Your entire health history. Nothing taken on trust.

Mandsager K, et al. Association of Cardiorespiratory Fitness With Long-term Mortality Among Adults Undergoing Exercise Treadmill Testing. JAMA Netw Open. 2018; 1(6):e183605. PMID: 30646252 · DOI: 10.1001/jamanetworkopen.2018.3605.

Retrospective cohort (n=122,007), observational. Risk-adjusted all-cause mortality fell with cardiorespiratory fitness and did not plateau at the top: elite vs. high adjusted HR 0.77 (0.63–0.95), low vs. elite 5.04 (4.10–6.20). An association measured in patients referred for treadmill testing — not a trial of training.

PMID 30646252·VO₂max

Leong DP, et al. Prognostic value of grip strength: findings from the PURE study. Lancet. 2015; 386(9990):266–273. PMID: 25982160 · DOI: 10.1016/S0140-6736(14)62000-6.

Prospective cohort (n≈140,000, 17 countries). Each 5 kg lower grip strength: all-cause mortality HR 1.16 (95% CI 1.13–1.20); a stronger predictor than systolic blood pressure.

PMID 25982160·Grip strength

Ference BA, et al. Association of Triglyceride- and LDL-C-Lowering Genetic Variants With Risk of Coronary Heart Disease. JAMA. 2019; 321(4):364–373. PMID: 30694319 · DOI: 10.1001/jama.2018.20045.

Mendelian randomization. CHD risk reduction proportional to absolute change in ApoB; associations of individual lipids became null after adjusting for ApoB — i.e. ApoB is the mechanistically relevant measure.

PMID 30694319·ApoB

Kronenberg F, et al. Lipoprotein(a) in atherosclerotic cardiovascular disease and aortic stenosis: a European Atherosclerosis Society consensus statement. Eur Heart J. 2022; 43(39):3925–3946. PMID: 36036785 · DOI: 10.1093/eurheartj/ehac361.

EAS consensus: "a causal and continuous association between Lp(a) concentration and cardiovascular outcomes," a risk factor even at very low LDL-C.

PMID 36036785·Lp(a)

Ridker PM, et al. Rosuvastatin to Prevent Vascular Events in Men and Women with Elevated C-Reactive Protein (JUPITER). N Engl J Med. 2008; 359(21):2195–2207. PMID: 18997196 · DOI: 10.1056/NEJMoa0807646.

RCT (n=17,802). Enrolment required LDL-C < 130 mg/dL and hs-CRP ≥ 2.0 mg/L; rosuvastatin cut the primary composite endpoint, HR 0.56 (95% CI 0.46–0.69). hs-CRP selected the population — the trial did not test lowering it. Rosuvastatin cut LDL-C 50% and hs-CRP 37%, so the benefit cannot be attributed to either alone. Read as: a risk marker good enough to enrich an RCT, not a demonstrated treatment target.

PMID 18997196·hs-CRP

Selvin E, et al. Glycated Hemoglobin, Diabetes, and Cardiovascular Risk in Nondiabetic Adults. N Engl J Med. 2010; 362(9):800–811. PMID: 20200384 · DOI: 10.1056/NEJMoa0908359.

Prospective cohort (ARIC). HbA1c predicts diabetes and cardiovascular risk in nondiabetic adults, with risk rising across the "normal" range.

PMID 20200384·HbA1c

Aune D, et al. Resting heart rate and the risk of cardiovascular disease, total cancer, and all-cause mortality — dose-response meta-analysis. Nutr Metab Cardiovasc Dis. 2017; 27(6):504–517. PMID: 28552551 · DOI: 10.1016/j.numecd.2017.04.004.

Meta-analysis (48 studies). Each +10 bpm resting heart rate: all-cause mortality RR 1.17 (95% CI 1.14–1.19).

PMID 28552551·Resting HR

Cappuccio FP, et al. Sleep duration and all-cause mortality: a systematic review and meta-analysis of prospective studies. Sleep. 2010; 33(5):585–592. PMID: 20469800 · DOI: 10.1093/sleep/33.5.585.

Meta-analysis of prospective cohorts (27 samples, n≈1,383,000). Short sleep RR 1.12 (1.06–1.18), long sleep RR 1.30 (1.22–1.38) for all-cause mortality; the authors conclude both are predictors of death. Duration was self-reported and the long-sleep arm carries significant heterogeneity, with reverse causation (illness lengthening sleep) not excluded. Our 7–9 h target reads these associations; no trial set it.

