Isolated Diastolic Hypertension in the UK Biobank: A Comparison of ACC/AHA and ESC/NICE Guideline Definitions

Abstract

The 2017 ACC/AHA guideline defines hypertension as a blood pressure (BP) ≥130/80mmHg, whereas the 2018 ESC and 2019 NICE guidelines use a ≥140/90mmHg threshold. Our objective was to study the associations between isolated diastolic hypertension (IDH), diagnosed using these two BP thresholds, and cardiovascular disease (CVD) in a large cohort of UK adults. We analysed data from UK Biobank, which enrolled participants between 2006-2010 with follow-up through March 2019. We excluded persons with systolic hypertension or baseline CVD. We defined incident CVD as a composite of non-fatal myocardial infarction, non-fatal stroke, or cardiovascular death. We used Cox regression to quantify associations between IDH and CVD, as well as the individual outcomes included in the composite outcome. We studied 151,831 participants with normal systolic BP (mean age 54 years, 40% male). Overall, 24.5% had IDH by the ACC/AHA definition compared with 6% by the ESC/NICE definition. Compared to normal BP, IDH by the ACC/AHA definition was not significantly associated with CVD risk (HR 1.08, 95% CI 0.98, 1.18) whereas IDH by the ESC/NICE definition was significantly associated with a modest increase in CVD (HR 1.15, 95% CI 1.04-1.29). Similar results were found by sex and among participants not taking baseline antihypertensives. Furthermore, neither IDH definition was associated with the individual outcomes of non-fatal myocardial infarction or stroke. In conclusion, the proportion of UK Biobank participants with IDH was significantly higher by the ACC/AHA definition compared to the ESC/NICE definitions; however, only the ESC/NICE definition was statistically associated with increased CVD risk.

INTRODUCTION

Hypertension is a common and modifiable cardiovascular disease (CVD) risk factor that currently affects over 1 billion people worldwide; a number that is estimated to exceed 1.5 billion by 2025. ¹ Historically, the most widely accepted threshold for defining hypertension has been a blood pressure (BP) greater than or equal to 140/90 mmHg. In 2017, the updated American College of Cardiology (ACC)/American Heart Association (AHA) blood pressure guideline reduced this threshold to 130/80mmHg, ² a change with numerous consequences. One of these is the redefinition of isolated diastolic hypertension (IDH), now defined as a diastolic BP ≥80mmHg with a systolic BP <130mmHg in the American guideline. This lower threshold increased the estimated prevalence of IDH among the US population from 1.3% to 6.5%, which translates into millions more Americans being newly eligible to be diagnosed with this condition. ³

In contrast to the American guideline, the 2018 European Society of Cardiology (ESC)¹ and 2019 National Institute of Clinical Excellence (NICE) ⁴ BP guidelines continue to recommended that adults in Europe and the United Kingdom (UK) be diagnosed with hypertension based on a threshold of ≥140/90mmHg. As such, the definition of IDH now differs between America and Europe. There is also uncertainty as to whether either definition of IDH is associated with CVD. Whereas some prior studies have demonstrated associations between the ESC/NICE definition of IDH and incident CVD^5–7, many others have not. ^8–11 We are aware of only two studies testing the ACC/AHA definition of IDH, with one epidemiological analysis from our group reporting no associations with CVD (Hazard Ratio [HR] of 0.94; 95% CI 0.78-1.12)³ and another analysis of a large clinical database suggesting a statistically marginal association with CVD (HR of 1.52; 95% CI 1.03-2.23; p=0.03). ⁷

The objective of the present study was to further examine, in a large epidemiological study of over 500,000 UK adults, whether IDH as defined by either the AHA/ACC definition or the ESC/2019 NICE definition is associated with a composite CVD outcome; in addition to leveraging the large sample size and sufficient power of the UK Biobank to test associations of IDH with the individual outcomes of non-fatal myocardial infarction or stroke.

METHODS

The UK Biobank study was approved by the North West Multi-center Research Ethics Committee. All participants provided written informed consent for the study. UK Biobank is a publicly available open database and the data are available for researchers after acceptance of a formal research proposal. The current study was conducted under application number 17712 (PI: Chatterjee). We will make our analytic methods and study materials available to other researchers on request to the corresponding author.

