Algorithm for environmental risk assessment of cosmetics to reduce their environmental impact

Abstract

Cosmetics, especially rinse-off personal care products (PCPs), such as shampoo, facial cleanser, and body wash, are composed of various chemicals and are one of the sources of chemicals released into aquatic ecosystems. Therefore, the cosmetic industry strives to reduce the impact of their products on the aquatic environment. In this study, we proposed an algorithm based on persistence, bioaccumulation potential, and toxicity (PBT) for the environmental risk assessment of cosmetics. PBT features are generally used in the evaluation of the environmental impact of chemicals. Based on the PBT assessment, it is possible to predict the short- and long-term effects of chemicals on the environment. Our algorithm derives substance and product scores from PBT features, allowing for the risk assessment of each ingredient in the product. Furthermore, we proposed a criterion for the environmental impact grade through which each component can be classified. We intend to use this grade and factors determined through the algorithm to manufacture products with low environmental impact.

Introduction

Numerous chemicals are released into aquatic ecosystems through various routes. Cosmetics, especially rinse-off personal care products (PCPs) which are composed of various chemicals such as organic compounds, botanical extracts, polymers, and fragrances, are one of the sources of chemicals released into aquatic ecosystems. They are released at low concentrations in vast areas over a long period of time [1]. The released chemicals could potentially affect the aquatic environment [2]. Recent studies show that oxybenzone and octinoxate, two of the ingredients in sunscreen, can cause coral reef bleaching. For this reason, the sale of sunscreen products containing these ingredients was banned in Hawaii [3]. In the European Union, Octamethylcyclotetrasiloxane (D4) and decamethylcyclopentasiloxane (D5) are classified as persistence, bioaccumulation potential and toxicity (PBT) or very persistent and very bioaccumulative (vPvB) substances, and their use is restricted [4]. However, there are few guidelines for reducing the environmental impact of PCPs. Therefore, the cosmetic industry strives to reduce the impact of their products on the aquatic environment [5].

Recently, the cosmetics industry is making efforts to reduce the environmental risks of not only packaging but also formulas. Typically, L’Oreal applies the gray water concept of the water footprint and evaluates the environmental risk of the formula using the two characteristics of biodegradability and environmental toxicity. And based on this, activities are being carried out to reduce the environmental risk of products [6]. In the case of Johnson & Johnson, an environmental risk assessment method called Global Aquatic Ingredient Assessment (GAIA) is applied through the concept of PBT, and a score is given to ingredients. It is also used as an index for better choices when developing new products [7].

In general, persistence, bioaccumulation potential, and toxicity are examined in the evaluation of the environmental impact of chemicals. Persistence is typically evaluated using half-lives in the environment (water, soil, and sediment) and biodegradability. Bioaccumulation potential is typically evaluated using octanol-water partition coefficients, and bioconcentration and bioaccumulation factors. Toxicity is typically evaluated on the basis of aquatic acute/chronic toxicity [8]. Based on the PBT assessment, it is possible to predict the short- and long-term effects of chemicals on the environment [9]. For this reason, many countries evaluate chemicals on the basis of PBT, for example, the criteria and test strategy for chemicals of the European Chemical Agency (ECHA) in the European Union, and the PBT profiler of the US Environmental Protection Agency (US EPA). Depending on their assessment, the use of certain chemicals is restricted or prohibited [10, 11]. In addition, chemical evaluation using this PBT assessment has been studied since the past, and the representative example is The Scoring and Ranking Assessment Model (SCRAM) [12]. In particular, the ingredients of PCPs have been subjected to numerous PBT tests. PBT can now be easily predicted using an in silico model based on the structure of chemicals. PBT assessment is suitable for the evaluation of the environmental risk of PCPs and their ingredients [7]. In this study, we propose an algorithm based on PBT. This algorithm uses PBT to first derive a substance score, which is then used to calculate the product score. Environmental risk assessment can be performed for each ingredient in the product. Furthermore, environmental impact grades A–E are assigned to the substances according to substance scores based on PBT related data to make determination of environmental safety easier during product development. The product score obtained through the algorithm indicates the product’s extent of improvement compared to other products. It is also possible to determine the ingredient with a significant impact on the environment, and reduce or replace the ingredient to lessen the environmental impact of the product.

