2160 • JID 2005:192 (15 December) • Ezeamama et al.

M A J O R A R T I C L E

Functional Significance of Low-Intensity PolyparasiteHelminth Infections in Anemia

Amara E. Ezeamama,1,3 Jennifer F. Friedman,1,4 Remigio M. Olveda,6 Luz P. Acosta,6 Jonathan D. Kurtis,1,5

Vincent Mor,2,3 and Stephen T. McGarvey1,3

1International Health Institute, 2Center for Gerontology Research and Health Care, and Departments of 3Community Health,4Pediatrics, and 5Pathology, Brown University, Providence, Rhode Island; 6Department of Immunology, Research Instituteof Tropical Medicine, Manila, The Philippines

Background. We wanted to quantify the impact that polyparasite infections, including multiple concurrent low-intensity infections, have on anemia.

Methods. Three stool samples were collected and read in duplicate by the Kato-Katz method in a cross-sectionalsample of 507 children from Leyte, The Philippines. The number of eggs per gram of stool was used to define 3infection intensity categories—uninfected, low, and moderate/high (M+)—for 3 geohelminth species and Schis-tosomiasis japonicum. Four polyparasite infection profiles were defined in addition to a reference profile that con-sisted of either no infections or low-intensity infection with only 1 parasite. Logistic regression models were usedto quantify the effect that polyparasitism has on anemia (hemoglobin level !11 g/dL).

Results. The odds of having anemia in children with low-intensity polyparasite infections were nearly 5-foldhigher ( ) than those in children with the reference profile. The odds of having anemia in children infectedP p .052with 3 or 4 parasite species at M+ intensity were 8-fold greater than those in children with the reference profile( ).P ! .001

Conclusion. Low-intensity polyparasite infections were associated with increased odds of having anemia. Inmost parts of the developing world, concurrent infection with multiple parasite species is more common thansingle-species infections. This study suggests that concurrent low-intensity infections with multiple parasite speciesresult in clinically significant morbidity.

Iron-deficiency anemia is the most prevalent nutritional

deficiency worldwide [1–5]. More than 90% of affected

individuals live in the developing world, where hel-

minth infections are highly prevalent and the parasites

are endemic [5]. Helminths are known to be significant

contributors to the overall anemia burden in the de-

veloping world [6–12]. The health sequelae associated

with anemia are the most pronounced in children and

women of reproductive age [3, 13]. In addition to the

morbidity in these groups, anemia generates profound

Received 3 June 2005; accepted 18 July 2005; electronically published 11November 2005.

Presented in part: annual meeting of the American Society of Tropical Medicineand Hygiene, Miami Beach, Florida, 7–11 November 2004 (abstract 951).

Potential conflicts of interest: none reported.Financial support: National Institutes of Health (grants RO1AI48123 and K23AI52125).Reprints or correspondence: Ms. Amara E. Ezeamama, Brown University, Dept.

of Community Health, International Health Institute, Box G, Providence, RI 02912(Amara_Ezeamama@brown.edu).

The Journal of Infectious Diseases 2005; 192:2160–70� 2005 by the Infectious Diseases Society of America. All rights reserved.0022-1899/2005/19212-0021$15.00

physiological and economic costs in the general pop-

ulation [5, 14]. The multiple insults due to helminths

in chronically exposed populations are believed to per-

sist throughout the life course. For example, chronically

infected children are likely to grow into adults with a

reduced physical capacity for work, which ultimately

translates into a diminished contribution to national

productivity [14]. The negative impact that high-in-

tensity helminth infections have on hemoglobin levels

has been convincingly demonstrated through obser-

vational and interventional studies of many popula-

tions [7, 13, 15–17].

Most investigators have examined the relationship

between morbidity and helminth infections, with a fo-

cus on single-species infection at varying intensities. It

is generally thought that low-intensity single-species in-

fections are associated with little, if any, measurable

morbidity [18, 19], and this hypothesis has influenced

the treatment and control recommendations of the

World Health Organization (WHO) [20–25]. Epidemi-

ologic studies, however, have shown that polyparasit-

by guest on May 14, 2011

jid.oxfordjournals.orgD

ownloaded from

Polyparasite Infections and Anemia • JID 2005:192 (15 December) • 2161

Table 1. Characteristics of the study population.

Characteristic Children, no. (%)

Age7–9 years 134 (26.4)10–12 years 164 (32.4)13–15 years 133 (26.2)16–18 years 76 (15.0)

Male sex 289 (57.0)Socioeconomic status

Low 162 (32.0)Medium 171 (33.7)High 174 (34.3)

Schistosomiasis japonicum infectionNone 105 (20.7)Low intensity 272 (53.7)Moderate/high intensity 130 (26.6)

Necator americanus infectionNone 225 (44.6)Low intensity 237 (46.9)Moderate/high intensity 43 (8.5)

Ascaris lumbricoides infectionNone 103 (20.4)Low intensity 117 (23.2)Moderate/high intensity 285 (56.4)

Trichuris trichiura infectionNone 36 (7.1)Low intensity 218 (43.2)Moderate/high intensity 251 (49.7)

Polyparasite infection profilea

Reference 32 (6.3)I 80 (15.8)II 152 (30.0)III 179 (35.4)IV 63 (12.5)

Anemiab 101 (19.9)Nutritional statusc

Weight-for-age z scoreNormal 65 (12.8)Mildly underweight 130 (25.6)Severely underweight 312 (61.5)

Body mass index z scoreNormal 259 (51.3)Mildly underweight 174 (34.5)Severely underweight 72 (14.3)

a The reference profile was no infection or infection with 1 parasite species atlow intensity; polyparasite infection profile I was concurrent infection with 2, 3, or4 parasite species at low intensity; polyparasite infection profile II was infectionwith 1 parasite species at moderate/high intensity and all other parasite speciespresent at low intensity or absent; polyparasite infection profile III was concurrentinfection with 2 parasite species at moderate/high intensity and all other parasitespecies present at low intensity or absent; and polyparasite infection profile IV wasconcurrent infection with 3 or 4 parasite species at moderate/high intensity.

b Hemoglobin level !11 g/dL.c Reference scores were z scores compiled by the Centers for Disease Control

and Prevention/National Center for Health Statistics for children in the UnitedStatesin 2000. Healthy, z score 1�1; mildly underweight, z score 1�3 to ��1; severelyunderweight, z score ��3.

ism is the norm in helminth-endemic regions [26–33]. It is

likely that individuals in such places have infections with mul-

tiple helminth species at intensities in various combinations,

including multiple low-intensity infections. The present em-

phasis on high-intensity infections is certainly not misplaced,

given the evidence of greater morbidity with high-intensity

infections [7, 12, 13, 34–36]. However, the connotation that

low-intensity infections have little or no functional significance

needs to be reexamined in light of the many areas of the world

in which multiple parasites are endemic [26–31, 37].

Studies specifically designed to understand the functional sig-

nificance of polyparasite infections in anemia and other mor-

bidities in human populations are lacking. Such epidemiolog-

ical studies permit the investigation of a question relevant to

public health: do individuals concurrently infected with mul-

tiple helminth species at low intensity have a measurable risk

of developing anemia? The objective of this study was to ex-

amine in children the impact that different helminth infection

intensity profiles have on anemia. We hypothesized that low-

intensity polyparasitism would be associated with increased

odds of having anemia and that the odds of having anemia

would increase with the number of species present at moderate/

high (M+) intensity.

