Inland Valley Rice Production Systems and Malaria Infection and Disease in The Savannah of Co Te D'ivoire

Tropical Medicine and International Health
volume 8 no 5 pp 449458 may 2003
Inland valley rice production systems and malaria infection

and disease in the savannah of Cote dIvoire
M.-C. Henry1, C. Rogier2, I. Nzeyimana1, S. B. Assi1, J. Dossou-Yovo1, M. Audibert3, J. Mathonnat3, A. Keundjian2,
E. Akodo4, T. Teuscher4 and P. Carnevale1
1
2
3
4
Institut P. Richet, Bouake, Cote dIvoire

Institut de Medecine Tropicale du Service de Sante des Armees, Marseille, France
CERDI/CNRS, Clermont-Ferrand, France
WARDA, Bouake, Cote dIvoire
Summary
In sub-Saharan Africa, lowlands developed for rice cultivation favour the development of Anopheles
gambiae s. l. populations. However, the epidemiological impact is not clearly determined. The
importance of malaria was compared in terms of prevalence and parasite density of infections as well
as in terms of disease incidence between three agroecosystems: (i) uncultivated lowlands, R0, (ii)
lowlands with one annual rice cultivation in the rainy season, R1 and (iii) developed lowlands with two
annual rice cultivation cycles, R2. We clinically monitored 2000 people of all age groups, selected
randomly in each agroecosystem, for 40 days (in eight periods of five consecutive days scheduled every
6 weeks for 1 year). During each survey, a systematic blood sample was taken from every sick and
asymptomatic person. The three agroecosystems presented a high endemic situation with a malaria
transmission rate of 139158 infective bites per person per year. The age-standardized annual malaria
incidence reached 0.9 malaria episodes per person in R0, 0.6 in R1 and 0.8 in R2. Children from 0 to
9-year-old in R0 and R2 had two malarial attacks annually, but this was less in R1 (1.4 malaria episodes
per child per year). Malaria incidence varied with season and agroecosystem. In parallel with
transmission, a high malaria risk occurs temporarily at the beginning of the dry season in R2, but
not in R0 and R1. Development of areas for rice cultivation does not modify the annual incidence of
malarial attacks despite their seasonal influence on malaria risk. However, the lower malaria morbidity
rate in R1 could be explained by socio-economic and cultural factors.
keywords malaria morbidity, rice cultivation systems, savannah, Cote dIvoire
Introduction
In sub-Saharan Africa, water resource development
projects affect many waterborne diseases. In the case
of malaria, irrigated rice cultivation favours the
multiplication of anopheline vectors. However, the
epidemiological impact of rice cultivation varies
according to the local malaria situation (Carnevale et al.
1999). It can be associated with an increase in
malaria transmission and morbidity, as in Burundi
(Coosemans 1985) and the uplands in Madagascar
(Laventure et al. 1996). Conversely, irrigated rice
cultivation does not seem to affect malaria transmission
or its incidence in northern Cameroon (Audibert et al.
1990), in the Senegal River valley (Faye et al. 1993,
1995), in the Kou valley in Burkina Faso (Boudin
et al. 1992) and in the Gambia River valley (Lindsay
et al. 1991).
2003 Blackwell Publishing Ltd
Most of these studies limit themselves in explaining

aspects of relations between irrigated rice cultivation,
transmission level, Plasmodium infections and malaria
morbidity at the local level. However, to predict the
consequences of rice cultivation development and to
control its possible negative aspects, it is necessary to
improve understanding of the interrelations between public
health, the environment and irrigated zones (Gioda 1992).
Thus, an interdisciplinary study of relationships between
lowland cultivation systems and malaria was conducted at
regional level in three West African important settings:
Sahel, savannah and forest. This study was carried out in a
savannah region of northern Cote dIvoire. Its objective
was to compare malaria pressure in three farming systems
in terms of prevalence and parasite density of infections,
and also in terms of clinical malaria incidence. Exposure to
transmission by Anopheles was the object of another study
(Dossou-Yovo et al., unpublished data).
449
M.-C. Henry et al. Inland valley rice production and malaria
Materials and methods

