European Journal of Clinical Nutrition (2006), 1–8

& 2006 Nature Publishing Group All rights reserved 0954-3007/06 $30.00 www.nature.com/ejcn

ORIGINAL ARTICLE

Prevalence of iron deficiency and anemia among healthy women of reproductive age in Bhaktapur, Nepal RK Chandyo1, TA Strand2, RJ Ulvik3, RK Adhikari1, M Ulak1, H Dixit4 and H Sommerfelt2 1 Department of Child Health, Institute of Medicine, Tribhuvan University, Kathmandu, Nepal; 2Centre for International Health, University of Bergen, Bergen, Norway; 3Institute of Medicine, Haukeland University Hospital, University of Bergen and Laboratory of Clinical Biochemistry, Norway and 4Department of Pediatrics, Kathmandu Medical College, Kathmandu University, Kathmandu, Nepal

Objective: To determine the prevalence of anemia and iron status as assessed by biochemical markers and to explore the associations between markers of iron status and iron intake. Study area and population: Five hundred healthy women of reproductive age from the Bhaktapur district of Nepal were included in the study. Methods: A cluster sampling procedure was applied for this cross-sectional study. Women without any ongoing infection aged 13–35 years were selected randomly from the population. We measured the plasma concentration of hemoglobin (Hb), ferritin and transferrin receptors. Dietary information was obtained by a food frequency questionnaire and two 24-h dietary recalls. Results: The prevalence of anemia (Hb concentration o12 g/dl) was 12% (n ¼ 58). The prevalence of depleted iron stores (plasma ferritin o15 mg/l) was 20% (n ¼ 98) whereas the prevalence of iron deficiency anemia (anemia, depleted iron stores with elevated transferrin receptor i.e. 41.54 mg/l) was 6% (n ¼ 30). Seven percent (n ¼ 35) of women were having irondeficient erythropoiesis (depleted iron stores and elevated transferrin receptor but normal Hb). Out of the 58 anemic women, 41 (71%) and 31 (53%) were also having elevated plasma transferrin receptor and depleted iron stores, respectively. Fifty-four percent of the women ate less than the recommended average intake of iron. The main foods contributing to dietary iron were rice, wheat flour and green and dry vegetables. Conclusions: The prevalence of anemia in our study was substantially lower than the national figure for non-pregnant women. Only about half of the women with anemia were also having depleted iron stores, suggesting that other causes of anemia may be prevalent in this population. Sponsorship: Norwegian Universities Committee for Development, Research and Education (NUFU).

European Journal of Clinical Nutrition advance online publication, 23 August 2006; doi:10.1038/sj.ejcn.1602508 Keywords: iron deficiency; anemia; plasma transferrin receptor; non-pregnant women; iron intake; Nepal

Introduction

Correspondence: Dr TA Strand, Centre for International Health, University of Bergen, Armauer Hansen Building, N-5021 Bergen, Norway. E-mail: [email protected] Guarantor: TA Strand. Contributors: All authors contributed to the study design and in preparation of the manuscript. The first draft of the manuscript was written by RKC. TAS was the principle investigator of the study and HS was the overall project coordinator. RKA and HD were the Nepalese coordinators. MU was the main study physician and supervised the fieldwork and data collection. RU supervised the laboratory analysis of iron parameters and its interpretation. Received 19 September 2005; revised 30 June 2006; accepted 6 July 2006

Iron deficiency (ID) is one of the leading nutritional deficiencies in the world, particularly in developing countries (DeMaeyer and Adiels-Tegman, 1985; WHO, 1998). When ID is sufficiently severe, the hemoglobin (Hb) concentration in the blood decreases, leading to iron deficiency anemia (IDA), which has negative health consequences, especially in children (Lozoff et al., 1991), adolescents (Bruner et al., 1996) and pregnant women (Scholl et al., 1992; Allen, 2000; Hinton et al., 2000). A substantial proportion of women in developing countries,

