Jet C. Van De Steene and Willy E. Lambert Development of an LC-ESI-MS/MS analysis for nine basic pharmaceuticals in environmental waters Jet C. Van De Steene and Willy E. Lambert Laboratory of Toxicology, Ghent University, Harelbekestraat 72, 9000 Gent, Belgium Jet.Vandesteene@UGent.be OVERVIEW 9 pharmaceuticals: flubendazole, rabeprazole, cinnarizine, domperidone, ketoconazole, miconazole, itraconazole, pipamperone and propiconazole (=pesticide). LC-ESI-MS/MS analysis in surface waters Development and optimisation of: * solid-phase extraction * chromatography Evaluation of matrix effect INTRODUCTION Pharmaceutical pollution in the environment (surface water, effluents, sediments…) is of growing concern. After excretion by humans and animals, and improper disposal of unused drugs, pharmaceuticals enter the waste water treatment plant, but are not fully degraded and are disposed in the aquatic environment. Research towards this pollution is necessary because of possible effects on fauna and flora, acute or chronic [1]. The aim of this work was to develop an LC-ESI-MS/MS method for identifiying and quantifiying these pharmaceuticals in effluent and surface water samples. EXPERIMENTAL Filtration was done onto a combination of a 1.5-µm glass fibre filter and a 0.45-µm membrane filter. Before extraction, the pH of the water samples was adjusted to 7 using a 5 % ammonia-solution or a 20% acetic acid-solution. SPE:*Speedisk phenyl for extraction (JTBaker): conditioning: 3 ml MeOH and H2O; 100 ml sample; wash: 3 ml 40%MeOH in H2O; dry 2 min; elute: 2 x 0.5 ml MeOH. *NH2 for sample clean-up (Varian): conditioning: 5 ml chloroform/MeOH (80:20); extract (MeOH) diluting with 4 ml chloroform, applying to column and collect directly into centrifugation tube HPLC: Agilent 1100 Series Column: In the first experiments a Chromolith C18 column (4.6 mm i.d.x 100 mm)(Merck) was used.Then a comparison was made with a pentafluorophenyl column (4.6 mm i.d x 100 mm; 5-µm particle size) (Varian). Gradient elution with (A) H2O/ ACN (95:5) and (B) H2O/ACN (5:95), both containing 2 mM ammoniumacetate and 2 mM acetic acid Detector: triple quadrupole MS: API 4000 (Applied Biosystems) with electrospray ionisation Experiments to evaluate matrix effect were in correspondence to the strategy applied by Matuszewski et al. [2]: MS/MS responses of known amounts of standards (A) were compared with those measured in a blank water extract spiked, after extraction, with the same analyte amount (B). Absolute matrix effect(ME%): B/A x 100 ME%>100%: signal enhancement ME%<100%: signal suppression Recoveries (RE) were calculated by spiking the samples, before extraction, at a concentration of 10 ng/l(C). The MS/MS responses were compared with B: RE= C/B x 100 Method of quantification: First, structural analogues were used as internal standards (domperidone analogue (a), itraconazole analogue (b), hexaconazole (c), cambendazole (d)). Finally, the standard addition method was used. Limit of detection: S/N=3 Limit of quantification: S/N=10. RESULTS and DISCUSSION Optimisation sample preparation [3] Several SPE tested in view of recovery and matrix effect: Oasis HLB (Waters), phenyl, C8 (Varian), polymeric and SCX sorbents (Phenomenex) Oasis HLB best for recoveries but major matrix effects Phenyl second best BUT: still prominent matrix suppression!! Sample clean-up necessary because of high matrix effect: NH2 column. No retention of the analytes if extract is applied in chloroform/MeOH (80:20). less matrix effect! 2. Optimisation chromatography Post-column split To diminish matrix effects, a split was installed after the column, so less flow/sample/interferences entered source. major reduction in matrix effects! 2 columns tested: *A Chromolith C18 (4.6 mm i.d.x 100 mm) *A pentafluorophenyl column (4.6 mm i.d.x 100 mm x 5 µm particle size). with same buffers and split. PFP-column showed same results for matrix effects, but chromatography was better for the basic compounds, so the PFP-column was further used. Figure 1: chromatogram with the pentafluorophenylcolumn. From left to right: rabeprazole (tr= 5.67 min), domperidone (tr=6.89 min), flubendazole (tr=7.05 min), pipamperone (tr=7.22 min), ketoconazole (tr=8.67 min), propiconazole (tr=8.88 min), itraconazole (tr=9.70 min), miconazole (tr=11.01 min), cinnarizine (tr=11.61 min) 3. Matrix effects and quantification Matrix effect for 3 different surface water samples: with and without structural analogues as internal standards: see Table 1 different results!! Accurate quantification? No labeled internal standards available, so the standard addition method must be applied. Table 1: Matrix effect in different surface water samples (area ratio: with structural analogues: see experimental) with CV% (n=5) 4. Results For quantification the standard addition method must be used. Different matrices (surface waters, effluent, influent) have to be examined. To be sure that there is linearity in every matrix, a 10-times diluted extract is also analysed. If the result is not more than a tenth of the original extract, linearity is showed for that matrix. Recoveries and limits of detection and quantification of the whole SPE-LC-ESI-MS/MS method are summarized in Table 2. Table 2: Recoveries (surface water spiked at 200 ng/l; n=5) and limits of detection and quantification CONCLUSION An LC-ESI-MS/MS method for the quantitative analysis of nine basic pharmaceuticals in surface water was developed: sample preparation and chromatography were optimised in view of matrix effects. Matrix effects could not be ruled out so for correct quantification the standard addition method is used. LODs in the range of 0.05 and 5 ng/l. LOQs in the range of 0.05 and 10 ng/l. References [1] C.G. Daughton and Ternes T.A., Environmenal health perspectives 107: 907-938 Suppl. 6 1999. [2] B.K. Matuszewski, M.L. Constanzer, C.M. Chavez- Eng. Analytical Chemistry 75:3019-3030 2003. [3] J.C. Van De Steene, K.A. Mortier and W.E. Lambert, Journal of Chromatography A 1123:71-81 2006.