States of anxiety and their induction by drugs - NCBI

Page created by Todd Hines
 
CONTINUE READING
Br. J. clin. Pharmac. (1986), 22, 251-261

States of anxiety and their induction by drugs

MALCOLM LADER & MALCOLM BRUCE
Department of Pharmacology, Institute of Psychiatry, De Crespigny Park, London SE5 8AF
1 Syndromes of anxiety include generalized anxiety states, various forms of phobic
disorder and panic attacks. It is unclear whether panic attacks are a separate syndrome
from anxiety states or a more severe form.
2 Drug-induced states of anxiety should provide useful models of the mechanisms of
anxiety and its treatment. High-risk populations might be identifiable.
3 Catecholamine infusions produce marked peripheral changes without fully reproducing
the central feelings. Lactate infusions also produce anxiety-like states lacking full credibility.
4 Experience with the benzodiazepine-receptor contragonists, the ,-carbolines, is limited
but panic states have been reproduced following their use. Caffeine produces an anxiety
state in high dose and some panic states have been induced.
5 The critical evaluation of drug-induced anxiety states is a promising way of elucidating
the mechanisms, psychological and physiological, associated with clinical anxiety.

Keywords anxiety panics yohimbine ,-carbolines caffeine
Introduction
Anxiety is a common human emotion, character-    some act'. For example: 'I am anxious to keep
ized by a feeling of overwhelming apprehension   your interest and attention'. The second meaning
and accompanied by many peripheral physio-       is 'being troubled in mind about some uncertain
logical changes. The latter, being fairly easy toevent, being in disturbing suspense, being fraught
measure, have provided a deceptively simple way  with worry'. The first type of anxiety is goal-
of studying anxiety but at the risk of oversimpli-
                                                 directed and controllable; the second type is dif-
fying the relationship between the central 'core'fuse, causing the person to feel passive and help-
properties of anxiety and the peripheral epiphen-less. The latter meaning includes the indescribable
omena. Attempts to induce anxiety have often     feeling of foreboding that is the core of anxiety;
failed to address themselves to this problem.    It is irreducible scientifically so one must rely on
Indeed, it is only partially soluble as use must be
                                                 subjective reports of this state.
made of subjective reports with inevitable scien-   Anxiety as part of everyday experience is a
tific reservations. In addition, the clinical classi-
                                                 normal emotion, as is fear. If the cause is clear,
fication of anxiety states and related conditionsthe emotion induced is labelled 'fear'; if the cause
remain rudimentary despite attempts to develop   is obscure, or although identifiable, its potential
rational schemata.                               impact is unpredictable, the more diffuse emotion
   This brief review will first outline the various
                                                 experienced is termed 'anxiety'. However, anxiety
syndromes of anxiety. Then it will review models may also be an abnormal emotion, and distinc-
of anxiety which have been developed using       tions between normality/abnormality and nor-
various pharmacological agents and evaluate them mal/clinical can be difficult. Clinical (morbid,
for their relevance to clinical syndromes. It will
                                                 pathological) anxiety can be defined operation-
not discuss animal models of anxiety (Lal &      ally by the need of the sufferer to seek relief from
Emmett-Oglesby, 1983; Kandel, 1983).             his anxiety. Such anxiety may be too severe, too
                                                 persistent or too pervasive for the person to
                                                 tolerate. He seeks medical advice, thereby be-
Definition of anxiety                            coming a 'patient' and the condition an 'anxiety
                                                 state'. Thus, a 'normal' subject may not seek
Anxiety has two main meanings. To be anxious help, although his anxiety levels might actually
can mean 'full of desire and endeavour to effect be 'greater' than those of the 'patient'.
252      M. Lader & M. Bruce
   Another distinction is between state and trait      dyspnoea, palpitations, chest pain, choking sen-
anxiety. State anxiety refers to anxiety felt at one   sations, dizziness, feeling of unreality, paraes-
particular time, the moment of enquiry. Trait          thesiae, hot and cold flushes, sweating, faintness
anxiety is the habitual tendency to feel anxious.      and trembling. Attacks usually last minutes but
Some individuals have such high levels of trait        may persist for hours. Nervousness and appre-
anxiety that they are chronically in a condition of    hension may develop between attacks. In gener-
state anxiety.                                         alized anxiety disorder, the anxiety does not
                                                       come in attacks but is persistent and accompanied
                                                       by symptoms of motor tension, autonomic hyper-
Clinical aniety                                        activity, apprehensive expectation, vigilance and
                                                       scanning of the environment. In the obsessive
Clinical anxiety can thus be defined operationally     compulsive disorder, recurrent ideas, thoughts,
as an emotional state, disproportionate to any         images or impulses incessantly invade conscious-
apparent cause, for which the sufferer seeks           ness and seem senseless or repugnant to the
help. However, several syndromes have been             person. The obsessions may be accompanied by
delineated since Freud's initial description in        ritualistic stereotyped behaviours (compulsions):
1894 (Freud, 1894). The American Psychiatric              The third main group are the post-traumatic
Association's Diagnostic and Statistical Manual,       stress disorders, acute, chronic or delayed. Diag-
third edition, (DSM-III), has a new and rather         nosis depends on recognizing the existence of a
complex classificatory scheme (Brown et al.,           stressor, such as accidents, natural disasters,
1984). The anxiety disorders are divided into          combat or torture, that would evoke significant
four categories; phobic disorder, anxiety states,      symptoms of distress in almost everyone. The
post-traumatic stress disorders and atypical           post-traumatic stress disorders are pathological
anxiety disorder. Hypochondriacal, hysterical          reactions to these events, with re-experiencing
and depressive disorders are placed elsewhere.         of the trauma as evidenced by recurrent and
   Phobic disorders are characterized by fear and      intrusive recollections of the event, recurrent
avoidance of certain situations or objects with        dreams of the event and suddenly feeling the
subsequent restriction of social or role function-     event was happening again. Numbing of respon-
ing. The phobic disorders are subdivided into          siveness to and reduced involvement with the
three types. In agoraphobia, the individual has        external world ensues. Other symptoms include
marked fear of, and thus avoids, being alone or        exaggerated startle response, sleep disturbance,
in public places from which escape might be            guilt about surviving, and impairment of memory
difficult or help unavailable, such as crowded         and concentration.
shops, tunnels and public transportation. The             Finally, atypical anxiety disorder is a ragbag
individual's normal activities become increasingly     residual category for individuals who cannot be
restricted. Agoraphobia may include discrete           placed elsewhere.
panic attacks but can occur in their absence. The         Of course, other diagnostic and taxonomic
essential feature of social phobia is that the         systems are extant but none include such carefully
individual has a persistent, irrational fear of, and   constructed diagnostic criteria as does the DSM-
compelling desire to avoid, situations in which        III. Nevertheless, some question the utility of
he may be exposed to public scrutiny. He also          the DSM-llI anxiety disorder categories in clinical
fears that he may behave in a manner that will         usage, in particular noting that mixed syndromes
be humiliating or embarrassing. As with agora-         are common both cross-sectionally and over the
phobia, anticipatory anxiety may be marked.            natural history of the condition (Tyrer, 1984).
