Brazilian Journal of Medical and Biological Research (2000) 33: 1317-1323 1317

Hyperventilation and panic

ISSN 0100-879X

Hyperventilation in panic disorder

patients and healthy first-degree
relatives
A.E. Nardi1, 1Laboratóriode Pânico e Respiração, Instituto de Psiquiatria, and
A.M. Valença1, 2Laboratóriode Fisiologia da Respiração, Instituto de Biofísica Carlos Chagas Filho,
I. Nascimento1, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ, Brasil
M.A. Mezzasalma1,
F.L. Lopes1
and W.A. Zin2

Abstract

Correspondence Our aim was to observe the induction of panic attacks by a hyperven- Key words
A.E. Nardi tilation challenge test in panic disorder patients (DSM-IV) and their · Panic attacks
Laboratório de Fisiologia da healthy first-degree relatives. We randomly selected 25 panic disorder · Respiration
Respiração, IPUB-UFRJ Anxiety disorder
patients, 31 healthy first-degree relatives of probands with panic ·
22410-003 Rio de Janeiro, RJ Ventilation
Brasil
disorder and 26 normal volunteers with no family history of panic ·

Fax: +55-21-523-6839 disorder. All patients had no psychotropic drugs for at least one week.
E-mail: aenardi@novanet.com.br They were induced to hyperventilate (30 breaths/min) for 4 min and
anxiety scales were applied before and after the test. A total of 44.0%
Research supported by CNPq (N = 11) panic disorder patients, 16.1% (N = 5) of first-degree
(No. 300500/93-9). relatives and 11.5% (N = 3) of control subjects had a panic attack after
hyperventilating (c² = 8.93, d.f. = 2, P = 0.011). In this challenge test
the panic disorder patients were more sensitive to hyperventilation
Received December 1, 1999
than first-degree relatives and normal volunteers. Although the hyper-
Accepted August 7, 2000 ventilation test has a low sensitivity, our data suggest that there is no
association between a family history of panic disorder and hyperreac-
tivity to an acute hyperventilation challenge test. Perhaps cognitive
variables should be considered to play a specific role in this associa-
tion since symptoms of a panic attack and acute hyperventilation
overlap.

Introduction tidal CO2 to conventionally accepted levels

of hypocapnia, seems to be a less convincing
Hyperventilation is an increase in the panicogenic challenge than CO2 inhalation
amount of inhaled and exhaled air per minute, (3,4). However, for a small group of panic
exceeding the amount necessary for normal disorder patients hyperventilation may be a
cell metabolism (1). In this condition, venti- safe and easy test for a more precise diagno-
lation exceeds cellular metabolism, resulting sis. Anyway, hyperventilation has been con-
in an elimination of CO2 greater than its sidered as a cause, a correlate and a conse-
production, which causes a decline in the quence of panic attacks (5). Stress-induced
arterial partial pressure of carbon dioxide hyperventilation produces symptoms that
(PCO2) (1,2). people tend to misinterpret as life threaten-
Hyperventilating at 30 breaths per minute, ing if they are unaware of the consequences
although causing a significant drop in end- of overbreathing (1,5). Misinterpretation of

