Comorbidity in panic disorder: II. Chronology of appearance and pathogenic comorbidity

285

Psychiatry Research, 46:285-293 Elsevier

Comorbidity in Panic Disorder: II. Chronology Appearance and Pathogenic Comorbidity Vladan Starcevic,

E.H. Uhlenhuth,

of

Robert Kellner, and Dorothy

Pathak

Received October 29, 1991; first revised version received March 31, 1992; second version received August 4, 1992; accepted November 20, 1992.

revised

Abstract. Ages of onset and the sequence of appearance of panic disorder (PD) and comorbid conditions were determined in a sample of 54 patients with the principal DSM-III-R diagnosis of PD. The onset of PD was earlier in patients with moderate to severe agoraphobia (AG) than in panic patients without AG. Patients with alcohol abuse and drug abuse before the onset of PD also had a tendency to develop PD earlier, which suggests that these conditions might have specifically predisposed to PD. All comorbid disorders, except for major depression, were more likely to precede the onset of PD so that, more often than not, PD appeared as a chronologically secondary condition. However, it was found that only for primary substance abuse such a temporal relationship might denote etiologic relatedness to PD, because of the reduced temporal distance between the onset of primary substance abuse and secondary PD. Key Words. Anxiety,

agoraphobia,

substance

abuse.

The typical ages of onset of most anxiety disorders have been fairly well established (Thyer et al., 1985; Ost, 1987), and simple phobia and social phobia were found to be significantly more likely to have an earlier onset when compared with the ages of onset of panic disorder (PD) and generalized anxiety disorder (GAD) (Thyer et al., 1985). It is not known, however, whether comorbidity of two or more anxiety or other disorders affects their ages of onset. The question of whether the presence of a particular anxiety or other disorder affects the age of onset of another disorder and increases the likelihood that it will appear is of relevance for etiology. The form of comorbidity that refers to an etiologic relationship between the two disorders has been termed pathogenic comorbidity by Kaplan and Feinstein (1974). The task here is to identify those chronologically primary disorders that also speciJi:cally predispose to the occurrence of a chronologically secondary disorder. Ideally, that can be studied by following a large number of patients with a particular disorder and determining how many of them develop other specific disorders. In the

Vladan Starcevic, M.D., is Research Fellow; E.H. Uhlenhuth,

M.D., is Professor of Psychiatry; Robert Kellner, M.D., Ph.D., is Vice-Chairman of the Department of Psychiatry and Professor of Psychiatry; and Dorothy Pathak, Ph.D., is Associate Professor of Biostatistics, Department of Family and Community Medicine, University of New Mexico School of Medicine, Albuquerque, NM. (Reprint requests to Dr. V. Starcevic, Dept. of Psychiatry, University of New Mexico School of Medicine, 2400 Tucker, N.E., Albuquerque, NM 87131, USA.) 016%1781/93/$06.00

@ 1993 Elsevier Scientific

Publishers

Ireland

Ltd.

