Effets sur la santé de la lumière artificielle

7. Are there potential health risks linked to artificial lights?

The SCENIHR opinion states:

4. OPINION

This opinion is based on a scientific rationale which has taken into account the relevant scientific literature and other accessible and reliable information on physical and technical characteristics of lighting technologies, principles of optical radiation, as well as biological and health effects of optical radiation. Health effects due to optical radiation have been considered both for the general population and for persons with photosensitive or other pathological conditions. Since the assignment also includes evaluation of possible health effects of various types of lighting technologies, additional data regarding lamp emissions was requested and some were obtained from stakeholders. In addition, for assessment purposes, data regarding exposure patterns was sought, but found to be virually lacking. This lack of information has seriously hampered efforts to perform specific risk assessments.

We have received some information regarding emission data, which has been used for our evaluation, for more than 180 different lamps. These lamps represent all major lamp types that are used for general lighting purposes (tubular fluorescent lamps; compact fluorescent lamps (CFLs) with and without a second envelope; halogen lamps that are either high or low voltage; high pressure discharge lamps (metal halide and sodium); light emitting diodes (LEDs); and incandescent lamps, although the degree of representativeness is uncertain. Regarding specific lamp types, CFLs are well represented in this collection, whereas LEDs for example have been measured in only a few cases. Based on the lamp emissions, the Standard EN 62471 (and also IEC 62471 and CIE S009, since they are all identical in this sense) categorizes the lamps according to the photo-biological hazard that they might pose. The different hazards are:

• Actinic UV-hazard for eye and skin.
• UVA-hazard for the eye.
• Blue-light hazard for the retina.
• Thermal retina hazard.
• IR-hazard for the eye.

Following the standards, emission measurements should be performed according to two approaches; namely at a distance where a light intensity of 500 lx is obtained and also at a distance of 20 cm. Based on these measurements, lamps are then classified according to the “Risk Group” (RG) to which they belong. RG0 (exempt from risk) and RG1 (minor risk) do not pose any hazards during normal circumstances. RG2 (medium risk) lamps also do not normally pose any hazards, due to our aversion responses to very bright light sources or due to the fact that we would experience thermal discomfort. RG3 (high risk) include only lamps where a short-term exposure poses a hazard. Importantly, this classification is based on acute exposure responses (a single day, up to 8 hours) and applies only to individuals of normal sensitivity. It should be noted, with respect to RG3 that the risk classification does not consider either long-term exposures or particularly sensitive persons in the population.

SCENIHR’s answers to the questions given in the Terms of Reference are given directly in connection with the questions below:

A: To explore and report scientific evidence on potential health impacts on the general public caused by artificial light of which the main purpose is to radiate in the visible range (as opposed to artificial light where the invisible part of the radiation is the main purpose, e.g. suntanning lamps or infrared lamps). The impacts of the light from all available electrical lighting technologies should be studied, both in the visible and invisible range (with specific analyses of the ultraviolet radiation subtypes UVA, UVB and UVC).

A combined assessment of natural and artificial light shows that adverse health effects due to optical radiation can either occur acutely at certain levels of exposure, or after long-term repeated exposures at lower levels. Depending on the effect (endpoint) of concern (e.g. skin burn, skin cancer, retinal damage, cataract) either the intensity or duration of exposure is of most relevance. In general, the probability that artificial lighting for visibility purposes induces any acute pathologic conditions is low, since expected exposures are much lower than the levels where effects are known to occur in healthy people and are also much lower than in typical summer daylight. The available lamp emission data show that for all investigated hazard outcomes, the absolute majority of lamps are classified as Risk Group 0 (RG0; "exempt from risk"). Most of the rare exceptions are classified as Risk Group 1 (RG1; "low risk"). The very few lamps assigned to higher Risk Groups were either measured without the required UV-shielding glass cover, or at a very short distance (20 cm) which is not the intended use distance for this lamp type.

Standard EN 62471 gives limits that are protective against acute effects, while long-term effects are only marginally considered. Thus the emissions in e.g. the UV range may comply with these limits, but may still have an effect on skin carcinoma incidences when a population is subjected to extensive and large scale exposure to these lamps. A common exposure situation, such as most household lighting, would involve an illumination level which is so low that exposure to potentially problematic radiation is considered negligible (with the possible exception of prolonged task lighting with a lamp close to the body which may lead to UV exposures approaching the current workplace limit set to protect workers from skin and retinal damage). However, according to a worst case scenario developed in the scientific rationale, the highest measured emissions of UV from fluorescent lamps used typically indoors in professional environments, although well below the limits for RG0, could be contributing to the number of squamous cell carcinomas in the EU population. This is in comparison to a hypothetical situation where the same population is not exposed to UV radiation from artificial light indoors. The annual erythemal UV dose expected from the worst case scenario approximately corresponds to the dose one would get from a half week Mediterranean holiday. Fluorescent lamps typically emit less than half of the UV radiation assumed in the worst case scenario. The vast majority of the CFLs tested emit erythemal UV at very low levels, amounting at the most to an extra day of sunbathing a year.

