Impact of Panic Attacks on Rehabilitation

and Quality of Life Among Persons

With Severe Psychotic Disorders
Renee Goodwin, Ph.D.
David A. Stayner, Ph.D.
Matthew J. Chinman, Ph.D.
Larry Davidson, Ph.D.

Objective: The study evaluated data from a sample of persons with severe psychotic disorders are relatively
psychotic disorders to determine whether those with and without comor- scarce.
bid panic attacks differed in rates of comorbidity of other psychiatric dis- Several studies have revealed high-
orders, in quality of life, and in rehabilitation outcomes. Methods: A total er-than-expected rates of panic attacks
of 120 individuals with psychotic disorders were assessed with the Center among patients with psychotic disor-
for Epidemiologic Studies–Depression scale, the Structured Clinical In- ders in psychiatric treatment settings
terview for DSM-III-R, the General Health Questionnaire, the Global As- (6–8). Studies have also shown that
sessment of Functioning scale, and several quality-of-life measures at persons with co-occurring panic at-
baseline and four and a half months after they had participated in a social tacks and psychosis may have different
rehabilitation program. Multivariate analyses of variance and Pearson’s responses to psychopharmacologic
chi square tests were used to compare baseline and follow-up scores be- treatment than persons who have a
tween individuals who did and did not have panic attacks. Results: Eigh- psychotic disorder only (9). In fact,
teen (15 percent) of the participants who had severe psychotic disorders some reports recommend the use of
also had panic attacks. Participants with this type of comorbidity had sig- alternative psychotherapies for opti-
nificantly higher rates of major depressive disorder, specific phobia, seda- mal treatment of individuals who have
tive abuse, polysubstance abuse, other substance abuse, and anorexia ner- co-occurring panic attacks and schizo-
vosa than participants who did not have panic attacks. Participants who phrenia (10). However, no studies
had panic attacks also had poorer rehabilitative outcomes and poorer have been published on the impact of
quality of life at baseline and at follow-up than participants who did not
have panic attacks. Conclusions: These data are the first to show that co-
morbid panic attacks are associated with poorer rehabilitative outcomes
and poorer quality of life among individuals with severe psychotic disor-
ders than among those who have psychotic disorders without panic at- Editor’s Note: This paper is part
tacks. Panic attacks may be a valuable prognostic indicator among persons of a series on anxiety disorders
with psychotic disorders and may have implications for treatment and re- edited by Kimberly A. Yonkers,
habilitation. (Psychiatric Services 52:920–924, 2001) M.D. Contributions are invited
that address panic disorder, ago-
raphobia, obsessive-compulsive

P
anic attacks are common ty (1,2), poorer treatment response disorder, social phobia, posttrau-
among individuals who have (3), and lower quality of life has been matic stress disorder, and gener-
lifetime major depression, anx- the focus of numerous investigations alized anxiety disorder. Papers
iety disorders, and substance use dis- (4,5). In contrast, data on the impact should focus on integrating new
orders (1–5). The relationship be- of comorbid panic attacks on the information that is clinically rele-
tween comorbid panic attacks and el- prognosis, course of illness, and qual- vant and that has the potential to
evated rates of psychiatric comorbidi- ity of life among individuals who have improve some aspect of diagnosis
or treatment. For more informa-
tion, please contact Dr. Yonkers
Dr. Goodwin is affiliated with the department of psychiatry at the College of Physicians and at 142 Temple Street, Suite 301,
Surgeons and the division of epidemiology of the Joseph L. Mailman School of Public Health New Haven, Connecticut 06510;
at Columbia University, 1051 Riverside Drive, Unit 43, New York, New York 10032 (e-mail, 203-764-6621; kimberly.yonkers
rdg66@columbia.edu). Dr. Stayner, Dr. Chinman, and Dr. Davidson are with the depart- @yale.edu.
ment of psychiatry at the Yale University School of Medicine in New Haven, Connecticut.