286 absence of such complex studies, another way to approach the problem is to study whether the presence of a chronologically primary condition is associated with an earlier onset of a secondary condition. If it is, the primary condition may be conceptualized as a predisposition to the secondary, because the temporal distance between the two etiologically related conditions can be assumed to be shorter than the temporal distance between the etiologically unrelated conditions. Certainly, etiologic relatedness and predisposition do not imply direct causation; they refer to a greater likelihood of developing another disorder relatively soon after the onset of the first one. In other words, the presence of one illness accelerates the appearance of another (e.g., the onset of atherosclerotic processes is earlier in diabetics than in individuals without diabetes). In the area of anxiety disorders, it would be important to investigate whether the presence of disorders that usually begin before the onset of PD, such as simple phobia and social phobia, makes an earlier onset of PD more likely. Such studies have not been carried out. There have been reports, though, that patients with a history of childhood separation anxiety and school phobia have an earlier onset of panic attacks and agoraphobia (AG) (Breier et al., 1986; Perugi et al., 1988). An earlier age of onset of AG with panic attacks was also found among patients with high levels of neuroticism (Kenardy et al., 1990). The first aim of the present study was to examine, on the basis of the retrospectively obtained ages of onset, the sequence of appearance of comorbid disorders in relation to the onset of PD. Our second aim was to investigate in an indirect manner pathogenic comorbidity among patients with a principal diagnosis of PD-by studying possible effects of the antecedent comorbid disorders on the age of onset of PD. The information obtained from studying the latter might shed more light on the specific forerunners of PD. On the basis of clinical observation and experience, we hypothesized that the onset of PD was earlier in patients with pathological anxiety, mainly in the form of simple phobia or social phobia, preceding the onset of PD. If correct, this might suggest that simple phobia or social phobia specifically predisposed to PD or constituted a risk factor for developing PD. As in our previous report on the comorbidity in PD and AG (Starcevic et al., 1992), the term comorbidity refers to the current or lifetime occurrence of two or more disorders in the same individual. Aprincipal disorder is defined as the one that is judged to interfere the most with the patient’s overall functioning and/or that prompted the patient to seek treatment. The principal disorder in all patients in this study was PD. A comorbid disorder may precede or follow the principal disorder, or its onset may be contemporaneous (within the same year) with that of the principal disorder. Thus, depending on the temporal relationship between them, the comorbid disorder may be chronologically primary or secondary to, or contemporaneous in onset with, the principal disorder.

Methods The demographic characteristics of patients in our sample and details of the method have been described elsewhere (Starcevic et al., 1992). A summary of the methods and a description of

287 the subjects follows: Subjects were all screened for possible participation in a drug treatment study of PD. They were recruited through newspaper advertisements, referrals from other physicians, flyers, and word of mouth. The inclusion criterion for participation was presence of DSM-III-R PD, with or without AG, in individuals between the ages of 18 and 61 (American Psychiatric Association, 1987). Because of the diagnostic hierarchical considerations, patients were excluded from the study if they had a lifetime history of schizophrenia, bipolar disorder, delusional disorder, or organic mental syndrome. Moreover, our sample did not include PD patients with a clinically significant depression or PD patients who were actively abusing alcohol or other psychoactive substance at the time of the assessment. Only those subjects whose PD was considered to be their principal diagnosis were included in the study. A total of 54 panic patients participated, 30 (55.6%) of whom were women and 24 (44.4%) of whom were men. The average age for women was 36.03 years (SD = 8.81), and for men it was 34.75 years (SD = 8.45). Nineteen (35.2%) patients had an uncomplicated PD (without AG), 13 (24.1%) patients met criteria for PD with mild AG, and 22 (40.7%) were diagnosed with PD with moderate and severe AG. The assessment procedures consisted of two structured interviews and several self-report measures. The instrument used in this study was the Structured Clinical Interview for DSMIII-R, Upjohn Version Revised (SCID-UP-R; Spitzer and Williams, 1988). The SCID-UP-R was developed for the purpose of diagnosing all DSM-III-R anxiety disorders (except for posttraumatic stress disorder), all mood disorders, and psychoactive substance abuse. The instrument also screens for past or present psychotic disorders. The interviews were conducted by two authors (V.S. and E.H.U.) at the same time. The interviewers were trained in the use of the SCID-UP-R, and one of them (E.H.U.) was unaware of the study hypothesis. The interrater reliability figures, presented elsewhere (Starcevic et al., 1992) showed good to excellent agreement. The ages of onset of the diagnosed disorders were defined as the ages when, judged retrospectively, all diagnostic criteria had been met. When the onset of any disorder was presumed to be before the age of 10, further evidence of such an early onset was sought (e.g., on the basis of the information provided by the family members). No attempt was made to establish chronological primacy when PD and a comorbid condition started in the same year (at the same age); in such cases, the onset of both disorders was considered to be contemporaneous. For comparisons of the mean ages of onset of PD in the three subtypes of PD, we used analysis of variance (ANOVA), followed by the Newman-Keuls multiple comparisons test. Comparisons of categorical variables (percentages of patients who developed PD in various age groups) were performed by means of ax* test (Zar, 1984). Results at or below the 5% level of probability (p < 0.05) were regarded as significant. All statistical analyses were carried out using standard procedures in the SAS statistical package for personal computers (SAS, 1985).

