- Email: [email protected]
Parenteral Nutrition-Associated Cholestasis in Small for Gestational Age Infants
Parenteral Nutrition-Associated Cholestasis in Small for Gestational Age Infants DANIEL T. ROBINSON, MD,
AND
RICHARD A. EHRENKRANZ, MD
Objective To identify small for gestational age (SGA) as an independent risk factor for parenteral nutrition-associated cholestasis (PNAC). Study design In a case-control study, records of infants treated in the neonatal intensive care unit from 1994 through 2003 with gestational ages (GA) <34 weeks and exposure to parenteral nutrition (PN) >7 days were reviewed. The primary outcome was the incidence of cholestasis in infants who were SGA. Secondary outcomes included PN duration, age at full enteral nutrition (FEN) and incidence of late-onset sepsis, necrotizing enterocolitis (NEC) and bronchopulmonary dysplasia (BPD). Analysis was by t test, logistic regression, and 2 analysis. Results Cases (n ⴝ 79) and control subjects (n ⴝ 152) had similar birth weights and GA (963 ⴞ 465 g versus 1090 ⴞ 463 g; 27 ⴞ 2 weeks versus 27 ⴞ 2 weeks; [mean ⴞ SD]). Of the infants who were SGA, 58% developed cholestasis (OR ⴝ 3.3, P < .01). Infants with cholestasis achieved FEN later (43 ⴞ 25 days versus 23 ⴞ 11 days) and had higher rates of sepsis (80% versus 34%), NEC (51% versus 7%), and BPD (65% versus 25%; P < .01). Of infants with cholestasis, infants who were SGA received fewer days of PN than infants who were appropriate for GA (49 ⴞ 24 days versus 68 ⴞ 36 days, P ⴝ .024). Conclusion Being SGA is an independent risk factor for PNAC. Infants who are SGA require less PN for cholestasis to develop. (J Pediatr 2008;152:59-62) ost preterm neonates receive parenteral nutrition (PN) as their initial nutritional support while they transition to enteral feedings. Although the mechanism is unclear, exposure to PN is one major factor in the development of cholestasis in preterm infants. Other known contributors to PN-associated cholestasis (PNAC) include sepsis, necrotizing enterocolitis (NEC), bowel surgery, and a lack of enteral feedings.1-4 In extremely low birth weight infants (birth weight ⬍1000 g), rates of PNAC have been documented to be as high as 50%.5 Subsequent liver failure, which results from long-term PN use, is the most significant complication. There is no preventive measure for PNAC. After cholestasis is established, the only clear therapy is the provision of full enteral nutrition. Studies with ursodeoxycholic acid have shown normalization of various liver function tests, but the numbers of subjects are small and methods and monitoring are varied.6,7 Alternative lipid emulsions have been shown to improve markers of cholestasis, but are not currently approved for use in the United States.8 Therefore, as the etiology and effective therapies are pursued, it remains essential to identify risk factors for the development of PNAC. Infants who are small for gestational age (SGA), or growth restricted, are known to have a number of metabolic abnormalities related to the liver as compared with their appropriate for gestational age (AGA) counterparts. For example, hepatic redox states were altered in rats with fetal growth restriction.9 Neonatal hypoglycemia occurs because of low glycogen stores. Autopsy results found SGA infants exposed to PN likely to have more severe liver damage.10 Selective assessments of small numbers of premature infants indicated that intrauterine growth restriction was a risk From the Yale University School of Medifactor for PNAC.11,12 cine, New Haven, Connecticut. We hypothesized that being born SGA was an independent risk factor for the Dr. Robinson was supported by NICHD development of PNAC. training grant HD 07094.
M
AGA BPD FEN GA NEC
Appropriate for gestational age Bronchopulmonary dysplasia Full enteral nutrition Gestational age Necrotizing enterocolitis
NICU PN PNAC SGA
Neonatal intensive care unit Parenteral nutrition Parenteral nutrition-associated cholestatsis Small for gestional age
Submitted for publication Jan 22, 2007; last revision received Apr 23, 2007; accepted Jun 1, 2007. Reprint requests: Daniel T. Robinson, MD, PO Box 208064, New Haven, CT 065208064. E-mail: [email protected]. 0022-3476/$ - see front matter Copyright © 2008 Mosby Inc. All rights reserved. 10.1016/j.jpeds.2007.06.002
59
METHODS A retrospective case-control study design, approved by the Yale University School of Medicine Human Investigation Committee, was used. Eligible patients were 1) cared for in the Yale New Haven Children’s Hospital Newborn Special Care Unit between January 1, 1994, and December 31, 2003, 2) born at ⬍34 weeks gestational age (GA), as determined by best obstetrical estimate and physical examination, and 3) exposed to ⱖ7 days of PN. Case subjects were defined as having cholestasis (direct bilirubin ⱖ2 mg/dL); control subjects did not have cholestasis. Exclusion criteria included cholestasis attributable to a congenital infection, an anatomic obstruction of the hepatobiliary tract, an inherited metabolic disease or metabolic syndrome, and a congenital gastrointestinal disorder requiring surgery. All infants with cholestasis underwent liver ultrasound scanning and testing for infectious hepatitis to exclude anatomic obstruction and infection as the cause of cholestasis. Control subjects were matched to case subjects on the basis of GA and date of birth. The first priority match was an infant of the same GA and a date of birth within 3 months of the case subject. If no match was found, then the next search was for an infant of the same GA and a date of birth within 6 months of the case subject. Finally, when no match was found with those criteria, a control subject with a GA within 1 week and a date of birth within 3 months of the case was used. Two control subjects were sought for each case. The primary outcome was the incidence of cholestasis among SGA infants. SGA was defined as a birth weight ⬍10% for GA.13 Secondary outcomes included the day of life at which cholestasis was detected and the level of direct bilirubin at the time of diagnosis. In addition, the number of days of PN exposure before the diagnosis, the total days of PN exposure, an enteral feeding history, and the incidence of late onset sepsis (diagnosed after 72 hours of life, either culture proven or culture negative but treated with antibiotics for ⱖ7 days), NEC (by using the modified Bell’s staging criteria14), and bronchopulmonary dysplasia (BPD; O2 treatment at 36 weeks postmenstrual age) were collected. The enteral feeding history included the number of days that infants were not offered enteral feeds (NPO) immediately after birth, the number of days with feeding interruptions (patient ordered NPO ⱖ12 consecutive hours within a 24-hour period) before and after full enteral nutrition (FEN) was achieved (ⱖ100 kcal/kg/day without PN support), and the day of life on which FEN was achieved.
