Is iron status associated with highly unsaturated fatty acid status among Canadian Arctic Inuit?

Abstract

Impaired fatty acid synthesis was noted in iron deficient animal models. Human data, however, are scarce. Although Canadian Inuit have a traditional diet rich in heme iron and long chain n-3fatty acids, recent literature has also indicated the presence of prevalent iron deficiency. We aimed to explore whether the presence of iron deficiency would affect fatty acid status and an estimate of the activity of desaturase 5 (Δ5), which is crucial in the biosynthesis of highly unsaturated n-3fatty acids among Canadian Inuit. Erythrocyte membrane fatty acid composition was utilized as an indicator of fatty acid status and serum ferritin and circulating hemoglobin level were measured as the indicators of iron status. Data analyzed were collected among 1511 Canadian Inuit adult participants in the International Polar Year Inuit Health Survey, 2007–2008. Only 13.7% of survey participants had iron deficiency; however, serum ferritin showed a moderate positive association with highly unsaturated n-3fatty acids after adjusting for age, waist and C-reactive protein (r = 0.172, P < .0001). Serum ferritin correlated significantly with Δ5 after further adjusting for highly unsaturated n-3fatty acids (r = 0.126, P < .0001). Although the current study only demonstrated a weak link between ferritin and Δ5, the latter association underscores a possible health risk caused by a nutrient interaction related to reduced iron intake and decreased highly unsaturated n-3 fatty acid biosynthesis. Future studies are recommended to evaluate iron status in relation to highly unsaturated n-3 fatty acid biosynthesis and status among indigenous people undergoing rapid dietary transitions.

---

1 Introduction

Despite the well-documented hormonal regulation of endogenous fatty acid metabolism, there is relatively little knowledge regarding nutritional regulators, particularly micronutrients. Experimental rat studies suggest that reduced iron status could be related to altered fatty acid metabolism as moderate diet-induced iron deficiency (ID) resulted in reduced conversion of shorter-chain polyunsaturated fatty acids to highly unsaturated fatty acids (HUFA).^1,2 These latter findings demonstrated decreased activity of iron-containing desaturases including Δ5, which are crucial in the de novolipogenesis of HUFA.³ The association between ID and HUFA status, however, has not been fully explored in the human context.

In the last few decades, Canadian Inuit have been experiencing a greatly accelerated dietary transition away from Inuit food to store food shipped from the southern regions,⁴ which raises public health concerns among this population. Compared to market food, Inuit food is more nutrient-dense and less energy-dense.⁴ The traditional n-3 HUFA rich Inuit diet is considered to be protective against cardiovascular disease,^5,6 particularly ischemic heart disease.⁷ The traditional Inuit diet mainly comprises of red meat from wild life, which is also an excellent source of heme iron. With the reduced consumption of Inuit traditional food, epidemiologic transitions have been noted in terms of increased obesity and chronic disease risk.⁸ Additionally, iron deficiency has been suggested to be more prevalent among the Arctic native population than the general population. An Alaskan study reported 28–39% and 4–10% of the Alaskan native adult females and males (≥18 years), respectively, had ID as opposed to a representative sample of 15% and 1.5% prevalence of ID among American females and males, respectively⁹ A later study observed that the prevalence of ID among Alaskan females and males was 4 and 13 times greater respectively than females and males in the American general population.¹⁰ In Canada, a recent health survey estimated an prevalence of 10.8–18.0% ID among Inuit preschoolers in Nunavik region,¹¹ which is higher than the reported prevalence of 4.5% ID among 3–5 year old American children.¹² Until recently, no information was available from a representative survey regarding the prevalence of ID among Canadian Arctic adult Inuit.

Given the suggested prevalence of ID among Inuit and the potentially important physiological regulatory function of iron on n-3 PUFA metabolism in humans, the impact of iron status on HUFA status among Canadian Inuit was investigated. The Inuit Health Survey 2007–2008 provided the opportunity of exploring the link between iron status and HUFA status in this population.

