Integrating target, nontarget analysis with machine learning to illuminate PFAS characteristics and health risks in Chinese cosmetics.

Summary

Using an integrated target/nontarget workflow with machine learning, the authors characterized PFAS in 31 Chinese cosmetics and found higher PFAS occurrence in “waterproof/long-lasting” products. Risk assessment suggested dermal exposure from two products could exceed acceptable daily intake thresholds, supporting stronger regulatory controls and disclosure.

Key Findings

• Target analysis detected 10 PFAS in 20/31 cosmetics (0.189–143 ng/g total).
• Nontarget screening identified 15 PFAS in 30/31 cosmetics (4.72–263 ng/g total).
• “Waterproof/sweatproof/long-lasting” products were more likely to contain PFAS.
• Dermal risk assessment indicated two products (a lotion and a sunscreen) could exceed acceptable daily intake.

Clinical Implications

Dermatologists and pharmacists can counsel patients to avoid “waterproof/long-lasting” products when concerned about PFAS exposure; regulators can set category-specific PFAS limits and mandate disclosure/testing.

Limitations

• Modest sample size (31 products) from a single national market may limit generalizability
• Dermal exposure modeling assumptions may not capture all real-world use scenarios

Future Directions

Expand surveillance to larger, multi-country product panels; validate exposure models with biomonitoring; establish enforceable PFAS thresholds and labeling requirements.