Transplantation: A Randomized Prospective Study

Nutritional Support after Liver

JAMES REILLY, M.D., REKHA MEHTA, M.D., LEWIS TEPERMAN, M.D., SAMUEL CEMAJ, M.D., ANDREAS TZAKIS, M.D., KATSUHIKO YANAGA, M.D., PAMELA RITTER, R.N., ABDUL REZAK, M.D., AND LEONARD MAKOWKA, M.D., PH.D. From the

Department of Surgery, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania

ABSTRACT. Nutritional support in patients with advanced cirrhosis is difficult due to protein, fluid and salt restrictions. Successful liver transplantation should improve nutrient tolerance. We randomly assigned 28 hypoalbuminemic cirrhotic patients to receive, immediately after liver transplantation, one of three regimens: group 1, no nutritional support (n 10); group 2, total parenteral nutrition (TPN) (35 kcal/kg/day) with standard amino acids (1.5 g/kg/day) (n 8); or group 3, isocaloric isonitrogenous TPN with added branched-chain amino acids (n 10). Therapy was continued for 7 days posttransplant. Jaundice resolution was unaffected by nutritional support. Nitrogen balance favored both TPN groups. =

=

=

Patients with advanced cirrhosis

are

often malnour-

ished.I-3 Nutritional restoration in such patients is often difficult due to salt and water intolerance; in many, encephalopathy necessitates severe protein restrictions which may further worsen nutritional status.44 When liver dysfunction is severe enough to warrant liver transplantation, the nutritional status of these patients often is a factor in determining outcome. The stress of this lengthy and difficult operation further increases energy requirements and protein catabolism.’ Preoperative malnutrition may substantially worsen under the duress of the transplantation procedure itself and the often stressful and prolonged postoperative course. In transplant candidates, preoperative nutritional therapy is further complicated by the necessity to continue therapy in an outpatient setting while awaiting a donor organ. The vagaries associated with such availability often makes an effective nutritional regimen im-

possible. If nutritional rehabilitation is difficult preoperatively, the postoperative period may be more conducive to nutritional support. The transplanted organ, if functional, should improve the very factors which limited effective nutritional support prior to transplant. If this is so, postliver transplant patients should demonstrate improved tolerance to fluid, salt, and protein, and should be able to achieve positive nitrogen balance and restore lean

body

mass.

In this randomized prospective, partially blinded study, we have employed total parenteral nutrition in a group of cirrhotics immediately after the conclusion of a

Branched-chain amino acid (BCAA)aromatic amino acid ratios were highest in group 3. Coma scores and serum ammonia levels were similar in all groups. Both TPN groups achieved respirator independence earlier; this difference was not statistically significant. Group 1 patients stayed longest in ICU; the difference was statistically significant. TPN with either standard or BCAA- enriched amino acids is tolerated well immediately after successful liver transplant. Positive nitrogen balance is achieved; large protein loads do not worsen encephalopathy. Nutritional support may improve respiratory muscle function, allowing earlier weaning from ventilatory support. A shortened Journal of length of ICU stay justifies the expense of TPN. ( Parenteral and Enteral Nutrition 14:386-391, 1990)

technically successful liver transplantation to assess tolerance to such therapy, and to determine whether any beneficial consequences of therapy could be perceived. MATERIALS AND METHODS

Patients were accessed to this study immediately before or after liver transplantation. All patients met clinical criteria for transplantation at the University of Pittsburgh; all were treated with a conventional immunosuppressive regimen of cyclosporine and steroids. All patients were hypoalbuminemic prior to the transplant (mean serum albumin, 2.52 ± 0.39 g %). After informed consent had been obtained, patients were randomized to receive one of three nutritional regimens. The clinical characteristics of the three patient groups are shown in Table I. In group 1, no specific nutritional therapy was given; these patients received standard isotonic intravenous glucose solutions as dictated by their clinical hydration status. In group 2, nutritional support consisted of &dquo;standard&dquo; total parenteral nutrition (TPN) supplying non-protein caloric intake at 35 kcal/kg/day and 1.5 g/ kg/day of protein intake. Dextrose intake was limited to 5 mg/kg/min; the balance of the energy intake was administered as intravenous fat emulsion (10% Intralapid solution). Crystalline amino acids (5%) in the &dquo;standard&dquo; formulation were administered in a 25% dextrose solution. Group 3 patients received isocaloric, isonitrogenous TPN; these patients received a branchedchain amino acid-enriched formula, adding Branchamin to a 3.5% amino acid base solution identical to that of the group 2 patients to achieve a final protein concentration of 5%. The total protein intake for this group was 386

