Bulletin of the World Health Organization

Elimination of endemic measles transmission in Australia

Anita E Heywood a, Heather F Gidding b, Michaela A Riddell c, Peter B McIntyre a, C Raina MacIntyre a & Heath A Kelly c

a. National Centre for Immunisation Research and Surveillance of Vaccine Preventable Diseases, The Children’s Hospital, Westmead, NSW, Australia.
b. Centre for Infectious Diseases and Microbiology – Public Health, Westmead Hospital, Westmead, NSW, Australia.
c. Victorian Infectious Disease Reference Laboratory, Locked Bag 815, Carlton South, Vic. 3053, Australia.

Correspondence to Heath Kelly (e-mail: heath.kelly@mh.org.au).

(Submitted: 27 July 2007 – Revised version received: 07 May 2008 – Accepted: 08 May 2008 – Published online: 04 November 2008.)

Bulletin of the World Health Organization 2009;87:64-71. doi: 10.2471/BLT.07.046375

Defining elimination

Several WHO regions have set target dates for the elimination of the transmission of endemic measles. The WHO Regional Office for the Western Pacific (WPRO) has nominated the target date of 2012.1 Since measles elimination was first proposed, definitions of elimination have progressed from requiring a reduction to zero in the incidence of infection in a defined geographical area,2 to the absence of endemic measles transmission and the lack of sustained transmission following an importation of measles virus in a large and well populated geographical area, as outlined in guidelines by WPRO.3 The indicators adopted by WPRO to monitor the progress towards measles elimination provide an operational definition of measles elimination.3,4

Papania & Orenstein have argued that elimination can be declared if multiple lines of evidence demonstrate the absence of endemic measles transmission.5 Several countries have declared elimination of endemic measles transmission using criteria that have become more rigorous over time (summarized in Table 1, available at: http://www.who.int/bulletin/volumes/87/1/07-046375/en/index.html), including the criteria we use here to declare elimination in Australia.

As with other countries that have declared elimination of measles, Australia’s national elimination plan included high two-dose immunization coverage and a disease surveillance system capable of a rapid response to potential measles outbreaks.16 Australia, like many other countries that have declared elimination, would have difficulty meeting the WPRO elimination criteria based on currently available reporting of the investigation of presumptive measles cases (Table 2, available at: http://www.who.int/bulletin/volumes/87/1/07-046375/en/index.html). However we believe multiple lines of evidence conclusively demonstrate the elimination of endemic measles transmission from Australia since 2005 at the latest, when notified confirmed cases were < 1 per million population. In this paper we outline how these criteria have been met, compare them with the WPRO criteria and justify their validity. We argue that a broader range of internationally accepted criteria for measles elimination is warranted.

Evidence of elimination

Low incidence

Notifications of measles cases in Australia were sporadic until the establishment of the National Notifiable Diseases Surveillance System (NNDSS) in 1991.17 Since then, public health legislation in all jurisdictions has included the mandatory reporting of measles cases by laboratories, clinicians and hospitals, to state and territory health departments. Notifications of confirmed cases are forwarded to the NNDSS. Since 2004, all Australian states and territories have adopted a case definition for a confirmed case of measles that requires laboratory evidence from an approved reference laboratory or an epidemiological link to a laboratory-confirmed case in conjunction with clinical evidence.18 These improvements mean that all confirmed measles cases notified to the NNDSS since 2004 are likely to represent true cases (WPRO criterion 1, Table 2).

Since a large national outbreak in 1993–4 (Box 1), there has been a progressive downward trend in measles notifications. The 10 confirmed cases of measles in 2005 (0.5 cases per million population) was the lowest annual figure ever reported on the NNDSS (Fig. 1).24 A total of 125 cases were reported in 2006 (6 cases per million).24 However, a large proportion (~54%) was attributable to a nationwide outbreak linked to the tour of a foreign spiritual group. Attendees at tour meetings were disproportionately opposed to vaccination and transmission was predominantly confined to one generation.25,26 In 2007, 11 cases were reported to the NNDSS (0.5 cases per million24). In both 2005 and 2007, Australia met the WPRO target of 1 case per million population but not in 2006, a year in which we believe that endemic measles transmission did not occur in Australia.

