Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

Haematological status of preschool and school-age children in urban and rural areas of Guatemala

María Eugenia Romero-Abal, Jesús Bulux, Iván Mendoza, Carlos Grazioso, and Noel W. Solomons

Abstract

We studied the prevalence of low haematocrit values (defined as <38%) in 1,253 children from urban and rural areas of Guatemala, to examine any urban-rural or age-related trends. Though the crude prevalences of low haematocrit for all the children showed a significant difference between urban and rural residents, the significance disappeared when these values were adjusted for differences in the age profiles of the two groups. As expected, preschool children had significantly more low haematocrits (32.0%) than school-age children (6.0%) (p < .05). Ferritin levels were available for 35.9% of the preschool children (one urban and one rural location); of these, 51.8% had levels below 12 mg/l, indicating iron deficiency. These values were used to determine the predictive value of haematocrit compared with ferritin values, and the cut-off at which haematocrit reaches optimum sensitivity and specificity to diagnose iron depletion. A cut-off of 39% had a sensitivity of 61% and a specificity of 45% in urban preschoolers, and a cut-off of 38% had a sensitivity of 75% and specificity of 42% in rural preschoolers.

Introduction

Urbanization is affecting developing and developed countries alike. In recent years, in developing nations, it has not necessarily been accompanied by industrialization and modernization. In fact, in tropical nations, it is often accelerated and out of control, with negative effects on health and nutrition status for both new urban migrants and the rural peasants left behind, due to their competition for natural resources, residential development of fertile arable land, and disruption of the agricultural labour force [1-5]. The consequences for both populations can be changes in availability of food, purchasing capacity, meal selection, and dietary variety [3, 6]. Although the preponderance of evidence shows nutrition status to be superior in urban compared with rural populations, the data are often not stratified or adjusted for socio-economic status. Rates of malnutrition are currently increasing faster in urban than in rural areas of developing countries [3-5].

The most widespread nutrition problem in the world is anaemia [7], affecting 10%-20% of the world population [8] but primarily pregnant women and preschool children in developing countries. The nature of the haematological deficiency in relation to urbanization in the Central American republic of Guatemala has not been examined. It was previously established that the majority of anaemia in this nation is related to iron deficiency [9]. Therefore, examination of the haematological status of a Guatemalan population should provide an index of its iron status.

Two diagnostic tests, haemoglobin concentration and haematocrit (packed-red-cell volume), are available for screening. Whereas the World Health Organization differentiated cut-off criteria for haemoglobin by age, sex, and physiological status [4], it established a single and universal cut-off point of <33% when using haematocrit. Since haematocrit (%) and haemoglobin (g/dl) roughly maintain a numerical relation of 3:1 [7], this raises a patent internal inconsistency. This is further complicated when the hypoxic effect of high altitude is considered, as a higher level of circulating red cells is maintained for a given iron reserve as one rises progressively above sea level [9, 11]. A criterion for 'anaemia" that is valid for a coastal population would not be valid in Guatemala City (1,500 m) or La Paz (3,700 m).

In recognition of this altitude effect, a correction scheme has been offered for shifting the haematocrit cut-off criterion with altitude [9, 12,] However, a poor degree of correspondence in diagnostic assessment was shown using one or the other format [13]. As a result, we chose to use a practical, and generous, cut-off of below 38% for what we consider low haematocrit to emphasize our uncertainty as to its diagnosis of anaemia.

We analysed haematocrit data sets from six different geographic areas, both rural and urban, and two age groups of children in absolute terms and in relationship to our diagnostic cut-off for low haematocrit. In two preschool populations in which simultaneous ferritin levels were measured, we probed for the haematocrit level that best resolved the assignment of risk of iron deficiency.

