Simultaneous determination of gemcitabine, taxol, cyclophosphamide and ifosfamide in wipe samples by high-performance liquid chromatography/tandem mass spectrometry: protocol of validation

RAPID COMMUNICATIONS IN MASS SPECTROMETRY

Rapid Commun. Mass Spectrom. 2007; 21: 1289–1296

) DOI: 10.1002/rcm.2960
Published online in Wiley InterScience (www.interscience.wiley.com
Simultaneous determination of gemcitabine, taxol,

cyclophosphamide and ifosfamide in wipe samples

by high-performance liquid chromatography/tandem

mass spectrometry: protocol of validation and uncertainty

of measurement

Cristina Sottani1*, Roberta Turci1, Rudolf Schierl2, Raffaella Gaggeri1, Anna Barbieri3,

Francesco Saverio Violante4 and Claudio Minoia1

1Laboratory for Environmental and Toxicological Testing, S. Maugeri Foundation, IRCCS, via Maugeri 10, Pavia, Italy2Institute for Occupational and Environmental Medicine, Ludwig-Maximilians-University, Munich, Germany3Safety, Hygiene and Occupational Medicine Service, University of Bologna, S.Orsola-Malpighi Hospital, via Palagi 9, 40138 Bologna Italy4Occupational Medicine Unit, S.Orsola-Malpighi Hospital, via Palagi, 9, 40138 Bologna Italy

Received 25 October 2006; Revised 29 January 2007; Accepted 31 January 2007

*CorrespoMaugeriE-mail: c

Measurable levels of anticancer agents are still detected on work surfaces in health-care settings.

However, application of recent guidelines for the protection of workers’ safety and health has

resulted in lowered contamination levels. To assess occupational exposure to antineoplastic agents,

very sensitive and specific procedures for environmental sampling and analysis are therefore needed.

In the present study an assay for simultaneous determination of gemcitabine, taxol, cyclopho-

sphamide, and ifosfamide in wipe samples, using two internal standards (trofosfamide and cepha-

lomannine), was developed and validated by high-performance liquid chromatography/tandem mass

spectrometry (HPLC/MS/MS). Solid-phase extraction (SPE) was used for sample concentration and

cleanup. The assay was found to be linear up to 1000 ng/wipe, with limits of quantitation of 25.0 ng/

wipe for gemcitabine and taxol, and 12.5 ng/wipe for cyclophosphamide and ifosfamide. In order to

investigate the effectiveness of the surface sampling, removal efficiency tests were repeated on

different types of surfaces. Recovery rates of between 62 and 81% were obtained at two contamination

levels (50.0 and 250 ng/100 cm2). Precision and trueness were determined on three different days. The

within-day precision was found to be always less than 12.1% for all the analytes. The overall

precision, expressed as relative standard deviation (RSD), was always less than 9.4%. Recoveries

varying from 75.0 (gemcitabine) to 95.0% (taxol) were obtained at three levels. In order to obtain a

quantitative indication of the quality of the result, the overall uncertainty of measurement (UOM)

was evaluated according to the EURACHEM/CITAC guide. The relative combined uncertainty was

found to be always less than 9.5%. The relative expanded uncertainty was also calculated, at three

contamination levels. Copyright # 2007 John Wiley & Sons, Ltd.

Antineoplastic and cytotoxic drugs (ADs) have been shown to

be carcinogenic and/or to have mutagenic and teratogenic

effects on humans.1 Health-care personnel handling these

agents at hospital pharmacies and departments may be

exposed to these drugs, mainly through inhalation and skin

contact. ADs have long been recognized as a potential health

hazard.2–5 In addition, several studies suggested that direct skin

contact should be considered as the main exposure route.6,7

In 1999, the Italian Ministry of Health issued the first

official guidelines8 for the safe handling of antineoplastic

drugs. These guidelines reflect and are consistent with most

international guidelines.9 Safe work practice and personal

ndence to: C. Sottani, Salvatore Maugeri Foundation, Via10, I-27100 Pavia, [email protected]

protective equipment, environmental and biological

monitoring, and education and training of health-care

personnel, are the main points contained in this document.

During the last decade, several methods have been devel-

oped and applied to environmental monitoring.10–13 Wipe

sampling is the common practice used to monitor surface

contamination in preparation and administration rooms at

hospitals.14–17 Gas chromatography/mass spectrometry

(GC/MS) and high-performance liquid chromatography

coupled with ultraviolet detection (HPLC/UV) or tandem

mass spectrometry (HPLC/MS/MS) are the most widely

used techniques in this field.18–24 Increasing numbers of

Copyright # 2007 John Wiley & Sons, Ltd.

