The impact of air pollution on the facial skin of Caucasian women using real‐life pollutant exposure measurements

Abstract

Background

To date, studies examining the effect of air pollution on skin characteristics have relied on regional pollution estimates obtained from fixed monitoring sites. Hence, there remains a need to characterize the impact of air pollution in vivo in real‐time conditions. We conducted an initial investigation under real‐life conditions, with the purpose of characterizing the in vivo impact of various pollutants on the facial skin condition of women living in Paris over a 6‐month period.

Materials and methods

A smartphone application linked to the Breezometer platform was used to collect participants’ individual exposures to pollutants through the recovery of global positioning system (GPS) data over a 6‐month period. Daily exposure to fine particulate matter (PM 2.5 µm and PM 10 µm), pollen, and air quality was measured. Facial skin color, roughness, pore, hydration, elasticity, and wrinkle measurements were taken at the end of the 6‐month period. Participants’ cumulated pollutant exposure over 6 months was calculated. Data were stratified into two groups (lower vs. higher pollutant exposure) for each pollutant.

Results

156 women (20–60 years‐old) were recruited, with 124 women completing the study. Higher PM 2.5 µm exposure was associated with altered skin color and increased roughness under the eye. Higher PM 10 µm exposure with increased wrinkles and roughness under the eye, increased pore appearance, and decreased skin hydration. Exposure to poorer air quality was linked with increased forehead wrinkles and decreased skin elasticity, while higher pollen exposure increased skin roughness and crow's feet.

Conclusion

This study suggests a potential correlation between air pollution and facial skin in real‐life conditions. Prolonged exposure to PM, gases, and pollen may be linked to clinical signs of skin ageing. This study highlights the importance of longer monitoring over time in real conditions to characterize the effect of pollution on the skin.

1. INTRODUCTION

Over 99% of the global population live in areas with ambient (outdoor) air pollution levels exceeding the World Health Organization's recommended limits. ¹ Ambient air pollution is made up of various organic and inorganic toxic compounds, including particulate matter (PM), heavy metals, and the gaseous pollutants sulfur dioxide, carbon monoxide, and nitrogen oxides. ²

Ambient air pollution is a known health hazard and its impact on internal organs is well documented. ¹ Higher concentrations of polycyclic aromatic hydrocarbons have been found in more polluted areas. ³ However, the impact of ambient air pollution on skin, the front‐line biological barrier, remains limited. ² Current evidence suggests that air pollution has deleterious effects on skin properties and accelerates skin ageing. ⁴ Various mechanisms are thought to contribute to air pollution‐induced skin ageing including the generation of free radicals and thus oxidative damage, and the induction of inflammatory cascade which promotes skin barrier dysfunction. ⁵

To date, studies examining the effect of air pollution on skin characteristics have largely relied on regional pollution estimates obtained from fixed monitoring sites. To better characterize skin exposure to pollutants and to determine whether there is a link between exposure to pollution and overall skin condition, there is a need to measure the individual subject's pollution exposure in real‐life and in real‐time. Recently, an application (app) was developed by Laboratoires Clarins in collaboration with Connected Physics (https://www.connectedphysics.com) to collect an individual's pollution exposure data. This App accesses an individual's precise location data and connects to the Breezometer platform (https://www.breezometer.com/) which collects data on pollutants in real‐time to deduce the concentration of various pollutants at the individual's location with high spatial and temporal resolution.

The primary objective of this study was to conduct an initial investigation under real‐life conditions, with the purpose of characterizing the in vivo impact of various pollutants on the facial skin on women living in an urban city (Paris). The Breezometer platform served as the means to capture and quantify individual exposures to said pollutants. Cumulative exposures over a 6‐month period were then calculated for the volunteers. Subsequently, distinct cohorts were established for each pollutant, and statistical analyses were applied to compare instrumental data between the designated groups.

2. METHODS

Evaluation of skin condition aimed to collect objective measurements characterizing facial skin appearance. Evaluations were carried out during measurement sessions conducted at 6 months. The study period ranged between June 2019 and December 2020. This study was conducted in accordance with the ethical principles of the Declaration of Helsinki, and good clinical practice guidelines (CPMP/ICH/135/95). The data were collected anonymously following the applicable reference methodology (MR001) issued by the French Data Protection Authority (CNIL), and each participant provided written informed consent and accepted mobile App terms and conditions prior to inclusion.