Nutritional Protocol During Study Period The amino acid and lipid solutions given during the study period were Trophamine (B. Braun Medical, Bethlehem, PA) and Liposyn (Hospira, Lake Forset, IL). The day of life on which PN was started changed in our unit during the decade; it was introduced between 24 and 48 hours of life in the earlier years, as compared with within the first 24 hours in the more recent years. All infants exposed in this study achieved maximum protein and lipid infusions between 3 and 60
Robinson and Ehrenkranz
3.5 g/kg per day. Lipid levels were monitored, and when evidence of lipid intolerance was detected, infusions were held and then reinstated at a lower rate until an acceptable level was obtained. Dextrose was initiated on day of life 1 to achieve a glucose infusion rate between 6 and 8 mg/kg per minute and advanced to a maximum of 15 mg/kg per minute by day of life 7. There were no alterations to the PN when cholestasis was detected (eg, trace elements were not withheld).
Statistics To detect an odds ratio of 3 for the development of cholestasis when infants were SGA, 71 cases would be needed for analysis given an estimated incidence of SGA in our population of 10%, a 2-tailed ␣ level of 0.05, and power of 80%. Statistics were calculated with the Student t test and 2 analysis for means and frequencies. Multivariate analysis was performed by using logistic regression. Calculations were made with SPSS software version 13.0 (SPSS, Chicago, IL). A P value ⬍.05 was considered to be significant.
RESULTS During the study period, 79 eligible subjects with cholestasis were identified and able to be matched to at least 1 control subject; 2 matches were available for 73 of the case subjects (Table I). Of the control subjects, 22 (14.5%) had a gestational age ⫾1 week of the case. Maternal characteristics, aside from age at delivery, including factors that may have contributed to uteroplacental insufficiency (tobacco use, cocaine use, chronic and pregnancy-induced hypertension), exposure to antenatal steroids, presence of chorioamnionitis, or delivery of infant by cesarean section, were not different in the study groups (data not shown). Except for birth weight and length of stay, infant characteristics were not significantly different in the patient groups. However, infants with cholestasis had higher incidences of comorbidities. Furthermore, in infants with culture-proven sepsis, the rates of gramnegative sepsis were higher in infants with cholestasis (45% versus 29%). Among the 40 SGA infants in the total population studied, cholestasis developed in 23 (58%). The odds ratio of developing cholestasis, when born SGA, was 3.3 (95% CI, 1.6-6.6). Neither the incidence of SGA nor that of cholestasis changed throughout the study period. Of the infants with cholestasis, infants who were SGA had significantly lower birth weights and were significantly more mature than the infants who were AGA (Table II). Rates of comorbidities were similar in the groups. Mothers of SGA infants were more likely to have had chronic hypertension and pregnancy-induced hypertension (data not shown). The day of life and postmenstrual age at which cholestasis was detected did not differ significantly between the SGA and AGA infants. There was also no difference in the value of direct bilirubin on detection of cholestasis. Liver failure that lead to referral for transplantation or to death did not develop in any infant. The Journal of Pediatrics • January 2008
Table I. Patient characteristics and nutrition history
Patient characteristics Maternal age (years) Birth weight (grams) Gestational age (weeks) Sex (male/female) Race (W/H/B/other) SGA Late-onset sepsis NEC Required surgery BPD LOS (days) Nutrition history NPO before fed (days) DOL FEN achieved Total days PN Days NPO before FEN Days NPO after FEN
Cholestasis (n ⴝ 79)
No cholestasis (n ⴝ152)
P value
Odds ratio (95% CI)
28 ⫾ 6* 963 ⫾ 465 27 ⫾ 2 49/30 37/28/10/4 23 (29)† 63 (80) 40 (51) 6 51 (65) 106 ⫾ 40
30 ⫾ 6 1090 ⫾ 463 27 ⫾ 2 80/72 84/41/20/7 17 (11) 52 (34) 11 (7) 0 38 (25) 72 ⫾ 40
⬍.01 .049 .362 .173 .453 ⬍.01 ⬍.01 ⬍.01
3.3 (1.6-6.6) 2.7 (1.2-6.0) 7.5 (3.0-18.9)
10 ⫾ 9 43 ⫾ 25 63 ⫾ 34 11 ⫾ 12 12 ⫾ 17
6⫾4 23 ⫾ 11 24 ⫾ 14 3⫾5 2⫾4
⬍.01 ⬍.01 ⬍.01 ⬍.01 ⬍.01
⬍.01 ⬍.01
2.8 (1.4-5.9)
W/H/B/other, White/Hispanic/black/other; LOS, length of stay; DOL, day of life. *Mean ⫾ SD. †n (%).