1290 C. Sottani et al.

anticancer drugs are being introduced into clinical practice,

and combination therapies are mostly used.

Monitoring surveys require cost- and time-effective tools.

Our objective was therefore to develop a reliable method for

the simultaneous determination of different drugs. HPLC/

MS/MS is the technique of choice for a variety of applications

due to its proven reliability. It was accordingly used in this

study since it provides accurate, precise analyses and high

throughput. Gemcitabine (GCA), taxol (TAX), cyclopho-

sphamide (CP) and ifosfamide (IF) can be identified in a

single run and quantified using trofosfamide (TRO) or

cephalomannine (CEP) as internal standards.

This assay was developed according to the ‘Guide for

Expression of Uncertainty in Measurement’ (GUM) issued by

EURACHEM in 1998 and 2000.25,26 The overall method was

then validated, including uncertainty of measurement (UOM).

Identification of sources of uncertainty, quantification of

uncertainty components, and calculation of total uncertainty

further improve the validity of the analytical results.

Special attention was given to the sampling step. Different

types of surfaces (e.g. laminate, stainless steel or linoleum)

were taken into consideration in order to investigate the

surface-sampling effectiveness.

EXPERIMENTAL

Chemicals and suppliesAnalytical reference standards of CP and IF were supplied by

USP (U.S. Pharmacopeia, Basel, Switzerland, catalog nos.

US15700-2 and US33620-5). TAX was commercially supplied

from Sigma Aldrich (St. Louis, MO, USA). GCA and the two

internal standards, TRO and CEP, were custom-synthesized.

For each compound, a certificate of analysis showing lot

number, purity, and expiration date was provided. GCA was

obtained from Eli Lilly Co. (Indianapolis, IN, USA). TRO was

supplied from Baxter Oncology GmbH (Frankfurt am Main,

Germany) and a powder sample of CEP was kindly provided

by Indena Italy (Milan, Italy). Methanol of HPLC grade, ethyl

acetate, acetonitrile and formic acid of analytical grade were

purchased from VWR International (Poole, UK). Filtered and

deionized water was obtained from a Milli-Q Plus system

(Millipore, Milford, MA, USA).

The analytical column was an Allure PFP Propyl

(150� 4.6 mm� 5mm, 60 A pore size) supplied by Restek

(Bellefonte, PA, USA). A 30� 4.6 mm� 3mm Hypersil ODS

C18 (Thermo Electron, Bellefonte, PA, USA) was used as guard

column. A Supelco (Sigma-Aldrich) VisiprepTM - DL (dis-

posable liner) solid-phase extraction (SPE) vacuum manifold

was used to simultaneously process up to 24 samples.

Pipette models P100 to P5000 were obtained from

Eppendorf (Hamburg, Germany), and disposable pipette

tips from Rainin Instruments (Woburn, MA, USA).

OASIS1 hydrophilic–lipophilic balance (HLB) cartridges

(0.2 g; 6 mL) for SPE (Waters, Milford, MA, USA) were used

and ashless filter paper circles (90 mm diameter) were chosen

as sample media (Whatman, Maidstone, UK).

Borosilicate glass bottles with blue screw-caps and

pouring rings (VWR International) were used to stock wipe

samples until analysis.


Preparation of standard solutionsSeparate stock solutions of GCA, TAX, CP, IF and the internal

standards (TRO and CEP) were prepared in methanol at a

concentration of 1.0 g/L. Two working standard solutions for

the four antineoplastic agents (GCA, TAX, CP, and IF) and

the two internal standards (TRO and CEP) were then created

by diluting the stock 1:100 to obtain a final concentration of

10 mg/L for each analyte.

Sampling procedure (removal efficiency tests)Ashless filter papers were used as sample media for taking

wipe samples. A 10 cm� 10 cm (100 cm2) plastic template

was used to define the area to be wiped. Prior to use, each

wipe filter was moistened with 0.5 mL deionized water.

Volumes of 500mL of the working standard solutions

containing either 50 ng or 250 ng of GCA, TAX, CP, and IF

were pipetted onto the surfaces to be tested. A volume of

500mL of the working internal standard solution (TRO

and CEP, 100mg/L) was also added to the sampling area.