2 Methods

2.1 Location and subjects

The current analyses are based upon data from the International Polar Year (IPY) Inuit Health Survey conducted in the late summer and fall of 2007 and 2008. It was a comprehensive health survey of Inuit residing in three jurisdictions covering a vast territory from the Canadian Eastern Arctic coast to the Western Arctic coast. Participants (≥18 yr), who self-identified as Inuk were recruited through random selection of households stratified by community with pregnant women being excluded. A total of 2796 households were approached from which 1901 household participated with a response rate of 68%. Among 2595 participants, 1511 individuals (58.2%) had available data on erythrocyte membrane fatty acids, serum ferritin and hemoglobin and whose C-reactive protein (CRP) level was less than 3.0 mg L⁻¹. The project was reviewed and approved by McGill Institutional Review Board and the three Inuit jurisdictions. In the comparison of the 1511 individuals included in the analyses with those excluded due to missing data, no clinical differences were observed in terms of average age and percent males.

2.2 Laboratory analyses

Weight and body fat percentage were measured using a bioelectrical impedance scale (Tanita Inc, Tokyo, Japan). Height was measured without shoes using a stadiometer with the patient standing on a hard surface. BMI was calculated (kg m⁻²). Waist circumference was measured at the end of a normal expiration with the tape placed horizontally between the last floating rib and the top of the hip.

Erythrocyte membrane fatty acid concentrations were determined by gas-liquid chromatography (Lipid Analytical Laboratories Inc, Guelph, Canada) as described previously.¹³ The fatty acid methyl esters were analyzed on a Varian 2400 gas-liquid chromatograph (Palo Alto, CA) with a 60 m DB-23 capillary column (0.32 mm internal diameter). The inter-assay coefficient of assay was less than 2%, 3% and 3% for primary SFA (C16:0 and C18:0), n-6 (C18:2 n-6, C20:3 n-6 and C20:4 n-6) and n-3fatty acids (C20:5 n-3 and C22:6 n-3), respectively.

Hemoglobin concentrations of the freshly collected whole blood were measured by HemoCue 201+ analyzers (HemoCue, Inc., Lake Forest, CA, USA). The daily coefficient of variation from 4 replicates of the quality control test was less than 3% and the daily inter-assay coefficient of variation was no more than 1.3%. Serum samples were separated from fasting blood samples and stored at −80 °C until analysis. CRP was measured by a highly sensitive Near Infrared Particle Immunoassay rate methodology with SYNCHRON Systems (Beckman Coulter, Mississauga, Canada) with a detection limit at 0.2 mg L⁻¹. The quality control sample was run every eight hours with the mean measured values of different levels of the standard always within expected ranges and the inter-assay coefficient of variation was less than 10%. Ferritin was measured by a sandwich chemiluminescence immunoassay on a LIAISON Analyzer (DiaSorin S.p.A., Saluggia (Vercelli), Italy) with detection limit <0.5 ng mL⁻¹. Mean daily inter-assay coefficient of variation was 1.4% (range = 0.2–4.1%) for quality control samples measured in duplicate.

2.3 Statistical analysis

As there were no gender related differences regarding the fatty acid composition of erythrocyte membranes and no observed interactions of gender and geographic location on the associations among variables of interest, the results were presented in a combined manner for all subjects. Spearman correlations were conducted, including the correlation between Δ5 and ferritin adjusted for age, waist, CRP and HUFAn-3. The age, waist and region adjusted differences between subjects with and without ID were evaluated by Generalized Linear Model (GLM) and post-hoc comparisons were conducted with Bonferroni adjustment. Waist was examined but not included in the final analyses. The activity of Δ5 was estimated by the ratio of C20:4 n-6/C20:3 n-6. Unlike in vitro experiments or animal studies, direct measure of the activity of human Δ5 is extremely difficult in clinical experiments, therefore the activity of Δ5 as well as the activity of other enzymes in fatty acid biosynthesis is typically estimated in human studies by the ratio of the fatty acid product to precursor fatty acids. ID was defined according to recommended cut-off values derived from NHANES.¹⁴ In order to avoid the results being biased by inflammation that can elevate serum ferritin levels, the data were evaluated by adjusting for serum CRP in correlation analyses and by excluding participants with a CRP greater than 3.0 mg L⁻¹ in all the analyses.¹⁵ All P values were obtained from 2-sided tests. Data were analyzed with the SAS software (version 9.2; SAS Institute, Cary, NC).