Results As reported previously (Starcevic et al., 1992), the comorbidity rate in the whole sample of PD patients was very high (83.3%): 45 out of 54 patients had at least one comorbid diagnosis. There was no significant difference in the comorbidity rates between the female and male patients (80% vs. 87.5%), and therefore, possible effects of gender were not studied further. Table 1 presents the ages of onset of PD, its subtypes, and comorbid disorders in chronological order. The onset of obsessive-compulsive disorder (OCD), simple phobia, and social phobia was usually in the early teens, and the onset of substance abuse was most common in the late teens and early twenties. GAD also had a tendency to start in the early twenties, while both mood disorders (major depression

288 and dysthymia) were most likely to begin in the mid-twenties. The mean ages of onset of PD and AG in the whole sample were in the late twenties. However, there was a statistically significant difference in the mean ages of onset of PD between panic patients without AG (34.7 years) and panic patients with moderate and severe AG (24.4 years). Table 2 shows the sequence of appearance of comorbid disorders relative to the onset of PD. All disorders except for major depression were more likely to precede the onset of PD. The first episode of major depression appeared before the onset of

Table 1. Ages of onset (in years) of panic disorder, agoraphobia, comorbid disorders, and subtypes of panic disorder Obsessive-compulsive Simple Social

disorder

phobia phobia

Drug abuse Alcohol

abuse

Generalized

anxiety

disorder

Dysthymia Agoraphobia’ First major depressive

episode

Panic disorder

n

Mean

SD

Median

6

13.3

6.0

12.5

6-23

24

14.6

10.0

12

5-40

22

14.9

8.9

13

6-44

17

17.3

4.1

17

12-26

19

15-33

Range

20

20.9

5.3

28

22.5

10.3

7

26.0

6.9

26

39

26.4

11.2

28

5-53

26

27.2

8.9

26

12-44

54

28.9

11.9

28

5-53

n

Mean2

SD

Median

Range

20.5

5-52 17-35

Panic disorder with moderate and severe

agoraphobia

22

24.4(a)

11.6

25.5

5-53

13

28.2(a,b)

10.6

29

7-44

19

34.7(b)

11.1

34

15-54

Panic disorder with mild agoraphobia Panic disorder without agoraphobia

1. Includes 4 patients without current agoraphobia, but with a history of past agoraphobia. 2. Means with different letters in parentheses are significantly different (analysis of variance + Newman-Keuls comparisons test: df = 51;p < 0.05).

multiple

Table 2. Sequence of appearance of disorders comorbid with panic disorder Preceded the onset of panic disorder

