Newly transmitted Plasmodium falciparum malaria in the central highland plateaux of Madagascar: assessment of clinical impact in a rural community.

Newly transmitted Plasmodium falciparum malaria in the central highland plateaux of Madagascar: assessment of clinical impact in a rural community J.P. Lepers,' P. Deloron,2 M.D. Andriamagatiana-Rason3, J.A. Ramanamirija,4 & P. Coulanges5 Plasmodium falciparum has recently reappeared in the central highland plateaux of Madagascar. To define its role in the overall malaria pathology occurring during the rainy season, we conducted a clinical and parasitological study among the 2776 consultants of a malaria field health centre which we established in the middle of the transmission area. From January to June 1988 the overall parasite rate was 63.2% and did not vary with age; the overall splenic rate was 46.9%, decreasing in individuals over 15 years of age. After the beginning of the main transmission season, an increase in parasite rates and parasite densities was observed, while febrile illnesses associated with malaria parasitaemia decreased. During the second part of the rainy season the parasite densities started to decrease, while parasite rates and malaria- associated fevers remained at the same level. The frequency of fevers was not related to the malaria parasite density owing to individual and seasonal variations. Malaria due to Plasmodium falciparum has recently reappeared in the central highland plateaux of Madagascar (1) and, according to some reports, caused very high morbidity and mortality among the rural inhabitants of this area. Very little is known concerning the epidemiology and morbidity associated with this disease, but control measures are urgently needed. Epidemiological data on the disease and its effects in the population are essential prior to developing control strategy guidelines. We report here the results of a study on the role of P.falciparum malaria in the centre of the transmission area. Subjects and methods Study are The study was conducted in Manarintsoa, a village at 1200m altitude with about 1550 inhabitants, in the ' Head, Malaria Research Unit, Institut Pasteur of Madagascar, Antananarivo, Madagascar. 2 Head of Research, INSERM Unite 13 and Institut de M6decine et d'Epidemiologie Africaines (IMEA), H6pital Claude Bernard, 75944 Paris, France. Requests for reprints should be sent to this address. 3 Biologist, Malaria Research Unit, Institut Pasteur of Madagas- car. 4 Physician, Malaria Research Unit, Institut Pasteur of Madagas- car. 5 Director, Institut Pasteur of Madagascar. Reprint No. 5065 central highland plateaux, 20 km west of Antanana- rivo, the capital city. The characteristic tropical mountain climate has a dry cold season (June to November) and a hot rainy season. Intensive antivec- torial control activities and mass drug administration conducted in the fifties led to the disappearance of Anopheles funestus in 1954. Since 1960, malaria eradication was obtained in nearly the whole plateaux area except for two small foci where seasonal transmission persisted at a reduced rate (2). In the mid-seventies, A. funestus reappeared. In early 1987, an unusual increase in the mortality of the population of Manarintsoa revealed the epidemic reappearance of falciparum malaria in the plateaux area (1). Since then, malaria seems to be highly endemic (P.falciparnm being by far the most encoun- tered species) and transmitted by both A. funestus and A. gambiae s.l. (3). Subjects Investigated In March 1987, we established in Manarintsoa a field health centre (the only one within an 8 km range) where the population has free access to medical advice and drugs. From January to June 1988, all invididuals presenting at this health centre were examined clinically, the axillary temperature was measured with an electronic thermometer, spleen size was recorded, and blood was collected by fingerprick for thick and thin blood smear examinations. Blood smears were Giemsa-stained and malaria parasites were looked for by examination of 1000 leukocytes. To assess the mortality related