Interpretation of IHA titres for the study of malaria epidemiology*

Bull. Org. mond. Sante 11973, 49, 485-492 Bull. Wid Hlth Org. Interpretation of IHA titres for the study of malaria epidemiology * HANS 0. LOBEL, HENRY M. MATHEWS, & IRVING G. KAGAN The results of IHA test surveys ofpersons with malaria parasitaemia in Ethiopia and the Philippines suggest that the antibody response may be influenced by the frequency and intensity of the antigenic stimulations and also be age-dependent. Antibody frequency dis- tribution curves from four different areas suggest that the shape of such curves can provide some information about the endemicity ofmalaria. Results ofsimilar and ofparasitological surveys in Bangladesh, Ethiopia, Haiti, and the Philippines were compared and related to available malaria surveillance information. The results indicate that a serologic population profile may provide an indication of the history and status of malaria. Technical aspects of the IHA test are reviewed; it may be desirable to use homologous antigens instead of a simian Plasmodium antigen. To obtain the most useful additional epidemiological infor- mation about a malaria situation, serologic data need to be age-related, and longitudinal surveys are usually more informative than a cross-sectional survey. Variables influencing the applicability of a sero- logic method for epidemiological purposes include the sensitivity and specificity of the test, the repro- ducibility of the test results, the development and persistence of the measured antibodies, and also the laboratory procedures necessary to analyse the specimens. Little is understood about the complex immuno- logical relationship between the human host and the Plasmodium antigens, but considerable progress has already been made in developing tests to detect malaria antibodies. The indirect fluorescent antibody (IFA) method is successfully used by several research laboratories but more adequate standardization is needed. Progress has also been made in the development of an indirect haemagglutination (IHA) procedure. For epidemiological purposes, the IHA test has two major advantages: a relatively large number of sera can be tested, and the technique can be automated. * From the Center for Disease Control, Public Health Service, US Department of Health, Education, and Welfare, Atlanta, Ga. 30333, USA. This study was supported in part by the State Depart- ment, Agency for International Development, Participating Agency Service Agreement (PASA) No. RA (HA) 5-68. The 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 IHA method, as developed by Rogers et al. (14), uses a P. knowlesi antigen that reacts with anti- bodies produced by P. falciparum, P. vivax, or P. malariae. The test results appear to be highly specific (3). Some limited field studies have been carried out with this method to explore its applica- bility (6, 9, 7, 3). This paper examines the antibody response in per- sons with parasitaemia in two different areas and the antibody patterns in populations with different levels of malaria endemicity and relates the results of serologic and parasitologic surveys conducted in 5 different countries to available malariometric infor- mation. The studies on which the data are based were carried out between 1968 and 1972. MATERIALS AND METHODS The IHA test developed by Rogers et al. (14) was used in the study; P. knowlesi was the antigen and human group 0 erythrocytes tanned with 1: 20 000 tannic acid solution were the carriers for the titra- tion of eluates of filter paper blood specimens. Antigen was prepared by exsanguinating Rhesus monkeys when the parasitaemia was at a peak and schizonts were the predominant stage. The recovered red cells were washed free of serum proteins and then lysed in distilled water containing 0.01 % Triton 3144 - 485 H. 0. LOBEL ET AL. X-100. The parasites were recovered by low-speed centrifugation and washed until they were free of haemoglobin. The washed parasites were frozen at -70'C until ready for further processing. For soluble antigen preparations, parasites were thawed in tepid water and suspended in 5 volumes of anti- gen diluent, pH 6.4 (equal volumes of pH 6.4 phos- phate-buffered saline (PBS) and 4% sodium chloride solution). This suspension was passed through a cooled Ribi Cell Fractionator at 1.17 x 108 Pa pres- sure. The recovered material was centrifuged at 12 000 x g for 15 min, and the supernatant stock anti- gen was decanted and stored at - 70°C. The optimum dilution of antigen was determined by sensitizing erythrocytes with dilutions of antigens made up in antigen diluent. Sensitized cells were suspended in PBS, pH 7.2, containing 1% normal rabbit serum and tested