n

Contemporaneous with the onset of panic disorder

n

%

n

%

Followed the onset of panic disorder

n

%

6

5

83.3

0

0.0

1

16.7

Simple phobia

24

20

83.3

0

0.0

4

16.7

Drug abuse

17

14

82.3

1

5.9

2

11.8

Alcohol abuse

20

15

75.0

0

0.0

5

25.0 13.6

Obsessive-compulsive

disorder

Social phobia Generalized

anxiety disorder

Dysthymia First maior deoressive

eoisode

22

15

68.2

4

18.2

3

28

19

67.8

5

17.9

4

14.3

7

4

57.1

1

14.3

2

28.6

26

11

42.3

4

15.4

11

42.3

289

PD as often as after its onset. Only eight patients (14.8%) had PD as an unequivocal primary condition. In another seven patients (13%), PD appeared as a primary disorder contemporaneously with AG (6 cases), GAD (2 cases), social phobia (1 case), and drug abuse (1 case). For the remaining 39 patients (72.2%) PD was clearly a secondary disorder. As for AG, there was no single patient in whom AG appeared as an unequivocal (only) primary condition. Table 3 presents age of onset of PD in relation to the sequence of appearance of comorbid disorders. The age of onset of PD in patients with specific comorbid conditions preceding the onset of PD is compared with the age of onset of PD in patients without the same comorbid disorders, and in patients with the same comorbid disorders appearing after the onset of PD. Analyses did not include those patients who developed PD and a comorbid condition contemporaneously, because in such cases the primary-secondary dichotomy could not be used. A significantly earlier age of onset of PD was not found in patients with antecedent simple phobia, social phobia, GAD, and major depression. On the other hand, patients with preexisting drug abuse and alcohol abuse were significantly more likely to develop PD earlier-throughout the third decade of life.

Table 3. A e of onset of panic disorder and chronology of appearance of comorbi 8 disorders

Age of

Comorbid

disorder

Simple phobia

Social phobia

Drug abuse

Alcohol abuse

Generalized anxiety disorder

Major depression

Comorbid disorder preceded onset of panic disorder

Absence of a comorbid disorder or comorbid disorder followed onset of panic disorder

onset (in years) of oanic disorder

n

%

n

%

X2

df

P

<20 20-29 2 30

4 7 9

20.0 35.0 45.0

6 14 14

17.6 41.2 41.2

0.20

2

NS

< 20 20-29 230

2 9 4

13.3 60.0 26.7

7 11 17

20.0 31.4 48.6

3.60

2

NS

< 20 20-29 230

0 11 3

0.0 78.6 21.4

9 10 20

23.1 25.6 51.3

12.63

< 20 20-29 2 30

0 10 5

0.0 66.7 33.3

10 11 18

25.6 28.2 46.2

8.39

2

0.02

<20 20-29 2 30

1 10 8

5.3 52.6 42.1

8 9 13

26.7 30.0 43.3

4.44

2

NS

<20 20-29 2 30

0 7 4

0.0 63.6 36.4

9 13 17

23.1 33.3 43.6

4.62

2

NS

2 0.002

290

Discussion The ages of onset found in this study are similar to the ages of onset of the same disorders, reported in various clinical and epidemiologic studies (Thyer et al., 1985; Breier et al., 1986; Ost, 1987; Robins et al., 1991). Since the groups of patients with comorbid diagnoses represented mutually overlapping subsets of the main sample of PD patients, analyses of the statistical significance of the differences between the ages of onset of comorbid disorders could not be performed. An earlier age of onset of PD in patients with moderate and severe AG remains to be explained. We could not selectively examine potentially specific risk factors for PD with AG, which might have accounted for an early onset of PD. This finding is not a consequence of AG, because in most patients AG followed the onset of PD. It is possible that early onset of PD in these patients was one of the factors that contributed to the development of AG, irrespective of the duration of PD, which was not significantly different in patients with and without AG at the time of the assessment. A significant difference between the ages of onset of an uncomplicated PD and PD with moderate and severe AG is another indicator that these subtypes of PD may represent essentially different conditions (Starcevic et al., 1992). Although the temporal relationships vary, our findings suggest that in most patients, PD follows the onset of other anxiety and substance abuse disorders. The same pattern seems to be even more characteristic of AG, indicating that in addition to panic attacks, other forms of psychopathology may play an important role in the development of AG. The truncated nature of our study precludes us from concluding that PD may be a culmination of or the last stage in the development of pathological anxiety. In other words, our cross-sectional design permitted us to study the diagnostic antecedents of PD better than the consequences of PD, especially in those patients who were younger or whose PD was of a relatively short duration at the time of the assessment. We sought to identify disorders whose presence before the onset of PD may play a role in the etiology of PD. Contrary to our expectations, presence of simple and social phobia did not significantly affect the age of onset of the subsequently appearing PD. This finding does not suggest that simple and social phobia might constitute specific risk factors for developing PD, although these disorders were more likely to precede PD, often by many years. Thus, the comorbidity of various anxiety disorders may denote nonspecific links rather than a specific etiologic relatedness. Furthermore, these conditions may all arise from common vulnerability at the genetic and/ or developmental level. An example of such an explanatory model for anxiety disorders is “behavioral inhibition,” postulated by Rosenbaum et al. (1988). In case of simple phobia, failure to demonstrate pathogenic comorbidity with PD may in part be a result of treating simple phobia as a single homogeneous entity: Himle et al. (1991) have recently suggested that the “situational” subtype of simple phobia (phobias of driving and flying, and claustrophobia) may predispose to AG. The relationship between social phobia and PD warrants further study, in view of their similarities and overlapping features. Although the results of this study do not suggest that social phobia may be a specific “precursor” of PD, the possibility of