3 Results

Overall 20.4% of women and 3.1% of men were defined as having ID. Overweight and obesity was prevalent in both genders (BMI: 25.7 ± 5.1 and 27.4 ± 5.6 for participants with ID and without ID, respectively). Serum ferritin showed marked differences between participants with and without ID: the median (25th percentile – 75th percentile) ferritin level was 6.6 (4.8–9.5) μg L⁻¹ and 35.7 (21.6–71.7) μg L⁻¹ for participants with ID and without ID, respectively.

In contrast to other fatty acids, n-3fatty acids showed strong and positive associations with age (r = 0.413 for n-3 HUFA, P < .0001) (Table 1). Similarly, Δ5 and serum ferritin also demonstrated strong correlations with age (both r > 0.3, both P<.0001). The correlation between hemoglobin and age, however, was too weak to be considered clinically important (r = −0.069, P < .01). After adjustment for age, waist and CRP, ferritin showed moderately strong correlations with n-3 HUFA (r = 0.172, P < .0001) and Δ5 (r = 0.195, P < .0001). After further adjusting for n-3 HUFA, ferritin and Δ5 still displayed a significant although weak correlation (r = 0.126, P < .0001).

Table 1 Spearman correlation coefficients (r_s) of fatty acid classes of erythrocyte membranes and Δ5 with serum ferritin and hemoglobin adjusted for age, waist and CRP

	Age			Ferritin ^a			Ferritin ^b
	r s	P		r s	P		r s	P
a Adjustment for age, waist circumference and CRP (n = 1507 unless indicated else). b Adjustment for age, waist circumference, CRP and HUFAn-3. c n = 1505. d n = 1506. e n = 1510.
Δ5	0.315	<.0001	Δ5	0.195	<.0001	Δ5	0.126 ^n = 1506	<.0001
SFA	0.041	ns	SFA	−0.089	<.01
MUFA	−0.066	<.05	MUFA	−0.146^c	<.0001
TFA	0.014	ns	TFA	−0.140	<.0001
n-6	−0.284	<.0001	n-6	0.077	<.01
HUFA n-6	−0.110	<.0001	HUFA n-6	0.154	<.0001
n-3	0.411	<.0001	n-3	0.167	<.0001
HUFA n-3	0.413^e	<.0001	HUFA n-3	0.172^d	<.0001
Ferritin	0.356	<.0001
Hemoglobin	−0.069	<.01

---

The HUFA composition of erythrocyte membranes for both n-3 and n-6 appeared to be lowest in the presence of ID and the HUFA status increased consistently with the elevation of iron status (both P < .0001) (Table 2). Such differences were not evident on the activity of Δ5, however, estimates of Δ5 in the presence of ID was non-significantly different from that without the presence of ID. Among individuals without ID, estimates of Δ5 was higher in the 3rd tertile of serum ferritin than the 1st tertile (P < .05) (Table 2). Erythrocyte membrane fatty acid profiles of the other fatty acid classes were similar at the different levels of ferritin.

