Int. J. Epidemiol. 1998 Albonjco 530 7

8/22/2019 Int. J. Epidemiol. 1998 Albonjco 530 7

1/8

C Internati ona l Epidemlological Association 1998 Printed in Great Britain International Journal of Epidemiology 1998 , 27 :530 - 537

Epidemiological evidence for a differentialeffect of hookworm species, Ancylostomaduodenale or Necator americanus, on ironstatus of childrenM Albonico,a RJ Stoltzfus,b L Savioli,c JM Tielsch,d HM Chwaya,e E Ercole f and G CancriniS

Background The hookworms, Ancylostoma duodenak an d Necator americanus, cause significantgastrointestinal blood loss. In clinical studies, greater blood losses have beenreported with A. duodenak. However, there has been no evidence that endemicA. duodenak infection has greater impact than N. americanus infection on the ironstatus of populations.

Methods In a sample of 525 school children in Pemba Island, Tanzania, we compared thedegree of anaemia and iron deficiency associated with the two hookworm speciesat the individual and community (i.e. school) levels. Multiple regression wasused to control for infection intensities and other child characteristics.

Results In the 492 children with hookworm positive faecal cultures, haemoglobin andferritin concentrations decreased with increasing proportions of A. duodenak.Among children with only N. americanus larvae, the prevalence of anaemia was60.5% and the prevalence of ferritin


2/8

HOOKWORM SPECIES AND IRON STATUS 5 3 1

From a morb idity standpoint, the most imp ortan t differencebetween the two species is that A. duodenale causes approx-imately five times greater blood loss than N. americanus.2 Thishas been demonstrated in clinical studies in which isotopicallylabelled red blood cells were injected into hookworm-infectedindividuals, and the faecal excretion of the label was measuredbefore and after deworming. Estimates from individual studieshave found 2- to 10-fold higher blood losses with A. duodenalethan with N. americanus.4

However, there has no t been evidence that endamic A. duodenaleinfection has greater impact than N. americanus infection on theiron status of populations. The population biology of hook-worm s, the host-parasite interaction ,10 and iron absorption inh u m a n s 1 1 may adapt in ways that prevent the burden of hook-worm anaemia from becoming overly great in stable humanpopula t ions . 1 2 Such compensatory strategies may preventspecies related differences in blood loss per hookworm frommanifesting into differences in the prevalence and intensity ofiron-deficiency anaemia in human populations.Because of the difficulty of measuring actual numbers of

worms in population studies, most studies of hookworm in-fection and iron-deficiency anaemia in communities have usedfaecal egg counts to determine the presence and severity ofhookworm infection. By this method, the two hookworm spe-cies cannot be distinguished, because their eggs have similarmicroscopic appearance.2 Thus, there has been no epidemio-logicaJ evidence that the species of hookworm being transmittedin a community has a meaningful impart on the iron status ofthe population.We have studied the epidemiology of hookworm infectionand iron-deficiency anaemia in schoolchildren in Pemba Island,Zanzibar.13 Hookworms are highly endemic on the island, 14and both species were believed to be present. 15 In the course ofan ongoing evaluation of the impart of school-based deworm-

ing on children's nutritional status, we determined the speciesof hookworm infection in a sample of 525 children from 10randomly selected communities. Infection intensities and ironstatus were also assessed. Our objectives were: (1) to describethe distribution of the two hookworm species across the island,(2) to determine whether children infected with A. duodenalehave worse iron status than those infected with N. americanus,and (3) to demonstrate w hether comm unities where A. duodenaleis highly endemic have greater burdens of iron-deficiencyanaemia than communities primarily infected with N. americanus,controlling for other factors.

MethodsStudy area and sam plePemba Island, the smaller of the two islands of Zanzibar, UnitedRepublic of Tanzania, lies 40 km off the East African coast, 5Sof the equator and is classed as a continental island with humidtropical climate. The northeast monsoon blows from October toDecember and is characterized by short rains. The southwestmonsoon brings the long rains from mid-March to mid-May.The temperatures rarely exceed 34 or fall below 21C.