Box 1. The evolution of measles elimination in Australia

Phase 1: pre-vaccine

  • Pre-vaccine epidemiology: similar to that reported from other comparable regions of the world – major epidemics occurring every 2 years and number of annual cases approximating the birth cohort.
  • 1966–1975: 146 deaths from measles were certified19
  • 1962–1971: 1 in every 22 hospitalized cases of measles complicated by encephalitis20
  • 1968: measles control efforts commenced with the licensing of a live attenuated monovalent measles vaccine19
  • 1975: measles vaccine incorporated into the national immunization schedule, as a single dose at 12 months of age
  • 1976–1985: recorded deaths from measles halved to 62 in decade following vaccine introduction, despite slow uptake19
  • Measles mortality decreases following the introduction of measles immunization, but there is ongoing transmission of a single endemic genotype (often as multiple strains of the same genotype).21,22

Phase 2: measles control

  • The incidence of measles and strain variation of the endemic genotype decline with expansion of immunization and outbreak control activities,21 but susceptible individuals accumulate and large outbreaks occur after longer inter-epidemic periods.
  • 1993–1994: large epidemic with > 10 000 notified cases
  • 1993: second dose of measles-containing vaccine (measles–mumps–rubella, MMR) for all 10–16 year olds recommended,19 prompted by large epidemic > 10 000 notified cases
  • Although this led to an overall reduction in measles cases, outbreaks continued in school-aged children and young adults.22

Phase 3: elimination

  • Progression into the elimination phase facilitated by the Australian Measles Control Campaign, in which 1.33 million children (96% of the target age group) were vaccinated.23
  • 1998: Australian Measles Control Campaign: re–scheduling of the second dose of MMR to 4 years of age; and catch–up vaccination for primary school children aged 5–12 years
Fig. 1. Measles notification rates per million population, Australia, 1991–200724

Quality surveillance

Local public health authorities are responsible for the active follow-up of all suspected cases of measles (defined as morbilliform rash with fever present at onset of rash and cough)27 to confirm the diagnosis and identify any additional cases. However, information about the investigation of suspected cases found not to be measles is not recorded at the national level, including laboratory performance indicators. Thus, due to reporting mechanisms, Australia is unable to provide data on the WPRO surveillance process criteria 3–7 (Table 2). However, enhanced surveillance from 1998–2003 in the state of Victoria (2005 population estimate = 5 million; approximately 25% of the Australian population)28 showed that 89% of suspected cases could be discarded after laboratory investigation, at a median annual rate of 2.9/100 000.29 This experience is likely to be applicable nationally and meets WPRO criterion 2.

High two-dose vaccine coverage

High vaccination coverage (greater than 95% for each new birth cohort) is required for herd immunity against measles and maintenance of measles elimination (WPRO criterion 8, Table 2).30 Since 1998, Australia’s measles elimination strategy has included vaccination coverage targets set to achieve 95% coverage with the first dose of measles-containing vaccine (MCV1) for children by 24 months and 90% two-dose (MCV2) coverage by school entry.31 In 1989, a child health survey indicated coverage with at least one dose of a measles-containing vaccine was 85%.32 Fig. 2 shows how Australia’s elimination strategy has resulted in increased vaccination coverage for both MCV1 and MCV2, as reported to the Australian Childhood Immunization Register (ACIR).

Fig. 2. Coverage rates of the first and second doses of measles-containing vaccine in Australia by assessment date as reported on the Australian Childhood Immunisation Registera
MCV, measles-containing vaccine
a Assessment date: MCV1 is scheduled at 12 months of age and assessed at 27 months of age; MCV2 is scheduled at 4 years of age and assessed at 6.25 years of age to allow for delayed notification.

The ACIR is estimated to capture more than 99% of Australian children aged less than 7 years.33,34 Coverage is reported for MCV1 at 2 years and 3 months of age and MCV2 at 6 years and 3 months of age to allow for delayed notification. In 2006, the ACIR recorded that 93.7–94.0% of children aged 2 years (born in 2004) had received at least one dose of MCV and 85.0–88.8% of children aged 6 years (born in 2000) had received both doses (Fig. 2). Substantial geographic homogeneity was demonstrated with coverage of one dose ranging from 92.7–96.2% and coverage of two doses ranging from 85.6–90.2% across all states and territories. These are minimum estimates, with parental recall surveys suggesting that the ACIR underestimates coverage for MCV2 by 5–10%33 and for vaccines scheduled at 12, 18 and 24 months of age by 3–5%.34 When corrected for estimated underreporting, the national and WPRO target of 95% coverage for one dose of measles-containing vaccine is exceeded; coverage with a second dose is likely to be > 90%.

