Association Between Menopausal Hormone Therapy and the Risk of Urologic Cancer: A National Population-Based Retrospective Cohort Study in Korea.

Many women experience various symptoms of menopause, including vasomotor symptoms (sweating and hot flushes), vaginal symptoms (sexual discomfort and vaginal dryness), psychological problems (anxiety and low mood), frequent urinary infections, and urinary incontinence, 1 which are caused by a natural decline in oestrogen and progesterone levels. Hence, menopausal hormone therapy (MHT) has been widely prescribed to alleviate symptoms. MHT comes in several formulations, the most common of which are oestrogen, combined oestrogen and progestin, and tibolone (a synthetic steroid). Although there are benefits of MHT use, 2 previous reports have revealed that MHT use is particularly associated with urologic cancer in women, including kidney 3 4 5 6 and bladder cancer. 7 8 9 10 Urologic cancers refer to malignancies that affect the urinary system, including the kidneys, bladder, prostate, ureters, and urethra. Kidney cancer is the 16th most common cancer worldwide, accounting for 2.2% of all newly diagnosed cases of malignancy. 11 Risk factors for kidney cancer include cigarette smoking, hypertension, obesity, and chronic kidney disease 12 ; however, the aetiology of the disease remains unknown. Bladder cancer is the ninth most frequently diagnosed cancer worldwide. In 2018, the estimated number of newly diagnosed cases of bladder cancer in men and women in the United States was 62,100 and 19,300, respectively. 11 The most common risk factors for bladder cancer are smoking and occupational exposure to aromatic amines. 13 Epidemiological studies have shown that the incidence of both types of urologic cancer is three- to four-fold times higher in men than in women, 11 which may be because more men smoke; however, the known risk factors are unlikely to explain the differences in the incidence of urologic cancer between men and women. Another possible factor is the difference in the levels of hormones between sexes, including hormonal changes after menopause and, particularly, using exogenous hormones in MHT. Previous research reports on the association between hormonal factors, including exogenous hormone use in MHT, and the incidence of urologic cancer in women, including kidney 3 4 5 6 and bladder cancer 7 8 9 10 ; however, the results were inconsistent. Therefore, this study aimed to evaluate the relationship between MHT and urologic cancer risk in women using data from the Korean Health Insurance Review and Assessment Service (HIRA).

South Korea has a universal system for health insurance coverage that serves approximately 98% (51 million) of Koreans, called National Health Insurance, which is managed by the Korean National Health Insurance Corporation (NHIC). 14 Detailed information on the Korean National Health Insurance Database is presented in Supplementary Data 1 . We performed a retrospective cohort study using health insurance data from the HIRA between January 1, 2002 and December 31, 2019. Diagnostic codes were extracted from the database using the International Classification of Diseases, 10th revision, surgery codes were extracted from the database using the Korea Health Insurance Medical Care Expenses (2012, 2016, 2019 version), and prescription codes were extracted from the database using the HIRA Drug Ingredients Codes. This retrospective cohort study was designed to emulate a target trial to estimate the causal effect of MHT on urologic cancer incidence. Eligible participants were postmenopausal women aged > 40 years who underwent national health examinations between 2002 and 2011 under South Korea’s National Health Insurance system, which covers approximately 98% of the population. The “time zero” for MHT users was defined as the date of first MHT prescription (with ≥ 6 months duration), and for non-users, the date of health examination during the same time window. Women who were still undergoing menopause or entered menopause during the first year of the study (2002) were excluded to ensure that MHT was initiated after menopause. To reduce immortal time bias and reverse causation, we excluded women diagnosed with any cancer (Cxx) or urinary tract disease (N00–N49) within 180 days following cohort entry. This lag period ensured inclusion of only incident cancer cases and reduced protopathic bias. MHT exposure was defined as use of MHT for more than 6 months between 2002 and 2011. Non-users were those who had no MHT prescriptions between 2002 and 2019. The follow-up period extended until a diagnosis of kidney (C64) or bladder cancer (C67), death, or the study end date (December 31, 2019). This structure incorporates core elements of a target trial—including eligibility criteria, treatment strategies, time zero alignment, outcome definition, and censoring—enhancing causal interpretability of the findings. Women with kidney cancer were defined as those who visited a medical institution three or more times with the diagnostic code for kidney cancer (C64), while women with bladder cancer were defined as those who visited a medical institution three or more times with the diagnostic code for bladder cancer (C67). Outcomes were defined based on at least three registrations of diagnostic codes or medical institution visits to ensure the accuracy and confirmation of cancer diagnoses. Cases with only one or two registrations were excluded from the analysis to reduce the risk of false positives and misclassification. MHT