J Mass Spectrom 2004; 39: 262-271
Chapter 2.1
Simultaneous quantification of cyclophosphamide, 4-hydroxycyclophosphamide, N,N,N-triethylenethio-
phosphoramide (thiotepa) and N,N,N-triethylene-phosphoramide (tepa) in human plasma by high-performance
liquid chromatography coupled with electrospray ionization tandem mass spectrometry (LC-MS/MS)
Milly E de Jonge, Selma M van Dam, Michel JX Hillebrand, Hilde Rosing,
Alwin DR Huitema, Sjoerd Rodenhuis, Jos H Beijnen
Abstract The alkylating agents cyclophosphamide (CP) and N,N,N-triethylenethiophosphoramide (thiotepa) are often co-administered in high-dose chemotherapy regimens. Since these regimens can be complicated by the occurrence of severe and sometimes life-threatening toxicities, pharmacokinetically guided administration of these compounds, to reduce variability in exposure, may lead to improved tolerability. For rapid dose adaptations during a chemotherapy course, we have developed and validated an assay, using liquid chromatography coupled with electrospray tandem mass spectrometry (LC-MS/MS), for the routine quantification of CP, thiotepa, and their respective active metabolites 4-hydroxycyclophosphamide (4OHCP) and N,N,N-triethylene-phosphoramide (tepa) in plasma. Because of the instability of 4OHCP in plasma, the compound is derivatized with semicarbazide (SCZ) immediately after sample collection and quantified as 4OHCP-SCZ. Sample pretreatment consisted of protein precipitation with a mixture of methanol and acetronitrile using 100 l plasma. Chromatographic separation was performed on an Zorbax Extend C18 column (150 x 2.1 mm ID, particle size 5 m), with a quick gradient using 1 mM ammoniumhydroxide in water and acetonitrile, at a flow rate of 0.4 ml/min. The analytical run time was 10 min. The triple quadrupole mass spectrometer was operating in the positive ion mode and multiple reaction monitoring was used for drug quantification. The method was validated over a concentration range of 200 to 40,000 ng/ml for CP, 50 to 5,000 ng/ml for 4OHCP-SCZ and 5 to 2,500 ng/ml for thiotepa and tepa, using 100 l of human plasma. These dynamic concentration ranges proved to be relevant in daily practice. Hexamethylphosphoramide was used as an internal standard. The coefficients of variation were less than 12% for both intra-day and inter-day precisions for each compound. Mean accuracies were also between the designated limits (15%). This robust and rapid LC-MS/MS assay is now successfully applied for routine therapeutic drug monitoring of CP, thiotepa and their metabolites in our hospital.
Chapter 2.1
84
Introduction High-dose chemotherapy with the alkylating agents cyclophosphamide (CP), carboplatin and thiotepa (CTC) is widely used in the treatment of advanced or metastatic breast, ovarian and testis tumors [1-6]. Because of the haematological stem cell support in this high-dose setting, the doses of the individual compounds can be increased up to 10-fold since bone marrow toxicity is no longer dose-limiting. However, high-dose CTC regimens can be complicated by the occurrence of severe and sometimes life-threatening non-marrow toxicities such as mucositis, veno-occlusive disease, oto- and cardiotoxicity [1-6]. The large interindividual variability of these toxicities may be due to interindividual pharmacokinetic variation of the compounds involved. Studies have shown that exposures (expressed as area under the plasma concentration versus time curve, AUC) of the compounds and their metabolites are related to toxicity [7-13]. Therefore, therapeutic drug monitoring (TDM) may be an important tool to minimize toxicity in the CTC regimen. To make routine and fast TDM in the CTC setting feasible, a fast and robust analytical method is required quantifying the necessary compounds. With this method, analytical results should become available within a few hours to make dose adaptations during a chemotherapy course possible. Since for carboplatin a simple method of analysis is available [14], it was intended to develop a method for simultaneously quantifying CP, thiotepa and their relevant metabolites.
Figure 1. Chemical structures of A) cyclophosphamide, B) 4-hydroxycyclophosphamide, C) thiotepa and D) tepa. CP is an oxazaphosphorine prodrug requiring activation by the cytochrome P450 enzyme system (CYP), to form its pharmacologically active metabolite 4-hydroxycyclophosphamide (4OHCP) (Figure 1A and B). 4OHCP can enter target cells where it is metabolised to form the final cytotoxic agent phosphoramide mustard. Since the 4OHCP level in plasma is an accurate determinant of the alkylating activity of CP [15],
O
O
NP
OH
NCH2
CH2CH2
CH2Cl
Cl
H
N PN
SN N P
N
ON
O
NPN
CH2
CH2CH2
CH2Cl
Cl
HO
A B
C D
Bioanalysis of cyclophosphamide, thiotepa and metabolites
85
it is necessary to quantify both CP and 4OHCP for TDM purposes. 4OHCP is a very unstable compound, with a plasma half-life of only a few minutes [16]. Accurate quantification of 4OHCP therefore requires instantaneous derivatization of the compound at the bedside to form a stable derivative which can subsequently be measured. Several methods for stabilizing 4OHCP have been described [17]. The method using stabilization with the derivatizing agent semicarbazide (SCZ), resulting in formation of 4OHCP-SCZ (Figure 2) [18,19], is robust, easy to perform and therefore feasible in clinical practice. The ethylenimine thiotepa is rapidly metabolized by oxidative desulfuration mediated by CYP to yield its metabolite tepa (Figure 1C and D). Thiotepa and tepa have a similar alkylating activity [20,21] and therefore it is important to quantify both compounds for TDM purposes. An assay for the simultaneous determination of CP, 4OHCP, thiotepa and tepa has not yet been developed. In our laboratory, a gas chromatographic assay was developed for the simultaneous quantification of CP, thiotepa and tepa in plasma [22]. However, sample pretreatment consisted of a labour-intensive liquid/liquid extraction, and in practice the assay appeared not to be very robust. We have also developed an assay for the determination of 4OHCP in plasma (quantified as 4OHCP-SCZ, Figure 2), using high-performance liquid chromatography (HPLC) with UV detection [18]. This assay, however, requires 1000 l of plasma per sample, has a lengthy sample pretreatment procedure and a run time of 20 min. To reduce the turn-around time of the samples, a combined assay with a simple sample pretreatment procedure and a short run time is necessary. This can be achieved by using tandem mass spectrometric (MS/MS) detection with multiple reaction monitoring (MRM). Because of the high selectivity and specificity of this technique, it permits the use of short run times and minimal sample-clean up procedures. Figure 2. Reaction of 4-hydroxycyclophosphamide with semicarbazide resulting in the formation of the semicarbazone derivative of 4-hydroxycyclophosphamide.
