How to read an announcement in this field
Longevity research produces more announcements than results, and the gap between the two is where most public misunderstanding lives. A useful discipline before reading any claim in this area is to establish four things: what species the work was done in, what was measured, over what period, and whether anything was randomised.
Those four questions dispose of a large share of what circulates. Work in short-lived model organisms tells you about mechanism and almost nothing about human outcome. Work measuring a marker rather than a health outcome tells you the marker moved. Work over weeks tells you nothing about a process measured in decades. The method set out in how to read a clinical trial applies here without modification.
What follows is a survey of five fields commonly described as transformative, each placed against that standard. None of this is a prediction about what will eventually work. It is a description of where the evidence currently is, which is a more useful thing for a reader to hold.
Senolytics and the clearance of senescent cells
Senescent cells stop dividing but remain metabolically active, and they secrete a mixture of inflammatory signals that appears to contribute to tissue dysfunction. Removing them in animal models has produced striking improvements in function and, in some models, in lifespan. The mechanism is coherent and the animal work has been replicated by multiple groups.
Human evidence is at a much earlier stage. The trials that exist are small, short, and mostly directed at specific disease indications rather than at ageing itself, with outcomes that are frequently functional or marker-based. That is a normal place for a field to be a decade into serious investigation, and it is not a criticism. It is simply not the same thing as a treatment.
Our review of senolytics sets out the grade this journal applies and what would change it. The short version is that animal evidence, however strong and however well replicated, does not by itself establish human benefit, and the translation record across medicine as a whole is poor enough to justify that caution.
Partial reprogramming, the most consequential and least ready
Cellular reprogramming is the technique of driving a differentiated cell back towards a stem-like state using a defined set of transcription factors. Applied partially and transiently, it appears to restore some youthful characteristics to cells and tissues in animal models without erasing cell identity, which is why it attracts more attention than any other line of work in the field.
It is also the furthest from a clinic. The central difficulty is control: the same process that rejuvenates a cell can, taken too far, produce a cell that behaves like a tumour. Delivering the intervention to the right tissues, for the right duration, with a means of stopping it, is an unsolved engineering problem as much as a biological one.
The accurate position is that this is a field with a genuinely new idea in it, being pursued seriously and funded heavily, with no human efficacy evidence for an ageing indication. A reader encountering a clinic offering anything described as reprogramming should understand that the term is being used loosely.
Regenerative injectables and what is already being sold
Unlike the two fields above, regenerative injectables are already in commercial use. Polynucleotides, biostimulatory materials, platelet preparations and growth factor products are offered in clinics across the United Kingdom, usually for skin quality, scarring or hair.
The evidence pattern is consistent across the category. Mechanistic plausibility is reasonable. Human studies exist but tend to be small, short, single-arm or inadequately controlled, and to report outcomes measured on instruments or on rater scores rather than on anything a patient would independently notice a year later. That places most of the category in the same evidential position this journal describes for any surrogate-only literature, and our explainer on surrogate endpoints explains why volume of publication does not fix it.
The regulatory question is separate and worth asking directly. Products in this category may fall under medicines, medical device or cosmetic rules depending on composition and claimed action, and the category has attracted regulatory attention in the United Kingdom. A clinic should be able to state which framework a product sits in and what it is certified or licensed for.
Exosome-derived preparations, the clearest example of the gap
Exosomes are small extracellular vesicles that carry proteins and nucleic acids between cells, and they are one of the mechanisms by which cells signal to each other. As a research subject they are legitimate and active, in wound healing, in immunology and in oncology.
As a clinic offering, the distance between the underlying science and the marketing is among the widest in aesthetic medicine. Preparations vary enormously in source, characterisation and processing, standardisation across suppliers is limited, and the human clinical literature for cosmetic indications is short, small and heavily surrogate-based. The regulatory position for injectable use differs by jurisdiction and has been the subject of specific warnings in several countries.
None of this means the biology is empty. It means that a patient being offered such a treatment is contributing to an evidence base rather than benefiting from one, and is entitled to be told so. Clinics that work in this area and describe it accurately are more useful to a reader than those that do not, and London practices operating in regenerative aesthetics, among them Mesglo London, tend to present these preparations as an emerging category rather than an established one. That framing is the minimum a patient should expect wherever the treatment is offered.
Machine learning, which changed measurement rather than treatment
The clearest practical effect of machine learning on ageing research so far has been on measurement. Models trained on molecular data return estimates of biological age, and similar approaches are applied to retinal images, movement data and routine clinical records. This is a real methodological advance and it has made large-scale ageing research tractable in ways it was not before.
What it has not done is produce a treatment, and the distinction is regularly blurred in commercial settings. A model output is an estimate of how closely a sample resembles a training set. Different models disagree on the same sample. No evidence establishes that acting to lower the number produced by any of them improves an individual outcome. Our explainer on biological age tests sets out where the chain of inference stops.
What is available now, and what it costs
Set against all of the above, the interventions with the strongest outcome evidence for healthspan are unchanged and unexciting: cardiorespiratory fitness, muscle mass maintained through resistance training, sleep, blood pressure and metabolic risk managed when they appear, tobacco avoided. Our reviews of VO2 max and of NAD precursors sit at opposite ends of our scale for exactly this reason, and the difference between them is the difference between outcome evidence and mechanistic enthusiasm.
For readers weighing what is currently sold, indicative UK ranges for 2026 give a sense of scale. Regenerative injectable treatments are commonly advertised between £300 and £800 per session, often in courses. Exosome-derived treatments frequently sit between £400 and £1,000 per session. Biological age testing is typically £200 to £500 per test, and comprehensive longevity assessments are commonly advertised from the low thousands upwards.
The reasonable position is neither dismissal nor enthusiasm. Several of these fields may produce something important within a decade. None of them has produced it yet, and a treatment sold today on the strength of what a field might eventually demonstrate is being sold on a promise rather than on a result.