A Validation Method Development for Simultaneous LC-ESI-TOF / MS Analysis of Some Pharmaceuticals in Tangkas River-Malaysia

Fármacos são substâncias químicas sintéticas ou naturais que podem ser encontrados em medicamentos de prescrição médica, drogas terapêuticas isentas de prescrição e medicamentos veterinários. A ocorrência de fármacos no meio ambiente e no ciclo da água em quantidades vestigiais (na faixa de nanogramas a poucos microgramas por litro) tem sido amplamente discutida na literatura na década passada. O aumento na frequência de detecção é em grande parte atribuída aos avanços nas técnicas analíticas e instrumentação. Este artigo descreve o desenvolvimento, otimização e validação de um método de análise simultânea de sete produtos farmacêuticos de diferentes classes cafeína (CAF), prazosina (PRZ), maleato de enalapril (ENL), carbamazepina (CBZ), nifedipina (NFD), levonorgestre l (GNL), sinvastatina (SMV) utilizando extração em fase sólida (SPE cartuchos Oasis HLB ) seguido por cromatografia líquida acoplada a espectrometria de massas com tempo de voo e ionização por electrospray (LC-ESI-TOF/MS). A faixa linear de calibração, 0,5-250 μg L, proporcionou coeficientes de correlação linear (R2) acima de 0,99 para todos os compostos. Os limites de quantificação instrumental (IQL) para todos os produtos farmacêuticos variou de 0,5-5 μg L no solvente por injeção direta de uma mistura padrão. A eficiência de extração (EE%), para a maioria dos compostos, foi superior a 40 e 60%, em água de rio e em água pura, respectivamente. O limite de quantificação (LOQ) para todos os produtos farmacêuticos variou de 13-800 ng L para água de rio. A precisão inter e intra dia do método foi calculado, como o desvio-padrão relativo (RSD%) de 2,33-22,3% e 0,6-9,9% , respectivamente, exce to para a cafeína, que apresentou um RSD% de 20,1% a 50 μg L. O efeito da matriz variou entre 10-41%. Dos sete produtos farmacêuticos, seis compostos farmacêuticos foram detectados na amostra de água de rio.


Introduction
][6] Several instruments have been used to analyse pharmaceutical residues in water samples.][12][13][14][15][16][17][18] Although environmental applications are still scarce, several authors have reported on the application of LC-Q-TOF for the screening, confirmation and quantitative analysis of target environmental contaminants, such as pharmaceuticals phenols and pesticides.SPE is widely used in concentration step of sample preparation in the determination of contaminants (e.g., pesticides and pharmaceuticals) from environmental water samples.Depending on the choice of sorbent, a wide range of polarities and chemical classes may be covered. 19he compounds studied belong to the groups of calcium channel blockers (nifedipine), reninangiotensin (enalpril), lipid modifying agent (simvastatin), antiepileptics (carbamazepine), antihypertensive (prazosin), sex hormones (levonorgestrel) and nervous system stimulant (caffeine).The structures for all compounds are given in Figure 1.
Occurrence of human pharmaceuticals pollution in Malaysian environment has never been studied before except of very few studies; 5 therefore conducting such study is crucial to have primary information about the pollution status in Malaysia.This study is the first study investigating human pharmaceuticals and synthetic hormones in Tangkas River, Malaysia.Tangkas River is a tributary of Langat River, a main river in the distinct of Hulu Langat in the state of Selangor, Malaysia.The river flows through settlement areas of residential, schools and restaurants.The pharmaceuticals as listed in Table 1 were selected on the basis of their consumption in Malaysia 20 and environmental occurrence and persistency reported in previous studies.
In the light of these concerns, the objective of the work is to develop a new, fast, selective and sensitive analytical method for detection and quantification of a broad range of pharmaceuticals in terms of polarity in river water.
The method is based on a single SPE extraction protocol for small sample volume of 100 mL for fast sample preparation followed by LC-ESI-TOF/MS instrument analysis with 16.1 min total run time despite of using long column 250 mm with i.d. 5 µm.
Several key points, such as optimization of collision energy and elution solvent for extraction to enhance the quantitative analysis in terms of signal-to-noise ratio (S/N ratio) were discussed.
The cartridges used for solid phase extraction (SPE) were Oasis HLB (3cc, Waters, USA).Individual stock standard solutions (1000 mg L -1 ) were prepared in HPLC-grade methanol and stored at -18 °C to minimise the degradation of the standard.A mixture of all pharmaceutical standards was prepared by appropriate dilution of the individual stock solutions.Further dilutions of this mixture were prepared in 0.1% FA in (MeOH-DIW (10:90, v/v)) before each analytical run and were used as the working standard solutions.