O
NP
OH
NCH2
CH2CH2
CH2Cl
Cl
HO
N NH2
O
H
H2N
O
O
NP
N
NCH2
CH2CH2
CH2Cl
Cl
H2 NNH2
O
4-hydroxycyclophosphamide
semicarbazide
semicarbazone derivative of 4-hydroxycyclophosphamide
Chapter 2.1
86
Assays quantifying CP and/or 4OHCP using HPLC with MS detection have been described [23-25]. Sottani et al [24] described the quantification of CP in urine using LC-MS/MS applying liquid/liquid extraction with ethylacetate as sample pretreatment. Baumann et al [23] described a LC-MS method, using single-ion monitoring, for the simultaneous determination of CP, 4OHCP and three other CP metabolites in 450 l human plasma. Methylhydroxylamide was used for derivatizing 4OHCP and a solid-phase extraction was used for sample pretreatment. Sadagopan et al [25] described an LC-MS/MS assay simultaneous determining CP and 4OHCP in mouse plasma. Methylhydroxylamide was used for derivatizing 4OHCP and protein precipitation followed by evaporation and reconstitution of the supernatant was used for sample pretreatment. To our knowledge, quantification of thiotepa and tepa using LC-MS/MS has not been described.
In this paper, we present the development and validation of a rapid, sensitive and specific method for the simultaneous quantification of CP, 4OHCP, thiotepa, and tepa in human plasma using HPLC coupled with electrospray ionization (ESI)MS/MS. The combination of a uniform, simple and above all fast sample pretreatment with a chromatographic system that provides a run time of 10 min allows high sample throughput and has proven very useful for routine TDM. The validation of the method was performed based on the most recent international guidelines for bioanalytical validation [26]. Experimental Chemicals CP, 4-hydroperoxycyclophosphamide (4OOHCP), and all other CP metabolites used for selectivity and specificity tests were a generous gift from Dr. Niemeyer, ASTA Medica, Frankfurt, Germany (purity >95%). Thiotepa (Ledertepa) was obtained from AHP Pharma (Hoofddorp, The Netherlands). Tepa was synthesized at the Faculty of Chemistry, University of Utrecht, according to the method described by Craig and Jackson [27] (purity >98%). The internal standard (IS) hexamethylphosphoramide (HMP) (Figure 3) originated from Sigma (Zwijndrecht, The Netherlands). Acetonitrile and methanol were HPLC-grade reagents and were obtained from Biosolve BV (Valkenswaard, The Netherlands). Potassium dihydrogenphosphate (suprapure grade) was from Merck (Darmstadt, Germany). SCZ hydrochloride (analytical reagent grade) was purchased from Acros (Geel, Belgium). A 2 M solution of SCZ in 50 mM potassium phosphate buffer (pH 7.4) was prepared (stored at 4C for a maximum of 6 months). Distilled water was used throughout the analysis and all other chemicals used were of analytical grade and used without further purification. Drug free human plasma originated from the Central Laboratory of the Netherlands Red Cross Blood Transfusion Service (Amsterdam, the Netherlands).
Bioanalysis of cyclophosphamide, thiotepa and metabolites
87
Figure 3. Chemical structure of the internal standard hexamethylphosphoramide. Chromatographic and mass spectrometric conditions An Agilent Technologies, (Palo Alto, CA, USA) HPLC system was used, consisting of a Model 1100 Series pump and cooled autosampler (10C). Separation was carried out on a Zorbax Extend C18 column (150 x 2.1 mm ID, particle size 5 m; Agilent Technologies) protected with an Agilent Extend C18 narrow-bore guard column (12.5 x 2.1 mm ID, particle size 5 m; Agilent Technologies). A stepwise gradient was used to elute the compounds from the column. At time zero a mixture of 96% eluent A (1 mM ammonium hydroxide in water) and 4% eluent B (100% acetonitrile) was flushed through the column. After 2 min, the fraction of acetonitrile was increased to 25% during 1 min. This mobile phase composition was maintained for 3 min. Subsequently, during 0.1 min, the mobile phase composition was set back at 96% eluent A and helt there for the final 4 min of the run. The flow-rate was 0.4 ml/min. The column outlet was connected directly to the electrospray sample inlet (Sciex, Thornhill, ON, Canada). The source temperature was set at 400C. Ions were created at atmospheric pressure and were transferred to an API 3000 triple-quadrupole mass spectrometer (Sciex). The curtain gas (1.1 ml/min) and the collision-induced dissociation (CID) gas (342 * 1015 molecules/cm2) consisted of nitrogen (grade 5.0) and the nebulizer and turbo gases (1.6 l/min and 7.0 l/min, respectively) were zero air. The electrospray source was operated in the positive ion mode. The electrospray voltage was +2.5 kV and the dwell time was 50 ms with a 5 ms pause between scans. Q1 and Q3 were operating at unit mass resolution. MRM was used for drug quantification. Precursor ions of analytes and IS were determined from spectra obtained during the infusion of standard solutions using an infusion pump connected directly to the electrospray source. As a result of the very soft ionization, provided by the electrospray ion source, only singly charged molecular ions were observed. Each of the precursor ions was subjected to CID to determine the product ions. The transitions of the protonated precursor/product ion pairs that were used for recording the selected-ion mass chromatograms are listed in Table 1. Data were processed by Analyst software (Applied Biosystems/ MDS Sciex, Analyst software version 1.2). Sample collection, pretreatment and processing Whole blood samples were collected in patients receiving high-dose CTC chemotherapy, including CP (6,000 or 4,000 mg/m2 in 4 days), carboplatin (1,600 or 1,067 mg/m2 in 4
N PN
ON
CH3CH3H3C
H3C
H3C CH3
Chapter 2.1
88
days) and thiotepa (480 or 320 mg/m2 in 4 days) [2-6]. Samples were collected in heparin tubes and immediately placed on ice. Plasma was immediately obtained by centrifuging the whole blood sample at 3,500 g for 3 min at 4 C. A 500 l volume of plasma was added to a polypropylene tube containing 50 l of SCZ solution. Samples were whirl-mixed for 10 s, placed in a water-bath at 35 C for 2 h and then stored at -70 C. For analysis, to 100 l of this solution 25 l of the IS solution was added (containing approximately 100 ng/ml HMP in ethanol). After whirl-mixing for 10 s, 300 l of the protein precipitation reagent methanol-acetonitrile (1:1, v/v) was added. The mixture was vortex mixed for 30 s, and automatically mixed (Labinco, Breda, The Netherlands) for 5 min. Subsequently, the samples were centrifuged at 7,000 g for 15 min. A volume of 50 l of supernatant was diluted with 400 l of a 1 mM ammonium hydroxide solution in water and whirl-mixed for 10 s. From this mixture, a volume of 10 l was injected on to the analytical column. Preparation of stock solutions, working solutions and plasma standards Two fresh stock solutions of CP, thiotepa