Different dose regimens of 5-fluorouracil and interferon-alpha in patients with metastatic colorectal carcinoma.

Three different 5-fluorouracil (5-FU)-interferon-alpha-2b (IFN)-containing regimens were designed for treatment of patients with advanced colorectal cancer. 87 patients with a Karnofsky index > or = 70 were included in three sequential non-randomised phase II trials. Regimen A consisted of 5-FU (750 mg/m2/day) given as a continuous infusion on days 1-5 followed by weekly 1-h intravenous infusions until week 8. IFN (5 MU) was given subcutaneously on days 1, 3 and 5 followed by injections (9 MU) every second day until week 8. The cycle was then repeated. Regimen B consisted of 5-FU (750 mg/m2/day) given as a continuous infusion on days 1-5 followed by 5-min intravenous injections on days 12 and 19. IFN (3 MU) was given subcutaneously on days 1-5 followed by injections (5 MU) on days 11-13 and 18-20. The cycle was repeated every fourth week. Regimen C consisted of 5-FU (750 mg/m2/day) given as a continuous infusion on days 1-5. IFN (3 MU) was given subcutaneously on days 1-5. The cycle was repeated every third week. The objective response rates (complete response (CR) and partial response (PR)) after approximately 4 months of therapy or longer were as follows: regimen A (n = 27) 22% (2 CR, 4 PR), regimen B (n = 33) 42% (4 CR, 10 PR) and regimen C (n = 27) 22% (1 CR, 5 PR). The corresponding response figures for previously untreated patients were regimen A 50%, regimen B 64% and regimen C 38%. Response durations varied from a few weeks up to 142 + weeks. Toxicities were generally mild and reversible, and the treatments were convenient for the patients and cost effective since the 5-day infusions could be given by a portable pump without hospitalisation. Our results are in agreement with those of others showing that 5-FU/IFN combinations can be highly effective in advanced colorectal cancer, and that a number of factors such as doses, dose intensities, infusion rates and timing of the two drugs may be crucial for the anti-tumour activity of this drug combination.

Ontogeny of a human polychlorinated biphenyl-binding protein. Level of expression in tracheal aspirates in bronchopulmonary dysplasia.

A human lung polychlorinated biphenyl binding protein (PCB-BP,M(r) 13 kd) has recently been purified from lavage fluid. Polyclonal monospecific antibodies against PCB-BP were produced and used for immunohistochemical staining in sections of human lung tissue. PCB-BP was found to localize preferentially to the nonciliated (Clara) cells of the lung, whereas the alveolar cells and ciliated cells of the larger airways were devoid of staining. Tracheal aspirates from infants receiving mechanical ventilation were subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis separation and Western immunoblotting. The antibodies to human PCB-BP detected a single band of the expected molecular weight, and a quantitative analysis of the ontogeny of PCB-BP in tracheal aspirates was performed by construction of Western immunoblot standard curves. A significant increase in the levels of PCB-BP in late gestation (gestational weeks 39 to 41) was demonstrated. In similar experiments, levels of PCB-BP in tracheal aspirates obtained from infants with bronchopulmonary dysplasia (BPD) at the postconceptional age of less than 38 weeks were found to be significantly elevated as compared with a normal study group of similar gestational age (21.8 +/- 4.8 vs 3.1 +/- 0.8 ng of PCB-BP per microgram of total protein, p < 0.005). It is suggested that the high levels of PCB-BP at the postconceptional age of less than 38 weeks observed in infants with BPD may reflect inflammatory injury and regeneration of airway epithelium, associated with proliferation of Clara cells.

Muscle protein synthesis after operation: effects of intravenous nutrition.

OBJECTIVE: To assess the effect of an elective abdominal surgical operation (open cholecystectomy) on the rate of protein synthesis in skeletal muscle in humans. DESIGN: Prospective random control trial. SETTING: University hospital. SUBJECTS: 17 Metabolically healthy patients who were to undergo elective open cholecystectomy. INTERVENTIONS: Patients randomised to receive either saline alone (n = 8) or total parenteral nutrition (n = 9) for three days after operation. The rate of protein synthesis in muscle was calculated from the increase in enrichment of (1-13C) leucine in protein after a flooding dose of (1-13C) leucine. RESULTS: Median (quartiles) rate of protein synthesis had decreased on the third postoperative day in the saline group by 49% (from 2.42 [2.03, 2.54] to 1.24 [0.99, 1.63]) and in the group that had received total parenteral nutrition by 54% (from 1.96 [1.90, 2.07] to 0.91 [0.79, 1.06]) (p < 0.01). CONCLUSIONS: The trauma associated with open cholecystectomy reduced the rate of protein synthesis in skeletal muscle by half in three days, and conventional total parenteral nutrition had no effect on these changes.

