The influence of left ventricular geometry on left atrial phasic function in hypertensive patients

Abstract

Objective. We aimed to investigate left atrial (LA) phasic function in hypertensive patients with different geometric patterns using two-dimensional (2DE) and three-dimensional (3DE) echocardiography. Methods. This cross-sectional study included 213 hypertensive subjects who underwent a complete 2DE and 3DE examination. The new updated criteria for left ventricular (LV) geometry, which consider LV mass index, LV end-diastolic diameter and relative wall thickness, were applied. According to this classification, the subjects were divided into six groups: normal geometry, concentric remodeling, eccentric non-dilated left ventricular hypertrophy (LVH), concentric LVH, dilated LVH and concentric-dilated LVH. Results. 2DE and 3DE LA volumes gradually increased from normal LV geometry to concentric and concentric-dilated LVH. LA reservoir and conduit functions, estimated by 2DE and 3DE LA total and passive emptying fractions, were decreased in subjects with concentric and concentric-dilated LVH. LA booster pump function was increased in patients with concentric, dilated and concentric-dilated LVH compared to subjects with normal LV geometry. The same results regarding LA phasic function were provided by 2DE strain analysis. Concentric, dilated and non-concentric dilated LVH were associated with LA enlargement independently of main demographic and clinical features. Conclusion. LV geometric patterns significantly influence LA phasic function. Concentric and dilated LVH patterns have the most prominent negative effect on LA enlargement assessed by both 2DE and 3DE.

Key Words::

• Hypertension
• left atrium
• left ventricular geometry
• three-dimensional echocardiography
• two-dimensional speckle tracking

Introduction

The left ventricular (LV) geometry, especially concentric left ventricular hypertrophy (LVH), has long been recognized as an important risk factor for cardiovascular morbidity and mortality in patients with hypertension (1,2). The authors of the Dallas study recently suggested a new version of LV geometry classification including a new parameter: LV dilatation (3). In this way, eccentric and concentric LVH groups were divided into two subgroups: dilated and non-dilated. Using the updated criteria, Bang et al. showed that patients with eccentric dilated and both concentric non-dilated and dilated LVH had increased risks of all-cause or cardiovascular mortality, compared with subjects with normal LV mass, while the eccentric non-dilated LVH group did not (4). The hazard ratio of the composite end-point, defined as the first event of stroke, myocardial infarction, heart failure or cardiovascular death, gradually increased from non-dilated LVH, through dilated LVH, to patients with dilated and non-dilated concentric LVH (4). More recently, two large studies revealed similar results using a linear method of LV dilatation assessment (5,6).

Previous studies in large populations demonstrated that LV geometry has a significant impact on left atrial volume (LAV) (7–9). The authors showed that abnormal LV geometric patterns, especially eccentric and concentric LVH, were independently associated with left atrial (LA) enlargement (9). However, there are also investigators who claim that LA size is determined by LV mass but not LV geometry in patients with recently diagnosed hypertension (10). All of the aforementioned studies investigated LA remodeling using solely two-dimensional echocardiography (2DE).

The purpose of the present study was to investigate LA phasic function and mechanics, by 2DE and three-dimensional echocardiography (3DE), in hypertensive patients with different LV geometric patterns according to the Dallas criteria, as well as the possible relationship between LA enlargement and different LV patterns.

Methods

This cross-sectional study enrolled 213 hypertensive patients. Arterial hypertension was diagnosed according to the current guidelines (11). Clinic blood pressure values were obtained on two separate visits, 3 weeks apart. Blood pressure was measured by a conventional sphygmomanometer in the morning by taking the average value of two consecutive measurements in the sitting position, 10 min apart. Blood pressure was calculated as average values of all the measurements.

Exclusion criteria were symptoms or signs of heart failure, coronary artery disease, previous cerebrovascular events, atrial fibrillation, congenital heart disease, valvular heart disease (more than mild), neoplastic disease, liver cirrhosis, kidney failure or endocrinological diseases including type 2 diabetes mellitus. Subjects with unsatisfactory 3DE acquisitions (nine participants) were also excluded from any further analyses.

Anthropometric measures (height and weight) and laboratory analyses (level of fasting glucose, blood creatinine, total cholesterol and triglycerides) were obtained from all the subjects included in the study. Body mass index (BMI) and body surface area (BSA) were calculated for each patient. The local ethics committee approved the study, and informed consent was obtained from all participants.

