Benzoyl peroxide: Percutaneous penetration and metabolic disposition

Benzoyl peroxide: Percutaneous penetration and metabolic disposition Sergio Nacht, Ph.D., David Yeung, M.S., Joseph N. Beasley, Jr., B.S., Mark D. Anjo, M.A., and Howard 1. Maibach, M.D. Mr. Vernon, NY, and San Francisco, CA The rransepidermal penetration and metabolic disposition of {'IC-benzoyl peroxide were assessed in vitro (excised human skin) and in vivo (rhesus monkey). In vitro, the benzoyl peroxide penetrated into the skin, through the stratum corneum or the follicular openings, or both, and was recovered on the dermal side as benzoic acid. In vivo, benzoic acid was recovered from urine in amounts equivalent to 45% and 98% of the radiolabel following, respectively, topical and intramuscular administration of small amounts of HC-benzoyl peroxide. We conclude that benzoyl peroxide penetrates as such into the skin layers and is converted therein to benzoic acid, which, in turn is absorbed into the systemic circulation. Renal clearance of the metabolite is sufficiently rapid as to preclude its hepatic conjugation with glycine, since following topical administration to rhesus monkeys, no hippuric acid was found in the urine, as could have been expected had a significant amount of benzoic acid passed through the liver. (J AM ACAD DERMATOL 4:31-37, 1981.)

Benzoyl peroxide, a commonly used antiacne agent, is a potent antibacterial; it is thought to penetrate into the follicles to eradicate Propionibacterium acnes.":" This agent also is recommended for use in a variety of other dermatoses. 4 Despite widespread use, we know of no detailed studies dealing with the absorption of benzoyl peroxide through normal skin and its eventual systemic disposition. (A recent investigation of the absorption of benzoyl peroxide from leg ulcers following the application of a 20% benzoyl peroxide formulation did reveal plasma levels of benzoic acid. 5) To fill this apparent gap in the literature, we present in vitro and in vivo studies on the From Vick Toiletry Research Division, Richardson-Merrell Inc., Mt. Vernon, NY, and the Department of Dermatology, University of California, San Francisco. Reprint requests to: Dr. Sergio Nacht, Vick Toiletry Research Division, Richardson-Merrell Inc., One Bradford Road, Mt. Vernon, NY 10553.

0190-9622/81/010031 +07$00.70/0 © 1981 Am Acad Dermatol

absorption and metabolism of benzoyl peroxide; such information is of obvious usefulness in assessing the safety of the topical use of benzoyl peroxide. MATERIALS AND METHODS

Transepidermal penetration In vitro, we used the diffusion cell technic with excised human skin,6-8 while in vivo data were obtained with radiolabel recovery procedures on urine specimens from rhesus monkeys."

Diffusion cell studies Strips of normal human skin, 700 to 800 }.Lm thick, were obtained from the abdominal area of white adult individuals, as plastic surgery specimens, and were either used immediately or stored in such a manner as to maintain their integrity." Circular discs, 4.3 em in diameter, were punched from these strips and were equilibrated for 12 to 20 hours in phosphate-buffered Ringer's solution containing streptomycin and penicil-

31

32

Journal of the American Academy of

Nacht et al

Dermatology

Buffer*

Skin soaks**

Skin rinse*** Reference standards

Solvent front

;'(~f,§ Rr; 0.78 .:.~,~~ Benzoyl peroxide

(

) Rf; 0.17

•

1

Benzoic acid Origin

c.:::.. .)Rr;O

Shaded area indicates presence of radioactivity

Hippuric acid

* Materiai extracted from the buffer bathing the dermis

** Material extracted from the skin disc at the end of the study *** Material obtained by rinsing the skin surfaceat the endof the study

Fig. 1. Thin layer chromatography-autoradiogram from buffer solution, skin soaks, and rinses obtained from skin diffusion studies.

