Vaccination in childhood and the risk of childhood cancer – a systematic review and meta-analyses

Several epidemiological studies investigated potential associations between vaccination in childhood and childhood cancer. However, results were inconsistent and a meta-analysis on this topic is lacking.

Systematic searches of the PubMed and Web of Science databases on childhood cancer and vaccination were conducted. We identified 4,163 studies, of which 52 studies fulfilled the inclusion criteria for this review. Pooled odds ratios (OR) were calculated using random effects models if three or more studies for a specific research question in this field were available.

Overall, we observed 75 effect estimates for decreased risks of different childhood cancer types after specific vaccinations, 16 for increased risks, and 351 for non-significant associations within the included studies. Ten studies presented data on associations between specific vaccines and childhood leukemia allowing meta-analyses of four associations. Compared to children without the specific vaccination, the pooled OR of leukemia risk was 1.05 [(95% CI: 0.74, 1.49), number of studies N = 5] for Bacille Calmette-Guérin vaccination, 0.92 [(0.75, 1.14), N = 3] for Hepatitis B vaccination and 1.04 [(0.84, 1.30), N = 4] for Poliomyelitis vaccination. Using the total number of vaccine injections as exposure variable, we observed an OR of 0.63 (0.26, 1.51) for childhood leukemia in children with a high number of injections (> 4 injections) compared to children without injections based on 3 studies.

The pooled analysis of previous studies does not support the hypothesis regarding an association between vaccination and the risk of childhood leukemia. However, all 52 studies had at least one of the following limitations: small sample size, invalid exposure assessment or lack of adjustment. Large studies with valid exposure assessment on immunization and detailed information on possible confounders (e.g. other vaccines, infections) are needed.