PMID 20469800·Sleep

Lewington S, et al. Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies (Prospective Studies Collaboration). Lancet. 2002; 360(9349):1903–1913. PMID: 12493255 · DOI: 10.1016/S0140-6736(02)11911-8.

Meta-analysis (~1M adults, 61 cohorts). Usual blood pressure is continuously and strongly associated with vascular mortality down to at least 115/75 mmHg, with no threshold; each +20 mmHg systolic ≈ doubles stroke and ischaemic-heart-disease death rates across middle and older age.

PMID 12493255·Blood pressure

SPRINT Research Group (Wright JT, et al.). A Randomized Trial of Intensive versus Standard Blood-Pressure Control. N Engl J Med. 2015; 373(22):2103–2116. PMID: 26551272 · DOI: 10.1056/NEJMoa1511939.

RCT (n≈9,300, high CV risk, non-diabetic). Targeting systolic < 120 vs. < 140 mmHg cut fatal + nonfatal major CV events and all-cause mortality — evidence for a tighter optimal than the classic 140 threshold (at the cost of more hypotension/AKI).

PMID 26551272·Blood pressure

C Reactive Protein Coronary Heart Disease Genetics Collaboration (CCGC). Association between C reactive protein and coronary heart disease: mendelian randomisation analysis based on individual participant data. BMJ. 2011; 342:d548. PMID: 21325005 · DOI: 10.1136/bmj.d548 · PMC3039696.

Mendelian randomization, 47 studies, n=194,418 (46,557 with CHD). Genetically raised CRP: risk ratio 1.00 (0.90–1.13) per 1 SD higher ln(CRP), against 1.33 (1.23–1.43) for circulating CRP in the same framework — discordant at P=0.001. The authors conclude CRP concentration itself "is unlikely to be even a modest causal factor in coronary heart disease." This is why hs-CRP is scored as a readout and never presented as something to lower directly.

PMID 21325005·hs-CRP (non-causality)

Ridker PM, et al. Antiinflammatory Therapy with Canakinumab for Atherosclerotic Disease (CANTOS). N Engl J Med. 2017; 377(12):1119–1131. PMID: 28845751 · DOI: 10.1056/NEJMoa1707914.

RCT (n=10,061, prior MI, hs-CRP ≥ 2 mg/L). Canakinumab (anti-IL-1β) cut the primary MACE endpoint at 150 mg — HR 0.85 (0.74–0.98, P=0.021) — with no reduction in lipids, isolating inflammation as a causal pathway rather than a correlate. Read with its limits: only the 150 mg dose met the multiplicity-adjusted threshold, all-cause mortality was neutral (HR 0.94, 0.83–1.06), and fatal infections rose. Together with ³⁰ it gives the honest position — the process is a target, the molecule is not.

PMID 28845751·Inflammation axis rationale

Manson JE, et al. Vitamin D Supplements and Prevention of Cancer and Cardiovascular Disease (VITAL). N Engl J Med. 2019; 380(1):33–44. PMID: 30415629 · DOI: 10.1056/NEJMoa1809944.

RCT (n=25,871, median 5.3 y). Vitamin D3 2000 IU/day was null on both primary endpoints: invasive cancer HR 0.96 (0.88–1.06) and major cardiovascular events HR 0.97 (0.85–1.12). Scope matters: participants were unselected and largely replete, so this is evidence against blanket supplementation — not against treating documented deficiency, which the trial did not test. Our 25-OH-D range is a status target, and this reference is why it is not also an instruction to supplement.

PMID 30415629·Vitamin D (limits of the target)

Jha P, et al. 21st-century hazards of smoking and benefits of cessation in the United States. N Engl J Med. 2013; 368(4):341–350. PMID: 23343063 · DOI: 10.1056/NEJMsa1211128.

Prospective cohort. 113,752 women and 88,496 men aged ≥ 25 interviewed in the US National Health Interview Survey (1997–2004), linked to deaths through 2006 (8236 and 7479 deaths). Current vs never smokers, ages 25–79: all-cause mortality HR 3.0 (women, 99% CI 2.7–3.3) and 2.8 (men, 99% CI 2.4–3.1); life expectancy shortened by more than 10 years; about 10, 9 and 6 years of life regained by quitting at 25–34, 35–44 and 45–54 respectively. Observational, with hazard ratios adjusted for age, education, adiposity and alcohol.

