Changes in Cardiovagal Baroreflex Sensitivity Are Related

Circulation Journal Vol.76, December 2012

Circulation JournalOfficial Journal of the Japanese Circulation Societyhttp://www.j-circ.or.jp

atients with liver cirrhosis (LC) frequently experience autonomic cardiovascular dysfunction, such as in-creased activity of the sympathetic nervous system and

reduced vagal cardiac function,1–3 which has important impli-cations for liver dysfunction and poor survival.4,5 Cardiovagal baroreflex has been shown to be an important determinant of electrical stability in the heart,6,7 a key mechanism of short-term (ie, beat-by-beat) arterial pressure (AP) control, and its sensi-tivity can predict increased mortality and end-organ damage.8–11 Cardiovagal baroreflex sensitivity (BRS) has been associated with changes in left ventricular (LV) morphology, and impaired BRS may be a marker of target organ damage in hypertensive patients.8,12

Patients with cirrhotic cardiomyopathy have enhanced ac-tivity of the sympathetic nervous system and hyperdynamic circulation showing increased cardiac output and reduced sys-temic vascular resistance.13,14 These changes may induce myo-cardial remodeling and LV hypertrophy (LVH), resulting in

systolic and diastolic functional abnormalities.14,15 Despite the associations between BRS and changes in LV morphology shown in animals and in patients,8,16,17 few studies have assessed the association of reduced cardiovagal BRS with the develop-ment of myocardial remodeling and increased LV mass index (LVMI) in LC patients. We therefore evaluated the relation-ship between the cardiovagal BRS control of heart rate (HR) and morphological and functional cardiac modifications in pa-tients with LC.

MethodsWe enrolled 82 cirrhotic patients scheduled for liver transplan-tation between July 2006 and June 2007. We excluded patients taking any cardiovascular medication, including β-blockers and calcium channel blockers; those with acute on chronic liver failure, hepatorenal syndrome, hepatopulmonary syndrome, or hepatic encephalopathy; and patients with diabetes mellitus or

P

Received March 20, 2012; revised manuscript received July 6, 2012; accepted July 27, 2012; released online August 28, 2012 Time for primary review: 21 days

Department of Anesthesiology and Pain Medicine, Asan Medical Center, University of Ulsan College of Medicine, Seoul, KoreaMailing address: Gyu-Sam Hwang, MD, PhD, Department of Anesthesiology and Pain Medicine, Laboratory for Cardiovascular Dynam-

ics, Asan Medical Center, University of Ulsan College of Medicine, 388-1 Pungnap-2dong, Songpa-gu, Seoul, 138-736, Korea. E-mail: [email protected]

ISSN-1346-9843 doi: 10.1253/circj.CJ-12-0380All rights are reserved to the Japanese Circulation Society. For permissions, please e-mail: [email protected]

Changes in Cardiovagal Baroreflex Sensitivity Are Related to Increased Ventricular Mass in

Patients With Liver CirrhosisJun-Gol Song, MD, PhD; Young-Kug Kim, MD, PhD;

Won-Jung Shin, MD, PhD; Gyu-Sam Hwang, MD, PhD

Background: Morphological and functional cardiac modifications attributable to neurohumoral activation and hy-perdynamic circulation have been found in patients with liver cirrhosis (LC). Cardiovagal baroreflex sensitivity (BRS) has been shown to inversely correlate with left ventricular (LV) morphology. It was hypothesized that, in patients with cirrhotic cardiomyopathy, reduced BRS is associated with myocardial remodeling and increased LV mass index (LVMI).

Methods and Results: Eighty-two LC patients scheduled to undergo liver transplantation were evaluated. Spectral analysis was done of beat-by-beat blood pressure and heart rate (HR) time series and BRS was derived from their cross-spectral gain. Echocardiography before liver transplantation was used to evaluate heart morphology and func-tion. BRS was inversely correlated with LV wall thickness (P=0.038), end-diastolic interventricular septum thickness (P=0.048), LVMI (P=0.005) and HR (P<0.001). On multivariate stepwise linear analysis LVMI and HR were inde-pendently associated with BRS. On tertile analysis of LVMI, compared with the lowest tertile of LVMI (75±11 g/m2), the highest tertile (118±13 g/m2) showed significantly impaired BRS (4.6±2.3 vs. 6.4±3.1 ms/mmHg, P=0.012).

Conclusions: Reduced cardiovagal BRS is associated with increased LV mass in patients with LC, suggesting a relationship between cardiovagal BRS control of HR and cardiac end-organ damage in patients with cirrhotic cardiomyopathy. (Circ J 2012; 76: 2807 – 2813)

Key Words: Baroreflex sensitivity; Left ventricular mass index; Liver cirrhosis

ORIGINAL ARTICLEHypertension and Circulatory Control


2808 SONG JG et al.

cardiovascular disease. Of 82 patients, 71 (87%) had chronic hepatitis B-related LC, 7 had alcoholic LC, 2 had hepatitis C-related LC, and 2 had cryptogenic cirrhosis. Stratification of patients according to modified Child-Turcotte-Pugh (CTP) showed that 20 patients (24%) were in Child class A, 34 (41%) in Child class B, and 28 (34%) in Child class C. LC was diag-nosed on liver biopsy and accepted clinical and biochemical criteria. In 41 patients (50%), ascites was confirmed on ultra-sonography.

