Check for updates
A method for the minimally invasive drug monitoring of mitotane by means of volumetric absorptive microsampling for a home-based therapeutic drug monitoring
Bettina Friedl1 . Max Kurlbaum 2,3 . Matthias Kroiss2,4 . Martin Fassnacht2,4 . Oliver Scherf-Clavel 1 D
Received: 30 January 2019 / Revised: 24 March 2019 / Accepted: 24 April 2019 C Springer-Verlag GmbH Germany, part of Springer Nature 2019 ☒
Abstract
Mitotane is the only currently approved treatment for adrenocortical carcinoma (ACC), a rare endocrine malignancy. Plasma levels within the range of 14 to 20 mg L-1 are correlated with higher clinical efficacy and manageable toxicity. Because of this narrow therapeutic index and slow pharmacokinetics, therapeutic drug monitoring is an essential element of mitotane therapy. A small step towards the therapeutic drug monitoring (TDM) by volumetric absorptive microsampling (VAMS) was made with this work. A simple method enabling the patient to collect capillary blood at home for the control of mitotane blood concentration was developed and characterized using MITRA™M VAMS 20 HL microsampler. Dried blood samples were extracted prior to HPLC- UV analysis. Mitotane and the internal standard dicofol (DIC) were detected at 230 nm by ultra-violet detection after separation on a C8 reversed phase column. The assay was validated in the range of 1 to 50 mg L-1. Dried samples were stable at room temperature and at 2-8 ℃ for 1 week. At 37 ℃, a substantial amount of the analyte was lost probably due to evaporation. Hematocrit bias, a common problem of conventional dried blood techniques, was acceptable in the tested range. However, a significant difference in recovery from spiked and authentic patient blood was detected. Comparison of mitotane concentration in dried blood samples (CDBS) by VAMS with venous plasma in patients on mitotane therapy demonstrated poor correlation of CDBS with the concentration in plasma (Cp). In conclusion, application of VAMS in clinical routine for mitotane TDM appears to be of limited value in the absence of a method-specific target range.
Keywords Volumetric absorptive microsampling . Mitotane . Adrenocortical carcinoma . Therapeutic drug monitoring . HPLC-UV
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s00216-019-01868-1) contains supplementary material, which is available to authorized users.
☒ Oliver Scherf-Clavel oliver.scherf-clavel@uni-wuerzburg.de
1 Institute for Pharmacy and Food Chemistry, University of Würzburg, Am Hubland, 97074 Würzburg, Germany
2 Department of Internal Medicine I, Division of Endocrinology/ Diabetology, University Hospital Würzburg, Oberdürrbacher Str. 6, 97080 Würzburg, Germany
3 Central Laboratory, Clinical Chemistry and Laboratory Medicine, University Hospital Würzburg, Oberdürrbacher Str. 6, 97080 Würzburg, Germany
4 Comprehensive Cancer Center, University of Würzburg, Josef-Schneider-Str. 6, 97080 Würzburg, Germany
Introduction
Adrenocortical carcinoma (ACC) is a rare and aggressive ma- lignancy of the adrenal gland. Surgery is the only curative treatment option but the risk of local recurrence even after complete resection is high and metastatic spread frequent. In advanced stages, ACC has a dismal prognosis with an overall survival of about 12-15 months [1]. Mitotane (o,p’-DDD), an orally administered chemical derivative of the insecticide di- chlorodiphenyltrichloroethane (DDT) [2-5], is the only ap- proved drug and applied both in an adjuvant setting and in metastatic disease [2-5]. Because of its high lipophilicity, mitotane exhibits exceptionally unfavorable pharmacokinet- ics. Accumulation of the compound in fat tissue may lead to elimination half-lives of 18 to 159 days [6]. Mitotane is almost exclusively eliminated in the form of metabolites via bile or urine. Main metabolites are the corresponding acid o,p’-DDA, which can be detected at tenfold higher concentration in blood
than mitotane itself, and the unsaturated derivative o,p’-DDE, which is barely detectable in most cases (Fig. 1) [6, 7]. Mitotane has been in use for treatment of ACC for decades and the substance and its metabolites have been the subject of numerous investigations over the past decades [7-9]. Although several studies demonstrated that a mitotane plasma level> 14 mg L-1 is associated with significant clinical benefit [10-12], many problematic issues of mitotane treatment, such as dose response effects, lack a certain level of evidence. Furthermore, there is no clear indication that plasma levels of the metabolites are correlated with clinical benefit or toxic- ity [13]. Due to the nature of the compound, typical toxic effects are common [14] and occur more frequently when high-dose regimens are applied (e.g., gastrointestinal disorder, which might be dose limiting) and at mitotane plasma levels of>20 mg L-(e.g., neurotoxicity, hepatic disorder, leucope- nia). However, plasma levels > 20 mg l are not associated with improved efficacy [11, 12]. The broad range of observable half-lives and the fact that mitotane plasma levels between 14 and 20 mg L-1 are considered as target concentration are reasons that underlie the strong recommendation for therapeu- tic drug monitoring (TDM) regardless of the selected dosing regimen [6, 15-17]. Furthermore, it is supposed that mitotane metabolism might be affected by autoinduction [18] and the inter-individual bioavailability of mitotane is highly variable [19, 20]. Therefore, mitotane plasma levels should be con- trolled at the initiation, change, or cessation of the therapy on a regular basis [6, 13]. Depending on the dosing scheme and the individual half-live, a few patients reach target levels within 4-6 weeks, but in the majority of patients, it takes several weeks to months to reach mitotane steady state [21]. Until attaining steady state, the plasma level should be
monitored more frequently in order to manage toxicity, espe- cially for high dosing regimens.
To date, the most common methods applied for the TDM of mitotane comprise a suitable sample preparation (protein pre- cipitation, liquid-liquid extraction, and/or solid-phase extrac- tion) followed by chromatographic separation and UV detec- tion [13, 22-25]. Patient plasma is the matrix used in all of these methods to analyze mitotane concentration, which re- quires periodic physician consultations for sampling. Since the introduction of the dried blood spot in 1963 by Guthrie and Susi [26], many advances in the analysis of dried matrices have been made [27]. One of the more recent milestones is the introduction of sampling devices, which promise to overcome the hematocrit dependence of the sampled blood volume [28-30]. The aim of this work was to test whether the use of one of these volumetric absorptive microsampling devices can replace venous plasma sampling for mitotane TDM. The method was characterized following the recommendations of the European Medicines Agency [31]. After developing and characterizing the method, it was applied to 51 leftover whole blood samples from patients, and mitotane concentrations ob- tained from these DBS samples were compared to those ob- tained from conventional plasma samples.