Materials and methods

Algorithm design

In this study, a PBT-based algorithm (Scheme 1) is developed to predict and assess the environmental risk of PCP ingredients. As PBT information is generally regarded as the intrinsic property of a substance, it does not vary with product type and manufacturing process.

Scheme 1.

Algorithm process

The content (%) of each ingredient in the product is also an essential factor. Exposure to very low levels of a high-risk substance is expected to represent a relatively low risk, whereas even a low-hazard substance may present a high risk if the exposure is significant. Therefore, the intention was to derive a comprehensive PBT evaluation score by evaluating the unique characteristics of the materials contained in the product and then applying the content (%) of the ingredients in the product to the substance score to derive environmental hazard scores for the final product. The algorithm process is shown in Scheme 1.

Selection of substance groups

The ingredients of the PCP selected for the study included organic, inorganic, and botanical extracts. PBT data is unavailable for most natural materials. They are therefore evaluated as worst-case scenarios when applied to the algorithm developed in this study. Thus, specific scores were assigned to botanical extracts to allow for the derivation of final scores in reasonable scenarios, even without PBT data.

Algorithms were designed to classify groups of the substance by first examining the materials contained in the product, and the groups of substances were categorized as water, organic, inorganic, and botanical extract.

Scoring concept

Score is between 0 and 100 that represents a substance’s environmental persistence in water, potential to bioaccumulate in food chains, and its direct toxicity to aquatic organisms (PBT characteristics). The higher the score, the more favorable the environmental safety characteristics of the substance (Scheme 1).

Persistence

To evaluate persistence, we examined the biodegradation data of ingredients. Through persistence assessment, chemicals can be classified into four groups [readily biodegradable (score 100); readily biodegradable, but failing 10-day window (score 30); inherently biodegradable (score 10); and not biodegradable (score 0)]. This score was assigned according to the degradation rate for each biodegradation group [13]. In the event that biodegradability cannot be determined owing to a lack of relevant data, 0 is assigned to the persistence assessment from a conservative perspective.

For substance group of the botanical extract without biodegradability data, a persistence evaluation score of 10 or 30 is assigned assuming a reasonable scenario. In the case of botanical extracts used in cosmetics, water and ethanol are mostly used as solvents. Water and ethanol are readily degradable. However, in the extraction process, a small amount of unintended substances may remain, so conservatively, a score corresponding to readily biodegradable, but failing 10-day window or inherently biodegradable was given.

Plant extracts can be used in cosmetics to beautify the human skin and maintain its physiological balance. Compared to synthetic cosmetic substances, herbal (plant) substances are milder and biodegradable, thus exhibiting low toxicity [14].

Bioaccumulation

Bioaccumulation data of each substance were examined. Referring to the EU PBT Guideline and the NIER Notification 2020-8 “Regulations on Classification, Labelling, etc. of Chemicals”, chemicals were classified according to Bioconcentration Factor(BCF) values of 5000 L/kg, 2000 L/kg, and 500 L/kg [15, 16].

Score 100 was assigned for the substance having BCF value < 500 L/kg, score 60 was assigned for the substance having BCF value ≥ 500 L/kg and < 2000 L/kg, score 40 was assigned for the substance having BCF value ≥ 2000 L/kg and < 5000 L/kg, and score 0 was assigned for the substance having BCF value > 5000 L/kg. In addition, in the absence of data, a score of 0 was assigned to the bioaccumulation assessment from the conservative perspective.

Toxicity

Calculation of predicted no effect concentration

In the aspect of water environmental toxicity of substances, the value of predicted no effect concentration (PNEC), which is the concentration of chemicals in each medium that is not assumed to affect the aquatic ecosystem and organisms of the habitat, is calculated. In the absence of environmental toxicity data, the ecotoxicological Threshold of Toxicological Concern (Eco TTC) concept was used. The Eco-TTC is similar to the TTC for human. The Eco-TTC applied in this study classify the substances into 6 classes according to the mode of action (MOA) framework (Verhaar scheme) based on the chemical structure and physicochemical properties of the target material. PNEC is applied according to the classification [17].