Study zone
The study was conducted in the savannah region of northern
Cote dIvoire, where population density was 2040
inhabitants per km2. All villages were classified according
to the farming systems in their surrounding valleys within a
two km radius: no (rice) cultivation (R0); no water control,
suitable for one cycle of rice cropping during the rainy
season (R1); partial or full water control that permits two
cycles of rice cropping per year (R2). The three categories
of farming practice are referred to below as agroecosystems. The 8 study villages per farming system were
randomly selected among villages pooled by agroecosystem. All R1 and R2 villages were situated in the
Departement of Korhogo where all lowlands were cultivated. The 8 R0 villages were situated in the Departement
of Katiola, where lowlands were not at all farmed. The R2
villages were Gbahouakaha, Kohotieri, Koumbolikaha,
Lamekaha, Nambekaha, Nombolo, Nongotchenekaha and
Zemongokaha; for R1, the villages were Binguebougou,
Fapaha, Kombolokoura, Kaforo, Karakpo, Kassoumbarga,
Katiorkpo, Tioro and for R0, Angolokaha, Doussoulokaha, Folofonkaha, Kabolo, Ounadiekaha, Petionara,
Serigbokaha, Timorokaha. There was a health centre in
Petionara (R0), Tioro (R1) and Gbahouakaha (R2) and a
village health post in Kohotieri (R2).
The total population required for a Poisson regression
analysis was estimated using Egret Siz (ver. 1, 1993) to
be 80 000 person-days to detect a relative risk below 0.5
with a power of 80%, a significance level of 5%, a
maximum incidence of 0.64 malarial attacks per personyear in the reference ecosystem (Trape et al. 1994) and
including 30% of participants lost to follow-up. This
total was attained by selecting 250 persons in each
village and monitoring them clinically for 40 days
distributed over eight periods of five consecutive days
scheduled every 6 weeks in the year.
People were selected from randomly sampled compounds by dividing the village in districts. Of these, six
districts were randomly chosen irrespective of the chiefs
district. Then, from the centre of each of these six districts
a cardinal direction was randomly selected. In this direction, the first six encountered compounds were selected.
If there were less than six compounds in this direction,
another direction was randomly selected. Each family head
and each person included in the study or their legal
guardian gave informed consent. Ethical approval for the
project was given by the Ivorian Ministry of Public Health.
During the monitoring periods, patients of villages participating or not in the study were treated free of charge by
the medical team.
450
Data collection
The active case detection (ACD) surveys for malaria
episodes started on 18 March 1997, 26 April, 10 June,
22 July, 2 September, 18 October, 25 November and
20 January 1998. During these monitoring periods, a nurse
assisted by two health workers from the village trained for
the purpose of the study, visited all households covered by
the study every day. A doctor provided permanent supervision of the teams. The presence, absence and health
condition of each included person were recorded daily by
the assistant nurse on a sheet meant for each household.
The nurse examined any detected sick person at home
and registered clinical observations on an ad-hoc sheet.
A blood sample was taken systematically and patients were
treated according to the clinical diagnosis made by the
nurse. When malaria was suspected, the patient was
treated with chloroquine at the dose of 25 mg/kg body
weight for 3 days in conformity with the National Program
for Malaria Control. ACD was scheduled 15 days after
the malaria transmission assessment carried out in 12 of
the 24 villages clinically monitored, in order to correlate
malaria infection and disease rates with transmission rates.
The cross-sectional surveys (CSS) were held during each of
the eight monitoring periods, a blood sample was taken
systematically from each asymptomatic person in the
study. This was carried out on the second day of each
period to make sure that a participant classified as
asymptomatic was free of illness during the days before
and after the blood sample was taken.
Laboratory procedures
Thick smears were made from blood samples and stained
with Giemsa in the field and examined using a microscope
(ocular 10, lens 100) at Institut Pierre Richet at Bouake.
Plasmodium species were identified and asexual forms of
each species counted on 200 leucocytes. The parasite
density was calculated by assuming an average concentration of 8000 leucocytes/ll of blood. The same experienced
technician, under the supervision of a parasitologist,
examined the smears from a given village. The six
technicians also compared the same set of blood samples.
Their rate of parasite detection and parasite density
estimates did not differ significantly. A randomly selected
10% sample of the thick smears was double-read for
quality control.
A urine study of antimalarial drugs was conducted in
January 1998 on more than 20 adults and children over
2 years old randomly selected in each village. The urine
collected was tested at the Institut de Medecine Tropicale
du Service de Sante des Armees at Marseille through a
high-performance liquid chromatography technique

(Brown et al. 1982), modified for simultaneous isolation
and measure of chloroquine and its metabolites, amodiaquine, sulfamide and quinine.
Analytical strategy
Demographic, clinical, parasitological and attendance data
were double-entered independently in an Access database
(ver. 7, 1995). Data were analysed using EpiInfo (ver. 6,
1995), STATA statistical package (StataCorp 2001) and
Egret (ver. 2, 1999) software programs.
The association between the parasite load and the
occurrence of clinical episodes was tested using a randomeffect logistic regression model for each agroecosystem
and taking clinical status (pathological episode vs.
asymptomatic state) as dependent variable, and parasite
density, age and season as independent variables. In this
type of model, a random intercept variable is allowed to
vary with subjects and this random subject-specific intercept allows taking into account the interdependency of
observations made on the same person. The independent
variables and their interaction terms were tested and kept
in the model when their effects were significant (likelihood
ratio statistic, P < 0.05). For each pathological period, the
probability that it was caused by malaria was estimated
through the attributable fraction calculated from the odds
ratios associated with the estimated parasite density in
each logistic model (Armstrong Schellenberg et al. 1994).
Pathological episodes considered were those characterized
by a high axillary temperature (37.5 C), a body hot to
the touch, sweat, shiver, headache, nausea or vomiting
(Rogier et al. 1999) or by a history of fever during the
48 h preceding the first day of ACD; or in cases of infants,
anorexia or any pathological condition described by the
mother (Smith et al. 1995). For individuals and given
periods, the number of malarial attacks was estimated by
the sum of probabilities of pathological episodes that were
caused by malaria, depending on the parasite load.
Malaria incidence density was calculated through the ratio
of pathological episodes attributable to malaria and
villagers person-days present during monitoring periods.
The clinical malaria incidence in a standardized population was calculated, for each agroecosystem, by multiplying age specific incidence densities estimated in each
category with the number of subjects belonging to age
groups of a virtual population of 1000 persons with an
age-distribution identical to the age-distribution of the
populations of the three studied agroecosystems as a
whole.
Parasitological data were analysed separately in terms of
(1) prevalence of Plasmodium falciparum, P. malariae and
P. ovale asexual blood forms, (2) density of P. falciparum