Materials and methods
Chemicals and reagents
Racemic mitotane (1-(2-chlorophenyl)-1-(4-chlorophenyl)- 2,2-dichlorethan), dicofol (2,2,2-trichloro-1,1-bis(4- chlorophenyl)ethanol), o,p’-DDE, potassium phosphate
CI
CI
CI
CI
CI
CI
H
OH
CI
CI
CI
Mitotane
Dicofol
CI
CI
HO
O
CI
CI
CI
CI
o,p’-DDE
o,p’-DDA
monobasic, ortho-phosphoric acid 85% for HPLC, and aceto- nitrile for HPLC gradient grade were purchased from Sigma- Aldrich Chemie GmbH (Taufkirchen, Germany), and zinc sulfate heptahydrate and methanol for HPLC gradient grade from VWR International S.A.S. (Fontenay-sous-Bois, France). o,p’-DDA was purchased from Alsachim SAS (Illkirch Graffenstaden, France). Other drug substances were obtained from the local drug bank at the Institute for Pharmacy (Würzburg, Germany). Ultrapure water was produced by a water purification system from Merck Millipore (Schwalbach, Germany). Venous blood with EDTA as antico- agulant was obtained from the blood donation service of the Bavarian Red Cross (Muenchen, Germany).
Equipment and chromatographic procedure
The samples were analyzed using an Alliance (Waters Corp, Milford, MA, USA) 2695 HPLC equipped with 995 photodi- ode array detector and column thermostat. A LUNA C8 5 um 150 × 4.6 mm reversed phase column (Phenomenex, Aschaffenburg, Germany) was used as stationary phase. The system was operated in isocratic mode with a flow rate of 1.2 mL min-1 and the column kept at 40 ℃. The autosampler temperature was set to 25 ℃. The mobile phase was com- posed of water containing 5% (v/v) acetonitrile, acetonitrile containing 5% (v/v) water, and 50 mM potassium dihydrogen phosphate buffer pH adjusted to 2.0 with phosphoric acid 85% at a ratio of 20:70:10 (v/v/v). Ultra-violet detection was per- formed at 230 nm using the photodiode array detector. The calculation of the concentrations was performed by peak height. The method run time was 15 min. A mixture of 2- propanol and water (50:50 v/v) was used as needle wash solution.
System control and data acquisition were performed with Empower® (Version 3) software. Calibration curves were cal- culated with the Empower program using a 1/x2 weighted linear least-square regression.
Preparation of stock solutions, calibration curve, and quality control samples
Two independent stock solutions for calibrators and quality control samples were prepared in dimethyl sulfoxide (DMSO) to obtain a final concentration of approx. 20 g L-1. These solutions were stored at - 80 ℃. Eight calibration and four quality control working solutions were prepared from the stock solutions by serial dilution with DMSO. Twenty-five microliters of working solution, or DMSO for the blank, was then spiked in 975 uL venous blood (hematocrit adjusted to 0.45) to achieve concentrations of 50, 30, 20, 10, 5, 2.5, 1.5, and 1 mg L-1 for the calibrators, and 45, 20, 7, and 1.5 mg L-1 for the quality control (QC) samples (QC high, QC medium, QC intermediate, QC low).
After incubation at 37 ℃ at 400 rpm for 30 min, spiked blood samples (calibrators and quality controls) were absorbed by a 20-uL MITRA™M VAMS device (Neoteryx, Torrance, CA, USA) and dried at ambient temperature (with desiccant) for at least 8 h. Silica gel in PE fleece bag (Tyvek 1059B) was used as desiccant (Wisepac Active Packaging Components Co., Ltd., Shanghai, China).
Sample preparation
Dried blood samples, calibration, and quality control stan- dards were stored at room temperature in a closed container (glass or polypropylene bag) containing desiccant until anal- ysis. The tip of the MITRA™M VAMS device was removed and placed in a polypropylene (PP) tube (2 mL) with round bottom (Eppendorf AG, Hamburg, Germany). After addition of 500 µL internal standard solution (mixture of methanol and 2% (m/V) aqueous zinc sulfate heptahydrate solution 4:1 (v/v) containing 2 mg L-1 dicofol), the samples were treated in an ultrasonic bath for 15 min and subsequently shaken for 1 h at 1400 rpm (Thermomixer comfort, Eppendorf AG, Hamburg, Germany). Four hundred fifty microliters of the extracted so- lution was transferred into a second tube and centrifuged for 5 min (4 ℃, 12.000 rcf). Afterwards, 300 uL of the superna- tant was transferred to a HPLC vial with PP insert for small volumes (Agilent Technologies, Waldbronn, Germany). One hundred microliters was finally injected onto the chromato- graphic column. The same extraction procedure was used for liquid blood samples using 20 uL of liquid blood instead of the VAMS device tip.
Method characterization
The method for dried samples was characterized following the recommendations of the European Medicines Agency [31].
The lower limit of quantification (LLOQ) was defined as the concentration of the lowest calibrator with accuracy be- tween 80 and 120% (n=5) and a precision within 20%. The upper limit of quantitation (ULOQ) was defined as the con- centration of the highest calibration level with accuracy be- tween 85 and 115% (n =5) and a precision within 15%.
Precision and accuracy were evaluated with QC-L (1.5 mg L-1), QC-I (7 mg L-1), QC-M (20 mg L-1), and QC-H (45 mg L-1). Within-run accuracy and precision were determined by measuring a calibration curve with five repli- cates of every QC in a single run. Between-run accuracy and precision (% relative standard deviation, %RSD) were assessed by performing a calibration curve with five replicates of each quality control on five different days. The limits for accuracy and precision were ± 15% and ≤15% (expressed as relative standard deviation), respectively.