Sampling and sample preparation
Samples are collected from Tangkas River, it is located in Kajang, a city in Selangor, Malaysia.Tangkas River is downstream of a main medicinal centre in this area.Samples were collected in March 2013 and there was no rain for at least two days prior to samples collection.All samples were collected in 1 L amber glass bottles using a Nylon polymer bucket previously rinsed with distilled water and methanol.The samples were vacuum filtered through 0.7 µm GF/F glass fiber filter and stored at 4 °C to minimize degradation of pollutants until SPE extraction.SPE of samples was carried out with a 10-sample GAST SPE vacuum manifold (DOA-P504-BN, USA).The SPE protocol was optimized through several experiments involving the following variables, sample loading flow rate, elution solvent, kind of SPE sorbent, sample size and the final solvent to reconstitute of analytes after drying (stream of nitrogen).In the light of the results of these preliminary trails, for further experiments, we select 9 mL min -1 , 5 mL of ethyl acetate, 3cc HLB Oasis cartridges, 100 mL of sample and 0.1% FA in MeOH-DIW (10:90, v/v) as a final solvent to reconstitute the analytes before injection.
The extraction efficiency of the target analytes from sample studied at two pHs, the cartridges were Oasis HLB, tested at neutral pH (without pH adjustment, pH = 7.2) and pH 2.5 using 1 mol L -1 HCl.The cartridges were preconditioned with 2 mL of ethyl acetate, 2 mL of MeOH and 2 mL of deionized water at a flow rate of 1 mL min -1 .
After the conditioning step aliquots of 100 mL of sample (without pH adjustment, pH = 7.2) were loaded into the cartridge.Samples were passed through the cartridges at a flow rate of 9.0 mL min -1 and then, rinsed with 1 mL of deionised water prior to elution.After that, the cartridges were dried under vacuum during approximately 15 min at a flow rate 14 mL min -1 to remove excess of water and finally the analytes retained were eluted with 5 mL of ethyl acetate at 1 mL min -1 .The extracts so obtained were evaporated to dryness by a gentle nitrogen stream and redissolved with 1 mL of 0.1% FA in MeOH-DIW (10:90, v/v).40 µL of the extract was automatically injected into LC-ESI-TOF/MS system for analysis.

LC-ESI-TOF/MS analysis
The LC analysis were performed using a Dionex Ultimate 3000/LC 09115047 (USA) system equipped with a vacuum degasser, a quaternary pump, an autosampler and a UV-Vis diode array detector.Chromatography was performed on a Thermo Scientific C 18 (250 mm × 2.1 mm, i.d.: 5 µm) column.The injection volume was 40 µL.All compounds were analysed in positive ion (PI) mode and eluted off the column with a mobile phase consisting of (A) 0.1% FA in DIW and (B) ACN-MeOH (3:1, v/v) at 0.3 mL min -1 .The elution started at 5% B and was then linearly increased to 60% B over 3 min, further increased to 97% B over 3 min and then kept isocratic for 5 min.Next, the elution was returned to its starting conditions over 11.1 min and allowed to equilibrate for 5 min prior to the next run.
The mass spectrometry was carried out on a TOF instrument (Bruker/Germany) equipped with a Z-spray electrospray interface.The results were obtained with the following settings: MS capillary voltages, 4000/3500 (PI/NI); collision energy for all analytes, 2-30 eV; drying-gas flow rate, 8.0 L min -1 ; drying gas temperature, 190 °C; set capillary, 4000 V; set end plate offset −500 V; set collision cell RF, 250 Vpp and nebuliser pressure, 4.0 bar.Two adduct ions, namely [M+H] + and [M+Na] + , were observed using TOF/MS analysis in positive-ion mode.The TOF results were collected between m/z 50-600 with low collision energy of 10 eV.All analytes were acquired using an independent reference spray via the LockSpray interference to ensure accuracy and reproducibility; mixture of sodium hydroxide and formic acid was used as the lock mass m/z 90.9766 -974.8132.The accurate mass was calculated using software MassLynx incorporated in the instrument.
The mass resolution (R) was calculated based on the full width at half maximum (FWHM).All pharmaceuticals have R ≥ 6000 at 1.0 µg L -1 level of spiking in river water for all pharmaceuticals.

