COMMUNICATIONS

The Effect of Methoxsalen Dose on Ultraviolet-A-InducedErythema

Sally H. Ibbotson, Robert S. Dawe,* and Peter M. FarrDepartment of Dermatology, Royal Victoria In®rmary, Newcastle upon Tyne, and *Photobiology Unit, University Department of Dermatology,

Ninewells Hospital and Medical School, Dundee, U.K.

There is considerable interindividual variation inbioavailability of Methoxsalen (8-methoxypsoralen)after ingestion of the standard dose used in photo-chemotherapy (psoralen plus ultraviolet A). A dosechange may be used to alter the degree of photo-sensitivity, although there is limited information onthe effect of 8-methoxypsoralen dose alterations onphototoxicity within individuals. We studied theeffect of changes of 8-methoxypsoralen dose over anarrow range in 15 subjects with psoriasis. Twohours after ingestion, serum 8-methoxypsoralen con-centration was determined and phototesting wasperformed at 350 6 30 nm (0.45±14 J per cm2). Theminimal phototoxic dose at 72 h was recorded,erythema was measured using a re¯ectance instru-ment, and dose±response curves were constructed.Each subject was tested on three occasions usingdoses of 25 mg per m2 (conventional dose) or con-ventional dose 6 10 mg. Median serum 8-methoxy-psoralen concentration increased from 96 to 143 to229 ng per ml with dose increases from conventional

dose ± 10 mg to conventional dose and conventionaldose + 10 mg, respectively (p < 0.001). The medianminimal phototoxic dose and D0.025 (the objectiveequivalent of the minimal phototoxic dose derivedfrom the dose±response curve) were signi®cantlyreduced with increasing 8-methoxypsoralen dosefrom conventional dose minus 10 mg (minimalphototoxic dose 1.7 J per cm2; D0.025 2.8 J per cm2)to conventional dose (1.2; 1.4 J per cm2) and con-ventional dose plus 10 mg (0.9; 1.0 J per cm2) (p< 0.001). Change in 8-methoxypsoralen dose had nodetectable effect on the maximum slope of the psor-alen plus ultraviolet A erythema dose±responsecurve. Thus, 8-methoxypsoralen dose changes withinindividuals, over a narrow but clinically relevantrange, signi®cantly altered the threshold response topsoralen plus ultraviolet A erythema but not the rateof increase in erythema with increasing ultraviolet Adose. Key words: intraindividual/8-methoxypsoralen/phototoxicity/photochemotherapy. J Invest Dermatol116:813±815, 2001

Controlled phototoxicity is used in the treatment ofchronic in¯ammatory dermatoses such as psoriasis.Cutaneous sensitization to ultraviolet A (UVA) canbe achieved by systemic ingestion of psoralens inpsoralen plus UVA (PUVA) photochemotherapy

(Fitzpatrick et al, 1955). In Europe, microcrystalline 8-methoxy-psoralen (8-MOP; Methoxsalen) is given at a standard dose of0.6 mg per kg body weight or, more accurately, 25 mg per m2

body surface area (Sakuntabhai et al, 1995). There are no clinicalstudies, however, to support this 8-MOP dosage regime as beingoptimal for therapeutic ef®cacy or for minimizing side-effects suchas erythema. There is considerable variation in interindividual and,to a lesser extent, intraindividual bioavailability of 8-MOP after astandard oral dose and serum concentration is a poor indicator oftissue levels (Lauharanta et al, 1982; SchaÈfer-Korting and Korting,1982; Herfst and de Wolff, 1983; Marrakchi et al, 1991; Donath etal, 1999). It appears that tissue 8-MOP concentration is a better

indicator of PUVA erythemal sensitivity than plasma levels (Yeo etal, 2000). In addition, there is an inverse relationship betweenserum 8-MOP concentration and UVA dose required to causeminimal erythema (minimal phototoxic dose, MPD) (Ljunggrenet al, 1981; Goldstein et al, 1982; Honigsmann et al, 1982). Noinformation exists, however, concerning the effect of 8-MOP dosechanges over a narrow, clinically relevant range on the nature of theUVA erythemal response in psoralen-sensitized skin withinindividuals and this was the purpose of the study.

MATERIALS AND METHODS

Subjects Fifteen adults with chronic plaque psoriasis, about to startsystemic 8-MOP PUVA (10 female; median age 44 y; range 24±77 y;skin types I (n = 1), II (n = 7), and III (n = 7)), were studied withethical approval. None of the subjects had a history of photosensitivity,and none was receiving medication or UV therapy prior to the study.Erythemal responses to PUVA were measured on the untanned normalskin of the mid-back, avoiding the midline, in each subject.