Stability of mitotane in liquid plasma is very well investi- gated [23, 25, 32]. The stability of standards in liquid blood
stored at 2-8 ℃ was evaluated by comparing 1-week-old standards with freshly prepared ones. Stability of stock solu- tions in DMSO was tested by comparing freshly prepared solutions in DMSO with 6-month-old solutions kept at - 80 °C.
Short-term stability of the analyte in the dried matrix was tested at different storage conditions. Dried samples were kept with 0.5 g desiccant per sample in tightly closed PP plastic bags for 1 week at 37 ℃ (93% r.h.), at room temperature (humidity not monitored), and 2-8 ℃ (humidity not monitored).
Post-preparative stability was assessed by keeping proc- essed samples in their glass vial with PP insert for 24 h in the autosampler at 25 ℃.
Six venous blood samples from six individual donors were analyzed as blanks. The absence of interfering peaks was characterized by blank responses <5% for the internal stan- dard (dicofol) and <20% of the LLOQ for mitotane. Additionally, the main mitotane metabolites o,p’-DDA and o,p’-DDE as well as likely co-medication (metoclopramide, haloperidol, hydrocortisone, fludrocortisone, prednisone, prednisolone, methylprednisolone, and dexamethasone) were investigated for potential interferences.
Carryover was tested by analyzing blank samples directly after the injection of the highest calibrator (50 mg L-1; n = 5). A mean signal <20% of the signal at the LLOQ was defined as the absence of carryover effects.
Quality control samples at four different concentrations (48, 20, 7, and 1.5 mg L-1) were prepared with high hemato- crit blood (adjusted to 0.55), medium hematocrit blood (ad- justed to 0.40), and low hematocrit blood (adjusted to 0.30). Dried blood samples were prepared at least in triplicate and quantified using a calibration curve spiked in blood with an intermediate hematocrit (0.45).
For extraction efficiency, dicofol (IS) and mitotane were diluted in six venous blood samples from six individual do- nors to yield quality control samples at two different concen- trations (MIT 20 mg L-1 and 6.5 mg L-1). Afterwards, they were sampled by MITRATM VAMS device, dried for at least 8 h and extracted. Additionally, dicofol (IS) and mitotane were diluted in a mixture of methanol and 2% (m/V) aqueous zinc sulfate heptahydrate solution 4:1 (v/v) to yield concentrations similar to the dried blood extract, assuming complete extrac- tion of analyte and IS (0.8 and 0.26 mg L-1 for MIT and 2 mg L-1 for DIC, respectively). All samples were measured using the described LC method (vide supra). The extraction efficiency was defined as the ratio of peak heights obtained from the processed blood samples (n =4 for each concentra- tion per patient) and the mean peak heights obtained from the solutions in neat solvent (n = 4 for each concentration).
To evaluate the importance of liquid blood incubation time and condition, blood was spiked with mitotane and dried sam- ples were prepared directly or after incubation at 37 ℃,
400 rpm for 30 min at 1.5 and 20 mg L-1. Five replicates for every condition and concentration were prepared. The relative recovery, calculated as IS-corrected peak height ratio (directly prepared divided by after incubation), was reported and interpreted using two-sided student’s t test.
Differences in recovery between spiked and patient blood was assessed by comparison of IS-corrected peak height ob- tained from liquid blood samples compared to dried samples prepared from the same sample at the same time. The relative recovery (dried sample divided by liquid sample) of spiked blood at 1.5 and 20 mg L-1 each in five replicates and of six authentic patient samples was calculated. The results were interpreted using ANOVA and Tukey’s post hoc test.
Paired plasma and venous blood samples
Fifty-one venous whole blood samples of 7.5 mL using EDTA S-Monovette® (Sarstedt AG & Co. KG, Nümbrecht, Germany) were obtained from patients with adrenal cancer undergoing current or previous treatment at mitotane doses ranging from 0.0 to 6.5 g/day. This study was part of the European Network for the Study of Adrenal Tumors (ENSAT) registry, which has been approved by the ethics committee of the University of Würzburg (approval number 86/03 and 88/11). All patients provided written informed con- sent. Whole blood was drawn for use in the current study and an aliquot of plasma sample submitted to the Lysosafe® TDM service provided on behalf of the manufacturer, HRA-Pharma (Paris, France). The plasma concentration was determined by a GC-MS method as reported by HRA-Pharma.
Data analysis
Comparison between mitotane MITRA™M and plasma concen- trations was realized according to the considerations from the microsampling working group of the International Consortium for Innovation and Quality in Pharmaceutical Development [33]. Patients with missing plasma concentra- tions, a hematocrit below 0.3, and a plasma or blood level < LLOQ (1 mg L-1) were excluded from analysis. CDBS was plotted against Cp to unveil any nonlinear relationship be- tween the data. A Bland-Altman plot including limits of agree- ment (mean difference d ± 1.96 SD) was used for the graphical approach to visualize agreement between the actual plasma concentrations and the predicted plasma concentrations based on MITRATM measurements. For this purpose, the difference between the predicted plasma concentration (CPred) and Cp was plotted against the mean of CPred and Cp [34]. CPred Was calculated by slope only, because the y-intercept was not sig- nificantly different from zero. Data analysis was carried out using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA) and R (version 3.4.2) [35] with ‘ggplot2,’ ‘dplyr,’ ‘readxl,’ ‘BlandAltmanLeh,’ ‘mcr,’ and ‘epiR’ packages. The
Pearson’s product moment correlation coefficient r and the concordance correlation coefficient (CCC) was calculated. CCC determines the agreement on a continuous measure ob- tained by two methods and was interpreted according to McBride [36] as follows: >0.99, almost perfect; 0.99-0.95, substantial; 0-0.95, moderate; and < 0.90 poor agreement. A Pearson’s product moment correlation was defined as follows: 0 to 0.3 or 0 to - 0.3, weak; 0.3 to 0.7 or - 0.3 to - 0.7, moderate; and 0.7 to 1 or - 0.7 to - 1, strong.
