Influence of Nighttime Blood Pressure on Left Atrial Size in Uncomplicated Arterial Systemic Hypertension

Abstract

The aim of the study was to determine the relations of 24-h blood pressure (BP) and its different phases with left atrial size. A total of 130 subjects (mean age 46 years) not taking cardiac drugs were studied by M-mode and Doppler echocardiography and ambulatory BP recording. Subjects (excluding those with coronary artery or valvular heart disease, heart failure, or diabetes) were classified into two groups: 25 normotensives and 105 hypertensives (history of antihypertensive treatment and office diastolic BP > 90 mm Hg). The two groups were comparable in terms of sex, age, and heart rate, whereas body mass index, ( < .01), office BP, average 24-h BP, and average daytime and nighttime BP (all P < .00001) were higher in hypertensives. Hypertensives also had increased left atrial dimension, left atrial dimension/aortic root ratio (both P < .001), and left ventricular mass (LV) indexed for height (P < .0001). Positive correlations of left atrial dimension were found with office BP, average 24-h, average daytime and nighttime systolic and diastolic BP, LV mass index, and Doppler-derived E/A ratio. In a multivariate model that included potentially confounding factors, only body mass index (standardized β coefficient = 0.41, P < .00001), average nighttime diastolic BP (β = 0.33, P < .00001), and male sex (β = 0.18, P < .01) were independent predictors of left atrial size in the pooled population. In conclusion, left atrial size is more closely related to ambulatory, rather than office, BP measurements, and high average nighttime BP is a powerful marker of left atrial enlargement in arterial hypertension. Am J Hypertens 1997;10:836–842 © American Journal of Hypertension, Ltd.

Left atrium, echocardiography, systemic hypertension, ambulatory blood pressure

The Framingham Heart Study has recently documented that echocardiographic left atrial enlargement is an independent predictor of stroke and death in the general population.¹ Factors such as greater predisposition to nonrheumatic atrial fibrillation,² thrombus formation, and embolization^3,4 are the possible mechanisms underlying the association between left atrial dilation and the subsequent development of stroke.

Left atrial enlargement may be an early sign of heart involvement in arterial systemic hypertension.^5,6 Although an abnormal left ventricular (LV) compliance may determine an increase of left atrial size, it is controversial whether left atrial dilation is a consequence of increased afterload itself or rather of changes in LV structure (that is, LV hypertrophy) and function that occur during the natural history of hypertension. Ambulatory blood pressure (BP) monitoring can provide useful information to clarify this issue, as BP overload detectable over 24 h has been demonstrated to be associated better than clinic BP with LV hypertrophy^7–10 and LV diastolic dysfunction.^11–13

The present study was designed to examine the association of 24-h BP profile, and its different time components, with left atrial size in a population including both normotensive and hypertensive subjects, with reference to clinical and echocardiographic variables.

Methods

Study Population

A total of 130 subjects (80 men, 50 women), aged 23 to 71 years, entered the study after their informed consent had been obtained. This population included both normotensive subjects recruited for a BP screening of the staff in our institution and hypertensive outpatients. None of these subjects had clinical or laboratory evidence of respiratory, endocrine, hepatic, renal, or hematologic disorders. Antihypertensive treatment, if any, was withdrawn ≥21 days before the study. Exclusion criteria were coronary artery disease (history of angina and effort treadmill electrocardiogram), valvular heart disease, heart failure, and diabetes mellitus. Based on history of antihypertensive treatment and office BP levels (diastolic BP >90 mm Hg as the mean of three different measurements in more than three visits at 1-week intervals), 25 were considered to be normotensive and 105 to have borderline or established hypertension.

Procedures

All subjects underwent Doppler echocardiographic examination on the same day as 24-h ambulatory BP recording.

