Pharmaceutical Development and Technology, 2010; 15(2): 169–177

R E S E A R C H A R T I C L E

Iontophoretic delivery of lisinopril: Optimization of process variables by Box-Behnken statistical design

Ramesh Gannu, Vamshi Vishnu Yamsani, Chinna Reddy Palem, Shravan Kumar Yamsani and Madhusudan Rao Yamsani

National Facilities in Engineering and Technology with Industrial Collaboration (NAFETIC), University College of Pharmaceutical Sciences, Kakatiya University, Warangal, India

Address for Correspondence: Prof. Y. Madhusudan Rao, National Facilities in Engineering and Technology with Industrial Collaboration (NAFETIC) University College of Pharmaceutical Sciences, Kakatiya University, Warangal-506 009 (A.P), India. Tel: + 91 870 2438844. Fax: +91 870 2453508. E-mail: ymrao123@yahoo.com

(Received 04 February 2009; revised 22 April 2009; accepted 01 June 2009)

Introduction

Lisinopril (LSP) is an angiotensin converting enzyme inhibitor used for the treatment of hypertension, con-gestive heart failure and to alleviate strain on hearts damaged as a result of a heart attack. LSP is slowly and incompletely absorbed after oral administration with a bioavailability of 25–30%.[1,2] LSP is available only in the form of oral tablets in the market. However, this formu-lation has a major disadvantage since it is incompletely absorbed from the gastrointestinal tract. To overcome the problem of incomplete absorption, improve bio-availability and for effective treatment of chronic hyper-tension, an alternative long-acting formulations could be beneficial. Transdermal route of administration may

be a good alternative to circumvent these problems. For certain drugs, transdermal delivery offers a number of advantages with respect to oral or parental administra-tion: improved patient compliance, reduced side-effects, elimination of first-pass effect, interruption or termina-tion of treatment when unnecessary, etc.[3] A classical transdermal administration would not be adequate because of the low permeability of the skin and the pro-longed lag time resulting from the excellent barrier prop-erties of the horny layer. However, only a small minority of drug molecules are able to passively penetrate the skin.[4] Ionic, neutral, and/or polar molecules typically show limited skin penetration ability.[5,6] LSP is an ideal candidate for the development of a transdermal dosage form because of its clinical need, low molecular weight

ISSN 1083-7450 print/ISSN 1097-9867 online © 2010 Informa UK LtdDOI: 10.3109/10837450903085418

AbstractThe objective of the investigation was to optimize the iontophoresis process parameters of lisinopril (LSP) by 3 × 3 factorial design, Box-Behnken statistical design. LSP is an ideal candidate for iontophoretic deliv-ery to avoid the incomplete absorption problem associated after its oral administration. Independent variables selected were current (X1), salt (sodium chloride) concentration (X2) and medium/pH (X3). The dependent variables studied were amount of LSP permeated in 4 h (Y1: Q4), 24 h (Y2: Q24) and lag time (Y3). Mathematical equations and response surface plots were used to relate the dependent and independ-ent variables. The regression equation generated for the iontophoretic permeation was Y1 = 1.98 + 1.23X1 − 0.49X2 + 0.025X3 − 0.49X1X2 + 0.040X1X3 − 0.010X2X3 + 0.58X1

2 − 0.17X22 − 0.18X3

2; Y2 = 7.28 + 3.32X1 − 1.52X2 + 0.22X3 − 1.30X1X2 + 0.49X1X3 − 0.090X2X3 + 0.79X1

2 − 0.62X22 − 0.33X3

2 and Y3 = 0.60 + 0.0038X1 + 0.12X2 − 0.011X3 + 0.005X1X2 − 0.018X1X3 − 0.015X2X3 − 0.00075X1

2 + 0.017X22 − 0.11X3

2. The statistical validity of the polynomials was established and optimized process parameters were selected by feasibility and grid search. Validation of the optimization study with 8 confirmatory runs indicated high degree of prognostic ability of response surface methodology. The use of Box-Behnken design approach helped in identifying the critical process parameters in the iontophoretic delivery of lisinopril.

Keywords: Lisinopril; optimization; Box-Behnken design; iontophoresis; factorial design

http://www.informahealthcare.com/phd

Phar

mac

eutic

al D

evel

opm

ent a

nd T

echn

olog

y D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Uls

ter

at J

orda

nsto

wn

on 1

1/09

/14

For

pers

onal

use

onl

y.