PARTICIPANTS, MATERIALS, AND METHODS

Study population and design. This cross-sectional study was

conducted in rice-farming villages in Leyte, The Philippines.

In this region, Schistosomiasis japonicum and intestinal hel-

minths are endemic [36, 38–40]. The present study involved

secondary data analysis and incorporated participants (7–30

years old) derived from 2 studies. These participants were part

of a longitudinal treatment and reinfection study focused on

the development of immunity to schistosomiasis. By design,

the population in the longitudinal study was restricted to in-

dividuals with S. japonicum infection. In addition, individuals

free of S. japonicum infection were then selected to serve as

control subjects in other studies of cognition [41]. The prev-

alence of S. japonicum infection in persons 7–30 years old in

the study area was 60%. The present study was restricted to

632 children �18 years old who provided stool samples and

informed consent/child assent. Girls were excluded from the

study on the basis of current pregnancy, as determined by urine

testing. The study population was further reduced because of

missing information on key covariates—namely, hemoglobin

level ( ), socioeconomic status ( ), and nutritionaln p 67 n p 52

status ( )—resulting in an final study population of 507n p 6

children. The institutional review boards of Brown University

and the Philippine Research Institute of Tropical Medicine ap-

proved the study.

Infection intensity. The parasite burden was determined

by examination of 3 stool samples from each study participant.

Each sample was examined in duplicate for the presence of S.

japonicum and 3 soil-transmitted helminths—Trichuris trichi-

ura, Ascaris lumbricoides, and Necator americanus—by the Kato-

Katz method [42, 43]. The number of parasite-specific eggs per

gram of stool (EPG) was used to define infections of low and

by guest on May 14, 2011

jid.oxfordjournals.orgD

ownloaded from

Table 2. Mean hemoglobin levels, according to subprofile classifications and infection categories.

Subprofile classification, infection categoryChildren,

no.

Meanhemoglobinlevel, g/dL

Childrenwith

anemia,% (no.)

Subprofile 0: 0 or 1 L infection (n p 32)0 infections 4 13.5 0Schistosomiasis japonicum 14 12.9 7.1 (1)Trichuris trichiura 10 13.2 0Ascaris lumbricoides 1 12.5 0Necator americanus 3 11.5 0

Subprofile 1: 2 L infections (n p 29)S. japonicum and T. trichiura 9 12.3 11.1 (1)S. japonicum and A. lumbricoides 3 12.7 0S. japonicum and N. americanus 1 13.2 0T. trichiura and A. lumbricoides 8 12.2 37.5 (3)T. trichiura and N. americanus 6 12.8 0A. lumbricoides and N. americanus 2 12.7 0

Subprofile 2: 3 L infections (n p 39)S. japonicum, T. trichiura, and A. lumbricoides 22 11.7 18.2 (4)S. japonicum, T. trichiura, and N. americanus 14 12.2 14.3 (2)S. japonicum, A. lumbricoides, and N. americanus 1 12.7 0A. lumbricoides, T. trichiura, and N. americanus 2 13.4 0

Subprofile 3: 4 L infections (n p 22)A. lumbricoides, T. trichiura, N. americanus, and S. japonicum 22 12.2 22.7 (5)

Subprofile 4: 1 M+ infection (n p 14)S. japonicum 3 12.6 33.3 (1)T. trichiura 8 13.7 0A. lumbricoides 2 12.8 0N. americanus 1 13.1 0

Subprofile 5: 1 M+ and 1 L infections (n p 42)S. japonicum (M+) and T. trichiura (L) 7 11.3 42.9 (3)S. japonicum (M+) and N. americanus (L) 2 13.5 0S. japonicum (M+) and A. lumbricoides (L) 1 11.6 0T. trichiura (M+) and S. japonicum (L) 6 10.7 33.3 (2)T. trichiura (M+) and A. lumbricoides (L) 7 13.1 0T. trichiura (M+) and N. americanus (L) 3 12.9 0A. lumbricoides (M+) and S. japonicum (L) 3 11.5 66.7 (2)A. lumbricoides (M+) and T. trichiura (L) 12 12.8 0A. lumbricoides (M+) and N. americanus (L) 0 … …N. americanus (M+) and S. japonicum (L) 0 … …N. americanus (M+) and T. trichiura (L) 1 13.3 0N. americanus (M+) and A. lumbricoides (L) 0 … …

Subprofile 6: 1 M+ and 2 L infections (n p 62)S. japonicum (M+), A. lumbricoides (L), and T. trichiura (L) 1 12.7 0S. japonicum (M+), A. lumbricoides (L), and N. americanus (L) 0 … …S. japonicum (M+), T. trichiura (L), and N. americanus (L) 7 12.2 28.6 (2)T. trichiura (M+), A. lumbricoides (L), and N. americanus (L) 3 13.0 0T. trichiura (M+), A. lumbricoides (L), and S. japonicum (L) 9 12.2 22.2 (2)T. trichiura (M+), N. americanus (L), and S. japonicum (L) 4 11.2 25.0 (1)A. lumbricoides (M+), T. trichiura (L), and N. americanus (L) 9 12.3 11.1 (1)A. lumbricoides (M+), T. trichiura (L), and S. japonicum (L) 28 12.0 7.1 (2)A. lumbricoides (M+), N. americanus (L), and S. japonicum (L) 1 12.7 0N. americanus (M+), A. lumbricoides (L), and T. trichiura (L) 0 … …N. americanus (M+), A. lumbricoides (L), and S. japonicum (L) 0 … …N. americanus (M+), T. trichiura (L), and S. japonicum (L) 0 … …

(continued)

by guest on May 14, 2011

jid.oxfordjournals.orgD

ownloaded from

Polyparasite Infections and Anemia • JID 2005:192 (15 December) • 2163

Table 2. (Continued.)

Subprofile classification, infection categoryChildren,

no.

Meanhemoglobinlevel, g/dL

Childrenwith

anemia,% (no.)

