Optimising the Squeeze

Compared to the total we were able to pull out of the kava powder using the gold standard for kavalactone extraction, a Thermo Scientific Dionex™ ASE™ 350 Accelerated Solvent Extractor, running HPLC grade organic solvents at 60 °C and 105 Bar, with 5 minutes of pre-incubation and 20 minutes of static hold, followed by 150% rinse volume, here’s how much of the combined 6 major kavalactones we were able to extract from the same kava powder on each “normal squeeze”:

Squeeze #1) 51.16%

Squeeze #2) 14.17%

Squeeze #3) 6.25%

Squeeze #4) 4.32%

Squeeze # 5) 3.12%

Squeeze #6) 2.16%

Squeeze #7) 1.68%

Squeeze #8) 1.44%

These 8 successive squeezes of the same kava yielded a total kavalactone extraction efficiency of 84.31%.

Bear in mind that the extraction efficiency in water (from a single wash) has been anecdotally reported elsewhere as being as low as only 15% of the available kavalactones, and there are numerous factors that affect this beyond those we’re exploring here. One extremely important (but often overlooked) example is how the kava was processed prior to packaging; This may be surprising to some, but factors such as the way kava is harvested, washed, peeled, chopped, dried, and milled can influence how easy it is to pull kavalactones out of it, so even if you buy two kava powders which test with similar kavalactone content, they may not extract the same at home – something that we have learned through many years of experience.

Nevertheless, there is no doubt that the best chunk of what is available is coming out in the first wash, but there may be some value in doing further washes if you’re looking to wring your kava out for everything its worth.

It’s also important to realise that the kava produced from subsequent washes was much more dilute than what was obtained from a single squeeze; Even though we continued to extract kavalactones with sequential washes, we ended up with much more liquid too. Kava derived from a single squeeze was more authentically similar to kava served in nakamals throughout Vanuatu, with its creamy and rich texture. Performing a second squeeze and combining it with the first resulted in a milder version which was still fairly enjoyable. Beyond this, we found it too watered down.

The extraction efficiency of each kavalactone under these conditions was not identical to one another, but the chemotype of the prepared beverages closely reflected the chemotype of the powder, especially in the first 2 squeeze cycles. Subsequent extractions saw the relative amounts of yangonin decrease, while the relative amounts of desmethoxyyangonin, methysticin, and dihydromethysticin slightly increased. The relative amount of flavokavains decreased substantially in subsequent washes, and after 5 washes we could no longer quantify the very small amount of extracted flavokavains. The relative amounts of kavain and dihydrokavain, the two most abundant kavalactones in our powder, remained fairly consistent from wash to wash.

So, should you do another wash?

Well, that’s entirely up to you!

The majority of the kavalactone content in traditionally prepared kava lies in the sediment that settles out of the drink. After scrutinising hundreds of kava preparations (and thousands of kavas) in our lab (with 36 unique samples prepared and analysed by UHPLC just for this investigation into the effect of water temperature on kavalactone extraction efficiency), we can say this with certainty. This is why stirring the natambia (serving bowl) is essential before dishing out each and every shell to distribute the kavalactones evenly from serving to serving.

Using hotter water during the squeeze objectively yields much higher sediment content than using colder water. This is abundantly evident when many samples prepared the same way (except for water temperatures) are lyophilised (controlled removal of the water by sublimation at low temperatures and pressures) – The volume of residual material in each vial noticeably rises from the one before, stepwise, in direct relation to the temperature used to prepare the sample.

When we centrifuge our samples at extreme g-forces for extended durations and subsequently separate and lyophilise the supernatant (examining the “water layer” instead of the sediment), we observe the same trend. Not only does hot water extract far more sediment, but it also extracts significantly more soluble material and nanometer-scale particles; Lyophilised supernatant from ice-cold extractions results in nearly empty vials, while lyophilised supernatant from very hot extractions results in vials that are still full to the brim, holding the shape of the material which was dissolved in the water even after the water is frozen and sublimated away.

Hot water extracts more material from traditional kava powder into the resulting beverage than using cold water does. There is no reasonable doubt or debate about it.

