TL; DR: We want to run a self-blinded RCT with 2-3 participants on intranasal administration of Orexin-A, testing whether it ameliorates the effects of mild sleep deprivation and/or reduces sleep need. We’re asking for $1k-$10k for the Orexin, some measurement devices to track health variables, and potentially other health tests with more funding.
Context
Orexin-A and Orexin-B (aka hypocretin) are neuropeptides that are involved in a number of different brain processes including in appetite, energy expenditure, and wakefulness. Type-1 narcolepsy is caused by a deficiency in Orexin-producing neurons which causes an Orexin-A deficiency.
Direct administration of orexins or administration of an orexin agonist (which activates the orexin receptors) increases wakefulness and reduces sleep. One study finds that giving monkeys exogenous Orexin ameliorates the effects of sleep deprivation. Administering an orexin antagonist (which blocks the receptors) causes sleepiness. Several orexin agonists are under trials as narcolepsy drugs. Several orexin antagonists have been approved as treatments for insomnia.
We are most interested in the potential for orexin to reduce sleep need and combat the effects of sleep deprivation. A stimulant effect alone from orexin without a reduction in sleeping time is not interesting.
The Plan
Procurement: Depending on the amount of funding we receive we’ll decide which supplier to order from; one of us isn’t comfortable with Suaway, so with low funding we’ll only have two participants.
Low funding: 2× Orexin-A+B from Suaway (~$500 in total with shipping)
Medium and high funding: 3× Orexin-A from Thermo Scientific (~$2000 in total with shipping)
High funding: 3× Orexin-A, 3× Orexin-B from Thermo Scientific (~$4000 in total with shipping), plus some other things like blood tests
Thermo Scientific is preferable as a supplier because the Orexin likely has higher purity and is available in powder form that can be stored in the freezer indefinitely (peptides such as Orexin can degrade in solution).
Experiment
The experiment will be a self-blinded randomized controlled trial, run in blocks of four days, ten blocks per participant. We’ll use a saline solution as a passive placebo.
Each participant will collect several active & passive measurements, including reaction speed via the psychomotor vigilance task, attention via the digit symbol substitution test, digit span, match-to-sample task, and whatever data is collected via a fitbit wearable. Data will be collected twice a day.
Data Analysis
We’ll collect and analyze the data from the three participants and analyze it by running some standard statistical tests (probably Welch’s t-test) on the resulting data. Things that will be tested for (for each measured variable):
How does performance on sleep deprivation+Orexin compare to sleep deprivation+placebo days?
How does performance on sleep deprivation+Orexin compare to normal sleep days?
Is there catch-up sleep in the night following Orexin days?
In the Case of Success
If our experiment shows promising results (strong effects on cognitive variables with minimal negative effects), we'll reach out to prospective funders and academic sleep labs that already study orexin, particularly at Stanford, Harvard, and Penn. These groups have existing research programs and IRB infrastructure, so our pilot data could help justify formal n=20-50 studies within their labs.
Longer-term, successful academic replication could lead to clinical development by companies already working in the orexin space (Jazz Pharmaceuticals, Takeda) or interest from nootropics companies for direct-to-consumer applications.
How the Money Will Be Used
1 Fitbit/person for data collection ($75/$150 in total with shipping)
Low funding: 2× Orexin-A+B from Suaway (~$500 in total with shipping)
Medium funding: 3× Orexin-A from Thermo Scientific (~$2000 in total with shipping)
High funding (in decreasing priority)
Health checkup at a doctor during the trial
Longer trial (more samples, different dosage)
Blood tests (1-2×/person, $100 per test)
Output
The output will be a write-up of a survey of the literature on the topic, a description of the experimental protocol, a statistical analysis, and the results. We will also publish all data collected during the time, and the code of the software written to collect that data. We’ll post the write-up of the results on LessWrong and one of our blogs, irrespective of whether we get a null result or find an effect, and potentially bioRxiv if we find a positive result.
Other details
Why hasn’t this been done before?
There are a few barriers to reducing sleep need with orexin:
You can’t patent a natural hormone, so you would have to design an orexin-like drug instead, which is expensive.
Stimulants like caffeine and nicotine are readily available and there are numerous stimulant medications so it’s hard to imagine a new stimulant being competitive.
Even if we had evidence that orexin agonist reduced sleep need, that isn’t sufficient for FDA approval. Sleep isn’t considered a disease, and the FDA won’t support a cure for something everyone has.
Essentially all research on Orexin has been in the context of treating narcolepsy, and not been focused on whether it can reduce sleep need in healthy humans.
All together, there’s little incentive for researchers or drug companies to pursue sleep need reduction therapies. We want to pursue this as a pure public good to get more people to work on sleep need reduction.
What Could Go Wrong
During the experiment, we find that Orexin is harmful to our health, and we abort the experiment
We find a null result in effectiveness over placebo
We find that Orexin improves performance (either up to full sleep baseline or above), but requires standard amounts of catch-up sleep, basically acting as a normal stimulant
Who's Involved
Appendices
Predictions on Outcomes
Manifold market on reaction speed effect size.
