A controlled evaluation of two school-based ... - Semantic Scholar

© International Epidemiological Association 1999

International Journal of Epidemiology 1999;28:591–596

Printed in Great Britain

A controlled evaluation of two school-based anthelminthic chemotherapy regimens on intensity of intestinal helminth infections Marco Albonico,a,c Rebecca J Stoltzfus,b Lorenzo Savioli,c Hababu M Chwaya,d Emmanuel d’Harcourte and James M Tielschb

Background School-based deworming programmes have been promoted as a cost-effective strategy for control of nematode infection in developing countries. While numerous efficacy studies have been conducted, there is little information on actual programme effectiveness in areas of intense transmission. Methods

A randomized trial of a school-based deworming programme was conducted in 12 primary schools on Pemba Island, Zanzibar. Four schools each were randomized to control, twice a year deworming with single dose mebendazole or three times a year deworming. Baseline and 12-month follow-up data on helminth infection using the Kato-Katz technique, demographic information and nutritional status were collected on 3028 children from March 1994 to May 1995.

Results

Intensity of infection measured as eggs per gram of faeces (epg) declined significantly for Ascaris lumbricoides, Trichuris trichiura and hookworm infections in both treatment groups. A. lumbricoides infection intensity declined 63.1% and 96.7% in the twice and three times per year treatment groups compared to the controls. T. trichiura infection intensity declined 40.4% and 75.9% respectively and hookworm intensity declined 35.3% and 57.2% respectively compared to control schools.

Conclusions These results suggest that school-based programmes can be a cost-effective approach for controlling the intensity of intestinal helminth infection even in environments where transmission is high. Keywords

Helminth infection, hookworm, deworming, school-based, randomized trial, Ascaris lumbricoides, Trichuris trichiura

Accepted

29 July 1998

Chemotherapy with single dose, broad-spectrum, safe, and lowcost anthelminthic drugs is the mainstay of programmes aimed at the control of morbidity due to intestinal nematodes.1 Chemotherapy targeted at school children has been considered one of the most cost-effective strategies to control helminthic infections in endemic areas.2,3 However, optimal programme models both in terms of delivery systems and frequency of dosing have not been established. Mathematical models predict that relatively short intervals between treatments (i.e. 2–6 months) are required in areas of a Scientific Committee, Ivo de Carneri Foundation, Milan, Italy. b Department of International Health, Johns Hopkins University School of

Hygiene and Public Health, Baltimore, MD, USA. c Schistosomiasis and Intestinal Parasites Unit, Division of Control of Tropical

Diseases, World Health Organization, Geneva, Switzerland. d Ministry of Health, Zanzibar, United Republic of Tanzania. e Children’s Hospital, Philadelphia, PA, USA.

Reprint requests: James M Tielsch, Department of International Health, Johns Hopkins School of Hygiene and Public Health, 615 N. Wolfe Street, Room 5515, Baltimore, MD 21205–2103, USA.

moderate to high transmission in order to reduce the prevalence of helminthic infections to very low levels.4 This has been confirmed by field experience in different countries,5,6 However, prevalence of infection is a poor indicator of the average intensity of infection (worm burden) as intestinal nematodes are highly aggregated in their distribution within a community.7 In this situation, large changes in average worm burden and the resulting decline in associated morbidity are not reflected by small observed changes in prevalence of infection. In Zanzibar, intestinal helminths are recognized to be a major public health problem with 99% of school children infected with one or more of the three major soil-transmitted nematodes: Ascaris lumbricoides, Trichuris trichiura and the hookworms (Necator americanus and Ancylostoma duodenale).8 This heavy burden of intestinal nematodes is believed to contribute to the widespread malnutrition of Zanzibari children.9,10 In response to this problem, the Ministry of Health of Zanzibar established a National Helminth Control Programme in June 1994. The proposed control strategy focused on periodic chemotherapy targeted at school children and delivered by

591

592

INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

teachers under the supervision of Ministry of Health staff. Previous research had established that mebendazole used as a single 500 mg dose was the most cost-effective anthelminthic treatment for this setting.11,12 The objective of this study was to compare two different regimens, every 4 months versus every 6 months, in the reduction of intensity of infection and associated morbidity.