asexual blood forms in parasite-positive thick smears, and
(3) prevalence of P. falciparum gametocytes.
Only one blood sample per person per monitoring period
was considered for the analysis. When a pathological
condition was detected, it was the blood sample taken
during the clinical episode that was retained. When several
blood samples from an asymptomatic period were available, one was randomly selected for analysis.
We used a generalized estimating equation (GEE)
approach for statistical analysis of repeated measures
(Zeger & Liang 1986), which can be used with normal
distributions and discrete data. We used an exchangeable
correlation structure in which the correlation between
observations made on the same person at different times is
assumed to be the same. The differences were tested by the
Wald test and 95% confidence intervals were calculated.
The prevalence of asymptomatic malaria infections was
analysed as a binomial response. The positive asymptomatic parasite density was log transformed and analysed
with a link function for a normally distributed response.
The GEE approach allows some departure from the
hypothesis about the distribution of the dependent variable
and gives robust estimates of regression coefficients taking
into account the interdependence of observations made
within the same person. Comparisons between prevalences
and between parasite densities were performed by
chi-square test and variance analysis. For parasite densities,
interactions between age and agroecosystem and between
season (dry season from November to April, and rainy
season from May to October) and agroecosystem were
tested using a multiple linear regression model. Clinical
malaria incidence densities observed for the different
categories, in the different age classes (09 and 10 years)
and in the different seasons were compared using the
likelihood ratio statistic in a Poisson regression model,
with the estimated number of malarial attacks as dependent variable and the cumulative number of monitoring days
as exposure variable. The village incidence mean rates were
compared between each pair of villages categories using
MannWhitney U-test. Statistical tests were considered as
significant when P < 0.05.
Results
Population description
From 18 March 1997 to 28 February 1998, 6184 people in
24 villages (2054 in eight villages of farming system R0,
2055 in eight villages of R1, and 2075 in eight villages
of R2) were clinically and parasitologically monitored.
Children born during the study were not included. The
451
Table 1 Distribution of population samples according to farming systems (R0, no rice cultivation; R1, single rice cropping;
R2, double rice cropping)
Age group
(year)
01
24
59
1019
2039
40
Total
R0
n (%)
R1
n (%)
R2
n (%)
164
246
411
392
411
221
144
185
370
514
411
431
164
164
308
473
411
555
(8)
(12)
(20)
(19)
(20)
(11)
2054 (100)
(7)
(9)
(18)
(25)
(20)
(21)
2055 (100)
(8)
(8)
(15)
(23)
(20)
(27)
2075 (100)
distribution of population samples by age is shown in

Table 1. The sample in R0 was younger than in R1 and R2.
The sex ratio was unbalanced for adults, particularly in R1
and R2. The female/male ratio was 1.2 in R0, and 1.7 in
R1 and R2. The population of the three village groups
belonged mostly to the Senufo ethnic group.
During the eight monitoring periods, 38 139 blood
samples were taken. Their distribution according to subjects clinical status and agroecosystems is shown in
Figure 1. Nine fever episodes (from the three farming
systems) without thick smears were excluded from the
study. Each fever syndrome corresponds to one illness
episode per person and per survey.
Irrespective of the farming system, population participation in the study was high. In fact, some 79% of
R1 and R2, and 77% of R0 populations took part in
38 139 blood samples
Exclusion of 705 blood samples obtained twice from the same person
during the same monitoring period
37 434 blood samples
36 387 in asymptomatic
subjects observed during the
cross sectional surveys
11 951 in
R0
12 306 in
R1
12 130 in
R2
1047 in febrile subjects found

by the active case detection
method
355 in
R0
344 in
R1
348 in
R2
Figure 1 Distribution of blood samples according to clinical

status and farming systems (R0, no rice cultivation; R1, single
rice cropping; R2, double rice cropping).
452
at least seven of the eight monitoring periods. An average

of six blood samples per person was taken from 75% of
R1 and R2, and 70% of R0 populations. Each person
of R0, R1 and R2 was visited at home on 27 days (9) on
average over the 40 active detection days planned in the
protocol.
Parasitological indexes of asymptomatic subjects
observed during CSS
Table 2 shows a considerable predominance of P. falciparum over P. malariae and P. ovale in the asymptomatic
infections in all the agroecosystems. The plasmodia distribution was comparable across the three agroecosystems:
P. falciparum, 99.3%; P. malariae, 5.17.9%; and
P. ovale, 0.51.5% according to agroecosystems. Annual
average prevalence of P. malariae was <5%, but at least
five times more important than P. ovale infections.
Differences between the three environments were significant in age group 09 years, where the highest percentages of these plasmodia were observed in R0 (P. malariae,
11.4% and P. ovale, 2.0%), and the lowest in R2
(P. malariae, 2.0% and P. ovale, 0.7%). Plasmodium
malariae and P. ovale prevalence did not differ significantly
with seasons.
Table 3 shows the mean rates of P. falciparum trophozoites and gametocytes and the mean parasite density in
positive thick films in asymptomatic subjects according to
age groups (09 and >9 years), seasons and agroecosystems. The mean annual prevalence of P. falciparum
infections ranged from 65% in R2 to 72% in R1 and R0.
Everywhere, more than 80% of children from 2 to 9 years
old were parasite-positive, with the highest percentage in
the age group 24 years. About 50% of adults aged 40 and
above were still infected. The mean annual gametocyte
rates reached about 7% everywhere. As observed in other
endemic areas, the three parasite indexes (trophozoite and
gametocyte rates and parasite density) were higher in
children than adults. Multivariate analysis showed that
P. falciparum asexual stages prevalence and parasite
densities decreased with age according to agroecosystems
(P < 0.001) (Table 4). The three parasite indexes were
higher in the rainy season than in the dry season in the
three agroecosystems. Seasonal variations in P. falciparum
asexual stages prevalence and parasite density differed
between agroecosystems (P < 0.001) (Table 5). This
increase in parasite densities during the rainy season was
less pronounced or absent among older children and
adults. With the onset of rains, parasitaemia increased
earlier in R0 than in R1 and R2 and with the onset of
the dry season, parasitaemia dropped faster in R0 and R1
than in R2.
Table 2 Plasmodium-positive infections in

