Hair Transplant Grafts: What the Number on Your Quote Means

Table of contents

• What hair transplant grafts actually are
• A graft is a unit of account, not a unit of hair
• The counting chain, in the profession's own terms
• How many hair transplant grafts do I need?
• Your donor area sets the ceiling, not your budget
• Why per-graft pricing and package pricing bend the count in opposite directions
• How many of your grafts actually grow
• What to get in writing before the deposit
• Frequently Asked Questions

What hair transplant grafts actually are

An intact graft usually contains one follicular unit: the natural grouping of hairs that emerges together from the scalp, along with its sebaceous glands and surrounding tissue. The anatomy was described by Headington in 1984, working from transverse sections of scalp biopsies (Arch Dermatol, PMID 6703750). A surgical team can also divide a unit into smaller grafts, as the experiment below shows. How many hairs an intact unit carries is a separate question, answered by the donor-scalp counts below rather than by Headington's paper.

The measured numbers come from Jimenez and Ruifernández, who photographed and counted the occipital scalp of 50 patients (Dermatol Surg 1999, PMID 10417585). In the donor region there are 65 to 85 follicular units per cm², carrying 124 to 200 hairs per cm², spaced 1.0 to 1.4 mm apart. They counted one, two and three hair units, and warn that variance is to be expected because the donor area is not uniformly dense. Dividing one of those published ranges by the other gives roughly 1.9 to 2.4 hairs per unit, which is arithmetic on two ranges rather than a ratio the study measured directly.

Other measurements sit at or below the bottom of that band. A single-centre study in Islamabad marked ten 1 cm² boxes in each of ten men, extracted with a 0.9 mm punch and counted every hair that came out, finding a graft-to-hair ratio of 1:2.3 with a per-patient range of 1:1.65 to 1:2.75 (World J Plast Surg 2018, PMID 30083502). Ten men at one clinic is a small sample, so read it as a measurement rather than a norm.

Hair type appears to matter too, though the published figures here are clinical observations rather than trial results. Writing in the ISHRS journal, one surgeon reported a hair-per-graft ratio of 1.9 in his Caucasian patients and 1.63 in his Asian patients, and cited Pathomvanich at 1.8 and Bernstein at 1.7, both also in Asian patients (Tsilosani, Hair Transplant Forum International 2009, 19(4):128-130).

So the conversion is real but it is a range, and it is yours, not a constant:

Grafts quoted	At 1.7 hairs/graft	At 2.0	At 2.3
1,500	2,550 hairs	3,000	3,450
2,000	3,400	4,000	4,600
2,500	4,250	5,000	5,750
3,000	5,100	6,000	6,900
4,000	6,800	8,000	9,200

Read the spread across a row rather than down a column. At 3,000 grafts the difference between the low and high ratio is 1,800 hairs, which is more hair than some people get transplanted in total. Any clinic can measure your ratio in five minutes with a densitometer before you pay for anything. Ask for the number.

A graft is a unit of account, not a unit of hair

The graft count can be increased without adding a single hair, because a multi-hair follicular unit can be dissected into two or more smaller grafts. The ISHRS has a formal name for this: external dissection, defined in its standard terminology as dividing a graft under the microscope "into singles or groups containing fewer follicles than the original intact group."

The clearest demonstration is a single-patient experiment published in the society's own journal. A surgeon took one 20 cm donor strip from a 32-year-old man with Norwood 4A loss, cut it in two, and handed each half to a separate four-person cutting team, instructing the second team to divide every unit it could (Tsilosani, Hair Transplant Forum International 2009, 19(4):128-130).

	Team 1: units kept intact	Team 2: units divided
Grafts produced	1,002	1,404
Follicles contained	1,834	1,676
Hairs per graft	1.83	1.19

Same strip, and the divided half yielded 40% more grafts out of slightly fewer follicles. Read that follicle row carefully. The author says the two halves were probably not cut exactly equal, so the 158-follicle gap is a quirk of the strip rather than evidence that dividing destroys hair. The graft row is the point: dividing units can increase the number a clinic bills for under per-graft pricing even when it adds no hair. In this case hairs per graft fell from 1.83 to 1.19.

His own conclusions are worth stating fairly, because they cut against the obvious reading. At five months he thought the divided side looked slightly fuller on close inspection, while adding that the difference was not noticeable at conversational distance, that the two sides were also implanted at different densities, that one case supports no firm conclusion, and that the patient was lost to follow-up well before the 12 to 18 months a final result takes. He also cites a 1999 Beehner comparison in which grafts from subdivided units survived a little better, a difference Beehner reported as not statistically significant, which is not the same as a demonstrated advantage. Splitting units is a legitimate technique, particularly for building a soft hairline out of single-hair grafts.

