Susceptibility to UV-A and UV-B Provocation Does Not Correlate With Disease Severity of Polymorphic Light Eruption FREE

Polymorphic light eruption (PLE) is a common photosensitivity disorder that is believed to be of immune origin.^1- 4 Patients with PLE show considerable interindividual variation in their disease presentation, including the severity of the disorder, but the factors influencing the clinical manifestations and severity are poorly understood. An objective way to determine the severity of the disease may be by artificial photoprovocation, in which the dose of UV irradiation needed to elicit PLE is recorded.

In this study, we assessed a range of clinical characteristics of PLE that may be indicative of PLE severity, against the ease of rash provocation by UV-A and UV-B radiation, in a large group of European patients. The main objective was to examine whether UV provocation response may be indicative of clinical disease severity.

This investigation was effected by a European Commission–funded collaboration between 3 European university dermatology departments during the period from January 1, 1999, to December 31, 2001. The centers involved were the Department of Dermatology, Leiden University Medical Center; the Photobiology Unit, Dermatological Sciences, University of Manchester and Hope Hospital; and the Department of Dermatology, University of Athens, Andreas Sygros Hospital.

The study received ethical approval from all local medical ethics committees, and it complied with the Declaration of Helsinki 2000. A total of 143 patients participated (48 Dutch, 29 Greek, and 66 English patients). Patients aged 18 to 70 years were included. The diagnosis of PLE was based on a strict set of inclusion and exclusion criteria (Table 1).

Provocation testing was performed in the winter and spring months, from November 1 to March 31, in the years 2002 to 2005 in the Leiden and Athens patients. Most provocation tests had been performed from January 1, 2001, to December 31, 2001, in the Manchester group.

Both UV-A and UV-B testing were performed on the forearm skin for up to 3 consecutive days but discontinued if a positive result was obtained sooner. All centers performed UV-A testing (Leiden: 48 patients; Athens: 29 patients; Manchester: 66 patients). According to standard practice, a daily dose of 20 J/cm² was applied to a 20 × 5-cm area of the ventral surface of the forearm with a broadband UV-A fluorescent source (Cleo Performance lamps, 0.7% UV output in the UV-B band, 280-315 nm; Philips Lighting, Eindhoven, the Netherlands).⁵ At the Manchester center, the whole forearm was exposed.

The UV-B testing was performed at the Leiden (47 patients) and Athens (13 patients) centers using broadband fluorescent lamps (TL12 tubes, 57.5% UV output in the UV-B band, 280-315 nm; Philips Lighting). The minimal erythema dose (just perceptible erythema at 24 hours) was first determined on the ventral side of the upper arm by use of a photoprovocation testing device containing 10 apertures with increasingly permeable meshes, as previously described.^6- 7 One minimal erythema dose was then applied daily to a 20 × 5-cm area of the contralateral ventral forearm.

A positive provocation response to UV-B or UV-A radiation was defined as the development of clearly visible and palpable erythematous papules or vesicles on the irradiated field. The last reading was performed 24 hours after the last provocation.

Standardized forms were constructed following a panel discussion of participating dermatologists experienced in photosensitivity diagnosis. The forms comprised 30 questions related to personal and disease characteristics. Patients were interviewed in person by a physician trained in photodermatology.

Five clinical characteristics selected from the standardized forms were used to devise a PLE severity score (PLESS), as described by Ling et al.⁸ The clinical characteristics included the total duration of symptoms in 1 year expressed in months, the possible involvement of the face, a visual analog score for severity of itching, the time taken for the rash to resolve, and the treatment history. As can be seen from Figure 1, the PLESS ranged from 0 to 11.

For statistical analysis, SPSS version 12 (SPSS Inc, Chicago, Ill) and SAS version 9.1 for Windows (SAS Institute Inc, Cary, NC) were used. Continuous data for the clinical characteristics were described using means, standard deviations, minimum, median, and maximum (Table 2). Dichotomous and categorical data for the clinical characteristics were described by absolute and relative frequencies (Table 2). Possible differences between patients with UV-A/UV-B–elicited rashes and patients without elicited rashes were tested (univariably) using the Wilcoxon–Mann-Whitney U test or the t test in cases of continuous data or scores and the Fisher exact test in cases of contingency tables. Multiple logistic regression analyses were performed for clinical characteristics with borderline significance (facial involvement, time needed for the rash to resolve, and the number of months per year a patient had PLE) to examine the common influence of disease characteristics (including age and sex) on the outcome of UV-A and UV-B provocation. The level of significance was set at P<.05.

