Secondary Logo

Journal Logo

Infection and Cholestasis in Neonates with Intestinal Resection and Long-Term Parenteral Nutrition

Sondheimer, Judith M.; Asturias, Edwin; Cadnapaphornchai, Melissa

Journal of Pediatric Gastroenterology & Nutrition: August 1998 - Volume 27 - Issue 2 - p 131-137
Original Articles

Objectives: This retrospective study was conducted to determine the incidence of cholestasis and liver failure in patients with intestinal resection in the neonatal period who subsequently become dependent on parenteral nutrition support and to assess the significance of associated clinical factors-gestational age, birth weight and length; length of bowel resected; presence of ileocecal valve; enteral feeding history; and infection to the incidence and severity of cholestasis.

Methods: Retrospective chart review of all patients in a single institution from May 1984 to February 1997 with neonatal small intestinal resection dependent on parenteral nutrition for at least 3 months.

Results: Forty-two patients fitting the inclusion criteria were the subjects of this review. Cholestasis developed in 28 (67%) while they were receiving parenteral nutrition (direct serum bilirubin more than 2mg/dl). In 21, the elevated direct bilirubin normalized while patients continued to receive parenteral nutrition. Seven patients progressed to liver failure. In 14 patients, serum direct bilirubin never rose above 2 mg/dl. The cholestatic patients did not differ from the noncholestatic in gestational age, birth weight, and length; primary diagnosis; length of bowel resected; or presence of ileocecal valve. The duration of dependence on parenteral nutrition was longer in noncholestatic (33.2 ± 9 months) than in cholestatic patients progressing to liver failure (19.4 ± 3 months) or in cholestatic patients who recovered (16.1 ± 1.9 months) (p < 0.05). Invasive fungal or bacterial infections occurred in all but one noncholestatic patient. The number of infections per patient was similar in all groups. The mean age (days) at first infection was significantly younger in cholestatic patients progressing to liver failure (28.5 ± 5) and cholestatic patients who recovered (48.2 ± 14.2) than in noncholestatic patients(167 ± 43.2) (p 0.01). Infection preceded the onset of cholestasis in all but 3 patients by an average of 13.5 days. Infecting organisms and site of first infection were similar in all patients.

Conclusions: Cholestasis is common in infants with neonatal intestinal resection. Liver failure develops in 16.6%. Bacterial infection early in life characterized the cholestatic patients, and cholestasis developed shortly after the first infection in 90% of patients.

Department of Pediatrics, Section of Gastroenterology, and Hepatology and Nutrition, University of Colorado Health Sciences Center and the Children's Hospital, Denver, Colorado, U.S.A.

Address correspondence and reprint requests to Dr. J. M. Sondheimer, Gastroenterology, Hepatology and Nutrition, The Children's Hospital, 1056 East 19th Avenue, Denver, CO 80218, U.S.A.

Submitted August 25, 1997; revised November 18, 1997; accepted December 31, 1997.

Presented in part at the annual meeting of the American Gastroenterological Association, San Francisco, May 1996.

Hepatic failure develops in 3% to 19% of children with short bowel syndrome acquired in the neonatal period, making it a leading cause of death in these patients (1-3). Cholestasis, which is thought to be the precursor of liver failure, is reported to occur in 30% to 60% of these patients (2,3). It is usually assumed that cholestasis in patients with short bowel is a result of chronic exposure to intravenous nutrition solutions (4-7). Data from animal studies support the view that intravenous nutrition solutions are hepatotoxic (8-10). However, liver failure has never been produced in experimental animals using intravenous nutrition alone.

In our patient population with neonatal intestinal resection, several observations suggested that exposure to intravenous nutrition was not a sufficient explanation for the development of cholestasis and liver failure. First, although all patients received intravenous nutrition, cholestasis did not develop in all. Second, there was no direct correlation between length of exposure to or total amount of intravenous nutrition and the prevalence or severity of cholestasis. To the contrary, we observed that cholestasis developed early in the course of intravenous nutrition therapy, usually resolved spontaneously, and did not recur despite continued use of intravenous nutrition. Finally, we noted that the onset of cholestasis in many patients seemed to follow directly on the heels of the first episode of bacterial or fungal infection.

