Detection of endometrial pathology using saline infusion sonography versus gel instillation sonography: a prospective cohort study

Objective To compare saline infusion sonography (SIS) with gel instillation sonography (GIS) in terms of feasibility and diagnostic accuracy. Design Prospective cohort study. Setting Leuven University bleeding clinic. Patient(s) A total of 804 patients: two consecutive cohorts of 402 women undergoing SIS or GIS. Intervention(s) Vaginal ultrasound (n = 804) followed by SIS (n = 402) or GIS (n = 402); office hysteroscopy in 685 patients, and endometrium sampling in 487 patients; surgery in 274 women: operative hysteroscopy (n = 230) or hysterectomy (n = 44). Main Outcome Measure(s) Patients' characteristics, technical failure rates, and final diagnosis. Pathology was defined as endometrial hyperplasia, polyp, cancer, or intracavitary myomas. Result(s) The technical failure rate (difference between proportions and confidence interval) was 5.0% for SIS versus 1.8% for GIS, respectively (3.21; [0.69–5.95]). Failure due to inadequate distension was 1.5% versus 0.3% for SIS and GIS, respectively (1.25; [–0.16–2.99]). Pathology was diagnosed in 180 patients (49%) of the SIS group versus 147 patients of the GIS group (40.2%) (8.88; [1.69–15.95]). The sensitivity was 77.8% and 85.0%, respectively (NS). The negative predictive value was 79.1% for SIS and 88.6% for GIS (9.54; [2.17–16.89]). Conclusion(s) Gel instillation sonography is a feasible, accurate alternative for SIS in the evaluation of women with abnormal bleeding, and has fewer technical failures.