Isolated diastolic hypertension and target organ damage: Findings from the STANISLAS cohort

Abstract Background Isolated diastolic hypertension (IDH) is defined as diastolic blood pressure (DBP) ≥80 mmHg and systolic blood pressure (SBP) <130 mmHg according to 2017 ACC/AHA guidelines. The effective cardiovascular risk linked to IDH is debated. Hypothesis IDH might contribute marginally to hypertension‐related target organ damage (TOD) development. Methods In this cross‐sectional analysis 1605 subjects from the STANISLAS cohort, a large familiar longitudinal study from Eastern France, were included. Participants were categorized according to average values at 24‐h ABP recording as having normal BP (SBP < 130/DBP < 80 mmHg); combined hypertension (SBP ≥130/DBP ≥80 mmHg or on antihypertensive treatment); IDH (SBP <130/DBP >80 mmHg); isolated systolic hypertension (ISH: SBP ≥130/DBP <80 mmHg). The association between hypertension status and TOD was assessed by multivariable‐adjusted logistic models. Results Using normotension as reference, IDH was not significantly associated with NTproBNP levels (adjusted odds ratio [OR] 1.04 [95%CI 0.82;1.32], p = .750), microalbuminuria (OR 0.99 [0.69; 1.42], p = .960), diastolic dysfunction (OR 1.53 [0.88; 2.68], p = .130), left ventricular (LV) mass index (OR per 10 g/m2 increase 1.07 [0.95; 1.21], p = .250), LV longitudinal strain (global: OR 1.07 [0.99; 1.14], p = .054; subendocardial: OR 1.06 [0.99; 1.13], p = .087), carotid intima media thickness (OR 1.27 [0.79; 2.06], p = .320), reduced ankle‐brachial index (<0.9; OR 1.59 [0.19; 13.55], p = .670) and pulse wave velocity (PWV; OR 1.07 [0.93; 1.23], p = .360). In contrast, combined hypertension and ISH were independently associated with LV mass index and PWV increase (all p ≤ .01). Conclusions IDH was not significantly associated with TOD. Further studies are needed to clarify the clinical role of IDH. Registration: URL: https://www.clinicaltrials.gov; Unique identifier: NCT01391442.


| INTRODUCTION
Hypertension is a major risk factor for cardiovascular morbidity and mortality. 1 Current guidelines classified hypertension into isolated diastolic (IDH), isolated systolic (ISH), and systolic and diastolic mixed (or combined) hypertension based on the elevation of systolic and/or diastolic blood pressure (DBP) values. 2,3 IDH is a less prevalent hypertension definition, 4 and is classified as elevated diastolic BP with a systolic BP within the normal range. 2 [5][6][7] A recent longitudinal analysis that included 8703 adults failed to show a significant association between IDH, as defined by the 2017 ACC/AHA guidelines, and increased risk for cardiovascular outcomes. 8 In line, previous observations demonstrated that IDH is usually not associated with cardiovascular outcomes independently of baseline systolic BP. 5,9,10 On the other side, some studies showed a slight but significant association between diastolic hypertension and cardiovascular risk, [11][12][13][14] although the larger part of these reports did not exclusively investigate IDH but diastolic hypertension in the setting of combined hypertension. Given this context, the effective cardiovascular risk linked to IDH is debatable. The Suivi Temporaire Annuel Non-Invasif de la Santé des Lorrains Assurés Sociaux (STANISLAS) cohort is a longitudinal transgenerational study from the Nancy region of France characterized by a familial structure and a long follow-up (up to 23 years). In this cohort, individuals underwent to an extensive cardiovascular evaluation and the hypertensive status was evaluated by ambulatory BP monitoring (ABPM), that provides a more accurate diagnosis and a better prediction of cardiovascular risk compared to office BP. 2,3 Moreover, several papers have suggested that 24-h average blood pressure (BP) is superior to office BP in relation to hypertension target organ damage. 15 The STANISLAS cohort therefore offers the unique opportunity of studying the early changes induced by cardiovascular risk factors in initially healthy subjects. 16 The aim of the present study is to determine the association between IDH, identified by 24-h ambulatory BP monitoring, and markers of target organ damage in a populational cohort with detailed cardiovascular phenotyping and long follow-up. survivors of the original cohort underwent their fourth examination (STANISLAS-V4) at our department, as previously described. 17 The research protocol was approved by the local Ethics Committee (Comité de Protection des Personnes Est III-Nancy-France) and all study participants gave a written informed consent to participate. The informed written consent was approved previously by the local ethics committee (ClinicalTrials.gov identifier NCT01391442). 16 For the present study, 1605 adult patients (i.e., ≥18 years old and with ambulatory BP measurements) attending STANISLAS-V4 were included in the analyses ( Figure 1). All participants were scheduled to attend the Centre d'Investigation Clinique Plurithématique Pierre Drouin at Nancy University Hospital at 8 AM after a 12-to 14-h fast. All subjects underwent blood and urine sampling for laboratory analysis. Medical history, medications, anthropometric parameters were also recorded.

