This paper reports a descriptive study of two outbreaks of mumps in Spanish towns in 1998 and an assessment of the efficacy of the vaccine used. It provides evidence of the low level of protection against the mumps virus that has been observed previously in subjects vaccinated with the Rubini strain.

The low efficacy of the Rubini strain of mumps used in measles, mumps, and rubella (MMR) vaccine has been shown by the sentinel surveillance system in Switzerland for 1986-93 (1) and in reports of outbreaks in Portugal, Italy, and Switzerland (2-7). As a result, Spain implemented a programme in 1997 to investigate outbreaks of mumps, assess vaccine potency and the cold chain, and measure the efficacy of vaccine in mumps outbreaks in Spain.

Outbreaks of mumps in two Spanish towns, Almoradí (Valencian region) and Manacor (Balearic Isles), affected large numbers of people in the second half of 1998. The attack rate was high in children aged 2 to 5 years, 93% of whom were vaccinated.

In 1981 MMR vaccine was introduced in Spain for children aged 15 months. In 1994, a second dose of MMR was added at age 11 years. Since 1999, this second dose has been offered to children aged 3 to 6 years. Two mumps vaccines were originally used in Spain – those containing the Jeryl-Lynn and the Urabe Am9 strains. In 1992, however, the Urabe strain was withdrawn, and the vaccine containing the Rubini strain was gradually introduced nationwide, albeit to different degrees in the various autonomous regions.

The childhood immunisation programme is not mandatory in Spain but has been administered free of charge ever since the first immunisation campaigns against poliomyelitis were introduced in 1963. From 1985 onwards, MMR vaccination coverage exceeded 80% at a national level. In 1998 an overall coverage of 93% for the first dose was recorded among children under the age of 2 years.

As a consequence of the consolidation of the childhood immunisation programme and the maintenance of high vaccination coverage, the incidence of mumps fell from a rate of 592 per 100 000 population in 1983 to 123/100 000 in 1986. This decline became progressively more marked in subsequent years, and the incidence of mumps fell by over 90% (figure 1). The incidence rose, however, in 1995-96, as a result of mumps outbreaks largely among unvaccinated adolescents over the age of 14 years, in regions where the Jeryl-Lynn strain had been used.

This increased incidence was associated with a shift in the age of presentation of the disease towards cohorts over 14 years of age – too old to have received the vaccine. Over 50% of cases were adolescents and 25% children aged 5 to 9 years (figure 2). Before the introduction of MMR vaccine, a high percentage of cases had occurred among children under the age of 5 years.

Figure 2 shows that a high percentage of cases in children under 9 years of age occurred among the vaccinated segment, a pattern possibly explained by the high vaccine coverage in such cohorts. From this age upwards an increase is seen in the number of cases among unvaccinated people. This situation illustrates the two types of outbreak currently occurring in Spain - one in late childhood and adolescence, mostly among an unvaccinated population, associated with large numbers of cases, the other in preschool or young schoolchildren affecting a percentage of the vaccinated population, but resulting in fewer cases.

The descriptive epidemiology of the two outbreaks, in Almoradí and Manacor, was studied and a vaccine efficacy study was then carried out. Information used for the descriptive study was obtained from a systematic survey of patients, and completed with information from clinical histories, vaccination registers, school records, vaccination cards, and family surveys.

The case definition contained in the protocols of the national epidemiological surveillance system (8) was adapted to the realities of both outbreaks, with ‘case’ defined as any individual resident in either epidemic territory who presented with acute uni- or bilateral inflammation of the parotid or other gland that lasted two or more days, between September 1998 and March 1999.

The term ‘vaccinated’ was applied to any person for whom information was documented in any of the sources used, the term ‘unvaccinated’ to any person for whom no information was documented in any of the above sources, and the term ‘unknown’ to anyone for whom no information was obtainable.

To estimate the efficacy of the vaccine, a retrospective cohort study was carried out. The following were taken as reference populations: in Almoradí, children born in 1990-7 (n=1018); and in Manacor, schoolchildren born from 1992 and enrolled at the primary school that registered the highest attack rate (n=162). Cohorts were constructed according to the respective time periods during which the different strains of vaccine had been used in each region and were assigned a reference population that, on the basis of its date of birth, had been scheduled for vaccination in those periods. All subjects in the respective cohorts were individually classified as case or non-case, and vaccination status and date were ascertained from vaccination cards, clinical histories, vaccination registers, and school records.

On the basis of this information, three cohorts were defined in Almoradí: a Urabe/Jeryl-Lynn cohort comprising 328 children, a Rubini cohort comprising 422 children, and a Jeryl-Lynn cohort comprising 268 children. In the case of Manacor, a Rubini cohort, based on a school population of 124 children, was considered.

The measure of vaccine efficacy (VE) was calculated on the basis of relative risk (RR), by applying the formula, VE=1-RR (9).

Both epidemics began in late September 1998 and lasted over seven months (cases were enrolled in the study until April 1999, although neither outbreak had ended by that time). The attack rates were 3% in Almoradí and 1% in Manacor.

The two epidemics were similar in the presentation of cases in time, age of people affected, and vaccination status. Spread took place through day care centres and schools, which acted as the true nodal centres, without any indication of clustering by home or neighbourhood.

The most notable feature of the two outbreaks was their distribution by age – children aged 2 to 5 years were affected, although 93% of cases were vaccinated. In Almoradí, the highest rates were in children aged 4 to 5 years, whereas in Manacor the population most affected was children aged 2 to 4 years (table 1).

