Recommendations

1.1.

Ferritin concentration is a good marker of iron stores and should be used to diagnose iron deficiency in otherwise apparently healthy individuals⁹ (strong recommendation,¹⁰ low certainty of evidence).

1.2.

In individuals with infection or inflammation, a ferritin concentration below 30 μg/L in children and 70 μg/L in adults may be used to indicate iron deficiency (conditional recommendation,¹¹ low certainty of evidence). In populations it is also possible to adjust ferritin values for infection/inflammation by applying correction factors as described for Question 4: Should ferritin be measured in combination with indicator(s) of infection or inflammation?

1.3.

A ferritin concentration exceeding 150 μg/L in menstruating women and 200 μg/L in men and non-menstruating women who are otherwise healthy may indicate a risk of iron overload (conditional recommendation, based on previous WHO recommendation). In adult, non-healthy individuals, a ferritin concentration exceeding 500 μg/L may indicate risk of iron overload (conditional recommendation, very low certainty of evidence).

1.4.

Ferritin concentration should not be used alone to identify risk of iron overload. Patients with elevated ferritin levels should receive clinical and laboratory evaluation to establish the underlying cause (strong recommendation, very low certainty of evidence).

Rationale

The available studies were not sufficient to justify a change in current ferritin cut-off values to define iron deficiency and risk of iron overload by sex or age groups. The recommended cut-off values for ferritin concentrations to define iron deficiency, including previous recommendations and new evidence when available, are presented in Table 1.

Table 1

Recommended cut-off values to define iron deficiency and risk of iron overload in apparently healthy and non-healthy individuals by age group.

Remarks

• In the absence of inflammation, the concentration of plasma/serum ferritin is positively correlated with the size of the total body iron stores. Ferritin levels are low in iron-deficient individuals and high in iron-loaded individuals.
• In populations, ferritin testing to ascertain the prevalence of iron deficiency or to determine the risk of iron overload is usually performed along with haemoglobin testing to assess the prevalence of anaemia. Measures of inflammation (e.g. C-reactive protein [CRP] and/or α-1 acid glycoprotein [AGP]) and additional iron indices, such as soluble transferrin receptor, are commonly used.
• The physiological changes occurring in hormones, blood composition and haemodynamics, as well as in inflammatory status during pregnancy, render it difficult to establish a fixed, unique ferritin concentration to define iron deficiency, especially when comparing to an invasive gold standard test such as bone marrow biopsy.
• Ferritin may be elevated due to iron overload or other causes, including liver disease, obesity, inflammation and malignancy. In cases of risk of iron overload, ferritin concentration only indicates the possibility of iron overload and the need for further assessment of the specific diagnosis, and the severity of the problem.
• Liver biopsies have commonly been used to report iron overload, because the liver is the dominant iron-storage organ, liver iron concentration correlates closely with the total iron balance, and the liver is the only organ in which the iron concentration is elevated in all forms of systemic iron overload. Non-invasive methods such as magnetic resonance imaging and computerized tomography are widely used to assess iron content in the liver.

Recommendation

2.1.

Ferritin concentration increases in response to iron-related interventions and may be used to monitor and assess the impact of interventions on iron status (strong recommendation, moderate certainty of evidence).

Remarks

• Iron interventions should be implemented in such a manner as to enable monitoring to be undertaken with the lowest number of blood draws.
• The interval following initiation of an iron intervention at which ferritin should be measured depends on the type of intervention and the amount of iron provided.
• Comprehensive planning, monitoring and evaluation of all simultaneous interventions that increase iron intake and/or utilization and/or reduce iron losses are required to account for the total amount of iron being received by populations that would result in ferritin changes in cases of iron deficiency, and to avoid risk of iron overload.
• Knowledge of the prevalence of infection/inflammation is critical for interpretation of ferritin concentrations in population surveys and to interpret changes after iron interventions.
• The inclusion of markers to diagnose iron-related genetic disorders is valuable, especially in regions where thalassaemias and other haemoglobinopathies are common.
• Cases of iron overload should be treated at individual level, since high ferritin concentrations are not sensitive to the effects of nutrition interventions.
• More research is needed to evaluate the effect of nutrition interventions on ferritin concentration through the life-cycle, especially during pregnancy, owing to changes in concentration, especially the typical decrease in concentration in late pregnancy.

Recommendations

3.1.

Ferritin may be measured using radiometric, nonradiometric and agglutination assays. One method does not appear to be superior to another and all methods are acceptable if a commutable material traceable to the WHO international reference standard is used to calibrate the assay. Once a method has been selected, that same method should be used for the follow-up of individuals and populations (strong recommendation, moderate/low certainty of evidence).

3.2.

Use of the WHO international reference standard of ferritin is recommended for calibration of all commercial kits and in regular laboratory practice, especially when following up individual cases, for population surveys or to measure the impact of public health interventions (strong recommendation, moderate certainty of evidence).

