Abstract
Objectives
Elevated resting heart rate (RHR) predicts mortality in middle-aged and older adults, primarily through cardiovascular disease (CVD). Whether RHR predicts mortality in young adults, which causes of death are involved and whether the association replicates are unclear. In young US adults, who fall below the starting age of current cardiovascular risk tools, we examined RHR and all-cause and cause-specific mortality across two cohorts.
Methods
We analysed 3291 adults aged 20–49 years from National Health and Nutrition Examination Survey (NHANES) 1999–2004 (primary cohort) linked to mortality through 2019 (median follow-up 17.8 years; 120 deaths) with replication in 8941 adults from NHANES III (1988–1994; 1317 deaths; median follow-up 27.6 years; temporal replication). RHR and heart rate reserve (HRR) were modelled per 10 beats per minute increment using Cox regression adjusted for demographic, lifestyle and comorbidity covariates.
Results
Each 10 beats per minute higher RHR was associated with higher all-cause mortality (primary cohort: HR, 1.26; 95% CI 1.07 to 1.50; validation: HR, 1.17; 95% CI 1.12 to 1.22). In the primary cohort most deaths (99 of 120) were non-CVD; the CVD-specific analysis was underpowered (21 deaths; HR, 1.15; 95% CI 0.77 to 1.71). In validation, RHR predicted non-CVD (HR, 1.17; 95% CI 1.11 to 1.23) and CVD (HR, 1.18; 95% CI 1.08 to 1.29) mortality about equally. The association strengthened with age, emerging by the early-to-mid 30 s in both cohorts, before the starting age of current risk tools. HRR, measured under submaximal conditions, was not associated with any outcome.
Conclusions
In young US adults, elevated resting heart rate, a free and routinely measured vital sign, predicts long-term mortality across two cohorts, marking broad survival vulnerability rather than cardiovascular risk alone. Detectable in the early-to-mid 30 s, below the starting age of current cardiovascular risk tools, resting heart rate may flag at-risk young adults before these tools apply. Submaximal HRR carried no prognostic value.
Keywords: Risk Factors, Epidemiology, Biostatistics
WHAT IS ALREADY KNOWN ON THIS TOPIC.
WHAT THIS STUDY ADDS
In two US cohorts of adults aged 20–49 years (National Health and Nutrition Examination Survey [NHANES] 1999–2004 and NHANES III), each 10 beats per minute higher resting heart rate predicted higher long-term mortality.
The association strengthened with age, becoming apparent only by the early-to-mid 30s, a decade before the age (around 40) at which current cardiovascular risk tools begin.
In the larger cohort, resting heart rate predicted non-cardiovascular and cardiovascular death about equally, marking broad survival vulnerability rather than a cardiovascular-specific risk.
Heart rate reserve measured under submaximal conditions carried no prognostic value.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Resting heart rate is free, universally recorded and informative from the early 30s, so it could help identify at-risk young adults who currently fall outside all formal cardiovascular risk assessment; whether it can usefully extend early risk stratification warrants prospective study.
Because elevated resting heart rate predicts non-cardiovascular as well as cardiovascular death, it is best regarded as a general survival marker; research should now identify the mechanisms linking it to these diverse causes, spanning cardiovascular, cancer and behavioural pathways.
Introduction
Elevated resting heart rate (RHR) is a well-established predictor of cardiovascular and all-cause mortality in middle-aged and older populations.1,4 Each 10 beats per minute increase in RHR carries a 10−20% higher mortality risk in adults older than 50 years, with proposed mechanisms centred on sympathetic overactivation, accelerated atherosclerosis and increased myocardial oxygen demand.5,7
In younger adults the evidence is more limited and less integrated. Within the National Health and Nutrition Examination Survey (NHANES) programme, resting pulse predicted coronary heart disease and death in the NHANES I Epidemiologic Follow-up Study,8 and a later NHANES III analysis related RHR to both cardiovascular and non-cardiovascular death9; the Chicago Heart Association Detection Project reported similar associations, including in younger men.10 These studies, however, examined broad adult age ranges in single cohorts. None focused specifically on young adults, characterised the age at which the association emerges, tested whether it reproduces across cohorts or evaluated heart rate reserve (HRR). This gap matters: adults aged 20–49 years fall below the starting age of contemporary cardiovascular risk tools, and their atherosclerotic burden is typically minimal.11 12 Because the RHR-mortality association in older adults is largely mediated by cardiovascular disease, whether it holds at these younger ages and whether any signal is specifically cardiovascular or broader, cannot be assumed from older cohorts.
HRR, defined as the difference between peak exercise heart rate and RHR, provides complementary information about cardiac chronotropic competence.13 Low HRR has been associated with mortality in clinical populations undergoing maximal exercise testing,14 15 but evidence from population-based samples using submaximal protocols is limited.
The NHANES 1999–2004 included a submaximal treadmill examination in adults aged 12−49 years, yielding two complementary measures: RHR, indexing resting autonomic tone and HRR, indexing submaximal chronotropic competence. We linked this cohort to national mortality data through 2019 and sought replication in the earlier NHANES III. Specifically, in adults aged 20–49 years, we examined whether RHR and HRR predict all-cause and cause-specific mortality, at what age any association emerges, and whether it reproduces in a second, earlier NHANES cohort.
Methods
Study design and population
This study used data from the NHANES 1999–2000, 2001–2002 and 2003–2004 cycles, nationally representative cross-sectional surveys conducted by the National Center for Health Statistics (NCHS). Examination data were collected during 1999–2004, and mortality follow-up extended through 31 December 2019. This analysis used de-identified, publicly available data and was exempt from further institutional review board review.
The cardiovascular fitness examination involved a submaximal treadmill test (Acuflex treadmill, Quinton Cardiology) using an individualised protocol targeting 75% of age-predicted maximum heart rate. Participants aged 12–49 years without contraindications were eligible. We restricted the analysis to adults aged 20–49 years with valid RHR measurements. Of 3359 participants who completed the treadmill examination, 68 (2.0%) were excluded for missing covariate data (primarily hypertension status (n=51), body mass index (BMI) (n=10) and smoking status (n=4)), yielding an analytical sample of 3291.
Exposures
RHR was measured as the seated, pre-exercise pulse rate recorded during the cardiovascular fitness examination. HRR was calculated as stage 2 peak heart rate minus RHR, reflecting submaximal chronotropic response. Both exposures were modelled continuously per 10 beats per minute increment.
Outcomes
Mortality was ascertained through probabilistic linkage with the National Death Index through 31 December 2019, performed by the NCHS. Vital status was determined using the MORTSTAT variable. Cause of death was classified using the NCHS 10-category underlying cause-of-death variable (UCOD_LEADING): heart disease (category 1), malignant neoplasms (category 2), chronic lower respiratory disease (category 3), accidents (category 4), cerebrovascular disease (category 5) and other causes. Cardiovascular disease (CVD) mortality was defined as heart disease or cerebrovascular disease (categories 1 and 5). Non-CVD mortality included all remaining causes.
