Introduction: The association of measles, mumps and rubella (MMR) vaccination with immune thrombocytopenia (ITP) occurrence has been shown. The risk of ITP with other vaccines is still not known. This study was aimed at assessing the association of recommended vaccinations in children with ITP occurrence.

Methods: We conducted a population-based study in France including all children newly diagnosed for primary ITP between July 2009 and June 2015. This cohort was built using a validated algorithm in the nationwide French health insurance database (SNDS). We assessed the risk of ITP with MMR vaccine, all combined vaccines containing diphtheria, tetanus and poliomyelitis (DTP) vaccines, pneumococcal and meningococcal C vaccines. We used two self-controlled designs: a case cross-over and a self-controlled case series. For the case cross-over, we compared the frequency of exposure to vaccines during a 6-week period immediately preceding the event (case period) with the frequency of exposure during a previous time period (control period, having the same duration as the case period). We performed sensitivity analyses using 8- and 12-week periods. Analyses were adjusted for exposure to other drugs known as inducers of ITP and seasonality. Odds ratios (OR) and their 95% confidence intervals (CI) were calculated. For the self-controlled case series, we compared the ITP incidence within periods of risk (following vaccination, named exposure period) with the incidence within the control period of non-exposure. The exposure period was defined by the 6 weeks after the vaccine dispensing in the principal analysis (8 and 12 weeks in sensitivity analyses). We further excluded the 2 weeks prior to vaccine dispensing from the non-exposure period to address selective survival bias (healthy vaccinee effect). The observation period was censored at ITP occurrence, due to variation of vaccination probability after ITP diagnosis and to the impossibility to distinguish ITP relapses from chronic ITP in the database. Analyses were adjusted for seasonality. Incidence rate ratios (IRRs) and their 95% CI were calculated. We assessed the exposure to each vaccine, and conducted subgroup analyses in patients without any concurrent vaccination during case and control periods for the case cross-over study and exposure periods for the self-controlled case series study. We also calculated the number of ITP cases occurring during the 6 weeks after vaccination divided by the number of vaccine doses dispensed in the French children population during the study period.

Results: We included 2,549 newly diagnosed primary ITP children. Among them, median age was 5.1 years and 46.5 % were females; 41.4% had been exposed to at least one studied vaccine before ITP onset. The results of the principal analysis are detailed in the Table. There was an increased occurrence of ITP following MMR vaccination (OR: 1.60, 95% CI: 1.09-2.34; IRR: 1.30, 95% CI: 0.95-1.80). Analyses excluding the patients with concurrent vaccination, notably meningococcal vaccination, led to similar results (OR: 1.66, 95% CI: 1.02-2.71; IRR: 1.39, 95% CI: 0.80-2.42). There was also an increased occurrence of ITP with the meningococcal C vaccine (OR: 1.92, 95% CI: 0.95-3.86; IRR: 1.40, 95% CI: 0.86-2.29). Analyses conducted in patients without any concurrent vaccination, notably MMR vaccination, confirmed these results with wide 95% CI because of fewer patients included (OR: 1.64, 95% CI: 0.57-4.71; IRR: 1.64, 95% CI: 0.69-3.86). No association was observed between other vaccines and ITP occurrence. The numbers of ITP cases occurring in the 6 weeks following vaccination per million doses dispensed were 8.2 for pneumococcal, 9.2 for DTP, 9.6 for meningococcal and 11.5 for MMR vaccines. Of note, these numbers overestimate the probability of vaccine-induced ITP. Indeed, they are ITP cases chronologically compatible with vaccine adverse reaction without any individual causality assessment (a worst-case scenario considering that all cases were triggered by vaccines).

Conclusion: This study showed an increased occurrence of ITP after MMR and meningococcal C vaccines. It is reassuring for other vaccines. We cannot exclude temporal association with MMR and meningococcal C vaccines due to the peak of ITP incidence at 12 months of age in the general population. However, vaccine-induced ITP is a very rare event, which does not cast doubt on the interest of vaccination.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.

Sign in via your Institution