Seroepidemiological studies by indirect haemagglutination tests for malaria*

Bull. Org. mond. S 1969, 41, 825-841 Bull. Wid Hith Org. 199418281 Seroepidemiological Studies by Indirect Haemagglutination Tests for Malaria Military Recruit Collections from Argentina, Brazil, Colombia, and the United States of America * IRVING G. KAGAN,1 Ph.D., HENRY M. MATHEWS,2 Ph.D., WALLACE A. ROGERS, Jr,3 M. D. & JANET FRIED 4 Seroepidemiological studies of 10 956 sera of military recruits from 4 countries in the Western Hemisphere suggest that the indirect haemagglutination test employing an antigen prepared from Plasmodium knowlesi adsorbed to human group 0 erythrocytes may be useful in epidemiological studies of malaria. Based on a titre of 1: 8, or higher, in a micro- titration of serum, the prevalence of serological reactors in the collection of sera from USA recruits was 0.9 %, in the Argentina collection 4.6 %, in the Brazil collection 20.8 %, and in the Colombia collection 21.3 %. In Brazil and Colombia the distribution of sero- logical reactors correlated with the distribution of malaria as determined by active and passive surveillance. In Argentina, serological reactors were found in states where active and passive surveillance indicated that malaria had been eradicated years ago. Epidemiological studies for the prevalence of malaria are being made by the classical methods of fever surveys, spleen-index surveys, or blood-slide examination for parasites. At the present time the indirect fluorescent antibody (IFA) and indirect haemagglutination (IHA) tests hold much promise as serological tools for evaluating plasmodial antibody in individuals and in populations. The IFA test is admirably suited for diagnostic purposes but, until the technique can be simplified or mechanized, it cannot be seriously considered for large-scale epidemiological studies. The IHA test, used for * From the US Department of Health, Education, and Welfare, Public Health Service, Health Services and Mental Health Administration, National Communicable Disease Center, Parasitology Section, Atlanta, Ga. 30333, USA. This study was supported in part by the US State Department, Agency for International Development, Participating Agency Service Agreement (PASA) No. RA (HA) 5-68. 1 Chief, Parasitology Section, National Communicable Disease Center, Atlanta, Ga., USA. 2 Head, Malaria Laboratory, Parasitology Section, NCDC, Atlanta, Ga. 'Former Head, Malaria Laboratory, Parasitology Sec- tion, NCDC, Atlanta, Ga. Present address: Department of Pathology and Oncology, University of Vermont College of Medicine, Burlington, Vt. 05401, USA. 4Biological Laboratory Technician, Malaria Laboratory, Parasitology Section, NCDC, Atlanta, Ga. field studies by Desowitz & Saave (1965) in Australia and New Guinea, showed promise as an epidemio- logical tool. In our laboratory the variables asso- ciated with the indirect haemagglutination technique were carefully examined, and a reproducible test of high sensitivity and specificity was developed (Rogers, Fried & Kagan, 1968). This technique employed a soluble extract of Plasmodium knowlesi adsorbed to human group 0 red cells. The antigen is prepared by using a French press essentially as described by Mahoney, Redington & Schoenbechler (1966). In the present study, this method was used to evaluate malaria antibody in 4 collections of sera obtained from military recruits in the Western Hemisphere. The test was shown to be feasible for epidemiological studies. MATERIALS AND METHODS Antigen preparation Splenectomized rhesus monkeys were infected with the Anopheles hackeri strain of P. knowlesi 5 by blood transfer. The animals were anaesthetized Laboratory of Parasite Chemotherapy, National Insti- tutes of Health, Chamblee, Ga., USA. 2437 825- 826 I. G. KAGAN AND OTHERS and exsanguinated when 10% or more of their red blood cells contained mature schizonts. Either hepa- rin or sodium citrate was used as an anticoagulant. All further procedures were carried out at 4°C. The blood cells were washed 3 times in 10 volumes of 0.85% saline solution by centrifuging at 1100 g for 5 minutes and lysed by adding at least 10 volumes of distilled water. After the mixture had been centri- fuged at 3000 g for 5 minutes, the sediment consisted of plasmodia, leucocyte nuclei, and some red-cell stroma. Washing the sediment 3 times with saline solution facilitated removal of most of the fluffy white stroma. The final sediment was stored at - 70°C before further processing. Blood from a monkey weighing 3 kg could be processed to yield 5 ml -20 ml of plasmodial sediment