Negative Longevity Trials: How to Read a Missed Endpoint Without Losing the Evidence
Missed primary endpoints constrain claims even when biomarkers or subgroups look promising. Three examples show why endpoint hierarchy and study context matter.
- Published
- Last updated
- Last reviewed
Selected null and negative studies, not a systematic estimate of the success rate of geroscience.
A negative trial is not an empty result. It tells readers which claim was not demonstrated under a particular design. The first task is to identify the prespecified question. The second is to keep exploratory explanations from replacing its answer. That discipline is useful whether the intervention eventually succeeds elsewhere or is discontinued. Sources: Effects of intermittent senolytic therapy on bone metabolism in postmenopausal women: a phase 2 randomized controlled trial.; Targeting the biology of ageing with mTOR inhibitors to improve immune function in older adults: phase 2b and phase 3 randomised trials.; Effect of Vitamin D Supplementation, Omega-3 Fatty Acid Supplementation, or a Strength-Training Exercise Program on Clinical Outcomes in Older Adults: The DO-HEALTH Randomized Clinical Trial..
A missed primary endpoint stays missed
In the randomized D+Q bone study of 60 postmenopausal women, the primary CTx endpoint at 20 weeks was not met: the between-group comparison had P = 0.611. Secondary and exploratory findings, including subgroup analyses, do not reverse that primary result. They can motivate another trial, with a newly prespecified hypothesis. Sources: Effects of intermittent senolytic therapy on bone metabolism in postmenopausal women: a phase 2 randomized controlled trial..
The useful reading is narrower than either “senolytics work” or “senolytics are disproved.” The study tested a particular regimen, population, time point and endpoint. A proposed responder subgroup needs prospective testing rather than retrospective promotion into the trial’s principal conclusion. Sources: Effects of intermittent senolytic therapy on bone metabolism in postmenopausal women: a phase 2 randomized controlled trial..
Changing the endpoint changes the test
The RTB101 report included Phase 2b and Phase 3 trials. The Phase 3 symptomatic respiratory-illness endpoint was not improved: odds ratio 1.07, 90% confidence interval 0.80–1.42, P = 0.65. The earlier programme used laboratory-confirmed infection measures, and populations and analyses also differed. The later null result cannot be concealed behind the earlier positive signal. Sources: Targeting the biology of ageing with mTOR inhibitors to improve immune function in older adults: phase 2b and phase 3 randomised trials..
Nor should the result be described as a perfect replication failure without naming those differences. An endpoint measuring reported illness and one requiring laboratory confirmation select different events. That does not excuse a failed trial; it identifies what a follow-up experiment would have to resolve. Sources: Targeting the biology of ageing with mTOR inhibitors to improve immune function in older adults: phase 2b and phase 3 randomised trials..
A biomarker analysis does not replace the parent trial
DO-HEALTH’s parent randomized trial enrolled 2,157 adults aged 70 or older and did not establish benefits on its six primary outcomes under its prespecified significance threshold. A later post hoc methylation analysis in 777 participants reported small clock changes. These are different analyses with different outcomes; the clock finding is not a reversal of the parent trial’s clinical results. Sources: Effect of Vitamin D Supplementation, Omega-3 Fatty Acid Supplementation, or a Strength-Training Exercise Program on Clinical Outcomes in Older Adults: The DO-HEALTH Randomized Clinical Trial.; Individual and additive effects of vitamin D, omega-3 and exercise on DNA methylation clocks of biological aging in older adults from the DO-HEALTH trial..
| Question | Record explicitly |
|---|---|
| What failed? | Primary endpoint, comparison, follow-up, estimate and uncertainty. |
| What looked favorable? | Secondary, subgroup or exploratory status; multiplicity and prespecification. |
| What changed between trials? | Population, endpoint definition, exposure, comparator and analysis. |
| What is still unknown? | Whether a new, testable explanation survives a prospective study. |
Sources: Effects of intermittent senolytic therapy on bone metabolism in postmenopausal women: a phase 2 randomized controlled trial.; Targeting the biology of ageing with mTOR inhibitors to improve immune function in older adults: phase 2b and phase 3 randomised trials.; Effect of Vitamin D Supplementation, Omega-3 Fatty Acid Supplementation, or a Strength-Training Exercise Program on Clinical Outcomes in Older Adults: The DO-HEALTH Randomized Clinical Trial..
Null is not the same as equivalent or harmless
Failure to show superiority is not proof of equivalence. A small or imprecise trial may leave important benefits and harms unresolved. Conversely, a clinically meaningful effect cannot be claimed merely because the confidence interval includes one. Interpretation should report what the interval permits and what the design was powered to detect, rather than translating “not significant” into a categorical biological verdict. Sources: Effects of intermittent senolytic therapy on bone metabolism in postmenopausal women: a phase 2 randomized controlled trial.; Targeting the biology of ageing with mTOR inhibitors to improve immune function in older adults: phase 2b and phase 3 randomised trials..
Safety requires its own reading. Absence of a serious event in a small study does not establish safety for broader use. Likewise, a biomarker response does not show that the benefit-risk balance is favorable. These are reasons to preserve the complete result, not reasons to dismiss experimentation. Sources: Effects of intermittent senolytic therapy on bone metabolism in postmenopausal women: a phase 2 randomized controlled trial.; FDA facts: biomarkers and surrogate endpoints.
When should the conclusion change?
Update the assessment when a preregistered replication tests the proposed subgroup, a correction changes the primary analysis, complete results resolve missing outcomes, or a better-powered trial answers the same clinical question. Another press release or a favorable mechanistic experiment is not, by itself, that update. The examples here are selected teaching cases, not a field-wide failure-rate calculation. Sources: Effects of intermittent senolytic therapy on bone metabolism in postmenopausal women: a phase 2 randomized controlled trial.; Targeting the biology of ageing with mTOR inhibitors to improve immune function in older adults: phase 2b and phase 3 randomised trials.; Effect of Vitamin D Supplementation, Omega-3 Fatty Acid Supplementation, or a Strength-Training Exercise Program on Clinical Outcomes in Older Adults: The DO-HEALTH Randomized Clinical Trial..
Related intelligence
Senolytics Pipeline Tracker: Human Evidence, Trial Status, and Translational Limits; What DO-HEALTH Actually Tested: Omega-3, Biomarker Aging, and Endpoint Limits; Longevity Trial Endpoints: Choosing Between Clinical Outcomes, Function and Biomarkers.