Covariates
Models were adjusted for age (continuous), sex, race/ethnicity (non-Hispanic white, non-Hispanic black, Mexican American, other Hispanic, other race), education (less than high school, high school/General Educational Development [GED], some college, college graduate or above), BMI (continuous, kg/m²), smoking status (never, former, current), self-reported physician-diagnosed diabetes (yes/no) and hypertension (yes/no).
Statistical analysis
Baseline characteristics were compared across RHR clinical categories (<60, 60–69, 70–79, ≥80 beats per minute) using Kruskal-Wallis tests for continuous variables and χ2 tests for categorical variables. Crude mortality rates were calculated per 1000 person-years.
The primary analysis used Cox proportional hazards regression with follow-up time (years from examination to death or 31 December 2019) as the time scale. Models were built sequentially: Model 1 (unadjusted), Model 2 (age, sex, race/ethnicity) and Model 3 (fully adjusted for all covariates). RHR was modelled primarily as a continuous variable (per 10 beats per minute increment) and, for display and trend testing, grouped into clinical categories (table 1; figure 1, Panels A–B), tertiles (figure 1, Panels C–D) and quintiles (test for linear trend; online supplemental eTable 3). The dose-response relationship was examined using restricted cubic splines with four knots placed at the 5th, 35th, 65th and 95th percentiles. Inverse probability weighted (IPW) Kaplan-Meier survival curves were constructed for RHR clinical categories and tertiles, with weights derived from multinomial propensity score models adjusting for all covariates. Weighted likelihood ratio test p values were used for group comparisons. Kaplan-Meier and dose-response analyses were presented for both all-cause and non-CVD mortality, with deaths from competing causes treated as censored events.
Table 1. Baseline characteristics by resting heart rate category among US adults aged 20–49 years, NHANES 1999–2004 (N=3359).
| Resting heart rate category | ||||||
|---|---|---|---|---|---|---|
| Characteristic | Overall N=3359* | <60 bpm N=421* | 60–69 bpm N=1028* | 70–79 bpm N=1112* | ≥80 bpm N=798* | P value† |
| Age (years) | 33.2 (8.6) | 34.1 (8.4) | 33.9 (8.6) | 33.0 (8.6) | 32.4 (8.5) | <0.001 |
| Sex | <0.001 | |||||
| Female | 1615 (48.1) | 105 (24.9) | 452 (44.0) | 583 (52.4) | 475 (59.5) | |
| Male | 1744 (51.9) | 316 (75.1) | 576 (56.0) | 529 (47.6) | 323 (40.5) | |
| Race/ethnicity | 0.075 | |||||
| Mexican American | 854 (25.4) | 108 (25.7) | 284 (27.6) | 274 (24.6) | 188 (23.6) | |
| Non-Hispanic black | 659 (19.6) | 95 (22.6) | 203 (19.7) | 203 (18.3) | 158 (19.8) | |
| Non-Hispanic white | 1571 (46.8) | 194 (46.1) | 466 (45.3) | 527 (47.4) | 384 (48.1) | |
| Other Hispanic | 162 (4.8) | 19 (4.5) | 46 (4.5) | 57 (5.1) | 40 (5.0) | |
| Other race | 113 (3.4) | 5 (1.2) | 29 (2.8) | 51 (4.6) | 28 (3.5) | |
| Education | 0.10 | |||||
| Less than high school | 761 (22.7) | 89 (21.2) | 246 (23.9) | 250 (22.5) | 176 (22.1) | |
| High school/GED | 806 (24.0) | 88 (21.0) | 247 (24.0) | 263 (23.7) | 208 (26.1) | |
| Some college | 1052 (31.3) | 132 (31.4) | 306 (29.8) | 348 (31.3) | 266 (33.4) | |
| College graduate or above | 738 (22.0) | 111 (26.4) | 229 (22.3) | 251 (22.6) | 147 (18.4) | |
| Unknown | 2 | 1 | 0 | 0 | 1 | |
| BMI (kg/m²) | 27.4 (5.9) | 26.3 (4.6) | 26.9 (5.3) | 27.7 (6.0) | 28.3 (6.9) | <0.001 |
| Unknown | 10 | 0 | 2 | 2 | 6 | |
| Waist circumference (cm) | 92.5 (14.4) | 90.1 (12.1) | 91.3 (13.1) | 93.1 (15.0) | 94.2 (16.0) | <0.001 |
| Unknown | 26 | 2 | 7 | 7 | 10 | |
| Smoking status | <0.001 | |||||
| Current | 890 (26.5) | 110 (26.2) | 257 (25.0) | 271 (24.4) | 252 (31.7) | |
| Former | 530 (15.8) | 82 (19.5) | 187 (18.2) | 155 (14.0) | 106 (13.3) | |
| Never | 1935 (57.7) | 228 (54.3) | 584 (56.8) | 685 (61.7) | 438 (55.0) | |
| Unknown | 4 | 1 | 0 | 1 | 2 | |
| Alcohol use | 0.054 | |||||
| Drinker | 788 (75.8) | 109 (83.8) | 245 (77.5) | 240 (73.6) | 194 (72.4) | |
| Never | 252 (24.2) | 21 (16.2) | 71 (22.5) | 86 (26.4) | 74 (27.6) | |
| Unknown | 2319 | 291 | 712 | 786 | 530 | |
| Diabetes | 54 (1.6) | 3 (0.7) | 5 (0.5) | 19 (1.7) | 27 (3.4) | <0.001 |
| Unknown | 1 | 0 | 0 | 0 | 1 | |
| Hypertension | 294 (8.9) | 22 (5.3) | 84 (8.3) | 97 (8.9) | 91 (11.6) | 0.003 |
| Unknown | 51 | 4 | 19 | 16 | 12 | |
| Resting heart rate (bpm) | 71.3 (10.7) | 54.8 (3.6) | 64.4 (2.8) | 73.7 (2.8) | 85.8 (5.4) | <0.001 |
| Peak heart rate (bpm) | 141.8 (17.1) | 139.6 (16.6) | 139.7 (18.3) | 142.4 (16.7) | 144.7 (15.9) | <0.001 |
| Absolute HRR (bpm) | 70.4 (19.2) | 84.9 (16.8) | 75.3 (18.7) | 68.7 (16.8) | 58.9 (16.6) | <0.001 |
| % age-predicted HR achieved | 75.9 (8.1) | 75.1 (7.8) | 75.0 (8.7) | 76.1 (8.0) | 77.1 (7.6) | <0.001 |
| Chronotropic index | 0.6 (0.1) | 0.6 (0.1) | 0.6 (0.1) | 0.6 (0.1) | 0.6 (0.1) | <0.001 |
| HR recovery at 1 min (bpm) | 10.6 (7.6) | 13.9 (8.0) | 11.7 (8.2) | 10.3 (6.5) | 7.9 (7.1) | <0.001 |
| Unknown | 359 | 55 | 88 | 137 | 79 | |
| HR recovery at 2 min (bpm) | 28.0 (11.7) | 35.0 (11.8) | 30.3 (11.2) | 27.4 (10.9) | 22.4 (10.7) | <0.001 |
| Unknown | 266 | 34 | 75 | 98 | 59 | |
| Estimated VO₂max (mL/kg/min) | 40.1 (10.0) | 45.2 (10.4) | 41.3 (9.7) | 38.9 (9.7) | 37.4 (9.3) | <0.001 |
| Unknown | 57 | 2 | 16 | 23 | 16 | |
| HbA1c (%) | 5.3 (0.7) | 5.2 (0.4) | 5.2 (0.5) | 5.3 (0.6) | 5.4 (1.0) | 0.13 |
| Unknown | 112 | 13 | 32 | 38 | 29 | |
| Total cholesterol (mg/dL) | 192.6 (39.1) | 190.7 (38.7) | 189.9 (37.6) | 193.5 (39.1) | 195.9 (40.8) | 0.011 |
