Hair Transplant: How Does It Work? The Biology, the Surgery and the Clock

Table of contents

• Hair transplant: how does it work at the level of the follicle?
• What a graft actually is, and why 3,000 grafts is not 3,000 hairs
• The six steps of the operation
• Why handling and time matter alongside technique
• The density arithmetic that sets your ceiling
• What happens in the months afterwards
• Two things a hair transplant does not do
• Three things the standard explanation gets wrong
• Who should be holding the instruments
• Frequently Asked Questions
• Where to take this next

Hair transplant: how does it work at the level of the follicle?

A hair transplant works because a hair follicle carries its own susceptibility to male pattern loss with it wherever it is moved. Norman Orentreich called this donor dominance in his 1959 paper on autografts in alopecia, published in the Annals of the New York Academy of Sciences. A follicle taken from the occipital scalp and placed in a bald frontal scalp behaves like occipital hair, not like the hair that used to grow there. That single observation is the foundation of every hair restoration procedure performed since.

The molecular reason came later. Sawaya and Price, in a 1997 study in the Journal of Investigative Dermatology, biopsied frontal and occipital scalp in 12 men and 12 women with androgenetic alopecia and measured the machinery that drives pattern loss. Frontal follicles carried higher levels of androgen receptor and of 5-alpha-reductase types I and II than occipital follicles from the same heads. Occipital follicles carried more aromatase, the enzyme that diverts testosterone away from the DHT pathway.

So the back of your head is not lucky. It is biochemically different. The follicles there are less responsive to the hormone that miniaturises the follicles at the front, and moving them does not change that.

Two consequences follow, and they shape everything else:

• The donor supply is finite and non-renewable. The ISHRS puts it bluntly on its own FUE information page, which also gives the micro punch sizes used to harvest, 0.7 to 1.2 mm: "Donor hair, once removed, is gone forever."
• Hair loss is a process, not an event. Surgery relocates follicles. It has no effect at all on the follicles it leaves behind.

If you have not placed your own loss on the standard scale yet, the Norwood stage finder is the faster first step. Stage determines how much area needs covering, and area determines whether the arithmetic below works in your favour.

What a graft actually is, and why 3,000 grafts is not 3,000 hairs

A graft is one follicular unit, and a follicular unit usually contains more than one hair. Human scalp hair does not emerge one strand at a time. It emerges in natural bundles, described and named by John Headington in a 1984 paper in Archives of Dermatology on the transverse microscopic anatomy of the scalp. Most units hold one to four terminal hairs along with their sebaceous glands and connective tissue sheath.

That means a graft count and a hair count are different numbers, and quotes are written in grafts.

Jimenez and Ruifernández measured the donor zone directly in 50 patients for a 1999 paper in Dermatologic Surgery. In occipital scalp they found 65 to 85 follicular units per cm² and 124 to 200 hairs per cm², with units spaced 1.0 to 1.4 mm apart. Pair the low figures together and the high figures together, since dense scalps are dense on both measures, and each unit carries roughly 1.9 to 2.4 hairs.

Run that against real session sizes:

Grafts quoted	Hairs actually moved (at 1.9 to 2.4 per unit)
1,000	1,900 to 2,400
2,000	3,800 to 4,800
2,347 (ISHRS mean first procedure)	4,460 to 5,630
3,000	5,700 to 7,200
4,000	7,600 to 9,600

The 2,347 figure is the mean first procedure recorded in the 2025 ISHRS Practice Census, with subsequent procedures averaging 1,637 grafts and a mean of 1.5 procedures per patient.

This cuts both ways. A graft moves roughly twice as much hair as the word suggests, but the same ambiguity is what lets a package quote a big-sounding number. Ask for the expected breakdown of one, two, three and four hair units: a 3,000 graft session made up mostly of singles moves far less hair than one weighted toward triples. Typical graft ranges by stage are on the individual Norwood pages, for example Norwood 4, where a first transplant is most commonly performed.

The six steps of the operation

The surgery itself is six steps, and the American Academy of Dermatology's patient page on hair transplants puts the whole thing at four to eight hours, with most patients awake throughout and needing only anaesthesia that numbs the scalp.

