Metabolic Response to Hypertonic Glucose Administration in Reye Syndrome Morey W. Haymond, M D , Irene E. Karl, PhD, James P. Keating, MD, and Darryl C. DeVivo, MD

Blood substrate and hormone concentrations were determined in 16 children with Reye syndrome prior to and following administration of hypertonic glucose. Baseline concentrations of lactate, pyruvate, alanine, glutamine, glutamate, proline, hydroxyproline, lysine, and aspartate were elevated ( p < 0.01), whereas citrulline and arginine were low. All substrate concentrations were below or within the normal range following 36 hours of therapy except those of lactate,pyruvate, and aspartate. Urea nitrogen excretion was reduced (p < 0.05) on the second day of therapy. Plasma concentrations of insulin and growth hormone increased and glucagon decreased during the first day. Cortisol remained elevated throughout the study period. We conclude that the high circulating concentrations of substrates are the result of both increased mobilization and decreased clearance and that hypertonic glucose infusion suppresses substrate mobilization. A primary abnormality of the mitochondria could explain the metabolic perturbations that occurred. A possible relationship between the encephalopathy in this disorder and an insult to both brain and brain capillary mitochondria is discussed. Haymond MW, Karl IE, Keating JP, et al: Mrtabolic response to hypertonic glucose administration in Reye syndrome. Ann Neurol 3:207-215, 1978

A complex variety of metabolic abnormalities have been observed in children with Reye syndrome, including elevations in the circulating concentrations of lactate, free fatty acids, alanine, glutamine, glutamate, and ammonia and decreased plasma concentrations of glucose, cholesterol, triglycerides, total lipids, lipoproteins, citrulline, and hepatic clotting factors [6]. More recently, decreased activities of hepatic ornithine transcarbamoylase (OTC), carbamoyl phosphate synthetase (CPS), pyruvate carboxylase, and pyruvate dehydrogenase have been documented in children during the acute phase of Reye syndrome [3,

33, 361. We report data o n circulating concentrations of metabolic intermediates and of glucoregulatory hormones obtained in 16 patients with Reye syndrome prior t o and for 60 hours following the institution of hypertonic glucose therapy. O u r interpretation of these data suggests that the high circulating concentrations of substrates in this disease are the result of both increased mobilization and decreased clearance, and hypertonic glucose corrects these metabolic abnormalities within 36 to 48 hours by suppression of substrate mobilization. A primary abnormality of the mitochondria could explain the metabolic perturbations just described.

From the Departments of Pediatrics, Medicine, and Neurology and Neurosurgery (Neurology), Washington University School of Medicine, St. Louis, MO.

Materials Sixteen children with clinical Re ye syndrome were treated and studied in a uniform manner. All patients had a history of an antecedent viral illness with a duration of 4 to 10 days (6.7 % 0.4 days, mean 2 1 SEM) and pernicious vomiting for 1 to 4 days (2.1 2 0.2 days) prior to admission. Associated viral illnesses included clinical varicella in 5 patients and serological evidence of recent influenza B in 6 patients and influenza A in 4 patients. When first seen, the children ranged in age from 3 to 15 years (mean, 8.8 t 1.0 years). Elevated serum glutamic oxaloacetic and pyruvic transaminase and creatine phosphokinase activities as well as decreased prothrombin times were documented in all patients [71. At the time of admission the neurological condition of each patient was estimated utilizing a modification 171 of the Huttenlocher staging criteria [17]. Ten patients were admitted in stage I1 comaand 6 in stage 111. Eight of the 10patients in stage I1 coma progressed to stage 111 within 2 to 19 hours after admission (12 ? 2 hours). Three of the 6 children in stage 111 coma progressed to stage IV between 9 and 15 hours after admission and subsequently died. At postmortem examination these 3 patients had fatty infiltration of the liver and. kidneys and cerebral edema. The 13 survivors were examined serially for the earliest signs of clinical improvement and recovery of consciousness, The earliest signs of improvement occurred 25 hours after admission (range, 5 to 68 hours), and recovery of

Accepted for publication Aug 17, 1977.

D~ ~ ~ st, ~~~i~~ Children’s ~ Address reprint requests Hospital, 500 S Kingshighway, St. Louis, M O 63 110.

0364-5 134/78/0003-0304SOl.25 @ 1978 by the American Neurological Association 207

~

consciousness occurred 88 hours after admission (range, 14 to 192 hours). Plasma values from normal children were derived from previously published studies from our laboratory [ 151. The data for normal urinary urea nitrogen were derived from 7 prepubertal normal children between 4 and 10 years old who were on a regular diet.