Braz J Med Biol Res 33(11) 2000

1318 A.E. Nardi et al.

these symptoms increases fear and activates potheses about panic disorder (10,11).
the autonomic nervous system, resulting in Since there is substantial evidence that
increasing respiratory frequency that will panic disorder runs in families (12), and that
further dissipate carbon dioxide and inten- affected subjects may share some biological
sify hypocapnic symptoms (1,5,6). sensitivity (7,8), the objective of the present
Respiratory abnormalities are associated study was to determine if there is any differ-
with anxiety, particularly with panic attacks ence in induced panic attacks using room air
(1,2). Symptoms such as shortness of breath, hyperventilation (30 breaths/min) over a pe-
“empty-head” feeling, dizziness, paresthesias riod of 4 min among panic disorder patients,
and tachypnea have been described in the some of their healthy first-degree relatives
psychiatric and respiratory physiology re- and a control group.
lated to panic disorder (2,3).
A high prevalence of anxiety disorders Patients and Methods
has been found among first-degree relatives
of subjects with panic attacks. Balon et al. We randomly selected 25 panic disorder
(7) infused sodium lactate in 45 control sub- patients, 14 women and 11 men (mean age ±
jects. An investigator, blind to the outcome SD: 33.8 ± 9.4 years), and 31 of their healthy
of the infusion study, obtained family histo- first-degree relatives (parents or children),
ries of 160 relatives. There was no signifi- 19 women and 12 men (mean age ± SD: 41.5
cant difference in the prevalence of mood ± 12.8 years), in the Laboratory of Panic and
disorders and substance abuse between first- Respiration of the Institute of Psychiatry,
degree relatives of subjects with (N = 45) Federal University of Rio de Janeiro. A group
and without (N = 115) panic attacks, but of 26 subjects with no family history of panic
there was a difference in the prevalence of disorder (14 women and 12 men, mean age ±
anxiety disorders, suggesting that individu- SD: 29.7 ± 7.9 years) was used for compari-
als with a family history of anxiety disorders son. The diagnosis was made by one rater
may be vulnerable to lactate-induced panic (I.N.) using the Structured Clinical Inter-
attacks. view Diagnostic (SCID) (13) for DSM-IV
Perna et al. (8) tested the hypothesis that (14). Patients who met DSM-IV criteria for
hyperreactivity to CO2 in healthy subjects current major depression, bipolar disorder,
represents an underlying familial vulnerabil- obsessive-compulsive disorder, schizophre-
ity to panic disorder. They studied 84 panic nia, delusional or psychotic disorders, or-
disorder patients, 23 healthy first-degree rela- ganic brain syndrome, severe personality dis-
tives of probands with panic disorder, and 44 order, epilepsy, or substance abuse or de-
healthy subjects with no family history of pendence (during the previous year) were
panic disorder. The first-degree relatives of excluded. Patients with comorbid dysthymia,
the probands with panic disorder reacted generalized anxiety disorder or past major
significantly more than the healthy subjects depression were included if panic disorder
without a family history of panic disorder. was judged to be the principal diagnosis.
These findings suggested an association be- The protocol was explained to the subjects,
tween a family history of panic disorder and who signed a voluntary written consent to
hyperreactivity to 35% CO2 inhalation. participate in the study. The subjects were
Multiple lines of evidence have shown informed that they would be asked to
biological, and especially respiratory, symp- hyperventilate room air and that the proce-
toms involved in anxiety disorders (3,4,9). dure was not dangerous but that anxiety or
The respiratory challenge test strategies have panic could occur during the session. Our
been especially fruitful in generating hy- local Ethics Committee approved the proto-

