Quantification of valproic acid in dried blood spots

Scandinavian Journal of Clinical and Laboratory Investigation

ISSN: 0036-5513 (Print) 1502-7686 (Online) Journal homepage: http://www.tandfonline.com/loi/iclb20

Quantification of valproic acid in dried blood spots Anton Pohanka, Margarita Mahindi, Michèle Masquelier, Lars L. Gustafsson & Olof Beck To cite this article: Anton Pohanka, Margarita Mahindi, Michèle Masquelier, Lars L. Gustafsson & Olof Beck (2014) Quantification of valproic acid in dried blood spots, Scandinavian Journal of Clinical and Laboratory Investigation, 74:7, 648-652 To link to this article: http://dx.doi.org/10.3109/00365513.2014.933360

Published online: 25 Jul 2014.

Submit your article to this journal

Article views: 95

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=iclb20 Download by: [Karolinska Institutet, University Library]

Date: 06 December 2015, At: 00:51

Scandinavian Journal of Clinical & Laboratory Investigation, 2014; 74: 648–652

TECHNICAL NOTE

Quantification of valproic acid in dried blood spots

Downloaded by [Karolinska Institutet, University Library] at 00:51 06 December 2015

ANTON POHANKA1, MARGARITA MAHINDI1, MICHÈLE MASQUELIER2, LARS L. GUSTAFSSON1,2 & OLOF BECK1 1Department of Laboratory Medicine, Clinical Pharmacology Unit, Karolinska Institute, Stockholm, Sweden, and 2Clinical Pharmacology, Karolinska University Hospital, Stockholm, Sweden

Abstract Background. Therapeutic drug monitoring (TDM) of the antiepileptic drug valproic acid (VPA) is recommended in patients with multiple drug therapy or with concomitant disabilities to ensure treatment efficacy and avoid adverse reactions in both adults and children. The use of sampling techniques compatible with home sampling, such as dried blood spot sampling could potentially facilitate this for patients. Aim. To assess the usefulness of a bioanalytical method for quantification of VPA in dried blood spots. Materials and methods. Quantification was based on liquid chromatography-mass spectrometry (LC-MS), both for the DBS method and the plasma-based reference method. Results. The method was validated in the range 10–1200 μmol/L. Total imprecision ranged from 4.9–8.9 (%CV) and accuracy was within ⫾ 14%. Conclusion. The validated method has potential for evaluation in therapeutic drug monitoring in combination with home sampling of DBS. The impact of spot size can be controlled through acceptance criteria and hematocrit in the range 30–60% can be accepted in sampling. Comparison of VPA levels between plasma and whole blood cannot be done without considering the blood-plasma ratio. Key Words: Blood-plasma ratio, dried blood spot, home sampling, LCMS, punching, selected ion monitoring, valproic acid

Introduction Valproic acid (VPA) is a first-line treatment for any form of epilepsy in monotherapy or as an adjunctive alternative in children as well as in adults and is increasingly used as an add-on therapy in forms of bipolar diseases [1–3]. There is a need to individualize valproic acid due to poor relationship between dose and valproic acid concentrations together with potential influence of important drug–drug interactions, in particular with lamotrigine [1,3,4]. Traditionally, plasma has been used as a matrix for monitoring antiepileptic drugs including VPA concentrations, necessitating invasive capillary or venous blood sampling at a health clinic or hospital [4]. The possibility of performing sampling at home will help patients by decreasing the number of visits to health care sites. Dried blood spots (DBS) as a matrix has a potential for patient home sampling and has a long standing as a sampling

technique, especially for qualitative methods as in national neonatal screening programs [5]. TDMmethods based on DBS have been published for several drugs [6], including several antiepileptic drugs [7–9], including notes on VPA. In DBS-based sampling, additional parameters of the bioanalytical method needs to be understood and validated [10,11], most noteworthy are the influence of hematocrit, dried blood spot volume and the role of punching size sampled for assayed concentrations [12,13]. Comparison with existing methods used for TDM need to be carefully evaluated as a shift of sample matrix from plasma to whole blood can be expected to cause different concentration levels necessitating the evaluation of the drug blood-plasma ratio. For VPA there is an expected bias in concentrations between plasma and whole blood, due to the VPA blood-plasma ratio of 0.64 [14]. Reported notes