Photoirradiation equipment An irradiation monochromator (Diffeyet al, 1984a), optically coupled to a liquid-®lled light guide, was used toachieve a uniform irradiation ®eld of 1 cm diameter.

Serum 8-MOP concentration Two hours after ingestion of 8-MOP(Methoxsalen, microcrystalline preparation, Crawford Pharmaceuticals,

Manuscript received August 11, 2000; revised January 31, 2001;accepted for publication February 15, 2001.

Reprint requests to: Dr. Sally H. Ibbotson, Photobiology Unit,University Department of Dermatology, Ninewells Hospital and MedicalSchool, Dundee DD1 9SY, U.K. Email: s.h.ibbotson@dundee.ac.uk

Abbreviation: MPD, minimal phototoxic dose.

0022-202X/01/$15.00 ´ Copyright # 2001 by The Society for Investigative Dermatology, Inc.

813

Milton Keynes, U.K.), a 20 ml blood sample was taken and serum wasprepared by centrifugation at 3000 rpm for 15 min at 4°C. Samples werekept at ±20°C until assayed and then extracted into hexane following theaddition of 5-methoxypsoralen as internal standard. The hexane extractwas evaporated to dryness under nitrogen, and the residue was dissolvedin a mobile phase (40% acetonitrile, 60% deionized water) andchromatographed on a reverse phase high performance liquid chromato-graphy system using a UV/vis detector at 247 nm (P. Barn®eld,Department of Chemical Pathology, St. Thomas's Hospital, London,U.K.).

8-MOP dose range Each patient was tested on three occasions using8-MOP doses of 25 mg per m2 (equivalent to 0.6 mg per kg), calculatedto the nearest 10 mg (Sakuntabhai et al, 1995) (conventional dose, CD),or CD 6 10 mg. The order of dose allocation was randomly ascribed foreach patient.

PUVA phototesting Two hours after ingestion of 8-MOP, 11adjacent sites on the mid-back were exposed to a geometric series ofdoses at 350 6 30 nm (dose range 0.45±14 J per cm2, increment factor1.4). Under these conditions, the spectral power distribution of theirradiation monochromator approximates that of a typical PUVA typeUVA ¯uorescent lamp (e.g., Philips TL-09) (Sakuntabhai et al, 1993a).The high irradiance of the monochromator (approximately 110 mW percm2) allowed considerably shorter exposure times than would have beenrequired had a conventional PUVA lamp been used.

Erythemal response The dose of UVA radiation required to causefaint but easily discernible erythema in psoralen-sensitized skin wasde®ned as the MPD and was measured at 72 h. Following each of thethree 8-MOP doses, the visual MPD was determined and erythemameasurements were made in triplicate at each irradiated site and atadjacent control sites, using a re¯ectance instrument (Diffey et al, 1984b).In each subject, at each of the three 8-MOP doses, the increase (DE) inmean erythema index compared with nonirradiated skin was plottedagainst the logarithm of the UVA dose and a logit function was ®tted tothe data to construct dose±response curves (Diffey and Farr, 1991). Thedose of radiation required to cause an increase in erythema index of0.025 (D0.025), clinically equivalent to the MPD, and the maximumslope were calculated for each dose±response curve (Fig 1).

Statistical analysis Values given are median (range). The serum8-MOP concentration, MPD, and D0.025 values for each of the threedoses were compared using within-subjects analysis of variance with log-transformed data. Follow-up comparisons were made using Tukey'sprocedure for multiple comparisons. Within-subjects analysis of variancewas used to compare the slopes of the dose±response curves at the three8-MOP doses. A value of p < 0.05 was taken to be of statisticalsigni®cance. A linear regression model, with no assumption ofindependence for within-subjects repeated measures and with robustvariance estimates, was used to assess how much of the variation in log-

transformed MPD and D0.025 could be explained by variation in the8-MOP serum level. Stata statistical software (Intercooled Stata forWindows, release 6; Stata, Texas, 1999) was used.

RESULTS

Serum 8-MOP concentrations Two hours after ingestion of8-MOP, serum concentrations of 8-MOP were measured in eachof the 15 subjects on three separate occasions after different oraldoses (Fig 2). 8-MOP concentration increased from a median(range) value of 96 (45±199) to 143 (65±359) to 229 (133±347) ngper ml with dose increases from CD ± 10 mg to CD andCD + 10 mg, respectively (p < 0.001). The wide interindividualvariation in bioavailability of 8-MOP after each of these standardoral doses is shown (Fig 2).

