GASTROENTEROLOGY
1991;101:900-805
The Effects of Chronic Endoscopic Variceal Sclerotherapy on Portal Pressure in Cirrhotics JACOB KORULA
and PHILIP
RALLS
Department of Medicine (Hepatology) and Radiology, University of Southern California School of Medicine, Los Angeles, California
The effect of obliterating esophageal varices by endoscopic sclerotherapy on portal pressure was prospectively studied in 11 cirrhotic patients with variceal hemorrhage. Portal venous pressure gradient, determined as the difference between transhepatic portal and hepatic vein pressure, increased by a mean of 31.1% + 14.5% in 8 (73%) and decreased by a mean of 30.1% + 11.7% in 3 (27%) patients, with no statistically significant change overall (P = 0.1). These changes in portal venous pressure gradient occurred despite an improvement in the laboratory and clinical parameters of hepatic function. Deep abdominal sonography with color flow imaging at variceal obliteration showed patent paraumbilical veins in 6 (55%) patients, 3 of whom had decreases in portal venous pressure gradient (29%, 19%, 42.5%) at variceal obliteration. In 5 (45%) patients without patent paraumbilical veins, a statistically significant increase in portal venous pressure gradient between initial endoscopic variceal sclerotherapy and variceal obliteration was noted (P = 0.008). Rebleeding (single episode in all 4 patients, before obliteration in 3 patients) occurred in those with an increase in portal venous pressure gradient; all patients with portal venous pressure gradient decreases were nonbleeders. No correlation between changes in portal venous pressure gradient and time to variceal obliteration, number of sclerotherapy treatments, or rebleeding episodes was observed. Thus, an increase in portal venous pressure gradient was noted in the majority of patients at variceal obliteration. Although the portal venous pressure gradient decrease may be explained by a patent paraumbilical vein, the mechanism of portal venous pressure gradient increase is not clear. It is speculated that this portal venous pressure gradient increase may be caused by an increase in collateral resistance or flow or a combination of
both, resulting from obliteration rices by endoscopic sclerotherapy.
of esophageal
va-
ndoscopic variceal sclerotherapy (EVS) is considered the definitive treatment of choice in variceal The initial results of EVS are thrombohemorrhage (1). sis and necrotizing inflammation of the submucosal veins of the esophagus (2). With repeated injections, obliteration of esophageal varices eventually results (2,3), consequently reducing rebleeding risk (3). The effect of obliterating esophageal varices on portal pressure in cirrhotic portal hypertension is not known. To determine the effects of EVS on portal pressure, a prospective study was carried out in patients receiving EVS for variceal hemorrhage.
E
Materials
and Methods
Patients were considered for the study if (a) they had bled from esophageal varices and were enrolled in a program of chronic sclerotherapy and (b) they were willing to undergo portal pressure measurements at the inception of EVS and at variceal obliteration. The EVS was carried out using flexible fiberoptic endoscopy and disposable needle injectors by methods described previously (2). Sodium tetradecyl sulfate, 1.5% (3% Sotradecol; Elkins-Sinn, Cherry Hill, NJ), in 25% dextrose in water was the sclerosant used in all patients. Methylene blue, 1% [American Quinine Co., Shirley, NJ), was added to the sclerosant mixture to aid in intravariceal injection, which was the predominant type of injection made at EVS. A maximum of 25 mL was injected at each EVS. Patients were offered treatments at weekly intervals until esophageal ulceration, stricture, or decreased variceal size to 5 2 + was observed, at which time EVS was
Abbreviations used in this paper: EVS, esophageal variceal sclerotherapy; HVP, hepatic vein pressure; PVP, portal vein pressure; PVPG, portal venous pressure gradient. o 1991 by the American Gastroenterological Association oom5085/91/$3.00
September 1991
