Iron deficiency anaemia prevalence in a population of immigrated women in Italy (original) (raw)

Abstract

Background: Martial deficiency and sideropenic anaemia are the most diffused deficiency pathologies in the world. WHO recommends preventive screening of the new immigrant population. No epidemiological data exist on its prevalence among migrant population in Italy. Methods: A transversal study was conducted at San Gallicano Hospital in Rome through laboratory screening on 821 migrant women and interviews on a sub-sample of 550 women (including socio-demographic, anamnestic and nutritional information). Results: The complete sub-sample (laboratory results and questionnaire) shows a 20.5% [95% confidence interval (CI) 16.8–24.3] prevalence of anaemia and a 22.7% (95% CI 18.9–26.6) prevalence of sideropenia. Sideropenic anaemia was found in 11.5% (95% CI 8.5–14.4) of cases. Results are similar in the rest of the sample. There is significant association between anaemia and the clinical conditions of haemorrhoids [odds ratio (OR) 3.8; P < 0.000], hypermenorrhoea (OR 3.3; P < 0.000) and metrorrhagia (OR 5.9; P < 0.000). Africans were found to be at highest risk of anaemia (OR 5.5; P < 0.000). Feeding habits have a milder effect. Unemployed and low educated people are more likely to be affected by non-iron deficiency anaemia. Conclusion: The observed prevalence of sideropenia and sideropenic anaemia is much greater than what the scientific literature reports for Western populations. Pathologies inducing bleeding and the country of origin (i.e. genetic factors, pre-existing conditions) appear to be associated with anaemia. Nutritional factors are less important because of an adequate nutritional income. Prevention programmes should then aim at screening larger samples for improving the access of migrants to health-care services.

Introduction

Iron is a fundamental micronutrient involved in the formation of haemoglobin, myoglobin and various oxygen-carrier enzymes. Physiological iron loss in women often causes unbalance, which can lead to iron deficiency and sideropenic anaemia. During pregnancy, these conditions can alter foetal development and may increase the maternal morbidity and mortality risk.1,2

The prevalence of these forms of anaemia is considerably high in developing countries and represents the main nutritional problem of some population groups.3 In the USA, the prevalence of sideropenic anaemia is estimated to be between 2 and 5% in non-pregnant women aged 12–49 years.4,5 According to World Health Organisation (WHO),3 the prevalence of anaemia in women aged 15–59 years is 10.3% in industrialized countries, whereas it reaches 42.3% in developing countries. Among pregnant women, who represent a high-risk group, the calculated rates are 22.7 and 52.0%, respectively.

Epidemiological data on the prevalence of sideropenic anaemia in developed and developing countries are available, but no data exist on migrant population in Italy. Studies carried out in Europe, North America and Australia demonstrated that iron unbalance is particularly frequent among migrants and socio-economically disadvantaged population groups.6–9 The WHO highlights the importance of implementing specific screening programmes for the above-mentioned population groups. As observed in other studies,10,11 sideropenic anaemia in migrant women can be favoured by environmental factors (i.e. insufficient feeding, disadvantaged social and economic conditions), underestimated diseases (i.e. gynaecological and gastrointestinal disorders) and difficulties in health-care utilization.

The ever-growing migration flow towards Italy poses health-care problems requiring effective prevention programmes based on reliable epidemiological data.12,13 Migrants may be affected by the most diffused diseases among their native population, which are often due to disadvantaged life conditions.14 The scientific literature only provides few epidemiological data on the issue.15 The Italian scientific literature does not include specific studies on the diffusion of anaemia among migrants in the country. Information could be derived from the Hospital Discharge Registers (HDR),16,17 since ICD-9-CM (adopted in all Italian public hospitals) codifies all types of anaemia and iron deficiency anaemia. Anyway, a more verisimilar picture of the prevalence of sideropenic anaemia among female migrants could be derived from the data of the out-patient department rather than those of the in-patient department. Indeed, in-patient women often suffer from severe anaemia whereas outpatients can also present milder levels of sideropenic anaemia. The authors considered it as important to focus on the latter condition, because early diagnosis can prevent hospitalization and future complications. Therefore, this study aims at assessing the prevalence of this pathology among out-patient women.