Table 2 Age and region adjusted fatty acid composition of erythrocyte membranes of Inuit Health Survey participants (least square mean ± standard error)

	ID ^a	Non-ID			P
		Ferritin tertile 1	Ferritin tertile 2	Ferritin tertile 3
	(n = 202)	(n = 435)	(n = 435)	(n = 435)
a ID: iron deficiency (ferritin < 12 μg L^−1). b Ferritin: median (25th percentile – 75th percentile). c a for individuals with ID, b, c and d for non-ID individuals at the 1st, 2nd and 3rd tertile of ferritin, respectively; ad indicates significant difference between individuals with ID and non-ID individuals at the 3rd ferritin tertile, and so on. d ^1 P < .05, ^2P < .01, ^3P < .001, ^4P < .0001.
Ferritin ^b (μg L^−1)	6.6 (4.8–9.5)	17.8 (14.7–21.6)	35.7 (30.0–45.3)	92.4 (71.7–132.6)
Age (yr)	33.8 ± 11.4	36.4 ± 12.1	40.0 ± 13.7	46.9 ± 14.9
BMI	25.7 ± 5.1	26.4 ± 5.8	26.9 ± 5.5	28.9 ± 5.2
Female (%)	91.6%	77.0%	50.3%	37.7%
Fatty acids
SFA	44.20 ± 0.38	43.74 ± 0.26	42.76 ± 0.25 ^ac^c^2^d^,bc1	43.29 ± 0.26	<.01
MUFA	25.74 ± 0.21	25.06 ± 0.14 ^ab1	24.66 ± 0.14 ^ac4	24.58 ± 0.15 ^ad4	<.0001
C18:2 n-6	14.35 ± 0.23	14.55 ± 0.15	14.90 ± 0.15	14.15 ± 0.16 ^cd2	<.01
C20:3 n-6	1.29 ± 0.03	1.28 ± 0.02	1.37 ± 0.02^bc1	1.28 ± 0.02	<.05
C20:4 n-6	6.87 ± 0.21	7.20 ± 0.14	7.78 ± 0.14 ^ac2^, bc1	7.99 ± 0.15 ^ad4, bd3	<.0001
HUFA n-6	7.83 ± 0.24	8.18 ± 0.17	8.86 ± 0.16 ^ac2^, bc1	9.08 ± 0.17 ^ad3, bd2	<.0001
C20:5 n-3	1.22 ± 0.08	1.42 ± 0.06	1.50 ± 0.06 ^ac1	1.71 ± 0.06 ^ad4,bd2	<.0001
C22:6 n-3	2.08 ± 0.10	2.36 ± 0.07	2.53 ± 0.07 ^ac3	2.58 ± 0.07 ^ad3	<.001
HUFA n-3	4.48 ± 0.20	5.10 ± 0.14	5.45 ± 0.14 ^ac3	5.81 ± 0.14 ^ad4, bd2	<.0001
Δ5	6.52 ± 0.38	5.86 ± 0.26	6.06 ± 0.26	6.87 ± 0.27^bd1	<.05

---

4 Discussion

The prevalence of iron deficiencies of 3.4% and 26.3% for Canadian Inuit men and women respectively was consistent with reports from previous studies in Alaska.^9,10 The findings of the current study indicate that iron status, as measured by serum ferritin, may have independent associations with Δ5 and thus the synthesis of long chain unsaturated fatty acids. The latter associations, however, were not strong enough to lead to markedly significant differences in the fatty acid profiles of erythrocyte membranes in the present study.

The higher consumption of traditional food in older age groups among Canadian Arctic Inuit has been detailed in the literature,⁴ and was reflected in the results of the current study. The traditional food of Arctic Inuit is mainly composed of meat from marine fish, marine mammals, land animals and birds, which are generally rich in HUFAn-3 and heme iron. The strong association of age with both HUFA n-3 in erythrocyte membranes and serum ferritin reflected the higher consumption of traditional food among the older relative to the younger Inuit. The competition between fatty acids from n-3 and n-6 classes likely resulted in the inverse associations of n-6 with age. The association between age and the estimated activity of Δ5 was driven by correlations of age with fatty acids, which was also noted in an earlier study among Cree of James Bay who are experiencing a similar dietary transition.¹⁶ Despite the rather strong correlation between serum ferritin and age (r = 0.356, P<.0001), hemoglobin was weakly correlated with age indicating that hemoglobin levels are a less relevant biomarker of traditional food consumption than ferritin.

