Cirrhotic cardiomyopathy is associated with poor outcomes in patients with cirrhosis. We investigated if subclinical cardiac morphologic and functional modifications can influence survival in patients with cirrhosis during follow-up. A series of patients with cirrhosis without cardiovascular or pulmonary disease underwent standard and tissue Doppler echocardiography to assess left ventricular geometry, systolic/diastolic function, and the main haemodynamic parameters. After baseline evaluation 115 patients with cirrhosis were followed up for at least 6 years. During follow-up 54 patients died (47%). On univariate analysis, age, body surface area (BSA), Model for End-Stage Liver Disease (MELD), mean arterial pressure, heart rate, cardiac index, systemic vascular resistance index, and the ratio of transmitral Doppler early filling velocity to tissue Doppler early diastolic mitral annular velocity (E/è) were associated with increased risk of death. In a Cox hazard regression analysis including these factors and other hypothesized important factors (but not MELD), increased age (P = 0.04) and left atrial dimension (P = 0.005) and lower BSA (P = 0.03) were the strongest predictors of death. When MELD was included in the analysis, the main predictors were MELD, age, and BSA. When multivariate analysis was performed incorporating only cardiovascular parameters, increased E/è (P = 0.003) and heart rate (P = 0.03) and reduced mean blood pressure (P = 0.01) were significantly associated with poor prognosis. Conclusion: In a large cohort of patients with cirrhosis and after a long follow-up, MELD, age, and BSA were the main predictors of death; among cardiovascular parameters, left atrium enlargement, increased heart rate and E/è, and reduced mean blood pressure were independent predictors of death. (Hepatology 2018).

Abbreviations

ASE: American Society of Echocardiography
BSA: body surface area
CI: cardiac index
CO: cardiac output
DD: diastolic dysfunction
EACVI: European Association of Cardiovascular Imaging
EAE: European Association of Echocardiography
E/e′: early diastolic transmitral and myocardial velocity on TDI ratio
EF: ejection fraction
HRS: hepatorenal syndrome
LA: left atrial
LV: left ventricular
LVH: LV hypertrophy
LVM: LV mass
MAP: mean arterial pressure
MELD: Model for End-Stage Liver Disease
SV: stroke volume
SW: stroke work
TDI: tissue Doppler imaging

Cardiovascular abnormalities play a marked role in the pathogenesis of major complications of cirrhosis such as ascites, hyponatremia, and hepatorenal syndrome (HRS). Cirrhotic cardiomyopathy is a very complex condition characterized by impaired myocardial contractility, left ventricular (LV) hypertrophy (LVH), diastolic dysfunction (DD), impaired chronotropic function, and electrophysiological abnormalities in the absence of other known causes of heart disease.1-3 In patients with cirrhosis, LV enlargement and a high prevalence of LVH despite low peripheral resistances and low to normal blood pressure have been reported and referred to the interaction between the high renin and aldosterone values and the work overload associated with the hyperdynamic circulatory state.3, 4 Regarding DD, its prevalence in patients with cirrhosis varies between 37% and 43% when assessed according to the classification proposed by the American Society of Echocardiography (ASE) and the European Association of Echocardiography (EAE) in 2009,5 being not associated with the stage of liver disease but strongly related to the increase in LV mass (LVM)3 but with no relevance in terms of prognosis and particularly with survival.6-8

Cirrhotic cardiomyopathy represents a condition of latent heart failure which manifests only under stress, resulting in a blunted increase in cardiac index (CI) and cardiac output (CO) during exercise or pharmacologic stimuli.2, 9 In fact, even if at rest no evidence of subclinical systolic dysfunction exists,3 along with the development and worsening of ascites, CO and CI do not further increase to compensate for the continuous lowering of peripheral vascular resistances and blood pressure; and therefore, their maintenance becomes critically dependent upon the heart rate. Moreover, a decreased CI has been identified as an independent predictor of development of HRS.10

In the literature, few studies have investigated the predictive value of cardiac parameters in patients with cirrhosis, with confounding results.6, 10-12

We performed a prospective study assessing cardiac dimensions, systolic and diastolic function, and hemodynamic parameters in a large cohort of patients with different stages of cirrhosis using a noninvasive state-of-the-art echocardiography technique, with the aim of investigating the predictive value of these parameters in the clinical course of these patients.

Patients and Methods

STUDY POPULATION

From 2009 to 2012 a large series of consecutive outpatients with cirrhosis underwent a detailed haemodynamic evaluation and standard transthoracic Doppler echocardiography in our Department. Exclusion criteria were arterial hypertension and history of cardiovascular disease, diabetes, or heart valve disease. Coronary heart disease was excluded in all participants on the basis of symptoms, negative family history, a normal standard 12-lead electrocardiogram, and normal wall motion on two-dimensional echocardiographic examination.

