CYP2D6: why codeine does nothing for some people and endangers others

Codeine is a prodrug. It is close to inert until CYP2D6 converts about 5-10% of it into morphine, which is the active analgesic. Your CYP2D6 genotype therefore decides what codeine does to you. The four phenotypes - Poor metabolisers (roughly 5-10% of people of European descent, carrying two non-functional alleles): little or no morphine is made, so codeine gives poor pain relief. Giving more just raises the codeine, not the morphine. - Intermediate and normal ("extensive") metabolisers: the expected conversion. - Ultra-rapid metabolisers (gene duplication; frequency varies widely by ancestry, up to ~10-30% in some North African and Middle Eastern populations): they make much more morphine, much faster. This is the dangerous end. Why it is on the FDA's radar The FDA issued a boxed warning after deaths in children given codeine after tonsillectomy, and in breastfed infants of ultra-rapid-metaboliser mothers, from morphine toxicity. Codeine and tramadol are now contraindicated in children under 12 and after tonsil/adenoid surgery under 18. Interactions that mimic a phenotype change Strong CYP2D6 inhibitors — fluoxetine, paroxetine, bupropion, quinidine — turn a normal metaboliser pharmacologically into a poor one, blunting codeine. The reverse cannot be induced; there is no clinical CYP2D6 inducer. Tramadol works the same way (CYP2D6 makes the active O-desmethyltramadol), so the same logic applies. Enzyme systems and the substrate/inhibitor map: https://wiki.soapbox.community/wiki/Cytochrome_P450_System_Inhibition_and_Induction Sources: FDA Drug Safety Communication (2017); CPIC guideline for CYP2D6 and codeine (Crews et al., Clin Pharmacol Ther).