2.1. Participants

Caucasian women living in the city center and suburbs of Paris, France, were recruited to the study. Assessments of participants’ skin were conducted 6 months after joining the study (assessments were done in January 2020). The women participating in this study had to meet the following inclusion criteria: were healthy and aged between 20 and 60 years old; used cosmetic products; planned to spend over 200 days in the same town; had healthy skin; were non‐smokers. Table 1. summarizes the volunteer's information.

TABLE 1.

Mean pollutant exposure of the two study groups in comparison to the WHO warning threshold. (n) is the number of participants in each group.

	Pollutant
	PM10		PM2.5		AQI		Pollen (tree)
	Higher exposure group	Lower exposure group	Higher exposure group	Lower exposure group	Higher exposure group	Lower exposure group	Higher exposure group	Lower exposure group
Total number of participants	59	62	57	64	60	61	61	62
Age (mean ± SD)	43.5 ± 10.4	43.5 ± 10.4	43.5 ± 10.6	43.5 ± 10.2	44.7 ± 10.2	42.3 ± 10.5	44.6 ± 9.9	42.3 ± 10.7
Photoype 1 (n)	16	17	15	18	15	16	12	23
Photoype 2 (n)	35	37	32	40	37	37	37	35
Photoype 3 (n)	5	2	6	1	4	3	4	3
Photoype 4 (n)	1	4	2	3	3	3	5	0
Photoype 5 (n)	2	2	2	2	2	2	2	1
Mean pollutant exposure (WHO Warning threshold)	35.8 ± 4.5 (45)	23.5 ± 3.8 (45)	18.5 ± 1.4 (15)	12.7 ± 2.5 (15)	67.8 ± 2.7 (59)	61.0 ± 2.4 (59)	1.1 ± 0.2 (NA)	0.9 ± 0.1 (NA)
Average time spent outdoor daily (hours ± SD)	3.8 ± 1.4	3.3 ± 1.6	3.8 ± 1.0	4.0 ± 1.9	3.8 ± 1.5	4.1 ± 2.1	3.9 ± 1.7	4.1 ± 1,8

Abbreviations: AQI, air quality index; PM, particulate matter; WHO, World Health Organization.

2.2. Lifestyle survey

The participants received questionnaires via an application. They reported their daily outdoor activity time.

2.3. Air pollution measurement

An application which collected participants’ location data via global positioning system (GPS) was connected to the Breezometer platform to collect data on the concentration of various pollutants with a spatial resolution of 250 m and temporal resolution of 12 min. ⁶ This platform leverages local information related to the air quality that are collected with sensors such as weather sensors, traffic information sensos, air quality sensors and satellite information. These data are combined thanks to an AI model to generate a dense grid of the distribution of different pollutants. The mean absolute error of the air quality index (AQI) provided by the solution is under 5%. The Breezometer application was used to evaluate participant's exposure to various pollutant types, summarized in Table 2.

TABLE 2.

Pollutants evaluated by the Breezometer.

Parameter	Description	Unit
AQI	Breezometer proprietary index which combines all solid and gaseous pollutants according to their hazardousness to health. Small values on this scale represent poor air quality	A.U. (1–100)
PM 10 µm	Air fine particle concentration smaller than 10 µm	µg/m3
PM 2.5 µm	Air fine particle concentration smaller than 2.5 µm	µg/m3
Pollen	Breezometer proprietary index built on health‐based research into the real impact of pollen on allergy sufferers.	A.U. (0–5)

Abbreviations: AQI, air quality index; A.U., arbitrary units; PM, particulate matter.

The AQI is a proprietary algorithm developed by Breezometer which is computed by combining the concentration of five main pollutants: Nitrogen dioxide (NO₂), Ozone (O₃), Sulfur dioxide (SO₂) and PM less than 10 µm (PM10) and less than 2.5 µm (PM2.5). The average of each pollutant type was calculated over a 24 h period. Each average concentration was converted into a sub‐AQI using the country official definition and the final AQI was determined by the worst sub‐AQI (being the most dominant) among them. It was expressed as a value ranging between 0 and 100.

The pollen Index is a proprietary algorithm which provides relative quantity of pollen (tree, grass or weed) in the air on a scale of 0 (None) to 5 (High).

The application we developed was set to collect the GPS information of the consumers every 5 min. However, data were only exported to an internal database when the application was opened.

