Am. J. Trop. Med. Hyg., 65(3), 2001, pp. 197–203
Copyright 
 2001 by The American Society of Tropical Medicine and Hygiene
INCIDENCE OF SYMPTOMATIC AND ASYMPTOMATIC PLASMODIUM FALCIPARUM
INFECTION FOLLOWING CURATIVE THERAPY IN ADULT RESIDENTS OF
NORTHERN GHANA
S. OWUSU-AGYEI, K. A. KORAM, J. K. BAIRD, G. C. UTZ, F. N. BINKA, F. K. NKRUMAH,
D. J. FRYAUFF, AND S. L. HOFFMAN
Navrongo Health Research Center, Navrongo, Ghana; Noguchi Medical Research Institute, Legon, Ghana;
Naval Medical Research Center, Silver Spring, Maryland; U.S. Naval Medical Research Unit No. 3, Cairo, Egypt
Abstract. Adult residents of holoendemic malaria regions in Africa have a naturally acquired immunity (NAI) to
malaria that renders them more resistant to new infections, limits parasitemia, and reduces the frequency and severity
of illness. Given such attributes, it is not clear how one might evaluate drug or vaccine efficacy in adults without
serious confounding. To determine symptomatic and asymptomatic malaria attack rates in adults of northern Ghana,
197 men and women underwent curative therapy for any pre-existing malaria infections at the start of the high
transmission (wet) season. They were monitored for first parasitemia and first clinical episode of infection by Plas-
modium falciparum over a 20-week period (May–October 1996). The cumulative incidence of primary infection by
P. falciparum was 0.98 and incidence density of infection was calculated to be 7.0 cases/person-year. Symptoms were
reported by 19.5% of the individuals at the time of first recurrent parasitemia. Incidence of infection, parasite density,
and the frequency of symptoms were comparable in males and females. The results suggest that NAI did not provide
these adults with significant defense against rapid re-infection and suggest that this population-infection design could
serve to demonstrate the efficacy of a drug or vaccine in preventing parasitemia.
INTRODUCTION data needed for valid and efficient design of malaria vaccine
and anti-malarial drug trials in the region.
Malaria is one of the leading causes of morbidity and
mortality in the tropics, affecting individuals of all ages, but
SUBJECTS AND METHODS
primarily very young children and pregnant women.1 Inte-
grated approaches against both the parasites and vectors are Study site and population. The Kassena-Nankana district
required to control this disease with vaccines offering the lies within the Guinea savanna woodlands between latitudes
best hope for sustained control in areas of intense transmis- 10
30 and 11
00 north and between longitudes 10
00 and
sion.2,3 10
30 west. It is located in the upper northeast region of
Malaria morbidity in adult Africans has long been consid- Ghana at its border with Burkina Faso. The Kassena-Nan-
ered of minor importance4 because adult residents of holo- kana district has an area of 1,675 km2 and a population of
endemic areas have usually developed high levels of protec- 140,000 living in roughly 13,000 dispersed compounds. The
tive immunity as a result of their long and continuous ex- main occupation is subsistence farming of predominantly
posure to malaria infections. The two forms of immunity millet and livestock. The Kassena-Nankana district has a
described are 1) an anti-disease immunity that allows one to mean monthly temperature range of 20–45
C and rainfall
carry parasite loads without symptoms, and 2) an anti-par- averages between 800 and 1,000 mm per year, occurring
asite immunity that is responsible for a marked reduction of mainly between May and October. Malaria transmission by
parasite densities after a certain age.5–7 Malaria slides from Anopheles gambiae and Anopheles funestus peaks at the
adults in highly endemic areas are usually positive regardless close of wet season (during October and November). Malaria
of the clinical context8 and adults are known to become rap- parasite prevalence among children  5 years old in the
idly re-infected with malaria parasites after treatment.9 Ap- Kassena-Nankana district was significantly greater in the wet
propriate criteria for deciding when adults should receive season (May to October) than in the dry season (November
treatment for malaria have never been agreed upon. It has to April).16 Malaria is the leading diagnosis of illness in the
been variably suggested that treatment should be based on district hospital, accounting for 60% of all admissions and
parasite density threshold,10 febrile condition,11,12 symptom 41% of hospital deaths.17
history,13 or an association between symptom history and Subject selection and consent. A representative sample
parasitemia.14,15 population of adult men and non-pregnant women between
The present study was conducted to measure the incidence the ages of 18 and 55 years was sought. The sample was
and features of symptomatic and asymptomatic infection by randomly selected using the Navrongo Demographic Sur-
Plasmodium falciparum in malaria-immune adult Ghanaians veillance System, a longitudinal census of all district resi-
who had been radically cured of all pre-existing Plasmodium dents that is updated every three months.18 We tallied the
infections. Effort was made to study newly incident infec- total number of eligible adults (48,032) within the four geo-
tions in comparably sized cohorts of males and females 18– graphic zones of the district and randomly selected 16 index
55 years old who were lifetime residents of the Kassena- compounds, with selections weighted by population size.