Examples include fears of speaking, performing            These different syndromes reflect the recog-
or eating in public, and using public lavatories.      nition by the clinicians of different types of
The third phobic disorder is simple phobia             anxiety with respect to symptom patterns and
('specific phobia') in which the individual fears      temporal relationships. Further support comes
and avoids animals such as dogs, snakes, birds,        from differing pharmacological responses: phobic
mice and insects, or closed spaces (claustro-          and panic disorders respond well to antidepres-
phobia), heights (acrophobia), elevators, etc. It      sants, but not to benzodiazepines, and generalized
is a very common condition but only occasionally       anxiety disorder conversely. These distinctions
leads to marked restriction or modification of         should be borne in mind when considering the
activities.                                            effects of drugs purporting to induce anxiety.
   In anxiety states no specific phobic object or
situation is identified. The essential features of
panic disorder are recurrent attacks of intense The rationale of inducing anxiety with drugs
apprehension, fear or terror, often associated
with feelings of impending doom, and accom- As anxiety is such an unpleasant feeling, studies
panied by a range of bodily symptoms such as attempting to deliberately induce it need scien-
States of anxiety and their induction by drugs             253
 tific justification to balance against ethical reser-   readily pass the blood-brain-barrier, such prob-
 vations. Several benefits might accrue:                 lems are very germane to the interpretation of
    The main reason is that the reliable and repro-      the observations.
 ducible induction of a clinical phenomenon is              Maranion (1924), in a descriptive report, em-
 usually very helpful in studying the pathological       phasizes the 'as if ' nature of the anxiety engen-
 mechanisms associated with that disorder. In            dered in normal volunteers. Physiological
most branches of medicine animal models have             changes such as palpitations and trembling are
been established but these are never convincing          noted but the emotional feeling is 'cold', partial
in psychiatry because of the absence of subjective       and lacks conviction to the subject. However,
reports (Crawley et al., 1984). Study of the             some subjects developed full blown anxiety and
mental and bodily states induced by reserpine            not just feeling 'as if I were anxious'. In another
has increased our understanding of mechanisms            study (Cantril & Hunt, 1932) some of the reac-
in depression and psychomotor retardation, as            tions to intramuscular adrenaline were described
has the use of amphetamine and L-dopa in schizo-         by normal volunteers as quite pleasant but
phrenia-like syndromes. An anxiety-inducing              others became anxious.
procedure should be even more heuristic because             Basowitz and his co-workers (1956) interviewed
of the resemblance between anxiety states in             12 normal subjects (medical interns) to ascertain
normal and clinical contexts.                            any past history of stress and to note any specific
   Such a model could be used to evaluate anti-          reactions to it such as palpitations, apprehension,
anxiety medications and non-pharmacological              tremor and perspiration. Using double-blind pro-
treatments. Much development work could take             cedures and a cross-over design, adrenaline was
place in normal human volunteers before tackling         infused intravenously in a dose of 5,ug kg-1 body
the greater problems of therapeutic trials in            weight h-1 and its effects compared with those of
patients. Induction of anxiety may itself be             saline. Adrenaline produced a distinct rise in
capable of exploitation as a therapeutic pro-            pulse pressure, averaging 20 mm Hg, and a
cedure in that patients can be taught strategies to      tachycardia, averaging 13 beats min-. Hand
deal with that anxiety, thereby gaining experience       steadiness and physical persistence were impaired.
and confidence in coping with their naturally-           The commonest subjective experience was pal-
occurring anxiety (Bonn et al., 1971).                   pitations. Symptoms were reported with the
   High risk populations might be identified using       saline testing as well as on the adrenaline
drug induction of anxiety as a challenge pro-            occasions, but only half as frequently and usually
cedure (Wearn & Sturgis, 1919). Thus, indivi-            when saline was the first treatment. Adrenaline
duals at risk of developing anxiety symptoms             produced symptoms which generally resembled
could be identified. Similarly, particular indiv-        those elicited at the initial interview as occurring
iduals might be especially sensitive to specific         in the subjects in response to stress. In emotion-
compounds, such as caffeine which may induce             ally labile subjects, excess symptoms but few
or exacerbate anxiety feeling (Greden, 1974).            cardiovascular changes were noted; in rigid per-
   Finally, insights might be gained into the tax-       sonalities, no symptoms but marked physiological
onomy of anxiety disorders, as to whether the            changes ensued.
subgroups outlined earlier are discrete entities.           Similar studies have been conducted in Sweden.
Thus, if the administration of an anxiogenic             In an initial experiment in six students, adrenaline
substance produced generalized anxiety but never         i.v. led to a greater distress than did noradrena-
precipitated panics, whereas another compound            line (Frankenhaeuser et al., 1961). An infusion
acted vice versa, this would support the indepen-        of both adrenaline and noradrenaline (0.28 mg
dence of the two syndromes.                              of each over 26-41 min lowered systolic and
                                                         diastolic blood pressures without affecting pulse
                                                         rate (Frankenhaeuser & Jarpe, 1962). This
Adrenergic induction of anxiety                          suggests that the adrenaline's effects predom-
                                                         inated. Amongst the 11 subjects, palpitations,
Catecholamines have long been known to be                tremor, and a general feeling of discomfort were
released in states of anxiety and suspected of           reported by most subjects: restlessness, appre-
reinforcing the state (Kopin, 1984). Many                hension, tenseness and dyspnoea by about half.
studies have used adrenaline and/or noradrena-           In a similar experiment, an infusion of adrenaline
line to induce anxiety but suffer from a number          produced a tachycardia, a rise in systolic and a
of drawbacks. Some reflect design problems but           drop in diastolic blood pressure (Frankenhaeuser
many fail to address the complex scientific and          & Jarpe, 1963). The intensities of the physio-
philosophical problems that beset the relation-          logical and the subjective reactions were un-
ship between the central and peripheral mechan-          related. Subjective estimates of effects declined
isms of anxiety. As the catecholamines do not            steadily as each infusion proceeded. Males
254       M. Lader & M. Bruce
seemed more aware of the body sensations follow-        4. False information about the injection may not
ing intramuscular injections of adrenaline than         have prevented the 'misinformed' subjects from
did females (Fast & Fisher, 1971).                      inferring correctly that their bodily symptoms
    The widely cited experiments of Schachter &         were caused by the injection.