asymptomatic and febrile subjects according to farming systems (R0, no rice cultivation; R1, single rice cropping; R2, double
rice cropping)
Asymptomatic subjects
(cross-sectional surveys)
Febrile subjects
(active case detection)
R0
R1
R2
R0
R1
R2
Thick blood films (n)

P. falciparum
P. malariae
P. ovale
P. falciparum + P. malariae
P. falciparum + P. ovale
P. malariae + P. ovale
P. falciparum +
P. malariae + P. ovale
11 951
66.3
0.4
0.1
5.2
0.8
0
0.2
12 136
68.1
0.5
0
3.7
0.4
0
0.1
12 130
61.3
0.4
0.1
3.2
0.2
0
0.1
355
76.9
0.3
0
6.5
0.6
0
1.1
344
71.2
0.3
0.3
4.1
1.5
0
0
348
70.1
0.3
0
5.2
0.6
0
0.3
All Plasmodium species (%)
73.0
72.8
65.3
85.4
77.3
76.4
Table 3 Plasmodium falciparum trophozoite and gametocyte rates and geometric mean parasite density in parasite-positive slides in
asymptomatic persons (observed during cross-sectional surveys) according to seasons, age groups and farming systems (R0, no rice
cultivation; R1, single rice cropping; R2, double rice cropping)
Prevalence
Parasite density (trophozoite/ll)
09 years
R0
Dry season
Wet season
Total
R1
Dry season
Wet season
Total
R2
Dry season
Wet season
Total
10 years
Total
Tr G
% % n
09 years
Tr G
% % n
Mean
(95% CI)
10 years
Tr G
%* % n
1810
3040
4850
83
90
87
11
14
13
2629 56 3
4472 66 5
7101 62 4
4439 67 7
7512 76 9
11951 72 8
1503 402 (370657) 1465 102 (96108) 2968 204 (193216)

2725 557 (523727) 2969 112 (107117) 5694 242 (231253)
4228 497 (472509) 4434 108 (104112) 8662 228 (220236)
1617
2636
4253
80
86
84
12
12
12
2989 61 3
5064 69 5
8053 66 4
4606 68 6
7700 75 7
12306 72 7
1299 402 (369698) 1827 113 (107119) 3126 191 (182202)

2263 435 (404635) 3516 117 (112122) 5779 196 (188204)
3562 423 (400475) 5343 115 (112119) 8905 194 (188201)
1382
2308
3690
75
81
79
10
11
11
3126 53 4
5314 62 5
8440 59 4
4508 60 6
7622 68 7
12130 65 6
1035 400 (362803) 1657 123 (116131) 2692 194 (183206)

1865 461 (426738) 3299 123 (118128) 5164 198 (190207)
2900 439 (412550) 4956 123 (119127) 7856 197 (190204)
Mean
(95% CI)
Total
n
Mean
(95% CI)
Percentage of * Trophozites and Gametocytes.
Clinical malaria incidence observed by ACD method

We considered only one fever episode per patient and
survey. If two or more fever episodes were observed during
two or more different monitoring periods in the same
patient, each fever episode per survey was taken in
account. Figure 2 shows the characteristics of the 1047
patients detected in the three agroecosystems. Among these
patients, there were 212 (nearly 20%) parasite-negative
subjects, 831 subjects with P. falciparum single or mixed
infection and four subjects with P. malariae or P. ovale
single infection. The distribution of Plasmodium species in
the patients according to the agroecosystems is reported in
Table 2. The three P. malariae single infections showed a
density <200 parasites/ll, and the only P. ovale single

infection showed a density of 4880 parasites/ll. Patients
had more parasites than healthy people everywhere
(Figure 3). Table 6 shows annual incidence density of fever
and malarial attack according to age and agroecosystem.
Mean annual frequency of fever was two episodes per
person per year in the three agroecosystems, but malaria
represented 43% of fevers in R0, 27% in R1, and 33% in
R2. The age-standardized annual malaria incidence
reached 0.9 (0.81.0) malarial attacks per person in R0,
0.6 (0.50.7) in R1, and 0.8 (0.71.0) in R2. In the three
agroecosystems, the burden of malaria was greatest among
children under 2 years of age. Annually, they had an
average of three malarial attacks in R1 and R2, and nearly
453
Table 4 Odds ratio for patent parasitaemia by age groups and

farming systems (R0, no rice cultivation; R1, single rice cropping;
R2, double rice cropping)
Age
group
(years)
217 R1
222 R2
677 patients with axillary

