1 Introduction

Phototesting in dermatology has been performed for almost a century with three main types of tests: (i) Determination of the minimal erythema dose (MED); (ii) 3–5 consecutive days of photoprovocation to provoke specific skin diseases [1], and (iii) photopatch testing to investigate if an eczematous reaction is caused by a chemical applied to the skin combined with subsequent UVR exposure [2,3,4].

If ultraviolet radiation (UVR) testing is only performed occasionally, the methodology tends to vary greatly from clinic to clinic. When phototesting is performed on a regular basis, a desire to standardize and simplify the procedure arises. When performing the MED test, small test fields are exposed to different incremental UVR doses. During the provocation test, a larger skin area is exposed to UVR for 3–5 consecutive days to provoke underlying diseases such as polymorphic light eruption, chronic actinic dermatitis, actinic prurigo, solar urticaria, or porphyria. The UVR dose used for the provocation test is determined by the MED test. The test procedure involves many determining factors: where on the body to test, the position of the person tested, the size of the irradiated area, and cover of the non-irradiated skin [5, 6]. The MED test must produce erythema with sharp borders to adjacent unirradiated skin as that is one of the definitions of MED [7].

Another important factor is the choice of light source, summed up by Ibbotson et al. in Appendix S3 [4]: solar simulated irradiation is commonly used as well as broadband UVB and UVA, or narrowband UVB. Small physical doses of UVB easily produce erythema, whereas UVA requires much higher irradiance to produce erythema-eliciting doses within a reasonable amount of time [8, 9]. Doses are often given in standard erythema dose (SED) which have equal erythema-provoking properties independent of wavelength. The test area must be homogeneously illuminated with direct irradiation from the light source or via a light guide. The use of monochromator-selected narrow wavebands is commonly used for MED testing of light sensitivity [4, 10] and has proven particularly useful for research purposes and in drug-induced photosensitivity testing [11]. Wavelengths of 300–305 nm and 360–370 nm are commonly used in phototesting [4]. This monochromator testing equipment is no longer commercially available, and new light sources are needed.

In this paper, we describe equipment and methods developed to overcome difficulties and impracticalities related to the MED test and UVR provocation test.

2 Current methodology

2.1 Background and procedure demand

Previously applied methodologies and requirements for optimizing the test procedure are reviewed in the following.

2.2 Test area

Different test areas have been in use, primarily the skin on the back, the buttocks, and the arms. The skin on the back is most frequently used as it is less curved than the arm and the buttocks and provides a large and relatively flat test field for performing both the MED test and provocation test on adjoining areas of skin with similar UVR sensitivity.

2.3 Position of patient

When using the skin on the back for testing, the patient is often placed in a sitting position. When irradiation takes long the patient tends to lean forward and away from the light source, and respiration movements may likewise change the distance to the skin when there is no fixed connection between light source and skin surface. A nurse may hold the lamp or a single light guide, but this will not solve the problem as it is difficult to keep it steady for an extended amount of time [6]. To address this, we have chosen to test our patients in a horizontal position lying face-down on an examination couch and construct light sources to be mounted directly on the skin or otherwise being able to accommodate body movements.

2.4 Demarcation of irradiated areas

The MED test areas are round or square, and about 1 cm2 in size (Fig. 1a, b). Light guides are frequently used to direct UVR to the skin (Fig. 1c). When hand-held, they are difficult to keep in direct contact with skin to achieve clear borders of erythema. The lamps must be able to irradiate at least six different UVR doses simultaneously to cover a broad dose range (Fig. 1b). A group of light guides have been used in MED testing (Fig. 1c), but these are unsuited for performing the provocation test where the irradiated area typically measures 5 × 5 cm.

