Hypertension and Obesity and the Risk of Kidney Cancer in 2 Large Cohorts of US Men and WomenNovelty and Significance
Kidney cancer incidence is increasing globally. Reasons for this rise are unclear but could relate to obesity and hypertension. We analyzed longitudinal relationships between hypertension and obesity and kidney cancer incidence in 156 774 participants of the Women’s Health Initiative clinical trials and observational studies over 10.8 years. In addition, we examined the effect of blood pressure (BP) on kidney cancer deaths for over 25 years among the 353 340 men screened for the Multiple Risk Factor Intervention Trial (MRFIT). In the Women’s Health Initiative, systolic BP (SBP) was categorized in 6 groups from <120 to >160 mm Hg, and body mass index was categorized using standard criteria. In age-adjusted analyses, kidney cancer risk increased across SBP categories (P value for trend <0.0001) and body mass index categories (P value for trend <0.0001). In adjusted Cox proportional hazards models, both SBP levels and body mass index were predictors of kidney cancer. In the MRFIT sample, there were 906 deaths after an average of 25 years of follow-up attributed to kidney cancer among the 353 340 participants aged 35 to 57 years at screening. The risk of death from kidney cancer increased in a dose–response fashion with increasing SBP (hazard ratio, 1.87 for SBP>160 versus <120 mm Hg; 95% confidence interval, 1.38–2.53). Risk was increased among cigarette smokers. Further research is needed to determine the pathophysiologic basis of relationships between both higher BP and the risk of kidney cancer, and whether specific drug therapies for hypertension can reduce kidney cancer risk.
The incidence of kidney cancer is increasing throughout the world among all age groups, races, and for all tumor sizes.1,2 Metastatic kidney cancer is one of the most treatment-resistant malignancies with a 5-year relative survival rate of 12.3% at time of diagnosis.3
Risk factors for kidney cancer include hypertension, increased body weight, and smoking.4–8 Hypertension has been found to increase risk for kidney cancer in numerous prospective studies.9–12 However, the duration of follow-up for many of these studies is short. Several studies have lacked measured blood pressure (BP) values.2,7 The few studies that have included data on women reported increased risk of kidney cancer with higher levels of BP.2,9,10,13 We previously reported a positive association between BP levels and kidney cancer in men >16 years in the Multiple Risk Factor Intervention Trial (MRFIT).14
Excess body weight has also been recognized as a risk factor for kidney cancer in general, and specifically in women,2,15–17 with overweight or obesity estimated to be causally related to >25% of US kidney cancer cases.5,16
Furthermore, few studies have examined obesity and hypertension together as risk factors for kidney cancer. Obesity is a risk factor for hypertension, thus they may represent a shared causal mechanism.2 One prospective US study identified a relative risk of 2.82 (95% confidence interval [CI], 1.97–4.02) for all kidney cancer cases in individuals who were both hypertensive and obese when compared with their lean normotensive counterparts.2 Another analysis, which simultaneously stratified by body mass index (BMI) and BP categories, found that BMI showed only a nonsignificant effect on renal cell carcinoma incidence when BP was markedly elevated.10
We examined the relationship between levels of hypertension and degree of obesity as risk factors for kidney cancer both separately and in relationship to one another over 10.8 years in a large racially diverse female population. We complemented these analyses with an evaluation of the long-term associations among BP, cigarette smoking, and kidney cancer–specific mortality in men, examining 25 years of follow-up in the 353 340 men screened for the MRFIT.18
The Women’s Health Initiative (WHI)19,20 and MRFIT18 are the 2 largest studies of the relationship of BP levels to risk of kidney cancer that include both the actual measurements of BP and the long-term follow-up to account, in part, for the incubation period from BP measurement to development of kidney cancer. The WHI provides more detailed information on measures, such as obesity, BMI, waist circumference (WC), drug therapy, and risk of kidney cancer. The MRFIT includes a larger sample of cigarette smokers and longer term follow-up to kidney cancer.