Methods

From January 2008 to April 2009, a screening programme was conducted in the Outpatient Department of San Gallicano Hospital in Rome, National Institute for Health Migration and Poverty (NIHMP). NIHMP is a privileged observatory for the health conditions of disadvantaged migrants since this institute has been committed for years with marginalized and weak Italian and migrant population groups.18,19 The study adopted a multi-disciplinary approach, since it is the result of the collaboration of doctors, nurses and cultural mediators dealing with communication.

A large sample of women attending the NIHMP Outpatient Department of Preventive Medicine of Migration was analysed in order to study the epidemiology and the determinants of martial deficiency in migrant women.

Each woman had the possibility of undergoing a complete blood cell count and a serum ferritin test. Neither clinical nor other selection was made. Out of 1257 women attending the out-patient department in the considered period, 821 blood samples were obtained for the haemochrome and ferritin test. Out of these women, 550 also agreed to answer a three-module questionnaire exploring socio-demographic, clinical and nutritional issues.

Information on the prevalence of anaemia and iron deficiency was derived from the results of haematochemical tests. The cut-offs for both conditions were fixed in accordance with WHO recommendations and the significance emerging from data.3

For the 453 persons with complete information (haematochemical tests results and epidemiological data), the Pearson’s correlation coefficient between haemoglobin and ferritin levels resulted 23.4%, explaining 39.1% (F1,451 = 26.1; P < 0.000) of haemoglobin variability.

A share of 50.5% of people with ferritin <15 ng ml−1 presented haemoglobin (Hb) <12 g dl−1, whereas at higher ferritin levels, <16% of the total had Hb <12 g dl−1. Persons with ferritin <15 ng ml−1 show 8.7 OR (chi-square = 18.5; _P_ < 0.000) of Hb <12 g dl−1, whereas for higher levels of ferritin, the OR was <1.6, and did not reach statistical significance. Therefore we considered anaemia as the cases where haemoglobin was <12 g dl−1 and iron deficiency when ferritin was <15 ng ml−1. These figures correspond to the limits indicated by WHO,3 respectively: serum ferritin level evidencing iron stores depletion, in absence of infection in >5 years females and Hb level below which anaemia is present for >15 years non-pregnant women.

The association between anaemia and iron deficiency is termed as iron deficiency anaemia (IDA). Anaemia without iron deficiency is referred to as non-iron deficiency anaemia (NIDA). These definitions are based on the conceptual framework adopted by WHO in the Guide for IDA Assessment, Prevention and Control.20

Out of the total epidemiological questionnaires administered, 453 were complete and therefore analysed. The information was gathered through accurate personal interviews carried out with the help of intercultural mediators. For the present study, only complete records have been considered, so that the response rate is 82.4%. The sample power of 453 cases (sp = 100%; P = 0.01) is the same of 550 records in refusing the hypothesis that the expected prevalence of anaemia in our sample (assumed as the mean of the figures for industrialized and developing countries) is equal to or greater than the one observed in developing countries.

Case history and medical examination can individuate possible causes for negative iron balance such as physiological (menstrual flow, pregnancy, etc.) or pathological (bleeding, genitourinary loss, gastroenteric or other) conditions.

The first part of the questionnaire concerns personal data (age, civil status, country of origin), social conditions (period of permanence in Italy, housing condition, education, job, permit of stay and cause of migration) and religious beliefs. No data on income or economic situation were collected, because the reliability of responses was unverifiable. Nevertheless, assessment of the working condition may provide significant data to assume that social disadvantage can determine a higher risk of anaemia.

Country of origin is a particularly important criterion, which can provide more verifiable information. We assumed Eastern European migrants as a base group for comparison analysis, due to their better living conditions on average, compared with other large migrant groups (i.e. Africans).

The second part of the questionnaire reports the case history and information derived from the medical examination carried out by a haematologist.

The third part regards eating habits and is aimed at gathering data on the number of meals per day, place of consumption and consumption frequency for the following foods:

(i) meat and fish, (ii) vegetables, (iii) fruits, (iv) legumes, (v) milk and derivatives. A relation between low consumption of foods with high iron bioavailability (i.e. meat, fish) and high risk of sideropenic anaemia is expected.

