494
TRANSACTIONSOF THE ROYAL SOCIETYOF TROPICALMEDICINE AND HYGIENE (1992) 86, 49&498
Morbidity from malaria and immune responses to defined Plasmodium falciparum antigens in children with sickle cell trait in The Gambia S. J. Allen’, S. Bennett2, E. M. Riley’, P. A. Rowe2, P. H. Jakobsen3, A. O’Donnell’ and B. M. Greenwood’ ‘Medical Research Council Laboratories, Fajara, The Gambia; 2Tropical Health Epidemiology Unit, London School of Hygiene and Tropical Medicine, London, UK; 3Department of Infection-Immunology, Division of Biotechnology, Statens Seruminstitut, Copenhagen, Denmark
Abstract Morbidity from Plasmodium falcivarum malaria and humoral and in vitro cellular immune resnonses to defined malaria antigens were measured in rural Gambian children with haemoglobin phenotype AS (HbAS) and in those with a normal haemoglobin (HbAA). In a survev undertaken during: the drv season. HbAS children had a higher parasite rate than HbAA children but a lower prevalence ofparasitakmia at a level of 50041 or greater. Malariometric indices measured during a rainy season survey were similar in the 2 grouns of children. During the rainy season, the incidence of infection with I’. falcinarum did not varv with haemoglobin phenotype. However; in children aged 6 years or less, a signiticantiy smaller propo&on of HbAS children who acauired infection develoned clinical svmutoms than did HbAA children. During both the dry season and rainy season surveys, hum&al and in vi&o iellular immune responses to defined antigens from the sporozoite and merozoite stages of P. falciparum were similar in the 2 groups of children. Thus, despite the differences in parasite indices and morbidity from malaria between the 2 groups of children, we found no evidence of an enhanced immune response to malaria infection amongst HbAS children compared with normal children
Introduction Many studies performed in malaria endemic areas have demonstrated that children who carry the sickle cell trait suffer less morbiditv and mortalitv from Plasmodium falciparum malaria than normal children (reviewed by l&NGELHANN et aZ., 1976). The mechanism(s) which underlie(s) this relative protection against morbidity from malaria shown by subjects with haemoglobin (Hb) AS have not been fully delineated. Protection may result from reduced develonment of narasites in ervthrocvtes containing sickle haemoglobin- compared with normal erythrocytes (LUZATTO et al., 1970; ROTH et al., 1978). Alternatively, protection may be effected by an enhanced host immune response against I’. falciparum in HbAS compared with HbAA children (RINGELHANN et al., 1976; BAYOUMI, 1987). To investigate this latter possibility, we measured a wide range of humoral and in vitro cell-mediated immune responses to defined malaria antigens in HbAS children and in children with HbAA.
Subjects and Methods Study area and subjects Details of the- study area, subjects and laboratory methods have been nublished nreviouslv (RILEY et al.. 1990). Briefly, children aged between 3 and 8 years who were resident in 2 groups of villages, the Fula and Kataba hamlets, situated near to the town of Farafenni on the north bank of the river Gambia, were studied. The area is one of dry Sudan savanna where intense malaria transmission occurs during a 3 month period at the end of the rainy season-September to November (GREENWOOD et al., 1987). Half of the inhabitants of these villages are of the Fula ethnic group, with the remainder divided approximately equally between Wollofs and Mandinkas. A trial of the protective effect of insecticideimpregnated bed nets against malaria was performed in some of the Fula villages during the year preceding this study (SNOW et al., 1988). Euidemioloaical methods -All childyen were examined during surveys held at the beginning (May) and end (October) of the rainy season. Any enlargement of the spleen was recorded and thick and thin blood films were examined for malaria parasites. Children found to have parasitaemia were given curative treatment. Venous blood samples were collected for the Address for correspondence: Dr S. J. Allen, Department of Tropical Paediatrics and International Child Health, Liverpool School of Tropical Medicine, Pembroke Place, Liverpool, L3 5QA, UK.