Braz J Med Biol Res 33(11) 2000

Hyperventilation and panic 1319

col, which complied with the principles of = very severe). Both self-rating scales, the
the Declaration of Helsinki. Inclusion crite- SUDS and the DSQ, had been tested by back
ria were: 18 to 55 years of age, occurrence of translation but not by test-retest examina-
at least three panic attacks in the two weeks tion. After the test the SUDS and DSQ were
before the challenge test day in panic disor- completed again. The subjects completed
der patients, no use of any psychotropic them independent of the raters. On the basis
drugs for at least one week by any subject, of the DSQ, a panic attack was defined as the
and a negative urine test for benzodiazepines following: 1) four or more symptoms of a
and other medications. Exclusion criteria panic attack from the DSM-IV; 2) at least
were: unstable medical condition, psycho- one of the cognitive symptoms of a panic
therapy during the study, use of any regular attack from the DSM-IV (e.g., fear of dying
antipsychotic, antidepressant, regular ben- or of losing sanity or control); 3) feeling of
zodiazepine or nonbenzodiazepine anxiolytic panic or fear, similar to spontaneous panic
medication for 4 weeks, or fluoxetine for 5 attacks recorded on a card which the raters
weeks before the test, or the presence of were not permitted to observe, and 4) agree-
suicidal risk. Subjects with a history of respi- ment at clinical diagnosis evaluation between
ratory disease and smokers were also ex- two diagnosis-blinded raters from the team
cluded. during the test. The feeling of a panic attack
The first-degree relatives (parents or chil- reported by the subjects was also examined
dren) and the comparison group were also in order to compare agreement between rat-
assessed by the SCID (13). They were free of ers and subjects about panic attacks.
any history of current panic disorder, major We explained to the subjects what they
mood disorder, schizophrenia, and current were expected to do and submitted them to a
substance abuse disorders. 30-s period of training for hyperventilation.
All subjects underwent a physical exami- The subject then relaxed for an additional 10
nation and laboratory exams to ensure they min, after which we induced hyperventila-
were healthy enough to participate in a hy- tion (30 respiratory movements per minute
perventilation challenge test. They had no over a period of 4 min) with a rater counting
respiratory or cardiovascular abnormalities aloud the movements measured with a chro-
and were free of caffeine ingestion for one nometer. Immediately after this period we
day. The test was conducted in the usual evaluated the level of anxiety and the induc-
examination room, with no changes made in tion of a panic attack.
the environment. All subjects were asked to
relax for 10 min. We then checked respira- Statistical analysis
tory frequency, pulse and blood pressure and
repeated the measurements 1 and 5 min after Panic rates for the three groups, also
the test. To measure the baseline anxiety separated according to rater and subject rat-
level, before hyperventilating subjects were ings and by gender, were compared by the
asked to complete the Subjective Units of chi-square test. Data concerning the effects
Disturbance Scale (SUDS), a semiquantita- of hyperventilation and time of observation
tive evaluation method ranging from 0 = no were tested by ANOVA. Pairwise compari-
anxiety, to 10 = maximum anxiety (15), and sons of the treatment groups were performed
the Diagnostic Symptom Questionnaire at end-point, using Fisher’s protected least
(DSQ) (15) adapted for DSM-IV in which significant difference method. The Kruskal-
the presence and level of discomfort of panic Wallis test was also applied to the main
symptoms experienced after the inhalations efficacy parameters. The level of signifi-
were rated on a 0-4 point scale (0 = none, 4 cance was set at 5%.