170 R. Gannu et al.

(405.5 g/mol), and low dose (2.5– 20 mg/day). LSP, with apparent dissociation constants (pKa) of 2.5, 4.0, 6.7 and 9.8, exhibits a negative charge (pKa 2.5 because of pro-line group).[7]

Ions permeate the skin at a much lower rate rela-tive to neutral compounds. Techniques such as ion pairing,[8] chemical enhancers[9] and electric current[10] were successfully used to enhance ionic drug permea-tion. Iontophoresis effectively delivers a large variety of compounds across the skin[11] and it involves the use of low-density electric current (approximately 0.5 mA/cm2) to drive charged molecules across the skin. The current cycle is constructed by placing an electrode in a drug-containing compartment in contact with skin; a ground electrode is placed on the body to complete the circuit.[12] Drug transport across the skin is enhanced by three mechanisms: Charged species are driven pri-marily by electric repulsion from the driving electrode, the flow of electric current may increase the perme-ability of skin, and electroosmosis may affect uncharged molecules. Efficiency of transport depends mainly on polarity, valency, and mobility of the charged species as well as on the electrical duty cycles and formulation components.[13]

Some important considerations include flux propor-tionality with respect to applied current density and the presence of ions other than drug (these decrease the efficiency of iontophoretic transport of the drug). Currently, up to 0.5 mA/cm2 is believed to be tolerable for patients. The onset of action with iontophoretic treat-ment is rapid, in contrast to hours for passive transder-mal delivery.[14] Since drug delivery is proportional to applied current, significant advantages of iontophoresis include the possibility of preprogramming the drug delivery, dose tailoring on an individual basis, or time tailoring in a constant or pulsatile fashion.

Traditional experiments require more effort, time, and materials when a complex process needs to be developed. Various experimental designs[15–17] are use-ful in developing process parameters for iontophoresis, requiring less experimentation and providing esti-mates of the relative significance of different variables. Response surface methodology (RSM) is a widely prac-ticed approach in the development and optimization of drug delivery devices.[18] In this investigation we explored the utility of RSM to the optimization of iontophoretic process parameters. Based on the principle of design of experiments, the methodology encompasses the use of various types of experimental designs, generation of polynomial equations, and mapping of the response over the experimental domain to determine the opti-mum process parameters.[18] The technique requires minimum experimentation and time, thus proving to be far more effective and cost effective than the conven-tional methods of formulating dosage forms.

The aim of this study was to optimize the process parameters for iontophoretic delivery of LSP by Box-Beheken statistical design. Independent variables selected were current density (X

1), salt concentration

(NaCl) (X2), and medium/pH (X

3) to evaluate their sepa-

rate and combined effects on percentage drug perme-ated Y

1 (Q

4), Y

2 (Q

24) and lag time (Y

3).

Materials and methods

Materials

Lisinopril was generously provided by Dr Reddy’s Laboratories, India. Ag (99.9% purity), AgCl (99%) were purchased from Sigma (Sigma-Aldrich-Química, Madrid) and Sd fine Chemicals, India, respectively. Dulbecco’s buffer (pH 7.4) was purchased from Himedia, India. High-performance liquid chromatography (HPLC) sol-vents, (methanol and acetonitrile) were purchased from Merck., India. All other reagents used were of analytical grade and used without further purification.

Preparation of rat abdominal skin

Albino rats weighing 150–200 g were selected for per-meation studies and the study was conducted with the approval of institutional ethical committee, University College of Pharmaceutical Sciences, Kakatiya University, India. The animals were sacrificed using anesthetic ether, hair of test animals was carefully trimmed short (< 2 mm) with electrical clippers and the full thickness skin was removed from the abdominal region. The epidermis was prepared surgically by heat separation technique,[19] which involved soaking the entire abdominal skin in water at 60°C for 45 s, followed by careful removal of the epidermis. The epidermis was washed with water and used for iontophoretic studies.

Preparation of electrodes

Silver/Silver chloride (Ag/AgCl) electrodes were used for the application of the iontophoretic current. Silver wire of 0.5 mm diameter (about 5 cm) was rinsed with concentrated hydrochloric acid, followed by water to remove surface contamination. The wire electrode was then immersed in molten silver chloride for 30 min. A gray silver chloride layer was gradually coated on the silver wire.

Iontophoresis of LSP

Iontophoresis experiments were conducted using Franz diffusion cells. The skin was clamped between the two halves of diffusion cell (3.56 cm2) so as to face the stratum

Phar

mac

eutic

al D

evel

opm

ent a

nd T

echn

olog

y D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Uls

ter

at J

orda

nsto

wn

on 1

1/09

/14

For

pers

onal

use

onl

y.

