Longevity & Aging · Metabolic & Cardiometabolic
major longevity-fitness cohorts Mandsager and VA studies differs in estimated cardiorespiratory fitness via treadmill not gas-exchange VO2max and wearable estimates carry 7-16 percent error
In plain terms: Did the big VO2max-and-longevity studies rely on estimated fitness rather than measured VO2max?
Part of: • Cardiorespiratory Fitness & Strength
Yes — they estimated fitness from treadmill time or METs, not gas-exchange VO2max, and wearable estimates carry 7-16% error. That is a methods caveat, not a refutation: estimated and measured fitness predict mortality near-identically.
📅 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)
How the studies fall
What the evidence shows
The large fitness-and-longevity cohorts estimated cardiorespiratory fitness from treadmill performance or METs rather than measuring gas-exchange VO2max, and wearable estimates carry 7 to 16 percent error, one Apple Watch validation underestimating by about 6 mL/kg/min. A methods caveat, not a refutation: estimated and measured fitness predict mortality near-identically.
The evidence (15)
| Source | Grade | Stance | Quality | Finding |
|---|---|---|---|---|
| Cheng R, Lin W, Song L, Pan J, Wang N, Li X 2025 · Frontiers in Physiology | observational | supports | low | 'Twenty male well-trained collegiate soccer players (Table 1) volunteered to participate in this study.' The paper validates the 30-15IFT field shuttle-run test's Buchheit (2008) VO2max prediction equation against gold-standard treadmill CPET in this cohort, finding systematic bias ('Despite large correlations with CT measures (r = 0.62-0.77), Bland-Altman analysis revealed significant systematic bias and wide limits of agreement'; VO2max overestimated by 4.08 mL/kg/min, 95% LoA -1.41 to 9.56) and testing a same-sample-derived population-specific equation (SEE 2.90 vs 4.91 mL/kg/min). No mention anywhere in the text of the Mandsager 2018 JAMA Network Open study, VA exercise-testing cohorts, or wearable-device fitness estimates -- off-scope for this claim's named subject. |
| Ainsworth BE, Cai Z 2025 · J Sport Health Sci | observational | supports | moderate | Commentary confirms estimated CRF (treadmill duration, submaximal, non-exercise algorithms) predicts mortality comparably to cardiopulmonary-exercise-tested CRF across test modalities. |
| Matsuo T, So R, Murai F 2022 · Journal of cardiovascular development and disease | observational | supports | moderate | A combined multiple-regression/linear-extrapolation heart-rate step-test method for estimating VO2max was compared to treadmill-measured VO2max in 128 adults, with standard errors of estimate of 4.15 (MRM) and 5.08 (LEM) mL/kg/min and signi |
| Patel PS, Heller S, Larson KF, Elfessi NM, Sydo N, Carta KG, Hussain N, Allison TG, Newman DB 2025 · Am J Cardiol | observational | mixed | moderate | 18,961-patient Mayo treadmill cohort scores functional aerobic capacity from treadmill performance, not gas exchange — illustrates Topol's point that large cardiology cohorts use treadmill-estimated capacity as the exposure. |
| Hansen MT, Rømer T, Sahl RE, Lange KK, Poggi AI, Furbo K, Augsburg JB, Chmura P, Nielsen B, Larsen MN, Krustrup P, Helge JW 2025 · European journal of sport science | observational | supports | moderate | A seismocardiography-based nonexercise VO2peak estimation method was compared to CPET-measured VO2peak in football players; bias was -1.0 mL/kg/min with wide 95% limits of agreement (+-9.1 mL/kg/min), SEE 4.8 mL/kg/min, mean absolute percen |
| Lambe R, O'Grady B, Baldwin M, Doherty C 2025 · PLoS One | observational | supports | moderate | Apple Watch VO2max validation vs indirect calorimetry: MAPE 13.31%, underestimated by about 6 mL/kg/min — confirms Topol's wearable-error claim (within his 7-16% range). |
