Hypothesis

Lactate is an unreliable indicator of tissue hypoxia in injury or sepsis

J Howard James, Fred A Luchette, Freda D McCarter, Josef E Fischer High blood lactate concentration (hyperlactacidaemia) in trauma or sepsis is thought to indicate tissue hypoxia and anaerobic glycolysis even when blood pressure, cardiac output, and urine output are within clinically acceptable ranges. However, mechanisms of lactate generation by well-oxygenated tissues have received little attention. Within cells, oxidative and glycolytic energy production can proceed in separate, independent compartments. In skeletal muscle and other tissues, aerobic glycolysis is linked to ATP provision for the Na+-K+ pump, the activity of which is stimulated by epinephrine. In injured patients, hypokalaemia may reflect increased Na+, K+-ATPase activity. We propose that increased blood lactate often reflects increased aerobic glycolysis in skeletal muscle secondary to epinephrine-stimulated Na+, K+-ATPase activity and not anaerobic glycolysis due to hypoperfusion. The hypothesis explains why hyperlactacidaemia often neither correlates with traditional indicators of perfusion nor diminishes with increased oxygen delivery. When other variables have returned to normal, continued attempts at resuscitation based on elevated blood lactate may lead to unnecessary use of blood transfusion and inotropic agents in an effort to increase oxygen delivery and lactate clearance. Increased blood lactate concentration (hyperlactacidaemia) in injury or sepsis is traditionally attributed to anaerobic glycolysis due to inadequate oxygen delivery. According to received wisdom, increased blood lactate signals hypoxia and hypoperfusion. Tissue hypoxia is thought to contribute to the development of multiple organ failure or death, and prompt restoration of cellular oxygen delivery is a fundamental goal of therapy. The usual indicators of adequate tissue perfusion have been normalisation of blood pressure, cardiac output, and urine output. However, these clinical variables are believed to be misleading when they have returned to normal yet circulating lactate remains high. The degree and duration of hyperlactacidaemia have been correlated with the subsequent development of organ failure.1,2 If high blood lactate concentration reveals local oxygen delivery failure that is masked by normalisation of traditional indicators, the monitoring of a physiological endpoint such as lactate might be a better guide to resuscitation.3 Clinical observations have challenged the accepted notion of equating increased blood lactate with hypoperfusion.4 Nonetheless, current thinking continues to interpret hyperlactacidaemia as hypoxia and to support stimulation of cardiac output and enhancement of oxygen delivery as therapy.5 It is accepted that blood lactate concentration may increase (albeit modestly) through illdefined mechanisms unrelated to tissue-oxygen debt. Such mechanisms have been described vaguely as involving disparity between peripheral glycolytic rate and mitochondrial oxidative capacity, either in the liver or elsewhere.5 Concepts of such ambiguity are difficult to evaluate experimentally or to apply clinically. Although hypoxic tissues have high rates of anaerobic glycolysis and lactate production, well-oxygenated tissues
Lancet 1999; 354: 50508
Department of Surgery, University of Cincinnati Medical Center, 231 Bethesda Avenue, Cincinnati, OH, USA (J H James PhD, F A Luchette MD, F D McCarter MD, J E Fischer MD) Correspondence to: Dr J H James (e-mail: jamesjh@ucmail.uc.edu)

can also generate lactate through aerobic glycolysisie, glycolysis not attributable to oxygen deficiency. In some situations, ATP production by aerobic glycolysis has been associated with the activity of membrane ion pumps such as the Na+, K+-ATPase. We propose that the epinephrine surge after injury and in sepsis stimulates the sarcolemmal Na+, K+-ATPase and greatly accelerates aerobic glycolysis and lactate production that is coupled to Na+, K+-ATPase activity in skeletal muscle. Consequently, a major proportion of the increase in blood lactate that occurs in injury or sepsis would be unrelated to poor tissue perfusion and unlikely to respond to supranormal oxygen delivery. Persistent hyperlactacidaemia in injured or septic patients who are haemodynamically stable may result more from epinephrine-stimulated aerobic glycolysis than from tissue hypoxia. If our hypothesis is correct, there are profound implications for the usefulness of lactate clearance as an endpoint of resuscitation.

