Measles Epidemic in The Netherlands,

Transcription

1 1483 CONCISE COMMUNICATION Measles Epidemic in The Netherlands, Susan van den Hof, 1 Marina A. E. Conyn van Spaendonck, 1 and Jim E. van Steenbergen 2 1 National Institute of Public Health and the Environment, Department of Infectious Diseases Epidemiology, Bilthoven, and 2 National Coordination Center for Communicable Disease Outbreak Management, Utrecht, The Netherlands In , a measles epidemic occurred in The Netherlands, with 3292 reported cases; 94% of the affected patients had not been vaccinated. Only 1 patient had received 2 doses of vaccine. Three patients died, and 16% had complications. For the unvaccinated population, the incidence per 1000 inhabitants 15 months to 14 years old increased from 83 (95% confidence interval [CI], ), in municipalities with vaccine coverage rates 90%, to 200 (95% CI, ), in municipalities with coverage rates 195%; for the vaccinated population, the incidence increased from 0.2 (95% CI, ) to 1.4 (95% CI, ). Unvaccinated individuals were 224 times (95% CI, times) more likely to acquire measles than were vaccinated individuals; the relative risk increased with decreasing vaccine coverage. Herd immunity outside unvaccinated clusters was high enough to prevent further transmission. More case patients came from the vaccine-accepting population living among unvaccinated clusters than from individuals who declined vaccination and who lived among the vaccine-accepting population. Although measles, mumps, and rubella (MMR) national vaccine coverage is high in The Netherlands (94% 96% for both doses), coverage is not homogeneously distributed within the country [1]. In 1999, 36 (7%) of 539 municipalities had a coverage rate of!90%; 12 of those municipalities had coverage rates of!80%. These municipalities are situated in a geographic belt from the southwest to the middle of the northern part of the country (figure 1A), where a relatively high number of members of orthodox-reformed churches reside. Most individuals from these sociogeographically clustered religious communities, estimated to comprise 2% of the total population, decline vaccination. Individuals who decline vaccination for their children for other reasons may be greater in numbers but live more or less scattered throughout the country. Before the introduction of measles vaccination in The Netherlands in 1976, national epidemics occurred every other year. Afterward, epidemics took place in 1983, , and [2 4]. Then, in June 1999, an outbreak of measles in a Received 31 May 2002; revised 31 July 2002; electronically published 29 October Previously published in part: van den Hof S, van den Kerkhof JHTC, ten Ham PBG, van Binnendijk RS, Conyn van Spaendonck MAE, van Steenbergen JE. Mazelen epidemie in Nederland, [Measles epidemic in The Netherlands, ]. Ned Tijdschr Geneesk 2001; 154: Permission granted for republication of these data from the editor of Ned Tijdschr Geneesk. Reprints or correspondence: Dr. Susan van den Hof, National Institute of Public Health and the Environment, Dept. of Infectious Diseases Epidemiology, PO Box 1, 3720 BA Bilthoven, The Netherlands (Susan.van.den.Hof@rivm.nl). The Journal of Infectious Diseases 2002;186: by the Infectious Diseases Society of America. All rights reserved /2002/ $15.00 municipality within a religious group with low vaccine coverage was reported. After serological confirmation, a case register was set up to closely monitor spread from the outbreak and to obtain information on complications, because this is not provided in the regular notification system. Here, we describe that measles epidemic and quantify the spread of measles to the vaccinated population. Materials and Methods Case register. Measles is a mandatory notification disease in The Netherlands. Physicians report cases to the regional Municipal Health Service (MHS), which, after verification, sends notification data with the personal identifiers removed to the Inspectorate of Health (IGZ) [1]. Since April 1999, the case definition for reports of measles has been a fitting clinical picture in combination with laboratory confirmation of the infection or contact (within 3 weeks) with a person with laboratory-confirmed infection. Immediately after a measles outbreak within a village with low vaccine coverage (78%) was serologically confirmed in June 1999, a case register was set up at the National Coordination Center for Communicable Disease Outbreak Management, under the auspices of the IGZ. After notification of a new case, the MHS completed a questionnaire by telephone interview of the reporting physician and the parents of the case patient. Two weeks later, the physician was asked to provide data on complications (pneumonia, otitis media, and other complications) and outcomes of measles infection (hospitalization and/or death). The register closed in May Data analysis. We calculated the age of the patient on the basis of the first day of illness and the date of birth. If the first day of illness was not known, we used as a proxy (in order of preference) date of the first day of fever, date of the first day of rash, date of