If used, β-blockers were stopped 48 hours before echocardiography. The diagnosis of liver cirrhosis was based on clinical, biochemical, imaging, and endoscopic findings in all patients. The presence of ascites was detected clinically and confirmed by an abdomen ultrasound examination. The diagnosis of refractory ascites was based on the criteria of the International Club of Ascites.13 In patients with refractory ascites, the hemodynamic evaluation was performed soon after the therapeutic paracentesis.

ECHOCARDIOGRAPHIC PARAMETER ANALYSIS

All echocardiographic examinations were performed by the same expert sonographer (M.C.) using a General Electric Vivid 7 ultrasound machine with a 2.5-MHz transducer. Echocardiograms were stored digitally and analyzed offline by a single experienced observer (M.C.). All measurements were made according to the guidelines of the ASE.14 LV end-diastolic and end-systolic diameters and wall thickness were assessed in M-mode. LV ejection fraction (EF) and fractional shortening were measured in biplane two-dimensional mode using the Simpson method.14 At a given contractile state, however, EF increases as LV geometry becomes more concentric.15 Thus, midwall fractional shortening was also calculated to assess underlying systolic dysfunction in the setting of concentric hypertrophy.16, 17 LVM was estimated using the anatomically validated formula of Devereux et al.14 LVM was normalized both by body surface area (BSA) and by height in meters to the power of 2.7. The latter has been demonstrated to maximize risk prediction.18 Criteria for LVH were LVM/height ≥48 g/m^2.7 for men and ≥44 g/m^2.7 for women according to the current recommendations.14

The following parameters were calculated: mean arterial pressure (MAP; systolic pressure + [2 × diastolic pressure]/3); stroke volume (SV) was computed as the difference between end-diastolic and end-systolic LV volume and used as a direct indicator of LV volume load; SV was also corrected by the BSA to obtain the SV index. CO was calculated as the product of SV and heart rate; CI was calculated as CO adjusted by the BSA; the systemic vascular resistance index was calculated as the product of MAP and 80/CI;19 the stroke volume-pulse pressure ratio was calculated as an index of arterial compliance; stroke work (SW), a measure of total cardiac workload, was calculated as the product of systolic blood pressure (pressure load) and SV (volume load) and converted into gram-meters per beat by multiplying by the conversion factor 0.0014.20

Pulsed Doppler recordings at the level of the mitral valve tips were obtained from apical four-chamber scans to measure early (E) and late (A) diastolic filling velocities, their ratio (E/A), and the early wave deceleration time (DT).5

The tissue Doppler imaging (TDI) program was set to pulse-wave Doppler mode. Filters were set to exclude high-frequency signals. Gains were minimized to allow a clear tissue signal with minimal background noise. The TDI of the diastolic velocities was obtained from the apical four-chamber view. The recorded wall was positioned in the center of the sector. A 1.5-mm sample volume was placed at the septal corner of the mitral valve annulus. The angle between the Doppler beam and the longitudinal motion of the septal mitral valve annulus was minimized as well. All Doppler parameters were recorded at a horizontal speed of 100 mm/second. The average values obtained from at least three consecutive cardiac cycles were taken into consideration. Early diastolic peak velocity of septal mitral annulus (è) was obtained, and the E/è ratio was derived, reflecting the LV filling pressure.21

Strain rate parameters were analyzed offline by the same operator (M.C.). Placing the region of interest on the medial corner of the mitral annulus, we assessed its peak systolic tissue velocity (septal). Placing the region of interest (6 × 4 mm) on the basal portion of the inferior interventricular septum, we assessed septal peak systolic strain, the systolic strain rate, and early and late diastolic strain rate parameters.22 Early and late diastolic strain rate parameters were derived, which is considered an index of segmental relaxation–segmental diastolic dysfunction.23

The diagnosis and grading of LV DD (Table 1) were based on both the ASE/EAE 2009 recommendations,5 the ASE and European Association of Cardiovascular Imaging (EACVI) 2016 recommendations,21 and the 2016 Thorax Center algorithm, which starts with assessment of the transmitral flow velocities (E/A ratio and DT) and subsequently the mitral annulus TDI parameters (E/e′) and the left atrial (LA) dimension (anteroposterior diameter).24 This algorithm was claimed to be much simpler, faster, and therefore more reproducible than the ASE/EAE algorithm.24 The high reproducibility of all echocardiographic parameters, in our laboratory, has also been reported.3

