Drug Testing

Research article and Analysis

Received: 8 November 2016 Revised: 9 February 2017 Accepted: 19 March 2017 Published online in Wiley Online Library: 12 May 2017

(www.drugtestinganalysis.com) DOI 10.1002/dta.2195

Development of a dried blood spot HPLC-PDA

method for the analysis of linezolid and
ciprofloxacin in hospital-acquired
pneumonia patients
Vincenzo Ferrone,a Maura Carlucci,b Roberto Cotellese,b Paolo Raimondi,b
Annadomenica Cichella,b Lorenzo Di Marco,a Salvatore Genovesea
and Giuseppe Carluccia*
This study developed a high performance liquid chromatography (HPLC) method involving dried blood spotting (DBS) as a
sampling method for the therapeutic drug monitoring of antimicrobial combination therapy with linezolid and ciprofloxacin.
DBS for standards, quality control samples, and patient samples was excised and then extracted using a mixture of
methanol/water/formic acid 80:20:0.1 (v/v/v), respectively. Linezolid (LZD) and ciprofloxacin (CPR) were measured by HPLC
using a Kinetex EVO C18 (100 × 4.6 mm I.D. 2.6 μm particle size). Mobile phase consisted of 10 mM ammonium acetate (A)
and a mixture of acetonitrile and methanol (B), both containing 0.1% triethylamine, in gradient elution. Detection was carried
out at 251 nm for linezolid (LZD) and 277 for ciprofloxacin (CPR). Ulifloxacin was used as an internal standard. The internal
standard, LZD, and CPR were eluted in 7.90, 7.18, and 8.89 min, respectively. Total run-time was 20 min. Calibration curves
were constructed in the range of 0.05–30 μg/mL for LZD and 0.02–10 μg/mL for CPR, respectively. The intra- and inter-day
precision (RSD values) did not exceed 9.43%, the intra- and inter-day accuracy (accuracy %) ranged between 96.2 and
106.2%. Haematocrit (Hct) effects were investigated to obtain a linear correlation between haematocrit values and volume
of blood. Copyright © 2017 John Wiley & Sons, Ltd.

Keywords: linezolid; ciprofloxacin; dried blood spot; method validation; HPLC-PDA

Introduction quinolone derivative obtained by addition of a fluorine atom in

Pneumonia is a respiratory infection of the lung affecting primarily
the microscopic air sacs known as alveoli.[1] It is defined as being
nosocomial or hospital-acquired pneumonia (HAP) if it becomes
apparent more than 48 h after hospital admission.[2] HAP affects
0.5 to 1% of inpatients and is the most common healthcare-
associated infection contributing to death.[3] Pneumonia in
ventilated patients is defined as ventilator-associated pneumonia
(VAP) and is considered a subset of HAP. The mortality rate of
patients with VAP is 24% to 50% which increases to 76% if infection
is caused by multidrug-resistant pathogens.[4] To prevent multidrug
resistance, an antimicrobial combination therapy may be used.
Linezolid (LZD) or (N-[[(5S)-3-[3-Fluoro-4-(4-morpholinyl)phenyl]-2-
oxo-5-oxazolidinyl]methyl]acetamide) (Figure 1a) is an
antimicrobial agent from the oxazolidinone class. It is especially
active against multidrug-resistant Gram-positive organisms such
as methicillin-resistant Staphylococcus aureus (MRSA) and certain
Gram-negative and anaerobic bacteria.[5,6]
LZD inhibits bacterial ribosomes by binding to the 23S ribosome
and preventing the 30S–50S fusion.[7] The achievement of a
minimum plasma concentration (Cmin) >2 μg/mL has been
proposed as theoretical threshold to ensure LZD efficacy.[8]
Ciprofloxacin (CPR) or N (1-cyclopropyl-6-fluoro-4-oxo-7-
position 6 and a piperazine substituent in position 7 with different
chemical structure and spectrum of activity.[9] CPR stabilizes
enzyme/cleaved DNA complexes within the cell, terminating DNA
replication. Ciprofloxacin, alone or in combination with other
antimicrobial drugs, provides effective treatment for a variety of
infections, in particular urinary, respiratory, and gastrointestinal
tract.[10] Fluoroquinolones generally show concentration-
dependent bacteria-killing activity that depends on the ratio of
maximum drug concentrations to minimum inhibitory
concentration.[11]
Different methods describing the determination of
fluoroquinolones or LZD concentrations in biofluids have been
reported.[12–20]
High performance liquid chromatography (HPLC) with UV,
fluorescence, mass spectrometry (MS), or tandem mass
* Correspondence to: Giuseppe Carlucci, Dipartimento di Farmacia, Università
degli Studi ‘G. d’Annunzio’ Chieti-Pescara, Via dei Vestini, 66100 Chieti, Italy.
E-mail: g.carlucci@unich.it
a Department of Pharmacy, Università degli Studi, G.d’Annunzio’ Chieti-Pescara,
via dei Vestini, 66100 Chieti, Italy
1611

b Department of Oral Health Sciences and Biotechnology, Università degli Studi, G.