Braz J Med Biol Res 33(11) 2000

1320 A.E. Nardi et al.

Results perventilating. Patients with panic disorder

tended to be more sensitive to the effects of
Table 1 shows the panic rates assigned by hyperventilation than first-degree relatives
the raters and by the subjects during the and normal controls. All subjects showed an
challenge test. Significantly more panic pa- increase in anxiety levels after hyperventi-
tients had a panic attack in response to hy- lating. The Kruskal-Wallis test (2.574, d.f. =
perventilation than first-degree relatives or 2, P = 0.276) was also applied to the main
the control group according to both rater efficacy parameters.
groups: 44.0% (N = 11) panic disorder pa-
tients, 16.1% (N = 5) first-degree relatives Discussion
and 11.5% (N = 3) control subjects had a
panic attack after hyperventilating. There This study compares panic disorder pa-
was no significant sex difference in any group tients and some of their first-degree relatives
(ANOVA, F = 1.42, d.f. = 2, P = 0.957). in a hyperventilation challenge test with a
Table 2 shows the SUDS subjective anxi- control group. In our sample the hyperventi-
ety level measurement before and after hy- lation challenge test (30 breaths/min over a
period of 4 min) differentiated panic disor-
Table 1 - Response of patients with panic disorder, healthy first-degree der patients from normal controls. We found
relatives and normal control subjects challenged with room air hyperventi- no difference in sex or anxiety response
lation.
among our groups. Perhaps these hyperven-
Raters evaluation of panic attacks: c² = 8.93, d.f. = 2, P = 0.011 (panic tilation-sensitive panic disorder patients will
disorder vs relatives; c² with Yates correction = 3.99, P = 0.045) (panic be clearly identified in the future with a more
disorder vs control; c² with Yates correction = 5.21, P = 0.022) (relatives vs
sophisticated methodology. Some method-
control; Fisher exact test, two-tailed, P = 0.715). Subject evaluation of
panic attacks: c² = 14.88, d.f. = 2, P = 0.0005 (panic disorder vs relatives; c² ological improvements can be made in order
with Yates correction = 8.24, P = 0.004) (panic disorder vs control; c² with to increase the sensitivity of the test.
Yates correction = 7.90, P = 0.004) (relatives vs control; Fisher exact test,
During voluntary hyperventilation, pa-
two-tailed, P = 1.00).
tients with panic disorder often recognize
Rating after Panic disorder First-degree relative Normal control some of the symptoms felt during a panic
hyperventilation (N = 25) (N = 31) (N = 26) attack (16-18). For this reason hypocapnia
Raters
may be an important factor in the anxiety
Panic 11 (44.0%) 5 (16.1%) 3 (11.5%) that occurs during a panic attack and hyper-
No panic 14 (56.0%) 26 (83.9%) 23 (88.5%) ventilation (4,11). These data suggest that
Subject hyperventilation occurs during some panic
Panic 13 (52.0%) 4 (12.9%) 3 (11.5%) attacks. Papp et al. (11), using basal arterial
No panic 2 (48.0%) 27 (87.1%) 23 (88.5%)
blood gas levels, found that only the patients
who had panic attacks after lactate infusion
Table 2 - Subjective anxiety levels rated by the subjects just before and
were chronic hyperventilators, as indicated
after hyperventilation (mean ± SD). by low levels of blood bicarbonate when
compared to controls. This study also shows
SUDS, Subjective Units of Disturbance Scale. ANOVA, F = 2.311, d.f. = 2,
that hyperventilation is a relatively mild
P = 0.106 (Fisher’s protected least significant difference: panic disorder vs
relatives, P = 0.058; panic disorder vs control, P = 0.070; relatives vs panicogenic agent.
control, P = 0.997). In a preliminary study (19) we observed
the induction of panic attacks by hyperventi-
Self-rating Panic disorder First-degree relative Normal control
(N = 25) (N = 31) (N = 26) lation in a group of panic disorder patients
(DSM-IV). Thirteen panic disorder patients
SUDS before 3.8 ± 3.1 2.7 ± 3.8 2.4 ± 1.9 and 11 normal volunteers were randomly
SUDS after 8.5 ± 4.6 4.1 ± 4.0 3.5 ± 3.2
selected and induced to hyperventilate (30

