Flavin-containing monooxygenases (FMO) are a family of microsomal enzymes that catalyse the oxygenation of xenobiotics [1]. Six FMO genes exist in humans, but only FMO3 enzyme is abundantly expressed in human liver. FMO3 contributes to the metabolism of several drugs including methimazole, nicotine, cimetidine, ranitidine, tamoxifen, and clozapine [1,2]. Dysfunctional FMO3 enzymes with deficient N-oxygenation of dietary trimethylamine cause trimethylaminuria (fish-odour syndrome).FMO3 is the main enzyme responsible for nicotine N ¢ -oxide formation, which is a minor pathway of nicotine metabolism. The FMO content of human liver microsomes is correlated with nicotine N ¢ -oxide formation, and N ¢ -oxygenase activity is abolished by FMO inhibitors in vitro [3]. Also, cDNA-expressed FMO3 is active in nicotine N ¢ -oxide formation [4]. A preliminary report on two siblings with fish-odour syndrome showed impaired urinary excretion of nicotine N ¢ -oxide compared with normal subjects [5]. Methimazole, an antithyroid agent and FMO3 inhibitor, has been reported to reduce nicotine N ¢ -oxide excretion [6]. Thus, formation of nicotine N ¢ -oxide can be used as a probe of human FMO3 activity [1].We recently published a study on pregnant smokers investigating the effect of pregnancy on nicotine and cotinine metabolism [7]. Ten subjects were given deuterium-labelled nicotine-d 2 and cotinine-d 4 infusion (30 or 45 m g/kg -1 of each compound) during pregnancy and postpartum. The systemic clearance was increased by 60% and 140% for nicotine and cotinine, respectively, in pregnancy compared with postpartum. This increase is most likely mediated by the induction of cytochrome P450 2A6, the main enzyme metabolizing nicotine and cotinine [8].Urine samples were collected and nicotine and metabolite concentrations were analysed by gas chromatography -mass spectrometry [7]. We have now reanalysed the data to examine the effects of pregnancy on the urinary ratio of nicotine N ¢ -oxide/nicotine, as an index of FMO3 activity. Ten subjects completed the study; a urine sample from one subject was lost. We measured both nicotine N ¢ -oxide-d 2 and nicotine-d 2 concentrations in urine samples collected over 8 h following the infusion in nine subjects during pregnancy and postpartum. In two subjects the calculation of the urine ratio during pregnancy was not possible because the concentration of nicotine (one subject) or concentration of nicotine N ¢ -oxide (one subject) was below the limit of quantification. In the remaining subjects the ratio of nicotine N ¢ -oxide/nicotine was substantially increased during pregnancy compared with postpartum (2.03 vs. 0.78; paired t-test, P = 0.017; 95% CI of the difference 0.5-2.0) ( Table 1).A possible confounding factor is that renal clearance of nicotine is pH-dependent, whereas nicotine N ¢ -oxide excretion is not. On average, pregnancy increases urine pH by 0.3-0.4 units [9] and might increase the nicotine N ¢ -oxide/nicotine ratio by decreasing nicotine excretion. Urine pH values were not recorded...