Substance and claimed effects

The substance is the rise in steroidal glycoalkaloids — primarily α-solanine and α-chaconine — that occurs when whole potato tubers (Solanum tuberosum) green or sprout in storage, and the kitchen practice of trimming versus discarding affected tubers in response. Cultivated potato flesh typically carries 10–150 mg total glycoalkaloids (TGA) per kg fresh weight; potato peel carries 3–10× more EFSA 2020. Light exposure drives chlorophyll synthesis (the visible green) and, by a separate but co-triggered pathway, drives glycoalkaloid synthesis in the outermost 1.5 mm of the tuber. Sprouting concentrates glycoalkaloids in the sprout itself and in the eye / shoulder regions of the tuber Friedman 2006. Claimed effects of the practice: prevention of acute gastrointestinal / cholinergic poisoning at the high end of the dose range, prevention of subclinical GI upset at the moderate end, and — speculatively, on animal-model and one Chinese case-control evidence — reduction of teratogenic risk during pregnancy. The entry covers all three.

Evidence by addressing question

Mechanism

Glycoalkaloids exert toxicity through two complementary mechanisms. First, both α-solanine and α-chaconine are reversible inhibitors of acetylcholinesterase and butyrylcholinesterase at micromolar concentrations Friedman 2006; the resulting acetylcholine excess explains the bradycardia, sweating, salivation, miosis, and CNS depression observed in severe poisoning. Second, the molecules intercalate into cholesterol-containing membranes and disrupt their integrity, lysing erythrocytes and gut enterocytes in vitro at low concentrations; α-solanine and α-chaconine act synergistically here, with combined cytotoxic potency greater than additive EFSA 2020.

Glycoalkaloid biosynthesis in the tuber proceeds from cholesterol through a series of hydroxylation, oxidation, and glycosylation steps. Both light and wounding upregulate the pathway. The reason the kitchen rule is "green skin = high alkaloid" is that the two signals — chlorophyll accumulation and glycoalkaloid accumulation — share the light trigger but proceed independently; chlorophyll is the visible marker, not the toxin Friedman 2006. Concentrations are highest immediately under the peel and in the eyes and sprouts, dropping ten- to twenty-fold across the next centimetre of flesh.

Evidence

The EFSA CONTAM Panel's 2020 risk assessment is the load-bearing regulatory document. It identified a lowest-observed-adverse-effect level (LOAEL) of 1 mg TGA per kg body weight per day for acute exposure in humans, drawn from case reports, outbreaks, and a small set of volunteer studies. Symptoms above this level: nausea, vomiting, diarrhoea, abdominal pain; at higher doses, neurological signs including drowsiness, confusion, vision disturbance, and in the worst cases coma and convulsions EFSA 2020. EFSA explicitly declined to validate the long-cited 200 mg TGA/kg raw-tuber "safety limit" (which traces to a 1924 German study by Bömer and Mattis) on the grounds that it lacks a modern scientific basis; Germany's BfR has recommended a lower 100 mg/kg threshold. No chronic-exposure reference point could be derived from the available human data.

The best-characterised human outbreak is the 1979 South London boarding-school incident: 78 schoolboys fell ill after eating boiled potatoes from a sack that had been left in storage over the summer term; 17 required hospitalisation, with vomiting, diarrhoea, fever, abdominal pain, and in the severe cases coma with convulsive twitching and brief hallucinations. All recovered within a week. Recovered potato samples contained 250–300 mg TGA/kg McMillan 1979. EFSA also flagged a 2015 Baden-Württemberg cluster where boiled-in-skin jacket potatoes at 236 mg TGA/kg caused stomach ache and vomiting in multiple diners EFSA 2020. Toxic-dose calculations from the literature converge on 2–5 mg/kg body weight for symptomatic poisoning and 3–6 mg/kg for potentially fatal exposure Friedman 2006.

Protocol

The handling rule has three pieces, all evidence-anchored.

