Soursop Safety: Annonacin and Neurotoxicity Research

Table of Contents

1. What Annonacin Actually Is
2. Complex I, and Why Neurons Are the Vulnerable Cell
3. The Guadeloupe Story, in Order
4. What the Laboratory Work Showed
5. What Is Genuinely Uncertain
6. Where the Acetogenins Are: Leaf, Seed, Bark, Pulp
7. How Much Reaches a Cup, a Fruit or a Can?
8. Why the Concern Is Cumulative, Not Acute
9. Dose and Duration: What Can and Cannot Be Said
10. Capsules, Extracts and Cancer Marketing
11. Who the Research Raises Concerns For
12. Drug and Condition Interactions
13. What the Guadeloupe Syndrome Looked Like
14. What Is Known for Long-Term Drinkers
15. Key Research Papers
16. Connections
17. Featured Videos

What Annonacin Actually Is

Annonaceous acetogenins are a chemical family found in the custard-apple family, Annonaceae. Structurally they are long fatty-acid-derived chains carrying one or more tetrahydrofuran rings and a lactone ring at one end. A 2015 review counted more than 100 acetogenins isolated from the leaves, bark, seeds, roots and fruit of Annona muricata, and described them as the plant's major constituents. Annonacin is the major acetogenin of A. muricata; related ones include squamocin and bullatacin.

Three properties are central to the safety question:

• They are lipophilic. They dissolve in fat far better than in water. In the 2004 rat study, annonacin given into the bloodstream was detected inside brain tissue.
• They are potent. In rat midbrain cultures, annonacin killed half of the dopaminergic neurons within 24 hours at about 0.018 micromolar — far more potent than MPP⁺ (1.9 micromolar) and about as potent as the pesticide rotenone (0.034 micromolar).
• They are present in prepared foods and drinks. Annonacin has been measured in leaf infusions and decoctions and in canned commercial soursop nectar, and acetogenins have been identified in commercial soursop juice and in a commercial alcoholic drink made from cherimoya fruit.

The same chemistry has agricultural uses: a commercial bioinsecticide formulated from Annona squamosa extract with the acetogenin annonin as its main active ingredient killed stored-grain insect pests in laboratory tests. Custard-apple seeds are also used in folk medicine against head lice, although a 2006 laboratory analysis attributed that anti-lice activity to oleic acid and an oleate triglyceride in the seed oil rather than to acetogenins. Seeds can harm people on contact too: a 2016 case series described six patients who developed toxic keratoconjunctivitis within 6–12 hours of custard-apple seed material getting into the eye.

Complex I, and Why Neurons Are the Vulnerable Cell

Most of a neuron's ATP is made in mitochondria by the electron transport chain, and the entry point to that chain is complex I — formally NADH:ubiquinone oxidoreductase. It accepts electrons from NADH and passes them onward, pumping protons across the inner mitochondrial membrane in the process. That proton gradient is what ATP synthase later converts into chemical energy.

Acetogenins inhibit complex I. In the 2004 rat study, annonacin lowered brain ATP by 44%. In cultured neurons, raising glucose or mannose — which lets cells make ATP by glycolysis instead — largely restored ATP and prevented the neurons from dying, which ties the toxicity to energy failure.

Dopaminergic neurons of the substantia nigra are considered especially vulnerable. They have long, very highly branched axons that are largely unmyelinated, which makes them energetically expensive to run; they handle dopamine, whose metabolism generates oxidative stress; and they do not divide, so a lost neuron is not replaced.

The main reason researchers took the soursop question seriously is that other complex I inhibitors have produced parkinsonism:

• MPTP. In 1983, Langston and colleagues reported four people who developed marked parkinsonism after injecting an illicit drug made in an attempt to synthesise a meperidine analogue; the substance was mainly MPTP. MPTP's metabolite MPP⁺ inhibits complex I. This is the clearest human evidence that a chemical of this class can cause parkinsonism.
• Rotenone. A plant-derived pesticide and complex I inhibitor. Given chronically to rats, it caused selective nigrostriatal degeneration with α-synuclein-containing inclusions (Betarbet 2000). In a case-control study nested in the US Agricultural Health Study, rotenone use was associated with Parkinson's disease (odds ratio 2.5; Tanner 2011).

Annonacin belongs to the same pharmacological category. Unlike MPP⁺, but like rotenone, it also reduced the survival of non-dopaminergic neurons in culture.

The Guadeloupe Story, in Order

Told in sequence, the record is an unresolved but serious signal rather than either a proven cause or a myth.

