The clinical and biomarker results differed
A blinded randomized trial studied NR in adults over fifty-five with subjective cognitive decline or mild cognitive impairment. Forty-six were randomized and thirty-seven completed the trial. Eight weeks of active treatment were assessed using a cognitive battery as the primary efficacy outcome and several plasma biomarkers as secondary outcomes. [1]
Between-group results showed lower pTau217 with NR, but no significant difference in the primary cognitive battery, other measured plasma biomarkers, digital cognitive scores or step counts. A within-individual analysis found a step-count difference. The authors called for larger and longer studies rather than concluding that NR improved cognition. [1]
Do not rename a marker as disease prevention
A marker associated with a disease process can be valuable for research. That does not mean every intervention-induced change in the marker proves that the disease is prevented, slowed or treated. The appropriate inference depends on evidence connecting the intervention, marker and clinical outcome.
Here, the contrast between the tau-related result and the cognitive result is the central story. A headline about Alzheimer’s prevention would skip that distinction and extend beyond the trial. A clear summary can acknowledge the marker signal while stating that cognition did not improve.
The enrolled group also matters. Subjective cognitive decline and mild cognitive impairment are the study’s stated categories. A broad claim about all healthy adults or every stage of dementia would require additional evidence in those populations.
Keep the analysis type attached to the result
The paper used both between-group and within-individual comparisons. Those approaches can contribute different information, but their results should not be mixed selectively into an apparent single positive outcome. If an activity finding appears in one analysis and not another, preserve that context.
Completion counts should also remain visible. Readers benefit from knowing how many people were randomized and how many finished, rather than seeing only the larger number. The next question is how the analysis handled the missing observations and whether they could influence interpretation.
For background, the earlier twenty-person pilot also separated higher blood NAD from unchanged cognition. These studies should encourage outcome-specific reading rather than a broad verdict based on the phrase “brain health.”
If a product cites this paper, ask for a sentence naming the measured marker and a separate sentence stating the cognitive result. That simple request makes overstatement easier to recognize. The study is a useful basis for further investigation, but it does not validate a retail treatment for cognitive decline or a prediction about an individual’s future neurological health. Our 2026 amnestic-MCI pilot adds another controlled cognitive-outcome study to that comparison.