Wellness, Actually · November 19, 2026
Do hair loss treatments actually work?
Short answer
Some do, modestly. Finasteride cuts scalp DHT by about two-thirds and added roughly 107 hairs to a baseline of 876 in a one-inch circle at one year, about 138 by year two, and most of that is loss prevented rather than hair regrown. Minoxidil works too, with about a 5% hair count improvement over placebo in a 381-woman trial. Rosemary oil, onion juice, and platelet-rich plasma have little or no credible evidence behind them, and nothing on the market turns a bald scalp into a full head of hair.
Hair loss is one of the few medical conditions where the patient notices it years before anyone else does. I know this because I have a childhood scar on my scalp that used to be under my hairline and now isn't. That scar is a sundial, and it keeps good time.
So the question is whether any of the things sold to people in my position actually work. Some do. Not as well as you'd like.
Why hair falls out, and why it falls out in a pattern
Hair grows in cycles. Anagen is the active growth phase, and about 90% of the hairs on your head are in it at any moment, producing keratin and pushing outward. That phase lasts somewhere between two and seven years, and how long is genetically set. Then comes catagen, a brief phase where the follicle involutes, then telogen, the shedding phase. A quiescent stretch, and the cycle restarts. The follicle doesn't die. It thins.
The follicle carries an androgen receptor, and that receptor is exquisitely sensitive to dihydrotestosterone, a metabolite of testosterone that you are going to hear about repeatedly. When DHT hits the follicle, the anagen phase shortens. Fewer hairs are actively growing at any given time, more are falling out, and the balance tips.
Pattern baldness happens because receptor sensitivity varies across the scalp. The crown, the top, the temples: very sensitive. The back and sides: not very. That map is the pattern. And because the androgen receptor gene sits on the X chromosome, which men get from their mothers, the relevant family photograph is of your maternal grandfather.
Women lose less hair to this mechanism for a simple reason. Less testosterone means less DHT.
The DHT story was pieced together over a very long time. Aristotle and Hippocrates both noticed that eunuchs don't go bald, wrote it down, and had no idea what to do with it. The modern version arrived in the 1970s with the Dominican children called guevedoces, XY individuals born with ambiguous genitalia and raised as girls who developed male characteristics at puberty. They had an inherited deficiency of 5-alpha-reductase, the enzyme that converts testosterone to DHT. DHT drives the development of external genitalia in utero; testosterone drives the muscle and the deep voice at puberty. Hence the apparent transformation at 12, which is what the name means. And they did not go bald. Roy Vagelos, who worked on that enzyme, would go on to help build a drug that blocks it.
Finasteride and minoxidil: real, small, and measured in hairs per square inch
That drug is finasteride. At 5 mg it treats the prostate; at 1 mg a day it's the hair loss dose. It cuts scalp DHT by about two-thirds.
Here's what that buys you, and the measurement matters. Trials take a one-inch circle of scalp and count the hairs in it. You have roughly 100,000 follicles on your head, so nobody is counting all of them. Baseline in the pivotal trial was 876 hairs in that circle. Finasteride added 107 over baseline at one year and about 138 at two. Mostly that's loss prevented, not regrowth, which is why men start taking it in their twenties at the first signs.
Emily's objection, which I think is fair even though I don't share her conclusion: who besides you is noticing 107 hairs out of 876? These trials use hair counts, not overall aesthetic judgment, and nobody has tested whether a stranger can tell the difference between 876 and 983 hairs in a circle. I'd like to see that study. It doesn't exist.
Finasteride is not appropriate for women who might become pregnant. Follow the guevedoces logic: suppress DHT in a pregnant woman carrying an XY fetus, and that fetus doesn't develop external genitalia. That's the reason for the warnings about handling broken pills.
Minoxidil is the other big one, also a repurposed drug, in this case a blood pressure medication whose side effect was hair growth. The mechanism isn't fully worked out. It appears to prolong anagen somewhat, and it shifts follicles out of telogen back into anagen fairly abruptly, which is why people often shed at first before anything regrows.
