The impact of weekly iron supplementation on the iron status and growth of adolescent girls in Tanzania


correspondence Professor Andrew Tomkins, Centre for International Child Health, Institute of Child Health, University College, 30 Guilford Street, London WC1N 1EH, UK. E-mail:


Summary We evaluated the effect of weekly doses of 400 mg of ferrous sulphate for 4  months on the iron status of adolescent girls in a controlled trial in Tanga, Tanzania. Supplementation led to a significantly greater increase in serum ferritin compared with the control group (+ 15.6 μg/l vs. 8.6 μg/l) (P= 0.002) but there was no significant difference in change in haemoglobin. Children given iron showed a significantly greater weight gain than controls (+ 2.4 kg vs. + 1.4 kg) (P= 0.03). Weekly iron supplementation may be an effective means of increasing iron stores and growth in children vulnerable to iron deficiency.


In recent years, a number of studies have suggested that weekly iron supplementation is as effective as daily supplementation in raising haemoglobin levels (Gross et al. 1994; Liu et al. 1995; Schultink et al. 1995; Ridwan et al. 1996; Angeles-Agdeppa et al. 1997; Berger et al. 1997; Palupi et al. 1997; Viteri et al. 1999). Although they have been subject to criticism (Hallberg 1998), the studies found that groups receiving intermittent iron supplementation achieved significant benefits equal to that of groups supplemented daily irrespective of age or culture (Beard 1998). In the past, many daily supplementation programmes in developing countries have been unsuccessful due to lack of supplies and compliance (Galloway & McGuire 1996) and as well as requiring fewer iron tablets, several studies found that weekly supplementation led to a reduction in these problems.

The majority of studies published so far have been conducted in Asia, with one (Berger et al. 1997) in South America. Due to the results of the studies, there is a growing momentum in several developing countries for a change in policy from daily to weekly or intermittent dosing (Galloway & McGuire 1996). However, there is little information on the impact of weekly iron supplementation in Africa where determinants of anaemia often differ markedly from those in Asia. This paper reports a study of weekly iron supplementation amongst adolescent girls living in the Tanga region of Tanzania.

Adolescent girls are an important target group for supplementation since their iron requirements are high (Brabin & Brabin 1992). Many are on the verge of becoming mothers for the first time and can readily be supplemented through school-based delivery programmes (Partnership for Child Development 1998), a measure which, if successful, could make the new technology even more cost-effective.

Materials and methods

Study area and population

This study was undertaken by Ushirikiano wa Kumwendeleza Mtoto Tanzania (UKUMTA) (Tanzania Partnership for Child Development), a partnership of government agencies, scientific institutions and donor agencies which aims to improve the health of school-aged children in Tanzania and which is part of an international research consortium (PCD 1998). UKUMTA's work began in Tanga Region and the study was conducted in a rural area of Muheza District in the villages of Misongeni, Ubembe and Kilometa Saba. Any child aged between 12 and 18 years and living in the villages was eligible to participate in the study, irrespective of school attendance. The study commenced in April 1996 and follow-up occurred in August 1996.

Inclusion criteria, study design and randomization

The criteria for inclusion into the study were completion of all baseline measurements and a haemoglobin concentration > 70 g/l. Children with severe anaemia were referred for treatment. Children included were randomly assigned to two groups matched for intensity of infection with hookworm and S. haematobium. Both were treated for infection with geohelminths and S. haematobium as part of a school-based mass treatment programme run within the region by UKUMTA. Treatment took the form of a single dose of albendazole (400 mg) and praziquantel (40 mg/kg body weight). During the 16 weeks following anthelminthic treatment, one group was given 12 doses of ferrous sulphate (2 × 200 mg per dose) (Weiders Ltd, Norway) whilst the second group was given 12 doses of a control treatment (2 × 50 μg cyanocobalamin (vitamin B12)) (Gold Shield Healthcare Ltd, Croydon, UK). The use of a strict placebo was not allowed by local and national ethical bodies. Whilst vitamin B12 deficiency can be considered as a contributory factor to anaemia, macrocytosis was observed in only 8% of blood films taken at baseline. Given this low prevalence, it was considered appropriate to use cyanacobalamin as a locally acceptable control. Girls were given treatments at intervals of at least 1 week. It was not possible to give children a weekly dose for the whole 16-week period since many were absent during the month-long school break in June. The design was single blind only, in which measurement staff did not know the treatment group.

Community and institutional approval

The study was approved by the Tanzania Commission for Science and Technology, the Tanzania Food and Nutrition Centre, Dar es Salaam and local officials of the Ministries of Health and Education in Tanga. After obtaining their agreement, meetings were held in all villages before the start of the trial to explain both its aims and procedures, and to gain community consent. The entry of children into the study was achieved after full explanation to parents and children. It was clearly explained to the children that they could decline to enter the study or remove themselves from the study without prejudice to their treatment.

