Determination of UVA protection factors by means of immediate pigment darkening in normal skin Kays H. Kaidbey, MD,a and Angelit Barnes, BAh Philadelphia, Pennsylvania A method is described for screening potentially useful photoprotective agents against UVA radiation by the use of immediate pigment darkening as an end point. Threshold doses ofimmediate pigment darkening showed a log normal distribution and the response was found to obey dose-reciprocity at irradiance levels below 50 mWjcm2• With this procedure, several marketed sunscreens containing benzophenone-3 as the only UVA absorber were found to have poor UVA protection factors, whereas those containing combinations ofbenzophenone3 and butyl methoxydibenzoyl methane or melanin were more effective. There was no correlation between the sun protection factor cited on the label and the calculated immediate pigment darkening-protection factor. (J AM ACAD DERMATOL 1991;25:262-6.) Despite mounting concerns and considerable publicity about exposure to UVA, it is still unknown whether these wavelengths do, in fact, contribute to the damaging effects of sunlight. Almost all studies clearly implicate UVB as being predominantly responsible for chronic actinic damage. l -S However, it has been argued that the relative abundance ofUVA in solar radiation and consequently the excessive doses received during many decades may contribute to photod(l.mage and carcinogenesis. Studies in hairless mice have shown that both dermal damage and tumor induction by UVA are dose-dependent. 6,7 Most currently available sunscreens offer little protection against UVA, as evidenced by their lack of effectiveness in patients with photosensitivity disorders. In contrast to the sun protection factor (SPF), there is as yet no standardized method for measuring photoprotection against UVA. The determination of a UVA protection factor poses special difficulties. Because large doses are required to produce erythema, it is impractical to use redness as an end point. Topical photosensitizers, notably psoralens, were therefore used to reduce the threshold dose for UVA erythema. 8- 10 However, these agents produce an artificial state of sensitivity to selected waveFrom the Department of Dermatology, Hospital of the University of Pennsylvania' and Ivy Laboratories,b Accepted for publication April 3, 1991. Reprint requests: Kays Kaidbey, MD, The Department of Dermatology, Hospital of the University ofPennsylvania, 3600 Spruce St., 2 Maloney Bldg., Philadelphia, PA 19104-4283.

16/1/29920 .

262

lengths in the UVA region and often result in painful phototoxic reactions or blistering. Furthermore, exaggerated protection factors can be obtained when the absorption spectrum of the sunscreen coincides with the action spectrum of the sensitizer. 9Jt can therefore be argued that factors derived by such assays, although perhaps applicable to patients receiving PUVA, are irrelevant to normal skin. Other end points that have been examined include UVA-induced pigmentation and immediate pigment darkening (IPD).9 The action spectrum for melanogenesis is similar to that for erythema, II whereas that for IPD encompasses both the UVA and visible regions, with a peak in the UVA,12 In a previous article, it was found that protection factors derived by the use of either pigmentation, IPD, or erythema as end points were essentially similar. 9 Because IPD has the lowest threshold dose and leaves no sequelae, we investigated the feasibility of using this response to determine protection factors against UVA (IPD-PF). MATERIAL AND METHODS UV source. The UV source was a 150 W xenon arc solar simulator. A 1 mm thick Schott WG345 filter and a 1 rom UG 11 filter were added to provide a continuous spectrum from 320 to 410 nm with a broad peak between 350 and 370 nm. 13 Irradiance at skin level was measured with a phototherapy UVA detector (modelIL 1350 radiometer, International Light Inc., Ne~buryport, Mass.) with peak sensitivity at 360 nm. The mean UVA irradi· ance at skin level was 48.8 mW jcm2 . Subjects. The subjects were healthy white students of both sexes between the ages of 18 and 48 years, with skin

Volume 2S Number 2, Part 1 August 1991

UVA protection factors 263

Table I. UVA protection factors (IPD-PFs) of selected sunscreens as determined by the IPD method Sunscreen

Active ingredients

LabeledSPF

IPD-PF

SD

No.

