The transthoracic and transhiatal resection techniques are compared using the 30-year experience of Oulu University Central Hospital. During the period 1960-1982 we favoured resections trough a transthoracic route, while during the period 1983-1989 a transhiatal route was preferred. This change, and the more enthusiastic attitude adopted towards resection, has lead to an increase in resectability from 23% (46/203) to 62% (43/69) (P less than 0.0001). The difference between radical resections, 50% (23/46) and 37% (16/43) has remained non-significant. Morbidity was higher after transthoracic than transhiatal resections, 57% (26/46) versus 42% (18/43), whereas mortality was nearly the same, 11% (5/46) and 9% (4/43). Postoperative pulmonary complications occurred in 28% (13/46) after transthoracic resection and in 14% (6/43) after transhiatal resection. No significant difference was detected in the development of late anastomotic strictures, 33% (15/46) and 30% (13/43), respectively. We conclude that transhiatal resection is as safe as transthoracic resection and seems to allow more resections to be carried out without any increase in mortality or morbidity, but long-term survival remains poor.