Subprofile 7: 1 M+ and 3 L infections (n p 56)S. japonicum (M+), T. trichiura (L), A. lumbricoides (L), and N. americanus (L) 8 12.3 25.0 (2)T. trichiura (M+), S. japonicum (L), A. lumbricoides (L), and N. americanus (L) 13 12.0 7.7 (1)A. lumbricoides (M+), T. trichiura (L), N. americanus (L), and S. japonicum (L) 30 11.9 20.0 (6)N. americanus (M+), T. trichiura (L), A. lumbricoides (L), and S. japonicum (L) 5 10.1 60.0 (3)

Subprofile 8: 2 M+ infections (n p 22)S. japonicum and T. trichiura 6 10.6 100.0 (6)S. japonicum and A. lumbricoides 0 … …S. japonicum and N. americanus 0 … …T. trichiura and A. lumbricoides 15 12.3 13.3 (2)T. trichiura and N. americanus 1 13.8 0A. lumbricoides and N. americanus 0 … …

Subprofile 9: 2 M+ and 1 L infections (n p 83)S. japonicum (M+), T. trichiura (M+), and A. lumbricoides (L) 5 10.9 40.0 (2)S. japonicum (M+), T. trichiura (M+), and N. americanus (L) 3 10.4 66.7 (2)S. japonicum (M+), A. lumbricoides (M+), and T. trichiura (L) 10 11.3 30.0 (3)S. japonicum (M+), A. lumbricoides (M+), and N. americanus (L) 1 10.6 100.0 (1)S. japonicum (M+), N. americanus (M+), and A. lumbricoides (L) 0S. japonicum (M+), N. americanus (M+), and T. trichiura (L) 2 8.8 100.0 (2)T. trichiura (M+), A. lumbricoides (M+), and S. japonicum (L) 40 11.6 25.0 (10)T. trichiura (M+), A. lumbricoides (M+), and N. americanus (L) 18 12.0 0T. trichiura (M+), N. americanus (M+), and S. japonicum (L) 3 9.9 66.7 (2)T. trichiura (M+), N. americanus (M+), and A. lumbricoides (L) 0 … …A. lumbricoides (M+), N. americanus (M+), and T. trichiura (L) 0 … …A. lumbricoides (M+), N. americanus (M+), and S. japonicum (L) 1 12.3 0

Subprofile 10: 2 M+ and 2 L infections (n p 93)S. japonicum (M+), T. trichiura (M+), A. lumbricoides (L), and N. americanus (L) 8 12.8 12.5 (1)S. japonicum (M+), N. americanus (M+), T. trichiura (L), and A. lumbricoides (L) 2 11.7 0S. japonicum (M+), A. lumbricoides (M+), T. trichiura (L), and N. americanus (L) 17 11.6 23.5 (4)T. trichiura (M+), N. americanus (M+), A. lumbricoides (L), and S. japonicum (L) 3 12.8 0T. trichiura (M+), A. lumbricoides (M+), N. americanus (L), and S. japonicum (L) 56 12.0 12.5 (7)A. lumbricoides (M+), N. americanus (M+), T. trichiura (L), and S. japonicum (L) 7 12.3 28.6 (2)

Subprofile 11: 3 M+ infections or 3 M+ and 1 L infections (n p 61)S. japonicum (M+), T. trichiura (M+), and N. americanus (M+) 4 8.1 100.0 (4)S. japonicum (M+), N. americanus (M+), and A. lumbricoides (M+) 0 … …S. japonicum (M+), T. trichiura (M+), and A. lumbricoides (M+) 22 11.9 22.7 (5)T. trichiura (M+), A. lumbricoides (M+), and N. americanus (M+) 2 11.5 50.0 (1)S. japonicum (M+), T. trichiura (M+), N. americanus (M+), and A. lumbricoides (L) 2 7.9 100.0 (2)S. japonicum (M+), N. americanus (M+), A. lumbricoides (M+), and T. trichiura (L) 1 10.7 100.0 (1)S. japonicum (M+), T. trichiura (M+), A. lumbricoides (M+), and N. americanus (L) 22 11.5 27.3 (6)A. lumbricoides (M+), T. trichiura (M+), N. americanus (M+), and S. japonicum (L) 8 10.9 50.0 (4)

Subprofile 12: 4 M+ infections (n p 5)S. japonicum, A. lumbricoides, T. trichiura, and N. americanus 5 10.1 40.0 (2)

NOTE. Subprofile categories were based on no. and intensity of infection. Anemia was defined as a hemoglobin level !11 g/dL. L, low intensity; M+, moderate/high intensity.

M+ intensity in accordance with WHO-established intensity

cutoff values for S. japonicum, T. trichiura, and A. lumbricoides

infections [7, 24]. The level of worm burden associated with

morbidity varies across populations and is dependent on the

age, sex, and nutritional iron intake of the person and the

species of hookworm present [10, 44, 45]. It is known that,

on average, a higher EPG for hookworm corresponds with

greater odds of having anemia, and yet the variation in specif-

ic EPG cutoff values used to define infection intensity levels

across studies is consistent with the premise that the critical

EPG for hookworm sufficient to induce anemia is not uniform

across populations [1, 10, 46–52]. Consequently, study-specific

EPG distributions have been used to define hookworm inten-

sity, so that the burden of hookworm associated with notable

by guest on May 14, 2011

jid.oxfordjournals.orgD

ownloaded from

2164 • JID 2005:192 (15 December) • Ezeamama et al.

changes in hemoglobin level in the study population is reflected

[10, 51–55]. In the present study, hookworm infections of low

and M+ intensity were defined as 1–999 EPG and �1000 EPG,

respectively, by use of empirical distributions. The hookworm

eggs were identified by polymerase chain reaction (PCR) as

being those of N. americanus. Details on the EPG measurement

for all parasites have been given elsewhere [41].

Primary determinant: parasite infection profiles. A par-

asite infection profile was assigned to each child on the basis

of the number of species present and the intensity of the in-

fections. Given 3 possible infection intensities (none, low, or

M+) and possible concurrent infection by up to 4 parasites at

1 of 3 intensity levels for each species, there were 81 possible

unique categories of polyparasite infection. The infection cat-

egory conferring the presumed lowest odds of having anemia

was no infections, and the infection category conferring the

presumed highest odds of having anemia was infection with

all 4 species at M+ intensity. A total of 68 of the 81 categories

of polyparasite infection were found in the study population.

To facilitate analyses, the 68 categories were collapsed into

12 subprofiles on the basis of the intensity and number of

concurrent infections and our a priori hypothesis that being

infected with more species at M+ intensity would be associated

with increased odds of having anemia. The subprofiles were

further collapsed into the following 5 infection profiles cor-

responding to putatively different risk levels for having anemia:

1. Reference profile ( ): no infection or infectionn p 32

with 1 parasite species at low intensity;

2. Polyparasite infection profile I ( ): concurrentn p 80

infection with 2, 3, or 4 parasite species at low intensity;

3. Polyparasite infection profile II ( ): infectionn p 152

with 1 parasite species at M+ intensity and all other parasite

species present at low intensity or absent;

4. Polyparasite infection profile III ( ): concurrentn p 179

infection with 2 parasite species at M+ intensity and all other

parasite species present at low intensity or absent;

5. Polyparasite infection profile IV ( ): concurrentn p 63

infection with 3 or 4 parasites at M+ intensity.

Anemia. Complete hemograms were determined on a Se-

rono Baker 9000 hematological analyzer (Serono Baker Diag-

nostics). Anemia was defined as a hemoglobin level !11g/dL [2].

Potential confounding factors. Socioeconomic status (SES),

nutritional status, sex, and age were considered to be poten-

tial confounding factors of the relationship between parasite

infection and anemia. SES was measured through a detailed

questionnaire evaluating 4 domains of SES [41]. In the pres-

ent study, only the summary SES variable, which was cal-

culated using principal components analysis to appropriately

weight questionnaire items as described by Filmer and Pritch-

ett [56], was used. The summary SES score was estimated for

45 (!10%) children with incomplete SES information by as-

signing to them the average summary SES score for children

of the same age and sex in the study population. SES was

divided into low, medium, and high categories by tertiles of

the distribution of the summary SES score. High SES was the

reference group for all analyses.

