Skip to content

Risk and outcomes

TL;DR — The exposure–outcome relationship is continuous, log-linear, and detectable to at least 115/75 mm Hg: in one million adults across 61 prospective studies, each 20 mm Hg higher usual systolic pressure at ages 40–69 was associated with more than a doubling of stroke and ischaemic heart disease mortality, with proportional effects halving by ages 80–89 while absolute differences grow (Lewington 2002, PMID 12493255). Randomised evidence matches: 5 mm Hg systolic reduction lowered major cardiovascular events by ~10% regardless of baseline pressure category or prior disease across 344,716 participants (BPLTTC 2021, PMID 33933205). Three things complicate the picture rather than overturn it. First, mean pressure is not the whole exposure — visit-to-visit variability and maximum systolic pressure predict stroke independently of the mean, with top-decile hazard ratios of 6.22 (95% CI 4.16–9.29) and 15.01 (6.56–34.38) respectively in the UK-TIA cohort (Rothwell 2010, PMID 20226988). Second, the J-curve is not entirely an artefact: low diastolic pressure is associated with subclinical myocardial damage and incident coronary events, particularly when systolic pressure is ≥120 mm Hg (McEvoy 2016, PMID 27590090). Third, exposure in young adulthood carries independent risk decades later (Zhang 2019, PMID 31319915).

The observational dose–response

Finding Estimate Source
Stroke and IHD death per 20 mm Hg usual systolic, ages 40–69 >2-fold each (Lewington 2002, PMID 12493255)
Threshold below which association disappears None detectable to 115/75 mm Hg (Lewington 2002, PMID 12493255)
Proportional effect at ages 80–89 vs 40–49 About half as extreme; absolute risk difference larger (Lewington 2002, PMID 12493255)
Peripheral arterial disease per 20 mm Hg usual systolic HR 1.63 (95% CI 1.59–1.66) in 4.2 million adults (Emdin 2015, PMID 26419648)
Atrial fibrillation per 20 mm Hg usual systolic HR 1.21 (1.19–1.22) overall; 1.91 (1.75–2.09) at ages 30–40 falling to 1.01 (0.97–1.04) at 80–90, in 4.3 million adults (Emdin 2017, PMID 27143136)
Pulse pressure vs mean of systolic and diastolic for predicting vascular mortality Mean is slightly more informative; pulse pressure much less (Lewington 2002, PMID 12493255)

The age-dependence of the atrial-fibrillation association is instructive: an exposure whose relative effect vanishes by the ninth decade is not thereby unimportant, because absolute event rates are highest there — the same pattern Lewington described for vascular mortality (Emdin 2017, PMID 27143136; Lewington 2002, PMID 12493255). General syntheses of the hypertension–cardiovascular disease relationship set out the full range of downstream conditions (Fuchs 2020, PMID 31865786).

What the randomised evidence adds

Observational dose–response could in principle reflect confounding or reverse causation. Randomised data settle direction and rough magnitude:

  • BPLTTC. 48 trials, 344,716 participants, median 4.15 years. Per 5 mm Hg systolic reduction: HR 0.91 (95% CI 0.89–0.94) in people without prior cardiovascular disease and 0.89 (0.86–0.92) in those with; no reliable heterogeneity across seven baseline systolic strata from <120 to ≥170 mm Hg (BPLTTC 2021, PMID 33933205). The clinical corollary the authors draw is that indication should be framed by cardiovascular risk, not by the pressure number.
  • Mendelian randomisation. Genetically predicted systolic pressure is associated with valvular heart disease (Nazarzadeh 2019, PMID 31290937), supporting causality for an outcome where observational confounding is plausible. A multivariable analysis within a 1.3-million-participant rare-variant study reported a surprising possible inverse effect of blood pressure on large-artery stroke, which the authors flagged rather than suppressed (Surendran 2020, PMID 33230300).

Lifetime and cumulative exposure

Residual lifetime risk of developing hypertension from age 55 or 65 is about 90%, and the lifetime probability of taking an antihypertensive is 60% (Vasan 2002, PMID 11866648). More usefully for prevention, exposure accumulated in young adulthood matters independently. Pooling six US cohorts (36,030 participants, median 17 years of follow-up, 4,570 coronary events, 5,119 heart-failure events, 2,862 strokes) with imputed risk-factor trajectories from age 18, systolic pressure ≥130 mm Hg during ages 18–39 was associated with a 37% higher later-life heart-failure risk and diastolic ≥80 mm Hg with 21%, independent of exposure after age 40 (Zhang 2019, PMID 31319915). Long-term visit-to-visit variability across 30 years likewise predicts subclinical kidney damage and albuminuria (Wang 2022, PMID 35360932). Blood pressure tracks from childhood with tracking coefficients around 0.38 systolic (Chen 2008, PMID 18559702), and reversion to normal pressure becomes unlikely after adolescence (Meng 2025, PMID 39495520).