Given these observations, and especially when taken in conjunction with the plethora of people who state the importance of using warm (or even hot) water during the squeeze, one could be forgiven for assuming that more sediment (and more dissolved material) equates to more kavalactones, but interestingly, we found that the total kavalactone content remained more or less unchanged, regardless of the amount of sediment or whether the squeeze was done with ice water or at temperatures so hot that starch gelatinisation occurred, resulting in thick, gooey kava that most would find truly unpalatable (or at temperatures anywhere in between); The ratio of kavalactones to sediment decreased with rising extraction temperatures.

Squeezing traditional kava powder at different water temperatures resulted in an essentially flat trendline, with no significant change to kavalactone extraction efficiency.

However, performing the squeeze with different water temperatures did result in making different kavas, for more reasons than just the sediment content:

The chemotype of the prepared beverages closely reflected the chemotype of the traditional powder used to make it, regardless of water temperature, but it was not an exact parallel; We noticed that the kavain to dihydromethysticin ratio (K:DHM) and the kavain to dihydrokavain ratio (K:DHK) showed a slight, but clear downward trend as water temperatures increased. The accompanying graph shows the smoothed trendlines.

Squeezing kava with ice water yielded a chemotype closer to one potentially associated with more euphoric effects, while hot water squeezes produced chemotype changes leaning towards the calming side of the spectrum, although the differences were not substantial.

Kavain is often characterised as being the compound most responsible for kava’s ability to induce “headiness”. Dihydromethysticin is often cited as being at the opposite end of the spectrum – it is metabolised more slowly and is generally regarded as being a major contributor to the “heavier” side of the subjective kava experience. In many respects, dihydrokavain is often thought of as being somewhere between kavain and dihydromethysticin in terms of its psychoactive effects. It is worth noting here, however, that the experience might best be viewed as a result of the synergy among the combined molecular orchestra at play, rather than attributing any specific effect to a single compound.

Nevertheless, these findings suggest that not only will kava squeezed in cold water be lighter in texture with substantially less sediment, but that it may also alter the resulting subjective psychoactive experience, perhaps nudging it slightly more in the direction of euphoric, whilst using hotter water may lean the imbiber slightly towards feelings which might be a bit closer to the soporific, although we did not follow this supposition up with pharmacological assays.

In any case, the overall chemotype of the beverage most closely approximated the parent powder when prepared in water somewhere in the temperature range of 25 to 45 °C (77 to 113 °F), although as we mentioned earlier, the observed variations to chemotype were subtle at all temperatures investigated.

Organoleptically, we found that not only did the texture change with rising squeeze temperatures (first becoming beautifully creamy as we rose through room temperature, but then thickening beyond desirability above 40 °C (104 °F)), but the taste changed too. The distinctive pepperiness of the kava coincided with the preparation temperature, becoming particularly pronounced above 30 °C (86 °F). At cold temperatures, there was no bitterness perceived at all, but it was abundantly evident by about the 35 °C (95 °F) mark, and by 42.5 °C (108.5 °F) we found the taste to be rather unpleasant. Our team described kava prepared at temperatures above this as, “nasty”, but only you can decide your own taste preferences.

We are accustomed to drinking kava at the local nakamals here in Vanuatu, where spring water, rainwater, or sometimes river water is used to prepare the kava. These all feel somewhat cool to the touch at first, but by the time they are collected and brought to the point where the squeeze takes place, they’ve usually warmed up to about the ambient temperature, which is typically around 28 °C (82.4 °F) towards the last half of the afternoon, when most kava sessions in this part of Northern Vanuatu begin to kick off.

Because of this, to us, the kava experience is most authentic when we can closely emulate what we’re used to, and this happens when we squeeze with water in the range of 25 to 30 °C (77 to 86 °F). If you want a lighter kava that some may find a little easier to drink, you can try using cooler water without worrying that you’re losing out on a significant amount of the available kavalactones, however, it needs to be said that there may be other compounds in the sediment which potentiate the experience, and we did not investigate the pharmacology of the finished products in this experiment, we just quantified the kavalactone content.

We also did not standardise the temperature of our prepared kavas before consumption, but we can tell you from experience that most people find kava easier to drink when chilled. At some nakamals in Vanuatu, or sometimes on special occasions, they will put a few bottles filled with frozen water into the serving bowl (after squeezing), thus chilling the kava without diluting it.

