Laboratory calibration of the polar organic chemical integrative sampler (POCIS) for passive sampling of pharmaceuticals in aquatic environments

(1)

Laboratory calibration of the polar organic chemical integrative

sampler (POCIS) for passive sampling of pharmaceuticals

in aquatic environments

Degree Project in Chemistry 30 credits, autumn 2010

By: Tu Tran Thanh Supervisor: Hanna Söderström Assistant supervisor: Jerker Fick

(2)

Summary

(4)

1. Introduction

Environmental pollutants have been monitored in many national and global projects in order to reduce both their acute and chronic effects on humans and ecosystems. One class of emergent environmental pollutants is pharmaceuticals which are released to the environment via mostly waste water treatment plants (WWTP) [1]. A great variety of these compounds have been detected in WWTP effluents. In United States antibiotics, hormones, personal care products, prescription and nonprescription drugs and steroids have been found in waste water

[2]

. Antimicrobials were detected in the WWTP effluents in Canada [3]. In Europe acidic drugs have been detected in both surface water and waste water [4], particularly human antibiotic substances were found at five Swedish sewage treatment plans [5]. However in order to assess and reduce their environmental contaminations it is necessary to sample a large volume of sample which may be representative for the whole system and to determine the time-weighted average (TWA) concentrations.

The most commonly utilized methods of sampling such as grab sampling give a snap-shot at the studied point of the system. The use of grab sampling can provide the approximation of TWA water concentration if multiple samples collected at different time points are combined. However it is time consuming and costly. Therefore there is a need to develop another type of sampler which is able to estimate TWA contaminant concentrations. Semi-permeable membrane device (SPMD) has been used to detect low level of hydrophobic pollutants with logKow greater than 2.5 such as polycyclic aromatic hydrocarbons (PAHs), polychlorinated

biphenyls (PCBs), polychlorinated dioxins and furans (PCDDs/PCDFs) over long period of time but it is not appropriate for hydrophilic contaminants (logKow < 2.5) [6]. In regard to

pharmaceuticals, Alvarez et. al. (2004) have developed a passive sampler called polar organic chemical integrative sampler (POCIS) which fulfills the requirement of TWA estimation for polar contaminants in water [7]. It consists of a sorbent, that is enclosed between two microporous membranes to avoid adsorption of large size particles, that facilitates linear uptake of hydrophilic contaminants up to 56 days (Figure 1) [7]. There are two different compositions of sorbent available depending on the target analytes, Oasis HLB (Poly(divinylbenzene)-co-N-vinylpyrrolidone) for sampling of pharmaceuticals, personal care products, prescription and non-prescription drugs; and triphasic admixture (80:20 [weight:weight] Isolute ENV+:Ambersorb 1500 dispersed on S-X3 Bio beads) for pesticides

(5)

sampling [7]. The membrane is made out of polyethersulfone (PES) which has been found to be the best for high uptake rates, minimal biofouling on the membrane surface and the suitability of long-term usage for integrative sampling [7]. In order to reduce the effect of field conditions such as water flow and biofouling; up to three (small sampling device) or six (large sampling device) POCISs can be mounted on one or two of carriers, respectively, and held inside a canister. The integrative sampling ability up to 56 days of POCIS (large sampling volume capacity) enhances the ability of detecting pharmaceuticals at trace levels and therefore enables the comprehensive screening of the system studied. However a drawback of POCIS is the demand for determination of laboratory calibration of sampling rates (Rs) for the

targets before field deployments. The sampling rate can be determined by either the losses of target compounds in laboratory calibration water (equation 1) or the laboratory uptakes by POCIS (equation 2) [8]. ) 1 ( * T V C C C R o o S − = * (2) T M C C R POCIS o POCIS S =

Where Rs is sampling rate (L/day), Co is initial concentration in water (µg/L), C is

concentration in water after POCIS exposure (µg/L), CPOCIS is concentration in POCIS (µg/g),

V is water volume (L), MPOCIS is sorbent mass (g) and T is exposure time (day).

One needs to keep in mind that the use of these two equations to calculate Rs may cause

overestimation in some cases due to the losses of target compounds not only by the uptakes of POCIS but also by degradation processes. Therefore it is important that sources to degradation are minimized and controlled during the experiments. In addition, laboratory calibrations cannot mimic exactly the field conditions and the field effects of environmental conditions are difficult to assess and adjust for. Therefore the calculations according to equation (1) and (2) just give approximate TWA water concentrations.

The development of performance reference compounds (PRC) has been studied to assess and adjust for the effects of environmental conditions in field. The idea is that the use of PRC will give a correction factor which compensates for environmental conditions in field and will therefore provide a better estimation of the TWA water concentrations [9]. To my knowledge, the Rs of POCIS for 68 pharmaceuticals have been calibrated under given conditions of

temperature, turbulence, salinity, sampling time and configuration (see Appendix 1).

(6)

There are four common calibration methods to quantify the amounts of analytes in a sample. These are external standard calibration, standard addition, internal standard and isotope dilution. In this project internal standard method was used. This method requires a proper internal standard which should have similarities to the analyte in molecular structure, retention time on the chromatographic column, losses in the sample treatment and instrument and should not be present in the samples. In term of pharmaceuticals analysis, group analogue internal standards have been used such as enrofloxacin as internal standard for other fluoroquinolones (norfloxacin, ciprofloxacin, ofloxacin); sulfamethazine for sulfamethoxazole; diaveridine for trimethoprim; cephalexin for cefadroxil, ampicillin, amoxicillin; 2-Me-5-Nitroimidazole for metronidazole; demeclocycline for doxycycline [12]. Some studies [10, 13, 14, 15] use deuterium, 13C or 15N labeled as internal standards due to their best similar structures which are expected to mimic ideally the analytes. In this method, a certain amount of the internal standard is added to all the standards as well as the unknown sample. Then the ratio of the detector responses of the analyte to the internal standard is plotted versus the amount of standard to allow the extrapolation of the analyte in the unknown samples.