Methods The patients were admitted to the intensive care unit and received hypertonic glucose solution through a central venous catheter. This solution contained 300 g m of glucose per liter and electrolytes delivered at a rate of 1,600 ml per square meter of body surface area per day. The plasma glucose concentrations were maintained between 200 and 400 mg per deciliter by adjusting the concentrations of glucose in the infusate. In the majority of patients the infused glucose concentration was decreased from 300 to 200 gm per liter between 24 and 36 hours. T h e arterial PO, was maintained between 100 and 200 torr by adjusting the F I O ~ . Further details regarding management, routine chemistry values, and blood gas data on these patients were published previously [7, 2 1I. Four to 5 ml of heparinized peripheral venous blood was obtained from 11 patients at the time of admission. Then arterial samples were taken at 4-hour intervals for the first 60 hours of therapy. Zero-time blood samples were not obtained in 5 patients; however, the remaining samples were obtained as described. The heparinized blood was divided into three tubes: 2 ml of blood was placed in 500 IU of Trasylol (aprotinin) (FBA Pharmaceuticals); 0.5 ml was precipitated with an equal volume of 3 M perchloric acid; and the remaining blood was placed in an additional tube. All samples were immediately placed on ice, centrifuged at 4°C. separated, and stored at -80°C until assayed. Twentyfour-hour urine collections were obtained during the first 48 hours from 2 patients and during the initial 72 hours of therapy from an additional 2 patients. Lactate, pyruvate, P-hydroxybutyrate, acetoacetate, glycerol, alanine, glucose, glutamine, and glutamate were determined utilizing microfluorometric methods [231. Carnitine was determined spectrophotometrically [251. T h e plasma samples obtained from 14 patients at 0,4,12,24,36, 48, and 60 hours following admission were analyzed on a Beckrnan 119 amino acid analyzer utilizing standard physiological methods [15]. Glucagon, insulin, and growth hormone were measured by radioimmunoassays; cortisol was determined by protein binding assay on plasma samples containing Trasylol as previously described [ 151. Urinary urea nitrogen and creatinine were determined utilizing standard methods [ I 51. When it is stated that paired data were analyzed, a paired Student t test was used. For all other comparisons, a twotailed Student t test for unpaired data was utilized [2]. All standard error. data in the text are expressed as mean

acetoacetate (280 -+ 67 pM), and glycerol (160 f 40 pM). Following institution of hypertonic glucose, mean plasma glucose concentrations rose and remained between 290 and 330 mg per deciliter for the first 24 hours of therapy and subsequently decreased to aconcentration between200 and 275 mgper deciliter (Fig 1).Mean plasma insulin values were 22 p U per milliliter at the time of admission, increased in response t o the hypertonic glucose infusion, and then F i g 1 . Glucose, P-hydroxybutyrate, acetoacetate, glyernl, and insulin responses in 16 patients with Reye slndrome prior to and for GO hours following hypertonic glucose therapy (see text for details). The zero-time sample represents only I I patients. Allz'alues are mean ? SE.

3wi

1

-za -

300-

W

$200I-W

Y IW-

a OJ

,

01

,

I

,

.

I

*

Results At the time of admission, the plasma glucose concentration was 137 5 17 mg per deciliter with mild elevations of /3-hydroxybutyrate (814 -t 212 pM),

208

Annals of Neurology

Vol 3 No 3 March 1978

-v

24 48 HOURS FOLLOWING ADMISSION

ranged between 100 and 200 pLJper milliliter (Fig 1). When considered as paired data, in the 11 patients who had zero-time blood obtained, concentrations of P-hydroxybutyrate and acetoacetate were decreased (p < 0.01) following4 hours of therapy (Table l),with maximal suppression of these substrates following 24 hours of hypertonic glucose (Fig 1). Although the mean glycerol concentration decreased, the change was not significant. Plasma concentrations of L-carnitine were 40.6 ? 5.7 pM prior to therapy and 39.2 + 4.4 p M 24 hours after therapy in 13 patients with Reye syndrome. Carnitine concentrations derived from plasma samples of 9 normal children were 39.2 ? 1.7 p M . O n admission the mean glucagon value was elevated (300 t 70 pg/ml) but was not in the concentration range observed in patients with diabetic ketoacidosis [39]. These concentrations were suppressed during

the initial 24 hours of therapy with hypertonic glucose (Fig 2). Growth hormone was detectable in all samples and rose from an initial concentration of 3.3 0.7 to 10.9 t 2.8 ng per milliliter after 4 hours of therapy. Subsequently, growth hormone release was erratic, with a general decrease in the mean plasma concentrations over the 60-hour study (Fig 2). O n e patient not included in Figure 2, who suffered respiratory arrest and died, had a marked increase in growth hormone concentrations (48 to 9 0 nglml) over the last 16 hours of study. This increase may have been secondary to uncal herniation and disruption of the pituitary somatotrophs. Plasma cortisol concentrations (64 2 10 pg/dl) were increased five to fifteenfold above normal physiological values at the time of admission. These concentrations slowly decreased to the mean value of 32 pg per deciliter by the end of the 60 hours of study (Fig 2).

*

Table 1. Paired Ddta Oaer the F i n t I2 Hours of Therupy" Normal Subjects"

Reye Syndrome Patients

Substrate Concentrations

Substrate Concentrations ( p M ) at: Substrate

No.