against a small battery of sera. Cells sensitized with the optimum antigen dilution were further tested against a larger battery of positive and negative sera to limit the variability of test results. The capillary blood from a finger-prick (about 0.1 ml) was absorbed onto a preprinted circle of filter paper (Ropaco grade No. 1023-.038). The papers were dried and packaged with " glassine" interleaves for shipment to the laboratory at ambient temperature. In the laboratory, a disk 10 mm in diameter was punched from the blood-soaked circle and eluted for at least 30 minutes in 0.2 ml of PBS. About 0.12-0.14 ml of eluate was expressed from the filter paper disk. The filter paper blood eluate corresponded approximately to a 1: 11 dilution of serum. In the test procedure, the filter paper eluates were diluted in 2-fold steps with loops in microtitration plates, and erythrocytes sensitized with antigen were added. After at least 1 h, the absence or presence of haemagglutination was determined. In the IHA test, titres of 1: 16 or greater were considered positive (3). The serologic indices used included the percentage of specimens with an IHA titre of 1: 16 or higher (seropositivity rate), the geo- metric mean titre ofthe specimens with a titre of 1: 16 or higher (mean positive titre), and/or the geometric mean titre of all specimens (mean titre). To calculate overall geometric mean titre values, we arbitrarily gave titres of less than 1: 16 a value of 1: 2. The sensitivity of the test was defined as the percentage of positive reactors among individuals with patent parasitaemia. The filter paper blood specimens were obtained during studies ofpopulations in Bangladesh, Ethiopia, Haiti, and the Philippines. Details regard- ing the characteristics of each population examined are given under " Results " below. Whenever blood was obtained for serologic testing, paired blood slides were made for microscopic examination to detect the presence of Plasmodium parasites. For evaluation of the serologic data, the test results have been compared with the results of the microscopic blood slide examinations and with malaria surveillance information. RESULTS Serologic indices and age To investigate the possibility of an age-related influence on the development of antibodies to ma- laria, seropositivity rates and mean positive titres were measured in two groups of patients, all of whom had patent parasitaemia as determined by blood slide examination. The IHA titrations were performed from filter-paper blood specimens. One group of 226 patients in Gambela, Ethiopia, came from a population in which the intensity of trans- mission was high: the parasite rate in 0-4-year-old children was 62.1 %. The other group of 197 patients came from populations in the Philippines in which the transmission level was much lower: the parasite rate was 7.4% in 0-4-year-old children. A total of 75% of the patients in Ethiopia and 65% of the patients in the Philippines were infected with P. falciparum. Fig. 1 shows that antibodies to malaria were detected in 70-90% of the malaria patients studied in Ethiopia and the Philippines, regardless of age. In Ethiopia, the mean positive titres were high in all age groups. In contrast, the n 0 I- 0 0 GAMBELA, ETHIOPIA 100 75 50- 25 PHILIPPINES 100 '-----~~~----75 =: 50 25 0 124 6810 20 40+ YEARS OF AGE 1024 512 1256 1128 64 32 PERCENT WITH HA TITER > 1i:16 MEAN POSITIVE TITER 1024 -512 1256 1128 64 32 Fig. 1. Serologic values in persons with patent para- sitaemia, by age. 486 IHA TITRES IN MALARIA 80- 70- BALIHAR, BANGLADESH 60 N-205 50 40 30 20 10-10 ,- . 40 RIO TUBA, PHILIPPINES 30 N-92 20 10 0 <16 32 12B 512 120481 8;921 16 64 256 1024 4096 16384 IHA TITER Fig. 2. Relative frequency distribution of IHA titres. mean positive titres from the Philippine patients were relatively low in the young age groups; titre levels rose gradually with increasing age but the maximum titre was not reached until adulthood. In this study, an age-related increase of antibody titres could readily be detected in patients with malaria who live in an area with a low level of malaria transmission, but not in patients living in an area with a high level of malaria transmission. Antibody patterns Graphic illustration of antibody titre frequency distributions may demonstrate the malaria experi- ence of the examined population. Four distinctly different frequency distributions of malaria antibody titres were obtained from titra- tions of filter-paper blood specimens collected in Bangladesh, Ethiopia, Haiti, and the Philippines (Fig. 2). Paired blood slides were examined for the presence of Plasmodium parasites. Only 38.1 % of the 441 specimens from Haiti had an IHA titre of 1 :16 or