High Plasma Levels of Phytosterols in Patients on Parenteral Nutrition: A Marker of Liver Dysfunction

Parenteral nutrition–associated cholestatic liver disease (PNAC) is considered to be multifactorial in origin, and its incidence has decreased during the past years (1). The pathogenesis of PNAC, however, remains poorly understood.Commercially available intravenous fat emulsions are soy-based preparations containing the phytosterols (PS) sitosterol, campesterol, and sigmasterol at a total concentration of approximately 1000 μmol/L (40 mg/dL). The observation of high plasma levels of PS in patients with PNAC, receiving intravenous fat emulsions, led to the hypothesis that accumulation of these substances contributes to cholestasis (2–4). An animal model, disclosing decreased bile flow induced by infusion of PS (3), supported the assumption. In this study, we monitored the sitosterol plasma concentrations in a heterogeneous group of patients with liver dysfunction during parenteral nutrition (PN). Plasma concentrations of sitosterol were measured, using accurate gas-liquid chromatography-mass spectrometry techniques, as previously described (5). In total, we analyzed five patients on PN. Three of the five patients developed steatosis with non-icteric hepatic dysfunction. Their course improved with reduced glucose infusion-rates while the dosage of lipid emulsions remained unchanged. Measurements of plasma sitosterol were continued in two patients with severe PNAC after stopping or reducing fat infusion. CASE DESCRIPTIONS Case 1 A 2-year-old boy with short bowel syndrome after neonatal volvulus. Residual intestinal length was 15 cm of jejunum; the ileocecal region had been resected. There had been several parental attempts to wean the child from PN and the parents had reduced the PN to twice weekly. The child was admitted with a history of three episodes of pancreatitis and one episode of cholangitis, and with growth retardation (body weight was 7 kg, length 69 cm) to evaluate the reason of recurrent pancreatitis. The endoscopic retrograde cholangiopancreatography (ERCP) disclosed a stenosed common channel, dilated ductus Wirsungianus, and stenosis of the ductus choledochus. The clearance of contrast medium from the bile duct was delayed. A papillotomy was performed. Oral nutrition was continued and PN was reassumed with 66 kcal/kg/day, non-protein calories were given through a daily infusion of lipid emulsion (Lipovenös 20% [Fresenius Kabi GmbH, Bad Homburg, Germany], 2 g/kg body weight [BW]) and glucose infusion (59% of total parenteral caloric intake; infusion rate 14.6 mg/kg/min). Amino acids were provided as Aminopaed (Pharmacia Upjohn GmbH, Erlangen, Germany) at a dose of 2 g/kg/day. The previously normalized serum levels of aspartate aminotransferase (AST), alanine aminotransferase (ALT), and γ-glutamyltransferase (GGT) rose again. Hepatic steatosis due to parenteral nutrition was diagnosed. The maximum glucose infusion rate was reduced to 8 mg/kg/min while total lipid dosage was maintained. Sonographic signs of liver steatosis disappeared and liver function tests normalized within 9 month. As shown in Figure 1A, the significantly elevated sitosterol plasma levels (maximum 6.6 mg/dL) declined along with normalization of liver function tests despite unaltered intravenous lipid infusion.FIG. 1.: A–C: Course of the plasma level of sitosterol (▪) in mg/dL, bilirubine (x) in mg/dL, aspartate aminotransferase (AST; □) times upper limit, γ glutamyltransferase (GGT) times upper limit of normal, the mean intravenous lipid dosage (g/kg/day), and the glucose infusion rate (mg/kg/minute) in cases 1 and 3. The time axis begins with hospital admission in case 1 and with reassumption and start of parenteral nutrition, respectively, at our institution in cases 2 and 3.Figure 1: ContinuedFigure 1: ContinuedCase 2 A 15-year-old boy had short bowel syndrome due to volvulus at 7 years of age. Fifteen centimeters of jejunum were left; the ileocecal region had been resected. He was admitted with growth retardation (body weight was 32 kg, length 149 cm). PN (30% of total caloric requirements) had been stopped due to lack of central venous access 3 month before. Sonography