Table 1. Definitions of DD

Algorithm

Parameters Considered

Diagnosis of DD

Grading of DD

ASE/EAE 20095

LA volume

Annular e′ velocity (septal and lateral e′)

Average E/e′

E/A

Septal e′ <8 cm/second or lateral e′ <10 cm/second and LA maximum volume index ≥34 mL/m²

E/A

Average E/e′

ASE/EACVI 201621

LA volume

Annular e′ velocity (septal and lateral e′)

Average E/e′

Peak velocity of tricuspid regurgitation

DD is present if more than half of the available parameters meet these cutoff values:

-Annular e′ velocity (septal e′ <7 cm/second, lateral e′ <10 cm/second)

-Average E/e′ ratio >14 (if only the lateral e′ or septal e′ velocity is available, a lateral E/e′ ratio >13 or a septal E/e′ >15 is considered abnormal)

-LA maximum volume index >34 mL/m²

-Peak velocity of tricuspid regurgitation >2.8 m/second

Thorax Center 201624

E/A

E/e′ (using septal e′)

LA diameter

DT ≥ 220 or < 220

DT ≥160 or <160

E/e′ <11, 11-15, or >15

LA diameter >40 mm

E/A

E/e′ (using septal e′)

LA diameter

Abbreviations: E/A, early and late diastolic velocity ratio; DT, early wave deceleration time.

STATISTICAL ANALYSIS

The data are expressed as mean ± SD (or median and range), as appropriate. All echocardiographic and hemodynamic parameters showed a normal distribution, which was formally verified by the Kolmogorov-Smirnov test. Survival free from death was estimated with the Kaplan-Meier method.

Potential predictors of mortality were analyzed with univariate Cox regression, and those statistically significant at the 10% level were introduced in a multivariate Cox regression model with backward selection of the variables. The significance level for the multivariate analysis was 5%.

Three multivariate models were considered: (1) with all predictors significant at the univariate analysis; (2) dropping Model for End-Stage Liver Disease (MELD); (3) dropping age, BSA, and MELD. Proportionality assumption was checked graphically. The results of the Cox regression were reported as P value, hazard ratio, and 95% confidence interval.

Results

We recruited 162 consecutive outpatients with cirrhosis. Patients were excluded if they had arterial hypertension (n = 11), history of cardiovascular disease (n = 5), diabetes mellitus (n = 16), heart valve disease (n = 7), or coronary heart disease (n = 5). Three patients were lost at follow-up. Therefore, for the prospective analysis we considered 115 patients with cirrhosis: 48 with refractory ascites, 31 with responsive ascites, 36 without ascites. The clinical, demographic, and biochemical data are reported in Table 2. Patients with cirrhosis were mostly male and in different stages of liver disease of different etiologies: 53 patients (46%) had alcoholic cirrhosis, 51 patients (44%) had postviral cirrhosis, 11 patients (10%) had other forms of cirrhosis.

Table 2. Comparison of Clinical, Demographic, Echocardiographic, and Hemodynamic Characteristics of Patients Divided According to Whether They Were Alive or Dead After 6 Years of Follow-Up

Variable	Alive (n = 61)	Dead (n = 54)	P
Age (years)	55 ± 11	62 ± 11	0.0004
Gender M/F (%)	75/25	68/32	NS
Etiology of cirrhosis (%) (alcohol/viral/other)	(44/48/8)	(49/40/11)	NS
Weight (kg)	78 ± 12	73 ± 14	0.02
BSA (kg/m²)	1.9 ± 0.2	1.8 ± 0.2	0.02
MELD	11 ± 5	16 ± 6	0.0001
Child-Pugh score	7 ± 2	9 ± 2	0.0001
Obesity (%)	10	12	NS
MAP (mm Hg)	95 ± 12	90 ± 11	0.01
Heart rate (beats/minute)	70 ± 11	77 ± 13	0.002
Creatinine (μmol/L)	82 (75-88)	122 (95-149)	0.008
Cardiac dimensions
Left atrial size (cm)	3.8 ± 0.4	4.0 ± 0.5	0.06
LVEDD (mm)	48 ± 4	48 ± 3	NS
LVEDVol (mL)	110 ± 23	107 ± 17	NS
LVM/height (g/m^2,7)	46 ± 10	46 ± 10	NS
LVH (%)	39	41	NS
LV systolic function
EF (%)	70 ± 6	69 ± 5	NS
MWFS	17 ± 2	17 ± 2	NS
Septal strain (%)	23 ± 4	23 ± 3	NS
Hemodynamic parameters
SV (mL/beat)	76 ± 16	74 ± 12	NS
SW (g/beat)	14 ± 4	13 ± 3	0.04
CI (L/minute/m²)	2.73 ± 0.6	3.06 ± 0.58	0.008
CO (L/minute)	5.3 ± 1.3	5.6 ± 1.0	NS
SVRI (dyn s/cm⁵ m²)	2,946 (2,750-3,142)	2,481 (2,285-2,677)	0.0009
LV diastolic function
E/A	1.14 ± 0.38	1.11 ± 0.4	NS
DT/HR	3.48 ± 1.02	3.35 ± 1.17	NS
Septal E′ (cm/second)	8.8 ± 2.1	8.2 ± 2.2	NS
E/e′	9.8 ± 2.9	11.3 ± 3	0.03
SR e/a	1.43 ± 0.65	1.50 ± 0.6	NS