We reviewed the records of all 45 patients in our center who had undergone intestinal resections as neonates from May 1984 to February 1997 and who were partially or totally dependent on intravenous nutrition for at least 3 months, to identify the factors in addition to intravenous nutrition that might correlate with the incidence and severity of cholestatic liver disease.

Back to Top | Article Outline


From May 1984 to February 1997, the six pediatric gastroenterologists at the University of Colorado Health Sciences Center managed the care of 45 infants with small bowel resections in the first 30 days of life who required intravenous nutritional support for at least 3 months. One patient was lost to follow-up while still receiving parenteral nutrition. Complete medical records from birth were not available in 2 patients born outside the Denver metro area. Therefore, 42 patients are the subjects of this evaluation. In 30 of the 42 patients, more than 50% of the estimated small bowel length was resected or atretic. The 12 remaining patients with lesser resections(ranging from 10% to 50% of estimated small bowel length) who required intravenous nutrition for more than 3 months were included, because there was a significant incidence of cholestasis among them.

Twenty-four of the 42 patients were born in eight referral hospitals in the metropolitan Denver area. Seven other patients were born at University Hospital in Denver, 10 were born in hospitals in the Rocky Mountain region, and 1 was born in Chile. All 35 out born patients and 1 patient born at University Hospital were transferred to Children's Hospital in the neonatal period. In 1990, after the affiliation of the pediatric services of Children's Hospital and University Hospital, all patients still receiving care at University Hospital were transferred to the nursery at Children's Hospital.

The medical records of the 42 patients were reviewed to obtain date of birth, gestational age, birth weight and height, and initial diagnosis resulting in bowel resection. We assessed the length of bowel remaining after initial surgery and the presence or absence of the ileocecal valve and colon. It was clear from operative notes in all patients whether the ileocecal valve had been removed. Unfortunately, there was no standard manner by which the length of remaining small bowel was estimated. When a massive resection was performed, the surgeon measured the remaining bowel along the antimesenteric border. Thus, the most accurate measurements of residual small bowel were obtained in the patients with the most extensive resections. When shorter resections were performed, surgeons generally measured the length of bowel removed and estimated the length or percent of small bowel remaining. Pathology reports frequently included a measurement of the length of resected bowel. However, because resected bowel was often damaged, atretic or in preservative, pathologist's measurements of resected bowel were not used. We used a standard reference for small bowel length relative to gestational age(11) to estimate the length in centimeters of small bowel in situ if the surgical report indicated only the length of small bowel resected or the percentage of small bowel remaining.

Cholestasis was defined as conjugated serum bilirubin concentration above 2 mg/dl. In all cases, elevations of direct bilirubin were accompanied by elevations in serum alkaline phosphatase and alanine aminotransferase. Levels of γ-glutamyltranspeptidase and aspartate aminotransferase were not regularly measured in all patients. We recorded the age (in days) at which serum conjugated bilirubin above 2 mg/dl was first detected, the maximum conjugated bilirubin level attained, and the age at which the maximum level occurred. The age at which the conjugated bilirubin returned to less than 2 mg/dl was recorded if applicable. Evidence of cholelithiasis was extracted from surgical and radiologic records. We recorded the age at which enteral feedings were introduced. We calculated the percentage of life that the patient had received enteral feedings at the onset of cholestasis by counting the number of days the child had received formula as a percentage of total days of life. We also assessed the percentage of daily energy intake supplied by the enteral route at the onset of cholestasis and at an arbitrarily selected age of 12 weeks.

Infections occurring while patients were receiving intravenous nutrition were identified by chart review, including microbiology reports, nursing notes, and physician progress notes. Isolation of bacterial or fungal pathogens from blood, urine, stool, peritoneal fluid, cellulitis aspirate, or cerebrospinal fluid were considered evidence of invasive infection. We also documented all stool and respiratory pathogens, including those detected by scanning electron microscope (stool). The dates of positive cultures were noted.