A

Ti02 OTS B-3 OMCX 507 B-3 OTS B"3 OMCX OTS B·3 OMCX OCTC OTS B-3 OMCX 507 1789 OTS B-3 OMCX Ti02 Melanin OTS B-3 1789 OCTC

4 30

1.6 1.6

0.1 0.3

10 20

15

2.0

0.5

20

30

2.0

0.3

20

45

2.4

0.9

20

29

2.6

0.7

20

15

3.0

0.7

20

15

4.5

1.0

10

15

4.7

1.5

20

B

C D E

F G

H

I

507, Octyl dimethyl p-aminobenzoic acid ester (Padimate 0); 1789, avobenzone (I-butyl methoxydibenzoyl methane); B-3, benzophenone·3 (oxybenzone); OCTC, octocrylene (2 ethylhexyl-2-cyano-3,3,-diphenylacrylate); OMCX, octyl p-methoxycinnamate; OTS, octyl salicylate; n02, titanium dioxide.

types III and IV. None was applying or receiving any medications. All exposures were given to the untanned midportion ofthe back, after the experimental procedures were explained and informed consent was obtained. Irradiation. All exposures to sunscreen-treated and unprotected skin were given in series of 25% dose increments. The threshold for IPD, defined as the smallest dose required to produce darkening of the skin with a clearly defined margin,9 was recorded immediately at the end of each exposure. Test sunscreens were applied at a topical dose of ~ mgjcm2 to rectangular areas measuring 15 X 5 em. They were spread as uniformly as possible with the finger and allowed to remain on the skin for 15 minutes before irradiation. The protection factor was the ratio of the IPD threshold dose in the protected skin to that in unprotected skin. An arithmetic mean IPD-PF was then calculated for each sunscreen from the individual factors. Sunscreens. Seven marketed and two investigational sunscreens were tested. These are listed in Table I. Reciprocity. To determine whether IPD obeys reciprocity, threshold doses were determined at three irradiances in groups of 15 subjects. In addition, the IPD-PFs

of two sunscreens were compared at 15 and at 50 mWI cm2• Wire-mesh screens were used to reduce the UVA intensity at skin level from 48.8 mWjcm2 to a mean of 14.9 and 5.0 mWjcm2• RESULTS The threshold IPD dose ranged from 0.7 joule/ cm2 to 3.6 joules/cm2 (mean 1.2 joules/cm 2 ). A normal distribution was closely approximated when the decimal logarithm of the threshold dose was plotted. Table I shows the mean IPD-PFs of the tested sunscreens, which ranged from 1.6 to 4.7. There was no apparent relation between the SPF cited on the label and the IPD-PF. Thus products with SPFs ranging from 4 to 30 had similar IPD-PFs, and the same was true of two other products with SPFs 15 and 30. Table II lists threshold IPD doses obtained at three irradiances. ANOVA revealed no significant difference among the three sets of values.

Journal of the American Academy of Dermatology

264 Kaidbey and Barnes

Table II. Threshold IPD dose at three irradiances (1) I-50 mWjcm2

Subject No.

1

2

3

4

5 6 7 8 9 10 11 12 13

14 15 Mean (SD)

1.3 1.1

1.6 1.7 1.4

3.0 0.9 2.0 1.3

1.7

1.7 2.6 0.7

1.1

0.8 1.2 0.7 1.0 0.7 1.0 1.7 1.4 (0.7)

1.0 0.9 0.8 0.7 0.9 0.9 1.4 1.7 1.2 (004)

1.3

Table III. IPD-PFs of two sunscreens determined at two intensities Subject No.*

1-15 mWjcm2

1- 50 rnWjcm 2

1.3 1.6 1.0 1.6 1.3 1.3 3.1 2.5 2.5 2.0 3.1 2.5

2.7 1.2

1 2 3

1.6 2.0 1.0

1.9 3.6 0.8 1.2 1.0 1.2 0.6 1.0 0.9 1.2 1.7

5 6 7

1.6 1.6 2.5 2.5 2.5 2.5 2.5 3.1

2.3 1.2

4

8

9

10

11 12

1.3

*Subjects I through 6 received sunscreen B; subjects 7 through ceived sunscreen G (see Table I).

