Aliment Pharmacol Ther 2010; 32: 105–112
Background Cirrhosis with arterial hypertension is not uncommon. Haemodynamic alterations in these patients and the effects of beta-blocker on hepatic venous pressure gradient (HVPG) and systemic haemodynamics have not been evaluated.
Aims To compare the systemic haemodynamic alterations in hypertensive and normotensive cirrhotics, and to investigate the effects of propranolol on these parameters.
Methods A retrospective analysis of consecutive hypertensive cirrhotic patients (n = 33) who underwent haemodynamic assessment and paired HVPG measurement was done. Normotensive cirrhotics (n = 50) served as controls.
Results Hypertensive patients had a significantly higher heart rate, systemic (SVRI), and pulmonary vascular resistance. There was a significant reduction in mean arterial pressure (MAP) in the hypertensive cirrhotic group from 112 (107–130) mmHg to 95 (77–114) mmHg (P < 0.01), but no change in the normotensives. SVRI remained the same in the hypertensive cirrhotic group, but it increased in the normotensives. There was no correlation between MAP reduction and HVPG reduction.
Conclusions The frequency of HVPG response with propranolol treatment in hypertensive cirrhotics is similar to normotensive cirrhotics. Propranolol treatment reduces MAP significantly in hypertensive patients with cirrhosis. Treatment with a nonselective beta-blocker is a good strategy for hypertensive cirrhotic patients.
It had been known for a long time that patients with cirrhosis seldom exhibit arterial hypertension.1–6 It had also been reported that hypertension disappears after the development of cirrhosis in humans3, 6 and experimental animals.7 However, in one of the large studies on this aspect,8 high blood pressure levels remained unaffected by the onset of cirrhosis. Hence, the coexistence of liver cirrhosis and essential hypertension is not rare in clinical practice, as both diseases have a high prevalence in mid and late life. Arterial hypertension is found in about 10% of patients with cirrhosis.5, 9
Patients with cirrhosis and portal hypertension eventually develop a hyperdynamic circulation, with high cardiac output and reduced systemic vascular resistance (SVR). This haemodynamic alteration is thought to be involved in the pathogenesis of sodium retention and ascites,10 which has been observed in 80% of cirrhotic patients after many years of disease.11 Whether the hyperdynamic circulation also occurs in cirrhotic patients with arterial hypertension remains unknown.
Nonselective beta-blockers (i.e. propranolol, nadolol, timolol) are the recommended first-line therapy for the prophylaxis of variceal haemorrhage in cirrhotic patients with varices at high risk for bleeding.12 They reduce portal pressure by decreasing portal and collateral blood flow. This effect is achieved via the blockade of both the beta-1-adrenoreceptors, causing a reduction in cardiac output and the beta-2-adrenoreceptors in the splanchnic vasculature, causing splanchnic vasoconstriction.13, 14 Near complete protection from variceal bleeding is reported when the hepatic venous pressure gradient (HVPG) is reduced to <12 mmHg or ≥20% reduction from the baseline values is achieved.15–17
The effects of beta-blocker on HVPG and systemic haemodynamics in patients with cirrhosis and arterial hypertension have not been evaluated. We hypothesize that hypertensive cirrhotics may behave differently from normotensive cirrhotics, which may be the consequence of a balance between different effects on the two determinants of portal pressure, which are influenced by beta-blockers, i.e. portal blood inflow and resistance to portal blood flow.
The aim of the study was to compare the systemic haemodynamic alterations in hypertensive and normotensive cirrhotics and to investigate the effects of propranolol on HVPG and systemic haemodynamics in these patients. The relevance of this study is to assess the haemodynamic alterations present in cirrhotics and to determine which of these alterations are corrected by treatment with nonselective beta-blockers.
Patients and methods
Retrospective analysis of records of consecutive cirrhotic patients, who underwent haemodynamic assessment and paired HVPG measurement in our department, from 2002 through 2008, was performed. From this group, consecutive patients with arterial hypertension, defined according to the JNC VII classification (diastolic blood pressure >90 mmHg and systolic blood pressure >140 mmHg),18 were selected. Only those patients whose records showed elevated blood pressure on at least two occasions were selected prior to the haemodynamic studies. The exclusion criteria were: current treatment with anti-hypertensive or portal pressure reducing drugs, Child-Pugh score ≥13, presence of any neoplastic disease, inability to attend follow-up, contraindications to beta-blockers (atrioventricular block, sinus bradycardia with heart rate <50 beats per minute, arterial hypotension with systolic blood pressure <85 mmHg, heart failure, asthma, peripheral arterial disease or diabetes mellitus needing insulin treatment), concomitant treatment for hepatitis B or C and HVPG <12 mmHg. The remaining patients were included in the study and comprised the hypertensive cirrhotic group. Normotensive age- and gender-matched patients of cirrhosis, who had underwent haemodynamic assessment and paired HVPG during the same period, were chosen randomly by computer for comparison (normotensive cirrhotic group).