Fig. 1
figure 1

Equipment previously used for phototesting: round erythema reactions (a). Different doses may be obtained by covering the six ports at various time points during irradiation (b). Different UVR doses may be obtained by six filtered light guides (c) or by inserting a grid into the light beam (d). Broad spectrum UVA, UVB, and blue light sources with fluorescent tubes (e)

Full size image

To ensure a clear demarcation of the MED test sites, the demarking material must be in a stable position and in direct skin contact to avoid unclear borders and to accommodate body movements. Reading of the erythema degree that defines MED is done by subjective assessment of barely perceptible erythema ( +) or uniform erythema with sharply defined borders (1 +), demonstrating the importance of the described demands. The ( +) reaction is easiest to evaluate and agree on by different observers and might be preferable to use [7].

2.5 Simultaneous incremental dosing

Different UVR doses are frequently obtained by manually covering MED test sites at different time intervals (Fig. 1b) or may be obtained by grids with different hole density or filters in front of light guides [6, 12] (Fig. 1d). We have constructed a printed label, the MED Test Patch, allowing six different test doses to be obtained by one irradiation session (Fig. 2a, b).

Fig. 2
figure 2

The MED Test Patch with six fields (ports), including five density filters with different UV transmission (a). Erythema obtained in square test fields (b). Solar simulator (c) fits to the MED Test Patch (d, e). UVB LED (309 nm) constructed to fit the MED Test Patch (f, g). UVA1 LED mini lamp (h) to be mounted directly onto the test patch on the skin (i). The LED lamps have maximal emission at 309, 370, or 415 nm (not shown)

Full size image

2.6 Light sources

Traditionally, MED testing is performed to establish the starting dose in phototherapy which explains the use of lamps with broad spectrum fluorescing UVB, UVA, and narrow band UVB tubes (Fig. 1e). During testing, minor body movements are not significant if the lamps are placed far from the skin surface. UVB lamps of this kind easily provoke erythema, but the UVA lamps with fluorescing tubes are too low-powered. For this reason, high-powered (Xenon and metal halide) lamp types have been used but they are heavy and generate much heat when used for provocation testing of larger skin areas. However, the xenon arc with filters is still in use for MED testing in combination with hand-held light guides. Low-cost, easy-handle light sources useable for both MED and provocation testing are needed and will be described in this paper.

2.7 Measuring method

All spectral measurements in relation to the development of the MED Test Patch and reproducibility hereof was performed with a spectrometer system (Fl3095, J&M Analytische Mess- und Regeltechnik GMBH, Aalen, Germany) consisting of a xenon lamp, and a monochromator and photo diode array [13]. Measurements were performed from 280 to 390 nm. Spectral irradiances of the light sources at every nm were measured by a Bentham DM 150 Double monochromator with fibre optics (Bentham Instruments Ltd., Reading, UK). Calculations of SED were based on the CIE erythema action spectrum [8] and one SED defined as 10 mJ/cm2 erythema-weighted radiant exposure [14,15,16]. In this paper, we follow the demands to visualisation of spectral irradiance curves as described by O’Mahoney et al. [17]. The uniformity of the irradiance was measured by SunSaver dosimeters in six spots, corresponding to the location of the six ports in the MED Test Patch [18], and by Sunpoint®Paper from Lawrence Hall of Science (University of California, Berkeley, CA, USA). SunSaver measurements of the irradiance during warming up time were performed every second for 10 min.

3 Development of a novel, simplified methodology for phototesting

3.1 Filter system

The MED Test Patch label consists of six test areas printed onto Trespaphan GND 0.05 mm polypropylene folio (Fig. 2a). Each of the six test areas measures 12 × 12 mm. One of the MED testing areas consist of an open area (no filter) and five areas with different layers of white pigmented enamel printed onto the folio to obtain density filters. The remaining folio is printed with a reflecting material with no light transmission. A linear relation between white enamel concentration and absorption of UVR was found for 310 nm (Fig. 3). The printing enamel consists of different concentrations of white enamel added to a transparent enamel base (Table 1). The chosen transmission intervals correspond to obtaining 20% decremental doses (linear logarithmic function). Figure 4 illustrates the absorption range between 24 measurements (12 test patches measured twice) of different labels including three production batches. The absorption range in average was about ± 3% between 300 and 390 nm (Table 2) Table 2 also illustrates how the transmission of erythema-weighted irradiance for the six MED Test Patch ports varies with the different light sources. For shorter and longer wavelengths, the absorption differs (Fig. 5). Without correction, broad spectrum UVR shorter than 300 nm, and longer than 390 nm should not be used along with the printed label. This soft, disposable label is fixed to the skin by tape and the test areas are of a square shape to ensure easy detection of sharp erythema borders (Fig. 2a, b).

