A Critical Appraisal of the Oura Ring’s Vascular-Ageing Claims, with Particular Reference to Women

Read published paper on Zenodo ↗

Abstract

Background. Arterial stiffness, measured as carotid–femoral pulse wave velocity (cfPWV), is an established independent predictor of cardiovascular events. Since 2024 the Oura Ring has reported a “Cardiovascular Age” (CVA), derived from nocturnal finger photoplethysmography (PPG). The ring describes CVA as an estimate of “how your heart and large arteries are aging compared to your chronological age” and classifies each user as below, aligned with, or above that age. Cardiovascular disease is the leading cause of death in women. Women’s arterial stiffness also follows a sex-specific trajectory, accelerating around the final menstrual period. A consumer metric of this kind therefore carries particular consequences for women.

Objective. To set out how arterial stiffness and vascular age are properly measured and validated, to examine what a finger PPG waveform can and cannot represent, and to appraise Oura’s CVA claims and evidence. Particular attention is given to women and to statements by Oura’s Chief Executive filed with the US Securities and Exchange Commission (SEC).

Findings. The ring does not measure pulse wave velocity, and Oura says so. Its only reference-standard validation is a company white paper. In 99 participants (63 women), it reported a correlation of r = 0.83 between estimated PWV and cfPWV, with no Bland–Altman analysis, no error metric and no results by sex. The ARTERY Society validation standard requires agreement statistics, which a correlation coefficient cannot supply. The only peer-reviewed study is an academic one in 160 mostly young adults. It did not use Oura’s algorithm, had no sex-stratified results and used chronological age rather than measured stiffness as its target. The ring-derived model had a mean absolute error of 7.25 years, which is wider than Oura’s ±5-year “aligned” band. Models trained to predict chronological age have a known regression-to-the-mean bias. That bias makes younger users appear older and older users younger, independent of their arteries. For women, the evidence is absent rather than adverse. No analysis addresses menstrual-cycle phase, pregnancy, the menopausal transition, hormone therapy, the effect of shorter stature on wave-reflection timing, or the higher prevalence of Raynaud’s phenomenon. CVA also cannot register women-specific risk enhancers such as pre-eclampsia or premature menopause. In a Rule 433 filing, Oura’s Chief Executive described the finger as “the first stop” of the pulse and claimed the ring’s accuracy allows it “to be right” about its predictions. The finger is the most distal site of the arterial tree, and no such accuracy has been demonstrated for CVA.

Conclusions. CVA is an unvalidated, PPG-derived age-prediction residual. It is presented to consumers in the language of arterial ageing and cardiovascular risk. It should not be used to reassure or alarm patients, and least of all women in midlife, whose cardiovascular risk rises as their stiffness trajectory changes. Risk should continue to be assessed with validated risk equations, blood pressure, lipids and women-specific history, and cfPWV where arterial stiffness itself is the question.

Keywords: arterial stiffness; pulse wave velocity; vascular age; cardiovascular age; photoplethysmography; wearable devices; women’s cardiovascular health; menopause; device validation; Oura Ring

1. Introduction

The idea of a “vascular age” is attractive and long established. In epidemiology it has been used chiefly as a communication device. It translates a multivariable risk estimate into the age of a person with the same risk and optimal risk factors.^1,2 Arterial stiffness offers a second, physiological route to the same idea. Large elastic arteries stiffen with age. Aortic stiffness, measured as cfPWV, predicts cardiovascular events and mortality independently of conventional risk factors and improves risk reclassification.^3–5

In May 2024 Oura Health Oy introduced Cardiovascular Age. The feature is derived from “age-related observations within a photoplethysmograph (PPG) signal, which carries information about estimated arterial stiffness and pulse wave velocity” and classifies users as trending below, above or within five years of their chronological age.⁶ Oura states that CVA “estimates the stiffening of large arteries, which has direct effects on the heart and also other major organs including the brain and kidneys.”⁶ It also states that the feature provides “an estimate of how your heart and large arteries are aging compared to your chronological age.”⁷ The company’s women’s heart-health material promotes its heart features to women directly, alongside the observation that “women experience more cardiovascular events during and after menopause.”⁸

Women are the population in which such a metric matters most and has been studied least. Cardiovascular disease is the leading cause of death in women and is under-recognised in them.⁹ Their arterial stiffness accelerates around the menopause.¹⁰ Their risk is modified by reproductive history that no sensor can detect.^11,12 This paper sets out the standard against which a vascular-age claim should be judged and examines the physiology of the finger PPG waveform. It then appraises Oura’s evidence and claims, including those made on the offering record of its initial public offering, with particular reference to women.