Previous evidence linking hyperlactacidaemia and epinephrine in shock

Nearly 30 years ago, several studies made the existence of an unqualified link between hyperlactacidaemia and tissue hypoxia doubtful. These studies showed that hyperlactacidaemia accompanying haemorrhage could be largely prevented by pretreatment with combined and adrenergic-receptor blockade.6 Similarly, adrenergic blockade abolished lactic acidosis induced by infusion of epinephrine.7 In dogs with haemorrhagic shock, plasma lactate correlated well with plasma catecholamines. Surprisingly, combined and -adrenergic blockade seemed to lessen hyperlactacidaemia not only by opposing catecholamine action on tissues but also by reducing the increase in plasma catecholamines.8 These early reports called for further studies into shock that would direct greater attention to the metabolic effects of epinephrine than to those of reduced tissue perfusion.6 Subsequent studies in dogs confirmed that elevated arterial lactate in shock was due not to lack of oxygen but to increased lactate production that could be mimicked by epinephrine infusion.9

THE LANCET Vol 354 August 7, 1999

505

Skeletal muscle as primary source of lactate in shock

In shocked dogs, the primary source of lactate was identified as skeletal muscle, in which lactate concentrations exceeded those in other tissues.10 When radiolabelled glucose was infused, the specific activity of lactate in arterial blood was similar to that in organs other than skeletal muscle, but lower in skeletal muscle itself.10 This result indicates that, in shock, skeletal muscle generates lactate more readily from its glycogen stores than from circulating glucose. Because skeletal muscle constitutes about 40% of the body-cell mass, changes in its metabolism in response to injury or infection allow it to become the main producer of lactate, perturbing overall carbohydrate metabolism.

Epinephrine and hyperlactacidaemia

In healthy individuals at rest, the plasma epinephrine threshold for producing an increase in circulating lactate is between 150 and 300 pg/mL, or about three to six times the normal concentration.11,12 In exercising individuals, increases in plasma lactate can be detected when the epinephrine concentration reaches about 200 pg/mL.13 As we shall discuss, epinephrine concentrations in sepsis or after injury frequently exceed these thresholds. During exercise of increasing intensity, plasma lactate concentration increases gradually at low work levels, but then increases rapidly as exercise increases to higher work intensity. The intensity at which lactate begins abruptly to increase is often called the lactate threshold or anaerobic threshold,14 suggesting that, at some work level, oxygen delivery becomes inadequate to meet metabolic demand. Circulating epinephrine concentration also rises with increasing exercise intensity; at maximum exercise intensity in trained athletes (runners or cyclists) epinephrine concentrations may reach 15002500 pg/mL.15 During exercise, oxygen saturation of muscle myoglobin remained stable at high exercise intensity and correlated poorly with circulating lactate concentration.16 However, several exercise studies have shown excellent correlation between concentrations of plasma lactate and epinephrine.13,1517 These observations suggest that the lactate threshold during exercise reflects increased aerobic glycolysis, stimulated by the rising epinephrine concentration rather than anaerobic glycolysis due to tissue hypoxia. Pronounced hyperlactacidaemia has also been associated with the high blood concentrations of epinephrine that occur in phaeochromocytoma18 or in patients treated with epinephrine after cardiopulmonary bypass.19 In acute bronchospasm or asthmatic attacks, 2-agonists can, rarely, cause lactic acidosis that is associated with hypokalaemia.20 Epinephrine secretion21 and plasma lactate22 can remain elevated for many weeks after major thermal injury. In burn-injured patients, there is little evidence for impairment of oxygen use,23 and lactate concentrations can be lowered by blockade with propranolol.22 In patients with shock due to sepsis, trauma, or haemorrhage, plasma epinephrine can be elevated for prolonged periods.24 In patients with septic shock, a group with high mortality, epinephrine was consistently high (>500 pg/mL) for 35 days preceding death. In most patients with traumatic or haemorrhagic shock, epinephrine concentrations were highest (>4000 pg/mL) immediately after injury or blood loss, and then fell to within the normal range in 1 or 2 days. However, in some of these patients, plasma
Stimulation by epinephrine of compartmentalised glycolysis coupled to Na+, K+-ATPase activity

epinephrine remained substantially elevated (5002000 pg/mL) for several days after injury.24 In one study, the haemodynamic effects of epinephrine were compared with those of dopamine in patients with severe bacterial sepsis or malaria.35 In 84% of these patients, epinephrine infusion was terminated earlier than planned because of pronounced lactic acidosis. Dopamine infusion was well tolerated, however, and both agents increased oxygen delivery and oxygen consumption. These observations do not necessarily identify skeletal muscle as the source of the lactate, although in some situations this inference can be made.16,26 Clearer evidence comes from studies showing that epinephrine stimulates lactate production by isolated skeletal muscle.27 Historically, attempts to understand epinephrinestimulated lactate production in skeletal muscle have focused on ATP supplyie, increased activity of glycogen phosphorylase and phosphofructokinase, which are enzymes thought to be the flux-controlling steps in glycogenolysis and glycolysis.28 Less attention has been given to ATP demandie, changes in cellular processes that may require increased ATP production. One such process is likely to be increased ion transport by the Na+, K+-ATPase.