Nankana district of northern Ghana. This study was con- Chiefs, elders, and potential subjects were then visited, the
ducted during the peak malaria transmission season to char- research work was explained and discussed, and opinions
acterize malaria in the area and to provide the foundation were sought before enrollment. Translation, narration, dis-
197
198 OWUSU-AGYEI AND OTHERS
cussion, and communal consent were the essential elements sidered negative if no parasites were observed within 100
of the informed consent process. high-power fields.
This research was approved by scientific and ethical re- Our working diagnosis of clinical malaria was a patent
view boards of the Ghanaian Ministry of Health and the U.S. asexual stage parasitemia accompanied by any one or more
Navy and was conducted in accordance with regulations of the following symptoms: fever, headache, chills, myalgia,
governing the protection of human subjects in medical re- dizziness, nausea, and diarrhea. All symptomatic subjects
search. with microscopy-confirmed malaria were promptly treated
Enrollment and screening. The screening process in- with chloroquine diphosphate (25 mg base/kg of body
volved clinical history, vital signs, physical examination by weight over a 48-hr period) and monitored for outcome.
a physician, collection of 1.0-ml venous blood for malaria Chloroquine failures were treated with three tablets of Fan-
smears, and hematology screening. Urine was collected from sidar (500 mg of sulfadoxine plus 25 mg of pyrimethamine;
all women for -human chorionic gonadotropin pregnancy F. Hoffman-La Roche, Basel, Switzerland).
testing. Laboratory screening involved microscopy of Gi- Data handling and analysis. Cumulative incidence (CI),
emsa-stained thick and thin blood smears, hemoglobin test- expressed as the risk of an outcome and/or probability of
ing, and qualitative testing for glucose-6-phosphate dehydro- infection during the 20 weeks, and incidence density (ID),
genase (G-6-PD) deficiency. Remaining blood and plasma which estimates new infections per unit of person-time at
were archived for future laboratory work. Specific exclusion risk, were used as quantitative tools in determining the fre-
criteria were pregnancy, allergic reaction to antimalarial quency of newly incident Plasmodium infections in this co-
drugs, and illness or a condition requiring hospitalization. hort. Due to the pre-erythrocytic stage incubation period of
Radical cure therapy of malaria. Just prior to the start the parasite, the first 10 days following curative therapy were
of the rainy season (April 1996), radical cure therapy was excluded from analyses as risk-free person-time. To identify
used to achieve complete clearance of any patent, dormant, predictors of a positive blood slide, and to determine wheth-
or incubating malaria infection from blood and liver. Vol- er the use of specific complaints or combinations of com-
unteers were treated with oral quinine sulfate (650 mg three plaints significantly increased the likelihood of correctly pre-
times a day for four days) followed by doxycycline, (100 dicting a positive blood smear, we used logistic regression
mg twice a day for 10 days). Subjects testing normal for G- models that fitted parasite positivity against symptoms re-
6-PD (NADP spot test; Sigma Chemical Co., St. Louis, corded. The multivariate interactions among reported symp-
MO) received primaquine (0.5 mg base/kg of body weight toms and parasitemia were investigated by graphical mod-
once a day for 14 days). Nineteen subjects who tested as eling19 using the MIM 3.0 software.20 The best fitting model
deficient for G-6-PD received only quinine and doxycycline was determined by means of a backwards elimination pro-
treatment. All doses were supervised and directly observed cedure and a less stringent cutoff value of P  0.1 was
over the 18 days of radical curative or 14 days of curative applied to improve the likelihood of finding associations giv-
therapy. en a small sample size.
