Controlled release of benzoyl peroxide from a porous microsphere polymeric system can reduce topical irritancy

Controlled release of benzoyl peroxide from a porous microsphere polymeric system can reduce topical irritancy Ronald C. Wester, PhD,a Rajesh Patel, PhD,b Sergio Nacht, PhD,b James Leyden, MD,c Joseph Melendres, BS,a and Howard Maibach, MDa San Francisco and Redwood City,

California, and Philadelphia, Pennsylvania Skin absorption of benzoyl peroxide from a topical lotion containing freely dispersed drug was compared with that from the same lotion in which the drug was entrapped in a controlledrelease styrene-divinylbenzene polymer system. In an in vitro diffusion system, statistically significant (p == 0.01) differences were found in the content of benzoyl peroxide in excised human skin and in percutaneous absorption. In vivo, significantly (p == 0.002) less benzoyl peroxide was absorbed through rhesus monkey skin from the polymeric system. This controlled release of benzoyl peroxide to skin can alter the dose relation that exists between efficacy and skin irritation. Corresponding studies showed reduced. skin irritation in cumulative irritancy studies in rabbits and human beings, whereas in vivo human antimicrobial efficacy studies showed that application of the formulations containing entrapped benzoyl peroxide significantly reduced counts of Propionibacterium acnes (p < 0.001) and aerobic bacteria (p < 0.001) and the free fatty acid/triglyceride ratio in skin lipids. These findings support the hypothesis that, at least for this drug, controlled topical delivery can enhance safety without sacrificing efficacy. (J AM ACAD DERMATOL 1991;24:720-6.)

Benzoyl peroxide (BPO) is clinically effective in acne primarily because of its bactericidal activity against Propionibacterium acnes and possibly through its mild keratolytic effect. I -3 The main site of pharmacologic action is the pilosebaceous cana1. 4 BPO penetrates through the follicular opening, probably by dissolving into sebaceous lipids, and there exerts its antimicrobial activity.5 Skin irritation is a common side effect, and a dose relation seems to exist between efficacy and irritation. 6 Theoretically, controlled release ofBPO from a delivery system to the skin could alter this dose relation by maintaining intrafollicular penetration while reducing percutaneous absorption. As an alternative approach, we entrapped BPO in a macroporous polystyrene delivery system (Styrenejdivinylbenzene Microsponge System, Advanced Polymer Systems, From the Department of Dermatology, University of California, San Francisco,' Advanced Polymer Systems, Redwood City,b and University of Pennsylvania, Philadelphia." Accepted for publication Dec. 17,1990. Reprint requests: R. C. Wester, University of California, San Francisco, CA 94143-0989.

16/1/27433

720

Inc., Redwood City, Calif.), whichwas then incorporated into standard vehicles for topical application. Our objective was to determine the skin absorption and irritation profiles of BPO when entrapped in the polystyrene system and dispersed in oilin-water vehicles compared with the same amount of BPO, but freely dispersed in similar vehicles, when applied to the skin. Release rates of BPO from the experimental formulations were measured in vitro using silicone elastomer (Silastic, Dow Chemical Corp., Midland, Mich.) membranes. Percutaneous absorption was determined in vitro with excised human skin and in vivo in the rhesus monkey. 7 Cumulative skin irritancy tests were conducted on white New Zealand strain rabbits and in human beings. The in vivo antimicrobial efficacy was determined in human volunteers with a quantitative culture technique. 8 METHODS

Test articles Tritium BPO (97% chemical and radiochemical purity), was formulated as a 5% lotion with the tritium BPO freely dispersed (formulation A) or entrapped in a

Volume 24 Number 5, Part 1

Controlled release of benzoyl peroxide 721

May 1991

Microsponge System (formulation B) and with a mean particle size of 25 p,m (range 5 to 50 ,urn). Five formulations, formulation C (simple oil-in-water emulsion vehicle [SOWE]), formulation D (2.5% free BPO in SOWE), formulation E (2.5% entrapped BPO in SOWE), formulation F (5% free BPO in SOWE), and formulation G (5.0% entrapped BPO in SOWE), were prepared for the rabbit skin irritancy and the antimicrobial efficacy studies. For cosmetic reasons the vehicle composition was somewhat modified, and proprietary BPO formulations containing 2.5% (formulation H) and 10.0% (formulation J) entrapped active ingredient were prepared and tested both in vitro for release rates and in vivo for human irritancy and antimicrobial efficacy. In these irritancy tests, commercial products containing BPO (formulation K, with 2.5% BPO, and formulation L, with 10.0% BPO) were used for comparison.