scheduled at intervals ranging from 2 weeks to 3 months. All rebleeding events were carefully documented. Portal pressure was measured using the percutaneous transhepatic method developed at this unit (4). Patients were placed supine on the radiographic fluoroscopy table, and an appropriate site on the right lateral chest wall corresponding to the 10th vertebral space was selected and infiltrated with local anesthetic. A thin needle (22-gauge Becton Dickinson, Rutherford, NJ) was inserted into the liver under fluoroscopic guidance and gradually withdrawn as contrast was injected. A portal venous branch was confirmed if arborization of the entered vessel was observed. Portal vein pressure (PVP) was recorded on a strain-gauge transducer (Gould-Statham P50, Oxnard, CA) only if blood was aspirated freely through the needle. A mean of three readings were obtained. At a separate puncture, the hepatic vein was identified if contrast injection into the cannulated vessel appeared to stream toward the hepatic outflow without branching. Hepatic vein pressure (HVP) was recorded only if blood was aspirated freely, and a mean of three readings were obtained. The portal venous pressure gradient (PVPG) was the difference between the PVP and HVP (the latter serving as zero reference]. Radiographs of the portal and hepatic vein were obtained as a permanent record of needle position, but portography was not carried out in each patient. Portal pressure studies were carried out at inception of EVS and again at variceal obliteration. Informed consent was obtained from all patients. Deep abdominal Doppler sonography of the splanchnic circulation (portal vein, splenic vein, collaterals) was carried on all patients at variceal obliteration by one of the authors (P.R.). We were unable to obtain Doppler sonography with color flow imaging at initial EVS because equipment was not available. Doppler sonography was carried out using the Acuson 128 computed sonography system [Mountain View, CA) operating at 3 MHz. The standard technique was to assess vessel location, morphology, presence, and direction of flow. A Doppler frequency of 3.5 MHz was used to obtain the flow information. Scanning was performed with the patients in the supine and left decubitus positions. When necessary, patients were given water orally to improve visualization of vessels. The Doppler sample volume was placed fully within the vessel, and the angle of insonation was kept to
1991;101:900-805
The Effects of Chronic Endoscopic Variceal Sclerotherapy on Portal Pressure in Cirrhotics JACOB KORULA
and PHILIP
RALLS
Department of Medicine (Hepatology) and Radiology, University of Southern California School of Medicine, Los Angeles, California
The effect of obliterating esophageal varices by endoscopic sclerotherapy on portal pressure was prospectively studied in 11 cirrhotic patients with variceal hemorrhage. Portal venous pressure gradient, determined as the difference between transhepatic portal and hepatic vein pressure, increased by a mean of 31.1% + 14.5% in 8 (73%) and decreased by a mean of 30.1% + 11.7% in 3 (27%) patients, with no statistically significant change overall (P = 0.1). These changes in portal venous pressure gradient occurred despite an improvement in the laboratory and clinical parameters of hepatic function. Deep abdominal sonography with color flow imaging at variceal obliteration showed patent paraumbilical veins in 6 (55%) patients, 3 of whom had decreases in portal venous pressure gradient (29%, 19%, 42.5%) at variceal obliteration. In 5 (45%) patients without patent paraumbilical veins, a statistically significant increase in portal venous pressure gradient between initial endoscopic variceal sclerotherapy and variceal obliteration was noted (P = 0.008). Rebleeding (single episode in all 4 patients, before obliteration in 3 patients) occurred in those with an increase in portal venous pressure gradient; all patients with portal venous pressure gradient decreases were nonbleeders. No correlation between changes in portal venous pressure gradient and time to variceal obliteration, number of sclerotherapy treatments, or rebleeding episodes was observed. Thus, an increase in portal venous pressure gradient was noted in the majority of patients at variceal obliteration. Although the portal venous pressure gradient decrease may be explained by a patent paraumbilical vein, the mechanism of portal venous pressure gradient increase is not clear. It is speculated that this portal venous pressure gradient increase may be caused by an increase in collateral resistance or flow or a combination of
both, resulting from obliteration rices by endoscopic sclerotherapy.