387

also 1.5 g/kg/day. The composition of the two TPN solutions is described in Table II. Patients in groups 2 and 3 received the parenteral nutrition solutions in a &dquo;blinded&dquo; fashion; ie, the protein composition of each individual’s prescription was known only to be the pharmacist performing the randomization and compounding. Group 1 patients received, of course, TPN. After randomization, patients began their respective nutritional regimens on the first postoperative day, once hemodynamic stability was achieved. On the first treatment day, one-half the total estimated nutritional regimen was given; from days 2 through 7 the full nutritional regimen was employed. TPN was administered through &dquo;dedicated&dquo; ports of triple lumen catheters inserted in the operating room prior to the transplant procedure. A clear liquid diet was allowed in all groups when postoperative ileus resolved. Standard liver function tests, electrolytes, glucose, calcium, phosphorus and magnesium, and plasma amino acids were monitored daily. All urine and aliquots of fluid from drainage catheters were collected daily for nitrogen balance calculation. The presence of encephalopathy was determined daily by clinical criteria (a modified Glasgow coma score). The time nec-

no

FIG. 1. Total

serum

bilirubin::!: SEM. Differences

were not

statisti-

cally significant.

TABLE I characteristics

Preoperative

TABLE II

Composition of TPN study solutions

~

~~~

~~~

~,

~~~~~

~

-~-

~

~

~

~

~

TABLE III

Operative stress FIG. 2. Average daily BUN and creatinine SE~1.. One patient in group 1 became azotemic due to cyclosporine toxicity. If this patient were

excluded, differences would

not

be statistically significant.

388

FIG. 3. Average daily modified Glasgow Coma Score ± SD. No significant differences were observed.

FIG. 5. Average daily nitrogen balance ± SEM in all three groups. The TPN groups 2 and 3 had better nitrogen balances compared to group 1, and the differences were statistically significant (p < 0.05).

FIG. 6. Cumulative FIG. 4. Average daily serum ammonia levels ± SEM. No differences were observed.

significant

essary postoperatively to achieve respiratory independence, extubation, and transfer from the intensive care

unit to the nursing floor was also noted. All data were collected and entered into a statistical database (RS1, BBN Software, Cambridge, MA). Comparative statistics between the study groups consisted of the Wilk-Shapiro test for normality and the unpaired Student’s t-test. Any p value less than 0.05 was assumed to be statistically significant. All data are reported as mean ± SEM. RESULTS

All patients in this study had advanced liver disease of diverse causes. All were Child’s C class patients, with

nitrogen balance.

ascites, muscle wasting, hypoalbuminemia and jaundice. No differences were noted between each randomization group with respect to age, sex, or preoperative serum albumin levels (Table I). The stress of the operative procedure was similar in all groups, judged by average length of the surgical procedure and units of blood transfused intraoperatively

(Table III). The response to successful liver transplantation was reflected by the observed fall in total serum bilirubin (Fig. 1). While the group 1 patients had, in general, a somewhat slower resolution of jaundice, this difference was

not

statistically significant.

Renal function was unaffected by nutritional support (Fig. 2). A single group 1 patient became transiently azotemic, altering the average BUN and creatinine values for this group. If this patient were excluded, no

389

logistical concerns: nonetheless. 10 of the 28 patients (2 in group 1. and 4 each in groups 2 and :3’ were noted on admission to be frankly encephalopathic by clinical

to

criteria. The high protein loads received by the group 2 and group 3 patients produced no discernible differences in mental function compared to group 1 (Fig. :3). Within 24 to 36 hr after transplantation. all patients were alert, and able to follow simple commands and answer questions. Serum ammonia levels remained normal and were unaffected by the type of nutritional support given (Fig.

4).

FIG. 7. Serum

3-methylhistidine levels ± SEM. Group 1 patients significantly higher (p < 0.05) on days 3 to 6 compared to group 2, and on days 3 to 4 compared to group 3.

were

