Secondary Logo

Journal Logo

Brief Reports

Small Randomized Trial Among Low–birth-weight Children Receiving Bacillus Calmette-Guérin Vaccination at First Health Center Contact

Biering-Sørensen, Sofie MSc*; Aaby, Peter DMSc*,†; Napirna, Bitiguida Mutna MD; Roth, Adam PhD†,§; Ravn, Henrik PhD; Rodrigues, Amabelia PhD*; Whittle, Hilton FRCP; Benn, Christine Stabell PhD*,†

Author Information
The Pediatric Infectious Disease Journal: March 2012 - Volume 31 - Issue 3 - p 306-308
doi: 10.1097/INF.0b013e3182458289


Low–birth-weight (LBW) children do not receive bacillus Calmette-Guérin (BCG) at birth in many low-income countries. Based on observational studies from West Africa and studies with randomized or alternate enrollment design from the United States and United Kingdom, we have formulated the hypothesis that BCG has nonspecific beneficial effects protecting against infections other than tuberculosis.14 The fact that LBW children do not receive BCG at birth made it possible for the Bandim Health Project (BHP) in Guinea-Bissau to test this hypothesis in a randomized clinical trial among LBW children. We tested the effect on overall infant mortality of BCG administered at birth. Compared with the usual practice of delayed BCG vaccination, BCG at birth was associated with a 45% (95% CI = 11%–66%) reduction in the neonatal period when very few of the controls had received BCG.5 With follow-up to 12 months of age, the reduction in the main outcome infant mortality was 17% (−8%–37%). Hence, LBW children may benefit in the neonatal period from receiving BCG at birth. Because LBW is the largest high-risk group in low-income countries, more information on the possible beneficial effect of BCG at birth is needed.

Before this trial,5 we had initiated a similar trial, which was stopped prematurely because of problems in randomization (available at:, NCT00146302). The trial had recruited children at 2 local health centers in the BHP study area and at the national hospital's maternity ward. We subsequently discovered that the problem in randomization had only occurred at the maternity ward and not at the local health centers. Therefore, we are reporting the results of BCG immunization among the LBW children recruited at the health centers, where there was no problem with randomization. These children were not included in the previous report of a similar trial.5



The LBW cohort was initiated to test whether BCG at birth would reduce infant mortality by 25% (available at:, NCT00146302). In this study, we recruited LBW children who came for their first vaccination at 2 of the 3 health centers in the BHP study area (Bandim, Belem). Furthermore, all LBW children from the city of Bissau born at the national hospital were offered participation in the trial. The BCG trial was initiated in November 2002. In 2004, after finding a significant difference in birth weight between the 2 groups, we discovered that the randomization procedures had been faulty in the hospital. The trial was stopped in November 2004, reorganized, and restarted.5 No similar problem occurred at the 2 local health centers. Therefore, we are reporting the effect of randomization to BCG at first health center contact.

Study Design.

BHP conducts a health and demographic surveillance system in 6 districts with a population of around 102,000, 30% of the inhabitants of the capital. All houses in the area are visited every month to register new pregnancies and births. If the child visiting the health center for first vaccination weighed less than 2.5 kg, the BHP field workers explained the purpose of the study, completed a questionnaire with background information, and obtained informed consent. Study procedures including consent, randomization, and follow-up home visits at day 3 and at 2, 6, and 12 months of age have been described in detail elsewhere.5 The visit at 12 months of age served to assess survival during the first year of life.

Statistical procedures which included control for twin pairs have been described previously.5 Pairs of same-sex twins were assigned to the same group to assure that there was no confusion related to who had received which treatment. Combined estimates for the 2 trials (Table 1) were calculated using a Cox model stratified for trial.

Table 1
Table 1:
Mortality Rate Ratios (MRR) at 3 Days, 4 Weeks, 2 Months, 6 Months, and 12 Months of Age in 2 Trials of BCG at Birth to LBW Children


In the period from November 2002 to November 2004, 105 LBW children were enrolled at the health centers, 51 and 54, respectively, in the BCG and the control groups. One child who had an LBW but weighed more than 2.5 kg when enrolled has been excluded. As given in Table, Supplemental Digital Content 1,, there was no difference in the anthropometric measurements between the children in the BCG and the control groups.