lin. All in vitro permeation studies were performed in duplicate. The diffusion cells consisted of two chambers: the lower one made of Plexiglas, the upper one of Teflon or Delrin. Each skin disc, with its stratum corneum side exposed to air, was mounted horizontally across the upper chamber of the cell by means of an O-ring seal. The disc formed a diaphragm between the two chambers of the diffusion cell, providing a diffusional area of 5.08 ern". When mounted, 40 ml of physiologic saline was added to the lower chamber. This buffer solution, bathing the dermal (receptor) side of the skin disc, was mixed continuously by a Teflon-coated magnetic rod driven by a 120-rpm electric synchronous motor located below the cell bath. The cell was equilibrated for 30 minutes in a water bath maintained by an immersion thermostatic pump at 30° ::': 004° C. 14C-Benzoyl peroxide was applied as a 10% suspension in an aqueous emulsion to the stratum corneum side of the skin disc in a dose of 900 J,Lg/cm2 and spread evenly with a Teflon policeman. This amount of the benzoyl peroxide preparation provided adequate coverage of the entire skin disc. The cell was stoppered to prevent solvent evapora-

tion, thus maximizing the delivery of 14C-benzoyl peroxide to the skin. At selected time intervals, aliquots of the receptor solution were withdrawn through a side tubing and analyzed for radioactivity by liquid scintillation counting. Care was taken to mai ntain "sink" conditions by preventing significant concentrations of tracer from accumulating on the receptor side of the disc, a condition which could lead to a backflow of radioactivity. Eight hours later, the receptor buffer solution was quantitatively transferred to a 200-ml separation funnel. The solution was brought to pH I with concentrated HCI, extracted three times with ethyl acetate, and the extracts pooled. Absence of radioactivity in the aqueous phase was confirmed by counting an aliquot of the buffer solution in a scintillation counter. * An aliquot of the pooled ethyl acetate was counted; the remainder of the extract was concentrated under vacuum in a rotary evaporatort to 3 to 5 ml. An aliquot was analyzed by thin layer chromatography (TLC) and by autoradiography. The applied material remaining on the surface of the skin discs was removed by carefully rinsing them with 50 ml of absolute ethanol. These rinses were passed through a Millipore filter, and a I.O-ml aliquot was analyzed by TLC and by autoradiography. The radioactive materials remaining within the skin layers were recovered by exhaustively extracting the skin membrane overnight with 50 ml of absolute methanol.f The methanol extract was filtered, a l.O-ml aliquot counted, and the remaining extract concentrated and analyzed by TLC and autoradiography. TLC was done on 20 x 20 em precoated Silica Gel 60, F254, O.25-mrn plates,§ using a mixture of toluene and methanol (9: 1) as the developing solvent. Authentic 14C-benzoyl peroxide and nonradioactive benzoic acid and hippuric acid were reference standards. After the TLC development, the plates were dried in air before examination under ultraviolet light. Benzoyl peroxide and benzoic acid spots were compared to those of the standard solutions and lightly circled (Fig. 1). Autoradiographs were obtained from TLC plates by exposing them to Eastman Kodak(TMl NS-5T x-ray film for 1 week. *Nuclear Chicago, Mark II Scintillation Counter, Searle Analytic, Inc., Des Plaines, lL. tBrinkman, Buchi Model R evaporator, Brinkman Instruments, Inc., Westbury, NY. :j:At! solvents used were of Spectrograde quality. §E. M. Laboratories, Elmsford, NY.

