Behav. Res. Ther. Vol. 30, No. 6, pp. 591-596, 1992 0005-7967/92 $5.00 + 0.00

Printed in Great Britain. All rights reserved Copyright 0 1992 Pergamon Press Ltd

AGORAPHOBIC AVOIDANCE AND PANIC FREQUENCY AS PREDICTORS OF LABORATORY INDUCED PANIC

REACTIONS

PATRICK LYNCH,’ DONALD BAKAL, ** WILLIAM WHITELAW,~ TAK FUNGI

and LINDA ROSE*

‘Division of Psychology, Foothills Hospital, 2Department of Psychology, ‘Faculty of Medicine and 4Academic Computing Services, University of Calgary, Calgary, Alberta, Canada T2N lN4

(Received 13 December 1991)

Summary-This study examined the importance of agoraphobic avoidance and frequency of panic as predictors of psychological and physiological responses of panic sufferers to a laboratory based provocation procedure. Psychophysiologic comparisons were made between 22 panic disorder patients and 15 controls, at baseline and across three periods of carbon dioxide gas inhalations (1, 3, 5%; balance oxygen). Subjective measures of anxiety, frightening cognitions and body sensations were obtained across the phases. Physiological measures of minute ventilation, breathing rate, tidal volume, end tidal CO, and heart rate were also obtained. Between group comparisons revealed significant differences between the groups on the subjective measures with no significant differences occurring on the physiological measures. Within group analyses revealed that pre-session questionnaire measures of agoraphobia avoidance and panic frequency predicted the degree of anxiety, frightening sensations and cognitions during baseline and 5% CO, inhalation. The results indicated that both self-reported agoraphobic avoidance and panic frequency are strong clinical predictors of psychological reactions of panic sufferers during laboratory provocation.

It has been known since the time of the first laboratory provocation studies, that the characteristics Ss bring to the laboratory setting play a major role in determining the magnitude of panic response to the specific provocation under investigation. Psychological explanations for the nature and degree of panic observed in the laboratory include anticipatory anxiety, fear of fear, expectancy, and uncontrollability (Clark, 1986; Margraf, Ehlers & Roth, 1986; McNally, 1990). It is possible that the majority of observed reaction is initiated by anticipatory anxiety to the setting. Consistent with this hypothesis is the fact that panic responses to the two most widely used methods of provocation, lactate infusion and carbon dioxide inhalation, have been linked to levels of baseline anxiety (Margraf et al., 1986; van den Hout, 1988).

Both agoraphobic avoidance and frequency of panic attacks are key determinants of the severity of panic disorder, as defined by DSM III-R (American Psychiatric Association, 1987). Conceptu- ally, avoidance behaviors and panic frequency are viewed as developing from ongoing panic attacks; i.e. fear of fear rather than fear of situations avoided (Goldstein & Chambless, 1978). Craske, Sanderson and Barlow (1987) presented data indicating that degree of avoidance and estimates of panic frequency were far from perfectly related and therefore capable of contributing differently to panic reactions during laboratory provocations.

This study examined, as part of a larger provocation study, the predictive power of agoraphobic avoidance and panic frequency in determining the psychological and physiological responses to carbon dioxide. The larger study examined the chronic hyperventilation hypothesis based on chemoreceptor sensitivity and respiratory after-discharge (Lynch, Bakal & Whitelaw, 1990). Although some between group (Panic disorder, Controls) data are presented, the focus is on the utility of agoraphobic avoidance and panic frequency as clinical predictors of psychological and physiological reactions of panic disorder patients during the laboratory baseline period and during respiratory stimulation by 5% CO, inhalation.

*Author for correspondence.

591

592 PATKICK LYNCH et al.

METHOD

Twenty-two Ss meeting DSM-III criteria for Panic Disorder with or without agoraphobia and 15 normal controls were recruited for the study. The Ss reported at least 4 of the 12 DSM-III symptoms to be moderately severe during episodes of anxiety and indicated a frequency of panic of at least once per week. The controls were selected from hospital and university staff, Both groups of Ss, all female, attended a clinical screening interview and completed a series of assessment questionnaires prior to their participation. Ss were required to undergo a brief medical examination to determine whether they were suffering from any physical condition that might be grounds for their exclusion from the study. All Ss provided informed consent and no Ss withdrew during the course of the study.