As a nephrologist I have a complicated relationship with this drug. Ask a physician who isn't a nephrologist whether they've ever prescribed oral minoxidil and the answer is almost certainly no. It's our bazooka. When a patient arrives on five blood pressure medications with heart disease and a pressure that won't come down, minoxidil is what we reach for. The reason it's last line is the side effect profile: substantial fluid retention, leg swelling, and pericardial effusion. Our dose is 10 to 40 mg. People taking it for hair are using 2.5 mg or rubbing it on their scalp, which is a different universe, but it still startles me a little.
I was taught that oral minoxidil outperforms topical for hair regrowth. I can't find data supporting that. A head-to-head randomized trial found daily oral 5 mg was not superior to topical 5% at 24 weeks. Oral is easier to take. It also grows hair in places you didn't ask for, and it nudges your blood pressure down, which you may not want if your blood pressure is fine.
The sexual side effects are real, smaller than the internet says, and heavily nocebo
Start with the easy question: while you are taking finasteride, does it cause sexual side effects? A systematic review found sexual adverse events in roughly 2% to 7% of treated men versus 1% to 4% on placebo. Most were mild and reversible on stopping. There is something there. Finasteride doesn't lower testosterone, but DHT is a more potent stimulator of the receptor than testosterone is, so it's biologically reasonable.
Then there's the 2007 study that should be required reading. Men on the 5 mg prostate dose were split into two groups. Everyone got the drug. One group was told it "may cause erectile dysfunction, decreased libido, problems of ejaculation, but these are uncommon." The other was told nothing. In the informed group, 43% reported a sexual side effect, versus 15% in the uninformed group. Erectile dysfunction specifically: 31% versus 10%. Same drug, same dose, a 30 percentage point gap produced entirely by a sentence.
That's the nocebo effect, and sexual function is about as susceptible to it as anything in medicine. It also means the loudest warnings on the internet are partly self-fulfilling.
Which brings us to post-finasteride syndrome, the claim that these effects persist indefinitely after stopping. Biologically we don't have a good explanation for why that would happen. The frequently cited evidence is a 2011 Journal of Sexual Medicine paper that interviewed 71 men who had reported persistent sexual problems more than three months after stopping. Among them, 94% reported low libido and 92% reported erectile dysfunction. Note the design: these are men who were recruited because they already had the complaint. Are there men who attribute lasting problems to finasteride? Yes. Is it conceivable? I think so. The fear of it has outrun the data by a considerable margin. One more complication: men taking hair drugs are often entering the stretch of life when these problems show up anyway. Though I hear the worry most from men in their twenties and thirties, which is not a group with a high background rate.
Surgery deserves a mention because it works on a different principle. Transplants harvest follicles from the DHT-insensitive back and sides and move them to the top. Those follicles keep their insensitivity. The Turkey packages you've seen, roughly $10,000 all-in including the flight and a nice hotel, are often cheaper than the same procedure here, and the outcome rates at the established clinics look legitimate. You do get what you pay for. Don't buy the discount version.
Women lose hair differently, and the research is thinner
About 90% of postpartum women report hair loss, and it is dramatic. The mechanism is telogen effluvium. During pregnancy, hairs that would normally have cycled out stay put, which is where the thick, shiny pregnancy hair comes from. After delivery they all enter telogen at once and shed together. It typically resolves within a year or so, though it can drag on. It is common and not a sign of anything wrong, other than that you miss your pregnancy hair.
Telogen effluvium follows any significant shock: high fever, complicated surgery, severe illness, crash dieting. GLP-1 drugs are the newest form of crash dieting, and hair loss is commonly reported on them. We don't know for certain it's causal, and the leading speculation is that rapid weight loss is the shock. Telogen lasts a couple of months, so the hair falls out two to three months after the triggering event, by which time most people have stopped connecting the two.
Chemotherapy is a different mechanism. Chemo targets fast-dividing cells, which is why it works on cancer, and hair follicles divide fast too. Scalp cooling genuinely works, with randomized trial evidence in women undergoing breast cancer chemotherapy. It preserves more hairs, not every hair. The cost is time: cooling starts before the infusion, runs through it, and continues for a couple of hours after. More insurers are covering it, but check first so you don't get billed out of pocket on the grounds that it's cosmetic, which is a genuinely irritating classification.