The WHI Observational Study (WHI-OS) and WHI Clinical Trial (WHI-CT) studies conducted recruitment across 40 US clinical centers, enrolling a combined total of 161 856 women aged 50 to 79 years from September 1993 to December 1998. Concentrated efforts were made to enroll a racially and ethnically diverse sample. All participants underwent a baseline physical examination and longitudinal follow-up with scheduled exams and questionnaire completion, as detailed in previous publications.19,20 The WHI provides detailed information on BMI, WC, drug therapy, and development of kidney cancer.
MRFIT methods have been described previously.18 The MRFIT screened 361 662 men aged 35 to 57 years at 22 clinical centers throughout the United States between 1973 and 1975. Information recorded from each participant included date of birth, Social Security number, race, cigarette smoking status, the use of medication for diabetes mellitus, and history of hospitalization for myocardial infarction. All participants of each study signed consent forms for participation, which were approved by the institutional review boards of collaborating institutions.
Inclusion Criteria and Measurements: WHI
We included women of the WHI-CT and WHI-OS with baseline height and weight measurements and self-reported race/ethnicity. We excluded women with a baseline BMI<18.5 kg/m2 (n=1383) because this group may include women with low BMI because of other health conditions. Race/ethnicity was categorized according to baseline self-report. A previous diagnosis of diabetes mellitus, diagnosis and treatment of hypertension, and current smoking status were self-reported. Height, weight, WC, and BP levels were assessed during baseline WHI examinations. Kidney cancer was assessed from baseline to August 2009. Cancer diagnoses were ascertained through questionnaire responses, and review of medical records that were locally (n=318) or centrally (n=97) adjudicated.
Analyses of Hypertension, Body Weight, and Kidney Cancer Incidence: WHI
We used descriptive statistics to examine the distribution of race/ethnicity, BP, BMI, WC, smoking status, and diabetes mellitus prevalence at baseline among women with and without incident kidney cancer. In unadjusted analyses, we assessed for trends of cancer incidence across categories of BP, BMI, and WC.
Cox proportional hazards modeling assessed whether systolic SBP (SBP) and BMI categories were independently associated with development of incident kidney cancer after adjusting for age, smoking status, race/ethnicity, and diabetes mellitus at baseline. The same analyses were repeated substituting diastolic BP (DBP) for SBP. We then used a similar modeling approach to determine how hypertension and central obesity (WC categories) influenced kidney cancer risk.
Inclusion Criteria and Measurements: MRFIT
For the MRFIT analyses, we include data from screened participants with measured baseline BP values (n=353 340). They included date of birth, Social Security number, serum cholesterol, race, cigarette smoking status, and the use of medication for diabetes mellitus, and history of hospitalization for myocardial infarction. The average of the second and third BP readings, measured with a standard mercury sphyngomanometer, was categorized into 5-mm (diastolic) and 10-mm (systolic) increments, as well as clinical hypertension categories. No information is available on obesity or the use of antihypertensive medications in the MRFIT sample.
Analyses of BP and Kidney Cancer Deaths in Men in MRFIT
We used Cox proportional hazards modeling with stratification by clinical center to determine the risk of mortality from kidney cancer–related to BP categories. We report unadjusted hazard ratios (HRs) and HRs adjusted for age, serum cholesterol, cigarette use, race (black versus nonblack), and diabetes mellitus status.
Because there is a known strong association between kidney cancer and end stage renal disease (ESRD),22 and hypertensive kidney damage is hypothesized to be a mechanism by which elevated BP increases renal cancer risk,23,24 we also used information obtained by matching data for MRFIT participants with the national registry of treated cases of ESRD of the Centers for Medicare & Medicaid Services.24 With this information, we were able to determine how many participants had ESRD before dying from kidney cancer.
