Longevity & Aging · Metabolic & Cardiometabolic
cardiorespiratory fitness VO2max predicts all-cause mortality risk
In plain terms: Does higher VO2max/fitness actually predict living longer?
Strong support Longevity & Aging 💰 Industry COI noted
Part of: • Cardiorespiratory Fitness & Strength
RefutedContestedStrong support
consensus score 1.00
Yes — higher fitness is one of the strongest and most reproducible predictors of lower all-cause mortality, though the data are observational and cannot prove fitness itself causes the added years.
📅 Last reviewed: 2026-07-14 ⓘ
Evidence ladder
How far up the ladder this claim has climbed. A high consensus on a low rung means "consistent so far," not "proven in people."
Top evidence so far: All trials, pooled (Meta-analysis)
MechanismIn-vitroAnimalObservationalRCTMeta-analysis
How the studies fall
9 support 0 contradict 0 tested null 0 mixed · 9 sources, 8 independent groups
The evidence (13)
| Source | Grade | Stance | Quality | Finding |
|---|---|---|---|---|
| Kokkinos 2017 · Mayo Clin Proc | observational | supports | moderate | Prospective VA cohort, 20,590 veterans with 2846 events over a median 11.3 years, whose sole analysed endpoint is a cardiovascular composite, not mortality: 'During follow-up (median, 11.3 years; range, 0.3-33.0 years), 2846 individuals experienced MACEs. The CRF-MACE association was inverse and graded.' Risk rose below the age-specific MET threshold (HR, 1.95; 95% CI, 1.73-2.21 and HR, 1.41; 95% CI, 1.27-1.56) and fell above it (HR, 0.77; 95% CI, 0.68-0.87 and HR, 0.57; 95% CI, 0.48-0.67), concluding 'Increased CRF is inversely and independently associated with the risk for MACE.' No mortality result is reported anywhere; the single occurrence of the word is an unfulfilled Methods line, 'Multivariate Cox models were used to estimate HR and 95% CIs for mortality across fitness categories.' Exposure is age-specific peak METs from a symptom-limited treadmill test, not gas-exchange VO2max. OFF-SCOPE for an all-cause-mortality claim; the earlier extract's 'each 1-MET' framing is also unreported - the paper models quintiles and threshold-anchored categories. |
| Han, Qie, Shi, Yang, Lu, Hu, Zhang, Zhang, Hu, Zhao 2022 · Br J Sports Med | meta-analysis | supports | high | Dose-response meta-analysis of 34 prospective cohort studies (PROSPERO CRD42020208883, search to 26 December 2019): 'The pooled relative risks (RRs) for all-cause, CVD and cancer mortality per one-metabolic equivalent increase in CRF were 0.88 (95% CI 0.83 to 0.93), 0.87 (95% CI 0.83 to 0.91) and 0.93 (95% CI 0.91 to 0.96), respectively.' Versus lowest CRF, all-cause mortality RR was 0.67 (95% CI 0.61 to 0.74) for intermediate and 0.47 (95% CI 0.39 to 0.56) for highest CRF. CRF was assessed by exercise stress test, so this is estimated fitness in METs rather than gas-exchange VO2max, and the substrate is observational (confounding and reverse causation not excluded). Competing interests: none declared. Read from abstract only. |
| Korpelainen 2016 · Ann Med | observational | supports | moderate | Prospective Finnish cohort of 3033 subjects referred for symptom-limited bicycle exercise testing, average follow-up 19 years: 'Low exercise capacity (1st quartile) was associated with a hazard ratio of 4.2 (95% CI: 1.7, 10.8) for CVD and 4.0 (95% CI: 2.5, 6.4) for all-cause mortality compared with high exercise capacity (4th quartile) among men and in women with a 5.4-fold (95% CI: 1.2, 24.0) risk for CVD and 2.3-fold (95% CI: 1.2, 4.3) risk for all-cause mortality, respectively.' Fitness was a workload surrogate — 'Exercise capacity was defined as the mean of last four minutes of exercise workload' — not measured VO2max, and the sample was clinically referred rather than general-population; abstract-only basis, so covariate adjustment is unverified. |
