Measles outbreak in a community with very low vaccine coverage, the Netherlands.

A 1999-2000 measles epidemic in the Netherlands started with an outbreak in an orthodox reformed elementary school with 7% vaccine coverage. The overall attack rate was 37%: 213 clinical cases among the 255 participating pupils (response 62%) and 327 household members. The attack rate ranged from 0% for the oldest groups of pupils to 88% for the youngest, who had not been exposed in previous measles epidemics. None of 25 vaccinated pupils had clinical symptoms. Among pupils with clinical symptoms, the self-reported complication rate was 25%. These data confirm that measles infection causes severe disease and that vaccination is the most effective means of preventing the disease and its complications. The data also show that clusters of persons refraining from vaccination interfere with measles elimination even in populations with very high overall vaccine coverage (96%).

A 1999-2000 measles epidemic in the Netherlands started with an outbreak in an orthodox reformed elementary school with 7% vaccine coverage. The overall attack rate was 37%: 213 clinical cases among the 255 participating pupils (response 62%) and 327 household members. The attack rate ranged from 0% for the oldest groups of pupils to 88% for the youngest, who had not been exposed in previous measles epidemics. None of 25 vaccinated pupils had clinical symptoms. Among pupils with clinical symptoms, the self-reported complication rate was 25%. These data confirm that measles infection causes severe disease and that vaccination is the most effective means of preventing the disease and its complications. The data also show that clusters of persons refraining from vaccination interfere with measles elimination even in populations with very high overall vaccine coverage (96%).
(preschool, which is voluntary) after vacation; therefore, we sent them the questionnaires in August only. Measles vaccination history of all 412 pupils of the school (grades 0 to 8) was verified at the Provincial Vaccination Administrations (PVA).

Laboratory Tests
The presence of specific serum IgM antibodies was determined with a commercially available IgM-capture enzyme-linked immunosorbent assay (ELISA) according to procedures recommended by the manufacturer (Meddens mucapture ELISA for measles, Biotest, Denville, NJ). IgG antibody concentrations were measured by an in-house ELISA (6).

Case Classification
We classified cases according to a modified Centers for Disease Control and Prevention case definition (7): Confirmed cases had >3 days of rash, fever >38.3°C, and either cough, conjunctivitis, or coryza; suspected cases had rash and fever according to questionnaire or recent measles according to form, with limited information on household members of pupils. We considered positive serologic results (positive IgM or a minimal fourfold rise in IgG titer) or virus isolation from blood or oropharyngeal swab to be laboratory evidence of measles infection.

Data Analyses
Attack rates for clinically confirmed and suspected measles cases were calculated by sex, year of birth, vaccination history, history of measles, and susceptibility, i.e., no vaccination, no history of measles, and birth in 1986 or later. Persons born before 1987 experienced measles epidemics in 1987-88 and in 1992-93. This was confirmed by the fact that we observed only one clinical case (the patient was born in 1986) among all persons born before 1987 (n = 226). Therefore, we considered all persons born before 1986 without information on history of measles to have had measles.
Vaccine efficacy estimates were based on the attack rate of measles among pupils who reported no history of measles in the questionnaire and by vaccination history as given by PVA. Symptoms and complications as reported in the questionnaires were described for those who had clinically confirmed or suspected measles and who had completed at least one questionnaire. We used the chi-square test to test differences in attack rates regarding categorical variables. A p value of <0.05 was considered statistically significant.

Response
Responses to questionnaires, limited information, and collected biological samples (from pupils and household members) are shown in Table 1. All families with one or more reported measles patients from June 21 through July 2, 1999, had elementary school pupils in their households. We obtained questionnaires on 299 persons and limited information on 283 of their household members from 123 families, and we obtained biological samples from 100 persons in 26 families.

Description of the Outbreak
In total, 213 cases of measles (110 confirmed and 103 suspected) were identified ( Table 2); 138 were in pupils. All suspected cases were epidemiologically linked to a confirmed case through school or family contacts. Therefore, we consider suspected cases true measles cases and describe our results for the confirmed and suspected cases together.
The epidemic curve is shown in Figure 1. Day 1 of rash was known in 137 of the 213 confirmed and suspected cases and occurred from June 15 to July 20, 1999. The number of   Table 4. c Here, 99 persons with laboratory case classification are shown instead of 100 as in Table 1, since 1 person did not provide blood and throat swab specimens, but only saliva and urine.  persons per household was 3 to 18 (median 6). The number of reported cases per household was 0 to 9 (median 2): 37 (30%) households reported no cases, including 12 (10%) households with children vaccinated against measles.

Attack Rates
The overall attack rate among confirmed and suspected cases was 37% (Table 3), 0% for the oldest pupils to 88% for the youngest (Figure 2). Two (1%) of the 213 patients were born in 1999; 166 (78%) from 1992 to 1998; and 43 (20%) from 1988 to 1991. Two (1%) patients were born before 1988 (1986 and 1987). The distribution of cases and attack rate by sex, vaccination history, history of measles, and susceptibility (i.e., no vaccination, no history of measles, and born in or after 1986), is shown in Table 3. Except for sex, all variables were associated with the attack rate (p <0.05).
The attack rate among susceptible pupils was 91% (133 of 146). Of the 28 nonpupils considered susceptible, 24 (86%) had clinical cases ( Table 3). Three of the four who did not become ill were probably protected by maternal antibodies (date of birth from December 1998 to April 1999).
Among the 69 pupils considered not susceptible because of reported history of measles, one had clinical symptoms and laboratory confirmation of measles infection (Table 3). According to the questionnaire, this grade 1 pupil had measles in 1998. No vaccination was registered at PVA. This child probably had another rash disease in 1998. No cases were observed among the 195 nonpupils considered not susceptible (Table 3).