Materials and methods

Subjects

Haematocrit values of ],253 children of both sexes (age range 6 144 months) were studied. These children participated in different types of surveys performed by the CeSSIAM staff from 1989 to 1992. In all cases, parents signed consent forms for the participation of their children. All the surveys were approved by the human subjects committee of CeSSIAM. The data were not stratified by sex since we did not include teenagers. In addition, the cut-off point did not differ between the sexes in the age range of our subjects.

Communities

All six geographic areas were higher than 1,300 m above sea level; therefore, we did not stratify the data in accordance with any altitude variable. The populations were school-age children from Guatemala City, Antigua Guatemala, and San Pedro Yepocapa and preschool children from the pert-urban community of Peronia, four hamlets in the department of Alta Verapaz, and three hamlets in the department of Santa Rosa.

The six groups were classified dichotomously as urban or rural in concordance with the United Nations criterion (a community with more than 20,000 inhabitants is considered urban).

Guatemala City, the capital of the country, with a population of 2 million, is 1,500 m above sea level and has an area of 2,126 km² and a population density of 1,073 inhabitants per km². Antigua Guatemala, 1,530 m above sea level, has more than 32,067 inhabitants. It is 45 km from Guatemala City and is connected with the capital by an asphalt highway. San Pedro Yepocapa is a township of 217 km², 1,400 m above sea level, in the department of Chimaltenango. It is 94 km from Guatemala City, 32 km of which arc difficult to travel.

Peronia is a pert-urban community on the outskirts of Guatemala City, with about 20,000 people. The four hamlets studied in Alta Verapaz, a department in the northern part of the country with an area of 8,686 km² Chamil (5,000 inhabitants), Pocola (1,433 inhabitants), La Esperanza (1,263 inhabitants), and Sehubu (489 inhabitants)-had an elevation of slightly over 1,300 m above sea level. In Santa Rosa department we surveyed three hamlets El Teocinte, El Naranjo, and Don Gregorio in the municipality of Santa Cruz Naranjo, 53 km from Guatemala City. The municipality covers an area of 67 km², 1,175 m above sea level, and has around 12,000 inhabitants.

Blood testing

Blood samples were taken regardless of whether or not the subject was in a fasting state. Samples were drawn from the antecubital vein or from the dorsal vein of the hand and transfered directly from the syringe to tubes for centrifuging. The amount of blood extracted depended on the type of survey being conducted. In Alta Verapaz and Peronia, whenever possible, an aliquot of plasma was obtained to determine the ferritin concentration.

Haematocrit determination

Heparinized capillary tubes were filled with blood up to two-thirds of their capacity directly from the syringe immediately after venipuncture They were then sealed and centrifuged at 1,500 rpm for five minutes on a microhaematocrit centrifuge. All the samples were run in duplicate no longer than one hour after being drawn. The coefficient of variation determination of haematocrit was 7%.

Measurement of plasma ferritin

After the capillary tubes were filled, the remaining blood samples were stored in heparinized 10-ml glass tubes. The plasma was separated by centrifuging no less than one hour after blood extraction, placed in cold-resistant vials, stored immediately on dry ice, and transported to Guatemala City, where they were stored at -70°C until shipment in dry ice to the University of Kansas laboratory for plasma ferritin measurement by an immunoradiometric assay [14].

Cut-off criteria

We considered a haematocrit below 38% as low because it represents the average of the mean values considered normal in preschool and school children [9]. This was based both on extrapolation and generalization from the experience in juvenile populations in highland Guatemala [9].

We considered a circulating ferritin level of below 12 µg/1 as diagnostic of iron deficiency [8].

Data analyses

Descriptive statistics of arithmetic mean, standard deviation, minimum and maximum values for age, haematocrit and ferritin, and percentage of values beyond the cut-off criteria for the nutritional indices were calculated. To compare mean values between the rural and urban groups as well as between age groups, we used analysis of variance and, when pertinent, Tukey's test [15]. To test the difference in the percentage of individuals falling within different nutritional categories, the chi-square test was employed [15]. Where age profiles differed across geography, we made statistical adjustments to normalize for age distribution [16].