Materials and Methods Study area The study was conducted on Pemba Island, the smaller of the two islands of Zanzibar, approximately 5 degrees south of the Equator and 50 km off the coast of mainland Tanzania. The island has a humid tropical climate with temperatures that rarely exceed 34°C or fall below 21°C. The rainfall pattern is characterized by a long rainy season lasting from March to May with heavy rains and by a short rainy season from October to December with light rains. Pemba is highly populated with 330 inhabitants per km2. About 15 000 people live in each of the three main towns, but the majority of the population is scattered in numerous rural villages. The economy is mainly based on farming and fishing, the main cash crops are cloves and coconuts. School attendance is estimated to be 80% among the first graders, with a steady decrease throughout the primary school years.

Study population and study sample The study was designed as a randomized programme evaluation to measure the change in prevalence and intensity of helminthic infections and morbidity 12 months after the initiation of the deworming programme. Twelve public elementary schools were randomly selected from the 72 schools on Pemba Island. From these schools, all 3944 children enrolled in 76 morning classes of the first four grades (standards 1–4) were invited to participate in the study. Participation in the study was voluntary and consent was obtained from the parents of all children participating. These schools were randomly allocated to three groups after stratification by district: Group A, as the four control schools; Group B, four schools in which a twice a year deworming regimen was implemented; and Group C, in which a three times a year deworming regimen was used. The study was conducted from March 1994 to May 1995 and was reviewed and approved by the institutional review boards of the Johns Hopkins University School of Hygiene and Public Health, the Ministry of Health of Zanzibar and the World Health Organization. Children were examined at baseline, from March to May 1994, with a nutritional and parasitological survey. Data were collected from 3605 children (91.4% of eligibles). After 6 months and 12 months nutritional and parasitological surveys were conducted in the same classes. Additional parasitological surveys assessing helminthic infections were also performed at 4 and 8 months in the Group C classes just before each treatment. On average, 92% of children received the correct dosing of mebendazole and only 7% of children received one dose less than assigned by the randomization. In the present paper, results are presented on 3028 children that were examined at baseline survey and followed after 12 months (84.0% of those surveyed at baseline and 76.8% of those originally eligible).

Assessments In every school, identification and demographic information, anthropometric indicators, faecal, urine and blood samples were collected from each child. The sample collection rates at baseline and at 12-month surveys were 100% for blood, 99% for urine, 95% and 90% for stool samples, respectively. Urine was tested with a reagent strip for semiquantitative values of haematuria (Hemastix, Ames-Miles Laboratories, Elkhart, IN). Microhaematuria indicates urinary schistosomiasis in this population with 69% sensitivity and 88% specificity.13 The children testing positive for visual haematuria were treated with praziquantel 40 mg/kg body weight. Haemoglobin (Hb), erythrocyte protoporphyrin and serum ferritin were also assessed as part of the evaluation of iron status.9 Malaria was investigated with thin and thick smears for each child. Children with fever were treated with pyrimethamine/sulfadoxine (Fansidar) at a dose of 25 mg/kg body weight of sulfadoxine. Those with severe anaemia (Hb,7 g/dl) were treated with mebendazole and oral iron, and were excluded from the analysis. All children were treated with mebendazole 500 mg as a single dose at the end of the trial. In the laboratory, stool samples were processed by the Kato-Katz quantitative technique according to WHO recommendations.14 Samples containing more than 10 000 eggs were re-examined by a modified Stoll dilution technique.15 The thick and thin blood smears were stained with 3% Giemsa solution and the malaria parasites were identified by species and counted against leucocytes to calculate parasitaemia per µl of blood.14 Quality control was performed by two laboratory technicians by re-examining 10% of the slides chosen at random for malaria and intestinal helminths, respectively. Agreement between readers was excellent for the presence of malaria parasitaemia (kappa = 0.93) and parasite density estimates were also highly reliable with an intraclass correlation of 0.85. Intraclass correlation between staff and quality control readings were 0.88 eggs per gram of faeces (epg) for A. lumbricoides, 0.79 epg for T. trichiura and 0.89 epg for hookworm.