| Ung GA, Nguyen KH, Hui A, Wong ND, Dineen EH 2024 · Am Heart J Plus 2024;42:100395 | observational | supports | low | NHANES III (1988–1994), 13,968 adults aged 20–<79, NDI mortality linkage to 31 Dec 2018, mean follow-up 17.5 years. Adjusted all-cause mortality HR for the highest vs lowest fitness tertile: 0.29 (95% CI 0.25–0.34) non-DM, 0.28 (0.22–0.37) pre-DM, 0.28 (0.21–0.38) DM — 'Regardless of DM status, there was lower all-cause mortality with higher CRF levels.' CRITICAL CAVEAT, from the paper's own Methods: VO2max was never measured. It was computed as 79.9 − 0.39(age) − 13.7(sex) − 0.127(weight in lbs) via the FRIEND non-exercise equation — three demographic inputs, no exercise test — and then 'Age, sex, and BMI were not included in our adjusted model due to the FRIEND equation already accounting for these variables.' The fitness tertiles are consequently age bands (mean 61 / 44 / 21 years) and sex bands (51% / 35% / 14% female); the authors concede they 'cannot rule out further confounding due to both unmeasured and measured factors related both mortality and VO2Max, including age, sex, and weight which are part of the FRIEND equation.' Also non-monotone: in pre-DM the MIDDLE tertile was null for all-cause mortality (HR 0.96, 95% CI 0.83–1.12, p=0.55) and for CV mortality (HR 1.02, p=0.89). Directionally consistent with the claim, evidentially near-empty — hence quality low, unchanged. The prior extract's 'replication in a US population sample' overstated it: nothing about fitness was independently measured here. |
| Laukkanen JA, Isiozor NM, Kunutsor SK 2022 · Mayo Clin Proc 2022 Jun;97(6):1054-1073 | meta-analysis | supports | high | Systematic review and random-effects meta-analysis of 37 unique prospective cohort studies (2,258,029 participants, 108,613 all-cause mortality events; MEDLINE/Embase/Web of Science, inception to December 2021; GRADE applied). 'The pooled multivariable-adjusted RR for all-cause mortality comparing the top vs bottom tertiles of CRF levels was 0.55 (95% CI, 0.50 to 0.61).' In MET units the pooled RR was 0.56 (0.50 to 0.62), and 'For every 1-MET increase in CRF, the RR for all-cause mortality was 0.89 (95% CI, 0.86 to 0.92)' — a graded dose-response, not a threshold. 'Strength of the association did not differ by publication year, age, sex, follow-up duration, CRF assessment method, or risk of bias', which answers the vault's standing caveat that most CRF sources use estimated METs rather than gas-exchange VO2max. Substrate is observational and the authors frame it as such — 'Aggregate analysis of observational cohort studies confirms a strong inverse and independent association between CRF and all-cause mortality risk' — so residual confounding and reverse causation (subclinical illness depressing measured fitness) are not excluded; the closing call to add CRF to standard risk panels is a recommendation, with no discrimination/reclassification statistic reported. Abstract-grade: Mayo Clin Proc is paywalled and not in PMC, so I2, publication-bias tests, the assigned GRADE rating, funding/COI and the 37-study list are unread. NOT independent of the claim's other meta-analytic votes: search window and per-MET estimate (0.89) essentially duplicate s35022163 (34 cohorts to Dec 2019, 0.88), and the individual cohorts already appraised on this claim may sit inside this pool. |
| Duggan 2024 · J Clin Med | observational | supports | moderate | 14,550 US Veterans, CABG at least 6 months after a symptom-limited treadmill test; mean follow-up 10.0 +/- 5.4 y, 6,502 deaths (43.0%). Inverse and graded: 'Each 1-MET increase in CRF was associated with an 11% lower mortality risk' (HR 0.89, CI 0.88-0.90), and vs Least-Fit, mortality was 22% lower in Low-Fit, 31% lower in Moderate-Fit, 52% lower in Fit and 66% lower in High-Fit, adjusted for age, BMI, race, CVD, prior PCI, medications and risk factors. Authors' own caveats: 'the association between fitness and survival, while compelling, does not indicate causality', population 'limited to primarily male Veterans', comorbidity burden higher in the Least-Fit quintile, and 'the cause of death of our subjects is not known'. Fitness is estimated METs from ACSM speed/grade equations, not gas-exchange VO2max. Same VA ETHOS registry as s27876315/s24821944. |
| Madden AM 2026 · BMC Geriatr | observational | supports | moderate | PRISMA systematic review, 94 studies in 104 manuscripts (Scopus/Cochrane/PubMed/Web of Science, 2016 to Feb 2024), adults >=65y; 17 studies reported all-cause mortality (13 physical-activity exposure, 4 cardiorespiratory-fitness exposure). 