1. The clinical observation

Through the 1990s, neurologists on the French Caribbean island of Guadeloupe saw an unusually high proportion of parkinsonian patients who did not have ordinary Parkinson's disease. In the 1999 series of 87 consecutive patients, 22 had Parkinson's disease, 31 had progressive supranuclear palsy (PSP) and 30 had atypical parkinsonism. The atypical patients had symmetrical rigidity and slowness; the PSP patients had vertical down-gaze palsy, severe gait and balance problems and a frontal-lobe syndrome.

The treatment response was a defining detail. In the 2002 series, three-quarters of the patients did not respond to levodopa — the drug that reliably helps true Parkinson's disease. A later account of 265 patients put atypical parkinsonism at 75% of cases and Parkinson's disease at 25%, the reverse of the proportions seen in Europe.

2. The case-control study

In 1999, Caparros-Lefebvre and Elbaz, for the Caribbean Parkinsonism Study Group, published a case-control study in The Lancet. Patients with PSP or atypical parkinsonism reported significantly more regular consumption of Annonaceae fruit and herbal tea than patients with Parkinson's disease (odds ratios 23.6 for fruit and 28.2 for herbal tea) and than hospital controls with benign symptoms (20.7 and 6.48). At that time the authors suspected the plants' benzyltetrahydroisoquinoline alkaloids; the focus moved to acetogenins in later work. A 2002 cell study from the same group found that an alkaloid extract of soursop root bark was also toxic to dopaminergic neurons.

3. The pathology

A case-control association based on dietary recall is a weak instrument on its own. In 2002, Caparros-Lefebvre, Sergeant, Lees and colleagues reported 220 consecutive patients in Brain and the brains of three PSP-like patients who came to autopsy. All three showed an accumulation of tau protein predominating in the midbrain, and the authors concluded that Guadeloupean parkinsonism may be a tauopathy identical or closely related to PSP. That gave the syndrome a molecular signature that a candidate toxin would need to reproduce.

4. The follow-up phenotyping

In 2007, Lannuzel, Höglinger, Verhaeghe and colleagues published a cross-sectional study of 160 parkinsonian patients in Brain. They classified 31% as Parkinson's disease, 32% as a PSP-like syndrome and 31% as a parkinsonism–dementia complex. Consumption of soursop was significantly greater in both atypical groups than in controls and in patients with Parkinson's disease, and the authors suggested the two atypical forms might be a single disease with a shared exposure to annonaceous neurotoxins, notably annonacin.

5. Later Caribbean and Pacific reports

• New Caledonia (2004). Two letters in Movement Disorders reported atypical parkinsonism in New Caledonia in connection with Annonaceae consumption and compared it with Guadeloupe. Neither letter has an abstract in PubMed, so their case details are not restated here.
• Severity in the French Caribbean (2022). Among 180 Caribbean patients with degenerative parkinsonism, including Parkinson's disease, lifelong Annonaceae consumption was associated with the more severe cluster of symptoms and cognitive deficits (odds ratio 3.76 for fruit or juice above 0.2 fruit-years; 2.91 for any herbal tea). The authors called for more restrictive public-health recommendations.
• French Guiana (2025). Of 372 patients, atypical syndromes made up 41.8% in French Guiana against 66.2% in the French West Indies; Annonaceae consumption was reported by 79.2% and 93% respectively.
• Caribbean versus mainland France (2026). In a comparison of 74 Caribbean and 104 mainland-French Parkinson's patients, 71 of 74 Caribbean patients but none of 20 mainland patients interviewed had eaten Annonaceae fruits, and the Caribbean group had more severe motor and cognitive scores after adjustment.
• Brain tissue (2025). In eight autopsied Caribbean patients, five had combined tau and α-synuclein pathology; in test-tube assays annonacin increased α-synuclein aggregation.

6. The risk-assessment response

A 2020 risk assessment of Annona muricata food supplements, carried out at Germany's Federal Institute for Risk Assessment within an EFSA-supported fellowship programme, found indications of neurotoxic potential in certain preparations, substantial uncertainty about safe use, and too few studies — particularly of long-term use — to establish a safe intake level. France's food-safety agency, today ANSES, has also examined Annonaceae products; the wording of its opinion could not be checked for this page, so its conclusions are not restated here.

What the Laboratory Work Showed

The laboratory work arrived in four steps, each addressing a gap in the argument.

