Immuno-epidemiology of malaria

Bull. Org. mond. Sante 1973, 49, 267-274Bull. Wid Hlth Org. Immuno-epidemiology of malaria A study in a tribal area of West Irian H. J. VAN DER KAAY,1 F. KLEIN,2 M. HAGENAAR-DE WEERDT,3 & J. H. E. T. MEUWISSEN' An investigation of malariometric indices in relation to immunoglobulin levels, rheuma- toid factors, and antithyroglobulins was carried out on 78 members of the Arfak tribe near Manokwari in Western New Guinea, in the course ofa WHO assessment ofmalaria control activities in that region. The population investigated had been exposed to a period of epidemic malaria, as indicated by the small differences in malariometric indices between consecutive age groups. Typically high spleen sizes were recorded, asfoundgenerally among Papuans in similar situations. Falciparum malaria was most prevalent, almost equal to cases of vivax and malariae malaria together. IgM levels were very high, while those ofIgG, IgA and IgD were not elevated. Total serum protein was rather low. No correlation between malariometric indices, autoantibodies, and immunoglobulin levels could be found. In particular there was no correlation between IgM levels and spleen indices, such as has beenfound in many other surveys. It is suggested that splenomegaly may show no correlation with the IgM level in Papuan populations without previous selection. This study among the Arfaks living in Manggo- Api near Manokwari was carried out during a WHO assessment of the malaria control activities in Irian Barat from February till April 1970. Besides obtain- ing malariometric indices, the laboratory examina- tions included measurements of indirect fluorescent antibody titres, levels of immunoglobulins A, G, M, and D, the titres of some rheumatoid factors and antithyroglobulins, and quantitative electrophoresis of the sera. The purpose of this communication is to provide material for comparison with similar investigations carried out in Papua-New Guinea and in various parts of Africa. The area around Manokwari, commonly inhabited by the Arfak tribe, is mountainous with only narrow stretches of lowland along the coast. This coastal area, including Manokwari and the nearby flat Keber valley, was formerly classified as hyperendemic for malaria (van Dijk, unpublished report, 1962 a). In 1 Chief Scientist, Institute of Tropical Medicine, Rapen- burg 33, Leyden, The Netherlands, to whom correspondence should be addressed. 2 Chief Scientist, ' Analyst, Department of Epidemiology, Medical Faculty, Erasmus University, Rotterdam, The Netherlands. 'Senior Lecturer, Institute for Medical Parasitology, University of Nijmegen, The Netherlands. the early 1950s, measures to control malaria were started on a small scale in Manokwari and were later expanded to include the surrounding villages. The methods used included oiling of breeding places in the central part of Manokwari, residual indoor spraying with DDT, and mass drug administration of a combination of chloroquine and pyrimethamine. Interruption of transmission was not achieved in the area (Meuwissen, unpublished report, 1960 a). During a precontrol survey in 1955 Plasmodium vivax was found to be the most prevalent parasite species, while in 1960 it was P. falciparum. This change in prevalence was also seen in other areas of Irian Barat where malaria control had some effect on the transmission of the disease (1; Meuwis- sen, unpublished report, 1960 a). P. malariae also remained present throughout the period 1955-60. The known malaria vectors of the area are Anopheles farauti, A. punctulatus, A. longirostris, and A. koliensis. All except A. koliensis were recorded during the survey as adult mosquitos, as larvae, or as both. The average annual rainfall for Manokwari for the 10-year period 1956-65 was 2 547 mm, spread evenly throughout the year ensuring perennial breed- ing of mosquitos. Daily minimum and maximum a Obtainable from the authors of this article. 