Step	What happens	What decides whether it goes well
1. Design and donor assessment	Hairline drawn on the patient, donor density and laxity measured, graft count and coverage area agreed	Whether the plan accounts for loss that has not happened yet
2. Anaesthesia	Local anaesthetic and tumescent fluid infiltrate donor and recipient zones	Epinephrine concentration and tumescence volume, both of which affect blood supply
3. Harvesting	FUE: micro punches of 0.7 to 1.2 mm score the skin around each unit, which is then lifted out. FUT: a strip 0.5 to 1.5 cm wide and 5 to 30 cm long is excised and the donor area closed with sutures or staples	Punch size, angle alignment with the hair, and how much of the donor is taken
4. Preparation	FUT strips are slivered and dissected under microscopes into 1, 2, 3 and 4 hair units. FUE grafts are sorted and trimmed. All are held in chilled solution	Hydration, temperature, and how long grafts wait
5. Recipient site creation	Thousands of incisions are cut at the angle, direction, depth and spacing the final hair will follow	This step, more than any other, is what makes a result look like hair or like a procedure
6. Placement	Units are inserted one at a time with fine forceps or a sharp implanter	Crush pressure on the graft and the number of attempts per site

Step 5 and step 6 are where the technique names live. Premade sites means every incision is cut first and the grafts go in afterwards. Stick and place means each site is cut and filled immediately, which is what a sharp implanter pen does in a single motion. That is the entire content of the DHI hair transplant label, which the ISHRS says is a placement technique rather than a separate operation.

Notice what is not on the list: the hair is not "implanted" into anything. It is placed into a fresh incision in skin that has to grow a new blood supply around it.

Why handling and time matter alongside technique

From the moment a follicular unit leaves the donor scalp it is an avulsed piece of living tissue with no blood supply, and everything that happens next is a race. Parsley and Perez-Meza's 2010 review of graft growth and survival in the Journal of Cutaneous and Aesthetic Surgery is the clearest published account of this, and it is the part of the procedure that almost no consumer page describes.

Three findings from it are worth knowing before you sit in a chair for eight hours.

Grafts take about three days to start regaining a blood supply. Until then they are perfused passively by the surrounding tissue. During that window the graft is vulnerable to dehydration, mechanical trauma, hypoxia, ATP depletion and cold injury.

Drying out is fast. Reported survival time for grafts sitting in a dry environment, on a glove or a gauze pad, runs from as little as three minutes to more than sixteen before significant graft death occurs. That is why well-run rooms keep grafts in chilled solution and why the review notes that mistakenly using sterile water as a holding solution can kill every graft in the batch.

Longer out-of-body time can reduce survival. One older in vivo study by Limmer, summarised in the review, tested follicular unit grafts stored in chilled saline. Its results are often condensed into a one-percent-per-hour rule of thumb, but the observed values were not perfectly linear and should not be treated as a universal prediction:

Time out of body	Graft survival
2 hours	95%
4 hours	90%
6 hours	86%
8 hours	88%
24 hours	79%
48 hours	54%

Those losses are not the only ones. Park and You examined 100 grafts from each of 42 patients harvested with a 1 mm sharp punch for a 2017 paper in Plastic and Reconstructive Surgery Global Open. Under a magnifying loupe they found an average transection rate of 7.40%, and under a microscope 6.34%, plus paring, fractures, dermal papilla injury and bulb injury in a further few percent of follicles that a loupe would call intact.

These studies describe separate failure points under different conditions, so their percentages cannot simply be added. Extraction injury, dehydration, storage time and placement trauma can all affect growth, but no published combined figure lets you calculate the survival rate of a specific clinic from those studies. When a clinic quotes "95% graft survival", ask how it was measured and which losses the number includes. That answer tells you more than any photograph in its before and after gallery.

A 2026 review of FUE complications in Frontiers in Medicine puts overall complication rates in large series at 1.2% to 4.7% and attributes poor growth mainly to follicular injury during harvesting, handling or implantation, prolonged out-of-body time, dehydration, smoking and vascular disease. A residual 0.5% to 1% grow poorly for no identifiable reason, known in the field as factor X.

