Non-exercise estimated cardiorespiratory fitness and mortality from all-causes, cardiovascular disease and cancer in the NIH-AARP Diet and Health Study

Abstract

Background:

Impractical methods and relatively small cohort have limited the applications of non-exercise estimated cardiorespiratory fitness (NEE-CRF). This study aimed to assess the association between a pragmatic NEE-CRF method and mortality outcomes in a large prospective cohort.

Methods:

A total of 330,769 participants [men (n=186,469) and women (n=144,300)] aged 50-71 years from the NIH-AARP Diet and Health Study were assessed at baseline (1995-1996) and prospectively followed until December 31, 2015 (14.9±2.1 years). NEE-CRF was estimated using pragmatic and previously validated equation, and Cox hazard analysis for mortality was conducted.

Results:

NEE-CRF was 9.9±1.5 METs in men and 7.2±1.6 METs in women. In total, 34,317 men and 20,295 women died during the follow-up. Higher NEE-CRF was associated with lower mortality risk from all-causes, cardiovascular disease and cancer. Compared to the lowest quartile of NEE-CRF, the HRs and 95% CI for all-cause mortality in the second, third and fourth quartiles were: 0.82 (0.79-0.84), 0.74 (0.72-0.77) and 0.70 (0.67-0.73) for men, and 0.84 (0.81-0.88), 0.78 (0.75-0.82) and 0.72 (0.68-0.77) for women (p trend <0.001 for all). For each 1-MET increase in NEE-CRF, risks for mortality due to cardiovascular disease and cancer were 0.85 (0.82-0.88) and 0.89 (0.87-0.91) in men, and 0.84 (0.81-0.88) and 0.89 (0.87-0.91) in women, respectively (p<0.001 for all).

Conclusion:

Higher NEE-CRF is independently associated with lower mortality risk in a large prospective cohort of men and women. These results support the utility of the applied NEE-CRF method for risk stratification, prevention and rehabilitation programs and application in large epidemiological studies.

Introduction

Cardiorespiratory fitness (CRF) is an established health marker that is strongly associated with mortality outcomes.^1–4 CRF is commonly quantified by maximal oxygen consumption (VO₂ max) and represents the integrated physiological capacity of the heart, lungs and skeletal muscles to supply the metabolic demands of the body during exercise.¹ Mounting evidence has demonstrated that higher CRF provides protective benefits against morbidity and mortality, while low CRF is an important modifiable risk factor associated with higher risk of all-cause, cardiovascular and cancer mortality.^1–4 A growing number of studies have suggested that low CRF is a more powerful predictor of mortality than established risk factors including smoking, hypertension, dyslipidemia, coronary artery disease and type-2 diabetes.^{1, 5–7} Unlike other health risk factors, CRF is not routinely measured in clinical and research settings,^{1, 8, 9} in part because the direct measurement of CRF with standardized exercise testing requires considerable time and expense, including laboratory equipment, qualified personnel to administer the test and medical supervision.^{1, 8–13}

Non-exercise estimated CRF (NEE-CRF) is a method that has been increasingly used in recent years to determine CRF using validated prediction equations.¹ NEE-CRF was recently recognized by Amercian Heart Association as alternative approach to standard exercise testing, when the latter is impractical.¹ In this approach, equations using self-reported or easily measurable variables such as age, weight, height and self-reported physical activity have been developed and validated against the gold standard, a directly measure of VO₂max from an exercise test.^{1, 14} Although direct measure of CRF is preferable, in many health-care and research setting this is not feasible particularly in low-resource clinics and in large epidemiological studies where exercise testing of large numbers of participants is impractical.^{1, 8–10} Even though early observations have demonstrated associations between NEE-CRF and mortality outcomes, methodological limitations in these studies remain challenging for a broad application.^{8, 9, 15–18} For instance, in the largest previous study (n=43,356) of NEE-CRF, all-cause and cardiovascular mortality included only 21% women.⁸ To our knowledge, the association between NEE-CRF and cancer mortality has been assessed in only one study with a relatively small sample size (n= 8,506) for epidemiological studies.¹⁵ A major limitation in previous studies of the association between NEE-CRF and mortality outcomes is their poor practicality. ^{8, 9, 15–18} Particulary, utilization of predicting equations for CRF requiring physical measurements and patient contact, for variables such as resting heart rate and waist circumference, which limit their utility in low-resource clinical settings and large epidemiological studies.^{8, 9, 15–18} Therefore, the aim of the present study was to assess the association between a pragmatic method of NEE-CRF and mortality outcomes from all-causes, cardiovascular disease and cancer in a large prospective cohort of men and women.

