Journal of Pharmaceutical and Biomedical Analysis 98 (2014) 327–333

Contents lists available at ScienceDirect

Journal of Pharmaceutical and Biomedical Analysis

journal homepage: www.elsevier.com/locate/jpba

Qualitative and quantitative analysis of PDE-5 inhibitors in

counterfeit medicines and dietary supplements by HPLC–UV
using sildenafil as a sole reference
Ida Fejős a , Gábor Neumajer b , Szabolcs Béni a,∗ , Péter Jankovics b,∗∗
a
Semmelweis University, Department of Pharmaceutical Chemistry, Hőgyes Endre u. 9, H-1092 Budapest, Hungary
b
National Institute for Quality and Organizational Development in Healthcare and Medicine, Directorate General of National Institute of Pharmacy,
Zrínyi u. 3, H-1051 Budapest, Hungary

a r t i c l e i n f o a b s t r a c t

Article history: Due to their popularity, medicinal products containing the phophodiesterase type 5 enzyme (PDE-5)
Received 27 March 2014 inhibitors sildenafil, vardenafil and tadalafil are often subject to counterfeiting. In addition, illicit herbal
Received in revised form 4 June 2014 dietary supplements adulterated with these substances or their analogs have appeared on the market
Accepted 8 June 2014
offering an easy and anonymous sale.
Available online 14 June 2014
This paper describes an analytical method for qualitative and quantitative screening of sildenafil,
vardenafil, tadalafil and 11 of their designer analogs in illegal erectile dysfunction products by high-
Keywords:
performance liquid chromatography with UV detection (HPLC–UV). Sildenafil served as a single external
Phosphodieserase-5 inhibitor analogs
Liquid chromatography
standard for both identification and quantification of all analytes. Relative retentions and reference UV
Counterfeit spectra were used for qualitative, and correction factors for quantitative analyses, respectively.
Correction factor The separation was performed on a Kinetex C18 reverse-phased column at 25 ◦ C using gradient elution.
Herbal dietary supplements Mobile phase A consisted of 200 mM ammonium acetate solution while mobile phase B was a 1:1 (v/v)
mixture of methanol and acetonitrile with a flow rate of 0.5 ml/min and injection volume of 5 ␮l. Detection
wavelength was set to 290 nm.
The method was validated in accordance with the appropriate guideline of the International Conference
on Harmonization (ICH) in terms of specificity, selectivity, precision, linearity, limit of quantitation, limit
of detection, accuracy, robustness and stability, and was successfully applied to the analysis of natural
dietary supplements and herbal remedies with an indication for enhanced male sexual potency. The
proposed method offers a cheap and simple alternative to LC–MS screening used by control laboratories
for routine analysis of suspicious products.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction
The phosphodiesterase-5 enzyme (PDE-5) inhibitor sildenafil
(SIL) was developed for the treatment of cardiovascular disease
(Revatio® , using in therapy of pulmonary arterial hypertension)
and was also found to be effective in erectile dysfunction disorders
(ED) therapy. SIL (Viagra® ) was the only PDE-5 inhibitor approved
for the treatment of ED until vardenafil (Levitra® , VAR) and tadalafil
(Cialis® , TAD) were also approved in 2003 by the FDA [1].
Erectile dysfunction is a highly prevalent inability to achieve
and maintain adequate erection and sexual performance. As the
∗ Corresponding author. Tel.: +36 1 217 0891; fax: +36 1 217 0891.
∗∗ Corresponding author.
E-mail addresses: beni.szabolcs@pharma.semmelweis-univ.hu,
beniszabi@gmail.com (S. Béni), jankovics.peter@gyemszi.hu (P. Jankovics).
PDE-5 is responsible for the hydrolysis of cGMP in the penis, inhib-
iting the enzyme results in vasodilatation and penile blood flow
maintenance due to the elevated level of cGMP [2]. Owing to the
adverse effects such as headache, facial flushing, nasal conges-
tion, back pain and visual disorders [3], and its interactions with
nitrates or ␣-blockers leading to hypotension or syncope [4], PDE-
5 inhibitors are under medical doctor’s prescription. Since PDE-5
inhibitors are contraindicated in many patients, alternative thera-
pies for ED should be taken into account. Furthermore, the taboo
associated with sexual disorders facilitated the increase of cheap,
easy-to-access dietary supplement consumption, supported by the
internet. There is a widespread belief that dietary supplements and
herbal products offer a safe, side effect free and healthy alternative
to medicinal products containing synthetic ingredients. However,
many of these herbal products were found to contain approved
PDE-5 inhibitors, such as SIL, VAR, TAD or unapproved designer
analogs with subtle changes in their structures [5–8] (see Fig. 1).