The z scores for each nutritional index (body mass index

[BMI], weight for age, and height for age) were calculated from

the Centers for Disease Control and Prevention/National Cen-

ter for Health Statistics reference values for children in the

United States in 2000 by use of EpiInfo software (version

2000; Centers for Disease Control and Prevention). To avoid

colinearity, only 1 nutritional status indicator was included

in multivariate models.

Statistical methods. Statistical analyses were performed us-

ing SAS (version 8.0; SAS Institute) and included univariate, bi-

variate, and multivariate analyses using logistic regression mod-

els. Because children from the same household were enrolled

in the present study, and because children within families were

expected to be similar with respect to infection status and he-

moglobin level, all models were specifically adjusted for po-

tential clustering by household [57–59]. Generalized estimating

equation models with exchangeable correlation matrix struc-

tures for dichotomous variables were constructed using SAS

Proc Genmod (version 8.0; SAS Institute). The multivariate

model estimated the odds ratios (ORs) of having anemia for

children with polyparasite infection profiles I–IV relative to that

for children with the reference profile. To show the range of

values consistent with reported OR point estimates, 95% con-

fidence intervals (CIs) were estimated using robust SE estimators.

Both bivariate and adjusted ORs were estimated. Goodness of

fit was assessed for all multivariate models by use of the Hosmer

and Lemeshow goodness-of-fit test [60, 61]. Multivariate mod-

els were deemed to have appropriate fit if the null hypothesis

of appropriate model fit was not rejected at the ap0.05 level

and if (a conservative cutoff value) for the associatedP � .20

Hosmer-Lemeshow test statistic. The contribution that the po-

lyparasite infection profile made in explaining the variability

in anemia was assessed in the multivariate models by use of

the likelihood-ratio test (LRT).

In addition, models were made to assess which parasite spe-

cies conferred the greatest risk of having anemia for each po-

lyparasite infection profile such that adjustment for its intensi-

ty in multivariate models significantly (110% change in effect

size) attenuated or strengthened the association between the

profile and anemia. For each parasite species that was adjust-

ed for, LRTs were conducted for parameter estimates associ-

ated with each intensity level, to determine which type of in-

fection was significantly associated with anemia. All analyses

were also adjusted for the potential confounding factors de-

scribed above.

by guest on May 14, 2011

jid.oxfordjournals.orgD

ownloaded from

Polyparasite Infections and Anemia • JID 2005:192 (15 December) • 2165

Table 3. Sociodemographic, anemia, and nutritional status of children, according to polypar-asite infection profile.

Parameter

Polyparasite infection profilea

Reference I II III IV

Average hemoglobin level, g/dL 12.9 12.2 12.1 11.7 11.2Anemia, % 3.5 17.5 16.9 21.0 34.9Weight-for-age z scoreb

Normal 6 (20.7) 9 (11.3) 28 (18.8) 18 (9.9) 4 (6.3)Mildly underweight 10 (34.5) 21 (26.3) 40 (26.0) 49 (27.1) 10 (15.9)Severely underweight 13 (44.8) 50 (62.5) 86 (55.8) 114 (63.0) 49 (77.8)

Body mass index z scoreNormal 18 (62.1) 46 (57.5) 78 (51.0) 98 (54.1) 19 (30.7)Mildly underweight 9 (31.0) 24 (30.0) 53 (34.6) 58 (32.0) 30 (48.4)Severely underweight 2 (6.9) 10 (12.5) 22 (14.4) 25 (13.8) 13 (21.0)

Age7–12 years 9 (31.0) 40 (50.0) 75 (48.7) 132 (72.9) 42 (66.7)13–15 years 9 (31.0) 28 (35.0) 51 (33.1) 32 (17.7) 13 (20.6)16–18 years 11 (37.9) 12 (15.0) 28 (18.2) 17 (9.4) 8 (12.7)

Male sex 9 (31.9) 50 (62.5) 77 (50.0) 104 (57.5) 49 (77.8)Socioeconomic status

Low 7 (24.1) 20 (25.0) 49 (31.8) 62 (34.3) 24 (38.1)Medium 6 (20.7) 30 (37.5) 47 (30.5) 65 (38.0) 23 (36.5)High 16 (55.2) 30 (37.5) 58 (37.7) 54 (29.8) 16 (25.4)

NOTE. Data are no. (%) of children, unless otherwise indicated.a The reference profile was no infection or infection with 1 parasite species at low intensity; polyparasite infection

profile I was concurrent infection with 2, 3, or 4 parasite species at low intensity; polyparasite infection profile IIwas infection with 1 parasite species at moderate/high intensity and all other parasite species present at lowintensity or absent; polyparasite infection profile III was concurrent infection with 2 parasite species at moderate/high intensity and all other parasite species present at low intensity or absent; and polyparasite infection profileIV was concurrent infection with 3 or 4 parasite species at moderate/high intensity.

b Reference scores were z scores compiled by the Centers for Disease Control and Prevention/National Centerfor Health Statistics for children in the United States in 2000 were used as the reference scores. Healthy, z score1�1; mildly underweight, z score 1�3 to ��1; severely underweight, z score ��3.

RESULTS

Univariate and bivariate results. Table 1 describes the so-

ciodemographic characteristics of the study population. Table

2 describes the unique infection categories and the 12 infection

subprofiles. Table 3 describes the study population with respect

to the final 5 polyparasite infection profiles.

The proportion of children with anemia in the study pop-

ulation was nearly 20%, and malnutrition was common (148%

of children). The prevalence of infections was high and exceed-

ed 50% for each parasite species. S. japonicum and N. ameri-

canus infections occurred most frequently at low intensity. The

majority of A. lumbricoides and T. trichiura infections occurred

at M+ intensity (table 1). There were patterns of decreasing mean

hemoglobin levels, BMIs, and weight-for-age z scores with in-

creasing intensities of infection and with increasing numbers of

concurrent infections (tables 2 and 3). In bivariate analyses (da-

ta not shown), the polyparasite infection profiles that included

more-severe infections with more parasite species were found to

be significantly associated with higher odds of having anemia.

Furthermore, a lower SES, male sex, and S. japonicum or N.

americanus infection at M+ intensity were associated with high-

er odds of having anemia. In addition, inverse associations with

anemia were found for better nutritional status and older age.

Overall pattern of association. After adjustment for SES,

nutritional status, age, and village of residence, a pattern of in-

creasing odds of having anemia was observed with increasing

severity in the polyparasite infection profile (figure 1 and table

4). For children concurrently infected with multiple parasite

species at low intensity, the odds of having anemia were nearly

5 times higher than they were for children with the reference

profile ( ). For children concurrently infected with 1 orP p .05

2 parasites at M+ intensity—that is, those with polyparasite in-

fection profiles II or III—the odds of having anemia were com-

parable to those for children with concurrent multiple low-in-

tensity infections. The highest odds of having anemia were

observed in children concurrently infected with 3 or 4 parasite

species at M+ intensity. Nutritional status, sex, and village of

residence were significantly and independently associated with

having anemia.

