British Journal of Dermatology 1997: 136: 699-705.

Photosensitivity in lupus erythematosus, UV photoprovocationresults compared with history of photosensitivityand clinical findings

T.HASAN. F.NYBERC* E.STEPHANSSON.* P.PUSKA.t M.HAKKINEN. S.SARNA4A-M.ROS* AND A.RANKIDepartment of Dermatology. University of Tampere tiiid Tampere Umversity Hospital. PLIOOI). FIN-i3>21 Tampen'. I'iiilaud'Department of Dermatology. Karoliuskii Institute and Hospital. Stockholm. Sweden'\ Department of Dermatology, Helsinki University Central Hospital Helsinki. Finland^ Department of Public Health. University of Helsinki. Helsinki, Fhilaiid

Accepted for publication 28 November 1996

Summary Photosensitivity. one of the presenting symptoms in lupus erythematosus (LEI. is still poorly definedand varying prevalence figures have been reported. The possibility of a coexisting photodermatosis.especially polymorphous light eruption (PLE), has often not been taken into account. We report theresults of ultraviolet A (UVA) and B (IIVB) photoprovocation tests in 67 clinically photosensitivepatients who had confirmed discoid LE (ULE). systemic LE (SLE) or subacute cutaneous LE (SCLE).The results are compared with a detailed history of photosensitivity and with clinical and serologicalfindings.

A pathological photoprovocation reaction, graded as weak, moderate or strong, was induced witheither UVA or UVB in 69% of patients with LE. in 100% of those with SCLE, in 70% of those with SLEand in 64% of those with DLE. but in none of 14 controls. Only 16% of the pathological reactionswere strong and long-lasting, resembling LE lesions, while 48% were moderate or weak andtransient, clinically like PLE. Fifty-three per cent of the provocation reactions which were biopsiedshowed a PLE-like histology or a non-specific inflammatory reaction, and most of them wereclinically moderate or weak reactions of short duration. In the remaining, mostly clinically strongor long-lasting reactions, the histology was consistent with LE. A history of sunlight sensitivity didnot predict a pathological photoprovocation result but a positive association between the presence ofSSA/Ro or SSB/La antibodies and a pathological photoprovocation reaction was found. We haveshown that PLE coexists with LE and that both PLE- and LE-like lesions can be induced with UVradiation in LE patients.

It is generally well established that sunlight canaggravate skin symptoms in patients with all subsetsof lupus erythematosus (LE). Photosensitivity occursin l l % - 9 4 % of patients with systemic LE (SLE).'"^ in52%-100% of those with subacute cutaneous LE(SCLE)^"^ and in 69%-90% of discoid LE (DLE).̂ '*̂Such wide ranges may be due to the use of differentdiagnostic criteria for LE in older studies and due toracial differences.^ However, the lack of a uniformdefinition is the most likely explanation. Furthermore,the possibility of a coexisting photodermatosis such aspolymorphous light eruption (PLE), is generally notconsidered. In our recent questionnaire study'" weshowed that as many as 49'Xt of Scandinavian patientswith LE also experience symptoms of PLE. This is more

than twice the overall prevalence of PLE, at 21%, in aScandinavian population.^'

Photoprovocation is an objective way to evaluatepossible photosensitivity. The capacity of both ultravio-let A (UVA) and B (UVB) radiation to reproduce LE andPLE skin symptoms has been shown in several studies,the frequency of pathological photoprovocation resultsvarying between 20% and 100%.^^''^ The criteria for apathological reaction have also varied and the possibi-lity of inducing a PLE reaction in LE patients has notbeen considered.

We wished to evaluate UVA and UVB photoprovoca-tion responses in correlation with a detailed history ofphotosensitivity. The type of symptoms previouslyexperienced were recorded in order to discriminate

© 1997 British Association of Dermatologists 699

"00 T.HASAN etitl

between PLE and UV aggravation of LE. The kinetics ofthe provoked reactions were followed up and the resultscorrelated with clinical, histological and serologicalfindings.