Study Population

The UK Biobank cohort is a national prospective cohort study that enrolled over 500,000 adult residents in the United Kingdom between 2006 and 2010 (baseline). ^{12, 13} In the United Kingdom, virtually all members of the general population are registered with a general practitioner through the National Health Service (NHS). Participants in the UK Biobank were invited on the basis of being identified from NHS patient registers and living within a reasonable travelling distance of an assessment centre. In order to recruit a widely generalisable population, the invitation mailing was stratified according to key demographic parameters (including age, gender and postcode areas as a measure of social deprivation), with over-sampling of particular groups as required.

All UK Biobank participants underwent a single baseline examination, at which time investigators collected a broad range of lifestyle, socioeconomic, environmental and genetic data. Questionnaires, physical assessments including blood pressure measurement, and phlebotomy were all used to collect this information during the baseline study visit.

For this analysis we excluded individuals due to genetic relatedness (n=150,832) or for being part of the very small subgroup that were not white (n=14,220). Unrelated individuals were determined from the quality control file provided by the UK Biobank. Other exclusion criteria were missing information on BP, the presence of systolic hypertension, or a baseline history of CVD (Figure 1).

Figure 1:

Flow chart of UK Biobank Participants Analyzed

Measurement of blood pressure

Two BP measurements were taken after 5 minutes in the seated position with a minimum of 1 minute between recordings. Measurements were performed using an appropriately sized cuff and an Omron 705IT automatic digital BP monitor. We calculated mean systolic BP and diastolic BP using both BP measurements. For individuals with just 1 automated BP reading (n=7886), we included these single values.

Measurement of covariates

Participant race/ethnicity were assessed using self-reported information on the study questionnaire. We calculated years of education based on the standardized International Standard Classification of Education of the United Nations Educational, Scientific and Cultural Organization. ¹⁴ Average annual household income was self-reported. The Townsend Deprivation Index (TDI) ¹⁵ was also derived for each participant, this is an area-based proxy measure for socioeconomic status composed of data on car ownership, household overcrowding, household owner-occupation, and unemployment (a higher TDI indicates worse deprivation). We classified smoking status based on self-report as current, former, or never smoking. Alcohol intake status, diabetes status, and antihypertensive drug use status were all self-reported. Estimated glomerular filtration rate (eGFR) was calculated based on serum creatinine using the CKD-EPI equation. ¹⁶ Serum total cholesterol, high density lipoprotein cholesterol, and triglycerides were measured using the Beckman Coulter AU5800 analyser; with low density lipoprotein cholesterol measured using a direct homogeneous Beckman assay. Genome-wide genotype data was collected on all UK Biobank participants, requiring centralized analysis of the following; genotype quality, relatedness of the genetic data, and properties of population structure (with the latter being accounted for in statistical models using principal components analyses).¹⁷

Outcomes

Methods for follow-up in the UK Biobank have been described elsewhere; ^{12, 13} in brief, incident disease diagnoses were acquired from linkage of study records to national hospital inpatient and outpatient records, death registrations, and primary care diagnoses. We defined CVD as non-fatal myocardial infarction, non-fatal stroke or cardiovascular death; with non-fatal diagnoses recorded on the basis of incident cases obtained from hospital inpatient data where it was documented as the primary diagnosis by ICD-10 nomenclature. We considered time to disease or time to death as the outcome. Follow-up for outcome was conducted through 11^th March 2019.

Statistical analysis

We defined IDH as a diastolic BP ≥80mmHg together with systolic BP <130mmHg according to the ACC/AHA definition and a diastolic BP≥90 with systolic BP <140mmHg by the ESC/NICE definition. All comparisons were to UK Biobank participants with normotension and persons with systolic hypertension were not included in the main analysis. Therefore, for analyses of IDH by the ACC/AHA definition the reference group was systolic BP<130 and diastolic BP <80 mmHg and for analyses of IDH by the ESC/NICE definition the reference group was systolic BP <140 and diastolic BP <90mmHg. Furthermore, persons with biologically implausible and likely erroneously recorded diastolic BP values were also not included in our models (i.e., for analyses of the ACC/AHA definition of IDH we excluded 433 participants with a diastolic BP over 120mmHg and for analyses of the ESC/NICE definition we excluded 54 participants with a diastolic BP over 130 mmHg). The percentage of UK Biobank participants meeting diagnostic criteria for IDH, based on both AHA/ACC and ESC/NICE definitions, were then calculated. Comparisons of demographics between those with normotension and those with IDH were made using the t-test to compare normally distributed continuous variables and the χ2 test for proportions.