1. Annonacin kills dopaminergic neurons by starving them of energy. Lannuzel, Michel, Höglinger and colleagues reported in Neuroscience in 2003 that annonacin killed rat mesencephalic dopaminergic neurons in culture, that longer incubation caused toxicity at lower concentrations, and that extra glucose or mannose restored ATP and prevented cell death. Withdrawing the toxin after a short exposure (under 6 hours) stopped the cells dying. Evidence tier: preliminary, cell culture.
2. Annonacin produces lesions in a living animal. Champy, Höglinger, Féger and colleagues reported in the Journal of Neurochemistry in 2004 that rats given annonacin intravenously by minipump for 28 days lost 31.7% of their dopaminergic neurons in the substantia nigra, along with cholinergic and GABAergic neurons in the striatum, without evident systemic toxicity. The distribution of lesions resembled that in the patients. Evidence tier: preliminary, animal.
3. Annonacin produces a tau change, not just cell death. Escobar-Khondiker, Höllerhage and colleagues reported in the Journal of Neuroscience in 2007 that annonacin lowered ATP in cultured rat striatal neurons and redistributed tau from the axons to the cell body. Höllerhage, Matusch, Champy and colleagues reported in Experimental Neurology in 2009 that 24 lipophilic complex I inhibitors all caused this tau redistribution, in proportion to how strongly they inhibited complex I, and proposed such compounds as candidate toxins for sporadic tauopathies. In a 2016 mouse study, 12 months of drinking commercial soursop juice increased phosphorylated tau in the brain, most clearly in mice carrying a mutant human tau gene. Evidence tier: preliminary, cell culture and animal.
4. The quantities in food. Champy and colleagues reported in Movement Disorders in 2005 the annonacin content of soursop fruit, canned nectar and leaf infusions (figures in the next sections), and estimated that an adult who consumes one fruit or one can of nectar a day would ingest, over a year, the amount of annonacin that induced brain lesions in the rats given purified annonacin intravenously. That estimate is the bridge between the epidemiology and the toxicology, and it is also the most contestable step, because it extrapolates across species and from an intravenous to an oral route.

Read together, the laboratory work shows that annonacin is a neurotoxin in cell and animal models, with a characterised mechanism, an anatomical target and a tau change resembling the human syndrome. Whether it caused the Guadeloupean cases is a separate question.

What Is Genuinely Uncertain

The research supports describing the hypothesis as plausible, biologically supported and unproven. The specific weaknesses:

• The epidemiology is observational and regionally narrow. The human studies are case-control and cross-sectional studies, almost all from the French Caribbean and French Guiana, plus two short letters from New Caledonia. None of those cited here followed people forward in time.
• Dietary recall is a weak instrument. Participants were asked about lifetime consumption of local foods, retrospectively, in a setting where the hypothesis eventually became locally known. Recall bias cannot be excluded after the fact.
• Other exposures were present. Guadeloupe has a history of agricultural pesticide use, including organochlorines used in banana cultivation, and soursop contains isoquinoline alkaloids as well as acetogenins. Disentangling one dietary exposure from a whole environment is difficult.
• Genetics cannot be excluded. The three autopsied patients in the 2002 study were all homozygous for the H1 tau haplotype, though none carried a tau gene mutation. An isolated population can carry risk variants that concentrate a rare phenotype.
• The rat lesion study used a different route. It infused purified annonacin into the bloodstream. Drinking an infusion or eating fruit involves gut absorption and a mixture of many compounds. In a rat pharmacokinetic study, only about 3.2% of an oral dose of annonacin reached the circulation; human pharmacokinetic data were not found.
• Annona fruit is widely eaten across Latin America, West Africa and Southeast Asia, yet a 2023 review of endemic parkinsonism listed only one Caribbean cluster (Guadeloupe and Martinique) among the ten endemic forms known worldwide. That is a real counterargument, although the same review concluded that both genetic and environmental factors contribute to these diseases.
• Not every review reaches the same verdict. A 2020 systematic review of A. muricata leaf extract listed neurotoxic effects among its findings but concluded overall that the extract had a favourable safety and tolerability profile, and noted that doses used in toxic animal studies may not translate directly to humans.
• Similar clusters elsewhere have had other explanations. The parkinsonism–dementia complex of Guam has been attributed at various times to cycad consumption and to other factors, and the history of that hypothesis is a cautionary tale about confident environmental attribution.

How researchers have weighed it. The Caribbean research groups have concluded that the evidence justifies public-health action: the 2022 authors proposed more restrictive preventive recommendations on Annonaceae products, and the 2026 authors wrote that health policy makers should raise public awareness of the risks. The 2020 German risk assessment concluded that no safe intake level could be set. On the benefit side, the uses for which people drink soursop leaf regularly are unsupported in humans (cancer), supported mainly in animals (blood sugar and blood pressure), or traditional (sleep and digestion), as the sibling articles on this hub describe.