BPO release rates from formulations Silastic membranes (#500-1) were mounted in static diffusion cells9 and the membrane surfaces dosed with 5% BPO lotion containing BPO either freely dispersed (formulation F) or entrapped in the polymer system (formulation G) at a rate of 5p,l (5 mg) offormulation/cm2• The formulations containing 2.5% and 10% BPO (formulations Hand J) were also tested and compared with commercial products containing similar drug concentrations (formulations K and L). Receptor fluid composed of distilled water/acetone (1:1) was added to the cell and maintained at 25 C. Because of the exceedingly low solubility of BPO in either water or normal saline solution, a water/acetone mixture was selected as receptor fluid to provide adequate "sink" conditions after preliminary experiments showed no interactions of this mixture with either the membrane or the mixtures placed on the "donor" side. Drug flux through the membrane was determined by periodically withdrawing the receptor phase and analyzing for permeant content by high-performance liquid chromatography (HPLC). Analyses were performed on a Perkin-Elmer HPLC system (Perkin-Elmer Corp., Norwalk, Conn.) fitted with a binary pump (model 250), autosampler (model ISS100), LC-95 spectrophotometer, and Nelson 2600 data station. A 5 p,m ultrasphere reversed-phase column in octadecylsilane form (4.6 rom inner diameter X 25 cm length) was used. The mobile phase was methanol/water (75:25), and the flow rate was set at 1.2 mlfmin. The spectrophotometer was set at 254 nm and injection volume was 10 J.'l. The retention times were 8.0 minutes for BPO and 5.2 minutes for benzophenone (internal standard). The results were reported as an average of two determinations. 0

In vitro diffusion system Skin was excised from human thigh with a dermatome set at 0.5 lll1'R thickness from (1) a 58-year-oldblackman,

(2) a 40-year-old black man, and (3) a 20-year-old white woman. Specimens were obtained shortly after death, stored refrigerated, and used within 7 to 10 days. Previous studies lO have shown that under these conditions the intrinsic skin permeability is not altered. In addition, when human skin specimens are used within this time, BPO during skin absorption is metabolized to benzoic acid and accumulated as such in the receptor fluid. Three or four replicates per human skin source were used, giving a total of 10 replicates. A 5 .ul volume of lotion containing 0.9 ,uCi (A) or 1.2 p,Ci (B) was spread over 1cm2 area of skin mounted on the diffusion cell. The diffusion cells were made of glass with flowthrough design. I I Buffered saline receptor fluid was maintained at 37 0 C and pumped at one reservoir volume (3 ml) per hour. Hourly reservoir samples were accumulated for 24 hours. After 24 hours ofapplication time, tritium content was determined in the accumulated reservoir fluid, skin surface washes and in the skin itself. The receptor phase samples from the permeation cells were analyzed for permeant content by scintillation counting (liquid scintillation counter, model 4640, Packard Instrument Co., Meriden, Conn.) within 24 hours after sampling. At the end ofthe monitoring period theskin surface was washed with soap (20% Ivory liquid) followed by two water rinses, and the wash solutions were analyzed for residual drug content by scintillation counting. The skin membrane itself was physically separated by heat l2 into epidermis and dermis, and each portion was completely solubilized with a tissue solubilizer (Soluene, Packard Instrument Co.). Aliquots of these homogenates were analyzed for drug content by scintillation counting. A mass comparison between the two test formulations was established for the amount of drug remaining on the skin surface, that which was retained within the skin, and that which permeated into the receptor phase. Material balance for BPO was determined for both formulations. 13 Statistical analyses were done by means of StatPlus software (Human Systems Dynamics, Northridge, Calif.) and an Apple lIe computer.