920 PSYCHIATRIC SERVICES ♦ July 2001 Vol. 52 No. 7

panic attacks on long-term treatment Table 1
outcomes, the effectiveness of com- Sociodemographic differences of 120 study participants who had psychotic disor-
munity-based rehabilitation programs, ders with and without comorbid panic attacks
and quality of life among persons with
psychotic disorders. Results of numer- Without panic attacks With panic attacks
ous previous studies suggest that panic (N=102) (N=18)
attacks influence the long-term treat- Characteristic N or mean % N or mean %
ment outcomes of mood and anxiety
disorders (3) as well as functional im- Age (mean±SD years) 40.8±10.4 39.0±6.2
pairment and suicidality among per- Sex
sons without a psychiatric diagnosis Female 51 50 11 56
Male 51 50 7 44
(11). Therefore it is likely that these Race
symptoms also affect the treatment African American 18 18 0 —
outcomes and well-being of persons Hispanic 1 1 0 —
who have severe psychotic illness. White 82 80 18 100
The goal of our study was threefold. Asian or Pacific Islander 1 2 0 —
Marital status
First, we sought to determine the Never married 68 67 12 67
prevalence of comorbid panic attacks Married once 7 7 1 5
among a group of persons who had se- Divorced 21 21 4 22
vere and persistent psychotic disor- Divorced and remarried 2 2 0 —
ders. Second, we wanted to compare Widowed 1 2 0 —
Widowed and remarried 1 1 1 5
the sociodemographic characteristics Have children 39 38 4 22
and the presence of mood, anxiety, Education
substance use, and eating disorders Grade 6 or less 1 1 0 —
among persons who had psychotic dis- Grade 7 to 12 18 18 4 22
orders with and without panic attacks. Graduated from high school1 49 48 3 17
Some college 27 26 6 33
Third, we wanted to compare the re- Graduated from two-year college 0 — 2 11
habilitation outcomes of participants Graduated from four-year college 6 6 3 17
who had psychotic disorders with and Some graduate school 2 2 0 —
without panic attacks in terms of qual- Hollingshead social class category
ity of life at baseline and four and a Unskilled or manual labor 82 80 13 75
Semiskilled labor 16 16 5 25
half months after they had participat- Skilled craftsman 2 2 0 —
ed in a social rehabilitation program. Medium-sized business or
On the basis of findings from previ- minor professional 4 3 0 —
ous studies in which the negative ef- Currently employed 6 24 3 19
fects of panic attacks on the function- 1 χ2=19.2, df=6, p=.004
ing of persons with psychosis were
documented (1–5), we hypothesized
that participants who had coexisting
psychotic disorders and panic attacks severe and persistent mental illness demiologic Studies–Depression (CES-
would have higher rates of comorbid- into the community by providing D) scale (16,17), subscales from
ity of other psychiatric disorders and weekly connections with community Lehman’s Quality of Life scale (18,
a poorer quality of life at follow-up volunteers who participated in leisure 19), the Global Assessment of Func-
than participants who had psychotic activities with the participants. The de- tioning (GAF) scale (20,21), and the
disorders without panic attacks. sign, methodology, and qualitative General Health Questionnaire (22)
data of the study, which was conduct- were used to evaluate participants at
Methods ed between 1991 and 1996, have been baseline and at follow-up.
The Partnership Project was a federal- described elsewhere (12,13). Pearson’s chi square tests were used
ly funded psychiatric rehabilitation Of the 260 individuals who partici- to determine the differences in so-
study aimed at facilitating participa- pated in the Partnership Project, those ciodemographic characteristics and
tion in a community-based treatment who had a diagnosis of a psychotic dis- psychiatric comorbidity between indi-
program for persons with serious and order (N=120) were included in the viduals who met criteria for psychotic
persistent mental illness (12,13). A to- study reported here. After the partici- disorders with and without panic at-
tal of 260 participants were recruited pants had been given a complete de- tacks. All tests were two-sided, and
from the Connecticut Mental Health scription of the study, written in- the significance level was set at .05.
Center outpatient programs and other formed consent was obtained. Multivariate analyses of variance were
community-based mental health facil- Diagnoses were made by trained used to compare scores on the Gener-
ities in New Haven, Connecticut. The interviewers using the Structured al Health Questionnaire, the GAF
aim of the program was to improve the Clinical Interview for DSM-III-R scale, the CES-D scale, and subscales
social integration of persons who had (SCID) (14,15). The Center for Epi- of the Quality of Life scale (purpose,
PSYCHIATRIC SERVICES ♦ July 2001 Vol. 52 No. 7 921
Table 2 of the measures between baseline
Psychiatric comorbidity among 120 patients who had psychotic disorders with and and follow-up, although this improve-
without panic attacks ment was not statistically significant.