Patients and methods

Patients

A photoprovocation test was performed in SO patientswith DIM, \0 with SLK and seven with SCLE patients (1 (Imen and 57 women). All were clinically photosensitive.They were followed-up at three dermatological clinics inScandijitivia. namely at the Karoiinska Hospital. Stock-holm. Sweden, the Helsinki Universitj' Central Hospital,Helsinki. Finland, and the Tampere University Hospital.Tampere. Finland. Fourteen volunteer non-photosensitivecontrols, three men and 11 women were also studied. Allwere Caucasian, belonging to nationalities closely asso-ciated genetically.''^ The study was approved hy the ethicscommittee of each hospital, and all patients gave writtenconsent before enrolment. The mean age of the patientswas 47 years (range 14-83) and the mean duration ofdisease was 12 years (range 1-37). The mean age of thecontrols was 58 years (range lfi-6()).

The diagnosis of DLE was based on clinical and histo-logical findings: the diagnosis of SCLF, on the originaldescription'' and skin histology. No patient with SCLH haddiscoid lesions. The diagnosis of SLE was based on the1982 revised criteria of the American College of Rheu-matology."" a combination of clinical, haematologicaiand serological features. All patients with SLE hadshown antinuclear antibodies (ANAs), and LE-specilicskin lesions;"' seven patients presented with acute skinlesions, e.g. a butterlly rash or widespread maculopapularrash, one had chronic discoid lesions, and two had bothacute and chronic skin lesions. The diagnostic criteria ofthe I,E subtypes were similar in the three clinics and basedon long-standing clinical collaboration.

Differentiation between UV aggravation or induction ofLE skin lesions and PLE-type photodermalosis'" wasemphasized with personal interviews. PLE was definedas a papular and/or vesicular pruritic eruption arising afew days after sun exposure on sun-exposed skin, some-times on a dilTerent skin area than LE lesions, and healingwithin a week. In the present study the term history ofphotosensitivity' refers either to a history of PLE or toaggravation or induction of LE skin lesions or both.

Most patients were on no systemic medication fortheir skin disease at the time of photoprovocation. Twopatients were taking hydroxychloroquine, 500 mg and

-)00 mg per day. respectively. Two patients received oralcorticosteroid. corresponding to 5 mg prednisolonedaily, one also took cyclosporin. 125 mg daily.

PhotoU'stiiHj cqidimu'ut diid photopivvocatUvi protocols

LIVA phototesting was performed with UVASUN 3{)()()with an emission spectrum of 540-400nm. ' ' IJVBphototesting was performed with four different lightsources: (i| Waldman 85-lOOW/UV (S (main emissionspectrum of 290-570 nm: UVA. UVB and UVC irradi-ance of l-5mW/cm". 0 5mW/cm" and O()mW/cm".respectively), (ii) Philips TL 2()W/12 (main emissionspectrum of 280-570nm: UVA. UVB and UVC irradi-ance of 0-7mW/cm". 1-3 mW/cm~ and 0-01 mW/cm .respectively), (iii) Osram Xenon XBO 1 50 W Schott WG29 5 filter (emission spectrum of 2 50-400 nm: UVA,UVB and VVC irradiaiice of lfvOmW/cm". 2-()mW/cm^and 0 2 mW/cm". respectively), and (iv) Osram HQ-TS 400 W/D (emission spectrum of 2 50-400 nm: UVA.UVB and IWC irradiance of 5 2 mVV/cm", ()• 5 mW/cm"and 0-2mW/cm". respectively).

The patients were tested in the autumn, winter orearly spring on intact skin of the upper back, upper armor forearm. Most patients were tested with both UVAand UVB. but 16 were only tested with UVB. First, theminimal erythema dose (MED) for both UVA and UVBwas determined. The MED. defined as barely perceptibleerythema with at least three visible sharp corners, wasassessed after 24 h. In most patients, reactions withUVA up to 80 I/cm" were normal and. thus, the exactUVA MED could not be stated. Photoprovocation wasthen performed, cither on 1 - i consecutive days (mostlythree times: 40 patients) or every 2 -5 days (mostly fiveto six times: 27 patients). The patients in these twogroups did not differ statistically with regard to age orsex, duration of LE skin symptoms, history or type ofphotosensitivity, and distribution or activity of LE skinlesions. UVA doses varied between 20 and lOOJ/cm^(maximum 2-7 MED) and UVB doses between 1 and 3MEIJ. so that moderate redness was achieved. Each UVdose was adjusted Individually according to the reactionof the previous provocation. The mean total UVA dosewas 247 J/cm~ and the mean total UVB dose 5-7 MED inthe patient group provoked on consecutive days. Therespective values were 182 J/cm" and 7-9 MED in thepatient group provoked every 2-5 days. The test areavaried from 5x8 cm to 2 x 2 cm. Provocation reactionswere evaluated about 24 h after each irradiation, andthereafter every 4 - 7 days up to 5 weeks. Fourteencontrols were provoked according to either protocol.