in 4 hours. For further preparation, the plasmodia were thawed in tepid water and placed in a cooled French pressure cell.' Two volumes of antigen diluent (equal volumes of 4% NaCl and M/15 phosphate buffer at pH 6.4) were then added. The cell was operated at 20000 lbf/in2 (1406 kgf/cm2). The antigen diluent used for rinsing the cell brought the final volume to 4 times the volume of the plasmodial sediment. The disrupted material was centrifuged at 8000 g for 15 min. The colour of the supernatant was tan to pink, and it was opalescent; the residue was black and appeared to consist of malaria pigment and amorphous debris. The supernatant was considered to be undiluted antigen and was used as such, or, when titration tests showed that dilution was compatible with suit- able cell sensitivity, it was diluted with antigen diluent. The undiluted antigen was stored at -70°C. Test procedure Human group 0 blood, either fresh or stored for less than 21 days in acid-citrate-dextrose at 4°C, was washed 3 times with phosphate-buffered saline (PBS) solution (equal volumes of 0.85% NaCl and M/15 phosphate buffer at pH 7.2). The buffy coat was discarded. The erythrocytes were suspended in 40 volumes of PBS, and an equal volume of 1: 20 000 (w/v) tannic acid in PBS was added. The mixture was incubated for 15 min in an ice-bath. After the cells had been washed twice with PBS, they were ready for sensitizing. 1 American Instrument Company, Inc., Silver Spring, Md., USA. Use of trade names is for identification only and does not constitute endorsement by the Public Health Service or by the US Department of Health, Education, and Welfare. The packed tanned cells were sensitized by adding 5 volumes of antigen, mixing, and incubating for 15 min at 37°C. To decrease non-specific agglutina- tion, an equal volume of PBS with 5% normal rabbit serum (NRS) was added, and this mixture was incubated for an additional S min. After 1 wash in PBS containing 1% NRS at pH 7.2, the cells were suspended in 50 volumes of this solution and used immediately. Cells with 1: 10 000 thiomersal were useful for 1 week if stored at 4°C. Before stored cells were used, they were resuspended in fresh PBS with 1% NRS. Plastic plates 2 with an 8 x 12 array of wells were used. Each well first received 0.05 ml of PBS with 1 % NRS. A 0.05-ml diluting loop filled with test serum was placed in the first well in each row, and the serum was serially diluted in 2-fold dilution, beginning with a 1: 2 dilution. One collection of sera (Argentina) was titrated in 3-fold steps at a starting dilution of 1: 3. During agitation, 0.025 ml of the 2% sensitized-cell suspension was dropped into each test well. Positive and negative control sera were titrated each day. The completed test was read 45 min-90 min later. A smooth mat of agglutinated cells, a + + reading according to Stavitsky (1954), was considered a positive reaction. In order to detect non-specific agglutination reactions, positive sera were tested with a 2% suspension of tanned red blood cells sensitized with 5 % NRS. Sera titrated The sera tested were collected on 4 occasions by the WHO Reference Serum Bank in New Haven, Conn., USA, from male military recruits. A total of 10 956 sera was evaluated: 3077 sera from Argentina; 2681 sera from Brazil; 2961 sera from Colombia; and 2237 sera from the United States of America. The USA collection was described by Paul et al. (1964), and the Brazil collection by Florey et al. (1967). The Colombia and Argentina collec- tions were described by Evans et al. (1969). In gen- eral, young men were bled during the months of June and July at the time they were inducted into the army. The sample was designed to include army camps from all areas of the countries. Some of these collections have been extensively studied for other parasitic antibodies (Feldman, 1965; Cuadrado & Kagan, 1967; Walls & Kagan, 1967; Walls, Kagan & Turner, 1967; Lamb & Feldman, 1968). I2Cooke Engineering Company, Alexandria, Va., USA. EPIDEMIOLOGICAL STUDIES BY INDIRECT HAEMAGGLUTINATION TESTS FOR MALARIA 827 FIG. 1 FREQUENCY DISTRIBUTION OF TITRES OBTAINED IN THE PASSIVE HAEMAGGLUTINATION TEST WITH SERA FROM MILITARY RECRUITS WITH AN ANTIGEN PREPARED FROM PLASMODIUM KNOWLESI - WHO- Br,I - 2 681 sera __ WHO-U S A -2 237 sera -_ WHO-Argentino-3 077 sera ..... WHO Colombio-2 961 sero Titre RESULTS The frequency distribution of titres obtained in the IHA tests with the 4 collections is shown in Fig. 1, and the titres obtained are shown in Table I. Based on the peak of the specific portion of the curves, a titre of 1: 8 or greater was considered positive for epidemiological purposes. Using this TABLE 1 DISTRIBUTION OF INDIRECT HAEMAGGLUTINATION TITRES FOR MALARIAL ANTIBODY OBTAINED WITH MILITARY-RECRUIT SERA T..Wl Reciprocal titres a Country no. _Itested USA 2 237 Brazil 2 681 Colombia 2 961 Argentina 3 077 . 