Introduction

Several cohort-based and epidemiological studies exploring cardiac diastolic performance have shown marked differences in the prevalence of left ventricular diastolic dysfunction (LVDD) in adult populations. The estimated rate of LVDD varies between 24% to 60%, and the reported inconsistencies may be partially explained by various factors including the assorted characteristics of studied populations as well as different diagnostic criteria applied [1–4]. The latter factor evolved considerably over the course of the last years [5,6]. As demonstrated by Negri et al. the prevalence of LVDD assessed by conventional vs. tissue Doppler imaging-based indices (TDI) was approximately two-fold higher when TDI-criteria were employed [3]. It is important to mention that a TDI-based evaluation is a standard in LVDD assessment in today’s guidelines [5,6].

Hypertension may lead to an increase in the left ventricular mass and a worsening of LV elasticity, which in turn results in impaired LV myocardial relaxation – a precedent of augmented LV filling pressures [7]. Hypertension is also one of the most important risk factors in the development of congestive heart failure (HF), a condition where at least half of HF patients are characterized by a preserved left ventricular systolic function (HFpEF) [8–10]. It needs to be underscored that clinical symptoms of HFpEF patients are tightly related to the progression of LVDD [7]. Therefore, there is a growing interest in a better understanding of early-stage and subclinical LVDD [7,11,12].

Although the prevalence of LVDD in the general population has been addressed in several cross-sectional and prospective studies, there are only a few analyses derived from large hypertensive cohorts [13–15]. Additionally, there is an evident shortage of data on LVDD occurrence documented in large hypertensive populations from recent years. Therefore, the aim of this study was to assess both the occurrence and the severity of LVDD in a large group of treated patients with established hypertension.

Methods

We retrospectively analyzed records of 610 patients (46% females) consecutively recruited from participants of the CARE NORTH Study between 2009 and 2011 who consented to echocardiographic recordings, and who did not meet exclusion criteria. At the time of the recruitment, participants of the CARE NORTH Study were regular ambulatory patients of the Department of Hypertension and Diabetology, Medical University of Gdańsk (MUG) Teaching Hospital, Gdańsk, Poland. In short, the CARE NORTH Study (N = 853) is a prospective observation aiming at clinical, biochemical, genetic, and metabolomic detailed phenotyping of a large cohort of patients with established hypertension. The recruitment to the core sample was random and was performed within the tertiary care outpatient clinic of the Department of Hypertension and Diabetology, MUG, Gdańsk. The exclusion criteria for the present study were as follows: established diagnosis of heart failure with reduced ejection fraction (HFrEF), asymptomatic left ventricular ejection fraction (LVEF) of <45%, hemodynamically significant valvular heart disease, coronary artery disease (after myocardial infarction, or requiring angioplasty/CABG), stroke, or atrial fibrillation (AFib).

All CARE-NORTH Study patients were subjected to structured medical surveys, which were followed by blood sampling for laboratory tests, and additional recordings. Ambulatory blood pressure was measured with SpaceLabs 90207 and SpaceLabs 90217 devices in 20 min daytime-, and 30 min nighttime intervals within the same month as the heart ultrasound.

Transthoracic echocardiographic (ECHO) recordings were performed with a Vivid 7 Pro™ GE device using a 2.0–3.6 MHz transducer. All ECHO measurements were obtained during relaxed breathing after several minutes of rest. All recordings included at least 3 cardiac cycles and were digitally stored for off-line analysis by two experienced cardiologists. Although our centre is not equipped with standard cardiovascular core laboratory, all echocardiographic data were recorded in the laboratory dedicated to research purposes solely; participants were not merged with regular patients; the recordings were performed with strict protocol and were used for presented research solely.

Measurements from 2D echocardiographic images were used to assess LV size, IVSd and PWd thickness according to European and American Society of Echocardiography recommendations [16]. LV linear measurements were also used to calculate LVEF. Relative wall thickness (RWT) was calculated with the following formula: 2xPWD/LV internal diameter at end-diastole.

LVMI was calculated according to ASE recommendations [16]. LA volume index (LAVI) was measured using the disk summation algorithm [17]. The mitral inflow was recorded from the apical four-chamber view with the sample volume at the tips of leaflets to obtain early (E) and atrial (A) trans-mitral flow velocities, the E/A ratio, and deceleration time of E (DecT E). The isovolumetric relaxation time (IVRT) was derived by placing the cursor Doppler cursor in the LV outflow tract to simultaneously display the end of aortic ejection and the onset of mitral inflow. Tissue Doppler imaging (TDI) was performed by the apical 4-chamber view. Early diastolic mitral annular velocity E’ was derived from averaged velocities of the lateral and septal curves. The ratio between E velocity and E’ velocity (E/E’ ratio) was calculated simultaneously.