measurement of immune responses to control and malaria antigens and for the determination of haemoglobin phenotype. During the rainy season children were visited weekly by a field worker to determine morbiditv from malaria. A clinical episode of malaria was diagnosed when an episode of fever (axillarv temuerature >37.5”C) was associated with P. falciparum parasitaemia at a level of 50001u.1 or greater and these children were treated with chloroquine. Children in whom parasitaemia was detected during the rainy season or at the October survey, or in whom spleen size had increased significantly between the 2 surveys, but in whom a clinical episode of malaria had not been recorded, were said to have acquired asymptomatic infection. For children in the Kataba villages, individual exposure to malaria infection was assessed by performing fortnightly searches of their bed nets for blood-fed Anopheles gambiae mosquitoes. Laboratory methods Antibodies to whole malaria parasites were measured using a standard immunofluorescent antibody test (IFAT) (RILEY et al., 1990). Antibodies binding to the surface of ring-infected erythrocytes were detected by indirect erythrocyte immunofluorescence (EMIF) (RILEY et al., 1991a). Antibodies to soluble exoantigens were measured in a semi-quantitative manner by crossed immunoelectrophoresis (JAKOBSEN et al., 1990). Enzyme-linked immunosorbent assays (ELISAS) were used to measure antibodies to the synthetic polymer (NANQa of the circumsporozoite protein (CSP) (RILEY et al., 1990); to specific B cell epitopes from the Pfl55IRESA molecule (RILEY et al., 1991a); to peptides p19OL and p190N representing conserved regions of the major merozoite surface antigen of I’. falciparum (PfMSPl); and to the BfMSBlss and PfMPS14z processing products of this antigen (RILEY et al., 1992). Defined I’. falciparum antigens employed in cellular assays are listed in Table 1. Donor mononuclear cells were incubated with each antigen in triplicate wells. Proliferative responses were measured by uptake of tritiated thymidine and expressed as the mean log stimulation index (LSI), and the proportion of children who responded to each antigen was calculated (RILEY et al., 1990). Purified vrotein derivative of Mvcobacterium tuberculos~s (PPD) a;d Candida albicans exiract were used as control antigens and phytohaemagglutinin (PHA) was used as a mitogen to test the viability of cell cultures. Interferon y (IFNy) production of donor cells in response
495 Table
1. Defined
P. falcipanun
Strain Circumsporozoite protein Wellcome LE5 7G8 Wellcome LE5 7G8 Pfl55lRESA
peptides
Description Th2R Th3R
S’repeat 3’repeat
Major merozoite surface antigens PFM(:!i) Wellcome PMEl PME2 PME6 PME7 PMEll PME14 P19OL P190N Soluble exoantigens F32
F32
Ad 49
employed
in cellular
assays
Amino acid sequencea
Reference
PSDKHIEQYLKKIQNSLSTE PSDQHIEKYLKTIQNSLSTE PSDKHIEQYLKKIKNSIS IKPGSADKPKDQLDYENDIE IKPGSANKPKDQLDYANDIE IKPGSANKPKDELDYENDIE
RILEY et al. (1990)
I’1 (390-411) TVAEEHVEEPTVAEE P2 (635-649) SLRWIFKHVAKTHLK I’3 (796-8 13) DVVGYIMHGISTINKEMK P4 (861-876) EENVEHDAEENVEENV
RILEY et al. (1991a)
(154-1047) (1047-1640) ‘(182-388)’ (349494) (1210-1614) (901-1498) (188-363) (188-363 + 1141-1287)
RILEY et al. (1992)
mixed soluble affinity purified affinity purified
RILEY et al. (1991b)
aNumbers in parentheses indicate the amino acid position of peptides. bOnly interferon y responses to the whole molecule were measured. to defined malaria antigens was measured by an ELBA (RILEY et al., 1990). Becauseof a limited supply of reagents, IFNy production was measured in cells obtained from children resident in the Kataba hamlets only. Control blood samples for the measurement of immune responsesto malaria antigens were collected from 30 Swedish children aged 3 to 10 years who had no previous exposure to malaria infection. Haemoglobin phenotype was determined by electrophoresis on cellulose acetatepaper in a Shandon electrophoresis tank using a discontinuous buffer system (Trisethylenediaminetetraacetateiborate). Statistical methods
Parasite indices were compared in HbAS and HbAA children by the Mantel-Haenszel test, stratifying for age. Morbidity from malaria in the 2 groups of children was compared by multiple logistic regression analysis allowing for the confounding variables of age, sex, ethnic group and, in the caseof children from the Fula hamlets, the use of insecticide-impregnated bed nets during the year before the study. For humoral and cellular responses, Student’s t test was used for continuous variables and the x2 test for categorical data. Ethical permission for both the Gambian and Swedish studies was obtained from the relevant ethical review committees. Results
