Resting Heart Rate and Risk of Cancer Mortality

Cancer is the second-leading cause of death in the United States (1, 2). The aging population as well as the high rates of unhealthy lifestyle behaviors, such as smoking, physical inactivity, and poor diet, contribute to increased cancer diagnoses (3). However, between 30% and 50% of all cancer cases are theoretically preventable (4, 5). Consequently, prevention offers the most cost-effective long-term strategy for reducing the burden of cancer (4), and modifiable lifestyle factors have been proposed as targets for preventative clinical and public health strategies (6).

Resting heart rate (RHR) is an essential vital sign measured as part of routine physical examinations in clinical practice (7). It is a simple and inexpensive procedure that does not require complex training and can provide important prognostic information (8–10). RHR is related to parasympathetic and sympathetic nerve activity and could be influenced by genetics, physiologic and psychologic stress, certain hormones (e.g., thyroid hormone, epinephrine), and/or lifestyle factors, such as physical activity (PA; refs. 10, 11). Higher RHR is observed in several cardiovascular-related conditions, such as hypertension and atherosclerosis, and is a predictor of cardiovascular morbidity and mortality (8, 9, 12, 13). Elevated RHR has also been observed in noncardiovascular-related mortality, such as respiratory and cancer-related deaths (12–14). Specifically, RHR has been associated with increased mortality of digestive cancers (13, 15–17), lung cancer (13, 15, 16), breast cancer (13), and kidney cancer (13). However, an important limitation is that these studies did not consider other factors that may directly impact RHR, such as cardiorespiratory fitness (CRF).

CRF is a health-related component of physical fitness which measures the ability of the circulatory, respiratory, and muscular systems to provide oxygen during sustained PA, and it is often used to represent consistent participation in recent aerobic PA (18). CRF is also a significant predictor of all-cause and cancer-related mortality (18–22). Furthermore, higher CRF is associated with lower RHR (23, 24). One study found a significant association between RHR and all-cause mortality, independent of CRF, suggesting that higher RHR not only reflects poor general fitness, but could also be an independent risk factor for all-cause mortality (25). Therefore, CRF may be a potential confounding factor in the previously reported associations between RHR and cancer mortality (13, 15–17).

The purpose of this study was to evaluate the association between RHR and cancer mortality, independent of and in combination with CRF. We hypothesized that there would be a significantly higher risk of cancer mortality among higher RHR groups and that this association would be attenuated after adjusting for CRF.

This study was conducted using the Aerobics Center Longitudinal Study (ACLS), a cohort of men and women who received comprehensive preventive medical assessments on a periodical basis at the Cooper Clinic in Dallas, Texas (26). Participants were ≥20 years old at baseline. Participants underwent preventive health examinations regarding diet, exercise, and other lifestyle factors associated with increased risk of chronic disease between 1974 and 2002 (27).

Participants with history of myocardial infarction, stroke, or cancer at baseline and those who had less than 1 year of follow-up for mortality were excluded. Participants were predominantly non-Hispanic whites, belonging to middle to upper socioeconomic strata, and employed in or retired from professional or executive positions (27).

The Cooper Clinic Institutional Review Board reviewed and approved the study annually. Written informed consent was obtained from participants before baseline data collection.

Participants completed a standardized medical history questionnaire on demographic characteristics, lifestyle habits (e.g., smoking, alcohol consumption, PA), medical conditions (e.g., hypertension, diabetes, hypercholesterolemia), and personal or parental history of cardiovascular disease (myocardial infarction or stroke) or cancer.

Participants underwent a comprehensive health evaluation after ≥12 hours fasting, as described in detail elsewhere (28, 29). Weight and height were measured with participants wearing light clothing without shoes using a standard clinical scale and stadiometer, respectively, from which body mass index (BMI) was quantified as weight in kilograms divided by height in meters squared (30).

Resting blood pressure (BP) was assessed with the standard auscultation method following the American Heart Association protocol (31), initiated after at least 5 minutes of seated rest. Hypertension was defined as a measured systolic BP of ≥130 mm Hg or diastolic BP of ≥80 mm Hg (32).

RHR was obtained from the electrocardiogram after a 5-minute rest with the patient in the supine position. Serum glucose and lipid levels were analyzed using automated bioassays after ≥12 hours of overnight fast, according to the Centers for Disease Control and Prevention Lipid Standardization Program.