The methanol was then allowed to evaporate completely

before sampling. The space inside the plastic template was

swept clean using vertical strokes. After folding in the

exposed side of the filter, the surface area was wiped a

second time using horizontal strokes. Finally, the wipe filters

were folded with the exposed side inward, transferred to

labelled borosilicate glass bottles, and stored at 48C until

processed.

Three different sites were selected in this study. Stainless

steel work surfaces of biological safety cabinets (BSC), linoleum

floors, and laminated bench tops were chosen as the most

representative work areas where contamination is usually

found.

Sample preparationWater (24.5 mL) was added to each bottle containing wipe

filters to obtain a final volume of 25 mL and the bottles were

shaken for 30 min in a vertical shaker.

Then 10-mL aliquots were loaded onto the SPE cartridges.

OASIS HLB (packing 0.2 g/6 mL column reservoir) SPE tubes

had been previously conditioned with 2 mL of ethyl acetate,

3 mL of methanol, and 1 mL of water. This step was followed

by column equilibration with 4 mL of deionized water. After

sample loading, a mild vacuum (approx. 15 mmHg) was

applied to the manifold and the HLB tubes were left to dry.

Analytes were then eluted at a flow rate of 0.05 mL/s with a

mixture of ethyl acetate/methanol (2:1, v/v). Eluates were

dried under a stream of nitrogen and redissolved in 200mL of

mobile phase. The reconstituted samples were vortexed for

1 min and 20mL of each sample were then injected into the

HPLC/MS/MS system.

In order to investigate whether the differences in the kind

of surface could affect the removal efficiency, additional

wipe samples were prepared in replicate (n¼ 6) at different

contamination levels (50 and 250 ng/100 cm2) and treated as

described above.

Note that during the validation study, the sampling step

was omitted, and the analytes and the internal standards

were added directly onto the filters.


DOI: 10.1002/rcm

Determination of antineoplastic agents in wipe samples 1291

Chromatographic conditionsAnalyses were performed on a Perkin Elmer-Sciex API 300

equipped with a Series 200 LC quaternary pump (Perkin

Elmer, Norwalk, CT, USA). An Allure PFP Propyl column

was used together with a guard column. Mobile phase A

consisted of 0.1% HCOOH in distilled water/acetonitrile

80:20 (v/v), and mobile phase B consisted of 0.1% HCOOH in

distilled water/acetonitrile 30:70 (v/v). An isocratic flow

(1 mL/min) was used for 4 min. The gradient then started at

100% A and linearly increased to 100% B over the course of

8 min, followed by an isocratic gradient of 100% B for 6 min,

before equilibration for 5 min at 100% A (total run time:

18 min). The HPLC eluate was split 1:20 (v/v) so that 50mL/

min entered the source.

The pentafluorophenyl column equipped with a guard

column proved to be the best choice since GCA can be

retained by the stationary phase (retention time

(Rt)¼ 2.7 min). The retention times for CP, IF and TAX were

10.7, 10.9 and 16.3 min, respectively.

Mass spectrometryAn API 300 triple quadrupole mass spectrometer (Perkin

Elmer-Sciex, Toronto, Canada), operating in positive ion mode,

was used to obtain both the full scan mass spectra (MS1) and

the product ion spectra (MS2). The instrument was equipped

with an electrospray ionization (ESI) interface. A Power

Macintosh 7500/100 with an Apple Macintosh System 7.5.3

was used for instrument control and data collection. MAC

QUAN (PE-SCIEX) software was used to process the

quantitative data. Poly(propylene glycol) was used to calibrate

the instrument in positive ion mode and to adjust the peak

width (FWHM) to 0.7 m/z units over the m/z 100–900 range for

HPLC/MS/MS spectral acquisition. The environmental

samples were analyzed by HPLC/MS/MS with an ionspray

needle voltage of 5800 V in the positive ion mode, and with a

cluster-breaking orifice voltage operating at 30 V. The MS/MS

measurements used collisionally activated dissociation (CAD)

in the closed-design Q2 collision cell operating with a collision

energy (DElab) of 5 eV. Nitrogen was used as the collision gas at

a setting of 8 (arbitrary units).

Validation studyLinear regression analysis was used to construct calibration

curves. Calculation of the concentrations was performed via

the peak area ratios from each analyte to the internal

standard (IS, TRO for CP, IF, and GCA; CEP for TAX).