Variability as a separate exposure

Rothwell's 2010 analysis reframed the exposure. In patients with previous transient ischaemic attack, top-decile visit-to-visit standard deviation of systolic pressure over seven visits carried HR 6.22 (95% CI 4.16–9.29) for stroke, rising to 12.08 (7.40–19.72) with ten visits — the dependence on measurement precision being itself evidence that the signal is real rather than noise. Maximum systolic pressure reached carried HR 15.01 (6.56–34.38) after adjustment for mean. In ASCOT-BPLA, residual on-treatment variability predicted stroke (top-decile HR 3.25, 2.32–4.54) and coronary events independently of mean clinic or ambulatory pressure; ambulatory variability was a weaker predictor, and all variability measures were most predictive in younger patients and at below-median mean pressure (Rothwell 2010, PMID 20226988).

Subsequent work is mixed but broadly supportive: variability predicted macrovascular and microvascular complications in ADVANCE and ADVANCE-ON (Hata 2013, PMID 23926207; Ohkuma 2017, PMID 28584014), and predicted events across baseline-risk strata in a large hypertension trial cohort (Mehlum 2018, PMID 29365085). Drug classes differ in the variability they produce — calcium channel blockers and diuretics reduce it, and this has been proposed as part of why they outperform on stroke (Webb 2011, PMID 21817143; Mehlum 2020, PMID 32336236; Webb 2010, PMID 20651263; Muntner 2014, PMID 24739073). Beat-to-beat variability after TIA and stroke has its own prognostic signal (Webb 2018, PMID 29229726), and genome-wide analysis has looked for its determinants (Yadav 2013, PMID 23929743). Whether variability is a modifiable target or a marker of arterial stiffness and adherence is unresolved. In SPRINT, higher visit-to-visit variability was associated with dementia and cognitive decline in post-hoc analyses (de Havenon 2021, PMID 34533059; Sible 2023, PMID 36448621; Yu 2023, PMID 37170806) — see vascular-dementia for the cognitive endpoint.

The J-curve

The J-curve argument is that below some pressure, events rise again — either because low pressure compromises coronary perfusion in diastole, or because sick people have low pressure (reverse causation).

  • Evidence for a mechanism, not just confounding. In 11,565 ARIC participants over 21 years, baseline diastolic pressure <60 mm Hg carried adjusted odds of 2.2 for high-sensitivity troponin-T ≥14 ng/L versus diastolic 80–89 mm Hg, and 60–69 mm Hg carried 1.5; low diastolic pressure predicted progressive troponin rise over six years and incident coronary heart disease and mortality but not stroke; the association was strongest with elevated baseline troponin (p for interaction <0.001) and most pronounced when systolic pressure was ≥120 mm Hg — i.e. at high pulse pressure (McEvoy 2016, PMID 27590090).
  • Evidence for confounding. J-shaped associations appear in observational registries (Staplin 2023, PMID 37156250) but not in the randomised comparison of targets, where the intensive arm did better despite lower achieved diastolic pressure (SPRINT 2021, PMID 34010531). Reviews set out both readings (Messerli 2009, PMID 19892233; Rahman 2017, PMID 28612327; Filippone 2021, PMID 38559601; Alderman 1996, PMID 8737854). Mechanistic work implicates subclinical myocardial injury and immune activation (Topel 2019, PMID 30217423), and registry data in revascularised diabetic patients show the pattern (Kai 2020, PMID 32015482; Kim 2024, PMID 39218965).
  • The practical residue. McEvoy's suggestion — when treating to systolic <140 mm Hg, avoid letting diastolic fall below 70 and particularly below 60 mm Hg — is a reasonable reading of the evidence; the 2026-09-01 PubMed and ClinicalTrials.gov searches located secondary analyses but no trial that randomised a diastolic floor (McEvoy 2016, PMID 27590090). In SPRINT-MIND analyses, low achieved diastolic pressure with systolic below 130 mm Hg was examined against cognitive outcomes with broadly reassuring results (Jiang 2023, PMID 36688305; Yang 2025, PMID 41070444).

