1. Introduction

Ask most patients when their face began to change and the answer is rarely a date. It is a mirror, a photograph, a video call, noticed on an ordinary day and dated only approximately — “the last year or two”, “since I turned forty-odd”. Ask the same question of the biology underneath the face, and a more specific answer is starting to emerge, not because the underlying mechanisms have changed, but because the tools used to study them have improved enough to show that “gradually, over the years” was always an approximation.

The working model in most aesthetic consultations is still implicitly linear: a face ages by a roughly constant amount each year, and treatment is scheduled reactively, when a particular feature becomes bothersome enough to act on. That model is administratively convenient. It is also, on the evidence set out below, a reasonably poor description of what the tissue is actually doing.

This review sets out the current state of that evidence, distinguishes the parts of it that are robust from the parts still being argued over, and translates it into something more useful than a caveat: a practical basis for deciding not only what to treat, but when to look harder, and when treatment done early is doing more than it appears to. It closes with a short worked framework, positioning four patient types at different points on the curve, of the kind a training clinic sees in an ordinary week.

2. The Evidence for Non-Linear Ageing

The idea that biological ageing might not be a straight line is not new. Cross-sectional and longitudinal studies of the blood proteome have described “waves” of change across the lifespan for some years; a 2019 analysis of plasma proteins identified three periods of pronounced change, centred around the ages of 34, 60 and 78. What has changed recently is the scale and resolution of the data available to test the idea.

The most widely reported recent contribution is a 2024 longitudinal study from Stanford, which tracked more than 135,000 molecular and microbial features across 108 adults aged 25 to 75, sampled every three to six months. Its central, more basic finding is on solid ground: using standard linear-regression and correlation methods, only 6.6% of the markers studied changed in a simple straight-line fashion with age. The great majority — 81.0% — changed non-linearly. That distinction alone is a reasonable basis for treating “ageing is gradual” as an oversimplification rather than a fact.

The more specific, more widely quoted claim from the same study — that molecular change clusters sharply around two ages, approximately 44 and 60 — rests on a more technically demanding sliding-window analysis, and it is this specific analysis that readers should treat with particular care. As of July 2026, the journal has attached an editor's note to the paper stating that, following review by the authors and independent researchers, the reliability of the peak-detection analyses used to identify those two ages is currently in question, with corrections pending. Independent commentators had already raised a related concern before that notice: with a median follow-up of only 1.7 years per participant, the design is arguably better suited to comparing groups of different ages than to demonstrating that any single person's biology accelerates at 44 or 60.

None of this means the underlying phenomenon is wrong — the earlier 2019 dataset, using different proteins, found waves at different but nearby ages, and a 2026 cross-disciplinary review concluded that the overall case for non-linear ageing, gathered across genomic, epigenetic and physiological evidence, is now substantial. It does mean that the precise ages 44 and 60 should currently be treated as a provisional finding under review rather than an established clinical fact, and clinicians citing this paper to patients or in teaching should say so.

For aesthetic practice, a separate and considerably older body of evidence is more directly useful, because it is tissue-specific, mechanistically clear, and not implicated in the dispute above. Dermal collagen content declines at a fairly steady rate of roughly 1–2% per year from the mid-twenties. In women, that decline is sharply accelerated by the loss of oestrogen around the menopausal transition: a foundational 1983 study found that postmenopausal women on hormone replacement carried substantially more skin collagen than untreated women of the same chronological age, and later work quantified the untreated decline at around 2% of collagen content per postmenopausal year over roughly fifteen years. More recent reviews describe a “collagen cliff” specifically in the first five years after the final period, with losses in the region of 30% over that window before the decline settles to its longer-term rate. Facial bone density and skin thickness appear to follow a broadly similar accelerated pattern across the same years.

Taken together, the two strands point in the same direction without depending on each other: general biological ageing appears to cluster rather than progress smoothly, and — independently of that — women experience a specific, hormonally driven, well-evidenced acceleration in the tissue most relevant to facial appearance during the menopausal transition. The staircase, not the ramp, is currently the better working picture.

3. What Injectable and Aesthetic Treatments Actually Change

It is worth being precise about what follows from this, because the temptation in aesthetic medicine is to reach for the stronger claim. Injectable and energy-based treatments do not alter the biological drivers described above. They do not slow the molecular shifts of early middle age, whatever their eventual resolution turns out to be, and they certainly do not alter the trajectory of circulating oestrogen at menopause. Nothing in a syringe changes hormonal physiology. Any claim to the contrary belongs with the other overclaims already catalogued in the aesthetic advertising literature, not in a clinical framework.

