Unscheduled bleeding and endometrial cancer in women on postmenopausal hormone replacement therapy and their matched controls: protocol for a descriptive cohort study using the Orchid-e database

Hormone replacement therapy (HRT) with oestrogen and progestogen is a common medical treatment for alleviating symptoms of menopause. Since 2015, its use has been increasing in the UK. Unscheduled bleeding can be a symptom of endometrial cancer, and guidelines state that women experiencing this should have an urgent referral for suspected endometrial cancer. However, unscheduled bleeding is also common in women taking HRT, particularly in the first few months after starting HRT or if there is a change in regimen. Current guidelines may result in women on HRT receiving referrals that are not necessary and undergoing unpleasant and invasive tests such as hysteroscopy. However, there is a lack of current information to guide recommendations. This protocol describes a cohort study in the ORCHID-e database of anonymised patient records from English primary care. We will use a cohort of women aged over 40 years starting on HRT with oestrogen and progestogen, age matched to women who have not started HRT. Exposure will be a prescription for oestrogen containing HRT with no previous prescription for oestrogen containing HRT in the previous year. Index date in each matched set will be the date of this prescription. Prescriptions for progestogen containing drugs will not be used to define the exposure, but this information will be extracted to describe the study population and for sensitivity analyses. Outcomes will be consultations for unscheduled bleeding, urgent referrals for suspected endometrial cancer, and diagnosis of endometrial cancer. Women will be followed up until they change exposure status or are otherwise censored. Women who start taking HRT in follow-up will re-enter the cohort in the exposed group. We will describe proportions of women with a code for consulting with unscheduled bleeding, proportions of those women referred for further investigation on the pathway for suspected endometrial cancer, and proportions diagnosed with endometrial cancer within one year of referral. We will investigate the diagnostic accuracy of unscheduled bleeding for endometrial cancer separately for women on HRT and those not on HRT. Analyses will be done by 6-month categories of time since index, age, calendar year, sociodemographic variables, risk factors for endometrial cancer, type of HRT.