Where the Acetogenins Are: Leaf, Seed, Bark, Pulp

Acetogenins have been isolated from every part of Annona muricata that has been studied — leaves, bark, seeds, roots and fruit. What matters for exposure is how much a serving delivers, and the measured figures run against a common assumption:

• Seeds. Seed extracts are the source of acetogenin-based insecticides, and the 2016 eye-injury case series involved seed material. The seeds are not eaten as food.
• Roots and bark. Used in some traditional decoctions. The 2002 cell study found an alkaloid extract of root bark toxic to dopaminergic neurons.
• Leaves. Sold as tea and packed into capsules. Commercial leaf products tested in 2023 contained 1.05–3.09 mg of annonacin per gram.
• Fruit pulp. Not a negligible source. The 2005 study estimated about 15 mg of annonacin in an average fruit, and a 2017 study measured 13.5–36.4 mg of the acetogenin squamocin per fruit of sugar apple (A. squamosa). In the 2022 study, even low cumulative fruit or juice consumption (above 0.2 fruit-years) was associated with more severe parkinsonism.

The 1999 and 2007 Guadeloupe studies and the 2022 and 2026 Caribbean studies all examined fruit (or fruit juice) as well as herbal tea; the human association has never been confined to the leaf.

One further complication is species. Annona squamosa (sugar apple, sweetsop), A. cherimola (cherimoya) and the North American pawpaw (Asimina triloba, another member of the family) all contain acetogenins: squamocin in sugar-apple pulp (2017), more than 80 acetogenins detected in a commercial cherimoya-based alcoholic drink (2014), and annonacin at about 0.07 mg per gram of pawpaw pulp, where it was toxic to rat cortical neurons in culture (Potts 2012). A label reading "custard apple" or simply "Annona" does not identify the species, but each of these species has been found to contain acetogenins.

How Much Reaches a Cup, a Fruit or a Can?

Acetogenins are lipophilic, and tea is water, so a reasonable expectation is that an infusion extracts relatively little. The 2005 measurements bear that out for the cup, and show that the fruit and its nectar deliver far more per serving:

Figures as reported in the 2005 study's abstract (and, for leaf products, the 2023 product analysis); bar lengths are drawn to scale against the 36 mg nectar figure.

Several factors affect how much a leaf preparation delivers:

• Preparation method. Traditional leaf preparations include both brief infusions and longer simmered decoctions; the 2005 study's 140-microgram figure covered a cup of either.
• Swallowing the leaf. Leaf powder in capsules or blended into drinks delivers the whole leaf rather than what passes into water.
• Alcohol-based preparations. Ethanol dissolves lipophilic compounds well, and acetogenins have been identified in an alcoholic drink made from cherimoya fruit (2014). The acetogenin content of soursop tinctures was not found in the literature checked for this page.
• Absorption. In rats, only about 3.2% of an oral dose of annonacin reached the bloodstream. How much of any of these amounts a person absorbs is not known.

Position on that range, multiplied by years of use, is what the researchers treat as cumulative exposure.

Why the Concern Is Cumulative, Not Acute

The research concern is not about a single cup or an occasional fruit producing a movement disorder. The patients in the Caribbean studies reported regular consumption over long periods, and the 2005 estimate was framed as a year of daily intake.

Chronic neurotoxicity differs from poisoning in three ways that make it harder for an individual to detect:

1. Exposure adds up. In cultured neurons, longer incubation caused toxicity at lower concentrations (2003), and the mouse juice study ran for 12 months (2016). How much annonacin builds up in human tissue is not known.
2. Neuronal loss is silent until it is large. The nigrostriatal system has substantial reserve; in Parkinson's disease, motor symptoms typically appear only after a large share of nigral dopaminergic neurons has already been lost.
3. The damage does not reverse. These neurons do not regenerate. Ending an exposure ends further injury from it; it does not recover what was lost.

Together, these mean that a daily habit can give no warning signal until the reserve is exhausted, which is why researchers give weight to the mechanism even without proof from the human data.

Dose and Duration: What Can and Cannot Be Said

There is no established safe dose of soursop leaf, fruit products or extracts, and there is no established therapeutic dose either. No adequate human trials exist from which either figure could be derived, and the 2020 German risk assessment concluded that a safe intake level could not be set. No figure on this page is a recommended amount.

What has been reported:

• Measured amounts per serving (2005): about 15 mg of annonacin in a fruit, 36 mg in a can of nectar, 140 micrograms in a cup of infusion or decoction.
• Product labels vary. The 2023 analysis of seven commercial leaf products found that labelled dosage recommendations varied between products, and one product contained less than half the content stated on its label. Label directions come from sellers, not from safety studies.
• Duration. The human associations involve long-term, regular consumption; the 2022 study measured exposure in cumulative fruit-years and found the association even at low cumulative fruit or juice intake and with any herbal-tea use.
• Concentrated extracts. No safety data were found for "standardised acetogenin" products, which are standardised to the constituent the toxicology concerns.