12 re-

1.5 (0.8)

*Threshold doses at 1= 5 roWjcm2 were obtained in a separate group of 15 subjects.

The IPD-PFs oftwo sunscreens (B and G) determined at two irradiances (15 and 50 mW I cm2) are shown in Table III. A paired t test showed no significant difference between the values at the two intensities (p = 0.2694). DISCUSSION

IPD, which was well described by Meriowsky, is a reversible oxygen-dependent reaction that appears immediately after exposure of human skin to UVA and to much larger doses of visible wavelengths. 14 Although the mechanism is not known, recent studies suggest that it probably represents a photochemical reaction that involves the oxidation of a low-molecular-weight form of melanin in melanosomes. IS, 16 The response is therefore best visualized in persons with darker complexion. The action spectrum, which extends from about 300 to 470 nm, shows a broad peak at 340 to 370 nm. 12, 17 Quantitative measurements have revealed that IPD shows first-order dose-response kinetics with saturation at high doses.1 7 Like the minimal erythema dose (MED), the minimal IPD dose shows a log normal distribution. As would be expected, the IPD threshold dose shows considerable interindividual and intraindividual variability and, like the MED, is strongly influenced by anatomic location. Within a defined area such as the untanned midportion of the back, how-

ever, IPD is a reproducible end point. 9 Because the action spectrum extends into the visible region, visible wavelengths should be removed from the light beam by appropriate filtration when IPD-PFs are determined. In the case of the solar simulator described herein, this results in a spectrum that closely parallels the action spectrum for IPD. Skin reactions must be evaluated immediately after exposure because the response at threshold levels disappears within 60 seconds. Fair skinned persons usually exhibit a poor or no IPD reaction and should not be used. 14, 18 The present findings support the common clinical impressionthat many currentlymarketed sunscreens provide little photoprotection against UVA as evidenced by their limited usefulness in patients with photodermatoses. One product that contains titanium dioxide was also disappointing. The low protection factors determined by the IPD procedure contrast with the significantly higher factors obtained when psoralens are used as photosensitizers. 9 Thus a sunscreen that contains the new UVA absorber butyl methoxydibenzoyl methane (sunscreen G) provided a protection factor of 4.8 with the psoralens method 19 compared to 3.0 with IPD. However, this sunscreen proved to be superior to the other tested benzophenone-containing products. To avoid the use of artificial photosensitizers, Stanfield et a1. 20 explored the feasibility of using delayed UVA erythema in normal skin to measure protection factors. Because the doses and exposure times required to elicit UVA erythema are exces-

Volume 25 Numbl:r 2, Part 1 August 1991

sively long, they modified their solar simulator to obtain high irradiances and compared the UVA MED and the protection factor of a sunscreen at 50 mW/cm 2 andat 100mW/cm2 • They found nosignificant differences at these two intensities. No comparisons were made, however, with lower irradiances closer to solar UVA levels. There is evidence that UVA erythema and delayed or true pigmentation do not obey dose-reciprocity. Kagetsu et a1. 2 ! who used a filtered xenon arc lamp, found that reciprocity did not hold when MEDs or minimal tanning doses were compared at 50 mW/ cm2 and at 5 mW/ cm2• The latter irradiance is similar to midday summer solar UVA at this latitude and hence is .the more relevant intensity. We have also observed a significant decrease in the UVA MED as progressively higher irradiances were used (unpublished observations). Furthermore, there still exists uncertainty about the thermal effects of high-intensity beams on the skin. Irradiances greater than 120 mW/ cm 2, for example, can evoke stinging, burning, and pain. Lack of reciprocity would invalidate UVA erythema and delayed pigmentation as end points. In this study, it was necessary to determine whether IPD obeyed dose-reciprocity. The minimal IPD doses were similar at 50 mW/ cm2 and at 5 mWI cm2, as were the IPD-PFs oftwo sunscreens at 50 mW/ cm2 and at 15 mW/ cm2• It is therefore concluded that the IPD reaction is probably not irradiance dependent at intensities less than 50 mW/ cm2. Recently, Sayre and Agin22 suggested that the photoprotective efficacy of a sunscreen against UVA can be predicted mathematically from the erythema action spectrum, the emission spectrum of the UV source, and the absorption spectrum of the test product. The predicted UVA factors of several marketed products similar to those examined in this study were higher than those obtained with the IPD method and ranged from 3.3 to 6.2. These differences are due, at least in part, to the fact that the action spectrum for erythema is different than that for IPD. They also found no differences in UVA protection factors of five products, the SPFs of which ranged from 15 to 39. This again indicates that there is no apparent correlation between the SPF and the UVA protection factor, at least when the xenon solar simulator is used as the source. This is contrary to the common belief that sunscreens with increas:.. ing SPF must block more UVA. Although the mathematical approach appears to be useful for

