Quantification of EPA 1694 Pharmaceuticals and Personal Care Products in Water at the ng/L Level Utilizing Online Sample Preparation with LC-MS/MSKevin McHale, Mark Sanders; Thermo Fisher Scientific, Somerset, NJ

IntroductionThere is growing environmental concern regarding the health impact of trace levels of pharmaceuticals and personal care products (PPCPs) in water resources. In response to this concern, the U.S. Environmental Protec-tion Agency (EPA) recently published Method 1694, which determines dozens of PPCPs in water, soil, sediment, and biosolids by high performance liquid chromatography combined with tandem mass spectrometry (HPLC-MS/MS).1 The method, which is yet to be promulgated, uses solid phase extraction (SPE) of water samples followed by HPLC-MS/MS analysis using a single transition for each compound to achieve low nanogram/liter (ng/L) limits of quantitation (LOQs).

The target analytes in the EPA method are divided into four groups, with each group representing one HPLC-MS/MS run. Three of the groups are extracted under acidic conditions; the fourth is extracted under basic conditions. These SPE methods can use up to 1 L of sample. Although not sample limited, the storage of large bottles of water requires a great deal of refrigeration space. In addition, manual SPE of 1 L of sample requires several hours of preparation.

One of the opportunities in the analysis of PPCPs in water is to reduce the time required for sample prepara-tion and analysis while maintaining the required sensitivity at the ng/L level and the selectivity to positively identify the analyte of interest. We describe a method for online sample prepara-tion and analysis using the Thermo Scientific EQuan system. This method couples a fast HPLC system with two LC columns – one for pre-concentra-tion of the sample, the second for the analytical analysis – and an LC-MS/MS instrument. Instead of processing 1 L of water by the manual, time-consuming process of SPE described in EPA Method 1694, this alternative approach incorporates online sample preparation in series with LC-MS/MS using smaller volumes of water (0.5-20 mL) to achieve ng/L quantita-tion limits.

GoalTo demonstrate a progressive approach to analyzing PPCPs in environmental sources of water at the ng/L level with online sample preparation using small volumes of water, thus saving time and reducing the cost of analysis.

Experimental ConditionsThe EQuan LC-MS/MS experimental setup is illustrated in Figure 1.

Sample Preparation Aqueous solutions containing 5% – 20% acetonitrile (ACN) and adjusted to pH 2.9, 6.6 or 11.3 were spiked with more than 60 PPCPs at the low ng/L level.

HPLCWater samples of 0.5 mL were directly injected onto a Thermo Scientific Hypersil GOLD aQ pre-concentration trapping column (2.1 x 20 mm, 12 µm) at 1.5 mL/min with H2O + 0.2% formic acid. After sufficient washing of the pre-concentration column, the target compounds were transferred to the Thermo Scientific Betasil C18 analyti-cal column (2.1 x 100 mm, 3 µm) for chromatographic separation by gradient elution prior to introduction into the mass spectrometer.

Application Note: 508

Key Words

• EQuan System

• TSQ Vantage

• PPCPs

• Water Analysis

Autosampler

3 µm Betasil 18Analytical Column

12 µm Hypersil GOLD aQPreconcentration

Column

LargeVolumeSampleLoop

6-PortValve

6-PortValve

Accela 600Pump

Accela 1250Pump

ConventionalVolume Sample Loop

6-PortValve

TSQ Vantage Triple Stage Quadrupole Mass Spectrometer

Figure 1. The EQuan pre-concentration LC-MS/MS experimental setup.

MSMS analysis was carried out on a Thermo Scientific TSQ Vantage triple stage quadrupole mass spectrometer. Two selected reaction monitoring (SRM) transitions per com-pound were acquired: one for quantitation and the other for positive confirmation. To maximize the performance of the triple stage quadrupole, time-specific SRM “win-dows” were employed at the retention times of the target compounds.

