What Deficiency Causes Dark Circles Under Your Eyes? The Studies Have Numbers

This is a question with an unusual property: it has a real answer, the real answer is partly yes, and
almost every page answering it gets the emphasis backwards. So here are the actual numbers, from studies
that counted.

The 116-patient study, factor by factor

A cross-sectional study in Iran examined 116 patients attending a dermatology outpatient department
between March and August 2022, and tested each proposed risk factor for significance. This is the most
useful single table on the subject.

Read the bottom two rows again. Poor sleep, the explanation offered by essentially everyone including
most articles on this exact search, did not reach significance in the study that measured it.

And the authors’ own conclusion named neither sleep nor a vitamin:

Two key factors associated with this cosmetic problem are rubbing eyes and prolonged exposure
to television.

So what about anaemia? That one is real

Anaemia is the nutritional finding that survives. It was present in 34.5% of those 116
patients at p=0.003, and a separate clinicopathological study of 50 patients in India found
anaemia in 22%, all of them female, noted as pallor on examination.

The mechanism is not mysterious and it is not about pigment. The skin under the eye is the thinnest on
the body, and directly beneath it sits a dense network of small blood vessels. The NHS lists
paler than usual skin among the symptoms of iron deficiency anaemia. Make the skin paler
and more translucent and you make what is behind it more visible. The circle darkens because the vessels
show through better, not because a pigment was deposited.

That is also why the dermatology literature classifies dark circles into four types, one of which is
explicitly vascular, appearing blue, pink or purple, as distinct from the
pigmented type which is brown, the structural type which is skin-coloured
shadow, and mixed. Anaemia acts on the vascular type. It will do nothing for the other three.

And vitamin D

A pilot case-control study compared 100 adults with periorbital melanosis against 100 age- and
sex-matched healthy controls:

• Vitamin D: 17.4 ± 6.5 ng/ml in patients vs 24.1 ± 4.8 ng/ml in controls,
  P<0.001
• HOMA-IR (insulin resistance): 2.20 ± 1.69 vs 0.77 ± 0.20,
  P<0.001
• Triglyceride to HDL ratio: 3.5 ± 2.3 vs 2.1 ± 0.3,
  P<0.001

Those are large, clean separations. And here is what the authors themselves wrote about them:

The exact causal link among POM, IR, and vitamin D needs to be established.

They describe their own work as a pilot study with a single-centre design and a relatively small sample
size, calling for controlled investigations in a large multicentre population. That is a research team
declining to claim what their own significant p-values might tempt someone else to claim on their behalf.

Keep the direction of the arrow open. Insulin resistance and low vitamin D travel together in the general
population for reasons that have nothing to do with eyelids, and a group of people with more metabolic
disturbance may have darker circles for a downstream reason rather than because vitamin D was the cause.

What the comprehensive review lists as causes

The published review of periorbital hyperpigmentation puts it this way:

Periorbital hyperpigmentation is caused by various exogenous and endogenous factors. The
causative factors include genetic or heredity, excessive pigmentation, postinflammatory hyperpigmentation
secondary to atopic and allergic contact dermatitis, periorbital edema, excessive vascularity, shadowing due
to skin laxity and tear trough associated with aging.

Seven causes. Not one of them is a vitamin.

That absence matters, because it tells you where the honest version of this article has to land. Anaemia
and vitamin D show up as associations in patients. They do not show up in the causal account
dermatology gives of the condition. Both things are true at once, and a page that reports only the first is
selling you something.

The part nothing you swallow can reach

A large share of dark circles are not pigment or blood at all. They are shadow.

The tear trough is the hollow running from the inner corner of the eye outward. As the midface ages, the
malar fat pad descends, orbital fat can herniate forward, and the orbital rim becomes more recessed, which
in some people is congenital rather than age-related. The result is a groove, and a groove under an overhead
light produces a dark line. Photograph the same face lit from below and the circle can vanish.

This is the structural type, and it is the reason so many people conclude that nothing
works. Nothing they tried was aimed at the right thing. A supplement cannot fill a hollow, and neither can a
cream. The interventions that address structural circles are fillers and surgery.

What the biopsies showed

The 50-patient study took skin biopsies, which is rare in this literature and worth knowing about. It
found dermal melanin deposition with melanophages along with predominantly increased epidermal
melanin, and increased melanisation at the base of vellus hair follicles in 14% of samples.

One negative result is especially informative: Perl’s Prussian blue stain for haemosiderin was
negative in all biopsies. Haemosiderin is iron pigment left behind when blood leaks into tissue,
and the folk explanation that dark circles are “old blood under the skin” would predict finding it. It was
not there in a single sample.

The authors’ conclusion was that the condition is multifactorial, and probably post-inflammatory
hyperpigmentation secondary to various causes, except in familial cases. In plain terms: for most people the
skin darkened in response to something that irritated it, and for a substantial minority it was inherited.

