Comparison of Dietary Protein with an Oral, Branched Chain-Enriched Amino Acid Supplement in Chronic Portal-Systemic Encephalopathy: A Randomized Controlled Trial

0270-9139/84/0402-0279$02.00/0

Vol. 4, No. 2, pp. 279-287, 1984 Printed in U.S. A.

HEPATOLOGY

Copyright 0 1984 by the American Association for the Study of Liver Diseases

Comparison of Dietary Protein with an Oral, Branched Chain-Enriched Amino Acid Supplement in Chronic Portal-Systemic Encephalopathy: A Randomized Controlled Trial DOUGLASHORST,NORMAND. GRACE,HAROLD 0. CONN, EUGENE SCHIFF, ALFREDOVITERI, DAVIDLAW,AND COLIN E. ATTERBURY

STEVEN SCHENKER,?

Lemuel Shattuck Hospital, Boston, Massachusetts; Veterans Administration Medical Center at West Haven, West Haven, Connecticut; Veterans Administration Medical Center at Nashville, Nashuille, Tennessee; University of Miami Medical Center, Miami, Florida; Scott and White Clinic, Temple, Texas; and University of New Mexico Medical Center, Albuquerque, New Mexico A randomized study was conducted in 37 hospitalized patients at six cooperating hospitals in which protein-intolerant cirrhotic patients were fed increasing amounts of either dietary protein or a branched-chain enriched amino acid solution (BCAA) until they attained an intake of 80 gm protein per day or equivalent or until they developed stage 2 encephalopathy. All patients initially received 20 gm of dietary protein for 1 week, after which 20 gm of protein or BCAA were added weekly. Nitrogen balance improved from negative to positive in all patients in whom it was measured and increased equally in both groups. Seven of the 20 patients in the protein group and 1 of 17 in the BCAA group developed encephalopathy of stage 2 or greater (p < 0.05). Changes in each component of the portal-systemic encephalopathy syndrome were compared, and differences were statistically significant for mental status grade (p < 0.01), asterixis (p < 0.05), Portalsystemic encephalopathy index (p < 0.01), but insignificant for Number Connection Test, EEG or ammonia. Plasma amino acid profiles showed an increase in BCAA in the study group. Thus, oral BCAA supplements appear to induce positive nitrogen balance to approximately the same degree as an equivalent amount of dietary protein without inducing encephalopathy as frequently.

Patients with chronic portal-systemic encephalopathy

(PSE) are faced with a nutritional dilemma. They require dietary protein to maintain nitrogen balance, but the ingestion of protein often precipitates encephalopathy (1-3). Despite treatment with lactulose or antibiotics, many cirrhotic patients are unable to tolerate sufficient protein to prevent longstanding negative nitrogen balance. This malnutrition may contribute to further deterioration of hepatic function and general health. Various ways of resolving this dilemma have been

Received March 2, 1983; accepted September 7, 1983. t Present address: University of Texas Health Science Center, Department of Gastroenterology, San Antonio, Texas 76501. Address reprint requests to: Douglas Horst, M.D., Department of Gastroenterology, Lemuel Shattuck Hospital, 185 Morton Street, Jamaica Plain, Massachusetts 02130.

attempted. Because not all types of protein appear to induce PSE to the same degree (4,5), some investigators have administered less encephalopathogenic proteins, such as vegetable protein, with some success (6, 7). In an attempt to correct the typical plasma amino acid profile of PSE, which is characterized by decreased levels of branched-chain amino acids (BCAAs) (leucine, isoleucine and valine) and increased levels of aromatic amino acids (AAAs) (phenylalanine, tyrosine, free tryptophan and methionine) (8-15), investigators have administered mixtures of amino acids rich in BCAAs and deficient in AAAs (16-24). It is extremely difficult to draw conclusions from these data since most of them were not controlled trials, and the few controlled trials evaluated different types and degrees of encephalopathy, using different mixtures of BCAAs administered in different ways, using different endpoints and finding different results (25). 279

280

HEPATOLOGY

HORST ET AL. PROTEIN GROUP

The present investigation was undertaken to compare in double-blind fashion the comagenic capacity and nutritional potential of dietary protein and an equivalent amount of a mixture of amino acids rich in BCAA and reduced in AAA administered orally to stable cirrhotic pajients with chronic PSE. MATERIALS AND METHODS Cirrhotic patients admitted to the six collaborating hospitals between April 4, 1978 and December 19, 1979 were studied. All had had chronic recurrent PSE and had required restriction of dietary protein to levels of 40 gm per day or less. In all, the cirrhosis was stable as determined by clinical and laboratory evaluations. All had either normal mental state or stable, Grade 1 encephalopathy according to a standard scale of grading hepatic coma (26). Patients with gastrointestinal bleeding, bacterial infection, the hepatorenal syndrome or acute alcoholic or fulminant hepatitis were excluded. All patients were placed on a standardized basal diet which contained 20 gm protein per day. At the end of 1 week on the basal diet, patients were selected using a sealed envelope technique based on a book of random numbers to enter the protein or BCAA groups. The investigators performing clinical assessments were unaware of the diet assignment. During the second week, patients in the dietary protein group were offered an additional 20 gm protein daily (total 40 gm per day) while the amino acid group received the basal 20 gm protein diet plus an equivalent amount of an amino acid solution enriched with BCAA and deficient in AAA (A662) (Table 1).The amino acids were dissolved in a flavored maltodextrin solution, and patients in the dietary protein group received the same amount of a similarly flavored glucose solution to insure equal intake of fluid, carbohydrate and calories. During the third and fourth weeks, patients received an additional 20 or 40 gm per day of protein, respectively, or the BCAA solution (Figure 1). Identical research menus were prepared by research dietitians at all hospitals, with occasional substitutions of comparable items. The 7-day menu apportioned 20 gm of protein (calculated as providing 3.2 gm Nz) and