The problem is not the scalpel. It is the invoice.

The society's own ethics column laid it out in the same journal (Williams, Hair Transplant Forum International 2018, 28(4):149). Writing about what a surgeon should do when a planned graft number is not achieved, he notes that "if charging by graft then, if there is a shortfall, the planned number can be artificially manufactured by splitting grafts," and asks directly: "Is this being honest since it does not in fact increase the number of hairs transplanted?" The same article says plainly that "it is important for patients to clearly understand the difference between 'follicles' and 'grafts,' and that a graft is usually equivalent to a follicular unit not a follicle."

When a profession's own ethics column has to spell that out for surgeons, it is unreasonable to expect patients to have worked it out from a price list.

The counting chain, in the profession's own terms

There is not one graft number in a hair transplant. There are at least six, and the ISHRS standard terminology defines each one separately. Its FUE terminology document states that to calculate the quality metrics, five things "must be observed, counted, and recorded in all surgeries": the number of punch insertions, all pieces of tissue removed regardless of appearance, the number of partially and completely transected follicles, the number of intact and transected follicles per graft, and pre-operative densitometry.

Term	What it counts	Why the distinction matters
Total punch insertions	Every punch attempt, whether it yields a graft or not	Includes attempts that produce no usable graft
Total grafts excised	Available plus unavailable grafts	Includes tissue that cannot be implanted
Grafts unavailable for transplant	Capped, completely transected and empty grafts	A "capped" graft is skin with no follicle in it; an "empty" graft is a punch into bald skin
Missing grafts (MGR)	Grafts that cannot be located because they were misplaced during surgery	Each one is a donor scar with nothing to show for it
Grafts available for transplant	Intact plus partially transected grafts	Available for insertion does not mean actually implanted
Calculated follicles per graft achieved (CFGA)	Intact follicles divided by grafts available plus completely transected grafts	A measure of excision quality, not a count of implanted hairs

Notice what falls out of that list. Excised, available and implanted grafts are different counts. The terminology document defines the first two; your clinic should say which count appears on the invoice and report how many grafts it actually placed. Missing grafts and follicles per graft add context that a single billed count cannot provide.

None of this is obscure. It has been published, free to read, on the professional body's website since 2019. The FUE clinical practice guidelines devote a section to quality control measures and name three: transection rate, average hairs per graft, and missing grafts. On transection the guideline is specifically about follicles rather than grafts: it says it is important to minimise follicle transection rates, and that "striving for a rate less than 10% would be prudent." On missing grafts it says the rate "should be as low as possible as an increased rate results in excessive and unnecessary donor scarring."

Ask a clinic for its follicle transection rate and its average hairs per graft on your case, and you are not asking for a trade secret. You are asking for two of the three numbers its own professional body says it should already be recording.

How many hair transplant grafts do I need?

The graft count is an area calculation, not a stage lookup. The arithmetic is: grafts needed = area to be covered in cm² × the density the surgeon plans to place. Published practice puts that density at roughly 30 to 40 follicular units per cm² in the frontal zone. A retrospective series of 820 men with Norwood 5 to 7 loss describes maintaining 35 to 40 FU/cm² at the front with a reducing gradient to 20 to 25 toward the vertex (J Cutan Aesthet Surg 2019), and packing above about 40 units per cm² is what the literature calls "dense packing", whose apparent advantages have been openly questioned (Unger, J Investig Dermatol Symp Proc 2005).

That is why a hairline rebuild and a crown can carry wildly different numbers for the same person. A 40 cm² frontal zone at 35 grafts/cm² is about 1,400 grafts. A crown whorl spirals outward and covers a lot of area, and in that same 820-case series it is also the zone planned at the lowest density, so graft count and finished appearance do not track each other there the way they do at the hairline. The density ceiling and the evidence behind it are covered in the guide to how a hair transplant works; the short version is that a transplanted area typically ends up carrying roughly half the density of scalp that never balded, and packing higher does not reliably buy more surviving hair.

Our per-stage planning ranges are published on the stage pages, so a Norwood 3 plan and a Norwood 6 plan start from different arithmetic. If you are not sure which pattern you match, the Norwood stage finder will place you in about a minute. Treat those ranges as the starting point of a conversation, never as a quote.

For scale, 173 members responding to the ISHRS 2025 Practice Census reported a mean of 2,347 grafts in a first procedure and 1,637 in a subsequent one. Asked separately about the average in a typical case, 185 members gave a mean of 2,262 grafts for FUE and 2,100 for FUT, and only 2.2% of them put their typical FUE case at 4,000 grafts or more. Those are self-reported averages from a voluntary survey of one professional society, not a ceiling for any individual, but they are a useful reality check: a quote for 5,000 grafts in one sitting sits in a band that very few practices describe as typical.