Pemba is densely populated with 330 inhabitants per km 2 .The island is divided into four districts of roughly equal size.About 15 000 peo ple live in each of the thre e main tow ns ofWete, Chake-Chake and Mkoani, the rest of the population

being scattered in numerous villages. The economy of theinhabitants is mainly based on farming and fishing. The primarycrops consist of subsistence agriculture, with some clove andcoconut plantations.The sample for the present study was selected from childrenincluded in the baseline evaluation survey for the Zanzibar

school-based deworming programme. For the evaluation, 12schools were randomly selected stratified by district, and 3605child ren in m orn ing classes of grad es 14 in tho se scho olsparticipated in the survey.13 The sample for the present studyincluded 570 children, drawn as a systematic random sub-sam ple. Specifically, every sev ent h child surve yed from 10 ofthe 12 schools was included. Children from th e first school werenot included because the method for culturing hookwormlarvae was not yet in place, and children from the last schoolwere not included because we had already exceeded ourintended sample size of 500 children.The study was approved by the internal review boards of TheJohns Hopkins University, the Ministry of Health of Zanzibar,and the World Health Organization.

AssessmentsSurvey data were collected in the schools by specially trainedstaff of the Ministry of Health of Zanzibar. On the day before thesurvey, plastic containers were distributed to each class and thechildren were asked to bring a stool sample to school the nextday. Over 95% of children returned a faecal sample. Faecalhookworm egg counts were determined by the Kato-Katz tech-n ique . 1 6 To determine hookworm species, cultures were suc-cessfully obtained for 525 of the 570 children selected for thestudy. The Harada-Mori techn ique 17 modified according to Garcia& Bruckne r18 was used in order to increa se the efficiency of theculture method. Approximately 1 g faeces was m ixed with activegranular charcoal, and some water was added in the case ofhard stools, to obtain an homogeneous paste. Two microscopeslides were folded in a piece of absorbent paper and placed in aPetri dish with a diameter of 9 cm. The stool specimens weresmeared carefully on the surface of the absorbent paper and5-6 ml of distilled water were added to the Petri dish, withoutreaching the stool level. The cultures were covered and labelledand left at room temperature (26-30C) in a shady place. Thepositivity of the cultures was checked with a disserting micro-scope and after 7-10 days the cultures were transferred to a testtube. The larvae were killed by heating (100C) in water for2-3 min and then fixed with 10% formalin. The sediment wasmeasured and a representative sample (usually 1/3 or 1/4) wasexamined for larvae that were identified and counted. Thecounts were then multiplied by 3 or 4 according to the pro-portion of sediment examined.

Urine samples were collected from 100% of children, andmicrohaematuria was diagnosed using Hemastix test strips(Ames Laboratories, Inc., Elkhart, IN). In this population micro-haem aturia screens for urinary schistosomiasis (Schistosoma haem-atobium infection) with 69% sensitivity and 89% specificity. 19

Blood samples were collected by venipuncture from allchildren to assess iron status. Haemoglobin was determined bythe HemoCue method (HemoCue AB, Angelhom, Sweden) anderythrocyte protoporphyrin was determined using a front-facehaem atoflu orom eter (Aviv Biomedical Inc., Lakewo od, NJ). Athin blood film was stained with Giemsa and malaria parasites


3/8

5 3 2 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

were quantitated against leukocytes using standard methods.The remaining blood was allowed to clot at room temperature,and serum was collected by centrifugation. Sera were stored at10C for up to 8 weeks, transported to Baltimore in liquidnitrogen and stored at -70C for up to 4 months. Serum ferritinwas determined on 442 samples (84% of study children) usinga fluorescence-linked immunoassay (DELFIA System by Wallac,Inc., Gaithersburg, MD). Missing data for serum ferritinwere mainly due to mislabelling of specimens, and losses intransport.Statistical analysisDistributions of faecal egg counts, serum ferritin and erythro-cyte protoporphyrin were skewed to high values, so geometricmeans are presented. Anaemia was defined as haemoglobin120 (imol/mol heme were considered indicativeof severe iron-deficient erythropoiesis. Serum ferritin 50 % A. duodenale larvae. The lattergroup includes 11 children w ith apparen tly pure A. duodenaleinfections (i.e. no N americanus larvae were found).Necator americanus was highly prevalent in all communitiessurveyed, but the prevalence of A. duodenale ranged from