Results and DiscussionThe current EPA Method 1694 describes three different LC methods for PPCPs from Groups 1, 2, and 4, which are amenable to positive electrospray ionization (ESI) MS/MS. To simplify the method and reduce the total analysis time, a single 10-minute LC-MS/MS method was developed, which included compounds from additional pharmaceuti-cal classes not included in EPA Method 1694, such as beta-blockers. In total, 67 compounds were analyzed by positive ESI-MS/MS (Table 1). Of these, 54 were from EPA Method 1694 Groups 1, 2, and 4.

Table 1. PPCPs analyzed

Compound Class Compound Class

Trimethoprim Antibiotic 4-epi-Chlorotetracycline Antibiotic, tetracycline

Cefotaxime Antibiotic, cephalosporin Demeclocycline Antibiotic, tetracycline

Norfloxacin Antibiotic, fluoroquinolone Chlorotetracycline Antibiotic, tetracycline

Ofloxacin Antibiotic, fluoroquinolone Doxycycline Antibiotic, tetracycline

Ciprofloxacin Antibiotic, fluoroquinolone Anhydrotetracycline Antibiotic, tetracycline

Lomefloxacin Antibiotic, fluoroquinolone Carbamazepine Anticonvulsant

Enrofloxacin Antibiotic, fluoroquinolone Fluoxetine Antidepressant

Sarafloxacin Antibiotic, fluoroquinolone Miconazole Antifungal

Flumequine Antibiotic, fluoroquinolone Thiabendazole Antihelmintic

Lincomycin Antibiotic, macrolide Diphenhydramine Antihistamine

Azithromycin Antibiotic, macrolide Acetaminophen Analgesic

Erythromycin Antibiotic, macrolide Codeine Analgesic, narcotic

Tylosin Antibiotic, macrolide Cimetidine Antiacid reflux

Anhydroerythromycin Antibiotic, macrolide Ranitidine Antiacid reflux

Clarithromycin Antibiotic, macrolide Digoxigenin Antiarrythmic

Roxithromycin Antibiotic, macrolide Digoxin Antiarrythmic

Ampicillin Antibiotic, penicillin Diltiazem Antiarrythmic, benzothiazepine

Penicillin G Antibiotic, penicillin Dextromethorphan** Antitussive

Penicillin V Antibiotic, penicillin Atenolol Beta-blocker

Oxacillin Antibiotic, penicillin Metoprolol Beta-blocker

Cloxacillin Antibiotic, penicillin Propranolol Beta-blocker

Metformin* Antidiabetic Albuterol Bronchodialator

Sulfadiazine Antibiotic, sulfa Midazolam Sedative, benzodiazepine

Sulfathiazole Antibiotic, sulfa 1-OH Midazolam Sedative, benzodiazepine

Sulfamerazine Antibiotic, sulfa 1-OH Alprazolam Sedative, benzodiazepine

Sulfamethazine Antibiotic, sulfa Alprazolam Sedative, benzodiazepine

Sulfamethizole Antibiotic, sulfa Nordiazepam Sedative, benzodiazepine

Sulfachloropyridazine Antibiotic, sulfa 1,7-Dimethylxanthine Stimulant

Sulfamethoxazole Antibiotic, sulfa Caffeine Stimulant

Sulfadimethoxine Antibiotic, sulfa Benzoylecgonine Stimulant

Minocycline Antibiotic, tetracycline Cocaine Stimulant

Oxytetracycline Antibiotic, tetracycline Cocaethylene Stimulant

4-epi-Tetracycline Antibiotic, tetracycline Cotinine Stimulant

Tetracycline Antibiotic, tetracycline

*Metformin was analyzed using HILIC

**PPCPs not included in EPA 1694

5.5 6.0 6.5Time (min)

0

10

20

30

40

50

60

70

80

90

100

0

10

20

30

40

50

60

70

80

90

100RT: 6.60AA: 13222

RT: 6.60AA: 10873

RT: 5.84AA: 2130

5.5 6.0 6.5Time (min)

RT: 6.61AA: 15386

RT: 6.59AA: 10320RT: 5.88

AA: 6596

5.5 6.0 6.5Time (min)

RT: 6.59AA: 227

RT: 5.87AA: 64258

4-epi-CTC

CTC

pH 2.9 pH 11.3pH 6.6

Figure 2. Chromatograms showing the pH effect on chlorotetracycline (CTC).