and tepa were prepared independently in ethanol, at concentrations of approximately 5 mg/ml (CP) and 500 g/ml (thiotepa and tepa). Stock solutions of 4OHCP were prepared by dissolving 4OOHCP in water, which decomposes after dissolution into hydrogen peroxide and 4OHCP, resulting in a final concentration of approximately 1 mg/ml. One solution was used to spike the plasma calibration samples and the other was used to prepare the quality control (QC) samples. Stock solutions were diluted further with water and the solutions of each analyte were added together to obtain working solutions. After spiking 1950 l of drug-free human plasma with 50 l of these working solutions, the following concentrations in plasma were obtained: 200, 400, 1,000, 4,000, 10,000, 20,000 and 40,000 ng/ml for CP, 50, 100, 250, 500, 1,000, 2,500, 5,000 ng/ml for 4OHCP and 5, 10, 50, 250, 500, 1,000 and 2,500 ng/ml for thiotepa and tepa. The choice for these calibration ranges was based on the concentrations achieved in patients in daily practice of TDM. QC samples at low, medium and high concentration levels were prepared in a similar way. To all of these mixtures 200 l of SCZ solution was added. The samples were whirl-mixed for 10 s, placed in a water-bath at 35 C for 2 h and subsequently stored at -70 C until analysis. After thawing, 100 l volumes of each calibration sample were processed as described for the blood samples. Validation procedures The validation of the assay was based on the FDA guidelines for Bioanalytical Method Validation [26]. Linearity Calibration standards were prepared and analysed in duplicate in three independent runs. Calibration curves (area ratio to the IS versus nominal analyte concentration) were fitted
Bioanalysis of cyclophosphamide, thiotepa and metabolites
89
by least-squares linear regression without weighting and using 1/x and 1/x2 (x=concentration) as weighting factors. In order to establish the best weighting factor, back-calculated concentrations were determined. The model with the lowest total bias and the most constant bias across the concentration range was considered to be the best fit. To assess linearity, deviations of the mean calculated concentrations over three runs should be within 15% from nominal concentrations for the non-zero calibration standards. At the lower limited of quantitation (LLQ) level a deviation of 20% was permitted. Accuracy and precision Inter-assay accuracy and intra- and inter-assay precisions of the method were determined by assaying five replicates of each of the QC samples with analyte concentrations around the LLQ, and in the low, medium and high concentration ranges in three separate analytical runs.
Inter-assay accuracy was determined as the percentage difference between the mean concentration after three analytical runs and the nominal concentration. Accuracy should be within 15% except at the LLQ concentration, where it should be less than 20%.
The coefficient of variation (CV) was used as a measure for intra- and inter-assay precision. Precision should not exceed 15% CV except for the LLQ where it should not exceed 20% CV.
To validate the accuracy and precision of the analysis of samples originally above the upper limit of quantification, an extra QC sample containing analyte concentrations two times higher than the concentrations of the highest QC sample, was diluted twofold with drug-free human plasma prior to analysis. Recovery The absolute recovery of the analytes after protein precipitation was determined by comparing the analytical results for plasma QC samples at the three concentration levels to a corresponding set of spiked plasma extracts (containing 100% of the theoretical concentration). Three replicates were analysed at each concentration level. Recovery should be consistent, precise and reproducible.
To determine the amount of ion suppression, the analytical results for spiked plasma extracts were compared with those for a corresponding set of diluted working solutions (in triplicate, at three concentration levels). Selectivity and specificity Out of six batches control human plasma, double blank samples (no analyte, no IS), blank samples (no analyte, with IS) and LLQ samples were prepared, processed and analysed to determine whether endogenous plasma constituents interfered with the assay. Interference may occur when co-eluting endogenous compounds produce ions at the same m/z values that are used to monitor the analyte and IS.
Chapter 2.1
90
To investigate the potential interference of other metabolites of CP and co-medication with the quantification of the analytes, the metabolites and co-medicated drugs were added to LLQ samples at therapeutic relevant concentrations. The samples were then processed and assayed according to the described method. The following metabolites were tested: Carboxyphosphamide (400 g/ml), 2-dechloroethylcyclophosphamide (175 g/ml), ketophosphamide (450 g/ml) and phosphoramide mustard (70 g/ml). The following drugs were tested: Aciclovir (100 g/ml), amphotericine B (20 g/ml), caffeine (1000 g/ml), carboplatin (1000 g/ml), ciprofloxacine (15 g/ml), dexamethasone (200 g/ml), fluconazole (60 g/ml), granisetrone (150 g/ml), itraconazole (15 g/ml), lorazepam (5 g/ml), mesna (150 g/ml), methoclopramine (1,5 g/ml), morphine (1000 g/ml), ondansetrone (150 g/ml), paracetamol (1000 g/ml), ranitidine (5 g/ml), roxithromycine (150 g/ml) and temazepam (15 g/ml). Areas of peaks co-eluting with the analyte peaks should not exceed 20% of the area at the LLQ level. At the IS retention time the interference should not exceed 5% of the IS peak area. Stability The stability of the analytes was investigated under various conditions including the stability in plasma, in the supernatant and in the final extract. The analytes were considered to be stable when 80-120% of the initial concentration was found.
Stability of freshly prepared QC samples, containing thiotepa, tepa, CP and 4OHCP-SCZ at a low, medium and high concentration level, was assessed after three freeze (-70 C)-thaw cycles. The concentrations of the analytes were related to the initial concentration as determined for the samples that were freshly prepared and processed immediately. The long-term stability of thiotepa, tepa, CP and 4OHCP-SCZ in plasma at -70 C was studied by re-analysing previously measured patient samples at different concentrations.
The processed sample stability of all analytes in the supernatant after storage for 10 days at 4 C was assessed at two different concentration levels. Also, the stability in the final extract was studied after 10 days storage at 4 C at two different concentration levels. The measured concentrations of the analytes in the stored processed samples were related to the measured concentrations of the same QC samples immediately after processing.