Contribution of individual parasite species to observed as-

sociations between polyparasite infection profiles and anemia.

Adjustment for intensities of N. americanus (model D) and S.

by guest on May 14, 2011

jid.oxfordjournals.orgD

ownloaded from

2166 • JID 2005:192 (15 December) • Ezeamama et al.

Figure 1. Prevalent anemia and polyparasite infection profiles. Values indicate the percentage of children with anemia for each polyparasite infection profile.Reference profile, no infection or infection with 1 parasite species at low intensity; polyparasite infection profile I, concurrent infection with 2, 3, or 4 parasitespecies at low intensity; polyparasite infection profile II, infection with 1 parasite species at moderate/high (M+) intensity and all other parasite species presentat low intensity or absent; polyparasite infection profile III, concurrent infection with 2 parasite species at M+ intensity and all other parasite species presentat low intensity or absent; polyparasite infection profile IV, concurrent infection with 3 or 4 parasite species at M+ intensity.

japonicum (model E) infections, in general, resulted in a weak-

ening of the association between the relevant polyparasite in-

fection profile and anemia (table 5). The association between

the polyparasite infection profile and anemia was attenuated

by 14%–22% and 20%–50% after adjustment for intensities of

N. americanus and S. japonicum infections, respectively. How-

ever, increases in the magnitude of the association between the

relevant polyparasite infection profile and anemia occurred with

adjustment for the intensities of A. lumbricoides (model A) and

T. trichiura (model C) infections. Regardless of the specific par-

asite intensity adjusted for, the association between concurrent

multiple low-intensity infections and anemia remained at least

marginally significant in multivariate models. Wald tests of var-

iable main effects in logistic regression models indicated that 4

of the polyparasite infection profiles significantly improved the

regression models’ ability to describe the variance in anemia. For

each polyparasite infection profile, (data not shown).P ! .02

DISCUSSION

In our study population, concurrent polyparasite infections were

associated with ∼5–8 times higher odds of having anemia, af-

ter adjustment for multiple confounding factors and cluster-

ing within household units. Of particular interest, even con-

current multiple low-intensity infections were associated with

higher odds of having anemia relative to those for children with

the reference profile—no infections or 1 low-intensity infection.

The results indicate that, for children with concurrent multiple

low-intensity infections, the odds of having anemia may be

similar to those for children concurrently infected with 1 or 2

parasites at M+ intensity. In addition, we found that S. japon-

icum, N. americanus, and A. lumbricoides infections play a sig-

nificant role in the association between the relevant polyparasite

infection profile and anemia.

Concurrent multiple parasite infections were found to be the

norm in our study population, as was the case in studies pub-

lished elsewhere [27, 29, 30, 62]. We extended the observations

in these studies by constructing profiles of polyparasitism for

the present study on the basis of the intensity of each infection.

Our data demonstrated that concurrent multiple infections at

M+ intensity increased the expected morbidity above that dem-

onstrated for high-intensity single-species infections. Children

infected with �3 parasites at M+ intensity had the highest odds

of having anemia, which underscores the importance of inten-

sity in determining the extent of morbidity in parasite infections

[7, 13, 15–17] and suggests a dose-response relationship be-

tween anemia and cumulative intensities in parasite infections.

The majority of studies suggest that low-intensity single-

species infections have negligible health consequences [18, 19,

24, 54, 63, 64]. The results of the present study challenge this

paradigm for the first time (to our knowledge) and suggest the

need for a reexamination of the consensus view that low-inten-

sity infections have little impact on morbidity. This reexami-

nation is important, because concurrent multiple low-intensi-

ty infections are common in areas in which multiple parasites

are endemic [20–24]. Globally, several million children could be

concurrently infected with multiple helminth species at low in-

tensity [11, 24, 25, 65]. This represents a large population of

children that may not be targeted for intervention, especially

by guest on May 14, 2011

jid.oxfordjournals.orgD

ownloaded from

Polyparasite Infections and Anemia • JID 2005:192 (15 December) • 2167

Table 4. Determinants of anemia from multivariate logistic re-gression model.

Parameter OR (95% CI) P

Polyparasite infection profilea‘I 4.61 (0.98–21.60) .052II 4.09 (0.74–23.90) .106III 4.32 (0.79–22.70) .099IV 7.85 (1.10–45.10) .030

Age 0.93 (0.83–1.05) .268Male vs. female sex 4.30 (2.30–8.20) !.0001Tanner stage (measure of sexual maturity) 0.65 (0.35–1.18) .136Socioeconomic statusb 2.03 (1.08–4.31) .029Nutritional status

Mildly underweight vs. normal 1.28 (0.13–7.03) .782Severely underweight vs. normal 3.21 (1.27–8.12) .013

Village of residenceBuri vs. Pitogo 0.40 (0.17–0.92) .031Macanip vs. Pitogo 2.23 (1.74–3.88) .004

NOTE. All estimates of effect were adjusted for potential clustering byhousehold units in logistic regression analyses. For the Hosmer and Leme-show goodness-of-fit test, (8 df) and . CI, confidence in-2x p 8.55 P p .386terval; OR, odds ratio.

a The reference profile was no infection or infection with 1 parasite speciesat low intensity; polyparasite infection profile I was concurrent infection with2, 3, or 4 parasite species at low intensity; polyparasite infection profile IIwas infection with 1 parasite species at moderate/high (M+) intensity and allother parasite species present at low intensity or absent; polyparasite infectionprofile III was concurrent infection with 2 parasite species at M+ intensityand all other parasite species present at low intensity or absent; and poly-parasite infection profile IV was concurrent infection with 3 or 4 parasitespecies at M+ intensity.

b Socioeconomic status was included as an ordinal covariate with 3 levels(low, medium, and high).

in the resource-limited settings of developing countries. A pub-

lication and research environment that minimizes the impor-

tance of low-intensity infections likely encourages the exclusion

of these children from interventions that may be beneficial to

their health.

We indirectly assessed the relative importance of individ-

ual species in the observed association between the polypara-

site infection profile and anemia by sequentially controlling for

the intensity of infection with each species. As we expected,

adjustment for the intensity of N. americanus or S. japonicum

infection led to significant weakening of the association be-

tween the relevant polyparasite infection profile and anemia,

suggesting that these species may have the largest effect on

anemia. This finding is consistent with those of many other

studies demonstrating a causal relationship between these par-

asites and anemia [10, 13, 65–68]. Conversely, adjustment for

the intensity of A. lumbricoides or T. trichiura infection signif-

icantly strengthened the association between the relevant po-

lyparasite infection profile and anemia. We also found that A.

lumbricoides infection was associated with anemia, but the di-

rection of the effect appeared to be protective, such that the

odds of having anemia decreased significantly with the in-

creasing intensity of the A. lumbricoides infection (data not

shown). We did not find T. trichiura infection to be significantly

associated with anemia, although other researchers [ 7, 69–75]

have noted an association between hemoglobin level and tri-

churis dysentery syndrome, a clinical subgroup not defined in

the present study. We note that, although the finding of sig-

nificant odds of having anemia for children with concurrent

multiple infections suggests the presence of biologically relevant

effect modifications, the analytic method of this study did not

provide information about the nature of any interaction be-

tween parasite species. The type of interaction between the

parasite species and the mechanisms of this interaction is the

subject of a separate manuscript in preparation (A. E. Ezea-

mama, S. T. McGarvey, L. P. Acosta, D. L. Manalo, W. Heiwei,

J. D. Kurtis, V. Mor, R. M. Olveda, and J. F. Friedman).

The present study had some limitations. The true magnitude

of the association between polyparasitism and anemia is likely

to have been underestimated, given how the reference profile

was defined. Children who were completely free of parasite in-

fections were rare in this population. As a result, the reference

group was defined to include children with any low-intensity

single-species infection, which made our estimate of effect more

conservative. Given the design of the study, we cannot exclude

the possibility that anemia predisposes children to concurrent

infections with multiple parasite species (reverse causality).

