The association between bisphenol A (BPA) exposure and adult health status is examined by measuring the urinary BPA concentration using a miniaturized technique based on dispersive liquid-liquid microextraction (DLLME) in combination with gas chromatography-mass spectrometry (GC–MS). Both the free bioactive and the glucuronide conjugated forms of BPA were measured, the glucuronide form usually being predominant. The main analogs of BPA, including bisphenol Z (BPZ), bisphenol F (BPF) and biphenol (BP) were also determined. Several parameters affecting enzymatic hydrolysis, derivatization by in-situ acetylation and the DLLME stages were carefully optimized by means of multivariate designs. DLLME parameters were 2 mL urine, 1 mL acetone and 100 μL chloroform, and hydrolysis was performed using β-glucuronidase and sulfatase at pH 5. No matrix effect was observed and quantification was carried out by aqueous calibration with a surrogate standard. Detection limits were in the range 0.01–0.04 ng mL−1. The intraday and interday precisions were lower than 11% in terms of relative standard deviation. Satisfactory values for all compounds were obtained in recovery studies (92–117%) at two concentration levels. Other bisphenols (BPF, BPZ and BP) were not detected in the urine samples, while BPA was the only bisphenol detected in the free form (creatinine adjusted) at concentration levels ranging from the detection limit to 15.9 ng g−1, and total BPA was detected at concentrations ranging from 0.46 to 24.5 ng g−1 levels. A comparison of the BPA content for both groups of patients revealed that slightly higher mean values were obtained for both free BPA and total BPA for diabetic patients, than for non-diabetic patients. However, a statistical comparison of the contents of BPA revealed that there were no significant differences. The procedure was validated using a certified reference material.