We identified 224 women aged 50 to 79 with incident breast cancer in the base population and 1009 controls matched by age, sex, general practice, duration of prescription drug history in the GPRD, and index date.
Characteristics of the cases and controls are listed in . Body mass index (BMI) was greater than 28
in a higher proportion of cases than controls. Although hormone replacement therapy (HRT) was used by approximately the same proportion of cases as controls, current HRT use was more frequent among cases. Cases were more likely than controls to have a history of benign breast disease.
Characteristics of cases and controls
lists the proportions of cases and controls with a recorded history of hyperlipidaemia and treatment with lipid-lowering drugs. Untreated hyperlipidaemia was more frequent among cases than controls.
Exposure classification of cases and controls
Multivariable model-adjusted relative risks for breast cancer according to exposure group are presented in . Compared with women without hyperlipidaemia, the relative risk (RR) for breast cancer among women currently treated with statins was 1.0 (0.6–1.6). Among the 200 current statin users, 26 were current users of pravastatin (two cases and 24 controls), the RR among women currently treated with pravastatin compared to women without hyperlipidaemia was 0.4 (0.1–1.8).
Multivariable model-adjusted relative risks of breast cancer with use of lipid-lowering drugs, untreated hyperlipidaemia, and other covariates
Compared to women without hyperlipidaemia, the risk was modestly increased among women with untreated hyperlipidaemia (RR 1.6 (1.1–2.5)) and among those with hyperlipidaemia currently treated with non-statin lipid-lowering drugs (RR 1.8 (0.9–3.4)). Compared with the same reference group, the RR for breast cancer among women with past statin use was similar to that of current statin users, and past non-statin users had a risk that was nearly the same as that of current non-statin users.
Among current statin users, the median duration of treatment was 1.8 years and the maximum duration was 8.6 years. The data show no trend in risk with increasing duration of current statin use: the adjusted RRs of breast cancer were 1.1 (0.4–2.8) among six cases and 28 controls with current statin use of duration 3 to 5 years and 1.1 (0.4–3.0) among five cases and 26 controls with current statin use of duration longer than 5 years.
Current long-term use of hormone replacement therapy (HRT) was independently associated with a cumulative increase in risk (up to RR 2.1 (1.2–3.7) for women with 24 or more prescriptions compared with non-users; see ), while past users had a somewhat decreased risk (RR 0.6 (0.4–1.0)). A history of benign breast disease was associated with an increased risk of breast cancer (RR 1.6 (0.8–3.1)). Women with BMI greater than 28
had a slightly increased RR of 1.4 (0.9–2.1) compared to women with BMI less than 24
We considered that an association with hyperlipidaemia might occur spuriously if hyperlipidaemia were diagnosed incidentally during the evaluation of patients with suspicious breast lesions. Therefore, we restricted a secondary analysis to case–control sets in which the first diagnosis of hyperlipidaemia (if any) or prescription for a lipid-lowering drug (if any) was recorded 6 months or more before the index date. In this analysis of 1095 subjects (208 cases and 887 controls), the RRs among women with untreated hyperlipidaemia (1.9 (1.1–2.6)) and women who had hyperlipidaemia treated with non-statin drugs either currently (1.8 (0.9–3.9)) or in the past (2.0 (0.9–4.4)) were similar to those in the primary analysis. This additional analysis provided further evidence that there was no increased risk among women treated with statins either currently (RR 0.8 (0.5–1.4)) or in the past (RR 1.2 (0.6–2.6)).