Journal ofNeurology, Neurosurgery, and Psychiatry 1993;56:589-598 NEUROLOGICAL EMERGENCY The management of medical coma David Bates The patient who is brought to the hospital casualty department, or seen on the Intensive Care Unit, though not having been exposed to evident trauma, may be harbouring delayed effects of trauma such as a subdural haematoma or meningitis arising from a basal skull fracture. The problems of raised intra- cranial pressure following a parenchymal haematoma in a hypertensive patient, the decompensation of a cerebral tumour or col- lection of pus, means that all possible causes of loss of consciousness must be considered by the physician when dealing with a patient in coma. Thus in the diagnosis of medical coma it is not easy to exclude the patient in coma following head injury. If one excludes patients with a transient loss of consciousness following seizure, syn- cope, cardiac dysrhythmia or hypoglycaemia and those unresponsive due to impending death, and considers patients who have been unconscious for some five to six hours, then 40% of such patients seen in medical practice will have taken some form of sedative drugs with or without alcohol.' Of the remainder just over 40% will have suffered hypoxic ischaemic insult as the result of cardiac arrest or anaesthetic accident, a third will be uncon- scious as a result of cerebrovascular acci- dents, either haemorrhage or infarction, and about a quarter will be unconscious as a result of metabolic coma including infection, renal failure, hepatic failure and complica- tions of diabetes mellitus. If one considers only those cases who are initially regarded as "of unknown aetiology" the proportion of drug overdoses is about 30%, mass lesions about 34% and diffuse metabolic causes account for 36%.2 Few problems are more difficult to manage than the unconscious patient because the potential causes of loss of consciousness are considerable and because the time for diagno- sis and effective intervention is relatively short. All alterations in arousal should be regarded as acute and potentially life threat- ening emergencies until vital functions are stabilised, the underlying cause of the coma is diagnosed and reversible causes are corrected. Delay in instituting treatment for a patient with raised intracranial pressure may have obvious consequences in terms of pressure coning but similarly the unnecessary investi- gation of patients in metabolic coma with imaging techniques may delay the initiation of appropriate therapy. It is therefore essen- tial for the physician in charge to adopt a sys- tematic approach initially to ensure resuscitation, and then to direct further tests towards producing the most rapid diagnosis and the most appropriate therapy. The devel- opment of such a systematic approach demands an understanding of the pathophysi- ology of consciousness and the ways in which it may be deranged. The causes of coma The phenomenon of consciousness depends upon an intact ascending reticular activating substance in the brainstem to act as the alert- ing or awakening element of consciousness together with a functioning cerebral cortex of both hemispheres which determines the con- tent of that consciousness. The ascending reticular activating substance is a continuous isodentritic core extending from the medulla through the pons to the mid-brain which is continuous caudally with the reticular inter- mediate grey lamina of the spinal cord and rostrally with the subthalamus, the hypothala- mus and the thalamus.3 Its functions and interconnections are considerable and its role greater than that of a simple cortical arousal system. There are named nuclei throughout the reticular formation and, although it was originally considered that cortical arousal depended upon projections from the reticular formation via the midline thalamic nuclei to the thalamic reticular nucleus and the cortex, it now seems unlikely that the thalamic retic- ular nucleus is the final relay and the specific role of the various links from the reticular for- mation to the thalamus has yet to be identi- fied. Similarly the neurotransmitters involved in this arousal system are not fully determined though it seems likely that, in addition to cholinergic and monoaminergic systems, gamma aminobutyric acid (GABA) may be important in controlling consciousness." It follows from recognition of the anatomy and pharmacology of the ascending reticular activating substance that structural damage to this pathway or chemical derangement of the neurotransmitters involved are mechanisms whereby consciousness may be impaired. Such conditions will occur with focal lesions in the brainstem, mass lesions in the posterior fossa impinging directly on the brainstem or Department of Neurology, University of Newcastle Upon Tyne, UK D Bates 589 on November 2, 2020 by guest. Protected by copyright. http://jnnp.bmj.com/ J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.6.589 on 1 June 1993. Downloaded from
Transcript
Journal ofNeurology, Neurosurgery, and Psychiatry 1993;56:589-598
NEUROLOGICAL EMERGENCY
The management of medical coma
David Bates
The patient who is brought to the hospitalcasualty department, or seen on the IntensiveCare Unit, though not having been exposedto evident trauma, may be harbouringdelayed effects of trauma such as a subduralhaematoma or meningitis arising from a basalskull fracture. The problems of raised intra-cranial pressure following a parenchymalhaematoma in a hypertensive patient, thedecompensation of a cerebral tumour or col-lection of pus, means that all possible causesof loss of consciousness must be consideredby the physician when dealing with a patientin coma. Thus in the diagnosis of medicalcoma it is not easy to exclude the patient incoma following head injury.
If one excludes patients with a transientloss of consciousness following seizure, syn-cope, cardiac dysrhythmia or hypoglycaemiaand those unresponsive due to impendingdeath, and considers patients who have beenunconscious for some five to six hours, then40% of such patients seen in medical practicewill have taken some form of sedative drugswith or without alcohol.' Of the remainderjust over 40% will have suffered hypoxicischaemic insult as the result of cardiac arrestor anaesthetic accident, a third will be uncon-scious as a result of cerebrovascular acci-dents, either haemorrhage or infarction, andabout a quarter will be unconscious as aresult of metabolic coma including infection,renal failure, hepatic failure and complica-tions of diabetes mellitus. If one considersonly those cases who are initially regarded as"of unknown aetiology" the proportion ofdrug overdoses is about 30%, mass lesionsabout 34% and diffuse metabolic causesaccount for 36%.2Few problems are more difficult to manage
than the unconscious patient because thepotential causes of loss of consciousness areconsiderable and because the time for diagno-sis and effective intervention is relativelyshort. All alterations in arousal should beregarded as acute and potentially life threat-ening emergencies until vital functions arestabilised, the underlying cause of the coma isdiagnosed and reversible causes are corrected.Delay in instituting treatment for a patientwith raised intracranial pressure may haveobvious consequences in terms of pressureconing but similarly the unnecessary investi-gation of patients in metabolic coma withimaging techniques may delay the initiation
of appropriate therapy. It is therefore essen-tial for the physician in charge to adopt a sys-tematic approach initially to ensureresuscitation, and then to direct further teststowards producing the most rapid diagnosisand the most appropriate therapy. The devel-opment of such a systematic approachdemands an understanding of the pathophysi-ology of consciousness and the ways in whichit may be deranged.
