0.19–1.14) Female: RR 0.23 (95% CI 0.04–1.27) Male: RR 0.62 (95% CI 0.22–1.80)	Age, HIV status of the mother, and presence at last examination
Roth et al. ^Citation39	Prospective cohort (Guinea- Bissau)	BCG vaccine	BCG skin test and scar: Age 2 months: 2332 children Age 6 months: 1817 children	Mortality	Tuberculin skin test at 2 & 6 months Both sexes: RR: 0.54 (95% CI 0.30–0.99) Male: 0.84 (95% CI 0.39–1.82) Female: 0.31 (955 CI 0.11–0.88) BCG scar Both sexes: RR:0.55 (95% CI 0.31–0.96) Male: 0.88 (95% CI 0.34–2.30) Female: 0.41 (95% CI 0.21–0.82)	Low birth weight, vaccination technique, BCG type, age vaccination, sex, vaccination place, birthplace, ethnicity, residential place, electricity, maternal education, and marital status
Aaby et al. ^Citation40	Prospective cohorts (Guinea- Bissau)	Measles vaccine and DTP	Hospitalized children in 2 periods; 1990–1996 and 2001–2002, who received measles vaccine before hospitalizations Period 1: 411 children received DTP before measles vaccine &and 63 received DTP with or after measles vaccine. The respective figures in period 2: 204 and 101 children	In-hospital case fatality ratio (CFR)	DTP with/after measles vaccine vs measles vaccine as the most recent Period 1: CFR 2.53 (95% CI 1.37–4.67) Period 2: CFR 1.77 (95% CI 0.92–3.41) Both periods: CFR 2.10 (95% CI 1.34–3.27)	Age, period, and sex
Aaby et al. ^Citation41	Randomized trials (Guinea- Bissau)	Measles vaccine, DTP, and IPV	9544 children enrolled in 4 trials Edmonston-Zagreb (EZ) measles vaccine trial 1: a medium dose of EZ or an IPV from 4 months of age and IPV at 9 months Control group: standard Schwarz measles vaccine EZ trial 2: intervention group high-titer EZ vaccine In the two dose trials: measles vaccine or IPV at 6 months and measles vaccine at 9 months	Mortality	IPV: RR 1.52 (95% CI 1.02–2.28) for girls; Measles vaccine: RR 1.01 (95% CI 0.69–1.46) girls IPV and measles vaccine at 9 months or later: RR 0.88 (95% CI 0.68–1.14). DTP < 3 doses and after measles vaccine or IPV: RR 1.3 (95% CI 0.97–1.73). Girls RR: 1.61 (95% CI 1.08–2.40). Boys RR: 1.02 (95% CI 0.67–1.54)	Randomization
Chan et al. ^Citation42	Prospective cohort (Philippines)	BCG and DTP	14,537 children aged <30 months BCG and DTP vaccinated children vs. children who received BCG vaccine but not DTP vaccine	Mortality	Both sexes: HR 0.43 (95% CI 0.21–0.88) Male: HR 0.32 (95% CI 0.14–0.73) Female: HR 0.86 (95% CI 0.18–4.23)	Age of BCG vaccination, maternal education, birth weight, ownership of TV/radio, prenatal care, sex, household cluster, and socioeconomic status
Roth et al. ^Citation43	Randomized trial (Guinea- Bissau)	BCG vaccine (revaccination)	2871 children aged 19 months to 5 years with low/no reactivity to tuberculin 1437 BCG revaccination group 1434 BCG unvaccinated, and control group	Mortality hospital admission	Mortality: HR 1.20 (95% CI 0.77–1.89) Hospitalization: RR 1.04 (95% CI 0.81–1.33)	Randomization
