Measles elimination in the WHO Region for Europe was already planned for 2010 but has not yet been achieved [1
]. The number of reported measles cases in the region increased 15 times between 2016 (the year with the lowest number of cases recorded) and 2018 [3
]. Several issues have contributed to this phenomenon, one of them being vaccine hesitancy- a factor identified as one of the 10 key health threats in 2019 [4
]. WHO recommends to diagnose the determinants of insufficient vaccine uptake and to tailor vaccination strategies [5
]. Susceptibility levels below 5% in children 10 years and older, below 10% in children age five to nine years, and below 15% in children 24 months to 4 years are needed to eliminate measles [6
In Austria, mandatory reporting of suspected measles cases has been implemented in 2006 [7
]. The first measles containing vaccines (MCV) were licensed in 1963 in the US [8
] and, since 2001, more than 2 billion doses were administered globally [9
]. Bivalent (with mumps) MCV was recommended in 1974 in Austria, and the use of two doses of trivalent (measles–mumps–rubella) vaccine (MMR) was recommended in 1994 [10
]. Since 1985, MCV was administered free of charge for children up to 15 years of age and since 2015 for all age groups [11
Vaccine coverage is estimated by the administrative method in Austria. Since 2016, an agent-based simulation model has been implemented to better assess the number of susceptible persons to measles [12
In Tyrol, one of Austria’s nine regions with 739,139 inhabitants in 2016 [13
], the estimated coverage of the birth cohorts 2000 to 2009 was 88.2% with one dose and 76.5% with two doses of measles containing vaccine in 2012 [14
]. Since 2000, an electronic vaccination registry has been fully implemented in Tyrol.
As of 2014, the Tyrolean government has required proof of two doses of MCV from all health care workers (HCW) in Tyrolean hospitals. Those who lack documented evidence of two doses of MMR are required to undergo serological testing. The same applied to all medical students.
The WHO provides guidance to seroprevalence surveys to better identify pockets of susceptible populations, and recommends their use if surveillance was only recently implemented and when coverage data are less reliable for older age groups [15
]. Nevertheless, serosurveys as recommended by WHO are costly, logistically difficult to conduct, and require substantial time commitment as well as ethical approval. The objective of our work was to assess the usefulness of retrospective seroprevalence data and to describe the proportion of measles susceptible persons by birth cohort and district to better understand sub-populations that are most at risk. This knowledge may help to tailor awareness raising activities as mentioned in the national measles elimination plan [16
We assessed the susceptibility levels for measles among the Tyrolean population using three sources of information: a retrospective analysis of already available serological data for measles IgG (serosurvey), vaccine coverage data from the electronic vaccination registry and surveillance data from the mandatory electronic reporting system.
We found that the WHO recommended susceptibility thresholds of maximum 15% in children 24 months–4 years, 10% in persons 5–9 years, and 5% in persons >10 years were not achieved. Susceptibility was age dependent and the threshold of less than 5% was only reached by birth cohorts 1976 (41 years of age as of 31 December 2016), 1978, and 1979 or older. Susceptibility levels for persons 10–41 years (birth cohorts 2007–1976) were twice as high as the recommended values of WHO of 5% for these age groups.