Surveillance and referral. Routine blood films were ob-
tained on day 15 of the initial treatment and were repeated RESULTS
at two-week intervals for the duration of the study (20
weeks) by trained fieldworkers permanently residing in each The study plan is chronologically depicted as a flow dia-
of the 16 compound clusters. Fieldworkers visited subjects gram in Figure 1. During the initial radical or curative ther-
three times a week at their homes to inquire of their general apy stage, 98.7% of the 16,758 drug doses were adminis-
health and check for fever by measuring axillary tempera- tered and witnessed with only one volunteer dropping out.
ture. Each visit began with a direct question: ‘‘Have you The baseline characteristics of the enrolled study population
been sick since the last visit?’’ If the answer was no, the that completed radical or curative treatment and contributed
interview ended. If the answer was yes, the subject was person-time and/or study endpoints are presented in Table 1.
asked to describe his or her symptoms and these were re- The mean age, weight, and hemoglobin concentration among
corded. If a subject complained of current illness to the field- females was significantly different from that of males; older
worker or had an axillary temperature 
 37.5
C, a non-rou- age was possibly an artifact that arose from the necessity of
tine blood film was obtained. Subjects were transported to restricting female enrollment to those not pregnant or nurs-
the district hospital where the blood smear was examined by ing. Patent malaria infections were identified in 110 (55.8%)
an expert malaria microscopist. Based upon presenting signs, of 197 subjects at enrollment and P. falciparum accounted
symptoms, and microscopy results, physicians decided upon for 102 (92.7%) of the 110 malaria infections, with the re-
a treatment plan. maining 7.3% due to infections by P. malariae (7) and P.
Study end points, treatment, and follow-up. The two ovale (1). Gametocytes were observed in 29.6% of these
primary outcome variables in this study were 1) time to the subjects. Glucose-6-phosphate dehydrogenase deficiency
primary post-cure parasitemia and 2) time to the first episode was detected in 19 (9.6%) of the 197 enrolled subjects (7.1%
of parasitemia associated with clinical symptoms. Blood of the women versus 11.6% of the men: P  0.28). Baseline
slides were prepared as Giemsa-stained thick and thin malaria prevalence was significantly greater among males
smears examined by 1,000 light microscopy using oil im- (58% versus 38%; P  0.01). All subjects belonged to the
mersion. Parasitemias were scored per microliter of blood two main ethnic groups (Kasssim and Nankam), in near bal-
by counting the number of asexual parasites per 200 white anced numbers, and had lived their entire lives in the area.
blood cells, assuming 8,000 white blood cells per microliter, Blood smear screening on day 15 following the start of
and multiplying the parasite count by 40. A slide was con- radical or curative therapy demonstrated parasite clearance
INCIDENCE OF MALARIA IN ADULT GHANAIANS 199
FIGURE 1. Study plan.
in all subjects. The first infections thereafter were identified fections per person-year of exposure. There were four fe-
during week 3 of post-cure follow-up. The cumulative in- males (4.7%) and three males (3.6%) who did not develop
cidence of primary P. falciparum infection over the 20 a detectable parasitemia during their 8–20-week participa-
weeks following cure are plotted for males and females in tion. All were G-6-PD normal. One of these subjects was
Figure 2. Both sexes demonstrated the most profound in- infected with P. falciparum at enrollment, two others with
crease in new infections during weeks 5–10, when the CI P. malariae, and six of the seven made voluntary clinic vis-
increased from 0.22 to 0.81 in males and 0.21 to 0.77 in its for diagnosis and treatment of illness during the study
females. Three subjects remained consistently aparasitemic period. Analysis by age, which divided the population into
to the 20-week endpoint. Four others dropped out of the two near equal groups,  40 years old (n  100) and 
 40
study (negative) at various earlier time points. The collective years old (n  97), revealed that mean times to parasitemia
20-week CI for P. falciparum infection in adults was 0.98. and symptomatic parasitemia in the older age group were
The respective 20-week wet season IDs for infection in both significantly delayed (P  0.01) relative to those in the
males and females was 7.2 and 6.7 (relative risk [RR]  younger group (Figure 4). As a result, wet season ID in the
1.07, 95% confidence interval  0.70–1.24) infections/per- older group was 5.8 infections/person-year compared with
son-year. The CI of illness coincident with parasitemia dur- 7.1 in the younger group (P  0.17). Parasite densities at
ing this period was 0.42 among males and 0.49 among fe- infection and at the time of symptoms were similar in the
males (Figure 2). Analysis by location (Figure 3) found the two groups.