Haemoglobinphenotype Haemoglobin phenotype was determined in 389 children. 309 children (79.4%) had HbAA, 76 (19.5%) had HbAS,, 3 had phenotype SC, and one child had phenotype SS. The distribution of haemoglobin phenotypes did not vary significantly with age or sex (data not shown). The prevalence of HbAS was greater in Mandinka children (25196; 26.0%) than in Wollof (21197, 21.7%) or Fula children (301196, 15.3%), but these dif-
ferenceswere not statistically significant (ns) (x*=5.09, 2 degreesof freedom [dfl). No significant difference was detected between the geometric mean (pm) number of blood-fed An. gambiae mosquitoes collected at each visit from the nets of HbAA children (gm=0.25) and HbAS children (gm=O.32; t test on ln(n+ 1): t=0.02, ns). Morbidity from malaria
During the dry seasonsurvey (May), the P. falciparum parasite rate was significantly greater in HbAS than HbAA children (Table 2). However, amongst children with parasitaemia, the incidence of parasitaemia at a level of 5OO/ulor greater was significantly lower in HbAS than HbAA children. A similar tendency was observed for higher levels of parasitaemia (5OOO/ulor greater) but, in this case,numbers were small and the difference between the groups was not statistically significant. This pattern of infection was present in children of all ages (data not shown). During the rainy seasonsurvey (October), when the overall prevalence of parasitaemia was much greater, no significant difference in parasite rate and density, or in the proportion of children who had acquired splenomegaly, was detected between the 2 groups. The proportion of children who acquired malaria infection during the rainy season was similar in HbAA children (209/281, 74.4%) and HbAS children (49/70, 70.0%). However, multiple logistic regression analysis showed that a lower proportion of HbAS children experienced one or more episodes of fever and high parasitaemia (>50001ul) than did HbAA children (19/49, 38.8% and 1031209,49.3% respectively; P5OO/ul >5OOO/yl Spleen rate Rainy season survey P. falciparum parasitaemia Any parasitaemia >5OOiyl >5OOO/yl
No.
HbAS No.
(n=309)
(n=76)
79 (25.6%)
(n=257)
13 (17.1%) (n=64)
105 (40.9%) 74 (70.5%)
29 (45.3%) 18 (62.1%)
Spleen rate Acquired splenomegalyd
38 (13.8%)
2a
X
10.32 4~27~ -
P
co.01 co.05 nsc ns
-
ns ns ns
-
ns ns
“Mantel-Haenszel test allowing for age. bTest for proportion with parasitaemia >5OO/yl out of those with any parasitaemia. ‘ns=Not significant. dSpleenpalpable during October survey but not in May, or an increase in spleen size of ~2 cm between the 2 surveys. An increase in spleen size of 1 cm between the 2 surveys was not considered to be clinically significant.
HbAA n - 94
HbAS 30
HbAA
HbAS
HbAA
HbAS
99
22
aa
18
25%
7-a
5-6
AGE (years) I
fever . high parasitaemia
u&Y ww
t IOW
paraaitaemia
0
asymptomstic infectmn
m
no *“idmOe of infection
Figure. Morbidity from malaria in Gambian children according to haemoglobin phenotype and age.
asexual blood stagesof P. falciparum between HbAS and HbAA children either at the May survey (a=75 and 306 respectively, median l/l0240 in both groups; Wilcoxon test, z=-0.22) or at the October survey (a=55 and 232 respectively, median= 1120480in both groups; Wilcoxon test, z=-1.058). A greater proportion of HbAA than HbAS children was seropositive during the May survey for antibody to the CSP repeat (NANlQs (351252,13.9% and 2167, 3.0% respectively; x2=5.12, ZVO.05) and to the RfMSRlg3 fragment of PfMSPl (261307,8.5% and 1176, 1.3% respectively; x2=3.37, P=O.O5). There was no significant difference in the proportion of children with either phenotype who were seropositive with regard to any of the other antigens tested. Measurements of humoral responses to CSP, Pfl55/RESA and PfMSPl antigens were repeated during the rainy season survey. The prevalence of responses tended to be higher than that at the dry seasonsurvey (for example 9119(47.4%) HbAS and 20/54 (37.0%) HbAA children responded to the (NAN&c epitope). No significant difference between HbAS and HbAA children in the prevalence of seropositivity