Table II. Patient characteristics and nutrition history for patients with cholestasis SGA (n ⴝ 23)
AGA (n ⴝ 56)
Patient characteristics Birth weight (grams) 707 ⫾ 221* 1068 ⫾ 498 Gestational age (weeks) 28 ⫾ 2 27 ⫾ 3 Late-onset sepsis 16 (70)† 47 (84) NEC 11 (48) 29 (52) Required surgery 1 5 BPD 12 (52) 39 (70) DOL cholestasis detected 32 ⫾ 14 34 ⫾ 22 Direct bilirubin level at 4.7 ⫾ 2.7 4.3 ⫾ 3.5 diagnosis (mg/dL) PMA at diagnosis 32 ⫾ 3 31 ⫾ 3 Nutrition history NPO before fed (days) 10 ⫾ 11 10 ⫾ 9 DOL FEN achieved 32 ⫾ 16 48 ⫾ 27 Total days PN 49 ⫾ 24 68 ⫾ 36 Days NPO before FEN 8⫾9 13 ⫾ 12 Days NPO after FEN 11 ⫾ 12 12 ⫾ 19
P value ⬍.01 .025 .149 .749 .191 .615 .681 .159 .753 ⬍.01 .024 .150 .901
PMA, Postmenstrual age. *Mean ⫾ SD. †n (%).
Infants with cholestasis had the first sustained enteral feeds introduced later than infants without cholestasis (Table I). Infants without cholestasis achieved FEN sooner. In infants in whom cholestasis developed, total exposure to PN was greater and there were more days with feeding interrup-
tions. The AGA infants achieved full feeds later than infants who were SGA, and they were exposed to PN for a longer duration (Table II). Although infants who were SGA had been exposed to PN for 4 days fewer at the time of diagnosis, this difference was not significant (data not shown).
DISCUSSION We report evidence that preterm infants who are SGA and exposed to PN for at least 7 days are at increased risk for PNAC. Although there has been much discussion of this topic, our study contains the largest sample and achieves the necessary statistical power to define SGA as an independent risk factor for PNAC. A number of findings in this study agree with observations previously reported, such as lower birth weights, higher morbidity rates, and the nutritional histories of infants with cholestasis. With sepsis-related damage to hepatocytes being associated with cytokine release triggered by lipopolysaccharides, we also analyzed the sepsis results using just those infants with proven gram-negative rod infection15,16; the incidence of gram-negative rod sepsis was higher in the group with cholestasis. Because many neonatologists handle advancing enteral nutrition in SGA infants with caution because of fears of an increase in NEC in that population,17 we anticipated the finding that infants with cholestasis had more feeding interruptions than infants without cholestasis. However, it was unexpected to find that, in patients with cholestasis, SGA infants achieved full enteral feeds at an earlier time and were exposed to less PN than AGA infants.
Parenteral Nutrition-Associated Cholestatsis in Small for Gestational Age Infants
61
Perhaps differences in hepatic metabolism attributable to impaired perfusion of the liver in utero will be identified between infants who are growth-restricted and infants who are not growth-restricted. Because the development and regulation of bile salt transporters in rats have been shown to markedly increase in the neonatal period, alterations in the transcription, translation, and function of proteins involved in bile production are likely.18 Another contributor may be altered metabolism of copper, a metal that may have dual impacts, protective as a cofactor in antioxidants but hepatotoxic when accumulated in excess.19,20 The absence of copper in postmortem livers of infants who are SGA and have PNAC suggests that insufficient antioxidant activity may be a contributing factor in the development of cholestasis in SGA infants.10 An additional area highlighted by our results is the striking difference in the incidence of BPD between infants with and infants without cholestasis. The bulk of vitamin A is stored in the liver. A significant reduction of BPD has been found in preterm infants to whom vitamin A was administered.21 A partial explanation of our findings may be that the liver damage consequent to the cholestasis decreases the bioavailability of vitamin A. The association may also be explained by the state of inflammation found in both cholestasis and BPD.22,23 This study is limited because of its retrospective design. The frequency of testing direct bilirubin levels was certainly affected by practice variation. This impairs the ability to accurately determine the time of onset of PNAC. Research on the basis of tests such as serum bile acid levels may more accurately pinpoint the time of onset of disease. Therefore, our data do not permit an accurate interpretation of the number of days of PN exposure before onset of disease and clinical detection. Because of the varied number of levels sent on each patient, patterns of bilirubin, such as timing of peak level and resolution of disease, and other variables including transaminases and coagulation profiles could not be studied for trends. A prospective study with standardized testing would help answer questions, including the number of days of exposure necessary before the development of cholestasis. Preterm infants who are SGA are an at-risk population in many respects. These data identify an increased susceptibility of cholestasis developing when exposed to PN. Because the mechanisms facilitating the development of PNAC are not fully understood, the identification of preventative and therapeutic modalities has been impaired. Investigations into the pathways involved in bile production in subjects who are growth restricted may reveal vital information about the
62
Robinson and Ehrenkranz
pathophysiology and effective therapies of this currently unavoidable complication.