of esophageal
va-
ndoscopic variceal sclerotherapy (EVS) is considered the definitive treatment of choice in variceal The initial results of EVS are thrombohemorrhage (1). sis and necrotizing inflammation of the submucosal veins of the esophagus (2). With repeated injections, obliteration of esophageal varices eventually results (2,3), consequently reducing rebleeding risk (3). The effect of obliterating esophageal varices on portal pressure in cirrhotic portal hypertension is not known. To determine the effects of EVS on portal pressure, a prospective study was carried out in patients receiving EVS for variceal hemorrhage.
E
Materials
and Methods
Patients were considered for the study if (a) they had bled from esophageal varices and were enrolled in a program of chronic sclerotherapy and (b) they were willing to undergo portal pressure measurements at the inception of EVS and at variceal obliteration. The EVS was carried out using flexible fiberoptic endoscopy and disposable needle injectors by methods described previously (2). Sodium tetradecyl sulfate, 1.5% (3% Sotradecol; Elkins-Sinn, Cherry Hill, NJ), in 25% dextrose in water was the sclerosant used in all patients. Methylene blue, 1% [American Quinine Co., Shirley, NJ), was added to the sclerosant mixture to aid in intravariceal injection, which was the predominant type of injection made at EVS. A maximum of 25 mL was injected at each EVS. Patients were offered treatments at weekly intervals until esophageal ulceration, stricture, or decreased variceal size to 5 2 + was observed, at which time EVS was
Abbreviations used in this paper: EVS, esophageal variceal sclerotherapy; HVP, hepatic vein pressure; PVP, portal vein pressure; PVPG, portal venous pressure gradient. o 1991 by the American Gastroenterological Association oom5085/91/$3.00
September 1991
scheduled at intervals ranging from 2 weeks to 3 months. All rebleeding events were carefully documented. Portal pressure was measured using the percutaneous transhepatic method developed at this unit (4). Patients were placed supine on the radiographic fluoroscopy table, and an appropriate site on the right lateral chest wall corresponding to the 10th vertebral space was selected and infiltrated with local anesthetic. A thin needle (22-gauge Becton Dickinson, Rutherford, NJ) was inserted into the liver under fluoroscopic guidance and gradually withdrawn as contrast was injected. A portal venous branch was confirmed if arborization of the entered vessel was observed. Portal vein pressure (PVP) was recorded on a strain-gauge transducer (Gould-Statham P50, Oxnard, CA) only if blood was aspirated freely through the needle. A mean of three readings were obtained. At a separate puncture, the hepatic vein was identified if contrast injection into the cannulated vessel appeared to stream toward the hepatic outflow without branching. Hepatic vein pressure (HVP) was recorded only if blood was aspirated freely, and a mean of three readings were obtained. The portal venous pressure gradient (PVPG) was the difference between the PVP and HVP (the latter serving as zero reference]. Radiographs of the portal and hepatic vein were obtained as a permanent record of needle position, but portography was not carried out in each patient. Portal pressure studies were carried out at inception of EVS and again at variceal obliteration. Informed consent was obtained from all patients. Deep abdominal Doppler sonography of the splanchnic circulation (portal vein, splenic vein, collaterals) was carried on all patients at variceal obliteration by one of the authors (P.R.). We were unable to obtain Doppler sonography with color flow imaging at initial EVS because equipment was not available. Doppler sonography was carried out using the Acuson 128 computed sonography system [Mountain View, CA) operating at 3 MHz. The standard technique was to assess vessel location, morphology, presence, and direction of flow. A Doppler frequency of 3.5 MHz was used to obtain the flow information. Scanning was performed with the patients in the supine and left decubitus positions. When necessary, patients were given water orally to improve visualization of vessels. The Doppler sample volume was placed fully within the vessel, and the angle of insonation was kept to