formulations were limited to cases where tibolone, combined oestrogen plus progestin by the manufacturer (CEPM), oral oestrogen alone, combined oestrogen plus progestin by the physician (CEPP), and topical oestrogen were prescribed. Supplementary Table 1 presents a detailed list of medications. Patients who sequentially received more than two MHT formulations were assigned to the subgroup of MHT formulation that was last used for more than 6 months. Patient characteristics, such as age, body mass index (BMI), Charlson comorbidity index (CCI), socioeconomic status (SES), region of residency, reproductive factors including parity, age at menarche and at menopause, and the period from menopause to study participation date, and habits, such as smoking, alcohol consumption, and physical exercise, were reviewed. The Asia-Pacific guidelines were followed when measuring BMI, 15 and a low SES was defined as the subject having received medical aid as medical insurance. An urban area was defined as an administrative district that was a large city. The CCI was calculated using the diagnostic codes for diseases that were the cause of visits to medical institutions within 1 year before the participation date in our study. 16 Smoking history was divided into “never,” “past,” and “current,” and drinking history was classified according to the number of alcoholic drinks per week. Exercise strength was classified based on the frequency of exercise lasting 30 minutes or more per week. During the data analysis, any missing data were carefully addressed to minimize bias. We performed a complete case analysis after excluding participants with missing data for key covariates. The proportion of missing data was low across all variables (less than 5%). Given the minimal overall extent of missingness, we did not apply any imputation procedures. The initiation date of treatment in the MHT and non-MHT groups was defined as the date when MHT was first prescribed and at which a health examination was performed, respectively. All continuous variables are described as medians (25th, 75th percentile), while all categorical variables are described as numbers (percentage). A Cox proportional hazard analysis was used to determine the risk of urinary tract cancer. In the Cox proportional hazard analysis, we included age, region, BMI, SES, CCI, parity, age at menarche, age at menopause, smoking, alcohol consumption, physical exercise, and period from menopause to inclusion as covariates to adjust for potential confounders. These adjustments allowed us to isolate the effect of MHT formulations on urologic cancer incidence. The proportional hazards assumption was evaluated using Schoenfeld residuals, and no violations were detected for the included variables. The last day of follow-up was defined as the date of death or the last confirmed date in the health insurance data. To confirm the robustness of this study, only cases wherein an obstetrician/gynaecologist prescribed MHT were selected and analysed. All statistics were two-tailed, and statistical significance was set at a P value of < 0.05. If a value was missing, the listwise deletion method was applied. All statistical analyses in this study were conducted using SAS Enterprise Guide version 6.1 (SAS Institute, Inc., Cary, NC, USA). This study was approved by the Institutional Review Board (IRB) of Sanggye Paik Hospital, Inje University (IRB No. 2020-08-002) for analysis using HIRA data. According to HIRA policy, research data are released only after IRB approval. All data were anonymized and analyzed via a secure virtual server provided by the NHIC, with no access to personal identifiers. Informed consent was waived under the Bioethics and Safety Act, and HIRA holds no responsibility for the study results.

Among the 2,506,271 women who underwent national health examinations between 2002 and 2011 and after excluding those with urinary tract diseases or cancer, 557,031 and 204,130 were assigned to the non-MHT and MHT groups, respectively ( Fig. 1 ). MHT = menopausal hormone therapy. The mean age and BMI were 56 (52–62) years and 23.8 (21.9–25.9) kg/m 2 , respectively. The follow-up periods for the non-MHT and MHT groups were 12.2 years and 13.4 years, respectively. Among the MHT group, 104,089 women were on tibolone, 65,597 on CEPM, 29,357 on oral oestrogen alone, 3,913 on CEPP, and 1,174 on topical oestrogen. Table 1 shows the detailed characteristics of all participants in this study. We compared the characteristics of women in the MHT group and the non-MHT group, including age, BMI, CCI, SES, smoking history, alcohol consumption, and physical activity. Additionally, Supplementary Table 1 provides detailed information on reproductive factors, including parity, age at menarche, age at menopause, and the period from menopause to inclusion. The analysis revealed that there were no statistically significant differences between the two groups across these characteristic. Data are expressed as the number (%) or median (25th percentile, 75th percentile). MHT = menopausal hormone therapy, BMI = body mass index, CCI = Charlson comorbidity index, SES = socioeconomic status. Table 2 presents HRs for MHT and included covariates from multivariable Cox proportional hazards models. The covariates—such as age, BMI, region, SES, smoking, alcohol consumption, physical activity, and reproductive variables—were incorporated solely for adjustment to account for potential confounding. These variables were not the primary exposure of interest and should