Bimatoprost

Latisse

Bimatoprost (Latisse) is a healing & recovery research compound. Prostaglandin F2-alpha analog — extends the anagen phase of the lash follicle, increasing length, thickness and darkness.

How Bimatoprost works

Prostaglandin F2-alpha analog — extends the anagen phase of the lash follicle, increasing length, thickness and darkness.

Proposed benefits

Longer, thicker, darker eyelashes — it is the prescription treatment approved for inadequate eyelashes (hypotrichosis). The same molecule is prescribed as an eye drop to lower pressure inside the eye in glaucoma, which is where the lash effect was first noticed, as a side effect.

Where to get Bimatoprost

The evidence for Bimatoprost

Graded by what exists behind each claim.

✅ Clinically validated

• FDA-approved twice — as Lumigan for glaucoma and as Latisse for eyelash hypotrichosis, both on randomized data. The lash trial showed significant increases in length, thickness and darkness.
• The label carries the known risks: permanent iris darkening in susceptible eyes, periorbital fat atrophy, and eyelid skin pigmentation.

📊 Correlative data

• Wide cosmetic use for lashes and brows, and off-label scalp use for hair loss where evidence is much weaker. The iris and orbital effects are consistently reported and are the reason application technique matters.

🧪 Theoretical / extrapolated

• A prostaglandin F2α analog. It extends the anagen growth phase of the hair cycle and increases follicle size — a genuinely different mechanism from minoxidil's vascular and channel effects.
• The same prostaglandin signaling on iris melanocytes explains the pigmentation, and on periorbital adipose tissue explains the fat atrophy. Both are the mechanism, not contamination, which is why they cannot be formulated away.

How to read these tiers: they say how much human evidence exists, not how well something works — and ✗ flags harm, never a disappointing trial. How the evidence tiers work →

What Bimatoprost actually does

Bimatoprost is not a prostaglandin. It is a prostamide, and the difference is a nitrogen atom that has generated twenty years of unresolved pharmacology.

Latanoprost and travoprost are esters. Corneal esterases hydrolyze them to a free carboxylic acid, and that acid is a selective agonist at the prostaglandin F2α receptor — the FP receptor Stjernschantz 2002. Bimatoprost is an amide: the same prostanoid skeleton carrying an ethylamide instead of an acid. Amides are far harder to hydrolyze than esters, and the question that follows is the entire mechanistic argument about this drug. Does bimatoprost work as the intact amide at something other than FP, or is it a prodrug that gets cut to the acid and works at FP like everything else?

In skin, that question has actually been answered, and almost nobody quotes the answer. Bimatoprost was applied to mouse skin at 0.01%, 0.03% and 0.06% and the tissue was assayed. Peak cutaneous concentrations were dose-dependent — 3.41, 6.74 and 12.3 µg per gram of tissue — and cutaneous drug was well maintained for 24 hours after a single dose. Critically: bimatoprost was recovered from the skin only as the intact molecule, with no detectable levels of metabolites. The authors’ conclusion is one sentence long and it is the load-bearing fact on this page: bimatoprost produces hypertrichosis as the intact molecule Woodward 2013.

So in skin, whatever the receptor is, the thing arriving at it is the amide.

Now the other half, which was published twelve years later and points the other way. Human hair follicles from pre-auricular facelift skin were grown in ex vivo organ culture for 9 days. PGF2α at 100 nM — the free acid, the classical FP agonist — stimulated terminal follicle fiber growth by 4.93% (p = 0.019) and intermediate follicle growth by 10.03% (p < 0.001), prolonged anagen in both, and the effect was blocked by an FP antagonist in both follicle types. RT-PCR confirmed FP gene expression in terminal and intermediate lower bulbs, and immunohistochemistry located FP protein in the dermal papilla and connective tissue sheath Miranda 2025.

Read those two results together and the honest position emerges. The receptor that is present in the human dermal papilla, and that is demonstrably required for prostanoid-driven hair growth, is FP Miranda 2025. The molecule that actually reaches the follicle when you apply bimatoprost is the unhydrolysed amide Woodward 2013. Those are compatible — an amide can be a weak direct FP agonist — but nobody has run the FP antagonist against bimatoprost itself in human follicle organ culture. That one experiment would settle prostamide-versus-FP in skin, it uses two reagents that both already exist, and it has not been done. A third group, working on a penetration-enhanced formulation, simply asserts the answer in its own conclusion — ‘direct prostamide F2α receptor-mediated stimulation of dermal papilla cells’ Subedi 2022 — without having tested it either.

