DEPRESSION AND ANXIETY 25:847–853 (2008)

Research Article
A CAFFEINE CHALLENGE TEST IN PANIC DISORDER
PATIENTS, THEIR HEALTHY FIRST-DEGREE RELATIVES
AND HEALTHY CONTROLS
Antonio E. Nardi, M.D.,1 Alexandre M. Valenc- a, M.D.,1 Isabella Nascimento, M.D.,1 Rafael C. Freire, M.D.,1
André B. Veras, M.D.,1 Valfrido L. de-Melo-Neto, M.D.,1 Fabiana L. Lopes, M.D.,1 Anna Lucia King, R.A.,1
Gastão L. Soares-Filho, M.D.,1 Marco A. Mezzasalma, M.D.,1 Arabella Rassi, M.D.,1 and Walter A. Zin, M.D.2

Our aim was to observe the induction of anxiety symptoms and panic attacks by a
caffeine challenge test in panic disorder (PD) patients (DSM-IV) and their healthy
first-degree relatives. We randomly selected 25 PD patients, 27 healthy first-degree
relatives of probands with PD, and 22 healthy volunteers with no family history of
PD. In a randomized double-blind experiment performed over two occasions 7 days
apart, 480 mg caffeine and a caffeine-free solution were administered in a coffee
form. Using specific panic attack criteria, 52.0% (n 5 13) PD patients, 40.7%
(n 5 11) first-degree relatives (v2 5 1.81, df 5 1, P 5 0.179), and none of the
control subjects had a panic attack after the test (v2 5 51.7, df 5 2, Po0.001). In
this caffeine challenge test, PD patients and their first-degree relatives were more
sensitive than healthy volunteers to the panic attack symptoms but less sensitive to
headache, increase in blood pressure, and insomnia. Our data suggest that there is
an association between panic attacks after the intake of 480 mg of caffeine in PD
patients and their first-degree relatives. There is a clear differentiation of PD
patients and their first-degree relatives by a caffeine test from the healthy group.
Depression and Anxiety 25:847–853, 2008. r 2007 Wiley-Liss, Inc.

Key words: panic attacks; coffee; anxiety disorder; comorbidity; stimulant drug

mobilization of intracellular calcium, phosphodiester-

INTRODUCTION
A high prevalence of anxiety disorders and some
biological sensitivity have been found among first-
degree relatives of subjects with panic attacks [Perna
et al., 1995]. Caffeine is a xanthine derivative that is
used worldwide as a psychostimulant and it may be a
useful biological model of induced panic attack [Lee
et al., 1988; Uhde, 1990, 1995]. There are well-
recognized anxiogenic potentials of caffeine, and
clinical researchers have been trying to establish
its capacity for discriminating panic disorder (PD)
patients [Uhde, 1995].
The highly complex mechanisms of caffeine action
have been the subject of investigation. The precise
mechanism underlying the caffeine panicogenic poten-
tials remains unknown, although the most likely
mechanism of caffeine’s stimulant effects is considered
to be the antagonism of central adenosine receptors
[Boulenger et al., 1984; Charney et al., 1985]. It
appears to act via multifaceted mechanisms including
ase inhibition (leading to an increase in intracellular
1
Laboratory of Panic & Respiration, Institute of Psychiatry,
Rio de Janeiro, Brazil
2
Laboratory of Respiration Physiology, Carlos Chagas Filho
Biophysics Institute, Federal University of Rio de Janeiro, Rio
de Janeiro, Brazil
Contract grant sponsor: Brazilian Council for Scientific and
Technological Development (CNPq); Contract grant number:
554411/2005-9.
Correspondence to: Antonio E. Nardi, M.D., Laboratory of Panic
& Respiration, Institute of Psychiatry, Federal University of Rio de
Janeiro, R. Visconde de Pirajá, 407/702, Rio de Janeiro-RJ-
22410–003 Brazil. E-mail: antonionardi@terra.com.br
Received for publication 18 October 2006; Revised 8 May 2007;
Accepted 23 May 2007
DOI 10.1002/da.20354
Published online 6 September 2007 in Wiley InterScience (www.
interscience.wiley.com).

r 2007 Wiley-Liss, Inc.

