Abstract

Assessing collective drug consumption based on the concentrations of illicit drugs and their metabolites in wastewater is a new technology. Currently, this technology is receiving attention in China, and methods for multiple illicit drug detection in wastewater are urgently needed. In our study, a method with a short runtime (7 min), a small solid-phase extraction (SPE) loading volume (50 mL) and high sensitivity (lower limits of quantitation (LLOQs) ranged from 0.2 to 5 ng/L) was developed for the simultaneous determination of amphetamines, ketamine, opiates, cocaine and their metabolites in wastewater. Samples were enriched by SPE on a mixed-mode sorbent (Oasis MCX) and analyzed by ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). The limits of detection (LODs) ranged from 0.1 to 2 ng/L, and the LLOQs varied between 0.2 and 5 ng/L. Moreover, the method developed was applied to real wastewater samples collected from 15 different wastewater treatment plants (WWTPs). In the results, the most abundant compounds were morphine (1.8–46.6 ng/L) and codeine (3.7–24.9 ng/L), which were detected in 13 WWTPs. After successful optimization of the UPLC-MS/MS conditions and sample loading pH, the method developed is able to meet the needs of common illicit drug monitoring and high-throughput analysis requirements.

HIGHLIGHTS

  • SPE-UPLC-MS/MS method was developed for the for determination of 11 illicit drugs in wastewater.

  • The method has a smaller SPE loading volume, a shorter UPLC runtime and higher sensitivity than other methods.

  • The method can meet the requirements for current drug abuse monitoring in China.

INTRODUCTION

In recent years, the use of illicit drugs has become more and more serious and has incalculable societal consequences. Monitoring of illicit drug consumption was first carried out through demographic surveys combining medical records and crime statistics (Hosogi et al. 2014). However, this approach had some shortcomings, including time lags and demographic limitations, which may lead to incorrect estimations or underestimations of drug use rate (Ohashi et al. 2015). In 2001, a new idea, that analysis of illicit drugs and their metabolites in wastewater could be used to assess collective drug consumption, was raised by Daughton (2001). In this idea, investigators can gain quantitative, real-time and objective measurements of target illicit drug consumption by monitoring selected illicit drugs in wastewater.

Since then, studies of illicit drugs and their metabolites in wastewater have been performed in various countries, mainly in Europe, North America, and Australia. Krizman-Matasic et al. developed a method for simultaneous analysis of opioid analgesics and their metabolites in wastewater in Yugoslavia (Krizman-Matasic et al. 2017). González-Mariño et al. developed a method of selective quantification of illicit drugs in wastewater in Spain (González-Mariño et al. 2012). Baker & Kasprzyk-Hordern used a method of analyse 65 illicit drugs in wastewater in England (Baker & Kasprzyk-Hordern 2011). Devault et al. evaluated illicit and licit drug consumption based on wastewater analysis in France (Devault et al. 2014). Lai et al. determined cocaine, MDMA and methamphetamine residues in wastewater in Australia (Lai et al. 2016). Centazzo et al. analysed for nicotine, cocaine, amphetamines, opioids and cannabis in the USA (Centazzo et al. 2019).

The drug use patterns and epidemics differ from region to region. Based on the China drug situation report (Pan et al. 2019), amphetamines have become the most popular illicit drugs, followed by opiates and ketamine. There have been several articles and applications regarding wastewater-based epidemiology studies in China (Du et al. 2017; Wang et al. 2019). Nevertheless, only few illicit drugs or class of illicit drugs have been covered and therefore do not meet the requirements for current drug abuse monitoring in China. A method for multiple illicit drug detection in wastewater is therefore urgently needed and is receiving attention in China. Many cities in China are planning to use wastewater analysis to improve monitoring and control capabilities, and evaluate the effectiveness of its drug-reduction programs. A multi-target method covering commonly abused drugs and their metabolites facilitates the comparison of drug uses between different cities.

In most cases solid-phase extraction (SPE) is conducted as a pretreatment method. A small SPE loading capacity is critical when sensitivity is satisfied because it can save a lot of time and better meet the high-throughput analysis requirement of illicit drug monitoring. The occurrence of illicit drugs and their metabolites in wastewater has been determined so far by liquid chromatography-tandem mass spectrometry. This procedure also needs a short runtime.

