Introduction
Hypertension and other cardiovascular diseases which have been among the leading cause of death worldwide, are preventable and manageable by medications such as antihypertensive, hypolipidemic, or anticoagulant agents.1 However, the increasing of non-adherence to antihypertensive and related drugs is a real menace to patient health and drug effectiveness. Several conventional methods have been applied to evaluate medication adherence including questionnaires, pharmacy dispense records, pill counts, or supervised administration.2 Besides, recently, drug testing in urine, oral fluid, or plasma using liquid chromatography tandem mass spectrometry (LC-MS/MS) has been proven as a valuable means for assessing the adherence of prescribed medications. The developed LC-MS/MS methods for drug adherence monitoring in general and studying cardiovascular medications in particular generally applied sample preparation processes employed solid-phase extraction or liquid-liquid extraction.3-8 This approach effectively cleans up and concentrates the analytes but significantly depends on the characteristics of the surveyed compounds as well as consumes labor, reagents, and time Nowadays, the enhancement in the sensitiveness of LC–MS/MS systems have allowed samples to be minimally diluted and then directly introduced into the analytical system. This offers a simple and faster sample preparation process (about 30 s) with minimal labor, time and reagent consumption and be able to screen the broader range of analytes in comparison to other mentioned techniques. For instance, “dilute-and-shoot” LC-MS/MS has been proven as an effective trend in doping control,9 analytical toxicology,10 or urine drug testing of a large number of antipsychotics, opioids, benzodiazepines, and other pain management medications and metabolites.11-13 As such, a limited number of antihypertensive, lipidlowering, antihyperglycemic, antithrombotic and other cardiovascular agents were successfully screened in urine applying “dilute-and-shoot” LC-MS/MS method.14-16 In which, the study of A.J. Lawson covered a largest number of antihypertensive medications but only 23 compounds.14
From the above overview, this study developed a “diluteand-shoot” LC-MS/MS method to detect a larger number of cardiovascular preventive compounds, covering 43 prescribed antihypertensive, lipid-lowering, and antithrombotic agents available worldwide. The design of experiment (DOE) was aslo applied through the method development process to achieve the effectively and reliably optimal LC-MS/MS condition with minimum experiments, time, cost, and labor consumption.17
Experimental
Material
43 surveyed cardiac drugs as well as atenolol-d7, and sulfameter (as internal standards (IS)) were provided from Sigma-Aldrich (St. Louis, MO, USA). Other IS including amlodipine-d4, clopidogrel-d4, diltiazem-d3, losartan-d4, telmisartan-d7 were supplied by TLC Pharmaceutical Standard. Formic acid, ammonium formate, HPLC-grade acetonitrile, and methanol were purchased from Daejung (Siheung, South Korea). Distilled water was prepared in the laboratory utilizing an Aqua Max water purification system supplied by Young Lin Instrument Co., Ltd. (Anyang, South Korea).
Instrumentation
The LC-MS/MS system included an Agilent 1200 series (Agilent Technologies) system combined with an API 3200 Q Trap triple-quadrupole mass spectrometer (AB SCIEX) operated with a Turbo V Ion Spray source. Analyst 1.6 software was employed for LC-MS/MS system management and data processing. The separation was performed on a Thermo Scientific Accucore RP-MS column (50 × 3.0 mm ID., 2.6 µm) combined with a C18 guard column (Phenomenex, 4.0 × 3.0 mm ID), both maintained at 50o C. Two separate gradient elution programs established with the same mobile phases: eluent A containing 8mM ammonium formate (HCOONH4) and 0.1% formic acid (HCOOH) in water, and eluent B containing 8mM HCOONH4 and 0.1% HCOOH in acetonitrile (ACN): water (90:10).
Drug calibrators and quality control samples preparation
A 1 mg/mL stock solution in methanol was made for each compound measured and IS, with the exception of 2 mg/mL for nicotinic acid and 5 mg/mL for HCTZ. Therefore, the concentration of nicotinic acid and HCTZ is correspondingly 2 times and 5 times higher than that mentioned the following solutions. Working standard mixtures of 4000 μg/L, 200 μg/L, 10 μg/L and IS working standard mixtures of 4000 μg/L were prepared by serial dissolving the stock solutions in water. All solutions were keeped at -20℃ and thawed at room temperature (25℃) before use. Fifteen calibration standards (0.25, 0.5, 1, 2, 5, 10, 20, 30, 50, 100, 200, 400, 600, 800, 1000 μg/L) were prepared by spiking an appropriate volume of the diluted standard solutions into an aliquot containing 250 μL of drug free human urine, and 200 µL of diluted IS solution, followed by dilution with water to attain a total volume of 1000 µL. Quality control (QC) samples correspond with three concentration levels (low, medium, and high) were independently prepared in the same way for all drugs measured. The sample was then vortexed and filtered using 0.45 μm filter before introducing into LCMS/MS system.