PMID 23343063·Smoking status (recorded, unscored)

Wood AM, et al. Risk thresholds for alcohol consumption: combined analysis of individual-participant data for 599,912 current drinkers in 83 prospective studies. Lancet. 2018; 391(10129):1513–1523. PMID: 29676281 · DOI: 10.1016/S0140-6736(18)30134-X.

Individual-participant data, 83 prospective studies across 19 high-income countries; 599,912 current drinkers without prior cardiovascular disease, 40,310 deaths over 5.4 million person-years. All-cause mortality rose curvilinearly with intake, with minimum risk at or below 100 g of ethanol per week; stroke was roughly linear (HR 1.14, 95% CI 1.10–1.17, per 100 g/week), as were heart failure (1.09) and fatal hypertensive disease (1.24), while myocardial infarction ran the other way (HR 0.94, 0.91–0.97). Scope matters: current drinkers only. It supports a threshold among people who drink and says nothing about drinking versus abstaining. Intake is reported in grams throughout — see Units on why "drinks" is not a unit.

PMID 29676281·Alcohol intake (recorded, unscored)

Sniderman AD, et al. A meta-analysis of low-density lipoprotein cholesterol, non-high-density lipoprotein cholesterol, and apolipoprotein B as markers of cardiovascular risk. Circ Cardiovasc Qual Outcomes. 2011; 4(3):337–345. PMID: 21487090 · DOI: 10.1161/CIRCOUTCOMES.110.959247.

Meta-analysis of 12 independent epidemiological reports, 233,455 subjects, 22,950 events, with all estimates converted to standardized relative risk ratios (RRRs). ApoB was the most potent marker (RRR 1.43, 95% CI 1.35–1.51), LDL-C the least (1.25, 1.18–1.33), non-HDL-C intermediate (1.34, 1.24–1.44); within-study head-to-head, apoB exceeded LDL-C by 12.0% (P < 0.0001) and non-HDL-C by 5.7% (P < 0.001). Complements ³: the Mendelian randomization gives the causal claim, this gives the head-to-head predictive margin.

PMID 21487090·ApoB

Kamstrup PR, Tybjærg-Hansen A, Steffensen R, Nordestgaard BG. Genetically elevated lipoprotein(a) and increased risk of myocardial infarction. JAMA. 2009; 301(22):2331–2339. PMID: 19509380 · DOI: 10.1001/jama.2009.801.

Mendelian randomization across three Copenhagen studies (CCHS n=8637; CGPS n=29,388; CIHDS n=2461). Observed plasma Lp(a) above the 95th percentile carried a multivariable-adjusted HR of 2.6 (95% CI 1.6–4.1) for myocardial infarction versus below the 22nd. The KIV-2 size polymorphism explained 21–27% of Lp(a) variation and tracked risk in the same direction, giving HR 1.22 (1.09–1.37) per doubling of genetically elevated Lp(a) — against 1.08 (1.03–1.12) for measured levels. This is the causal evidence the EAS consensus ⁴ rests on; note that it is a white Danish population, and that Lp(a) here is measured in mg/dL.

PMID 19509380·Lp(a) causality

Cholesterol Treatment Trialists' (CTT) Collaborators (Mihaylova B, et al.). The effects of lowering LDL cholesterol with statin therapy in people at low risk of vascular disease: meta-analysis of individual data from 27 randomised trials. Lancet. 2012; 380(9841):581–590. PMID: 22607822 · DOI: 10.1016/S0140-6736(12)60367-5 · PMC3437972.

Individual-participant meta-analysis of 27 randomized trials (22 statin vs control, n=134,537; 5 more- vs less-intensive, n=39,612). Major vascular events fell with RR 0.79 (95% CI 0.77–0.81) per 1.0 mmol/L LDL-C reduction — about a fifth — largely irrespective of age, sex, baseline LDL-C or prior vascular disease. The proportional reduction was at least as large in the two lowest-risk categories (RR 0.62 and 0.69) as in the highest, and in people without vascular disease all-cause mortality fell too (RR 0.91, 0.85–0.97). This moves LDL-C from guideline-cited ¹⁰ to primary-literature-cited, and makes it one of the few markers here demonstrated to be a lever rather than a readout — the intervention is statin therapy, which lowers more than LDL-C alone, but the dose-response runs with the LDL-C change.

PMID 22607822·LDL-C

García-Hermoso A, et al. Muscular Strength as a Predictor of All-Cause Mortality in an Apparently Healthy Population: A Systematic Review and Meta-Analysis of Data From Approximately 2 Million Men and Women. Arch Phys Med Rehabil. 2018; 99(10):2100–2113.e5. PMID: 29425700 · DOI: 10.1016/j.apmr.2018.01.008.