(piperazin-1-yl)-quinoline-3-carboxylic acid) (Figure 1b) is a d’Annunzio’ Chieti-Pescara, via dei Vestini, 66100 Chieti, Italy

Drug Test. Analysis 2017, 9, 1611–1619 Copyright © 2017 John Wiley & Sons, Ltd.
 Drug Testing
and Analysis
Figure 1. Chemical structure of (a) linezolid, (b) ciprofloxacin, and (c)
ulifloxacin, the internal standard.
spectrometry (MS/MS) detection are the techniques most used.
Conventional solid-phase extraction (SPE) and liquid–liquid
extraction (LLE), although being efficient methods to extract LZD
and CPR from biological fluids, are impractical in routine analyses
as they are laborious and solvent- and time-consuming. In addition,
some of them have deleterious effects on the environment due to
the large quantities of organic solvents that need to be used and
discarded.
Due to the limitation of performing therapeutic drug monitoring
in hospitals which do not have a therapeutic drug monitoring
(TDM) laboratory, dried samples spot devices are becoming a valid
option to overcome these problems. In recent years, dried blood
spots (DBS) have gained increasing attention due to the
development of new applications. Bang was the first to make
practical and effective use of DBS for the determination of blood
sugar levels. However, this sampling technique did not achieve its
breakthrough as a cheap screening method until the 1960s, when
it was applied for the detection of phenylketonuria.[21] Less invasive
sampling, simplified sample preparation, and easier storage are the
reasons for the recent increase of interest in DBS techniques.
Despite the numerous applications using DBS,[22–26] the exact
quantification is still a frequently discussed issue.[27–31]
A first concern is if and how venous and capillary blood
concentrations of analyte correlate.[29] Another challenge in DBS-
based sampling is the spot volume, because in clinical samples
obtained from a finger-prick, the blood volume is not controlled
and will vary.[28,30] Other issues to be investigated are the site of
punching and the haematocrit (Hct) effect. The latter is the most
widely discussed DBS-related problem due to the accuracy of the
Hct measurement and to the blood/plasma distribution ratio of
an analyte. To date, the spot volume, the site of punching, and
the Hct effect must be evaluated when a DBS sampling technique
is performed. [28–31]
At present, no methods for the quantification of both CPR and
LZD in DBS are reported. The aim of this study is to describe the first
set-up, application, and validation of an HPLC method to quantify
the TDM of LZD and CPR involving dried blood spot as a
biosampling procedure.
Experimental design
Chemicals and reagents
LZD, (CAS 165800–03-3) and CPR (CAS 85721–33-1) were
purchased from Sigma-Aldrich (Milan, Italy), while ulifloxacin (CAS
1612
no. 112984–60-8), used as the internal standard (I.S.), was supplied
wileyonlinelibrary.com/journal/dta Copyright © 2017 John Wiley & Sons, Ltd.
V. Ferrone et al.
by Suzhou Bichal Biological Technology Co. Ltd (Jiangsu, China).
Methanol, acetonitrile, triethylamine, and ammonium acetate were
obtained from Carlo Erba Reagents (Milan, Italy). Hydrochloric acid
was from VWR International (Milan, Italy). All chemicals were of
analytical-reagent grade or better.
All dilutions were prepared with HPLC-grade water obtained by
passage through an Elix 3 and Milli-Q Academic water purification
system (Millipore, Bedford, MA, USA). Bond Elut DMS card was
supplied by Agilent Technologies (Santa Clara, CA, USA). Blank
whole blood from healthy donors was kindly supplied by S.S.
Annunziata Hospital Transfusion Centre (Chieti, Italy).
Chromatographic conditions
The chromatographic apparatus for the determination of these
analytes was a Waters system consisting of a model 600 pump
and an UV–Vis photodiode array detector model 2996 (Waters,
Milford, MA, USA). A Rheodyne model 7125i injector (Rheodyne,
Cotati, CA, USA) equipped with 20 μL loop and a degasser system
model DG-4400 (Phenomenex, Torrance, CA, USA) was used.
Chromatographic separation was achieved using a Kinetex EVO
C18 (100 × 4.6 mm I.D. 2.6 μm particle size) column protected by
a disposable Security Guard Gemini precolumn (3.0 × 4.0 mm)
(Phenomenex, Torrance, CA, USA) maintained at 25 ± 1°C, using a
thermostatically-controlled column heater. Mobile phase consisted
in 10 mM ammonium acetate buffer adjusted to pH 3.5 with acetic
acid (phase A) and a mixture of acetonitrile and methanol in a ratio
of 80/20 (v/v) (phase B), both phases were added with 0.1% (v/v) of
triethylamine. To perform the best separation of the analytes, a
linear gradient elution program was used; the gradient began from
0 to 4 min with 90% and 10% of phase A and B, respectively, from 5
to 10 min became 65% and 35% of phase A and B, respectively, and
returned to the original conditions between 11 and 13 min
followed by 7 min of re-equilibration of the column to the initial
condition. The flow rate was set at 1.0 mL/min. The solvents were
filtered before use through a 0.45 μm WTP membrane, while
ammonium acetate solution was filtered through a WCN 0.5 μm
membrane (Whatmann, Maidstone, UK). Detection was carried
out at 251 nm for LZD and 277 for CPR. Empower v.2 Software
(Waters Spa, Milford, MA, USA) was used for setting-up the analysis
and for data management.
Preparation of calibration standards and quality controls
Stock solutions of 2 mg/mL of LZD, CPR and I.S. were prepared by
appropriate dissolution of the reference powders into methanol.
Then the stock solutions were diluted with Milli-Q water to obtain
the working solutions. Calibration curve was prepared by spotting
on filter paper spiked control blood with LZD, CPR, and I.S. to obtain
a concentration of 0.05, 0.15, 0.50, 2.00, 5.00, 15.00, and 30.00 μg/
mL for linezolid and 0.02, 0.10, 0.25, 0.50, 1.00, 2.50, and 10.00 μg/
mL for ciprofloxacin. Quality control samples were prepared in a
similar way at the concentration of 0.10, 1.00, and 25.00 μg/mL
for LCD and 0.05, 0.75, and 7.50 μg/mL for CPR, the I.S. (Figure 1c)
concentration was 5 μg/mL. The I.S. was chosen because it is not
co-administrated with the investigated analyte and is not abused
by the patients.
Sample preparation
Venous blood was taken and collected in EDTA-coated 7.5
ampoules and was used for the DBS-specific validation except for
Drug Test. Analysis 2017, 9, 1611–1619