However, this direction of causality is unlikely, given the evi-

dence from prospective [76, 77] and randomized [18, 40, 78]

clinical trials demonstrating that helminth infections cause

anemia. In addition, despite careful control for measured

confounding factors, it remains possible that the associations

were biased by unmeasured confounding factors or residual

confounding on the basis of measured covariates; however, we

have no reason to believe that any unmeasured confounding

factors would be differentially distributed by polyparasite in-

fection profile and anemia status. Finally, the present study was

conducted in the context of a larger study that specifically

sampled individuals with S. japonicum infection. It is possible

that the sampling scheme of the larger study somewhat limits

the generalizability of our findings to this community, because

our sampling scheme might have overemphasized the role that

S. japonicum plays in the context of polyparasite infection.

In the present study, we examined the relationship between

anemia and the more naturally occurring phenomenon of con-

current multiple low-intensity infections, rather than isolated

analyses of low-intensity single-species infections. We found

that concurrent multiple low-intensity infections conferred an

increased risk of having anemia, suggesting that this com-

mon pattern of infection is not clinically benign and, thus,

should not be ignored. Interventional studies targeting indi-

viduals with low-intensity infections are necessary to inform

or refute our findings. In the meantime, inclusion of children

with low-intensity infections may need to be considered in

by guest on May 14, 2011

jid.oxfordjournals.orgD

ownloaded from

Table 5. Significance of individual species in the association between polyparasite infection profile and anemia in fully adjusted multivariate models.

Category Model A

Model B,Ascaris lumbricoides

intensity adjusted

Model C,Trichuris trichiuraintensity adjusted

Model D,Necator americanusintensity adjusted

Model E,Schistosomiasis

japonicumintensity adjusted

Polyparasite infection profilea

Reference 1.00 1.00 1.00 1.00 1.00

I 4.61 (.052) 6.06 (.018) 5.73 (.029) 4.03 (.087) 3.78 (.068)

II 4.09 (.106) 8.13 (.014) 6.39 (.039) 3.96 (.126) 3.86 (.094)

III 4.32 (.099) 11.71 (.005) 7.63 (.024) 3.82 (.126) 3.05 (.176)

IV 7.85 (.030) 21.60 (.001) 14.18 (.006) 6.26 (.047) 3.71 (.162)

Helminth significance in explaining anemia variance in study population, x2, 1 df (P)b

Low-intensity infection NA 2.23 (.135) 0.84 (.359) 2.81 (.094) 6.44 (.011)

Moderate/high-intensity infection NA 9.64 (.002) 3.58 (.059) 9.62 (.002) 10.54 (.001)

Overall model fit statistic, x2, 8 df (P) c 8.54 (.383) 5.40 (.713) 8.06 (.428) 6.33 (.607) 8.28 (.407)

NOTE. Data are odds ratios (P), unless otherwise indicated. Model A was adjusted for sex, nutritional status, village, socioeconomic status, and age. Models B, C, D, and E were adjusted for all variables in model Ain addition to the shown helminth species. All models are clustered by household units. NA, not applicable.

a The reference profile was no infection or infection with 1 parasite species at low intensity; polyparasite infection profile I was concurrent infection with 2, 3, or 4 parasite species at low intensity; polyparasite infectionprofile II was infection with 1 parasite species at moderate/high (M+) intensity and all other parasite species present at low intensity or absent; polyparasite infection profile III was concurrent infection with 2 parasitespecies at M+ intensity and all other parasite species present at low intensity or absent; and polyparasite infection profile IV was concurrent infection with 3 or 4 parasite species at M+ intensity.

b Likelihood ratio tests were used to test the null hypothesis that the specified intensity level had no effect on anemia. The hypothesis was rejected when .P ! .05c Hosmer and Lemeshow goodness-of-fit test.

by guest on May 14, 2011 jid.oxfordjournals.org Downloaded from

Polyparasite Infections and Anemia • JID 2005:192 (15 December) • 2169

helminth-control initiatives in the many regions of the devel-

oping world where polyparasitism is the norm.

Acknowledgments

We thank the study participants from Macanip, Buri, and Pitogo inLeyte, The Philippines, for making this study possible; Blanca Jarilla, MarioJiz, Archie Pablo, Raquel Pacheco, Patrick Sebial, Mary Paz Urbina, andJemaima Yu (the field staff), for their diligence and energy; and Dr. JosephHogan (Department of Biostatistics and Community Health, Brown Uni-versity, Providence, Rhode Island), for his advice.

References

1. Stoltzfus RJ, Chwaya HM, Tielsch JM, Schulze KJ, Albonico M, SavioliL. Epidemiology of iron deficiency anemia in Zanzibari schoolchildren:the importance of hookworms. Am J Clin Nutr 1997; 65:153–9.

2. Zetterstrom R. Iron deficiency and iron deficiency anaemia during in-fancy and childhood. Acta Paediatr 2004; 93:436–9.

3. Viteri FE. The consequences of iron deficiency and anaemia in preg-nancy on maternal health, the foetus and the infant. SCN News 1994;11:14–8.

4. van den Broek N. Anaemia and micronutrient deficiencies. Br MedBull 2003; 67:149–60.

5. Stephenson LS, Latham MC, Ottesen EA. Malnutrition and parasitichelminth infections. Parasitology 2000; 121(Suppl):S23–38.

6. Traore M, Traore H, Kardorff R, et al. The public health significanceof urinary schistosomiasis as a cause of morbidity in two districts inMali. Am J Trop Med Hyg 1998; 59:407–13.

7. Stephenson LS, Holland CV, Cooper ES. The public health significanceof Trichuris trichiura. Parasitology 2000; 121(Suppl):S73–95.

8. Shubair ME, Yassin MM, al-Hindi AI, al-Wahaidi AA, Jadallah SY, AbuShaaban Na-D. Intestinal parasites in relation to haemoglobin leveland nutritional status of school children in Gaza. J Egypt Soc Parasitol2000; 30:365–75.

9. Robertson LJ, Crompton DWT, Sanjur D, Nesheim MC. Haemoglobinconcentrations and concomitant infections of hookworm and Trichuristrichiura in Panamanian primary schoolchildren. Trans R Soc TropMed Hyg 1992; 86:654–6.

10. Olsen A, Magnussen P, Ouma JH, Andreassen J, Friis H. The contri-bution of hookworm and other parasitic infections to haemoglobinand iron status among children and adults in western Kenya. Trans RSoc Trop Med Hyg 1998; 92:643–9.

