Use of statins and risk of uterine leiomyoma: A cohort study in the UK Biobank

Abstract

Statins are primarily used to treat hypercholesterolemia and for the secondary prevention of coronary artery disease. Besides their lipid-lowering effect, statins demonstrate other pleiotropic effects. For example, experimental studies of human cell lines and animal models have demonstrated that statins could suppress uterine leiomyoma growth and may prevent uterine leiomyoma.^{1, 2}

However, only two clinical studies have investigated the association between use of statins and risk of uterine leiomyoma, and their results are inconsistent. The first one was a nested case-control study. The study indicated that use of statins was associated with a reduced risk of uterine leiomyoma, which was consistent with previous experimental studies.³ The second study on this topic was published by our team. In contrast to the nested case-control study, by analyzing data from FDA Adverse Event Reporting System (FAERS), we found that uses of simvastatin, rosuvastatin, and fluvastatin might be associated with increased risk of uterine leiomyoma, which suggested that this issue was not definitively resolved.⁴

The UK Biobank is a large-scale database containing extensive sociodemographic, lifestyle, and clinical data on half a million participants.⁵ Leveraging this database, we conducted a cohort study to explore the association between use of statins and risk of uterine leiomyoma.

Details of the cohort study are provided in the Supplementary Materials. In brief, the study population were premenopausal female participants in the UK Biobank. We restricted the study population to premenopausal females because uterine leiomyomas were less likely to develop after menopause.⁶ Based on long-term treatment with a statin or not, the participants were divided into statin users and nonusers. The outcome was the first inpatient diagnosis (either main or secondary) of uterine leiomyoma during follow-up.

Female participants were followed from baseline visit until uterine leiomyoma diagnosis, death, or the last linkage date with hospital inpatient data, whichever came first. In order to minimize the potential for reverse causality, we excluded the first year of follow-up (for all individuals).

The covariates included age, race, Townsend deprivation index, smoking status, alcohol use, vigorous physical activity, age at menarche, number of childbirth, number of abortion, any comorbidity at baseline (hyperlipidemia, ischemic heart disease, ischemic cerebrovascular disease, hypertension, diabetes, obesity, or pelvic inflammatory disease), and oral contraceptive. These covariates were factors known to be correlated with risk of uterine leiomyoma in previous literatures,^7-9 or indications for use of statins.¹⁰

We used Cox proportional hazards regression to analyze the association between use of statins and risk of uterine leiomyoma, with results expressed as hazard ratios (HRs) and 95% confidence intervals (95% CI). We developed a multivariable model with adjustment for the above covariates. Subgroup analyses calculated HRs by stratifying the study cohort according to covariates and individual comorbidities. Moreover, we also performed subgroup analysis stratified by statin type.

We conducted several sensitivity analyses. First, to balance baseline differences between statin users and nonusers, we performed 1:1 propensity score matching (PSM). Second, to further minimize the potential for reverse causality, we performed a sensitivity analysis by excluding the first 2 years of follow-up (for all individuals). Third, to test the robustness of our results in relation to the definition of uterine leiomyoma, we performed a sensitivity analysis by ascertaining uterine leiomyoma solely through the main diagnosis reported in the linked hospital inpatient data. Fourth, as the average age of menopause in UK women is 51 years,¹¹ we performed a sensitivity analysis by censoring the follow-up at age 51.

After selection, a total of 60,635 female participants were included in our study (1687 statin users and 58,948 nonusers) (Figure S1). The mean age of the included participants was 46.18 years (standard deviation, 4.26) at baseline. Table S1 shows the baseline characteristics of participants according to use of statins.

In the main analysis, during a median follow-up of 8.12 years (interquartile range, 7.44–8.72), 1807 female participants (64 statin users and 1743 nonusers) had a first inpatient diagnosis of uterine leiomyoma. After adjustment for the covariates, we observed no significant association between use of statins and risk of uterine leiomyoma (adjusted HR, 1.04; 95% CI, 0.80–1.34) (Table 1). Moreover, no individual statin type was significantly associated with risk of uterine leiomyoma (Table 1).

Figure S2 shows stratified analyses by covariates. The associations between use of statins and risk of uterine leiomyoma did not significantly differ by age, race, Townsend deprivation index, smoking status, alcohol use, vigorous physical activity, any comorbidity at baseline, or oral contraceptive.

Figure S3 shows stratified analyses by individual comorbidities. The associations between use of statins and risk of uterine leiomyoma did not significantly differ by hyperlipidemia, ischemic cerebrovascular disease, hypertension, diabetes, obesity, or pelvic inflammatory disease.