Baseline measurements

Stool, urine and blood samples were collected from each child. Stool samples were examined microscopically once for the eggs of species of intestinal worms using the rapid Kato-Katz technique and the concentration of eggs was expressed as eggs/g of faeces. A 10-ml aliquot of urine was filtered through 12 mm diameter polycarbonate membranes with a 12-μm pore size (Costar, UK) held in pop-top membrane holders (Millipore Ltd). The presence of eggs of S. haematobium was noted and the concentration expressed as eggs/10 ml of urine. Venous blood was collected using a closed collection system (Vacutainer Ltd) and the haemoglobin concentration estimated using a photometer (Hemocue Ltd); 1-ml aliquots of blood were centrifuged and serum collected (Becton Dickinson Microtainer, MSE Microcentaur from Fisons). Ferritin, C-reactive protein and α1-glycoprotein concentrations were determined by ELISA at the Tanzania Food and Nutrition Centre. Ferritin and CRP were measured by sandwich ELISAs using capture and horseradish peroxidase-conjugated antibodies to ferritin and CRP (Dako, Cambridge, UK). Ferritin standards were manufactured by Dako and CRP standards by Behring Diagnostics (Milton Keynes, UK). The method used to measure α1-glycoprotein has been described previously (Filteau et al. 1994). Thick and thin blood films were prepared and stained using Giemsa stain and then examined for the presence of Plasmodium spp. The concentration of infected red blood corpuscles was expressed as parasites per 200 white blood cells. Height and weight were measured to the nearest 0.1 cm and 0.1 kg, respectively, using a stadiometer ('Leicester' Model, Children's Growth Foundation) and electronic scales (Soehnle). Z-scores of height for age and weight for age were calculated using NCHS reference standards and EpiInfo version 6.0 software. A socioeconomic score was developed by scoring children for reported family possession (1) or lack (0) of the following items: house construction composing concrete walls, concrete floors and tin roofs, latrine, bicycle, and radio. In addition children were also scored for presence or absence of shoes at interview. A socio-economic score was formed by summing the scores (maximum 7, minimum 0).


Fourteen to 15 weeks after the start of treatments, children were asked to provide stool and urine samples again for parasitological examination and 16 weeks after treatment were weighed, measured for height and asked to give a venous blood sample to estimate concentration of haemoglobin and to estimate the intensity of infection with Plasmodium spp. Children then received antihelminthic treatment as necessary and all children with a haemoglobin level < 120 g/l were given 28 tablets of 200 mg ferrous sulphate.

Data entry and statistical methods

Double data entry was done using FoxPro (Microsoft, USA) and analysed using SPSS software. Because most variables were not normally distributed, χ2 tests were used to compare prevalence and Kruskal–Wallis tests to compare differences between continuous variables. Correlations were calculated using Spearman rank correlation. Tables show arithmetic means of variables with standard deviations.


In total, 261 girls volunteered to participate in the baseline survey and of these 235 completed all baseline measurements. A total of 234 girls met the inclusion criteria and were recruited to the study. Of the girls included in the study, 119 (51% of those recruited) received 12 doses of iron or control treatment and completed all follow-up measures. The remaining 115 (49% of those recruited) either dropped out of the study or due to absences failed to receive all 12 doses of treatment. Girls who completed the study did not differ from those who dropped out with respect to any of the variables measured at baseline. Data for serum ferritin and acute phase protein concentrations were available for 107 of the girls who completed the study. Girls for whom these data was available did not differ from others with respect to any of the variables measured.

The mean age of both treatment groups was 13.8 years (SD = 1.5 years) and did not differ significantly (P= 0.93). Helminth infections were common (hookworm 81%, T. trichiura 57%, A. lumbricoides 22%, S. haematobium 56%). No significant differences existed at baseline between groups in terms of estimated prevalence or intensity of infection. Nor did the groups differ significantly on decline of these variables following de-worming. The two groups did not differ with respect to socio-economic status, having similar mean scores (treatment score = 3.05, control score = 3.19 P= 0.89).

A highly significant difference was observed with respect to change in serum ferritin concentration, that of the treatment group increasing more than that of the control treatment (Table 1). Although it is clear that ferritin concentrations correlated positively with acute phase protein concentrations (Spearman Rank correlations: C-reactive protein = 0.39 = 107 P= 0.00001; α1-glycoprotein = 0.26 = 107 P= 0.007), the effect of supplementation on ferritin concentrations was observed whether children with inflammation (indicated by C-reactive protein > = 5 mg/l and/or α1-glycoprotein ≥ 0.7 g/l) were excluded from analysis or not. Mean haemoglobin concentrations did not differ between the treatment groups at either baseline or follow-up and there was no difference in change in haemoglobin during the study (Table 2).

Table 1.  Change in mean serum ferritin ± SD (μg/l) (all children irrespective of inflammation as indicated by acute phase protein concentrations) Thumbnail image of
Table 2.  Change in mean haemoglobin (g/l) ± SD Thumbnail image of

Girls given iron were found to put on more weight during the study than girls given control treatments (Table 3). There was no difference between groups at either baseline or follow-up with respect to anthropometric indices or height.