I’HA, control and malaria antigens measured during the dry seasonsurvey did not differ significantly according to haemoglobin phenotype (see Table 3). During the rainy season survey, mean LSI to the P4 antigen of Pfl55IRESA was significantly greater for HbAS than HbAA children (n=l8, mean LSI=O.45, standard error=0.17 versus 58, -0.02 and 0.10 respectively; t=2.31, P
TRANSACTIONSOF THE ROYAL SOCIETYOF TROPICALMEDICINE AND HYGIENE (1992) 86, 49&498
Morbidity from malaria and immune responses to defined Plasmodium falciparum antigens in children with sickle cell trait in The Gambia S. J. Allen’, S. Bennett2, E. M. Riley’, P. A. Rowe2, P. H. Jakobsen3, A. O’Donnell’ and B. M. Greenwood’ ‘Medical Research Council Laboratories, Fajara, The Gambia; 2Tropical Health Epidemiology Unit, London School of Hygiene and Tropical Medicine, London, UK; 3Department of Infection-Immunology, Division of Biotechnology, Statens Seruminstitut, Copenhagen, Denmark
Abstract Morbidity from Plasmodium falcivarum malaria and humoral and in vitro cellular immune resnonses to defined malaria antigens were measured in rural Gambian children with haemoglobin phenotype AS (HbAS) and in those with a normal haemoglobin (HbAA). In a survev undertaken during: the drv season. HbAS children had a higher parasite rate than HbAA children but a lower prevalence ofparasitakmia at a level of 50041 or greater. Malariometric indices measured during a rainy season survey were similar in the 2 grouns of children. During the rainy season, the incidence of infection with I’. falcinarum did not varv with haemoglobin phenotype. However; in children aged 6 years or less, a signiticantiy smaller propo&on of HbAS children who acauired infection develoned clinical svmutoms than did HbAA children. During both the dry season and rainy season surveys, hum&al and in vi&o iellular immune responses to defined antigens from the sporozoite and merozoite stages of P. falciparum were similar in the 2 groups of children. Thus, despite the differences in parasite indices and morbidity from malaria between the 2 groups of children, we found no evidence of an enhanced immune response to malaria infection amongst HbAS children compared with normal children
Introduction Many studies performed in malaria endemic areas have demonstrated that children who carry the sickle cell trait suffer less morbiditv and mortalitv from Plasmodium falciparum malaria than normal children (reviewed by l&NGELHANN et aZ., 1976). The mechanism(s) which underlie(s) this relative protection against morbidity from malaria shown by subjects with haemoglobin (Hb) AS have not been fully delineated. Protection may result from reduced develonment of narasites in ervthrocvtes containing sickle haemoglobin- compared with normal erythrocytes (LUZATTO et al., 1970; ROTH et al., 1978). Alternatively, protection may be effected by an enhanced host immune response against I’. falciparum in HbAS compared with HbAA children (RINGELHANN et al., 1976; BAYOUMI, 1987). To investigate this latter possibility, we measured a wide range of humoral and in vitro cell-mediated immune responses to defined malaria antigens in HbAS children and in children with HbAA.
Subjects and Methods Study area and subjects Details of the- study area, subjects and laboratory methods have been nublished nreviouslv (RILEY et al.. 1990). Briefly, children aged between 3 and 8 years who were resident in 2 groups of villages, the Fula and Kataba hamlets, situated near to the town of Farafenni on the north bank of the river Gambia, were studied. The area is one of dry Sudan savanna where intense malaria transmission occurs during a 3 month period at the end of the rainy season-September to November (GREENWOOD et al., 1987). Half of the inhabitants of these villages are of the Fula ethnic group, with the remainder divided approximately equally between Wollofs and Mandinkas. A trial of the protective effect of insecticideimpregnated bed nets against malaria was performed in some of the Fula villages during the year preceding this study (SNOW et al., 1988). Euidemioloaical methods -All childyen were examined during surveys held at the beginning (May) and end (October) of the rainy season. Any enlargement of the spleen was recorded and thick and thin blood films were examined for malaria parasites. Children found to have parasitaemia were given curative treatment. Venous blood samples were collected for the Address for correspondence: Dr S. J. Allen, Department of Tropical Paediatrics and International Child Health, Liverpool School of Tropical Medicine, Pembroke Place, Liverpool, L3 5QA, UK.