Iontophoretic delivery of lisinopril 171

corneum towards the donor compartment, cathodal chamber of the cell in which 3 mL ( 30 mg) of drug solu-tion was placed. A power supply (Sharada Electronics Inc., India) was used to deliver a constant direct current for 4 h via Ag/AgCl electrodes. LSP, has a low molecular weight (Mol. wt. 405.5), which is sufficiently small for successful iontophoretic transdermal administration. LSP, with apparent dissociation constants (pKa) of 2.5, 4.0, 6.7 and 9.8, were found to have electrically enhanced drug movement under the influence of cathodal ionto-phoresis. The receptor solution (anodal chamber) was 12 mL of phosphate buffer (pH 7.4) and was magneti-cally stirred. The samples were collected from receptor compartment at predetermined intervals and replen-ished with fresh buffer. The drug content in the samples was determined by HPLC and the concentration was corrected for sampling effects according to the following equation:[20]

C C (V V V ) (C C )n1

n T T S n 11

n 1= / − /− − (1)

where C1n is the corrected concentration of the nth sam-

ple, Cn is the measured concentration of LSP in the nth

sample, C n-1

is the measured concentration of the LSP in the (n -1)th sample, V

T is the total volume of the receiver

fluid and VS is the volume of the sample drawn.

HPLC Analysis of LSP

Analysis of samples was performed with a Shimadzu HPLC system equipped with LC-10AT pump, UV-Vis spectrophotometric detector (SPD-10A) and C18 col-umn (Phenomenex; 250 × 4.6 mm; 5 m) at ambient temperature. The mobile phase used was a mixture of phosphate buffer (25 mM potassium dihydrogen ortho phosphate, pH 5.0) and acetonitrile (88:12). A flow rate of 1 mL min−1 was maintained and the detection wave-length was 215 nm. A calibration curve was plotted for LSP in the range of 50–2500 ng mL−1. A good linear relationship was observed between the concentration of LSP and the peak area of LSP with a correlation coef-ficient (r2 = 0.999). The required studies were carried out to estimate the precision and accuracy of the HPLC method. Sample preparation briefly involved the filtra-tion of iontophoretic sample through 0.45 µ membrane filter, diluted with mobile phase and 20 L was spiked into column.

Experimental design

Box-Behnken statistical design was used to statistically optimize the iontophoresis process parameters and evaluate main effects, interaction effects and quad-ratic effects of the process parameters on the amount permeated in 4 h (Q

4) and amount permeated in 24 h

(Q24

). Response surface methodologies such as the

Box-Behnken model possible and Central Composite Design (CCD), curvature in the response function.[21,22] A 3-factor, 3-level Box-Behnken design was used to explore quadratic response surfaces and constructing second order polynomial models with Design Expert (Version 7.1, Stat-Ease Inc., Minneapolis, MN, USA). The Box-Behnken design was specifically selected since it requires fewer runs than a CCD in cases of three or four variables. This cubic design is characterized by set of points lying at the mid point of each edge and a replicate centre point of the multidimensional cube.[23] A design matrix comprising of 13 experimental runs was con-structed. The non-linear computer generated quadratic model is given as

Y b b X b X b X b X X b X X

b X X b X b X0 1 1 2 2 3 3 12 1 2 13 1 3

23 2 3 11 12

22 2

= + + + + +

+ + + 2233 3

2b X+ (2)

where Y is the measured response associated with each factor level combination; b

0 is an intercept; b

1 to b

33 are

regression coefficients computed from the observed experimental values of Y; and X

1, X

2 and X

3 are the coded

levels of independent variables. The terms X1X

2 and X

i2

(i = 1, 2 or 3) represent the interaction and quadratic terms, respectively.[22,23] The dependent and independ-ent variables selected are shown in Table 1 along with their low, medium and high levels, which were selected based on the results from preliminary experimenta-tion. The current density (X

1), ionic strength (X

2) and

medium/pH (X3) used to prepare the 13 experimental

trials and the respective observed responses are given in Table 2.

Optimization data analysis and optimization model validation

A total of 13 runs were generated using Design Expert software and the polynomial equations generated were statistically validated using ANOVA provision in the soft-ware. The models were evaluated in terms of statistically

Table 1. Variables in Box-Behnken design.

Factor

Levels used, Actual (coded)

Low (−1) Medium (0) High (+1)

Independent variables

X1 = Current density

(mA)0.05 0.28 0.5

X2 = Ionic strength

(Nacl, mM)0 50 100

X3 = Medium/pH Buffer (pH 5.8) Water Buffer (pH 7.4)

Dependent variables Constraints

Y1 = Q

4 (Cumulative amount of LSP permeated in

4 h, mg)1 ≤ Y

1 ≤ 5

Y2 = Q

24 (Cumulative amount of LSP permeated in

24 h, mg)3 ≤ Y

2 ≤ 15

Y3 = Lag time (h) 0.4 ≤ Y

3 ≤ 0.9

Phar

mac

eutic

al D

evel

opm

ent a

nd T

echn

olog

y D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Uls

ter

at J

orda

nsto

wn

on 1

1/09

/14

For

pers

onal

use

onl

y.