Volume 4

Benzoyl peroxide

Number I January, 1981

Urinary excretion studies The metab olic disposition of topically applied 14Cbenzoyl peroxide was determined by analyzing urine specimens collected at 0 to 6, 6 to 12, and 12 to 24 hours, and then dail y for 7 days after the topical application of 14C-benzoyl peroxide . All specimens were frozen until analyzed . Three monkeys, two males and one female, trained for metabolic studies, were used . Each animal was seated in a metabolism chair , and 139 J-tg of the carrier-free radiolabeled benzoyl peroxide dissolved in acetone was applied to the lightly clipped and shaved area of the ventral forearm . The monkey's hands were tied to the sides of the ch airs so that the medication could not be rubbed off. To prevent the labeled materi al from droppin g into the urine specimen, a physic al barrier sep arated the site of appl ication from the urine collecting cont ainer. Th e monke ys remained in the metabolism chairs for 24 hour s, during which time they were hand-feel and supplied with water. After this , the remaining 14C_ benzoyl peroxide wa s removed from the skin surface by carefull y washing the forearm with soap and water. The monkeys then we re transferred to indiv idual metabolism cages for the rest of the study . When the topical pha se of the in vivo stud y was completed, the an imals were rested for 4 weeks to allow any remaining urin ary radioactivity to return to background level. Then, as an internal control, each of the three monkeys was given an intramuscular injection of 91.3 u.g !'IC-benzoyl peroxide in propylene glycol to measure its elimination kinetics and to correct for possible storage in organs of the body. Urine specimens were collected for 7 days and frozen until analyzed. After thawing and before analy sis, the total volume of each specimen was recorded and a 1.0 m! aliqu ot counted by liquid scintillation .

Isolation and identification of metabolites Only those urine samples with the highest radio activity levels were used for metabolite identification . These specimens were extracted three times , using five parts of ethyl acet ate to o ne of urine , after the pH of the specimen was measured . Th e organic extracts were pooled, concentrated in a rotary evaporator, and an aliquo t of the concentrate and ano ther of the remaining aqueous phase were counted . Ne xt , the aqueo us phase was adju sted to pH 1 by the dropwise addition of I N HCl and it was extracted three times, using the same ratio of ethyl acetate to sample as before. The aqueous and organic extracts were measured for radioactivity.

33

Urine s amp les~ pH9

t

EtAc /5: 1) extraction Ix 31 I

41

'lr

OrganiCfraclion* Ipooledl

Aqueous fradi on*

~

Adj usted to pH I with IN Hel

t

EtAc (5:11 extraction (x 31

AqUeOUs~raction*

I

Organi c i ad ion* (pool ed)

-It

Evaporation under nitrogen (weight measured)

-It

TLC toluene/ met hanol (9:1)

t

Autorad iograms 14 days to 2 weeks exposure)

-It

EI ulion of radioactive areas" from TLC plates for quantitation

.J,

.

HPLC of the eluted fractions for final

identification ~

Aliquots counted for radioactivity by liquid scintillation for quantltauon

Fig. 2. Analytic procedure for the identification of 1·IC_ ben zoyl peroxide metabolites in urine from rhes us monkeys . The pooled organic fractio n was evaporated to dryne ss , the residue weighed and dissolved in 0.2 ml methanol, and the resulting solution spotted on a TLC plate, along with standards of benzoic acid, hippuric acid , and benzoyl peroxide. After the plates had been developed with the 9: 1 mixture of toluene-methanol , they were dried and autoradiograms obtained from each plate for 4 days to 2 weeks , depending on the am ount of radioactivity present. The R, for each radioact ive area was measured on the autoradiogram and comp ared to the values obtained for the authentic standards . The material from the ra dioactive are as was eluted from the plates with eth anol , and the ethanol solution was passed through a sintered glass filter and counted . An aliquot was injected into a High Performance Liquid Chrom atograph y (HPLC)* unit. Th e mobile phase was a 1 : I mixture of methanol and water; the column was a Microbondapack-Phenyl Reverse Phasej ; the flow rate was adjusted at 2 .0 mllm in; and the UV detector was set at 254 nm. R etention times were measured and compared to those of the refer ence sta ndards to verify the identity of the compounds recovered. The ' Spectra-Physics Inc., Santa Clara , CA . t Waters Associates, Milford , MA.

Journal of the American Academy of Dermatology

34 Nacht et al

Table I. Percutaneous penetration of benzoyl peroxide through excised hum an skin

___________.. .I Rinse (skin surface) Soaks (skin layers) Buffer (dermal side) Total

J

JL_g_*_ _

%t

4,350

I

Chemical identity

95.5 2.6

120 86

Benzoyl peroxide Benzoyl peroxide and benzoic acid Benzoic acid

1.9

4,556

100.00

*Average of two diffus ion cells. tpercent of the total amount recovered.