Table 1. Epidemic outbreaks of mumps in Almoradí and Manacor. Attack rate (%) by age

AGE

ALMORADÍ

MANACOR

< 16 months

0

0.3

16 to 24 months

0

3

2 years

0.7

6

3 years

2

15

4 years

23

7

5 years

22

4

6 to 9 years

5

3

10 to 14 years

1

1

> 14 years

0.03

0.02

TOTAL

1

0.6

In Almoradí, the results obtained for the total study population showed a relative risk (RR) of 0.85 for vaccinated compared with unvaccinated subjects and a vaccine efficacy of 15%; the efficacy found for the cohort vaccinated with the Urabe/Jeryl-Lynn strain was 85%. For children included in the Rubini cohort, however, no significant difference in incidence was observed between vaccinated and unvaccinated subjects (RR=4,95% confidence interval 0.6 to 27).

In Manacor, the study was restricted to the school, with a relative risk of 0.67 and a vaccine efficacy of 33% being obtained for the total study population. No significant differences were observed between the group vaccinated with the Rubini strain and the unvaccinated population, with the vaccine efficacy for the Rubini vaccinated cohort being 40% (table 2).

Vaccinated cohorts 

Cases

Population

 Attack rate (%)

Relative risk 
(CI 95%)

Vaccine efficacy (%) (CI 95%)

Total population 

106*

1018*

10

ALMORADÍ

Vaccinated

100

964

10

0.85 (0.4:2)

15 (-97:63)

Unvaccinated

5

41

12

URABE/JERYL

16

328*

5

Vaccinated

12

311

4

Unvaccinated

4

16

25

0.15 (0.06:0.4)

85 (57:94)

RUBINI

90

422*

21

Vaccinated

88

395

22

4 (0.6:27)

(neg)

Unvaccinated

1

18

5

JERYL-LYNN

0

268*

Vaccinated

0

258

Unvaccinated

0

7

Total population 

43

162*

27

MANACOR

Vaccinated

41

154

26

0.67 (0.2:2)

33 (-101:78)

Unvaccinated

2

5

40

RUBINI

38

124

31

Vaccinated

36

120

30

0.6 (0.2:2)

40 (-66:78)

Unvaccinated

2

4

50

Most cases of mumps in Spain are aged 15 to 19 years, which is characteristic of a disease targeted by a childhood immunisation programme. The main disease burden thus affects a population group that is too old to have received the vaccine and whose exposure to wild virus has reduced as a result of herd immunity produced by high coverage decreasing circulation of virus.

The outbreaks detected in Almoradí and Manacor are noteworthy because they affected a vaccinated population, aged 2 to 5 years. The expected incidence in this population group should have corresponded to individual instances of vaccine failure, but the high incidence encountered suggests generalised failure of the vaccine administered.

The difference observed between the two regions in the pattern of distribution of cases by age was linked to the respective immunisation programmes in place in the two communities. MMR vaccine containing the Rubini strain was administered in the Valencian region from 1994 to 1945 and in the Balearic Isles from 1994 until the outbreak occurred.

The results of the vaccine efficacy studies conducted in these two outbreaks confirm the low efficacy of mumps vaccine afforded by the Rubini strain, as has become increasingly apparent to the national surveillance system and the findings reported by studies from Switzerland, Portugal, and Italy (1-7).

In 1999, Spain's Ministry of Public Health and Consumer Affairs decided to restrict the use of mumps vaccine containing the Rubini strain to children allergic to one of the components of other vaccines. A further ministerial policy recommendation aims to ensure that all children receive one dose, at least, of MMR vaccine containing a vaccine strain other than Rubini.

1. Matter HC, Cloetta J, Zimmermann H, Sentinella Arbeitsgemeinschaft. Measles, mumps, and rubella: monitoring in Switzerland through a sentinel network, 1986-1994. J Epidemiol Community Health 1995; 49(suppl 1): 4-8.

2. Paccaud MF, Hazeghi P, Bourquin M, Maurer AM, Steiner CA, Seiler AJ, et al. A look back at 2 mumps outbreaks. Soz Präventivmed 1995; 40: 72-9.

3. Gonçalves G, de Arújo A, Montero Cardoso ML. Outbreak of mumps associated with poor vaccine efficacy – Oporto, Portugal, 1996. Eurosurveillance.1998; 3: 115-21.

4. Germann D, Ströhle A, Eggenberger K, Steiner CA, Matter L. An outbreak of mumps in a population partially vaccinated with the Rubini strain. Scand J Infect Dis 1996; 28: 235-8.

5. Toscani L, Batou M, Bouvier P, Schlapfer A. Comparison of the efficacy of various strains of mumps vaccine: a school survey. Soz Präventivmed 1996; 41: 341-7.

6. Chamot E, Toscani L, Egger O, Germann D, Bourquini C. Estimation de l’efficacité de trois souches vaccinales ourliennes au cours d’une épidémie d’oreillons dans le canton de Genêve (Suisse). Rev Epidem et Santé Publ 1998, 46: 100-7.

7. The Benevento and Compobasso Pediatricians Network for the Control of Vaccine-Preventable Diseases. Field evaluation of the clinical effectiveness of vaccines against pertussis, measles, rubella and mumps. Vaccine 1998; 16: 818-22.

8. Centro Nacional Epidemiología. Protocolos de enfermedades de declaración obligatoria. Madrid: Ministerio de Sanidad y Consumo, 1996.

9. Orenstein W, Bernier R, Hinman A. Assessing vaccine efficacy in the field. Further observations. Epidemiol Rev 1998; 10: 212-41.