Echocardiographic examinations were performed with subjects in the partial left decubitus position, using a 2.5 MHz transducer connected to a Sector Imager 5000 (OTE-Biomedica, Florence, Italy) and recorded on strip-chart paper. At the end of each study, a physician blinded to the echocardiographic examination estimated (mean of three different measure-ments) heart rate and cuff BP, using a mercury sphygmomanometer. Consecutively coded echocardiographic tracings were examined by two experienced observers, who were unaware of BP values of any of the subjects, on three or more cardiac cycles. M-mode measurements were made by use of the American Society of Echocardiography leading-edge-to-leading-edge convention.¹⁴ Left atrial anterior–posterior dimension was measured at end-systole and the aortic root dimension at end-diastole at the level of the aortic valve. The ratio of left atrial dimension to the aortic root was derived. Left atrial dimension was measured as its largest diameter and the exact location of the posterior wall was confirmed by two-dimensional echocardiography. M-mode and two-dimensional measurements of left atrial size were compared in our laboratory on a sample of 40 subjects (10 normotensives, 30 hypertensives) of the study population. M-mode left atrial anterior–posterior dimension was related to the following two-dimensional measurements of left atrium: both anteroposterior and superior–inferior dimension in parasternal long-axis (r = 0.57, P < .0001 and r = 0.49, P < .001, respectively); medial–lateral dimension (r = 0.44, P < .01); superior–inferior dimension (r = 0.39, P < .01); and left atrial area (r = 0.50, P < .001) in apical four-chamber view. M-mode anteroposterior dimension of left atrium has been found to correlate well with the angiographic anteroposterior minor axis of the atrium,¹⁵ with the angiographic area of the left atrium recorded in the right anterior oblique projection¹⁶ and with biplane angiographic volumes.¹⁵ LV mass was estimated by the Penn convention¹⁷ and corrected for height in meters (LV mass index).¹⁸ Fractional shortening was calculated as the percentage of change in the internal LV dimension between systole and diastole. Relative diastolic wall thickness was determined by the sum of posterior wall and septal thickness divided by the LV internal diastolic diameter. Pulsed Doppler evaluation of LV inflow tract was performed from the apical four-chamber view, with the sample volume placed at the level of the mitral annulus, and a strip-chart speed of 50 or 10 mm/sec. The early filling phase peak transmitral flow velocity (peak E) and the late diastolic filling phase peak transmitral flow velocity A (peak A) were measured, and the flow velocity E/A ratio was derived.

The 24-h ambulatory BP recording was performed using the Pressurometer Takeda TM-2420 (A & D Co. Ltd., Tokyo, Japan). Accuracy and reproducibility of TM-2420 has been described.¹⁹ Before starting the recording, the reliability of the instrument was tested by a mercury sphygmomanometer. When the BP values provided by the two instruments differed by more than ± 3 mm Hg and could not be corrected by changing the position of the microphone, the recording was not considered available for the study. With the TM-2420 device, systolic and diastolic BP and heart rate were recorded at 15-min intervals during the daytime (7 AM to 11 PM) and at 30-min intervals during the night (11 PM to 7 AM). Readings that showed an inconsistent increase or decrease in systolic or diastolic BP > 20 mm Hg and readings with a pulse BP (systolic BP − diastolic BP) < 5 mm Hg were deleted. Further analysis of tracings was performed only if ≥75% of the total readings per 24 h passed the deletion criteria.

Data Analysis

Data are expressed as mean ± SD. The Student's t test for unpaired data was used to analyze the intergroup differences. The average systolic and diastolic BP and heart rate obtained by ambulatory BP recording were calculated for the 24-h period and, separately, for daytime and nighttime periods. Day–night BP and heart rate differences were derived. The means of the individual SDs and the mean of the individual coefficients of variation (SD × 100/mean) for the 24-h recordings were taken as measurements of BP variability. Linear regression analyses and partial correlation tests using Pearson's method were used to assess the univariate relation between left atrial dimension and potential determinants. Prediction of left atrial dimension was obtained by stepwise, forward, multiple linear regression analysis that included potential confounding variables not obviously related to each other (for example, 24-h BP and daytime BP). The null hypothesis was rejected at P < .05. Statistical analyses were made by the SPSS/PC for Windows (release 6.0) statistical package (SPSS, Chicago, IL) and by following the recommended procedures.