2.4. Colorface measurements

Facial wrinkles, skin color, skin roughness, and pore measurements were obtained by image analysis. Standardized photos of the face were taken with the Colorface (Newtone Technologies) using the following lighting modalities: diffuse lighting, 45° directional lighting, 60° directional lighting, parallel polarized lighting, crossed polarized lighting, and blue fluorescent lighting. Images were analyzed independently by BrightTex Bio‐Photonics (https://www.btbp.org/index.html). Parameters related to the skin surface were analyzed from the parallel‐polarized images and those related to skin color from cross‐polarized images. Lab* measurements were calculated on the forehead, on the cheeks, and under the eye area. Wrinkles were segmented and classified into three categories: deep, emerging, lines and their average severity was calculated from the intensity difference (visibility) of each wrinkle detected. The following wrinkles were analyzed: crow's feet, glabellar, under the eye wrinkles. Pores were segmented on the cheeks and forehead and their average size was calculated. Skin roughness severity was obtained from the local variation of the pixel color intensity within the region of interest (cheeks, forehead, and under the eye).

2.5. Spectrophotometer measurements

The CM 700D (Konika Minolta) spectrophotometer was used to measure skin color and reflectance spectra in the range of 360–740 nm, as previously described. ⁷ , ⁸ , ⁹ Six main parameters of interest were examined: oxyhemoglobin, hemoglobin, melanin, measurements of the color space L*a*b*, to characterize skin color; where L* represents the skin's perceptual brightness (lightness), A* red/green vision, and B* yellow/blue vision. Each measurement was taken three times consecutively, and an average was calculated. Three measurements were taken at the right cheek and under eye area. Measurements sites are shown in Figure 1.

FIGURE 1.

Measurement sites for each parameter measured.

2.6. Cutometer measurements

A cutometer MPA 580 (Courage & Khazaka) with a probe with an aperture size of 2 mm was used to measure the biomechanical properties of the upper layers of the epidermis by applying negative pressure (suction) which deforms the skin. The measurement method was based on the suction method. Negative pressure (350 mbar) was created in the device and the time/strain mode was used with three consecutive cycles: the skin was sucked into the probe measuring cup during 5‐s, then the skin was released within the probe during 1‐s. Thus, this measurement provides characteristics such as firmness and elasticity of the skin. All measurements were taken three times at the mouth‐ear line (with a space of 1 finger after the nose). To avoid possible effects of the repetitive measurements, the probe was displaced for about 5 mm after each measurement. The following parameters were analyzed: R0 (skin firmness), R2, R5, R7 (skin elasticity). Measurements sites are shown in Figure 1.

2.7. Corneometer measurements

A corneometer CM 825 (Courage & Khazaka) was used to measure the level of moisturization of the superficial layers of the epidermis (stratum corneum) by measuring its electrical properties. The measurement method was based on determining the capacitance of the dielectric medium. Any change in moisturization at the surface of the skin lead to changes in these dielectric constants and, consequently, the capacitance measurement. The measurement was reported in arbitrary digital data units from 0 to 125. Three successive and adjacent measurements were taken below the line of the underside of the nose and the mean value was calculated.

2.8. Vapometer measurements

A Vapometer (Delfin Technologies), a closed unventilated chamber system, was used to measure transepidermal water loss (TEWL), which measures evaporation of water contained in the skin (g/m²h)). Three successive and adjacent measurements were taken at the right cheek and under eye area.

2.9. Sebumeter measurements

A sebumeter SM 815 (Courage & Khazaka) was used to measure the quantity of sebum at the skin's surface. The measurement was taken with a probe, applied to the skin and covered with a sebum‐sensitive clear film. The measurement was given in microgram per square centimeter and corresponds to the total quantity of sebum taken up on the film. Three measurements were taken on the left, middle, and right sections of the forehead.

2.10. Data analysis

Participants were stratified into two pollutant exposure groups (lower and higher exposure), using the median of the pollutant exposure distribution cumulated over 6 months as a threshold (Table 1). Distinct cohorts were established for each pollutant. Student T‐tests were performed where data had normal distribution and equal variance. If data were not normally distributed, Mann–Whitney U tests were performed. p < 0.05 was considered statistically significant.

3. RESULTS

3.1. Participants

A total of 156 Caucasian women living in urban areas of Paris, France, were recruited to the study, and 124 women completed the study. Pollution data collection was reviewed every month. Participant ages ranged between 20–60 years old. There was no significant difference in age between the lower exposure and higher exposure pollutants groups.

No statistical differences were shown for the time spent outdoor between the pollutant groups (see Table 1).