2 1484 van den Hof et al. JID 2002;186 (15 November) Figure 1. A, Municipal vaccine coverage of first measles, mumps, and rubella (MMR) dose (1996 birth cohort at 1 January 1999). B, Municipal incidence of reported measles cases among children!15 years old in The the visit to the general practitioner (GP), date of confirmation of the diagnosis, or date of reporting to the MHS. The vaccination status of all Dutch inhabitants is routinely registered in a computerized database in which data on births, deaths, and migration are processed continuously. Coverage on a municipal, provincial, and national level for each vaccine is reported annually [1]. For example, coverage of the first MMR dose for the 1996 birth cohort is assessed at 1 January We estimated age-specific municipal numbers of vaccinated and unvaccinated children 15 months to 14 years old by applying the 1996 birth cohort vaccine coverage to this age group, because municipal coverage levels had been consistent for 15 years. We estimated age-specific municipal incidence rates by dividing the total number of cases among vaccinated and unvaccinated children by the total number of vaccinated and unvaccinated children. Finally, we estimated age-specific municipal relative risks by dividing the incidence rate among unvaccinated children by the rate among vaccinated children. Incidence rates and relative risks were averaged over the municipalities and weighted by number of inhabitants, and corresponding 95% confidence intervals (CIs) were estimated. The association between municipal vaccine coverage and incidence rates among vaccinated and unvaccinated children, adjusted for age and population size, was assessed by means of a generalized linear model with a binomial distribution and logit as link function. Results Between June 1999 and May 2000, 3292 cases were reported, for an overall incidence of 0.21 cases/1000 inhabitants. A cluster of 5 cases with onset of disease in April 1999 was reported retrospectively in July 1999, whereas there were no reports of cases in May The peak of the epidemic was in November All isolates collected were a D6-type measles virus, a genotype widely distributed in Europe. The index case and, therefore, the exact origin of the imported virus could not be identified. Of the 3286 (99.8%) case patients whose age was known, the median age was 6 years (range, 0 52 years). Four case patients (0.1%) were!6 months old, 191 (6%) were 6 14 months old, 985 (30%) were 15 months to 4 years old, 1456 (44%) were 5 9 years old, 452 (14%) were years old, 103 (3%) were years old, and 95 (3%) were 119 years old. Attack rates were highest among those 4 7 years old. Of the 3182 (97%) patients with known sex, 1618 (51%) were male. Complications. One or more complication was reported for 519 (16%) case patients (table 1). Three children (2, 3, and 17 years old) died as a result of complications [5]; all 3 were unvaccinated for religious objections. The proportion of patients with 1 complication was 22% among patients!15 months old, 19% among patients 15 months to 4 years old, 16% among pa-

3 JID 2002;186 (15 November) Measles in the Netherlands 1485 Table 1. Measles vaccination status, reasons for not being vaccinated, and complications in reported patients with measles in The Characteristic No. (%) of patients Vaccination status Vaccinated 158 (5) Unvaccinated 3092 (94)!15 months old Positive intention to vaccinate 96 (3) Religious objection 54 (2) 15 months old Religious objection 2657 (81) Other reason 170 (5) Born before introduction of vaccination 57 (2) Unknown 58 (2) Unknown 42 (1) Complications (16) Deceased 3 (0.1) Hospital admission 72 (2) Otitis media at home 187 (6) Pneumonia at home 140 (4) Otitis media and pneumonia at home 29 (1) Other complications at home 88 (3) None 2368 (72) Unknown 405 (12) Total 3292 (100) tients 5 9 years old, 11% among patients years old, and 15% among patients 119 years old. Hospital admission was more frequently reported for adults (i.e., patients 119 years old; 7% of cases) than for all younger age groups (2% of cases). Respiratory complications such as pneumonia were the most prevalent reasons for hospital admission in all age categories. Otitis media and pneumonia were the most common complications in patients who were treated at home (table 1). Patients with other complications at home predominantly had respiratory infections. Vaccination status. Vaccination status of the patients is shown in table 1. Most patients (94%) were unvaccinated; 83% reported religious or fundamental objections to vaccination. Of the 158 vaccinated patients (5%), 1 had received 2 doses of MMR or measles vaccine, 139 had received 1 dose, and 18 reported having received an unknown number of doses. On the basis of year of birth, all of these 18 case patients were not (yet) eligible for a second dose, and we can assume that they had been vaccinated only once. Besides the 158 vaccinated case patients (5%), another 96 (3%) were from the vaccinating population but had been too young to be eligible for the first MMR dose, which usually is given at age 14 months. Incidence related to vaccine coverage. A third (176/539) of all municipalities reported measles cases. However, 75% of all cases were reported by 37 municipalities (7%), with the spatial incidence distribution reflecting the geographical belt with low vaccine coverage (figure 1). In the 162 municipalities with reported cases in the age group 15 months to 14 years (accounting for 88% of all patients), the incidence was higher, because vaccine coverage was lower for both the unvaccinated and vaccinated populations (table 2). Unvaccinated individuals were 224 (95% CI, ) times more likely to acquire measles than vaccinated individuals, and the relative risk increased with decreasing vaccine coverage. This effect remained when smaller age intervals and population size were taken into account in a statistical model, although the effect was greatest in the age groups!10 years old (data not shown). Discussion During this epidemic, 3292 case patients, predominantly unvaccinated individuals from areas with low vaccine coverage, were reported to the case register. Most of the cases that occurred in vaccinated persons were sporadic and, like cases that occurred in unvaccinated persons, occurred in an area with low vaccine coverage. The spread of measles to areas with high vaccine coverage was hardly observed, which indicates that herd immunity in the general population was sufficient to prevent further transmission. The effect of herd immunity was shown even more clearly by the increased risk of measles in unvaccinated individuals in municipalities with decreased vaccine coverage. The paradoxical result of herd immunity is that more cases were reported from the group of vaccine-accepting individuals who live among clusters of unvaccinated individuals than from those who decline vaccination for other than religious reasons who live among vaccinated individuals. It has been observed before in the United States [6, 7] that staying in an area with low vaccine coverage is the most important risk factor for vaccinated children to contract measles. Transmission rates were relatively low during the summer holidays, and the number of patients increased rapidly after the start of school. This and the increase in attack rate from age 4 years onward emphasizes the importance of schools in the transmission of measles [8]. Like the poliomyelitis epidemic in 1992 [9], the virus was spread to Canada by visiting relatives. As a result of strict measures, such as closure of the school, the number of patients in Canada was restricted to 17. Vaccination status. The low proportion (5%) of vaccinated case patients and the lack of spread in the vaccinating population underlines the effectiveness of vaccination in protection Table 2. Municipal incidence rates per 100,000 inhabitants of reported patients with measles in unvaccinated (Inc unvacc ) and vaccinated (Inc vacc ) populations 15 months to 14 years old and corresponding confidence intervals (CIs), by municipal vaccine coverage rates, in The Vaccine coverage No. of patients Inc unvacc (95% CI) Inc vacc (95% CI) RR (95% CI) 90% (53 113) 0.2 ( ) 507 ( ) 90 95% (62 160) 0.5 ( ) 206 ( ) 195% ( ) 1.4 ( ) 72 (39 501) NOTE. RR, relative risk.

4 1486 van den Hof et al. JID 2002;186 (15 November) against measles and its complications. In 1987, a 2-dose strategy was implemented to increase the proportion of immune people. In 1999, an equal number of birth cohorts were vaccinated once and twice against measles. The fact that only 1 (0.6%) of 158 vaccinated case patients had received a second dose shows the additional protection afforded by the second vaccination. Complications. Sixteen percent of all reported case patients developed 1 complication. It seems plausible that a higher proportion of case patients with complications is seen and reported by their physicians than patients with uncomplicated cases; therefore, the true proportion of case patients with complications may be lower than the 16% observed in reported cases. Still, measles infection causes substantial morbidity: 3 patients died, and 72 others were reported to have been hospitalized. By means of a capture-recapture approach (data not shown), we estimated that the true number of case patients hospitalized between April 1999 and May 2000 was 157 (95% CI, ) [10]. Upper and lower respiratory tract complications were most often reported, both in patients who were hospitalized and those who were treated at home. In accordance with other reports [11, 12], we observed a higher complication rate in infants and adults. The hospitalization rate was elevated only in adults. Incomplete reporting. Not all patients consult a physician, and not all patients seeking consultation are reported; thus, the true number of case patients will be higher than the 3292 reported. The proportion of case patients seeking medical advice was 60% in Denmark just before the introduction of vaccination [13]. This proportion probably is lower in the orthodox-reformed group, which generally views measles as a common, mild childhood disease. The GP practice is a typical example in the municipality where the measles epidemic was first noticed; 50 (30%) of 164 identified patients