Phosphatidylethanol is a "pathological" phospholipid, formed via the action of phospholipase D only in the presence of ethanol. The present study was made to elucidate how different levels and patterns of alcohol intake affect blood levels of phosphatidylethanol in comparison with other markers of abuse. We used a new HPLC-evaporative light-scattering detection technique for phosphatidylethanol quantitation. This method had a total coefficient of variation of <20% at the detection limit of 0.2 nmol, equaling 0.8 μmol/liter of whole blood. Two groups were studied. (a) Five healthy volunteers were given 32 to 47 g of ethanol in a single dose, to give blood ethanol levels of ∼25 mmol/liter after 30 to 60 min. Phosphatidylethanol, carbohydrate-deficient transferrin (CDT), and blood ethanol were measured before and after the intake. (b) Twelve student volunteers were studied during a 3 week period of prolonged alcohol consumption (total estimated intake: 1334 ± 488 g, mean ± SD) and phosphatidylethanol, serum-CDT, gamma-glutamyltransferase, and blood ethanol were measured at the start of the period (day 1) and twice at the end of the period (days 18 and 21). In group (a), no phosphatidylethanol was detected at any time after ethanol dosage/intake. In group (b), no blood phosphatidylethanol or blood ethanol could be demonstrated at the start, and serum-CDT was below the discrimination limit (1.3%) in all persons. No phosphatidylethanol was detected in those four persons with the lowest intake (742 ± 150 g). However, the remaining eight persons had detectable levels of phosphatidylethanol (1.0 to 2.1 μmol/liter), and these had a higher total intake (1630 ± 389 g). There was a statistically significant (p = 0.02) increase in serum CDT for 3 weeks. However, only 3 of 12 persons increased above the discrimination limit. The present results indicate that a substantial alcohol intake is needed to elevate blood phosphatidylethanol. In comparison with serum-CDT, blood phosphatidylethanol appears more sensitive.