Although the chemical composition of intravenous nutrition solutions changed, the nutritional goals of enteral and parenteral nutrition in patients with intestinal resection remained relatively stable during the study. The goal for total daily non-protein energy intake was 377 to 502 kJ/kg (90 to 120 kcal) per day, adjusted to achieve the desired rate of weight gain. The goal for carbohydrate intake was 60% to 70% of total energy intake with an intravenous dextrose concentration limit of 25 g/dl. The goal for total protein intake in patients under 6 months of age was 2 to 3 g/kg per day. When cholestasis was detected, the average daily intake of intravenous amino acids was decreased by alternating amino acid intakes of 1 to 1.5 and 2 to 2.5 g/kg per day. Copper and manganese were reduced or eliminated at the onset of cholestasis with further changes in supplementation dictated by serum levels. Crystalline amino acid solutions were used at a maximum concentration of 2.5 g/dl. The amino acid solutions used from 1984 to 1988 were Travasol (Baxter/Clintec, Glendale, CA, U.S.A.), Novamine(Pharmacia, Clayton, NC, U.S.A.), Aminosyn (Abbott Laboratories, Abbott Park, IL, U.S.A.) or Trophamine (McGaw, Irvine, CA, U.S.A.) with supplemental cysteine HCl. Since 1988, all patients were given Trophamine with supplemental cysteine HCl. The intravenous lipid source was soybean oil (Intralipid, Baxter/Clintec, Deerfield, IL, U.S.A.) in either 10% or 20% solution. The goal for lipid administration was 30% to 40% of total daily energy intake with a maximum of 3 g/kg per day, depending on the lipid tolerance determined by serum triglyceride levels. Administration of intravenous nutrition was timed to allow infants some time free of the intravenous catheter each day. Enteral feedings were instituted at the earliest possible time and were advanced aggressively to the limit of the patient's tolerance, defined by such parameters as diarrhea, hematochezia, emesis, and lactic acidemia. All but three patients received most of their enteral feedings in the first 6 months by gastrostomy drip. Protein hydrolysate formula (Pregestimil, Mead Johnson and Company, Evansville, IN, U.S.A.) or chemically defined elemental diets(3232A, Ross, Columbus, OH, U.S.A.; Tolerex, Vivonex, Vivonex-Pediatric, Sandoz Nutrition Corp., East Hanover, NJ, U.S.A.) were used in all patients with small amounts of long chain fats added to enteral formulas if needed to prevent essential fatty acid deficiency. No supplemental glutamine or fiber was used in these patients.

Biochemical monitoring of hepatic function and/or toxicity in patients receiving parenteral nutrition have generally become less frequent and less detailed since 1984. Our present protocol for monitoring hepatic toxicity in infants on parenteral nutrition includes evaluation of serum total protein and albumin, serum total and direct bilirubin, alanine and aspartate amino transferases, γ-glutamyltranspeptidase and alkaline phosphatase every month. Between 1984 and 1990, liver functions were performed weekly in most hospitalized infants on intravenous nutrition. More frequent or detailed testing was performed in patients in whom jaundice developed or in those who had biochemical abnormalities detected on routine screening. Less frequent monitoring was used in outpatients as the ratio of enteral to total daily energy intake increased.

The protocol for evaluation of the febrile infant with a central venous catheter was to obtain a blood culture through the intravenous line and a urine culture. Other cultures were obtained, depending on the judgment of the attending physician. Broad spectrum antibacterial agents were given through the central venous catheter in all febrile patients for 48 hours. Antibiotics were then modified or discontinued depending on the results of initial cultures or clinical complications such as continued fever or disseminated intravascular coagulation. There was no routine screening for intestinal bacterial overgrowth. In the presence of such clinical indicators as distension, vomiting, diarrhea, poor weight gain, lactic acidosis or cholestasis, patients received either gentamycin or metronidazole by mouth. Only 4 patients were treated for more than 1 month with rotating oral antibiotics. Urso deoxycholate was occasionally administered to patients after the onset of cholestasis. However, it uniformly caused increased diarrhea and was not used as a routine medication.

