Increased pulse pressure is associated with left atrial enlargement in resistant hypertensive patients

Abstract

Resistant hypertension (RH) is frequently associated with a high prevalence of target organ damage, which impairs the prognosis of these patients. Considering cardiac alterations in RH, most attention has been devoted to left ventricular hypertrophy (LVH), but data concerning left atrial enlargement (LAE) is less known. This cross-sectional study assessed the factors associated with LAE, with special focus on blood pressure (BP) estimates obtained by ambulatory blood pressure monitoring (ABPM), in 250 patients with RH, aged 64 ± 11 years. LAE and LVH were observed in 10.0% (95% CI 6.3–13.7) and 57.1% (95% CI 50.8–63.5) of patients, respectively. Compared with patients with normal atrium size, those exhibiting LAE were older, more frequently women, had elevated pulse pressure (PP) measured both at the office and by ABPM, and showed higher prevalence of LVH (83% vs 54%; p = 0.016). In a logistic regression analysis, adjusting for age, gender, body mass index, left ventricular mass index and BP pressure estimates, night-time PP was independently associated with LAE (OR for 5 mmHg = 1.28, 95% CI 1.24–1.32; p = 0.001). In conclusion, besides classical determinants of LAE, such as age and LVH, an elevated night-time PP was independently associated with LAE in patients with RH.

Key Words::

• Ambulatory blood pressure monitoring
• left atrial enlargement
• pulse pressure
• resistant hypertension

Introduction

Resistant hypertension (RH) is defined as blood pressure (BP) that remains above the goal despite the use of three or more antihypertensive drugs, one of them a diuretic. Several studies have shown that RH is frequently associated with a high prevalence of target organ damage and other comorbid conditions (1), which in turn impair the prognosis of this group of patients.

Target organ damage in RH has been assessed at different levels, including renal alterations, such as microalbuminuria and reduced glomerular filtration rate, and cardiac abnormalities. In the latter, most attention has been devoted to left ventricular hypertrophy (LVH) (2), but data concerning left atrial enlargement (LAE) is less well known. There is strong evidence that LAE, as determined by echocardiography, is a robust predictor of cardiovascular outcomes in patients without atrial fibrillation or significant valvular disease (3–5). These outcomes appear to be partially mediated by the presence of LVH. On the other hand, atrial remodelling, a process characterized by structural and electrophysiological atrial changes, is known to play a central role in atrial fibrillation initiation and maintenance, as well as in LAE (6,7).

Previous studies (8–10) have shown that LAE was associated with other clinical and echocardiographic determinants of high cardiovascular risk in hypertension, such as older age, obesity and LVH. It is a point of controversy whether LAE is a direct consequence of changes in left ventricular structure or directly derives from a chronically increased blood pressure load. Although one study suggested that clinical parameters besides BP were more important in developing LAE (11), there are no studies regarding the relationship of left atrial size and BP as measured by means of ABPM, which includes average of 24-h, daytime and night-time BP, as well as the circadian pattern in RH patients. In order to explore the relationship between BP estimates and LAE further, the aim of this study was to evaluate possible factors associated with LAE, with special focus on blood pressure estimates, measured by both office and ambulatory blood pressure monitoring in subjects with RH, a condition characterized by sustained BP elevation at the clinical setting.

Patients and methods

Study population

Patients with clinically documented RH were consecutively recruited from Hypertension Units spread throughout Spain. Inclusion criteria: patients aged ≥ 18 years, office RH diagnosis (BP ≥ 140 and/or 90 mmHg despite a therapeutic plan with at least three drugs at appropriate doses, including a diuretic).

Only patients with a good adherence to the antihypertensive treatment were included. All patients underwent a standard protocol and those with secondary hypertension were excluded. Demographic and anthropometric characteristics, cardiovascular risk factors and cardiovascular and renal diseases were recorded: 261 patients on the Spanish Resistant Hypertension Registry, consecutively visited in the centres where echocardiographic evaluation was available, were collected. Eleven patients with atrial fibrillation were excluded in this analysis. Urinary albumin excretion (UAE) was determined as the average of the urinary albumin/creatinine ratio evaluated from three fresh first-morning-void urine samples, obtained on separate days within a 1-month period (12).