The vital status of each participant enrolled in the MRFIT cohort was determined through 1999 by using the National Death Index and records from the Social Security Administration. Death certificates from respective departments of health were obtained, and the cause of death was determined by a qualified nosologist.25 For the purposes of this study, kidney cancer is defined as a primary malignancy of the kidney, renal pelvis, or ureter (International Classification of Diseases, Ninth Revision, code 189 and International Classification of Diseases,Tenth Revision, codes C64, C65, and C66).
All analyses were completed using SAS version 9.1 (SAS Institute, Cary, NC). For all results, a significant finding was defined by a P value of <0.05.
Analyses of Kidney Cancer Incidence in Women
A total of 156 774 women were eligible for inclusion in these analyses from the WHI-CT and WHI-OS samples. Baseline characteristics revealed a racially/ethnically diverse population with more than a third of participants reporting a history of hypertension, ≈70% overweight or obese, and <10% reporting a history of diabetes mellitus or current smoking (Table 1). During the average 15-year follow-up, 407 incident kidney cancers occurred. In unadjusted analyses for trend, kidney cancer incidence significantly increased with higher SBP and DBP, previous hypertension diagnosis or treatment, and higher BMI or WC categories.
All SBP categories ≥140 mm Hg (ie, women with systolic hypertension) showed higher point estimates of kidney cancer risk when compared with SBP levels of ≤120 mm Hg. Women with SBP of 120.1 to 130.0 mm Hg also had a significant (HR, 1.33; 95% CI, 1.01–1.75) increase in kidney cancer incidence when compared with those who had SBP ≤120 mm Hg (Table 2). Elevated DBP (≥90 mm Hg; HR, 1.56; 95% CI, 1.06–2.29) and BMI were independently associated with kidney cancer. In addition, risk of kidney cancer increased with increasing age, SBP, BMI, and cigarette smoking.
In analyses stratified by BMI categories after adjustment for age, smoking, race/ethnicity, there was an increased risk of kidney cancer associated with elevated SBP in each strata of BMI although consistently significant only for women with a BMI≥30 (Table 3). This was likely, in part, because of the small number of women with elevated BP and normal (<25 kg/m2) or overweight (25.0–29.9 kg/m2) BMI. Among obese women, increasing SBP showed a dose–response association with kidney cancer risk, with women in the highest BP category (≥160 mm Hg) having an HR of 1.94 (95 CI%, 1.07–3.50). A similar pattern was found when categorical DBP was modeled within each BMI category.
Applying the same modeling strategy with WC quartiles instead of BMI categories, the point estimates of kidney cancer risk generally increased with WC, with risk almost doubling (HR, 1.91; 95% CI, 1.38–2.63) for women in the highest WC category (>97.9 cm). Both elevated SBP and DBP remained significant predictors of kidney cancer in models that included WC (Table S1 in the online-only Data Supplement).
The rate of women developing kidney cancer was higher for women on treatment for hypertension, varying from 2.2/1000 person-years (PYs) for no hypertension to 3.6/1000 PYs for women on antihypertensive drug therapy. In general, among women treated for hypertension, the percentage developing kidney cancer increased with their SBP levels, especially at the extremes. There were few women with elevated BP (SBP>140 mm Hg) but no drug therapy who developed kidney cancer. Within each BP category, there was generally a higher risk of kidney cancer among women treated when compared with those not treated, but the number of kidney cancers is relatively small within any single SBP category (Table S2).
Analyses of Kidney Cancer Mortality in Men
During the 25 years of follow-up in MRFIT, 906 deaths from kidney cancer were identified (0.26% of participants) with 840 of the deaths among white men (0.26%), 46 deaths were among black men (0.20%), and 20 deaths were among men of other races (0.16%).
On average, men who died from kidney cancer had greater SBP at entry (133.0 mm Hg) when compared with those who did not (130.0 mm Hg). The percentage dying of kidney cancer increased with higher BP levels even throughout the normotensive range (Table 4). When compared with men with SBP<120 mm Hg, the adjusted HRs ranged from 1.29 for those with SBP of 120 to 129 mm Hg to 1.87 for participants with SBP>160 mm Hg. The adjusted HR corresponding to a 1 SD higher SBP level was 1.18 (P<0.0001).