Therefore, the purpose of our study was to develop a method with a short runtime, a small SPE loading volume, and high sensitivity to simultaneously analyze amphetamines, opiates, ketamine, cocaine and their metabolites in wastewater. These include amphetamine (AMP), methamphetamine (MAMP), 3,4-methylenedioxyamphetamine (MDA), 3,4-methylenedioxymethamphetamine (MDMA), ketamine (K), norketamine (NK), morphine (MOR), 6-acetylmorphine (6-AM), codeine (COD), cocaine (COC) and benzoylecognine (BZE). This method was applied to real wastewater samples collected from 15 wastewater treatment plants (WWTPs).

METHODS

Chemicals and reagents

MDA, MDMA, K, NK, MOR, COC, BZE, AMP, MAMP, COD, 6-AM, AMP-d5, MAMP-d5, NK-d4, 6-AM-d6, K-d4, MDMA-d5, MOR-d3, BZE-d8, MDA-d5 and COC-d3 were acquired from Cerilliant (Round Rock, TX, USA) as 0.1 or 1 mg/mL standards in methanol. The mixed stock standard solution (1 μg/mL) and internal standard solution (10 μg/mL) were diluted with methanol. Working standard solutions with concentrations of 50 and 5 ng/mL in methanol were prepared through serial dilutions of the stock standard solutions. A working internal standard solution of 20 ng/mL was generated by dilution in methanol.

High-performance liquid chromatography-grade acetonitrile and methanol were acquired from Sigma-Aldrich (St Louis, MO, USA). Hydrochloric acid (analytical grade) was obtained from Yonghua Chemistry Co. (Jiangsu, China). Formic acid (98%) was obtained from Fluka (Buchs, Switzerland). Ammonium hydroxide solution (25%) was purchased from Aladdin Chemistry Co. (Shanghai, China). Ultrapure water was prepared by an in-house Milli-Q water system (Millipore, MA, USA). Oasis MCX (3 cc/60 mg) cartridges were purchased from Waters Corporation (Waters, USA). Glass microfiber filter papers (GF/C, 1.2 μm) were acquired from Whatman (Kent, UK). PTFE filters (0.22 μm) were purchased from Shanghai Guoyao Chemical Reagent Co. (Shanghai, China).

Sample collection and preparation

Wastewater samples of 24 hour composites were collected in 600 mL PET bottles from different WWTPs. After collection, all wastewater samples were acidified with hydrochloric acid to pH 2 and then transported to the laboratory for analysis. Samples were stored at −20 °C until pretreatment.

The method was optimized and validated using drug-free wastewater that was collected from a WWTP in a small town in China.

Sample pretreatment

All samples were thawed to room temperature, and then filtered through glass microfiber filter papers, and the pH was adjusted if necessary. A mixture of isotope-labeled internal standards (100 μL of 20 ng/mL) was added to the 50 mL wastewater samples prior to SPE.

The SPE of the wastewater samples was carried out with a Gilson SPE manifold (GX-274). Initially, Oasis MCX cartridges were conditioned with methanol (4 mL) and acidified water (4 mL, pH 2 by hydrochloric acid) both at a flow rate of 4 mL/min. Then, the wastewater samples passed through the cartridge at flow rate of 2 mL/min. Before elution, the cartridges were washed with water (4 mL) and methanol (4 mL) at a flow rate of 2 mL/min and dried under vacuum. Analytes were eluted from the cartridge using 4 mL of 5% ammonium hydroxide in methanol at a flow rate of 1 mL/min, and then the extracts were evaporated to approximately 200 μL under nitrogen stream at 50 °C. Finally, the samples were filtered through 0.22 μm PTFE filters before injection into the ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) system. The wastewater samples were analyzed in duplicate.

UPLC-MS/MS analysis

Chromatographic separation was performed using an Acquity Ultra Performance LC (Waters, USA). In the final procedure, the target compounds were separated on a Waters Acquity TM UPLC HSS T3 column (100 mm × 2.1 mm, 1.8 μm) at a flow rate of 0.3 mL/min. Formic acid (0.1%) in water was mobile phase A. Acetonitrile was mobile phase B. A gradient elution program was applied as follows: 0–1 min, 4% B; 1–5 min, from 4% to 50% B; 5–5.4 min, from 50% to 90% B; 5.4–5.9 min, 90% B, 5.9–6 min, from 90% to 4% B, 6–7 min, 4% B. The total runtime was 7 min. The injection volume was 20 μL.

Mass spectrometric analysis was performed using a Sciex 6500 Plus Q-trap quadrupole mass spectrometer (AB Sciex, USA). The parameters of the ion source were as follows: temperature, 500 °C; collision activation dissociation (CAD) gas, high; curtain gas (CUR), 35 psi; ion spray voltage, 5,500 V; nebulizing gas (GS1), 50 psi; heater gas (GS2), 50 psi. The analytes were analyzed in positive mode using multiple reaction monitoring (MRM) mode. The MRM parameters are shown in Table 1. The data were evaluated by MultiQuant 3.0.2.