3111 - 267- H. J. VAN DER KAAY ET AL. temperatures were in the range between 22.4°C and 31.9°C. MATERIALS AND METHODS The present malaria survey was carried out in the village of Manggo-Api where a group of 150 people of the Arfak tribe had settled during the previous 6 months. A sample of 99 persons was drawn from this population on a family basis. Serum could be obtained from 88 persons, but only 78 samples were available in sufficient quantity to allow deter- mination of some nonmalarial factors. Spleen examination. This was carried out on all subjects, while lying on their backs with their knees flexed. The sizes of spleens were recorded using Hackett's classification. Spleen rates and a weighted average, the average enlarged spleen (AES), were calculated as recommended by the World Health Organization (2). Blood examination. Thick blood smears from fin- ger pricks were dried, stained with Giemsa stain within 24 h, and examined under the microscope for malaria parasites in 200 oil-immersion fields. To calculate the parasite densities, parasites were counted against 200 leucocytes and then multiplied by 40, an average of 8 000 leucocytes being taken as a standard. These data were used to calculate the parasite rate (2) for the relevant age groups, as well as the positive parasite density index (PPDI) by means of the scale proposed by Bruce-Chwatt (3). Venous blood samples were collected from the 88 individuals for further tests in Vacutainer tubes, which were placed in racks for 6-12 h to allow clot formation and separation. The serum was then stored in liquid nitrogen. The indirectfluorescent antibody test (IFA). Malar- ial antibody levels were measured by the micro test method, developed by Meuwissen (4) and modified by Voller and O'Neill (5), P. fieldi being used as the antigen. The individual titre index was found by calculating the exponent to the basis of 2 after hav- ing divided the titre by 10. The titre of <1: 20 (= 1: lOx 20) thus corresponded to a titre index of 0, while a titre of 1: 20 (= 1: 1Ox 21) corresponded to a titre index of 1, and so on. The mean titre index (MTI) was calculated by dividing the sum of indi- vidual titre indices by the number of observations. Immunoglobulin determinations. These were car- ried out by the single radial diffusion method of Mancini, as modified by Kalff (6). The results are expressed in international units instead of mg/ml, following the recommendations of Rowe et al. (7). International Reference Preparation 67/97 was used as a standard for IgG, IgA and IgM, while standard preparation 67/37 was used for IgD (8, 9). Mono- specific antisera against IgG, IgA and IgM were prepared in our laboratory in rabbits, and a rabbit antiserum against IgD was obtained from the Cen- tral Laboratory of the Dutch Red Cross Blood Transfusion Service, Amsterdam. For the determi- nation of IgD it was necessary to treat the Mancini plates for 1 h with 4% tannic acid to obtain better visualisation of the precipitation rings. The standard deviation of the method was ±3.7% for IgG, ±3.1% for IgA, ±3.7% for IgM, and ±1.2% for IgD. The validity of the Mancini method for the determination of IgM is somewhat questionable. Monomeric IgM has been found in some patients with tropical diseases in Africa (10) and its presence could give rise to false values for the IgM concen- tration, since it reacts differently with anti IgM sera. Owing to insufficient serum, the presence of mono- mers could not be assessed in our sample. However, even polymeric and polyclonal IgM preparations may give widely divergent results in the Mancini test when equal quantities of IgM from different individuals are analysed (de Bruyn & Klein, unpub- lished observations, 1973). No satisfactory method to eliminate all these errors has yet been developed and IgM figures should therefore be regarded with some reserve. Electrophoresis. Electrophoresis of serum samples was performed on cellulose acetate membranes using the Beckman Microzone apparatus according to the instructions supplied by the manufacturer. It proved to be necessary to dilute