0 2 4 8 16 32 64 128 256 512 1 024 2 048 4096 9 27 81 243 2203 7 - 7 13 - 7 1 979 76 - 68 153 - 134 - 129 59 - 34 - 27 12 2162 62 - 108 134 - 166 - 146 93 - 46 - 31 8 2 877 58 100 30 10 2 a Dilution schema. 3000 6 1800 t; 'B 160 E m 8 2 2 1 2 I IL ----- - - 1. G. KAGAN AND OTHERS FIG. 2 GEOGRAPHICAL DISTRIBUTION OF THE ARGENTINIAN RECRUITS criterion, the following prevalence of positive reac- tors was obtained: USA, 20 out of 2237 (0.9%) specimens positive; Argentina, 142 out of 3077 (4.6%) specimens positive; Brazil, 558 out of 2681 (20.8%) specimens positive; and Colombia, 629 of 2961 (21.3%) specimens positive. Argentina The geographical distribution of the recruits sam- pled is shown in Fig. 2. The greatest concentration occurred in the north-eastern part of the country, intensifying toward Buenos Aires. The public health areas are shown in Fig. 3. Table 2 shows the number FIG. 3 STATES AND PUBLIC HEALTH AREAS IN ARGENTINA" a Roman numerals (I-VI) refer to public health areas, each comprised of more than 1 state. Small numbers refer to the individual states listed in Tables 2, 3 and 5. of positive serological reactors for each state within the 6 public health areas. Areas II and V had the highest serological prevalence rates. The prevalence of antibodies varied from state to state, ranging from zero in 7 states to a maximum of 28% sero- logical positive in the State of Tucuman. The sample for Tucuman (50 sera) was, however, small. Table 3 shows the number of slide-positive malaria cases in selected states from 1962 to 1965. These figures were obtained from active and passive sur- veillance records. These rates also show marked variation, ranging from zero positive in Catamarca to 11 % positive in Chaco. Fig. 4 shows the geographical distribution of re- actors obtained by slide and serological methods. The discrepancy in the distributions will be discussed later. Table 4 shows the distribution of positive reactors among blood groups in the ABO system. The distri- bution of positive serological reactors in the A, B, and 0 subgroups, respectively, was not significantly 828 829EPIDEMIOLOGICAL STUDIES BY INDIRECT HAEMAGGLUTINATION TESTS FOR MALARIA TABLE 2 ARGENTINA: THE NUMBER OF POSITIVE SEROLOGICAL REACTORS TO A PLASMODIUM KNOWLESI ANTIGEN FOR MALARIA ANTIBODY BY STATE AND PUBLIC HEALTH AREA State Name of state Public code Ihealth area 1 Buenos Aires 7 Entre Rios I 20 Santa Fe Total for area I 3 10 18 C6rdoba La Pampa San Luis II Total for area II 4 Corrientes 5 Chaco 8 Formosa 13 Misiones Total for area 1I1 2 Catamarca 9 Jujuy 16 Salta I 21 Santiago del Estero 22 Tucuman Total for area IV 11 12 17 6 14 15 19 La Rioja Mendoza San Juan Total for area V Chubut Neuquen Rio Negro Santa Cruz V V VI No. positive No. tested % positive 53 9 7 69 22 5 3 30 O 0 0 2 3 2 10 0 14 27 0 9 0 9 0 1 3 0 4 1 147 154 155 1 456 236 183 60 479 71 38 3 45 157 94 64 186 177 50 571 6 84 24 114 72 73 150 5 300 4.6 5.8 4.5 4.7 9.3 2.7 5.0 6.3 1.4 0 0 4.4 1.9 2.1 1.6 5.4 0 28.0 4.7 0 10.7 0 7.9 0 1.4 2.0 0 1.3 14 Total for area VI Total for country 11 142 3 077 4.6 1. G. KAGAN AND OTHERS FIG. 4 GEOGRAPHICAL DISTRIBUTION OF SLIDE AND SEROLOGICAL RESULTS BY PUBLIC HEALTH AREAS IN ARGENTINA a Vertical numerals refer to slide rotes Ito/ic numerals refer Ito percentage sero/ogico/ly positive a Roman numerals (I-VI) refer to public health areas. TABLE 3 ARGENTINA: NUMBER OF SLIDE-POSITIVE MALARIA CASES, 1962-65, FROM ACTIVE AND PASSIVE SURVEILLANCE BY STATE AND PUBLIC HEALTH AREA Name of state Publicahealth area Corrientes Chaco Formosa Misiones Total for area Ill Catamarca Jujuy Salta Santiago del Estero TucumAn Total for area IV Total III IV No. positive No. tested 16 4 974 570 246 5 806 0 211 441 11 24 687 6 493 12 297 45 375 14 957 33 344 105 973 17 191 125 938 316 100 38 799 91 506 589 534 695 507 State code 4 5 8 13 2 9 16 21 22 % positive 0.13 10.96 3.81 0.74 5.5 0 0.17 0.14 0.03 0.03 0.12 0.93 830 EPIDEMIOLOGICAL STUDIES BY INDIRECT HAEMAGGLUTINATION TESTS FOR MALARIA TABLE 4 ARGENTINA: DISTRIBUTION OF POSITIVE SEROLOGICAL REACTORS TO A PLASMODIUM KNOWLESI ANTIGEN FOR MALARIA ANTIBODY BY BLOOD GROUPS Blood Serologically positive Distribution of sample group No. No. % a A 42 29.6 985 32.0 B 19 13.4 277 9.0 0 78 54.9 1 738 56.5 AB 3 2.1 77 2.5 Total 142 100.0 3 077 100.0 a 2= 3.22 (P-0.05). different from the distribution in the samples as a whole. Table 5 shows a comparison of serological and surveillance results for malaria for states from which both data are available. Of the 9 states with high serological prevalence, 6 are also high in the list of states with high slide-positivity rates. Brazil The geographical distribution of Brazilian recruits tested is shown in Fig. 5. The largest group of recruits came from states along the eastern seaboard. TABLE 5 ARGENTINA: COMPARISON OF SEROLOGICAL FIG. 5 GEOGRAPHICAL DISTRIBUTION OF BRAZILIAN RECRUITS Few samples were obtained from the central zone, since this area is sparsely populated. Fig. 6 shows the delineation of public health areas and identifies each of the states in Brazil. The number of positive serological reactors by state and public health area is shown in Table 6. The serological prevalence in public health areas varied from 10% to 40%, with an average of 21 %. AND SLIDE RESULTS FOR MALARIA Serology Slide examination Rak State Nmofsaea % State tate code SName of state positive code Name of state positive 1 22 Tucumdn 28.0 5 Chaco 10.96 2 16 Salta 5.4 8 Formosa 3.81 3 13 Misiones 4.4 13 Misiones 0.74 4 2 Catamarca 2.1 9 Jujuy 0.17 5 9 Jujuy 1.6 16 Salta 0.14 6 4 Corrientes 1.4 4 Corrientes 0.13 7 5 Chaco 0 22 TucumSn 0.03 8 8 Formosa 0 21 Santiago del Estero 0.03 9 21 Santiago del Estero 0 2 Catamarca 0.00 a States from which both data were prevalence rates. available. States are ranked fromn highest to lowest 831 I. G. KAGAN AND OTHERS TABLE 6 BRAZIL: NUMBER OF POSITIVE SEROLOGICAL REACTORS TO A PLASMODIUM KNOWLESI ANTIGEN FOR MALARIA ANTIBODY BY STATE AND PUBLIC HEALTH AREA State Name of state Public F No. positive No. tested % positive code health area 2 Acre 3 8 37.5 3 Roraima 2 2 100.0 4 Amazonas 103 259 39.8 5 Para 94 239 39.3 Total for area 202 508 39.8 7 Maranh&o II 35 127 27.6 Total for area II 35 127 27.6 24 Mato Grosso III 69 272 25.4 Total for area 1I1 69 272 25.4 25 Goias 10 39 25.6 IV 26 Distrito Federal 0 5 0 Total for area IV 10 44 22.7 8 Piaui 7 30 23.3 9 Ceara 13 280 4.6 10 Rio Grande do Norte 3 19 15.8 11 Paraiba 3 34 8.8 V 12 Pernambuco 72 214 33.6 13 Alagbas 1 4 25.0 15 Sergipe 0 8 0 16 Bahia 44 254 17.3 Total for area V 143 843 17.0 17 Minas Gerais 37 199 18.6 18 Espirito Santo 0 6 0 19 Rio de Janeiro VI 2 24 8.3 20 Sao Paulo 16 167 9.6 27 Guanabara 0 45 0 Total for area VI 55 441 12.5 21 Parana 9 209 4.3 22 Santa Catarina VIl 3 24 12.5 23 Rio Grande do Sul 32 213 15.0 Total for area VIl 44 446 9.9 Total 558 2 681 20.8 832 EPIDEMIOLOGICAL STUDIES BY INDIRECT HAEMAGGLUTINATION TESTS FOR MALARIA FIG. 6 STATES AND PUBLIC HEALTH AREAS IN BRAZILa a Roman numerals (I-VII) refer to public health areas, each comprised of I or more states. Small numbers refer to indi- vidual states listed in Tables 6-8. FIG. 7 GEOGRAPHICAL DISTRIBUTION OF SLIDE AND SEROLOGICAL RESULTS BY PUBLIC HEALTH AREAS IN BRAZILa The number of slide-positive malaria cases, as determined by active and passive surveillance for the last half of 1963 and the first half of 1964 by state and public health area, is shown in Table 7. Slide-positive rates in the states varied from 0.01 % to 16.8%. Public health area IV showed the highest rate of positive slides, 15.7%, whereas only 2% of slides from area V were positive. The over-all slide-positivity rate was 4.7%. The geographical distribution of slide and sero- logical results is shown in Fig. 7. The serological percentages of positives were higher than the slide positivity percentages and did not vary as much. Table 8 shows the ranking, from highest to lowest prevalence according to serological and slide exami- nations, of states for which both types of results were available. Examination of this table shows that 10 of the top 12 states in the " serology " column are also in the top 12 of the " slide examina- tion " column. These results suggest a high degree of correlation. Colombia Fig. 8 indicates the geographical distribution of the sampled recruits in Colombia, and Fig. 9 shows FIG. 8 GEOGRAPHICAL DISTRIBUTION OF COLOMBIAN ARMY RECRUITS a Roman numerals (I-VII) indicate public health areas. 