The pattern most closely matching the exposure in the Caribbean studies is regular consumption over many years.

Capsules, Extracts and Cancer Marketing

Capsules and extracts differ from tea in ways the research makes relevant:

• A capsule delivers the whole leaf, not an extract of it. With tea, acetogenins that stay in the leaf are discarded with the spent leaves; swallowing leaf powder discards nothing. Leaf products tested in 2023 contained 1.05–3.09 mg of annonacin per gram.
• Solvent extracts concentrate lipophilic constituents. Organic-solvent extraction draws out lipophilic compounds such as acetogenins more than water does.
• Content varies and is rarely stated. The seven products tested in 2023 ranged from 1.05 to 3.09 mg of annonacin per gram and varied in which acetogenins were present; one exceeded the United States Pharmacopoeia limit for total aerobic microbial count.
• Species and plant part are often unstated. A label reading "graviola 1000 mg" does not say whether it contains leaf, bark, seed or a mixture, or which Annona species supplied it.

Cancer claims and regulatory action

According to the 2020 risk assessment, A. muricata supplements are marketed particularly for cancer and parasitic infections. The anticancer evidence for soursop and its acetogenins comes from cell-culture and animal studies; no clinical trial has shown that soursop treats or prevents cancer in people. The evidence is examined in the article on soursop and cancer. In September 2008 the US Federal Trade Commission announced eleven law-enforcement actions against sellers of products marketed as cancer cures; one concerned a soursop (guanabana) extract advertised to prevent and treat melanoma. The settlements barred the sellers from claiming that their products prevent, treat or cure any type of cancer unless the claim is supported by competent and reliable scientific evidence.

Who the Research Raises Concerns For

The research raises a particular concern for the following groups. For most of them there are no human studies of soursop at all; the concern rests on the mechanism and the animal data.

• People with an existing parkinsonian or movement disorder. Parkinson's disease, progressive supranuclear palsy, multiple system atrophy, corticobasal syndrome, essential tremor or an undiagnosed movement problem. This is the one group with direct human data: in the 2022 and 2026 Caribbean studies, Annonaceae consumption was associated with more severe disease, including in Parkinson's disease.
• People with a family history of parkinsonism or other neurodegenerative disease. Researchers describe these diseases as arising from both genetic and environmental factors, and complex I inhibitors are among the environmental factors studied.
• People with a known or suspected mitochondrial disorder. Their respiratory chain is already impaired, and acetogenins inhibit complex I.
• Pregnancy and breastfeeding. There is no human safety data.
• Infants and children. There is no safety data, and the developing nervous system is not studied for this exposure.
• People with other neurodegenerative diagnoses, including Alzheimer's disease and other tauopathies, given that annonacin and soursop juice produced tau changes in laboratory models.
• Seeds. Seed material is the source of acetogenin insecticides, and eye contact with custard-apple seed material has caused toxic keratoconjunctivitis.

Decisions about supplements and herbal products belong with a clinician who knows the person's history.

Drug and Condition Interactions

No clinical interaction studies of soursop were found. The points below are what the mechanism and the animal data raise.

• Levodopa, dopamine agonists and other antiparkinsonian drugs. A theoretical conflict: a dopaminergic neurotoxin acting on the neurons the drugs are meant to support. No interaction study exists.
• Metformin. Metformin's own action involves a time-dependent, self-limiting inhibition of mitochondrial complex I (Owen 2000). Combining it with a more potent complex I inhibitor has not been studied, and additive glucose lowering is also possible. See the blood sugar article.
• Insulin and sulfonylureas. Animal data suggest possible additive glucose lowering; no human interaction data exist.
• Antihypertensives — ACE inhibitors, ARBs, calcium-channel blockers, diuretics, beta-blockers. Animal data suggest possible additive blood-pressure lowering; dizziness on standing is the symptom such an effect would produce.
• Chemotherapy and radiotherapy. Chemotherapy-induced peripheral neuropathy is already a common, dose-limiting problem, and a neurotoxin adds to the same concern; several cancer drugs also act on or depend on mitochondrial function.
• Sedatives, benzodiazepines, sleep medication and alcohol. Possible additive sedation, given the traditional sleep use.
• Liver or kidney impairment. Human clearance of acetogenins has not been characterised; in rats, the elimination half-life of oral annonacin was about 4.8 hours.

