Bimatoprost (Latisse) is FDA-approved to grow eyelashes, and latanoprost grows them too. But do these prostaglandin analogs regrow scalp hair? A full look at every preclinical and clinical study for both drugs, and the wider prostaglandin class.
There is a drug that grows hair so reliably the FDA approved it. It is bimatoprost, sold as Latisse, and it grows eyelashes. Latanoprost, its cousin, does the same. So the obvious question has been asked for twenty years: if these prostaglandin drops grow eyelashes, will they regrow the hair on your head?
The honest answer is that we still do not know, and the reason is more interesting than either the hype or the horror stories.
The short answer: Bimatoprost and latanoprost are prostaglandin analogs. For eyelashes the evidence is strong and bimatoprost is the winner: it is FDA-approved and in its pivotal trial 78 percent of users grew fuller lashes versus 18 percent on vehicle. For the scalp, the picture is more nuanced than "it works" or "it fails," and neither drug has a large trial behind it. Latanoprost, the less famous one, has one small controlled study that was cleanly positive, plus preclinical data roughly comparable to minoxidil, which is why it is the one still used off-label. Bimatoprost, the famous eyelash drug, actually did grow hair in men: in its maker's Phase 2 program it beat the placebo vehicle by about 9 hairs per cm² across hundreds of men. But it underperformed generic minoxidil in the same trials, did nothing in women, and was abandoned without an efficacy paper, probably as much for commercial reasons as anything. Even a drug built to attack the pathway from the opposite side, setipiprant, failed a baldness trial. The pathway is real biology with real but modest human signals that has never produced a decisive scalp trial.
Prostaglandins are signaling lipids your body makes everywhere, including in skin. In the hair follicle they do not all pull the same way. Two of them, PGE2 and PGF2-alpha, stimulate hair growth, while a third, PGD2, inhibits it (Nieves and Garza 2014).
That last one matters for baldness. In 2012 a Johns Hopkins group found that PGD2 is elevated roughly tenfold in bald scalp compared with haired scalp in the same men, about 16.3 versus 1.5 nanograms per gram of tissue, and that PGD2 actively suppresses hair growth through a receptor called GPR44 (Garza 2012). So pattern baldness is, in part, a prostaglandin balance problem: too much of the brake (PGD2), and drugs that push the accelerator (PGF2-alpha analogs) became an obvious idea.
The growth-promoting signal runs through the FP receptor, the prostaglandin-F receptor, a G-protein-coupled receptor. What makes it relevant to hair is where it sits: in the dermal papilla and the outer root sheath, the control hub at the base of the follicle, and not in the surface skin cells (Colombe 2007; Khidhir 2012). A drug that reaches the papilla and switches on FP is talking directly to the cells that decide whether a follicle grows.
When FP is activated, the follicle does two things that matter: it prolongs the anagen, or growth, phase, and it nudges resting follicles in telogen back into growth. In mice, bimatoprost pulled that switch forward from day 38 to day 17; in cultured human follicles it grew the fibers about 35 percent longer, and an FP-blocking drug abolished the effect, which proves it is receptor-mediated and not just irritation (Khidhir 2012; Sasaki 2005). That is exactly why these drugs grow eyelashes so well: a longer growth phase means longer, thicker, darker, and more numerous lashes.
They aim at the same target and produce the same downstream effect. The difference is the chemistry that gets them there.
Latanoprost, along with its cousins travoprost and tafluprost, is a prodrug. It is applied as an inactive ester, and enzymes in the skin (esterases) have to cut it into latanoprost acid before anything happens. Only that acid is the active FP-receptor agonist, so latanoprost needs one activation step to work.
Bimatoprost is a prostamide, an amide rather than an ester. It acts directly on the FP and prostamide receptors without being converted first. Pharmacologists have argued for years about whether bimatoprost hits a distinct "prostamide receptor" or is simply an FP agonist, but functionally both routes converge on the same growth pathway. Being direct-acting, with no activation step, is the usual explanation for why bimatoprost is the more potent eyelash grower.
