Finding the Goldilocks zone
I closed Part 2 of this series with a teaser: how do we know whether we’re deficient or depleted in iron, and what is an ‘optimal’ level of storage iron, or ferritin?
It’s an important question to answer accurately, because on the one hand, above-optimal iron stores are associated with a host of disease conditions (more on that, shortly), and on the other hand, suboptimal iron stores are damaging too, even in people who don’t have iron deficiency anaemia:
“There is a hierarchy of iron requirements as iron becomes deficient. Hemoglobin production is protected at the expense of other tissue iron requirements, so iron required for brain use, for example, may be compromised even though hemoglobin concentrations are within the normal range. This state is known as iron deficiency without anemia and is indicated in subjects by normal hemoglobin concentrations but low iron stores (reflected by low serum ferritin concentrations). Iron deficiency without anemia is of concern because it affects cognition and endurance and may have negative consequences for immune function.”
Iron: Physiology of Iron, from Encyclopedia of Food and Health
So, what is the iron ‘Goldilocks zone’ – the level of iron stores that’s not too little, not too much, not only for optimising people’s current physical and cognitive function, but also for protecting them against the development of chronic diseases in which iron overload is implicated? When I began writing this post, I assumed that the answer to that question would be relatively easy to find. Boy, was I wrong. It turns out that both the lower and upper bounds of what constitutes a healthy range of serum ferritin concentration, are hotly contested.
In the absence of a definitive answer to this vital question, I decided to attempt to triangulate from several different types of data – reference ranges, absorption studies, and cohort studies in which ferritin levels are correlated with health outcomes.
Let’s start with…
Ferritin reference ranges
There are huge discrepancies between how various authorities define the optimal range for ferritin.
The World Health Organisation defines iron deficiency as a serum ferritin level of less than 15 micrograms per litre (µg/L) in apparently healthy adults, and less than 70 µg/L in adults with infection or inflammation. Increased risk of iron overload is defined as serum ferritin above 150 µg/L in healthy females and above 200 µg/L in healthy males, or above 500 µg/L in females and males with infection or inflammation:
Lifeblood, the division of the Australian Red Cross that collects and supplies blood products, accepts blood donations from female donors with ferritin levels between 15 and 400 µg/L, and male donors with 30–500 µg/L:
Medscape, a US-based website which provides continuing education for physicians and other health professionals, sets ferritin reference ranges of 12-300 µg/L for males and 10-150 µg/L for females.
The Royal College of Pathologists in Australia considers that a serum ferritin level of less than 30 µg/L is diagnostic of iron deficiency in both female and male adults, and that iron deficiency is still possible with ferritin levels of 30-100 µg/L in adults with coexisting inflammatory disease. Hence, most pathology laboratories in Australia have adopted 30 µg/L as their cut-off point for defining iron deficiency.
And a 2024 systematic review of guidelines and papers that determined ferritin reference intervals in healthy adults, concluded that guidelines that set the lower limit of normal ferritin at anything below 30 µg/L are derived from poor quality evidence because (among other reasons) they were based on studies that did not exclude people at risk of iron deficiency due to blood loss:
“Ferritin reference intervals reported in the literature and used by laboratories in the real world are substantially less than 30 µg/L, particularly for females, despite robust evidence suggesting that a ferritin less than 30 µg/L is sensitive (92%) and specific (98%) for the diagnosis of iron deficiency… Therefore, the use of current laboratory ferritin reference intervals will result in underdiagnosis and undertreatment of iron deficiency.”
The origin of ferritin reference intervals: a systematic review
However, a 2021 Cochrane review slammed the paucity and low quality of evidence on which all reference intervals for ferritin are based:
“Our systematic review of ferritin thresholds to define iron deficiency and iron overload revealed very limited data to yield a reliable estimation of the accuracy of serum ferritin as a test for iron deficiency in otherwise healthy populations (i.e. without inflammation and comorbid disease), and essentially no data from applicable studies as a test for iron overload.