With such a diverse range of chemical classes, the chal-lenge was in developing a single LC-MS/MS method with-out compromising the target ng/L sensitivity. Both sample pH and the % ACN in the sample affected the response of PPCPs in water when employing the online sample prepa-ration approach with the EQuan system. To determine the best method for achieving ng/L sensitivity on the TSQ Vantage™ mass spectrometer, the effects of sample pH and %ACN were investigated.

Effects of Sample pHSample pH was found to affect the response of some PPCPs in water based on chemical reactivity. During the method development, PPCPs were added to aqueous solutions at three different pHs: 2.9, 6.6, and 11.3. As shown in the chromatograms in Figure 2, chlorotetracy-cline (CTC) was readily observed at pH 2.9 and pH 6.6. However, at pH 11.3, CTC completely disappeared, being converted to 4-epi-CTC. It is important to note that no 4-epi-CTC was added to the water samples prior to LC-MS/MS analysis. All of the 4-epi-CTC detected was due to the conversion of CTC, which has been shown to have a short half-life in solutions at pH 11.2. A similar effect was observed with erythromycin, which reacted quickly in acidic solution and converted to anhydroerythromycin at pH 2.9.

The pH also affected the solubility of some PPCPs, even within the same compound class. Figure 3 displays the area response for cloxacillin and penicillin. For cloxa-

cillin, the area response at pH 2.9 and pH 6.6 is evident in the bar chart at the top left; whereas at pH 11.3, cloxacil-lin was not observed. A similar effect was seen for ampicil-lin, oxacillin, cefotaxime, and diltiazem. However, the opposite effect was observed for penicillin V (and G), as seen in the bar chart in the bottom right. The same trends were observed with LC-MS/MS (5 µL injection) as with the EQuan method (0.5 mL injection), indicating that this is a sample solubility effect.

The pH effect on the MS response was also observed with several other PPCPs when using the EQuan system. Using ranitidine as an example, the MS response was much greater at pH 11.3 than at pH 2.9 or 6.6, as shown in the chart at the top left of Figure 4. However, this pH effect was not observed when using a 5 µL injection of the water samples directly onto the analytical column at the same mass loading of ranitidine, as seen in the bar chart in the lower right of Figure 4. This difference in response is believed to be attributed to the change in the local partitioning chemistry between ranitidine and the station-ary phase of the pre-concentration column. With a 5 µL injection directly onto the analytical column, the partition-ing chemistry was not affected for a long enough period to change the retention of ranitidine. Nevertheless, under the right sample solution conditions, namely pH 11.3 and 5%-10% ACN, ranitidine and other basic PPCPs, such as cimetidine, codeine, and lincomycin, yielded quantitative trapping recovery using the EQuan system.

Effects of %ACNThe effect on the LC-MS/MS response for the PPCPs was examined as a function of the % ACN in the water sam-ples. Many of the larger, more lipophilic compounds, such as the macrolide antibiotics, showed a significant increase in area response as a function of increasing %ACN in the water sample. For tylosin and roxithromycin, the increased response was most dramatic between 5% and 10% ACN at pH 2.9. The area response increased by a factor of 3 for roxithromycin and a factor of 10 for tylosin when the %ACN was increased from 5% to 10%. The same trend was observed with LC-MS/MS (5 µL injection) as with the EQuan system, indicating that this is a sample solubility ef-fect due to the compounds’ lipophilic nature.

Although increasing the %ACN in the water sample helped the response of certain PPCPs, it caused a signifi-cant decrease in response in others if the percentage was too high (Figure 5). This effect, observed for ciprofloxacin, trimethoprim, fluoroquinolones, and sulfa drugs, was at-tributed to a loss of compound retention on the trapping column, where compounds have a greater affinity for the solvent than the trapping column stationary phase. This effect is similar to compound “break-through” on an SPE cartridge. No fall-off in MS response was observed with a 5 µL injection onto the analytical column.