DBS-HPLC-PDA analysis of linezolid and ciprofloxacin
the correlation of the capillary dried blood spot (CDBS) with venous
dried blood spot (VDBS) where punched discs of VDBS and CDBS
from patients were analyzed for their CPR and LZD levels.
DBS samples were collected at the same time as routine TDM
samples for plasma and did not require extra visits to the clinic or
additional fingerpricks.
Fingerprick blood was directly spotted onto the paper cards by
placing two-to-three drops of blood on the DBS card, omitting a
direct contact between the finger and the card. To prepare DBS
spiked samples, 25 μL of blood, previously spiked with the analytes
and the I.S., was transferred onto a Bond Elut DMS card by an
Eppendorf pipette, and was left to dry 3 h at room temperature
and then transferred into a sealed plastic bag containing desiccant
and stored at
20C° until analysis. Validation was performed using
blood with Htc value of 45% except for the evaluation of the Htc
percentage effect on the analysis. Extraction from DBS was
performed using a mixture of methanol/water/formic acid
80:20:0.1 (v/v/v), respectively.
A 9-mm diameter disk was punched out from DBS and
transferred into a 2.0 mL Eppendorf containing 300 μL of extraction
solution and was sonicated for 25 min at 35°C followed by
centrifugation at 4000 rpm for 10 min to obtain a clear supernatant.
The supernatant was then transferred into a clean tube, evaporated
to dryness under a gentle stream of nitrogen and reconstituted to
25 μL with mobile phase, 20 μL was injected into the HPLC system.
Method validation
The proposed method was validated according to the guidance for
industry on the validation of bioanalytical methods.[32]
Figure 2. A typical chromatogram of LZD, CPR, and I.S. after injection of: (a) a blank DBS extract, (b) a sample of patient in standard therapy with intravenous
linezolid (Zyvox 600 mg twice daily) and ciprofloxacin (ciproxin 400 mg twice daily) 4 h after administration of these drugs, (c) a blank DBS extract spiked with
5 μg/mL of internal standard and both LZD and CPR.
Drug Test. Analysis 2017, 9, 1611–1619 Copyright © 2017 John Wiley & Sons, Ltd.
Drug Testing
and Analysis
Linearity
Linearity was evaluated using freshly prepared spiked matrix
samples in the range of 0.05–30 μg/mL for LZD and 0.02–10 μg/
mL for CPR. Each calibration curve consisted in a seven-calibrator
concentration. Daily calibration curves were constructed using the
ratios of the observed peak areas of LZD and CPR to the I.S., to
describe the relationships between concentration of the analytes
and the detector response, a linear regression analysis with
weighting factors consisted in 1/x2 values.
Precision and accuracy
Accuracy and precision were evaluated in six replicates at lower
limit of quantification (LLOQ), low-level quality control (LQC),
medium-level quality control (MQC), and high-level quality control
(HQC) on the same day and three analytical batches of QCs level
on six consecutive days, respectively. Accuracy was determined
by the mean of the measured QC concentration to the theoretical
QC value and reported in percentage. Precision was defined by
the relative standard deviation percentage (RSD%).
Limit of detection and limit of quantification
The limit of detection (LOD) was defined as the concentration
which gave a signal-to-noise ratio of 3 while the limit of
quantification was determined by injecting six times a spiked
sample with decreasing concentration of the analyte. LOQ was
regarded as the minimum amount of analytes that gave a precise
measurement (accuracy within ±20% and RSD% <20%).
1613
wileyonlinelibrary.com/journal/dta
 Drug Testing
and Analysis V. Ferrone et al.

Selectivity form of pretreatment, which can retain several matrix compounds

The selectivity of the method was tested by comparing the
chromatograms of six different batches of blank plasma.
Figure 2 shows a representative chromatogram of a blank plasma
sample. There was no chromatographic interference from
endogenous compounds at the retention times of linezolid and
ciprofloxacin. The chromatographic peaks were well resolved to
the baseline. During the selectivity assay, other co-administered
drugs in spiked blood which were not found to interfere with
the assay.
Recovery
Recovery of LZD and CPR was evaluated by comparing the mean
peak areas of different post extracted QC samples with those
obtained by spiking the post extracted blank matrix with the
analytes at the QCs levels of concentration.
Stability
Stability of LZD and CPR were investigated during sample
collection, after long-term storage and short-term storage, and
through several freeze and thaw cycles according to the guidance
for industry on the validation of bioanalytical methods.[32]
in the paper.
Choosing the most suitable extraction solvent is of primary
importance to achieve good extractability and selectivity of the
target analytes in DBS. First, to select the extraction solvent,
methanol ethanol and acetonitrile were tested using 300 μL for
each organic solvent. Methanol was chosen due to the best
extraction yield compared to acetonitrile and ethanol. The
extraction solution was chosen by testing different solvent
compositions, at various percentages of organic composition;
methanol/water/formic acid 80:20:0.1 (v/v/v), respectively, gave
the best extraction recovery, as shown in Figure 3. Different
extraction times were tested, from 5 to 50 min every 5 min. A
plateau in the extraction yield was reached after 25 min
(Figure 4), therefore 25 min were chosen as optimum for the
extraction time. The last parameter tested was the extraction
temperature, in the range of 25–50°C, every 5°C; the optimum
was found to be 35°C.
Experiments regarding the homogeneity of drug distribution
were carried out, and no significant differences were found, which
was due to the low blood volume spotted (25 μL).