Prevalence of iron deficiency and anemia among healthy women RK Chandyo et al

2 such as Nepal, enter pregnancy with inadequate iron stores. Women with inadequate stores are at increased risk of developing IDA during pregnancy. Early detection and treatment of sub-clinical ID among non-pregnant women may play a pivotal role in preventing development of IDA during pregnancy. Studies from South-Asian countries (Agarwal et al., 1987; WHO, 1998; Kapil et al., 1999; Bondevik et al., 2000; Dreyfuss et al., 2000) indicate that two-thirds of pregnant women suffer from IDA, which is the highest prevalence in the world. Most studies of IDA have been undertaken in pregnant women or hospitalized patients, where the markers were influenced by physiological hemodilution or metabolic and pathophysiological changes. Pre-pregnancy evaluation provides information that may prove critical for the development of effective intervention strategies against anemia and IDA in developing countries. The aim of this cross-sectional study, in a random sample of non-pregnant women from an agriculture-based urban community in Nepal, was to determine the prevalence of anemia and ID. Furthermore, we also estimated the intake of iron and identified the main dietary sources of iron.

sampling units. A total of 23 Toles and five CF were randomly selected from the municipality. We obtained a list of all 2 736 women between 13 and 35 years living in these selected clusters (Figure 1). Seven hundred and ninety-two women were then selected randomly from these clusters. Two hundred and ninety-two women whom we approached could not be enrolled, mainly because they had moved (167 women), 141 of whom were from the CF. The women were weighed using an UNICEF electronic scale (SECA, Hamburg, Germany) with an accuracy of 100 g. A locally made height board measured height to the nearest 0.1 cm.

Blood sampling and storage Blood was collected from the cubital veins at least 2 h after any food intake and placed into micronutrient-free heparinized polypropylene tubes (Sarstedt, Germany). The Hb concentration was analyzed immediately following blood ˚ ngelholm, Sweden). The heparsampling with Hemocue (A inized blood was centrifuged (760 g, 10 min, room temperature), separated and transferred into micronutrient-free polypropylene vials (Eppendorf, Hinz, Germany). These vials were stored at 451C in Nepal until transfer to Norway.

Subjects and methods Study area The Kathmandu valley includes the districts of Kathmandu, Lalitpur and Bhaktapur. Bhaktapur, the study site, has a population of approximately 80 000, and the majority of the people work in agriculture. The study site includes an urban area where most of the population has a relatively low income. The indigenous people of Bhaktapur municipality are the Hindu/Buddhist Newar. Around the city of Bhaktapur, there are approximately 50 carpet factories (CF) in which migrant families live and work for various lengths of time. The workers belong mostly to the Tamang and Magar ethnic groups and come from various regions of Nepal. The study included women aged 13–35 years that were non-pregnant, without any ongoing disease, who lived in Bhaktapur municipality, and who gave informed consent. We excluded women with acute or chronic illness and those taking vitamins, minerals or drugs (with the exception of contraceptive pills or injections). The Nepal Health Research Council in Katmandu and the Human Research Ethical Committee of the Medical Faculty at the University of Bergen, Norway approved the study. The implementation of all aspects of the project was in agreement with the International Ethical Guidelines for research involving human subjects, as stated in the latest version of the Helsinki Declaration.

Subjects and selection procedures There are 128 neighborhoods called ‘Toles’ in the municipality. We used the Toles or the CF as the clusters or primary European Journal of Clinical Nutrition

Biochemical analyses of markers of iron status and of C-reactive protein The blood samples were analyzed at the Laboratory of Clinical Biochemistry, Haukeland University Hospital in Bergen, Norway. The concentration of soluble plasma transferrin receptors (P-TfR) was quantified by particleenhanced nephelometry using polystyrene particles coated with a monoclonal antibody in a Behring Nephelometer Analyzer II (Dade Behring Marburg GmbH, Germany). The plasma ferritin (P-Ferritin) was analyzed by a turbidimetric immunoassay and C-reactive protein (CRP) was analyzed by immunoassay in a Roche/Hitachi Modular Analyzer, Roche Diagnostics (Roche, Mannheim, Germany).

Definitions The stages of negative iron balance with the respective cutoff limits for the analytical tests used in this study, are shown in Table 1. Depleted iron stores, which is an early sign of developing ID, was defined as P-Ferritin o15 mg/l, which corresponds to absence of stainable iron in bone marrow smears (Hallberg et al., 1993). Iron-deficient erythropoiesis (IDE) was defined as normal Hb concentration with depleted iron stores combined with a elevated P-TfR 41.54 mg/l which is the 97.5 percentile derived from a healthy Norwegian female population. According to the World Health Organization (WHO), Hb concentration o12 g/dl in non-pregnant women is defined as anemia (CDC/MMWR 1998). IDA was defined as depleted iron stores and elevated P-TfR concurrent with anemia.