What the Guadeloupe Syndrome Looked Like

There is no early warning sign specific to acetogenin exposure. The atypical parkinsonism described in Guadeloupe had features that differ from classic Parkinson's disease, as reported in the 1999, 2002, 2007 and 2008 papers:

• Early postural instability and falls, rather than after years of disease.
• Difficulty with vertical eye movement, particularly looking down, in the PSP-like group (absent by definition in the parkinsonism–dementia group).
• Symmetrical rigidity and slowness. Tremor was not absent: in the 2007 study it occurred in more than half of the PSP-like patients, which distinguished them from classical PSP.
• Frontal and subcortical cognitive change and dementia, and pseudobulbar signs affecting speech and swallowing.
• Hallucinations (59% of the PSP-like group and 52% of the dementia group in 2007), and in many patients dysautonomia, myoclonus and REM sleep behaviour disorder.
• Little or no benefit from levodopa in most patients.

These features are not specific to soursop — they describe atypical parkinsonism from any cause, and most people who have them have never consumed an Annona product. Assessing such symptoms is a matter for a neurologist, and dietary and supplement history is part of the information a neurological assessment can use.

What Is Known for Long-Term Drinkers

For readers who have been drinking soursop leaf tea, or taking graviola products, regularly for a long time, the research says the following:

• The exposure is the variable that changes going forward. In cultured neurons, removing annonacin after a short exposure stopped further cell death (2003). Acetogenins are not known to cause physical dependence or withdrawal.
• Long-term use does not mean damage has occurred. The human evidence is an association from specific populations with regular long-term intake, not a dose–response curve on which an individual can be placed. Most people who have consumed Annona foods have not developed a movement disorder.
• There is no test for cumulative acetogenin exposure, so a person's own history of use is the only record of it.
• Capsules are not a lower-exposure form. They deliver whole leaf, at 1.05–3.09 mg of annonacin per gram in the products tested.
• The fruit is part of the same question. The Caribbean studies associated fruit and juice consumption, as well as herbal tea, with atypical parkinsonism and with more severe disease.
• The evidence for each reason people drink it — sleep, blood pressure, glucose, cancer — is reviewed in the other three articles on this hub, alongside better-studied options for each.