Results
Method characterization
The chromatographic separation was performed within 15 min with dicofol (IS) and mitotane retention times of about 7.9 and 8.8 min, respectively (Fig. 2). Linear regression (y = ax + b; mean ± SD of a 0.0143 ±0.0009, and b 0.0003 ± 0.0014; n = 5) with y being the ratio of mitotane peak height to dicofol peak height and x being the concentration in milli- gram per liter using 1/x2 as weighting factor fulfilled the de- fined criteria. The coefficient of determination was >0.992 (n=5) and back-calculated calibrator concentrations were within ± 15% (±20% for the LLOQ) of the nominal concen- tration. Precision and accuracy were within the acceptance criteria: At the LLOQ (1 mg L-1), 2.37% and 100.4% respec- tively; at the ULOQ (50 mg L-1), 7.30% and 106.8% respec- tively. The relative standard deviations of quality controls (Table 1) were between 5.5 and 14% for both intra- and inter-day precision. Investigation of accuracy of quality con- trols revealed a deviation of less than 15% from the theoretical
| QC sample | Imprecision (RSD, %) | Accuracy (%) |
|---|---|---|
| Inter-day (n =25) | ||
| QC-L | 14.0 | 107.4 |
| QC-I | 13.8 | 104.9 |
| QC-M | 9.7 | 98.4 |
| QC-H | 9.9 | 102.1 |
| Intra-day (n=5) | ||
| QC-L | 5.5 | 108.1 |
| QC-I | 8.0 | 103.0 |
| QC-M | 9.4 | 104.0 |
| QC-H | 8.0 | 103.3 |
concentration at each tested concentration level (98.4- 108.1%; Table 1).
The analysis of six individual blank samples did not show any interference at the retention time of dicofol (IS) nor at the retention time of mitotane (Fig. 3). Neither did the metabolites o,p’-DDA, o,p’-DDE, nor any of the other investigated drug substances interfere with the peaks due to dicofol (IS) or mitotane.
The analyte and internal standard were stable post- preparation stored for 24 h in the autosampler. Mitotane stock solution was stable for 6 months at - 80 ℃. Mitotane (QC-L, QC-M, and QC-H) was stable in liquid blood for at least 1 week at 2-8 ℃ (accuracy, 103.0-108.1%; precision, 5.5- 9.4%; n = 5 for every QC level). Dried samples with desiccant in a closed PP bag were stable at room temperature and at 2- 8 ℃ for 1 week. At 37 ℃, a substantial amount of the analyte was lost (accuracy, 54.9-72.3%; Table 2).
0,20
0,18
o,p’-DDA
0,16
0,14
0,12
₹ 0,10
0,08
o,p’-DDE
0,06
0,04
DIC
0,02
MIT
0,00
2,00
4,00
6,00
8,00
10,00
12,00
14,00
Minutes
0,010
DIC
0,009
0,008
0,007
0,006
マ
0,005
MIT
0,004
0,003
0,002
0,001
0,000
2,00
4,00
6,00
8,00
10,00
12,00
14,00
Minutes
No carryover effects were observed. There were no peaks present in the blank samples at the retention time of dicofol (IS) nor at the retention time of mitotane when the blank samples were injected directly after the highest calibrator.
Samples prepared from blood with high, medium, or low hematocrit were within the general acceptance criteria for quality control samples (accuracy, 89.8-113.0%; precision, 4.5-10.1%; Table 3).
Values obtained for analyte extraction efficiency of sam- ples with high mitotane concentration ranged from 67.0 to 71.5% for mitotane (RSD 5.1-12.7%, n =4 per patient) and from 98.2 to 99.8% for the internal standard dicofol (Table 4).
| Storage temperature | QC sample | Accuracy (%) | RSD (%) |
|---|---|---|---|
| Room temperature | QC-L | 109.9 | 4.4 |
| QC-I | 110.1 | 7.0 | |
| QC-M | 113.9 | 5.9 | |
| QC-H | 102.7 | 5.1 | |
| 2-8 ℃ | QC-L | 101.3 | 13.1 |
| QC-I | 92.2 | 9.2 | |
| QC-M | 91.4 | 1.3 | |
| QC-H | 112.0 | 7.1 | |
| 37 °℃ | QC-L | 54.9 | 10.7 |
| QC-I | 64.5 | 5.3 | |
| QC-M | 63.6 | 6.8 | |
| QC-H | 72.3 | 25.3 |
Values obtained for analyte extraction efficiency of samples with low mitotane concentration ranged from 84.1 to 94.0% for mitotane (RSD 5.2-9.0%, n = 4 per patient) and from 99.7 to 100.7% for the internal standard dicofol (Table 5).
The relative recoveries of directly sampled VAMS vs. blood samples incubated at 37 ℃ for 30 min prior to sampling were 94.8 and 106.2% for the 20 mg L-1 and the 1.5 mg L-1 sample, respectively. Differences in peak height ratio mitotane/IS between the groups were not statistically signifi- cant (p=0.385 and 0.392).
| Hematocrit | QC sample | Accuracy (%) | RSD (%) |
|---|---|---|---|
| Low (0.30)ª | QC-H | 104.2 | 6.0 |
| QC-M | 107.3 | 8.7 | |
| QC-I | 107.2 | 9.1 | |
| QC-L | 109.4 | 4.5 | |
| Medium (0.40)b | QC-H | 102.8 | 7.6 |
| QC-M | 90.6 | 8.0 | |
| QC-I | 89.8 | 4.9 | |
| QC-L | 113.0 | 7.3 | |
| High (0.55)b | QC-H | 91.0 | 5.9 |
| QC-M | 92.6 | 7.0 | |
| QC-I | 90.1 | 6.4 | |
| QC-L | 105.3 | 10.1 |
ª n = 3 replicates per concentration level
b n = 5 replicates per concentration level
| Sample | Mitotane | Dicofol | ||
|---|---|---|---|---|
| EE (%) | RSD (%) | EE (%) | RSD (%) | |
| Patient no. 1 | 67.4 | 5.1 | 99.2 | 0.6 |
| Patient no. 2 | 67.0 | 4.6 | 99.8 | 0.4 |
| Patient no. 3 | 66.6 | 12.7 | 99.0 | 0.7 |
| Patient no. 4 | 71.5 | 9.8 | 98.8 | 0.5 |
| Patient no. 5 | 70.6 | 8.1 | 98.2 | 1.4 |
| Patient no. 6 | 70.4 | 8.1 | 98.3 | 1.2 |
| Average | 68.9 | 98.9 | ||
The relative recoveries of dried samples compared to liquid blood samples were 64.7 and 66.3% for the spiked samples at 20 mg L-1 and 1.5 mg L-1, respectively. Mean relative recov- ery of six individuals was 77.3 ±4.8% (see Fig. 4). The rela- tive recovery in the patient samples was significantly in- creased (patients vs. high spiked concentration, p = 0.013; pa- tients vs. low spiked concentration, p=0.029), whereas no difference between high vs. low spiked concentration was detected (p=0.909).
