Introduction

Pharmaceutically active compounds (PhACs) are monomer compounds with medical efficacy or physiological activity. They are widely used as a common treatment for diseases in human and veterinary medicine [1]. Besides, the preventive use of antibiotic feed additives still has been existed in livestock [2]. A 2013 surveillance report by the European Centre for Disease Prevention and Control stated that β-lactams, macrolides, lincosamides and streptogramins, and tetracyclines accounted for 83.5% of the total human antibiotic sales in all 30 European countries [3]. As exogenous environmental contaminants, an increasing amount of PhACs has been released into aquatic environment with discharged human and animal excretions, causing a worldwide PhACs’ pollution [1]. Contaminated water can be introduced back into the human body through food chain. Therefore, increasing concerns have been raised for PhACs’ bioaccumulation and biomagnification in the aquatic organisms, which can lead to various health and environmental risks [4, 5]. Previous literature has demonstrated that excess PhACs might cause diseases such as prostate and breast cancer [6]. Notably, the misuse of PhACs such as antibiotics has permeated our lives through a variety of channels. Research found that nearly 60% of the residue found comes from environmental and food residues rather than drugs [7]. A 2016 report from a urine sample of 586 children from Shanghai (city located in Southeast China) supported this finding and detected 21 antibiotic residues with the overall detection frequency of 79.6% [8].

Prior research investigated the residues of PhACs in the aquatic environment. Recent studies demonstrated that compounds including sulfadiazine, sulfamethoxazole, oxytetracycline, tetracycline, trimethoprim, amoxicillin, quinolones, analgesics, and antianxietics are frequently detected at trace levels (nanograms to low micrograms per liter) in waste water [9,10,11], surface water [12,13,14,15], raw water [16, 17] and drinking water [18, 19]. For example, Tamtam et al. detected maximum contents of sulfamethoxazole at 544 ng/L in Seine River [20]. Similarly, Shen et al. [21] found that concentrations of sulfonamides and tetracyclines were up to 2680 ng/L and 1470 ng/L in Huangpu River (Shanghai City, China). The investigation conducted by Grujic et al. showed that the maximum content of azithromycin in ground water was 140 ng/L [22]. Moreover, residues of multiple pharmaceuticals had been detected in tap water in Germany [23] and China [9]. Table 1 shows a summary of LC-MS and LC-MS/MS methods for determination of relevant PhACs in water samples.

Table 1 Summary of LC-MS and LC-MS/MS methods for quantitative determination of PhACs in environmental samples
Full size table

Despite the extent evidence, the previous studies mostly focused on a particular class or several classes of pharmaceuticals in water. Little has been done to simultaneously detect frequently used PhACs. It is essential to develop accurate methods for simultaneous analyses of human and veterinary PhACs at trace levels in a wide range of aquatic environmental matrices. Research shows that simultaneous analysis allows fast, accurate, and reliable data collection on the sources of target analytes from the environment [29]. Furthermore, tests based on a limited number of PhACs classes lack scientific basis, which prohibits proper evaluation of environmental contamination and implementation of safety measures. In order to overcome this methodological drawback in research, we determine 44 PhACs including four macrolides, seventeen sulfonamides, four quinolones, chloramphenicol, eight β-lactams, four tetracyclines, lincomycin, amantadine, 4-acetamidophenol, phenylbutazone, trimethoprim, clenbuterol, and hydrocortisone. The choice of these 44 PhACs was based on the commonly consumed compounds [13, 16] and prescription data in China [30,31,32]. Furthermore, to our best knowledge this study is the first to test amantadine in aquatic environment, thus expanding the understanding of PhACs’ pollution status.

Due to the varying physicochemical properties of these chosen 44 analytes, appropriate sample pretreatment techniques must be employed. To date, a number of techniques have been used in sample pretreatment procedure to detect PhACs in aqueous environment matrices, such as liquid-liquid extraction [10], hollow fiber liquid-phase microextraction [13], and dispersive solid-phase microextraction [9]. However, solid-phase extraction (SPE) has been frequently reported as a more efficient method to concentrate, separate, and screen target analytes from aqueous samples [15, 33, 34]. Moreover, new extraction materials have been invented including Oasis HLB (styrenedivinylbenzene-vinylpyrrolidone copolymer) [35] and Oasis MCX (styrenedivinylbenzene-vinylpyrrolidone copolymer-SO3H) [36], which are proven to have achieved improved degree of precision and recoveries. Further, research has also suggested advantages of residue analytical technologies such as ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS), including improved resolution, shortened time for analysis, and direct detection of target compounds [37,38,39]. Although these methods (i.e., SPE-UPLC-MS/MS) have been widely employed in various fields, hitherto the approach of detecting 44 PhACs by UPLC-MS/MS remained to be explored.