More than one-third of the population was considered to be immune with susceptibility levels below the requested age-dependent thresholds. The highest susceptibility levels were identified among the most recent birth cohorts until birth cohort 2011, with susceptibility levels differing between 15–50% from the WHO target. Ten birth cohorts revealed susceptibility levels between 5% and 9% and further 22 birth cohorts levels between 1% and 4% above the expected WHO targets. Many member states of the European Union did not achieve the requested WHO goals and susceptible persons accumulated over time [18
]. Even a slight increase in MCV uptake was associated with significant reduction of overall burden of disease [19
Significantly more males were susceptible to measles compared to females. This uneven gender distribution was also confirmed by surveillance data. Such inequities are not yet addressed through vaccination policies [20
] and studies suggest there are no major sex differences in IgG production following vaccination [21
We identified several limitations of our seroprevalence study: due to the very long observation period, the likelihood of losing detectable antibody levels—especially of persons who were vaccinated with only one dose of MCV—has to be considered and might have biased our results. This would probably not apply for persons who had either experienced natural infection, received two doses of MCV or were re-exposed. This may have been of less influence, as participants from nearly all birth cohorts were investigated recently. In fact, the long time span of the serosurvey resulted in a large sample size and thus allowed more precise estimates. Estimates also became more reliable due to careful exclusion of ambiguous data and they were based on well validated test kits in an accredited laboratory. Despite the fact the only few seroprevalence data of recent birth cohorts were available, the estimation of susceptibles by the two different approaches, retrospective seroprevalence data, and estimation of susceptibles using coverage data corresponded quite well. As expected, coverage with one dose of MCV was constantly higher compared to coverage with two doses and thus susceptibility was inversely correlated. The long observation period may have also resulted in an underestimation of susceptibility in more recent birth cohorts due to very small numbers. Using coverage data from the electronic registry has resulted in a much better estimate of susceptibles of more recent birth cohorts. For the assessment of susceptibility with one or two doses of MCV we used the reported vaccine effectiveness for each dose as reported by the Centre for Diseases Control (CDC), Atlanta [22
]. The assumption of 93% vaccine effectiveness for the first dose and 97% of the second dose was very conservative as higher vaccine effectiveness was reported by other authors [23
Another limitation of our seroprevalence data was that despite the entire region being covered, districts that are physically closer located to the laboratory were overrepresented and districts further away sent fewer samples. This is important as Tyrol is known to be a region with a high prevalence of vaccine hesitant persons [25
]. This may have resulted in a greater inaccuracy from districts further away from the capital. Nevertheless, the highest incidence districts calculated from surveillance data were in line with the highest susceptibility levels by district from the seroprevalence data. As vaccine coverage data were only available in aggregated format, we could not compare incidence with coverage by district.
The differences of susceptibility by birth cohort with seroprevalence and estimated adjusted susceptibility data from the electronic vaccination registry corresponded well in their pattern, and the mean varied between −5% for the second dose and 5% with the first dose. Susceptibility levels of seroprevalence data were higher in recent birth cohorts until 2009 compared to susceptibility calculated using coverage data. The latter was probably due to the small sample sizes of recent birth cohorts in the seroprevalence survey.
The estimates of susceptible persons from the agent-based simulation model at a national level identified that approximately 70% of 19–30 year-old persons were protected with two doses of MCV as of 2018. In our data the increase of susceptibility using the second dose adjusted susceptibility started already from the birth cohort 2005 until birth cohort 1998 [12
]. This may be due to regional differences.
Limitations of our coverage data were the different sources of numerator and denominator data. Participants who moved out of the region or died were still included in the numerator but were excluded from the denominator, in which only resident persons were reported. This might have led to an overestimation of vaccine coverage.
Limitations of our surveillance data were that detailed information was only available since 2009. As older age groups were more affected in recent years in many countries, this may have resulted in underestimation of attack rates of younger individuals. If a virus is introduced to a susceptible population, cluster cases are known to occur in similar age groups. Therefore, the short observation period of our surveillance data may have reflected this phenomenon. Nevertheless, birth cohorts with more than one case reported were significantly positively associated with susceptibility levels > 10%.
Targeted catch-up campaigns of the identified birth cohorts focusing on high incidence districts would definitely be crucial to close immunization gaps and reduce the number of susceptibles but raising awareness and social mobilization of the target groups were shown to be difficult. In our experience, when mandatory vaccination has not been implemented and no incentives for vaccination were provided, participation in catch-up campaigns remained unsatisfactory [25
We conclude the Tyrolean population is not sufficiently protected against measles and additional measures such as targeted supplementary immunization activities (e.g., targeted catch-up campaigns) are necessary to achieve elimination. This was also underlined by the fact that the majority of recently reported measles cases corresponded well with the birth cohorts with highest susceptibility levels. Priorities should be given to districts and birth cohorts with the lowest susceptibility, considering the WHO recommended values.
Our seroprevalence data suggested that additional immunization activities should target birth cohorts between 2011 and 2015 as a first priority and birth cohorts between 1998 and 2010 as a second priority.
Furthermore, our results suggest that birth cohorts before 1976 are not target groups for supplementary immunization activities as the majority of them are probably immune due to natural infection. In 2019, we conducted several activities mainly during the European Immunization Week to increase awareness and enhance uptake targeting the identified birth cohorts and districts, such as social mobilization and information campaigns in schools and universities, including health care workers.
We conclude that using different sources of information such as vaccine coverage data, surveillance data, and retrospective seroprevalence data is useful to identify pockets of susceptibles and retrospective seroprevalence data could be an additional cheap and useful source of information to better identify sub-populations, such as birth cohorts and geographical areas with increased measles susceptibility. This will especially apply for older birth cohorts, whose coverage data are lacking.