DBS-specific validation
A spot volume range of at least 15–40 μL should be investigated
during the validation. The spotting volumes should be tested at a
minimum of two concentration levels (low and high) in triplicate.
In all cases, the spot volume for calibration standard should be
equal to the expected spot volume of the clinical samples. When
these tests show that the effect of spot volume falls within
acceptable ranges, there is no need to use accurately spotted
volumes for clinical application. [30]
Possible differences between punches taken from the centre or
from the periphery of the spot should be investigated during DBS
method development and validation. Spot homogeneity was
investigated by preparing DBS at two concentration levels (low
Figure 3. Evaluation of the extraction yield as a function of the percentage
and high) in triplicate and punching from the centre and the of methanol in the extraction solution.
periphery of the DBS. When deviated analysis results are observed
from the peripheral punches, a bigger puncher could be tested to
overcome this problem. Whole blood was centrifuged at 1500 g
for 15 min to separate the plasma and red blood cells. Increasing
volumes of plasma were added to whole blood providing
decreasing Hct levels within 30–55%. The non-centrifuged whole
blood with 45% Hct served as 100% control. These were then
analyzed for %Hct levels on an ABX Pentra XL 80. The blood was
then spiked with CPR and LZD, providing 3 QC concentration
(low, medium, and high) at each investigated Hct value, and
spotted onto collection cards. Punches (9 mm) taken from the
centre of each spot were subsequently analyzed.

Results and discussion

Optimization of extraction from DBS
The use of DBS is a valid alternative compared to traditional blood
sampling; the analysis of a few drops of blood can also be suitable
for pharmacokinetic studies, TDM, and analyses involving children.
1614

The simple blood adsorption onto the paper can be considered a Figure 4. Extraction yield % as a function of the extraction time.

wileyonlinelibrary.com/journal/dta Copyright © 2017 John Wiley & Sons, Ltd. Drug Test. Analysis 2017, 9, 1611–1619

DBS-HPLC-PDA analysis of linezolid and ciprofloxacin
Evaluation of haematocrit effect, spot volume, and punch
localization
Haematocrit (Hct) may have an important role on the spot area and
in the matrix effect, therefore these parameters should be
evaluated. As the Hct levels change, the viscosity of the blood also
changes. The variability of viscosity leads to differences in flux and
diffusion properties of blood, which can directly impact the size of
the blood spots formed. The use of a non-cellulose paper as Bond
Elut DMS card may decrease the bioanalytical bias introduced as
a haematocrit variability, but it must be evaluated. According to a
previous study,[34] different Hct values: 30% to 55%, respectively,
were investigated to obtain a linear correlation between
haematocrit values and volume of blood. The haematocrit effect
is described by the following equation:
V Std
C corrected ¼ C observed
V Std þ bðHct
45Þ
where Ccorrected is the concentration after the correction for Hct
value, Cobserved is the found concentration before its correction,
Vstd is the blood volume at the standardized condition (45% Hct)
and b is the regression coefficient between blood volume and
haematocrit percentage.
The equation obtained between Vest and Hct is:
Vest = 20.93 + 0.1133 x (Hct-45) where b value is the slope
(0.1133) and Vstd the intercept (20.93).
Using uncorrecting concentration between the high and low Hct,
the error was approximately ±27%, while correcting values for Hct
percentage using Eqn (1), the bias were lowered to not more than
10% according to the Food and Drug Administration (FDA)
guidelines which request ±15%.[32]
The accuracy and precision of quantitation of LZD and CPR
extracted from a 9-mm diameter punch taken from a DBS sample
derived from 15, 25, and 40 μL aliquots of human blood were
investigated. For all examined blood samples, accuracy and
precision were less than the defined acceptance criteria of 15%.
Furthermore, compared to those from 25 μL, the difference
between the accuracy values for the extracts from 15 and 40 μL
were all less than 7.3%. This indicates that there is no notable
difference in distribution of analyte and blood across spots when
a 9-mm punch is taken from the centre of the blood spot.
Chromatographic condition
During the method development, four reversed-phase columns
Spherisorb ODS (150 × 4.6 mm I.D. 5 μm), Gemini C18
(150 × 4.6 mm I.D. 5 μm), Supelcosil LC-18 (150 × 4.0 mm I.D.
5 μm), and Kinetex EVO C18 (100 × 4.6 mm I.D. 2.6 μm) were tested
with different mobile phase compositions. LZD and CPR were not
separated with Spherisorb ODS and Supelcosil columns. Using
Gemini C18, the peak shapes for all the analysed compounds were
Table 1. Mean linear regression parameters for linezolid and ciprofloxacin obtained using seven calibrators (n = 6)
Compound Slope (± S.D) Intercept (± S.D)
Linezolid 0.093 ± (0.004) 0.0019 ± (0.0001)
Ciprofloxacin 0.182 ± (0.008) 0.0257 ± (0.0010)
Drug Test. Analysis 2017, 9, 1611–1619 Copyright © 2017 John Wiley & Sons, Ltd.
Drug Testing
and Analysis
very broad, which made their integration rather difficult. Good
separation of all analytes and internal standard was obtained using
a Kinetex EVO C18 column. This semi-porous silica column (packed
with core-shell particles) provided a shorter diffusion path for
analytes, minimised peak broadening at flow rates and had a high
permeability. Compared to the results obtained using other
columns, the application of Kinetex EVO C18 afforded higher
separation efficiency. Hence, it was possible to analyse LZD and
CPR more rapidly, with good resolution and peak shape. To achieve
the efficient separation of the analyzed compounds and the IS,
different mobile phases (methanol, acetonitrile, water with additive
ammonium acetate), first in isocratic mode, then with a gradient
programme, were tested. Isocratic separation was found the be
not suitable for the analysis of linezolid and ciprofloxacin due to
the overlapping of the drug peaks. It was determined that the
optimal mobile phase consisted of a mixture of methanol-
acetonitrile and 10 mM ammonium acetate buffer (pH 3.5) in a
(1) gradient elution programme. Gradient elution is typically used to
improve the peak separation and eliminate significant matrix
effects due to co-elution of the analyte and the endogenous plasma
components. Furthermore, to increase the peak shape, different
triethylamine (TEA) percentages in the range of 0.5 and 1% (v/v)
were added to the mobile phase. The optimum was achieved by
using 1% (v/v) of TEA. Under this condition, the peak shape of the
analytes was symmetric and the retention times of ciprofloxacin
and linezolid were well separated. To detect the analytes with high
sensitivity, each drug was monitored at its maximum wavelength.
By applying the chromatographic condition herein reported, the
total run time in this assay was 20 min and the retention times for
CRP, LZD were 7.1 min and 8.9 min, respectively. The retention time
for internal standard was 7.9 min. The retention times had
consistently excellent reproducibility of less than 0.16%. The
resolution of the developed HPLC-PDA method was satisfactory,
as all analytes were well base-line separated. The peaks were
symmetric with a symmetry factor As ≤1.34. All the evaluate
parameters were within the required range, as established in FDA
guidelines.[32]
Linearity
Linearity was evaluated using freshly prepared spiked matrix
samples in the range of 0.05–30 μg/mL for LZD and 0.02–
10 μg/mL for CPR. The correlation between LZD and CPR
concentration (x) and the ratios of the observed peak areas of
linezolid and ciprofloxacin to the internal standard (y) led to a
mean slope (±S.D.) of 0.093 (± 0.004) for LZD and 0.182
(± 0.008) for CPR, respectively, and a mean y intercept (±S.D.) of
0.0019 (± 0.0001) for LZD and 0.00257 (±0.00010) for CPR. Each
calibration curve consisted of a seven-calibrator concentration.
To describe the relationships between concentration of the
analytes and the detector response, a linear regression analysis
with weighting factors consisting in 1/x2 values was used. The
results are shown in Table 1.
2
r Range (μg/mL) Weighting LOQ (μg/mL) LOD (μg/mL)
2
0.9995 0.05–30 1/x 0.05 0.015
2
1615
0.9998 0.02–10 1/x 0.02 0.007
wileyonlinelibrary.com/journal/dta
 Drug Testing
and Analysis V. Ferrone et al.