Iron deficiency
From the five studies in apparently healthy populations, the relative low DOR [diagnostic odds ratio] of 1.6 indicated an almost uninformative test. Moreover, there was no clear consensus on appropriate thresholds to detect iron deficiency (thresholds ranged from 16 µg/L to 70 µg/L)…
Iron overload
Concerning iron overload, either for one threshold for males and females (36 studies) or one threshold per gender (nine studies), meta‐analyses found a low to moderate mean DOR of 8 and 10, respectively. For both cases, the very low quality of the studies by GRADE assessment diminishes the validity of the estimates.
It seems unlikely that a single serum ferritin threshold could discriminate with high accuracy between either iron‐replete/non‐iron‐overload versus iron‐deficient/iron‐overload participants across all disease groups. So, it is necessary to develop corresponding trade‐offs for each specific condition.”
Serum or plasma ferritin concentration as an index of iron deficiency and overload
The Cochrane reviewers concluded:
“Ferritin concentrations appear consistently lower in patients with iron deficiency compared with iron overload, and low ferritin concentrations therefore indicate iron deficiency. A ferritin threshold below 15 to 30 µg/L appears to indicate absent bone marrow iron stores in healthy populations, but there are too few studies to confidently recommend a particular threshold. Data in critical patient subgroups (e.g. pregnancy, children) are too sparse to be sure that the same broad conclusion holds in all these groups.
The sensitivity of ferritin for diagnosis of iron overload is around 80%, meaning that a negative test correctly excludes iron overload in most people. The specificity appears low, meaning that positive tests do not definitively indicate iron overload.”
Dang. So what we can conclude from the Cochrane review is that if your serum ferritin level is below 15 µg/L, you’re definitely iron deficient; if it’s below 30 µg/L, you’re almost certainly iron deficient, but no one knows exactly where to draw the line below which iron deficiency is a certainty. And as for the upper bound, above which iron overload is a certainty, science says… buggered if we know!
OK. Maybe we can learn something useful about the optimal level of iron stores, from
Iron absorption studies
In studies of absorption rates of highly bioavailable iron that was labelled with a radioiron tracer, conducted in both healthy men and young women, iron absorption rates were inversely correlated with serum ferritin up to a ferritin level of 60 µg/L, meaning that the lower the ferritin level, the higher the rate of iron absorption. At ferritin levels above 60 µg/L,
“absorption decreased to a level just sufficient to cover basal iron losses, implying that at a serum ferritin concentration 60 µg/L no further accumulation of iron stores will occur by dietary iron absorption.” [my emphasis]
Iron absorption from the whole diet in men: how effective is the regulation of iron absorption?
These findings suggest that there is a natural ceiling of around 60 µg/L of ferritin, and that the human body has adaptations to prevent it from accumulating diet-derived iron stores above this level. The implication is that ferritin levels above 60 µg/L are not due to increased absorption of iron from the diet, but instead are related to some pathological process, which may explain the association between various disease states and higher ferritin levels (see next section). However, these iron absorption studies were quite small, enrolling only 40 men and 21 women, limiting the confidence with which any conclusions drawn from them can be held.
OK, so what could we learn if we had access to data from many thousands of people with a wide range of ferritin levels? That brings us to…
Cohort studies tracking health outcomes
Cohort studies that recruit large numbers of people who are relatively healthy at baseline, and follow them up for many years or even decades, are an excellent way to explore the relationship between iron stores and various health outcomes.
Type 2 diabetes
Frequent blood donors have lower iron stores (as measured by serum ferritin), higher insulin sensitivity, and a lower risk of developing type 2 diabetes. Conversely, people who have iron overload due to hereditary haemochromatosis are well known to be at increased risk of developing type 2 diabetes.
Excess iron induces oxidative damage to the insulin-secreting cells, impairs insulin sensitivity and increases insulin resistance. Multiple cohort studies suggest that these mechanisms may cause type 2 diabetes in people who don’t have hereditary haemochromatosis, but do have high iron stores:
- In 1613 men enrolled in the Kuopio Ischemic Heart Disease Risk Factor study, who were aged 42-60 and were initially free of type 2 diabetes and hereditary haemochromatosis, those who remained diabetes-free had an average serum ferritin of 151 µg/L, while those who went on to develop type 2 diabetes had an average serum ferritin of 191 µg/L. A marked increase in risk began at around 185 µg/L. However, while higher body iron stores were associated with increased risk of developing diabetes, iron depletion was not protective, probably because iron plays a crucial role in glucose metabolism.
