Poliomyelitis: epidemiology and prophylaxis

Bulletin of the World Health Organization, 55 (6): 747-753 (1977) Poliomyelitis: epidemiology and prophylaxis 4. Serological and virological surveys conducted after a mass vaccination campaign for the control of a threatening poliomyelitis epidemic D. METSELAAR,' K. McDONALD,2 W. GEMERT,3 B. NOTTAY,'4 & J. M. MULI B In 1973, a type 1 poliomyelitis epidemic in Kenya was curtailed at an early stage by two mass distributions of trivalent oral vaccine. It was considered useful to know the immunity status of the child population that had resultedfrom the vaccine distributions and that had presumably contributed to its control. We also wished to know to what extent wild and vaccine virus strains were in circulation after the mass vaccination campaign. Anal swabs and blood were collectedfrom a sample of the children in four areas where the efficiency of vaccine distribution had varied, and the results of virus isolation attempts and antibody tests are reported. Three poliovirus strains were isolated. It was surprising that, in general, the herd immunity after two vaccination rounds did not substantially differ from that found in Kenya on other occasions. Possible reasons for these results are discussed. Since 1954, the major poliomyelitis epidemics in Kenya have been caused by type 1 virus and have occurred nearly every third year (1, 2). This regularity has allowed both the prediction of epidemics and attempts to prevent them. A cam- paign to control an epidemic in the early stages of development during 1972-1973 is described by Koinange et al. (3). Two rounds of mass vaccination were conducted on that occasion when 1.7 and 1.3 million doses of trivalent oral poliomyelitis vaccine were distributed. To obtain an insight into poliovirus circulation and the immunity status of the children some time after the campaign, virological and serological sur- veys were conducted in four areas in Kenya. These surveys were also intended to show whether the poliovirus strains isolated would be of wild or vaccine origin. 1 Head, Department of Virology, Netherlands Medical Research Centre and National Public Health Laboratory Services of Kenya, P.O. Box 20752, Nairobi, Kenya. 'Assistant Virologist, Medical Research Centre, Nairobi, Kenya. 2 Statistician, Medical Research Centre, Nairobi, Kenya. ' Research Fellow, Enteric Virology Branch, Center for Disease Control, Atlanta, GA 30333, USA. ' Senior Technologist, National Public Health Laboratory Services, Nairobi, Kenya. MATERIAL AND METHODS Areas The four geographical areas chosen for the viro- logical and serological surveys are widely separated and belong to the more densely populated regions of Kenya. In general, the people were living on their own land rather than in villages. The interval be- tween the last mass vaccination and the surveys varied. During this interval, relatively few vaccine doses were distributed at health centres and dispen- saries. These vaccinations were part of the normal routine and were not related to the mass vaccination rounds. In west Kenya an area east of Kisumu township was chosen. Specimens were collected 4 months after the second vaccination round and during that inter- val, vaccine had been given to a few children only. In Kwale District, situated in the eastern coastal region of Kenya, specimens were collected 7 months after the last vaccination round. During that interval, 3160 doses of vaccine had been distributed among 40 000 children under 4 years of age living in the district. In Nithi Division, north of central Kenya, the surveys were conducted 8 months after the last vaccination round. During the interval, 2750 doses 3660 -747- D. METSELAAR ET AL. of vaccine were distributed among approximately 20 000 children under 4 years of age. In Kiambu District in central Kenya, the interval between the last mass distribution round and the survey was 11 months. During the interval, 18 580 doses of vaccine were distributed in the district where approximately 75 000 children under 4 years of age were living. In Kiambu and east Kisumu, medical students collected the specimens as part of their training in community health techniques. Only children under 3 years of age were included. In the two other areas, trained laboratory staff collected the specimens