Correlation between paired plasma and venous blood samples
Four of the venous blood samples had a concentration below the defined LLOQ, and in four additional samples, hematocrit level was below the defined hematocrit range. Hence, these samples were excluded from analysis. Mean plasma concen- tration of the evaluated patient samples was 13.73 mg L-1 (SD ±5.85) and mean whole blood concentration was 9.19 mg L-1 (SD±4.41). Mean hematocrit was 0.37 and the mean blood to plasma ratio was about 66.9% (SD ±15.2) (Table 6; for additional data, see Electronic Supplementary
| Sample | Mitotane | Dicofol | ||
|---|---|---|---|---|
| EE (%) | RSD (%) | EE (%) | RSD (%) | |
| Patient no. 1 | 85.1 | 5.2 | 99.7 | 0.1 |
| Patient no. 2 | 84.1 | 5.2 | 100.4 | 0.2 |
| Patient no. 3 | 86.5 | 6.3 | 100.7 | 0.4 |
| Patient no. 4 | 94.0 | 9.0 | 99.9 | 0.6 |
| Patient no. 5 | 88.6 | 6.0 | 99.4 | 1.2 |
| Patient no. 6 | 86.1 | 8.4 | 100.4 | 0.9 |
| Average | 87.4 | 100.1 | ||
Material (ESM) Table S1). Furthermore, Table S1 gives data on the difference (%) between patient plasma concentrations and the predicted plasma concentrations, which was - 0.81 ± 22.97%. Agreement between CDBS and Cp was poor (r= 0.87, p<0.0001; CCC=0.60). The slope of the Passing-Bablok regression (CDBS VS. Cp) was 0.72 (Fig. 5). Bland-Altman plots including limits of agreement are shown in Fig. 6.
Furthermore, covariates like hematocrit and lipid values (triglycerides, HDL, LDL, and cholesterol levels) were eval- uated, as mitotane might alter lipid metabolism of the patients. It has been known for many years that mitotane induces high LDL cholesterol, but also HDL cholesterol and sometimes triglyceride concentrations [37]. The actual mechanism under- lying these lipid abnormalities is unknown. There was no correlation between the triglyceride level of the patient and the concentration of mitotane in whole blood as well as in plasma. The same results were obtained for HDL, whereas a negative correlation between LDL values and plasma concen- tration was observed. However, due to multiple testing and the small sample size, this finding should be considered carefully. No correlation was observed between hematocrit or lipid values, except for LDL, and the measured mitotane concen- trations in dried blood or in plasma (Table 7; for graphical representation, see ESM Fig. S1). In addition, it seems that hematocrit values have no impact on the partitioning of mitotane between blood and plasma in the evaluated hemato- crit range. However, patients with pathologically low hemat- ocrit level had a high blood to plasma partition ratio showing a moderate correlation. After exclusion of those extreme values, for which the assay was not tested, there was no significant correlation left between blood to plasma ratio and hematocrit (for graphical representation, see ESM Fig. S2).
Discussion
Since the first methods for the monitoring of mitotane were published using gas chromatography in the late 1980s of the last century [12, 32, 38], modern instrumentation has led to a decrease in the volume of matrix needed for the determination. Compared with other methods, the current procedure uses a very low volume of venous blood yielding a dried blood sam- ple (20 µL compared to at least 100 uL plasma) and hence microsampling may represent a novel and more convenient way to monitor mitotane in patients who live far away from the treating cancer center. The MITRA™M VAMS, an approved in vitro diagnostic (IVD), was selected for the present study because of its status as IVD and the favorable handling char- acteristics. The device is built from a small plastic stick with an attached sphere of absorption agent on the tip of the stick. This sphere turns red as soon as it gets in contact with capillary blood and does not collect more than the intended volume of 10, 20, or 30 µL when handled correctly. After drying, the
Relative Recovery: VAMS/liquid blood
p = 0.013
p = 0.029
0.8
0.7
0.6
Spiked low
Spiked high
Patients
samples are stable for at least 1 week when shipped at ambient temperature in sealed PP plastic bags containing desiccant (see ESM Fig. S3).
However, in the summer season or in regions with hot climate, the samples should be shipped on ice to reduce loss of analyte due to evaporation (vapor pressure of mitotane is about 0.26 mPa at 30 ℃ [39]). Another reason for the reduced accuracy after storage at elevated temperature could be due to aging of the sample and related extractability issues not nec- essarily limited to mitotane [40]. However, since samples stored at room temperature and at 2-8 ℃ did not show a decrease in accuracy, unlike described by Xie et al. [40], the loss of analyte due to evaporation seems more likely.
Although the samples prepared directly after spiking yielded the same results, incubation of blood at 37 ℃ for 30 min prior to VAMS sampling was adopted to better reflect the situation in vivo. Furthermore, if the method is expanded to capillary blood, warming of the matrix prior to sampling for CR and QC samples could be crucial to simulate the same viscosity and density of the absorbed blood.
The recovery was comparatively high and no individual influence on extraction efficiency was found in spiked patient samples, expressed by a low RSD. Nevertheless, the recovery from high and low concentration was substantially different (68.9 vs. 87.4%). Since linearity was not an issue,
concentration-dependent effects, like saturation of the extrac- tion medium, were unlikely. An explanation could be related to the batch to batch variability of the VAMS device itself. During the experiments, two different batches of devices were used and the “calculated average blood wicking volume” (see CoA) varied as much as from 20.8 to 22.5 µL from batch to batch. The recovery from low concentration was carried out with the latter batch, whereas all the other experiments were conducted using the first batch of VAMS. This should also be considered when measuring patient samples. An in spec var- iation of wicked blood volume of more than 10% from the nominal value could result in substantial bias, when the VAMS used for calibration comes from a different batch than the VAMS device used for collecting patient samples.