measurement of immune responses to control and malaria antigens and for the determination of haemoglobin phenotype. During the rainy season children were visited weekly by a field worker to determine morbiditv from malaria. A clinical episode of malaria was diagnosed when an episode of fever (axillarv temuerature >37.5”C) was associated with P. falciparum parasitaemia at a level of 50001u.1 or greater and these children were treated with chloroquine. Children in whom parasitaemia was detected during the rainy season or at the October survey, or in whom spleen size had increased significantly between the 2 surveys, but in whom a clinical episode of malaria had not been recorded, were said to have acquired asymptomatic infection. For children in the Kataba villages, individual exposure to malaria infection was assessed by performing fortnightly searches of their bed nets for blood-fed Anopheles gambiae mosquitoes. Laboratory methods Antibodies to whole malaria parasites were measured using a standard immunofluorescent antibody test (IFAT) (RILEY et al., 1990). Antibodies binding to the surface of ring-infected erythrocytes were detected by indirect erythrocyte immunofluorescence (EMIF) (RILEY et al., 1991a). Antibodies to soluble exoantigens were measured in a semi-quantitative manner by crossed immunoelectrophoresis (JAKOBSEN et al., 1990). Enzyme-linked immunosorbent assays (ELISAS) were used to measure antibodies to the synthetic polymer (NANQa of the circumsporozoite protein (CSP) (RILEY et al., 1990); to specific B cell epitopes from the Pfl55IRESA molecule (RILEY et al., 1991a); to peptides p19OL and p190N representing conserved regions of the major merozoite surface antigen of I’. falciparum (PfMSPl); and to the BfMSBlss and PfMPS14z processing products of this antigen (RILEY et al., 1992). Defined I’. falciparum antigens employed in cellular assays are listed in Table 1. Donor mononuclear cells were incubated with each antigen in triplicate wells. Proliferative responses were measured by uptake of tritiated thymidine and expressed as the mean log stimulation index (LSI), and the proportion of children who responded to each antigen was calculated (RILEY et al., 1990). Purified vrotein derivative of Mvcobacterium tuberculos~s (PPD) a;d Candida albicans exiract were used as control antigens and phytohaemagglutinin (PHA) was used as a mitogen to test the viability of cell cultures. Interferon y (IFNy) production of donor cells in response
495 Table
1. Defined
P. falcipanun
Strain Circumsporozoite protein Wellcome LE5 7G8 Wellcome LE5 7G8 Pfl55lRESA
peptides
Description Th2R Th3R
S’repeat 3’repeat
Major merozoite surface antigens PFM(:!i) Wellcome PMEl PME2 PME6 PME7 PMEll PME14 P19OL P190N Soluble exoantigens F32
F32
Ad 49
employed
in cellular
assays
Amino acid sequencea
Reference
PSDKHIEQYLKKIQNSLSTE PSDQHIEKYLKTIQNSLSTE PSDKHIEQYLKKIKNSIS IKPGSADKPKDQLDYENDIE IKPGSANKPKDQLDYANDIE IKPGSANKPKDELDYENDIE
RILEY et al. (1990)
I’1 (390-411) TVAEEHVEEPTVAEE P2 (635-649) SLRWIFKHVAKTHLK I’3 (796-8 13) DVVGYIMHGISTINKEMK P4 (861-876) EENVEHDAEENVEENV
RILEY et al. (1991a)
(154-1047) (1047-1640) ‘(182-388)’ (349494) (1210-1614) (901-1498) (188-363) (188-363 + 1141-1287)
RILEY et al. (1992)
mixed soluble affinity purified affinity purified
RILEY et al. (1991b)
aNumbers in parentheses indicate the amino acid position of peptides. bOnly interferon y responses to the whole molecule were measured. to defined malaria antigens was measured by an ELBA (RILEY et al., 1990). Becauseof a limited supply of reagents, IFNy production was measured in cells obtained from children resident in the Kataba hamlets only. Control blood samples for the measurement of immune responsesto malaria antigens were collected from 30 Swedish children aged 3 to 10 years who had no previous exposure to malaria infection. Haemoglobin phenotype was determined by electrophoresis on cellulose acetatepaper in a Shandon electrophoresis tank using a discontinuous buffer system (Trisethylenediaminetetraacetateiborate). Statistical methods
Parasite indices were compared in HbAS and HbAA children by the Mantel-Haenszel test, stratifying for age. Morbidity from malaria in the 2 groups of children was compared by multiple logistic regression analysis allowing for the confounding variables of age, sex, ethnic group and, in the caseof children from the Fula hamlets, the use of insecticide-impregnated bed nets during the year before the study. For humoral and cellular responses, Student’s t test was used for continuous variables and the x2 test for categorical data. Ethical permission for both the Gambian and Swedish studies was obtained from the relevant ethical review committees. Results