Low levels of UV emissions may occur from certain lamp types (quartz halogen lamps, single- and double-capped fluorescent lamps as well as incandescent light bulbs). These emissions may, in some cases, in particular for certain halogen lamps with poor UV filtering, include UVC in addition to UVA and UVB. UVC is not naturally present on Earth due to the blocking action of the earth’s atmosphere, so any emissions from lamps would provide a novel type of exposure. However, most action spectra on skin and eye effects include UVC. Hence, biologically effective doses take UVC into account and are thus considered in the categorization of the Risk Group, as discussed above. However, detectable levels of UVC do signal a considerable overall output of biologically harmful short wavelength UV radiation. Regarding a possible need for separate UVA, UVB or UVC radiation limits for tungsten halogen lamps and other light sources that emit UV radiation, the Scientific Committee considers that there is no scientific basis for making such specific recommendations beyond the established dose limits.

Evidence from in vitro experiments suggests that blue light at 10 W/m² induces photochemical retinal damage (Class II) upon acute (hours) exposure, and animal experiments and in vitro studies suggest that cumulative blue light exposure below the levels causing acute effects also can induce photochemical retinal damage. There is no consistent evidence from epidemiological studies regarding the effect of long-term exposure to sunlight (specifically the blue component of sunlight) on photochemical damage to the retina (particularly to the retinal pigment epithelium), which may contribute to age-related macular degeneration (AMD) later in life. Whether exposure from artificial light could have effects related to AMD is uncertain. There is no evidence that artificial light from lamps belonging to RG0 or RG1 would cause any acute damage to the human eye. Studies dedicated to investigating whether retinal lesions can be induced by artificial light during normal lighting conditions are not available. Lamp types belonging to RG2 and higher are usually meant to be installed by professionals in locations where they do not pose a risk. Chronic exposure to blue light from improperly used lamps could, in theory, induce photochemical retinal damage. There is however no evidence that this constitutes a risk in practice. It is unlikely that chronic exposures to artificial light during normal lighting conditions could induce damage to the cornea, conjunctiva or lens. Besides the beneficial effect of light, e.g. through synchronising the day-night rhythm, there is mounting evidence suggesting that exposure to light at night while awake (especially during shiftwork), may be associated with an increased risk of breast cancer and also cause sleep, gastrointestinal, mood and cardiovascular disorders possibly through circadian rhythm disruption. Importantly, these effects are associated with light, without any specific correlation to a given lighting technology.

B: To update the SCENIHR report on Light Sensitivity (from 23 September 2008) in light of further evidence, and to examine the aggravation of the symptoms of pathological conditions in the presence of lamp technologies other than compact fluorescent lamps (including conventional incandescent and halogen lamps, halogen lamps with improved efficiency and light emitting diode lamps).

The previous SCENIHR opinion on Light Sensitivity (SCENIHR 2008) identified that some pre-existing conditions (epilepsy, migraine, retinal diseases, chronic actinic dermatitis, and solar urticaria) could be exacerbated by flicker and/or UV/blue light. However, at that time there was no reliable evidence that compact fluorescent lamps (CFLs) could be a significant contributor. This conclusion needs updating as more recent studies indicate a negative role for certain CFLs and other artificial light sources (including sometimes incandescent bulbs) in photosensitive disease activity. There are no published data on the effect of exposure of a photosensitive patient to light from halogen lamps.

There is strong evidence that UV and, in some patients visible light, can induce skin lesions of true photodermatoses. Although sunlight is reported by most patients as the main trigger of disease activity, occasionally severely affected patients over the range of endogenous (and exogenous) diseases report a provocative role for artificial lighting. There is a lack of controlled skin provocation studies relating effects to the magnitude and the wavelength components of the light source, although there is evidence that the shorter wavelength light components (blue or UV) tend to be more effective than the longer wavelength components (red) in aggravating pre-existing conditions. Some research work has been conducted in particularly severely affected individuals suffering from photodermatoses such as lupus erythematosus, chronic actinic dermatitis and solar urticaria. This provides good evidence for the aggravation of symptoms related to these pre-existing skin diseases. Such work needs to be confirmed, and also extended using a range of lamp types over a wider range of diseases in greater numbers of patients. Particular attention seems justified for the individual variability of the conditions for aggravation of such diseases. Until such data exist, it seems reasonable to assume that the UV, and in some cases the blue radiation component of artificial lighting in an as yet undefined number of patients, may contribute to the aggravation of symptoms related to their skin disease, and in the case of lupus erythematosus possibly also to the aggravation of their systemic disease.

