In a prospective trial with and without laser therapy the outcome on 105 patients was evaluated according to a defined protocol and terminology. Only patients with actively bleeding lesions during early endoscopy had laser therapy. Patients showing evidence of recent bleeding or no bleeding had none. Defects of the laser equipment or impossibility of reaching a bleeding lesion prevented its use in some patients (control group). There was no difference concerning rate of recurrent bleeding or mortality. Controlled trials with defined groups of patients are urgently needed.