Each calibration curve consisted of a blank sample, a zero

sample (wipe sample processed with internal standard) and

six (GCA and TAX) to seven (CP and IF) calibration

standards (12.5, 25.0, 50.0, 100, 500, and 1000 ng/wipe).

Each blank wipe filter was spiked with 0.5 mL of the working

standard solution (25, 50, 100, 200, 500, 1000, and 2000mg/L),

and with 0.5 mL of the working internal standard solution

containing TRO and CEP at 100mg/L. Calibration curves

were prepared in duplicate.

Quality control samples (QCs) were prepared in replicate

(n¼ 8) at low (40 ng/wipe), medium (312.5 ng/wipe) and

high (625.0 ng/wipe) levels. All the samples were also spiked

with 500mL of the working internal standard solution

(100mg/L). QCs were then treated in exactly the same


manner as the calibration standards, as described in the

‘Sample preparation’ section. Control samples were stored at

48C for future assays, including stability tests.

Additional samples were prepared to investigate removal

efficiency depending on the kind of surface. Three different

surfaces were spiked at the target concentrations of 50 and

250 ng/100 cm2. Samples were collected in replicate (n¼ 6)

using the technique described in the ‘Sampling procedure’

section and analyzed.

The stability of stock solutions of the target compounds

and the internal standards was evaluated at 48C after 1 and 6

months. The effects of temperature on the short-term stability

of the analytes were also assessed at two concentration levels

(100 and 1000mg/L). The spiked wipe samples were kept at

48C for 4 to 24 h.

The limit of detection (LoD) and the limit of quantitation

(LoQ) were determined by analyzing ten replicates of the

blank wipe filters spiked with the working and the internal

standard solutions. LoD is defined as three times the

standard deviation of the blank. LoQ is the lowest

concentration of analyte that can be determined with an

acceptable level of repeatability precision and trueness.25

Linearity, trueness and precision were determined accord-

ing to the EURACHEM guide.25 An increasingly common

expression of accuracy is uncertainty of measurement. It

provides a more powerful figure of accuracy (precision and

trueness) and defines the range of the values that could

reasonably be attributed to the measurand.

The uncertainty was thus evaluated and reported so that a

quantitative indication of the quality of the result can be

given. For a measurement result y, all the uncertainty

components, however evaluated, are combined using the law

of propagation of uncertainty, to give the total uncertainty

u(y), termed combined standard uncertainty.

In this study, the parameters calibration and repeatability

were considered. Since both the uncertainty components are

proportional to the analyte concentration, they are converted

into relative standard deviations (RSDs). The relative

combined standard uncertainty _ucðyÞ is therefore given by:

_ucðyÞ ¼uðyÞ

y¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi_u2

calib þ _u2rep

q

Finally, the relative expanded uncertainty _UðyÞ was

obtained by multiplying _ucðyÞ, the relative combined

standard uncertainty, by a coverage factor k and by ym,

which is the best estimate of the true value. The choice of the

factor is based on the level of confidence desired. For an

approximate confidence level of 95%, k is 2.

_UðyÞ ¼ ½ _ucðyÞ� � k � ym

RESULTS AND DISCUSSION

In ESI positive ion mode the compounds yielded

abundant MHþ ions. These protonated molecules at m/z

262, 161, 264 and 854 for CP, IF, GCA and TAX, respectively,

were selected as the precursor ions for the MS/MS

fragmentation of all the analytes. The product ion spectra

are shown in Fig. 1, together with tentative structures for the

major product ions.


DOI: 10.1002/rcm

Figure 1. ESI-MS/MS spectra of the MHþ ions of (A) CP, (B) IF, (C) GCA, and (D) TAX. The analyte

concentration was 1mg/mL, the flow injection carrier fluid was 0.1% HCOOH/CH3CN (30:70 v/v, pH 3.0),

and the flow rate was 0.05mL/min.


