Non-exercise estimated cardiorespiratory fitness and cancer incidence: The NIH-AARP diet and health study

Highlights

• •Higher non-exercise estimated cardiorespiratory fitness (NEE-CRF) was associated with lower risk of total cancer incidence in men and women.
• •Higher NEE-CRF was associated with lower risk of colorectal cancer incidence in men and lower risk of breast cancer incidence in women.
• •These novel findings suggest that higher fitness levels may provide preventive benefits against the development of cancer, while low fitness could potentially serve as a modifiable cancer risk factor.

Abstract

Background

Non-exercise estimated cardiorespiratory fitness (NEE-CRF) has been shown to be associated with mortality, although its association with cancer incidence is unknown. The study aimed to assess the prospective association between NEE-CRF and cancer incidence in a large cohort of men and women.

Methods

The National Institutes of Health-American Association of Retired Persons diet and health study is a prospective cohort that included 402,548 participants aged 50–71 years who were free from cancer at baseline (1995–1996) (men (n = 238,835) and women (n = 163,713)) and were followed until December 31, 2015. The exposure variable was NEE-CRF expressed in metabolic equivalents. NEE-CRF was estimated using a validated equation of self-reported predictors on demographics and lifestyle behaviors derived from baseline questionnaires. Primary outcomes were total cancer incidence and incidence of prostate, breast, lung, and colorectal cancers. Cox proportional hazards models were analyzed for the association between NEE-CRF and cancer incidence outcomes adjusted for established cancer risk factors.

Results

During 13.7 ± 3.2 years of follow-up (mean ± SD), 64,344 men and 31,315 women developed a new cancer. For every 1-metabolic equivalent higher NEE-CRF, the hazard ratios and 95% confidence intervals (95%CIs) were 0.96 (95%CI: 0.94–0.97) and 0.88 (95%CI: 0.84–0.92) of total and colorectal cancer incidence among men, and 0.95 (95%CI: 0.93–0.97) and 0.94 (95%CI: 0.91–0.97) of total and breast cancer incidence among women, respectively (all p < 0.001). NEE-CRF was not associated with incidence of prostate and lung cancers in men or colorectal and lung cancers in women.

Conclusion

These results suggest that higher CRF levels, as assessed by the applied non-exercise estimated method, may provide preventive benefits against the development of cancer, while low CRF could potentially serve as a modifiable cancer risk factor. Integrating NEE-CRF into screening paradigms and referring low-fit individuals to improve CRF could complement the public health prevention strategy against cancer.

1. Introduction

Cancer remains one of the leading causes of morbidity and mortality worldwide and in the United States.^1,2 Approximately 19 million new cancer cases are diagnosed annually worldwide and 1.9 million in the United States alone.^3,4 The American Cancer Society estimates a lifetime probability for developing any type of cancer is 40.5% for men and 38.9% for women.⁵ Incident cancer cases are projected to increase by approximately 70% worldwide and by 45%–50% in the United States over the next few decades, which will have substantial socioeconomic impact in addition to the human toll.6, 7, 8 Globally, the economic burden of cancer is estimated to be $895 billion, which represents 1.5% of the world's gross domestic product.⁹ In 2015, direct medical costs for cancer in the United States were estimated to be $80.2 billion, while healthcare expenditures in Europe in 2014 reached $110 billion.¹⁰

Despite the fact that the burden of cancer continues to rise, a substantial number of cancers are related to modifiable lifestyle risk factors and, thus, are potentially preventable.10, 11, 12, 13 These include cigarette smoking, poor diet (low consumption of vegetables, fruits, and dietary fibers; high consumption of sugary drinks and processed foods), obesity, and physical inactivity, all of which are well-established risk factors for other chronic diseases as well.10, 11, 12, 13 While physical inactivity is a well-known risk factor for cancer,^10,14, 15, 16 accumulating evidence also suggests that low cardiorespiratory fitness (CRF) is associated with higher incidence and mortality from cancer.17, 18, 19, 20 CRF reflects the integrated physiological capacity of the heart, lungs, and skeletal muscle to supply the required energy during maximal aerobic exercise, commonly termed maximal oxygen uptake (VO_2max).¹⁸ Although CRF is a powerful health marker that is strongly associated with mortality outcomes,^18,20 it is not routinely measured in clinical, health, or research settings.¹⁸ This is in part due to the fact that the direct measurement of CRF with standardized exercise and metabolic testing requires resources that are unavailable in many centers.^18,21