The relationship between baseline DBP and mortality from kidney cancer was weaker than for SBP (Table 4). The adjusted HRs ranged from 1.16 to 1.49 and did not increase in a graded manner. Average levels of DBP for those who died from kidney cancer and those who did not were 84.9 and 83.8 mm Hg. The adjusted HR for DBP corresponding to a 1 SD higher level was 1.13 (P<0.0001). For comparison, the corresponding HRs for cardiovascular disease and all cancers are 1.38 (P<0.0001) and 1.06 (P<0.0001), respectively.
After 10 years, the cumulative percentage dying in the first (<117 mm Hg) and the fifth (>141 mm Hg) quintiles were 0.043% and 0.086%. After 25 years, these percentages were 0.20 and 0.40. The adjusted HR for death from kidney cancer for the upper versus lower quintile of SBP was 1.73 (95% CI, 1.38–2.16; P<0.0001).
Among participants who died from kidney cancer, 45.5% reported smoking cigarettes at entry when compared with 36.4% for men who did not die from kidney cancer. The adjusted HR for death from kidney cancer associated with smoking was 1.75 (95% CI, 1.53–2.00; P<0.0001). The effects of smoking and SBP on risk of death from kidney cancer were additive. Risk was lowest among nonsmokers with SBP<120 mm Hg and highest among smokers with SBP>120 mm Hg. HR comparing nonsmokers with SBP<120 mm Hg to smokers with SBP>140 mm Hg was 2.68 (Table 5).
Covariates not associated with risk of kidney cancer–specific mortality in the Cox model included: black race (adjusted HR, 0.75; 95% CI, 0.56–1.02), serum cholesterol (adjusted HR per 40 mg/dL higher level, 1.01; 95% CI, 0.94–1.08), and the use of medication for diabetes mellitus (adjusted HR, 1.15; 95% CI, 0.69–1.92).
Only 19 MRFIT participants who died of kidney cancer (2.1%) were identified in the ESRD registry when compared with 1.2% of those who did not die from kidney cancer entered the registry. For the 19 participants, the average length of time from entering the ESRD registry to kidney cancer death was 2.5 years; average SBP and DBP at entry were 140 and 89 mm Hg. None had a history of diabetes mellitus and the average age at screening was 45.5 years.
We observed an excess risk of kidney cancer in both men and women with increasing BP levels. These relationships were independent of the association of kidney cancer with elevated body weight (measured by either BMI or WC) or cigarette smoking.
The results of both the MRFIT and the WHI are consistent with the previous literature. Although a recent study did not find an independent risk of hypertension and kidney cancer,26 the study evaluated a single BP measurement in late adolescence (age 17 years) at the time of recruitment for military service. More than 900 000 recruits were followed up >14 years, linked to the Israeli Cancer Registry. The relative risk of kidney cancer was 1.28 (95% CI, 0.17–9.50) among 4223 recruits with established hypertension by age of 17 years. This is consistent with the HR of 1.18 in the MRFIT sample for a 1 SD increase in SBP. It would be important to determine whether the known increase in BP from adolescence to adult age is a risk factor for kidney cancer and whether independent of weight gain.
Obesity and hypertension, to some extent, may represent a shared causal mechanism in the development of kidney cancer. Obesity is associated with increased glomerular filtration rate and increased renal plasma flow, which may render the kidneys more susceptible to kidney damage and carcinogenesis.23,27 Furthermore, patients with hypertension and chronic renal hypoxia caused by the upregulation of hypoxia-inducible factors that may in turn play a role in oncogenesis.10,28 Both obesity and hypertension have also been associated with oxidative stress29 and lipid peroxidation,23,30 which is hypothesized to play a role in kidney cancer pathogenesis. In addition, obesity could increase cancer risk through increased levels of insulin and insulin-like growth factor I.21,31,32 Sporadic clear-cell kidney cancer commonly involves mutations in the von Hippel-Lindau (VHL) tumor suppressor gene; with as many as 91% of clear-cell kidney cancers containing an alteration in the VHL gene.33 The VHL gene is an important regulator of hypoxia-inducible factors, fibronectin assembly, and overall cell cycle regulation.