Table II. Distribution of radioactivity in extracts of urine from rhesus monkeys following intramuscular and topical admini stration of HC-benzoyl peroxide Intramuscular* administration Fraction

pH

Sex

9 9 9

M M F

Mean M M F

Mean

Aqueous

I

Organict

I

Topicalt administration

I

I

Total recovery (%)§

Aqueous

93.0 92.4 93.1

7 .0 7.6 6.9

91.8 89 .0 97.2

Organic

Total recovery (%)

93.0 90.1 94.6

7.0 5.4

96.1 96.0 93.7

92 .6

7.4

95.3

92.8

7.2

92.7

12.0 0.0 3.3

88.0 100.0 96 .7

92.6 73.0 96.2

6.0 0.0 3.0

94.0 100 .0 97 .1

92.3 71.0 96.6

5.1

94 .9

87.3

3.0

97 .0

86.6

9.9

- Intramuscular injection of a sm all dose of He-benzoyl peroxide; 0-6 hour speci me ns studied. t Topica\ application of a tracer dose of He-b enzoy l peroxide in acetone; 6-12 hour spec imen from female and 6-24 hou r urines from male monkeys were studied. tEthyl acetate fraction. § B,t~ ed

on the total amount of radioactivity present originally in the urine sample .

extraction and analytic schemefor these samples is presented in Fig. 2.

RESULTS Penetration through excised human skin All radioactive spots matched exactly areas that had been circled on the TLC plates under UV light (Fig. 1). These areas were identified as either benzoyl peroxide or benzoic acid by comparing their R, values with the standards . Of the 4 ,556 fLg of radiolabeled benzoyl peroxide applied to the surface of the skin discs , 1.9 % (86 j.tg) penetrated through the skin and was re covered as benzoic acid from the receptor buffer solution . The skin discs contained 2.6% of the radiolabel (120 fLg), approximately one half of which was benzoyl peroxide and one half benzoic acid. Thus, the total amount of benzoyl peroxide

delivered to the skin under maximal "steady state" conditions was 206 f-I.-g in an 8-hour period, or 4.5 % of the applied dose. This represents a penetration of 5 .1 j.tg/cm 2/hour . Most (95 .5% ) of the benzoyl peroxide originally applied topically was recovered unchanged from the skin surface at the end of the 8-hour period (Table I).

Metabolic studies Plots of the urinary excretion of radioactivity with time are presented in Fig. 3 . In all cases , excretion occurred more rapidl y and peaked earlier following intramuscular injection than after the topical administration of tolC-benzoyl peroxide. The largest amount of l-le was excreted between 0 and 6 hours after injection. (These samples were used for the identification of metabolites of benzoyl peroxide.) In all cases, the urine spec-

Volume 4 Number j January, 1981

Benzoyl peroxide

106

Monkey 'iH (female)

Monkey *2 (male)

35

Monkey #3 (male)

-Intramuscular injection --0

Topical application

1\\

, / I

\

v~ 104

\ \

Fig, 3. Urinary excretion of He after the intramuscular injection and the topical application of radio labeled benzoyl peroxide to two male and one female monkey.

imens collected on day 7 were found to contain no radioactivity. The metabolic disposition of benzoyl peroxide was similar in all three monkeys used in these studies. The data presented in Table II are the results in the two male and in the one female monkey. No sex-related differences were observed. Overall, 95.3% of the radiolabel excreted in urine following intramuscular injection was recovered after ethyl acetate extraction at the natural pH of the urine. Of this, 92.6% was in the aqueous phase and 7.4% in the ethyl acetate fraction. When the aqueous phase was re-extracted after adjusting at pH 1.0, 94.9% of the radiolabel was transferred to the organic fraction, leaving 5.1 % in the aqueous phase. Analysis of the second organic extract by TLC and autoradiography revealed only a single radioactive entity. The relative R,. of this metabolite was 0.4, the same as that for the benzoic acid standard. Under these conditions the R, for intact benzoyl peroxide was 0.81, while hippuric acid remained at the origin (R, = 0). HPLC analysis of this metabolite eluted from the TLC plates yielded a single metabolite with a retention time identical to that of benzoic acid. Therefore, ben-