UVA protection factors 265

screening purposes, a human efficacy test is still required for the validation of predicted values. The two investigational sunscreens provided LPDPFs of 4 to 5 and hence appear to be significantly more effective than the tested marketed products. Because the. maximal UVA dose that can be received during an entire day is approximately 120 joules/ cm2, a protection factor of4orgreater should provide adequate protection for normal p~rsons. This, however, may not be the case for patients with a photosensitivity disorder because the provocative dose of UVA can be low. Further studies are required to assess the therapeutic usefulness of the new UVA photoprotectants. We suggest that IPD provides a rapid and reproducible end point for screening potentially useful agents or formulated products for photoprotection against UVA. This method avoids the unpleasant and painful phototoxic reactions associated with the use of psoralens and provides more realistic UVA protection factors.

REFERENCES 1. Kligman LH, Akin FJ, Kligman AM. The contributions of

UVA and UVB to connective tissue damage in hairless mice. J Invest Dermatol 1985;84:272-6. 2. Cole CA, Forbes D, Davies RE. An action spectrum for UV photocarcinogenesis. Photochem Photobiol 1986;43:27584. 3. Sterenborg HM, Van der Leun JC. Action spectra for tumorigenesis by ultraviolet radiation. In: Passchier WF, Bosnjakovic BFM, eds. Human exposure to ultraviolet radiation, risks and regulations. Amsterdam: Elsevier, 1987: 173-90. 4. Bissett DL, Harmon DP, Orr TV. Wavelengths dependence of histological, physical and visible changes in chronically UV-irradiated hairless mouse skin. Photochem Photobiol 1989;50:763-9. 5. Wu1fHD, Poulsen T, Davies RE, Urbach F. Narrow-band .UV radiation and induction of dermal elastosis and skin cancer. Photodermatology 1989;6:44-51. 6. Kligman LH, Kaidbey KH, Hitchins VM, et al. Long wavelength (>340 nm) ultravio1et-A induced skin damage in hairless mice is dose dependent. In: Passchier WF, Bosnjakovic BFM, eds. Human exposure to ultraviolet radiation, risks and regulations. Amsterdam: Elsevier, 1987:7781. 7. Van WeeldenH,deGruijl FR, Van der LeunJC. Carcinogenesis by UVA, with an attempt to assess the carcinogenic risks of tanning with UVA and UVB. In: Urbach F, Gange RW, eds. The biological effects of UVA radiation. New York: Praeger 1986:137-46. 8. Jarratt M, Hill M, Smiles K. Topical protection against long-wave ultraviolet A. J AM ACAD DERM(\TOL 1983;9: 354-60. 9. Kaidbey K, Gange RW. Comparison of methods of assessing photoprotection against ultraviolet A in vivo. J AM ACAD DERMATOL 1987;16:346-53. 10. Kaidbey KH, Kligman AM. Laboratory methods for

lournal of the American Academy of Dermatology

Kaidbey and Barnes

11.

12. 13. 14. 15. 16.

