Atrial fibrillation is associated with increased central blood pressure and arterial stiffness

Abstract Atrial fibrillation (AF) is the most common arrhythmia in clinical practice and beta blockers (BBs) are the drugs of choice for rate or rhythm control in these patients. The purpose of this study was to describe differences in arterial stiffness (AS), central blood pressure (cBP), and the role of BBs on cBP in patients with AF compared to healthy individuals. The authors included 76 patients with paroxysmal/persistent AF. Carotid‐femoral pulse wave velocity (PWV) and cBP were measured and compared with data from 75 healthy individuals. Patients with AF had higher PWV (8.0 m/s vs. 7.2 m/s, p < .001), central systolic blood pressure (cSBP) (118 mm Hg vs. 114 mm Hg, p = .033), central pulse pressure (cPP) (39 mm Hg vs. 37 mm Hg, p = .035) and lower pulse pressure amplification (PPA) (1.24 vs. 1.30, p = .015), without differences in peripheral blood pressure (pBP) and heart rate (HR). AF patients had significantly increased PWV (β= 0.500, p = .010, adjusted R² = 0.37) after adjustment for confounding factors. The use of BBs significantly reduced PPA (β = ‐0.059, p = .017, adjusted R² = 0.30). AF patients have higher PWV, cSBP, cPP, and lower PPA, compared to healthy patients. These findings support the role of AS in the development of AF. Use of BBs is related to a potential adverse effect on cBP.


Study protocol
Blood samples were collected from the antecubital fossa after an overnight fast. BP and carotid-femoral pulse wave velocities (PWV) were measured and pulse wave analysis (PWA) was performed within 1 day in patients after successful restoration of SR with cardioversion. Measurements of BP, PWV, and PWA were made before the procedure in patients who were hospitalized for PVI. All measurements were performed after 15 min of rest in a quiet, temperaturecontrolled room in a supine position. All patients were in SR during the study.

Hemodynamic measurements
BP was measured, using a validated digital oscillometric device (A&D UA-767; A&D Company Ltd., Tokyo, Japan), at least twice and mean BP was recorded.
Carotid-femoral PWV was measured and PWA was performed, using a Sphygmocor device (Sphygmocor Xcel and Sphygmocor Px, AtCor Medical, Sydney, Australia), at least twice and mean values were recorded.
The quality of measurements for PWA and PWV were controlled using the Sphygmocor Xcel's build in quality control (QC) indicator.
All measurements that did not meet the QC requirements (at least peripheral waveform quality above or equal to 75%) were dismissed and repeated.
PPA was calculated as a ratio of peripheral pulse pressure to central pulse pressure (pPP/cPP).
All measurements were performed in a dedicated, temperature controlled, study room for both the study, and control patients.

Echocardiography
Echocardiography was done as part of the clinical management of the patients elected for cardioversion or PVI in the study group. Echocardiographic information was available for all patients of the study group (76) and for 29 patients of the control group. The investigations were performed by experienced personnel.

Laboratory analysis
C-reactive protein, creatinine, estimated glomerular filtration rate (eGFR), plasma glucose levels, complete blood count, total cholesterol, LDL-cholesterol, HDL-cholesterol, and triglycerides were measured by standard laboratory methods in the local clinical laboratory.

Statistical analysis
The statistical programs Statistica 10

RESULTS
The study population consisted of 76 patients with AF and 75 age matched controls. The general characteristics of the AF patients and the control patients are described in Table 1.

Hemodynamics
Patients with a history of AF had higher PWV (

Echocardiography
Patients with AF had mild dilatation of the left atria (LA) compared to the control group (23.1 ml/m 2 vs. 36.4 ml/m 2 , p < .001). Analyzing the AF patients and controls together revealed a positive correlation of LA diameter (r = 0.38, p < .001) and indexed LA volume (r = 0.33, p = .001) with PWV. No statistically significant correlations were observed when the patients of the AF group and the controls were analyzed separately.
There was also a positive correlation of LA diameter with pSBP (r = 0.56, p = .002), cSBP (r = 0.51, p = .005), pPP (r = 0.42, p = .024) and cPP (r = 0.38, p = .043) in the control group. No significant correlation between BP and LA diameter was seen in the AF group. The echocardiography data is presented in Table 4.