Correspondence: Anton Pohanka, TDM Laboratory, Clinical Pharmacology, C1:68, SE-141 86 Stockholm, Sweden. Tel: ⫹ 46 8 58581195. Fax: ⫹ 46 8 58581050. E-mail: [email protected] (Received 13 August 2013 ; accepted 30 May 2014 ) ISSN 0036-5513 print/ISSN 1502-7686 online © 2014 Informa Healthcare DOI: 10.3109/00365513.2014.933360

Quantification of valproic acid of quantification of VPA in DBSs were based on GC-MS [15] and LC-MS/MS [16]. For plasmabased quantification of VPA, several methods are published, either quantifying valproic acid alone [17,18] or in combination with other analytes [19,20]. The aim of this work was to assess the influence of hematocrit and spot size on quantification as well as bias between whole blood and plasma concentrations in order to establish DBS for home-based TDM.


Materials and methods Sampling material and chemicals Whatman 903 paper sheets of A4 size (GE Healthcare, Wakuesha, WI) were cut to convenient sizes and used for all DBS work. For punching, stainless steel custom made punches (Gerdins Components, Mjällom, Sweden) and commercial punches from Biltema (Biltema, Helsingborg, Sweden) were used. Valproic acid was purchased from Sigma (SigmaAldrich, St. Louis, MA), heptanoic acid, methanol (HPLC gradient grade) ammonia solution 25% (Suprapur), acetic acid (p.a.) and hydrochloric acid (Suprapur) from Merck KGaA (Darmstadt, Germany). De-ionized water was produced with a Milli-Q plant (Millipore, Billerica, MA). Blank blood and blank plasma were obtained from anonymized leftover TDM-samples at the TDM laboratory in the Department of Clinical Pharmacology, mainly from patients receiving immunosuppressive treatment. Plasma-based reference method The method is used 7 days a week at the TDM Laboratory in the Department of Clinical Pharmacology, Karolinska University Hospital, for quantification of valproic acid plasma concentrations with an approximate throughput of 2500 samples/year. The quantification range was 100–1500 μmol/L. Internal quality control samples at 150 and 700 μmol/L exhibit a total CV% of 4.8% and 1.5%, respectively. The method was launched in spring 2011 and takes part in the UKNEQAS/LGC proficiency test program. During this period all reported results were within the accepted accuracy levels set by the program.

Preparation of DBS calibrators and quality control (QC)-samples Valproic acid stock and working solutions in 0.01 mol/L HCl (aq) were spiked into pooled blank blood with an approximate hematocrit of 40. Spiking volumes never exceeded 5% of the final volume. The spiked blood was pipetted onto filter paper in 20 μL aliquots and left to dry overnight in a ventilated lab

649

hood and then kept in dry conditions at room temperature until use. Average spot sizes were 7 mm diameter. For testing of basic DBS parameters whole blood was spiked with 490 or 600 μmol/L of valproic acid and spots between 20–100 μL were produced by manual pipetting onto filter paper. DBS sample processing Validation was performed with fixed volume punching, i.e. an oversized punch was used (diameter 8 or 9 mm), to ensure collection of the whole spot including a discernible white paper rim. For experiments based on fixed area samples, punching was centered on the spot, using 5 or 6 mm punches. The spots were transferred to HPLC vials and 150 μL extraction solution (methanol containing internal standard) was added. The samples were agitated for 60 min, followed by centrifugation at 2100 g for 5 min. The filter paper discs were not removed from the HPLC vials during analysis. Validation of the DBS method Validation was performed in accordance with European Medicines Agency guidelines with the exception of using three concentration levels in accuracy and precision assessment, and assessment of sample stability limited to room temperature and ⫹ 8°C. Analysis of processed DBS samples was performed with identical instrumental parameters as for the plasma-based reference method. Assessment of basic parameters in DBS The influence of hematocrit on DBS concentration was investigated by spiking valproic acid at 600 μmol/L in whole blood with different hematocrit in ten-unit increments between 20% and 70%. The spiked blood was aliquoted in 20 μL spots. The spots were centrally punched with a 5 mm punch and quantified with an identically punched calibration curve based on 20 μL blood spots with blank blood of an approximate hematocrit of 40%. The hematocrit 40 sample was set as reference. Influence of sample volume on spot homogeneity was performed by spiking whole blood of hematocrit 40% to a VPA concentration of 600 μmol/L and, after agitation, aliquoting blood spots of 20, 30, 40, 50 and 100 μL size. The spots were punched with a 6 mm punch and analyzed as above. The 20 μL sample was set as reference. Method comparison A total of 34 venous blood samples from anonymized patients treated with valproic acid were quantified with the reference plasma method and the DBS method. Triplicate dried blood spots of 20 μL were