Without panic With panic Discussion

attacks (N=102) attacks (N=18) This study is the first to provide data
Mental disorder N % N % χ2†
suggesting that comorbid panic at-
p
tacks have a detrimental impact on
Major depressive disorder 27 26 13 68 13.5 .001 rehabilitation outcomes and quality
Bipolar disorder 10 9 4 21 2.7 .1 of life among persons with severe psy-
Dysthymia 4 4 2 6 ns chotic disorders. These data indicate
Alcohol use disorder 42 41 11 61 ns
Sedative abuse 4 4 5 24 9.5 .009
that panic attacks are prevalent—af-
Cocaine abuse 7 7 4 22 5.6 .06 fecting about 15 percent of persons
Cannabis abuse 24 23 6 33 3.5 .2 with psychotic disorders—and are as-
Stimulant abuse 6 6 3 12 ns sociated with a significantly elevated
Opioid abuse 4 4 4 18 5.5 .066 risk of comorbid psychiatric disorders
Hallucinogen abuse 3 3 0 — 4.3 .12
Polysubstance abuse 6 6 5 24 6.2 .033
among patients with severe psychotic
Other substance abuse 3 3 4 18 7.0 .031 illness in community based treatment
Social phobia 0 — 1 5 6.1 .1 settings (6–8).
Specific phobia 0 — 2 11 11.7 .02 The reasons for the higher-than-ex-
Obsessive-compulsive pected prevalence of panic attacks
disorder 11 10 3 16 ns
Generalized anxiety
among individuals with severe psy-
disorder 0 — 1 5 5.9 .1 chotic disorders are unknown. It is
Anorexia nervosa 0 — 2 11 12.2 .019 possible that panic attacks lead to psy-
chosis, or that psychosis increases the
† df=1 risk of panic attacks. It is also possible
that a third factor increases the risk of
both panic attacks and psychosis. It is
mastery, and growth as aspects of Comorbidity of psychotic disorders not inconceivable—given the differ-
well-being; satisfaction with school; Participants who had psychotic disor- ences in comorbid disorders, re-
and satisfaction with health) at base- ders and comorbid panic attacks were sponse to treatment, and course of ill-
line and at follow-up. significantly more likely than those ness among persons with comorbid
who did not have panic attacks to panic attacks and psychosis—that co-
Results have major depressive disorder, spe- morbidity of panic attacks and psy-
Comorbidity of panic attacks cific phobia, sedative abuse, polysub- chosis reflects a different form or
Of the 120 participants who had psy- stance abuse, other substance abuse, subtype of disorder within the schizo-
chotic disorders, 18, or 15 percent, or anorexia nervosa (Table 2). Rates phrenia spectrum (23,24). These po-
also had panic attacks. Of these, nine of bipolar disorder, dysthymia, alco- tential pathways are not known and
had schizophrenia and nine had hol use disorder, cannabis abuse, may merit further investigation.
schizoaffective disorder. One partici- stimulant abuse, hallucinogen abuse, The mechanism through which
pant also met the criteria for brief re- social phobia, obsessive-compulsive panic attacks lead to poorer outcomes
active psychosis. disorder, and generalized anxiety dis- in quality of life among individuals
order were also higher among partic- who have psychotic disorders is un-
Sociodemographic characteristics ipants who had comorbid panic at- clear. The lower scores in several key
No significant differences were ob- tacks than among those who did not. quality-of-life domains among partic-
served in sociodemographic charac- However, these differences were not ipants in our study who had panic at-
teristics between participants with statistically significant. tacks are consistent with the associa-
schizophrenia, schizoaffective disor- tion between comorbid panic attacks
der, or other psychotic disorders, so Rehabiliation and quality of life and the related impairment in func-