Algorithm-based LV diastolic performance assessment

The algorithm for the LV diastolic dysfunction diagnosis was based on a document categorizing DD published in 2009 [5]. As the 2009 guidelines did not include recording of tricuspid regurgitation, we were unable to comply with the most recent standards, which define a peak TR velocity as one of the prerequisites to assess LVDD. We used the mitral inflow E/A, TDI velocities (E’septal, E’lateral, E/E’) and left atrium volume index (LAVI) to assess the stages of LV diastolic dysfunction (LVDD). Except for two exemptions which are detailed below, patients with LAVI ≥34 mL/m², and either E’septal <8 cm/s or E’lateral <10 cm/s were furhter classified as LVDD. The classification by specific subgroups of LVDD was as follows: stage 1 LVDD was defined as (all inclusive): (1) LAVI equal of greater than 34 mL/m² (the only exemption was LAVI <34 mL/m² coinciding with increased IVRT >100 ms, and/or DecT E > 200 ms), (2) transmitral E/A ratio of <0.8, (3) normal LV filling pressure denoted by (E/E’ ≤ 8). Patients were stratified to stage 2 LVDD based on the following criteria (all conditions had to must be met): (1) E/A ratio between 0.8 and 2.0, (2) LAVI equal to or greater than 34 ml/m², (3) E/E’ ≥ 9 but <13. The diagnostic criteria for LVDD stage 3 were as follows: (1) E/E’ ratio ≥13 and E/A ratio ≥2.

The so-called “athlete’s heart” characteristic was defined as E' septal ≥8, and E' lateral ≥10 and the LAVI equal to or greater than 34 mL/m² (these patients were classified as normal LV diastolic function). Lastly, there was a subgroup of patients where it was impossible to clearly determine their LV diastolic status based on the recommendations. These patients where left as such, and labeled as “non-defined diastolic dysfunction NDDD” (e.g. female patients with: (1) E/E’ suggesting increased LAP, (2) E/A suggesting mitral inflow pseudonormalisation, and (3) LAVI within normal range).

Results

In the entire group, mean hypertension duration was 13.7 ± 9.7 years, and antihypertensive treatment initiation delay was 1.9 ± 4.5 of years. The clinical characteristic of the study group is presented in Table 1.

The reproducibility of data was assessed with relation to key parameters used in the LV diastolic function evaluation i.e. left atrium volume, E-wave, A-wave, E’septal, and E’lateral. Both ICC and ρc were greater than 0.95 indicating the very good agreement between the observers (these results practically ruled out a potential bias related to data acquisition).

The detailed echocardiographic characteristic with relation to sex is summarized in Table 2. Overall, 49.7% of the patients had normal diastolic function. Grade 1 LVDD was observed in 24.4% whereas grade 2 LVDD was found in 19.3% of the patients. None of the studied patients were characterized by grade 3 LVDD, however, a subset of patients scored ambiguously and were classified as NDDD (6.6%). There was a clear-cut difference in the distribution of abnormal diastolic function between females and males, which is summarized in the Figure 1.

Prevalence and distribution of left ventricular diastolic dysfunction in treated patients with long-lasting hypertension

All authors

Ewa Świerblewska, Jacek Wolf, Katarzyna Kunicka, Beata Graff, Katarzyna Polonis, Michał Hoffmann, Marzena Chrostowska, Anna Szyndler, Piotr Bandosz, Beata Graff & Krzysztof Narkiewicz

https://doi.org/10.1080/08037051.2018.1484661

Published online:

21 August 2018

Figure 1. Distribution of left ventricle diastolic function phenotypes with relation to sex, and the presence of diabetes mellitus (DM). Legend: N – normal diastolic function; N1 – athletic heart (regarded as a norm variation; see Ref.No.5); LVDD 1 – LV diastolic dysfunction grade 1; LVDD 2 – LV diastolic dysfunction grade 2; NDDD – unclassified diastolic dysfunction (see methods for details).

Display full size

Both males and females with LVDD were older with longer duration of hypertension as compared to patients with normal diastolic function (Table 3). There was a consistent difference observed in both sexes related to comorbid type 2 diabetes mellitus. Diabetes coincided approximately 2-to-3 fold more often in patients with LVDD as compared to hypertensive patients with normal LV function (Table 3).

Several other differences in clinical characteristics related to diastolic function status were found between females and males (Table 3). Also noteworthy was a gradual increase in the number of antihypertensive drugs administered alongside with deteriorating LV diastolic function, in both sexes.

