Examination of the complete uterus at hysterectomy is the most accurate means of detecting endometrial pathology. Hysterectomy is the gold standard against which other diagnostic techniques are to be judged. In one study, comparing dilatation and curettage (D&C) under general anesthesia to hysterectomy, D&C sampled less than half of the endometrial surface in 60% of specimens (Stock). In a similar comparison of both the Pipelle and Vabra devices to hysterectomy, the Pipelle generally left a single tract in the endometrium and sampled, on average, only 2.4 out of 8 possible sectors of the uterine cavity, examining only 4.2% of the endometrium. The Vabra aspirator sampled, on average, 7.4 out of 8 possible sectors of the uterine cavity, and 41.6% of the endometrium (Rodriguez). Endometrium from the cornual regions of the uterus is likely to be missed by suction curettage (Hale).

Suction curettage often fails to remove endometrial polyps detected by dilatation and curettage (Fothergill, Goldchmit, Grimes).

The histologic diagnosis from suction curettage is identical to histology from D&C or hysterectomy in 67% to 98% of biopsies (Guido, Ferry, Fothergill, Goldchmit, Greenwood, Koonings, Lipscomb, Lutz, Rodriguez, Stovall). Endometrial cancer or hyperplasia may be missed by suction curettage in up to 30% of cases, especially if the lesion is localized to a polyp or involves less the 5% of the endometrium (Guido, Ferry, Fothergill, Goldchmit, Greenwood, Lutz, Stovall). Conversely, suction biopsy may diagnose endometrial hyperplasia missed by D&C (Fothergill, Goldchmit). Inability to interpret an endometrial biopsy because of insufficient tissue should not be viewed as a negative biopsy (Greenwood, Heller). Greenwood followed 871 symptomatic women for almost 3 years after an endometrial biopsy. In 78 of these women, the initial biopsy was insufficient to interpret. Cancer or atypical hyperplasia was later diagnosed in 15 of these 78 patients (Greenwood).