Plamed Group Logo

Pueraria Extract for Oil Control: Four Assays, Four Numbers, One Question

November 15, 2026Ingredient Insights

A pueraria extract sold on an oil-control claim usually arrives with one number attached to it. Depending on which laboratory ran the work, that number can be 96.71%, 57.08%, 35%, or 20.63% — and none of the four is comparable with the others. Plamed standardises PCR® Supermolecule Pueraria Lobata Extract at ≥15% pueraria flavones, and this note puts all four measurements side by side on one material so that a formulator can see what each one is entitled to say.

Specification

Value

Trade name

PCR® Supermolecule Pueraria Lobata Extract

Marker

Pueraria flavones ≥15% (UV)

INCI

Pueraria Lobata Extract, Hydroxypropyl Cyclodextrin

Composition

Water-soluble powder, typically 30% extract on 70% hydroxypropyl cyclodextrin; declared ranges 10–40% and 60–90%

Recommended use level

0.1–1.0%, leave-on and rinse-off

Solubility

Completely soluble in water at room temperature

Stability

pH 4–6; six months at 45 °C with no colour change

Endpoints available

5α-reductase inhibition; SZ95 sebocyte lipid droplets; ACC gene expression; two 30-person finished-emulsion studies; a 30-person patch test; HET-CAM

Four Numbers, Four Different Measurements

Set the figures in a row and the trap becomes visible. At 1%, the extract inhibits 5α-reductase by 96.71%. In a sebocyte model challenged with dihydrotestosterone, 0.5% reduces lipid-droplet staining by 57.08%. At 1%, it lowers ACC gene expression by 35%. In a 30-person study conducted by Plamed using a finished emulsion containing 1%, porphyrin-covered sebum area on the cheek falls 20.63% over 28 days.

Every one of those is a real result from a real report by Plamed, and every one of them is measuring something different: an enzyme's activity in a tube, a droplet inside a cultured cell, a transcript, and a photograph of a face. They differ in unit, in control, in duration, and in what would have to be true for the number to mean what a reader assumes. Placing them on one scale is the single most common error in this category, and it is made in both directions — by suppliers who quote the highest figure and by buyers who discount the material because the lowest one looks weak.

The Enzyme Assay: The Strongest Number and the Weakest Claim

The 5α-reductase activity assay is one of the most commonly cited indicators for evaluating oil-control potential. Plamed's PCR® Supermolecule Pueraria Lobata Extract has also been tested using this assay, which follows a straightforward procedure. 5α-reductase converts testosterone to dihydrotestosterone, an androgen that drives sebaceous output. The assay measures enzyme activity by monitoring NADPH consumption at 340 nm after a 20-minute incubation at 37 °C. The blank control consists of an untreated enzyme solution.

Two things about that result get misread. The first is the comparison with finasteride: the extract exceeds the control in the same run at the same endpoint, which is a statement about that plate. It is not a potency ranking against a drug, because the two are dosed in different units in a cell-free system and neither has crossed a stratum corneum. The second is what the top of the curve implies. A 96.71% figure describes near-complete inhibition of an enzyme in buffer; it says nothing about how much of the material reaches a sebaceous gland in a formulation, and a cell-free assay has no opinion on delivery.

What the series does establish is direction and dose dependence, over a wider range than most suppliers show. Many pueraria extracts are documented at one concentration. Having 0.1%, 0.5% and 1% on the same plate with the same control makes it possible to see that the response is still climbing at the top of the range rather than sitting on a plateau — which matters when a brief asks for the lowest dose that still moves the target.

The Sebocyte Assay: Where the Curve Flattens

The second system in Plamed's test is closer to skin and less flattering. SZ95 human sebocytes are challenged with 1 μM dihydrotestosterone, stained with Nile red, and quantified by fluorescence; the result is lipid-droplet content relative to the challenged control. Isotretinoic acid at 3 μg/mL is the positive control (57.93%), and PCA zinc runs in the same study (Shanghai Weipu Testing).