Statistical analysis Data were entered into a microcomputer using the Epi-Info package. Intensity of intestinal helminth infections was measured by the mean epg count. The geometric mean was used to normalize the typical over-dispersed distribution of worm intensity in endemic areas. Infections were classified as light, moderate and heavy according to WHO criteria.16 The thresholds for moderate and heavy infections were 5000 and 50 000 epg for A. lumbricoides, and 1000 and 10 000 epg for T. trichiura. For hookworm infection, categories with thresholds of 2000 and 5000 epg were used to define moderate and heavy infections respectively. Estimates of variance were adjusted to account for the clustered randomization scheme using the Generalized Estimating Equation approach.17

Results The prevalence of A. lumbricoides, T. trichiura and the hookworms at baseline were 79%, 96% and 94% respectively (Table 1). Malaria parasites were found in 61% of children, 99% of which were Plasmodium falciparum infections. Thirty-one per cent of children tested positive for haematuria. Baseline comparisons of

A CONTROLLED EVALUATION OF SCHOOL-BASED DEWORMING Table 1 Characteristics of the study sample at baseline Control Twice a year Three times a year (n = 1037) (n = 980) (n = 1011) Age (mean ± SD)

10.5 ± 1.6

10.6 ± 2.0

10.5 ± 1.8

50.9

50.2

49.8

73.0%

66.5%

76.2%

238

149

330

None

27.0%

33.5%

23.8%

Light

52.9%

47.7%

53.4%

Moderate

20.1%

18.7%

22.7%

0%

0.1%

0.1%

94.8%

96.6%

97.2%

531

584

627

None

5.2%

3.4%

2.8%

Light

54.0%

56.0%

56.0%

Moderate

40.2%

39.4%

40.3%

0.6%

1.3%

0.9%

91.1%

94.6%

95.8%

334

510

603

None

8.9%

5.4%

4.2%

Light

75.7%

75.6%

73.0%

Moderate

13.3%

14.5%

18.7%

2.1%

4.5%

4.2%

Sex (% boys) Ascarisa Prevalence Geometric mean epg

Heavy Trichurisa Prevalence Geometric mean epg

Heavy Hookworma Prevalence Geometric mean epg

Heavy

a Helminth prevalence and intensity based on 993 observations in control

group, 922 in twice a year group and 975 in three times a year group.

treatment groups showed that age and gender distributions were similar in the three treatment groups (Table 1). The three times a year treatment group had a slightly higher prevalence and intensity of infection for all three geohelminths. Potential biases were examined by comparing children that were followed at 12 months and the 577 children that were lost to follow-up. Slightly more children were lost in the control

593

and in the three times a year treatment groups (21.5% and 22.1%) than in the twice a year group (19.3%), (P , 0.05). Boys (22.7%) and non-stunted children (21.5%) were more likely to be lost than girls (15.3%) and stunted children (16.5%), (P , 0.01). However, there was no difference in the baseline prevalence or intensity of geohelminth infections between respondents and those lost to follow-up. The intensity of infection 12 months after treatment showed a dramatic decline in both treatment groups compared to the control group for all three geohelminths (Table 2). In the control group, the intensity of infection rose between 46.0% and 179.4% from baseline to 12 months. In contrast, intensity decreased in both treatment groups (P , 0.001), with the decline significantly higher in the three times a year group compared with the twice a year group (P , 0.001) (Table 2). The pattern of intensity of infection over time was different for the three geohelminths (Table 2). For A. lumbricoides infection, there was a marked decrease in both treatment groups at 6 months (87.9% and 99.1% in the twice a year and three times a year treatment group, respectively) and maintenance of low levels of infections at 12 months. T. trichiura infection showed a significant difference at 6 months (31.0% reduction in the twice a year group and 58.2% in the three times a year group) and at 12 months with almost twice the impact in the three times a year compared to the twice a year treatment group. A modest increase for hookworm infections in the twice a year (16.1%) and three times a year (5.6%) treatment groups was detected at 6 months. However, this unexpected trend must be compared to a large increase in intensity of infection in the control group. At 12 months, intensity of hookworm infection was significantly reduced (P , 0.001) for both treatment groups, with a larger decline in the more frequent treatment schedule (P , 0.001). We also examined changes in categories of intensities according to WHO criteria16 between the treatment groups at 12 months. For A. lumbricoides infection, there was a 12% and 50% reduction in prevalence in the twice a year and three times a year treatment groups respectively. Most infections were cured or shifted to light infections, with an 86.4% reduction of moderate-heavy infections in the three times a year treatment