Many people find they enjoy being served kava this way, although it is somewhat less common in Vanuatu than simply drinking the kava warm.

On a related note, despite kava’s documented antimicrobial properties, and despite the fact that pathogens do not tend to grow in kava powder if it was processed properly, appropriately pasteurised and with the moisture content reduced to a sufficiently low concentration, once the drink has been prepared, certain bacterial species can colonise the mixture, turning it sour. This happens much more slowly when the kava is cold, so if you are creating a large batch which you intend to serve over the course of hours or longer, keeping it chilled may be worth considering.

57.89% was the maximum kavalactone extraction efficiency we were able to achieve in a single squeeze/wash, regardless of how long we massaged our strainer bag. Chemotypes of the beverages were largely unaffected by squeeze times.

Our lyophilised samples made it abundantly clear that more material had been extracted as squeeze duration increased, but despite the progressively larger amounts of sedimentation, the amount of kavalactones extracted did not continue to increase beyond a certain point.

How long is long enough?

Within just 80 seconds of squeezing (followed by strong handwringing of the strainer bag), we had already extracted an average of over 45% of the available kavalactone content, with extraction efficiencies increasing steadily up until about the 4-and-a-half-minute mark and plateauing shortly thereafter. No squeeze longer than 404 seconds (6.73 minutes) resulted in higher kavalactone concentrations in the resulting beverage, even if we massaged the strainer bag for an hour straight.

Short squeezes lasting only a few minutes gave us potent kava that was very light and easy to drink, even seeming a bit too “watered down” to our seasoned taste testers (who have become accustomed to the rich and creamy kava as served locally in Vanuatu). Longer squeezes made the mixture thicker, becoming more like nakamal style kava around the 7-minute mark and beyond, and giving us the initial subjective perception that we were drinking “seriously strong” kava, but it wasn’t typically any more abundant in kavalactones than squeezes that lasted only 5 minutes.

These results were based on squeezing 62.5 g of traditional kava powder in an R&P strainer bag in 1 L of 28 °C water, using our automated squeeze system for consistent results. 30 unique analyses were performed for these squeeze-time trials. While this sample size is relatively small, we feel it was sufficient to identify general trends. Our experiments and analyses are ongoing.

So how long should you squeeze your kava for?

That’s entirely up to you, but we find that about 5 or 6 minutes works well for us.

The blender method has a fairly substantial following, so we’ll try not to rain on that parade too heavily, and any extraction technique might suit particular preferences, but it’s unlikely anyone on our team will be putting their own kava in a blender anytime soon.

Traditionally, kava is prepared by sealing the ground plant material in a strainer bag (historically made from woven plant fibres, such as pandanus leaves or coconut husks), immersing the bag in a large bowl (natambea/tanoa) of water, and thoroughly squeezing and kneading it. The bag is then removed from the water, wrung out to extract the last drops, and the makas (used kava powder) is discarded.

An alternative method involves using a high-speed blender (which could be a food processor, smoothie maker, hand blender, or similar device) to impart high-sheer stresses to the kava in the water, simultaneously stirring and shredding it, then filtering the mixture through a strainer bag (and wringing it out afterwards).

Both methods work well enough to prepare a beverage that can impart the desirable effects we are all so familiar with, and using a blender might make kava stronger tasting, but is it really a better extraction method?

Support for the blender method:

Putting kava in a blender and then straining the resultant mixture is said to offer a potent shell of kava, and anecdotally, it has garnered some support. People have been dabbling with this for decades, but it really gained prominence back in 2015 with the CTHAR method (Gautz, Loren D., Rachel Li, and H. C. Bittenbender. 2015. Preparing Kava: Optimizing kavalactone extraction in water. Proceedings of Kava 2015 Conference, July 25-26, 2015 at Chaminade University, Honolulu, Hawaii).

CTAHR is a method proposed by Bittenbender et al., which aims to maximise kavalactone extraction. The technique calls for processing kava with water in a blender, but many people overlook the fact that this research was not conducted on commercially available traditional kava powder, which is a very different substance to what their team was working with. When it comes to blenders and kava, there have been many references to CTAHR over the years.