The 42 patients were divided into two groups-those developing cholestasis(direct bilirubin concentration 2 mg/dl or more) at any point while receiving intravenous nutrition (Chol), and those whose conjugated bilirubin never reached 2 mg/dl (Non-Chol). We subdivided the cholestatic group into those whose cholestasis resolved while receiving intravenous nutrition (Chol-R) and those in whom liver failure developed (Chol-LF). The criteria defining liver failure were progressive cholestatic jaundice, coagulopathy not correctable with parenteral vitamin K, and increased serum ammonia.

Mean values ± SEM are shown in text and tables. Student'st-test, analysis of variance and chi-square tests were used for statistical analysis of data.

Back to Top | Article Outline


Cholestasis developed in 28 of the 42 patients (Chol) while they were receiving intravenous nutrition (67%). Liver failure developed in 7 of the cholestatic patients (16.6% of the total; Chol-LF). Four died (at 9, 14, 19, and 24 months of age) and three are listed for combined liver-intestinal transplant (present ages 10, 17, and 20 months). In the remaining 21 patients, cholestasis resolved (Chol-R) while patients continued to receive intravenous nutrition. There were 14 patients (33%) who never had cholestasis(Non-Chol). The gestational age, birth weight, birth length, cause of neonatal intestinal resection, and duration of intravenous nutrition in the three groups are listed in Table 1. There were no significant differences in birth characteristics between the groups with and without cholestasis. The duration of intravenous nutrition was significantly shorter in both groups of cholestatic patients when compared with those in whom cholestasis never developed. The birth dates of the Non-Chol patients ranged from May 1984 to May 1995 (median January 1988). The birth dates of the Chol-R patients ranged from August 1984 to March 1995 (median October 1991). The birth dates of the Chol-LF patients ranged from January 1992 to April 1996 (median September 1994).



The surgical characteristics and cholestasis histories of the Chol and Non-Chol groups are summarized in Table 2. There was no difference between groups in presence of the ileocecal valve. Although the mean length of residual small bowel did not differ among the groups, five of the seven Chol-LF patients had had massive small bowel resections with an estimated 9 to 51 cm of small bowel remaining (mean, 27.2 ± 7.1 cm). The remaining two Chol-LF patients had lesser resections with 154 cm and 70 cm of small bowel remaining. The former patient had gastroschisis with profound dysmotility and the latter had necrotizing enterocolitis (NEC) with no motility disorder.



The mean age at the onset of cholestasis was 38.6 ± 6.9 days in the Chol-R patients and 55.4 ± 10.0 days in the Chol-LF group (p< 0.1). By linear regression analysis, no significant correlation(R2 = 0.18) between the age of onset of cholestasis and the residual bowel length was determined (data not shown). The mean age at which the maximum conjugated bilirubin occurred in Chol-R patients was 100.7± 17.1 days and the mean age at which conjugated bilirubin normalized was 155.1 ± 21.5 days. In all Chol-R patients, the serum bilirubin normalized while the patient continued to receive intravenous nutrition. The mean maximum conjugated bilirubin level in the Chol-R patients was 7.9± 1.7 mg/dl (range, 2.2-31 mg/dl; median, 5.1 mg/dl). The mean maximum conjugated bilirubin attained by the Chol-LF group was 18.4 ± 1.3 mg/dl (range, 15.1-25 mg/dl; median, 18.6 mg/dl). As expected, this mean was significantly higher than that of the Chol-R group (p< 0.01). The maximum conjugated bilirubin in the Chol-LF group occurred just prior to death in the four patients dying of liver failure, and at the latest serum sample in the three living patients. The mean age at maximum conjugated bilirubin was significantly greater in Chol-LF than in Chol-R patients (p < 0.01). The incidence of cholelithiasis was similar in all three groups.