Mechanisms for differences in lipolysis between human subcutaneous and omental fat cells.

Catecholamine-regulation of lipolysis and beta-adrenoceptor binding isoterms were studied in human sc and omental isolated fat cells from 24 subjects undergoing elective cholecystectomy. The lipolytic sensitivity of the nonselective beta-agonists epinephrine and isoprenaline as well as the selective agonists norepinephrine (beta 1) and terbutaline (beta 2) was significantly increased 5-10 times in omental fat cells. On the other hand, no regional difference in antilipolytic sensitivity was seen for the alpha 2 agonist clonidine. No regional difference in lipolytic action was seen when measuring the effect of forskolin, (Bu)2cAMP or enprofylline, which act at different postadrenoceptor steps in the lipolytic cascade. Lipolysis data showed no sex differences. A beta 1-pattern was seen in both regions when lipolysis dose-response curves were arranged in order of potency. Radioligand saturation experiments with the nonselective beta-antagonist 125I-cyanopindolol and competition experiments between this radioligand and the selective antagonists CGP-20,712-A (beta 1) and ICI-118551 (beta 2) showed a 2-fold increase in the amount of beta 1- and beta 2-adrenergic receptors in omental as compared to sc fat cells (P less than 0.02). Competition studies with the same radioligand and the nonselective beta-agonist isoprenaline showed no regional differences in terms of receptor affinity (Kd high 10 nM and Kd low 1 microM) or in relative fraction of receptors in the high affinity state (35%). It is concluded that an increased lipolytic sensitivity for beta 1- and beta 2-agonists can be due to an increase in the amount of the two adrenoceptor subtypes in omental fat cells and thereby explain why catecholamines are more lipolytic in omental cells than in sc fat cells.

Free amino acids in plasma and muscle during 24 hours post-operatively--a descriptive study.

Intracellular amino acids in skeletal muscle show a specific concentration pattern on the third post-operative day. The temporal development of these changes has not been clarified. Here the amino acid concentrations in skeletal muscle were studied during the first post-operative day in fourteen patients undergoing elective abdominal surgery. Muscle amino acids were determined pre-operatively, as well as at 12 and 24 h post-operatively. In muscle the concentrations of glutamine and the basic amino acids decreased gradually during the first 24 h after surgery to 79% (P less than 0.001) and 67% (P less than .001) respectively. The sum of the essential amino acids decreased to 73% (P less than 0.001) at 12 h, but thereafter rose to 91% (P less than 0.05) at 24 h. The sum of the BCAA decreased to 84% (P less than 0.05) at 12 h but then increased to 116% (P less than 0.05) at 24 h. The alanine concentration increased to 122% (P less than 0.001) during the first post-operative day. In plasma the alanine concentration increased at 12 h while most other amino acids declined. At 24 h post-operatively the plasma concentrations of all amino acids had returned to normal or showed a tendency towards normalization except for phenylalanine, which increased. At the end of the first post-operative day the concentrations of amino acids in muscle were consistent with the alterations previously observed three days after surgery. The changes in plasma amino acid concentrations only partly reflected those in muscle.

Quantitative comparison of insulin-like growth factor mRNA levels in human and rat tissues analysed by a solution hybridization assay.

