Lack of efficacy of the standard potency Edmonston-Zagreb live, attenuated measles vaccine in African infants.

Lack of efficacy of the standard potency Edmonston-Zagreb live, attenuated measles vaccine in African infants P. Kiepiela,1 H.M. Coovadia,2 W.E.K. Loening,3 P. Coward,4 G. Botha,5 J. Hugo,6 & P.J. Becker7 The efficacy of standard potency Edmonston-Zagreb (E-Z) measles vaccine was tested in a randomized trial of Black infants in a rural area of South Africa where a measles epidemic was occurring. The following immunization schedules were used: 48 infants aged 4-8.5 months who received 3.9 log 50 in- fectious units of E-Z vaccine (group A); 48 infants aged 4-8.5 months who received 3.28log50 infec- tious units of Schwarz vaccine (group B); and 28 infants aged >9 months who received 3.28log50 in- fectious units of Schwarz vaccine and served as controls (group C). For infants aged less than 23 weeks who were given either the E-Z or Schwarz vaccine, the number of seropositives was low (28%), irrespective of the pre-vaccination level of measles antibody. There was a higher number of seropositives (68%) among those in the age range >23 weeks to <36 weeks who received the E-Z vaccine rather than the Schwarz vaccine (36%). When administered to children aged >36 weeks, the Schwarz vaccine produced a satisfactory, though suboptimal response rate (61 %). There was no correlation between seropositivity and pre-vaccination measles antibody status. Use of the standard dose of E-Z vaccine may have been one of the factors for this poor response, and this supports the WHO recommendation that titres higher than the standard potency vaccine are needed if 6-month-old infants are to be successfully immunized against measles. Introduction Measles remains one of the most severe and important infectious diseases of children who live under poor socioeconomic conditions (1). Until recently, the WHO Expanded Pro- gramme on Immunization (EPI) recommended that children in developing countries be immunized against measles as soon as possible after 9 months of age. For children of this age seroconversion rates of Research Technician, Department of Paediatrics and Child Health, Faculty of Medicine, University of Natal, Durban, South Africa. 2 Professor and Head, Department of Paediatrics and Child Health, Faculty of Medicine, University of Natal, P.O. Box 17039, Congella, 4013, Durban, South Africa. Requests for reprints should be sent to this author. 3Professor, Department of Paediatrics and Child Health, Faculty of Medicine, University of Natal, Durban, South Africa. 4Medical Technologist, Department of Paediatrics and Child Health, Faculty of Medicine, University of Natal, Durban, South Africa. 5Registrar, Manapo Hospital, Qwa-Qwa, South Africa. 6 Medical Superintendent, Manapo Hospital, Owa-Qwa, South Africa. 7 Senior Consultant, Institute of Biostatistics, Medical Research Council of South Africa. Tygerberg, South Africa. Reprint No. 5166 75-98% are achieved with the Schwarz vaccine.a These rates are significantly lower if the conven- tional Schwarz vaccine is used to immunize children aged <9 months, because of the presence of circulating maternal antibodies (2). In the south- eastern border region of South Africa, 20-45% of cases of measles occur in infants aged less than 8 months (3); effective methods of immunizing the very young against measles in high-density regions of developing countries need to be found. Attempts have been made to solve this problem bv immunizing infants aged 4-8 months. Several studies that administered Schwarz vaccine to young infants consistently reported poor clinical results (4, 5). However, vaccination at such an early age with a measles vaccine grown in human diploid cells (Edmonston-Zagreb (E-Z) vaccine) is highly effective (6-12). High titre E-Z vaccine has been reported to provide superior seroconversion than the Schwarz vaccine in 4-7-month-olds, not- withstanding the presence of maternal antibodies. The differences between South Africa and the developing countries where these trials were con- ducted (Mexico (4, 7, 8), the Gambia (10, 12), and a Clements, C.J. et al. Research into alternative measles vaccines in the 1990s. Unpublished WHO document EPI/ GEN/88. 