The causes ofcomaThe phenomenon of consciousness dependsupon an intact ascending reticular activatingsubstance in the brainstem to act as the alert-ing or awakening element of consciousnesstogether with a functioning cerebral cortex ofboth hemispheres which determines the con-tent of that consciousness. The ascendingreticular activating substance is a continuousisodentritic core extending from the medullathrough the pons to the mid-brain which iscontinuous caudally with the reticular inter-mediate grey lamina of the spinal cord androstrally with the subthalamus, the hypothala-mus and the thalamus.3 Its functions andinterconnections are considerable and its rolegreater than that of a simple cortical arousalsystem. There are named nuclei throughoutthe reticular formation and, although it wasoriginally considered that cortical arousaldepended upon projections from the reticularformation via the midline thalamic nuclei tothe thalamic reticular nucleus and the cortex,it now seems unlikely that the thalamic retic-ular nucleus is the final relay and the specificrole of the various links from the reticular for-mation to the thalamus has yet to be identi-fied.
Similarly the neurotransmitters involved inthis arousal system are not fully determinedthough it seems likely that, in addition tocholinergic and monoaminergic systems,gamma aminobutyric acid (GABA) may beimportant in controlling consciousness."
It follows from recognition of the anatomyand pharmacology of the ascending reticularactivating substance that structural damage tothis pathway or chemical derangement of theneurotransmitters involved are mechanismswhereby consciousness may be impaired.Such conditions will occur with focal lesionsin the brainstem, mass lesions in the posteriorfossa impinging directly on the brainstem or
Department ofNeurology, UniversityofNewcastle UponTyne, UKD Bates
589 on N
ovember 2, 2020 by guest. P
rotected by copyright.http://jnnp.bm
j.com/
J Neurol N
eurosurg Psychiatry: first published as 10.1136/jnnp.56.6.589 on 1 June 1993. D
mass lesions involving the cerebral hemi-spheres causing tentorial pressure coning andconsequently comproming the ascendingreticular activating substance either by directpressure or by a process of ischaemia. Inaddition toxins, most commonly ingesteddrugs, may have a significant depressanteffect upon the brainstem ascending reticularactivating substance and thereby result in lossof consciousness.The content of consciousness resides in the
cerebral cortex of both hemispheres. Unlikethose discrete cortical functions such as lan-guage or vision which are focally locatedwithin the cortex, the content of conscious-ness can best be regarded as the amalgum ofall cognitive function. Coma arising from dis-ruption of this cortical activity requires a dif-fuse pathology such as generalised anoxia orischaemia, commonly seen after cardiac arrestor anaesthetic accidents, or the effects of pre-sumed cortical vasospasm seen in infectivemeningitis or the chemical meningitis follow-ing subarachnoid haemorrhage where gener-alised cortical ischaemia is believed to be thecause of disruption of function.
For the physician attempting to diagnosethe cause of coma consideration must begiven to:-A) Supra or infra tentorial mass lesions.Typically these will provide evidence ofraised intracranial pressure and commonlyproduce focal signs. Pathologies such as neo-plasm or haematoma, infarction with cerebraloedema, abscess, focal encephalitis andvenous sinus thrombosis should beconsidered.B) Subtentorial destructive lesions or the localeffect of toxin. These pathologies will directlydamage the ascending reticular activatingsubstance as in brainstem infarction,rhombencephalitis, brainstem demyelinationand the much more common effects of self-poisoning with sedative drugs.C) Diffuse damage to the cerebral cortex.Bilateral cortical injury is most commonlyseen in states of hypoxia and ischaemia butmay be mimicked by hypoglycaemia, keto-acidosis, electrolyte abnormalities, bacterialmeningitis, viral encephalitis and diffuse post-infectious encephalomyelitis. It is also thelikely pathology of coma following subarach-noid haemorrhage.
Table 1 The Glasgow coma scale
Eyes Open Spontaneously 4To verbal command 3To pain 2
No response 1Best motor response To verbal command Obeys 6
Best verbal response Orientated 5Disorientated 4Inappropriate words 3Incomprehensible sounds 2No response 1
DefinitionsThere is a continuum from the individual infull consciousness to the patient in deepcoma. The terminology which is most usuallyemployed derives from the Brain InjuriesCommittee of the MRC7:-A) Confusion-"disturbance of consciousnesscharacterised by impaired capacity to thinkclearly and with customary repetition and toperceive, respond to and remember currentstimuli; there is also disorientation".Confusion involves a generalised disturbanceof cortical cerebral function which is usuallyassociated with considerable EEG abnormali-ties. Some authors describe an interveningstate between normal consciousness and con-fusion, that of clouding of consciousness.2B) Delirium-"a state of much disturbed con-sciousness with motor restlessness, transienthallucinations, disorientation and perhapsdelusions".C) Obtundation-"a disorder of alertnessassociated with psychomotor retardation".D) Stupor-"a state in which the patient,though not unconscious, exhibits little or nospontaneous activity". Although the individ-ual appears to be asleep he or she will awakento vigorous stimulation but show limitedmotor activities and usually fail to speak.E) Coma-"a state of unarousable psycho-logic unresponsiveness in which the subjectslie with eyes closed and show no psychologi-cally understandable response to externalstimulus or inner need". This may be short-ened to "a state of unarousable unresponsive-ness" which implies both the defect in arousaland in awareness of self or environment man-ifest as an inability to respond. A more usefulassessment of coma is derived from the hier-archical Glasgow Coma Scale8 in whichpatients who fail to show eye opening inresponse to voice, perform no better thanweak flexion in response to pain and make, atbest, only unrecognisable grunting noises inresponse to pain, are regarded as being incoma. This allows the patients to have an eyeopening response of two or less, a motorresponse of four or less and verbal response oftwo or less. The sum Glasgow score of eightshould not be regarded as being definitive ofcoma since the total score can be achieved inseveral different ways (table 1).F) Vegetative state. When the cortex of thecerebral hemispheres of the brain recovermore slowly than the brain stem or when thecortex is irreversibly damaged there may arisea situation in which the patient enters a vege-tative state without cognitive function. It maybe a transient phase through which patientsin coma pass as they recover or deterioratebut, and commonly after anoxic injuries tothe brain, there develops a state in which thebrain stem recovers function but the cerebralhemispheres are not capable of recovery.When this occurs the patient enters a "persis-tent vegetative state" described by Jennettand Plum.9 Such patients may survive forlong periods, on occasion for decades, butnever recover outward manifestations ofhigher mental activity and the condition,
590 on N
ovember 2, 2020 by guest. P
rotected by copyright.http://jnnp.bm
j.com/
J Neurol N
eurosurg Psychiatry: first published as 10.1136/jnnp.56.6.589 on 1 June 1993. D
which is comparatively newly recognised,relates to the development of modem resusci-tative techniques. Other terms have beenused in the past to identify similar conditions.These include coma vigil, the apallic syn-drome, cerebral death, neocortical death, andtotal dementia.G) Akinetic mutism has been defined as asimilar condition of unresponsiveness butapparent alertness, as demonstrated by reac-tive alpha and theta electroencephalographicrhythms in response to stimuli. The majordifference from the vegetative state, in whichthere is tone in the muscles and extensor orflexor responses, is that patients with akineticmutism have flaccid tone and are unrespon-sive to peripheral pain. It is thought that thisstate is due to bilateral frontal lobe lesions,diffuse cortical lesions or lesions of the deepgrey matter.'0H) The locked in syndrome. Feldman"described a de-efferented state due to bilat-eral ventral pontine lesion involving damageto the corticospinal, corticopontine and corti-cobulbar tracts. The patient has total paraly-sis below the level of the third nerve nucleiand although able to open, elevate anddepress the eyes has no horizontal eye move-ments and no other voluntary eye movement.The diagnosis depends upon the physicianbeing able to recognise that the patient canopen the eyes voluntarily and can signalassent or dissent by responding numericallywith eye closure. Similar states are occasion-ally seen in patients with severe polyneurop-athy, myasthenia gravis and after the use ofneuromuscular blocking agents.I) Pseudo coma. Rarely, patients who appearin coma without structural, metabolic, toxicor psychiatric disorder being apparent, can beshown by tests of brainstem function, to haveintact brainstem activity and corticopontineprojections and not to be in coma.