In this study a solution hybridization assay was evaluated for its application to the measurement of levels of specific mRNAs. The evaluation included parameters such as incubation time, hybridization stringency and probe concentration/structure. Both short (50 bases derived from synthetic oligonucleotides) and long (125-147 bases) RNA probes, derived from cloned sequences, could be used to obtain quantitative information on specific mRNA species. The solution hybridization assay was used to compare the levels of insulin-like growth factor-I (IGF-I) and IGF-II mRNAs in various rat and human tissues. In the rat the liver was the main source of IGF-I mRNA (approximately 400 molecules/cell), but significant levels were also found in extrahepatic tissues such as fat and muscle (3-50 molecules/cell). Human liver contained approximately 100-fold less IGF-I mRNA than rat liver. Human fat, muscle and placenta contained levels of IGF-I mRNA (2-8 molecules/cell) similar to those in the liver. Levels of IGF-II mRNA in rat and human tissues were similar, in that the expression was greatest in the placenta (approximately 200 molecules/cell). Species differences were evident, however, since human liver and fat contained significant amounts of IGF-II mRNA (15-20 molecules/cell), while the rat counterparts had almost undetectable levels. Young and old rats were used to examine the influence of age on the expression of IGF-I and GH receptor mRNAs in the liver. Levels of both IGF-I mRNA and GH receptor mRNA were found to decrease with age (2.8-fold and 1.7-fold respectively). It is concluded that low levels of IGF mRNAs can be detected using the solution hybridization assay and that there are considerable species differences within and between tissues with regard to steady-state levels of IGF-I and IGF-II mRNAs.

Glucocorticoid receptor mRNA expression in pulmonary alveolar macrophages in sarcoidosis.

The presence of the glucocorticoid receptor was demonstrated by immunocytochemistry in pulmonary alveolar macrophages obtained by bronchoalveolar lavage. Also, GR mRNA content was determined by solution hybridization in PAM from 12 healthy volunteers and in 6 patients with sarcoidosis. No significant differences with regard to GR mRNA expression was detected between the two groups examined. For comparison, lung tissue from three patients undergoing thoracic surgery was examined and found to contain GR mRNA levels in the same range. As an indication of GR function, we also determined the mRNA levels of a glucocorticoid-regulated gene, metallothionein IIA, during basal conditions and after in vitro incubation of PAM with dexamethasone. Neither the control sample nor the dexamethasone-stimulated MTII mRNA values in PAMs differed significantly between the two groups. Solution hybridization is a rapid, sensitive and convenient assay which enables accurate and specific quantitation of GR mRNA in PAM. The GR mRNA content and basal as well as dexamethasone-induced MTII mRNA levels in PAM from patients with sarcoidosis is not significantly different from those in healthy volunteers.

Decreased insulin-stimulated 3-0-methylglucose transport in in vitro incubated muscle strips from type II diabetic subjects.

Peripheral insulin resistance in type II diabetes mellitus has been attributed to alterations in skeletal muscle glucose metabolism. However the direct dose-response relationship between insulin and glucose transport has not yet been studied in human skeletal muscle. We investigated 3-0-methylglucose transport in in vitro incubated skeletal muscle strips from eight healthy controls (age 61 +/- 6 yrs) and six lean type II diabetic patients treated with oral antidiabetic medication (age 73 +/- 3 yrs). Rectus abdominis muscle samples (approximately 1 g), obtained during elective abdominal surgery, were clamped at their resting length in vivo, whereupon strips (20-50 mg) were prepared for in vitro incubation. Measurements of high-energy phosphates and glycogen levels revealed that the muscle strips maintained energy levels during the incubation period. Glucose transport responded to insulin in a dose-response manner in the control group, with a 2-fold increase following maximal stimulation. Muscle strips from the diabetic group demonstrated a marked decrease in the insulin dose-response curve (P less than 0.01), when compared to healthy muscle strips. At a maximal insulin concentration (10,000 microU x ml-1), the response of the diabetic muscle tissue was 50% less than that of the healthy control tissue (P less than 0.05). This report demonstrates a dose-response curve for insulin stimulated 3-0-methylglucose transport in in vitro incubated human skeletal muscle strips. Furthermore, in type II diabetic muscle, our results provide evidence for one or several defects at a postreceptor level.

Metformin increases insulin-stimulated glucose transport in insulin-resistant human skeletal muscle.