Enteral feedings were introduced at a mean age of 29.7 ± 9.2 days in the Non-Chol patients, at 27.3 ± 11.4 days in the Chol-R group, and at 32.7 ± 9.8 days in the Chol-LF group (p < 0.1). At the onset of cholestasis, the Chol-R group had received enteral formula feedings for 30 ± 6.4% of the days of life. Similarly, in the Chol-LF group, enteral feedings had been administered for 37.5 ± 13.2% of the days of life at the onset of cholestasis (p > 0.1). At the onset of cholestasis, the percentage of total daily energy intake received enterally was low in both cholestatic groups (14.2 ± 5.1% in Chol-R and 8.9 ± 6.4% in Chol-LF) (p > 0.1). Taking an arbitrary age of 12 weeks, we assessed the percentage of daily energy intake that was administered enterally in all three study groups. In the Non-Chol and Chol-R groups, the enteral intakes were 32.9 ± 8.3% and 35.9± 7.2% of the total respectively, whereas the enteral intake was 13.3± 5.5% of total in the Chol-LF group (p < 0.01).

The infection history of the 3 groups is summarized inTable 3. Invasive bacterial or fungal infection occurred in all Chol patients. Blood, urine, peritoneal cavity, and abdominal wall were the sites of first infection, with blood the most common. The mean age at first infection in the Chol-R group was 48.2 ± 14.2 days, and in the Chol-LF group 28.5 ± 5 days (p > .1). Bacterial or fungal infection occurred in 13 of the 14 Non-Chol patients. A positive blood culture was the first infection in all Non-Chol patients. The mean age at which the first infection occurred in this group was 167 ± 43 days, significantly greater than the mean age of first infection in Chol-R and Chol-LF groups (p < 0.001). (If the total "infection free" time on intravenous nutrition of the single Non-Chol patient with no episodes of infection is included in the infection-free mean, the mean age of first infection increases to 182 ± 42 days.) The mean number of bacterial or yeast infections occurring in each patient during the course of intravenous nutrition was similar in all three groups, ranging from 3.8 episodes per patient in Non-Chol patients to 4.9 per patient in the Chol-R and Chol-LF patients. Episodes of infection per month of parenteral nutrition were fewer in the Non-Chol patients, probably because of the higher average duration of parenteral nutrition in this group. In the Chol-LF group, 61.7% of all infections occurred in the first 6 months of life. In the Chol-R group, 41.3% of infections occurred in the first 6 months, whereas in the Non-Chol group, only 30.1% of infections occurred in the first 6 months of life.



The onset of cholestasis followed the first infection by a mean of 13.5± 3.4 days in 25 of the 28 cholestatic patients. In the three remaining cholestatic patients, all in the Chol-R group, conjugated hyperbilirubinemia preceded the first infection by 151, 165, and 169 days. The first episode of infection was a positive blood culture in all three. Review of these three patient records revealed that two had acute superior vena cava obstruction with congestive hepatopathy coincident with the onset of cholestasis. No other patient in this study experienced superior vena cava obstruction. In the third patient in whom cholestasis preceded the first infection, no particular clinical event appeared to precede the onset of cholestasis. This patient had profound dysmotility of the upper small bowel secondary to gastroschisis and intestinal atresia.

No viral infections were documented in any of the Chol patients before the onset of cholestasis. After the onset of cholestasis, viral pathogens were seen in the stool of seven Chol patients by scanning electron microscope(rotavirus in six, enteric adenovirus in one). Respiratory syncytial virus pneumonia and Influenza type B septicemia occurred in one patient each. Only the systemic Influenza B infection, which occurred in a Chol-LF patient, was associated with an exacerbation of cholestatic jaundice. Viral pathogens were also found in the stool of eight Non-Chol patients (rotavirus in six, enteric adenovirus in one, and small round virus in one). Salmonella was cultured from the stool of one Chol patient (after the onset of cholestasis) and in two Non-Chol patients. Clostridium difficile toxin was detected in the stools of four Chol and one Non-Chol patient. No positive results in tests for hepatitis B surface antigen, or hepatitis C antibodies were obtained. No hepatitis C RNA was visualized in polymerase chain reaction.