Data are reported as mean ± SD or median (and interquartile range), as appropriate.
Abbreviations: DT, early wave deceleration time; E/A, early and late diastolic velocity ratio; HR, heart rate; LVEDD, LV end-diastolic diameter; LVEDVol, LV end-diastolic volume; MWFS, midwall fractional shortening; NS, nonsignificant; SR e/a, early and late diastolic TDI velocities ratio; SVRI, systemic vascular resistance index.

All patients with cirrhosis showed an increase in LV end-diastolic dimensions and volumes, in wall thickness, and in LVM, with an important increase in LVH (40%). This increase in LVM was less evident in patients with more advanced liver disease. Patients with alcoholic cirrhosis did not present the echocardiographic features of alcoholic cardiomyopathy (i.e., increased LV volumes, subclinical systolic dysfunction, or reduced EF) compared with other forms of cirrhosis.

In the overall population the prevalence of LV DD was 30% using ASE/EAE 2009, 17% using ASE/EACVI 2016, and 48% using the Thorax Center algorithm.

SURVIVAL

During a median follow-up of 5 years (range 20 days-6 years), 54 patients died (47%): 20 patients in the first 6 months and 27 patients in the first year. Main causes of death were complications of cirrhosis (gastrointestinal hemorrhage, HRS, liver failure, sepsis, hepatocellular carcinoma).

On univariate analysis the parameters associated with death at follow-up were age, BSA, MELD, MAP, heart rate, serum creatinine, SW, CI, systemic vascular resistance index, and E/e′ (Table 2).

In a Cox hazard regression analysis incorporating these parameters (but not MELD) and all ecochardiographic and hemodynamic parameters (model 2), the most significant predictors of death were age, BSA, and LA dimension (Table 3 and Fig. 1). When MELD was included in the Cox analysis (model 1), MELD (P < 0.0001), age (P = 0.008), and BSA (P = 0.03) were the best predictors and the prognostic value of LA dimension was blunted.

Table 3. Clinical, Echocardiographic, and Hemodynamic Parameters Associated With Death During Follow-Up (Cox's Regression Analysis)

Predictors	Hazard Ratio (95% Confidence Interval)	P
MELD	1.15 (1.08-1.23)	0.0001
BSA (kg/m²)	0.09 (0.01-0.81)	0.03
Age (years)	1.06 (1.02-1.10)	0.008
Dropping MELD
LA dimension (cm)	3.42 (1.45-8.06)	0.005
BSA (kg/m²)	0.11 (0.01-0.81)	0.03
Age (years)	1.04 (1.00-1.07)	0.04
Dropping age, BSA, and MELD
E/e′	1.22 (1.07-1.38)	0.003
MAP (mm Hg)	0.96 (0.93-0.99)	0.011
HR (beats/minute)	1.04 (1.00-1.08)	0.03

Abbreviation: BSA: body surface area; E/e' early diastolic transmitral and myocardial velocity on TDI ratio; MAP, mean arterial pressure; HR, heart rate.

Details are in the caption following the image — **Figure 1**
Open in figure viewer PowerPoint

Kaplan-Meier curves showing the probability of survival at 6 years of patients with cirrhosis according to the E/e′ ratio (A) and LA dimension (B) assessed at baseline.

When multivariate analysis was performed incorporating only cardiovascular parameters (model 3), increased E/e′ and heart rate and reduced MAP were significantly associated with the occurrence of death (Table 3 and Fig. 1).

Regarding LV DD, if defined according to the ASE/EAE 2009 recommendations, the prevalence in the alive and dead groups was 28% and 31%, respectively (nonsignificant). If defined according to the ASE/EACVI 2016 recommendations, the prevalence in the alive and dead groups was 15% and 21%, respectively (nonsignificant). At variance, when assessed by the newly proposed Thorax Center algorithm, the prevalence in the alive and dead groups was 42% and 63%, respectively (nonsignificant).

No association with prognosis emerged regarding LV DD when assessed with both the ASE/EAE 2009 and ASE/EACVI 2016 recommendations or with the Thorax Center algorithm.