Key Research Papers

1. Caparros-Lefebvre D, Elbaz A (1999). Possible relation of atypical parkinsonism in the French West Indies with consumption of tropical plants: a case-control study. Caribbean Parkinsonism Study Group. The Lancet 354(9175):281-6. The original epidemiological signal. Case-control. — PubMed PMID: 10440304
2. Caparros-Lefebvre D, Sergeant N, Lees A, et al. (2002). Guadeloupean parkinsonism: a cluster of progressive supranuclear palsy-like tauopathy. Brain 125(Pt 4):801-11. The neuropathological characterisation. — PubMed PMID: 11912113
3. Lannuzel A, Michel PP, Caparros-Lefebvre D, et al. (2002). Toxicity of Annonaceae for dopaminergic neurons: potential role in atypical parkinsonism in Guadeloupe. Movement Disorders 17(1):84-90. Alkaloids from soursop root bark. Cell culture. — PubMed PMID: 11835443
4. Lannuzel A, Michel PP, Höglinger GU, et al. (2003). The mitochondrial complex I inhibitor annonacin is toxic to mesencephalic dopaminergic neurons by impairment of energy metabolism. Neuroscience 121(2):287-96. The mechanism paper. Cell culture. — PubMed PMID: 14521988
5. Champy P, Höglinger GU, Féger J, et al. (2004). Annonacin, a lipophilic inhibitor of mitochondrial complex I, induces nigral and striatal neurodegeneration in rats: possible relevance for atypical parkinsonism in Guadeloupe. Journal of Neurochemistry 88(1):63-9. The animal lesion study. Animal. — PubMed PMID: 14675150
6. Champy P, Melot A, Guérineau Eng V, et al. (2005). Quantification of acetogenins in Annona muricata linked to atypical parkinsonism in Guadeloupe. Movement Disorders 20(12):1629-33. Annonacin content of fruit, nectar and infusions. — PubMed PMID: 16078200
7. Caparros-Lefebvre D, Steele J (2005). Atypical parkinsonism on Guadeloupe, comparison with the parkinsonism-dementia complex of Guam, and environmental toxic hypotheses. Environmental Toxicology and Pharmacology 19(3):407-13. Comparison with Guam. — PubMed PMID: 21783505
8. Escobar-Khondiker M, Höllerhage M, Muriel MP, et al. (2007). Annonacin, a natural mitochondrial complex I inhibitor, causes tau pathology in cultured neurons. Journal of Neuroscience 27(29):7827-37. Links the compound to the specific pathology. Cell culture. — PubMed PMID: 17634376
9. Lannuzel A, Höglinger GU, Verhaeghe S, et al. (2007). Atypical parkinsonism in Guadeloupe: a common risk factor for two closely related phenotypes? Brain 130(Pt 3):816-27. Cross-sectional. — PubMed PMID: 17303592
10. Höllerhage M, Matusch A, Champy P, et al. (2009). Natural lipophilic inhibitors of mitochondrial complex I are candidate toxins for sporadic neurodegenerative tau pathologies. Experimental Neurology 220(1):133-42. Cell culture. — PubMed PMID: 19682988
11. Angibaud G, Gaultier C, Rascol O (2004). Atypical parkinsonism and Annonaceae consumption in New Caledonia. Movement Disorders 19(5):603-4. Letter; no abstract indexed. — PubMed PMID: 15133832
12. Langston JW, Ballard P, Tetrud JW, Irwin I (1983). Chronic Parkinsonism in humans due to a product of meperidine-analog synthesis. Science 219(4587):979-80. The MPTP precedent. — PubMed PMID: 6823561
13. Coria-Téllez AV, Montalvo-Gónzalez E, Yahia EM, Obledo-Vázquez EN (2018). Annona muricata: A comprehensive review on its traditional medicinal uses, phytochemicals, pharmacological activities, mechanisms of action and toxicity. Arabian Journal of Chemistry 11(5):662-691. Not indexed in PubMed. — doi:10.1016/j.arabjc.2016.01.004
14. Moghadamtousi SZ, Fadaeinasab M, Nikzad S, et al. (2015). Annona muricata (Annonaceae): A Review of Its Traditional Uses, Isolated Acetogenins and Biological Activities. International Journal of Molecular Sciences 16(7):15625-58. Review. — PubMed PMID: 26184167
15. Caparros-Lefebvre D (2004). Atypical parkinsonism in New Caledonia: comparison with Guadeloupe and association with Annonaceae consumption. Movement Disorders 19(5):604. Letter; no abstract indexed. — PubMed PMID: 15133833
16. Lannuzel A, Ruberg M, Michel PP (2008). Atypical parkinsonism in the Caribbean island of Guadeloupe: etiological role of the mitochondrial complex I inhibitor annonacin. Movement Disorders 23(15):2122-8. Review. — PubMed PMID: 18816693
17. Cleret de Langavant L, Roze E, Petit A, et al. (2022). Annonaceae Consumption Worsens Disease Severity and Cognitive Deficits in Degenerative Parkinsonism. Movement Disorders 37(12):2355-2366. Observational; 180 patients. — PubMed PMID: 36210778
18. Nasri A, Chaumont H, Arnaud S, et al. (2025). Atypical parkinsonian syndromes in French Guiana: Similarities and differences with Caribbean variants. Journal of Parkinson's Disease 15(7):1287-1297. — PubMed PMID: 40961237
19. Ribeyron L, Petit A, Edragas R, et al. (2026). The Clinical Effect of Annonaceae Fruit Consumption on Caribbean Parkinson's Disease Severity. Behavioural Neurology 2026(1):e8897550. — PubMed PMID: 42192274
20. González-Lizárraga F, Boluda S, Hidalgo JR, et al. (2025). Protein Coaggregation in Caribbean Atypical Parkinsonism: The Contribution of Annonacin. Neuropathology and Applied Neurobiology 51(4):e70026. — PubMed PMID: 40702589
21. Rottscholl R, Haegele M, Jainsch B, et al. (2016). Chronic consumption of Annona muricata juice triggers and aggravates cerebral tau phosphorylation in wild-type and MAPT transgenic mice. Journal of Neurochemistry 139(4):624-639. Animal. — PubMed PMID: 27569447
22. Bonneau N, Schmitz-Afonso I, Brunelle A, et al. (2015). Method development for quantification of the environmental neurotoxin annonacin in Rat plasma by UPLC-MS/MS and application to a pharmacokinetic study. Journal of Chromatography B 1004:46-52. Oral bioavailability in rats. — PubMed PMID: 26444335
23. Bonneau N, Baloul L, Bajin Ba Ndob I, et al. (2017). The fruit of Annona squamosa L. as a source of environmental neurotoxins: From quantification of squamocin to annotation of Annonaceous acetogenins by LC-MS/MS analysis. Food Chemistry 226:32-40. — PubMed PMID: 28254016
24. Le Ven J, Schmitz-Afonso I, Lewin G, et al. (2014). Identification of the environmental neurotoxins annonaceous acetogenins in an Annona cherimolia Mill. alcoholic beverage using HPLC-ESI-LTQ-Orbitrap. Journal of Agricultural and Food Chemistry 62(34):8696-704. — PubMed PMID: 25088119
25. Potts LF, Luzzio FA, Smith SC, et al. (2012). Annonacin in Asimina triloba fruit: implication for neurotoxicity. Neurotoxicology 33(1):53-8. Pawpaw fruit pulp, not supplements. Cell culture. — PubMed PMID: 22130466
26. Jocelin Chan WJ, Harnett JE, Meroni A, et al. (2023). An evaluation of the quality of Annona muricata leaf products. Journal of Pharmacy and Pharmacology 75(10):1357-1365. Seven commercial leaf products. — PubMed PMID: 37440207
27. Chan WJ, McLachlan AJ, Hanrahan JR, Harnett JE (2020). The safety and tolerability of Annona muricata leaf extract: a systematic review. Journal of Pharmacy and Pharmacology 72(1):1-16. Systematic review. — PubMed PMID: 31659754
28. Raclariu-Manolica AC, Bakhiya N, Hirsch-Ernst KI (2020). Risk assessment regarding the use of Annona muricata in food supplements. EFSA Journal 18(Suppl 1):e181112. German Federal Institute for Risk Assessment, EU-FORA programme. — PubMed PMID: 33294051
29. Betarbet R, Sherer TB, MacKenzie G, et al. (2000). Chronic systemic pesticide exposure reproduces features of Parkinson's disease. Nature Neuroscience 3(12):1301-6. Rotenone in rats. Animal. — PubMed PMID: 11100151
30. Tanner CM, Kamel F, Ross GW, et al. (2011). Rotenone, paraquat, and Parkinson's disease. Environmental Health Perspectives 119(6):866-72. Nested case-control. — PubMed PMID: 21269927
31. Owen MR, Doran E, Halestrap AP (2000). Evidence that metformin exerts its anti-diabetic effects through inhibition of complex 1 of the mitochondrial respiratory chain. Biochemical Journal 348(Pt 3):607-14. — PubMed PMID: 10839993
32. Ribeiro LP, Zanardi OZ, Gonçalves GLP, et al. (2018). Toxicity of an Annonin-Based Commercial Bioinsecticide Against Three Primary Pest Species of Stored Products. Neotropical Entomology 47(1):145-151. — PubMed PMID: 28353149
33. Intaranongpai J, Chavasiri W, Gritsanapan W (2006). Anti-head lice effect of Annona squamosa seeds. Southeast Asian Journal of Tropical Medicine and Public Health 37(3):532-5. — PubMed PMID: 17120975
34. Nagaraja H, Kugar T, Shivanna Y, et al. (2016). Ocular toxicity by seeds of Annona squamosa (custard apple). Indian Journal of Ophthalmology 64(8):611-3. Case series. — PubMed PMID: 27688289
35. Menšíková K, Steele JC, Rosales R, et al. (2023). Endemic parkinsonism: clusters, biology and clinical features. Nature Reviews Neurology 19(10):599-616. Review. — PubMed PMID: 37684518
36. US Federal Trade Commission (2008, September 18). FTC sweep stops peddlers of bogus cancer cures. Press release. — ftc.gov

