The effects of caffeine consumption do not decay with a ~5 hour half-life — LessWrong

epistemic status: confident in the overall picture, substantial quantitative uncertainty about the relative potency of caffeine and paraxanthine

tldr: The effects of caffeine consumption last longer than many assume. Paraxanthine is sort of like caffeine that behaves the way many mistakenly believe caffeine behaves.

You’ve probably heard that caffeine exerts its psychostimulatory effects by blocking adenosine receptors. That matches my understanding, having dug into this. I’d also guess that, insofar as you’ve thought about the duration of caffeine’s effects, you’ve thought of them as decaying with a ~5 hour half-life. I used to think this, and every effect duration calculator I’ve seen assumes it (even this fancy one based on a complicated model that includes circadian effects). But this part is probably wrong.

Very little circulating caffeine is directly excreted.^[1] Instead, it’s converted (metabolized) into other similar molecules (primary metabolites), which themselves undergo further steps of metabolism (into secondary, tertiary, etc. metabolites) before reaching a form where they’re efficiently excreted.

Importantly, the primary metabolites also block adenosine receptors. In particular, more than 80% of circulating caffeine is metabolized into paraxanthine, which has a comparable^[2] binding affinity at adenosine receptors to caffeine itself. Paraxanthine then has its own 3-5 hour half-life as it’s metabolized into a handful of other things.

Since paraxanthine is by far the dominant primary metabolite, and its further metabolites are mostly either very short lived or have poor affinity for adenosine receptors, we can fruitfully use a simplified model of Caffeine ⟶ Paraxanthine ⟶ Elimination. The upshot is an effective concentration curve with a broader peak and slower decline — about twice as long to reach half of peak effective concentration, assuming paraxanthine and caffeine are equipotent — than that given by the simple elimination model. When I say effective concentration, I mean the concentration of caffeine that would be needed to produce the same effect.

Below is a simulator that models caffeine and paraxanthine metabolism following ingestion of 100mg caffeine or paraxanthine (link to full version). Note that the relative potency of paraxanthine can be adjusted; I am uncertain about how it compares to caffeine within that 4-fold window (see the section on relative potency).

Paraxanthine supplements

If the above model is correct, paraxanthine itself is sort of like caffeine that behaves the way I mistakenly used to believe caffeine behaves. Directly using paraxanthine as a stimulant would have two major differences/advantages compared to caffeine:

1. Effects wear off a lot faster, which means you can take it later in the day without affecting sleep as much (or in the morning without affecting a siesta as much)
2. For a given peak level of stimulation you probably develop less tolerance, since active molecules don’t hang around as long (really just another implication of point 1)^[3]

(It might also have some other differences, like a somewhat different profile of effects.)

In the US you can buy paraxanthine itself as a supplement (I believe it only became available recently, in 2022). I’ve been using 100mg capsules intermittently in the last few weeks.^[4] Some early impressions:

• Taking one or two capsules in the morning has a similar wakefulness-promoting effect to drinking a cup or two of coffee in the morning
• I’ve taken one capsule around 4-5pm a handful of times, which did not interfere with going to sleep at 10-11pm
• Taking a capsule at 7pm gave me enough energy to do some work in the evening, and did not interfere with going to sleep at 11:30pm
• Subjectively the peak effects of 100mg paraxanthine feel similar or weaker than those of 100mg caffeine, while 200mg paraxanthine feels stronger than 100mg caffeine
• The effects seem to peak within an hour of dosing (maybe 30-45 minutes)
• I sometimes feel a bit of an energy crash 2-3 hours after dosing when I take paraxanthine in the morning, but not in the afternoon/evening

The supplements sold in the US appear to exclusively use enfinity branded paraxanthine, perhaps due to them holding some very broad patents on the use of paraxanthine as a supplement.^[5] On their website they emphasize that paraxanthine has a “cleaner” effect than caffeine, is supposedly safer, and has a somewhat shorter half-life that isn’t affected by slow caffeine metabolism.^[6] The somewhat shorter half-life undersells this point: it’s effectively much shorter due to the lack of active downstream metabolites. They also don’t say anything about reduced tolerance on the main page (though it’s mentioned in the FAQ).

Exactly how potent is paraxanthine compared to caffeine?

By “how potent”, I mean the binding affinity at adenosine receptors. The binding affinity (Ka) is equivalent to the odds ratio that a ligand will be bound to some particular receptor at some point in time, divided by the concentration of the ligand (since the odds ratio is proportional to ligand concentration).^[7] If one substance has twice the affinity as another, you only need half as much to get the same receptor occupancy.