Epinephrine stimulates Na+, K+-ATPase activity in skeletal muscle

Epinephrine has acute effects on muscle physiology. Muscles ability to conduct an action potential and contract depends on the membrane potential and concentration gradients across the membrane for sodium and potassium. These ion gradients are maintained by the Na+, K+-ATPase, which consumes one molecule of ATP to transport three sodium ions out of the cell and two potassium ions into the cell. At rest, muscle uses less than 10% of its total Na+, K+-ATPase activity to maintain sodium and potassium gradients. During muscle activity, ion movements associated with action potentials dissipate sodium and potassium gradients. If unopposed, this dissipation of ion gradients will result in loss of muscle membrane excitability. When muscle activity is high, the large reserve of Na+, K+-ATPase activity in muscle preserves membrane excitability and muscle contractility.29 The activity of the Na+, K+-ATPase in muscle is stimulated acutely by insulin, epinephrine, increased intracellular

506

THE LANCET Vol 354 August 7, 1999

sodium concentration, and contractile activity itself, among other factors.30 Thus skeletal muscle has a large latent capacity to transport sodium and potassium, thereby consuming ATP, in response to various stimuli. However, maximum stimulation of the Na+, K+ pump in skeletal muscle by epinephrine is greater than that by insulin. Epinephrine stimulates Na+, K+-ATPase activity in isolated skeletal muscles, thereby increasing potassium uptake and sodium excretion and hyperpolarising the membrane potential. This effect is mimicked by other -adrenergic agonists, and by analogues of cyclic AMP, and is blocked by propranolol or ouabain. These observations suggest that epinephrine stimulates the Na+, K+-ATPase (figure) by binding to adrenergic 2-receptors and raising cyclic AMP production.31 The reduction of circulating potassium concentration after administration of insulin or of -adrenergic agonists is probably due, at least partly, to stimulation of the Na+, K+-ATPase in skeletal muscle.32

selected trauma patients present with hypokalaemia, the degree of which is associated with the severity of trauma and with subsequent mortality.41 An acute, transient reduction in plasma potassium also occurs immediately after severe head trauma and has been attributed to massive catecholamine discharge.42 These observations suggest that stimulation by epinephrine of muscle Na+, K+-ATPase activity in injured patients has been detected many times in the guise of hypokalaemia.

Hypothesis
After injury and/or haemorrhage and during sepsis, neuroendocrine and cardiovascular stimuli combine to trigger and sustain release of epinephrine. High epinephrine concentrations stimulate adrenergic receptors in skeletal-muscle cell membranes and, among other effects, increase cyclic AMP production. This increase leads to the coordinated stimulation both of Na+, K+ATPase activity and of glycogenolysis. Increased Na+-K+ pump activity results in accelerated aerobic glycolysis that is sustained mainly by glycogen-derived glucose-6phosphate. Rapid ATP production fuelled by glycogen causes hyperlactacidaemia and muscle glycogen depletion, and intracellular accumulation of potassium in muscle results in hypokalaemia. Therapeutic measures to hasten lactate clearance by raising perfusion and oxygen delivery to supranormal levels would be expected to have little direct effect on these metabolic processes. However, such measures may reduce glycolysis indirectly, by reducing the stimuli for epinephrine release. Unnecessary use of blood products (red-cell transfusion) subjects the patient to increased risk of infectious viral agent transmission. Inappropriate use of inotropic agents in an effort to raise tissue oxygen delivery at the expense of increased myocardial oxygen consumption may have little benefit for the patient. However, if resuscitation is inadequate and if hypotension, hypovolaemia, or hypoxia persist and continue to stimulate epinephrine release,43 treatments aimed at improving tissue perfusion may reduce blood lactate by slowing epinephrine secretion. If epinephrine secretion persists for reasons unrelated to inadequate resuscitation, aerobic glycolysis would be expected to persist. Although hypotension, hypovolaemia, or hypoxia can stimulate epinephrine secretion and thus raise lactate concentration, continued efforts at resuscitation in the presence of normal blood pressure, PO2, pulse, and urine ouput may be indefensible and possibly harmful.