to malaria we Bulletin of the World Health Organization, 68 (2): 217-222 (1990) C} World Health Organization 1990 217 J.P. Lepers et al. reviewed all the deaths in Manarintsoa since the reappearance of malaria. A similar survey was con- ducted in Ambohimandry, a village located in the same area with 14 000 inhabitants, where the condi- tions of malaria transmission are supposed to be similar, but without any intervention measures. Results A total of 2776 patients were examined, 53.1% males and 46.9% females. The mean age of the population was 18.8 years with the following age distribution: < 2 years (167), 2-9 years (725) 10-14 years (445), 15-19 years (362), 20-29 years (480), 30-39 years (344), and >40 years (253). As no other health centre was available within an 8 km range, and as the consulta- tions and treatment were free, most (66.2%) of the inhabitants of Manarintsoa presented at least once during the 6-month period of the study. Each month we examined an average of 22% of the total popula- tion of Manarintsoa. In addition, a few patients came from the neighbouring villages. Prevalence and density of parasites In all, 1754 (63.2%) individuals presented with P. falciparum parasitaemia and the parasite rate was similar among the various age groups (Table 1). The three other human Plasmodium species were observed at a much lower rate: P. vivax in 369 (13.3%) blood slides (associated with P. falciparum in 107 cases); P. ovale in 28 (1%) blood slides (associated with P. falciparum in 6 cases); and P. malariae in 3 (0.11%). Prevalence of P. falciparum parasitaemia varied during the follow-up period (X2 test, P < 10-6), increas- ing from 44.6% in January to a peak of 75.2% in April (X2 test for trend, P< 10-6), then decreasing to 5 1.5% in June (X2 test for trend, P< 10-6) (Table 2). The geometric mean parasite densities followed a similar pattern as the parasite rates with an initial value of 8682 parasites per p1, a peak value of 11 193 per p1 in April, and decreasing by June to 3810 parasites per p1 (Fig. 1). Clinical findings and parasitological status A total of 1299 (46.8%) patients exhibited spleen enlargment. Splenic rate varied with age (X2 test, P<10-6). It was above 65% in all groups under 15 years, and gradually decreased to 3.9% in subjects older than 40 years (%2 test for trend, P< 10-6)(Table 1). Splenic rates were consistently higher in individuals infected with P. falciparum than in those who were not, but followed a similar evolution during the 6- month period ofthe study (X2 test, P< 10-6) (Table 2). It gradually increased from 27.1% in January to 55.9% Fig. 1. Evolution, by month, of geometric mean parasite density (--- -) and mean axillary temperature (-) In Plasmodlum falclparunmlnfected Individuals seen In the Manarintsoa health centre. 38,6 38.4C 2 38.2 0. E 38.0 x37.8 37.6 37.4 12000 la0 0 *10000 800 0 8000 *w 0. 0 4000 a 2000 January February March April May June Month in May (X2 test for trend, P< 10-6), then remained at this value (X2 test, NS). A total of 1297 (46.8%) subjects seen at the health centre were febrile (axillary temperature > 37.5°C). The mean axillary temperature varied monthly from 37.4 °C to 38.0 °C, with an average value of 37.6 'C. When the population was split according to the parasitological status of the subjects, the mean axillary temperature of individuals present- ing with P. falciparum decreased monthly from 38.4 'C in January to 37.6 'C in March, then remained constant (Fig. 1). The percentage of febrile subjects followed similar variations (X2 for trend, P< 10-6) with an initial decrease from 70.8% in Jan- uary to 42.5% in March (X2 test for trend, P< 10-6) stabilizing thereafter (X2 test for trend, NS) (Table 2). In individuals not infected with P. falciparum the mean axillary temperature and percentage of febrile subjects did not follow a similar temporal evolution but remained constant with an average value of 37.3 'C and 37.7%, respectively, for the whole follow- up period (one way ANOVA and x2 test, NS) (Fig. 1). Malaria-related mortality Our investigations showed that 37 of the 1550 