REFERENCES 1. Drongowski RA, Coran AG. An analysis of factors contributing to the development of total parenteral nutrition-induced cholestasis. JPEN J Parenter Enteral Nutr 1989;13:586-9. 2. Kelly DA. Liver complications of pediatric parenteral nutrition-epidemiology. Nutrition 1998;14:153-7. 3. Kaufman SS, Gondolesi GE, Fishbein TM. Parenteral nutrition associated liver disease. Semin Neonatol 2003;8:375-81. 4. Moss RL, Das JB, Raffensperger JG. Necrotizing enterocolitis and total parenteral nutrition-associated cholestasis. Nutrition 1996;12:340-3. 5. Beale EF, Nelson RM, Bucciarelli RL, Donnelly WH, Eitzman, DV. Intrahepatic cholestasis associated with parenteral nutrition in premature infants. Pediatrics 1979;64:342-7. 6. Al-Hathlol K, Al-Madani A, Al-Saif S, Abulaimoun B, Al-Tawil K, El-Demerdash A. Ursodeoxycholic acid therapy for intractable total parenteral nutrition-associated cholestasis in surgical very low birth weight infants. Singapore Med J 2006;47:147-51. 7. Chen CY, Tsao PN, Chen HL, Chou HC, Hsieh WS, Chang MH. Ursodeoxycholic acid (UDCA) therapy in very-low-birth-weight infants with parenteral nutritionassociated cholestasis. J Pediatr 2004;145:317-21. 8. Gura KM, Duggan CP, Collier SB, Jennings RW, Folkman J, Bistrian BR, et al. Reversal of parenteral nutrition-associated liver disease in two infants with short bowel syndrome using parenteral fish oil: implications for future management. Pediatrics 2006;118:e197-201. 9. Lane RH, Flozak AS, Ogata ES, Bell GI, Simmons RA. Altered hepatic gene expression of enzymes involved in energy metabolism in the growth-retarded fetal rat. Pediatr Res 1996;39:390-4. 10. Zambrano E, El-Hennawy M, Ehrenkranz RA, Zelterman D, Reyes-Mugica M. Total parenteral nutrition induced liver pathology: an autopsy series of 24 newborn cases. Pediatr Dev Pathol 2004;7:425-32. 11. Baserga MC, Sola A. Intrauterine growth restriction impacts tolerance to total parenteral nutrition in extremely low birth weight infants. J Perinatol 2004;24:476-81. 12. Aucott SW, Donohue PK, Northington FJ. Increased morbidity in severe early intrauterine growth restriction. J Perinatol 2004;24:435-40. 13. Alexander GR, Himes JH, Kaufman RB, Mor J, Kogan M. A United States national reference for fetal growth. Obstet Gynecol 1996;87:163-8. 14. Walsh MC, Kliegman RM. Necrotizing enterocolitis: treatment based on staging criteria. Pediatr Clin North Am 1986;33:179-201. 15. Green RM, Beier D, Gollan JL. Regulation of hepatocyte bile salt transporters by endotoxin and inflammatory cytokines in rodents. Gastroenterology 1996;111:193-8. 16. Nathanson MH, Boyer JL. Mechanisms and regulation of bile secretion. Hepatology 1991;14:551-66. 17. Yu VYH, Upadhyay A. Neonatal management of the growth-restricted infant. Semin Fetal Neonatal Med 2004;9:403-9. 18. Tomer G, Ananthanarayanan M, Weymann A, Balasubramanian N, Suchy FJ. Differential developmental regulation of rat liver canalicular membrane transporters Bsep and Mrp2. Pediatr Res 2003;53:288-94. 19. Cordano A. Clinical manifestations of nutritional copper deficiency in infants and children. Am J Clin Nutr 1998;67:1012-6S. 20. Peña M, Lee J, Thiele DJ. A delicate balance: homeostatic control of copper uptake and distribution. J Nutr 1999;129:1251-60. 21. Tyson JE, Wright LL, Oh W, Kennedy KA, Mele L, Ehrenkranz RA, et al. Vitamin A supplementation for extremely-low-birth-weight infants. National Institute of Child Health and Human Development Neonatal Research Network. N Engl J Med 1999;340:1962-8. 22. Chess PR, D’Angio CT, Pryhuber GS, Maniscalco WM. Pathogenesis of bronchopulmonary dysplasia. Semin Perinatol 2006;30:171-8. 23. Jobe AH, Bancalari E. Bronchopulmonary dysplasia. Am J Respir Crit Care Med 2001;163:1723-9.