Validation of the analytical procedure
Each compound was identified based on mass value (m/z) and retention times.Quantitation was carried out using the TOF mode, by extracting the narrow window extracted ion chromatogram (nwXIC) of the molecular ion for each compound (typically extracted using a 0.03 Da window) as reported in one study 0.02 Da. 24 Positive identification of the target compounds was based on (i) accurate mass measurement of the base ion with an error of ≤ ± 10.5 part per million (ppm) for most of compounds; (ii) LC retention time of the analyte compared to that of a standard with an error of ≤ ± 0.7% for most of compounds.
The reproducibility and repeatability of the method were evaluated from run-to-run experiments (three successive injections of a standard solution) with three different concentrations 10, 50 and 300 µg L -1 and week-to-week experiments (three successive weeks) with 100 µg L -1 .The precision of the method (in terms of peak areas and retention time) was expressed as the relative standard deviation (RSD%) of replicate measurements.All results were presented in Tables 2 and 3.
Calibration curves were generated of each analyte by injecting pooled solutions prepared from the standard mixtures (0.5-250 µg L -1 ).Calibration curves were built for each compound by plotting the peak area of each analyte against the concentration of each analyte using linear regression analysis and the concentration range that gave good fit (determination coefficients, R 2 > 0.99).The IQLs were estimated from the injection of a standard solution successively diluted until reaching a concentration level corresponding to the least concentration in calibration curves for each compound.LOQ was defined as the minimum detectable amount of an analyte in spiked river water extract giving a signal-to-noise ratio of ≥ 10.

Matrix effects
To evaluate the matrix effects (ME%) on signal intensity.SPE extracts of river water was spiked with pharmaceuticals (spike level of 200 µg L -1 ).The spiked samples were injected to LC-ESI-TOF/MS.
The matrix effect could be calculated with the following equation: (1) Vol. 25, No. 2, 2014   By this equation, the effect of sorbent is ignored so only the effect of matrix is considered, as reported elsewhere. 25here A S is the peak area of the analyte in pure standard solution, A SP is the peak area in the spiked matrix extract (after drying with nitrogen before injection) and A USP is the peak area in the un-spiked matrix extract.
In this procedure the losses of analytes caused during ionization can be evaluated.Yet excluding any losses caused by SPE and further sample preparation.ME% (+) suggests ionization suppression and ME% (-) suggests ionization enhancement.

Selection of the mobile phase and elution program
Several experiments were provided to select the best elution program and mobile phase based on running time, resolution, and signal-to-noise ratio (S/N).The optimal separation of 7 compounds detected in PI mode was achieved using a gradient elution (16.1 min) and a mobile phase ACN-MeOH (3:1, v/v).Total analysis in the PI mode (including the equilibration to the initial mobile phase conditions) was 16.1 min, which represented an approximate three-fold reduction in the analysis time compared to another study (45-min run). 26An example of the total ion chromatogram (TIC) showing the separation of compounds detected in PI mode is showing in Figure 2.