not be interpreted as independent risk factors. Rather, their inclusion strengthens the robustness of the MHT effect estimate. Similarly, reproductive variables—such as parity and menopausal timing—were included in Supplementary Table 2 for additional adjustment and are not considered risk predictors in this context. Covariates were included for adjustment purposes. Their HRs are shown for transparency but should not be interpreted as causal or independent risk factors. HR = hazard ratio, CI = confidence interval, BMI = body mass index, SES = socioeconomic status, CCI = Charlson comorbidity index. a HRs were adjusted for age group, BMI, SES, region, CCI, parity, age at menarche, age at menopause, smoking, alcohol, physical exercise, period from menopause to inclusion. Kidney cancer occurred in 997 (0.2%) women in the non-MHT group. Meanwhile in the MHT group, kidney cancer occurred in 190 (0.2%), 106 (0.2%), 75 (0.3%), 4 (0.1%), and 7 (0.6%) women treated with tibolone, CEPM, oral oestrogen alone, CEPP, and topical oestrogen, respectively. Among the women in the non-MHT group, bladder cancer occurred in 910 (0.2%) patients. Meanwhile in the MHT group, bladder cancer occurred in 104 (0.1%), 60 (0.1%), 35 (0.1%), 5 (0.1%), and 5 (0.4%) women treated with tibolone, CEPM, oral oestrogen alone, CEPP, and topical oestrogen, respectively. The median duration of hormone therapy in the MHT group was 23 (10–55) months. Additionally, 77.2% of women in the MHT group received hormone therapy within 5 years after menopause ( Supplementary Table 3 ). In Table 3 , HRs for urologic cancer incidence associated with different MHT formulations were adjusted for significant variables, including age, region, BMI, SES, CCI, and lifestyle factors (smoking, alcohol consumption, and physical activity). These adjustments ensure that the associations between MHT use and cancer risk are independent of potential confounding factors. HRs were adjusted for age group, body mass index, socioeconomic status, region, Charlson comorbidity index, parity, age at menarche, age at menopause, smoking, alcohol, physical exercise, period from menopause to inclusion. HR = hazard ratio, CI = confidence interval, MHT = menopausal hormone therapy. In the Cox proportional hazard analysis that adjusted for variables, the incidence of kidney cancer was significantly increased in women treated with oral oestrogen alone (HR, 1.36; 95% confidence interval [CI], 1.062–1.735) and topical oestrogen (HR, 2.84; 95% CI, 1.270–6.344) ( Fig. 2 ). However, the incidences of kidney cancer among women treated with tibolone, CEPM, and CEPP were not different ( Table 3 ). HR = hazard ratio, CI = confidence interval. Notably, tibolone significantly decreased the incidence of bladder cancer (HR, 0.69; 95% CI, 0.548–0.858) ( Fig. 2 ), whereas other formulations did not affect the incidence of bladder cancer ( Table 3 ). In the dose-dependent analysis, tibolone had no preventive effect on bladder cancer when 1.25 mg (half-dose) was given ( Supplementary Table 3 ).

Our study reveals that the use and type of MHT is associated with kidney cancer risk among postmenopausal women; oral and topical oestrogen formulations are significantly associated with the incidence of kidney cancer in women on MHT. However, other types of MHT were not significantly associated with the incidence of kidney cancer. Previous studies revealed that the risk of kidney cancer in postmenopausal women differs depending on the type of MHT. Among the types of MHT, studies that described the association of oestrogen alone with kidney cancer showed conflicting results. 3 4 5 Our study’s findings align with some studies but differ from others. Several potential factors may explain these differences. For instance, population differences, particularly in terms of ethnicity and geography, may result in variations in baseline cancer risks and MHT usage patterns. Additionally, differences in study design, MHT formulations, and follow-up durations could account for the conflicting results across studies. Molokwu et al. 5 suggested that the risk of kidney cancer was significantly increased in women using oestrogen alone. Similarly, McCredie and Stewart 6 revealed that women using oestrogen alone had a significantly increased risk of kidney cancer. Additionally, Zhang et al. 17 performed a meta-analysis of previously reported case–control and cohort studies and demonstrated that oestrogen alone is associated with an increased risk of kidney cancer, which is consistent with our findings. Two other studies reported on the relationship between the risk of kidney cancer and combined oestrogen plus progestin; however, the relationships were nonsignificant. 3 4 Setiawan et al. 3 demonstrated that among 106,036 women, those on combined oestrogen plus progestin therapy had a nonsignificant 27% increase in the risk of kidney cancer. Meanwhile, among a population of 118,219 United States nurses, another study reported that women using combined oestrogen plus progestin had a nonsignificant decrease in the risk of kidney cancer. 4 The underlying mechanisms regarding oestrogen therapy affecting the risk of kidney cancer remain unclear; however, several studies have proposed potential mechanisms. Several studies have proposed that modulating hormone levels and receptor activities in kidney tissue may partially influence the risk of kidney cancer. Tanaka et al. 18 discovered that oestrogen and progesterone receptors in women are expressed in both normal and cancerous kidney tissue, suggesting that hormonal changes could affect kidney cancer progression. Moreover, several experimental studies have suggested that the estrogenic effect is related to the occurrence of renal cell tumours in animals. 