Meta-analysis of 38 cohorts, 1,907,580 participants, 63,087 deaths. Higher handgrip strength: all-cause mortality HR 0.69 (95% CI 0.64–0.74) versus lower, slightly stronger in women (0.60) than men (0.69). Knee-extension strength gave an independent HR 0.86 (0.80–0.93) — the signal is not specific to the hand, which is the point: grip is a cheap probe of whole-body strength, not itself the thing that matters. Corroborates ² at roughly fourteen times the sample, and is why the grip row is worded as a predictor rather than a lever.

PMID 29425700·Grip strength

American Diabetes Association. Standards of Care in Diabetes. Diagnostic thresholds: HbA1c 5.7–6.4% and fasting glucose 100–125 mg/dL define prediabetes; ≥ 6.5% / ≥ 126 mg/dL define diabetes.

HbA1c, fasting glucose

ESC/EAS Dyslipidaemia Guidelines and ACC/AHA Blood Cholesterol Guideline. Basis for LDL-C and triglyceride risk stratification and the lower-is-better principle for atherogenic lipids.

LDL-C, triglycerides, HDL

NCEP ATP III. Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III) final report. Circulation. 2002; 106(25):3143–3421. PMID: 12485966. Defines low HDL cholesterol sex-specifically: < 40 mg/dL (men) and < 50 mg/dL (women). Cutoffs cross-checked against the AHA/IDF-harmonised metabolic-syndrome criteria (PMC2675814). Also defines non-HDL cholesterol as total − HDL and sets its goal 30 mg/dL above the paired LDL-C goal, tabulating the pairs: LDL < 100

PMID 12485966·HDL sex stratification, non-HDL-C range

KDIGO Clinical Practice Guideline for CKD. eGFR < 60 mL/min/1.73m² sustained ≥ 3 months defines chronic kidney disease; ≥ 90 is normal filtration.

eGFR

Delanaye P, Jager KJ, Bökenkamp A, et al. CKD: A Call for an Age-Adapted Definition. J Am Soc Nephrol. 2019; 30(10):1785–1805. PMID: 31506289 · PMC6779354.

Argues the fixed KDIGO eGFR < 60 threshold over-diagnoses CKD in older adults and under-diagnoses it in the young; proposes age-adapted reduced-GFR thresholds of 75 (< 40 y), 60 (40–65 y) and 45 (> 65 y) mL/min/1.73m². Not sex-specific.

PMID 31506289·eGFR age stratification

Endocrine Society Clinical Practice Guideline (Vitamin D). Basis for sufficiency thresholds; both deficiency and very high 25-OH-D levels are suboptimal.

Vitamin D

NGSP / IFCC HbA1c master equation. The calibration relating the IFCC scale (mmol/mol) to the NGSP/DCCT scale (%): NGSP% = (0.09148 × IFCC) + 2.152. Maintained by the National Glycohemoglobin Standardization Program with the IFCC Working Group on HbA1c Standardization. Unlike the molar-mass conversions in Units, this is a calibration standard rather than arithmetic.

HbA1c unit conversion

Fasting insulin / insulin resistance supported by the HOMA-IR literature showing hyperinsulinemia precedes dysglycemia; no single landmark trial defines the optimal range, so our threshold is a preventive-medicine heuristic pending stronger evidence.

Fasting insulin

Body composition body-fat % and visceral adiposity outperform BMI as predictors of metabolic risk in the obesity/adiposity literature; our ranges are coarse, not yet sex-specific, and flagged as placeholders.

Body fat

Standard laboratory reference intervals (CBC + comprehensive metabolic panel; conventional adult clinical-lab ranges). The albumin, creatinine, lymphocyte %, MCV, RDW, ALP, and WBC ranges are standard adult clinical intervals — not tightened longevity targets — because these markers were originally ingested purely as PhenoAge inputs. Several carry independent mortality signal in the literature (RDW and WBC most notably), which is why WBC, lymphocyte %, and RDW now populate the Inflammation axis and albumin and ALP the Foundational axis. Note that they remain clinically calibrated while the other axes are longevity calibrated — see Limitations.