In vivo percutaneous absorption Female rhesus monkeys (n = 4) were placed in metabolic chairs and each topical dose was applied to a premeasured area of abdominal skin. A 75 ,ul volume containing 3.4 j.tCi (A) or 3.3 ,uCi (B) was spread over a 12 cm2 area of abdominal skin. The monkeys had free access to food and water but were restricted from touching their abdominal area by barrier plates. Urine was collected in·a pan under the metabolic chair. A barrier plate was placed between the dose application skin area on the monkey and the urine collection pan below the monkey to prevent contamination from applied material that might fall off the skin because of skin desquamation. Urine was collected for 24 hours in the metabolic chair. The skin site of application was washed with soap (20% Ivory liquid)

722

Journal of the American Academy of Dermatology

Wester et at. 1.12

0.90

C\i

E

,U

0

FOOM.LATIO'I F

0

FOOM.LATia'l G

0.67

OJ

.5 0

0,45

0.22

0.00 I ' l - - - - . - - J ' - - - - - - - - - ' - - - - - - ' - - - - - - ' - - - - - - ' 5,40 0.00 3.60 7.20 9.00 1.80

TIME .(hrs)

Fig. 1. Release profiles of BPO from formulations F and G. Results are the average of two determinations.

Table 1. In vitro percutaneous absorption of benzoyl peroxide in human skin % Applied doset Paramcter*

Formulation A (freely dispersed)

I Formulation B (entrapped)

Skin Epidermis Dermis

6.1 ± 4.4 5.2 ± 4.1 0.4 ± 0.4

1.4 ± 0.8 1.1 ± 0.7 0.1 ± 0.1

Skin edge

0.6 ± 0.7

0.2

Receptor fluid Surface wash Total

17.1 ± 15.7 3.5 85.6 ± 29.0 83.2 108.9 ± 22.3 88.1

Statistical comparison+

p = 0.01 p = 0.02 p > 0.05 (NS) ± 0.3 p > 0.05 (NS) ± 3.4 p < 0.05 ± 19.9 p> 0.05 (NS) ± 18.4 p > 0.05 (NS)

NS, Not statistically significant; skill edge. edge where system is clamped together. *N = 10 for each parameter. tData presented as mean ± SD. +Student's t test.

and two water rinses, and the monkeys transferred to metabolic cages for continued urine collection during the next 6 days.? Aliquots of urine were analyzed for tritium content by scintillation counting. Statistical analyses were done as described earlier.

Cumulative skin irritancy In rabbits. Two groups of New Zealand strain rabbits (six animals/group), male or female, weighing between 2 and 4 kg, were used. Group 1 received formula-

tions C, D, and E and group 2 received formulations C, F, and G. The hair from the backs of the animals was clipped 24 hours before the first test material application; 0.025 ml of each formulation was applied over a previously marked 2 cm2 area for 21 days,S days a week. Irritancy observations were recorded daily with the following scale: Grade 0, no skin reaction; grade 0.5, barely perceptible, spotty erythema; grade 1.0, mild, even erythema covering the whole area of application; grade 2.0, intense erythema with clearly defined edge; grade 3.0, intense erythema plus edema, raised edge; grade 4.0, intense erythema, edema, and bullae; grade 5.0, necrosis. In human beings. Formulations Hand J, containing 2.5% and 10.0% BPO, respectively, were tested on the backs of human volunteers in a 14-day cumulative irritancy open-patch test. 14 A panel of 25 volunteers was enrolled by a commercial testing laboratory (Derma-Test Laboratories, Inc., Long Island City, N.Y.). In this study, the test materials were applied to the subjects' backs daily, Monday through Friday, for 2 weeks. The dosage was 150 ,ul per patch. Patches were made ofnonwoven cotton cloth (Webril, Curity) 20 mm in diameter, covered with paper adhesive tape (Scanpor, Norgeplaster, Oslo, Norway), and held in place with occlusive plastic tape (Blenderm Surgical Tape, 3M Co., St. Paul, Minn.). Irritancy was scored daily, on a blind basis, according to the following scale: 0, no visible reaction; 1, mild erythema; 2, moderate erythema with edema; 3, intense erythema with edema; 4, intense erythema, edema, and vesicular erosion. The results were expressed as (1) the percentage of the total subjects showing a response and (2) the cumulative response index (eRI), calculated as the sum of the mean daily scores of all the subjects tested.

