Results
A total of 642 hysteroscopies for 577 patients aged 70 years or older were performed during the study period. The median patient age at the time of the procedure was 76 years (interquartile range [IQR] 73-80). Two-hundred and thirty hysteroscopies were performed in the outpatient hysteroscopy clinic, and 412 hysteroscopies were performed in the OR. Twelve patients underwent repeat hysteroscopic procedures unrelated to the primary procedure, and those repeat procedures were excluded from the analysis. The final analysis consisted of a total of 630 hysteroscopies, of which 225 were performed in the office and 405 in the OR. Five-hundred seventy-seven hysteroscopies were first interventions and another 53 were second interventions. Two-hundred and twenty-two of the former patients were included in the office hysteroscopy group, and 355 patients were included in the OR hysteroscopy group.
Figure 1A illustrates the volume of procedures among the two groups of women during the study period. While the total number of hysteroscopies more than doubled during this period, the number of procedures for this population in the outpatient hysteroscopy clinic increased by approximately 16 times. Thus, by the end of the follow-up period, an office procedure became the primary hysteroscopic intervention for patients 70 years of age or older.
The study population included 416 (72.1%) patients aged 70-79 years, 155 (26.9%) patients aged 80-89 years, and six (1.0%) patients aged 90 years and over (the oldest was 93 years old). The baseline characteristics of the two groups are detailed in Table 1 . The median age at the time of hysteroscopy was 76 years (IQR 73-80), and it was similar for the two groups. More women were referred to office procedures due to abnormal sonographic findings (67.1% vs. 49.6%, p<0.001: Table 1), while more women with active PMB were referred to procedures in the OR (47% vs. 31.1% p<0.001: Table 1).
The procedure was successful in 194 (87.4%) patients whose procedure was performed in the office compared with 342 (96.3%) patients whose procedure was performed in the OR under anesthesia. The diagnosis of cervical stenosis or difficulty in penetrating the cervix was significantly higher in the office group. Out of 54 (24.3%) cases in the office group with cervical stenosis, the surgeon managed to enter the uterine cavity in 32 (59.2%). The surgeon failed to penetrate the uterine cavity, mainly due to cervical stenosis, obstructive cervical mass, and patient intolerance, in 28 (12.6%) of office hysteroscopies. Forty-five (20.3%) of the hysteroscopies in our clinic were not completed in a one-step procedure, mainly due to large or multiple findings and uterine bleeding. Only 7 (3.15%) procedures were abandoned due to patient discomfort.
A total of 47 (21.2%) patients who underwent office hysteroscopy required a second hysteroscopy, 19 due to failure in the first visit, 27 due to incomplete procedures, and one due to continuous bleeding that precluded the performance of the intervention. In 44 of those 47 cases, the second hysteroscopy was performed in the OR under general anesthesia. A total of six patients who underwent OR hysteroscopy under general anesthesia required a second procedure (two cases due to failure and four cases due to non-diagnostic histopathologic samples taken at the first procedure).
Overall, 150 (67.6%) of office hysteroscopies had a positive finding. Moreover, 43 (43/69=62.3%) of the women with PMB and 106 (106/149=71.1%) of those with abnormal sonographic findings also had a positive hysteroscopic finding(s). The histopathological findings by groups are shown in Table 2 . PMB was a risk factor for malignancy for the entire cohort, with 73/236 (30.9%) patients with PMB diagnosed as having a malignancy compared to 25/341 (7.3%) patients with no PMB (p < .001, odds ratio (OR) = 5.66; confidence interval (CI) 3.46-9.26). In the repeat hysteroscopy subgroup, 7/17 (41.2%) patients with PMB were diagnosed as having a malignancy compared to 6/36 (16.7%) patients with no bleeding (p = .053, odds ratio (OR) = 3.5; confidence interval (CI) 0.95-12.89).
Figure 1B illustrates the relative contribution of each of the hysteroscopic approaches to the diagnosis of new abnormal findings among the 577 patients included in our cohort.
While samples taken in the outpatient hysteroscopy clinic during the first years did not contribute to a new diagnosis, their contribution increased during the following years.
All patients who underwent hysteroscopy in our outpatient clinic were discharged on the same day. The median length of hospitalization was three days (range 0-28 days) for the patients who underwent the procedures in the OR. Three-hundred and twenty-nine cases (92.7% of all procedures under anesthesia) were performed in an inpatient setting, while 26 (7.3% of all procedures under anesthesia) were performed in our outpatient ORs.
Short-term complications among patients treated in our outpatient hysteroscopy clinic included one case (0.45%) of uterine perforation that was managed expectantly, and one case of bleeding that necessitated admission a few days later for a repeat procedure under general anesthesia. Four of the 355 (1.13%) patients who underwent an OR hysteroscopy sustained uterine perforation. Emergent abdominal surgeries were required in three of those cases, and they included two abdominal hysterectomies, including one case of left colectomy. Three patients (0.85%) sustained complications that required admission to the intensive care unit (ICU): there was one cardiac arrest, and two cases of severe sepsis (secondary to pyometra). Unfortunately, the latter two patients died during the hospitalization, yielding a mortality risk of 0.56% for patients aged 70 years or older in our cohort who underwent OR hysteroscopy.