The Journal of Pediatrics • January 2008
AND
RICHARD A. EHRENKRANZ, MD
Objective To identify small for gestational age (SGA) as an independent risk factor for parenteral nutrition-associated cholestasis (PNAC). Study design In a case-control study, records of infants treated in the neonatal intensive care unit from 1994 through 2003 with gestational ages (GA) <34 weeks and exposure to parenteral nutrition (PN) >7 days were reviewed. The primary outcome was the incidence of cholestasis in infants who were SGA. Secondary outcomes included PN duration, age at full enteral nutrition (FEN) and incidence of late-onset sepsis, necrotizing enterocolitis (NEC) and bronchopulmonary dysplasia (BPD). Analysis was by t test, logistic regression, and 2 analysis. Results Cases (n ⴝ 79) and control subjects (n ⴝ 152) had similar birth weights and GA (963 ⴞ 465 g versus 1090 ⴞ 463 g; 27 ⴞ 2 weeks versus 27 ⴞ 2 weeks; [mean ⴞ SD]). Of the infants who were SGA, 58% developed cholestasis (OR ⴝ 3.3, P < .01). Infants with cholestasis achieved FEN later (43 ⴞ 25 days versus 23 ⴞ 11 days) and had higher rates of sepsis (80% versus 34%), NEC (51% versus 7%), and BPD (65% versus 25%; P < .01). Of infants with cholestasis, infants who were SGA received fewer days of PN than infants who were appropriate for GA (49 ⴞ 24 days versus 68 ⴞ 36 days, P ⴝ .024). Conclusion Being SGA is an independent risk factor for PNAC. Infants who are SGA require less PN for cholestasis to develop. (J Pediatr 2008;152:59-62) ost preterm neonates receive parenteral nutrition (PN) as their initial nutritional support while they transition to enteral feedings. Although the mechanism is unclear, exposure to PN is one major factor in the development of cholestasis in preterm infants. Other known contributors to PN-associated cholestasis (PNAC) include sepsis, necrotizing enterocolitis (NEC), bowel surgery, and a lack of enteral feedings.1-4 In extremely low birth weight infants (birth weight ⬍1000 g), rates of PNAC have been documented to be as high as 50%.5 Subsequent liver failure, which results from long-term PN use, is the most significant complication. There is no preventive measure for PNAC. After cholestasis is established, the only clear therapy is the provision of full enteral nutrition. Studies with ursodeoxycholic acid have shown normalization of various liver function tests, but the numbers of subjects are small and methods and monitoring are varied.6,7 Alternative lipid emulsions have been shown to improve markers of cholestasis, but are not currently approved for use in the United States.8 Therefore, as the etiology and effective therapies are pursued, it remains essential to identify risk factors for the development of PNAC. Infants who are small for gestational age (SGA), or growth restricted, are known to have a number of metabolic abnormalities related to the liver as compared with their appropriate for gestational age (AGA) counterparts. For example, hepatic redox states were altered in rats with fetal growth restriction.9 Neonatal hypoglycemia occurs because of low glycogen stores. Autopsy results found SGA infants exposed to PN likely to have more severe liver damage.10 Selective assessments of small numbers of premature infants indicated that intrauterine growth restriction was a risk From the Yale University School of Medifactor for PNAC.11,12 cine, New Haven, Connecticut. We hypothesized that being born SGA was an independent risk factor for the Dr. Robinson was supported by NICHD development of PNAC. training grant HD 07094.
M
AGA BPD FEN GA NEC
Appropriate for gestational age Bronchopulmonary dysplasia Full enteral nutrition Gestational age Necrotizing enterocolitis
NICU PN PNAC SGA
Neonatal intensive care unit Parenteral nutrition Parenteral nutrition-associated cholestatsis Small for gestional age
Submitted for publication Jan 22, 2007; last revision received Apr 23, 2007; accepted Jun 1, 2007. Reprint requests: Daniel T. Robinson, MD, PO Box 208064, New Haven, CT 065208064. E-mail: [email protected]. 0022-3476/$ - see front matter Copyright © 2008 Mosby Inc. All rights reserved. 10.1016/j.jpeds.2007.06.002
59
METHODS A retrospective case-control study design, approved by the Yale University School of Medicine Human Investigation Committee, was used. Eligible patients were 1) cared for in the Yale New Haven Children’s Hospital Newborn Special Care Unit between January 1, 1994, and December 31, 2003, 2) born at ⬍34 weeks gestational age (GA), as determined by best obstetrical estimate and physical examination, and 3) exposed to ⱖ7 days of PN. Case subjects were defined as having cholestasis (direct bilirubin ⱖ2 mg/dL); control subjects did not have cholestasis. Exclusion criteria included cholestasis attributable to a congenital infection, an anatomic obstruction of the hepatobiliary tract, an inherited metabolic disease or metabolic syndrome, and a congenital gastrointestinal disorder requiring surgery. All infants with cholestasis underwent liver ultrasound scanning and testing for infectious hepatitis to exclude anatomic obstruction and infection as the cause of cholestasis. Control subjects were matched to case subjects on the basis of GA and date of birth. The first priority match was an infant of the same GA and a date of birth within 3 months of the case subject. If no match was found, then the next search was for an infant of the same GA and a date of birth within 6 months of the case subject. Finally, when no match was found with those criteria, a control subject with a GA within 1 week and a date of birth within 3 months of the case was used. Two control subjects were sought for each case. The primary outcome was the incidence of cholestasis among SGA infants. SGA was defined as a birth weight ⬍10% for GA.13 Secondary outcomes included the day of life at which cholestasis was detected and the level of direct bilirubin at the time of diagnosis. In addition, the number of days of PN exposure before the diagnosis, the total days of PN exposure, an enteral feeding history, and the incidence of late onset sepsis (diagnosed after 72 hours of life, either culture proven or culture negative but treated with antibiotics for ⱖ7 days), NEC (by using the modified Bell’s staging criteria14), and bronchopulmonary dysplasia (BPD; O2 treatment at 36 weeks postmenstrual age) were collected. The enteral feeding history included the number of days that infants were not offered enteral feeds (NPO) immediately after birth, the number of days with feeding interruptions (patient ordered NPO ⱖ12 consecutive hours within a 24-hour period) before and after full enteral nutrition (FEN) was achieved (ⱖ100 kcal/kg/day without PN support), and the day of life on which FEN was achieved.
