Introduction

Deficiency of a number of vitamins and minerals required for normal erythropoiesis (haematinics) is associated with anaemia. Deficiency of iron, B12 and folate (the most common haematinics) are the most prevalent forms of vitamin deficiency worldwide [1], [2], [3]. Individually and collectively, deficiency of iron, B12 and folate are significant causes of morbidity in the population. The estimated prevalence of haematinic deficiency varies widely with studies from Europe ranging from 5% to 46% [4],[5],[6]. In a recent Canadian population study the estimated prevalence of folate deficiency was close to zero (following flour supplementation) while B12 deficiency was estimated at approximately 5% of the population [7]. By contrast, in the NHANES 2005–2006 study in the US the prevalence of folate deficiency was estimated at 4.5% using a more sensitive measure of red cell folate [8].

While methods for the measurements of B12 and folate deficiency are reasonably well defined, the measurement of iron deficiency is more complex. Diagnosis of iron deficiency based on ferritin (the current standard test) may be enhanced by the addition of serum soluble transferrin receptor (sTfR) assay. Soluble transferrin receptor (sTfR), a transmembrane protein expressed abundantly on erythroid precursors in the bone marrow [9], is of value in identifying iron deficiency [10],[11],[12]. When iron stores are depleted, sTfR levels rise indicating iron-deficient erythropoiesis which precedes the development of anaemia and in some instances precedes the decline in serum ferritin concentrations [13]. Thus sTfR has the potential to identify latent iron deficiency. However, the sTfR assay is relatively expensive and is not widely used.

Anaemia is associated with a number of classic changes in red cell morphology, depending on the underlying cause. Iron deficiency is associated with small red cells (microcytosis) whereas deficiency of B12 or folate is associated with increased red cell size (macrocytosis), expressed as mean cell volume (MCV). Laboratory evidence of microcytosis and macrocytosis and/ or iron, B12 and folate deficiency requires further assessment (including expert morphological assessment of a peripheral blood smear) to investigate a wide range of potential underlying causes. These causes include nutritional deficiency, malabsorption, blood loss, occult malignancy, excess alcohol intake, liver disease, hypothyroidism, hereditary causes of anaemia and rare bone marrow disorders such as myelodysplasia.

While data on the prevalence of haematic deficiency and associated red cell morphological abnormalities are well documented in general population samples worldwide, data on the prevalence and relative importance of underlying conditions are relatively sparse. Data from the Republic of Ireland are needed to inform national policy on nutrition and health including the issue of folate supplementation of flour. There is a dearth of population-based data on the prevalence and determinants of macrocytosis -a frequent finding in clinical practice and associated with significant resource use in follow-up investigations.

The aim of this study was to estimate the prevalence and major determinants of haematinic deficiency and macrocytosis in linked, anonymised blood samples drawn from a national population-based study of men and women aged 45 years and older, including expert morphological assessment of red cell morphology and assessment of a wide range of potential underlying causes.

Methods

Results

Table 1 shows the breakdown of core socio-demographic characteristics of the study sample and relevant clinical variables by gender. The men were younger and there were significant differences in social class, educational attainment, smoking status and alcohol intake by gender. The data showed that 31.3% of men and 13.2% of women had evidence of abnormal liver function based on elevated gamma glutamyl transferase (GGT) concentrations (GGT> 48 u/L). A total of 3.9% of the sample had TSH levels consistent with hypothroidism and 0.9% of the sample tested positive for coeliac disease, (Table 1).

Table 2 shows estimated dietary intakes of iron, folic acid and vitamin B12 by gender and age group. Iron intakes were higher in men than in women and a substantial proportion of the population sample had estimated iron intakes below the Recommended Daily Allowance (RDA) ranging from 31% in men aged 45-64 years to 70% in women in the 45-64 age group. Small but statistically significant gender differences in estimated dietary folic acid and vitamin B12 intakes were also observed (Table 2).

Mean haemoglobin concentrations were higher in men than in women. However the prevalence of anaemia was also higher in men (4.6%) than in women (1.0%) reflecting the different cut-off points for anaemia in men and women, (Table 3). As expected, the prevalence of iron deficiency based on low serum ferritin was considerably higher in women (8.7%) than in men (3.7%). Similarly, the prevalence of raised sTfR (>21nmol/ml) consistent with iron deficiency was higher in women (7%) than in men 3.2%. In men with estimated iron intakes below the RDA, 68.4% had low serum ferritin and 80.8% had an elevated sTfR. Similarly, 63.1% of women with estimated iron intakes below the RDA had low serum ferritin and 67.9% had an elevated sTfR. Based on both low ferritin (<17ng/ml for men and <11ng/ml for women) and raised sTfR, 2.3% (0.9% men and 3.8% women) were iron-deficient on both criteria, 3.9% (2.8% men and 5.0% women) were iron-deficient by ferritin but not sTfR and 2.8% (2.3% men and 3.2% women) were classified as deficient by sTfR but not ferritin (latent iron deficiency) –see Figure 1. The estimated prevalence of iron deficiency defined on the basis of either low ferritin or raised sTfR was 9%.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Distribution of ferritin in subjects with normal sTfR (=<21nmol/L) and those with elevated sTfR (>21nmol/L).