ResuscitationAlthough resuscitation is commonly per-formed by the casualty officer or the anaes-thetist in the Intensive Care Unit rather thanby the neurologist it is appropriate that theneurologist remembers that, in patients whoare unconscious, protection of the airway,respiration, support of the circulation andprovision of an adequate supply of glucoseare all important in stabilising the patient. Itis frequently necessary to intubate the tracheain a patient in coma, not only to ensure anadequate airway but also to prevent the aspi-ration of vomit. It is also important to notethe respiratory rate and pattern before intuba-tion and certainly before instituting mechani-cal ventilation; since depressed respiration isa frequent clue to drug overdose or metabolicdisturbance, increased respiration to hypoxia,hypercapnia or acidosis and fluctuating respi-ration may indicate a brainstem lesion. Thepossibility that respiratory failure is the causeof coma should always be considered in apatient with disordered respiration.
Once adequate oxygenation and circulation
are ensured and monitored, blood should bewithdrawn for the determination of bloodglucose, biochemical estimations and toxicol-ogy. It is then reasonable to give a bolus of25-50g of dextrose despite the present con-troversy about the use of intravenous glucosein patients with ischaemic or anoxic braindamage. It can be argued that extra glucosein this situation may augment local lactic acidproduction by anaerobic glycolosis andpotentially worsen ischaemic or anoxic dam-age. In practice in the situation of ischaemicor anoxic brain damage and even in the pres-ence of a diabetic ketoacidosis the administra-tion of such a quantity of glucose will not beimmediately harmful and in the hypogly-caemic patient it may well be life saving. Areasonable compromise would be to obtainan early assessment of the level of blood glu-cose by dextrostix testing but these are notsufficiently accurate to preclude the need forformal laboratory assessment. When glucoseis given in this situation an argument can bemade for giving a bolus of thiamine at thesame time to prevent precipitation ofWemicke's encephalopathy."2An essential part of resuscitation includes
the establishment of baseline, blood pressure,pulse, temperature, the establishment of anintravenous line and the stabilisation of theneck together with an examination for menin-gitis. It may be difficult in those patients whohave sustained some degree of trauma in theircollapse to assess the stability of the neck butthe establishment of an adequate airway cer-tainly takes precedence and the identificationof meningismus in a febrile patient probablytakes precedence over the stabilisation ofneck movements. In a comatose febrilepatient with meningismus seen outside thehospital environment the intramuscular injec-tion of penicillin before transfer is now recog-nised to carry a significant advantage.
HistoryOnce the patient is stable it is important toobtain as much information as possible fromthose who accompanied the patient to hospi-tal or who watched the onset of coma. Thecircumstances in which consciousness waslost are of vital importance in helping to iden-tify the diagnosis. Generally, coma is likely topresent in one of three ways; as the pre-dictable progression of an underlying illness;as an unpredictable event in a patient with apreviously known disease; as a totally unex-pected event. Distinctions between these pre-sentations are often achieved by the history ofthe circumstances in which consciousness waslost. In the first category are patients follow-ing focal brainstem infarction who deteriorateor those with known intracranial mass lesionswho show similar deterioration. In the secondcategory patients with recognised cardiacarrhythmia or the known risk factor of sepsisfrom an intravenous line. In the final categoryit is important to determine whether there hasbeen a previous history of seizures, trauma,febrile illnesses, or focal neurological distur-
591 on N
ovember 2, 2020 by guest. P
rotected by copyright.http://jnnp.bm
j.com/
J Neurol N
eurosurg Psychiatry: first published as 10.1136/jnnp.56.6.589 on 1 June 1993. D
bances. The history of a sudden collapse inthe midst of a busy street or office indicatesthe need for different investigations fromthose necessary for the patient who is discov-ered at home in bed surrounded by emptybottles of sedative tablets.
Examination and monitoringThe third phase of the management of thepatient in coma involves a rapid but system-atic examination to identify possible causes ofthe coma.
TemperatureFever usually indicates infection and rarely abrainstem or diencephalic lesion affecting thetemperature centres."3 Most commonly thecombination of fever and coma indicates sys-temic infection such as pneumonia, septi-caemia or a cerebral cause such as meningitis,encephalitis or abscess. When seizures occurtogether with fever the possibility ofencephalitis or cerebral abscess is greatlyincreased. Heat stroke may present as afebrile comatose patient when the clue to thediagnosis is in the environment.
Hypothermia is most commonly seen as acomplication of an accident or cerebrovascu-lar disease when an elderly patient is discov-ered having lain for hours or days in anunderheated room. It may also be seen fol-lowing intoxication with alcohol or barbitu-rates, with peripheral circulatory failure andrarely with profound myxoedema.
Heart rateA tachyarrhythmia or bradyarrhythmia maybe significant in identifring the cause of cere-bral hypoperfusion. Irregularity of the pulsealways raises the question of atrial fibrillationand associated embolic disease.
Blood pressureHypotension might indicate shock, myocar-dial infarction, septicaemia or intoxication. Itmay also indicate diabetes mellitus orAddison's disease. Hypertension is of lesshelp in the diagnosis of the patient in coma asit may be the cause, as in cerebral haemor-rhage or hypertensive encephalopathy, but itcan also be the result of the cerebral lesion.