TABLE 1. AMINOACID CONTENT Amino acid

L-Tryptophan L-;wleucine L-Leucine L-Lysine acetate L-Methionine L-Phenylalanine L-Threonine L-Valine L- Alanine L- Arginine L-Histidine L-Proline L-Serine Glycine a

Hepatic Aid’”.

OF A662“

Grams specific amino acid/ZO gm total amino acids

0.16 2.20 2.68 2.10 0.24 0.24 1.10 2.04 1.86 1.46 0.56 1.94 1.22 2.18

I

0

I

WEEKS 2

3

a A 6 6 2

,BASAL P E R , O D ~RANDOMIZATION

BCAA GROUP

80

I

60

DIETARY 60 PROTEIN 4o g per day 20

40 A 6 6 2

2o gperday 0

0 I

2

3

4

FIG.1. Schedule ofprotein and A662 intake. During the basal period (first week), all patients received dietary protein of 20 gm per day. After randomization the protein group ( t o p ) received 40, 60 and 80 gm per day dietary protein, respectively, during the second, third and fourth weeks. After randomization, the BCAA group (bottom) received 20 gm dietary protein for the next 3 weeks of the study. They received 20, 40 and 60 gm per day A662 solution (cross-hatched area), during the second, third and fourth weeks, respectively.

approximately 2,300 nonprotein calories, 1,800 as carbohydrate and 500 as fat, into three meals each day. The food contained animal, dairy and vegetable protein in the same proportions found in a normal hospital diet. This diet remained unchanged for patients in the A662 group, but for patients in the protein group, the quantities of food and glucose solution were increased weekly to provide 20 gm additional protein and approximately 300 additional nonprotein calories. Food was weighed before and after each meal and protein consumption was calculated and recorded. The amino acid solution employed was an experimental solution (A662) provided by McGaw Laboratories. It contained small amounts of the essential AAAs and methionine and relatively large amounts of BCAAs as well as several nonessential amino acids (Table 1).The solution also contained maltodextrins, sucrose, and lipid as partially hydrogenated soy bean oil, lecithin and mono- and diglycerides. The A662 solution was prepared by mixing 20 gm of amino acid powder, providing 2.9 gm of nitrogen, in 450 ml of water, with an osmolarity of 940 mOsm per liter and providing approximately 750 calories, 150 derived from fat and 520 from carbohydrate. All urine and stools were collected, weighed and coded daily, except a t one hospital where only urines were collected. Aliquots were frozen and sent to a commercial laboratory where total urinary and fecal nitrogen were measured by code number. Nitrogen balance was calculated weekly based on samples collected during the final 5 days of each experimental period. Cutaneous and other losses of nitrogen were estimated to be 0.5 gm per day in all patients (27).

Vol. 4,No. 2, 1984

COMPARISON OF DIETARY PROTEIN IN CHRONIC ENCEPHALOPATHY

Patients were assessed daily for the grade of encephalopathy on a 0 to 4 scale (26). Mental state was evaluated for each patient throughout the study by the same physician who did not know what the patient was receiving and who, except at one center, did not participate in the clinical management of the patient. Each examination was performed after the food trays had been removed from the patient’s room to prevent the physician from learning which diet the patient was receiving. If PSE of Grade 2 or more developed, the treatment the patient was receiving was considered a failure, the study terminated and appropriate therapy administered. Patients were assessed daily for the presence of asterixis on a semiquantitative scale from 0 to 4 (26). The number connection test (NCT) (28) was performed daily and the time in seconds required to complete the test recorded. Four variations of the NCT of equal degree of difficulty were used in rotation to minimize the learning effect. The score in seconds was converted to a 0 to 4 scale (26). EEGs were obtained at the end of each week of observation whenever possible. The EEG tracings for each patient were read as a group on a 0 to 4 score by the neurologist who had devised and calibrated the grading system (26) and who was unaware of the patients’ identities or therapy. No EEGs were available for the patients from one hospital. Fasting ammonia concentrations were measured three times per week using the method of Seligson and Ishihara (29). Three centers used arterial blood and two used venous blood samples throughout the study. The blood ammonia concentration was converted to a 0 to 4 arbitrary scale for use in the PSE Index (26). The PSE Index is the sum of the grade of mental state (multiplied by 3 because of its relative importance in the clinical assessment of encephalopathy) plus the grades of asterixis, NCT, EEG, and blood ammonia concentration, each expressed on a 0 to 4 scale (26). NCT is arbitrarily graded as follows: Grade 0, 120 sec. The arterial ammonia concentration is arbitrarily graded as follows: Grade 0 , 4 5 0 pg per 100 ml; Grade 1, 150 to 200; Grade 2,201 to 250; Grade 3, 251 to 300; Grade 4, >300. The total of the weighted scores, the maximum of which is 28, is termed the PSE Sum. The PSE Sums are not always comparable because individual components of the index may be missing. The PSE Index, which is the sum divided by the highest possible PSE Sum for that particular calculation is expressed in per cent. This composite score permits comparison of the severity of PSE between different patients and in the same patient a t different times, even when one or more of the components are missing or are untestable, as long as the mental state grade is known. Fasting blood was drawn to measure the concentration of amino acids in plasma at the end of each weekly period of therapy. Blood was drawn a t least 10 hr after the last dose of the amino acid solution had been ingested. Samples were immediately centrifuged, the plasma aspirated and stored at -10°C until analyzed as unidentified specimens at a commercial laboratory.

