Measles remains an important cause of childhood mortality, especially in developing countries. In 2000, measles killed 770,000 children worldwide, accounting for nearly half of vaccine preventable deaths [1]. Failure to deliver at least one dose of measles vaccine to all infants remains the primary reason for high measles mortality and morbidity in developing countries [2]. Measles vaccination coverage among infants in Southeast Asia and Africa is still low, ranging between 54–55% in 1999 to 65–67% in 2003 [3]. However, Sri Lanka [4], Latin America [5], Romania [6] and South Korea [7], experienced outbreaks of measles in spite of sustained high coverage with single-dose vaccination strategy. Thus, the 2001–2005 WHO/UNICEF strategic plan for measles mortality reduction and regional elimination recommended achieving high routine vaccination coverage (>90%) in every district and ensuring that all children receive a second opportunity for measles immunization [2].

Population movement and high population densities facilitate transmission of the measles virus. Thus, outbreaks of measles are common among refugees and displaced populations [8–11]. Poor nutritional status, also common in these settings, has been associated with an increased risk of death following measles [11, 12]. As a result, measles outbreaks are a major killer during complex emergencies. Immunization of children against measles is considered a highly cost-effective priority among displaced populations housed in camps [13]. WHO and UNICEF recommend vaccinating all children from six months through 14 years of age along with vitamin A supplementation during emergencies [14]. At the minimum, children from six months through 4 years must be vaccinated, while vaccine availability, funding, human resources and local measles epidemiology may influence the choice of the age groups covered [14].

On 26 December 2004, a tsunami struck the coast of the state of Tamil Nadu, India, affecting more than 896,000 individuals and resulting in 7,983 deaths [15]. Affected individuals were provided with shelter in temporary camps. On 28 December 2004, the Field Epidemiology Training Programme of the National Institute of Epidemiology, Chennai, Tamil Nadu, sent three teams to the districts of Cuddalore, Nagapatinam and Kanyakumari that had been affected by the tsunami to assist district health authorities in (1) stimulating the routine surveillance system so that it would meet the needs of the emergency and (2) responding to outbreaks. On 29 December 2004, health authorities initiated supplemental measles immunization targeting children aged six to 60 months in all tsunami-affected villages in Tamil Nadu. On 30 December 2004, the emergency surveillance system set-up by the Indian Field Epidemiology Training Programme detected a cluster of measles in one of the relief camps in the district of Cuddalore. The estimated measles vaccine coverage exceeded 95% since 1988 in Tamil Nadu. (Director of Public Health, Govt. of Tamil Nadu, unpublished data). We investigated this cluster to estimate the magnitude of the problem and to propose recommendations for control.

We identified 101 cases that met the WHO case definition for measles. There were no deaths. We collected eleven blood samples from the case-patients for serological testing. Of these, eight were positive for IgM antibodies against measles virus. We took throat swabs from two children with active measles for the isolation of the virus. Phylogenetic analysis indicated that both the sequences of measles virus belonged to genotype D8.

Cases were reported with dates of onset between December 2004 and January 2005 (Figure 1) from tsunami-affected and unaffected areas. In both areas, the number of cases reported peaked during the second week following the first case. In tsunami-affected areas, 30 of the 71 cases (42%) occurred within 14 days of the tsunami. 71 measles cases were reported from 20 tsunami-affected villages (2005 population: 53,104) while the remaining 30 cases were from nine unaffected villages (2005 population: 17,357). The attack rate in tsunami-affected areas was lower than in tsunami un-affected areas, although the difference was not statistically significant (p = 0.24). The median ages of case-patients in tsunami-affected and un-affected areas were 54 months (range: 4 months-12 years) and 60 months (range: 6 months-21 years), respectively (p = 0.471). There was no major difference in the age-specific attack rates in tsunami-affected and unaffected areas (Table-1). 26 of the 71 case patients (36.2%) from tsunami-affected areas and 14 of 30 (46.7%) cases in tsunami unaffected areas were above 60 months of age. Among the 101 cases, three (3%) had vaccination records indicating they had received measles vaccine, three (3%) had vaccination records indicated that they had not received measles vaccine and 95 (94%) had no vaccination records.

Measles virus circulated between December 2004 and February 2005 in tsunami-affected and un-affected areas of the district of Cuddalore, Tamil Nadu. Transmission occurred despite high one-dose measles vaccine coverage in the state. Active surveillance in tsunami affected areas with early case detection and management might have contributed to the absence of fatality. In addition, supplemental measles immunization targeting children between six and sixty months might have decreased the population of susceptible children in tsunami affected areas which might have limited the spread of this cluster.

The findings of the present investigation suggest that measles transmission in the affected area was unrelated to the tsunami for a number of reasons. First, attack rates were not statistically different in tsunami affected and unaffected areas. Second, the attack rate over time showed a similar pattern in both areas. Third, a large proportion of cases occurred within one incubation period after the occurrence of the tsunami. Fourth, the age distribution of the cases did not differ in tsunami affected and unaffected areas. Finally, the size of shelters in tsunami affected areas in Cuddalore ranged from 115 to 2500 (average: 656)[17]. It was much smaller than the major population concentration usually leading to outbreaks in refugee setting [22].