Back to Top | Article Outline


The cause of cholestasis in patients with neonatal intestinal resection who require long-term intravenous nutrition is not known. It has been a convenient shorthand to refer to this common complication as "TPN cholestasis" (total parenteral nutrition cholestasis) although it is obvious that the term does not fully define the source of the problem. Some factor related to the intestinal resection itself must also be of importance. In adults, there is a fivefold increase in the risk of cholestasis in patients with small bowel resection receiving intravenous nutrition compared with that of patients with intact intestines on intravenous nutrition(12). Although transient abnormalities of transaminases, alkaline phosphatase, and γ-glutamyltranspeptidase are common in adults receiving intravenous nutrition, cholestatic jaundice is rare and is associated with very long term intravenous nutrition in combination with near-total small bowel resection (13). Results of a recent study of 16 adults with extensive small bowel resection showed that cholestatic liver disease developed in 25% of patients 46 to 88 months after the initiation of intravenous nutrition. The entire small bowel had been resected from the ligament of Treitz to the midcolon in all patients who had cholestatic jaundice (14).

In 67% (28 patients) of our patients with neonatal intestinal resection, conjugated hypeprbilirubinemia developed. In 7 patients, cholestasis progressed to overt liver failure (16.6% of the total group). In contrast to adults however, cholestasis and liver failure did not correlate with the duration of intravenous nutrition. In fact, the onset of cholestasis occurred early in life (between 6 and 123 days of age-median 32 days). There was no statistical difference between the Chol and Non-Chol patients in length of residual small bowel nor was there a statistically significant linear correlation between the onset of cholestasis and residual bowel length in the two Chol groups. It should be noted, however, that 5 of the 7 patients with liver failure had very short residual intestinal length.

We noted an increase in gastroschisis among the patients with cholestasis(8 of 28 Chol vs. 2 of 14 Non-Chol patients). Although this partition did not reach statistical significance, the clustering of patients with gastroschisis in the cholestatic groups suggests a link between intestinal stasis and liver disease. After intestinal resection, some degree of dilation and slowed transit develops in the residual small bowel, which is thought to be an adaptive response to improve absorption. Patients with gastrochisis have the highest incidence of intestinal dilation and dysmotility as a consequence of antenatal obstruction of the upper small bowel. Rather than a positive adaptive response, bowel dilation and slow transit may be a pathologic response, increasing the risk of hepatic exposure to enterotoxins, translocated bacteria, and endotoxins. Recent experiments in rats have shown that both lipopolysaccharide endotoxin, from Gram-negative organisms, and peptidoglycan-polysaccharide endotoxin, from Gram-negative and Gram-positive organisms, promote the release of tumor necrosis factor-α and interleukin-1 from hepatic Kupffer cells, causing acute and chronic hepatic inflammation and fibrosis (15). Thus, Gram-negative or Gram-positive flora, in the gut or in the systemic circulation, may be the source of toxins causing cholestasis in infants with short bowel. It is tempting to speculate that protection of the liver from toxins may require antioxidant function, which is probably depleted by the administration of intravenous nutrition, especially in newborn infants(16).

We cannot explain the apparent increase in frequency of cholestasis and liver failure in the patients born in the latter half of the survey. We do not think that changes in antibiotic usage or changes in the chemical composition of parenteral nutrition solutions that occurred during the study are responsible, Indeed, product development, both in the area of parenteral nutrition solutions and antibiotics has been focused on the production of products with less cholestatic potential. We can only speculate that the increased number of secondary and tertiary care nurseries in our referral area and the increased availability of high-quality home care services over time have resulted in the transfer of a potentially sicker population of patients at higher risk of cholestasis, with more straightforward patients remaining at their birth hospital.