© 1497Britisli Associatiiin nl' isls. HiHish loitrmi! otDcniuiiolofiii. t Jd.

PHOTOSENSITIVITY IN LUPUS ERYTHEMATOSUS 7(31

Biopsy specimens were obtained from the pathologicalprovocation reaction sites for routine histology. Sectionswere stained with haematoxylin and eosin and most alsowith periodif-acid-SchiiT (Py\SI reagent. Biopsy specimenswere examined by experienced pathologisls. The histolo-gical diagnosis of LE and PLE were based on generallyaccepted criteria.^ *'~"̂ The main histological criteria forLE were hydropic degeneration of the epidermal basalcells, and a perivascular and periappendageal lymphoidcell intiltrate. The features of PLH were a perivascuiarlymphoid cell inliitrate. with parakeratosis, focal spon-giosis and slight vaeuolization of basal keratinocytes.

Table 1. I'liotnprDvocation resultservtheniiitosiis lLE| and in controls

in piitieiils with [iipiis

Diagnosis

DLESLESCLETotalControls

H

5010

7ft714

18*4*h

28*0

UVA

(55%)

Paihological reactions

UVB

31ft7

44(66%)0

UVA or UVB

U 164%)7|7U%)7(100%)

4ft (69%)0

" Five SIJE and 11 DLE patients were not pholoprovoked with U\'ADLE, Discoid LE; SLE, svsteinic LE: SCLE, subacute cutaneous 1.1'].

Sewiogkal evahmtkm

ANA assay WHS performed with routine indirect immu-nofluorcscence using rat fixed liver tissue or Hep-2 cellsas the antigens. In sera from 51 patients, antibodies toSSA/Ro and SSB/La extraclable nuclear antigens (ribo-nucleoproteins) were assayed by a standard immuno-dilfusion technique with the mixture of soluble proteinsof rabbit thymus and pig spleen as antigen. In 16 cases.a commercial enzyme immnnoassay lEIA) methodIENA4 ELISA. Quanta Lite. Inova Diagnostics, SanDiego. CA. U.S.A.) was used for screening Sm. RNP.SSA/Ro and SSB/La IgG antibodies. In case of a positivescreening result |2() units), the specificity of the anti-bodies was determined with a specilic EIA assay (SSAHLISA, SSB ELISA: Quanta Lite. Inova Diagnostics).

were considered pathological either with UVA. UVB orboth. All seven patients with SCLE, seven of 10 (70%)with SLE and 32 of SO (fS4'X,) with DLE gave patholo-gical provocation reactions. UVB induced a pathologicalprovocation reaction slightly more often than UVA(Table 1).

Seventy per cent of ail pathological reactions wereweak or moderate, the rest being strong. Eighty-four percent of pathological reactions appeared within 1 week ofthe last provocation and the remaining 16% more than Sdays after the last provocation. In one patient with SIJi.the reaction first appeared 3 weeks after the provocation.Eifty-nine reactions were followed up for at least 21 daysor until they subsided. Most reactions (6 3%) disappearedwithin 14 days after the last provocation (transient

Stiilistiai} auaUjsis

for testing associations with clinical and serologicaldata, Fisher's exact and Maiilel-Haenszel two-sidedtests were used, respectively. The statistical analysiswas performed with StatXact (Statistical Software forExact Nonparametric Inference),^"' Statistical signifi-cance was considered as y'<0-05. Confidence intervals(CI) were determined with the Confidence IntervalAnalysis program."''

Results

Results of UVA and ilVH pliolopnworatiou and kinetics ofthe mictions produced in the three subijwups of lupuscriithenuilosus piitienls

The photoprovocation results of 67 patients and 14controls are presented in Table 1. Altogether, in 46patients {b^'Yu. 95% CI S6%-79%) 73 provocation sites

40

35

co

cti

ra<u"oa>qui

3Z

30

25

20

15

10

5

0

n long-lasting reactions

n transient reactions

strong moderate/weak

Clinical appearance of photoprovocated reaction

Figure 1, Reitctioii kinetics ol 59 pliotiiprnviikwi skin reatlions. Siroiigand long-ltistiiig reactions had a statistifally sii^niticant association(P<0-001).