• Store dark and cool. 4–10 °C in opaque bags or unlit cupboards; refrigerator (below ~4 °C) reduces alkaloid synthesis but converts starch to reducing sugars (a separate quality problem). Storage at 20 °C for several months without sprout inhibitors elevates TGA several-fold over storage at 1–4 °C, and storage under light induces glycoalkaloid synthesis more than storage in the dark in matched conditions Friedman 2006. Concentrations also rise with storage duration regardless of conditions.
• Trim green and sprouted areas generously. 30–80% of a normal tuber's glycoalkaloid load sits in the outer ~1.5 mm under the peel; peeling alone removes >90% of TGA in non-green tubers. For green or sprouted tubers, cutting at least 3–5 mm below the green margin and removing sprouts at their base reduces flesh TGA to background levels in tubers where greening is superficial Friedman 2006.
• Discard heavily green / extensively sprouted tubers. When greening penetrates beyond the peel into the flesh, when sprouts are long (> ~1–2 cm) and multiple, or when the cooked potato tastes bitter or causes any tongue / throat burning, discard the whole tuber. Cooking does not rescue. α-Solanine is thermally stable to ~243 °C, well above boiling-water temperature; boiling reduces TGA by only 1–4%, microwaving by ~15%, deep-frying at 210 °C by ~40% — none enough to bring a 250 mg/kg tuber under the LOAEL for a normal serving EFSA 2020.

Misconceptions

Three widely repeated rules are wrong or incomplete. (1) "Just cook them well — heat destroys solanine." False: glycoalkaloids are heat-stable across normal cooking temperatures; only sustained high-temperature deep-frying meaningfully reduces them Friedman 2006. (2) "If it's green you can just peel it." Partially true: deep peeling (≥ 3 mm below the green) handles superficial greening, but greening that has penetrated the flesh and tubers with multiple long sprouts carry alkaloids beyond what peeling captures. (3) "Green = solanine." The green pigment is chlorophyll, which is itself harmless; light exposure is what produces both chlorophyll and glycoalkaloids in parallel, so green colour is a strong but imperfect visual proxy for glycoalkaloid elevation. A non-green but sprouted, bitter tuber may have elevated alkaloids without the visual warning Friedman 2006.

Failure modes

The two ways the practice goes wrong are (1) trusting cooking to fix a green or sprouted tuber and (2) ignoring the bitter-taste / mouth-burning warning. Both contributed to the 1979 outbreak: the potatoes were boiled (standard preparation, ineffective), and several boys later reported the meal had tasted unpleasant McMillan 1979. Bitter-taste threshold sits around 140 mg TGA/kg; mouth/throat burning emerges around 200 mg TGA/kg — both below the typical poisoning dose but high enough to flag the tuber as discardable Friedman 2006. The taste signal is the body's last line of defence; spit out and discard the meal.

Contraindications and special populations

Pregnancy is the one population where caution beyond the general rule is warranted. Glycoalkaloids are teratogenic in frog, hamster, and mouse embryos at concentrations producing anencephaly, microcephaly, neural tube defects, and orofacial clefts; mechanism involves inhibition of cholesterol synthesis at early development and direct membrane disruption EFSA 2020. Human evidence is limited but suggestive: a 2018 case-control study from Shanxi Province (622 NTD cases, 858 controls) found maternal periconceptional consumption of sprouted potato ≥ 4 meals/week associated with ~2.2× higher odds of neural tube defect, ~2.5× for anencephaly, and ~3.5× for orofacial clefts, versus < 1 meal/week reference Ni et al. 2018. The result is one observational study and may reflect confounders (the high-sprout-consumption group also had lower folate intake), but the biology is plausible enough that pregnant readers should be strict about the discard rule rather than the trim rule. Children at lower body weight reach the toxic mg/kg threshold from smaller portions and warrant the same strictness.

Stakes

For an average adult (~70 kg) the LOAEL of 1 mg/kg body weight corresponds to ~70 mg TGA. A 200 g serving of a 250 mg/kg green / sprouted tuber delivers ~50 mg — at the threshold; a 200 g serving from a 350 mg/kg heavily affected tuber delivers ~70 mg, at or above LOAEL EFSA 2020. Most cases are subclinical GI upset that the reader doesn't connect to the meal because onset is delayed by 1–25 hours (peak 4–8 h); α-solanine elimination half-life is ~11 h, α-chaconine ~19 h, so symptoms persist for several days Friedman 2006. Severe outbreaks are rare but documented and follow a consistent pattern: long-stored, visibly affected potatoes, boiled or baked (ineffective), eaten as a normal portion.