The used methodology provides information on the usual diet in the long run, generally 1 year. Data on the individual eating habits in the country of origin of 175 women living in Italy for <12 months were also collected. We assumed that migrants usually adapt their eating habits to that of the native population after 1 year of immigration. In addition, within the first year of immigration, the iron balance can still be influenced by the eating habits of their countries of origin. For example, female migrants coming from poorer countries can show malnutrition (i.e. low consumption of meat and derivatives).

All those women who have found to be anaemic underwent diagnostic investigation and received nutrition advice and therapeutic support.

A descriptive analysis was carried out in order to observe the socio-demographic, clinical and nutritional characteristics of the sample. The prevalence of anaemia, IDA and NIDA was then calculated by estimating binomial exact CIs.

The study design was transversal and allowed to analyse the association between anaemia and the above-mentioned potential risk factors by calculating the ORs in univariate analysis per each significant independent variable. The significant associations for anaemia and each subgroup dependent variable (IDA and NIDA) were then analysed within a multivariate logistic regression model.

Results

The 453 cases provided with complete haematochemical and epidemiological information from the questionnaire were analysed. Prevalence observed in the sub-sample was 20.5% for anaemia (95% CI 16.8–24.3) and 11.5% for IDA (95% CI 8.5–14.4), respectively. Iron deficiency was detected in 22.7% of the total cases (95% CI 18.9–26.6). The Hb range of variation was 8.1–15.7 g dl−1 (mean = 12.8; SD = 1.2), whereas ferritin range was 1–270 ng ml−1 (mean = 43.4; SD = 40.1). These figures differ not significantly to those of the participants who did not agree to fill in the questionnaire (n = 271) and those whose questionnaires are not completed (n = 97).

These 368 women showed a prevalence of anaemia of 19.6% (95% CI 15.6–24.2). Iron deficiency was observed in 23.1% (95% CI 18.7–27.9) of the cases, determining a IDA prevalence equal to 11.0% (95% CI 7.9–14.7).

The average length of stay in Italy was 18 months: 61.4% live in Italy for >12 months and 38.6% for <12 months. Among all participants, 76.4% declared having a high school diploma or university degree and 59.8% were employed (53.6% as home helpers). Those and other socio-demographic information are given in table 1.

Table 1

Absolute and percentage socio-demographic and clinical distribution N = 453 women

Table 1

Absolute and percentage socio-demographic and clinical distribution N = 453 women

Concerning the number of meals consumed per day, the majority declared to eat three or more meals per day (80.6%), whereas only 2.4% reported one meal per day (table 2). The main place of consumption of breakfast, lunch and dinner resulted to be mainly home, their own or the one of the employer. Less than 10% declared they receive at least one out of three meals from public assistance services, mainly at lunch.

Table 2

Nutritional habits and feeding frequency (N = 453 women)

Table 2

Nutritional habits and feeding frequency (N = 453 women)

Food consumption frequency by nourishment groups does not highlight nutritional deficiencies, particularly regarding iron. The participating women declared they eat meat or fish once a day or three to four times a week in 47.7 and 32.2% of cases, respectively. Nevertheless, a relevant number of women reported to consume meat or fish less than two times a week (18.1%) or just once a month (2.0%).

Food consumption in the country of origin was also reported for the 175 women living in Italy for <12 months. No relevant difference was observed with respect to the rest of the sample: 58.8% reported to consume meat and fish at least once a day, 12.6% twice a week and 2.3% once a month.

An analysis of the risk factors that can lead to anaemia and IDA has been carried out on the basis of the socio-demographic, clinical and nutritional data derived from the epidemiological questionnaire.

The risk of anaemia was estimated for each observed factor through univariate analysis. According to these analyses, iron deficiency is significantly associated with anaemia (table 3): a subject with ferritin <15 ng ml−1, has >7-fold greater risk (OR) of being anaemic in comparison with a person having higher levels of iron. As regards socio-demographic factors, being aged between 20 and 44 years results significantly associated with anaemia, since this age group has a 1.9-fold risk of anaemia compared with the base group (age > 44 years).