A. Pohanka et al.

Results


Validation results The method was validated for quantification in the range 10–1200 μmol/L and met standard European Medicines Agency criteria with total imprecision below 8.9 (%CV) and accuracy within ⫾ 14% for all QC-levels. Matrix effects, recovery and process efficiency were all found to be within 100 ⫾ 13%. No interfering peaks were seen in any of the tested blank blood matrices. Carry-over was found to be below 0.2%, equaling a carry-over of highest calibration point of less that 25% of lower limit of quantification. Repetitive quantification of QCsamples up to 42 days was within ⫾ 15% of initial concentrations. Method comparison Calculated patient sample concentrations ranged from 140–750 with the reference plasma method and from 120–569 with the DBS method. See Figure 1 for plots of the method comparison data. DBS basic parameters Influence of hematocrit on DBS concentration. Spiked whole blood with high hematocrit (60 and 70) resulted in visually smaller and darker colored 20 μL spots compared to low hematocrit blood (20 and 30). Using a ⫾ 15% accuracy criterium, only valproic acid levels of the 70% hematocrit samples were significantly different from the reference 40% hematocrit sample. For results, see Figure 2. Influence of sample volume on spot homogeneity. No spot volumes showed any significant differences from the reference 20 μL spot volume. Results of spot size are presented in Figure 3.

Discussion Method development and validation The validation was based on controlled volume dried blood spots by using 20 μL spots for calibrators and QC-samples, punching the whole dried blood spot for extraction The same approach was applied for the method comparison. This approach was chosen to eliminate the possible impact of uncontrolled spot volume uncertainty if punching a small fixed diameter of the blood spot.

(A) DBS blood concentration (µM)

collected before separation of plasma and analysis of plasma levels. DBS samples were kept at room temperature for up to 42 days until analysis.

Relative difference (DBS minus plasma)%

650

Plasma conc vs DBS conc y = 0.8045x -15.855 R2 = 0.9013

600 400 200 0

0

(B)

200 400 600 Plasma concentration (µM) Relative difference between DBS and plasma

10% 0% –10% –20% –30% –40% –50%

0

200 400 600 Average concentration (µM)

Figure 1. Method comparison between venous dried blood spots and plasma sampling for valproic acid. (A) Linear regression of data. (B) Relative Bland-Altman plot of method comparison data. The middle dotted line represents the average bias between methods while the upper and lower dotted lines indicate the 95% confidence interval. (In both graphs n ⫽ 32 after removing two samples resulting in low-range non-quantifiable concentrations in either method).

Method comparison Comparison between the reference plasma-based method and DBS is a comparison between valproic acid plasma levels and valproic acid venous blood levels. Linear regression of the method comparison data exhibits a slope of 0.80 with a correlation of r2 ⫽ 0.90 (Figure 1A) and an average bias of ⫺ 27% (95% CI –49 to ⫺ 5, Figure 1B) for the DBS method compared to the reference plasma method. Kong et al. [15] suggested use of a conversion formula for DBS vs plasma concentrations based on the found consistent bias in their comparison. An argument against using a factor would be the possible impact of hematocrit on the bias between plasma and whole blood concentrations. Assessment of basic parameters The impact of spot size on concentration (Figure 3) shows that there is a trend of higher concentrations for larger volume spots but not even the 100 μL spot