C' 1497 British Association of Dermalologi.sls. Brilisli Jinininl ofDcniuiloloqij. 1 ?6,

702 T.HASAN etal.

Figure 2. (al A normal UVA and a pathological UVB provocationreaction 10 days after the last provocation in a 34-year old man withdiscoid lupus erythematosus. The large rectangular ligures are provo-cation areas and the small squares are graded exposures of UV fordetermining the minimal erythema dose, (bl The same UVB provocationreaction as in (a| shown in close-up S months later. The reactionresembles a spontaneous lesion of discoid lupus erythematosus.

reactions), while 3 7% were visible for more than 21 days(long-lasting reactions). There was a strong statisticallysignificant association between strong and long-lastingreactions (Fig. 1). Forty-eight per cent of the pathologicalreactions were both moderate/weak and transient, clini-cally resembling PIJi more than LE. Sixteen per cent ofreactions were both strong and long-lasting, clinicallylike LE lesions. A classic lesion of D\£. was provoked inonly 14 of 50 (28%) pathological reactions in patientswith DLE (Fig. 2). Controls developed pigmentation.erythema or slight oedema, the latter two subsidingwithin 3 days.

The irradiation protocol used did not particularlyaffect the final provocation results as 73% of patientsprovoked with UVB every day and 63% of patientsprovoked with UVB every other to every third daygave a pathological reaction. Comparison of the resultsbetween protocols using UVB devices with eithernarrow or wide UV spectrum did not differ statistically,neither did the strength and persistence of the reactionsbetween the three clinics.

Histology uf the pathological photoprovocation results

Thirty-six of 73 reactions (49%) were evaluated histo-logically. 10 of them in serial biopsies so that 51 biopsieswere examined altogether. Fifty-three per cent of thereactions biopsied showed histological features of PLE ora dermal perivascular lymphocytic infiltrate (Table 2).In the remaining 47% of the reactions, the histologicaldiagnosis was consistent with LE. LE histology cor-related significantly with the strong as well as long-lasting clinical reaction types. In some cases, thehistological picture changed from PLE to LE in serialbiopsies during follow-up (Table 2). The detailed timepoints of the biopsies are shown in Fig. 3. UVA- andUVB-provoked skin lesions showed the histology of LE in38% and 52%. respectively (statistically not significant).Most of the pathological reactions not examined histo-logically were weak or moderate and transient, manylasting only a few days.

Association of pbotoprovocation results with demographic

data, clinical findings and autoantibodies

In 84% (95% Cl 73%-92%) of patients. LE skin symp-toms were exacerbated on exposure to sunlight. Sixty-seven per cent (95% CI 55%-78%) of the patientsreported a history of typical PLE. Most patients (68%)with moderate or weak and transient provocation reac-tions had had a PLE-like rash.

1997 British Association of Dermatologists. British journal nj Deruuilologii. 1 ib. fi99-7()5

PHOTOSENSITIVITY IN LUPUS ERYTHEMATOSUS 703

Table 2. The histology of 36 pathological provocation reactions compared with the appearance and duration of the reaction

LE-like histology * PLE-like or non-specific histology Statistical significance t

Clinical appearance of the reactionstrong Jweak/moderate §

The duration of the reactionmore than 21 days (long-lasting)lesK than 14 days (transientinot followed-up

All pathological reactions examined

8 (80%)9(35%)

12(75%)2{137n)3

17(47%!

2 (20%)17(65%)

4(25%)13 (87%)2

19 (53%)

P = 0-001

' In six reactions the histologicai picture changed from PLE-Iike/non-specific to LE-like in the repeated biopsies. These are included in the group 'LE-like histology', t Fisher's exact, two-tailed. $ strong = either erythema, papules/plaques and DLE-like scaling or erythema with papules and markedoedema, § weak = just erythema lasting for at least t week, moderate —erythema with papules lasting longer than 3 days.LE, lupus erythematosus: PLE, polymorphous light eruption: DLK. discoid lupus erythematosus.