Table 1

Optimized MRM transitions and retention times for selected compounds and their isotope-labeled internal standards

CompoundRetention time (min)Precursor ion ([M + H]+)Product ionDelustering potential (V)Collision energy (eV)
AMP 3.78 136.1 119.1 40 11 
91.1 40 22 
MAMP 3.94 150.1 119.1 40 14 
91.1 40 23 
MDA 3.90 180.1 105.0 50 30 
135.1 50 24 
MDMA 4.02 194.2 163.4 40 16 
105.0 40 31 
4.17 238.1 179.1 60 24 
125.1 60 35 
NK 4.09 224.1 207.1 60 15 
125.1 60 28 
MOR 3.04 286.1 201.2 80 35 
165.3 80 52 
6-AM 3.85 328.1 211.3 80 36 
165.3 80 48 
COC 4.69 304.1 182.2 60 28 
150.2 60 35 
BZE 4.25 290.2 168.3 70 26 
105.2 70 43 
COD 3.60 300 199.2 80 40 
165.3 80 52 
AMP-d5 3.77 141.7 93.3 38 22 
MAMP-d5 3.92 155.2 91.1 40 23 
MDA-d5 3.88 185.0 168.0 60 15 
MDMA-d5 4.01 199.1 165.2 40 18 
K-d4 4.15 242.2 183.1 60 26 
NK-d4 4.07 228.3 183.0 60 23 
MOR-d3 3.03 289.2 201.1 80 35 
6-AM-d6 3.84 334.2 165.2 60 50 
COC-d3 4.69 307.1 185.3 60 27 
BZE-d8 4.24 298.3 171.3 60 27 
CompoundRetention time (min)Precursor ion ([M + H]+)Product ionDelustering potential (V)Collision energy (eV)
AMP 3.78 136.1 119.1 40 11 
91.1 40 22 
MAMP 3.94 150.1 119.1 40 14 
91.1 40 23 
MDA 3.90 180.1 105.0 50 30 
135.1 50 24 
MDMA 4.02 194.2 163.4 40 16 
105.0 40 31 
4.17 238.1 179.1 60 24 
125.1 60 35 
NK 4.09 224.1 207.1 60 15 
125.1 60 28 
MOR 3.04 286.1 201.2 80 35 
165.3 80 52 
6-AM 3.85 328.1 211.3 80 36 
165.3 80 48 
COC 4.69 304.1 182.2 60 28 
150.2 60 35 
BZE 4.25 290.2 168.3 70 26 
105.2 70 43 
COD 3.60 300 199.2 80 40 
165.3 80 52 
AMP-d5 3.77 141.7 93.3 38 22 
MAMP-d5 3.92 155.2 91.1 40 23 
MDA-d5 3.88 185.0 168.0 60 15 
MDMA-d5 4.01 199.1 165.2 40 18 
K-d4 4.15 242.2 183.1 60 26 
NK-d4 4.07 228.3 183.0 60 23 
MOR-d3 3.03 289.2 201.1 80 35 
6-AM-d6 3.84 334.2 165.2 60 50 
COC-d3 4.69 307.1 185.3 60 27 
BZE-d8 4.24 298.3 171.3 60 27 

The quantifier ions are in bold.

Method validation

Method validation included linearity, limits of detection (LODs), lower limits of quantitation (LLOQs), extraction recovery, matrix effect, precision and accuracy.

Linearity was determined by linear regression with 1/x weighting. The linearity range was acceptable if the correlation coefficient (r2) was greater than 0.99. The LODs and LLOQs were determined with spiked wastewater having decreasing concentrations of the analytes. The LODs were defined as the concentration yielding a signal-to-noise ratio (S/N) of at least 3:1, while the LLOQs were the lowest concentration yielding an S/N of at least 10:1.

The extraction recovery and matrix effect of the target compounds were determined using wastewater spiked with the analytes at two concentrations (5 and 50 ng/L) with 12 replicates of each concentration in the following groups: set 1, wastewater spiked with the standard solution before SPE; set 2, extracts spiked with the standard solution after SPE; set 3, the standard solution mixed into solvent. Extraction recovery and matrix effect were calculated using the following equations:
formula
formula
where Aset1, Aset2 and Aset3 all represent analyte peak areas.