T 37 C during ACD
Odds ratio (95% CI)

R0
238 R0
R1
117 R0
127 R11
126 R2
370 patients with a fever history

during 48 h preceding ACD
R2
1047 detected patients
age group
1
1.83 (1.372.44)
1.25 (0.981.60)
0.52 (0.420.66)
0.30 (0.240.38)
0.23 (0.190.29)
farming system
1
1
1
1.97
1.66
0.96
0.51
0.40
(1.512.58)
(1.332.09)
(0.781.18)
(0.420.63)
(0.330.50)
1
2.16
1.74
1.13
0.56
0.43
(1.672.80)
(1.402.16)
(0.921.37)
(0.460.67)
(0.360.52)
0.76 (0.660.87) 0.54 (0.470.62)

1.18 (1.081.27) 0.85 (0.790.92)
Table 5 Wet/dry season odds ratio for patent parasitaemia by age

groups and farming systems (R0, no rice cultivation; R1, single rice
cropping; R2, double rice cropping)
Odds ratio (95% CI)
R0
R1
R2
All ages
1.66 (1.531.81) 1.50 (1.401.62) 1.27 (1.181.37)
<10 years 1.28 (1.191.36) 1.09 (1.021.17) 1.08 (1.011.16)
10 years 1.13 (1.061.21) 0.97 (0.911.04) 0.96 (0.891.03)
five in R0. Between age 2 and 9 years, they had one to two
malarial attacks per year. From 10 years on, villagers had
<1 malarial attack every 2 years. The weak incidence of
malarial attacks in the age group of 10 and above does not
allow an accurate analysis of risk factor effects on
morbidity. In children aged up to 9 years comparison of
malaria annual risk, which was 1.9 malarial attacks per
person per year in R0, 1.4 in R1 and 2.2 in R2, showed
that it was similar in R0 and R2 (P 0.5). It was lower in
R1 than in R0 and R2 (P 0.001). In an additional
analysis taking the individual villages as units of analysis,
the annual incidence densities in children showed little
variation among villages in R0 (SD 0.43), but was larger
for R1 and R2 (SD 0.66 and 0.79, respectively). Annual
incidence densities were significantly lower in R1 villages
than in R0 (P 0.04) or R2 (P 0.03), but no significant
difference was observed between R0 and R2 villages
(MannWhitney U-test).
The incidence rate was, respectively, during the rainy
season, 2.6 malarial attacks per child per year in R0, 1.9 in
R1 and 2.5 in R2 and during the dry season 0.8 malarial
attacks per child per year in R0, 0.6 in R1 and 1.6 in R2.
In all the systems, malaria morbidity was higher in the
454
835 Plasmodium positive infection
766 single infections
212 Plasmodium negative

infection
69 mixed infections
Figure 2 Characteristics of patients found by the active case

detection method in the three farming systems (R0, no rice cultivation; R1, single rice cropping; R2, double rice cropping).
95% CI10 Logarithm of parasite density
Effect of
<2
24
59
1019
2039
40
Effect of
<10
10
3.8
3.6
3.4
3.2
3.0
2.8
2.6
2.4
2.2
2.0
R0 R1 R2
R0 R1 R2
Asympt.
Febrile
<10 years old
R0 R1 R2
R0 R1 R2
Asympt.
Febrile
10 years old
Figure 3 Annual geometric mean density (95% CI) in parasitepositive asymptomatic (observed during cross-sectional surveys)
and febrile (found by active case detection method) persons
according to age groups and farming sytems (R0, no rice cultivation; R1, single rice cropping; R2, double rice cropping).
rainy season than in the dry season, with a very significant

risk in R0 and R1 (P < 0.001), but just significant in R2
(P 0.05). Figure 4 clearly shows that malarial attacks
persisted in high numbers in R2 at the beginning of the dry
season in December, before decreasing. This situation was
quite different in R0 and R1, where malaria incidence fell
as soon as the rainy season ended.
A cut-off value above which parasite density associated
with fever defines a symptomatic malaria case was calculated through the attributable fraction method (Smith
et al. 1994). It was 2500, 1000 and 100 P. falciparum
Table 6 Annual incidence density of fever and malarial attack (found by the active case detection method) according to age groups and
farming systems (R0, no rice cultivation; R1, single rice cropping; R2, double rice cropping)
Observations
R0
R1
R2
Age group
(years)
Person-day
01
24
59
1019
2039
40
Total
01
24
59
1019
2039
40
Total
01
24
59
1019
2039
40
Total
1116
1687
2870
2667
2968
2997
14 305
1065
1356
2516
3479
2868
3257
14 541
970
1132
2113
3026
2703
4028
13 972
4688
7157
11 407
10 009
11 248
11 754
56 263
4549
5640
9697
12 867
10 746
13 018
56 517
4172
4867
8064
11 055
9964
15 248
53 370
Fever
Malaria fever in the standard