Aaby et al. ^Citation44	Randomized controlled trial (Guinea-Bissau)	Measles vaccine	6648 children aged 4.5 months who received 3 doses of DTP vaccine ≥4 weeks before enrollment were randomized into Group A: Edmonston-Zagreb measles vaccine at age 4.5 and 9 months; Group B: no vaccine at age 4.5 months and Edmonston-Zagreb vaccine at age 9 months; Group C: no vaccine at age 4.5 months and Schwarz measles vaccine at age 9 months	Main outcome: mortality between age 4.5 and 36 months	Mortality between 4.5–36 months: group A vs. B and C RR 0.78 (95% CI 0.59 − 1.05). Female: RR 0.64 (95% CI 0.42 to 0.98), Male: RR 0.95 (95% CI 0.64–1.42) (interaction test, P = .18)	Randomization
Aaby et al. ^Citation45	Randomized trial (Guinea-Bissau)	BCG vaccine	2320 LBW infants randomly assigned to receive (N = 1168) BCG vaccine at birth (early) or later (current practice) (N = 1152)	Infant mortality	RR 0.83 (95% CI 0.63–1.08)	Randomization
Biering-Sorensen et al. ^Citation46	Randomized trial (Guinea- Bissau)	BCG vaccine	105 LBW children presenting for first vaccination at health center were randomized to receive BCG vaccine immediately (N = 51) or later (usual practice N = 54)	Mortality	3 days of enrollment: RR 0.17 (95% CI 0.02–1.35) Age 1 month: RR 0.28 (95% CI 0.06–1.37) After age 2 months: RR 0.27 (95% CI 0.07–0.98) Infancy: RR 0.41 (95% CI 0.14–1.18)	Randomization
Aaby et al. ^Citation47	Prospective cohort within a randomized trial (Guinea- Bissau)	DTP and BCG	2320 LBW children 1212 DTP vaccinated children vs. 618 DTP unvaccinated at 2 months of age. 1182 received BCG early vs. 1161 received BCG later	Mortality up to age 6 months	DTP vs. no DTP age 2 months: RR 2.45 (95% CI 0.93–6.45) for girls RR 0.53 (95% CI 0.23–1.20) for boys. DTP and early BCG: RR 4.33 (95% CI 1.54–12.2) DTP and delayed BCG: RR 1.71 (95% CI 0.73–4.01)	MUAC, socioeconomic status, nutritional status, health, birth weight study area, sex, twinning, birth order, breastfeeding status, consultations, maternal MUAC, mother’s age, ethnic group, mother’s schooling, living with the father, BCG arm, and vitamin A.
Hirve et al. ^Citation48	Prospective cohort (India)	BCG, DTP, and MV	4138 children BCG only, N = 896 DTP after BCG, N = 827 BCG + DTP1 simultaneously, N = 918 DTP1 only, N = 1797 DTP after BCG + DTP simultaneously, N = 717 BCG after DTP, N = 647 MV only, N = 830 MV and DTP simultaneously, N = 82 DTP after MV, N = 46	Mortality	BCG and DTP simultaneously or BCG as most recent vaccination: RR 0.15 (95% CI 0.03–0.70) vs. DTP the last vaccine. BCG only vs. DTP recommended schedule: RR 0.37 (95% CI 0.07–2.11). BCG most recent vs. unvaccinated: RR 0.27 (95% CI 0.06–1.14). DTP most recent vs. unvaccinated: RR 0.76 (95% CI 0.37–1.55). BCG most recent vs. DTP most recent: RR0.24 (95% CI 0.05–1.20). MV and DTP simultaneously: RR 4.77 (95% CI 0.33–70.2). DTP after MV: RR 15.9 (95% CI 2.12–119)	Nutritional status, birth weight, recent weight, and age
Benn et al. ^Citation49	Prospective cohort (Guinea- Bissau)	DTP	474 children 300 received MV or IPV at 6 months of age. 455 received MV at 9 months of age. 31% (141/455) with missing ≥1 DTP doses	Mortality	Missing DTP: RR 3.55 (95% CI 1.23–10.26) for girls RR 0.97 (95% CI 0.34–2.8) for boys	Weight of age, mid-upper arm circumference, sex, and age