To derive PNEC values, the lowest toxicity value on organisms in the media is required and is expressed as EC₅₀ or LC₅₀ for acute toxicity and NOEC or LOEC for chronic toxicity. In the case of the lowest effect concentration, the assessment factor (AF) used to describe uncertainty and variability in the assessment process of the environment is selected. In the absence of the separate existence of results of the study on uncertainty, the Registration, Evaluation, Authorization and Restriction of Chemicals (REACH) assessment factor of the EU is used.

AF depends on whether toxicity data are acute or chronic, or how much data is available at the trophic level (Table 1). In general, three trophic levels(fish, invertebrates (preferred Daphnia) and algae) are used.

Table 1.

Assessment factor (AF) according to available toxicity data on aquatic organisms (EU) [18]

Data combination	AF
1 acute trophic levela	10,000
2 acute trophic level; use the most sensitive taxon	5000
3 acute trophic level; use the most sensitive taxon	1000
3 acute trophic level; 1 chronic organism available (fish or invertebrate) but not on most sensitive acute taxon	1000
3 acute trophic levels; 1 chronic organism available (fish or invertebrate) which is also most sensitive acute taxon	100
3 acute trophic levels; 2 chronic organisms available including most sensitive acute taxon	50
3 acute trophic levels; 3 chronic trophic levels including most sensitive taxon	10

^aThree trophic levels : fish, invertebrates (preferred Daphnia) and algae

PNEC_water is elicited using the following equation.

$$\text{PNECwater}=\frac{\text{Lowest LC50 or NOEC}}{\text{AF}}$$

• LC₅₀ or NOEC: Most sensitive impact concentration in the media.

• AF: Assessment Factor.

Scoring of toxicity

To evaluate toxicity, the toxicity data for each substance in the product were reviewed, and then the PNEC value or threshold value was derived. For the scoring of toxicity on aquatic environment media, the aquatic environment hazard (acute, chronic) classification criteria of the NIER Notification 2020-8 “Regulations on Classification, Labelling, etc. of Chemicals” were referenced [16].

The PNEC_water values can be calculated based on the toxicity values applied to the GHS classification ranging from 0.00001 to 2 mg/L. Toxicity scores were derived using each calculated PNEC value. This equation is shown in Fig. 1.

Fig. 1.

Equation elicited for scoring toxicity. The equation is derived from the relation between the value of PNEC water and the toxicity score. The values of PNEC water in the range of 0.00001–2 mg/L derived from the GHS classification criteria were assigned to an arbitrary score from 25 to 80. A The graph between the value of PNEC water and the toxicity score. B The graph of the value PNEC water in log scale and the arbitrary score

Calculation of substance score

Substance score

The overall PBT score was calculated using the developed algorithm. The substance groups were classified as water, organic, inorganic, and botanical extract, and the total score was derived by applying different weight to each item according to the substance group.

Organic chemicals were given high weight because organics are persistent and have long-term impacts on the environment if they do not break down to simple compounds found in nature. Inorganic chemicals were given higher weight to toxicity because inorganic compounds (e.g., metals) are often used in forms that are already simple and naturally occurring [7]. The extract is also a substance derived from nature, and most of the substances used as solvents are biodegradable substances, and a high weight is given to toxicity. However, since there is uncertainty about the solvent, a slightly higher weight was given to persistence than inorganic chemicals. The substance score (Score(S))1 extraction method is shown in Table 2.

Table 2.