RespirationFor those reasons already given, assessmentof respiration may be compromised by theneeds of resuscitation but generally, slow andshallow breathing raises the question of drugintoxication. Deep, rapid respiration suggestspneumonia or acidosis which may also occurin brainstem lesions causing central neuro-genic hyperventilation.
IntegumentThe appearance of the skin and mucousmembrane may identify anaemia, jaundice,cyanosis or raise the possibility of carbonmonoxide poisoning. Bruising over the scalpor mastoids, the presence of blood in theexternal auditory meati or nostrils will raise
the possibility of a basal skull fracture andbruising elsewhere in the body raises thequestion of significant trauma. An exanthemmay indicate the presence of a viral infectioncausing meningoencephalitis, meningococcalsepticaemia or raise the question of haemor-rhagic disease. Hyperpigmentation raises thepossibility of Addison's disease, and the pres-ence of bullous skin lesions is frequently seenin barbiturate intoxication. Evidence ofKaposi sarcoma, anogenital herpetic lesionsor oral candodiasis would raise the questionof an acquired immune deficiency syndrome(AIDS) with the consequent plethora ofpossible CNS disease.
BreathThe odour of the breath of an unconsciouspatient may indicate the presence of intoxica-tion with alcohol, raise the question of dia-betes or suggest that the cause of coma isuraemic or hepatic.
CardiovascularAuscultation and examination of the heartmay indicate valvular disease and raise thepossibility of endocarditis. Bruits over thecarotid vessels might indicate the presence ofcerebrovascular disease and splinter haemor-rhages seen in the nail bed would raise thepossibility of sub-acute bacterial endocarditisor collagen vascular diseases.
AbdomenExamination of the abdomen may revealsigns of trauma or rupture of viscera,hepatomegaly or splenomegaly may indicatethe possibility of a portocaval shunt and thefindings of polycystic kidneys would raise thepossibility of subarachnoid haemorrhage.
MeningismusExamination of the skull and spine is impor-tant and the physician should always look forneck stiffness. Kernig's test in which the resis-tance of flexion of the thigh with the legextended is examined or Brudzinski's tests inwhich flexing of one thigh is noted to causeflexion of the other thigh, should be per-formed to help in differentiating neck stiff-ness, due to meningeal irritation, from thatdue to a developing tonsillar pressure cone. Ifthe Kemig and Brudzinski tests are positivetogether with neck stiffness this impliesinflammation in the lumbar theca and sug-gests a diffuse meningitic process. If thesetests are negative, however, then the neckstiffness alone is more suggestive of a forami-nal pressure cone.'4
Fundal examinationThe presence of papilloedema, fundal haem-orrhage or evidence of emboli, together withthe findings of hypertensive, vascular or dia-betic retinopathy are important. The fundalappearances may be diagnostic as in the find-ing of subhyaloid haemorrhage but morecommonly only help to confirm or refute evi-dence of raised intracranial pressure. Theabsence of papilloedema does not necessarily
Bates on N
ovember 2, 2020 by guest. P
rotected by copyright.http://jnnp.bm
j.com/
J Neurol N
eurosurg Psychiatry: first published as 10.1136/jnnp.56.6.589 on 1 June 1993. D
mean that there is no increased intracranialpressure.
Neurological examinationThe position, posture and spontaneous move-
ments of the unconscious patient should benoted. The formal neurological examinationconsists of the elicitation of various reflexresponses."3 The most important aspects ofneurological examination are those whichdefine the level of consciousness, identify theactivity of the brainstem and search forevidence of lateralisation (table 2).A) The level of consciousnessThe Glasgow Coma Scale8 provides the mostuseful hierarchical assessment of the level ofconsciousness. The response to commands,calling the patient's name and painful stimuliare observed for eye opening, limb movementand voice. Painful stimuli such as supraorbitalpressure for central stimulation and nail bedpressure for peripheral stimulation are usefuland reproducible. Eye opening is relativelyeasy to assess though the fixed and unrespon-sive opening of the eyes sometimes seen indeep coma must not be confused with thevolitional or reflex opening of the eyes from a
closed position in response to stimuli. All fourlimbs are tested individually for movementand the best response is recorded in assessingthe Glasgow Coma Scale but an asymmetrybetween responses may be of importance inthe overall assessment (vide infra). Patients inlighter grades of coma still retain the ability tovocalise and may grimace and withdraw theirlimbs from pain. These responses are pro-
gressively lost as the coma deepens and it isimportant to test pain bilaterally in theperiphery and cranially since patients may
only vocalise or respond to painful stimuli on
one side raising the possibility of hemianaes-thesia and providing evidence for a focallesion. A grimace response to painful stimula-tion is believed to indicate intact corticobul-bar function2 but there are patients in coma,
particularly after hypoxic ischaemic insults,who show grimace in response to minorperipheral stimulation yet have no associatedperipheral motor response. When this situa-tion is seen it always raises the question of a
ventral pontine lesion or of a cervical cordinjury but more commonly evolves into a veg-
etative state and is, generally, a poor prognos-tic sign.The level of coma should be documented
Table 2 Neurological assessment ofcoma
Glasgow coma scale Eye openingMotor responseVerbal response
Motor function Motor responseMuscle toneTendon reflexesSeizures
serially and is one of the most important indi-cators of the need for further investigation.Thus when the level of consciousness can beseen to be improving there is no need tomake urgent decisions but when deteriorationoccurs then management decisions must bemade. It may of course be correct, when theprognosis is recognised to be hopeless, tomake a decision not to undertake furtherinvestigation or therapy.B) Brainstem functionThe brainstem reflexes are particularly impor-tant in helping to identify those lesions whichmay affect the reticular activating substance,explain the reason for coma and potentiallyhelp in identifying the viability of the patient.The reflexes used are predominantly relatedto the eyes and the pattern of respiration:-
(1) PUPILLARY REACTIONSThe size, equality and reaction of the pupilsto light is recorded. Unilateral dilatation ofthe pupil with loss of the light response sug-gests uncal hemiation or a posterior com-municating artery aneurysm. Midbrainlesions typically cause loss of the light reflexwith midposition pupils, pontine lesionscause miosis but a retained light response.Fixed dilatation of the pupils is an indicationof central diencephalic herniation and may bedifferentiated from the fixed dilatation due toatropine like agents by the use of pilocarpineeye drops which will cause miosis if thedilatation is due to loss of parasympatheticinnervation but be ineffective if it is pharma-cological. A Homer's syndrome may be seenipsilateral to a lesion in the hypothalamus,thalamus or brainstem when it will be associ-ated with anhydrosis of the ipsilateral side ofthe body, but can also be due to diseaseaffecting the wall of the carotid artery whenanhydrosis will only affect the face."5 Hepaticor renal failure and other forms of metaboliccoma may make the light reflexes appearunduly brisk and the pupils therefore relative-ly small. Most drug intoxications tend tocause small and sluggishly reactive pupils anda pontine haemorrhage will cause pin pointpupils due to parasympathetic stimulation.16
(2) CORNEAL RESPONSESThe comeal reflex is usually retained untilcoma is very deep. If it is absent in a patientwho is in otherwise light coma then the possi-bility of drug induced coma or of local causesof anaesthesia to the cornea should be consid-ered. The loss of the comeal response whendrug overdose is excluded is a poor prognos-tic sign.