http://dx.doi.org/10.1016/j.jpba.2014.06.010
0731-7085/© 2014 Elsevier B.V. All rights reserved.
328 I. Fejős et al. / Journal of Pharmaceutical and Biomedical Analysis 98 (2014) 327–333
Fig. 1. Chemical structures of the PDE-5 inhibitors examined.
Adulteration with PDE-5 inhibitors and their analogs poses serious
health risks in the case of patients who turn to alternative ther-
apies. Further health issues are also evoked by the lack of safety
and toxicity data of the numerous unapproved analogs. Due to the
popularity of PDE-5 inhibitor drugs, their counterfeit pharmaceu-
tical preparations have appeared in various websites offering an
easy and anonymous sale. Therefore, official control laboratories
need simple, easy-to-use, inexpensive routine methods to screen
and determine the illicit adulterants in herbal dietary supplements
and in suspicious pharmaceutical preparations.
Until now, several authors have reported methods for screening
and confirmation of synthetic PDE-5 inhibitors in illicit sex-
ual performance enhancer products, including HPLC–MS [1,9,10],
HPLC–MS/MS [11–14], UHPLC [15], and UHPLC–MS/MS [2,16]. Even
though these methods allowed unambiguous identification of the
analytes, quantification required the use of expensive reference
standards. In order to overcome this practical difficulty, Poplawska
et al. used charged aerosol detection (CAD) [17] to determine the
amount of PDE-5 inhibitor analogs in herbal dietary supplements
using SIL and TAD as the only reference substances. However, an
independent off-line analytical technique (TOF-MS) was necessary
to identify the compounds.
Realizing that an easy and cost-effective method capable of both
identifying and quantifying PDE-5 inhibitors and their designer
analogs in herbal matrices would facilitate the work of control
laboratories responsible for screening illicit medicinal products
and food supplements, we aimed at designing a simple HPLC–UV
method for this purpose. SIL was used as the only external standard
to overcome the repeated use of expensive references of the analogs
in routine analysis. The method was fully validated based on the
guidelines of the International Conference on Harmonization (ICH).
2. Materials and methods
2.1. Standards and reagents
Reference standards of sildenafil citrate, tadalafil and vardenafil
dihydrochloride trihydrate were provided by Intas Pharmaceuticals
Ltd. (Ahmedabad, India), Eli Lilly and Co. (Indianapolis, IN, USA) and
Bayer AG (Leverkusen, Germany), respectively. Certified reference
standards of homosildenafil (HS), hydroxyhomosildenafil (HHS),
thiosildenafil (TS), thiohomosildenafil (THS), N-desethylvardenafil
(NDV), pseudovardenafil (PV) and acetildenafil (ACE)) were
purchased from Toronto Research Chemicals (Toronto, ON,
Canada) while hydroxythiohomosildenafil (HTHS), dimethylsilde-
nafil (DMS), hydroxyvardenafil (HV) and thiodimethylsildenafil
(TDS) were obtained from TLC PharmaChem Inc. (Vaughan, ON,
Canada).
Illicit herbal dietary supplement samples used in this work were
confiscated and sent for analysis to the laboratory of the National
Institute of Pharmacy, Budapest by the National Tax and Customs
Office.
Acetonitrile (ACN), methanol (MeOH) (both of gradient purity
grade) and ammonium acetate of analytical grade were all
purchased from Merck KGaA. (Darmstadt, Germany). Acetic
acid (puriss.) was obtained from Sigma–Aldrich GmbH (Seelze,
Germany). Water was produced by a Millipore Elix3 (Billerica, MA,
USA) water purifying system.
2.2. Instrumentation
HPLC experiments and validation were performed on an Agilent
1200 HPLC system equipped with a diode array detector (Agilent
Technologies, Santa Clara, CA, USA). A Kinetex C18 100 Å 2.6 ␮m
100 mm × 4.6 mm reverse-phased column (Phenomenex, Torrance,
CA, USA) was used. Column temperature was maintained at 25 ◦ C.
200 mM ammonium acetate solution was used as mobile phase
A. Mobile phase B consisted of a mixture of equal volumes of MeOH
and ACN. The gradient program of the optimized method was as
follows: 0–9 min 40–50% B; 9–17 min 50–80% B; 17–20 min 80%
B; 20–20.5 min 80–40% B; 20.5–25 min 40% B. The flow rate of
the mobile phase was 0.5 ml/min. Autosampler temperature was
set to 20 ◦ C. The injection volume was 5 ␮l. The detection of the
measurements was performed at 290 nm.
2.3. Standard and sample preparation
2.3.1. Stock and working solutions
Stock solutions were prepared by dissolving precisely weighed
amounts of each studied substance in mobile phase B to obtain
a concentration of about 1 mg/ml. Working solutions used in the
 I. Fejős et al. / Journal of Pharmaceutical and Biomedical Analysis 98 (2014) 327–333
validation study were obtained by dilution of the stock solutions to
the desired concentration with mobile phase B.
2.3.2. External standard solution
A 0.5 mg/ml solution of SIL in mobile phase B was prepared and
diluted 25-fold with mobile phase B (20 ␮g/ml SIL).
2.3.3. System suitability solution
The system suitability solution containing 20 ␮g/ml SIL and
20 ␮g/ml VAR was prepared from a 0.5 mg/ml SIL solution and
0.5 mg/ml VAR solution applying a 25-fold dilution.
2.3.4. Sample preparation
Illicit herbal dietary supplements received for PDE-5 screening
were either tablets or capsules. The powdered tablets, or the con-
tent of the capsules were subjected to analysis. For each sample, a
quantity equivalent to the average dosage unit was agitated with
mobile phase B for 30 min at a rate of 300 rpm, sonicated for 5 min
and centrifuged for 10 min applying 2400 × g. An aliquot of the clear