acquisition of primary infection to be most rapid in the west- There were 190 primary, post-cure infections by P. fal-
ern compounds and slowest in the south (mean time to in- ciparum during the 20 weeks following cure. Symptoms and
fection: west  7.2 weeks versus south  8.7 weeks; P  physical complaints characterized the first appearance of par-
0.03). Mean times to the first episode of confirmed symp- asites in 37 (19.5%) of these 190 infections. Comparable
tomatic malaria were similar among the four locations stud- proportions of males and females were symptomatic at this
ied. first parasitemia (males: 22 of 109  20.2%; females: 15 of
Table 2 shows the characteristics of these primary infec- 81  18.5%). Fever was present in eight (21.6%) of the 37
tions in male and female subjects. There were no statistically primary infections that presented with clinical illness, but
significant differences between sexes for either CI, ID, the was not associated with the level of parasitemia. Among
frequency of symptomatic primary infections, or the geo- individuals with asymptomatic primary parasitemias who
metric mean (GM) density of parasitemias. Irrespective of were not treated, asexual stage parasites were variably pre-
sex, adults experienced a 20-week wet season ID of 7.0 sent in subsequent follow-up slides. Parasitemia was detect-
(95% confidence interval  5.1–6.8) new P. falciparum in- ed in 65% of 566 follow-up slides from 113 representative
TABLE 1
Baseline characteristics of the Ghanaian study population*
Characteristic Males Females P
No. of subjects (%) 112 (56.8) 85 (43.1) –
Mean age (years) (95% CI) 35.4 (33.3–37.5) 43.3 (41.3–45.2) 0.01
Mean weight (kg) (95% CI) 57.4 (56.0–58.8) 50.3 (49.0–51.6) 0.01
Mean Hb level (g/dl) (95% CI) 12.1 (11.8–12.4) 10.8 (10.3–11.2) 0.01
No. Plasmodium falciparum positive (%) 64 (57.1) 33 (38.8) 0.01
GM parasitemia/l (95% CI) 135 (103–177) 98 (78–123) 0.07
No. febrile (%) 2 (1.0) 1 (0.5) –
* CI  confidence interval; Hb  hemoglobin; GM  geometric mean.
200 OWUSU-AGYEI AND OTHERS
FIGURE 2. Cumulative incidence after radical cure of primary Plasmodium falciparum infections and of P. falciparum infections with
symptoms in adult males and females in the wet season (May–December 1996) in northern Ghana.
subjects and usually not detected over more than three con- modeling. This approach showed dizziness and myalgia to
secutive examination dates. During the course of study time be directly linked to parasitemia, with diarrhea independent
applied to the 153 infected-asymptomatic volunteers, there of all other variables. The other reported symptoms (chills,
were 242 occasions, judged to be clinical malaria on the nausea, headaches, and fever) were independent of the par-
basis of signs and symptoms that prompted blood smear ex- asitemia given a presentation of dizziness and/or myalgia.
amination. Asexual stage parasites of P. falciparum were Patients who reported chills or nausea were likely to also
present at 62 of these 242 occasions and those of P. malariae report dizziness or myalgia, but rarely both. Headache and
were present in one case. Asexual stages of P. falciparum fever were independent of the other symptoms given the
were thus associated with 99 of 279 cases of malaria-like presence of chills.
illness in these previously cured adults. The frequencies of Table 3 shows the range and GM densities of primary
specific complaints/symptoms associated with these infec- post-cure P. falciparum parasitemias that were measured in
tions are plotted against those of 180 other cases where par- symptomatic and asymptomatic cases. Relatively low-den-
asites were undetected (Figure 5). The frequency of dizzi- sity parasitemias characterized more than 85% of the infec-
ness, chills, and headache were significantly greater among tions and were associated with physical complaints. The den-
those with a confirmed parasitemia. sity of asymptomatic parasitemias at first appearance of in-
Stepwise logistic regression that was applied to the data fection was no different than the density of parasitemia later
revealed dizziness and myalgia as potential (P  0.1) pre- associated with the first occasion of clinical illness following
dictors of a positive blood slide, while diarrhea was unre- curative treatment. There were no significant differences be-
lated. The multivariate dependencies among parasitemia and tween sexes, or between symptomatic and asymptomatic
reported symptoms were further investigated using graphical cases in the frequency of high parasitemias or in the GM
densities of parasitemias.