In this paper, we aimed to develop a rapid, accurate, and sensitive analytical method to extract the 44 target compounds belonging to different therapeutical classes, which greatly facilitates the simultaneous UPLC-MS/MS determination process. We demonstrated a robust and efficient analytical method that is capable of determining 44 PhACs. This method is proven to improve extraction efficiency and minimize interferences, which raises the detection sensitivities and facilitates the studies of residue analysis in water samples. The optimized method was successfully applied to fifteen samples including raw water, treated water, and tap water, offering important evidence for distribution of selected human and veterinary drug pollution through water environment in China.

Experimental

Reagents and samples

A total of 44 PhACs and 10 isotopically labeled compounds were selected as the model analytes and surrogates and/or internal standards. The abbreviations, CAS numbers, purities, sources, and usage are displayed in Table 2. HPLC-grade methanol (MEOH), acetonitrile (MECN), and formic acid were supplied by Merck (Darmstadt, Germany). Phosphoric acid (H3PO4) was purchased from Titan (Shanghai, China). Na2EDTA was obtained from Sigma-Aldrich (St. Louis, MO). Ultrapure water was obtained by means of a Milli-Q apparatus by Millipore (Milford, MA, USA). The solid extraction columns filled with Oasis HLB (200 mg, 6 mL), MCX (200 mg, 6 mL), and WAX (200 mg, 6 mL) were supplied by Waters (Wexford, Ireland). Stocked solutions for each standard were prepared at a level of 1000 μg/mL in MEOH except for ceftiofur dissolved in MEOH/water (1:1, v/v) to get uniform solution. A calibration curve of standard solutions (0.1, 1, 5, 10, 50, 100, 500 ng/mL) was prepared from stock solutions by serial dilution with MECN/water (1:9, v/v). The internal standards mentioned before were prepared at 200 ng/mL in MEOH/water (1:1, v/v). By the means of adding blank water extracts to each serially diluted standard solution, the matrix-matched standard solutions were similarly arranged (0.1, 1, 5, 10, 50, 100, and 500 ng/mL). All standard solutions were retained in the dark at 4 °C.

Table 2 Major information of 44 target analytes and 10 isotopically labeled standards
Full size table

Raw water and treated water samples were obtained from five water companies in Hangzhou City in China. The samples of tap water were attained from five different households. All samples were collected in March 2019. Permissions for collecting the water samples were obtained.

Apparatus

The Acquity Ultra Performance LC system (Waters, Milford, USA) consisted of a degasser, a binary gradient pump, an autosampler (10 °C), and a column oven (40 °C). The analytes were detected using a Xevo TQD with Masslynx™ software (version 4.1). A 150 mm × 2.1 mm i.d. ACQUITY UPLC® BEH C18 column with 1.7-μm particles (Waters) was used for separation of 44 PhACs.

Chromatographic conditions

A mixture of (A) ultrapure water with 0.1% formic acid and (B) MECN with 0.1% formic acid was chosen as a mobile phase. The flow rate of the mobile phase was 0.3 mL/min. The following gradient program was used for the analysis: 90% A (initial), 90–80% A (1.0–1.5 min), 80–60% A (1.5–3.5 min), 60% A (3.5–6.0 min), 60–40% A (6.0–6.5 min) , 40% A (6.5–7.0 min), 40%–0% A (7.0–7.5 min), 0% A (7.5–10.5 min), 0–90% A (10.5–11.0 min). 2.5 min of equilibration was executed before the next injection. The injection volume was 10 μL, and the column temperature was maintained at 40 °C.

