An increase in idiopathic thrombocytopenic purpura (ITP) cases in the 6 weeks following the first dose of the measles, mumps and rubella (MMR) vaccine has been established, with absolute risks estimated as 1 in 22 3001 and 1 in 21 000 vaccine doses,2 with two in every three cases attributable to the vaccine. However, the risk after a second dose of MMR vaccine has not been investigated.

Hospital admissions for children aged from 3 to <6 years with a discharge diagnosis of ITP (ICD-code D693 in any diagnosis field) were identified from computerised hospital episode data from North, East and South London, Essex, East Anglia, Sussex and Kent for the period from 1 April 1997 to 31 December 2005. These admissions were then linked to second MMR dose records held on population-based child-health database systems. Only successfully linked admissions were used for the analysis.

A re-admission for ITP within 10 days was classed as a continuation of the previous episode Validation of the coding was not undertaken because comprehensive validation in an earlier study using the same hospital discharge data set confirmed the diagnosis coding to be accurate.1 Analysis was carried out using the self-controlled cases series method, which only uses cases and automatically controls for fixed individual level confounding.3 The method enables us to estimate the relative incidence, which is the ratio of the rate of events in the 6 weeks after vaccination to the rate of events in the absence of this exposure, with adjustment for age.

A total of 106 ITP admissions in 78 individuals who had received a second dose of MMR vaccine were identified; 14 admissions were excluded because they occurred closer than 10 days to an earlier admission, leaving a total of 92 admissions satisfying the study criteria. Of these 92 admissions, four took place within the period of interest (fig 1). The relative incidence of an ITP episode during the 6-week post-vaccination period compared to the control period was estimated as 1.04 (95% CI 0.37 to 2.92). To identify the size of the risk excluded by this study the upper end of the 95% CI for the attributable risk was calculated as follows: the expected number of cases in the risk period is 4/1.04  =  3.85; the expected number attributable to MMR if the relative incidence is 2.92 is 3.85*(2.92−1)  =  7.4; the attributable fraction of all cases is 7.4/92  =  8%; the annual incidence of ITP in 3–6 year olds in England is 14 per 100 000 (Hospital Episode data for the period April 1997 to March 2005); the upper 95% CI for the risk is therefore 0.08*14  =  1 per 100,000.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Episodes of ITP in children aged 3 to 6 years: interval from second mmR vaccination.

This study found no evidence of an increased risk of ITP within 6 weeks of the second dose of MMR and excludes risks of more than about 1 per 100 000. There have been isolated reports of ITP after both a first dose and second dose of a measles-containing vaccine 4 5 but we were unable to identify a cohort of children with ITP within 6 weeks of the first dose who then went on to receive the second dose. Indeed there may be few such children because it is recommended6 that children who develop ITP within 6 weeks of the first dose are tested for antibodies and only receive the second dose if susceptible to one or more of the viruses. Of the 33 children who were admitted for ITP after the age of 3 years and before their second dose of MMR, none had a recurrence within 6 weeks of the second MMR.

In conclusion, our study provides reassurance that among children receiving a second dose of MMR vaccine there is no evidence of an increased risk of ITP.