Haemoglobinphenotype Haemoglobin phenotype was determined in 389 children. 309 children (79.4%) had HbAA, 76 (19.5%) had HbAS,, 3 had phenotype SC, and one child had phenotype SS. The distribution of haemoglobin phenotypes did not vary significantly with age or sex (data not shown). The prevalence of HbAS was greater in Mandinka children (25196; 26.0%) than in Wollof (21197, 21.7%) or Fula children (301196, 15.3%), but these dif-
ferenceswere not statistically significant (ns) (x*=5.09, 2 degreesof freedom [dfl). No significant difference was detected between the geometric mean (pm) number of blood-fed An. gambiae mosquitoes collected at each visit from the nets of HbAA children (gm=0.25) and HbAS children (gm=O.32; t test on ln(n+ 1): t=0.02, ns). Morbidity from malaria
During the dry seasonsurvey (May), the P. falciparum parasite rate was significantly greater in HbAS than HbAA children (Table 2). However, amongst children with parasitaemia, the incidence of parasitaemia at a level of 5OO/ulor greater was significantly lower in HbAS than HbAA children. A similar tendency was observed for higher levels of parasitaemia (5OOO/ulor greater) but, in this case,numbers were small and the difference between the groups was not statistically significant. This pattern of infection was present in children of all ages (data not shown). During the rainy seasonsurvey (October), when the overall prevalence of parasitaemia was much greater, no significant difference in parasite rate and density, or in the proportion of children who had acquired splenomegaly, was detected between the 2 groups. The proportion of children who acquired malaria infection during the rainy season was similar in HbAA children (209/281, 74.4%) and HbAS children (49/70, 70.0%). However, multiple logistic regression analysis showed that a lower proportion of HbAS children experienced one or more episodes of fever and high parasitaemia (>50001ul) than did HbAA children (19/49, 38.8% and 1031209,49.3% respectively; P5OO/ul >5OOO/yl Spleen rate Rainy season survey P. falciparum parasitaemia Any parasitaemia >5OOiyl >5OOO/yl
No.
HbAS No.
(n=309)
(n=76)
79 (25.6%)
(n=257)
13 (17.1%) (n=64)
105 (40.9%) 74 (70.5%)
29 (45.3%) 18 (62.1%)
Spleen rate Acquired splenomegalyd
38 (13.8%)
2a
X
10.32 4~27~ -
P
co.01 co.05 nsc ns
-
ns ns ns
-
ns ns
“Mantel-Haenszel test allowing for age. bTest for proportion with parasitaemia >5OO/yl out of those with any parasitaemia. ‘ns=Not significant. dSpleenpalpable during October survey but not in May, or an increase in spleen size of ~2 cm between the 2 surveys. An increase in spleen size of 1 cm between the 2 surveys was not considered to be clinically significant.
HbAA n - 94
HbAS 30
HbAA
HbAS
HbAA
HbAS
99
22
aa
18
25%
7-a
5-6
AGE (years) I
fever . high parasitaemia
u&Y ww
t IOW
paraaitaemia
0
asymptomstic infectmn
m
no *“idmOe of infection
Figure. Morbidity from malaria in Gambian children according to haemoglobin phenotype and age.
asexual blood stagesof P. falciparum between HbAS and HbAA children either at the May survey (a=75 and 306 respectively, median l/l0240 in both groups; Wilcoxon test, z=-0.22) or at the October survey (a=55 and 232 respectively, median= 1120480in both groups; Wilcoxon test, z=-1.058). A greater proportion of HbAA than HbAS children was seropositive during the May survey for antibody to the CSP repeat (NANlQs (351252,13.9% and 2167, 3.0% respectively; x2=5.12, ZVO.05) and to the RfMSRlg3 fragment of PfMSPl (261307,8.5% and 1176, 1.3% respectively; x2=3.37, P=O.O5). There was no significant difference in the proportion of children with either phenotype who were seropositive with regard to any of the other antigens tested. Measurements of humoral responses to CSP, Pfl55/RESA and PfMSPl antigens were repeated during the rainy season survey. The prevalence of responses tended to be higher than that at the dry seasonsurvey (for example 9119(47.4%) HbAS and 20/54 (37.0%) HbAA children responded to the (NAN&c epitope). No significant difference between HbAS and HbAA children in the prevalence of seropositivity
I’HA, control and malaria antigens measured during the dry seasonsurvey did not differ significantly according to haemoglobin phenotype (see Table 3). During the rainy season survey, mean LSI to the P4 antigen of Pfl55IRESA was significantly greater for HbAS than HbAA children (n=l8, mean LSI=O.45, standard error=0.17 versus 58, -0.02 and 0.10 respectively; t=2.31, P
