Cardiac dysfunction as an early predictor of portal hypertension in chronic hepatitis C

Introduction

Cirrhotic cardiomyopathy is a type of chronic cardiac dysfunction observed in patients with cirrhosis; it is characterized by blunted contractile responsiveness to stress and/or altered diastolic relaxation with electrophysiological abnormalities, in the absence of any other known cause of cardiac disease [1]. Systolic dysfunction is the result of an impaired sensitivity to sympathetic activation, leading to a blunted increase in cardiac output and decreased contractility with exercise, pharmacological stress, and volume challenge [2]. Diastolic dysfunction in cirrhotic patients has been associated with increased left ventricular (LV) wall thickness, subendocardial edema, fibrosis and altered collagen structure, ultimately leading to altered relaxation [3]. Finally, electrophysiological abnormalities have been observed, including QTc prolongation and impaired electromechanical coupling [4].

With regard to the pathogenesis, it has been postulated that the marked splanchnic arterial vasodilatation observed in liver cirrhosis is initially compensated for by a hyperdynamic circulation; later, the progression of liver disease and portal hypertension leads to further splanchnic vasodilatation and reduction of the effective arterial blood volume, leading to activation of the renin-angiotensin-aldosterone system and the sympathetic nervous system. The overexposure to sympathetic effector molecules leads to myocardial injury, receptor internalization and downregulation, resulting in an impaired response of β-adrenergic receptors [5-9]. The increased production of endogenous cardiodepressant mediators, such as nitric oxide, carbon monoxide, endogenous cannabinoids and inflammatory cytokines, has also been proposed as a possible causative mechanism [10-12].

Cardiovascular changes in cirrhosis have prognostic implications [13-15]. Therefore, the evaluation of subclinical cardiac involvement could contribute to the early detection of those cirrhotic patients at higher risk of decompensation and development of complications, allowing better tailoring of follow up.

The present study aimed to evaluate the relationship between cirrhotic cardiomyopathy investigated by standard echocardiography, and the severity of liver fibrosis evaluated by transient elastography, in a group of patients affected by hepatitis C virus (HCV)-related chronic liver disease at different stages of fibrosis. In addition, we aimed to investigate the correlation between the echocardiographic findings and endoscopic complications of portal hypertension.

Patients and methods

A consecutive series of patients with HCV-related chronic liver disease was studied. HCV infection was established by positive serology for HCV and quantifiable HCV RNA over the 6 months prior to recruitment. We excluded patients suffering from liver disease of viral etiology other than HCV, or from decompensated cirrhosis complicated by ascites, hepatopulmonary or hepatorenal syndrome, or portal-systemic encephalopathy. Patients with a history of cardiovascular disease, including ischemic cardiomyopathy, atrial fibrillation, severe valvular defects, severe arterial hypertension or previous hospitalizations for heart failure, were also excluded.

A total of 70 patients and 20 healthy controls were enrolled. All patients underwent a thorough clinical and laboratory evaluation. Patients’ data were recorded according to local ethical guidelines.

Results

Patients’ characteristics

We report the clinical characteristics of the study population in Table 1. Patients were classified, as explained above, according to LS measured by liver transient elastography; controls and patients were comparable for all the clinical variables considered, except for the prevalence of Type 2 diabetes mellitus, which was higher in patients with advanced liver disease. Consequently, a higher proportion of anti-diabetic therapy was reported in this subgroup of patients. In addition, the patients were more frequently taking antihypertensive drugs, especially β-blockers.

Table 1 Characteristics of the population

Echocardiographic findings

As shown in Table 2, patients with LS>12.5 kPa had a significantly larger left atrium than patients with LS<12.5 kPa and controls, whether measured as anterior-posterior (AP) diameter (P=0.03) or as area and volume index (P=0.007 and P=0.003, respectively). Furthermore, LS>12.5 kPa was the only independent predictor of LA dilatation (Table 3). LV diameters, thicknesses and volumes did not differ significantly between groups, although a tendency towards larger cardiac mass index, LV volume and wall thickness was observed in the group of patients with advanced fibrosis. All groups showed an LV ejection fraction within the normal range and mild diastolic dysfunction, according to the ASE Guidelines, while the peak systolic velocities recorded with TDI (S¢) were significantly lower in patients with LS suggestive of liver cirrhosis (Table 2).