Bittenbender’s team investigated 8 extraction variables, with 2 possible conditions each (fresh or dry kava, 20 or 45 °C (60 or 113 °F) water, blender or hand kneaded, 60 or 120 second agitation, 1 or 3 agitation cycles, 1:3 or 1:1 kava to water dilution ratios, fine or large particle sizes, and lateral roots or rhizomes.

These variable conditions lead to 256 possible extraction combinations, however, only a single test each of just 16 unique extraction combinations were trialled to develop the CTAHR method, and only half of those experiments were performed on dry kava (as opposed to fresh green plants), and even the dry material wasn’t “traditional kava powder” like most consumers are used to seeing these days. This doesn’t leave a lot of room to recognise outlying data points, and it also requires quite a bit of inference to arrive at the “optimised” method.

This isn’t a critique of Bittenbender’s work by any means, and our experimental methods have many limitations also, but it is worth highlighting that the CTAHR method may have been reported as something quite different if more combinations had been trialled, or if the experiments were based on extracting traditional kava powders.

Even with so few experiments, ostensibly useful information can be obtained, especially with the help of some statistical analysis, and we thank Bittenbender for his contribution to the growing pool of kava knowledge – the more people experimenting (and reporting their results), the better; His team has added a tremendous amount of value to the kava scene. Nevertheless, it is not easy to isolate with confidence which factors are responsible for which outcomes when multiple variables have been changed for each extraction effort, and the trialled experimental conditions left ample room for further investigation. We’ve performed hundreds of squeeze experiments, and we still have much to learn.

What we did:

Based on commonly reported settings and durations, we tried to emulate something similar to what fans of the blender method were seemingly doing at home; We set our high-powered commercial blender at 7.7/10 (which gave us plenty of blending power without feeling like we were summoning the apocalypse) and ran it for 15 seconds, 30 seconds, 1 minute, 2 minutes, 3 minutes, and 4 minutes, blending batches of 62.5 g of traditional kava powder with 1 L of water, then pouring the mixture into an R&P strainer bag and wringing it out over a bowl. We then analysed the prepared kava by UHPLC in our state-of-the art analytical laboratory.

It should be noted that we didn’t blend continuously for 4 minutes straight, not only because we were trying to avoid significantly over-heating the kava, but we didn’t want to torture our blender either, so we intermittently paused on the longer duration experiments to let things cool off for a minute or so. The times reported represent the time spent actually blending, not the total duration the mixture resided in the blender, although that would have only added a couple extra minutes at most.

Experiments were performed in triplicate and the results were compared to what we could achieve by hand using the same kava to water ratios, trying out a few different styles and intensities of hand-kneading and squeezing (5 minutes each). Squeezing kava is pretty simple, but techniques vary, so we thought we’d take that into account, just in case it made a difference (it can).

The results:

We were underwhelmed by the results of the blender method, both in terms of taste, and perhaps more surprisingly, in terms of kavalactone extraction efficiency.

Our subjective findings:

We found that after 15 seconds of blending, the kava tasted “fine”, but slightly bitter. Bitterness, along with darkening of the beverage, was proportional to the time spent blending, and after 1 minute in the blender, we found the kava became much more bitter. Some might still find it quite tolerable, but we’re pretty fussy about our kava, and we didn’t particularly care for it. After 2 minutes, it became quite warm, much darker, and fairly unenjoyable to drink. After 3 minutes, the kava was very dark, definitely unpleasant to drink (for us), and quite a bit hotter. After 4 minutes of blending, the most common word used to describe the taste was, “gross”, and it also became rather spicey, which we didn’t find meshed well with our taste buds, but some might enjoy the more peppery version of kava that 4 minutes of blending can provide (although getting past the substantial bitterness would likely be difficult for most).

There is no doubt that hand squeezing, regardless of technique or intensity, provided for a beverage that was more enjoyable to consume than anything we could produce with the blender.

Kavalactone concentrations and ratios:

When we averaged out all of our various hand-squeeze efforts and compared those values to the average we could get from blending, we found the blender extracted only 85.92% of the kavalactones that could be extracted on average by hand. That said, the kavalactone extraction efficiency of the blender increased as more time was spent blending, and the longest duration blend was finally able to pull about as much kavalactone content into the drink as a half-decent hand-squeeze, however, powering through the taste and texture would likely rule out all but the most iron-mouthed kava lovers from blending for 4 minutes.