3.2. Effect of pollutants on wrinkles

The mean severity of under eye wrinkles was significantly higher in women with higher exposure to PM 10 µm levels than lower exposure (9302 ± 325 vs. 8401 ± 245, respectively; p < 0.05) (Figure 2A). The mean severity of deep under eye wrinkles was also significantly higher in women with higher PM 10 µm exposure than lower PM 10 µm exposure (10180 ± 283 vs. 9498 ± 193, respectively; p < 0.05) (Figure 2B). Women exposed to poorer air quality (lower AQI) exhibited significantly greater forehead wrinkle severity than those exposed to better air quality (5864 ± 189 vs. 5400 ± 136, respectively; p < 0.05) (Figure 2C).

FIGURE 2.

Effect of pollution on facial wrinkles in women living in Paris. (A) Under eye average wrinkle severity at lower versus higher PM 10 µm exposure, (B) Deep under eye average wrinkle severity at lower versus higher PM 10 µm exposure, and (C) Forehead wrinkle severity at lower versus higher AQI. Data represents mean ± standard deviation. ^* p < 0.05. y‐Axis units are arbitrary units. AQI, air quality index; PM, particulate matter.

3.3. Effect of pollutants on skin color

The skin color values measured from the images are shown in Figure 3. The b* value (yellowness) was significantly higher in the under eye area of women exposed to higher PM 2.5 µm than lower PM 2.5 µm (22.1 ± 0.4 vs. 21.1 ± 0.4, respectively; p < 0.05). Moreover, women with higher tree pollen exposure exhibited significantly greater cheek b* value than those with lower tree pollen exposure (21.4 ± 0.5 vs. 19.9 ± 0.4, respectively; p < 0.01). Conversely, women with higher tree pollen exposure exhibited lower under eye L* value (darkness) than those with lower exposure (51.7 ± 0.8 vs. 49.5 ± 1.0, respectively; p = 0.0506).

FIGURE 3.

Effect of pollution on the facial skin color of women living in Paris. (A) Under eye B* value at lower versus higher PM 2.5 µm exposure, (B) Cheek B* value at lower versus higher pollen exposure, and (C) Under eye L* value at lower versus higher pollen exposure. Data represents mean ± standard deviation. ^* p < 0.05; ^** p < 0.01. y‐Axis units are arbitrary units. PM, particulate matter.

There were not statistical differences regarding the spectrophotometer results.

3.4. Effect of pollutants on skin hydration and elasticity

Facial skin hydration, measured by the corneometer, was significantly lower in women exposed to more PM 10 µm than in women less exposed (36.3 ± 1.8 vs. 42.9 ± 1.5, respectively; p < 0.01) (Figure 4A). Moreover, women exposed to poorer air quality (lower AQI) exhibited significantly lower R2 parameter related to skin elasticity (0.57 ± 0.01 vs. 0.60 ± 0.01, respectively; p < 0.05) (Figure 4B).

FIGURE 4.

Effect of pollution on facial skin hydration and elasticity in women living in Paris. (A) Facial hydration at lower versus higher PM 10 µm exposure, and (B) facial elasticity at poorer versus better AQI. Data represents mean ± standard deviation. ^* p < 0.05; ^** p < 0.01. y‐Axis units are arbitrary units. PM, particulate matter; AQI, air quality index.

3.5. Effect of pollutants on skin texture (roughness)

The mean severity of under eye skin roughness was significantly higher in women exposed to more PM 2.5 µm than in women exposed to less PM 2.5 µm (902 ± 36 vs. 818 ± 35, respectively; p < 0.05) (Figure 5A). Similarly, the mean severity of under eye skin roughness was significantly higher in women exposed to more PM 10 µm than in women exposed to less PM 10 µm (921 ± 38 vs. 797 ± 32, respectively; p < 0.05) (Figure 5A).

FIGURE 5.

Effect of pollution on the facial skin texture of women living in Paris. (A) Under eye roughness at lower versus higher PM 2.5 µm and 10 µm exposure, (B) front forehead roughness at lower versus higher tree pollen exposure, and (C) Crow's feet roughness at lower versus higher tree pollen exposure. Data represents mean ± standard deviation. ^* p < 0.05; ^** p < 0.01. y‐Axis units are arbitrary units. PM, particulate matter.

Front forehead skin roughness (Figure 5B) significantly higher in women exposed to more tree pollen than those exposed to less (700 ± 31 vs. 600 ± 24, respectively; p < 0.01).

Crow's feet roughness (Figure 5C) were also significantly higher in women exposed to more tree pollen than those exposed to less (744 ± 28 vs 648 ± 19, respectively; p < 0.01).

