- Email: [email protected]
Measles antibody levels in a vaccinated population in Brazil
TRANSACTIONS
OFTHEROYALSOCIETYOFTROPICALMEDICINEANDHYGIENE(1998)92,227-230
Measles antibody
levels in a vaccinated
population
227
in Brazil
‘Department of Biological Sciences, UniM. J. Coxl, R. S. Azevedo2, E. Massad2, A. R. Fooks3 and D. J. Nokesl versity of Warwick, Coventr CV4 7AL, UK; 2Faculdade de Medicina da Universidade de Sco Paul0 and LIM-01 HCFMUSFj Silo Paula, Brazil; PCentre for Applied Microbiology and Research (CAMR), Porton Down, Salisbury, Wiltshire, SF’4 OJG, UK Abstract An epidemiological study of measles-specific immunoglobulin G antibody levels was conducted using a representative sample of a vaccinated suburban population in SPo Paulo State, Brazil. The study aimed to determine immunity status in relation to age and infection or vaccination experience. 549 age-structured samples of sera, collected in 1990, were screened and calibrated to the international reference serum, using measles nucleoprotein in an enzyme immunoassay. In the age group with direct experience of vaccination (9 months to 15 years), whether routine or campaign, over 90% had detectable antibody250 miu/mL. However, 14% of these had antibody concentrations between 50 and 100 miu/mL and 30% between 50 and 255 miu/mL. In those over 15 years of age, 94% had antibody levels >255 miuiml, assumed to be the result of past infection. The study suggested that, within highly vaccinated populations, a proportion of individuals had measles antibody levels which may be insufficient to protect against reinfection or clinical disease. The implications of these results, and similar findings elsewhere, in relation to the persistence of measles requires investigation; this has particular relevance in SHo Paulo following the recent measles outbreak. Keywords:
measles, immunization, seroprevalence, Brazil
Introduction Seroepidemiological studies reporting age-specific prevalence of antibodies to specific viral infections are increasingly being used in infectious disease surveillance, such as measuring the impact of vaccination on population levels of immunity or, more specifically, identifying age groups with low levels of antibody from which to assess the risk of infection outbreak (E&BAD et al., 1995; GAY et al., 1997a, 1997b; MILLER, 1997). Furthermore, such data yield key information by means of which to summarize the epidemiology of infectious diseases in host populations, such as the basic reproductive number, R,,, the per-person force of infection, and the level of vaccine-induced immunity required to gen(ANDERSON & MAY, 1991; erate herd protection AZEVEDO NETO et al., 1994). There is an increasing body of evidence, however, to suggest that seropositivity (the presence of antibody above some defined cut-off level) does not necessarily signify protective immunity and alone may not provide sufficient information to characterize the status of a population’s immunity and epidemiology. Instead, studies indicate that the level of existing antibody is an important correlate of the outcome in terms of infection, clinical disease or a booster response following re-exposure by vaccination or infection. In a report of a measles outbreak in the USA (CHEN et al., 1990), individuals with a pre-exposure antib,ody titre < 1: 120, using the plaque reduction neutralization (PRNT) test, were found to be susceptible to clinical disease, although infection did occur at even higher pre-exposure titres. A recent study in the UK ~CALVERT et al.. 1996) found a booster resnonse in 75% of previously vaccinated schoolchildren wiih low levels of antibody. These levels (50-100 miulml, calibrated using the international antimeasles reference semm [FORSEY et al., 19911) have been used as a threshold of clinical protection for surveillance of the UK immunization campaign (GAY et al., 1997a). A study during a measles vaccine trial in Senegal reported high infection rates in individuals with titres between 40 and 125 miu/mL, although a proportion of seronegative but vaccinated individuals were protected (SAMB et al., 1995). The imnlication of these studies mav be imnortant h that s&optimal protection against clinicalhisease or reversion to susceptibility following immunization could result in the failure of a vaccination programme. They may also provide a mechanism for the persistence of viral transmission in highly vaccinated populations. Further studies which characterize quantitatively the
age-specific antibody profile of measles in vaccinated communities would be of use in establishing the potential significance of low-level immunity to infection persistence, to clinical disease distribution, and to vaccination policy. In this study we used a serum collection made in 1990 from randomly selected individuals representative of a highly vaccinated suburban communitv, to relate immunitv levels as defined bv antibodv titre to age and vaccination or infection experience. Measles immunization in Brazil began in the late 1960s. became part of the National Immumzation Program& in 1973, and has been compulsory since 1977. Vaccination coverage in 1986 was about 65% (unpublished report, Ministry of Health, Sao Paulo State, 1987). In 1987, a mass programme was carried out in the greater metropolitan area of Sao Paulo, reaching 86% of children aged 9 months to 14 years (PANNUTI et al., 1991). Subsequently, in 1992, a major vaccination strategy against measles-mumps-rubella was introduced, including a ‘pulse’ immunization campaign aimed at all children aged l-10 years followed by their inclusion in the immumzation calendar at 15 months (~&ASSAD et al., 1995). However, according to the Centre of Eoidemiological Surveillauce of the-State of Sao Paul0 Ipersonal communication) over 12 000 confirmed measles cases were reported in the first 10 months of 1997, mostlv in the citv of Slo Paulo. with about 15% in children under 5 years old, 25% in those aged 5-20 years, and 50% in those 20-30 years old. Materials and Methods Collection of sera The study used samples collected in 1990 from the town of Caieiras, in northern Slo Paulo State, Brazil, for the investigation of rubella. The collection of sera as well as demographic details and methods used in sampling the community have been described previously (AZEVEDO NETO et al., 1994). The study population was chosen as representative of a suburban Brazilian community. Standard theory (COCHRAN, 1977) was applied to achieve a random two-level cluster sample from families within randomly selected administrative regions. The age structure of the samples used in this study is shown in Table 1. These included 47 1 sera from all age groups in the initial community survey and 39 maternal-cord blood uairs collected from Hosaital Regimul de Caieiras (EMED). Sera were stored ai -20°C. No information regarding measles vaccination status was available for individuals.