Mass spectrometry

Compounds were detected by multiple reaction monitoring (MRM) using a Micromass Quattro Ultima triple-quadrupole mass spectrometer (Micromass, Manchester, UK) equipped with an electrospray ionization (ESI) source. The parameters were attained for the mass spectrometry in positive and negative ion modes. In order to achieve maximum sensitivity for identification and detection of 44 PhACs, we set the typical ESI parameters as follows: the source temperature was 150 °C, while the capillary voltage was set at 3.5 kV and − 3.0 kV for positive ions and negative ions; the cone gas was 50 L/h; desolvation temperature and desolvation gas were held at 500 °C and 800 L/h respectively; nitrogen was preformed both as nebulizing and desolvation gas.

Sample pretreatment and extraction

2.5-L amber glass bottles were used for sampling, which were previously cleaned in the laboratory. All samples were immediately acidified with H3PO4 to pH 2.0 and stored at 4 °C. One-liter samples were filtered through 0.45-μm MCM membrane filters (Agela, Tianjing, China). Each water sample was spiked with 200 μL of the internal standard (200 ng/mL), then added 0.2 g Na2EDTA to prevent PhACs from complexation with metal ions before solid-phase extraction (SPE) [40]. Extraction of 44 PhACs was performed using 200-mg Oasis HLB SPE cartridges. The columns were set up in series and preconditioned successively with 6 mL of MEOH, 6 mL of Milli-Q water, and 6 mL of H3PO4 solution at pH 2.0. The samples were loaded at a flow rate of 5 mL/min. After loading, the HLB cartridge was rinsed with 6 mL deionized water. The cartridge was dried for 15 min under vacuum, and elution of the retained targets was performed with 10 mL of 2% formic acid solution in MEOH/MECN (4:1, v/v). This volume was evaporated until dryness under a nitrogen stream and redissolved in 1.0 mL of water/MECN (9:1, v/v). The extracts were centrifuged for 10.0 min at 12,000 rpm before the UPLC-MS/MS analysis. Each test corresponds at least to three individual experiments, which was executed in triplicate in each experiment.

Validation study

To evaluate the performance of the established techniques, method validation criteria such as linear dynamic range, recovery, precision, and limits of detection and quantification were determined. Matrix-matched calibration curves were obtained by the ratios of peak areas for standard and internal standard solutions at seven concentrations, ranging from 0.1 to 500 ng/mL for three runs. A total of six blank water samples were executed to verify interference from the matrix. We followed Niessen et al.’s equation (n = 6) [41] and calculated the matrix effect in order to assess the level of matrix-induced signal suppression/enhancement (ME) caused by water matrix, and the ion suppression/enhancement due to matrix effects was determined:

$$ \mathrm{Matrix}\ \mathrm{effect}\ \left(\%\right)=\left(C-B\right)/A\times 100\% $$

where A is the responses of the 44 PhAC standards in solvent, B is the responses of target analytes in unspiked water effluent extracts, and C is the responses of that spiked in water effluent extract. The LOD and the LOQ were detected as the lowest injected concentrations that produced the signal-to-noise (S/N) ratios of 3 and 10, respectively.

The extraction recoveries were determined for three replicates by analyzing spiked samples which consisted of three different concentrations of standard mixture. The precision of the analytical method was evaluated by calculating intra- and inter-day precision and accuracy, expressed as relative standard deviation (RSD %) values, as well as recoveries achieved from the spiked samples [42]. Intra-day (n = 6) and inter-day (n = 9) precision were obtained by analyzing spiked samples at different times on the same day and on the consecutive days at the concentrations of 5, 50, and 200 ng/L.

Results and discussion

Optimization of MS/MS conditions

To evaluate the mass spectral fragmentation pattern of each compound, individual standard solution (500 ng/mL) of each compound was optimized by direct injection in the spectrometer with MEOH/water (1:1, v/v) as the solvent [43]. In both positive and negative ion modes, the precursor ion with the best relative intensity was obtained by full-scan data acquisition, and its daughter ions were selected with the help of collision energy. The multiple reaction monitoring (MRM) mode was used to heighten the sensitivity and selectivity of the detection in order to monitor for each analyte [44]. The most abundant and stable daughter ion was selected as the quantitative ion and another is the qualitative ion for each target. Experimental data showed that higher precursor ion signal intensities and better fragmentation patterns were derived for PEN G, PEN V, CLX, and CAP in negative mode, which deprotonated molecular ion [M-H]–. The rest of targets were determinated in positive mode and generated an intense protonated molecular ion [M+H]+. Table 3 lists an overview of the data obtained for the 44 PhACs under the various ionization conditions.

