keywords eradication, Madagascar, mass vaccination campaigns, poliomyelitis, seroprevalence

Transcription

1 Tropical Medicine and International Health volume 6 no 12 pp 1032±1039 december 2001 Mass vaccination campaigns to eradicate poliomyelitis in Madagascar: oral poliovirus vaccine increased immunity of children who missed routine programme M. Rakoto Andrianarivelo 1, P. Boisier 2, L. Rabarijaona 2, M. Ratsitorahina 2, R. Migliani 2 and H. Zeller 1 1 Unite de Virologie, Institut Pasteur de Madagascar, Antananarivo, Madagascar 2 Unite d'epideâmiologie, Institut Pasteur de Madagascar, Antananarivo, Madagascar Summary To assess the impact of mass vaccination campaigns using oral poliovirus vaccine (OPV) in Madagascar, serum neutralizing antibodies and geometrical mean titres (GMTs) to poliovirus were measured among 472 children aged up to 59 months, before and after the mass campaign, regardless of their previous history of routine vaccination. In this study, overall coverage with three routine and two mass campaign OPV doses was 69.9 and 93.4%, respectively. Seroprevalences to all poliovirus types were signi cantly higher after the mass campaign among the children who were not vaccinated through routine programme: 67.5% vs. 90.2% (P < 0.001) for type 1; 66.7% vs. 95.1% (P < 0.001) for type 2; and 55.3% vs. 82.9% (P < 0.001) for type 3. Geometrical mean titres to all poliovirus types also signi cantly increased after the mass campaign among the same study group: 34.5 vs (P < 0.001) for type 1; 35.1 vs (P < 0.001) for type 2; and 13.3 vs (P < 0.001) for type 3. Post-mass campaign seroprevalences and GMTs for poliovirus, especially types 1 and 3, among children who received up to two routine and two mass campaign OPV doses were signi cantly higher than pre-mass campaign seroprevalences among children who received three routine OPV doses. Reasons for lack of adherence to the vaccination programme and the mass campaign are discussed. The ndings strongly support the WHO strategy of conducting mass campaign in all endemic countries. However, as the mass campaign strategy now has been discontinued, it is crucial to increase the routine coverage and to improve acute accid paralysis surveillance in order to ful l the goal of poliomyelitis eradication. keywords eradication, Madagascar, mass vaccination campaigns, poliomyelitis, seroprevalence correspondence M. Rakoto Andrianarivelo, Unite de Virologie, Institut Pasteur de Madagascar, BP 1274, Antananarivo 101, Madagascar. Fax: (261 20) ; mala@pasteur.mg Introduction The strategies recommended in the polio eradication programme by the World Health Assembly in May 1988 are as follows: achieving and maintaining a high level of routine vaccination coverage with oral poliovirus vaccine (OPV); implementation of a surveillance system for acute accid paralysis (AFP) and wild poliovirus circulation; conducting of national mass vaccination campaigns and localized campaigns directed against the nal reservoirs for poliovirus transmission (Hull et al. 1994). Those strategies led to the eradication of poliomyelitis in the Americas in 1994 (Robbins & de Quadros 1997), and recently in the Western Paci c Region in 2000 [Expanded Programme on Immunization (EPI) 2000]. Despite reported outbreaks occurring in many parts of Africa, progress towards eradication of poliomyelitis takes place where extensive mass vaccination programmes are conducted. During mass vaccination campaigns, doses of OPV are given to children, usually <5 years of age, over a short period of time regardless of their previous vaccination status, with the primary objective to interrupt wild poliovirus circulation. Mass vaccination campaigns are conducted as two rounds separated by a 4±6-week interval. The last con rmed case of poliomyelitis associated with wild poliovirus in Madagascar was reported in 1997 (Centers for Disease Control and Prevention 1999). Wild poliovirus circulation among healthy children under 1032 ã 2001 Blackwell Science Ltd