0 Hr

4 Hr

12 H r

Lactate' Pyruvate' P - H y drox ybutyrate' Acetoacetate' Hy drox yproline Asparrate Threonine Serine Proline Glutamate'! Glutamine' Citrulline Glycine Alanine' Valine Isoleucine Leucine Tyrosine Ornithine Phen ylalanine Lysine Histidine Arginine

11 11 11 11

9,090 t 1,090 278 f 50 814 f 212 280 f 67 352 f 63 68 ? 19 I66 2 l h 203 ? 25 654 t 115 278 f 71 1,557 288 9 2 5 391 ? 27 1,262 362 267 f 37 138 t 30 196 ? 30 262 t 34 104 t 18 187 -C 24 902 f 91 113 f 10 39 t 17

8,880 2 800 322 f 44 439 f 108' 189 f 51' 337 f 54 53 f 10 141 ? 13 142 f 25 663 t 132 364 ? 54 2,043 f 277 16 f 7 349 39 1,919 ? 350' 200 f 35' 90 34' 124 t 24' 231 f 36 88 f 9 140 f 21' 924 f 112 113 15 70 f 3 1

8,170 840 232 f 34" 211 ? 47"' 81 f 26('.' 255 f 37 47 2 8 114 16' 138 f 21 473 _t 104' 213 f 45d 1,441 ? 235d 18 2 5 285 f 46' 1,467 373" 149 f 26' 76 f 17' 101 f 16' 169 f 22' 98 18 137 ? 20' 655 f 127' 99 t 12 38 f 10

9 9 9 9 9 11 11

9 9 11 9

9 9

9 9 9 9 9 9

*

*

*

*

*

*

No.

(pM)

15 15 15 15

1,110 f 160 77 +- 9 323 ? 100 125 +- 38 217 ? 32 14 f 2 126 f 12 149 +. 14 182 +- 11 62 f 16 571 2 6 1 100 f 5 194 f 15 305 f 38 246 +. 24 72 f 5 109 f 12 58 f 10 53 ? 8 48 ? 4 168 f 24 78 f 5 82 f 16

6 6 6 6

6 15 15

6 6 15 6 6

6 6

6 6 6 6 6

aIncludesdata on patients who had blood samples drawn prior to institution of hypertonic glucose therapy. Data represent mean ? SE. Statistics performed by paired I. test. bNormal children following an overnight fast. 'Determined microfluorometrically. dp < 0.01 from 4-hour value. 'p < 0.01 from 0-hour value. p ' < 0.05 from 0-hour value. *p < 0.05 from 4-hour value.

Haymond et al: Hypertonic Glucose in Reye Syndrome

209

tional 8 hours of hypertonic glucose administration. The mean circulating concentrations of these amino acids continued to fall to within the normal range by 36 hours of therapy (see Table 2 , Fig 3). Blood pyruvate concentrations increased over the first 8 hours of therapy, although the rise was not statistically significant (see Table 1, Fig 3 ) . The pyruvate concentration was 322 44 p M at 4 hours but was significantly lower 12 hours after admission ( 2 3 2 f 34 p M ; p < 0.05)(Table 1 ) . Blood lactate concentrations slowly decreased following institution of hypertonic glucose, and the concentration after 24 hours of therapy was lower than the admission value (9.09 ? 1.09 m M a t zero time, 5.69 -t 1.01 m M a t 2 4 hours;) < 0.05)(see Fig 3 ) . Throughout the period of study, lactate concentrations remained higher than normal values (3.20 ? 0.29 m M at 60 hours vs 1.00 0.76 m M for normal children;) < 0.001), whereas pyruvate concentrations were similar to normal values at 60 hours. The lactate/pyruvate ratios were consistently higher in the patients with Reye syndrome compared with normal children (Reye, 43.2 +. 8.0 at zero time and 43.7 ? 9.4 at 12 hours, vs normal, 15.2 ? 1.4;p < 0.01).In contrast, the P-hydroxybutyrate/ acetoacetate ratios in the patients with Reye syndrome were similar to those of the normal children throughout the study period. Mean plasma concentrations of phenylalanine, tyrosine, threonine, glycine, isoleucine, leucine, and valine were mildly elevated at the time of admission, did not increase following institution of hypertonic glucose, and were significantly lower 4 to 12 hours after admission (see Tables 1 , 2 ) . All these amino acids were within the normal limits 24 hours after admission with the exception of tyrosine and phenylalanine. Mild elevations of plasma histidine and ornithine were observed at zero time and returned to normal over the initial 24 hours of therapy. Plasma arginine and citrulline concentrations were very low, and no striking change occurred during the period of study. Plasma aspartate concentration was elevated fivefold above normal at the time of presentation. Following institution of therapy the mean concentration decreased, but at no time was the fall statistically significant, and the concentrations remained elevated twofold to threefold above normal values even after 60 hours of therapy (30 ? 8 pM at 60 hours vs 14 ? 2 p M for normal chi1dren;p < 0.05) (see Table 2 ) . Urinary urea nitrogen excretion was greatest for each child studied during the initial 24 hours of therapy (Fig 4 ) and was reduced significantly during the second day (292 39 and 128 5 2 1 mmoles per 24 hours, respectively; p < 0.05 by paired t test). When measured in 2 patients, a further decrease in the excretion of urea nitrogen was observed on the third +_

I I-

3

*

u

o

r

0

1

24 48 HOURS F O L L O W I N G A D M I S S I O N

Fig 2 . Plasma glut-agon,groumth hormone, and cortisol re.rponses in patients with Reye Jyndrorne prior to and for 60 hours folloiring hypertonil. glucose therap.y (see text for details). The zero-time sample represents only 11 patients. One patient who died was excludedfrom the growth hormone data beiause her ralues were 48 to 90 ngper mzlliliter o w r the last 16 hours of study. All zmalues are mean S E .