Recently, the American Heart Association has recognized non-exercise estimated CRF (NEE-CRF) methods as alternative approaches to standard exercise testing when the latter is unavailable or not feasible.¹⁸ The NEE-CRF method uses validated equations from self-reported variables such as age, sex, weight, height, and physical activity to estimate directly measured CRF from an exercise test.¹⁸ Previous relatively small epidemiological studies (8506–43,356 participants) utilizing various prediction equations have demonstrated inverse associations between NEE-CRF and mortality outcomes, including cancer mortality.22, 23, 24, 25, 26 We recently studied a large prospective cohort of men and women showing that higher NEE-CRF was strongly and independently associated with lower risk of all-cause, cardiovascular, and cancer-related mortality in both men and women.²⁷ To our knowledge, the risk association between NEE-CRF and incidence of cancer has not been previously explored. Given the simplicity and strong predictive value of the method we utilized, along with the worldwide epidemic of cancer and the recommendations for estimating CRF as a part of a general health examination,¹⁸ the current study provides additional important information for cancer screening and prevention practice. To test this hypothesis, the present study aimed to assess the association between a method of NEE-CRF and total cancer incidence in a large prospective cohort of men and women. Additionally, the study sought to assess the prospective association between NEE-CRF and incidence of some of the most common cancers in the United States, which account for 46% and 50% of all cancer incidence in men (prostate, lung, and colorectal) and women (breast, lung, and colorectal), respectively.³

2. Methods

2.1. Study population and design

The National Institutes of Health-American Association of Retired Persons (NIH-AARP) Diet and Health Study (www.clinicaltrials.gov; NCT00340015) has been previously described.²⁸ In brief, between 1995 and 1996 questionnaires asking about demographics, medical history, dietary and lifestyle behaviors were mailed to 3.5 million AARP members aged 50–71 years. In total, 617,119 questionnaires were returned and 566,398 were satisfactorily completed (92%).²⁸ Excluded from the present study were participants whose questionnaires were completed by a spouse or other surrogate (n = 15,760), participants with baseline self-reported diagnosis of any cancer (n = 50,591), and those with missing, out of normal range, or incomplete information for estimating CRF (n = 97,499). The resulting analytic cohort included 402,548 participants (238,835 men and 163,713 women) (Fig. 1). The study was approved by the Special Studies Institutional Review Board of the U.S. National Cancer Institute, and all participants gave informed consent by virtue of completing and returning the questionnaire.

Fig. 1.

Flowchart of the study design. CRF = cardiorespiratory fitness.

2.2. Exposure assessment of NEE-CRF

NEE-CRF was estimated using a previously developed and validated equation:²⁹

VO_2max (mL/kg/min) = 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).²⁹

NEE-CRF was expressed as a continuous variable in metabolic equivalents (METs) (1 MET ≈ 3.5 mL/kg/min O₂) as well as in a categorical variable conveying low (<25th percentile), moderate (25th–75th percentiles), and high (>75th percentile) NEE-CRF levels, which is consistent with previous reports.^19,21,30,31 The corresponding MET levels were <8.9, 8.9–10.9, and >10.9 for men, and <6.1, 6.1–8.2, and >8.2 for women, respectively.