Table 3 MRM conditions for 44 PhACs under study
Full size table

Optimization of chromatographic separation

Four chromatographic columns including Waters Acquity UPLC BEH C18 (150 mm × 2.1 mm, 1.7 μm), Waters Acquity UPLC CSH C18 (100 mm × 2.1 mm, 1.7 μm), Waters Acquity UPLC HSS T3 (100 mm × 2.1 mm, 1.8 μm), and Agilent ZORBAX Eclipse XDB-C18 (100 mm × 2.1 mm, 1.8 μm) were tested. Ultimately, BEH C18 column was selected due to good separation and retention behavior of the compounds with high sensitivity and good resolution. All the analytes were eluted at less than 10.5 min with a 12 min runtime. Consistent with literature [11, 16], MEOH and MECN with formic acid at various concentrations were investigated as mobile phases. The results indicated that MECN led to superior elution strength and decreased retention time. It is suggested that formic acid can enhance ionization efficiency in mobile phases [17]. Therefore, we eventually selected a solvent system as the mobile phases using gradient elution. Such solvent system consists of 0.1% formic acid in MECN and 0.1% formic acid aqueous solution, which afforded the most satisfied chromatographic response and promoted the ionization efficiency of mass spectrometry. The chromatograms for each target compound under optimized condition are displayed in Fig. 1.

Fig. 1
figure 1figure 1

MRM chromatogram for the 44 PhACs standards under optimum UPLC-MS/MS conditions

Full size image

Optimization of SPE cartridges

When analyzing water samples, it was highly desirable to concentrate and generate efficient extracts using SPE. Feng et al. [45] and Barbara et al. [36] had successfully applied HLB and MCX columns to extract some kinds of antibiotics. In the test, we analyzed the 44 PhACs simultaneously. Three different commercially SPE cartridges including Oasis HLB, MCX, and WAX were used. The 44 compounds investigated were from different types, such as sulfonamides, quinolones, tetracyclines, macrolides, β-lactam antibiotics, lincomides, adrenomimetics, antipyretic analgesics, and glucocorticoids. The reported values of pKa for sulfonamides, quinolones, tetracyclines, macrolides, and β-lactam antibiotics were in range of 1.4–8.4, 5.5–8.5, 3.3–9.3, 7.1–8.8, and 2.7–7.1. The pKa values of CAP were 9.5 and 11.0, while TMP were at 1.32 and 7.12 [36]. The known pK1 values of APAP and CLEN were 9.5 and 9.6. These target analytes existed as cationic, zwitterionic, and anionic groups at acidic, neutral, and basic pH, correspondingly [44]. Therefore, the 2.0–10 pH range was covered in the present research. One hundred milliliters of tap water was spiked with the mixed standard solution containing all 44 targets at 50 ng/mL. It then was adjusted to pH 2.0 with H3PO4 and passed through the three widely available SPE cartridges. The performance data of the study are summarized in Fig. 2. Excellent repeatabilities had been attained for the three SPE columns, whose RSDs for all analysis were between 0.4 and 9.6%. By comparing the contents of individual components, we found that except for APAP, CLEN, SFC, EFC, DOXY, CTC, and ROX, the recoveries of other compounds varied markedly using different cartridges. The recoveries of MCX and WAX cartridges ranged from 3.6 to 98.5% and from 3.7 to 96.4% respectively, while those for HLB cartridge varied from 76.4 to 108%. The uppermost extraction rates were gained by means of HLB cartridges, which might attribute to its sorbent. It is combined with hydrophilic-lipophilic polymer, which provided reversed-phase capacity with a special polar hook for superior capture of polar compounds [45, 46]. Guo et al. also reported the HLB column displayed upper adsorption capacity for antibiotics [47]. The recoveries obtained with WAX columns were very low as these were inferior to 20% for AMA, PHE, SBA, PEN G, PEN V, AMOX, AZ, CLX, PIP, EFT, MY, and ERY. On the contrary, MCX provided good recoveries for the majority of those mentioned before (up to 50%) but failed in recovering SDX, SBA, TMP, AMOX, and MY. In view of the best recoveries for all targets, HLB column was opted for further study.