greater, with a mean pos- itive titre of 1: 86. The highest titre was 1: 8 192. The parasite rate was 2.0%, and only P. falciparum was diagnosed. In Balihar, Bangladesh, only 23.4% of the 205 specimens had a positive serologic reaction. The mean positive titre was 1: 78, and no titre was higher than 1: 512. All blood slides were negative. The 396 specimens in Ethiopia were collected after the peak of the transmission season and posi- tive IHA titres were found in 77.3% of the speci- mens, with a mean positive titre of 1: 530. The highest titre measured was 1: 32 768. The slide positivity rate was 58.6 %. An outbreak of falciparum malaria had occurred in a lumber camp in Rio Tuba, Philippines. All but 5 of the 92 persons (94.6 %) who had had symptoms suggestive of malaria infection had an IHA titre of 1: 16 or greater. The mean positive titre was 1: 2 180. Since everyone had been treated with chloroquine several days before the specimens were collected, only 17.4% still had a detectable parasit- aemia at the time of the survey. These data illustrate that some information on the endemicity of malaria can be obtained from the type of the frequency distribution curve. Malaria surveys In an effort to determine whether serologic data from population surveys can be used to estimate the occurrence of malaria transmission, serologic and parasitologic surveys were carried out in Gambela, Ethiopia; in a cluster of villages on Palawan, the Philippines; in 2 localities in Bangladesh; and in a group of 6 villages in Haiti. In one of these localities (Balihar, Bangladesh), the malaria eradication campaign had interrupted the transmission of malaria, and in the other areas, varying levels of malaria transmission were found. From each individual included in the surveys, a blood slide was obtained for microscopic examina- tion to detect the presence of Plasmodium parasites and a filter paper blood specimen was obtained for serologic testing. The parasitologic and serologic survey results were compared, and they were related to the available malaria surveillance information. The age-related parasite and seropositivity rates are shown in Fig. 3, and Table 1 indicates the mean titre and the mean positive titre values for the age groups 0-4, 5-14, 15-29, and 30 years and older. A. A survey was conducted in January 1970 in Gambela, Ethiopia, which has a high level of malaria endemicity. The survey included 396 per- manent residents of Gambela, i.e., 30% of the population. Plasmodium parasites were found in 62% of the children under 5 years of age. The parasite rate reached its peak in the 6-7-year-old children and then decreased with increasing age, reflecting the effect of the acquired immunity. P. falciparum accounted for 70% of the malaria infections. The seropositivity rates were 70-75% in young children and increased to a maximum of 90% in adults. The mean positive titre level was high in all age groups: 1: 333 in the 0-4-year-old group and 1: 765 in the 30 years and older group. 80 70 HAITI 60 N-441 Z 50 w 40 ° 30 C) 20 U. O 10 U 401 GAMBELA, ETHIOPIA 30 N. 396 20 10 <6 3 i218 1 2048 892 132768 16 64 256 1024 4096 16384 IHA TITER 487 H. 0. LOBEL ET AL. IHA GAMBELA, ETHIOPIA 100 75 50 25 0q PALAWAN, PHI LI PPINE100 75 50 I OR. > 0F_ KALAPARA, BANGLADE c) 100 0QL 75 F- 50 z LU 25 tr 0 LLI BALIHAR, BANGLADI 75 50 25 0 HAITI 75 50 25 r Tl 0246810 20 30 40+ 024681 YEARS OF AGE * HA TITER OF 1:16 IS POSITIVE ** DETERMINED BY BLOOD SMEAR EXAMINATIC Fig. 3. Serologic and parasitologic value age. B. The survey conducted in Septei the coast of Palawan, Philippines, incl of 7 adjacent villages with a popula and almost the entire population wa, Of the children under 5 years of a patent parasitaemia, and similar para found in the older age groups. P. counted for 65% of the infections. The seropositivity rate was 39% in 5 years of age and increased gradua PARASITEMIA the oldest age group. The mean positive titre was 1: 91 in the 0-4-year-old group and 1: 420 in the 30 years and older group. The surveillance methods of the malaria eradica- tion programme on Palawan included active and passive case detection (ACD and PCD). In 1970 a total of 832 blood slides had been examined in this population of 1 643 (51 %) and in 1971 a total of S 1 641 slides were examined (100%). The number of detected malaria cases per 1 000 population (annual parasite incidence, API) was 94.9 in 1970, as compared with 153.8 in 1969. P. falciparum was diagnosed in 83.3% of the cases found in 1970. C. A survey was carried out in February 1970 in Kalapara, Bangladesh (population 5 300), which included 995 individuals (19% of the population). Of the children in the 0-4-year age group, 10% had a patent parasitaemia and the parasite rates were somewhat lower in the older age groups. Only , . .