This equation utilizes variables that are easily available by self-report, a common limitation of previous NEE-CRF stu?>dies.22, 23, 24, 25, 26 The equation was developed in a sample of approximately 800 men and women aged 19–79 years and validated with a direct measurement of VO_2max using respiratory gas exchange analysis, a gold-standard method for CRF.^18,29 This yielded one of the highest explained variations (R² = 0.74, standard error of estimate = 5.64 mL/kg/min O₂) of direct VO_2max measurement,^18,29 and the utility of this method has been confirmed for its association with mortality outcomes.²⁷

Baseline self-reported age, height, weight, and physical activity status were used to calculate NEE-CRF.^18,29 Physical activity status was classified using the question, “Did you participate in physical activity ≥20 min (in the past 12 months) that caused increases in breathing or heart rate, or worked up a sweat”. Responses to this question fell into 6 categories: 0 = “never”, 1 = “rarely”, 2 = “1–3 times/month”, 3 = “1–2 times/week”, 4 = “3–4 times/week”, and 5 = “≥5 times/week”.²⁸ For calculating NEE-CRF, physical activity status from the questionnaires (6 categories) was adjusted to match the scale used in the prediction Matthews equation (8 categories).²⁹ 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 “≥5 times/week”.^28,29

2.3. Follow-up and cancer outcomes ascertainment

Participants were followed from baseline (1995–1996) until the date of first primary cancer diagnosis, death, or end of follow-up (December 31, 2015). Cancer incidence (the first diagnosed cancer other than non-melanoma skin cancer) and type of cancer were identified by probabilistic linkage to cancer registries (Surveillance, Epidemiology, and End Results) of the 6 baseline recruitment states (California, Florida, Louisiana, New Jersey, North Carolina, and Pennsylvania) and 2 metropolitan areas (Atlanta, GA, and Detroit, MI). Additionally, 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 were used for tracking. Vital status was determined through linkage to the Social Security Administration Death Master File³² and the National Death Index.³³ The primary outcomes were total cancer incidence (any diagnosed cancer excluding basal and squamous skin cancers) and incidence of prostate (C619), breast (C500–C509), lung (C340–C349), and colorectal (C180–C189, C260, C199, and C209) cancers. International Classification of Diseases for Oncology (ICD-O), Third Edition, codes were used for cancer incidence outcomes.³⁴

2.4. Statistical analysis

Data management and statistical analyses were conducted during November 2021 using IBM SPSS Statistics Software Version 23.0 (IBM Corp., Armonk, NY, USA). The significance level was set at p < 0.05. Data report and presentation followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.³⁵ Due to established sex differences in CRF,^18,21 analyses were conducted for men and women separately. Baseline participants’ characteristics are expressed as mean ± SD or percentage (%) and are presented by NEE-CRF categories. Differences across categories were tested using one-way analysis of variance for continuous variables and χ² 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. Tests for linear trend were performed for NEE-CRF categories using χ² tests. Multivariable Cox hazard models were used to quantify the risk association between NEE-CRF and cancer incidence outcomes. The risk models took into account competing events (death occurred before cancer diagnosis)^36,37 and were analyzed both as continuous (expressed in METs) and categorical variables (expressed as low, moderate, and high NEE-CRF levels).

2.4.1. Covariates

Covariates were selected based on previous reports showing an association with outcomes of interst.^{15,19,30,31,38} All risk models were adjusted for race, ethnicity, marital status, education level, age, obesity (body mass index of ≥30 kg/m²), first-degree relatives with history of cancer, diabetes, hypertension, dyslipidemia, smoking status, total alcoholic drinks per day, total energy intake, fiber intake, fruit intake, vegetable intake, red meat intake, and meeting the physical activity guidelines¹⁴ as determined by physical activity status 4–5 times/week and 6–7 times/week (which correspond to questionnaire responses of “3–4 times/week” and “≥5 times/week”, respectively^28,29). For women only, additional adjustments were made for hormonal and reproductive factors, including menopausal status and age of menopause start, years of oral contraceptive use, hormonal replacement therapy status and years of use. Additional analysis was conducted excluding age, physical activity, and obesity (body mass index of ≥30 kg/m²) variables from the risk models.