https://doi.org/10.1371/journal.pone.0077743.g001

Overall, 3.0% and 2.7% of the sample had low serum folate (<2.3ng/ml) and serum B12 (<120ng/ml) concentrations respectively with no significant gender differences in folate concentrations but a higher prevalence of Vitamin B12 deficiency in women (Table 3). Mean serum folate in those with folate intakes below the RDA was 2.0ugm/day versus 8.6ugm/day in those with estimated intakes above the RDA. Mean serum B12 in those with estimated dietary B12 intakes below the RDA was 99.0ug/day versus 298.8ug/day in those with estimated intakes above the RDA. Subjects aged 65 or over had a significantly higher prevalence of Vitamin B12 deficiency compared to subjects aged 45-64 (5.0% vs. 1.6% respectively, p=0.01). On cross tabulation of the serum folate and serum Vitamin B12 data, only two subjects had both low B12 (<120ng/l) and folate (<2.3ng/ml) concentrations.

Clinically significant macrocytosis (MCV>99fl) was detected in 8.4% of participants whereas microcytosis was uncommon, Table 3. No evidence of myelodysplasia or other bone marrow disorders were detected from morphological assessment of blood films. Macrocytosis was more common in men and in younger participants. In univariate and multivariate analyses,(Table 4) strong and highly significant associations with macrocytosis were observed for lower social status, current smoking status and alcohol intake above the recommended limits, elevated gamma glutamyl transferase (GGT) levels, deficiency of folic acid and vitamin B12, hypothyroidism and coeliac disease. The association between smoking and macrocytosis was not attenuated on adjustment for alcohol and GGT. The population attributable fractions (PAF) for major factors associated with macrocytosis were as follows: alcohol intake above recommended limit 6.3% (95% CI: 4.3% - 7.6%), elevated GGT 25.0% (95% C.I: 22.5% - 27.5%), current smoking status 24.6% (95% CI: 22.1% - 27.0%), deficiency of folic acid 10.5% (95% C.I: 8.8% - 12.3%) and vitamin B12 deficiency 3.4% ( 95% C.I: 2.4% - 4.5%).

Discussion

This study provides data on the prevalence of haematinic deficiency and macrocytosis and the relative importance of major underlying causes of relevance to both clinical practice and public health nutrition policy. Based on the laboratory reference ranges for haemoglobin, the estimated prevalence of anaemia was higher in men (4.6%) than in women (1%). However the prevalence of iron deficiency was as expected substantially higher in women than in men. The findings highlight the importance of nutritional deficiency as a potential cause of iron, B12 and folate deficiency in a developed country setting. Consistent with previous work, coeliac disease is not an important contributor to the burden of iron deficiency in Ireland [22]. In the context of iron deficiency, the findings also highlight the impact of using the soluble transferrin receptor assay in addition to serum ferritin on the estimated prevalence of iron deficiency. The findings confirm the role of excess alcohol intake, abnormal liver function, hypothyroidism and coeliac disease as determinants of macrocytosis in addition to serum B12 and folic acid deficiency. However the findings on smoking and macrocytosis were unexpected. In this study smoking was on a par with abnormal liver function as a contributor to the population prevalence of macrocytosis, accounting for approximately 25% of cases, substantially higher than excess alcohol intake (6.3%), deficiency of folic acid (10.5%) or vitamin B12 (3.4%.).

The prevalence of anaemia is high in Ireland, especially in men. In Ireland laboratories use different lower cut off points below which anaemia is defined –varying from 13-14gm/dl in men and 11.5-12.5gm/dl in women. Based on both low ferritin and raised soluble transferrin receptor assay (sTfR), 2.3% of the sample were iron-deficient on both criteria whereas 9.0% could be classified as iron-deficient on the basis of either low ferritin or raised sTfR. Thus it is clear that the addition of sTfR to current clinical screening protocols would substantially increase the prevalence of diagnosed iron deficiency. However, if sTfR was added to serum ferritin as a confirmatory test for iron deficiency prior to referral for specialist investigation (such as endoscopy) the number investigations would decrease with potential reductions in patient morbidity and cost to health services. The sTfR assay is an expensive test and not as well validated in large population samples as ferritin [12]. Nonetheless, if the sTfR assay was introduced into routine clinical practice it would be widely requested as suspected iron deficiency is common. However, if the test was introduced in parallel with a reduction in the lower limit of haemoglobin ranges and a sex specific normal range for ferritin, it would add a margin of safety to ensure that genuinely latent iron deficiency was not ignored – an important issue in the elderly where cancer becomes a more common cause of anaemia. Because ferritin levels rise as part of the acute phase response, with fatty liver and with increased alcohol consumption, this masks co-existing iron deficiency, further highlighting the potential utility of a sTfR. Our data suggest that further analysis of the use, economic benefit and impact on clinical outcomes of the sTfR assay is warranted.