Method accuracy and precision were determined by analyzing two concentrations of wastewater: 5 ng/L and 50 ng/L, with 12 replicates of each. The accuracy was calculated based on the percentage ratio of the measured concentration to the nominal concentration. The precision was evaluated by the relative standard deviation (RSD) of the analysis of spiked samples.

RESULTS AND DISCUSSION

UPLC-MS/MS method optimization

Three different high-performance liquid chromatography columns were tested: a Waters Acquity TM UPLC HSS T3 column (100 mm × 2.1 mm, 1.8 μm), an Allure PFPP column (100 mm × 2.1 mm, 5 μm), and an Agilent Eclipse Plus C18 column (100 mm × 2.1 mm, 3.5 μm). Additionally, different mobile phases were tested: acetonitrile or methanol and water, with different modifiers, including: formic acid, ammonium formate, acetic acid, ammonium acetate, among others. The peak shapes, retention time, separation and sensitivity of the chromatographic procedure were evaluated. Finally, a Waters Acquity TM UPLC HSS T3 column and a mobile phase consisting of water (0.1% formic acid) and acetonitrile were used because of good peak shapes and separation. Figure 1 shows the chromatographic separation of each compound in this method.

Figure 1

MRM chromatogram of standard mixture of compounds (5 ng mL).

Figure 1

MRM chromatogram of standard mixture of compounds (5 ng mL).

Optimization of samples loading pH

Three different pH values (pH = 2, pH = 5 and pH = 7) of the samples were tested for SPE cartridge loading. Figure 2 shows the extraction recovery of each compound at the three different pH values. Overall, the poorest extraction recovery was observed at pH 7. The highest extraction recovery for 9 of the 11 substances was provided with samples acidified to pH 2. Therefore, a pH of 2 was used for loading samples.

Figure 2

Comparison of extraction recoveries of target compounds at different pH.

Figure 2

Comparison of extraction recoveries of target compounds at different pH.

Method validation

Method parameters, including linearity, LODs, LLOQs, extraction recovery, matrix effect, precision and accuracy, were validated for wastewater samples, and the results are shown in Table 2. The linear range was found to be 0.2–200 ng/L for COC, 0.5–200 ng/L for K, NK and BZE, 1–200 ng/L for MAMP, MDMA and MOR, 1.5–200 ng/L for COD and 6-AM, 2–200 ng/L for AMP, and 5–200 ng/L for MDA, and the correlation coefficients (r2) were all greater than 0.99. Regarding the LODs and LLOQs in wastewater, the LODs ranged from 0.1 to 2 ng/L and the LLOQs ranged from 0.2 to 5 ng/L.

Table 2

Method validation parameters for the determination of target compounds in wastewater (n = 12)

CompoundLinearity ranger2LOD (ng/L)LLOQ (ng/L)Spiking level 5 ng/L
Spiking level 50 ng/L
Extraction recovery (%)Matrix effect (%)Accuracy (%)Precision (%)Extraction recovery (%)Matrix effect (%)Accuracy (%)Precision (%)
AMP 2–200 0.99897 1.0 2.0 105 −10 100 99 −10 94 
MAMP 1–200 0.99865 0.3 1,0 99 14 95 98 95 
MDA 5–200 0.99898 2.0 5.0 96 102 10 99 102 
MDMA 1–200 0.9993 0.5 1.0 99 −9 100 95 −10 98 
0.5–200 0.99909 0.2 0.5 103 −11 100 99 −13 99 
NK 0.5–2 0.99977 0.2 0.5 98 −8 101 97 −17 100 
MOR 1–200 0.99716 0.5 1.0 99 −18 99 97 −15 102 
6-AM 1.5–200 0.99919 1.0 1.5 98 −3 106 89 −9 104 
COC 0.2–200 0.9994 0.1 0.2 99 −2 101 98 103 
BZE 0.5–200 0.99895 0.2 0.5 94 −33 100 98 −29 104 
COD 1.5–200 0.99844 1.0 1.5 95 −3 96 97 −6 100 
CompoundLinearity ranger2LOD (ng/L)LLOQ (ng/L)Spiking level 5 ng/L
Spiking level 50 ng/L
Extraction recovery (%)Matrix effect (%)Accuracy (%)Precision (%)Extraction recovery (%)Matrix effect (%)Accuracy (%)Precision (%)
AMP 2–200 0.99897 1.0 2.0 105 −10 100 99 −10 94 
MAMP 1–200 0.99865 0.3 1,0 99 14 95 98 95 
MDA 5–200 0.99898 2.0 5.0 96 102 10 99 102 
MDMA 1–200 0.9993 0.5 1.0 99 −9 100 95 −10 98 
0.5–200 0.99909 0.2 0.5 103 −11 100 99 −13 99 
NK 0.5–2 0.99977 0.2 0.5 98 −8 101 97 −17 100 
MOR 1–200 0.99716 0.5 1.0 99 −18 99 97 −15 102 
6-AM 1.5–200 0.99919 1.0 1.5 98 −3 106 89 −9 104 
COC 0.2–200 0.9994 0.1 0.2 99 −2 101 98 103 
BZE 0.5–200 0.99895 0.2 0.5 94 −33 100 98 −29 104 
COD 1.5–200 0.99844 1.0 1.5 95 −3 96 97 −6 100 