the groups were combined. Compared with participants who did tioning that has frequently been de-
No significant differences were ob- not have comorbid panic attacks, scribed among persons with major
served in gender, race, marital status, those who did have panic attacks re- depression and anxiety disorders (25).
parenthood status, social class, or ported a poorer quality of life at base- It also is conceivable that panic at-
employment status between partici- line and at four and a half months’ fol- tacks interfere with the response to
pants who did and did not have pan- low-up on all Quality of Life sub- psychopharmacologic treatment or
ic attacks (Table 1). Participants who scales, as shown in Table 3. More with an individual’s ability to tolerate
had comorbid panic attacks were sig- careful inspection of the data re- various medications, resulting in
nificantly less likely to have graduat- vealed some interesting trends. Par- poorer outcomes. Alternatively, it
ed from high school than those who ticipants who had comorbid panic at- could be that poorer outcomes are
did not. tacks appeared to improve on several predicted by higher rates of comorbid
922 PSYCHIATRIC SERVICES ♦ July 2001 Vol. 52 No. 7
disorders among persons who have Table 3
panic attacks, although the reasons Mean±SD scores on health and quality-of-life measures for 120 patients who had
for these comorbidities remain un- psychotic disorders with and without panic attacks
known.
The differences we observed in Without panic With panic
well-being and quality of life between Item attacks (N=102) attacks (N=18) F† p
baseline and follow-up, most of which Well-being1
were not statistically significant, show Purpose
several potentially interesting trends Baseline 59.8±14.5 47.0±16.8
that may provide useful direction for 4.5 month follow-up 58.9±16.1 52.5±10.7
future study. Individuals who had co- Mastery
Baseline 56.6±15.6 47.8±9.99
morbid panic attacks, although they 4.5 month follow-up 57.4±13.5 54.3±10.5
had considerably lower scores at Growth
baseline and at follow-up, tended to Baseline 65.5±10.3 54.5±11.9 5.4 .027
show improvement from baseline to 4.5 month follow-up 64.9±12.5 56.5±10.9
follow-up—in some cases by several Depression2
Baseline 19.37±12.6 33.0±11.14
points—on several measures, such as 4.5 month follow-up 19.7±11.7 30.3±9.8 6.07 .019
mastery as an aspect of well-being on General health3
the CES-D scale. Baseline 56.2±14.4 70.8±18.5 4.25 .047
In contrast, the scores of individu- 4.5 month follow-up 55.9±12.8 64.3±19.0 4.66 .038
als who did not have panic attacks Functioning4
Baseline 49.2±14.1 44.2±8.9
changed very little; in some cases, the 4.5 month follow-up 45.5±10.5 47.2±10.9
scores worsened slightly. However, Satisfaction with school5
because the sample was small and Baseline 6.5±.8 3.7±0 12.1 .018
these changes may have reflected re- 4.5 month follow-up 6.3±.7 4.7±0 5.4 .069
gression to the mean, our findings Satisfaction with health5
Baseline 4.9±1.1 4.3±.96
should be interpreted with caution. 4.5 month follow-up 5.2±.05 4.6±.1
Notable, however, is the consistency
with which this pattern was seen 1 From the Quality of Life scale. Possible scores range from 0 to 100, with higher scores reflecting
throughout the quality-of-life scores better functioning.
2 From the Center for Epidemiologic Studies scale. Possible scores range from 0 to 60, with higher
and the diversity of domains in which scores reflecting higher rates of depressive symptoms.
well-being was investigated—for ex- 3 From the General Health Questionnaire. Possible scores range from 0 to 120, with higher scores

ample, depression and satisfaction reflecting better general health.