Diagnostic sensitivity and specificity [16] were applied to the prediction of those with and without iron deficiency (as determined by ferritin) by haematocrit values, and the sensitivity and specificity values for each level of haematocrit were determined. The haematocrit level that simultaneously optimized both sensitivity and specificity was computed; this intersection of the cumulative sensitivity and cumulative specificity curves is regarded as the best estimate of correspondence of diagnostic capacity between a reference standard test and a screening procedure [17].

Results

Age profile and haematological status of the subjects

We examined a total of 1,253 children of both sexes, ranging in age from 6 to 144 months. Thirty-three per cent (383) were from urban and 69.4% (870) from rural areas. The mean age of the rural sample was 59.5 ± 14.4 months, with 81.0% preschoolers and 19.0% school children. The mean age of the urban sample was 80.6 ± 14.6 months, with 26.4% preschoolers and 73.6% school children (table 1).

The mean haematocrit was 40.5 ± 2.7% in the urban areas and 39.2 ± 4.3% in the rural areas. The crude prevalences of low haematocrit values were 13.3.% and 26.9% respectively (table 1), which were significantly different (p < 05, x2 = 25.4). Given the differences in age profiles, however, prevalences adjusted for the global age profile of the whole population sample were also generated. The age-adjusted prevalences were 20.6% in the urban areas and 22.1 % in the rural areas, and with these adjusted prevalences the significant geographical differences disappeared.

Haematocrit status

Preschool children

The overall mean haematocrit for the 806 preschool children was 38.1 ± 4.3%. By community, these were 37.9 ± 2.9% in Peronia, 38.6 ± 4.1% in Alta Verapaz, and 37.7 ± 6.0% in Santa Rosa (table 2). The prevalence of low haematocrit was 32% (258 of 806), ranging from 36.9% in Peronia to 32.8% in Alta Verapaz and 26.6% in Santa Rosa (table 3).

Table 2 provides a breakdown of haematocrit for the preschoolers by community and age group. In Peronia we found a significant difference between the 12-23-month and 24-35-month intervals and the other three intervals. In Alta Verapaz and Santa Rosa there were no significant differences between age groups in the mean distribution of haematocrit values. We did not find significant differences between communities.

TABLE 1. Description of the children studied by community, age, and prevalence of low haematocrit

TABLE 2. Mean haematocrit values in preschool children by age group

No statistically significant differences were found between communities for any age group.

Values in the same column marked with a different superscript-a or b are significantly different (p < .05) by ANOVA between age groups within the same community.

TABLE 3. Prevalence (percentage) of low haematocrit values among preschool children from urban and rural areas of Guatemala

Infants, under 12 months of age, were excluded from the prevalence analysis, as the 38% cut-off used for the other children would not be appropriate for those in the first year of life.

Values in the same horizontal row marked with a different superscript a or b are significantly different (p < .05) by the chi-square test between communities for the same group.

Values in the same column marked with a different subscript x, y, or z are significantly different (p <.05) by the chi-square test between age groups within the same community.

Table 3 illustrates the prevalence of low haematocrit as stratified by community and 12-month age intervals. Significant differences were observed between communities, and within communities between age groups. It should be noted that we have included data on children in the 6-11-month age range as a category in table 2 but excluded them in table 3. It is recognized that the cut-off criterion for haemoglobin in infants to diagnose anaemia is lower than that for older children [10] and that our criterion may overestimate anaemia in this subgroup. Since infants represented less than 1.2% of the under-72-month-olds in the two rural areas, their inclusion or exclusion from the prevalence table would make a negligible impact on the conclusions.

School-age children

The overall mean haematocrit for the 447 school children was 41.6 ± 2.7%. By community, means were 41.0 ± 2.3% in Guatemala City, 42.5 ± 3.0% in Antigua, and 41.3 ± 2.8% in Yepocapa. The total prevalence of low haematocrit in school children was 8.7% (39 of 447); it was 10% in Guatemala City, 6% in Antigua, and 8% in Yepocapa. There were no significant differences across geographic sites. No significant differences were seen in mean haematocrit by community and by 12month age interval for the school-age populations.

