In 21 patients who had undergone resection and reconstruction for esophageal carcinoma, the postoperative platelet aggregability was measured and the correlation between the occurrence of anastomotic leakage and platelet aggregability was investigated. There was no statistical difference in the clinical features of the patients between those with (n = 5) and without (n = 16) anastomotic leakage. Platelet aggregability was measured by the turbidimetric method from blood samples taken preoperatively, and at 1, 3, and 7 postoperative days (POD). The average values of platelet aggregability in patients without anastomotic leakage were 81.2, 70.4, 80.1, and 81.8%, while those with leakage were 81.3, 47.6, 52.3, 70.6% preoperatively, and 1, 3, and 7 POD, respectively. Thus, platelet aggregability significantly decreased in patients with anastomotic leakage on the first postoperative day (p < .05), and then gradually recovered postoperatively as time passed. Therefore, the measurement of platelet aggregability is considered to be one of the parameters predicting the occurrence of anastomotic leakage, and a prevention in the decrease of platelet aggregability as well as its activation could become a treatment for preventing anastomotic leakage.