Hysterectomy, Oophorectomy, and Risk of Renal Cell Carcinoma

Renal cell carcinoma (RCC) is the most common type of kidney cancer, accounting for about 90% of all cases (1). The incidence rates of RCC have been increasing since 1975 (2), although the increase appears to have slowed or leveled off in recent years and the overall incidence of RCC remains at 11.3 per 100,000 person-years (3). Some of the increase, such as the increase in RCC diagnosed at localized stage, may have resulted from increases in advanced imaging. However, the highest increase was seen in grade III tumors—a more aggressive disease, indicating that the increase may also be attributable to changes in other risk factors (2). Despite improvements in diagnosis and management during the last two decades, RCC remains one of the most lethal urological malignancies with an estimated 14,830 deaths/year in the United States (4). Well-established risk factors for RCC include sex, obesity, smoking, and hypertension (4).

Although RCC incidence rates are higher in men than in women worldwide, this difference disappears with advancing age, from a ratio of 2:1 at ages 41–60, to almost 1:1 for adults older than 70 (5–8). It has been proposed that higher risk of RCC among men may be related to smoking and occupational exposures (9). Female hormones, including factors that may affect their levels, such as reproductive, and menstrual factors and exogenous hormone use, may also play protective roles. Studies have shown that estrogen-activated estrogen receptor β acts as a tumor suppressor, and this may contribute to the different RCC incidence rates between men and women (10). However, epidemiologic evidence on reproductive and hormonal factors and risk of RCC is inconsistent (4). For example, parity has been reported to be associated with an increased risk of RCC among women in several cohort studies (11–13), but not in other studies (14–16). Similarly, associations are not consistently observed for other reproductive and hormonal factors (4).

Of note, epidemiologic studies have suggested an increased risk of RCC after hysterectomy (17). However, the evidence is mostly derived from case–control studies (15, 18–20). Among six cohort studies, four (with a total number of cases < 250) did not observe any association between hysterectomy and RCC (13, 14, 21, 22), but two large studies observed a significant positive association with relative risks in the range of 1.28–1.50 (16, 23). A Swedish study based on registry databases observed higher risk of RCC after hysterectomy (HR, 1.50; 95% CI, 1.33–1.69) and a more pronounced association in women with hysterectomy at age 44 years or younger (HR, 2.36; 95% CI, 1.49–3.75; ref. 23). However, the study lacked information on oophorectomy and was unable to adjust for major confounders, such as body mass index (BMI) and smoking. Hysterectomy is one of the most commonly performed surgical procedures in US women, undertaken in more than one in every three women by age 60 years (24). Also, bilateral salpingo-oophorectomy (BSO) is commonly performed at the time of hysterectomy (25). Therefore, it is of public health importance to clarify the associations between hysterectomy with and without oophorectomy and risk of RCC. Further, hormone therapy, commonly used among women who have undergone a hysterectomy, might modify the association between hysterectomy and risk of RCC (26, 27). Thus, it is important to examine the association between hysterectomy, oophorectomy status, and risk of RCC stratified by exogenous hormone use to shed light on the possible impact of endogenous or exogenous hormones on the associations.

In the present study, we used the Women's Health Initiative (WHI), a large prospective study in the United States, with detailed information on potential confounders and centrally adjudicated RCC cases, to investigate associations between hysterectomy, oophorectomy, and risk of RCC. In addition, we examined the association between hysterectomy and risk of RCC by age at time of the surgical procedures and by time since hysterectomy. Our study is the first to examine the association by tumor stage. We further tested whether exogenous estrogen use was an effect modifier of the associations between hysterectomy, BSO, and risk for RCC.

The WHI is a large prospective study, designed to address the major causes of morbidity and mortality among postmenopausal women in the United States (28). Details of the study's design and recruitment are described elsewhere (29). Briefly, 161,808 women ages 50 to 79 were recruited from 40 clinical centers throughout the United States between 1993 and 1998. The WHI includes four overlapping clinical trials (CT; including estrogen-alone trial, estrogen plus progestin trial, dietary modification trial, and calcium and vitamin D trial) and an observational study (OS). All participants in the WHI-OS were followed annually. CT participants were followed-up every 6 months through 2005 and annually thereafter. After the protocol-specified clinical trial completion date of March 31, 2005, subsequent outcome assessment required re-consent, which was obtained from 82% of surviving participants for follow-up through 2010 and 86% of surviving participants for further follow-up. The baseline characteristics were similar by re-consent status with slightly higher participation by white women compared with other ethnic groups (30). The WHI study was approved by institutional review boards at all 40 clinical centers and at the clinical coordinating center. All participants provided written informed consent.