848
cAMP), antagonism of adenosine receptors, and
benzodiazepine receptor antagonism [Nehlig et al.,
1992]. In addition, caffeine increases plasma levels of
the tryptophan metabolite kynurenine, a neuroactive
metabolite of tryptophan in health volunteers [Orlikov
and Ryzov, 1991], suggesting involvement of 5-hydro-
xytryptamine systems in caffeine action. Imaging
studies of cerebral blood flow using positron emission
tomography show significant decreases with caffeine
infusion in healthy subjects as well as PD patients, but
also associated increases in glucose utilization [Camer-
on et al., 1990]. These studies suggest that decreased
cerebral blood flow is a relatively nonspecific effect, not
solely responsible for caffeine-induced panic.
The pharmacological challenge test in PD involves
giving a provoking agent under controlled rules to
clarify some aspect of behavioural or biological
function [Uhde, 1995]. In addition to caffeine, various
agents such as sodium lactate, carbon dioxide, yohim-
bine, isoprenaline, and cholecystokinin have been used
as provoking agents in healthy volunteers as well as in
PD patients [Gorman et al., 2000]. These pharmaco-
logical models of PD can provide researchers with
insight into the endogenous neurocircuity instrumental
to the expression of panic attacks and stimulate the
development of pharmacological interventions.
Our aim was to observe if PD patients—DSM-IV
[American Psychiatric Association, 1994]—and their
first-degree relatives respond in a similar way to the
induction of panic attacks by a caffeine challenge test.
We also decided to discriminate PD patients from their
first-degree relatives as they might have some physio-
logical differences in relation to the anxiety response
level to the caffeine challenge test. We expected that
the PD patients would have a greater sensitivity to this
challenge test than the other two groups.
EXPERIMENTAL PROCEDURES
From an initial group of 93 PD patients, we
randomly selected 25 PD patients, 17 women and 8
men; mean age7SD, 33.9713.0 years; and 27 of their
first-degree relatives, 20 women and 7 men; mean
age7SD: 37.2711.5 years from the Laboratory of
Panic & Respiration of the Institute of Psychiatry,
Federal University of Rio de Janeiro. All the subjects
who participated in this study were not recruited for
any other study. A healthy group of 22 subjects with no
first-degree relative history of moderate or severe
anxiety or mood disorder (16 women and 6 men, mean
age7SD: 31.7712.9 years) was used for comparison.
The PD group first received a clinical diagnosis made
by a study psychiatrist and was then interviewed by a
second clinician with the Structured Clinical Interview
Diagnostic(SCID) [First et al., 1997] for DSM-IV. If
the two clinicians disagreed on the diagnosis, they met
to confer. If a consensus on the diagnosis could not be
reached, the subject was not enrolled. Patients who met
DSM-IV criteria for current mood disorder, obsessive-
Depression and Anxiety
Nardi et al.
compulsive disorder, psychotic disorders, organic brain
syndrome, severe personality disorder, epilepsy, or
substance abuse or dependence (during the previous
year) were excluded. Patients with comorbid dysthy-
mia, generalized anxiety disorder, or past major
depression were included, if PD was judged to be the
principal diagnosis. The protocol was explained to the
subjects, who signed a voluntary written consent to
participate in the study. The subjects were informed
that they would be asked to drink different coffee
solutions; the procedure was not dangerous but the
anxiety symptoms could occur during the session. Our
local Ethics Committee approved the protocol. The
inclusion criteria were 18–55 years of age, occurrence
of at least three panic attacks in the 2 weeks before the
first challenge test day in PD patients, no use of any
psychotropic drugs for at least 4 weeks and fluoxetine
for 5 weeks by any participant, and a negative urine test
for benzodiazepines and other medications before the
test. Exclusion criteria were any medical condition that
could influence the results, cognitive-behavior psy-
chotherapy during the study, or the presence of suicidal
risk. Smokers were also excluded.
The first-degree relatives and the comparison group
were also assessed by the SCID [First et al., 1997].
They were free of any history of PD, major mood
disorder, psychotic disorder, and current substance use
disorders. Most of the PD patients (n 5 25) brought at
least one first-degree relative (n 5 27) while a few of
them brought two relatives.
All subjects underwent a physical examination and
laboratory exams to ensure that they were healthy
enough to participate in the challenge test. They had
no respiratory or cardiovascular abnormalities and
were free of caffeine ingestion for 1 week. All the
subjects received a list containing the main soft-drinks,
coffee and tea presentations, and over-the-shelf drugs,
which contain caffeine so that they avoid them during
this period. The compliance was obtained by their
retrospective report.
The test was conducted in the usual examination
room, with no changes made in the environment. All
subjects were asked to relax for 10 min. Then we
checked their respiratory frequency, pulse, and blood
pressure and repeated the measurements 15 and 30 min
after the challenge test. To measure the baseline anxiety
level, before drinking the coffee, the subjects were
asked to complete the Subjective Units of Disturbance
Scale (SUDS), a semiquantitative evaluation method
ranging from 0 5 no anxiety to 10 5 maximum anxiety
[Bech et al., 1986], and the Diagnostic Symptom
Questionnaire (DSQ) [Sanderson et al., 1989] adapted
for DSM-IV, in which the presence and level of
discomfort of panic symptoms experienced after
drinking the coffee solutions were rated on a 0–4 point
scale (0 5 none , 4 5 very severe). Both self-rating
scales, the SUDS and the DSQ had been tested by back
translation. The SUDS and the DSQ were completed
again 30 min after the coffee solution intake. The
 Research Article: A Caffeine Challenge Test in Panic Disorder Patients 849