CRF (in metabolic equivalents or METs) was quantified by a maximal treadmill test using a modified Balke protocol (33). Participants were encouraged to reach their maximum effort, estimated as reaching ≥85% of age-predicted heart rate maximum, and the test was terminated after participant request due to exhaustion or according to physician's criteria due to medical reasons. CRF values were dichotomized as unfit (low) and fit (high) corresponding to the lower 20% and the upper 80%, respectively, of the age- and sex-specific distribution of treadmill exercise duration in the entire ACLS population following our earlier studies (34–38).

Participants were followed from baseline examination until date of death for decedents or December 31, 2003, for survivors. Vital status was evaluated with the National Death Index and death certificates. Cancer deaths were determined using the International Classification of Diseases, Ninth Revision (codes 140–208; ref. 39), for deaths occurring before 1999, and Tenth Revision (codes C00-C97) for deaths occurring between 1999–2003. For cancer in specific sites, the following ICD-9/10 codes were used: breast, 174–175/C50; lung, 162.2–163.0/C34; prostate, 186/C61; colon, 153/C18; and rectum, 154/C19-C21.

Participants were grouped into four RHR categories based on RHR at their baseline examination: (i) <60 beats/minute, (ii) 60–69 beats/minute, (iii) 70–79 beats/minute, and (iv) ≥80 beats/minute, based on previous studies (40, 41). Baseline characteristics were described for all participants (men and women combined) across the established RHR categories, given that there was no significant interaction between RHR and sex (P = 0.389). Groups were compared using χ² analysis for categorical variables and general linear models for continuous variables.

Follow-up time was calculated from the date of baseline medical examination until death date for decedents, or December 31, 2003, for survivors. Cox proportional hazard models were used to estimate HRs, 95% confidence intervals (95% CI), and cancer mortality rates per 10,000 person-years of follow-up according to RHR categories, using <60 beats/minute as the reference group. HRs estimates were calculated for overall cancer mortality and site-specific lung, breast, prostate, colorectal, and other (esophagus, stomach, intestine, liver, pancreas, gall, kidney, and lymph nodes) cancer mortality.

Multivariable Cox models included the following covariates: age (years), sex, examination year, BMI (kg/m²), current smoking (yes/no), heavy alcohol drinking (>7 drinks per week for women, and >14 drinks per week for men), meeting the 2018 US aerobic PA guidelines (≥500 metabolic equivalent [MET]-min/week), parental history of cancer (yes/no), abnormal electrocardiogram (ECG, yes/no based on presence of abnormalities in heart rhythm, position/axis, and/or impression obtained from ECG during supine rest or during the maximal treadmill test), diabetes (yes/no based on physician-diagnosed diabetes, insulin use, or measured fasting glucose ≥126 mg/dL), measured hypertension (systolic BP ≥130 mm Hg or diastolic BP ≥80 mm Hg), self-reported hypertension (yes/no based on physician-diagnosed hypertension), hypercholesterolemia (yes/no based on physician-diagnosed hypercholesterolemia or measured total cholesterol ≥240 mg/dL), and CRF (as a continuous variable, in METs). Additionally, Cox models were developed using a stratified analysis for current smoking status, obesity status (BMI ≥30 kg/m²), meeting aerobic PA guidelines, and self-reported hypertension (42).

Finally, a joint analysis was conducted by dichotomizing RHR (<80 beats/minute or ≥80 beats/minute) and CRF status [fit (upper 80%) or unfit (lower 20%)] and categorizing participants into “fit and <80 beats/minute,” “unfit and <80 beats/minute,” “fit and ≥80 beats/minute,” and “unfit and ≥80 beats/minute,” using “fit and <80 beats/minute” as the reference group. Statistical analyses were performed using SAS software (version 9.4 SAS Institute Inc.) and two-sided P < 0.05 were deemed significant.

A total of 50,108 men and women were included in this study, with mean follow-up of 15.0±8.6 years, and 1,090 overall cancer deaths. Table 1 shows baseline characteristics for the overall sample and for each RHR category. The individuals in the lower RHR groups were more fit, more physically active, leaner, and had a lower prevalence of hypertension, diabetes, and hypercholesterolemia compared with the higher RHR categories.

Table 2 shows the HRs of overall cancer mortality for the four categories of RHR. There was a significant positive linear trend for cancer mortality along with increasing levels of RHR. Compared with RHR <60 beats/minute, individuals with RHR 70–79 and ≥80 beats/minute had a significantly greater risk of cancer mortality after adjusting for age, sex, examination year, BMI, current smoking, heavy alcohol drinking, meeting the 2018 US aerobic PA Guidelines, parental history of cancer, abnormal electrocardiogram findings, diabetes, measured hypertension, self-reported hypertension, and hypercholesterolemia (Table 2). After further adjustment for CRF, individuals with RHR ≥80 beats/minute remained at significantly increased risk of cancer mortality (Table 2).