The effect of decreased analyte retention with increas-ing %ACN in the water sample was also observed with cotinine using a 5 µL injection on the analytical C18 column. As Figure 6 shows, the LC peak splitting for cotinine was readily observed in acidic (red) and neutral (green) water samples. However, at pH 11.3, the cotinine peak was virtually unchanged, even at 20% ACN. This is likely due to the fact that the basic compound cotinine is uncharged at pH 11.3, which increases its affinity for the C18 stationary phase.

As seen with cotinine, the biggest challenge in develop-ing an EQuan method for PPCPs was the small, highly-

polar organic compounds. Different trapping columns and mobile phases were tested, but as expected, compromises had to be made to allow the largest breadth of PPCPs in one LC-MS/MS run. Metformin was the clearest example. Despite many approaches, no satisfactory reverse-phase LC method could be discovered because of its very high polarity. Hence, as described in EPA Method 1694, hydrophilic interaction liquid chromatography (HILIC) was used for the successful LC separation of metformin in water. Again, pH had a dramatic effect on the response of metformin (and other Group 4 PPCPs). The best response for metformin was with the water sample adjusted to pH 11.3 prior to injection on the reverse-phase EQuan trap-ping column.

EQuan Method SummaryDespite all of the challenges in the development of one single LC/MS method for this diverse group of compo-nents, a balance was found that allowed the measurement of the 67 PPCPs in water by the EQuan system, with a large majority being quantified at or below 10 ng/L using a 0.5 mL injection volume with detection on the TSQ Vantage mass spectrometer.

The best compromise for the online sample prepara-tion method was to run an acidified and a basified water sample containing 10% ACN. Figure 7 shows example chromatograms for the PPCPs in water at the ng/L level us-ing this approach. The red chromatograms were the water samples at pH 2.9, and the blue chromatograms were the water samples at pH 11.3. In general, basic conditions were preferable for analyzing the smaller, more polar com-pounds, and acidic conditions were preferable for analyz-ing the larger, more lipophilic compounds.

Figure 3. Area response plots demonstrating the pH effect on the sample solubility.

Figure 4. Area response plots for ranitidine demonstrating the pH effect on the preconcentration column.

Figure 5. Area response plots showing effect of decreased retention with increasing %ACN.

pH 2.9 pH 11.3pH 6.6

0.0 0.5 1.0 1.5 2.0 2.5Time (min)

RT: 1.84AA: 141940

RT: 1.81AA: 137918

RT: 1.84AA: 138950

RT: 1.07AA: 85007

RT: 1.84AA: 56218

0.0 0.5 1.0 1.5 2.0 2.5Time (min)

RT: 1.86AA: 133742

RT: 1.88AA: 140409

RT: 1.84AA: 137257

RT: 1.82AA: 134771

0.0 0.5 1.0 1.5 2.0 2.5Time (min)

0

20

40

60

80

100

0

20

40

60

80

100

0

20

40

60

80

100

0

20

40

60

80

100

RT: 1.83AA: 146721

RT: 1.83AA: 135126

RT: 1.41AA: 71260

RT: 1.84AA: 68484

RT: 1.06AA: 96122

RT: 1.87AA: 42653

20%

15%

10%

5%

Figure 6. Chromatograms showing the %ACN effect on LC column retention for cotinine.

4.0 4.5 5.0 5.5 6.0 6.5 7Time (min)

RT: 6.04AA: 17975

RT: 6.13AA: 37475

RT: 6.29AA: 19715

RT: 6.14AA: 43273

RT: 6.20AA: 8392

RT: 6.19AA: 17142

RT: 6.20AA: 16556

RT: 6.31AA: 71593

3.5 4.0 4.5 5.0 5.5 6.0 6Time (min)

RT: 5.41AA: 16000

RT: 5.47AA: 12712

RT: 5.50AA: 4994

RT: 5.62AA: 76608

RT: 5.63AA: 4510

RT: 5.84AA: 37897

RT: 5.95AA: 8263

2 3 4 5Time (min)