The density arithmetic that sets your ceiling

Transplanted hair is placed at roughly half the density it grew at in the donor area, and understanding that number prevents most disappointment. Surgical placement density in the literature runs 30 to 40 grafts per cm². Native occipital scalp runs 65 to 85 units per cm². So a restored area carries somewhere around 40% to 55% of the density that never-balded scalp carries.

It looks like hair anyway, because the eye reads contrast between hair and scalp rather than counting strands. That is the illusion the operation is built on, and it works better on fine light hair against pale skin than on coarse dark hair against pale skin.

Why not pack the grafts in tighter? Because the evidence is genuinely split, and it is one of the honest uncertainties in this field. The original density study by Mayer, summarised in the Parsley review, planted two-hair units at four densities and reported survival of 97% at 10 grafts/cm², 92% at 20, 72% at 30 and 78% at 40. Nakatsui and colleagues found the opposite in a 2008 Dermatologic Surgery study, reporting 98.6% growth at 72 grafts/cm² against 95.6% at 23 using a lateral slit technique. That second study, the one most often cited to justify dense packing, was conducted in a single patient. The 2026 complications review takes the cautious side, recommending densities below 50 units/cm² to avoid recipient-site necrosis.

The donor side has its own arithmetic, and it is the one that limits lifetimes rather than sessions. The 2026 review notes that many authors recommend extracting no more than 10% to 20% of baseline follicular unit density per session, and that exceeding it produces visible thinning, a moth-eaten appearance or permanent donor depletion. Take a donor area at 75 units per cm² and harvest at 15%: that is about 11 units per cm², so an average 2,347 graft session has to be spread across roughly 210 cm² of donor scalp to stay inside the safe rate.

That is why surgeons talk about a lifetime donor supply rather than a per-session maximum, and why advanced stages such as Norwood 6 are a coverage decision rather than a density decision.

What happens in the months afterwards

Transplanted hair falls out before it grows, and this catches people who were not warned. The AAD's patient page states that between two and eight weeks after surgery the transplanted hair will fall out, that this is normal, and that by the third month the hair may look thinner than it did before the transplant. Results typically appear between six and nine months, and for some patients it takes twelve.

The shaft is shed. The follicle stays, resets into a resting phase and then re-enters growth, which is why the calendar is measured in months.

The month by month version, including what is normal at each point and what warrants a phone call to the surgeon, is on the hair transplant timeline, and the dip most people panic about is covered on the one month after a hair transplant page.

Two things a hair transplant does not do

It does not stop pattern hair loss. The follicles that were going to miniaturise are still going to miniaturise, on the same schedule, in the skin between the new grafts. The five-year placebo arm of the finasteride registration trials, reported in the FDA-approved Propecia label, gives the scale: 75% of untreated men were rated visibly worse than baseline at five years, 19% unchanged and 6% improved. A hairline built in front of hair that follows that trajectory ends up as an island.

This is why medication comes up in surgical consultations. In a randomised, double-blind, placebo-controlled trial of 79 men aged 20 to 45 who took finasteride 1 mg from four weeks before surgery to 48 weeks after, published in Dermatologic Surgery in 2005, 94% of the finasteride group showed visible increases in frontal and superior scalp hair at week 48 against 67% of the placebo group. Finasteride is prescription-only in most countries and is not indicated for women. The same Propecia label's year-one adverse event table lists decreased libido in 1.8% of treated men against 1.3% on placebo, erectile dysfunction in 1.3% against 0.7% and ejaculation disorder in 1.2% against 0.7%, with 1.2% stopping treatment for a sexual adverse experience against 0.9% on placebo. The fuller picture, including what is and is not established about persistent symptoms, is in the guide to finasteride side effects.

It does not add hair to your head. The total number of follicles is unchanged, and some grafts may not survive. Coverage at the front is bought with density at the back: a fair trade when the donor is dense and the recipient area is modest, a poor one when it is the other way round. That is the real content of the phrase "you are not a good candidate".

Three things the standard explanation gets wrong

Fact-checking this article against the highest-ranking patient explanations turned up three recurring errors. No page is named, because the errors are inherited rather than invented.