2% to 88% (Figure 1). To examine the effect of species on com-munity burden of iron-deficiency anaemia, children in fourschools with >20% prevalence of A. duodenale were comparedto children in six schools with 5 0 % A. duodenale. In the community level analyses, the hook-worm species variable represented low prevalence versus highprevalence schools. The models included all child characteristicsthat differed between hookworm species groups, or that weknew to be associated with iron status from previous ana-lyses .13 Adjusted mean values and standard deviations for ironstatus measures were calculated by analysis of variance.Whether and how to control for the intensity of hookworminfection presented a dilemma. Because the blood loss causedby hookw orm infection is directly proportional to the n um berof adult worms in the intestine, differences in intensity (i.e.worm burdens) of A. duodenale an d N. americanus infections couldconfound the comparison of iron status between individuals orcommunities infected with the two species. We measuredintensity of infection by faecal egg counts, but these cannot bedirectly compared betw een the two species because A. duodenaleproduces 2-3 times more eggs per adult worm than doesN americanus. Furthermore, there was not a simple way to

100rs: 80

60ra 40a- 20

1 .ill i|

E3cof

Cracw

N. americanusA. duodenale

School

Figure 1 Prevalence ol N. americanus an d A duodenale infections inchildren from 10 schools on Pemba Island, Tanzania. Number ofchildren surveyed in each school ranges from 35 to 60

correct the faecal egg counts from the Kato-Katz method toreflect the relative worm burdens of the two species.2Because a given number of eggs/g faeces represents fewer

A. duodenale worms than N. americanus worms, controlling forhookworm eggs/g faeces might bias against finding a truespecies-related difference in iron status. However, it protectsus from attributing to species a difference in iron status thatwas truly caused by greater worm burdens. We carried out ana-lyses with and without adjustment for hookworm eggs/g faeces,and found it changed the results very little. We present resultsof analyses with adjustment for hookworm eggs/g faeces, aconservative a pproach. To adjust for hoo kw orm eggs/g faeces,an ordinal categorical variable was included in the multivariateregression models, coded as 0 if 0 eggs/g, 1 if 1-1999 eggs/g,2 if 2000-3999 eggs/g, and 3 if 4000-5999 eggs/g, and 4 if s*6000eggs/g faeces. The multiple R 2 for the models were highestwhen we used the variable in this form, rather than geometricmean eggs/g or other categorical forms. Data were analysedusing Systat statistical software (SPSS, Inc., Chicago, IL).

ResultsPrevalence and geographical distribu tionof hookworm speciesBased on the faecal cultures, 33 children (6.3%) were notinfected with hookworms, 11 (2.1%) were infected with onlyA. duodenale, 340 (64.8%) were infected with only N. americanus,and 141 (26.9%) were infected with both species of hookworm(Table 1). The prevalence of hook wor m infection based onfaecal cultures was 93.7%, and based on Kato-Katz method was9 6 . 2 % . The overall agreement for hookworm positivitybetween the faecal culture and Kato-Katz methods was 91.4%(480/525). Only four of the 525 children studied, less than 1%of the sample, were hookworm negative by both methods.


4/8

HOOKW ORM SPECIES AND IRON STATUS 5 3 3

Table 1 Number of faecal samples diagnosed with hookworm infeaion by Kato-Katz method and faecal culture method

a Column per cents.b Row per cents.

Kato-Katz resultNegativePositiveAll

Faecal culture resultNegative

429

33 (6.3%)b

Positive forA . duodenal! only

38

11 (2.1%)

N americanus only11

32 9340 (64.8%)

A . duodenale + N . americanus2

13 9141 (26.9%)

A ll20 (3.8%)a

505 (96.2%)52 5

The prevalence of N. americanus infection was uniformly highin all 10 schools surveyed, but the prevalence of A. duodenalevaried greatly, from 2% to 88% (Figure 1). Four schools (Mich-enzani, Tumbe, Msuka, and Mgogoni) were considered highlyendemic for A. duodenale with prevalences of 42-88%, while theremaining six schools had prevalences below 18% . The fourschools with high prevalence of A. duodenale were locatedat opposite ends of the island, and did not share any obviousgeographical characteristics (Figure 2). Because A. duodenale

H O O K W O R M S P E C IE SI D E N T I F IC A T I O N S U R V E YPemba, U nited Republic o( Tanzania

K W A L E M A p dVrtSHA 1t-13ApdiCHOKOCHO 1* -! 8Ao riNffOUBENI- l -21Afr tMKHBCAN1 O-UWlTIDJKUSUKAMSOGOMnZDJBANISNSJUU 24-M* Etaokutog of rainy a n o n

Figure 2 Map of Pemba Island, Tanzania, showing geographic locationof the 10 survey schools and the dates of the survey in each school

transmission may exhibit seasonality, the dates of the survey ineach school are included in Figure 2. The heavy rains began asthe survey team moved from Chokocho to Ngombeni school,and continued for the duration of the study. The four schoolswith high prevalence of A. duodenale were surveyed during theheavy rains, but so were Ngombeni, Kizimbani and Shengejuuwhich had low prevalences.

