School of Pharmacy & Biomedical Sciences, University of Portsmouth, Portsmouth, Hampshire, UK

Alopecia is a non-life-threatening condition, which may seem trivial to the unaffected. However, those physicians who see patients with hair loss know all too well the devastating impact it can have on an individual's quality of life. The studyof Kantoret al in this issue of the journal (2003) not only has significant implications for dermatology but other areas of biology. While they show a strong trend for those with alopecia to be at the low end of the serum ferritin scale, they raise some fascinating questions about the role of iron in basic biology. Why women with alopecia have low serum ferritin concentrations is of concern to dermatologists, other groups with an interest in conditions that have rapid cell proliferation should also consider the implication of a low serum ferritin. While the effect of iron on blood is relatively well known, its role in other systems is only just emerging.

Although the association between iron and hair loss was first postulated by^{Hård (1963)}, it was not until the early 1990s that the significance of low iron stores as assessed by serum ferritin concentrations in women with diffuse hair loss (androgenetic alopecia, AGA) (^{Rushton et al, 1990}) was first demonstrated (^{Rushton and Ramsay, 1992;Rushton and Fenton, 1992}). Since then the role of serum ferritin in female hair loss has been a controversial topic fiercely debated at international hair meetings. The report by Kantor et al is the first study, based on sound epidemiological principles, to clearly demonstrate an association between low serum ferritin and alopecia. Some may argue the sample sizes are too small, but significant differences were found, adding to the growing body of evidence that serum ferritin is important in hair biology and supporting the largely anecdotal evidence of its role in persistent excessive hair shedding (chronic telogen effluvium, CTE). Whether a low serum ferritin is causative remains to be seen.

Iron is stored mainly in the liver, within the iron storage proteins ferritin and hemosiderin. Many of the key biological functions of iron in living systems rely on the high redox potential enabling rapid conversion between Fe²⁺ and Fe³⁺. Iron stimulates the liver to make ferritin, and serum ferritin provides a reliable estimate of body iron stores (^{Cook et al, 1974}). The majority of functional iron within the body is present in haem proteins, which are involved in oxygen transport and mitochondrial electron transfer. Total body iron averages approximately 3.8 g in men and 2.3 g in women. This difference is reflected in the lower reference ranges for hemoglobin and serum ferritin in "normal" adult females which we have argued could represent pathological iron deficiency (^{Rushton et al, 2001}). This hypothesis has been supported by a large-scale US study that found 38% of women in the San Diego area to be iron deficient (^{Rushton et al, 2002}), confirming a problem far more common than many physicians appreciate.^{Hobbs (1961)}, in a different area of medicine, sug-gested the "normal" adult female range for hemoglobin was not physiological and iron therapy should be instituted in all women with a hemoglobin level below 136 g/L; the current lower limit is 120 g/L. Kantor et al's findings of a significantly lower hemoglobin and serum ferritin concentrations in women with telogen effluvium (TE) under the age of 40 compared with controls is noteworthy.

The deleterious effects of iron deficiency are partly due to impaired delivery of oxygen to the tissues and to a deficiency of iron-containing compounds (^{Hallberg, 1982}). Clinical features include restlessness and irritability (^{Dillmann et al, 1979}), lower IQ scores in adolescent girls (^{Bruner et al, 1996;Nelson, 1999}), fatigue in nonanaemic women (^{Verdon et al, 2003}), and perhaps more significant, abnormalities in response to infection and impaired T-cell proliferation (^{Dallman, 1986}). The latter is particularly relevant in view of the surprise finding of a significantly lower mean serum ferritin concentration in women with alopecia areata (AA), when they would have been expected to have had higher values due to AA being regarded as an inflammatory condition. This certainly raises some interesting questions, not least: why do those with alopecia totalis/universalis have higher serum ferritin concentrations? The authors speculate on this observation and explain their findings by a "threshold theory," but further investigation is needed. Whether increasing the serum ferritin level will have any meaningful impact on the natural course of AA or if there is a beneficial influence on therapeutic regimens needs to be studied. Further, what are the frequencies of AA, TE, CTE, and AGA in female patients with iron excess, e.g., hemochromatosis? Clearly, answers to these questions are eagerly awaited.

The importance of the Kantor et al work and the potential impact for female alopecia sufferers is clear to see. The evidence that serum ferritin plays an important role in hair loss is becoming more evident. Any guidance that helps the dermatologist deal with patients complaining of hair loss is most welcome, as is understanding the mechanisms involved. The optimal serum ferritin range in women with alopecia has yet to be established. While we await this data, it might be prudent to use parameters obtained from studies of iron staining in the bone marrow, which suggest a serum ferritin concentration of between 30 and 70 g/L, in the absence of inflammation, would be appropriate. Kantor et al provide a firm basis for future research and, if their findings are confirmed, current dermatological practises and investigations involving therapeutic agents will need to be reviewed.