Chronic kidney disease could represent a likely alternative explanation for associations between BP and kidney cancer.24 Chronic kidney disease secondary to elevated BP is an important risk factor for kidney cancer.22,24 Subclinical kidney damage secondary to elevated BP may be in the pathway from elevated BP to kidney cancer. Alternatively, environmental agents may contribute to both kidney injury (leading to cancer onset) and elevated BP. These studies do not prove a causal association of elevated BP and kidney cancer.
For all women combined, we found higher rates of kidney cancer among women with treated (versus untreated) hypertension and slightly lower point estimates of kidney cancer incidence among those without hypertension. In addition, within BP strata, point estimates for kidney cancer incidence were typically lower for pharmacologically treated versus untreated individuals. These findings potentially suggest that antihypertensive medications may contribute to kidney cancer risk. However, severity or duration of hypertension alone could explain such findings. For example, women whose elevated BP is not treated pharmacologically are more likely to have mild hypertension controlled with lifestyle alone or hypertension of shorter duration (eg, recently diagnosed and attempting lifestyle change before initiation of drug therapy) when compared with those who are prescribed antihypertensive medications. Adjusting for the presence of hypertension has eliminated excess risk associated with pharmacotheraphy in multiple studies.23,34 Furthermore, 1 large study found that antihypertensive use did not modify the relationship between BP and kidney cancer incidence, whereas among individuals taking antihypertensive drugs, only those with poorly controlled BP showed a significantly increased cancer risk.10 As a result, studies have typically concluded that associations between antihypertensive medication and kidney cancer are unlikely to be causal, reflecting instead confounding by the presence of hypertension.1,20,23,35
Additional prospective studies are needed to determine whether treatment and control of hypertension can reduce the risk of kidney cancer and if so, whether specific drug therapies would differ in preventive efficacy. The low incidence of kidney cancer precludes the use of data from any specific previous hypertension trial to answer this question. However, it is possible that pooling the results from many large clinical trials may provide a clue as to whether BP lowering reduces kidney cancer risk. It is possible that combination of hypertension or obesity, renal damage and host, genetic markers could identify a high risk of kidney cancer and lead to specific treatment strategies.
This study has several limitations, which are, in part, addressed by the use of 2 complementary samples. Because cancer development occurs for prolonged periods of time, 15 to 25 years may still be a short duration of follow-up time. The WHI sample comprises an ethnically diverse population; however, a relatively low number of incident kidney cancer occurred in ethnic groups other than white women precluding evaluation of ethnic-specific patterns. A similar problem was found for MRFIT in which almost all cases occurred in white men. The geographic and ethnic diversity of the samples add an important element of external validity to the literature on kidney cancer. Only a single measurement of risk factors was obtained for the men screened for MRFIT. Results from WHI used only baseline measurement of BP. Experimental studies that evaluate effects of change in BP on risk of kidney cancer would prove further evidence of possible role of hypertension in the pathogenesis of kidney cancer. We are unable to evaluate the association between hypertension or obesity and specific subtypes of kidney cancer from these data. However, because >85% of kidney cancers arise from the renal parenchyma,1 it is likely that the outcomes assessed here primarily reflect renal cell carcinoma.
We found that obesity and hypertension were independently associated with the development of kidney cancer, and that hypertension shows a robust association with kidney cancer in long-term follow-up in US men and women. The mechanisms of these associations are not known or are the role of specific antihypertensive drug therapy in reducing kidney cancer risk.
We acknowledge the Women’s Health Initiative (WHI) and Multiple Risk Factor Intervention Trial (MRFIT) for the use of the data that was examined in these analyses.