Braz J Med Biol Res 33(11) 2000

Hyperventilation and panic 1321

breaths/min) for 3 min. Anxiety scales were and 12 patients with other anxiety disorders
applied before and after the test. Nine (69.2%) (obsessive-compulsive disorder and gener-
panic disorder patients and one (9.1%) of the alized anxiety disorder) the authors found no
control subjects had a panic attack after significant differences in base excess or bi-
hyperventilating. The difference observed in carbonate levels between patients with panic
these two studies may be due to a better disorder, other anxiety disorders and healthy
training of the research group in the hyper- controls. This result also suggests that there
ventilating challenge test, the size of the is no chronic hyperventilation in patients
samples or the severity of the disorder. with panic disorder.
Several other studies do not give support Panic disorder patients exhibit both be-
to the theory that hyperventilation arouses haviorally and physiologically abnormal re-
panic attacks or contributes to their severity sponses to respiratory challenge tests (10,11).
(11,20,21). As hyperventilation frequently Panic patients reported significantly more
occurs, it may not be a cause or an important panic attacks and anxiety during the chal-
component of panic attacks. It may be that lenges than normal volunteers (1,2,9). In an
hyperventilation is a consequence of panic anxiety disorder sample (9), patients with
attacks in hyperventilation-predisposed in- DSM-III agoraphobia, panic disorder, gen-
dividuals (10,21,22). A variety of studies eralized anxiety disorder, social phobia and
have shown that differences between indi- normal controls underwent a series of exper-
viduals with and without panic disorder in imental procedures and measurements to de-
measurements of panic during hyperventila- termine whether panic attack patients show a
tion challenges are considerably lower than greater tendency towards hyperventilation
those observed in CO2 challenges (23,24). In that is independent from their anxiety levels.
addition, panic disorder individuals appeared The agoraphobia and panic disorder patients
to be less compliant with the hyperventila- did not show significantly lower levels of
tion instructions than were other groups (23). expired PCO2 at rest than the other anxious
Gorman et al. (17) conducted a study on or non-anxious groups. However, the panic
12 patients with panic disorder or agorapho- attack patients did show significantly higher
bia with panic attacks (DSM-III). The pa- levels of anxiety and hyperventilatory symp-
tients were asked to hyperventilate (30 respi- toms during a hyperventilation test or during
ratory movements/min) for 15 min. After the breathing 5% CO2 in air. A strong relation-
test the patients filled in the Acute Panic ship was found between hyperventilatory
Inventory, which measures the severity of symptoms and anxiety in all anxiety disorder
the typical panic symptoms. Three (25%) of groups of patients and in the controls. On the
the 12 patients had panic attacks after this basis of these results, Holt and Andrews (9)
procedure, demonstrating that the induction concluded that agoraphobia and panic disor-
of respiratory alkalosis would not be enough der patients do not show a unique tendency
to provoke panic attacks in all patients. Acute toward hyperventilation, but rather that their
hyperventilation is a probable cause of panic hyperventilatory symptoms and perhaps in-
attacks in a percentage of patients with panic termittent overbreathing episodes are a func-
disorder (1,4,21). tion of the high levels of anxiety they experi-
Zandbergen et al. (21) looked for evi- ence. Their method probably did not distin-
dence of chronic hyperventilation in 18 pa- guish panic disorder patients from social
tients with panic disorder (DSM-III-R) by phobics because they relied only on the pa-
measuring their arterial PCO2, base excess, tients’ subjective judgement for the diagno-
bicarbonate and pH. Comparing the data sis of a panic attack, the hyperventilation
with those obtained for 18 healthy controls lasted just 1.5 min and 36% of the panic

