Towards elimination: measles susceptibility in Australia and 17 European countries

Abstract

Objective

To evaluate age-specific measles susceptibility in Australia and 17 European countries.

Methods

As part of the European Sero-Epidemiology Network 2 (ESEN2), 18 countries collected large national serum banks between 1996 and 2004. These banks were tested for measles IgG and the results converted to a common unitage to enable valid intercountry comparisons. Historical vaccination and disease incidence data were also collected. Age-stratified population susceptibility levels were compared to WHO European Region targets for measles elimination of < 15% in those aged 2–4 years, < 10% in 5–9-year-olds and < 5% in older age groups.

Findings

Seven countries (Czech Republic, Hungary, Luxembourg, Spain, Slovakia, Slovenia and Sweden) met or came very close to the elimination targets. Four countries (Australia, Israel, Lithuania and Malta) had susceptibility levels above WHO targets in some older age groups indicating possible gaps in protection. Seven countries (Belgium, Bulgaria, Cyprus, England and Wales, Ireland, Latvia and Romania) were deemed to be at risk of epidemics as a result of high susceptibility in children and also, in some cases, adults.

Conclusion

Although all countries now implement a two-dose measles vaccination schedule, if the WHO European Region target of measles elimination by 2010 is to be achieved higher routine coverage as well as vaccination campaigns in some older age cohorts are needed in some countries. Without these improvements, continued measles transmission and outbreaks are expected in Europe.

Introduction

Live attenuated measles vaccines have been available since the early 1960s and are now in use worldwide. They have the potential to achieve highly effective measles control and elimination, as observed in the Americas.¹

In 1998, the WHO European Region agreed to eliminate measles in Europe by 2007.² By 2002, the incidence of measles in Europe was estimated to be below 5 per 100 000 and a strategic plan was developed which outlined an approach for achieving elimination by the revised year of 2010.^3–5 The approach focused on each member state delivering two doses of measles vaccine through the routine programme at very high (> 95%) coverage, undertaking catch-up campaigns to address older susceptible cohorts, strengthening surveillance through case-based reporting and laboratory confirmation of suspect cases, and improving communication about the benefits and risks of vaccination.

To measure progress towards elimination and to identify populations for vaccination campaigns, age-group specific susceptibility targets were established that corresponded to an effective reproduction number less than one, and hence elimination.^6,7 These age-specific susceptibility levels could be estimated from high-quality historical vaccine coverage data (but only in populations with no measles transmission) or from population serological surveillance data.⁸ Progress towards elimination can also be assessed from age-specific incidence data, but this is less useful when close to elimination because it is possible for susceptible age cohorts to go unnoticed for many years. Outbreaks in older susceptible cohorts have occurred in Europe in recent years and are serious because of the greater morbidity caused by the disease in older individuals.^9,10 The size of outbreaks generated by imported measles cases can also be used to determine the effective reproductive number if cases are confirmed and extensive investigation to identify all cases in a cluster is performed.⁷

In many countries, high-quality historical vaccine coverage and disease incidence data are not available so serological surveillance is an essential part of assessing population immunity. Even in countries with good vaccine coverage and disease incidence, data serological surveillance can help identify older susceptible cohorts and also problems with vaccine effectiveness. Although serological surveillance has clear potential, in the past it has been difficult to compare countries because they have used different methods for testing serum antibody levels.

To obtain standardized serological data, countries participated in the measles work-package of the European Sero-Epidemiology Network 2 (ESEN2).¹¹ The ESEN2 project was a continuation of the original ESEN project with the same purpose of coordinating and harmonizing serological surveillance in Europe.^8,12,13 The measles component of the original project included seven countries, and identified four with a low risk of outbreaks (England and Wales, Finland, France and the Netherlands) and three with an intermediate/high risk of measles outbreaks (Denmark, Germany and Italy). Germany and Italy have since experienced outbreaks, highlighting the importance of seroepidemiological surveys and the need for targeted action based on the results.^14,15

In this paper, the results from measles serological surveillance in participating countries, as well as data on measles vaccine coverage and disease incidence, are presented and compared to the WHO European Region elimination targets. The results are used to identify susceptible cohorts to help inform future vaccination strategies as well as to identify discordance with routine coverage estimates suggesting possible problems with vaccine effectiveness or coverage data.