Echocardiography

Echocardiographic examinations were performed using a commercially available Vivid 7 (GE Vingmed, Horten, Norway) ultrasound machine equipped with both a 2.5 MHz transducer and a 3 V matrix probe for 3DE data-set acquisitions.

Reported values of all 2DE parameters were obtained as the average value of three consecutive cardiac cycles. Left ventricular end-diastolic (LVIDd) and end-systolic diameters, posterior wall (PWT) and interventricular septum thickness (IVS) were measured according to the recommendations for chamber quantification (12). Relative wall thickness (RWT) was calculated according to the formula: (2 × posterior wall thickness)/LV end-diastolic diameter. LV ejection fraction (EF) was calculated using the biplane method. Left ventricular mass (LVM) was calculated using the corrected American Society of Echocardiography method: 0.8 × (1.04 × [(LVIDd + IVSd + PWTd)³ - LVIDd³]) + 0.6 (13), and indexed for the BSA (LVMI).

Cut-off values for LVIDd, LVMI and RWT were taken from the recently published guidelines (13). The upper limit of LVIDd for women is 5.2 cm and for men is 5.8 cm; the cut-off for LVMI for women is 95 g/m² and 115 g/m² for men; and the cut-off for RWT is 0.42.

Although the original Dallas classification of LV geometric patterns includes determination of LV volume (3), we decided to use a linear method of LV dilatation estimation that has been recently assessed in large cohorts of hypertensive subjects and demonstrated no differences in mortality from the original Dallas criteria (5,6). All subjects were divided into six groups: (i) normal LV geometry (normal LVMI, LVIDd and RWT); (ii) concentric remodeling (normal LVMI and LVIDd, increased RWT); (iii) eccentric non-dilated LVH (increased LVMI, normal LVIDd and RWT); (iv) concentric LVH (increased LVMI and RWT, normal LVIDd); (v) dilated LVH (increased LVMI and LVIDd, normal RWT); and (vi) concentric-dilated LVH (increased LVMI, LVIDd and RWT). However, owing to the relatively small study group and further statistical analysis, we merged dilated and concentric-dilated LVH into one group.

LAV was measured just before mitral valve opening. LAV was determined according to the biplane area-length method in four- and two-chamber views, and indexed for BSA.

Pulsed-wave Doppler assessment of transmitral LV was obtained in the apical four-chamber view according to the guidelines (14). Tissue Doppler imaging was used to obtain LV myocardial velocities in the apical four-chamber view, with a sample volume placed at the septal and lateral segments of the mitral annulus during early and late diastole (e′ and a′). The average of the peak early diastolic relaxation velocity (e′) of the septal and lateral mitral annulus was calculated, and the E/e′ ratio was computed.

Two-dimensional echocardiographic assessment of left atrial volumes and function

LAVs were measured in three different sequences of the cardiac cycle: maximal LAV was measured just before mitral valve opening, pre-atrial contraction (pre-A) LAV was determined at the onset of atrial systole (peak of P wave in ECG) and minimal LAV was measured at mitral valve closure. All the volumes were determined according to the biplane method in four- and two-chamber views, and all the values were indexed for BSA. The total emptying volume, which represents the LA reservoir function, was calculated as the difference between maximum and minimum LAV; the passive emptying volume, which represents the conduit function, was calculated as the difference between maximum and pre-A LAV; and the active emptying volume, which corresponds to LA booster function, was calculated as the difference between pre-A and minimum LAV (15). Accordingly, the total emptying fraction was calculated as the ratio between total emptying volume and maximum LAV; the passive emptying fraction was computed as the ratio between passive and maximum; and the active emptying fraction was calculated as the proportion between active and pre-A LAV.

2DE strain imaging was performed in the apical four- and two-chamber views (16,17), and commercially available software (EchoPAC 112; GE-Healthcare, Horten, Norway) was used for the 2DE strain analysis. LA strain was calculated from the peak of the P wave to the peak of the P wave in the next cardiac cycle. The LA endocardium was manually traced. An average longitudinal strain curve was automatically generated, and it included a negative deflection (LA negative longitudinal strain) representing LA active contraction, followed by a positive one during LA filling (LA positive longitudinal strain). Their summation represented total LA longitudinal strain. LA strain (positive, negative and total) was calculated by averaging the values obtained in four- and two-chamber apical views.