subjects completed them independently of the raters. previous session (i.e. those given caffeine in the first
On the basis of the DSQ, a panic attack was defined by session received caffeine-free solution in the second
objective raters as the following: (1) four or more and vice versa).
symptoms of a panic attack from the DSM-IV; (2) at
least one of the cognitive symptoms of a panic attack STATISTICAL ANALYSIS
from the DSM-IV (e.g. fear of dying or of losing
control); (3) feeling of panic or fear, similar to Panic rates for the three groups were compared by
spontaneous panic attacks recorded on a card, which the w2 test. Data concerning the effects of caffeine and
the raters were not permitted to see; and (4) agreement time of observation were tested by two-way ANOVA
with clinical diagnosis evaluation between two diag- with repeated measures for time and independent
nosis blinded raters from the team during the test. The groups for SUDS (before and after) and heart rate
comparison of the two raters score was done after the (before and after). Pair-wise comparisons of the
test. The feeling of a panic attack reported by the treatment groups were performed at end-point, using
subjects was also examined in order to compare Fisher’s-protected least significant difference method.
agreement between raters and subjects. The results For identifying heart rate differences among groups, a
were divided into raters’ score (all the four items of the one-way ANOVA was used for time difference using
panic attack criteria) and subjects score (just the first Fisher’s protected least significant difference method.
three items of panic attack criteria). The level of significance was set at 5%.
All raters were research members of our laboratory and
trained for the instruments used in this trial. Before staring RESULTS
the trial and twice during it, we had theoretical and
practical sessions of training to standardize our recordings. Due to our inclusion and exclusion criteria, the
Caffeine is a very important substance in our Brazilian comorbidity observed in our PD sample was very low.
dietary pattern [Camargo et al., 1999]. We calculated Only six (24%) patients had previous major depressive
approximately the individual daily intakes (mg/kg) of episodes, five (20%) had diagnosis of generalized
caffeine from the consumption data generated by the anxiety disorder (GAD), and three (12%) had diagnosis
sample and the caffeine content of the most consumed of dysthymia. The mean age of onset of PD was
products. We did not include participants with a 27.578.1 years (7SD), the duration of the illness was
diagnosis of abuse or dependence on caffeine. 15.4711.0 years (7SD), and 13 (52%) PD patients
We explained to the subjects what they were expected had never been treated with psychotropic drugs before.
to do and they relaxed for an additional 10 min, after Our three groups (PD, first-degree relatives, and
which we gave them the coffee solution. In a healthy controls) were matched according to gender,
randomized double-blind experiment performed over age, and body mass index (Table 1). The consumption
two occasions 7 days apart, an oral dose of 480 mg of pattern of caffeine of all the subjects was 74.3%
caffeine (Cafés, Brazil, about 460 ml of solution), or a consumed soft drinks regularly, 67.6% coffee, 60.8%
caffeine-free solution (Cafés, caffeine-free, Brazil, also chocolate products, and 32.4% tea. The average and
about 460 ml) was administered in the form of instant median potential daily intakes of caffeine by the sample
coffee. The amount of caffeine in the solution was
calculated by our local biochemical laboratory. The
TABLE 1. Demographic characteristics of our three
decaffeinated coffee contained a small amount of groups: panic disorder patients, first-degree relatives,
caffeine as more than 97% of the caffeine was removed, and healthy controls
resulting in less than 5 mg of caffeine intake. The
source of caffeine used did not guarantee that the doses Male Female
given in the challenges were exact. Most caffeine
Gender-n (%)
products (e.g. caffeine citrate) were quite bitter. To deal
PD n 5 25 8 (32.0) 17 (68.0)
with the caffeine-free drink (placebo), we used FDR n 5 27 7 (25.9) 20 (74.1) w2 5 1.26 df 5 2
two tablets of a low-calorie sweetener sucralose P 5 0.863
(0.2 calories/tablet) in the caffeine and in the caffeine-free HC n 5 22 6 (27.3) 16 (72.7)
drinks. The patient was requested to drink the coffee Age (years7SD)
solution within a period of 15 min, after which a PD 33.9713.0
30-min period followed, so that the caffeine could FDR 37.2711.5
reach its peak level in the blood [Blanchard and Sawers, HC 31.7712.9 Mann–Whitney,
1983]. All the session procedures were in the morning P 5 0.762
and the subjects had least time (8 hr) of fasting before Body mass index (mean7SD)
PD 23.474.3
the experiment. They were randomly divided into two
FDR 24.273.6
groups: in the first session, subjects received the oral HC 23.673.8 Mann–Whitney,
dose of caffeine or caffeine-free solution, and in the P 5 0.539
second session the same procedure was performed
using the drug that had not been administered in the PD, panic disorder; FDR, first-degree relatives; HC, healthy controls.