with the pursuit of goals such as at-
tending school—in an effort to reveal
this pattern in both psychiatric and
nonpsychiatric domains.
Interpretation of these data is
largely speculative, and further em-
pirical testing will be required. Limi-
tations of this study include the use of
a small sample of persons with co-
morbid panic attacks, which is likely
to have contributed to the absence of
statistical significance of many of the
observed differences in quality of life
between the two groups. For the
same reasons, the quality-of-life out-
comes were not corrected for multi-
ple comparisons, which is also a sig-
nificant limitation of the data. Our
findings may be generalizable only to
patients who receive community-
based outpatient psychiatric treat-
ment for psychosis.
The co-occurrence of panic attacks
and psychosis has been relatively neg-
lected as a target of study in clinical
settings, for several possible reasons.
First, clinicians may believe that pan-
4 From the Global Assessment of Functioning scale. Possible scores range from 0 to 100, with high-
er scores reflecting better functioning.
5 From the Quality of Life scale. Possible scores range from 1 to 7, with 1 representing “terrible”
and 7 representing “delighted.”
† df=1, 119
ic attacks do not occur among persons
who have schizophrenia and other
psychotic disorders. Second, this par-
ticular form of comorbidity may be
considered inconsequential in treat-
ment outcomes among patients who
have psychosis. Also, it is possible that
the use of hierarchical diagnostic cri-
teria obscures the prevalence of co-
morbid panic attacks among persons
who have schizophrenia and other
psychotic disorders in clinical and re-
search settings (23).
Conclusions
Our data suggest that panic attacks
may affect the course of illness, the
response to psychosocial interven-
tion, and quality of life among per-
sons with severe psychotic disorders,
suggesting that this diagnostic pattern
is clinically important. These findings
should alert clinicians to the potential
impact of comorbid panic attacks on
treatment outcomes and quality of
life among persons with psychotic dis-
orders, highlighting the importance
of diagnosing and treating both disor-
ders when they co-occur (2,26). Addi-
tional research is needed to illumi-
nate the sequence and timing of on-
sets, as well as the potential etiologic,
neurobiological, and psychosocial
mechanisms involved in these rela-
tionships. ♦
References
1. Kessler RC, Stang PE, Wittchen HU, et al:
Lifetime panic-depression comorbidity in
the National Comorbidity Survey. Archives
of General Psychiatry 55:801–808, 1998
2. Andrade L, Eaton WW, Chilcoat HD: Life-
time comorbidity of panic attacks and ma-