Comparisons across age groups

An increase in the average haematocrit and a decrease in the prevalence of low hematocrit occurred with increasing age for children of preschool age. The values for haematocrit and haemoglobin were stable in the school-age population. However, we found differences in a global comparison between the 38.1 ± 4.3% mean haematocrit for all preschoolers and 41.6 ± 2.7% for all school children (p > .05) and in the prevalences of low haematocrit of 32.1% and 8.7% respectively (p < .05). These age differences remain valid when the adjustments for differential urban and rural participants are entered (data not shown). This age-dependent gradient in anaemia rates is consistent with our understanding of the epidemiology of nutritional anaemias [7, 9, 13].

Plasma ferritin levels and the diagnostic reliability of haematocrit for iron deficiency

Plasma ferritin levels were determined in a subsample of 290 (65.2%) of the 445 preschool children in rural Alta Verapaz and 42 (41.6%) of the 101 in urban Peronia. With less than 12 µg/l used as the cut-off, 54% of the former and 37% of the latter were classified as iron-deficient. These prevalences were significantly different by chi-square analysis (p < .05)

In Alta Verapaz the haematocrit level that simultaneously optimized both diagnostic sensitivity and specificity was 39%, which provided a sensitivity of 61% and a specificity of 45%. In Peronia a haematocrit of 38% provided a sensitivity of 75% and a specificity of 42%.

Discussion

Extensive surveys of nutritional anaemia around the world have revealed salient features of its epidemiology. For instance, gender differences in anaemia rates do not appear in the pre-adolescent period [9, 18]. For that reason, we did not present our preadolescent population data separately by gender. Also, the age trends for anaemia in childhood have been well established 19. 13, 18, 19]. Nutritional anaemia is a widespread problem in children in developed and developing countries alike [8, 9, 20-30]. However, intercountry comparisons are unreliable because of non-uniform diagnostic criteria across surveys.

Strictly speaking, anaemia is a condition in which the body can no longer produce and maintain the levels of circulating red blood cells required for optimum transport of oxygen to the tissues. Physical signs and symptoms appear, and work capacity is severely curtailed [31]. Iron deficiency, which is the most common nutritional precursor of anaemia, has adverse effects on performance in physical activity [32] and in cognitive [33, 34] and immunological [35] domains, even in advance of a frank red-cell deficit. Moreover, other functions can recover with iron therapy well before haemoglobin values improve significantly. Thus, a more relevant screening question for a population than the prevalence of anaemia is the status of iron nutriture. The haematocrit is an immediate and instantaneous inexpensive test, requiring a minimal blood sample, and would have many advantages in nutritional assessment if we could refine our approaches to its diagnostic use.

For regions in which the population is dispersed over a variety of altitudes from coastal to mountainous locations, differences in ambient oxygen tension and erythropoietic stimulus do not permit a uniform diagnostic criterion for haematocrit or haemoglobin [9, 11, 13]. More fundamentally, however, the conventional approach to the definition of anaemia is as an end-stage process. If we had used the cut-off criterion of the ICNND (33.5%) or of the WHO (33%), we would have estimated that Guatemalan preschool children had an anaemia prevalence of less than 6% and would have assumed a minimal public health problem in haematological nutrition. This is contradicted, of course, by the 52% of plasma ferritin levels that indicated iron depletion in the 61% subsample of preschoolers for whom this biochemical determination was performed.

