The vehicle is a dose

This register has spent forty-five reviews asking companies for concentrations, on the reasonable grounds that you cannot assess a formula you cannot read. Adding four approved tretinoin and tazarotene products has forced a correction: the concentration is necessary and it is not sufficient, and the gap between those two statements is where most of the actual pharmacology lives.

Tretinoin appears in this register in four manufactured vehicles, all approved, all with published trials.

Read those as concentrations alone and Renova is the weakest by a factor of two and a half. Read them as products and Renova is the one indicated for indefinite nightly use on ageing facial skin, precisely because the low strength and the emollient base were chosen together to make that possible.

The number that makes the case

Altreno is full-strength tretinoin at 0.05% — the concentration that built the drug's reputation for peeling. In two trials totalling 1,640 subjects, application-site dryness was reported by 4% of users and stinging or burning by 3%.

Ninety-six per cent of people using full-strength tretinoin did not report dryness. The variable that changed was not the dose.

The base carries sodium hyaluronate, soluble collagen, glycerin and mineral oil. None of that is there to dissolve the drug. It is there because the limiting factor on topical retinoids has never been potency — it has been how many people stop in week three.

Retin-A Micro makes the same argument by a different route. The tretinoin sits inside a porous methyl methacrylate crosspolymer that releases it gradually rather than delivering it all at once, and the label publishes what that achieves systemically: bioavailability of 0.82% after a single application and 1.41% with repeated daily use, with plasma tretinoin essentially unchanged from baseline. A vehicle that meters release is doing dose control that no instruction can.

Two compounders have understood this

The point is not that manufactured always beats compounded. Two compounded products in this register have made better vehicle decisions than several approved ones.

• ▸Alloy's rosacea cream is anhydrous — silicones, alkanes and esters, no water phase — which means no preservative is required and none is present, on skin that is more likely than most to react to preservatives. It also carries two genuine penetration enhancers appropriate to driving azelaic acid from a waterless base.
• ▸Dear Brightly's Night Shift is a silicone elastomer base, also anhydrous, also preservative-free, published down to the trade names of the raw materials. Silicone elastomers spread thin and release slowly, which is a reasonable design for a first-time retinoid user.
• ▸Skin + Me publishes two base creams to full INCI and declares the correct antioxidant systems for each active: BHT with tretinoin, ascorbic acid and sodium metabisulfite with hydroquinone.

All three are formulation decisions of real quality. What none of them has is a trial showing what the vehicle achieved — which is exactly what Altreno's 4% dryness figure is, and what makes it citable.

What this changes about reading a product page

A personalised tretinoin at 0.018% tells you almost nothing on its own. Dissolved in an alcohol-heavy gel it may be more irritating than 0.05% in a hyaluronate lotion. Held in a slow-release reservoir it may deliver less drug per hour and more drug per week. Suspended in an anhydrous silicone base it may sit on the surface and release slowly, or may drive deeper if penetration enhancers are present — and you cannot tell which without the excipient list.

So the question to put alongside 'what strength' is 'what is it in'. In this register, the answer is published for every approved product and for three compounded ones. Everywhere else, the vehicle is the part of the formula nobody has been asked about.

The scoring consequence

Formulation integrity has always been a criterion here, weighted at 20%. From this point it explicitly rewards a published vehicle and evidence that the vehicle was designed rather than defaulted to — and it will keep marking down a published concentration that arrives with no excipient list, because half a specification is not a specification.