Singer (1962) highlight the interactions between           These problems illustrate the difficulties in
cognitive factors and physiological arousal. They       designing and executing informative experi-
injected small doses of adrenaline, some sub-           ments in this area. Surprisingly, despite the
jects knowing what effects to expect, others            pivotal importance of this theory in the area of
remaining in ignorance. The subjects were then          emotions, replications have been few. Erdmann
placed with a stooge who acted either in a              & Janke (1978) tried to obviate as many as
euphoric manner or angrily. Observation of the          possible of the above objections. Drug admini-
subject and his self-report both indicated that         stration (ephedrine) was disguised so the sub-
subjects ignorant of the effects of their injection     jects were unaware they had been administered
showed and felt more emotional experience               anything. Four situational conditions were used
(euphoria or anger) than informed subjects. In a        - a neutral condition and 'anger', 'happiness'
second experiment, subjects given either adrena-        and 'anxiety'. The subject read a report on his or
line, a placebo or chlorpromazine watched a             her performance on a previously completed in-
comedy film. The adrenaline subjects showed             telligence test and answered supplementary
more amusement than those given the cx-                 questions designed to elicit angry or happy
adrenoceptor blocking agent, chlorpromazine             responses, or was threatened with electric shocks.
(Schachter & Wheeler, 1962). Thus awareness             Even so, the four conditions differed in respects
of physiological arousal seemed to be the sub-          other than the emotion to be induced. Mood was
strate on which cognitive clues induced a specific      assessed using an adjective checklist which
emotion.                                                yielded a scale of general well-being, an anxiety
    Schachter (1966) presented a new theory of          scale, and an excitement scale.
emotions based on this work. According to this             The results were only partly in accord with
theory, emotions arise from an interaction              Schachter's theory. The effects of the 'happiness'
between a state of physiological arousal and            and 'anger' situations were consistent with the
cognitive information derived from the situation.       theory and the ephedrine increased the emotional
Three assumptions underlie the theory:                  ratings. With the 'anxiety' situation, which is our
 1. An individual labels an inexplicable state of       present concern, the results were unexpected.
physiological arousal according to situational          Differences between anxiety scale ratings with
cognitive clues.                                        the anxiety and the neutral situations were found
2. If an immediate and totally congruent explan-        after placebo administration but not following
ation is forthcoming, the person will not feel an       ephedrine, when the two states were indistin-
emotion.                                                guishable. This is contrary to Schachter's theory
3. Even if cognitive clues are forthcoming, an          but the authors were at a loss to explain their
emotion will result only insofar as the underlying      findings. It could be speculated that anxiety was
physiological arousal is raised.                        already maximal under the placebo condition.
    However, several problems of method include:           In a further experiment, Erdmann & van
1. The placebo and adrenaline groups did not            Lindem (1980) administered the 3-adrenoceptor
differ in the degree of emotional reaction,             antagonist, oxprenolol, or placebo, or the ,1-
whether euphoria or anger was induced. Thus,            adrenoceptor agonist, orciprenaline, to students
clear evidence was lacking that the degree of           in one of two situations; one designed to induce
emotional reaction depends on the level of evoked       anger, the other neutral. The expected physio-
arousal. Against this must be set the limitation        logical changes were induced by the drugs, and
that adrenaline does not induce a credible              the 'anger' situation led to a significant rise in
arousal state; as many studies have shown, sub-         blood pressure, especially diastolic. The 'anger'
jects report only feeling 'as if ' they were anxious.   situation led to an increase in self-reports of
2. The emotional situations were poorly stan-           anger only in the placebo subjects but to increases
dardized in that the confederate's behaviour was        in anxiety ratings both in the placebo and the
not strictly controlled, and varied with the sub-       orciprenaline subjects.
ject's reactions, possibly confounding effects.            The circulatory effects of the 3-adrenoceptor
3. The emotional situations were also poorly            stimulant, orciprenaline, more closely resem-
matched. The euphoria-inducing situation and            bled those accompanying anxiety than those
the anger-inducing situation differed with respect      associated with anger. This implies that self-
to duration, the type of subject's activity and the     perception of the physiological pattern out-
type of behaviour rating as well as the intended        weighed the cognitive prompting provided by
affective differences.                                  the anger-inducing situation as anxiety, not
States of anxiety and their induction by drugs           255
anger, ratings were increased.                          caused patients to feel uncomfortable, restless,
   A recent study exemplifies the current interest      and irritable, and to appear tense and anxious;
in the biological basis of panic states. Nesse et al.   and heart rate and systolic blood pressure were
(1984), used increasing bolus doses of i.v. iso-        elevated. Yohimbine produced more physio-
prenaline in eight patients with panic attacks and      logical hallmarks of anxiety (13/13) than did
six normal control subjects. Patients had higher        adrenaline (6/13). The second infusion of saline
resting heart rates and increased plasma con-           produced considerable anxiety implying rapid
centrations of adrenaline, cortisol and growth          conditioning to the test room situation.
hormone than did controls. Plasma noradrenaline            As the pharmacological properties of yohim-
levels were only marginally increased and the           bine became clearer, namely, that it was primarily
heart rate response to the infusions were de-           an a2-adrenoceptor antagonist, yohimbine
creased relative to the normals. It was suggested       administration was used as an index of such
that the 1-adrenoceptor response may actually           autoreceptor function. A 20 mg dose had signi-
be diminished in patients with panic attacks. If        ficant anxiogenic effects in eight normal volun-
this is so, it might reflect a down-regulation of       teers (Charney et al., 1982). A 30 mg challenge
these receptors consequent on the repeated              in 10 volunteers produced in addition increases
release of catecholamines during states of anxiety      in systolic blood pressure and autonomic symp-
and panic.                                              toms such as piloerection and rhinorrhea. Plasma
                                                        free MHPG (3-methoxy 4-methoxy phenylethy-
                                                        lene glycol), a noradrenaline metabolite, con-
Yohimbine                                               centrations were substantially increased. Both
                                                        diazepam and clonidine antagonized yohimbine-
As mentioned above, catecholamines are not              induced anxiety but only clonidine attenuated
ideal as anxiogenic substances because they act         the increase in plasma MHPG, blood pressure
peripherally and the anxiety induced is secondary       and autonomic symptoms (Charney et al., 1983).
to those physiological changes. Yohimbine is a          Yohimbine can precipitate panic attacks in
more satisfactory sympathomimetic agent but             patients diagnosed as suffering from panic dis-
with complex actions. At low dose it appears to         orders (Uhde et al., 1983). The effects of
block ct2-adrenoceptors thereby enhancing               yohimbine were assessed in 39 drug-free patients
neural release of noradrenaline; at higher doses        with agoraphobia and panic attacks, as compared
it blocks a1-adrenoceptors producing an                 with 20 normal controls (Charney et al., 1984a).
adrenolytic action. It was used extensively by          Yohimbine (20 mg orally) was associated with
Gershon's group. In the first study, 51 male            greater self-ratings of anxiety, nervousness,
mental hospital subjects and 9 normal volunteers        palpitations, hot and cold flushes, restlessness,
were given 0.5 mg kg-' (Holmberg & Gershon,             tremors, piloerection, and sitting systolic blood
1959). Psychological changes were noted                 pressure in the patients than in the normals.
