Inclusions and exclusions
Participants were eligible for the study if at the time of study entry, they had been registered as an active patient for at least one year, and their health records had been accessed at least once within the past ten years from their study entry date. To avoid bias due to ‘immortal periods’, actively registered patients who were prescribed HRT after the acceptable mortality reporting (AMR) date (a starting date from which practice recorded mortality was close to age- and sex-standardized national mortality rates) of the corresponding GP practice were selected for the study. Patients with a previous history of any kind of cancer, acute myocardial infarction, serious heart failure, stroke (except transient ischaemic attack), chronic kidney disease (stage 3 to 5), dementia, premature ovarian insufficiency, surgically induced menopause before 45, and premature menopause were excluded from the study at baseline.
The analyses included patients who were born between 1921 and 1960 and started HRT at the selected age from 1984 until the study end date which was 1st January 2017, and their matched controls. Participants were followed up from the date of first HRT prescription until death, or transferred out, or study end date whichever comes first. Patients who were transferred out during the study period were no longer followed up, and their observation time was censored at that point of time. The SQL server 2016 was used to extract data from THIN.