Aerobic glycolysis and the Na+, K+-ATPase

Increased activity of the Na+, K+-ATPase leads to increased lactate production under well-oxygenated conditions in various cells, including erythrocytes, vascular smooth muscle, neurons, glia, and skeletal muscle.3336 Conversely, inhibition of the Na+, K+ pump with ouabain reduces lactate production. To explain this feature, it has been proposed that clusters of enzymes consisting of the complete glycolytic cascade may be associated with membranes in proximity to ion transporters.33,37 Further research is necessary to clarify whether, among other possibilities, only a distinct subgroup of membrane Na+, K+ pumps preferentially receives glycolytically derived ATP or whether hormones, energy demand, and nutritional state influence the ratio of oxidatively derived to glycolytically derived ATP supplied to membrane ion pumps. It is nonetheless clear that glycolysis provides ATP to sustain Na+, K+-ATPase activity in cells with intact oxidative capacity, including skeletal muscle. In a subcellular compartment to which access by diffusion from the bulk cytoplasm was limited, local ADP availability would regulate local ATP production (figure). In this compartment, ATP consumption by the Na+, K+ATPase would be the main influence on ADP availability. Close association between ATP-consuming membrane ion pumps and ATP-producing glycolytic enzymes would tightly couple the use and synthesis of ATP such that the rate of ATP utilisation regulated the rate of ATP production. Furthermore, the flux-controlling enzyme in glycolysis, phosphofructokinase, is activated by ADP and inhibited by ATP. Stimulation of the Na+, K+-ATPase would generate ADP, thereby raising phosphofructokinase activity and accelerating aerobic glycolysis. In keeping with this proposed mechanism, either -receptor blockade by propranolol38 or inhibition of the Na+, K+-ATPase by ouabain inhibits epinephrine-stimulated lactate production by skeletal muscle in vitro.27

Testing the hypothesis

The hypothesis predicts that epinephrine is a primary stimulus to lactate production through stimulation of Na+, K+ pump activity in skeletal muscle. These predictions can be tested in healthy individuals by an assessment of the net flux of lactate and potassium across a limb. If this hypothesis is correct, we would expect to find the following associations over a range of epinephrine concentrations: (i) lactate flux should be positively correlated with epinephrine concentration; (ii) potassium flux should be negatively correlated with epinephrine concentration; and (iii) lactate flux should be proportional to potassium flux. Moreover, -blockade would be expected to dissociate epinephrine concentrations from lactate and potassium fluxes. We are unaware of any study that has examined all of these relations.

Hypokalaemia after injury indicates increased Na+, K+-ATPase activity

Muscle potassium uptake is stimulated by epinephrine or other 2 agonists, and that stimulation is inhibited by ouabain, indicating involvement of the Na+-K+ pump.31,39 Administration of 2 agonists results in a prompt decrease in circulating potassium.32,40 More than half of randomly

THE LANCET Vol 354 August 7, 1999

507

Patients with severe head trauma, but with relatively minor systemic injuries, have increased metabolic expenditure, a hyperdynamic cardiovascular state,44 and increased circulating epinephrine45 and lactic acid concentrations.46 In these patients, propranolol reduces blood pressure, heart rate, cardiac work, and epinephrine concentrations.47 Limb flux studies like those described above could be done in these patients. Studies in which plasma concentrations of potassium, lactate, and epinephrine were to be measured concurrently should show positive correlations between epinephrine and lactate and negative correlations between potassium and lactate as well as between potassium and epinephrine. In these patients without major systemic injury, -blockade might be of use in dissociating plasma lactic acid from epinephrine concentrations.
Supported by grants from Shriners Hospitals for Children and the US Public Health Service.