Blood pressure measurements

Office BP was measured in the outpatient clinic by a trained nurse. After 5 min of rest in the sitting position, BP was measured using a validated oscillometric device (Omron 705I, Kyoto, Japan) and appropriate sized cuffs. We obtained three measurements spaced by 2 min, and the average of the two last BP measurements was considered the definitive office BP value. Twenty-four-hour ambulatory BP monitoring (ABPM) was carried out in all patients with a validated Spacelabs-90207 device (Issaquah, WA, USA) and suitably sized cuffs. The ABPM was carried out on a working day, starting at around 08:00–10:00 h. Ambulatory BP readings were obtained at 20-min intervals throughout both the awake and asleep periods. Awake and asleep periods were considered from 10:00 to 20:00 h and from 0:00 to 06:00 h, respectively. All patients included in the study had recordings of good technical quality (at least 80% of valid readings).

Echocardiography

Echocardiography examinations were carried out with the patients in the partial left decubitus position. Left atrium size, end-diastolic ventricular internal diameter, end-diastolic ventricular septum thickness and posterior wall thickness were measured. Left ventricular mass calculation was performed according to the American Society of Echocardiography recommendations (13). LVH was defined as a left ventricular mass index (LVMI) ≥ 125 g/m² in men or ≥ 110 g/m² in women. LAE was defined as a left atrium diameter indexed to body surface area > 26 mm/m².

Statistical analyses

Statistical methods were performed with the statistical package SAS for Windows version 9.2 (Cary, NC, USA). Continuous variables are summarized as mean ± SD and categorical data are presented as percentages. Bivariate comparisons between patients with LAE and patient with normal left atrium diameter were carried out by Student t-test in continuous data, and by χ² in categorical data. Pearson's correlations analysis was used to identify possible relationships between left atrium diameter and age, LVMI and ABPM estimates. A logistic regression analysis was then performed, with LAE used as the dependent variable. Possible confounders (age, gender, office BP and LVMI) were forced to enter in the model. Then, variables of interest (ambulatory BP estimates) were examined in a backward stepwise procedure. Multivariate odds ratios (OR) and their 95% confidence intervals (95% CI) were calculated for each independently associated variable.

The study protocol was approved by the local institutional ethics committee, and written informed consent was obtained from all patients. The investigation conforms to the principles outlined in the declaration of Helsinki.

Results

A total of 250 subjects with RH subjects were included. Mean age was 64 ± 11 years, 39% were women; 39.6% of them diabetics, 52.8% obese (body mass index ≥ 30 kg/m²) 65.3% had metabolic syndrome and 12% were active smokers. LAE was found in 25 subjects (10.0%; 95% CI 6.3–13.7) and LVH was observed in 132 subjects (57.1%; 95% CI 50.8–63.5) (Table I).

Table I. Clinical and demographic characteristics of the patients.

n	250
Age, years	64 ± 11
Gender, females	97 (39)
Office systolic BP, mmHg	159 ± 16
Office diastolic BP, mmHg	89 ± 12
Associated cardiovascular risk factors
Current smokers	30 (12.0)
Diabetes mellitus	99 (39.6)
BMI ≥ 30 kg/m^2	132 (52.8)
Metabolic syndrome	160 (65.3)
Target organ damage
Microalbuminuria^a	88 (48.3)
Renal function impairment^b	54 (24.3)
Left ventricular hypertrophy^c	132 (57.1)
Left atrial enlargement^d	25 (10.0)

Data given as mean ± SD or number and percentage: n (%). ^aUrinary albumin excretion ≥ 22 mg/g in males, ≥ 31 mg/g in females. ^bGlomerular filtration rate (MDRD) < 60 ml/min/1.73 m². ^cLeft ventricular mass index ≥ 125 g/m² in males and ≥ 110 g/m² in females. ^dIndexed left atrium diameter ≥ 26 mm/m². BP, blood pressure; BMI, body mass index.