281

Blood was drawn three times per week for analysis of 19 biochemical tests, including serum bilirubin (direct and total), alkaline phosphatase, SGOT and SGPT, albumin, total protein, glucose, urea, creatinine and electrolytes and six standard hematologic assays. Physical examination was performed before and at weekly intervals during treatment. Vital signs and body weight were measured and recorded daily. All medications, adverse effects and patients’ complaints were recorded daily. This protocol was approved by the Human Investigation Committee at each hospital and informed written consent was obtained from each patient prior to randomization. Data were analyzed using Student’s paired and unpaired tests, Fisher’s exact test and analysis of variance. When analyzing serial data, both absolute values and differences between the initial and subsequent values were used. Basal values were defined as the mean of the daily measurements (mental state, asterixis and NCT) during the final 3 days of the baseline period, and were compared with the mean of the last 3 days of each subsequent period to avoid hangover effects from the previous period. Blood ammonia levels represented the mean of the last two measurements of the basal period of each week. EEG grades were determined from the weekly EEGs. In patients who developed PSE and in whom the study was terminated, the observations made on the day of termination were considered the final values. Sample size calculations were based on the assumption that dietary protein would induce PSE in 50%of patients taking 80 gm daily while A662 would do so in 20%. To achieve statistical significance (p < 0.05) while reducing error to p < 0.10 would require 40 patients in each group. RESULTS Between April 4, 1978 and December 19, 1979, 37 patients were randomized, 20 to the protein group and 17 to the BCAA group. Ten were entered from the Scott and White Clinic, 9 from the Lemuel Shattuck Hospital, 6 from the West Haven Veterans Administration and Yale Medical Centers, 6 from the University of Miami Medical Center, 4 from the Vanderbilt Medical Center and 2 from the University of New Mexico Medical Center. Two additional patients began the basal period, but were withdrawn before randomization, one because PSE had developed and one because he would not adhere to the protocol. They were not included in any calculations. The data were analyzed for the total group of 37 patients (20 in the protein group and 17 in the BCAA group) and for a subgroup of 26 patients (12 in the protein group and 14 in the BCAA group) after the exclusion of 11 patients who had been able to tolerate more than 40 gm protein per day prior to the study and who, therefore, did not satisfy all criteria of inclusion. Data were also analyzed for a subgroup of 18patients (eight patients in the protein group and 10 in the BCAA group) whose physicians adhered to all details of the protocol and who consumed >75% of the prescribed diets. The results in the three analyses which dealt with 37, 26 and 18 patients, respectively, were similar in all respects. The

282

HORST ET AL.

results from the 26 patient subgroup are presented in detail. Antiencephalopathic agents, such as lactulose or neomycin had been discontinued at the beginning of the basal period in all 26 patients. Two patients, one in each group, became encephalopathic during the basal period on 20 gm dietary protein per day. Each was restarted on lactulose, which was continued throughout the study, and the basal period was repeated. The patients in the protein and BCAA groups were similar in age, sex, weight, type and duration of iiver disease and severity and duration of PSE (Table 2). They had similar physical signs of chronic liver disease and their laboratory data were similar (Table 3). Although the mean total serum bilirubin, alkaline phosphatase, transaminase and prothrombin levels were slightly more abnormal in the protein group and the serum albumin slightly lower in the BCAA group, only the SGOT was significantly different (p < 0.05).