11. Rev. 1. Bulletin of the World Health Organization, 69(2): 221-227 (1991) © World Health Organization 1991 221 P. Kiepiela et al. Table 1: Age range of the study children at vaccination No. of children (age range in weeks) Vaccination group >16 to 23 >23 to 28 >28 to 36 >36 to 40 >40 to 44 Total Aa 26 14 8 - - 48 Bb 20 17 1 1 - - 48 cc - - - 27 1 28 a Edmonston-Zagreb vaccine. b Schwarz vaccine given at 636 weeks of age. c Schwarz vaccine given at >36 weeks of age. Guinea-Bissau (13, 14)) are, however, consider- able. We therefore carried out a study with the E-Z vaccine in a southern African setting to determine the serological responses and side-effects and to measure the clinical protection afforded to children given the E-Z vaccine (dose, 3.9 log 50 infectious units) or the conventional Schwarz vaccine (dose, 3.28 log 50 infectious units) at 16-34 weeks of age. A group of children who received the Schwarz vaccine (3.28 log 50 infectious units) at >9 months of age served as the controls. The study was undertaken during a measles epidemic in order to enhance the possibility of measuring the protection given, de- spite a small sample size. Subjects and methods The study was approved by the Ethics Committee of the Faculty of Medicine, University of Natal. In- formed consent was obtained from the parent(s) or guardian of each child after explanation of the object of the trial and the number of blood specimens required. Location The study was carried out in Qwa-Qwa, an impover- ished rural area of South Africa, with little industry and no agriculture (area, 480 kmi2; estimated popu- lation, 385 145 in 1988, 2.9% of whom were aged <1 year; average number of persons per household, 7). Study design The study began on 1 February 1988 with the random immunization (using a table of random numbers) of 96 children aged 4-8.5 months: 48 received the E-Z vaccine (3.91og50 infectious units, group A) and 48 the Schwarz vaccine (3.28log50 infectious units, group B); 28 children aged >9 months who received the Schwarz vaccine (3.28log50 infectious units, group C) (Table 1) served as controls. Blood samples were obtained from the children by venepuncture before vaccination, and at 6-8 weeks, 5-9 months, and 10-14 months post-vaccina- tion. The mother or guardian was instructed to bring the child to hospital if any adverse effects such as fever, rash, cough, conjunctivitis, coryza, diarrhoea, vomiting, ear or skin infections, or neurological complications occurred within 2 weeks of immuniza- tion. The children were visited monthly by a health care worker, who also provided parents with health education information, especially advice on nutri- tion, and, if necessary, were subsequently seen by one of the authors, who prescribed treatment. Vaccines used The vaccines outlined below were used in the study. * E-Z live, attenuated vaccine (lot No. 84/2; date of manufacture, December 1986; expiry date, December 1988; titre prior to vaccination (manu- facturer's data), -3.0 log 50 infectious units, and from field samples at the end of the study, 3.9log5O)b grown in human diploid cells.C The vaccine was reconstituted in distilled water and a 0.5-ml dose was given subcutaneously. * Schwarz live, attenuated measles vaccine (lot No. M115N41A; date of manufacture, November 1986; expiry date, November 1988; titre prior to study (manufacturer's data) ¢3.0 log 50 infectious units, and from field samples after vaccination 3.28 log 50 grown in chick embryo fibroblasts. d This vaccine was also given subcutaneously as a 0.5-ml dose. Serology Measles IgG antibody was detected using an en- zyme-linked immunosorbent assay (ELISA) kit.e b Institute of Immunology, Zagreb, Yugoslavia. c Diplovax, African Subequatorial Pharmaceuticals. d Rimevax. Smith Kline, Rixensort, Belgium. * Behring, Frankfurt, Germany. 222 WHO Bulletin OMS. Vol 69 1991. Lack of efficacy of Edmonston-Zagreb measles vaccine in African infants ELISA plates coated with measles antigen were supplied from the same batch (OSOK 02103, lot No. 406446A), as was the anti-human IgG-alkaline phosphatase conjugate (OSDH 04/05, lot No. 410955A) and supplementary agents (lot 19604). Results are expressed as mIU/ml, with respect to WHO standard anti-measles serum. An absorbance of <0.2 (equivalent to <34 mIU/ml) was considered negative. The pre- and post-vaccination sera of each child and the respective mother's serum were assayed in duplicate and always included in the same tests. The same number of samples from each vaccine group was assayed simultaneously to avoid any bias. As controls, international and local standards (two of which were kindly donated by Dr R. Glass, Centers for Disease Control) were always included in each batch. Seroresponse was defined according to the persistence of antibody (12), while seropositivity was defined as >200 mIU/ml at any time post-vaccination (Dr N. Halsey, personal