population of 1000 persons
Malaria fever
Incidence
per personyear
Attributable
fraction*
Incidence
per personyear
Population
(%)
Incidence
per year
(95% CI)
112
69
54
38
49
33
355
97
52
66
45
38
46
344
73
71
65
47
41
51
348
8.7
3.5
1.7
1.4
1.6
1
2.3
7.8
3.4
2.5
1.3
1.3
1.3
2.2
6.4
5.3
2.9
1.6
1.5
1.2
2.4
62.2
45
15.7
15.7
10.6
6.6
155.8
36.8
17.7
21.8
6
4.3
3
89.6
34.6
40.6
26
10.4
3.6
2.2
117.4
4.8
2.3
0.5
0.6
0.3
0.2
1
3
1.1
0.8
0.2
0.1
0.1
0.6
3
3
1.2
0.3
0.1
0.1
0.8
7.4
9.8
17.5
21.4
19.9
24
100
7.4
9.8
17.5
21.4
19.9
24
100
7.4
9.8
17.5
21.4
19.9
24
100
356
224
88
123
69
49
909
217
118
144
37
29
20
559
223
297
210
74
26
12
842
(129766)
(70467)
(18225)
(30283)
(12204)
(5172)
(7861045)
(46564)
(11358)
(41318)
(1135)
(0133)
(0110)
(464668)
(50615)
(81634)
(75445)
(12208)
(0143)
(094)
(720977)
* Fever fraction attributable to malaria was estimated by logistic regression model.

Ratio of the total study population in the age group to the study population as a whole.
800
600
400
200
0
10
8
6
4
2
0
Malaria
attacks/1000
child-day
R1
1000
10
8
6
4
2
0
Infective
bites/1000
man-night
R2
1000
800
600
400
200
0
D- J- F M A M J J A S O N D J- F
96 97
98
Months of survey
Malaria
attacks/1000
child-day
10
8
6
4
2
0
Malaria
attacks/1000
child-day
Infective
bites/1000
man-night
Infective
bites/1000
man-night
R0
1000
800
600
400
200
0
Figure 4 Incidence density of malaria fevers (point) in children

09 years old (found by active case detection method) as a consequence of the entomological inoculation rate (bar) in the three
farming systems (R0, no rice cultivation; R1, single rice cropping;
R2, double rice cropping).
trophozoites per microlitre, respectively, for people aged

04, 59, and 10 years and above. The sensitivity of
malaria case definition varied between 0.7 and 0.9 whereas
the specificity was equal to 0.8, in each of the three
agroecosystems.
Antimalarial drug consumption was very low in the
three agroecosystems. No traces of sulphamide, amodiaquine and quinine were found in urine. Chloroquine
and/or its metabolites were detected in 1.7% of R1
(2/172) and R0 (4/236) samples, and 4.7% of R2 (7/149)
samples. This higher rate in R2 was due to the fact that
20% of the samples (5/24) of Kohotieri village were
positive.
Discussion
The parasitological indices showed that in the three
agroecosystems malaria is highly endemic. Even if the
mean annual prevalence was 70% among infants, parasitaemia slowly reduced between 2- and 5-year olds, and at
least one of two adults was parasite-positive. This level of
455
high endemicity is compatible with the perennial transmission with pronounced seasonal peaks observed in the
three agroecosystems. The mean annual inoculation rate
was similar in all of the three agroecosystems, with,
respectively, 158, 139 and 155 infective bites per person
per year, in R0, R1 and R2 (Dossou-Yovo et al.,
unpublished data).
The clinical study showed that, irrespective of the
system, inhabitants experience on average two fever
attacks per year. However, the age-standardized annual
malaria incidence was 0.9 malarial attacks per person in
R0, 0.6 in R1 and 0.8 in R2. It is mainly P. falciparum that
causes malaria fever episodes. If it is assumed that
pyrogenic parasitic densities are similar for all plasmodia
species among people of a given age in highly endemic
conditions (Trape et al. 1994), only one malarial attack
with a P. ovale single high parasitic infection was detected.
The three subjects with a P. malariae single infection had a
very low parasite density, apparently not compatible with
diagnosis of malarial attack.
To quantify malaria morbidity, we used the method of
estimation of fever fractions attributable to malaria. This
method can be applied even if distribution of parasite
infection on asymptomatic subjects includes a number of
highly infected individuals (Smith et al. 1994). In our
study, 8% of asymptomatic children had an infection
above 5000 P. falciparum trophozoites per microlitre,
whereas this was the case for 37% of the sick ones. Eight
per cent of healthy subjects aged 10 years or older had
more than 500 parasites/ll whereas this was the case for
16% of sick subjects of the same age group.
These high asymptomatic parasitaemias observed are
attributable to recent malaria cases which occurred before
the monitoring period. Malaria fevers are short especially
in adults (Rogier et al. 1999), and they may not all have
been reported. Finally, self-medication is common. Indeed
in the Korhogo region, traditional medicine, most often
plants, is systematically used in the first instance to treat
febrile people (De Plaen et al., personal communication).
This may have been effective against fever but not against
parasites. Smith et al. (1994) observed that parasite
distribution in non-feverish subjects is different, depending
on place of detection in the village or in the health centre.
Also they attributed high asymptomatic parasitaemia
observed in the village to recent cases of malaria.
The three agroecosystems present a distinct epidemiological behaviour concerning immunity acquisition in
children from 0 to 9 years old. In this age group, annual
parasite prevalence and parasitaemia in asymptomatic
subjects were slightly higher in R0 than in R1 and R2. The
annual incidence of clinical malaria reached on average
two malarial attacks per child in R0 and R2, whereas
456
children in R1 suffered less from malaria, with 1.4 episodes