Welaga et al. ^Citation50	Prospective cohort (Ghana)	MV and DTP	3082 children aged 6–35 months 1989 children had health card 1093 children without health card	Mortality	Measles vaccine at enrollment vs. only DTP RR 0.51 (95% CI 0.27–0.97)	Age, zone, weight-for-age, ownership of radio, and sex
Krishnan et al. ^Citation51	Retrospective cohort (India)	BCG, DTP,and measles	12,142 children, N = 5479 females, N = 6663 males	Mortality	DTP vaccination vs. no vaccination Female: HR 1.65 (95% CI 1.17–2.32). Male: HR 2.21 (95% CI 1.24–3.93). BCG vs. no BCG: HR 1.06 (95% CI 0.67–1.67) for girls Measles vaccination vs. no vaccination: HR 1.34 (95% CI 0.85–2.12) for girls	Wealth index, access to health care, parental education, type of family, birth, age, and sex
Fisker et al. ^Citation52	Observational study (Guinea- Bissau)	MV, YF, and DTP	2331 children 685 children received MV only 358 received MV + DTP, 940 received MV+ YF only 348 received MV+ YF + pentavalent	Mortality	Follow-up 6 months: Live and inactivated vs. live vaccines only: RR 3.24 (95% CI 1.20–8.73); MV vs. MV + DTP RR 1.56 (95% CI 0.39–6.29) MV + YF vs. MV + YF + penta RR 7.73 (95% CI 1.79 − 33.4) Follow-up 12 months: Live and inactivated vs. live vaccines only: RR 1.86 (95% CI 0.89–3.89) MV vs. MV + DTP: RR 1.14 (95% CI 0.44–2.95) MV+ YF + pentavalent vs. MV + YF only: RR 4.85 (95% CI 1.46–16.2)	Age, sex, ethnicity, urban/rural, season, morbidity, maternal education, whether the mother signed the form, and stunting
Fisker et al. ^Citation53	Observational cohort study (Guinea-Bissau)	Measles vaccine	Data from 1999 to 2006 were used to compare mortality between vaccinated and unvaccinated children	Mortality after age 12 months	RR: 0.71 (95% CI 0.56–0.90) in vaccinated vs. unvaccinated children Female: 0.59 (95% CI 0.43–0.80) Male: 0.87 (95% CI 0.62–1.23)	Ethnicity, maternal age and schooling years, and village
Aaby et al. ^Citation54	Prospective cohort (Senegal)	BCG, DTP, and Measles vaccine	4133 children. BCG-first, BCG + DTP1-first, or DTP1-first. DTP vs. measles vaccine between 9 and 24 months of age, as 9 months is the minimum age for measles vaccine	Mortality	BCG + DTP1-first vs. unvaccinated: RR 0.69 (95% CI 0.53–0.89) DTP with or after measles vaccine vs. measles vaccine-only as the most recent RR 2.59 (95% CI 1.32–5.07). Differential effect by sex	Sex, age, birth year, season, and village
Lund et al. ^Citation55	Randomized trial (Guinea-Bissau)	Oral polio vaccine	7012 healthy normal-birth-weight neonates were randomized to BCG only (N = 3467 intervention group) or OPV0 with BCG (N = 3494 usual practice). All children were to receive OPV with pentavalent vaccine (DTP, H. influenzae type b, and hepatitis B) at age 6, 10, and 14	Infant mortality	Both sexes HR 0.83 (95% CI 0.61–1.13). Male: HR 0.72 (95% CI 0.47–1.10) Female: HR 0.97 (95% CI 0.61–1.54) Enrollment in the first 2 days of life HR 0.58 (95% CI 0.38–0.90) From enrollment in OPV campaigns HR 0.68 (95% CI 0.45–1.00)	Randomization