The major product ion of the MHþ ion of CP (m/z 261) is at

m/z 140, originating21 by cleavage of the N–P bond. As can be

seen in Fig. 1(A), the mass resolution in the third quadrupole

analyzer is not particularly high, but this is due to a choice of

enhanced sensitivity rather than resolution. The MHþ ion of IF

(m/z 161) forms its major product ion at m/z 92, by cleavage of

the N–P and C–C bonds with H rearrangement, as shown in

Fig. 1(B). For GCA the product ion spectrum of the MHþ ion at

m/z 264 (Fig. 1(C)) shows only one major ion at m/z 112, due to


the cleavage of the glucosidic bond with H rearrangement. For

TAX the base peak of the product ion spectrum of the MHþ ion

(m/z 854) is detected at m/z 286 (Fig. 1(D)). This involves

cleavage of the O–C bond,27,28 where TAX has been protonated

on the side chain (S), accompanied by H rearrangement,28 to

yield m/z 286 as the Sþ2H ion.

One product ion was monitored for each analyte

ensuring the maximum sensitivity and selectivity. With a

scan time for each fragmentation of 500 ms (dwell time), the


DOI: 10.1002/rcm


precursor-product ion transitions were measured in the

MRM (multiple reaction monitoring) mode. The MRM

parameters and the respective precursor-product ion

combinations used for quantification are summarized in

Table 1. GCA, TAX, CP and IF were quantified by the use of

two internal standards (TRO and CEP). The protonated

molecules of TRO and CEP at m/z 832.6 and 324.4 were

selected as precursor ions with product ions at m/z 164.2

and 153.8, respectively.

Figure 2 shows a typical MRM chromatogram obtained from

the analysis of a processed wipe sample spiked with GCA,

TAX, CP, IF, and the internal standards TRO and CEP. The

concentration of the analytes was 12.5 ng/wipe. This means

that the working range can be considered to be suitable for the

assessment of surface contamination at sites where ADs are

handled.

Calibration curves were linear in the range 0–1000 ng/wipe

with correlation coefficients greater than 0.998. As the method

was intended to quantify more than one analyte, each analyte

was tested to ensure that there was no interference. To this end,

analyses of zero samples showed that no other substances with

similar retention times to those of the target compounds were

detected (data are not shown). Therefore, possible interferences

cannot affect the detection of the four compounds. Similarly,

the analyses of blank samples showed that no substances were

Table 1. MRM parameters for each specific precursor and produ

Compound Molecular structure

GCA

TAX

CP

IF

CEP

TRO


present at the retention times of the four antineoplastic agents

or of the internal standard.

The method was found to be precise with a within-day

precision always less than 12.08%. Day-to-day precision was

assessed over an extended period of time, approximately 15

days, and met acceptance criteria with RSD <9.4%.

Trueness, expressed as relative error (RE%), ranged from

�10 to 2. All the validation parameters (precision and trueness)

obtained at low, medium, and high concentrations during the

3-day measurement period are summarized in Table 2.

Extraction recoveries for the target analytes were evaluated

at QC levels. At the lowest level (40 ng/wipe), mean recoveries

varied between 75 and 85%. All recovery values, with the

relative standard deviations (%RSD), are listed in Table 3.

In the removal efficiency tests, it can be seen (Table 4)

that the lowest recovery percentages were found for all

the analytes when wipe samples had been taken from the

laminated bench top. These results are consistent with the

principle that an ideal surface is extremely smooth and

non-porous, since stainless steel and linoleum are less porous

than laminate. Mean recoveries ranged from 62.7 (TAX, low

level) to 73.2% (CP, high level).

Stock solutions were found to be stable for 6 months when

stored at 48C, with the exception of gemcitabine and taxol

which were shown to degrade over time. The nominal

ct ion combination and their respective retention times (RT)

RT(min)

Precursor ion(m/z)

Product ion(m/z)

2.7 264.1 112.2

15.3 854.0 286.2

10.9 261.0 140.1

10.7 261.0 92.0

14.8 832.6 264.2

13.5 323.4 153.8


DOI: 10.1002/rcm

Figure 2. MRM chromatogram of an extracted wipe sample containing GCA, IF, CP, and TAX with TRO and CEP as

internal standards (12.5 ng/wipe).