the sera with an equal volume of saline to obtain a satisfactory degree of linearity for scanning with the Joyce Loebl Chromo- scan MK II densitometer. The results were recorded as percentages of the total protein, which was deter- mined by the biuret method. Since the P,2 fraction, if present, was almost completely incorporated in the nearly always large y fraction, P2+Y were taken together as the total immunoglobulin fraction. The standard deviation of the method was ±6.1% for the albumin fraction, ± 13.3% for the a+# fraction, and ± 10.7% for the P22+Y fraction. The latex fixation test. This test with Cohn frac- tion 2 (16% solution obtained from the Central Laboratory of the Dutch Red Cross Blood Trans- 268 IMMUNO-EPIDEMIOLOGY OF MALARIA fusion Service, Amsterdam) was carried out as described by Valkenburg (11). The results are given for untreated sera with titres >640 and for sera previously heated to 560C for 30 min (reactors irrespective of titre). The latter procedure inactivates an inhibitor that can mask weak reactions. It should therefore give a better estimate of rheumatoid factors (RF) than the test with untreated serum in which the net effect of RF and inhibitor is measured (12). However, some agglutinators in the latex test proved to be thermolabile, contrary to common experience with European sera, and this led to a slight decrease in the sensitivity of this test. The sera with titres > 640 without inactivation are considered " clinically positive " by rheumatoid serologists and are only given for comparison with data in the literature because such borderline titres have little meaning here. The Waaler-Rose test (HEAT). This was carried out with human ORh+ erythrocytes sensitized with rabbit whole anti-human serum (13). Sheep cells would have been less suitable here, because of the frequent occurrence of heterophile agglutinins. For every test the total number of reactors, irrespective of titre value, has been given as well as the number of sera counted as "clinically positive" in the serologic diagnosis of rheumatoid arthritis, i.e., with titres of > 32. Antithyroglobulins. These were measured using the test kit supplied by Wellcome Laboratories Ltd., England. Since this test system uses tanned sheep red cells coated with thyroglobulin, difficulties were sometimes encountered when the amount ofuncoated cells used for the absorption of heterophile agglu- tinins proved to be insufficient. Some 15 sera could not be read with certainty, but all of these could have had only marginal titre values at most (i.e. <1: 5). RESULTS Blood examination. Table 1 shows the highest parasite rate (about 75 %) to be in the 2-9-year age group, gradually decreasing to about 37% for adults. P. falciparum was the most prevalent species for the whole group examined, with P. vivax second. P. ma- lariae was found particularly among those up to 20 years of age. The gametocyte rate of the P. falci- parum carriers was 19% for the 2-9-year age group. Of the 4 mixed infections 3 were among children of 5 years and younger. The positive parasite density index was 2.6 for the 24-year age group, gradually decreasing to about 1 for those of 20 years and older. Spleen examination. The spleen rate, also given in Table 1, was 90% for the whole group and remained remarkably high since even for those of 40 years and older it was still 70 %. A spleen class 3-5 was found in about two-thirds of the individuals examined. The AES, being 3.1 in the 5-9-year age group, was still 2.1 for those 40 years and older. Table 1. Malariometric data from a survey of the Arfak tribe, West Irian Blood examination for parasites Spleen examination Age No + /No Crude Parasite species No. Spleen groups No.x /No para- PPDI enlarged/ rt E(years) exmned site P acaPvvxP.m iedNo. rae E rate (%) .vvxP a.mxdexamined (% 2-4 5/6 83 4 (2) 2 1 2 2.6 6/6 100 2 5-9 18/25 72 12 (4) 2 5 1 2.2 24/25 96 3.1 10-19 12/19 63 5 (1) 4 4 1 1.9 18/19 95 3 20-29 6/19 32 3 (0) 3 0 0 1.2 17/19 89 2.6 30-39 7/14 50 4 (1) 1 2 0 1.0 13/14 93 2.6 > 40 5/16 31 2 (0) 3 0 0 1.1 11/16 69 2.1 total 53/99 53 30 15 12 4 89/99 90 a Within parentheses is the number with P. falciparum gametocytes. 