Calculation of substance score with different weight according to chemical categories

Score(S) calculation equation		Note
Organic : 0.50(Persistence) + 0.25(Bioaccumulation) + 0.25(Toxicity)		WATER score is calculated as 100
Inorganic : 0.25(Persistence) + 0.15(Bioaccumulation) + 0.60(Toxicity)
Botanical extract : 0.30(Persistence) + 0.20(Bioaccumulation) + 0.50(Toxicity)
The rest : 0.50(Persistence) + 0.25(Bioaccumulation) + 0.25(Toxicity)

Selection of relatively hazardous substances in products

After deriving the Score(S) via the method described above, the substance is examined for hazard relative to other substances in the product. The average Score(S) of each product category (shampoo, conditioner, foam cleansing, etc.) is derived, and compared with each Score(S) contained in the product. In this case, if the individual Score(S) is less than the product category average Score(S), the substance is screened as a relatively hazardous substance within the product.

$$\text{Hazard analysis}:\text{Individual Score}\left(\text{S}\right)-\text{Product category average Score}\left(\text{S}\right)$$

• Hazard analysis > 0: Relatively less hazardous substance.

• Hazard analysis ≤ 0: Relatively more hazardous substance.

Calculation of product score

The product score (Score(P))2 is derived to determine how the target substance affects the product. Score(P) is calculated by considering the components of content and use factor to the Score(S) derived using PBT.

Among PCPs, highly concentrated shampoo and body wash are often developed to reduce the use of water and packaging materials. Therefore, it is necessary to make corrections for the use factor to compare Score(P) of general and highly concentrated shampoos. For example, if the factor for general shampoo is 1.00, factor for highly concentrated shampoo can be given a factor of 0.50 in consideration of the amount used.

Overall, the substance score in product (Score(S in P))3 can be explained by the concept of [Score(S) × (content(%) × use factor). The Score(P) is derived from the sum of each Score(S in P) and is comparable within the same product category. The closer the Score(P) is to 100, the less hazardous the product is. Score(P) is calculated using the following equation.

$$\text{Score}\left(\text{S in P}\right)=\text{Score}\left(\text{S}\right)\times \text{Content}\left(\text{\%}\right)\times \text{Use factor}$$

$$\text{Score}\left(\text{P}\right)=\sum [\text{Score}\left(\text{S in P}\right)]$$

• Use factor can be changed by the product. (e.g., based on general shampoo 1.00, highly concentrated shampoo 0.50, etc.)

Risk of substance in product

This algorithm did not calculate the amount of exposure resulting from the use of the product but rather calculated the risk by replacing the value of the substance’s content with the exposure amount.

In chemical environmental risk assessment, the risk quotient (RQ) is generally calculated using the exposure/toxicity method. By applying this concept, the substance_RQ (RQ(S)) was calculated using the content (exposure) / Score(S) in this algorithm. The content used when calculating RQ is relative content to total ingredients amount excluding water in the product.

The calculated RQ(S) values were summed to derive the overall product_RQ (RQ(P)). The substance risk contribution was calculated by comparing the RQ(S) values of each substance with the RQ(P) values of the entire product (RQ(S)/RQ(P)), to determine the ratio (%) of risk of each substance in the product.

If RQ(S) contribution is greater than the average RQ(S) contribution, it is evaluated as “substance subject to reduction”.

$$\begin{array}{cc}\hfill \text{RQ}\left(\text{S}\right)& =\text{Relative content}÷\text{Score}\left(\text{S}\right)\hfill \\ \hfill \text{RQ}\left(\text{P}\right)& =\sum \left[\text{RQ}\left(\text{S}\right)\right]\hfill \\ \hfill \text{RQ}\left(\text{S}\right)\text{contribution of the entire product}& =\left[\text{RQ}\left(\text{S}\right)÷\text{RQ}\left(\text{P}\right)\right]\times 100\hfill \\ \hfill \end{array}$$

• RQ(S) contribution > Average RQ(S) contribution : “substance subject to reduction”.

Algorithm interpretation

When the individual Score(S) was less than the average Score(S), the substance could be considered as a relatively hazardous substance within the product. In addition, whenever an RQ(S) contribution exceeded the average RQ(S) contribution of the entire product, it was assessed as a substance subject to reduction.