Remarks

• The risks of radioactive contamination and the high cost of equipment are important drawbacks of radiometric assays.
• For follow-up of individuals and populations, the same method for ferritin determination should be used, to minimize variability. It is also important to control other sources of error in laboratory testing related to handling of samples; transport and storage conditions; the use of manual versus automated procedures; and differences in equipment performance and those inherent to the operator.
• Ferritin may be measured in either serum, plasma or other biological fluids, but the same sample matrix should be used when measuring the impact of interventions in individuals and at the population level.
• International reference materials for ferritin have been developed for calibrating working standards in the routine ongoing assays performed in laboratories and also for evaluating and standardizing new tests for quantification of ferritin. A WHO international standard of ferritin from the National Institute for Biological Standards and Control, WHO International Laboratory for Biological Standards, United Kingdom of Great Britain and Northern Ireland (NIBSC code 94/572), is commercially available and recommended for use with all assays.
• It is important that reference materials are commutable and traceable to the WHO reference standard, so the results are equivalent among procedures and to avoid calibration bias.
• Quality controls should be included with every run, or at least daily, on instruments measuring ferritin. The inclusion of quality controls of low, medium and high ferritin concentrations is desirable.
• Laboratories performing ferritin determinations for patient care or for public health assessments should participate in external quality assurance programmes.

Recommendations

4.1.

In areas of widespread infection or inflammation, serum ferritin should be assessed with the concurrent measurement of two acute phase response proteins, CRP and AGP (strong recommendation, moderate certainty of evidence).

4.2.

The increase in ferritin values caused by inflammation should be accounted for in individuals and populations. One method is to raise the cut-off value that defines deficiency, to 30 μg/L or 70 μg/L, depending on the age group (see Table 1). Another method is to exclude individuals with elevated concentrations of CRP or AGP from prevalence calculations based on ferritin. Alternatively, arithmetic or regression correction approaches may be used to adjust ferritin concentrations for inflammation and apply the cut-off points recommended for healthy populations. The adjustment that best suits the country reality should be selected and used as long as those conditions prevail (strong recommendation, moderate certainty of evidence).

Remarks

• The need for and magnitude of infection/inflammation correction depends on the population group, geographic region and other factors.
• The application of different adjustment approaches will result in a high degree of variability in the estimated prevalence of depleted iron stores. The selected adjustment based on country conditions should be used as long as those conditions prevail.
• Determination of both CRP and AGP concentrations may be important because they reflect different phases of the acute phase response that range from acute infection to chronic inflammation.
• Possible adjustments include the following:

  • the higher ferritin cut-off adjustment approach uses a higher ferritin-concentration cut-off value for individuals with infection/inflammation, e.g. <30 μg/L;
  • the exclusion approach uses the inflammation, malaria-biomarker information, or both, to exclude individuals with elevated inflammation (as defined by a CRP concentration >5 mg/L, AGP concentration >1 g/L, or both) or individuals with malaria infection;
  • the arithmetic correction factor approach applies an arithmetic correction factor by grouping inflammation into groups, e.g. (i) reference (both CRP concentration <5 mg/L and AGP concentration <1 g/L); (ii) incubation (CRP concentration >5 mg/L and AGP concentration <1 g/L); (iii) early convalescence (both CRP concentration >5 mg/L and AGP concentration >1 g/L); and (iv) late convalescence (CRP concentration <5 mg/L and AGP concentration >1 g/L);
  • the regression correction approach uses linear regression to adjust ferritin concentrations by the CRP and AGP concentrations on a continuous scale, and malaria infection as a dichotomous variable. The adjusted ferritin equation is calculated by subtracting the influence of CRP, AGP and malaria as follows:

    $${\text{Ferritin}}_{\text{adjusted}}={\text{ferritin}}_{\text{unadjusted}}-{\mathrm{\beta }}_1({\text{CRP}}_{\text{obs}}-{\text{CRP}}_{\text{ref}})-{\mathrm{\beta }}_2({\text{AGP}}_{\text{obs}}-{\text{AGP}}_{\text{ref}})-{\mathrm{\beta }}_3\text{\hspace{0.17em}malaria}$$

    where β₁ is the CRP regression coefficient, β₂ is the AGP regression coefficient, β₃ malaria is the malaria regression coefficient, obs is the observed value, and ref is the external reference value generated to define low inflammation.

Remarks

• Owing to the scarcity and dispersion of data, it was not possible to make a recommendation for population prevalence ranges to define the magnitude of iron deficiency as a public health problem using ferritin concentrations.
• The population prevalence ranges established for determining the magnitude of anaemia as a public health problem could be suitable as a guide to determine the prevalence ranges for defining the severity of iron deficiency as a public health problem based on adjusted ferritin concentrations. To classify iron deficiency as severe, moderate, mild or no public health problem (measured by ferritin concentration below the recommended cut-off values), the prevalence of iron deficiency could be ≥40.0%, 20.0–39.9%, 5.0–19.9% or ≤4.9%, respectively (see Table 2).
• Initiating iron interventions in populations with a mild, moderate and/or severe prevalence of iron deficiency could help prevent anaemia, as well as adverse consequences of iron deficiency without anaemia.

Table 2

Population prevalence ranges to define the magnitude of iron deficiency as a public health problem using ferritin concentrations.