Female pattern hair loss, the thinning at the crown, gets less research attention than male pattern loss. What exists is mostly minoxidil. A 2023 meta-analysis comparing minoxidil to spironolactone, another repurposed blood pressure drug known to grow hair, found about half of women in both arms reported improved growth. That's a subjective endpoint. The largest double-blind randomized trial, from 2004 in 381 women, found minoxidil improved hair count over placebo by around 5%. Not awesome. Something.
Finasteride isn't the answer here either, and not only because of the pregnancy issue. Female pattern hair loss doesn't appear to be androgen-driven. Even when you measure testosterone and DHT in women, the levels don't correlate with the hair loss. Anti-androgen therapy aimed at a mechanism that isn't operating shouldn't be expected to deliver much.
Nutrafol comes up constantly. It's a supplement containing ashwagandha among other things, the brand has run its own randomized trials, and a 2023 JAMA Dermatology review placed it on a list of things with possible benefit. The effects in the company-run trials are small, maybe 10% or a bit less, in a couple dozen people. It's expensive. The company also states that it works best if you rest more and lower your stress, which are things that help anyway.
Onion juice, rosemary oil, and the rest of the internet
Men have been rubbing substances into their scalps for a very long time. The Ebers papyrus, from around 1550 BCE, prescribes the fat of lions, hippos, crocodiles, cats, snakes, and ibex. Hippocrates, our eunuch-observing friend, recommended a mixture of opium, horseradish, pigeon droppings, and beetroot. Julius Caesar was self-conscious enough about his balding head, per Suetonius, that he combed his hair forward and liked the laurel wreath for the cover it provided. So the influencer telling you to grind onions onto your scalp is working in an ancient tradition. There is no evidence it does anything, and he will smell like a bad salad. Watch for the product link at the end of the clip. It's always there.
Rosemary oil is the more interesting case because there is an actual trial behind it. Published in Skinmed in 2015, 100 patients randomized to rosemary oil or 2% minoxidil for six months. The stated conclusion was that both groups increased hair counts with no difference between them, which influencers have translated into "rosemary oil is as good as minoxidil and it grows in your garden."
Read the figure instead of the abstract. The bar graph of hair counts at baseline, month three, and month six is essentially flat in both arms. The rosemary group sits around 120 hairs at baseline and maybe reaches 130 at six months. Minoxidil starts near 130 and stays near 130. The minoxidil group had more hair at every timepoint, including before the trial started. The study is small and the bars overlap heavily. "No difference between groups" is doing a lot of unearned work.
Which is the broader problem. Whether it's on your head or on a graph, ten hairs is difficult to see.
Red light therapy has one trial showing increased hair density versus sham, with the sham group's hair thinning by 15 micrometers over 16 weeks. The effect size is larger than most of what's in this article. It also hasn't been replicated, and I'm wary of a single impressive result. If you try it, the device matters.
Platelet-rich plasma injections into the scalp, a vampire facial for your head, have a meta-analysis of randomized trials covering about 400 patients and suggesting a gain of roughly 27 hairs per square centimeter. The authors describe their own included trials as highly heterogeneous, of low quality, with evident publication bias. That reads as: when scalp injections fail, the results go in a drawer, and when they happen to work, they get published.
What's actually in the pipeline
If you are a small-cap biopharma company with a hair loss candidate, you will find investors. Keep that in mind while reading what follows.
An extended-release oral minoxidil is coming, and I expect it to work, because there's no reason it wouldn't. Smoothing the blood levels should reduce the blood pressure dips. The business logic is that plain oral minoxidil is generic and cheap, while the extended-release version will be branded and expensive.
Clascoterone is a new topical anti-androgen already used for acne at other concentrations. The company announcement touts a 539% relative improvement in target area hair count without saying what the counts actually were. Did it go from one hair to 5.39 hairs? The announcement is from December 2025 and there's still no published paper, which I find puzzling.