The HRs for site-specific cancer mortality indicated a positive linear trend across RHR categories for death due to lung cancer, colorectal cancer, and other cancers in Model 1 after adjusting for age, sex, and examination year (Table 3). However, these trends as well as the trends in other site-specific cancer mortalities were no longer significant after adjusting for other possible confounders including CRF in Model 3.

Figure 1 shows HRs for overall cancer mortality by combined RHR and CRF categories. Compared with the “fit and <80 beats/minute” reference group, the “unfit and ≥80 beats/minute” group was at the greatest risk of cancer mortality (HR, 1.78; 95% CI, 1.30–2.43) after adjusting for potential confounders. Also, compared with the “fit and <80 beats/minute” group, the “fit and ≥80 beats/minute” group had 51% increased risk of cancer mortality (HR, 1.51; 95% CI, 1.11–2.04), and the “unfit and <80 beats/minute” group had 41% increased risk of cancer mortality (HR, 1.41; 95% CI, 1.20–1.66).

In additional analyses, the potential moderating effects of current smoking, obesity, meeting aerobic PA guidelines, and self-reported hypertension on the association between RHR and cancer mortality were analyzed by including interaction terms as well as stratified analyses of these subgroups in the multivariable Cox regression. No significant interactions were observed, suggesting that the associations between RHR and cancer mortality were similar among these potential effect modifiers. Some hypertension medications (e.g., beta-adrenergic blocking agents) can lower RHR. Although we did not have data on hypertension medications in this study, we found consistent trends showing higher HRs in RHR ≥80 beats/minute compared with RHR <60 beats/minute in individuals with (HR, 1.34; 95% CI, 0.83–2.15) and without (HR, 1.34; 95% CI, 1.01–1.77) self-reported hypertension in the stratified analyses.

Lower RHR has been associated with higher fitness (7, 23, 24) and less susceptibility to chronic diseases (10, 41, 43); however, its relationship with cancer is inconsistent (16, 17, 44–48). We found that elevated RHR (≥80 beats/min) was associated with higher risk of overall cancer mortality even after adjusting for CRF in this cohort (Table 2). In the joint analysis (Fig. 1), we observed the highest risk of cancer mortality in the unfit with elevated RHR (≥80 beats/minute) group; however, this risk of cancer mortality was slightly attenuated in the fit with high RHR group, suggesting a possible protective effect of CRF on cancer mortality despite high RHR.

These results are comparable to findings from previous studies. Jouven and colleagues (17) analyzed healthy middle-aged men in the Paris Prospective Study-1 and found a consistent and graded association between RHR and cancer mortality. They observed a significantly higher risk of cancer mortality for the highest quartile of RHR compared to the lowest quartile (HR, 2.40; 95% CI, 1.90–2.90) after adjusting for current PA and traditional cardiovascular risk factors. Furthermore, Anker and colleagues (16) analyzed a clinical population of patients with colorectal carcinoma, non-small cell lung cancer, and pancreatic cancer, and found that increased RHR (≥75 beats/minute) was an independent predictor of all-cause mortality (HR, 1.66; 95% CI, 1.00–2.76) after accounting for other clinically relevant mortality predictors, such as BMI, type of tumor, and tumor stage. In addition, they reported RHR as the strongest predictor of overall survival among advanced cancer patients. However, other epidemiologic studies have also reported different findings. Kristal-Boneh and colleagues (46) did not find a significant association between RHR and cancer mortality among 3,527 men after 8 years of follow-up (RR, 1.13; 95% CI, 0.40–3.00). Similarly, Mensik and Hoffmeister (48) followed 4,756 healthy participants for 12 years and did not find statistically significant associations between RHR and cancer mortality. Both of these studies had relatively smaller sample sizes with shorter follow-up times and fewer cancer cases than this study, which may have contributed to lack of significant results. In addition, none of the earlier studies considered CRF in their analyses.

Some pathophysiologic mechanisms have been proposed to provide a biological explanation underlying the association between high RHR and increased cancer mortality. Elevated RHR reflects increased sympathetic nerve activity, which is influenced by the activity of the autonomic nervous system, levels of circulating hormones, and CRF (49). It has been shown that increased sympathetic activation might play an essential role in the relationship between RHR and cancer mortality, potentially due to the effect of β-adrenergic signaling on the regulation of multiple cellular processes that contribute to the initiation and progression of cancer (50). The effect of prolonged sympathetic nervous system activation on cancer progression has also been associated with dysregulation of the hypothalamic-pituitary-adrenal axis (51). Activation of the sympathetic nervous system and hypothalamic-pituitary-adrenal axis increases circulating levels of catecholamines and glucocorticoids (51, 52), which can dysregulate the immune system over time (53, 54). An impaired immune response may enable tumor growth through intracellular pathways, initiated on β-adrenergic and glucocorticoid receptors, that alter the tumor microenvironment by promoting inflammation (e.g., elevations in IL6; refs. 55, 56), cellular resistance to apoptosis, and cancer progression (50, 57).