When two clinics quote you numbers 2,000 apart, ask each to specify the recipient area, planned graft density, whether units will be split and what its graft count means. Area multiplied by density gives the planned placement count. Splitting units can increase the graft count without adding hairs; the hairs-per-graft ratio changes the expected hair yield. Without those definitions, the two quotes are not comparable.

Your donor area sets the ceiling, not your budget

Donor supply is finite, non-renewable and individual. Extraction moves hair; it does not create it. The ISHRS says so bluntly on its own FUE patient page: "Donor hair, once removed, is gone forever."

How much can come out in one session is a percentage, not a fixed graft number. A 2026 narrative review of FUE complications reports that many authors recommend limiting extraction to 10% to 20% of baseline follicular unit density per session, and that donor overharvesting, with its moth-eaten appearance and permanent depletion, is "increasingly observed in high-volume sessions exceeding 3,000 to 4,000 grafts" (Front Med 2026, PMID 41709896). That is a pattern the reviewers describe, not a measured rate, and the same section puts the prevention target at 15% to 20% elsewhere, so treat it as roughly a tenth to a fifth rather than a precise threshold.

The Islamabad study above counted something different again: a mean donor density of 154.76 hairs/cm², of which 54.85 hairs/cm² were extracted, or 35.44% of the hairs, with a per-patient range of 28.9% to 42.8%. Hairs and follicular units are not the same denominator, so that 35% does not sit on the same scale as the 10% to 20% above and the two should not be read as one exceeding the other. Its authors' conclusion was one sentence long: "As the donor density varies, the FUE should be performed with caution."

A man with 85 follicular units per cm² across a wide safe zone has a materially different lifetime graft budget from a man with 60 per cm² across a narrow one, and no chart can tell them apart. Both will be quoted from the same price list.

The donor area also cannot do the thing patients most want it to. A transplant redistributes hair; it does not treat the condition that thinned it, and native hair around the grafts keeps miniaturising on its own schedule. That surrounding native hair is what one randomised, double-blind, placebo-controlled trial set out to measure. In 79 men aged 20 to 45, finasteride 1 mg taken daily from four weeks before until 48 weeks after surgery improved scalp hair around the transplant against placebo at week 48, on both an expert dermatologist's review of global photographs and macrophotographic hair counts (p < .01 for each). Visible increases in superior and frontal scalp hair were recorded in 94% of the finasteride group and 67% of the placebo group, a descriptive split the paper reports without a significance test, and that 67% is a reminder that these men had all just had surgery (Leavitt et al, Dermatol Surg 2005, PMID 16188178). The trial measured the hair around the grafts, not the grafts themselves. Finasteride is prescription-only in most countries and is not suitable for women who may become pregnant; in the first year of the controlled trials the FDA label for Propecia records drug-related decreased libido in 1.8% of men against 1.3% on placebo, erectile dysfunction in 1.3% against 0.7% and ejaculation disorder in 1.2% against 0.7%, and the broader and genuinely contested picture is set out in the guide to finasteride side effects. The decision belongs with a prescriber, not a price list, but the graft plan and the medication question are not separable.

Why per-graft pricing and package pricing bend the count in opposite directions

Both common pricing models create an incentive that points at the graft number rather than at your hair, and they point opposite ways.

Per-graft pricing rewards a higher count. If the surgeon falls short of the quoted number, splitting units closes the gap without adding hair, which is exactly the scenario the ethics column above describes. Per-graft rates vary enormously by market, from the US band down, and the spread is large enough that a few hundred grafts either way is real money in some countries and rounding error in others.

Fixed-price package deals reward a higher count too, but for a different reason. Where a clinic advertises "maximum grafts" for a flat fee, the count becomes a marketing figure with no price attached to it, and the pressure runs toward harvesting more than the donor area can spare. That trade-off is set out in detail on the Turkey cost page, and the wider question of what actually moves a male patient's bill is in hair transplant cost for men.

Neither model is dishonest in itself. But comparing two quotes means comparing two incentive structures as well as two prices. The clean way to neutralise it: ask both clinics to quote the same recipient area at the same planned density, and to state the assumed hairs per graft. If one is still 2,000 grafts higher, something substantive differs and they can tell you what.

How many of your grafts actually grow

Not all of them, and the losses stack in places the graft count does not show.