Within the first 3 days of enrollment, 6 children died (1 BCG recipient [female] and 5 controls [4 male, 1 female]), the mortality rate ratio (MRR) for BCG versus controls being 0.17 (0.02–1.35) (Table 2). Eight children died in the neonatal period (within 28 days of birth), 2 BCG recipients (1 male and 1 female) and 6 controls (5 male and 1 female), with the MRR being 0.28 (0.06–1.37). At 2-month follow-up, the BCG group had significantly lower mortality (MRR = 0.27 [0.07–0.98]; Table 2). During the first year of life, 16 children died, 5 in the BCG group and 11 in control group, the MRR being 0.41 (0.14–1.18) (P = 0.098). The effect of BCG in infancy was stronger in boys (MRR = 0.21 [0.04–1.03]; P = 0.054) than in girls (MRR = 0.81 [0.19–3.36]; P = 0.77) (P = 0.21 for interaction between early BCG administration and gender).

Table 2
Table 2:
Mortality at 1, 2, 6, and 12 Months of Age According to Randomization Status

As was the case in the previous study,5 the neonatal deaths, most of which occurred within 3 days of enrollment, were ascribed to neonatal sepsis (2 BCG recipients, 3 controls), and respiratory distress and hypoglycemia (3 controls), according to the verbal autopsies. The deaths between 1 and 5 months of age were due to respiratory infection (2 BCG recipients, 1 control), severe malaria (1 BCG recipient, 1 control), and 2 could not be determined because the family had moved (2 controls). There was only 1 death due to acute diarrhea (1 control) between 6 and 12 months of age.

Since the present and the previous trial5 were identical in design, we calculated combined estimates from the 2 trials, as given in Table 1. It can be observed that BCG at birth had a significant beneficial effect on mortality at all time points up to 6 months of age.


The small trial of BCG at the first health center contact supported the beneficial effect of BCG administration in the first months of life among LBW children. It is worth noting that tendencies were similar in both the present and the previous trial.5 In the combined analysis of these trials, BCG at birth reduced mortality significantly after 3 days of follow-up, and at 1, 2, and 6 months of age, compared with the current policy of delayed BCG vaccination.

The strongest beneficial effect of BCG was in the first months of life, in which few of the controls had received BCG (Table 2). This may represent the “true” effect of BCG if it is not modified by other vaccines. The difference between the effect on mortality of early and delayed BCG vaccination was reduced once the controls started receiving BCG and children received diphtheria-tetanus-pertussis (DTP) and oral polio vaccine (OPV) after 6 weeks of age (Table 2). The 2 trials combined provide evidence that early BCG administration can reduce overall infant mortality up to 20%, the MRR at 12 months of age being 0.79 (0.61–1.02). Because LBW children are the largest high-risk group and neonatal mortality has been particularly difficult to reduce in low-income countries, there are good reasons to further study the potential beneficial effects of BCG administration at first contact.

It should be emphasized that although BCG vaccination is recommended at birth to children with normal birth weight, this vaccination is very often delayed for logistic reasons, the median age at vaccination often being more than 1 month.5 One common logistic reason is that health care workers do not like to open a vial of 20 doses of BCG for just one child and therefore provide BCG only on certain days for more children to be present. If BCG has beneficial effects also among children with normal birth weight, it would certainly be cost-effective to open a bottle for just one child. In evaluating and designing vaccination schedules, scientists and policy makers need to take the nonspecific effects of our current and future vaccines into consideration.


1. Kristensen I, Aaby P, Jensen H. Routine vaccinations and child survival: follow up study in Guinea-Bissau, West Africa. BMJ. 2000;321:1435–1438.
2. Roth A, Jensen H, Garly ML, et al.. Should low birth weight infants receive BCG vaccination at birth? Community study from Guinea-Bissau. Pediatr Infect Dis J. 2004;23:544–550.
3. Roth AE, Garly ML, Jensen H, et al.. Bacille Calmette Guerin vaccination and infant mortality. Expert Rev Vaccines. 2006;5:277–293.
4. Shann F. Heterologous immunity and the non-specific effects of vaccines. A major medical advance? Pediatr Infect Dis J. 2004;23:555–558.
5. Aaby P, Roth A, Ravn H, et al.. Randomised trial of BCG vaccination at birth to low-birth-weight children: beneficial non-specific effects in the neonatal period? J Infect Dis. 2011;204:245–252.

BCG; infant mortality; neonatal mortality; nonspecific effects of vaccines

Supplemental Digital Content

© 2012 by Lippincott Williams & Wilkins, Inc.