We also enrolled an age-matched control group without liver disease for the evaluation of cardiovascular autonomic func-

tion. This control group consisted of 17 individuals (11 men, 6 women; mean age, 49.7±7.5 years). None of these individu-als was taking any cardiac medication, had cardiovascular dis-eases or had diabetes. The study was approved by the Institu-tional Review Board, and all subjects provided written informed consent.

Subjects were placed in the supine position for at least 10 min in an environmentally controlled experimental room with an ambient temperature of 23–25°C in the morning. HR was mea-sured continually using a standard electrocardiogram (ECG). Non-invasive beat-to-beat AP in the middle finger was deter-

Figure 1. Representative examples of RR interval (RRI), systolic arterial pressure (SAP), heart rate variability (HRV), SAP vari-ability (SAPV), coherence, and baroreflex sensitivity (BRS) in the supine position in a (blue) control subject and (red) patient with liver cirrhosis. Note the changes in time series of RRI and SAP and changes in frequency ranges of HRV and SAPV. BRS calcu-lated from frequency ranges showing coherence >0.5 is also illustrated. The shaded areas represent the low-frequency (LF; 0.04–0.15 Hz) and high-frequency (HF; 0.15–0.40 Hz) regions, respectively.


2809Baroreflex Sensitivity and LV Mass

mined using photoplethysmography (Finometer, FMS, The Netherlands), a technique that provides results correlating with directly measured radial AP.18 An appropriate-sized Finometer blood pressure cuff was placed on the middle finger and ad-justed to heart level. After stabilizing the finger pressure sig-nal, ECG and AP signals were continuously recorded for 10 min in the supine position with spontaneous breathing. Beat-by-beat ECG and AP signals were digitized and collected at 500 Hz using an online personal computer interfaced with an analog-to-digital converter (DI-720U; DATAQ Instruments, Akron, OH, USA). Offline analysis was performed using signal pro-cessing software (CODAS, DATAQ; DADiSP/Adv DSP, DSP Development, Cambridge, MA, USA) and custom-written MATLAB (MathWorks, Natick, MA, USA) scripts.19,20

To measure cardiovagal BRS, we first evaluated RR inter-val (RRI) variability and systolic AP (SAP) variability (SAPV) using frequency domain analysis. The amplitude of low-fre-quency (LF) SAP oscillations (0.04–0.15 Hz) was evaluated as a sympathetic control of the vasculature, and the amplitude of high-frequency (HF) RRI oscillations (0.15–0.40 Hz) as a vagal control of the heart.21 Stationary 6-min time series of beat-by-beat RRI and SAP were interpolated to 5 Hz to provide equi-distant samples. The fast-Fourier-transformed power spectrum density of the signals, which was based on Welch’s algorithm of averaging periodograms, was calculated using a sliding win-dow with a width of 500 points and an overlap of 250 points after detrending and application of a Hanning filter. This meth-od yields a frequency resolution of 0.01 Hz. The resulting 5 periodograms were averaged to produce the estimated spec-trum.

Frequency-domain transfer function analysis was used to quantify cardiovagal BRS, as described.22 Briefly, the transfer function gain and coherence (the squared coherence function) between SAP and RRI were estimated using the cross-spectral

method.19 LF gain, for which the coherence exceeds 0.5, was used to calculate the cardiovagal BRS from the transfer func-tion gain. This band area is believed to derive from the blood pressure regulatory system, thus reflecting the function of the beat-to-beat SAP control system and baroreflex.22

Prior to surgery, patients underwent comprehensive echo-cardiographic evaluation using a Hewlett-Packard Sonos 2500 or 5500 imaging system to evaluate heart morphology and func-tion. Parameters measured included end-systolic LV internal diameter (LVIDs), end-diastolic LV internal diameter (LVIDd), end-systolic LV posterior wall thickness (LVPWs), end-dia-stolic LV posterior wall thickness (LVPWd), end-systolic in-terventricular septum thickness (IVSs), end-diastolic interven-tricular septum thickness (IVSd), LVMI, diameter of the left atrium (LA), diameter of the aorta, end-diastolic volume (EDV) and end-systolic volume (ESV). LV wall thickness (LVWT) was calculated as IVSd + LVPWd and LV relative wall thick-ness (RWT) was calculated as (IVSd + LVPWd) / LVIDd.3 Parameters of systolic function included LV ejection fraction (LVEF), stroke volume (SV) and fractional shortening (FS), with the latter calculated as [(LVIDd – LVIDs) / LVIDd] × 100 (%). Diastolic function was assessed on 2-D echocardiog-raphy as E/A ratio, where E is the early maximal ventricular filling velocity, and A is the late diastolic or atrial velocity, and deceleration time. LVMI was calculated by dividing LVM by the body surface area. LVH was defined as LVMI >118 g/m2 for men and LVMI >108 g/m2 for women.23

