Endometrial biopsy and curettage are widely used uterine sampling procedures. Occasionally, the amount of tissue obtained is so minimal that diagnosis cannot be reached. In published studies, insufficient samples comprise 2% to 60% of endometrial samples and are influenced by multiple contributory factors. To date, specific quantitative criteria for an adequate endometrial sample have not been established in the pathology community. The decision to classify a sample as nondiagnostic is subject to great interobserver variability, especially where elderly patients are concerned. Meanwhile, whether or not to repeat the procedure is the dilemma for clinical management. Herein we studied the clinicopathologic features of 1120 endometrial samples designated as insufficient for diagnosis. Such samples were more commonly encountered in elderly patients than younger ones (14.6% vs. 5.8%). Our pathologists generally required one intact tissue fragment containing both glands and stroma for premenopausal patients and 5 to 10 strips of atrophic endometrial epithelium for postmenopausal patients. By the 12-mo follow-up, 38% of patients with nondiagnostic samples had second sampling procedures and 7% underwent hysterectomy. The second sample was adequate in 75% of patients, 10% of which showed malignant tumor. If any worrisome histologic findings were present in the initial nondiagnostic sample, a high percentage of these patients were found to have uterine malignancy on second procedures (43%). Aims of this study are to bring awareness to this commonly overlooked topic, to validate our diagnostic criteria, and to outline important clinical implications.