PATIENTS WITH ALCOHOL-RELATED CIRRHOSIS

When we performed the univariate analysis only in the 53 patients with alcohol-related cirrhosis, the parameters associated with death at follow-up were increased MELD (P = 0.0003), reduced systolic blood pressure (P = 0.04) and MAP (P = 0.05), and lower LV SW (P = 0.03). At multivariate Cox regression analysis the main predictors of poor outcome were increased MELD (P < 0.0001) and reduced LV SW (P = 0.003); if MELD was excluded, the best independent predictor remained low LV SW (P = 0.03).

TYPE 1 HRS DEVELOPMENT DURING FOLLOW-UP

During the observation period 12 patients developed type 1 HRS. The median time between the initial evaluation and the development of the complication was 3 months (range 20 days-3.5 years). We compared the 48 patients with refractory ascites who did and did not develop type 1 HRS. Only patients with refractory ascites were compared because no patients without ascites and only 1 patient with responsive ascites developed the syndrome. The 11 patients with refractory ascites who developed type 1 HRS showed significantly lower body mass index, higher MELD and Child-Pugh scores, and increased systolic blood pressure, pulse pressure, LV SW, and CI (Table 4). A multivariate analysis identified increased MELD (P = 0.006) and CI (P = 0.01) as the main predictors of type 1 HRS.

Discussion

In a previous study we performed a cross-sectional evaluation of our cohort to investigate for the occurrence of LV systolic and/or diastolic dysfunction. We observed that patients with cirrhosis showed a reduction in peripheral vascular resistance; a compensatory hyperdynamic syndrome; and significant increases in CI, CO, and cardiac work, with consequent increases in the prevalence of LVH and associated LV DD. The latter was strongly related to increased age and LVM but not related to the etiology or the severity of the liver disease. Systolic dysfunction was not evident at rest, becoming evident as a lack of an appropriate LV contractile response in the advanced stage of the liver disease, when CO became critically dependent upon the heart rate (thus suggesting a possible detrimental effect of β-blockers in these patients). To better define the importance of these cardiovascular parameters as predictors of poor outcomes, we performed a prospective study with a long follow-up. To our knowledge, we studied the largest cohort of patients with cirrhosis (115 patients) for the longest follow-up (at least 6 years). The long follow-up justifies the high incidence of mortality (47% at the end of follow-up). In smaller cohorts and with shorter follow-up (1-2 years) a similar incidence of mortality was reported compared with our findings (27% in the first year in our study, 21% in the study of Ruiz-del-Arbol et al.,12 27% in the study of Merli et al.11).

Previous studies which evaluated the prognostic role of cardiac parameters in patients with cirrhosis have shown conflicting results (Table 5). Merli et al. investigated the prognostic value of cardiovascular parameters in 90 patients with cirrhosis followed for 24 months. An increased LA dimension and a decreased LVM, but not E/e′ ratio, correlated with survival.11 Ruiz-del-Arbol in 80 patients with cirrhosis followed up for 12 months identified the E/e′ ratio as an independent predictive factor of mortality.12 Survival was significantly greater in the E/e′ <10 group compared to the E/e′ >10 group. Of interest, basal LVM was significantly increased in patients who presented type 1 HRS and in patients dead at 12 months compared with alive patients, in contrast with the results of Merli et al.

Table 4. Comparison of Clinical, Demographic, Echocardiographic, Hemodynamic, and Biochemical Features of Patients With Refractory Ascites Who Presented Type 1 HRS and Who Did Not Develop Type 1 HRS After 6 Years of Follow-Up

Variable	With Type 1 HRS (n = 11)	Without Type 1 HRS (n = 37)	P
Age (years)	65 ± 9	60 ± 11	NS
Weight (kg)	68 ± 6	78 ± 14	NS
BMI (kg/m²)	23 ± 2	26 ± 4	0.03
MELD	21 ± 5	16 ± 6	0.03
Child-Pugh score	11 ± 2	10 ± 2	<0.05
SBP (mm Hg)	126 ± 13	116 ± 12	0.03
PP (mm Hg)	55 ± 13	44 ± 10	0.003
Cardiac dimensions
LA size (cm)	4.2 ± 0.4	3.9 ± 0.5	NS
LVEDVol (mL)	112 ± 21	107 ± 19	NS
LVM/height (g/m^2.7)	44 ± 5	44 ± 9	NS
Hemodynamic parameters
SW (g/beat)	15 ± 4	12 ± 2	0.005
CI (L/minute/m²)	3.40 ± 0.50	2.83 ± 0.59	0.03
CO (L/min)	6.2 ± 1.3	5.5 ± 1.0	0.08
SVRI (dyn s/cm⁵ m²)	2,122 (1,803-2,440)	2,547 (2,262-2,831)	NS
Biochemical parameters
Creatinine (μmol/L)	122 (91-154)	118 (90-146)	NS
Plasma renin activity (ng/mL/hour)	32 (26-37)	29 (6-52)	NS
Plasma aldosterone (pmol/L)	2,686 (933-5,903)	2,308 (406-7,066)	NS