The findings of the present study also indicate an independent but weak association between ferritin and Δ5. This relationship remained significant after adjustment for n-3 HUFA suggesting that it is not attributable to diet. In the current study, the presence of ID was related to the lower levels of n-3 and n-6 HUFA, which may be related to the fact that 1–5% of body iron is present as a component of enzymes such as desaturases.¹⁷ Interestingly, in a study of children 6–11 yrs of age with ID (ferritin 8.2 ± 5.4 μg L⁻¹), iron supplementation for 15-weeks was associated with significant increases in the percent composition of n-3 in erythrocyte membranes but not in estimates of desaturase activity.¹⁸ In a cross-sectional study, adults with low iron status (serum iron <12 μmol L⁻¹ for men, <10 μmol L⁻¹ for women) were shown to have lower ratios of HUFA to short chain PUFA than adults with normal iron levels.¹⁹ More studies regarding other dietary or metabolic variables are needed to evaluate the lack of a significant relationship of ID status with Δ5 in the present study.

The finding that ID is more prevalent among Arctic Inuit than the general population may seem to be counterintuitive given the putatively high consumption of red meat among Inuit. Dietary transition away from nutritious traditional food has raised the concern about inadequate iron intake, which was virtually non-existent previously. Earlier dietary studies in Canadian Arctic Inuit communities showed that about 10% of Inuit females at child-bearing age reported their iron intake below the estimated average requirement.^20,21 On the other hand, adequate iron intake alone may not be sufficient to prevent the occurrence of ID. Although adult Alaskan natives demonstrated adequate intake of biological iron compared to the US general population, ID was higher among the former, which was explained by excessive gastrointestinal tract iron loss as revealed from stool samples from the Alaskan natives.⁹ This conclusion was supported by further studies that showed a high prevalence of chronic active gastritis with H. pylori infection and pervasive occult gastrointestinal bleeding among adult Alaskan Inuit.¹⁰ Furthermore, robust evidence from in vitro and animal models showed that inflammation could trigger the synthesis of hepcidin, a liver derived hormone and a key regulator in iron metabolism, which blocks the iron duodenal absorption and traps iron inside macrophages,²² reduces iron transportation and results in iron-restricted erythropoiesis.²³ Inflammation among Inuit can be caused by H. pylori infection, which was noted to be prevalent among Arctic native populations in a limited number of studies.^9,10,24,25 Similarly, ear and respiratory infections have been commonly reported among Inuit children.¹¹ Household crowding, poor sanitation and low social economic status, which were considered to be risk factors for increased person-to-person spread of infections, were prevalent in Canadian Inuit communities.²⁶

Cautious interpretation of the results from the current study and earlier studies is needed as various confounding factors could limit data interpretation in the present work. The two previous human studies discussed above had small sample sizes. For the current study, the sample size of Inuit with ID may not have been sufficiently large to detect significant differences. Moreover, the use of the conservative approach using Bonferroni adjustment for post-hoc comparison in the present work might have led to an inflated type 2 error. The cross-sectional nature of the current study cannot establish a causal relationship between iron status and fatty acid metabolism.

In summary, the current Inuit health survey results provide weak evidence regarding relationships of serum ferritin with Δ5 and erythrocyte membranen-3 HUFA. The above association, if confirmed in future studies, underlines the double burden upon Canadian Inuit facing rapid health transition. In that regard, reduced consumption of traditional Inuit food high in heme-iron concurrent with a low n-3 HUFA intake could exacerbate chronic disease risk via a potentially reduced level of biosynthesis of n-3 HUFA. Findings of the current study also indicate that factors related to disease such as replacing nutritious traditional food by low quality market foods and poor hygiene conditions can work in tandem to increase disease risk among indigenous people undergoing rapid transitions.