in- habitants of Manarintsoa died during the 2 months before our arrival in March 1987. This represents a 11.9%, monthly mortality rate. From March to December 1987, one of us was present (on average, once or twice a week) and treated all fevers sus- pected to be due to malaria with antimalarial drugs. Monthly mortality rates were 4.5%° in March-July 1987 (malaria transmission season) and 1.55%. in WHO Bulletin OMS. Vol 68 1990.218 Plasmodlum falciparum malaria In the central highland plateaux of Madagascar Table 1: Clinical and parasitological findings In Individuals seen In the Manarintsoa health centre, January-June 1988, by age No. with P. falciparum No. of fevers in P. falciparurm- Age (years) No. of subjects No. with splenomegaly parasitaemia infected subjects 0-2 167 108 (64.7)' 87 (52.1) 65 (74.7) 2-9 725 560 (77.2) 474 (65.4) 263 (55.6) 10-14 445 313 (70.3) 284 (63.8) 156 (54.9) 15-19 362 143 (39.5) 249 (68.8) 125 (50.2) 20-29 480 86 (17.9) 305 (63.5) 160 (52.5) 30-39 344 80 (23.3) 226 (65.7) 105 (46.5) > 40 253 9 (3.6) 129 (51.0) 38 (29.5) Total 2776 1299 (46.8) 1754 (63.2) 912 (52.0) Figures in parentheses are percentages. Table 2: Clinical and parasitological findings In Individuals seen In the Manarlntsoa health centre January-June 1988, by month No. with P. falciparum No. of fevers in P. falciparum- Month No. of subjects No. with splenomegaly parasitaemia infected subjects January 269 73 (27.1)' 120 (44.6) 85 (70.8) February 402 170 (42.3) 244 (60.7) 167 (68.4) March 624 285 (45.7) 402 (64.4) 171 (42.5) April 636 315 (49.5) 478 (75.2) 266 (55.6) May 472 264 (55-9) 318 (67.4) 137 (43.1) June 373 192 (51.5) 192 (51.5) 86 (44.8) Total 2776 1299 (46.8) 1754 (63.2) 912 (52.0) ' Figures in parentheses are percentages. August-December (when malaria is supposed to be not transmitted). Since January 1988, the physician remained in the village on a permanent basis. The monthly mor- tality rate was 1.1%o from January to June 1988, and no deaths were suspected to be due to malaria. By comparison, during the same semester of 1988, the average monthly mortality rate was 3.7%/, in Ambo- himandry, the control village with no medical inter- ventions. Discussion Malaria transmission In this study, we investigated the importance of malaria in an area where P. falciparum malaria reappeared only recently after an absence of some years. This reappearance was detected one year before we initiated our study. Nevertheless, our results indicate that the level of endemicity was already high. Among individuals presenting to our health centre, P. falciparum parasite rates were con- sistently greater than 44% during the 6-month follow-up period. Splenic rates were consistently higher than 60% in the 2-9-year-old children. These results are in agreement with our previous findings where bi-monthly malaria surveys, conducted in the same area from October 1987 to March 1988 and in October 1988, demonstrated both parasite and splenic rates remaining consistently higher than 55% in 5-15-year-old healthy schoolchildren (1). (J.P. Lepers, unpublished results). These data allow us to consider this area of the highland plateaux of Madagascar, where schistosomiasis is absent, as hyperendemic for malaria. The lack of age-related variations in P. fal- ciparum prevalence (Table 1) is a peculiar feature of this area which is probably a consequence of the new malaria transmission. Adults, like young children, did not have the time to develop protective immunity. In this context, the inverse relationship between frequency of spleen enlargement and age may be related to the persistence of some degree of immune memory in adults who lived in a malaria trans- mission area when they were young. Alternatively, in older individuals, spleen tissues may be mechanically less susceptible to enlargement. WHO Bulletin OMS. Vol 68 1990. 