2 5 years was observed in a study conducted in 1995±1996 (Rakoto Andrianarivelo et al. 1999). In Madagascar routine OPV doses are given at 6 (OPV-1), 10 (OPV-2) and 14 weeks (OPV-3), with a booster dose at 12 months. The routine vaccination coverage with OPV-3 in 1997 was 60% according to the national EPI unit. As a consequence, additional efforts towards polio eradication were considered necessary and followed by the implementation of a revised strategy. Mass vaccination campaigns using OPV were initiated in Madagascar in The rst series was conducted from 1 to 4 October and the second, from 5 to 8 November. It is important to estimate the impact of such a vaccination campaign in order to justify the substantial costs, due to purchase and distribution of vaccine, that are required to implement and maintain the long term eradication strategy. The impact of the mass campaign strategy depends on the vaccination coverage by which the outcome is measured through the AFP surveillance and monitoring of the wild poliovirus circulation (Birmingham et al. 1997). However, the sensitivity of the AFP surveillance in Madagascar has been as low as <0.5 per children aged up to 15 years (Centers for Disease Control and Prevention 1999). Therefore, a seroprevalence study has been considered useful to obtain more information about the impact of the current anti-polio strategy. The aim of this study was to assess the impact of the rst mass vaccination campaigns in Madagascar in 1997 among children up to 59 months regardless of their previous history of routine vaccination by monitoring changes in the immune status against poliovirus types 1, 2 and 3. Patients and methods Study sites and design The study was conducted in three distinct urban and rural settings in the highlands characterized by low socioeconomic status and poor sanitation: (1) two districts in the capital, Antananarivo (47 32 E, S) located in the central highlands at an altitude of 1500 m; (2) Miantsoarivo village (47 22 E, S), 60 km southwest of Antananarivo, which is dif cult to access by road and (3) Ihosy village (46 64 E, S), 600 km from Antananarivo at an altitude of 1200 m and close to the national road. From each of the three study sites, 300 children aged up to 59 months were selected to participate in the study. The size of the study samples was determined from an expected increase of at least 10% of the routine vaccination coverage, which was estimated to be 60% before the study was conducted. In Antananarivo, the study group was composed by random sampling technique in two districts. A house was randomly selected and individuals from this household and the neighbouring households meeting the age criteria were recruited until the number of 300 children was reached. In Ihosy and Miantsoarivo, the targeted children were selected from all the households in addition to those who were in close proximity, in order to reach the expected size of the study samples. Enrolment A baseline campaign questionnaire, which included date of birth, sex, site of enrolment, previous routine vaccination based on health cards, and reasons for lack of or incomplete routine vaccination, was completed for each child 2 weeks before the rst round of the mass campaigns. After 3±4 weeks following the second round, the study sites were revisited in an attempt to locate the original child. A second questionnaire was completed, including additional information about the number of mass campaign OPV doses, and reasons for lack of or incomplete participation in the mass campaign. Blood collection Blood was collected from the children after obtaining the mothers' consent. Blood (3±5 ml) was taken by venipuncture using Vacutainer tubes (Becton Dickinson, Aalst- Erembodegem, Belgium) on two occasions during the preand the post-mass campaign survey. After centrifugation (200 g ± 10 min), sera were transported within 24 h in containers at +4 C to the Institut Pasteur and stored at )20 C before laboratory processing. Serology The methods used for determination of the neutralizing antibody titres against poliovirus types 1, 2 and 3 have been described elsewhere (WHO 1997). Before laboratory testing, sera were inactivated for 30 min at 56 C in a water bath. Afterwards, twofold dilutions of sera ranging from 1/8 to 1/1024 were incubated for 3 h at 36 C in duplicate with 100 TCID 50 of the corresponding Sabin-type poliovirus. HEp-2 cells suspension were added and results were scored after 7 days. The neutralizing antibody titre is the end point dilution of sera, which protects 50% of the cell culture. A titre > 8 for any poliovirus serotype was considered to be positive. The GMT to each poliovirus serotype was calculated for all samples: any negative sample was converted into log 1, and a titre over 1024 was interpreted equal to ã 2001 Blackwell Science Ltd 1033

