The purpose of this study was to investigate the utility of plasma pharmacokinetic and pharmacodynamic measures including plasma deoxynucleosides, homocysteine and methylmalonic acid concentrations in understanding the time course and extent of the inhibition of thymidylate synthase (TS) by pemetrexed in the context of a phase I/II combination study with vinorelbine. Eighteen patients received supplementation with folic acid and Vitamin B(12) 1 week before beginning treatment with pemetrexed and vinorelbine administered in a dose-escalating manner on a 21-day cycle. Heparinised blood samples were collected from consenting patients in the first cycle for pharmacokinetic analyses and in the first two cycles for determination of plasma thymidine, deoxyuridine, homocysteine and methylmalonic acid concentrations. These values were correlated with response and toxicity. Plasma deoxyuridine was used as a measure of TS inhibition, and concentrations of deoxyuridine were significantly elevated relative to baseline on days 1 (P<0.01), 2 (P<0.001) and 3 (P<0.05) after treatment at all pemetrexed dose levels (400-700 mg m(-2)). The magnitude of deoxyuridine elevation correlated with pemetrexed area under the plasma concentration-time curve (AUC) (r(2)=0.23, P<0.05). However, deoxyuridine concentrations returned to baseline between 8 and 15 days after treatment with pemetrexed, suggesting that inhibition of TS was not durable. Pemetrexed AUC correlated with the percentage decline (relative to baseline) in both platelets (r(2)=0.58, P<0.001) and leucocytes (r(2)=0.26, P<0.05) at day 8. Baseline homocysteine was also significantly correlated with these measures of haematological toxicity (r(2)=0.37, P<0.01 and r(2)=0.39, P<0.01, respectively). In addition, there was a significant reduction of plasma homocysteine on days 8 (P<0.005) and 15 (P<0.05) in cycle 1 compared to baseline values. The results suggest that the TS inhibitory effects of pemetrexed are short-lived and make the case for a more frequent schedule of administration such as every 2 weeks. The lack of protracted TS inhibition may be due to concomitant vitamin administration, and this may be the mechanism by which vitamins prevent life-threatening toxicity from pemetrexed. Baseline homocysteine concentration remains a predictive marker for haematological toxicity even following folate supplementation.