The relationship between serum ferritin and diabetes risk persisted after correcting for serum C-reactive protein, an inflammatory marker. You may remember from Part 2 that ferritin is an acute phase reactant, meaning that it goes up in response to inflammation and hence, is not an accurate indicator of iron stores during inflammatory episodes. But these results indicate that it was high iron stores, not inflammation, that mediated the association between elevated ferritin and diabetes risk. - Among 32 826 women aged 30 to 55 years in 1989-1990, who participated in the Nurses’ Health Study, those who developed type 2 diabetes during 10 years of follow-up had an average serum ferritin of 109 µg/L, vs 71.5 µg/L in diabetes-free participants who were matched to cases on age, race, fasting status for the blood test, and on body mass index for those in the top decile for BMI. Once again, the association persisted after adjusting for C-reactive protein. And the women who went on to develop type 2 diabetes also had a lower ratio of transferrin receptors to ferritin, which is another indicator that their raised ferritin was truly indicative of higher iron stores, rather than inflammation (see Part 2 for a refresher on the relationship between ferritin, transferrin receptors, and inflammation).Figure 1. Relative Risks of Type 2 Diabetes According to Ferritin Concentration, from ‘Body Iron Stores in Relation to Risk of Type 2 Diabetes in Apparently Healthy Women’
Examination of dietary records revealed that participants who became diabetic had a higher intake of haem iron, trans fat, red and processed meats and total calories at baseline, and a lower intake of cereal fibre and magnesium. - In a meta-analysis of 15 prospective studies comprising 77,352 participants from the US, UK, Europe and Asia, 18,404 of whom developed type 2 diabetes during follow-up, there was a dose-response relationship between serum ferritin levels and the risk of developing diabetes. There was also a significant sex difference, such that among participants with elevated ferritin, women had twice the risk of becoming diabetic as men, at the same ferritin level. Once again, the association held even after adjusting for C-reactive protein and other markers of inflammation and metabolic dysfunction, indicating that the accumulation of iron was directly related to increased risk of developing type 2 diabetes.
Figure 3: Sex-specific analysis of circulating ferritin levels and type 2 diabetes risk. (A and B) Forest plots of type 2 diabetes risk per 100-μg/L increase in circulating ferritin level in (A) men and (B) women. (C and D) Forest plots of type 2 diabetes risk per 100-μg/L increase in circulating ferritin level in (C) men and (D) women. Sizes of gray diamond represent the statistical weight that each study contributed to the overall estimate. (E) Sex-specific dose-response analysis of the nonlinear association between circulating ferritin concentration and the risk of type 2 diabetes. Green dashed line represents reference line that RR=1. From ‘Sex-Specific Association of Circulating Ferritin Level and Risk of Type 2 Diabetes: A Dose-Response Meta-Analysis of Prospective Studies’.In men, the risk of developing type 2 diabetes rose with ferritin levels above 140 μg/L, whereas in women, risk began rising with any ferritin level above 10 µg/L.
Cardiovascular mortality
As with type 2 diabetes, it has long been observed that regular blood donors have a lower risk of cardiovascular disease. Furthermore, the doubling of cardiovascular risk that occurs after women go through menopause, has been partially attributed to the fact that once menstrual blood loss ceases, women’s ferritin levels begin to rise, reflecting greater body stores of iron.
- In the Copenhagen City Heart Study of 8988 individuals, 6364 of whom died during a median follow-up period of 23 years, cardiovascular mortality was 20 per cent higher in participants with ferritin of 200 µg/L and above, vs those with ferritin less than 200 µg/L, and there was a stepwise increase in cardiovascular mortality with increasing ferritin concentrations.
- A sample of 5471 participants aged 52 years or over, enrolled in the English Longitudinal Study of Ageing, found that apparently healthy men with serum ferritin concentrations over 193 µg/L had more than double the risk of cardiovascular mortality, compared to men with ferritin levels between 119 and 193 µg/L, even after adjusting for inflammatory markers.