'simulating an anxiety state': the subjects             MHPG plasma concentrations were higher in
became tense, anxious, restless, irritable and          the patients, especially those with frequent panic
tremulous. The degree of autonomic response             attacks. It was suggested that panicky patients
correlated highly with the subject's initial anxiety    have heightened brain noradrenergic activity.
level. Imipramine (1 mg kg-1) was given three              Thus, yohimbine is more effective than adrena-
times daily for 3 weeks to some of the sub-             line which in turn is more effective than nor-
jects. Marked potentiation of yohimbine effects         adrenaline in inducing both central and peripheral
ensued with tremor and restlessness amounting           anxiety-like symptoms.
to an acute panic. Ingram (1962) used a                    Other drugs acting on the adrenergic system
yohimbine challenge (0.5 mg kg-' infused over 5         used as anxiomimetic agents are isoprenaline
min) to evaluate the effectiveness of anti-anxiety      (Frohlich et al., 1969) and piperoxane (Bunney,
drugs. Reserpine and amylobarbitone reduced             1981). Often, however, the emotion is a cold and
the psychological and autonomic effects of              unconvincing feeling. Where anxious patients
yohimbine whereas atropine decreased the                are involved or other subjects have had extensive
pressor response. Paradoxically, chlorproma-            experience of anxiogenic substances, secondary
zine enhanced the responses to yohimbine.               conditioning is believed to occur linking drug-
   In a direct appraisal of how much yohimbine-         induced peripheral changes to central feelings of
and adrenaline-induced states resemble clinical         anxiety (Tyrer, 1976).
anxiety, Garfield and his colleagues (1967)
presented a battery of behavioural tests to 12
psychiatric patients before and after infusing Lactate infusions
these compounds and saline, two administrations
of each. Both yohimbine (0.5 mg kg-' over 6 Patients with anxiety states have a less efficient
min), and adrenaline (0.2 ,ug kg-' over 20 min exercise response than normal controls
256      M. Lader & M. Bruce
(McFarland & Huddleson, 1936; Jones &                 original findings of Pitts & McClure (1967) have
Mellersh, 1946; Holmgren & Strom, 1961). In a         been replicated several times, with EEG changes
standard exercise task, the patients show greater     (Fink et al., 1969) and alterations in forearm
rises in heart rate and blood lactate level, and      blood flow (Kelly et al., 1971), providing objec-
following the exercise they take up oxygen more       tive support. The data from these and other
than normals, suggesting the repayment of a           studies are quite consistent (Haslam, 1974;
larger oxygen debt.                                   Gorman et al., 1981; Bonn et al., 1971; Liebowitz
   From this, Pitts & McClure (1967) developed        et al., 1984). Over 80% of patients with panic
 the idea that perhaps the lactate ion itself could   disorder but less than 20% of normal subjects
 produce anxiety attacks in susceptible persons.      experience a panic attack when given lactate.
 To test this hypothesis, the following were intra-      Propranolol does not block all the effects
 venously infused double-blind in random order        (Arbab et al., 1971), but the lactate-induced
 into 14 patients with anxiety neurosis and 10        panic in patients is prevented by prior successful
 normal controls: 500 mmol sodium (±) lactate,        treatment with an antidepressant such as imipra-
 500 mmol sodium (±) lactate with 20 mmol             mine (Rifkin et al., 1981). Plasma adrenaline
 calcium chloride and 555 mmol glucose in 167         concentrations, but not plasma prolactin, tes-
 mmol sodium chloride. These solutions, of simi-      tosterone, noradrenaline or cortisol, are higher
 lar osmolarities, were given as 20 ml kg-1 body      among panickers than non-panickers (Fink et
 weight over 20 min to each subject.                  al., 1969).
 Symptoms were rated.                                    However, the mechanism of induction of
    Sodium lactate produced symptoms which            symptoms may be related to adrenaline release.
 'were markedly similar or identical' to those        In the accounts of lactate-induced symptoms,
 experienced in their 'worst attacks' by the          one is struck by the similarity of the symptoms
 anxious patients. Such reports were fewer from       produced to those following the infusion of
normal subjects. Symptoms were much less              adrenaline (Hawkins et al., 1960). Animal studies
frequent when lactate plus calcium chloride was       show that lactic acid releases adrenaline and
infused, and the glucose in saline infusion           noradrenaline from the adrenal medulla (Cannon
produced almost no symptoms in either patients        etal., 1924; Woods etal., 1956), and preliminary
or controls. The authors suggested that anxiety       reports suggest the same in humans for adrenaline
symptoms were related to hypocalcaemia pro-           (Appleby et al., 1981). Anxious patients may
duced by lactate infusion.                            have learned to associate such symptoms with
   Grosz & Farmer (1969) pointed out how              anxiety and these subjects might be more stressed
tenuous the link was between anxiety symptoms         by the non-specific aspects of the experimental
and hypocalcaemia. Anxiety can occur without          situation anyway. Also, the relationship between
high blood lactate concentrations and high blood      anxiety, lactate, glucose and adrenaline release
lactate levels without anxiety. An infusion of        would appear to be a fruitful area of research
sodium lactate produces a metabolic alkalosis         (Stanaway & Hullin, 1973).
whereas the endogenously produced lactate ion
shifts the acid-base balance of the body towards      Pentylenetetrazol
metabolic acidosis. Sodium bicarbonate levels
rise with sodium lactate infusion and the com-        This drug was first found to be anxiogenic in man
pensatory respiratory acidosis (adaptive hypo-        during research into its proconvulsant effects
ventilation) could be accompanied by feelings of      (Rodin, 1958; Rodin & Calhoun, 1970). How-
discomfort. Grosz & Farmer (1969) add that the        ever, no systematic work has been done in
rise in lactate produced by the infusion would        humans despite much animal work (Lal &
cause only a trivial change in the ionized calcium    Enmett-Oglesby, 1983).
concentration in the blood.
   In a further study, Grosz & Farmer (1972)
repeated Pitts & McClure's (1967) experiment          Carbon dioxide
but included a control infusion of sodium bicarb-
onate. Symptoms produced by the bicarbonate           Carbon dioxide was initially used as an anxiolytic
and the lactate infusions were the same, and          (Wolpe, 1958), but replication attempts more
were not associated with a rise in blood lactate      recently have not confirmed this. Indeed, it has
concentrations. Thus, the association of lactate      instead produced anxiomimetic effects (Van den
production and anxiety was not upheld by the          Hout & Griez, 1984; Griez, 1984). Further work
evidence.                                             implies that carbon dioxide-induced sensations
  Nevertheless the infusion of alkalinizing solu-     are labelled as either pleasant or unpleasant,
tions does seem to induce feelings of anxiety,        depending on prior expectations (Van den Hout
especially in patients with anxiety states. The       & Griez, 1982). Thus, the increased respiration
States of anxiety and their induction by drugs            257
rate and rise in catecholamines (Lambertsen,       all blocked by prior administration of the benzo-
1965), cause peripheral symptoms which in          diazepine antagonist, Ro 15-1788. More graded
patients with panic disorder are associated with   responses were obtained with lower doses (25-
their panic thereby creating a positive feedback   500 ,ug kg-') (Insel et al., 1984), with behavioural
loop. There is no evidence, as yet, of central     effects similar to those seen in monkeys facing a
effects.                                           threat such as the approach of an unfamiliar
                                                   investigator.