5 Conflict of interest

None of the authors had any financial or personal conflict of interest to disclose.

Non-standard abbreviations

ID	Iron deficiency
HUFA n-3	Highly unsaturated n-3fatty acids

6 Acknowledgement

This study was funded by the Canadian Institute of Health Research, Inuvialuit Settlement Region, Nunavut, and Nunatsiavut IPY Inuit Health Survey Steering Committees. The above sponsors were not involved in study design. The authors thank the IPY Inuit Health Survey research staff and participants. YEZ carried out the data analysis and drafted the manuscript. SK supervised the manuscript drafting. GME designed the study and supervised the data analysis and manuscript drafting. All authors contributed to the interpretation and revision of the manuscript.

7 References

1. S. C. Cunnane and K. R. McAdoo, Iron Intake Influences Essential Fatty Acid and Lipid Composition of Rat Plasma and Erythrocytes, J Nutr., 1987, 117(9), 1514–9.
2. G. I. Stangl and M. Kirchgessner, Different degrees of moderate iron deficiency modulate lipid metabolism of rats, Lipids, 1998, 33(9), 889–95.
3. M. T. Nakamura and T. Y. Nara, Structure, function, and dietary regulation of delta6, delta5, and delta9 desaturases, Annu. Rev. Nutr., 2004, 24, 345–76.
4. H. V. Kuhnlein, O. Receveur, R. Soueida and G. M. Egeland, Arctic indigenous peoples experience the nutrition transition with changing dietary patterns and obesity, J Nutr, 2004, 134(6), 1447–53.
5. E. Dewailly, C. Blanchet, S. Lemieux, L. Sauve, S. Gingras and P. Ayotte, et al. n-3 Fatty acids and cardiovascular disease risk factors among the Inuit of Nunavik, Am J Clin Nutr, 2001, 74(4), 464–73.
6. E. Counil, P. Julien, B. Lamarche, M. L. Chateau-Degat, A. Ferland and E. Dewailly, Association between trans-fatty acids in erythrocytes and pro-atherogenic lipid profiles among Canadian Inuit of Nunavik: possible influences of sex and age, Br. J. Nutr., 2009, 102(5), 766–76.
7. M. Kjaergaard, S. Andersen, M. Holten, G. Mulvad and J. J. Kjaergaard, Low occurrence of ischemic heart disease among Inuit around 1963 suggested from ECG among 1851 East Greenland Inuit, Atherosclerosis, 2009, 203(2), 599–603.
8. M. Gracey and M. King, Indigenous health part 1: determinants and disease patterns, Lancet, 2009, 374(9683), 65–75.
9. K. M. Petersen, A. J. Parkinson, E. D. Nobmann, L. Bulkow, R. Yip and A. Mokdad, Iron deficiency anemia among Alaska Natives may be due to fecal loss rather than inadequate intake, J Nutr, 1996, 126(11), 2774–83.
10. R. Yip, P. J. Limburg, D. A. Ahlquist, H. A. Carpenter, A. O'Neill and D. Kruse, et al. Pervasive occult gastrointestinal bleeding in an Alaska native population with prevalent iron deficiency. Role of Helicobacter pylori gastritis, JAMA, J. Am. Med. Assoc., 1997, 277(14), 1135–9.
11. A. Pacey, H. Weiler and G. M. Egeland, Low prevalence of iron-deficiency anaemia among Inuit preschool children: Nunavut Inuit Child Health Survey, 2007–2008, Public Health Nutr., 2010, 5, 1–9.
12. V. M. Cardenas, Z. D. Mulla, M. Ortiz and D. Y. Graham, Iron deficiency and Helicobacter pylori infection in the United States, Am. J. Epidemiol., 2006, 163(2), 127–34.