Methods

Study population and design

The NIH-AARP Diet and Health Study (www.clinicaltrials.gov; NCT00340015) has been previously described.¹⁹ In brief, between 1995 and 1996 a questionnaire on demographics, medical history, dietary and lifestyle behaviors was mailed to 3.5 million American Association of Retired Persons (AARP) members. The cohort included AARP members aged 50-71 years who resided in six states (California, Florida, Louisiana, New Jersey, North Carolina, and Pennsylvania) and two metropolitan areas (Atlanta, GA and Detroit, MI) with existing population-based cancer registries. In total, 617,119 questionnaires were returned and 566,398 were satisfactorily completed (92%). The satisfactory criteria included indication of sex, skipping only small portions of the questionnaire, having less than 10 recording errors and consumtion of more than 10 foods.¹⁹

We excluded participants whose questionnaires were completed by a spouse or other surrogates (n=15,767), participants with baseline self-reported diagnoses of cancer (n=50,591), cardiovascular disease (n=80,254), stroke (n=12,812), end-stage chronic kidney disease (n=1,299), those with missing information for estimating CRF (n=43,450) and participants with less than 5 years follow up (n=31,456). The resulting analytic cohort included 330,769 participants (186,469 men and 144,300 women) (Figure 1). The NIH-AARP Diet and Health Study was approved by the Special Studies Institutional Review Board of the National Cancer Institute, and all participants gave informed consent by virtue of completing and returning the questionnaire.

Figure 1.

Flowchart of Study Design.

Exposure Assessment of Non-Exercise Estimated Cardiorespiratory Fitness (NEE-CRF)

Cardiorespiratory fitness (expressed as maximal oxygen consumption) was estimated using a previously developed and validated equation by Matthews et al.¹⁰ VO₂ max (mL•kg⁻¹•min⁻¹) = 34.142+1.463 (Physical activity status 0–7) +0.133 (age in years)–0.005 (age²) +11.403 (sex, male=1 and female=0)–0.254 (weight in kilograms) +9.170 (height in meters).¹⁰ This equation is practical in that it is simple and utilizes variables that are easily available by self-report, a common limitation of previous NEE-CRF studies.^{8, 9, 15–18} The equation was developed in a sample of approximately 800 men and women aged 19-79 years, and validated with a direct measurement of VO₂ max using respiratory gas exchange analysis, a gold-standard method for CRF.^{1, 10} This yielded one of the highest explained variations (R²=0.74) of direct VO₂ max measurement compared to other non-exercise estimates of CRF in the literature.^1,10

For estimating CRF in the current study, baseline self-reported age, height, weight and physical activity status were used to calcualte NEE-CRF.^{1, 10} Physical activity status was clasified using the simple question, “did you participate in physical activity ≥ 20 minutes (in the past 12 months) that caused increases in breathing or heart rate, or worked up a sweat”. The report to this question included six categories: 0=“never”, 1=“rarely”, 2=“1-3 times/month”, 3=“1-2 times/week”, 4=“3-4 times/week” and 5=“≥5times/week”.¹⁹ For calculating NEE-CRF, physical activity status from the questionnaires (six categories) was adjusted to match the scale used in the Matthews et al. equation (eight categories)¹⁰. Status= 0 for a report of “never/rarely”, status=1 for a report of “1-3 times/month”, status = 2-3 for a report of “1-2 times/week”, status = 4-5 for “3-4 times/week” and status = 6-7 for a response of “≥5times/week”.^{10, 19} NEE-CRF was expressed as continuous variable in (mLO₂•kg⁻¹•min⁻¹) and metabolic equivalents (METs) (1 MET=3.5 mLO₂•kg⁻¹•min⁻¹) as well as a categorical variable in quartiles of VO₂ max .^{11, 12}

Cohort Follow Up and Mortality Outcomes Ascertainment

Participants were followed from baseline (1995-1996) until the date of death or December 31, 2015, by means of linkage to the National Change of Address database maintained by the U.S. Postal Service, specific change of address requests from participants, and updated addresses returned during other mailings. Vital status was determined through linkage with the Social Security Administration Death Master File²⁰ and determinations of vital status and causes of death were made by using the National Death Index.²¹ The primary outcomes were all-cause mortality, and mortality due to cardiovascular disease and cancer. ICD-9 and ICD-10 codes were used to classify the underlying cause of death obtained from death certificates. Codes of 390-448 for ICD-9 and I00-78 for ICD-10 were utilized for cardiovascular disease related death. Codes of 140-239 for ICD-9 and C00-C97 and D00-D48 for ICD-10 were utilized for cancer related death.