Pre-session measures

A number of questionnaire measures were used to validate the clinical diagnosis and to operationalize agoraphobic avoidance and panic severity. Agoraphobic avoidance was assessed with two questionnaires; the Fear Questionnaire (FQ; Marks & Mathews, 1979) and the Mobility Inventory for Agoraphobia (MIA; Chambless, Caputo, Jasin, Gracely & Williams, 1985). The FQ provides scores on three categories of clinical phobias: agoraphobia (FQ-AG), blood injury (FQ-BI) and social phobia (FQ-S). The MIA is a 27-item instrument which assesses a wide range of situations that may provoke anxious avoidance, either alone or in the presence of a trusted companion. The combination score was used in this study. The MIA also has an item which allows for the assessment of reported frequency of panic attacks during the previous week. Ss also completed the Agoraphobic Cognitions Questionnaire (ACQ) and the Body Sensations Question- naire (BSQ; Chambless, Caputo, Bright & Gallagher, 1984). The ACQ consists of 14 items dealing with thoughts of physical catastrophe due to anxiety symptoms and with thoughts reflecting social or behavioral disaster from loss of control. Ss rate the thoughts in terms of how often each occurs when they are nervous or frightened. The BSQ is a 17-item scale concerning the degree to which patients fear somatic symptoms commonly associated with panic. Ss rate the degree of fear usually associated with the various sensations. The Beck Depression Inventory (BDI; Beck, Ward, Mendeisohn, Mock & Erbaugh, 1961) was also included.

Psychological and ~~y.~io~ogica~ measures obtained during CO, ~ro~ocat~o~

Assessment of panic-related cognitions and sensations across the experimental phases was achieved with a modified version of the ACQ and BSQ. Ss were asked to rate how persistent/intense the various thoughts/sensations were during the actual phases of the laboratory procedures (baseline, 1, 3. 5%) rather than in terms of general or usual fearfulness. Subjective anxiety/panic across the experimental phases was assessed with a IO-point Likert scale ranging from 0 (very relaxed and sleepy) through 5 (moderately severe anxiety) to 10 (full-blown panic attack).

Expired air was continuously sampled from a mouthpiece through which the Ss breathed. End tidal CO, and minute ventilation were both assessed across the baseline and CO, inhalation phases of the study. End tidal COz was monitored with a Beckman LB-II Infra-Red Gas Analyzer. Minute ventilation was determined on the basis of expired flow, which was measured by a Fleisch No. 2 pneumotachograph and a Validyne differential pressure transducer attached to the output side of an Otis-McKerrow valve, with the signal sent to a Hewlett-Packard Carrier amplifier. The gas mixtures were administered from a rubber reservoir bag connected to the input side of the Otis-McKerrow valve. The reservoir was kept partially full of gas through a continuous flow from the calibrated gas cylinder. With each change in gas concentration, the reservoir and connecting tubes were thoroughly flushed and filled with the next gas concentration.

Procedure

The clinical interview and questionnaires were administered in a session separate from the laboratory session. Ss were informed that this was a study of respiratory parameters and that they would be required to lie supine and breathe different mixtures of CO,-0,. In order to reduce anxiety, they were told that this was not a study of panic attacks, but of their general physiology,

Predictors of laboratory induced panic reactions 593

and were reassured that measures would be taken to reduce the possibility that they would panic in the laboratory. They were further informed that any physiological changes occurring during the procedure would be a result of breathing the gases.

The laboratory session began with the measurement of the Ss Forced Vital Capacity (FVC) and Forced Expiratory Volume in 1 set (FEVl). Next, the S was required to lie on a bed and the EKG electrodes were attached. The S was familiarized with the use of the mouthpiece and noseclip. Once everything was in place, the 5-min baseline period began during which time the Ss breathed room air through the mouthpiece. At the end of this period, the ACQ, BSQ and Anxiety rating scale were completed. The baseline was followed by three successive 5-min CO, gas inhalation phases (1, 3, 5%). Ss completed the three subjective rating scales at the end of each phase. The instructional set given to Ss to complete the rating scales was in reference to their experiences during the immediately proceding phase.