Conventionally, we have viewed the diagnostic strategy for iron nutriture in terms of the sequential nature of tissue responses [8]. Thus, it is not surprising that long periods of time would be required, for instance, for a person who normally maintains a 46% haematocrit to suffer the 13% reduction in red-cell number to be classified as "anaemic" by the WHO criterion of less than 33%. We propose, moreover, that at altitudes above 1,500 m that prevail in the Guatemalan highlands-or even at sea level-iron-compromised individuals come into red-cell equilibria with their depleted iron stores without passing the severe diagnostic cut-off level of the international standards [10, 12]. Evidence for this relativistic trial was gathered from studies conducted in women in Gothenburg, Sweden [36], which showed that many women whose haemoglobin levels were not in the range defined as anaemic responded to the haematinic nutrient with a rise in this index. The concept that arises from this and other similar experiences is that having any level of circulating red cells below the body optimum for oxygen transport would be a true form of anaemia.

We should re-emphasize that we have adopted the cut-off criterion of less than 38% to define a condition that can be expressed as a prevalence, i.e., a rate of low haematocrit, as we have done here, on somewhat a priori empirical grounds [13]. However, what we hope for is a means to perform such a simple and inexpensive test as the packed-red-cell volume determination and interpret it in terms of the distribution underlying iron stores. Given the sequential nature of development of the iron-deficiency syndrome [8], it would seem to be a reversal of logic to ask a red-cell index (haematocrit) to predict an iron-storage index (ferritin). If, however, a sufficiently firm and fixed predictive relationship were to exist, it would greatly strengthen the capacity of surveys based on haematocrits alone to guide policy decisions. In our sensitivity-specificity analyses, the best prediction of iron-reserve status by the red-cell index was precisely around that chosen empirically as our cut-off level. On the other hand, the predictive power of haematocrit for risk of iron deficiency as defined by low iron stores for the individual must be conceded to be exceptionally weak. In other words, the preschool children in our sample could have had haematocrit values of 40, 42, or higher and still have had depleted iron stores.

It should be recognized that only one-third of the total sample, and exclusively children in the preschool age range, were available for the sensitivity-specificity assay. If speculation is allowed as to why there was a dissociation between iron status and red-cell volume, altitude may again be the key. There is probably population variance in haematological adaptation to lower oxygen, such that not all persons expand their red-cell volume in the same way. If we consider two children with identical total-body iron content, and one produces more red cells, i.e., generates a higher haematocrit, he or she will have a correspondingly smaller iron reserve in store. This would account for a reciprocal relationship between haematocrit and ferritin in the Guatemalan context, which is one of increased altitude.

However, even if the within-individual prediction of one's iron status by haematocrit is poor, can the prevalence of low haematocrits be used to assess the prevalence of iron deficiency in the same population? If we take the rate of low haematocrit in urban preschoolers (37%) and that of ferritin less than 12 µg/dl (37%), we see an opening for a correspondence at the level of the group. For the rural preschoolers, however, the rate of low haematocrits was lower (31 %) but the prevalence of deficient ferritin was 54%. This latter relation is more consistent with the sequential nature of evolution of nutritional anaemia [8]. So, even at the level of the population, the power of haematocrit to reflect the underlying iron nutriture is likely not to be high. We would caution that all such relationships are conditioned by altitude, patterns of endemic diseases such as parasitoses, and associated micronutrient deficiencies in the population.

In conclusion, if ever there was a tip-of-the-iceberg relationship between overt clinical signs and subclinical malnutrition, it is that of haematological status and iron deficiency. In opting for an empirically chosen criterion of less than 38% for low haematocrit, we naturally amplify the rate of classification of the haematological abnormality. This level of circulating red cells, however, is more reflective of the underlying state of iron reserves than are the more severe criteria of international standards for preschool children of the Guatemalan highland zone. Further evidence for the validity of the less restrictive haematological criterion to reflect the population's true status is the comparability of the age-dependent prevalence pattern with what is expected from classic haematological epidemiology [7].

Finally, it is necessary to examine the urban poor as subjects of nutrition surveys. The majority of Latin Americans now live in towns or cities [2], but research has concentrated on rural populations [37].