Table 3.  Change in mean weight (kg) ± SD Thumbnail image of

Treatment groups did not differ with respect to prevalence of infection with P. falciparum during the study (Table 4). At baseline, treatment groups showed a trend towards difference with respect to parasitaemia (P= 0.06). Change in parasitaemia differed significantly between groups during the study (P= 0.01). Acute phase protein concentrations did not differ between groups during the study (Table 5).

Table 4.  Prevalence (%) and intensity (parasites/200 wbc) of infection with P. falciparum Thumbnail image of
Table 5.  Change in mean acute phase proteins (C-reactive protein and a1-glycoprotein) ± SD Thumbnail image of


Serum ferritin levels of girls given weekly iron increased significantly more (+ 50%) than those of girls given control treatments (+ 18%). This suggests that weekly iron supplementation may be a highly effective means of raising iron stores in a community vulnerable to the development of iron deficiency. Similar results have recently been described among Malaysian schoolgirls (Tee et al. 1999). Raising iron stores may be of benefit for a number of reasons. Most girls included in our study were of the age when first pregnancies occur in Tanzania and raising stores could both protect them from iron deficiency during pregnancy and have a beneficial effect on fetal development. Iron deficiency has also been associated with low cognitive function and educational achievement, reduced physical fitness and immunity to infection (reviewed by Nokes et al. 1998). Raising iron stores could reduce the incidence of such problems.

While an increase was observed in the mean haemoglobin concentrations of both treatment groups, no difference was observed between treatment and control. This is in marked contrast to several other studies of weekly supplementation (Angeles-Agdeppa et al. 1997; Berger et al. 1997; Palupi et al. 1997). Our results highlight the need to include a control group in order to draw valid conclusions. Lack of control has made difficult the interpretation of several studies which have observed an increase in haemoglobin on weekly iron supplementation (Hallberg 1998).

Since at baseline 73% of our girls were not anaemic (Hb < 110 g/l), the lack of an effect of supplementation on haemoglobin is unsurprising. Similarly, haemoglobin levels of non-anaemic Malaysian schoolgirls were also unaffected by weekly iron-folate supplementation (Tee et al. 1999). Our results suggest that the provision of weekly supplementary iron did not increase haemoglobin levels in our population. This could be attributed to coexisting micronutrient deficiencies or problems with metabolism of iron due to chronic infections such as malaria which is holo-endemic in the study area (Lyimo et al. 1991). In a placebo-controlled trial in the Gambia, another area of holo-endemic malaria, Bates et al. (1987) found no haematological effect of giving a twice weekly multinutrient drink containing iron (100–200 mg), thiamine, riboflavin and vitamin C to 5–14-year-old children during a period of 3 months. In our study, the failure of supplementation to affect haemoglobin levels might also be attributed to the fact that the end of the intervention period occurred at a time of high dietary iron availability, coinciding both with the main harvest, when food is plentiful, and also with the orange harvest.

In our study, supplementation led to a significant increase in weight gain. Previous studies have shown that daily supplementation with iron is associated with increased growth (Chwang et al. 1988; Latham et al. 1990) and improved appetite (Lawless et al. 1994). In our study, it is likely that follow-up occurred too early to detect any potential gains in height.

A potential concern of giving iron is the possibility that supplementation may increase the risk of infection with malaria (Oppenheimer et al. 1986; Smith et al. 1989). Other studies (Bates et al. 1987; Harvey et al. 1989) have not demonstrated such an effect. In our study, a significant difference was observed with respect to change in parasitaemia (Table 4). However, since the groups were quite dissimilar at baseline (P= 0.06), it is not possible to determine whether weekly supplementation had any significant effect on parasitaemia. Levels of acute phase proteins did not differ between groups during the study suggesting that supplementation did not lead to any increase in inflammation. Data were not available to determine if supplementation had any clinical impact.

The results of this study imply that weekly iron supplementation could provide a highly effective means of raising the iron stores of adolescent girls. By protecting them from the cognitive effects of iron deficiency, weekly iron supplementation could help girls to get the most out of school. As well as being members of a population vulnerable to iron deficiency, the schoolgirls included in the study were of an age when pregnancies would be expected to begin to arise (Partnership for Child Development 1998). Supplementation could also help both girls and their babies achieve improved health. The study has shown for the first time that weekly iron supplementation can have an effect on growth. These benefits, attributable to iron supplementation, occurred at the same time as girls were treated for helminth infections through the activities of the Tanzanian Partnership for Child Development. It suggests that weekly iron supplementation could produce effects over and above those already attributed to school based mass de-worming programmes (Beasley et al. 1999).


We are grateful for the support of the Wellcome Trust and the Department for International Development (UK). These results and interpretations do not necessarily reflect their policies. We should also like to thank Juana Willumsen and Suzanne Filteau of the Institute of Child Health, London and laboratory staff of the Tanzania Food and Nutrition Centre for assays of serum ferritin and acute phase proteins.