269 H. J. VAN DER KAAY ET AL. Table 2. Distribution in different age-groups of IFA titres in Arfak sera and of mean titre indices Age groups (years) IFA titre Total 5-9 10-19 20-29 30-39 >40 1:20 1 1 0 0 0 2 1:40 5 1 1 2 3 12 1:80 4 2 4 2 2 14 1:160 11 13 7 7 4 42 1:320 1 2 5 2 5 15 1:640 0 0 1 0 2 3 total 22 19 18 13 16 78 MTI 3.3 3.7 4.1 3.7 4.1 The indirect fluorescent antibody test. The results are given in Table 2 as a frequency distribution of IFA titres for each age group. Each group showed the same range of individual titres. The MTI showed only a slight gradual increase from 3.3 in the 5-9-year age group to 4.1 for those of 20 years and older, no convincing differences being found in our sample among the various age groups. Immunoglobulin determinations. The results ex- pressed in international units per ml are summarized in Table 3 by age groups. Our number of samples is small, the range of the results is wide, and accord- ingly the standard deviations are considerable. This applied to each Ig class determined and to each age group presented in the table. Unfortunately no samples from people below the age of 5 years were available. The mean IgM level for males was higher than for females, with a significance of P=0.06 in Wilcoxon's test and 0.025< P< 0.050 in the test of Student-Welch. In the levels of the other immuno- globulin classes no significant differences between the sexes were found. The relationship between immunoglobulins and malarial indicators was investigated by studying the scatter diagrams relating the levels of the immuno- globulin classes to spleen indices and fluorescent antibody titres. No significant correlation could be found, either through regression analysis or by test- ing by the methods of Student-Welch and Wilcoxon the differences of mean Ig levels between groups with high and low spleen indices or IFA titres. A scatter diagram of spleen size against IFA titres did not give a significant correlation in linear regres- Table 3. Immunoglobulin levels in 78 Arfak sera expressed in international units/ml serum 'g Age group No. of Mean S.D. Rangeor sex sera level IgA 5- 9 16 63 22 26- 101 10-19 16 69 25 24- 119 20-29 1 7 56 1 4 25- 75 30-39 13 77 34 38- 142 > 40 16 79 32 36- 128 total 78 68 27 24- 142 IgG 5- 9 16 197 37 133- 277 10-19 16 174 27 126- 216 20-29 17 174 49 65- 272 30-39 13 189 33 142- 250 > 40 16 170 34 109- 225 total 78 181 38 66- 277 gM 5- 9 16 694 1 098 111-3699 10-19 16 696 539 139-2 507 20-29 17 1 436 1 630 170-6862 30-39 13 1 294 1 116 413-4 614 > 40 16 1 556 1 211 548-4438 M 44 1 365 1 470 111-6 862 total F 34 833 664 127-3 000 M + F 78 1 133 1 211 111-6862 IgD 5- 9 16 31 27 S3- 83 10-19 16 34 37 63- 149 20-29 17 21 23 .3- 91 30-39 13 19 18 .3- 66 > 40 16 40 44 .3- 135 total 78 29 32 <3- 149 sion analysis (Fig. 1). A x2 test for the groups with spleen size 0-2/3-5 and IFA titres 20-80/160-640 yielded a P value between 0.05 and 0.10, to which no great importance can be attached in view of the rather arbitrary divisions between high and low in the respective groups. Electrophoresis. Table 4 shows that rather low values for total protein and in particular for albumin were found. No monoclonal components were found in the y-region of any of the sera investigated. 270 IMMUNO-EPIDEMIOLOGY OF MALARIA Table 4. Total protein and electrophoretic fractions (g per litre of serum) of 78 Arfak sera Mean SD Range total 62 14 37-99 protein 02+Y 26 7 15-46 globulin a + ,8 14 4 8-25 globulin albumin 22 6 11-37 5' SPLEEN 4ZE t4. 3. 2. 1. 