Selectivity and robustness
The selectivity of the methods was investigated by analyzing chromatograms obtained from standards individually, standards mixture and solvent without standards.The retention times for all standards were same.Robustness was studied by changing mobile phase, change place of instrument and volume injection, at all changes, the retention time and peak purity still same without differences (Table 4)

Optimization of collision energy
The collision energy was studied from 2 to 30 eV to identify the optimum value for all analytes.(Table 5) indicates that signal-to-noise ratio for all analytes was best at 10 eV, the only exception being caffeine and  enalapril, which show the best response at 5 eV.In the case of 20 eV, the S/N ratio was high for four compounds, (prazosin, carbamazepine, nifedipine and simvastatin); meanwhile, caffeine, enalapril and levonorgestrel were fragmented completely (94, 99.3 and 95.4%, respectively).This variation may be attributed to the structure of the compounds and abbility of these compounds to resist the high collision energy.Thus, 10 eV was selected as the optimum value as a good compromise between the best response for all analytes and an acceptable response for caffeine and enalapril.

Elemental composition
The accurate mass data for the molecular ions were processed using Microsoft Excel program, which provided the elemental formulas and mass errors (i.e., differences between the accurate masses and the theoretical values).Table 6 lists the exact mass measurements and mass errors obtained in the TOF mode for molecular ions.The errors obtained were between 0.26-4.08ppm or 0.05-1.8mDa which is within the widely accepted accuracy threshold of 5 ppm. 24The intensities of the mass of the compounds decreased as the collision energy increased.
Solid phase extraction: effect of pH, choice of eluent and sorbent.
Commercially available Oasis HLB polymeric sorbent is a copolymer that contains lipophilic divinylbenzene units and more hydrophilic N-vinylpyrrolidone units.The pharmaceuticals were spiked in deionised water at a concentration of 1.0 µg L -1 and the pharmaceuticals were eluted with studied solvents to select the best eluent.The studied elution solvents were 5 mL of ethyl acetate as (eluent A), 3 mL of ethyl acetate plus 3 mL of methanol as (eluent B) and 3 mL of acetone plus 3 mL of methanol as (eluent C).Ethyl acetate gave the highest extraction efficiency for all analytes with some exceptions.In average, the analytes were extracted between 121 and 28.5% with (eluent A), between 172 and 42% with (eluent B) and between 142 and 22% with (eluent C).Therefore, ethyl Based on these results, caffeine is the most polar compound and simvastatin is the least polar.The extraction efficiency of organic compounds by SPE is highly dependent on the polarity of the eluents and compounds.
The results show that the extraction efficiency for caffeine was the lowest with (Eluent A) but highest with (Eluent B and C).The extraction efficiency of enalapril was the highest with eluent C. In the case of prazosin, carbamazepine, nifedipine, the highest extraction efficiency was found with (Eluent A) rather than (Eluent B) or (Eluent C).This finding may be attributed to the polarity of the compounds versus the polarity strength of the elution solvents.However, these results were consistent with previous studies using different cartridges. 5,24ased on this result, ethyl acetate (Eluent A) was selected as a good compromise, extracting all pharmaceuticals with an extraction efficiency ranged from 30% to 127% for 100 mL river samples fortified with 1.0 µg L -1 .Ethyl acetate (5 mL) was used as the eluent in the pH optimization experiments.The extraction efficiency of the pharmaceuticals was studied at pH 2.5 and without pH adjustment (pH = 7.2).The pharmaceuticals were spiked at level of concentration 1 µg L -1 in river water and deionised water.For most of the compounds, pH did not have a pronounced effect on the extraction efficiency, with the exception of nifedipine which was poorly recovered at low pH (5% at pH 2.5).Nifedipine could be extracted at fairly high yield at pH without adjustment (40%).At pH 2.5, enalapril was extracted at high level of extraction efficiency (70%).This suggests improvement of extraction efficiency related to stability of carboxylic group in the structure of enalapril at pH 2.5.However, surface water without pH adjustment was considered for further work because of the basicity of the most compounds (pKa > 7).Therefore, no pH adjustment was performed in this study as shown in Figure 3.
Different cartridges were used to study the extraction efficiency of target compounds, including Supelclean TM ENVI-Chrom P (highly crosslinked, neutral, specially cleaned styrene-divinylbenzene co-polymer resin used to retain hydrophobic compounds with some hydrophilic functionality under reversed phase conditions), Oasis HLB cartridges (universal polymeric reversed-phase sorbent developed for the extraction of a wide range of acidic, basic, and neutral compounds) and Supelclean TM LC-SAX (quaternary amine, Cl -counter-ion, ion exchanger and reverse-phase sorbent cartridge).Thus, multiple interior structures were used among the different cartridges.As Shown in Figure 4, better extraction efficiencies were obtained for the target compounds with Oasis HLB compared to ENVI-CHROM P, despite the differences in size (3cc for Oasis HLB and 6cc for ENVI-CHROM P).LC-SAX was only effective for simvastatin and prazosin, not the other target compounds.Thus, this cartridge is recommended for the extraction of more hydrophobic and weakly acidic compounds.Based on these results, 3cc HLB Oasis cartridges were selected for further experiments.