DISCUSSION
The results of this study show that wet season malaria
transmission in the Kassena-Nakana district of northern Gha-
na is intense and that the incidence of infection and clinical
illness associated with malaria in adults is virtually identical
in males and females. Despite a natural immunity derived
from a lifetime of intense exposure, our results demonstrate
that adults become rapidly re-infected and parasitemic fol-
lowing effective radical cure treatment. Neither genetic re-
sistance nor long-acquired immunity appears to be sufficient-
ly strong or widely present in this population to completely
inhibit parasitemia. Rapid re-infection and patency among
FIGURE 3. Cumulative incidence after radical cure of primary
Plasmodium falciparum infections by location in the Kassena-Nan- treated malaria-immune adults was recognized more than a
kana District of northern Ghana during the wet season (May–De- decade ago in western Kenya.9 This general susceptibility
cember 1996). and rapid appearance of low-density blood stage infections
INCIDENCE OF MALARIA IN ADULT GHANAIANS 201
TABLE 2
Characteristics of first infections with Plasmodium falciparum after radical cure in adult male and female study subjects during the wet season
(May–December 1996) in northern Ghana*
Characteristic Males Females P
Person-weeks at risk 777 629 –
No. with primary P. falciparum infections/total subjects (%) 109/112 (97.3) 81/85 (95.3) 0.45
No. symptomatic at first positive slide/total positive subjects (%) 22/109 (20.2) 15/81 (18.5) 0.77
Cumulative incidence/20 weeks of the wet season 0.99 0.97 –
Incidence density/person-year 7.3 6.7 0.56
GM parasitemia/l—first appearance—Asymptomatic (95% CI) 129 (96–169) 162 (118–223) 0.27
GM parasitemia/l—first appearance—Symptomatic (95% CI) 451 (171–1,190) 347 (131–921) 0.72
* GM  geometric mean; CI  confidence interval.
may result from the ineffectiveness of anti-sporozoite and tion-matched cohort of infants (6–24 months old) who were
anti-liver stage immunity or to an asexual blood stage effect treated and followed through 20 weeks of wet season in the
coupled with a pre-erythrocytic stage immunity that reduces following year (Adult ID  7.03 infections/person-year ver-
but does not completely eliminate all infected hepatocytes.2,9 sus Infant ID  7.06 infections/person-year; RR  0.99,
In spite of comparable malaria attack rates measured in 95% confidence interval  0.82–1.19).21 The cumulative in-
adult males and females, the relatively small size and eco- cidence profiles of adult and infant cohorts indicate the same
logical uniformity of the Kassena-Nakana district study area, rapid increase and plateau effect. Among infants, the most
and seasonal climatic conditions that impinge evenly, there- rapid increase in incidence occurred between five and 10
by smoothing over ecologic and hydrologic differences, our weeks with a plateau reached at week 10 that incremented
analysis identified a significant difference between locations only slightly more by the 20-week endpoint. In an analogous
in the mean time to infection. This difference may relate to manner, new infections among adults increased most steeply
the geographic concentration of well-established irrigation from weeks 5 (male CI  0.22, female CI  0.21) to 10
systems in the western compounds. If this was not due to (male CI  0.81, female CI  0.77). Among infants, only
chance, it is likely that even more highly significant differ- 18 new infections occurred in weeks 11–20, leaving 2.3%
ences between locations will exist in the Kassena-Nakana of the wet season cohort parasite-free. In a similar manner,
district during dry season conditions. This observation may the adult CI was 0.88 in week 13 and only 18 additional
be supported by location analysis of data from recent studies subjects (9.5% of the infected cohort) developed patent in-
conducted during the wet and dry season using infant co- fections during the ensuing seven weeks of study.