The age of the patient, the duration of LE. and theoverall activity of skin lesions at the time of testing didnot correlate with the photoprovocation results. UVprovocation was pathological in 76% (95% CI 60%-89'X)) of patients with a disseminated distrihution of skinlesions compared with 5y% (95% CI 4O%-77%) ofpatients with lesions localized to the face or scalp(statistically not significant).

Patients with SSA/Ro and/or SSB/La antibodies gavea pathological provocation reaction significantlymore often (83%. 95% CI 63%-95%) than thoselacking these antibodies (60%. 95% CI 44%-75%)

20

18

16

a>0}

a.0n"o1 .

E2

14

12

10

8

6

4

2

0

DLE-likehistology

D PLE-like ornon-specifichistology

< 4 days 4-7 days 8-14 days > 21 days

Time-point of the biopsy after the lastprovocation

Figure 3. Histological diagnosis in relation to the timing of the biopsy.Fifty-one biopsies from 36 reactions were evaluated, 1 5 of these wereadditional serial samples from 10 pathological reactions.

( P = 0-048. common odds ratio (OR) 3-50). The ser-ological method was not found as a confounding factor( O R E U S A = 3 - 0 0 . ORi,,,munudinuMon= 3-85). The fewpatients belonging to different LE subgroups did notallow us to study further associations, ANA results didnot correspond to photoprovocation results.

Discussion

Photoprovocation with both UVA and UVB radiationwas performed in 67 patients with either discoid, sub-acute or systemic LE to confirm the type of photosensi-tivity and to evaluate reaction patterns and kinetics. In69% of the patients provocation yielded a pathologicalreaction. Only about 16'Ki of reactions were strong andlong-lasting, resembling LE lesions. The clinical patternand kinetics in 48% of the reactions were of doubtful LElesions that more resembled PLE.

Our observation that ail patients with SCLE. 70% ofthose with SLE and 64'X. of those with DLE give apathological response to photoprovocation. UVB beingslightly more effective than UVA. is in agreement withrecent provocation studies.'' '" UVB-provoked skinlesions slightly more often gave histology typical of LEthan did lesions provoked by UVA.

Early LE may be difficult to distinguish from PLE. asthere are cases with strikingly similar clinical, histologi-cal, photobiological and serological lindings.'•''*"•*"''""'We recently conducted a questionnaire-based study, inwhich we showed that 49% of LE patients also have ahistory suggestive of PLE.'" This is more than twice thereported prevalence of PLE in a corresponding Scandina-vian population.'' Coexistence of PLE and LE seemscommon, although PLE has not been identified as apredisposing factor for LE in previous studies. * ' ' " This

1997 British Association of Dermatologists, British journal ofDermatoIoiiii. 1 36, 699-705

704 T.HASAN I'tal.

coexistence may have led to diiliculties in the interpreta-tion of photoprovocation reactions and may be onereason for diiTering results.'""'"

As about half of the pro\'ocation reactions in thepresent study were clinically non-specitic. histologicalexamination was performed. In only 47% of the reac-tions biopsied was the histology consistent with LE(Table 2). fn some long-lasting reatUons. only repeatedbiopsies revealed the true nature of the skin lesion.because the bistological picture changed from non-specitic or PI.K-like lo that resembling LE. Most reactionswhich showed a histological picture of PLE disappearedin few days, excluding the possibility of later developingthe changes of LE. In these cases the histology of thesingle biopsy was considered reliable.

The pathomechanism of LlV-induced LE lesions is notknown, UVB radiation has been shown to induceexpression of the nuclear autoantigen Ro/SSA at thecell surface iu vitro.^"^ but contradictory results havebeen reported about possible associations betweenphotosensitivity and the occurrence of circulatingSSA/Ro antibodies in LE patients."' '" '"^""^ Our obser-vation about a statistically significant associationbetween occurrence of SSA/Ro and/or SSB/La antibo-dies and a positive photoprovocation test resultstrengthens the role of these autoantigens in the patho-mechanism of UV-induced LE skin lesions.

Both PLE- and LE-!ike lesions can be induced with UVradiation in LE patients. This is important to considerwhen further investigations into the pathomechanismof LE are planned.