This cluster of measles occurred despite a high one-dose measles vaccine coverage in Tamil Nadu (Director of Health Services, unpublished data). An independent LQAS vaccine coverage survey conducted in one primary health care center area also confirmed high one-dose coverage in the state [23]. The results of the phylogenetic analysis indicated that both the sequences of measles virus belonged to genotype D8. Genotype D4 and D8 have been isolated from different districts of Tamil Nadu over the last four years from sporadic cases as well as during outbreaks (National Institute of Virology, unpublished data). The absence of importation of a new genotype further supports the hypothesis that a measles virus of that genotype circulates in the area despite high one-dose vaccination coverage. Measles transmission in areas with high one-dose vaccination coverage points to the limitation of the one dose measles vaccination strategy in measles control. Our model suggested that even with a high coverage scenario, the one-dose measles immunization strategy in place in Tamil Nadu allowed an accumulation of a population of susceptible children that exceeded the size of two birth cohort, thereby allowing transmission. Outbreaks of measles may occur when the population of susceptible in a community exceeds one birth cohort [20]. Measles outbreaks in areas with high coverage with a single-dose strategy have been reported in Sri Lanka [4], Latin American countries [5] and in Romania [6]. A second opportunity for measles immunization in areas with high one-dose coverage may provide additional immunity needed for effective measles control [24]. Our model also suggested that children in the district could have benefited from a second opportunity for measles vaccination. To reduce the measles mortality in India, the Government of India developed a strategic plan in 2005 with the objective of (1) reducing the measles mortality by two-thirds by 2010, compared to 2000 estimates and (2) achieving at least 90% coverage with measles vaccine in 80% of the districts of the country by 2009 [25]. In addition, states like Tamil Nadu where one-dose coverage exceeds 90%, the existing network of the national polio surveillance project will be used to collect good quality epidemiological data through active surveillance and outbreak investigations.

During this cluster, the number of cases increased and decreased sharply over time. When measles affects more than one village, transmission is expected to last for number of incubation periods. The decrease of incidence observed in the present investigation is unlikely to be due to surveillance bias. The passive surveillance system continued to be stimulated for a number of weeks following the tsunami and was progressively included as a pilot project in the Integrated Disease Surveillance project (IDSP), the new national surveillance initiative. Supplemental immunization activities that took place at the time of the outbreak might have contributed to the rapid decrease of incidence. These included (1) the post-tsunami immunization campaign that targeted children of 6–60 months between 29 December 2004 and 9 January 2005 in the affected areas and (2) a practice referred to as 'ring immunization' under which children aged 6–60 months of age living within a radius of 5 kilometers of the case of measles. Our investigation however was not able to document any age shift in the cases following the campaign as the outbreak ended rapidly. More generally, we were not able to quantify the impact of supplementary immunization activities on the dynamic of the outbreak.

We did not examine possible missed opportunities for measles immunization in this investigation. Missed opportunities for immunization are important to identify to increase immunization coverage [26]. As a consequence, we were unable to propose specific recommendations to eliminate them. However, the high one-dose measles coverage in the state of Tamil Nadu suggests that addressing the need of a second opportunity for measles vaccination is of higher priority that preventing missed opportunities in this specific setting.

The results of our investigation suggested that measles virus transmission continued despite high one-dose measles vaccine coverage and that the cluster was unrelated to the tsunami. The World Health Organization recommends vaccinating children aged six months to 14 years in refugee and internally displaced persons camps [14] but also mentions that age group to be targeted during such campaigns depends on the local epidemiology of the disease. In Tamil Nadu, the high one-dose coverage may have contributed to the decision to restrict the target age group to children 6 to 60 month of age after the tsunami emergency. However, vaccinating children only up to five years of age might have left some children susceptible to measles as more than 36% of the cases were older than five years of age. The results of our investigation support the recommendation to vaccinate children up to 14 years of age during complex emergencies, even if the one-dose coverage is high. It is also necessary that information, education and communication activities targeting mothers as well as health care workers emphasize the importance of vaccination cards to allow appropriate evaluation of vaccination coverage in Tamil Nadu. In the longer term, the ministry of health of the government of India will be examining how best to address measles as a public health problem given the heterogeneity of the country with respect to measles vaccination coverage. While some states will need to focus on increasing the one-dose vaccination coverage, a number of southern states might be considered for second opportunity of measles vaccination.

The present cluster of measles was unrelated to accumulation of susceptible population on account of tsunami. Measles transmission despite high one-dose vaccination coverage points towards the need of second opportunity of measles immunization. During complex emergencies, measles immunization should be offered to children from 6 month to 14 years of age.

The author(s) declare that they have no competing interests.