19 Oestrogen alone is only prescribed for patients who have underwent a hysterectomy. The most common indications for a hysterectomy are uterine fibroids and adenomyosis. Therefore, there might be an association between uterine fibroids, adenomyosis, hysterectomies, and kidney cancer. Zucchetto et al. 20 also reported an increased incidence of kidney cancer in patients who underwent a hysterectomy. In the analysis that adjusted for variables, the elevated incidence of kidney cancer was associated with older age, obesity, and current smoking status, which is consistent with previously known factors. 12 Obesity is associated with an increased risk of kidney cancer. 12 Additionally, previous studies demonstrated that smoking is a significant risk factor for kidney cancer, 12 21 with men and women who smoke having a 1.54- and 1.22-times higher risk, respectively, than non-smokers. 21 Moreover, our analysis found the incidence of kidney cancer was lower in women living in rural areas. Sadowski et al. 22 discovered that residents in rural areas had a lower incidence of kidney cancer but a higher mortality from the disease than those in urban areas, which is consistent with our findings and may be attributed to a disparity in available diagnostic utilization and hospital access between rural and urban areas. Tibolone significantly decreased the incidence of bladder cancer, whereas other MHT formulations were not significantly associated with the incidence of bladder cancer. However, although it was not significant, CEPM decreased the incidence of bladder cancer, which is consistent with previous studies. 7 10 Research has suggested that MHT for postmenopausal women affects the risk of bladder cancer, but the results have been inconsistent. Several studies have discovered no significant association between MHT and bladder cancer; however, some cohort studies have reported that the risk of bladder cancer may vary depending on the MHT formulation. 7 8 9 10 Studies have discovered that oestrogen alone was not significantly associated with an increased risk of bladder cancer; however, there may still be an association, albeit nonsignificant. 7 8 9 Daugherty et al. 10 reported on an inverse association between combined oestrogen plus progestin and bladder cancer, and Davis-Dao et al. 7 discovered similar results. However, a meta-analysis by Xu et al. 23 revealed that combined oestrogen plus progestin was negatively but not significantly associated with bladder cancer. Additionally, McGrath et al. 9 reported that women on combined oestrogen plus progestin therapy had a lower risk of bladder cancer compared with women not using this therapy, although it was not significant. However, a 2020 study with longer observation periods and more participants conducted by the same researcher observed no inverse association between combined oestrogen plus progestin and the risk of bladder cancer. The differences in findings across studies may be attributed to several factors. Differences in the types and formulations of MHT used, variations in the baseline characteristics of the study populations, and methodological differences, such as follow-up durations and data collection methods, could all contribute to the observed discrepancies. The exact mechanisms underlying the results of previous observational studies are unknown; however, previous experimental research discovered that sex hormonal signalling plays an important role in the incidence and progression of bladder cancer. 24 An underlying mechanism could be that sex hormones have important physiological effects in maintaining the structures and functions of the female lower urinary tract. 25 The female genital and lower urinary tracts have a common embryonic origin arising from the urogenital sinus; therefore, significant alterations of sex hormone levels, such as during pregnancy and menopause, can cause significant modifications in the lower urinary tract, causing various symptoms, including urgency, incontinence, and recurrent urinary tract infections. 26 Moreover, frequent and recurrent urinary tract infections may be followed by chronic irritation of the bladder epithelium, resulting in a significantly increased risk of bladder cancer. 27 Yu et al. 28 demonstrated that sex hormones are essential in maintaining the bladder’s structure and that both oestrogen and androgen can reverse bladder muscle atrophy caused by ovariectomy. Similarly, Yang et al. 29 suggested that deprivation of sex hormones negatively affects the bladder structure and histology, and that oestrogen administration can preserve bladder function by inhibiting collagen hyperplasia and increasing smooth muscle density. In our study, tibolone significantly decreased the incidence of bladder cancer. There have been few studies on the association between tibolone use for postmenopausal women and the risk of bladder cancer. Tibolone has estrogenic, progestogenic, and androgenic effects with differential metabolism in each tissue, making it unique for treating menopausal symptoms. 