| Latanoprost | Bimatoprost | |
|---|---|---|
| Chemical class | PGF2-alpha analog (isopropyl ester) | Prostamide (amide) |
| Needs activation first? | Yes, esterases convert it to latanoprost acid | No, acts directly |
| Receptor | FP receptor (agonist) | FP / prostamide receptor |
| Where it acts | Dermal papilla and outer root sheath | Dermal papilla and connective-tissue sheath |
| Downstream effect on the follicle | Prolongs anagen; telogen to anagen | Prolongs anagen; telogen to anagen |
| Relative potency for eyelashes | Effective, but the weaker of the two | More potent; the FDA-approved lash drug |
Mechanistically, then, the two are the same drug wearing different clothes. Latanoprost is a prodrug that becomes an FP agonist; bimatoprost is a direct-acting prostamide. Both end up prolonging the follicle's growth phase from the dermal papilla. The gap between them is one of chemistry and potency, not a fundamentally different mechanism, which is worth keeping in mind: the two behave alike, so the fact that they diverge so sharply in the clinic (approved for lashes, abandoned or barely tested for scalp) is about delivery and dose, not about one having a secret mechanism the other lacks.
Here is a nuance that is easy to get wrong: because these drugs are receptor agonists, they do not really change how much prostaglandin your scalp makes. They supply an outside molecule that switches the FP receptor on directly, mimicking the growth-promoting PGF2-alpha signal. That means they add "accelerator," but they do nothing to the elevated PGD2 "brake" sitting in bald scalp. They work around the prostaglandin imbalance rather than correcting it, which is a different strategy from raising the whole prostaglandin pool (what a 15-PGDH inhibitor would do) or removing the brake (what a PGD2 blocker like setipiprant tried to do). It also means the whole effect hinges on getting enough agonist to the receptor, which is precisely the problem on the scalp.
Barely, and briefly. Both are built to act locally and clear fast. After topical application, latanoprost is absorbed as its acid, peaks in the blood within about 5 minutes, and is eliminated with a plasma half-life of roughly 17 minutes (Sjoquist 2002). Bimatoprost clears a bit more slowly, with a plasma half-life of about 45 minutes, but still under an hour. Whatever reaches the bloodstream is gone quickly, so systemic levels stay very low, and the well-known side effects, iris darkening and periorbital fat atrophy, are local to where the drug is applied, not body-wide. The one scalp-specific worry is surface area: a scalp is far larger than an eyelid margin, so more could be absorbed. That is exactly why the bimatoprost scalp program ran dedicated "local pharmacokinetics and tolerability" trials to measure how much stayed put, and the drugs' rapid clearance is the reason systemic exposure is expected to stay low even across a whole scalp.
This is where prostaglandin analogs look their best. Here is essentially the entire preclinical record for hair, for both drugs and the class.
| Study | Drug | Model | What it found | Evidence tier |
|---|---|---|---|---|
| Uno 2002 | Latanoprost | Bald stumptail macaque (the primate model of pattern baldness) | 0.05% latanoprost induced "moderate to marked" regrowth, converting vellus to terminal hairs; vehicle did nothing. Roughly matched 5% minoxidil in the same lab. | Animal, primate |
| Sasaki 2005 | PGF2-alpha, latanoprost, unoprostone | Mouse follicles | All stimulated hair and pushed follicles from resting (telogen) into growth (anagen). | Animal, mouse |
| Khidhir 2012 | Bimatoprost | Human scalp hair follicles in organ culture, plus mice | Grew human follicles about 35 percent longer; in mice it pulled the switch into growth phase forward from day 38 to day 17. Blocking the receptor abolished the effect. | Ex vivo human tissue + mouse |
| Colombe 2007 | (mechanism) | Human follicle tissue | The FP receptor sits in the dermal papilla and outer root sheath, explaining how PGF2-alpha could act on hair. | Ex vivo human |
| Miranda / Tobin 2025 | PGF2-alpha | Human scalp follicles ex vivo (independent lab) | Confirmed PGF2-alpha grows human follicles, but the effect was small, about 5 to 10 percent. | Ex vivo human |
Two honest caveats. First, there has never been a head-to-head test of bimatoprost versus latanoprost at the actual hair follicle, so the common claim that "bimatoprost is more potent" is extrapolated from eye pharmacology, not measured on hair. Second, the best-looking bimatoprost study (Khidhir 2012) was authored by Allergan, the company that owns the drug. Impressive preclinical data from a single interested lab is a reason for interest, not proof.