of children under 4 years of age. Sampling Sampling was done by random selection of a number of administrative units. In these units, a number of heads of household were randomly chosen from the lists of taxpayers in the areas. The house of each of the selected taxpayers served as a starting point from which a cluster of houses was visited (4). Serum was obtained from blood from a finger prick. The blood was collected in small glass tubes marked for 0.5 ml and containing 0.25 ml of diluent. Blood was dropped into the tube until the 0.5-ml mark was reached. After centrifugation, the serum was separated from the clot. The serum was intended to be diluted 1 in 3. With the exclusion of Kiambu District, enquiries were made at the time the speci- mens were taken concerning the number of times the child had already received oral poliomyelitis vaccine (not necessarily during the mass vaccination cam- paign). Faeces were collected on anal swabs and were kept in 1 ml of Hanks' balanced salt solution with antibiotics. Both serum and swabs were frozen and transported to the laboratory in liquid nitrogen. Laboratory methods Virus isolation was attempted in three tubes of primary baboon kidney cells (BKC). To each tube was added 0.2 ml of the transport medium. Isolates producing cytopathogenic effects compatible with those of poliovirus were identified against rabbit poliomyelitis antisera. The identification was verified by one of the authors (B.N.) in the Enteric Virology Branch of the Centre for Disease Control in Atlanta, GA, USA. The polioviruses were then tested for their relationship with vaccine strains in Wecker's plaque reduction test as modified by Nakano (5). The reproductive capacity at supraoptimal tempera- ture (RCT) was assessed as described by Nakano et al. (6). Antibodies were estimated by the micrometabolic inhibition test as described by Kyriazopoulou & Bell (7) using Vero instead of HeLa cells. A tube neutralization test with BKC was used only for the Kiambu sera and by mistake, these sera were diluted 1 in 30. They were tested against type 1 only. The sera from the other areas were diluted 1 in 9 and tested against all three types of poliovirus. Approximately 100 TCID50 of virus was used as the antigen in all the tests. Control titrations were included and sera were inactivated for 30 min at 56°C. RESULTS Virus isolation The results of virus isolation given in Table 1 show isolation rates that are low for tropical coun- Table 1. Virus isolations and results of Wecker and RCT tests Months after No. of All enteroviruses RCT markerArea mascampaign specimens Polioviruses Wecker testss campai specimens No. % 39.5° 39.9C Kwale 7 233 28 12.0 East Kisumu 4 86 7 8.1 Nithi 8 239 41 17.2 2 x type 2 vaccine-like -a Kiambu 11 176 32 18.2 1 x type 1 non-vaccine- ±b like Total 734 108 14.7 a More than 5 logio difference with a control incubated at 35.5eC. b More than 2 but less than 5 logio difference with the control. 748 POLIOMYEL1TIS: EPIDEMIOLOGY AND PROPHYLAXIS Table 2. Mean vaccine dose received for age and area Age Kwale Kisumu Nithi Overall d.f. b p (months) 1 2 3 6-11 0.8 (25) a 0.4 (17) 1.0 (25) 0.7 (67) 2 cs0.1 0 12-23 0.6 (44) 0.5 (28) 2.1 (38) 1.1 (110) 2 <0.01 24-35 0.9 (78) 0.5 (41) 1.9 (74) 1.2 (193) 2 < 0.01 36-47 1.3 (74) - 1.9 (73) 1.6 (147) 1 < 0.01 Overall 1.0 (221) 0.5 (86) 1.9 (210) 1.2 (517) a The number of children is in parentheses. b Degrees of freedom. tries. Of 108 viral isolates, only 3 were polioviruses effect significant at the 5% level was in the group (two type 2 and one type 1), of which two were aged 6-11 months and then only for types 2 and 3. antigenically vaccine-like. In spite of the differences in the number of doses given, no significant differences could be demon- Doses of vaccine and antibody responses strated in the proportions positive for antibodies in The mean number of doses received by the chil- the three areas. The figures for the three areas were dren according to information received from therefore combined, regardless of whether the chil- mothers is given in Table 2. The mean number of dren received vaccine or not, with the exception of doses in the three areas differed significantly. The those aged 6-11 months for whom the influence of children in Nithi received the highest mean dose. the dose was