com- munication, 1989). Antibody levels are reported as follows: - % of seropositives at 6-8 weeks and at 40-56 weeks post-vaccination; - % of seropositives at 6-56 weeks post-vaccina- tion, according to age at vaccination; and - % of seropositives at 6-56 weeks post-vaccina- tion, according to both pre-vaccination (i.e., maternal) antibody status and the age at vaccina- tion. Statistical methods The relationship between the level of measles IgG in mIU/ml and the absorbances was determined using a polynomial regression. The relationship between \/iHW7il and the absorbance readings was de- scribed very well by the equation: mIU/ml = 3.0031 + 12.794 (absorbance) + 5.791 (absorbance)2. The adjusted coefficient of determination was R = 0.97, i.e., 97% of the variation in \/IU7il is accounted for by the variation in absorbance. The fit for the above equation was limited to the observed range of absorbance values (0-1.5). The mIU/ml value for a single specimen was calculated by squaring the result obtained after substitution of the appropriate absorbance value into the above equation (control runs were under- taken with each run). Results are reported to the nearest integer. Results Rate of follow-up Most children returned for the 6-8 weeks post-vac- cination examination (41/48 for group A; 46/48 for group B; and 25/28 for group C). This decreased to 28/48, 30/48, and 23/28 for groups A, B, and C, respectively, for the examination at 5-9 months post-vaccination, and to 17/48, 21/48, and 18/28, respectively, for the third examination at 10-14 months post-vaccination. At least two follow-up samples of blood were obtained from no less than 90% of the sample population. Health monitoring A total of 93/124 children (33/48 in each of groups A and B, and 27/28 in group C) were monitored monthly by the health care worker. Development of measles. During 26 months post- vaccination follow-up, one of the children in group A who was seronegative developed measles 19 months after vaccination. Monitoring of the study children is continuing. Children who were persist- ently seronegative were revaccinated with the Schwarz vaccine after samples of their blood had been analysed, i.e., 16 months after the initial vac- cination. Deaths. There were two deaths in group A: a 6-month-old male who died of unknown causes (pre-immunization antibody level, 224 mIU/ml); and a male who died of pneumonia aged 15 months (pre-immunization antibody level, 25 mIU/ml). The latter child had persistent measles antibody. A 6-month-old male who had diarrhoea died in group B (pre-immunization antibody level 55 mIU/ml). There were no deaths in group C. None of the deaths was associated with measles. Seropositivity Seroresponse 6 weeks to 14 months post-vaccination. The proportion of seropositives 6-8 weeks and 10-14 months post-vaccination are shown in Fig. 1 for groups A, B, and C. In group A two of the five children who underwent late seroresponse had had blood taken for antibody determination at 6-8 weeks post-vaccination. Similarly, two of three late seroresponders in both groups B and C had blood taken at the 6-8-week follow-up visit. All but three WHO Bulletin OMS. Vol 69 1991. 223 P. Kiepiela et al. Fig. 1. Seroresponse of study children at various periods after vaccination with Edmonston-Zagrob (group A) or Schwarz vaccine (groups B and C). Shown at the top of each bar is the number of seropositives/total number of children in the group. 75 50 25' group A E group B S group C 17/28 13/25 av- 4'U-cc Time after vaccination (weeks) (two in group A and one in group B) of the children who were late seroresponders were older than 23 weeks, and all except six (four in group A and two in group B) had a pre-vaccination antibody level of <40 mIU/ml. Seroresponse according to age at vaccination. As expected, the seropositivity rate was highest in group C (61% (17/28)); in group A and group B it was lowest for children aged >16 weeks to <23 weeks (group A, 31% (8/26) and group B, 25% (5/20)). For children aged >23 weeks to <36 weeks, the seropositivity rate increased to 68% (15/22) in group A and to 36% (10/28) in group B (Fig. 2). Seroresponse 6-56 weeks post-vaccination. Because the number of children in each study group is small, the conclusions drawn from the findings must be treated with reserve (Table 2). The general points outlined below can, however, be made. * Children in group C exhibited protective levels (61%), irrespective of their pre-vaccination anti- body titre. * For infants <23 weeks of age who were given either the E-Z or the Schwarz