Aaby et al. ^Citation56	Natural experiment within randomized controlled trial (Guinea-Bissau)	OPV and measles vaccine	N = 2129 had 2 doses of measles vaccine at age 4.5 and 9 months, N = 4288, 1 dose of measles vaccine at 4.5 or 9 months. 854 children did not receive OPV	Mortality (2-dose/1-dose measles vaccine) by OPV status	RR (2-dose/1-dose measles vaccine): 0.70 (95% CI 0.52–0.94) No OPV at birth: RR 1.04 (95% CI 0.53–2.04) OPV in the 1^st week of life: RR 0.45 (95% CI 0.29–0.71) OPV after age 1 month: RR 3.63 (95% CI 0.87–15.2)	District and calendar time
Fisker & Thysen ^Citation57	Prospective cohort (Guinea-Bissau)	Sequence of vaccines measles/third pentavalent vaccine (DTP-Hib-Hepatitis B) (Penta)	7094 measles vaccinated children aged 9 months to 5 years 5426 children received Penta before measles vaccine 1032 children received Penta with or after measles vaccine	Mortality	Receipt of Penta after MV vaccine: HR 1.19 (95% CI 0.84–1.69) vs. measles after Penta. Receiving missing Penta doses on the visit date: HR 1.87 (95% CI 0.96–3.65)	Period, region, and maternal education
Biering-sorensen et.al.^Citation58	Randomized controlled trial (Guinea-Bissau)	BCG vaccine	Low-weight neonates were randomized to receive early BCG vaccination (N = 2083) vs. control (current policy, later N = 2083) BCG vaccination later	Mortality Infectious disease mortality	Mortality: RR 0.70 (95% CI 0.47–1.04) RR 0.66 (95% CI 0.44–1.00) censoring OPV campaigns. Infectious disease mortality RR 0.57 (95% CI 0.35–0.93)	Randomization
Byberg et al. ^Citation59	Prospective cohort (Guinea-Bissau)	Measles campaign vaccination	5633 children aged 9–59 months received measles vaccine vs. 1006 did not	Mortality	Both sexes HR 0.28 (95% CI 0.10–0.77) Female: HR 0.17 (95% CI 0.05–0.59) Male: 0.92 (95% CI 0.10–8.82)	Region, sex, maternal schooling, electricity, ethnicity, vaccination status at time of campaign and previous routine measles vaccine
Upfill-Brown et al. ^Citation15	Randomized trial (Bangladesh)	Polio vaccine	Following vaccination with tOPV at age 6, 10, and 14 weeks, infants were randomly assigned to receive IPV (N = 299) or tOPV (N = 315) at age 39 weeks	Diarrhea through age 52 weeks	Proportion with diarrhea was not different between the groups (P = .18). The mean number of diarrhea days was lower in the tOPV group 5.9 vs. 6.7 in the IPV group (P = .0037)	Randomization
Byberg et al. ^Citation60	Prospective cohort (Guinea-Bissau)	DTP and measles vaccine, as the most recent vaccine	The most recent vaccine: measles vaccine only 775 children, DTP 611 children, DTP≥measles vaccine 614 children, and other/no vaccines 132 children	Mortality age 12–29 months	Compared to measles vaccine only: DTP: HR: 0.92 (95% CI 0.59–1.43) DTP≥measles HR 1.44 (95% CI 0.97–2.15) Other/no vaccines: HR 2.58 (95% CI 1.54–4.33)	Age, cohort, and sex
Benn et al. ^Citation61	Prospective study (Guinea-Bissau)	OPV at birth	4345 children 3383 received OPV at birth vs. 962 who did not	Infant mortality	Original results: OPV vs. no OPV at birth Both sexes HR 1.45 (95% CI 0.97–2.17) Female: HR 0.87 (95% CI 0.53–1.44) Male: HR 2.82 (95% CI 1.41–5.65) Before OPV campaign Both sexes: HR 0.98 (95% CI 0.60–1.60) Female: HR 0.64 (95% CI 0.35–1.19). Male: HR 1.58 (95% CI 0.74–3.35) After OPV campaign Both sexes: HR 1.87 (95% CI 0.75–4.64) Female: HR 0.95 (95% CI 0.33–2.72). Male: HR 8.26 (95% CI 0.1.02–66.8) Risk reduction in mortality after campaign: Female: 67% (95% CI 42%–81%) Male: 75% (37–90%)	Season, suburb, maternal education, place of enrollment, and arm circumference