Albumin, lymphocyte %, RDW, ALP, WBC

Kaminsky LA, et al. Reference Standards for Cardiorespiratory Fitness Measured with Cardiopulmonary Exercise Testing: Data from the Fitness Registry and the Importance of Exercise National Database (FRIEND). Mayo Clin Proc. 2015 (updated 2021; global standards 2020). PMC4919021

VO₂max percentiles by sex and age decade, 20–79 y, from maximal CPX testing. 50th percentile falls from 48.0 (men) / 37.6 (women) at 20–29 y to 24.4 / 18.3 at 70–79 y — ~10% decline per decade.

VO₂max stratification

Dodds RM, et al. Grip Strength across the Life Course: Normative Data from Twelve British Studies. PLoS One. 2014; 9(12):e113637. PMID: 25474696 · DOI: 10.1371/journal.pone.0113637.

Pooled normative grip data, n=49,964 (26,687 female). Peak median grip 51 kg (men, 29–39 y) and 31 kg (women, 26–42 y).

PMID 25474696·Grip strength stratification

EWGSOP2 (Cruz-Jentoft AJ, et al.). Sarcopenia: revised European consensus on definition and diagnosis. Age Ageing. 2019.

Probable sarcopenia at grip strength < 27 kg (men) / < 16 kg (women). See also the German National Cohort grip analysis (n > 200,000, Age Ageing 2023) comparing population cutoffs to EWGSOP2.

Grip strength clinical thresholds

Gallagher D, et al. Healthy percentage body fat ranges: an approach for developing guidelines based on body mass index. Am J Clin Nutr. 2000; 72(3):694–701. PMID: 10966886.

Sex-, age- and ethnicity-specific healthy body-fat ranges derived by linking predicted %fat to NIH/WHO BMI limits. Healthy ranges ≈ 10–20% (men) and 18–28% (women); essential fat ≈ 2–5% and 10–13% respectively.

PMID 10966886·Body fat stratification (healthy band + floor)

Body-fat obesity cut-off points. No single authoritative definition of obesity by body-fat percentage exists — the WHO expert committee noted no formal agreement — but the literature converges on ≈ 25% (men) and ≈ 35% (women). De Lorenzo A, et al. New obesity classification criteria. World J Gastroenterol. 2016; PMID: 26811617 (states 23–25% men, 30–35% women). Oliveira BR, et al. Nutrients. 2023; PMID: 37447300 (cardiometabolic-risk cut-offs of 25–28% men, 37–40% women).

PMID 26811617PMID 37447300·Body fat upper borderline

Voss A, Schroeder R, Heitmann A, Peters A, Perz S. Short-Term Heart Rate Variability — Influence of Gender and Age in Healthy Subjects. PLoS One. 2015; 10(3):e0118308. PMID: 25822720 · DOI: 10.1371/journal.pone.0118308 · PMC4378923.

Normative short-term (5-min supine) HRV in n=1906 healthy adults. Mean RMSSD by age band and sex falls from 42.9 ms (women 25–34) and 39.7 ms (men 25–34) to 19.1 ms in both by 65–74 y — a large age effect and a small, time-domain-non-significant sex effect. The metric is RMSSD; window and device must be stated with any threshold.

PMID 25822720·HRV stratification and metric definition

Coronary artery calcium (CAC), MESA. Adding CAC to the MESA risk score raises 10-year CHD prediction from AUC 0.75 to 0.80, and CAC predicts events even among those at low Framingham risk; CAC progression independently predicts all-cause mortality.

Not currently a marker — recorded as the strongest identified candidate addition to Cardio Risk, pending a decision on imaging-derived inputs.

Ashwell M, Gunn P, Gibson S. Waist-to-height ratio is a better screening tool than waist circumference and BMI for adult cardiometabolic risk factors: systematic review and meta-analysis. Obes Rev. 2012; 13(3):275–286. DOI: 10.1111/j.1467-789X.2011.00952.x. See also Ashwell M, Hsieh SD (Int J Food Sci Nutr), and the systematic review supporting 0.5 as a global boundary value, PMID: 20819243.

Meta-analysis of > 300,000 adults across multiple ethnic groups. WHtR outperformed both waist circumference and BMI for cardiometabolic risk (mean AUROC 0.704 vs 0.693 and 0.671; 4–5% discrimination improvement over BMI, p < 0.01). Mean boundary value was 0.50 for both men and women across fourteen countries — one of the few anthropometric thresholds that does not require sex stratification.