Statistical analysis

Due to established sex differences in CRF,^{11, 22, 23} all analyses were conducted for men and women separately. Baseline participants’ characteristics expressed as mean ± standard deviation or absolute number and percent (%) are presented by NEE-CRF quartiles. Differences across quartiles were tested using one-way analysis of variance for continuous variables and Chi-square tests for categorical variables. Utilizing scaled Schoenfeld residuals and graphical evaluation of the Kaplan-Meier curves, no major violation of the proportional hazards assumption was evident.

Multivariable Cox hazard models and 95% confidence intervals were used to quantify the risk association between NEE-CRF and mortality outcomes both as continuous variables (expressed in mLO₂•kg⁻¹•min⁻¹ and in METs) and categorical variables (expressed in quartiles of VO₂max). Model were adjusted for race ethnicity, marital status, education level, age, body mass index first-degree relatives with history of cancer, diabetes, hypertension, dyslipidemia, smoking status, total alcohol drinks per day, total energy intake, fiber intake, fruit intake, vegetable intake and red meat intake in men, and included also menopausal status in women. Tests for linear trend were performed for NEE-CRF quartiles using Chi-square tests. In addition, stratified analyses for the association between NEE-CRF and mortality outcomes were performed for age categories (50-60 and 61-71 years), body mass index categories (18.5-24.9, 25-29.9 and ≥30) smoking status (never, former and current) and physical activity status (never or rarely and 1-3 times per month, 1-2 times per week, 3-4 times per week and 5 or more times per week). Data report and presentation followed the “Strengthening the Reporting of Observational Studies in Epidemiology” (STROBE) guidelines.²⁴ Little’s MCAR test was conducted for potential pattern in the few missing data. The test showed that missingness was complete at random and analyzed as complete case method.²⁵ Data management and statistical analyses were conducted between October 2019 and March 2020 using IBM SPSS Statistics Software version 23.0 (IBM, Armonk, NY, USA). The significance level was set at p<0.05.

Results

The cohort included 330,769 participants [men (n=186,469) and women (n=144,300)] with a mean age of 61.1±5.4 years at recruitment (1995-1996). Mean NEE-CRF was 34.7±5.1 mLO₂•kg⁻¹•min⁻¹ (9.9±1.5 METs) for men and 25.3±5.6 mLO₂•kg⁻¹•min⁻¹ (7.2±1.6 METs) for women. Table 1 presents baseline participants’ demographics and characteristics by NEE-CRF quartiles. In both men and women, higher NEE-CRF quartile was associated with a higher prevalence of college education, lower body mass index, lower prevalence of diabetes and higher physical activity level. During 14.9±2.1 years follow up, 34,317 men died from all-causes (5,262 deaths from cardiovascular disease, 12,629 deaths from cancer) (Table 2), and 20,295 women died from all-causes (3,371 deaths from cardiovascular disease, 7,548 deaths from cancer) (Table 3).

Table 1.

Baseline Characteristics in the NIH-AARP Diet and Health study by VO₂ max Quartiles of Non-Exercise Estimated Cardiorespiratory Fitness (N=330,769).

