Blood pressure and cardiovascular disease: are diastolic values 80–89 mmHg enough to define hypertension?

This editorial refers to ‘A cohort study and meta-analysis of isolated diastolic hypertension: searching for a threshold to guide treatment’†, by A.P. Jacobsen et al., on page 2119. Simplified view of the meta-analysis of the association between isolated diastolic hypertension by the 2017 ACC/AHA definition and cardiovascular disease events in epidemiological studies published to date including >10 million participants. Studies with research-grade measurement of blood pressure (BP) highlighted in red show no association, and only studies in which office BP measurements were obtained by routine clinical care showed a modest association (highlighted in green). CI, confidence interval. Simplified view of the meta-analysis of the association between isolated diastolic hypertension by the 2017 ACC/AHA definition and cardiovascular disease events in epidemiological studies published to date including >10 million participants. Studies with research-grade measurement of blood pressure (BP) highlighted in red show no association, and only studies in which office BP measurements were obtained by routine clinical care showed a modest association (highlighted in green). CI, confidence interval. The continuous relationship between blood pressure (BP) and cardiovascular (CV) and renal events makes it difficult to distinguish between normal and high BP. In clinical practice, pragmatic BP cut-offs between normality and disease are needed to simplify the diagnosis, clinical evaluation, and treatment of subjects with disease. Epidemiological associations between BP and CV risk begin at very low BP levels [i.e. systolic BP (SBP) >115 mmHg and diastolic BP (DBP) >75 mmHg],1,2 but are not sufficient to establish disease-defining thresholds. Knowledge of the putative benefits of medical interventions in reducing silent arterial damage progression induced by high BP values in a given BP range and the benefits in reducing fatal and non-fatal CV events and all-cause mortality is crucial.3,4 Therefore, hypertension, as a disease, has historically been defined as the arbitrary BP level at which the benefits of treatment (lifestyle interventions or BP-lowering drugs) unequivocally outweigh the risks of treatment in clinical trials. After carefully reviewing the evidence, the 2018 European Society of Cardiology/European Society of Hypertension (ESC/ESH) guidelines5 maintained the existing thresholds to define hypertension in adults aged ≥16 years as sustained office BP ≥140/90 mmHg, as did the 2020 International Society of Hypertension (ISH) guidelines.6 However, the 2017 American College of Cardiology/American Heart Association (ACC/AHA) guidelines7 lowered the BP cut-off values, defining hypertension as BP ≥130/80 mmHg, based mainly on the epidemiological risk associated with these lower BP values rather than on the risk–benefit balance of reducing BP to lower values by treatment. The immediate consequence was a significant increase in the crude prevalence of hypertension of up to 45.6% of US adults,8 with similar estimates in countries such as Canada, China, India, and Italy. As this unilateral modification of the classical criteria was not agreed with the whole scientific community, studies have investigated whether subjects with the new ACC/AHA stage 1 (grade 1) hypertension, including subjects with SBP 130–139 mmHg and/or DBP 80–89 mmHg, a BP range still considered as ‘high-normal’ by the 2018 ESC/ESH guidelines,5 have an increased risk for CV events, CV mortality, and all-cause mortality. Of concern is a small group of subjects with increased DBP and normal SBP, defined as isolated diastolic hypertension (IDH), which is more frequent in younger patients but may be detected at any age. The prevalence of IDH depends on the thresholds that define it, and a recent report using data from 151 831 participants with normal SBP from the UK Biobank found a prevalence ranging from 6% (ESC/ESH definition) to 24.5% (ACC/AHA definition).9 It is unclear whether IDH is independently associated with increased CV risk and mortality. Some studies using the classical ESC/ESH definition (DBP >90 mmHg) have found no association with CV disease (CVD),10 while others found a modest 15% increase in CV risk,9 but no increased risk when the ACC/AHA guidelines were used to define IDH.9 In fact, the ACC/AHA reduction in the DBP cut-off to 80 mmHg, based on expert opinions but not robust evidence, not only results in an increased prevalence of IDH but also increases uncertainty about its association with long-term fatal and non-fatal CV events. Knowledge of the long-term CV risk of subjects with the new ACC/AHA grade 1 hypertension is important, but so are the benefits of interventions targeted at lowering BP to below 130/80 mmHg in these subjects. Focusing the problem on IDH defined by the 2017 ACC/AHA cut-off, few cohort studies of CV risk have been reported, and the results are conflicting; although most found no increased CV risk, a few found only a modest increase. These conflicting results may be explained by the methodologies used, especially the way in which BP was measured, and the statistical power. In fact, studies showing no increase in CV risk were those using the best, most reliable BP readings, whereas those showing modest increases in risk used less reliable and reproducible office BP readings. In addition, besides clinical CV events and mortality, the surrogate endpoints of silent hypertensive disease progression with respect to DBP values are also important in