OBJECTIVE--By audit from January to June 1989 to quantify, separately for hearts, kidneys, liver, lungs and corneas, the possible increases in transplantable organs from brain stem dead potential donors in intensive care units and to compare them with the increases achieved in October-November 1989, during intense, national publicity about transplantation. DESIGN--Prospective audit of all deaths in intensive care units in England from 1 January to 30 June 1989 and subsequent case study of the impact of publicity on offers and donations during October-November 1989. SETTING--15 regional and special health authorities in England. PATIENTS--5803 patients dying in intensive care units, of whom 497 were confirmed as brain stem dead and had no general medical contraindication to organ donation. MAIN OUTCOME MEASURES--Organ specific suitability for transplantation (as reported by intensive care units); consent for donation of specific suitable organs; and procurement of specific organs reported as suitable for transplantation and offered. RESULTS--In the 497 (8.6%) brain stem dead potential donors were estimated the organ specific suitability for heart as 63%, kidneys 95%, liver 70%, lungs 29%, and corneas 91%. Refusal of relatives (30%) accounted for major losses of suitable organs of all types. For kidneys the loss was equivalent to 44% of brain stem dead actual kidney donors. No discussion of organ donation was the second most important reason for missed kidney donors, the loss being equivalent to 10% of brain stem dead actual donors. Non-procurement or difficulties with allocating organs was the second most notable cause of missed suitable liver and lung donors; 29% (55) of the offered total of 189 liver donors and 27% (21) of 78 offered suitable lung donors in six months. Non-procurement of suitable, offered organs was rare for kidneys and modest, of the order of 13% and 10% respectively, for heart and corneas. Corneal donation from brain stem dead potential donors might be improved nearly as much (that is, a 78% increase in brain stem dead actual corneal donors) by specific measures to promote corneal donation when other organs are offered as by reducing the overall refusal rate. Restricted offers, non-procurement, and no discussion of donation accounted for nearly equal numbers of lost donations of hearts (each equivalent to 15% of donated hearts). During October-November 1989 when there was intense, positive publicity about transplantation the rates of refusal and non-discussion fell compared with during January-June (22%, 36/163 v 30%, 138/460; 7%, 33/497 v 2%, 4/167 respectively). Offers of suitable donors increased significantly (p less than 0.02) compared with the first six months of 1989, most notably for heart donors (80 v 60.1 expected) and kidney donors (122 v 102.1 expected) but only for kidneys was there a noticeable 17% increase in actual donors (118 actual audited donors v 100.8 expected donors; p = 0.09). CONCLUSIONS--Four strategies to increase the supply of transplantable organs from brain stem dead potential donors in intensive care units were identified: (a) reducing refusal of relatives (b) avoiding non-procurement of actually suitable organs (by logistical initiatives) and deterioration of initially suitable organs (by donor care initiatives); (c) converting restricted offers to unrestricted offers; and (d) ensuring discussion with families. Early referral to the transplant team or coordinator gives time for discussion about donor care and agreement on medical suitability for donation of specific organs. Solving some of the logistical problems of non-procurement may be a prerequisite for increased offers to be translated into increased donations. The impact of publicity therefore needs to be measured on offers of suitable donors as well as by actual donations.