Statistical analysis

All the parameters were tested for normal distribution using the Kolmogorov-Smirnov test. Continuous variables were presented as mean ± standard deviation and compared by the analysis of equal variance (ANOVA), as they showed a normal distribution. Bonferroni post hoc analysis was used for the comparison between different groups. The differences in proportions were compared using the chi-squared test or Fisher's exact test. The correlation between different LV geometric patterns and impaired 2DE and 3DE strains, independent of age, BMI, systolic blood pressure level and LV mass, was determined by multivariate logistic regression [odds ratio (OR) and 95% confidence interval (CI)]. The cut-off value for the 2DE left atrial volume index (LAVI) was 34 ml/m², according to the current recommendation (13). Because of the lack of guidelines regarding a cut-off value for the 3DE LAVI, we used data from an available study with relatively large number of subjects (LAV_max/BSA = 33 ml/m²) (18). A p value < 0.05 was considered statistically significant.

Results

Using the modified Dallas criteria, we identified 91 hypertensive subjects with normal LV geometry, 31 with concentric LV remodeling, 44 with eccentric non-dilated LVH, 33 with concentric LVH and 14 with dilated or concentric-dilated LVH. Patients with concentric, dilated and concentric-dilated LVH were older than subjects with normal geometry or concentric remodeling (Table I). There were no gender differences among groups. BMI was higher in concentric LVH than in the normal geometry group. Blood pressure values progressively increased from normal LV geometry to dilated or concentric-dilated LVH. The percentage of treated subjects also increased in the same direction. Serum creatinine and total cholesterol levels were similar among the different groups, while patients with concentric LVH had higher levels of glucose than those with other LV geometric patterns (Table I).

Table I. Demographic characteristics and clinical parameters of the study population.

	Normal LV geometry (n = 91)	Concentric LV remodeling (n = 31)	Eccentric non-dilated LVH (n = 44)	Concentric LVH (n = 33)	Dilated and concentric-dilated LVH (n = 14)	p
Age (years)	48 ± 8^a,b	49 ± 7^c	51 ± 9	54 ± 8^a	58 ± 7^b,c	< 0.001
Female (%)	45 (49)	12 (39)	28 (64)	15 (45)	4 (29)	0.107
BMI (kg/m²)	25.9 ± 3.2^p	26.6 ± 3	27.1 ± 3.4	27.9 ± 3.7^p	28.1 ± 4.3	0.017
BSA (m²)	1.91 ± 0.18^d	1.94 ± 0.19	2 ± 0.17^d	1.99 ± 0.18	2.01 ± 0.22	0.031
Clinic systolic BP (mmHg)	136 ± 12^a,b,e	139 ± 14^c,s	145 ± 13^e,r	150 ± 15^a,s	157 ± 15^b,c,r	< 0.001
Clinic diastolic BP (mmHg)	81 ± 8^a,b	84 ± 8^f	84 ± 7^r	89 ± 10^a	92 ± 9^b,f,r	< 0.001
Antihypertensive drugs (%)	16 (18)^a,b,e	7 (23)^c,g	19 (43)^e	20 (60)^a,g	10 (71)^b,c	< 0.001
Plasma glucose (mmol/l)	5 ± 0.8^a	5.1 ± 0.7^s	5.2 ± 0.6^u	5.7 ± 0.8^a,s,u	5.6 ± 0.9	< 0.001
Serum creatinine (μmol/l)	66 ± 9	65 ± 8	68 ± 9	66 ± 8	72 ± 9	0.091
Triglycerides (mmol/l)	1.7 ± 0.3^a	1.9 ± 0.4	1.8 ± 0.4	2 ± 0.5^a	2 ± 0.4	< 0.001
Total cholesterol (mmol/l)	5.1 ± 0.8	5.2 ± 0.7	5.2 ± 0.9	5.5 ± 1	5.6 ± 1.1	0.097

LV, left ventricular; LVH, left ventricular hypertrophy; BMI, body mass index; BSA, body surface area; BP, blood pressure.

^ap < 0.01 for Normal LV geometry vs Concentric LVH; ^bp < 0.01 for Normal LV geometry vs Dilated LVH; ^cp < 0.01 for Concentric LV remodeling vs Dilated and concentric-dilated LVH; ^dp < 0.05 for Normal LV geometry vs Eccentric non-dilated LVH; ^ep < 0.01 for Normal LV geometry vs Eccentric non-dilated LVH; ^fp < 0.05 for Concentric LV remodeling vs Dilated and concentric-dilated LVH; ^gp < 0.01 for Concentric LV remodeling vs Concentric LVH; ^pp < 0.05 for Normal LV geometry vs Concentric LVH; ^rp < 0.05 for Eccentric non-dilated LVH vs Dilated and concentric-dilated LVH; ^sp < 0.05 for Concentric LV remodeling vs Concentric LVH; ^up < 0.05 for Eccentric non-dilated LVH vs Concentric LVH.