Prevalence of iron deficiency and anemia among healthy women RK Chandyo et al

3

Figure 1

Selection procedure and study profile of ID prevalence study in Bhaktapur, Nepal.

Table 1 Stages of negative iron balance with the respective cutoff limits of the analytical tests among 500 women in Bhaktapur, Nepal Stages

Normal iron status Anemia Depleted iron stores Iron-deficient erythropoiesis Iron deficiency anemia

Hemoglobin Plasma transferrin Plasma (g/dl) receptor (mg/l) ferritin (mg/l) X12.0 o12.0 — X12.0 o12.0

p1.54 — — 41.54 41.54

415 — o15 o15 o15

Anemia of chronic disease, caused by functional deprivation of iron in the bone marrow, was considered to be present when iron was withheld in the stores with normal P-TfR but P-Ferritin 450 mg/l (Wordwood, 1980). As the study area is located about 1400 m above sea level, a more relevant cutoff value would be Hb o12.3 g/dl (CDC/MMWR, 1998). We therefore also report data based on a cutoff of Hb adjusted for the altitude.

Twenty – four hour dietary recall and food frequency questionnaires The dietary recall methods were included only after enrolment of the first 106 women in the study. Therefore, the information on dietary intake was obtained from the last 394 women. Two 24-h dietary recalls were obtained from each woman, approximately 1 week apart. The female field workers were trained to undertake the dietary recall interviews before the study. The first recall was carried out at the clinic, whereas for the second recall, the field workers visited the women in their homes. Models of local foods were used during the interviews to estimate the portion sizes. Three hundred and seventy-nine women (96%) of the 394 women completed two 24-h dietary recalls. The analysis of nutrient intake was restricted to these 379 women, using the mean intake of the two recalls. For the food frequency questionnaire (FFQ), a total of 53 locally consumed food items were listed, and the women were requested to recall how often they consumed each of them. The frequency of consumption was categorized into never in the past 6 months, 1–5 European Journal of Clinical Nutrition

Prevalence of iron deficiency and anemia among healthy women RK Chandyo et al

4 times in the past 6 months, 1–3 times a month, once a week, 2–4 times a week and once or more times per day. The study physician filled in the questionnaires. The daily intakes of the various nutrients from the 24-h dietary recalls were calculated using Indian food tables from the Wfood2 program version 1.0 (WorldFood, 1996). The total iron contribution by the different foods was derived from the nutritive values of the 24-h dietary recalls, whereas the frequency of reported consumption was derived from the FFQ. The estimated average requirement of iron for nonpregnant women of reproductive age is 8.1 mg/day (NAP, 2003).

Table 2 General characteristics of 500 women included in a study evaluating iron status in Bhaktapur, Nepal Variables Mean age, years (s.d.) Mean age of menarche, years (s.d.) Numbers of women with one or more para (%) Numbers of women who smokes (%)a Numbers of vegetarians (%)a Mean family size (s.d.) Mean BMI (s.d.), kg/m2 Numbers of women with BMI o18.5 kg/m2 Numbers of illiterate women (%) Numbers of women using contraceptives (%)b

Value 23 14.5 268 26 10 6.5 21.8 55 166 209

(5.8) (2.2) (54) (7) (3) (3.2) (3.0) (11) (33) (71)

a

Data from 394 women from whom we obtained dietary recalls. Among 296 married women.

b

Statistical analysis The Hb concentrations as well as the intake of iron and energy were normally distributed. The P-Ferritin and P-TfR, however, were skewed to the higher values. The descriptive statistics included means, medians, inter quartile ranges and 95% confidence intervals. The Hb concentration was used as a dependent variable in simple and multiple linear regression analyses to determine the association with the relevant explanatory variables. The statistical analyses were undertaken using Stata, version 8 (STATA Corp, Houston, TX, USA) and the standard errors were adjusted for the design effect of the study design using standard STATA commands for analysis of survey data, such as svymean and svyreg (Etlinge and Sribney, 1996). The graphs describing the relationship between Hb level with P-TfR and iron intake were constructed using generalized additive models (GAM) in the statistical software R version 1.9.0 (Wood, 2000). Because adolescents are growing their iron requirements differ from adults. We therefore present the values for nutrient intake, indicators of iron status and the association between iron intake and Hb concentration separately for women p18 years and 418 years of age.