0

50

100

0

50

100

0

50

100

0

50

100

0

50

100

0

50

100

0

50

100

RT: 3.18AA: 17332

RT: 4.60AA: 4059

RT: 4.73AA: 2342

RT: 4.77AA: 14267

RT: 4.75AA: 10378

RT: 4.87AA: 10805

RT: 4.96AA: 10283

Cotinine(100 ng/L)

Albuterol(10 ng/L)

Cimetidine(1 ng/L)

Atenolol(10 ng/L)

Acetaminophen(100 ng/L)

Dimethylxanthine(100 ng/L)

Ranitidine(1 ng/L)

Codeine(10 ng/L)

Sulfadiazine(100 ng/L)

Lincomycin(1 ng/L)

Caffeine(100 ng/L)

Ampicillin(100 ng/L)

Cefotaxime(100 ng/L)

Sulfathiazole(100 ng/L)

Trimethoprim(1 ng/L)

Minocycline(100 ng/L)

Sulfamerazine(10 ng/L)

Norfloxacin(10 ng/L)

Ofloxacin(1 ng/L)

Benzoylecgonine(1 ng/L)

Oxytetracycline(100 ng/L)

Figure 7 (a,b,c). Example chromatograms of the PPCPs in water at the ng/L level. The LLOQ for each compound is listed in parentheses.

a

4.5 5.0 5.5 6.0 6.5 7.0Time (min)

0

50

100

0

50

100

0

50

100

0

50

100

0

50

100

0

50

100

0

50

100

RT: 6.29AA: 5797

RT: 6.43AA: 8520

RT: 6.55AA: 8058

RT: 6.25AA: 2716

RT: 6.61AA: 22391

RT: 6.59AA: 18998

RT: 6.79AA: 7998

RT: 6.81AA: 25414

RT: 6.02AA: 1639

6.0 6.5 7.0 7.5 8.0 8.5Time (min)

RT: 7.70AA: 7901

RT: 7.74AA: 8205

RT: 7.81AA: 8448

RT: 8.10AA: 13156

RT: 8.22AA: 42335

RT: 8.39AA: 8141

RT: 8.43AA: 4026

5 6 7 8Time (min)

RT: 6.98AA: 31162

RT: 7.02AA: 18888

RT: 7.08AA: 4554

RT: 7.14AA: 5093

RT: 7.25AA: 6110

RT: 7.33AA: 15417

RT: 7.63AA: 44411

Ciprofloxacin(1 ng/L)

Lomefloxacin(1 ng/L)

Tetracycline(10 ng/L)

Sulfamethazine(10 ng/L)

Enrofloxacin(1 ng/L)

epi-Chlorotetracycline(100 ng/L)

Sulfamethazole(10 ng/L)

Metoprolol(10 ng/L)

Demeclocycline(100 ng/L)

Sarafloxacin(1 ng/L)

Penicillin G(100 ng/L)

Digoxigenin(10 ng/L)

Cocaine(1 ng/L)

Azithromycin(10 ng/L)

Sulfachloropyridazine(10 ng/L)

Penicillin V(100 ng/L)

Doxycycline(10 ng/L)

Sulfamethoxazole(10 ng/L)

Cocaethylene(1 ng/L)

Digoxin(100 ng/L)

Propranolol(1 ng/L)

Figure 7. Example chromatograms of the PPCPs in water at the ng/L level. The LLOQ for each compound is listed in parentheses. (continued)

8 9 10

Time (min)

RT: 9.32

AA: 10169

RT: 9.45

AA: 7681

RT: 9.46

AA: 58538

RT: 9.50

AA: 5640

RT: 9.57

AA: 19522

RT: 9.79

AA: 1628

RT: 10.29

AA: 9502

7.0 7.5 8.0 8.5 9.0 9.5

Time (min)

RT: 8.76

AA: 11698

RT: 8.83

AA: 57545

RT: 9.01

AA: 9582

RT: 9.12

AA: 9402

RT: 9.15

AA: 74405

RT: 9.27

AA: 19291

RT: 9.33

AA: 3972

6.5 7.0 7.5 8.0 8.5 9.0

Time (min)