Depression and Anxiety

850 Nardi et al.
were 2.7470.31 for PD, 2.9170.45 for first-degree
relatives, and 2.8070.52 mg/kg for the healthy group
(Mann–Whitney, P 5 0.618).
The panic rates assigned by the raters and by the
subjects after the challenge test are recorded in Table 2.
Significantly more PD and first-degree relatives had a
panic attack in response to caffeine than the healthy
group according to both rater groups. The panic attack
symptoms generally started after 30 to 40 min of the
coffee intake. The mean duration of the anxiety
symptoms after the caffeine intake did not differ
among the groups (mean7SD): 48.5720.2 min for
the PD patients, 50.3718.6 min for the first-degree
relatives, and 43.9719.4 min for the healthy group
(ANOVA F 5 0.531 df 5 2, 73 P 5 0.265). In the case of
all the first-degree relatives who had a panic attack,
their PD relative also had the same attack with a similar
response to the test (Table 2). No subject of any group
had a panic attack or a significant increase in anxiety/
symptoms after the consumption of the caffeine-free
solution.
The frequency of panic attack symptoms and other
anxiety symptoms after the 480 mg caffeine challenge
test in PD patients, their first-degree relatives, and
healthy subjects are recorded in Table 3. We can
observe that PD patients and their first-degree relatives
had a higher frequency of panic symptoms than the
healthy group with the exception of paresthesias,
depersonalization/derealization (PD and normal con-
trols higher than first-degree relatives), and sweating
(normal controls higher than the other two groups). Of
the other anxiety symptoms induced by caffeine,
headache, increased blood pressure, and insomnia, the
TABLE 2. Response of patients with panic disorder,
first-degree relatives and healthy controls after the
caffeine challenge test. Raters () and subjects ()
evaluation of panic attacks (w2 test).
PD FDR HC
After 480 mg caffeine intake n 5 25 (%) n 5 27 (%) n 5 22 (%)
Raters
Panic 13 (52.0) 11 (40.7) 0 (0.0)
No panic 12 (48.0) 16 (59.3) 22 (100.0)
Subject
Panic 12 (48.0) 12 (44.4) 0 (0.0)
No panic 13 (52.0) 15 (55.6) 22 (100.0)
PD, panic disorder; FDR, first-degree relatives; HC, healthy
controls.
w2 5 51.7, df 5 2, Po0.001.
PD versus relatives. w2 with Yates correction 5 1.81, P 5 0.179.
PD versus control. w2 with Yates correction 5 50.1, Po0.001.
Relatives versus control. Fisher exact test, 2-tailed Po0.001.
w2 5 50.7, df 5 2, Po0.001.
PD versus relatives. w2 with Yates correction 5 0.707, P 5 0.400.
PD versus control. w2 with Yates correction 5 50.6, Po0.001.
Relatives versus control. Fisher exact test, 2-tailed, Po0.001.
The group included all the same 12 PD patients who had a panic
attack identified by the raters.
Depression and Anxiety
healthy group had a higher frequency than the other
groups. No participant in any group reported olfactory
hallucinations after the caffeine intake as this is
described as intravenous caffeine in normal volunteers
[Nickell and Uhde, 1994/1995].
The SUDS level measurement before and 30 min
after the caffeine solution demonstrated that PD
patients (before 5 3.172.7; after 5 7.574.9) were not
more sensitive to the effect of caffeine than relatives
(before 5 2.872.6; after 5 7.074.3) or normal con-
trols (before 5 3.672.4; after 5 7.174.7). All subjects
showed a similar increase in anxiety levels after the
caffeine test (ANOVA; F 5 0.344, df 5 2, 73,
P 5 0.641).
All groups had a similar heart rate before (1 min
before) the caffeine intake (one-way ANOVA;
F 5 0.283, df 5 2, 73, P 5 0.716). There was no time
effect on heart rate and a time group interaction,
demonstrating no heart rate–time interaction between
groups (two-way ANOVA, group  time interaction
with Greenhouse–Geisser correction: F 5 0.460,
df 5 2, 146, P 5 0.591).
DISCUSSION
Our initial hypothesis was not confirmed as the PD
group responded in a way similar to their first-degree
relatives. Perhaps, these caffeine-sensitive PD patients
and some of their first-degree relatives will be clearly
identified in the future with a more sophisticated
methodology.
The strength of our data can be based on the
inclusion of a healthy group, the caffeine-free placebo
solution, the sample size, and the urinary tests to
preclude the possibility of drug reception prior to
experimentation. A cognitive vulnerability, i.e. a greater
tendency to misinterpret intense somatic sensations
[Chambless et al., 2000; Clark, 1986], or/and learned
responses acquired through interoceptive conditioning
procedures [Wolpe and Rowan, 1988], may have the
central underlying mechanism that led some PD
patients to increase their anxiety levels or even to have
a panic attack in response to the caffeine challenge.
The present study cannot dismantle the biological and
psychological mechanisms underlying the sensitivity
shown by a proportion of PD patients and their
relatives to the caffeine challenge. Our study did not
have a more sophisticated measurement of cognitive
factors associated with PD. The source of caffeine used
also did not guarantee that the doses given in the
challenges were exact.
In our trial the PD patients and their first-degree
relatives were less sensitive to headache, insomnia,
and increase in blood pressure than healthy controls.
Other studies [Charney et al., 1985; Uhde, 1990]
found an opposite pattern. Perhaps this difference
was due to our methodological features in the caffeine
solution, its intake procedure, and time to check
the changes.
 Research Article: A Caffeine Challenge Test in Panic Disorder Patients 851

