J. psychio~.Res., Vol. 24, No. I. PP. I-8, 1990
0022-3956/W 13.00 + .OO 0 1990 Pergamon Press plc
Printed inGreat Britain.
SEQUENCE
ARTHUR
OF IMPROVEMENT IN AGORAPHOBIA PANIC ATTACKS
RIFKIN*, JOHN C. PECKNOLD~,
BALLENGER~, GRAHAM
D. BURROWSI,
RICHARD
WITH
P. SWINSON~I, JAMES C.
RUSSELL NOYES\\, ROBERT
L. DUPONT**
and
IRA LESSERtt *Department of Psychiatry, Queens Hospital Center, Jamaica, NY, U.S.A.; TDepartment of Psychiatry, St. Mary’s Hospital Center, Montreal, Quebec, Canada; *Department of Psychiatry, Toronto General Hospital, Toronto, Ontario, Canada; $Department of Psychiatry, Medical University of South Carolina, Charleston, SC, U.S.A.; !Department of Psychiatry, University of Melbourne, Heidelberg, Victoria, Australia; 1(Department of Psychiatry, University of Iowa, Iowa City, IA, U.S.A.; **Institute of Behavior and Health, Inc. Rockville, MD, U.S.A. and WDepartment of Psychiatry, Harbor-UCLA Medical Center, Torrance, CA, U.S.A. (Received 10 January
1989; revised 31 March 1989; re-revised22 June 1989)
Summary-In a multi-center comparison of alprazolam to placebo in the treatment of agoraphobia with panic attacks, the sequence of sustained remissionin both treatment groups, was panic attacks before phobias. This may suggest that phobias are secondary to panic attacks in the pathogenesis of the disorder, although other explanations may account for these data and are discussed. INTRODUCTION
A LARGE, multi-site study comparing alprazolam to placebo in the treatment of panic disorder provided the opportunity to study the sequence of improvement of the major symptoms of this disorder: panic attacks, anticipatory anxiety, and agoraphobia, since most patients with frequent panic attacks also have the other two symptoms. The results of this study concerning efficacy (BALLENGER, BURROWS, DUPONT, LESSER, NOYES, PECKNOLD, RIFKIN, & SWINSON, 1988), safety and side-effects, (NOYES, DUPONT, PECKNOLD, RIFKIN, RUBIN,S~INSON,BALLENGER,&BIJRROWS, 1988)and discontinuation (PECKNOLD, SWJNSON, KUCH, & LEWIS, 1988), have been presented. Details of the protocol and a description of the subjects have been presented there and will only be summarized here. The sequence of improvement is important because it may provide guidelines to clinicians (and patients) about what to expect, and perhaps shed some light on theories of pathogenesis of this disorder. In particular we wish to indirectly address the hypothesis proposed by KLEIN (1981) that agoraphobia with panic attacks generally follows a specific pattern. He postulates that the initial event is unexpected panic attacks. Secondarily this leads to anticipatory anxiety and agoraphobia because the person is especially fearful of panic attacks in situations from which escape is difficult. Klein’s hypothesis was based on his clinical observations on the development of symptoms of Panic Disorder with agoraphobia. This paper assesses whether longitudinal and symptomatic observations on the response to alprazolam (or placebo) support Klein’s hypothesis. The logic of our partial test of this hypothesis is that the sequence of improvement gives information about the sequence of onset. If phobia is secondary
