You Don't Look Tired. You Have Three Separate Things Happening Under Your Eyes.

One complaint, at least three diseases

Ask men what they would change about their face and, after hair, the answer clusters hard around the eyes. The phrasing is almost always the same: I look tired even when I'm not. It is a description of how other people react, not of a lesion, and it hides the single most useful fact about the region — that under-eye darkness is not a condition. It is a symptom shared by several unrelated ones.

The most cited attempt to sort them examined periorbital hyperpigmentation in a large clinic population and proposed four categories. Vascular was the commonest at 41.8%: visible bluish veins and telangiectasia, concentrated at the inner corner. Constitutional came second at 38.6%: a curved band of genuine brown-to-black melanin following the shape of the orbital rim, often involving the upper lid as well. Post-inflammatory pigment accounted for 12%, and shadow effects — darkness produced by contour rather than colour — for 11.4%. Many people have two at once.

That distribution is the whole argument for reading before buying. Roughly four in ten men with this complaint have a blood problem, roughly four in ten have a pigment problem, and about one in ten has an optical illusion produced by bone and fat. A single product cannot address all three, and the market sells a single product.

The mirror test, which takes about ten seconds

Three manoeuvres separate them, and you can run all three in the bathroom.

First, the stretch. Place a fingertip beside the outer corner and pull the skin gently outward, flattening it. Vascular darkness typically becomes more obvious under tension, because stretching the skin thins it further over the vessel bed; the classification literature explicitly uses this — bluish discolouration that becomes more prominent when the overlying skin is stretched is a vascular descriptor. True melanin does not intensify when you pull on it. It travels with the skin.

Second, the horizontal test. Lie flat for five minutes, then look immediately. Venous pooling is gravity-dependent and partially drains; pigment is not and does not. If the crescent lightens noticeably after lying down, you are looking at a circulatory phenomenon and no amount of brightening serum will touch it.

Third, the light-angle test. Stand under a downward ceiling light, then hold a phone torch at eye level directly in front of your face. A structural shadow — the trough where the lower lid meets the cheek — largely disappears under flat frontal light, because a shadow needs a light source above it to exist. Colour does not care where the light comes from. This one is not a validated clinical test; it is simply optics, and it is the fastest way to find out whether you have been trying to treat geometry with a cream.

Most men who have cycled through four eye creams without result get their answer here. The creams were not weak. They were aimed at the wrong category.

Thicker skin, thinner in the only place it counts

There is a genuine sex difference in skin construction, and it is usually deployed to explain why men age later. Measurements across ages 15 to 93 found skin collagen density consistently lower in women than in men, with male skin thickness declining linearly from around age 20 while female thickness held roughly steady until about 50 and then fell. Broadly, male facial skin is thicker and denser.

The eye is where that advantage evaporates. Systematic ultrasound and cadaver mapping of facial skin thickness puts the eyelid at the bottom of the range — the thinnest skin on the face and among the thinnest on the body, on the order of half a millimetre at the lid, and uniquely lacking a subcutaneous fat layer beneath it. A twenty-something percent structural advantage in a region already at the floor of the distribution buys very little. The vessel is close to the surface for everyone.

There is a second asymmetry that is purely mechanical and rarely mentioned. A man who shaves is running a blade across the lower third of his face most mornings, which exfoliates that region continuously whether he intends it or not. The periorbital area receives none of that. It is the one part of the male face that gets no routine physical intervention at all — which is part of why the contrast between a smooth jawline and a crepey lower lid tends to announce itself abruptly rather than gradually.

Sleep, alcohol and salt, quantified rather than moralised

The folk explanation is that dark circles mean you are not sleeping enough, and it is partly true in a narrow and testable way. In a controlled study, ten people were photographed after a normal night and again after 31 hours of sleep deprivation following a night cut to five hours. Forty observers, who were not told which was which, rated the sleep-deprived faces as having darker circles under the eyes, more hanging eyelids, redder eyes, more swollen eyes, paler skin and more fine lines. The effect was detectable by strangers, which is the part worth taking seriously.