Results

Demographics and Ambulatory BP of the Study Population

The characteristics of the study groups are listed in Table 1. The two groups were comparable in terms of sex, age, and heart rate, with greater body mass index in hypertensives (P < .01). Obviously, office and ambulatory BP (all P < .00001) were higher in hypertensives. Average heart rate in any time component of the ambulatory recording, day–night BP decrease, and measurements of BP variability during the 24 h did not differ significantly between the two groups (data not shown).

Table 1

Characteristics and Ambulatory BP of the Study Population

Variable	Normotensives (n = 25)	Hypertensives (n = 105)	P
Sex (% men)	60	62	NS
Age (years)	43.4 ± 9.7	46.9 ± 8.6	NS
BMI (kg/m2	24.9 ± 2.3	26.8 ± 3.6	< .01
Office systolic BP (mm Hg)	134.2 ± 12.9	155.7 ± 17.2	< .00001
Diastolic BP (mm Hg)	80.6 ± 5.8	99.5 ± 6.3	< .00001
Heart rate (beats/min)	71.6 ± 11.3	75.1 ± 12.4	NS
24-h systolic BP (mm Hg)	124.0 ± 10.9	143.7 ± 15.6	< .00001
Diastolic BP (mm Hg)	77.9 ± 3.0	91.7 ± 7.6	< .00001
Daytime systolic BP (mm Hg)	128.6 ± 11.9	148.1 ± 15.8	< .00001
Diastolic BP (mm Hg)	80.7 ± 3.3	94.5 ± 7.8	< .00001
Nighttime systolic BP (mm Hg)	110.4 ± 9.3	130.4 ± 18.1	< .00001
Diastolic BP (mm Hg)	69.4 ± 4.1	82.8 ± 8.9	< .00001

Values are mean ± SD or percent.

BMI, body mass index; BP, blood pressure.

Doppler Echocardiographic Determinations

Doppler echocardiographic differences between normotensives and hypertensives are reported in Table 2. Left atrial dimension and left atrial dimension / aortic root ratio were higher (both P < .001) in hypertensives. They also exhibited higher septal and posterior wall thicknesses (both P < .00001), relative wall thickness and LV mass index (both P < .0001) and lower E/A ratio (P < .00001). The prevalence of LV hypertrophy (defined as LV mass index ≥141 g/m in men and ≥ 98 g/m in women²⁰) in hypertensives was 36% (38 of 105). None of the normotensive subjects had LV hypertrophy.

Table 2

M-Mode and Doppler Echocardiographic Measurements

Variable	Normotensives	Hypertensives	P
LA (mm)	32.8 ± 3.4	36.2 ± 4.9	< .001
LA/AO	1.14 ± 0.1	1.25 ± 0.1	< .001
LVIDD (mm)	49.7 ± 4.2	50.9 ± 5.5	NS
LVIDS (mm)	33.7 ± 3.9	34.6 ± 5.1	NS
FS (%)	32.4 ± 3.1	31.8 ± 4.9	NS
IVST (mm)	8.8 ± 1.5	11.2 ± 2.3	< .00001
PWT (mm)	8.5 ± 1.2	10.3 ± 1.6	< .00001
RDWT	0.35 ± 0.05	0.42 ± 0.1	< .0001
LVMI (g/m)	92.7 ± 24.4	130.9 ± 44.5	< .0001
E/A ratio	1.2 ± 0.4	0.95 ± 0.2	< .00001

Values are mean ± SD.

LA, left atrial dimension; LA/AO, left atrial to aortic root ratio; LVIDD, left ventricular internal diameter in diastole; LVIDS, left ventricular internal diameter in systole; FS, fractional shortening; IVST, interventricular septal thickness; PWT, posterior wall thickness; RDWT, relative diastolic wall thickness; LVMI, left ventricular mass index; E/A, peak velocity E wave/peak velocity A wave.

Determinants of Left Atrial Size

The univariate relations of left atrial size with clinical and echocardiographic variables in the entire study population are reported in Table 3. Heart rate, either measured clinically or calculated as the average of any ambulatory recording period, and measurements of BP variability were not significantly related to left atrial size.