Selectivity Table 3. Recoveries of linezolid and ciprofloxacin in dried blood spot

The analytical method should be able to differentiate the analytes Nominal Mean Measured Recovery RSD%
of interest and I.S. from endogenous components in the matrix or concentration concentration (%)
(μg/mL) ± S.D.
other components in the sample. Normally, absence of interfering
components is accepted where the response is less than 20% of LINEZOLID
the lower limit of quantification for the analytes and 5% for the 0.10 0.096 ± 0.006 96.7 6.25
internal standard. Six independent blank dried blood spot 1.00 0.97 ± 0.08 97.3 8.24
samples were analyzed and no interferences were observed at 25.00 24.31 ± 0.34 97.4 1.39
the retention time of the selected analytes and of the internal CIPROFLOXACIN
standard. Resolution values ranged from 1.87 (± 0.04) for CPR/I. 0.05 0.046 ± 0.004 92.4 8.64
S. to 2.33 (±0.06) for LZD/I.S. at the higher point of the calibration 0.75 0.71 ± 0.05 94.67 7.04
curve. A representative chromatogram of a blank dried blood 7.50 7.19 ± 0.44 95.88 6.11
spot sample is shown in Figure 2a, while a blank spiked with
the internal standard and both LZD and CPR are shown in
Figure 2c.
except for LLOQ which should be within 20% of the nominal value.
Precision, expressed as relative standard deviation percent (RSD%),
Limit of detection (LOD) and limit of quantification (LOQ) should be demonstrated within a single run (intra-day precision)
The limit of detection (LOD) was the spiked blood concentration and in different runs (inter-day precision) using the same runs
that had a signal-to-noise ratio > 3, while the limit of quantification and data for the demonstration of accuracy. RSD% should not
(LOQ) was considered the lowest standard of the calibration curve, exceed 15% for the QC samples, except for the LLOQ which should
determined by injecting a spiked sample six times with decreasing not exceed 20%. Intra-day precision and accuracy were evaluated
concentrations of the analytes. The LOQs of the method were by analyzing six times in the same day blank samples spiked with
0.02 μg/mL and 0.05 μg/mL for CPR and LZD, respectively, with the studied analytes at 3 different concentration levels (LLOQ 0.05
good precision (RSD ≤ 9.43 for LZD and ≤7.42 for CPR) and accuracy QCL 0.10, QCM 1.00, and QCH 25.00 μg/mL) for LZD and (LLOQ
(within 92.6 and 106.3 for LZD and 95.4 and 104.0 for CPR). The 0.02, QCL 0.05, QCM 0.75, and QCH 7.5 μg/mL) for CPR, respectively.
LODs were 0.008 μg/mL for CPR and 0.015 μg/mL for LZD. The For inter-day precision, the assays described above were repeated
results are shown in Table 2. six times over six different days. Table 2 displays the overview of
the intra-day (RSD%), inter-day (RSD%) and mean relative errors
of the methods. The intra- and inter-day accuracies were within
Precision and accuracy
96% and 106.2%. The intra- and inter-day (RSD%) were not more
The data for intra- and inter-day precision and accuracy obtained than 9.43%. These results indicated that the validated assay was
from QC samples at four different levels of analytes (LLOQ, LQC, precise, accurate and reproducible.
MQC, and HQC) are shown in Table 2. The QC samples were
analyzed against the calibration curve, and the obtained
Extraction recovery
concentrations were compared with the nominal value. The
accuracy and precision are the most important criteria for judging Recovery yield of linezolid and ciprofloxacin from blank DBSs
the performance of an analytical method. Accuracy was reported spiked with the analytes at the QCs concentrations was evaluated;
as percentage of the nominal value and evaluated for the values recoveries for linezolid were within 96.7% and 97.3% while for
of the QC samples obtained within a single run (intra-day accuracy) ciprofloxacin were within 92.4% and 95.88%. Recovery of LZD and
and in different runs (inter-day accuracy). The mean concentration CPR was evaluated by comparing the mean peak areas of different
should be within 15% of the nominal value for the QC samples, QC samples post-extracted with those obtained by spiking the