The dermal papilla is the target, and the size of the follicle predicts the response. The dermal papilla is the mesenchymal control center that decides how long anagen runs and how thick a fiber the matrix builds. FP protein sits there Miranda 2025. And the response was twice as large in intermediate follicles as in matched terminal follicles from the same person — 10.03% against 4.93%. Intermediate, vellus-like follicles are exactly what androgenetic alopecia produces. That is the strongest mechanistic argument anyone has for putting this drug on a scalp rather than only on a lash line, and it comes from a 2025 organ-culture paper rather than from any product page.

Why this topical works when the peptide topicals on this site do not. Bimatoprost has a molecular weight of about 415 Da. The 500 Dalton rule — that a molecule above 500 Da does not cross the corneal layer, argued from contact allergens, from the topical pharmacopoeia and from every marketed transdermal system Bos 2000 — puts it comfortably inside. This is a small lipophilic drug doing what small lipophilic drugs do. Hold that thought when you read the SYN-AKE and decapeptide-12 pages in this same catalog.

Cell, rodent, human — and where it stops

Step one, human tissue in a dish, and it is unusually good evidence for a cosmetic-adjacent compound. Matched terminal and intermediate follicles from human female pre-auricular facelift skin, gold-standard ex vivo organ culture, 9 days, PGF2α 100 nM with and without an FP antagonist, plus RT-PCR and immunohistochemistry for the receptor Miranda 2025. Note the limit: the agonist tested was PGF2α, not bimatoprost.

Step two, human dermal papilla cells. Bimatoprost maximally activated human dermal papilla cell proliferation at 5 µM, which the authors report as equivalent to 10 µM minoxidil Subedi 2022. That is a like-for-like comparison against the only topical with a scalp indication, in the cell type that controls the follicle, and bimatoprost was twice as potent on a molar basis.

Step three, rodents, twice, and the second one is a formulation study rather than a drug study. Bimatoprost at the clinical 0.03% once daily for 14 days increased pelage hair growth in C57BL/6 mice — earlier onset of regrowth in shaved animals and faster time to complete regrowth Woodward 2013. Separately, an optimized vehicle (ethanol, diethylene glycol monoethyl ether, cyclomethicone, butylated hydroxyanisole) produced 4.60-fold higher human skin flux and a 529% increase in dermal drug deposition against bimatoprost in ethanol; in an androgenic alopecia mouse model it increased hair-covered area at day 10 by 585% against vehicle, and at day 14 the 5% formulation gave hair weight 8.45-fold greater than 5% bimatoprost in ethanol and 1.30-fold greater than 5% minoxidil, with significantly increased follicle number and diameter in the deep subcutis Subedi 2022. Read the 8.45-fold number carefully: it is the vehicle, not the drug. Same molecule, same percentage, eight-fold difference in outcome depending on what it was dissolved in.

Step four, randomized humans — and here the result is flatter than the mechanism promises. 60 people with eyebrow hypotrichosis (Global Eyebrow Assessment grade 1 or 2), block randomized to three arms of 20: bimatoprost 0.03% gel, bimatoprost 0.01% gel, or minoxidil 2% gel, once daily. All three arms improved significantly (P ≤ 0.001). There was no statistically significant difference between the three arms — P = 0.091, 0.102 and 0.663 for the three pairwise comparisons Zaky 2023. Bimatoprost was as good as minoxidil and no better, with 0.03% nominally ahead of 0.01%.

Step five, pooled. Six randomized controlled trials of topical prostaglandin analogs against placebo for hair loss. Prostaglandin analogs significantly improved hair length and density (p < 0.001), with no significant difference in adverse events against control. The authors’ own caveat is the important one: the best dose and frequency require further study Jiang 2023.

The obstacles, named one at a time. (1) Lashes and eyebrows are not scalp. The randomized human evidence is periocular; there is no adequately powered randomized trial of bimatoprost for androgenetic alopecia of the scalp, and the review literature treats scalp use as an extension rather than an established indication Zeppieri 2024. (2) The best animal result belongs to a vehicle that is not sold. BIM-TF#5 was never taken into humans Subedi 2022; what people decant onto their scalps is an ophthalmic solution formulated for a cornea. (3) The receptor question is open for this molecule in this tissue, as above. (4) Everything reverses on stopping except the iris Stjernschantz 2002, so this is an indefinite commitment, and the periorbital risk is duration-dependent Altin Ekin 2025. (5) The pooled meta-analysis mixes sites, molecules and doses across six small trials Jiang 2023.

Bimatoprost pharmacokinetics — how much of it actually gets in

The catalog says ‘Local’. That is true and it hides the most useful arithmetic on the page — because for this compound, unusually, the numbers exist.