Risk-based rather than pressure-based treatment

Because relative risk reduction is roughly constant per 5 mm Hg while absolute benefit scales with baseline risk, the same pressure reduction is worth much more in a high-risk person (BPLTTC 2021, PMID 33933205). This logic underlies the 2017/2025 US and 2024 ESC decision to make treatment thresholds partly risk-conditional rather than purely pressure-conditional — see blood-pressure targets and guidelines. Sustained control is associated with slower multimorbidity progression in older adults (Bowling 2020, PMID 32501546), and apparent treatment resistance combined with poor adherence carries a distinctly worse prognosis (Lee 2025, PMID 39543414).

Downstream conditions covered elsewhere

Outcome Where it lives
Stroke, its subtypes and secondary prevention stroke
Left ventricular hypertrophy, HFpEF, hypertensive cardiomyopathy hypertensive-heart-disease
Coronary artery disease ischemic-heart-disease
Cognitive decline and dementia prevention vascular-dementia
Diabetes co-management type-2-diabetes

Open questions

  • Is visit-to-visit variability a modifiable target, or a marker of stiffness and adherence? The 2026-09-01 searches located trials that changed variability and post-hoc outcome analyses, but no trial that deliberately reduced visit-to-visit variability and used a hard cardiovascular endpoint (Rothwell 2010, PMID 20226988; Webb 2011, PMID 21817143).
  • Does the diastolic floor McEvoy proposes (>70 mm Hg, certainly >60) change outcomes if applied prospectively, or would it simply undertreat systolic pressure? (McEvoy 2016, PMID 27590090; SPRINT 2021, PMID 34010531)
  • Given independent effects of ages-18–39 exposure on later heart failure, is there a case for treating stage 1 hypertension in young adults on lifetime-risk grounds when 10-year risk is negligible? The 2026-09-01 searches located lifestyle, mechanistic and special-cause trials in young adults, but no event-powered trial of pharmacological treatment selected for uncomplicated stage 1 hypertension at ages 18–39 (Zhang 2019, PMID 31319915).
  • Why does the blood pressure–atrial fibrillation association attenuate to null by age 80–90 while the absolute burden peaks there? (Emdin 2017, PMID 27143136)
  • Should relative-risk-constant, absolute-risk-scaled reasoning displace threshold-based treatment entirely, and what would guidelines look like if it did? (BPLTTC 2021, PMID 33933205)