	Men					Women
	All Men (n=186,469)	VO2 max Q-1 (<31.5 mL•kg−1•min−1) (n=46,934)	VO2 max Q-2 (31.5–35 mL•kg−1•min−1) (n=47,133)	VO2 max Q-3 (>35–38.2 mL•kg−1•min−1) (n=45,537)	VO2 max Q-4 (>38.2 mL•kg−1•min−1) (n=46,865)	All Women (n=144,300)	VO2 max Q-1 (<21.8 mL•kg−1•min−1) (n=36,292)	VO2 max Q-2 (21.8–25.7 mL•kg−1•min−1) (n=36,384)	VO2 max Q-3 (>25.7–29.2 mL•kg−1•min−1) (n=35,314)	VO2 max Q-4 (>29.2 mL/kg/min) (n=36,310)
Age	61.1±5.4	62.9±5	61.5±5.3	61±5.3	58.8±5.1	61±5.4	63.5±4.9	62.2±5	60.6±5.1	57.7±4.7
Race/Ethnicity, %
Non-Hispanic White	92.9	93.8	93.3	92.6	91.9	90.5	89	90.3	90.9	92
African American	2.7	2.9	2.7	2.6	2.4	5.1	7.3	5.3	4.5	3.2
Other	4.4	3.3	4	4.8	5.7	4.4	3.7	4.4	4.6	4.8
Education
<High school/high school graduate	18.5	25	20.3	17.4	15.3	18.5	25	20.3	17.4	15.3
Post high school/some college	29.2	24.3	31.9	29	29.2	29.2	24.3	31.9	29	29.2
College or graduate degree	47.2	31.5	45.7	49.6	54.5	47.2	31.5	45.7	49.6	54.5
Married or living as married (%)	85.3	84.8	86.6	86.1	83.7	85.3	84.8	86.6	86.1	83.7
Body Mass Index (kg/m2)	27.1±4.1	30.9±4.6	27.4±3	26±2.7	24.3±2.7	27.1±4.1	30.9±4.6	27.4±3	26±2.7	24.3±2.7
Diabetes (%)	7.7	12.7	7.8	6	4.4	7.7	12.7	7.8	6	4.4
Energy Intake (kcal/day)	2067.6±1014	2097.3±1072	2030.3±998	2041.8±963	2100.5±1016	1589±759	1639.7±822	1574±744	1562.4±706	1579.6±756
Fibers (g/day)	20.3±10.7	19.3±10.3	19.6±10.1	20.4±10.5	22±11.6	17.9±9.9	17.3±9.7	17.4±9.3	17.9±9.4	19.1±10.8
Vegetables (cups/day)	2±1.3	2±1.3	2±1.3	2±1.3	2.1±1.4	1.9±1.3	1.9±1.3	1.9±1.3	1.9±1.3	2.1±1.5
Fruits (cups/day)	2.1±1.7	1.9±1.7	2±1.7	2.1±1.7	2.3±1.9	2 ±1.7	1.9±1.6	1.9±1.6	2±1.6	2.2±1.8
Red Meat (g/day)	81.5±74	91.8±80	82±75	78±67.4	74±72	81.5±74	91.8±80	82±75	78±67.4	74±72
Alcohol (drinks/day)	1.4±3.4	1.4±3.8	1.4±3.4	1.3±3.1	1.3±3.3	0.4±1.3	0.3±1.2	0.4±1.2	0.5±1.3	0.5±1.3
Smoking Status (never/former/current, %)	31.8/55/10.3	27.2/60.3/9.4	30.3/56.3/10.4	33.2/53.4/10.6	36.7/49.8/10.8	45.4/38.3/13.8	46.6/39.7/11/1	45.5/37.4/14.8	44.8/37.6/15.1	44.7/38.6/14.1
Physical Activity Status
Never or Rarely (%)	13.7	34.8	13.1	5.5	1	21.1	50.1	22.9	9.7	1.6
1-3 times per month or 1-2 times per week (%)	26.6	49.6	48.1	39.3	13.2	35.4	40.1	48.4	38.9	18.4
3-4 times per week (%)	28.5	13.4	30.5	32.5	32.7	25.8	8.6	23.8	35.7	35.5
5 or more times per week (%)	21.2	2.3	22.3	18.4	52.1	16.6	1.2	4.9	15.6	44.6
Postmenopausal (%)						93.5	96.5	95.3	93.5	88.8

^*All comparisons between VO₂max quartiles both in men and women are significantly different (p<0.001).

Q-quartile.

Table 2.

Hazard Models of Non-Exercise Estimated Cardiorespiratory Fitness in the NIH-AARP Diet and Health Study of Men Participants (n=186,469).