All data are presented as mean ± SD, with groups compared using the chi-square test, t-test, 1-way analysis of variance (ANOVA) with Bonferroni correction for multiple compari-sons, the Kruskal-Wallis test and the Scheffe or Dunn’s test, as appropriate. Because the autonomic function data were skewed, these data were analyzed after natural logarithmic transformation. Correlations between independent variables

Table 1. Clinical and Hemodynamic Characteristics vs. LVMI and BRS in LC Patients

All (n=82)Tertiles of LVMI

P-valueTertiles of BRS

P-value1 st (n=27) 2nd (n=28) 3rd (n=27) 1 st (n=27) 2nd (n=28) 3rd (n=27)

Sex (M/F) 59/23 18/9 18/10 21/6 0.650 16/11 21/7 19/8 0.179

Age (years) 51.1±7.7 50.5±7.1 50.8±9.1 52.0±6.9 0.767 51.7±7.7 52.7±6.2 48.9±8.9 0.173

BMI (kg/m2) 24.1±3.5 23.4±3.2 24.1±4.0 24.8±3.3 0.308 24.0±4.0 23.6±3.0 24.7±3.4 0.483

CTP score 8.7±2.1 8.0±1.9 8.5±2.3 　 9.6±2.0* 0.017 9.3±1.9 8.5±2.0 8.4±2.4 0.269

MELD score 14.2±6.3 10.6±3.7 　14.9±6.2* 　17.2±6.7* <0.001　 15.9±7.0 13.9±7.0 13.0±4.2 0.227

TB (mg/dl) 4.8±8.3 1.9±1.1 4.6±7.7 　 7.8±11.6* 0.037 5.9±10.1 4.7±8.3 3.7±6.2 0.619

Albumin (mg/dl) 2.5±0.6 2.6±0.7 25±0.5 2.5±0.5 0.946 2.5±0.5 2.7±0.5 2.5±0.6 0.308

Creatinine (mg/dl) 0.8±0.3 0.8±0.2 0.8±0.5 0.9±0.5 0.464 0.8±0.4 0.9±0.3 0.8±0.1 0.508

PT (INR) 1.5±0.4 1.3±0.2 　 1.6±0.4* 　 1.6±0.4* 0.043 1.5±0.4 1.5±0.4 1.4±0.3 0.453

Hct (%) 31.2±5.1 34.0±4.8 　29.3±3.4* 　30.4±5.8* 0.010 30.2±5.0 29.9±4.8 　　　33.5±4.9*,† 0.015

Na (mmol/L) 136.1±5.0　　 136.4±4.1　　 136.1±5.8　　 135.8±5.0　　 0.736 135.0±6.0　　 135.7±4.7　　 137.7±3.7　　 0.126

Ammonia (μmol/L) 86.8±48.2 70.0±26.4 92.1±57.6 98.1±38.4 0.050 87.1±39.7 85.0±57.0 88.3±33.4 0.963

AST (IU/L) 59.8±35.8 53.9±22.7 53.6±23.5 71.9±51.7 0.072 70.6±50.0 57.1±29.3 51.5±19.9 0.130

ALT (IU/L) 35.7±18.9 38.7±19.4 28.2±13.7 　 40.5±21.1* 0.040 34.6±21.7 35.1±16.0 37.3±19.2 0.857

QTc interval (ms) 442.2±30.3 430.3±26.7 444.4±36.8 451.2±20.9 0.050 454.9±32.4 445.1±27.9 441.4±26.0 0.064

SAP (mmHg) 110.3±15.8 109.5±14.9 109.7±14.5 111.6±18.4 0.655 112.4±17.4 107.4±17.5 111.0±12.1 0.487

DAP (mmHg) 52.3±11.5 54.2±13.6 53.9±6.9 48.6±12.6 0.094 52.5±13.4 50.7±12.8 53.7±7.6 0.621

HR (beats/min) 70.4±10.9 71.5±10.9 69.4±9.2 70.4±12.5 0.709 76.9±11.0 72.3±8.0 　　　62.1±7.8*,† <0.001　

Data given as mean ± SD. *P<0.05 vs. tertile 1, †P<0.05 vs. tertile 2.ALT, alanine aminotransferase; AST, aspartate aminotransferase; BMI, body mass index; BRS, baroreflex sensitivity; CTP, Child-Turcotte-Pugh; DAP, diastolic arterial pressure; Hct, hematocrit; HR, heart rate; LC, liver cirrhosis; LVMI, left ventricular mass index; MELD, model for end-stage liver disease; PT (INR), prothrombin time (international ratio); QTc interval, QT interval corrected for HR; SAP, systolic arterial pres-sure; TB, total bilirubin.


2810 SONG JG et al.

were analyzed using Pearson’s or Spearman’s rank correlation test, as appropriate. Variables with a significance level of P<0.1 on correlation analysis were included in a stepwise multiple linear regression analysis to determine variables that could independently predict BRS. P<0.05 was considered signifi-cant. Cumulative survival was estimated using the Kaplan-Meier method, and among-group survival was compared on log-rank test. All statistical analysis was done using SPSS version 12.0 (SPSS, Chicago, IL, USA).