Table 2 Intensity of geohelminth infections (geometric mean (±SD)) at baseline, 4, 6, 8 and 12 months Baseline epg

4 months epg

6 months epg

% change from baseline

8 months epg

12 months epg

% change from baselinea

Ascaris Control

238 (±35)

410 (±35)

+72.3

665 (±22)

+179.4

Twice a year

149 (±42)

18 (±35)

–87.9

55 (±35)

–63.1

Three times a year

330 (±32)

3 (±10)

–99.1

11 (±25)

–96.7

3 (±8)

4 (±12)

Trichuris Control

531 (±6)

821 (±6)

+54.6

775 (±6)

+46.0

Twice a year

584 (±5)

403 (±7)

–31.0

348 (±8)

–40.4

Three times a year

627 (±5)

262 (±8)

–58.2

151 (±11)

–75.9

378 (±9)

164 (±12)

Hookworm Control

334 (±9)

950 (±5)

+184.4

803 (±6)

+140.4

Twice a year

510 (±7)

592 (±5)

+16.1

330 (±9)

–35.3

Three times a year

603 (±6)

637 (±4)

+5.6

258 (±10)

–57.2

889 (±5)

591 (±6)

a Difference between % change from baseline between twice a year versus control and twice a year versus three times a year are all significant at the P < 0.001

level.

594

INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

group (Figure 1). There was only a modest effect on prevalence of T. trichiura infection, with a 4.8% and 12.4% reduction in the twice a year and three times a year treatment groups. However, there was a 27.5% reduction in moderate-heavy infections for the twice a year group and a 60.4% reduction in the three times a year group (Figure 2). Similarly, the reduction in prevalence of hookworm infection was largest in the moderate to heavy infections, 23.2% in the twice a year group and 61.1% in the three times a year group (Figure 3). The impact of treatment was even more evident when looking at the 45.6% increase in moderate to heavy hookworm infections in the control group at 12 months (Figure 3). As expected during actual programme operation, not all children received all doses assigned due to school absences,

Figure 1 Prevalence of moderate and heavy infection of Ascaris lumbricoides at baseline and 12-month follow-up by treatment group

Figure 2 Prevalence of moderate and heavy infection of Trichuris trichiura at baseline and 12-month follow-up by treatment group

Figure 3 Prevalence of moderate and heavy infection of hookworm at baseline and 12-month follow-up by treatment group

refusals, etc. In addition, some children received more doses than assigned because severely anaemic children were treated with mebendazole regardless of treatment group. To explore further the dose-response relationship between chemotherapy and intensity of infection, we broke the randomization and plotted the proportionate change in intensity from baseline by number of doses received (Figures 4, 5 and 6). For all three helminth infections, there was a significant dose-response effect on intensity of infection with those who received three or four doses having the largest proportionate reductions in the face of an overall expected increase in intensity as demonstrated by those who received no drug.