The effect of metformin (0.1 mM) on glucose transport was investigated in healthy control and in insulin-resistant human skeletal muscle. Muscle samples (200-400 mg) were obtained from the rectus abdominis muscle (abdominal surgery) or from the vastus lateralis portion of the quadriceps femoris muscle (open biopsy technique) from 8 healthy controls (age 38 +/- 4 yrs, BMI 23 +/- 1) and from 6 insulin-resistant subjects (age 53 +/- 5 yrs, BMI 30 +/- 2). Metformin had no effect on basal or insulin-stimulated (100 microU/ml) 3-0-methylglucose transport in incubated muscle strips from healthy subjects. Muscle tissue from the insulin resistant group did not respond to 100 microU/ml of insulin (0.73 +/- 0.17 for basal and 0.81 +/- 0.22 mumol x ml-1 x h-1 for insulin-stimulation, NS). Basal glucose transport was unaffected by metformin, whereas insulin-stimulated (100 microU/ml) glucose transport was increased by 63% in the insulin-resistant muscles (0.73 +/- 0.17 in the absence vs 1.19 +/- 0.18 mumol x ml-1 x h-1 in the presence of metformin, p less than 0.05). In conclusion, metformin abolishes insulin-resistance in human skeletal muscle by normalizing insulin-stimulated glucose transport accross the muscle cell membrane. The mechanism for this effect remains to be elucidated.

Vasoactive intestinal peptide and secretin: effects of combined and separate intravenous infusions on bile secretion in man.

The effects of intravenously administered vasoactive intestinal peptide (VIP) and secretin on bile secretion were studied in 12 patients with complete biliary fistulas. The two peptides were administered both simultaneously and separately. During VIP infusion bile volume increased by 60%, and during the combined VIP and secretin infusion bile volume increased by another 70%. VIP increased bile bicarbonate concentration by some 30%. Although secretin did not increase the concentration, bicarbonate output increased threefold during secretin infusion but only twofold during VIP infusion. The outputs of bile acids were not significantly affected by the two peptides, whereas the concentrations decreased by 40% and 70% after VIP and secretin, respectively. The canalicular bile flow, measured by [14C]erythritol, was unaffected by VIP infusion, whereas secretin alone and the combination of the two peptides increased the canalicular clearance by 80%. The choleretic effect of VIP thus seems to occur only at the ductular level. Secretin exerts its effect at the ductular level and possibly also at the canalicular level. It is concluded that the two peptides have additive effects on the ductular bile flow.

Beta-adrenoreceptor subtype expression in human liver.

The pharmacological and gene expressions of beta 1- and beta 2-adrenoceptor subtypes (BAR1 and BAR2) were investigated in human liver by radioligand binding assays, adenylate cyclase experiments, and RNA excess solution hybridization. [125I]Cyanopinodolol, nonlabeled adrenergic agents, and BAR1/BAR2 cRNA were used as probes. The relationship between binding sites for BAR1 and BAR2 was markedly different from that between the basal mRNA expression for the two receptor subtypes. Plasma membranes as well as a microsomel-enriched fraction contained binding sites only for BAR2. The potency of BAR agonists and antagonists in stimulating adenylate cyclase activity of plasma membranes was typical of a BAR2 response. Northern blot analysis of total cellular RNA isolated from liver tissue showed hybridization of the BAR1 probe to a mRNA species of 2.5-2.6 kilobases and of the BAR2 probe to a mRNA species of 2.2-2.3 kilobases. The basal level of BAR1 mRNA was 5-fold higher than of BAR2 mRNA, as assayed by solution hybridization. No difference in BAR subtype mRNA stability was observed, as indicated by a mRNA half-life of approximately 5.5 h for both receptor subtypes. It is concluded that specific factors are involved in the steady state regulation of BAR subtype expression in human liver. This tissue contains solely BAR2 owing to a posttranscriptional block of basal BAR1 expression.

Inhibition of lipolysis by agents acting via adenylate cyclase in fat cells from infants and adults.

The in vitro lipolytic effect of catecholamines is poor during infancy because of enhanced alpha 2-adrenoceptor activity. The mechanisms behind this were investigated in isolated fat cells obtained from 1- to 4-mo-old infants and from adults. The cells were incubated with agents that inhibit lipolysis through distinct receptors coupled to adenylate cyclase via the inhibitory GTP binding coupling protein, Gi. The sensitivity to the alpha 2-adrenoceptor agonist clonidine was 14 times higher in the infant group as compared to the adults, whereas that to an adenosine analogue was 14 times lower. The sensitivities to prostaglandin E2 and nicotinic acid were similar in both age groups. Preincubation of the adipocytes with pertussis toxin abolished the antilipolytic effects of all agents. The density of alpha 2-adrenoceptor binding sites determined with [3H]yohimbine was increased by about 25% in the infants. In conclusion, the antilipolytic sensitivity of adenosine and alpha 2-adrenoceptors develops separately and may play different roles in the regulation of lipolysis in man. Furthermore, the enhanced alpha 2-adrenoceptor sensitivity during infancy seems at least in part to be due to an increase in the number of receptors.