Do et al. ^Citation62	Randomized controlled trials (Guinea-Bissau)	Measles vaccine 2 doses at age 18 weeks and 9 months	Randomly assigned children to receive 2 doses of measles vaccine at age 18 weeks and 9 months (N = 1068 intervention) or 1 dose (N = 557, current practice control) at age 9 months	Symptoms	Diarrhea HR 0.89 (95% CI 0.82–0.97) Vomiting HR 0.86 (95% CI 0.75–0.98) Fever HR 0.93 (95% CI 0.87–1.00)	Randomization
Schoeps et al. ^Citation18	Randomized controlled trial (Burkina Faso)	Measles vaccine and DTP	Randomly assigned children to receive standard measles vaccine at age 9 months (N = 2238 control) or a measles vaccine 4 weeks after dose 3 of DTP-containing pentavalent vaccine and the standard measles vaccine at age 9 months (N = 2258 intervention)	Composite end point admission or mortality age 4.5–36 months	HR 1.00 (95% CI 0.83–1.20). Results from the per-protocol and intention-to-treat analysis were similar	Randomization
Fisker et al. ^Citation63	Randomized trial (Burkina Faso & Guinea-Bissau)	Measles vaccine	8309 children aged 4–7 months were randomized to an extra early standard dose of measles vaccine (Edmonston-Zagreb strain) (N = 4153) or no extra dose 4 weeks after DTP3-hepatitis B-Hib vaccine (N = 4156)	Mortality	HR 1.05 (95% CI 0.75–1.46).	Randomization
Thysen et al. ^Citation64	Prospective cohort (Guinea-Bissau)	BCG vaccine during the neonatal period	33,137 children received neonatal BCG vaccine and 6284 children did not receive BCG during the neonatal period (87% were vaccinated later)	Mortality	TB-exposed: HR 0.57 (95% CI 0.26–1.27) TB unexposed: HR 0.57 (95% CI 0.47–0.69)	Twin status, maternal age and year of birth
Jensen et al. ^Citation65	Randomized controlled trials (Guinea-Bissau)	BCG vaccine	In 3 clinical trials, 6525 infants were randomized to receive BCG at birth or delayed BCG vaccine	Mortality	November-January: HR 0.22 (95% CI 0.11–0.45) February-October: HR 0.86 (95% CI 0.61–1.21) P = .001 for interaction BCG ˟month High malaria transmission months: Pooled HR 0.41 (95% CI 0.25–0.66) Low malaria transmission-months: HR 0.83 (95% CI 0.56–1.23) P = .02 for interaction BCG ˟period	Randomization
Steiniche et al. ^Citation17	Randomized controlled trial (Guinea-Bissau)	Measles vaccine	3548 children randomly assigned to receive measles vaccine at age 4.5 and 9 months or at 9 months.	Out/inpatient consultations, growth up to age 9 months	No effect of early measles vaccine vs. no vaccine on mid-upper-arm-circumference for-age, weight-for-age Z score. Consultations: HR 1.03 (95% CI 0.92–1.16).	Randomization
Prentice et al. ^Citation16	Randomized control trial (Uganda)	BCG vaccine	560 healthy infants were randomly assigned to receive BCG at birth (N = 280) or age 6 weeks (N = 280)	Nontuberculous infectious disease	During the first 6 weeks of life, HR 0.71 (95% CI 0.53–0.95) Age 6–10 weeks follow-up: HR 1.10 (95% CI 0.87–1.40)	Randomization