The relative recovery comparing liquid and dried blood of QC low and QC medium supports a recovery of about 70%, assuming a near 100% recovery from liquid samples. Mean relative recovery of patient samples was increased compared to spiked blood. Although the difference was not more than 15%, a substantial amount of variability could be attributed to the difference between spiked and authentic blood samples. As the samples came from critically ill patients, influence of plasma albumin or partial hemolysis cannot be ruled out and the matter needs further attention. In particular, additional in- vestigations are needed whether or not the discrepancy in
| Plasma concentration [mg L~]] | Whole blood concentration [mg L-1] | Hematocrit | Ratio blood to plasma [%] | |
|---|---|---|---|---|
| n | 41 | 41 | 41 | 41 |
| Mean (± SD) | 13.73 (±5.85) | 9.19 (±4.41) | 0.37 (±0.04) | 66.85 (±15.24) |
| Median | 14.4 | 9.54 | 0.37 | 65.6 |
| Range | 1.40-31.50 | 1.07-21.02 | 0.29-0.44 | 30.64-96.35 |
30
Slope
0.72 (0.64-0.96)
Intercept -0.88 (-3.34-0.18)
VAMS Concentration [mg/L]
25
20
15
10
..
5
Pearson’s r = 0.867
0
0
5
10
15
20
25
30
Plasma Concentration [mg/L]
relative recovery is limited to mitotane and the used method of extraction or if this is a general problem of the VAMS device. However, not the full extent of variation could be attributed to the altered recovery since, in general, the recovery seems to be increased, which does not explain the poor correlation of blood and plasma concentrations. Velghe et al. [41] recently summarized advances in DBS analysis avoiding hematocrit bias and presented other devices for collecting dried samples. Future studies investigating the differences in recovery of spiked vs. authentic samples comparing all available devices are needed to collect evidence whether or not the presented findings are device related, and if so, strategies to cope with the problem will be required.
The blood to plasma ratio found was below 1, indicating that MIT concentration in erythrocytes is below the plasma concentration. This is in concordance with earlier findings in small samples (n = 1 and n = 6) where the mitotane concentra- tion in red blood cells reached 18 to 25% of the plasma con- centration [24, 25]. There is a moderate correlation between CDBS and Cp but the prediction of plasma levels based on Mitra™M measurements is not reliable due to high variability of blood to plasma ratio. As illustrated in Bland-Altman plot (Fig. 6a), there is no obvious relation between the differences of predicted vs. actual plasma concentrations and the mean of predicted vs. actual plasma concentrations. The 95% confi- dence interval for the predicted plasma concentration ranges from - 7.6 to 6.6 mg L-1, which would be unacceptable for clinical purpose. Figure 6b also underlines the large propor- tion of samples with a clinically unacceptable bias of more than ± 15% when plasma concentrations were predicted from DBS concentrations.
In conclusion, a nonlinear model may be necessary to relate Mitra™M and plasma concentrations. However, stan- dard deviation for blood to plasma ratio is high (0.67 ±
0.15). Analytical imprecisions of both methods and other unidentified sources of variance might contribute to the observed variation. Additionally, the existence of further covariates, which might lead to a higher binding of mitotane to the erythrocytes in some of the patients, could be another source of uncertainty. For two patients, three subsequent plasma and blood concentration levels were available. Interestingly, in those two patients, plasma and blood levels seem to converge over time after dosage adjustment (see ESM Fig. S4). The hypothesis that blood levels might respond more slowly to dose adjustments (due to partitioning and/or erythrocyte life cycle) com- pared to plasma concentrations needs to be evaluated in a prospective trial since multiple measurements were only available for these two individuals. Although actual rec- ommendations for the target range refer to plasma concen- trations, it is not evident that plasma is the most suitable matrix for the therapeutic drug monitoring of mitotane. The putatively slower reacting blood concentrations might correlate better with adverse reactions or drug response.
In a recently published work by Velghe and Stove [42], the application of VAMS to measure anti-epileptic drugs on left- over patient samples was described. They also demonstrated a high %RSD on the observed blood to plasma ratios, which was between 14 and 20%. The authors concluded that there is a significant level of uncertainty when trying to calculate se- rum concentrations from VAMS concentrations with an aver- age conversion coefficient. Thus, it should be considered if there is even a need and a possibility to convert blood levels to plasma concentrations and/or how to establish reference ranges in dried blood. Therefore, future studies analyzing paired patient samples from capillary blood, liquid blood, and plasma in steady state are needed to satisfactorily answer these questions.
a
+ 1.96 sd = 6.6
5
Difference between actual and predicted plasma concentration [mg/L]
0
-5
- 1.96 sd = - 7.6
0
10
20
30
Mean of actual and predicted plasma concentration [mg/L]
b
+ 1.96 sd = 57.4 %
Relative difference between actual and predicted plasma concentration [%]
50
+ 15 %
0
- 15 %
-50
- 1.96 sd = - 54.3 %
0
10
20
30
Mean of actual and predicted plasma concentration [mg/L]
Fig. 6 a Bland-Altman plot comparing results obtained from MITRA™M tips and plasma samples. The solid line represents the mean difference between predicted plasma concentration based on VAMS measurement and the actual plasma concentration. The dashed lines represent the upper and lower limit of agreement (±1.96 SD). b Bland-Altman plot using relative difference vs. mean concentrations. The clinically acceptable limits of ± 15% based on acceptable imprecision of a bioanalytical assay are represented by the dotted lines
Conclusion
In this work, a new method for the quantification of mitotane from dried blood samples is described and the applicability of the developed method using VAMS on leftover real-life pa- tient samples could be demonstrated. In theory, this technique could enable patients and clinicians to perform therapeutic drug monitoring of mitotane in a more convenient way. Patients could collect the samples and send them to the labo- ratory themselves; thus, there would be no need for regular consultation at the hospital or physician’s office for blood or plasma collection purposes, which might reduce treatment costs. Especially during the initial phase of mitotane treat- ment, more frequent TDM than current standard of care is desirable. More intense sampling could facilitate the progno- sis whether a patient will reach satisfactory mitotane concen- trations in an acceptable period of time. In combination with model-based simulation, this new method could be a way to decide whether a patient should stay on mitotane treatment or not in order to minimize the number of patients receiving this highly toxic substance while having only minimal chances of reaching therapeutic plasma levels. However, reference ranges in capillary blood would have to be established for a reliable interpretation of the observed concentrations since a simple conversion does not seem to be possible. As of today, thera- peutic drug monitoring of mitotane using VAMS is not feasi- ble, because there are too many unanswered questions. Further studies investigating the sources of variation and the clinical implications associated are necessary until the present- ed concept could become standard in routine care.