Fig. 2
figure 2

Effect of different SPE materials on the recoveries of 44 PhACs (spiked at 50 ng/L) from water samples at pH 2.0 (n = 3)

Full size image

Eluent optimization

A series of solvents were employed to evaluate the appropriate eluent efficiency. Target analytes were extracted as described previously in the “Sample pretreatment and extraction” section. After loading the water samples onto HLB cartridges, 10 mL of several solvents was eluted, including (A) MEOH, (B) MECN, (C) MEOH/MECN (1:1, v/v), (D) MEOH/MECN (2:1, v/v), (E) MEOH/MECN (4:1, v/v), and (F) 2% formic acid solution in MEOH/MECN (4:1, v/v). Recoveries of 44 targets attained by different eluents are presented in Table 4. Initial trials MEOH and MECN (A and B) were conducted as individual eluent. Both showed poor extraction results for most compounds (less than 60%). Accordingly, different eluents (C, D, and E) were tested. Results showed that the recoveries of these three eluted targets had increased significantly, but did not vary greatly except for AMA, PHE, and TYL. As the proportion of MECN increased, the recoveries of AMA, PHE, and TYL improved steadily, especially the ones of TYL, which went from 12.0 to 57.3%. Although the extraction results attained by eluent E for the majority of the 44 compounds exceeded 70%, the extraction rates of some compounds (i.e., PHE, AMOX, AZ, PIP, CN, SFC, EFC, CFC, DFC, OTC, TET AZM, ROX, TYL, and MY) only increased significantly after adding 2% formic acid solution. In summary, excellent recoveries collected from eluent F (76.4% ± 5.8% to 102% ± 6.6%) were found to be superior to the other five solvents for all the analytes (n = 3).

Table 4 Recoveries of 44 target compounds attained by different eluents
Full size table

Method validation

Matrix effect

The ESI source is greatly impressionable to ingredients in the matrix, which may lead to ion suppression or enhancement. The results listed in Fig. 3 show that signal suppression or enhancement (ME) did exist in real water samples. The matrix effects of major compounds were from − 56.8 to 53.3% (except for PEN G, CTC, and AZM), and approximately 70% analytes were interfered by weak matrix effects (− 20 to 20%) [35]. To compensate for ME of compounds and low SPE recoveries, ten internal/surrogate standards were utilized. The choice of standards above was made on the basis of similar structure and performance in the established method. In addition, the standards of APAP, PHE, CLEN, and HC were diluted directly, for which IS/SS did not compensate for ME [1].

Fig. 3
figure 3

Matrix effect of 44 PhACs in water samples (n = 6)

Full size image

Linearity of calibration and limits of detection and quantification

The matrix-matched calibration curves were for all targets. The results of linearity, linear range, the LOD, and the LOQ are reported in Table 5. Satisfactory linearity and coefficients of determination (r2 > 0.99) were attained over the concentration range of 0.1–500 ng/mL for AMA, PHE, SPD, SDZ, SMZ, STZ, SMR, SMT, SQX, SAA, SBA, SM2, SMM, SMD, YMP, PEN G, PIP, DOXY, CTC, OTC, ROX, TYL, and MY. The rest of compounds were achieved over the range of 1.0–500 ng/mL. The results indicated approved sensitivity for the proposed method. The LOD and LOQ, determined as the minimum concentration of compounds in the spiked blank samples with a signal-to-noise ratio (S/N) of 3 and 10, ranged from 0.0111 to 0.966 ng/L and from 0.0382 to 3.26 ng/L, respectively. These results are consistent with the findings from previous studies which tested 10 and 11 antibiotics simultanously (e.g., Sergiane et al. [12]; Soparat et al. [13]).