2.4.2. Sensitivity analysis

Sensitivity analyses were performed for potential biases and confoundings of the observed association between NEE-CRF and incident cancer outcomes. E-value was calculated to measure an association's robustness for potential uncontrolled confounders. The E-value is a continuous variable where 1 is the lowest possible value, indicating no unmeasured confounding is needed to explain away the observed association between exposure and outcome. On the other hand, the higher the E-value, the stronger the confounder associations must be to explain away this relationship between exposure and outcome.³⁹ Stratified analyses by age categories, obesity status, smoking status, and physical activity status were conducted. In order to address potential reverse causality, ana?>lysis excluding participants with fewer than 5 years of follow-up was performed. For selection bias, Little's missing completely at random test was conducted to assess potential patterns in the missing data. The test showed that missingness was random and analyzed as complete case method.⁴⁰

3. Results

The cohort included 402,548 participants (men (n = 238,835; 61.6 ± 5.4 years) and women (n = 163,713; 62.1 ± 5.4 years)) at baseline. Mean NEE-CRF was 9.8 ± 1.5 METs for men and 7.2 ± 1.6 METs for women. Table 1 presents baseline participant demographic characteristics by NEE-CRF categories. During 13.7 ± 3.2 years of follow-up, 64,344 men and 31,315 women developed any type of cancer (Table 2). In continuous models, for every 1-MET higher NEE-CRF, there was a 4% (hazard ratio (HR) = 0.96, 95% confidence interval (95% CI): 0.94–0.97) and 12% (HR = 0.88, 95%CI: 0.84–0.92) lower risk of total and colorectal cancer incidence in men, and a 5% (HR = 0.95, 95%CI: 0.93–0.97) and 6% (HR = 0.94, 95%CI: 0.91–0.97) lower risk of total and breast cancer incidence in women, respectively (all p < 0.001) (Fig. 2). In categorical models, compared to the lowest NEE-CRF category, moderate and high categories were respectively associated with a 6% and 9%, and a 12% and 30% lower risk of total and colorectal cancer incidence in men. In women, the risks for total and breast cancer incidence were 10% and 11%, and 10% and 11% lower, respectively (Table 2).

Table 1.

Baseline characteristics in the NIH-AARP diet and health study by NEE-CRF categories (n = 402,548).

	Men				Women
	All men (n = 238,835)	Low NEE-CRF <25th percentile <8.9 METs (n = 59,617)	Moderate NEE-CRF 25th–75th percentiles 8.9–10.9 METs (n = 127,500)	High NEE-CRF >75th percentile >10.9 METs (n = 51,718)	All women (n = 163,713)	Low NEE-CRF <25th percentile <6.1 METs (n = 39,953)	Moderate NEE-CRF 25th–75th percentiles 6.1–8.2 METs (n = 83,157)	High NEE-CRF >75th percentile >8.2 METs (n = 40,603)
Age (year)	61.6 ± 5.4	64.0 ± 5.4	61.9 ± 5.1	58.1 ± 4.8	62.1 ± 5.4	64.5 ± 4.8	62.6 ± 5.1	58.7 ± 4.9
Non-Hispanic White (%)	93.1	94	93.2	91.9	90.4	88.8	90.5	91.9
College or graduate degree (%)	45.6	37.8	46.1	53.4	30.4	22.9	29.8	46.3
Married or living as married (%)	85.2	84.4	86.1	83.1	44.6	37.6	45.0	50.7
BMI (kg/m2)	27.2 ± 4.2	31.1 ± 4.7	26.7 ± 3.0	24.2 ± 2.7	26.8 ± 5.7	32.8 ± 6.2	26 ± 3.8	22.7 ± 3.1
Obesity (%)	21.0	55.2	13.0	1.4	23.6	64.0	14.4	1.4
Smoking status (never/former/current, %)	29.7/57.1/10.3	24.5/62.6/9.7	29.8/56.8/10.6	35.3/51.4/10.6	44.4/39/14.1	44.9/41.2/11.3	44.2/38.1/15.3	44.3/38.7/14.5
Meeting the physical activity guidelines (%)	49.9	15.0	51.2	86.9	41.7	9.1	38.9	79.7
Dyslipidemia (%)	49.2	48.2	48.1	52.9	50.9	44	48.9	60.9
Hypertension (%)	45.6	58.9	44.5	33.2	39.3	57.9	38.1	24.1
Diabetes (%)	9.9	16.5	8.8	5.3	7.2	15.3	5.6	2.4
Energy intake (kcal/day)	2048 ± 1009	2078 ± 1074	2021 ± 979	2082 ± 1003	1594 ± 790	1650 ± 859	1571 ± 747	1586 ± 802
Vegetables (cups/day)	2.0 ± 1.3	1.97 ± 1.3	2.0 ± 1.7	2.3 ± 1.9	1.9 ± 1.4	1.9 ± 1.3	1.9 ± 1.3	2.1 ± 1.5
Fruits (cups/day)	2.1 ± 1.8	1.95 ± 1.7	2.0 ± 1.7	2.1 ± 1.7	2.0 ± 1.7	1.9 ± 1.7	2.0 ± 1.6	2.2 ± 1.8
Red meat (g/day)	79.5 ± 73.0	90.3 ± 78.0	77.9 ± 72.0	71.1 ± 70.0	48.1 ± 48.0	56.6 ± 52.0	47.2 ± 44.0	41.6 ± 50.7
Fibers (g/day)	20.7 ± 11.1	21.2 ± 12.5	19.8 ± 11.0	20.9 ± 11.0	18 ± 10.1	17.5 ± 10.0	17.6 ± 9.5	19.2 ± 11.1
Alcohol (drinks/day)	1.3 ± 3.4	1.4 ± 3.7	1.3 ± 3.3	1.3 ± 3.3	0.4 ± 1.3	0.3 ± 1.2	0.5 ± 1.3	0.5 ± 1.3
Postmenopausal (%)	–	–	–	–	93.9	96.7	94.8	89.2