If you want to see prostaglandin analogs work in humans, look at eyelashes, not scalp.
| Study | Drug | Design | Result |
|---|---|---|---|
| Smith 2012 (pivotal, FDA registration) | Bimatoprost 0.03% (Latisse) | Phase 3 RCT, n=278, 16 weeks | 78.1% grew at least one grade of fuller lashes vs 18.4% on vehicle (P<.0001); length, thickness and darkness all improved; number-needed-to-treat 1.67 |
| Wirta 2015 | Bimatoprost 0.03% | Pooled safety, 6 RCTs | Confirmed the safety profile behind the 2008 FDA approval for eyelash hypotrichosis |
| Coronel-Perez 2010 | Latanoprost | Open-label, eyelash alopecia in alopecia areata, n=44 | "Acceptable" regrowth in 45 percent; not randomized, not blinded, not FDA-approved for this |
Head to head for eyelashes, bimatoprost wins. It is the one with FDA approval, the one with a 78 percent registration-grade trial, and pharmacologically it does not need to be activated the way latanoprost (a prodrug) does. A small direct comparison trial was registered but never published its numbers, so the exact margin is unproven, but the weight of evidence clearly favors bimatoprost for lashes.
A safety note that matters for anyone applying these near the eyes or hairline. Long-term use around the eye can cause prostaglandin-associated periorbitopathy, a sinking or hollowing of the eye socket from fat loss, and it is worst with bimatoprost (seen in about 93 percent of long-term users in one comparison versus 41 percent for latanoprost). Latanoprost can permanently darken a light-colored iris if it reaches the eye surface. Both can darken eyelid skin. These are eyelid-margin and eye-surface risks; they are part of why "just put the eyelash drug on your scalp" is not as simple as it sounds.
For all the talk, the entire controlled human scalp evidence for latanoprost is a single pilot.
Blume-Peytavi 2012 enrolled 16 men with mild pattern baldness and applied latanoprost 0.1% and placebo daily to two small "minizones" on the scalp for 24 weeks. At the end, hair density was higher on the latanoprost side than on placebo (P=.0004).
That is a genuine positive signal, and it is why latanoprost keeps getting mentioned. But read the fine print the authors themselves put in: it was explicitly a "pilot," it was only 16 young men with mild loss, only 8 of the 16 were rated a good clinical response, the growth-to-rest ratio actually fell slightly on the drug side, and it was measured on tiny patches, not a whole balding scalp. It has never been replicated in a larger trial. One 16-man pilot on postage-stamp zones is not proof that latanoprost regrows a balding scalp.
This is the part almost no one mentions, and it is the most telling.
Allergan, the company that makes Latisse and therefore had every commercial reason to want bimatoprost to work on scalps, ran a full Phase 2 program in pattern baldness. Not a pilot: more than 1,100 men and women across multiple trials.
| Trial (ClinicalTrials.gov) | Population | Size | Status |
|---|---|---|---|
| NCT01325337 | Men, pattern baldness | 307 | Completed |
| NCT01904721 | Men, pattern baldness | 244 | Completed, results posted |
| NCT01325350 | Women, female pattern hair loss | 306 | Completed |
| NCT02170662 | Mechanism study | 33 | Completed |
| NCT02676310 | Dose-escalation follow-on | 53 | Terminated (2017) |
The posted results tell the story, and it is worse than "inconclusive." The primary endpoint in these trials was the change in target-area hair count at 6 months.
In men, bimatoprost added about 12.7 hairs where the placebo vehicle already added 5.8 (NCT01904721). A small edge. But in the other men's trial (NCT01325337), the minoxidil 5% comparator running in the very same study added 21.9 hairs, roughly double bimatoprost's best arm. In women (NCT01325350), bimatoprost did essentially nothing: its arms ranged from minus 3.5 to plus 4.3 hairs versus plus 1.1 for vehicle, doing worse than placebo in one arm, while minoxidil 2% added 13.6. So the trials could clearly detect a hair drug. The minoxidil arms worked, and so, more modestly, did bimatoprost in men, beating the placebo vehicle by about 9 hairs per cm². It just could not match minoxidil, and it did nothing in women.
Then the trail goes cold. Allergan published no dedicated efficacy paper for any of these scalp trials. The design survives mostly in the company's own patents. The follow-on dose-escalation study was terminated in 2017. Bimatoprost never advanced to Phase 3 for the scalp and was never approved for it.
The company that proved bimatoprost grows eyelashes tested it on more than a thousand scalps, saw it barely beat placebo, and walked away. That is the single most useful fact in this entire comparison.
Allergan never publicly stated why it stopped, so part of this is inference. But the pieces fit, and it is worth separating what is documented from what is reasoned.