Note what moved. Every reliable cue on that list is vascular, fluid or pallor. Sleep loss dilates and congests periorbital vessels, promotes fluid retention in a tissue plane with nowhere to put it, and drains colour from the surrounding skin so the contrast reads harder. Alcohol and dietary sodium act on the same fluid and vasodilation axis, over hours rather than months.

What sleep does not do is deposit melanin. Constitutional pigment is inherited and stable; it was there at 19 and it will be there after a fortnight of nine-hour nights. So the test is cheap and definitive: take two proper nights and look again. If the crescent halves, you have a reversible vascular problem and your intervention is behavioural, not chemical. If it does not move at all, stop blaming your schedule and re-run the mirror test.

What topicals can realistically move, and by how much

Here the honest answer is narrower than the shelf suggests, and it splits along the same three lines.

For the vascular lane, caffeine is the standard ingredient and the evidence for this specific use is genuinely thin. The most-cited controlled work applied a counter pad containing caffeine and vitamin K in an emulsified base over 28 days and reported a modest response rate in dark-circle reduction — a real result, but a small one, and the reproducible caffeine literature sits in other indications entirely. Treat it as a mild, temporary vasoconstrictor that improves the look of a congested morning rather than a treatment. A light gel such as ↗Lumin DragonRay Eye Gel is the format that makes sense for it; a heavy occlusive cream on a puffy lid can make the fluid component look worse before anything else improves. Our ↗caffeine page lays out where the evidence actually is.

For the pigment lane, the actives are the ordinary ones and they work on the ordinary timescale — months. Vitamin C, niacinamide, azelaic acid and tranexamic acid all have a mechanism for reducing melanin production and transfer, and none of them care that the skin in question happens to be under an eye. ↗Bioderma Pigmentbio C-Concentrate is formulated squarely for this problem; the ↗vitamin C page covers stability and concentration, and the ↗hyperpigmentation pathway sets out the full escalation ladder including what a clinician can add.

For the structural lane, retinoids are the only topical with a plausible route in, because thickening the dermis is the one thing that could plausibly change how much vessel and shadow show through. It is slow, it is partial, and the eyelid tolerates it badly if you rush. Start at low frequency with a formulation built for the site, such as ↗Anua Retinol 0.1 + Caffeine Revitalizing Eye Cream, and expect to judge it at four months, not four weeks. ↗The retinol page explains the tolerance curve. If irritation is the limiting factor, a barrier-first product like ↗CeraVe Eye Repair Cream used on off nights does more than most people expect, because a compromised lid looks darker independent of everything above. Our ↗eye cream ranking is a reasonable place to start narrowing.

When the answer is a needle, a laser, or nothing at all

A shadow is not a colour, and no topical corrects a contour. If the light-angle test collapsed your darkness, the honest options are volume replacement or acceptance.

Hyaluronic acid filler in the tear trough is the volume answer and the data is better than its reputation. A meta-analysis pooling 31 reports and 2,556 participants found an overall satisfaction rate of 91.0%. The same pooling found swelling or oedema in 19.2% and bruising in 18.4% — high-frequency, self-limiting events — with contour irregularity at 5.3% and the blue-grey Tyndall effect at 0.9%. The tear trough is also among the least forgiving sites in aesthetic medicine: over-correction here reads as puffiness rather than youth, and product placed too superficially can stay visible for far longer than the marketing half-life suggests. That it is reversible with hyaluronidase is the reason it remains a reasonable thing to try, and the reason to insist on an injector who keeps it on hand.

For stubborn constitutional pigment that topicals have plateaued against, pigment-targeting lasers are the next rung, and they belong in a clinic that can assess your phototype honestly, because the same devices that clear pigment can create it.

And then there is the answer nobody sells: nothing. An orbital rim that casts a shadow is inherited architecture. A faint constitutional band that has been there since your teens is not a sign of decline, a failure of self-care, or evidence that you need a routine. The men who end up most frustrated with this region are usually the ones treating an anatomical constant as a solvable defect, and buying a new cream every quarter to prove it. Working out which of the three you actually have is the entire intervention — ↗the men's routine covers the sensible baseline around it, and the my-skin profile will keep recommendations pointed at the right lane rather than the loudest one.