There are four subtypes of adenosine receptors: A1, A2A, A2B, A3. The psychostimulatory effects of caffeine are thought to be mediated by the blocking of A1 and A2A receptors, with A1 receptors probably being more important. I was surprised to learn that the affinities of caffeine and paraxanthine for human A1 and A2A receptors are not well established in the literature. Below are my takeaways from researching this:

• Paraxanthine has about 1.5-2x the affinity of caffeine for A1 and A2A in most studies where both have been estimated, across multiple species
• Paraxanthine’s affinity for human A1 and A2A receptors has only been measured in one study that I was able to find
• • This result concurs with paraxanthine having ~2x caffeine’s affinity
  • This study used a different experimental method for the estimation (adenosine-cAMP dose-response curves) than most other studies (radioligand binding)
  • The human receptors were expressed by CHO cells (a cell line from hamsters that’s easy to work with)
• Human measurements of caffeine’s affinity for A1 and A2A range over a factor of 5 and 20 respectively^[8]
• • The 5x range for A1 does seem to reflect genuine reported measurements, while I’m less sure about the 20x range for A2A (e.g. maybe the most extreme values were misreported in that giant linked table)

From the above we might tentatively guess that paraxanthine is 1.5-2x as potent a stimulant as caffeine, while noting that measurements seem to be all over the place. This is in contrast to my subjective experience so far, where an equal dose of paraxanthine feels similar or weaker than caffeine.

Note that I’m comparing paraxanthine to coffee, which could be confounded by several things such as other stuff in the coffee having effects, placebo, or the coffee pods I was using having more or less caffeine than claimed. A blind comparison between caffeine and paraxanthine capsules would be much better, and in hindsight I wish I’d done that right away.

Some other hypotheses for what could be going on here:

• My subjective impressions are unreliable
• The relevant receptors behave differently in human brains than in the assay conditions used by these studies
• • The measurements being all over the place is suggestive of conditions making a big difference
• The A1/A2A antagonism model is incomplete
• • E.g. my impression is due to the lack of other subjective effects of caffeine besides the primary stimulatory effects
  • E.g. caffeine has additional stimulatory effects mediated by non A1/A2A pathways
• Less absorbed paraxanthine reaches A1/A2A receptors in the brain for some reason
• Paraxanthine’s peak effects are reduced by slower absorption
• • I doubt this: subjectively effects seem to peak quickly (maybe 30-45 minutes after dosing), and wear off fairly quickly (a few hours)
• The bioavailability (absorption) is worse than that of caffeine
• • Probably not if we think the absorption is fast, since fast absorption generally implies high bioavailability
• There’s less paraxanthine in the capsules than claimed

Concluding thoughts

I recommend trying paraxanthine as a stimulant. For me personally, two use cases stand out so far:

1. Taking a 100mg capsule in the late afternoon/evening can give me enough energy to work when I’d otherwise be too tired, without noticeably affecting my sleep
2. Taking 100-200mg after waking can maybe replace coffee for getting me going in the morning, probably with less resulting tolerance buildup

Considering how widely used caffeine is, I was surprised to learn through independent research that paraxanthine considerably extends the duration of its effects, which almost no one seems to understand. I was also surprised by how poorly characterized the interactions of caffeine and paraxanthine with adenosine receptors seem to be. There’s probably a civilizational inadequacy story here.

If such an important point about caffeine — the one ubiquitous nootropic — was unknown to even the nerds, that’s some evidence towards there being other low hanging cognitive enhancement fruit. That is, we expect more low hanging fruit in this world than in the worlds where the nerds already knew about this.

Paraxanthine-based stimulants look to me like a pretty darn low-hanging fruit that took forever to be picked; science has known about caffeine metabolism and paraxanthine’s adenosine receptor antagonism since at least the early 1980s, yet the paraxanthine supplements only became available a few years ago.^[9]

1. ^{^}

   Caffeine can easily cross between the brain and bloodstream. This means that, because of diffusion, the concentration in the brain (where psychostimulatory effects are mediated) closely tracks the concentration in the bloodstream. When I say “excreted” I mean the molecule is removed from circulation, such that its concentration in the bloodstream (and brain) is reduced. This is mostly done by the kidneys, with the molecules ending up in urine.

2. ^{^}

   Numbers are surprisingly hard to pin down here; see the section on relative potency.
3. ^{^}

   My guess is that equilibrium tolerance is roughly a function of the (per day) area under the curve (AUC) of equilibrium effective concentration.

4. ^{^}

   The label says not to exceed 300mg per day. My guess is this comes from the maximum recommended daily caffeine intake of 400mg, which is metabolized to about 300mg paraxanthine in vivo.

5. ^{^}

   The linked patent seems to claim any use of paraxanthine as a dietary supplement with a dose in the range 2-800mg.
6. ^{^}

   Might not there be slow paraxanthine metabolizers as well? I suppose enfinity isn’t incentivized to ask this question.

7. ^{^}

   Note that reported values are usually Ki/Kd, which is the reciprocal of this definition. The meaning of these values is the concentration needed for 50% receptor occupancy by the ligand (i.e. 1:1 odds ratio).

8. ^{^}

   Why such big discrepancies? I don’t know. One thought is that, since adenosine receptors are G-coupled protein receptors (GPCRs), it might not be reasonable to summarize a ligand’s binding affinity with a single number. This is because GPCRs have two different states, and ligand binding affinities for the states can differ.
9. ^{^}

   This is also evidence that some circumstances recently changed. For example: a regulatory change, or cheap mass production of paraxanthine being enabled by other tech developments.