Fig. 3
figure 3

Illustration of the linear relation between white enamel concentration and the transmission of 310 nm wavelength

Full size image
Table 1 Printed MED Test Patch with six test areas (ports)
Full size table
Fig. 4
figure 4

Spectral transmission of each of the five ports of the MED Test Patch in relation to unfiltered radiation. The intended transmissions are given as 80%, 64%, 51%, 41%, and 33% (20% decrements)

Full size image
Table 2 The upper table section describes the intended transmission of the MED Test Patch ports compared to the measured transmission and standard deviation (SD) at 310, 360 nm, and 300–390 nm (see Fig. 4)
Full size table
Fig. 5
figure 5

Spectral transmission measurements through the five ports of one MED Test Patch using our solar simulator with each port in front. The transmission is relative to unfiltered solar simulator irradiation (port 1)

Full size image

3.2 Light sources

We describe four light source constructions, illuminating a 4 × 6 cm area, designed to fit the MED Test Patch, and adhering tightly to the Test Patch to prevent direct or scattered light from reaching the surrounding skin. In this way, shielding of the skin around the Test Patch becomes unnecessary (Fig. 1e). Without the Test Patch, the lamps are used directly for the photoprovocation test.

The first of four different light sources are a solar simulator consisting of a 150 W Xenon arc lamp and a reflector constructed to distribute light evenly to the 4 × 6 cm test area positioned 15 cm from the lamp within a housing (Fig. 2c). To compensate for motion due to respiration, a distance up to 6 mm can be accommodated without any light scatter reaching the surrounding skin. As the light source is rather large, it is fitted with wheels and mounted on a motorized telescopic arm to adjust it to skin level (Fig. 2d, e). The base is flat to fit under an examination couch (Fig. 2c). Adjustments of the spectral range may be performed by mounting a filter in an exchangeable frame into the light path to remove UVC and most wavelengths below 300 nm (Schott WG 305, Mainz, Germany), or to transmit visible light only (filter Schott GG 400, Mainz, Germany). The spectral irradiance is seen in Fig. 6a. For the unfiltered solar simulator, it takes 85 s to deliver 1 SED and filtered with WG 305 it takes 15 times longer (Table 3). The time to obtain a stable irradiance and the uniformity of the irradiance for the six ports can be seen in Table 4. A stable irradiance is defined as irradiance with no significant change over time.

Fig. 6
figure 6

Spectral irradiance on a linear and log scale of four different light sources and their erythema effective spectral irradiance: a WG 305 filtered solar simulator (Fig. 2c); b 309 nm LED lamp (Fig. 2f); c 370 nm LED lamp (Fig. 2h); d 415 nm LED lamp

Full size image
Table 3 Irradiation time to obtain six different UVR doses in the given dose ranges depending on light source.
Full size table
Table 4 Irradiance was measured in six spots corresponding to the location of the ports in the MED Test Patch
Full size table

The three other sources use LEDs for testing with UVB, UVA1, and blue light. They are constructed to fit the irradiation area of 4 × 6 cm and lightweight enough to mount directly onto the Test Patch.

The UVB source (Fig. 2f, g) consists of 35 UV LEDs with a stated peak wavelength of 308 nm and a half-width of 12 nm (PFF-H10-F35 PhotonWave Corp., Woburn MA, USA). With the chosen setting, it delivers 1 SED in 82 s. The spectral irradiance is illustrated in Fig. 6b. The measured peak wavelength is 309 nm with a half-width of 14 nm. The diodes were placed to optimize the uniformity of irradiance in the field (Table 4) with a distance from the LED to the skin of 10 mm. There is hardly any warmup time (Table 4).