BCG = Bacillus Calmette–Guérin; CI = Confidence interval; DtaP = Diphtheria, tetanus and pertussis; HR = Hazard ratio; IPV = Inactivated polio vaccine; Hib = Haemophilus influenzae type b; LBW = Low born weight; OPV = Oral polio vaccine; RR- Relative risk (Rate ratio).

Live attenuated vaccines

Measles vaccine

Studies published from 2000 onward^{12,13,25,26,34,53,59} supported earlier publications^5–10 regarding the potential protective effect of measles containing vaccines on child survival (Table 1). Some noted this effect among females but not among males.^44,53,59 Veirum et al. found that the in-hospital case fatality, especially among children hospitalized for pneumonia or malaria, was significantly lower among children who had a measles vaccine than those who were not vaccinated for measles.³⁰ Studies from Senegal and Malawi showed no significant association between measles vaccination and child mortality.^36,38

In a randomized controlled trial conducted in Guinea-Bissau, 6648 children aged 4.5 months who received three doses of DTP vaccine ≥4 weeks before enrollment were randomized to receive Edmonston-Zagreb measles vaccine at age 4.5 and 9 months (group A), no vaccine at age 4.5 months and Edmonston-Zagreb vaccine at age 9 months (groups B), or no vaccine at age 4.5 months, and Schwarz measles vaccine at age 9 months (groups C). Intention-to-treat analysis showed no significant difference in mortality between age 4.5 and 36 months in group A compared to groups B and C, RR = 0.78 (95% CI 0.59–1.05); however, a significant association was found in females but not males (interaction test, P = .18) and in the per-protocol analysis.⁴⁴ In a two-center more recent randomized controlled trial from Guinea-Bissau and Burkina Faso,^18,63 children were randomly assigned to receive standard measles vaccine at age 9 months (control group) or an additional measles vaccine 4 weeks after recipient of the third dose of pentavalent vaccine (DTP-hepatitis B,H.influenzae type b vaccines) showed no significant difference between the groups in mortality.⁶³ An analysis of a composite endpoint of mortality/hospitalizations in the trial from Burkina Faso also showed null results.¹⁸ The effect of the number of measles vaccine doses on morbidity was assessed in a sub-study within a randomized controlled trial in which children were randomly assigned to receive two doses at age 18 weeks and nine months (intervention group) or one dose (control group) at age nine months.⁶² This study showed 7–14% lower risk for maternal reports on their child having diarrhea, vomiting, and fever, but not other symptoms, among the intervention vs. the control group. An analysis of secondary endpoints from a clinical trial showed that an early additional measles vaccine dose had no significant benefit on child nutritional status or medical consultations.¹⁷

Altogether, these studies mostly showed an advatnage in child survival in relation to measles vaccination.

BCG vaccine

In an analogy to measles vaccine, it was postulated that BCG vaccine, another live attenuated vaccine, might also have a beneficial effect on child survival. Kristensen et al.¹² conducted a prospective cohort study in Guinea-Bissau during 1990–1996, in which the recommended vaccination schedule included BCG and polio at birth; DTP and OPV at six, 10, and 14 weeks; and measles at age nine months. They showed that children who were vaccinated with a BCG vaccine had significantly 28% lower risk of mortality by age  sixmonths compared to unvaccinated children.¹² Breiman et al. in a large prospective cohort study of 37,894 children from Bangladesh showed 41% significantly lower risk of mortality in relation to BCG vaccination during the first six months of life vs. later vaccination.¹³ The inverse significant association between BCG vaccination and child mortality was demonstrated in other cohort studies from Burkina Faso¹⁴ and Papua New Guinea,³⁴ but such association was not significant in a study from Malawi.³⁸ A study from Guinea-Bissau showed a protective significant effect of BCG vaccine, compared to no BCG vaccination during the neonatal period (unvaccinated or later vaccinated children) among children classified as unexposed to tuberculosis. The association was not significant in children classified as exposed to tuberculosis, although it was of similar direction and magnitude.⁶⁴

Prospective studies showed that among BCG-vaccinated children, those who developed a scar following vaccination had significantly lower mortality than those who did not develop a scar^27,33,39 as well as children who had a positive tuberculin test than those with a negative result.^27,39 BCG scar was associated with a significant reduction in the risk of malaria mortality RR = 0.32 (95% CI 0.13–0.76) (Table 1).

The formation of BCG scar might be affected by multiple factors, e.g. vaccination technique, dose of vaccine injected, and type of vaccine,⁶⁶ but this does necessarily imply vaccine take or failure. Decay in tuberculin reactivity following BCG vaccination has been reported;^66,67 however, it does not affect vaccine efficacy (reviewed in Ref. ⁶⁶). Scar formation and positive tuberculin after BCG vaccination might be a proxy of the immune response to the vaccine, and therefore, these studies^27,39 strengthened earlier findings of inverse association between BCG vaccination and mortality.