Table 3

Risk factors association with anaemia, IDA and NIDA (N = 453 women)

Table 3

Risk factors association with anaemia, IDA and NIDA (N = 453 women)

African women show a remarkably higher risk of anaemia; correspondently, the average values of Hb (12.0 g dl−1) and ferritin (32 ng ml−1) differ significantly (P < 0.001) from those of the other women (Hb 12.8 g dl−1; ferritin 44 ng ml−1).

Compared with the other participants, African women are more likely to be younger, unmarried, without children and present in Italy for a long time, whereas, from a clinical-anamnestic perspective, they do not differ from the remaining population.

Occupational conditions constitute another risk factor when comparing unemployed or household helpers to the group of housewives, students and others.

As regard health conditions, gastric pathologies are significantly (P < 0.05) associated to anaemia even though diseases causing bleeding such as haemorrhoids, menorrhagia and metrorrhagia involve a higher risk of Hb <12 g dl−1. In general, the presence of haemorrhoids, menorrhagia, metrorrhagia or gastric and intestinal pathologies are significant risk factors for developing anaemia. The presence of menstrual flow increases the probability of having anaemia more than gastric pathologies.

Skipping lunch implies a higher risk of anaemia, particularly IDA (OR = 2.9; P < 0.05). The frequency of consumption of determined foods do not influence iron deficiency. Nevertheless, there is a significant association between the consumption of meat (or fish), vegetables and milk and the presence of NIDA.

In fact, irregular consumption of vegetables (less than once a day), even if accompanied with regular consumption of milk and derivatives (once a day) or vice versa, provokes a higher risk of NIDA. Similarly, a regular consumption of meat (or fish) and vegetables, at least once a week and once a day respectively, seems to be a significant protective factor to prevent NIDA when compared with an irregular consumption of these nourishments.

Other risk factors significantly associated with NIDA are an age <45 years, African origin, primary-school education or no education, and unemployment.

Country of origin and skipping lunch increase the risk of IDA too, but the most remarkable effects derive from the presence of menstrual flow and bleeding pathologies such as menorrhagia, metrorrhagia and haemorrhoids.

Risk factors were analysed through a multivariate approach in order to avoid reciprocal confounding effects. The most influencing factors of each disease were weighted for every significant stratification by using the Mantel–Haenszel method. As expected, iron deficiency was confirmed as the principal factor causing anaemia, with an OR reaching 19.5 (P < 0.0001). On the contrary, menstrual flow was no longer the main factor for IDA. The presence of at least one pathology among the above mentioned ones, resulted to be the principal cause of IDA with an OR equal to 4.4 (P < 0.02). Latin Americans are at the highest risk of getting NIDA (OR = 5.2; P < 0.01), followed by Africans (OR = 4.6; P < 0.02).

Mathematical logistic regression models were developed in order to calculate OR by taking into account all the significant causes of disease simultaneously (table 4). By applying the regression model for IDA that considers the non-compound variables only (menorrhagia, metrorrhagia, haemorrhoid, origin) and joining these ones in linear addictive combination we obtained an OR of 82.6 (P < 0.0001).

Table 4

Multiple logistic regression models for anaemia, IDA and NIDA (N = 453 women)

a: Log likelihood = –164.8 likelihood ratio chi-square10 = 130.4 P = 0.000; generalized _R_2 = 0.284; goodness of fit: N. groups = 10 Hosmer–Lemeshow chi-square8 = 5.3 P = 0.722

b: Log likelihood = 132.3 likelihood ratio chi-square6 = 58.4 P = 0.000; generalized _R_2 = 0.181; goodness of fit: N. groups = 7 Hosmer–Lemeshow chi-square5 = 3.8 P = 0.573

c: Log likelihood = 116.6 likelihood ratio chi-square6 = 42.0 P = 0.000; generalized _R_2 = 0.153; goodness of fit: N. groups = 10 Hosmer–Lemeshow chi-square8 = 4.3 P = 0.82

