Michihiro Sumino

Echo‐Doppler measurements of portal vein and superior mesenteric artery blood flow in humans: Inter‐ and intra‐observer short‐term reproducibility

The reproducibility of echo‐Doppler measurements of portal vein and superior mesenteric artery blood flow has not been extensively studied. In the present study, two groups of subjects were examined to test inter‐ and intra‐observer reproducibility. Each study population consisted of 15 non‐portal hypertensive and 15 portal hypertensive subjects. With a standardized technique, the cross‐sectional area and velocity of blood flow in the portal vein and superior mesenteric artery were recorded in triplicate by skilled operators. The flow volume of each vessel was calculated by multiplying the cross‐sectional area by the velocity of blood flow. The measurements were performed in a blind fashion over a 60 min period. The reproducibility of measurements was assessed by calculation of intraclass correlation coefficients and coefficients of variation. The intra‐observer intraclass correlation coefficient was 0.77 for portal vein blood flow and 0.84 for superior mesenteric artery blood flow, suggesting good reproducibility. The intra‐observer coefficient of variation was 11 and 9%, respectively. In contrast, the interobserver intraclass correlation coefficient was calculated to be 0.49 for portal blood vein blood flow and 0.57 for superior mesenteric artery blood flow, indicating fair reproducibility. In addition, the interobserver coefficients of variation were calculted to be 20 and 18%, respectively. These data suggest that intra‐observer reproducibility in echo‐Doppler measurements of portal vein and superior mesenteric artery blood flow is acceptable but inter‐observer reproducibility is not. Examination by a single operator, rather than multiple operators, is therefore advisable. Even when measurements are performed by a single investigator an approximate variance of 10% in the measurement in a single subject should be expected.

Gastric mucosal blood flow after smoking in healthy human beings assessed by laser Doppler flowmetry

We measured regional gastric mucosal blood flow by laser Doppler flowmetry before and after control (n = 8) or cigarette smoking (n = 8) in healthy human beings. The control group showed no change in both antrum (from 1.15 +/- 0.32 to 1.20 +/- 0.39 V, NS) and corpus gastric mucosal blood flow (from 1.15 +/- 0.32 to 1.12 +/- 0.28 V, NS). In contrast, cigarette smoking caused a significant reduction in gastric mucosal blood flow in the antrum (from 1.08 +/- 0.31 to 0.71 +/- 0.22 V, p < 0.01) and in the corpus (from 0.99 +/- 0.26 to 0.66 +/- 0.24 V, p < 0.01). The magnitude of reduction in gastric mucosal blood flow was similar between the antrum and the corpus (-34% +/- 11% versus -33% +/- 15%, NS). We conclude that cigarette smoking induces a significant reduction in gastric mucosal blood flow and that no heterogeneous response occurs in regional gastric mucosa. In addition, the laser Doppler flowmeter appears to be a sensitive method to assess rapid change in gastric mucosal blood flow in human beings.

Portal pressure after prophylactic sclerotherapy in patients with high-risk varices

Portal hemodynamics and transhepatic portal venographic findings were studied before and after prophylactic sclerotherapy (mean duration = 40 +/- 14 days) in 16 patients with high-risk esophageal varices. Portal pressure, evaluated by the portal venous pressure gradient, increased by a mean of 21% in eight patients (50%) and decreased by a mean of 20% in eight patients (50%) with no statistically significant change overall. The two groups were further analyzed separately to identify the mechanism of the change in portal pressure. Intrahepatic vascular resistance did not change significantly in either group. However, the prevalence of extravariceal portosystemic shunts was greater in patients with decreased portal pressure than in those with increased portal pressure (88% vs. 25%, p < 0.05). Further, the enlargement of extravariceal portosystemic shunts was more marked in patients with decreased portal pressure than in those with increased portal pressure (88% vs. 0%, p < 0.01). In addition, liver function, assessed by intrinsic clearance, was not modified in the two groups. We conclude that prophylactic sclerotherapy increases or decreases portal pressure without modifying liver function. Although the mechanism of these portal pressure changes is not clear, intrahepatic vascular resistance does not play an important role and the presence of extravariceal portosystemic shunts may prevent further increases in portal pressure.