The extraction recoveries were higher than 85% for all compounds in wastewater samples at 5 ng/L and 50 ng/L (89–105%). In addition, during the evaporation step, a large amount of loss was observed if the extracts were evaporated to dryness, especially for AMP and MAMP, which had smaller molecular weights. Therefore, the extracts were evaporated to approximately 200 μL to reduce losses.

The matrix effect was tested for all target compounds at 5 ng/L and 50 ng/L. For most compounds, signal suppression was observed, and only two compounds, MAMP and MDA, showed signal enhancement (from 2 to 14%). Eight compounds, 6-AM, MDMA, MOR, AMP, COC, COD, K and NK, showed signal suppression between −2 and −18%, and only BZE exhibited a higher matrix effect (−33%).

The method showed good accuracy and precision for the target compounds at 5 ng/L and 50 ng/L with isotope-labeled internal standards. The precision ranged from 2% to 7% at 50 ng/L, and increased to 3% to 10% at 5 ng/L. The accuracy ranged from 94% to 106% for all compounds.

Our method meets the needs of common illicit drug monitoring after optimization of the sample loading pH and UPLC-MS/MS conditions. A 50 mL SPE loading volume, 7 min runtime, and high sensitivity (LLOQs ranged from 0.2 to 5 ng/L) were achieved with this method. González-Mariño et al. used a 200 mL wastewater SPE loading volume and 40 min run time with LLOQs ranging from 5 to 25 ng/L for selective quantification of illicit drugs in wastewater (González-Mariño et al. 2012). In a method developed by Devault et al. for the evaluation of illicit drug consumption based on wastewater analysis, the SPE loading volume was 200 mL, the liquid chromatography runtime was 30 min, and LLOQs ranged from 10 to 40 ng/L (Devault et al. 2014). A SPE loading volume of 100 mL, liquid chromatography runtime of 13 min, and LLOQs of 10–250 ng/L were reported by Pedrouzo et al. for determining illicit drugs and their metabolites in wastewater (Pedrouzo et al. 2011). In most other methods with most of the same target compounds (Baker & Kasprzyk-Hordern 2011; Pedrouzo et al. 2011; González-Mariño et al. 2012; Borova et al. 2014; Devault et al. 2014; Lopes et al. 2014; Senta et al. 2015; Boles & Wells 2016), the SPE loading volume was larger than 100–200 mL, and the liquid chromatography runtime was long.

Compared to the published methods, our method achieved a smaller SPE loading volume (50 mL of wastewater), shorter UPLC runtime (7 min), and higher sensitivity ((LLOQs ranged from 0.2 to 5 ng/L). Although same SPE loading volume (50 mL) was utilized in some studies (Centazzo et al. 2019; Mercan et al. 2019), longer UPLC runtimes (10 and 14 min) were used or lower sensitivities (LLOQs ranged from 3.12 to 82.1 ng/L) were achieved in these studies compared to in ours. Therefore, our method can better meet the high-throughput analysis requirement of illicit drug monitoring compared to previously developed methods.

Method application

The occurrence and level of the target compounds in 15 WWTPs were evaluated using the method developed (Table 3). The measured concentrations varied widely, from <LLOQ to 46.6 ng/L. The most abundant compounds were MOR (1.8–46.6 ng/L) and COD (3.7–24.9 ng/L), which were detected in 13 WWTPs. As shown in Figure 3, MOR and COD had the highest concentrations compared to other compounds in the wastewater samples, and 6-AM was not detected. In wastewater analysis, MOR and 6-AM are the target residues of heroin for estimating the consumption of heroin. However, according to a report (Chen et al. 2013), only 0.01% of MOR in wastewater is produced by illegal use of heroin. In addition, relatively high concentrations of MOR and COD may be due to the ingestion of opiate-containing medications, including brown mixture (Liu et al. 2009). The ingestion of these products can cause the detection of morphine and codeine in urine. Therefore, reasonable estimates of the consumption of heroin should be made in conjunction with relevant hospital data. In China, Du et al. (2017) estimated heroin consumption for the first time and used MOR as a target residue. In the process, therapeutic MOR and COD were considered and subtracted.

