Relation of urinary bisphenol concentration and diabetes or prediabetes in French adults: A cross-sectional study

**Background:** Bisphenol A (BPA) is a known endocrine-disrupting chemical widely found in food packaging, thermal paper, and other consumer products. Experimental studies suggest BPA may contribute to insulin resistance, adipogenesis, and pancreatic β-cell dysfunction. While meta-analyses have linked BPA exposure to diabetes risk, fewer data exist for bisphenol S (BPS) and bisphenol F (BPF), which are increasingly used as BPA substitutes. This study aimed to examine associations between urinary concentrations of BPA, BPS, and BPF and the prevalence of diabetes or prediabetes in a nationally representative sample of French adults.

**Methods:** Data were drawn from the Esteban cross-sectional study conducted in mainland France between April 2014 and March 2016. The analytic sample included 852 adults aged 18–74 years with complete questionnaire, medical examination, fasting serum glucose, and urinary bisphenol measurements. Diabetes and prediabetes were defined using a combination of self-reported diagnosis, diabetes medication reimbursement (ATC codes beginning with A10) from the National Health Data System (SNDS), and fasting serum glucose (≥7.0 mmol/L for undiagnosed diabetes; 6.1–7.0 mmol/L for prediabetes, per WHO criteria). Urinary total bisphenols were quantified by gas chromatography-tandem mass spectrometry. BPA, BPS, and BPF were detected and quantified in 100% of samples. Logistic regression models were used to estimate odds ratios (OR) for an increase of 0.1 units in log-transformed bisphenol concentration (μg/L). Models were adjusted for age, sex, education level, body mass index (BMI), hypertension, hypercholesterolemia, sedentariness level, smoking duration, daily energy intake, and urinary creatinine concentration.

**Key Results:** The prevalence of diabetes or prediabetes in the sample was 17.8% (95% CI: 15.3–20.4). In bivariate analyses, participants with diabetes or prediabetes were significantly older (mean 58.0 vs. 50.2 years, p<0.0001), more likely to be male (65.8% vs. 40.7%, p<0.0001), had lower education levels (38.2% vs. 24.9% less than high school, p<0.001), higher BMI (35.1% vs. 11.0% obese, p<0.0001), higher prevalence of hypertension (61.8% vs. 26.1%, p<0.0001) and hypercholesterolemia (38.8% vs. 25.9%, p=0.004), and were more likely to be long-term smokers (29.1% vs. 15.2% smoking >20 years, p<0.001). In crude logistic regression, the OR for diabetes or prediabetes per 0.1-unit increase in log-transformed urinary BPA was 1.08 (95% CI: 1.03–1.13, p<0.001). After full adjustment, the association remained significant (OR=1.12; 95% CI: 1.05–1.19, p<0.001). No significant associations were found for BPS (adjusted OR=0.99; 95% CI: 0.96–1.02, p=0.44) or BPF (adjusted OR=1.04; 95% CI: 0.98–1.11, p=0.20).

**Clinical Implications:** This study adds to the growing evidence that BPA exposure is independently associated with dysglycemia, even after rigorous adjustment for established diabetes risk factors including diet. The null findings for BPS and BPF are notable given their increasing use as BPA substitutes, though the authors caution that data were collected before widespread substitution (2014–2016). Key limitations include the cross-sectional design (precluding causal inference), single-spot urine measurement (which may not capture chronic exposure), unavailability of family history of diabetes, and lack of a second confirmatory glucose test or HbA1c. The study's strengths include the use of multiple data sources to identify diabetes and prediabetes, adjustment for dietary intake (rarely included in prior studies), and standardized biomonitoring procedures. The authors conclude that while BPA regulation in France (banned in food contact materials since 2015) appears justified, prospective longitudinal studies with repeated exposure measurements are necessary to establish causality and to monitor the health effects of BPA substitutes like BPS and BPF.