showed hepatic steatosis. Serum levels of AST, ALT, GGT, bile acids, and bilirubin were within normal limits. In addition to oral nutrition and nighttime continuous enteral nutrition via percutaneous endoscopic gastrostomy tube, PN was restarted. Non-protein calories were given as daily infusion of Lipovenss 20% (1.8 g/kg/BW) and glucose infusion (54% of total parenteral caloric intake; infusion rate 13 mg/kg/min). Amino acids were provided as Aminosteril (1.5g/kg/day; Fresenius Kabi GmbH). PN was administered 6 days a week. Five months later, levels of AST, ALT, bile acids, and GGT rose. Although sonographic signs of hepatic steatosis existed before reassuming PN, the abnormal liver function tests were attributed to PN-associated hepatic steatosis. The initial low levels of sitosterol (0.11 mg/dL) increased to 10.5 mg/dL concomitant with the impaired liver function. Maximum glucose infusion rate was reduced to 8 mg/kg/min and later to 6 mg/kg/min. Normal liver function was restored 9 months later and the sitosterol plasma levels declined, though total lipid dosage was maintained (Fig. 1B). Case 3 A 3-month-old female infant, who was born at 35 gestational weeks, was admitted with short bowel syndrome due to necrotizing enterocolitis at 1 month of age. Fifteen centimeters of jejunum and 10 cm terminal ileum were left with the ileocecal region intact. Body weight was 3.2 kg; length was 51 cm. PN (up to 70% of total caloric requirements) contained daily infusion of Lipovenss 20% (1.8–3.5 g/kg BW.) and glucose infusion (54% of total parenteral caloric intake; infusion rate 13 mg/kg/min). Amino acids were provided as Aminopaed at a dose of 2g/kg/day. Serum levels of AST, ALT, GGT, bile acids and bilirubin were within normal limits. At 4 month of age, during a sepsis episode, levels of AST, ALT, bile acids and GGT rose. Bilirubin was 2 mg/dL. Levels of AST remained elevated for 6 weeks. At the onset of sepsis, the caloric intake by PN was reduced to 30% of daily requirements for 2 weeks and restarted with 60% of daily requirements on 5 days a week. Daily infusion of lipid emulsion ranged from 2.3 to 2.9 g/kg BW. The amount of oral caloric intake was increased and PN was gradually reduced to 2 days a week and terminated at the age of 15 months. Normalization of liver function tests occurred within 5 weeks from onset of sepsis. During sepsis, sitosterol plasma levels were highly elevated (maximum 11.4 mg/dL). Although intravenous lipid dosage was increased between month 6 and 8 of PN (Fig. 1C), concentrations of plasma sitosterol declined with normalization of liver function tests. Case 4 A 3-year-old boy with microvillus atrophy, who was on total PN, developed severe PNAC in the third year of life. Liver biopsy showed cirrhosis. To improve PNAC, intravenous fat infusions were reduced to 4 g/kg weekly, sufficient to maintain the supply with unsaturated fatty acids and fat-soluble vitamins. Plasma levels of albumin, cholesterol, and clotting factors steadily declined. The dramatically high sitosterol plasma levels (36 mg/dL), despite massive reduction of intravenous lipid administration, declined slowly (Fig. 2). Normal basal values as observed in uncomplicated PN (Fig. 3) were not measured, even after an observation interval of 140 days.FIG. 2.: Decrease in Sitostrol plasma level in patients with cholestatic icterus after total stop of intravenous lipid (Pt5) or reduction to 4 g/kg/week (Pt4).FIG. 3.: Sitosterol plasma levels in long-term parenteral nutrition patients with normal liver function (n = 8).Case 5 A 2-month-old female infant, who was born at 35 gestational weeks with multiple intestinal atresia, had enteral nutrition and PN providing 67 kcal/kg on 5 days a week, and developed cholestatic liver disease. Lipovenss was administered at an average dose of 1.2 g/kg/day. PN was required due to functional intestinal obstruction, limiting enteral intake. A liver biopsy showed periportal fibrosis, accumulation of bile pigments within the small bile ducts and giant cell hepatitis. Bilirubin serum level was 6.5 mg/dL. No biliary excretion was found by Tc-Iod-iminodiacetic acid-scintigraphy. There were normal serum levels of alpha-1-antitrypsin, transferrin