Hookworm species and individuals' iron statusThe characteristics of children with only N. americanus infectiondid not differ significantly from those infected with A. duodenale(Table 2). The children were on average 10 years old, and abouthalf were girls. About one-fourth of children had microhaem-aturia, an indicator of urinary schistosomiasis, and aro und 60 %were positive for Plasmodium faldparum. Trichuris trichiura andAscaris lumbricoides infections were common. The prevalenceand int ens ity of these infections did not differ significantlybetween hookworm species groups.

The prevalence and intensity of hookworm infection, assessedby the Kato-Katz method, did not differ significantly betweenthe hoo kw orm species groups, but were significantly low erin children with negative faecal cultures (Table 2). The distribu-tion of faecal egg counts assessed by the Kato-Katz method wassimilar between the hookworm species groups (Figure 3).

By all indicators, the iron status of children infected withA. duodenale was worse than those infected with only N. americanus,and haemoglobin and serum ferrit in values became progres-sively worse according to the percentage of A. duodenale larvaefound (Table 3). Overall there is a consistent p attern of wors eiron status in children with more A. duodenale infection. The dif-ferences between groups was most significant for serum ferritin,the indicator most specific to iron deficiency.Hookworm species and community burdenof iron deficiency and anaemiaCompared to children in schools with low prevalence ofA. duodenale, children in the four schools with high A duodenaleprevalence were slightly older and had a higher prevalence ofmicrohaematuria (Table 4). The prevalence and intensity ofT. trichiura or A. lumbricoides infeaion did not differ betweenchildren in the two groups of schools. However, the prevalenceof hookworm infection and hookworm faecal egg countswere significantly higher in children in the schools with highA. duodenale prevalence. The difference in the h ookw orm faecalegg counts was apparent both in the mean values (Table 4) andin the distribution of values (Figure 4).

After adjusting for these differing characteristics between thegroups, children in the high prevalence schools had significantly


5/8

5 3 4 I N T E R N A T I O N A L J O U R N A L OF E P I D E M I O L O G Y

Table 2 Characteristics of study children by species of hookworm infeaion determined by faecal cultures

Characterist ic Ne gat ive c ul tur e(n = 33)N. americanuslarvae only(n = J40)

A . duodenale*50% of larvae(n = 39)

Age (years)Se x (% girls)Microhaematuna cMalaria parasitaemiaT. trichuria

prevalenceeggs/g faeces

A lumbricoidesprevalenceeggs/g faeces

Hookworms^prevalenceeggs/g faeces

10.4 1.5b42.4%24.2%60.6%

93.9%525 (73, 3763) d

72.7%23 9 (6, 9072)

87.9%25 8 (44, 1499)

10.5 1.951.0%23.4%63.3%

97.0%63 3 (117, 3439)

71.5%24 0 (6, 9691)

96.8%608 (102, 3607)

11.0 2.145.9%24.3%60.7%

9 6 . 4 %577 (108, 1856)

71.2%2 33 (6, 9576)

9 8 . 2 %8 4 4 (188, 3790)

10.5 2.648.7%23.1%54.1%

100.0%602 (193, 1882)

76.9%326 (10, 10 290)

92.3%420 (56, 3168)

a Includes 11 samples In which only A duodenak larvae were found, and 141 samples Inwhich both spedes were found.b Mean SDc An Indication of urinary schlstosomiasis.d Geometric mean (-1 SD, +1 SD)c Prevalence and intensity of hookworm Infeaion by Kato-Katz method

D Culture negativeH N. americanus only < 5 0 % A. duodenale > 5 0 % A. duodenale

90 70 60g 50~ 40 30o 20

00 1-1999 2000-39 99 4000-5999 6000-7 999 8000+Hookworm eggs/g faeces

Kato-Katz methodFigure 3 Distributions of hookwo rm faecal egg counts by hookwo rm spedes category determined by larval culture. Numbers of children inspecies categories, culture negative, 33; N americanus only, 340;