*

Eleven potential gluconeogenic substrates were observed to be significantly higher () < 0.01) than the respective concentrations in 6 normal children who fasted overnight (see Table 1).These were lactate, pyruvate, glutamine, glutamate, alanine, proline, lysine, phenylalanine, tyrosine, glycine, and aspartate. Three patterns of substrate response to hypertonic glucose occurred: ( 1 ) a transient rise in substrate concentrations, followed by suppression (pyruvate, alanine, glutamine, glutamate, proline, and lysine); ( 2 ) a decrease in substrate concentrations (lactate, /3-hydroxybutyrate, acetoacetate, glycerol, hydroxyproline, threonine, glycine, isoleucine, leucine, valine, tyrosine, phenylalanine, ornithine, and histidine); and ( 3 ) no significant change (aspartate, citrulline, and arginine) (Tables 1 , 2 ; Figs 1 , 3). When just paired data are considered, the concentrations of plasma glutamine, glutamate, alanine, proline, and lysine increased over the first 4 hours of therapy. However, only the rise in alanine was significant () < 0.05)(see Table 1).When compared with the 4-hour peak concentrations, all these substrates were significantly lower ( p < 0.001)following an addi-

210 Annals of Neurology

Vol 3

No 3

March 1978

*

Table 2. Plasma A m i n o A c i d Re.rponse

to

Hypertonic GLucnse i n I 4 Patients with Reye Syndrnme

Hours of Therapy

Substrate

0"

4

12

24

H ydroxyproline Aspartate Threonine

317 t 4 9 68 f 19 166 rt 16 203 rt 25 654 ? 116 1,610 t 310 278 rt 71 13 f 5 391 rt 27 1,262 f 362 267 f 37 138 rt 30 187 f 25 262 f 34 187 f 25 104 5 18 902 f 91 113 2 10 39 4 17

339 rt 40 43 2 8 130 2 13 131 f 17 614 2 107 1,834 4 232 305 f 44 18 2 5 329 f 34 1,636 f 268 195 f 29 78 rt 15 122 2 20 221 f 30 131 rt 17 109 rt 20 1,085 f 200 132 f 23 66 f 23

227 f 28 36 f 7 105 14' 132 % 16" 436 2 88 1,296 rt 192 186 f 33 18 2 4 289 2 40 1,213 2 276 139 f 21" 66 4 15" 94 & 14' 157 f 20h 127 -t 17 103 rt 13 625 f 94 92 f 9 33 f 7

185 40

%

65

f

Serine

Proline Glutamine" Glutamated Citrulline GIycine Alanine Valine Isoleucine Leucine Tyrosine Phen ylalanine Ornithine Lysine Histidine Arginine

36

25'' 8 69 t 6c 105 & 8' 150 rt 35' 712 2 150' 115 t 19 12 f 4 207 f 18" 532 f 127b 91 -e 7' 46 f 12' &

5'

8 0 f 9' 99 ? Y 8 6 k 10 273 -+ 45' 74 t 7' 22 t- 7"

*

181 14' 33 f 6 52 i 4" 91 & 7" 100 2 13' 389 f 44" 74 f 10" 8 f 3 173 f 15' 274 f 36" 70 f 6' 42 5 11' 53 2 6 ' 56 2 6' 8 0 t- 5' 66 -+ 7 158 f 22' 63 & 6' 23 5 6

48

60

131 2 11' 31 2 5 56 f 5' 91 C _ 7" 77 f 9 ' 318 37' 68 f 14' 7t-4 188 t 18' 250 ? 26" 79 -t 9' 35 28' 56 f 6 ' 46 2 4' 77 f 5" 74 rt_ 7 135 f 19' 62 f 3' 32 * 4

128 f 13' 30 5 8 79 c 11' 1 1 4 ? 11' 8 1 f 10'~ 348 f 44' 51 2 5' 5 2 2 228 f 23' 293 2 39' 94 f 11' 42 f 10' 66 ? 13" 46 4 ' 73 5 5' 7 1 t 12 156 & 19' 67 f 5' 4 6 c 11

'Zero-time value represents only 9 patients. All data are pM ( 2 S E ) . ' p < 0.05 from admission value. 'p < 0.01 from admission value. "Determined microfluorometrically ( N = 1 1 for zero time; N = 14 for subsequent values).

day (Fig4).A fail in total plasmaamino acid concentration in these 4 patients paralleled the decrease in urinary urea nitrogen excretion during the period of study (Fig 4). (For comparative purposes, the urinary urea nitrogen excretion in 7 normal children on a 2 to 3 gm per kilogram protein diet was 6.92 2 0.92 mg per milligram of creatinine per 2 4 hours, or 169 ? 27 mmoles per 24 hours.) Because of the small number of nonsurviving patients in this study, no correlation could be made between substrate response and clinical outcome. However, the patients who died generally had higher mean concentrations of lactate, pyruvate, alanine, glutamine, and glutamate. Patients admitted in stage I1 coma had substrate concentrations which were lower but not significantly different from levels in patients in srage 111 coma (Table 3). Similarly, there was no correlation between the admission Pao, and lactate concentrations. In the surviving patients, no correlation was found between the time when neurological improvement or complete recovery was first noted and substrate concentrations o r substrate response to therapy.