Nutritional Protocol During Study Period The amino acid and lipid solutions given during the study period were Trophamine (B. Braun Medical, Bethlehem, PA) and Liposyn (Hospira, Lake Forset, IL). The day of life on which PN was started changed in our unit during the decade; it was introduced between 24 and 48 hours of life in the earlier years, as compared with within the first 24 hours in the more recent years. All infants exposed in this study achieved maximum protein and lipid infusions between 3 and 60
Robinson and Ehrenkranz
3.5 g/kg per day. Lipid levels were monitored, and when evidence of lipid intolerance was detected, infusions were held and then reinstated at a lower rate until an acceptable level was obtained. Dextrose was initiated on day of life 1 to achieve a glucose infusion rate between 6 and 8 mg/kg per minute and advanced to a maximum of 15 mg/kg per minute by day of life 7. There were no alterations to the PN when cholestasis was detected (eg, trace elements were not withheld).
Statistics To detect an odds ratio of 3 for the development of cholestasis when infants were SGA, 71 cases would be needed for analysis given an estimated incidence of SGA in our population of 10%, a 2-tailed ␣ level of 0.05, and power of 80%. Statistics were calculated with the Student t test and 2 analysis for means and frequencies. Multivariate analysis was performed by using logistic regression. Calculations were made with SPSS software version 13.0 (SPSS, Chicago, IL). A P value ⬍.05 was considered to be significant.
RESULTS During the study period, 79 eligible subjects with cholestasis were identified and able to be matched to at least 1 control subject; 2 matches were available for 73 of the case subjects (Table I). Of the control subjects, 22 (14.5%) had a gestational age ⫾1 week of the case. Maternal characteristics, aside from age at delivery, including factors that may have contributed to uteroplacental insufficiency (tobacco use, cocaine use, chronic and pregnancy-induced hypertension), exposure to antenatal steroids, presence of chorioamnionitis, or delivery of infant by cesarean section, were not different in the study groups (data not shown). Except for birth weight and length of stay, infant characteristics were not significantly different in the patient groups. However, infants with cholestasis had higher incidences of comorbidities. Furthermore, in infants with culture-proven sepsis, the rates of gramnegative sepsis were higher in infants with cholestasis (45% versus 29%). Among the 40 SGA infants in the total population studied, cholestasis developed in 23 (58%). The odds ratio of developing cholestasis, when born SGA, was 3.3 (95% CI, 1.6-6.6). Neither the incidence of SGA nor that of cholestasis changed throughout the study period. Of the infants with cholestasis, infants who were SGA had significantly lower birth weights and were significantly more mature than the infants who were AGA (Table II). Rates of comorbidities were similar in the groups. Mothers of SGA infants were more likely to have had chronic hypertension and pregnancy-induced hypertension (data not shown). The day of life and postmenstrual age at which cholestasis was detected did not differ significantly between the SGA and AGA infants. There was also no difference in the value of direct bilirubin on detection of cholestasis. Liver failure that lead to referral for transplantation or to death did not develop in any infant. The Journal of Pediatrics • January 2008
Table I. Patient characteristics and nutrition history
Patient characteristics Maternal age (years) Birth weight (grams) Gestational age (weeks) Sex (male/female) Race (W/H/B/other) SGA Late-onset sepsis NEC Required surgery BPD LOS (days) Nutrition history NPO before fed (days) DOL FEN achieved Total days PN Days NPO before FEN Days NPO after FEN
Cholestasis (n ⴝ 79)
No cholestasis (n ⴝ152)
P value
Odds ratio (95% CI)
28 ⫾ 6* 963 ⫾ 465 27 ⫾ 2 49/30 37/28/10/4 23 (29)† 63 (80) 40 (51) 6 51 (65) 106 ⫾ 40
30 ⫾ 6 1090 ⫾ 463 27 ⫾ 2 80/72 84/41/20/7 17 (11) 52 (34) 11 (7) 0 38 (25) 72 ⫾ 40
⬍.01 .049 .362 .173 .453 ⬍.01 ⬍.01 ⬍.01
3.3 (1.6-6.6) 2.7 (1.2-6.0) 7.5 (3.0-18.9)
10 ⫾ 9 43 ⫾ 25 63 ⫾ 34 11 ⫾ 12 12 ⫾ 17
6⫾4 23 ⫾ 11 24 ⫾ 14 3⫾5 2⫾4
⬍.01 ⬍.01 ⬍.01 ⬍.01 ⬍.01
⬍.01 ⬍.01
2.8 (1.4-5.9)
W/H/B/other, White/Hispanic/black/other; LOS, length of stay; DOL, day of life. *Mean ⫾ SD. †n (%).