Table 2. Intra-day and inter-day accuracy and precision for linezolid and ciprofloxacin at the LLOQ and QCs value

Theoretical Intra-day (n = 6) Inter-day (n = 6)

concentration
(μg/mL) Mean Measured Accuracy (%) Precision Mean Measured Accuracy (%) Precision (RSD%)
concentration (RSD%) concentration
(μg/mL) ± S.D. (μg/mL) ± S.D.

LINEZOLID
0.05a 0.046 ± 0.003 92.6 6.49 0.053 ± 0.004 106.2 9.43
0.10 0.096 ± 0.004 96.3 4.15 0.095 ± 0.005 95.2 5.25
1.00 1.03 ± 0.06 103.45 5.82 1.05 ± 0.06 105.3 5.71
25.00 24.87 ± 0.57 99.48 2.29 25.4 ± 1.15 101.6 4.61
CIPROFLOXACIN
0.02a 0.019 ± 0.002 95.3 7.42 0.021 ± 0.03 104.0 7.17
0.05 0.048 ± 0.003 96.3 6.25 0.053 ± 0.005 106.2 9.43
0.75 0.74 ± 0.02 98.66 2.70 0.78 ± 0.07 104.3 8.97
7.50 7.53 ± 0.19 100.4 2.52 7.51 ± 0.43 100.1 5.72
1616

a
LLOQ concentration.

wileyonlinelibrary.com/journal/dta Copyright © 2017 John Wiley & Sons, Ltd. Drug Test. Analysis 2017, 9, 1611–1619
 Drug Testing
DBS-HPLC-PDA analysis of linezolid and ciprofloxacin and Analysis

Table 4. Stability of linezolid and ciprofloxacin in dried blood spot at Short-term stability in DBSs
different temperature
Spiked blank DBS were stored at room temperature and
20°C,
QCs Storage Mean Accuracy Precision +4°C and 35°C up to one month from the sampling date, then
Concentration temperature measured (%) (RSD%)
(μg/mL) (°C) concentrationa
subjected to the extraction treatment previously described.
(μg/mL) ± S.D. Stability studies showed no significant differences between room
temperature and refrigerated temperature. Nor were any
Linezolid significant differences shown when DBSs were stored at 37°C.
0.10
20 0.098 (±0.002) 98.00 2.04 These results, shown in Table 4, are very satisfactory due to the very
1.00
20 0.97 (±0.05) 97.00 5.15 low difference between DBSs stored at
20°C with the same
25.00
20 24.91 (±0.23) 99.64 0.93 sample stored at +35°C.
0.10 4 0.097 (±0.001) 97.00 1.03 The maximum loss of the two analytes being less than 7% is
1.00 4 0.99 (±0.03) 99.00 3.03 attributable to the water loss during the blood spot drying that
25.00 4 24.59 (±0.37) 98.36 1.50 minimizes most enzymatic processes responsible for the loss in
0.10 Room 0.097 (±0.003) 97.00 3.09 analyte compared to other matrices. Freeze–thaw stability of LZD
1.00 Room 0.98 (±0.06) 98.00 6.12 and CPR was determined by assaying the three QC samples in
25.00 Room 23.44 (±0.24) 93.76 1.01 triplicate over three freeze–thaw cycles.
0.10 35 0.095 (±0.003) 95.00 3.16 Three aliquots at each concentration were stored at
20° for 24 h
1.00 35 0.95 (± 0.04) 95.00 4.21 then thawed unassisted at room temperature. When completely
25.00 35 24.68 (±0.33) 98.70 1.33 thawed, the samples were refrozen under the same conditions for
Ciprofloxacin 24 h. No degradation was observed after three freeze–thaw cycles.
0.05
20 0.047 (±0.002) 94.00 4.26
0.75 -20 0.73 (±0.02) 97.33 2.74
7.50 -20 7.28 (±0.26) 97.07 3.57 Carry-over
0.05 4 0.046 (±0.003) 92.00 6.52
Carry-over should be addressed and minimised during method
0.75 4 0.71 (±0.06) 94.67 8.45
development. It was assessed by injecting blank samples after a
7.50 4 7.33 (±0.29) 97.73 3.96
calibration standard at the upper limit of quantification.32 The
0.05 Room 0.045 (±0.002) 90.00 4.44
carry-over of the blank sample following the high concentration
0.75 Room 0.73 (±0.07) 97.33 9.59
was <6% of the lower limit of quantification for LZD and CPR while
7.50 Room 7.24 (±0.15) 96.53 2.07
it was <2% for the I.S.
0.05 35 0.047 (±0.004) 94.00 8.51
0.75 35 0.69 (±0.02) 92.00 2.90
7.50 35 7.12 (±0.21) 94.93 2.95 Clinical application
a
Results are expressed as average of the analysis in triplicate. Before being applied to patients, a correlation between VDBS and
CDBS needs to be established. Hence, 9-mm punched discs of VDBS
and CDBS were analyzed for their CPR and LZD levels. The
post-extracted blank matrix with the analytes at the QCs levels of correlation coefficients (r2) of 0.964 for CPR and 0.951 for LINZ show
concentration. The. a very good comparability, indicating that finger-prick sampling is
results are reported in Table 3. The mean (±SD) extraction sufficient to measure both compound (Figure 5). In routine TDM,
recovery for IS was 93.36 ± 1.24% (n = 6), The use of an IS in the most of the drug levels are reported as plasma values. Therefore,
extraction procedure is crucial to compensate for variability in if DBS levels are reported as whole blood values, a correlation
extraction efficiency. In the present study, ulifloxacin was chosen between DBS and plasma is needed.
as the I.S. Ulifloxacin exhibits good absorbance at the wavelength To demonstrate the applicability of this assay and to
of the other analytes. It also displayed appropriate compare it against a validated method[33] with plasma matrix
chromatographic retention with its peak sufficiently separated from as the ‘gold standard’, matched plasma and DBS samples were
that of LZD and CPR (Figure 2c). analyzed.
1617