Substances that corresponded to both “relatively hazardous substance” and “substance subject to reduction " were preferentially managed, that is, the content of the substance was adjusted in the relevant product, or, if test data were lacking, the Score(S) was adjusted through additional tests such as biodegradation test.

Results

Examples and case study of substance and product assessment

To examine the validity of the algorithm for product evaluation, the actual product was selected. The components in the shampoo representing function such as surfactant, humectant, emollient, preservative, pH adjuster, anti-oxidant, and oil were applied to the algorithm (Table 3).

Table 3.

Target Product’s composition

Substance	CASRN	Group	Content
Sodium laureth sulfate	68891-38-3	Organic	15.0
Polyquaternium-10	68610-92-4	Organic	0.5
Glycerin	56-81-5	Organic	10.0
Caprylic/Capric Triglyceride	73398-61-5	Organic	1.0
Hydrogenated coconut oil	84836-98-6	Botanical extract	1.0
Trisodium EDTA	150-38-9	Organic	0.4
Sodium chloride	7647-14-5	Inorganic	1.0
Citric acid	77-92-9	Organic	0.1
Cocamide methyl MEA	866889-75-0	Organic	3.0
Water	7732-18-5	Water	68.0

The PBT scores for the target product (shampoo) are shown in Tables 4 and 5.

Table 4.

Substance PBT information in the target product

Substance	Biodegradation	BCFa	Toxicity	Type/Eco-TTCb	Assessment factor	PNECc/TTC
Sodium laureth sulfate	Readily biodegradable	–	0.14 mg/L	Chronic	10	0.014 mg/L
Polyquaternium-10	–	–	2.4 mg/L	Acute	5000	0.00048 mg/L
Glycerin	Readily biodegradable	3.162	885 mg/L	Acute	1000	0.885 mg/L
Caprylic/Capric Triglyceride	Readily biodegradable	60.32	0.01 mg/L	Chronic	50	0.0002 mg/L
Hydrogenated coconut oil	Readily biodegradable (failing 10 day window)	–		Class 5		0.000044 mg/L
Trisodium EDTA	Inherently biodegradable	3.162	25 mg/L	Chronic	100	0.25 mg/L
Sodium chloride	–	3.162	252 mg/L	Chronic	50	5.04 mg/L
Citric acid	Readily biodegradable	3.162	440 mg/L	Acute	5000	0.088 mg/L
Cocamide methyl MEA	–	–	33.4 mg/L	Acute	10,000	0.00334 mg/L
Water	–	–	–	–	–	–

^aBioconcentration factor

^bThreshold of Toxicological Concern

^cThe value of predicted no effect concentration

Table 5.

Score calculation in the target product

Substance	P-Score	B-score	T-score	Score(S)
Sodium laureth sulfate	100	0	68	67
Polyquaternium-10	0	0	57	15
Glycerin	100	100	76	94
Caprylic/Capric triglyceride	100	100	60	90
Hydrogenated coconut oil	30	0	57	58
Trisodium EDTA	10	100	72	48
Sodium chloride	0	100	79	63
Citric acid	100	100	72	93
Cocamide methyl MEA	0	0	65	16
Water	–	–	–	100
Product category average score (S)				60.56

Score(S) derived through PBT data was compared to product category average Score(S). The algorithm evaluates water as Score(S) 100. In other words, the closer the Score(S) is to 100, the less hazardous the component is expected to be. Products included shampoo, conditioner, foam cleansing, and body cleanser. The average Score(S) for each product category can be used as the reference value for the evaluation of the Score(S) according to component. A substance whose Score(S) is lower than the average Score(S) for each product category can be regarded as being relatively hazardous within the product.

In this assessment, for example, the average Score(S) for the target product (shampoo) category was set as 60.56, and a component whose Score(S) was less than 60.56 was considered to be relatively hazardous (polyquaternium-10, hydrogenated coconut oil, trisodium EDTA, cocamide methyl MEA). The hazardous substances in the target product (shampoo) are presented in Table 6.

Table 6.