M. J. COX ETAL.
228
Table. Details of serum samples collected in December 1990, including the maternal-cord blood pairs Age class
No. of sera
Months 0a l-2 3-4 5-6 7-8 .9-10 11-12 13-18 19-23 Years 2 i 5 6-10 11-15 16-20 21-30 31-40 Total
39 27 22 27 48 37 27 29 25 50 44 33 14 35 12 20 43 17 549
protective immunity: (i) 255 miu/mL, estimated for a PRNT titre of 1:120 (RATNAM et al., 1995), reported as the minimum antibody level (titre) conferring protection from clinical disease (CHEN et al., 1990), (ii) 100 miu/mL and (iii) 50 miu/mL, the range for a threshold of clinical protection which has been used previously (GAY et al., 1997a). Other statistical analysis was carried out using the statistical program SPSS@ release 6.0 (1993). Results The coefficients of variation for the 4 internal control sera in the EIA ranged from 1.7% to 6.3% and there was significant homogeneity of variance between the controls (~~‘9.2, df=9, PcO.05). There was a significant correlation between antibody titres measured using EL4 and HI (n=50, ~0.83, P~O.0001) (Fig. 1).
aCord blood samples. Nucleoprotein enzyme immunoassay (EIA) Microtitre plates (96 wells; Dynatech, Immulon II@) were coated with recombinant measles nucleoprotein expressed in insect cells using a baculovirus expression system (FOOKS et al., 1993) as antigen in carbonate buffer (pH 9.6) overnight at 4°C. Optimal dilution of antigen (c. 3.3 WmL) was determined by chequerboard titration. Plates were then blocked with 2% skimmed milk powder (SMP) in phosphate-buffered saline (PBS) for 1 h at room temperature (RT). Test sera were added at l/200 in SMP-PBS+ 0.05% Tween 2@ (Sigma) for 2 h at RT. Plates were washed 3 times in PBS-Tween 20@, after which conjugate (rabbit anti-human immunoglobulin G (IgG)-horse-radish peroxidase, Dako) was added at l/2000 for a further 2 h at RT. Plates were developed using o-phenylenediamine substrate in phosphate-citrate buffer (pH 5) for 30 min and read at 490 nm. Haemagglutination inhibition test (HI) In order to investigate whether antibody measured using nucleoprotein correlated with functional antimeasles antibody, 50 sera, with a range of antibody titres as determined by EIA using nucleoprotein, were screened by the measles HI. The method was based on that previously described by NORRBY (1962) and used rhesus-positive red blood cells and 4 haemagglutinating units of measles antigen. Data analysis All EIA plates contained a local standard of high titre measles IszG antibodv in serial dilution for intra- and inter-plate standard&ation. For both EL4 and HI, the local standard serum was calibrated against the international antimeasles serum (5 international units [iu], FORSEY et aZ., 1991) and used to quantify results. All test sera results are presented as loglo miu/mL or geometric mean titres of seropositive persons (those with measles antibody levels 250 miuiml). Four internal control sera of low to high titre were included in all EIA plates to monitor consistency and tested using Bartlett’s (SOKAL & ROHLF, test of homogeneity of variance 1981). Three cut-off antibody concentrations were used to demonstrate age-specific serological status of individuals within the population. These were based on different published sensitivity points or levels associated with
1.5
2 Nucleoprotein
2.5 3 immunoassay
3.5 4 titres (log,, miulml)
4.5
Fig. 1. Correlation between measles antibody titres in 50 Brazilian sera, as measured using recombinant measlesnucleoprotein in an enzyme immunoassay and measles haemagglutination inhibition test. Scatter plots illustrating individual antibody concentrations for all samples (n=549, Table) are shown in Fig. 2 and summarized as age-prevalence profiles in Fig. 4. Geometric mean titres of those persons with antibody levels 250 miu.mL are presented in Fig. 3. There was a clear decline in maternally-derived antibody from birth over the range O-9 months (Fig. 2 B, C). The titres of antibody in cord blood sera were very high and remained high in the first 2 months after birth, with levels 2255 miu/mL in over 95% of individuals. Thereafter, titres rapidly declined so that only 40% exceeded 255 miu/mL after 4 months and only 5% did so at 6 months, even though much lower levels of antibody persisted. From 6 months onwards through childhood an increasing proportion of subjects showed evidence of high titre antibody above one or more of the cut-off levels (Figs 2-4) reportedly associated with clinical immunity and assumed to result from infection or vaccination. In individuals over the age of 20 years (240 months) there was little change with age in the proportion with tin-es above 255 miuiml. Antibody titres in cord sera (mean = 3.36 loglo miu/mL, SD = 0.4) were highly correlated with titres in their respective maternal pairs (mean = 3.26 loglo miu/mL, SD = 0.45) (39 pairs examined; T = O-964, PCO.0001); no significant difference was detected between the mean titres. Prevalence of detectable antibody in adults was similar to that in infants (O-4 months) and children over one year of age. However, antibody titres in children aged 218 months and Cl5 years were lower: only 70% had titres greater than 255 miu/mL. These age classes correlated with the probable influence of immunization: seroconversion in those older than 15-20 years would almost certainly have been due to natural infection, whereas seroconversion in younger children would, to a greater or lesser extent (probably correlating with age) be due to imunization. A ‘window’ of lower antibody titres clearly existed in children aged 5-10 years where, although
MEASLES SEROLOGYIN BRAZIL
229
2.5
I 5
14: 0
:
1: 10
I : I 15 20
:
I 30
I: 25
:
/ 35
:I 40
Age (years)
Age class
Fig. 3. Age-specific geometric mean measles antibody titres with 95% confidence intervals (horizontal lines) of individuals with 50 miu/mL of measles antibody. Age classes(months [m] and years [y] are as given in the Table.