We tested associations of IDH, by both definitions, with incident CVD using Cox regression models using time to event data. The proportionality of hazards was visually confirmed. Model 1 was adjusted for age; sex; education, categorized as higher (college/university degree or other professional qualification), upper secondary (second/final stage of secondary education), lower secondary (first stage of secondary education), vocational (work-related practical qualifications), or other; socioeconomic status (with categories derived from Townsend deprivation index quintiles 1, 2 to 4, and 5); enrolment center; and the first 20 principal components of ancestry (to adjust for population structure and given that BP variability is partly hereditable). Model 2 included the Model 1 variables in addition to smoking status, alcohol consumption status, high density lipoprotein cholesterol, low density lipoprotein cholesterol, triglycerides, diabetes status, body mass index, antihypertensive drug use status, and eGFR. A high rate of missingness for anti-hypertensive use (30%), necessitated the creation of an ‘Unknown Category’ to include all those with missing information for antihypertensive medication. Model 3 included the Model 2 variables in addition to baseline systolic BP value. Unadjusted and adjusted Kaplan Meier cumulative incidence curves were constructed according to IDH category. For the adjusted Kaplan Meier curves, we set the exposure variables in model 2 of the Cox regression to a common value across the IDH groups. The common value was estimated with continuous factors set to their average value and categorical variables set to the most frequent category.

We conducted a number of sensitivity analyses. First, we assessed for an association between IDH and CVD on the basis of median age and on the basis of sex (also testing for effect modification using multiplicative interaction terms). Second, we also conducted analyses of IDH among the subgroup of UK Biobank participants recorded as not being on antihypertensive medication at baseline. Third, we tested for associations of IDH with the individual outcomes of non-fatal myocardial infarction and non-fatal stroke (persons who died in follow-up were right censored for this analysis). All analyses were performed using R (version 3.6.1) and we considered a p<0.05 (two sided) as statistically significant.

RESULTS

UK Biobank investigators enrolled 502,536 participants aged 37 to 70 years at 22 centers, with 94% of the sample being white and of European ancestry. A total of 98.25% of participants were born in the UK and therefore are assumed to have complete information on their medical history available in the NHS electronic health record. For the analysis of the ACC/AHA definition of IDH, we studied 89,126 participants with systolic BP below 130 mmHg (mean age 53 years, 34% male). Among these, 75.5% (67,254) were normotensive and 24.5% (21,872) had IDH. For the analysis of the ESC/NICE definition of IDH, we studied 151,831 participants with systolic BP below 140 mmHg (mean age 54 years, 40% male). Of these, 93.7% (142,225) were normotensive and 6.3% (9,606) had IDH (Figure 1). Redefining IDH using the ACC/AHA criteria therefore lead to an absolute increase in prevalence of 18.2 percentage points.

Participants with IDH by both definitions were more likely to be younger, male, overweight, have higher baseline systolic BP, have less favourable lipid values, and have a lower eGFR (Table 1). Using both definitions, patients with IDH also had higher Townsend Deprivation Index scores and were more likely to have completed O levels or A levels as their ceiling of education- whereas participants who held a degree were less likely to have IDH by either criterion. Using the ESC/NICE definition, patients with IDH were more likely to diabetic or be current or past smokers, whereas there was no significant difference in the prevalence of either diabetes or smoking by the ACC/AHA definition of IDH.

Table 1:

Characteristics of UK Biobank participants at the baseline visit based on the presence or absence of IDH, both by ESC/NICE and ACC/AHA definitions ^*