The following participants were excluded from the WHI cohort of 161,808 for this analysis: 12,655 women who had a history of cancer (except nonmelanoma skin cancer) at baseline; 635 who had no follow-up information; 1,553 women who had missing information on the main exposures of hysterectomy or age of hysterectomy or oophorectomy, collected at enrollment; and 2,366 women with missing information on major covariates, including hormone use, BMI, history of diabetes, history of hypertension, and diet quality. After exclusions, 144,599 women remained for further analysis.

Surgical histories of hysterectomy, oophorectomy, and age at surgery were collected through self-reported questionnaires at WHI enrollment.

Hysterectomy status at baseline was determined by asking the following questions: “Did you ever have a hysterectomy? (This is a surgery to take out your uterus or womb.)” Oophorectomy status at baseline was based on a question “have you ever had an operation to remove one or both of your ovaries.” The response to this question was categorized as no; yes, one was taken out; yes, both were taken out; yes, part of an ovary was taken out; yes, unknown number taken out; and don't know. Age at hysterectomy and age at oophorectomy were also collected at baseline by responding to predefined categories: <30 years old, 30–34, 35–39, 40–44, 45–49, 50–54, 55–59, 60, or older. We also estimated time since hysterectomy until the end of follow-up by using age at baseline, subtracting age at hysterectomy (a median age was assigned for each specific age category), and then adding follow-up time since enrollment.

Information on hysterectomy status during follow-up was also collected by asking the question: “Since your last medical update, which of the following exams, tests, or procedures have you had?” Information on hysterectomy during follow-up was based on response to an option of “removal of the uterus or womb (hysterectomy).” Oophorectomy status during follow-up was not collected. The validity of self-reported hysterectomy and oophorectomy has been previously confirmed with sensitivity of 91% and positive predictive value of 97% for hysterectomy status and sensitivity of 73% and positive predictive value of 100% for oophorectomy status (31).

The outcome was incidence of RCC. In the WHI, participants initially self-reported cancer diagnoses. All self-reported cancer cases were adjudicated by obtaining medical records and pathology reports, and all records were adjudicated by centrally trained physicians and coded in accordance with the Surveillance Epidemiology and End Results (SEER) coding guidelines. We restricted our outcome cases to RCC with histology codes (ICD-O-2), including 8032, 8140, 8260, 8270, 8290, 8310, 8312, 8316, and 8317 through 8320; ref. 32).

We considered the following potential risk factors for RCC measured at baseline, including demographics (age, race/ethnicity, education), family history of any cancer, lifestyle factors (diet quality, BMI, smoking, physical activity, and alcohol consumption), exogenous hormone use, several medical conditions linked to RCC (hypertension, diabetes, and kidney disease), and participation in different WHI study subcohorts (observational study or CT and different treatment assignments for all 3 CT). Diet quality was derived from the Food Frequency Questionnaire to obtain the Healthy Eating Index (HEI-2015) score based on the 2015 Dietary Guidelines for Americans (33). Physical activity was estimated as metabolic equivalent task (MET)-hours per week. Information on smoking status (never, former, and current), intensity, and duration was collected by self-administered questionnaires. For former and current smokers, pack-years of smoking were derived by multiplying the total years of smoking by the number of cigarettes smoked per day divided by 20. Information on lifetime use of menopausal hormones was obtained at baseline using structured questionnaires and charts displaying colored photographs of various hormone preparations.

Our primary analysis focused on the associations of RCC with hysterectomy and oophorectomy status at baseline, because the WHI did not collect information on oophorectomy status during follow-up. Hysterectomy and oophorectomy status was also combined into a single variable with six categories: no hysterectomy and no bilateral oophorectomy, no hysterectomy but with one or partial oophorectomy, no hysterectomy but bilateral oophorectomy, hysterectomy without bilateral oophorectomy, hysterectomy with one or partial oophorectomy, and hysterectomy with bilateral oophorectomy.