6/8

HOOKWORM SPECIES AND IRON STATUS 5 3 5Table 3 Iron status of hookworm-infected children according to degree of A. duodenale infection determined by faecal culture

Iron status Indicator*N . americanuslarvae on ly(n = 340)

A. duodenale< 5 0 % of larvae( n = 1 1 3 ) > 5 0 % of larvae(n = 39)

Haemoglobin (g/dl)Mean SD b% < l l b% 1 2 0

Ferritln (ug/1)Geometric mean (+1 SD, -1 SD) C% < 1 2 C

10.4

97 (58,

14.5 (7.7,

1.560.514,0162)29.6

27.5)33.1

10.3

109 (65,

13.4 (7.2,

1.564.518.3183)36,9

25.4)50.7

9 8

103 (62,

9.6 (5.1,

1.80 .22 .

.5

.6

.8

173)3 2 .

18.1.9

I)58.9

' Values are predlrted from analysis of vanan ce, adjusted for hookworm eggs/g faeces determined by Kato-Katz method, age, helght-for-age Z score, malariaparasitaemia, and sex.

b Linear trend In Indicator by hookworm speaes category significant, P *; 0 05c Linear trend in indicator by hookworm speaes category significant, P < 0.001.

Table 4 Characteristics of study children in schools with low or highprevalence of A duodenale infection determined by faecal culturesA. duodenale preva lence

Characteristic Low, 2 0%(6 schools, 308 children) (4 schools, 217 children )Age (years)Sex (% girls)MicrohaematuriacMalaria parasitaemiaT trichiura

prevalenceeggs/g faecesA, lumbricoides

prevalenceeggs/g faeces

Hookwormsprevalenceeggs/g faeces

10.4 1.7"'u48.5%

20. 9% d61.6%

96.7%637 (110, 3689)e

71.5%234 (6, 9499)

9 4 . 8 % b486 (69, 3439)b

11.0 2.350.2%27.6%62.3%

9 7 . 2 %575 (123, 2 6 9 2 )

7 2 . 6 %2 59 (7, 9789)

9 8 . 2 %8 1 1 (180, 3652)

* Mean SD.b Significantly different from high prevalence schools, P < 0 05.c An indication of urinary schlstosomiasis.d Significantly different from high prevalence schools, P < 0.10c Geometric mean (-1 SD, +1 SD).

Prevalence and intensity of hookworm infeoion by Kato-Katz method.

differences in the impaa on iron status of individuals would bebased on measurement of all of these factors. In reality, such astudy is nearly impossible to conduct. To account for the dura-tion of infection, one would have to begin with an uninfectedindividual, and follow longitudinally the course of infectionboth in terms of species and number of worms, the iron intakeand bioavailability of the diet, and the iron status of the indi-vidual. Such a study would illuminate our understanding of

CJ Culture negativeB N. americanus only 50% A. duodenale

90* 8 0g 70 60f 50i 4 08 30o 20

1-1999 2000-3999 4000-5999 6000-7999 8000+Hookworm eggs/g faecesKato-Katz method

Figure 4 Distributions of hookworm faecal egg counts in childrenfrom schools with low (


7/8

5 3 6 I N T E R N A T I O N A L J O U R N A L O F E P I D E M I O L O G Y

Table 5 Iron status of Zanzibari children by prevalence of A duodenalein the schoolA.

Iron statusIndicator* (6Haemoglobin (g/dl)

Mean SD% < 1 1% < 9

duodenale prevalenceLow, 1 2 0

Ferritin (jig/1)Geometr ic mean(+1 SD, -1 SD)% < 1 2

93 (55, 156) d

27 .1 b

14.8 (7.8, 28.0) c

31.0d

High,>20%(4 schools, 217 chUdren)

10.2 1.565.118.3

109 (65, 184)35.2

13.1 (6.9, 24.9)48.4

' Values are predicted from analysis of variance, adjusted for hookwormeggs/g faeces determined by Kato-Katz me thod, age, height-for-age Z score,malaria parasitaemia, haematuna and sex.b Significantly different from high prevalence schools, P < 0 0 5c Significantly different from high prevalence schools, P < 0.10.d Significantly different from high prevalence schools, P < 0.001.

relationships between hookworm load, egg counts, and bloodlosses are difficult to es timate reliab ly'2 for a num ber of reasons.By simply including categories of hookworm eggs/g faeces inour analyses of the species variables, we have compared theiron status of individuals and communities within those categ-ories and thus have likely underestimated the impact of greaterblood loss per worm with A. duodenale. Although an unbiasedestimate of the species difference based on numbers of worms isdesirable, any m athematical model would be highly speculative,and this answer awaits a future study in which worms areexpelled and c ounted. In the me antim e, i t is reasonable to inferthat the true effect on iron status by the different species is atleast as great as our results would indicate.