supernatant was diluted 25-fold with mobile phase B.
2.4. Method validation
The method was validated in accordance with the correspond-
ing ICH guideline [18] in terms of specificity, selectivity, precision,
linearity, limit of detection (LOD), limit of quantitation (LOQ), accu-
racy, robustness and stability.
2.4.1. Specificity and selectivity
Selectivity was examined by comparing the chromatograms of a
blank solution (mobile phase B), various herbal matrices and solu-
tions spiked with a mixture of the 14 compounds. Chromatograms
were checked for interfering peaks and sufficient separation of the
analytes.
2.4.2. Precision
Method precision was evaluated in terms of the relative
standard deviation percentage (RSD%) of the retention times and
the peak areas of the 14 examined PDE-5 inhibitors. The concen-
trations were kept constant at 3 concentration levels (1 ␮g/ml,
10 ␮g/ml, 100 ␮g/ml) during the six parallel measurements.
2.4.3. Linearity
Linearity was studied within the concentration range of
1–100 ␮g/ml for all the 14 compounds. Each concentration was
injected three times. Calibration curves for each analyte were estab-
lished with 7 concentrations. The regression line was calculated by
the method of least squares.
2.4.4. LOD and LOQ
LOD of PDE-5 inhibitors was determined as the concentration
yielding a signal three times the noise of the baseline, while the
LOQ was calculated as ten times the noise.
2.4.5. Accuracy
To determine the accuracy of the method, recovery measure-
ments were performed at three concentration levels (2.5 ␮g/ml,
25 ␮g/ml, 75 ␮g/ml) for all the studied compounds. 20 ␮g/ml SIL
was used as external standard. The experiment was conducted
three times for each concentration and the average percentage
recovery was determined for each compound. The recovery was
calculated using the correction factors determined during method
optimization, the concentration of the SIL external standard and
the peak areas of the compounds.
To evaluate matrix effects, accuracy measurements were also
conducted in the presence of various herbal matrices for TAD and
329
VAR. Three dietary supplements of different labeled composition
were subjected to the sample preparation procedure described
under Section 2.3.4. The extracts were then spiked with known
amounts of TAD and VAR and recovery of the 2 compounds were
calculated.
2.4.6. Robustness
As response functions, relative retentions and relative peak
areas (with reference to SIL) were observed. Three variables
(methanol-to-acetonitrile ratio, ammonium acetate concentration
in the mobile phase, and column temperature) which were deemed
to affect the results were investigated at the upper and lower lev-
els with regard to the optimum (see Table S1 for the full-factorial
experimental design in supplementary material).
2.4.7. Stability
In order to check the stability, stock solutions were kept in
refrigerator (2–8 ◦ C) for seven days. Clarity of the solutions was
examined visually and chromatograms obtained from the freshly
prepared and stored solutions were compared. A similar procedure
was carried out with a working solution which comprised a mixture
of all the analytes at 10 ␮g/ml and was stored at room temperature
for 24 h.
3. Results and discussion
3.1. Choice of the target molecules
The target group of the substances to be examined contained
SIL, VAR and TAD as the active substances of approved medicinal
products. Most of the analogs were chosen from those previously
detected in illicit medicinal products and dietary supplements ana-
lyzed by the laboratory of the National Institute of Pharmacy. In
addition, some compounds were purchased to represent a group
of molecules with a structure characteristic to PDE-5 inhibitor
analogs. Thus, the 14 analytes covered a relatively wide range of
chemical structures.
3.2. Method optimization
During method optimization the diversity and complexity of
dietary supplements and herbal remedies should be taken into
account. Furthermore, the different solubility of the analytes
required the use of a universal solvent which allowed complete
extraction of the compounds to be determined from the various
products. According to our experiments, most of the investigated
compounds were soluble in methanol. TAD was the only exception
which had a poor solubility in MeOH, but was freely soluble in ACN.
As a compromise, a 1:1 (v/v) mixture of MeOH and ACN was found
to be an acceptable solvent mixture for all the studied compounds
and it was used for sample extraction.
Based on their chemical structures, the compounds were clas-
sified into 5 groups: sildenafil-type, thiosildenafil-type, vardenafil-
type compounds, acetildenafil and tadalafil (for the structures see
Fig. 1). UV spectra are similar within each group but show sig-
nificant differences between the groups (see Fig. S1 in online
supplementary data). The differences between UV spectra can
aid the adequate identification of the compounds, but also raise
difficulties for a single-wavelength detection method. Therefore,
290 nm was found to be the best compromise at which all the
compounds have significant absorbance close to an absorption
maximum. Vardenafil analogs do not show real maxima but all
derivatives exhibit absorbance. In the case of low adulterated quan-
tities of PDE-5 inhibitors 230 nm may provide a better signal to
noise ratio, however, herbal components would be detrimental to
the chromatographic profile.
330 I. Fejős et al. / Journal of Pharmaceutical and Biomedical Analysis 98 (2014) 327–333