Baseline characteristics were summarized using percentages for categorical variables and means (SD) for continuous variables. Differences by hysterectomy status (yes or no) were tested using chi-square tests for categorical variables and t tests for continuous variables.

Multivariable Cox proportional hazards models were used to assess the associations between baseline hysterectomy/oophorectomy status, age at hysterectomy, and RCC risk. All models were stratified by study cohort (participation in OS or CTs, and different treatment assignments for all four CTs). We used time on the study as the time scale for the Cox models. Survival time was defined from enrollment to date of diagnosis of RCC, date of death, loss to follow-up, or March 1, 2019, whichever came first. Updated hysterectomy status and age at hysterectomy were analyzed as time-varying variables using time-dependent covariate Cox models. In all multivariable models, potential confounders were considered including age, race/ethnicity, BMI, education, diet quality, smoking pack-years, alcohol intake, physical activity, parity, history of estrogen-alone use, history of estrogen plus progestin use, family history of cancer, diabetes, hypertension, and kidney. Missing data were included in the regression models as a separate category for the respective variable. Details on these variables are listed in Table 1.

We explored potential effect modification of the association between hysterectomy and oophorectomy at baseline and risk of RCC by major risk factors for RCC, including hypertension, exogenous hormone use, BMI, smoking status, and race/ethnicity. Also, we examined whether time since hysterectomy was associated with risk of RCC. Finally, we performed several sensitivity analyses: (i) to examine whether intervention of exogenous estrogen-alone was associated with risk of RCC in the WHI hormone trial among women with hysterectomy; (ii) to examine whether associations between hysterectomy and risk of RCC differ by tumor stage; (iii) to examine association between hysterectomy and risk of RCC by excluding women in the intervention arm of the estrogen-alone trial or in the estrogen plus progestin trial; (iv) to examine the association by excluding the first two years of follow-up; (v) to examine the association by excluding non-clear cell RCC; and (vi) to examine the association by using attained age instead of time on study as the time scale. SAS 9.4 was used for all the analyses. All statistical significance tests reported were two-tailed. P values ≤ 0.05 were considered statistically significant.

At baseline, 40.4% of participants reported a history of hysterectomy. Among women with hysterectomy, 45.1% reported that they had had a BSO. Compared with women with no hysterectomy, women with hysterectomy were more likely to be older, minorities, less educated, nonsmokers, have higher BMI, be physically inactive, have poor diet quality intake, and to be never or past alcohol drinkers. They were also more likely to report a family history of cancer and a history of hormone use, and they had a higher prevalence of diabetes, hypertension, and kidney or bladder stones (Table 1).

Over a mean of 15.9 years of follow-up, 583 women developed RCC. In the multivariable-adjusted models, hysterectomy at baseline was borderline significantly associated with an increased risk of RCC. When updated hysterectomy during follow-up was considered, hysterectomy was significantly associated with an increased risk of RCC (HR, 1.28; 95% CI, 1.01–1.60). Younger age (<40 years) or older age at hysterectomy (≥ 55) appeared to be associated with higher risk of RCC (HR, 1.35; 95% CI, 1.01–1.80 and HR, 1.41; 95% CI, 1.04–1.91, respectively) compared with women without hysterectomy (Table 2) when age at hysterectomy was analyzed as a time-varying variable. Results were similar when age at hysterectomy at baseline was analyzed. There was no statistically significant linear trend for age at hysterectomy (restricted to women with hysterectomy; P for trend = 0.96).

Oophorectomy was not associated with risk of RCC in the multivariate-adjusted models irrespective of whether participants reported a partial oophorectomy, or removal of one or two ovaries. When hysterectomy and oophorectomy were analyzed jointly, compared with women without hysterectomy and oophorectomy, women with hysterectomy alone had a borderline increased risk of RCC (HR, 1.31; 95% CI, 0.99–1.73; Table 2).

Table 3 shows the association between hysterectomy and RCC risk after stratification by hypertension, hormone therapy use, BMI, smoking status, and race/ethnicity. The association between hysterectomy and risk of RCC appeared more pronounced among women without hypertension (P value for interaction = 0.12), among women who were past smokers (P for interaction = 0.51), and among non-Hispanic white women (P for interaction = 0.24). However, none of the tests for interaction reached statistical significance (Table 3).