TABLE 3. Frequency of panic attack symptoms and other anxiety symptoms after the 480 mg caffeine challenge test in
panic disorder patients, their first-degree relatives, and healthy controls

PD FDR HC
Symptoms n 5 25 (%) n 5 27 (%) n 5 22 (%) Analysis

Fear of dying 14 (56.0) 15 (55.6) 6 (27.3) w2 5 11.7, df 5 2, P 5 0.003

Chest pain/discomfort 12 (48.0) 11 (40.7) 4 (18.2) w2 5 22.0, df 5 2, Po0.001
Shortness of breath 11 (44.0) 13 (48.1) 4 (18.2) w2 5 21.0, df 5 2, Po0.001
Paresthesias 9 (36.0) 7 (25.9) 6 (27.3) w2 5 25.8, df 5 2, Po0.001
Feelings of choking 10 (40.0) 11 (40.7) 3 (11.1) w2 5 34.0, df 5 2, Po0.001
Dizziness/lightheadedness 11 (44.0) 12 (44.4) 7 (31.8) w2 5 5.05, df 5 2, P 5 0.080
Depersonalization/derealization 8 (32.0) 8 (29.6) 8 (36.4) w2 5 18.5, df 5 2, Po0.001
Losing control/going crazy 6 (24.0) 7 (25.9) 5 (22.7) w2 5 40.0, df 5 2, Po0.001
Chills/hot flushes 11 (44.0) 12 (44.4) 8 (36.4) w2 5 5.05, df 5 2, P 5 0.080
Nausea/abdominal distress 9 (36.0) 11 (40.7) 6 (27.3) w2 5 16.0, df 5 2, Po0.001
Palpitations 14 (56.0) 12 (44.4) 5 (22.7) w2 5 16.3, df 5 2, Po0.001
Sweating 6 (24.0) 5 (18.5) 7 (31.8) w2 5 40.0, df 5 2, Po0.001
Trembling/shaking 6 (24.0) 7 (25.9) 5 (22.7) w2 5 40.0, df 5 2, Po0.001
Other anxiety symptoms
Headache 7 (28.0) 9 (33.3) 8 (36.4) w2 5 19.0, df 5 2, Po0.001
Increase in systolic blood pressure (more than 5 mmHg) 8 (32.0) 10 (37.0) 11 (50.0) w2 5 9.86, df 5 2, P 5 0.007
Insomnia in the night after the test 8 (32.0) 9 (33.3) 9 (40.9) w2 5 13.7, df 5 2, P 5 0.001

PD, panic disorder; FDR, first-degree relatives; HC, healthy controls.