Discussion This evaluation demonstrated that a school-based deworming programme using periodic treatment with mebendazole 500 mg as a single dose can significantly reduce the intensity of infection of three common intestinal nematodes: A. lumbricoides, T. trichiura and the hookworms. The geometric mean intensity of infection for children that were treated every 4 months for a year, as compared to children in the control group, was 96.7% lower for A. lumbricoides, 75.9% lower for T. trichiura and 57.2% lower for hookworms. These declines in intensity of infection were made despite the large increases in intensity of infection observed among children in the control group. The impact was greatest on A. lumbricoides, confirming the results of previous studies showing that mebendazole is most effective against A. lumbricoides and less effective against the hookworms and T. trichiura.10,18,19 In Zanzibar, where intensity of transmission is extremely high, there was a significant advantage to using a three times a year rather than a twice a year dosing schedule. These results also reinforce the importance of a control group in community trial design especially for infections where the intensity of transmission can vary from season to season and year to year. Without an untreated group, we would have concluded that mebendazole had no impact on hookworm infections at 6 months, a clearly inaccurate inference. The increase in burden of hookworm infection is particularly striking in these data. Both major hookworm species Necator americanus and Ancylostoma duodenale, were prevalent in our study population20 and some proportion of the increase in hookworm burden may have been due to a reactivation of developmentally arrested larvae21 as opposed to reinfection. While our data provide some support for the concept of arrested development, especially in Ancylostoma duodenale, they cannot directly address this interesting hypothesis. These results suggest that the seasonality of transmission of geohelminths may be an important factor to consider when planning a schedule for mass chemotherapy.21,22 To reduce further the worm burden of hookworm in a highly endemic area such as Pemba, it is worth considering alternative treatment schedules with mebendazole or the choice of a more effective, but more costly, drug for hookworm infection such as albendazole. For example, giving two 500 mg doses of mebendazole a week apart after the rainy season (during high transmission) may significantly reduce worm burden. Another strategy would be to plan more frequent dosing schedules at the end of the peak transmission season, to hit the worms at a time when they are maturing in the intestine, and to minimize the incidence of

A CONTROLLED EVALUATION OF SCHOOL-BASED DEWORMING

Figure 4 Proportional change in geometric mean eggs per gram of faeces from baseline examination for Ascaris lumbricoides by number of doses of mebendazole received

Figure 5 Proportional change in geometric mean eggs per gram of faeces from baseline examination for Trichuris trichiura by number of doses of mebendazole received.

Figure 6 Proportional change in geometric mean eggs per gram of faeces from baseline examination for hookworms by number of doses of mebendazole received

post-treatment reinfection. Whatever additional strategies are considered, the different costs of the benzamidazoles11 and the logistics of delivery must be carefully evaluated in the design of cost-effective delivery schedules. Our study is one of only a handful of trials of anthelminthic therapy in school children,3,23–28 in spite of the magnitude of

595

helminthic infections in this age group across the world.29 In comparison to other trials of anthelminthic therapy, our study had several strengths. First, because our study took place within the context of a school-based programme, we measured the impact of a public health programme in practice in sub-Saharan Africa, and not just the efficacy of a chemotherapeutic agent. Secondly, the large size and the use of a control group permitted statistically efficient comparisons between groups. Zanzibar, with its high prevalence of growth retardation, anaemia, malaria and other parasitic infections, is representative of coastal East Africa and many other developing regions of the world. Thus, in addition to providing valuable information to the Government of Zanzibar, the results of this evaluation are likely to be of use to public health planners in other areas where intestinal helminthic infections are endemic. Several aspects of our trial suggest that we may have underestimated the impact that a similar programme may have in other areas where infection is less intense and less prevalent. First of all, high reinfection rates likely decreased the impact of our periodic treatment. The fact that intensity of infection increased so significantly in the control group over the course of the study, suggests that reinfection rates may have been particularly high during this time. Furthermore, because the intensity of infection was so high at baseline, eggs per gram of faeces, used in this study to measure intensity of infection, may have been a less accurate index of worm burden. As worm load increases in an individual child, each worm becomes less efficient at producing eggs.30 Conversely, as the worm burden is lowered, each worm’s egg production is increased. Thus, measuring egg output may have underestimated the actual reduction in worm burden. It is also worth considering that the egg counts at 4, 6 or 12 months after treatment do not give a comprehensive picture of the decrease of the worm burden. The decline is not linear but sharp after treatment with a nadir at 2–3 weeks after which egg counts gradually increase up to the post-treatment survey 4 or 6 months later.12 In summary, our study showed that school-based deworming programmes can be an effective strategy in decreasing the intensity of helminthic infection in areas of high endemicity, and that there is significant advantage to a more intense, three times a year treatment schedule. It has also demonstrated that prevalence is a poor indicator for monitoring the impact of chemotherapy-based helminth control programmes in highly endemic areas.7 More work is needed, however, to improve the impact of chemotherapy. Our study and others11,18,19 show that mebendazole given as a single dose, although safe, effective and affordable, has limits, particularly in the treatment of hookworm in highly endemic areas. The impact of other affordable anthelminthic agents or the development of alternative treatment schedules in helminth control programmes in different epidemiological settings should be evaluated to determine their relative cost-effectiveness. Finally, more operational research is needed to measure the impact of deworming programmes on morbidity in school children, since the ultimate objective of school-based programmes is not only to reduce worm burden, but also to improve growth, physical fitness and school performance. Several studies have been conducted considering these issues. However, they are relatively small, involving less than 300 children,31–33 and show the impact of deworming only in an ideal research setting.