Braz J Med Biol Res 33(11) 2000

1322 A.E. Nardi et al.

disorder patients and 31% of the social pho- is rapidly accumulating. Panic disorder pa-
bic patients were on medication during the tients seem to be more sensitive to hyperven-
test. In our hyperventilatory challenge test tilation than normal volunteers. Our data
the results were statistical significant for support Klein’s theory (26) proposing the
panic attacks (Table 1) and the anxiety levels existence of a “suffocation false alarm” mech-
(Table 2) tended to show that panic disorder anism. According to this theory, one of the
patients may be more sensitive to respiratory mechanisms that signal suffocation is a rise
tests than first-degree relatives and volun- in CO2 concentration. The suffocation false
teers. In our study the panic attack diagnostic alarm as well as the CO2 hypersensitivity
criteria were clearly defined, patients had may act in conjunction with cognitive fac-
not been taking any drug for at least one tors to induce panic attacks in panic disorder
week and they were induced to hyperventilate patients. There still are a number of critical
for 4 min. Another consideration is that our areas in need of exploration before such
sample consisted of moderate and severe assertion can be fully validated. These areas
panic disorder patients. include the need to establish in which group
The present study has some limitations. of panic disorder patients hyperventilation
First, the proband group tested, possibly be- can be used as a specific test, whether it
cause of the exclusion criteria, might not be varies with patient age or sex and to what
representative of other series of patients with extent it is cognitively mediated (27). Our
panic disorder. Second, only the healthy rela- results are different from those of Balon et
tives who accepted to participate in the test al. (7) and of Perna et al. (8), who found a
were studied. Some more anxious and per- high prevalence of increased anxiety and
haps more vulnerable subjects may not have panic attacks in the relatives of panic disor-
been included. der patients (8) and in healthy subjects, sug-
According to the cognitive approach (25), gesting that lactate-induced panic in panic
the catastrophic interpretation of bodily sen- disorder relatives (7) and CO2 challenge test
sations induced by hyperventilation may play (8) may be a trait marker for genetic vulner-
a more important role in the panic attack. ability to anxiety disorders (27). Our data
The erroneous interpretation of bodily sen- also suggest that the hyperventilation test is a
sations as dangerous and threatening would specific marker for some patients but not
result in an increase in anxiety, which in turn strong enough to be a trait marker. Cognitive
would lead to new sensations, and thus result factors are also clearly present in the hyper-
in a vicious cycle until the panic attack. The ventilatory test (25,27). The CO2 and lactate
respiratory re-training techniques may re- infusion tests can be easily compared with a
duce the severity of symptoms during the placebo test but the difficulty of using a
panic attack by decreasing the hyperventila- placebo test in the hyperventilatory test makes
tory response (8,25). We also observed that its results weaker.
some symptoms of this syndrome (anxiety, The respiratory challenge test needs more
faster or deeper breathing, palpitations, ta- research with a large number of patients to
chycardia, dizziness, tremors, feelings of help elucidate the neurobiologic vulnerabil-
panic, uneasiness, fatigue, crying spells, ity of panic disorder patients and may lead to
paresthesias) are similar to the clinical find- the development of a more specific diagno-
ings of panic disorder, and many of them are sis and effective treatments. Our data sug-
part of the diagnostic criteria for panic at- gest that there is no association between a
tacks on the basis of the DSM classification. family history of panic disorder and hyperre-
Evidence implicating respiratory abnor- activity to an acute hyperventilation chal-
malities in the pathogenesis of panic attacks lenge test. Perhaps cognitive variables should