The UVA source (Fig. 2h, i) consists of 35 (5 × 7) UV LEDs with a stated peak wavelength at 365 nm and a half-width of 9 nm (NCSU033A, NICHIA EUROPE, NL). The LEDs are selected high power editions with nearly identical intensity. The distance from the diodes to the skin is 15 mm. As the diodes generate heat, the lamp has a built-in ventilator. As the UVA source is equipped with high-power diodes it only takes 6 min and 15 s to deliver 1 SED. The spectral irradiance is illustrated in Fig. 6c. The measured peak wavelength is 370 nm with a half-width of 10 nm. Due to equipment age, the uniformity of the intensity is rather uneven, and the warmup time is 3 min (Table 4).

The blue light source consists of LEDs with a peak wavelength at 415 nm and a half-width of 14 nm (LHUV-0415-A065, LUXEON UV1, Lumileds Holding, USA). The distance between the LEDs and the skin is 19 mm. This light source is mainly used for photoprovocation testing, and not intended to be used together with the MED Test Patch. The irradiance is 1648 W/m2 and the spectral irradiance is shown in Fig. 6d. The uniformity of the irradiance is very good with practically no warmup time (Table 4).

The exposure time to obtain six different UVR doses in one session for the described light sources is seen in Table 3. The highest dose range is used to test persons with normal UVR sensitivity, whereas the lower dose range is used for light-sensitive persons. The degree of erythema will often not only give one MED but also include higher degrees of erythema, severe erythema, and oedema. When just perceptible UVR dose is needed the exposure time may be up to three times shorter (Table 3).

4 Discussion

The first attempt to measure sun tolerance was made just before the year 1900 by Niels Finsen who may indirectly have begun the tradition of using solar simulators: xenon arc lamps equipped with a Schott WG 305 filter which removes most wavelengths below 300 nm, emulating the solar spectrum. Additionally, our solar simulator fits the MED Test Patch label and its absorption characteristics (Table 2). The solar simulator may also be filtered for testing with visible light using a 3 mm Schott GG 400 filter. Testing with visible light nowadays is preferably done by LEDs emitting maximally at 415 nm to eliminate most UVR (Fig. 6d). The light source housing is almost identical for all the LED irradiation sources used for UVA1, UVB, and blue light testing (Fig. 2f, h).

Broadband UVB such as Phillips TL12 (Eindhoven, The Netherlands), and Waldmann UV6 (Villingen-Schwenningen, Germany), and narrowband UVB such as Philips TL01 are often used for testing and may be preferable if the MED test is performed to determine the starting dose for phototherapy with fluorescent tubes of similar bandwidth. Narrowband UVB (TL01) can be emitted by a single, short, fluorescent tube, particularly convenient for testing on the arm. The drawback of larger light sources is the need for extensive cover of the patient’s body during testing (Fig. 1c, e). Narrowband UVB (TL01) is usable, but it is important to be aware that fewer SEDs (factor 0.6) is needed to elicit erythema than calculated from the CIE erythema spectrum [19]. This may also be the case for the 309 nm light source but is not documented.

MED testing with UVA demands a high-power light source to provide enough energy to elicit all desired degrees of erythema. A metal halide lamp UVA SUN 3000 (340–400 nm) (MUTZHAS, Munich, Germany) has been in use but it produces much heat, illuminates a large area, and is big and heavy [20, 21]. In 2012, Wan et al. [22] described a handheld UVA LED source which can irradiate a small spot for MED testing but not larger areas. We chose to construct a very lightweight high-intensity LED lamp which can be mounted directly onto the Test Patch without illuminating adjacent skin areas. The illumination area is also large enough to use for photoprovocation. Our lamp lives up to the known LED stability and has now been in use for 17 years without needing repair. The uniformity of the irradiance, however, has decreased with age (Table 4). The dose to deliver one MED can still be correct as the irradiance at each port of the Test Patch and the absorption of each port can be considered. None of the LED light sources resemble natural daylight which does not pose any problem except when phototesting is performed to determine the sun protection factor (SPF) of a sunscreen.