The first factor is efficacy, and it is more subtle than "it failed." In Allergan's own Phase 2 trials, bimatoprost did beat placebo in men, adding about 13 hairs per cm² versus 4 for vehicle, a real effect across hundreds of men. The problem was the company running the comparison: minoxidil in the same study added 22, and in women bimatoprost did essentially nothing while minoxidil worked. So the issue was not that bimatoprost had no effect, it was that it could not beat a cheap generic you cannot out-market. For an expensive new prescription product, that is often reason enough to walk.
The commercial context explains why even a modest result was not worth chasing, and here an old patent matters. As early as February 1997, a patent by Murray Johnstone already covered using prostaglandins, latanoprost included, to grow hair, and it explicitly named the scalp. That meant Allergan could never own the broad idea of "a prostaglandin for scalp hair." It could only patent specific formulations, which it did, and formulation patents are narrow and comparatively easy to work around. So the commercial ceiling was low: even a Phase 3 win would have faced fast generic and compounded competition, because no one could lock up the space.
Two more things were closing that window at the wrong moment. The Johnstone patent itself was set to expire around 2019, which would open the scalp-prostaglandin field to everyone. And Allergan was heading into its acquisition by AbbVie, announced in 2019, exactly the kind of corporate transition where non-core, weakly differentiated dermatology assets get shelved. It is the same dynamic that has stranded other promising-but-unownable hair molecules.
Put the documented and the inferred together and the decision looks obvious. A drug has to be clearly excellent to justify a nine-figure Phase 3 on a molecule you cannot own, against a market about to open to generics, inside a company being acquired. Bimatoprost on the scalp was not excellent. It lost to minoxidil. So it was dropped, and, tellingly, no one has picked the idea back up since. The honest version is that the weak efficacy is the documented reason; the 1997 prior art, the expiring exclusivity, and the acquisition are why even a mediocre result was not worth a fight.
Two papers get quoted constantly as proof that prostaglandins do not work for hair. It is worth knowing what they actually tested, because it is not scalp baldness.
Alopecia areata is an autoimmune disease in which the immune system attacks the follicle. It is a completely different problem from pattern baldness, and it is in eyebrow and eyelash skin, not scalp. These studies show these drugs failing in autoimmune eyelash loss. They do not show that prostaglandins fail on a balding scalp. Fair is fair, in both directions.
To be complete, here is every other prostaglandin-pathway approach that has been tried for hair.
So on both sides of the pathway, accelerator and brake, the clinic has said the same thing: not yet.
The most likely explanation is boring and mechanical, not mysterious. Eyelashes sit in the thinnest skin on the body, and the drug is dropped essentially right onto the follicle. The scalp is different: thick skin, a large surface area, and follicles that are actively being miniaturized by DHT. Getting enough of a prostaglandin analog down to the dermal papilla, across a whole balding scalp, at a concentration that is both effective and tolerable, is an unsolved delivery problem. A pro-growth nudge may simply not be strong enough to overcome active androgen-driven miniaturization the way it can grow a healthy eyelash longer.
So the two are closer than the eyelash story suggests. Bimatoprost has the larger men's dataset and did beat placebo, it just could not match minoxidil and was dropped without a paper. Latanoprost has a smaller but cleanly positive result and is the one still used off-label today. Either way, both showed real, modest human signals, and the honest bottom line is that no prostaglandin drug yet has a large, decisive scalp trial behind it.
Here is the entire human and preclinical record for both drugs in one place, sorted by drug, with the main endpoint and the actual result versus placebo. Each row is explained in the sections above.
