Purpose and status of this register
This is Longevity Magazine's standing register of evidence grades for claims about ageing, healthspan, mortality, function and related biomarkers. It is intended as a citation point for journalists, librarians, clinicians and readers who need to distinguish a claim's current evidential status from its public profile. It should be read with the individual review named in each register entry, which contains the claim definition, the evidence considered and the reasons for the grade.
A grade belongs to a specific claim, not to an intervention, test or behaviour in all circumstances. For example, evidence that an exposure is associated with lower mortality does not establish that deliberately changing that exposure will reduce mortality. Evidence for a medicine in people with a diagnosed condition does not automatically establish an anti-ageing effect in people without that condition. These distinctions determine both the wording of a claim and its grade.
The register was assessed on 19 November 2026. "Last looked at" means the date on which this register last considered the grade against the journal's published appraisal. It does not mean that every underlying study was published on that date, nor that the claim will retain the same grade indefinitely. A material new trial, systematic review, safety signal or correction can prompt reassessment.
The grades are evidence judgements, not treatment advice. They do not replace discussion with an appropriate clinician, particularly where a claim concerns prescribed medicines, chronic disease, frailty, pregnancy, eating disorders, disability or a substantial change to diet or exercise.
The A to D grading scale
The scale asks whether the available evidence supports the exact claim being made in people relevant to that claim. It gives greatest weight to outcomes that matter directly to people, such as mortality, disease incidence, physical function, cognition or quality of life. Biomarkers can contribute, but they normally cannot carry a claim about longer or healthier life on their own.
| Grade | What it means here | Evidence normally needed |
|---|---|---|
| A | Strong support for a narrowly defined claim. | More than one well-conducted trial, or comparably decisive evidence where trials are not feasible, with relevant populations and meaningful clinical outcomes. |
| B | Supportive evidence with important limits. | At least one persuasive line of human evidence, usually including trials or robust prospective evidence, but with limits in replication, duration, population or outcomes. |
| C | Suggestive or indirect evidence. | Early human trials, observational associations, intermediate outcomes, or findings that do not yet establish the proposed healthspan claim. |
| D | Insufficient support for the stated claim. | Little relevant human evidence, conflicting findings, serious design limitations, or evidence restricted to models, mechanisms or surrogate measures. |
A higher grade does not mean an intervention is suitable, safe or appropriate for every person. A lower grade does not prove that a claim is false. It records that the current evidence does not justify stronger wording. No grade in this register is a recommendation to begin, stop or alter treatment.
What moves a claim up or down
A claim may move upward when new evidence tests the same intervention, dose or exposure in a population close to the people for whom the claim is being made. The most informative studies prespecify outcomes, retain participants adequately, report harms as well as benefits and make it possible to compare the intervention with a credible alternative. Replication by independent research groups matters because a single positive finding can be unstable.
Outcome choice is central. A finding that changes a blood measure, imaging value or composite biological-age score may be biologically interesting, but it does not by itself demonstrate fewer fractures, preserved independence, lower dementia incidence or longer life. A grade can therefore remain C or D even where a mechanism is plausible and a biomarker changes in the expected direction.
Claims move downward, or fail to move upward, when the evidence relies on uncontrolled before-and-after results, selective follow-up, short duration for a long-term claim, unrepresentative participants or outcomes substituted after results are known. Confounding is particularly important in observational longevity research: people who train, sleep well or maintain a certain weight can differ in income, illness, healthcare access, smoking, disability and many other factors that also affect later outcomes.
The following decision rule is used when wording a conclusion: if the evidence directly measures the claimed outcome in a relevant human population and has meaningful replication, consider A or B; if it chiefly measures a proxy, association or mechanism, do not grade above C without additional direct evidence. Safety uncertainty can also limit confidence, especially for medicines used outside their established clinical indication.
Current register of graded claims
The entries below record the journal's current grades as at 19 November 2026. Review titles are included so that the reasoning can be located in the corresponding Longevity Magazine appraisal. The grade applies only to the claim stated in the first column. It should not be transferred to broader claims about "anti-ageing", disease prevention or life extension.