172 R. Gannu et al.

significant coefficients and r2 values. Various feasibility and grid searches were conducted to find the optimum process parameters. Eight optimum checkpoint for-mulations were selected over the whole experimental region by intensive grid search to validate the chosen experimental model and polynomial equations. The optimized checkpoint iontophoretic parameters were evaluated for various response properties. The resultant experimental values of the responses were quantita-tively compared with the predicted values to calculate the percentage prediction error.

Results and discussion

Iontophoresis of LSP

Iontophoresis has been proposed as a technique for enhancing skin penetration of several drugs.[24] The present study was aimed at identifying the key factors controlling electro transport of LSP. A set of preliminary trials were undertaken to establish the range of each process variable with the aim to optimize the ionto-phoretic delivery across rat abdominal skin. The prelim-inary trials conducted have revealed that iontophoretic process parameters with high current density (0.5 mA) resulted in more drug permeation (Q

4) whereas with low

current density (0.05 mA) resulted in low drug permea-tion (Q

4). Based on these observations the lower and

higher levels of current densities were retained at 0.05 and 0.5 mA, respectively and the center region (0.28 mA) was also used during the run.

Thirteen experiments were required for the response surface methodology based on the Box-Behnken design. The design has the advantage of requiring fewer experi-ments (13 experiments) than would a full factorial design

(27 batches). The experimental runs and the observed responses for the 13 trials are given in Table 2. Based on the experimental design, the factor combinations resulted in different responses. The range of response were found to be 1.10 mg in P03 to 4.68 mg in P07; 3.45 mg in P03 to 14.07 mg in P07 and 0.41 h in P07 to 0.82 h in P03 as Y

1, Y

2 and Y

3 respectively. Iontophoretic

delivery profiles of all the 13 formulations were shown in Figures 4a–b.

The theoretical (predicted) values and the observed values were in reasonably good agreement as seen from Table 3. The significance of the ratio of mean square variation due to regression and residual error was tested using analysis of variance (ANOVA). The ANOVA indi-cated a significant (P < 0.05) effect of factors on response. The relationship between the dependent and independ-ent variables was further elucidated using contour and response surface plots. The effects of X

1 and X

2 and

their interaction on Y1 at a fixed level of X

3 are given in

Figures 1a and 1d. At low levels of X2 (sodium chloride), Y

1

was found to be increased from 1.47 mg (P01) to 4.68 mg (P07) when the current density (X

1) was increased from

0.05–0.5 mA. Similarly, at high levels of X2, Y

1 increases

Table 3. Composition of checkpoint formulations the predicted and experimental values of response variables and percentage prediction error.

Check point formula composition (X

1:X

2:X

3)

Response variable

Experimental value

Predicted value

% prediction error

0.50:8:Wa Y1 (µg) 4.68 5.01 −7.05

Y2 (µg) 14.07 13.95 +0.85

Y3 (h) 0.41 0.38 +7.32

0.49:4:5.8b Y1 (µg) 3.95 3.78 +4.30

Y2 (µg) 10.58 11.10 −4.91

Y3 (h) 0.60 0.58 +3.33

0.40:12:Wa Y1 (µg) 3.29 3.45 −4.86

Y2 (µg) 10.33 11.27 −9.10

Y3 (h) 0.48 0.52 −8.33

0.38:12:7.4c Y1 (µg) 2.81 3.10 −10.32

Y2 (µg) 7.35 8.06 −9.66

Y3 (h) 0.45 0.42 +6.67

0.40:25:5.8b Y1 (µg) 2.78 3.06 −10.07

Y2 (µg) 7.49 8.20 −9.48

Y3 (h) 0.61 0.58 +4.92

0.45:4:Wa Y1 (µg) 4.03 3.92 +2.73

Y2 (µg) 9.11 10.36 −13.72

Y3 (h) 0.59 0.63 −6.78

0.46:9:7.4c Y1 (µg) 4.19 4.28 −2.15

Y2 (µg) 12.11 12.64 −4.38

Y3 (h) 0.51 0.48 +5.88

0.41:16:7.4c Y1 (µg) 3.90 4.11 −5.38

Y2 (µg) 11.10 11.68 −5.23

Y3 (h) 0.51 0.59 −15.69

Note: awater; bphosphate buffer pH 5.8; cphosphate buffer pH 7.4.

Table 2. Observed responses in the iontophoresis of LSP by Box-Behnken design.

