1

Photoacoustic evaluation of the penetration of piroxicam gel through in

vivo skin applied with phonophoresis

Fabiola Leite Frana Dias da Silveira
1
, Fernanda Roberta Marciano
1
, Paulo Roxo Barja
1
,
Daniel Acosta-Avalos
2
1
Instituto de Pesquisa e Desenvolvimento (IP&D), Universidade do Vale do Paraba, Av.
Shishima Hifumi 2911, Urbanova, CEP 12244-000, So Jos dos Campos, SP, Brasil

2
Centro Brasileiro de Pesquisas Fsicas (CBPF), Ministrio da Cincia e Tecnologia, R.
Xavier Sigaud 150, Urca, CEP 22290-180, Rio de Janeiro, RJ, Brasil

Author for correspondence:
Daniel Acosta-Avalos
Centro Brasileiro de Pesquisas Fsicas (CBPF)
R. Xavier Sigaud 150, Urca, CEP 22290-180
Rio de Janeiro, RJ, Brasil
Tel: (55-21) 21417167
e-mail: dacosta@cbpf.br

2
Abstract

Phonophoresis is a technique that uses ultrasound (US) waves to facilitate drug
penetration through human skin. Previous works have used the photoacoustic technique
to study drug penetration in human skin in vivo, showing that phonophoresis indeed
enhances drug penetration through skin when compared to manual massage. The present
work investigates the effect of the phonophoresis application time in drug penetration.
The drug utilized in our study was Piroxicam, commonly applied in skin through
phonophoresis. The US equipment was operated at 3 MHz, in the continuous mode, with
intensity of 0.5 W/cm
2
; 3 mg of Piroxicam were applied into the skin using three different
application times: 1 min, 2 min and 3 min. A typical photoacoustic setup was used to
analyze the drug penetration. A background in the theory required for data analysis is
offered in the text. The results obtained do not show statistical difference in the quantity
of Piroxicam that penetrates the skin, for the different times of phonophoresis. This
indicates that, for the conditions employed in our study, 1 min of phonophoresis
application is sufficient to help drug penetration through skin.

3
Key words: photoacoustic, phonophoresis, piroxicam, human skin

4
Introduction

In the phonophoresis therapy, the therapeutical properties of the ultrasound (US) waves
are utilized to help drug delivery trough the skin to soft tissues. However, the mechanism
of action for this procedure remains incompletely defined, though cavitation and thermal
processes have been strongly implicated [1]. Some studies have been done to analyze the
real enhance of transdermal drug delivery. Cagnie et al. [2] compared the concentration
of Ketoprofen in synovial and fat tissue after application of continuous and pulsed
phonophoresis and compared with the topical application through sham phonophoresis.
After biopsies of these tissues they observed that in fact phonophoresis increase the
Ketoprofen concentration when compared with topical application. Recently, an in vivo
evaluation of phonophoresis drug delivery was done using photoacoustic (PA) technique
[3]. The PA technique is based on the PA effect, which consists in the generation of
sound waves after the absorption of chopped radiation by a material enclosed in a sealed
cell. In 1976, Rosencwaig and Gersho developed a model [4] that explains the PA effect
in solids through the diffusion of the heat generated after absorption of the incident
radiation by the analyzed material, to the external media surrounding this material,
showing that the PA signal depends on its thermal properties. Using this technique, Barja
et al. [3] compared the penetration through the skin of the topically applied drug
diclofenac resinate (Cataflan ), using two different techniques: manual massage and
phonophoresis with US in continuous mode. They observed that the concentration of
Cataflan is higher with phonophoresis than that with manual massage, showing that US
therapy enhances transdermal drug delivery through the use of a noninvasive technique.
In the present report, a study of drug penetration was done comparing three different
phonophoresis application times, to analyze if drug delivery enhances with the increase
of the phonophoresis application time.

