P21 (P021)

The upside is concentrated in mouse mechanism, not in anything proven in people. P21's genuine asset is a clean, broad, brain-penetrant neurogenic profile across rodent models, raising BDNF, lifting hippocampal neurogenesis, and lowering tau pathology. Its weakness is that none of this has ever been demonstrated in a human, and the one independent in-vivo test of the core mechanism failed. Breadth and bioindividuality score modestly because the preclinical footprint is wide but entirely animal. Evidence is the dimension that drags the total down hardest.

Efficacy (2.5/5.0): Efficacy is low because the effects are large and consistent in rodents but completely undemonstrated in humans. In normal adult mice, peripherally administered P021 enhanced learning and both short-term and spatial reference memory while inducing dentate-gyrus neurogenesis, per Li 2010, and in Alzheimer's-model mice it rescued cognition while lowering tau, per Kazim 2014. Aged rats showed reversed memory decline and restored synaptic markers after chronic oral dosing, per Bolognin 2014. Those are real animal effect sizes, but they are not human cognitive outcomes. Mouse neurogenesis and memory rescue predict essentially nothing about a healthy human's cognition, and the Alzheimer's translational record, where dozens of neurotrophic agents worked in mice and failed in people, is the base rate here. A confirmed rodent effect without any human outcome cannot score as effective for the marketed human use, which is why the number stays low despite how clean the animal data looks.

Breadth of Benefits (2.8/5.0): Breadth is moderate because the preclinical footprint is genuinely wide, wider than most grey-market peptides, but it is entirely in animals. Across studies, P021 touched Alzheimer's pathology, normal aging, Down-syndrome models, traumatic brain injury, and even retinal degeneration, with neurogenesis and synaptic rescue as the common thread, per Baazaoui 2017 and Chohan 2015. In Down-syndrome model mice, early treatment rescued developmental and memory deficits while raising BDNF and lowering GSK-3beta, per Kazim 2017, which shows the same core mechanism reaching across very different disease contexts. The boundary is that breadth of mechanism is not breadth of proof. None of these systems has a controlled human endpoint, so the wide reach is a reason the molecule is interesting, not a reason to expect any of these benefits in people. A peptide that does many things in mice and nothing measured in humans has broad potential and zero demonstrated human range.

Evidence Quality (1.5/5.0): Evidence quality is the weakest dimension and the reason the score sits in caution. There are zero human trials of any kind, so start there. Worse, the single independent in-vivo test of P021's signature BDNF mechanism failed: a University of Bologna group found it rescued cells in culture but did not raise BDNF or improve neuroanatomy in live Cdkl5-knockout mice, per Mottolese 2024. On top of that, nearly every positive paper has the inventor as senior author, and he co-founded the company licensing the peptide and holds the patents, an explicit conflict disclosed in the work, per Kazim 2017. Preclinical-only, inventor-dominated, with a failed independent replication of the core claim, this is floor-level evidence and the dimension I refuse to raise on mouse data alone.

Speed of Onset (2.5/5.0): Speed is unknown in humans and slow in the animal record. In rodents, the benefits appeared only after chronic dosing over weeks to months, including in-diet treatment lasting up to a year, per Kazim 2014. The mechanism itself argues against a quick result: building new hippocampal neurons and shifting tau biology are slow biological processes, not same-day effects, per Baazaoui 2017. No acute effect has been demonstrated in any species on a validated endpoint, so vendor claims of a fast acute response are unsupported. With no human pharmacology at all, any timeline for a person is guesswork, which is why this scores middling rather than as a strength.

Durability (2.5/5.0): Durability is unknown in humans and conditional in rodents. The animal benefits were shown under continuous chronic dosing, and some prevention paradigms started treatment before pathology and maintained it for the whole study, per Bolognin 2014. Whether any effect persists after stopping, and over what human timescale, has never been measured. Neurogenesis-driven changes could in theory outlast dosing, but there is no evidence either way, so this is scored from absence of data, not from a demonstrated lasting effect.

Bioindividuality Upside (2.5/5.0): Human inter-individual response is entirely unknown because no human has been studied. In rodents, aged animals with more baseline deficit tended to show larger relative responses, per Bolognin 2014, which hints that people with more cognitive room to recover might respond more. The disease-model work points the same way, with the clearest rescue showing up in animals that started with the most pathology, per Kazim 2014. That is a reasonable mechanistic guess, not data. Without any human variability picture, this lands at the middle, since there are plausible modifiers but no way to predict who would respond, and a healthy person with an intact memory system might be the least likely to notice anything at all.