17.

appraising the efficacy of sunscreens. J Soc Cosmet Chern 1978;29:525-36. Gange RW, Park YK, Auletta M, et al. Action spectra for cutaneous responses to ultraviolet radiation. In: Urback F, Gange RW, eds. The biological effects of UVA radiation. New York: Praeger, 1986:57-65. Henschke U, Schultze R. Uber pigmentierung durch langwelliges ultraviolett. Strahlentherapie 1939;64: 14-42. Berger DS. Specification and design of solar ultraviolet simulators. 1 Invest Dermatol 1969;53:192-9. Beitner H. Immediate pigment-darkening reaction. Photodermatology 1988;5:96-100. Honigsmann H, Schuler G, Aberer W, et al. Immediate pigment darkening phenomenon: a re-evaluation of its mechanisms. 1 Invest DermatoI1986;87:648-52. Beitner H, Wennersten G. A qualitative and quantitative transmission electronmicroscopic study of the immediate pigment darkening reaction. Photodermatology 1985;2: 273-8. Rosen CF, Jaques SL, Stuart ME, et al. Immediate

18.

19.

20. 21. 22.

pigment darkening: visual and reflectance spectrophotometric analysis of action spectrum. Photochem Photobiol 1990;51:583-8. Poh Agin P, Desrochers DL, Sayre RM. The relationship of immediate pigment darkening to minimal erythemal dose, skin type, and eye color. Photodermatology 1985; 2:288-94. Gange RW, Soparkar A, Matzinger E, et al. Efficacy of a sunscreen containing butyl methoxydibenzoylmethane against ultraviolet A radiation in photosensitized subjects. 1 AM ACAD DERMATOL 1986;15:494-9. Stanfield lW, Feldt PA, Csortan ES, et a!. Ultraviolet A sunscreen evaluations in normal subjects. 1 AM ACAD DERMATOL 1989;20:744-8. Kagetsu N, Gange RW, Parrish lA. UVA-induced erythema, pigmentation and skin surface temperature changes are irradiance dependent. 1 Invest Dermatol 1985;85:445-7. Sayre RM, Agin PP. A method for the determination of UVA protection for normal skin. J AM ACAD DERMATOL 1990;23:429-40.

Trachyonychia associated with alopecia areata: A clinical and pathologic study Antonella Tosti, MD, Pier Alessandro Fanti, Federico Bardazzi, MD Bologna, Italy

MD, Rossella Morelli, MD, and

Forty of 1095 patients (3.65%) with alopecia areata had severe nail changes that fulfilled the clinical criteria for the diagnosis of trachyonychia. Twelve of these patients had a nail biopsy. A mild to moderately dense lymphocytic infiltrate associated with exocytosis and spongiosis was detected in the proximal nailfold, nail matrix, nail bed, and hyponychium of 11 patients. One patient showed the pathologic changes of lichen planus; lichen planus of the skin developed 6 months after the nail biopsy. Immunohistochemical characterization on paraffin-embedded sections showed that the inflammatory infiltrate consisted of peripheral T lymphocytes. Immunophenotyping on frozen sections was performed in four cases. The results revealed a T4 ITS ratio of 2: 1and the presence of Langerhans cells in the nail matrix. Ourresults show that trachyonychia is an uncommon nail manifestation of alopecia arcata. Distinctive pathologic features of a mild to moderately dense lymphocytic infiltrate associated with exocytosis and spongiosis characterize trachyonychia as well as the other nail abnormalities caused by alopecia areata. The clinical association of trachyonychia with alopecia arcata does not exclude that the nail abnormality can be due to other diseases such as lichen planus. (J AM ACAD DERMATOL 1991;25:266-70.) The term trachyonychia (from the Greek word trakos meaning rough) originally coined by Alkiewicz1 in 1950 well describes the broad specFrom the Department of Dermatology, University of Bologna. Accepted for publication March 18. 1991.

Reprint requests: Antonella Tosti, Clinica Dermato1ogica, Universita di Bologna. via G. Massarenti. 1,40138 Bologna, Italy. 16/1/29533

266

trum of nail disorders that can be observed in alopecia areata. Although the clinical features of trachyonychia associated with alopecia areata have been extensively and accurately described,2-8 the prevalence and the pathologic features of this condition are still unknown. Spongiotic changes have been detected in the few patients who have undergone nail biopsy,3,9, 10 but pathologic studies on a large series of patients are lacking. The aim of this study was to