291 etiologic relatedness between these two conditions remains open, particularly in panic patients with AG. The studies of the temporal relationship between GAD and PD have not produced uniform results. In at least two studies, GAD was more likely to precede PD, although the figures varied substantially: this pattern was found in 65% and 100% of panic patients in the studies by Fava et al. (1988) and Cloninger et al. (1981), respectively. In the present study, GAD preceded PD in nearly 68% of panic patients. Breier et al. (1986) found that in 60% of patients with both GAD and PD with AG, the onset of GAD occurred approximately at the same time as the onset of PD with AG. In view of the relatively short intervals between the onset of GAD and the onset of PD, GAD may in some cases be conceived of as a prodrome of PD, or perhaps as a different, early mode of presentation of PD, as suggested by Garvey et al. (1988). Our findings do not support the notion that GAD specifically predisposes to PD. The data on temporal relationships between major depression and PD also vary substantially, and there does not seem to be a consistent pattern. In some studies, depression was found to be more likely to precede PD (Breier et al., 1984; de Ruiter et al., 1989; Sanderson et al., 1990). Other investigators reported the opposite trend: for example, Thompson et al. (1989) found that anxiety disorders, especially AG, usually preceded the onset of depression. The latter finding is in accordance with the hypothesis that depression is a result of the AG-induced handicap and demoralization. The prevalence of major depression secondary to PD cannot be estimated adequately, because most studies were performed on truncated samples (i.e., in patients who have not passed through an entire age-based period of risk for developing depression). Our data on the sequence of appearance of major depression and PD do not indicate that either of these disorders tends to have chronological primacy. Similar findings have also been reported by Stein et al. (1989, 1990). It is of interest to note, though, that consistent with the demoralization hypothesis, our patients with secondary major depression tended to have an early onset of PD and, in particular, an early onset of AG. The most interesting finding of this study is an earlier onset of PD in patients with preexisting drug and alcohol abuse. It would be erroneous to conclude that this finding implies a direct etiologic link between psychoactive substance abuse and PD, as there are pathways to PD other than through drug and alcohol abuse, and not all patients with substance abuse go on to develop PD. Nevertheless, it does appear that the risk for developing PD is increased in individuals with drug and alcohol abuse. Such a conclusion is supported by studies that suggest that repeated alcohol withdrawals or other consequences of alcoholism may predispose to panic attacks in vulnerable individuals (Cox et al., 1989; George et al., 1990). Moreover, a number of drugs, most notably cocaine, have been implicated in the precipitation of panic attacks (Cox et al., 1990). Primary substance abuse may increase susceptibility to develop PD on a long-term basis, so that PD would appear even after the symptoms of substance abuse are no longer present. The examination of temporal relationships between alcohol abuse and specific anxiety disorders in the present study produced results that could be expected on the