HEPATOLOGY

TABLE 2. COMPARISON

OF

CLINICAL FEATURESOF PROTEIN

AND

A662 GROUPS

No. of patients Mean age in years (range) Men/women Type cirrhosis Alcoholic Posthepatitic Cryptogenic Miscellaneous Duration of cirrhosis (mean) in years Portal-systemic anastomosis Duration of protein intolerance in months (mean and range) Maximal protein tolerance (no. of patients) 20 gm BCAAs 4 1 per day dietary protein, worsened abruptly on the twentieth day, while taking an average of 17 gm protein and a NS, not statistically significant. 25 gm A662 solution. The patient, who had chronic active hepatitis, had been on long-term prednisone and neomycin therapy. Neomycin was discontinued prior to the orous therapy of the urinary tract infection, diabetes and basal period and the steroid dosage tapered to 10 mg hyperosmolar state, the patient remained febrile and died prednisone on alternate days at the end of the basal 10 days later. In retrospect, there had been a progressive period. The sudden development of encephalopathy was increase in serum BUN, creatinine and bilirubin concenassociated with fever and severe hyperglycemia (600 mg trations, alkaline phosphatase activity and prolongation per dl), presumably precipitated by a urinary tract infec- in prothrombin time which began after the end of the tion. Despite withdrawal of all nitrogen intake and vig- basal period. Autopsy revealed inactive, macronodular cirrhosis, subacute portal vein thrombosis, hydronephrosis and chronic interstitial pyelonephritis, cholelithiasis and an incidental meningioma, which apparently did not contribute to his death. It is not known whether encephalopathy was precipitated by ingestion of the protein-A662 mixture, urinary tract infection, hyperosmolar hyperglycemia, portal vein thrombosis, tapering of adrenocortical therapy, or whether the BCAA solution aggravated derangements of carbohydrate metabolism which occur in cirrhotic patients with diabetes (30-34). Mental status, which is depicted graphically for each patient in Figure 3, worsened in most of the patients in NITROGEN the protein group and remained unchanged or improved BALANCE in all except one of the BCAA group. In the protein 9 Per day group, the mean mental state grade increased from 0.6 to 1.4 (p < 0.05) while in the BCAA group the mean mental state decreased insignificantly from 0.8 to 0.5. The posttherapy mean level was significantly worse in the protein group than in the BCAA group (p < 0.025). In the protein group, asterixis worsened progressively from a mean grade at randomization of 0.2 to a final grade of 1.0 (p < 0.05) (Table 6). In the A662 group, the degree of asterixis remained relatively low (mean score 0.3 before, 0.2 after). There were no significant differences in the NCT I 0' I I 1 during the study in either the protein or BCAA groups. BASAL I 2 3 4 The mean test time became slightly longer in both groups WEEKS (Table 6). FIG. 2 . Comparative effect of dietary protein and A662 on nitrogen There were no statistically significant differences in balance. Mean ( 2 S.D.) nitrogen balance is shown for the protein group the mean EEG grades in either group during the period (solid circles) and for the A662 group (open circles). Although nitrogen of study, although the mean EEG grade tended to worsen balance is more positive with dietary protein at 3 weeks (40 gm per day) and 4 weeks (60 gm per day), the differences are not statistically in the protein group and improve in the BCAA group (Table 6). significant.

-3i I '

HORST ET AL.

284

Mean blood ammonia levels were similar in the two groups at the end of the basal period and had increased slightly in both groups by the end of the treatment period. None of the differences is statistically significant (Table 6). At the time of randomization, the mean PSE Index was 21 and 24% in the protein and BCAA groups, respectively, an insignificant difference. After therapy, the PSE Index had risen significantly in the protein group (p < 0.01) and had fallen slightly in the BCAA group (Table 6). The difference between the PSE Index in the two groups at the termination of the study is highly significant statistically (p < 0.005). Data were available for 16patients, eight in each group (Table 7). The baseline amino acid profiles, which were similar in the two groups, showed the characteristic increased concentrations of tyrosine, phenylalanine, and methionine and low levels of leucine, isoleucine and valine.

PROTEIN GROUP

MENTAL STATE GRADE

i

31

TABLE6. EFFECTOF PROTEIN AND A662 ON VARIOUS OF THE PSE SYNDROME COMPONENTS Component

Mental state (grade) Protein BCAAs Asterixis (grade) Protein BCAAs NCT (seconds) Protein BCAAs EEG (grade) Protein BCAAs Blood ammonia (rg) Protein BCAAs PSE Index (%) Protein BCAAs

Mean value at randomization

Mean value at end of study

Statistical significance (before vs. after)

** c NS

0.6 f 0.5 NS” 0.8 f 0.5

1.4 f 1.0 0.5 f 0.7

* b

0.2 f 0.7 NS 0.3 f 0.5

1.0 & 1.2 0.2 & 0.4

***d

78 f 49 NS 67 f 29

82f43 NS 65 & 65

NS NS

0.7 f 0.9 NS 1.3 f 0.9

1.4 & 1.3 NS 1.0 & 0.8

NS NS

166 f 49 NS 169 f 51

1 8 5 k 1.3 NS 187 2 98

NS NS

21 f 13 NS 2 4 k 13

3 8 k 23 17 17

***d

NS

* b

**c

NS

NS, not statistically significant. *, p < 0.01. ‘**, p < 0.005. ***, p < 0.05.