Results Subject characteristics The mean age of the 500 women was 23 years and 27% were p18 years old (n ¼ 136). Fifty-two percent of the women lived in joint families with a mean family size of 6.5 (Table 2). Two hundred and ninety-six (59%) women were married, and among these, 209 (71%) used contraceptives, of these 140 (47% of the married women) used Depo-Provera (depotmedroxyprogesterone acetate) injections. One of the spouse in 38 (13%) of the married couples was sterilized. The mean age of menarche was 14.5 years. Forty-six percent of the women were nullipara, 13% were para 1, 24% para 2 and 17% were para X3. Three hundred and thirty-four (67%) women said that they could read and write, whereas among the married women, 258 (87%) reported that their husbands could do so. European Journal of Clinical Nutrition

Nutrient intake Seventy-six percent of the women had energy intakes less than the recommended dietary allowances of 2200 Kcal. The mean intake and the corresponding 95% confidence intervals of iron, heme iron and vitamin C intakes are presented in Table 3. The interquartile range for iron intake was 6.0– 10.3 mg and 54% of the women consumed inadequate amounts of iron. One-third of the total dietary recalls did not include any animal products. The main foods providing iron and their frequency of intake are shown in Table 4.

Normal iron status and prevalence of ID The mean (s.d.) Hb, P-Ferritin and P-TfR were 13.2 g/dl (1.3), 35.2 mg/l (30.3) and 1.6 mg/l (0.86), respectively (Table 5). Among the total 499 blood samples, which were adequate for the analysis of iron parameters, 53% (n ¼ 264) were from women with normal iron status. The 2.5th–97.5th percentile ranges among these women for Hb, P-Ferritin and P-TfR was 12.2–15.2 g/dl, 16.6–93.5 mg/l and 0.89–1.5 mg/l, respectively. Depleted iron stores were found in 20%, whereas IDE was found in 7%. The prevalence of anemia was 12%, whereas the combination of anemia and depleted iron stores was found in 6% (n ¼ 31). The prevalence of anemia was 16% (n ¼ 79), when we used altitude adjusted cutoff of Hb (o12.3 g/dl). The combination of anemia plus elevated P-TfR concentration was identified in 8% (n ¼ 41). The majority of the women with depleted iron stores (68%) or with elevated P-TfR (78%) were not anemic. Severe anemia (o7 g/dl) was found in only one woman, thus most of the anemia in this population was mild or moderate. Forty-eight percent of anemic women were below 20 years of age. Only five anemic women (1%) had P-Ferritin 450 mg/l but normal P-TfR, indicating anemia of chronic disease. According to our definition, the prevalence of IDA was 6% (n ¼ 30). The GAM plot depicting the relationship between the P-TfR and Hb concentration is shown in Figure 2. The graph shows that there is a linear and negative correlation between

Prevalence of iron deficiency and anemia among healthy women RK Chandyo et al

5 Table 3 Mean intake of energy, protein, fat, iron and vitamin C among women of reproductive age in Bhaktapur, Nepal Variable

RDA

Mean or percent (95% CI) Age p18 (years) (n ¼ 86)

Energy (kcal) Protein (g) Fat (g) Iron (heme and non-heme) (mg) % o8.1 mga Heme iron (mg) Vitamin C (mg)

2200 50 20 8.1

1694 46 13 7.6 63 0.36 42

Age 418 (years) (n ¼ 293)

(1608, 1781) (42, 49) (12, 15) (6.9, 8.3) (52, 73) (0.21, 0.52) (42, 56)

1984 54 17 8.6 51 0.61 48

All (n ¼ 379)

(1925, 2042) (52, 56) (15, 19) (8.2,9.0) (45, 57) (0.48, 0.73) (45, 52)

1918 52 16 8.4 54 0.55 48

(1867, 1969) (50, 54) (15,18) (8.1, 8.7) (49, 59) (0.45, 0.65) (45, 51)

Abbreviations: CI, confidence interval. a Estimated average requirement for iron set by the National Academy of Science, USA.