At 3 minutes or less, the blender resulted in lower kavalactone content in the drink than any of our hand-squeeze tests, except for one. The one exception was the gentlest squeeze we could possibly do and still say it was a squeeze rather than a steep – not really what most would consider a squeeze at all, kind of flopping the bag over in the bowl and softly prodding it with the fingers. Every other squeeze we tried, from “mild” to “aggressive” (think Conan the Barbarian meets Attack of the Giant Anaconda), resulted in stronger kava than using the blender.

On a slight side note, we found that a moderate squeeze was best. A massively intense squeeze extracted marginally (about 3%) more kavalactones than a relaxing “making dough” squeeze, but the more intense the squeeze, the more bitter the kava became. It required substantially more effort to get those meager kavalactone gains, at the expense of taste. Also, although most kavalactones are extracted with relative consistency independent of squeeze exertion, the more intense the squeeze, the lower the K:DHM and the lower the K:FK ratio became – nudging those tudei-esque compounds up, and our beloved kavain down, relative to the other kavalactones.

Blending also did a much better job at pulling out the “less desirable” compounds than hand squeezing. The extraction efficiency of flavokavains increased slightly in proportion to hand-squeeze intensity, but it rose considerably with the blender, pulling on average 117.48% of the flavokavains compared to hand squeezing, and noticeably less kavain (around 7% less, depending on duration), relative to the other kavalactones. Blending managed to pull out a bit more dihydromethysticin than hand-squeezing too (110.44% compared to hand-squeezing, on average), so if you’ve ever felt any residue from last night’s kava session the following morning, you’ve got even more reason to avoid the blender method.

Traditional kava powder has quite a bit of indigestible fibre, and anyone who’s tried kava without straining the makas out (spilled the bag in the bowl, and didn’t bother to filter it, maybe?) has probably told you it’s rather unpleasant to consume. When the particle size is reduced, as happens in a blender, more of this material can make its way through the strainer bag, and that’s something people with a sensitive stomach may want to avoid as well.

Let’s wrap it up:

In our controlled laboratory experiments, blending adversely impacted taste, shifted the chemotype in an undesirable way (for most), imparted more fibre to the drink, the noise wasn’t particularly conducive to relaxation, and the cleanup was more than enough hassle to outweigh any efforts saved over a traditional squeeze (we didn’t find blending to be easier, quicker, or less effort, but some folks might). We’ll add to this that our production team uses a lot of blades around here, and processing a root as tough as kava is terribly hard on equipment; If you’re committed to the blender, be prepared for its life to be shortened significantly.

All these sacrifices could be worth it to a handful of people when weighed against potential improvements in kavalactone extraction efficiency, but that would require significant gains, which we didn’t see. In fact, even a half-hearted squeeze was enough to pull about the same amount of kavalactones into the drink as 3 minutes in our blender, and the hand-squeezed kava always tasted much better.

For people who are unable to perform a traditional kava squeeze, or people who prefer their kava a bit on the goopier, hotter, or more bitter side, with a little more indigestible fibrous content, and the chemotype shifted slightly closer to that of tudei kava, the blender method might be great, but for us, it’s a hard no.

We are not aware of any indigenous cultures who historically added fats during kava preparation. Our controlled experiments found no benefits to adding fats during the squeeze, and some potential downsides. After thorough examination of the data, we’ll continue to stick with plain water.

Some background:

Ni Vanuatuan peoples have been drinking kava prepared with plain water for thousands of years, a method deeply rooted in their culture and tradition. On very rare occasions we have been told of uncommon instances where coconut water may have been added, although we’ve never seen it firsthand at a nakamal or in any village we’ve visited, and it isn’t clear if this was added during or after squeezing, or why. In days of yore, their kava was strained through woven plant materials such as pandanus or banana leaves, coconut palm fibres, bark cloth, or sometimes through compound-containing materials such as hibiscus bark, but this is very atypical nowadays, and they never add milk, cream, fats, or oils to their kava, even though these are all available to them. Although there are invariably many untapped improvements to any given process, we try to learn from those who came before us, especially when in doubt, and we thank the people of Vanuatu for extending their knowledge of kava to us.