Waist-to-height ratio

Liu Z, et al. A new aging measure captures morbidity and mortality risk across diverse subpopulations from NHANES IV. PLoS Med. 2018; 15(12):e1002718. PMID: 30596641 · DOI: 10.1371/journal.pmed.1002718. See also Levine ME, et al. (DNAm PhenoAge), Aging (Albany NY) 2018; PMID: 29676998, and Belsky DW, et al. (DunedinPACE), eLife 2022; PMID: 35029144 for the epigenetic clocks we would ingest rather than compute. Exact coefficients cross-checked against the reference R implementation (dayoonkwon/BioAge).

Operationalizes blood-biomarker PhenoAge on NHANES; the algorithm we implement.

PMID 30596641·PhenoAge / biological age

Mavrommatis C, et al. An unbiased comparison of 14 epigenetic clocks in relation to 174 incident disease outcomes. Nat Commun. 2025; 16:11164. PMID: 41402269 · DOI: 10.1038/s41467-025-66106-y.

Large-scale (n = 18,859) comparison of 14 widely used clocks against 174 incident disease outcomes and all-cause mortality over 10 years of follow-up. Second- and third-generation clocks outperform first-generation ones; of 176 significant associations, 27 diseases show a stronger association with a clock than that clock's own all-cause-mortality association. Cited for the structural point that clocks are not interchangeable — the comparison itself is of epigenetic clocks, which Evum ingests rather than computes.

PMID 41402269·More than one clock

Munoz ML, et al. Validity of (Ultra-)Short Recordings for Heart Rate Variability Measurements. PLoS One. 2015; 10(9):e0138921. PMID: 26414314 · DOI: 10.1371/journal.pone.0138921 · PMC4586373.

n = 3,387 adults (mean age 53), supine; intervals from finger arterial pressure (Portapres), not ECG. Against a 240–300 s reference (full text, Table 4, ln scale): SDNN r = 0.859 at 30 s and 0.956 at 120 s, with biases of 0.240 and 0.064; RMSSD r = 0.932 and 0.986. "RMSSD outperformed SDNN" at every length; 60 s was not tested.

PMID 26414314·`hrv_sdnn` criterion 2

Baek HJ, et al. Reliability of ultra-short-term analysis as a surrogate of standard 5-min analysis of heart rate variability. Telemed J E Health. 2015; 21(5):404–414. PMID: 25807067 · DOI: 10.1089/tmj.2014.0104.

n = 467 healthy volunteers aged 8–69, 5-min R-R series segmented down to 10 s. Minimum length to estimate the 5-min value reliably: 30 s for RMSSD, 240 s for SDNN.

PMID 25807067·`hrv_sdnn` criterion 2

O'Grady B, et al. The Validity of Apple Watch Series 9 and Ultra 2 for Serial Measurements of Heart Rate Variability and Resting Heart Rate. Sensors (Basel). 2024; 24(19):6220. PMID: 39409260 · DOI: 10.3390/s24196220 · PMC11478500.

n = 39 healthy adults, 316 morning measurements, supine, 5-min Breathe-app sessions, against a Polar H10 chest strap with Kubios (not a clinical ECG). SDNN: Apple lower by 8.31 ms (95% CI 5.59–11.04), MAPE 28.88 %, limits of agreement −53.8 to +37.2 ms (full text); not equivalent within the pre-specified ±10 ms. Resting HR MAPE 5.91 %. Note: the paper's Table 1 lists means of 85 ms (Kubios) and 93.3 ms (Apple), which contradicts its own sign; only the difference statistics are used here.

PMID 39409260·`hrv_sdnn` criterion 1

Bonneval L, et al. Validity of Heart Rate Variability Measured with Apple Watch Series 6 Compared to Laboratory Measures. Sensors (Basel). 2025; 25(8):2380. PMID: 40285070 · DOI: 10.3390/s25082380 · PMC12031371.

n = 78 adults aged 20–75 against a 3-lead ECG, Breathe-app recordings. R-R intervals and heart rate near-perfect at rest (MAPE 1.15 %); successive-interval differences only moderate (MAPE 31.31 % at rest). No SDNN reported.

PMID 40285070·`hrv_sdnn` criterion 1

Hernando D, et al. Validation of the Apple Watch for Heart Rate Variability Measurements during Relax and Mental Stress in Healthy Subjects. Sensors (Basel). 2018; 18(8):2619. PMID: 30103376 · DOI: 10.3390/s18082619 · PMC6111985.

n = 20, against a Polar H7 chest strap. Interval series agreement > 0.9; about 5 gaps per recording; time-domain indices not significantly affected by them.