2. Measuring arterial stiffness and vascular age

2.1 The reference standard

Carotid–femoral pulse wave velocity is the accepted non-invasive reference measure of aortic stiffness.^13,14 It is the distance travelled by the pulse between the carotid and femoral arteries divided by the transit time, measured with applanation tonometry, mechanotransducers or Doppler. The expert consensus threshold for target-organ damage is 10 m/s. Path-length conventions must be standardised, because a few centimetres of error in surface distance alter the result materially.¹³ Reference values rise steeply with age and blood pressure, from about 6 m/s in young normotensive adults to more than 10 m/s in the elderly.¹⁵ The prognostic gradient is substantial. In a meta-analysis, each 1 m/s increase in aortic PWV was associated with a 14% higher risk of cardiovascular events and a 15% higher risk of cardiovascular and all-cause mortality.³ An individual-participant meta-analysis of 17,635 subjects confirmed that aortic PWV improves event prediction beyond standard risk factors.⁴

2.2 How a device is shown to estimate PWV

The ARTERY Society guideline sets the minimum requirements for validating a non-invasive PWV device.¹⁶ The device must be compared with a reference method in an adequate and representative sample, with prescribed age and blood-pressure distributions. Agreement must be expressed as the mean difference and its standard deviation (Bland–Altman analysis), and accuracy graded against fixed thresholds. A correlation coefficient does not satisfy this requirement. Correlation measures association across individuals, not agreement within them. It is inflated by the range of the sample, and a sample spanning young adults to the elderly will show high correlation with almost any age-sensitive variable.¹⁴

2.3 What “vascular age” can mean

The term is used in at least three senses, and the distinctions matter (Table 1).² A risk-equivalent vascular age, such as Framingham “heart age,” is calculated from risk factors and inherits the predictive validity of the risk equation.¹ A stiffness-equivalent vascular age maps a measured cfPWV onto the age at which that value is typical in a healthy reference population.¹⁵ A predicted-age vascular age is the output of a model trained to estimate chronological age from a physiological signal. The difference between predicted and actual age is then interpreted as accelerated or decelerated ageing. Oura’s CVA belongs to the third class, although its marketing borrows the language of the second.

Table 1. Three meanings of “vascular age” and the evidence each requires.

3. What a finger PPG waveform represents

3.1 The signal and its age-related features

The finger PPG pulse is a volumetric optical signal from the digital arteries and arterioles. Its contour changes with age in recognisable ways. The diastolic inflection becomes less distinct, the reflected wave arrives earlier, and indices derived from the second derivative of the waveform shift. Millasseau and colleagues showed that a stiffness index can be derived from the digital volume pulse, calculated as subject height divided by the interval between the systolic and diastolic peaks. That index correlates with large-artery stiffness and increases with age.¹⁸ Takazawa and colleagues showed that an ageing index from the second derivative of the PPG correlates with age.¹⁹ The age-related information in the finger PPG is therefore real. The question is what else that information contains.

3.2 The finger is the end of the arterial tree, not “the first stop”

Aortic stiffness, the quantity with prognostic value, is a property of the large elastic arteries. The digital arteries are small muscular vessels at the extreme periphery. In a stiffness gradient that runs from aorta to periphery, they are among the least affected by age-related elastic degeneration and the most affected by vasomotor tone. Age-related change in the finger waveform is transmitted mainly through the timing and magnitude of reflected waves. Those waves depend on central stiffness but also on the distance to reflection sites, and hence on body height, as well as on peripheral resistance and heart rate.^18,20 The pulse arrives at the wrist before it reaches the finger. The finger is the last accessible point on the arm, not the first.