Gastric mucus generation in cirrhotic patients with portal hypertension : Effects of tetraprenylacetone

We have evaluated gastric mucus generation (study 1) and the effects of tetraprenylacetone on gastric mucus generation (study 2) in cirrhotic patients with portal hypertension. Study 1: Included were 50 noncirrhotics (group A), 25 cirrhotics without portal hypertension (group B), and 25 cirrhotics with portal hypertension (group C). The antrum, corpus, and fundus mucus generation was assessed by hexosamine concentration using biopsy specimens. In groups A and B, the antrum hexosamine concentration was significantly higher compared with the corpus (P<0.01,P<0.01) and the fundus (P<0.01,P<0.01). In contrast, the hexosamine concentration at each location was similar in group C. Furthermore, the antrum hexosamine concentration of group C was significantly lower compared with that of group A (P<0.05). In study 2, a double-blind design, 300 mg of tetraprenylacetone was administered for four weeks in 10 cirrhotics with portal hypertension and placebo in 10. The regional hexosamine concentrations were measured before and after drug administration. Placebo administration did not change hexosamine concentration at each location. In contrast, tetraprenylacetone increased the antrum and corpus hexosamine concentration (P<0.01,P<0.05), although the fundus concentration did not change. These data suggest that cirrhotics with portal hypertension have reduced gastric antral mucus generation and tetraprenylacetone normalizes this.

Wedged hepatic venous pressure reflects portal venous pressure during vasoactive drug administration in nonalcoholic cirrhosis

Hepatic venous catheterization is widely used to assess portal pressure. However, it remains unclear whether wedged hepatic venous pressure is a close indicator of portal venous pressure during vasoactive drug administration in nonalcoholic cirrhosis. To address this issue, we analyzed the data from our previous published studies. Forty patients with nonalcoholic cirrhosis (HBV infection in five, HCV infection in 28, and cryptogenic in seven) were available in this analysis. A vasoconstrictor (N=14), vasodilator (N=10), or combination (N=16) was administered. The agreement of the changes between portal and wedged hepatic venous pressures during pharmacological manipulation was assessed by an intraclass correlation coefficient. The intraclass correlation coefficient in each subgroup was more than 0.60 (0.62 in vasoconstrictor group, 0.87 in vasodilator group, and 0.73 in combination group). When the analysis was performed according to the cause of liver disease, the values were 0.67 in HBV infection, 0.73 in HCV infection, and 0.74 in cryptogenic cirrhosis. These results suggest that wedged hepatic venous pressure reflects portal venous pressure during vasoactive drug administration in patients with nonalcoholic cirrhosis.

Vasopressin plus oxygen vs vasopressin alone in cirrhotic patients with portal-hypertensive gastropathy : Effects on gastric mucosal haemodynamics and oxygenation

The effects of vasopressin plus oxygen and vasopressin alone on gastric mucosal perfusion and oxygenation were studied using reflectance spectrophotometry and laser Doppler velocimetry in 23 cirrhotic patients with portal‐hypertensive gastropathy. The measurements were performed under basal conditions and after double‐blinded administration of placebo (n= 7), vasopressin (0.3 U/min; n= 8) or vasopressin (0.3 U/min) plus nasal oxygen (4 L/min; n= 8). No significant effects on gastric mucosal haemodynamics and oxygenation were observed after placebo. In contrast, vasopressin and vasopressin plus oxygen induced a similar reduction in haemoglobin content (‐26 ± 2 and ‐21 ± 4%, respectively P < 0.01) and laser Doppler signal (‐23 ± 2 and ‐22 ± 2%, respectively, P < 0.01). Although each treatment induced a significant reduction in oxygen saturation (‐21 ± 2 and ‐7 ± 1%, respectively P < 0.01), the effect was less pronounced in patients receiving the combination than in those receiving vasopressin alone (P < 0.01). These data suggest that vasopressin and vasopressin plus oxygen reduce gastric mucosal hyperaemia and that the oxygen supplement partially protects against gastric mucosal hypoxia during vasopressin infusion in cirrhotic patients with portal‐hypertensive gastropathy.

Reduced portosystemic hemodynamic responsiveness after orthostasis in patients with cirrhosis

We studied portosystemic hemodynamic responsiveness after 1 min orthostasis in nine patients with cirrhosis and nine age-matched normal subjects. Orthostasis increased diastolic arterial pressure, which is a close indicator of arterial tone, in normal subjects (+17%,P<0.01). In contrast, no significant change in diastolic arterial pressure was observed in patients with cirrhosis (−3%, NS). The increase in heart rate was less in patients with cirrhosis than in normal subjects (+15% vs +28%,P<0.05). Orthostasis also decreased portal blood flow, which was assessed by an echo-Doppler flowmetry, in normal subjects (−27%,P<0.01), but in patients with cirrhosis it was not modified (−3%, NS). Plasma noradrenaline concentration showed similar increase in both groups (normal vs cirrhosis; +61% vs +55%, NS). Although the change in plasma noradrenaline concentration was related with that in diastolic arterial pressure (r=0.71,P<0.05) and inversely with that in portal blood flow (r=−0.69,P<0.05) in normal subjects, no such significant correlation was found in patients with cirrhosis. We conclude that (1) a reduced hemodynamic responsiveness to sympathetic stimulation exists on both systemic and portohepatic vascular beds and (2) such a blunted baroreflex function is probably located at the receptor or effector level in patients with cirrhosis.