The aims of this project were (1) to determine the Rs of 31 human pharmaceuticals and (2) to

study the effects of water temperature (5oC and 25oC), flow conditions (stirring and non-stirring) and the POCIS sampler design (with and without the use of canister and two different types of sorbent, pharmaceutical POCIS and pesticide POCIS). The pharmaceutical levels in the POCISs and in water samples were extracted and analysed by LC-MS/MS and quantified with the internal standard method.

2. Materials and methods

2.1. Chemicals and reagents

Pharmaceutical- and pesticide- POCISs with the sequestering diameter of 5.4 cm and 0.200±0.004 g of the sequestering media mass were used. These POCISs and two canisters that holds three POCISs in each were purchased and rented, respectively, from ExposMeterAB Sampling Technology, Sweden. The triphasic admixture (80:20 (weight:weight) Isolute ENV+:Ambersorb 1500 dispersed on S-X3 Bio beads were bought from ExposMeterAB Sampling Technology, Sweden.

(7)

later at -18oC. Low difference in degradation was however seen between storage at 2oC and -18oC, respectively.

Another set of stock solution was prepared for constructing two separate calibration curves. Eight calibration points ranging from 5 ng/ml to 2 µg/ml were used for the ten times diluted POCIS extracts and five other points from 0.4 ng/ml to 2 ng/ml were used for the quantification of the water samples. The density of methanol at 20oC (0.7914 g/mL) was used to correct for the pipette volume. These stock solutions and working calibration standards were stored in dark at -18oC.

2.2. Selection of target pharmaceuticals

This project focused on the human pharmaceuticals that were present at high levels in aquatic environment in Sweden, Asia and Africa. Other criteria of pharmaceutical selection were the potential to cause environmental effects and that their Rs data were not available yet,

particularly at 5oC. Some of the pharmaceuticals that are consumed a lot in Africa such as antimalarial drugs (amodiaquine, artesunate, lumefantrine), antibiotics (amoxicillin, flucloxicinllin, penicillin) are not included in this project since they are well known of being highly degradable [16]. Trimethoprim was chosen to serve as a control due to its stability and availability of Rs data [10, 14]. The complete list of 31 target pharmaceuticals can be found in

Appendix 2 together with their Chemical Abstract Service (CAS) numbers, Anatomical Therapeutic Chemical (ATC) codes, molecular structures, molecular weights and internal standards used during quantification.

2.3. Experimental setup

The RS calibration experiments were designed according to Figure 2 and performed by the

static renewal of tap water spiked with the target pharmaceuticals for 7, 14 and 21 days, respectively. The first set (A, B and C) was to determine the Rs of the target pharmaceuticals

(8)

taken after one and two days for non-stirring experiment and after one day for the stirring conditions, twice a week, to assess if decreases in the water concentrations of the pharmaceuticals due to POCIS adsorption as well as stability and adsorption onto the surfaces such as beaker walls and metal rings of POCISs occurred between renewals. At the end of 7, 14 and 21 days, respectively, the POCISs were collected. All POCISs and water samples were frozen and stored at -18oC until extraction and LC-MS/MS analysis.

The second set (D and E) was performed as the description above for set B and C with the differences that pesticide-POCISs were used. The purpose of these experiments was to study

5oC 25oC

7 days 14 days 21 days

Canister Stirring

Pharm-POCIS Every day

renewal

(B) (A) (C) Non-stirring Pharm-POCIS Every second day renewal

(E) Non-stirring Pest-POCIS Every second day renewal Canister

(D) 21 days 21 days F lo w a n d s a m p le r d e s ig n ( u s e o f c a n is te r a n d t y p e s o f s o rb e n t) e ff e c ts Temperature effect Blank controls Non-stirring Non-spiking 21 days 21 days (G) (H) Pharm-POCIS Pest-POCIS Pharm-POCIS Pest-POCIS

(9)

if the effect on Rs at different temperatures differs between pesticide- and

pharmaceutical-POCIS. These experiments were taken into account since Arditsoglou et. al. (2008) have shown that pesticide-POCIS might have higher uptake rates than pharmaceutical-POCIS for some pharmaceuticals such as estrone (E1), 17α-ethynylestradiol (EE2) and mestranol (MeEE2) [13].

Two different POCISs (one pharm- and one pest-) were exposed to non-fortified water at each studied temperature (G and H) to serve as POCIS blank controls. This was to compensate for the presence of the analytes in the POCISs and non-fortified water. The non-fortified water was renewed every week until the POCIS were collected after 21 days. Blank water samples were collected initially and at the end of each exposure period (day 7, day 14 and day 21).