2
A. RIFKINet al.
to panic attacks, then phobia should remit after panic attacks. This logic is not air-tight, which will be discussed further after presentation of the data. METHOD
This study was done at five sites in the United States, two in Canada, and one in Australia. Subjects had agoraphobia with panic attacks or panic disorders. Informed consent was obtained after the nature of the procedures had been fully explained. After a seven day drug-free period, subjects were randomized to receive alprazolam or placebo under doubleblind conditions, for eight weeks. Outcome evaluations were done at baseline, and at the end of the first, second, third, fourth, sixth, and eight weeks. In this analysis the results from all eight sites are pooled. The maximum dosage was 10 mg/day. At the end of four weeks the mean dose was 4.9 + 1.5 mg/day, and at the end of the eighth week it was 5.7 + 2.2 mg/day. Of the 506 subjects who entered the study, 351 completed eight weeks (88% of those randomized to alprazolam, and 51% to placebo). For both treatment groups the mean age was 37 years, and 67% were women. The mean duration of the present episode in months, for those treated with alprazolam was 76; and 72 for those given placebo. Agoraphobia was present in 83% of the patients in the alprazolam group, and in 84% of those given placebo. Entry into the study required at least one panic attack weekly. A panic attack had to have three or more symptoms. All subjects in this analysis of sequence had to have other criteria in addition to the entry criteria to the study to be sure there were sufficient levels of severity of anticipatory anxiety and phobia. In this study of the sequence of improvement, subjects were included who met specific criteria for severity of symptoms at baseline and who completed all eight weeks of the study. Remission was felt to have occurred if the severity of the symptoms was markedly reduced and remained so throughout the remaining weeks of the study (at least two). For each relevant symptom there was an entrance criterion, the minimal severity required for the subject to be eligible, and a criterion for remission. Panic attacks were categorized as spontaneous, situational and total categories. In this analysis of sequence, three main symptoms (with subdivisions) are used: panic attacks, phobias, and anticipatory anxiety. For details on the instruments used to measure these symptoms, see BALLENGER et al. (1988). Briefly, panic attacks were defined as sudden, severe anxiety that peaked within ten minutes and generally abated within minutes to two hours. Spontaneous panic attacks were considered unexpected by the patient. Those that occurred in a situation where panic attacks had occurred often were considered situational. It might appear that situational panic attacks are similar and hard to differentiate from anticipatory anxiety. However, much effort was spent educating the patient about the distinction. The major difference concerns onset. Situational panic attacks occur as a surge of severe symptoms occurring within seconds to minutes; whereas anticipatory anxiety does not surge. To emphasize this point, graphs were shown to the subjects illustrating this distinction, to supplement the oral explanation. At each visit the patient and the investigator jointly determined the number and type of attack since the previous appointment. Total panic attacks were the sum of spontaneous and situational attacks. The entrance criterion was at least one attack per week, and remission was none.
IMPROVEMENT
3
IN AGORAPHOBIA
Phobia was measured in several ways. In this paper the four most important measures are used. 1. Main phobia was the phobia that bothered the patient the most and the one for which he or she most wanted treatment. It was measured on a ten point scale. 2. Avoidance of main phobia is a measure of how much the subject stayed away from the main phobic situation, assessed on a five point scale from 4 (‘very often’) to 0 (‘never’). 3. Fear of main phobia refers to the fear evoked by the need to confront the main phobic situation. It was measured on a ten point scale from ‘extremely’ to ‘not at all’. 4. Overall phobia: the patient was asked to measure on a ten point scale the severity of all phobias, combining fear and avoidance. Anticipatory anxiety was measured by asking the subject to rate mean intensity for the past week on a ten point scale. For each of these measures, criteria were set, a priori, for minimum severity to be included in this analysis and maximum severity to be considered remitted. These criteria were set after the data were collected. This analysis of sequence was not part of the original design of the study. But, the criteria were established before the analysis of these data. On the ten point scales for main phobia, fear of main phobia, overall phobia, and anticipatory anxiety, the entrance criterion was at least five (moderate), and remission was at most two (mild). On the five point scale of avoidance, the entrance criterion was at least three (often), and remission was one (never).
Statistical analysis A subject was considered remitted if the pre-determined level of remission was reached and sustained for at least two weeks, and until the conclusion of the study period. Each relevant symptom was plotted as remitted or not for each session. The mean week of reaching sustained remission was compared using a t-test. A non-parametric test, the Mann-Whitney, was also used. The results were the same for each test. Only the results of the t-test are reported in this paper. RESULTS
Table 1 shows the sequence of improvement with the mean weeks of sustained remission for all subjects (placebo and alprazolam). The sample sizes differ according to the number
TABLE 1. WEEK OF SUSTAINEDREMISSION Symptom
N
Spontaneous panic attacks Situational panic attacks Anticipatory anxiety Avoidance of main phobia Total panic attacks Main phobia Overall phobia
298 328 303 345 359 400 398
Fear of main phobia occur.
is not listed because
Mean (SD) 4.5 6.2 6.4 6.5 6.6 7.1 7.2 remission
(3.5) (3.2) (3.1) (3.0) (2.9) (2.6) (2.5) did not
A. RIFKIN et al.
4
who met the entrance criteria and completed eight weeks. Except for spontaneous panic attacks, the weeks of sustained remissions bunch together, and for all symptoms except spontaneous and situational panic attacks most subjects did not achieve sustained remission. Table 2 shows the two-way comparisons for all the symptoms, with one exception, by the statistical significance of the difference in time to sustained remission. The exception is ‘fear of main phobia’, which is not listed because no one achieved sustained remission. The major finding is that all measures of panic attacks, total, spontaneous and situational, remitted before ‘overall phobia’.
TABLE 2. STATISTICAL SIGNIFICANCE OFDIFFERENCE
SpPA SitPA AA AvMPh TotPA MPh OvPH
SpPA
SitPA
AA
-
< 0.05 -
< 0.05 NS -
AvMPh < 0.05 NS NS -
IN TIME TO SUSTAINED REMISSION*
TotPA < 0.05 < 0.05 NS NS -
MPh < < < <
0022-3956/W 13.00 + .OO 0 1990 Pergamon Press plc
Printed inGreat Britain.