What treatment can change is narrower, but genuinely useful, and worth separating into two categories that are frequently and unhelpfully treated as one.

The first category is corrective: treatments that address the visible consequence of tissue change without altering the tissue's underlying trajectory. Volume replacement in a deflated cheek, support along a descended jawline, and botulinum toxin reducing the mechanical force of a hyperactive muscle all sit here. They are legitimate, often the correct clinical choice, and entirely honest — provided they are presented as what they are: management of a visible consequence, not a change to the process producing it.

The second category is genuinely biostimulatory, and this is where the more interesting clinical argument sits. Controlled resurfacing (microneedling, selected peels), autologous and injectable skin-quality treatments (platelet-rich plasma, polynucleotides, hyaluronic acid–amino acid combinations), and biostimulatory fillers (poly-L-lactic acid, calcium hydroxylapatite) have evidence for genuinely increasing local fibroblast activity and collagen synthesis, not merely disguising its absence. This is a real, local, measurable effect. It is also a modest one relative to the systemic forces described in Section 2, and the evidence base — while growing — remains product- and protocol-specific rather than settled.

The honest clinical argument, then, is not that these treatments “flatten the curve”. It is that they can raise the tissue's starting position before a known acceleration window arrives. A dermis carrying a better collagen and hydration reserve going into the menopausal transition — because it has spent several years on a foundation of SPF, a tolerated retinoid, and appropriately timed biostimulatory treatment — will very plausibly show less of the resulting decline than a dermis that enters that transition already depleted by accumulated photodamage and years of untreated neglect. That is a claim about starting position and accumulated reserve, not about altering the mechanism, and it is the claim this review is willing to defend.

There is a second, more immediate clinical implication. A treatment begun just as a patient enters a steep segment of the curve can appear to underperform relative to the same treatment given during a plateau year, simply because the tissue is losing ground at the same time treatment is trying to add it. Patients are not, as a rule, told this in advance. They should be: a result that looks like it “faded fast” during a known acceleration window is not necessarily evidence that the treatment, or the injector, failed.

4. Four Patients, Four Points on the Curve

The following four vignettes are the kind familiar to any Harley Street practice; the first is illustrated with one documented case, reported in full elsewhere.

4.1 The Plateau Patient (twenties to mid-thirties). This patient sits, biologically, on the flat part of the staircase. Intrinsic ageing has barely begun to show; the dominant risk to her future face is accumulated extrinsic damage — ultraviolet exposure, unmanaged pigment, inconsistent care — rather than any structural loss. A documented case from this practice illustrates the point precisely: a woman in her early thirties, previously treated at several clinics with unsequenced botulinum toxin and filler, presenting with generalised loss of skin vitality rather than any single correctable feature.

The clinical priority at this stage is almost entirely foundational: broad-spectrum SPF, a tolerated retinoid, barrier support, and — where indicated — light biostimulatory treatment rather than structural correction. The temptation to treat this patient reactively, addressing whichever feature she raises in isolation, should be resisted; it was precisely that pattern, repeated across several previous providers, that produced her presenting complaint. The correct question for a plateau patient is not “what would you like changed” but “what is the least we need to do now, done consistently, to arrive at the next acceleration window in the best possible condition.” Note the asymmetry: doing too little here is the common failure, not doing too much.

4.2 The Approaching Patient (early-to-mid forties). This patient is entering, or about to enter, the first widely reported window of accelerated change — whatever its precise boundaries turn out to be once the current review of the underlying dataset concludes. Clinically, she often presents with vague, hard-to-localise concerns: a sense that something has changed without a clear feature to point to, tiredness that photographs worse than it feels, early static change in previously dynamic lines.

This is the point at which consultation cadence deserves to change, independent of what the patient has booked in for. A full five-domain read — skin, movement, volume and support, laxity and descent, proportion — is more valuable here than at any point since her twenties, because early identification of which domain is moving first allows treatment to stay ahead of visible change rather than correct it after the fact. Foundational skin care, if not already well established, should be treated with more urgency than it was five years earlier; biostimulatory treatment earns a stronger case here than in the plateau years. Structural correction, by contrast, is usually still premature — the risk at this stage is treating an early skin-quality signal as though it were a volume problem.

4.3 The Accelerating Patient (perimenopause to established menopause, roughly mid-forties to mid-fifties). This is the patient for whom the evidence in Section 2 is least contested and most actionable. The mechanism is specific — falling oestrogen, reduced fibroblast collagen output, increased collagenolytic activity — and the timeframe is comparatively well defined: the sharpest phase of dermal collagen loss in the published literature sits in the first five years following the final menstrual period. Bone density and skin thickness are declining across a similar window, which is part of why facial support and skin quality often deteriorate together at this stage rather than in the sequence a linear model would predict.