Fig. 3. Comparison of the chromatograms applying 200 mM ammonium acetate (pH

7.0) as mobile phase A and as mobile phase B: (a) acetonitrile or (b) methanol.

Fig. 2. Effect of pH on the chromatograms applying mixtures of ammonium acetate

and acetic acid as mobile phase A, and (a) acetonitrile or (b) methanol as mobile
phase B. system suitability criterion was based on the resolution of these
two peaks. The thiosildenafil derivatives were the most retained
forming a separate group at the end of the chromatogram.
As for the chromatographic separation, reversed phase and
HILIC type columns were tested. Finally, C18 stationary phase was
chosen, because it showed considerably higher separation poten- 3.3. Method validation
tial compared to bare silica. The use of core–shell particles was
preferred in order to achieve improved performance as a result of 3.3.1. Specificity and selectivity
the short diffusion path length. A successful separation of the standards was achieved by the
Owing to subtle differences in the structures, separation of all optimized method (Fig. 4). The blank (mobile phase B) showed
the compounds under isocratic conditions was not feasible. There- no interference. To evaluate any matrix-derived interference, 3
fore, various gradient elution programs were tried using ACN, different herbal dietary supplements of known composition were
MeOH and their mixtures as the organic phase. Because all the extracted. No interfering peaks were detected in the obtained chro-
examined compounds (apart from TAD) contain one or more basic matograms.
nitrogen centers, the pH of the mobile phase had to be optimized.
The effect of pH was studied in the range from 2.7 through 7.0 3.3.2. Precision
using mixtures of ammonium acetate and acetic acid. Fig. 2 shows The results are summarized in Table 1. RSD% calculated for peak
chromatograms as a function of pH of mobile phase A. Regardless areas was below 1.0% for all compounds with ACE being the only
of the organic phase used, increasing the pH of mobile phase A exception. Maximum RSD% of retention times was as low as 0.31%.
resulted in an improvement of peak shape and resolution. For fur-
ther method optimization pH 7.0 (200 mM ammonium acetate) was
applied. Comparing the two systems at pH 7.0, neither ACN, nor
MeOH alone as mobile phase B were able to separate all the com-
pounds (see Fig. 3). Using ACN as mobile phase B, ACE and NDV as
well as SIL and VAR co-eluted, while the use of MeOH resulted in
a co-elution of VAR and HS. The elution order was also altered, the
most significant being the case of ACE and TAD. The latter could be
explained by the different solubility of these substances in ACN and
MeOH as compared to the other compounds. The poor separation
with both 100% ACN and 100% MeOH suggested the use of their
mixtures as mobile phase B. The most suitable system contained
a 1:1 (v/v) mixture as mobile phase B, applying 200 mM ammo-
nium acetate as mobile phase A using a gradient elution program
as described in Section 2. It is noted that low salt concentration
in mobile phase A decreased the resolution between ACE and NDV.
Elevated column temperatures showed a negative effect on the sep-
aration of all compounds, therefore, no column heating was applied
and room temperature (25 ◦ C) was chosen. 5 ␮l injection volume
and 0.5 ml/min flow rate was found to be optimal. Fig. 4 shows a
representative chromatogram obtained with the optimized system.
SIL eluted in the middle of the separation window, which was Fig. 4. The chromatogram obtained with the optimized system: 200 mM ammo-
advantageous for qualitative and quantitative analysis, as it was nium acetate (pH 7.0) as mobile phase A, methanol:acetonitrile 1:1 as mobile phase
used as external standard. SIL and VAR eluted consecutively, so the B, 0.5 ml/min, 25 ◦ C, 5 ␮l injection volume, 290 nm.
 I. Fejős et al. / Journal of Pharmaceutical and Biomedical Analysis 98 (2014) 327–333 331