219 J.P. Lepers et al. Clinkal Impact of malaria This study was initiated after the first rains, a few weeks before the peak of malaria transmission (D. Fontenille, personal communication). During the first part of the follow-up, from January to April, monthly parasite rates and splenic rates gradually increased. Such augmentation was also observed for the geometric mean parasite densities (GMPDs) in P. falciparum-infected individuals, while their mean axillary temperatures and fever rates were decreasing. During the second part of the study, from April to June, both parasite and splenic rates remained stable in the population, as well as fever rates and mean temperatures in P. falciparum-infected individuals. Moreover, GMPDs decreased. During the whole follow-up period, monthly mean temperatures and percentages of fevers did not vary in individuals non- infected with P. falciparum, indicating that the observed variations were probably related to fal- ciparum malaria. Thus, the clinical impact of P. falciparum malaria varied during the season of malaria transmission. Although endemicity remained high during the whole period of the study (as demon- strated by the persistence of elevated parasite rates) we observed a decrease in the clinical symptoms associated with malaria, followed two months later by a decrease in parasite densities. Such variations could be explained partly by an increase in uncon- trolled antimalarial drug consumption by the population. However, we do not have any indication for such an increase in drug consumption. The other hypothesis suggests that, after a few months of malaria transmission, individuals developed mechan- isms effective in partially controlling the disease- associated pathology. This fact is of particular interest as P. falciparum malaria was reintroduced very recently in Manarintsoa (1). The importance of malaria as a cause of death in the plateaux area is difficult to evaluate. During the six months of our survey, no death appeared to be due to malaria among the inhabitants of Manarant- soa, but all individuals had free access to the medical centre, where clinical malaria attacks received prompt curative treatment. Under these circumstan- ces, the mortality rate was similar to that during the dry and cold season, when malaria is not transmitted. In the control village, this mortality rate was more than three times greater than in Manarantsoa. If we consider that the difference is due only to malaria, then the malaria-related mortality rate can be esti- mated as 2.6%, per month during the transmission season or 18.2%. per year. Assessing the importance of malaria as a cause of morbidity in a rural community where malaria is endemic is very difficult. A high proportion of healthy individuals present with a positive blood smear, and many febrile subjects will be infected with P. falciparum regardless of whether malaria is the cause of their illness or not. Various authors have suggested that this difference could be identified more easily by assessing the blood parasite density. We arbitrarily chose an axillary temperature of 37.5 °C to define important clinical episodes associated with malaria, as this value had been shown to be a useful threshold by various authors (4).' Among P. falciparum-infected patients, 52% were febrile; this proportion was lower (37.7%) in the other indi- viduals (X2 test, P< 10-6). The axillary temperature of individuals did not vary with their parasite count except for highly elevated axillary temperatures (> 39.5 °C) or high blood parasite densities (> 50 000 per p1) (Table 3). This relation was observed in the various age groups. Conversely, when malaria-infec- ted subjects were classified according to their parasite density, a progressive increase in fever rates with the level of parasitaemia (%2 test for trend, P < 106) was observed (Fig. 2). Such an association between tem- perature and malaria parasitaemia has already been reported (4,5); however, one third of non-infected individuals were febrile, and only half of the subjects with a P.falciparum parasite density as high as 10 000 to 20 000 per p1 of blood were presenting with a febrile illness; even patients with higher parasite den- sities were not all febrile. A similar observation was reported in Nigeria (5), and suggests an individual susceptibility to the pyrogenic