| Unknown | 143 | 18 | 44 | 43 | 38 | |
| C-reactive protein (mg/dL) | 0.3 (0.7) | 0.2 (0.3) | 0.3 (0.4) | 0.4 (0.7) | 0.5 (0.9) | <0.001 |
| Unknown | 128 | 16 | 39 | 40 | 33 | |
| Follow-up (years) | 17.6 (2.2) | 17.6 (2.3) | 17.7 (2.2) | 17.7 (2.2) | 17.5 (2.2) | 0.027 |
| Unknown | 3 | 0 | 3 | 0 | 0 | |
| All-cause death | 124 (3.7) | 13 (3.1) | 32 (3.1) | 43 (3.9) | 36 (4.5) | 0.4 |
| CVD death | 24 (0.7) | 2 (0.5) | 8 (0.8) | 7 (0.6) | 7 (0.9) | 0.8 |
| Cancer death | 42 (1.3) | 7 (1.7) | 12 (1.2) | 10 (0.9) | 13 (1.6) | 0.4 |
Baseline characteristics are shown for all 3359 participants who completed the treadmill examination; the mortality analyses in Table 2 used the complete-case analytic sample of 3291 participants (120 deaths) after excluding 68 with missing covariate data.
Bold P values indicate statistical significance (P < 0.05).
Mean (SD); n (%).
Kruskal-Wallis rank sum test; Pearson’s χ2 test.
bpm, beats per minute; CVD, cardiovascular disease; GED, General Educational Development; HbA1c, glycated haemoglobin; HRR, heart rate reserve; LRT, likelihood ratio test; NHANES, National Health and Nutrition Examination Survey; RHR, resting heart rate.
Figure 1. IPW Kaplan-Meier survival curves by resting heart rate: all-cause versus non-cardiovascular mortality. (A) and (B) IPW Kaplan-Meier estimates stratified by RHR clinical category (<60, 60–69, 70–79, ≥80 bpm) for all-cause mortality (A, 120 events) and non-cardiovascular (non-CVD) mortality (B, 99 events). (C) and (D) The corresponding estimates stratified by RHR tertile (T1 (lowest), T2, T3 (highest)) for all-cause (C) and non-CVD (D) mortality. Weights were derived from a multinomial propensity-score model including age, sex, race/ethnicity, education, BMI, smoking status, diabetes and hypertension, with truncation at the 99th percentile. Weighted likelihood-ratio test p values are annotated; the y-axis is truncated to 0.85–1.00. IPW numbers at risk reflect the reweighted pseudo-population, so each group approximates the full analytical sample. Non-CVD mortality treats cardiovascular deaths as censored. The analytical cohort comprised 3291 adults aged 20–49 years from NHANES 1999–2004, with follow-up through 31 December 2019 (median, 17.8 years; IQR, 16.3–19.2). BMI, body mass index; bpm, beats per minute; CVD, cardiovascular disease; IPW, inverse probability weighted; RHR, resting heart rate.
Cause-specific Cox models were fitted for CVD mortality, cancer mortality, accidental death and non-CVD mortality, with deaths from other causes treated as censored events. Subgroup analyses were conducted by age group (20–29, 30–39, 40–49), sex, race/ethnicity, smoking status, BMI category and hypertension status; interaction p values were obtained from likelihood ratio tests of cross-product terms. Fine-grained age-stratified analyses used 5-year age bands. The proportional-hazards assumption was verified using scaled Schoenfeld residuals (global p=0.17; RHR p=0.38). The cause-specific and subgroup analyses were considered exploratory; for the cause-specific outcomes we additionally report Benjamini-Hochberg false discovery rate q values.
Sensitivity analyses included exclusion of deaths within 2 and 5 years of baseline, use of an alternative RHR measurement (treadmill warm-up heart rate), additional adjustment for physical activity and laboratory values, survey-weighted Cox regression incorporating NHANES complex survey design and substitution of individual comorbidities (diabetes, hypertension) with a modified Charlson Comorbidity Index constructed from available NHANES questionnaire data,16 and Fine-Gray competing-risks models for the cause-specific outcomes.
Model discrimination was assessed using Harrell’s C-statistic, and internal validation was performed using 200 bootstrap resamples. Predicted 15-year mortality probabilities were estimated from the fully adjusted Cox model for representative demographic profiles to illustrate risk heterogeneity.
All analyses were conducted in R V.4.5.2. Two-sided p<0.05 was considered statistically significant.
Replication cohort (NHANES III)
We sought to replicate the primary findings in NHANES III (1988–1994), an earlier survey conducted under the same NCHS programme. RHR was measured as a 60 s radial pulse recorded at rest during the physical examination (NHANES III examination variable PEP6DR). Mortality was ascertained through the identical probabilistic National Death Index linkage through 31 December 2019, and cause of death was classified using the same NCHS 10-category underlying cause-of-death variable (UCOD_LEADING), with CVD mortality defined as heart disease or cerebrovascular disease and non-CVD mortality comprising all remaining causes. We applied the same age restriction (20–49 years) and the same covariate set, and we fitted identical sequentially adjusted Cox proportional hazards models, with RHR modelled per 10 beats per minute increment. HRR could not be assessed because NHANES III did not include an exercise treadmill examination.
Results
Study population
Of the 3359 participants who completed the treadmill examination (mean (SD) age, 33.2 (8.6) years; 51.9% male; mean (SD) RHR, 71.3 (10.7) beats per minute; table 1; online supplemental eFigure 1), 3291 had complete covariate data and constituted the analytical sample. Those with higher RHR tended to be younger, were more likely to smoke and had higher BMI (online supplemental eTable 1). Over a median follow-up of 17.8 years, 120 deaths occurred in the analytical sample.