One series put numbers on the harvesting damage. It took 100 grafts at random from each of 42 patients whose follicles had been extracted with a 1 mm sharp punch, and inspected them under a ×5.5 loupe and under a ×60 binocular microscope (Park and You, Plast Reconstr Surg Glob Open 2017, PMID 28458974). Per 100 grafts:

Finding	×5.5 loupe	×60 microscope
Transection rate	7.40%	6.34%
Follicles pared	4.31	9.07
Follicles fractured	1.90	1.95
Dermal papilla injury	1.52	0.79
Partial bulb injury	0.43	1.24

The microscope found about twice as much paring and nearly three times as much bulb injury, which is the authors' own headline. It did not find more of everything: transection and dermal papilla injury both scored lower under magnification, and the paper does not explain that inversion. These are one clinic's figures, not settled rates, and the reason to quote them at all is that none of this appears anywhere in a graft count.

Graft quality itself has a published grading scale, the Graft Quality Index, which sorts FUE grafts into four morphological grades from undamaged with full perifollicular tissue down to transected with irregular margins (True, Hair Transplant Forum International 28(2):45). It exists precisely because "3,000 grafts" describes quantity and says nothing about condition.

Handling subtracts more again, before a single hair grows. So the rule for reading any quote is short: a survival percentage with no stated denominator is not a claim about your result. Ninety-five percent of the grafts that reached the holding dish is a different sentence from 95% of the punch insertions.

The honest yardstick is the photograph at twelve months, not the count on the day, and how to read one is covered in hair transplant before and after. Graft counts have been used as a sales figure before, which is the story in hair plugs vs transplant.

What to get in writing before the deposit

None of these questions require medical knowledge to ask, and each one has a factual answer the clinic already holds.

1. What is my measured donor density, in follicular units and hairs per cm², and what is my hairs-per-graft ratio?
2. What recipient area, in cm², and what placement density does this graft number assume?
3. Will follicular units be divided, and if so roughly how many, and where?
4. What count will be billed: punch insertions, grafts excised, grafts available, or grafts actually implanted?
5. What happens if the count falls short or runs over? Refund, credit, or nothing?
6. What follicle transection rate and missing graft rate does this clinic record, and will mine be given to me afterwards?
7. Who holds the punch and who places the grafts, by name and role?
8. Is a second procedure anticipated, and what would it cost?

Question 4 is the one that most often changes the answer to everything else. Question 7 is the one that most often changes the outcome, and the reasons why are set out in the piece on the implanter pen and who should be holding it.

Frequently Asked Questions

How many hairs are in one hair transplant graft?

An intact follicular unit usually contains one to four hairs, with averages around 1.6 to 2.4 in the donor measurements discussed above. A unit divided into smaller grafts can yield fewer hairs per graft. Ask for your own measured hairs-per-unit ratio and whether the surgical team plans to split units.

How many grafts do I need for a hair transplant?

It depends on the area being covered and the density the surgeon plans to place, typically 30 to 40 grafts per cm², not on your Norwood stage alone. Surveyed surgeons reported a mean of 2,347 grafts in a first procedure in the ISHRS 2025 census, and our Norwood 4 planning range and the other stage pages give a starting point, but only an in-person donor assessment produces a real number.

Can a clinic inflate the graft count?

Yes, without lying, by dividing multi-hair follicular units into smaller grafts. In a published single-patient experiment, one half of a donor strip was cut into 1,404 grafts from 1,676 follicles while the other half yielded 1,002 grafts from 1,834 follicles, the author noting the halves were probably not exactly equal in size. The graft count rose about 40%; the hair did not. Asking what is counted as a graft, and what the assumed hairs per graft is, closes most of that gap.

Is 5,000 grafts in one session realistic?

It is uncommon. Only 2.2% of surgeons in the ISHRS 2025 census described their typical FUE case as 4,000 grafts or more, and a 2026 review of FUE complications reports that donor overharvesting is increasingly observed in sessions above 3,000 to 4,000 grafts. Large single sessions exist, but a very high number on a first quote is worth asking about rather than admiring.

Does a higher graft count mean a better result?

No. Coverage is decided by where the hairs go and how much hair each graft carries, not by the count alone. Whether the trade-off is worth making at all is examined in is a hair transplant worth it, and the surgical risks are in hair transplant side effects.

Where this leaves you

The graft number is the least informative figure in a hair transplant quote and the only one most patients are given. Area, density, hairs per graft and donor supply are what decide the result, and all four are measurable before any money changes hands.

A dermatologist or a hair restoration surgeon can measure your donor density and map the recipient area in one appointment. Those measurements, not a graft number, are what a second opinion should be compared against. Take the eight questions above with you and get the answers in writing.