Discussion Mitochondria1 disruption is a constant finding in Reye syndrome. These ultrastructural abnormalities have been observed in various body organs including the

brain [30], liver [31], skeletal muscle [29], heart, and pancreas [6]. We believe that the chemical abnormalities presented in this study represent the metabolic correlates of these histological observations. T h e disturbed plasma concentrations of amino and organic acids could represent the net result of overproduction by extrahepatic tissues (e.g., brain, muscle, and adipose tissue) and decreased clearance by the liver. Normally, skeletal muscle releases alanine and glutamine in excess of its tissue content [9, 341, and the rate of release of these two amino acids is increased in catabolic states [8, 10, 20, 341. In addition, in vitro muscle incubation studies have documented augmented release of alanine, glutamine, and glutamate in the presence of glucose and ammonium chloride [12]. Patients with Reye syndrome are in a catabolic state as predicted from the clinical circumstances of an antecedent viral infection and poor oral intake, the documented high plasma concentrations of cortisol, glucagon, and growth hormone (see Fig 2), and the elevated urea nitrogen excretion (see Fig 4 ) [36]. Therefore, the elevated plasma concentrations of alanine, glutamine, and glutamate observed in our patients prior to therapy probably reflect, in part, accelerated amino acid release from muscle tissue. The transient increase in plasma concentrations of

Haymond et al: Hypertonic Glucose in Reye Syndrome 2 1 1

2000

J

a

Ti

1

1

300

A

EOW

-H -a

-

w

c

1m-

0

c 4

1 . 1 . 1

I

24

48

i

72

H # m Following Admirrion

3 0

OJ

-H =

300-

I

.

F i g 4 . Total plasma amino ai-id cot~t-r)~tratio~~ ( A A ) (sum of allaniino acids nieusured), in nrnioles prr litrr, total 24-hour urea nitrogen exrrrtion ( U U N ) ,in niniolrs per 24 hours. and urinary ureu nitrogen, in milligruni.! per nig creatinine per 24 hours, drtcrniiued it] 4 patirrrts with Rejr spdronie during hjprrtonii.glut-osr therap.]. Cross-hatched areas represent + 1 SE. Synihob represent indiiiduul pdtient duta.

I

2

A

," 200a 3

LL

a 100-,

-

L L

F i g 3 . Plasma alanine, glutamate. glutaniine. lactate, and pjrui,ate responses i n 16 patient.[ with Reye syndrome prior t o and for 6 0 hours following hypertonic glucose therap.y (see text for details). The zero-time sample represents o d j 11 patients. All iulues are mean & SE. Cross-hatchedareas represent mean 2 SE for plasma imalues i n 15 normal children folloicing an cwernightfast.

these three amino acids and lysine after administration of hypertonic glucose (see Table 1, Fig 2) may be analogous to the stimulated release of amino acids in the presence of glucose and ammonia observed in vitro. This interpretation also could explain the previously described decrease in blood ammonia concentrations [ 17, 181which have been observed during the

212

Annals of Neurology Vol 3

No 3 March 1978

first 24 to 48 hours after hospitalization, since virtually every child with Reye syndrome receives a glucosecontaining solution intravenously. T h e subsequent reduction of plasma amino acid concentrations and the decreased urinary urea nitrogen excretion (see Fig 4 ) also support our impression that in this clinical setting, hypertonic glucose administration opposes muscle protein catabolism, presumably through stimulation of insulin secretion. We also suspect that accelerated production by extrahepatic tissues partly explains the elevated blood concentrations of lactate and pyruvate. Shannon and co-workers [ 3 5 ] demonstrated a net production of lactate by the brain in patients with Reye syndrome, particularly when the blood ammonia concentrations exceeded 300 p g per deciliter. A disturbance in the facilitated transport of organic anions across the mitochondrial membrane following acute ammonia challenge was reported recently in rats with chronic portacaval shunts [ 161. It is possible, therefore, that the increased cerebral production of lactate in Reye syndrome is the result of a structural mitochondria1 injury, a functional disturbance in the malate-aspartate shuttle, or both. The hepatic activities of CPS and O T C have been

Table 3 . Substrate Concentrations of Patients with Reye Syndrome in Stage I1 a n d Stage 111 Coma a t T i m e of Admission

Coma

No.of

Patients

Lactate (mM)

Pyruvate (PM)