                                                      One study in man used N-methyl-,B-3-carbox-
,-carbolines                                       amide (FG 7142) (Dorow et al., 1983). Five
                                                   healthy male volunteers received increasing oral
The development of adrenaline and yohimbine-       doses, 100-200 mg and in one case 400 mg. On
induced models of anxiety stemmed from the         two occasions (out of 12) severe anxiety was
well-known sympathomimetic concomitants of         observed and in both instances the plasma con-
anxiety such as tachycardia and tremor. Another    centration of FG 7142 was over 150 ng ml-'. One
starting point has been the benzodiazepines and    volunteer reported severe anxiety with intense
the discovery of specific high-affinity receptors  inner strain and excitation with inability to speak.
in the brain for these and related drugs (Petersen Flushes of the face and the extremities were
et al., 1986). In the search for endogenous sub-   accompanied by a feeling of warmth. Blood
stances binding to these receptors with physio-    pressure and pulse rate rose. Agitation increased
logical effects, a class of 13-carboline compounds and the subject paced around the ward with the
was studied. These substances are related to       feeling of precordial pressure and palpitations.
plant alkaloids and derivatives such as harmine    Muscle tension and reflexes increased. These
and harmol with well-studied psychotropic prop-    effects peaked 1 h after administration and lasted
erties (Ho, 1977). Braestrup and his coworkers     for 2 h. The second volunteer was given 400 mg
isolated and identified the methyl and ethyl       of FG 7142 after taking doses of up to 200 mg
esters of ,B-carboline-3-carboxylic acid from      without incident. After 10 min, he developed
human urine and showed that they bound to          facial flushes, tremor and cold sweat. Fifteen
benzodiazepine binding sites (Nielsen et al.,      minutes later a stronger attack started. He felt a
1979; Braestrup et al., 1981). Although these      sense of impending doom. Yet a third attack
particular compounds were artefacts of extraction  started which was terminated by an intravenous
(Squires, 1981), the 3-carbolines in general are   benzodiazepine.
proving to be useful tools in our elucidation of      On the basis of two such fragmentary accounts
neurochemical mechanisms involved in anxiety.      it is not clear whether a true anxiety syndrome
   In animals, ,-carbolines antagonize the anti-   was induced. A slow infusion with an attempt to
convulsant actions of the benzodiazepines          measure sustained anxiety levels would be more
(Tenen & Hirsch, 1980; Cowen et al., 1981).        convincing. The waves of anxiety experienced
They also reduce the dose of a convulsant such as  by the second subject are not typical of panic
pentylenetetrazol needed to induce convulsions     attacks which, although they can be repeated,
without producing convulsions themselves, a        tend not to be in waves. They are more reminis-
proconvulsant effect (Oakley & Jones, 1980).       cent of hallucinogens such as LSD. The relation-
The sedative/antianxiety properties (as inferred   ship between ,B-carbolines and harmaline alka-
from animal models) of benzodiazepines are         loids cannot be overlooked. Nevertheless, the
antagonized by substituted ,B-carbolines (Cowen    interactions, biochemical and clinical, with the
et al., 1981; Petersen et al., 1982; Mendelson et  benzodiazepines suggest that this model of anxiety
al., 1983). ,-carbolines also possess intrinsic    is useful.
actions opposite in nature to the benzodia-
zepines. For example, they significantly increase
sleep latency (Mendelson et al., 1983). These Caffeine
compounds have been termed 'inverse agonists'
or 'contragonists'.                                Caffeine, a methylxanthine, is so well-known
   In primates, several P-carbolines induce syn- and widely used that its anxiogenic properties
dromes resembling anxiety. Following a large are often overlooked. It has been studied in the
dose of drug (2.5 mg kg-1 of ,B-carboline-3- laboratory since the last century (Hoch &
carboxylic acid ethyl ester, 13-CCE), Rhesus Kraepelin, 1895; Mosso, 1893), and a monu-
monkeys rapidly became agitated and struggled mental study was published by Hollingsworth
in the restraining chair, vocalization increased, (1912). Case reports indicate that caffeine either
they urinated and defecated (Ninan et al., 1982). in high dose or in withdrawal may produce symp-
Heart rate, blood pressure, plasma cortisol and toms of anxiety (Greden, 1974; MacCallum,
catecholamines were elevated. These effects were 1979), such as insomnia, tremor (Wharrad et al.,
258       M. Lader & M. Bruce
 1985), irritability, palpitations, nausea and        taining beverages, especially late at night. This
 diarrhoea.                                           area of research is probably the most useful at the
    One controlled double-blind study yielded         moment involving, as it does, a widely used drug.
 complex results (Goldstein et al., 1965).
 Caffeine, 300 mg, made subjects feel more alert      Conclusions
 mentally, more active physically, and more
nervous than did placebo. However, reports of         Guttmacher et al. (1983) set out the require-
increased arousal and of nervousness tended to        ments for an ideal model of anxiety:
occur in different individuals. The lower dose        1. It should mimic naturally occurring anxiety
(150 mg) gave similar but non-significant results.    but pharmacologic models engender 'as if' or
Another study showed quite consistently that          'cold' anxiety.
caffeine citrate in doses of 300 and 600 mg           2. It should therefore produce not only the
increased ratings of alertness, elation and feeling   physical signs of motor tension and autonomic
of quickness (Lader, 1969}. A recently completed      hyperactivity but also the dysphoria, worry,
study (Bruce et al., 1986) on nine normal subjects    fear, hypervigilance, irritability, and dread of
given placebo, 250 and 500 mg of anhydrous            which anxious patients complain.
caffeine, in a double-blind balanced design,          3. It should be replicable.
showed that caffeine significantly increased skin     4. The phenomena should be either short lived
conductance, mood ratings of alertness and            or readily reversible.
bodily symptom ratings of severe shaking and          5. It should differentiate between normal and
trembling. These results indicate increased           those with pathology, e.g. reflect the potential of
arousal in normal subjects, but not the induction     a trait response.
of clinical anxiety, at the doses of caffeine used.   6. It should reflect the potential for a state
A previous study produced similar findings            response. Those who have been successfully
(Charney et al., 1984b), but found that MHPG          treated clinically should not respond or respond
levels were not altered in a consistent fashion by    far less than those who have had no treatment.