Generally, double envelope CFLs emit much less UV radiation than single envelope CFLs. Most LEDs in general use emit little or no UV radiation. However, with the considerable variability of UV/blue light components for lighting technologies, even of the same or similar kind, no general advice can be given and individual optimisation of the lighting technology is advised for these patients.

The effect of light is variable depending on the genetic alterations that are causing inherited retinal degeneration. In specific conditions like Stargart disease, the retinal pigment epithelial (RPE) cells are particularly sensitive to Class II photochemical damage, which is induced by peaks at shorter wavelengths. In other retinal dystrophies, light does not exert any aggravating effect. However, since the causative mutation is seldom known to the patient or their family, and because there is no clear correlation between genotype and phenotype, it is recommended for all patients with retinal dystrophy to be protected from light by wearing special protective eyeware that filter the shorter and intermediate wavelengths.

The previous SCENIHR opinion on Light Sensitivity stated that modern CFLs are basically flicker-free due to their electronic high frequency ballasts. However, it was also noted that studies indicated that residual flicker can occur during certain conditions, at times also related to other circuitry like dimmers operated with the light source, in both CFLs and incandescent bulbs. In principle, there can be a residual sinusoidal modulation of the light of any light source at twice the line frequency of e.g. 50-60 cycles. Any light source operated on DC, after transformation from the AC line, is flicker-free. This has been the predominant case for LED operation, but is also applicable to other lighting technologies, e.g. halogen and incandescent lamps. Flicker cannot typically be observed in static settings above about 60-80 cycles, while in conjunction with dynamic scenes, the effect is still visible at higher frequencies. There is no scientific evidence available to evaluate if conditions such as Irlen-Meares syndrome, myalgic encephalomyelitis, fibromyalgia, dyspraxia, autism, and HIV infection are influenced by the lighting technologies considered in this opinion.

C: If health risks are identified under points A or B, to estimate the number of EU citizens who might be at risk and identify the level of emission/exposure safeguarding the health of citizens and/or means to mitigate or entirely prevent the impact of the problematic parameter of the light technology in question.

All healthy individuals may be at some risk from UV radiation and blue light from indoor lighting, albeit to different degrees due to differences in genetic background and in the type of light source used. Short-term UV effects on healthy people are thought to be negligible. A proper assessment of long-term risks due to daily low level UV exposure is not possible because data on actual personal indoor UV exposure are lacking. Due to this knowledge deficit, it would appear advisable to be cautious and to develop worst case scenarios. The worst case scenario examined in this opinion involved workplace/school exposure to double- or single-capped fluorescent lamps in ceiling-mounted open luminaires. This scenario assumes the validity of extrapolating from studies on animals with short lifespans to life-time human exposures. Furthermore, it assumes the appropriateness of dose-level extrapolation from experimental studies to real human exposures and that all individuals in a population experience the same risk independent of susceptibility factors. If we take lamps with the highest measured UV output (still well within Risk Group 0), such exposure adds the equivalent of 3 to 5 days vacation in a sunny location to the average annual UV dose. Although this would lead to an increase in the personal risk of squamous cell carcinoma, such an increase would remain small (a few % over a lifetime in Denmark). Population-wide exposure to such lamps could, however, add approximately 100 cases of squamous cell carcinomas a year to a base line of 900 cases/year in Denmark. It should be stressed that the UV output of most of the fluorescent lamps tested fall well below this level, and are not expected to affect squamous cell carcinoma incidences. Improper use of lamps belonging to Risk Groups 1- 3 (due to missing or disregarded user information, non-professional installation) could cause retinal damage. While no such cases are known, appropriate measures could be considered to ensure that these lamps are not misused.

The current standardization of lighting lamps and luminaires in four risk categories appears sufficient to limit the personal short-term risk. However, RG0, as it is based on acute effects, should not be taken to imply adequate protection of the general population as a whole from effects after long-term exposure to UV radiation. Nevertheless, it would be useful to communicate information on risk categories to the consumer.

The previous SCENIHR opinion (SCENIHR 2008) stated that a number of patients are exceptionally sensitive to UV/blue light exposure. The number of EU citizens with light- associated skin disorders that would be affected by exposures from CFLs was estimated in the report to be around 250,000. Clearly, the risk for this group of patients is not limited to CFL, but includes all light sources with significant UV/blue light emissions. The lack of proper data precludes any improvement of the estimate of the size of the affected group.