MS analyte parameters
Precursor and product ion transitions of each compound were determined by direct infusion of standard solution with positive and negative electrospray ionization (ESI) source. The multiple reaction monitoring (MRM) transitions and compound tuning parameters are shown in Table 1. According to optimization results, the optimal mode for each compound which created the higher intensity signal was selected (i.e. 39 compounds were detected in a positive ESI method and 4 compounds in a negative ESI method).
Table 1. MRM transitions, Compound tuning parameters, and tR.
In scouting phase, five MS parameters including ion spray voltage, capillary temperature, curtain gas, ion source gas 1, and ion source gas 2 were screened to identify the significant factors by applying fractional factorial design. Peak areas of poorly sensitive compounds (Amlodipine, Atenolol, Captopril, Losartan, Lovastatin, Moxonidine, Nicotinic acid, and Spironolactone) were chosen as responses. Analysis of variance (ANOVA) was utilized to assess the impacts of factors. Selected important factors were then optimized by Box-Behnken design with 15 runs including 3 centre points.
LC parameters
As the analytical column is stable at temperature below 60℃, the influence of the column temperature was studied in a range from 20℃ to 50℃ with a step of 5℃. Three LC related parameters namely flow rate, ammonium formate concentration, and percentage of eluent B at 0 min were also optimized by I-optimal design with 20 runs. Intensities of poor sensitive compounds were chosen as responses.
Method validation
Selectivity
The selectivity of method was studied by comparing six drug-free urine samples from six individual sources and drug-free urine samples spiked with a surveyed medications mixture at lower limit of quantification (LLOQ) concentrations. The absence of interfering peaks at retention times of analytes indicated satisfactory selectivity.
Sensitivity
The limit of detection (LOD) was assessed by the analyte concentration with the signal-to-noise (S/N) ratio was > 3. The LLOQ concentration was determined at which the S/N ratio was ≥ 10 as well as the precision (assessed by relative standard deviation, RSD) and variance of accuracy (relative error, RE) were ≤ 20%.
Carryover
The carryover was tested by analyzing the blank samples right away the upper limit of quantification (ULOQ) samples (n = 3). The carryover should ideally be < 20%.
Matrix effect
The matrix factors of the analytes were assessed by comparing the analyte/IS ratio in urine samples and water (solvent) at low, medium, and high concentration in three separate experiments (n = 3). Average percentage difference between the two should preferably be between - 20% and 20%.
Linearity
The linearity was tested within the concentration range from LLOQ to ULOQ concentration using a weighting factor of 1/x in the linear regression analysis. Linearity was evaluated basing on the coefficient of determination (R2) in five replicates. R2 value of >0.95 indicated acceptable linear.
Precision and accuracy
The intra-day, inter-day precisions, and accuracy were assessed by analyzing five replicates on same day, and over three different days of four concentrations: LLOQ, low of quantification (LQC), medium of quantification (MQC), and high of quantification (HQC). Standard curves for each batch were prepared and analysed on the same day to determine the concentration of each QC sample. RSD and RE were also calculated to evaluate the precision and accuracy.
Stability
The stability of all compounds in urinary samples was investigated at 3 QC concentrations (LQC, MQC, and HQC) in three replicates. The QC samples were stored under 4 different storage conditions before analyzing: 24 h at room temperature (25℃), 2 weeks at -20℃, three cycles of freezing (-60℃ for 12 h) and thawing (room temperature), and autosampler 5℃ for 24 h. An analyte was considered to be stable in urine when the calculated concentrations were 85–115% of those of the freshly prepared samples.
Results
Method development
Preliminary experiments were conducted with the following gradient LC condition proposed by Lawson et al.: eluent A including 1mM HCOONH4 and 0.1% HCOOH in water, and eluent B including 1mM HCOONH4 and 0.1% HCOOH in 90% ACN.14 Some analytes such as captopril, losartan, lovastatin, moxonidine, nicotinic acid, hydrochlorothiazide (HCTZ) or spironolactone showed the poor sensitivity and chromatographic performance, so further experiments were conducted to obtain the more suitable condition.
Optimization of MS parameters
At first, five MS parameters including ion spray voltage, capillary temperature, curtain gas, ion source gas 1, and ion source gas 2 were screened to identify the significant factors by applying fractional factorial design. Since pvalue < 0.05, ion spray voltage, capillary temperature, and curtain gas were demonstrated the more importance and selected for optimization step. These MS selected factors were optimized by Box-Behnken design with 15 runs including 3 centre points. From the results of Box-Behnken design, optimal MS conditions were revealed. The desirability values were 0.954 and 0.427 for negative and positive mode, respectively (Table 2).