| Plasma | Blood | |||
|---|---|---|---|---|
| Correlation | p value | Correlation | p value | |
| Hematocrit | 0.071 | 0.657 | 0 | 0.999 |
| Triglycerides | 0.296 | 0.067 | 0.188 | 0.251 |
| HDL | 0.040 | 0.813 | 0.083 | 0.616 |
| LDL | -0.347 | 0.031 | -0.247 | 0.129 |
Funding information This study was funded by Horphag Research (Europe) Ltd. and in part by the German Research Council (DFG, German Research Foundation) Projektnummer: 314061271 - TRR 205 as well as by an individual grant to M.F. (FA466/4-2) and M.K. (KR4371/ 1-2).
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Research involving human participants This study was part of the European Network for the Study of Adrenal Tumors (ENSAT) registry, which has been approved by the ethics committee of the University of Würzburg (approval number 86/03 and 88/11).
Informed consent All patients provided written informed consent.
References
1. Fassnacht M, Terzolo M, Allolio B, Baudin E, Haak H, Berruti A, et al. Combination chemotherapy in advanced adrenocortical carci- noma. N Engl J Med. 2012;366(23):2189-97.
2. Cueto C, Brown JH. The chemical fractionation of an adrenocorticolytic drug. Endocrinology. 1958;62(3):326-33. https://doi.org/10.1210/endo-62-3-326.
3. Cueto C, Brown JH, Richardson AP Jr. Biological studies on an adrenocorticolytic agent and the isolation of the active components. Endocrinology. 1958;62(3):334-9. https://doi.org/10.1210/endo-62-3- 334.
4. Terzolo M, Angeli A, Fassnacht M, Daffara F, Tauchmanova L, Conton PA, et al. Adjuvant mitotane treatment for adrenocortical carcinoma. N Engl J Med. 2007;356(23):2372-80. https://doi.org/ 10.1056/NEJMoa063360.
5. Berruti A, Fassnacht M, Baudin E, Hammer G, Haak H, Leboulleux S, et al. Adjuvant therapy in patients with adrenocortical carcinoma: a position of an international panel. J Clin Oncol. 2010;28(23): e401-2. https://doi.org/10.1200/JCO.2009.27.5958.
6. European Medicines Agency. Lysodren: summary of product char- acteristics. 2009, http://www.ema.europa.eu/docs/en_GB/ document_library/EPAR _-_ Product_Information/human/000521/ WC500047235.pdf. Accessed Apr 2018.
7. Kasperlik-Zaluska AA, Cichocki A. Clinical role of determination of plasma mitotane and its metabolites levels in patients with adre- nal cancer: results of a long-term follow-up. J Exp Ther Oncol. 2005;5(2):125-32.
8. Hermsen IG, Fassnacht M, Terzolo M, Houterman S, den Hartigh J, Leboulleux S, et al. Plasma concentrations of o,p’DDD, o,p’DDA, and o,p’DDE as predictors of tumor response to mitotane in adre- nocortical carcinoma: results of a retrospective ENS@T multicenter study. J Clin Endocrinol Metab. 2011;96(6):1844-51. https://doi. org/10.1210/jc.2010-2676.
9. Hescot S, Paci A, Seck A, Slama A, Viengchareun S, Trabado S, et al. The lack of antitumor effects of o,p’DDA excludes its role as an active metabolite of mitotane for adrenocortical carcinoma treat- ment. Horm Cancer. 2014;5(5):312-23. https://doi.org/10.1007/ s12672-014-0189-7.
10. Megerle F, Herrmann W, Schloetelburg W, Ronchi CL, Pulzer A, Quinkler M, et al. Mitotane monotherapy in patients with advanced adrenocortical carcinoma. J Clin Endocrinol Metab. 2018;103(4): 1686-95. https://doi.org/10.1210/jc.2017-02591.
11. Baudin E, Pellegriti G, Bonnay M, Penfornis A, Laplanche A, Vassal G, et al. Impact of monitoring plasma 1,1- dichlorodiphenildichloroethane (o,p’DDD) levels on the treatment of patients with adrenocortical carcinoma. Cancer. 2001;92(6):1385-92.
12. van Slooten H, Moolenaar AJ, van Seters AP, Smeenk D. The treatment of adrenocortical carcinoma with o,p’-DDD: prognostic implications of serum level monitoring. Eur J Cancer Clin Oncol. 1984;20(1):47-53.
13. Feliu C, Cazaubon Y, Guillemin H, Vautier D, Oget O, Millart H, et al. Therapeutic drug monitoring of mitotane: analytical assay and patient follow-up. Biomed Chromatogr. 2017;31(11):e3993. https:/ doi.org/10.1002/bmc.3993.
14. Daffara F, De Francia S, Reimondo G, Zaggia B, Aroasio E, Porpiglia F, et al. Prospective evaluation of mitotane toxicity in adrenocortical cancer patients treated adjuvantly. Endocr Relat Cancer. 2008;15(4): 1043-53. https://doi.org/10.1677/ERC-08-0103.
15. Terzolo M, Pia A, Berruti A, Osella G, Ali A, Carbone V, et al. Low-dose monitored mitotane treatment achieves the therapeutic range with manageable side effects in patients with adrenocortical cancer. J Clin Endocrinol Metab. 2000;85(6):2234-8. https://doi. org/10.1210/jcem.85.6.6619.
16. Faggiano A, Leboulleux S, Young J, Schlumberger M, Baudin E. Rapidly progressing high o,p’DDD doses shorten the time required to reach the therapeutic threshold with an acceptable tolerance: preliminary results. Clin Endocrinol. 2006;64(1):110-3. https:/ doi.org/10.1111/j.1365-2265.2005.02403.x.
17. Mauclere-Denost S, Leboulleux S, Borget I, Paci A, Young J, Al Ghuzlan A, et al. High-dose mitotane strategy in adrenocortical carcinoma: prospective analysis of plasma mitotane measurement during the first 3 months of follow-up. Eur J Endocrinol. 2012;166(2):261-8. https://doi.org/10.1530/EJE-11-0557.