Evidence that vasoactive intestinal peptide induces ductular secretion of bile in humans.

The effect of intravenously administered vasoactive intestinal polypeptide on bile secretion was studied in 11 patients with complete biliary drainage. After infusion of vasoactive intestinal polypeptide, bile volume increased by 65%. In the 2 patients investigated, the output of bicarbonate increased by approximately 250% and the concentration by 50%-70%. Vasoactive intestinal polypeptide thus caused a bicarbonate-rich choleresis. The output of biliary lipids was not affected by infusion of vasoactive intestinal polypeptide, whereas the concentration decreased by approximately 40%. The canalicular bile flow, measured by the clearance of [14C]erythritol, was not affected by infusion of vasoactive intestinal polypeptide. The choleretic effect of vasoactive intestinal polypeptide thus seems to occur only at the ductular level. The ductular bile flow was calculated to be stimulated threefold to fourfold.

Characterization of insulin resistance after surgery.

The effects of insulin on glucose utilization were investigated in seven nonobese patients before and 24 hours after elective cholecystectomy. Surgery was followed by a significant increase in the circulating levels of glucose and insulin. The hypoglycemic action of insulin was reduced by one third (p less than 0.01) after surgery. In isolated fat cells after surgery there was a significant overall reduction of 35% to 50% of the effects of insulin on 3-0-methylglucose transport and lipogenesis at 1 mumol/L of glucose (where hexose transport is rate-limiting for insulin action). However, there was no change in insulin sensitivity in these cells. The effects of insulin on lipogenesis in adipocytes incubated with 100 mumol/L of glucose (where glucose metabolism is rate-limiting for insulin action) and adipocyte insulin receptor binding were not influenced by surgery. Insulin action in vivo and in vitro was not altered in five nonoperated control subjects 24 hours after they were given the same type of nutritional support as the cholecystectomy patients postoperatively. It was concluded that an elective moderate surgical trauma induces a rapid and marked insulin resistance that is not the result of postoperative nutritional restriction and involves a postreceptor binding alteration of glucose transport.

Changes in catecholamine-induced lipolysis in isolated human fat cells during the first year of life.

Catecholamine-induced lipolysis in isolated human adipocytes during the first year of life was investigated. During this period fat cell size increased markedly. Basal and catecholamine-induced glycerol release were positively correlated with age when lipolysis was expressed per cell. However, when lipolysis was expressed per unit of cell surface area (micrometer squared), this correlation was observed only for noradrenaline. Basal lipolysis and the effect of the pure beta-agonist, isoprenaline, were identical in infants and adults. From 0 to 2 mo of age noradrenaline had very little lipolytic effect. The addition of the alpha-2-adrenoceptor antagonist, yohimbine, to noradrenaline equalized lipolysis per micrometer squared in infants and adults and the alpha-2-adrenoceptor sensitivity was significantly enhanced in infants. In both groups the lipolytic adrenoceptor was of the beta-1 type. In conclusion, adipocytes from infants have a poor lipolytic response to noradrenaline partly because of the small fat cells but mainly because of an enhanced alpha-2-adrenoceptor activity.

Endoscopic fluorescein flowmetry in the evaluation of gastric and duodenal mucosal blood flow.

Fluorescein flowmetry (FF) implies the measurement of a relative capillary blood flow, expressed as an index--that is, the ratio between the maximum fluorescence, obtained during the first circulatory passage of sodium fluorescein (Na-F), and the rise time, defined as the time interval between 10% and 90% of the maximum fluorescence. The aim of this study was to develop FF to be used during endoscopy for the evaluation of ventricular and duodenal mucosal blood flow. FF was applied during gastroduodenoscopy in 38 patients with normal mucosa, as verified by histopathologic examination. The blood flow index did not differ significantly for the mucosa of the fundus, corpus, and antrum. However, when compared with the duodenal mucosa, the blood flow index of the ventricular mucosa was almost twice as high (P less than 0.01). This, in turn, was due to a significantly higher maximum fluorescence in the ventricular mucosa (P less than 0.01), indicating a denser capillary network than in the duodenal mucosa. Since FF is a reliable method, easy to perform and without complications, it is suitable whenever mucosal blood flow warrants measurement during endoscopy.