3 Statistical analysis The statistical signi cance in type-speci c polioviruses before and after the mass campaign was performed with Mac Nemar test for pairwise comparisons of the seroprevalences and a non-parametric test (Wilcoxon rank sum test) for the GMTs. Results Sample sizes and vaccination coverage A total of 929 children were enrolled in the study: 319 from Antananarivo, 319 from Miantsoarivo, and 291 from Ihosy. A total of 457 (49.2%) children were excluded, mainly because the parents refused blood sampling (n ˆ 384) and because the children could not be followed-up (n ˆ 73). Table 1 summarizes the characteristics of the 472 remaining study children. Children in the Antananarivo study group were older than those in the Miantsoarivo and Ihosy groups. The routine vaccination coverage was signi cantly higher in the Ihosy study group compared with the Miantsoarivo and Antananarivo groups, whereas the mass campaign coverage was signi cantly higher in Miantsoarivo compared with Ihosy and Antananarivo. In total, 68.9% (325 of 472) children had received less than three OPV doses and 26.1% (123 of 472) were not vaccinated during the routine programme. Overall coverage with two mass campaign OPV doses was 93.4% (441 of 472). Reasons for vaccination failure The mean age of the children, who were not routinely vaccinated, and those who received one to two, and more than three doses was 37.3, 31.9 and 35.3 months, respectively (P ˆ 0.235). For the 123 children who had not received any OPV routine dose, 106 (86.2%) parents could not give a reason for not attending the vaccination programme, 15 (12.2%) reported that the vaccination centre was too far away from their habitation, and 2 (1.6%) were not aware of the vaccination programme. For the 31 children who were partially or not at all vaccinated during the mass campaign, no reason was given by 10 (32.3%) parents. Other common explanations were: migration of the family (n ˆ 9, 29.0%); illness of the child (n ˆ 5, 16.1%); perception that OPV vaccination in the mass campaign was not required for children previously vaccinated during the routine programme (n ˆ 4, 12.9%); unawareness of the existence of the mass campaign (n ˆ 3, 9.7%). Sources of mass vaccination campaign information Four hundred and twenty-six households reported different sources of information related to the mass campaign. Some households reported one or more sources. Overall, the health care centre (42.6%), the radio (31.5%), public oral announcement (30.4%), and posters (16.2%) were the major sources of information about the mass campaign. The frequency of the different reported sources of information varied between the three study sites. In Ihosy, the health care centre was reported as the main source of information by 97.7% of the households. In Antananarivo, the radio (54.5%), public oral announcement (44.1%), and posters (36.6%) were reported as the main sources, whereas none of the households reported that they had obtained information from the health care centre. In Miantsoarivo, public oral announcement (58.7%), the radio (35.6%), and the town council (26.9%) were the most important sources Characteristics Antananarivo Miantsoarivo Ihosy P-value Table 1 Characteristics of the 472 children according to the study sites Number of children Sex ratio (male : female) Mean age (months) Number of routine OPV doses received 0 43 (25.9) 47 (38.8) 33 (17.8) 1±2 13 (7.8) 3 (2.5) 8 (4.3) < ³ (66.3) 71 (58.7) 144 (77.8) Number of mass campaign OPV doses received 0±1 18 (10.8) 3 (2.5) 10 (5.4) (89.2) 118 (97.5) 175 (94.6) OPV: Oral poliovirus vaccine. P-values based on v 2 -test for comparison of the mean age and the number of OPV doses received between study sites; Figures in brackets are percentages ã 2001 Blackwell Science Ltd

4 of information. Television was rarely reported as an information source at all three study sites. Changes in immune status according to study site The overall seroprevalences for all poliovirus types increased signi cantly among the 472 study children after the mass campaign: 78.2% vs. 94.3% (P < 0.001) for poliovirus type 1; 86.0% vs. 98.5% (P < 0.001) for poliovirus type 2; and 69.1% vs. 89.8% (P < 0.001) for poliovirus type 3. The GMTs also increased signi cantly after the mass campaign: 67.9 vs (P < 0.001) for poliovirus type 1; vs (P < 0.001) for poliovirus type 2; and 24.8 vs (P < 0.001) for poliovirus type 3. Before the mass campaign, the lowest level of seroprevalences and GMTs for all serotypes was observed in the Miantsoarivo group. However, the highest increase in seroprevalence and GMT after the mass campaign was observed in the same study group, especially for poliovirus types 2 and 3. Increases in the proportion of protected children according to the study site ranged from 11.5 to 19.8% for type 1, from 6.0 to 20.7% for type 2, and from 15.7 to 29.7% for type 3. The GMTs increased according to the study site by 3.7± 7.9-fold for type 1, by 4.1±8.9-fold for type 2, and by 4.4±10.5-fold for type 3. In all three study sites, the seroprevalences and the GMTs were