18. Arshad U, Taubert M, Kurlbaum M, Frechen S, Herterich S, Megerle F, et al. Enzyme autoinduction by mitotane supported by population pharmacokinetic modelling in a large cohort of adreno- cortical carcinoma patients. Eur J Endocrinol. 2018;179(5):287-97.
19. Kerkhofs TM, Derijks LJ, Ettaieb H, Den Hartigh J, Neef K, Gelderblom H, et al. Development of a pharmacokinetic model of mitotane: toward personalized dosing in adrenocortical carcinoma. Ther Drug Monit. 2015;37(1):58-65.
20. Kerkhofs T, Baudin E, Terzolo M, Allolio B, Chadarevian R, Mueller H, et al. Comparison of two mitotane starting dose regi- mens in patients with advanced adrenocortical carcinoma. J Clin Endocrinol Metab. 2013;98(12):4759-67.
21. Fassnacht M, Kroiss M, Allolio B. Update in adrenocortical carci- noma. J Clin Endocrinol Metab. 2013;98(12):4551-64.
22. Andersen A, Kasperlik-Zaluska AA, Warren DJ. Determination of mitotane (o,p-DDD) and its metabolites o,p-DDA and o,p-DDE in plasma by high-performance liquid chromatography. Ther Drug Monit. 1999;21(3):355-9.
23. Garg MB, Sakoff JA, Ackland SP. A simple HPLC method for plasma level monitoring of mitotane and its two main metabolites in adrenocortical cancer patients. J Chromatogr B Anal Technol Biomed Life Sci. 2011;879(23):2201-5. https://doi.org/10.1016/j. jchromb.2011.06.001.
24. De Francia S, Pirro E, Zappia F, De Martino F, Sprio AE, Daffara F, et al. A new simple HPLC method for measuring mitotane and its two principal metabolites tests in animals and mitotane-treated pa- tients. J Chromatogr B Anal Technol Biomed Life Sci. 2006;837(1- 2):69-75. https://doi.org/10.1016/j.jchromb.2006.04.005.
25. Mornar A, Sertic M, Turk N, Nigovic B, Korsic M. Simultaneous analysis of mitotane and its main metabolites in human blood and urine samples by SPE-HPLC technique. Biomed Chromatogr. 2012;26(11):1308-14. https://doi.org/10.1002/bmc.2696.
26. Guthrie R, Susi A. A simple phenylalanine method for detecting phenylketonuria in large populations of newborn infants. Pediatrics. 1963;32:338-43.
27. Enderle Y, Foerster K, Burhenne J. Clinical feasibility of dried blood spots: analytics, validation, and applications. J Pharm Biomed Anal. 2016;130:231-43. https://doi.org/10.1016/j.jpba. 2016.06.026.
28. De Kesel PM, Lambert WE, Stove CP. Does volumetric absorptive microsampling eliminate the hematocrit bias for caffeine and paraxanthine in dried blood samples? A comparative study. Anal Chim Acta. 2015;881:65-73. https://doi.org/10.1016/j.aca.2015.04. 056.
29. Denniff P, Spooner N. Volumetric absorptive microsampling: a dried sample collection technique for quantitative bioanalysis. Anal Chem. 2014;86(16):8489-95. https://doi.org/10.1021/ac5022562.
30. Kok MGM, Fillet M. Volumetric absorptive microsampling: current advances and applications. J Pharm Biomed Anal. 2018;147:288- 96. https://doi.org/10.1016/j.jpba.2017.07.029.
31. European Medicines Agency. Guideline on Bioanalytical Method Validation. 2015, https://www.ema.europa.eu/documents/scientific- guideline/guideline-bioanalytical-method-validation_en.pdf. Accessed Mar 2018.
32. Inouye M, Mio T, Sumino K. Use of GC/MS/SIM for rapid deter- mination of plasma levels of o,p’-DDD, o,p’-DDE and o,p’-DDA. Clin Chim Acta. 1987;170(2-3):305-14.
33. Evans C, Arnold M, Bryan P, Duggan J, James CA, Li W, et al. Implementing dried blood spot sampling for clinical pharmacoki- netic determinations: considerations from the IQ Consortium Microsampling Working Group. AAPS J. 2015;17(2):292-300. https://doi.org/10.1208/s12248-014-9695-3.
34. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1(8476):307-10.
35. R Core Team. R: a language and environment for statistical com- puting. Vienna: R Foundation for Statistical Computing; 2017.
36. McBride GB. A proposal for strength of agreement criteria for Lin’s Concordance Correlation Coefficient. NIWA Client Report; 2005.
37. Shawa H, Deniz F, Bazerbashi H, Hernandez M, Vassilopoulou- Sellin R, Jimenez C, et al. Mitotane-induced hyperlipidemia: a ret- rospective cohort study. Int J Endocrinol. 2013;2013:624962. https://doi.org/10.1155/2013/624962.
38. Benecke R, Vetter B, De Zeeuw RA. Rapid micromethod for the analysis of mitotane and its metabolite in plasma by gas chroma- tography with electron-capture detection. J Chromatogr. 1987;417(2):287-94.
39. U.S. Department of Health and Human Services Agency for Toxic Substances and Disease Registry. Toxicological profile for DDT, DDE, and DDD. 2002, https://www.atsdr.cdc.gov/toxprofiles/ tp35.pdf. Accessed Mar 2018.
40. Xie I, Anderson M, Wang M, Xue L, Breidinger S, Goykhman D, et al. Extractability-mediated stability bias and hematocrit impact: high extraction recovery is critical to feasibility of volumetric ad- sorptive microsampling (VAMS) in regulated bioanalysis. J Pharm Biomed Anal. 2018;156:58-66.
41. Velghe S, Delahaye L, Stove CP. Is the hematocrit still an issue in quantitative dried blood spot analysis? J Pharm Biomed Anal. 2019;163:188-96.
42. Velghe S, Stove CP. Volumetric absorptive microsampling as an alternative tool for therapeutic drug monitoring of first-generation anti-epileptic drugs. Anal Bioanal Chem. 2018;410(9):2331-41.
Publisher’s note Springer Nature remains neutral with regard to jurisdic- tional claims in published maps and institutional affiliations.