Table 4 Main sources of iron in Bhaktapur, Nepal based on 24 h dietary recalls Foods

Iron contribution (%)

Frequency of consumption Number of recalls with particular foods (%)a

Rice grain, flake or flour Green or dry vegetables (mustard, radish, spinach etc.) Wheat grain or flour Pulses or beans (lentil, black and red gram) Buffalo meat Local beer

31% 29% 20% 6% 5% 2%

758 517 314 294 142 139

(100%) (68%) (41%) (39%) (19%) (18%)

Number of women who reported to consume at least once a week (%)b 394 319 255 256 159 118

(100%) (81%) (65%) (65%) (40%) (30%)

Iron content (mg) per 100 gc

0.6–1.5 0.8–3.6 2.8–3.9 1.5–3.3 1.3 0.3

a

Total 758 dietary recalls (two recalls from each woman). Data based on food frequency questionnaires from 394 women. c Data from Indian and Nepali food tables in Wfood2 program for the analysis of nutritive values. b

Table 5 Mean hemoglobin, plasma ferritin and transferrin receptors Variable

Mean or n (s.d. or percent) Age p18(years) (n ¼ 136)

Mean hemoglobin level g/dl (s.d.) Number of women with hemoglobin o12 g/dl (%) Mean plasma ferritin, mg/l (s.d.) Number of women with ferritin o15 mg/l (%) Mean plasma transferrin receptor, mg/l (s.d.) Number of women with transferrin receptor 41.54 mg/l (%)

P-TfR and Hb concentration when the transferrin receptor is more than approximately 1.5 mg/l. The GAM plot of iron intake and Hb concentration is shown in Figure 3. The Hb concentration was negatively and linearly associated with the iron intake when the iron intake was less than the recommended intake (8.1 mg/day).

Multiple regression analysis For women 13–18 years of age, working in CF strongly predicted the Hb concentration (Table 6). In the analyses that included older women, using Depo-Provera contraceptives was associated with a substantially higher Hb

12.9 25 26.9 40 1.9 56

(1.5) (18%) (35.6) (29%) (1.31) (41%)

Age 418 (years) (n ¼ 364) 13.4 33 38.3 58 1.5 130

(1.2) (9%) (27.5) (16%) (0.60) (36%)

All (n ¼ 500) 13.2 58 35.2 98 1.6 186

(1.3) (12%) (30.3) (20%) (0.86) (37%)

concentration. Women with parity of X2 also had higher mean Hb concentration than the nulliparous women. Also in this age group, women working in CF had a substantially lower mean Hb concentration than the local resident women.

Prevalence of acute inflammation The prevalence of acute inflammation indicated by elevated CRP concentration (410 mg/l) was 2.2%. Among the 11 women with elevated CRP, eight women also had P-Ferritin levels 415 mg/l. The CRP was positively correlated with P-Ferritin in the Spearman’s correlation analysis (r ¼ 0.20, European Journal of Clinical Nutrition

Prevalence of iron deficiency and anemia among healthy women RK Chandyo et al

6

Figure 2 Graph showing the relation between Hb and P-TfR among 500 women of reproductive age in Bhaktapur, Nepal. Graph has been made using a GAM in R (Wood S, 2000), The solid line depict the regression line form the GAM analysis, the area between the broken lines represent the 95% confidence interval (CI) of this regression line. The small bars on the x axis show the distribution of individual observations and the vertical lines indicate the 25th, 50th and 75th percentiles of the P-TfR concentration.

Pp0.0005) but was not associated with P-TfR (r ¼ 0.05, P ¼ 0.26) or Hb levels (r ¼ 0.04, P ¼ 0.31).

Discussion In this cross-sectional study from randomly selected women, we found that the prevalence of anemia and IDA was lower than the WHO estimates for non-pregnant women in South Asia (WHO, 2000). The prevalence of anemia among adolescent and non-pregnant women in Nepal has been estimated to be B60% (Tiwari and Seshadri, 2000; Shah and Gupta, 2002) with even higher figures among pregnant women (Bondevik et al., 2000; Dreyfuss et al., 2000). In contrast to most of the studies mentioned above, the women included in this study were selected randomly from the community. However, our attempts to obtain a representative sample were compromised by frequent migration of the women from CF. As more than half of the screened women from this stratum moved away (n ¼ 129), we had to recruit women of the same age that were not initially part of the survey. Therefore, this stratum may not constitute a perfect random sample of the population from CF. In the other stratum that consisted of 80% of the women, however, the random sampling procedure worked well. Furthermore, the age distribution of non-respondent women was not different from those who responded. The sub-sample of women from whom we obtained dietary information did not differ from the total sample with regard to iron status (data not shown). European Journal of Clinical Nutrition

Figure 3 Graphs showing the relations between Hb and iron intake among women 13–18 and 418 years of age in Bhaktapur, Nepal. The graphs were made from GAM in R (Wood S, 2000). The solid lines depict the regression lines form the GAM analyses, the area between the broken lines represent the 95% CI of these regression lines. The small bars on the x axis show the distribution of individual observations and the vertical lines indicate the 25th, 50th and 75th percentiles of iron intake.