vaccine, the seropositivity rate was poor (28%), irrespective of their pre-vaccination antibody titre. * For low pre-immunization antibody levels (<40 mIU/ml), children aged >23 weeks to <36 weeks who received the E-Z exhibited better seropositivity than those who received the Schwarz vaccine (60% versus 29%). This advan- tage of the E-Z vaccine was maintained even if the pre-vaccination antibody levels were 40-90 mIU/ml. * Although measles antibody levels were high (112- 2235 mIU/ml) only two were <200 mIU/ml) in all the samples of maternal blood (n = 104) taken when the infants were vaccinated, seropositivity did not correlate with pre-vaccination antbody status. Maintenance of seroresponse 5-14 months post-vaccination The length of time that an antibody level ¢200 mIU/ml was maintained was determined for those children from whom at least three samples of blood were taken. A reduction in measles antibody levels to below the protective threshold was detected 5-11 months after vaccination in six of 12 children from group A, in five of 11 children from group B, and in four of 16 children from group C. Fig. 2. Post-vaccination seroresponse depending on age at vaccination with Edmonston-Zagreb (group A) or Schwarz vaccine (groups B and C). Shown at the top of each bar is the number of seropositives/total number of children in the group. 0 a2 , 16,5 23 > 23, s28 > 25, s 36 > 36, 5 40 > 40,. 44 Age range at vaccination (weeks) WHO Bulletin OMS. Vol 69 1991. 0 CL0 o, 224 Lack of efficacy of Edmonston-Zagreb measles vaccine in African infants Table 2: Seroconversion among the study children 6-56 weeks after vaccination, according to pre-vaccination antibody status and age at vaccination Pre-vaccination Proportion who seroconverted, by age at vaccination (weeks)b measles antibody Vaccination status (mlU/ml) groupa >16 to 23 >23 to 28 >28 to 36 >36 to 40 >40 to 44 Total A 1/8 (13)c 3/4 (75) 3/6 (50) - - 7/18 (39) <40 B 0/2 (0) 2/7 (29) 2/7 (29) - - 4/16 (25) C - - - 12/21 (57) - 12/21 (57) A 2/3 (67) 4/5 (80) - - - 6/8 (75)d >40 to 90 B 1/2 (50) 1/4 (25) 0/2 (0) - - 2/8 (25) C - - - 1/2 (50) - 1/2 (50) A 1/5 (20) 2/4 (50) - - - 3/9 (33) >90 to 199 B 3/8 (38) 1/3 (33) 0/1 (0) - 4/12 (33) C - - - 0/1 (0) 1/1 (100) 1/2 (50) A 4/10(40) 1/1 (100) 2/2(100) - - 7/13 (54) >200 B 1/8 (13) 3/3(100) 1/1 (100) - - 5/12 (42) C - - - 3/3 (100) - 3/3 (100) A = Edmonston-Zagreb vaccine; B= Schwarz vaccine given to infants s36 weeks of age; and C = Schwarz vaccine given to infants >36 weeks of age. b Shown are the number of seropositivesAotal number in each category. c Figures in parentheses are percentages. d A versus B S 36 weeks, P< 0.05, x2 = 4000. Discussion Our results indicate that the E-Z vaccine may be superior to the Schwarz vaccine for infants under 9 months of age, but that at a titre of 3.9 log 50 infectious units it does not produce a satisfactory seroresponse. We also investigated the effects on seroresponse of pre-vaccination antibody levels and age at vaccination. Age remained a major barrier to effective immunization: for infants aged <23 weeks who received either vaccine, the seropositivity rate was poor (28%), irrespective of pre-vaccination anti- body level. This was disappointing, since other studies of the E-Z vaccine have reported conversion rates of 84-100%, and protection against measles for infants in this age range. A better rate of seropositivity (68%), although unsatisfactory for effective control of measles, was achieved for infants in the age range >23 weeks to <36 weeks. Results from other studies suggest that both these deficien- cies with the E-Z vaccine could be overcome by using a higher titre formulation. The seropositivity rate produced by the Schwarz vaccine in infants aged >36 weeks was less than optimum (61%), and for children below this age the rate was too low to offer protection against measles. Although some children showed an inadequate antibody response, only one developed measles over the 26-months follow-up period. Since the attack rate of measles among susceptibles is almost 100%, the single case possibly arose because of an in- creased herd immunity effect (15); our findings show that fewer measles cases were reported in Qwa-Qwa during 1988 and 1989, although in the latter year this was due to underreporting. It is also likely that, in addition to antibody, cellular factors are necessary for immunity against measles. Antibodies may have been detectable by the plaque neutralization test. For example, Black et al. have shown that many children who had low levels of antibody