A significant correlation was observed between left atrium diameter and both age (r = 0.33; p < 0.0001) and LVMI (r = 0.33; p < 0.0001). With respect to BP estimates, a significant correlation was also observed between left atrium diameter and pulse pressure (PP) measured at the office (r = 0.27; p < 0.0001), during 24 h (r = 0.32; p < 0.0001), daytime (r = 0.32; p < 0.0001) and night-time (r = 0.29; p < 0.0001). There were no significant correlations with other BP estimates.

Compared with patients with normal atrium size, those with LAE were older (72.1 ± 9 vs 62.9 ± 11 years; p < 0.0001), more frequently women (72% vs 35.1%; p = 0.0003) and had a higher prevalence of LVH (83.3% vs 54.0%; p = 0.016). With respect to BP estimates, patients with LAE had higher office PP (83 ± 26 vs 69 ± 16 mmHg; p = 0.0086) and ambulatory PP (24-h: 75 ± 17 vs 60 ± 14 mmHg; daytime: 75 ± 18 vs 60 ± 14 mmHg; night-time: 76 ± 19 vs 60 ± 16 mmHg; p < 0.001 for all comparisons). There were no significant differences in either the number or the types of antihypertensive drugs between subjects with or without LAE (Table II).

Table II. Factors associated with left atrial enlargement (LAE).

Factor	With LAE, n = 25	Without LAE, n = 225	p-value
Age (years)	72.1 ± 9	62.9 ± 11	< 0.0001
Gender: females	79 (72)	18 (35.1)	0.0003
Number of antihypertensive drugs	4.1 ± 1.03	4.0 ± 1.02	ns
Patients with ACE inhibitors or ARB	84.0	78.5	ns
Diabetes mellitus	11 (44.0)	88 (39.1)	ns
Obesity (BMI ≥ 30 kg/m^2	10 (40.0)	122 (54.2)	ns
Metabolic syndrome	18 (75.0)	142 (64.2)	ns
Left ventricular hypertrophy	15 (83.3)	109 (54.0)	0.016
Microalbuminuria	8 (44.4)	80 (48.8)	ns
eGFR < 60 ml/min/ 1.73 m^2	5 (29.4)	49 (23.9)	ns
Office systolic BP, mmHg	165 ± 21	159 ± 15	ns
Office diastolic BP, mmHg	82 ± 13	90 ± 12	0.004
Office PP, mmHg	83 ± 26	69 ± 16	0.0086
24-h systolic BP, mmHg	148 ± 22	139 ± 17	ns
24-h diastolic BP, mmHg	73 ± 11	79 ± 12	0.0158
24-h PP, mmHg	75 ± 17	60 ± 14	0.0001
Daytime systolic BP, mmHg	150 ± 23	141 ± 17	ns
Daytime diastolic BP, mmHg	75 ± 13	81 ± 12	0.0113
Daytime PP, mmHg	75 ± 18	60 ± 14	0.0002
Night-time systolic BP, mmHg	144 ± 24	135 ± 21	ns
Night-time diastolic BP, mmHg	68 ± 11	74 ± 13	0.0241
Night-time PP, mmHg	76 ± 19	60 ± 16	0.0001
Nigh-to-day systolic BP ratio	0.96 ± 09	0.95 ± 09	ns
Night-to-day diastolic BP ratio	0.92 ± 0.1	0.92 ± 0.09	ns

Data given as mean ± SD or number and percentage: n (%). ACE inhibitors, angiotensin-converting enzyme inhibitors; ARB, angiotensin receptor blockers; BMI, body mass index; eGFR, estimated glomerular filtration rate; BP, blood pressure; PP, pulse pressure.