11. Public health impact of schistosomiasis: disease and mortality. WHOExpert Committee on the Control of Schistosomiasis. Bull World HealthOrgan 1993; 71:657–62.

12. Brooker S, Peshu N, Warn PA, et al. The epidemiology of hookworminfection and its contribution to anaemia among pre-school childrenon the Kenyan coast. Trans R Soc Trop Med Hyg 1999; 93:240–6.

13. Crompton DW, Whitehead RR. Hookworm infections and human ironmetabolism. Parasitology 1993; 107(Suppl):S137–45.

14. Gilgen DD, Mascie-Taylor CG, Rosetta LL. Intestinal helminth infec-tions, anaemia and labour productivity of female tea pluckers in Ban-gladesh. Trop Med Int Health 2001; 6:449–57.

15. Stephenson LS. Helminth parasites, a major factor in malnutrition.World Health Forum 1994; 15:169–72.

16. Stephenson LS. Optimising the benefits of anthelmintic treatment inchildren. Paediatr Drugs 2001; 3:495–508.

17. Stoltzfus RJ, Chway HM, Montresor A, et al. Low dose daily ironsupplementation improves iron status and appetite but not anemia,whereas quarterly anthelminthic treatment improves growth, appetiteand anemia in Zanzibari preschool children. J Nutr 2004; 134:348–56.

18. Despommier DD, Gwadz RW, Hotez PJ, Knirsch CA. Parasitic diseases.4th ed. New York: Apple Tree Productions, 2000.

19. Hinz E. Human helminthiasis in the Philippines: the epidemiologicaland geomedical situation. Berlin: Springer-Verlag, 1984.

20. Savioli L, Stansfield S, Bundy DAP, et al. Schistosomiasis and soil-transmitted helminth infections: forging control efforts. Trans R SocTrop Med Hyg 2002; 96:577–9.

21. Jinabhai CC, Taylor M, Coutsoudis A, Coovadia HM, Tomkins AM,Sullivan KR. Epidemiology of helminth infections: implications forparasite control programmes, a South African perspective. Public HealthNutr 2001; 4:1211–9.

22. Urbani C, Palmer K. Drug-based helminth control in western Pacificcountries: a general perspective. Trop Med Int Health 2001; 6:935–44.

23. Gyorkos TW. Monitoring and evaluation of large scale helminth controlprogrammes. Acta Trop 2003; 86:275–82.

24. Prevention and control of schistosomiasis and soil-transmitted hel-minthiasis: report of a WHO expert committee. World Health OrganTech Rep Ser 2002; 912:i–vi, 1–57, back cover.

25. WHO. Communicable diseases: control of schistosomiasis and soiltransmitted helminth infections. Report by the Secretariat A54/10. Ge-neva: WHO, 2001.

26. Ashford RW, Craig PS, Oppenheimer SJ. Polyparasitism on the Kenyacoast. 2. Spatial heterogeneity in parasite distributions. Ann Trop MedParasitol 1993; 87:283–93.

27. Buck AA, Anderson RI, MacRae AA. Epidemiology of poly-parasitism.I. Occurrence, frequency and distribution of multiple infections in ruralcommunities in Chad, Peru, Afghanistan, and Zaire. Tropenmed Par-asitol 1978; 29:61–70.

28. Chunge RN, Karumba N, Ouma JH, Thiongo FW, Sturrock RF, But-terworth AE. Polyparasitism in two rural communities with endemicSchistosoma mansoni infection in Machakos District, Kenya. J TropMed Hyg 1995; 98:440–4.

29. Keiser J, N’Goran EK, Traore M, et al. Polyparasitism with Schistosomamansoni, geohelminths, and intestinal protozoa in rural Cote d’Ivoire.J Parasitol 2002; 88:461–6.

30. Tchuem Tchuente LA, Behnke JM, Gilbert FS, Southgate VR, Ver-cruysse J. Polyparasitism with Schistosoma haematobium and soil-trans-mitted helminth infections among school children in Loum, Camer-oon. Trop Med Int Health 2003; 8:975–86.

31. Thiong’o FW, Luoba A, Ouma JH. Intestinal helminths and schisto-somiasis among school children in a rural district in Kenya. East AfrMed J 2001; 78:279–82.

32. Rietveld E, Vetter JC, Stilma JS. Concurrent parasitic infections amongpatients with onchocerciasis and controls in Sierra Leone, West Africa.Doc Ophthalmol 1987; 67:25–32.

33. Raso G, Luginbuhl A, Adjoua CA, et al. Multiple parasite infectionsand their relationship to self-reported morbidity in a community ofrural Cote d’Ivoire. Int J Epidemiol 2004; 33:1092–102.

34. Tshikuka JG, Gray-Donald K, Scott M, Olela KN. Relationship of child-hood protein-energy malnutrition and parasite infections in an urbanAfrican setting. Trop Med Int Health 1997; 2:374–82.

35. McGarvey ST, Wu G, Zhang S, et al. Child growth, nutritional status,and Schistosomiasis japonica in Jiangxi, People’s Republic of China. AmJ Trop Med Hyg 1993; 48:547–53.

36. McGarvey ST, Aligui G, Daniel BL, Peters P, Olveda R, Olds GR. Childgrowth and Schistosomiasis japonica in northeastern Leyte, the Phil-ippines: cross-sectional results. Am J Trop Med Hyg 1992; 46:571–81.

37. al-Agha R, Teodorescu I. Intestinal parasites infestation and anemia inprimary school children in Gaza governorates—Palestine. Roum ArchMicrobiol Immunol 2000; 59:131–43.

38. Olveda RM, Daniel BL, Ramirez BD, et al. Schistosomiasis japonica inthe Philippines: the long-term impact of population-based chemo-therapy on infection, transmission, and morbidity. J Infect Dis 1996;174:163–72.

39. Olveda RM, Tiu E, Fevidal P Jr, de Veyra F Jr, Icatlo FC Jr, DomingoEO. Relationship of prevalence and intensity of infection to morbidityin Schistosomiasis japonica: a study of three communities in Leyte,Philippines. Am J Trop Med Hyg 1983; 32:1312–21.

40. McGarvey ST, Aligui G, Graham KK, Peters P, Olds GR, Olveda R.

by guest on May 14, 2011

jid.oxfordjournals.orgD

ownloaded from

2170 • JID 2005:192 (15 December) • Ezeamama et al.

Schistosomiasis japonica and childhood nutritional status in north-eastern Leyte, the Philippines: a randomized trial of praziquantel versusplacebo. Am J Trop Med Hyg 1996; 54:498–502.

41. Ezeamama AE, Friedman JF, Acosta LP, et al. Helminth infection andcognitive impairment among Filipino children. Am J Trop Med Hyg2005; 72:540–8.

42. Katz N, Chaves A, Pellegrino J. A simple device for quantitative stoolthick-smear technique in Schistosomiasis mansoni. Rev Inst Med TropSao Paulo 1972; 14:397–400.

43. Katz N, Zicker F, Chaves A, Antunes CM. Clinical trials with pyrantelpamoate in intestinal parasitoses. Rev Inst Med Trop Sao Paulo 1972;14:212–21.