833 0 i-5 834 1. G. KAGAN AND OTHERS TABLE 7 BRAZIL: NUMBER OF SLIDE-POSITIVE MALARIA CASES, 1963-64, FROM ACTIVE AND PASSIVE SURVEILLANCE BY STATE AND PUBLIC HEALTH AREA Sctadte Name of state Public No. positive No. tested % positive 2 Acre 130 2 749 4.7 3 Roraima 416 2 480 16.8 4 Amazonas 582 10 624 5.5 Para 2670 18930 14.1 Total for area 3 798 34 783 10.9 7 MaranhAo II 3825 24635 15.5 Total for area 11 3 825 24 635 15.5 24 Mato Grosso Il 1 024 28 917 3.5 Total for area III 1 024 28 917 3.5 25 Goias 4187 26 589 15.7 IV 26 Distrito Federal _ _ Total for area IV 4 187 26 589 15.7 8 Piaui 878 13195 6.7 9 Ceara 470 20 236 2.3 10 Rio Grande do Norte 5 58 240 0.01 11 Paraiba 101 27 265 0.37 V 12 Pernambuco 223 17 336 1.3 13 Alag6as 2 4 808 0.04 15 Sergipe 13 22114 0.06 16 Bahia 2 131 24 471 8.7 Total for area V 3 823 187 665 2.0 17 Minas Gerais 1 560 45 757 3.4 18 Espirito Santo 611 13307 4.6 19 Rio de Janeiro VI 71 12 664 0.6 20 Sao Paulo _ _ 27 Guanabara 20 6 312 0.3 Total for area VI 2 262 78 040 2.9 21 Parana 364 17 828 2.0 22 Santa Catarina VIl 935 28679 3.3 23 Rio Grande do Sul - - - Total for area VII 1 299 46 507 2.8 Total 20 218 427 136 4.7 EPIDEMIOLOGICAL STUDIES BY INDIRECT HAEMAGGLUTINATION TESTS FOR MALARIA TABLE 8 BRAZIL: COMPARISON OF SEROLOGICAL AND SLIDE RESULTS FOR MALARIA Serology State Name of state acode ~ Roraima Amazonas Para Acre Pernambuco Maranh&o Goias Mato Grosso Alag6as Piaui Minas Gerais Bahia Rio Grande do Norte Santa Catarina Paraiba Rio de Janeiro Ceard ParanA Sergipe Espirito Santo Guanabara % State positive code 100 39.8 39.3 37.5 33.6 27.6 25.6 25.4 25.0 23.3 18.6 17.3 15.8 12.5 8.8 8.3 4.6 4.3 0.0 0.0 0.0 3 25 7 5 16 8 4 2 18 24 17 22 9 21 12 19 11 27 15 13 10 Slide examination Name of state a Roraima Goias Maranh&o Para Bahia Piaui Amazonas Acre Espirito Santo Mato Grosso Minas Gerais Santa Catarina Ceard Paran& Pernambuco Rio de Janeiro Paraiba Guanabara Sergipe Alag6as Rio Grande do Norte a States from which both data were available. States are ranked from highest to lowest prevalence rates. the states and public health areas. Recruits were drawn mainly from the western part of the country. The eastern half of the country (public health areas II and IV) is sparsely populated and consequently con- tributed few recruits. The data from the whole Colombian survey (Table 9), based on the state of longest residence, indicate that public health areas III and IV, in the interior of the country, had the highest prevalence of positive antibody levels. The central public health area (area 11) had the lowest, whereas the coastal areas (areas I and V) showed inter- mediate antibody levels. Of the 4 countries included in this study, the recruits from Colombia showed the highest percent- age of positive titres to malaria, 21.3 %. Note (Fig. 1) that the peak titre of the Colombian recruits was 1: 16, whereas the peak titre of recruits from Argentina, Brazil, and the USA was 1: 8. This may indicate more intense exposure to malaria than is found in the other countries studied. The number of slide-proven malaria cases for the years 1963, 1964, and 1965 obtained by active and passive surveillance, is shown in Table 10. The over- all prevalence by slide examination was 3 %. The geographical distribution of slide and serological results by public health area is shown in Fig. 10. Table 11 shows the ranking, from highest to lowest prevalence, according to serology and slide examination. The table shows that 8 of the top 10 states in the " serology " column rank in the Rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 3 4 5 2 12 7 25 24 13 8 17 16 10 22 11 19 9 21 15 18 27 positive 16.8 15.7 15.5 14.1 8.7 6.6 5.5 4.7 4.6 3.5 3.4 3.3 2.3 2.0 1.3 0.6 0.4 0.3 0.1 0.0 0.0 835 I. G. KAGAN AND OTHERS FIG. 9 STATES AND PUBLIC HEALTH AREAS IN COLOMBIAa a Roman numerals (I-V) refer to public health areas, each comprised of 1 or more states. Small numbers refer to individual states listed in Tables 9-11. first 10 states in the " slide examination " column. These data again support the results obtained by serological testing in the country. Table 12 shows that the incidence of malaria antibody is related to altitude, as one would expect. Malaria, as evidenced by antibody prevalence, has the highest prevalence in altitudes between sea level and 200 m. Antibody in the recruit population declined sharply at higher altitudes and levelled off above 1500 m. Malaria transmission is believed to occur up to the 1500 m-2000 m levels. This altitude factor may explain the relatively low antibody activ- ity among persons from public health area II. This region, which lies along the Andes chain, is largely mountainous. No specific correlations were found between blood types in the ABO system and the presence of malaria antibody (Table 13). The data from lifelong residents of the same area follow the same patterns as data from the entire sample which is based on the state of longest FIG. 10 GEOGRAPHICAL DISTRIBUTION OF SLIDE AND SEROLOGICAL RESULTS BY PUBLIC HEALTH AREAS IN