zoic acid was the only species recovered from urine after intramuscular administration of benzoyl peroxide. When 14C-benzoyl peroxide dissolved in acetone was applied in a dose of 4 p.,g/cm 2 to a 35 ern" area of the ventral forearm of these same monkeys, about 45% of the applied radioactivity was recovered in the urine. This represents a percutaneous absorption of about 1.8 p.,gl em! of benzoyl peroxide in a 24-hour period since, at the end of this time, the remaining nonabsorbed drug was removed from the surface of the skin by washing the site of application with soap and water. The excretion pattern of radioactivity (Fig. 3) resembled that seen following the intramuscular injection of benzoyl peroxide, but peak excretion was delayed. The largest amounts of 14C were found in the 12- to 24-hour and the 24- to 48-hour urine collections. These samples were pooled and used in the metabolite identification studies. When the urine collected following topical application was extracted with ethyl acetate at pH 9.0, 7.2% of the 14C label was recovered in the organic fraction; 92.8% remained in the aqueous phase (Table II). Re-extracting the aqueous fraction at pH 1.0

36 Nacht et at

transferred 97.0% of the radioactivity to the organic fraction; only 3.0% remained in the aqueous phase. When this second organic extract was analyzed by TLC and autoradiography, one major metabolite with a relative R, identical to benzoic acid was found. There were three minor, more polar metabolites, but the quantity of these was too small-less than 2% of the total-to permit their identification. However, their mobility did not coincide with that of hippuric acid. To assess potential hydrolysis of hippuric acid during the extraction procedure, this compound was added to urine samples, which were then submitted to the same analytic scheme as in the I~C study. All the hippuric acid added was recovered as such, and no benzoic acid was detected in these samples. This indicates that no hydrolysis took place during sample processing. DISCUSSION

The results from cell diffusion studies using excised human skin indicate that topically applied benzoyl peroxide penetrates unchanged through the stratum corneum or into the follicles. The drug diffuses into the epidermis and the dennis , where it is converted, perhaps with the intervention of follicular bacteria, to benzoic acid . This metabolic conversion is apparently completed within the skin because the receptor buffer solution bathing the dermis contained only benzoic acid, the end product of benzoyl peroxide . We conclude that only benzoic acid enters the systemic circulation and does so by diffusing into the blood vessels of the dermis. In vivo studies in rhesus monkeys confirmed the conversion of benzoyl peroxide to benzoic acid after intramuscular and topical administration. At least 95% of the radioactivity in urine following either mode of administration was recovered as benzoic acid. Three minor, more polar, and as yet unidentified metabolites were found after topical, but not after intramuscular, administration; these accounted for less than 5% of the J 4 C label in urine. The delayed excretion of the radiolabel following topical application , as compared with intramuscular injection, might be attributed to the longer lag time required for percutaneous absorp-

Journal of the American Academy of Dermatology

tion than for intracutaneous metabolism of the drug. Holzmann et al" found that after the application of large amounts of benzoyl peroxide to leg ulcers (20 mg/cm'Vday) , benzoic acid was identified in plasma. However, it took 3 days of continuous application for the benzoic acid plasma levels to become detectable. This unusually long delay could be due in part to the low sensitivity of the analytic technic used (100 u.g benzoic acid/ml plasma). In OUI' diffusion studies, the maximum flux of benzoyl peroxide through intact human skin was 2.1 ,ug/cm 2/hour. If the drug was applied to 20 ern" of skin, even assuming that all the benzoic acid that penetrated in I hour was still present in plasma, at an average plasma volume of 3,000 ml , this would only resu It in a benzoic acid plasma concentration of about 10 ng/ml, much too low to be detected chemicall y . It is also possible that, as suggested by Holzmann et aI, ~ impaired blood flow in the area of the ulcer might have contributed to delayed metabolism and systemic absorption of the benzoyl peroxide. In this study, we found that the benzoic acid generated in the metabolic disposition of benzoyl peroxide was excreted in urine as such , and not conjugated with glycine as hippuric acid (benzoyl glycine), as it is after oral ingestion in humans, in the rhesus monkey, and in certain other mammalian species.!" This is not a totally unexpected finding. Drugs not absorbed from the stomach, intestines, or rectum can avoid first passage through the liver, where the conjugation of benzoic acid with glycine takes place. In fact, the portal circulation is deliberately bypassed in some cases by administering a drug sublingually or by the buccal route, so that it is absorbed directly into the systemic circulation through the oral mucosa, 11 thereby circumventing first passage through the liver. In the case of benzoic acid formed by the intracutaneous metabolism of topically applied benzoyl peroxide, absorption occurs as benzoate through the blood vessels in the dermis. Once this metabolite enters the systemic circulation , it is promptly transported to the kidneys where it is excreted in the urine unchanged. None is recircu-