ResultsRepresentative tracings of RRI, SAP, HR variability (HRV), SAPV, coherence, and transfer function gain from a patient with LC and from a control subject are shown in Figure 1. Subjects were classified by tertiles of LVMI and BRS, with Table 1 summarizing the distribution of clinical and hemody-namic parameters of LC patients along with tertiles of LVMI and BRS. The mean LVMI in LC patients was 97±20 g/m2 (range, 54–153 g/m2), and the prevalence of LVH was 13.4% (11/82; 6 men, 5 women). Patients in the highest LVMI tertile (118±13 g/m2) differed significantly from those in the lowest tertile (75±11 g/m2) in CTP score, model for end-stage liver disease (MELD) score, total bilirubin concentration, prothrom-bin time, hematocrit, and alanine aminotransferase concentra-tion. The SAP, diastolic AP (DAP) and HR, however, did not differ significantly among the 3 groups. On classification of

LC patients into BRS tertiles, hematocrit and HR differed significantly among groups (P<0.05 each; Table 1).

Table 2 lists echocardiographic data in groups of LC pa-tients classified by tertiles of LVMI and BRS. As expected, patients in the highest LVMI tertile had significantly higher LVIDs, LVIDd, LVPWd, IVSs, IVSd, LVWT, LVMI, LA, ESV, and EDV than patients in the lowest LVMI tertile. Echo-cardiographic parameters of LV systolic function such as SV differed significantly among the LVMI tertiles, whereas LVEF and FS did not. Parameters of LV diastolic function, such as peak E and peak A velocities, differed significantly among the LVMI tertiles. Patients in the lowest BRS tertile had signifi-cantly higher LVMI, LVPWd, IVSd, and LVWT than those in the lowest BRS tertile. In contrast, parameters of LV systolic and diastolic function did not differ significantly among the BRS tertiles.

When we compared the autonomic data in LC patients and controls, we found that HRVLF, HRVHF, SAPVLF, total power of SAPV and BRS were significantly lower in LC patients (P<0.05 each; Table 3). Moreover, BRS in the highest LVMI tertile was significantly lower than that in the lowest LVMI tertile (4.6±2.3 ms/mmHg vs. 6.4±3.1 ms/mmHg, P=0.012; Figure 2).

When we examined the relationship between BRS and car-diovascular parameters, including age, sex, body mass index, HR, SAP, DAP, and functional and morphological cardiac pa-rameters, we found that BRS was significantly correlated with

Table 2. Echocardiographic Characteristics vs. LVMI and BRS in LC Patients

Tertiles of LVMI P-value (ANOVA)

Tertiles of BRS P-value (ANOVA)1 st (n=27) 2nd (n=28) 3rd (n=27) 1 st (n=27) 2nd (n=28) 3rd (n=27)

Cardiac morphology

LVIDs (mm) 27.4±3.9 　28.9±3.9* 　　　32.2±3.8*,† <0.001　 29.6±4.2 29.2±3.3 29.7±5.4 0.919

LVIDd (mm) 46.8±4.4 　50.6±3.1* 　55.4±3.7* <0.001　 51.0±5.4 50.3±4.6 51.6±5.5 0.667

LVPWs (mm) 14.5±1.8 15.6±1.9 15.6±2.1 0.08　　 15.1±1.7 15.0±2.1 15.6±2.1 0.459

LVPWd (mm) 8.5±0.9 9.4±0.9 　9.6±1.0* <0.001　 9.6±1.0 　 8.8±1.0* 9.1±1.1 0.039

IVSs (mm) 13.5±1.2 14.3±1.5 　14.5±1.4* 0.009 14.1±1.2 14.0±1.2 14.3±1.8 0.063

IVSd (mm) 8.0±1.1 　 9.3±0.8* 　9.7±0.7* <0.001　 9.4±0.9 　8.6±1.1* 9.0±1.3 0.028

LVWT (mm) 16.5±1.7 　18.7±1.3* 　19.3±1.6* <0.001　 19.0±1.6 　17.5±1.8* 18.1±2.1 0.012

RWT 0.37±0.06 0.37±0.05 0.35±0.05 0.222 0.38±0.05 0.36±0.05 0.36±0.06 0.179

LVMI (g/m2) 75.0±10.9 　98.3±3.4* 　　　117.9±12.9*,† <0.001　 104.3±18.2 　 90.9±18.2* 96.4±22.1 0.036

LA (mm) 38.3±4.1 　39.8±4.1* 　　　43.4±4.8*,† <0.001　 40.1±4.7 40.5±4.6 40.9±5.2 0.831

Aorta (mm) 32.2±3.2 31.8±2.7 33.3±3.2 0.212 32.3±3.2 32.9±3.1 32.0±2.9 0.496

ESV (ml) 31.1±10.9 35.7±9.6 　42.1±15.2* 0.004 34.0±10.5 34.4±9.3 40.5±16.8 0.109

EDV (ml) 83.3±21.9 　102.2±21.2* 　　　119.8±31.3*,† <0.001　 100.0±28.0 96.9±24.4 108.5±33.9 0.315