(3) SPONTANEOUS EYE MOVEMENTThe resting position of the eyes and the pres-ence of spontaneous eye movements shouldbe noted. Conjugate deviation of the eyesraises the question of an ipsilateral hemi-sphere or contralateral brainstem lesion.Abnormalities of vertical gaze are less com-mon with a patient in coma but depression ofthe eyes below the meridian may be seen withdamage at the level of the mid brain tectum
593 on N
ovember 2, 2020 by guest. P
rotected by copyright.http://jnnp.bm
j.com/
J Neurol N
eurosurg Psychiatry: first published as 10.1136/jnnp.56.6.589 on 1 June 1993. D
and in states of metabolic coma. The restingposition of the eyes is normally conjugate andcentral but it may be dysconjugate whenthere is damage to the oculo-motor orabducens nerves within the brainstem oralong their paths.
Roving eye movements seen in light comaare similar to those of sleep. They cannot bemimicked and their presence excludes psy-chogenic unresponsiveness.13 Periodic alter-nating gaze or "ping pong" gaze is a repetitiveconjugate horizontal ocular deviation which isof uncertain aetiology."7 Spontaneous nystag-mus is rare in coma since it reflects interac-tion between the oculovestibular system andthe cerebral cortex. Retractory nystagmus inwhich the eyes jerk irregularly back into theorbit and convergence nystagmus may beseen with mid brain lesions.'8 Ocular bob-bing, an intermittent jerking downward eyemovement, is seen with destructive lesions inthe low pons and with cerebellar haematomaor hydrocephalus. 19
(4) REFLEX EYE MOVEMENTSThese are tested by the oculocephalic andoculovestibular responses. The oculocephalicor doll's head response is tested by rotatingthe patient's head from side to side andobserving the eyes. In coma with an intactbrainstem the eyes will move conjugately andin a direction opposite to the head move-ment. In a conscious patient such a responsecan be imitated by deliberate fixation of theeyes but is not common. In patients withpontine depression the oculocephalicresponse is lost and the eyes remain in themid position of the head when turned.The oculovestibular response is more accu-
rate and useful. It is elicited by instillingbetween 50-200ml of ice cold water into oneexternal auditory meatus. The normalresponse in the conscious patient is the devel-opment of nystagmus with the quick phaseaway from the side of stimulation. A tonicresponse with conjugate movement of the eyetowards the stimulated side indicates anintact pons and suggests a supratentorialcause for the coma. A dysconjugate responseor no response at all indicates brainstemdamage or depression. Both ears should bestimulated separately and if unilateral irriga-tion causes vertical eye movement the possi-bility of drug overdose arises because manydrugs affect lateral eye movement.The value of oculovestibular testing in
patients without lateralising eye signs is con-siderable because they identify the intactnessof the brainstem and corticopontine connec-tions but may also reveal the presence of anintrinsic brainstem lesion by causing dyscon-jugate eye posturing. In addition they are thedefinitive way of identifying patients in psy-chogenic coma who will show normal nystag-mus and frequently be distressed by themanoeuvre.
(5) RESPIRATIONModern techniques of assisted respirationand the need to examine patients in intensive
care units where their respiration is controlledcomplicates the assessment of normal respira-tory functions. If, however, the patient is seenbefore respiration is controlled then the pres-ence of long cycle periodic respiration sug-gests a relatively high brainstem lesion,central neurogenic hyperventilation implies alesion at the level of the upper pons and shortcycle periodic respiration, which carries apoor prognosis, is seen with lesions lower inthe brainstem. In general the presence ofregular rapid breathing correlates with pul-monary complications and a poor prognosisrather than with the site of neurologicaldisease in patients with coma.20
C) MotorfunctionAs part of the assessment of the GlasgowComa Scale it may have been appreciatedthat there is lateralisation in the individualpatient which implies a focal cause for thecoma. The observation of involuntary move-ment affecting the face or limbs and asym-metry of reflexes will help to support thispossibility. Focal seizures are an importantindicator of a focal cause for the coma andthe observation of more generalised seizuresor of multifocal myoclonus would raise thepossibility of a metabolic or ischaemic-anoxiccause for the coma with diffuse cortical irrita-tion. The testing of tone as part of the assess-ment of muscle function can be useful in thecomatose patient where it is possible to detectasymmetry of tone not only in the limbs butalso in the face.
By this stage of the management of thepatient in coma it should be possible to iden-tify those patients who are unconscious withfocal signs, those who are unconscious with-out focal signs but with the presence ofmeningismus and those who have loss of con-sciousness without either focal signs ormeningismus (table 3).
Investigations ofthe patient in comaThe relevant investigations to be undertakenin the individual patient will be identified bythe differential diagnosis. In general the roleof investigation in the patient in coma is tohelp establish the aetiology of that coma and
Table 3 Classification of differential diagnosis ofcoma
Coma without focal or lateralising signs and withoutmeningismus1 Anoxic-ischaemic conditions2 Metabolic disturbances3 Intoxications4 Systemic infections5 Hyperthermia/Hypothermia6 EpilesyComa without focal or lateralising signs but with meningealirritation1 Subarachnoid haemorrhage2 Meningitis3 EncephalitisComa with focal brainstem or lateralising cerebral signs1 Cerebral tumour2 Cerebral haemorrhage3 Cerebral infarction4 Cerebral abscess
594 on N
ovember 2, 2020 by guest. P
rotected by copyright.http://jnnp.bm
j.com/
J Neurol N
eurosurg Psychiatry: first published as 10.1136/jnnp.56.6.589 on 1 June 1993. D
will vary from simple blood tests throughmore complex blood tests, examination of theCSF, electrophysiological tests and imaginginvestigation. Although the electroencephalo-gram has some hierarchical value in theassessment of the depth of coma and hasbeen used to an extent to identify a prognosisof coma,2'22 its major role is in identifyingpatients who are in subclinical status epilepti-cus or have complex partial seizures becausethis will significantly alter their manage-ment.2' It may also be useful in distinguishingbetween feigned or psychiatric coma, inwhich it will be normal, and genuine cerebraldisease when it may show diffuse abnormali-ties or help to identify a focal lesion. Theprognostic value of the-EEG is probably notas great as that obtained from careful obser-vation of clinical signs.2'
Evoked potentials, predominantly brain-stem evoked potentials and somatosensoryevoked potentials, may give information relat-ing to the intactness of brainstem pathwaysand to the existence of a cortical component.Theoretically the use of brainstem evokedresponses could provide evidence for thepresence and site of brainstem disease and, asthey are relatively unaffected by drug coma,they may provide evidence on the aeti-ology.2425 Regrettably there is as yet little cor-relation between evoked response studies incoma and prognosis but it seems likely thatthe use of somatosensory evoked potentialsand brainstem auditory evoked potentials willbecome of value in identifying the prognosisof patients in coma. One technical problem isthe need to undertake these recordings in thebusy premises of an intensive care unit whereconsiderable other electrical interference isoccurring.
