Table 3 Narrative synthesis of the association between PFAS mixtures and birth weight with different mixture methods.
Reference | Modeling strategy | Mixture methods | Other componentsa | Association direction |
|---|---|---|---|---|
Single methods | ||||
Govarts 2016 [58] | Dimension reduction | PCA | OCP, PCB, heavy metals, phthalates, particulate matter | Negative |
Gao 2022 [83] | Index modeling | WQS | No | Negative |
Padula 2023 [84] | Bayesian weighted sum | No | Negative | |
Svensson 2021 [61] | WQS | OCP, phthalates, plasticizer, antibacterial, phenols, PAH, pesticides | Negative | |
Wang 2023 [85] | WQS | No | Null | |
Shen 2022 [86] | Quantile g-computation | No | Null | |
Song 2023 [87] | Quantile g-computation | No | Negative | |
Hu 2021 [88] | Response-surface modeling | BKMR | OCP, PCB, heavy metals, phthalates, phenols | Negative |
Wang 2023 [89] | BKMR | No | Positive | |
Lazarevic 2022 [39] | Bayesian structured GAM w/ spike–slab | No | Null | |
Luo 2021 [90] | BKMR | No | Negative | |
Pearce 2021 [40] | Exposure continuum mapping (ECM) | PBDE, OCP, PCB | Positive | |
Multiple methods | ||||
Eick 2022 [69] | Response-surface modeling + index modeling | BKMR + Quantile g-computation | perceived stress and depression | Negative |
Eick 2023 [68] | BKMR + Quantile g-computation | perceived stress, depression, gendered racial stress and anxiety | Negative | |
Marks 2021 [91] | BKMR + WQS | OCP, PCB | Negative | |
Kalloo 2020 [92] | Dimension reduction | PCA + K-means clustering | PBDE, OCP, PCB, phenols, parabens, phthalates, diammonium phosphate, pyrethroid insecticides, cotinine | Null with one PC negative |
Zhuang 2021 [93] | Response-surface modeling + dimension reduction | BKMR + Bayesian factor analysis | PCB | Negative |
Eick 2023 [94] | Response-surface modeling + index modeling + dimension reduction | BKMR + Quantile g-computation + PCA | PBDE, OCP, bisphenols, pyrethroid insecticides, phthalates and marijuana and tobacco | Negative |