Batch

Independent variables Dependent variables (mean ± SD)

X1

X2

X3

Y1 (mg)a Y

2 (mg)a Y

3 (h)a

P01 −1 −1 0 1.47 ± 0.100 4.90 ± 0.261 0.45 ± 0.012

P02 −1 0 1 1.40 ± 0.159 4.39 ± 0.421 0.50 ± 0.022

P03 −1 1 0 1.10 ± 0.057 3.45 ± 0.655 0.82 ± 0.018

P04 0 −1 1 2.31 ± 0.142 7.35 ± 0.319 0.45 ± 0.016

P05 0 0 0 1.95 ± 0.068 6.58 ± 0.213 0.60 ± 0.018

P06 0 1 −1 1.70 ± 0.042 5.49 ± 0.131 0.61 ± 0.021

P07 1 −1 0 4.68 ± 0.299 14.07 ± 0.775 0.41 ± 0.011

P08 1 0 −1 4.12 ± 0.247 12.88 ± 0.621 0.60 ± 0.023

P09 1 1 0 2.34 ± 0.221 7.41 ± 0.695 0.80 ± 0.027

P10 1 0 1 4.19 ± 0.145 12.11 ± 0.460 0.51 ± 0.018

P11 0 1 1 1.68 ± 0.027 5.16 ± 0.173 0.52 ± 0.021

P12 −1 0 −1 1.38 ± 0.195 4.37 ± 0.760 0.45 ± 0.011

P13 0 −1 −1 2.29 ± 0.156 7.33 ± 0.608 0.48 ± 0.013aValues represented are mean ± SD (n = 3).

Phar

mac

eutic

al D

evel

opm

ent a

nd T

echn

olog

y D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Uls

ter

at J

orda

nsto

wn

on 1

1/09

/14

For

pers

onal

use

onl

y.

Iontophoretic delivery of lisinopril 173

from 1.10 mg (P03) to 2.34 mg (P09) when X1 increases

from 0.05–0.5 mA. As shown in Figures 2a–d, with a low amount of sodium chloride, the response Y

2 was found to

be increased from 4.90 mg (P01) to 14.07 mg (P07) when X

1 increased from 0.05–0.5 mA. The cumulative amount

of LSP permeated in 4 h (Q4) and 24 h (Q

24) were found to

1.00

1.00

0.50

0.50

0.00

0.00

−0.50

−0.50−1.00

−1.00Current (mA)

Current (mA)

Nac

l (m

M)

Nacl (mM)

Nacl (mM)

Nacl (mM)

Q4 (mg)

Q4

(mg)

5

4.73.7752.85

1.9251

1.00 1.000.50 0.50

0.00 0.00−0.50 −0.50

−1.00 −1.00

1.00

1.00

0.50

0.50

0.00

0.00

−0.50

−0.50−1.00

−1.00Current (mA)

Current (mA)

Med

ium

/pH

Medium/pH

Medium/pH

Med

ium

/pH

Q4 (mg)

Q4

(mg)

4.23.475

2.752.025

1.3

1.00 1.000.50 0.50

0.00 0.00−0.50 −0.50

−1.00 −1.00

1.00

1.00

0.50

0.50

0.00

0.00

−0.50

−0.50−1.00

−1.00

Q4 (mg)

Q4

(mg)

2.62.2751.95

1.6251.3

1.00 1.000.50 0.50

0.00 0.00−0.50 −0.50

−1.00 −1.00

1.675272.26321

2.85116

3.43911

4.02705

1.78204 2.234392.68673

3.13907

3.59141

5

2.32141

2.32141

2.12157 1.92173 1.522041.72189

5

(a) (d)

(e)

(f)

(b)

(c)

Figure 1. Contour plot showing effect of (a) current density (X1) and Nacl (X

2); (b) current density (X

1) and medium/pH (X

3); (c) Nacl (X

2) and

medium/pH (X3) on response Y

1 (Q

4); Corresponding response surface plots (d–f ).

Phar

mac

eutic

al D

evel

opm

ent a

nd T

echn

olog

y D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Uls

ter

at J

orda

nsto

wn

on 1

1/09

/14

For

pers

onal

use

onl

y.

174 R. Gannu et al.

1.00

1.00

0.50

0.50

0.00

0.00

−0.50

−0.50−1.00

−1.00Current (mA)

Current (mA)

Nac

l (m

M)

Nacl (mM)

Nacl (mM)

Nacl (mM)

Q24 (mg)

Q24

(mg)

5

1512963

1.00 1.000.50 0.50

0.00 0.00−0.50 −0.50

−1.00 −1.00

1.00

1.00

0.50

0.50

0.00

0.00

−0.50

−0.50−1.00

−1.00Current (mA)

Current (mA)

Med

ium

/pH

Medium/pH

Medium/pH

Med

ium

/pH

Q24 (mg)

Q24

(mg)

12.210.175

8.156.125

4.1

1.00 1.000.50 0.50

0.00 0.00−0.50 −0.50

−1.00 −1.00

1.00

1.00

0.50

0.50

0.00

0.00

−0.50

−0.50

−1.00

−1.00

Q24 (mg)

Q24

(mg)

8.37.3756.45

5.5254.6

1.00 1.000.50 0.50

0.00 0.00−0.50 −0.50

−1.00 −1.00

5.42256.69375 7.965 9.23625

10.50755

7.65919 7.0656 5.878436.47201

5.28484

5

(a) (d)

(e)

(f)

(b)

(c)

5.53167 7.145838.76

10.3742

11.9883

Figure 2. Contour plot showing effect of (a) current density (X1) and Nacl (X

2); (b) current density (X

1) and medium/pH (X

3); (c) Nacl (X

2) and

medium/pH (X3) on response Y

2 (Q

24); Corresponding response surface plots (d−f ).