Due to our inclusion and exclusion criteria, the
comorbidity in our PD sample was very low. Only the
GAD could influence the response but even this
diagnosis was present in only 20% of the PD patients.
Bruce et al. [1992] measured the effects of anxiety
from two doses of oral caffeine (250 and 500 mg)
and placebo was compared in 12 patients with GAD,
12 patients with PD, and 12 healthy subjects. Caffeine
produced significantly less decrease in electroencepha-
lographic a wave activity, greater decrease in N1-P2
auditory evoked potential amplitude and greater
increase in skin conductance level, systolic and diastolic
blood pressure, critical fusion flicker frequency,
and self-ratings of anxiety and sweating in patients
with GAD than in normal patients. PD patients
showed different reactivity than normal patients did
with respect to electroencephalographic a waves,
N2 latency, N2-P2 auditory evoked potential ampli-
tude, and physical tiredness but were less reactive than
patients with GAD on several variables.
The precise and clinical criteria for induced panic
attack may be the crucial point for our results. In our
opinion, the precise criteria for laboratory panic attack
should include participants’ and raters’ evaluation. Our
definition of a laboratory-induced panic attack is based
on the DSQ [Sanderson et al., 1989] adapted for DSM-
IV. In addition, it requires confirmation by the patient
and two raters. The subjective level of anxiety after any
anxiety-induced test in PD patients is higher than
before the tests (Table 2) in the groups. The anxiety
level is not directly linked to the panic attack developed
by some patients. Our criteria for laboratory-induced
panic attack involve the presence and severity of
the symptoms with patients’ and raters’ evaluations
irrespective of general anxiety level or total number
of symptoms manifested. This has a high clinical
significance as a patient can increase his or her anxiety
level after the test but cannot have enough symptoms
or similarity of subjective/cognitive symptoms for a
precise panic attack diagnosis.
Our PD patients and first-degree relatives reported
similar symptoms and we observed a similar heart rate
suggesting that panic disorder patients perceive these
symptoms to be more aversive, perhaps resulting in a
greater likehood to panic. This is where the distinction
between PD and the non-panic subjects would be
relevant as one might expect greater catastrophic
cognitions in the PD group. In a previous respiratory
trial with a sample similar to this one, Nardi et al.
[2000] observed the induction of panic attacks by a
hyperventilation challenge test in PD patients and their
healthy first-degree relatives. They randomly selected
25 PD patients, 31 healthy first-degree relatives of
probands with PD, and 26 healthy volunteers. A total
of 44.0% (n 5 11) PD patients, 16.1% (n 5 5) of first-
degree relatives, and 11.5% (n 5 3) of control subjects
had a panic attack after hyperventilating. In this
respiratory challenge test, the PD patients were more
sensitive to hyperventilation than first-degree relatives
and healthy volunteers. Our group also [Nardi et al.,
2002] observed the induction of panic attacks symp-
toms by a breath-hold test in PD patients and their
healthy first-degree relatives. They randomly selected
26 PD patients, 28 healthy first-degree relatives of
probands with PD, and 25 healthy volunteers. Using
specific panic attack criteria, 46.1% (n 5 12) PD
patients, 7.1% (n 5 2) first-degree relatives, and 4.0%
(n 5 1) control subjects had a panic attack after the test.
There was no heart rate, anxiety levels or breath-hold
time differences among the groups. Both data [Nardi
Depression and Anxiety
852 Nardi et al.
et al., 2000, 2002] described the first-degree relatives
reacting in a similar way as the healthy group. This did
not happen in our caffeine challenge test where the PD
group and the first-degree relatives responded in a
similar pattern.
The most interesting aspect of our study is that the
PD patients had a similar response as their first-degree
relatives in the caffeine test suggesting a genetic
vulnerability. Perna et al. [1995] described treatment-
seeking PD probands, their first-degree relatives, and