SEQUENCE
ARTHUR
OF IMPROVEMENT IN AGORAPHOBIA PANIC ATTACKS
RIFKIN*, JOHN C. PECKNOLD~,
BALLENGER~, GRAHAM
D. BURROWSI,
RICHARD
WITH
P. SWINSON~I, JAMES C.
RUSSELL NOYES\\, ROBERT
L. DUPONT**
and
IRA LESSERtt *Department of Psychiatry, Queens Hospital Center, Jamaica, NY, U.S.A.; TDepartment of Psychiatry, St. Mary’s Hospital Center, Montreal, Quebec, Canada; *Department of Psychiatry, Toronto General Hospital, Toronto, Ontario, Canada; $Department of Psychiatry, Medical University of South Carolina, Charleston, SC, U.S.A.; !Department of Psychiatry, University of Melbourne, Heidelberg, Victoria, Australia; 1(Department of Psychiatry, University of Iowa, Iowa City, IA, U.S.A.; **Institute of Behavior and Health, Inc. Rockville, MD, U.S.A. and WDepartment of Psychiatry, Harbor-UCLA Medical Center, Torrance, CA, U.S.A. (Received 10 January
1989; revised 31 March 1989; re-revised22 June 1989)
Summary-In a multi-center comparison of alprazolam to placebo in the treatment of agoraphobia with panic attacks, the sequence of sustained remissionin both treatment groups, was panic attacks before phobias. This may suggest that phobias are secondary to panic attacks in the pathogenesis of the disorder, although other explanations may account for these data and are discussed. INTRODUCTION
A LARGE, multi-site study comparing alprazolam to placebo in the treatment of panic disorder provided the opportunity to study the sequence of improvement of the major symptoms of this disorder: panic attacks, anticipatory anxiety, and agoraphobia, since most patients with frequent panic attacks also have the other two symptoms. The results of this study concerning efficacy (BALLENGER, BURROWS, DUPONT, LESSER, NOYES, PECKNOLD, RIFKIN, & SWINSON, 1988), safety and side-effects, (NOYES, DUPONT, PECKNOLD, RIFKIN, RUBIN,S~INSON,BALLENGER,&BIJRROWS, 1988)and discontinuation (PECKNOLD, SWJNSON, KUCH, & LEWIS, 1988), have been presented. Details of the protocol and a description of the subjects have been presented there and will only be summarized here. The sequence of improvement is important because it may provide guidelines to clinicians (and patients) about what to expect, and perhaps shed some light on theories of pathogenesis of this disorder. In particular we wish to indirectly address the hypothesis proposed by KLEIN (1981) that agoraphobia with panic attacks generally follows a specific pattern. He postulates that the initial event is unexpected panic attacks. Secondarily this leads to anticipatory anxiety and agoraphobia because the person is especially fearful of panic attacks in situations from which escape is difficult. Klein’s hypothesis was based on his clinical observations on the development of symptoms of Panic Disorder with agoraphobia. This paper assesses whether longitudinal and symptomatic observations on the response to alprazolam (or placebo) support Klein’s hypothesis. The logic of our partial test of this hypothesis is that the sequence of improvement gives information about the sequence of onset. If phobia is secondary