At the end of therapy, the BCAA group showed statistically significant increases of the BCAAs and threonine and ornithine as well as significant decreases in tyrosine and glutamine. The BCAA/AAA molar ratio had risen from 1.48 to 2.95, a significant increment (p < 0.001). In the protein group, there were significant increases of phenylalanine and arginine only, and the molar ratio remained unchanged (1.36 before to 1.37 after). SGPT activity, which had been higher in the protein group than in the BCAA group at randomization, increased further in the protein group and decreased in the BCAA group during therapy. The SGOT decreased slightly in both groups. As expected, the mean BUN increased by 5 mg per dl in the protein group (p < 0.01) and by 12 mg per dl in the BCAA group (p < 0.01). The magnitude of these increments was not significantly different. Creatinine levels remained constant in both groups. No other biochemical or hematological tests changed significantly. Three patients in the protein group and seven in the BCAA group developed nausea, a statistically insignificant difference. Vomiting occurred in two patients in the BCAA group. Several patients in each group complained of abdominal fullness, cramps and diarrhea. Hyperglycemia was observed in one patient, whose course and subsequent death are previously described.

WEEKS BCAA GROUP

MENTAL STATE GRADE I

0 PERIOD

HEPATOLOGY

’

3

WEEKS FIG.3. Comparative changes in mental state in dietary protein and BCAA groups. Each line depicts the mental state grade of one patient, at the end of the basal period (date of randomization)connected to the mental state grade at the end of 3 weeks of therapy or the time of development or encephalopathy.

DISCUSSION This investigation shows that A662, a mixture of amino acids enriched with BCAAs administered as an oral supplement to patients with chronic PSE is nutritionally effective and precipitates PSE less frequently than equivalent amounts of dietary protein. Using diets

Vol. 4, No. 2,1984

COMPARISON OF DIETARY PROTEIN IN CHRONIC ENCEPHALOPATHY

285

TABLE7. FASTINGPLASMAAMINOACID LEVELS Protein group (n = 8 )

Alanine Arginine Glutamic acid Glycine Histidine Isoleucine Leucine Lysine Methionine Ornithine Phenylalanine Proline Serine Taurine Threonine Tyrosine Valine Molar ratio (BCAAs:AAAs)

Prestudy

33.9 13.4 38.5 34.6 8.9 5.3 8.6 12.5 5.4 8.2 8.9 21.1 15.0 5.7 14.5 10.5 12.1 1.36

S'E'

Poststudy

3.4 1.2 9.0 2.1 0.6 0.4 0.7 0.8 0.4 1.3 0.8 1.8 1.o 0.6 1.4 0.6 1.2 0.11

32.7 15.5 37.9 31.8 9.2 5.7 8.4 15.1 5.2 9.3 9.9 24.8 15.3 6.7 14.1 11.8 14.8 1.37

BCAA group (n = 8) S'E.

Prestudy

4.9 0.8" 8.4 1.7 0.5 0.4 0.9 1.7 0.4 1.0 0.9" 2.4 1.6 0.7 2.3 1.1 1.7 0.18

41.6 12.3 40.1 36.1 10.0 4.8 9.0 18.4 5.7 7.9 8.9 25.5 15.8 5.9 17.0 12.2 14.2 1.48

"*'

Poststudy

S.E.

3.3 1.6 11.0 1.6 0.8 0.2 0.9 1.5 0.4 0.4 1.2 2.5 1.6 0.6 1.6 1.3 1.3 0.25

36.8 13.2 31.7 40.7 11.2 7.0 11.5 22.4 5.2 12.4 7.6 34.1 18.4 5.8 22.8 7.4 21.0 2.95

2.8 2.3 10.4 4.3 1.2 0.7" 1.2" 2.8 0.4 1.8" 0.9 4.7" 2.4 1.0 2.6" 1.1" 1.7' 0.41"

Normal (Mean f S.D.)

33.5 +- 7.1 9.2 & 2.1 1.9 f 1.2 24.5 & 7.1 8.0 & 0.8 6.8 k 2.6 12.7 & 3.0 16.5 -+ 3.4 2.9 f 0.6 4.3 f 1.2 5.3 f 0.8 20.0 -+ 6.3 11.7 & 2.6 4.3 & 0.5 14.3 & 3.3 6.0 f 1.0 22.1 & 4.4 3.7 & 0.7

All values are expressed as milligrams per deciliter. Before and after difference p < 0.05. * Before and after difference p < 0.001. a