PMID 30103376·`hrv_sdnn` criterion 1 (reference not ECG)

Miller DJ, et al. A Validation of Six Wearable Devices for Estimating Sleep, Heart Rate and Heart Rate Variability in Healthy Adults. Sensors (Basel). 2022; 22(16):6317. PMID: 36016077 · DOI: 10.3390/s22166317 · PMC9412437.

n = 53 (aged 25 ± 6), one lab night with PSG and ECG. Apple Watch HRV, as RMSSD exported through a third-party app with an unstated sampling period (full text): bias −9.6 ms, ICC 0.67, limits ±55.2 ms. Sleep/wake agreement 88 % for Apple Watch.

PMID 36016077·`hrv_sdnn` criterion 1

Dekker JM, et al. Heart rate variability from short electrocardiographic recordings predicts mortality from all causes in middle-aged and elderly men. The Zutphen Study. Am J Epidemiol. 1997; 145(10):899–908. PMID: 9149661 · DOI: 10.1093/oxfordjournals.aje.a009049.

Dutch men, 878 aged 40–60 and 885 aged 65–85. SDNN from the resting 12-lead ECG. 5-year age-adjusted all-cause mortality, SDNN < 20 vs 20–39 ms: RR 2.1 (1.4–3.0) in middle-aged and 1.4 (0.9–2.2) in elderly men.

PMID 9149661·`hrv_sdnn` criterion 3

Dekker JM, et al. Low heart rate variability in a 2-minute rhythm strip predicts risk of coronary heart disease and mortality from several causes: the ARIC Study. Circulation. 2000; 102(11):1239–1244. PMID: 10982537 · DOI: 10.1161/01.cir.102.11.1239.

Case-cohort within 14,672 adults aged 45–65 without CHD. Time-domain HRV from a 2-min strip, in tertiles: low HRV carried higher risk of incident CHD and death, not explained by other risk factors. The abstract does not name the individual indices, and the full text was not accessible.

PMID 10982537·`hrv_sdnn` criterion 3

Hillebrand S, et al. Heart rate variability and first cardiovascular event in populations without known cardiovascular disease: meta-analysis and dose-response meta-regression. Europace. 2013; 15(5):742–749. PMID: 23370966 · DOI: 10.1093/europace/eus341.

8 studies, n = 21,988. Lowest vs highest SDNN: RR 1.35 (1.10–1.67) for a first CV event; about 1 % lower risk per 1 % higher SDNN. Recording lengths are pooled, not limited to short-term.

PMID 23370966·`hrv_sdnn` criterion 3

Jarczok MN, et al. Heart rate variability in the prediction of mortality: A systematic review and meta-analysis of healthy and patient populations. Neurosci Biobehav Rev. 2022; 143:104907. PMID: 36243195 · DOI: 10.1016/j.neubiorev.2022.104907.

32 studies plus two individual-participant datasets, 38,008 participants. Lower HRV predicted higher mortality across ages, sexes, populations and recording lengths. Lowest quartile of 5-min RMSSD vs the others: HR 1.56 (1.32–1.85).

PMID 36243195·HRV (prognosis)

Plews DJ, Laursen PB, Kilding AE, Buchheit M. Heart rate variability in elite triathletes, is variation in variability the key to effective training? A case comparison. Eur J Appl Physiol. 2012; 112(11):3729–3741. PMID: 22367011 · DOI: 10.1007/s00421-012-2354-4.

Two elite triathletes, 77 days. The 7-day rolling average of ln RMSSD declined towards non-functional overreaching in one and stayed stable in the other. A case comparison (n = 2), cited for the averaging method only.

PMID 22367011·Body Battery (HRV window)

Plews DJ, et al. Monitoring training with heart rate-variability: how much compliance is needed for valid assessment? Int J Sports Physiol Perform. 2014; 9(5):783–790. PMID: 24334285 · DOI: 10.1123/ijspp.2013-0455.

Trained triathletes, ln RMSSD averaged over 1–7 random days per week: agreement with the full-week value plateaued after 3–4 days. Recommendation: at least 3 valid data points per week.

PMID 24334285·Body Battery (HRV minimum)

Kiviniemi AM, et al. Endurance training guided individually by daily heart rate variability measurements. Eur J Appl Physiol. 2007; 101(6):743–751. PMID: 17849143 · DOI: 10.1007/s00421-007-0552-2.

Randomized, 26 moderately fit men, 4 weeks. Morning HRV judged against an individual reference (10-day mean − SD); HRV-guided training improved VO₂peak, the predefined plan did not.