2.4. Extraction procedure

The extraction and filtration of POCIS extracts were performed according to the standard operating procedures of the Environment Sampling Technology (EST) Lab (Figure 3). Briefly, a glass column (1 cm internal diameter and 30 cm length) for each POCIS was first fitted with a plug of glass wool (about 2 cm) and then rinsed with methanol. A 100-mL flask was inserted at the end of the column following by disassembling the sampler and carefully transferring the sorbent into the column with methanol. The membranes were rinsed to transfer all attached sorbent to the column. The 100 µL internal standard mixture (1 µ g/mL) was then added onto the sorbent. This high amount of internal standard was added to compensate for the dilution of the final extracts as described later. After the addition of the internal standard mixture a second layer of glass wool was added on top of the sorbent to prevent splashing of the sorbent when adding the extraction solvent. The pharmaceuticals were eluted by 2x20 mL of methanol for pharmaceutical POCISs and 50 mL of toluene:methanol:dichloromethane (1:1:8) for pesticide POCISs. The eluate was reduced in volume to about 2 mL by a rotary evaporator (BUCHI Rotavapor R-114) and then filtered through a glass fiber filter paper (0.45 µm) which was fitted inside a glass Pasteur pipette (150 mm length). The evaporated extract was rinsed many times with a total volume of 10 mL methanol following by adding the rinses to the filtered pipette. Afterward the filtered extract was evaporated by a gentle stream of air to approximate 200 µL for pharmaceutical POCIS extracts and dryness for pesticide pharmaceutical POCIS’s. Since the remaining toluene in the pesticide POCIS extracts might destroy the aqua gold column and therefore disturb the chromatograms in the LC-MS/MS analysis. The extracts were then reconstituted to the final volume of 1 mL water/methanol (75/25). Prior to the analysis, the final extracts were diluted 10 times with water/methanol (75/25) due to high concentrations of the target pharmaceuticals in the extracts.

(10)

Figure 3. Extraction and filtration procedures of pharmaceutical and pesticide POCISs

2.5. Evaluation of recovery efficiency

The evaluation of recovery on pharmaceutical-POCIS was carried out on solid phase extraction (SPE) cartridges containing Oasis HLB sorbent (200 mg). Triplicates of 50 mL tap water were fortified with 126 ng of the selected pharmaceuticals. This level corresponded to the estimated amount taken up by pharmaceutical-POCIS during a seven-day exposure (2 µg/L) and at the lowest Rs found in literature [7]. The Oasis HLB cartridges were first washed

(11)

the level of 126 ng of each analyte in methanol/water (25/75) to be able to calculate the extraction recovery. 100 µL of the internal standard mixture (100 ng/mL) was added to each sample prior to LC-MS/MS analysis to be able to calculate the extracted yields not influenced by instrumental matrix suppression or enhancement effects.

The recovery on pesticide-POCIS was evaluated by weighing triplicates of 200±0.1 mg of the triphasic admixture (80:20 (weight:weight) Isolute ENV+:Ambersorb 1500 dispersed on S-X3 Bio beads). Each triplicate of the sorbents was transferred to a glass column using methanol after fitting the column with about 2 cm of glass wool. These columns were then conditioned with 15 mL methanol, 15 mL methanol/water (50/50) and 15 mL water. After that 50 mL tap water fortified with the standard mixture containing 126 ng of each target pharmaceutical was loaded onto the sorbent. Extraction was then performed by adding 50 mL of toluene:methanol:dichloromethane (1:1:8) and the eluate was evaporated as described above for pesticide-POCIS extracts. The final extract was reconstituted using 150 µ L methanol, 100 µ L internal standard mixture (100 ng/mL) and 750 mL water prior to the LC-MS/MS analysis without further dilutions.

2.6. Selection of appropriate internal standards

A mixture of 14 available internal standards was prepared at the level of 100 ng/mL. The non-fortified tap water sample, standard mixture and triplicates of non-fortified tap water were treated as described in the recovery efficiency test for pharmaceutical POCIS but with the addition that 100 µL of the internal standard mixture was added to the samples prior to extraction. The suitable internal standard for each analyte was chosen when the ratio of analyte/internal standard of the extracted sample to that of the standard was closest to 100% ± 20% and there was similarity in the retention times of the analyes and the internal standards (± 3 min). The list of chosen internal standards can be found in Appendix 2.

2.7. Pharmaceutical analysis

Two individual MS/MS methods were developed for paracetamol and sulfadoxine, respectively, by introducing each tuning solution (1 µg/mL) to the mass spectrometer by direct infusion technique using the syringe of 500 µL at the flow rate of 10 µL/min with the aid of the mobile phase consisting of 50% water (0.1% formic acid), 20% acetonitril (0.1% formic acid) and 30% methanol (0.1% formic acid). The other pharmaceuticals in the list have been tuned previously. Furthermore, a LC-MS/MS method including a mixture of 31 target pharmaceuticals was developed by injecting 10 µL (for 75/25 water/methanol extracts) or 100 µ L (for 100% water samples) of 52.83 ng/mL to 196.23 ng/mL to optimize the chromatographic separation efficiency and MS/MS sensitivity for POCIS and water samples, respectively.

The LC-MS/MS system included a loop of 10 µ L (for POCIS samples) or 100 µ L (for water samples), an Autosampler Finnigan Thermo PALTM, a Surveyor LC Pump Plus, a Hypersil Gold aQ C18-column (50x2.1 mm i.d., particle size 5 µm ThermoScientific) with the mobile

(12)

(0.1% formic acid) and a MS/MS TSQ Quantum Ultra (ThermoScientific). Detail about the gradient and flow rate can be found in Table 1.

Table 1. Gradient of the mobile phase for pharmaceutical analyses

Time (min) % water % acetonitril % methanol Flow rate (µL/min)

0 100 0 0 250 1 100 0 0 250 8 20 20 60 300 11 0 40 60 350 13 0 100 0 400 13.6 0 100 0 400 13.61 100 0 0 250 17.6 100 0 0 250

3. Results and discussion

3.1. Method development

The three most intense transitions of paracetamol and sulfadoxine were observed (see Appendix 3). The transitions from 152.0 to 110.2 and 311.0 to 156.2 for paracetamol and sulfadoxine, respectively, were found to be the most intense transitions that were used for quantitative purpose. These transitions could be explained by the loss of C(OH)-CH3 group

according to McLafertty rearrangement in case of paracetamol and the cleavage at S-N bond with respect to sulfadoxine (Figure 4).

