The Annona fruits are oft-staples throughout the neotropics. Every rural house in Puerto Rico, where I am currently compiling this, has a soursop or three. Annona squamosa & Annona reticulata are slightly less common. These fruits are planted widely, are varying degrees of domesticated, and are cherished by the peoples who grow them. Let us start by saying that yes, they all have some form & degree of annonaceous acetogenin (a potent neurotoxin — in a lab). Annonacin, bullaticin, squamocin, or one of the countless novel ones. This is a wide phytochemical family present throughout the Annonaceae — from the seed spices of West Africa (Xylopia, Monodora), to the Rollinias & Duguetias of Central America the Pond Apples of Florida & the Caribbean, to the autumn pawpaws of Ohio.

If you’ve been around the internet in fruit-growing circles you will have probably noticed a fair amount of people who readily warn you how these fruits are toxic, or potentially so. When you approach these fruits with a fear-based kilter, you unknowingly invoke the entire family Annonaceae (and Lauraceae too, which also contains acetogenins). This is problematic as humans have been eating from these families safely for a very long time. Words like "cytotoxic" and "neurotoxic" inspire fear, and fear is bias. Some people eat avocadoes, which contain cardiotoxic acetogenins, on a daily basis. So why don't they get sick? The answer, at least partly, is in how humans digest things.

I’m writing this to help address an important cultural discrepancy — and to show how people, piloted often by fear, have been misinformed by thier own biases. There’s a deep-seated fear of and distrust of the indigenous, the exotic, and it sometimes presents in the form of heightened, verging on obsessive "scientific" scrutiny with regard to food & food purity. This is especially prevalent in the western, English-speaking world. I have been keeping an eye on this “discussion” - in quotes as it scarcely changes - since 2008 or so when I became aware of it under the pretext of cherimoya and soursop (Annona cherimola & A. muricata respectively). It has since moved along into other Annonaceae fruits, including our beloved temperate pawpaws (Asimina triloba), which may contain 6-8x as much acetogenin (annonacin specifically) as a soursop. There are plenty of people afraid to eat tropical soursops, but the argument with pawpaws goes something like “If soursops made people sick at X amount, then pawpaws surely will at 8x that amount!?” Nature, of course, is hardly ever so straightforward. For now, let’s not get into the West African seed spices, that’s a no man’s land.

Awhile back, there was a publication referred to as “The Guadeloupe Study” (https://pubmed.ncbi.nlm.nih.gov/10440304/) which implicated consumption of decocted soursop leaf matter in neurodegeneration, observed in an isolated population of the French Guadeloupe. Then came another study from New Caledonia, with similar observations. The consumption was mainly of leaf tea, believed to provide health benefits and/or cancer treatment. Like many individual studies it has it’s limitations, and people who raise fear about the subject appear to be reading too far into it, into things the study never really said. Firstly, cooked leaf matter is an unconventional use of the plant - this is not how the plant has been consumed typically. Second, variable control — what were the other dietary, environmental, genetic factors involved? (Further studies have implicated a genetic factor to be likely involved. This is important to acknowledge - it may be a situation similar to celiac disease.) Third, were the participants drinking alcohol? Acetogenins are soluble in alcohol — not in water. And so on. The original study is a broad epidemiological observation of the effects of eating soursop leaves as tea, which identifies annonacin (*or it's immediate phytochemical family) as the culprit.

Note how the Guadeloupe Study does not concern consumption of the fresh fruit, or of heating / cooking the ingested fresh fruit material. That there is assumption being made based on the mere presence of said toxin, regardless of the form or part of plant consumed. Constituents of leaves and bark and fruit can vary, cooking changes chemistry, sometimes profoundly. Note also that there have been several studies since clarifying the activity of these compounds in living organism, not just in a laboratory. We don’t need comparatively vague epidemiology to understand the biochemical activity.

The absorption factors for acetogenins like annonacin are probably around 3%. Yes, it was measured, in rodents at least (this is a moderately reliable measure for humans, generalist). Despite how much neurotoxin the fruits may contain, only a small fraction is ultimately absorbed into the blood (https://www.thieme-connect.com/products/ejournals/abstract/10.1055/s-0034-1394993). In humans this may be an even lower number, since their destruction takes place during the gastric phase. There is also the question of how humans metabolize it, which may differ from rodents. The cited study on bioavailability above came as a result of the Guadeloupe study, after. It's important to note also that the French Government (Agence française de sécurité sanitaire des produits de santé) did not conclude that consuming soursop leaf tea was harmful in and of itself, that there was more to it. They aren’t idiots, nor are they being negligent.

”But it made people sick!”
A quick google will get you the approximate problematic annonacin dosage (one example -- 3.8 and 7.6 mg per kg per day for 28 days; from Wiki FWIW). This is something we know.

And how much each pawpaw fruit contains (or, in general, it's 6-8x that of soursop, circa 100-120mg) (bear in mind these numbers are tentative; inexact; some studies may put it a bit higher)

With an absorption/bioavailability circa 3% (meaning low, generally) we aren't anywhere near the problem "dosages," at least in amounts people normally would eat-- in fact, much of the pharmaceutical study has been on how to increase the low absorption, via nanocarriers and such. There's some hope that it will be cancer medicine.