caffeine and that there was no correlation with          As argued above, catecholamines, lactate
anxiety rating scales.                                infusions and carbon dioxide, fail the second
   A preliminary survey comparing patients with       requirement of the model. They have no specific
panic disorder and normal controls found that         central manifestations, but rather induce periph-
the patients had increased sensitivity to caffeine    eral changes which, in anxiety disorder subjects,
(Boulenger & Uhde, 1982). Other evidence              sets up a cognitive positive feed back loop leading
suggests that patients with anxiety disorders have    to the development of the central manifestations
increased caffeine sensitivity which leads to de-     of anxiety. Caffeine has yet to be tested so as to
creased consumption (Lee et al., 1985). At high       satisfy item 6, but otherwise fits the model. The
dose (720 mg orally), panic attacks were induced      a2-adrenoceptor antagonist, yohimbine, meets
in two normal volunteers (Uhde et al., 1983,) a       all the criteria of the model and is the best
finding of great relevance to the controversy as      current pharmac logical model of anxiety. The
to whether panic attacks and generalized anxiety      1-carbolines havJ not been adequately tested in
are separate syndromes or whether panic attacks       man although anipal data have much promise.
are a severe variant of anxiety.                         Of the anxiety disorders defined in DSM III,
   Caffeine and its related compounds interact        drug induction of panic disorder and generalized
weakly with the benzodiazepine receptor site          anxiety disorder has been achieved. Clinical
(Boulenger et al., 1982). They also stimulate         dispute concerns categorization of panic disorder
central noradrenergic activity (Berkowitz et al.,     as a separate condition from generalized anxiety
1970). In caffeine-naive subjects they produce a      disorder or as a more severe form. However,
dose related increase in plasma noradrenaline         inferences can be made from the pharmacological
(Robertson et al., 1978), although not in habitual    induction of anxiety. The drugs which satisfy
caffeine users (Robertson et al., 1984). How-         Guttmacher's model for anxiety, that is yohim-
ever, the most potent central effect is at the        bine, caffeine and 3-carbolines, suggest that
adenosine receptor site (Boulenger et al., 1982).     both generalized anxiety disorder and panic dis-
   To date caffeine provides a fairly convincing      order can be induced in a dose related manner.
model of generalized anxiety. Literally thous-        The implication therefore is that panic disorder
ands of people are aware that caffeine produces       is a more severe form of generalized anxiety
anxiety and insomnia and avoid caffeine con-          disorder.
States of anxiety and their induction by drugs                  259
References
                                                                 (1984a). Noradrenergic function in panic anxiety,
American Psychiatric Association (1980). Diagnostic              effects of yohimbine in healthy subjects and
    and Statistical Manual of Mental Disorders. 3rd ed.          patients with agoraphobia and panic disorder.
    Washington, D. C.                                           Arch. gen. Psychiat., 41, 751-763.
Appleby, I. L., Klein, D. F., Sachar, E. J. et al.           Charney, D. S., Galloway, M. P. & Heninger, G. R.
    (1981). Biochemical indices of lactate-induced               (1984b). The effects of caffeine on plasma MHPG,
    panic: A preliminary report. In Anxiety: New                subjective anxiety, autonomic symptoms and
    research and changing concepts, eds Klein, D. F. &          blood pressure in healthy humans. Life Sci., 35,
    Rabkin, J. pp. 411-423. New York: Raven Press.               135-144.
Arbab, A. G., Bonn, J. A. & Hicks, D. C. (1971).             Cowen, P., Green, A., Nutt; D. & Martin, I. (1981).
    Effect of propranolol on lactate induced phenom-             Ethyl-beta-carboline-carboxylate lowers seizure
    ena in normal subjects. Br. J. Pharmac., 41, 430.            threshold and antagonizes flurazepam-induced
Basowitz, H., Korchin, S. J., Oken, D.,                          sedation in rats. Nature, 290, 54-55.
    Goldstein, M. S. & Gussack, H. (1956). Anxiety and       Crawley, J. N., Skolnick, P. & Paul, S. M. (1984).
    performance changes with a minimal dose of epine-            Absence of intrinsic antagonist actions of benzo-
    phrine. AMA Arch. Neurol. Psychiat., 76, 98- 105.            diazepine antagonists on an exploratory model of
Berkowitz, B. A., Travers, J. H. & Spector, S. (1970).           anxiety in the mouse. Neuropharmacology, 23,
    Release of norepinephrine in the CNS by theo-                531-537.
    phylline and caffeine. Eur. J. Pharmac., 10, 64-71.      Dorow, R., Horowski, R., Paschelke, G., Amin, M.
Bonn, J. A. Harrison, J. & Rees, W. L. (1971).                   & Braestrup, C. (1983). Severe anxiety induced by
    Lactate induced anxiety; therapeutic application.           FG 7142, a beta-carboline ligand for benzodi-
    Br. J. Psychiat., 119, 468-471.                              azepine receptors. Lancet, H, 98-99.
Boulenger, J. P. & Uhde, T. W. (1982). Caffeine              Erdmann, G. & Janke, W. (1978). Interaction
    consumption and anxiety: preliminary results of a           between physiological and cognitive determinants
    survey comparing patients with anxiety disorders            of emotion: Experimental studies of Schachter's
    and normal controls. Psychopharmac. Bull., 18,              theory of emotion. Biol. Psychol., 6, 61-74.
    53-57.                                                   Erdmann, G. & van Lindern, B. (1980). The effects of
Boulenger, J. P., Patel, J. & Marangos, P. J. (1982).           beta-adrenergic stimulation and beta-adrenergic
    Effects of caffeine and theophylline on adenosine           blockade on emotional reactions. Psychophysi-
    and benzodiazepine receptors in human brain.                ology, 17, 332-338.
    Neurosci. Lett., 30, 161-66.                             Fast, G. J. & Fisher, S. (1971). The role of body
Braestrup, C., Nielsen, M., Skovbjerg, H. et al.                 attitudes and acquiescence in epinephrine and
    (1981). Beta-carboline-3-carboxylates and benzo-             placebo effects. Psychosom. Med., 33, 63-84.
    diazepine receptors. In GABA and benzodiaze-             Fink, M., Taylor, M. A. & Volavka, J. (1969).
   pine receptors, Costa, E. et al. New York: Raven              Anxiety precipitated by lactate. New Engl. J.
    Press.                                                       Med., 281, 1429.
Brown, J. T., Mulrow, C. D. & Stoudemire, G. A.              Frankenhaeuser, M., Jarpe, G. & Matell, G. (1961).