Volume 4 Number J January, 1981

lated to the liver for conversion to hippuric acid. In this study , und er practically " occlusive" conditions , the maximal rate of transfer of benzoyl peroxide to the skin was 5, I fLg/cm 2 /hour; of this, only 2 . \ I-tg/cm 2 /hour appeared on the dermal side. Calculations indicate that , even under unrealistic conditions of excessive topical application , the systemic absorption of benzoyl peroxide (or of its metabolite, benzoate) cannot be greater than 500 mg per day, an amount easily handled metabolically by the mammalian organism .!" Based on these find ings , topical applications of benzoyl perox ide c an be expec ted to be free of toxic systemic effects and should not result in the accumulation of metabolites in body tissues . Holzmann et aP reached a similar conclusion based on their results obtained with a 20% benzoyl peroxide preparation . In conclusion, we hav e sho wn that benzoyl peroxide applied topically penetrated unchanged through the stratum corneu m or follicular openings of excised human skin and was converted metabolically to benzoic acid within the skin. We further have shown in rhe sus monkeys in vivo that this benzoic acid is systemicall y absorbed as ben zoate and rapidly excreted in the urine unchanged , without being conjugated to hippuric acid, as would occur foll owing oral administration.

Benzoyl peroxide

37

REFERENCES I. Nacht S, Gans EH , McGinley KJ , Kligman AM: Comparative activity of hexachlorophen e and of benzoyl peroxide against Propionibacterium acn es . (In press.) 2. Fulton JE, Bradley S: Th e choice of vitamin A acid, erythromycin, or benzoyl peroxide for the topical treatment of acne. Cutis 17:560-564 , 1976 . 3. Lyon s RE: Comparative effe ctiveness of benzoyl peroxide and tretinoin in acne vulgaris. In t J DermatoJ 17:246-25 1, 1978 . 4. Kligman AM, Leyden 11, Stewart R: New uses for benzoyl peroxide: A broad spectrum antimicrobial agent. Int J Dermatol16:413-417, 1977. 5. Holzmann H, Morsches B, Benes P: The absorption of benzoyl peroxide from leg ulcers. Arzneim Forseh 29: 1180-1183 , /979 . 6. Nacht S, Yeung D, Beasley IN Jr, Anjo MA, Maibach HI: Benzoyl peroxide: In vi vo and in vitro skin penetration and metabolic disposition . Clin Res 27: 533A , 1979 , 7. Marzu lli FN: Barriers to skin penetration , J Invest Dermatol 39:387-393 , 1962. 8. Galey WR, Lonsdale HK , Nacht S: The in vitro permeability of skin and buccal mucosa to selected drugs and tritiated water. J Invest Derm atol 67:713-717, 1976. 9. Weste r RC. Noonan PK, Maibach HI: Recent advances in percutaneous absorption using the rhesus monkey model. J Soc Cosmet Chern 30: 297-307 , 1979. lO. La DlI BN. Mandel HG, W ay EL: Fundamentals of drug metabolism. and drug disposition. Baltimore, 1971, The Williams & Wilkins Co, 11 . GibaJdi M, Kanig JL: Absorption of drugs through the oral mucosa. J Oral Ther Pharrnacol 1:440-450, 1965.