Table 2. Validation parameters: precision and trueness

Nominal conc. Validation parameters

GCA TAX CP IF

Day 1 Day 2 Day 3 Day 1 Day 2 Day 3 Day 1 Day 2 Day 3 Day 1 Day 2 Day 3

Mean (Xm) (n¼ 8) 40.0 39.3 38.8 41.0 41.3 41.3 37.8 39.5 40.0 39.5 41.3 41.0(%RSD; within-day)(n¼ 8)

4.3 8.0 8.6 6.5 8.6 7.8 4.6 12.1 7.3 5.4 9.0 9.1

40.0 ng/wipe between days (n¼ 24) between days (n¼ 24) between days (n¼ 24) between days (n¼ 24)(% RE) (n¼ 24) 1.8 �2.8 1.9 1.4(%RSD) (n¼ 24) 8.5 7.4 8.8 8.0Mean (Xm) (n¼ 8) 305.3 333.0 329.5 289.5 330.3 319.3 340.0 339.0 346.8 311.5 350.5 344.3(%RSD; within-day)(n¼ 8)

8.2 7.0 8.2 8.4 7.3 4.5 6.2 10.8 7.0 3.9 10.6 7.9

between days between days between days between days312.5 ng/wipe (% RE) (n¼ 24) �3.2 �0.2 �8.6 �7.4

(%RSD) (n¼ 24) 8.5 8.6 8.0 9.4

Mean (Xm) (n¼ 8) 593.8 616.5 619.75 590.0 627.8 626.3 631.5 627.5 641.8 629.3 629.3 635.8(%RSD; within-day)(n¼ 8)

4.8 6.7 6.1 4.2 5.4 3.2 6.1 7.8 6.7 2.8 9.2 7.6

between days between days between days between days625.0 ng/wipe (% RE) (n¼ 24) 2.0 1.7 �7.4 �1.0

(%RSD) (n¼ 24) 6.0 5.1 6.7 6.7

Table 3. Recovery percentages (%RSD) from the wipe samples spiked with the four analytes

Compound

Recovery (%RSD)

Spiking level 40.0 ng/wipe Spiking level 312.5 ng/wipe Spiking level 625 ng/wipe

GCA 75 (8.2) 83 (7.9) 88 (8.5)TAX 78 (7.5) 92 (5.5) 95 (4.1)CP 85 (8.1) 87 (8.5) 90 (10.1)IF 85 (11.3) 89 (5.7) 92 (4.8)

Copyright # 2007 John Wiley & Sons, Ltd. Rapid Commun. Mass Spectrom. 2007; 21: 1289–1296

DOI: 10.1002/rcm


Table 4. Removal efficiency tests: recoveries of the target

analytes (%RSD) from wipe samples taken from sites where

antineoplastic agents are typically handled (stainless steel work

tray of the BSC, laminated bench top and linoleum flooring)

Compound Surface

Removal efficiency (%RSD)

Level(50 ng/100 cm2)

Level(250 ng/100 cm2)

stainless steel hoodwork tray

68 (25.2) 81 (10.5)

GCA laminated bench top 62 (21.2) 69 (8.5)linoleum flooring 73 (16.2) 76 (15.5)


65 (20.5) 76 (11.1)

TAX laminated bench top 55 (16.5) 71 (9.1)linoleum flooring 68 (23.7) 81 (10.5)


75 (11.1) 80 (10.1)

CP laminated bench top 65 (8.1) 70 (12.1)linoleum flooring 71 (17.2) 78 (8.5)


68 (15.3) 79 (13.8)

IF laminated bench top 63 (12.3) 63 (7.8)linoleum flooring 68 (21.2) 81 (9.5)


concentration of GCA and TAX decreased by 20% over the

course of 3 months. GCA, TAX, CP and IF were found to be

stable in matrix (spiked wipe samples stored into glass

bottles) at 48C for at least 24 h.

Gemcitabine: Uncertainty Components

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

Sta

ndar

d U

ncer

tain

ties

(%)

u Combined 4.2 3.0 2.2

u repeatability 3.0 2.9 2.1

u calibration 2.9 0.5 0.3

Level 40.0 ng/wipe Level 312.5 ng/wipe Level 625.0 ng/wipe

Taxol: Uncertainty Components

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

Sta

nd

ard

Un

cert

ain

ties

(%

)

u Combined 5.9 1.8 1.3

u repeatability 2.8 1.6 1.1

u calibration 5.2 0.8 0.4


Figure 3. Relative uncertainty values (%) obtained for GCA


The LoD was found to be 12.5 ng/wipe for GCA and TAX,

and 6.25 ng/wipe for CP and IF. The LoQ was assessed

at 25 ng/wipe for GCA and TAX, and at 12.5 ng/wipe for

CP and IF.

Uncertainty of measurement in methodvalidationThe analytical method was fully validated and uncertainty of

measurement (UOM) was also evaluated. The relevant

uncertainty sources were identified, quantified, and then

combined to determine the overall uncertainty.