5
Theory
The Rosencwaig-Gersho (RG) model [4] states that the PA signal is produced after the
transformation of electromagnetic energy in heat, through non-radiative pathways. The
heat diffusion through the absorber and environment can be analyzed using the diffusion
equation for each media involved. In this way, the RG model produces a complete
solution for the PA signal that is, however, a complex mathematical expression. For a
better comprehension of the phenomena, two thermal and two optical regimes are
analyzed. In the optical case, materials are classified as transparent or opaque. In the
thermal case, materials can be classified as thin or thick, depending of the relation
between the thermal diffusion length () and the sample thickness (l). The thermal
diffusion length is defined as:
f t
o
= (1)
where o is the thermal diffusivity and f is the chopped frequency of the incident
electromagnetic radiation. As it can be observed, the increase in f produces a decrease in
. If > l, the material is considered thermally thin; if < l, the material is considered
thermally thick. The continuous increase in f can produce the transition from thermally
thin to thermally thick. The complete mathematical expression for the PA signal can be
simplified in the different optical and thermal regimes. For the case of optically opaque
and thermally thin samples (sample is the material that absorbs the electromagnetic
radiation), the PA signal can be written as [3]:
PA signal ~
b
g b
k
Y
2

(2)
where Y is a constant,
b
is the thermal diffusion length for the base material (material
over the sample, in the opposite face to that illuminated by the electromagnetic radiation),
k
b
is the thermal conductivity of the base material and
g
is the thermal diffusion length
for the gas (generally, air) facing the sample on the side illuminated by the
electromagnetic radiation. Eq. (2) can be rewritten using the definition for the thermal
effusivity (c): o c / k = . The thermal effusivity is a parameter related with the capacity
of a material to interchange heat with the environment in opposition to the thermal
6
diffusivity that is related with the capacity of a material to permit the passage of heat. Eq.
(2) can then be rewritten as:
PA signal ~
f
Y
b
g
t c
o
(3)
It can be observed in eq. (3) that the PA signal in the analyzed case does not depend on
thermal parameters of the sample, but it depends on the thermal parameters of the gas and
the base. So, it would be influenced by the changes in thermal parameters of both
materials.
Figure 1 shows the scheme of the PA cell employed. It can be seen that the
modulated light impinges first on a glass window that closes one face of the PA chamber.
Inside this chamber there is air. On the other face the PA chamber is closed with an
aluminum foil. In contact with the Al foil is the in vivo human skin. In this scheme, the
sample corresponds to the aluminum foil, the gas corresponds to the air and the
base corresponds to the human skin. To apply eq. (3) the Al foil must be in the
optically opaque and thermally thin regimes. For that, the modulation frequency f used
was 17 Hz, so = 0.13 cm, and as the thickness of the foil was 65 m, the thermally thin
requirement is satisfied. It can be observed in the Eq. (3) that the PA signal then depends
on the thermal effusivity of the human skin.
Gutierrez-Juarez et al. [5] also studied the drug penetration through human skin
using PA techniques. They evaluated the thermal effusivity of a mixture of substances
(c
MIX
), and proposed the following expression for c
MIX
:
c
MIX
= c
1
(c
2
/ c
1
)
X
(4)
where X is the relative concentration of substance 2 on substance 1. As we are interested
in evaluating the penetration of drug into the human skin through the estimation of X, in
eq. (4) the subscripts are identified as follows: MIX for the complex skin plus drug, 1 for
the clean skin and 2 for the pure drug. In this way, X can be written as follows:
) log(
) log(
SKIN DRUG
SKIN MIX
X
c c
c c
=
(5)

7
where log stands for logarithm in any base. In this work we use log as the decimal
logarithm. As c
DRUG
and c
SKIN
are constants and c is inversely proportional to the PA
signal (eq. 3), so the concentration X would be proportional to the following expression:
|
|
.
|

\
|
~
MIX
SKIN
Signal PA
Signal PA
X log (6)
So, the measurement of the PA signal of the clean skin and of the skin plus drug after
phonophoresis is sufficient to estimate the value of the concentration of the drug inside
the skin. This value is approximate because of the multiplicative constant factor
) / log(
SKIN DRUG
c c that does not depend on the phonophoresis treatment.