A logistic regression analysis was performed to ascertain the association of BP estimates obtained during 24-h ABPM and LAE. After adjustment for possible confounders (age, gender, body mass index, LVMI, office BP and other 24-h ABP estimates), elevated night-time PP was independently associated with LAE (OR for 5 mmHg = 1.28; 95% CI 1.24–1.32; p = 0.001).

Discussion

The main finding of this study carried out in resistant hypertensive subjects was that among different BP estimates obtained during 24-hour ABPM, night-time PP was independently associated with LAE, after adjusting for other well known associated factors (age, gender, LVMI) and also for office BP. This strengthens the relationship between night-time BP and other markers of target organ damage previously found in RH patients (12).

The prognosis of patients with RH compared with better controlled hypertensive subjects has not been widely evaluated. The prognosis is possibly impaired because these patients have a long standing history of poorly controlled hypertension (1) and they usually have associated cardiovascular risk factors and a high prevalence of target organ damage (14–17). Previous studies (2,12,16,17) have shown the factors associated with renal damage or LVH, but the relationship between BP estimates and LAE has not been previously reported in RH subjects.

A significant positive relationship between BP and left atrium size has been shown in some previous studies (18,19), but not in others (20,21). Cuspidi et al. (18) reported in a study carried out in 2500 untreated and treated uncomplicated essential hypertensive patients, consecutively attended in a hypertension clinic, that subjects with LAE had an office systolic BP significantly higher in comparison with those without LAE. Moreover, in a population-cohort based prospective study carried out in 1849 male and 2152 female participants of the Framingham Heart Study and Framingham Offspring Study (19), increasing levels of systolic and PP were significantly associated with LAE: subjects with an 8-year average systolic BP of 140 mmHg or higher were twice as likely to have LAE as those with values of 110 mmHg or lower. In this study, increased levels of systolic and PP (but not diastolic or mean BP) were significantly associated with left atrial size. On the other hand, Tedesco et al. (20) showed in 164 hypertensive subjects that age and LVMI were the only significantly associated factors with LAE. Cuspidi et al. (21) did not observe significant differences either in office BP or 48-h ambulatory BP between subjects with or without LAE.

The results of the present study are focused on an unselected population of RH patients, and give strength to PP, an indirect measure of arterial rigidity, as the most important BP determinant of LAE. We found clear differences in PP, measured both at the office and by ABPM, between patients with and without LAE. On the contrary, the relationship with diastolic BP was negative. Although systolic BP values were higher in patients with LAE, differences did not reach statistical significance.

Staessen et al. (22) have previously shown that in patients with isolated systolic hypertension, higher PP estimated by means of 24-h ABPM was a better predictor of an adverse outcome than office PP, whereas conventional and ambulatory mean pressures had no influence on it. Elevated PP is associated with advanced age and predisposes to LVH, impaired ventricular relaxation and increased left atrium size. In the Framingham population (23), in subjects aged 35 years or older, PP was an important risk factor for incident atrial fibrillation.

In a previous study carried out in 1140 subjects ≥ 40 years old with sustained hypertension (24), it was reported that non-dippers patients had a higher percentage of LAE than dipper ones, and had a twofold greater risk of developing atrial fibrillation, but data about PP were not shown. In the present study, there were not significant differences in the systolic or diastolic night-day ratio. Other previous works (25,26) have shown that nocturnal hypertension is better associated with cardiovascular target organ damage as compared with the non-dipping pattern. On the other hand, Triantafyllidi et al. (27) reported that office PP was the only significant determinant of left atrium size, but data about ambulatory BP or PP were not reported. Our data are in agreement with these previous results, as among different office and ambulatory BP measurements, nocturnal PP was the most closely associated factor with LAE in our cohort of RH.

Considering the relationship between PP and target organ damage in RH, and more than that, to the best of our knowledge, data about its association with LAE have not been reported previously.