establishing the need for interventions in IDH, particularly in low CV risk subjects.5–7 In this issue of the European Heart Journal, Jacobsen et al.11 analyse the association of IDH with different endpoints in 5104 subjects aged 45–84 years from the Multi-Ethnic Study of Atherosclerosis (MESA), examining the risk for hypertension-mediated organ damage (HMOD), and the incidence of clinical CVD and new-onset systolic hypertension. Specifically, the study analysed the association between IDH and: (i) the coronary artery calcium (CAC) score, one of the strongest surrogate markers of silent progression from atherosclerosis to clinical CV events,12 as it may be handled as a continuous outcome variable using the CAC score in units; (ii) the incidence of systolic hypertension, with the effect of transforming IDH subjects into classical hypertensive subjects with a modified CV risk; and (iii) the incidence of CV events during 13 years of follow-up. Of the subjects included, 5059 (99.1%) were normotensive and 45 (0.9%) were diagnosed with IDH defined by the ESC/ESH criteria, compared with 3752 considered normotensive and 305 diagnosed with IDH using the ACC/AHA criteria. The prevalence of IDH increased from 0.9% using the ESC/ESH criteria to 7.5% using the ACC/AHA criteria. The authors also performed a meta-analysis of the adjusted association of IDH defined by the 2017 ACC/AHA criteria and CV events, including the MESA subjects in a database with those from seven previous cohort studies, making a total of eight studies with >10 million participants, in which two large cohorts used office BP measurements and six used more accurate research-grade BP measurements. The main findings were the lack association between IDH and CAC [adjusted odds ratio for CAC >0.88, 95% confidence interval (CI) 0.66–1.17], incident CVD [hazard ratio (HR) 1.19, 95% CI 0.77–1.84], and death (HR 0.94, 95% CI 0.65–1.36) in middle-aged and older subjects during 13 years of follow-up. As expected, IDH was significantly associated with incident SBP ≥130 mmHg, which was detected in 41 persons not taking antihypertensive medication at baseline (HR 3.76, 95% CI 2.38–5.94). In addition, as shown in Graphical Abstract, in the meta-analysis IDH was not consistently associated with CVD in cohorts using the best research-grade BP measurements (adjusted HR 1.04, 95% CI 0.98–1.10) and there was a very modest association when all cohorts (including those with routine office BP measurements) were combined (adjusted HR 1.09, 95% 1.03–1.15). The association between IDH and the composite endpoint of CVD, chronic kidney disease (CKD), atrial fibrillation (AF), and heart failure (HF) in the post-hoc sensitivity analyses was not significant, and IDH was not associated with all-cause death. The study results have important implications because they clearly show the lack of an increased risk for CVD and for the composite of CVD, CKD, AF, and HF in middle-aged and older subjects with IDH diagnosed by the ACC/AHA threshold of DBP ≥80 mmHg. The implications are double-edged: (i) for the definition of IDH, the results further indicate that the 2018 ESC/ESH guideline decision5 to keep the diastolic BP threshold of ≥90 mmHg for the diagnosis of hypertension was reasonable and should be maintained, as was done by the 2020 ISH guidelines;6 and (ii) the need or not for therapeutic interventions. This is more complicated, because physicians’ decisions in patients with high BP are based not only on BP levels but also on the total CV risk conferred by the association of other CV risk factors such as diabetes, dyslipidaemia, smoking, obesity, sedentarism, HMOD, clinical CV, and cerebrovascular or renal disease. In addition to lifestyle changes, antihypertensive drug treatments are indicated in subjects at high risk, particularly those with clinical CVD, even in the range of ‘high-normal BP’ (grade 1 hypertension for the ACC/AHA), as recommended by all guidelines.5–7 The grey area concerns subjects at low or moderate risk with high-normal BP (ACC/AHA grade 1), where no evidence of the benefit of pharmacological treatment on the incidence of CV events, CV mortality, and all-cause mortality has been reported.13,14 Nevertheless, subjects with IDH, even when diagnosed by ACC/AHA thresholds, have a significant risk for systolic hypertension over time, and the combination of BP ≥130/80 mmHg, even in grade 1, is not a benign condition, particularly in the long term. All guidelines agree that subjects in this BP range at low/moderate CV risk should strictly adhere to healthy lifestyle habits to delay or avoid progressive BP increases and the progression of silent HMOD leading to clinical CV events and premature death. Unfortunately, the MESA cohort included subjects aged ≥45 years and cannot shed light on the natural history of IDH in younger people or the management of IDH in these subjects. When IDH is detected early in the young, there are many years to possible CV events; however, even in a prospective clinical trial with a very long follow-up, the risk of CV events may be confounded by the incidence of systolic hypertension over time, as shown by the current study. In summary, physicians should assume that healthy lifestyles rather than antihypertensive drug therapy are sufficient in young, middle-aged, and older adults who meet the 2017 ACC/AHA criteria for IDH, which we continue to define as high-normal BP in the ESC/ESH guidelines. Conflict of interest: none declared. The opinions expressed in this article are not necessarily those of the Editors of the European Heart Journal or of the European Society of Cardiology.