1 0.8
.g
c ::
.
0.6 -
g 0.4
0.2
i-8; -2m
:* :a
3.5
l
.
: -
-
.
3
!.
.
-I
.
:
! .
2.5
55'
l .
:-
1.5
.
1 14: 0
I: 1
/::: 2 3
.
.
!
i
50
.
.
.
-
l
. :
-
100
2
.
. . .
.
. -8
0
.
8 !
.
-! I::! 4 5
!
. ;
I
:-j@
. i
:
7 I I I: 6 7
j
:
0’ j
f
.
I::::: 8 9
/ II 101112
Age (months) Fig. 2. Scatter plots of measles antibody enzyme immunoassay tirres for all 549 samples showing (A) all ages, (B) those aged O-4 years, and (C) those aged O-12 months. The horizontal lines on each graph represent published cut-off values (miui mL) for possible thresholds of immunity to reinfection and clinical disease (see text). prevalence was high, there was a proportion below 255 miuiml.
Ii-i: 9.10m
1 3-18r Age
ClaSS
Fig. 4. Age-prevalence profiles of the proportion of individuals whose antibody tines were between the different immunity thresholds. Solid bars, 2255 miuiml; shaded bars, loo-254 miu/mL; open bars, 50-99 miuiml. Age classes (monms[m] and years [y] are as given in the Table.
.
-
.
l
5.6m
with titres
Discussion The objective of this study was to determine the immune status to measles of a randomly selected and representative sample of a population during the course of an immunization programme and thus highlight potential age-specific risk of susceptibility to reinfection or disease, based on the titres of antimeasles specific immunoglobulin G (IgG). The PRNT is considered to be the most sensitive assay available for the detection of antibodies to measles virus (ALBRRCHT et al., 1981) and, like HI, it measures functional antibody. However, neither HI nor, especial-
ly, the PRNT are practical for large-scale seroepidemiological studies. Indirect EIA for IgG has previously been shown to correlate well with neutralizing antibody (ERDMAN et al., 1991) and its sensitivity was improved by the use of recombinant measles nucleoprotein as antigen (HUMMEL et al., 1992). In our study, recombinant measles nucleoprotein proved a suitable antigen in an EIA for population-based epidemiological studies. A significant correlation between antibody measured using the antigen and that measured in the same sera by HI suggested that the EIA tines can be used to reflect functional antimeasles antibody, supporting previous findings (WARNES et al., 1994). Further evaluation and comparison with the PFWT is currently underway. At the time of the study, a significant proportion of the population, mainly those at the age influenced by measles immunization (9 months-15 years), had antibody titres which may not protect against infection and clinical disease. Antibody titres in adults, who were very unlikely to have been immunized before natural infection, were significantly higher that those in the vaccinated age group. These results are consistent with studies which showed either that vaccine-induced measles antibody levels were lower than those naturally acquired (e.g., CHRISTENSON & BOTTIGER 1994) or that the decline in immunity was rapid following immunization. Our results also support the recent observation in Canada that 28% of vaccinated schoolchildren had PBNT titres less than 1: 120 (BATMAN et aZ., 1996). Detailed interpretation of age-specific seroprevalence with respect to known immunization policy changes is
M. J. COX
230
difficult. The ‘window’ of lower antibody titres in 5-l 0 years old children may be the result of low immunization coverage in the Caieiras area during previous years or during the mass immunization programme of 1987. Whilst it is not possible confidently to relate the recent increase in measles transmission in SPo Paulo to our observations, the group of individuals with low titres identified in the survev of 1990 would have been vounp: adults in 1997 and”it was within this age group that at: tack rates were high. Previous epidemiological and theoretical studies have demonstrated that a proportion of susceptible young adults (those uninfected at the start of the programme who may be too old to have received vaccine) can be generated when mass immunization is introduced to a population as a result of reduced virus circulation (ANDERSON & MAY, 1991; UKKONEN, 1996).
The question arises whether the virus is transmitted by individuals with lower immunity status, such as those who have been vaccinated, and in what way this may affect the population dynamics of measles transmission. Subclinical virus transmission within vaccinated communities has been reported by PEDERSEN et al. (1989). This studv highlighted the imuortance of defining quantitative6 lev
ETAL.
clonal antibody-based capture enzyme immunoassays for detection of specific antibodies to measles virus. Journal of Clinical Microbiologv. 29, 1466-147 1. Fooks, A. R., Stephenson,J. R., Warnes, A., Dowsett, B.A., Rima, B. K. &Wilkinson, G. (1993). Measles virus nucleocapsid protein expressed in insect cells assembles into nucleocapsid-like structures. Journal of General Virology, 74, 1439-1444. Forsey,T, Heath, A. B. & Minor, I? D. (1991). The first international standard for anti-measles serum. Biologicals. Y _ 19.I 237-241. Gay, N., Ramsey, M., Cohen, B., Hesketh, L., Morgan-Capner, I’., Brown, D. & Miller, E. (1997a). The epidemiology of measles in England and Wales since the 1994 vaccination campaign. Communicable Disease Report CDR Review, 7, R17-R21.
Gay, N., Miller, E., Hesketh, L., Morgan-Capner, I’., Ramsey, M., Cohen, B. & Brown, D. (1997b). Mumps surveillance in England and Wales supports introduction of two dose vaccination schedule. CommunicableDisease Report CDR Review. 7, R2LR26.