Cause-of-death distribution
Cancer was the leading cause of death (42 (35.0%)), followed by NCHS residual causes (31 (25.8%); a category encompassing suicide, chronic liver disease and HIV) and accidents (20 (16.7%)). Cardiovascular disease accounted for only 21 deaths (17.5%). Overall, non-CVD causes accounted for 99 of 120 deaths (82.5%; online supplemental eTable 2).
RHR and all-cause mortality
In the fully adjusted model, each 10 beats per minute increase in RHR was associated with 26% higher all-cause mortality (HR, 1.26; 95% CI 1.07 to 1.50; p=0.007; table 2). IPW-adjusted Kaplan-Meier curves showed visible survival divergence across RHR categories (weighted likelihood ratio test [LRT] p<0.001; figure 1, Panels A–B), with the same separation when RHR was dichotomised at 80 beats per minute (online supplemental eFigure 4), and the association followed a monotonic gradient across RHR quintiles (Q5 vs Q1: HR, 1.94; 95% CI 1.07 to 3.53; p for trend=0.014; online supplemental eTable 3). Restricted cubic spline analysis indicated a linear dose-response relationship (p for non-linearity=0.88; online supplemental eFigure 7, Panel A).
Table 2. Association of resting heart rate and heart rate reserve with all-cause and cause-specific mortality among US adults aged 20–49 years, NHANES 1999–2004.
| Events/N | Person-years | Rate* | HR (95% CI) | P value | |
|---|---|---|---|---|---|
| All-cause mortality | |||||
| RHR, per 10 bpm | 120/3291 | 58 025 | 2.07 | 1.26 (1.07 to 1.50) | 0.007 |
| RHR categories, bpm | |||||
| <60 (reference) | 13/416 | 7334 | 1.77 | 1 (Ref) | — |
| 60–69 | 31/1004 | 17 761 | 1.75 | 1.11 (0.58 to 2.13) | 0.752 |
| 70–79 | 41/1093 | 19 351 | 2.12 | 1.49 (0.79 to 2.80) | 0.219 |
| ≥80 | 35/778 | 13 578 | 2.58 | 1.78 (0.92 to 3.42) | 0.086 |
| P for trend | 0.030 | ||||
| Cause-specific mortality† | |||||
| Non-CVD mortality | 99/3291 | — | — | 1.28 (1.07 to 1.55) | 0.009 |
| Cancer | 42/3291 | — | — | 1.19 (0.89 to 1.58) | 0.248 |
| Accidental death | 20/3291 | — | — | 1.42 (0.95 to 2.14) | 0.089 |
| Other non-CVD causes | 37/3291 | — | — | 1.30 (0.96 to 1.77) | 0.091 |
| CVD mortality | 21/3291 | — | — | 1.15 (0.77 to 1.71) | 0.508 |
| Exploratory composites‡ | |||||
| Behavioural/external | 51/3291 | — | — | 1.35 (1.05 to 1.75) | 0.021 |
| Disease-oriented | 48/3291 | — | — | 1.19 (0.91 to 1.57) | 0.204 |
| Heart rate reserve | |||||
| HRR, per 10 bpm | 120/3291 | 58 025 | 2.07 | 1.01 (0.90 to 1.14) | 0.807 |
| Mutual adjustment§ | |||||
| RHR, per 10 bpm | — | — | — | 1.39 (1.14 to 1.70) | 0.001 |
| HRR, per 10 bpm | — | — | — | 1.13 (0.99 to 1.30) | 0.063 |
All HRs are from the fully adjusted model (age, sex, race/ethnicity, education, body mass index, smoking status, diabetes and hypertension). The em dash (—) indicates not estimated; person-years and mortality rates are reported only for the all-cause and resting heart rate category analyses. For the all-cause association, the unadjusted and partially adjusted (age, sex and race/ethnicity) HRs per 10 bpm were 1.19 (1.01–1.40) and 1.31 (1.11–1.54), respectively.
Per 1000 person-years.
Cause-specific Cox models treated deaths from other causes as censored events.
Behavioural/external composite: accidental deaths plus NCHS residual-category deaths (UCOD category 10, predominantly suicide, chronic liver disease and other behavioural causes); disease-oriented composite: cancer, CLRD, DM and nephritis.
RHR and HRR entered simultaneously in the fully adjusted model.
bpm, beats per minute; CLRD, chronic lower respiratory disease; CVD, cardiovascular disease; DM, diabetes mellitus; HRR, heart rate reserve; NCHS, National Center for Health Statistics; RHR, resting heart rate; UCOD, underlying cause of death.
RHR and cause-specific mortality
Because most deaths in this young cohort were non-cardiovascular (99 of 120), the RHR-associated mortality signal was carried predominantly by non-CVD causes (figure 2). RHR was associated with non-CVD mortality (HR, 1.28; 95% CI 1.07 to 1.55; p=0.009). The CVD-specific estimate was positive but non-significant, limited by the small number of cardiovascular deaths (21 events; HR, 1.15; 95% CI 0.77 to 1.71; p=0.51); this wide CI is compatible with effect sizes close to the non-CVD estimate, so these data cannot establish that RHR selectively spares cardiovascular death. The cause-specific estimates were essentially unchanged in Fine-Gray competing-risks models (online supplemental eTable 8e).
Figure 2. Cause-specific mortality: resting heart rate and heart rate reserve. Forest plot of HRs (per 10 bpm) and 95% CIs for resting heart rate (red circles) and heart rate reserve (blue triangles), organised in three sections. Upper section (primary decomposition): all-cause (120 events), cardiovascular (CVD; 21 events) and non-CVD (99 events) mortality, with non-CVD subcategories (cancer, 42 events; accidental, 20 events; other non-CVD, 37 events). Middle section (composite reclassification): a behavioural and external-cause composite (51 events), non-CVD excluding cancer (57 events) and a disease-related non-CVD composite (48 events). Lower section (age-stratified, ≥35 years): non-CVD (57 events) and cancer (28 events) mortality. All models were adjusted for age, sex, race/ethnicity, education, BMI, smoking status, diabetes and hypertension. BMI, body mass index; bpm, beats per minute; CVD, cardiovascular disease; HRR, heart rate reserve; RHR, resting heart rate.