Results and discussion
Mass spectrometry The most sensitive mass transitions for MRM analysis are depicted in Table 1. Representative chromatograms of control human plasma and a QC sample are presented in Figure 4. In Figure 5, tandem mass spectra of the four compounds are shown. For CP, the most abundant fragment in the product ion mass spectrum was seen
Bioanalysis of cyclophosphamide, thiotepa and metabolites
91
at m/z ratio of 140. This product ion is a result of cleavage of the nitrogen-phosphorus bond connecting the ring structure with the amine, resulting in the loss of di(2-chloroethyl)amine. For 4OHCP-SCZ, the most abundant fragment at m/z 221 was probably the result of cleavage of the carbon-oxygen group in the middle of the molecule resulting in the loss of phosphoramide mustard. Detected product ions for thiotepa and tepa are both the result of loss of azacyclopropane (m/z 43) resulting in products with m/z 147 and 131, respectively. Formation of the product ion of HMP at m/z 135 is based on the same mechanism with concomitant loss of dimethylamine (m/z 45). The MS/MS settings were adjusted to maximize the response of each of the precursor-product ion combinations. Table 1. Retention times, capacity factors (k) and monitored transitions of the analytes and internal standard.
Compound Retention time (min)
k Precursor (m/z) Product (m/z)
CP 9.2 8.2 261 140 4OHCP-SCZ 7.2 6.2 334 221 TT 7.7 6.7 190 147 T 2.4 1.4 174 131 HMP 7.0 6.0 180 135 CP= cyclophosphamide; 4OHCP-SCZ= semicarbazone derivative of 4-hydroxycyclophosphamide; TT= thiotepa; T= tepa; HMP= hexamethylphosphoramide.
Sample pretreatment During the development of the sample pretreatment procedure we focussed on non-labor-intensive methods to accelerate sample processing. Moreover, because of the potential volatility of thiotepa, but especially tepa, a heating or evaporating step was not preferred in sample pretreatment. A protein precipitation method was then validated using a mixture of organic solutions (methanol-acetonitrile (1:1, v/v)) as precipitation reagent.
The kinetics of the derivatization procedure of 4OHCP in plasma were investigated in more detail. Applying a 2M SCZ solution for the reaction [18], we found that the derivatization reaction was only completed after 5 h at room temperature (20 C) or after 1-2 h at 35 C at all low, medium (Figure 6) and high concentration levels of 4OHCP. Cooling the mixture to 4 C during derivatization slows down the reaction dramatically and heating the mixture more than 35 C does not increase the rate of the derivatization reaction (data not shown). The heating process at 35 C appeared not to influence the amount of measured thiotepa, tepa and CP in the mixture. Therefore, it was concluded that the most optimal derivatization conditions were at 35 C for 2 h, which is feasible in clinical TDM practice.
Chapter 2.1
92
05000
1000015000200002500030000
0 1 2 3 4 5 6 7 8 9 10
Time (min)
Inte
nsity
(cps
)
0
5000
10000
15000
20000
25000
0 1 2 3 4 5 6 7 8 9 10
Time (min)
Inte
nsity
(cps
)
0
20000
40000
60000
80000
100000
0 1 2 3 4 5 6 7 8 9 10
Time (min)
Inte
nsity
(cps
)
0
300
600
900
1200
1500
0 1 2 3 4 5 6 7 8 9 10
Time (min)
Inte
nsity
(cps
)
0
200
400
600
800
1000
0 1 2 3 4 5 6 7 8 9 10
Time (min)
Inte
nsity
(cps
)
0
10000
20000
30000
40000
50000
0 1 2 3 4 5 6 7 8 9 10
Time (min)
Inte
nsity
(cps
)A B C D E F
TIC double blank XIC cyclophosphamide (m/z 261) XIC 4-hydroxycyclophosphamide-semicarbazide (m/z 334) XIC thiotepa (m/z 190) XIC tepa (m/z 174) XIC hexamethylphosphoramide (m/z 180)
Figure 4. Total ion-chromatogram (TIC: A) of a processed control plasma sample and the extracted single ion chromatograms (XIC: B to F) of a processed quality control sample at the medium concentration level. Cyclophosphamide (B, 4.03 g/ml); semicarbazone derivative of 4-hydroxycyclophosphamide (C, 1.01 ng/ml); thiotepa (D, 265 ng/ml); tepa (E, 251 ng/ml); hexamethylphosphoramide (F, 25.0 ng/ml).
Bioanalysis of cyclophosphamide, thiotepa and metabolites
93
0
50000
100000
150000
200000
250000
300000
350000
0 50 100 150 200 250 300
m/z
Inte
nsity
(cps
)
m/z=261
m/z=140
Figure 5. MS/MS spectra of cyclophosphamide (A, m/z 261) semicarbazone derivative of 4-hydroxycyclophosphamide (B, m/z 334) thiotepa (C, m/z 190) and tepa (D, m/z 174).
