One year prevalence of panic disorder (PD) is 2.1% and lifetime prevalence is 5.1% in United States of America. One year and lifetime prevalence of agoraphobia are 0.6% and 1.1%, respectively (1). Panic attack is described as abrupt beginning of 4 of 13 symptoms listed in DSM-IV-TR and a period of intense fear and distress (2). PD is a variable disorder in terms of physical, physiological and cognitive symptoms of panic attack. It has been suggested that, since there are panic attacks consisting of very different symptom clusters, there may be subtypes of this disorder and that failure to distinguish these subtypes may lead to theoretical, methodical and treatment errors (3). For these reasons, subtyping studies based on factor analysis of symptoms during panic attacks have been conducted (4-8).
Several features have been taken into account in order to detect subtypes of PD. Presence of agoraphobia (9), nocturnal panic attacks (10), prominent symptoms during panic attacks (11,12), timing of panic attacks (13,14), biological features of the disorder (15), genetic liability (16) are among these features.
PD has been subtyped into different groups, regarding prominent panic attack symptoms as PD with heart and pulmonary symptoms (17), vestibular symptoms (18), gastrointestinal symptoms (19), depersonalization/derealization (20) or without subjective sense of fear (21).
Starcevic et al. (9) divided panic attack symptoms as more important (first rank) and those with secondary importance (second rank). First rank symptoms include; palpitations, tachycardia, dyspnea, dizziness or feeling faintness and trembling. Second rank symptoms are choking, shortness of breath, chest pain, numbness, hot flashes, depersonalization/derealization and nausea. Studies in this field have reported that the broadest part of the variance is explained by cardiorespiratory and vestibular factors (8,22-24).
In the literature, other groups detected with factor analysis are cognitive symptoms, depersonalization, numbness and autonomic activation. It has been argued that this clinical heterogeneity may have important implications on etiology, course and treatment of the disorder (3).
In the literature, when compared to other psychiatric disorders, there are fewer studies which investigated quantitative evaluation of PD psychopathology. Segui et al. (5) using factor analysis, found four factors among panic attack symptoms; cardiorespiratory, vestibular, mixed symptoms and general arousal. The authors also reported that factor analysis is a very valuable method to study subtyping and that more studies are necessary on this issue.
In our study, our aim was to detect subtypes of PD, based on factor analysis, in our country.
159 consecutive patients who applied to Bakirkoy Training and Research Hospital for Psychiatry, Neurology and Neurosurgery secondary care outpatient clinics and who volunteered to participate to the study were included. Thirty-two patients did not accept to participate in the study. Seventeen patients were excluded since they did not fulfill inclusion criteria. Inclusion criteria were; at least having panic attack during the last month, being 18-64 years of age, PD disorder per DSM IV-TR diagnostic criteria (2) and having necessary cognitive abilities to fill self-report forms in the study. Exclusion criteria was having PD diagnosis associated with general medical condition or direct physiological effects of alcohol/substance per DSM IV-TR diagnostic criteria.
PD diagnosis were validated and comorbid conditions, particularly agoraphobia, were evaluated by the researchers by using Structural Clinical Interview for DSM IV Axis I Disorders (SCID-I) since original diagnosis of the patients were made by clinicians who were not included in the study. Tests in the test battery were administered in a random order during a single session. Ethical board approval from Bakirkoy Training and Research Hospital for Psychiatry, Neurology and Neurosurgery instutional review board before the study. Sociodemographical data of the study are formerly used in another study, “Panic disorder and comorbidity” (25).
Sociodemographical Data Form: The form is developed by the researchers, by adding factors which can influence etiology and course of PD to sociodemographical data form provided in SCID-I interview manual. Besides, we detected symptoms experienced by the patients during their latest panic attacks, by adding symptoms which are thought to be features of panic attack and which have been frequently stated by our patients during our clinical practice, such as “fear of paralysis” (5) and the symptom, “restlessness, need to escape”, which was suggested to be included in DSM diagnostic criteria by Cox et al. (26).