596

INTERNATIONAL JOURNAL OF EPIDEMIOLOGY

Acknowledgements The authors thank the Pemba Public Health Team for their excellent work in collecting these data. The study was funded by cooperative agreement DAN-5116–1–00–8051–00 between the Office of Health and Nutrition, United States Agency for International Development and the Johns Hopkins University. Additional support was provided by the WHO Program of Control of Intestinal Parasitic Infections and the Government of Italy.

References 1 Savioli L, Bundy D, Tomkins A. Intestinal parasitic infections: a

soluble public health problem. Trans Roy Soc Trop Med Hyg 1992;86: 353–54. 2 Anderson RM, Medley GF. Community control of helminth infections

of man by mass and selective chemotherapy. Parasitology 1985;90: 629–60. 3 Bundy DAP, Wong MS, Lewis LL et al. Control of geohelminths by

delivery of targeted chemotherapy through schools. Trans Roy Soc Trop Med Hyg 1990;84:115–20. 4 Anderson RM. Transmission dynamics of Ascaris lumbricoides and

the impact of chemotherapy. In: Crompton DWT, Nesheim MC, Pawlowski ZS (eds). Ascariasis and its Prevention and Control. London and Philadelphia: Taylor and Francis, 1989, pp.44–54. 5 Seo B, Chai J. Comparative efficacy of various interval mass treatment

on Ascaris lumbricoides infection in Korea. Korean J Parasitol 1980; 18:145–51. 6 Thein Hlaing. Ascaris lumbricoides infections in Burma. In: Crompton

DWT, Nesheim MC, Pawlowski ZS (eds). Ascariasis and its Public Health Significance. London and Philadelphia: Taylor and Francis, 1985, pp.83–122. 7 Bundy DAP, Hall A, Medley GF et al. Evaluating measures to control

intestinal parasitic infections. World Health Stat Q 1992;45:168–79. 8 Albonico M, Chwaya HM, Montresor A et al. The burden of parasitic

infection in school children of Zanzibar: a need for integrated control. E Africa Med J 1997;74:294–98. 9 Stoltzfus RJ, Chwaya HM, Tielsch J et al. Epidemiology of iron

deficiency anaemia in Zanzibari school children. Am J Clin Nutr 1997;65:153–59. 10 Stoltzfus RJ, Albonico M, Tielsch J et al. Linear growth retardation in

Zanzibari school children. J Nutr 1997;127:1099–105. 11 Albonico M, Smith PG, Hall A et al. A randomised controlled trial

comparing Mebendazole 500 mg and Albendazole 400 mg against Ascaris, Trichuris and the hookworms. Trans Roy Soc Trop Med Hyg 1994;88:585–89. 12 Albonico M, Smith PG, Ercole E et al. Rate of reinfection with

intestinal nematodes after treatment of children with mebendazole or albendazole in a highly endemic area. Trans Roy Soc Trop Med Hyg 1995;89:538–41. 13 Savioli L, Hatz C, Dixon H et al. Control of morbidity due to Schistosoma

haematobium on Pemba Island: egg excretion and haematuria as indicators of infection. Am J Trop Med Hyg 1990;43:289–95. 14 Basic Laboratory Methods in Medical Parasitology. Geneva: World Health

Organization, 1991. 15 de Carneri I. Aggiornamento di un metodo per il conteggio di alte

cariche di uova di elminti nelle feci. Giornale di Malattie Infettive e Parassitarie 1992;44:854–55.