The records were analysed for baseline characteristics of included patients, comprising age, gender, aetiology of cirrhosis, Child-Pugh score and serum creatinine. Haemodynamic data analysed pre- and post-therapy were heart rate, mean arterial pressure, HVPG, cardiac index, SVRI and PVRI.
Measurement of HVPG
Hepatic venous pressure gradient was measured before administering beta-blockers (baseline) and after at least 6 weeks, but not more than 52 weeks on beta-blocker treatment (on therapy). It was performed after an overnight fast and under antibiotic cover. Using local anaesthesia, a 7F central venous catheter (Arrow Medical, Athens, TX, USA) was placed in the right femoral vein or internal jugular vein under fluoroscopic guidance, using the Seldinger technique. HVPG was measured by the standard technique19 in which a balloon catheter was introduced into the right hepatic vein under fluoroscopic guidance and connected to a transducer. The zero reference point of pressure was set at the mid-axillary point. The free hepatic venous pressure (FHVP) was obtained by keeping the catheter free in the lumen of the hepatic vein. The balloon of the catheter was then inflated to wedge the lumen of hepatic vein. The pressure tracing at this juncture showed absence of wave forms and the pressure was labelled as wedged hepatic venous pressure (WHVP). Presence of wedging was further confirmed by injection of 2 mL of intravenous contrast which showed absence of reflux of contrast into the inferior vena cava and appearance of a sinusoidogram. HVPG was determined by subtracting free hepatic venous pressure from wedged hepatic venous pressure (HVPG = WHVP−FHVP). All measurements were performed in triplicate and mean was taken. If the difference between any two HVPG readings was more than 1 mmHg, all the three readings were discarded and fresh set of triplicate readings were taken. The normal value of the HVPG in our haemodynamic laboratory is between 1 and 4 mmHg.
Measurement of cardiac output and systemic vascular resistance and pulmonary vascular resistance
After measuring HVPG, the balloon catheter was then advanced into the right atrium, pulmonary artery and then pulmonary capillaries for measurement of the right atrial pressure (RAP), pulmonary arterial pressure (PAP) and pulmonary capillary wedge pressure (PCWP) respectively. Mean arterial pressure (MAP) was measured simultaneously by continuous non-invasive blood pressure monitoring. Blood samples were obtained from the pulmonary artery and the femoral artery for estimating oxygen saturation. Heart rate was derived from continuous electrocardiogram monitoring. Cardiac output was calculated by Fick’s oxygen method20 as follows: oxygen consumption (mL/min) ÷ Arterio-venous oxygen difference (mL/L). Cardiac output was then indexed by dividing by body surface area and expressed as cardiac index (CI; L/min/m2). The systemic vascular resistance (SVRI, in dynes s/m2/cm5) was calculated as: (MAP in mmHg – RAP in mmHg) × 80 ? CI (in L/min/m2). The pulmonary vascular resistance index (PVRI, in dynes s/m2/cm5) was calculated as: (PAP in mmHg – PCWP in mmHg) × 80 ? CI (in L/min/m2).
The standard protocol of incremental dosing of beta-blockers was used to achieve the target heart rate. Propranolol was started at a dose of 20 mg twice daily. The dose of propranolol was increased on alternate day to achieve target heart rate of 55/min or to the maximal dose to 320 mg/day if the medication was well tolerated and the systolic blood pressure remained ≥90 mmHg.
On the occurrence of intolerable side effects, systolic blood pressure <90 mmHg or pulse rate <55 mmHg, the dose of the medication was decreased step wise and eventually stopped, if these adverse events persisted. Re-introduction of the medication was attempted if cessation of the medication did not result in improvement of the reported side-effect.
Compliance was assessed by interview with the patients as well as corroboration of the patient’s history by a member of the family, if present, during consultation. Noncompliance was defined as at least one episode of stopping the medication for more than 3 days.
On-therapy HVPG was performed 6 weeks after the full tolerated dose of propranolol was achieved, to assess the haemodynamic response to propranolol.
