Determination of human-health pharmaceuticals in filtered water by chemically modified styrene-divinylbenzene resin-based solid-phase extraction and high-performance liquid chromatography/Mass Spectrometry

In 1999, the Methods Research and Development Program of the U.S. Geological Survey National Water Quality Laboratory began the process of developing a method designed to identify and quantify human-health pharmaceuticals in four filtered water-sample types: reagent water, ground water, surface water minimally affected by human contributions, and surface water that contains a substantial fraction of treated wastewater. Compounds derived from human pharmaceutical and personal-care product use, which enter the environment through wastewater discharge, are a newly emerging area of concern; this method was intended to fulfill the need for a highly sensitive and highly selective means to identify and quantify 14 commonly used human pharmaceuticals in filtered-water samples. The concentrations of 12 pharmaceuticals are reported without qualification; the concentrations of two pharmaceuticals are reported as estimates because long-term reagent-spike sample recoveries fall below acceptance criteria for reporting concentrations without qualification.

The method uses a chemically modified styrene-divinylbenzene resin-based solid-phase extraction (SPE) cartridge for analyte isolation and concentration. For analyte detection and quantitation, an instrumental method was developed that used a high-performance liquid chromatography/mass spectrometry (HPLC/MS) system to separate the pharmaceuticals of interest from each other and coextracted material. Immediately following separation, the pharmaceuticals are ionized by electrospray ionization operated in the positive mode, and the positive ions produced are detected, identified, and quantified using a quadrupole mass spectrometer.

In this method, 1-liter water samples are first filtered, either in the field or in the laboratory, using a 0.7-micrometer (μm) nominal pore size glass-fiber filter to remove suspended solids. The filtered samples then are passed through cleaned and conditioned SPE cartridges at a rate of about 15 milliliters per minute. Excess water is eliminated from the cartridge sorbent bed by passing air through the cartridges, and the analytes retained on the SPE bed are eluted from the cartridge sequentially, first with methanol, followed by acidified methanol, and combined in collection tubes. This sample extract then is reduced from about 10 milliliters (mL) to about 0.1 mL (or 100 microliters) under a stream of purified nitrogen gas with the collection tubes in a heated (40°C) water bath. The reduced extracts then are fortified with an internal standard solution (when using internal standard quantitation), brought to a final volume of 1 mL with an aqueous ammonium formate buffer solution, and filtered through a 0.2-μm Teflon syringe filter as they are transferred into vials for instrumental analysis.

Instrumental analysis by the HPLC/MS procedure permits determination of individual pharmaceutical concentrations from 0.005 to 1.0 microgram per liter, based on the lowest and the highest calibration standards routinely used. The reporting levels for this method are compound dependent, and have been experimentally determined based on the precision of quantitation of compounds from eight fortified organic-free water samples in single-operator experiments. The method detection limits and interim reporting levels for the compounds determined by this method were calculated from recoveries of the pharmaceuticals from reagent-water samples amended at 0.05 microgram per liter, and ranged between 0.0069 and 0.0142 microgram per liter, and 0.015 and 0.10 microgram per liter, respectively. Concentrations for 12 compounds are reported without qualification, and for two compounds are reported as qualified estimates. After initial development, the method was applied to more than 1,800 surface-, ground-, and wastewater samples from 2002 to 2005 and documented in a number of published studies. This research application of the method provided the opportunity to collect a large data set of ambient environmental concentrations and also permitted the collection of an extensive set of reagent blanks and spike quality-control (QC) samples. This multiple-year set of QC samples enabled further evaluation of method performance under multiple operator and multiple instrument conditions typical of routine laboratory operation. These results are an important part of the entire data set contained in this report because they document method performance over an extended time. Because sample matrix can substantially affect method performance, inclusion of environmental matrix-spike samples is required as a routine component of study plan quality control.

Method performance has been measured by long-term tracking of observed recoveries from fortified organic-free water samples processed with environmental samples (laboratory reagent spikes), as well as by observed recoveries from multiple fortified environmental water samples. The fortified environmental samples included surface water, wastewater effluent-dominated surface water, and ground water, fortified at two environmentally relevant concentrations and corrected for ambient environmental concentrations.

Because the responses of individual pharmaceuticals vary as a function of proton affinity, the ionization efficiency, and thus relative response, of each pharmaceutical, the quality-control surrogate compounds, and the quantitation internal standard can be suppressed or enhanced by the presence of the sample matrix. As a result, several quality-control sample types are required to properly interpret the ambient environmental concentrations of pharmaceuticals in aqueous samples. The quality-control sample types and results include laboratory reagent spikes and laboratory reagent blanks to provide insight into the performance of the method in the absence of a sample matrix, and matrix-spike recovery samples and replicate environmental samples, collected from representative sample matrix types within the aquatic system under study.