Background-Polyethylene glycol (PEG) 3350 is a non-absorbable, nonmetabolised osmotic agent used in lavage solutions for gut cleansing. Aims-To compare the eYcacy of PEG and lactulose in chronic constipation. Methods-A total of 115 patients with chronic constipation entered a multicentre, randomised, comparative trial. They initially received two sachets containing either PEG (13 g/sachet) or lactulose (10 g/sachet) and were given an option to change the dose to one or three sachets/ day, depending on response. Results-Ninety nine patients completed the trial. After four weeks, patients in the PEG group (n=50) had a higher number of stools and a lower median daily score for straining at stool than patients in the lactulose group (n=49). Overall improvement was greater in the PEG group. Clinical tolerance was similar in the two groups, but flatus was less frequently reported in the PEG group. The mean number of liquid stools was higher in the PEG group but the diVerence was significant only for the first two weeks. There were no serious adverse events and no significant change in laboratory tests in either group. At the end of the study, the number of sachets used by the patients was 1.6 (0.7)/day in the PEG group and 2.1 (0.7)/day in the lactulose group. Sixty one patients completed a further two months open study of one to three sachets PEG daily; there was no loss of eYcacy and no serious toxicity. Conclusion-Low dose PEG 3350 was more eVective than lactulose and better tolerated. (Gut 1999;44:226-230)