0
2000000
4000000
6000000
8000000
10000000
12000000
0 50 100 150 200 250 300 350 400
m/z
Inte
nsity
(cps
)
m/z 334
m/z=221
0
200000
400000
600000
800000
1000000
1200000
0 20 40 60 80 100 120 140 160 180 200
m/z
Inte
nsity
(cps
)
m/z=190
m/z=147
0
500000
1000000
1500000
2000000
2500000
3000000
3500000
0 20 40 60 80 100 120 140 160 180 200
m/z
Inte
nsity
(cps
)
m/z=174
m/z=131
O
NPN
Cl CH2 CH2
H
CH2CH2Cl
O
O
NPN
Cl CH2 CH2
H2O
CH2CH2Cl
N N
ONH2
P N
N
N
S
O
N
N
NP
A B C D
m/z=140
m/z=221
m/z=147
m/z=131
Chapter 2.1
94
Figure 6. Percentage of the formation of the semicarbazone derivative of 4-hydroxycyclophos-phamide (4OHCP-SCZ) versus time in a quality control sample at medium concentration level at 20 C (!) and 35 C (!) (taking the maximum measured value of 4OHCP-SCZ as 100%). Chromatography HPLC separation of the four analytes was performed under basic conditions using an eluent composed of aqueous 1 mM ammonium hydroxide and acetonitrile together with a column containing a base-stable stationary phase (the pH of the aqueous component is approximately 10). This approach was chosen since it resulted in excellent chromatography for thiotepa and tepa compared to neutral or acidic conditions. Moreover, thiotepa and tepa appeared to be stable at pH 10 [28-30]. The most hydrophilic compound, tepa, could be retained on the stationary phase when the aqueous content of the mixture was high in the first 2 min of the run (96% ammonium hydroxide and 4% acetonitrile). Thereafter, the fraction of acetonitrile was increased to 25% and in the final 4 min of the run the initial eluent composition was again used (column conditioning period). This chromatographic system resulted in adequate and reproducible retention for tepa to separate it from endogenous components. All other analytes were eluted from the column in the column conditioning period due to the extended lag-time of the HPLC system used. The run time could be limited to 10 min (Figure 4). Validation procedures The assay was linear over the validated concentration ranges of 200-40,000 ng/ml for CP, 50-5,000 ng/ml for 4OHCP and 5-2,500 ng/ml for thiotepa and tepa. The best fit for the calibration curves was obtained by using a weighting factor of 1/concentration2 for CP and tepa and 1/concentration for 4OHCP and thiotepa. The deviations from the nominal concentration were
Bioanalysis of cyclophosphamide, thiotepa and metabolites
95
similarities with thiotepa and tepa, HMP was chosen as IS. HMP also proved useful as IS for the quantification of CP and 4OHCP. Its structure is shown in Figure 3. The intra- and inter-assay performance data are presented in Table 2. Accuracies were within 3.9% for the LLQ and within 12.4% for the other concentrations. The inter-assay precision, expressed as CV, was
Chapter 2.1
96
Selectivity and specificity MRM chromatograms of six batches of control human plasma contained no endogenous peaks co-eluting with any of the analytes. Figure 4A shows a representative TIC of a control human plasma sample. LLQ samples, prepared in these six batches of human plasma, could be quantified within the required 20% deviation from the nominal concentration (data not shown). No chromatographic interferences were found from the CP metabolites and co-medicated drugs tested in control human plasma; LLQ samples spiked with the tested drugs and CP metabolites could be quantified within the required 20% deviation. Stability Thiotepa, tepa, CP and 4OHCP-SCZ appeared to be stable after three freeze-thaw cycles (Table 4). Moreover, the stability of the analytes in the supernatant and final extract at 4 C is guaranteed for at least ten days (Table 4). The long-term stability of thiotepa, tepa, CP and 4OHCP-SCZ in plasma was established in previously analysed patient samples stored at -70 C that were 2, 5 and 7 months old. All measured concentrations were within 80-120% limits (data not shown). More stability data for the four compounds at various conditions have been published previously by our group [18,22]. Table 3. Mean protein precipitation recovery and CV% for the analytes and internal standard after three workups.
Compound Nominal concentration
(ng/ml)
Recovery (%)
CV
(%)
CP 402.90 100.00 0.741 4,029.00 92.00 8.570 20,147.00 97.00 4.980TT 10.22 96.00 8.370 255.50 86.00 13.800 1,022.00 97.00 11.800T 10.03 105.00 6.450 250.70 95.00 4.940 1,003.00 95.00 2.260HMP 100.00 99.00 4.200CP= cyclophosphamide; TT= thiotepa; T= tepa; HMP= hexamethylphosphoramide.
Bioanalysis of cyclophosphamide, thiotepa and metabolites
97
Table 4. Stability of the analytes. Compound Matrix Conditions Nominal conc.
(ng/ml) Deviation
(%) CV (%)
No. of replicates
CP Final extract 4 C, 10 days 402.90 -0.520 0.129 3 20,150.00 5.150 4.050 3 Supernatant 4 C, 10 days 402.90 -0.074 1.860 3 20,150.00 8.140 3.190 3 Plasma 3 freeze (-70 C)-
thaw cycles 402.90 0.099 5.840 3
4,029.00 5.310 3.970 3 20,150.00 4.570 0.725 3 4OHCP-SCZ Final extract 4 C, 10 days 110.60 8.250 10.400 3 2,764.00 -10.700 4.190 3 Supernatant 4 C, 10 days 110.60 10.100 8.010 3 2,764.00 2.480 4.960 3 Plasma 3 freeze (-70 C)-
thaw cycles 110.60 1.000 6.600 3
552.80 -9.850 5.120 3 2,764.00 -8.100 4.150 3 TT Final extract 4 C, 10 days 10.22 -1.200 1.050 3 1,022.00 -7.950 7.090 3 Supernatant 4 C, 10 days 10.22 -6.910 6.800 3 1,022.00 -12.900 1.890 3 Plasma 3 freeze (-70 C)-
thaw cycles 10.22 -4.570 2.210 3
255.50 3.330 3.470 3 1,022.00 9.260 0.517 3 T Final extract 4 C, 10 days 10.03 3.500 4.660 3 1,003.00 7.960 1.530 3 Supernatant 4 C, 10 days 10.03 10.800 6.800 3 1,003.00 9.240 3.530 3 Plasma 3 freeze (-70 C)-
thaw cycles 10.03 -5.770 6.410 3
250.70 8.250 1.850 3 1,003.00 11.700 0.893 3 CP= cyclophosphamide; 4OHCP-SCZ= semicarbazone derivative of 4-hydroxycyclophosphamide; TT= thiotepa; T= tepa. Conclusion A validated assay is described for the simultaneous quantification of CP, thiotepa, and their respective metabolites 4OHCP and tepa in human plasma using LC-MS/MS, requiring only 100 L plasma per sample. The unstable metabolite 4OHCP is converted into the stable 4OHCP-SCZ derivative, using a single-step derivatization procedure. This sensitive and selective LC-MS/MS method allowed minimal matrix interference, and therefore the use of a simple sample clean-up procedure. Accuracies and precisions
Chapter 2.1
98
were well within the predefined limits. The relatively short run time allows high-throughput analysis.
The method described is routinely applied in our hospital to monitor patients receiving the high-dose CTC chemotherapy regimen. Analytical results are available within a short time-window, making dose adaptations during a chemotherapy course possible. The validated ranges have proved to be suitable for routine TDM measurements (Figure 7). So far, approximately 500 samples have been analysed using the described method.
Figure 7. Plasma concentrations of A) cyclophosphamide (!) and 4-hydroxycyclophos-phamide (") and B) thiotepa (!) and tepa (") in a patient treated with a 1-h infusion of CP (1000 mg/m2) followed by a 1-h infusion of carboplatin (265 mg/m2) and a 30 min infusion of thiotepa (40 mg/m2).