Phar

mac

eutic

al D

evel

opm

ent a

nd T

echn

olog

y D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Uls

ter

at J

orda

nsto

wn

on 1

1/09

/14

For

pers

onal

use

onl

y.

Iontophoretic delivery of lisinopril 175

be increased from 1.10 and 4.68 mg to 3.45 and 14.07 mg in P03 and P07, respectively, using water as medium (pH 7.0). The results suggest that increase in the ionic strength decreased the iontophoretic permeation of LSP. Chloride ions might be competing with the transport of LSP anions; therefore, the permeation of LSP ions was decreased with the increasing ionic strength of sodium chloride.

Data analysis

The experiments, P07, P08 and P10 had the highest Q4

and Q24

. Table 3 shows the observed and predicted val-ues with residuals and percent error of responses for all the batches. The Q

4 and Q

24 (dependent variables)

obtained at various levels of the three independent variables (X

1, X

2 and X

3) was subjected to multiple

regression to yield a second-order polynomial equa-tion (full model):

Y (Q ) 1.98 1.23X 0.49X 0.025X 0.49X X

0.040X X 0.01 4 1 2 3 1 2

1 3

= + − + −

+ − 110X X 0.58X 0.17X

0.18X2 3 1

222

32

+ −

−

(3)

Y (Q ) 7.28 3.32X 1.52X 0.22X 1.30X X

0.49X X 0.092 24 1 2 3 1 2

1 3

= + − + −

+ − 00X X 0.79X 0.62X

0.33X2 3 1

222

32

+ −

−

(4)

The value of the correlation coefficient (r2) of Equation 3 was found to be 0.9763, indicating good fit (Table 4). The Q

4 values measured for the different experiments

showed wide variation (i.e. values ranged from a mini-mum of 1.10 in P03 to a maximum of 4.68 mg in P07). The results clearly indicate that the Q

4 value is strongly

affected by the variables selected for the study. This is also reflected by the wide range of values for coef-ficients of the terms of Equation 2. The main effects of X

1, X

2, and X

3 represent the average result of changing

1 variable at a time from its low level to its high level. The interaction terms (X

1X

2, X

1X

3, X

2X

3, X

12, X

22, and

X3

2) show how the Q4 changes when two variables are

simultaneously changed. The negative coefficients for all three independent variables indicate an unfavora-ble effect on the Q

4, while the positive coefficients for

the interactions between two variables (X1X

3) indicate

a favorable effect on Q4. Among the three independent

variables, the lowest coefficient value is for X2 (−0.49),

indicating that this variable is insignificant in predic-tion of Q

4. The Q

24 values (equation 4) of P07, P08 and

P10 were found to be more among the experimental tri-als. More amount of LSP is deposited in the skin layers during application of current, have permeated across skin and this was clearly evidenced from the experi-ments, where in P03 the Q

24 value was low among the

experiments.The lag time was found to be in the range of 0.41–

0.82 h. A polynomial Equation 5 was also developed for lag time (response Y

3):

Y (lag time) 0.60 0.0038X 0.12X 0.011X

0.005X X 0.013 1 2 3

1 2

= + + − + − 88X X 0.015X X

0.00075X 0.017X 0.11X1 3 2 3

12

22

32

−

− + −

(5)

Among the independent variables selected and their interactions, only X

1 and X

2 (Equation 5) were found to

be significant (P < 0.05), indicating a major contribut-ing effect of X

1 and X

2 on lag time. A positive value of

the coefficient for X1 (current) and X

2 (sodium chloride)

indicates a favorable effect on lag time. The experiments conducted at high current level and low salt concentra-tion yielded low lag time.

Contour plots and response surface analysis

Two-dimensional contour plots and three-dimensional response surface plots are presented in Figures 1a–f, 2a–f and 3, which are very useful to study the interaction effects of the factors on the responses. These types of plots are useful in the study of the effects of two factors on the response at one time. In all the presented Figures, the third factor was kept at a constant level. All the relationships among the three variables are non-linear, although Figures 1a–b; 1d–e exhibits a linear relationship of factor X

1 with factors X

2 and X

3, in the form of almost

straight lines up to the high level of current density. At medium to high level of current density the curves were found to be linear. Factors X

2 have curvilinear rela-

tionship at all the levels of two variables (Figure 2a–b) whereas factor X

3 has non-linear relationship (Figure

2c). The corresponding response surface plots (Figures

Table 4. Summary of results of regression analysis for responses Y1, Y

2 and Y

3 for fitting to quadratic model.