References
1 Bakker J, Gris P, Coffernils M, Kahn RJ, Vincent JL. Serial blood lactate levels can predict the development of multiple organ failure following septic shock. Am J Surg 1996; 171: 22126. Abramson D, Scalea TM, Hitchcock R, Trooskin SZ, Henry SM, Greenspan J. Lactate clearance and survival following injury. J Trauma 1993; 35: 58488. Porter JM, Ivatury RR. In search of the optimal end points of resuscitation in trauma patients: a review. J Trauma 1998; 44: 90814. Gore DC, Jahoor F, Hibbert JM, Demaria EJ. Lactic acidosis during sepsis is related to increased pyruvate production, not deficits in tissue oxygen availability. Ann Surg 1996; 224: 97102. Astiz ME, Rackow EC. Septic shock. Lancet 1998; 351: 150105. Irving MH. The sympatho-adrenal factor in haemorrhagic shock. Ann R Coll Surg Engl 1968; 42: 36786. Halmagyi DF, Irving MH, Gillett DJ, Varga D. Effect of adrenergic blockade on consequnces of sustained epinephrine infusion. J Appl Physiol 1967; 23: 17177. Halmagyi DF, Kennedy M, Varga D. Combined adrenergic receptor blockade and circulating catecholamines in hemorrhagic shock. Eur Surg Res 1971; 3: 37888. Liddell MJ, Daniel AM, MacLean LD, Shizgal HM. The role of stress hormones in the catabolic metabolism of shock. Surg Gynecol Obstet 1979; 149: 82230. Daniel AM, Shizgal HM, MacLean LD. The anatomic and metabolic source of lactate in shock. Surg Gynecol Obstet 1978; 147: 697700. Clutter WE, Bier DM, Shah SD, Cryer PE. Epinephrine plasma metabolic clearance rates and physiologic thresholds for metabolic and hemodynamic actions in man. J Clin Invest 1980; 66: 94101. Staten MA, Matthews DE, Cryer PE, Bier DM. Physiological increments in epinephrine stimulate metabolic rate in humans. Am J Physiol 1987; 253: E32230. Podolin DA, Munger PA, Mazzeo RS. Plasma catecholamine and lactate response during graded exercise with varied glycogen conditions. J Appl Physiol 1991; 71: 142733. Riley M, Maehara K, Porszasz J, et al. Association between the anaerobic threshold and the break-point in the double product/work rate relationship. Eur J Appl Physiol 1997; 75: 1421. Mazzeo RS, Marshall P. Influence of plasma catecholamines on the lactate threshold during graded exercise. J Appl Physiol 1989; 67: 131922. Richardson RS, Noyszewski EA, Leigh JS, Wagner PD. Lactate efflux from exercising human skeletal muscle: role of intracellular PO2. J Appl Physiol 1998; 85: 62734. Hughson RL, Green HJ, Sharratt MT. Gas exchange, blood lactate, and plasma catecholamines during incremental exercise in hypoxia and normoxia. J Appl Physiol 1995; 79: 113441. Bornemann M, Hill SC, Kidd GS. Lactic acidosis in pheochromocytoma. Ann Intern Med 1986; 105: 88082. Totaro RJ, Raper RF. Epinephrine-induced lactic acidosis following cardiopulmonary bypass. Crit Care Med 1997; 25: 169399. Assadi FK. Therapy of acute bronchospasm. Complicated by lactic acidosis and hypokalemia. Clin Pediatr 1989; 28: 25860. Goodall M. Sympatho-adrenal medullary response to thermal burn. Ann N Y Acad Sci 1968; 150: 68589.