Table 3

Univariate Relations of Left Atrial Size in the Study Population

Variable	Correlation Coefficient	P
Age	0.20	.02
BMI	0.52	.00001
Office SBP	0.29	.0008
Office DBP	0.28	< .001
Average 24-h SBP	0.41	< .00001
Average 24-h DBP	0.32	.002
Average daytime SBP	0.39	< .00001
Average daytime DBP	0.27	.002
Average nighttime SBP	0.43	< .00001
Average nighttime DBP	0.45	< .00001
LV mass index	0.40	< .00001
RDWT	0.23	.007
E/A ratio	−0.21	< .01

BMI, body mass index; SBP, systolic blood pressure; DBP, diastolic blood pressure; LV, left ventricular; RDWT, relative diastolic wall thickness.

Multiple linear regression analysis for the pooled population was performed, expressing the left atrial size as a function of sex, age, body mass index, office systolic and diastolic BP, daytime systolic and diastolic BP, nighttime systolic and diastolic BP, LV mass index, and E/A ratio. Body mass index (P < .00001), average nighttime diastolic BP (P < .00001), and male sex (P < .01) were independent predictors of left atrial dimension (Table 4). After removing the effect of body mass index, nighttime diastolic BP, and sex, the partial correlation coefficients for the other variables versus left atrial dimension were not significant. Furthermore, by forcing LV mass index into the model, the three above-mentioned variables remained still independent predictors of left atrial size, whereas the β coefficient (β = 0.11) of LV mass index was not significant. In sex-specific analyses, body mass index was the main contributor of left atrial size in men (n = 80, β = 0.44, P < .00001) and women (n = 50, β = 0.39, P = .002). However, average nighttime diastolic BP confirmed its independent association to left atrial size, the standardized β coefficient being 0.31 (P < .001) in men and 0.38 (P = .003) in women.

Table 4

Predictors of Left Atrial Size in the Study Population (Multiple Linear Regression Analysis)

Dependent Variable	Predictors	Standardized Coefficient	P
Left atrial dimension	BMI	0.41	< .00001
	Average nighttime DBP	0.33	< .00001
	Male sex	0.18	< .01

R² = 0.42; SEE = 3.73 mm; P < .0001.

BMI, body mass index; DBP, diastolic blood pressure; SEE, standard error of estimate.

Discussion

The present study demonstrates that 1) left atrial size is more closely related to ambulatory BP than to office BP, and 2) along the time components of 24-h BP, only average nighttime BP is an independent predictor of left atrial size. The influence of arterial systemic hypertension on left atrial enlargement has been extensively investigated in case-control studies.^5,6,21–23 In one of these,²² increased left atrial dimensions were found only in 5% of patients with diastolic hypertension, whereas the studies that included systolic hypertensives^5,6,21,23 found a significant relation of BP overload and left atrial abnormalities. An association of high levels in systolic or pulse pressures with increased left atrial size has been also described in a population-based Framingham cohort,²⁴ even after adjustment for age and body mass index.

The present study extends these observations by examining the impact of ambulatory BP monitoring on left atrial size in uncomplicated arterial hypertension. Left atrial size was related to both systolic and diastolic BP, determined either by office measurement or as average of any time component during the whole recording, in a population that included normotensives as hypertensives and a wide range of left atrial dimensions. Ambulatory BP monitoring has been already demonstrated to successfully predict organ damage, because pressures detectable during different daily activity correspond to the real hemodynamic overload imposed to the heart.^7–13 In our study, left atrial dimension was positively related more to average nocturnal BP than to daytime BP. This result is not surprising, because elevated nocturnal BP and decreased ambulatory BP decline from day to night were previously identified as major determinants of LV mass increase²⁵ and diastolic dysfunction^26,27 in arterial systemic hypertension. Measurements of BP variability, such as means of standard deviations and coefficients of variation of 24-h BP, were not associated with left atrial dimension. The role of these variables was assessed because of the close relation of BP variability to target organ damage described in arterial hypertension.^28,29