stimulus of the malaria infection. This relationship between fever and para- site density varied also during the course of the transmission season, since in January 70.8% of all P. falciparum-infected individuals were febrile whatever their blood parasite density, and even parasitaemias lower than 500 per pl were associated with fever in more than 60% of occasions. In various endemic areas, it has been proposed to determine a pyrogenic threshold of parasitaemia, above which fever and clinical symptoms are likely to be a consequence of the P. falciparum infection. Trape et al. (6) found that a parasite count of 5000 per p1 in sick children from the Congo allowed one to separate those in whom malaria was probably the cause of their illness from other children. Similarly, Baudon et al. (7) defined the parasitological threshold for fever as 10 000 parasites per p1 in a rural savannah area of Burkina Faso. In the central highland plateaux of Madagascar, such a threshold appears to be of poor predictive value because of individual and seasonal variability. Delfini, L.F. The relationship between body temperature and malaria parasitaemia in rural forest of Western Nigeria. Unpub lished document WHO/MAL/64.654 (1964). 220 WHO Bulletin OMS. Vol 68 1990. Plasmodlum falciperum malaria In the central highland plateaux of Madagascar Table 3: ParasIte density of Individuals seen in the Manarintsoa health-centre January-June 1988, by age group and by temperature group Axillary temperature (°C) Age group Parasite density' <37 37.1-37.5 37.6-38.5 38.6-39.5 >39.5 0-4 years { 5-15 years >15 years 0 <500 501-5000 5001-50 000 > 50 000 0 < 500 501-5000 5001-50 000 > 50 000 0 < 500 501-5000 5001-50 000 > 50 000 46.5b 11.8 12.5 20.8 8.3 37.3 11.7 17.3 29.0 4.7 42.4 10.5 14.5 26.5 6.2 50.0 8.8 10.3 25.0 5.9 49.2 6.2 15.4 23.8 5.4 54.5 11.0 10.5 17.8 6.3 46.8 8.3 15.6 12.8 16.5 35.4 7.6 16.2 27.3 13.6 31.3 10.9 16.6 34.3 6.8 37.0 5.0 7.0 22.0 29.0 27.8 6.7 13.9 33.9 17.8 25.0 6.3 18.1 38.9 11.8 28.0 2.0 6.0 16.0 48.0 34.2 2.6 10.5 30.3 22.4 23.0 3.4 12.6 32.2 28.7 ' Number of asexual P. falciparum parasites per p1 of blood. b Percentage of individuals in a temperature group presenting with a given parasite density. In the central highland plateaux of Madagascar, malaria is seasonally transmitted. After the beginning of the main transmission season, an increase in parasite rates and parasite densities was observed, while febrile illnesses associated with malaria were decreasing. During the second part of the rainy season parasite densities started to decrease, while parasite rates and fevers associated with malaria remained at the same level. This suggests that the clinical impact of malaria occurred mainly during the former part of the transmission season, although the parasite and splenic rates had not reached their maximum value. Thus, intervention measures aiming at a decrease in malaria-associated morbidity should Fig. 2. The relationship between Plasmodlum fakiperum blood parasite density and episodes of fever (axillary temperature >37.5°C) In Individuals seen In the Manarlntsoa health centre. 100 I 80 60 -, he oh a, 40 20 WHO Bulletin OMS. Vol 68 1990. 0 500 2000 5000 10000 20000 30000 50000 100000 200000 Parasite density per cu. mm. of blood 221 J.P. Lepers et al. mainly be conducted in the beginning of the rainy season. Control campaigns (insecticidal house-spray- ing and mass drug administration) were started in the whole plateaux area in the early part of 1989 and we are following the evolution of malaria endemy. Acknowledgements This work was supported in part by a grant from AUPELF/ UREF. R6sumb R6apparitlon du paludisme A Plasmodlum Ial- ciparum dans les hauts plateaux du centre de Madagascar: estimation de son Impact clilnique dans une communaute rurale. Le paludisme a Plasmodium falciparum, qui avait disparu depuis plus de 20 ans de la region des hauts plateaux centraux de Madagascar, y a recemment ete reintroduit. Afin de preciser les repercussions du paludisme dans la population, nous avons 'tudie le role de P. falciparum dans la pathologie des consultants