Type A uncertainties are the standard uncertainty of

calibration ( _ucalib) and the standard uncertainty of repeatability

( _urep). In addition, there are two type B uncertainties:

– t

u C

u r

u c

u

u

u

, T

he uncertainty in the certified internal volume of the grade

A volumetric flask which was used to prepare the working

standard solutions, and

– t
he uncertainty stated by the manufacturer for the 200mL
pipette which was used to reconstitute the dry residue

obtained after SPE.

A 100 mL grade A volumetric flask is certified to

within� 0.10 mL. The relative standard uncertainty ( _u

glassware tolerance) is 0:1=ffiffiffi3

p� �=100 � 0:0006.

A 0.2 mL pipette is certified to within� 0.3mL. The relative

standard uncertainty ( _u volume) is 0:3=ffiffiffi3

p� �=200 � 0:00086.

As can be seen, type B uncertainties can be ignored as they

are negligible compared with the other sources. The different

Ifosfamide: Uncertainty Components

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

9.0

10.0

Sta

nd

ard

Un

cert

ain

ties

(%

)

ombined 9.4 2.9 2.7

epeatability 3.2 2.8 2.7

alibration 8.8 0.8 0.5


Cyclophosphamide: Uncertainty Components

0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

5.0

Sta

ndar

d U

nce

rtai

ntie

s (%

)

Combined 4.4 2.5 2.4

repeatability 2.6 2.5 2.4

calibration 3.5 0.4 0.2


AXl, CP and IF in wipe samples at three QC levels.


DOI: 10.1002/rcm

Table 5. Relative expanded uncertainties

QC1�U(ng/wipe)

QC2�U(ng/wipe)

QC3�U(ng/wipe)

IF 40.0� 9.4 312.5� 23.0 625.0� 40.9GCA 40.0� 3.6 312.5� 22.9 625.0� 31.7TAX 40.0� 5.7 312.5� 12.7 625.0� 17.2CP 40.0� 3.9 312.5� 20.1 625.0� 36.1


uncertainty contributions can easily be visualized in

histograms, as shown in Fig. 3. The spreadsheet filled in

with the appropriate values is shown as a table below the

diagram.

At QC level 2 (QC2, 312.5 ng/wipe) and QC level 3 (QC3,

625 ng/wipe), the contribution of the repeatability uncer-

tainty is by far the largest. At QC level 1 (QC1, 40 ng/wipe),

the relative uncertainty of calibration was found to be the

main contribution, except for GCA. The highest combined

uncertainty was measured for IF (9.4% at QC level 1).

The expanded uncertainty U was also evaluated, so that

the results can be expressed following the recommended

form: Result y¼ x� (U) (Table 5). The expanded uncertainty

is required to provide an interval which may be expected to

include a large fraction of the values that could reasonably be

attributed to the measurand. In addition, knowledge of the

overall uncertainty and its major components allows to

understand where improvements can be made.

CONCLUSIONS

A novel procedure for the simultaneous determination of

GCA, TAX, CP and IF in wipe samples was developed using

HPLC/MS/MS and validated, including uncertainty of

measurement. In addition, collection efficiencies for different

types of surfaces were evaluated.

Wipe sampling is one of the most important tools of

worksite analysis. It enables the evaluation of significant

exposures caused by surface contamination. In addition, it

can be used to assess the effectiveness of decontamination

procedures.

Based on our field experience, this method can be

considered to be suitable for the assessment of residual

contamination in work areas where ADs are handled and

safety procedures are in place.

However, variables such as sampling media, moistening

agents, and collection procedures should be studied and

optimized in an attempt to further enhance collection

efficiency. Future work will not only focus on the develop-

ment of a standard wipe sampling protocol, but will also

include a further increase in the number of drugs monitored.


REFERENCES

1. International Agency for Research on Cancer. Monographson the Evaluation of the Carcinogenic Risk of Chemicals toHumans: Pharmaceutical Drugs. Lyon, France. Available:http://monographs.iarc.fr/

2. Falck K, Grohn P, Sorsa M, Vainio H, Heinonen E, Holsti LR.Lancet 1979; 1: 1250.

3. Donner AL. Am. J. Hosp. Pharm. 1978; 35: 900.4. Sorsa M, Anderson D. Mutat. Res. 1996; 355: 253.5. Sessink PJM, Bos RP. Drug Safety 1999; 4: 347.6. Kromhout H, Hoek F, Uitterhoeve R, Huijbers R, Overmars

RF, Anzion R, Vermeulen R. Ann. Occup. Hyg. 2000; 44: 551.7. Fransman W, Vermeulen R, Kromhout H. Ann. Occup. Hyg.