In further analysis, we assessed the association between age at hysterectomy and RCC risk stratified by history of hypertension, estrogen-alone use, and race/ethnicity (Table 4). We observed that a more pronounced increased risk of RCC was associated with younger age at hysterectomy among women without hypertension, and a more pronounced increased risk was associated with older age at hysterectomy among women with hypertension (P value for interaction = 0.17). No significant interaction was observed between age at hysterectomy and estrogen-alone use or race/ethnicity (Table 4).

Finally, we observed that risk of RCC did not vary significantly by time since hysterectomy [P for trend = 0.41. HR, 1.37 (0.93–2.04) for <10 years, HR, 1.34 (0.97–1.86) for 10–<20 years, HR, 1.17(0.87–1.58) for 20–<30 years, and HR, 1.28 (0.97–1.68) for 30 or more years]. Also, no association was observed between use of estrogen alone and risk of RCC in women with hysterectomy (HR, 1.07; 95% CI, 0.65–1.77) based on the WHI estrogen-alone trial. Results were similar when attained age instead of time on study was used as the time scale. Results were also similar after excluding women in the intervention arm of estrogen-alone trial or estrogen plus progestin trial, or after excluding the first two years of follow-up, or after excluding non-clear cell RCC (Supplementary Table S1). However, in the analysis by RCC stage, we observed similar results for localized RCC, but no association for regional or advanced RCC (Supplementary Table S1).

In this large prospective study of postmenopausal women, we observed that women with a history of hysterectomy had an approximately 28% higher risk of RCC relative to women without hysterectomy. The association between age at hysterectomy and risk of RCC appeared to be a U-shaped. Women with earlier age at hysterectomy had a more pronounced risk of RCC in women without hypertension while women with older age at hysterectomy had a more pronounced risk of RCC in women with hypertension, although the interaction test did not reach statistical significance (P = 0.17). Contrary to our original hypothesis, BSO was not associated with risk of RCC. We found no evidence that exogenous estrogen use modified the association between hysterectomy and risk of RCC.

Our findings regarding the association of hysterectomy with risk of RCC are in line with those of Karami and colleagues, based on two large US cohorts (NIH-AARP and PLCO; ref. 16). Similar to us, Karami and colleagues observed that women who had hysterectomy either with or without an oophorectomy had an approximately 30% to 40% increased risk of kidney cancer (16). However, Karami's study reported no difference by age at hysterectomy. A Swedish study based on register databases also observed that hysterectomy was associated with increased risk of RCC and a more pronounced association in women with earlier age at hysterectomy (23). Other cohort studies (13, 14, 21, 22) that reported no association between hysterectomy and risk of RCC included relatively fewer cases (< 250) and were unable to perform in-depth analysis, for example, examining age at hysterectomy or potential effect modification.

Our study was the first to observe that the association between age at hysterectomy and risk of RCC may be a U-shaped, and that the association may depend on hypertension status. We observed a more pronounced association in women with earlier age at hysterectomy (<40 years) and without hypertension, and in women with older age at hysterectomy (≥55 years) and with hypertension. Hypertension is a well-established risk factor for RCC. An elevation in systolic but not diastolic pressure is the most prevalent type of hypertension in those ages 50 or over (34). This form of hypertension may result in chronic renal hypoxia and chronic renal damage that increases RCC risk in interaction with hysterectomy-induced renal damage (35, 36). Furthermore, a study reported that risk of hypertension after hysterectomy increases with age at hysterectomy (37). The more pronounced association with earlier age at hysterectomy may be associated with underlying conditions or genetic predisposition toward hysterectomy at an earlier age (38). In addition, the U-shaped association with age at hysterectomy may be due to residual confounding. For example, those with family history of kidney cancer may develop RCC at a younger age while those with higher chemical exposures (or advanced kidney diseases) may get RCC at an older age. However, we have not measured these potential confounders. Finally, the observed U-shape relationship may be a chance finding. More studies are needed to verify our findings.

Our data did not support the hypothesis that the observed association between hysterectomy and risk of RCC may be modified by hormone therapy. First, our data did not show that the association between hysterectomy and risk of RCC differed by either estrogen-alone use (commonly used among women who have undergone a hysterectomy). Second, although the study power may have been limited, a sensitivity analysis based on the WHI estrogen-alone use clinical trial showed that use of estrogen alone was not associated with risk of RCC among women with hysterectomy.

