Pregnant women are increasingly considered a priority group for influenza vaccination, but the evidence in favor relies mainly on observational studies, subject to the "healthy-vaccinee bias". Propensity score methods-sometimes applied-reduce but cannot eliminate residual confounding. Meta-analyses of observational studies show relative risks far from the thresholds that would confirm the efficacy of universal vaccination for pregnant women without needing randomized controlled trials (RCTs). Critical articles have shown that in the four RCTs investigating the outcomes of this vaccination, there was a tendency towards higher offspring mortality. In the largest RCT, there was a significant excess of presumed/serious neonatal infections, and also significantly more serious adverse events. Many widely acknowledged observational results (about hormone replacing therapy, vitamin D, omega-3 fatty acids, etc.) were confuted by RCTs. Therefore the international drive to consider this vaccination a "standard of care" is not justified yet. Moreover, there is the risk of precluding further independent RCTs for "ethical considerations", so as "to not deny the benefits of influenza vaccinations to pregnant women of a control group". Instead, before promoting national campaigns for universal vaccination in pregnancy, further large, independent, and reassuring RCTs are needed, even braving challenging a current paradigm. Until then, influenza vaccination should be offered to pregnant women only once open information is available about the safety uncertainties, to allow truly informed choices, and promoting also other protective behaviors.
Keywords: healthy-vaccinee bias; influenza vaccination; informed consent; offspring deaths; pregnant women; real-world vs. randomized evidence; serious adverse events increase.