Also photosensitivite patients undergoing photodynamic therapy might be expected to react to CFL and LED sources to a greater extent than to incandescent lighting. This is due to a combination of greater sensitivity of porphyrins to blue light (soret band), coupled with an enhanced blue light emission of these sources. However, such patients are aware of their extreme photosensitivity which needs careful management. For patients with light-associated skin disorders, the previous SCENIHR opinion recommended that, when using CFLs, a double envelope type is preferable. The current opinion supports that position. Double envelope CFLs generally emit much less UV radiation than single envelope CFLs and are much less likely to induce a reaction in patients with light-associated skin disorders. While a second envelope undoubtledly reduces the UV component of any particular lamp, the currently available data, however, documents the high variability of UV and blue light emissions due to different internal design parameters, even for the same externally visible architecture (i.e. also when a second envelope is present). While some compact fluorescent lamps are in the same category, retrofit LED lighting, which does not emit UVR on the physical grounds of the light generation therein, would potentially provide an even better option for such patients. However, for patients whose sensitivity extends into the visible part of the spectrum, it may be necessary to exclude LEDs which have a significant blue component. The UV/blue light irradiation from halogen lamps is also highly dependent on the lamp type. With lamps other than incandescent retrofit halogen bulbs, attention needs to be given to the proper installation, as they are at times sold by the manufacturer to be installed at larger distance or in conjunction with special luminaires or filters against e.g. UV or IR irradiation or to prevent other hazards like fires. While it is unlikely that there would be a significant UV risk from halogen lamps for the general public, provided that protective measures are complied with, the UV content can rise to levels which are of concern for patients with light-associated skin disorders at close operating distances and long exposure times, which is not a very common use pattern for this lamp type.

For individuals with photosensitive skin diseases a list of lamp models (not only types) that are specifically suitable for their situation is needed. The non-representative sample spanning across a wide range of lighting technologies which is provided by Schulmeister et al. (2011) provides a first try. However, important issues like the modificationcurrently assessed. In view of the large number of patients affected by photosensitive diseases it may be advisable to make sufficient information on the emitted spectrum for individual lamp models available to the healthcare professionals and their patients to allow them to choose their lighting solutions optimally.

D: To identify potential research needs related to the areas where the lack or scarcity of scientific evidence prevents SCENIHR from coming to firm conclusions.

The scientific rationale has identified a number of areas where relevant data are lacking regarding the effects of specific lighting technologies on medical conditions. The most important areas where knowledge gaps have to be filled in order to be able to draw firm conclusions related to this opinion include:

• Emission data (ranging from UVC up to 800 nm) characterizing the different lighting technologies – a challenge due to the variation of manufacturing parameters, and a database of these characteristics of specific lamps on the European market.
• Exposure database on indoor visible light radiance to the eye and personal UV exposures from various lamp types compared to ambient outdoor exposure. The database should be established in view of the potential conditioning of the eye due to the largely different voluntary exposure to sunlight from one individual to another, and for the also very different use patterns for UV/ light protective eyewear between individuals and populations.
• Attention should be paid to develop a risk group categorisation that takes into account potential chronic effects like SCC.
• Eye conditions: a. epidemiologic studies of artificial light exposure and ocular pathologies (including AMD); and b. retinal effects of chronic exposure to artificial light for visibility purposes (animal studies).
• The role of various types of artificial lighting sources in photosensitive skin diseases (provocation studies).
• Mechanisms and consequences of exposure to artificial light in the late evening, at night and in the early morning, including circadian disruptions in both shift-workers and in the general population.
• Flicker induced health effects from the residual high frequency (100-120 Hz) intensity modulations.
• The particular role of UVC components in artificial lighting for skin diseases taking into account especially sensitive populations and the role of prior exposure to sunlight.
• The effects of non-incandescent light sources, in particular those with very inhomogenous or biased spectral distribution on colour rendition, fatigue, and other components of the human visual perception.

5. COMMENTS RECEIVED DURING THE PUBLIC CONSULTATION ON THE HEALTH EFFECTS OF ARTIFICIAL LIGHT

A public consultation on this opinion was opened on the website of the EU non-food scientific committees from 19 July to 30 September 2011. Information about the public consultation was broadly communicated to national authorities, international organisations and other stakeholders.

In total, 16 contributions were received of which four were from public authorities, five from industry, one from academia, two from NGOs and six from individuals and three others.

Most of the material submitted was relevant, contained specific comments and referred to peer-reviewed scientific literature. As a result, each submission was carefully considered by the Working Group. Only three submissions from industry disagreed with the preliminary opinion and the submission from academia showed some disagreement. The document has been revised to take account of the relevant comments and the literature has been updated with relevant publications. The scientific rationale was clarified and strengthened in certain respects. The opinion, however, remained essentially unchanged.

Source & ©: , Health effects of artificial light, 19 March 2012,
 4. Opinion and 5. Comments, pp. 79-85.