Jesper Kjærgaard 2016

Kjærgaard, Jesper. 2016. Bacillus Calmette-Guérin vaccination at birth: Effects on early childhood infections, growth, and development.

The Bacillus Calmette-Guérin vaccine (BCG), which is used to protect against tuberculosis, has been associated with a variety of other effects since it was developed almost 100 years ago. Most notably, observational studies and randomized clinical trials from low-income countries indicate that it protects against unrelated infections, i.e. a so-called non-specific effect. The Danish Calmette Study was conducted to study these effects in a high-income population. 

The immune response to BCG is not fully understood but involves a pro-inflammatory profiling of the immune system, also when exposed to unrelated pathogens. Immune changes have been implicated in changes in both child growth and child development and for that reason we also studied these outcomes.

We randomized 4262 children at birth to receive BCG vaccination at birth or to a no-intervention control group. We had pre-specified subgroup analyses of child sex, prematurity, and maternal BCG vaccination. The statistical analysis plan was finalized prior to unblinding of the data. 

Follow-up for the outcomes reported in this thesis consisted of telephone interviews and clinical examination at age 3 and 13 months, as well as online developmental questionnaires distributed to the parents at 12 months and additionally to the parents of premature children at age 6 and 22 months. The outcomes of this thesis were number of parent reported infections, child growth and body composition, and child psychomotor development.

Overall, there was no effect of BCG on either incidence of infections, growth, body composition or psychomotor development. A subgroup analysis of children of mothers who were BCG vaccinated showed a reduced incidence of infections from 0 to 3 months among BCG vaccinated children (incidence rate ratio = 0.62, CI: 0.39 to 0.98), but there was no effect from 3 to 13 months. Previous research has shown that maternal exposure to BCG or mycobacteria can alter the effect of BCG in the offspring, and thus the unexpected lack of effect on overall infections can possibly be explained by the lack of maternal exposure to BCG in our study, as only 17% of the mothers were BCG vaccinated. In the studies where non-specific protective effects of BCG have previously been found, most of the mothers were BCG vaccinated or exposed to mycobacteria. 

Premature children had a non-significant increased risk of infection, which was corroborated by an analysis of hospitalizations for infections (not reported in this thesis). This was also unexpected as previous research indicated a more beneficial effect among low birth weight children. The study did not have power enough to exclude a negative effect of BCG on the development of premature children, and thus a cautious approach to vaccinating premature children may be prudent in a high-income setting.

We succeeded in recruiting the planned number of participants and had high follow-up rates for most outcomes. A limitation is that it was not feasible to blind the parents to the randomization group.

In conclusion, BCG did not have any public health benefit on incidence of infections and did not affect child growth or child development in the present study.