Hummel, K. B., Erdman, D. D., Heath, J. & Bellini, W. J. (1992). Baculovirus expression of the nucleoprotein gene of measles virus and utility of the recombinant protein in diagnostic enzyme immunoassays. Journal of Clinical Microbiology, 30, 2874-2880.
Massad, E., Azevedo-Neto, R. S., Burattini, M. N., Zanetta, D. M. T., Coutinho, F. A. B.,Yang, H. M., Moraes, J. C., Pannuti, C. S., Souza,V. A. U. F., Silveira, A. S. B., Struchiner, C. J., Oselka, G. W., Camargo, M. C. C., Omoto,T M. & Passos.S. D. 11995).Assessinathe efficacv of a mixed vaccination strategy against rubellg in SHo Paula, Brazil. Znarnational .@urnal of Epidemiology, 24, 842-850.
Miller, E.,Waight, E-Gay, N.,Ramsey, M., Vurdien, J., Morgan-Capner, I?, Hesketh, L., Brown, D.,Tookey, P. & Peckham, C. (1997) .The epidemiology of rubella in England and Wales before and after the 1994 measles and rubella vaccination campaign: fourth joint report from the PHLS and the National Congenital Rubella Surveillance Programme. Communicable Disease Report CDR Review, 7, R26-R32.
Acknowledgements
This work was supported by the Royal Society and FAPESP. R.S.A. and E.M. are supported by CNPq. D.J.N. is a Royal Society University Research Fellow. We thank D. Brown for comments on the text, B. Cohen for supply of measles HI antigen and W. J. Edmunds for helpful discussion. References
Albrecht, I’., Herman?, I<. & Burns, G. R. (1981). Role of virus strain in conventional and enhanced measlesplaque neutralisation test. Journal of Virological Methods, 3,25 l-260. Anderson, R. M. & May, R. M. (1991). Infectious Diseases of Humans: Dynamics and Control. Oxford: Oxford University Press.
-
Azevedo Neto, R. S., Silveira, A. S. B., Nokes, D. J.,Yang, H. M., Passos,S. D., Cardoso, M. R. A. & Massad, E. (1994). Rubella seroepidemiology in a non-immunized population of Sao Paulo State, Brazil. Epidemiology and Infection, 113, 161-173.
Babad, H. R., Nokes, D. J., Gay, N. J., Miller, E., MorganCapner, I? &Anderson, R. M. (1995). Predicting the impact of measles vaccination in England and Wales: model validation and analysis of policy options. Epidemiology and Infection, 114,319-344. Calvert, N., Cutts, F. T., Irving, R., Brown, D., Marsh, J. & Miller, E. (1996). Measles immunity and response to revaccination among secondary school children in Cumbria. Epidemiology and Infection, 116, 65-70.
Chen, R. T., Markowitz, L. E., Albrecht, I’., Stewart, J. A., Mofenson, L. M., Preblud, S. R. & Orenstein, W. A. (1990). Measles antibody: reevaluation of protective titres. JournaZof Infectious Diseases, 162, 1036-1042.
Christenson, B. & Bottiger, M. (1994). Measles antibodv-comoarison of lona-term vaccination titers. earlv vaccination tit&s and naturally acquired-immunity to and booster effects on the measles-virus. Kzccine,12, 129-133. Co&an, W. G. (1977). Sampling Techniques, 3rd edition. New York: John Wiley & Son. Erdman, D. D., Anderson, L. J., Adamms, D., Stewart, J., Markowitz, L. & Bellini, W. J. (1991). Evaluation of mono-
Norrby, E. (1962). Haemagglutination by measles virus. 4. A simple procedure for production of high potency antigen for haemagglutination-inhibition (HI) tests. Proceedings of the 111, Society for Experimental Biology and Medicine, 8114-8118.
Pannuti, C. S., Moraes, J. C., Souza, V., Camargo, M. & Hidalgo, N. (199 1). Measles antibody prevalence after mass immunization in Slo Paulo. Bulletin of the World Health Orsanitation, 69.557-560. Pedersen, I.‘R., Modrhorst, C. H., Glikmann, G. & von Magnus, H. (1989). Subclinical measles infection in vaccinated se&positive individuals in arctic Greenland. ‘Vaccine,7, 345-348.
Ramam, S., Gadag,V., West, R., Burris, J., Oates, E., Stead, F. & Bouillane, N. (1995). Comparison of commercial enzyme immunoassay kits with plaque reduction neutralization-test for detection of measles virus antibody. Journal of Clinical Microbiology, 33, 811-815.