Table 2. The optimization of MS parameters
Optimization of LC parameters
The results of column temperature investigation showed that high temperatures faster elution of analytes, improved peak shapes, and obtained the acceptable sensitivity (peak area and peak height). Therefore, the temperature of analytical column was stabled at 50°C in following experiments.
Three other LC related parameters namely flow rate, ammonium formate concentration, and percentage of eluent B at 0 min were also optimized by I-optimal design with 20 runs. Intensities of poor sensitive compounds were chosen as responses. At optimal condition, the desirability values were 0.943 and 0.466 for negative and positive mode, respectively (Table 3).
Table 3. The optimization of LC condition.
Overall, there were the significant differences in desirability values between positive and negative mode since the number of responses of positive mode (8) was higher than that of negative one (4). Despite the low desirability, the sensitivity and chromatographic performance of almost surveyed compounds was acceptable and good enough for drug screening method. Therefore, the finally optimal LCMS/MS was selected following DOE results (Table 2 and 3). The complete chromatograms all analytes were shown in Figure 1.
Figure 1. Chromatograms of 40 analytes in a positive ESI mode (a) and 4 analytes in a negative ESI mode (b): 1. Moxonidine, 2. Nicotinic acid, 3. Atenolol, 4. Lisinopril, 5. Clonidine, 6. Nadolol, 7. Triamterene, 8. Enalapril, 9. Pindolol, 10. Terazosin, 11. Captopril, 12. Acebutolol, 13. Metoprolol, 14. Celiprolol, 15. Labetalol, 16. Bisoprolol, 17. Doxazosin, 18. Propranolol, 19. Perindopril, 20. Diltiazem, 21. Bevantolol, 22. Betaxolol, 23. Indapamide, 24. Ramipril, 25. Carvedilol, 26. Amlodipine, 27. Losartan, 28. Olmesartan, 29. Rosuvastatin, 30. Irbesartan, 31. Pitavastatin, 32. Telmisartan, 33. Spironolactone, 34. Warfarin, 35. Atorvastatin, 36. Fluvastatin, 37. Mevastatin, 38. Lovastatin, 39. Clopidogrel.
Method validation
Selectivity and sensitivity
There were no considerable interfering peaks observed at the retention times expected for the analytesf or IS. The extracted ion chromatograms of 43 interested compounds and IS were shown in Supporting Information (Figure S1). The LODs were from 0.1 to 50 ppb, and the LLOQs ranged from 0.25 to 100 ppb (Table 4).
Table 4. QC concentrations, sensitivity, linearity, carry over, matrix effect validation.
Carryover
The carryover of the all surveyed compounds was less than 19.48% of the LLOQ (Table 4).
Matrix effect
Mean percentage difference of the analyte/IS ratio between human urine and water samples was from - 19.92% to 18.92% for all but three analytes (bevantolol, carvedilol, nicotinic acid) (Table 4).
Linearity
The coefficient of determination (R2) of all compounds was more than 0.9870 showing the acceptable linearity of the developed method.
Precision and accuracy
The good precision and accuracy were observed for all compounds (Table 5). The RSD% was not more than 19.87% and 18.54% for intra-assay and inter-assay precision, respectively. The recovery of each compound was in the range from 84.54 to 119.78%.
Table 5. Precision, and accuracy results.
Stability
The results of stability validation are shown in Figure 2 and Supporting Information (Table S1). Under four storage conditions, the mean of recoveries and RSD satisfied the acceptance criteria (± 15% of the control values) for all analytes but carvedilol (recovery of 73.07% at LQC). No significant degradation was detected, so most analytes were assessed to be stable in urine under all described conditions.
Figure 2. Stability validation.
Discussion
A quick, cost-effective, and specific “dilute-and-shoot” LC–MS/MS method with minimal sample preparation process was investigated and validated for the determination of 43 prescribed antihypertensive and related drugs in human urine. The optimal mass spectrometric and chromatographic parameters were investigated by applying experimental design approach. The validation results indicated that this screening LC-MS/MS method was specific, reproducible, and sensitive with the limit of detection from 0.1 to 50.0 µg/L. For now, this dilute-andshoot LC–MS/MS method has simultaneously screened a largest number of hypertensive and related drugs in human urine. In comparison with other related literatures, of the 24 drugs compared, 11 were improved the sensitivity and 10 had higher concentration of detection (Table 6). The less sensitivity of these compounds could be due to the simultaneously screening a larger number of analytes in different structures. The assay could be optimized for concurrently analysis 43 drugs but difficult to obtain the best solution for each compound. In particular, 4 of 10 less sensitive drugs belong statin group, which has a more specialized dilute-and-shoot LC–MS/MS method developed by Jang et al. 2018.16
Table 6. Comparison with related literatures.