3 4

5 6 7

10 11

12

13

14

15

16

17

18 19 20 21

22 Gore DC, Honeycutt D, Jahoor F, Barrow RE, Wolfe RR, Herndon DN. Propranolol diminishes extremity blood flow in burned patients. Ann Surg 1991; 213: 56873. 23 Wolfe RR, Jahoor F, Herndon DN, Miyoshi H. Isotopic evaluation of the metabolism of pyruvate and related substrates in normal adult volunteers and severely burned children: effect of dichloroacetate and glucose infusion. Surgery (St Louis) 1991; 110: 5467. 24 Benedict CR, Grahame-Smith DG. Plasma noradrenaline and adrenaline concentrations and dopamine-beta-hydroxylase activity in patients with shock due to septicaemia, trauma and haemorrhage. Q J Med 1978; 47: 120. 25 Day NP, Phu NH, Bethell DP, et al. The effects of dopamine and adrenaline infusions on acid-base balance and systemic haemodynamics in severe infection. Lancet 1996; 348: 21923 (see also Department of Error. Lancet 1996; 348: 902). 26 Fryburg DA, Gelfand RA, Jahn LA, et al. Effects of epinephrine on human muscle glucose and protein metabolism. Am J Physiol 1995; 268: E5559. 27 Luchette FA, Friend LA, Brown CC, Upputuri RK, James JH. Increased skeletal muscle Na+, K+-ATPase activity as a cause of increased lactate production after hemorrhagic shock. J Trauma 1998; 44: 796801. 28 Dietz MR, Chiasson JL, Soderling TR, Exton JH. Epinephrine regulation of skeletal muscle glycogen metabolism. Studies utilizing the perfused rat hindlimb preparation. J Biol Chem 1980; 255: 230107. 29 Nielsen OB, Clausen T. The significance of active Na+, K + transport in the maintenance of contractility in rat skeletal muscle. Acta Physiol Scand 1996; 157: 199209. 30 Clausen T. Long- and short-term regulation of the Na+-K+ pump in skeletal muscle. News Physiol Sci 1996; 11: 2430. 31 Clausen T, Flatman JA. Beta 2-adrenoceptors mediate the stimulating effect of adrenaline on active electrogenic Na-K-transport in rat soleus muscle. Br J Pharmacol 1980; 68: 74955. 32 Clausen T. Adrenergic control of Na+-K+-homoeostasis. Acta Med Scand 1983; 672 (suppl): 11115. 33 Parker JC, Hoffman JF. The role of membrane phosphoglycerate kinase in the control of glycolytic rate by active cation transport in human red blood cells. J Gen Physiol 1967; 50: 893916. 34 Paul RJ, Bauer M, Pease W. Vascular smooth muscle: aerobic glycolysis linked to sodium and potassium transport processes. Science 1979; 206: 141416. 35 Pellerin L, Magistretti PJ. Glutamate uptake into astrocytes stimulates aerobic glycolosis: a mechanism coupling neuronal activity to glucose utilization. Proc Natl Acad Sci USA 1994; 91: 1062529. 36 James JH, Fang CH, Schrantz SJ, Hasselgren PO, Paul RJ, Fischer JE. Linkage of aerobic glycolysis to sodium-potassium transport in rat skeletal muscle. Implications for increased muscle lactate production in sepsis. J Clin Invest 1996; 98: 238897. 37 Paul RJ, Hardin CD, Raeymaekers L, Wuytack F, Casteels R. Preferential support of Ca2+ uptake in smooth muscle plasma membrane vesicles by an endogenous glycolytic cascade. FASEB J 1989; 3: 2298301. 38 Luchette FA, Robinson BRH, Friend LA, McCarter FD, Frame SB, James JH. Adrenergic antagonists reduce lactic acidosis in response to hemorrhagic shock. J Trauma 1999; 46: 87380. 39 Flatman JA, Clausen T. Combined effects of adrenaline and insulin on active electrogenic Na+-K+ transport in rat soleus muscle. Nature 1979; 281: 58081. 40 Brown MJ, Brown DC, Murphy MB. Hypokalemia from beta2-receptor stimulation by circulating epinephrine. N Engl J Med 1983; 309: 141419. 41 Vanek VW, Seballos RM, Chong D, Bourguet CC. Serum potassium concentrations in trauma patients. South Med J 1994; 87: 4146. 42 Pomeranz S, Constantini S, Rappaport ZH. Hypokalaemia in severe head trauma. Acta Neurochir 1989; 97: 6266. 43 Lewis GP. Physiological mechanisms controlling secretory activity of adrenal medulla. In: Greep RO, Astwood EB, eds. Handbook of physiology: endocrinology. Vol VI. Adrenal Gland, ed. Washington, DC: American Physiological Society, 1975: 30919. 44 Clifton GL, Robertson CS, Kyper K, Taylor AA, Dhekne RD, Grossman RG. Cardiovascular response to severe head injury. J Neurosurg 1983; 59: 44754. 45 Hamill RW, Woolf PD, McDonald JV, Lee LA, Kelly M. Catecholamines predict outcome in traumatic brain injury. Ann Neurol 1987; 21: 43843. 46 Hadfield JM, Little RA, Jones RA. Measured energy expenditure and plasma substrate and hormonal changes after severe head injury. Injury 1992; 23: 17782. 47 Robertson CS, Clifton GL, Taylor AA, Grossman RG. Treatment of hypertension associated with head injury. J Neurosurg 1983; 59: 45560.

508

THE LANCET Vol 354 August 7, 1999