Table 1
Precision, linearity, accuracy, LOD and LOQ of the 14 compounds.

Precision: (RSD%) Peak area Retention time Linearity:

Compound name 1 ␮g/ml 10 ␮g/ml 100 ␮g/ml 1 ␮g/ml 10 ␮g/ml 100 ␮g/ml Regression line R2

ACE 2.15 0.47 0.88 0.10 0.02 0.31 y = 18.248x − 25.168 0.9981
TAD 0.86 0.25 0.49 0.10 0.01 0.31 y = 10.203x − 4.2874 0.9994
HHS 0.70 0.22 0.51 0.06 0.01 0.21 y = 9.3212x − 2.4598 0.9997
DMS 0.46 0.96 0.48 0.01 0.05 0.09 y = 9.1972x + 0.4119 0.9994
SIL 0.75 0.19 0.52 0.04 0.01 0.14 y = 9.1930x − 3.6671 0.9993
HS 0.24 1.07 0.52 0.02 0.04 0.06 y = 10.071x + 0.9192 0.9993
NDV 1.34 0.59 0.57 0.03 0.10 0.17 y = 4.3081x − 1.0492 0.9995
HV 1.39 0.74 0.49 0.02 0.08 0.13 y = 4.2432x + 0.0797 0.9993
VAR 1.25 0.33 0.14 0.03 0.01 0.12 y = 4.1344x − 3.2226 0.9982
PV 1.23 0.37 0.51 0.02 0.01 0.09 y = 4.7694x − 1.2493 0.9997
HTHS 1.18 1.11 0.51 0.01 0.03 0.04 y = 6.032x + 0.1290 0.9994
TDS 1.30 0.17 0.51 0.02 0.01 0.08 y = 6.438x − 2.3460 0.9996
TS 0.77 1.14 0.55 0.01 0.02 0.04 y = 6.9632x + 0.1826 0.9994
THS 0.93 0.22 0.50 0.02 0.01 0.09 y = 7.3974x − 3.2034 0.9993

Compound name Accuracy: recovery (%) LOQ/LOD:

2.5 ␮g/ml 25 ␮g/ml 75 ␮g/ml Sample A Sample B Sample C LOQ (␮g/ml) LOD (␮g/ml)

ACE 94.0 95.4 98.5 0.060 0.020

TAD 104.7 103.1 100.1 96.7 95.9 95.7 0.050 0.017
HHS 107.3 106.5 101.7 0.025 0.008
DMS 96.2 96.9 105.6 0.030 0.010
SIL 101.5 102.7 100.0 0.030 0.010
HS 96.5 95.7 104.1 0.025 0.008
NDV 91.1 90.7 100.4 0.065 0.022
HV 95.3 95.7 104.8 0.065 0.022
VAR 98.7 97.1 96.7 94.0 93.7 95.6 0.065 0.022
PV 102.9 103.1 98.2 0.065 0.022
HTHS 90.8 92.6 101.3 0.060 0.020
TDS 100.4 101.1 97.8 0.060 0.020
TS 90.8 92.6 101.2 0.065 0.022
THS 99.9 99.4 96.0 0.065 0.022

Values for the reference compound sildenafil are listed in boldface.