Future expansion of the assay could include the addition of drug metabolites, because some drugs have short halflife as well as are metabolised and excreted as metabolites in urine, such as spironolactone, aspirin, ramipril, or fluvastatin. The assay also could be applied to the analysis of actual urine samples to validate its clinical effectiveness in further experiments.
Conclusions
In conclusion, the developed method could be a promising approach for screening the presence of prescribed cardiovascular drugs in human urine.
Supporting Information
Supporting information is available at https://drive. google.com/file/d/1QHBrI7yTj0MhxCK1U-8tcBhLXTn__ uGv/view?usp=sharing.
Acknowledgments
This research did not receive any specific grant from public, commercial, or non-profit funding agencies. The authors thank the Institute of New Drug Development Research and the Central Laboratory of Kangwon National University for the use of their analytical equipment.
References
- World Health Organization. Cardiovascular diseases (CVDs), https://www.who.int/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds)
- Abegaz, T. M.; Shehab, A.; Gebreyohannes, E. A.; Akshaya, Bhagavathula, A. S.; Elnour, A. A. Medicine 2017, 96, e5641, DOI: 10.1097/MD.0000000000005641.
- Punt, A. M.; Stienstra, N. A.; van Kleef, M. E. A.; Lafeber, M.; Spiering, W.; Blankestijn, P. J.; Bots, M. L.; Maarseveen, E. M. V. J. Chromatogr. B 2019, 1121, 103, DOI: 10.1016/j.jchromb.2019.05.013.
- Richter, L. H. J.; Jacobs, C. M.; Mahfoud, F.; Kindermann, I.; Bohm, M.; Meyer, M. R. Anal. Chim. Acta 2019, 1070, 69, DOI: 10.1016/j.aca.2019.04.026.
- Dias, E.; Hachey, B.; McNaughton, C.; Nian, H.; Yu, C.; Straka, B.; Brown, N. J.; Caprioli, R. M. J. Chromatogr. B 2014, 937, 44, DOI: 10.1016/j.jchromb.2013.08.010.
- Tomaszewski, M.; White, C.; Patel, P.; Masca, N.; Damani, R.; Hepworth, J.; Samani, N. J.; Gupta, P.; Madira, W.; Stanley, A.; Williams, B. Heart 2014, 100, 855, DOI: 10.1136/heartjnl-2013-305063.
- Gonzalez, O.; Alonso, R. M.; Ferreiros, N.; Weinmann, W.; Zimmermann, R.; Dresen, S. J. Chromatogr. B. 2011, 879, 243, DOI: 10.1016/j.jchromb.2010.12.007.
- Murray, G. J.; Danaceau, J. P. J. Chromatogr. B 2009, 877, 3857, DOI: 10.1016/j.jchromb.2009.09.036.
- Guddat, S.; Solymos, E.; Orlovius, A.; Thomas, A.; Sigmund, G.; Geyer, H.; Thevis, M.; Schanzer, W. Drug Test Anal. 2011, 3, 836, DOI: 10.1002/dta.372.
- Sanchis, Y.; Coscolla, C.; Yusa, V. Talanta 2019, 202, 42, DOI: 10.1016/j.talanta.2019.04.048.
- Feng, S.; Enders, J. R.; Cummings, O. T.; Strickland, E. C.; McIntire, T.; McIntire, G. J. Anal. Toxicol. 2020, 44, 331, DOI: 10.1093/jat/bkz098.
- Cao, Z.; Kaleta, E.; Wang, P. J. Anal. Toxicol. 2015, 39, 335, DOI: 10.1093/jat/bkv024.
- Dahlin, J. L.; Palte, M. J.; LaMacchia, J.; Petrides, A. K. The JALM, 2019, 3, 974, DOI: 10.1373/jalm.2018.027342.
- Lawson, A. J.; Shipman, K. E.; George, S.; Dasgupta, I. J. Anal. Toxicol. 2016, 40, 17, DOI: 10.1093/jat/bkv102.
- Truong, Q. K., Mai X. L.; Lee, J. Y.; Rhee, J.; Vinh, D.; Hong, J.; Kim, K. H. Arch. Pharm. Res. 2018, 41, 530, DOI: 10.1007/s12272-018-1011-9.
- Jang, H.; Mai, X. L.; Lee, G.; Ahn, J. H.; Rhee, J.; Truong, Q. K.; Vinh, D.; Hong, J.; Kim, K. H. Mass Spectrom. Lett. 2018, 9, 95, DOI: 10.5478/MSL.2018.9.4.95.
- Sahu, P. K.; Ramisetti, N. R.; Cecchi, T.; Swain, S.; Patro, C. S.; Panda, J. J. Pharm. Biomed. Anal. 2018, 147, 590, DOI: 10.1016/j.jpba.2017.05.006.