Outbreaks

The available surveillance data for Australia confirms that a high proportion of cases are imported or linked to an imported case, transmission from imported cases is quickly interrupted, and outbreaks following importation of the measles virus are self-limiting and contained. Enhanced surveillance in the state of Victoria identified 58 outbreaks between 1998 and 2006, with a total of 262 cases ranging from 1 to 75 cases.35 Among outbreaks in which the source case could be identified as imported, 33 did not result in transmission, while 22 were associated with secondary cases. Only 3 cases could not be directly linked to importation.35 Satisfying WPRO criterion 9, 91% of measles outbreaks or transmission foci in Victoria between 1998 and 2006 involved < 10 cases. Although such detailed data are not available nationally, the patterns are likely to be similar to those in Victoria. This is based on several lines of evidence. First, results for serosurvey and vaccination coverage are similar between regions. Second, a smaller period of enhanced surveillance in Western Australia between March 1999 and October 2000 identified 28 cases of measles all resulting from nine importations, all of < 10 cases.36

Absence of an endemic genotype

Variability in the nucleotide sequence of the measles virus, of which there are eight clades (A–H) and 23 currently assigned genotypes (A, B1–B3, C1, C2, D1–D10, E, F, G1–G3, H1, H2), can be exploited for molecular epidemiological purposes.37,38 Molecular analysis in routine case and outbreak investigations during the elimination phase of measles control is critical to document the genotype of each new cluster and demonstrate the absence of sustained transmission of one genotype (WPRO criterion 10, Table 2), to identify the source of the measles virus in outbreak situations and to confirm vaccine-associated fever/rash illness (genotype A), which can also assist outbreak investigations.39

Retrospective molecular analyses of measles cases in Victoria from 1973 to 1998 and prospective molecular analyses nationally from 1999 to 2001 were conducted at the Victorian Infectious Diseases Reference Laboratory.40,41 Fig. 3 shows the steady decline in measles-associated hospitalizations in Victoria since the introduction of a vaccine and the succession of measles genotypes identified during this time, consistent with the WHO phases of measles epidemiology.30 The D1 genotype was identified in the earliest samples available for analysis suggesting that genotype D1 may have been the endemic genotype in Australia in the prevaccine era. By 1985 the D7 genotype appeared to replace D1 as the endemic genotype. In the early 1990s, outbreaks of genotypes C2 and H1 were subsequently identified suggesting Australia had moved to the WHO-defined measles control phase of genotypic replacement. Since this time several genotypes have been identified, but none repeatedly, suggesting there have been no endemic genotypes in circulation since this time. Fig. 4 illustrates the measles virus genotypes identified in outbreaks in Australia between 2001 and 2006. Source countries of measles virus importations could be identified for the majority of clusters and are indicated on the map.

Fig. 3. Measles-associated hospitalizations and measles virus genotypes isolated in Victoria (1962–2004) during WHO-defined measles elimination phases20,40,41,a
MCC, measles control campaign; MMR, measles-mumps-rubella vaccine; m.o., months' old; y.o., years' old.
Arrows reflet changes to measles immunization policy. Boxes show genotypes detected retrospectively (pre-1999) from measles virus isolates and clinical samples during the measles control phase. Measles virus isolates prospectively (1999-2004) detected duringn the elimination phase represent genotypes associated with imported cases or vaccine-related illness (now shown in figure). Prospectively isolated genotypes (number of isolates) include A (4 isolates detected), D3 (1), D4 (2), D5 (4), D7 (1), D8 (4), D9 (2), G2 (2), G3 (2), H1 (2) and H2 (1).
Fig. 4. Measles virus genotypes detected in Australia, 2001–200641,a
ACT, Australian Capital Territory; NSW, New South Wales; NT, Northern Territory; Qld., Queensland; SA, South Australia; Tas., Tasmania; Vic, Victoria; WA, Western Australia; USA, United States of America
a Boxes list confirmed country of importation (where available), genotype detected and year of detection in parenthesis for each Australian state as indicated by the arrow. There were no reports of measles disease in the NT or ACT during the period 2001-2006. "Source not identified" refers to individual cases of measles for which the source of importation (index case) could not be identified. Sequence interrogation of the measles virus nucleic acid detected did not substantiate the ongoing circulation of any genotype strain detected.