The role of enteral feedings in promoting or in protecting against cholestasis has not been clarified. It has been suggested that lack of enteral feedings promotes cholestasis by compromising hepatic and gall bladder bile flow (17). There has been some encouraging experience with the use of secretogogues such as cholecystokinin and ursodeoxycholic acid in the prevention and treatment of cholestasis in infants and children on intravenous nutrition (18-21). Conversely, enteral feedings may actually facilitate bacterial translocation into the portal circulation, thereby possibly increasing the risk of hepatic endotoxin exposure and injury (22,23). In the current study, we found no statistical difference in the age at which enteral feedings were initiated between Chol and Non-Chol groups. Both Chol-R and Chol-LF groups had some formula feedings administered for approximately one third of their lives at the onset of cholestasis. Neither group was receiving more than 15% of daily energy intake by the enteral route at the onset of cholestasis, however. It is clear when the infants with and without cholestasis were compared at an arbitrary time point early in life that the Non-Chol and the Chol-R groups were receiving significantly more of their daily energy requirements by the enteral route than was the Chol-LF group. These findings suggest that a reduced proportion of daily energy given enterally may be associated with a poor prognosis for the progression of cholestasis. However, more detailed, prospective studies will be needed to evaluate the role of enteral feedings-their quantity and quality-in the prevention or promotion of cholestasis in infants with intestinal resection.

The association of cholestasis with bacterial infection has long been recognized in young infants (24). Cholestasis is most commonly reported in Gram-negative infections (particularly of the urinary tract), or in pneumococcal and Candida infections. The possibility that infection, not parenteral nutrition, was the important factor in producing cholestasis in sick neonates receiving intravenous nutrition was first suggested by Manginello and Javitt in 1979 (25). Wolf and Pohlandt recently evaluated the incidence of cholestasis in a large number of neonates receiving short term parenteral nutrition. The incidence of cholestasis in 92 patients who did not have infection was zero. The incidence of cholestasis was 26% in 152 patients who had infection while receiving parenteral nutrition (26). In our patients, we were impressed by the close association between invasive bacterial or fungal infection and the onset of cholestasis. In only 3 of the 28 cholestatic patients did cholestasis occur before the first infection. The first infection occurred significantly earlier in the cholestatic patients than it did in the Non-Chol patients. Although the mean number of infections per patient did not differ between the groups, the Chol groups experienced more of their total infections in the first 6 months of life than did the Non-Chol patients. The organisms cultured during the initial infection were a mixed group of Gram-negative, Gram-positive and fungal infections and did not differ significantly between the Chol and Non-Chol groups. Once established, cholestasis did not resolve with treatment of infection. Rather, the bilirubin rose steadily despite treatment of infection, reached a peak, and gradually resolved or rose progressively until overt liver failure developed. We could not identify any event that predicted the timing of peak bilirubin or the rapidity of its resolution. The results of this study show that cholestasis is not simply a function of the duration of exposure to intravenous nutrition. The data suggest that infection early in life, when the developing liver may be uniquely sensitive to cholestatic injury and stressed by the administration of intravenous nutrition may play an important role in the cholestasis of neonates with intestinal resection.