3.3 State dependence: tone is not structure

Takazawa’s study is instructive for a second reason. The same second-derivative ageing index that correlated with age shifted acutely when vasoactive drugs were administered. Nitroglycerin made the waveform look younger and angiotensin made it look older.¹⁹ Structural arterial stiffness does not change within minutes, but the waveform does. A finger-PPG ageing metric is therefore sensitive to vasomotor state: ambient and skin temperature, sympathetic activity, sleep stage, alcohol, caffeine, hydration, medication and the phase of the menstrual cycle. Oura acknowledges that “cold hands” and loose fit degrade the signal.²¹ Averaging over 14 or more nights reduces random night-to-night noise. It does not remove systematic effects such as seasonal temperature, a change of medication or a change of hormonal state, all of which shift the baseline.

3.4 Rapid change is implausible as arterial remodelling

Oura’s own material illustrates the problem. A member reports that, after seven weeks of exercise, her CVA fell from +7.5 to +3 years and her estimated PWV from 7.6 to 7.3 m/s.²² A reduction of 0.3 m/s is small relative to the short-term variability of cfPWV measurement itself.¹⁴ Converting it into 4.5 years of “younger” arteries implies a scaling of about 0.07 m/s per year, which magnifies a small signal change into a large change in apparent age. Aerobic training can reduce arterial stiffness. A change of this size over seven weeks, however, is more parsimoniously explained by altered vasomotor tone, resting heart rate or measurement variance than by structural arterial rejuvenation. Presenting it as the latter is a claim the evidence does not support.

4. Appraisal of Oura’s evidence

4.1 The company white paper

Oura’s white paper reports a model based on pulse wave analysis of the high-resolution PPG signal. In 99 participants (36 men and 63 women; mean age 46.6 ± 16.6 years), the PWV estimated from that model correlated with cfPWV at r = 0.83.²¹ The paper reports no mean difference, no standard deviation of differences, no mean absolute error and no Bland–Altman plot, so it cannot be graded against the ARTERY criteria.¹⁶ It reports no results by sex, age band, blood pressure or skin tone. It does not state whether the participants who generated the correlation were also those on whom the model was trained. The algorithm “takes into account the impact of individual demographics” and requires the user’s sex assigned at birth.^21,23 If age and sex enter the model, part of the correlation with cfPWV, which itself rises strongly with age, may be supplied by the demographic inputs rather than by the waveform. The paper does not report how much of the predictive value remains once age is accounted for.

The correlation can be put in context with an illustrative calculation. If cfPWV in a sample of this age range has a standard deviation of about 2 m/s, r = 0.83 implies a residual standard deviation of about 1.1 m/s (2.0 × √(1 − 0.83²)). That residual is equivalent to roughly a decade of typical age-related stiffening. By the meta-analytic gradient, it also corresponds to a difference of about 14% in cardiovascular event risk.³ A correlation that sounds strong is thus compatible with individual errors of a size that would move a person across Oura’s categories. The assumption of a 2 m/s standard deviation is ours, and the actual figure would be determined by publishing the agreement statistics.

Oura further states that its science team conducted “clinical studies with 600 participants in collaboration with external partners, such as the Kuopio Research Institute of Exercise Medicine (KuLTu) and UCLA.”⁷ We identified no peer-reviewed publication reporting CVA or PWV agreement from these studies.

4.2 The National University of Singapore study

Oura promotes a 2026 study from the National University of Singapore as research that “validates Oura’s vascular age estimation.”²⁴ The paper itself is more limited.²⁵ Researchers recorded overnight PPG from an Oura Ring and a fingertip sensor during polysomnography in 160 adults (78 men; median age 31 years; range 20–70). They built their own models to predict chronological age, using ten-fold cross-validation. The ring-derived model achieved a mean absolute error of 7.25 years (r = 0.80; R² = 0.59). The fingertip model achieved 6.28 years. The age gap (ΔAge) was associated with systolic blood pressure. The authors note a predominance of younger participants, single-night data for most, and proportional bias, with age underestimated in older adults. No results are reported by sex. The principal investigator serves on Oura’s medical advisory board and two authors belong to the Oura–NUS Joint Lab, although the authors state that the work was conducted independently of Oura.²⁵

Four features of this study limit what it can validate. First, it did not test Oura’s CVA algorithm. It tested the researchers’ own models applied to ring PPG. Second, its reference was chronological age, not cfPWV, so it shows that the ring signal contains age information but not that the residual reflects arterial stiffness. Third, its error exceeds the width of Oura’s categories (Section 4.3). Fourth, its population was young and healthy, and the age range in which cardiovascular risk and the menopausal transition converge was under-represented.