30 The estrogenic effects of tibolone are primarily expressed in the vagina, bone tissue, and brain, with less expression in the endometrium. Considering the results of previous research, selective tissue-specific effects of various sex hormones through tibolone are thought to maintain the bladder structure and function, which may prevent urinary tract symptoms and infections and thus significantly lower the risk of bladder cancer, which is consistent with our findings. In our study, older age and current smoking status significantly increased the incidence of bladder cancer, which is consistent with previously known factors. 13 In contrast, living in rural areas and parity significantly decreased the incidence of bladder cancer. Previous studies have shown that smoking is the most important risk factor for developing bladder cancer, with a 50% increased risk for smokers. 13 Similar to kidney cancer, several studies have discovered that women in rural areas have a lower incidence of bladder cancer than those in urban areas, 31 which is consistent with our results. Previous studies on the relationship between reproductive factors and the risk of bladder cancer did not show consistent results; however, several studies report that parity was negatively associated with the risk of bladder cancer. 7 8 10 Pregnancy is associated with dramatic changes in oestrogen and progesterone levels, which may persist for months. However, the exact mechanisms by which oestrogen and progesterone influence the risk of bladder cancer later in life remain unclear. In our study, other reproductive factors, such as age at menarche and at menopause, BMI, low SES, and physical exercise were not associated with the incidence of bladder cancer. Our study has several strengths. First, our study included over 1 million postmenopausal women; the number of participants in the study is comparable to that of the existing observational studies. Second, we used tibolone as the main drug for MHT. In previous studies, tibolone was not a mainstream prescription drug; however, it is the most prescribed medication in Korea. Third, we investigated most of the various combinations of drugs used for MHT, including several CEPMs (Angeliq ® , Climen ® , Clian ® , and Femoston ® ), providing a robust understanding of the potential interactions and their impact on the risk of kidney and bladder cancers. Supplementary Table 4 shows detailed information on the MHT drugs included in this study. Finally, the study’s analysis was adjusted for age, BMI, SES, CCI, smoking and drinking history, physical exercise, and reproductive factors such as parity, age at menarche and at menopause, and the period from menopause to study inclusion. This study has some limitations. First, we could not confirm the detailed CEPM drug list associated with the NHIC policy. Second, although we adjusted for several potential confounders including age, BMI, SES, and lifestyle factors, residual confounding may persist. In particular, detailed information on occupational exposures, second-hand smoking, or family history of cancer was unavailable in the dataset. Third, we could not perform a medical record review because the information provided by the HIRA dataset is categorized based on prescription and diagnostic codes. Forth, while a dose-dependent analysis for tibolone was conducted and showed no preventive effect on bladder cancer at a 1.25 mg (half-dose) level, we did not comprehensively analyze the cumulative dose or duration of other MHT formulations to investigate potential dose-response relationships. This represents a limitation of the study, and future research should consider exploring cumulative exposure to MHT to better understand its long-term effects. Fifth, we did not have detailed information on hysterectomy indications (benign vs. malignant) or oophorectomy status in this cohort. Since oophorectomy often accompanies hysterectomy, and given that women receiving oestrogen-only therapy are likely to have undergone hysterectomy, this may have influenced the observed association between oestrogen-only therapy and renal cancer risk. The lack of this surgical information represents a limitation of the study, and future research should consider incorporating these factors to better understand the relationship between hormone therapy and cancer risk. Sixth, because the HIRA database records malignant bladder tumors mainly under the C67 code, some very early non-invasive (Ta or Tis) cases coded as D09.0 might not have been included. However, given the coding practice in Korea, the C67 code captures the vast majority of clinically relevant bladder cancers, and this limitation is unlikely to affect the overall association patterns observed. Given the limitations of this study, future research should investigate the potential dose-response relationship of cumulative MHT exposure and explore the differential effects of various MHT formulations. Additionally, understanding how demographic factors, surgical histories, and specific cancer risk profiles influence the relationship between MHT and urologic cancer will be essential. This will provide more nuanced insights into the long-term impacts of MHT use. In conclusion, this study observed significant associations between certain MHT formulations and the incidence of kidney and bladder cancers. While these findings add valuable information to the existing literature, further investigation is needed to fully understand the mechanisms underlying these associations, given the observational nature of this study. Considering the study’s large population size of over 1 million, we expect the results to be considered when treating postmenopausal women with MHT. Additional studies are needed to investigate the possibility of other unknown confounding factors.