Participant Characteristics	Normotension by ESC/NICE Definition N=142225	IDH by ESC/NICE Definition N=9606	P-value†	Normotension by ACC/AHA Definition N=67254	IDH by ACC/AHA Definition N=21872	P-value‡
Percentage %	93.7	6.3		75.5	24.5
Male (%)	39.0	49.9	<0.001	31.8	39.4	<0.001
Age (Years)	54.66 (8.08)	53.20 (7.70)	<0.001	53.80 (8.08)	52.73 (7.63)	<0.001
Townsend Deprivation Index	-1.54 (2.93)	-1.30 (3.02)	<0.001	−1.50 (2.95)	−1.37 (3.00)	<0.001
Qualification			<0.001			<0.001
O levels, CSE or equivalent	17.6	18.3		17.5	18.6
A levels, professional or equivalent	33.7	36.9		32.9	35.1
Degree	35.8	31.7		38.3	35.0
None	12.9	13.0		11.3	11.4
Smoking Status			0.034			0.249
Never	56.5	55.1		57.2	56.6
Previous	32.5	33.5		31.3	31.8
Current	11.0	11.4		11.5	11.6
Alcohol consumption (%)			0.043			<0.001
Never	3.0	2.7		3.1	2.6
Previous	3.6	3.3		4.0	3.5
Current	93.3	93.9		92.9	93.9
Antihypertensive (%)			<0.001			<0.001
Yes	0.001	0.0		0.007	0.01
No	69.3	73.3		68.0	71.0
NA category	30.6	26.7		31.9	29.0
SBP (mmHg)	123.19 (9.29)	131.96 (5.19)	<0.001	116.34 (7.51)	122.59 (4.63)	<0.001
Pulse pressure (mmHg)	47.44 (7.62)	41.16 (5.18)	<0.001	45.22 (6.68)	40.23 (4.94)	<0.001
BMI (kg/m2)	26.39(4.43)	29.62 (5.44)	<0.001	25.45 (4.08)	27.83 (4.98)	<0.001
Diabetes (%)	3.7	4.5	<0.001	2074 (3.1)	713 (3.3)	0.199
Direct LDL (mmol/L)	3.51 (0.84)	3.67 (0.86)	<0.001	3.44 (0.82)	3.59 (0.84)	<0.001
HDL Cholesterol (mmol/L)	1.48 (0.38)	1.37 (0.36)	<0.001	1.51 (0.38)	1.43 (0.37)	<0.001
Triglycerides (mmol/L)	1.60 (0.94)	1.92 (1.10)	<0.001	1.47 (0.86)	1.70 (0.99)	<0.001
eGFR (ml/min/1.73m2)	9.72 (10.30)	9.08 (9.96)	<0.001	10.05 (10.6)	9.42 (10.37)	<0.001

^*Values are either % or mean (standard deviation)

^†P value comparing those with normotension (systolic BP <140 mmHg and diastolic BP <90 mmHg) to those with IDH (systolic BP <140 mmHg and diastolic BP ≥90 mmHg)

^‡P value comparing those with normotension (systolic BP <130 mmHg and diastolic BP <80 mmHg) to those with IDH (systolic BP <130 mmHg and diastolic BP ≥80 mmHg)

There were 3,451 CVD events amongst 89,126 participants over a 10-year median follow-up in the analysis of the ACC/AHA definition of IDH. The CVD incidence rate (per 1,000 person years) was 4.5 for participants defined as normotensive and 4.3 for those defined as having IDH in the analyses of the ACC/AHA definition. There were 6,475 CVD events amongst 151,831 participants over a median 9.8 years follow-up in the analysis of the ESC/NICE definition of IDH. The CVD incidence rate (per 1,000 person years) was 4.8 for participants defined as normotensive and 5.2 for those defined as having IDH in the analyses of the ESC/NICE definition. The respective Kaplan Meier cumulative incidence curves are provided in Figure 2 (and the adjusted Kaplan Meier curves in Figure S1, please see http://hyper.ahajournals.org).

Figure 2.

a: Cumulative incidence Kaplan Meier curves for ESC/NICE definition of IDH.

b: Cumulative incidence Kaplan Meier curves for AHA/ACC definition of IDH.

Compared to normal blood pressure, IDH by the ACC/AHA definition was not statistically associated with a significantly increased risk of CVD (HR 1.08, 95% CI 0.98-1.18, Model 3, Table 2). However, despite the lack of statistical significance in the traditional sense, the confidence intervals include the possibility of up to an 18% increase in risk for IDH by the ACC/AHA definition. On the other hand, IDH by the ESC/NICE criteria was significantly associated with a modestly higher risk of CVD (HR 1.15, 95% CI 1.04-1.29, Model 3). The association between IDH and ASCVD was also assessed in subgroups stratified on the basis median age and sex, with similar findings (Figure 3, all p-interaction >0.1). In addition, among the 72.5% of UK Biobank participants who were not taking BP medication at baseline and who did not have missing information on this variable, IDH by the ACC/AHA definition was not significantly associated with a statistically increased risk of CVD (HR 1.05, 95% CI 0.80-1.37; Model 3) whereas IDH by the ESC/NICE definition was again associated with an increased risk (HR 1.17, 95% CI 1.04-1.32; Model 3). As with the analysis of the sample overall, there was nonetheless overlap in the confidence intervals of the associations between both IDH definitions and CVD among persons not on baseline antihypertensive therapy.