Note: Both in men and women, all comparisons between fitness categories are significantly different (p < 0.001).

Abbreviations: BMI = body mass index; METs = metabolic equivalents; NEE-CRF = non-exercise estimated cardiorespiratory fitness; NIH-AARP = National Institutes of Health-American Association of Retired Persons.

Table 2.

Incidence and risk analyses of NEE-CRF in men (n = 238,835) and women (n = 163,713).

Outcome/NEE-CRF categories	Men					Women
	Total cases	Low NEE-CRF	Moderate NEE- CRF	High NEE-CRF	p trend	Total cases	Low NEE-CRF	Moderate NEE- CRF	High NEE-CRF	p trend
Total cancer incidence cases (n (%))	64,344 (26.9)	17,570 (7.4)	35,045 (14.7)	11,729 (4.9)	<0.001	31,315 (19.1)	8709 (5.3)	15,995 (9.8)	6611 (4.0)	<0.001
HR (95%CI)		1 (Reference)	0.94 (0.91–0.97)	0.91 (0.87–0.96)	0.001		1 (Reference)	0.90 (0.86–0.95)	0.89 (0.83–0.95)	<0.001
Prostate cancer incidence cases in men and breast cancer incidence cases in women (n (%))	26,096 (10.9)	6039 (2.5)	14,571 (6.1)	5486 (2.3)	<0.001	10,997 (6.7)	2874 (1.8)	5601 (3.4)	2522 (1.5)	<0.001
HR (95%CI)		1 (Reference)	1.05 (0.99–1.1)	1.09 (1.0–1.2)	0.09		1 (Reference)	0.89 (0.82–0.96)	0.88 (0.79–0.99)	0.01
Lung and bronchus cancer incidence cases (n (%))	7740 (3.2)	2376 (1.0)	4150 (1.7)	1214 (0.5)	<0.001	4624 (2.8)	1173 (0.7)	2509 (1.5)	942 (0.6)	<0.001
HR (95%CI)		1 (Reference)	1 (0.9–1.1)	1.1 (0.94–1.3)	0.28		1 (Reference)	0.94 (0.83–1.1)	0.94 (0.78–1.1)	0.61
Colorectal cancer incidence cases (n (%))	5732 (2.4)	1880 (0.8)	3025 (1.3)	827 (0.3)	<0.001	3028 (1.8)	965 (0.6)	1516 (0.9)	547 (0.3)	<0.001
HR (95%CI)		1 (Reference)	0.88 (0.79–0.99)	0.70 (0.59–0.84)	<0.001		1 (Reference)	0.87 (0.75–1.00)	0.89 (0.71–1.1)	0.14