A randomized controlled trial of BCG-vaccinated children with low/no reactivity to the tuberculin test showed no significant difference in mortality or hospitalizations between BCG-revaccinated children and those who did not receive BCG revaccination.⁴³

Low birth weight (LBW) newborns comprise a main risk group for infant mortality, therefore, if BCG has beneficial effects on child survival, it is essential to deliver the vaccine early to this group. Roth et al. in a prospective study showed that BCG-vaccinated LBW babies had survival benefits, which were greater when the vaccine was given during the first week of life compared to unvaccinated children: RR = 0.07 (95% CI 0.01–0.62).³¹ A randomized controlled trial of LBW newborns (N = 2320) from Guinea-Bissau showed that BCG given at birth vs. later vaccination (local practice) was associated with 17% lower risk of infant mortality, although the effect was not statistically significant RR = 0.83 (95% CI 0.63–1.08).⁴⁵ Another trial of 105 LBW babies who presented to health centers in Guinea showed that immediate BCG immunization was associated with significantly reduced risk of mortality after age two months.⁴⁶ Biering-sorensen et al.⁵⁸ showed that the beneficial effect of early BCG vaccination on survival of LBW infants in Guinea Bissau was stronger in relation to infectious disease mortality. A pooled analysis of data from three clinical trials also supported the protective effect of BCG vaccine on child survival, but this impact varied significantly by calendar month and malaria high-transmission season.⁶⁵ An analysis of cytokines in one of the trials showed overall increased innate cytokine responses particularly of interleukin (IL)-1β, IL-6, tumor necrosis factor (TNF)-α, and interferon (IFN)-γ in relation to BCG vaccination.⁶⁸ BCG-related cytokine enhancement and vaccine-specific response resembled the seasonal related effect of BCG-child survival association.⁶⁵ A randomized trial of infants from Uganda¹⁶ showed 29% lower risk of physician-diagnosed nontuberculous infectious disease for BCG vaccinated at birth only during the first few weeks of life.

Collectively, these findings indicate that early vaccination with BCG of infants, including LBW babies, from high-child mortality settings likely has benefits on overall child survival. Evidence on the impact of BCG vaccine on nontuberculosis infectious disease morbidity is scarce and should be studied further.

OPV

OPV is usually given at the same age/visit as DTP vaccine, and thus, deciphering the nonspecific impact of OPV vs. DTP vaccination on children’s health has been challenging, and less abundant evidence exists on the assocation between OPV and nonspecific vaccine effects (Table 1).

Kristensen and colleagues showed that children in Guinea-Bissau who received one dose of OPV had increased risk of mortality compared to those who were not vaccinated.¹² Another cohort study showed a significant difference in child mortality between children who were vaccinated with OPV compared to unvaccinated children, although significant inverse relationships were noted between OPV vaccination before age six and mortality and hospitalizations.³⁵ In a randomized study, healthy normal-birth-weight neonates were randomized to receive BCG only (intervention group) or first dose of OPV with BCG (usual practice) and showed no significant difference between the groups in infant mortality, although an inverse significant association was found when children were enrolled during the first two days of life and until OPV vaccination campaigns.⁵⁵ During a clinical trial with two doses of measles vaccines, shortage in OPV doses occurred. The authors showed a significant inverse association between Receipt of two doses of measles vaccine vs. one dose among children who received OPV during the first week of life, but not among those who were vaccinated with the first dose of OPV after age one month.⁵⁶ Another analysis showed no significant difference in child mortality between children who received OPV at birth compared to those who did not, neither before or after OPV campaigns.⁶¹ This study showed risk reduction in child mortality following OPV campaigns compared to the preceding period, in both sexes.⁶¹

A randomized controlled trial conducted in Bangladesh assessed differences in the incidence of diarrheal diseases between children who were randomized to receive inactivated polio vaccine (IPV) or OPV at age 39 weeks.¹⁵ There was no significant difference between the groups in the proportion of children who had diarrhea; however, the OPV group had significantly fewer diarrhea days.¹⁵