Brain imaging techniques including CTand MRI are important in coma in providingevidence of the diagnosis.26 The former has avery significant role to play in identifyingthose patients who have a structural cause forcoma though the latter has not yet been for-mally evaluated in this respect and there areproblems in inserting the patient in comatogether with necessary life support systemsinto the field of the MRI scan.
Other more complex techniques such asintracranial pressure monitoring and cerebralblood flow studies are rarely of help in thediagnosis of medical coma and their role inprognosis is not fully evaluated though theyare likely to be limited by their invasiveness.2'Measures of biochemical parameters in comaare predominently diagnostic but some mea-sures such as brain type creatinine kinase andneuron specific enolase in the cerebrospinalfluid may help in determining prognosis.28On clinical grounds patients can be allocat-
ed to one of three varieties of coma29:-1) Coma with focal signsExcept in those patients in whom an underly-ing and irreversible terminal disease is recog-nised, it is obligatory that CT scan or MRIscan be undertaken to identify the cause ofthe coma. This will define whether or not astructural abnormality is present and in many
instances give a clue to the underlying nature.If the CT scan is normal then the possibilityof a nonstructural focal abnormality antedat-ing the onset of coma or being part of thecoma, as occasionally happens with hypogly-caemia or hepatic encephalopathy, must beconsidered. If there is no focal structuralabnormality on a CT scan then other investi-gations including metabolic and CSF exami-nation should be carried out.
Once the image has been obtained thequestion of more definitive therapy, be itneurosurgical, the reduction of intracranialpressure by the use of steroids and mannitol,the application of a specific antibacterial orantiviral agent, or the use of chemotherapymay be considered.2) Coma with meningeal irritation but withoutphysical signsPatients in this group will usually be sufferingfrom subarachnoid haemorrhage, acute bac-terial meningitis, or viral meningo-encephali-tis. The distinction between infective andnoninfective can usually be made on the basisof fever and a lumbar puncture will beexpected to reveal the cause. It is a counsel ofperfection that, because of the theoreticalpotential of a collection of pus or of identify-ing the site of the subarachnoid haemorrhage,a CT scan should be undertaken before lum-bar puncture. In practice in many hospitalsthroughout the United Kingdom, CT headscanning is not easily available and the pres-ence of meningismus, particularly if associat-ed with fever, raises the possibility ofmeningitis and indicates the need for anassessment of the CSF.When CSF examination is undertaken by
lumbar puncture it is important to rememberthat an inadequate lumbar puncture does notpreclude the possibility of a pressure cone butmay prevent proper assessment of the CSF.Although some authorities still recommendthat only a few ml of fluid need be obtainedfor bacterial culture and cell count,'2 inpractice once the dura and arachnoid arebreached by a lumbar puncture needle thepossibility of herniation does not dependsolely upon the fluid which is collected butrather upon that which may leak out duringsubsequent hours. It is therefore importantthat when a decision to undertake a lumbarpuncture is made sufficient CSF is obtainedto enable an adequate assessment of the cellcount, a gram stain and provide fluid for cul-ture and antibody analyses together with ameasure of the total protein and sugar.
In those centres in which a CT scan is avail-able the detection of blood in the subarach-noid space at CT scan precludes the need forlumbar puncture but whether or not lumbarpuncture has been carried out to identify thepresence of subarachnoid haemorrhage thepatient should then be transferred to a neuro-surgical unit, probably be given intravenousnimodipine, and be subjected to angiographyand surgery. In general those patients who arein coma from subarachnoid haemorrhage areless of a surgical emergency than those whohave higher states of consciousness.
595 on N
ovember 2, 2020 by guest. P
rotected by copyright.http://jnnp.bm
j.com/
J Neurol N
eurosurg Psychiatry: first published as 10.1136/jnnp.56.6.589 on 1 June 1993. D
3) The presence of coma without focal signs ormeningismusThese patients are likely to have a metabolicor anoxic cause for their coma. One of thecommonest causes remains that of drug over-dose and it is appropriate to withdraw bloodto send to the toxicology laboratories inpatients presenting in this way. In generalthere will be a clue from the circumstances inwhich the patient was discovered and fromthe previous history. Reliance is placed uponthe assessment of metabolic and toxicmetabolites in the blood and evidence shouldbe sought for hepatic failure, renal failure,hyperglycaemia, hypoglycaemia, and distur-bances of electrolytes or acidosis. The ma-jority of commonly available drugs can nowbe assayed within blood and serum enzymesshould also be estimated. Problems inevitablyarise when patients who are unconscious havebeen consuming alcohol and an assessment ofthe relevant importance of this in causing theunconsciousness may be difficult. Again theproblem may be helped by the expedient ofmeasuring blood alcohol levels.
Perhaps the most important single cause ofunresponsiveness, which is directly treatableand correctable, is that of hypoglycaemia andthis should have already been covered duringthe initial resuscitation of the patient. By thistime in management the formal level of bloodsugar will have been estimated and appropri-ate treatment for hypo and hyperglycaemiamay be instituted. The treatment of acid baseabnormalities will require not only the routinebiochemistry but arterial blood gas analysis tomonitor progress. Usually a patient who hassuffered from hypoxia or ischaemia will havebeen identified by the mode of presentationand by the normality of investigations thusfar. The possibility of poisoning with carbonmonoxide should be considered and excludedby the measurement of carboxyhaemoglobin.In general, patients who have suffered anoxicor ischaemic insult should be given 100%oxygen and the monitoring of PAO2 will beimportant together with the maintenance ofadequate circulation and oxygenation.
Patients who are in shock or hypertensiveencephalopathy will be diagnosed by the levelof blood pressure and those with disturbedtemperature regulation by use of the ther-mometer, though a rectal thermometer maybe required. These causes can then be cor-rected.