Acknowledgments

We thank Ms Kirsti Leszczynski. Ph.L and Mr ReijoVisuri. M.Sc. Finnish Centre for Radiation and NuclearSafety. Helsinki, for spectral measurement results andMs Laura Huurto. Ph.L., Turku University CentralHospital. Turku, for dosimetric calculations of photo-testing equipment In Helsinki and Tampere UniversityHospitals. Finland. We also thank Ms Liisa Kannas.Helsinki University Central Hospital lor skilful practicalassistance. This study was supported by grants from theMedical Research Fund of Tampere University Hospital,Tampere. Finland, the Finnish Rheumatism Associa-tion. Finland, the Edvartl Welaiider Foundation, Swedenand the Finsen Foundation, Sweden.

References1 'JulTanclii DL. Dubois IIL. Cutaneous nianifestations of systemic

lupus erylhematosus. Arch Di'mmlol 19fi4; 90: i77-Hfi,

2 Harvey AM, Shulmini LK. Tumulty AP t'( ill. Systemic lupuserythemiitosuK. Keview of ihe literature iinci clinical analysis of15S cases. Medinm- t954; 3J: 291-457.

3 Wysenbeek AJ, Block DA. Eries IJi. Pre\'alence and expression ofphotoscnsitivity in systemic lupus erythematosus. Ann Rheum Dis1989: 48: 4(nl-5.

4 t^istlner M. Wallace Df, Nessim S <•( ill. l.upus t-rythematosus In the1981)s: ii survey of 570 patients. Scuiin Arihrit'is Rhi'imi 1991: 21:55-64.

5 HcutiitT EH. Blaszczyk .VI. jablonskii S ct ul. Studies on criteria of theKuropcHii Academy of Dermatology and Venereoiogy for the ciassi-lication of cutaneous lupus erythematosus. 1. Selection of clinicalgroups and study factors, liit / Deriitalol 1991: 30: 411-17.

(i Sontheimer RD, Thomas JR, Gitliam |N. Subacute t'utiiiieous lupuserythemalosus. A cutaneous marker for 9 distinct lupus erythe-matosus subsc-!. Anh Denualol 1979: 1 H : 14(19-15.

7 fallen |l ' Kulick MD. Stel/er G. Fowler ]F. Suhacute cutaneouslupus erylhematosus. Clinical, senilogic. anci immunogeneticstudies of forty-nine patients seen in a nonreferral setling. / AmAcmi Dermilal 1986: t 5 : 1227-37.

8 Calien |1'. Discoid lupus erythematosus: variants and clinicalassociations. CUu DcniuUol 198S; 3: 49-57.

9 Ward MM, Studcnski S. Clinical manifestations of systemic lupuserythemalosus. Identilicatioii of racial and socioeconomicinlluences./\n7r/i(IfniAM199(): ISO: S49-5J.

10 Nylicrg F. Hasan T. Puska P I'l (il. Occurrence of polymorphouslight eruption in lupus erythematosus. Hr j Dmualo! 1997; 1 36:217-21 .

I I Ros AM. Wennersten G. Current aspects of polymorphous lighteruptions in Sweden. Photodenmi(olo(iii 198(S: 3: 298-302.

12 Epstein jH. Tuffanelli DL. Dubois EL. Lighl sensitivity and lupuserytheniiitiisus. Anh Dcnimlol iy(S5: 9t ; 4 8 i - 5 .

1 i Freeman RC. Knox [M, Owens DW. Cutaneous lesions of lupuserythematosus induced hy monochromatic light. Arch Dcriiwloli9(i9: ]0(l: (i77-82.

14 Cripps D|, Rankin ]. Action spectra of lupus erythemalosus and

experimental immunolluorescence. Arch Deniuitol I97J: 107:Sh5-7.

is Holzle K. Plewig G. Lehmann P. Photodermatoses—diagnosticprocedures and their interpretation. Phouxifrimilolog)! 1987: 4:H19-14.

16 Wolska H, Blaszczyk M, jablonska S. I'hototests in patients withvarious forms of lupus t'rythematosus. liil / Dernuitol 1989: 28:98- l ( ) i .

I 7 van Weelden H, V'elthuis P|, Baart de la Faille H. Light-inducedskin lesions In lupus erythematosus: photobiological studies. ArchDenmtol Res 1989: 28t: 470-4.