3.3.3. Linearity was concluded that particular care should be taken during prepa-
The calculated regression lines showed good linearity with ration of mobile phase B.
determination coefficients (r2 ) higher than 0.998 for each analyte
(see Table 1). The slopes of the regression lines were used to calcu- 3.3.7. Stability
late the correction factors for quantitative analysis. Working and stock solutions were stored at room temperature
for 24 h and in refrigerator (2–8 ◦ C) for 7 days. All solutions were
3.3.4. LOD and LOQ found to be stable under these conditions.
The LOD was in the range of 0.008–0.022 ␮g/ml, while LOQ was
between 0.025 and 0.065 ␮g/ml. See Table 1 for the exact values. 3.4. Qualitative and quantitative analysis

3.4.1. Qualitative analysis

3.3.5. Accuracy
The accuracy data were in the acceptable 90–110% recovery
range (Table 1) at each concentration level and for each compound.
To evaluate any matrix effect, known amounts of TAD and VAR
were added to extracts of three different herbal dietary supple-
ments of different compositions. No significant change in recovery
values was observed indicating that the method was accurate in
the presence of herbal matrix components (see Table 1).
3.3.6. Robustness
The choice of the three factors examined in the robustness test
was based on our experience during method optimization. Upper
and lower levels for each factor were determined with respect to
practical aspects (e.g. realistic errors in column temperature or in
mobile phase composition during preparation).
Assessment of the results revealed that the MeOH:ACN ratio
was critical in terms of relative retention values while relative peak
areas were not affected (see Fig. S2 in online supplementary data).
The effect of ammonium acetate concentration and column tem-
perature appeared to be less significant, even though resolution
between ACE and NDV decreased at higher temperatures. Thus, it
Our approach to use SIL as the only external standard for both
identification and quantification of different PDE-5 inhibitors was
based on the idea that an HPLC–UV system is capable of delivering
orthogonal data, which – when combined in an appropriate way
– can be characteristic for a molecule within a certain group. In
this case, relative retention of any of the chosen compounds with
reference to SIL and comparison of the UV spectrum of the same
compound to a reference spectrum provide sufficient information
to identify the analyte. In order to render the method reliable it
was essential to ensure that (1) the robustness of the method is
well studied and (2) the acceptance criteria for a compound to
be declared as identified are carefully established. Relative reten-
tion intervals are presented in Table 2. From the robustness data
it appeared that for certain analytes (e.g. ACE, NDV, HV) relative
retention to SIL is more sensitive to small changes in experimen-
tal conditions than for others (e.g. VAR, DMS, HS). However, it was
considered to be reasonable to use an interval of the same width
for all analytes. The ±0.15 min range was found to be an accept-
able compromise. Even though there are some overlaps between
the intervals, this leads to ambiguity only in the case of TDS and TS
where UV spectra are similar.
332 I. Fejős et al. / Journal of Pharmaceutical and Biomedical Analysis 98 (2014) 327–333