We did not measure duration of infection nor iron intake andbioavailability. This does not threaten our inferences unlessduration of infection or iron intake were different for childrenor school communities in different species categories. We canthink of no scenario in which the duration of infection mighthave differed by species, since both species likely have beenendemic on Pemba Island for many years and children's risk ofexposure with age is not known to differ for the two species.It is somewhat more probable that dietary iron intake mightbe associated with species in our community level analyses. Ifchildren infected with A. duodenale or schoolchildren in com-munities with high levels of A. duodenale infection consume lessiron than their counterparts infected with N. americanus, thenour results could be spurious. Lacking dietary intake data, wecannot prove that this is not the case. However, we consider itto be unlikely that in this environment where essentially allchildren are infected with hookworms, the determinants ofexposure to A. duodenale as opposed to N. americanus would alsobe associated with poor diet.

The two approaches we took to address the question of spe-cies impact have com plementary strengths and weaknesses. Thecomparison of individuals by degree of A. duodenale infectiondoes not associate the species variable with school, an advant-age because school communites may vary in level of povertyand thus diet. Rather children are compared to other children,some of whom come from the same school community. Ourresults from this analysis are also strengthened by the dose-reponse pattern seen in haemoglobin and ferritin values bydegree of A. duodenale infection (Table 3).

The comparison of children grouped by community preval-ence of A. duodenale infection represents the practical questionthat confronts public health planners. Are comm unities or areaswh ere A. duodenale is highly endemic in greater need of inter-vention than communities where N. americanus predominates?Based on our results, the answer to this question is 'Yes.' Thestrength of this analysis is that in reality the risk of A. duodenaleinfection (or the force of transmission of that parasite) is morecharacteristic of a community than of an individual. The chil-dren in school communities with high prevalence of A. duodenalewho happened not to be infected with that species at the timeof our survey, have likely experienced A. duodenale in the pastor will do so in the future. The community also represents thelevel at which public health decisions are made. The predom-inate hookworm species being transmitted may be an appro-priate means of characterizing and targeting communities, butwill rarely influence the course of treatment of an individual.

Most importantly, the weight of our evidence is strengthenedby the consistency of the findings from complementaryapproaches to the problem. We conclude that, as predicted fromearlier clinical studies, the species of hookworm being trans-mitted in a community influences the burden of iron-deficiencyanaemia in the community.

The full implications of hookworm species for the planningand implementation of interventions are yet to be determined.Our results suggest that w here iron-deficiency an aem ia is apublic health problem, hookw orm control is particularly urgentif A. duodenale is the predominant species. However, this shouldnot be interpreted to mean that endemic N. americanus infectionis of low priority. We and others have demonstrated previouslythat hookworms can be an important cause of anaemia in pop-ulations whe re N. americanus p red o m in a t es .23 '2 4 Where anaemiais prevalent and hookworms are known to be endemic, lackof knowledge about hookworm species should not impedeinterventions.

A potentially important question is whether the optimal regi-men for anthelminthic control of hookworms in populationsdiffers for the two spedes. Because transmission of A. duodenalehas a distinctly seasonal patte rn in some envir onm ents , it ispossible that the efficacy of public health deworming regimenswould depend on the seasonal timing of treatments. It is alsopossible that the rate of reinfection following treatment differsfor the two spedes, with implications for the frequency of treat-ments. Further epidemiological research is needed to answerthese questions.

AcknowledgementsThe authors thank the Helminth Control Team, espedallyMr Faki BF, for their excellent work in collecting these data.


8/8

H O O K W O R M S P E C I E S AND IRON ST AT US 53 7

The work was supported by cooperative agreement DAN-5 1 1 6 -1 -0 0 -8 0 5 1 -0 0 b e tween The Johns Hopkins Universityand the Office of Health and Nutrition, United States Agency forInternational Development.