Our data showed that the association between hysterectomy and risk of RCC is unlikely to be due to major risk factors for RCC, because we were able to adjust for comprehensive potential risk factors for RCC. However, our data could not completely rule out that the increased risk of RCC associated with hysterectomy may be due to residual confounding by indication for hysterectomy. For example, uterine fibromyomas are one of leading indications for hysterectomy in the United States (39). Although we were able to adjust for many confounders that may be related to uterine bleeding, such as race/ethnicity, obesity, and hypertension, there may be residual unmeasured confounding factors that may have affected our findings.

Another possible reason for the association between hysterectomy and risk of RCC may be due to surgically induced renal damage or pelvic anatomy changes after a hysterectomy. Previous studies have reported a high incidence of post-renal obstruction following hysterectomy, and hydronephrosis has been observed in patients after hysterectomy, even without recognized injury to the ureter (16, 35, 40). Urinary tract stones also would be a possible indicator of ureteral obstruction. Our data indicated that prevalence of kidney or bladder stone disease was higher in women with hysterectomy (47.9%) than women without hysterectomy (39.8%).

Finally, one of our sensitivity analyses showed that hysterectomy was associated with risk for localized RCC but not with regional or advanced RCC, although the study power for assessing regional or advanced RCC may have been limited. These results indicate that it is also possible that women with hysterectomy were more likely to undergo later abdominal imaging that could have incidentally detected small indolent tumors. However, if findings reflect on imaging artifact, it should apply to all ages, where our results indicated that the association was more pronounced in women with age at hysterectomy younger than 40 years or older than 55 years.

The strengths of the present study include the prospective design with a large sample size and long follow-up, availability of comprehensive information on potential confounders, and adjudicated outcomes. Other strengths of the present study include that we had updated information on hysterectomy status during follow-up and were able to consider hysterectomy as a time-dependent variable in the model. Further, we were able to explore whether the associations between hysterectomy, age at hysterectomy, and risk of RCC differed by tumor stage or tumor histology, and whether the associations were modified by hormone use, hypertension, or other potential effect modifiers. However, several limitations also need to be acknowledged. First, misclassification due to self-reported information on hysterectomy status is possible. Such misclassification, however, most likely would be nondifferential, which may have biased estimates toward the null. Second, there may have been detection bias due to more intensive medical surveillance in more recent years after hysterectomy, although the association between hysterectomy and risk of RCC did not change substantially by the time since hysterectomy. Third, information on types of hysterectomy (such as vaginal or abdominal hysterectomy, laproscopic) or underlining indications for hysterectomy (such as uterine fibroids, endometriosis, ovarian tumors, or uterine prolapse) was not collected. Finally, our study only included postmenopausal women, along with relatively low proportions of minority groups other than black women. We were unable to assess whether the findings would differ among younger women or other minority groups. Thus, similar study including premenopausal women and other minorities is needed.

In conclusion, our large prospective study showed that women with a history of hysterectomy had an approximate 28% increased risk of RCC. Because hysterectomy is one of the most commonly performed surgical procedures among women in the United States, this finding merits further study. If confirmed, women should be made aware of increased risk of RCC when considering hysterectomy. Our data further indicate that the association may not be fully explained by common hysterectomy risk factors or by exogenous hormone use, which is common among women who have undergone a hysterectomy. BSO was not associated with risk of RCC. Future studies with detailed data on types of hysterectomy and underlying indication for hysterectomy may be helpful to elucidate this association.

No disclosures were reported.

J. Luo: Conceptualization, formal analysis, methodology, writing–original draft. T.E. Rohan: Writing–review and editing. M.L. Neuhouser: Writing–review and editing. N. Liu: Writing–review and editing. N. Saquib: Writing–review and editing. Y. Li: Writing–review and editing. A.H. Shadyab: Writing–review and editing. L. Qi: Writing–review and editing. R.B. Wallace: Writing–review and editing. M. Hendryx: Conceptualization, writing–review and editing.

The WHI program is funded by the National Heart, Lung, and Blood Institute, NIH, U.S. Department of Health and Human Services through contracts HHSN268201600018C, HHSN268201600001C, HHSN268201600002C, HHSN268201600003C, and HHSN268201600004C. A short list of WHI investigators is in a Supplementary file.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.