Laboratory Results
The diagnosis was laboratory confirmed for 39 of the 51 clinically confirmed and suspected cases with one or two biological samples, the first of which was collected at or just after Day 1 of rash (IgM postive or IgG titer rise). We had collected only one sample in each of the remaining 12 cases; measles rash did not develop in these patients until 3 to 20 days later. As expected, IgM antibodies could not be detected in these cases. Five of 48 asymptomatic persons who had provided biological samples had laboratory evidence of measles infection (Tables 2, 4).

Vaccination History and Vaccine Efficacy
Of all 412 pupils, 28 (7%) had been vaccinated, according to PVA records. Of the 255 participating pupils, 25 (10%) had been vaccinated: 20 had one dose of MMR vaccine, and 5 had had two doses. None of the 25 vaccinated pupils reported measles symptoms (Table 3). Four (one parent and three young children) (10%) of the 42 nonpupils with a questionnaire reported vaccination against measles. None reported symptoms.

Conclusion
We have described an outbreak of measles in a mostly unvaccinated population. From this outbreak, measles spread and affected mainly (94%) unvaccinated persons from orthodox reformed communities. By May 2000, 3,292 cases of measles were reported to the national registry, including three measles-related deaths and 72 hospitalizations.

Attack Rates
The susceptibility levels and attack rates were closely related to the number of previous epidemics encountered; those persons born after 1992, when the last epidemic began, had the highest susceptibility levels and attack rates. The 1999 birth cohort and part of the 1998 birth cohort are exceptions because they were partially protected by maternal antibodies. Sex was not associated with the attack rate, which is in accordance with previous reports (8). The infectivity of the measles virus is shown by the high attack rate (90%) among those considered susceptible (i.e., those with no history of measles or vaccination).

Import and Export of Measles Virus
Measles viruses isolated from patients showed that the epidemic was caused by a D6 type measles virus, a genotype widely distributed throughout Europe (9). Genotype D6 had frequently been isolated from unrelated cases in the Netherlands between 1993 and 1999 (van Binnendijk et al., unpub. data). During this period, the number of measles cases reported in the Netherlands decreased to one of the lowest rates in Europe (<1 per million in 1998). However, because of low vaccine coverage in orthodox reformed communities, the number of susceptible persons increases. Consequently, measles epidemics still occur, despite high national vaccine coverage and population immunity (1,6). Previously, we showed that measles is not endemic in the Netherlands, not even in areas with low vaccine coverage (10). This was confirmed in this 1999-2000 epidemic; no more cases were reported within 1 year after the start of the outbreak. Therefore, we assume that the epidemic was initiated by import from another country. Until the measles virus is eradicated, circulation will continue worldwide and epidemics will occur. During this epidemic, visiting relatives exported measles to Canada. The outbreak was restricted to 17 cases within an orthodox reformed community in Canada as a result of stringent measures (e.g., closing the school) (11).

Laboratory Results
We observed five asymptomatic persons with serologic proof of measles infection. All had been in close contact with one or more measles patients. Two were children, one vaccinated (#5 in Table 4) and one without recorded measles vaccination or history of measles disease (#4). Incomplete immunity in the presence of residual maternal antibodies may have developed in the latter child during the 1992 measles epidemic (12). Two adults (#2 and #3) reported history of measles, the third (#1) reported no history of measles but might have had measles, on the basis of the year of birth. However, this person might also have had subclinical primary infection.
We assume that the increase in specific IgG (#2-#5) reflects secondary immune response in persons reexposed to measles virus, as has been demonstrated (13-15). We have not been able to detect virus, either by virus culture or RT-PCR from blood or oropharyngeal swab (data not shown), from any of these subclinically reinfected persons, as was recently shown for an immune mother of an adult measles patient (16). However, even if virus can be detected in blood, urine, or saliva, the critical issue is whether the virus load in these subclinically reinfected persons is high enough to transmit the measles virus.

Vaccination History and Vaccine Efficacy
We observed low vaccine coverage (7% to 10%), but excellent vaccine effectiveness (100%) for the measles component of the MMR vaccine; none of the vaccinated persons had measles symptoms. In the measles epidemic following this outbreak, 5% of the reported cases patients were vaccinated; almost all of them had received one dose (5).  The real percentage of vaccinated patients is probably smaller. We expect that more vaccinated than unvaccinated persons with measles symptoms are seen and reported by GPs.

Symptoms and Complications
Measles is sometimes thought of as a mild disease. However, we observed a self-reported complication rate of 25% for all patients, 68% of whom consulted a GP. We do not know whether children who did not complete a form on complications consulted a GP. The percentage of consultations for uncomplicated measles cases could be smaller than that for complicated cases. Therefore, the percentage of consultations for all cases may be overestimated.
The complication rate of 25% is based on self-reported complications, and the diagnosis was not always confirmed by a physician. This could explain why the complication rate is somewhat higher than expected for measles (8,12). Still, burden of disease was very high in the participating measles patients. During the following epidemic (1999)(2000), three measlesrelated deaths and 72 hospitalizations were reported (5).
In this descriptive study of a measles outbreak with an attack rate of 90% among susceptible persons, we have shown that measles disease is severe, even in an industrialized country. Vaccination is the most effective means of preventing the disease and its complications. The national vaccine coverage of 96% for both doses of MMR is theoretically high enough to eliminate measles (17). However, despite this very high coverage, measles epidemics still occur as a result of areas with low vaccine coverage. In these sociodemographically clustered, mainly unvaccinated communities, the number of susceptible people increases, and consequently epidemics occur periodically. The clustering of unvaccinated persons is the critical factor for measles elimination in the Netherlands.