0. Fig. 1 sizes. Tes the re reacto not yi larly r rheurm of par parasi no in variab Malar The cate t] * *I Table 5. Autoantibodies in Arfak sera Test No. investigated No. positive (%) Waaler-Rose reactors 77 10.4 Waaler- Rose " clinically positive "a 77 5.2 Latex reactors 67 47.8 Latex " clinically positive" b 77 4,4 Antithyroglobulin reactors 76 3.9 a titres > 32. b titres > 640. + *tj* * experienced a period of high transmission of malaria. More than 60% of children over the age of 5 years - - as well as adults of all ages have a spleen of at least class 3 following Hackett's classification, both sexes being equally affected. This is not uncommon among * * populations living in areas of New Guinea with very intensive malaria transmission (1, 14). The parasite rate decreased with age, though there were only a .40 . . . few cases with high parasite densities. The PPDI IFA was highest (2.6) among the 2-4-year age group. No IFA titres were found over 1: 640; this could .Scatter diagram of IFA titres against spleen be due to the use of P.fieldi as the antigen. Moreover the MTI showed only a slight increase by age. A prob- able explanation is that this group of Arfak people had settled in this village near Manokwari only aboutts for rheumantoid factors. Table 5 summarizes 6 months prior to our survey. In this area the inten- sults. x2 analysis of the distribution of latex sity of transmission was lower, owing to the control ers versus IFA titres above or below 1: 160 did measures carried out by the health department inield a significant result (0.20<P<0.30). Simi- Manokwari and its immediate surroundings, than iorelation was found between the presence of in the area inhabited by this tribe prior to their atoid factor in the latex test and the presence settlement. However, the intensity of transmission asitaemia (0.05<P<0.10).Scatterdiagramsof in the mountainous area is known to show con- te counts against immunoglobulin levels gave siderable variation (Meuwissen, unpublished report, dication of any correlation between these 1960). Migration from a relatively malaria-free les. highland area across malarious valleys to the coastal area, could have subjected these people temporarily DISCUSSION to epidemic malaria. This might explain why the malariometric data and IFA titres did not differia assessment much in individuals of consecutive age groups. This high spleen rate and the AES (Table 1) indi- could have been due to a relatively incomplete hat this group of the Arfak tribe must have immunity before migration. 271 H. J. VAN DER KAAY ET AL. Immunoglobulin levels A comparison of the results of immunoglobulin levels with the published data is hampered by the fact that most authors expressed their results in abso- lute units (mg/100 ml), prior to the demonstration by Rowe et al. (7) that purified immunoglobulin gave a much larger spread in the estimation of immunoglobulin levels than properly chosen stan- dard sera. Nevertheless our data were recalculated in mg per 100 ml of serum, using the factors given by Rowe et al. (7). Thus Table 6 allows a rough comparison of our values with those of Wells (15) from New Guinea, and with the results of surveys conducted in Tanzania and Nigeria (16, 17), and in the Netherlands on data drawn from the work of Kalff (18). The IgG and in particular the IgA levels observed were remarkably low, which may partly be the consequence of a low total protein value reflecting itself in a smaller total globulin fraction. The ranges for IgG and IgA were also lower than those for West and East African populations. The ranges for IgM however are among the highest published so far. As in the group in New Guinea with splenomegaly, the relative quantity of IgM in our sera is considerable (15). In West African populations the level of IgG is very high, while IgM and IgA are about equal. This " African " immuno- globulin pattern was confirmed for Tanzania by Voller et al. (16), for Liberia by Capucinelli et al. (19), and for Liberia and Nigeria by Muller (20). Contrary to what has been observed elsewhere (18, 20, 21) but in agreement with Wells (15), we found higher mean IgM values for males than for females. Although this difference is of borderline significance, Table 6. Mean approximate values of serum immuno- globulins, expressed in mg/100 ml, in various popula- tion samples