which were essentially equinitrogenous, equicaloric, equiosmolar and equivolumic and which contained adequate calories from carbohydrate and fat, we compared A662 with dietary protein in a selected group of cirrhotic patients who had had recurrent episodes of PSE that had required the restriction of dietary protein and the administration of lactulose and/or antibiotic agents. At the time of study, PSE was minimal or absent. Gradually increasing amounts of protein precipitated or worsened PSE in about half of the patients on dietary protein and in only one of those receiving A662. Worsening of mental state, asterixis, and the PSE Index occurred in the dietary protein group while a tendency toward improvement was observed in the A662 group. These data should not be interpreted to mean that A662 is effective in the treatment of PSE. The investigation was specifically designed to compare the comagenic potential of dietary protein and A662 in patients who were susceptible to hepatic encephalopathy. Indeed, these data demonstrate objectively the coma producing properties of protein. Furthermore, the ingestion of increasing amounts of both protein and A662 progressively reversed the negative nitrogen balance of these patients to a similar degree despite the lower amount of available nitrogen and the theoretically lower nutritional efficiency of A662. Despite complaints about taste and mild gastrointestinal symptoms in both groups, which were attributed to the hyperosmolarity of the A662 and the "placebo" solution, most patients tolerated both diets well. The most serious complication of the study occurred in a diabetic patient receiving A662, who became septic, hyperglycemic, and encephalopathic and subsequently died. These observations were made using a blind format in randomly selected patients who remained in the hospital throughout the trial. There were no known breaches of

blindness in any physician. Diets were administered by research dietitians, and the amounts of protein or protein plus A662 ingested, which were determined by the weight consumed, were similar for the two groups at each stage of the study. This positive analysis does not imply that the investigation was free of flaws. First, too few patients were studied. Our sample size calculations indicated that 80 patients would be required, but the appearance of A662 (Hepatic Aid'") on the commercial-market in effect terminated the investigation. Second, several violations of the protocol occurred. Patients who were able to tolerate more than the arbitrary 40 gm protein per day were erroneously admitted; in some, the amount of protein administered in the basal period was arbitrarily doubled; in others, medications for PSE, such as lactulose or neomycin, were continued throughout the period of therapy. All such violations were excluded from the analysis of 26 patients who satisfied all criteria. Third, the 7-day periods of observation may have been too short to observe the full effect of changes in protein intake, but longer periods were not considered practical. Fourth, a significant number of data were missing or unusable. EEGs were not obtained each week at all institutions, were not always performed on the last day of each period and were not available for blind evaluation at one hospital. At several institutions, the blood ammonia assay was performed on venous blood, a less reliable assay than arterial blood (35), and was not quantified as frequently as the protocol required. Because of technical difficulties with the blood ammonia assay, all blood ammonia determinations from one hospital, which contributed a large fraction of the patients, were excluded. These problems may have accounted for the failure of the blood ammonia levels to increase signifi-

286

HORST ET AL.

cantly in the patients with protein-precipitated PSE, an almost invariable association (1-4). The NCT was administered to some patients using the same pattern repeatedly, a procedure which permits the patient’s performance to improve progressively and thus diminishes the comparative value of the test (28). Furthermore, in most instances, the test was terminated at 120 sec, a practice which systematically minimizes the test score in the presence of encephalopathy. These artifacts are probably responsible for the apparent failure of the NCT to confirm more gross deterioration of mental state. Finally, this investigation was conducted for the most part without the use of other forms of therapy, such as lactulose or neomycin, that would probably have been successfully administered in actual clinical practice. Until recently, the use of BCAAs in the management of PSE has been limited to uncontrolled, anecdotal observations (16,18,21).Only three randomized, controlled clinical trials of the intravenous administration of BCAAs have been reported, two of them in preliminary form. Rossi-Fanelli et al. compared a BCAA solution (42 to 56 gm leucine isoleucine valine per day) with lactulose in 40 patients with acute hepatic encephalopathy precipitated primarily by infection or hemorrhage from varices (19). They found BCAA administration to be at least as effective as lactulose. Unfortunately, the groups were not nutritionally equivalent. Furthermore, there is no true control group since the use of antibiotics may have adversely affected the efficacy of lactulose (Conn, H. O., Arch. Intern. Med. 1981; 141:846; Deresinski, S. C., Arch. Intern. Med. 1982; 142:1406,Correspondence) and the administration of large volumes of hypertonic glucose may have affected patients bleeding from varices. Wahren and co-workers compared a BCAA solution (40 gm per day) with simple glucose administration for 5 days in 52 cirrhotic patients with acute PSE (36). Improvement in encephalopathy was observed in half of both groups (52 and 48%, respectively), but survival was better in the placebo group (80%) than in the BCAA group (52%)(p < 0.05). More recently, Cerra and associates reported on 59 patients in whom intravenous F080, an intravenous solution of amino acids identical to A662, was compared with neomycin in the management of acute PSE (Hepatology 1982; 2:699, Abstract). Thirty cirrhotic patients received 40 to 120 gm of the BCAA-enriched solution by central venous catheter and 29 received neomycin (4 gm per day) plus glucose by central venous catheter for 4 to 11 days. Hepatic encephalopathy was effectively treated with F080 more frequently than with neomycin (53 vs. 1796, respectively; p < 0.01) and survival was greater in the F080 group (83 vs. 45% respectively; p < 0.01). Negative nitrogen balance persisted in the neomycin group (-8 gm per day), but was restored to normal with the amino acid solution (+0.1 gm per day). The typical PSE plasma amino acid pattern, which was present in both groups before therapy, remained abnormal in the neomycin group, but was restored to normal by the administration of F080.