Over 200 compounds have been isolated from kava, but it is the 6 major kavalactones which are believed to be responsible for the overwhelming majority of its desirable effects. These kavalactones are primarily produced by epithelial cells lining the resin ducts, which are abundant in the parenchyma tissue of the lateral roots and rhizomes (also known as basal roots or “stumps”) of Piper methysticum. While the lateral roots are more potent in their effects, they are more difficult to process, particularly in harvesting and peeling, and they impart undesirable flavours to the drink. Therefore, it is primarily the underground stumps that are used to prepare kava for consumption at nakamals in Vanuatu. Kavalactones are lipophilic molecules, meaning they dissolve readily in fats and poorly in water. A number of studies have assessed the partition coefficients of kavalactones, demonstrating that they favour organic phases over aqueous ones. Organic solvents have also proven to extract kavalactones more efficiently than water.

Thus, it seems logical that some people might believe adding fats to the squeeze during kava preparation would improve kavalactone extraction efficiency when using a strainer bag. In our trials, we found this was not the case.

Support for the blender method:

Putting kava in a blender and then straining the resultant mixture is said to offer a potent shell of kava, and anecdotally, it has garnered some support. People have been dabbling with this for decades, but it really gained prominence back in 2015 with the CTHAR method (Gautz, Loren D., Rachel Li, and H. C. Bittenbender. 2015. Preparing Kava: Optimizing kavalactone extraction in water. Proceedings of Kava 2015 Conference, July 25-26, 2015 at Chaminade University, Honolulu, Hawaii).

CTAHR is a method proposed by Bittenbender et al., which aims to maximise kavalactone extraction. The technique calls for processing kava with water in a blender, but many people overlook the fact that this research was not conducted on commercially available traditional kava powder, which is a very different substance to what their team was working with. When it comes to blenders and kava, there have been many references to CTAHR over the years.

Bittenbender’s team investigated 8 extraction variables, with 2 possible conditions each (fresh or dry kava, 20 or 45 °C (60 or 113 °F) water, blender or hand kneaded, 60 or 120 second agitation, 1 or 3 agitation cycles, 1:3 or 1:1 kava to water dilution ratios, fine or large particle sizes, and lateral roots or rhizomes.

These variable conditions lead to 256 possible extraction combinations, however, only a single test each of just 16 unique extraction combinations were trialled to develop the CTAHR method, and only half of those experiments were performed on dry kava (as opposed to fresh green plants), and even the dry material wasn’t “traditional kava powder” like most consumers are used to seeing these days. This doesn’t leave a lot of room to recognise outlying data points, and it also requires quite a bit of inference to arrive at the “optimised” method.

This isn’t a critique of Bittenbender’s work by any means, and our experimental methods have many limitations also, but it is worth highlighting that the CTAHR method may have been reported as something quite different if more combinations had been trialled, or if the experiments were based on extracting traditional kava powders.

Even with so few experiments, ostensibly useful information can be obtained, especially with the help of some statistical analysis, and we thank Bittenbender for his contribution to the growing pool of kava knowledge – the more people experimenting (and reporting their results), the better; His team has added a tremendous amount of value to the kava scene. Nevertheless, it is not easy to isolate with confidence which factors are responsible for which outcomes when multiple variables have been changed for each extraction effort, and the trialled experimental conditions left ample room for further investigation. We’ve performed hundreds of squeeze experiments, and we still have much to learn.

Our results:

Out of 36 unique samples of kava prepared with any kind of fats/oil/milk added during squeezing, when analysed by UHPLC, none showed statistically significant higher kavalactone content than kava powder squeezed with water alone. Interestingly, almost all fatty additives resulted in lower total kavalactone extraction efficiency, decreasing the total amount of kavalactones extracted by up to 17%, and by 9% on average, compared to kava prepared using water alone.

We did not investigate the mechanism for the observed decrease in extraction efficiency, so it’s anyone’s guess at this stage whether it can be attributed to oils binding to the kava powder and preventing some particles from being released, fats clogging the pores of the strainer bag to some extent, or something entirely different. There was an increase in the total amount of material in some of the lyophilised samples of supernatant after centrifuging, but fats weren’t the secret ingredient to extracting more kavalactones in our tests, and the excess material was comprised primarily of inactive constituents, or components of the additives themselves. The chemotypes of the beverages also did not appear to be influenced by adding oil-containing products to the squeeze.

