The most recent issue of the Lancet has an interesting case study: a child overdosing on morphine acquired through the breast milk of her mother (who was taking codeine, a chemical precursor to morphine). Normally, codeine is considered safe for nursing mothers. What happened here?

If turns out that newborns have trouble metabolizing morphine, but in general the amount of morphine in breast milk is low enough to avoid problems. In this case, however, the mother was heterozygous for a variant in the gene CYP2D6, which encodes the enzyme that converts codeine to morphine. As a heterozygote, she was an "ultra-rapid metaboliser", meaning the codeine she took was converted quickly to morphine. Confirming this, the concentration of morphine in here milk was higher than normal: "A morphine concentration of 87 ng/mL was found—the typical range of milk concentrations after repeated maternal codeine is 1·9–20·5 ng/mL at doses of 60 mg every 6 h."

The frequency of the fast metabolism allele varies between population--the case report claims a frequency of 1% in Finland and Denmark, 10% in Greece and Portugal, and 29% in Ethiopia. So codeine may generally be safe for Finnish mothers, but less so for Ethiopian ones.

Further, the fast metabolism phenotype itself varies between populations as well. One study found that Caucasian fast metabolisers had a faster rate of metabolism than Chinese fast metabolisers. So one gene is not enough to model the phenotype of codeine metabolism--one may need to take into account entire pathways.