What to actually do, in order

1. Get the blood test before you buy the supplement. Anaemia and vitamin D are the two
findings with numbers behind them, and both are measurable. The NHS is explicit that iron deficiency anaemia
is diagnosed with a full blood count and that iron tablets should be taken when a doctor recommends them
after diagnosis, not on suspicion. There is a reason for that. Iron supplements taken without deficiency
cause constipation and stomach upset for no benefit, and in people with an undiagnosed iron-loading
condition they are actively harmful.

2. Stop rubbing your eyes. This was tied for the strongest factor in the 116-patient
study at 40.5% and p=0.002, it costs nothing, and it is the one item on the entire list fully under your
control. Chronic friction is a classic cause of post-inflammatory hyperpigmentation, which is exactly what
the biopsy study concluded most cases are. If you rub because your eyes itch, the itch is the thing to
treat.

3. Protect the area from sun. Exposure over five hours daily was significant at p=0.032,
and ultraviolet light drives melanin production wherever it lands. The eyelid area is routinely missed by
both sunscreen and hats.

4. Work out which type you have before spending more. Pull the skin under your eye
gently sideways. If the darkness stretches with the skin, it is pigment. If it stays put and the hollow is
what you see, it is structural. Press gently: if it lightens noticeably, there is a vascular component.
Three different problems, three different answers, and a product for one will do nothing for the other
two.

Frequently Asked Questions

What deficiency causes dark circles under the eyes?

Anaemia has the strongest evidence, present in 34.5% of patients in a 116-patient study with a significant
association (p=0.003), and in 22% of a separate 50-patient series. Vitamin D was significantly lower in
patients than matched controls in a case-control study, 17.4 versus 24.1 ng/ml at P<0.001, though those
authors state they did not establish a causal link. Note the framing: these are deficiencies found more often
in people with dark circles, not deficiencies shown to cause them. The published review of causes lists
genetics, pigmentation, post-inflammatory change, eyelid swelling, blood vessel visibility and tear-trough
shadowing, and no vitamin at all.

Does lack of sleep really cause dark circles?

The study that tested it says no. In 116 patients, sleep quality returned p=0.681, well short of
significance, while family history, eye rubbing, anaemia, thyroid disorder, diabetes, sun exposure and
television time all reached it. Poor sleep does produce fluid retention and eyelid swelling that can deepen
the shadow temporarily, which is probably why the belief is so durable. But as an explanation for persistent
dark circles it did not survive measurement.

Will iron supplements get rid of my dark circles?

Only if you are actually anaemic, and only for the vascular component. Anaemia makes skin paler, and paler
skin over the thinnest skin on the body makes the vessels beneath it more visible. Correcting the anaemia
addresses that. It does nothing for melanin deposits or for the shadow cast by a tear trough. The NHS
diagnoses iron deficiency anaemia with a full blood count and advises taking iron tablets when a doctor
recommends them, not on suspicion, because unnecessary iron causes side effects and is harmful in people with
undiagnosed iron overload.

Is vitamin D deficiency the cause?

Not established, and the researchers who found the association say so themselves. Their pilot case-control
study of 100 patients and 100 controls found vitamin D of 17.4 ng/ml versus 24.1 ng/ml at P<0.001, and
also found markedly higher insulin resistance in the same patients. Their stated position is that the exact
causal link needs to be established, and they list a single-centre design and small sample as limitations.
Low vitamin D and insulin resistance cluster together generally, so the eyelid finding may sit downstream of
metabolic health rather than of vitamin D specifically.

Why do my dark circles run in the family?

Because that is the single strongest association measured. Family history was present in 49.1% of patients
at p=0.002, the highest frequency of any significant factor in that study, and the biopsy study found a
positive family history in 16% of its patients and concluded the condition is post-inflammatory pigmentation
except in familial cases. Inherited facial anatomy determines orbital rim position, skin thickness and
baseline pigmentation, all three of which produce dark circles directly. Genetics is the honest answer for a
large share of people, which is also why it is the answer least often given.

Are dark circles caused by old blood under the skin?

The biopsies say no. Perl’s Prussian blue stain detects haemosiderin, the iron pigment left behind when
blood leaks into tissue, and it was negative in every single biopsy in the 50-patient study. What the same
biopsies did find was dermal melanin with melanophages and increased epidermal melanin. It is melanin, not
residue from blood.

How do I know which type of dark circle I have?

Dermatology classifies them as pigmented (brown), vascular (blue, pink or purple), structural (skin
coloured shadow from a hollow) and mixed. Stretch the skin sideways: pigment moves with the skin, a hollow
does not. Press lightly: a vascular component blanches. Change the lighting: a structural circle can
disappear under light from below. This is worth two minutes because the treatments do not overlap. Depigmenting
agents act on brown, correcting anaemia acts on blue, and only fillers or surgery act on a hollow.

Do eye creams work?