Watut Valley, Dutch Usam- Mano- Eastern popu- bara, lbadan,Ig kwari New Guinea latiun bara- Nigeradan(Arfak) (Splenomegaly latiol Tan. Nigeria c group) a sml ai IgG 1 452 1 882 1 246 3395 2 500 1gM 963 1 210 137 238 189 IgA 97 152 178 not done 183 a Data from Wells (15). b Data from Voller et al. (16). c Data from Sagoe (17). it suggests that this possibility cannot be excluded and should be kept in mind in future work. As for IgD, our mean value (Table 3) is only a little higher than that for a group of 30 normal Europeans 18-42 years old (Zegers et al., unpub- lished data) but the range is much larger (European mean 22 IU/ml, range 7-70 IU/ml). Similar findings were reported by Rowe et al. for Gambians (22). Colwell et al. reported high levels for IgD in a Viet- Namese sample but not in Cambodians living in the same village (23). As these figures are given as abso- lute values, read from an IgD paraprotein standard, they should be regarded as rather inconclusive (see also Hobbs et al. (24)). In agreement with us, Colwell et al. found no significant correlation between IgD levels and malarial indicators. In this study, immunoglobulin levels for the 4 immunoglobulin classes could not be related to any of the malarial indicators. This is contrary to what has been found for IgM levels and IFA titres (16), IgM levels and splenomegaly (15), and IgG levels and parasitaemia (21). However, the scatter diagram for spleen size against the IgM level given by Sagoe (17) showed a similar lack of correlation for a group of selected Nigerian patients. This was also observed in some cases in Papua-New Guinea by Crane et al. (14), who suggested that spleen size and the IgM level behaved to some extent as indepen- dent variables. Nevertheless, there can be little doubt about the causal relationship between splenomegaly and malaria in view of the effect of treatment with antimalarials. If such a relationship does not become apparent in the immunoglobulin pattern, it may be that in splenomegaly the hyperstimulation of the immune apparatus causes atypical reactions that could border on malignancy, as suggested by Sagoe (17). Such independent responses may obscure any correlations that are found in the absence of this form of immune condition. The correlations between splenomegaly and immunoglobulin levels by Wells (15) in Papua-New Guinea populations may be explained by differences in malaria trans- mission, selection of the samples, and genetic con- stitution. Metselaar & Van Thiel (25) had previously drawn attention to the fact that in Papuans the spleen rate very often ran a course deviating from other malarial indicators and suggested that it was not suitable as a criterion for classification. It should also be realized that arranging spleen sizes, or IFA titres, or both in groups below or above a certain value means drawing arbitrary lines in a continuum and accordingly the significance of X2 tests may 272 IMMUNO-EPIDEMIOLOGY OF MALARIA depend on the position of such lines. This can be judged from Fig. 1, where there is a cluster in the middle of the diagram, caused by the two coinciding maxima of spleen size and IFA titre distributions. The result of the x2 test in similar cases may depend on the field in which such a cluster happens to fall, thus giving a false suggestion of significance. Our total serum protein values are rather low compared with other surveys in New Guinea (15, 26, 27, 28). To a large extent this could be due to the low albumin fraction. These results could pos- sibly be explained by the less satisfactory nutritional status of the group and the high degree of various infections. Autoantibodies Rheumatoid factors and antithyroglobulins were found much less frequently than in the surveys by Wells (15) in New Guinea, and by Houba & Alli- son (29) and Shaper et al. (30) in Africa. The number of " clinically positive " titres in the latex fixation test is comparable with that found previously by Valkenburg (31) in another Western New Guinea survey, but the