After oral administration, diethylpropion is rapidly and extensively metabolized in man by N-de-ethylation and stereoselective carbonyl reduction. The unchanged drug excreted in acidic urine represents about 2% of the dose, while the total metabolites determined account for some 85%. The excretion curves indicate that the probable contribution of the parent compound to the observed activity is small. The major metabolites, together representing about 70% of the dose, are N-ethylaminopropiophenone, (+)-N-diethylnorpseudoephedrine, (+)-N-ethylnorpseudoephedrine, (-)-norephedrine and (-)-norpseudoephedrine. The other stereoisomers of the three amino-alcohols, and aminopropiophenone, are present in minor amounts.Diethylpropion (Amfepramone D.C.I.) (I, Fig. l), an anorectic agent (e.g. Boissier, 1962;Eiden, 1970) is completely absorbed from the human gastrointestinal tract and excreted exclusively via the renal pathway (Schreiber, Bozian & others, 1965). It undergoes extensive metabolism, the main metabolic routes being N-dealkylation, keto reduction and deamination, with para-hydroxylation occurring only to a small extent (Schreiber, Min & others, 1968). However, because the pooled urine samples were made alkaline (pH 12) and extracted over 24 h, some qualitative and quantitative aspects of these studies are questionable, since diethylpropion (I) and its aminoketone metabolites (11 and EI) are unstable under alkaline conditions (Hossie, 1970). Consequently, some of the "metabolites" detected by Schreiber & others (1968) are probably artifacts.Hossie (1970) showed that the two major metabolic pathways are N-dealkylation (compounds II, 111, V and VI account for some 60% of the dose) and reduction (compounds IVY V and VI account for about 40% of the dose). No N-oxides or glucuronides could be detected. The amino-ketones contain one asymmetric carbon atom, but after metabolic reduction a second asymmetric centre is created; little is known of the stereochemistry of the derived amino-alcohols IVY V and VI excreted by man. Hossie (1970) showed that the ratio VI threo: VI erythro compound was approximatively 3:1, but g.1.c. conditions did not give a suitable separation of the diastereoisomers to allow full quantitative assessment. Banci, Cartoni & others (1971) found IV to be present only in the threo form, the V erythro form to be possibly present with V threo form, and VI to be composed of more threo than erythro forms.We have investigated in man the quantitative aspects of the excretion of I and its metabolites I1 to VI, and the stereochemical characteristics of the excreted aminoalcohols IV, V and VI.