Notes: Risk models were adjusted for race, ethnicity, marital status, education level, first-degree relatives with a history of cancer, age, obesity (BMI of ≥30 kg/m²), meeting physical activity guidelines, diabetes, hypertension, dyslipidemia, smoking status, total alcoholic drinks per day, total energy intake, fiber intake, fruit intake, vegetable intake, and red meat intake. For women only, additional adjustments were made in the risk models for menopausal status and age of menopause start, years of oral contraceptives use, hormonal replacement therapy status and years of use.

Abbreviations: 95%CI = 95% confidence interval; BMI = body mass index; HR = hazard ratio; NEE-CRF = non-exercise estimated cardiorespiratory fitness.

Fig. 2.

Continuous risk models of NEE-CRF per 1-MET increase in men and women. MET = metabolic equivalent; NEE-CRF = non-exercise estimated cardiorespiratory fitness.

Risk associations between NEE-CRF and incidence of prostate and lung cancers among men and lung and colorectal cancers among women were not significant in the fully adjusted models (Table 2), but they were significant in models that excluded age, physical activity status, and obesity (Supplementary Table 1). Trend analysis showed that in moderate and high NEE-CRF categories there were fewer total cancer incidence cases in both sexes; there were also fewer incident cases of lung and colorectal cancers in men and fewer incident cases of breast, lung, and colorectal cancer in women (Table 2). Sensitivity analyses showed similar results with respect to risk association between higher NEE-CRF and lower cancer incidence risks, suggesting low risk of confounding, bias, and reverse causality (Supplementary Tables).

4. Discussion

Using a large prospective cohort of men and women, the current study demonstrated that higher NEE-CRF was independently associated with lower risk of total cancer incidence in both sexes. Additionally, higher NEE-CRF was associated with lower risk of colorectal cancer incidence in men and lower risk of breast cancer incidence in women (Table 2 and Fig. 2).

The results were further supported by sensitivity analyses, suggesting low risk of bias, confounding, and reverse causality (Supplementary Tables 2–4). Overall, these findings suggest that higher CRF, as assessed by the applied non-exercise estimated method, could provide protective benefits against the development of cancer, while low CRF levels could be used as a modifiable risk factor for cancer prevention. In agreement with previous recommendations to incorporate CRF as a “vital sign” in health and clinical settings,¹⁸ the current data support the assessment of NEE-CRF as part of cancer screening and the referral of low-fit individuals to exercise-based programs as a public health and disease prevention strategy.

The results of the current study are consistent with previous reports utilizing a direct CRF measure, which have generally shown that higher CRF provides protective benefits against the development of cancer.^17,19,30,38,41, 42, 43, 44, 45 A meta-analysis of 7 prospective studies demonstrated that compared to low CRF, moderate and high CRF levels were associated with a 14% and 19% lower risk of total cancer incidence and a 26% and 27% lower risk of colorectal cancer incidence among men, respectively.¹⁷ The Cooper Center Longitudinal Study of approximately 14,000 men showed that compared to low CRF, moderate and high CRF levels were associated with a 33% and 44% lower risk of colorectal cancer incidence, respectively.³⁰ The UK Biobank cohort study demonstrated a 4% lower risk of colorectal cancer incidence for every 1-MET higher CRF among 65,127 participants.¹⁹ A recent small study among 184 women demonstrated that each 1-MET increase in CRF was associated with a 20% lower risk of total cancer incidence.⁴⁵

To our knowledge, the current study is the first to demonstrate the protective benefits of higher NEE-CRF against the development of cancer. These findings extend previous reports of the association between directly measured CRF and cancer incidence^17,19,30,38,41, 42, 43, 44, 45 as well as studies on the association between NEE-CRF and mortality outcomes.22, 23, 24, 25, 26 The current study supplements existing knowledge with important insights on the association between NEE-CRF and cancer incidence among men and women. While the data are observational and do not prove cause and effect, the findings may have significant societal implications for public health, including cancer screening and prevention programs. Screening for CRF using the NEE-CRF method and referring low-fit individuals to exercise programs for CRF improvement could reduce the risk of cancer and enhance public health. Such programs are effective at improving CRF¹⁸ and have comparatively low costs, which could complement existing cancer-prevention strategies.⁴⁶