2004; 48: 237. DOI: 10.1093/annhyg/meh017.8. Provvedimento 5 agosto 1999. Documento di linee-guida per

la sicurezza e la salute dei lavoratori esposti a chemioterapiciantiblastici in ambiente sanitario, G.U. 236, 7.10.1999 (ItalianGuidelines).

9. OSHA. Work Practice Guidelines for Personnel Dealing withCytotoxic (Antineoplastic Drugs). US Dept of Labor: Washing-ton, DC, 1986. OSHA Instruction Publication 8-1.1, Office ofOccupational Medsicine.

10. Connor TH, Anderson RW, Sessink PJ, Broadfield L, PowerLA. Am. J. Health Syst. Pharm. 1999; 56: 1427.

11. Hedmer M, Jonsson BA, Nygren O. J. Environ. Monit. 2004; 6:979. DOI: 10.1039/b409277e.

12. Minoia C, Turci R, Sottani C, Schiavi A, Perbellini L, AngeleriS, Draicchio F, Apostoli P. Rapid Commun. Mass Spectrom.1998; 20: 1485.

13. Turci R, Sottani C, Spagnoli G, Minoia C. J. Chromatogr.B Analyt. Technol. Biomed. Life Sci. 2003; 789: 169.

14. Baker ES, Connor TH. Am. J. Health Syst. Pharm. 1996; 53: 27.15. Mason HJ, Blair S, Sams C, Jones K, Garfitt SJ, Cuschieri MJ,

Baxter PJ. Ann. Occup. Hyg. 2005; 49: 603. DOI: 10.1093/annhyg/mei023.

16. Shultz H, Bigelow S, Dobish R, Chambers CR. J. Oncol.Pharm. Pract. 2005; 11: 101.

17. Connor TH, Sessink PJM, Harrison BR, Pretty JR, Peters BG,Alfaro RM, Bilos A, Beckmann G, Bing MR, Anderson LM,De Cristoforo R. Am. J. Health Syst. Pharm. 2005; 62: 475.

18. Schreiber C, Radon K, Pethran A, Schierl R, Hauff K, GrimmCH, Boos KS, Nowak D. Int. Archives Occup. Environ. Health2003; 76: 11. DOI: 10.1007/s00420-002-0385-6.

19. Schmaus G, Schierl R, Funck S. Am. J. Health Syst. Pharm.2002; 59: 956.

20. Acampora A, Castiglia L, Miraglia N, Pieri M, Soave C, Liotti F.Ann. Occup. Hyg. 2005; 7: 611. DOI: 10.1093/annhyg/mei029.

21. Sabatini L, Barbieri A, Tosi M, Violante FS. J. Mass Spectrom.2005; 40: 669. DOI: 10.1002/jms.840.

22. Zeedijk M, Greijdanus B, Steenstra FB, Uges DRA. Eur. J.Hosp. Pharm. 2005; 1: 18.

23. Rubino FM, Floridia L, Pietropaolo AM, Tavazzani M,Colombi A. Med. Lav. 1999; 4: 572.

24. Micoli G, Turci R, Arpellini M, Minoia C. J. Chromatogr.B Biomed. Sci. Appl. 2001; 750: 25.

25. The Fitness for Purpose of Analytical Methods, EURACHEM/CITAC Guide (1st edn). EURACHEM/CITAC: Teddington,UK, 1998. Available: www.eurachem.ul.pt

26. Quantifying Uncertainty in Analytical Measurement, EURA-CHEM/CITAC Guide (2nd edn), Final Draft, EURACHEM/CITAC: Teddington, UK, 2000. Available: www.eurachem.ul.pt

27. Lee ML, Kerns EH. Mass Spectrom. Rev. 1999; 18: 187.28. Royer I, Alvinerie P, Armand JP, Ho LK, Wright M,

Monsarrat B. Rapid Commun. Mass Spectrom. 1995; 9: 495.


DOI: 10.1002/rcm

Simultaneous determination of gemcitabine, taxol, cyclophosphamide and ifosfamide in wipe samples by high-performance liquid chromatography/tandem mass spectrometry: protocol of validation

Documents