+

+

HEPATOLOGY

Thus, the three studies of BCAA administration in acute hepatic encephalopathy used different experimental designs, administered different amounts of different types of BCAA solutions, and came to different conclusions. A similar pattern exists between the results of the present investigations and the only other randomized controlled clinical trials in which oral BCAAs were administered to cirrhotic patients with chronic PSE (23, 24). In the first investigation, Ericksson and her associates studied seven cirrhotic patients with minimal PSE at the time of study, most of whom were hospitalized and taking 40 to 100 gm dietary protein per day (23). Their patients were given either 30 gm of powdered BCAA (21 gm leucine, 6 gm valine and 3 gm isoleucine) or a placebo powder for 14 days in a cross-over design, followed by the “other” therapy for an additional 14 days. Improvement was noted in two patients after placebo (29%)and three (43%) after BCAAs, an insignificant difference. One patient worsened with each type of treatment. No difference was noted in the EEG patterns or in psychometric tests with the two forms of therapy. There were no changes in the typical PSE amino acid pattern when measured 12 hr after the last dose of BCAAs, although a transient postprandial increase of plasma BCAA levels indicated that the amino acids had been absorbed. Nitrogen balance was not measured. McGhee et al. studied four patients with chronic PSE, each of whom received either a 50 gm casein diet or 20 gm casein plus 30 gm of a BCAA-enriched solution, the same one used in the present study, for 11 days each in a cross-over study (24). They noted no significant changes in clinical status, EEG, or psychometric tests between the two diets. The fasting amino acid pattern became normal in the BCAA-enriched diet, and nitrogen balances were equal on the two diets. These investigations of oral BCAA solutions come to different conclusions. This may be due to several factors. First, our study was designed to challenge patients by gradually increasing the amounts of two different sources of nitrogen until encephalopathy developed, whereas Ericksson and McGhee used a constant level of protein intake. Second, Eriksson and McGee used a cross-over design and studied smaller numbers of patients. Third, the components of the diets were different. Eriksson’s patients received a larger dose of BCAA than did ours (30 gm vs. 8 to 21 gm per day), significantly more leucine (21 gm vs. 3 to 8 gm per day), and two to five times as much dietary protein. One fundamental difference between the negative Swedish and positive American studies is evident. Wahren and Eriksson and their associates administered large amounts of pure leucine, isoleucine and valine, whereas Cerra et al. and our group administered a mixture of essential and nonessential amino acids that contained increased proportions of leucine, isoleucine and valine but in smaller amounts. Whether or not the improved tolerance reported in Cerra’s study and ours is related to the particular balance of the amino acids administered is not known. Further investigations are needed to confirm and explain these diverse observations.

Vol. 4, No. 2, 1984

COMPARISON OF DIETARY PROTEIN IN CHRONIC ENCEPHALOPATHY

Acknowledgments: We would like to acknowledge the assistance of Margie Vena, dietitian at the Lemuel Shattuck Hospital for her design and administration of the hospital diets, as well as to Eula Fields, dietitian at the West Haven Veterans Administration Medical Center, for administration of the diets. We would also like to acknowledge Dr. Thomas Noble for his help in the design and initiation of the study. REFERENCES 1. Phillips GS, Schwartz R, Gabuzda GJ, et al. The syndrome of impending hepatic coma in patients with cirrhosis of the liver given certain nitrogenous substances. N Engl J Med 1952; 247239-246. 2. Summerskill WHJ, Wolfe SJ, Davidson CS. The management of hepatic coma in relation to protein withdrawal and certain specific measures. Am J Med 1957; 2359-76. 3. Mutchnick MG, Lerner E, Conn HO. Portal-systemic encephalopathy and portal anastomosis: a prospective, controlled investigation. Gastroenterology 1974; 66:1005-1019. 4. Bessman AN, Mirick GS. Blood ammonia levels following the ingestion of casein and whole blood. J Clin Invest 1958; 37:990998. 5. Fenton JCB, Knight EJ, Humpherson PL. Milk and cheese diet in portal-systemic encephalopathy. Lancet 1966; 1364-166. 6. Greenberger NJ, Carley J , Schenker S, et al. Effect of vegetable and animal urotein diets in chronic enceDhaloDathv.Am J Die. Dis 1977; 22:945-855. 7. De Bruijn KM, Blendis LM, Zilm DH, et al. Effect of dietary manipulations in subclinical portal-systemic encephalopathy. Gut 1983; 2453-60. 8. Wu C, Bollman JL, Butt HR. Changes in free amino acids in the plasma during hepatic coma. J Clin Invest 1955; 34845-849. 9. Iber FI, Rosen H, Levenson SH, et al. The plasma amino acids in patients with liver failure. J Lab Clin Med 1957; 50417-425. 10. Ning M, Lowenstein LJ, Davidson CS. Serum amino acid concentrations in alcoholic hepatitis. J Lab Clin Med 1967; 70554-562. 11. Fischer JE, Yoshimura N, Aguirre A, et al. Plasma amino acids in patients with hepatic encephalopathy: Effects of amino acid infusions. Am J Surg 1974; 127:40-47. 12. Fischer JE, Funovics JM, Aguirre A, et al. The role of plasma amino acids in hepatic encephalopathy. Surgery 1975; 78:276-290. 13. Morgan MY, Hilson JP, Sherlock S. Plasma ratio of valine, leucine, and isoleucine to pheylalanine, and tyrosine in liver disease. Gut 1978; 19~1068-1073. 14. Cascino A, Cangiano C, Calcaterra V, et al. Plasma amino acids imbalance in patients with liver disease. Am J Dig Dis 1978; 23591598. 15. Marchesini G, Zoli M, Dondi C, et al. Prevalence of subclinical hepatic encephalopathy in cirrhotics and relationship to plasma amino acid imbalance. Dig Dis Sci 1980; 25763-768. 16. Fischer JE, Rosen HM, Ebeid AM, et al. The effect of normalization of plasma amino acids on hepatic encephalopathy in man. Surgery-1976;8077-91. 17. Eeberts EH. Hamster W. Jureens. . et al. Effect of branched chain amino acids' on latent portal systemic encephalopathy. In: Walser M, Williamson J , eds. Metabolism and clinical implications of branched chain amino and ketoacids. Amsterdam: ElseviedNorth Holland, 1980: 453-463.