PSYCHIATRIC SERVICES ♦ July 2001 Vol. 52 No. 7 923

 jor depression in a population-based study: 17. Husaini BA, Neff JA, Harrington JB: De- 22. Goldberg D, Cooper B, Eastwood MR: A
age of onset. Psychological Medicine pression in rural communities: validating standardized psychiatric interview for use
26:991–996, 1996 the CES-D scale. Journal of Community in community surveys. British Journal of
Psychology 8:20–27, 1980 Preventive and Social Medicine 24:18–23,
3. Frank E, Shear MK, Rucci P, et al: Influ- 1970
ence of panic-agoraphobic spectrum symp- 18. Lehman AF: A quality of life interview for
toms on treatment response in patients the chronically mentally ill. Evaluation and 23. Bermanzohn PC, Porto L, Arlow PB, et al:
with recurrent major depression. American Program Planning 11:51–62, 1988 Hierarchical diagnosis in chronic schizo-
Journal of Psychiatry 157:1101–1107, 2001 phrenia: a clinical study of co-occurring
19. Lehman AF, Postrado LT, Rachuba LT: syndromes. Schizophrenia Bulletin 26:517–
4. Markowitz JS, Weissman MM, Ouellette R, Convergent validation of quality of life as- 525, 2000
et al: Quality of life in panic disorder. sessments for persons with severe mental
Archives of General Psychiatry 46:984–992, illness. Quality of Life Research 2:327–333, 24. Hofmann S: Relationship between panic
1989 1993 and schizophrenia. Depression and Anxiety
9:101–106, 1999
5. Kessler RC, McGonagle KA, Zhao S, et al: 20. Endicott J, Spitzer RL, Fleiss JL, et al: The
Lifetime and 12-month prevalence of Global Assessment Scale: a procedure for 25. Roy-Byrne PP, Stang P, Wittchen HU, et al:
DSM-III-R psychiatric disorders in the measuring the overall severity of psychi- Lifetime comorbidity of panic-depression
United States: results from the National atric disturbance. Archives of General Psy- in the National Comorbidity Survey: associ-
Comorbidity Survey. Archives of General chiatry 33:766–771, 1976 ation with symptoms, impairment, course,
Psychiatry 51:8–19, 1994 and help-seeking. British Journal of Psychi-
21. Bodlund O, Kullgren G, Ekselius L, et al: atry 176:229–235, 2000
6. Cassano GB, Pini S, Saettoni M, et al: Mul- Axis V–Global Assessment of Functioning
tiple anxiety disorder comorbidity in pa- Scale: evaluation of a self-report version. 26. Cutler J: Panic attacks and schizophrenia:
tients with mood spectrum disorders with Acta Psychiatrica Scandinavica 90:342–347, assessment and treatment. Psychiatric An-
psychotic features. American Journal of 1994 nals 24:473–476, 1994
Psychiatry 156:474–476, 1999
7. Cosoff SJ, Hafner RJ: The prevalence of co-
morbid anxiety in schizophrenia, schizoaf-
fective disorder, and bipolar disorder. Aus-
tralian and New Zealand Journal of Psychi-
atry 32:67–72, 1998
8. Cutler JL, Siris SG: “Panic-like” sympto-
motology in schizophrenic and schizoaffec-
tive patients with postpsychotic depression:
observations and implications. Compre-
hensive Psychiatry 32:465–473, 1991
9. Galynker I, Ieronimo C, Perez-Acquino A,
Journal Invites Annual Meeting Papers
et al: Panic attacks with psychotic features.
Journal of Clinical Psychiatry 57:402–406, Authors of material presented at the American Psychi-
1996
atric Association’s 2001 annual meeting, held May 5–10
10. Arlow PB, Moran ME, Bermanzohn PC, et in New Orleans, are invited to submit their papers for
al: Cognitive-behavioral treatment of panic
attacks in schizophrenia. Journal of Psy- possible publication in Psychiatric Services.
chotherapy Practice and Research
6:145–150, 1997 Material must be unpublished and not under review
for publication elsewhere. Papers must be in final man-
11. Klerman GL, Weissman MM, Ouellette R,
et al: Panic attacks in the community. uscript form, ready for blind peer review.
JAMA 265:742–746, 1991
Papers must conform to the length, authorship, and
12. Davidson L, Haglund KE, Stayner DA, et other requirements outlined in Psychiatric Services’
al: A qualitative study of supported social-
ization. Psychiatric Rehabilitation Journal Information for Contributors (see pages 613–614 of
24:275–292, 2001 the May 2001 issue or visit the journal’s Web site at
13. Davidson L, Stayner DA, Rakfeldt J, et al: http://psychservices.psychiatryonline.org). Papers should
Friendship and the restoration of commu- be mailed to John A. Talbott, M.D., Editor, Psychiatric
nity life: supported socialization for people Services, American Psychiatric Association, 1400 K
with psychiatric disabilities, in Social Sup-
ports and Psychiatric Rehabilitation. Edit- Street, N.W., Washington, D.C. 20005. For more infor-
ed by Stastny P, Campbell J. New York, Wi- mation, contact the editorial office: phone, 202-682-
ley, in press 6070; fax, 202-682-6189; e-mail, psjournal@psych.org.
14. Spitzer RL, Williams JB, Gibbon M, et al:
The Structured Clinical Interview for
DSM-III-R (SCID): I. history, rationale,
and description. Archives of General Psy-
chiatry 49:624–629, 1992
15. Williams JB, Gibbon M, First MB, et al:
The Structured Clinical Interview for
DSM-III-R (SCID): II. multisite test-retest
reliability. Archives of General Psychiatry
49:630–636, 1992
16. Radloff LS: The CES-D Scale: a self-report
depression scale for research in the general
population. Applied Psychological Mea-
sures 1:385–401, 1977

924 PSYCHIATRIC SERVICES ♦ July 2001 Vol. 52 No. 7