In our sensitivity analyses (Table 1), the association between use of statins and risk of uterine leiomyoma remained: (1) when we performed PSM and the baseline characteristics of two matched groups were all balanced (Table S2 and Figure S4); (2) when we excluded the first 2 years of follow-up; (3) when we used only the main diagnoses to identify uterine leiomyoma; and (4) when we censored the follow-up at age 51 years old.

Thus, in this large-scale cohort study, we observed no reduced risk of uterine leiomyoma among premenopausal participants with long-term use of statins. Such association was independent of many sociodemographic, lifestyle, and clinical characteristics. Moreover, no significant association was found for all individual type of statins and similar results were yielded in a number of sensitivity analyses.

Our results were inconsistent with previous clinical studies. Borahay et al. conducted a nested case-control study using data from a large commercial health insurance program and found that use of statins was associated with a reduced risk of uterine leiomyoma.³ We could not completely explain the discrepancies between Borahay et al. study and the present study, but it should be noted that some differences in study design exist. First, due to the limited information contained in health insurance data, Borahay et al. study only analyzed the potential confounding effects of age, region and comorbidity, and were unable to analyze the potential confounding effects of other sociodemographic, lifestyle, and clinical factors. In our study, UK Biobank contained extensive sociodemographic, lifestyle, and clinical information. Thus, compared with Borahay et al. study, we further analyzed the potential confounding effects of race, Townsend deprivation index, smoking status, vigorous physical activity, age at menarche, number of childbirth, number of abortion, and oral contraceptive. All these factors had been reported to be correlated with the occurrence of uterine leiomyoma. By adjusting for these potential confounding factors, our study might provide more reliable results than Borahay et al. study. Second, the uterine leiomyoma cases in Borahay et al. study contained some postmenopausal females. However, it was very unlikely that uterine leiomyoma would develop after menopause.⁶ The fact might be that those postmenopausal cases diagnosed with uterine leiomyoma might have already had uterine leiomyomas years ago, and this might lead to a reverse causality bias in Borahay et al. study. In order to avoid this bias, we restricted our study population to premenopausal females.

In our previous pharmacovigilance study, by disproportionality analyses using the FAERS database, we found that uses of simvastatin, rosuvastatin, and fluvastatin might be associated with increased risk of uterine leiomyoma.⁴ However, the results of disproportionality analyses could only demonstrate statistical associations and not causations. As exposure was identified before the outcome, our present cohort study might be able to establish a causation between use of statins and risk of uterine leiomyoma.¹² Thus, the results of our present cohort study were more reliable than those of our previous pharmacovigilance study.

Our study had several strengths. First, the UK Biobank contained extensive sociodemographic, lifestyle, and clinical information, which enabled us to adjust for a wide range of confounders and conduct comprehensive subgroup analyses. Second, the associations between use of statins and risk of uterine leiomyoma remained consistent in most subgroup and sensitivity analyses, which further confirmed the robustness of our results.

Our study also had some limitations. First, information on incident uterine leiomyoma was obtained by linking to hospital inpatient data. That said, uterine leiomyomas diagnosed at the outpatient clinic, and asymptomatic/undiagnosed ones were not captured in our data. This misclassification might be differential between statin users and nonusers because users had more frequent healthcare visits and were subjected to surveillance bias. However, when using only the main diagnoses from linked inpatient hospital data to identify uterine leiomyoma, the result remained unchanged. Second, potential reverse causality might exist in our study as it took years for uterine leiomyoma to develop. However, the result remained unchanged when we excluded the first 2 years of follow-up. Third, although we adjusted for all main indications for statins in the statistical model and further performed subgroup analyses, indication bias could not be completely avoided.

In conclusion, in this cohort study of UK Biobank female participants, use of statins is not associated with reduced risk of uterine leiomyoma. These findings do not support that use of statins may prevent uterine leiomyoma. However, our findings should be interpreted with cautions due to indication and surveillance biases.

XFJ, LZ, DL, and WZ were responsible for the study concept and design. HY, YH, YQ, and WC did the data and project management. XFJ did the data cleaning and analysis. XFJ and LH made the figures and tables. HL, LZ, and YS interpreted the data. XFJ drafted the manuscript. DL and LZ revised the manuscript. All the authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

The authors have no conflicts of interest to disclose.

This study was supported by Natural Science Foundation of Sichuan Province (grant number 2022NSFSC0644). The funders had no role in the development of this article (i.e., in the study design; collection, analysis, and interpretation of data; report writing; or decision to submit the paper for publication).