Payoff

Compliance with the rule has two payoffs. First, near-elimination of the small annual risk of a "mystery stomach bug" from a stored potato — a benefit invisible to the reader because nothing happens. Second, the elimination of a class of pregnancy-period exposure with at least suggestive teratogenic signal. Both are protective rather than constructive; the entry is a not-getting-poisoned entry, not a feel-better-tomorrow entry.

Out-of-scope

This entry focuses on the post-harvest kitchen practice. Adjacent topics that warrant separate coverage: the broader glycoalkaloid entry covering tomato (tomatine), aubergine, and the dietary-source picture; cultivar-level breeding for low-alkaloid varieties; and the home-garden question of green-shouldered tubers from sun exposure during growth (which often have peel-only greening that responds well to peeling).

The credibility range

Optimist case. The mechanism is well-characterised (cholinesterase inhibition + membrane disruption + cholesterol-synthesis interference), the dose-response is mapped at the population level by EFSA, the case-report literature contains a coherent set of documented outbreaks across decades, and the in vivo teratology in three species is consistent. The 200 mg/kg legacy threshold and the modern 1 mg/kg body-weight LOAEL converge on the same practical conclusion: discard heavily affected tubers, trim mildly affected ones, store dark and cool. A practice that prevents acute poisoning at no cost has an asymmetric payoff worth following even if the per-event probability is small.

Skeptic case. Documented severe poisonings are rare relative to potato consumption; most people eat occasional slightly-greened or sprouted potatoes without consequence. The bitter-taste defence is an effective filter at the dose ranges that matter. The teratogenicity signal in humans rests on one observational case-control study with potential confounding by folate and other dietary factors; the cited animal models used parenteral or much higher doses than dietary exposure. The 200 mg/kg legacy guideline was derived from 1920s data on minor symptoms and may overstate routine risk; EFSA itself declined to validate it. Strict application risks unnecessary food waste at scale.

Author's call. The asymmetry is the deciding factor: trimming or discarding takes seconds, costs cents, and removes a low-frequency but documented risk that includes hospitalisation in the severe tail and a plausible (if not yet proven) reproductive-period signal. The skeptic position is right about average-day risk; the optimist position is right about the worth of a near-free hedge. Recommend the practice as written: store dark and cool, trim moderate, discard severe, never override a bitter or burning taste. Confidence in the rule itself is high; precision on where exactly the "trim vs discard" line sits is the soft spot. Evidence score 4 (regulatory opinion + case reports + mechanism + one human teratology study, no large RCT and unlikely to be one); controversy score 1 (EFSA's downward revision of the legacy threshold is a refinement, not a fight).

Stakeholder and incentive map

• Regulatory: EFSA (EU), BfR (Germany), FDA (US), Health Canada — all advise discard of green and sprouted tubers; EU Commission Recommendation 2022/561 mandates monitoring. Incentive: public-health risk reduction.
• Producer / breeder: seed-potato breeders incentivised to keep cultivar-level TGA low (commercial varieties capped < 200 mg/kg by industry custom); growers incentivised to avoid post-harvest light exposure that damages saleability.
• Cultural / household: waste-aversion culture (older relatives' "just cut the green off") pulls in the trim-not-discard direction; food-safety culture (parenting subreddits, school cafeteria policy after 1979) pulls in the discard direction.
• Counter: no real commercial counter-pressure — this isn't a controversial supplement, just a kitchen practice — though some heritage / heirloom variety advocates flag that wild-type potatoes naturally carry higher TGA and the rule should be cultivar-aware.

Population variability

• Body weight. Acute toxic dose is a mg/kg figure; children reach the same threshold from smaller portions. A 20 kg child crosses 1 mg/kg at 20 mg TGA — under a small serving of a moderately affected tuber.
• Pregnancy. See contraindications. Strict-discard rule applies; routine trimming is not enough where the tuber is visibly green or sprouted.
• Cultivar. Modern commercial cultivars (Russet, Yukon Gold, Désirée) are bred low; heritage and wild-type cultivars (some Andean landraces) can carry baseline TGA approaching the legacy 200 mg/kg threshold even in unaffected tubers Friedman 2006.
• Mechanical damage. Cuts, bruises, and frost damage during harvest or transport induce a wound-response glycoalkaloid spike independent of light exposure.