The lower prevalence of anemia in our study compared to the national estimate (NFHS, 1996; WHO, 2000) may partly be explained by the fact that we enrolled apparently healthy women living in an urban setting, which is not endemic for malaria, hookworm infestation and hemoglobinopathies. We found less than 3% of the study participants to be vegans, which is less than in the general population of Nepal. The 11% prevalence of chronic energy deficiency (body mass index o18.5 kg/m2) in our study was also lower than that observed in the Nepal Family Health Survey 1996, which revealed a figure of 28.3% (NFHS, 1996). The prevalence estimates of ID and IDA vary depending upon the cutoff values of the laboratory tests used for diagnosis (Dallman et al., 1981; Gibson, 1990). As pointed out by Hallberg et al. (1993) the commonly used cutoff value of 12 mg/l for P-Ferritin, was established before an international calibration standard for the ferritin assay was available. In their extensive study on ID in women, using an assay calibrated against an international standard, they found that P-Ferritin o15 mg/l was the best cutoff value with a diagnostic efficiency of 91–95% for correct classification of depleted iron stores, compared to 86%, when 12 mg/l was

Prevalence of iron deficiency and anemia among healthy women RK Chandyo et al

7 Table 6 Multiple linear regression analysis for the association of age, Depo-Provera uses, parity, illiteracy and women working in carpet factories with hemoglobin concentrations in 500 women of Bhaktapur, Nepal Age p18 (years)

Variables Coeff Age (continuous) Working in carpet factories No Yes Use of Depo-Provera No Yes Parity 0 1 2 X3 Illiterate No Yes

Age 418 (years) 95% CI

Coeff

95% CI

(0.18, 0.19)

0.02

(0.06, 0.22)

1 1.24

— (1.80, 0.69)*

1 0.54

— (0.87, 0.21)**

— —

— —

1 0.45

— (0.17, 0.73)**

— — — —

— — — —

1 0.07 0.48 0.61

— (0.32, 0.47) (0.07, 0.89)** (0.07, 1.1)**

— —

— —

1 0.21

— (0.06, 0.48)

0.004

Abbreviations: Coeff, coefficient; CI, confidence interval; Depo-Provera, depot-medroxyprogesterone acetate. *Po0.0005, **Po0.05.Adjusted regression coefficients. The models included age, working in carpet factories, use of Depo-Provera, parity and illiteracy. The latter three variables were not included in the models with women o18 years of age.

used as cutoff. Moreover, the cutoff of 15 mg/l for P-Ferritin is also recommended by the Center for Disease Control, USA (CDC/MMWR, 1998). The smoothed curve depicting Hb versus P-TfR (Figure 2) indicates that at the population level the cutoff value for P-TfR determined from our Norwegian female reference population, was appropriate. ID is usually defined by a multiple criteria model, in which the presence of X2 iron status markers beyond the cutoff values is considered to be indicative of ID and IDA is simply when anemia occurs concurrently (Cook et al., 1976). ID is often described as a clinical stage, which includes the two stages of depleted iron stores and IDE. However, according to a strict pathophysiological definition, ID should be restricted to the stage of IDE, which is followed by IDA. If IDA were defined as anemia with low P-Ferritin, the prevalence would have been identical with our findings, where we also included P-TfR level in the definition. This suggests that inclusion of Hb and P-Ferritin might be sufficient to define IDA. Our results indicate that the prevalence of elevated P-TfR concentration was higher than that of depleted iron stores, which is in accordance with a study from the Ivory Coast (Asobayire et al., 2001). This apparent inconsistency may be explained by the fact that IDE appears before the iron stores are completely emptied, at P-Ferritin levels of 30–40 mg/l, as shown by Hallberg et al. (Hallberg et al., 1993). The P-TfR concentration is usually increased owing to ID, leading to insufficient supplies of iron to the bone marrow. Alternatively, it may be due to a hyperproliferative production of red blood cells caused by various other reasons (Punnonen et al., 1997). IDE, as judged from high P-TfR and low P-Ferritin levels, can be regarded as a temporary, intermediate step in