Table 2 Echocardiographic parameters

Table 3 Predictors of left atrial enlargement

EGD findings

Forty-nine patients had undergone endoscopy within two years prior to recruitment; of these, 25 patients showed abnormal endoscopic findings (esophageal varices F1 to F4 and/or congestive gastropathy). As shown in Table 4, LS was significantly higher in patients with endoscopic signs of portal hypertension, while the peak systolic velocity, recorded as the TDI S¢ wave at the mitral annulus, was significantly lower. Logistic regression confirmed that the S¢ wave measured by TDI was the only independent predictor of the presence of endoscopic signs of portal hypertension (Table 5). We further considered a cutoff value for LS of 19 kPa, previously described as highly predictive for the diagnosis of large esophageal varices [18], according to which we built a further model. Again, LS was not predictive for portal hypertension in logistic regression analysis (P=0.13). On the other hand, S¢-wave velocity (P=0.02, odds ratio [OR] 0.61, 95%CI 0.40-0.95), LV mass (P=0.04, OR 1.05 95%CI 1.0-1.1) and LA volume (P=0.04, OR 0.90, 95%CI 0.82-0.99) were significantly associated with endoscopic findings of portal hypertension.

Table 4 Characteristics of the subpopulation of liver disease based on the presence of endoscopic signs of portal hypertension

Table 5 Predictors of portal hypertension

Discussion

The present study indicates that echocardiographic parameters (LA volume, LV mass index and TDI S¢-wave velocity) may merit consideration as predictors of the presence of esophageal varices. These data need to be discussed in the light of the existing literature. We confirm that patients with compensated cirrhosis are characterized by a significantly larger left atrium, whether measured by AP diameter or by volume index; moreover, LS is the only independent predictor of LA enlargement. LA enlargement in cirrhotic patients has already been reported by Merli et al [19], in association with an increase in LV diameters, but not volumes. On the other hand, we did not find any difference in ventricular volumes or thicknesses, but the latter observation could have been partially biased by the exclusion from our cohort of alcoholic cirrhosis, which can be associated with altered LV diameters [20].

We failed to correlate the severity of liver fibrosis with diastolic function assessed by Doppler transmitral flow analysis, through parameters such as the E/A ratio, deceleration time and E/E¢. These parameters are notoriously affected by different factors, such as the overload state which is typical of cirrhosis, therefore mirroring a functional diastolic defect rather than a primary cardiac impairment [21]. On the other hand, the LA enlargement reflects the effects of increased filling pressures over time, independently of hemodynamic alterations, being a better indicator of chronic diastolic dysfunction. LA enlargement might be considered a diagnostic criterion of diastolic dysfunction in the course of cirrhotic cardiomyopathy; its accuracy as predictor of heart failure, ascites and hepatorenal syndrome in cirrhotic patients should be assessed by prospective studies.

When systolic function was assessed by conventional echocardiography, no significant differences in LV ejection fraction were observed between patients with chronic liver disease and controls. More complex techniques, such as TDI, were able to identify a reduction in the longitudinal peak systolic velocity (S¢) and to detect signs of early damage to LV longitudinal function in patients with chronic liver disease, even in earlier stages of fibrosis. Our results are similar to those obtained by Kazankov et al [22], who applied the ASE Guidelines as we did.

Concerning electrophysiological abnormalities, though we found a longer QTc in HCV-related liver diseases, no correlation was detected with the degree of liver fibrosis. The relationship between QTc prolongation and the severity of liver disease is still debated and it would be useful to rule out the possible impact of β-blockers in prospective studies [23].

Finally, we evaluated the association between echocardiographic and EGD findings. Nowadays, EGD is the screening method for all cirrhotic patients, since no other noninvasive techniques have shown a satisfactory predictive value for signs of portal hypertension, even though numerous studies have been conducted [24]. According to our data, the cardiovascular changes characteristic of cirrhotic cardiomyopathy appear to be closely related to the presence of esophageal varices. A reduction of the peak S¢-wave velocity on TDI, an increase in LV mass and LA dilatation could be early signs of portal hypertension, and thus may be useful for the early detection of patients with a higher risk of portal hypertension and associated complications. Furthermore, echocardiography seems to perform better as a predictor of endoscopic findings of portal hypertension compared to liver elastography. In fact, even though LS was significantly higher in patients with abnormal endoscopy, as already reported by other groups [18], when considered as a predictor in a more complex model including echocardiographic findings, it lost its significance. This is particularly relevant, suggesting that the S¢ wave could be considered a better and earlier predictor of portal hypertension than LS. It is possible that a more complex evaluation of heart function, including strain analysis, might provide further clues for the early detection of subclinical myocardial dysfunction in cirrhotic patients.

A major advantage of our study is the high homogeneity of the population; the restriction to HCV-related liver disease significantly reduced the impact of potential confounding factors, such liver disease etiology, which affected previous reports. Limitations of the present study include the relatively small sample size and the absence of a stress test (physical activity or pharmacological stress), which could have better disclosed subclinical systolic dysfunction. Finally, we still do not have prospective data on the follow up of patients affected by portal hypertension, which would help us evaluate the potential role of echocardiographic findings as prognostic markers.

In conclusion, our results show that LA enlargement and peak systolic velocity measured by TDI are suitable instruments for detecting diastolic and systolic dysfunction in chronic liver disease. In the absence of reliable methods to identify cirrhotic patients who should undergo endoscopy for screening of esophageal varices, echocardiographic parameters might be worth considering.