Structural Clinical Interview for DSM IV Axis I Disorders (SCID-I): It is a structured clinical interview to evaluate Axis I disorders by First et al. (27). SCID-I was adapted to Turkish by Corapcioglu et al. (28). SCID-I detects lifetime presence of axis one disorders.
Panic Agoraphobia Scale (PAS): It is developed by Bandelow et al. (29). There are clinical observer and patient self-report forms. It can be applied to PD patients regardless of the presence of agoraphobia. There are five subscales.
A- Features of panic attack: Three items are rated and one item is unrated,
B-Agoraphobia and avoidance behavior: Three items,
C-Anticipatory anxiety: Two items,
D-Disability: Three items,
E- Worries about health: Two items.
Each subscale score is computed by adding item scores in that particular subscale and total score is obtained by summing all items. It is adapted to Turkish by Tural et al. (30).
Hamilton Rating Scale for Depression (HAM-D): It is a 17-items scale, developed by Williams et al. (31) to evaluate the level and change of severity of depression. It is adapted to Turkish by Akdemir et al. (32).
Beck Anxiety Inventory (BAI): It is a 17-items scale, developed by Beck et al. (33) to measure the severity of anxiety experienced by the individual. Higher scores correspond to increased levels of anxiety. The scale is adapted to Turkish by Ulusoy et al. (34).
Principal Component Analysis were computed on the variables including 15 items; 13 items in DSM IV-TR panic attack diagnostic criteria and 2 additional items which were reflecting the experiences of the patients in their last panic attack. Common factors were computed by using ortogonal factor model on the detected principal components. After this, factor loadings and weight of each item on the factor were computed by applying Varimax rotation. Factors with eigenvalues higher than 1 are included in the analysis. Pearson chi-square test is used to compare categorical variables. p<0.05 was accepted as statistically significant. Statistical analysis were computed with SPSS for Windows 11.0 software.
Hundred twenty of the participants were women (75.5%), and 39 were men (24.5%). Mean age of the participants was 39.16±11.24 years. Sociodemographical features of the patients were summarized in Table 1 and scale scores were summarized in Table 2. Symptoms of the last panic attack experienced by the participants were shown in Table 3. Principal Components Analysis revealed 6 factors with eigenvalues larger than 1, which explained 55.94% of the variance, cumulatively (Table 4).
First factor (variance=15%) consisted of symptom cluster hot flushes or chills, sweating, loose of control and nausea or stomach ache, these last two were distributed to first, fifth and sixth factors by borderline values. Second factor (variance=9.38%) included dizziness, lightheadness feelling faint, trembling or shaking, and depersonelization/derealization symptom cluster. Third factor (variance=8.89%) included palpitations, tachycardia or perception of heartbeats, chest pain and discomfort symptom cluster. Fourth factor (variance=7.95%) consisted of numbness/tingling, fear of paralysis, and dyspnea and sensation of being smothered, which were mainly loaded in the fifth factor. Fifth factor (variance=7.5%) included dyspnea or sensation of being smothered, choking and fear of loosing control or going insane, which were mainly loaded in the first factor, and nausea or stomach pain, which were distributed to three factors. Sixth factor (variance=7.1%) consisted of fear of dying, need to escape, restlessness and nausea or stomach ache symptoms (Table 4).
In their study on phobic avoidance, Turgeon et al. (35) found a significant difference between women and men only in phobic avoidance and there were no significant gender differences in the other components of agoraphobia. In our study, while there was a gender difference in the rate of agoraphobia, this was not statistically significant.
When compared with other psychiatric disorders, subtyping studies, based on symptom profile, by using factor analysis is less common in PD. Using DSM-IV-TR panic attack diagnostic criteria in their phenomenological study, Segui et al. (5) found 4 groups which explained 57% of the total variance. The biggest group explained 26.1% of the total variance, cardiorespiratory symptoms factor, and included chest pain, palpitation, numbness, dyspnea and shaking, trembling and fear of dying. The second factor explained 15.1% of the total variance, vestibular symptoms factor, and consisted of dizziness, faintness, fear of loosing control and going insane. The third factor explained 8.5% of the total variance, mixed symptoms factor, and included symptoms such as sweating, depersonalization/derealization, sensation of smothering and hot flushes. The fourth and last factor explained 7.2% of the total variance, general arousal symptoms factor, and included gastrointestinal symptoms (stomachache, nausea, vomiting), shaking, trembling, hot flushes and numbness. Factor with these high variance rates suggested that PD was not a homogeneous group and might involve subtypes.