Volume 24 Number 5, Part I May 1991

Controlled release 0/ benzoyl peroxide 723 5 -A-

4 C\I

b3

.....

Ol

~

a

2

FORM. L

..,..,- FORM.

./

J

./'

-.- FORM. K

../

/'

-.- FORM. H

A

...---'

/~'

.----'~

---=,=====.

/ "/..'----=:.::::::::==' :::::---, .:::.---+

o o

~{~,~. :::::-:.~

3

2

4

5

6

7

TIME (hrs) Fig. 2. Release profiles of BPO from formulations Hand J compared with commercial products K and L. Results are the average of two determinations.

In vivo antimicrobial efficacy

Table II. In vivo percutaneous absorption of

Formulation G (5.0% entrapped BPO in SOWE) was initially studied.. Twelve healthy male and female informed. volunteers older than 18 years of age were selected.. The subjects were free of any cutaneous or internal diseases, and each had no known allergies or sensitivities to topical products and specifically to antiacne preparations. Subjects were screened bacteriologically to ascertain that they had P. acnes counts equal to or greater than 100,000 organisms per square centimeter. The formulation was applied twice daily, 5 days a week for 2 weeks under the supervision of the investigator (at Ivy Laboratories, Philadelphia, Pa.). Each subject was instructed to avoid the use of any other topical product on the face during the study period and against the use of medicated soaps. Quantitative bacteriologic cultures were obtained. according to the method of Williamson and Kligman 8 from the forehead before starting and again on days 7 and 14. Culture samples were obtained 24 hours after the last application of BPO and the culture fluids were monitored for drug carryover. Subsequently, formulations Hand J containing entrapped BPO at 2.5% and 10.0% concentrations were also tested for antimicrobial efficacy against P. acnes. The reduction in the number of organisms recovered from the skin surface and the changes in the free fatty acids/triglyceride ratio (FFA/TG) in skin surface lipids were measured..

benzoyl peroxide in the rhesus monkey

RESULTS The cumulative release of BPO from formulations F and G through the Silastic membrane is shown in Fig. 1. Apparent "zero order" release of BPO from these formulations is observed. This

% Dose absorbed

FormUlation A (freely dispersed)

Formulation B (entrapped)

1

16.7

4

13.9 13.4 14.8* ± 1.5

8.5 9.7

Monkey

2 3

Mean ± SD

15.3

8.4

7.7

8.6* ± 0.8

*p = 0.002.

result is expected because, as a result of the very low solubility of this drug in the formulation, maximal thermodynamic activity is maintained in the donor side throughout the experiment. When freely dispersed, BPO steady-state flux was 0.09 mg/cm 2/hr. When entrapped, the initial flux was practically identical (0.1 mg/cm2/hr). This similarity most likely reflects the small amount of BPO adsorbed onto the outside of the polymeric spheres. This hy~ pothesis was confirmed by observing the polymeric powder under the microscope with polarized light; a small amount of crystalline refringent material could be detected. When this pool of free BPO was depleted, the flux changed abruptly to 0.04 mg/ cm2/hr and remained at this rate for the next 30 hours. This flux represents the release of entrapped drug from the macroporous polymer system. The inflection point suggests initiation of controlled release. Fig. 2 represents the release profiles of BPO

Journal of the American Academy of Dermatology

724 Wester et af.

ill

600 _Form. C

I-

0

u

(fJ

(0.0% SPO)

500

~ Form. E

),

u c

o

....III 400 ~ ill

> :;:;

300

(5.0% BPO)

~ Form. F

"5

E

::J

(2.5% BPO)

~ Form. G

III

U

(2.5% BPO)

Form. D

(5.0% BPO)

200

),

III

0

,...

100

N

0 Group 1

Group 2

Fig. 3. Cumulative 21-day irritancy scores for various BPO formulations in rabbits. Results are the sum of total irritancy scores in each group. Six animals per group were used.

