Page 97 - The diagnostic work-up of women with postmenopausal bleeding

Statistical analyses
Statistical analysis was performed according to the intention-to-treat principle.
Differences in dichotomous outcomes were analysed with the chi-square test,
or Fisher’s exact test when the expected frequencies fell below five. Continuous
variables were tested for normal distribution, and since none of the variables were
normally distributed, a Mann-Whitney U-test for univariate analysis was used. The
primary and secondary outcomes were compared by calculating relative risks (RRs)
and their 95% confidence intervals (CIs).Women who were lost to follow-up were
excluded in the analysis. Kaplan-Meier curves were conducted to present the time
to recurrent bleeding in both groups and log-rank test was used to test differences
in time to recurrent bleeding. Sensitivity and specificity for SIS were calculated,
with visual hysteroscopy results (polyp yes/no) as reference standard.Women with
inconclusive SIS were considered as suspicious for having endometrial pathology
and have an indication for a hysteroscopy.Therefore, all women with inconclusive SIS
were considered as positive result and counted as such in the sensitivity calculation.18 5 For analysis, the Statistical Package for the Social Sciences (IBM Corp, Armonk, NY,
USA) version 20.0 was utilised. Statistical significance was set at p < 0.05.
Results
During the study period 201 women presenting with postmenopausal bleeding, an endometrial thickness of > four millimetres and a benign histology result in the 12 participating hospitals, agreed to participate in the study. One woman was excluded because of use of Tamoxifen, 98 were randomly allocated to SIS and hysteroscopy and 102 women to expectant management.
The baseline characteristics of the participants in the two groups were comparable (Table 1).Of the 98 women in the SIS/hysteroscopy group,87 underwent hysteroscopy (89%) and 11 women declined further invasive diagnostics and opted for expectant management (Figure 1). In six of these 87 women SIS was not performed, because of protocol violation. In the 102 women allocated to expectant management, all women followed the study protocol.Two of these women were lost to follow-up; we could not contact them after one year and they were excluded from analysis.
Diagnostic work-up: a RCT
95