Basically, this doesn’t add up in a basic way — it shouldn’t be making people sick. This invokes genetics, diet and/or environment.

It probably means there's some other variable(s) influencing the neurotoxicity first observed and documented in Guadeloupe and New Caledonia. Some part of it may the unconventional use of the plant (leaf tea) - but it also suggests there's some dietary, genetic, environmental element that's making people more susceptible -- either by increasing exposure, bioavailability, or by making people more sensitive to it. https://academic.oup.com/brain/article/130/3/816/277881

"About 50% of the atypical patients were not heavy consumers of annonaceaous products, suggesting that other factors, environmental or genetic, might affect the vulnerability of patients to the neurotoxins in Annonaceae."

The studies even pointed this out, too.

The less charitable among us have missed this part of it; instead they have conflated in-vitro toxicity (in a lab) with in-vivo digestion/absorption (in an organism). Many seem to flatly regard what you eat as the "dosage" and spread around a basic misunderstanding of human digestion, of science generally. Even the Wikipedia on soursop was guilty of this (until recently — it has been revised).

Moreover, and on principle, indigenous peoples aren't & weren't a bunch of ignoramuses poisoning themselves over & over during the domestication of all these different custard apples, which are routine fruits all over the neotropics. These implied attitudes, and the quick-to-quarrel sort of people who seem to find them are exhibiting a kind of anti-indigenous perspective, which is disheartening. This is the unfortunate cultural element I’m talking about, rooted in the separation of modern, insulated suburbanites from thier own past. Dare I point out how it’s mostly a provincial overconcern of temperate-climate-dwelling English-speaking people? (The very people least familiar with eating & everyday use of said fruits). This is an observed aspect, not an overriding point. From ptaquilicide in bracken ferns, to solanaceous toxins of the nightshade family, Indigenous food fearmongering has a long history as well.

Sociological points aside, it’s important to note that these fruits have a long history of fresh consumption, and also to realize the myriad of variables that come along with discussions regarding toxicity. It appears that temperate suburbanites do not grasp how often and widely soursops are eaten pantropically — it’s a very common dooryard and market fruit everywhere it can be grown. Also note, these studies come out of isolated populations, it has not been observed to a degree relative to the popularity or general amount of fruit consumed. We would certainly see far more reports of this clinically if it were a straightforward 1:1 phenomenon. The concept that “soursops are toxic” (therefore pawpaws must be too) is itself fundamentally flawed — and I will argue that all of this points towards some additional component(s) in the mix. We must also contend with the lack of reports from widely-eaten fruits — Biriba, Sugar Apple, Corazon — which contain novel acetogenins related to, but notably more potent than annonacin.

Among the tropical semi-domesticate Annona species (which have been profiled, anyhow) many contain acetogenins that are more potent -- annonacin is just the most accounted for. Bullaticin is the most obvious example (Annona reticulata). Biriba fruit (Rollinia mucosa), Annona crassifolia, A. senegalensis, are some other examples -- these species' principal acetogenins are significantly more potent too (cell lines depending, making such comparisons is of course difficult). Annona squamosa, the common and more domesticated sugar apple — also implicated for it’s phytotoxin, squamocin. This idea that pawpaws can be singled out for a heightened toxicity is itself based in an inadequate comparison that does not account for the larger phytochemistry of the family. I have to argue that acetogenins of varied potencies and concentrations are characteristic of the Annonaceae family itself, it is important to regard this in its context.

Rolliniastatins — in Rollinia mucosa
Bullaticin — in Annona reticulata
Squamocin — in Annona squamosa
…and it goes on like this.

Note that all have seen more domestication than the pawpaw, too. This is a phytochemical family characterized by annonacin, not just a single toxin.

Note how the “discussion” mostly surrounds annonacin — when we have semi-domesticated Annona fruits with even more potent acetogenins in them. Are we going to purvey fear about every new acetogenin that’s detected? Just the ones we know about? At what point might we acknowledge the scope & context has maybe been misunderstood? I maintain the point that these these are indigenous foods, not random wild fruits.

Bear in mind that none of this is intended to express unconcern with public health — it is an attempt to give people more context, which this whole “discussion” desperately needs. The past few years have been a lesson, for me, in just how unfamiliar the temperate population is with these fruits — and how much fear comes out of that unfamiliarity.

In the wider context of nature there are many other fruits with similar (or different) toxicities (underripe commercial lychees? Ackee fruit?), of which we are simply unaware, or at which we barely bat an eye. After all, there is a hint of risk in everything, no? Some of us even seek out a potent neurotoxin, ethanol, on a daily basis.

The Annonas & company are no exception. Enjoy your fruits.

—- UPDATE (9/2025) —-

For more information on the neurological aspects of acetogenins, and to better understand why these fruits are not toxic in amounts humans would normally eat, please view Dr. Loren Kirshner’s talk a the 5th International Pawpaw Conference, which corroborates in more detail much of what has been previously said in this article.