Serum 8-MOP concentrations and MPD and D0.025 Withincreasing serum 8-MOP concentration, the MPD and D0.025

values fell signi®cantly (p = 0.002, p < 0.001, respectively). Only30% of the variation in MPD and D0.025 could be explained byvariation in serum 8-MOP concentration, however.

MPD There was a signi®cant reduction in median (range) MPDwith increasing oral 8-MOP dose from CD ± 10 mg (1.7; 0.9±5.0)to CD (1.2; 0.8±1.7) and CD + 10 mg (0.9; 0.5±1.7 J per cm2) (p< 0.001) (Fig 3a). These data also demonstrate the variation inMPD values at each of the 8-MOP doses.

PUVA erythema dose±response curves A signi®cantreduction in median (range) D0.025 value occurred when the oral8-MOP dose was increased from CD ± 10 mg (2.8; 0.8±8.2) toCD (1.4; 0.7±2.6) and CD + 10 mg (1.0; 0.4±2.5 J per cm2)(p < 0.001) (Fig 3b). The wide variation in individual D0.025 valuesat each of the oral 8-MOP doses is also shown (Fig 3b). There wasno detectable effect of the 8-MOP dose on the maximum slope ofthe PUVA erythema dose±response curves [CD ± 10 mg (median217; range 77±617) to CD (249; 180±372) and CD + 10 mg (256;184±393) (n = 15; p = 0.07)].

Figure 1. A typical PUVA erythema dose±response curve in onesubject. Each point represents the increase in erythema index comparedwith unirradiated sites (DE; mean 6 SD). The line drawn through thedata points is the logit function obtained by regression analysis. TheD0.025 (as indicated) is the dose of UVA radiation required to cause anincrease in erythema index of 0.025.

Figure 2. Serum 8-MOP concentration increased with increasingoral 8-MOP doses. There was a signi®cant difference in concentrationsacross the three doses (p < 0.001). The differences in concentrations forthe doses CD ± 10 mg and CD (p = 0.004), and CD ± 10 mg andCD + 10 mg (p < 0.001) were also highly unlikely to be chance®ndings. The difference in 8-MOP concentrations between CD andCD + 10 mg was less clearly signi®cant, however (p = 0.15).

814 IBBOTSON ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

DISCUSSION

We have assessed the effects of small, clinically relevant 8-MOP dosechanges on the characteristics of PUVA erythema 72 h after UVAexposure. The ®ndings relate to the use of microcrystalline 8-MOPpreparations. There was a signi®cant reduction in the dose of UVArequired to cause threshold erythema in psoralen-sensitized skinwhen the dose of 8-MOP was increased by 10 mg increments. Thiswas supported by the ®ndings of signi®cant reductions in D0.025

values, which are the dose±response curve equivalent values of theMPD.

There is limited information that, over a wide 8-MOPconcentration range, there is an inverse correlation with theMPD, and the ratio of 8-MOP concentration to MPD is constantat higher 8-MOP doses (Ljunggren et al, 1981). In that study theeffect of large variations in oral 8-MOP dose on the MPD wasexamined within individuals, however, and at the lowest 8-MOPdose the ratio of serum concentration and MPD was not constant,which may re¯ect the strong ®rst pass metabolism of 8-MOP(Brickl et al, 1981). Additionally, a signi®cant association existsbetween erythemal sensitivity, 8-MOP dose, skin type, and pasthistory of PUVA treatment (Sakuntabhai et al, 1994).

The ®ndings of this study suggest that the common practice ofgiving a 10 mg greater dose of 8-MOP to those in whom aphototoxic response is undetectable, on phototesting after the CD,is appropriate as the MPD and D0.025 would be expected to reduceby 25% and 29%, respectively.

The susceptibility to burning in the clinical setting is related to thesteepness of an individual patient's PUVA dose±response curve(Sakuntabhai et al, 1993b). Our ®ndings did not show a signi®canteffect of these relatively small dose changes on the maximum slope ofthe UVA erythema dose±response curve, however. The study wassmall and we may not have detected a psoralen dose effect on the rateof increase in erythema response to increasing UVA doses, for thisreason. The absence of a detectable, large effect on the slope ofresponse, however, calls into question the value of a small (10 mg)reduction in 8-MOP dose for patients experiencing painfulerythema (``burning'') episodes during treatment. A larger dosereduction may be of bene®t, or a greater reduction in UVA dose maybe more appropriate.

In conclusion, we have found that small variations in 8-MOP doselead to changes in the threshold erythema response (MPD; D0.025)that are likely to be clinically important but do not have detectableeffects on the maximum slope of the PUVA erythema dose±responsecurve.