The transcapillary exchange of sodium fluorescein in ischaemic limbs measured by fluorescein flowmetry.

The transcapillary exchange of sodium fluorescein in the skin of 36 extremities (19 patients) with arterial occlusive disease was measured using two different ways of analysing the tissue fluorescence, fluorescein flowmetry (FF) and dynamic fluorescein angiography (FA). 7 mg sodium fluorescein/kg body weight was given intravenously. With FF, the transcapillary exchange of sodium fluorescein is expressed in arbitrary units as a fluorescence index. With FA, the time interval from the bolus injection to the first appearance of the fluorescence in the tissue is measured. There was an inverse correlation between pain and fluorescence index. All eight extremities with a fluorescence index of 0.001 density units/s or lower needed amputation, whereas all extremities with higher indices were viable. All extremities requiring amputation had an appearance time of 150 s or more, which contrasted with those remaining viable, having an appearance time of 120 s or less. The results indicate that FF as well as FA are appropriate methods to measure a 'nutritive' transport of solutes in the skin. Theoretically, however, if quantitative values for blood flow are required, we recommend FF, this method being more adequate, since it is influenced both by the fraction of cardiac output distributed to the tissue and by cardiac output itself. If, however, only a prognosis of a limb's viability is needed, FA is sufficient, being easier to perform, only requiring the measurement of the appearance time.

Role of alpha-adrenoceptors for adipocyte size in man.

Catecholamines have dual effects on lipid mobilization in man. Lipolysis is stimulated via beta-adrenoceptors and inhibited via alpha 2-adrenoceptors. The relationship between fat-cell size and catecholamine-induced lipolysis (expressed per cell surface area) was investigated in vitro in subcutaneous adipocytes of thirty-five non-obese subjects between 1 month and 45 years of age. Fat-cell volume showed a positive correlation with noradrenaline-induced lipolysis (r = 0.7). Furthermore, fat-cell size showed a negative correlation with the alpha 2-effect of noradrenaline (r = 0.8) but no correlation with the beta-effect of the hormone (r less than 0.1). Although age showed a positive correlation with noradrenaline-induced rate of lipolysis (r = 0.6) it did not contribute to the relationship between cell size and the catecholamine effect. There was no relationship between adipocyte size and the basal (unstimulated) rate of lipolysis (r less than 0.1). In conclusion, ageing in non-obese subjects is associated with enlargement of fat-cell size and enhancement of the lipolytic effect of catecholamines. The latter is due to diminished alpha 2-anti-lipolytic effect of the hormones. This modulation of the alpha 2-adrenoceptor activity may be of importance for the regulation of adipocyte size in man.

Nonresectable malignant bile duct obstruction. Surgical bypass or endoprosthesis?

Forty-one patients with malignant obstructive jaundice, not amenable to radical surgery, had biliary decompression by either surgical bypass or endoscopically introduced endoprosthesis. The two treatment groups were statistically comparable with reservation for age, which was significantly lower in the operated group. Judged by the effect on bilirubin and alkaline phosphatase levels the two methods were equally efficient. Major complications were more common and initial hospital stay was significantly longer in the surgical bypass group. Neither early mortality nor survival differed between the groups. Implications of these data in the management of patients with unresectable malignancy are discussed.

Peritoneovenous shunts for malignant ascites.

The intractable malignant ascites of 27 patients was treated by insertion of a peritoneovenous shunt. Eight had a Le Veen shunt, 17 a Denver shunt, one patient had a Denver shunt subsequently replaced by a Le Veen shunt and one a Denver shunt which was replaced by another Denver shunt. The operations were safely performed under local anesthesia. Shunt occlusion occurred in eight patients, in two of whom, however, the shunt was cleared with the flushchamber (Denver shunt) and by anticoagulant treatment (Le Veen shunt). In two further patients the shunts were replaced. All patients with a satisfactory functioning shunt had good palliation. Thanks to the shunts, 16 patients were able to leave hospital to spend their remaining time at home. Survival times were generally very short due to the patients' terminal malignancy, all patients dying within 4 months. In three patients there was a possible relationship between shunt placement and death. In conclusion, peritoneovenous shunting is indicated in selected patients with malignant ascites.