8
Materials and Methods

Measurements were performed in four young volunteers (21 to 32 years old) that did not
present allergy to the drug employed. Furthermore, volunteers should not present any
injury or metallic implant in the region of the skin chosen for the experiment. Initially,
PA measurements were performed in the clean skin. After that, 3g of Piroxicam gel were
applied in the right arm using phonophoresis. Three different regions of the same arm
were chosen to apply the drug with different phonophoresis application times (1, 2 and 3
minutes). All the volunteers were analyzed the same day. The experimental procedure
was repeated after 24h, because this period corresponds to the half life time of the drug in
the body. Ten repetitions were done in all volunteers. The US equipment utilized in the
phonophoresis was operated in the continuous mode, at 3MHz and intensity of 0.5W/cm
2
.
For PA measurements, light from a 250 W tungsten lamp was modulated at 17 Hz with a
mechanical chopper (Stanford Research Systems, mod.SR540) and directed to a PA cell
closed by a glass window in one side and, in the other, by an Al foil (about 65m
thickness). Skin was pressed against the external face of the Al foil. This configuration
made it possible to perform measurements in vivo. The PA cell contained an electret
microphone employed to detect the acoustic waves generated in the PA chamber. The
microphone signal was detected with a lock-in amplifier (Stanford Research Systems
mod. SR530). A computer was interfaced to the lock-in through the RS232 port to control
the system and storage the generated data. The software Microcal Origin was utilized
to analyze the obtained data.
Every measurement was done as follows: after phonophoresis, the volunteer pressed the
skin against the Al foil on the PA cell and the PA signal was registered at three second
intervals, up to a total of 100 points, from which the average value was taken as the PA
signal value. For each phonophoresis application period (1, 2 and 3 minutes), this
procedure was repeated subsequently 10 times and a statistical analysis was done on the
total set of average values. In some cases, some of the obtained values were discarded
due to experimental problems during the measurement procedures.
The statistical ANOVA or Students t tests were done using GraphPad InStat version 3.0,
GraphPad Software, San Diego California USA, www.graphpad.com.
9
Results and Discussion

Each PA signal measured corresponds with the average of 100 simultaneous
measurements. Every averaged PA signal was considered as one measurement. This
procedure was repeated 10 times for every experimental situation in the 4 volunteers: 1
min phonophoresis, 2 min phonophoresis and 3 min phonophoresis. In all situations, the
PA signal of the clean skin was obtained before phonophoresis and the mixture of skin
plus Piroxicam after application of phonophoresis. In this way, for every situation a set of
averaged PA signals was created and analyzed using the appropriated statistical tests. Not
all the generated sets have 10 values, because some measurements showed low signal-to-
noise ratio and were discarded. Tables I to IV show the average values for the PA signal,
measured in volts (V), and its corresponding standard deviation, for every volunteer.
It can be observed in Tables I to IV that the PA signal for the clean skin is significantly
higher than the PA signal of skin plus Piroxicam (MIX) (paired Students t test, p < 0.05
for all the analyzed times in all volunteers). This is related to the fact that the penetration
and storage of the drug in the skin decrease the PA signal when compared to the signal of
the clean skin, which is expected for the calculation of X (eq. 6). Following, the averaged
values for the PA signals of clean skin and skin plus Piroxicam were used to estimate the
value of X, for each volunteer and for each phonophoresis application time. The results
are showed in Table V. For each volunteer, an ANOVA test was done to compare the X
values obtained for the three different application times. In all cases, the ANOVA test
showed that all the obtained X values are statistically similar in the different times
analyzed (p > 0.05). This suggests that higher phonophoresis application times do not
increase the quantity of drug that penetrates the skin; in this case, 1min of phonophoresis
application would be sufficient to enhance transdermal drug penetration. Considering the
skin as a barrier that cumulates the drug to a later (and slower) diffusion to deeper tissues,
then the result obtained indicates that the drug does not penetrate increasing its
concentration with the increase in phonophoresis time, but its concentration maintains
approximately the same value and perhaps the phonophoresis promotes lateral drug
diffusion when the vertical drug penetration saturates. However, another interpretation is
possible if we think that the skin behaves as a dynamical barrier when phonophoresis is
10
used, allowing continuous drug delivery to deeper tissues without drug accumulation in
the stratum corneum (which is the layer actually responsible for the photoacoustic signal
in our experiment). More experiments must be done to elucidate which interpretation is
right.