Hepatic venous pressure gradient response, defined as a reduction in HVPG to <12 mmHg or ≥20% reduction from the baseline values, was considered as the primary endpoint.
Effect of propranolol on heart rate, mean arterial pressure, cardiac index, systemic and pulmonary vascular resistance indices were defined as the secondary endpoints.
Results were expressed as median (range) for continuous data and number (%) for categorical data. Mann–Whitney U-test, Wilcoxon signed-rank test and Fisher’s exact test were used for comparison of variables. Correlation between mean arterial pressure reduction and HVPG reduction was performed by Spearman’s rank correlation. Statistical significance was established at a P value <0.05. Statistical analysis was performed using the spss 15.0 statistical package (SPSS Inc., Chicago, IL, USA).
From 2002 through 2008, one thousand one hundred and twenty-seven patients with cirrhosis underwent HVPG in our department. Of these patients, nine hundred and fifty-eight patients underwent complete haemodynamic assessment (including SVR, PVR and cardiac index) while the remaining 169 patients did not undergo full assessment and were excluded. The complete haemodynamic assessment was performed as part of various clinical research protocols running in the department, notably, early-primary, primary and secondary prophylaxes of variceal bleeding. Of these, seventy-eight (8%) patients were hypertensive as per the JNC VII criteria. Forty-five patients were excluded from the study for the following reasons: Taking antihypertensive or portal pressure reducing drugs – 39, bronchial asthma – 3, and diabetes mellitus – 3. Hence, thirty-three patients were included in the hypertensive cirrhotic group. Fifty normotensive age- and gender-matched patients of cirrhosis were included in the normotensive cirrhotic group.
The baseline characteristics of included patients are given in Table 1. There was no difference in age, gender, aetiology of cirrhosis, Child-Pugh score and serum creatinine values between the two groups. Hypertensive patients had significantly higher heart rate, SVRI and PVRI, while cardiac indices in the two groups were comparable (Table 1).
|Parameter||Hypertensive cirrhotics (n = 33)||Normotensive cirrhotics (n = 50)||P value|
|Age, years||52 (33, 70)||50 (36, 72)||0.473|
|Male||28 (85%)||40 (80%)||0.772|
|Female||5 (15%)||10 (20%)|
|Alcohol||12 (36%)||16 (32%)||0.912|
|Viral||12 (36%)||20 (40%)|
|Others||9 (28%)||14 (28%)|
|Child-Pugh score||8 (5, 12)||8 (5, 12)||0.800|
|A||14 (43%)||13 (26%)||0.160|
|B||11 (33%)||27 (54%)|
|C||8 (24%)||10 (20%)|
|Serum creatinine, mg/dL||1.0 (0.4, 1.4)||1.0 (0.4, 1.4)||0.539|
|Large||20 (61%)||26 (52%)||0.503|
|Small||13 (39%)||24 (48%)|
|Previous variceal bleeding||10 (30%)||8 (16%)||0.174|
|Resting HR, beats/min||88 (70, 112)||84 (70, 112)||0.050|
|MAP, mmHg||112 (107, 130)||90 (71, 98)||<0.01|
|HVPG, mmHg||17 (12, 29)||16 (12, 29)||0.926|
|Cardiac index, L/min/m2||4.7 (2.5, 7.1)||4.9 (2.4, 8.7)||0.382|
|SVRI, dyn s/m2/cm5||1745 (1156, 3276)||1417 (665, 2521)||<0.01|
|PVRI, dyn s/m2/cm5||126 (84, 353)||84 (34, 173)||<0.01|
Propranolol was started after baseline HVPG measurement. The median propranolol dose used was similar in both the groups: 160 (120–320) mg in the hypertensive cirrhotic group vs. 160 mg (80–320) mg in the normotensive cirrhotic group; P = 0.996. The second HVPG was performed in median 18 (6–52) weeks from the first HVPG to determine the haemodynamic response to propranolol. The duration between first and second HVPG was similar in the two groups [hypertensive cirrhotic group 19 (8–50) weeks vs. normotensive cirrhotic group 18 (6–52) weeks; P = 0.911).