one year after having received measles vaccine were able to mount a secondary response to revaccination (15). In view of the regular monitoring and reliable recognition of measles by the local population, it is unlikely that any cases were missed during fol- low-up. In some children a delayed measles antibody response after vaccination can arise because of their young age, persistence of maternal antibody, or malnutrition (4, 17). We detected such a delay in 11 children for whom none of these factors were directly implicated. Several explanations have been proposed for the greater seroconversion rates after vaccination of infants aged 7-11 months in developing countries compared with those in industralized countries (18). The contribution of prematurity can be discounted since such children were excluded from the study. Furthermore, children in developing countries ex- perience immuno-acceleration caused by frequent and occasionally serious infections at an extremely young age (19). WHO Bulletin OMS. Vol 69 1991. 225 P. Kiepiela et al. Black et al. have documented that in most developing countries women have similar antibody levels to those of women in the USA (20). All the mothers in our study were from a low-income group and had high levels of measles antibody. Despite this, the number of seropositives had no bearing on the pre-vaccination antibody status. However, our results cannot strictly be compared with those of Black et al. since we used a different test to measure the antibody level and reported the results in mIU/ml and not as titres. Nutritional status and subsequent illnesses did not influence the seropositivity rates, since in all three study groups equal numbers of well-nourished and undernourished children responded (data not included). This is in accord with observations made by other workers (4, 18). Inadequate immunity to measles has been demonstrated in children vaccinated at an early age (16). Our data tend to support this finding, since fewer children who were vaccinated with the Schwarz vaccine when they were less than 36 weeks of age maintained their serological status. The E-Z vaccine is as safe as the Schwarz vaccine. Furthermore, an equal number of children in groups A and B who were seropositive still had protective levels of antibody 5-14 months post-vac- cination. Both the E-Z vaccine at standard potency and the Schwarz vaccine appeared to offer clinical protection against measles during the follow-up period. Our study confirms the results reported by Khanum et al. (5), and this endorses the recent WHO recommendation that titres greater than standard potency vaccine are needed if 6-month-olds are to be successfully immunized against measles (21). Acknowledgements We thank Mrs E. Steyn for typing the manuscript; Mrs E. Phule for the follow-up data; and Mr J. Swanepoel for his invaluable help in setting up the computer program for analysing the ELISA results. Professor J. Moodie is thanked for kindly determining the titres of field samples of E-Z vaccine. Our study was supported entirely by a grant from the Anglo American Corporation. Resume Vaccin antirougeoleux vivant att6nu6 Edmonston-Zagreb: manque d'efficacit6 de la dose standard chez les nourrissons afri- cains On a teste l'efficacite de la dose standard du vaccin antirougeoleux Edmonston-Zagreb (E-Z) au cours d'un essai randomise men6 chez des nourrissons noirs d'une r6gion rurale d'Afrique du Sud a l'occa- sion d'une epidemie de rougeole. Le calendrier vac- cinal suivant a e applique: 48 nourrissons de 4 a 8,5 mois ont requ 3,91og50 unites infectieuses de vaccin E-Z (groupe A); 48 nourrissons de 4 a 8,5 mois ont requ 3,280og50 unit6s infectieuses de vac- cin Schwarz (groupe B); et 28 nourrissons de plus de 9 mois ont requ 3,281og50 unit6s infectieuses de vaccin Schwarz (groupe C) et ont servi de temoins. L'efficacite du vaccin a e mesuree par le degre de protection contre la rougeole et par les r6ponses en anticorps seriques, d6termin6es par titrage immu- noenzymatique (ELISA). On a evalue les effets de l'age au moment de la vaccination et de la presence d'anticorps avant celle-ci. Un controle a ensuite e effectu6 dans 90% des cas sur au moins deux 6chantillons de sang preleves entre 6 et 56 semaines apres la vaccination. Aucun des nourrissons ayant requ le vaccin Schwarz n'a eu la rougeole; en revanche, un des nourrissons vaccine avec le vaccin E-Z l'a contrac- t6e. Deux nourrissons du groupe A et un nourrisson d'un groupe Schwarz sont deced6s. Toutefois, ces deces etaient sans rapport avec la rougeole. Chez les