292

basis of previous research, as summarized by Kushner et al. (1990). Thus, all 10 of our patients with multiple comorbidity of PD, social phobia, and alcohol abuse had an onset of social phobia before the onset of excessive drinking. Alcohol abuse was also more likely to follow the onset of simple phobia (in 5 of 7 patients with PD, simple phobia and alcohol abuse), whereas in cases of comorbidity of PD, GAD, and alcohol abuse, alcohol abuse preceded the onset of GAD (5 cases) almost as often as it followed the onset of GAD (6 cases). Finally, alcohol abuse preceded the onset of PD in 75% of our patients with comorbid alcohol abuse. Except for one case in whom the onset of PD occurred at the same age as the onset of marijuana abuse, the onset of PD in our patients was temporally separate from the onset of drug abuse. In most patients, drug abuse preceded the onset of PD: this was the temporal pattern for 11 of 14 patients with marijuana abuse, for six of seven patients with cocaine abuse, for all four patients with amphetamine abuse, for one of two patients with LSD abuse, and for the single patient with heroin abuse. of the Study. The main limitation of this study, shared with studies of a similar nature, is its retrospective method, which may entail errors in patients’ recollections of the ages of onset of various disorders, particularly if they began in childhood. The cross-sectional design of the study precluded adequate examination of the disorders that appear after the onset of PD or as a consequence of PD. Moreover, studies of the temporal relationships between PD and comorbid disorders should include additional samples of patients with principal diagnoses of other anxiety disorders, major depression, and substance abuse. Prospective studies in clinical populations should ideally be initiated in children with overt manifestations of pathological anxiety and in those who are considered to be at risk for developing anxiety disorders (e.g., children of individuals with PD). Future research in this area should be carried out in PD subjects drawn from the community. This would provide a test of the applicability of our findings to nonclinical settings. Perhaps the most interesting aspect of such research would involve comparisons between the rates of PD and ages of onset of PD in the two distinct groups: (1) individuals with primary substance abuse and (2) individuals with a primary diagnosis of one of the other anxiety disorders. Limitations

Acknowledgment.

This work was supported

by the Upjohn

Company,

Kalamazoo,

MI.

References Psychiatric Association. DSM-III-R: Diagnostic and Statistical Manual of Mental 3rd ed., revised, Washington, DC: American Psychiatric Press, 1987. Breier, A.; Charney, D.S.; and Heninger, G.R. Major depression in patients with agoraphobia and panic disorder. Archives of General Psychiatry, 41:1129-1135, 1984. Breier, A.; Charney, D.S.; and Heninger, G.R. Agoraphobia with panic attacks: Development, diagnostic stability, and course of illness. Archives of General Psychiatry, American

Disorders.

43:1029-1036,

1986.

Cloninger, C.R.; Martin, R.L.; Guze, S.B.; and Clayton, P.J. A blind follow-up and family study of anxiety neurosis: Preliminary analysis of the St. Louis 500. In: Klein, D.F., and Rabkin, J., eds. Anxiety: New Research and Changing Concepts. New York: Raven Press, 1981. pp. 137-150.