Among exploratory composites, a behavioural and external-cause grouping (accidental deaths plus NCHS residual-category deaths; 51 events) was associated with RHR (HR, 1.35; 95% CI 1.05 to 1.75; p=0.02), whereas a disease-oriented grouping (cancer, chronic lower respiratory disease, diabetes and nephritis; 48 events) was not (HR, 1.19; p=0.20; online supplemental eTable 5). Individual non-CVD subcategories showed consistently elevated but individually non-significant HRs: accidental death (HR, 1.42; p=0.09), other non-CVD causes (HR, 1.30; p=0.09) and cancer (HR, 1.19; p=0.25). Among the 111 deaths without diabetes or hypertension on the death certificate, the RHR association remained significant (HR, 1.26; 95% CI 1.05 to 1.50; p=0.01). Across the cause-specific outcomes, the non-CVD association remained significant after Benjamini-Hochberg correction (q=0.044), whereas the exploratory composites did not.
When restricted to participants aged 35 years and older, the RHR association with non-CVD mortality strengthened (57 events; HR, 1.52; 95% CI 1.19 to 1.94; p<0.001) and cancer-specific mortality also became significant (28 events; HR, 1.50; 95% CI 1.05 to 2.13; p=0.03; figure 2).
The IPW-adjusted Kaplan-Meier curves for non-CVD mortality showed clear separation across RHR categories (figure 1, Panels B and D), and dose-response analyses showed significant linear associations for RHR with both all-cause (p linear=0.007) and non-CVD mortality (p linear=0.009), with no evidence of non-linearity (online supplemental eFigure 7, Panels A–B).
Age-dependent emergence
The RHR-mortality association strengthened with age. Among participants younger than 35 years, the association was not statistically significant (HR, 0.98; 95% CI 0.74 to 1.30; p=0.88), whereas among those aged 35–49 years it was strong (HR, 1.48; 95% CI 1.19 to 1.83; p<0.001). The association was strongest at ages 35–39 (HR, 2.60; 95% CI 1.46 to 4.64; p=0.001; online supplemental eFigure 5), with continued significance at ages 40–44 (HR, 1.43; 95% CI 1.02 to 2.01; p=0.038; online supplemental eTable 10).
Heart rate reserve
HRR was not associated with all-cause mortality (HR, 1.01; 95% CI 0.90 to 1.14; p=0.81; online supplemental eTable 4) or any cause-specific outcome (all p>0.18). RHR and HRR were weakly correlated (onlinesupplemental eFigures 2 3). This null finding was consistent across all age strata, sex, race/ethnicity and comorbidity subgroups (online supplemental eFigure 8; online supplemental eTable 9).
Risk heterogeneity across demographic profiles
To translate HRs into absolute terms, we estimated 15-year all-cause mortality at RHR=80 beats per minute for 30 demographic profiles defined by age, sex and race/ethnicity (online supplemental eFigure 9; online supplemental eFigure 6). Predicted mortality rose with age, was roughly two times as high in men as in women, and varied further by race/ethnicity, spanning a more than sixfold gradient at a single RHR value (from 1.0% in a 25-year-old Mexican American woman to 6.9% in a 45-year-old non-Hispanic black man). A fixed RHR threshold therefore cannot capture individual-level risk, which should be interpreted in demographic context.
Sensitivity analyses
Deaths were distributed across the entire follow-up period without early clustering (online supplemental eTable 8c), and the RHR-all-cause mortality association was robust across sensitivity analyses (online supplemental eTable 8), including exclusion of early deaths (≤2 and ≤5 years), an alternative RHR measurement and additional adjustment for physical activity and laboratory values (HR range, 1.24–1.38). Substituting a modified Charlson Comorbidity Index for individual comorbidities yielded virtually identical results (HR, 1.28; online supplemental eTable 8d). Survey-weighted analysis was directionally consistent but attenuated (HR, 1.18; 95% CI 0.99 to 1.42; online supplemental eTable 6). The E-value for unmeasured confounding was 1.84 (lower bound, 1.33), and bootstrap-corrected discrimination was modest (C-statistic, 0.682; online supplemental eTable 6 and 8b).
Replication in NHANES III
In NHANES III (1988–1994), an earlier sample from the same survey programme linked to mortality through 2019, the analytical cohort comprised 8941 adults aged 20–49 years, with 1317 deaths over a median follow-up of 27.6 years. The all-cause association was replicated: each 10 beats per minute increase in RHR was associated with higher mortality (HR, 1.17; 95% CI 1.12 to 1.22), with a graded relationship across the RHR distribution. Unlike the smaller main cohort, this larger sample was powered to detect a cardiovascular signal, and RHR predicted both non-CVD (HR, 1.17; 95% CI 1.11 to 1.23) and CVD mortality (HR, 1.18; 95% CI 1.08 to 1.29) at similar magnitude, including cancer death (HR, 1.16; 95% CI 1.06 to 1.26). The age pattern mirrored the primary analysis: the association strengthened with age and was strongest in the mid-to-late 30s (35–39 years: HR, 1.26), yet was already detectable by the early 30s (30–34 years: HR, 1.19; 95% CI 1.05 to 1.35). Estimates were consistent across strata of sex, race and ethnicity and smoking status (range, 1.10–1.21). Across cohorts, the all-cause point estimates were closely aligned (figure 3).
Figure 3. Resting heart rate and mortality: primary cohort and NHANES III replication cohort. Forest plot of the fully adjusted HR per 10 bpm higher resting heart rate for all-cause, non-CVD and CVD mortality in the primary cohort (NHANES 1999–2004; N=3291; 120 deaths) and the replication cohort (NHANES III, 1988–1994; N=8941; 1317 deaths). Both cohorts used identical Cox proportional-hazards models adjusted for age, sex, race/ethnicity, education, body mass index, smoking status, diabetes and hypertension, with resting heart rate modelled per 10 bpm increment. The dashed vertical line indicates an HR of 1.00; horizontal bars are 95% CIs. bpm, beats per minute; CVD, cardiovascular disease.
Discussion
In this cohort of 3291 US adults aged 20–49 years, elevated RHR predicted all-cause mortality. Three features stand out. First, RHR behaved as a broad survival marker rather than a cardiovascular-specific one: although non-cardiovascular causes dominated, NHANES III showed RHR predicting cardiovascular and non-cardiovascular death to a similar degree. Second, the association strengthened with age, emerging by the early-to-mid 30s, well before the age (around 40) at which current cardiovascular risk tools apply. Third, submaximal HRR carried no prognostic value.
Resting heart rate as a broad marker of survival
At these ages the prognostic reach of RHR runs mainly through non-cardiovascular death, which accounted for 99 of the 120 deaths; the cardiovascular-specific estimate was positive but imprecise (21 deaths; HR, 1.15), with a CI compatible with an effect as large as the non-cardiovascular one. We therefore read RHR as a general survival marker rather than a specifically cardiovascular one, consistent with Greenland et al,10 who reported that RHR predicts both cardiovascular and non-cardiovascular mortality, including in younger men. In older populations the RHR-mortality association is largely mediated by coronary disease, heart failure and sudden cardiac death1,4; in young adults, whose atherosclerotic burden is minimal,11 12 elevated RHR may instead mark broader health vulnerability.