Consistent with previous reports of directly measured CRF, in the current study NEE-CRF was associated with colorectal cancer incidence among men but not among women in the fully adjusted models (Table 2).³¹ These sex differences are possibly due to hormonal and body-composition variance between men and women, and they require further investigation.³¹ Similar to studies using a directly measured CRF, in the current study NEE-CRF was not associated with prostate cancer incidence in men in the fully adjusted models,^17,42,44 although a trend toward higher risk was observed in the categorical (but not continuous) model, which requires future research.

While most previous studies have found an association between directly measured CRF and lung cancer incidence in men,^30,38,42,43 to our knowledge the current study is the first to analyze this association using NEE-CRF in both men and women. Although our findings align with data from the UK Biobank cohort,¹⁹ the results showed a lack of association between NEE-CRF and lung cancer incidence in both sexes when analyzed in the fully adjusted risk models. In contrast, significant association was found in risk models when age, phy?>sical activity, and obesity status variables were excluded (Supplementary Table 1), which warrants further research using both a direct measure and non-exercise estimate of CRF.

Cancer is a broad and varied group of diseases. The exact protective mechanisms of higher CRF against developing cancer are not fully understood and probably differ between cancer types and sites. Yet several potential biological pathways may explain the role of CRF in mediating cancer risk.47, 48, 49, 50 Higher levels of CRF are associated with improved insulin sensitivity, reduced chronic inflammation, enhanced regulation of sex steroid hormones and growth-related hormones, decreased adipose tissue, optimized immune function, elevated antioxidant capacity, enhanced DNA repair, cell proliferation and apoptosis, all of which may act as potential protective mechanisms against the initiation and progression of cancer. These mechanisms likely interact in a complex manner by blocking cancer cell initiation and countering cancer cell replication among individuals with higher CRF.47, 48, 49, 50 However, despite strong observational evidence supporting the preventive role of CRF in cancer,17, 18, 19 experimental randomized controlled studies addressing these potential mechanisms are needed.

4.1. Strengths

The study has several strengths, including a large sample size with over 400,000 men and women, prospective evaluation of incident cancer direct endpoints, a lengthy follow-up time (13.7 ± 3.2 years), and a wide range of analyses that were adjusted for many cancer risk factors in order to extract the independent association between NEE-CRF and cancer incidence. Additional strengths include sensitivity analyses for addressing potential biases and confounding as well as the use of the Surveillance, Epidemiology, and End Results national cancer registry and International Classification of Diseases for Oncology (ICD-O) codes for cancer outcomes ascertainment.³⁴ Finally, utilization of a validated NEE-CRF equation that does not require patient contact or physical measurements is an additional strength given its practical and widespread application.^18,27,29

4.2. Limitations

The study has several limitations. First, although the study is a large cohort with a relatively diverse U.S. population,²⁸ and although the cohort's demographics are similar to previous studies,^19,22, 23, 24, 25, 26^,30 most participants (>90%) were Non-Hispanic Whites; thus, future studies should focus on more diverse populations. Second, consistent with previous studies using the NEE-CRF method,22, 23, 24, 25, 26 physical activity status was self-reported, thereby 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.22, 23, 24, 25, 26 It is an established method for assessing physical activity, with strong predictive value for numerous health outcomes.^14,21 Third, although consistent with previous reports of NEE-CRF22, 23, 24, 25, 26 and the majority of epidemiological studies in this field, participants were assessed at baseline only, which did not account for potential changes in NEE-CRF over time.17, 18, 19, 20 Fourth, although NEE-CRF was predicted by a validated equation with high explained variability of VO_2max,^18,29 all non-exercise estimated methods are inherently imprecise with a relatively high standard error of estimate.^18,51 Finally, NEE-CRF was found to be strongly and independently associated with cancer incidence outcomes; however, the findings are observational in nature, and causal association needs further investigation.