--T 5% of the infections were with P. falciparum. ESH Antibodies to malaria were found in 30% of the children under 5 years old, and the seropositivity rate increased gradually to 76% in the 30 years and older group. The mean positive titre was 1: 80 in the 0-4-year age group and 1: 223 in those 30 years and older. ~' ' ' In December 1968 the malaria eradication pro- gramme conducted a parasitologic survey of 1 315 children under 10 years of age; the survey showed a parasite rate of 9.4 %. The malaria eradica- tion campaign started in May 1970 with residual indoor insecticide spraying. 0 20 30 40+ D. The serologic and parasitologic survey in Balihar, Bangladesh (population 1 400) was con- ducted in February 1970. No Plasmodium parasites were found in any of the 205 persons (15% of the population) examined during the survey. s in 5 areas, by Antibodies to malaria could not be detected in children under 8 years of age. The seropositivity rates then rose rapidly with increasing age, and 77% of the persons older than 30 had antibodies to mber 1970 on malaria. The mean positive titre was 1: 90 in this luded a cluster age group. tion of 1 643, The malaria surveillance activities had been initi- s tested. ated in Balihar in 1964 with ACD. Between 1964 and ge, 6% had a 1969 the number of slides examined annually per 100 site rates were population ranged between 7.4 and 26.7. The API falciparum ac- declined from 2.8 in 1964, to 1.4 in 1965, 0.1 in 1966, 0.1 in 1967, and 0.07 in 1968. No malaria cases were children under found in 1969. P. falciparum accounted for 61.5% lly to 84% in of the infections in 1964 and 10% of the infections 488 IHA TITRES IN MALARIA Table 1. Parasitologic and serologic survey data, by ages of persons sampled Parasitology I HA test results Area Age Number Parasite Percent Mean Mean examined rate (%) positive a titre positivetitre Gambela, 0-4 95 62.1 70.5 1 :74 1 :333 Ethiopia 5-14 101 78.2 68.3 1 :121 1 :405 15-29 100 46.0 83.0 1 :229 1 :657 30+ 100 48.0 87.0 1 :350 1 :765 Total 396 58.6 77.3 1:149 1 :530 Palawan, 0-4 206 5.7 38.8 1: 9 1: 91 Philippines 5-14 572 7.2 47.9 1:14 1 :114 15-29 326 7.4 74.5 1:107 1 :417 30+ 440 2.9 84.3 1:182 1:420 Total 1545 5.8 62.7 1:42 1 :257 Kalapara, 0-4 115 9.6 29.6 1: 6 1 :80 Bangladesh 5-14 477 6.7 51.9 1:12 1 :65 15-29 190 5.3 64.7 1 :28 1 :121 30+ 213 2.8 75.6 1 :71 1 :223 Total 995 5.9 55.7 1:18 1 :105 Balihar, 0-4 94 0 0 1 :2 - Bangladesh 5-14 51 0 13.7 1 :3 1 :29 15-29 26 0 57.7 1:19 1:97 30+ 34 0 76.5 1 :37 1:90 Total 205 0 23.4 1 :5 1 :78 Haiti 0-4 89 3.4 29.2 1 :5 1 :49 5-14 214 1.9 29.9 1:5 1:55 15-29 65 1.5 52.3 1:14 1:77 30+ 73 1.4 60.3 1 :36 1:244 Total 441 2.0 38.1 1:8 1 :86 alHAtitre>1 :16. in 1965; no falciparum parasites were detected in any of the subsequent years. The malaria eradication campaign had been initi- ated in Balihar in 1964 with bi-annual cycles of residual indoor insecticide spraying. E. In April 1970 a parasitologic and serologic survey was conducted in 6 localities (population 4775) on Haiti that had a history of unstable malaria. A total of 441 persons were examined, and the average parasite rate was 2.0%, without a marked concentration in any of the age groups. The sero- positivity rate was 29% in children under 5 years of age; it increased gradually to 60 % in the oldest age group. The mean positive titres were less than 1: 80 up to the age of 30 years and reached a maximum of 1: 244 in those aged 30 years and over. The sur- veillance activities in these localities included ACD and PCD. During the preceding 3 years, the number of detected cases had varied from 25 to 261 per year. The antimalaria measures had consisted of residual indoor insecticide spraying and mass drug adminis- tration. A comparison of the age-related population pro- files in the 5 study areas (Fig. 3) indicates that dif- ferences' in the malaria experience are detectable from the serologic observations, in contrast to the information provided by the parasitologic examina- tions. DISCUSSION Serologic observations on malaria infections in adults indicate that the antibody response is related to the frequency, level, and duration of parasit- aemia. However, no data are available on the influence of age on the development and persistence of antibodies. The serologic studies of people with parasitaemia in Gambela, Ethiopia, and in the Philippines suggest that the possibility of an age- related influence on the antibody development may 489 490 H. 0. LOBEL ET AL. have to be considered in the interpretation of sero- logic indices. The malaria antibody levels of the children with parasitaemia in the two areas comple- ment the