Linearity and limit of quantifications
The effective linear dynamic ranges (R 2 > 0.99) determined for pure standards are presented in Table 8.The LC-ESI-TOF/MS method was found to be linear up to the concentration of 250 µg L -1 .Higher concentrations were not analyzed.The LOQs were determined for every compound in river water.The instrumental quantification limit (IQL) was determined to be the concentration that has the signal to noise ratio of ≥ 10.The IQLs ranged from 0.5-5 µg L -1 .The LOQs was determined to be the concentration that spiked in surface water and had the signal to noise ratio of ≥ 10.In surface water, the LOQ ranged from 13-800 ng L -1 .

Matrix Effect
Co-eluting matrix components may cause suppression of the analyte signal during electrospray ionization and therefore, the suppression of the signals of the studied components was evaluated.In the chromatogram of the river water extracts, some signal suppression was noticed for some compounds at the end of the chromatographic run (> 8 min).This indicates that the matrix constituents that elute at higher proportions of acetonitrile can severely suppress the ionization of the analytes eluting at retention times longer than 8 min.
To quantify the matrix effect, a set of experiments were performed.SPE extracts of river water were spiked (200 µg L -1 ) with the studied compounds and analyzed.The peak areas of the individual compounds were compared to the peak areas of the pure standards made in the solvent.The enhancement or suppression of the signal was calculated according to equation (1).The ion suppression was ranged from (10-41%) in river sample extracts.Some signal suppression was noticed for carbamazepine, nifedipine and levonorgestrel, i.e., 37, 41 and 24%, respectively.The ion suppression for the rest of the compounds was less than 20%.More severe signal suppression was observed during the analysis of river water, where over 24% of the signal intensity was lost for the compound having retention times longer than 8 min.At higher proportions of acetonitrile, the more lipophilic (hydrophilic) matrix components elute from the column and these were probably responsible for most of the signal suppression.The results were in line with the previous study that has been reported by Hernando et al.. 27 All method validation parameters were presented in Table 8.