Table 4

Multiple logistic regression models for anaemia, IDA and NIDA (N = 453 women)

a: Log likelihood = –164.8 likelihood ratio chi-square10 = 130.4 P = 0.000; generalized _R_2 = 0.284; goodness of fit: N. groups = 10 Hosmer–Lemeshow chi-square8 = 5.3 P = 0.722

b: Log likelihood = 132.3 likelihood ratio chi-square6 = 58.4 P = 0.000; generalized _R_2 = 0.181; goodness of fit: N. groups = 7 Hosmer–Lemeshow chi-square5 = 3.8 P = 0.573

c: Log likelihood = 116.6 likelihood ratio chi-square6 = 42.0 P = 0.000; generalized _R_2 = 0.153; goodness of fit: N. groups = 10 Hosmer–Lemeshow chi-square8 = 4.3 P = 0.82

Discussion

Within the considered sample of migrant women aged 14–60 years, a relevantly higher prevalence of anaemia was observed compared with what was expected for women aged 15–59 years living in industrialized countries. The prevalence observed in the present study is in fact 20.5%, a percentage 2-fold higher than estimated by WHO for women aged 15–59 and living in industrialized countries (10.3%) but half of what expected for non-industrialized countries (42.3%).3 Similarly, the registered prevalence of sideropenic anaemia is ∼12% whereas the US National Health and Nutrition Examination Survey observed a 3% prevalence for non pregnant women aged 12–49 (95% CI 2–4).5

The findings of our study are similar to those of the US study, especially when comparing the results for iron deficiency referring to the ethnic subgroups. Our study resulted in a prevalence of 23% (95% CI 19–27) in migrant women in Italy, the US study resulted in prevalence rates of 19% (95% CI 14–24) in Afro-American females and of 22% (95% CI 17–27) in Mexicans, respectively.

Moreover, our study showed a sideropenia prevalence of 36% (95% CI 19–56) in sub-Saharan African females, 23% (95% CI 15–32) in Latin American women and of 22% (95% CI 18–27) in Eastern Europeans.

The above mentioned values largely exceed the 10% (95% CI 7–13) prevalence of iron deficiency resulted in Caucasian women in the US study.

In our sample anaemia derives from sideropenia in about half of cases. In fact, the prevalence of sideropenic anaemia is 11.5% while the prevalence of anaemia is 20.3%. Therefore, further research is needed to assess the prevalence of NIDA in migrant women (i.e. folate-deficiency and B12-deficiency anaemia, haemolytic anaemia and haemoglobinopathy).

The socio-demographic characteristics of the analysed sample (table 1) are similar to those of the hospital inpatient women and of the migrant female population in the metropolitan area of Rome. Among the considered sample, 92.9% of women were undocumented migrants; their socio-demographic distribution was therefore highly concordant with the one of recently regularized people in the area.19

The observed sample did not show evident nutritional deficiency and the consumption frequency of each type of food resulted quite regular. Therefore, the factors influencing the presence of IDA do not reside in feeding conditions.

Occupational conditions (employment versus unemployment) were not included in the IDA regression model, because unemployed people did not result at greater risk of IDA and anaemia. Nevertheless, household helpers, representing almost every employed person in the considered sample, or unemployed people seem to be at a higher risk of getting anaemia than housewives and students. Given that most household helpers in Italy, especially migrants, are not regularly employed, this difference can be assumed as an indicator of the effect of social instability on the risk of anaemia.

The regression model for IDA recognized as determinants: (i) the presence of menorrhagia, metrorrhagia or haemorrhoids; (ii) the presence of at least one pathology in >44-year-old subjects; (iii) the interaction of the habit of skipping lunch and the irregular consumption of vegetables, milk and derivatives; (iv) the African or Latin American origin.

The IDA multivariate model cannot consider all the possible causes of the pathology, but it identifies important factors. Therefore, it identifies some significant risk factors for IDA, the most important of which is the presence of a bleeding pathology. In this model the positive (cases) predictive value (PPV) is 76.9% and the negative predictive value (NPV) is 90.5%. It means that, as regards IDA, 90.1% of predictions are correct. Therefore, it can be a useful anamnesis instrument, but it is not complete because of its still insufficient sensitivity (19.2%).