Group/CRF Categories	Total	VO2 max Q-1 (<31.5 mL•kg−1•min−1) (n=46,934)	VO2 max Q-2 (31.5-35 mL•kg−1•min−1) (n=47,133)	VO2 max Q-3 (>35-38.2 mL•kg−1•min−1) (n=45,537)	VO2 max Q-4 (>38.2 mL•kg−1•min−1) (n=46,865)	P trend	1-mL•kg−1•min−1 increase	P value
All-cause mortality (n/%)	34,317 (18.4)	12,976 (27.6)	9,250 (19.6)	6,965 (15.3)	5,126 (10.9)	<0.001	<0.001	<0.001
HR 95% (CI)		1 (Reference)	0.82 (0.79-0.84)	0.74 (0.72-0.77)	0.7 (0.67-0.73)	<0.001	0.96 (0.95-0.96)
Cardiovascular disease mortality (n/%)	5,262 (2.8)	2,283 (4.9)	1,397 (3)	952 (2.1)	630 (1.3)	<0.001	<0.001	<0.001
HR 95% (CI)		1 (Reference)	0.81 (0.75-0.87)	0.72 (0.66-0.79)	0.69 (0.61-0.77)	<0.001	0.95 (0.94-0.96)
Cancer mortality (n/%)	12,629 (6.8)	4,281 (9.1)	3,431 (7.3)	2,779 (6.1)	2,138 (4.6)	<0.001	<0.001	<0.001
HR 95% (CI)		1 (Reference)	0.84 (0.8-0.89)	0.79 (0.74-0.83)	0.71 (0.66-0.76)	<0.001	0.97 (0.96-0.97)

CI; confidence intervals, HR; hazard ratio, MET; metabolic equivalent. N/A; not applicable. Q-quartile. Risk models were adjusted for race ethnicity, marital status, education level, age, body mass index, first-degree relatives with history of cancer, diabetes, hypertension, dyslipidemia, smoking status, total alcohol drinks per day, total energy intake, fibers intake, fruits intake, vegetables intake and red meat intake.

Table 3.

Hazard Models of Non-Exercise Estimated Cardiorespiratory Fitness in the NIH-AARP Diet and Health Study of Women Participants (n=144,300).

Group/CRF Categories	Total	VO2 max Q-1 (<21.8 mL•kg−1•min−1) (n=36,292)	VO2 max Q-2 (21.8-25.7 mL•kg−1•min−1) (n=36,384)	VO2 max Q-3 (>25.7-29.2 mL•kg−1•min−1) (n=35,314)	VO2 max Q-4 (>29.2 mL•kg−1•min−1) (n=36,310)	P trend	1- mL•kg−1•min−1 increase	P value
All-cause mortality (n/%)	20,295 (14.1)	7,676 (21.2)	5,479 (15.1)	4,182 (11.8)	2,958 (8.1)	<0.001	<0.001	<0.001
HR 95% (CI)		1 (Reference)	0.84 (0.81-0.88)	0.78 (0.75-0.82)	0.72 (0.68-0.77)	<0.001	0.95 (0.94-0.96)
Cardiovascular disease mortality (n/%)	3,371 (2.3)	1,503 (4.1)	919 (2.5)	606 (1.7)	341 (0.9)	<0.001	<0.001	<0.001
HR 95% (CI)		1 (Reference)	0.84 (0.77-0.93)	0.75 (0.66-0.84)	0.63 (0.54-0.73)	<0.001	0.95 (0.94-0.96)
Cancer mortality (n/%)	7,548 (5.2)	2,411 (6.6)	2,047 (5.6)	1,698 (4.8)	1,392 (3.8)	<0.001	<0.001	<0.001
HR 95% (CI)		1 (Reference)	0.88 (0.82-0.94)	0.82 (0.76-0.89)	0.77 (0.7-0.85)	<0.001	0.97 (0.96-0.98)

CI; confidence intervals, HR; hazard ratio, MET; metabolic equivalent. N/A; not applicable. Q-quartile. Risk models were adjusted for race ethnicity, marital status, education level, age, body mass index, first-degree relatives with history of cancer, diabetes, hypertension, dyslipidemia, smoking status, total alcohol drinks per day, total energy intake, fibers intake, fruits intake, vegetables intake, red meat intake and menopausal status.