Quantification of valproic acid

1.4 1.2 1 0.8 0.6 0.4 0.2 0

20%

30%

40% 50% Hematocrit

60%

70%

Figure 2. Quantification of a centered 5 mm punch from 20 μL dried blood spots of different hematocrit, all n ⫽ 4 and spiked to a nominal concentration of 600 μmol/L of valproic acid. Data was normalized to hematocrit 40% as the calibration curve was based on blood spots with an approximate hematocrit of 40%. Dotted lines represent ⫾ 15% from nominal concentration while vertical bars represent 95% confidence interval.

had significantly different concentrations compared to the 20 μL spot. This suggests a moderate chromatographic effect of the blood distribution during wetting and drying of the filter paper. Adjusting calibrator and QC-sample spot sizes to patient sampling volumes with clear acceptance criteria of spot sizes could minimize spot size variation impact on method uncertainty. The marginal effect of spot size makes patient sampling of unknown volumes feasible. Hematocrit levels show a more significant impact with higher concentrations for higher hematocrit (Figure 2). The main reason is probably the higher viscosity of a high hematocrit resulting in smaller spots. Still, only spiked blank blood with extreme hematocrit values fall outside ⫾ 15% of the

nominal value. The influence of hematocrit in spot size variation need to be considered but hematocrit has a greater impact when comparing concentrations in plasma versus whole blood. The blood plasma ratio of valproic acid, 0.64, means that valproic acid is mainly present in the plasma fraction of the blood. A higher hematocrit will theoretically result in a greater difference between plasma and DBS concentrations. The tendency that high hematocrit will result in smaller DBSs will counteract this relation but counts for an insignificant compensation. Extreme hematocrit levels can also be found in conjunction with several diseases and conditions. Thus, for rare cases there is a need to evaluate patient hematocrit levels to avoid misinterpretation of DBS analysis.

Conclusion An analytical method for valproic acid in venous dried blood spots was developed and validated. The use of blood spot sizes between 20 and 100 μL did not result in significantly different concentrations of VPA and the method was found to be robust in the hematocrit range 30–60%. The blood-plasma ratio of VPA results in significantly higher plasma concentrations than DBS concentrations but the use of a conversion factor needs to be carefully evaluated, as diverging hematocrit levels influences the differences. With consideration taken of these parameters the method can be evaluated in therapeutic drug monitoring with home sampling of DBSs, in comparison with traditional plasma sampling.

Acknowledgements None

Comparison of different blod spot volumes (n=4) VPA conc. relative to sample vol. 20 µL


VPA conc relative to hematocrit 40%

Comparison of different hematocrit levels (n=4) 1.6

651

1.4

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the article. This work was supported by grants awarded by Stockholm County Council (ALF 2011-0619 and ALF 2012-0413) and by the Swedish Research Council (Lars L. Gustafsson (VR 2011-3440 and VR 2011-7381).

1.2 1 0.8 0.6 0.4 0.2 0

20

30

40 50 µL sample volume

75

100

Figure 3. Quantification of valproic acid concentration for a centered 6 mm punch from dried blood spots of different volumes. All spots were aliquoted from blank blood spiked to a nominal concentration of 600 μmol/L of valproic acid. Data was normalized to the 20 μL results as the calibration curve was based on this blood spot size. Dotted lines represent ⫾ 15% from nominal concentration while vertical bars represent the 95% confidence interval.