4.3 The categories are narrower than the error

Oura classifies users as “aligned” when CVA is within five years of chronological age, and as “below” or “above” when it differs by six years or more.²³ The published ring-PPG error is wider than this band. If prediction errors are approximately normal with a mean absolute error of 7.25 years, their standard deviation is about 9.1 years. Only about 42% of predictions would then fall within ±5 years of the true value. A person whose arteries are exactly typical for her age would be classified as “above” or “below” about half the time on error alone (Table 2). These figures derive from an academic model rather than Oura’s own, which has published no error metric. They nevertheless show the scale of uncertainty attached to a three-category output whose central band is narrower than the instrument’s error.

Table 2. Illustrative classification consequences of the published ring-PPG age-prediction error (MAE 7.25 years), assuming normally distributed errors.²⁵

Calculations: SD = 7.25 × 1.2533; P(|e| ≤ 5) = 2Φ(5/9.09) − 1; P(|e| ≥ 6) = 2[1 − Φ(6/9.09)]. Illustrative; actual error distributions were not published.

4.4 Regression to the mean in predicted-age metrics

Any model trained to predict chronological age with imperfect accuracy regresses toward the mean age of its training sample. Its predictions are compressed, so the age gap is negatively correlated with true age. This is well characterised for “brain age” from neuroimaging, where uncorrected age gaps systematically make the young appear older and the old younger.¹⁷ The NUS authors observed exactly this proportional bias.²⁵ For a least-squares model with R² = 0.59 and a training mean of about 40 years, the expected age gap from this artefact alone is about −0.41 × (age − 40). That is roughly +6 years at age 25, −4 years at 50 and −12 years at 70, before any arterial difference is considered. Oura reports that “the average CVA among Oura members is 1.76 years younger than their chronological age.”²² That figure is compatible with a healthier-than-average user base, with this bias, or with both. Oura has not reported whether CVA is corrected for it. If it is not, an older woman may be reassured and a younger one alarmed for reasons unrelated to her arteries.

5. Particular considerations in women

Oura addresses women’s cardiovascular health directly in its marketing.⁸ Its white paper enrolled more women than men.²¹ Yet neither the white paper nor the NUS study reports a single result by sex. The omission matters because female physiology bears on every step from waveform to interpretation (Table 3).

Table 3. Women-specific factors affecting a finger-PPG estimate of vascular age, and their treatment in Oura’s published evidence.

Three consequences follow. First, the period of fastest arterial change in a woman’s life, the year around the final menstrual period, is the period in which a cross-sectional model trained on mixed ages is least likely to track her. That is because the model encodes an average trajectory, not a sex- and stage-specific one.¹⁰ Second, a midlife woman shown a “below” or “aligned” CVA may take reassurance at the moment her risk is rising. The regression-to-the-mean bias that flatters older users compounds this. Third, the most important determinants of a woman’s cardiovascular risk beyond conventional factors are historical: pregnancy complications, age at menopause and autoimmune disease. No waveform can register them, so a metric presented as a summary of cardiovascular ageing omits them by construction.¹¹

6. The Chief Executive’s statements on the offering record

On 24 September 2026, Oura Inc. filed with the SEC an issuer free writing prospectus under Rule 433 (Registration Statement No. 333-298734; Accession No. 0001193125-26-401224). Its Exhibit A is the transcript of Robinhood’s Chief Executive interviewing Oura’s Chief Executive, Thomas Hale.²⁹ A free writing prospectus may not conflict with the registration statement.³⁰ Under FDA regulation, the oral and written statements of a manufacturer’s representatives are evidence of a product’s intended use.³¹ The interview does not name Cardiovascular Age. It does, however, make statements about the physiological basis and accuracy of the ring’s cardiovascular measurements, and about ageing as a use case, that bear directly on CVA (Table 4).

