Background Endoscopic submucosal dissection (ESD) has come to be widely performed for reduced invasiveness; however, its safety in patients with co-morbidities is not fully examined. We aimed to evaluate the safety and efficacy of gastric ESD with co-morbidities categorized according to ASA Physical Status Classification. Methods Two hundred and forty patients of ASA 1 (no co-morbidities), 268 of ASA 2 (mild), and 19 of ASA 3 (severe) were treated by ESD for gastric neoplasms. We retrospectively compared clinicopathological features and treatment results of these three groups. Results Cases (by percent) treated with anticoagulant/ platelet agents were more common in the higher ASA grades (ASA 1, 5.8%; ASA 2, 29.1%; ASA 3, 31.6%; P \ 0.0001). There were no significant differences in case numbers treated under guideline criteria, curative resection (ASA 1, 79.6%; ASA 2, 79.9%; ASA 3, 78.9%), or complications related to the ESD procedure (e.g., postoperative bleeding, perforation, thermal injury). By a patient risk prediction model on surgery, i.e., P-POSSUM, morbidity was halved, and no patients died compared to a predicted death rate of 0.5-2%; however, total and complications unrelated to ESD procedure (e.g., aspiration pneumonia, ischemic heat attack) were more common in higher ASA grades (ASA 1, ASA 2, ASA 3: 15.4, 23.9, 26.3%, respectively, P = 0.014; 0.4, 7.1, 0%, respectively, P = 0.00087). Deviation rates from clinical pathway were more frequent and hospital stay (days) longer in higher ASA grades (ASA 1, ASA 2, ASA 3: 11.3, 17.9, 26.3%, respectively, P = 0.014; 8, 8, 9%, respectively, P = 0.0053). Conclusions ESD is an efficient treatment for gastric neoplasms with co-morbidities. However, additional caution is required because co-morbidity is a risk factor for both total complications and complications unrelated to the ESD procedure, and may cause deviations in the clinical course and prolonged hospital stay.