Alanine

Stage

I1

9 6

9.44 t 1.71 9.46 t 1.40

247 f 37 363 2 70

895 t 271 1.631 t 530

111

reported to be decreased by 25 to 85% in children with Reye syndrome [3, 361. These decreases in enzyme activities may become apparent clinically in the presence of high rates of substrate delivery [361. Decreased activation of CPS and O T C with normal activities of the cytosolic urea cycle enzymes [3, 361 could account for the decreased plasma concentrations of citrulline and arginine and the increased plasma concentrations of aspartate, glutamine, alanine, and ammonia which have been observed in Reye syndrome. In vitro hepatic perfusion studies demonstrated that the kinetics of gluconeogenic substrate uptake are not saturated until rates of delivery exceed physiological rates by fifty [24]. Therefore, as mentioned, it seems likely that the increased plasma concentrations of lactate, pyruvate, and alanine in our patients result from a defect in hepatic substrate uptake, metabolism, or both, coupled with increased rates of substrate delivery, rather than from the saturation of otherwise intact hepatic mechanisms. There is substantial evidence suggesting a defect in hepatic gluconeogenesis in Reye syndrome. The similarity between the elevated plasma concentrations of gluconeogenic precursors found in Reye syndrome and in children with primary deficiencies of hepatic gluconeogenic enzymes [ 5 , 271 and pyruvate dehydrogenase [191 is noteworthy in this regard. Recently, Robinson and co-workers [33] demonstrated decreased activities of pyruvate dehydrogenase and pyruvate carboxylase in 6 children with Reye syndrome. These enzyme abnormalities probably explain the complex apparent dissociation of the oxidation-reduction potentials between the cytosolic and mitochondrial compartments which is implied by the elevated blood lactate/pyruvate ratios and the normal blood P-hydroxybutyratelacetoacetate ratios. We observed a similar discordance between these two ratios in a child with congenital deficiency of pyruvate carboxylase [ 5 ] .In vitro incubation studies of liver tissue obtained from children with Reye syndrome also documented a defect in glucose formation from pyruvate [13]. These in vivo and in vitro observations suggest that the predisposition for hypoglycemia in patients with this syndrome reflects a disturbance in hepatic gluconeogenesis. A mitochondrial defect also could underlie the observed abnormalities in fatty acid metabolism. Our patients demonstrated only modest ketosis at the time

(PM)

Glutamine (PM) 1,219 ? 280 1.823 t 408

Glutamate (PM) 216 t 72 296 t 61

of admission despite a 36- to 48-hour history of anorexia and vomiting. The characteristic hepatic steatosis and the high plasma concentrations of free fatty acids suggest a hepatic mitochondrial defect in &oxidation. It has been speculated that a deficiency in hepatic biosynthesis of carnitine could explain the fatty acid abnormalities in Reye syndrome [4]; however, our observations of normal plasma carnitine concentrations would argue against this possibility. Administration of hypertonic glucose should oppose peripheral lipolysis by stimulating insulin secretion (see Fig l), resulting in decreased free fatty acid concentrations in plasma. W e interpret the decrease of plasma glycerol concentrations (Fig 1)as a measure of this effect. The pathogenesis of the encephalopathy remains unclear. Some investigators speculate that the cerebral disturbance is primary and associated with variable insults to other organs, while others view the encephalopathy as secondary to hepatic failure. Postulated mechanisms underlying the cerebral disturbance include hyperammonemia [ 181, short-chain fatty acidemia [381, and the de novo synthesis of false neurotransmitters [22]. Fischer and associates [ 111 have presented evidence in support of the false neurotransmitter theory to explain the cerebral disturbance associated with chronic liver failure. They observed an excellent correlation between the plasma (valine + isoleucine + leucine)/(phenylalanine + tyrosine) ratio and the grade of encephalopathy both in animals with portacaval shunts and in humans with cirrhosis. W e observed this ratio to be low also in our patients with Reye syndrome at the time of admission. However, it did not change throughout the clinical course and did not correlate with the clinical staging, as shown in Table 4. Therefore, we conclude that this

+

Table 4 . Ratio of (Valine lsoleucine t Leucine) t o (Tyrosine + PhenylaLanine) i n Reye Syndrome

Time of Determination

Ratio

Admission (N = 15) Clinical nadir (N = 15) Neurological improvement (N = 12) Neurological recovery (N = 12) Normal (N = 6) “p value < 0.001 of Reye syndrome values.