Exploratory cause-specific analyses did not implicate any single cause of death: the behavioural and external-cause composite was associated with RHR whereas the disease-oriented composite was not (online supplemental eTable 5), although both were underpowered and require confirmation in larger cohorts with detailed cause-of-death data.
Several mechanisms could plausibly link elevated RHR to non-cardiovascular death in young adults. Among the 31 deaths in the NCHS residual category, none had baseline diabetes and only 6.5% had hypertension; national vital statistics indicate that this category at these ages is composed largely of suicide, homicide and chronic liver disease.17 18 For external-cause deaths specifically, elevated RHR plausibly serves as a physiological readout of heightened sympathetic and reduced vagal tone, the same autonomic state that accompanies chronic psychological distress, impulsivity, substance and alcohol use, poor sleep and psychiatric comorbidity and that predicts subsequent suicide in large cohort studies.19,23 Through this shared substrate, the conditions that raise resting sympathetic tone may also heighten risk-taking and impair judgement, while heavy alcohol use and poor sleep can degrade coordination and reaction time; these are plausible pathways to drug overdose, motor vehicle incidents and suicide, the leading external causes of death at these ages.24 RHR may thus mark a behavioural-risk phenotype that conventional covariates capture poorly. Elevated RHR is also a marker of chronic low-grade inflammation, which may contribute to cancer mortality25 26; higher RHR has itself been associated with cancer mortality in a meta-analysis,27 and the RHR-cancer association reached significance only at ages 35 and older (HR, 1.50; p=0.03).
In the larger NHANES III cohort (8941 adults, 1317 deaths over a median 27.6 years), the greater statistical power clarified the cause-specific picture: RHR predicted non-cardiovascular (HR, 1.17; 95% CI 1.11 to 1.23) and cardiovascular death (HR, 1.18; 95% CI 1.08 to 1.29) to a nearly identical degree, with a parallel age gradient. This concordance does not show that RHR spares cardiovascular causes; rather, it reinforces that elevated RHR is a broad, robust marker of survival vulnerability rather than a specifically cardiovascular one, extending earlier within-NHANES observations8 9 to young adults across two cohorts examined a decade apart.
The stronger non-cardiovascular signal in the primary cohort may also reflect period effects: its follow-up captured the opioid-era rise in external-cause deaths, whereas the older NHANES III cohort accrued more deaths at higher attained ages, where cardiovascular causes predominate. The weaker cardiovascular signal in the primary cohort likely reflects its limited events, not its shorter follow-up: a follow-up-matched 18-year truncation of NHANES III left the association unchanged (HR, 1.18; 95% CI 1.04 to 1.33; online supplemental eTable 11).
Age-dependent emergence
This age-dependent strengthening likely reflects the gradual accumulation of subclinical risk that turns elevated RHR into a clinically meaningful marker12; consistent with this, a large cohort study reported more pronounced per-beats per minute RHR-mortality associations at younger ages.28 The association emerged a decade before the starting age of current risk-stratification tools (both Systematic Coronary Risk Evaluation 2 (SCORE2) and the Pooled Cohort Equations begin at age 40),29,31 suggesting a potential gap in preventive assessment for adults in their 30s. Even the recent PREVENT equations, which lower quantitative risk estimation to age 30, do not incorporate RHR.32
HRR under submaximal conditions
The consistently null HRR finding carries practical implications. HRR measured at submaximal intensity (75% of age-predicted maximum) may lack sufficient physiological stress to unmask chronotropic incompetence, which has been linked to mortality in studies using maximal exercise testing.14 15 HRR from submaximal protocols, as used in population surveys and primary care, should therefore be interpreted with caution.
Clinical implications
These findings suggest that elevated RHR in young adults need not be interpreted solely through a cardiovascular lens. Adding RHR to a model already dominated by age barely improved discrimination (C-statistic, 0.709–0.710); its value therefore lies less in sharpening individual prediction than in serving as an inexpensive, universally recorded signal in routine clinical encounters and in supporting population-level risk communication, consistent with a marker of vulnerability rather than an individual risk calculator. A persistently elevated reading is simple to measure and may flag younger adults warranting attention to contributors beyond cardiovascular risk factors, such as psychological distress, substance use, sleep disorders or chronic inflammation. RHR also partly reflects cardiorespiratory fitness, yet additionally adjusting for estimated VO₂max from the treadmill test attenuated the association only modestly (HR, 1.21; 95% CI 1.02 to 1.45; online supplemental eTable 8), consistent with prognostic information beyond fitness, although this estimate is an imperfect index of fitness and residual confounding cannot be fully excluded.
Predicted 15-year mortality ranged from 1% to 7% across demographic profiles at RHR=80 beats per minute, so interpretation should be individualised rather than tied to a single universal cut-point. A pragmatic prompt nonetheless helps at the bedside: an RHR persistently at or above 80 beats per minute, roughly the upper quartile to upper decile for age and sex (online supplemental eTable 12), in an otherwise apparently healthy adult in their late 30s or 40s is a reasonable trigger to evaluate modifiable contributors such as physical inactivity, poor sleep, excess alcohol and thyroid dysfunction. To make the absolute stakes explicit, predicted 15-year mortality across RHRs from 65 to 90 beats per minute (online supplemental eTable 13) rises only modestly in low-risk young adults but substantially in those with adverse profiles. Although the relative hazard associated with RHR did not differ significantly by sex or race/ethnicity (all p interaction >0.67; online supplemental eTable 9), absolute risk varied widely because baseline mortality differs across these groups. These differences reflect well-documented social and structural determinants of health,33,35 not the RHR-mortality association, which was consistent across groups.
Strengths and limitations
This study has several strengths. It draws on two large, nationally representative US cohorts with standardised National Death Index linkage, replicating the primary finding in a second, earlier NHANES cohort. The consistency of the all-cause association across cohorts, age, sex, ethnicity and smoking status, with a graded dose-response, supports its robustness. It is also, to our knowledge, the first study to focus on adults aged 20–49 years, to identify the age at which the association emerges and to evaluate submaximal HRR.