-

. "

287

18. Freund H, Dienstag J , Lehrich J, et al. Infusion of branched-chain enriched amino acid solution in patients with hepatic encephalopathy. Ann Surg 1982; 196209-220. 19. Rossi-Fanelli F, Riggio 0, Cangiano C, et al. Branched-chain amino acids vs lactulose in the treatment of hepatic coma: a controlled study. Dig Dis Sci 1982; 27929-935. 20. Milliken W, Henderson M, Warren D, et al. Total parenteral nutrition with F080Rin cirrhotics with subclinical encephalopathy. Ann Surg 1983; 197294-304. 21. Freund H, Yoshimura N, Fischer J. Chronic hepatic encephalopathy-Long-term therapy with a branched-chain amino acid enriched elemental diet. JAMA 1979; 242:347-349. 22. Schafer K, Winther MB, Ukida M, et al. Influence of an orally administered protein mixture enriched in branched chain amino acids on the chronic hepatic encephalopathy (CHE) of patients with liver cirrhosis. Z. Gastroenterol 1981; 19356-362. 23. Eriksson LS, Persson A, Wahren J. Branched-chain amino acids in the treatment of chronic hepatic encephalopathy. Gut 1982; 23~801-806. 24. McGhee A, Henderson M, Millikan W, et al. Comparison of the effects of hepatic-aid and a casein modular diet on encephalopathy, plasma amino acids, and nitrogen balance in cirrhotic patients. Ann Surg 1983; 197:288-293. 25. Conn HO. Nutritional management of advanced liver disease. In: Winters RW, Greene HL, eds. First Bristol-Myers Symposium on Nutritional Research Nutritional support of the seriously ill patient. New York Academic Press, 1982. 26. Conn HO, Leevy CM, Vlahcevic ZR, et al. Comparison of lactulose and neomycin in the treatment of chronic portal-systemic encephalopathy. Gastroenterology 1977; 72573-583. 27. Blackburn GL, Bistrian BL, Mainis B, et al. Nutritional and metabolic assessment of the hospitalized patient. JPEN 1970; 1:1122. 28. Conn HO. The trailmaking and Number Connection Tests in the assessment of mental state in portal systemic encephalopathy. Am J Dig Dis 1977; 22541-550. 29. Seligson D, Ishihara K. The measurement of ammonia in whole blood, erythrocytes, and plasma. J Lab Clin Med 1957; 49:962-974. 30. Conn HO, Schreiber W, Elkington SG. Association of impaired glucose tolerance with portal systemic shunting in Laennaec's cirrhosis. Dig Dis 1979; 16227-239. 31. Sherwin RS, Fisher M, Bessoff J , et al. Hyperglucagonemia in cirrhosis: altered secretion and sensitivity to glucagon. Gastroenterology 1978; 74:1218-1223. 32. James JH, Jeppson B, Ziparo V, et al. Hyperammonaemia, plasma, amino acid imbalance, and blood-brain amino acid transport: a unified theory of portal systemic encephalopathy. Lancet 1979; 2~772-775. 33. Marchesini G, Forlani G, Zoli M, et al. Insulin and glucagon levels in liver cirrhosis. Relationship with plasma amino acid imbalance of chronic hepatic encephalopathy. Dig Dis Sci 1979; 24594-601. 34. Soeters PB, Fischer JE. Insulin, glucagon, amino acid imbalance and hepatic encephalopathy. Lancet 1976; 2:880-882. 35. Conn HO, Lieberthal MM. Blood ammonia concentration. In: The hepatic coma syndromes and lactulose. Baltimore, Maryland Williams and Wilkins, 1979 69-76. 36. Wahren J , Denis J, Desurmont P, et al. Is intravenous administration of branched chain amino acids effective in the treatment of hepatic encephalopathy? A multicenter study. Hepatology 1983; 3~475-480.