Population immunity

Although serological evidence of population immunity (measuring the proportion of sera samples that are positive for measles IgG antibodies) is not listed as a WPRO criterion, high population immunity is important evidence of elimination. WPRO provides three indirect measures of immunity (criteria 8–10) but population-based serosurveillance is the gold standard for assessing immunity. In Australia, national serosurveillance programmes conducted by the National Centre for Immunization Research and Surveillance of Vaccine Preventable Diseases, have included measles serosurveys during 1996–1998 before the Measles Catch-up Campaign, in 1999 to evaluate the success of the campaign and again in 2002 to evaluate a campaign targeting young adults.23,42 The 2002 serosurvey estimated that 93.9% of the Australian population was immune to measles, with immunity > 90% in all age groups, except 1 year-olds (64.9% positive, 95% CI: 59.7–69.8%), 2–4 year-olds (88.5% positive, 95% CI: 85.3–91.3%) and the 20–24 year age group (87.2% positive, 95% CI: 84.3–89.8%) (Fig. 5).42 More than 97% of people born before 1968, when measles vaccine first became available in Australia, had evidence of measles immunity.43

Fig. 5. Percentage of sera samples that are positive for measles IgG antibody by birth cohort, Australian national serosurvey, 200242

Evidence from this source is robust as the serosurveys all used the same methods and demonstrated comparable results to prospectively collected, random-cluster sampling of school-age groups.43,44 We therefore believe that Australia’s national serosurveys are an accurate measure of population level measles immunity. In 2002, population immunity was well above 90%.

Estimation of R

In disease modelling, infectious disease elimination is defined as the maintenance of the reproductive number, R below unity (R < 1).45 The reproductive number, R, summarizes the susceptibility of the population, its mixing patterns and the contagiousness of the disease, and represents the average number of secondary cases produced by a typical case.5,45 When R is > 1, the number of cases increase from one generation to the next and an epidemic ensues. When R is < 1, case numbers decrease from one generation to the next. If R is maintained constantly < 1 (the epidemic threshold), endemic transmission is considered to be eliminated. Using this definition, transmission of infective agents can still occur following elimination but endemic transmission is not re-established at a population level.46 Therefore, calculating R is a useful tool for monitoring the progress of elimination efforts. Estimates of R in the Australian setting have been obtained using several different methods with consistent results.35,36,42,47

The most robust method of calculating R is to use serosurveillance data to estimate susceptibility in each age group.45 Using data from the 1996–1998 Australian national serosurvey (before the catch-up campaign), R was estimated as 0.90.47 Since then, estimates of R using serological data have been well below the epidemic threshold; R was estimated as 0.57 from the 1999 serosurvey, and 0.69 in 2002, and modelled to remain below 0.8 until at least 2012.42,47 These estimates from serological data provide evidence that sustained transmission is unlikely to have occurred since 1999.

Besides using serosurveillance data, R can be estimated using enhanced disease notification data. There are three such methods. The first method uses the proportion of all cases that are identified as imported.48 As the recording of whether a measles case is imported is incomplete on the NNDSS database, this method is likely to overestimate R. Using 2001–2006 NNDSS data, 44 measles cases were recorded as imported out of a total of 446 cases and R was calculated as 0.90. Using data from the state of Western Australia, with a more complete follow-up of cases, R was estimated to be 0.62 between March 1999 and October 2000.36 The two other methods of calculating R rely on data for the distribution of the size and duration of outbreaks.46,49 Using outbreak data from enhanced surveillance in Victoria between 1998 and mid-2003, R was calculated to be 0.85–0.87 using the size of outbreak method and 0.73–0.76 using data on the duration of outbreaks.35 The similarity of the estimates of R using serosurveillance data to these estimates from case surveillance data (national and state estimates) supports the reliability of these estimates.