Inactivated vaccines

DTP whole cell vaccine

The association between DTP whole cell vaccine and child survival was examined in multiple observational studies with inconsistent findings^{12,13,34,38,47,51} (Table 1). It was shown that vaccination with one dose of DTP was associated with significantly increased risk of child mortality; however, for two-three doses, such association was not significant.¹² Additional studies also suggested elevated risk for child mortality in relation to DTP immunization,^28,51 while other studies showed significantly reduced risk for mortality among children who were vaccinated with DTP compared to unvaccinated ones.^13,14,34,38 A meta-analysis of observational study showed that vaccination with DTP was not significantly related to all-cause child mortality RR 1.38 (95% CI 0.92–2.08).²⁴ A reanalysis that systematically assessed all DTP studies included in the above-mentioned metal-analysis showed similar non-significant difference in all-cause child mortality: 1.32 (95% CI 0.83–2.08). The authors identified in a meta-regression several significant factors that affected the results including that study location, studies using the landmark approach, and studies with high risk of exposure misclassification, which were significantly associated with increased RR estimates. In contrast, studies with high risk of selection bias showed borderline significance.⁶⁹

Other inactivated vaccines

Limited evidence exists on the association between other inactivated pediatric vaccines and child mortality. Lehmann et al. found no significant differences between child mortality and recipient of hepatitis B vaccine and pneumococcal vaccine.³⁴ Future studies are needed to explore potential beneficial effects of inactivated vaccines on children’s health.

Discussion

Our review showed that potential nonspecific effects of different vaccines on children’s health have gained growing attention over the past two decades. Utmost evidence is based on observational studies. Most studies provided supportive evidence of an inverse association between measles vaccination and child mortality in populations with high child mortality, while impact on non-measles morbidity and children’s growth might be small or nonexistent. Vaccination with a BCG vaccine was related to lower risk of mortality among young children, including LBW babies. Evidence on the impact of BCG vaccine on nontuberculosis infectious diseases morbidity is scarce but suggests a potential beneficial impact. The association between DTP whole cell vaccine and child survival was examined in multiple observational studies with inconsistent findings, and results might be affected by bias and confounding. Evidence on the role of OPV remains elusive, showing the potential beneficial impact on child survival.

The exact mechanism that might explain potential nonspecific beneficial effects of vaccines on child survival is not fully clear. Kandasamy and colleagues in their systematic review of the nonspecific immunological effects of childhood vaccines⁷⁰ demonstrated consistent patterns of nonspecific immunological responses in some BCG and measles vaccines studies. It was proposed that heterologous T cell effects and priming of innate immunity play a role.⁷¹ For example, a human challenge study demonstrated that BCG vaccination induced genome-wide epigenetic reprogramming of monocytes and protected against experimental infection with an attenuated yellow fever virus vaccine strain. Epigenetic reprogramming was accompanied by functional alterations suggestive of “trained immunity”, with a main role for IL-1β as a mediator of trained immunity responses.⁷² The production of TNFα and IL-1β to mycobacteria or unrelated pathogens was higher two weeks and three months after BCG vaccination, but these differences were less marked one year following vaccination. The heterologous production of T helper 1 (Th1) (IFN-γ) and Th17 (IL-17 and IL-22) immune responses to nonmycobacterial stimulation remained strongly elevated one year after BCG vaccination.⁷³ Jensen et al. showed that among LBW infants from Guinea-Bissau, BCG significantly increased the in vitro cytokine responses to purified protein derivative of Mycobacterium tuberculosis. BCG was also related to increased responses to heterologous innate stimulation, mainly of the cytokines IL-1β, IL-6, TNF-α, and IFN-γ, thus suggesting that BCG vaccination may contribute to the development of the neonatal immune system.⁶⁸

Most evidence on nonspecific effects of vaccines relies on observational studies, which are prone to confounding and bias. Namely, delayed or missed immunization is likely related to environmental, socioeconomic, and parental characteristics that affect the child survival and health status. The potential effect of healthy vaccinee bias cannot be completely eliminated. Nonetheless, consistent evidence across populations and several well-designed large prospective studies suggests that some vaccines, i.e. BCG and measles vaccines, were inversely related to child mortality. More evidence is needed to assess the role of these and other vaccines in morbidity and children’s growth and development and explore potential biological mechanisms.

Acknowledgments

This work was undertaken in partial fulfillment of the requirements for a PhD degree of Ms. Muna Omar, at the School of Public Health, Sackler Faculty of Medicine, Tel Aviv University.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

The author(s) reported that there is no funding associated with the work featured in this article.