horts. Drug and vaccine efficacy trials might be intentionally We assumed that adult life-long residents of Navrongo, in
focused within particular areas of the Kassena-Nakana dis- contrast to their young children, would manifest a lower or
trict to achieve rapid endpoints and tight confidence inter- delayed incidence of patent re-infection, lighter parasitemias
vals. at the time of re-infection, and mild or negligible illness. We
Contrary to expectations, the ID of primary parasitemia found lighter parasitemias and decreased clinical manifes-
in this adult cohort was virtually identical to that of a loca- tations in adults, but a near identical rate of re-infection fol-
lowing radical cure. A lower attack rate in adults might have
been observed had both cohorts been studied simultaneously,
but was not apparent in our comparison of wet season IDs
for adults during 1996 and for infants during 1997. Height-
ened transmission impacting upon adults, or reduced trans-
FIGURE 4. Mean time to primary infection by Plasmodium fal- FIGURE 5. Comparative physical complaints/symptoms reported
ciparum and primary symptomatic infection by age group in adult by adult Ghanaians with positive or negative malaria blood smears.
Ghanaians. Error bars indicate 95% confidence intervals. * Indicates a statistically significant difference between groups.
202 OWUSU-AGYEI AND OTHERS
TABLE 3
Range and geometric mean densities of first Plasmodium falciparum infections after radical cure in adult Ghanaians
Asymptomatic Symptomatic
Parasitemia range (/l blood) Males Females Males* Females†
2,000 82 64 40 46
2,000–9,999 3 1 3 5
10,000–19,999 2 0 2 0
	20,000 0 1 2 1
Combined 87 66 47 52
GM parasitemia (range) 129 (40–12,560) 162 (40–31,640) 199 (40–97,680) 200 (40–22,000)
* Males: 22 symptomatic at first positive slide; 25 of 87 others (initially asymptomatic-positive) later became symptomatic-positive.
† Females: 15 symptomatic at first positive slide; 37 of 66 others (initially asymptomatic-positive) later became symptomatic-positive.
mission experienced by the infant cohort may have obscured symptoms and parasitemia were investigated. Parasitemia
real differences attributable to age-dependent immunity. was directly linked only to dizziness and myalgia, which in
Interestingly, in contrast to the near equal incidence den- turn were likely to occur in combination with chills or nau-
sity seen in adults and children, we observed among adults sea. This is in contrast to other studies that did not find
that the 	 40-year-old age group had a lower ID of re-in- symptom reports useful in diagnosing clinical malaria.8 Our
fection compared with those  40 years old, suggesting an preliminary findings suggest that adults presenting with diz-
effective age-dependent, naturally-acquired immunity. Under ziness or myalgia in association with chills and headache are
identical conditions of risk, an anti-parasite immunity ca- likely to be parasitemic. We believe that this graphical ap-
pable of barring certain genotypes and suppressing parasit- proach to the identification of key physical complaints and
emia to sub-clinical levels in adults would be expected to symptoms associated with parasitemias should be tested
prevent or delay the patency of primary infections relative more rigorously.
to those sustained by infants. The incidence data herein reported for adult males and
As expected, adults experienced parasite densities and fe- females provides the necessary foundation for rational de-
brile episodes that were considerably reduced from those sign of trials to evaluate drug or vaccine effects. Although
recorded in the infant cohorts.21 It was interesting to note African infants and pregnant women will certainly be the
that many of the infected Ghanaian adults in the present target population for vaccination when an effective product
study experienced episodes of illness that corresponded with, becomes deliverable, adults and school children will proba-
and may have arisen from, their P. falciparum infections. bly be the earliest recipients, in whom safety, tolerance, and
We concede that the exclusive role of the parasitemias as a infection-blocking qualities can be readily assessed. In this
cause of their illness cannot be ascertained with certainty regard, it was important to demonstrate near universal sus-
since the same signs and symptoms were reported frequently ceptibility to re-infection and its rapid manifestation as pat-
by these same individuals in the absence of parasites. How- ent parasitemias in adult Ghanaian males and females. Ad-
ever, the 37 occasions of illness that corresponded precisely ditionally important in planning such studies is our docu-
with the first appearance of parasites following radical cure mentation of radical cure compliance and dropout rates, the
are most probably the true result of infection. Literature on proportion of radically cured subjects who dropped out prior
the subject of adult malaria implies that adults with lifelong to yielding an endpoint, and the frequency and nature of
exposure to malaria do not become ill when they become
illness that developed coincident with re-infection. Applying
re-infected with malaria parasites.4,12 Our data suggest that
our findings of an 80% attack rate within 12 weeks to a
this may not be the case when parasites have been complete-
prospective 12-week trial of a drug/vaccine with predicted
ly cleared. Interesting recent consideration has been given to
75% protective efficacy (PE), and an 80% probability (pow-
the impact of liver schizonticidal therapy upon protective
immunity elicited by immunization of mice with radiation- er) of having the lower 95% confidence limit for that PE 

attenuated sporozoites,22 and to the potential loss of premu- 70%, we would require a population of 1,004 adults. With
nition in clinically immune adults who have undergone such these variables kept the same, but with the lower 95% con-
curative therapy to eliminate their ‘‘benign’’ chronic malaria fidence limit relaxed to 
 60% for the PE, we would need
infections.23 only to enroll 151 subjects.