ROC curve analysis identified BMI >29.9 (27.4-31.8 [95%CI]) kg/m² as the most accurate cut-off point to identify the presence of LVDD (AUC =0.613 ± 0.02[SE]; p < 0.001).

The logistic regression analysis, which aimed to identify factors associated with the presence of LVDD is presented in Table 4. Variables pushed into the model included age, female sex, obesity (BMI ≥30), diabetes mellitus, 24-hour ambulatory SBP, hypertension duration, LVMI, number of antihypertensive drugs, and pharmacological treatment including thiazide diuretics, aldosterone antagonists, beta₁-blockers, ACEIs or ARBs, calcium channel blockers, and alfa-blockers (Table 4).

Discussion

There are three main findings from our study. First, diastolic dysfunction is common in treated hypertensives sans overt cardiovascular disease. Second, there are sex-specific differences in the prevalence and predictors of diastolic dysfunction. Finally, very high prevalence of LVDD occurs despite long-term antihypertensive treatment and reasonable blood pressure control, yet part of this phenomenon may be ascribed to highly prevalent diabetes in our cohort.

Previous studies in different populations have shown that LV performance can be influenced by several factors including female sex, obesity, high blood pressure, diabetes mellitus, LV geometry, and certain drugs [4,7,10,20–23]. However, the clustering of these factors, and in particular, their interaction with aging may cause interpretation of their independent contribution to LVDD to be more difficult. First, aging might affect the link between gender and LVDD. Such a relationship was observed in younger subjects [10,20], but not in persons over the age of 65 years [2]. Second, it is not always easy to discriminate the direct effects of hypertension per se from blood pressure elevation reflecting aging process [4,21]. Lastly, both the use of antihypertensive treatment reducing cardiovascular risk and the level of blood pressure control [7,13] should be taken into account. We decided to address these issues in our study performed in treated hypertensive patients within relatively narrow age range.

In accordance with previous studies, we observed a very close independent relationship between LVDD and aging despite ongoing antihypertensive treatment. In the study by Abhayaratna et al. who assessed the prevalence and predictors of left ventricle diastolic dysfunction in an older population [11], approximately half of this studied population had hypertension. These data clearly showed that advancing age is one of the main predictors of the presence of diastolic dysfunction. Also the results of a population-based study from Central Italy support our findings [2]. In another study, Negri et al. have shown that factors involved in diastolic dysfunction development may differ across various life-decades [14]. In the study by Wu et al., age, history of hypertension, BMI, TG, IVST and PWV⩾ 16ms⁻¹ were the independent predictors of mild LVDD [20].

In our study, the prevalence of diastolic dysfunction was especially evident in hypertensive women in whom only 38.8% of them were characterized by a normal diastolic function. Furthermore, grade 2 diastolic dysfunction was much more frequent in females (24.9%) than in males (14.6%). Lack of third degree LVDD in our cohort may be related to reasonable control of blood pressure and common treatment with RAA system blocking agents including both ACEIs or ARBs, beta₁-blockers, and aldosterone receptor antagonists (Table 1).

Importantly, there were sex-specific differences in the factors associated with diastolic dysfunction. The impact of obesity was more evident in the more advanced LVDD-group in females which was not evident in males. These findings are in line with the results from the Framingham study showing that women are at similar risk of HFpEF and HFrEF, and that men, by contrast, are at increased risk of HFrEF [24]. The mechanisms underlying these sex-related differences are unclear. However, the prospective MESA study demonstrated that women and men undergo different patterns of LV remodeling, which might be explained by lower LV chamber volume, smaller number of myocytes, increased sensitivity for diffuse interstitial fibrosis or different LV responses to activation of the renin-angiotensin aldosterone system in women and men [25]. Furthermore, serum levels of several biomarkers were distinctly different in women compared with men and had differential effects on left ventricular structure and function. It has been suggested that coronary microvascular inflammation, promoting fibrotic processes which may lead to diastolic dysfunction development and progression in subjects with obesity, hypertension, diabetes mellitus, chronic obstructive pulmonary disease, anemia, and chronic kidney disease [26]. Whether this link is related to gender remains to be elucidated.