O H NH O CH₃ H N H₂ S O O NH N N O O CH₃ CH₃

Figure 4. Proposed fragmentation pathways of the most paracetamol (left) and sulfadoxine (right) transitions

(13)

with 5 µm particle size and an electrospray coupled to an ion trap mass spectrometer (ESI-MS) [12]. However this system was not used in this project due to the lack of time.

3.2. Evaluation of recovery efficiency

Methanol and the mixture of toluene:methanol:dichloromethane (1:1:8) have been found to be the suitable solvents for the extraction of a wide range of pharmaceuticals and pesticides in pharmaceutical-POCIS and pesticide-POCIS, respectively [7]. These solvents were used in this work without any further optimizations. The analyses of the effluents from the SPE columns and the glass columns when the fortified water samples were loaded did not result in any detected peaks. This ensured the binding efficiency of the sorbents to the target pharmaceuticals. The recoveries of the target pharmaceuticals on pharm- and pest-POCIS were shown in Table 2. In general the recoveries of target pharmaceuticals from pharm-POCIS were found to be higher than those of pest-pharm-POCIS. Some compounds; beclomethasone, diclofenac, oxazepam, paracetamol, sulfadoxine; were recovered at an acceptable level for both pharm- and pest-POCIS. Some of the others showed low recoveries such as duloxetine, ketoconazole, miconazole, perphenazine, tamoxifen. Another research has used the mixture of methanol and acetyl acetate for the extraction on pest-POCIS [18]. Therefore the optimization to find the best extraction solvent (mixture) should be performed to yield a better recovery on pest-POCIS.

Table 2. The recovery of target pharmaceuticals on pharm- and pest-POCIS

Pharmaceutical Pharm-POCIS Pest-POCIS Pharmaceutical Pharm-POCIS Pest-POCIS Recovery RSD (%) Recovery RSD (%) Recovery RSD (%) Recovery RSD (%) Atorvastatin 12.4 45.2 95.0 66.7 Levonorgestrel 62.6 16.1 37.7 5.4 Beclomethasone 66.2 16.8 79.2 7.1 Miconazole 32.1 7.9 13.4 7.1

Bezafibrate 123.7 18.6 25.0 3.6 Norfloxacin n.a. n.a. n.a. n.a.

Bupropion 74.7 9.8 21.4 3.2 Oxazepam 95.4 8.0 62.5 10.9

Ciprofloxacin n.a. n.a. n.a. n.a. Oxytetracycline 24.6 38.0 1.2 106.7

Codeine 71.7 6.5 34.6 10.8 Paracetamol 117.6 5.9 46.2 8.5

Diclofenac 149.3 6.2 40.6 2.2 Perphenazine 1.8 1.0 18.7 11.7

Doxycycline n.a. n.a. n.a. n.a. Promethazine 19.6 11.5 17.5 14.3

Duloxetine 2.9 0.1 12.0 17.9 Risperidone 57.2 4.1 32.2 12.2

Enrofloxacin n.a. n.a. n.a. n.a. Sulfadoxine 81.1 2.5 45.8 4.6

(14)

3.3. Calculation of water concentrations and sampling rates

None of the target pharmaceuticals was found to be present in the non-fortified tap water and blank POCIS extracts. Therefore no blank corrections were made for the calculation of water and POCIS extract concentrations. The concentrations of pharmaceuticals in water were calculated as the integral of the average concentrations from day 0 to day 1 and from day 0 to day 2 following by dividing by 1 and 2 for stirring and non-stirring conditions, respectively. The deviation within each day of water sample collection (d0, d1, d2) was less than 20% with at least 5 points of replicates. Most of the pharmaceuticals were removed due to the uptake of POCIS and other adsorbing phenomena, with less than 10% for non-stirring after 2 days of exposure and 20% for stirring condition after 1 day. Degradation of sulfadoxine was shown as an example in Figure 5. This indicated good maintenance of water concentrations to ensure the integrative uptakes of POCIS.

Figure 5. Degradation of Sulfadoxine in the calibration experiment at 25oC

The accumulation of the amount of each pharmaceutical in POCIS over 7, 14, 21 days of exposure under the specified conditions (equation 2) was used to calculate the Rs of the target

pharmaceuticals. The Rs were given by the slopes of the linear curves which were generated by plotting the concentration factor of each POCIS (Cf/POCIS) which corresponded to ratio

between the amount of each pharmaceutical detected in POCIS extract (CPOCIS*MPOCIS) and

the corresponding water concentration (Co) versus the exposure time (Figure 6). By doing so,

the Rs of totally 19 of the 31 target pharmaceuticals were determined with the correlation

coefficients (R2 values) higher than 0.909 for stirring condition at 25oC; 0.874 and 0.824 for non-stirring condition at 25oC for pharm- and pest-POCIS, respectively; 0.822 and 0.856 for non-stirring condition at 5oC for pharm- and pest-POCIS, respectively. In addition to the linear uptake of most of the pharmaceuticals, the uptakes of some others by POCIS were found to be binomial such as oxazepam, diclofenac, tramadol, haloperidol and verapamil (Figure 6). A summary of the determined Rs was given in Table 3. The Rs of trimethorpim

were found to be 0.385 L/day and 0.078 L/day at 25oC for stirring and non-stirring conditions, respectively, which were consistent with the finding of Sherri et. al. (0.360 L/day for stirring condition and 0.090 L/day for non-stirring condition at 25oC) [9]. Erythromycin has been shown to be unstable in aqueous solution which could explain its undetected level in water samples and POCIS extracts, consequently its Rs was not able to be determined [17].