Group

Relative lipid-droplet inhibition

Significance vs challenged control

Extract 0.008%

40.78%

p < 0.01

Extract 0.016%

50.66%

p < 0.01

Extract 0.25%

54.90%

p < 0.01

Extract 0.5%

57.08%

p < 0.01

PCA zinc 0.008%

34.75%

p < 0.05

PCA zinc 0.016%

48.20%

p < 0.01

Isotretinoic acid 3 μg/mL

57.93%

p < 0.01

SZ95 sebocytes, 1 μM dihydrotestosterone challenge, Nile red, three wells per group. Percentages calculated against the challenged control, as printed in the report.

The shape of this curve is the useful information. Between 0.016% and 0.5% the response moves from 50.66% to 57.08%. Almost all of the effect is present at the low end, and the top of the range buys a few points rather than a step change. That is not an argument against using 0.5%; it is an argument against paying for 1% on the strength of this assay alone, and it is the reason the human data in the next section are worth reading on their own terms.

PCA zinc is in the same study, and the comparison is more interesting than a straight win. At 0.016% the two materials sit close together — 50.66% against 48.20% — while the MTT screen shows where they diverge. The extract passes at 0.5% with 90.00% cell viability and at 1.0% with 85.06%; PCA zinc falls to 55.22% viability at 0.031%. The limitation on zinc in this model is the width of its usable window rather than its potency at a matched low dose, which is the reason a formulator might pair the two rather than substitute one for the other.

The Gene-Expression Assay: 35% Is a Partial Return

The third measurement sits one level upstream. ACC — acetyl-CoA carboxylase — catalyses the first step of fatty-acid synthesis, and the assay reads its relative expression by qPCR against β-actin in human sebocytes challenged with 18.8 μM dihydrotestosterone (Hunter Biotechnology). At 1%, the extract returns ACC expression to 0.646 ± 0.045 against a challenged control set at 1.000, a down-regulation of 35% (p < 0.001). Isotretinoic acid at 10 μM returns 0.556, or 44%. The unstimulated control in the same run reads 0.542.

Those three numbers together are more honest than the headline. A 35% reduction does not mean the gene is switched off, and it does not mean the model is restored: the treated group sits above the unstimulated control, not at it. What the report supports is a partial normalisation of one transcript in the fatty-acid pathway at 1%, in a system where the positive control only reaches 44% — which is a statement about the model's ceiling as much as about the material.

The distinction matters because ACC expression is not sebum. It is one enzyme's message, in one pathway, in a dish. A supplier who presents 35% as "35% less oil" has crossed from a transcript to a secretion without any measurement in between.

The Human Reading: Two Emulsions, One Dose Apart

The fourth number comes from faces. Two finished oil-control emulsions, one containing 0.5% and one containing 1% of the extract, were tested on 30 volunteers each over 28 days, with readings at 4 and 8 hours and at days 14 and 28. Sebum was read as porphyrin-covered area and spot count by VISIA CR on the cheek; pore area was read from the same images; acne lesion volume came from 3D imaging (PONY Testing).

Endpoint, day 28

0.5% emulsion

1% emulsion

Significance

Lesion volume

−80.90%

−81.63%

p < 0.001 both

Lesion count

−74.27%

−82.68%

p < 0.001 both

Porphyrin-covered sebum area

−10.95%

−20.63%

p < 0.001 / p < 0.01

Porphyrin spot count

−13.30%

−21.76%

p < 0.001 / p < 0.001

Pore area

−15.12%

−20.93%

p < 0.01 / p < 0.01

Thirty valid volunteers per emulsion, cheek, VISIA CR and 3D imaging, baseline to day 28. Change from baseline within each group. The reports test the sebum endpoints against an untreated control area as well; lesion and pore endpoints are tested against baseline.

Read across the rows and the dose response is not uniform. Doubling the extract in the emulsion roughly doubles the sebum-area and spot-count effects, from 10.95% to 20.63% and from 13.30% to 21.76%, and moves pore area by five points. Lesion volume does not move at all — 80.90% against 81.63% — while lesion count moves by eight points. A single "oil control" claim would flatten all of that into one sentence and lose the only part a formulator can act on: the dose buys sebum suppression, and it does not buy more lesion-volume reduction beyond 0.5%.