None of the adult patients became severely ill, but they In conclusion, this study has shown that Ghanaian adults
reported physical complaints that limited their capacity to of the northern Kassena-Nakana district experience a sea-
function normally, prompting them to interrupt their usual sonal ID of infection by P. falciparum that was comparable
routine and seek medical care. Most of the symptoms ob- with that of infants living in the same communities. Within
served were rather nonspecific. However, there were signif- the 20-week wet season, virtually all adults had manifested
icant differences between parasitemic and non-parasitemic a patent primary infection and many reported illness that
patients that may not have been due entirely to chance. Al- coincided with parasitemia. Adults experienced predomi-
though we did not undertake or power this investigation to nantly low-density infections, characterized by mild clinical
test a specific hypothesis of adult clinical predictors of par- illness. Dizziness, myalgia, chills, fever, and headache were
asitemia, we found dizziness and myalgia to be the best mul- the most commonly reported symptoms that prompted health
tivariate predictors of parasitemia. This pattern was con- clinic visits seeking medical care. This study provides the
firmed when multivariate interactions between different foundation for designing efficient, practical clinical trials and
INCIDENCE OF MALARIA IN ADULT GHANAIANS 203
preventive strategies against malaria in this holoendemic re- 8. Trape JF, Peelman P, Mourault-Peelman B, 1985. Criteria for
gion. diagnosing clinical malaria among a semi-immune population
exposed to intense and perennial transmission. Trans R Soc
Trop Med Hyg 79: 435–442.
Acknowledgments: We thank the people of the Kassena-Nankana 9. Hoffman SL, Oster CN, Plowe CV, Woollet GR, Beier JC, Chu-
District in northern Ghana for their cooperation, which made this lay JD, Wirtz RA, Hollingdale MR, Mugambi M, 1987. Nat-
study possible. Sincere appreciation also goes to the field and lab- urally acquired antibodies to sporozoites do not prevent ma-
oratory staff of the Navrongo Health Research Center and Noguchi laria: vaccine development implications. Science 237: 639–
Memorial Institute for Medical Research, and especially to Charles 642.
Attiogbe, for technical assistance. We also thank Drs. Kweku Enos 10. Rogier C, Commenges D, Trape JF, 1996. Evidence for an age-
and Mensah Afful (War Memorial Hospital, Navrongo). Sincere dependent threshold effect of malaria parasite density on the
thanks also go to Dr. Trevor Jones (Naval Medical Research Center) occurrence of fever suggesting an inverse relationship be-
and Drs. Thomas Smith and Ivo Muller (Swiss Tropical Institute) tween anti-parasite and anti-toxic immunity. Am J Trop Med
for their helpful comments and statistical advice. Hyg 54: 613–619.
Financial support: The work was supported by the Naval Medical 11. Rougemont A, Breslow N, Brenner E, Moret AL, Dumbo O,
Research Center independent research initiative funds and the STO Dolo A, Soula G, Perrin L, 1991. Epidemiological basis for
F6.1 work unit 61102AA0101.BXF.1431 of the Military Infectious clinical diagnosis of childhood malaria in endemic zone in
Disease Research Program. West Africa. Lancet 338: 1292–1295.
12. Smith T, Hurt N, Teuscher T, Tanner M, 1995. Is fever a good
Disclaimer: The views of the authors expressed herein do not purport sign for clinical malaria in surveys of endemic communities?
to reflect those of the Ghanaian Ministry of Health, the U.S. Navy, Am J Trop Med Hyg 52: 306–310.
or the U.S. Department of Defense. 13. Bassett MT, Taylor P, Bvirakare J, Chiketa F, Govere E, 1991.