Several limitations merit discussion. First, in the primary cohort the modest number of deaths (120, of which 21 were cardiovascular) limited power for cause-specific estimates, and the exploratory composites should be interpreted cautiously; the larger NHANES III cohort, with 354 cardiovascular deaths, directly addressed this. Second, NHANES public-use mortality files group cause of death into 10 broad categories (UCOD_LEADING), and the residual category 10 cannot be split into individual causes (eg, suicide, homicide, chronic liver disease) without restricted-use data, so we characterised its composition from national vital statistics for younger age groups.17 18 Third, RHR was measured once at baseline; single-visit RHR has acceptable reproducibility in epidemiological studies,36 and any regression dilution from this single measurement would bias estimates toward the null, so the true association is likely stronger than reported, making our findings conservative.37 Fourth, residual confounding by unmeasured behavioural and psychosocial factors cannot be excluded, although the E-value of 1.84 (lower bound, 1.33) indicates that explaining away the association would require a moderately strong unmeasured confounder. Estimates were unchanged when a modified Charlson Comorbidity Index replaced individual comorbidities (HR, 1.26–1.28; all p<0.01; online supplemental eTable 8d)16 and when analysis was restricted to deaths without diabetes or hypertension on the death certificate (111 events; HR, 1.26; p=0.01). Because the Cox models adjusted for the demographic factors underlying NHANES sample weighting, unweighted estimation can provide valid and more efficient estimates38; the survey-weighted analysis was directionally consistent (HR, 1.18; online supplemental eTable 7), and the large validation cohort provided precise confirmation. Fifth, although the central finding replicated in NHANES III, both samples share the same survey programme, examination procedures and mortality-linkage methodology, so this represents temporal replication within a single survey framework rather than fully independent external validation; generalisability to non-US populations, which differ in background external-cause mortality and physical-activity patterns, remains to be established. Finally, amid growing concern over formulaic analyses of public datasets,39 this study was hypothesis-driven and replicated across two cohorts.
Conclusions
Across two national cohorts of US adults aged 20–49 years (NHANES 1999–2004 and NHANES III), elevated RHR predicted long-term all-cause mortality. The association strengthened with age, becoming evident a decade before the starting age of current cardiovascular risk tools, and HRR measured under submaximal conditions carried no prognostic value. These findings suggest that RHR in young adults may serve as a broad marker of survival vulnerability rather than a specifically cardiovascular risk factor. Because it is simple and free to measure, the resting pulse could help identify younger adults who merit closer attention before they reach the ages targeted by existing risk tools.
Supplementary material
Footnotes
Funding: This work was supported by the Non-communicable Chronic Diseases: National Science and Technology Major Project (grant number 2023ZD0509400). The funder had no role in the design and conduct of the study; the collection, management, analysis or interpretation of the data; the preparation, review or approval of the manuscript; or the decision to submit the manuscript for publication.
Provenance and peer review: Not commissioned; externally peer reviewed.
Patient consent for publication: Not applicable.
Ethics approval: This study involves human participants and was approved by National Center for Health Statistics (NCHS) Research Ethics Review Board (ERB). The NHANES 1999–2004 protocol was approved by the NCHS ERB (Protocol No. 98-12); NHANES III (1988–1994) was approved by the NCHS Institutional Review Board. The present study is a secondary analysis of de-identified, publicly available NHANES data and was exempt from further institutional review board review; all NHANES participants provided written informed consent.
Data availability free text: The NHANES 1999–2004 and NHANES III data used in this study are publicly available from the National Center for Health Statistics (https://www.cdc.gov/nchs/nhanes/). The linked mortality files are available from the NCHS (https://www.cdc.gov/nchs/data-linkage/mortality-public.htm). The complete analysis code and de-identified analytic datasets are provided as an online supplementary reproducibility archive (R scripts, analytic datasets, result tables and a README).
Data availability statement
Data are available in a public, open access repository. All data relevant to the study are included in the article or uploaded as supplementary information.
References
- 1.Zhang D, Shen X, Qi X. Resting heart rate and all-cause and cardiovascular mortality in the general population: a meta-analysis. CMAJ. 2016;188:E53–63. doi: 10.1503/cmaj.150535. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Aune D, Sen A, Ó’Hartaigh B, et al. Resting heart rate and the risk of cardiovascular disease, total cancer, and all-cause mortality - A systematic review and dose-response meta-analysis of prospective studies. Nutr Metab Cardiovasc Dis. 2017;27:504–17. doi: 10.1016/j.numecd.2017.04.004. [DOI] [PubMed] [Google Scholar]
- 3.Kannel WB, Kannel C, Paffenbarger RS, Jr, et al. Heart rate and cardiovascular mortality: The Framingham study. Am Heart J. 1987;113:1489–94. doi: 10.1016/0002-8703(87)90666-1. [DOI] [PubMed] [Google Scholar]
- 4.Woodward M, Webster R, Murakami Y, et al. The association between resting heart rate, cardiovascular disease and mortality: evidence from 112,680 men and women in 12 cohorts. Eur J Prev Cardiolog. 2014;21:719–26. doi: 10.1177/2047487312452501. [DOI] [PubMed] [Google Scholar]
- 5.Palatini P, Julius S. Elevated heart rate: a major risk factor for cardiovascular disease. Clin Exp Hypertens. 2004;26:637–44. doi: 10.1081/ceh-200031959. [DOI] [PubMed] [Google Scholar]
- 6.Böhm M, Reil J-C, Deedwania P, et al. Resting heart rate: risk indicator and emerging risk factor in cardiovascular disease. Am J Med. 2015;128:219–28. doi: 10.1016/j.amjmed.2014.09.016. [DOI] [PubMed] [Google Scholar]
- 7.Olshansky B, Ricci F, Fedorowski A. Importance of resting heart rate. Trends Cardiovasc Med. 2023;33:502–15. doi: 10.1016/j.tcm.2022.05.006. [DOI] [PubMed] [Google Scholar]
- 8.Gillum RF, Makuc DM, Feldman JJ. Pulse rate, coronary heart disease, and death: the NHANES I Epidemiologic Follow-up Study. Am Heart J. 1991;121:172–7. doi: 10.1016/0002-8703(91)90970-s. [DOI] [PubMed] [Google Scholar]
- 9.Alhalabi L, Singleton MJ, Oseni AO, et al. Relation of Higher Resting Heart Rate to Risk of Cardiovascular Versus Noncardiovascular Death. Am J Cardiol. 2017;119:1003–7. doi: 10.1016/j.amjcard.2016.11.059. [DOI] [PubMed] [Google Scholar]