In patients with drug overdose the possibil-ity of using specific antidotes should be con-sidered. The use of Naloxone in patients inwhom there is a high index of suspicion ofopioid poisoning and Benzodiazepine antago-nists in self-poisoning with Benzodiazepine.The use of analeptic agents in barbituratepoisoning cannot now be supported.'0 Con-sideration should also be given to clearing theingested toxin from the stomach, the passageof a nasogastric tube should usually be con-sidered and this is one indication for intuba-tion of the trachea to prevent the risk ofaspiration. The importance of the diagnosisof drug overdose coma is that such patients
have a good prognosis provided they aregiven adequate respiratory and circulatorysupport during their unconsciousness. Theyare, however, particularly liable to showdepression of brainstem responses and if thepossibility of drug overdose is not consideredtheir level of coma may be misinterpreted andtheir prognosis might be thought undulypessimistic.
The prediction of outcome in comaHaving made an assessment of the cause ofcoma, established its severity and introducedappropriate treatment, the physician shouldbe able to identify the likely outcome to col-leagues and to friends and relatives of thepatient. Sedative drugs or alcohol overdoseare not usually lethal and carry a good prog-nosis provided that circulation and respira-tion is protected. The physician canreasonably give a good prognosis in patientssuffering from self-poisoning with sedativedrugs provided that those complications ofcardiac arrhythmia, aspiration pneumoniaand respiratory arrest are avoided or correct-ed. In non-traumatic coma other than thatwhich is drug induced those factors whichdetermine the outcome have been defined"3and include the cause of the coma, the depthof the coma, the duration of coma and certainclinical signs, among the most important ofwhich are brainstem reflexes. Overall only15% of patients in non-traumatic coma formore than six hours will make a good ormoderate recovery; the other 85% will die,remain vegetative, or reach a state of severedisability in which they remain dependent.Patients whose coma is due to metabolicreasons, including infection, organ failure andbiochemical disturbances, have a better prog-nosis. Thirty five per cent of patients willachieve moderate or good recovery; of thosewhose coma follows hypoxic ischaemic insultonly 11% make such a recovery; of those incoma due to cerebrovascular disease only 7%can be expected to make such a recovery.Twenty per cent of patients in coma follow-ing hypoxic ischaemic injury will enter thevegetative state due to the likelihood ofhypoxic ischaemia resulting in bihemisphericdamage with relative sparing of the brain-stem.
Apart from the diagnosis the depth ofcoma affects the individual prognosis. Thosepatients not showing eye opening after sixhours of coma have only a 10% chance ofmaking a good or moderate recovery whereasthose whose eyes opened in response topainful stimuli have a 20% chance of makinga good recovery. The longer the coma persiststhe less likely there is to be recovery; 15% ofpatients in coma for six hours make a good ormoderate recovery compared with only 3%who remain unconscious at one week.3"The study of 500 patients reported by
Levey et al 31 using prospective data frompatients with clearly defined levels of coma,diagnoses and outcomes, showed that someclinical signs are significantly associated witha poor prognosis: in the total cohort of 500
Bates on N
ovember 2, 2020 by guest. P
rotected by copyright.http://jnnp.bm
j.com/
J Neurol N
eurosurg Psychiatry: first published as 10.1136/jnnp.56.6.589 on 1 June 1993. D
patients corneal reflexes were absent 24 hoursafter the onset of coma in 90 patients and thissign was incompatible with survival. In a
more uniform group who suffered anoxicinjury there were 210 patients: 52 of thesehad no pupillary reflex at 24 hours, all ofwhom died. By the third day 70 were left witha motor response poorer than withdrawal andall died. By the seventh day the absence ofroving eye movements was seen in 16 patientsall of whom died. The confidence intervalsfor all of these criteria were 0 95 (table 4).
At the opposite end of the scale more than25% of patients who show roving conjugateeye movements within six hours of the onsetof coma or who show withdrawal responses topain or eye opening to pain, will recover inde-pendence and make a moderate or goodrecovery. The use of combinations of clinicalsigns help to improve the accuracy of progno-sis: at 24 hours the absence of a comealresponse, pupillary light reaction or caloric or
doll's eye response is not compatible withrecovery to independence. Patients who are
able to speak words within 24 hours or whoshow nystagmus on caloric testing are likelyto make a good recovery (table 5)."The most accurate prediction of outcome
in a patient in medical coma is still that which
Table S Prediction ofoutcome ofcoma at 24 hours by a combination of clinical signs
Percentage ofpatients with different outcomes
DIPVS SD MDIGR
500 patients
+Any two reacting:
PupilsCorneals > No (120) 97 2 1Oculovestib
i Yes
Motor better than flaccid No (183) 80 8 12
Yes
Motor withdrawal No (135) 69 14 17
+ Yes
Verbal moans No (106) 58 19 23
No- Yes (56) 46 13 41
Summarised from Levy et al.3'
is obtained from the use of clinical signs andthere is little to be added by more sophisti-cated testing other than in identifying thecause of the coma. It is possible to predictthose patients who will not make a recoveryand who will die in coma or who will enter avegetative state within the first week of coma.It is rare for patients in medical coma who arein a vegetative state at one month to showany form of recovery.33
Continuation of careThe long term care of patients in comamaybe undertaken in an intensive care unit,on a specialist ward or later in a long stayhospital. It is important that those patients inwhom prognosis is hopeless should not bepermanently exposed to the rigors of inten-sive care medicine, but should continue toreceive basic care within routine hospitalwards. So long as patients are considered tohave a potential for recovery they should belooked after in intensive care units or on spe-cialist wards. Their respiration, skin, circula-tion and bladder and bowel function needattention, seizures controlled and the level ofconsciousness regularly assessed and moni-tored. It is important that the mobility ofjoints and circulation to pressure areas aremaintained during the long term care of thepatient and the possibility of aspiration pneu-monia, peptic ulceration and other complica-tions of long term intensive care beconsidered and avoided. Techniques such asmechanical ventilation and the use of steriodtherapy are not to be used routinely in themanagement of the comatose patient as theydo not improve prognosis and may specifi-cally compromise recovery.'4
The persistent vegetative stateThe relative resilience of the brainstem allowsit to survive injuries which may create irre-versable damage to the cerebral hemispheresand then the patient will enter that statedefined as vegetative. Retrospectively, afterpostmortem examination, it may be possibleto identify massive neocortical damage whichwill indicate that the patient was permanentlyin the vegetative state,35 but there are noclinical or laboratory means of confirmingthis before postmortem, and therefore theterm persistent vegetative state rather thanpermanent vegetative state is used clinically.Specialists in rehabillitation are concernedthat physicians may take the attitude thatthere is no point in treating such patientstherefore creating a self fulfilling prophecy ofpoor prognosis, no treatment and pooroutcome.36
There is continuing debate as to the poten-tial for recovery for patients who are vegeta-tive. In patients who have suffered nontraumatic injury such as anoxia andischaemia, the prognosis for recovery fromthe vegetative state is poor after the first fewweeks. There are some reports of patientswho have suffered coma as a result of headtrauma in whom an improvement from the
597 on N
ovember 2, 2020 by guest. P
rotected by copyright.http://jnnp.bm
j.com/
J Neurol N
eurosurg Psychiatry: first published as 10.1136/jnnp.56.6.589 on 1 June 1993. D
vegetative state has been recognised aftermonths, but these anecdotal cases of recoveryfrom vegetative state are difficult to validateand it seems possible that such patients werenot truly vegetative, rather in a state of pro-found disability but with cognition, at thebeginning of the observation.37-39
Investigations do not help to identify vege-tative state because many types of EEG pat-tern have been recorded from near normalityto a flat record and CT scans usually showconsiderable cortical atrophy with ventriculardilatation. Somato-sensory evoked responsesare said to show loss of the cortical compo-nent and PET scans to show cortical meta-bolic under activity. None is diagnostic intheir results.40
Patients in a persisting vegetative state willoften have received artificial ventilation atsometime during their initial coma or resusci-tation, but they are not truly respiratordependent and, since they are able to breathe,are only dependant upon carers for the supplyof liquid and nutrients and the prevention ofcomplications. The management of individ-ual patients will depend upon circumstancesand other aspects of the diagnosis and consid-ered prognosis.4'
1 Bates D, Cartlidge NEF. The Prognosis of MedicalComa. In: Tunbridge WMG, ed. Advanced medicine.London: Pittman Medical, 1981.