Table 2 qualitative analysis and correction factors for quantitative calcu-

The calculated relative retention intervals and the correction factors of the 14
lations the sample contains approximately 18 mg thiodimethyl-
components.
sildenafil/dosage unit. This result has also been confirmed using
Compound name Relative retention Correction factor MS detection as an independent technique.
(referring to sildenafil)
In the case of further adulterated dietary supplements, the fol-
ACE 0.685–0.715 0.5 lowing compounds were detected and quantified: Sample A (see
NDV 0.705–0.735 2.1 Fig. S4A) contained 60 mg HTHS and 67 mg THS per capsule, Sam-
TAD 0.755–0.785 0.9
ple B (see Fig. S4B) contained 9 mg PV per capsule, Sample C (see
HV 0.815–0.845 2.2
HHS 0.865–0.895 1 Fig. S4C) contained 9 mg TAD and 33 mg SIL per capsule, Sample D
DMS 0.935–0.965 1 (see Fig. S4D) contained 65 mg HTHS and 20 mg THS per capsule.
SIL 0.985–1.015 1 Sample E (see supporting Fig. S4E) contained 18 mg HV and
VAR 1.025–1.055 2.1
a PDE-5 analog not included in our sample set. The retention
HS 1.065–1.095 0.9
PV 1.165–1.195 1.9 of this compound could not be matched with any of the com-
HTHS 1.185–1.215 1.5 pounds examined, however the UV data suggested a vardenafil
TDS 1.245–1.275 1.4 analog. Using mass spectrometric detection in conjunction with
TS 1.265–1.295 1.3 our chromatographic system, this compound was identified as
THS 1.335–1.365 1.2
hydroxythiovardenafil (HTV). Using the response factor of the par-
ent compound vardenafil (2.1) an approximate amount of 20 mg
HTV was found in the adulterated sample. The latter case clearly
Although it is realized that this approach is not suitable for iden-
shows the advantage of our method, namely the fast and robust
tification of molecules other than those of the target group, it could
identification and approximate quantitation of PDE-5 inhibitors in
be used to tentatively predict the general structure of any new PDE-
adulterated herbal dietary supplements.
5 inhibitor based on its relative retention to SIL and its UV spectrum.
Of course, in such cases as well as in case of ambiguity, other meth-
ods should be used to confirm the structure. Because the mobile 4. Conclusion
phase used in this method is MS compatible, direct coupling of an
MS to the HPLC–UV system may easily solve this problem. In this paper, a fully validated HPLC–UV method is described
enabling the simultaneous separation of sildenafil, vardenafil,
3.4.2. Quantitative analysis tadalafil and their 11 unapproved designer analogs. The optimized
In HPLC assays, the lack of reference substance for a compound method allows a rapid screening and determination of PDE-5
to be determined implies that its amount can be estimated by the inhibitor content of illegal erectile dysfunction products utilizing
use of a different external standard. This is a common practice in the only sildenafil as a widely available external standard for the quali-
European Pharmacopeia especially in tests for related substances. tative and quantitative analysis, for the first time. The identification
In this case, the peak area of the analyte is compared to that of is based on relative retentions and UV spectra, furthermore, it is
the external standard. To correct for any significant differences in possible to perform a tentative preliminary identification of unde-
response factors, the peak area of the substance to be determined clared substances in complex herbal matrices with the developed
is multiplied by a correction factor. method. The quantification is easily achievable applying the cor-
Correction factors used in this method were calculated as fol- rection factors, the SIL external standard concentration and sample
lows. For each compound, the slope of the regression lines obtained dilution.
in the linearity study was divided by the slope obtained for SIL and The optimized and validated method has been successfully
the reciprocal of the fraction was taken as the corresponding cor- applied to the analysis of real samples of natural dietary supple-
rection factor (listed in Table 2). For practical purposes, the results ments and herbal remedies. Due to the easy use and the low costs
were rounded to one decimal place. of the developed method, it could serve as a routine procedure in
control laboratories.
3.4.3. System suitability
Because the key identification parameter is the relative reten- Acknowledgments
tion with respect to SIL, it is essential to verify the suitability of
the method prior to each analysis. System suitability criteria were The financial support from OTKA PD 109373 is highly appreci-
based on the results obtained during method validation. As a result, ated. Sz. Béni thanks the Hungarian Academy of Sciences for the
the following requirements should be met: in the chromatogram financial support under the János Bolyai Research Scholarship.
obtained with the system suitability solution, resolution between
SIL and VAR must be above 3.25 and relative retention of VAR Appendix A. Supplementary data
should be in the range of 1.036–1.040. Furthermore, the RSD for
retention time and peak area of SIL should not exceed 0.5% and Supplementary material related to this article can be found, in
1.0%, respectively, as calculated from six subsequent injections of the online version, at http://dx.doi.org/10.1016/j.jpba.2014.06.010.
external standard solution.
References
3.4.4. Real sample analysis