One of the possibilities which could explain very high prevalence of the LVDD reported in our study is relatively high representation of diabetes in the CARE NORTH cohort. The recruitment of the participants to the core CARE NORTH group was carried out in the outpatient clinic of our Department (Department of Hypertension and Diabetology) where we see hypertensive and diabetic patients requiring tertiary care. Thus, the percentage of diagnosed diabetes does not necessarily reflect the overall prevalence of diabetes in the general and hypertensive population. The link between the diabetes and LVDD is strongly supported by our results. The multivariate analysis showed the highest odds ascribed to the diagnosed diabetes when compared to any other covariate in the logistic regression model (Table 4). Also the prevalence of different forms of LVDD is the highest in the group of diabetics affecting approximately two thirds of patients (Figure 1, lower panel). Diabetes or hyperglycemia itself have been long-recognized conditions to promote negative processes implicated in both vessels’, and cardiac structural, and functional remodeling [27,28]. Several detrimental pathogenic mechanisms which link dysglycemia to the development of LVDD were identified including interstitial deposits of advanced glycosylation end products (AGEs) within the myocardium [29]. Autopsies of the humans’ hearts revealed greater cardiac weight, interstitial fibrosis, replacement fibrosis, and perivascular fibrosis; the processes most severe evident when diabetes and hypertension coexist [30]. Additionally to collagen perivascular and in between myofibers’ accumulation, the left ventricle of diabetics is characterized by enhanced triglyceride and cholesterol concentrations [31]. Moreover, cardiomyocytes lipotoxicity and direct effects of insulin may lead to cellular apoptosis, endothelial dysfunction and chronic adrenergic stimulation [32]. Finally, the insulin resistance [33], excess in visceral adiposity [34], or the enhanced activities of circulating enzymes (DPP-4) and fatty acids-binding proteins may further complicate matters [35].

We observed a very high prevalence of LVDD despite long-term treatment and reasonable blood pressure control. The LIFE sub-study has shown that antihypertensive treatment in patients with hypertension and with electrocardiographic LV hypertrophy, has been shown to improve transmitral flow patterns, but it was not associated with reduced cardiovascular morbidity and mortality [13]. The ASCOT sub-study analysis of a cohort of patients with well-controlled hypertension and three cardiac risk factors showed that E/E’ ratio is an important predictor of cardiac outcomes [36]. In our study the presence of LVDD was related to the on-treatment blood pressure level (Table 4), but not to the duration of hypertension therapy. Taken together, earlier studies showed that hypertensive individuals with controlled blood pressure have higher cardiovascular risk compared to normotensive persons presenting similar blood pressure. Our findings suggest that this residual risk may be attributable, at least in part, to a high burden of diastolic dysfunction despite antihypertensive treatment.

The incidence and prevalence of heart failure with preserved left ventricular ejection fraction EF (HFpEF) have dramatically increased during the last 15 years [4,37]. Interestingly, the survival rates improved among HF patients with reduced EF, probably reflecting continuous progress in both invasive and pharmacological treatment of coronary heart disease (CHD). In contrast, mortality in HFpEF [9] prompted more studies focusing on both pre-clinical and clinical LVDD. Patients with HF with preserved EF are older, are more likely to be female, and have more severe hypertension, obesity, and anemia than those with HF with reduced EF [38]. In our study, we excluded patients with overt HF. Nevertheless, we observed diastolic dysfunction in almost half of our patients. Previous studies assessing diastolic dysfunction focused primarily on patients with overt CHF [39,40], the elderly [2,11], population-based cohorts [1,39], untreated hypertensives [41] or solely patients with left ventricular hypertrophy [13,41], and the reported prevalence of LVDD varied between 24% and 65%. These differences in the reported prevalence might be related not only to the clinical characteristics of the various cohorts, but might reflect a sensitivity to the applied LVDD assessment [2,11,22]. The majority of these studies used previous [2,37,42] or partially modified criteria [1,21] of LVDD. At the time of our study initiation we had decided to apply the 2009 LVDD diagnostic criteria, however, these algorithms were updated in 2015, and 2016. This would be interesting to see to what particular extent the older vs. newer criteria influence the grouping of the patients, however, it was impossible to perform such analysis as peak TR velocity was not recorded. Nevertheless, we would like to underline that in our study performed in patients sans overt cardiovascular complications, we have applied well-defined criteria of grading LVDD using parameters derived from Tissue Doppler Imaging (TDI) [5]. This allowed for more precise identification of subjects with impaired diastolic function as compared to studies based on conventional echocardiography protocols [43].

Limitations

Our Centre was not equipped with a core laboratory at the time of the patients recruitment. Another formal shortcoming of our study is lack of the CARE NORTH registration at ClinicalTrials.gov. Although, there is an ongoing vivid discussion related to the proposed LVDD diagnostic algorithms [44,45] we are aware of the fact that the applied 2009 diagnostic recommendations may be perceived as the sizable study limitation.

Conclusions

There is an unexpectedly high prevalence of different forms of diastolic dysfunction in middle-aged treated hypertensive patients who are free of overt cardiovascular disease. Whether such derangements may predict the development of congestive heart failure remains to be determined in prospective studies.