Table 2:

Hazard ratios (95% CI) for IDH, by both ESC/NICE and ACC/AHA definitions, and incident cardiovascular disease events^*

Definition of IDH	Adjustment Model	Cardiovascular Disease (Fatal and non-fatal)	Non-Fatal Stroke	Non-Fatal MI
IDH by ESC/NICE definition	Model 1†	1.12 (1.02,1.23)	1.43 (1.12,1.83)	1.31 (1.05,1.64)
	Model 2‡	1.10 (0.99, 1.22)	1.23 (0.93, 1.62)	1.13 (0.89,1.45)
	Model 3§	1.15 (1.04, 1.29)	1.17 (0.88, 1.56)	1.08 (0.84,1.39)
IDH by ACC/AHA definition	Model 1†	0.98 (0.91, 1.06)	1.25 (1.01, 1.56)	1.14 (0.93, 1.39)
	Model 2‡	1.00 (0.92, 1.10)	1.29 (1.02, 1.64)	1.10 (0.88, 1.38)
	Model 3§	1.08 (0.98,1.18)	1.26 (0.98, 1.63)	1.10 (0.86, 1.39)

^*All comparisons are to UK Biobank participants with normal blood pressure (when studying IDH by the ESC/NICE definition this group consists of those with systolic BP <140 mmHg and diastolic BP <90 mmHg; when studying IDH by the ACC/AHA definition this group consists of those with systolic BP <130 mmHg and diastolic BP <80 mmHg)

^†Model 1 is adjusted for age, sex, qualification, assessment center, Townsend deprivation index, and first 20 genetic principal components

^‡Model 2 is adjusted for model 1 + smoking, alcohol consumption, HDL, LDL, triglycerides, body mass index, antihypertensive medication, diabetes, eGFR

^§Model 3 is adjusted for model 2 + baseline systolic BP

Figure 3: Subgroup Analyses of Hazard Ratios* of the association of IDH, by both ESC/NICE and ACC/AHA definitions, with ASCVD events According to (a) Median Age, (B) Sex, or (C) Not Being on Baseline Antihypertensive Medications.

* All comparisons are to UK Biobank participants with normal blood pressure (when studying IDH by the ESC/NICE definition this group consists of those with systolic BP <140 mmHg and diastolic BP <90 mmHg; when studying IDH by the ACC/AHA definition this group consists of those with systolic BP <130 mmHg and diastolic BP <80 mmHg). The model is adjusted for age, sex, qualification, assessment center, Townsend deprivation index, first 20 genetic principal components, smoking, alcohol consumption, HDL, LDL, triglycerides, body mass index, antihypertensive medication, diabetes, eGFR, and baseline systolic BP.

Finally, analyses of each of the individual endpoints in the composite outcomes were assessed (Table 2). There was no statistically significant association between IDH and non-fatal MI by either the ACC/AHA (HR 1.10, 95% CI 0.86, 1.39; Model 3) or the ESC/NICE definitions (HR 1.08, 95% CI 0.84-1.39; Model 3). Similarly, no statistically significant association was found between IDH and non-fatal stroke by the ACC/AHA (HR 1.26, 95% CI 0.98-1.63; Model 3) or the ESC/NICE definitions (HR 1.17, 95% CI 0.88-1.56; Model 3). These results were similar when the sample was analysed by age and gender (Table S1, please see http://hyper.ahajournals.org).

DISCUSSION

Defining IDH by the ACC/AHA threshold substantially increased the prevalence of IDH in the UK Biobank cohort when compared to the ESC/NICE definition. However, there was no statistically significant association between IDH by the ACC/AHA definition and CVD events; whereas IDH as defined by ESC/NICE was significantly associated with a modest increased risk of CVD. The excess risk for the ESC/NICE definition of IDH was most evident among women and participants not on baseline antihypertensive therapies.

Previous epidemiological studies have found that both systolic and diastolic hypertension are associated with CVD risk, ¹⁸ although the two are highly correlated and systolic hypertension has been shown to have by far the stronger association with CVD of the two parameters. ⁷ Therefore, the focus of our study was not on whether diastolic BP as a continuous variable is harmful per se, but rather on the prognostic implication of the specific BP phenotype of IDH. In persons with IDH, CVD risk is being driven by diastolic BP elevation (because they have normal systolic BP). As such, the IDH phenotype that may be used to test - in isolation - the diastolic BP cut-point for hypertension recommended in a given guideline. Furthermore, there has been controversy about whether IDH is associated with increased CVD risk; with this controversy being particularly true for the lower BP cut-points recommended by ACC/AHA. ^{3, 19}