nourrissons de moins de 23 semaines ayant requ soit le vaccin E-Z, soit le vaccin Schwarz, le nombre des seroconversions est faible (28%), quel que soit le taux d'anticorps antirougeo- leux avant vaccination. On a observe une s6rocon- version plus importante (68%) avec le vaccin E-Z qu'avec le vaccin Schwarz (36%) chez les nourris- sons ayant entre 23 et 36 semaines. Administr6 aux enfants de plus de 36 semaines, le vaccin Schwarz a donne un taux de reponse satisfaisant bien que sous-optimal (61%). Aucune corr6lation n'a pu 6tre etablie entre la s6roconversion et la presence d'anti- corps avant vaccination, dans aucun groupe. La dose standard de vaccin E-Z est peut-etre a l'origine de cette faible r6ponse, ce qui corroborerait la recommandation de l'OMS pr6conisant l'emploi de doses plus fortes pour vacciner correctement les enfants de 6 mois. References 1. Ferrinho, P. National workshop on measles immuniza- tion. South African medical joumal, 76: 82-83 (1989). 2. Measles immunity in the first year after birth and the optimum age for vaccination in Kenyan children. Colla- borative study by the Ministry of Health of Kenya and the World Health Organization. Bulletin of the World Health Organization, 55: 21-31 (1977). 226 WHO Bulletin OMS. Vol 69 1991. Lack of efficacy of Edmonston-Zagreb measles vaccine in African infants 3. Loening, W.E.K. & Coovadia, H.M. Age-specific occur- rence rates of measles in urban, peri-urban, and rural environments: implications for time of vaccination. Lan- cet, 2: 324-326 (1983). 4. Sabin, A.B. et al. Successful immunization of children with and without maternal antibody by aerosolized measles vaccine. 1. Different results with undiluted human diploid cell and chick embryo fibroblast vaccines. Journal of the American Medical Association, 249: 2651 -2662 (1983). 5. Khanum, S. et al. Comparison of Edmonston-Zagreb and Schwarz strains of measles vaccine given by aero- sol or subcutaneous injection. Lancet, 1: 150-153 (1987). 6. World Health Organization. Measles immunization be- fore the age of 9 months? Position statement by the Expanded Programme on Immunization of the WHO. Lancet, 2: 1356-1357 (1988). 7. Sabin, A.B. Immunization against measles by aerosol. Reviews of infectious diseases, 5: 514-523 (1983). 8. Sabin, A.B. et al. Successful immunization of infants with and without maternal antibody by aerosolized measles vaccine. Journal of the American Medical Association, 251: 2363-2371 (1984). 9. Beck, M. et al. Immune response to Edmonston- Zagreb measles virus strain in monovalent and com- bined MMR vaccine. Developments in biological standardization, 65: 95-100 (1986). 10. Whittle, H.C. et al. Immunisation of 4-6-month-old Gambian infants with Edmonston-Zagreb measles vac- cine. Lancet, 2: 834-837 (1984). 11. Whittle, H.C. et al. Effects of dose and strain of vaccine on success of measles vaccination of infants aged 4-5 months. Lancet, 1: 963-966 (1988). 12. Whittle, H.C. et al. Trial of high-dose Edmonston-Za- greb measles vaccine in the Gambia: antibody response and side-effects. Lancet, 2: 811-814 (1988). 13. Aaby, P. et al. Trial of high-dose Edmonston-Zagreb measles vaccine in Guinea-Bissau: protective efficacy. Lancet, 2: 809-814 (1988). 14. Aaby, P. Child mortality related to seroconversion or lack of seroconversion after measles vaccination. Pe- diatric infectious diseases journal, 8: 197-200 (1989). 15. Black, F.L. The role of herd immunity in control of measles. Yale journal of biology and medicine, 55: 351-360 (1982). 16. Black, F.L. et al. Inadequate immunity to measles in children vaccinated at an early age: effect of revaccina- tion. Bulletin of the World Health Organization, 62: 315-319 (1984). 17. Wesley, A. et al. Immunization against measles in children at risk for severe disease. Transactions of the Royal Society of Tropical Medicine and Hygiene, 73: 710-714 (1979). 18. Halsey, N.A. et al. Response to measles vaccine in Haitian infants 6 to 12 months old. Influence of maternal antibodies, malnutrition, and concurrent illnesses. New Englandjournal of medicine, 331: 544-549 (1985). 19. Kiepiela, P. et al. Age-related lymphocyte sub-popula- tion changes among healthy Africans from birth to adulthood. Annals of tropical paediatrics, 9: 199-205 (1989). 20. Black, F.L. et al. Geographic variation in infant loss of maternal measles antibody and in prevalence of rubella antibody. American journal of epidemiology, 124: 442-452 (1986). 21. EPI Global Advisory Group: Part I. Weekly epidemio- logical record, 65(2): 5-10 (1990). WHO Bulletin OMS. Vol 691991. 227