the pathophysiological process leading to IDA. In this case, the iron stores are depleted which causes inadequate iron supply to the erythropoetic cells of the bone marrow. Then, the existing population of normal erythrocytes in the circulation will slowly be replaced by microcytic and hypochromic erythrocytes and anemia will gradually appear. The prevalence of anemia and IDA were 12 and 6% respectively, indicating that only about half of the anemia was owing to ID. The high prevalence of anemia among women without depleted iron stores (47%) suggests that other anemia-causing nutritional deficiencies may be prevalent in this population. This is supported by the high prevalence of elevated P-TfR concentration (37%), which could be explained by hyperproliferative production of red blood cells owing to deficiencies of micronutrients like folate, vitamin A and vitamin B12 (Suharno et al., 1992; Bondevik et al., 2000). Our results also showed that the use of Hb measurement as a proxy for ID and IDA clearly underestimates other causes of anemia. The use of hormonal contraception has also previously been found to be associated with iron status in women of the reproductive age (UNDP, 1998). In the present study, we found that women who used Depo-Provera injections had substantially higher concentrations of Hb, which is in accordance with another study (Milman et al., 1992). Although menstrual disturbance (irregular bleeding, spotting and amenorrhea) after Depo-Provera is a well-known side effect (Belsey, 1991), published data on its effect on Hb concentration and anemia is limited. The high (47%) proportion of women using Depo-Provera in the present study corresponds to the national figure for Nepal (HMG, 2001). Three fourths of all Depo-Provera users in our study were X2 para and we have information indicating that a European Journal of Clinical Nutrition

Prevalence of iron deficiency and anemia among healthy women RK Chandyo et al

8 substantial proportion of women in this setting take iron and folic acid supplements during pregnancy and lactation, which is in accordance with the recommendations by the national health authorities (District Public Health Office, Bhaktapur, Personal communication). The positive correlation between age and Hb (r ¼ 0.19, Pp0.0005) and the high coverage of iron supplementation during pregnancy and lactation may explain our findings of a positive relation between parity and Hb concentration. In conclusion, the prevalence of anemia and IDA in our study population was lower than the Nepali national figures. The use of Depo-Provera, working in CF and parity were the major predictors of Hb concentration. IDA accounted only about half of the total anemia; therefore, other causes are likely to be important in the development of anemia in this population.

Acknowledgements We thank all the women who participated in this study. We are also very grateful for the contribution of field, computer and administrative staff of the Child Health Research Project in the Pediatric Department, Institute of Medicine, Kathmandu. We thank the staff and the founder Shyam Sunder Dhaubhadel of the Siddhi Memorial Hospital in Bhaktapur for their cooperation. Finally, we thank the staff at the Laboratory for Clinical Biochemistry at Haukeland University Hospital, in particular Tone Minde and Hans Henriksen for their proficient processing of the plasma specimens.

References Agarwal DK, Agarwal KN, Tripathi AM (1987). Nutritional status in rural pregnant women of Bihar and Uttar Pradesh. Ind Pediatr 24, 119–125. Allen LH (2000). Anemia and iron deficiency: effects on pregnancy outcome. Am J Clin Nutr 71 (5 Suppl), 1280S–1284S. Asobayire FS, Adou P, Davidsson L, Cook JD, Hurrell RF (2001). Prevalence of iron deficiency with and without concurrent anemia in population groups with high prevalences of malaria and other infections: a study in Cote d’Ivoire. Am J Clin Nutr 74, 776–782. Belsey EM (1991). Menstrual bleeding patterns in untreated women and with long-acting methods of contraception. Task force on long-acting systemic agents for fertility regulation. Adv Contracept 7, 257–270. Bondevik GT, Eskeland B, Ulvik RJ, Ulstein M, Lie RT, Schneede J et al. (2000). Anaemia in pregnancy: possible causes and risk factors in Nepali women. Eur J Clin Nutr 54, 3–8. Bruner AB, Joffe A, Duggan AK, Casella JF, Brandt J (1996). Randomised study of cognitive effects of iron supplementation in non-anaemic iron-deficient adolescent girls. Lancet 348, 992–996. CDC/MMWR (1998). Centers for Disease Control. CDC, USA. Criteria for anaemia in children and child bearing aged women. CDC MMWR 1998. http://www.cdc.gov/mmwr//PDF/rr/rr4703. pdf. Cook JD, Finch CA, Smith NJ (1976). Evaluation of the iron status of a population. Blood 48, 449–455.