COLOMBIA a 3.9 1/9.2 4#.6 Vertical numerals refer to slide rates Italic numerals refer to percentage serologicallpy positive a Roman numerals (I-V) indicate public health areas. residence. The over-all percentage of positives is somewhat, but not significantly, lower. The lower percentage may be due to a smaller number of specimens being available in this category. United States of America The geographical distribution of the 30 serological reactors to the malaria antigen in the USA, by state of birth, gave no indication of a pattern. Individuals whose sera had reacted in the test lived from Maine to California. This suggests that the reactions observed were probably non-specific. The fact that almost all of the sera were of negative titre also attests to the test's high degree of specificity and non-reactivity. Of the 2237 USA sera tested, 34 (1.5%) reacted to malaria antigen, but only 20 reacted in serum dilutions of 1: 8 or greater. Of the positive reactors, 12 out of 20 (60%) were from a rural environment, 6 out of 20 (30%) were from towns, and 2 out of 20 (10%) were from urban areas. 836 EPIDEMIOLOGICAL STUDIES BY INDIRECT HAEMAGGLUTINATION TESTS FOR MALARIA 837 TABLE 9 COLOMBIA: THE NUMBER OF POSITIVE SEROLOGICAL REACTORS TO A PLASMODIUM KNOWLESI ANTIGEN FOR MALARIA ANTIBODY BY STATE AND PUBLIC HEALTH AREA Public State of longest residence Birthplace State Name of state health code area No No. tested o. No. tested % 2 Antioquia 43 200 21.5 16 91 17.6 4 Atlantico 4 43 9.3 3 24 12.5 5 Bolivar 32 145 22.0 23 92 25.0 11 C6rdoba 6 21 28.6 2 10 20.0 14 Guajira 1 8 12.5 1 1 100.0 16 Magdalena 23 150 15.3 4 81 4.9 Total for area 109 567 19.2 49 299 16.4 7 Caldas 47 364 12.9 25 205 12.2 12 Cundinamarca 45 404 11.1 19 207 9.2 15 Huila 13 100 13.0 1 48 2.1 19 Norte de Santander 33 105 31.4 24 74 32.4 21 Santander 49 179 27.4 24 95 25.3 23 Tolima 35 281 12.5 19 149 12.8 Total for area 11 222 1 433 15.5 112 778 14.4 3 Arauca 30 45 66.7 4 8 50.0 6 Boyaca III 101 368 27.4 40 174 23.0 17 Meta 10 22 45.5 1 4 25.0 Total for area 1I1 141 435 32.4 45 186 24.2 1 Amazonas 2 7 28.6 0 1 0 8 Caqueta IV 12 34 35.3 2 14 14.3 20 Putumayo 22 33 66.7 5 9 55.6 Total for area IV 36 74 48.6 7 24 29.2 9 Cauca 24 124 19.4 14 79 17.7 10 Choco 45 55 81.8 28 33 84.8 18 Narino V 11 63 17.5 5 28 17.9 24 Valle (del Cauca) 42 213 19.7 15 87 17.2 Total for area V 122 455 26.8 62 227 27.3 Total sample 630 ] 2964 | 21.3 275 - 1 514 { 18.2 1. G. KAGAN AND OTHERS TABLE 10 COLOMBIA: NUMBERS OF SLIDE-PROVEN MALARIA CASES (1963, 1964, 1965) BY STATE AND PUBLIC HEALTH AREA NametofastateePublicof state health Antioquia Atlantico Bolivar C6rdoba Guajira Magdalena Total for area I 7 Caldas 12 Cundinamarca 15 Huila I 1 19 Norte de Santander 21 Santander 23 Tolima Total for area 11 Arauca Boyaca Meta Total for area IlIl Amazonas Caqueta Putumayo Total for area IV No. positive 10 658 182 1 215 9 608 204 657 22 524 588 1 804 52 3 996 6 551 1 030 14 021 980 2 930 2 383 6 293 0 164 757 921 Cauca 2 260 Choco 1 790 V Narino 2 478 Valle (del Cauca) 774 Total for area V 7 302 Total sample 51 061 1 484 641 3.4 State code 2 4 5 11 14 16 No. tested 210 055 27 587 91 289 119 617 26 245 99 738 574 531 61 040 91 269 34 059 61 432 73 921 81 982 403 703 22 371 84814 30 824 138 009 348 32 171 11 976 44 495 72 077 57 886 108 756 85 184 323 903 % positive 5.1 0.6 1.3 8.0 0.8 0.6 3.9 1.0 2.0 0.2 6.5 8.9 1.2 3.5 4.4 3.4 7.7 4.6 0 0.5 6.3 2.1 3.1 3.1 2.3 0.9 2.2 8 20 838 EPIDEMIOLOGICAL STUDIES BY INDIRECT HAEMAGGLUTINATION TESTS FOR MALARIA 839 TABLE 11 COLOMBIA: COMPARISON OF SEROLOGICAL AND SLIDE RESULTS FOR MALARIA Serology Slide examination Rank State of longest residence Birthplace Stte Name of state a State Name of state a % State Name of state a % code positive code positive code positive Choco Arauca Putumayo Meta Caqueta Norte de Santander Amazonas C6rdoba Boyaca Santander Bolivar Antioquia Valle (del Cauca) Cauca Narino Magdalena Huila Caldas Guajira Tolima Cundinamarca Atlantico 14 10 20 3 19 21 17 5 6 11 18 9 2 24 8 23 4 7 12 16 15 Guajira Choco Putumayo Arauca Norte de Santander Santander Meta Bolivar Boyaca C6rdoba Narino Cauca Antioquia Valle (del Cauca) Caqueta Tolima Atlantico Caldas Cundinamarca Magdalena Huila Amazonas 100.0 84.4 55.6 50.0 32.4 25.3 25.0 25.0 23.0 20.0 17.8 17.7 17.6 17.2 14.3 12.8 12.5 12.2 9.2 4.9 2.1 0.0 21 11 17 19 20 2 3 6 9 10 18 12 5 23 7 24 14 4 16 8 15 1 Santander C6rdoba Meta Norte de Santander Putumayo Antioquia Arauca Boyaca Cauca Choco Narino Cundinamarca Bolivar Tolima Caldas Valle (del Cauca) Guajira Atlantico Magdalena Caqueta Huila Amazonas 8.9 8.0 7.7 