Systolic function

LVEF (%) 63.3±4.8 65.2±5.0 65.6±4.5 0.207 66.0±4.3 64.4±5.3 63.7±4.6 0.224

FS (%) 41.5±5.2 43.0±6.0 42.0±5.0 0.544 42.1±4.6 41.8±4.7 42.6±6.9 0.86　　 SV (ml) 51.6±11.8 　 66.2±14.2* 　　　 77.7±18.2*,† <0.001　 66.1±19.1 62.5±17.4 68.0±18.4 0.529

Diastolic function

E (cm/s) 65.2±12.4 74.5±20.9 　 82.0±20.5* 0.005 77.5±25.2 67.5±12.4 77.4±17.8 0.093

A (cm/s) 62.3±14.3 　73.7±19.7* 66.0±18.2 0.041 68.6±15.5 66.3±18.4 67.4±20.5 0.902

E/A ratio 1.1±0.3 1.1±0.4 1.4±0.6 0.073 1.2±0.5 1.1±0.4 1.2±0.4 0.394

Dec time (ms) 215.3±37.8 221.0±33.4 207.4±42.6 0.478 214.6±44.8 213.5±27.5 216.0±41.7 0.974

Data given as mean ± SD. *P<0.05 vs. tertile 1, †P<0.05 vs. tertile 2.Dec time, deceleration time; E/A ratio, peak filling velocity during early ventricle diastole/peak filling velocity during early ventricle systole; EDV, end-diastolic volume; ESV, end-systolic volume; FS, fractional shortening; IVSd, end-diastolic interventricular septum thickness; IVSs, end-systolic interventricular septum thickness; LA, left atrium; LVIDd, end-diastolic left ventricular internal diameter; LVIDs, end-systolic left ventric-ular internal diameter; LVPWd, end-diastolic left ventricular posterior wall thickness; LVPWs, end-systolic left ventricular posterior wall thick-ness; LVEF, left ventricular ejection fraction; LVWT, left ventricular wall thickness; RWT, relative wall thickness; SV, stroke volume. Other abbreviations as in Table 1.



LVWT (r=–0.233, P=0.038), IVSd (r=–0.224, P=0.048), LVMI (r=–0.315, P=0.005), and HR (r=–0.431, P<0.001; Table 4). Multivariate stepwise linear regression analysis using the vari-ables with P<0.1 on univariate analysis showed that only LVMI and HR were significant independent determinants of BRS in LC patients (Table 4).

During a mean follow-up of 4.0±0.9 years (range, 1.0–4.9 years), 7 patients (8.5%) died, 5 (71.4%) of them from re-curred hepatocellular carcinoma, 1 from sepsis and the other from pneumonia. Survival, however, was not different among BRS tertiles (log-rank test, P=0.852).

DiscussionThe major finding of this study is that BRS is inversely cor-related with LV morphology, including LVWT, IVSd, and LVMI, in LC patients. Multivariate regression analysis showed

Table 3. Autonomic Function vs. LVMI in LC Patients

Autonomic function Control (n=17)

LC patients (n=82)

P-value (t-test)

Tertiles of LVMI P-value (ANOVA)1 st (n=27) 2nd (n=28) 3rd (n=27)

HRVLF (ln ms2) 5.8±0.9 4.3±1.1 <0.001　 4.6±1.1 4.3±1.1 4.1±1.2 0.19　HRVHF (ln ms2) 5.1±1.0 3.6±1.2 <0.001　 3.6±1.3 3.7±1.0 3.6±1.3 0.98　HRVLF/HF 2.6±1.6 3.2±3.3 0.49　　 4.0±3.7 3.0±2.8 2.9±3.6 0.17　SAPVLF (ln mmHg2) 1.9±0.8 1.0±0.9 <0.001　 0.9±0.7 0.9±0.8 1.1±1.1 0.585

SAPVHF (ln mmHg2) –0.33±0.94　　 –0.30±0.91　　 0.907 –0.3±0.9　 –0.5±0.8　 –0.1±1.0　 0.427

SAPVTP (ln mmHg2) 3.0±0.7 1.3±1.2 <0.001　 1.1±1.2 1.5±1.2 1.4±1.2 0.356

BRSLF (ms/mmHg) 8.3±5.1 5.5±2.8 0.037 6.4±3.1 (5.0–7.6) 5.5±2.8 (4.5–6.8) 4.6±2.3 (3.6–5.5) 0.104

BRSLF (ln ms/mmHg) 1.9±0.7 1.6±0.6 0.022 1.8±0.4 (1.6–1.9) 1.6±0.5 (1.5–1.8) 1.3±0.7 (1.1–1.6) 0.012

Data given as mean ± SD and 95% confidence interval. *P<0.05 vs. tertile 1.HF, high frequency; HRV, HR variability; LF, low frequency; SAPV, systolic arterial pressure variability; TP, total power. Other abbreviations as in Table 1.

