The Longevity Research Evidence Landscape

The evidence ladder

Animals: useful, but not proof for people

The National Institute on Aging runs the Interventions Testing Program, which is designed to identify agents that extend lifespan and healthspan in mice. It tests in genetically varied mice at three sites. [5] In one report, rapamycin fed late in life lengthened median and maximum lifespan in mice at three independent sites. [9] That is a mouse result. A review of cancer research found the average rate of successful translation from animal models to clinical trials was under 8%, a reminder that animal findings often fail to carry over, though that figure comes from cancer and not from aging research. [10] Label for people: preclinical (cells/animals only).

Observational studies: patterns, not proof

Observational studies do not test an intervention. They watch groups of people and compare them over time. [11] A pooled analysis of 94 cohorts with over 30 million participants found higher physical activity was associated with lower risk of death and disease. [6] People who exercise may differ from others in other ways, so the result is an association. Even so, the pattern is consistent, which is why activity is one of the better supported habits.

Trials: the strongest design, with their own limits

Clinical trials test a medical, surgical or behavioral intervention in people. [11] In the CALERIE trial, 218 adults aged 21 to 51 were randomized to aim for 25% calorie restriction or usual eating for two years. The authors reported larger decreases in cardiometabolic risk factors in the restriction group without adverse effects on quality of life, and said the findings suggest potential benefits that need further human studies. [7] The trial measured risk factors, not how long people lived.

A trial can also show no benefit. In ASPREE, 19,114 healthy older adults were randomized to aspirin or placebo. Aspirin did not prolong disability-free survival over about five years and caused more major bleeding. [12]

Small early trials are common in this field. An open-label phase 1 pilot of dasatinib plus quercetin had nine participants and looked at cell markers. [13] Its authors reported that patents on senolytic drugs are held by Mayo Clinic, which is why funding and conflicts of interest are part of reading any study. [13]

Markers versus outcomes

A surrogate endpoint is a trial endpoint used as a substitute for a direct measure of how a patient feels, functions or survives. [14] Many aging studies report markers such as DNA methylation clocks. DunedinPACE, for example, was associated with illness, disability and death in its development study. [15] But a 2024 review said no consensus exists on how biomarkers of aging should be validated before clinical use. [16] A change in a marker is not a change in health. See Surrogate Endpoints and Aging Biomarkers in Trials.

Evidence labels used on this site

• Approved medicine: a regulator has approved it, and the page names the regulator and the approved condition.
• Established: consistent, strong evidence in people.
• Promising but limited: early human data, small, short or marker-based.
• Mixed: studies disagree.
• Insufficient evidence: too little to judge.
• Preclinical (cells/animals only): no human outcome data.
• No good evidence of benefit and evidence of harm: trials found no gain or found damage.
• Marketing claim: a statement from sellers, not a study.

Where we looked and found nothing usable, the page says Proper evidence not found and names what was checked. Fuller notes are on How We Label Evidence in Longevity Research.

Examples of labels in practice