Authors’ addresses: S. Owusu-Agyei and F. N. Binka, Navrongo Clinical diagnosis of malaria: can we improve? J Trop Med
Health Research Center, Navrongo, Ghana. K. A. Koram and F. K. Hyg 94: 65–69.
Nkrumah, Noguchi Medical Research Institute, Legon, Ghana. J. K. 14. Mkawagile DSM, Kihama CM, 1986. Relationship between
Baird, and D. J. Fryauff, Naval Medical Research Center, 503 Robert clinical diagnosis of malaria and parasitaemia in adult patients
Grant Avenue, Silver Spring, MD, 20910-7500. G. C. Utz, U.S. attending Mwananyamala dispensary, Dar es Salaam. Cent
Naval Medical Research Unit No. 3, Cairo, Egypt. S. L. Hoffman, Afr J Med 32: 2–5.
Celera Genomics, 45 West Gude Drive, Rockville, MD 20850. 15. Stein CM, Gelfand M, 1985. The clinical features and labora-
tory findings in acute Plasmodium falciparum malaria in Har-
are, Zimbabwe. Cent Afr J Med 31: 166–170.
REFERENCES 16. Binka FN, Morris SS, Ross DA, Arthur P, Aryeetey ME, 1994.
Patterns of malaria morbidity and mortality in children in
1. World Health Organization, 1997. Tropical Disease Research northern Ghana. Trans. R Soc Trop Med Hyg 88: 381–385.
Progress 1995–1996. Thirteenth Program Report UNDP/ 17. Ministry of Health Report, 1996. Report of the War Memorial
World Bank/WHO Special Program for Research & Training Hospital, Kassena-Nankana District of Ghana.
in Tropical Diseases (TDR). 18. Binka FN, Ngom P, Phillips JF, Adazu K, MacLeod BB, 1999.
2. Hoffman SL, ed, 1996. Malaria Vaccine Development: A Multi- Assessing population dynamics in a rural African society: the
Immune Response Approach. Washington, DC: American So- Navrongo Demographic Surveillance System. J Biosoc Sci
ciety for Microbiology Press. 31: 375–391.
3. Miller LH, Hoffman SL, 1998. Review: research towards vac- 19. Whittaker J, 1990. Graphical Models in Applied Multivariate
cines against malaria. Nature Med 4: 520–524. Statistics. Chichester, United Kingdom: John Wiley & Sons.
4. Wilson DB, 1950. A review of hyperendemic malaria. Trop Dis 20. Edwards D, 1995. Introduction to Graphical Modeling. New
Bull 47: 677–698. York: Springer Verlag.
5. Daubersies P, Sallenave-Sales S, Magne S, Trape JF, Contamin 21. Baird JK, Owusu-Agyei S, Utz GC, Koram K, Barcus MJ, Bin-
H, Fandeur T, Rogier C, Mercereau-Puijalon O, Druilhe P, ka FN, Jones TR, Hoffman SL, Nkrumah FK, 2001. Seasonal
1996. Rapid turnover of Plasmodium falciparum populations malaria attack rates in infants and young children in northern
in asymptomatic individuals living in a high transmission Ghana. Am J Trop Med Hyg (In press).
area. Am J Trop Med Hyg 54: 18–26. 22. Scheller LF, Azad AF, 1995. Maintenance of protective immu-
6. Trape JF, Rogier C, Konate L, Diagne N, Bouganali H, Canque nity against malaria by persistent hepatic parasites derived
B, Legros F, Badji A, Ndiaye G, Ndiaye P, Brahimi K, Faye from irradiated sporozoites. Proc Natl Acad Sci USA 92:
O, Druilhe P, da Silva LP, 1994. The Dielmo Project: a lon- 4066–4068.
gitudinal study of natural malaria infection and the mecha- 23. Smith T, Felger I, Tanner M, Beck H-P, 1999. The epidemiology
nisms of protective immunity in a community living in a ho- of multiple Plasmodium falciparum infections. II. Premuni-
loendemic area in Senegal. Am J Trop Med Hyg 51: 123–137. tion in Plasmodium falciparum infection: insights from the
7. Day KP, Marsh K, 1991. Naturally acquired immunity to Plas- epidemiology of multiple infections. Trans R Soc Trop Med
modium falciparum. Immunol Today 12: A68–A71. Hyg 93 (suppl 1): 59–64.