In total, 143 patients with PLE were tested with UV-A lamps. The overall positive response of UV-A provocation was 78.3%, ranging from 68.8% in Leiden to 87.9% in Manchester. The number of exposures on consecutive days necessary to induce PLE and the number of elicited papules are given in Table 3. In most cases, papular responses with a variable intensity of erythema were observed (Figure 2).

A total of 60 patients with PLE were tested with UV-B lamps; a papular rash after UV-B provocation was achieved in 28 patients (46.6%), comprising 9 (69.2%) of 13 Athens patients and 19 (40.4%) of 47 Leiden patients. Twenty-three (38.3%) of 60 patients with PLE had a positive UV-A and UV-B provocation. Thirteen patients (21.7%) did not develop a rash with either source, and 5 patients (8.3%) responded only to the UV-B source. There was no difference in the UV-A or UV-B provocation results between countries, but there was a trend for a higher number of positive UV-A provocations in England (where a larger surface area was challenged) and a lower number in the Netherlands.

As depicted in Table 2, a high percentage of patients (n = 64; 47.4%) was found in the moderate category, 33 patients (24.4%) belonged to the mild category, and the remaining 38 patients (28.1%) were in the severest category. The distribution of PLESS categories differed between the countries. On average, the PLESS was higher in Dutch than in Greek patients (mean ± SD, 6.3 ± 2.0 vs 5.1 ± 2.1; P = .02) and English patients (mean ± SD, 6.3 ± 2.0 vs 4.7 ± 2.0; P<.001), whereas the severity scores of Greek and English patients were similar (P = .36).

An overview of the disease characteristics of the evaluated patient population is given in Table 2. The described data are a selection of the most relevant features obtained from the standardized interview. Skin type showed similar distribution between the countries (P = .22 using the χ²test; data not shown).

As shown in Table 2, we examined the association between UV provocation and PLESS. Overall, neither the UV-A nor UV-B provocation response showed any significant association with PLESS. In UV-A non-responsive patients, the PLESS did not differ significantly from that of UV-A responsive patients (mean, 4.7 vs 5.5; P = .09). Similar results were also found in cases of UV-B provocation.

Analysis of individual features included in the PLESS revealed a borderline significant association between facial involvement and time needed for the rash to resolve with UV-A provocation. A high number of months that a patient was affected per year was associated with a positive UV-B provocation (P = .04).

Further analysis of clinical characteristics not included in the PLESS showed a significant association only between a positive UV-A provocation and sun-reactive skin type I. This correlation was not found in cases of UV-B provocation. The examination of a potential relationship between the disease characteristics revealed that facial involvement was associated with a longer rash resolution time (P = .03) and a higher number of months per year that the patient had PLE (P = .02).

Polymorphic light eruption is a common disorder of unknown pathogenesis that shows a high variability in its clinical characteristics and severity. We have examined whether the ease of disease provocation by UV-A and/or UV-B irradiation correlates with clinical features of the disorder, including those indicating disease severity.

In our examined group, PLE induction was found in 78.3% of patients after UV-A and 46.7% after UV-B irradiation. The UV-A results are consistent with data from published studies, but our number of UV-B responses was higher and there are large variations in irradiation protocols.^3,9- 13 The distribution of PLESS in our PLE population was normal, with 24.4%, 47.4%, and 28.1% found in the mild, moderate, and severe categories of PLE severity, respectively. We originally assumed that the outcome of photoprovocation would be a reliable tool to demonstrate an individual's clinical severity. To our surprise, the PLESS showed no significant association with either UV-A or UV-B provocation. Therefore, our findings indicate that the provocation results are not predictive of the clinical severity of PLE.