    (1984). The anxiety disorders. Ann. int. Med., 100,          Effects of intravenous infusions of adrenaline and
    558-564.                                                     noradrenaline on certain psychological and physio-
Bruce, M., Scott, N., Lader, M. & Marks, V. (1986).              logical functions. Acta Physiol. Scand., 51, 175-
    The psychopharmacological and electrophysio-                 186.
    logical effects of single doses of caffeine in healthy   Frankenhaeuser, M. & Jarpe, G. (1962). Psycho-
    human subjects. Br. J. clin. Pharmac., 22, 81-87.            physiological reactions to infusions of a mixture of
Bunney, W. E. (1981). Current biological strategies              adrenaline and noradrenaline. Scand. J. Psychol.,
    for anxiety. Psychiatric Annals, 11, 11-15.                  3, 21-29.
Cannon, W. B., Linton, J. R. & Linton, R. R. (1924).         Frankenhaeuser, M. & Jarpe, G. (1963). Psycho-
    Conditions of activity in endocrine glands. IV.              physiological changes during infusions of adrenaline
    The effects of muscle metabolites on adrenal                 in various doses. Psychopharmacologia, 4, 424-
    secretion. Am. J. Physiol., 71, 153-162.                     432.
Cantril, H. & Hunt, W. A. (1932). Emotional effects          Freud, S. (1894). The justification for detaching from
    produced by the injection of adrenaline. Am. J.              neurasthenia a particular syndrome: The anxiety
    Psychol., 44, 300-307.                                       neurosis. Collected works, 1, 76-109. London:
Carr, D. B. & Sheehan, D. V. (1984). Panic anxiety: A            Hogarth Press and Institute of Psychoanalysis.
    new biological model. J. clin. Psychiat., 45, 323-       Frohlich, E. D., Tarazi, R. C. & Dustan, H. P. (1969).
    330.                                                         Hyperdynamic beta-adrenergic circulatory state.
Charney, D. S., Heninger, G. R. & Sternberg, D. E.               Increased beta-receptor responsiveness. Arch.
    (1982). Assessment of the alpha-2 adrenergic auto-           intern. Med., 123, 1-7.
    receptor function in human: effects of oral              Garfield, S. L., Gershon, S., Sletten, I. et al. (1967).
    yohimbine. Life Sci., 30, 2033-2041.                         Chemically induced anxiety. Int. J. Neuro-
Charney, D. S., Heninger, G. R. & Redmond, D. E.                psychiatry, 3, 426-433.
    (1983). Yohimbine induced anxiety and increased          Goldstein, A., Karzer, S. & Warren, R. (1965).
    noradrenergic function in humans: effects of                 Psychotropic effects of caffeine in man. II Alert-
    diazepam and clonidine. Life Sci., 33, 19-26.                ness, psychomotor coordination and mood. J.
Charney, D. S., Heninger, G. R. & Breier, A.                     Pharmac. exp. Ther., 150, 146-151.
260       M. Lader & M. Bruce
Gorman, X M., Fyer, A. F., Gliklick, J., King, D. &      Kopin, I. J. (1984). Avenues of investigation for the
    Klein, D. F. (1981). Effect of sodium lactate on         role of catecholamines in anxiety. Psycho-
    patients with panic disorder and mitral valve pro-       pathology, 17, Suppl 1, 83-97.
    lapse. Am. J. Psychiatry, 138, 247-249               Lader, M. (1969). Comparison of amphetamine
Greden, J. F. (1974). Anxiety or caffeinism: A               sulphate and caffeine citrate in man. Psycho-
    diagnostic dilemma. Am. J. Psychiatry, 131, 1089-        pharmacologia, 14, 83-94.
    1092                                                 Lal, H. & Emmett-Oglesby, M. W. (1983) Be-
Griez, E. (1984). Experimental models of anxiety.            havioural analogues of anxiety: Animal models.
    Problems and perspectives. Acta Psychiatr. Belg.,        Neuropharmacology, 22, 1423-1441.
    84, 511-532.                                         Lambertsen, C. J. (1965). Theraputic gases, oxygen,
Grosz, H. J. & Farmer, B. B. (1969). Blood lactate in        carbon dioxide, helium. In Drill's pharmacology in
    the development of anxiety symptoms. A critical          medicine, ed. Dipalma, J. R. New York: McGraw
    examination of Pitts and McClure's hypothesis and        Hull.
    experimental study. Arch. gen. Psychiatry, 21,       Lee, M. A., Cameron, 0. G. & Greden, J. F. (1985).
    611-619.                                                 Anxiety and caffeine consumption in people with
Grosz, H. J. & Farmer, B. B. (1972). Pitts and               anxiety disorders. Psychiatric Research, 15, 211-
    McClure's lactate-anxiety study revisited. Br. J.        217.
    Psychiatry, 120, 115-118.                            Liebowitz, M. R., Fyer, A. F., Gorman, J. M., Dillon,
Guttmacher, L. B., Murphy, D. L. & Insel, T. R.              D. & Appleby, G. (1984). Lactate provocative of
    (1983). Pharmacologic models of anxiety. Comp.           panic attacks. I. Clinical and behavioral findings.
    Psychiatry, 24, 312-326.                                 Arch. gen. Psychiatry, 41, 764-770
Haslam, M. T. (1974). The relationship between the       Marafion, G. (1924). Contribution a l'etude de l'action
    effect of lactate infusion on anxiety states, and        emotive de l'adrenaline. Rev. Fr. Endocrinol., 2,
    their amelioration by carbon dioxide inhalation.         301-325.
    Br. J. Psychiatry, 125, 88-90.                       MacCallum, W. A. G. (1979). Excess coffee and
Hawkins, D. R., Monroe, J. D., Sandifer, M. G. &             anxiety states. Int. J. Soc. Psychiatry, 25, 209-210.
    Vernon, C. R. (1960). Psychological and physio-      McFarland, R. A. & Huddleson, J. H. (1936). Neuro-
    logical responses to continuous epinephrine              circulatory reactions in psychoneuroses: studies by
    infusion. Psychiatric Research Reports of the            the Schneider method. Am. J. Psychiatry, 93, 567-
    American Psychiatric Association, 12, 40-52.             599.
Ho, B. (1977). In Current developments in psycho-        Mendelson, W. B., Cain, M., Cook, J. M., Paul, S. M.
    pharmacology, 4, 153-175, ed. Essman, W. New             & Skolnick, P. (1983). A benzodiazepine receptor
    York: Spectrum.                                          antagonist decreases sleep and reverses the hypnotic
Hoch, A. & Kraepelin, E. (1895). Uber die Wirkung            actions of flurazepam. Science, 219, 414-416,
    der Theebestandtheile auf korperliche und geistige   Mosso, U. (1893). Action des principes actifs de la
    Arbeit. Psychologische Arbeiten, 1, 379-488.             noix de kola sur la contraction musculaire. Arch.
Hollingsworth, H. L. (1912). The influence of caffeine       Ital. Biol., 19, 241-256.
    on mental and motor efficiency. Arch. Psychol.,      Nesse, R. M., Cameron, 0. G., Curtis, G. C.,
    22, 1-166.                                               McCann, D. S. & Huber-Smith, M. J. (1984).