1. Graft sizes of 10 to 15 hairs, and sessions of about 50 grafts. Several widely cited pages still describe round punch grafts holding 10 to 15 hairs, mini-grafts of 2 to 4 and strip grafts of 30 to 40, with roughly 50 punch grafts per session. Those are the specifications of the operation performed before the follicular unit era, and they are the reason hair plugs looked the way they did. The 2025 ISHRS census records male harvesting as 85.4% FUE, 12.5% strip and 2.1% a combination of the two, in sessions averaging over 2,000 grafts.

2. Scalp reduction, flap surgery and tissue expansion listed as current options. These are real reconstructive operations. They are not what is meant today by a hair transplant for pattern loss, and they do not appear as categories in the professional body's own practice census.

3. "The transplanted hair is permanent." Donor dominance means the follicle keeps its own programming. It does not mean the hair is exempt from ageing, and it certainly does not protect the hair around it. The honest version is that transplanted follicles behave like the donor hair they came from, for as long as that donor hair would have lasted.

Who should be holding the instruments

Creating recipient sites is surgery, and the professional body says so explicitly. The ISHRS position statement on qualifications for scalp surgery, revised in July 2024, states that any procedure involving a skin incision to prepare the scalp to receive tissue, including creating recipient sites by any means, is a surgical procedure that must be performed by a properly trained and licensed physician, with delegation recognised only to licensed professionals under supervision.

That matters because the same census found 59.4% of members reporting black-market clinics in their own city, and repair of another physician's or a black-market transplant accounting for 6.9% of patients.

Questions worth asking, none of which require you to know anything about surgery:

• Who personally cuts the recipient sites, and who performs the extraction? Ask for names and credentials, not "the team".
• How many grafts, over how many cm², at what density? Grafts alone is not a plan.
• What is the expected breakdown of one, two, three and four hair units?
• How long will grafts be out of the body, and in what holding solution?
• What percentage of my donor density are you extracting in this session?
• What is the plan for the hair I still have?
• Is a second session anticipated? The mean is 1.5 procedures per patient, so a quote assuming one may be incomplete.

Price varies by an order of magnitude across markets, and the variable is rarely the biology. The hair transplant cost calculator prices a given graft count across nine countries, the breakdown of hair transplant costs for men turns those ranges into a realistic total, and the methodology page documents where the ranges come from.

Frequently Asked Questions

How does a hair transplant work in simple terms?

Hair follicles from the back and sides of the scalp are genetically less sensitive to DHT than the follicles at the front and crown, and they keep that property when they are moved. A surgeon removes them in their natural clusters of one to four hairs, then places them into tiny incisions in the balding area at the angle and direction the hair should grow. The hair sheds within two months, then regrows from the transplanted follicle over the following six to twelve.

Is transplanted hair permanent?

Transplanted follicles retain the characteristics of the donor area they came from, so they are not lost to pattern balding the way frontal hair is. That is not the same as permanent: the hair still ages, and it gives no protection to the untransplanted hair around it, which continues to miniaturise on its own schedule.

How many grafts will I need?

It depends on the area to be covered and the density placed in it, not on a stage number alone. At the usual placement density of 30 to 40 grafts per cm², covering 60 cm² takes roughly 1,800 to 2,400 grafts. Typical clinic ranges for each Norwood stage are on the individual stage pages, and the ISHRS records a mean first procedure of 2,347 grafts.

Does a hair transplant hurt?

Most patients are awake for the whole procedure with only local anaesthetic numbing the scalp, according to the AAD, and the injections themselves are the part most people describe as uncomfortable. The 2026 Frontiers in Medicine review puts pain beyond routine discomfort at around 6% of cases in the first 24 to 48 hours. Pain that worsens after day three is a reason to contact the surgeon rather than wait.

Where to take this next

The mechanism is simple and the execution is not. Donor dominance makes the operation possible, follicular unit anatomy sets how much hair each graft carries, placement density shapes how it looks, and careful extraction, storage and placement all influence survival. A clinic that can answer on all four is describing an operation. One that answers only in graft counts and package prices is describing a sale.

If you are weighing this up, take your donor assessment to a board-certified dermatologist or a hair restoration surgeon and ask them to talk through lifetime donor supply before anything else, including price. That number is the ceiling on every decision that follows it.