PubMed Topic Searches

1. PubMed: Annonacin neurotoxicity
2. PubMed: Annonaceae and atypical parkinsonism
3. PubMed: Acetogenins and complex I inhibition
4. PubMed: Graviola supplement adverse effects
5. PubMed: Annona muricata and reproductive toxicity
6. PubMed: PSP and environmental risk factors
7. PubMed: Annona muricata subchronic toxicity studies
8. PubMed: Acetogenin extraction into infusions

Connections

• All Herbs
• Soursop Leaf Benefits Hub
• Soursop and Cancer: The Claim Examined
• Soursop Leaf for Blood Sugar and Blood Pressure
• Soursop Leaf for Sleep and Digestive Comfort
• Soursop Leaf (Annona muricata)
• Progressive Supranuclear Palsy
• Parkinson's Disease
• Multiple System Atrophy
• Corticobasal Syndrome
• Essential Tremor
• Alzheimer's Disease
• Peripheral Neuropathy
• Neurology
• Toxins

Safety and disclaimer. This page is health information, not medical advice. Annonacin and related acetogenins in soursop are mitochondrial complex I inhibitors and are neurotoxic to dopaminergic neurons in cell and animal studies. Regular long-term consumption of Annona fruit, juice and herbal tea has been associated with atypical parkinsonism in observational studies from the French Caribbean; that association is not proven to be causal and has not been ruled out. No safe long-term intake level has been established for leaf, fruit products or concentrated extracts. The research raises particular concern for people with movement disorders, neurodegenerative or mitochondrial conditions, a family history of parkinsonism, and for pregnancy, breastfeeding and childhood. Decisions about supplements belong with a clinician, and neurological symptoms are a matter for a proper assessment rather than a web page.

Back to Table of Contents

Featured Videos