relatives of healthy controls. Although relatives with a
personal history of any mental disorder were excluded,
22% of those related to PD probands developed a
panic attack following one inhalation of 35% CO2.
Fifty-one percent of the PD probands but only 2% of
the relatives of normal controls developed panic attacks
as well. Thus, high-risk relatives were significantly less
likely to develop panic attacks with CO2 than ill
probands. Bellodi et al. [1998] conducted a twin study
that generated string evidence for the role of genetic
factors in sensitivity to CO2 inhalation. If it could be
established that an abnormal sensitivity to CO2
comprises a trait marker for PD, the test could be
used to identify individuals with a relatively high risk
for the onset of spontaneous panic attacks.
Another important genetic link may be associated
with the adenosine receptor system, which mediates the
psychoactive effects of caffeine, and is also thought to
be involved in the regulation of anxiety [Alsene et al.,
2003]. There is a significant association between self-
reported anxiety after caffeine administration and two
linked polymorphisms on the A2a receptor gene, the
1976C4T and 2592C4Tins polymorphisms [Alsene
et al., 2003]. The variations in anxiogenic responses to
caffeine and polymorphisms in the adenosine receptor
genes have been studied [Lam et al., 2005]. Lam et al.
[2005] tested the hypothesis that the A2aAR 1976T4C
genetic variant confers susceptibility to PD using a
Chinese population. The positive association was
found between polymorphism and PD in a Western
population but not in an Asian population suggesting
that this association could be ethnicity dependent. The
1976C4T polymorphism may be in linkage disequili-
brium with a functional variant that affects PD, and the
extent of this linkage disequilibrium is not similar for
all ethnic populations.
As to other panic-inducing challenge agents, not all
individuals with PD display a panic attack after a
caffeine challenge. Perhaps this may establish the
existence of subgroups of patients with PD, possibly
with differences at clinical, psychological, and physio-
logical levels [Bandelow et al., 1996; Briggs et al.,
1993]. A large amount (e.g. 480–720 mg) of oral
administration of caffeine safely produces symptoms
of anxiety and panic attacks with the complication of
headache only. Caffeine-induced panic attacks include
both somatic and affective symptoms of a panic attack
[Lee et al., 1988; Uhde, 1990]. The core symptoms of
caffeine-induced anxiety include a number of nonpanic
Depression and Anxiety
symptoms, including restlessness, excitement, insom-
nia, and diuresis. Unlike other panicgens, caffeine-
provoked panic is associated with long lasting anxiety
symptoms that may last many hours.
The PD and their first-degree relatives are being
followed up by our research group and we will check
the longer-term risks for the caffeine induction of panic
attacks in the first-degree relatives with no prior history
of panic attacks.
CONCLUSION
Our data may contribute to improve the knowledge
about neurobiology of PD, suggesting that there may
be a genetic association between panic attacks after the
intake of 480 mg of caffeine in PD patients and their
first-degree relatives. Our main finding is the clear
differentiation of PD patients and their first-degree
relatives by a caffeine test from the control group. The
precise and clinical criteria for induced PA may be the
crucial point for our results. The determination of a PA
using numerous objective and subjective measures is
essential as a PA is not a sum of physiological measured
symptoms (blood pressure, heart rate, breathing rate)
but rather the physiological measures changed equally
in all groups.
Acknowledgments. This work is supported by the
Brazilian Council for Scientific and Technological
Development (CNPq), grant no.554411/2005-9.
REFERENCES
Alsene K, Deckert J, Sand P, de Wit H. 2003. Association between
A2a receptor gene polymorphisms and caffeine-induced anxiety.
Neuropsychopharmacology 28:1694–1702.
American Psychiatric Association. 1994. Diagnostic and Statistical
Manual for Mental Disorders. 4th edition. DSM-IV. Washington,
DC: American Psychiatric Press.