| Blood pressure measurements
Office BP was measured three times in all participants, at 1-min intervals, using an electronic sphygmomanometer after the participant had rested for at least 10 min and calculated as the mean of the three measurements. As previously described, 16 all participants underwent a 24-h recording of ABPM using the Spacelabs 90207 ambulatory monitor (Spacelabs Medical, WA). The monitoring cuff was placed around the participant's nondominant arm. The BP system was programmed to acquire measurements every 15 min from 6 AM to 10 PM and every 30 min from 10 PM to 6 AM. Self-reported sleep-wake times were used to divide ABP monitoring data into daytime and nocturnal periods. BP indices were calculated from 24-h, daytime, and nighttime averaged measurements according to current guidelines. 18 In addition, participants had to complete a diary describing their main daily activities (e.g., eating, sleeping) and were asked to avoid excessive exercise during the 24-h recording. Central reading of the recordings was performed by a trained technician blinded to the participants' clinical features. Data were considered for further analysis if they met the following criteria: the recording lasted ≥24 h, ≥70% of the expected number of readings were available, no missing data for >2 consecutive hourly intervals, and ≥2 valid measurements were and DBP reduced more than 10% between daytime and nighttime, otherwise he was considered a nondipper. 18

| Assessment of markers of target organ damage
Echocardiographic exams were performed by an experienced sonographer, in the left lateral decubitus position with a commercially available standard ultrasound scanner (Vivid 9, General Electric Medical Systems, Horten, Norway) using a 2.5 MHz phased-array transducer (M5S), as previously described. 16 The echo/Doppler examination was performed according to EAE/ASE recommendations 19 and longitudinal left ventricular (LV) deformation parameters (strain) were obtained using speckle tracking echocardiography. LV diastolic dysfunction (DD) was assessed according to the 2009 ASE/EACVI recommendations. 20 We used this algorithm instead of the most recent 2016 ASE/EACVI scoring system because the new criteria commonly results in a marked reduction of DD, as also recently demonstrated by our group. 21 Image acquisitions were shown to be highly reproducible. 17 Carotid intima media thickness (cIMT) measurements were performed by high-resolution echo-tracking (Wall Track System; Pie Medical, Maastricht, The Netherlands) on the right common carotid artery at 1-2 cm below the carotid bifurcation and the retained value was obtained as the mean of four measurements. 22 Both reproducibility and agreement (intra/interoperator/devices) were excellent. 16 Carotid to femoral pulse wave velocity (PWV) measurement was performed with Complior ® (Alam Medical, Vincennes, France) and Sphygmocor ® CVMS (AtCor Medical, Sydney, NSW, Australia) devices according to the European Network for Noninvasive Investigation of Large Arteries recommendations. 23 The protocol was previously discussed in detail. 16 The ankle-brachial index (ABI) was calculated as the systolic pressure at the ankle, divided by the systolic pressure at the arm. Systolic BP was measured with a continuous Doppler machine using a BP cuff.
An ABI between 0.9 and 1.2 is considered normal, while an index <0.9 suggests the presence of arterial disease. 24 The measurements were performed according to a standardized protocol 25 on the side (arm and ankle) with the highest SBP.
N-terminal pro-B-type natriuretic peptide (NT-proBNP) was measured using a multiplex assay (CVDII panel, Olink Proteomics AB ® , Uppsala, Sweden). The final assay readout was expressed as normalized protein expression values, which is an arbitrary unit on a log2 scale in which a higher value corresponds to higher protein expression.