For NIDA, the significant identified causes are low education, unemployment, presence of at least one bleeding pathology, country of origin, irregular consumption of vegetables and milk, interaction of different variables such as education, job and country of origin. In this model, education and occupational conditions play the main role, followed by the presence of pathologies and the country of origin. Nourishment resulted to have a much lower effect. Educational, occupational and pathological factors in linear combination with the country of origin and the consumption of vegetables and milk show an OR of 84.4 (P < 0.0001).

Iron deficiency is the main factor determining anaemia in the observed sample, but, in most cases, it does not explain the variation of haemoglobin levels. Therefore, other factors (i.e. folate-deficiency, B12 deficiency and haemoglobinopathy) should be also taken into account.

Almost all the considered variables for IDA and NIDA are included in the anaemia regression model.

As expected, also from a multivariate analysis, iron deficiency is confirmed as the main factor causing anaemia with an OR of 8.1 (P < 0.0001). Other important factors are bleeding pathologies (i.e. haemorrhoids and metrorrhagia), occupational conditions, and education. Nevertheless, it should also be considered that the low risk related to gastric diseases and intestinal pathologies can be explained, apart from the sample power, by their underestimation during the process of anamnesis.

By joining non-compound variables in linear combinations, clinical factors (pathologies) account for OR = 46.6 (P < 0.0001), socio-demographic causes for OR = 11.6 (P < 0.0001) and nutritional variables for OR = 3.4 (P < 0.01). The model has PPV = 66.7% and NPV = 85.7%, so that, by using this model, 83.4% of the observed cases are correctly classified. According to this mathematical model, bleeding pathologies are the main determinant of anaemia, and socio-demographic causes result more influent than nutritional habits.

Among the 821 migrant women screened at NIHMP, San Gallicano Hospital in Rome an important prevalence of anaemia (20.1%) and IDA (11.3%) was observed. The prevalence of IDA resulted in the present study is significantly higher than the maximum expected level for the female adult population in Western countries as reported in the scientific literature (5%).3–5 This observation confirms the WHO indications that newcomer migrants in Western Countries are particularly at risk and supports the WHO recommendation on the need for preventive screening.3 This study provides the first data on the prevalence of this condition among migrant population in Italy.

In the analysed sample, ∼50% of the anaemic patients are affected by IDA; its large prevalence indicates the need for screening procedures based on the combination of two low-cost tests such as blood count and serum ferritin test.

The study design allowed to estimate the prevalence on a wider sample using laboratory parameters, but also medical examination results and other epidemiological information collected by a questionnaire within a subsample, were available. Therefore, this information can also be used for exploring the factors determining the cases of anaemia observed at the NIHMP.

The chosen study design was properly intended to estimate the prevalence of anaemia (IDA or NIDA), which was the main objective of the study. It does not allow to get to definitive conclusions on the factors determining anaemia. Nevertheless, the epidemiological information obtained from the questionnaire allowed the identification of the effects of some significant causes of anaemia.

In conclusion, multivariate analysis suggests that factors significantly associated with anaemia can be identified, such as the presence of pathologies inducing bleeding and the country of origin (i.e. genetic factors, pre-existing conditions). Education and occupational condition (social stability) also play a role whereas nutritional factors seem to be less important for the genesis of anaemia in the analysed subjects, who anyway showed an adequate nutritional income.

This information is useful for clinical practice, especially in the anamnesis, and in the epidemiological analysis representing the necessary reference for further studies specifically oriented to investigate the determinants of anaemia in migrant population.

Acknowledgements

The authors wish to thank all those women who consented to answer the questionnaires for their fundamental contribution in this study. They are also grateful to cultural mediators and to the NIHMP staff. They would particularly like to thank Cecilia Fazioli for linguistically revising and translating this article.

Conflicts of interest: None declared.

• A transversal study has been conducted on 821 migrant women at the outpatient department of San Gallicano Hospital in Rome.
• A total of 20.1% (95% CI 17.4–23.1) prevalence of anaemia and 11.3% (95% CI 9.1–13.6) prevalence of sideropenic anaemia were found. The observed prevalence of sideropenic anaemia is much greater than what reported in scientific literature for Western populations.
• Pathologies inducing bleeding and the country of origin appear to be significantly associated with anaemia.