- 10.Greenland P, Daviglus ML, Dyer AR, et al. Resting heart rate is a risk factor for cardiovascular and noncardiovascular mortality: the Chicago Heart Association Detection Project in Industry. Am J Epidemiol. 1999;149:853–62. doi: 10.1093/oxfordjournals.aje.a009901. [DOI] [PubMed] [Google Scholar]
- 11.Virani SS, Alonso A, Aparicio HJ, et al. Heart Disease and Stroke Statistics-2021 Update: A Report From the American Heart Association. Circulation. 2021;143:e254–743. doi: 10.1161/CIR.0000000000000950. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.McGill HC, Jr, McMahan CA. Determinants of atherosclerosis in the young. Pathobiological Determinants of Atherosclerosis in Youth (PDAY) Research Group. Am J Cardiol. 1998;82:30T–36T. doi: 10.1016/s0002-9149(98)00720-6. [DOI] [PubMed] [Google Scholar]
- 13.Brubaker PH, Kitzman DW. Chronotropic incompetence: causes, consequences, and management. Circulation. 2011;123:1010–20. doi: 10.1161/CIRCULATIONAHA.110.940577. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Cole CR, Blackstone EH, Pashkow FJ, et al. Heart-rate recovery immediately after exercise as a predictor of mortality. N Engl J Med. 1999;341:1351–7. doi: 10.1056/NEJM199910283411804. [DOI] [PubMed] [Google Scholar]
- 15.Lauer MS, Francis GS, Okin PM, et al. Impaired chronotropic response to exercise stress testing as a predictor of mortality. JAMA. 1999;281:524–9. doi: 10.1001/jama.281.6.524. [DOI] [PubMed] [Google Scholar]
- 16.Charlson ME, Pompei P, Ales KL, et al. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–83. doi: 10.1016/0021-9681(87)90171-8. [DOI] [PubMed] [Google Scholar]
- 17.Curtin SC, Tejada-Vera B, Bastian BA. Deaths: Leading Causes for 2020. Natl Vital Stat Rep. 2023;72:1–115. [PubMed] [Google Scholar]
- 18.Tejada-Vera B, Bastian BA, Curtin SC. Deaths: Leading Causes for 2023. Natl Vital Stat Rep. 2025;74:1–115. doi: 10.15620/cdc/174607. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Thayer JF, Lane RD. A model of neurovisceral integration in emotion regulation and dysregulation. J Affect Disord. 2000;61:201–16. doi: 10.1016/s0165-0327(00)00338-4. [DOI] [PubMed] [Google Scholar]
- 20.Alvares GA, Quintana DS, Hickie IB, et al. Autonomic nervous system dysfunction in psychiatric disorders and the impact of psychotropic medications: a systematic review and meta-analysis. J Psychiatry Neurosci. 2016;41:89–104. doi: 10.1503/jpn.140217. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Raine A. Autonomic nervous system factors underlying disinhibited, antisocial, and violent behavior. Biosocial perspectives and treatment implications. Ann N Y Acad Sci. 1996;794:46–59. doi: 10.1111/j.1749-6632.1996.tb32508.x. [DOI] [PubMed] [Google Scholar]
- 22.Wilson ST, Chesin M, Fertuck E, et al. Heart rate variability and suicidal behavior. Psychiatry Res. 2016;240:241–7. doi: 10.1016/j.psychres.2016.04.033. [DOI] [PubMed] [Google Scholar]
- 23.Chang S-S, Bjørngaard JH, Tsai MK, et al. Heart rate and suicide: findings from two cohorts of 533 000 Taiwanese and 75 000 Norwegian adults. Acta Psychiatr Scand. 2016;133:277–88. doi: 10.1111/acps.12513. [DOI] [PubMed] [Google Scholar]
- 24.Heron M. Deaths: Leading Causes for 2019. Natl Vital Stat Rep. 2021;70:1–114. [PubMed] [Google Scholar]
- 25.Whelton SP, Narla V, Blaha MJ, et al. Association between resting heart rate and inflammatory biomarkers (high-sensitivity C-reactive protein, interleukin-6, and fibrinogen) (from the Multi-Ethnic Study of Atherosclerosis) Am J Cardiol. 2014;113:644–9. doi: 10.1016/j.amjcard.2013.11.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Coussens LM, Werb Z. Inflammation and cancer. Nature. 2002;420:860–7. doi: 10.1038/nature01322. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Pozuelo-Carrascosa DP, Cavero-Redondo I, Lee IM, et al. Resting Heart Rate as a Predictor of Cancer Mortality: A Systematic Review and Meta-Analysis. J Clin Med. 2021;10:1354. doi: 10.3390/jcm10071354. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Raisi-Estabragh Z, Cooper J, Judge R, et al. Age, sex and disease-specific associations between resting heart rate and cardiovascular mortality in the UK BIOBANK. PLoS One. 2020;15:e0233898. doi: 10.1371/journal.pone.0233898. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Hageman S, Pennells L, Ojeda F, et al. SCORE2 risk prediction algorithms: new models to estimate 10-year risk of cardiovascular disease in Europe. Eur Heart J. 2021;42:2439–54. doi: 10.1093/eurheartj/ehab309. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Goff DC, Jr, Lloyd-Jones DM, Bennett G, et al. 2013 ACC/AHA guideline on the assessment of cardiovascular risk: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2014;129:S49–73. doi: 10.1161/01.cir.0000437741.48606.98. [DOI] [PubMed] [Google Scholar]
- 31.Arnett DK, Blumenthal RS, Albert MA, et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease. Circulation. 2019;140:e596–e646. doi: 10.1161/CIR.0000000000000678. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Khan SS, Matsushita K, Sang Y, et al. Development and Validation of the American Heart Association’s PREVENT Equations. Circulation. 2024;149:430–49. doi: 10.1161/CIRCULATIONAHA.123.067626. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Wong MD, Chung AK, Boscardin WJ, et al. The contribution of specific causes of death to sex differences in mortality. Public Health Rep. 2006;121:746–54. doi: 10.1177/003335490612100615. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Cunningham TJ, Croft JB, Liu Y, et al. Vital Signs: Racial Disparities in Age-Specific Mortality Among Blacks or African Americans - United States, 1999-2015. MMWR Morb Mortal Wkly Rep. 2017;66:444–56. doi: 10.15585/mmwr.mm6617e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Geronimus AT, Hicken M, Keene D, et al. “Weathering” and age patterns of allostatic load scores among blacks and whites in the United States. Am J Public Health. 2006;96:826–33. doi: 10.2105/AJPH.2004.060749. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Stanforth PR, Gagnon J, Rice T, et al. Reproducibility of resting blood pressure and heart rate measurements. The HERITAGE Family Study. Ann Epidemiol. 2000;10:271–7. doi: 10.1016/s1047-2797(00)00047-8. [DOI] [PubMed] [Google Scholar]
- 37.Clarke R, Shipley M, Lewington S, et al. Underestimation of risk associations due to regression dilution in long-term follow-up of prospective studies. Am J Epidemiol. 1999;150:341–53. doi: 10.1093/oxfordjournals.aje.a010013. [DOI] [PubMed] [Google Scholar]
- 38.Korn EL, Graubard BI. Epidemiologic studies utilizing surveys: accounting for the sampling design. Am J Public Health. 1991;81:1166–73. doi: 10.2105/ajph.81.9.1166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Suchak T, Aliu AE, Harrison C, et al. Explosion of formulaic research articles, including inappropriate study designs and false discoveries, based on the NHANES US national health database. PLoS Biol. 2025;23:e3003152. doi: 10.1371/journal.pbio.3003152. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data are available in a public, open access repository. All data relevant to the study are included in the article or uploaded as supplementary information.