Figure 2. Baroreflex sensitivity (BRS) in individual liver cirrhosis patients ac-cording to left ventricular mass index (LVMI) tertile. Straight line, average.

Table 4. Determinants of BRS in LC Patients

BRS

Univariate Multivariate

r P-value B SE P-value

HR –0.431 <0.001　 –0.467 0.003 <0.001

LVPWd –0.19　　 0.09　　LVMI –0.315 0.005 –0.351 0.005 <0.001

RWT –0.12　　 0.09　　LVWT –0.233 0.038

IVSd –0.224 0.048

LVEF –0.206 0.07　　

R2 for model=0.334, P<0.001.Abbreviations as in Tables 1,2.


2812 SONG JG et al.

that only LVMI was independently associated with impaired BRS. The echocardiographic parameters of LV systolic and diastolic function, however, did not differ significantly among the BRS tertiles. These findings indicate that BRS is associ-ated with cardiac morphology in patients with LC, suggesting that cardiovagal baroreflex control of HR may play an impor-tant role in increasing LV mass in patients with LC.

Several studies have shown associations between impaired autonomic function and increased LV mass. For example, ab-normal baroreceptor-cardiac reflex sensitivity in rats has been associated with increased LV mass.24 In patients with end-stage renal disease,12 LVH has been associated with deranged car-diac parasympathetic regulation, showing that the vagal tachy-cardic reserve was depressed in patients with severe LVH, irrespective of their uremic status. Associations between BRS and LV morphology and function have been demonstrated in patients with essential hypertension, including associations of BRS with LV geometry and LVMI.8 In addition, decreased BRS was inversely correlated with changes in LVMI in chil-dren with obstructive sleep apnea.17 Consistent with these re-sults, we found that reduced cardiovagal BRS was associated with increased LVMI in LC patients. Although the mechanisms underlying this association remain unclear, they may be at least partly due to the relationship between reduced BRS and hemo-dynamics and humoral systems.1,25 Reduced BRS in cirrhosis is related to determinants of the hyperdynamic circulation and to the renin-angiotensin-aldosterone system, including with increased baseline plasma concentrations of renin, aldosterone and norepinephrine.25 Moreover, significantly impaired urinary sodium and water excretion was observed in patients with vagal dysfunction compared with patients with normal cardiovascu-lar tests, with these impairments associated with higher circu-lating concentrations of noradrenaline, renin, and anti-diuretic hormones after water loading.1

Several factors may explain the cause of cardiac morpho-logical changes. In hypertensive patients, changes in heart mor-phology may be due to neuroendocrine activation and auto-nomic dysfunction, with chronic sympathetic overactivity and norepinephrine release being found to play a crucial role in the determination of human cardiac hypertrophy.26 In addition, reduced BRS has been associated independently with LV mor-phology in hypertensive patients,8 consistent with the present findings. Taken together, these findings indicate that impaired BRS may play an important role in LV morphological modi-fications in both hypertensive and LC patients. Patients with LC have shown enhanced sympathetic nervous system activ-ity, including increases in sympathetic nerve burst frequency and circulating catecholamine concentrations.13 Because most LC patients do not have hypertension, changes in LC patients may be due to stimulation mediated by baroreceptors and vol-ume receptors, resulting from low AP and reduced central blood volume.13,14 Therefore, in LC patients, cardiac remodeling may be caused primarily by sympathetic overactivity together with the induction of an autonomic imbalance and relative para-sympathetic withdrawal and, finally, with BRS impairment.8 Alternatively, the hyperdynamic circulation in patients with advanced LC and portal hypertension may be due primarily to vagal autonomic dysfunction.27 In rats with portal hyperten-sion, the central cardiovascular regulatory nuclei initiate hy-perdynamic circulation in response to a gut signal associated with portal hypertension, suggesting that central neural activa-tion depends on vagal afferent nerves and thus plays a primary role in the development of hyperdynamic circulation.28 Further studies are required to determine the mechanisms and links between impaired autonomic function and hyperdynamic cir-

culation resulting in cardiac morphological modifications in patients with cirrhotic cardiomyopathy.

Determination of cardiac morphological changes in LC pa-tients has yielded somewhat different results depending on the methods used.29 In an animal study, eccentric LVH developed in bile duct-ligated rats in conjunction with the development of a hyperdynamic syndrome.30 LC patients with hyperdy-namic circulation have normal or increased LV mass, and high-er LV end-systolic and -diastolic diameters than controls.29,31 In the present study, when we stratified LC patients by LVMI tertiles, we found that the highest tertile had a mean LVMI of 118±13 g/m2, meeting the inclusion criteria (LVMI >108 g/m2 for women).

