The effect of different membranes, membrane modules, feed temperatures, flow rates and solute concentrations on the permeate flux and salt rejection in direct contact membrane distillation (DCMD) was first studied with synthetic seawater and compared to distilled water. After optimizing these operating conditions, DCMD was tested with real feed samples, namely river water (RW-R), seawater (SW-R), and a secondary effluent from a municipal wastewater treatment plant (MW-R). The permeate flux achieved with MW-R was significantly lower than those obtained with the other feed samples. Two membrane module configurations (H-cell and W-cell) were then studied using SW-S, spiking diphenhydramine (DP) as model organic pollutant in some experiments. The H-cell performed better in terms of permeate quality for the same volume of permeate collected. A long experiment (500 h) was conducted with SW-R employing a larger H-cell. Severe fouling was observed, but high rejections of ion species (> 99%) were recorded together with complete rejections of pharmaceuticals (diclofenac, azithromycin, clarithromycin and erythromycin) detected in SW-R at 9.53-73.53 ng L-1 (detection limits < 0.16 ng L-1). Colonies of Escherichia colt or enterococci were not detected in 100 mL of permeate (distillate) solution, complying with the European Directive for drinking water.