44. Prevention and control of intestinal parasitic infections: report of aWHO expert committee. World Health Organ Tech Rep Ser 1987; 749:1–88.

45. Montresor A, Crompton DWT, Hall A, Bundy DAP, Savioli L. Guide-lines for the evaluation of soil transmitted helminthiasis and schisto-somiasis at the community level. Geneva: WHO, 1998.

46. Shield JM, Vaterlaws AL, Kimber RJ, et al. The relationship of hook-worm infection, anaemia and iron status in a Papua New Guinea high-land population and the response to treatment with iron and mebenda-zole. P N G Med J 1981; 24:19–34.

47. Latham MC, Stephenson LS, Hall A, Wolgemuth JC, Elliot TC, Cromp-ton DWT. Parasitic infections, anaemia and nutritional status: a studyof their interrelationships and the effect of prophylaxis and treatmenton workers in Kwale District, Kenya. Trans R Soc Trop Med Hyg 1983;77:41–8.

48. Hercberg S, Chauliac M, Galan P, et al. Relationship between anae-mia, iron and folacin deficiency, haemoglobinopathies and parasiticinfection. Hum Nutr Clin Nutr 1986; 40:371–9.

49. Pritchard DI, Quinnell RJ, Moustafa M, et al. Hookworm (Necatoramericanus) infection and storage iron depletion. Trans R Soc Trop MedHyg 1991; 85:235–8.

50. Allemann A, Bauerfeind P, Gyr N. Prevalence of hookworm infection,anaemia and faecal blood loss among the Yupno people of Papua NewGuinea. P N G Med J 1994; 37:15–22.

51. Stephenson LS, Latham MC, Kurz KM, Kinoti SN. Single dose me-trifonate or praziquantel treatment in Kenyan children. II. Effects ongrowth in relation to Schistosoma haematobium and hookworm eggcounts. Am J Trop Med Hyg 1989; 41:445–53.

52. Stephenson LS, Kinoti SN, Latham MC, Kurz KM, Kyobe J. Singledose metrifonate or praziquantel treatment in Kenyan children. I. Ef-fects on Schistosoma haematobium, hookworm, hemoglobin levels, sple-nomegaly, and hepatomegaly. Am J Trop Med Hyg 1989; 41:436–44.

53. WHO. Field studies and the relation between intestinal infections andhuman nutrition. NUT/81.3 PDP/82.4. Geneva: WHO, 1981.

54. Beaver PC. Biology of soil-transmitted helminths: the massive infection.Health Lab Sci 1975; 12:116–25.

55. Stephenson LS, Latham MC, Kurz KM, Kinoti SN, Oduori ML, Cromp-ton DW. Relationships of Schistosoma hematobium, hookworm andmalarial infections and metrifonate treatment to hemoglobin level inKenyan school children. Am J Trop Med Hyg 1985; 34:519–28.

56. Filmer D, Pritchett LH. Estimating wealth effects without expendituredata—or tears: an application to educational enrollments in states ofIndia. Demography 2001; 38:115–32.

57. Forrester JE, Scott ME, Bundy DA, Golden MH. Clustering of Ascaris

lumbricoides and Trichuris trichiura infections within households. TransR Soc Trop Med Hyg 1988; 82:282–8.

58. Drake LJ, Bundy DA. Multiple helminth infections in children: impactand control. Parasitology 2001; 122(Suppl):S73–81.

59. Shapiro AE, Tukahebwa EM, Kasten J, et al. Epidemiology of helminthinfections and their relationship to clinical malaria in southwestUganda. Trans R Soc Trop Med Hyg 2005; 99:18–24.

60. Hosmer DW, Hosmer T, Le Cessie S, Lemeshow S. A comparison ofgoodness-of-fit tests for the logistic regression model. Stat Med 1997;16:965–80.

61. Lemeshow S, Hosmer DW Jr. A review of goodness of fit statistics foruse in the development of logistic regression models. Am J Epidemiol1982; 115:92–106.

62. Howard SC, Donnelly CA, Kabatereine NB, Ratard RC, Brooker S.Spatial and intensity-dependent variations in associations betweenmultiple species helminth infections. Acta Trop 2002; 83:141–9.

63. WHO. Helminth control in school-age children: a guide for managersof control programmes. Geneva: WHO, 2002.

64. WHO. The new worm control strategy. Geneva: WHO, 2003.65. Utzinger J, Keiser J. Schistosomiasis and soil-transmitted helminthiasis:

common drugs for treatment and control. Expert Opin Pharmacother2004; 5:263–85.

66. Hawdon JM, Hotez PJ. Hookworm: developmental biology of the in-fectious process. Curr Opin Genet Dev 1996; 6:618–23.

67. Brooker S, Bethony J, Hotez PJ. Human hookworm infection in the21st century. Adv Parasitol 2004; 58:197–288.

68. Hotez PJ, Brooker S, Bethony JM, Bottazzi ME, Loukas A, Xiao S.Hookworm infection. N Engl J Med 2004; 351:799–807.

69. Ramdath DD, Simeon DT, Wong MS, Grantham-McGregor SM. Ironstatus of schoolchildren with varying intensities of Trichuris trichiurainfection. Parasitology 1995; 110:347–51.

70. Diniz-Santos DR, Jambeiro J, Mascarenhas RR, Silva LR. Massive Tri-churis trichiura infection as a cause of chronic bloody diarrhea in achild. J Trop Pediatr (in press).

71. Noorizan AM, Mahendra RS. Trichuris dysentery syndrome: evidencethat it may be underdiagnosed in Kelantan. Med J Malaysia 2001; 56:53–7.

72. Callender JE, Grantham-McGregor SM, Walker SP, Cooper ES. Treat-ment effects in Trichuris dysentery syndrome. Acta Paediatr 1994; 83:1182–7.

73. Layrisse M, Aparcedo L, Martinez-Torres C, Roche M. Blood loss dueto infection with Trichuris trichiura. Am J Trop Med Hyg 1967; 16:613–9.

74. Kamath KR. Severe infection with Trichuris trichiura in Malaysian chil-dren: a clinical study of 30 cases treated with stilbazium iodine. AmJ Trop Med Hyg 1973; 22:600–5.

75. Lotero H, Tripathy K, Bolanos O. Gastrointestinal blood loss in Tri-churis infection. Am J Trop Med Hyg 1974; 23:1203–4.

76. Beasley NM, Tomkins AM, Hall A, et al. The impact of populationlevel deworming on the haemoglobin levels of schoolchildren in Tanga,Tanzania. Trop Med Int Health 1999; 4:744–50.

77. de Silva NR. Impact of mass chemotherapy on the morbidity due tosoil-transmitted nematodes. Acta Trop 2003; 86:197–214.

78. Jinabhai CC, Taylor M, Coutsoudis A, Coovadia HM, Tomkins AM,Sullivan KR. A randomized controlled trial of the effect of antihel-minthic treatment and micronutrient fortification on health status andschool performance of rural primary school children. Ann Trop Pae-diatr 2001; 21:319–33.

by guest on May 14, 2011

jid.oxfordjournals.orgD

ownloaded from