Our results differed from those reported by Palmer et al,¹⁴ who concluded that the clinical severity of PLE was correlated with the ease of provocation to solar-simulated radiation. The difference between the results could be attributable to large differences in group size (n = 9 vs n = 143), differences in the UV source used (with elicited PLE lesions in different subsets of patients with PLE), or the method of severity scoring. The number and types of questions included in our scores differed substantially from the method of Palmer and colleagues. Nevertheless, both their and our questions on the severity score showed internal consistency.^8,14

This lack of correlation between UV provocation and PLESS was observed in all 3 participating countries. The outcome of provocation testing was similar in the 3 centers, but the PLESS scores were higher in Dutch patients. This was attributable to higher visual analog scores for itching, more facial involvement among Dutch patients, and a relatively large proportion of Dutch patients who scored high on treatment history. Differences in answers about previous treatment could have been influenced by the moment of inclusion in the study in relation to the duration of the disease. The mean duration of PLE at inclusion in this assessment among Dutch patients was 19 years, whereas among English and Greek patients it was 12 years.

Clinical features of PLE that have previously been reported to correlate with UV-A photoprovocation testing are the duration of the disease, the persistence of naturally occurring lesions, and the interval from accidental exposure to the development of lesions.¹⁵ In this study, we found correlations only between UV provocation and facial involvement of PLE, a high number of months a patient was affected with PLE during 1 year, and sun-reactive skin type I. Positive photoprovocation in patients with PLE has varied among reported studies between 0% and 100% during the last 20 years.^{9- 12,16- 21} This large variation can be attributed to differences in UV source used, number and total dose of exposures, and possibly some patient characteristics. Our data confirm that UV-A lamps give the highest number of positive responses and are therefore the most suitable sources for confirming the disease. However, because a large proportion of false-negative responses remains in patients with a classic history of PLE, the outcome of UV-A testing cannot be regarded as the gold standard and essential for making the diagnosis, but it can be helpful in some cases. The gain for clinical practice of UV-B provocation testing is even smaller, although the results of UV-B provocation testing might provide information that helps to unravel pathogenic mechanisms of the disease. We already showed in another study that the UV-B–induced migration of epidermal Langerhans cells and neutrophils differs between patients who develop PLE rashes after UV-B and UV-A provocation.⁶

On the basis of our results, we suggest that UV-A and UV-B testing have only a limited role in clinical practice. The clinical severity and the provocation test outcome were both regarded as important for therapeutic advice to patients with PLE. However, since they are not correlated with each other, how these measures could contribute to therapeutic decision making (eg, choice of light source for UV hardening and/or preventive measures) should be investigated.

Correspondence: Artiena Soe Janssens, MD, Department of Dermatology, B1-Q, Leiden University Medical Center, PO Box 9600, 2300 RC Leiden, the Netherlands (a.s.janssens@lumc.nl).

Financial Disclosure: None reported.

Accepted for Publication: August 24, 2006.

Author Contributions: Dr Janssens had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Janssens, Pavel, Antoniou, Diepgen, de Gruijl, and Rhodes. Acquisition of data: Janssens, Ling, Winhoven, Anastasopoulou, Stratigos, and Rhodes. Analysis and interpretation of data: Janssens, de Gruijl, and Rhodes. Drafting of the manuscript: Janssens, Winhoven, Anastasopolou, Diepgen, and Rhodes. Critical revision of the manuscript for important intellectual content: Janssens, Pavel, Ling, Stratigos, Antoniou, de Gruijl, and Rhodes. Statistical analysis: Janssens. Obtained funding: de Gruijl. Administrative, technical, and material support: Janssens, Antoniou, Anastasopoulou, de Gruijl, and Rhodes. Study supervision: Pavel, Stratigos, Diepgen, de Gruijl, and Rhodes.

Funding/Support: This study was supported by grant QLK4-CT01-0015 from the European Union Framework V Programme.

Role of Sponsor: The funding organization had no role in the design and conduct of the study, in the collection, analysis, and interpretation of the data, or in the preparation of the manuscript, review, or approval of the manuscript.

Acknowledgment: We are grateful to Thomas Bruckner, MD, from the Department of Social Medicine, Center of Occupational and Environmental Dermatology, University Hospital of Heidelberg, Heidelberg, Germany, for his statistical assistance. We thank all members of the SUNALL consortium for their intellectual contributions during panel discussions and Rein Willemze, MD, PhD (Leiden, the Netherlands), for his valuable comments.