|| Checking for direct PDF access through Ovid
Postmenopausal bleeding (PMB) always requires further investigation because about 1/10 affected women will prove to have endometrial cancer. The risk level is accurately discerned by measuring endometrial thickness by transvaginal ultrasonography using a cutoff of 5 mm or higher to define diseased endometrium. The present retrospective study sought to determine the frequency of endometrial and cervical pathology in women with PMB whose initial endometrial thickness was less than 4.4 mm, denoting thin endometrium. A total of 332 women were seen in the years 1992–2002 with PMB, defined as any vaginal bleeding in a postmenopausal woman not taking hormone replacement therapy (or unscheduled bleeding in a woman on hormones). The 313 traceable women had a median age of 63 years when PMB began, and were followed up for a mean of 7.5 years.Two cases of cervical cancer but no endometrial cancers were detected during follow-up. Using a reference population of women from southern Sweden, the incidence of cervical cancer was higher than expected. The standard incidence ratio was 8.7, with a 95% confidence interval of 1.1–31.4. The frequency of a pathological Pap smear during follow-up was 3.9%, compared to an expected prevalence of 4.0%. Twenty-five of 41 women who were admitted because of recurrent bleeding had an endometrial biopsy, and 4 of them (16%) were found to have endometrial cancer. Repeat Pap smears disclosed abnormal findings in 3/28 women with rebleeding. Cervical cancer was not significantly more prevalent in women with recurrent bleeding than in the general population without PMB. Endometrial and cervical diseases were similarly prevalent in women whose endometrium was less than 4.4 mm thick and those whose endometrial thickness exceeded 4.5 mm.These findings suggest that diagnostic efforts focus on ruling out cervical disease in women with PMB and thin endometrium. Because recurrent bleeding appears to correlate with an increased risk of both cervical and endometrial pathology, invasive endometrial sampling and repeated cervical assessment are indicated.