For the pigmented type there is a rationale, since the review lists topical depigmenting agents including
hydroquinone, kojic acid, azelaic acid and retinoic acid among treatment alternatives, alongside chemical
peels, laser therapy and surgical correction. For the structural type there is no mechanism by which a cream
could work, because the problem is the shape of the bone and fat beneath the skin rather than the skin. Most
disappointment with eye creams comes from applying a pigment treatment to a shadow.

Which doctor should I see?

A dermatologist for the circles themselves, and a GP first if you have not had bloods done, because
anaemia, thyroid disorder and diabetes all reached significance in the study above and all three matter well
beyond your appearance. Thyroid disorder was present in 18.1% of patients (p=0.004) and diabetes in 13.8%
(p<0.001). If a cosmetic complaint is what finally gets those checked, it has done something useful.

Sources

• The Frequency of Periorbital Hyperpigmentation Risk Factors. Cross-sectional study of 116 patients at a
  skin outpatient department in Iran, March to August 2022. Family history 57 patients (49.1%, p=0.002); eye
  rubbing 47 (40.5%, p=0.002); anaemia 40 (34.5%, p=0.003); sun exposure over 5 hours 41 (35.3%, p=0.032);
  thyroid disorders 21 (18.1%, p=0.004); television over 5 hours 17 (14.7%, p=0.005); hormonal pills 16 (13.8%,
  p=0.023); diabetes 16 (13.8%, p<0.001). Sleep quality/insomnia p=0.681 and atopy p=0.432, both not
  significant. Authors’ conclusion that two key factors are rubbing eyes and prolonged exposure to television.
  Checked August 2026 —
  PMC11807928
• A Study of Clinicopathological Correlation of Periorbital Hyperpigmentation. Fifty patients, 33 female and
  17 male, mean age of onset 30.44 years with 70% presenting in the third and fourth decades; 16% positive family
  history; 18% atopic; 22% with anaemia, all female, noted as pallor on examination. Histopathology showed dermal
  melanin deposition with melanophages along with predominantly increased epidermal melanin; increased
  melanisation of vellus hair follicle basal layer in 14% of biopsies; Perl’s Prussian blue stain for
  haemosiderin negative in all biopsies. Authors’ conclusion of multifactorial origin, probably
  post-inflammatory hyperpigmentation secondary to various causes except in familial cases. Checked August 2026
  —
  PMC6042186
• Periorbital melanosis and its possible association with insulin resistance and vitamin D deficiency: a
  pilot case-control study. 100 adult patients with periorbital melanosis and 100 age- and sex-matched healthy
  controls. Vitamin D 17.4 ± 6.5 ng/ml in cases versus 24.1 ± 4.8 ng/ml in controls (P<0.001);
  HOMA-IR 2.20 ± 1.69 versus 0.77 ± 0.20 (P<0.001); triglyceride to HDL-c ratio 3.5 ± 2.3
  versus 2.1 ± 0.3 (P<0.001). Authors state the exact causal link among POM, insulin resistance and
  vitamin D needs to be established, and list single-centre design and relatively small sample size among
  limitations. Checked August 2026 —
  PMC11334248
• Periorbital Hyperpigmentation: A Comprehensive Review. PMID 26962392. Causative factors listed as genetic
  or heredity, excessive pigmentation, postinflammatory hyperpigmentation secondary to atopic and allergic
  contact dermatitis, periorbital edema, excessive vascularity, and shadowing due to skin laxity and tear trough
  associated with aging; no nutritional or vitamin cause appears among them. Treatment alternatives listed
  include topical depigmenting agents such as hydroquinone, kojic acid, azelaic acid and topical retinoic acid,
  and physical therapies including chemical peels, surgical corrections and laser therapy. Checked August 2026
  —
  PubMed 26962392
• Infraorbital Dark Circles: A Review of the Pathogenesis, Evaluation and Treatment. PMID 27398005.
  Multifactorial, resulting from deep facial anatomy, soft tissue changes and contributions from the skin;
  classification into pigmented, vascular, structural and mixed types; shadowing from contour deformities
  including maxillary hypoplasia with age-related or congenital recessed inferomedial orbital rims, herniation
  of orbital fat, and malar fat pad deflation and descent. Checked August 2026 —
  PubMed 27398005
• NHS, Iron deficiency anaemia. Symptoms include tiredness and lack of energy, shortness of breath, heart
  palpitations and paler than usual skin; diagnosis via a full blood count; iron tablets to be taken when
  recommended by a GP after diagnosis rather than self-prescribed. Checked August 2026 —
  NHS

All percentages and p-values on this page are quoted from the named
studies and were checked against the published papers in August 2026. Where this article says a study found
an association, that is what it means: an association observed in patients, not a demonstrated cause. The
vitamin D authors’ own statement that the causal link remains to be established applies to every nutritional
figure here. This page summarises published research and is not a diagnosis of your own dark circles.