Whether it was whole dairy milk, almond milk, soy milk, olive oil, coconut milk, coconut cream, or something else, and whether it constituted just 0.3% or up to 10% of the total liquid volume, we found none of the resulting beverages to be more concentrated in kavalactones. We did not investigate emulsifiers, partially because they may be implicated in leaky gut syndrome (although this is outside our area of expertise), but also because we couldn’t find any at the limited markets available nearby when we decided to perform these experiments, and we didn’t want to wait to order them in from overseas before jumping in the deep end with this one.

When we centrifuged our samples to isolate the sediment from the supernatant (the “liquid layer”), we saw that some of these additives influenced how the extracted kavalactones were partitioned in the beverage; Very oily compounds, such as dairy milk, coconut milk, and olive oil all shifted the supernatant towards higher kavalactone concentrations, sometimes close to doubling the amount of kavalactone content normally found outside the sediment, however, the overwhelming majority of kavalactone content still resided in the sediment, and the total kavalactone content of the prepared kava remained unimproved, regardless of type or quantity of fat.

We found that if the extraction water contained less than 1% coconut milk or olive oil (by volume), the resulting kava was still enjoyable. Outside of these 2 exceptions, however, using virtually any amount of almost any kind of milk or fat during the squeeze substantially increased the bitterness. When large amounts (10% of the total liquid volume) of fat-containing additives were used, our team found the overall taste became much worse than when prepared with water alone, and for many additives just 1% was enough to ruin the taste of the kava for us.

We did not investigate how adding these substances to already prepared kava might influence the flavour, or how they may have influenced the taste of other kavas prepared using different methods, and we acknowledge that everyone’s taste preferences vary. Consider doing a side-by-side comparison if in doubt – the perception of taste can change with environmental conditions and a person’s physiology at any given time, but we suspect most people would easily detect an increase in bitterness when milk or other fat-containing additives are added to the squeeze.

Unsurprisingly, even small amounts of oils made for slippery gloves, and larger amounts resulted in more hassle during cleanup. When true oils comprised 0.3 – 1% of the total liquid volume, there was an oily texture to the kava, but it still appeared homogenous, with no obvious oil floating on top. At 3% oil content and above, the surface had noticeable oil separating from the mixture, even after thorough squeezing, and cleaning our automated squeezing machine became a real pain.

We did not investigate absorption or other pharmacological attributes; We only quantified the kavalactone content of the kava, but it should be noted that snacks are often available at nakamals, and some locals enjoy small nibbles of finger food after a shell. Generally, they say it is to cleanse the palate, not to potentiate the effects, but anecdotally some people do report that the kava “kicks” after they follow up their shell(s) with something to eat. As far as getting more kavalactones from the powder into your shell goes though, based on our research, adding fats during squeezing isn’t likely going to help.

Traditional methods, refined over eons by the Ni Vanuatuan peoples, remain the gold standard for a reason. In our view, this study reinforces the wisdom of sticking to plain water for the kava squeeze. If you enjoy adding fats (or anything else) to your kava, don’t let us stop you! Taste cannot be disputed, and we all have our own preferences.

We’ve seen comments online suggesting that kava may be stronger if prepared the evening beforehand. Others have speculated that the chemotype shifts, potentially altering the experience. The results of our investigation did not support these postulations.

Experimental conditions:

We prepared traditional kava powder using 28 °C (82.4°F) water, kneading it for 5 minutes in an R&P strainer bag within our automated squeeze machine, then transferred it to our natambea (tanoa) and let it sit uncovered at room temperature in our well-lit laboratory. We gave it a stir and collected a small sample every 15 minutes for the first few hours, then half-hourly, then hourly, then twice daily, regularly testing the kava for a week in total. After the first 24 hours, we transferred it from the natambea into a sterile Schott bottle, which we sealed and kept in the fridge, opening it only to collect aliquots after giving it a good shake. We tested the kava over the course of a week, then scrutinised the UHPLC data.

Our results:

No significant changes in kavalactone content or chemotype were observed throughout the study. The kavalactone profile remained stable at all time points, suggesting that kava’s strength and chemotype do not degrade or shift under the conditions tested.