statistically higher after the mass campaign. Changes in immune status according to previous routine vaccination Vaccination coverage for receipt of two mass campaign OPV doses was 91.9% (113 of 123) for the children without a history of previous routine dose, 83.3% (20 of 24) for children with a history of one or two routine dose(s), 93.2% (150 of 161) for children with three routine doses, and 96.3% (158 of 164) for children with four routine doses. The time interval between the last routine OPV dose received and the rst serum collection was 19.1 months for children who were vaccinated with one to two doses, 24.6 months for those who had received three doses, and 21.4 months for those who had received four doses. In all subgroups including those who were not vaccinated through routine programme, the seroprevalences and the GMTs were higher after the mass campaign (Figure 1). Pre- vs. post-mass campaign seroprevalences among the children without a history of previous routine vaccination and those with a history of one or two routine dose(s) were: 67.5% vs. 90.2% (P < 0.001) and 75.0% vs. 91.7% (P ˆ 0.2), respectively, for type 1; 66.7% vs. 95.1% (P < 0.001) and 79.2% vs. 100% (P ˆ 0.06), respectively, for type 2; and 55.3% vs. 82.9% (P < 0.001) and 62.5% vs. 95.8% (P ˆ 0.005), respectively, for type 3. Geometrical mean titres among the children without a history of previous routine vaccination and those with a history of one or two routine dose(s) were: 34.5 vs (P < 0.001) and 32.8 vs (P ˆ 0.02), respectively, for type 1; 35.1 vs (P < 0.001) and 49.2 vs (P ˆ 0.08), respectively, for type 2; and 13.3 vs (P < 0.001) and 12.9 vs (P ˆ 0.001), respectively, for type 3. The same pattern was observed for the children who had received three and four doses during the routine vaccination programme (Figure 1). The post-mass campaign seroprevalences and the GMTs for 133 children who had received zero to two routine OPV doses and two mass campaign OPV doses were compared with the pre-mass campaign seroprevalences and GMTs for 325 children, who had received three routine OPV doses. The post-campaign seroprevalences for all three serotypes were higher in the subgroup of children who were not or inadequately vaccinated under the routine programme, compared with the precampaign seroprevalences for those receiving three routine OPV doses: 91.7% vs. 82.5% (P ˆ 0.01) for type 1, 97.0% vs. 93.8% (P ˆ 0.2) for type 2, and 86.5% vs. 74.8% (P ˆ 0.006) for type 3. The post-campaign GMTs for all three serotypes were statistically higher in the subgroup of children who were not or inadequately vaccinated under the routine programme, compared with the pre-campaign GMTs for those receiving three routine OPV doses: vs (P < 0.001) for type 1, vs (P < 0.001) for type 2, and vs (P < 0.001) for type 3. Changes in immune status according to mass vaccination campaigns The mean age of the children who were not vaccinated during mass campaigns (n ˆ 18), and those who received one (n ˆ 13) or two doses (n ˆ 441) was 43.7, 32.5, and 35.4 months, respectively (P ˆ 0.06). Seroprevalences and GMTs to all poliovirus types increased signi cantly after a mass campaign among the children who had received two mass campaign OPV doses (data not shown). In the subgroup of children who had received one dose only, seroprevalences and GMTs to all poliovirus types also increased after mass campaign, but there was no statistical difference before and after the mass campaign. In the subgroup of children who had not received any dose, seroprevalences and GMTs were similar for type 1 before and after the mass campaign, and increased for types 2 and 3 after mass campaign, but in any case there was no statistical difference. ã 2001 Blackwell Science Ltd 1035

5 1036 ã 2001 Blackwell Science Ltd

6 Figure 1 Pre- vs. post-mass campaign seroprevalences (Line: s, pre-mass survey; d, post-mass survey) and geometrical mean titres (GMTs) (Bar: h, pre-mass survey;, post-mass survey) to poliovirus types 1, 2 and 3 according to the number of OPV doses received through the routine vaccination programme among the 472 children. Vertical bars represent standard errors of the mean values. Sample sizes are as follows: 0 dose, n ˆ 123; 1±2 doses, n ˆ 24; 3 doses, n ˆ 161; 4 doses, n ˆ 164. Seroconversion after the mass vaccination campaigns A high proportion of the children who were seronegative to poliovirus types 1, 2 or 3 before the mass campaign seroconverted after the receipt of two mass campaign OPV doses (Table 2). The seroconversion rates for each poliovirus serotype in children who were not vaccinated through the routine programme were similar to