Discussion

The criteria used to justify our declaration of measles elimination in Australia are as follows:

Based on the number of notified cases, the most conservative year for declaration of elimination in Australia is 2005, although multiple lines of evidence suggest interruption of the endemic transmission of measles since 1999. The set of interim criteria defined by WPRO for the documentation of elimination of endemic measles transmission in a region may not be practical in many countries such as Australia, despite the existence of adequate surveillance systems, due to varying capacity in reporting these criteria at a national level. This particularly applies to the extensive documentation on the investigation of suspected cases at national level, which currently includes the discard rate, laboratory performance indicators and obtaining virologic samples from every presumptive chain of transmission (WPRO criteria 2–7, Table 2).

Australia satisfied the WPRO criterion of < 1 case per million population in 2005 and 2007. Although very low incidence is a significant criterion in defining measles elimination, we believe incidence rates > 1 case per million should not exclude declaration of elimination, especially if the cases are acquired outside the country and if other evidence suggests that sustained transmission has not occurred.

Adequate disease surveillance is an important criterion for establishing and monitoring measles elimination. The completeness of reporting, the sensitivity of the surveillance system, the use of laboratory confirmation, adequate epidemiological investigation of suspected measles cases and adequate genotyping of outbreaks are all important surveillance performance indicators.4 Although Australia’s surveillance reporting mechanisms do not currently record the investigation of presumptive measles cases at the national level as required by the WPRO criteria, we believe that surveillance is adequate for investigation of isolated cases of measles. Enhanced surveillance in the state of Victoria between 1998 and 2003 demonstrated reporting of non-measles suspected cases much higher than the WPRO target discard rate.29 However, higher rates of discard (i.e. more clinically suspected cases of measles) were reported during the period directly after notification of a case compared to periods when no measles case had been reported. The discard rate may therefore be a useful criterion at the beginning of the elimination phase but, when measles is rare, other diseases may be more likely to be suspected clinically and measles testing may not be requested by clinicians.

Australia’s national one-dose coverage of a measles-containing vaccine satisfies the WPRO criteria, while Australia’s national two-dose coverage is likely to be > 90%. However, Australia has provided additional data from a series of serosurveys that demonstrate 90% of the Australian population is immune to measles, providing the opportunity to identify population groups in need of targeted programmes. Additionally, modelling seroprevalence and surveillance data provides further evidence of elimination, with the reproductive number being maintained < 1. Finally, comprehensive molecular analyses provide substantial evidence of the absence of an endemic measles virus in Australia.

The declaration of the elimination of endemic measles from a region is not static and requires commitment to maintaining coverage, surveillance and outbreak control. Although England and Wales declared endemic measles eliminated in 2003 (Table 1), sustained transmission has recurred due to a reduction in vaccination rates,50 highlighting the requirement of maintenance as fundamental in declaring the elimination of endemic measles.

Under strict application of the WPRO criteria for case investigation, Australia would find it difficult to demonstrate measles elimination, as would most other countries that have previously declared elimination. However, the Australian criteria for the elimination of the transmission of endemic measles satisfy and extend the other WPRO criteria and lack only the WPRO surveillance process criteria 2–7. We believe the data presented confirm measles elimination in Australia and point to the need to broaden the current criteria for elimination of endemic measles transmission. ■


Acknowledgements

We thank the staff of the laboratories who provided the sera for the national serosurveys and laboratory staff at the Institute of Clinical Pathology and Medical Research for their help in processing and testing these sera. The following individuals and groups have contributed data to this paper: Doris Chibo and colleagues at the Victorian Infectious Diseases Reference Laboratory (genotyping data); Neils Becker from the National Centre for Epidemiology and Population Health, Australian National University and James Wood from NCIRS (estimating R); Brynley Hull from NCIRS (ACIR data); James Fielding, from the Communicable Disease Control Unit, Department of Human Services, Victoria (Victorian measles outbreak data); the Communicable Disease Network Australia (advice and NNDSS data); and the Public Health Laboratory Network and state and territory health departments (notification data).

Funding: CIDM-Public Health is supported by an infrastructure grant from the New South Wales Health Department. The National Centre for Immunization Research and Surveillance of Vaccine Preventable Diseases is supported by the Australian Government Department of Health and Ageing, the New South Wales Health Department and The Children’s Hospital at Westmead. Victorian Infectious Diseases Reference Laboratory acknowledges ongoing support from the Department of Human Services, Victoria. The Western Pacific Measles Regional Reference Laboratory at VIDRL receives support from WHO.

Competing interests: None declared.

References

Share