(15)

Figure 6. Linear uptake (at 25oC, stirring condition, pharm-POCIS) and binomial uptake (at 25oC, non-stirring condition, pest-POCIS) of pharmaceuticals in POCIS over time

3.4. Effect of the POCIS sorbent: pharmaceutical and pesticide configuration

Generally the sampling rates of 15 pharmaceuticals by pharm-POCIS were determined at non-stirring condition whereas 14 compounds were detected in case of pest-POCIS. The effect of two different types of sorbent of POCIS was noticed by comparing the sampling rates at non-stirring condition of set B and D at 5oC and set C and E at 25oC (Figure 7). The result showed that the Rs of the five compounds paracetamol, trimethorpim, haloperidol, verapamil and

promethazine increased while the other seven compounds; codeine, sulfadoxine, tramadol, risperidone, bupropion, beclomethasone and bezafibrate; decreased when it came to the use of pest-POCIS instead of pharm-POCIS at 5oC. In the other hand, the Rs of the six compounds

(paracetamol, codeine, trimethoprim, sulfadoxine, risperidone and promethazine) increased at 25oC when using pest-POCIS compared to pharm-POCIS whereas one compounds decreased (bezafibrate). The largest increase was found to be from 0.045 L/day to 0.168 L/day for verapamil at 5oC and from 0.009 L/day to 0.101 L/day for paracetamol at 25oC. This suggested the use of pest-POCIS as a better option for some pharmaceuticals, particularly verapamil and paracetamol, since higher Rs can make the detection at lower water

(16)

(17)

POCIS so the increase in Rs could be even higher if the recovery is improved. However the

drawback of low extraction recovery should be taken into account when choosing pest-POCIS because it might limit the number of pharmaceuticals to be detected.

Figure 7. The effect of POCIS sorbent at 5oC and 25oC

Another interesting phenomenon was observed for diclofenac. The uptake of diclofenac by pharm-POCIS was found to be binomial at both 5oC and 25oC but it became linearly in case of pest-POCIS with the correlation coefficient of 0.975 at 5oC and 0.931 at 25oC (see Table 3). This finding might play an important role when it came to the study of diclofenac using POCIS in particular.

3.5. Effect of temperature on sampling rate of pharm- and pest-POCIS

The effect of temperature on sampling rate of POCIS was generally observed to be positive for most of the compounds ranging from 4.6% (codeine) to 300% (verapamil) for pharm-

(18)

Figure 8. The effect of temperature on sampling rate

POCIS and from 52% (bupropion) to 516% (diclofenac) for pest-POCIS except for the observation that the uptake of paracetamol was found to be negatively affected by temperature on pharm-POCIS (-45%) but positively on pest-POCIS (123%) (Figure 8). One could be seen that the sampling rates by pest-POCIS were positively affected by temperature with larger extent which might suggest the limit in capacity of pharm-POCIS compared to pest-POCIS. However it can be better to choose pharm-POCIS when sampling pharmaceuticals in water that had high variation in temperature since pharm-POCIS was less affected by temperature. Another observation was noticed for the effect on the uptake modes of tramadol and verapamil by pest-POCIS. This was seen as linear uptake at 5oC but binomial at 25oC (see Table 3).

3.6. Effect of the flow conditions and the use of canister

The effect of the flow on pharm-POCIS was studied by changing from non-stirring (set C) to stirring conditions (set A) at 25oC and adding canister which held three POCIS to each set.

(19)

Figure 9. Effect of the flow with and without canister

It should be noted that since the two canisters were exposed for 21 days, hence the effect could only be studied for the pharmaceuticals that showed linear uptake in POCIS for 21 days. The calculation was made by taking the average of the concentration factors of pharmaceuticals in the triplicates of POCIS per day. The results showed that the sampling rate was significantly affected by the flow without the use of canister with a positive factor ranging from 109% for sulfadoxine up to 1756% for verapamil (Figure 9). In the situation of canister usage, it was found that the canister reduced the effect of flow by a factor of 16% (sulfadoxin) to 64% (verapamil). No differences in sampling rates were statistically observed when using a canister at non-stirring condition according to paired t test (0.95). The reduced effect of using a canister played a major role when it came to field deployments where the field conditions varied a lot such as water flow and biofouling.

4. Conclusion and future remarks

The project was successful to laboratory determine the Rs of the target pharmaceuticals under

specific conditions. The Rs of totally 19 of 31 target pharmaceuticals were determined mostly

ranging from 0.2 L/day to 1 L/day for stirring condition at 25oC, from 0.03 L/day to 0.1 L/day for non-stirring condition at 5oC and from 0.05 L/day to 0.2 L/day for non-stirring condition at 25oC. There was a clear trend in increasing Rs from low (5oC) to high temperature (25oC). Pest-POCIS generally showed higher uptake rate compared to pharm-POCIS especially at 25oC. However the test showed low recovery in general, not only for pest-POCIS, may require an improvement. The behavior of binomial uptake should also be noted when applying POCIS for pharmaceutical sampling. The effect of water flow has been found to be positive but significantly reduced by the use of canister.

This work showed that the use of POCIS is a suitable sampling device for pharmaceuticals. However there are still numerous of things that need to overcome to bring POCIS from

(20)

qualitative to quantitative. One of those is the great number of variable environmental factors such as water flow, temperature, pH and biofouling that need to be minimized or reduced. Therefore the use of PRC for correcting of field conditions should be further studied to enhance the ability of quantitative determination of water concentration.