Ramam, S., West, R., Gadag, V., Williams, B. & Oates, E. (1996). Immunitv aaainst measles in school-aaed children: implications for measles revaccination strategi&.. Canadian .yournal of Public Health, 87, 407-410. Samb, B., Aaby, P.,Whittle, H. C., Seek, A., Rahman, S., Bennett, J., Markowitz, L. & Simondon, F. (1995). Serologic status and measles attack rates among vaccinated and unvaccinated children in rural Senegal. Pediatric Infectious Disease Journal, 14,203-209. Sokal, R. R. & Rohlf, F. J. (1981). Biometry, 2nd edition. New York: W. H. Freeman. Ukkonen, P. (1996). Rubella immunity and morbidity: impact of different vaccination programs in Finland 1979-1992. Scandinavian .yournal of InfectiousDiseases, 283 l-35. Warnes, A., Fooks, A. R.-& Stephenson, J. R. (1994). Production of measles nucleoprotein in different expression systems and its use as a diagnostic reagent.Journal ofVirologicalMethads, 49,257-268. Received 27 August cepted for publication
1997; revised 17 November 18 November 1997
1997; ac-
OFTHEROYALSOCIETYOFTROPICALMEDICINEANDHYGIENE(1998)92,227-230
Measles antibody
levels in a vaccinated
population
227
in Brazil
‘Department of Biological Sciences, UniM. J. Coxl, R. S. Azevedo2, E. Massad2, A. R. Fooks3 and D. J. Nokesl versity of Warwick, Coventr CV4 7AL, UK; 2Faculdade de Medicina da Universidade de Sco Paul0 and LIM-01 HCFMUSFj Silo Paula, Brazil; PCentre for Applied Microbiology and Research (CAMR), Porton Down, Salisbury, Wiltshire, SF’4 OJG, UK Abstract An epidemiological study of measles-specific immunoglobulin G antibody levels was conducted using a representative sample of a vaccinated suburban population in SPo Paulo State, Brazil. The study aimed to determine immunity status in relation to age and infection or vaccination experience. 549 age-structured samples of sera, collected in 1990, were screened and calibrated to the international reference serum, using measles nucleoprotein in an enzyme immunoassay. In the age group with direct experience of vaccination (9 months to 15 years), whether routine or campaign, over 90% had detectable antibody250 miu/mL. However, 14% of these had antibody concentrations between 50 and 100 miu/mL and 30% between 50 and 255 miu/mL. In those over 15 years of age, 94% had antibody levels >255 miuiml, assumed to be the result of past infection. The study suggested that, within highly vaccinated populations, a proportion of individuals had measles antibody levels which may be insufficient to protect against reinfection or clinical disease. The implications of these results, and similar findings elsewhere, in relation to the persistence of measles requires investigation; this has particular relevance in SHo Paulo following the recent measles outbreak. Keywords:
measles, immunization, seroprevalence, Brazil
Introduction Seroepidemiological studies reporting age-specific prevalence of antibodies to specific viral infections are increasingly being used in infectious disease surveillance, such as measuring the impact of vaccination on population levels of immunity or, more specifically, identifying age groups with low levels of antibody from which to assess the risk of infection outbreak (E&BAD et al., 1995; GAY et al., 1997a, 1997b; MILLER, 1997). Furthermore, such data yield key information by means of which to summarize the epidemiology of infectious diseases in host populations, such as the basic reproductive number, R,,, the per-person force of infection, and the level of vaccine-induced immunity required to gen(ANDERSON & MAY, 1991; erate herd protection AZEVEDO NETO et al., 1994). There is an increasing body of evidence, however, to suggest that seropositivity (the presence of antibody above some defined cut-off level) does not necessarily signify protective immunity and alone may not provide sufficient information to characterize the status of a population’s immunity and epidemiology. Instead, studies indicate that the level of existing antibody is an important correlate of the outcome in terms of infection, clinical disease or a booster response following re-exposure by vaccination or infection. In a report of a measles outbreak in the USA (CHEN et al., 1990), individuals with a pre-exposure antib,ody titre < 1: 120, using the plaque reduction neutralization (PRNT) test, were found to be susceptible to clinical disease, although infection did occur at even higher pre-exposure titres. A recent study in the UK ~CALVERT et al.. 1996) found a booster resnonse in 75% of previously vaccinated schoolchildren wiih low levels of antibody. These levels (50-100 miulml, calibrated using the international antimeasles reference semm [FORSEY et al., 19911) have been used as a threshold of clinical protection for surveillance of the UK immunization campaign (GAY et al., 1997a). A study during a measles vaccine trial in Senegal reported high infection rates in individuals with titres between 40 and 125 miu/mL, although a proportion of seronegative but vaccinated individuals were protected (SAMB et al., 1995). The imnlication of these studies mav be imnortant h that s&optimal protection against clinicalhisease or reversion to susceptibility following immunization could result in the failure of a vaccination programme. They may also provide a mechanism for the persistence of viral transmission in highly vaccinated populations. Further studies which characterize quantitatively the
age-specific antibody profile of measles in vaccinated communities would be of use in establishing the potential significance of low-level immunity to infection persistence, to clinical disease distribution, and to vaccination policy. In this study we used a serum collection made in 1990 from randomly selected individuals representative of a highly vaccinated suburban communitv, to relate immunitv levels as defined bv antibodv titre to age and vaccination or infection experience. Measles immunization in Brazil began in the late 1960s. became part of the National Immumzation Program& in 1973, and has been compulsory since 1977. Vaccination coverage in 1986 was about 65% (unpublished report, Ministry of Health, Sao Paulo State, 1987). In 1987, a mass programme was carried out in the greater metropolitan area of Sao Paulo, reaching 86% of children aged 9 months to 14 years (PANNUTI et al., 1991). Subsequently, in 1992, a major vaccination strategy against measles-mumps-rubella was introduced, including a ‘pulse’ immunization campaign aimed at all children aged l-10 years followed by their inclusion in the immumzation calendar at 15 months (~&ASSAD et al., 1995). However, according to the Centre of Eoidemiological Surveillauce of the-State of Sao Paul0 Ipersonal communication) over 12 000 confirmed measles cases were reported in the first 10 months of 1997, mostlv in the citv of Slo Paulo. with about 15% in children under 5 years old, 25% in those aged 5-20 years, and 50% in those 20-30 years old. Materials and Methods Collection of sera The study used samples collected in 1990 from the town of Caieiras, in northern Slo Paulo State, Brazil, for the investigation of rubella. The collection of sera as well as demographic details and methods used in sampling the community have been described previously (AZEVEDO NETO et al., 1994). The study population was chosen as representative of a suburban Brazilian community. Standard theory (COCHRAN, 1977) was applied to achieve a random two-level cluster sample from families within randomly selected administrative regions. The age structure of the samples used in this study is shown in Table 1. These included 47 1 sera from all age groups in the initial community survey and 39 maternal-cord blood uairs collected from Hosaital Regimul de Caieiras (EMED). Sera were stored ai -20°C. No information regarding measles vaccination status was available for individuals.