European Journal of Clinical Nutrition

Dallman PR, Reeves JD, Driggers DA, Lo EY (1981). Diagnosis of iron deficiency: the limitations of laboratory tests in predicting response to iron treatment in 1-year-old infants. J Pediatr 99, 376–381. DeMaeyer E, Adiels-Tegman M (1985). The prevalence of anaemia in the world. World Health Stat Q 38, 302–316. Dreyfuss ML, Stoltzfus RJ, Shrestha JB, Pradhan EK, LeClerq SC, Khatry SK, et al. (2000). Hookworms, malaria and vitamin A deficiency contribute to anemia and iron deficiency among pregnant women in the plains of Nepal. J Nutr 130, 2527–2536. Etlinge JL, Sribney WM (1996). Some basic concepts for design based analysis of complex survey data. Stata Tech Bull 31, 3–6. Gibson R (1990). Principle of Nutritional Assessment. Oxford University Press: Oxford. Hallberg L, Bengtsson C, Lapidus L, Lindstedt G, Lundberg P-A, Hulten L (1993). Screening for iron deficiency: an analysis based on bone-marrow examinations and serum ferritin determinations in a population sample of women. Br J Haematol 85, 787–798. Hinton PS, Giordano C, Brownlie T, Haas JD (2000). Iron supplementation improves endurance after training in iron-depleted, nonanemic women. J Appl Physiol 88, 1103–1111. HMG, Nepal (2001). His Majesty Government. National population census 2001, Central Bureau of Statistics HMG, Nepal. Kapil U, Pathak P, Tandon M, Singh C, Pradhan R, Dwivedi SN (1999). Micronutrient deficiency disorders amongst pregnant women in three urban slum communities of Delhi. Indian Pediatr 36, 983–989. Lozoff B, Jimenez E, Wolf AW (1991). Long-term developmental outcome of infants with iron deficiency. N Engl J Med 325, 687–694. Milman N, Kirchhoff M, Jorgensen T (1992). Iron status markers, serum ferritin and hemoglobin in 1359 Danish women in relation to menstruation, hormonal contraception, parity, and postmenopausal hormone treatment. Ann Hematol 65, 96–102. NAP (2003). Dietary References Intakes: Application in Dietary Planning. National Academy Press 2003. NFHS (1996). Nepal Family Health Survey, 1996, His Majesty’s Government of Nepal, Ministry of Health, Department of Health Service. Punnonen K, Irjala K, Rajamaki A (1997). Serum transferrin receptor and its ratio to serum ferritin in the diagnosis of iron deficiency. Blood 89, 1052–1057. Scholl TO, Hediger ML, Fischer RL, Shearer JW (1992). Anemia vs iron deficiency: increased risk of preterm delivery in a prospective study. Am J Clin Nutr 55, 985–988. Shah BK, Gupta P (2002). Anemia in adolescent girls: a preliminary report from semi-urban Nepal. Indian Pediatr 39, 1126–1130. Suharno D, West CE, Muhilal Logman MH, de Waart FG, Karyadi D, Hautvast JG (1992). Cross-sectional study on the iron and vitamin A status of pregnant women in West Java, Indonesia. Am J Clin Nutr 56, 988–993. Tiwari K, Seshadri S (2000). The prevalence of anemia and morbidity profile among school going adolesecent girls of urban Kathmandu, Nepal. J Nepal Med Assoc 39, 319–325. UNDP (1998). Task force for epidemiological research on reproductive health, UNDP, effects of contraceptives on hemoglobin and ferritin. Contraception 58, 262–273. WHO (2000). Nutrition in South-East Asia. New Delhi, World Health Organization. Regional Office for South-East Asia: New Delhi. WHO U (1998). World Health Organization, UNICEF & UNU Iron Deficiency: Indicators for Assessment and Strategies for Prevention 1998. World Health Organization: Geneva. Wood S (2000). Modelling and Smoothing Parameter Estimation with Multiple Quadratic Penalties. J R Statist Soc B 62, 413–428. Wordwood M (1980). Serum Ferritin. In: Jacob A, Wordwood M (eds) Iron in Biochemistry and Medicine, II. Academic Press: London. WorldFood (1996). WorldFood Dietary Assessment System. A Computer Software Package. The Regents of the University of California: Berkeley, CA.