Bandelow B, Amering M, Benkert O, Marks I, Nardi AE, Osterheider
M, Tannock C, Tremper J, Versiani M. 1996. Cardio-respiratory
and other symptom clusters in panic disorder. Anxiety 2:99–101.
Bech P, Kastrup M, Rafaelsen OJ. 1986. Mini-compendium of rating
scales for states of anxiety, depression, mania, schizophrenia with
corresponding DSM III syndromes. Acta Psychiatr Scand 73:1–37.
Bellodi L, Perna G, Caldirola D, Arancio C, Bertani A, Di Bella D.
1998. CO2-induced panic attacks: a twin study. Am J Psychiatry
155:1184–1188.
Blanchard J, Sawers SJA. 1983. Chronic comparative pharmacoki-
netics of caffeine in young and elderly men. J Pharmacokin
Biopharmacol 11:109–126.
Boulenger JP, Uhde TW, Wolff EA III, Post RM. 1984. Increased
sensitivity to caffeine in patients with panic disorder. Preliminary
evidence. Arch Gen Psychiatry 41:1067–1071.
Briggs AC, Stretch DD, Brandon S. 1993. Subtyping of panic
disorder by symptom profile. Brit J Psychiatry 163:201–209.
Bruce M, Scott N, Shine P, Lader M. 1992. Anxiogenic effects of
caffeine in patients with anxiety disorders. Arch Gen Psychiatry
49:867–869.
Camargo MC, Toledo MC, Farah HG. 1999. Caffeine daily intake
from dietary sources in Brazil. Food Addit Contam 16:79–87.
 Research Article: A Caffeine Challenge Test in Panic Disorder Patients
Cameron OG, Model JG, Hariharan M. 1990. Caffeine and human
cerebral blood flow: a positron emission tomography study. Life
Sci 47:1141–1146.
Chambless DL, Beck AT, Gracely EJ, Grisham JR. 2000. Relation-
ship of cognitions to fear of somatic symptoms: a test of the
cognitive theory of panic. Depress Anxiety 11:1–9.
Charney DS, Heninger GR, Jatlow PI. 1985. Increased anxiogenic
effects of caffeine in panic disorders. Arch Gen Psychiatry
42:233–243.
Clark DM. 1986. A cognitive approach to panic. Behav Res Ther
24:461–470.
First MB, Spitzer RL, Gibbon M, Williams JBM. 1997. Structured
Clinical Interview Diagnostic (SCID) for DSM-IV axis I Dis-
orders-Clinician Version (SCID-CV). Washington, DC: American
Psychiatric Press.
Gorman JM, Kent JM, Sullivan GM, Coplan JD. 2000. Neuroana-
tomical hypothesis of panic disorder, revised. Am J Psychiatry
157:493–505.
Lam P, Hong CJ, Tsai SJ. 2005. Association study of A2a adenosine
receptor genetic polymorphism in panic disorder. Neurosci Lett
18:98–101.
Lee MA, Flegel P, Greden JF, Cameron OG. 1988. Anxiogenic effects
of caffeine on panic and depressed patients. Am J Psychiatry
145:632–635.
Nardi AE, Nascimento I, Valenca AM, Lopes FL, Zin WA,
Mezzasalma MA, Versiani M. 2002. A breath-holding challenge
in panic disorder patients, their healthy first-degree relatives, and
normal controls. Respir Physiol Neurobiol 23:43–47.
853
Nardi AE, Valenca AM, Nascimento I, Mezzasalma MA, Lopes FL,
Zin WA. 2000. Hyperventilation in panic disorder patients and
healthy first-degree relatives. Braz J Med Biol Res 33:1317–1323.
Nehlig A, Daval JL, Debry G. 1992. Caffeine and the central nervous
system: mechanisms of action, biochemical, metabolic and
psychostimulant effects. Brain Res Rev 17:139–170.
Nickell PV, Uhde TW. 1994/1995. Dose-response effects of
intravenous caffeine in normal volunteers. Anxiety 1:161–168.
Orlikov A, Ryzov I. 1991. Caffeine-induced anxiety and increase of
kynurenine concentration in plasma of healthy subjects: a pilot
project. Biol Psychiatry 29:391–396.
Perna G, Cocchi S, Bertani A, Arancio C, Bellodi L. 1995. Sensitivity
to 35% CO2 in healthy first-degree relatives of patients with panic
disorder. Am J Psychiatry 152:623–625.
Sanderson WC, Rapee RM, Barlow DH. 1989. The influence of
an illusion of control on panic attacks induced via inhalation of
5.5% carbon dioxide-enriched air. Arch Gen Psychiatry 46:
157–162.
Uhde TW. 1990. Caffeine provocation of panic: a focus on biological
mechanisms. In: Ballenger J, editor. Neurobiology of panic
disorder (frontiers of clinical neuroscience). Vol. 8. New York,
NY: Alan R. Liss. p 365–376.
Uhde TW. 1995. Caffeine-induced anxiety: An ideal chemical model
of panic disorder. In: Asnis GM, van Praag HM, editors. Einstein
Monograph Series in Psychiatry. New York: Wiley-Liss.
p 181–205.
Wolpe J, Rowan VC. 1988. Panic disorder: a product of classical
conditioning. Behav Res Ther 26:441–450.
Depression and Anxiety