We would like to thank Peter Barn®eld of St. Thomas's Hospital in London for the

performance of high performance liquid chromatography analysis of serum 8-MOP

levels. We would also like to thank the Newcastle University Hospitals Special

Trustees for funding this project.

REFERENCES

Brickl R, Schmid J, Koss FW: Considerations of pharmacokinetics of psoralens afteroral administration and conclusions for therapy. Proc Int Psoralens SIR 263±268,1981

Diffey BL, Farr PM: Quantitative aspects of ultraviolet erythema. Clin Physiol Meas12:311±325, 1991

Diffey BL, Farr PM, Ive FA: The establishment and clinical value of a dermatologicalphotobiology service in a district general hospital. Br J Dermatol 110:187±194,1984a

Diffey BL, Oliver RJ, Farr PM: A portable instrument for quantifying erythemainduced by ultraviolet radiation. Br J Dermatol 111:663±672, 1984b

Donath P, Bethea D, Amici L, et al: Low and irreproducible methoxsalen levels inpatients receiving photochemotherapy. Arch Dermatol 135:604±606, 1999

Fitzpatrick TB, Hopkins CE, Blickenstaff DD, Swift S: Augmented pigmentation andother responses of normal human skin to solar radiation following oraladministration of 8-methoxypsoralen. J Invest Dermatol 25:187±190, 1955

Goldstein DP, Carter DM, Ljunggren B, Burkholder J: Minimal phototoxic doses and8-MOP plasma levels in PUVA patients. J Invest Dermatol 78:429±433, 1982

Herfst MJ, De Wolff FA: Intraindividual and interindividual variability in 8-methoxypsoralen kinetics and effect in psoriatic patients. Clin Pharmacol Ther34:117±124, 1983

Honigsmann H, Jaschke E, Nitsche V, Brenner W, Rauschmeier W, Wolff K: Serumlevels of 8-methoxypsoralen in two different drug preparations: correlationwith photosensitivity and UV-A dose requirements for photochemotherapy. JInvest Dermatol 79:233±236, 1982

Lauharanta J, Juvakoski T, Kanerva L, Lassus A: Pharmacokinetics of 8-methoxypsoralen in serum and suction blister ¯uid. Arch Dermatol Res273:111±114, 1982

Ljunggren B, Bjellerup M, Carter DM: Dose±response relations in phototoxicity dueto 8-methoxypsoralen and UV-A in man. J Invest Dermatol 76:73±75, 1981

Marrakchi S, Decloquement L, Pollet P, et al: Variation in 8-methoxypsoralenpro®les during long-term psoralen plus ultraviolet A therapy and correlationsbetween serum 8-methoxypsoralen levels and chromametric parameters.Photodermatol, Photoimmunol Photomed 8:206±211, 1991

Sakuntabhai A, Farr PM, Diffey BL: PUVA erythemal sensitivity depends on plasmapsoralen concentration and UV sensitivity. Br J Dermatol 128:561±565, 1993a

Sakuntabhai A, Sharpe GR, Farr PM: Response of psoriasis to twice weekly PUVA.Br J Dermatol 128:166±171, 1993b

Sakuntabhai A, Matthews JNS, Farr PM: Improved prediction of the minimalphototoxic dose in PUVA therapy. Br J Dermatol 130:604±609, 1994

Sakuntabhai A, Diffey BL, Farr PM: Calculation of 8-methoxypsoralen dose accordingto body surface area in PUVA treatment. Br J Dermatol 133:919±923, 1995

SchaÈfer-Korting M, Korting HCh: Intraindividual variations of 8-methoxypsoralenplasma levels. Arch Dermatol Res 272:1±7, 1982

Yeo U-C, Shin J-H, Yang J-M, Park K-B, Kim M-M, Bok H-S, Lee E-S: Psoralen-ultraviolet A-induced erythema: sensitivity correlates with the concentrationsof psoralen in suction blister ¯uid. Br J Dermatol 142:733±739, 2000

Figure 3. The MPD and D0.025 values reducedwith increasing oral 8-MOP doses. (a) MPD and(b) D0.025. There was a signi®cant difference inMPD and D0.025 across the three doses (p< 0.001) and between CD ± 10 mg and CD (p< 0.001; p = 0.002 for MPD and D0.025,respectively). The difference between CD andCD + 10 mg for MPD and D0.025 was less clearlysigni®cant (p = 0.03; p = 0.07, respectively).

VOL. 116, NO. 5 MAY 2001 METHOXSALEN DOSE EFFECT ON PUVA ERYTHEMA 815