3.6. Effect of pollutants on pores

The average pore surface area (Figure 6A) was significantly higher in women exposed to more PM 10 µm than in women exposed to less (218 ± 12 vs 176 ± 10, respectively; p < 0.01). The average pore diameter (Figure 6B) was also higher in women exposed to more PM 10 µm than in women exposed to less (0.558 ± 0.008 vs. 0.539 ± 0.006, respectively; p = 0.0503).

FIGURE 6.

Effect of pollution on facial pores in women living in Paris. (A) Pore surface area at lower versus higher PM 10 µm exposure, and (B) pore diameter at lower versus higher PM 10 µm exposure. Data represents mean ± standard deviation; ^** p < 0.01. PM, particulate matter.

3.7. Effect of pollutants on TEWL and sebum

No statistical differences were shown between pollutant groups on the TEWL and sebum measurements (data not shown).

4. DISCUSSION

This is the first observational study to examine the effect of air pollution on skin ageing in real‐time. This study addressed some of the methodological limitations of previous studies (which typically gathered general regional pollutant exposure data rather than specific individual pollutant exposure data) by collecting participants’ air pollutant exposure in real‐time via the Breezometer platform. Collectively, the findings of this study suggest a potential correlation between various air pollutants and the manifestation of skin aging characteristics, potentially associated with increased wrinkles, skin roughness, enlarged pores, and decreased skin hydration and elasticity. In this study, smaller PM (PM 2.5 µm) seemed to influence skin color and roughness under the eye. A decrease in overall air quality coincided with an increase forehead wrinkles and skin elasticity, while pollen was associated with skin roughness and crow's feet. Furthermore, higher exposure to large PM (PM 10 µm) was linked to increases in both wrinkles and roughness under the eye, pore appearance, along with a decrease of skin hydration. The average pollution exposure levels of women exposed to lower versus high PM 10 µm in this study were 23.7 and 35.6 µg/m³, respectively. The impact of PM 10 µm at a concentration of 30 µg/m³ was previously found to increase the expression of pro‐inflammatory genes and autophagy, which are indicators of skin inflammation and skin ageing in human dermal fibroblasts. ¹⁰

Previous studies have shown that prolonged exposure to high levels of air pollution can dysregulate skin homeostasis, thus accelerating extrinsic skin ageing. ⁵ , ¹¹ A study of 400 women by Vierkotter et al., ⁴ found elderly women living in rural areas with lower air pollution generally had fewer wrinkles and pigmented spots compared to those living in urban areas. Fuks et al., ¹² showed that ambient O₃ are positively associated with coarser facial wrinkles in 803 women and 1207 German men >60 years old, which is independent of NO₂, PM, and ultraviolet radiation. Furthermore, we previously demonstrated that air pollution negatively affects the skin color of Chinese women, whereby women living in a city with higher air pollution exhibited significantly different a* and b* values and melanin levels than those living in a city with lower air pollution. ⁹

Current scientific evidence suggests that there are three mechanisms by which ambient air pollutants promote skin ageing: free radical generation; inflammatory cascade activation; and impairment of the skin barrier. These processes are thought to produce direct and indirect toxicity to the skin (reviewed by Mancebo and coworkers ² , ⁵ , ¹³ ).

Limitations of this study which should be considered include: Other environmental (e.g., temperature and history of UV radiation exposure, or season), lifestyle factors (e.g., nutrition, sleep deprivation, sunscreen, or skincare routine), and stress may also contribute to the differences in skin parameters measured ¹⁴ ; 6 months may be too short to assess deeper skin condition alterations; indoor air quality or indoor pollution was not considered in this study; pollen analysis was not performed at peak pollen exposure periods throughout the year, and prior information on volunteers’ allergies was not collected.

In conclusion, this study suggests a negative impact of air pollution on the facial skin of Caucasian women in real‐life conditions. A perspective for further exploration involves extending the duration of pollution exposure measurement and incorporating cofactors in the statistical analysis. Additionally, the clinical significance of variations attributed to pollution exposure remains to be fully examined. The findings of this exploratory study highlight the need for dermatological products which either prevent or reduce pollution‐related skin damage.

CONFLICT OF INTEREST STATEMENT

All authors were employees of Laboratoires Clarins at the time of the study.

Robic J, Lata W, Nkengne A, Bigouret A, Vie K. The impact of air pollution on the facial skin of Caucasian women using real‐life pollutant exposure measurements. Skin Res Technol. 2024;30::e13669. 10.1111/srt.13669

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.