| Statistical methods
Proportions were compared using χ 2 test and were expressed as number (proportion as percentage). Continuous variables were expressed as mean ± SD or median (interquartile range [IQR: Q1; Q3]) and compared using one-way ANOVA and Kruskal-Wallis depending on the normality of the distribution. We focused on the association between noninvasive markers of target organ damage and IDH, with a cutoff of DBP ≥80 mmHg and SBP <130 mmHg at 24-h ABP recording.
Descriptive statistics were computed on all patients' characteristics in the sample overall and according to hypertension status. Multinomial logistic regression models were used to assess the associations between the dependent variable hypertension (HT) categories (No HT [reference], IDH, ISH, and SDH) and independent variables (echocardiographic, vascular and biological markers of target organ damage).
The log linearity assumptions of the relationship between HT status and the continuous variables of target organ damage were assessed using restricted cubic splines according to the Harrell's rule. 26 We use the Wald test for assessing the linearity assumptions. The variables were then categorized, as required, to meet the model assumptions.
Each model was adjusted for gender, age, waist circumference, smoking status, total cholesterol, glycemia, lipid lowering agents, selfdeclared hypertension status, glomerular filtration rate, and hemoglobin. The two-tailed significance level was set at p < .05. All analyses were performed using SAS version 9.4.6 (SAS Institute Inc., Cary, NC) and R version 3.6.1 (2019-07-05).

| Participants' baseline characteristics
The characteristics of the population are given in Table 1

| Associations between hypertension classes and markers of target organ damage
In models adjusted for gender, age, waist circumference, smoking status, total cholesterol, glycemia, lipid lowering agents, self-declared hypertension status, glomerular filtration rate and hemoglobin, the majority of associations in the "crude model" between IDH and markers of target organ damage became no longer significant (Table 2). In particular, with normotension as reference, IDH was not significantly associated with NTproBNP levels (p = . NTproBNP concentrations (all p ≤ .010). Using spline-based analyses, we did not find any evidence of a nonlinear association between hypertension classes (compared to normal BP) and markers of target organ damage, except for the association between SDH and NTproBNP ( Figure 2). Sensitivity analysis based on office BP measurements showed similar findings (Table S1).

| DISCUSSION
The key finding of our study was that an isolated increase in diastolic BP, without a concurrent rise in systolic values at 24-h ABP recording, was not significantly associated with markers of target organ damage.
Concurrently, we showed that combined hypertension was associated T A B L E 1 (Continued)  We also demonstrated that, with normotension as reference, combined hypertension was associated with a higher NTproBNP concentration, meanwhile both SDH and ISH, but not IDH, were correlated with LV mass index and PWV increase. Prior observations suggested that LV mass is more closely related to systolic BP, whereas LV wall thickness correlates better with diastolic BP. 32,35,36 Our finding is in line with previous evidences and corroborates the hypothesis that wall stress is mainly related to systolic BP and is a key determinant of LV hypertrophy development. 37  In our analysis we found a trend toward the association between GLS change and IDH. GLS has been demonstrated as a sensitive tool to recognize early subclinical systolic dysfunction in newly diagnosed hypertensive patients without LVH, 42 even when ejection fraction and other strain components are normal. 43 Interestingly, a recent echocardiographic study conducted on STANISLAS cohort identified a significant association between layer-specific strain variables and self-reported dyspnea, suggesting that refined strain components (i.e., subendocardial strain) could help identify early stages of increased LV filling pressure. 44 Altogether, our results appear to be consistent with previous observations showing a lack of association between IDH and target organ damage. These findings might also provide an explanation for the inconclusive association between IDH and adverse cardiovascular outcomes frequently reported in the scientific literature.

| Limitations
The main limitation of our study is its observational design; therefore, it is not possible to establish a causal link to the results obtained.
In addition, given our sample size, we could not adjust our analysis for every possible cardiovascular risk variable. The cross-sectional design allowed us to assess only intermediate signs of target organ damage; therefore, our findings cannot be extrapolated to the incidence of hard cardiovascular or renal end-points. We had no assessment of microvascular damage besides microalbuminuria and therefore cannot rule out the impact of IDH on this target. The reclassification of our population according to ESC/ESH BP thresholds 3 identified a very small number of IDH patients (N = 11), preventing any reliable sensitivity analysis in this setting ( Figure S1). Different duration of hypertension among subgroups might underestimated the effect of BP on TOD in patients with a shorter follow-up, although it was not statistically significant. Finally, the results of this analysis cannot be extended to general hypertensive population, as they refer to an initially healthy population with a low cardiovascular risk at the time of evaluation and constituted by a mix of already diagnosed and newly-diagnosed hypertensives during STANISLAS research visit.

| Conclusions
In our study IDH was not significantly associated with target organ damage. This finding might suggest that the clinical significance of IDH in the absence of elevated systolic BP is questionable. Further studies are needed to clarify the causative role of IDH in the development of target organ damage and cardiovascular outcomes.