5. Conclusion

This is the first and largest prospective study to demonstrate that higher NEE-CRF is associated with lower risk of total cancer incidence in men and women. A lower risk among those with higher NEE-CRF was also found for colorectal cancer incidence in men and breast cancer incidence in women. These results suggest that higher CRF, as assessed by the applied non-exercise estimated method, may have preventive benefits against the development of cancer in both sexes, while low CRF could be used as a modifiable cancer risk factor. Given the simplicity of obtaining NEE-CRF and the predictive value for cancer incidence, implementation of the applied method in health, clinical, or research settings could be beneficial for screening and prevention. The findings support the assessment of CRF directly or with the applied NEE-CRF method for cancer screening and the referral of low-fit individuals to exercise-based programs as a public health and cancer prevention strategy.

Acknowledgments

We are indebted to the participants in the NIH-AARP Diet and Health Study for their outstanding cooperation. We gratefully acknowledge the contributions of David Campbell and Michael Spriggs at Information Management Services, Linda Liao at the Division of Cancer Epidemiology & Genetics, and Virginia DeSeau at Technology Transfer Center for their important assistance with this study. We also thank Sigurd Hermansen and Kerry Grace Morrissey from Westat for study outcomes ascertainment and management and Leslie Carroll at Information Management Services for data support and analysis.

This research was supported (in part) by the Intramural Research Program of the NIH, National Cancer Institute. Cancer incidence data from the Atlanta metropolitan area were collected by the Georgia Center for Cancer Statistics, Department of Epidemiology, Rollins School of Public Health, Emory University, Atlanta, Georgia. Cancer incidence data from California were collected by the California Cancer Registry, California Department of Public Health's Cancer Surveillance and Research Branch, Sacramento, California. Cancer incidence data from the Detroit metropolitan area were collected by the Michigan Cancer Surveillance Program, Community Health Administration, Lansing, Michigan. The Florida cancer incidence data used in this report were collected by the Florida Cancer Data System (Miami, Florida) under contract with the Florida Department of Health, Tallahassee, Florida. The views expressed herein are solely those of the authors and do not necessarily reflect those of the Florida Cancer Data System or Florida Department of Health. Cancer incidence data from Louisiana were collected by the Louisiana Tumor Registry, Louisiana State University Health Sciences Center School of Public Health, New Orleans, Louisiana. Cancer incidence data from New Jersey were collected by the New Jersey State Cancer Registry, The Rutgers Cancer Institute of New Jersey, New Brunswick, New Jersey. Cancer incidence data from North Carolina were collected by the North Carolina Central Cancer Registry, Raleigh, North Carolina. Cancer incidence data from Pennsylvania were supplied by the Division of Health Statistics and Research, Pennsylvania Department of Health, Harrisburg, Pennsylvania. The Pennsylvania Department of Health specifically disclaims responsibility for any analyses, interpretations, or conclusions. Cancer incidence data from Arizona were collected by the Arizona Cancer Registry, Division of Public Health Services, Arizona Department of Health Services, Phoenix, Arizona. Cancer incidence data from Texas were collected by the Texas Cancer Registry, Cancer Epidemiology and Surveillance Branch, Texas Department of State Health Services, Austin, Texas. Cancer incidence data from Nevada were collected by the Nevada Central Cancer Registry, Division of Public and Behavioral Health, State of Nevada Department of Health and Human Services, Carson City, Nevada.

Authors’ contributions

BV contributed to study design and conception, data collection, statistical analysis, results interpretation, drafting, writing, and revising the submitted manuscript; JM contributed to the conception and design of the study, results interpretation, and to drafting the article and revising it critically for important intellectual content; CEM served as principle investigator and contributed to the conception and design of the study, data collection, results interpretation, and to drafting the article and revising it critically for important intellectual content. All authors have read and approved the final version of the manuscript, and agree with the order of presentation of the authors.

Competing interests

The authors declare that they have no competing interests.