11
Conclusions

The results obtained in this work support that the PA technique can be helpful in studying
drug delivery through the in vivo human skin. In spite of other evidences, we believe that
phonophoresis applications times higher than 1 min do not promote the penetration of
drug at concentrations higher than a saturation concentration. The value of this saturation
level must be determined by skin parameters and by US parameters, because preliminary
experiments done in our laboratory showed that US in pulsed mode promotes higher
concentration of drug in the skin than US in continuous wave mode.

12
References
1. Merigno G, Kalia Y N, Guy R H. Ultrasound-enhanced transdermal transport . J Pharm
Sci 2003: 92: 1125 1137.

2. Cagnie B, Vinck E, Rimbaut S, Vanderstraeten G. Phonophoresis versus topical
application of Ketoprofen: comparison between tissue and plasma levels. Phys Ther
2003: 83: 701 712.

3. Barja P R, Acosta-Avalos D, Rompe P C B, Anjos F H, Marciano F R, Silva M D. In
vivo evaluation of drug delivery after ultrasound application: a new use for the
photoacoustic technique. J Phys IV France 2005: 125: 789 791.

4. Rosencwaig A, Gersho A. Theory of photoacoustic effect with solids. J Appl Phys
1976: 47: 64 69.

5. Gutierrez-Juarez G, Vargas-Luna M, Cordova T, Varela J B, Bernal-Alvarado J J, Sosa
M. In vivo measurement of human skin absorption of topically applied substances by a
photoacoustic technique. Phys Meas 2002: 23: 521 532.

13
Table I: Average value and standard deviation (SD) of the PA signal (in V) of the clean
skin and the mixture (MIX) of skin plus Piroxicam, in function of the phonophoresis time
application, for volunteer 1.

1 min
(N=10)
Time
2 min
(N=10)

3 min
(N=8)
Average SD Average SD Average SD
Clean Skin 0,0023 0,0005 0,0022 0,0005 0,0024 0,0006
MIX 0,0019 0,0003 0,0018 0,0004 0,0018 0,0004
N: Number of repetitions

14
Table II: Average value and standard deviation (SD) of the PA signal (in V) of the clean
skin and the mixture (MIX) of skin plus Piroxicam, in function of the phonophoresis time
application, for volunteer 2.

1 min
(N=10)
Time
2 min
(N=10)

3 min
(N=9)
Average SD Average SD Average SD
Clean Skin 0,0020 0,0006 0,0022 0,0006 0,0022 0,0004
MIX 0,0017 0,0004 0,0019 0,0004 0,0020 0,0003
N: Number of repetitions

15
Table III: Average value and standard deviation (SD) of the PA signal (in V) of the clean
skin and the mixture (MIX) of skin plus Piroxicam, in function of the phonophoresis time
application, for volunteer 3.

1 min
(N=10)
Time
2 min
(N=10)

3 min
(N=9)
Average SD Average SD Average SD
Clean Skin 0,0021 0,0006 0,0021 0,0006 0,0024 0,0006
MIX 0,0017 0,0004 0,0018 0,0004 0,0019 0,0003
N: Number of repetitions

16
Table IV: Average value and standard deviation (SD) of the PA signal (in V) of the clean
skin and the mixture (MIX) of skin plus Piroxicam, in function of the phonophoresis time
application, for volunteer 4.

1 min
(N=7)
Time
2 min
(N=6)

3 min
(N=5)
Average SD Average SD Average SD
Clean Skin 0,0025 0,0005 0,0024 0,0004 0,0022 0,0004
MIX 0,0020 0,0003 0,0022 0,0003 0,0019 0,0002
N: Number of repetitions

17
Table V: Average value (X
m
) and standard deviation (SD) calculated for the approximate
values of X, calculated using eq. 6, for each volunteer and for each phonophoresis
application time.

1 min
Time
2 min

3 min
Volunteer X
m
SD X
m
SD X
m
SD
1 0,082 0,045 0,092 0,052 0,125 0,044
2 0,076 0,049 0,064 0,036 0,041 0,029
3 0,067 0,053 0,062 0,023 0,083 0,046
4 0,086 0,048 0,035 0,026 0,077 0,058
18

Figure Legend

Figure 1: Drawing of the PA cell used to measure the PA signal of in vivo human skin.