The applied dose, computed. Latisse is 0.03% w/v, which is 0.3 mg/mL. A single drop delivered from the applicator is roughly 30 µL, so about 9 µg of bimatoprost per eyelid per night. Spread along an upper lash line about 30 mm long and 1–2 mm wide, that is an area near 0.3–0.6 cm² and an applied dose of roughly 15–30 µg per cm². For scale: 1 mL of an ophthalmic solution decanted over a 500 cm² scalp is 300 µg total, or about 0.6 µg per cm² — roughly 30 times less drug per unit area than the lash line receives. That single ratio explains more about why lash results are reliable and scalp results are not than any argument about receptors.

What actually gets in, and how much is there. This is the rare topical where somebody measured the tissue rather than modeled it. At the clinical 0.03%, peak skin concentration was 6.74 µg per gram of tissue, and it was well maintained for 24 hours after a single application Woodward 2013. Bimatoprost’s molecular weight is about 415 g/mol, so 6.74 µg/g is approximately 16 µM in skin.

Now put that next to what the biology needs. Human follicles in organ culture responded to 100 nM Miranda 2025. Human dermal papilla cells were maximally stimulated at 5 µM Subedi 2022. A skin concentration of 16 µM is roughly 160-fold above the organ-culture concentration and about three-fold above the dermal papilla optimum. Delivery is not the limiting step for this drug at the lash line. That conclusion runs against the reflex assumption for every topical on this site, and it is the reason bimatoprost belongs in a different category from the cosmetic peptides — 415 Da clears the 500 Dalton barrier Bos 2000 with room to spare.

What degrades it — and the answer is: in skin, not much. Esterase hydrolysis is how latanoprost and travoprost are activated Stjernschantz 2002; bimatoprost is an amide, and amidase activity in skin is evidently insufficient, because no metabolites were detectable in the tissue at all Woodward 2013. Systemic elimination of whatever is absorbed goes by hepatic oxidative metabolism and conjugation, but the applied mass is measured in micrograms and systemic exposure from ocular or lash-line dosing is correspondingly tiny.

‘Local’ does not mean ‘confined to the target’, and the orbit proves it. Periorbital fat atrophy occurs in roughly half of long-term unilateral users Altin Ekin 2025, which means clinically meaningful drug reaches orbital adipose tissue several millimetres from where the drop landed. Anyone applying this to a hairline should expect the same lateral spread into the skin around it.

The oral barrier and the injection route, for completeness. There is no oral bimatoprost and no dermatologic injection of it; swallowed, a prostanoid amide faces gut-wall and hepatic first-pass metabolism and would be pharmacologically pointless at these masses. Unlike almost everything else in this catalog, there is no injectable fallback when penetration is the question — which is precisely why the vehicle work matters so much Subedi 2022.

What would have to be true, and how you would know it was not

This is a topical drug with no blood read-out, so the falsification is structured the only honest way it can be: what to watch, how long before it means anything, and what will fool you. Four predictions, and the second and third argue against the product.

1. On a scalp, the response should be biggest in the miniaturized areas — and if it is not, the dermal papilla story is wrong. Intermediate follicles outgrew matched terminal follicles two to one in human organ culture, 10.03% against 4.93% Miranda 2025. What to watch: photograph the frontal hairline and a fully terminal occipital patch under identical lighting at a fixed distance, monthly. The prediction is conversion of vellus hairs at the receding margin outpacing any change in the dense area. Diffuse thickening everywhere with nothing at the margin would be the falsifier — it would mean the effect is on fiber diameter in follicles that were already working, which is a different and less useful drug.

2. It should be no better than minoxidil, and expecting more is the error. The one randomized head-to-head in humans found bimatoprost 0.03% and 0.01% statistically indistinguishable from minoxidil 2%, with all three improving (P ≤ 0.001) and none separating (P = 0.091, 0.102, 0.663) Zaky 2023. How long before it means anything: the lash and brow trials run 16 weeks or longer, and the follicle cycle sets the floor — nothing observed before 8 weeks is a drug effect, and a judgement made before 16 is premature. This is the prediction that cuts against the compound: it is a second mechanism, not a stronger one.

3. Iris color must be watched in hazel and mixed-color eyes, and the change is not reversible. Predisposition sits with hazel or heterochromic irides; the pigmentation is likely permanent or very slowly reversible; and the clear-cut disadvantage is heterochromia between the eyes in unilaterally treated people Stjernschantz 2002. What to watch: a close, in-focus daylight photograph of both irides at baseline and every three months, same light, no flash. This is the one adverse effect on the page that cannot be undone by stopping.