0

50

100

0

50

100

0

50

100

0

50

100

0

50

100

0

50

100

0

50

100

RT: 8.49

AA: 23044

RT: 8.56

AA: 12492

RT: 8.56

AA: 22100

RT: 8.61

AA: 6460

RT: 8.69

AA: 5490

RT: 8.71

AA: 24841

RT: 8.74

AA: 28204

Dextromethorphan(1 ng/L)

Erythromycin(10 ng/L)

Midazolam(1 ng/L)

1-OH Midazolam(1 ng/L)

Anhydrotetracycline(10 ng/L)

Diphenhydramine(1 ng/L)

Tylosin(10 ng/L)

Sulfadimethoxine(1 ng/L)

Diltiazem(1 ng/L)

Anhydroerythromycin(1 ng/L)

1-OH Alprazolam(10 ng/L)

Carbamazepine(1 ng/L)

Roxithromycin(1 ng/L)

Clarithromycin(1 ng/L)

Oxacillin(100 ng/L)

Alprazolam(1 ng/L)

Flumequine(1 ng/L)

Fluoxetine(1 ng/L)

Cloxacillin(100 ng/L)

Miconazole(1 ng/L)

Nordiazepam(1 ng/L)

Figure 7. Example chromatograms of the PPCPs in water at the ng/L level. The LLOQ for each compound is listed in parentheses. (continued)

c

b

Part of Thermo Fisher Scientific

www.thermoscientific.comLegal Notices: ©2010 Thermo Fisher Scientific Inc. All rights reserved. All trademarks are the property of Thermo Fisher Scientific Inc. and its subsidiaries. This information is presented as an example of the capabilities of Thermo Fisher Scientific Inc. products. It is not intended to encourage use of these products in any manners that might infringe the intellectual property rights of others. Specifications, terms and pricing are subject to change. Not all products are available in all countries. Please consult your local sales representative for details.

Thermo Fisher Scientific, San Jose, CA USA is ISO Certified.

AN63352_E 11/10S

In addition to these

offices, Thermo Fisher

Scientific maintains

a network of represen­

tative organizations

throughout the world.

Africa-Other +27 11 570 1840Australia +61 3 9757 4300Austria +43 1 333 50 34 0Belgium +32 53 73 42 41Canada +1 800 530 8447China +86 10 8419 3588Denmark +45 70 23 62 60 Europe-Other +43 1 333 50 34 0Finland/Norway/ Sweden +46 8 556 468 00France +33 1 60 92 48 00Germany +49 6103 408 1014India +91 22 6742 9434Italy +39 02 950 591Japan +81 45 453 9100Latin America +1 561 688 8700Middle East +43 1 333 50 34 0Netherlands +31 76 579 55 55New Zealand +64 9 980 6700Russia/CIS +43 1 333 50 34 0South Africa +27 11 570 1840Spain +34 914 845 965Switzerland +41 61 716 77 00UK +44 1442 233555USA +1 800 532 4752

ConclusionThe current EPA Method 1694 describes three different LC methods for PPCPs from Groups 1, 2, and 4, which are amenable to positive ESI-MS/MS. To simplify the method and reduce total analysis time, a single 10-minute LC-MS/MS method was developed on the EQuan system includ-ing compounds from additional pharmaceutical classes not included in the EPA method, such as beta-blockers and benzodiazepines.

The EQuan system significantly reduced sample preparation and analysis time while providing quantifica-tion of PPCPs in water at low ng/L levels. Online sample preparation of the water samples eliminated the need to use two different offline SPE methods on 1 L of water. This reduced the total analysis time from hours to minutes. The sensitivity of the TSQ Vantage mass spectrometer, using time-dependent SRMs to maximize detector duty cycle, provided low- or sub-ng/L limits of quantitation for the targeted PPCPs in water.

References1. EPA Method 1694: Pharmaceuticals and personal care products in water,

soil, sediment, and biosolids by HPLC/MS/MS, December 2007, EPA-821-R-08-002.

2. Loftin, K.A.; Adams, C.D.; Meyer, M.T.; Surampalli, R. “Effects of Ionic Strength, Temperature, and pH on Degradation of Selected Antibiotic” J. Environ. Qual., 2008, 37, 378-386.