McCormack's endoscopic signs for diagnosing portal hypertension: Comparison with gastroesophageal varices

To assess the significance of McCormacks gastric mucosal signs for diagnosing portal hypertension, 100 controls and 100 patients with cirrhosis and portal hypertension underwent endoscopy. Each endoscopic recording was reviewed by multiple blinded observers to reduce bias. Individual signs more frequently observed in patients with cirrhosis and portal hypertension than in controls were fine pink speckling (20% versus 8%, p < 0.05), the snakeskin pattern (30% versus 5%, p < 0.01), and cherry-red spots (15% versus 3%, p < 0.01). In contrast, the prevalence of superficial reddening was similar in the two groups (7% versus 13%, NS). Overall, these gastric mucosal signs also appeared more commonly in patients with portal hypertension than in controls (54% versus 27%, p < 0.01); the sensitivity, specificity, and accuracy of McCormacks signs (overall assessment) for diagnosing portal hypertension were 54%, 73%, and 64%, respectively. Corresponding figures for modified McCormacks signs (exclusion of superficial reddening) were 50%, 85%, and 68%. However, these figures were still lower than those for gastroesophageal varices (72%, 100%, and 86%). We conclude that (1) superficial reddening is not a specific finding in patients with portal hypertension, and (2) gastric mucosal findings are of low sensitivity and specificity for diagnosing portal hypertension compared with gastroesophageal varices.

Arterial oxygen desaturation during non-sedated diagnostic upper gastrointestinal endoscopy in patients with cirrhosis☆☆☆★★★♢

Oxygen saturation was studied with a pulse oximeter in 80 patients with cirrhosis (44 Pugh-Childs class A, 25 class B, and 11 class C) and 80 controls undergoing diagnostic esophagogastroduodenoscopy (EGD). No narcotic agent was used during the procedure. Baseline SaO2 was significantly lower in cirrhotics than in controls (97.7 +/- 1.0% versus 98.4 +/- 0.9%, p < 0.01). However, nadir SaO2 during EGD was similar for controls and cirrhotics (94.7 +/- 3.0% versus 94.9 +/- 3.3%, NS). Significant hypoxia was found in 29 (36%) control patients: mild hypoxia (95% > nadir SaO2 > or = 90%) in 22 patients and severe hypoxia (nadir SaO2 < 90%) in 7. Similarly, significant hypoxia was noted in 28 (35%) cirrhotic patients: mild hypoxia in 21 and severe hypoxia in 7. The mean duration of significant hypoxia during total EGD time was also similar for controls and cirrhotics (7.4 +/- 6.3% versus 9.2 +/- 10.7%, NS). When the degree of hypoxia during EGD was correlated with the severity of liver disease, analysis of variance (ANOVA) failed to show a significant relationship between Pugh-Childs class and nadir SaO2 or duration of significant hypoxia during total EGD time. These results suggest that oxygen desaturation during EGD occurs both in cirrhotic patients and in controls. We therefore conclude that a population of patients with cirrhosis does not have an increased risk of oxygen desaturation during non-sedated EGD.

Effects of exercise-induced sympathoadrenergic activation on portal blood flow.

We examined the relationship between portal venous blood flow and sympathoadrenergic activation after muscle exercise. For this purpose, we used echo Doppler and measured plasma noradrenaline concentration before and after mild (7 metabolic units,N=8) and maximal exercise (14 metabolic units,N=8) in 16 patients without significant disease. Portal venous flow did not change after mild exercise. In contrast, a significant reduction in portal venous flow was observed after maximal exercise (P<0.01). This was due to reductions in both cross-sectional area of the portal vein (P<0.01) and portal venous velocity (P<0.01). Overall, there were significant inverse relationships between the change in plasma noradrenaline concentration and that in cross-sectional area of the portal vein [r=−0.44,P<0.01 (absolute change);r=−0.47,P<0.01 (relative change)], that in portal venous velocity (r=−0.63,P<0.01;r=−0.61,P<0.01), and that in portal venous flow (r=−0.54,P<0.01;r=− 0.59,P<0.01). These results suggest that the reduction in portal venous flow after exercise is related to the degree of sympathoadrenergic activation. This reduction may be due mainly to splanchnic vasoconstriction.