number of positive Waaler-Rose tests is higher. Both tests are higher than in a Dutch control population, where 1.2% of the latex tests and 1.8% of the Waaler-Rose tests were " clinically positive". In the latex test a rather high number of low " subclinical " titres was found. The very large difference with the figures of Wells (15) may be partially explained by his use of a slide latex test, which could give false positive results caused by Qq (12). No such explanation presents itself for the Waaler-Rose test or for the antithyroglobulin test. The lower frequency of rheumatoid factor reactions seems to accord with the results from experi- mental malaria in monkeys, where such reactions appeared to be rather transitory and became fewer as the frequency of infection increased. Antithyro- globulins were found infrequently in these experi- ments (Klein, Meuwissen, & Kornman-van den Bosch, unpublished data). Any correlations of rheumatoid serologic results with other indicators, whether immunoglobulin levels or malaria, were conspicuous by their absence. In an African group with splenomegaly, Klein & Mattern (32) found a significant difference in the frequency of positive Waaler-Rose tests in individuals with and without an elevated IgM level, but this was not observed in our sample. An examination of the scat- ter diagrams relating latex reactors with parasite rates, spleen indices, and immunoglobulin levels clearly demonstrates this absence of correlation. In summary, the results discussed here indicate that, contrary to many observations, immunoglobu- lin levels, in particular IgM, may not correlate with splenomegaly when measured in a population with- out selection. There is also no suggestion here of any correlation of rheumatoid factors with malarial indicators as was claimed by Greenwood et al. (33). The irregular occurrence of autoantibodies may be explained by their transitory appearance and sub- sequent depression as demonstrated in experimental malaria (see also Klein et al. (34)). ACKNOWLEDGEMENTS We thank the Government of Indonesia and in partic- ular Professor J. Sulianti Saroso, Director General of CDC, Djakarta, for facilities extended during the assess- ment in West Irian. We are grateful to Dr B. K. Kawen- gian, Director, CDC, Province of West Irian, and the staff of the Malaria Unit, and Dr R. 0. Darwish, Regional entomologist, WHO Regional Office for South-East Asia, for assistance during the fieldwork. Financial and material assistance was provided by the World Health Organization. We thank Mrs B. Gommans-Kuyer for the IFA examinations and Miss J. M. B. Smulders and Miss M. Demmendal for carrying out the rheumatoid serologic tests. Thanks are due to Professor H. A. Valkenburg and Dr A. Voller for their critical reading of the manuscript. RltSUMIe IMMUNO-tPIDIMIOLOGIE DU PALUDISME: ETUDE DANS UNE TRIBU DE L'IRIAN OCCIDENTAL Une 6tude de certains indices paludom6triques (indice plasmodique, esp&ce et pr6valence des parasites, volume de la rate, titres d'anticorps fluorescents) et de leur relation avec les titres d'immunoglobulines, de facteurs rhumatoldes et d'antithyroglobulines a et6 faite parmi les 78 membres de la tribu Arfak, en Irian Occidental. Des r6sultats de l'analyse 6lectrophor6tique de s6rums sont aussi presentes. 273 274 H. J. VAN DER KAAY ET AL. On n'a releve que de faibles differences des indices paludom6triques dans les groupes d'age successifs, ce qui semble indiquer que cette population a ete expos&e occasionnellement a une 6pid6mie de paludisme, proba- blement au cours d'une migration. Plasmodiumfakciparum 6tait l'espece pr6dominante, presque aussi fr6quente que P. vivax et P. malariae reunis. Des indices spl6niques 6lev6s ont ete enregistr6s, comme il est de regle chez les Papous en cas de trans- mission intense du paludisme. Les taux de prot6ines s6riques totales, et en particulier les taux d'albumine, 6taient peu eleves comparativement a ceux constat6s en d'autres occasions en Nouvelle-Guin6e. Les taux d'IgM etaient tres