Methods

Serum bank collection

Each participating country was required to test a serum bank representative of the general population in their country using their usual measles assay for measuring antimeasles IgG antibody. ESEN project guidelines recommended that approximately 100 sera be tested in each 1-year age band of those < 20 years of age and 200 in each 5-year age band in those aged ≥ 20 years. Although it was preferable that countries collected and tested the bank during the ESEN2 study (2001–2003), some countries had already collected and tested such banks between 1996 and 2000. Each country obtained ethical approval from the appropriate national authorities for the serum collections.

Vaccine programme, coverage and measles incidence

A questionnaire was distributed to each country and completed in 2001/2002 to obtain information on current and historical measles vaccine programme organization, vaccine coverage estimates by age since 1970 and measles incidence by age group (clinical notifications and laboratory confirmations) since 1970. This information was subsequently updated in 2006 using data from the WHO centralized information system for infectious diseases (CISID).¹⁶

Standardization and reference assay

A standardization panel of 151 sera was prepared by the measles, mumps and rubella (MMR) vaccine reference centre (Robert Koch Institute, Berlin, Germany) and tested by each participating country at the same time as the serum bank. Each country was required to use the same measles assay for testing the panel as used for testing the national serum bank. Standardization equations were obtained by regression of local results against the reference centre and were used to convert the titres of the national serum bank to the unitage of the reference centre (ESEN2 units). Further details of the standardization methodology, including the back-standardization method used for countries that had already tested their national serum bank, are given by Kafatos.¹⁷ Details of the measles assays used by the participating countries and the selection of the standardization equations are given by Tischer.¹⁸ The assay of the reference laboratory to which results were standardized was the enzyme immunoassay (EIA, Enzygnost, Dade Behring). The equivocal range for this assay was 0.15–0.35 IU/ml. After standardization, the results were classified as negative, equivocal or positive using these cut-offs. Comparisons of panel results obtained by this EIA and by the gold standard plaque neutralization test suggested that the equivocal titres could be regarded as positive.¹⁸

Data analysis

The proportion seropositive or equivocal (antimeasles antibody concentration > 0.15 IU/ml) was calculated in each age group along with 95% exact confidence interval and plotted to form a seroprofile. Reported first-dose vaccine coverage at 24 months for each age group was added along with an indication of the number of doses and type of measles vaccine scheduled for each age group. Second-dose vaccine coverage was not generally available and is therefore not shown.

The proportion seronegative in each country by age was compared to the WHO elimination targets of < 15% in those aged 2–4 years, < 10% in those aged 5–9 years and < 5% in those aged 10–19, 20–39 and 40+ years. Countries were grouped into those with low susceptibility to outbreaks (WHO elimination targets met for 2–4 year-old and 5–9 year-old age groups and at least two of the three older age groups), intermediate susceptibility (targets missed in two of the following age groups: 10–19, 20–39 or 40+) and higher susceptibility (targets missed in either 2–4 or 5–9 year-old age groups). More emphasis is placed on the younger age groups because of increased spread of disease in the young and also because this is likely to reflect recent problems in vaccine coverage.

Results

Serum bank collection

Australia and seventeen countries in the WHO European Region undertook testing for measles IgG antibodies of serum banks collected between 1996 and 2004 (Table 1). Serum banks were obtained either through residual sera collected during routine laboratory testing (11 of 18 countries), by population-based random sampling (6 of 18), or a combination of these two methods (1 of 18). Sera were collected from all age groups, were evenly distributed between males and females, and were geographically representative of each country. Although using residual sera raises the possibility of bias it is unlikely that for measles immunity this would be large. Not all countries met the sample size targets with numbers too small for evaluation (n < 75) in under 2 year-olds in three countries (Bulgaria, the Czech Republic and Malta). In Spain, there was no sampling in under 2 year-olds, and in Luxembourg there was no sampling in under 4 year-olds and only 37 samples from 4 year-olds.

Table 1. Year and number of samples collected in national serum banks of participating countries.

Country or area	Type of sample	Year of collection	Age range collected (years)	No. samples aged < 20 years	No. samples aged ≥ 20 years
Australia	Residual	2002	1–34	2496	1278
Belgium	Residual	2002/2003	1–60+	1953	1421
Bulgaria	Residual	2001–2004	1–60+	969	697
Cyprus	Residual/Population	2003	1–50	1901	1000
Czech Republic	Population	2001	1–60+	1695	1318
England and Wales	Residual	2000	1–60+	1814	1756
Hungary	Residual	2003	1–60+	2014	1476
Ireland	Residual	2003	1–60+	1214	1376
Israel	Residual	1998	1–60+	1866	1484
Latvia	Population	2003	1–60+	1594	1432
Lithuania	Residual	2003	1–60+	1872	1480
Luxembourg	Population	2000/2001	4–60+	1381	1298
Malta	Residual	2003	1–60+	820	1047
Romania	Residual	2002	1–60+	2304	1535
Slovakia	Population	2002	1–60+	2080	1560
Slovenia	Residual	1999/2000	1–60+	2000	1399
Spain	Population	1996	2–39	1926	1679
Sweden	Population	1996/1997	a	994	398

^a Sera collected from 2, 5, 8, 10, 14, 17, 20–34 and 65+ age groups.