M. J. COX ETAL.
228
Table. Details of serum samples collected in December 1990, including the maternal-cord blood pairs Age class
No. of sera
Months 0a l-2 3-4 5-6 7-8 .9-10 11-12 13-18 19-23 Years 2 i 5 6-10 11-15 16-20 21-30 31-40 Total
39 27 22 27 48 37 27 29 25 50 44 33 14 35 12 20 43 17 549
protective immunity: (i) 255 miu/mL, estimated for a PRNT titre of 1:120 (RATNAM et al., 1995), reported as the minimum antibody level (titre) conferring protection from clinical disease (CHEN et al., 1990), (ii) 100 miu/mL and (iii) 50 miu/mL, the range for a threshold of clinical protection which has been used previously (GAY et al., 1997a). Other statistical analysis was carried out using the statistical program SPSS@ release 6.0 (1993). Results The coefficients of variation for the 4 internal control sera in the EIA ranged from 1.7% to 6.3% and there was significant homogeneity of variance between the controls (~~‘9.2, df=9, PcO.05). There was a significant correlation between antibody titres measured using EL4 and HI (n=50, ~0.83, P~O.0001) (Fig. 1).
aCord blood samples. Nucleoprotein enzyme immunoassay (EIA) Microtitre plates (96 wells; Dynatech, Immulon II@) were coated with recombinant measles nucleoprotein expressed in insect cells using a baculovirus expression system (FOOKS et al., 1993) as antigen in carbonate buffer (pH 9.6) overnight at 4°C. Optimal dilution of antigen (c. 3.3 WmL) was determined by chequerboard titration. Plates were then blocked with 2% skimmed milk powder (SMP) in phosphate-buffered saline (PBS) for 1 h at room temperature (RT). Test sera were added at l/200 in SMP-PBS+ 0.05% Tween 2@ (Sigma) for 2 h at RT. Plates were washed 3 times in PBS-Tween 20@, after which conjugate (rabbit anti-human immunoglobulin G (IgG)-horse-radish peroxidase, Dako) was added at l/2000 for a further 2 h at RT. Plates were developed using o-phenylenediamine substrate in phosphate-citrate buffer (pH 5) for 30 min and read at 490 nm. Haemagglutination inhibition test (HI) In order to investigate whether antibody measured using nucleoprotein correlated with functional antimeasles antibody, 50 sera, with a range of antibody titres as determined by EIA using nucleoprotein, were screened by the measles HI. The method was based on that previously described by NORRBY (1962) and used rhesus-positive red blood cells and 4 haemagglutinating units of measles antigen. Data analysis All EIA plates contained a local standard of high titre measles IszG antibodv in serial dilution for intra- and inter-plate standard&ation. For both EL4 and HI, the local standard serum was calibrated against the international antimeasles serum (5 international units [iu], FORSEY et aZ., 1991) and used to quantify results. All test sera results are presented as loglo miu/mL or geometric mean titres of seropositive persons (those with measles antibody levels 250 miuiml). Four internal control sera of low to high titre were included in all EIA plates to monitor consistency and tested using Bartlett’s (SOKAL & ROHLF, test of homogeneity of variance 1981). Three cut-off antibody concentrations were used to demonstrate age-specific serological status of individuals within the population. These were based on different published sensitivity points or levels associated with
1.5
2 Nucleoprotein
2.5 3 immunoassay
3.5 4 titres (log,, miulml)
4.5
Fig. 1. Correlation between measles antibody titres in 50 Brazilian sera, as measured using recombinant measlesnucleoprotein in an enzyme immunoassay and measles haemagglutination inhibition test. Scatter plots illustrating individual antibody concentrations for all samples (n=549, Table) are shown in Fig. 2 and summarized as age-prevalence profiles in Fig. 4. Geometric mean titres of those persons with antibody levels 250 miu.mL are presented in Fig. 3. There was a clear decline in maternally-derived antibody from birth over the range O-9 months (Fig. 2 B, C). The titres of antibody in cord blood sera were very high and remained high in the first 2 months after birth, with levels 2255 miu/mL in over 95% of individuals. Thereafter, titres rapidly declined so that only 40% exceeded 255 miu/mL after 4 months and only 5% did so at 6 months, even though much lower levels of antibody persisted. From 6 months onwards through childhood an increasing proportion of subjects showed evidence of high titre antibody above one or more of the cut-off levels (Figs 2-4) reportedly associated with clinical immunity and assumed to result from infection or vaccination. In individuals over the age of 20 years (240 months) there was little change with age in the proportion with tin-es above 255 miuiml. Antibody titres in cord sera (mean = 3.36 loglo miu/mL, SD = 0.4) were highly correlated with titres in their respective maternal pairs (mean = 3.26 loglo miu/mL, SD = 0.45) (39 pairs examined; T = O-964, PCO.0001); no significant difference was detected between the mean titres. Prevalence of detectable antibody in adults was similar to that in infants (O-4 months) and children over one year of age. However, antibody titres in children aged 218 months and Cl5 years were lower: only 70% had titres greater than 255 miu/mL. These age classes correlated with the probable influence of immunization: seroconversion in those older than 15-20 years would almost certainly have been due to natural infection, whereas seroconversion in younger children would, to a greater or lesser extent (probably correlating with age) be due to imunization. A ‘window’ of lower antibody titres clearly existed in children aged 5-10 years where, although
MEASLES SEROLOGYIN BRAZIL
229
2.5
I 5
14: 0
:
1: 10
I : I 15 20
:
I 30
I: 25
:
/ 35
:I 40
Age (years)
Age class
Fig. 3. Age-specific geometric mean measles antibody titres with 95% confidence intervals (horizontal lines) of individuals with 50 miu/mL of measles antibody. Age classes(months [m] and years [y] are as given in the Table.