Among men, higher NEE-CRF was associated with lower risk of mortality due to all-causes, cardiovascular disease and cancer. In categorical models in which the first quartile was the reference, NEE-CRF in the second, third and fourth quartiles was associated with 18%, 26%, and 30% reduced risks for all-cause mortality, 19%, 28%, and 31% reduced risks for cardiovascular disease mortality and 16%, 21% and 29% reduced risks for cancer mortality, respectively (p trend<0.001 for all) (Table 2). In continuous models, for every 1-MET higher NEE-CRF there were 15%, 15% and 11% reductions in risk for all-cause, cardiovascular disease and cancer mortality, respectively. The corresponding hazard ratios (HR) and 95% confidence intervals (CI) were: 0.85 (0.84-0.86), 0.85 (0.82-0.88), 0.89 (0.87-0.91), respectively (all p<0.001) (Figure 2). In women, compared to the first quartile of NEE-CRF, the second, third and fourth quartiles of NEE-CRF were associated with 16%, 22%, and 28% reduced risks for all-cause mortality, 16%, 25%, and 37% reduced risks for cardiovascular disease mortality and 12%, 18% and 23% reduced risks for cancer mortality, respectively (p trend<0.001 for all) (Table 3). In continuous models, for every 1-MET higher NEE-CRF there were 16%, 16% and 11% reductions in risk for all-cause, cardiovascular disease and cancer mortality, respectively. The corresponding hazard ratios (HR) and 95% confidence intervals (CI) were: 0.84 (0.83-0.85), 0.84 (0.81-0.88), 0.89 (0.87-0.91), respectively (all p<0.001) (Figure 2). Trend analyses showed significantly lower incidence of all-cause, cardiovascular and cancer mortality in both sexes (Tables 2 and 3).

Figure 2.

Hazard Ratios and 95% Confidence Intervals of NEE-CRF per 1-MET Increase.

CVD; cardiovascular disease, MET; metabolic equivalent, NEE-CRF; non-exercise estimated cardiorespiratory fitness.

In stratified analyses by age groups, body mass index category, smoking status and physical activity levels, NEE-CRF was associated with reduced risk of mortality due to all-causes, cardiovascular disease and cancer (Table 4). In men, for each 1-MET increase in NEE-CRF the risk reduction ranged from 6% for cancer mortality in current smokers to 35% for cardiovascular disease mortality in individuals who were physically active 3-4 times a week (Table 4). In women, for each 1-MET increase in NEE-CRF, the risk reduction ranged from 5% for cancer mortality in the age group 50-60 years to 38% for cardiovascular disease mortality in individuals with normal body mass index (Table 4).

Table 4.

Stratified Continuous Models of Non-Exercise Estimated Cardiorespiratory Fitness (METs) in the NIH-AARP Diet and Health Study of Men and Women (N=330,769).

	Age (years)		Body Mass Index (kg/m2)			Smoking Status			Physical Activity
	50-60	61-71	18.5-24.9	25-29.9	≥30	Never	Former	Current	Never or rarely and 1-3 times per month	1-2 times per week	3-4 times per week	5 or more times per week
HR 95% (CI) All-cause mortality-men	0.85 (0.83-0.87)	0.85 (0.84-0.86)	0.85 (0.83-0.87)	0.87 (0.85-0.88)	0.85 (0.83-0.87)	0.89 (0.86-0.91)	0.86 (0.85-0.88)	0.89 (0.87-0.92)	0.81 (0.76-0.86)	0.88 (0.84-0.92)	0.84 (0.79-0.9)	0.83 (0.77-0.9)
HR 95% (CI) All-cause mortality-women	0.85 (0.82-0.88)	0.84 (0.82-0.85)	0.83 (0.81-0.86)	0.83 (0.8-0.86)	0.87 (0.84-0.9)	0.9 (0.87-0.92)	0.85 (0.83-0.88)	0.86 (0.83-0.89)	0.75 (0.7-0.8)	0.89 (0.83-0.95)	0.82 (0.73-0.92)	0.83 (0.71-0.96)
HR 95% (CI) CVD mortality-men	0.8 (0.75-0.86)	0.86 (0.83-0.89)	0.85 (0.8-0.91)	0.88 (0.84-0.93)	0.82 (0.78-0.87)	0.83 (0.77-0.89)	0.85 (0.82-0.89)	*0.96 (0.89-1.04)	0.68 (0.59-0.79)	0.82 (0.73-0.92)	0.76 (0.64-0.9)	0.8 (0.65-0.97)
HR 95% (CI) CVD mortality-women	0.81 (0.74-0.9)	0.86 (0.81-0.89)	0.84 (0.78-0.9)	0.83 (0.78-0.9)	0.88 (0.82-0.94)	0.91 (0.85-0.97)	0.84 (0.79-0.9)	0.85 (0.78-0.93)	0.8 (0.68-0.95)	0.85 (0.72-0.99)	*0.84 (0.63-1.1)	*0.87 (0.62-1.2)
HR 95% (CI) Cancer mortality-men	0.89 (0.86-0.93)	0.89 (0.87-0.91)	0.9 (0.87-0.94)	0.92 (0.89-0.94)	0.86 (0.83-0.9)	0.92 (0.87-0.96)	0.91 (0.88-0.94)	0.94 (0.9-0.99)	0.8 (0.72-0.9)	0.86 (0.8-0.93)	0.8 (0.72-0.9)	0.85 (0.75-0.97)
HR 95% (CI) Cancer mortality-women	0.91 (0.87-0.96)	0.89 (0.85-0.91)	0.9 (0.86-0.94)	0.87 (0.82-0.91)	0.9 (0.86-0.96)	0.94 (0.89-0.99)	0.92 (0.88-0.97)	0.91 (0.86-0.96)	0.76 (0.67-0.86)	0.87 (0.77-0.97)	0.72 (0.59-0.86)	0.78 (0.61-0.99)