In a similar study, Briggs et al. (7) used Principal Components Analysis and found five factor, which explained 49% of the total variance. Among these factors, two factors in which respiratory symptoms were dominant or not were significant. There were no significant differences between these groups in terms of sociodemographical variables, only symptom profiles were significantly different. Phenomenological and treatment comparisons revealed that spontaneous panic attacks were more common in patients with respiratory complaints and that this group responded to imipramine better. Situational panic attacks were more common in the group without respiratory complaints and they responded better to alprozolam. Authors claimed that symptom based subtyping might predict treatment response.
Shioiri et al. (36) added “agoraphobia” and “anticipatory anxiety” to the 13 DSM-III-R panic attack criteria and studied these 15 items in 247 patients. They compared panic attacks with panic attacks with restricted symptoms in cluster analysis. There were 3 clusters in the panic attack group and 4 clusters in the restricted symptom attack group. They found that there were differences between the groups in terms of symptom distribution and frequency, thus, attacks with restricted symptom represent a different group than panic attacks.
Cox et al. (26) added “hopelessness” and “restlessness, need to escape” to the 13 items included in DSM-III-R panic attack criteria and studied these 15 items in 212 patients. Principal Components Analysis revealed three factors. First factor included mainly symptoms associated with dizziness, second factor included mainly cardiorespiratory symptoms and the third factor included mainly cognitive symptoms. Authors stated that PD was a heterogenous disorder and that DSM diagnostic classification was limited.
In our study, we conducted Principal Components Analysis with the 15 variables, that was formed up by adding two items which have been frequently stated by our patients during our clinical practice, (5) to DSM-IV-TR panic attack diagnostic criteria; “fear of paralysis”, which is thought to be a feature of panic attack and and the“restlessness, need to escape” symptom which was suggested to be included in DSM diagnostic criteria by Cox et al. (26). Six factors explained 55.94% of the total variance. These factors were labeled as autonomic hyperactivity, vestibular, cardiovascular, pseudoneurological, respiratory and fear of dying groups, based on the loaded items.
First group, labeled autonomic hyperactivity consisted of symptom cluster hot flushes or chills, sweating, loose of control and nausea or stomach ache, which were closer to lower limit. Second group which was labeled as vestibular, included dizziness, lightheadness feelling faint, trembling or shaking, and depersonelization/derealization symptom cluster. Third group was labeled as cardiovascular group and included palpitations, tachycardia or perception of heartbeats, chest pain and discomfort symptom cluster. Fourth group consisted of pseudoneurological symptoms, such as numbness/tingling, fear of paralysis, and dyspnea and sensation of being smothered. We belive that numbness triggered fear of paralysis and formed this factor. Fifth group (variance=7.5%) included dyspnea or sensation of being smothered, choking and fear of loosing control, nausea or stomach pain, which were closer to lower limit. This group was labeled as respiratory group. The sixth and last group consisted of fear of dying, need to escape, restlessness and associated borderline symptoms of nausea or stomach ache. This group was labeled as fear of dying group since the main loading was on this fear.
There were some limitations of the present study. Having a medical disorder comorbid with PD may have an effect on panic symptoms. Although having PD due to general medical conditions was an exclusion criteria in our study, we did not specifically investigate medical disorders which could be associated with PD. We did not evaluate axis II comorbidities by SCID-II. Therefore, we did not evaluate whether patients had pure panic disorder. Since this was a retrsopective study, other limitations included exaggeration and difficulties in remembering the symptoms.
Results of our study indicated that there were very significant phenomenological subgroups among panic attack symptoms, which were based on Principal Components Analysis, a very important method to make this evaluation. The subgroups in our country seemed to be consistent with subgroups reported from other studies. Detecting panic disorder subtypes may help to predict treatment variablity, treatment response and selection of treatment in resistant cases. Subtyping via investigation of symptom structure variability may also set an example for other anxiety disorders.