Back to Top | Article Outline


1. Cooper A, Floyd TF, Ross AJ. Morbidity and mortality of short bowel syndrome acquired in infancy: An update. J Pediatr Surg 1984; 19:711-7.
2. Caniano DA, Starr J, Ginn-Pease ME. Extensive short-bowel syndrome in neonates: Outcome in the 1980s. Surgery 1898;105:119-24.
3. Galea MH, Holliday H, Carachi R., Kapila L. Short-bowel syndrome: A collective review. J Pediatr Surg 1992;27:592-6.
4. Mullick FG, Moran CA, Ishak KG. TPN: A histopathologic analysis of the liver changes in 20 children. Mod Pathol 1994;7:190-4.
5. Cohen C, Olsen MM. Pediatric total parenteral nutrition - Liver histopathology. Arch Pathol Lab Med 1981;105:152-6.
6. Benjamin DR. Hepatobiliary dysfunction in infants and children associated with long term TPN. A clinico-pathologic study. Am J Clin Pathol 1981;76:276-83.
7. Vileisis RA, Inwood RJ, Hunt CE. Prospective controlled study of parenteral nutrition associated cholestatic jaundice: Effect of protein intake. J Pediatr 1980;96:893-7.
8. Bucuvalas JC, Goodrich AL, Blitzer BL, Suchy FJ. Amino acids are potent inhibitors of bile acid uptake by liver plasma membrane vesicles isolated from suckling rats. Pediatr Res 1985;19:1298-304.
9. Belli DC, Fournier LA, Lepage G, et al. Total parenteral nutrition-associated cholestasis in rats: Comparison of different amino acid mixtures. JPEN J Parenter Enteral Nutr 1986;11:67-73.
10. Narkewicz MR, Jones G. Cysteine supplementation and reduction of TPN-induced hepatic lipid accumulation in the weanling rat.J Pediatr Gastroenterol Nutr 1985;21:18-24.
11. Touloukian RJ, Walker-Smith GJ. Normal intestinal length in preterm infants. J Pediatr Surg 1983;18:720-23.
12. Messing B, Zarka Y, Lemann M, Iglicki F, Coffin B, Rambaud J. Chronic cholestasis associated with long-term parenteral nutrition. Transplantation Proceedings 1994;26:1438-9.
13. Stanko RT, Nathan G, Mendelow H Adibi SA. Development of hepatic cholestasis and fibrosis in patients with massive loss of intestine supported by prolonged parenteral nutrition.Gastroenterology 1987;92:197-202.
14. Ito Y, Shils ME. Liver dysfunction associated with long-term total parenteral nutrition in patients with massive bowel resection. JPEN J Parenter Enteral Nutr 1991;15:271-6.
15. Lichtman SN, Wang J, Schwab JH, Lemasters JJ. Comparison of peptidoglycan-polysaccharide and lipopolysaccharide stimulation of kupffer cells to produce tumor necrosis factor and interleukin-1.Hepatology 1994;19:1013-22.
16. Sokol RJ, Taylor SF, Devereaux MW, et al. Hepatic oxidant injury and glutathione depletion during total parenteral nutrition in weanling rats. Am J Physiol 1996;270:G691-700.
17. Quigley EMM, Marsh MN, Shaffer JL, Markin RS. Hepatobiliary complications of total parenteral nutrition.Gastroenterology 1993;104:286-301.
18. Spagnuolo MI, Iorio R, Vegnente A, Guarino A. Urosdeoxycholic acid for treatment of cholestasis in children on long-term total parenteral nutrition: A pilot study. Gastroenterology 1996;111:716-9.
19. Sitzmann JV, Pitt HA, Steinborn PA, Pashaz R, Sanders RC. Cholecystokinin prevents parenteral nutrition induced biliary sludge in humans. Surg Gynecol Obstet 1991;170:25-31.
20. Teitlebaum DH, Han-Markey T, Schumacher RE. Treatment of parenteral nutrition associated cholestasis with cholecystokininoctapeptide. J Pediatr Surg 1995;30:1082-5.
21. Rintala RJ, Lindahl H, Pohjavuori M. Total parenteral nutrition associated cholestasis in surgical neonates may be reversed by intravenous cholecystokinin: A preliminary report. J Pediatr Surg 1995;30:827-30.
22. Steinwender G, Schimp G, Sixl B, et al. Effect of early nutritional deprivation and diet on translocation of bacteria from the gastrointestinal tract in the newborn rat. Pediatr Res 1996;39:415-20.
23. Alverdy JC, Aoys E, Moss GS. TPN promotes bacterial translocation from the gut. Surgery 1988;104:185-90.
24. Rooney JC, Hill DJ, Danks DM. Jaundice associated with bacterial infection in the newborn. Am J Dis Child 1971;122:39-41.
25. Manginello FP, Javitt NB. Parenteral nutrition and neonatal cholestasis. J Pediatr 1979;94:296-8.
26. Wolf A, Pohlandt F. Bacterial infection: The main cause of acute cholestasis in newborn infants receiving short term parenteral nutrition. J Pediatr Gastroenterol Nutr 1989;8:297-303.

Cholestasis; Infection; Liver failure; Neonate; Short bowel syndrome

© Lippincott-Raven Publishers