1.28 5 0.14 1.17 ? 0.12 1.33 t 0.14 1.23 ? 0.09 3.94 f 0.21a

Haymond et al: Hypertonic Glucose in Reye Syndrome 213

amino acid ratio has little or no effect in the setting of Reye syndrome. This discrepancy may also imply that other analogies between Reye syndrome and chronic hepatic encephalopathy are equally tenuous. The principal evidence supporting a primary mechanism for the encephalopathy is the uniform nature of the histopathological lesions in brain, liver, and skeletal muscle [29-3 11. Other, less compelling evidence is the poor correlation between the hyperammonemia and the encephalopathy [38] and the fact that the encephalopathy may develop before the onset of the hepatopathy [ 11. In summary, it would appear that most, if not all, of the systemic metabolic abnormalities that we and others have observed in Reye syndrome could be explained by a primary mitochondrial insult involving liver, muscle, and other organs, including the brain. Whether the mitochondrial ipjury also impairs oxidative phosphorylation remains to be clarified. Greene and co-workers [ 141 measured hepatic adenosine triphosphate (ATP) concentrations in children with Reye syndrome (1.59 0.24 pmoleslgm wet tissue) and found these concentrations to be insignificantly lower than in controls (1.81 ? 0.16 pmoles/gm wet tissue). Unfortunately, the concentrations of adenosine diphosphate (ADP) and monophosphate were not measured in these biopsy specimens. Furthermore, the values reported by Greene and associates [ 141 are lower than the hepatic ATP concentrations we have obtained in our laboratory (2.53 2 0.17 pmoleslgm wet tissue) [5,28]. A decrease in the tissue ATPlADP ratio would be an appropriate intracellular stimulus to accelerate glycolysis, with overproduction of pyruvate and lactate in muscle and brain. Decreased ATP synthesis might account for the cerebral edema. Perhaps a more interesting speculation regarding the pathogenesis of the cerebral edema revolves around initial observations of Reye and associates [32] of lipid droplets in the endothelial cells of cerebral capillaries and the recent observation that brain capillaries are particularly rich in mitochondria [26, 321. Oldendorf and co-workers [26] have associated the increased mitochondrial content of capillary endothelial cells in the brain with the large apparent work capability of the blood-brain barrier. If their speculations are correct, a primary insult to mitochondria could lead to a breakdown in the integrity of the blood-brain barrier and explain the cerebral edema seen in Reye syndrome. An impairment in the facilitated transport of glucose across the bloodbarrier might exist in Reye syndrome, analogous to the disparity between the blood and brain glucose concentrations that have been demonstrated in experimental salicylism [37]. Therefore, maintaining a higher plasma glucose concentration to increase diffu-

*

214

Annals of Neurology

Vol 3 No 3 March 1978

sion of glucose into the central nervous system may have obvious benefit. The pathogenesis of the mitochondrial insult in Reye syndrome remains obscure. The low circulating concentrations of hepatic clotting factors and lipoproteins, also seen in this disease, might suggest that the pathogenesis is associated with a transient defect in protein synthesis. Such a defect would be expressed only in those plasma and tissue proteins which have short half-lives (lipoproteins, various coagulation factors, and selected tissue enzymes). Theoretically, such a defect could be secondary to virus-related disruption of normal translational events or consequent to defective production of high-energy phosphate compounds which are necessary to support metabolic and protein synthetic processes within affected tissues. Further investigations along these lines might well provide a better understanding of the pathogenesis of the ubiquitous mitochondrial injury in Reye syndrome. Supported in part by General Clinical Research Grant RR00036,

US Public Health Service Grants H D AM 04355 and AMD 192 1, and a Jerry Lewis Neuromuscular Disease Research Center Granr. We wish to thank Karen Strobel, Dr William Clarke, Thomas Howard, D a n Dallas,Joy Brothers, Dale Osborne. Mary Pat Leckie, and Cathy Whitehead for technical help and Janet North for secretarial assistance. We also thank the St. Louis Children's Hospital house staff. the nursing personnel in the Intensive Care Unit and the Clinical Research Unit, and the clinical laboratory personnel who made an invaluable contribution to the care of these critically ill children.

References 1. Applebaum MM, Thaler MM: Reye syndrome without initial hepatic involvement. Am J Dis Child 131:295-296, 1977 2. Bahn AK: Basic Medical Statistics. New York, Grune & Stratton, 1972 3. Brown T. Hug G , Lansky L, et al: Transiently reduced activity of carbamyl phosphate synthetase and ornithine transcarbamylase in liver of children with Reye's syndrome. N Engl J Med 294:861-867, 1976 4. Cornelio F, DiDonato S,Peluchetti D, et al: Fatal cases of lipid storage myopathy with carnitine deficiency. J Neurol Neurosurg Psychiatry 40:170-178, 1977 5. DeVivo DC, Haymond M W , Leckie PM, et al: The clinical and biochemical implications of pyruvate carboxylase deficiency. J Clin Endocrinol Metab 45:1281-1296, 1977 6. DeVivo DC, Keating JP: Reye's syndrome. Adv Pediatr 22:175-229, 1976 7. DeVivo DC, Keating JP, Haymond MW: Reye syndrome: result ofintensivesupportive care.JPediatr 87:875-880,1975 8. Felig P, Owen OE. Wahren J, et al: Amino acid metabolism during prolonged starvation. J Clin Invest 48:584-594, 1969 9. Felig P, Pozefsky T, Marliss E, er al: Alanine: key role in gluconeogenesis. Science 167: 1003-1004, 1970 10. Felig P, Wahren J: Amino acid metabolism in exercising man. J Clin Invest 50:2703-2714, 1971 11. FischerJE. Rosen HM, Ebeid AM, et al: The effect of normalization of plasma amino acids on hepatic encephalopathy in man. Surgery 80:77-91, 1976