Braz J Med Biol Res 33(11) 2000

Hyperventilation and panic 1323

be considered to play a specific role in this with cognitive therapy control are needed.
association since symptoms of a panic attack
and acute hyperventilation overlap. The im- Acknowledgments
portance of cognitive, genetic and environ-
mental factors in the reactivity to respiratory The authors thank Ronir R. Luiz for the
tests remains unclear. Studies on twins and statistical analysis.

References

1. Papp LA, Martinez JM, Klein DF, Coplan Psychiatry, 152: 623-625. 861.
JD, Norman RG, Cole R, de Jesus MJ, 9. Holt PE & Andrews G (1989). Hyperventi- 18. Ley R (1985). Agoraphobia, the panic at-
Ross D, Goetz R & Gorman JM (1997). lation and anxiety in panic disorder, social tack, and the hyperventilation syndrome.
Respiratory psychophysiology of panic phobia, GAD and normal controls. Behav- Behaviour Research and Therapy, 23: 79-
disorder: three respiratory challenges in iour Research and Therapy, 27: 453-460. 81.
98 subjects. American Journal of Psychia- 10. Bataglia M & Perna G (1995). The 35% 19. Nardi AE, Valença AM, Nascimento I,
try, 154: 1557-1565. CO2 challenge in panic disorder: optimiza- Mezzasalma MA & Zin WA (1999). Panic
2. Gorman JM, Papp LA, Coplan JD, tion by receiver operating characteristic disorder and hyperventilation. Arquivos de
Martinez JM, Lennon S, Goetz RR, Ross (roc) analysis. Journal of Psychiatry Re- Neuropsiquiatria, 57: 932-936.
D & Klein DF (1994). Anxiogenic effects search, 29: 111-119. 20. Garssen B, Buikhuisen M & Dick R (1996).
of CO2 and hyperventilation in patients 11. Papp LA, Klein DF & Gorman JM (1993). Hyperventilation and panic attacks. Ameri-
with panic disorder. American Journal of Carbon dioxide hypersensivity, hyperven- can Journal of Psychiatry, 153: 513-518.
Psychiatry, 151: 547-553. tilation and panic disorder. American Jour- 21. Zandbergen J, Lousberg HA, Pols H, Loof
3. Rapee R, Brown T, Antony M & Barlow D nal of Psychiatry, 150: 1149-1157. C & Griez E (1990). Hypercarbia versus
(1992). Response to hyperventilation and 12. Crowe RR, Noyes R, Pauls DL & Slymen hypocarbia in panic disorder. Journal of
inhalation of 5.5% carbon dioxide-en- D (1983). A family study of panic disorder. Affective Disorders, 18: 75-81.
riched air across the DSM-III-R anxiety Archives of General Psychiatry, 40: 1065- 22. Bass C (1997). Hyperventilation syn-
disorders. Journal of Abnormal Psychol- 1069. drome: a chimera? Journal of Psychoso-
ogy, 101: 538-552. 13. First MB, Spitzer RL, Gibbon M & Wil- matic Research, 42: 421-426.
4. Maddock RJ & Carter CS (1991). Hyper- liams JBM (1997). Structured Clinical In- 23. Antony MM, Brown TA & Barlow DH
ventilation-induced panic attacks in panic terview Diagnostic (SCID) for DSM-IV Axis (1997). Response to hyperventilation and
disorder with agoraphobia. Biological Psy- I Disorders - Clinician Version (SCID-CV). 5.5% CO2 inhalation of subjects with
chiatry, 29: 843-854. American Psychiatric Press, Washington. types of specific phobia, panic disorder,
5. Papp LA, Klein DF & Gorman JM (1998). 14. American Psychiatric Association (1994). or NO mental disorder. American Journal
Carbon dioxide hypersensitivity, hyper- Diagnostic and Statistical Manual for Men- of Psychiatry, 154: 1089-1095.
ventilation, and panic disorder. In: Bellodi tal Disorders. DSM-IV. 4th edn. American 24. Humble M (1987). Aetiology and mechan-
L & Perna G (Editors), The Panic Respira- Psychiatric Press, Washington. isms of anxiety disorders. Acta Psychia-
tion Connection. Medical Media, Milano. 15. Bech P, Kastrup M & Rafaelsen OJ (1986). trica Scandinavica, 76 (Suppl 335): 15-30.
6. Bandelow B, Amering M, Benkert O, Mini-compendium of rating scales for 25. Salkovskis PM, Jones DO & Clark DM
Marks I, Nardi AE, Osterheider M, Tan- states of anxiety, depression, mania, (1986). Respiratory control in treatment
nock C, Tremper J & Versiani M (1996). schizophrenia with corresponding DSM of panic attacks: replication and extension
Cardio-respiratory and other symptom III syndromes. Acta Psychiatrica Scandi- with concurrent measurement of behav-
clusters in panic disorder. Anxiety, 2: 99- navica, 73 (Suppl 326): 1-37. iour and PCO2. British Journal of Psychia-
101. 16. Sanderson WC, Rapee RM & Barlow DH try, 148: 526-532.
7. Balon R, Jordan M, Pohl R & Yeragani VK (1989). The influence of an illusion of con- 26. Klein DF (1996). Panic disorder and agora-
(1989). Family history of anxiety disorders trol on panic attacks induced via inhalation phobia: hypothesis hothouse. Journal of
in control subjects with lactate-induced of 5.5% carbon dioxide-enriched air. Ar- Clinical Psychiatry, 57 (Suppl 6): 21-27.
panic attacks. American Journal of Psy- chives of General Psychiatry, 46: 157-162. 27. Coryell WH (1998). Familial risk for panic
chiatry, 146: 1304-1306. 17. Gorman JM, Askanazi J, Liebowitz MR, disorder and CO2 sensitivity. In: Bellodi L
8. Perna G, Cocchi S, Bertani A, Arancio C & Fyer AJ, Stein J, Kinney JM & Klein DF & Perna G (Editors), The Panic Respiration
Bellodi L (1995). Sensitivity to 35% CO2 (1984). Response to hyperventilation in a Connection. Medical Media, Milano.
in healthy first-degree relatives of patients group of patients with panic disorder. A-
with panic disorder. American Journal of merican Journal of Psychiatry, 141: 857-

Braz J Med Biol Res 33(11) 2000