Kavalactone degradation discussion:

Despite rigorous analysis, there weren’t even subtle variations in kavalactone content of noteworthy mention, countering the idea that letting kava sit overnight (or longer) is likely to enhance or alter its effects. With that in mind, our study focused solely on kavalactone stability, not other factors like microbial growth, pH changes, or other differences which may potentially alter the experience. Although these other aspects could still have an influence, kavalactones have always been hailed as kava’s most important constituents (in terms of psychoactivity), and we can now confirm that they’ll likely be unchanged between the time you squeeze your kava and the time you down your shell.

Why share “boring” results?

Even when “nothing happened,” sharing null results is crucial for scientific progress. Documenting stable outcomes helps confirm the reliability of previous findings and directs future research away from unproductive paths. Including null results in the scientific record also contributes to addressing the replication crisis, ensuring that our understanding of kava is as accurate and balanced as possible.

While many journals and reviewers tend to favour positive or novel results, we believe that all findings, including null results, are valuable. Thank you for supporting our ongoing research into the kava squeeze. We’ll continue to share our findings, whether they’re surprising or not!

These were the experimental conditions for our multiple wash investigations, and for the most part they apply to our other experiments too:

For consistency in technique, we used an automated system (essentially a glorified portable washing machine) to gently squeeze our kava in 28 °C (82.4°F) water, for 5 minutes per squeeze cycle, followed by strong hand-wringing of the strainer bag between each successive squeeze.

We used 62.5 grams (about 2.2 ounces, for our American friends) of kava in 1 L (about 1.06 US liquid quarts) of water, collecting a sample and draining the liquid after each squeeze cycle, then cleaning our squeeze machine, returning to it the wrung-out strainer bag with the partially extracted kava powder still inside, and replacing the liquid with 1 L of fresh water for each subsequent squeeze (for a total of 8 liters of prepared kava, which made for some seriously dilute drinks by the end of our experiment).

We conducted our experiments using a traditional grind kava powder, derived from a blend of fresh green Vanuatu noble cultivars with a net chemotype of 423165 and containing 6.661% kavalactone content by weight (including the moisture content of the powder). Note that we report our kavalactone concentrations based upon the powder as packaged, rather than on the dried weight. Although this may be somewhat uncommon in the industry and results in reporting lower kavalactone levels, we feel it is a more accurate reflection of the powder as used, and it makes it easier for people to understand what they’re really working with.

Almost all water has impurities (or additives such as chlorine), and these can alter both the kavalactone content and the taste of the finished beverage. When we prepare kava for ourselves to drink, we usually use filtered rainwater, but we were looking to minimise variables in these tests, so our experiments were conducted using ultrapure water (measured at 18.2 MΩ of resistance).

Our samples were weighed to within 100 µg (a microgram is one millionth of a gram) on analytical balances calibrated with certified class OIML E2 weights with uncertainty +/- 0.000016 g (NATA accredited for compliance with ISO/IEC 17025, by laboratory No.3279), and all of the other instrumentation used for these experiments was also modern, fit for purpose, and well cared for, even down to our pipettes, which are serviced and calibrated by Eppendorf to ISO 8655-6:2022.

Kavalactone concentrations were analysed by qualified experts on our Thermo Scientific Vanquish Horizon Ultra-High-Performance Liquid-Chromatography system, comprised of VF-A10-A Split Sampler, VF-P10-A Binary Pump, VFD11-A Diode Array Detector, and VH-C10-A Column Compartment, fitted with a 200 x 2.1 mm Hypersil GOLD, 1.9 µm particle size column, running a gradient of isopropanol and water as the eluent, at 60 °C, with active pre-heating.

UV detection was at 362, 341, 246, and 218 nm, with peak identification assisted by elution time and spectrum matching, and relative quantification calculations were based on peak areas at 246 nm.

Correlation coefficients for all identified compounds were greater than 99.995% on a 20-point calibration curve derived by serial dilution of ampoules of Cerilliant certified analytical reference standards. We don’t want to toot our own horns too loudly, but this kind of precision is not possible without both high quality apparatus and exceedingly skillful staff.

We used Chromeleon 7.2.10 software to operate our UHPLC, and Chromeleon 7.3.2 to process the data. Our lower and upper confidence probabilities were 99.5%.