Selection of relatively hazardous substances in the target product

Substance	Score(S)	Gap with Product category average score (S)	Result
Sodium laureth sulfate	67	6.44	–
Polyquaternium-10	15	–45.56	Relatively hazardous
Glycerin	94	33.44	–
Caprylic/Capric Triglyceride	90	29.44	–
Hydrogenated coconut oil	38	– 2.56	Relatively hazardous
Trisodium EDTA	48	– 12.56	Relatively hazardous
Sodium chloride	63	2.44	–
Citric acid	93	32.44	–
Cocamide methyl MEA	16	– 41.07	Relatively hazardous
Water	100	–	–
Product category average score (S)	60.56

After the hazard assessment of the substance, the environmental exposure effect of product use was also investigated. Assuming that the consumption rate of each product (category) is the same, the content (%)(except water) of the components was replaced by the exposure quantity, and the risk of each component was calculated by dividing the content (%) by the previously derived Score(S). After comparing with the average risk value of each component, if the value was large, it was evaluated as an adjustment target. In the product, sodium laureth sulfate, glycerin and cocamide methyl MEA were substances subject to reduction because their RQ(S) contributions are larger than average of RQ(S) contribution. An example of risk contribution in the product (shampoo) is shown in Table 7.

Table 7.

Calculation of risk contribution in the target product

Substance	Relative content (%) (except water)	Score(S)	RQ(S)	RQ(S) contribution	Result
Sodium laureth sulfate	46.885	67	0.70	36.48	Substance subject to reduction
Polyquaternium-10	1.5625	15	0.10	5.43	–
Glycerin	31.25	94	0.33	17.33	Substance subject to reduction
Caprylic/Capric triglyceride	3.125	90	0.03	1.81	–
Hydrogenated coconut oil	3.125	38	0.08	4.29	–
Trisodium EDTA	1.25	48	0.03	1.36	–
Sodium chloride	3.125	63	0.05	2.59	–
Citric acid	0.3125	93	0.003	0.18	–
Cocamide methyl MEA	9.375	16	0.59	30.54	Substance subject to reduction
RQ(P)			1.92
Average of RQ(S) contribution:				11.11

Cocamide methyl MEA were categorized under “relatively hazardous” and “substance subject to reduction”. Therefore, this ingredient should be managed first. For this ingredient, the content of the ingredient in the product should be reduced, or if the test data is insufficient, the Score(S) should be adjusted through an additional test such as biodegradation test.

Score(P) was derived for product-specific comparisons within the same product category. The points were derived by applying the content and use factor to the Score(S) derived earlier, and the product scores were compared by combining the Score(S in P) of the materials in the product. For example, if the evaluation product (shampoo-1) consisted of only ten substances (including WATER) as presented in Table 8, the total score of the product would be 90.20. If another product was assessed in the same manner and yielded a total score below 90.20, shampoo-1 was considered to be less hazardous to the environment than another product.

Table 8.

Score(P) calculation for product(Shampoo-1) and improved product(Shampoo-2)

Substance	Content (%)	Use factor	Score(S)	Score(S in P)
Shampoo-1
Sodium laureth sulfate	15.0	1.00	67	10.05
Polyquaternium-10	0.5	1.00	15	0.08
Glycerin	10.0	1.00	94	9.40
Caprylic/Capric Triglyceride	1.0	1.00	90	0.90
Hydrogenated coconut oil	1.0	1.00	38	0.38
Trisodium EDTA	0.4	1.00	48	0.19
Sodium chloride	1.0	1.00	63	0.63
Citric acid	0.1	1.00	93	0.09
Cocamide methyl MEA	3.0	1.00	16	0.48
Water	68.0	1.00	100	68.00
Score(P)				90.20
Shampoo-2
Sodium laureth sulfate	13.0	1.00	67	8.71
Polyquaternium-10	0.5	1.00	15	0.08
Glycerin	10.0	1.00	94	9.40
Caprylic/Capric triglyceride	1.0	1.00	90	0.90
Hydrogenated coconut oil	1.0	1.00	40	0.38
Trisodium EDTA	0.4	1.00	48	0.19
Sodium chloride	1.0	1.00	63	0.63
Citric acid	0.1	1.00	93	0.09
CocamideMEA	3.0	1.00	92	2.76
Water	70.0	1.00	100	70.00
Score(P)				93.14

