Code and aggregate results for urinary DEHP metabolites and fasting insulin, glucose and C-peptide in NHANES

Abstract

Version 1.0.0 code and aggregate-results archive supporting complementary NHANES analyses of urinary di(2-ethylhexyl) phthalate metabolites and fasting insulin, glucose, HOMA-IR and C-peptide in adults without diagnosed diabetes. The principal cross-sectional analysis uses NHANES 2005–2018; the complementary historical analysis uses NHANES 2003–2004. These are different cycles of the same survey, not an external validation cohort. The historical C-peptide result is exploratory and does not meet the prespecified within-revision multiple-testing criterion. The coefficient decomposition is algebraic and does not establish a biological mediation mechanism or a causal exposure effect. The revision protocol was recorded after earlier analyses and is not a prospective preregistration. This archive contains analysis and verification code, source-file retrieval information, aggregate numerical results, figure-source data and documentation. Individual participant records, raw NHANES XPT files, participant-level RDS files, internal submission records and the article manuscript are excluded. Rebuilding participant-level analysis inputs requires obtaining the public-use source files from CDC/NCHS. README and verification documentation define the exact tested scope and remaining reproducibility limitations; no new end-to-end cold-download or independent laboratory replication is claimed. Licensing is component-specific: original code is MIT; original aggregate results, figures and documentation are CC BY 4.0. Third-party software, source facts and CDC materials retain their original terms and are not relicensed. Consult the package license and rights documentation before reuse. Creators, in order: Jie Li, Xiubo Sun, Jing Zhang, Lijie Zhai and Ling Yu. Corresponding author: Ling Yu, Department of Pharmacy, The Second Hospital of Jilin University. Funding: Jilin Province Tianhua Health Public Welfare Foundation Project, grant J2025JKJ012 to Ling Yu. AI assistance through Codex was used for source review, code development, statistical checking and drafting during the revision. The authors reviewed and approved the revised work and take responsibility for its accuracy and interpretation. This is a code/results archive, not an article, preprint or preregistration.