Application for real samples
The studied compounds are commonly used pharmaceuticals in Malaysia (Figure 1).This new  developed method was applied to assess the occurrence of studied pharmaceuticals in Tangkas River.To make sure that the internal standard (caffeine 13 C 3 ) was not present in the water sample blank, sample without any internal standard was analyzed.The internal standard (caffeine 13 C 3 ) was not detected in real sample so this method considered more selective for studied compounds.This is the first report of human pharmaceutical pollutants in samples collected from Tangkas River, Malaysia.All pharmaceuticals detected in Tangkas River were presented in Figure 5. Results of samples analysis is presented in Table 9.In March 2013, six target analytes were determined in Tangkas River at concentrations of 257, 351, 20, 70, 50 and 85 ng L -1 for caffeine, prazosin, enalapril, carbamazepine, nifedipine and simvastatin, respectively.The rest of compounds were detected but at level of concentration < LOQ.
In this study prazosin is reported for the first time in the aquatic environment worldwide.It was found in the river with the concentration of 351 ng L -1 .The high level of caffeine (257 ng L -1 ) is not only due to the amount present in pharmaceuticals but also to its presence in some products such as chocolate, coffee, tea, or sports drinks.This widespread use results in caffeine be detected in effluent and influent of sewage treatment plants at high level of concentration in µg L -1 . 28,29ntihypertensive class such as nifedipine, was detected at low concentration, 50 ng L -1 .This may be attributed to the fact that nifedipine is easily degraded and is light sensitive so it does not show high persistence in the aquatic environment.Nifedipine was not detected in 11 rivers samples collected from Germany which is consistent with our results that showed the scarce occurrence of nifedipine in Tangkas river water. 30Enalapril was present at low concentration (20 ng L -1 ).Levonorgestrel was detected at concentration < LOQ.Carbamazepine and simvastatin were detected at 70 and 85 ng L -1 , respectively.The confirmation of the detection of caffeine, prazosin, enalapril, carbamazepine and simvastatin in the surface water was performed by comparison to the mass spectra and retention times of the standards for each compound (Table 9).The concentration of target compounds in Tangkas River; b the matrix is surface water in different countries according to the references.N.A.: not available.

Conclusion
The developed SPE-LC-ESI-TOF/MS method was precise, sensitive and accurate allowing extraction and determination of caffeine, prazosin, enalapril, carbamazepine, nifedipine, levonorgestrel and simvastatin from water samples.The recovery obtained for all target compounds using (3cc Oasis HLB) cartridges were good relative to previous studies.
The optimization of mobile phase, gradient elution program and collision energy plays an important role in enhancing the signal-to-noise ratio in terms of quantification analysis for all compounds.TOF/MS is a very sensitive detector able to detect extremely low concentrations in real samples with high accuracy and high resolution in terms of m/z value.The method performance data indicate that the techniques applied to routine analysis of surface water samples for pharmaceuticals is selective and sensitive for the majority of compounds tested with LOQ down to 13 ng L -1 .The developed method was successfully applied for the detection of six pharmaceutical residues at 40 µL volume injection with low matrix effect in Tangkas River, Malaysia.

Figure 1 .
Figure 1.Chemical structures of the studied compounds and surrogate.

Figure 3 .
Figure 3. Influence of pH adjustment, pH 2.5 versus pH without adjustment, on the extraction efficiency obtained using Oasis HLB cartridges (level of spiking is 1.0 µg L -1 in surface water).

Figure 4 .
Figure 4. Effect of cartridge sorbent on the extraction efficiency of all target compounds (level of spiking is 1.0 µg L -1 in surface water).

Figure 5 .
Figure 5. Detection of six pharmaceuticals residues in Tangkas river sample.

Table 2 .
Repeatability for all target compounds with different concentrations (n = 3)

Table 3 .
Reproducibility for all pharmaceuticals within three weeks using 100 µg L -1 of standards

Table 4 .
Effect of injection volume, lab movement and mobile phase on the robustness a Mobile phase composition (75% ACN-MeOH,v/v); b mobile phase composition (77% ACN-MeOH,v/v).

Table 6 .
Elemental composition and mass measurements for all pharmaceuticals

Table 7 .
Optimization of elution solvent to extract all pharmaceuticals (level of spiking is 1.0 µg L -1 in DIW)

Table 8 .
13nearity, instrumental quantification limits (IQLs), limit of detection (LOD), matrix effect, extraction efficiency and limit of quantification (LOQ) for pharmaceuticals in the LC-ESI-TOF/MS.Level of spiking is 1.0 µg L -1 of pharmaceutical standards and surrogate (CAF13C 3 ); b level of spiking is 200 µg L -1 of pharmaceutical standards and internal standard (CAF 13 C 3 ). a

Table 9 .
Confirmation data of all target pharmaceuticals in real sample