'Similar to findings for PA, higher levels of CRF were significantly associated with lower all-cause mortality risk'. Largest input, Kokkinos ETHOS n=137,626 veterans: 'Mortality risk decreased with increasing level of CRF' and 'in those aged 80-95: extremely fit participants had 40.6% mortality compared to 75.7% among least fit participants (HR 0.27, 95% CI 0.24-0.29)'; Phan n=14,499: 'Patients aged 80-90 with low fitness had the highest risk of all-cause mortality (HR 10.6, 95% CI 7.8-14.4'. Of the 16 observational studies on this outcome, 11 were good and 1 poor quality by Newcastle-Ottawa. MEASUREMENT CAVEAT: only 1 of the 4 CRF studies (Woo 2019, n=1176) used measured peak oxygen uptake -- 'those who died had lower VO2peak (women: 19.5 [SD 3.8] vs. 17.1 [SD 3.6];' -- the rest used '6MWT/distance, 6-minute walk speed, peak oxygen uptake using the cycle ergometer, or the exercise tolerance test and treadmill speed and grade'. SYNTHESIS CAVEAT: 'Meta-analysis was not performed because of the heterogeneity of the PA exposure and of the various outcomes.' -- narrative synthesis only; 'This review was not prospectively registered.'; funded by the AXA Research fund with a co-author employed by Vitality Group and holding equity in Discovery Ltd. |
| Mandsager, Harb, Cremer, Phelan, Nissen, Jaber 2018 · JAMA Netw Open | observational | supports | high | Retrospective single-center cohort of 122,007 consecutive adults referred for symptom-limited treadmill testing (Cleveland Clinic 1991-2014; 13,637 deaths over 1.1 million person-years, median follow-up 8.4 y). Risk-adjusted all-cause mortality fell monotonically with cardiorespiratory fitness and was lowest in elite performers (>=97.7th percentile for age and sex): elite vs low adjusted HR 0.20 (95% CI 0.16-0.24), elite vs high 0.77 (0.63-0.95) - no plateau and no U-shaped harm at the top of the distribution, holding at both 95th- and 99th-percentile elite cutoffs. Read in the risk direction, it is LOW vs ELITE fitness that carries adjusted HR 5.04 (4.10-6.20). The paper's own comparison is that reduced fitness is 'comparable to or greater than' traditional clinical risk factors - CAD adjusted HR 1.29 (1.24-1.35), smoking 1.41 (1.36-1.46), diabetes 1.40 (1.34-1.46) - and the like-for-like adjacent-band contrast (below average vs above average) is 1.41, equal to smoking rather than worse than it. Elite-over-high survival was significant only in patients >=70 y (0.71, 0.52-0.98) and in hypertension (0.70, 0.50-0.99). Scope limits the authors state themselves: exposure is ESTIMATED METs from treadmill grade and speed at a SINGLE test, not gas-exchange VO2max; 'the association between CRF and mortality does not prove causation'; and the referral population 'may not reflect the general population'. |
| Ezzatvar 2021 · J Sport Health Sci | meta-analysis | supports | moderate | Meta-analysis of 20 prospective cohorts (156,371 patients with established CVD, mean age 61.4 y): all-cause mortality HR 0.42 (95%CI 0.28-0.61) for high vs low CRF, with a significant dose-response of HR 0.81 (0.74-0.88) per 1-MET increase; CAD subgroup HR 0.32 (0.26-0.41). Direction is unambiguous, but quality is moderate not high: the pooled all-cause estimate carries I2 = 94.6% and the paper's own test reports 'Asymmetry suggestive of small-study effects was observed for all-cause mortality (LFK index = -5.07)' - major asymmetry by its own threshold - over observational primaries with study-specific CRF cut-offs and acknowledged misclassification bias. |
| Faselis 2014 · Hypertension | observational | supports | moderate | 2153 hypertensive men aged 70+ (mean 75), 1039 deaths over mean 9.0 y: all-cause mortality 11% lower per 1-MET (HR 0.89, 95% CI 0.86-0.93). Vs <=4.0 METs: low-fit HR 0.82, moderate-fit HR 0.64, high-fit (>8 METs) HR 0.52 - the halving is the TOP-vs-BOTTOM fitness contrast, and the exposure is estimated peak METs, not measured VO2max. |
| Cozma D, Gaita D, Crisan S, Tudoran C, Dumitrescu AS, Vacarescu C 2026 · J Clin Med | observational | supports | low | Narrative review (no pooling, no new data, PRISMA declined), searching 1966-2024: 'Cardiorespiratory fitness (CRF), indexed by maximal oxygen uptake (VO 2 max), demonstrates the strongest and most linear dose-response relationship with all-cause mortality identified in preventive medicine, with every 1 metabolic equivalent of task (MET) increment associated with a 12-15% reduction in mortality risk.' The numbers are secondhand from cited primaries: Mandsager 2018 (n=122,007) - Elite fitness had an 80% lower all-cause mortality vs Low, Above-Average 45% lower; Myers VETS 2002 (n=6,213 men) - each 1-MET increase associated with a 12% improvement in survival. The paper disclaims causation unprompted: 'These associations derive from observational cohort data and are subject to residual confounding, healthy volunteer bias, and reverse causation; causal inference cannot be established from the current evidence base', and 'no randomized trial has demonstrated that raising VO 2 max through exercise training directly reduces mortality risk in proportion to the associations reported in cohort studies.' Our claim reads PREDICTS, so that caveat leaves the stance intact - but this source would CONTRADICT any causal restatement. Zero-weight per CONVENTIONS 5 (a review is not evidence); it restates primaries the vault should count directly. J Clin Med 2026;15(12):4597, MDPI; no external funding, authors declare no conflicts; revised and accepted the same day. |