Table 3

Concentration of target compounds in 15 WWTPs

Concentration (ng/L)
WWTPCODMDAMDMACOCBZEMOR6-AMMAMPAMPKNK
3.7 <LLOQ <LLOQ <LLOQ <LLOQ 4.7 <LLOQ 2.5 <LLOQ <LLOQ <LLOQ 
11.7 <LLOQ <LLOQ <LLOQ <LLOQ 19.7 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
<LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 1.8 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
7.2 <LLOQ <LLOQ <LLOQ <LLOQ 25.6 <LLOQ 4.5 <LLOQ 1.4 <LLOQ 
6.0 <LLOQ <LLOQ <LLOQ <LLOQ 23.8 <LLOQ 2.6 <LLOQ 3.8 <LLOQ 
12.0 <LLOQ <LLOQ <LLOQ <LLOQ 19.5 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
13.7 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
10.5 <LLOQ <LLOQ <LLOQ <LLOQ 26.8 <LLOQ 2.4 <LLOQ <LLOQ <LLOQ 
<LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 10.0 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
10 22.5 <LLOQ <LLOQ <LLOQ <LLOQ 46.6 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
11 24.9 <LLOQ <LLOQ <LLOQ <LLOQ 41.8 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
12 20.6 <LLOQ <LLOQ <LLOQ <LLOQ 45.9 <LLOQ 3.9 <LLOQ <LLOQ <LLOQ 
13 10.2 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
14 14.1 <LLOQ <LLOQ <LLOQ <LLOQ 26.6 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
15 4.9 <LLOQ <LLOQ <LLOQ <LLOQ 22.8 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
Concentration (ng/L)
WWTPCODMDAMDMACOCBZEMOR6-AMMAMPAMPKNK
3.7 <LLOQ <LLOQ <LLOQ <LLOQ 4.7 <LLOQ 2.5 <LLOQ <LLOQ <LLOQ 
11.7 <LLOQ <LLOQ <LLOQ <LLOQ 19.7 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
<LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 1.8 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
7.2 <LLOQ <LLOQ <LLOQ <LLOQ 25.6 <LLOQ 4.5 <LLOQ 1.4 <LLOQ 
6.0 <LLOQ <LLOQ <LLOQ <LLOQ 23.8 <LLOQ 2.6 <LLOQ 3.8 <LLOQ 
12.0 <LLOQ <LLOQ <LLOQ <LLOQ 19.5 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
13.7 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
10.5 <LLOQ <LLOQ <LLOQ <LLOQ 26.8 <LLOQ 2.4 <LLOQ <LLOQ <LLOQ 
<LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 10.0 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
10 22.5 <LLOQ <LLOQ <LLOQ <LLOQ 46.6 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
11 24.9 <LLOQ <LLOQ <LLOQ <LLOQ 41.8 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
12 20.6 <LLOQ <LLOQ <LLOQ <LLOQ 45.9 <LLOQ 3.9 <LLOQ <LLOQ <LLOQ 
13 10.2 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
14 14.1 <LLOQ <LLOQ <LLOQ <LLOQ 26.6 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
15 4.9 <LLOQ <LLOQ <LLOQ <LLOQ 22.8 <LLOQ <LLOQ <LLOQ <LLOQ <LLOQ 
Figure 3

Comparison of target compound concentrations in 15 WWTPs.

Figure 3

Comparison of target compound concentrations in 15 WWTPs.

Other relatively abundant compounds found in the WWTPs included MAMP and K. Amphetamines have become the most popular illicit drugs in China (Pan et al. 2019). MAMP was detected in five WWTPs, and the concentration varied between 2.4 ng/L and 4.5 ng/L; however, AMP was not detected. Additionally, MDA and MDMA were not detected in the WWTPs.

According to a previous study (Li et al. 2014), the concentrations of MAMP and AMP are strongly correlated, and AMP excretion is 4–7% in urine after MAMP ingestion. In addition, there are no medicines containing components with the potential to metabolize directly into AMP in China. Therefore, AMP was not detectable with the current methods in our results according to the concentration of MAMP. Shao et al. (2020) determined the concentration of MAMP in 22 Chinese cities. MAMP concentrations ranged from 42.6 ng/L (Harbin) to 700 ng/L (Xi-an) in wastewater samples. AMP concentrations varied between 14.2 ng/L (Yinchuan) and 173 ng/L (Lanzhou). In Cookeville (USA), MAMP concentrations varied between 27 ng/L and 60.3 ng/L and AMP concentrations ranged from 54.1 ng/L to 86.4 ng/L (Boles & Wells 2016). In different countries and regions, illicit drugs exhibit different epidemic characteristics in wastewater.