And an example of the improved product (Shampoo-2) based on the previous evaluation is shown in Table 8. Cocamide methyl MEA classified as “relatively hazardous” and “substance subject to reduction” was replaced with Cocamide MEA having the same function. Cocamide MEA has biodegradation and BCF data, and Score(S) is 92. In addition, sodium laureth sulfate containing a high content was reduced to 13% as an ingredient classified as “substance subject to reduction”. As a result, Score(P) was improved to 93.14.

Grading of ingredients

In this study, the EU REACH’s PBT and vPvB criteria and GHS classification were applied to create the grading criteria. Then, the range of Score(S) corresponding to each criterion was derived through the algorithm described in the Score(S) calculation method (2.3, 2.4). Based on this, the ingredients were classified as grades A to E, and the criteria and definitions of each grade are listed in Table 9. Ingredients with no data were classified as grade C.

Table 9.

Criteria and definition of ingredient grade

Criteria	Score(S)	Grade	Definition
Readily biodegradable BCF < 500 NOEC > 1/LC(EC)50>100	92 ≤ score ≤ 100	A	Environmental risk-free ingredient
Readily biodegradable (failing 10-day window) BCF < 2000 or No data (only botanical extracts)	47 < score < 92	B	Ingredient with low environmental impact
Readily biodegradable (failing 10-day window) 500 ≤ BCF < 2000 NOEC ≤ 1/LC(EC)50≤100 or No data (excluding botanical extracts)	32 < score ≤ 47	C	Chronic toxicity potential ingredient
Inherently biodegradable 2000 ≤ BCF < 5000 NOEC ≤ 0.1/LC(EC)50≤1	17 < score ≤ 32	D	PBT potential ingredient
Non-biodegradable 5000 ≤ BCF	0 < score ≤ 17	E	vPvB potential ingredient

The use of grade A and B ingredients is recommended, as they are less likely to cause negative environmental impacts. Ingredients categorized under classes D and E that could potentially be PBT or vPvB substances should be progressively restricted or banned.

Discussion

This study developed an algorithm to assess the effect of various chemicals used in cosmetics on the environment and devise countermeasures for the use of substances that are highly hazardous to the environment. The inherent characteristics of the substances, Persistence (P), Bioaccumulation (B), and Toxicity (T) data applied to all substances, at the same time, were used to calculate the basic score (Score(S)) and to apply concept of exposure to conduct environmental risk assessment of the substance. The use of well-established PBT data and logical algorithms helps to achieve consistent endpoints, even when evaluated by other assessors or non-toxicologists.

This algorithm performs an environmental risk assessment of products and ingredients based on data. In the absence of data, the most conservative score is given. Therefore, if there are ingredients without data, product evaluation results will be inaccurate. In addition, in the case of an ingredient with insufficient data, there is a possibility of being classified into a lower grade. For more accurate evaluation, it is necessary to test or collect a lot of PBT data through an open database. In addition, exposure to soil and air should be considered in addition to exposure to water of cosmetics when conducting environmental risk assessment.

The main difference between this algorithm and other similar environmental risk assessment methods is that it can calculate both absolute and relative risk for ingredients in a product. The algorithm can be used to calculate the environmental risk contribution of each ingredient in a product, making it possible to identify ingredients that should be reduced or replaced in products. Furthermore, a relative risk assessment allows us to continuously improve the environmental impact of our products. Second, by assigning a grade to each ingredient, even non-specialists can easily understand the environmental impact of the target ingredient. This will be of great help in selecting ingredients with low environmental impact when developing products. Product improvement activities based on the study findings will contribute to the reduction of the environmental impact of cosmetic products. We hope this algorithm could be used towards manufacturing greener consumer products.

Funding

The authors have not disclosed any funding.

Conflict of interest

The authors have no conflict of interest to disclose.