2 Plum F, Posner JB. 7he diagnosis of stupor and coma, 3rded. Philadelphia: Davis, 1980.
3 Brodal A. Neurological anatomy in relation to clinicalmedicine, 3rd ed. Oxford: Oxford University Press,1981.
4 Jouvet M. The role of monoamines and acetyl cholinecontaining neurones in the regulation of the sleep/wakecycle. Rev Physid 1972;64:166-307.
5 Defeudis FV. Cholinergic roles in consciousness. In:Defeudis FV, ed. Central chonergic systems and behav-iour. London: Academic Press, 1974.
6 Tinuper P. Idiopathic recurring stupor: A case withpossible involvement of the gamma aminobutyricacid(GABA)ergic system. Ann Neurol 1992 31:503-6.
7 Medical Research Council Brain Injuries Committee. Aglossary of psychological terms commonly used in casesof head injury. MRC War Memorandum. London:HMSO, 1941.
8 Teasdale G, Jennett B. Assessment of coma and impairedconsciousness: a practical scale. Lancet 1974;2:81-4.
9 Jennett B, Plum F. The persistent vegetative state: a syn-drome in search of a name. Lancet 1972;1:734-7.
10 Cairns H. Disturbances of consciousness with lesions ofthe brain stem and diencephalon. Brain 1952;75:109-46.
11 Feldman MH. Physiological observations in a chronic caseof locked in syndrome. Neurol 1971;21:459-78.
12 Harris JO, Berger JR Clinical approach to stupor andcoma. In: Bradley WG, Daroff RB, Fenichel GM,Marsden CD, eds. Neurology in nical practice. London:Butterworth, 1991.
13 Fisher CM. The neurological examination of thecomatose patient. Acta Neurol Scand 1969;45(suppl 46):1-56.
14 De Jong RN. The neurological examination, 4th ed.Haggerstown: Harper and Rowe, 1979.
15 Crill WE. Homer's syndrome secondary to deep cerebrallesions. Neurology 1966;16:325-7.
16 Walsh FB, Hoyt WF. Clinical neuro-opthalmology, 3rded. Baltimore: Williams and Wilkins, 1969.
17 Stewart JD, Kirkham TH, Mathieson G. Periodic alter-nating gaze. Neurology 1979;29:222-4.
18 Daroff RB, Hoyt FF. Supranuclear disorders of ocularcontrol systems in man. In: Bach-y-Rita, Collins, Hyde,eds. The control of eye movements. New York: AcademicPress, 1971.
19 Fisher CM. Ocular bobbing. Arch, Neurol 1964;11:543-6.20 Leigh RJ, Shaw DA. Rapid regular breathing in uncon-
scious patients. Arch Neurol 1976;33:356-61.21 Jorgensen EO, Malchow-Moloer A. Natural history of
global and critical brain ischaemia. Resuscioion1981;9:133-91.
22 Synek VM. EEG abnormality grades and subdivisions ofprognostic importance in traumatic and anoxic coma inadults. Clin Electroencephalogr 1988;19:160-6.
23 Engle J, Ludwig BI, Fetell M. Prolonged partial complexstatus epilepticus: EEG and behavioural observations.Neurology 1978;28:863-9.
24 Papanicolaou AC, Loring DW, Eisenberg HM, et al.Auditory brain stem evoked responses in comatosedhead injured patients. Neurosurgery 1986;18:173-5.
25 Walser H, Emry M, Janzer R Somatosensory evokedpotentials in comatosed patients: correlation with out-come and neuropathological findings. J Neurol 1968;233:34-40.
26 Tasker RC, Matthew DJ, Kendall B. Computed tomo-graphy in the assessment of raised intracranial pressurein non-traumatic coma. Neuropaediatrics 1990;21:91-4.
27 Jaggi JL, Obrist WD, Jennarelli TA, Langfitt TW.Relationship of early cerebral blood flow and metabo-lism to outcome inacute head injury. J. Neurosurg1990;72:176-182.
28 Roine RO, Somer H, Caste M, et al. Neurological out-come after out of hospital cardiac arrest: prediction bycerebrospinal fluid enzyme analysis. Arch Neurol 1989;46:753-6.
29 Adams RD, Victor M. Principles of neurology, 4th ed. NewYork: McGraw Hill, 1989.
30 Schwartz GR. Emergency toxicology and general prin-ciples of medical management of the poisoned patient.In: Schwartz, Safar, Stone, et al, ed. Principles and prac-tice of emergency medicine. London: WB Saunders, 1978.
31 Levy DE, Bates D, Caronna JJ, et al. Prognosis in non-traumatic coma. Ann Intern Med 1981;94:293-301.
41 ANA Committee on Ethical Affairs. Persistant vegetativestate: report of the American Neurological AssociationCommittee on Ethical Affairs. Ann Neurol 1993;33:386-90.
598 on N
ovember 2, 2020 by guest. P
rotected by copyright.http://jnnp.bm
j.com/
J Neurol N
eurosurg Psychiatry: first published as 10.1136/jnnp.56.6.589 on 1 June 1993. D