A B
10
100
1000
10000
100000
0 5 10 15 20 25Time (h)
Con
cent
ratio
n (n
g/m
l)
10
100
1000
10000
0 1 2 3 4 5 6 7 8Time (h)
Con
cent
ratio
n (n
g/m
l)
Bioanalysis of cyclophosphamide, thiotepa and metabolites
99
Acknowledgements This work was supported with a grant from the Dutch Cancer Society (project NKI 2001-2420). References 1. Van der Wall E, Beijnen JH, Rodenhuis S. High-dose chemotherapy regimens for solid tumors. Cancer Treat
Rev 1995; 21: 105-132 2. Rodenhuis S, Baars JW, Schornagel JH, et al. Feasibility and toxicity study of a high-dose chemotherapy
regimen for autotransplantation incorporating carboplatin, cyclophosphamide and thiotepa. Ann Oncol 1992; 3: 855-860
3. Rodenhuis S, Richel DJ, Van der Wall E, et al. Randomised trial of high-dose chemotherapy and haemopoietic progenitor-cell support in operable breast cancer with extensive axillary lymph-node involvement. Lancet 1998; 352: 515-521
4. Rodenhuis S, Westerman A, Holtkamp MJ, et al. Feasibility of multiple courses of high-dose cyclophosphamide, thiotepa, and carboplatin for breast cancer or germ cell cancer. J Clin Oncol 1996; 14: 1473-1483
5. Rodenhuis S, De Wit R, De Mulder PHM, et al. A multi-center prospective phase II study of high-dose chemotherapy in germ-cell cancer patients relapsing from complete remission. Ann Oncol 1999; 10: 1467-1473
6. Schrama JG, Baars JW, Holtkamp MJ, et al. Phase II study of a multipe-course high-dose chemotherapy regimen incorporating cyclophosphamide, thiotepa, and carboplatin in stage IV breast cancer. Bone Marrow Transplant 2001; 28: 173-180
7. Ayash LJ, Wright JE, Tretyakov O, et al. Cyclophosphamide pharmacokinetics: correlation with cardiac toxicity and tumor response. J Clin Oncol 1992; 10: 995-1000
8. Petros WP, Broadwater G, Berry D, et al. Association of high-dose cyclophosphamide, cisplatin, and carmustine parmacokinetics with survival, toxicity, and dosing weight in patients with primary breast cancer. Clin Cancer Res 2002; 8: 698-705
9. Huitema ADR, Spaander M, Matht RAA, et al. Relationship between exposure and toxicity in high-dose chemotherapy with cyclophosphamide, thioTEPA and carboplatin. Ann Oncol 2002; 13: 374-384
10. Przepiorka D, Madden T, Ippoliti C, et al. Dosing of thiotepa for myeloablative therapy. Cancer Chemother Pharmacol 1995; 37: 155-160
11. Hussein AM, Petros WP, Ross M, et al. A phase I/II study of high-dose cyclophosphamide, cisplatin, and thiotepa followed by autologous bone marrow and granulocyte colony-stimulating factor-primed peripheral-blood progenitor cells in patients with advanced malignancies. Cancer Chemother Pharmacol 1996; 37: 561-568
12. Colby C, Koziol S, McAfee SL, et al. High-dose carboplatin and regimen-related toxicity following autologous bone marrow transplant. Bone Marrow Transplant 2002; 26: 467-472
13. Kloft C, Siegert W, Beyer J, et al. Toxicity of high-dose carboplatin: ultrafiltered and not total plasma pharmacokinetics is of clinical relevance. J Clin Pharmacol 2002; 42: 762-773
14. Van Warmerdam LJC, Van Tellingen O, Maes RAA, et al. Validated method for the determination of carboplatin in biological fluids by Zeeman atomic absorption spectrometry. Fresenius J Anal Chem 1995; 351: 1820-1824
15. Moore MJ. Clinical Pharmacokinetics of cyclophosphamide. Clin Pharmacokinet 1991; 20: 194-208 16. Kwon CH, Maddison K, LoCastro L, et al. Accelerated decomposition of 4-hydroxycyclophosphamide by human
serum albumin. Cancer Res 1987; 47: 1505-1508 17. Baumann F, Preiss R. Cyclophosphamide and related anticancer drugs. J Chromatogr B 2001; 764: 173-192 18. Huitema ADR, Tibben MM, Kerbusch T, et al. High-performance chromatographic determination of the stabilized
cyclophosphamide metabolite 4-hydroxycyclophosphamide in plasma and red blood cells. J Liq Chromatogr Rel Technol 2000; 23: 1725-1744
19. Belfayol L, Guillevin L, Louchahi K. Measurement of 4-hydroxycyclophosphamide in serum by reversed phase high-performance liquid chromatography. J Chromatogr B 1995; 663: 395-399
20. Cohen BE, Egorin MJ, Kohlhepp EA, et al. Human plasma pharmacokinetics and urinary excretion of thiotepa and its metabolites. Cancer Treat Rep 1986; 70: 859-864
21. Van Maanen MJ, Smeets CJM, Beijnen JH. Chemistry, pharmacology and pharmacokinetics of NNN-triethylenethiophosphoramide (ThioTEPA). Cancer Treat Rev 2000; 26: 257-268
Chapter 2.1
100
22. Huitema ADR, Tibben MM, Kerbusch T, et al. Simultaneous determination of N,N',N"-triethylenethio-phosphoramide, cyclophosphamide and some of their metabolites in plasma using capillary gas chromatography. J Chromatogr B 1998; 716: 177-186
23. Baumann F, Lorenz C, Jaehde U, et al. Determination of cyclophosphamide and its metabolites in human plasma by high-performance liquid chromatography-mass spectrometry. J Chromatogr B 1999; 729: 297-305
24. Sottani C, Turci R, Perbellini L, et al. Liquid-liquid extraction procedure for trace determination of cyclophosphamide in human urine by high-performance liquid chromatography tandem mass spectrometry. Rapid Commun Mass Spectrom 1998; 12: 1063-1068
25. Sadagopan N, Cohen L, Roberts B, et al. Liquid chromatography-tandem mass spectrometric quantitation of cyclophosphamide and its hydroxy metabolite in plasma and tissue for determination of tissue distribution. J Chromatogr B 2001; 759: 277-284
26. U.S. Food and Drug Administration: Center for Drug Evaluation and Research: Guidance for Industry: Bioanalytical Method Validation. May 2001; http://www.fda.gov/cder/guidance/4252fnl.htm
27. Craig AW, Jackson H. The metabolism of 32P-labelled triethylenephosphoramide in relation to its anti-tumor activity. Br J Pharmacol 1955; 10: 321-325
28. Van Maanen MJ, Brandt AC, Damen JM, et al. Degradation study of thiotepa in aqueous solutions. Int J Pharm 1999; 179: 55-64
29. Van Maanen MJ, Tijhof IM, Damen JM, et al. The degradation of N,N',N"-triethylenephosphoramide in aqueous solutions: a qualitative and kinetic study. Int J Pharm 2000; 196: 85-94
30. Cohen BE, Egorin MJ, Nayar MSB, et al. Effects of pH and temperature on the stability and decomposition of N,N'N"-triethylenethiophosphoramide in urine and buffer. Cancer Res 1984; 44: 4312-4316
Ther Drug Monit 2003; 25: 261-263
Chapter 2.2
Sorption of thiotepa to polyurethane catheter causes falsely elevated plasma levels
Milly E de Jonge, Ron AA Matht, Selma M van Dam, Sjoerd Rodenhuis, Jos H Beijnen
Abstract Central venous access catheters are commonly used in clinical oncology. The double lumen variant is applied in pharmacokinetic studies for simultaneous administration and blood sampling when frequent blood collections are necessary. Occlusion of one lumen, a common complication, necessitates the investigator to perform blood sampling through the administration lumen after interrupting the infusion. Plasma concentrations measured in this sample can be influenced by sorption of the previously infused compound to the catheter lumen. In this study, the quality of cyclophosphamide, thiotepa, and carboplatin plasma concentrations is investigated when sampling is performed through the administration lumen.