13. K. D. Stark and B. J. Holub, Differential eicosapentaenoic acid elevations and altered cardiovascular disease risk factor responses after supplementation with docosahexaenoic acid in postmenopausal women receiving and not receiving hormone replacement therapy, Am J Clin Nutr, 2004, 79(5), 765–73.
14. A. C. Looker, P. R. Dallman, M. D. Carroll, E. W. Gunter and C. L. Johnson, Prevalence of Iron Deficiency in the United States, JAMA. 1997 March 26, 1997, 277(12), 973–6.
15. T. A. Pearson, G. A. Mensah, R. W. Alexander, J. L. Anderson, R. O. Cannon III and M. Criqui, et al. Markers of Inflammation and Cardiovascular Disease: Application to Clinical and Public Health Practice: A Statement for Healthcare Professionals From the Centers for Disease Control and Prevention and the American Heart Association, Circulation. 2003 January 28, 2003, 107(3), 499–511.
16. Y. E. Zhou, S. Kubow, E. Dewailly, P. Julien and G. M. Egeland, Decreased activity of desaturase 5 in association with obesity and insulin resistance aggravates declining long-chain n-3 fatty acid status in Cree undergoing dietary transition, Br. J. Nutr., 2009, 102(6), 888–94.
17. S. A. S. Gropper, J. K. Smith, J. L. Groff. Advanced Nutrition and Human Metabolism. 5th ed.Australia; United States: Wadsworth/Cengage Learning; 2009.
18. C. M. Smuts, H. Y. Tichelaar, P. J. van Jaarsveld, C. J. Badenhorst, M. Kruger and R. Laubscher, et al. The effect of iron fortification on the fatty acid composition of plasma and erythrocyte membranes in primary school children with and without iron-deficiency, Prostaglandins, Leukotrienes Essent. Fatty Acids, 1994, 51(4), 277–85.
19. M. Krajcovicova-Kudlackova, J. Klvanova and M. Dusinska, Polyunsaturated Fatty Acid Plasma Content in Groups of General Population with lowvitamin B6 or low iron serum levels, Ann. Nutr. Metab., 2004, 48(2), 118–21.
20. H. V. Kuhnlein, O. Receveur, R. Soueida and P. R. Berti, Unique patterns of dietary adequacy in three cultures of Canadian Arctic indigenous peoples, Public Health Nutr., 2008, 11(4), 349–60.
21. P. R. Berti, R. Soueida and H. V. Kuhnlein, Dietary assessment of Indigenous Canadian Arctic women with a focus on pregnancy and lactation, Int J Circumpolar Health, 2008, 67(4), 349–62.
22. M. D. Knutson, Iron-sensing proteins that regulate hepcidin and enteric iron absorption, Annu. Rev. Nutr., 2010, 30, 149–71.
23. T. Ganz and E. Nemeth, Hepcidin and disorders of iron metabolism, Annu. Rev. Med., 2011, 62, 347–60.
24. I. McKeown, P. Orr, S. Macdonald, A. Kabani, R. Brown and G. Coghlan, et al. Helicobacter pylori in the Canadian arctic: seroprevalence and detection in community water samples, Am. J. Gastroenterol., 1999, 94(7), 1823–9.
25. C. N. Bernstein, I. McKeown, J. M. Embil, J. F. Blanchard, M. Dawood and A. Kabani, et al. Seroprevalence of Helicobacter pylori, incidence of gastric cancer, and peptic ulcer-associated hospitalizations in a Canadian Indian population, Dig. Dis. Sci., 1999, 44(4), 668–74.
26. A. L. Jenkins, T. W. Gyorkos, K. N. Culman, B. J. Ward, G. S. Pekeles and E. L. Mills, An overview of factors influencing the health of Canadian Inuit infants, Int J Circumpolar Health, 2003, 62(1), 17–39.

---