Figure 5. Concentrations of (a) CPR and (b) LZD determined from punched 9-mm DBS discs of CDBS versus VDBS.

Drug Test. Analysis 2017, 9, 1611–1619 Copyright © 2017 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/dta
 Drug Testing
and Analysis
Figure 6. Correlation between plasma concentrations, obtained using a validated method, of (a) CPR and (b) LZD, and DBS concentrations.
As shown in Figure 6, a good correlation between the
DBS and the plasma concentration was found. By applying
[CPR calculated plasma] = ([CPR DBS] – 0.02869)/ 1.417 and
[LZD calculated plasma] = ([LZD DBS] -0.3749) / 1.201 it is possible to
correlate the plasma concentration of the investigated analytes
with their DBS concentration for the TDM.
The proposed method was successfully applied for the
quantification of linezolid and ciprofloxacin in human dried
blood spot samples of patients in standard therapy with
intravenous LZD (Zyvox 600 mg twice daily) and CPR (ciproxin
400 mg twice daily) afferent to S.S. Annunziata Hospital (Chieti, Italy)
(Figure 2b). Written informed consent was obtained before blood
sampling.
Conclusions
A simple, precise, accurate HPLC-DAD method for the
quantification of linezolid and ciprofloxacin in human dried blood
spot sample was developed and fully validated. The proposed
method ensures satisfactory sensibility and specificity for the
therapeutic drug monitoring of both LZD and CPR. The main goal
of this work is the development of a method ensuring the
simultaneous determination of two analytes and the application
of a sample preparation with the DBS technique. Furthermore,
DBS disregards the need of blood samples, facilitating the shipping
and storage to prevent bacterial degradation. Additionally, high
efficiency and reproducibility using a common available DAD
detection indicate that the present method can be an important
analytical tool during routine determination of ciprofloxacin and
linezolid.
References
[1] A. McLuckie. Respiratory Diseases and its Management. Springer Verlag,
2009.
[2] American Thoracic Society. Infectious Diseases Society of America.
Guidelines for the management of adults with hospital-acquired,
ventilator-associated, and healthcare-associated pneumonia. Am. J.
Respir. Crit. Care Med. 2005, 171, 388.
[3] P. Gross, H. Neu, P. Aswapokee, C. Van Antwerpen, N. Aswapokee.
Deaths from nosocomial infections: Experience in a university
hospital and a community hospital. Am. J. Med. 1980, 68, 219.
[4] J. Chastre, J. Y. Fagon. Ventilator associated pneumonia. Am. J. Respir.
Crit. Care Med. 2002, 165, 867.
1618
[5] G. M. Eliopoulos, R. C. Moellering. in Antibiotics in Laboratory Medicine,
3rd Edn. The Williams & Wilkins Co., Baltimore, 1991.
wileyonlinelibrary.com/journal/dta Copyright © 2017 John Wiley & Sons, Ltd.
V. Ferrone et al.
[6] A. Marchese, G. C. Schito. The oxazolidinones as a new family of
antimicrobial agent. Clin. Microbiol. Infect. 2001, 7, 66.
[7] D. Livermore. Linezolid in vitro: mechanism and antibacterial spectrum.
J. Antimicrob. Chemoter. 2003, 51, 9.
[8] F. Pea, P. Viale, P. Cojutti, B. Del Pin, E. Zamparini, M. Furlanut.
Therapeutical drug monitoring may improve safety outcomes of
long-term treatment with linezolid in adult patients. J. Antimicr.
Therapy. 2012, 67, 2034.
[9] P. C. Sharma, A. Jain, S. Jain. Fluoroquinolone antibacterials: a review on
chemistry, microbiology and therapeutic prospects. Acta Pol. Pharm.
2009, 66, 587.
[10] A. Forrest, D. E. Nix, C. H. Ballow, M. C. Birmingham, J. J. Schentag.
Pharmacodynamics of Intravenous Ciprofloxacin in Seriously Ill
Patients. Antimicrob. Agents Chemother. 1993, 37, 1073.
[11] G. Hayem, P. X. Petit, M. Levacher, C. Gaudin, M. F. Kahn, J. J. Pocidalo.
Cytofluorometric analysis of chondrotoxicity of fluoroquinolone
antimicrobial agents. Antimicrob. Agents Chemoter. 1994, 38, 243.
[12] G. La Marca, F. Villanelli, S. Malvagia, D. Ombrone, S. Funghini,
M. De Gaudio, S. Fallani, M. I. Cassetta, A. Novelli, E. Chiappini,
M. De Martino, L. Galli. Rapid and sensitive LC-MS/MS method for the
analysis of antibiotic linezolid on dried blood spot. J. Pharm. Biomed.
Anal. 2012, 67-68, 86.
[13] J. Zander, B. Maier, M. Zoller, D. Teupser, M. Vogeser. Quantification
of linezolid in serum by LC-MS/MS using semiautomated sample
preparation and isotope dilution internal standardization. Clin.
Chem. Lab. Med. 2014, 52, 381.
[14] O. A. Phillips, M. F. Abdel-Hamid, N. A. Al-Hassawi. Determination of
linezolid in human plasma by LC–MS–MS. Analyst. 2001, 126, 609.
[15] G. Carlucci. Analysis of fluoroquinolones in biological fluids by high
performance liquid chromatography: Review. J. Chromatogr. A. 1998,
812, 343.
[16] V. Ferrone, M. Carlucci, P. Palumpo, G. Carlucci. Development and
validation of an MEPS-UHPLC-PDA method for determination of
ulifloxacin in human plasma and urine of patients with peripheral
arterial disease. J. Pharm. Biomed. Anal. 2016, 128, 313.
[17] M. Simon, N. Thuo, N. Karisa, A. Mutori, G. O. Kokwaro, K. Maitland.
Determination of ciprofloxacin in human plasma using high
performance liquid chromatography coupled with fluorescence
detection: Application to a population pharmacokinetics study in
children with severe malnutrition. J. Chromatogr. B 2011, 879, 146.
[18] J. Vella, F. Busuttil, N. Sammut Bartoloa, C. Sammut, A. Serracino-Inglott,
L. M. Azzopardi, G. LaFerla. A simple HPLC-UV method for the
determination of ciprofloxacin in human plasma. J. Chrom. B. 2015,
989, 80.
[19] V. Ferrone, M. Carlucci, P. Palumbo, G. Carlucci. Bioanalytical method
development for quantification of ulifloxacin, fenbufen and felbinac
in rat plasma by solid-phase extraction (SPE) and HPLC with PDA
detection. J. Pharm. Biomed. Anal. 2016, 123, 205.
[20] W. Li, F. L. Tse. Dried blood spot sampling in combination with LC-MS/
MS for quantitative analysis of small molecules. Biomed. Chromatogr.
2010, 24, 49.
[21] R. Guthrie, A. Suzi. A simple phenylalanine method for detecting
phenylketonuria in large population of newborn infants. Pediatrics.
1963, 32, 338.
[22] D. Blessborn, S. Romsing, Y. Bergqvist, N. Lindegardh. Assay for
screening for six antimalarial drugs and one metabolite using dried
Drug Test. Analysis 2017, 9, 1611–1619