References1 Warren KS, Bundy DAP, Anderson RM et al Helminth infection. In.

Jamison DT, Mosley WH, Measham AR, Bobadilla JL (eds). DiseaseControl Priorities in Developing Countries New York. Oxford UniversityPress, 1993, pp.131-60.

2 Pawlowski ZS, Schad GA, Stott GJ. Hookworm Infection and Anaemia:Approaches to Prevention and Control. Geneva: World Health Organ-ization, 1991.

3MillerTA. Hookworm infection in man. AdvParasitol 1978:17:315-84.4 Roche M, Layrisse M. The natu re and causes of 'hoo kwor m anem ia'.

Am J Trop Med Hyg 1966:15:1031-100.5GiUes HM, Watson Williams EJ, Ball PAJ. Hookw orm infection and

anaemia: an epidemiological, clinical, and laboratory study. Q J Med1964:33:1-24.

6 Stephenson LS. Impact of Helminth Infections on Human Nutrition. NewYork: Taylor fr Francis, 1987.

7 Schad GA. Presidential address. Hooked on hookworm: 25 years ofattachment. J Parasitol 1991:77:177-86.

8 Schad GA, Chowdhury AB, Dean CG et al Arrested developmentin human hookworm infections: an adaptation to a seasonallyunfavorable external environment Science 1973,180:502-04.

9 Anderson RM, Schad GA. Hookworm burdens and faecal egg counts:an analysis of the biological basis of variation Trans R So c Trop Med Hy g1985;79:812-25.

10Prit char d DI. The survival strategies of hookworms . Parasitology Today1995:11:255-59.

1 1 Cook JD. Adaption in iron metabolism. Am J Chn Nutr 1990:51:301-08.

12 Crompton DWT, Whitehead RR. Hookworm infections and humaniron metabolism Parasitology 1993,107:S137-145.

13Sto ltzf us RJ, Chwaya HM, Tielsch JM, Schulze KJ, Albonico M,Savioli L. Epidemiology of iron deficiency anemia in Zanzibari schoolchildren. Am J Chn Nutr 1997:65:153-59.

1 4 Renganath an E, Ercole E, Albonico M et al. Evolution of operationalresearch studies and development of a national control strategyagainst intestinal helminths in Pemba Island experien ce, 1988-1992Bull World Health Organ 1995:73:183-90.

1 5de Camen I, Albonico M, Di Matteo L, Savioli L, Kisumku UM. SiaNecator sia Ancylostoma parassitano l'uomo a Pemba (Zanzibar).Parassitologia 1992;34(Suppl.l):190.

1 6 World Health Organization Bench Aids for the Diagnosis of IntestinalParasites Geneva: WHO, 1994.

17 Har ada Y, Mori O. A new me thod for culturing hoo kwor m. YonagoAda Medica 1955;l:177-95.

1 8 Garcia LS, Bruckner DA Diagnostic Medical Parasitology. New York;Elsevier, 1988.

1 9 Savioli L, Hatz C, Dixon H, Kisumku UM, Mot t KE. Control ofmorbidity due to Schistosoma haematobium on Pemba Island: eggexcretion and haematuria as indicators of infection. Am J Trop Med Hyg1990;43:289-95.

20 Tr ape Jp. Rapid evaluation of malaria parasite density and standa rd-ization of thick smear examination for epidemiological investigations.Trans R Soc Trop Med Hyg 1985;79:181-84.

2 1 Johnson-Spea r MA, Yip R. Hemoglobin difference be tween black andwhite women with comparable iron status: justification for race-specific anemia criteria. Am J Clin Nutr 1994;60:117-21

2 2 Gilles HM. Selective primary health care: Strategies for control of dis-ease in the developing world. XVH. Hookworm infection and anemi a.Rev Infect Di s 1985:7:111-18.

2 3 Stoltzfus RJ, Albonico M, Chwaya HM el al HemoQuant determ-ination of hookworm-rel ated blood loss and its role in iron deficiencyin African children Am J Trop Med Hyg 1996:54:399^104.

2 4 Layrisse M, Roche M. The relationship between anemia and hook-worm infection. Am J Hyg 1964:79:279-301.

Documents

Int. J. Epidemiol. 1998 Albonjco 530 7