DISCUSSION
In this study patients with AF had higher PWV, cSBP, cPP, and lower PPA compared to healthy controls, without differences in HR, pSBP, or pPP. Treatment with BBs was associated with lower PPA.
AS is a recognized marker of cardiovascular risk, 3 while carotidfemoral PWV measurement has been widely accepted as the gold standard for assessing AS. 4 AS has been found to be an independent predictor of primary coronary events 12 and stroke in hypertensive patients. 5 The interactions between AF and AS are not fully understood and information about the importance of AS in association with AF is scarce and contradictory. An earlier study with 34 patients and 31 controls did not find any difference in PWV between patients with first episode of AF and healthy patients. 13 The reasons for the differences from our study might be that the mean age of the participants was younger (49 years vs. 57 years in our study), the prevalence of hypertension was lower (39% vs. 65%) and the study excluded patients with LA diameter over 40 mm. These differences indicate that the patients of our study group were in a more advanced stage of the disease, which might explain the discrepancy between the results.
Another study by Kizilirmak and coworkers 14 compared cBP and AS in patients with paroxysmal AF and in the control group. They found that patients with paroxysmal AF had higher cBP and increased PWV.
Also, there was a significant difference in pBP (133/83 mm Hg, vs. 120/75 mm Hg, p < ,001) between the patients and the control group, which correlated with difference in cBP. 14 In the present study, despite the absence of a difference in pBP, the AF group showed higher cSBP and cPP. The importance of assessing cSBP and cPP was demonstrated in the Strong Heart Study, 6 where cBP proved to be a better predictor for cardiovascular events than pBP in participants without clinical cardiovascular disease at baseline. Similar results were confirmed in a meta-analysis by Vlachopoulus and coworkers 3 A reduction in cBP with antihypertensive drugs better predicts further cardiovascular events than pBP. 8 These findings support the theory that, compared to pBP, cBP reflects better the loading conditions for the heart, brain, and other organs. The higher cBP in our study population compared to the control might indicate a higher residual cardiovascular risk irrespective Use of beta blockers (n (%)), metoprolol (n (%)), bisoprolol (n (%)), nebivolol (n (%)) Values are presented as mean ± SD or count (%). Abbreviations: SD, standard deviation; ACEIs, angiotensin-converting enzyme inhibitors; ARBs, angiotensin II receptor blockers.
of having normal brachial BP levels. This is also supported by increased AS in the study patients compared to the control.
Lamante and coworkers showed that PWV and PP, a surrogate marker for AS, is correlated with LA size in hypertensive patients without previous AF episodes. 15 This was also confirmed by another study with 111 hypertensive patients 16 and by a larger prospective, community-based observational study, where peripheral PP was predictive of AF incidence. 17 These results demonstrate that increased AS may cause atrial enlargement, a known risk factor for AF. 18-20 AS influences cardiac remodeling and left ventricular geometry and has an important role in the diastolic function of the ventricle, 21 all of which are considered major determinants of LA size and hence contribute to their relationship with AF. 22 In our study patients with AF had a larger LA diameter, indexed LA end systolic volume, and increased AS, compared to the control group. We found a positive correlation between PWV and size and volume of LA when AF patients and controls were analyzed together; however, no significant correlation was seen in the AF group. The reason for this might be that all AF patients were managed using a rhythm control strategy, which is usually opted for patients with milder structural changes in the heart.
Hypertension is the most prevalent, independent, and potentially modifiable risk factor for AF. 9 Also, 65% of our study patients had a diagnosis of hypertension. As the prevalence of AF and hypertension increase with age, it is common to see AF patients with concomitant was also confirmed in a meta-analysis by Manisty and Hughes. 24 However, the inferior performance on cBP does not seem to be a class effect. Our recent study showed that the vasodilating BB nebivolol reduced cBP, cPP, and left ventricular wall thickness significantly more than metoprolol, with comparable reduction in pBP and HR. 25 The superior effect of nebivolol compared to atenolol regarding cBP reduction was also reported by Dhakam and coworkers 26 There is some evidence that, through reducing HR, non-vasodilating BBs may be associated with augmentation of cBP, thereby reducing the lowering effects on cBP. 8 .007 Values are presented as mean ± SD or count (%). Avbbreviation: SD, standard deviation.
with sick sinus syndrome who had lower HR (60 beats per minute) versus higher heart rate (90 beats per minute), did not show any increase in cBP. 27 In addition, Teeäär and coworkers showed that atenolol's inferior ability to reduce central BP in an acute setting may be related to heart rate-dependent and -independent mechanisms. 28 The current AF and hypertension guidelines 1,29 recommend BBs as the first-line drugs for rate or initial rhythm control for patients with AF. According to our study, BBs were significantly linked to lower PPA.
These findings coincide with the results of the CAFE study 8 and a metaanalysis, 24 where BBs had a smaller effect on cBP compared to other antihypertensive drugs. This further confirms the need to assess cBP in order to better manage the higher cardiovascular risk of AF patients.

CONCLUSIONS
Patients with AF have higher cSBP, cPP, PWV, and lower PPA compared to healthy patients, without differences in peripheral BP. These findings support the hypothesis that AS may play an important role in the development of AF. The use of BBs is related to the potential adverse effect on cBP, which may have an impact on the higher residual cardiovascular risk in patients with AF.

CONFLICTS OF INTEREST
The authors declare no conflicts of interests.

AUTHOR CONTRIBUTIONS
Priit Pauklin acquired data, interpreted the results, drafted and revised the manuscript, approved the final version. Jaan Eha interpreted the results, revised the manuscript, approved the final version. Kasper