Data are reported as mean ± SD or median (and interquartile range), as appropriate.
Abbreviations: BMI, body mass index; LVEDVol, LV end-diastolic volume; NS, nonsignificant; PP, pulse pressure; SBP, systolic blood pressure; SVRI, systemic vascular resistance index.

Table 5. Studies Which Evaluated the Prognostic Role of Cardiac Parameters in Patients With Cirrhosis

Reference	Number of Patients	Follow-Up (Years)	Cardiovascular Prognostic Parameters
Krag et al.10	24	1	↓ CI (<1.5 L/minute/m²)
Merli et al.11	90	2	↑ LA dimension ↓ LV mass
Ruiz-Del Arbol et al.12	80	1	↑ E/e′ LVDD
Sampaio et al.6	98	0.5	↓ MAP
Cesari et al.^{(present study)}	115	6	↑ LA dimension ↑ E/e′ ↑ heart rate ↓ MAP

Sampaio et al. evaluated the prognostic impact of cardiac systolic and diastolic dysfunction (assessed by tissue Doppler and speckle tracking analysis) in 98 patients with cirrhosis followed up for 6 months.6 None of the echocardiographic parameters were associated with the occurrence of death. The only independent predictors of mortality were a Child score >10 points and a MAP below the median (Table 5). Krag et al. reported that a low CI (below 1.5 L/minute/m²) assessed by gated myocardial perfusion imaging was associated with the development of HRS type 1 within 3 months in 24 patients with advanced cirrhosis and ascites, more important than MELD score to predict prognosis (Table 5).10

As expected, in our study MELD, age, and BSA were the main predictors of poor outcome. Of interest, when MELD was excluded by the Cox analysis, the increased LA dimension entered in the model and was an independent predictor of death (Table 3 and Fig. 1).

In our models, LV volume, mass, and prevalence of hypertrophy did not predict death, as well as the parameters of LV systolic function (midwall fractional shortening, EF, and septal strain).

When multivariate analysis was performed incorporating only cardiovascular parameters, increased E/e′ and heart rate and reduced MAP were significantly associated with the occurrence of death (Table 3 and Fig. 1).

In line with other studies (Table 5), we observed a predictive value of increased LA dimension, increased E/e′ (Fig. 1), and low MAP, whereas no effect of LVM was evident. Because increased LA dimension and E/e′ are strongly associated with increased LV filling pressure,25 we suggest that in patients with cirrhosis with normal systolic function the reduction of MAP and the increase in LV filling pressure are the main cardiovascular predictors of death.

At variance with previous results, the systolic function and SV being equal, in patients with a poor outcome we observed an increase in heart rate and a reduction of BSA and MAP, therefore inducing a consequent concomitant reduction of SW and systemic vascular resistance index and an increase of CI (all parameters which are related to heart rate, BSA, and MAP). Also, Ruiz-del-Arbol reported a reduction (not significant) of MAP and SW in 14 patients with cirrhosis with moderate ascites who presented type 1 HRS during 12 months of follow-up.12

The prevalence of LV DD showed wide variation depending on the algorithm considered, being lower using the ASE/EACVI 2016 recommendations and higher using the newly proposed Thorax Center algorithm. However, no association with prognosis emerged regarding LV DD when assessed with all of the above algorithms.

Regarding the development of type 1 HRS, we identified increased MELD (P = 0.006) and CI (P = 0.01) as independent predictors of development of this complication. This is in contrast with the results of Krag et al.,10 who reported a higher incidence of type 1 HRS in patients with low CI (below 1.5 L/minute/m²), and with the results of Ruiz-del-Arbol et al.,12 who observed an association between plasma renin activity and E/e′ ratio, with the development of this complication in 14 patients with cirrhosis and moderate ascites who presented type 1 HRS during 12 months of follow-up.

In conclusion, among echocardiographic and hemodynamic parameters, an increased LA dimension and E/e′ (as indexes of increased LV filling pressure) and low MAP with related compensatory increased heart rate seem to be useful in identifying patients with cirrhosis at increased risk of dying.