Numerous studies have demonstrated that vitamin B12 deficiency increases with age, [23],[24], with reports of between 10% to 15% of people aged over 60 years with low vitamin B12 levels [25]. By comparison the prevalence of B12 deficiency in this Irish population sample is low at 2.7%. Levels Vitamin B12 of between 120 to 135 ng/l which are reported as indeterminate in the participating laboratory were classified as normal in this study. However, only 1.6% of participants had vitamin B12 concentrations within this indeterminate range. Although published data suggests that the age related decrease in B12 levels may be attributed to poor intake or increased losses [24], estimated mean dietary intakes in both middle-aged and older participants in this study were over 5.0 µg/day which is more than twice the recommended daily intake of 2.4ug/day for men and women [26]. Thus it is likely that the low prevalence of B12 deficiency observed in this study reflects relatively high intakes of cobalamin in the middle-aged and older Irish population.

Low folate levels were also uncommon among the Irish population. However, in contrast to vitamin B12, this estimate concurs with other internationally published data on the prevalence of folate deficiency [27],[24]. The estimated mean daily intake of folic acid amongst this Irish cohort was 339µg. This is well above the recommended daily allowance of 200ug/day [26], for male and females (excluding peri-conceptional females). This study found that folate levels did not increase with age such that only 2.1% of 65 years and over were found to be folate-deficient in contrast to 3.5% with low folate levels in the 45-64 age group. In Ireland, folic acid fortification is voluntary but widespread. One of the concerns surrounding mandatory folate fortification is the potential masking of B12 deficiency in those afflicted with cobalamin deficiency [6]. This is unlikely to be a significant issue in Ireland given the current findings on vitamin B12 intakes and the low prevalence of B12 deficiency. Concern has also been raised regarding the role of unmetabolized folic acid with the possibility that excess levels of oxidised unmetabolised folate could promote cancer growth [28]. The assay used in this study detects all forms of folate including the unmetabolised fractions and 0% of participants had folate levels above the current reference range of 20ng/ml.

The estimated prevalence of macrocytosis at 8.4% is higher than previously reported estimates of between 1.7% and 3.6% [29]. Of those participants found to be B12 and folate-deficient an extremely small proportion (0.1%) had a raised MCV, a finding which is significantly lower than reports from other groups [30]. While the association between smoking and macrocytosis is under-appreciated it has been previously described. In a paper from Meade’s group published in 1978 a small independent effect of smoking, alcohol and the oral contraceptive pill on mean cell volume (MCV) was noted [31]. Subsequently, Eschwege et al in 1979 described a more marked elevation of MCV in patients with alcoholic liver disease who were also smokers [32]. A later study describes small, but significant increases in MCV with smoking [33]. In the current study the risk of macrocytosis was increased approximately three-fold in smokers relative to non smokers. It may be objected that this association is confounded by alcohol intake. However in further multivariate analyses (not shown) the association between smoking and macrocytosis was essentially unchanged (marginally stronger) on adjustment for age, gender, alcohol intake and GGT, with the latter fitted as a continuous variable to minimise residual confounding. A possible effect of smoking on aminolevulinic acid synnthetase has been postulated little research has been done in this area [34].

This study has significant strengths and some important limitations which should be noted. The study sample is nationally representative and data are available on a wide range of relevant nutritional, lifestyle, socio-demographic and clinical variables which have allowed for a relatively detailed and comprehensive analyses of the determinants of haematinic deficiency and macrocytosis. The findings provide important reference data for other studies drawing on general population samples. A particular strength of our study is the availability of morphological analysis of the blood smears by expert reviewers, allowing us to exclude significant evidence of haemolytic anaemia and thalassaemia (conditions associated with enhanced erythropoietic activity) and increased sTfR levels not associated with iron deficiency. However, the sample size is relatively small for a national survey and no data on young adults were available. In addition, data on clinical follow-up for occult malignancy are lacking and we do not have data on inflammatory markers (C-reactive protein or erythrocyte sedimentation rate), which are important in the interpretation of serum ferritin concentrations. Mild EDTA change was widespread on the films examined. However, this mirrors the situation in most haematology laboratories which accept samples referred from community settings.

Summary

We have described the prevalence and major nutritional, lifestyle and clinical determinants of haematinic deficiency and macrocytosis in a representative sample of Irish middle-aged men and women. The use of the soluble transferrin receptor assay has considerable potential in the routine assessment of iron deficiency and there is a need for additional studies addressing the overall cost effectiveness of this assay in different clinical settings. The findings on macrocytosis, if replicated in further cross-sectional and cohort studies, suggest that smoking should be added to the list of major causes of this common haematological abnormality.

Acknowledgments

We would like to acknowledge the input of Dr. Muhammad Khan, Department of Haematology, Cork University Hospital, for assistance with reviewing the blood films.