1 0.8
.g
c ::
.
0.6 -
g 0.4
0.2
i-8; -2m
:* :a
3.5
l
.
: -
-
.
3
!.
.
-I
.
:
! .
2.5
55'
l .
:-
1.5
.
1 14: 0
I: 1
/::: 2 3
.
.
!
i
50
.
.
.
-
l
. :
-
100
2
.
. . .
.
. -8
0
.
8 !
.
-! I::! 4 5
!
. ;
I
:-j@
. i
:
7 I I I: 6 7
j
:
0’ j
f
.
I::::: 8 9
/ II 101112
Age (months) Fig. 2. Scatter plots of measles antibody enzyme immunoassay tirres for all 549 samples showing (A) all ages, (B) those aged O-4 years, and (C) those aged O-12 months. The horizontal lines on each graph represent published cut-off values (miui mL) for possible thresholds of immunity to reinfection and clinical disease (see text). prevalence was high, there was a proportion below 255 miuiml.
Ii-i: 9.10m
1 3-18r Age
ClaSS
Fig. 4. Age-prevalence profiles of the proportion of individuals whose antibody tines were between the different immunity thresholds. Solid bars, 2255 miuiml; shaded bars, loo-254 miu/mL; open bars, 50-99 miuiml. Age classes (monms[m] and years [y] are as given in the Table.
.
-
.
l
5.6m
with titres
Discussion The objective of this study was to determine the immune status to measles of a randomly selected and representative sample of a population during the course of an immunization programme and thus highlight potential age-specific risk of susceptibility to reinfection or disease, based on the titres of antimeasles specific immunoglobulin G (IgG). The PRNT is considered to be the most sensitive assay available for the detection of antibodies to measles virus (ALBRRCHT et al., 1981) and, like HI, it measures functional antibody. However, neither HI nor, especial-
ly, the PRNT are practical for large-scale seroepidemiological studies. Indirect EIA for IgG has previously been shown to correlate well with neutralizing antibody (ERDMAN et al., 1991) and its sensitivity was improved by the use of recombinant measles nucleoprotein as antigen (HUMMEL et al., 1992). In our study, recombinant measles nucleoprotein proved a suitable antigen in an EIA for population-based epidemiological studies. A significant correlation between antibody measured using the antigen and that measured in the same sera by HI suggested that the EIA tines can be used to reflect functional antimeasles antibody, supporting previous findings (WARNES et al., 1994). Further evaluation and comparison with the PFWT is currently underway. At the time of the study, a significant proportion of the population, mainly those at the age influenced by measles immunization (9 months-15 years), had antibody titres which may not protect against infection and clinical disease. Antibody titres in adults, who were very unlikely to have been immunized before natural infection, were significantly higher that those in the vaccinated age group. These results are consistent with studies which showed either that vaccine-induced measles antibody levels were lower than those naturally acquired (e.g., CHRISTENSON & BOTTIGER 1994) or that the decline in immunity was rapid following immunization. Our results also support the recent observation in Canada that 28% of vaccinated schoolchildren had PBNT titres less than 1: 120 (BATMAN et aZ., 1996). Detailed interpretation of age-specific seroprevalence with respect to known immunization policy changes is
M. J. COX
230
difficult. The ‘window’ of lower antibody titres in 5-l 0 years old children may be the result of low immunization coverage in the Caieiras area during previous years or during the mass immunization programme of 1987. Whilst it is not possible confidently to relate the recent increase in measles transmission in SPo Paulo to our observations, the group of individuals with low titres identified in the survev of 1990 would have been vounp: adults in 1997 and”it was within this age group that at: tack rates were high. Previous epidemiological and theoretical studies have demonstrated that a proportion of susceptible young adults (those uninfected at the start of the programme who may be too old to have received vaccine) can be generated when mass immunization is introduced to a population as a result of reduced virus circulation (ANDERSON & MAY, 1991; UKKONEN, 1996).
The question arises whether the virus is transmitted by individuals with lower immunity status, such as those who have been vaccinated, and in what way this may affect the population dynamics of measles transmission. Subclinical virus transmission within vaccinated communities has been reported by PEDERSEN et al. (1989). This studv highlighted the imuortance of defining quantitative6 lev
ETAL.
clonal antibody-based capture enzyme immunoassays for detection of specific antibodies to measles virus. Journal of Clinical Microbiologv. 29, 1466-147 1. Fooks, A. R., Stephenson,J. R., Warnes, A., Dowsett, B.A., Rima, B. K. &Wilkinson, G. (1993). Measles virus nucleocapsid protein expressed in insect cells assembles into nucleocapsid-like structures. Journal of General Virology, 74, 1439-1444. Forsey,T, Heath, A. B. & Minor, I? D. (1991). The first international standard for anti-measles serum. Biologicals. Y _ 19.I 237-241. Gay, N., Ramsey, M., Cohen, B., Hesketh, L., Morgan-Capner, I’., Brown, D. & Miller, E. (1997a). The epidemiology of measles in England and Wales since the 1994 vaccination campaign. Communicable Disease Report CDR Review, 7, R17-R21.
Gay, N., Miller, E., Hesketh, L., Morgan-Capner, I’., Ramsey, M., Cohen, B. & Brown, D. (1997b). Mumps surveillance in England and Wales supports introduction of two dose vaccination schedule. CommunicableDisease Report CDR Review. 7, R2LR26.