We have previously reported no statistically significant excess risk for CVD with either the ESC/NICE or the more liberal ACC/AHA definitions of IDH in the Atherosclerosis Risk in Communities (ARIC) Study.³ We now confirm in a far larger cohort from UK Biobank that IDH, when defined by ACC/AHA, was not significantly associated with excess CVD risk. By contrast, the present analysis of the UK Biobank suggests that the ESC/NICE definition of IDH was significantly associated with a modest increase in ASCVD risk. While these respective associations do differ in their statistical significance (at least from a traditional perspective), it is nonetheless worth noting that the confidence intervals for both overlap and as such our data do not exclude the possibility that the true associations between both definitions of IDH and CVD are in fact the same. Similarly, one difference between our prior ARIC study³ and the current study is that the larger sample size in the UK Biobank may have allowed for the detection of the modestly increased CVD risk seen among those with IDH as defined by ESC/NICE that was not previously evident in ARIC. This finding of increased risk for CVD by the ESC/NICE definition of IDH in UK Biobank is in line with the report of Flint et al, who studied a large administrative database and also found excess CVD risk with IDH by the ESC/NICE definition. ⁷

In contrast to the present results from UK Biobank and our previous results from ARIC, neither of which demonstrated a statistically significant increased risk of CVD when IDH was defined by ACC/AHA criteria³, the Flint data also suggested a marginally significant association between IDH and events when defined by the ACC/AHA definition.⁷ However, the Flint study reported risk related to IDH in a way that differed from our analysis. While we modelled IDH was a binary (yes/no) exposure, which is intuitive, Flint et al. modelled the risk of IDH as a per unit increase in diastolic BP Z-score among persons who met criteria for a diagnosis of IDH. Therefore, the two associations cannot be directly compared. Furthermore, Flint et al used a large administrative dataset derived from an electronic health records where BP measurements were performed for reasons that were potentially clinically informative (i.e., confounding by indication) and where both the measurement and recording of BP may have been less accurate than in epidemiologic studies like the UK Biobank (which measured BP using a rigorous protocol, a concern that is particularly operative for the difficult measurement of diastolic BP²⁰).

Leveraging the large sample size of UK Biobank, we were also able to examine the associations of both IDH definitions with the individual non-fatal myocardial infarct and stroke endpoints comprising the CVD composite used as our primary outcome. We found no statistically significant association between IDH and these individual outcomes. However, while not statistically significant, the point estimate for non-fatal stroke suggested a possible trend for increased risk with IDH by the ACC/AHA definition (HR 1.26, 95% CI 0.98, 1.63; Model 3). Furthermore, in our present study, CVD events were derived from hospital inpatient data and ICD 10 codes. As such, further studies with more rigorous adjudication of stroke and more stroke events should assess this potential association in other cohorts.

This UK Biobank study has several limitations. Firstly, the population studied represents a homogenous racial group from the United Kingdom and, as such, the results may not be generalisable to other populations. Secondly, the potential for a selection bias exists in that healthier participants may have been more likely to enrol in the study. Third, the UK Biobank population was restricted to adults aged between 37 to 70 years at baseline making its application to younger adults difficult, a group in whom IDH is particularly common. ²¹ Finally, it is important to note that the focus of this analysis was IDH and we did not assess the relative merits of the systolic BP thresholds recommended in BP guidelines.

PERSPECTIVES.

In the UK Biobank sample, IDH was substantially more prevalent when defined by the ACC/AHA definition compared to the ESC/NICE definitions. However, IDH was only associated with an increased CVD risk when defined by the ESC/NICE definition. Taken together, these results provide observational data to support the higher diastolic BP threshold for diagnosing hypertension as recommended by the 2018 ESC and 2019 NICE BP guidelines.

NOVELTY AND SIGNIFICANCE

What is new?

• There has been conflicting evidence as to whether IDH is associated with adverse CVD outcomes

• This study sought to further assess the impact of IDH by both the ESC/NICE and ACC/AHA definitions on CVD outcomes in a large epidemiological cohort from the UK Biobank.

What is relevant?

• This study showed a significant association between IDH by ESC/NICE definition and CVD outcomes, whereas no association was found when testing IDH by the ACC/AHA definition.

• Relative to the ACC/AHA cut-point of ≥80 mmHg, these UK Biobank findings provide observational evidence to support the ESC/NICE guideline cut-off for diastolic hypertension of ≥90 mmHg

Summary:

• In the UK Biobank population, IDH was substantially more prevalent when defined by the ACC/AHA guidelines compared to the ESC/NICE guidelines. However, only IDH as defined by the ESC/NICE guidelines was associated with an increase in CVD