those who had received one or more routine OPV doses. In total, 8.1% (38 of 472) had no detectable antibodies against any of the three serotypes before the mass campaign. Vaccination coverage of those triple seronegative children was 31.6% with three routine OPV doses, and 92.1% with two mass campaign OPV doses. After the mass campaign, 39.5% (15 of 38) children seroconverted for all three serotypes, 44.7% (17 of 38) for two serotypes, 13.2% (5 of 38) for one serotype, and 2.6% (1 of 38) remained triple seronegative. After the mass campaign, the proportion of children, who were immunized against all three serotypes, increased from 58.1% (274 of 472) to 85.4% (403 of 472) (difference, 27.3%). Discussion In the present study, the seroprevalences and the GMTs at each site were statistically higher after mass campaign regardless of the number of OPV doses received previously through the routine programme. The relative low seroprevalence, especially type 3, among children, who had not been vaccinated earlier or who had received only one or two routine OPV doses, increased signi cantly after the mass campaign for all three poliovirus types. Interestingly, the seroprevalences for poliovirus types 1 and 3 were signi cantly higher in children, who had received one or zero OPV dose in the routine programme and two OPV doses during the mass campaign, than in children who had received as many as three or four routine doses before the mass campaign. These ndings suggest that a high proportion of children who had not been vaccinated or inadequately vaccinated following the routine programme, were susceptible to poliovirus infection, and by adding the OPV doses during the mass vaccination campaign their immune status improved considerably. The results of the study correspond to previous observations that mass vaccination campaigns are an imperative component in the control programme of poliomyelitis (Mas Lago et al. 1994; Richardson et al. 1995; Reichler et al. 1997). A total of 93.4% of the children adhered to the two OPV doses during the mass campaign, and there was an increase of 27.3% of children who were subsequently found to be seropositive against all three poliovirus types. At national level these gures indicate that approximately children under 5 years in Madagascar would bene t from such a control programme with an estimated cost of 2.4 US$ per child vaccinated. In the group of triple seronegative children who received two mass campaign doses, a substantial percentage seroconverted to all three serotypes. In the same context, a high proportion of children, who were seronegative to poliovirus types 1, 2 or 3 seroconverted to all three types regardless of the number of routine OPV dose received. These observations were in particular signi cant for children who were not previously vaccinated. However, it should be emphasized that a proportion of children remains susceptible to poliovirus infections, probably as a Table 2 Seroconversion rates after mass vaccination campaigns among children who were seronegative for poliovirus types 1, 2 or 3 before the mass campaign Number of routine OPV doses* Number of seronegative Mass campaign coverage children for poliovirus type 0 1±2 3 4 with two OPV doses ² Type 1 (n ˆ 103) 28/40 (70.0) 4/6 (66.7) 25/31 (80.6) 20/26 (76.9) 94/103 (91.3) Type 2 (n ˆ 66) 35/41 (85.4) 5/5 (100) 8/9 (88.9) 11/11 (100) 60/66 (90.9) Type 3 (n ˆ 146) 34/55 (61.8) 8/9 (88.9) 29/41 (70.7) 28/41 (68.3) 135/146 (92.5) OPV: Oral poliovirus vaccine. *Data are the number of children who seroconverted after receipt of two mass campaigns OPV doses/number of seronegative prior mass vaccination (% seroconverted). ²Data are the number of children who received two mass campaigns OPV doses/total number of seronegative prior mass vaccination (% mass campaign coverage). ã 2001 Blackwell Science Ltd 1037

7 result of inadequate seroconversion. Because poliomyelitis outbreaks may appear despite high routine coverage (Kim-Farley et al. 1984; Sutter et al. 1991; van Niekerk et al. 1994; Mpabalwani et al. 1996), our ndings support WHO recommendations that poliovirus infection in endemic areas can be prevented by conducting mass vaccination campaigns in addition to the routine programme. Several studies conducted in the tropical and subtropical regions have demonstrated that children may not seroconvert adequately to OPV, especially in relation to poliovirus types 1 and 3 (Patriarca et al. 1991; World Health Organization Collaborative Study Group on OPV 1995; Maldonado et al. 1997). In our study, low seroprevalences and GMTs prior to the mass campaign, especially to type 3, were observed