A host of theoretical reasons exist for expecting an improvement of iron status in the transition from traditional rural lifestyles to urban residence. Nevertheless, present evidence suggests that haematological status in Guatemala does not conform to the generalization of improved nutrition status of urban populations compared with rural. The age-adjusted prevalences of low haematocrits were equivalent, and there was only a slight predominance of deficient ferritin levels in the preschoolers from the rural provinces. Moreover, whatever may be the urban-rural contrasts, rates of low haematocrits upwards of 30% in preschool children, corresponding to depletion of iron stores in almost half, are a cause for alarm and a call to action. Given the potential physical-performance, cognitive-function, and disease-resistance consequences of iron deficiency, public health planners must make efforts to address the disorder in children throughout the country, regardless of the density of settlement of the population.

Acknowledgements

We acknowledge the participation of Marjorie Haskell, Aura-Marina de Guerrero, Marylena Arita, Hector Gamero, Carlos Rivera, Rosalba Perez, Carmen Yolanda Lopez, Alejandrina Vasquez, Carlos Valdez, Sarah Obermeier, Blanca Arevalo, Ana Margarita Isalgue, Julieta Quan de Serrano, Guillermo Segura, and Kelley Cavan in the collection phases of this study.

Financial support was provided by the Nestle Foundation, the United States Agency for International Development, the Procter & Gamble Co., the International Eye Foundation, and the Thrasher Research Fund.

References

1. Hussain AM, Lunven P. Urbanization and hunger in the cities Food Nutr Bull 1987;9(4)4:50-61.

2. Popkin BM, Bisgrove EZ Urbanization and nutrition in low-income countries Food Nutr Bull 1988;10(1):3-28.

3. Viteri FE Nutrition-related health consequences of urbanization Food Nutr Hull 1987;9(4):33-49.

4. Mott KE, Dejjeux A, Manoyo A et al. Parasitic diseases and urban development Bull WHO 1990;68(6): 691 -8.

5. Rossi-Espagnet A Health services and environmental factors in urban slums and shanty towns of the developing world. Food Nutr Bull 1987;9(4):4-20

6. Musgrove P. Basic food consumption in north-east Brazil: effects of income, price, and family size in metropolitan and rural areas. Food Nutr Bull 1988;10(1):29-38.

7. De Maeyer EM. Preventing and controlling iron deficiency anaemia through primary health care: a guide for health administrators and program managers. Geneva: WHO, 1989.

8. Cook JD, Clement A, Finch CA. Assessing iron status of a population. Am J Clin Nutr 1979;32:2115-19.

9. Viteri FE, De Tuna V, Guzman MA. Normal haematological values in the Central American population. Brit J Haematol 1972;23:189-204.

10. World Health Organization. Nutritional anemia: report of a WHO scientific group. Technical Report Series, no. 405. Geneva: WHO, 1968

11. Tufts DA, Haas JD, Beard JL, Spielvogel H. Distribution of hemoglobin and functional consequences of anemia in adult males at high altitude. Am J Clin Nutr 1985;42:111.

12. Interdepartmental Committee for Nutrition in National Development. Evaluación nutricional de la población de Centro América y Panama: Guatemala. Guatemala City: INCAP, 1969.

13. Ruz M, Rosas A, Bulux J et al. Haematological status of school children in two regions of Guatemala: relevance of normality standards. Int J Food Nutr 1992; 43:89-95.

14. Miles LEM, Lipschitz DA, Bieber CP, Cook JD. Measurement of serum ferritin by a 2-site immunoradiometric assay. Anal Biochem 1974;61:20924.

15. Colton T. Statistics in medicine. Boston, Mass, USA: Little Brown, 1974.

16. Dawson-Saunder B. Trapp RG. Basic and clinical biostatistics. San Mateo, Calif, USA: Appleton & Lange, 1990.

17. Mausner JS, Bahn AK. Epidemiology: an introductory text. Philadelphia, Pa, USA: W. B. Saunders, 1994.

18. WHO Expert Scientific Working Group. Summary of a report on assessment of the iron nutritional status of the United States population. Am J Clin Nutr 1985;42: 1318-30.

