2.1. Study area and sample Dhaka, with a metropolitan area of 815.8 km2, is surrounded by the three major rivers Turag, Buriganga and Shitalakha, which are currently used for industrial waste disposal. Dhaka is one ofthe most densely populated cities in the world with 12 M people, of which fewer than 25%are served by sewage treatmentfacilities (Islamet al., 2014c). The most consumed foods for Bangladeshi people, i.e. cereals, pulses, vegetables, fruits, fish, meat, eggs and milk, were collected during February–March, 2012 and August–September, 2013. About 173 food samples were collected from three rivers and their adjacent areas around the Dhaka city metropolitan area, Bangladesh. A composite sample for each food item was prepared and homogenized in a food processor, and 50 g test portions were stored at 20 0C in the laboratory of the Institute of Nutrition and food Science (INFS), University of Dhaka, Bangladesh. The prepared samples were brought to Yokohama National University, Japan, for chemical analysis. 2.2. Sample analysis All chemicals were analytical grade reagents; Milli-Q water (Elix UV5 and MilliQ, Millipore, Boston, MA, USA) was used for the preparation of solutions. The Teflon vessel and polypropylene containers were cleaned, soaked in 5% HNO3 for more than 24 h, then rinsed with Milli-Q water and dried. For metal analysis, 0.3– 0.5 g of the food sample was treated with 6 mL 69% HNO3 (Kanto Chemical Co, Tokyo, Japan) and 2 mL 30% H2O2 (Wako Chemical Co, Tokyo, Japan) in a closed Teflon vessel and was digested in a Microwave Digestion System (Berghof speedwave1, Eningen, Germany). The digested samples were then transferred into a Teflon beaker, and total volume was made up to 25 mL with Milli-Q water. The digested solution was then filtered by using syringe filter (DISMIC1 – 25HP PTFE, pore size = 0.45 mm; Toyo Roshi Kaisha, Ltd., Tokyo, Japan) and stored in 50 mL polypropylene tubes (Nalgene, New York, NY, USA). 2.3. Instrumental analysis and quality control For trace metals, samples were analyzed using inductively coupled plasma mass spectrometer (ICP-MS, Agilent 7700 series, Santa Clara, CA, USA). Instrument operating conditions and parameters for metal analysis are presented in Table S2 (Supplementary material). The detection limits of ICP-MS for the studied metals were 0.7, 0.6, 0.8, 0.4, 0.06 and 0.09 ng/L for Cr, Ni, Cu, As, Cd and Pb, respectively. Multi-element Standard XSTC-13 (Spex CertiPrep1, Metuchen, NJ, USA) solutions were used to prepare calibration curves. Internal calibration standard solutions containing 1.0 mg/L of indium (In), yttrium (Y), beryllium (Be), tellurium (Te), cobalt (Co) and thallium (TI) were purchased from Spex Certi Prep1 (Metuchen, NJ, USA). During the procedure, 10 mg/L internal standard solution was prepared from the primary standard and added to the digested samples. Multi-element solution (purchased from Agilent Technologies) was used as tuning solution covering a wide range of masses of elements. All test batches were evaluated using an internal quality approach and validated if they satisfied the defined Internal Quality Controls (IQCs). Before starting the analysis sequence, relative standard deviation (RSD, <5%) was checked by using tuning solution purchased from Agilent Technologies. The certified reference materials INCT-CF-3 (corn flour) and DORM-2 (dogfish muscle), both from the National Research Council (Canada), were analyzed to confirm analytical performance and good precision (relative standard deviation bellow 20%) of the applied method (Table S3; Supplementary material).