Two boundaries belong with these figures. The first is that both studies tested finished emulsions, not the raw material, so the reported changes are the performance of a formula containing the extract. Nothing in the reports separates the extract's contribution from that of the rest of the base, and no supplier can make that separation from this data. The second is the early time points. Lesion volume is already down 18–20% at 4 hours in both groups, which is faster than a sebaceous mechanism would suggest and is better read as an instrument-and-surface effect than as a treatment response; the day 14 and day 28 columns are the ones to quote.

What the human studies do not say. They do not establish that the extract alone reduces sebum, that 1% is better than 0.5% for acne, or that either emulsion outperforms a competitor — there is no comparator arm in either report. They establish that two emulsions containing the extract produced these changes in 30 people over 28 days.

Where the Numbers Stop Being Comparable

Four endpoints, four answers to "does it control oil". The table below is the one worth keeping beside a supplier's data sheet, because it shows which comparisons are legitimate and which are arithmetic on incompatible things.

Assay

What is actually measured

Legitimate comparison

5α-reductase inhibition

Enzyme activity in buffer, NADPH consumption at 340 nm

Same method, same plate, same substrate batch

Sebocyte lipid droplets

Nile red fluorescence per well, challenged against unopposed DHT

Same model, same challenge, same staining run

ACC expression

mRNA relative to β-actin, against a challenged control

Same qPCR run and reference gene; not against protein or activity

Sebum imaging

Porphyrin-covered area and spot count on cheek

Same instrument, site and time point; against baseline or an untreated area

Three misuses follow from that table, and all three are common enough to check for in any deck. Quoting 96.71% as "96.71% oil control" converts an enzyme rate into a skin outcome with nothing in between. Subtracting one laboratory's percentage from another's assumes the controls were equivalent, which they are not: an enzyme blank, a DHT-challenged control and a within-subject baseline are three different references. And quoting a 4-hour reading beside a 28-day reading treats a short-term surface change as if it were the treatment effect.

Specification and the Dose That Was Actually Tested

Plamed's PMOil-control® Pueraria (PCR® Supermolecule Pueraria Lobata Extract) is a water-soluble powder carrying 30% pueraria extract on 70% hydroxypropyl cyclodextrin, with declared ranges of 10–40% and 60–90%. The marker is pueraria flavones at ≥15% by UV, and the batch on file reads 16.75%. Recommended use level is 0.1–1.0% for both leave-on and rinse-off, which brackets the concentrations used in every study above: 0.1%, 0.5% and 1% in the enzyme work, 0.5% and 1% in the finished emulsions. Solubility is complete in water at room temperature, the material is stable at pH 4–6, and a six-month study at 45 °C shows no colour change.

Safety documentation is at the same standard as the efficacy work. A 30-person patch test read at 0.5, 24 and 48 hours produced no reactions at any reading, and a HET-CAM test at 3% scored non-irritant. Both were run on the raw material rather than on a finished formula, which is the correct scope for a raw-material safety file and not a substitute for a formulation-specific assessment.

What to Ask For With the Sample

The question is not which pueraria extract has the highest number. It is which number was measured on something that can go into a formula, at a dose the formula can carry. Three documents settle most of it, and they are worth requesting before the sample is weighed.

  1. The full report behind each percentage, including the cytotoxicity screen. Without the MTT result from the same study, a 57% lipid-droplet figure cannot be placed: it may sit well inside the tolerated range, or at its ceiling, and the two mean different things for a formula.

  2. The base formula and the raw-material percentage in the human study. A finished-emulsion result is the performance of a product; the percentage tells you what was in it. A report that gives the outcome without the loading cannot be used to set a dose.

  3. The batch certificate for the marker. A ≥15% flavone specification is a floor, and a floor is not a value. The certificate should show what the delivered lot measured, which is the number a formulator needs when the brief is tight.

Plamed supplies all three with the specification, and the study reports in full rather than as summary sheets, so that the controls and the statistics can be read alongside the results. That is the difference between a number and a claim: the number is only as useful as the conditions printed next to it.

Ingredient Insights - Pueraria Extract | Plamed's Insight