16 Prevention and control of intestinal parasitic infections. Report of a WHO

Expert Committee. World Health Organ Tech Rep Ser 749. Geneva: WHO, 1987, pp.12–13. 17 Zeger SL, Liang KY, Albert PS. Models for longitudinal data: a gen-

eralized estimating equation approach. Biometrics 1988;44:1049–60. 18 Muttalib MA, Khan MU, Haq JA. Single dose regime of mebendazole

in the treatment of polyparasitism in children. J Trop Med Hyg 1981; 84:159–60. 19 Ismail MM, Premaratne UN, Suraweera MG. Comparative efficacy of

single dose anthelminthics in relation to intensity of geohelminth infections. Ceylon Med J 1991;36:162–67. 20 Albonico M, Stoltzfus R, Cancrini G et al. Epidemiologic evidence for

a differential effect of hookworm species A. duodenale or N. americanus, on iron status of children. Int J Epidemiol;1998;27:530–37. 21 Schad GA, Chowdhury AB, Dean CG et al. Arrested development

in human hookworm infections: an adaptation to a seasonally unfavourable external environment. Science 1973;180:502–04. 22 Kobayashi A. Studies on the mode of Ascaris infection in Japan. In:

Yokogawa M et al. (eds). Collected Papers on the Control of Soil-transmitted Helminthiases Vol I. Tokyo: APCO, 1980, pp.109–18. 23 Shield J, Anian G, Ostwald R et al. Reinfection with intestinal

helminths after treatment with mebendazole and fluctuations in individual Ascaris lumbricoides infections with time. PNG Med J 1984; 27:89–94. 24 Sinniah B. Prevalence, treatment and reinfection of intestinal

helminths among schoolchildren in Kuala Lumpur, Malaysia. Public Health Lond 1984;98:38–42. 25 Bundy DAP, Wong MS, Lewis LL et al. Control of geohelminths by

delivery of chemotherapy through schools. Trans Roy Soc Trop Med Hyg 1990;84:115–20. 26 Sinniah B, Chew PI, Subramaniam K. A comparative trial of

albendazole, mebendazole, pyrantel pamoate and oxantel pyrantel pamoate against soil-transmitted helminthiases in school children. Trop Biomed 1990;7:129–34. 27 Stephenson LS, Latham MC, Adams EJ et al. Physical fitness, growth,

and appetite of Kenyan school boys with hookworm, Trichuris trichiura and Ascaris lumbricoides infections are improved four months after a single dose of albendazole. J Nutr 1993;123:1036–46. 28 Stephenson LS, Kinoti SN, Latham MC et al. Single dose metrifovate

or praziquantel treatment in Kenyan children. I. Effects on Schistosoma haematobium, hookworm, hemoglobin levels, splenomegaly, and hepatomegaly. Am J Trop Med Hyg 1989;41:436–44. 29 Bundy DAP. The global burden of intestinal nematode disease. Trans

Roy Soc Trop Med Hyg 1994;88:259–61. 30 Hall A. Intestinal helminths of man: the interpretation of egg counts.

Parasitology 1982;85:605–13. 31 Robertson LJ, Crompton DWT, Sanjur D et al. Haemoglobin

concentrations and concomitant infections of hookworm and Trichuris trichiura in Panamanian primary schoolchildren. Trans Roy Soc Trop Med Hyg 1992;86:654–56. 32 Nokes C, Grantham-McGregor SM, Sawyer AW et al. Moderate to

heavy infections of Trichuris trichiura affect cognitive functions in Jamaican schoolchildren. Parasitology 1992;104:539–47. 33 Stephenson L, Latham MC, Kinoti SN et al. Improvements in physical

fitness of Kenyan schoolboys infected with hookworm, Trichuris trichiura and Ascaris lumbricoides following a single dose of albendazole. Trans Roy Soc Trop Med Hyg 1990;84:277–82.