6.5 6.3 5.1 4.4 3.4 3.1 3.1 2.3 2.0 1.3 1.2 1.0 0.9 0.8 0.6 0.6 0.5 0.2 0.0 a States from which both data were available. States are ranked from highest to lowest prevalence rates. TABLE 12 TABLE 13 COLOMBIA: DISTRIBUTION OF POSITIVE SEROLOGICAL DISTRIBUTION OF POSITIVE SEROLOGICAL REACTORS REACTORS BY ALTITUDE BY BLOOD GROUPS FOR COLOMBIA Cumulative Altitude No. No. % percentage(m) positive examined positive of total no. I positive 0-200 174 516 33.7 28.3 200-500 112 404 27.7 46.6 500-1000 49 254 19.3 54.6 1000-1500 105 527 19.9 71.7 1500-2000 85 590 14.4 85.5 2000-3000 84 575 14.6 99.2 3000-4000 5 32 15.6 100.0 Total 614 2 898 I 21.2 Serologically positive No. %a Distribution of sample No. %a 152 24.1 744 25.1 62 9.8 260 8.8 405 64.3 1 905 64.3 11 1.7 55 1.9 630 2 964 a x = 0.89 (P >0.05). L3] 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 10 3 20 17 8 19 11 6 21 5 2 24 9 18 16 15 7 14 23 12 4 81.8 66.7 66.7 45.4 35.2 31.4 28.6 28.6 27.8 27.4 22.0 21.5 19.7 19.4 17.5 15.3 13.0 12.9 12.5 12.5 11.1 9.3 I ~~~~~~~~~~~~- I. G. KAGAN AND OTHERS DISCUSSION The indirect haemagglutination test using P. know- lesi antigen has been shown to be a sensitive and reproducible method for demonstrating antibody to human malaria (Rogers, Fried & Kagan, 1968). Desowitz & Saave (1965) used an IHA test employ- ing formolized tanned sheep cells sensitized with P. cynomolgi and P. coatneyi antigens to study areas of different malaria endemicity. Their results showed reduced antibody levels in protected popula- tions over the antibody levels in populations exposed to malaria. The serological data obtained in this study could not be related directly to the malarial history of the donor. For this reason an epidemiological approach was made to the evaluation of the data. Except for Argentina, serological data on the distribution of malaria in the 4 countries sampled compare well with data obtained by active surveillance. One would normally expect the seropositivity rate to exceed the slide-positivity rate several-fold, since the former may be measuring lifelong experience with the parasite, whereas the latter measures active infection at the moment of examination. In the states of Chaco, Formosa, and Santiago del Estero in Argentina, the slide-positivity rates were higher than the sero- positivity rates. These differences may be due, in part, to differences in sampling, since recruits are of restricted ages, and the people in the surveillance surveys are of all ages. The extremely large differ- ence between the serologically positive samples (28 %) and the slide-positive (0.02%) figures in the state of Tucuman is puzzling. These data suggest either that the microscopical techniques were inade- quate in this area or that the seropositivity index is erroneous. The lack of reactivity in the serological tests with the USA sample suggests that the test used did not yield a high rate of false positivity. A comparison of data obtained by serology and by active and passive surveillance in Brazil and Colom- bia strongly suggests that both methods are measur- ing prevalence of malaria, since with a few excep- tions, states with the highest slide-positive rates also ranked high in serological reactivity. Discrepancy in the 2 sets of data for Tucuman is probably due to the fact that the IHA test for malaria detects anti- body from experience with malaria as far back as 5 years or more. Active surveillance detects only the cases of malaria presenit at the time of sampling. One would logically expect much higher levels by the serological method. Years ago, Tucuman was an area of high endemicity, and this may be reflected in the serological data. The over-all distribution of serological reactors in Argentina obtained in this study is plausible. Bejarano (1962) reviewed the epidemiology of malaria in Argentina for the years 1897-1960, and recorded 9 cases of malaria in the State of Mendoza in 1959 (the last year reviewed). In our study we found a seropositivity rate of 10.7% among the 84 samples from Mendoza. In San Luis, 5% of the 60 samples were seropositive. Malaria was common in this state in the 1940s and 1950s. The last case was reported in 1958. The geographical distribution in Argentina of the 3 anopheline vectors, Anopheles pseudopunctipennis, A. albitarsis, and A. darlingi, coincides very closely with the serological range of reactors obtained in this study (Bejarano, 1959). Malaria may not exist in these areas today, but it did exist in the not too distant past, and the sero- logical method may be detecting this event. Athreya & Coriell (1967) discussed