Cardiovascular oscillation is dependent on very complex interactions among hemodynamic, humoral and electrophysi-ological variables, which are integrated by the autonomic and central nervous systems.18,20,21 BRS is an essential part of this regulatory system involved in the maintenance of circulatory stability. A decrease in BRS may result in blood pressure in-stability, contributing to organ damage and other undesirable outcomes.32 Of the parameters reflecting cardiovascular auto-nomic function in LC patients, only BRS differed significantly among the tertiles of LVMI in this study. In contrast, none of the autonomic parameters, such as LF SAP and HF RRI oscil-lations, which reflect sympathetic and vagal control of the vas-culature and heart, respectively, differed. Therefore, the pres-ent findings suggest that baroreceptors play more important roles than other autonomic function parameters such as HRV and BPV in maintaining circulatory stability and preventing cardiac morphological changes in LC patients.

The present study had several limitations. First, although we measured spontaneous cardiovagal BRS under basal resting conditions using spectral analysis, cardiovagal BRS can also be determined during a blood-pressure-perturbing period, by a modified Oxford technique or the Valsalva maneuver. The main limitation of spontaneous baroreflex gain based on con-ventional frequency domain approaches in the closed-loop and the non-perturbed environment is that it cannot identify wheth-er the observed changes in SAP and RRI are the result of negative or positive feedback mechanism between SAP and RRI. Moreover, the effects of other variables such as respira-tion or chemoreceptors on the estimate of the baroreflex gain are not clearly considered.33 Second, we measured indirect pa-rameters of the sympathetic nervous system rather than direct measurements of sympathetic nerve activity and norepineph-rine concentration. Third, owing to the small sample size of the present study, we could not determine whether liver trans-plantation would cause alterations in cardiac function nor whether BRS could predict outcome among BRS tertiles, be-cause 1- and 5-year survival rates of liver transplantation at Asan Medical Center are approximately 93.5% and 83.2%, re-spectively.34 Therefore, further study using a large number of patients is needed to confirm these issues.

In conclusion, the present results demonstrate that impair-ment of the baroreflex is related to worsening in cardiac re-modeling, resulting in increased LV mass. These findings may suggest a relationship between cardiovagal BRS control of HR and change in heart morphology of LC patients.

DisclosuresSpecific Author Contributions: Principal investigator, patient recruitment, patient evaluation, data collection, manuscript preparation and editing: Gyu-Sam Hwang; data collection, patient evaluation, manuscript writing, and statistics: Jun-Gol Song; data collection and patient evaluation: Young-Kug Kim; data collection and background research: Won-Jung Shin. Fi-



nancial Support: None. Potential Conflicts of Interest: None.

References 1. Hendrickse MT, Triger DR. Vagal dysfunction and impaired urinary

sodium and water excretion in cirrhosis. Am J Gastroenterol 1994; 89: 750 – 757.

2. Hendrickse MT, Triger DR. Peripheral and cardiovascular autonom-ic impairment in chronic liver disease: Prevalence and relation to hepatic function. J Hepatol 1992; 16: 177 – 183.

3. Valeriano V, Funaro S, Lionetti R, Riggio O, Pulcinelli G, Fiore P, et al. Modification of cardiac function in cirrhotic patients with and without ascites. Am J Gastroenterol 2000; 95: 3200 – 3205.

4. Ates F, Topal E, Kosar F, Karincaoglu M, Yildirim B, Aksoy Y, et al. The relationship of heart rate variability with severity and prognosis of cirrhosis. Dig Dis Sci 2006; 51: 1614 – 1618.

5. Hendrickse MT, Thuluvath PJ, Triger DR. Natural history of auto-nomic neuropathy in chronic liver disease. Lancet 1992; 339: 1462 – 1464.

6. Billman GE, Schwartz PJ, Stone HL. Baroreceptor reflex control of heart rate: A predictor of sudden cardiac death. Circulation 1982; 66: 874 – 880.

7. Cerati D, Schwartz PJ. Single cardiac vagal fiber activity, acute myo-cardial ischemia, and risk for sudden death. Circ Res 1991; 69: 1389 – 1401.

8. Milan A, Caserta MA, Del Colle S, Dematteis A, Morello F, Rabbia F, et al. Baroreflex sensitivity correlates with left ventricular morphol-ogy and diastolic function in essential hypertension. J Hypertens 2007; 25: 1655 – 1664.

9. Lantelme P, Khettab F, Custaud MA, Rial MO, Joanny C, Gharib C, et al. Spontaneous baroreflex sensitivity: Toward an ideal index of cardiovascular risk in hypertension? J Hypertens 2002; 20: 935 – 944.

10. Okada N, Takahashi N, Yufu K, Murozono Y, Wakisaka O, Shinohara T, et al. Baroreflex sensitivity predicts cardiovascular events in pa-tients with type 2 diabetes mellitus without structural heart disease. Circ J 2010; 74: 1379 – 1383.

11. Yufu K, Takahashi N, Okada N, Wakisaka O, Shinohara T, Nakagawa M, et al. Gender difference in baroreflex sensitivity to predict car-diac and cerebrovascular events in type 2 diabetic patients. Circ J 2011; 75: 1418 – 1423.

12. Mircoli L, Rivera R, Bonforte G, Fedele L, Genovesi S, Surian M, et al. Influence of left ventricular mass, uremia and hypertension on vagal tachycardic reserve. J Hypertens 2003; 21: 1547 – 1553.