Ketamine is a popular illicit drug in Southeast Asian countries (Pan et al. 2019). K was detected in two WWTPs and the concentrations were 1.4 and 3.8 ng/L. No NK was found. K use was monitored by tracking concentrations of K and NK. NK is the metabolites of K, so NK is directly linked to human consumption of K. However, when the use of K is low, K is more suitable because a larger amount of K is excreted unchanged, and it is more likely to be detected. González-Mariño et al. (2012) determined the concentration of illicit drugs in a city in northwest Spain, but K was not detected. In New York State (USA), K was also not detected (Asimakopoulos et al. 2017). Khan et al. (2014) collected and analyzed wastewater samples from four cities in China: Shanghai, Shenzhen, Beijing, and Guangzhou. In their study, the concentrations of K ranged from 6 ng/L to 500 ng/L, which were higher than the concentrations observed in our study. This may have been dependent on the location of the wastewater samples.

In addition, COC and its metabolite BZE were not detected in the WWTPs. COC is not widely used in China because of its high cost.

Factors of results

The concentrations of illicit drugs in wastewater might be affected by many factors, such as sampling process, sewer environment, the diffusion and dispersion effects of the target compounds. Sampling is a crucial step in wastewater analysis, and scientific sampling is a prerequisite for getting accurate monitoring data. There are many ways to sample, including the flow-proportional continuous sampling method, the constant continuous sampling method, and the grab sample method (Ort et al. 2010). In our study, we took 500 mL samples every 2 hours, then mixed 12 samples collected on the same day, and finally collected 24-hour composite samples in 600 mL PET bottles. This sampling method can effectively reduce sampling errors compared with other methods, and lead to accurate monitoring data.

The transformation, degradation, stability and masking effect of illicit drug residues in the sewer system is an important aspect that should be considered. Thai et al. simulated the sewer environment in Australia in the laboratory to investigate the stability of illicit drugs (Thai et al. 2014). The results showed that the degradation rates of some illicit drugs increased significantly in sewers, compared to the values measured in wastewater alone. However, the sewer environment is different in different countries and regions. No article studied the impact of the sewer environment in China on the transformation, degradation, stability and masking effect of illicit drugs. Our next study is to explore the transformation, degradation, stability and masking effect of amphetamines, ketamine, opiates, cocaine, and their metabolites in China's sewer environment.

Moreover, different sampling heights may also affect the accuracy of monitoring data because of the diffusion and dispersion effects of the some compounds in wastewater. Fenet et al. found that carbamazepine, oxcarbazepine, and their main metabolites have different concentrations at different heights in the vicinity of its submarine outfall (Fenet et al. 2014). But the effect of different sampling heights on the concentrations of illicit drugs has not been reported yet. In our study, we used autosamplers to gather samples from 50 cm underwater. Therefore the diffusion and dispersion effects of the analytes can be eliminated when comparing the concentrations of the analytes in different WWTPs.

In addition, an online sensors for wastewater quality monitoring are a new methodology for characterizing illicit intrusions in a sanitary or combined sewer system. Compared to laboratory analysis, online sensors can get complete information that is completely representative of the wastewater pollutant dynamics (Banik et al. 2016). But online sensors are also affected by many factors, such as sensor location and the problems related to the mathematical modeling (Leopardi et al. 2020). Few online sensors are currently used for illicit drug monitoring in wastewater, but they are a promising monitoring tool for the future.

CONCLUSIONS

In the study, a method with a short runtime, small SPE loading volume, and high sensitivity was developed. Overall, the developed method can meet the requirements for current drug abuse monitoring in China. And this method was successfully applied to wastewater samples from 15 WWTPs, as a result of which, MOR, COD MAMP and K were detected. Determination of illicit drugs and their metabolites in wastewater is expected to be a significant tool for monitoring the consumption of illicit drugs.

ACKNOWLEDGEMENTS

This work was funded by the Ministry of Science and Technology of the People's Republic of China (81871531), Shanghai Science and Technology Commission (19DZ1200600), Shanghai Key Laboratory of Forensic Medicine (17DZ2273200), and Shanghai Forensic Service Platform (19DZ2292700).

DATA AVAILABILITY STATEMENT

All relevant data are included in this paper.