| Lee K, Kim D, Shin S, Choi H, Jang A, Chung J 2025 · Healthcare (Basel, Switzerland) | observational | supports | moderate | A smartwatch-derived VO2max prediction model (heart rate + GPS from a 12-min Cooper test) was compared to treadmill-measured VO2max in Korean adults; best model SEE was 3.02-3.18 mL/kg/min (R2=0.85-0.86), still leaving unexplained error ver |
| Schembre SM, Riebe DA 2011 · Measurement in physical education and exercise science | observational | supports | moderate | A non-exercise VO2max estimation equation derived from the IPAQ self-report questionnaire explained only 43% of variance in measured VO2max (SEE=5.45 mL/kg/min); estimated VO2max fell within submaximal-testing error limits (20%) for only 87 |
| Molina-Garcia P, Notbohm HL, Schumann M, Argent R, Hetherington-Rauth M, Stang J, Bloch W, Cheng S, Ekelund U, Sardinha LB, Caulfield B, Brønd JC, Grøntved A, Ortega FB 2022 · Sports medicine (Auckland, N.Z.) | meta-analysis | supports | high | INTERLIVE meta-analysis of 14 validation studies found consumer wearables using resting-condition algorithms overestimated VO2max by a mean of 2.17 mL/kg/min (limits of agreement -13.07 to 17.41), while exercise-based algorithms had smaller |
| Singh B, Cadenas-Sanchez C, da Costa BGG, Castro-Piñero J, Chaput JP, Cuenca-García M, Maher C, Marín-Jiménez N, McGrath R, Molina-García P, Myers J, Gower B, Ortega FB, Lang JJ, Tomkinson GR 2024 · J Sport Health Sci | meta-analysis | mixed | high | Meta-analysis of 42 studies/3.8M observations: objectively-measured, exercise-estimated, and non-exercise-estimated CRF show similar associations with all-cause and CVD mortality — supports Topol's factual point that cohorts used estimated CRF, but undercuts the implication that this weakens the mortality signal. |
| Chappell H, Gilliam Z, Madero B, Springer J, Pipoly M, Baller KL, Oehler C, Long JD, Pierce GL, Voss MW 2026 · Exercise, sport & movement | observational | supports | moderate | Nonexercise CRF prediction equations (using waist circumference, body roundness index, self-reported activity) were validated out-of-sample against measured VO2max in older adults, explaining only 42-52% of variance (R2) with constant error |
| Qiu S, Cai X, Sun Z, Wu T, Schumann U 2021 · Atherosclerosis | meta-analysis | supports | moderate | Meta-analysis found algorithm-estimated CRF predicts mortality, underscoring that much cohort evidence relies on estimated rather than directly measured fitness. |
| Wolf C, Blackwell TL, Johnson E, Glynn NW, Nicklas B, Kritchevsky SB, Carnero EA, Cawthon PM, Cummings SR, Toledo FGS, Newman AB, Forman DE, Goodpaster BH 2023 · medRxiv | observational | supports | moderate | SOMMA cohort methods paper details that directly measured VO2peak via CPET is feasible but resource-intensive, explaining why large cohorts often substitute estimates. |
| Dakule K, Veseta U, Arnis V, Grinberga K, Kalejs O, Tomsone S 2025 · Journal of functional morphology and kinesiology | observational | supports | moderate | In masters basketball players, indirect VO2max protocols (Astrand-Ryhming, YMCA, Polar OwnIndex) were compared to gold-standard CPET; MAE was ~17-18% for the best-performing tests and the Polar OwnIndex overestimated VO2max by ~30% (MAE 32% |
| Carrier B, Marten Chaves S, Navalta JW 2025 · Sensors | observational | supports | low | Independent Garmin fenix 6 VO2max validation in general population confirms non-trivial estimation error, corroborating wearable-inaccuracy claim. |
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.