The present study should be interpreted in the context of several limitations. These results are based on a cross-sectional design. On the other hand, 513 patients were included in the Resistant Hypertension Registry of the Spanish Society of Hypertension, but an echocardiographic study was performed only in 250 patients from those hypertension units where the echocardiography was available. We used in the analysis the left atrium diameter, because we had no data about the left atrium volume. Furthermore, the proportion of patients with LAE was small, and the observed differences must be confirmed in larger studies.

In conclusion, we have observed for the first time an independent association of night-time PP with LAE in resistant hypertensive patients. Given the fact that LAE predisposes to chronic atrial fibrillation, elevated night-time PP should alert physicians attending patients with RH.

Investigators of the Spanish Society of Hypertension-Resistant Hypertension Registry

Scientific Committee

Pedro Aranda, Pedro Armario, Carlos Calvo, Juan Antonio Divisón, Pablo Gómez, Raquel Hernández del Rey, Anna Oliveras, Alejandro Roca-Cusachs, Luis M. Ruilope, Carlos Sanchís, Julián Segura, Alejandro de la Sierra, Carmen Suárez.

Investigators

María Abad, María Jesús Adrián, Javier Andrés, Pedro Armario, Jesús María Arteaga, Mario Luis Ávila, Sonia Blanco, Josep Bonet, Carlos Calvo, Mari Cruz Carreño, Eva María Cotilla, Luis Miguel Cuadrado, Alejandro de la Sierra, Leyre Díez, Juan Antonio Divisón, Mónica Doménech, María Jesús Eiris, Vicente Esteve, Angela Felip, Gema Fernández-Fresnedo, Patrícia Fernández-Llama, Celia Fernández, José Antonio García, Vicente Giner, Pablo Gómez, Javier Guerediaga, Fernando Gutiérrez, Raquel Hernández-del Rey, Pedro Horcajo, Julio Jorge, Pedro Jesús Labrador, Ignacio Lekuona, Antonio Liébana, María Dolores López, Dolores Lorenzo, Antonia Maldonado, María José Manzanera, Nieves Martell-Clarós, Jesús Martín, Isabel Martínez, Juan Carlos Martínez-Acitores, Carmen Morata, Anna Oliveras, Manuel Olmos, Aitor Parra, José Manuel Pascual, Antonio Pedreira, Gonzalo Pia, Mar Piedecausa, Luis Angel Prieto, José Neldo Prieto, Josep Redón, Nicolás Roberto Robles, Alejandro Roca-Cusachs, Jordi Roig, Eduardo Rovira, José Saban, Juan Francisco Sánchez, Pilar Sánchez, Julián Segura, Javier Sobrino, Carmen Suárez, Javier Tisaire, Pere Torguet, Jesús Toril, Santiago Torreira, Susana Vázquez, Ana Veiga, Luis Vigil, José Ovidio Yáñez.

Acknowledgements

We thank the Statistical support of the senior biostatistician, Marta Figueras Balsells from TFS Spain.

Sources of funding

The Spanish Resistant Hypertension Registry is funded by Novartis Pharmaceuticals.

Conflict of interest: The Spanish Resistant Hypertension Registry is funded by Novartis Pharmaceuticals. Pedro Armario has received honoraria for development of educational presentation from Menarin. Anna Oliveras has received honoraria for development of educational presentation from Menarini. Alejandro de la Sierra has received honoraria for lectures from Daiichi-Sankyo, Novartis and Merck Sharp & Dohme and honoraria for development of educational presentations from Menarini. Luis M Ruilope has received honoraria for lectures or advisory from Bayer, Sanofi-Aventis, Takeda, Daiichi-Sankyo, Otsuka, Relypsa, Roche, oehringer-Ingelheim, Novartis, Astra-Zeneka, and BMS and fees for lectures and participation in reviews activities and development of educational presentations. The remaining authors declare that they have no conflict of interest.