Table III. 14-Day cumulative irritancy in humans

Formulation

1. 2.5% BPO

Commercial product (K) Entrapped BPO (H) Vehicle II. 10% BPO Commercial product (L) Entrapped BPO (J) Vehicle

Cumulative % Total subjects with response positive response index*

36 12 0

1.04 (1) 0.24 (2) 0.0 (3)

52 24 0

2.59 (4) 1.64 (5) 0.0 (6)

*Duncan's Multiple Range test showed significant difference (p < 0.05) between 1 and 2, 1 and 3, 4 and 5, 4 and 6, 5 and 6; but no significant difference (p > 0.05) between 2 and 3.

from formulations Hand J compared with their commercial counterparts (formulations K and L). Table I presents the results for the in vitro percutaneous absorption of BPa through human skin. There was a statistically significant difference (p = 0.01) between formulations A and B in the percent of applied dose retained in the skin and accumulated in the receptor fluid. This difference supports the concept of controlled release of the BPa from the polymer delivery system into and through the skin. In both cases most of the applied dose was recovered in the surface washes. Most of the tritium content within the skin was in the epidermis, and total material accountability for both systems was good.

Table II reflects the in vivo percutaneous absorption of BPa in the rhesus monkey. Absorption from formulation A, with BPa freely dispersed in the lotion, was 14.8% ± 1.5%. Absorption of BPa entrapped in the macroporous polymer, fonnulation B, was significantly (p = 0.002) less at 8.6% ± 0.8%. For both formulations, most of the tritium was excreted in the first 24-hour urine sample. Fig. 3 presents the 21-day cumulative irritancy scores of the various formulations tested in rabbits. Entrapping BPa in a controlled-release polymer system significantly reduced the drug skin irritation potential both at 2.5% (p < 0.001) and at 5.0% (p < 0.001) BPa concentrations in the formulations. The results from the human cumulative irritancy test are presented in Table III. At each BPa concentration, the formulation containing entrapped BPa was significantly less (p < 0.05) irritating than its commercial counterpart. All the test samples (scrubs) obtained in vivo from the human volunteers at days 0, 7, and 14 were cultured for P. aenes and for aerobic bacteria. The number of P. aenes, expressed as 10glO per cm2 at baseline and on days 7 and 14 of treatment, are shown in Fig. 4 for the initial test. The mean P. aenes counts were significantly (p < 0.001) reduced over the baseline at both 7 and 14 days after treatment, with an average reduction of at least two orders of magnitude in the counts. Fig. 4 also shows approximately two orders of magnitude reduction in the aerobic bacteria (p < 0.001). When the reformulated products (formulations

Volume 24 Number 5, Part 1 May 1991

Controlled release ofbenzoyl peroxide 725 (II

E

o

10

0

...... (f)

:2 Z <{ l'J IT: 0 0 (f)

8

Aerobes

4

~

2

S1

..J

•

6

IT: U

l'J 0

P. acnes

0 BASELlf\E

1 'Ni:EK

2 WEEK

Fig. 4. Reduction in P. acnes and aerobe counts after treatment with BPO formulation G. Table IV. In vivo antimicrobial efficacy of formulations containing 2.5% and 10% entrapped BPO P. acnes count* (IOgl0 bacteria/cm2)

Formulation

2.5% BPO (H) 10% BPO (J)

Baseline

6.04

I

Week 2

I

2.75t

FFA/TG ratio*

Week 4

2.25t

Baseline

0.74

I

Week 2

0.45t

I

Week 4

0.32t

(± 0.65)

(± 1.82)

(± 2.01)

(± 0.28)

(± 0.19)

(± 0.23)

(± 0.51)

(± 1.89)

(± 1.85)

(± 0.19)

(± 0.17)

(± 0.11)

5.58

3.04t

1.92t

0.71

0.53t

0.31 t

*Data presented as mean ± SD. tDifference between this value and baseline is significant at the 95% confidence level (Dunnett's t tests).