blood-slide findings of the populations from which the individuals were drawn and suggest that the level of exposure was much higher for the subjects with parasitaemia in Gambela than for those in the Philippines. The relatively low antibody levels of the infected children in the Philippines are in contrast to the high titres of the adults in the same population. Similar observations have led to the suggestion by Voller et al. (16) and J. Lelyveld (unpublished observations, 1971) that the antibody levels in adults may be influenced by the cumulative effect of multiple infections and that adults may also have a greater immunological capacity than children to respond to malaria infections. The most notable parasitological difference be- tween the two populations described was that the parasite rates of the children were much higher in Gambela than in the Philippines; this difference indicates that the Gambela population was subject to more frequent antigenic challenges. This could explain why the children with malaria in Gambela had higher antibody titres than those in the Philip- pines. It may be that not only the frequency but also the variety of antigenic stimulations determines an age-related increase in the antibody titres, or that an age-dependent change in the ability to produce antibodies after infection is more apparent in indivi- duals who live in areas with a relatively low level of transmission. For instance, the high infection rates with the accompanying high antibody titre levels in children in an area such as Gambela may obscure an age-dependent immunological capability. The frequency distribution curves obtained in Haiti and in Balihar, Bangladesh, are characteristic of an area where a parasitic disease is absent or where the prevalence of infection is very low (5). These curves consist of a single low-titred peak resembling an exponential curve. The practically unimodal curves obtained for Gambela, Ethiopia, and for Rio Tuba, Philippines, indicate that these populations experienced a high level of exposure. The peak of the curve for Rio Tuba was at a higher titre level than that for Gambela, which suggests more recent or frequent infections in the Philippine individuals. The data from the serologic and parasitologic malaria surveys in 5 separate areas suggests that a serologic population profile can provide an indica- tion of the history and status of malaria in a popula- tion. Relating the survey results to the available malaria surveillance data is difficult because the sur- veillance information from these different areas is not readily comparable. However, the absence of malaria transmission during several years in Balihar, Bangladesh, is evident from both the serologic data and the surveillance information. In Gambela, Ethi- opia, the serologic and the parasitologic population profiles indicated a high level ofmalaria transmission. The serologic profiles from Kalapara, Bangladesh, from Palawan, and from Haiti, fall between these two extremes, and it is possible that, especially in such areas, with a " moderate " amount of malaria, the serological technique will be a useful complement to the parasitological parameters for characterizing the intensity and distribution ofmalaria transmission. The IHA method permits the testing of a large number of specimens in a relatively short period of time by a small laboratory staff. A recent paper (10) contained a report on several of the technical aspects that may affect the IHA test results for malaria. A relatively low degree of reproducibility (3, 15) is an inherent limitation of many IHA tests. Some factors influencing reproducibility include the anti- serum diluent and the microtitration loops used; variations in erythrocyte lots and their fragility and susceptibility to lysis and metabolic changes during storage (12, 1, 8); the concentration of the erythrocytes added to the diluted serum; the type, molarity, and pH of the buffer; the nature and con- centration of the protein stabilizer; and the length and temperature of the incubation. The use of com- plex antigenic material for sensitization also influ- ences the test results. Variability of results can be reduced by using several groups of positive and negative control sera, but this is time-consuming and inefficient. Farshy & Kagan (2) reported that red blood cells that are treated with pyruvic aldehyde, tanned, and fixed with glutaraldehyde,a and sub- sequently sensitized with a batch of antigen can be stored for several months without any substantial change in reactivity. The use of such a method in the IHA test for malaria may reduce the variability of the test results. A single nonhuman primate antigen, P. knowlesi, was used in the