Table II. Patient characteristics and nutrition history for patients with cholestasis SGA (n ⴝ 23)
AGA (n ⴝ 56)
Patient characteristics Birth weight (grams) 707 ⫾ 221* 1068 ⫾ 498 Gestational age (weeks) 28 ⫾ 2 27 ⫾ 3 Late-onset sepsis 16 (70)† 47 (84) NEC 11 (48) 29 (52) Required surgery 1 5 BPD 12 (52) 39 (70) DOL cholestasis detected 32 ⫾ 14 34 ⫾ 22 Direct bilirubin level at 4.7 ⫾ 2.7 4.3 ⫾ 3.5 diagnosis (mg/dL) PMA at diagnosis 32 ⫾ 3 31 ⫾ 3 Nutrition history NPO before fed (days) 10 ⫾ 11 10 ⫾ 9 DOL FEN achieved 32 ⫾ 16 48 ⫾ 27 Total days PN 49 ⫾ 24 68 ⫾ 36 Days NPO before FEN 8⫾9 13 ⫾ 12 Days NPO after FEN 11 ⫾ 12 12 ⫾ 19
P value ⬍.01 .025 .149 .749 .191 .615 .681 .159 .753 ⬍.01 .024 .150 .901
PMA, Postmenstrual age. *Mean ⫾ SD. †n (%).
Infants with cholestasis had the first sustained enteral feeds introduced later than infants without cholestasis (Table I). Infants without cholestasis achieved FEN sooner. In infants in whom cholestasis developed, total exposure to PN was greater and there were more days with feeding interrup-
tions. The AGA infants achieved full feeds later than infants who were SGA, and they were exposed to PN for a longer duration (Table II). Although infants who were SGA had been exposed to PN for 4 days fewer at the time of diagnosis, this difference was not significant (data not shown).
DISCUSSION We report evidence that preterm infants who are SGA and exposed to PN for at least 7 days are at increased risk for PNAC. Although there has been much discussion of this topic, our study contains the largest sample and achieves the necessary statistical power to define SGA as an independent risk factor for PNAC. A number of findings in this study agree with observations previously reported, such as lower birth weights, higher morbidity rates, and the nutritional histories of infants with cholestasis. With sepsis-related damage to hepatocytes being associated with cytokine release triggered by lipopolysaccharides, we also analyzed the sepsis results using just those infants with proven gram-negative rod infection15,16; the incidence of gram-negative rod sepsis was higher in the group with cholestasis. Because many neonatologists handle advancing enteral nutrition in SGA infants with caution because of fears of an increase in NEC in that population,17 we anticipated the finding that infants with cholestasis had more feeding interruptions than infants without cholestasis. However, it was unexpected to find that, in patients with cholestasis, SGA infants achieved full enteral feeds at an earlier time and were exposed to less PN than AGA infants.
Parenteral Nutrition-Associated Cholestatsis in Small for Gestational Age Infants
61
Perhaps differences in hepatic metabolism attributable to impaired perfusion of the liver in utero will be identified between infants who are growth-restricted and infants who are not growth-restricted. Because the development and regulation of bile salt transporters in rats have been shown to markedly increase in the neonatal period, alterations in the transcription, translation, and function of proteins involved in bile production are likely.18 Another contributor may be altered metabolism of copper, a metal that may have dual impacts, protective as a cofactor in antioxidants but hepatotoxic when accumulated in excess.19,20 The absence of copper in postmortem livers of infants who are SGA and have PNAC suggests that insufficient antioxidant activity may be a contributing factor in the development of cholestasis in SGA infants.10 An additional area highlighted by our results is the striking difference in the incidence of BPD between infants with and infants without cholestasis. The bulk of vitamin A is stored in the liver. A significant reduction of BPD has been found in preterm infants to whom vitamin A was administered.21 A partial explanation of our findings may be that the liver damage consequent to the cholestasis decreases the bioavailability of vitamin A. The association may also be explained by the state of inflammation found in both cholestasis and BPD.22,23 This study is limited because of its retrospective design. The frequency of testing direct bilirubin levels was certainly affected by practice variation. This impairs the ability to accurately determine the time of onset of PNAC. Research on the basis of tests such as serum bile acid levels may more accurately pinpoint the time of onset of disease. Therefore, our data do not permit an accurate interpretation of the number of days of PN exposure before onset of disease and clinical detection. Because of the varied number of levels sent on each patient, patterns of bilirubin, such as timing of peak level and resolution of disease, and other variables including transaminases and coagulation profiles could not be studied for trends. A prospective study with standardized testing would help answer questions, including the number of days of exposure necessary before the development of cholestasis. Preterm infants who are SGA are an at-risk population in many respects. These data identify an increased susceptibility of cholestasis developing when exposed to PN. Because the mechanisms facilitating the development of PNAC are not fully understood, the identification of preventative and therapeutic modalities has been impaired. Investigations into the pathways involved in bile production in subjects who are growth restricted may reveal vital information about the
62
Robinson and Ehrenkranz
pathophysiology and effective therapies of this currently unavoidable complication.