eleves avec des variations tres etendues, tandis que les taux d'IgG et d'IgA restaient remarquable- ment faibles. Les valeurs d'IgD pr6sentaient une gamme plus large, mais leur moyenne 6tait comparable A celle observ6e en Europe. Les observations relatives aux immunoglobulines diff6raient de celles faites en Afrique ou les taux d'IgG sont plus 6lev6s, mais concordaient plus ou moins avec les r6sultats d'autres enquetes en Nouvelle- Guinee. On n'a not6 aucune correlation entre la teneur en immunoglobulines et les indices paludom6triques et, en particulier, entre les taux d'IgM et le volume de la rate. Les facteurs rhumatoldes et les antithyroglobulines ont ete rarement d6cel6s; ils 6taient sans rapport avec les taux d'immunoglobulines et les indices paludom6triques. Les auteurs estiment que dans une population non selectionnee la splenomegalie peut se d6velopper comme un processus plus ou moins ind6pendant de l'evolution de l'infection palud6enne et sans relation directe avec les taux d'immunoglobulines. REFERENCES 1. MErSELAAR, D. Transactions of the Royal Society of Tropical Medicine and Hygiene, 54, 523 (1960). 2. WoRLD HEALTH ORGANIZATION, Terminology of malaria and of malaria eradication, Geneva, 1963. 3. BRUCE-CHWATr, L. J. Transactions of the Royal Society of Tropical Medicine and Hygiene, 52: 389 (1958). 4. MEUWISSEN, J. H. E. TH. ET AL. In: Abstracts and Reviews 8th Int. Congresses on Tropical Medicine and Malaria, Teheran, 7-15 Sept. 1968, p. 1432. 5. VOLLER, A. & O'NEELL, P. Bulletin of the World Health Organization, 45: 524 (1971). 6. KALFF, M. W. Clinical biochemistry, 3: 91 (1970). 7. RowE, D. S. ET AL. Bulletin of the World Health Organization, 46: 67 (1972). 8. RowE, D. S. ET AL. Bulletin of the World Health Organization, 42: 535 (1970). 9. RoWE, D. S. ET AL. Bulletin of the World Health Organization, 43: 607 (1970). 10. KLEIN, F. ET AL. Immunology, 13: 641 (1967). 11. VALKENBURG, H. A. In: Kellgren, J. H., ed. The Epidemiology of Chronic Rheumatism. Oxford, Blackwell, 1963, vol. 1, p. 337. 12. KLEIN, F. ET AL. Immunology, 10: 87 (1966). 13. VALKENBURG, H. A. In: Kellgren, J. H., ed. The Epidemiology of Chronic Rheumatism. Oxford, Blackwell, 1963, vol. 1, p. 330. 14. CRANE, G. G. ET AL. Transactions of the Royal Society of Tropical Medicine and Hygiene, 66: 724 (1972). 15. WELLS, J. V. Transactions of the Royal Society of Tropical Medicine and Hygiene, 64, 53 (1970). 16. VOLLER, A. ET AL. Journal of tropical medicine and hygiene, 74: 45 (1971). 17. SAGOE, A. S. British medical journal, 3: 378 (1970). 18. KALFF, M. W. Clinica chimica acta, 28: 277 (1970). 19. CAPUCCINELLI, P. ET AL. Journal of tropical medicine and hygiene, 75: 180 (1972). 20. MULLER, A. S. Population studies on the prevalence of rheumatic diseases in Liberia and Nigeria. Thesis, Leiden, 1970. 21. MACGREGOR, I. A. ET AL. Clinical and experimental immunology, 7: 51 (1970). 22. RowE, D. S. ET AL. Clinical and experimental immunology, 3: 63 (1968). 23. COLWELL, E. J. ET AL. Transactions of the Royal Society of Tropical Medicine and Hygiene, 65: 310 (1971). 24. HOBBS, J. R. ET AL. Lancet (2): 614 (1966). 25. METSELAAR, D. & THIEL, P. H. VAN. Tropical & geographical medicine, 11: 157 (1959). 26. LUYKEN, R. ET AL. Tropical & geographical medicine, 10: 385 (1958). 27. JANSEN, A. A. J. Nutrition, infection and serum proteins in Papuans of Netherlands New Guinea. Thesis, Utrecht, 1959. 28. CURTAIN, C. C. ET AL. American journal of tropical medicine and hygiene, 14: 678 (1965). 29. HOUBA, V. & ALLISON, A. C. Lancet (1): 848 (1966). 30. SHAPER, A. G. ET AL. Lancet (1): 1342 (1968). 31. VALKENBURG, H. A. In: Kellgren, J. H., ed. The Epidemiology of Chronic Rheumatism. Oxford, Blackwell, 1963, vol. 1, p. 74. 32. KLEIN, F. & MATTERN, P. Annals of the rheumatic diseases, 24: 458 (1965). 33. GREENWOOD, B. M. ET AL. Clinical and experimental immunology, 9: 161 (1971). 34. KLEIN, F. ET AL. In: Grubb, R. & Samuelsson, G., eds. Human anti-human gammaglobulins. Oxford, Pergamon Press, 1971, p. 63.