293 Cox, B.J.; Norton, G.R.; Dorward, J.; and Fergusson, P.A. The relationship between panic attacks and chemical dependencies. Addictive Behaviors, 1453-60, 1989. Cox, B.J.; Norton, G.R.; Swinson, R.P.; and Endler, N.S. Substance abuse and panicrelated anxiety: A critical review. Behaviour Research and Therapy, 28:385-393, 1990. de Ruiter, C.; Rijken, H.; Garssen, B.; van Schaik, A.; and Kraaimaat, F. Comorbidity among the anxiety disorders. Journal of Anxiety Disorders, 3:57-68, 1989. Fava, G.A.; Grandi, S.; and Canestrari, R. Prodromal symptoms in panic disorder with agoraphobia. American Journal of Psychiatry, 145: 1564-1567, 1988. Garvey, M.J.; Cook, B.; and Noyes, R. The occurrence of a prodrome of generalized anxiety in panic disorder. Comprehensive Psychiatry, 291445-449, 1988. George, D.T.; Nutt, D.J.; Dwyer, B.A.; and Linnoila, M. Alcoholism and panic disorder: Is the comorbidity more than coincidence? Acta Psychiatrica Scandinavica, 81:97-107, 1990. Himle, J.A.; Crystal, D.; Curtis, G.C.; and Fluent, T.E. Mode of onset of simple phobia subtypes: Further evidence of heterogeneity. Psychiatry Research, 36:37-43, 199 1. Kaplan, M.H., and Feinstein, A.R. The importance of classifying initial comorbidity in evaluating the outcome of diabetes mellitus. Journal of Chronic Diseases, 27:387-404, 1974. Kenardy, J.; Oei, T.P.S.; and Evans, L. Neuroticism and age of onset for agoraphobia with panic attacks. Journal of Behavior Therapy and Experimental Psychiatry, 21:193-197, 1990. Kushner, M.G.; Sher, K.J.; and Beitman, B.D. The relation between alcohol problems and the anxiety disorders. American Journal of Psychiatry, 147:685-695, 1990. Ost, L.-G. Age of onset in different phobias. Journal of Abnormal Psychology, 96:223-229, 1987. Perugi, G.; Deltito, J.; Soriani, A.; Musetti, L.; Petracca, A.; Nisita, C.; Maremmani, I.; and Cassano, G.B. Relationships between panic disorder and separation anxiety with school phobia. Comprehensive Psychiatry, 29:98-107, 1988. Robins, L.N.; Locke, B.Z.; and Regier, D.A. An overview of psychiatric disorders in America. In: Robins, L.N., and Regier, D.A., eds. Psychiatric Disorders in America: The Epidemiologic Catchment Area Study. New York: Free Press, 1991. pp. 328-366. Rosenbaum, J.F.; Biederman, J.; Gersten, M.; Hirshfeld, D.R.; Meminger, S.R.; Herman, J.B.; Kagan, J.; Reznick, J.S.; and Snidman, N. Behavioral inhibition in children of parents with panic disorder and agoraphobia. Archives of General Psychiatry, 451463-470, 1988. SAS. SAS User’s Guide Statistics. Cary, NC: SAS Institute, Inc., 1985. Sanderson, W.C.; Beck, A.T.; and Beck, J. Syndrome comorbidity in patients with major depression or dysthymia: Prevalence and temporal relationships. American Journal of Psychiatry, 147: 1025-1028, 1990. Spitzer, R. L., and Williams, J.B.W. Structured Clinical Interview for DSM-III-R, Upjohn Version Revised (SCZD-UP-R). New York: New York State Psychiatric Institute, 1988. Starcevic, V.; Uhlenhuth, E.H.; Kellner, R.; and Pathak, D. Patterns of comorbidity in panic disorder and agoraphobia. Psychiatry Research, 42: 17 I- 183, 1992. Stein, M.B.; Shea, C.A.; and Uhde, T.W. Social phobic symptoms in patients with panic disorder: Practical and theoretical implications. American Journalof Psychiatry, 146:235-238, 1989. Stein, M.B.; Tancer, M.E.; and Uhde, T.W. Mcrjor depression in patients with panic disorder: Factors associated with course and recurrence. Journal of Affective Disorders, 19:287-296, 1990. Thompson, A.H.; Bland, R.C.; and Orn, H.T. Relationship and chronology of depression, agoraphobia, and panic disorder in the general population. Journal of Nervous and Mental Disease, 177:456-463, 1989. Thyer, B.A.; Parrish, R.T.; Curtis, G.C.; Nesse, R.M.; and Cameron, O.G. Ages of onset of DSM-III anxiety disorders. Comprehensive Psychiatry, 26: 113-122, 1985. Zar, J.H. Biostatistical Analysis. 2nd ed. Englewood Cliffs, NJ: Prentice-Hall, 1984.