18 lehmann P. llolzle E. Kind P et ill. Kxperimentai reproduction ofskin lesions in lupus erythematosus by II\A and rVI3 radiation./ Am Aim! Ihrimiiol 1990: 22: 181-7.

19 Cavalli-Slorza LI.. Piazza A. Human genoniic diversity in Europe: asummary of recent research and prospects for the future. Eitr jHum Gaiei 199J: 1: 3-18.

20 Tan EM. Cohen AS, Fries IF.-(ah The 1982 revised criteria for theclassification of systemic lupus erythemalosus. Arthritis Rheum1982: 118:412-16.

21 ("lilliam |N. Sontheimer RD. Distinctive cuianeous subsets in thespectrum of lupus erythematosus. / Am Aciid DiTiiuilol 1981: 4:471-5.

22 .Mutzhas MF. Holzle K. Hofmann C. Piewig GA. A new apparatuswith high radiation energy between 320-460 tun: physicaldescription and dermatological applications. / Invest Denimta!1981: 76 :42-7 .

1997 British Association of Dermatologists. HriOsh jouniiil of Dfrnmlolofiii. 1 36, 699-705

PHOTOSENSITIVITY IN LUPUS ERYTHEMATOSUS 70S

2} Lever W F, Schaumburg-l^vcr G. Histopcitlmhijii of thf Skin, /ihcdn. I'hiiaddphid: l.B.I.Ippincoli. 19')1).

24 van Braag MCCl, Boom BW. Vernieer B). Diagnosis and trcatmeiilof polymorphous lighl eruption. Int I Dermato! 1994: 33: 233-9.

2 S Slatistical Software for Exact Nonparametric Inference. Cytel Soft-ware Corporation 1992. Cambridge, MA. U.S.A.

2f) Gardner \1 | , Gardner SB. Winter I'D. Coiifidfticf liili'rvn! Aiuiliisis.version T-II: London: British Medical Associalion, I99fi.

27 Murphy G, Hawk J. The prevaleiKe uf antinuclear antibodies Inpatients with apparent polymorphic liglit eruption. Br / DITIIKIIOI1991; 125:448-51.

28 Ortel B. Tanew A. Woll'f K, Hoiiigsmanii H. Polymorplious li(̂ hteruption: action spectrum and photoprotectlon. / Am Attid Dcniui-tol 1986; 14: 748-53,

29 Petzelbauer P. Binder M. Nikolakis 1' cI (d. Severe sun sensitivityand the presence of antlnuclear antibodies in patienis withpolymorphous lif̂ lit eruption-like lesions. A forme fruste ofphotosensitive lupus erylliemato.sus.- / Am Acad Denmiln! 1992:26: f)8-74.

Jll Kiss M. Husz S. Dobozv A. The occurrence of antiimclear. anli-

SSA/Ko and anti-SSB/La antibodies in patienis with polymor-phous light eruption. Aciti Derm Venerfo! 1991: 71; J4] - 5.

51 Cahn M. Ixvy li\. Shalfer B, Polymorphous light eruption. A ten-year lollow-up and evaluation. Arih Dcrmalol ]9li5; 88: 75(1-8.

il Jansen CT, Karvonen |. Polymorphous light eruption. A seven-year-lbllow-up evaluation of 114 patients. Arcli Oenmitol 1984:120: Sf>2-S.

5 3 Furukawa K Kashihara-Sawaini M. Lyons MB. Norris DA. Hiruiin;;of antibodies to the extractable nuclear antigens SS-A/Ko and SS-B/La is induced on the surface of human keratinocytes by ultra-violet light (UVL): implication.s for the pathogenesis of photosen-sitive cutaneous lupus. / Invest Dcrwalol 1990: 94: 77-8S.

34 Mond CB. Peterson MGE, Rothfield NF. Correlation of anti-Koantihody with photosensitivity rasli in systemic lujnis erythema-losus patieiits. Arthrilis Rheum 1489: J2: 202-4.

3 5 Kind P, Lehmann 1'. i'lewig C. Photolesting in lupuserythematosus. / Invest Dmmito! 1995: 100: 55-7S.

56 Lee LA. Roberts CM. Frank MB ci al. The autoantibody response toRo/SSA in cutaneous lupus erythcmatosus. Arch Uermaiol 1994;130: 12(i2-S.

© 1997 British Associalion of Dermatologists. Briiisli jouniti! oj . 1 36.