Quadratic model R2 Adjusted R2 Predicted R2 SD % CV

Response (Y1) 0.9763 0.9459 0.6269 0.24 10.66

Response (Y2) 0.9771 0.9477 0.6387 0.63 8.76

Response (Y3) 0.7628 0.5035 0.3856 0.08 14.45

Regression equations of the fitted quadratic model:Y

1 = 1.98 + 1.23X

1 − 0.49X

2 + 0.025X

3 − 0.49X

1X

2 + 0.040X

1X

3 − 0.010X

2X

3 + 0.58X

12 − 0.17X

22 − 0.18X

32;

Y2 = 7.28 + 3.32X

1 − 1.52X

2 + 0.22X

3 − 1.30X

1X

2 + 0.49X

1X

3 − 0.090X

2X

3 + 0.79X

12 − 0.62X

22 − 0.33X

32;

Y3 = 0.60 + 0.0038X

1 + 0.12X

2 − 0.011X

3 + 0.005X

1X

2 − 0.018X

1X

3 − 0.015X

2X

3 − 0.00075X

12 + 0.017X

22 − 0.11X

32.

Phar

mac

eutic

al D

evel

opm

ent a

nd T

echn

olog

y D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Uls

ter

at J

orda

nsto

wn

on 1

1/09

/14

For

pers

onal

use

onl

y.

176 R. Gannu et al.

1d–f and 2d–f) showed the relationship between these factors even more clearly. The results reveal that the Q

4

and Q24

were found to be increased with increasing cur-rent density at low level of sodium chloride. At higher level of sodium chloride iontophoretic permeation was less than that of lower level of sodium chloride. It indi-cates that the negatively charged chloride ions might be competing for the negatively charged LSP for permea-tion. Therefore as the concentration of sodium chloride increases the iontophoretic permeation was found to be decreased. The relationship between sodium chloride and current density (Figure 3) showed that the salt and current density from low level to high level, the lag time was found to be decreased however the difference was statistically insignificant (P > 0.05).

Checkpoint analysis

Eight checkpoint experiments were conducted and evaluated as shown in Table 3. The results indicate that the measured Q

4, Q

24 and lag time values were as

expected. When measured Q4, Q

24 and lag time values

were compared with predicted Q4, Q

24 and lag time val-

ues using Student’s t- test, the differences were found to be insignificant (P < 0.05). Thus, we can conclude that the obtained mathematical equation is valid for predict-ing the responses.

Optimization

The optimum process parameters were selected based on the criteria of maximum Q

4 and Q

24 values, by apply-

ing constraints on Y1 (1 ≤ Y

1 ≤ 5) and Y

2 (3 ≤ Y

2 ≤ 15). Upon

trading of various response variables and extensive grid search, the process parameters with current density of 0.50 mA, sodium chloride of 8 mM and water as medium were found to fulfill the maximum requisite of an

optimum iontophoretic process because of maximum Q

4 and Q

24.

Conclusions

The penetration was influenced by ionic strength (i.e. sodium chloride concentration) of the donor solution and current density. The results of the experimental study confirm that the factors X

1 and X

2 significantly

influence the dependent variable Q4 and Q

24. The

permeation of LSP across rat abdominal skin might be due to electrostatic repulsion. The application of Box-Behnken experimental design technique for opti-mization of process parameters helps in reaching the

Current (mA)Nacl (mM)

Lag

time

(h)

0.820.71750.615

0.51250.41

1.00 1.000.50 0.50

0.00 0.00−0.50 −0.50

−1.00 −1.00

Figure 3. Response surface plot showing effect of current density (X

1) and medium/pH (X

3) on response Y

3 (Lag time, h). 0.00

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

9.00

24.0

Time (h)

Cum

ulat

ive

amou

nt o

f LS

P p

erm

eate

d (m

g)

P1 P2

P3 P4

P5 P6

(a)

0.00

2.00

4.00

6.00

8.00

10.00

12.00

14.00

16.00

Time (h)

Cum

ulat

ive

amou

nt o

f LS

P p

erm

eate

d (m

g) P7 P8P9 P10P11 P12P13

(b)

0.0 4.0 8.0 12.0 16.0 20.0

24.00.0 4.0 8.0 12.0 16.0 20.0

Figure 4. (a−b) Iontophoretic delivery of LSP across rat abdominal skin.

Phar

mac

eutic

al D

evel

opm

ent a

nd T

echn

olog

y D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Uls

ter

at J

orda

nsto

wn

on 1

1/09

/14

For

pers

onal

use

onl

y.

Iontophoretic delivery of lisinopril 177

optimum point in the shortest time with minimum experiments.