Hummel, K. B., Erdman, D. D., Heath, J. & Bellini, W. J. (1992). Baculovirus expression of the nucleoprotein gene of measles virus and utility of the recombinant protein in diagnostic enzyme immunoassays. Journal of Clinical Microbiology, 30, 2874-2880.
Massad, E., Azevedo-Neto, R. S., Burattini, M. N., Zanetta, D. M. T., Coutinho, F. A. B.,Yang, H. M., Moraes, J. C., Pannuti, C. S., Souza,V. A. U. F., Silveira, A. S. B., Struchiner, C. J., Oselka, G. W., Camargo, M. C. C., Omoto,T M. & Passos.S. D. 11995).Assessinathe efficacv of a mixed vaccination strategy against rubellg in SHo Paula, Brazil. Znarnational .@urnal of Epidemiology, 24, 842-850.
Miller, E.,Waight, E-Gay, N.,Ramsey, M., Vurdien, J., Morgan-Capner, I?, Hesketh, L., Brown, D.,Tookey, P. & Peckham, C. (1997) .The epidemiology of rubella in England and Wales before and after the 1994 measles and rubella vaccination campaign: fourth joint report from the PHLS and the National Congenital Rubella Surveillance Programme. Communicable Disease Report CDR Review, 7, R26-R32.
Acknowledgements
This work was supported by the Royal Society and FAPESP. R.S.A. and E.M. are supported by CNPq. D.J.N. is a Royal Society University Research Fellow. We thank D. Brown for comments on the text, B. Cohen for supply of measles HI antigen and W. J. Edmunds for helpful discussion. References
Albrecht, I’., Herman?, I<. & Burns, G. R. (1981). Role of virus strain in conventional and enhanced measlesplaque neutralisation test. Journal of Virological Methods, 3,25 l-260. Anderson, R. M. & May, R. M. (1991). Infectious Diseases of Humans: Dynamics and Control. Oxford: Oxford University Press.
-
Azevedo Neto, R. S., Silveira, A. S. B., Nokes, D. J.,Yang, H. M., Passos,S. D., Cardoso, M. R. A. & Massad, E. (1994). Rubella seroepidemiology in a non-immunized population of Sao Paulo State, Brazil. Epidemiology and Infection, 113, 161-173.
Babad, H. R., Nokes, D. J., Gay, N. J., Miller, E., MorganCapner, I? &Anderson, R. M. (1995). Predicting the impact of measles vaccination in England and Wales: model validation and analysis of policy options. Epidemiology and Infection, 114,319-344. Calvert, N., Cutts, F. T., Irving, R., Brown, D., Marsh, J. & Miller, E. (1996). Measles immunity and response to revaccination among secondary school children in Cumbria. Epidemiology and Infection, 116, 65-70.
Chen, R. T., Markowitz, L. E., Albrecht, I’., Stewart, J. A., Mofenson, L. M., Preblud, S. R. & Orenstein, W. A. (1990). Measles antibody: reevaluation of protective titres. JournaZof Infectious Diseases, 162, 1036-1042.
Christenson, B. & Bottiger, M. (1994). Measles antibodv-comoarison of lona-term vaccination titers. earlv vaccination tit&s and naturally acquired-immunity to and booster effects on the measles-virus. Kzccine,12, 129-133. Co&an, W. G. (1977). Sampling Techniques, 3rd edition. New York: John Wiley & Son. Erdman, D. D., Anderson, L. J., Adamms, D., Stewart, J., Markowitz, L. & Bellini, W. J. (1991). Evaluation of mono-
Norrby, E. (1962). Haemagglutination by measles virus. 4. A simple procedure for production of high potency antigen for haemagglutination-inhibition (HI) tests. Proceedings of the 111, Society for Experimental Biology and Medicine, 8114-8118.
Pannuti, C. S., Moraes, J. C., Souza, V., Camargo, M. & Hidalgo, N. (199 1). Measles antibody prevalence after mass immunization in Slo Paulo. Bulletin of the World Health Orsanitation, 69.557-560. Pedersen, I.‘R., Modrhorst, C. H., Glikmann, G. & von Magnus, H. (1989). Subclinical measles infection in vaccinated se&positive individuals in arctic Greenland. ‘Vaccine,7, 345-348.
Ramam, S., Gadag,V., West, R., Burris, J., Oates, E., Stead, F. & Bouillane, N. (1995). Comparison of commercial enzyme immunoassay kits with plaque reduction neutralization-test for detection of measles virus antibody. Journal of Clinical Microbiology, 33, 811-815.
Ramam, S., West, R., Gadag, V., Williams, B. & Oates, E. (1996). Immunitv aaainst measles in school-aaed children: implications for measles revaccination strategi&.. Canadian .yournal of Public Health, 87, 407-410. Samb, B., Aaby, P.,Whittle, H. C., Seek, A., Rahman, S., Bennett, J., Markowitz, L. & Simondon, F. (1995). Serologic status and measles attack rates among vaccinated and unvaccinated children in rural Senegal. Pediatric Infectious Disease Journal, 14,203-209. Sokal, R. R. & Rohlf, F. J. (1981). Biometry, 2nd edition. New York: W. H. Freeman. Ukkonen, P. (1996). Rubella immunity and morbidity: impact of different vaccination programs in Finland 1979-1992. Scandinavian .yournal of InfectiousDiseases, 283 l-35. Warnes, A., Fooks, A. R.-& Stephenson, J. R. (1994). Production of measles nucleoprotein in different expression systems and its use as a diagnostic reagent.Journal ofVirologicalMethads, 49,257-268. Received 27 August cepted for publication
1997; revised 17 November 18 November 1997
1997; ac-