Table 4. Statements in Exhibit A to Oura’s Rule 433 filing of 24 September 2026 bearing on Cardiovascular Age, quoted verbatim, with comment.²⁹

Outside the filing, Hale has been reported as describing the prospect of telling a user: “In 10 years, if you don’t change your course, you’re going to have a heart attack.”³² That statement is not part of the SEC record, and we cite it only from a secondary summary. It nevertheless illustrates the destination toward which ageing and cardiovascular metrics are being directed: individual event prediction. No consumer PPG metric has been shown to perform this, and a validated risk equation already does it better.^1,11

Oura’s CVA documentation states that the ring “is not a medical device and is not intended to diagnose, treat, cure, monitor, or prevent medical conditions.”²³ The FDA’s January 2026 general wellness policy allows non-invasive physiological estimates to be offered for wellness provided that they do not imply clinical use.³³ We express no legal conclusion. The scientific point is that statements describing arterial stiffness, damage to “the brain and kidneys,” risk of “heart disease, stroke, and heart failure,” and predictions that are “right” are claims about clinical measurement and risk.^6,7,29 They require the evidence appropriate to such claims.

7. Clinical implications

False reassurance. A woman of 55 with hypertension and a history of pre-eclampsia who is told her cardiovascular age is “below” her chronological age has learned nothing about her risk. She may nonetheless conclude otherwise.

False alarm. A healthy 25-year-old classified as “above” may be responding to model bias, a cold bedroom or a change in contraception rather than to her arteries.

Spurious tracking. Short-term changes in CVA attributed to lifestyle change may reflect vasomotor tone and measurement variance rather than arterial remodelling (Section 3.4). Over-reading them risks misplaced confidence in an intervention, or in the absence of one.

Displacement of validated assessment. Risk should be estimated with validated equations, measured blood pressure, lipids, glycaemia, smoking and women-specific history.¹¹ Arterial stiffness, where clinically relevant, should be measured by cfPWV according to consensus methods.¹³

8. Questions the manufacturer should answer

Bland–Altman agreement between CVA-derived PWV and cfPWV, graded against ARTERY criteria, in an independent test set.

Results by sex, age band, blood pressure category and skin tone,³⁴ and for women by menopausal status and use of hormone therapy or hormonal contraception.

How much of the model’s association with cfPWV remains after accounting for age and sex as inputs.

Whether CVA is corrected for regression-to-the-mean bias, and the method of correction.

Test–retest reproducibility of CVA over weeks in stable individuals, and its sensitivity to ambient temperature, menstrual phase and vasoactive medication.

Prospective evidence that CVA predicts cardiovascular events independently of established risk factors, in women as well as men.

Published results of the stated 600-participant studies with KuLTu and UCLA.

9. Recommendations

9.1 For clinicians

Do not interpret Oura CVA as a measure of arterial stiffness or as a cardiovascular risk estimate.

Assess risk with validated tools, including women-specific risk enhancers, irrespective of a patient’s CVA.¹¹

Where arterial stiffness is clinically relevant, measure cfPWV by consensus methods.¹³

Counsel midlife women that rising cardiovascular risk around the menopause is not reflected reliably by consumer metrics.¹⁰

9.2 For researchers and the manufacturer

Validate against cfPWV per ARTERY criteria, with sex-stratified reporting as a minimum.¹⁶

Correct predicted-age residuals for regression to the mean before presenting them to individuals.¹⁷

Study women across the menstrual cycle, pregnancy, the menopausal transition and hormone therapy.

Test incremental prognostic value over validated risk equations before presenting CVA as a summary of cardiovascular ageing.

10. Conclusion

Arterial stiffness is a powerful predictor of cardiovascular events, and its measurement is well standardised. Oura’s Cardiovascular Age borrows that standing without meeting that standard. It is an age-prediction residual derived from the most distal, most vasomotor-sensitive segment of the arterial tree. Its only reference-standard evidence is an unpublished correlation without agreement statistics. The best published error for a ring-PPG age model is wider than the categories presented to users, and no result has been reported for women as a group. For women in particular, the metric is silent on the years of fastest arterial change and blind to the reproductive history that most distinguishes their risk. The statements made to investors in Oura’s SEC filing, that the finger is the pulse’s “first stop” and that the ring’s accuracy lets it “be right,” are not supported by the physiology or by the published evidence. Until CVA is validated against cfPWV with sex-stratified agreement statistics, and shown to add prognostic information, it should be regarded as a wellness number and not as a measure of how a woman’s, or anyone’s, arteries are ageing.