Average daily nitrogen intake in group 2 and group 3, randomized to received postoperative TPN, was 10.06 ± 5.73 g/day and 10.40 ± 5.67 g/day respectively. This difference was not statistically significant. From a practical viewpoint, all patients had just progressed to a clear liquid diet by the fifth or sixth postoperative day, but intake remained poor up to the 10th day. Nitrogen balance was significantly improved in both TPN groups compared to group 1 (group 1 us group 2, p < 0.0001; group 1 us group 3, p < 0.0011) (Fig. 5). While the average daily nitrogen balance in group 2 was superior to that of group 3, the difference was not statistically significant (p < 0.52). Cumulative nitrogen balance over the 7 study days (group 1, -278.83 g; group 2, -10.65 g; group 3, -62.4) favored groups 2 and 3 in a statistically significant fashion (Fig. 6). Plasma amino acid levels varied from day to day and from patient to patient. Serum 3-methylhistidine levels were significantly increased in the group 1 patients from days 2 to 6 of the study, perhaps reflecting increased skeletal muscle breakdown in this unfed group (Fig. 7). Urinary 3-methylhistidine excretion was not measured. The branched-chain amino acid/aromatic amino acid ratio was significantly elevated in the group 3 patients receiving supplemental branch chain amino acids compared to both the group 1 and group 2 patients ( p < 0.05) (Fig. 8). Both TPN groups achieved

independence

from venti-

latory support and extubation earlier than the control group 1 patients (Table IV), but these differences were not statistically significant. Because length of ICU stay was primarily dependent upon the need for ventilatory support, the TPN groups also left the ICU earlier (Table FIG. 8. Branched-chain (BCAA) to aromatic amino acid (AA.A) ± SEM. Group 3 levels demonstrate statistically significant differences compared to both groups 1 and 2 (p < 0.05) on day 2 to 6. TABLE IV

Postoperative data

*

t

Group 1 vs group 2, group 3: p < 0.05. LOS, length of postoperative hospital stay (days).

differences

were

IV); this advantage for the nutritionally treated groups was not statistically significant (group 1 us group 2 or group 3, p < 0.05). Because of the expense associated with parenteral nutrition, we examined the total hospital charges of all study patients to determine whether the expense of TPN was reflected here. Total hospital charges were highest in the control group 1 patients, hut the differences between study groups was not statistically significant (Table IV). Three deaths occurred, none during the studv ’&dquo;’~’-~ al. Two group 1 patients died at 17 and 54 r -: transplantation, of rejection and sepsis resph, 1 B l> 1.B. A single group 3 patient died of recurrent acute hepatitis and HIV infection 65 days postoperative.

observed.

Signs of encephalopathy were monitored using a modified Glasgow Coma Score. Formal preoperative scoring could not be accomplished prior to transplantation due

DISCUSSION

Patients who are

are

candidates for liveer

frequently malnourished, but

the

transplantation

impact

of malnu-

390 trition upon their perioperative course is speculative. While malnourished patients in general are at higher risk of complications and even death after surgical procedures, this risk has not been quantitated in liver trans-

plant recipients. Restoring nutritional status prior to transplant is complicated by numerous pathophysiologic and logistical factors. Classic nutritional parameters, such as body weight are often confounded by the presence of edema and ascites; thus we chose serum albumin as the major nutritional entry criterion to this study, recognizing that the etiology of hypoalbuminemia in the cirrhotic is multifactorial. Cirrhotic

transplant candidates are typically protein intolerant; encephalopathy limits effective protein intake, often to less than 40 g of protein/day. Fluid and salt restrictions further compromise nutritional intake. Finally, the availability of a suitable donor organ defines the time available for nutritional repletion. Thus, the immediate postoperative period may be more conducive to nutritional therapy. If the newly transplanted liver can assume its metabolic role promptly, the need for protein, salt, and water restriction may be lessened, and effective nutritional support may be feasible. Calne’s group6,7 reported the feasibility of TPN in achieving positive nitrogen balance in 14 post-liver transplant patients. We have demonstrated in this study that patients undergoing liver transplantation can in fact tolerate protein intake far in excess of that typically achieved in the setting of advanced liver failure. Both nutritional regimens were designed to provide 1.5 g of protein/kg/day by the second postoperative day. In both TPN groups, positive nitrogen balance was achieved without any clinical evidence of protein intolerance, such as encephalopathy. Serum ammonia levels remained normal, and BUN levels were unchanged. The group 3 patients receiving the branched-chain amino acid-enriched formula did not evidence any improvement in nitrogen balance compared to those in group 2. Their branched chain/aromatic amino acid ratios were, however, higher. The ratio of branched-chain amino acids to aromatic amino acids has been implicated in the pathogenesis of hepatic encephalopathy,’ and we have reported normalization of this ratio spontaneously after successful liver transplantation.’ Cerra et all’,&dquo; have suggested that increased branched-chain amino acid intake promotes improved nitrogen balance, particularly in stressed surgical patients. We cannot confirm their observations in our patients. Daly et al 12 and Bonau et all’ have also found no significant benefit for postoperative patients supported with branched-chain amino acid solutions. Rogers et all’ and Arora et allj have suggested that protein depletion and malnutrition may interfere with numerous aspects of pulmonary function, including surfactant and mucous production, and diaphragmatic strength. We can only speculate what impact nutritional support had upon such parameters in the immediate

post-liver transplant period. However, improved energy and protein intake may have played a role in facilitating

weaning from ventilatory support, perhaps by enhancing diaphragmatic and accessory respiratory muscle function.

Finally, an unexpected benefit of nutritional support in the immediate post-transplant period proved to be an economic one. Because the major determinant of length of ICU stay at our institution is the need for respiratory support, the extubated, nutritionally repleted patients left the ICU an average of 2.3 days earlier than did those receiving no nutritional support. This resulted in an estimated cost savings of $20,500, more than enough to justify the costs of the TPN. We conclude from this study that malnourished liver transplantation patients can receive aggressive nutritional support in the immediate postoperative period and achieve positive nitrogen balance. Nitrogen balance is unaffected by the addition of supplemental branch chain amino acids. Plasma amino acid levels are generally physiologic during nutritional support; the branchedchain to aromatic amino acid ratio is increased in patients receiving supplemental branched-chain amino acids. Hepatic tolerance to nitrogen in these patients seems virtually normal with respect to metabolizing large protein loads. Urea nitrogen, creatinine, and ammonia levels remain normal; encephalopathy does not occur secondary to nitrogen intake. Nutritional support may enhance muscle strength in the early postoperative period and facilitate weaning from ventilatory support and shortening ICU stay. Following liver transplantation, nutritional support does not increase total hospital costs. REFERENCES 1. Hehir

D, Jenkins R, Bistrian B,

et al: Nutrition in

patients

undergoing orthotopic liver transplantation. JPEN 9:695-700, 1985 2. DiCecco S, Wieners E, Wieners R, et al: Assessment of nutritional status of patients with end stage liver disease undergoing liver transplantation, Mayo Clin Proc 64:95-102, 1989 3. Fischer J, Bower R: Nutritional support in liver disease. Surg Clin North Am 61:653-660, 1981 4. Shronts E, Teasley K, Theole S, et al: Nutritional support of the adult liver transplantation candidate. J Am Diet Assoc 87:442-448. 1987 5. Shanbhogue R, Bistrian B, Jenkins R, et al: Increased protein catabolism without hypermetabolism after human orthotopic liver transplantation. Surgery 101:146-149, 1987 6. O’Keefe S, Williams R, Calne R: Catabolic loss of body protein after human liver transplantation. Br Med J 280:1107-1108, 1980 7. Johnson P, O’Grady J, Calvey H, and Williams R: Nutrition in liver transplantation. IN Liver Transplantation, Calne R (ed). Grune and Stratton, New York, 1987, pp 113-117 8. McCullough A, Czaja A, Donald Jones J, et al: The nature and prognostic significance of serial amino acid determinations in severe chronic and acute liver disease. Gastroenterology 81:645-

652, 1981 9.

Reilly J, Halow G, Gerhardt A, et al: Plasma transplantation: Correlation with clinical

amino acids in liver outcome.

Surgery

97:263-269, 1985 10. Cerra F, Mazuski J, Chute E, et al: Branched chain metabolic support, a prospective double-blind trial in surgical stress. Ann Surg 199:286-291, 1984 11. Cerra F, Upson D, Angelico R, et al: Branched chains support postop protein synthesis. Surgery 92:192-199, 1982

391 et al: Effects of post-op infusion of branched chain amino acids on nitrogen balance and fore arm muscle substrate flux. Surgery 94:151-158, 1983 13. Bonau R. Jeevanandam M, Moldawer L, et al: Muscle amino acid reflux in patients receiving BCAA solutions after surgery. Surgery

12.

Daly J. Mihranian M, Kehoe J,

101:400-407, 1987

et al: Nutrition and COPD: Stateof-the-art mini-review. Chest 85:63:-66s. 1984 15. Arora N. Rochester D: Respiratory muscle strength and maximal voluntary ventilation in undernourished patients. Am Rev Respir Dis 126:5-8, 1982

14.

Rogers R, Dauber J. Sanders M.