Experiment aborted due to health concerns:
Niplav: 5%
NMP: 10%
Sam: 5%
Null result
Niplav: 65%
NMP: 75%
Sam: 45%
Stimulant result but catch-up sleep
Niplav: 15%
NMP: 5%
Sam: 40%
Positive result without catch-up sleep
Niplav: 5%
NMP: 2%
Sam: 10%
Negative result, worsening compared to placebo:
Something else:
BOTEC for return on investment:
If orexin works: 1 hour less sleep = ~4% life improvement
QALYs per person per year: 0.02
Potential users: 4.5B × 20% = 900M people
// Probability chain: positive result → eventual adoption
P(positive): 4.5%
P(adoption|positive): 0.17% (academic→clinical→approval→adoption)
P(discovery never happens otherwise): 30%
P(we accelerate discovery by 5-10 years): 70%
Average acceleration: 7.5 years
Total discovery QALYs: 4.5% × 0.17% × 900M × 0.02 × 30y = 54k QALYs
Acceleration QALYs: 4.5% × 0.17% × 900M × 0.02 × 7.5y = 13.5k QALYs
Expected QALYs: 30% × 54k + 70% × 13.5k = 25.7k QALYs
Expected value: 25.7k × $100k = $2.6B
ROI: $2.6B / $5k = 520:1 (hundreds)
Safety
Summary
There is little to no research on the safety intranasal administration of Orexin-B because Orexin-B’s poor ability to pass through the blood-brain barrier (BBB) makes it less attractive for research purposes, whereas Orexin-A readily passes through the BBB and is “more potent” than Orexin-B (Mieda et al. 2013).
We believe that given:
Orexin-B’s lower BBB diffusion rate
Orexin-B’s lower potency to Orexin-A
The many Orexin-A-wasn’t-harmful-to-me anecdotes
that intranasal administration of Orexin-A is the first thing we want to test if we have the funding to buy pure Orexin-A. This assumes that:
The peptide source purity is sufficiently high
The synthesis of the intranasal spray is performed correctly correctly
The administration of the intranasal spray is performed correctly
Additionally, with enough funding to buy Orexin-B we believe we will be one of the first humans to intranasally administer Orexin-B.
Literature
Kastin & Akerstrom 1999 found that:
Orexin B, moreover, was rapidly degraded in blood, so no 125I-orexin B could be detected in intact form in brain when injected peripherally. Thus, although orexin B is rapidly metabolized in blood and has low lipophilicity, orexin A rapidly crosses the BBB from blood to reach brain tissue by the process of simple diffusion.
Calvin & Fadel 2018 state that:
While we are not aware of any papers describing intranasal administration of [Ala11,D-Leu15]-OxB in humans, Weinhold et al. (Weinhold et al., 2014) observed enhanced wakefulness and attention in human narcoleptic patients following intranasal administration of 435 nmol [1261 ug] OXA. Thus, given the limited comparative information available, our orexin doses [200 nmol for humans] seems reasonable.
Baier et al. 2011 found no harmful effects of intranasal OxA.
Weinhold et al. 2013 found no harmful effects of intranasal OxA.
Anecdotes
There are many internet anecdotes that show either benefits or null results of OxA:
So I tried a dose of several mg in one session. It did not produce any noticeable acute psychoactive effects, but made me not able to sleep for about 70 hours or so. I did feel less tired than I normally would have after not sleeping for a night or two, but no real stimulatory effects.
I was suffering from quite bad insomnia at that time, and after that dose I was able to sleep an hour or two more per night. Do not really know why, but it seemed to have done something to my brain.
I tried it a few times. I'm pretty anxious overall but Orexin-A didn't have any impact on my anxiety levels. … Honestly, I didn't feel any noticeable effects [in response to “Did it have any effect on your sleep quality or daytime alertness?”]
I have all of the symptoms you have described other than waking at night. I did a full bottle of orexin a at a variety of dosages and did not experience anything. Ability to nap was unhindered. I had gotten it when ceretropic was still around so the source should have been reliable.
The first day was not a significant difference, however I must admit that today I was able to focus and study without the strong urge to fall asleep, something that I was not able to do for a long time.
I absolutely feel great results. … Near instant effect, and bypasses the blood stream by being absorbed through the nerves directly to the brain.
Our scores on the tests during conditions of sleep deprivation improved to be commensurate with baseline.
Experimental Protocol
It doesn't matter when a block starts, it can be any day of the week, and one can take a couple of days off from the experiment when something (e.g. a holiday) comes in the way.
Measurements
We'd like to keep the measurements manageable and scalable: There is a core of measurements performed in each block, but if I feel like I have slack I might decide to do more extensive measurements for one block.
If done per block, this will not impact the quality of the data, but please don't decide within a block to switch the detail of measurements.
Measurements were selected according to how much they degrade with sleep deprivation, informed by an auto-generated report using Elicit.
Most of the datapoints will be collected with an application using pygame.
Active measurements
Passive measurements