PMID 17849143·Body Battery (personal baseline)

Buchheit M. Monitoring training status with HR measures: do all roads lead to Rome? Front Physiol. 2014; 5:73. PMID: 24578692 · DOI: 10.3389/fphys.2014.00073 · PMC3936188.

Review. Individual changes should be read against the measure's error and smallest important change, and in training context; HR measures cannot inform on all of wellness and fatigue.

PMID 24578692·Body Battery (interpretation)

Busso T, Chalencon S. Validity and Accuracy of Impulse-Response Models for Modeling and Predicting Training Effects on Performance of Swimmers. Med Sci Sports Exerc. 2023; 55(7):1274–1285. PMID: 36791017 · DOI: 10.1249/MSS.0000000000003139.

11 swimmers, 61 weeks, daily load. Banister-type models fit well, but their prediction of future performance was "not satisfactory for individual training planning". The closest published load model; daily, not intraday.

PMID 36791017·Body Battery (A2 is a heuristic)

Caserman P, et al. Assessing the Accuracy of Smartwatch-Based Estimation of Maximum Oxygen Uptake Using the Apple Watch Series 7: Validation Study. JMIR Biomed Eng. 2024; 9:e59459. PMID: 39083800 · DOI: 10.2196/59459 · PMC11325102.

n = 19. Lab VO₂max (cycle ergometer, gas analysis) 45.88 ± 9.42 vs Apple Watch 41.37 ± 6.5 ml/kg/min; ICC(2,1) 0.47.

PMID 39083800·Apple VO₂max (open question)

Lambe R, et al. The accuracy of Apple Watch measurements: a living systematic review and meta-analysis. NPJ Digit Med. 2026; 9(1):63. PMID: 41513748 · DOI: 10.1038/s41746-025-02238-1 · PMC12823594.

82 studies, 14 metrics, searched to 2025-09-24. Heart rate bias −0.27 bpm (LoA −7.19 to 6.64). Full text: heart-rate variability had one study, O'Grady ⁴²; VO₂max one (n = 30, Apple lower by 6.07 ml/kg/min, "clinically significant"); the authors call for validation of respiratory rate and wrist temperature.

PMID 41513748·`hrv_sdnn` criterion 1 (evidence base)

Natarajan A, et al. Heart rate variability with photoplethysmography in 8 million individuals: a cross-sectional study. Lancet Digit Health. 2020; 2(12):e650–e657. PMID: 33328029 · DOI: 10.1016/S2589-7500(20)30246-6.

8,203,261 Fitbit users, one day each, 5-min windows. Age × sex benchmark tables for RMSSD and SDRR at two times of day; strong diurnal variation; SDRR differs by sex, RMSSD does not. Cross-sectional, with no outcomes.

PMID 33328029·`hrv_sdnn` (lead, not sufficient)

Theurl F, et al. Smartwatch-derived heart rate variability: a head-to-head comparison with the gold standard in cardiovascular disease. Eur Heart J Digit Health. 2023; 4(3):155–164. PMID: 37265873 · DOI: 10.1093/ehjdh/ztad022 · PMC10232241.

Garmin vivoactive 4 PPG against 1000 Hz ECG, 30 min, 263 subjects (mostly after myocardial infarction or stroke). SDANN concordance 0.96, rMSSD 0.66: "caution is warranted with HRV markers that predominantly assess short-term variability".

PMID 37265873·PPG RMSSD (lead)

Gunter EW, Lewis BG, Koncikowski SM. Laboratory Procedures Used for the Third National Health and Nutrition Examination Survey (NHANES III), 1988–1994. Atlanta, GA, and Hyattsville, MD: Centers for Disease Control and Prevention; 1996. Section VII-R, Serum C-Reactive Protein. wwwn.cdc.gov/nchs/data/nhanes3/manuals/labman.pdf. No PMID: a CDC laboratory manual, read at the source. "This method quantifies C-reactive protein (CRP) by latex-enhanced nephelometry … on a Behring Nephelometer"; under Reportable range of results: "Report values <0.3 as <0.3 mg/dL." NHANES III is the cohort Phenotypic Age was fit on: "For step 1, NHANES III was used to generate a measure of phenotypic age" (Levine ME, et al., Aging (Albany NY) 2018, PMID 29676998). Neither source says how values below the floor were coded in the analysis, and nothing here relies on it.

hs-CRP and standard CRP