Sources of Funding
The Women’s Health Initiative (WHI) program is funded by the National Heart, Lung, and Blood Institute (NHLBI), National Institutes of Health (NIH), US Department of Health and Human Services through contracts HHSN268201100046C, HHSN268201100001C, HHSN268201100002C, HHSN268201100003C, HHSN268201100004C, and HHSN271201100004C. Dr Sanfilippo’s time on this project was supported by the University of Pittsburgh Internal Medicine Residency program. The Multiple Risk Factor Intervention Trial was contracted by the NHLBI, NIH, Bethesda, MD. Follow-up after the end of the trial was supported with NIH/NHLBI grants R01-HL-43232 and R01-HL-68140.
The online-only Data Supplement is available with this article at http://hyper.ahajournals.org/lookup/suppl/doi:10.1161/HYPERTENSIONAHA.113.02953/-/DC1.
- Received December 6, 2013.
- Revision received December 27, 2013.
- Accepted January 31, 2014.
- © 2014 American Heart Association, Inc.
- Setiawan VW,
- Stram DO,
- Nomura AM,
- Kolonel LN,
- Henderson BE
- 3.↵National Cancer Institute. SEER Stat Fact Sheets: Kidney and Renal Pelvis. 2013. Bethseda, MD: National Cancer Institute. http://seer.cancer.gov/statfacts/html/kidrp.html. Accessed March 4, 2014.
- Schouten LJ,
- van Dijk BA,
- Oosterwijk E,
- Hulsbergen-van de Kaa CA,
- Kiemeney LA,
- Goldbohm RA,
- Schalken JA,
- van den Brandt PA
- Cote ML,
- Colt JS,
- Schwartz KL,
- Wacholder S,
- Ruterbusch JJ,
- Davis F,
- Purdue M,
- Graubard BI,
- Chow WH
- Weikert S,
- Boeing H,
- Pischon T,
- et al
- Fraser GE,
- Phillips RL,
- Beeson WL
- Bjørge T,
- Tretli S,
- Engeland A
- Luo J,
- Margolis KL,
- Adami HO,
- Lopez AM,
- Lessin L,
- Ye W
- D’Agostino RB,
- Sullivan L,
- Massaro J
- Eberly LE,
- Stamler J,
- Kuller LH,
- Neaton JD
- Curb JD,
- McTiernan A,
- Heckbert SR,
- Kooperberg C,
- Stanford J,
- Nevitt M,
- Johnson KC,
- Proulx-Burns L,
- Pastore L,
- Criqui M,
- Daugherty S
- Kaelin WG Jr..
- Nickerson ML,
- Jaeger E,
- Shi Y,
- et al
- Shapiro JA,
- Williams MA,
- Weiss NS,
- Stergachis A,
- LaCroix AZ,
- Barlow WE
Novelty and Significance
What Is New?
Previous research on obesity, hypertension, and kidney cancer has been limited by small sample sizes13 and self-reported predictors.2,13,14 Furthermore, of the prospective studies we identified with mixed-sex or female samples,2,7–9,14 only 2 measured blood pressure.9,15 Both of these studies are European, thus their findings may not generalize to the more racially/ethnically diverse US population. The current study examines the relationship between levels of hypertension and degree of obesity as risk factors for kidney cancer both separately and in relationship to one another for a prolonged period of time in a large racially diverse female population. It also analyzes the association between blood pressure and kidney cancer deaths over 25 years of follow-up in men.
What Is Relevant?
An understanding of the pathogenesis of kidney cancer is essential for developing better preventive measures in this era of escalating incidence.
Among 156 774 participants of the Women’s Health Initiative, blood pressure and body mass index were independent predictors of kidney cancer incidence in adjusted models. Multiple Risk Factor Intervention Trial (MRFIT) data showed that the risk of death from kidney cancer increased in a dose–response fashion with increasing blood pressure for an average of 25 years of follow-up. In addition, smoking and systolic blood pressure were independently associated with increased on risk of death from kidney cancer.