References

  1. Lewington S, et al. Age-specific relevance of usual blood pressure to vascular mortality. Lancet. 2002;360:1903-13. PMID 12493255
  2. Blood Pressure Lowering Treatment Trialists' Collaboration. Pharmacological blood pressure lowering... individual participant-level data meta-analysis. Lancet. 2021;397:1625-1636. PMID 33933205
  3. Rothwell PM, et al. Prognostic significance of visit-to-visit variability, maximum systolic blood pressure, and episodic hypertension. Lancet. 2010;375:895-905. PMID 20226988
  4. McEvoy JW, et al. Diastolic Blood Pressure, Subclinical Myocardial Damage, and Cardiac Events. J Am Coll Cardiol. 2016;68:1713-1722. PMID 27590090
  5. Zhang Y, et al. Associations of Blood Pressure and Cholesterol Levels During Young Adulthood With Later Cardiovascular Events. J Am Coll Cardiol. 2019;74:330-341. PMID 31319915
  6. Emdin CA, et al. Usual blood pressure, peripheral arterial disease, and vascular risk: cohort study of 4.2 million adults. BMJ. 2015;351:h4865. PMID 26419648
  7. Emdin CA, et al. Usual blood pressure, atrial fibrillation and vascular risk: evidence from 4.3 million adults. Int J Epidemiol. 2017;46:162-172. PMID 27143136
  8. Fuchs FD, et al. High Blood Pressure and Cardiovascular Disease. Hypertension. 2020;75:285-292. PMID 31865786
  9. Nazarzadeh M, et al. Systolic Blood Pressure and Risk of Valvular Heart Disease: A Mendelian Randomization Study. JAMA Cardiol. 2019;4:788-795. PMID 31290937
  10. Surendran P, et al. Discovery of rare variants associated with blood pressure regulation. Nat Genet. 2020;52:1314-1332. PMID 33230300
  11. Vasan RS, et al. Residual lifetime risk for developing hypertension in middle-aged women and men. JAMA. 2002;287:1003-10. PMID 11866648
  12. Wang Y, et al. Associations of Long-Term Visit-to-Visit Blood Pressure Variability With Subclinical Kidney Damage and Albuminuria in Adulthood. Hypertension. 2022;79:1247-1256. PMID 35360932
  13. Chen X, et al. Tracking of blood pressure from childhood to adulthood. Circulation. 2008;117:3171-80. PMID 18559702
  14. Meng Y, et al. Tracking and Transition Probability of Blood Pressure From Childhood to Midadulthood. JAMA Pediatr. 2025;179:34-45. PMID 39495520
  15. Hata J, et al. Effects of visit-to-visit variability in systolic blood pressure... ADVANCE. Circulation. 2013;128:1325-34. PMID 23926207
  16. Ohkuma T, et al. Prognostic Value of Variability in Systolic Blood Pressure... ADVANCE-ON. Hypertension. 2017;70:461-468. PMID 28584014
  17. Mehlum MH, et al. Blood pressure variability and risk of cardiovascular events and death. Eur Heart J. 2018;39:2243-2251. PMID 29365085
  18. Webb AJ, et al. Effect of dose and combination of antihypertensives on interindividual blood pressure variability. Stroke. 2011;42:2860-5. PMID 21817143
  19. Mehlum MH, et al. Blood Pressure-Lowering Profiles and Clinical Effects of ARBs Versus CCBs. Hypertension. 2020;75:1584-1592. PMID 32336236
  20. Webb AJ, et al. Blood pressure variability and risk of new-onset atrial fibrillation. Stroke. 2010;41:2091-3. PMID 20651263
  21. Muntner P, et al. Effect of chlorthalidone, amlodipine, and lisinopril on visit-to-visit variability of blood pressure (ALLHAT). J Clin Hypertens. 2014;16:323-30. PMID 24739073
  22. Webb AJS, et al. Prognostic Significance of Blood Pressure Variability on Beat-to-Beat Monitoring After TIA and Stroke. Stroke. 2018;49:62-67. PMID 29229726
  23. Yadav S, et al. Genome-wide analysis of blood pressure variability and ischemic stroke. Stroke. 2013;44:2703-2709. PMID 23929743
  24. de Havenon A, et al. Increased Blood Pressure Variability and the Risk of Probable Dementia or MCI: SPRINT MIND. J Am Heart Assoc. 2021;10:e022206. PMID 34533059
  25. Sible IJ, et al. Blood Pressure Variability and Cognitive Decline: A Post Hoc Analysis of the SPRINT MIND Trial. Am J Hypertens. 2023;36:168-175. PMID 36448621
  26. Yu J, et al. Visit-to-Visit Blood Pressure Variability and Cardiovascular Outcomes in the STEP Trial. Hypertension. 2023;80:1507-1516. PMID 37170806
  27. Staplin N, et al. Relationship between clinic and ambulatory blood pressure and mortality. Lancet. 2023;401:2041-2050. PMID 37156250
  28. SPRINT Research Group. Final Report of a Trial of Intensive versus Standard Blood-Pressure Control. N Engl J Med. 2021;384:1921-1930. PMID 34010531
  29. Messerli FH, et al. The J-curve between blood pressure and coronary artery disease or essential hypertension. J Am Coll Cardiol. 2009;54:1827-34. PMID 19892233
  30. Rahman F, et al. The J-shaped Curve for Blood Pressure and Cardiovascular Disease Risk. Curr Atheroscler Rep. 2017;19:34. PMID 28612327
  31. Filippone EJ, et al. The diastolic blood pressure J-curve revisited: An update. Am Heart J Plus. 2021;12:100065. PMID 38559601
  32. Alderman MH. Blood pressure J-curve: is it cause or effect? Curr Opin Nephrol Hypertens. 1996;5:209-13. PMID 8737854
  33. Topel ML, et al. Mechanisms underlying the J-curve for diastolic blood pressure. Int J Cardiol. 2019;276:255-260. PMID 30217423
  34. Kai H, et al. Low blood pressure and cardiovascular events in diabetic patients with coronary artery disease after revascularization. Hypertens Res. 2020;43:715-723. PMID 32015482
  35. Kim BS, et al. Association of office blood pressure with ischemic and bleeding events in patients undergoing PCI. Sci Rep. 2024;14:20310. PMID 39218965
  36. Jiang C, et al. Diastolic Blood Pressure and Intensive Blood Pressure Control on Cognitive Outcomes: SPRINT MIND. Hypertension. 2023;80:580-589. PMID 36688305
  37. Yang R, et al. Diastolic Blood Pressure and Cognitive Function in Adults With Achieved Systolic Blood Pressure Below 130 mm Hg. Circ Cardiovasc Qual Outcomes. 2025;18:e011902. PMID 41070444
  38. Bowling CB, et al. Association of Sustained Blood Pressure Control with Multimorbidity Progression Among Older Adults. J Am Geriatr Soc. 2020;68:2059-2066. PMID 32501546
  39. Lee CJ, et al. Prognosis of apparent treatment-resistant hypertension and poor adherence: a nationwide cohort study. Hypertens Res. 2025;48:49-59. PMID 39543414