Primary endpoint: HVPG response
The frequency of HVPG response was similar in both the groups: Of 33 patients in hypertensive cirrhotic group, 11 (33%) achieved primary end point, while of normotensive cirrhotics 18/50 (36%) achieved primary endpoint (P = 1.000) (Figure 1). Overall, the HVPG decreased in 51/83 (61%) patients, remained the same in 5/83 (6%) patients and increased in 27/83 (33%) patients. Overall, there was a median reduction in HVPG of 9% (hypertensive cirrhotic 10% vs. normotensive cirrhotic 8.6%; P = 0.675). In hypertensive cirrhotic group, the median post-therapy HVPG in responders was 13 (8–21) mmHg, while in nonresponders it was 18 (13–22) mmHg. In normotensive cirrhotic group, the median post-therapy HVPG in responders was 12 (7–22) mmHg, while in nonresponders, it was 18 (13–30) mmHg (Figure 1).
Secondary endpoints: systemic haemodynamic response
The median resting heart rate achieved was 58 (range 55–65) beats/min. There was no difference in the median resting heart rate achieved in the two groups (hypertensive cirrhotic group 58 [55–64] beats/min vs. normotensive cirrhotic group 58 [55–65] beats/min; P = 0.932) (Table 2).
|Hypertensive cirrhotics (n = 33)||Normotensive cirrhotics (n = 50)|
|Baseline||On-therapy||P value||Baseline||On-therapy||P value|
|Resting HR, beats/min||88 (70, 112)||58 (55, 64)||<0.01||84 (70, 112)||58 (55, 65)||<0.01|
|MAP, mmHg||112 (107, 130)||95 (77, 114)||<0.01||90 (71, 98)||88 (64, 105)||0.202|
|Cardiac index, L/min/m2||4.7 (2.5, 7.1)||3.7 (2.1, 5.2)||<0.01||4.9 (2.4, 8.7)||4.2 (2.4, 9.0)||<0.01|
|SVRI, dyn s/m2/cm5||1745 (1156, 3276)||1767 (1098, 3577)||0.728||1417 (665, 2521)||1587 (758, 2531)||0.032|
|PVRI, dyn s/m2/cm5||126 (84, 353)||118 (64, 297)||0.085||84 (34, 173)||75 (38, 195)||0.904|
There was a significant reduction in mean arterial pressure in the hypertensive cirrhotic group from 112 (107–130) mmHg to 95 (77–114) mmHg (P < 0.01). However, there was no reduction in mean arterial pressure in the normotensive cirrhotic group [baseline 90 (71–98) mmHg vs. post-therapy 88 (64–105) mmHg; P = 0.202] (Table 2).
Cardiac index decreased significantly in both the groups. SVRI, which was significantly lower in the normotensive cirrhotic group in baseline, increased significantly with therapy, while it remained the same in the hypertensive cirrhotic group. There was no significant change in PVRI in both groups (Table 2).
Correlation of mean arterial pressure reduction with HVPG reduction
In the hypertensive cirrhotic group, the median reduction in mean arterial pressure was 15% (4–32%), while the median reduction in HVPG was 10% (−50–41%). However, there was no correlation between mean arterial pressure reduction and HVPG reduction [correlation coefficient (Spearman’s rho) 0.209; P = 0.242] (Figure 2).
In the normotensive cirrhotic group, as there was no reduction in the mean arterial pressure, correlation with HVPG reduction could not be performed.
The results of our study clearly show that the frequency of HVPG response with propranolol treatment in hypertensive cirrhotics is similar to normotensive cirrhotics. Moreover, with propranolol treatment, there was a significant reduction in mean arterial pressure in the hypertensive cirrhotic group; however, there was no reduction in mean arterial pressure in the normotensive cirrhotic group. Cardiac index decreased significantly in both groups. SVRI, which was significantly lower in the normotensive cirrhotic group in baseline, increased significantly with therapy, while it remained the same in the hypertensive cirrhotic group. The HVPG reduction was an independent effect and we did not find any correlation between mean arterial pressure reduction and HVPG reduction in hypertensive cirrhotics. Thus, nonselective beta-blocker is a good treatment option for hypertensive cirrhotic patients as it is as effective as in normotensives in reducing the HVPG; it also corrects various haemodynamic abnormalities in hypertensive cirrhotics.
In our group of cirrhotic patients who were eligible for haemodynamic studies for beta-blocker therapy, 8% were found to be hypertensive. In patients with liver disease, the blood pressure is most often low normal or decreased.21, 22 Studies on arterial hypertension in patients with cirrhosis show that the frequency is reduced, with a prevalence of raised arterial blood pressure of 3–7%.3–6, 9 Our study reported a frequency of 8%. This apparently high frequency may be due to referral bias. Our hospital is a tertiary care institute and patients of cirrhosis with co-existing hypertension are more often likely to be referred here than patients with cirrhosis alone.