Table 5 Calibration parameters of 44 PhACs
Full size table

Recoveries and precision

Extraction recoveries and precision assays were conducted at three different concentrations (5, 50, 200 ng/L) of the 44 target compounds. The results of these assays are reported in Table 6. RSD values lower than 10% were attained for all the samples. Generally, good recoveries (> 70%) were achieved. Recoveries ranged from 75.7 ± 4.3% (SFC) to 108 ± 9.4% (SDM). The developed methodology presented acceptable reproducibility and less interferences and background noise (n = 3). Similar recoveries of SMZ, TMP, DOXY, CTC, OTC, TET, AZM, CAP, and MY were reported by Boix et al., Liang et al., and Vergeynst et al. [8, 25, 48]. The intra- and inter-day precision were also satisfactory with RSDs being always lower than 9.6% for all compounds.

Table 6 Validation parameters of target analytes
Full size table

Application of the method

The fully optimized and validated experimental procedure was later applied to the analysis of the 44 targets from the fifteen real samples containing raw water, treated water, and tap water in Hangzhou, China. The model analytes were confirmed by the retention time, accurate mass, and MS/MS spectrum following the criteria described in European Commission Decision 2002/657/EC [49]. The levels of the 44 PhACs were quantified by the matrix-matched calibration curves. The chromatograms for a real water sample are presented in Fig. 4. It was possible to detect the 44 PhACs in all the considered aquatic environmental matrices. In fact, the results demonstrated the presence of the 17 PhACs in the water samples. There were remarkable differences among the three types of samples. Thirteen of them (AMA, SPD, SMZ, SQX, SMM, SMD, TMP, AZ, PIP, EFT, DOXY, ROX, and MY) were found in raw waters (Table 7). It was clear that all the examined samples in raw water contained one or more target contaminants, in which AMA, SMZ, ROX, and MY were more representative. The highest contents of those analytes reached 374.72, 8.86, and 3.53 ng/L, respectively. Meanwhile, the concentrations of other compounds were almost lower than 9 ng/L in raw water samples. Results show that the concentrations of sulfonamides in raw water are close to the ones found by Li et al. [10], Tang et al. [50], and Cahill et al. [51], but less than those in a study from Batt et al. [43]. Although high consumption of tetracyclines played significant role in human and veterinary medicine, we only detected trace concentration of DOXY in raw water samples, while the highest level was 1.97 ng/L. Similar findings were reported by a number of publications [2]. This phenomenon can be attributed to the formation of stable complexes by tetracyclines and bivalent or trivalent cations. Therefore, tetracyclines are more likely to remain in soil surface or combine with suspended matter or sewage slugged during the wastewater treatment. In relation to treated water and tap water, up to 13 compounds were detected in the samples analyzed. The data collected from treated water were similar to those from tap water. AMA and ROX were the compounds most frequently detected, being present in 100% and 60% of the treated water samples, respectively, while AMA was present in 80% of tap water samples analyzed. Overall, most compounds detected are veterinary PhACs, in which AMA is the most representative, which may be explained by the fact that veterinary drugs have been more extensively used for treating diseases and as feedstuff additives in the livestock industry [52]. These representations implied current status of the use of pharmaceuticals.

Fig. 4
figure 4figure 4

MRM chromatogram for the 44 PhACs in a real water sample

Full size image
Table 7 Concentration (ng/L) of 17 target analytes detected in water samples
Full size table

Conclusions

This study makes notable contributions to knowledge in the areas of methodology in pharmaceutically active compounds in water environment. To the best of our knowledge, this is the first paper to implement an efficient method for the simultaneous determination of the 44 PhACs. In particular, amantadine was detected in aquatic environment for the first time, thus expanding the understanding of PhACs’ pollution status. The whole optimized SPE-UPLC-ESI/MS process for the extraction, separation, and determination of various target analytes was further verified in water samples, with low limits of detection, satisfactory linearity, and good recoveries and reproducibility. Crucially, compared with previous studies that detected one or limited numbers of classes of pharmaceutical, we have also obtained the similar detection limits for simultaneous determination of the 44 targets. Furthermore, the proposed method was successfully applied to the analysis of the 44 PhACs in water samples, which presented great potential in the analysis of the target compounds detected. Overall, most compounds detected are veterinary PhACs, in which AMA is the most representative. This technique can be employed as a large-scale tool for monitoring exposure of the water population to the 44 PhACs. Although some PhACs found in water environment are a result of clinical or veterinary use, more epidemiological research should explore further exposure assessment of pharmaceuticals detected in water samples. This is particularly important for consumers in areas where water quality is deemed poor.