and iron, a normal sweat test, and negative cultures for cytomegalovirus in urine and enterovirus in stool. Serologic findings were negative for CMV, Ebstein-Barr virus, herpes simplex virus, varizella, HHV-6. Adenovirus and Enterovirus. Improvement in intestinal propulsive function allowed abrupt withdrawal of PN and liver function tests normalized within 6 weeks. These observations finally confirmed the suspected diagnosis of PNAC. Sixteen days after cessation of PN, sitosterol levels had only declined from 8.8 mg/dL to 5.7 mg/dL, whereas liver function tests were not yet improved. Bilirubin levels were 3.9 mg/dL and 5.3 mg/dL, respectively, AST rose from 48 to 94 U/L at the same time (Fig. 2). DISCUSSION The pathogenesis of PNAC is still unclear. Recently it has been hypothesized that accumulation of plant sterols may be one important factor. However, PNAC has been reported several years before the introduction of lipid emulsions into parenteral nutrition (6), indicating that factors other than lipids might be relevant in the pathogenesis of PNAC. To date, there is no defined toxic level of plant sterols in PN. Exact kinetic data are lacking. In patients receiving PN for more than 1 week and with normal liver function tests, we never observed sitosterol levels exceeding 100 μmol/L (4 mg/dL). We observed a linear correlation between intravenous lipid dosage and plasma levels of sitosterol (Fig. 3). In adults, the daily infusion of 500 mL of Intralipid 20% on 3 consecutive has been reported to produce a rise in plasma phytosterol levels, which returned to baseline with a half-life of 5 days (6). Because phytosterols are exclusively eliminated in bile, plasma levels represent the equilibrium of intravenous input and biliary excretion. Our observations in two patients with cholestasis (patients 4 and 5) show that the half life of phytosterols can be prolonged to several weeks (Fig. 2). The fact, that sitosterol levels in patient 5 were not extremely elevated may be explained by the short history of cholestasis at the time of cessation of PN and by a relatively low lipid dose (1.2 g/kg/day). Since the plasma levels of phytosterols are not only dependent from biliary excretion but also from the administered dose, a close correlation between GGT, bile acids or bilirubin and phytosterol levels is not very likely and has not been tested for. Regression analysis would require stratification for various dosages. Our observations show, that high levels of plant sterols are an obligatory finding in patients with PNAC or any other cholestatic liver disease, receiving intravenous fat emulsions. The observation of high levels of plant sterols in patients with PNAC therefore does not allow the conclusion of a causal relationship. The plasma levels of sitosterol (4.9–11.4 mg/dL; 118–275 μmol/L) observed in our anicteric patients with PN-associated steatosis are comparable to those of the patients with “mild liver dysfunction” (6 + 1.6 mg/dL; 152 + 41 μmol/L) reported by Clayton et al. (2) in 1993. We continued the infusion of intravenous lipids, thus providing additional amounts of phytosterols, comparable to those resulting in reduced bile flow in the animal model published by Iyer et al (3). If accumulation of PS exerted a toxic effect, the continued infusion of fat emulsions should induce or worsen preexisting liver dysfunction and cholestasis in patients with steatosis and slightly impaired ability to excrete plant sterols. However, in our three patients with abnormal liver function tests due to steatosis, improvement of liver function was achieved with a reduction in glucose infusion rate, while the dosage of intravenous lipids was maintained. We conclude that elevated levels of plant sterols in patients on PN are a symptom of liver dysfunction. The plasma levels depend on the degree of cholestasis and on the dosage of intravenous lipid emulsions. These data do not exclude the possibility that elevated levels of PS may aggravate overt PNAC or that plant sterols may contribute to the development of PNAC in susceptible individuals. In patients with PN-associated steatosis, the withdrawal of intravenous fat emulsions at a usual dosage cannot be recommended.