References

1

Manifestations of iron deficiency

,

Semin Hematol

,

1982

, vol.

19

(pg.

19

-

30

)

2

et al.

Long-lasting neural and behavioral effects of iron deficiency in infancy

,

Nutr Rev

,

2006

, vol.

64

Suppl

(pg.

34

-

43

)

3

WHO, UNICEF, ONU

Iron Deficiency Anemia: Assessment, Prevention and Control

,

A guide for programme managers

,

2001

Geneva (CH)

World Health Organization

Report No.: WHO/NHD/01.3

4

Centers for Disease Control and Prevention

Recommendations to prevent and control iron deficiency in the United States

,

MMWR Recomm Rep

,

1998

, vol.

47(RR-3)

(pg.

1

-

29

)

5

Centers for Disease Control and Prevention

Iron deficiency - United States, 1999–2000

,

MMWR

,

2002

, vol.

51

(pg.

897

-

9

)

6

Hematological problems in immigrants from Southeast Asia

,

Hematol Oncol Clin North Am

,

2004

, vol.

18

(pg.

1405

-

22

)

7

et al.

Dietary intake does not account for differences in low iron stores among Mexican American and non-hispanic White women: Third National Health and Nutrition Examination Survey, 1988-1994

,

J Nutr

,

2002

, vol.

132

(pg.

996

-

1001

)

8

et al.

Nutrition and health among migrants in the Netherlands

,

Public Health Nutrition

,

2001

, vol.

4

(pg.

659

-

64

)

9

et al.

Iron deficiency in Australian-born children of Arabic background in central Sydney

,

Med J Aust

,

2001

, vol.

174

(pg.

165

-

8

)

10

et al.

The role of blood loss and diet in the aetiology of mild iron deficiency in premenopausal adult New Zealand women

,

Public Health Nutr

,

2001

, vol.

4

(pg.

197

-

206

)

11

Risk of suboptimal iron and zinc nutriture among adolescent girls in Australia and New Zealand: causes, consequences and solutions

,

Asia Pac J Clin Nutr

,

2002

, vol.

11

Suppl 3

(pg.

S543

-

52

)

12

Italian National Institute of Statistics – ISTAT, Italy: Available at http://demo.istat.it (1 October 2009, date last accessed)

13

Caritas/Migrantes

Dossier Statistico Immigrazione 2009 (Immigration Statistical Dossier 2009)

,

2009

Rome (IT)

Idos Centro Studi e Ricerche

Rapporto: XIX

14

Rapporto tra disagio culturale e salute: alcune riflessioni a margine di una ricerca interdisciplinare [Relationship between cultural uneasiness and health: thoughts within a interdisciplinary research]

,

Cultura Salute Immigrazione [Culture Health Immigration]

,

1995

Rome

Armando Editore

(pg.

62

-

74

)

15

Public health research among immigrant populations: still a long way to go

,

Eur J Epidemiol

,

2003

, vol.

18

(pg.

841

-

2

)

16

Il ricovero ospedaliero degli stranieri in Italia nell’anno 2000 [Hospital admission of foreign citizens in Italy in 2000]

,

2002

Rome (IT)

Ministry of Health

17

La sfida della medicina delle migrazioni [The challenge of migration medicine]

In: Caritas-Migrantes, editor. Dossier Statistico Immigrazione 2005 [_Immigration Statistical Dossier 2005_]. Rome (IT): Idos Centro Studi e Ricerche; 2005 Report No. XV: 179–88

18

Health systems confront poverty

,

2003

Copenhagen (DK): WHO Regional Office for Europe; 2003 Public Health Case Studies No.: 1

19

,

Gli immigrati nella provincia di Roma - Rapporto 2005 [The immigrants in the Province of Rome - Report 2005]

,

2005

Milan

Franco Angeli Editore

20

Iron nutritional status defined

,

Dietary iron: birth to two years

,

1989

New York

Raven Press

(pg.

19

-

36

)

© The Author 2010. Published by Oxford University Press on behalf of the European Public Health Association. All rights reserved.