Limitation of the study is lack of examining the possible effects of DSM-IV TR axis II disorders, since these disorders were not excluded. New studies which assess axis II disorders may yield important results.
In conclusion, results indicated that PD symptoms in our patients supported DSM-IV-TR diagnostic criteria and that quantitative methods differentiates PD into meaningful subgroups in a natural setting. DSM diagnostic system describes 13 symptoms associated with panic attack and implies that at least four of these symptoms must occur simultaneously. Physical and psychological symptoms in this broad spectrum are seen in certain loadings in most of the patients and presence of one symtpom seems to be a predictor of another symptom. This variable grouping may guide the researcher and clinician to form the treatment of therapy of the disorder, which also shows a variability in treatment response.
1. Grant BF, Hasin DS, Stinson FS, Dawson DA, Goldstein RB, Smith S, Huang B, Saha TD. The epidemiology of DSM-IV panic disorder and agoraphobia in the United States: results from the National Epidemiologic Survey on Alcohol and Related Conditions. J Clin Psychiatry 2006; 67:363-374.
2. American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders (DSM-IV). Koroglu E (Translation editor). Forth Edition, Ankara: Hekimler Yayın Birliği, 2001. (Turkish)
3. Ley R. The many faces of PAN: psychological and physiological differences among three types of panic attacks. Behav Res Ther 1992; 30:347-357.
4. Seguí J, Salvador-Carulla L, García L, Canet J, Ortiz M, Farré JM. Subtyping panic disorders according to their symptoms. Med Clin (Barc) 1998; 110:524-528.
5. Seguí J, Salvador-Carulla L, García L, Canet J, Ortiz M, Farré JM. Semiology and subtyping of panic disorder. Acta Psychiatr Scand 1998; 97:272-277.
6. Aronson TA, Logue JM. Phenomenology of panic attacks: a descriptive study of panic disorder patients’ self reports. J Clin Psychiatry 1988; 49:8-13.
7. Briggs AC, Stretch DD, Brandon S. Subtyping of panic disorder by symptom profile. Br J Psychiatry 1993; 163:201-209.
8. Shioiri T, Someya T, Murashita J, Takahashi S. The symptom structure of panic disorder: a trial using factor and cluster analysis. Acta Psychiatr Scand 1996; 93:80-86.
9. Starcevic V, Kellner R, Uhlenhuth EH, Pathak D. The phenomenology of panic attacks in panic disorder with and without agoraphobia. Comp Psychiatry 1993; 34:36-41.
10. Sarisoy G, Boke O, Arik AC, Sahin AR. Panic disorder with nocturnal panic attacks: symptoms and comorbidities. Eur Psychiatry 2008; 23:195-200.
11. Meuret AE, White KS, Ritz T, Roth WT, Hofmann SG, Brown TA. Panic attack symptom dimensions and their relationship to illness characteristics in panic disorder. J Psychiatr Res 2006; 40:520-527.
12. Biber B, Alkin T. Panic disorder subtypes: differential responses to CO2 challenge. Am J Psychiatry 1999; 156:739-744.
13. Mellman TA, Uhde TW. Sleep panic attacks; new clinical findings and theoretical implications. Am J Psychiatry 1989; 146:1204-1207.
14. Labbate LA, Pollack MH, Otto MW, Langenauer S, Rosenbaum JF. Sleep panic attacks: an association with childhood anxiety and adult psychopathology. Biol Psychiatry 1994; 36:57-60.
15. Katon WJ. Panic disorder and somatization: review of 55 cases. Am J Med 1984; 77:101-106.
16. Garssen B, de Beurs E, Buikhuisen M, van Balkom A, Lange A, van Dyck R. On distinguishing types of panic. J Anxiety Disord 1996; 10:173-184.
17. Nardi AE, Nascimento I, Valença AM, Lopes FL, Mezzasalma MA, Zin WA, Versiani M. Respiratory panic disorder subtype: acute and long-term response to nortriptyline, a noradrenergic tricyclic antidepressant. Psychiatry Res 2003; 120:283-293.