in Miantsoarivo. Although we have not studied the various factors that may affect the immunogenicity of OPV administered through the routine programme, a relatively low OPV-3 routine coverage compared with the other study sites could in part explain the poor seroresponse to type 3 in this population. Nevertheless, the level of seroconversion and rises in antibodies titres to types 2 and 3 after mass campaign in this area was higher than has been observed in the other sites. Because of its strong immunogenicity compared with the other antigen components, humans probably seroconvert to type 2 with relatively higher ef ciency (Maldonado et al. 1997). A recent study in Jordan has demonstrated that the immunogenicity induced by OPV, especially type 3, was consistently higher after mass campaign compared with what was achieved through the routine programme alone with an equal number of OPV doses received (Reichler et al. 1997). Our study provides additional data that the seroprevalences and the antibodies titres for all poliovirus types strongly increase following mass vaccination campaigns although children are not vaccinated through routine programme. However, in all study groups a long interval between the last routine OPV dose received and the rst serum collection may explain the low antibody titres compared with what has been observed within a short time after mass campaign. A higher level of immunity acquired after mass campaigns compared with the level achieved through routine programme has been previously described (Richardson et al. 1995). This phenomenon is probably because of an effect of transmission of OPV to close contacts (Heymann et al. 1987). In our study, the receipt of two mass campaign OPV doses was associated with high antibody response for all three serotypes. In the other subgroups of children, the levels of seroprevalence and GMT to poliovirus slightly increased for all serotypes among children who had received one mass campaign dose, and for types 2 and 3 among children who did not receive any dose in the mass campaign. As measurement of viral excretion was not studied and because the sample size was low, it was not possible to evaluate if any subsequently passive transfer of poliovirus vaccine had in uenced the antibody response among children who were only partially or not at all vaccinated in the mass campaign. Successful mass vaccination campaigns using OPV were for the rst time carried out in Madagascar in 1997 by which the overall nationwide coverage was reported to be more than 100% by the Ministry of Health. In our study, the overall mass campaign coverage was 93.4%. A vaccination coverage of more than 100% during a mass campaign is very likely because of an underestimation of the target population and due to an enrolment of children > 5 years. The high coverage achieved during this mass vaccination campaign can be ascribed to excellent social mobilization as a result of ef cient information system directed to the households. However, a high percentage of the parents whose children were not vaccinated in the mass campaign were unable to give an explanation for why they had not attended. This tendency could be interpreted as the result of lack of information. A high level of information is essential in order to achieve a satisfactory coverage rate. In the information that is provided it is also important to emphasize to the parents that they have the possibility of having their children vaccinated in other vaccination site and that the child may receive mass campaign OPV doses even in the presence of illness, and irrespectively of their previous vaccination status. In many countries where mass vaccination programme have been implemented poliomyelitis was eradicated (Robbins & de Quadros 1997; Expanded Programme on Immunization 2000). However, despite excellent coverage during mass campaign, the introduction of imported wild poliovirus from neighbouring countries, where polio is endemic, has been reported in China (Centers for Disease Control and Prevention 2000). Furthermore, a massive outbreak of poliomyelitis occurred in Angola, mainly because of migration of unvaccinated population to urban settings, low routine and mass campaign coverage, and political instability (Valente et al. 2000). For those reasons, a high level of routine coverage should be achieved and maintained to ensure individual protection of children, and high-quality mass vaccination campaigns should be sustained with a greater emphasis on synchronized mass campaign across national borders where poliovirus transmission still occurs. The ndings of this study support the strategy of conducting mass campaigns aiming to achieve a higher rate of protection among children who have been missed in the routine programme and to reduce the number of 1038 ã 2001 Blackwell Science Ltd