d'un dispensaire que nous avions installe a Mianarintsoa, au centre de la region de transmission. Durant la deuxieme saison de transmission, de janvier a juin 1988, tous les individus se presentant a ce dispensaire ont fait l'objet d'un examen clinique et d'un pre- levement sanguin pour confection d'une goutte epaisse. Au total, 2776 sujets ont ete etudies dont 1754 (63,2%) etaient parasites par P. falciparum; la proportion de sujets parasites ne variait pas avec I'Age mais avec la saison: la prevalence est pas- s6e de 44,6% en janvier a 75,2% en avril, puis est redescendue a 51,5% en juin. Les densites para- sitaires moyennes ont suivi des variations iden- tiques. Mille deux cent quatre vingt dix neuf sujets (46,8%) etaient porteurs d'une splenomegalie; l'in- dice splenique restait en plateau a 65% jusqu'a 15 ans, puis diminuait progressivement jusqu'a 3,9% chez les individus de plus de 40 ans. Cet indice est passe de 27,1% en janvier a 55,9% en mai, puis est reste constant. II etait plus eleve chez les individus presentant une infestation a P. falciparum que chez les autres, mais suivait les memes varia- tions dans le temps. Au total, 1297 (46,8%) con- sultants du dispensaire etaient febriles. Parmi les sujets infestes par P. falciparum la proportion de sujets febriles est passee de 70,8% en janvier a 42,5% en mars et leur temperature axillaire moyenne de 38,4 0C a 37,6 °C pendant la mime periode; ces deux variables sont ensuite restees constantes. Chez les individus ayant une goutte epaisse negative, la proportion de sujets ftbriles et la temperature moyenne sont restees constam- ment autour de 37,7% et de 37,300 respectivement. L'impact clinique du paludisme a P. falciparum varie donc au cours de la saison de transmission. Bien que le niveau d'endemicite reste 6lev6 (pen- dant toute la duree de cette saison, la fr6quence des fi6vres associ6es au paludisme a diminu6 rapidement, suivie 2 mois plus tard d'une diminu- tion des densit6s parasitaires. Le taux de mortalite par paludisme a egalement et6 6valu6 en exami- nant les registres de mortalite de Mianarintsoa et d'endroits situes en dehors de notre zone d'in- tervention. Ce taux a et6 estim6 a 2,6%, par mois au cours de la saison de transmission, et a 18,2%, par an. L'expression clinique du paludisme, qui s'observe surtout au debut de la saison de trans- mission de la maladie (alors que les indices para- sitaires et spl6niques n'ont pas encore atteint leurs valeurs maximales) est un 6lement important dont il faut tenir compte lors de l'elaboration de strate- gies d'intervention visant a diminuer la morbidit6 liee au paludisme a P. falciparum sur les hauts plateaux de Madagascar. References 1. Lepers, J.P. et al. Reappearance of falciparum malaria in central highland plateaux of Madagascar. Lancet, 1: 586 (1988). 2. Ramanamirlja, J.A. La lutte antipaludique a Madagas- car. In: S6minaire Inter-lies Oc6an Indien, Souris- Chaude, lie de R6union, 15-20 June 1981. 3. Fontenille, D. & Rakotoarivony, I. Reappearance of Anopheles funestus as a malaria vector in the Antan- anarivo region, Madagascar. Transactions of the Royal Society of Tropical Medicine and Hygiene, 82: 644-645 (1988). 4. Greenwood, B.M. et al. Mortality and morbidity from malaria among children in a rural area of The Gambia, West Africa. Transactions of the Royal Society of Tropical Medicine and Hygiene, 81: 478-486 (1987). 5. Hendrickse, R.G. et al. Malaria in early childhood. An investigation of five hundred seriously ill children in whom a "clinical" diagnosis of malaria was made on admission to the children's emergency room at Univer- sity College Hospital, Ibadan. Annals of tropical medicine and parasitology, 65: 1-20 (1971). 6. Trape, J.F. et al. Criteria for diagnosing clinical malaria among semi-immune population exposed to intense and perennial transmission. Transactions of the Royal Society of Tropical Medicine and Hygiene, 79: 435-442 (1985). 7. Baudon, D. et al. Morbidite palustre en milieu rural au Burkina Faso. Etude de 526 acces f6briles. M6decine d'Afrique noire, 33: 767-776 (1986). 222 WHO Bulletin OMS. Vol 68 1990.