Knowledge gaps

• No modern RCT or controlled-feeding study sets the true human LOAEL; the EFSA reference point is anchored to outbreak case data that reads as an upper bound on the no-effect threshold rather than a precise toxic dose.
• The human teratogenicity signal rests on one case-control study; replication in other populations with different baseline folate status would clarify the contribution.
• The "trim depth" rule (3 mm? 5 mm? whole tuber?) is mechanistically informed but not formally calibrated against post-trim residual TGA across the realistic range of greening / sprouting severity.
• Population-level estimates of how often dietary exposure exceeds the LOAEL across normal household practice are missing — EFSA estimated mean dietary exposure well below LOAEL, but tail exposures from individual affected tubers are not well sampled.

Scope relative to brief. Brief named: glycoalkaloid rise on greening/sprouting, solanine toxicity, bitter-taste warning, dark-and-cool storage, discard threshold, complementarity with the broader glycoalkaloid entry. All covered. The article frames the substance as the kitchen practice (storage + trim/discard decision) rather than the toxin itself, since the broader glycoalkaloid entry is the natural home for the full chemistry / nightshade-family picture.

Action choice. Went with respond over do or avoid: the reader's behaviour is triggered by an observed event (greening, sprouting, bitter taste). The storage habit is daily-ish, but the load-bearing call is the response protocol when something visibly affected appears. as-needed cadence follows from that.

Rating difficulties.

• health_short_term: 2 — debated upward to 3 then back down. Severe outbreaks are documented but rare; most readers' lifetime exposure is the avoided subclinical GI upset they would never have connected to the meal. The non-zero score is earned by the severity of the tail (the 1979 outbreak, the EFSA acute LOAEL) and the per-event preventability. A 3 would overstate how often the typical reader actually escapes a problem because of this rule.
• evidence: 4 — anchored on the EFSA 2020 risk assessment plus the case-report literature plus the mechanism. No RCT (would be unethical), no chronic-exposure reference point — both block a 5.
• applicability: 4 — used the avoidance / recognition-of-warning-sign lift from meta.md: every adult cooks potatoes occasionally, every adult should know the bitter-taste warning. A naive prevalence count would land at 2 or 3 because severe poisoning is rare; the wider decision audience earns the 4.
• pull: 1 — chore-tier. There is no felt reward to the act of glancing at the bag and binning the bad ones; the win is purely the absence of a problem that would not have visibly happened. A 0 would imply aversive; trimming a potato is not aversive, just unremarkable.

Dream narrative omitted. Overall score lands well below the 40 floor at which a dream narrative becomes obligatory. The honest hook is clarity / relief / not-being-poisoned, not aspiration; a projected-life-cascade dream would ring false on an entry whose payoff is the meal that did not put you on the floor. Dek and tagline written straight, per dream-narrative.md §1.

Pregnancy framing. The teratogenicity section deliberately gives equal weight to the animal-model consistency and the single human case-control study (Ni et al. 2018), and names "one study is not proof" before recommending strictness. The recommendation rides on asymmetric downside, not on overclaiming the evidence. Considered scoping the entry to non-pregnant adults only; rejected — the wider rule is the same, and the pregnancy carve-out is one paragraph plus a warning callout.

Future links. Should cross-link to the broader glycoalkaloid entry once it exists (covers tomato / aubergine / dietary picture, cultivar selection, baseline exposure modelling). Also worth linking to a future home-garden potato-growing entry if one is added (covers the prevent-greening-at-the-source angle: hilling, sun-shielding during growth).

Separate-entry candidates surfaced during the write. A standalone entry on heritage / wild-type cultivars and their elevated baseline glycoalkaloids would be worth flagging if the catalogue gets into Andean potato varieties. A separate entry on the food-poisoning recognition heuristic ("delayed-onset GI symptoms — what to consider beyond stomach bug") could be worth backlogging.

Excluded by editorial choice. Did not cover commercial seed-potato breeding programs, glycoalkaloid extraction for medicinal use, or the in vitro anti-cancer literature on α-solanine — all off-spec for a kitchen-practice entry and the wrong audience.