REFERENCES

Banik
B. K.
Di Cristo
C.
Leopardi
A.
de Marinis
G.
2016
Illicit intrusion characterization in sewer systems
.
Urban Water J.
14
(
4
),
416
426
.
Centazzo
N.
Frederick
B. M.
Jacox
A.
Cheng
S. Y.
Concheiro-Guisan
M.
2019
Wastewater analysis for nicotine, cocaine, amphetamines, opioids and cannabis in New York City
.
Forensic Sci. Res.
4
(
2
),
152
167
.
Chen
C.
Kostakis
C.
Irvine
R. J.
Felgate
P. D.
White
J. M.
2013
Evaluation of pre-analysis loss of dependent drugs in wastewater: stability and binding assessments
.
Drug Test. Anal.
5
(
8
),
716
721
.
Daughton
C. G.
2001
Illicit Drugs in Municipal Sewage: Proposed New Non-Intrusive Tool to Heighten Public Awareness of Societal Use of Illicit/Abused Drugs and Their Potential for Ecological Consequences
.
American Chemical Society
,
Washington
, pp.
348
364
.
Du
P.
Zhou
Z.
Bai
Y.
Xu
Z.
Gao
T.
Fu
X.
Li
X.
2017
Estimating heroin abuse in major Chinese cities through wastewater-based epidemiology
.
Sci. Total Environ.
605–606
,
158
165
.
Fenet
H.
Arpin-Pont
L.
Vanhoutte-Brunier
A.
Munaron
D.
Fiandrino
A.
Martinez Bueno
M. J.
Boillot
C.
Casellas
C.
Mathieu
O.
Gomez
E.
2014
Reducing PEC uncertainty in coastal zones: a case study on carbamazepine, oxcarbazepine and their metabolites
.
Environ. Int.
68
,
177
184
.
Khan
U.
van Nuijs
A. L.
Li
J.
Maho
W.
Du
P.
Li
K.
Hou
L.
Zhang
J.
Meng
X.
Li
X.
Covaci
A.
2014
Application of a sewage-based approach to assess the use of ten illicit drugs in four Chinese megacities
.
Sci. Total Environ.
487
,
710
721
.
Lai
F. Y.
O'Brien
J. W.
Thai
P. K.
Hall
W.
Chan
G.
Bruno
R.
Ort
C.
Prichard
J.
Carter
S.
Anuj
S.
Kirkbride
K. P.
Gartner
C.
Humphries
M.
Mueller
J. F.
2016
Cocaine, MDMA and methamphetamine residues in wastewater: consumption trends (2009–2015) in South East Queensland, Australia
.
Sci. Total Environ.
568
,
803
809
.
Li
J.
Hou
L.
Du
P.
Yang
J.
Li
K.
Xu
Z.
Wang
C.
Zhang
H.
Li
X.
2014
Estimation of amphetamine and methamphetamine uses in Beijing through sewage-based analysis
.
Sci. Total Environ.
490
,
724
732
.
Liu
H.-C.
Ho
H.-O.
Liu
R. H.
Lin
D.-L.
2009
Determination of codeine and morphine in brown mixture (opium preparations) in Taiwan using SPE and GC/MS
.
Forensic Sci. J.
8
(
1
),
1
12
.
Mercan
S.
Kuloglu
M.
Tekin
T.
Turkmen
Z.
Dogru
A. O.
Safran
A. N.
Acikkol
M.
Asicioglu
F.
2019
Wastewater-based monitoring of illicit drug consumption in Istanbul: preliminary results from two districts
.
Sci. Total Environ.
656
,
231
238
.
Shao
X. T.
Liu
Y. S.
Tan
D. Q.
Wang
Z.
Zheng
X. Y.
Wang
D. G.
2020
Methamphetamine use in typical Chinese cities evaluated by wastewater-based epidemiology
.
Environ. Sci. Pollut. Res. Int.
27
(
8
),
8157
8165
.
Thai
P. K.
Jiang
G.
Gernjak
W.
Yuan
Z.
Lai
F. Y.
Mueller
J. F.
2014
Effects of sewer conditions on the degradation of selected illicit drug residues in wastewater
.
Water Res.
48
,
538
547
.
Wang
Z.
Shao
X. T.
Tan
D. Q.
Yan
J. H.
Xiao
Y.
Zheng
Q. D.
Pei
W.
Wang
Z.
Wang
D. G.
2019
Reduction in methamphetamine consumption trends from 2015 to 2018 detected by wastewater-based epidemiology in Dalian, China
.
Drug Alcohol Depend.
194
,
302
309
.