Holmberg, G. & Gershon, S. (1961). Autonomic and             Adrenergic function in patients with panic anxiety.
    psychic effects of yohimbine hydrochloride.              Arch. gen. Psychiatry, 41, 771-776.
    Psychopharmacologia, 2, 93-106.                      Nielsen, M., Gredal, 0. & Braestrup, C. (1979). Some
Holmgren, A. & Strom, G. (1959). Blood lactate               properties of 3-H-diazepam displacing activity with
    concentration in relation to absolute and relative       human urine. Life Sci., 25, 679-86.
    work load in normal men, and in mitral stenosis,     Ninan, P. T., Insel, T. R., Cohen, R. M., Cook, J. M.,
    atrial septal defect and vasoregulatory asthenia.        Skolnick, P. & Paul, S. M. (1982). Benzodiazepine
    Acta med. Scand., 163, 185-193.                          receptor-mediated experimental 'anxiety' in
Ingram, C. G. (1962). Some pharmacologic actions of          primates. Science, 218, 1332-1334.
    yohimbine and chlorpromazine in man. Clin.           Oakley, N. R. & Jones, B. J. (1980). The procon-
    Pharmac. Ther., 3, 345-352.                              vulsant and diazepam reversing effects of ethyl-
Insel, T. R., Ninan, P. T., Aloi, J., Jimerson, D. C.,       beta-carboline-3-carboxylate. Eur. J. Pharmac.,
    Skolnick, P. & Paul, S. M. (1984). A benzodiaze-         68, 381-382.
    pine receptor-mediated model of anxiety. Studies     Petersen, E. N., Pschelke, G., Kehr, W., Neilsen, M.
    in nonhuman primates and clinical implications.          & Braestrup, C. (1982). Does the reversal of the
    Arch. gen. Psychiatry, 41, 741-750.                      anticonffict effect of phenobarbital by beta-CCE
Jones, M. & Mellersh, V. (1946). A comparison of the         and FG 7142 indicate benzodiazepine receptor
     exercise response in anxiety states and normal          mediated anxiogenic properties? Eur. J. Pharmac.
     controls. Psychosom. Med., 8, 180-187.                  82, 217-221.
Kandel, E. R. (1983). From metapsychology to mole-       Petersen, E. N., Jensen, L. H., Drejer, J. & Honore,
     cular biography: Explorations into the nature of        T. (1986). New perspectives in benzodiazepine
     anxiety. Am. J. Psychiatry, 140, 1277-1293.             receptor pharmacology. Pharmacopsychiat., 19,
Kelly, D. H. W., Mitchell-Heggs, N. & Sherman, J.            4-6.
     (1971). Anxiety and the effects of sodium lactate   Pitts, F. N. & McClure, J. N. (1967). Lactate meta-
     assessed clinically and physiologically. Br. J.         bolism in anxiety neurosis. New Engl. J. Med., 277,
     Psychiatry, 119, 129-141.                               1329-1336.
States of anxiety and their induction by drugs                    261
Rifkin, A., Klein, D. F., Dillon, D & Levitt, M.              Med., 3, 343-349.
    (1981). Blockade by imipramine or desipramine of       Tenen, S. S. & Hirsch, J. D. (1980). Beta-carboline-3-
    panic induced by sodium lactate. Am. J. Psychiatry,       carboxylic acid ethyl ester antagonizes diazepam
    138, 676-677.                                             activity. Nature, 288, 609-610.
Robertson, D., Frolich, J. C., Carr, R. K., Watson,        Tyrer, P. (1976). The role of bodily feelings in anxiety.
    J. T., Hollifield, J. W., Shand, D. G. &                  London: Oxford University Press.
    Oates, J. A. (1978). Effects of caffeine on plasma     Tyrer, P. (1984). Classification of anxiety. Br. J.
    renin activity, catecholamines and blood pressure.        Psychiatry, 144, 78-83.
    New Engl. J. Med., 298, 181-186.                       Uhde, T. W., Boulenger, J. P., Vittone, B., Siever,
Robertson, D., Hollister, A. S., Kincaid, D.,                 L. J. & Post, R. M. (1983). Human anxiety and
    Workman, R., Goldberg, M. R., Tung, C. &                  noradrenergic function: Preliminary studies with
    Smith, B. (1984). Caffeine and hypertension. Am.          caffeine, clonidine and yohimbine. In Proceedings
    J. Med., 77, 54-60.                                       of the 7th world congress ofpsychiatry, Vienna, 2,
Rodin, E. A. (1958). Metrazol tolerance in a 'normal'         693-698. New York: Plenum Press.
    and volunteer population. Electroencephalogr.          Van den Hout, M. A. & Griez, E. (1982). Cognitive
    clin. Neurophysiol., 10, 433-446.                         factors in carbon dioxide therapy. J. psychosom.
 Rodin, E. A. & Calhoun, (1970). Metrazol in a                Res., 26, 209-214.
     'normal' volunteer population. J. nerv. ment.         Van den Hout, M. A. & Griez, E. (1984). Panic
     Dis., 150, 438-450.                                      symptoms after inhalation of carbon dioxide. Br. J.
 Schachter, S. & Singer, J. E. (1962). Cognitive, social      Psychiatry, 144, 503-307.
     and physiological determinants of emotional state.    Wearn, J. T. & Sturgis, C. C. (1919). Studies on
     Psychol. Rev., 69, 379-399.                              epinephrine. Effects of the injection of epinephrine
 Schachter, S. & Wheeler, L. (1962). Epinephrine,             in soldiers with irritable heart. Arch. intern. Med.,
    chlorpromazine, and amusement. J. abnorm. soc.            24, 247-268.
    Psychol., 65, 121-128.                                 Wharrad, H. J., Birmingham, A. T., MacDonald,
 Schachter, S. (1966). The interaction of cognitive and       I. A., Inch, P. J. & Mead, J. L. (1985). The
    physiological determinants of emotional states. In        influence of fasting and of caffeine intake on finger
    Anxiety and Behavior, ed. Spielberger, C. D. New          tremor. Eur. J. clin. Pharmac., 29, 37-43.
    York: Academic Press.                                  Wolpe, J. (1958). Psychotherapy by reciprocal
Squires, R. F. (1981). GABA receptors regulate the            inhibition. Stanford, California: University Press.
    affinities of anions required for brain specific       Woods, E. F., Richardson, J. A., Richardson, A. K.
    benzodiazepine binding. Adv. biochem. Psycho-             & Bozeman, R. F. (1956). Plasma concentrations
    pharmac., 26, 129-138.                                    of epinephrine and arterenol following the action
Stanaway, R. G. & Hullin, R. P. (1973). The relation-         of various agents on the adrenals. J. Pharmac. exp.
    ship of exercise response to personality. Psychol.         Ther., 116, 351-355.
                                                                                   (Received 27 January 1986,
                                                                                       accepted 10 May 1986)
You can also read