Start with “compared with what?”
A person can feel better during a study without the study demonstrating that the assigned supplement caused the improvement. That is not a contradiction. It is a question about the comparison. I read an NAD headline by locating the treatment group, the relevant control and the measurement time before deciding what the result supports.
Three NR studies illustrate different designs. They should not be combined into an average benefit score: they tested different participants and outcomes. Their practical value here is to show why a change from baseline, a controlled difference and a biological observation need different wording.
A baseline improvement is not the controlled result
In a long-COVID trial, NR increased blood NAD, but randomized comparisons did not find significant differences in cognition, fatigue, sleep or mood. Later exploratory analyses pooled NR-exposure periods and found some improvements from baseline. Those analyses were post hoc and unadjusted for multiple comparisons. [1]
The paper’s Results distinguish 61 enrolled and randomized from 58 completing baseline testing. Dropout also matters when interpreting later findings. The full trial explainer preserves that distinction. I would not turn an exploratory within-group improvement into a claim that NR beat placebo for symptom recovery. Nor does a nonsignificant result prove every possible benefit is absent.
Crossover means the comparison happens within participants
In a randomized, placebo-controlled crossover study, thirteen adults received NR and placebo in separate six-week periods. The study reported changes in certain muscle metabolites and minor body-composition measures, but no improvement in insulin sensitivity or mitochondrial function. [2]
Having people receive both conditions can help compare responses while accounting for stable differences between participants. It does not remove questions about sequence, time between conditions or whether the measured outcome answers the consumer claim. For the specific findings, read the body-composition study; do not borrow one favorable endpoint to describe all metabolic health.
Twins are informative without becoming untreated controls
A five-month study gave NR to both members of twenty BMI-discordant identical-twin pairs. It reported biological changes including muscle mitochondrial measures, without improvement in adiposity or metabolic health. Because both twins received NR, the design did not provide an untreated twin comparison. [3]
The twin relationship helps researchers explore similarities and differences. It does not make an exposed co-twin equivalent to a placebo group. Our twin-study analysis explains the longer observation period without turning it into proof that a longer retail subscription will deliver a health benefit.
Keep three comparisons on three lines
| Design | Useful comparison | Do not silently substitute |
|---|---|---|
| Randomized parallel groups | Assigned treatment versus its concurrent control | Change from one group’s own baseline |
| Randomized crossover | Treatment and control periods within participants | A different outcome or an unrelated formula |
| Both twins exposed | Changes during exposure and differences between twins | An untreated placebo-controlled contrast |
Turn the headline into a usable worksheet
In the study claim reader, put the claimed benefit in the first field and the actual measurement in the endpoint field. Write the controlled result before any exploratory result. If the paper has several populations or timepoints, name the one used in the claim. The worked example shows how to keep a symptom claim separate from a blood marker.
Next ask whether the marketed product uses the studied ingredient and route. A paper about oral NR cannot by itself establish the benefit of an NAD+ injection, NMN capsule or multi-ingredient blend. The ingredient overview is useful for that second check.
For a purchase comparison, keep Life Extension’s NR bottle and Tru Niagen’s selected product tied to their labels and terms. A clearer label or lower cost can improve a shopping comparison. Neither turns a study’s missing clinical result into a demonstrated benefit.