Echocardiographic conventional parameters

The definition of various LV geometric patterns, according to the Dallas classification, determined the difference in LV diameters, interventricular septum, posterior and relative wall thickness, as well as LVMI between the observed groups (Table II). Participants with dilated and concentric-dilated LVH had somewhat lower LV ejection fraction than those with normal LV geometry (Table II).

Table II. Two-dimensional echocardiographic parameters of left ventricular (LV) structure and function in the study population.

	Normal LV geometry (n = 91)	Concentric LV remodeling (n = 31)	Eccentric non-dilated LVH (n = 44)	Concentric LVH (n = 33)	Dilated and concentric-dilated LVH (n = 14)	p
LV end-diastolic diameter (mm)	49.2 ± 4	46.8 ± 3.4	53.4 ± 3.9	51.9 ± 4	56.6 ± 4.3	< 0.001*
LV interventricular septum thickness (mm)	9.1 ± 0.8	10.5 ± 1	10.4 ± 0.9	11.6 ± 1.1	12.3 ± 1.1	< 0.001^†
Posterior wall thickness (mm)	9 ± 0.7	10.2 ± 0.9	10.1 ± 1	11.5 ± 1.2	12.1 ± 1.1	< 0.001^†
Relative wall thickness	0.37 ± 0.03^a,b,h	0.44 ± 0.02^h,m	0.38 ± 0.03^m,n,o	0.45 ± 0.03^a,n	0.43 ± 0.03^b,o	< 0.001
LV mass (g)	155.1 ± 12.1	171.2 ± 13.5	209.7 ± 13.2	235 ± 21	292 ± 27	< 0.001^‡
LV mass index (g/m²)	81.2 ± 6.2	88.2 ± 6.2	104.8 ± 7.4	118.2 ± 11	145.2 ± 14	< 0.001^‡
LV ejection fraction (%)	65 ± 5^a,q	63 ± 4	64 ± 3	62 ± 4^a	61 ± 4^q	0.001
E/A ratio	1.09 ± 0.29^a,b,e	1.02 ± 0.27^c,g,y	0.81 ± 0.24^e,y	0.75 ± 0.26^a,g	0.72 ± 0.25^b,c	< 0.001
E/e′ ratio	7.4 ± 2.6	8.9 ± 3.2	9.5 ± 3.4	12.2 ± 3.9	13.7 ± 3.8	< 0.001**

LVH, left ventricular hypertrophy; E, early diastolic mitral flow (pulsed Doppler); A, late diastolic mitral flow (pulsed Doppler); e′, average of the peak early diastolic relaxation velocity (e′) of the septal and lateral mitral annulus (tissue Doppler).

*p < 0.05 for all comparisons except for Eccentric non-dilated LVH vs Concentric LVH or Dilated and concentric-dilated LVH; †p < 0.01 for all comparisons except for Concentric LV remodeling vs Eccentric non-dilated, and Concentric LVH vs Dilated and concentric-dilated LVH; ‡p < 0.01 for all comparisons; **p < 0.05 for all comparisons except for Normal LV geometry vs Concentric LV remodeling, Concentric LV remodeling vs Eccentric non-dilated, and Concentric vs Dilated and concentric-dilated LVH; ^ap < 0.01 for Normal LV geometry vs Concentric LVH; ^bp < 0.01 for Normal LV geometry vs Dilated and concentric-dilated LVH; ^cp < 0.01 for Concentric LV remodeling vs Dilated and concentric-dilated LVH; ^ep < 0.01 for Normal LV geometry vs Eccentric non-dilated LVH; ^gp < 0.01 for Concentric LV remodeling vs Concentric LVH; ^hp < 0.01 for Normal LV geometry vs Concentric LV remodeling; ^mp < 0.01 for Concentric LV remodeling vs Eccentric non-dilated LVH; ⁿp < 0.01 for Eccentric non-dilated LVH vs Concentric LVH; ^op < 0.01 for Eccentric non-dilated LVH vs Dilated and concentric-dilated LVH; ^qp < 0.05 for Normal LV geometry vs Dilated and concentric-dilated LVH; ^yp < 0.05 for Concentric LV remodeling vs Eccentric non-dilated LVH.

The mitral E/A ratio progressively decreased, whereas the E/e′ ratio increased, from normal LV geometry to the dilated LVH group (Table II).