RHR can also be influenced by CRF. Higher CRF has been associated with decreased sympathetic outflow and increased parasympathetic activity, which translates to a lower RHR (58). As a result, increased CRF may protect against cancer mortality through reduced activation of the β-adrenergic pathway that promotes cancer initiation and progression. In addition, higher CRF has also been associated with lower levels of hypothalamic–pituitary–adrenal axis activity that may protect against damage to the immune system that would favor the tumor microenvironment and cancer progression (51, 59). Moreover, greater CRF may protect against cancer mortality via anti-inflammatory effects. Higher CRF has been shown to be associated with reduced levels of proinflammatory cytokines (e.g., IL6) linked to tumor growth and malignant progression (60). In summary, decreased sympathetic activity, reduced hypothalamic-pituitary-adrenal axis activity, improved immune function, and decreased inflammation may contribute to the possible protective role of higher CRF in the relationship between RHR and cancer mortality.

This cohort provided a large analytical sample including both men and women with objectively-measured exposures (RHR through electrocardiogram and CRF by maximal treadmill test) and hard cancer mortality outcomes from the National Death Index. Some limitations should be outlined. Only baseline data of RHR and CRF assessment were included in the analysis, therefore, concomitant changes in RHR and CRF over time were not considered. Also, data were not available for prescription medications that might affect RHR, such as beta-blockers that are sometimes prescribed for hypertension. However, we attempted to account for medications by including physician-diagnosed hypertension and measured hypertension as separate covariates, since individuals with physician-diagnosed hypertension are more likely to be on medications. Furthermore, the cohort is comprised of highly-educated non-Hispanic whites from middle to upper socioeconomic strata. Consequently, our results may not necessarily apply to other populations. On the other hand, homogeneity in ethnicity and socioeconomic status may reduce potential confounding due to race/ethnicity, education, and income. Another limitation was the lack of incident cancer cases as opposed to cancer mortality; thus, the effect of CRF on RHR and cancer incidence was not assessed. Relatedly, information such as the date, patient age, or cancer stage at the time of diagnosis were not available, all of which could affect survival. Medical advancements in cancer treatment during the long assessment period (1974–2003) also affect survival rates and may influence the observed relationships between RHR, CRF, and cancer mortality even after adjusting for examination year. Finally, the current results are derived from an observational study, thus causation cannot be inferred.

In conclusion, RHR ≥80 beats/minute was associated with increased risk of overall cancer mortality independent of CRF. In addition, higher levels of CRF attenuated the magnitude of this association between RHR and cancer mortality. These findings provide evidence of RHR as a possible independent risk factor of cancer mortality. Furthermore, these results support the importance of PA promotion through clinical and public health interventions to improve CRF and lower RHR for the reduction of cancer mortality risk as well as for prevention of cardiovascular disease.

S.N. Blair reports grants from NIH during the conduct of the study. No disclosures were reported by the other authors.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

L. Gutierrez-Martinez: Formal analysis, visualization, writing–original draft. A.G. Brellenthin: Formal analysis, supervision, visualization, methodology, writing–review and editing. E.C. Lefferts: Methodology, writing–review and editing. D. Lee: Conceptualization, data curation, supervision, writing–review and editing. X. Sui: Conceptualization, data curation, writing–review and editing. C.J. Lavie: Writing–review and editing. S.N. Blair: Conceptualization, resources, funding acquisition, writing–review and editing.

Research reported in this publication was supported by the National Institute on Aging and the National Heart, Lung, and Blood Institute of the NIH awarded to S.N. Blair under award numbers R01AG06945 and R01HL62508, and to D. Lee under award number R01HL133069. We thank the Cooper Clinic physicians and technicians for collecting the baseline data and staff at the Cooper Institute for data entry and data management. The authors thank Dr. Aditi Narsale for her help in early data analysis. This study was supported by the NIH grants (AG06945, HL62508, and HL133069). S.N. Blair has received unrestricted research grants from The Coca-Cola Company, but these grants were not used to support this manuscript.

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.