13. Henriksen JH, Moller S, Ring-Larsen H, Christensen NJ. The sym-pathetic nervous system in liver disease. J Hepatol 1998; 29: 328 – 341.

14. Moller S, Henriksen JH. Cirrhotic cardiomyopathy: A pathophysio-logical review of circulatory dysfunction in liver disease. Heart 2002; 87: 9 – 15.

15. Braverman AC, Steiner MA, Picus D, White H. High-output conges-tive heart failure following transjugular intrahepatic portal-systemic shunting. Chest 1995; 107: 1467 – 1469.

16. Xie HH, Miao CY, Liu JG, Su DF. Importance of blood pressure variability in organ protection in spontaneously hypertensive rats treated with combination of nitrendipine and atenolol. Acta Pharma-col Sin 2002; 23: 1199 – 1204.

17. McConnell K, Somers VK, Kimball T, Daniels S, VanDyke R, Fenchel M, et al. Baroreflex gain in children with obstructive sleep apnea. Am

J Respir Crit Care Med 2009; 180: 42 – 48.18. Parati G, Casadei R, Groppelli A, Di Rienzo M, Mancia G. Com-

parison of finger and intra-arterial blood pressure monitoring at rest and during laboratory testing. Hypertension 1989; 13: 647 – 655.

19. Cho SK, Hwang GS, Kim YK, Huh IY, Hahm KD, Han SM. Low-dose atropine amplifies cardiac vagal modulation and increases dy-namic baroreflex function in humans. Auton Neurosci 2005; 118: 108 – 115.

20. Lee K, Picard G, Beske SD, Hwang GS, Taylor JA. Effects of fitness and age on the response to vagotonic atropine. Auton Neurosci 2008; 139: 60 – 67.

21. Heart rate variability: Standards of measurement, physiological in-terpretation and clinical use: Task Force of the European Society of Cardiology and the North American Society of Pacing and Electro-physiology. Circulation 1996; 93: 1043 – 1065.

22. Robbe HW, Mulder LJ, Ruddel H, Langewitz WA, Veldman JB, Mulder G. Assessment of baroreceptor reflex sensitivity by means of spectral analysis. Hypertension 1987; 10: 538 – 543.

23. de Simone G, Devereux RB, Roman MJ, Schlussel Y, Alderman MH, Laragh JH. Echocardiographic left ventricular mass and electrolyte intake predict arterial hypertension. Ann Intern Med 1991; 114: 202 – 209.

24. Minami N, Head GA. Relationship between cardiovascular hyper-trophy and cardiac baroreflex function in spontaneously hypertensive and stroke-prone rats. J Hypertens 1993; 11: 523 – 533.

25. Moller S, Iversen JS, Henriksen JH, Bendtsen F. Reduced baroreflex sensitivity in alcoholic cirrhosis: Relations to hemodynamics and humoral systems. Am J Physiol Heart Circ Physiol 2007; 292: H2966 – H2972.

26. Schlaich MP, Kaye DM, Lambert E, Sommerville M, Socratous F, Esler MD. Relation between cardiac sympathetic activity and hyper-tensive left ventricular hypertrophy. Circulation 2003; 108: 560 – 565.

27. Moller S. Vagal mediation of systemic cardiovascular response to portal hypertension: From experimental studies in animals to treat-ment of humans. Gut 2008; 57: 884 – 885.

28. Liu H, Schuelert N, McDougall JJ, Lee SS. Central neural activation of hyperdynamic circulation in portal hypertensive rats depends on vagal afferent nerves. Gut 2008; 57: 966 – 973.

29. Wong F, Liu P, Lilly L, Bomzon A, Blendis L. Role of cardiac struc-tural and functional abnormalities in the pathogenesis of hyperdy-namic circulation and renal sodium retention in cirrhosis. Clin Sci (Lond) 1999; 97: 259 – 267.

30. Inserte J, Perello A, Agullo L, Ruiz-Meana M, Schluter KD, Escalona N, et al. Left ventricular hypertrophy in rats with biliary cirrhosis. Hepatology 2003; 38: 589 – 598.

31. Friedman HS, Fernando H. Ascites as a marker for the hyperdynamic heart of Laennec’s cirrhosis. Alcohol Clin Exp Res 1992; 16: 968 – 970.

32. Frattola A, Parati G, Cuspidi C, Albini F, Mancia G. Prognostic value of 24-hour blood pressure variability. J Hypertens 1993; 11: 1133 – 1137.

33. Porta A, Baselli G, Rimoldi O, Malliani A, Pagani M. Assessing baro-reflex gain from spontaneous variability in conscious dogs: Role of causality and respiration. Am J Physiol Heart Circ Physiol 2000; 279: H2558 – H2567.

34. Lee SG, Hwang S, Kim KH, Ahn CS, Moon DB, Ha TY, et al. To-ward 300 liver transplants a year. Surg Today 2009; 39: 367 – 373.

Documents

Changes in Cardiovagal Baroreflex Sensitivity Are Related