All-cause mortality
- The Copenhagen City Heart Study mentioned above also examined the relationship between ferritin levels and total mortality (the risk of dying of any cause, within the follow-up period), and found that each incremental increase in ferritin concentration above 200 µg/L was associated with an incremental increased risk of premature death overall. They concluded that “Moderately to markedly increased ferritin concentrations represent a biological biomarker predictive of early death in a dose-dependent linear manner in the general population.”
- Another Danish study used deidentified health data from 161,921 individuals in the greater Copenhagen area, who were followed for up to roughly nine years, to track the association between ferritin and transferrin saturation, and mortality risk. They found that both low and high ferritin levels were associated with an increased risk of death from all causes, with females being at highest risk of death from high ferritin levels, and males having equally increased all-cause mortality with both high and low ferritin values.
For females without concurrent inflammation (i.e. C-reactive protein [CRP] below 10 mg/L), the ferritin level associated with lowest all-cause mortality was 60 µg/L – which, as you’ll recall, is the exact same ferritin level that was found to be the natural ‘ceiling’ in the iron absorption studies mentioned above. Ferritin levels below 37 µg/L and above 101 µg/L were associated with increased risk of death.
Females with concurrent inflammation (CRP > 10 mg/L) had their lowest risk of death at ferritin levels of 52 µg/L, while risk of death was heightened at ferritin below 16 µg/L and above 68 µg/L.
For males without concurrent inflammation, the ‘mortality sweet spot’ for ferritin was 125 µg/L, while concentrations below 79 µg/L and above 193 µg/L were associated with increased risk of death. Males with inflammation had their lowest risk of death at 118 µg/L, with increased risk at ferritin concentrations below 90 µg/L and above 220 µg/L.Figure 2. Relative hazards for the associations between ferritin level and death for females and males, given absent (CRP < 10 mg/L) or present (CRP > 10 mg/L) inflammation. The dashed lines represent 95% confidence intervals. CRP: C-reactive protein. From ‘Association of ferritin and transferrin saturation with all-cause mortality, and the effect of concurrent inflammation: a danish cohort study’.For transferrin saturation, which you might recall from Part 2 is decreased in iron deficiency, inflammation, and anaemia of chronic disease, and increased in iron overload, the ‘sweet spots’ for lowest mortality were 33.9 per cent and 28.7 per cent for females with and without inflammation, respectively, while for males, it was 32.3 per cent and 30.6 per cent with and without inflammation, respectively.Figure 3. Relative hazards for the associations between TS level and death for females and males, given absent (CRP < 10 mg/L) or present (CRP > 10 mg/L) inflammation. The dashed lines represent 95% confidence intervals. TS: transferrin saturation; CRP: C-reactive protein. From ‘Association of ferritin and transferrin saturation with all-cause mortality, and the effect of concurrent inflammation: a danish cohort study’.Note that in Australia, a ferritin concentration between 25 and 155 µg/L is considered normal in menstruating adult females, while 40-260 µg/L is considered normal in adult males. Transferrin saturation between 20 and 50 per cent is considered normal in both sexes. These ranges include values at which the Danish study found increased risk of death.
Pulling it all together…
The precise bounds of the ‘Goldilocks zone’ for serum ferritin – the most widely-used marker of body iron stores – are still not known, but commonly-used reference ranges are almost certainly too wide, if you’re aiming for optimal health and longevity. Many people with ferritin levels at the lower end of the reference range will be falsely reassured, when in fact they may have early iron deficiency. Conversely, many people may unwittingly be at higher risk of type 2 diabetes, cardiovascular disease and premature death, even though their ferritin levels are well within the upper half of the reference range.
My tentative conclusion is that for females, ferritin levels between 50 and 70 µg/L might be optimal, and for men, 90-140 µg/L.
Finally, no single biomarker can definitively identify either iron deficiency or iron overload. A full blood examination (known as a complete blood count in the US), iron studies, soluble transferrin receptor assay and C-reactive protein will help to clarify your iron status.