Vaccine programme, coverage and measles incidence

Routine measles vaccination has been in place for two or more decades in all the participating countries. All countries have now adopted a two-dose MMR vaccine schedule (Table 2). The first country to introduce a two-dose schedule was Slovenia in 1974 and the last was Spain in 1996. Most countries moved from one measles dose to a two-dose MMR vaccine schedule. In England and Wales and Romania large catch-up campaigns in older children were also performed when the MMR vaccine was introduced.^19,20

Table 2. Measles vaccination policies, reported incidence in 2001and year of previous major outbreak.

Country or area	Year of introduction			Year two-dose MMR vaccine introduced		Age of vaccination in 2001			Mean first dose coverage 1997–2001 (%)	2001 reported measles incidence (per 100 000)	Year incidence last exceeded 20 per 100 000
	First routine dose	Second routine dose				First dose (months)	Second dose (years)
Australia	1975	1993		1993		12	4		91	0.7a	1994
Belgium	1985	1994		1994		15	11/12		80	N/A	1997
Bulgaria	1969	1983		2001		13	12		93	0.1a	1992
Cyprus	1974	1989		1989		13–15	4–6		88	0b	1983
Czech Republic	1969	1975		1987		15	2		96	0.1b	1990
England and Wales	1968	1995		1995		12–15	4		88	0.1a	1994
Hungary	1974	1990		1991		15	11		100	0.2a	1989
Ireland	1985	1992		1992		15	4–5		77	6.3a	2000
Israel	1967	1994		1994		12	6		94	0.3a	1994
Latvia	1968	1987		2002		15	7		97	<0.1b	1987
Lithuania	1966	1992		1998		15	12 6–7c		97	0.2a	1987
Luxembourg	1986d	1994		1994		15–18	5–7		91	1.6a	1987
Malta	1983	1992		1995		15	7		90	0.5a	1986
Romania	1979	1994		2005		12–15	7		98	<0.1a	1998
Slovakia	1969	1977		1992		14	11		99	0a	1984
Slovenia	1968	1978		1990		12–18	5–7		92	0b	1995
Spain	1978	1996		1996		12–15	3–6		94	0.1b	1995
Sweden	1971	1982		1982		18	12		94	<0.1a	1983

MMR, measles, mumps and rubella; N/A, not available.
^a Notifications.
^b Laboratory confirmations.
^c Age in 2002.
^d First year of routine MMR vaccine, measles vaccine used previously.

Measles vaccine coverage data were usually obtained by routine administrative assessment (which may be subject to bias) or special surveys (which may be imprecise). Coverage data were only consistently available for the first dose as assessed at 24 months. The five-year mean reported first-dose coverage level (1997–2001) varied from 77% in Ireland to nearly 100% in Hungary (Table 2).

By 2001, reported measles incidence had declined to very low levels with clinical and laboratory notifications (< 2/100 000) in all countries except for Ireland where the incidence of notifications was 6/100 000 (Table 2). The last year before 2001 with significant measles incidence (> 20/100 000) varied from 1983 in Cyprus and Sweden, to 1997 in Belgium and 2000 in Ireland (Table 2).

Seroprofiles and comparisons to WHO age-specific targets for measles susceptibility

Measles seroprofiles, first-dose measles vaccine coverage and the vaccine(s) recommended for each age cohort for the 18 countries are available from the ESEN2 pages of the Health Protection Agency website: http://www.hpa.org.uk/esen2. These seroprofiles can be used to help identify susceptible age cohorts in each country that may have arisen through low vaccine coverage, poor effectiveness of the vaccine or a reduction in measles transmission in unvaccinated cohorts. The seroprofiles show that the proportion seropositive or equivocal increases from age 1–3 years in most countries, reflecting vaccine administration. In the older prevaccination cohorts (typically aged over 35 years), high seroprevalence is generated through natural exposure, whereas in the younger cohorts seroprevalence should reflect mostly vaccination. In the age cohorts not scheduled for more than one dose, reported vaccine coverage agrees approximately with measles seroprevalence in all countries, with the exceptions of Bulgaria, Latvia and Romania where reported official age-specific coverage is much higher than the proportion seropositive.