CVD; cardiovascular disease, CI; confidence intervals, HR; hazard ratio expressed per 1-MET (metabolic equivalent) increase. All risk models were statistically significant (p<0.05) except models with *, where statistical significance was not reached. Risk models were adjusted for race ethnicity, marital status, education level, age, body mass index, first-degree relatives with history of cancer, diabetes, hypertension, dyslipidemia, smoking status, total alcohol drinks per day, total energy intake, fibers intake, fruits intake, vegetables intake, red meat intake and menopausal status for women only.

Discussion

The current study aimed to assess the association between a comparatively pragmatic method of NEE-CRF and mortality outcomes in a large prospective cohort of men and women. The results showed that higher NEE-CRF was independently associated with lower risks of mortality due to all-causes, cardiovascular disease and cancer in men and women. Higher quartiles of NEE-CRF were associated with a significantly lower incidence of death in both sexes. For every 1-MET increase in NEE-CRF the reductions in risk for mortality ranged from 11% to 15% in men and from 11% to 16% in women (Figure 2). The stratified analyses by age groups, body mass index categories, smoking status and physical activity levels further supported the inverse association between NEE-CRF and mortality outcomes in men and women (Table 4). These novel findings suggest a strong prognostic value of the utilized NEE-CRF approach, suggesting it may have broader applications for health care and research settings. The results support the application of this method for general health assessment and clinical risk stratification particularly in low-resource settings. Health-care professionals can easily estimate CRF using this method, and refer low fit individuals to prevention and rehabilitation programs. The data also provide an opportunity to estimate CRF in large, questionnaire-based epidemiological studies.

The results of the current study are consistent with previous observations in terms of participants’ characteristics, estimated CRF levels and impact on mortality risk among individuals with higher NEE-CRF.^{8, 9, 15–18} For instance, Stamatakis et al.⁹ reported comparable NEE-CRF quartiles (<9.8 METs in the lowest quartile to>12 METs in the highest quartile for men and <6.8 METs in the lowest quartile to >9.3 METs in the highest quartile for women) with similar risk reductions among 32,319 adults from the United Kindom.⁹ The authors found that for every 1 standard deviation increase in NEE-CRF (corresponding to 1.6-1.7 METs) there were 15% and 12% reductions in all-cause mortality, and 25% and 27% reductions in cardiovascular mortality among men and women, respectively.⁹ Similarly, data from the Aerobics Center Longitudinal Study (ACLS) (n=43,356) demonstrated 15%, and 19% reductions in risk among men, and 13% and 16% reductions in risk among women for all-cause and cardiovascular disease mortality per 1-MET increase in NEE-CRF, respectively.⁸ A smaller (n=8,506) but representive sample of US population derived from the Third National Health and Nutrition Examination Survey (NHANES III) observed comparable mean NEE-CRF values of 11.8±0.1 among men and 8.9±0.1 among women.¹⁵ Additionaly, reductions of 30% and 28% in risk for cancer motality per each 1-MET increase in NEE-CRF among men and women, respectively.¹⁵ Our findings also align with a meta-analysis of 33 studies (102,980 participants) demonstrating a 13% reduction in risk for all-cause mortality with every 1-MET increase in measured CRF using an exercise testing.^{4, 26}