4. The upper eyelid sulcus should deepen with duration, and it should partially recover if you stop. Among 86 unilateral users, 67.4% had at least one periorbital change; orbital fat atrophy was the commonest at 48.8%, then deepening of the upper eyelid sulcus 38.4%, involution of dermatochalasis 32.6% and enophthalmos 26.7%; independent risk factors were age over 60, bimatoprost use, travoprost use, and therapy longer than one year, all p < 0.05 Altin Ekin 2025. In nine people who stopped after at least a year, the changes had visibly improved one year later Abalo-Lojo 2023. What will fool you: almost everything. Photographs at different angles, different lighting, different expressions, and normal ageing all move eyelid appearance. The reason the periorbitopathy literature is credible is that it is built on unilateral users Altin Ekin 2025 Abalo-Lojo 2023, where each person is their own control — and that is exactly the design a lash-line user cannot replicate if they treat both eyes.

What nobody has tested yet

Four experiments. The first settles a twenty-year-old pharmacology argument and could be run in a month.

1. Nobody has run an FP antagonist against bimatoprost itself in human hair follicle organ culture. The model exists, the antagonist exists and both were used — but with PGF2α, not with bimatoprost Miranda 2025. Meanwhile the drug is known to reach skin as the unhydrolysed amide Woodward 2013. If antagonist blocks bimatoprost, the prostamide is simply a weak FP agonist and the whole prostamide-receptor argument dissolves; if it does not, there is a second receptor in the dermal papilla nobody has characterized. Either result is publishable and neither has been produced.

2. Nobody has imaged the orbits of cosmetic lash users. Every periorbitopathy dataset comes from glaucoma patients instilling drops into the eye, mostly unilaterally, for years Altin Ekin 2025 Abalo-Lojo 2023. Cosmetic users apply a smaller dose to a different site, usually bilaterally, often for years as well. Nobody has measured whether the 48.8% orbital fat atrophy figure applies to them at all, in either direction. A cross-sectional imaging study of long-term Latisse users against matched non-users is straightforward, obviously interesting, and absent.

3. Nobody has taken a penetration-optimized bimatoprost into a human scalp trial. The vehicle that produced an 8.45-fold improvement over ethanol in mice, and beat 5% minoxidil, stopped at the mouse Subedi 2022. A randomized 24-week trial of that formulation against 5% minoxidil in androgenetic alopecia is the obvious next step and has not been run — which means the strongest preclinical hair result this molecule has ever produced has never been tested in a person.

4. Nobody has tested whether follicle size predicts who responds. The organ-culture data says intermediate follicles respond twice as strongly as terminal ones Miranda 2025. That predicts that early, patchy miniaturization responds and late, fully-shed scalp does not — a responder rule that could be tested with trichoscopy at baseline and has never been examined in any prostaglandin hair trial, including the six in the meta-analysis Jiang 2023.

Bimatoprost — its own safety story, not its class's

This page carries a shared healing-and-recovery safety block, and the compound card even describes bimatoprost’s benefits as ‘soft-tissue and gut repair… angiogenesis’. None of that is this drug. Bimatoprost is a prostanoid with three specific, measured, well-documented adverse effects, and they are all consequences of the mechanism rather than contamination — which is why no formulation change removes them.

1. Iris pigmentation, and the mechanism is transcriptional rather than proliferative. Latanoprost, unoprostone, travoprost and bimatoprost all increase iris pigmentation in some patients; predisposition sits with hazel or heterochromic eyes; and because latanoprost and travoprost are selective FP agonists, the phenomenon is very likely FP-mediated. Several studies show melanogenesis stimulated in iridial melanocytes with upregulated transcription of the tyrosinase gene; histopathologic work found no evidence of harmful consequences; the one clear-cut disadvantage is potential permanent heterochromia in unilaterally treated people Stjernschantz 2002. The cell work separates the two possible mechanisms cleanly: latanoprost and PGF2α greatly increased tyrosinase activity in murine melanoma lines and produced small increases in human uveal and cutaneous lines, while failing to raise the mitotic index of any human line Dutkiewicz 2000. So the iris darkens because existing melanocytes make more pigment, not because there are more of them — which is reassuring about melanoma risk and is exactly why the color change does not reverse.