Declarations

Author contributions: Bomi Joseph: Conceptualization, article conception and design, development of the manuscript outline, and writing of the manuscript.

Darshana Goakar: Research and investigation, source identification, collection of supporting documents, and compilation of the reference literature.

Both authors reviewed and approved the final manuscript.

Funding: This work was supported by Peak Health Center, San Francisco, California, USA (ROR: https://ror.org/04b23x642).

Competing interests: Bomi Joseph is the inventor of the Deep Health® Device, a health-measurement technology.

Darshana Goakar declares no competing interests.

Data availability

No new data were generated or analyzed specifically for this article. All data, figures, performance measures, and other quantitative information discussed in the manuscript derive from the publicly available sources cited in the article.

Use of artificial-intelligence tools

No artificial-intelligence tools were used in the research, analysis, preparation, or writing of this manuscript.

Note on sources

Descriptions of Oura products and features, and performance figures attributed to Oura, are derived from company publications, press releases, and other publicly available Oura materials current as of September 2026. Unless otherwise stated in the manuscript, these company- reported claims and figures have not been independently verified by the authors.

References

1. D’Agostino RB Sr, Vasan RS, Pencina MJ, et al. General cardiovascular risk profile for use in primary care: the Framingham Heart Study. Circulation. 2008;117(6):743–753.

2. Groenewegen KA, den Ruijter HM, Pasterkamp G, Polak JF, Bots ML, Peters SAE. Vascular age to determine cardiovascular disease risk: a systematic review of its concepts, definitions, and clinical applications. Eur J Prev Cardiol. 2016;23(3):264–274.

3. Vlachopoulos C, Aznaouridis K, Stefanadis C. Prediction of cardiovascular events and all-cause mortality with arterial stiffness: a systematic review and meta-analysis. J Am Coll Cardiol. 2010;55(13):1318–1327.

4. Ben-Shlomo Y, Spears M, Boustred C, et al. Aortic pulse wave velocity improves cardiovascular event prediction: an individual participant meta-analysis of prospective observational data from 17,635 subjects. J Am Coll Cardiol. 2014;63(7):636–646.

5. Mitchell GF, Hwang SJ, Vasan RS, et al. Arterial stiffness and cardiovascular events: the Framingham Heart Study. Circulation. 2010;121(4):505–511.

6. Oura Health Oy. ŌURA announces two new heart health features, moving deeper into preventative health [press release]. Business Wire; 10 May 2024. Available from: https://www.businesswire.com/news/home/20240510769393/en/

7. Oura Health Oy. All about Cardiovascular Age. The Pulse Blog; 1 April 2025 (modified 28 April 2026). Available from: https://ouraring.com/blog/cardiovascular-age/

8. Oura Health Oy. Heart health for women: what you should know. The Pulse Blog; 28 May 2024 (modified 4 November 2024). Available from: https://ouraring.com/blog/heart-health-for-women/

9. Garcia M, Mulvagh SL, Merz CN, Buring JE, Manson JE. Cardiovascular disease in women: clinical perspectives. Circ Res. 2016;118(8):1273–1293.

10. Samargandy S, Matthews KA, Brooks MM, et al. Arterial stiffness accelerates within 1 year of the final menstrual period: the SWAN Heart Study. Arterioscler Thromb Vasc Biol. 2020;40(4):1001–1008.

11. Arnett DK, Blumenthal RS, Albert MA, et al. 2019 ACC/AHA guideline on the primary prevention of cardiovascular disease. Circulation. 2019;140(11):e596–e646.

12. Bellamy L, Casas JP, Hingorani AD, Williams DJ. Pre-eclampsia and risk of cardiovascular disease and cancer in later life: systematic review and meta-analysis. BMJ. 2007;335(7627):974.

13. Van Bortel LM, Laurent S, Boutouyrie P, et al. Expert consensus document on the measurement of aortic stiffness in daily practice using carotid-femoral pulse wave velocity. J Hypertens. 2012;30(3):445–448.

14. Townsend RR, Wilkinson IB, Schiffrin EL, et al. Recommendations for improving and standardizing vascular research on arterial stiffness: a scientific statement from the American Heart Association. Hypertension. 2015;66(3):698–722.