2
A. RIFKINet al.
to panic attacks, then phobia should remit after panic attacks. This logic is not air-tight, which will be discussed further after presentation of the data. METHOD
This study was done at five sites in the United States, two in Canada, and one in Australia. Subjects had agoraphobia with panic attacks or panic disorders. Informed consent was obtained after the nature of the procedures had been fully explained. After a seven day drug-free period, subjects were randomized to receive alprazolam or placebo under doubleblind conditions, for eight weeks. Outcome evaluations were done at baseline, and at the end of the first, second, third, fourth, sixth, and eight weeks. In this analysis the results from all eight sites are pooled. The maximum dosage was 10 mg/day. At the end of four weeks the mean dose was 4.9 + 1.5 mg/day, and at the end of the eighth week it was 5.7 + 2.2 mg/day. Of the 506 subjects who entered the study, 351 completed eight weeks (88% of those randomized to alprazolam, and 51% to placebo). For both treatment groups the mean age was 37 years, and 67% were women. The mean duration of the present episode in months, for those treated with alprazolam was 76; and 72 for those given placebo. Agoraphobia was present in 83% of the patients in the alprazolam group, and in 84% of those given placebo. Entry into the study required at least one panic attack weekly. A panic attack had to have three or more symptoms. All subjects in this analysis of sequence had to have other criteria in addition to the entry criteria to the study to be sure there were sufficient levels of severity of anticipatory anxiety and phobia. In this study of the sequence of improvement, subjects were included who met specific criteria for severity of symptoms at baseline and who completed all eight weeks of the study. Remission was felt to have occurred if the severity of the symptoms was markedly reduced and remained so throughout the remaining weeks of the study (at least two). For each relevant symptom there was an entrance criterion, the minimal severity required for the subject to be eligible, and a criterion for remission. Panic attacks were categorized as spontaneous, situational and total categories. In this analysis of sequence, three main symptoms (with subdivisions) are used: panic attacks, phobias, and anticipatory anxiety. For details on the instruments used to measure these symptoms, see BALLENGER et al. (1988). Briefly, panic attacks were defined as sudden, severe anxiety that peaked within ten minutes and generally abated within minutes to two hours. Spontaneous panic attacks were considered unexpected by the patient. Those that occurred in a situation where panic attacks had occurred often were considered situational. It might appear that situational panic attacks are similar and hard to differentiate from anticipatory anxiety. However, much effort was spent educating the patient about the distinction. The major difference concerns onset. Situational panic attacks occur as a surge of severe symptoms occurring within seconds to minutes; whereas anticipatory anxiety does not surge. To emphasize this point, graphs were shown to the subjects illustrating this distinction, to supplement the oral explanation. At each visit the patient and the investigator jointly determined the number and type of attack since the previous appointment. Total panic attacks were the sum of spontaneous and situational attacks. The entrance criterion was at least one attack per week, and remission was none.
IMPROVEMENT
3
IN AGORAPHOBIA
Phobia was measured in several ways. In this paper the four most important measures are used. 1. Main phobia was the phobia that bothered the patient the most and the one for which he or she most wanted treatment. It was measured on a ten point scale. 2. Avoidance of main phobia is a measure of how much the subject stayed away from the main phobic situation, assessed on a five point scale from 4 (‘very often’) to 0 (‘never’). 3. Fear of main phobia refers to the fear evoked by the need to confront the main phobic situation. It was measured on a ten point scale from ‘extremely’ to ‘not at all’. 4. Overall phobia: the patient was asked to measure on a ten point scale the severity of all phobias, combining fear and avoidance. Anticipatory anxiety was measured by asking the subject to rate mean intensity for the past week on a ten point scale. For each of these measures, criteria were set, a priori, for minimum severity to be included in this analysis and maximum severity to be considered remitted. These criteria were set after the data were collected. This analysis of sequence was not part of the original design of the study. But, the criteria were established before the analysis of these data. On the ten point scales for main phobia, fear of main phobia, overall phobia, and anticipatory anxiety, the entrance criterion was at least five (moderate), and remission was at most two (mild). On the five point scale of avoidance, the entrance criterion was at least three (often), and remission was one (never).
Statistical analysis A subject was considered remitted if the pre-determined level of remission was reached and sustained for at least two weeks, and until the conclusion of the study period. Each relevant symptom was plotted as remitted or not for each session. The mean week of reaching sustained remission was compared using a t-test. A non-parametric test, the Mann-Whitney, was also used. The results were the same for each test. Only the results of the t-test are reported in this paper. RESULTS
Table 1 shows the sequence of improvement with the mean weeks of sustained remission for all subjects (placebo and alprazolam). The sample sizes differ according to the number
TABLE 1. WEEK OF SUSTAINEDREMISSION Symptom
N
Spontaneous panic attacks Situational panic attacks Anticipatory anxiety Avoidance of main phobia Total panic attacks Main phobia Overall phobia
298 328 303 345 359 400 398
Fear of main phobia occur.
is not listed because
Mean (SD) 4.5 6.2 6.4 6.5 6.6 7.1 7.2 remission
(3.5) (3.2) (3.1) (3.0) (2.9) (2.6) (2.5) did not
A. RIFKIN et al.
4
who met the entrance criteria and completed eight weeks. Except for spontaneous panic attacks, the weeks of sustained remissions bunch together, and for all symptoms except spontaneous and situational panic attacks most subjects did not achieve sustained remission. Table 2 shows the two-way comparisons for all the symptoms, with one exception, by the statistical significance of the difference in time to sustained remission. The exception is ‘fear of main phobia’, which is not listed because no one achieved sustained remission. The major finding is that all measures of panic attacks, total, spontaneous and situational, remitted before ‘overall phobia’.
TABLE 2. STATISTICAL SIGNIFICANCE OFDIFFERENCE
SpPA SitPA AA AvMPh TotPA MPh OvPH
SpPA
SitPA
AA
-
< 0.05 -
< 0.05 NS -
AvMPh < 0.05 NS NS -
IN TIME TO SUSTAINED REMISSION*
TotPA < 0.05 < 0.05 NS NS -
MPh < < < <