| Drug | Study or trial | Design and size | Primary endpoint | Result vs placebo / vehicle |
|---|---|---|---|---|
| Latanoprost | Uno 2002 | Preclinical, bald macaque scalp | Hair regrowth (vellus to terminal) | 0.05%: 5-10% conversion to terminal hairs; vehicle none. Roughly matched 5% minoxidil. |
| Latanoprost | Blume-Peytavi 2012 | 16 men, 24 wk RCT, small "minizones" | Hair density (TrichoScan) at 24 wk | Significantly higher than the placebo site (P=.0004); but the anagen ratio fell (1.98 to 1.57), only 8/16 responded, and the size of the gain is shown only in a figure. Never repeated. |
| Latanoprost | Coronel-Perez 2010 | 44 patients, open-label (AA eyelash) | Eyelash regrowth (categorical) | ~45% "acceptable" regrowth. Uncontrolled, no placebo comparison. |
| Latanoprost | Ross 2005 | 11 patients, 12 wk (AA eyebrow) | Eyebrow regrowth | No significant regrowth (P=1.0). Negative; different disease. |
| Bimatoprost | Khidhir 2012 | Preclinical, human follicles + mice | Hair-fiber length | +35% vs control (ex vivo). |
| Bimatoprost | Smith 2012 (NCT00693420, FDA pivotal) | 278, Phase 3 RCT, 16 wk | >=1-grade eyelash prominence (GEA) at wk 16 | 78.1% vs 18.4% vehicle (P<.0001). Basis for Latisse approval (2008). |
| Bimatoprost | Additional eyelash RCTs (pooled in Wirta 2015) | 6 RCTs, pooled | Safety (not efficacy) | Safety record behind the approval. |
| Bimatoprost | NCT01325337 | 307 men, Phase 2, 6 mo | Change in target-area hair count, month 6 | Bimatoprost +13.1 vs vehicle +4.1; minoxidil 5% in the same trial +21.9. |
| Bimatoprost | NCT01904721 | 244 men, Phase 2, 6 mo | Change in target-area hair count, month 6 | Bimatoprost +12.7 / +9.3 vs vehicle +5.8. |
| Bimatoprost | NCT01325350 | 306 women, Phase 2, 6 mo | Change in target-area hair count, month 6 | Bimatoprost −3.5 to +4.3 vs vehicle +1.1; minoxidil 2% +13.6. Bimatoprost essentially failed. |
| Bimatoprost | NCT02170662 | 33, Phase 2 | Mechanism / biomarkers | No efficacy endpoint. |
| Bimatoprost | NCT01189279 | Phase 1 | Safety and PK of a new formulation | No efficacy endpoint. |
| Bimatoprost | NCT02848300 | 11, Phase 1 | Local PK and tolerability (scalp) | No efficacy endpoint. |
| Bimatoprost | NCT02676310 | 53, Phase 1 | Dose-escalation safety | Terminated (2017); no results. |
| Both (head to head) | NCT00187577 | 14, RCT (AA eyelash), lata vs bima | Eyelash prominence | Completed; results never published. |
| Both | Roseborough 2009 | 11, 16 wk (AA eyelash) | Eyelash regrowth | No regrowth for either drug. Negative; different disease. |
That is every registered scalp trial and every eyelash trial we could find for both drugs, with the endpoints. Read the bimatoprost scalp rows together and the picture is stark: in men, bimatoprost barely edged the placebo vehicle and was roughly half as good as the minoxidil arm running in the same study; in women, it did nothing, while minoxidil worked. The trials could clearly detect a real hair drug. Bimatoprost just was not one on the scalp.
Neither is proven to regrow scalp hair by a large trial. Latanoprost has one small controlled study (16 men) that beat placebo on hair density. Bimatoprost beat placebo in men in a Phase 2 program, by about 9 more hairs per cm², but it underperformed minoxidil, did nothing in women, and was abandoned. Both show real but modest signals, not definitive proof.
No. Bimatoprost is FDA-approved only for eyelash growth, as Latisse, not for scalp hair. Its maker ran a Phase 2 scalp program in more than 1,100 people, never published an efficacy paper, and never advanced it to Phase 3. Using it on the scalp is off-label and unproven.
For eyelashes, bimatoprost is clearly better and is the only one FDA-approved. For the scalp, latanoprost has the cleaner positive result, a small controlled study, and is the one still used off-label, while bimatoprost's larger program was abandoned after it lost to minoxidil. Neither is definitively proven for scalp hair.
Its Phase 2 showed only a modest effect in men and nothing in women, and it lost to generic minoxidil in the same trials. On top of that, using prostaglandins for hair had been prior art since a 1997 patent, so there was little exclusivity left to protect a costly Phase 3. The reason to walk was likely commercial as much as clinical.
The known side effects, iris darkening and periorbital fat atrophy, are local to where the drug is applied near the eye, not body-wide, and these drugs clear from the blood within about an hour, so systemic exposure is low. They are prescription drugs and scalp use is off-label, so use them only under a clinician's care.
This article is educational and is not medical advice. Bimatoprost and latanoprost are prescription drugs. Bimatoprost is FDA-approved only for eyelash growth; using either drug on the scalp for hair loss is off-label and not proven, and should only be considered with a qualified clinician who can weigh the risks, including the eye-related side effects described above. Nothing here is a recommendation to use any of these compounds.