19. Viteri FE, Guzman MA. Hematological status of the Central American population: prevalence of individuals with haemoglobin levels below normal. Brit J Haematol 1972;23:725-35.

20. Lira PIC, Cartagena HA, Mello R de et al. Estado nutricional de crianças menores de seis anos, segundo posse da terra, em areas rurais do estado de Pernambuco, nordeste do Brasil. Arch Latinoam Nutr 1985; 35(2):247-57

21. Sigulem DM, Tudisco ES, Paiva de Rodriguez E, Guerra de Campos CC. Anemia nutricional y parasitosis intestinal en menores de 5 años. Rev Paul Med 1985; 103(6):308-12.

22. Tershakovee AM, Weller SC. Iron status of inner-city elementary school children: lack of correlation between anemia and iron deficiency. Am J Clin Nutr 1991;54:1071-6.

23. Anour E, Hasin M, Yodfat Y. [Anemia in Jerusalem children aged 3-6 years]. Harefuah 1992;123:373-6.

24. Raman L, Pawashe AB, Ramalakshmi BA. Iron nutritional status of preschool children. Indian J Ped 1992; 59:209-12.

25. Jalaludin B. Taylor R. Levy S. Montaville B. Gee K. Prevalence of anaemia and iron deficiency at different levels of urbanization in Vanuatu. Papua New Guinea Med J 1992;35:128-35.

26. Molla A, Khurshid M, Molla AM. Prevalence of iron deficiency anaemia in children of the urban slums of Karachi. J Pakistan Med Assoc 1992;42:118-21.

27. Hongo T. Suzuki T. Ohba T et al. Nutritional assessment of a group of Japanese elementary school children in Tokyo: with special emphasis on growth, anemia and obesity. J Nutr S Vitaminol 1992;38:177-96.

28. Westhuyzen J. Steyn NP. Hematological nutrition of schoolchildren in the far northwestern Cape. Trop Geogr Med 1992;44:47-51.

29. Abellan-Ripoli A, Alcaraz-Quinonero M, Mengual-Cos M et al. [Prevalence of iron deficiency and iron deficiency anemia in infancy and related factors in a community of the Murcia region]. Anales Españoles de Pediatría 1992;36:265-8.

30. Lehmann F. Gray-Donald K, Mongeon M, Di Tommaso S. Iron deficiency anemia in l-year-old children of disadvantaged families in Montreal. J Can Med Assoc 1992;146:1571-7.

31. Basta SS, Soekiman MS, Karyadi D, Scrimshaw NS. Iron deficiency anemia and the productivity of adult males in Indonesia. Am J Clin Nutr 1979;32:916-25.

32. Viteri FE, Torun B. Anemia and physical work capacity. Clin Hematol 1974;3:609-24.

33. Pollitt E, Saco-Pollitt C, Rudolph LL, Viteri EF. Iron deficiency and behavioral development in infants and preschool children. Am J Clin Nutr 1986;43:555-65.

34. Lozoff B. Brittenham GM, Wolf AW et al Iron deficiency anemia and iron therapy effect on infant development test performance. Pediatrics 1987;79:98195.

35. Dallman PR. Iron deficiency and the immune response. Am J Clin Nutr 1987;46:329-34.

36. Garby L, Irnell L, Werner I. Iron deficiency in women of fertile age in a Swedish community: III. Estimation of prevalence based on response to iron supplementation. Acta Med Scand 1969;185:113-17.

37. Solomons NW. Urban nutrition in the tropics: a neglected subject of scientific inquiry. In: Gross R. Solomons NW, eds. Tropical urban nutrition. Eschborn, Germany: Gesellschaft für Technische Zusammenarbeit, 1987:15-30.

Contents - Previous - Next