per year. However, reduction of malaria risk between 04
and 59 age groups seemed more important in R0 (6.6
factor) than in R1 and R2 (2.5 factor). Observations might
suggest that, between 0 and 4 years, children in R0
encounter a higher malaria risk than those in R1 and R2.
The values of risk reduction fall into the range observed
between the same age groups in other studies (Snow et al.
1999). Plasmodium malariae and P. ovale were more
frequently found in R0 than in R1 and R2, especially in the
04 age group. However, their prevalence did not differ
with the season. As Anopheles funestus was more frequently captured in R0 than in R1 and R2, P. malariae
could be preferably transmitted by this vector (Boudin
et al. 1991). However, our observations do not allow the
verification of this hypothesis.
The three agroecosystems also present a distinct epidemiological seasonal pattern. Generally, in all three,
asymptomatic infections were stronger and more frequent
during the rainy season than in the dry season, especially in
children up to 9 years old. There were also more malarial
attacks during the rainy season. However, in R1 the
increase of parasite load in sick children at the beginning of
the rainy season was lower than in R0 and R2 and it
remained low during the rest of the rainy season. Malarial
attacks were also less frequent, although the transmission
rate was the same as in the two other agroecosystems.
Neither chloroquine consumption, which was very low, nor
protection with mosquito nets, which were used only by
about 4% of the population in the three agroecosystems,
can explain these findings. The use of fumigation, mosquito
coils or aerosol insecticides was more frequently reported in
R1 than in R2 and R0 (M. Audibert, personal communication). According to Snow et al. (1998), their use does not
protect against mild but rather against severe malaria
disease. Moreover the R1 families were richer than those of
R2 and most of R0 (Audibert et al. 2003). Further investigations are needed to determine how the family richness
might have an effect on the reduced incidence in R1.
In R2, with irrigated rice cultivation, malarial attacks
persisted for a short period in the beginning of the dry
season before decreasing. Conversely, in R0 and R1, they
reduced abruptly at the end of the rains. The temporarily
maintained level of malaria morbidity, which closely
coincided with the prolongation of transmission period in
R2, was almost non-existent in R0 and R1. The prolongation of intense hostparasite contact in R2 resulted in a
smaller increase in malaria risk in R2 than in R0 and R1 in
the rainy season.
Our findings are somewhat different from the observations made in other irrigated rice cultivation zones in West
Africa. In an irrigated zone of The Gambia (Lindsay et al.
1991) malaria fever incidence, highest during the rainy

season, persisted during the first week of the dry season
before decreasing. In Burkina Faso in dry savannah
(Boudin et al. 1992), plasmodia infection incidence became
maximal at the beginning of the dry season and then
decreased in villages with irrigated rice cultivation,
whereas incidence was highest during the rainy season in
the villages without rice cultivation. In Mali (Sissoko et al.,
unpublished data), malaria fever incidence was fairly
constant over the seasons at a low level in an irrigated zone
whereas incidence was low during the dry season and high
at the end of the rainy season in a non-irrigated zone.
It is clear that in northern Cote dIvoire savannah, the
double cropping rice system extends malaria risk during a
short period, in the beginning of the dry season. It does not
modify annual malaria incidence in the developed lowlands
zone, which is comparable with the uncultivated lowlands
zone, with the same annual transmission rate. These results
confirm that in a stable malaria zone, rice cultivation does
not significantly affect malaria pressure, contrary to what
may occur in an unstable malaria zone (Carnevale et al.
1999).
In conclusion, in the savannah region of northern Cote
dIvoire, lowland rice cultivation does not significantly
influence malaria risk, but socio-economic and cultural
factors might reduce malaria pressure. Because of the
large scale study and the methodology focused on
malaria, estimates in the context of farming systems rather
than in single villages, the data presented contribute to a
better knowledge of the malaria risk related to rice
cultivation.
Acknowledgements
This study was undertaken within the framework of the
WARDA/WHO-PEEM/IDRC/DANIDA/Government of
Norway Health Research Consortium on the Association
between irrigated Rice Ecosystems and Vector-Borne
Diseases in West Africa. The Consortium received technical assistance from the International Development
Research Centre (IDRC), Ottawa, Canada and financial
support from the International Development Research
Centre (IDRC), Ottawa, Canada, the Danish International
Development Agency (DANIDA) and the Royal Government of Norway.
We thank the Korhogo District Health Officer,
Dr Richard Kohou and his deputy, Dr Felix Bledi, as well
as Dr Aboudramane Konate, Head of Niakara Hospital.
We also thank the village chiefs and population for their
warm welcome during each of our visits. Finally, we are
grateful to the team of nurses, health workers and
microscopists who made this study possible. The principal
investigator thanks the French Institut de Recherches