From the clinical standpoint, these results suggest performing an accurate echocardiographic evaluation in all patients with cirrhosis, paying specific attention to the assessment of LA dimensions and E/e′ ratio, to better identify patients with a worse prognosis which requires a lower threshold to initiate workup for liver transplantation. Further research is needed to collect follow-up outcome data in these patients to support this contention.

It might be argued that we should have used LA volume, which might have a stronger prognostic value for cardiovascular events, instead of LA diameter. However, the latter is more readily measurable and was available in all our patients, including those in the older studies, and provides an accurate estimate of LA size in 94% of patients.24

Strengths of this study to be underscored include the long follow-up and the large sample size of phenotypically well-characterized patients with cirrhosis, the use of state-of-the art echo-Doppler techniques entailing TDI for LV assessment, recording of echocardiograms, as well as centralized readings by a single experienced cardiologist.

REFERENCES

1 Theocharidou E, Krag A, Bendtsen F, Moller S, Burroughs AK. Cardiac dysfunction in cirrhosis—does adrenal function play a role? A hypothesis. Liver Int 2012; 32: 1327-1332.
10.1111/j.1478-3231.2011.02751.x
PubMed Web of Science® Google Scholar
2 Moller S, Henriksen JH. Cirrhotic cardiomyopathy. J Hepatol 2010; 53: 179-190.
10.1016/j.jhep.2010.02.023
PubMed Web of Science® Google Scholar
3 Cesari M, Fasolato S, Rosi S, Angeli P. Cardiac dysfunction in patients with cirrhosis: is the systolic component its main feature? Eur J Gastroenterol Hepatol 2015; 27: 660-666.
10.1097/MEG.0000000000000340
PubMed Web of Science® Google Scholar
4 Cesari M, Letizia C, Angeli P, Sciomer S, Rosi S, Rossi GP. Cardiac remodeling in patients with primary and secondary aldosteronism: a tissue Doppler study. Circ Cardiovasc Imaging 2016; 9: e004815.
10.1161/CIRCIMAGING.116.004815
PubMed Web of Science® Google Scholar
5 Nagueh SF, Appleton CP, Gillebert TC, Marino PN, Oh JK, Smiseth OA, et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography. Eur J Echocardiogr 2009; 10: 165-193.
10.1093/ejechocard/jep007
PubMed Web of Science® Google Scholar
6 Sampaio F, Pimenta J, Bettencourt N, Fontes-Carvalho R, Silva AP, Valente J, et al. Systolic dysfunction and diastolic dysfunction do not influence medium-term prognosis in patients with cirrhosis. Eur J Intern Med 2014; 25: 241-246.
10.1016/j.ejim.2014.01.011
PubMed Web of Science® Google Scholar
7 Sampaio F, Pimenta J, Bettencourt N, Fontes-Carvalho R, Silva AP, Valente J, et al. Systolic and diastolic dysfunction in cirrhosis: a tissue-Doppler and speckle tracking echocardiography study. Liver Int 2013; 33: 1158-1165.
10.1111/liv.12187
CAS PubMed Web of Science® Google Scholar
8 Nazar A, Guevara M, Sitges M, Terra C, Sola E, Guigou C, et al. Left ventricular function assessed by echocardiography in cirrhosis: relationship to systemic hemodynamics and renal dysfunction. J Hepatol 2013; 58: 51-57.
10.1016/j.jhep.2012.08.027
PubMed Web of Science® Google Scholar
9 Wong F, Liu P, Lilly L, Bomzon A, Blendis L. Role of cardiac structural and functional abnormalities in the pathogenesis of hyperdynamic circulation and renal sodium retention in cirrhosis. Clin Sci (Lond) 1999; 97: 259-267.
10.1042/cs0970259
CAS PubMed Web of Science® Google Scholar
10 Krag A, Bendtsen F, Henriksen JH, Moller S. Low cardiac output predicts development of hepatorenal syndrome and survival in patients with cirrhosis and ascites. Gut 2010; 59: 105-110.
10.1136/gut.2009.180570
CAS PubMed Web of Science® Google Scholar
11 Merli M, Torromeo C, Giusto M, Iacovone G, Riggio O, Puddu PE. Survival at 2 years among liver cirrhotic patients is influenced by left atrial volume and left ventricular mass. Liver Int 2017; 37: 700-706.
10.1111/liv.13287
PubMed Web of Science® Google Scholar
12 Ruiz-del-Arbol L, Achecar L, Serradilla R, Rodriguez-Gandia MA, Rivero M, Garrido E, et al. Diastolic dysfunction is a predictor of poor outcomes in patients with cirrhosis, portal hypertension, and a normal creatinine. Hepatology 2013; 58: 1732-1741.