15. Reference Values for Arterial Stiffness’ Collaboration. Determinants of pulse wave velocity in healthy people and in the presence of cardiovascular risk factors: ‘establishing normal and reference values’. Eur Heart J. 2010;31(19):2338–2350.

16. Wilkinson IB, McEniery CM, Schillaci G, et al. ARTERY Society guidelines for validation of non-invasive haemodynamic measurement devices: Part 1, arterial pulse wave velocity. Artery Res. 2010;4(2):34–40.

17. Smith SM, Vidaurre D, Alfaro-Almagro F, Nichols TE, Miller KL. Estimation of brain age delta from brain imaging. NeuroImage. 2019;200:528–539.

18. Millasseau SC, Kelly RP, Ritter JM, Chowienczyk PJ. Determination of age-related increases in large artery stiffness by digital pulse contour analysis. Clin Sci (Lond). 2002;103(4):371–377.

19. Takazawa K, Tanaka N, Fujita M, et al. Assessment of vasoactive agents and vascular aging by the second derivative of photoplethysmogram waveform. Hypertension. 1998;32(2):365–370.

20. Smulyan H, Marchais SJ, Pannier B, Guerin AP, Safar ME, London GM. Influence of body height on pulsatile arterial hemodynamic data. J Am Coll Cardiol. 1998;31(5):1103–1109.

21. Ohukainen P, Rantanen A, Clausel A. Cardiovascular Age unlocks new dimensions of health beyond heart rate [white paper]. Oura Health Oy; 2024. Available from: https://ouraring.com/blog/wp-content/uploads/2024/05/FINAL_Oura_Cardiovascular_Age_White_Paper.pdf

22. Oura Health Oy. How to lower your Cardiovascular Age. The Pulse Blog; 31 January 2025 (modified 11 June 2025). Available from: https://ouraring.com/blog/how-to-lower-cardiovascular-age/

23. Oura Health Oy. Cardiovascular Age [support article]. Available from: https://support.ouraring.com/hc/en-us/articles/28451491040019-Cardiovascular-Age

24. Oura Health Oy. New NUS research validates Oura’s vascular age estimation, a key indicator of cardiovascular health. The Pulse Blog; April 2026. Available from: https://ouraring.com/blog/new-nus-research-validates-ouras-vascular-age-estimation/

25. Yilmaz G, Ghorbani S, Ong JL, Aghayan Golkashani H, Zhang C, Yeo BTT, Chee MWL. Vascular age estimation using a consumer wearable sleep tracker. PLOS Digit Health. 2026;5(3):e0001329. doi:10.1371/journal.pdig.0001329

26. Robb AO, Mills NL, Din JN, et al. Influence of the menstrual cycle, pregnancy, and preeclampsia on arterial stiffness. Hypertension. 2009;53(6):952–958.

27. Hayward CS, Kelly RP. Gender-related differences in the central arterial pressure waveform. J Am Coll Cardiol. 1997;30(7):1863–1871.

28. Garner R, Kumari R, Lanyon P, Doherty M, Zhang W. Prevalence, risk factors and associations of primary Raynaud’s phenomenon: systematic review and meta-analysis of observational studies. BMJ Open. 2015;5(3):e006389.

29. Oura Inc. Issuer free writing prospectus filed pursuant to Rule 433, dated 24 September 2026, relating to preliminary prospectus dated 21 September 2026. Registration Statement No. 333-298734; SEC Accession No. 0001193125-26-401224. Exhibit A: transcript of Vlad Tenev’s interview of Thomas Hale, Chief Executive Officer.

30. US Securities and Exchange Commission. Conditions to permissible post-filing free writing prospectuses. 17 CFR §230.433.

31. US Food and Drug Administration. Meaning of intended uses. 21 CFR §801.4.

32. AI vs. doctors: Oura CEO on health data [interview summary of Pioneers of AI episode]. StartupHub.ai; 2026. Available from: https://www.startuphub.ai/ai-news/artificial-intelligence/2026/ai-vs-doctors-oura-ceo-on-health-data

33. US Food and Drug Administration. General wellness: policy for low risk devices. Guidance for industry and FDA staff. 6 January 2026.

34. Bent B, Goldstein BA, Kibbe WA, Dunn JP. Investigating sources of inaccuracy in wearable optical heart rate sensors. NPJ Digit Med. 2020;3:18.