12. Garber AJ, Karl IE, Kipnis DM: Alanine andglutamine synthesis and release from skeletal muscle: I. Glycolysis and amino acid release. J Biol Chem 251:826-835, 1976 13. Glasgow AM, Cotton RB, Dhiensiri K: Reye's syndrome: 111. The hypoglycemia. Am J Dis Child 125:809-811, 1973 14. Greene HL, Wilson FA, Glick AD, et al: Hepatic ATP concentrations and glycolytic enzyme activities in Reye syndrome. J Pediatr 89:777-780, 1976 15. Haymond MW, Karl IE, Weldon VV, et al: The role of growth hormone and cortisone on glucose and gluconeogenic substrate regulation in fasted hypopituitary children. J Clin Endocrinol Metab 425346-855, 1975 16. Hindfelt B, Plum F, Duffy 'IT: Effect of acute ammonia intoxication on cerebral metabolism i n rats with portacaval shunts. J Clin Invest 59:386-396, 1977 17. Huttenlocher PR: Reye's syndrome: relation of outcome to therapy. J Pcdiatr 80:845-850, 1972 18. Huttenlocher PR, Schwartz AD, Klatskin G: Reye's syndrome: ammonia intoxication as a possible factor in the encephalopathy. Pediatrics 43:433-454, 1969 19. Israels S, Haworth JC, Dunn H G , et al: Lactic acidosis in childhood. Adv Pediatr 22:267-303, 1975 20. Karl IE, Garber AJ, I(lpnis DM: Alanine andglutamine synthesis and release from skeletal muscle: 111. Dietary and hormonal regulation. J Biol Chem 258:844-850, 1976 2 1. Keating JP, Haymond MW, Karl IE, et al: Hypophosphatemia in Reye's syndrome, in Pollack JD (ed): Reye's syndrome. New York, Grune & Stratton, 1975, pp 255-259 22. Lloyd KF,Davidson L, Gall DG,etal: Octopamine in the brains of children with acute hepatic encephalopathy (Reye's syndrome). Proc SOCNeurosci 5:394A, 1975 23. Lowry OH, Passonneau JV: A flexible system of enzymatic analysis. New York, Academic Press, 1972 24. Mallette LE, Exton J H , Park CR: Control of gluconeogenesis from amino acids in the perfused rat liver. J Biol Chem 244:57 13-17 19, 1969 25. Marquis MR, Fritz IB: Enzymological determination of free carnitine concentrations in rat tissue. J Lipid Res 5:184-187, 1964 26. Oldendorf WH, Cornford ME, Brown WJ: The large apparent

27. 28.

29. 30.

3 1.

32.

33.

34.

35.

36.

37.

38.

39.

work capability of the blood-brain barrier: a study of the mitochondria1 content of capillary endothelial cells in brain and other tissues of the rat. Ann Neurol 1:409-417, 1977 Pagliara AS, Karl IE, Haymond M W , et al: Hypoglycemia in infancy and childhood: Part 11. J Pediatr 82:558-577, 1973 Pagliara AS, Karl IE, Keating JP, et al: Hepatic fructose 1.6 diphosphatase deficiency: a cause of lactic acidosis and hypoglycemia in infancy. J Clin Invest 51:2115-2123, 1972 Partin JC, Partin JS, Schubert W K Muscle ultrastructure in Reye's syndrome. Pediatr Res 11:564, 1977 Partin JC, Partin JS, Schubert WK, etal: Brain ultrastructure in Reye's syndrome (encephalopathy and fatty alteration of the viscera). J Neuropathol Exp Neurol 34:425-445, 1975 Partin JC, Schubert WK, Partin JS: Mitochondria1 ultrastruccure in Reye's syndrome (encephalopathy and fatty degeneration of the viscera). N Engl J Med 285:1339-1343, 1971 Reye RDK, Morgan G , Baral J: Encephalopathy and fatty degeneration of the viscera-a disease entity in childhood. Lancet 2:749-752, 1963 Robinson BH, Gall DG, Cutz E: Deficient activity of hepatic pyruvate dehydrogenase and pyruvate carboxylase in Reye's syndrome. Pediatr Res 11:279-281, 1977 Ruderman NB, Houghton CRS, Hens R: Evaluation of the isolated perfused rat hindquarter for the study of muscle metabolism. Biochem J 124:639-651, 1971 Shannon DC, DeLong GR, Bercu B, et al: Studies on the pathophysiology of encephalopathy in Reye's syndrome: hyperammonemia in Reye's syndrome. Pediatrics 56:9991004, 1975 Snodgrass PJ, DeLong GR: Urea cylce enzyme deficiencies and an increased nitrogen load producing hyperammonemia in Reye's syndrome. N Engl J Med 2942355-860, 1976 Thurston JH, Pollack PG, Warren SK, et al: Reduced brain glucose with normal plasma glucose in salicylate poisoning. J Clin Invest 49:2139-2145, 1970 Trauner DA, Stockard JJ, Sweetman L EEG correlations with biochemical abnormalities in Reye syndrome. Arch Neurol 34:116-118, 1977 Unger RH: Glucagon and the insu1in:glucagon ratio in diabetes and other catabolic illnesses. Diabetes 20:834-838, 197 1

Haymond e t al: H y p e r t o n i c Glucose in R e y e S y n d r o m e 215