2. Periorbital fat atrophy, and it is common rather than rare. In 86 unilaterally treated patients, 67.4% showed at least one periorbital change; orbital fat atrophy 48.8%, deepening of the upper eyelid sulcus 38.4%, involution of dermatochalasis 32.6%, enophthalmos 26.7%. On multivariate analysis, age over 60, bimatoprost use, travoprost use and therapy longer than one year were each independent risk factors for higher grades, all p < 0.05 Altin Ekin 2025. The cell biology behind it is specific: PGF2α agonists caused concentration-dependent down-sizing of three-dimensional adipocyte organoids made from human orbital fibroblasts, increased their physical solidity, and changed expression of collagen 1, TIMP2, MMP2 and MMP9 in a pattern inversely correlated with organoid size Itoh 2021. This is the drug remodeling fat and matrix through the receptor it is meant to use.

3. But it is partly reversible, which the warnings usually omit. Nine patients treated for at least a year stopped for cosmetic reasons; one year later, the deepened sulcus and reduced eyelid fat pad had visibly improved Abalo-Lojo 2023. Partial, slow, and real. Iris color is the exception that does not come back Stjernschantz 2002.

4. Eyelid skin hyperpigmentation and hypertrichosis where you did not want it. The same melanogenic mechanism that darkens the iris darkens periocular skin, and the same follicular mechanism that grows lashes grows hair wherever the solution runs. Both are reversible on stopping. The practical control is application technique: this is applied along the lash line, never into the eye, and drug that runs onto the lower lid or the cheek is drug doing exactly what it does at the lash line, in the wrong place.

5. What the record is unusually strong on. Unlike most of this catalog, bimatoprost has randomized human trials Zaky 2023, a meta-analysis Jiang 2023, decades of ophthalmic post-marketing surveillance and a mechanistic literature for its main adverse effects Stjernschantz 2002 Itoh 2021. The pooled hair-trial data found no significant difference in adverse events against control Jiang 2023. The risks on this page are known, quantified and manageable — which is a genuinely different situation from the unmeasured compounds elsewhere on this site, and it should be said as clearly as the warnings.

Sources read for this page

• Woodward DF, et al. The biodisposition and hypertrichotic effects of bimatoprost in mouse skin. Experimental Dermatology 2013 · PMID 23278986
• Miranda BH, et al. Prostaglandin F2alpha stimulates the growth of human intermediate hair follicles in ex vivo organ culture with potential clinical relevance. Frontiers in Physiology 2025 · PMID 40606226
• Subedi L, et al. Preparation of topical bimatoprost with enhanced skin infiltration and in vivo hair regrowth efficacy in androgenic alopecia. Drug Delivery 2022 · PMID 35040730
• Zaky MS, et al. Comparative study of the efficacy and safety of topical minoxidil 2% versus topical bimatoprost 0.01% versus topical bimatoprost 0.03% in treatment of eyebrow hypotrichosis: a randomized controlled trial. Archives of Dermatological Research 2023 · PMID 37517060
• Jiang S, et al. The efficacy of topical prostaglandin analogs for hair loss: A systematic review and meta-analysis. Frontiers in Medicine 2023 · PMID 36999072
• Stjernschantz JW, et al. Mechanism and clinical significance of prostaglandin-induced iris pigmentation. Survey of Ophthalmology 2002 · PMID 12204714
• Dutkiewicz R, et al. Effects of latanoprost on tyrosinase activity and mitotic index of cultured melanoma lines. Experimental Eye Research 2000 · PMID 10870514
• Altin Ekin M, et al. Predictors of severity in prostaglandin-associated periorbitopathy. Cutaneous and Ocular Toxicology 2025 · PMID 40055901
• Abalo-Lojo JM, et al. Improvement of Prostaglandin-Associated Periorbitopathy after Discontinuing Treatment. Turkish Journal of Ophthalmology 2023 · PMID 36847619
• Itoh K, et al. Prostaglandin F2alpha agonists induced enhancement in collagen1 expression is involved in the pathogenesis of the deepening of upper eyelid sulcus. Scientific Reports 2021 · PMID 33903711
• Bos JD, Meinardi MM. The 500 Dalton rule for the skin penetration of chemical compounds and drugs. Experimental Dermatology 2000 · PMID 10839713
• Zeppieri M, et al. From Eye Care to Hair Growth: Bimatoprost. Pharmaceuticals 2024 · PMID 38794131

Bimatoprost — safety, predicted from mechanism

Bimatoprost — interference & stacking

Bloodwork to run alongside Bimatoprost

Baseline first, then again at 8–12 weeks.

The Inflammation Deep Dive panel covers these in one order — 10 markers, $248.35 with the discount applied.

Check results you already have → · All 103 markers A–Z

Bimatoprost — frequently asked questions

What Bimatoprost is used for

Bimatoprost appears under 1 goal in the goal router.

Related Healing & Recovery compounds

← Explore the full Protocol Vault

↑ Back to on this page