18. Jacob RG, Møller MB, Turner SM, Wall C 3rd. Otoneurological examination in panic disorder and agoraphobia with panic attacks: a pilot study. Am J Psychiatry 1985; 142:715-720.
19. Lydiard RB, Greenwald S, Weissman MM, Johnson J, Drossman DA, Ballenger JC. Panic disorder and gastrointestinal symptoms: findings from the NIMH epidemiologic catchment Area Project. Am J Psychiatry 1994; 151:64-70.
20. Cassano GB, Petracca A, Perugi G, Toni C, Tundo A, Roth M. Derealization and panic attacks: a clinical evaluation of 150 patients with panic disorder/agoraphobia. Compr Psychiatry 1989; 30:5-12.
21. Boer JA. Defining panic: a diagnostic dilemma. Hum Psychopharmacol Clin Exp 1997; 12 (Suppl.1): 3-6.
22. Cox BJ, Hasey G, Swinson RP, Kuch K, Cooke R, Warsh J, Jorna T. The symptom structure of panic attacks in depressed and anxious patients. Can J Psychiatry 1993; 38:181-184.
23. de Beurs E, Garssen B, Buikhuisen M, Lange A, van Balkom A, Van Dyck R. Continuous monitoring of panic. Acta Psychiatr Scand 1994; 90:38-45.
24. Bandelow B, Amering M, Benkert O, Marks I, Nardi AE, Osterheider M, Tannock C, Tremper J, Versiani M. Cardio-respiratory and other symptom clusters in panic disorder. Anxiety 1996; 2:99-101.
25. Konkan R, Yacinkaya S, Erkiran M, Erkmen H. Panic disorder and comorbidity. Düşünen Adam The Journal of Psychiatry and Neurological Sciences 2003;16:219-222. (Turkish)
26. Cox BJ, Endler NS, Norton GR. Levels of “nonclinical panic”. J Behav Ther Exp Psychiatry 1994; 25:35-40.
27. First MB, Spitzer RL, Gibbon M. Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I) Clinical Version, Washington DC: American Psychiatric Pres Inc, 1997.
28. Corapcioglu A, Aydemir O, Yildiz M, Esen A, Koroglu E. Structured Clinical Interview for DSM-IV Axis I Disorders (SCID-I), Clinical Version. Hekimler Yayin Birligi, Ankara, 1999. (Turkish)
29. Bandelow B, Brunner E, Broocks A, Beinroth D, Hajak G, Pralle L, Rüther E. The use of the Panic and Agoraphobia Scale (PAS) in a Clinical Trial. Psychiatry Res 1998; 77:43-49.
30. Tural U, Fidaner H, Alkin T, Bandelow B. Validity and reliability of the Turkish version of the Panic and Agoraphobia Scale. Turk Psikiyatri Derg 2000; 11:29-39. (Turkish)
31. Williams JB. A structured interview guide for Hamilton Depression Rating Scale. Arch Gen Psychiatr 1988; 45:742-747.
32. Akdemir A, Orsel S, Dag I, Turkcapar H, Iccan N, Ozbay H. Clinical use and the reliability and validity of the Turkish version of the Hamilton Depression Rating Scale (HDRS). Journal of Psychiatry Psychology and Psychopharmacology 1996; 4:251-259. (Turkish)
33. Beck AT, Epstein N, Brown G, Steer RA. An inventory for measuring clinical anxiety: psychometric properties. J Consult Clin Psychol 1988; 56:893-897.
34. Ulusoy M, Sahin N, Erkmen H. Turkish Version of The Beck Anxiety Inventory: psychometric properties. J Cognitive Psychotherapy: Int Quaterly 1998; 12:28-35.
35. Turgeon L, Marchand A, Dupuis G. Clinical features in panic disorder with agoraphobia: a comparison of men and women. J Anxiety Disord 1998; 12:539-553.
36. Shioiri T, Someya T, Fujii K, Noguchi T, Takahashi S. Differences in symptom structure between panic attack and limited symptom panic attack: a study using cluster analysis. Psychiatry Clin Neurosci 1997; 51:47-51.