Longevity stack compounds show uneven results, each tested on its own

Do elamipretide, MOTS-c, epitalon and NAD+ precursors add up in a stack?

Nobody knows whether elamipretide, MOTS-c, epitalon and NAD+ precursors add up, because nobody has run the combination. Longevity stack write-ups follow the same recipe: name a compound, name a pathway, move on. Elamipretide stabilizes cardiolipin in the inner mitochondrial membrane. MOTS-c is a mitochondrial-derived peptide proposed to signal through AMPK.

Epitalon is marketed on a proposed link to telomerase regulation. NAD+ precursors feed the salvage pathway that regenerates NAD+ from nicotinamide. Stated this way, the four sound complementary: different organelles, different enzymes, no overlap.

That framing claims something it has not earned. "These pathways are biologically distinct" is a true statement about mechanism, while "therefore stacking them is additive" is a guess wearing the clothes of a conclusion. Mitochondrial signaling networks cross-talk constantly, and a peptide that shifts AMPK activity and a cofactor that shifts NAD+ availability are not guaranteed to reinforce each other because they enter through different doors. Nothing rules out that they compete for downstream resources or blunt each other's signal, and no one has measured which happens.

What did elamipretide's Phase 3 trial find?

Elamipretide's Phase 3 trial, MMPOWER-3, missed both co-primary endpoints in adults with primary mitochondrial myopathy ^[2]. Elamipretide is the only compound in the stack with a completed, randomized, placebo-controlled Phase 3 trial. MMPOWER-3 randomized 218 participants to 40 mg/day subcutaneous elamipretide or placebo for 24 weeks and measured six-minute walk distance and total fatigue score. Neither endpoint reached significance in the full, genetically mixed cohort ^[2].

Two other analyses complete the picture. An earlier, smaller crossover trial, MMPOWER-2 (n=30), found a 19.8-meter difference in walk distance favoring elamipretide that did not reach significance (95% CI −2.8 to 42.5, p=0.083), alongside significant reductions in patient-reported fatigue ^[3]. A post hoc analysis of MMPOWER-3 by genotype found that participants with nuclear DNA mutations affecting the mitochondrial DNA replisome, mostly POLG and TWNK variants, trended toward improvement on the walk test even though the trial as a whole did not ^[1].

That is a mixed grade, not a clean one, and it applies only to a disease population treated at a specific subcutaneous dose. None of these trials studied elamipretide alongside any other compound, and none enrolled healthy adults using it for longevity. Elamipretide's broader clinical record is analyzed separately.

What did the FDA approve elamipretide for?

The FDA granted accelerated approval to Forzinity, an elamipretide formulation, for a specific Barth syndrome indication in September 2025. That indication does not establish a longevity benefit or validate a multi-compound stack. It also does not make a separately supplied research formulation equivalent to the approved medicine. See the FDA announcement.

Has MOTS-c been tested in humans?

Native MOTS-c has never completed a human clinical trial. MOTS-c and epitalon sit at a lower evidence tier than elamipretide. The foundational MOTS-c work is cell and rodent biology, and the only human trial that reached registration studied a chemically modified analog of the peptide, not the native sequence people are buying and injecting.

No completed clinical trial of native MOTS-c in humans is documented, and that gap should be stated plainly rather than papered over with a mechanism paragraph. MOTS-c's human evidence is covered in a separate analysis.

Is epitalon's evidence independently replicated?

Epitalon's evidence has never been broadly replicated: the bulk of its published dataset traces back to a single laboratory group. That leaves the evidence base without the broad, independent replication that gives a compound's literature real weight. An independent in vitro replication attempt was published and subsequently required a correction, which is worth knowing before treating that replication as confirmation.

No human trial has measured telomere length before and after epitalon dosing, so any claim about telomere effects in people is, at best, extrapolated from cell-culture and animal work. Epitalon's human evidence gets its own analysis.

Neither the MOTS-c gap nor the epitalon gap means the compounds do nothing. The accurate label for both is "mechanistically proposed, not clinically demonstrated in humans," a different claim from the one implied by a protocol page that lists a dose next to a citation-free mechanism paragraph.

What do NAD+ precursor trials show?

NAD+ precursors reliably raise blood and tissue NAD+ in randomized trials ^[4]⁠^[6]⁠^[7]⁠^[8], but the functional and cognitive outcomes measured in the same trials have been mostly flat ^[4]. Among the four compounds, NAD+ precursors have by far the deepest randomized-trial record. Nicotinamide riboside raised blood NAD+ 2.6-fold in older adults with mild cognitive impairment ^[4], elevated the skeletal muscle NAD+ metabolome in aged men ^[6], and increased cerebral NAD+, measured by phosphorus magnetic resonance spectroscopy, in a Parkinson's disease trial ^[7]. A head-to-head comparison of nicotinamide, nicotinamide riboside, and nicotinamide mononucleotide found that NR and NMN raise circulatory NAD+ comparably, while plain nicotinamide does not sustain the same chronic increase ^[8].

Functional outcomes tell a different story. In the cognitive impairment trial, NAD+ rose 2.6-fold but cognition scores stayed flat, and cerebral blood flow in the default mode network decreased ^[4]. In a COPD trial, six weeks of NR lowered sputum interleukin-8 by an estimated 52.6% (95% CI −75.7% to −7.6%, p=0.030), a real anti-inflammatory signal, but in a disease population with active airway inflammation, not in healthy adults using NR for longevity ^[5].

The pattern across the NR trials is consistent. Raising blood or tissue NAD+ is well established and repeatable; converting that biochemical change into a clinically meaningful outcome in a general, healthy population is not yet established.

Has anyone tested these longevity compounds together?

No published trial, case series, or pharmacokinetic study has looked at any two of these compounds taken together, let alone all five. Every cited result describes one compound studied alone, usually in a specific patient population, at a specific dose and route, and none of it transfers automatically to a mixture.

Elamipretide's trials used subcutaneous dosing in adults with a diagnosed mitochondrial disease ^[2]⁠^[3]. The NR trials used oral dosing up to 1 g/day in cognitively impaired, COPD, or Parkinson's populations, not in people simultaneously taking a mitochondrial-targeted peptide, an AMPK-signaling peptide, and a compound with almost no independent human data ^[4]⁠^[5]⁠^[7].

No published pharmacokinetic interaction study has asked whether an NAD+ precursor changes the exposure or clearance of elamipretide, or whether elamipretide changes NAD+ salvage kinetics. No one has measured a shared biomarker, such as fasting glucose, inflammatory markers, or six-minute walk distance, in the same subjects taking all five components at once versus each alone. The absence of a warning on a protocol page is not evidence that the combination is safe; it is evidence that the combination has never been studied.

How strong is the evidence behind a five-compound longevity stack?

No evidence exists for the five-compound stack as a combination. Grading elamipretide, NAD+ precursors, MOTS-c, and epitalon one at a time gives four different answers:

• Elamipretide. A real Phase 3 trial that missed both co-primary endpoints, with a possible signal in one genetic subgroup ^[1]⁠^[2]⁠^[3].
• NAD+ precursors. They reliably move a blood biomarker without reliably moving function ^[4]⁠^[5]⁠^[6]⁠^[7]⁠^[8].
• MOTS-c. The native molecule has never completed a human trial.
• Epitalon. A thin, largely single-source literature with no human telomere data behind it.

Combining them does not average those grades into one respectable score. It stacks four separate uncertainties, plus an entirely untested fifth variable, the interaction itself, which no trial anywhere has measured.