Two-dimensional echocardiographic left atrial phasic function and deformation

Maximum, minimum and atrial precontraction LAVIs progressively increased from normal geometry to dilated and concentric-dilated LVH (Table III). Total LAV, which represents LA reservoir function, was larger in dilated and concentric-dilated LVH than in normal LV geometry (Table III). Passive LAV, a parameter of LA conduit function, reduced from normal LV geometry to concentric-dilated LVH. On the other hand, active LAV, an index of LA booster pump function, increased in the same direction. Therefore, the total and passive LA emptying fraction decreased, while the active LA emptying fraction increased from normal LV geometry to concentric-dilated LVH (Table III).

Table III. Two-dimensional echocardiographic left atrial volume (LAV) and mechanical analysis in the study population.

	Normal LV geometry (n = 91)	Concentric LV remodeling (n = 31)	Eccentric non-dilated LVH (n = 44)	Concentric LVH (n = 33)	Dilated and concentric-dilated LVH (n = 14)	p
2DE LAV analysis
LAVmax (ml)	51.4 ± 9.6^a,b	53.5 ± 11.2^f	56.8 ± 14	61.1 ± 15.4^a	66.7 ± 16.8^b,f	< 0.001
LAVmax/BSA (ml/m²)	26.9 ± 4.8^a,b	27.6 ± 5.7^c	28.4 ± 7.1^r	30.7 ± 7.8^a	33.2 ± 8.5^b,c,r	< 0.001
LAVmin (ml)	19 ± 4.3	20.2 ± 4.5	22.2 ± 4.8	25.6 ± 5	29.3 ± 5.3	< 0.001**
LAVmin/BSA (ml/m²)	9.9 ± 2.2	10.4 ± 2.3	11.1 ± 2.3	12.9 ± 2.6	15.6 ± 2.7	< 0.001**
LAVpre-a (ml)	27.9 ± 5.2	30.8 ± 5.5	34.1 ± 5.9	41.3 ± 6.3	47.7 ± 6.7	< 0.001*
LAVpre-a/BSA (ml/m²)	14.6 ± 2.7	15.9 ± 2.9	17 ± 3	20.8 ± 3.2	23.7 ± 3.4	< 0.001*
LA TotEV (ml)	32.4 ± 5.8^q	33.3 ± 5.6	34.6 ± 6.1	35.5 ± 6.5	37.4 ± 7.2^q	0.011
LA PassEV (ml)	23.5 ± 3^a,b,d	22.7 ± 2.8^c,g	21.7 ± 3^d,r	19.8 ± 3.4^a,g	19 ± 3.8^b,c,r	< 0.001
LA ActEV (ml)	8.9 ± 2.6	10.6 ± 2.9	11.9 ± 3.1	15.7 ± 3.5	18.7 ± 4	< 0.001*
LA TotEF (%)	63 ± 7^a,b	62 ± 7^f	61 ± 6	58 ± 5^a	56 ± 5^b,f	< 0.001
LA PassEF (%)	46 ± 6	42 ± 6	40 ± 5	32 ± 5	28 ± 4	< 0.001¶
LA ActEF (%)	32 ± 5^a,b	34 ± 5^d	35 ± 6^d	38 ± 7^a	39 ± 7^b	< 0.001
2DE LA mechanics
Total longitudinal strain (%)	38 ± 6^a,b,d	37 ± 6^f,s	35 ± 5^d	33 ± 4^a,s	32 ± 4^b,f	< 0.001
Positive longitudinal strain (%)	22 ± 4	21 ± 4	18 ± 3	14 ± 3	12 ± 3	< 0.001*
Negative longitudinal strain (%)	−16 ± 4^a,q	−16 ± 4^f	−17 ± 4	−19 ± 5^a	−20 ± 5^f,q	< 0.001

LV, left ventricular; LVH, left ventricular hypertrophy; LAV_max, maximal left atrial volume; LAV_min, minimal left atrial volume; BSA, body surface area; LAV_pre-a, left atrial volume before atrial contraction; LA, left atrium; TotEV, total emptying volume; PassEV, passive emptying volume; ActEV, active emptying volume; TotEF, total emptying fraction; PassEF, passive emptying fraction; ActEF, active emptying fraction.