Chapter 2.2
102
Introduction The use of semipermanent central venous access devices has become a common practice in clinical oncology, in particular when high-dose chemotherapy regimens are used. Surgically implanted central venous catheters are used for administration of chemotherapy, reinfusion of hematopoietic cells, and intensive supportive care management, including frequent blood sampling and infusion of antibiotics, analgesics, anti-emetics, blood products, and parenteral nutrition [1]. Prolonged venous access may improve comfort and quality of life of cancer patients by circumventing the need of frequent peripheral venous punctures.
In pharmacokinetic studies, multilumen catheters are usually applied, allowing blood sampling during and after administration of the chemotherapy without interrupting the infusion. Catheter occlusion is a complication that frequently occurs, precluding blood sampling through the plugged lumen. Formation of clots inside the catheter is probably the result of the interaction between the catheter material and the physiologic mechanisms of blood clotting [2]. When lumen occlusion occurs during pharmacokinetic studies, the investigator is forced to collect blood samples from a lumen that is not blocked. Most of the times this will be the same lumen through which administration is realized. When the infused drug interacts with the material of the lumen, measured plasma concentrations of the previously infused drug may be falsely elevated.
Cyclophosphamide, carboplatin, and thiotepa are chemotherapeutic agents that are frequently used in high-dose combination regimens in the treatment of patients with cancer. For drug concentration monitoring, blood samples are drawn during and after administration of the respective compounds. Concentrations measured in the plasma provide information for treatment optimization and, therefore, need to be of high quality.
The goal of this study was to investigate if and to what extent cyclophosphamide, carboplatin, and thiotepa exhibit sorption to the catheter and how this interaction contaminates the collected plasma samples. Patients and methods Patients had a double-lumen polyurethane Arrow-Howes Central Venous Catheter with Blue FlexTip (Arrow International Inc., Reading, PA) located in the right subclavian vein. The infusion was connected to the lumen with the end most downstream. Whole blood samples were collected from the proximal opening [3].
Three patients were treated for 4 consecutive days with cyclophosphamide (1,000 or 1,500 mg/m2/day in 500 ml 0.9% NaCl) during 1 h, followed by carboplatin (265 or 400 mg/m2/day in 500 ml glucose 5%) during 1 h, and thiotepa (40 or 60 mg/m2/day in 100 ml
Sorption of thiotepa to catheter
103
0.9% NaCl) in two 30-min infusions every 12 h. After the end of the thiotepa infusion, other supportive fluids were slowly infused.
For therapeutic drug monitoring, blood samples were taken immediately after end of cyclophosphamide, carboplatin, and thiotepa infusions and 30 min after end of infusion of each compound. Samples were simultaneously drawn from both lumina: first from the proximal ending lumen not used for the infusions (collection lumen), and second from the distal ending lumen used for administration of the chemotherapy (administration lumen). Sampling through the administration lumen was carried out after removing the drug-containing fluid from the lumen by rinsing the lumen with 5 ml saline solution. After withdrawing and discarding 3 ml of blood, the actual sample of 5 ml whole blood is withdrawn using a 5-ml vacuum heparinized tubing. Collection of 5 ml whole blood from the collection lumen (control sample) was performed likewise after withdrawing and discarding 3 ml of blood. After sampling, the lumen was flushed with 5 ml of saline solution.
After collection, the samples were immediately placed on ice. Plasma was subsequently separated by centrifuging the sample at 3,500 rpm for 3 min at 4C. Plasma ultrafiltrate, for carboplatin determination, was prepared immediately using the Amicon micropartition system with a YMT-14 membrane (Millipore Corporation, Bedford, MA). Ultrafiltrate was prepared by transferring 0.5 ml plasma in the micropartition system and centrifuging the system at 2,500 rpm for 15 min. Plasma and ultrafiltrate were stored at -70 C until analysis. Samples were analyzed within 1 week after collection. Analyses of thiotepa and cyclophosphamide were performed using a GC-NPD method [4], and carboplatin was determined using AAS [5]. Within-day and between-day precision and accuracy were less than 10% for both analysis methods.
The patients participated in clinical studies that were approved by the Committee on the Medical Ethics of the Netherlands Cancer Institute. Written informed consent to participate in the pharmacokinetic study was obtained from all patients. Results Plasma concentrations of cyclophosphamide, carboplatin, and thiotepa in samples collected immediately after end of their respective infusions are presented in Table 1. Concentrations are classified according to the lumen of collection. For thiotepa, plasma concentrations observed in the sample collected from the administration lumen just after end of infusion were approximately 100% higher when compared with concentrations in the samples taken from the collection lumen. However, the cyclophosphamide and carboplatin concentrations were similar after sampling through both administration and collection lumen. In the samples taken 30 min after the end of infusion, during which the lumen has been constantly perfused with supporting fluids, thiotepa concentrations in samples taken from both lumina were similar (data not shown).