8 susceptible children. Being an island, Madagascar is less at risk of having polio introduced from outside. However, as the mass campaign strategy has now been discontinued in Madagascar, vaccination will be undertaken through the routine programme only. It is therefore crucial to increase the routine coverage and to conduct more aggressive AFP surveillance in order to ful l the goal of eradicating poliomyelitis on the island. Acknowledgements We are indebted to the families of the children for participation in this study; the laboratory staff of the Virology Unit, Institut Pasteur de Madagascar, for processing blood samples; and to Dr Peter Leutscher for critical review the manuscript. References Birmingham M, Aylward B, Cochi S & Hull H (1997) National immunization days: state of the art. Journal of Infectious Diseases 175, S183±S188. Centers for Disease Control and Prevention (1999) Progress toward poliomyelitis eradication, African Region, 1988±April Morbidity and Mortality Weekly Report 48, 513±518. Centers for Disease Control and Prevention (2000) Importation of wild poliovirus into Qinghai province, China, Morbidity and Mortality Weekly Report 49, 113±114. Expanded Programme on Immunization (2000) Certi cation of poliomyelitis eradication, WHO Western Paci c Region. Weekly Epidemiological Record 75, 399±400. Heymann D, Murphy K, Brigaud M, Aymard M, Tembon A & Maben G (1987) Oral poliovirus vaccine in tropical Africa: greater impact on incidence of paralytic disease than expected from coverage surveys and seroconversion rates. Bulletin of the World Health Organization 65, 495±501. Hull H, Ward N, Hull B, Milstien J & de Quadros C (1994) Paralytic poliomyelitis: seasoned strategies, disappearing disease. Lancet 343, 1331±1337. Kim-Farley R, Rutherford G, Lich eld P et al. (1984) Outbreak of paralytic poliomyelitis, Taiwan. Lancet 2, 1322±1324. Maldonado Y, Pena-Cruz V, de la Luz Sanchez M et al. (1997) Host and viral factors affecting the decreased immunogenicity of Sabin type 3 vaccine after administration of trivalent oral polio vaccine to rural Mayan children. Journal of Infectious Diseases 175, 545±553. Mas Lago P, Ramon Bravo J, Andrus J et al. (1994) Lessons from Cuba: mass campaign administration of trivalent oral poliovirus vaccine and seroprevalence of poliovirus neutralizing antibodies. Bulletin of the World Health Organization 72, 221±225. Mpabalwani E, Monze M, Saijo M, Terunuma H & Luo N (1996) Poliomyelitis outbreak in Zambia. Lancet 347, Niekerk A van, Vries J, Baard J, Schoub B, Chezzi C & Blackburn N (1994) Outbreak of paralytic poliomyelitis in Namibia. Lancet 344, 661±664. Patriarca P, Wright P & John T (1991) Factors affecting the immunogenicity of oral poliovirus vaccine in developing countries: review. Review of Infectious Diseases 13, 926±939. Rakoto Andrianarivelo M, Rabarijaona L, Boisier P, Chezzi C & Zeller H (1999) Wild poliovirus circulation among healthy children immunized with oral polio vaccine in Antananarivo, Madagascar. Tropical Medicine and International Health 4, 50±57. Reichler M, Kharabsheh S, Rhodes P et al. (1997) Increased immunogenicity of oral poliovirus vaccine administered in mass vaccination campaigns compared with the routine vaccination program in Jordan. Journal of Infectious Diseases 175, S198±S204. Richardson G, Linkins R, Eames M et al. (1995) Immunogenicity of oral poliovirus vaccine administered in mass campaigns versus routine immunization programmes. Bulletin of the World Health Organization 73, 769±777. Robbins F & de Quadros C (1997) Certi cation of the eradication of indigenous transmission of wild poliovirus in the Americas. Journal of Infectious Diseases 175, S281±S285. Sutter R, Patriarca P, Brogan S et al. (1991) Outbreak of paralytic poliomyelitis in Oman: evidence for widespread transmission among fully vaccinated children. Lancet 338, 715±720. Valente F, Otten M, Balbina F et al. (2000) Massive outbreak of poliomyelitis caused by type-3 wild poliovirus in Angola in Bulletin of the World Health Organization 78, 339±346. World Health Organization (1997) Manual for the Virological Investigation of Polio. WHO/EPI/GEN WHO, Geneva, p. 66. World Health Organization Collaborative Study Group on Oral Poliovirus Vaccine (1995) Factors affecting the immunogenicity of oral poliovirus vaccine: a prospective evaluation in Brazil and the Gambia. Journal of Infectious Diseases 171, 1097±1106. ã 2001 Blackwell Science Ltd 1039