the significance of high B-gene frequency in population groups in various parts of the world in relation to endemic malaria. They postulated that the presence of blood group-A substance in malaria may be responsible for increased pathogenicity and mortality in individ- uals in group A thus causing a shift in the gene fre- quency in the population. Raper (1968) analysed data from 3500 patients experimentally infected with malaria in the treatment of neurosyphilis and other disorders. Raper could not find any correlation between parasitaemia and the presence of anti-A or anti-B in the blood of the patients. Our inability to find any correlation between serological reactivity and the ABO status of the 6038 young men sampled in Argentina and Colombia confirms this finding. ACKNOWLEDGEMENTS We wish to express our appreciation to the WHO Reference Serum Bank, Yale University, New Haven, Conn., USA, for making the military recruit serum collections available for study. We wish to thank Dr Raul Cuadrado for his consultation and help. 840 EPIDEMIOLOGICAL STUDIES BY INDIRECT HAEMAGGLUTINATION TESTS FOR MALARIA 841 RtSUME ETUDES StRO-IPIDI-MIOLOGIQUES DU PALUDISME PAR L'EPREUVE D'HEMAGGLUTINATION INDIRECTE: PRELEVEMENTS DE StRUMS CHEZ DES RECRUES D'ARGENTINE, DU BRE-SIL, DE COLOMBIE ET DES ETATS-UNIS D'AMERIQUE On a eu recours a l'epreuve d'hemagglutination indirecte, utilisant comme antigene Plasmodium knowlesi adsorbe sur des erythrocytes humains du groupe 0, pour titrer les anticorps antipaludeens dans 10 956 serums preleves sur des recrues de quatre pays de l'h6mi- sphere occidental. Un titre de 1/8 ou plus a e choisi comme seuil de positivite dans cette enquete epidemiolo- gique. Sur la base de ce critere, la prevalence des reactions serologiques positives a et la suivante: Etats-Unis d'Am6rique: 20 serums positifs sur 2237 (0,9%); Argen- tine: 142 sur 3077 (4,6o%); Bresil: 558 sur 2681 (20,8%); Colombie: 629 sur 2961 (21,3 %). Aux Etats-Unis, l'analyse de la provenance des serums positifs a montre la diversite de leur origine et l'absence de tout foyer, ce qui donne a penser qu'il pourrait s'agir de reactions non specifiques. En Argentine, on a decele la presence d'anticorps chez des recrues dans des regions ou aucun cas de paludisme n'avait e signale depuis 10 ans au moins; il semble que l'epreuve d'hemag- glutination mette en evidence des anticorps residuels dans une population plusieurs ann6es apres l'6radication du paludisme. Au Bresil, on a releve une correlation satisfai- sante entre les resultats des examens serologiques et la repartition du paludisme determinee par les examens parasitologiques directs. C'est en Colombie que les taux de positivite ont ete les plus elev6s, les titres de 1/16 etant les plus nombreux alors que dans les trois autres pays le titre de 1/8 etait le plus frequemment observe. Dans ce pays egalement, on a note une concordance satisfaisante entre la prevalence du paludisme etablie par l'enqu8te serologique et la prevalence attestee par les resultats des examens parasitologiques. Plus de 50% des serums positifs provenaient de recrues colombiennes stationnees dans des camps situes 'a moins de 1000 m d'altitude. En Argentine et en Colombie, aucune correlation n'est apparue entre la presence des anticorps antipaludeens et l'appartenance a l'un ou l'autre groupe sanguin du systeme ABO. REFERENCES Arthreya, B. H. & Coriell, L. L. (1967) Amer. J. Epidem., 86, 292-304 Bejarano, J. F. R. (1959) Prim. Jornadas Entomoepidem. argent., 1, 305-329 Bejarano, J. F. R. (1962) Rev. Asoc. med. fuerz. Arm., 3, 28-58 Cuadrado, R. R. & Kagan, I. G. (1967) Amer. J. Epidem., 86, 330-340 Desowitz, R. S. & Saave, J. J. (1965) Bull. Wld Hlth Org., 32, 149-159 Evans, A. S., Casals, J., Opton, E. M., Borman, E. K., Levine, L. & Cuadrado, R. R. (1969) Amer. J. Epidem., 90, 292-303 Feldman, H. A. (1965) Amer. J. Epidem., 81, 385-391 Florey, C. du V., Cuadrado, R. R., Henderson, J. R. & de Goes, P. A. (1967) Amer. J. Epidem., 86, 314-318 Lamb, G. A. & Feldman, H. A. (1968) Amer. J. Epidem., 87, 323-328 Mahoney, D. F., Redington, B. C. & Schoenbechler, M. J. (1966) Milit. Med., 131, Suppl., pp. 1141-1151 Paul, J. R., Niederman, J. C., Pearson, R. J. C. & Florey, C. du V. (1964) Amer. J. Hyg., 80, 286-292 Raper, A. B. (1968) Trans. roy. Soc. trop. Med. Hyg., 62, 158-159 Rogers, W. A., Fried, J. A. & Kagan, I. G. (1968) Amer. J. trop. Med. Hyg., 17, 804-809 Stavitsky, A. B. (1954) J. Immunol., 72, 360-367 Walls, K. W. & Kagan, I. G. (1967) Amer. J. Epidem., 86, 305-313 Walls, K. W., Kagan, I. G. & Turner, A. (1967) Amer. J. Epidem., 85, 87-92 7