Hand J) containing 2.5% and 10.0% BPO were tested for antimicrobial efficacy, significant reductions (p < 0.05) were observed both in the number of P. acnes and in the FFAjTG ratio in skin surface lipids (Table IV). No significant differences between treatments were detected at any time point. DISCUSSION

Because BPO pharmacologic activity is a consequence of its ability to penetrate into the skin preferentially through the follicular openings, a controlled-release topical delivey system might reduce the percutaneous absorption of BPO without affecting its intrafollicular penetration, thereby reducing the irritancy of the drug without sacrificing efficacy. Therefore we compared the skin absorption of BPO from topical formulations in which the drug was either freely dispersed or was entrapped in a controlled-release macroporous polymer delivery system. Statistically significant differences were found between these formulations in BPO content in the skin, receptor fluid accumulation, and in in vivo

skin penetration in the rhesus monkey. Release studies of BPO through artificial Silastic membranes and through excised human skin indicate a controlled release of BPO from the macroporous polymer system into and through the skin. Corresponding studies of rabbit skin irritation with freely dispersed and entrapped BPO showed reduced skin irritation with the entrapped/controlled-release system. This effect was further confirmed when formulations containing entrapped BPO at 2.5% and 10.0% concentrations were tested in comparison to their commercial counterparts for cumulative irritancy in human beings. Here again the controlledrelease formulation demonstrated a significantly lower irritancy potential than the conventional vehicles. However, when the experimental formulations were evaluated for antimicrobial activity in vivo, their efficacy was in line with that of conventional products (1. 1. Leyden, unpublished results). These results demonstrate that controlled release can provide the means for dissociating the dose relation between efficacy and irritation for BPO.

Journal of the American Academy of Dermatology

726 Wester et al. REFERENCES

1. Nacht S, Gans EH, McGinley KJ, et al. Comparative activity of benzoyl peroxide and hexachlorophene. In vivo studies against Propionibacterium acnes in humans. Arch Dermatol 1983;119:577-9. 2. Fulton JE, Bradley S. The choice of vitamin A acid, erythromycin, or benzoyl peroxide for the topical treatment of acne. Cutis 1976;17:560-4. 3. Kligman AM, Leyden JJ, Stewart R. New uses of benzoyl peroxide; a broad spectrum antimicrobial agent. Int 1 Dermatol 1977;16:413-7. 4. Nacht S. Methods to assess the transepidermal and intrafollicular penetration of anti-acne agents. In: The Proprietary Association, ed. Proceedings of the 1980 Research and Scientific Development Conference. New York City, 1981:88-91. 5. Leyden J1, McGinley KJ, Mills OH, et al. Topical antibiotics and topical antimicrobial agents in acne therapy. In: Juhlin LA, Rossman H, Strauss JS, cds. Symposium in Lund. Uppsala, Sweden: Uppland Grafisker AB, 1980:151-

8. 9.

10. 11.

12.

13.

M.

6. Fulton JE, Bradley S. Studies on the mechanism of action of topical benzoyl peroxide in acne vulgaris. J eutan Pathol 1974;1:191-4. 7. Wester RC, Maibach HI. Percutaneous absorption in the

14.

rhesus monkey compared to man. Toxicol Appl Pharmacol 1975;32:394-8. Williamson P, Kligman AM. A new method for the quantitative investigation of cutaneous bacteria. J Invest Dermatol 1965;45:498-503. Franz TJ. Percutaneous absorption. On the relevance of in vitro data. J Invest Dermatol 1975;64:190-5. Nacht S, Young D, Beasley IN Jr, et al. Benzoyl peroxide: percutaneous penetration and metabolic disposition. JAM ACAD DERMATOL 1981;4:31-7. Reifenrath WG, Spencer TS. Evaporation and penetration from skin. In: Bronaugh R, Maibach HI, eds. Percutaneous absorption: mechanisms-methodology I drug delivery. New York: Marcel Dekker, 1985:305-35. Shah VP, Skelly JP, Maibach HI, et al. Principles and practices of in vitro percutaneous penetration studies: relevance to bioavailability and bioequivalence determinations. Pharm Res 1987;4:265-7. Wester RC, Maibach HI. In vitro testing of topical pharmaceutical formulations. In: Bronaugh R, Maibach HI, eds. Percutaneous absorption: mechanisms-methodology, drug delivery. New York: Marcel Dekker, 1989:653-9. Nacht S. Efficacy of anti-acne drugs and formulations. Cosmetics & Toiletries 1986;101:47·55.