The Rs of some pharmaceuticals were not calibrated in this project due to their undetectable

levels in water samples and POCIS extracts (miconazole, oxytetracycline, tetracycline). One could re-measure them with a lower dilution factor to be able to detect them. It should be noted that the others that were successfully determined should be taken away from the list of MS-analyzed compounds in order to gain sensitivity and avoid overloading the detector. Another part of this work that could be considered in future work is to develop a method for the chromatographic separation and MS-analysis of the compounds (enrofloxacin, norfloxacin, ciprofloxacin and doxycycline) that were not detected by the developed LC-MS/MS method.

Many studies have been conducted to figure out the relationship between physical and chemical properties of pharmaceuticals to their Rs. To the best of my knowledge, none of

those studies have shown a good correlation and thus it is difficult to predict the sampling rate of an unknown pharmaceutical. The reason could be due to the studies only consider one property of the pharmaceutical such as logKow, retention time and pKa at a time. One possible

option is the consideration of a group of properties and followed by the use of multivariate analysis such as partial least square (PLS) or orthogonal partial least square (OPLS).

Acknowledgement

(21)

References

1. Jones O. A. H., Voulvoulis N., and Lester J. N., Critical Reviews in Environmental Science and Technology 35 (2005) 401.

2. Ternes T.A., Trends in Analytical Chemistry 20 (2001) 419.

3. Miao X. S., Bishay F., Chen M., Metcalfe C. D., Environmental Science and Technology 38 (2004) 3533.

4. Quintana J. B., Reemtsma T., Rapid Commun Mass Spectrom 18 (2004) 765.

5. Lindberg R., Wennberg P., Johansson M., Tysklind M. and Andersson B., Environmental Science and Technology 39 (2005) 3421.

6. Johnson A., Environmental Assessment Program: Standard Operating Procedure for using Semipermeable Membrane Devices to Monitor Hydrophobic Organic Compounds in Surface Water (2007) version 2.0.

7. Alvarez D. A., Petty J. D., Huckins J. N., Jones-Lepp T. L., Getting D. T., Goddard J. P., and Manahan S. E., Environmental Toxicology and Chemistry 23 (2004) 1640. 8. Togola A., Budzinski H., Analytical Chemistry 79 (2007) 6734.

9. Mazzella N., Lissalde S., Moreira S., Delmas F., Mazellier P., and Huckins J. N., Environmental Science Technology 44 (2010) 1713.

10.Li H., Helm P. A., and Metcalfe C. D., Environmental Toxicology and Chemistry 29 (2010) 751.

11.Sherri L. M., Evelyn L. M., and Charles S. W., Environmental Toxicology and Chemistry 26 (2007) 2517.

12.Lindberg R., Jarnheimer P., Olsen B., Johansson M., Tysklind M., Chemosphere 57 (2004) 1479.

13.Arditsoglou A., Voutsa D., Environmetal Pollution 156 (2008) 316.

14.Bueno M. J. M., Hernando M. D., Aguera A., Fernandez-Alba A. R., Talanta 77 (2009) 1518.

15.Harman C., Boyum O., Tollefsen K. E., Thomas K., and Grung M., Journal of Environmental Monitoring 10 (2007) 239.

16.Germer J., Sinar E., Journal of Agriculture and Rural Development in the Tropics and Subtropics 111 (2010) 41.

17.Zhang Z., Hibberd A., Zhou J. L., Analytica Chimica Acta 607 (2008) 37.

18.Pesce S., Morin S., Lissalde S., Montuelle B. and Mazzella N., Environmental Pollution 159 (2011) 735.

19.Brisaert M., Heylen M. and Plaizier-Vercammen J., Pharmacy World and Science 18 (1996) 182.

(22)

Appendices

Appendix 1. List of available Rs of pharmaceuticals by POCIS [20]

Pharmaceutical Quiscent (Q) Rs Flowing (F) Rs Exp. Q/Fa Temp. Q/Fb Ref.

(L day-1) (L day-1) (days) (°C)

(23)

Erythromycin 0.183 0.911 29 / 25 22 / 28 [11] Erythromycin n.t. 0.163 1,3,7 21 [14] 17α-Estradiol (α E2) n.t. 0.1451j/0.1216k 7,14,28 23.5 [13] 17β-Estradiol n.t. 0.037 10 15 [17] 17β-Estradiol 0.334 0.580g/0.596h/0.693i 8 5,15,25/25 [10] 17β-Estradiol (βE2) n.t. 0.1144j/0.1145k 7,14,28 23.5 [13]

Estrogens estriol (E3) n.t. 0.1305j/0.1571k 7,14,28 23.5 [13]

(24)

(25)

Sulfamethazine 0.049 0.114 29 / 25 22 / 28 [11] Sulfamethoxazole n.d. n.d. 29 / 25 22 / 28 [11] Sulfamethoxazole 0.202 0.291g/0.348h/0.339i 8 5,15,25/25 [10] Sulfamethoxazole n.t. n.d. 1,3,7 21 [14] Sulfapyridine 0.041 0.051 29 / 25 22 / 28 [11] Sulfapyridine 0.201 0.213g/0.411h/0.436i 8 5,15,25/25 [10] Sulfisoxazole n.d. 0.536 29 / 25 22 / 28 [11] Temazepam 0.128 0.421 29 / 25 22 / 28 [11] Tetracycline n.t. 0.071 1,3,7 21 [14] Triclosan 0.753 1.006g/1.442h/1.929i 8 5,15,25/25 [10] Triclosan n.t. n.d. 28 18 [15] Trimethoprim 0.090 0.360 29 / 25 22 / 28 [11] Trimethoprim 0.215 0.267g/0.319h/0.462i 8 5,15,25/25 [10] Trimethoprim n.t. n.d. 1,3,7 21 [14] Venlafaxine 0.104 0.167g/0.388h/0.521i 8 5,15,25/25 [10] a_{Days of exposure.}

b_{Temperature during experiments.}

c_{No data available.}

d_{Not tested.}

e_{L day}-1_g-1_{sorbent, at 15 °C, lower or equal to given value.}

f

L day-1g-1 sorbent, at 21 °C, lower or equal to given value.