significant. The results are shown in When the children less than 35 months of age are columns 8 and 9 of Table 3. In general, the propor- considered, those in Kwale received a higher mean tion of children with antibodies increased with age, dose than those in Kisumu. For the three areas, the and in those aged 6-11 months the effect of the differences in the group aged 6-11 months are small. vaccine was clear. In Table 3, the percentages of children with antibodies against the three types of poliovirus (for DISCUSSION Kiambu against type 1 only) are given for each age group. Melnick & Wenner (8) have estimated that the When the vaccine dosages were correlated with the isolation rate from anal swabs may be no more than proportion of children with antibodies in each age 50% of that from faecal specimens. This may be group for the three areas combined, the only dose responsible for the low isolation rates found in our Table 3. Percentage of children with antibodies per age group and per virus type Kwale East Kisumu Nithi All three areas Kiambu Percentage Percentage Percentage Percentage Age No. with anti- No with anti- No with anti- No with anti- No Percentage(months) exam- bodies of exam- bodies of exa° bodies of iexam bodies of with anti- mned type: ined type: ined type: ined type: ined typies1f 1 2 3 1 2 3 1 2 3 1 2 3 6-11 25 8 44 12 17 11 35 11 25 24 28 17 36 not 6 17 3 14 28 vacc. 31 vacc. 26 58 26 12-23 44 34 58 27 28 32 35 17 38 43 60 25 110 35 52 24 48 40 24-35 78 42 65 43 41 39 41 34 74 50 61 33 193 45 59 37 64 52 36-47 74 76 71 58 - 73 73 66 55 147 74 68 56 - - 749 D. METSELAAR ET AL. survey. Moreover, many specimens were contami- nated with fungi and had to be filtered and this may have resulted in loss of virus. Metselaar et al. (9), in another survey in Kenya, obtained 19% non-polio- viruses and 7% polioviruses (probably vaccine strains) from 205 swabs taken in rural areas. Nottay (unpublished observations, 1971) found 31.3% non-polioviruses and 7.6% polioviruses (partly vac- cine strains) in 2925 faecal specimens from children attending child health clinics in Nairobi. These latter percentages are more in accordance with the usual carrier rates in children in tropical countries. Only three poliovirus strains were isolated; two of these (both type 2) had all the characteristics of vaccine strains and, although the results of the RCT test leave some possibility of its being of vaccine origin (6), the third one, a type 1 strain, was most probably a wild virus. There is no reason to suppose that polioviruses were more affected by the swab technique than were other enteroviruses and therefore one wild poliovirus isolate among 108 enteroviruses is a low incidence. It has been suggested (10, 12) that in tropical countries during interepidemic periods, wild strains of low virulence regularly circulate and provide natural immunization to the great majority of susceptible people with little serious disease. Epidemics should, according to this theory, be the result of the intro- duction of virulent strains. Metselaar (2) has expres- sed doubts whether this large-scale circulation of low-virulence strains in interepidemic periods still occurs and whether, if ever, epidemics in tropical countries still result from the introduction of virulent strains. He suggested that a gradual selection of virulent strains at the expense of less virulent ones is taking place and this may be responsible for the changing epidemiological pattern of poliomyelitis in tropical countries. He compared the present pattern of poliomyelitis in many tropical countries with that of measles; in the absence of vaccination, epidemics of the latter disease occur when there are enough susceptible people in the population to allow large- scale circulation of the virus. In interepidemic periods, the virus is, at most, circulating at a low level. The scarcity of wild polioviruses at the time of our surveys tends to favour this view. Compared with types 1 and 3, type 2 vaccine- derived strains have relatively stable antigenic char- acteristics. Furthermore, in the Wecker test, wild type 2 poliovirus strains are rarely vaccine-like. Type 2 vaccine strains also change their RCT char- acteristic from negative to positive less frequently than the other types (6). It is, therefore, not possible to