*p < 0.05 for all comparisons except for Normal LV geometry vs Concentric LV remodeling, and Concentric LV remodeling vs Eccentric non-dilated; **p < 0.05 for all comparisons except for Normal LV geometry vs Concentric LV remodeling, Concentric LV remodeling vs Eccentric non-dilated, and Concentric vs Dilated and concentric-dilated LVH; ¶p < 0.05 for all comparisons except for Concentric LV remodeling vs Eccentric non-dilated, and Concentric vs Dilated and concentric-dilated LVH; ^ap < 0.01 for Normal LV geometry vs Concentric LVH; ^bp < 0.01 for Normal LV geometry vs Dilated and concentric-dilated LVH; ^cp < 0.01 for Concentric LV remodeling vs Dilated and concentric-dilated LVH; ^dp < 0.05 for Normal LV geometry vs Eccentric non-dilated LVH; ^fp < 0.05 for Concentric LV remodeling vs Dilated and concentric-dilated LVH; ^gp < 0.01 for Concentric LV remodeling vs Concentric LVH; ^qp < 0.05 for Normal LV geometry vs Dilated and concentric-dilated LVH; ^rp < 0.05 for Eccentric non-dilated LVH vs Dilated and concentric-dilated LVH; ^sp < 0.05 for Concentric LV remodeling vs Concentric LVH.

Total LA longitudinal strain, which represents LA reservoir function, progressively decreased from normal LV geometry to concentric-dilated LVH (Table III). This was also the case for LV positive strain, an LA conduit function parameter. LA negative strain, a parameter of LA pump function, became less negative in the concentric-dilated group than in other groups, which means that pump function gradually increased from normal LV geometry to concentric-dilated LVH (Table III).

Three-dimensional echocardiographic left atrial phasic function

All 3DE-derived LAVs were higher than the corresponding 2DE-derived LAVs (Table IV). The trend of 3DE LAVs change remained the same as for 2DE LAVs. The major difference was a more pronounced gradual decrease in total and passive LA emptying fractions, representing LA reservoir and conduit functions, from normal LV geometry to dilated and concentric-dilated LVH (Table IV). The same was noticed for the progressive increment of the active LA emptying fraction, an LA booster pump function (Table IV).

Table IV. Three-dimensional echocardiographic left atrial volume (LAV) and mechanical analysis in the study population.

	Normal LV geometry (n = 91)	Concentric LV remodeling (n = 31)	Eccentric non-dilated LVH (n = 44)	Concentric LVH (n = 33)	Dilated and concentric-dilated LVH (n = 14)	p
3DE LAV analysis
LAVmax (ml)	61.3 ± 15.4^a,b	62.4 ± 16.1^c,s	65.4 ± 17	72 ± 19.1^a,s	76.5 ± 21.6^b,c	< 0.001
LAVmax/BSA (ml/m²)	32 ± 7.6^p,q	32.2 ± 8.2	32.7 ± 8.8	36.2 ± 9.8^p	38 ± 11.5^q	0.001
LAVmin (ml)	23.8 ± 4.9^a,b	24.3 ± 5.3^c,g	26.2 ± 5.8^o,n	31.7 ± 6.3^a,g,n	34.4 ± 6.8^b,c,o	< 0.001
LAVmin/BSA (ml/m²)	12.4 ± 2.5^a,b	12.6 ± 2.8^c,g	13.1 ± 2.7^o,n	15.9 ± 3.2^a,g,n	17.1 ± 3.4^b,c,o	< 0.001
LAVpre-a (ml)	34.5 ± 5.9	35.7 ± 6.3	39.1 ± 7.4	48.8 ± 9.1	54.6 ± 9.5	< 0.001**
LAVpre-a/BSA (ml/m²)	18.1 ± 3.6^a,b	18.4 ± 3.9^c,g	19.6 ± 4.2^o,n	24.5 ± 5.5^a,g,n	27.2 ± 6.8^b,c,o	< 0.001
LA TotEV (ml)	37.3 ± 7.6	38.1 ± 7.9	39.2 ± 8	40.3 ± 8.2	42.1 ± 8.7	0.057
LA PassEV (ml)	26.6 ± 5.5^p,q	26.7 ± 5.3^t	26.3 ± 5.2	23.2 ± 4.7^p	21.9 ± 4.2^q,t	0.001
LA ActEV (ml)	10.7 ± 3	11.4 ± 3.2	12.9 ± 3.6	17.1 ± 4.1	20.2 ± 4.8	< 0.001**
LA TotEF (%)	61 ± 6^a,b	61 ± 6^c,g	60 ± 5^r,x	56 ± 5^a,g,x	55 ± 4^b,c,r	< 0.001
LA PassEF (%)	44 ± 5	43 ± 6	40 ± 5	32 ± 4	29 ± 4	< 0.001**
LA ActEF (%)	31 ± 5^a,b	32 ± 6^t	33 ± 5	35 ± 6^a	37 ± 6^b,t	< 0.001

LV, left ventricular; LVH, left ventricular hypertrophy; LAV_max, maximal left atrial volume; LAV_min, minimal left atrial volume; BSA, body surface area; LAV_pre-a, left atrial volume before atrial contraction; LA, left atrium; TotEV, total emptying volume; PassEV, passive emptying volume; ActEV, active emptying volume; TotEF, total emptying fraction; PassEF, passive emptying fraction; ActEF, active emptying fraction.