(26)

Appendix 2. 31 target pharmaceuticals (*) Pharmaceutical CAS number Category (ATC code) Molecular structure (Molecular weight) (MW g/mol)

Atorvastatin IS: Amytriptiline D6 134523-00-5 HMG CoA reductase inhibitors (C10AA05) NH N O C H3 CH3 F OH O OH OH (MW = 558.64) Beclometasone IS: Carbamazepine D10 4419-39-0 - Corticosteroids acting locally (A07EA07) - Corticosteroids, potent

(group III) (D07AC15)

- Corticosteroids (R01AD01) - Glucocorticoids (R03BA01) O CH3 Cl OH CH3 CH3 OH OH O (MW = 521.04) Bezafibrate IS: Oxazepam D5 41859-67-0 Fibrates (C10AB02) NH O O CH3 C H₃ O OH ( MW = 361.82) Bupropion IS: Risperidone D4 34841-39-9 Other antidepressants (N06AX12) Cl CH3 N H CH3 C H3 CH3 O (MW = 239.74) Ciprofloxacin IS: n.a. 85721-33-1 - Fluoroquinolones (J01MA02, S01AX13) - Anti-infectives (S02AA15, S03AA07) N N H N F O O OH (MW = 331.35) Codeine IS: Sulfamethoxazole 13C6 76-57-3

(27)

Diclofenac

IS: Oxazepam D5 15307-86-5

- Other dermatologicals

(D11AX18)

- Acetic acid derivatives

and related substances (M01AB05)

- Anti-inflammatory

preparations, non-steroids for topical use

(M02AA15) - Anti-inflammatory agents, non-steroids (S01BC03) Cl Cl NH OH O (MW = 296.15) Doxycycline IS: n.a. 564-25-0 - Tetracycline (J01AA02) - Anti-infectives and

(28)

Furosemide IS: Risperidone D4 54-31-9 Sulfonamides, plain (C03CA01) O NH O O H Cl S O O NH2 (MW = 330.74) Haloperidol IS: Amytriptiline D6 52-86-8 Butyrophenone derivatives (N05AD01) F O N OH Cl (MW = 375.9) Irbesartan IS: Risperidone D4 138402-11-6 Angiotensin II antagonists, plain (C09CA04) N N O CH3 N N N N H (MW = 428.53) Ketoconazole IS: Risperidone D4 65277-42-1 Imidazole derivatives (J02AB02, D01AC08, G01AF11) O O N N Cl Cl H N N O C H₃ O (MW = 531.43) Levonorgestrel IS: Risperidone D4 797-63-7 Progestogens (G03AC03) O H H H H OH CH₃ CH (MW = 312.45) Miconazole IS: Risperidone D4 22916-47-8

- Imidazole and triazole

derivatives (D01AC02)

- Anti-infectives and

(29)

Oxytetracycline IS: Risperidone D4 79-57-2 - Tetracycline and derivatives (D06AA03, J01AA06, QJ51AA06) - Antibiotics (G01AA07, S01AA04, QG51AA01) O N H₂ O H N CH₃ C H3 O OH H OH O OH OH O H H (MW = 460.434) Paracetamol IS: Sulfamethoxazole 13C6 103-90-2 Anilides (N02BE01) _H_O NH O CH₃ (MW = 151.17) Perphenazine IS: Risperidone D4 58-39-9 Phenothiazines with piperazine structure (N05AB03) N N N S O H Cl (MW = 403.97) Promethazine IS: Promethazine D7 60-87-7 58-33-3 (HCl)

- Antihistamines for topical

use (D04AA10) - Phenothiazine derivatives (R06AD02) S N C H3 N CH3 CH3 (MW = 284.42) Risperidone IS: Risperidone D4 106266-06-2 Other antipsychotics (N05AX08) N O F N N N O C H3 (MW = 410.485) Sulfadoxine IS: Sulfamethoxole 13C6 2447-57-6 Sulfonamides (QJ01EJ13) N H2 S O O NH N N O O CH3 CH3 (MW = 310.33) Tamoxifen IS: Tamoxifen 13C2 10540-29-1 Anti-estrogens (L02BA01) CH3 O N C H3 CH3 (MW = 371.52) Tetracycline 60-54-8 64-75-5 (HCl) - Anti-infectives and

(30)

QG51AA02)

Tramadol 27203-92-5 Other opioids

(N02AX02) O H O CH3 N CH3 C H3 H (MW = 263.4) Trimethoprim 738-70-5 Trimethoprim and derivatives (J01EA01, QJ51EA01) N N O O O CH3 CH3 CH3 NH2 N H2 (MW = 290.32) Verapamil 52-53-9 Phenylalkylamine derivatives (C08DA01) O O CH3 C H3 N CH3 C H3 CH3 N O O CH3 CH3 (MW = 454.60) * _{Source: http://en.wikipedia.org}