We also found a significant direct association between left atrial size and LV mass index, but this relation disappeared in the multivariate model, also by forcing LV mass index into the model. This finding demonstrates that left atrial enlargement that develops with LV hypertrophy is the consequence of increased afterload more than of LV mass increase itself. An association between atrial abnormalities and LV hypertrophy was documented in studies that used office BP as an index of afterload.^24,30–32 To our knowledge, the present study is the first to suggest that the relation of LV mass to left atrial dimension is mediated by BP pattern during the course of the entire day. Also, Doppler-derived E/A ratio and age were not independently related to left atrial size. E/A ratio is widely used as an index of LV diastolic compliance when LV filling pressures are normal,³³ and abnormalities in LV filling have been found to correlate with left atrial dimension in hypertensive patients.³⁴ However, the multivariable model of our study demonstrates that the relationship between LV diastolic compliance and left atrial size is mediated by confounders such as body size, nocturnal BP, and sex. An association between advancing age and left atrial dilation was previously reported in echocardiographic studies^24,35,36 but could not be expected in our population sample, in which few elderly subjects were included. This may have blunted the association of aging and left atrial enlargement, representing a study limitation. On the other hand, the inclusion of several subjects with middle age and some with younger age leads us to consider how the increase in left atrial size can develop as an early sign of hypertensive heart disease in patients with no other discernible cause of left atrial dilation.

By adjusting for clinic and echocardiographic variables, only body mass index, average nighttime diastolic BP, and male sex were independently associated with left atrial dimension. According to previous reports,^24,37 body mass index was the main predictor of left atrial size. Increased blood volume and preload are the main hemodynamic changes associated with overweight and they promote the development of LV diastolic dysfunction.^38–40 They may be also the mechanisms underlying left atrial enlargement in obese hypertensives. The sex-dependence of left atrial dimension was expected,^24,36 being probably linked to the same body size, which is generally increased in men. Previous echocardiographic studies have reported a similar increase of LV mass in men than in women at every BP level.^41,42 The magnitude of contribution of average nighttime diastolic BP to degree of left atrial size was lower than that of body mass index, but substantially higher compared with that of male sex. This independent contribution was confirmed in sex-specific analyses where increased body mass index was again the main predictor. Office BP and BP during the day had no independent relations with left atrial size. Therefore, nighttime diastolic BP suggests itself as a powerful marker of left atrial enlargement.

Why left atrial enlargement was associated with BP overload recorded during the night is uncertain. The duration of exposure to increased BP levels over the 24 h more than the high nocturnal BP itself is probably responsible for the observed atrial abnormalities. Furthermore, the exclusion of subjects with congestive heart failure may have blunted the degree of association between daytime BP and left atrial size by eliminating patients with severe hypertension, who are more prone to left atrial enlargement. However, it is remarkable that a previous report⁴³ identified greater susceptibility of hypertensives, whose BP fails to decrease during the night, to cardiovascular complications including stroke and transient cerebral ischemic attack. The demonstration of a substantial link between left atrial enlargement and stroke¹ provides a basis for explaining the incidence of cardiovascular sequelae in hypertension. It is noteworthy that left atrial enlargement is often associated with LV diastolic dysfunction, which participates in the development of congestive heart failure,⁴⁴ another main complication of hypertension.

Implications

The current investigation suggests that much of the left atrial abnormalities found in uncomplicated hypertension are explained by increased body weight often coexisting with arterial hypertension. However, based on the evidence of an independent association of average nighttime diastolic BP and changes in left atrial size, our findings support the hypothesis that the duration of exposure to BP loading over a 24-h period may be a key determinant of left atrial enlargement. Previous reports had shown an association of elevated nocturnal BP with LV hypertrophy²⁵ and diastolic dysfunction.^26,27 The current study confirms that increased afterload recorded during the night is a powerful marker of cardiac involvement in arterial systemic hypertension. A stable reduction of BP during a 24-h period, including the nocturnal phase, should prevent left atrial enlargement and subsequent development of atrial fibrillation and stroke.