The findings from the current study extend these earlier reports and add several important and novel insights that have significant public health implications for health screening, prevention and rehabilitation programs. To our knowledge, the current cohort is the largest study of men and women worldwide showing the risk association between a simple and practical method of NEE-CRF and mortality outcomes. Although previous studies have demonstrated associations between the NEE-CRF and mortality outcomes, these are limited because relatively small cohorts and utilization of impractical variables which challenging a broad application.^{8, 9, 15–18} The current study improves upon previous efforts by using a prospective evaluation of a significantly larger cohort (n=330,769) and utilization of a more pragmatic NEE-CRF equation that does not require patient contact or physical measurements. The utilized equation in the current study was developed and validated with direct measurements of VO₂ max, yielding a high level of explained variance (R²=0.74).^{1, 10} These findings provide further evidence for the clinical and prognostic value of CRF, and concur with the recent call for “fitness as a vital sign” to be assessed along with traditional risk factors.¹

Although the exact mechanisms involved in which higher NEE-CRF might mediate death is not fully understood, there are several potential explanations. CRF is a well-established health marker associated with many chronic conditions and mortality outcomes.^{1, 2, 4} Higher CRF is associated with lower blood lipids, enhanced glucose utilization, improved cardiovascular, pulmonary and muscle function, lower body fat%, reduced chronic inflammation, balanced hormonal regulation, improved immune system activity and healthier cardiometabolic risk factor profile.^{1, 11, 27} People with higher CRF exhibit reduced incidence of many chronic conditions such as cardiovascular disease, stroke, heart failure, type 2 diabetes, metabolic syndrome, Alzheimer’s disease, dementia, cancer, disability and perioperative complications.¹ Given that the applied NEE-CRF method used in the current study was validated with direct measure of VO₂ max, the prediction equation provides a good surrogate for the gold-standard metric of CRF, which is strongly associated with health outcomes.^{1, 10}

The strengths of the study include a large sample size with over 330,000 men and women, prospective evaluation of hard-endpoints (mortality), lengthy follow up time (14.9±2.1 years), analysis of both all-cause and specific causes of mortality, and adjusting the models for many potential confounders to extract the independent association between NEE-CRF and mortality. Additional strengths include the exclusion of individuals with less than 5 years follow up to address the potential for reverse casualty, stratified analyses for several sub-groups, and ascertainment of mortality outcomes through Social Security Administration Death Master File and the National Death Index. ^{20, 21}

The study has also some limitations. First, although the NIH-AARP Diet and Health Study is a large cohort with a relatively diverse US population¹⁹ and the cohort’s demographics are similar to previous studies in this area,^{8, 9, 15–18} most participants (>90%) were from Non-Hispanic white ancestries; thus, future studies should focus on more diverse populations. Second, consistent with previous studies using the NEE-CRF method,^{8, 9, 15–18} physical activity status was self-reported, raising the potential for over or under estimation and recall bias. Nonetheless, self-reported physical activity has been applied in all previous NEE-CRF studies,^{8, 9, 15–18} it is an established method to assess physical activity, having strong predictive value for numerous health outcomes. ^28–31 Third, although consistent with previous reports of NEE-CRF,^{8, 9, 15–18} and the majority of epidemiological studies in this field, participants were assessed at baseline only, the potential changes in NEE-CRF over time were not accounted.^{1–4, 30–32} Finally, as is the case in all epidemiological studies the findings are observational and do not suggest cause and effect relationship.

In summary, findings from this large prospective study demonstrate that the applied pragmatic method for assessing NEE-CRF is associated with mortality outcomes in men and women, suggesting its prognostic value for clinical and research settings. Higher NEE-CRF was independently associated with lower risk of death from all-causes, cardiovascular disease and cancer. The results support the potential of this approach for clinical application in health screening, risk stratification, prevention and rehabilitation programs and utilization in large questionnaire-based epidemiological studies.