| Claim assessed | Current grade | Reason in brief | Review record | Last looked at |
|---|---|---|---|---|
| Resistance training is associated with healthier ageing and lower mortality risk. | B | Human evidence supports important functional and risk associations, while mortality causation remains difficult to isolate. | Resistance training and mortality | 19 November 2026 |
| Higher grip strength predicts lower mortality and better later-life outcomes. | B | It is a useful prognostic marker, but improving grip strength itself is not proven to produce the same outcome change. | Grip strength: healthy ageing evidence graded | 19 November 2026 |
| Higher cardiorespiratory fitness is a marker of lower mortality risk. | B | The prognostic evidence is substantial; marker status is not equivalent to proof of a single life-extending intervention. | VO2 max as a longevity marker | 19 November 2026 |
| Calorie restriction extends healthy human life. | C | Human findings include relevant metabolic and risk-factor data, but direct long-term healthspan or lifespan evidence is limited. | Calorie restriction in humans | 19 November 2026 |
| Time-restricted eating extends healthspan. | C | Human studies are often short and rely heavily on intermediate outcomes, with uncertainty about durability and comparison diets. | Time-restricted eating | 19 November 2026 |
| Sleep architecture can be modified to slow ageing. | C | Associations between sleep measures and later outcomes do not yet establish that changing architecture alters ageing trajectories. | Sleep architecture and ageing | 19 November 2026 |
| GLP-1 receptor agonists improve healthspan in people generally. | C | Clinical benefits in defined medical settings cannot yet be treated as a general healthspan claim. | GLP-1 receptor agonists and healthspan | 19 November 2026 |
| Metformin extends healthspan in non-diabetic adults. | D | The proposed broad preventive claim lacks sufficient direct outcome evidence in the stated population. | Metformin: healthy ageing trials in non-diabetic adults | 19 November 2026 |
| Rapamycin-based treatment extends healthy human life. | D | Mechanistic rationale and non-human findings do not establish the claimed human outcome. | Rapamycin and mTOR | 19 November 2026 |
| Senolytics improve human healthspan. | D | Human evidence is not yet sufficient for the broad claim. | Senolytics: current state | 19 November 2026 |
| NAD precursors extend human healthspan. | D | Changes in biochemical measures do not establish clinically meaningful ageing outcomes. | NAD precursors: NMN and NR | 19 November 2026 |
No claim currently listed holds an A grade. That is a statement about this register's threshold for direct, replicated human outcome evidence, not a finding that no potentially useful ageing-related intervention exists.
How to cite and interpret a grade
When citing an entry, state the claim, the grade and the register date together. A usable formulation is: "Longevity Magazine graded the claim that higher grip strength predicts lower mortality and better later-life outcomes as B, last assessed on 19 November 2026." This preserves the distinction between prediction and intervention. It is not accurate to shorten that entry to "grip training reduces mortality".
For claims involving treatment, diagnosis or disease management, the grade should never be used as a substitute for clinical guidance. The register does not set prescribing standards, assess an individual's contraindications or calculate personal benefit and harm. It also does not certify products, validate commercial testing services or endorse a practitioner.
Where a review uses language such as "associated with", "predicts", "changes a marker" or "has not yet shown", those words are part of the conclusion rather than stylistic caution. Replacing them with "prevents", "reverses" or "extends life" changes the proposition being evaluated. A citation should therefore retain the original level of certainty and the population limits.
Grades are not averaged across claims. A medicine can have strong evidence for one diagnosed condition and a D grade for a distinct claim about healthspan in people without that condition. Likewise, a biomarker can be useful for risk assessment while remaining unsuitable as proof that an intervention has delivered durable clinical benefit.
Review process and revision triggers
Each register entry begins by defining the smallest claim that can be tested. The appraisal then separates evidence by design: randomised trials, prospective observational studies, cross-sectional analyses, mechanistic work and non-human research. It checks whether the population, exposure, comparator and outcome match the wording of the claim. This prevents a result in a clinical subgroup, or a change in a laboratory measure, from being silently generalised to ageing as a whole.
Revision is considered when there is a completed human trial with a relevant clinical outcome, a substantial synthesis that changes the balance of evidence, credible evidence of harm, a correction affecting a key study or a meaningful change in how the claim is framed. A new paper alone does not automatically change a grade. Its design, completeness, outcome relevance and consistency with the existing record matter.
The register is deliberately conservative about broad claims such as "reverses ageing", "adds years" or "extends healthspan". Such phrases encompass multiple outcomes over long periods. A short study of a proxy cannot answer them. Conversely, a carefully defined claim can be evaluated more clearly and may earn a higher grade than a vague umbrella claim based on the same general area of research.
Grades reflect evidence available at the assessment date, not a permanent verdict. Readers using an older citation should check the date in the register before treating it as current.
Limits of this register
This register does not cover every factor relevant to ageing, every disease-prevention intervention or every outcome that matters in later life. It is not a directory of services, a catalogue of products, a screening guide or a personal risk calculator. It does not determine whether a particular person should take a medicine, adopt a dietary pattern, change exercise or undergo a test.
It also does not grade the quality of every individual study in isolation. A grade is a synthesis judgement about a defined claim and an overall body of evidence. Important evidence may exist outside the entries listed here, especially in populations with specific illnesses, disabilities or treatment needs. Those findings may not apply to a general healthspan claim.
Absence from the register is not a grade and should not be read as evidence for or against a claim. D means insufficient support for the wording assessed, not proof of no effect. A and B do not settle questions of feasibility, equity, cost, long-term adherence, personal values or net benefit for an individual. Those questions require context that a register cannot supply.