References [1] Tomson T, Dahl ML, Kimland E. Therapeutic monitoring of antiepileptic drugs for epilepsy. Cochrane Db Syst Rev 2007:2. [2] Haymond J, Ensom MHH. Does valproic acid warrant therapeutic drug monitoring in bipolar affective disorder? Ther Drug Monit 2010;32:19–29. [3] Patsalos PN, Berry DJ, Bourgeois BFD, Cloyd JC, Glauser TA, Johannessen SI, Leppik IE, Tomson T, Perucca E. Antiepileptic drugs – best practice guidelines for

652

[4]

[5] [6]


[7]

[8]

[9]

[10]

[11]

A. Pohanka et al. therapeutic drug monitoring. A position paper by the subcommission on therapeutic drug monitoring, ILAE Commission on Therapeutic Strategies. Epilepsia 2008; 49:1239–76. Patsalos PN, Berry DJ. Therapeutic drug monitoring of antiepileptic drugs by use of saliva. Ther Drug Monit 2013;35:4–29. Marsden D, Larson C, Levy HL. Newborn screening for metabolic disorders. J Pediatr 2006;148:577–84. Spooner N, Lad R, Barfield M. Dried blood spots as a sample collection technique for the determination of pharmacokinetics in clinical studies: considerations for the validation of a quantitative bioanalytical method. Anal Chem 2009;81:1557–63. AbuRuz S, Al-Ghazawi M, Al-Hiari Y. A simple dried blood spot assay for therapeutic drug monitoring of lamotrigine. Chromatographia 2010;71:1093–9. la Marca G, Malvagia S, Filippi L, Fiorini P, Innocenti M, Luceri F, Pieraccini G, Moneti G, Francese S, Dani FR, Guerrini R. Rapid assay of topiramate in dried blood spots by a new liquid chromatography-tandem mass spectrometric method. J Pharmaceutand Biomed 2008;48: 1392–6. la Marca G, Malvagia S, Filippi L, Luceri F, Moneti G, Guerrini R. A new rapid micromethod for the assay of phenobarbital from dried blood spots by LC-tandem mass spectrometry. Epilepsia 2009;50:2658–62. Edelbroek PM, van der Heijden J, Stolk LML. Dried blood spot methods in therapeutic drug monitoring: methods, assays, and pitfalls. Ther Drug Monit 2009;31:327–36. Timmerman P, White S, Globig S, Luedtke S, Brunet L, Smeraglia J. EBF recommendation on the validation of bioanalytical methods for dried blood spots. Bioanalysis 2011;3:1567–75.

[12] O’Mara M, Hudson-Curtis B, Olson K, Yueh Y, Dunn J, Spooner N. The effect of hematocrit and punch location on assay bias during quantitative bioanalysis of dried blood spot samples. Bioanalysis 2011;3:2335–47. [13] Denniff P, Spooner N. The effect of hematocrit on assay bias when using DBS samples for the quantitative bioanalysis of drugs. Bioanalysis 2010;2:1385–95. [14] Paixao P, Gouveia LF, Morais JAG. Prediction of drug distribution within blood. Eur J Pharm Sci 2009;36:544–54. [15] Kong ST, Lim WH, Wang HY, Ng YL, Ho PC, Lim S. Comparison of dried blood spots and plasma valproic acid and carbamazepine levels. Epilepsia 2012;53:116. [16] Janis GC, Walker K, Collins JA. Determination of anticonvulsant drug concentrations in dried capillary blood spotted on filter paper. Epilepsia 2004;45:145. [17] Fernandez-Campos F, Calpena A, Soy D, Colom H. Determination of total and unbound concentrations of valproic acid in human plasma by liquid chromatographytamdem mass spectrometry. J Liq Chromatogr R T 2012; 35:1171–83. [18] Matsuura K, Ohmori T, Nakamura M, Itoh Y, Hirano K. A simple and rapid determination of valproic acid in human plasma using a non-porous silica column and liquid chromatography with tandem mass spectrometric detection. Biomed Chromatogr 2008;22:387–93. [19] Shibata M, Hashi S, Nakanishi H, Masuda S, Katsura T, Yano I. Detection of 22 antiepileptic drugs by ultraperformance liquid chromatography coupled with tandem mass spectrometry applicable to routine therapeutic drug monitoring. Biomed Chromatogr 2012;26:1519–28. [20] Hwang K, Song SS, Lee JH, Shin MC, Seo JS. Development and validation of an LC/MS/MS method for determination of valproic acid and its metabolite 2-propyl-4-pentenoic acid in monkey plasma. J Korean Soc Appl Bi 2012;55:13–7.