test. However, the use of homo- logous antigens would be desirable to increase the reactivity of the test, especially since the filter paper blood collection method results in some loss of sensitivity. A multi-species antigen would then gen- a 1,5-pentanedial. lIHA TITRES IN MALARIA 491 erally have to be used, because in population surveys it is seldom possible to know which Plasmodium species is responsible for, or has contributed to, the development ofmalaria antibodies. However, a single homologous antigen could be used to test specimens from areas where one Plasmodium species is known to predominate markedly. Serologic data reflect the accumulated malaria experience, i.e., the period prevalence (11), and the presence of antibodies can only indicate that the examined population is, or has been, infected with malaria. More meaningful estimates of malaria trans- mission can be obtained from age-related serologic population profiles. The ability of serologic surveys to detect short- term or temporary changes in the level of malaria transmission needs to be determined. A decline in the transmission level may be especially difficult to detect by serologic means and may necessitate per- iodic surveys of very young children or surveys to determine the seroconversion rates. The experience with the IHA test for malaria sug- gests that it may become an important procedure in the armamentarium of malaria epidemiologists. Further laboratory investigations will be necessary to improve the standardization of test results. Field studies need to be carried out in areas with different malariological conditions to permit better inter- pretation and understanding of the results in relation to other malariometric measurements and to deter- mine the optimum ways in which the method can be applied to the study of malaria epidemiology. RJSUM1 INTERPRETATION DES TITRES D'HEMAGGLUTINATION INDIRECTE DANS L'ETUDE DE L'EPIDEMIOLOGIE DU PALUDISME Les auteurs exposent la technique de 1'6preuve d'h6mag- glutination indirecte au moyen d'un antigene Plasmodium knowlesi et examinent les r6sultats obtenus au cours d'une serie d'enquetes 6pid6miologiques sur le paludisme. L'6tude comparative des donn6es serologiques et parasito- logiques conduit a admettre une influence de I'age sur la production des anticorps dont il faut tenir compte dans l'interpretation des r6sultats. On etablit un parallele entre les r6ponses en anticorps et les taux de positivit6 des examens parasitologiques et autres donn6es paludom6- triques dans des populations du Bangladesh, d'Ethiopie, de Haiti et des Philippines. II semble, d'apres les courbes de r6partition des anticorps en fonction de l'age, que la s6rologie puisse fournir des informations int6ressantes sur les ant6c6dents et la situation presente en matiere d'infec- tion palud6enne dans une population. Les critares 6pidemiologiques utilises au cours des programmes de lutte antipaludique sont fond6s sur une m6thode parasitologique dont les possibilites sont limi- tees: elle n'indique en effet que l'absence ou la pr6sence de parasitemie a un moment determin6. On a donc propose d'avoir recours aux epreuves s6rologiques pour 1'etude de 1'epid6miologie du paludisme. Plusieurs techniques de re- cherche des anticorps peuvent etre utilis6es a cet effet. Leur avantage principal est leur aptitude a determiner l'intensi- te et la r6partition de la transmission du paludisme au sein d'une population. La collecte des 6chantillons de sang sur papier filtre facilite beaucoup les enquetes sur le terrain. REFERENCES 1. DANON, D. ET AL. Transfusion, 4, 339-342 (1964). 2. FARSHY, D. C. & KAGAN, I. G. Amer. J. trop. Med. Hyg., 21, 868-872 (1972). 3. KAGAN, 1. G. Amer. J. trop. Med. Hyg., 21, 683-689 (1972). 4. KAGAN, I. G. Experimental Parasitol., 31, 126-135 (1972). 5. KAGAN, I. G. Serological epidemiology, New York, Academic Press, 1973, Ch. 13. 6. KAGAN, I. G. ET AL. Bull. Wld Hlth Org., 41, 825-841 (1969). 7. KAGAN, I. G. ET AL. Bull. N.Y. Acad. Med., 45, 1027-1042 (1969). 8. MARKS, P. A. & JOHNSON, A. B. J. clin. Invest., 37, 1542-1548 (1958). 9. MATHEWS, H. M. ET AL. Amer. J. Epid., 92, 376-381 (1970). 10. MEUWISSEN, J. H. E. T. ET AL. Bull. Wld Hlth Org., 46, 771-782 (1972). 11. PAYNE, A. Milbank Mem. Fund. Quart., 43 (2), Part 2, 245-305 (1965). 492 H. 0. LOBEL ET AL. 12. POULIK, M. D. & LAUF, P. K. Clin. exp. Immunol., 4, 165-175 (1969). 13. PULL, J. H. Amer. J. trop. Med. Hyg., 21, 651-637 (1972). 14. ROGERS, W. A., JR ET AL. Amer. J. trop. Med. Hyg., 17, 804-809 (1968). 15. SCHUBERT, J. H. & CORNELL, R. G. J. Lab. clin. Med., 52, 737-743 (1958). 16. VOLLER, A. ET AL. J. trop. Med. Hyg., 74, 45-52 (1971). 17. Wld Hlth Org. techn. Rep. Ser., 1971, No. 467 p. 32.