Appendix 3. Tuning parameters of target compounds

Pharmaceutical Mode Parent

(m/z) Daughter (m/z) Collision energy (V) TubeLens offset (V) Type

Atorvastatin ESI+ 559.2 440.4 20.0 120.0 Quantitative

ESI+ 559.2 250.0 43.0 120.0 Qualitative

Beclomethasone ESI- 453.2 377.3 19.0 92.0 Quantitative

ESI- 453.2 297.2 27.0 92.0 Qualitative

Bezafibrate ESI- 360.0 274.1 21.0 101.0 Quantitative

ESI- 360.0 154.1 30.0 101.0 Qualitative

Bupropion ESI+ 240.0 184.1 12.0 77.0 Quantitative

ESI+ 240.0 131.2 25.0 77.0 Qualitative

Ciprofloxacin ESI+ 332.0 231.1 35.0 117.0 Quantitative

ESI+ 332.0 288.2 16.0 117.0 Qualitative

Ciprofloxacin* _ESI+ _336.0 _318.0 _20.0 _106.0 _Quantitative

Codeine ESI+ 300.1 215.1 23.0 102.0 Quantitative

ESI+ 300.1 165.2 41.0 102.0 Qualitative

Diclofenac ESI- 294.0 250.0 15.0 96.0 Quantitative

Doxycycline ESI+ 445.0 428.3 17.0 107.0 Quantitative

ESI+ 445.0 200.9 27.0 107.0 Qualitative

Duloxetine ESI+ 298.1 44.3 12.0 74.0 Quantitative

ESI+ 298.1 123.5 50.0 74.0 Qualitative

Enrofloxacin

ESI+ 360.1 316.2 17.0 112.0 Quantitative

ESI+ 360.1 245.1 25.0 112.0 Qualitative

(31)

Erythromycine ESI+ 734.3 576.6 19.0 154.0 Quantitative

ESI+ 734.3 558.5 18.0 154.0 Qualitative

Erythromycine* ESI+ 736.4 578.6 17.0 129.0 Quantitative

ESI+ 736.4 160.1 29.0 129.0 Qualitative

Felodipine ESI+ 384.0 338.1 5.0 84.0 Quantitative

ESI+ 384.0 352.1 10.0 84.0 Qualitative

Furosemide ESI- 328.9 285.0 18.0 81.0 Quantitative

ESI- 328.9 205.0 23.0 81.0 Qualitative

Haloperidol ESI+ 376.0 165.1 22.0 88.0 Quantitative

ESI+ 376.0 123.1 36.0 88.0 Qualitative

Ibersartan ESI+ 429.2 207.1 22.0 110.0 Quantitative

ESI+ 429.2 180.1 38.0 110.0 Qualitative

Ketoconazole ESI+ 531.1 254.9 34.0 134.0 Quantitative

ESI+ 531.1 244.0 33.0 134.0 Qualitative

Levonogestrel ESI+ 313.0 245.2 17.0 103.0 Quantitative

ESI+ 313.0 109.2 29.0 103.0 Qualitative

Miconazole ESI+ 416.9 159.1 29.0 114.0 Quantitative

ESI+ 416.9 161.0 31.0 114.0 Qualitative

Norfloxacin ESI+ 320.0 302.1 20.0 114.0 Quantitative

ESI+ 320.0 233.1 23.0 114.0 Qualitative

Oxazepam ESI+ 287.0 241.1 22.0 83.0 Quantitative

ESI+ 287.0 269.1 15.0 84.0 Qualitative

Oxazepam* _ESI+ _292.0 _246.1 _22.0 _Quantitative

Oxytetracycline ESI+ 461.0 426.2 18.0 130.0 Quantitative

ESI+ 461.0 443.2 13.0 130.0 Qualitative

Paracetamol

ESI+ 152.0 93.3 10.0 81.0 Qualitative

ESI+ 152.0 65.4 32.0 81.0 Qualitative

Perphenazine ESI+ 404.1 171.2 21.0 113.0 Quantitative

ESI+ 404.1 143.2 27.0 113.0 Qualitative

Promethazine ESI+ 285.1 86.3 16.0 65.0 Quantitative

ESI+ 285.1 198.0 26.0 65.0 Qualitative

Promethazine* _ESI+ _292.1 _89.3 _16.0 _Quantitative

Risperidone ESI+ 411.1 191.1 27.0 94.0 Quantitative

ESI+ 411.1 110.2 44.0 94.0 Qualitative

Risperidone* ESI+ 415.1 195.1 27.0 Quantitative

Sulfadoxine

ESI+ 311.0 156.2 18 90 Quantitative

ESI+ 311.0 108.2 24 90 Qualitative

ESI+ 311.0 92.2 26 90 Qualitative

Tamoxifen ESI+ 372.2 72.4 22.0 113.0 Quantitative

ESI+ 372.2 129.1 26.0 113.0 Qualitative

Tamoxifen* _ESI+ _375.2 _75.2 _22.0 _110.0 _Quantitative

Tetracycline ESI+ 445.1 427.4 12.0 109.0 Quantitative

(32)

ESI+ 445.1 154.1 27.0 109.0 Qualitative

Tramadol ESI+ 264.1 58.4 16.0 82.0 Quantitative

ESI+ 264.1 246.2 10.0 82.0 Qualitative

Tramadol* ESI+ 268.1 58.4 16.0 Quantitative

Trimethoprim ESI+ 291.0 230.1 23.0 106.0 Quantitative

ESI+ 291.0 123.2 25.0 106.0 Qualitative

Trimethoprim* ESI+ 294.1 233.2 22.0 101.0 Quantitative

ESI+ 294.1 126.2 24.0 101.0 Qualitative

Verapamil ESI+ 455.2 165.1 28.0 118.0 Quantitative

ESI+ 455.2 303.3 23.0 118.0 Qualitative

Laboratory calibration of the polar organic chemical integrative sampler (POCIS) for passive sampling of pharmaceuticals in aquatic environments