Monochromator-equipped light sources are standard in the UK giving a choice of many different wavelengths to determine an action spectrum for a specific disease [11]. Such equipment is, however, rather expensive and may be replaced by LEDs as only a few wavelengths are generally used (300–305 nm and 360–370 nm) [4], accommodated by our chosen LED devices with measured maximum emission at 309 nm, and 370 nm. The risk of blurred MED test fields when using handheld light guides to direct light from the monochromator to the skin is avoided by mounting the LED sources directly on the skin.

The most common way of producing incremental doses of UVR is to cover small windows in a template at different times or to open small fields at different times points [23] (Fig. 1b). Both tasks demand complete focus, especially when time differences between each procedure become short. This problem is solved using the MED Test Patch where all incremental doses are obtained by fixed absorption. The layer thickness of the white enamel, however, may vary from batch to batch, as illustrated in Fig. 4 and Table 2, and descriptions of its practical use was published in 2007 and 2008 by Faurschou and Wulf [24,25,26]. A similar approach was described in 2018 by Conant et al. [27]. In 1969, Berger et al. [28] obtained the same effect by equipping a solar simulator with light guides, each providing six different intensities (Fig. 1c).

Reading of the erythema degree that defines MED is done as just perceptible erythema (degree ( +)) or uniform erythema with sharply defined borders (degree 1 +). Sharp borders rely on the template, e. g. the MED Test Patch, adhering tightly to the skin surface. A gap will blur the borders, a problem which increases with the size of the light source. We have chosen square test areas assuming that it is easier to see sharp borders compared to circular test areas as seen where light guides are used (Figs. 1a, 2a). A grid can also be used as an alternative to the printed test patch (Fig. 1d) with the advantage of not changing the light spectrum but must be placed at a certain distance to the skin for the light distribution to be even. This reduces the possibility of sharp borders of the irradiated fields unless combined with a perforated patch on the skin. We have constructed the light sources to align tightly with the MED Test Patch, ensuring sharp erythema borders and a complete definition of the distance from the light source to the skin. As no light can escape the intended exposure area, the patients need not be covered with a protective fabric on skin outside the test field when room temperature allows. In most other cases, it has proven necessary for both patient and examiner [24]. We have chosen the use of an open square as 100% dose instead of unprinted polypropylene material. Since the Test Patch material does not have precisely the same absorption throughout the UV spectrum, it may introduce some uncertainty, especially for wavelengths below 300 nm, and alter the percentual absorption of the other five dose levels, particularly at shorter and longer wavelengths (visible light, short UVB, and UVC) (Table 2).

In this paper, the time to deliver one SED is given. The number of SEDs to elicit MED in healthy individuals is totally dependent on skin pigmentation (skin type). About 3–6 SEDs will elicit MED in white-skinned people of Northern Europe, and as high as 20 SED in black-skinned people [29]. This should be taken into consideration when choosing the test dose range (Table 3).

The Test Patch is disposable, and the light sources are not in contact with the patient’s skin, hence no need for cleaning or fabric washing facilities. The Test Patch is made of polypropylene (PP) which converts to CO2 and H2O when burned. It takes around 20 years for PP to biodegrade in nature.

The photoprovocation test is performed with light sources irradiating an area of 4 × 6 cm, traditionally using 70% of the MED dose. The blue 415 nm LED lamp is mainly used for this purpose, not for MED testing, and for this reason, the altered absorption by the MED Test Patch in wavelengths > 400 nm is of no importance.

It took years to develop all the equipment needed for phototesting. We realize that resources and many years of frequent use are required to justify the expenses. However, we hope to inspire others, having now successfully used the equipment for up to 17 years.

The advantage is having a standardized UVA and UVB MED test method which is easy and performable within an hour. It improves the accuracy of the results and performing the tests does not entirely depend on skilled personnel, except for reading and interpreting the degree of erythema.