An N-of-1 trial (or single-subject design) is a personalised clinical trial where a single patient acts as their own study population. Instead of testing a drug on a large group of people, an individual receives sequential, randomised treatments or placebos over time to determine the optimal care. N-of-1 trials are a core tool in precision medicine to evaluate customised and off-label therapies when a traditional large-scale clinical trial isn’t possible, especially in ultra-rare diseases or chronic conditions.
The ultra-rare condition, in my case, is an obsession with self optimisation. This trait did emerge from a long history of trying to get to the bottoms of some pretty murky women’s health and/or metabolic symptoms. I use and/or deliberately here, because the boundary between reproductive hormones, metabolism, inflammation, stress physiology, and energy regulation… well there isn’t a boundary. Polycystic ovary syndrome (PCOS) has just this year been renamed polyendocrine metabolic ovarian syndrome (PMOS). In the past 5 years, I have self-experimented a lot. Attempting to optimise performance as a university athlete, and in a high-stress early-career environment. I documented along the way. (*side-note: This is one of the topics I am most excited to implement into my second-brain knowledge base. A properly structured personal data layer (symptoms, interventions, labs, training, nutrition, sleep, cycle phase, stress, environment) could reveal relationships that are too subtle to notice anecdotally but too individualised to show up cleanly in population averages.) Over the next few weeks, I am going to structure some of these self-experiments and results into a series of posts.
Alongside documenting the experiments here, I’ll be entering findings into StuffThatWorks, a patient-reported platform built to expose patterns that emerge across similar patients. The goal is not to pretend personal data is the same as randomised evidence. The goal here is to place personal experience inside a broader, structured context.
Important to stress: this is not medical advice. Any positive result from my own experience should not be extrapolated casually to another person’s physiology. Equally, any negative result should not be treated as evidence that an intervention doesn’t work. An N-of-1 result is a signal from one biological system, under one set of conditions, measured against one set of outcomes.
The first topic in my self-experimentation series: peptides.
—
A peptide is, at its simplest, a short chain of amino acids. In drug development, the field is usually focused on therapeutic peptides in the rough range of 4 to 100 amino acid residues, sitting somewhere between small molecules and larger biologics: more programmable and target-specific than many small molecules, but usually smaller, more synthetically tractable, and less structurally complex than antibodies or full proteins.
The history is not fringe. It begins with some of modern medicine’s most important drugs. Insulin therapy transformed diabetes care in the 1920s, making peptide therapeutics part of clinical practice. Oxytocin became one of the first peptide hormones to be sequenced and synthesised, and Bruce Merrifield’s development of solid-phase peptide synthesis in 1963 made it practical to build peptides rather than just extract them from tissue. Peptides often mimic or directly modify signalling systems the body already uses. They are vulnerable to degradation, short half-lives, often require injection, and can struggle with bioavailability and tissue penetration. Much of modern peptide chemistry is therefore a game of preserving the biological signal while engineering around the molecule’s weaknesses.
Small enough to be designed and modified, large enough to carry complex signalling instructions, and close enough to endogenous biology that their effects can be highly state-dependent.
Part 1. BPC-157
I tested BPC-157 during a return-to-competition training phase, increasing volume and intensity, and potentially increased recovery speed from a minor running injurty (ITB). The primary hypothesis if BPC-157 meaningfully supported tissue repair, I might see better recovery tolerance as training load increased. The secondary hypothesis was more interesting: gut-mediated recovery. BPC-157 was studied as a stable gastric peptide, and much of the preclinical literature frames it around cytoprotection, wound healing, angiogenesis, nitric oxide signalling, and repair across gastrointestinal and musculoskeletal tissues. Human evidence remains thin, so the “gut healing” claim should be treated as mechanistically plausible but not clinically proven.
Mechanistically, the gut hypothesis is that improving mucosal repair and local inflammatory signalling could reduce systemic stress load, indirectly changing recovery, sleep, and energy regulation. That is not the same as saying BPC-157 heals the gut; it is a testable hypothesis about whether a repair signal at the barrier/inflammatory interface shows up downstream in physiology.
Note: The safety, efficacy, and appropriate dosage of BPC-157 are not well-established in humans.
Main finding
I expected the most obvious readout to show up in gut symptoms, and/or, in training recovery. Instead, the clearest pattern appeared in my REM sleep data: the first week of bpc-157 coincided with a doubling of my REM sleep (from an avg >40-50 min at baseline to avg 2 hours in week 1). This was notably the highest REM I had ever seen tracked by my garmin. Now, confounding variables are a thing,. Of note, the REM increase occured hbefore increasing training volume, which commensed in Week 2. This does not prove BPC-157 increased REM sleep. Consumer wearables are useful for longitudinal self-tracking, but sleep-stage classification is still imperfect; validation studies generally find better performance for sleep/wake and total sleep time than for precise staging, and Garmin stage-level estimates should be treated as directional rather than definitive.
But it is interesting!
If the REM change was real, there are a few possible interpretations:
Improved recovery tone: BPC-157 may have reduced some background repair/inflammatory burden, allowing deeper sleep architecture to normalise.
Gut-brain axis effect: If gut irritation or barrier stress was contributing to sympathetic tone, improved gut signalling could plausibly show up as better REM continuity.
Wearable artefact: Garmin may have reclassified sleep stages differently due to changes in heart rate, movement, HRV, or sleep timing.
Early verdict: interesting, not proven.
BPC-157 is best treated here as an investigational repair signal, not a confirmed healing intervention. The strongest personal finding was the REM pattern, but because the experiment was not blinded, not isolated, and occurred during a changing training block, the result is anecdotal only.
That said, the most common criticism toward BPC-157 in particular, has been placebo effect. This may well be the case. I first encountered the peptide in crossfit communities around 2 years prior, where it was commonly referenced as a wolverine healing miracle that reduced time out from injury. This may well be true, especially when the off-label usage has proliferated first through a specific community using it for a specific context. However, if I were able to bottle the placebo effect, it would have efficacy enough to warrant it’s use. Placebo has been seen to outperform certain approved drugs. But the interesting case here, I had no expectation for any improvement to my sleep data.
Thank you for reading. I’ll be brainstorming more interesting and quantitative ways to record these self-experiments, and I’ll share more in due course.