**p < 0.05 for all comparisons except for Normal LV geometry vs Concentric LV remodeling, Concentric LV remodeling vs Eccentric non-dilated, and Concentric vs Dilated and concentric-dilated LVH; ^ap < 0.01 for Normal LV geometry vs Concentric LVH; ^bp < 0.01 for Normal LV geometry vs Dilated and concentric-dilated LVH; ^cp < 0.01 for Concentric LV remodeling vs Dilated and concentric-dilated LVH; ^gp < 0.01 for Concentric LV remodeling vs Concentric LVH; ⁿp < 0.01 for Eccentric non-dilated LVH vs Concentric LVH; ^op < 0.01 for Eccentric non-dilated LVH vs Dilated LVH; ^pp < 0.05 for Normal LV geometry vs Concentric LVH; ^qp < 0.05 for Normal LV geometry vs Dilated and concentric-dilated LVH; ^rp < ^{0.05 for Eccentric non-dilated LVH vs Dilated and concentric-dilated LVH; s}p < 0.05 for Concentric LV remodeling vs Concentric LVH; ^tp < 0.05 for Concentric LV remodeling vs Dilated and concentric-dilated LVH; ^xp < 0.05 for Eccentric non-dilated LVH vs Concentric LVH.

Logistic regression analysis

Using cut-off values of the 2DE and 3DE LAV maximum indices from the recommendation and the literature (7,12), multivariate logistic regression analysis demonstrated that only concentric (OR 4.7, 95% CI 1.6–10.6; and OR 4.3, 95% CI 1.5–11.2, respectively), dilated and concentric dilated LVH (OR 7.5, 95% CI 1.5–12.5; and OR 6.9, 95% CI 1.9–12.6, respectively) were associated with 2DE and 3DE LAV maximum indices independently of age, BMI, systolic blood pressure and LV mass.

Discussion

This investigation provides several new findings. (i) 2DE and 3DE LAVs gradually increased from normal LV geometry to concentric and concentric-dilated LVH. (ii) LA reservoir function, estimated by the total 2DE LA emptying fraction, was decreased in subjects with concentric and concentric-dilated LVH compared with patients with normal geometry and concentric remodeling. 3DE demonstrated a more gradual mode of deterioration of LA reservoir function than 2DE. (iii) LA conduit function, illustrated with 2DE and 3DE LA passive emptying fractions, progressively reduced from normal LV geometry to concentric-dilated LVH. (iv) LA booster pump function, evaluated by 2DE and 3DE, was increased in patients with concentric, dilated and concentric-dilated LVH compared to subjects with normal LV geometry. (v) The same results regarding LA phasic function were provided by 2DE strain analysis. (vi) Concentric, dilated and non-concentric dilated LVH were associated with LA enlargement independently of the main demographic and clinical features.

In our investigation, maximum, minimum and precontraction 2DE and 3DE LAVIs progressively increased from normal geometry to concentric-dilated LVH. However, there was no significant difference between normal geometry and concentric remodeling, or between concentric LVH and dilated LVH. These findings are in line with the literature (7–9). Patel et al. investigated a large sample of patients with preserved ejection fraction, and also found no difference in maximum LAVI between normal geometry and concentric remodeling, and no difference between eccentric and concentric LVH (8,9). Cho et al. (7) and Tsioufis et al. (10) came to the same conclusions in significantly smaller samples of hypertensive patients. Considering the findings from the Dallas Heart Study, which showed that LA enlargement is associated with increased mortality in the general population (19), it is possible that the higher mortality of patients with concentric and dilated LVH, found in previous studies (4), could be partly explained by larger LAVs in these patients. In addition, Wu et al. showed that 3DE LAVs are powerful predictors of future cardiac events, although the 3DE minimum LAVI tended to have a stronger and additive prognostic value compared to the 3DE maximum LAVI (18). These findings, in combination with the results of our study regarding significantly higher 3DE LAVIs in concentric and dilated LVH, also support the hypothesis of a relationship between LA remodeling and increased mortality in patients with these geometric patterns.