In order to prove the applicability of the developed method, the [1] S.R. Gratz, C.L.K.A. Flurer, Analysis of undeclared synthetic phosphodiesterase-
described procedure has been applied to real samples in various 5 inhibitors in dietary supplements and herbal matrices by LC–ESI-MS and
LC–UV, J. Pharm. Biomed. Anal. 36 (2004) 525–533.
cases including dietary supplements and herbal remedies for the [2] P. Proença, C. Mustra, M. Marcos, J.M. Franco, F. Corte-Real, D.N. Vieira,
treatment of erectile dysfunction. As a result of the screening, vari- Validated UPLC–MS/MS assay for the determination of synthetic phosphodie-
ous samples were found to contain SIL, thiosildenafil-type analogs sterase type-5 inhibitors in postmortem blood samples, J. Forensic Leg. Med.
20 (2013) 655–658.
(HTHS, TS, THS, TDS), vardenafil analogs (PV, HV) and tadalafil. [3] A.W. Shindel, 2009 update on phosphodiesterase type 5 inhibitor therapy part
Supplementary Fig. S3 illustrates the confirmatory identifica- 2: updates on optimal utilization for sexual concerns and rare toxicities in this
tion of the adulterants. Applying relative retention, UV spectra for class (CME), J. Sex. Med. 6 (2009) 2352–2364.
 I. Fejős et al. / Journal of Pharmaceutical and Biomedical Analysis 98 (2014) 327–333
[4] S. Gur, P.J. Kadowitz, A. Gokce, S.C. Sikka, U. Lokman, W.J.G. Hellstrom, Update
on drug interactions with phosphodiesterase-5 inhibitors prescribed as first-
line therapy for patients with erectile dysfunction or pulmonary hypertension,
Curr. Drug Metab. 14 (2013) 265–269.
[5] S. Singh, B. Prasad, A.A. Savaliya, R.P. Shah, V.M. Gohil, A. Kaur, Strategies for
characterizing sildenafil, vardenafil, tadalafil and their analogues in herbal
dietary supplements, and detecting counterfeit products containing these
drugs, Trends Anal. Chem. 28 (2009) 13–28.
[6] D.N. Patel, L. Li, C.L. Kee, X. Ge, M.Y. Low, H.L. Koh, Screening of synthetic
PDE-5 inhibitors and their analogues as adulterants: analytical techniques and
challenges, J. Pharm. Biomed. Anal. 87 (2014) 176–190.
[7] B.J. Venhuis, D. de Kaste, Towards a decade of detecting new analogues of
sildenafil, tadalafil and vardenafil in food supplements: a history, analytical
aspects and health risks, J. Pharm. Biomed. Anal. 69 (2012) 196–208.
[8] P. Jankovics, S. Lohner, A. Darcsi, J. Németh-Palotás, S. Béni, Detection and struc-
ture elucidation of hydroxythiovardenafil as an adulterant in a herbal dietary
supplement, J. Pharm. Biomed. Anal. 74 (2013) 83–91.
[9] X. Zhu, S. Xiao, B. Chen, F. Zhang, S. Yao, Z. Wan, D. Yang, H. Han, Simultaneous
determination of sildenafil, vardenafil and tadalafil as forbidden components
in natural dietary supplements for male sexual potency by high-performance
liquid chromatography–electrospray ionization mass spectrometry, J. Chro-
matogr. A 1066 (2005) 89–95.
[10] Y. Cai, T.G. Cai, Y. Shi, X.L. Cheng, L.Y. Ma, S.C. Ma, R.C. Lin, W. Feng, Simulta-
neous determination of eight PDE5-IS potentially adulterated in herbal dietary
supplements with TLC and HPLC–PDA-MS methods, J. Liq. Chromatogr. Relat.
Technol. 33 (2010) 1287–1306.
[11] E.S. Lee, J.H. Lee, K.M. Han, J.W. Kim, I.S. Hwang, S. Cho, S.Y. Han, J. Kim, Simul-
taneous determination of 38 phosphodiestrase-5 inhibitors in illicit erectile
dysfunction products by liquid chromatography–electrospray ionization-
tandem mass spectrometry, J. Pharm. Biomed. Anal. 83 (2013) 171–178.
333
[12] P. Zou, S.S.Y. Oh, P. Hou, M.Y. Low, H.L. Koh, Simultaneous determi-
nation of synthetic phosphodiesterase-5 inhibitors found in a dietary
supplement and pre-mixed bulk powders for dietary supplements using high-
performance liquid chromatography with diode array detection and liquid
chromatography–electrospray ionization tandem mass spectrometry, J. Chro-
matogr. A 1104 (2006) 113–122.
[13] Y. Zhang, Z. Huang, L. Ding, H. Yan, M. Wang, S. Zhu, Simultaneous determi-
nation of yohimbine, sildenafil, vardenafil and tadalafil in dietary supplements
using high-performance liquid chromatography–tandem mass spectrometry,
J. Sep. Sci. 33 (2010) 2109–2114.
[14] H. Lee, B.J. Lee, A novel approach to simultaneous screening and con-
firmation of regulated pharmaceutical compounds in dietary supplements
by LC/MS/MS with an information-dependent acquisition method, Food
Addit. Contam. Part A: Chem. Anal. Control Expo Risk Assess. 28 (2011)
396–407.
[15] P.Y. Sacré, E. Deconinck, P. Chiap, J. Crommen, F. Mansion, E. Rozet, P. Courselle,
J.O. De Beer, Development and validation of a ultra-high-performance liquid
chromatography–UV method for the detection and quantification of erectile
dysfunction drugs and some of their analogues found in counterfeit medicines,
J. Chromatogr. A 1218 (2011) 6439–6447.
[16] Y. Ren, C. Wu, J. Zhang, Simultaneous screening and determination of 18 illegal
adulterants in herbal medicines and health foods for male sexual potency by
ultra-fast liquid chromatography–electrospray ionization tandem mass spec-
trometry, J. Sep. Sci. 35 (2012) 2847–2857.
[17] M. Poplawska, A. Blazewicz, K. Bukowinska, Z. Fijalek, Application of high-
performance liquid chromatography with charged aerosol detection for
universal quantitation of undeclared phosphodiesterase-5 inhibitors in herbal
dietary supplements, J. Pharm. Biomed. Anal. 84 (2013) 232–243.
[18] Guideline Q2(R1), Validation of analytical procedures: text and methodology,
in: International Conference on Harmonisation, 2005 http://www.ich.org