Table 3 shows the percentage seronegative in five age groups for each country and the classification of countries into low, medium and high risk of measles outbreaks. Three countries (the Czech Republic, Luxembourg and Spain) met the WHO targets for elimination in all age groups and a further four (Hungary, Slovakia, Slovenia and Sweden) only missed the target in the 20–39 year-old age group by < 3.5%. In these countries the risk of measles outbreaks is low. Four countries (Australia, Israel, Lithuania and Malta) had met the recommended WHO susceptibility targets in the under 10 year-old age groups but not for older children/adults and were classified as having intermediate susceptibility. The remaining seven countries (Belgium, Bulgaria, Cyprus, England and Wales, Ireland, Latvia and Romania) had not met the WHO targets for susceptibility in the 2–4 or 5–9 year-old age groups or, with the exception of Romania, in some older age groups either and were therefore classified as having high susceptibility.

Table 3. Percentage measles seronegative (antibody titre < 0.15 IU/ml) by age group compared to WHO targets.

Country or area	Measles seronegative (%)
	2–4 years	5–9 years	10–19 years	20–39 years	40+ years
WHO target	< 15	< 10	< 5	< 5	< 5
Low susceptibility
Czech Republic	1.0	0.8	1.5	3.2	0.2
Hungary	2.9	3.8	3.5	8.5	0.3
Luxembourg	5.4a	4.7	5.0	2.6	0.2
Slovakia	3.8	4.8	3.3	6.1	0.3
Slovenia	4.0	3.2	4.2	6.1	1.5
Spain	5.0	7.2	4.0	0.8	–
Sweden	1.0	5.8	4.7	5.9	0.5
Intermediate susceptibility
Australia	10.8	8.0	7.9	9.4	–
Israel	9.1	6.9	5.2	7.0	1.4
Lithuania	4.7	9.8	9.6	12.4	0.3
Malta	9.6	4.3	6.9	5.3	3.1
Higher susceptibility
Belgium	12.4	14.1	13.3	4.6	2.0
Bulgaria	30.4	25.9	20.7	10.1	9.0
Cyprus	21.8	21.8	13.2	5.9	1.0
England and Wales	18.9	10.2	6.9	2.8	0.2
Ireland	14.2	11.8	8.6	7.8	7.6
Latvia	19.0	42.9	39.8	30.8	3.5
Romania	24.3	11.4	4.2	1.4	0.3

^a Based on only 37 sera.

Discussion

This paper provides an overview of measles seroepidemiology in Australia and a large number of countries throughout Europe for the period 1996–2004. The results of the ESEN2 project and those of the earlier ESEN project, provide invaluable information about progress towards the WHO measles elimination target for 2010 in the WHO European Region.⁴

The results of the serological surveys illustrate the heterogeneity of measles control in the region. This reflects the wide range in current and historical measles vaccine policy and vaccine coverage. Three groups of countries can be distinguished. Seven countries (the Czech Republic, Hungary, Luxembourg, Slovakia, Slovenia, Spain and Sweden) have age-specific susceptibility levels congruent with having achieved or approaching measles elimination. All these countries have had two-dose measles vaccine programmes from the 1970s onwards with very high reported first-dose measles vaccine coverage (mostly > 95%) for at least the previous five years. The notable exception is Spain, which only implemented a second dose in 1996 at the time the serosurvey was undertaken. Reflecting this high population immunity, all these countries except Spain had very low reported measles incidence with an average incidence from notifications of less than 4/100 000/year and incidence from confirmed cases (where available) of < 0.5/100 000/year in the decade before the serosurveys. In Spain, the population immunity comes from both vaccination and disease. To ensure measles elimination is achieved in 2010, it will be critical that these countries maintain very high routine coverage (> 95%) with two doses of measles vaccine.