Then there's PP405 from Pelage, which is the one generating real hype. The pitch is that it reactivates dormant scalp stem cells and grows hair in areas that were effectively dead. It's a topical, and the mechanism involves pyruvate metabolism in the mitochondria. What exists so far is a press release on a phase 2a trial: 78 men and women, applied daily for four weeks, followed to week 12. The primary endpoint was safety, which it met. Hair growth was exploratory. The reported result is that at week eight, among men with more advanced hair loss, 31% had a greater than 20% increase in hair density versus 0% on placebo.
Count the qualifiers. They enrolled men and women, then reported men. Then men with more advanced loss. Then the proportion of them clearing a 20% threshold, rather than the average effect. That construction usually means the average effect wasn't much. Phase 3 will tell us. This is not investment advice.
Bottom line
Finasteride and minoxidil work. They mostly preserve what you have rather than restore what you've lost, and the gains in trials are counted in dozens of hairs per square inch against a baseline in the high hundreds. Transplant surgery works on real biology and has legitimate outcomes. Scalp cooling works during chemotherapy, at the cost of hours. Most everything else, onion juice, rosemary oil, PRP, is either unsupported or supported by trials too weak to lean on.
The sexual side effects of finasteride are real but uncommon, usually mild, and usually reversible, and a large share of what men report comes from having been warned. The permanent-damage story that dominates search results rests on a self-selected survey of 71 men.
And if you're wondering whether to skip all of it and shave your head, there is evidence that men with digitally removed hair are rated taller, stronger, and more dominant. They're also rated slightly less attractive and nearly four years older. I had an AI generate a shaved version of myself before committing. Strongly recommend that order of operations.
I covered this in depth on Wellness, Actually, listen below.
Frequently asked questions
How much hair does finasteride actually grow?
In the pivotal trial, men had about 876 hairs in a one-inch circle of scalp at baseline and gained roughly 107 hairs over that baseline at one year, rising to about 138 at two years. Most of that effect is loss prevented rather than new regrowth. Finasteride reduces scalp DHT by about two-thirds, which shortens the hormonal pressure on the follicle.
Does finasteride cause permanent sexual side effects?
The evidence for permanent effects is weak. A systematic review found sexual adverse events in about 2% to 7% of treated men versus 1% to 4% on placebo, most of them mild and reversible on stopping the drug. The main cited evidence for post-finasteride syndrome is a 2011 Journal of Sexual Medicine paper interviewing 71 men who were recruited precisely because they already reported persistent problems, which is not a design that can establish causation.
Is oral minoxidil better than topical minoxidil for hair loss?
There is no good data showing oral is superior. A head-to-head randomized trial found daily oral minoxidil 5 mg was not superior to topical 5% after 24 weeks. Oral is easier to take, but it can grow hair in unwanted places and lower your blood pressure, which matters if your pressure is already well controlled.
Does rosemary oil work as well as minoxidil for hair regrowth?
No. The claim traces to a 2015 Skinmed trial of 100 patients randomized to rosemary oil or 2% minoxidil for six months, which reported no difference between groups. Looking at the actual hair count figure, both bars are essentially flat across six months and the minoxidil group had more hair at every timepoint including baseline. The study is too small to support the conclusion influencers draw from it.
Why do you lose hair after pregnancy and does it grow back?
About 90% of postpartum women report hair loss, caused by telogen effluvium. During pregnancy, hairs that would normally shed stay in place, and after delivery they all enter the shedding phase at once. It typically resolves within about a year, though it can last longer for some people.
Does scalp cooling prevent hair loss from chemotherapy?
Yes, it works, with randomized trial evidence in women undergoing breast cancer chemotherapy. It preserves more hairs but not every hair. The main cost is time, since cooling runs before, during, and for a couple of hours after the infusion. More insurers are covering it, but confirm coverage in advance because some still classify it as cosmetic.
Wellness, Actually Podcast
"What's the deal with hair regrowth?" — Listen to the full episode, including the week's health news and listener Q&A.
About the author
F. Perry Wilson, MD MSCE is a nephrologist, clinical researcher, and Associate Professor of Medicine and Public Health at Yale University, where he directs the Clinical and Translational Research Accelerator. He hosts the Wellness, Actually podcast with Emily Oster, writes the weekly Impact Factor column on Medscape, and is the author of How Medicine Works and When It Doesn't (Grand Central, 2023).