Home » Blog » Clinical Articles » Testing For Iron

Testing For Iron

Blood tests for iron status are among the most common requested in clinical medicine. This may be largely justified because of the prevalence of iron deficiency combined with a relatively common genetic condition of haemochromatosis.

In Australia, iron deficiency, defined by the Royal College of Pathologists of Australasia (RCPA) as a ferritin level below 30 ug/L, affects only 3.4% of men but 22.3% of women according to the Australian Bureau of Statistics survey in 2011-2012. The issue in women is particularly related to premenopausal women (16-44 years) where 34.1% are iron deficient.

This is not surprising when nutrition surveys show that 40% of premenopausal women have inadequate dietary iron intake.

Despite this high prevalence, screening with iron studies is not currently recommended in any demographic. While many hospitals include a ferritin in the shared care antenatal panel, most antenatal guidelines assume that an FBE will detect iron deficiency (which is probably wrong). Anaemia is a late stage of iron deficiency and ideally not a stage we should be waiting for.

While it is true that microcytosis of red cells is often found in iron deficiency this is unreliable as

The unevenness (or high red cell distribution width / RDW) is a more sensitive test of early iron deficiency. The association of B12 deficiency and iron deficiency, especially in vegetarians, is so important that clinicians should always think of the other when the other is detected.

It is estimated that one in eight Australians carry the predisposition to haemochromatosis. It is most common in British / Celtic peoples (C282Y or H63D are the common HFE gene mutations).

When two HFE heterozygotes have children, one in four of the offspring will be homozygote therefore roughly (1/8 * 1/8 * 1/4 =) 1/256 Australian are homozygote – but only half develop disease. This may be because many have been protected from iron overload through diet or blood loss, such as blood donation. Even at a ‘disease’ prevalence of 1:400 to 1:500, haemochromatosis is a relatively common condition with significant potential morbidity that must be considered, especially in all relatives (first degree relatives can be gene tested without iron studies).

‘Iron overload’ is a little more awkward to define than iron deficiency. Serum ferritin levels above the population norms are not necessarily harmful, but if we waited for serum ferritin levels to reach dangerous levels (eg >1000 ug/L), we would not be preventing the sequelae of iron overload such as liver disease, but also a higher risk of cardiovascular disease and premature arthropathy.

Most labs have upper clinical decision limits for ferritin of between 200 and 500 ug/L as a sensitive early warning for the possibility of haemochromatosis. Should a high ferritin level be confirmed, gene testing can be rebated according to Medicare Benefit Schedule (MBS) requirements.

The pathology tests

I have been discussing serum ferritin as the marker of iron stores however clinicians in Australia commonly request ‘iron studies’. Indeed, ferritin is the storage protein for iron that ‘leaks’ out of cells and most accurately reflects cellular iron stores.

What is the value of the other two measurements?

One of the other measurements is serum iron and it is a bad measure of iron status (we probably shouldn’t report it at all). The serum iron level depends on meals, depends on the time of day (lower in the afternoon) and most importantly, depends on the concentration of the protein that chaperones iron in the circulation: serum transferrin.

Patients with higher transferrin levels will generally have higher serum iron levels. What is important is how iron is the transferrin carrying and this is calculated as the ‘transferrin saturation’ (a ratio of serum iron to transferrin).

Typically transferrin saturation is at least 10% full, and uncommonly more than 45% full and levels outside this are supportive of iron deficiency and iron overload respectively. While the transferrin saturation calculation corrects some of the unreliability of serum iron, saturation is still subject to diet and supplements and diurnal variation.

Clinicians in Australia are used to requesting the full iron study panel of tests. This is useful in iron overload because in haemochromatosis, the earliest change is a high transferrin saturation which may be found years before the ferritin rises above the upper decision limit. A confirmed elevation of transferrin saturation is also allows haemochromatosis gene testing to be MBS rebated.

Iron deficiency can be identified by a low serum ferritin (less than 30 ug/L) and the rest of the iron studies may also be altered with low serum iron saturation and higher levels of transferrin. Low serum iron saturation is non-specific (eg diet and afternoon samples) and high transferrin is also non-specific (eg OCP and pregnancy). Unfortunately there are some patients that are misidentified with iron deficiency because of these non-specific tests even when ferritin was clearly normal and there is discussion of banning the ability to request serum iron and transferrin when looking for iron deficiency because of the potential harms in misinterpretation.

For clinicians there are even more important confounders than the physiological effects on serum iron and transferrin saturation because when inflammation (the ‘acute phase reaction’) is present the body actually hides away its iron stores by decreasing iron release (low serum iron), decreasing transferrin production (ie negative acute phase reactant), and because iron is no longer being mobilised, it starts accumulating in cells (ferritin rise as if it were an acute phase protein).

Unfortunately all the iron studies are therefore unreliable in the presence of inflammation and if there is some suspicion a serum CRP is the most sensitive and specific test to detect inflammation. All we can say otherwise is

There is a test that helps separate true iron deficiency in anaemic patients with inflammatory disorders called ‘soluble serum transferrin receptors’ but it is not covered in the Medicare Benefits Schedule although the RCPA have made a submission to government.


General Practice Pathology is a new regular column each authored by an Australian expert pathologist on a topic of particular relevance and interest to practising GPs.

The authors provide this editorial, free of charge as part of an educational initiative developed and coordinated by Sonic Pathology.