Shu Fu

Aconitine-type diester alkaloids are sodium-channel openers, the pharmacological opposite of a class I blocker, and the arrhythmia they cause is triggered and catecholamine-sensitive rather than re-entrant. Aconite poisoning produces bradyarrhythmia, nodal tachycardia, bidirectional ventricular tachycardia, intraventricular block and ventricular fibrillation; ventricular tachyarrhythmia occurs in roughly 18 percent of poisoned patients and refractory ventricular tachyarrhythmia and asystole are the leading causes of death. Of 31 patients admitted to Hong Kong public hospitals with aconite root poisoning between 1989 and 1991, two died of ventricular arrhythmia, and in reported fatal cases ventricular tachycardia progressed to fibrillation and cardiac arrest refractory to resuscitation within 12 hours of ingestion. There is no antidote; amiodarone and flecainide have been used with variable success and no agent is reliably effective.

Aconitine and mesaconitine prolong action potential duration and provoke early and delayed afterdepolarisations in cardiac myocytes. Adding a drug that independently delays repolarisation increases the probability that a subclinical residual alkaloid load becomes a sustained ventricular tachyarrhythmia. The interaction is additive pharmacodynamic proarrhythmia rather than a metabolic one, so it does not require the two exposures to peak together.

Digoxin inhibits the sodium-potassium ATPase and raises intracellular calcium; aconitine-type alkaloids force sodium channels open and raise intracellular sodium, which drives calcium in through the sodium-calcium exchanger. The two converge on calcium overload and delayed afterdepolarisations from different directions. The toxic presentations also overlap closely (nausea, vomiting, visual and sensory disturbance, ventricular ectopy, bidirectional ventricular tachycardia), so a clinician can attribute early aconitine toxicity to digoxin and lose time.

One of the five drugs named in the classical shi ba fan (eighteen incompatibilities) rhyme as opposing the Aconitum group, and for this genus the classical rule has a real pharmacological basis rather than only a textual one. Ge and colleagues (2024) showed in rats that Pinellia ternata alters cytochrome P450 activity and thereby the metabolism of the toxic Aconitum compounds, with the incompatibility attributed to increased absorption of those compounds in vivo. Neither herb is acutely toxic at its own dose; the hazard is created by the pairing.

Another of the shi ba fan five, and the best mechanistically characterised of them. When processed aconite root was combined with Fritillaria bulb, the aconite marker benzoylmesaconine showed higher intestinal permeability; Fritillaria inhibited both P-glycoprotein function and P-glycoprotein expression, with the alkaloids peimine, peimisine and imperialine identified as the inhibiting constituents. The classical incompatibility here reads as an efflux-transporter interaction that increases how much alkaloid reaches the circulation from the same oral dose.

The remaining three drugs of the shi ba fan Aconitum group, completing the classical set of Ban Xia, Gua Lou, Bei Mu, Bai Lian and Bai Ji. None is toxic on its own, but experimental work reports that the cardiotoxicity of Hei Shun Pian increases when it is combined with these five drugs. The mechanistic detail for these three is thinner than for Pinellia and Fritillaria, so this should be treated as a classical prohibition with supporting rather than conclusive modern pharmacology.

Zhi Fu Zi is the lateral (daughter) root, processed. Chuan Wu is the mother root of the same plant and Cao Wu is Aconitum kusnezoffii; both carry the same diester alkaloids and are separately dosed drugs. Prescribing them together stacks diester exposure while each individual dose looks acceptable. Hong Kong surveillance attributed herb-induced aconitine poisonings principally to Chuan Wu and Cao Wu used for musculoskeletal complaints, and identified larger-than-recommended dose and inadequate processing as the main determinants of risk. Substitution of raw for processed material collapses the entire safety margin: raw lateral root carries aconitine, mesaconitine and hypaconitine at hundreds of mg/kg, against a pharmacopoeial ceiling of 0.020 percent (200 mg/kg) for the sum of all three in the processed slice.

The diester-to-monoester hydrolysis that detoxifies aconite continues in the pot, so how long the formula is boiled determines how much diester alkaloid the patient actually swallows. Diester alkaloid concentration in the aqueous extract falls rapidly in the first phase of decoction and stabilises at a low level only after about 4 hours, with monoester alkaloid content peaking between 4 and 6 hours. Standard practice is therefore to boil the aconite slices for 1-2 hours before the rest of the formula is added. Any instruction that shortens the boil, including combining aconite in a single pot with volatile aromatics that are conventionally added in the last few minutes, undermines that. Surveillance of herb-induced aconitine poisoning identifies excess dose and inadequate processing or preparation as the two determinants of poisoning, and notes that patients need to be able to follow the decoction instructions at home.