Three things distinguish good practice here. First, visit frequency: a patient in this window benefits from closer-spaced review than the standard interval, because the rate of change within the window is itself elevated and a plan set six months ago may already be out of date. Second, the honest expectation-setting described in Section 3 — that a given treatment may appear to hold less well here than it did five years earlier, through no fault of the treatment. Third, and most distinctively, this is the stage at which the correct clinical action may sit outside the room entirely. Because the dominant driver is systemic hormonal change, a conversation about menopause care with the patient's GP or a menopause specialist is a legitimate and sometimes the single most effective intervention available — hormone replacement therapy has been directly shown, in the foundational literature on this subject, to preserve measurably more dermal collagen than no treatment. An aesthetic practice with no comfortable way of raising this is, for a meaningful proportion of patients in this window, withholding the most effective lever it has access to.

4.4 The Second-Wave Patient (early-to-mid sixties). Whatever the eventual resolution of the specific age attached to it, a second period of broadly reported biological change centres on the early sixties, and clinically this tends to present differently from the perimenopausal window: less about skin quality in isolation, more about laxity, descent, and the visible consequences of several decades of accumulated bone remodelling and fat repositioning finally reaching a threshold.

The central clinical judgement at this stage is the one addressed at length elsewhere in this series: distinguishing what a non-surgical treatment can honestly achieve from what has become surgical territory, and saying so plainly rather than offering an escalating series of non-surgical treatments as a substitute for that conversation. Energy-based remodelling and continued biostimulatory support remain legitimate tools for skin quality and modest laxity; they are not a substitute for surgical repositioning where descent is established, and presenting them as one is the least defensible version of the camouflage-versus-biostimulation distinction drawn in Section 3. This is also the stage at which the population-level caveat matters most in practice: patients arriving at their early sixties carry markedly different amounts of accumulated sun exposure, smoking history and prior treatment, and the range of what is normal at this stage is wide.

5. Clinical Implications: Reading the Curve in Practice

Three practical changes follow from the material above, none of which require new equipment or a change to what is actually injected.

The first is consultation cadence. A patient plan made once and topped up indefinitely — treating whatever is loudest at each visit, the pattern already identified in this series as the central failure of unplanned aesthetic care — will always eventually be describing a face that has moved on. Scheduling a full five-domain reassessment at the approach of each known window, rather than only when the patient raises a new concern, is a defensible clinical policy on the evidence set out here, not a marketing device.

The second is where foundational treatment sits in the sequence. Broad-spectrum SPF and a tolerated retinoid are usually presented as the entry level of the treatment pyramid — true, but potentially undersold by that framing. On the evidence above, foundational care is most valuable precisely when it looks least necessary: in the plateau years and the approach to a known acceleration window, not as a rescue measure once decline is already visible.

The third is a matter of honesty in both directions. Practitioners should not claim that any injectable or energy-based treatment alters the biological drivers of ageing; the correct claim is narrower — that treatment changes what is visible, and that well-timed biostimulatory treatment can improve a patient's tissue reserve ahead of a window in which reserve matters more than usual. Equally, practitioners should not treat every consultation as an opportunity to find something to correct. A patient standing on a flat section of the curve, as several will be in any given week, is very often best served by very little.

Two limitations deserve stating plainly. Everything in Section 2 is population-level evidence; individual variation in genetics, sun exposure, smoking and general health is substantial, and the curve is a reason to look more carefully at certain ages, not a forecast for any specific patient. And a meaningful part of the general-biology evidence — specifically, the precise ages attached to the first acceleration window — is, at the time of writing, under active editorial review. Clinicians using this material in patient-facing education should represent that honestly, citing the menopause-specific evidence, which is not in dispute, as the firmer ground.

6. Conclusion

The central argument of this review is not that aesthetic medicine has been getting the treatments wrong. It is that treatment timing has been under-examined relative to treatment selection, and that a growing, if unevenly settled, body of evidence gives clinicians a more specific basis for that timing than “come back if something bothers you”.

Facial reading — skin, movement, volume and support, laxity and descent, proportion — has always been the discipline this journal argues for. What the ageing curve adds is a fifth dimension to that reading: not only what is happening to a given face, but where, on a trajectory now reasonably well understood to be uneven, that face currently sits. Understanding the face, understanding the person and understanding the problem was never only a question of anatomy. Increasingly, it is also a question of timing.