DBS-HPLC-PDA analysis of linezolid and ciprofloxacin
blood spot sampling, sequential extraction and ion-trap detection.
Bioanalysis. 2010, 2, 1839.
[23] S. Aburuz, J. Millership, J. Mc Elnay. Dried blood spot liquid
chromatography assay for therapeutic drug monitoring of metformin.
J. Chromatogr. B. 2006, 832, 202.
[24] E. Kralj, J. Tronteli, T. Paji, A. Kristl. Simultaneous measurement of
imatinib, nilotinib and dasatinib in dried blood spot by ultra high
performance liquid chromatography tandem mass spectrometry.
J. Chromatogr. B. 2012, 903, 150.
[25] G. Lawson, E. Cocks, S. Tanna. Bisoprolol, ramipril and simvastatin
determination in dried blood spot using LC-HRMS for assessing
medication adherence. J. Pharm. Biomed. Anal. 2013, 83, 1.
[26] M. Hatami, E. Karimnia, K. Farhadi. Determination of salmeterol in dried
blood spot using ionic liquid based dispersive liquid-liquid
microextration coupled with HPLC. J. Pharm. Biomed. Anal. 2013, 85, 283.
[27] T. Panchal, N. Sponeer, M. Barfield. Ensuring the collection of high
quality dried blood spot samples across multisite clinical studies.
Bioanalysis. 2017, 9, 209.
[28] M. O’Mara, B. Hudson-Curtis, K. Olson, Y. Yueh, J. Dunn, N. Spooner. The
effect of hematocrit and punch location on assay bias during
quantitative bioanalysis of dried blood spot samples. Bioanalysis.
2011, 3, 2335.
Drug Test. Analysis 2017, 9, 1611–1619 Copyright © 2017 John Wiley & Sons, Ltd.
Drug Testing
and Analysis
[29] A. Sharma, S. Jaiswal, M. Shukla, J. Lal. Dried blood spots: Concepts,
present status, and future perspectives in bioanalysis. Drug Test. Anal.
2014, 6, 399.
[30] N. G. L. Jager, H. Rosing. Procedures and practices for the validation of
bioanalytical methods using dried blood spots: A review. Bioanalysis.
2014, 6, 2481.
[31] P. M. De Kesel, N. Sadones, S. Capiau, E. Lambert, C. P. Stove. Hemato-
critical issues in quantitative analysis of dried blood spot: Challenges
and solutions. Bioanalysis. 2013, 5, 2023.
[32] FDA Food and Drug Administration of the United States. Guidance for
Industry - Bioanalytical Method Validation. US Department of Health
and Human Services, Center for Drug Evaluation and Research and
Center for Veterinary Medicine, 2003.
[33] V. Ferrone, M. Carlucci, R. Cotellere, P. Raimondi, A. Cichella, L. Di Marco,
G. Carlucci. Development and validation of a fast micro-extraction by
packed sorbent UHPLC-PDA method for the simultaneous
determination of linezolid and ciprofloxacin from patients with
hospital-acquired pneumonia. Talanta. 2017, 164, 64.
[34] D. H. Vu, R. A. Koster, J. W. Alffenaar, J. R. Brouwers, D. R. Uges.
Determination of moxifloxacin in dried blood spots using LC-MS/
MS and the impact of hematocrit and blood volume.
J. Chromatogr. B 2011, 879, 1063.
1619
wileyonlinelibrary.com/journal/dta