However, previous investigations of LA remodeling in different types of LV geometry used the traditional classification of LV patterns, without consideration of LV dilatation, and the authors provided only values of maximum 2DE LAVIs (7–10). Our investigation is the first to have provided comprehensive analysis of 2DE and 3DE phasic function, with all volumes, emptying fractions and strains, using updated criteria of LV geometry. All 3DE LAVs were higher than corresponding 2DE LAVs in the present study, as found in the previous investigations (18). This could be explained by the fact that 3DE estimation avoids foreshortening of the LA cavity and geometric assumptions, providing measures comparable with magnetic resonance (20). Our decision to use 3DE along with the 2DE technique in LA evaluation was made because we aimed to accurately determine LAVs in our patients.

The present findings revealed that LA reservoir function, estimated by total 2DE and 3DE LA emptying fractions, as well as total longitudinal strain, was decreased in subjects with concentric and concentric-dilated LVH compared with normal geometry and concentric remodeling. Furthermore, 3DE estimation has indicated more gradual deterioration of LA reservoir function than 2DE. Previous analysis showed that decreased LA reservoir function was associated with increased mortality, and even superior and incremental to LA enlargement (19). LA reservoir function correlates with LV longitudinal strain, a good predictor of outcome superior to ejection fraction (21,22), which also contributes to worse outcome in patients with concentric and dilated LVH.

Our results indicated reduced LA conduit function, evaluated by 2DE and 3DE volumetric and 2DE strain analyses, in hypertensive subjects with concentric and dilated LVH. Similar investigations with different LV geometric patterns have not been carried out so far; however, a recent investigation showed that hypertension leads to abnormal LA reservoir and conduit functions (23).

In our study, booster pump LA function, assessed by volumetric and strain analyses, was increased in patients with concentric and dilated LVH compared with normal geometry. A recent investigation showed that hypertensive patients with LVH have decreased conduit and increased booster pump function of the LA compared with hypertensive patients without LVH (24). This could explain why subjects with concentric and dilated LVH and higher LVMI have amplified LA pump function than those without LVH (normal geometry and concentric remodeling). In a condition of reduced LA reservoir and conduit function, the compensatory increment of LA booster pump function is necessary to maintain normal LV diastolic filling. In the later course of hypertensive disease, LA pump function gradually decreases, which elevates the risk of occurrence of atrial fibrillation (25,26) and further increases cardiovascular risk in patients with LVH. LA pump function is associated with heart failure symptoms (27), which could also explain the worse prognosis of patients with concentric and dilated LVH.

Our investigation indicates a relationship between concentric and dilated LVH and LA enlargement. This agrees with previous studies that showed a relationship between concentric LVH and LA dilatation (7–9). These authors used only 2DE LA assessment, traditional LV geometry classification and the significantly lower cut-off value for normal LAVI (28 ml/m²) recommended by previous guidelines (12). A recently published recommendation significantly raised the cut-off for normal LAVI to 34 ml/m² (13), which could have an impact on the final results of the aforementioned studies (7–10).

Limitations

Our research has several limitations. First, all echocardiographic analysis, especially 3DE assessment, can be significantly influenced by the quality of ultrasound images. Secondly, we excluded all the patients with comorbidities such as diabetes, coronary artery disease, heart failure and renal failure, which reduces the potential generalization of our results. Thirdly, we could not determine the causal relationship between LV geometric patterns and LA phasic function because this was a cross-sectional study. Fourthly, no guidelines exist for normal 3DE LAV, and therefore we used the existing data from a study that included 124 healthy subjects (18). Fifthly, we could not determine differences between dilated and concentric-dilated LVH owing to the small sample size.

Conclusion

LAVs estimated by 2DE and 3DE were higher in subjects with concentric, dilated and concentric- dilated LVH than in hypertensive patients with normal geometry or concentric remodeling. Reservoir and conduit functions of the LA were reduced, whereas booster pump function was elevated in compensation in individuals with concentric and dilated LVH in order to maintain diastolic filling of the left ventricle. We reached similar conclusions about LA phasic function using 2DE and 3DE volumetric and strain analysis. Concentric and dilated LVH were LV patterns that were associated with LA enlargement independently of the main demographic and clinical characteristics. Additional prospective investigations are needed to evaluate the importance of LA phasic function in hypertensive patients with different geometric patterns.

Declaration of interest: The authors report no conflicts of interest.