10.1002/hep.26509
CAS PubMed Web of Science® Google Scholar
13 Salerno F, Gerbes A, Gines P, Wong F, Arroyo V. Diagnosis, prevention and treatment of hepatorenal syndrome in cirrhosis. Gut 2007; 56: 1310-1318.
CAS PubMed Web of Science® Google Scholar
14 Lang RM, Bierig M, Devereux RB, Flachskampf FA, Foster E, Pellikka PA, et al. Recommendations for chamber quantification. Eur J Echocardiogr 2006; 7: 79-108.
10.1016/j.euje.2005.12.014
CAS PubMed Web of Science® Google Scholar
15 Aurigemma GP, Gaasch WH, McLaughlin M, McGinn R, Sweeney A, Meyer TE. Reduced left ventricular systolic pump performance and depressed myocardial contractile function in patients >65 years of age with normal ejection fraction and a high relative wall thickness. Am J Cardiol 1995; 76: 702-705.
10.1016/S0002-9149(99)80201-X
CAS PubMed Web of Science® Google Scholar
16 De Marco M, Chinali M, Romano C, Benincasa M, D'Addeo G, D'Agostino L, et al. Increased left ventricular mass in pre-liver transplantation cirrhotic patients. J Cardiovasc Med (Hagerstown) 2008; 9: 142-146.
10.2459/JCM.0b013e3280c7c29c
PubMed Web of Science® Google Scholar
17 de Simone G, Devereux RB, Koren MJ, Mensah GA, Casale PN, Laragh JH. Midwall left ventricular mechanics. An independent predictor of cardiovascular risk in arterial hypertension. Circulation 1996; 93: 259-265.
10.1161/01.CIR.93.2.259
PubMed Web of Science® Google Scholar
18 de Simone G, Devereux RB, Maggioni AP, Gorini M, de Divitiis O, Verdecchia P; MAVI (MAssa Ventricolare sinistra nell'Ipertensione) Study Group. Different normalizations for body size and population attributable risk of left ventricular hypertrophy: the MAVI study. Am J Hypertens 2005; 18: 1288-1293.
10.1016/j.amjhyper.2005.05.027
PubMed Web of Science® Google Scholar
19 Torregrosa M, Aguade S, Dos L, Segura R, Gonzalez A, Evangelista A, et al. Cardiac alterations in cirrhosis: reversibility after liver transplantation. J Hepatol 2005; 42: 68-74.
10.1016/j.jhep.2004.09.008
PubMed Web of Science® Google Scholar
20 de Simone G, Devereux RB, Kimball TR, Mureddu GF, Roman MJ, Contaldo F, et al. Interaction between body size and cardiac workload: influence on left ventricular mass during body growth and adulthood. Hypertension 1998; 31: 1077-1082.
10.1161/01.HYP.31.5.1077
PubMed Web of Science® Google Scholar
21 Nagueh SF, Smiseth OA, Appleton CP, Byrd BF 3rd, Dokainish H, Edvardsen T, et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography: an update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J Am Soc Echocardiogr 2016; 29: 277-314.
10.1016/j.echo.2016.01.011
PubMed Web of Science® Google Scholar
22 Palmieri V, Capaldo B, Russo C, Iaccarino M, Di Minno G, Riccardi G, et al. Left ventricular chamber and myocardial systolic function reserve in patients with type 1 diabetes mellitus: insight from traditional and Doppler tissue imaging echocardiography. J Am Soc Echocardiogr 2006; 19: 848-856.
10.1016/j.echo.2006.02.011
PubMed Web of Science® Google Scholar
23 Takemoto Y, Pellikka PA, Wang J, Modesto KM, Cauduro S, Belohlavek M, et al. Analysis of the interaction between segmental relaxation patterns and global diastolic function by strain echocardiography. J Am Soc Echocardiogr 2005; 18: 901-906.
10.1016/j.echo.2005.05.008
PubMed Web of Science® Google Scholar
24 van Dalen BM, Strachinaru M, van der Swaluw J, Geleijnse ML. A simple, fast and reproducible echocardiographic approach to grade left ventricular diastolic function. Int J Cardiovasc Imaging 2016; 32: 743-752.
10.1007/s10554-015-0832-6
PubMed Web of Science® Google Scholar
25 Andersen OS, Smiseth OA, Dokainish H, Abudiab MM, Schutt RC, Kumar A, et al. Estimating left ventricular filling pressure by echocardiography. J Am Coll Cardiol 2017; 69: 1937-1948.
10.1016/j.jacc.2017.01.058
PubMed Web of Science® Google Scholar

Author names in bold designate shared co-first authorship.

Citing Literature

Volume68, Issue1

July 2018

Pages 215-223

Cardiovascular predictors of death in patients with cirrhosis