decide whether the type 2 isolates are derived from vaccine distributed 8 months earlier during the mass vaccination campaign or from recent vaccina- tions. The latter is the more likely possibility. The children concerned were siblings, and according to the information given by relatives, one child was 3 years of age and had never received vaccine, while the other was 1 year old and had received vaccine three times. The age-immunity curves of the four areas show a rise with age (Table 3). The percentages with anti- bodies would probably have been somewhat higher if the tests had been started from lower dilutions (13). There are no significant differences for the areas. When three of the areas are combined (columns 8 and 9) the percentages in the various age groups are of the same order as those found by Metselaar & Nottay and by Masar & Christensen some years ago in other areas of Kenya (14). Appar- ently, the two-round mass vaccination campaign did not appreciably reduce the age at which immunity resulting from infections by wild virus is achieved. Poliomyelitis is endemic in Kenya with epidemics (1, 2). The increase with age of the proportion of children with antibodies results from immunizations by circulating wild virus and, for the last 10 years or so, also from vaccine distribution. Vaccine is given routinely at health centres and dispensaries and occasionally during mass vaccination campaigns. Only a proportion of the children are reached by routine vaccination; this proportion varies over the country but rarely exceeds 35% (15). It was therefore assumed that the doses of vaccine mentioned by the mothers as received by the chil- dren were, for the greater part, given at the time of the mass vaccination campaign preceding the sur- veys. It was expected that shortly after the mass distribution a significant influence of the number of doses on immunity rates would be easily demon- strable. That this influence was restricted to the youngest age group was surprising. It is unlikely that this was a chance observation, as comparable results were obtained in another experiment in Kenya where vaccine was given on a house-to-house basis to pre- registered children (16). It is equally unlikely that high proportions of immune children among those over 11 months of age before the mass campaign masked the influence of the number of doses, since the mass distributions were initiated because of an impending epidemic of poliomyelitis type 1 (3). Also, immunity rates, at least for this type, must have been 750 POLIOMYELITIS: EPIDEMIOLOGY AND PROPHYLAXIS low, not only in the 6-1 1-month age group but also in those aged 1 and 2 years. A more likely possibility to be considered is that the vaccine virus had spread to some extent in the community immediately after vaccine distribution and that such spread is more favoured in infants, who have more contact with each other, than in babies. This spread may have masked differences in conversion in relation to dose in those over 11 months of age because children not receiving vaccine at the time of distribution were nevertheless vaccinated. Also, in industrialized countries such spread accounts for a number of immunizations (17). This spread is only likely to be of significance soon after distribution because vaccine viruses have limited circulation potential (18). The few isolations made by ourselves are in accordance with that observation. Temporary circulation of vaccine virus could also explain the absence of significant differences between areas in the proportion of children with antibodies, notwithstanding differences in the mean doses. This would imply, at least under Kenyan conditions, that above a certain minimum coverage by mass vaccina- tion shortcomings tend to disappear. Gold et al. (13) compared the vaccination history as given by informants with records of physicians and found that information given from memory was frequently incorrect. This is also likely to be true for our survey and may have influenced the results. In none of the age-groups did the number of vaccine doses have a significant influence on immu- nity rates for poliovirus type 1. Moreover, a substan- tial proportion of the children remained negative for antibodies against one or more of the three types, but especially against type 1. It may