Acknowledgements

The financial support in the form of National Doctoral Fellowship (NDF), sanctioned by AICTE, New Delhi, India, is highly acknowledged. The authors are thankful to Dr Reddy’s Laboratories, Hyderabad, India, for pro-viding gift samples of lisinopril.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.

References

1. Lancaster SG, Todd PA. Lisinopril: A preliminary review of its pharmacodynamic and pharmacokinetic properties and thera-peutic use in hypertension and congestive heart failure. Drugs 1988;35:646–669

2. Marc H. Angiotensin converting enzyme inhibitors, angiotensin and calcium channel blockers. In: Williams AD, Thamas LL, edi-tors. Foye’s principles of medicinal chemistry. 5th ed. Baltimore, MA: Lippincott Williams and Wilkins; 2002:533–561.

3. Vaddi H, Wang P, Chan S. Effect of some enhancer on the per-meation of haloperidol through rat skin in vitro. Int J Pharm 2001;212:247–255.

4. Thacharodi D, Rao KR. Transdermal absorption of nifedipine from microemulsions of lipophilic skin penetration enhancers. Int J Pharm 1994;111:235–240.

5. Rao G, Guy RH, Glikfel P, La Course WR. Reverse iontophoresis: Noninvasive glucose monitoring in vivo in human. Pharm Res 1995;12(12):1869–1873.

6. Alvarez-Figueroa MJ, Delgado-Charro MB, Blanco-Méndez J. Passive and iontophoretic transdermal penetration of meth-otrexate. Int J Pharm 2001;212:101–107.

7. Clarke’s analysis of drugs and poisons. London: The Pharmaceutical Press. www.medicinecomplete.com

8. Valenta C, Siman U, Kratzel M, Hadgraft J. The dermal deliv-ery of lignocaine: Influence of ion pairing. Int J Pharm 2000;197:77–85.

9. Smith JC, Irwin WJ. Ionisation and the effect of absorption enhancers on transport of salicylic acid through silastic rubber and human skin. Int J Pharm 2000;210:69–82

10. Anliker M, Maibach HI, Singh P. Transdermal iontophoresis and solute penetration across excised human skin. Curr Prob Dermatol 1995;22:184–188.

11. Banga AK, Chien YW. Iontophoretic delivery of drugs: Fundamentals, developments and biomedical applications. J Control Rel 1988;7:1–24.

12. Guy RH. Iontophoresis, recent developments. J Pharm Pharmacol 1998;50:371–374.

13. Guy RH, Kalia YN, Delgado-Charro MB, Merino V, Lopez A, Marro D. Iontophoresis: Electrorepulsion and electroosmosis. J Control Rel 2000;64:129–132.

14. Singh P, Maibach HI. Iontophoresis: An alternative to the use of carriers in cutaneous drug delivery. Adv Drug Deliv Rev 1996;18:379–394.

15. Gohel MC, Amin AF. Formulation optimization of controlled release diclofenac sodium microspheres using factorial design. J Control Rel 1998;51:115–122.

16. Nazzal S, Khan MA. Response surface methodology for the opti-mization of ubiquinone self-nanoemulsified drug delivery sys-tem. AAPS PharmSciTech 2002;3(1):1–9.

17. Li S, Lin S, Chien YW, Daggy BP, Mirchandani HL. Statistical optimization of gastric floating system for oral controlled deliv-ery of calcium. AAPS PharmSci Tech 2001;2:Art. 1.

18. Lewis GA, Mathieu D, Phan-Tan-Luu R. Pharmaceutical experi-mental design. New York: Marcel Dekker; 1999.

19. Zhao K, Singh J. Mechanism of percutaneous absorption of tamoxifen by terpenes: Euginol, D-limonen and menthone. J Control Rel 1998;55:359–366.

20. Hayton WL, Chen T. Correction of perfusate concentration for sample removal. J Pharm Sci 1982;71:820–821.

21. Govender S, Pillay V, Chetty DJ, Essack SY, Dangor CM, Govender T. Optimisation and characterisation of bioadhe-sive controlled release tetracycline microspheres. Int J Pharm 2005;306:24–40.

22. Chopra S, Pati GV. Motwani SK. Release modulating hydrophilic matrix systems of losartan potassium: Optimisation of formula-tion using statistical experimental design. Eur J Pharm Biopharm 2007;66(1):73–82.

23. Box GEP, Behnken DW. Some new three level designs for the study of quantitative variables. Technometrics 1960;2:455–475.

24. Green PG. Iontophoretic delivery of peptide drugs. J Control Rel 1996;41:33–48.

Phar

mac

eutic

al D

evel

opm

ent a

nd T

echn

olog

y D

ownl

oade

d fr

om in

form

ahea

lthca

re.c

om b

y U

nive

rsity

of

Uls

ter

at J

orda

nsto

wn

on 1

1/09

/14

For

pers

onal

use

onl

y.