On comparing the baseline circulatory parameters in the hypertensive cirrhotics with normotensive cirrhotics, hypertensive patients had significantly higher heart rate. It has already been shown that patients with hypertension generally have a higher heart rate.23 Moreover, a higher heart rate predicts the risk of developing hypertension in a normotensive screened cohort.24, 25 Cirrhosis is also associated with increased heart rate and blood pressure and heart rate obtained during the 24-h monitoring were positively and closely correlated.26 Combination of cirrhosis and hypertension may have an additive effect on heart rate.
Among the baseline haemodynamic parameters hypertensive cirrhotic patients had significantly higher SVRI and PVRI, while cardiac indices in the two groups were comparable. SVRI is decreased in most patients with cirrhosis. The pathogenesis of decreased SVRI in cirrhosis is multifactorial.27 There is presence of a surplus of circulating vasodilators. Combined with a decreased sensitivity to pressor substances, this leads to vasodilatation and reduced vascular resistance.21, 27 The coexistence of decreased splanchnic resistance due to cirrhosis and increased peripheral resistance due to arterial hypertension is likely and may result in increased, normal, or decreased systemic vascular resistance. However, most evidence indicates that a decreased overall systemic vascular resistance with a low arterial blood pressure is most often the outcome, even in patients with primary essential hypertension.3, 4, 6 In our study too, the SVRI in hypertensive cirrhotics was below the normal range of 1970–2390 dyn s/m2/cm5.
One of the important findings in our study is that the frequency of HVPG response with propranolol treatment in hypertensive cirrhotics is similar to normotensive cirrhotics. Nonselective beta-blockers are the recommended first-line therapy for the prophylaxis of variceal haemorrhage in cirrhotic patients with varices at high risk for bleeding.12 Near complete protection from variceal bleeding is reported when the hepatic venous pressure gradient (HVPG) is reduced to <12 mmHg or ≥20% reduction from the baseline values is achieved.15–17 Previous studies had shown this response to vary from 33 to 40% with beta-blockers and consistent with previous studies, in our study, 36% normotensive cirrhotics responded to propranolol. However, no previous study had evaluated the haemodynamic response in hypertensive cirrhotic patients. We found a similar response rate in this group of patients too.
The other important finding of our study was that with beta-blockers, there was a significant reduction in mean arterial pressure in the hypertensive cirrhotic group; however, there was no reduction in mean arterial pressure in the normotensive cirrhotic group. Risk of development of hypotension has been an important concern while using beta-blockers and in many series a significant proportion of patients are unable to tolerate beta-blockers because of hypotension. However, in our study, it was interesting to note that in spite of median 160 (80–320) mg of propranolol, there was no reduction in mean arterial pressure. In a few previous studies also, it was noted that beta-blockers did not have a significant systemic hypotensive effect.28–30 The reason for this variation in effect on mean arterial pressure in various studies remains unknown. One important limitation of our study is that our study is not able to show the effect of propranolol in a patient who is receiving an antihypertensive treatment different from non-selective beta-blockers, as we included patients not on any anti-hypertensives.
In the present study, as expected, the cardiac index decreased significantly in both groups with beta-blocker. SVRI, which was significantly lower in the normotensive cirrhotic group at baseline, increased significantly with therapy, while it remained the same in the hypertensive cirrhotic group. There was no significant change in PVRI in both groups. Beta-adrenergic blockade causes a significant reduction in cardiac output and a significant increase in systemic vascular resistance in normotensive cirrhotics.31 However, in our group of hypertensive cirrhotics, it failed to increase the SVRI further.
In our study, there was no correlation between mean arterial pressure reduction and HVPG reduction in the hypertensive cirrhotic group. Thus, the HVPG reduction was an independent effect from MAP reduction. This may be due to the fact that reduction in arterial hypertension is mainly dependent on beta-1-blockade, while reduction in HVPG depends on effect of both beta-1- and beta-2-blockade.
In conclusion, the results of the present study demonstrate that the frequency of HVPG response with propranolol treatment in hypertensive cirrhotics is similar to normotensive cirrhotics. Moreover, with propranolol treatment, there was a significant reduction in mean arterial pressure in the hypertensive cirrhotic group; however, there was no reduction in mean arterial pressure in the normotensive cirrhotic group. The HVPG reduction is an independent effect and there was no correlation between MAP reduction and HVPG reduction in hypertensive cirrhotics. Thus, non-selective beta-blocker is a good treatment option for hypertensive cirrhotic patients. Prospective studies are needed to confirm these findings further.
Declaration of personal and funding interests: None.