well be that, of the three types of vaccine, type 1 caused least conversions. Low conversion rates by live poliomyelitis vac- cines are common in tropical countries. Interference by non-poliomyelitis enteroviruses, which circulate in large numbers in tropical countries, and the probable presence of an inhibitor in the intestinal tract of many children in these countries (19) are put forward as explanations. Sabin et al. (20) reduced interference in Mexico by giving the vaccine to as many children as possible in an area within a few days. Apparently, this method did not work as well in Kenya, at least not when the age-antibody rates are considered. Nevertheless, the 1972-1973 epidemic was, in all probability, halted by the mass vaccination cam- paign. This is indicated by the number of specimens from poliomyelitis patients received in the labora- tory at that time; these increased until 2 weeks after completion of the first round of vaccination and then dropped abruptly (3). Three possible explanations are put forward for this: 1. In some of the children negative for type 1 antibodies the intestine may have been colonized by vaccine types 2 and 3 and, through interference, may have prevented infection by wild type 1 virus. If this is true, then mass vaccination campaigns in which live trivalent poliomyelitis vaccine is distributed con- trol epidemics not only by immunization against the epidemic type but also by interference with the circulation of that type. This view is supported by the results of mass vaccination campaigns in Singa- pore (21) and Kingston-upon-Hull (22) during which live type 2 vaccine was distributed to control a type 1 epidemic. 2. Type 2 poliovirus shares antigens with type 1 and to a lesser extent with type 3. Ashkenazi & Melnick (23) observed heterotypic antibody re- sponses after feeding monovalent live vaccine. Not- tay (unpublished observations, 1971) observed that in 255 sera collected in Kenya from children under 12 years of age, type 1 and type 2 antibodies were found together in individual sera more frequently than would be expected from random distribution. The coefficient of correlation was 0.39 (P<0.05) in those aged 6-11 months and 0.57 (P<0.01) in the 1-2-year-old group. It ceased to be significant in the older age groups. No correlation was found for other antibody combinations. Ashkenazi & Mel- nick (23) suggest that sometimes traces of such het- erotypic antibodies may be present that are not detected in serological tests. If so, these traces could nevertheless prevent infection with either vaccine or wild virus. On the one hand, therefore, they could contribute to low conversion rates after vaccination and on the other hand they could explain why an epidemic is halted notwithstanding low antibody rates in the youngest age groups. 3. The herd immunity required to prevent the emergence of a poliomyelitis epidemic seems to be considerably lower than is the case for measles and rubella (17, 24). Presumably an epidemic of polio- myelitis may also come to an end with a larger residue of suceptible people in the population than do epidemics of measles and rubella. John (25) recommended that at least five doses of 751 752 D. METSELAAR ET AL. triple vaccine be given to obtain satisfactory conver- sion rates in tropical countries. This is an impossible target for many of these countries. Schwartz et al. (26) stressed the importance of starting vaccination at an early age. Our results seem to give support to this recommendation. ACKNOWLEDGEMENTS We are grateful to the Director of Medical Services, Kenya, for permission to publish the paper and to Dr M. H. Hatch, Chief, Enteric Virology Branch, CDC, Atlanta, GA, USA for making staff, facilities, and materials availkIble for the identification, typing, and differentiation of viruses. RtSUMt POLIOMYELITE: EPIDEMIOLOGIE ET PROPHYLAXIE 4. ENQUETES SEROLOGIQUES ET VIROLOGIQUES CONSECUTIVES A UNE CAMPAGNE DE VACCINATION DE MASSE POUR PARER AU RISQUE D'EIPIDEMIE DE POLIOMYELITE Un debut d'epidemie de poliomy6lite due au virus de type 1 a ete rapidement enraye en 1973 grace a deux distributions de vaccin buccal triple. On a ensuite procedd dans 4 zones a des prelevements rectaux en vue de l'isolation du virus, ainsi qu'au recueil de lames de sang pour examen serologique, le laps de