Introduction

Sunscreens are known to protect in the ultraviolet (UV) domain (UVB and UVA). However, until recently visible light (VL) (400–700 nm) was considered as devoid of any photobiological effects on cutaneous tissue. It has now been demonstrated that VL stimulates pigmentation, especially in melano-competent subjects.^1-3 Protection from VL combined with UV photoprotection significantly prevents hyperpigmentation disorders such as melasma,^{4, 5} post-inflammatory hyperpigmentation⁶ and actinic lentigos.⁷ More specifically, the high energy visible light (blue and violet light) is responsible for skin pigmentation by inducing a dose-correlated hyperpigmentation, while red light does not induce pigmentation in vivo.¹ Tinted sunscreens containing opaque pigments with large particle sizes, such as iron oxides, are the only currently available, efficient topical photoprotection products against visible light-induced hyperpigmentation. Several methods have been developed to evaluate the VL protection capability of sunscreens, but in vivo evaluation remains the most representative of real-life use. Recently, a group of experts agreed and proposed the calculation of a VL protection factor (VL-PF).⁸ This corresponds to the ratio of change of colorimetry measurements (individual typology angle, ITA°) between unprotected areas and protected areas after several days of visible light exposure using a solar simulator equipped with appropriate filters and measured using a chromameter.⁸ However, in vivo methods have their limitations due to ethical concerns over the photo-exposure of subjects, and related time and costs. Consequently, these methods are not ideal for screening multiple prototypes, whereas in vitro methods would be more ethically acceptable and flexible. However, no correlation has yet been demonstrated between in vitro and in vivo methods. The aim of this study was to evaluate the correlation between these two types of method.

Materials and methods

Results

The transmission spectra of the tinted sunscreens showed that the most effective protection was against high energy visible light (Fig. 1). Twenty subjects (13 women and seven men) with a mean age of 34.9 years (range: 18–49 years) were included in the in vivo study; all completed the study. The calculated VL-PF was equal to 1 for the untreated zone and close to 1 for the zone treated with the non-tinted sunscreen A (Fig. 2a). For the non-tinted sunscreen A, a marked hyperpigmentation was observed on the exposed area compared to the unexposed (Fig. 2c). Despite the claim of protection against VL, we observed differences in the VL-PF between the tinted sunscreens. The sunscreens E and F presented the highest VL-PFs, 3.28 and 3.40, respectively, (Fig. 2a) with no visible hyperpigmentation compared to unexposed skin (Fig. 2c). The VL-PF was reinterpreted as a percentage using the formula (1-(1/VL-PF)) ×100 in order to facilitate comprehension by dermatologists and patients, and called pVL-PF. The pVL-PF was calculated for each sunscreen formulation; 0% corresponds to an untreated exposed area and 100% corresponds to complete protection from the VL (equivalent to an unexposed area). The lowest factor was for the non-tinted sunscreen A at 9.7%, and the highest were for sunscreens E and F with 65.7% and 66.4% respectively (Fig. 2b). The coefficient of determination, R², was then calculated between the percentage of blocked light and the pVL-PF of each test area, for every wavelength from 400 to 700 nm, and for every wavelength range starting from the wavelength 400 nm. The correlations were the highest at 420 nm (R² = 0.9910) for separate wavelengths, and between 400 and 469 nm (R² = 0.9904) for wavelength ranges (Fig. 3). For statistical analysis, the Pearson correlation coefficient was used. Correlations were found to be statistically significant for every wavelength from 400 nm to 700 nm, and for every wavelength range (P < 0.001).

Discussion

Our results show a strong linear relationship between the sunscreen in vivo pVL-PF measurement and the in vitro percentage of blocked light. To our knowledge, only one other study has assessed both in vitro and human in vivo evaluations of sunscreen VL photoprotection but no correlation between these two approaches was demonstrated.⁹ Moreover, the in vivo assessment did not correspond to the standard method that is now proposed. We are proposing the pVL-PF calculation which is a new interpretation of the original VL-PF and allows simply evaluating from 0% to 100% the performance on VL-induced pigmentation in order to easily compare different formulations. It is also easier to understand for dermatologists and patients looking for high VL photoprotection. We previously demonstrated that melanocytes directly sense the light through direct stimulation of the opsin-3 receptor.¹⁰ Interestingly, the best correlation between in vivo pigmentation and the in vitro transmittance was observed from 400 to 469 nm, which corresponds to the absorption spectrum of opsin-3. These data support the activation of melanogenesis by the direct interaction of photons with the opsin-3 receptor and that so far, the best way to prevent VL-induced hyperpigmentation is to prevent the photons of the high energy VL to reach the opsin-3 receptor.

In conclusion, the transmittance measurements from 400 to 469 nm were found to be a reliable predictive tool to evaluate sunscreen VL photoprotection efficacy which could be used to screen products during formulation development.