In four countries (Australia, Israel, Lithuania and Malta), the proportion susceptible had reached the WHO susceptibility targets in children under 10 years of age, but was above the target for adolescents and young adults. These countries have employed routine two-dose measles programmes since the early 1990s, with high reported routine vaccine coverage for the previous five years (> 90%). The susceptible cohorts in young persons in these countries either reflect those who were scheduled only for a single dose or for two doses, but presumably delivered at a lower coverage than the target of > 95%. The levels of coverage were high enough to interrupt measles circulation leading to the accumulation of a pool of susceptible older persons. These countries need to ensure routine two-dose coverage is at least 95% and also consider implementing catch-up campaigns. Such strategies have been successfully undertaken in other countries, such as the Republic of Korea.²¹

It is of concern that in seven (Belgium, Bulgaria, Cyprus, England and Wales, Ireland, Latvia and Romania) of the 18 countries, there is high susceptibility in several age groups, including young children. In Belgium, Cyprus, Ireland and England and Wales, first-dose measles vaccine coverage in the five years up to 2001 was below 90%. This has led to the proportion of children susceptible to measles exceeding the WHO susceptibility targets. This lower coverage has only been a recent phenomenon in England and Wales, and to some extent in Ireland (which also has high susceptibility in adults), reflecting the impact of parental concern on the safety of MMR vaccine on uptake.²² The decreased coverage in England and Wales can also be seen in the reduced seropositivity in children compared to the ESEN 1996 survey.⁸ Since these ESEN2 surveys, outbreaks in Ireland and England and Wales have been reported in young children.²³ In Belgium and Cyprus, in addition to high susceptibility in young children, a significant proportion of adolescents remain susceptible to measles reflecting lower historical measles vaccine uptake.

In Bulgaria and Latvia, there is evidence of susceptible age groups in both young children and young adults, and only in young children in Romania. This disagrees with the officially reported measles coverage data, which in each case is > 90% in the last five years. In Latvia, MMR vaccine coverage estimates agree with the observed rubella seroprofiles,²⁴ suggesting that the reported vaccine coverage is accurate and that there may either be a problem with the vaccine, the samples or the standardization process. If there was a problem with the standardization then this would need to be age specific because the oldest age cohorts showed high levels of antibody; also, samples were tested immediately upon collection and the testing of the standardization panel was satisfactory. The affected cohorts received the Leningrad strain of measles vaccine during the 1980s and 1990s. If the problem is the vaccine and true seroprevalence is as low as reported then it is surprising that large outbreaks have not already occurred. However, large outbreaks in older populations that had been highly vaccinated with the Leningrad strain have been reported from neighbouring countries, such as Ukraine.²⁵ The discrepancy requires further study.

In Bulgaria, the discrepancy between vaccine coverage and seroprevalence is smaller and more consistent across age groups, including vaccinated adults. This could represent a possible problem with sample storage or with the assay, but this seems unlikely because seroprevalence is high in the oldest age groups and the standardization panel results were good. A large outbreak in 1992 in children and young adults suggests that routine vaccine coverage could be overestimated²⁶ and that significant pools of susceptibles may exist, but recent incidence has been very low with no cases reported in 2002–2004 and no indigenous spread following an imported case from China in 2005.²⁷

In Romania, the low susceptibility in the adolescent and adult age groups presumably reflects the impact of the 1998 catch-up campaign that targeted 7–18 year-olds and after which the reported incidence of measles was very low. However, levels of susceptibility are particularly high in the younger (< 8 years) age groups, and the discrepancy with reported routine vaccine coverage suggests that it is lower than reported. Indeed Romania recently reported a large, national outbreak, which particularly affected younger age groups from marginalized populations.^16,28

To achieve elimination, all these higher susceptibility countries will need to strengthen their routine measles programmes to achieve > 95% with both doses and address older susceptible age groups through catch-up campaigns. These activities will need to be supported by information campaigns highlighting the importance and safety of MMR vaccine.

In this paper, we have employed a fixed cut-off of 0.15 IU/ml. There is only limited observational data to support this arbitrary cut-off.¹⁸ In several countries with long-standing high coverage two-dose measles programmes (Hungary, Slovakia, Slovenia and Sweden), the proportion classified as “seronegative” is marginally > 5% in young adults (20–39 year-olds). These “seronegatives” may reflect waning antibody over time in highly vaccinated (and protected) cohorts, who have not had an opportunity to be exposed to natural measles infection.²⁹ This has been observed previously and there is no evidence to date to suggest secondary vaccine failure occurs in these cohorts despite follow-up for up to three decades.^30,31 Further analysis of the quantitative antibody titres using mixture modelling may be used to examine waning immunity and also estimate the proportion of the population falling into groups such as vaccinated, past infection, recent infection and negative.³²

In conclusion, the ESEN2 project has demonstrated which countries have met or are approaching measles elimination and what other countries need to do to reach this target. It is critical that all countries in Europe achieve and maintain very high vaccine coverage with two doses of measles vaccine and address older susceptible groups, if the target of measles elimination by 2010 in the WHO European Region is to be reached. ■