consulted a GP, of whom 15 (30%) were reported (C. E. D. van Isterdael, personal communication). Thus, 9% of the total number of cases were reported by the GPs. Fortyseven percent (74/157) of the estimated total number of hospitalized case patients were reported as such to the case register. Under the assumption that 30% of the case patients visit a physician and that 30% 47% of the visiting patients are reported by the physician, then the estimated total number of case patients would be 23,000 37,000. These estimates indicate that the true number of case patients was much higher than reported. In conclusion, this epidemic once more shows that, even in a wealthy country, measles is not just a mild childhood disease but can cause severe complications and even death. Vaccination offers effective protection against measles disease and complications. Although the population of those who decline vaccination for religious reasons in The Netherlands is not large enough to sustain transmission of measles virus [4], the number of susceptible people in this population will again be sufficient in several years to cause a similar outbreak after importation of measles virus. These unvaccinated clusters pose a risk to themselves and to surrounding vaccine-accepting individuals, although herd immunity in the vaccinated population is high enough to prevent further transmission. As a consequence, more case patients were from the vaccinated population that lives among unvaccinated religious groups than from the individuals who declined vaccination for other reasons who live among vaccinated groups. Acknowledgments The case register was realized through the cooperation of physicians, especially public health nurses of the respective municipal health services. We thank Hans van den Kerkhof, Peter ten Ham, and Rob van Binnendijk, for comments on the Dutch manuscript; Rick Grobbee and Hester de Melker, for suggestions on the English version; and Laurens Zwakhals, for help in creating the figure. References 1. Inspectorate of Health. Immunisation status in the Netherlands per 1 January 1999 [in Dutch]. The Hague: Inspectorate of Health, Bijkerk H, Bilkert-Mooiman MA, Houtters HJ. The immunization status of patients registered with measles during the epidemic [in Dutch]. Ned Tijdschr Geneeskd 1989;133: van der Zwan CW, Plantinga AD, Rümke HC, Conyn van Spaendonck MAE. Measles in The Netherlands: epidemiology and the effect of vaccination [in Dutch]. Ned Tijdschr Geneeskd 1994;138: Wallinga J, van den Hof S. Measles epidemiology in The Netherlands: an exploratory analysis of notification [in Dutch]. Ned Tijdschr Geneeskd 2000;144: Centers for Disease Control and Prevention. Measles outbreak Netherlands, April 1999 January MMWR Morb Mortal Wkly Rep 2000;49: Salmon DA, Haber M, Cangarosa EJ, et al. Health consequences of religious and philosophical exemptions from immunization laws. JAMA 1999;282: Feikin DR, Lezotte DC, Hamman RF, et al. Individual and community risks of measles and pertussis associated with personal exemptions to immunization. JAMA 2000;284: Finkenstädt BF, Grenfell BT. Time series modelling of childhood diseases: a dynamical systems approach. Appl Stat 2000;49: Centers for Disease Control and Prevention. Isolation of wild poliovirus type 3 among members of a religious community objecting to vaccination Alberta, Canada, MMWR Morb Mortal Wkly Rep 1993;42: van den Hof S, Smit C, van Steenbergen JE, de Melker HE. Hospitalizations during a measles epidemic in The Pediatr Infect Dis J (in press). 11. Miller DL. Frequency of complications of measles, BMJ 1964;2: Redd SC, Markowitz LE, Katz SL. Measles vaccine. In: Plotkin SA, Mortimer EA Jr, eds. Vaccines. 3rd ed. Philadelphia: WB Saunders, 1999: Horwitz O, Grunfeld K, Lysgaard-Hansen B, et al. The epidemiology and natural history of measles in Denmark. Am J Epidemiol 1974;100:

5 1704 ERRATUM In an article in the 15 November 2002 issue of the Journal (van den Hof S, Conyn-van Spaendonck MAE, van Steenbergen JE. Measles epidemic in The J Infect Dis 2002;186:1483 6), there are errors in the abstract and in table 2. In the abstract, the fourth sentence should read as follows: For the unvaccinated population, the incidence per 1000 inhabitants 15 months to 14 years old increased from 83 (95% confidence interval [CI], ), in municipalities with vaccine coverage 195% [not 90%], to 200 (95% CI, ), in municipalities with vaccine coverage 90% [not 195%]; for the vaccinated population, the incidence increased from 0.2 (95% CI, ) to 1.4 (95% CI, ). The corrected table 2 is shown in the next column. The authors regret these errors. Table 2. Municipal incidence rates per 100,000 inhabitants of reported patients with measles in unvaccinated (Inc unvacc ) and vaccinated (Inc vacc ) populations 15 months to 14 years old and corresponding confidence intervals (CIs), by municipal vaccine coverage rates, in The Vaccine coverage No. of municipalities Inc unvacc (95% CI) Inc vacc (95% CI) RR (95% CI) 195% (53 113) 0.2 ( ) 72 (39 501) 90 95% (62 160) 0.5 ( ) 206 ( ) 90% ( ) 1.4 ( ) 507 ( ) NOTE. RR, relative risk. The Journal of Infectious Diseases 2002;186:1704