| Lang 2024 · Br J Sports Med | meta-analysis | supports | low # adjudicated calibration #10 (2026-08-19): the authors' OWN supplemental GRADE rates this outcome "very low certainty" — downgraded on 4/5 domains incl. publication bias (funnel asymmetry + Egger's), I2=99.4% unexplained | Overview of systematic reviews (PROSPERO CRD42022370149), 26 meta-analyses / >20.9M observations / 199 unique cohorts. Directionally clear: 'CRF had the largest risk reduction for all-cause mortality when comparing high versus low CRF (HR=0.47; 95% CI 0.39 to 0.56)', with dose-response 'for every 1-metabolic equivalent of task (MET) higher level of CRF ... a 11%-17% reduction in all-cause mortality (HR=0.89; 95% CI 0.86 to 0.92, and HR=0.83; 95% CI 0.78 to 0.88)'. But it pools nothing itself - 'We summarised results using a narrative synthesis approach using summary of findings tables. Results were reported as described by the systematic review authors.' - and those all-cause-mortality estimates are Han 2022 s35022163 (the HR=0.47) and Laukkanen 2022 s35562197, both already independent sources on this claim ('Estimates from Laukkanen (2022), Han (2022), Kodama (2009) and Aune (2020) were reported as RR'). Counted as review, not a third primary meta-analysis. Authors' own certainty: 'The certainty of the evidence across all studies ranged from very low-to-moderate', downgraded largely for indirectness - 'there were 1 858 274 male participants compared with 180 202 female participants for all-cause mortality'. [Supplemental table 3: 'There is very low certainty' of the protective effect for high-vs-low CRF; 'Publication bias: -1 point, asymmetry was identified using funnel plots and' Egger's test.] |
| Weeldreyer NR, De Guzman JC, Paterson C, Allen JD, Gaesser GA, Angadi SS 2025 · Br J Sports Med 2025;59(5):483- | meta-analysis | supports | moderate | Systematic review and meta-analysis, 20 prospective cohorts, 398,716 observations analysed (67% male), follow-up 7.7-26 y; three-level REML random-effects model with robust variance estimation, PROSPERO CRD42023392979. Against a normal weight-fit reference, all-cause mortality was raised in every unfit stratum - normal weight-unfit HR 1.92 (1.43-2.57), overweight-unfit 1.82 (1.47-2.24), obese-unfit 2.04 (1.54-2.71) - while fit strata were not - overweight-fit 0.96 (0.61-1.50), obese-fit 1.11 (0.88-1.40). Median statistical power for the fit-unfit contrasts 98-100%; results held after removing every influential trial and cluster. Authors: 'Our results show that CRF remains a potent predictor of mortality risk independent of BMI.' Quality held at moderate, not high: the underlying evidence is observational, the ACLS/CCLS cluster supplies most of the sample, CRF is dichotomised (fit usually = above the study's 20th percentile) so no gradient is estimable, heterogeneity is substantial (I2 level-2 59.7-81.6% in the unfit models), publication bias was assessed by eye only ('Visual inspection of the funnel plots revealed some asymmetry') with no Egger test or GRADE rating, the search was English-only, and the cohorts are predominantly white, US and higher-SES. No specific grant; competing interests 'None declared'. |
Disagree, or know a study we missed?
We grade by evidence, not opinions. The way to weigh in is to point us to a study we haven't cited (check the evidence table above first), or to flag a problem with one we have. Every submission is reviewed; if it holds up, the grade updates and shows in Science Changes Its Mind.
Opens a short form. You'll sign in with Google so submissions are tied to a real account — we don't display your identity, and we only accept a link we can verify (PubMed, DOI, ClinicalTrials.gov).
Educational only, not medical advice. Grades and scores reflect published evidence weighted by study design and quality; see the methodology.