temps ecoule entre ces pr6levements et la distribution de vaccin n'etant pas le meme pour les 4 zones. Dans 3 de celles-ci, on s'est efforce de determiner le nombre des doses de vaccin buccal contre la poliomyelite requ par chaque enfant dont un 6chantillon de sang a e prelev6, soit au cours des deux distributions soit en provenance d'autres sources. On a pu isoler a partir de 734 specimens deux souches correspondant au vaccin du type 2 et une souche du type 1 - probablement sauvage. Il ne semble pas que le virus sauvage ait ete tres repandu apres l'endiguement de 1'e'pide'mie. On a classe les serums selon la zone de pr6levement et I'age des enfants. Bien que le nombre de doses adminis- trees dans les diverses zones ait e tres variable, on n'a pas enregistre entre les zones de differences importantes dans le titre d'anticorps. On peut en conclure qu'a partir d'un minimum ddtermine de couverture par la vaccination de masse, la circulation de souches de virus provenant du vaccin peut provoquer l'accroissement du taux d'anticorps chez les sujets non vaccines. En combinant les chiffres obtenus pour 3 zones, on n'a releve une diffe- rence sensible entre les enfants vaccines et non vaccines, en ce qui concerne la proportion de ceux qui possedaient des anticorps, que dans le groupe d'age de 6 a 11 mois et pour les virus de type 2 et 3. Les titres d'anticorps apres les vaccinations de masse ne differaient pas sensiblement de ceux releves au Kenya a d'autres occasions en l'absence de vaccination. Et pourtant l'epidemie a ete endiguee... Parmi les hypotheses avancees en guise d'explication figurent l'interference du virus provenant du vaccin avec le virus sauvage, la propagation du virus provenant du vaccin dans la collectivite et l'existence d'anticorps a reaction crois6e non decel6s au moyen de l'epreuve effectuee. REFERENCES 1. NOTTAY, B. & METSELAAR, D. Bulletin of the World Health Organization, 48: 421-427 (1973). 2. METSELAAR, D. Lancet, 1: 174-176 (1976). 3. KOINANGE, W. ET AL. Bulletin of the World Health Organization, 48: 543-545 (1973). 4. GESER, A. ET AL. Bulletin of the World Health Organization, 43: 521-537 (1970). 5. NAKANO, J. H. & GELFAND, H. M. American journal of hygiene, 75: 363-376 (1962). 6. NAKANO, J. H. ET AL. Progress in medical virology, 24 (1978) (in press). 7. KYRIAZOPOULOU, V. G. & BELL, E. G. Bulletin of the World Health Organization, 47: 171-175 (1972). 8. MELNICK, J. L. & WENNER, H. A. In: Lennette, E. H. & Schmidt, N. J. ed. Diagnostic procedures for viral and rickettsial infections, 4th ed. New York, Ameri- can Public Health Association, 1969, p. 558. 9. METSELAAR, D. ET AL. Bulletin of the World Health Organization, 48: 429-433 (1973). 10. RHODES, A. J. Bulletin ofhygiene, 22: 353-385 (1947). 11. SABIN, A. B. American journal ofpublic health, 41: 1215-1237 (1951). POLIOMYELITIS: EPIDEMIOLOGY AND PROPHYLAXIS 753 12. SABIN, A. B. In: Poliomyelitis. Geneva, World Health Organization, 1955 (Monograph Series, No. 26). 13. GOLD, E. ET AL. New England journal of medicine, 289: 231-235 (1973). 14. METSELAAR, D. & NoTEAY, B. In: Vogel, L. C. et al. ed. Health and disease in Kenya. Nairobi, E.A. Literature Bureau, 1974, pp. 255-260. 15. SCHOFELD, F. D. ET AL. In: Seminar on immuniza- tions in Africa, Kampala, December 1971. Paris, Centre intemationale de l'Enfance, 1972, pp. 293- 299. 16. METSELAAR, D. ET AL. Bulletin of the World Health Organization, 55, 755-759 (1977). 17. HORSTMANN, D. M. American journal of epidemio- logy, 97: 173-186 (1973). 18. Fox, J. P. In: Cumakov, M. P. ed., Fourth Scientific Conference on Oral Live Poliovirus Vaccine, Moscow, May 1960. Moscow, Institute of Poliomyelitis and Virus Encephalitis, 1961, pp. 531-545. 19. DOMOK, I. ET AL. Bulletin of the World Health Organization, 51: 333-347 (1974). 20. SABIN, A. B. ET AL. Journal of the American Medical Association, 173: 1521-1526 (1960). 21. HALE, J. H. ET AL. British medical journal, 2: 1541- 1549 (1959). 22. LANCET. Oral vaccine in an emergency. Lancet, 1: 439 (1963). 23. ASHKENAZI, A. & MELNICK, J. L. New England journal of medicine, 267: 1228-1230 (1962). 24. OBERHOFER, T. R. ET AL. American journal of epide- miology, 101: 333-339 (1975). 25. JOHN, T. J. British medical journal, 1: 812 (1976). 26. SCHWARZ, T. A. ET AL. Journal of hygiene, 70: 719- 726 (1972).