REFERENCES 1. Drongowski RA, Coran AG. An analysis of factors contributing to the development of total parenteral nutrition-induced cholestasis. JPEN J Parenter Enteral Nutr 1989;13:586-9. 2. Kelly DA. Liver complications of pediatric parenteral nutrition-epidemiology. Nutrition 1998;14:153-7. 3. Kaufman SS, Gondolesi GE, Fishbein TM. Parenteral nutrition associated liver disease. Semin Neonatol 2003;8:375-81. 4. Moss RL, Das JB, Raffensperger JG. Necrotizing enterocolitis and total parenteral nutrition-associated cholestasis. Nutrition 1996;12:340-3. 5. Beale EF, Nelson RM, Bucciarelli RL, Donnelly WH, Eitzman, DV. Intrahepatic cholestasis associated with parenteral nutrition in premature infants. Pediatrics 1979;64:342-7. 6. Al-Hathlol K, Al-Madani A, Al-Saif S, Abulaimoun B, Al-Tawil K, El-Demerdash A. Ursodeoxycholic acid therapy for intractable total parenteral nutrition-associated cholestasis in surgical very low birth weight infants. Singapore Med J 2006;47:147-51. 7. Chen CY, Tsao PN, Chen HL, Chou HC, Hsieh WS, Chang MH. Ursodeoxycholic acid (UDCA) therapy in very-low-birth-weight infants with parenteral nutritionassociated cholestasis. J Pediatr 2004;145:317-21. 8. Gura KM, Duggan CP, Collier SB, Jennings RW, Folkman J, Bistrian BR, et al. Reversal of parenteral nutrition-associated liver disease in two infants with short bowel syndrome using parenteral fish oil: implications for future management. Pediatrics 2006;118:e197-201. 9. Lane RH, Flozak AS, Ogata ES, Bell GI, Simmons RA. Altered hepatic gene expression of enzymes involved in energy metabolism in the growth-retarded fetal rat. Pediatr Res 1996;39:390-4. 10. Zambrano E, El-Hennawy M, Ehrenkranz RA, Zelterman D, Reyes-Mugica M. Total parenteral nutrition induced liver pathology: an autopsy series of 24 newborn cases. Pediatr Dev Pathol 2004;7:425-32. 11. Baserga MC, Sola A. Intrauterine growth restriction impacts tolerance to total parenteral nutrition in extremely low birth weight infants. J Perinatol 2004;24:476-81. 12. Aucott SW, Donohue PK, Northington FJ. Increased morbidity in severe early intrauterine growth restriction. J Perinatol 2004;24:435-40. 13. Alexander GR, Himes JH, Kaufman RB, Mor J, Kogan M. A United States national reference for fetal growth. Obstet Gynecol 1996;87:163-8. 14. Walsh MC, Kliegman RM. Necrotizing enterocolitis: treatment based on staging criteria. Pediatr Clin North Am 1986;33:179-201. 15. Green RM, Beier D, Gollan JL. Regulation of hepatocyte bile salt transporters by endotoxin and inflammatory cytokines in rodents. Gastroenterology 1996;111:193-8. 16. Nathanson MH, Boyer JL. Mechanisms and regulation of bile secretion. Hepatology 1991;14:551-66. 17. Yu VYH, Upadhyay A. Neonatal management of the growth-restricted infant. Semin Fetal Neonatal Med 2004;9:403-9. 18. Tomer G, Ananthanarayanan M, Weymann A, Balasubramanian N, Suchy FJ. Differential developmental regulation of rat liver canalicular membrane transporters Bsep and Mrp2. Pediatr Res 2003;53:288-94. 19. Cordano A. Clinical manifestations of nutritional copper deficiency in infants and children. Am J Clin Nutr 1998;67:1012-6S. 20. Peña M, Lee J, Thiele DJ. A delicate balance: homeostatic control of copper uptake and distribution. J Nutr 1999;129:1251-60. 21. Tyson JE, Wright LL, Oh W, Kennedy KA, Mele L, Ehrenkranz RA, et al. Vitamin A supplementation for extremely-low-birth-weight infants. National Institute of Child Health and Human Development Neonatal Research Network. N Engl J Med 1999;340:1962-8. 22. Chess PR, D’Angio CT, Pryhuber GS, Maniscalco WM. Pathogenesis of bronchopulmonary dysplasia. Semin Perinatol 2006;30:171-8. 23. Jobe AH, Bancalari E. Bronchopulmonary dysplasia. Am J Respir Crit Care Med 2001;163:1723-9.
The Journal of Pediatrics • January 2008