Chapter 2.2
104
Other sampling procedures for accurately measuring thiotepa from the administration lumen immediately after end of thiotepa infusion have been investigated. Rinsing the lumen with 5 ml saline followed by withdrawing and discarding 10 or 20 ml whole blood (in stead of 3 ml), still resulted in thiotepa plasma concentrations that were 35% to 50% higher in samples drawn from the administration lumen compared with those drawn from the collection lumen. Table 1. Concentrations of thiotepa, cyclophosphamide, and carboplatin at the end of infusion after blood sampling through both collection and administration lumen.
Patient Thiotepa (ng/ml) Cyclophosphamide (g/ml) Carboplatin (M) a b a b a b I 1472 2551 49.09 50.62 65.51 61.99 II 1086 2107 35.06 34.55 55.07 54.10 III 919.4 1917 45.48 44.90 90.12 83.97 a= collection lumen; b= administration lumen Discussion and conclusion Catheter occlusion, a result of thrombus formation, is a frequently seen problem in semipermanent central venous catheters, preventing blood sampling through the plugged lumen and necessitating urokinase instillation [2,6]. Even with effective flushing and the use of a heparin lock in the lumen that is not frequently used, catheter occlusion cannot always be prevented. With persistent occlusive problems, the catheter has to be removed [2]. Pharmacokinetic sampling from the lumen through which administration of the drug is realized is possible after interruption of the infusion. According to our results, this can sometimes be complicated by falsely elevated plasma levels of the drug. Thiotepa appears to interact with the polyurethane catheter, leading to absorption onto the surface of the lumen. Because the interaction is reversible, thiotepa is slowly released from the lumen. In subsequent collected blood samples, thiotepa concentrations are elevated. Cyclophosphamide and carboplatin have apparently minor affinity to the catheter.
Cyclosporin A [7-11], tacrolimus [12], and isosorbide dinitrate [13-15] are examples of drugs that also have been shown to exhibit sorption to infusion systems. The extent of sorption appears to depend on the lipophilicity of the infused drug, its vehicle [12,13], and the material of the catheter [8,12-15]. For the highly lipophilic drugs cyclosporin A and tacrolimus, concentrations in plasma samples, taken from the catheter previously used for their administration, were significantly greater than concentrations obtained peripherally. Sorption of cyclosporin A and tacrolimus persisted for several days [7-12], depending on catheter material [8,12] and continued use of the catheter for administration of other fluids [9-11].
Sorption of thiotepa to catheter
105
Thiotepa sorption to the polyurethane catheter is not as extensive as reported for cyclosporin A and tacrolimus [8,11,12]. The latter compounds are accordingly much more lipophilic than thiotepa, necessitating the use of the vehicles Cremophor EL and Cremophor HC, respectively. Thiotepa sorption to the catheter appeared to be completely neutralized after 30-min infusion of fluids through the lumen. However, applied rinsing methods after end of thiotepa infusion appeared not to eliminate thiotepa from the lumen completely.
Based on these findings, we recommend that a catheter lumen used for infusion of thiotepa should not be used to monitor thiotepa plasma concentrations during and immediately after end of infusion. This is an important finding needing recognition by researchers since otherwise clinical interventions will be based on falsely elevated plasma levels. It is likely that contamination of drug-monitoring samples by sorption of the respective drug to the catheter is relevant for many other, particularly lipophilic, drugs. Investigators should be aware of this and should avoid the need of sampling through the same lumen by which the drug is administered. Alternatives, if possible, could be sampling from a peripheral vein or the initial use of a triple-lumen catheter. However, preventing lumen occlusion to occur is, off course, the best strategy.
Acknowledgements This work was supported with a grant from the Dutch Cancer Society (project NKI 2001-2420).
References 1. Reed WP, Newman KA, De Jongh C, et al. Prolonged venous access for chemotherapy by means of the
Hickman catheter. Cancer 1983; 52: 185-192 2. Barzaghi A, DellOrto M, Rovelli A, et al. Central venous catheter clots: Incidence, clinical significance and
catheter care in patients with hematologic malignancies. Ped Hematol Oncol 1995; 12: 243-250 3. Huitema ADR, Holtkamp, M, Tibben MM, et al. Sampling technique from central venous catheters proves critical
for pharmacokinetic studies. Ther Drug Monit 1999; 21: 102-104 4. Huitema ADR, Tibben MM, Kerbush T, et al. Simultaneous determination of N,N,N-triethylenethiophos-
phoramide, cyclophosphamide and some of their metabolites in plasma using capillary gas chromatography. J Chromatogr B 1998; 716: 177-186
5. Van Warmerdam LJC, Van Tellingen O, Maes RAA, et al. Validated method for the determination of carboplatin in biological fluids by Zeeman atomic absorption spectrometry. Fresenius J Anal Chem 1995; 352: 777-781
6. Horne MK, Mayo DJ. Low-dose urokinase infusions to treat fibrinous obstruction of venous access devices in cancer patients. J Clin Oncol 1997; 15: 2709-2714
7. De Witte T, Hoitsma A, Manoiu MM, et al. Monitoring of cyclosporin during continuous intravenous administration. Bone Marrow Transplant 1986; 1: 147-150
8. Carreras E, Lozano M, Deulofeu R, et al. Influence of different indwelling lines on the measurement of blood cyclosporin A levels. Bone Marrow Transplant 1988; 3: 637-639
9. Lorenz RG, Garrett N, Turk JW, et al. Problems with therapeutic monitoring of cyclosporine using silicone central venous line samples. Transplantation 1991; 6: 1109-1110
Chapter 2.2
106
10. Leson CL, Bryson SM, Giesbrecht EE, et al. Therapeutic monitoring of cyclosporine following pediatric bone marrow transplantation: problems with sampling from silicone central venous lines. Ann Pharmacother 1989; 23: 300-302
11. Blifeld C, Ettenger RB. Measurement of cyclosporine levels in samples obtained from peripheral sites and indwelling lines. N Engl J Med 1987; 317: 509
12. Shaeffer MS, Collier DS, Haven MC, et al. Falsely elevated FK-506 levels caused by sampling through central venous catheters. Transplantation 1993; 56: 475-476
13. De Muynck C, Remon JP, Colardyn F. The sorption of isosorbide dinitrate to intravenous delivery systems. J Pharm Pharmacol 1988; 40: 601-604
14. De Muynck C, Colardyn F, Remon JP. Influence of intravenous administration set composition on the sorption of isosorbide dinitrate. J Pharm Pharmacol 1991; 43: 601-604
15. De Muynck C, Vanderbossche GMR, Colardyn F, et al. Sorption of isosorbide dinitrate to central venous catheters. J Pharm Pharmacol 1993; 45: 139-141