pour le Developpement, especially Dr F. Riviere and
Dr B. Philippon, for their support.
References
Armstrong Schellenberg JRM, Smith T, Alonso PL & Hayes RJ
(1994) What is clinical malaria? Finding case definitions for
field research in highly endemic areas. Parasitology Today 10,
439442.
Audibert M, Josseran R, Josse R & Adjidji A (1990) Irrigation,
schistosomiasis, and malaria in the Logone Valley, Cameroon.
American Journal of Tropical Medicine and Hygiene 42,
550560.
Audibert M, Mathonnat J & Henry MC (2003) Malaria and
property accumulation in rice production systems in the
savannah zone of Cote dIvoire. Tropical Medicine and
International Health 8, 471483.
Boudin C, Robert V, Carnevale P & Ambroise TP (1992) Epidemiology of Plasmodium falciparum in a rice field and a savanna area
in Burkina Faso. Comparative study on the acquired immunoprotection in native populations. Acta Tropica 51, 103111.
Boudin C, Robert V, Verhave J-P, Carnevale P & AmbroiseThomas P (1991) Plasmodium falciparum and P. malariae
epidemiology in a west African village. Bulletin World Health
Organization 69, 199205.
Brown NT, Bing TP & Chulay JD (1982) Determination of
chloroquine and its de-ethylated metabolites in human plasma
by ion-pair high-performance liquid chromatography. Journal of
Chromatography 229, 248254.
Carnevale P, Guillet P, Robert V et al. (1999) Diversity of malaria
in rice growing areas of the Afrotropical region. Parassitologia
41, 273276.
Coosemans MH (1985) Comparaison de lendemie malarienne
dans une zone de riziculture et dans une zone de culture de coton
dans la plaine de la Rusizi, Burundi. Annales de la Societe Belge
de Medecine Tropicale 65, 187200.
Faye O, Fontenille D, Gaye O et al. (1995) Malaria and rice
growing in the Senegal river delta (Senegal). Annales de la
Societe Belge de Medecine Tropicale 75, 179189.
Faye O, Gaye O, Herve J-P, Diack PA & Diallo S (1993) Malaria
in the Sahelian zone of Senegal. 2. Parasitic indices. Annales de
la Societe Belge de Medecine Tropicale 73, 3136.
Gioda A (1992) Identical causes but various effects: irrigation,
health and development. Secheresse 3, 227234.
Laventure S, Mouchet J, Blanchy S et al. (1996) Rice, a source of
life and death on the Madagascar plateaux. Cahiers dEtudes et
de Recherches Francophones Sante 6, 7986.
Lindsay SW, Wilkins HA, Zieler HA, Daly RJ, Petrarca V & Byass
P (1991) Ability of Anopheles gambiae mosquitoes to transmit
malaria during the dry and wet seasons in an area of irrigated
rice cultivation in The Gambia. Journal of Tropical Medicine
and Hygiene 94, 313324.
Rogier C, Ly AB, Tall A, Cisse B & Trape JF (1999) Plasmodium
falciparum clinical malaria in Dielmo, a holoendemic area in
Senegal: no influence of acquired immunity on initial
457
symptomatology and severity of malaria attacks. American

Journal of Tropical Medicine and Hygiene 60, 410420.
Smith T, Armstrong Schellenberg J & Hayes R (1994) Attributable
fraction estimates and case definitions for malaria in endemic
areas. Statistics in Medicine 13, 23452358.
Smith T, Hurt N, Teuscher T & Tanner M (1995) Is fever a good
sign for clinical malaria in surveys of endemic communities?
American Journal of Tropical Medicine and Hygiene 52,
306310.
Snow RW, Craig M, Deichmann U & Marsh K (1999) Estimating
mortality, morbidity and disability due to malaria among
Africas non-pregnant population. World Health Organization
77, 624640.
Snow RW, Peshu N, Forster D et al. (1998) Environmental

and entomological risk factors for the development of clinical
malaria among children on the Kenyan coast. Transactions
of the Royal Society of Tropical Medicine and Hygiene 92,
381385.
Trape J-F, Rogier C, Konate L et al. (1994) The Dielmo project:
a longitudinal study of natural malaria infection and the
mechanisms of protective immunity in a community living
in a holoendemic area of Senegal. American Journal of Tropical
Medicine and Hygiene 51, 123137.
Zeger SL & Liang K (1986) Longitudinal data analysis for discrete
and continuous outcomes. Biometrics 42, 121130.
Authors
Elena Akodo and Thomas M. Teuscher, West Africa Rice Development Association, Bouake, Cote dIvoire. Tel.: +225 3163 8983;
E-mail: teuschert@who.ch
Marie C. Henry (corresponding author), I. Nzeyimana, S. B. Assi, J. Dossou-Yovo and P. Carnevale, Institut P. Richet, BP 1500,
Bouake 01, Cote dIvoire. Tel.: +225 31 63 37 46; E-mail: depauw.henry@fasonet.bf; inzevimana2003@yahoo.fr; assisergi@yahoo.fr;
dossoup2002@yahoo.fr; Pierre.Carnevale@mpl.ird.fr
M. Audibert and J. Mathonnat, CERDI, 65 Boulevard F. Mitterand, F-63000, Clermont-Ferrand. E-mail: M.Audibert@cerdi.uclermont1.fr; J.Mathonnat@cerdi.u-clermont1.fr
C. Rogier and A. Keundjian, BP46, Parc du Pharo, 13998 Marseille-Armees, France. Tel.: +33 491 15 01 50/52;
E-mail: christophe.rogier@wanadoo.fr; imtssa.pra@wanadoo.fr
458

Inland Valley Rice Production Systems and Malaria Infection and Disease in The Savannah of Co Te D'ivoire

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Inland Valley Rice Production Systems and Malaria Infection and Disease in The Savannah of Co Te D'ivoire

Uploaded by

Copyright:

Available Formats

Tropical Medicine and International Health

volume 8 no 5 pp 449458 may 2003