RESULTS

Pre-experimental compurisons

There were no significant differences between the panic Ss and controls in terms of age, height, weight, FVC, or FEVl (Table 1). The groups were significantly different in terms of their agoraphobic avoidance scores as measured by the MIA (t = 6.91, d.f. = 25, P < 0.001) and the FQ-AG (t = 5.40, d.f. = 23, P < 0.001). The panic Ss exhibited significantly more phobic behavior on the FQ-S subscale (t = 5.80, d.f. = 28, P < 0.001) and the FQ-BI subscale (t = 4.24, d.f. = 29, P < 0.001) of the Fear Questionnaire. Panic patients also indicated greater presence of frightening cognitions and sensations as measured by the ACQ (t = 7.57, d.f. = 31, P < 0.001) and the BSQ (t = 5.37, d.f. = 28, P < 0.001). Panic patients were more depressed in terms of their BDI scores (t = 6.10, d.f. = 24, P < 0.001). The intercorrelations of the pre-session measures are presented in Table 2. The two measures of agoraphobic avoidance, the MIA and FQ-AG, correlated 0.84 while each correlated 0.30 and 0.24 with the measure of panic frequency.

Group d@erences in psychological and physiological responses across baseline and CO, provocation

The Panic Group and Controls were compared in terms of their subjective and physiological responses across the baseline and CO, inhalation phases. A two way (Groups, Phases) MANOVA performed on the BSQ, ACQ, and Anxiety rating scale revealed significant main effects for Groups [F(3,33) = 5.22, P < 0.011 and Phases [F(9,27) = 5.99, P < O.OOl]. The Group univariate Fs for BSQ scores [F(1,35) = 13.84, P <O.OOl], and ACQ scores [F(1,35), = 12.15, P <O.Ol] were significant. The univariate F for ACQ [F(1,35) = 5.33, P < 0.031 was also significant. The Phases univariate Fs for BSQ, ACQ and Anxiety scores were also significant (P < 0.01). There were a

Table I. Comparison of Panic Disorder patients and Controls on pre-experimental physical and psychological characteristics*

Panic Disorder Controls

M SD M SD

Age (yr) 33.1 9.1 33.7 8.4 Height (cm) 160.1 6.8 161.3 7.2 Weight (kg) 61.4 13.5 64.3 13.7 FVC (liters) 3.9 0.6 4.1 0.6 FEVI (liters) 3.4 0.5 3.5 0.6 FQ-AG 14.7 II.0 I.7 2.2 FQ-BI I I.8 8.1 3.7 3.2 FQ-S 17.1 8.6 5.6 2.9 MIA 119.7 38.8 59.6 10.5 BDI 19.2 II.3 3.9 2.7 Panic Freq. 4.7 3.1 0.0 0.0 ACQ (PM 34.0 8.9 17.7 4.1 BSQ (P=) 43.8 17.2 22.3 6.3

‘FVC, forced vital capacity; FEVI, forced expiratory volume in I set; FQ-AG. FQ-BI, FQ-S, agoraphobia, blood-injury, social phobia subscales of Fear Questionnaire; MIA, Mobility Inventory for Agoraphobia; BDI, Beck Depression Inventory; Panic Freq., number of panic attacks in week preceding assess- ment; ACQ, Agoraphobic Cognitions Questionnaire; BSQ, Body Sensations Questionnaire.

Table 2. Correlations between w-session assessment measures

FO-S

FQ-AG 0.41 FQ-S FQ-BI MIA BDI Panic Freq. ACQ (pre)

Panic ACQ BSQ FQ-BI MIA BDI Freq. (pre) (pre)

0.22 0.84 0.34 0.24 0.53 0.29 0.27 0.43 0.18 0.49 0.37 0.35

0.14 0.37 0.09 0.44 0.25 0.50 0.30 0.47 0.37

0.29 0.42 0.20 0.20 0.13

0.31

594 PATRICK LYNCH el al.

number of significant differences observed with the comparison between the phases, with the largest differences occurring in comparisons made with the subjective ratings obtained following 5% CO, inhalation. The ratings obtained at the end of this phase were significantly greater for ACQ, BSQ, and Anxiety than the ratings obtained at baseline, 1% CO,, and 3% CO, inhalation (P < 0.001). The Groups x Phase MANOVA interaction was nonsignificant.

The physiologic data were also analyzed across the baseline and three CO2 inhalation phases. The data were reduced to I-min averages for the time periods defining each phase. The data for the last minute of each phase were used in the analysis. A two-way repeated measures MANOVA (Groups, Phases) was carried out using a total of five physiological measures (minute ventilation, breathing rate, tidal volume, end-tidal CO2 and heart rate). The multivariate effects for Phases were significant [F(13,22) = 49.20, P < O.OOl]. The univariate Fs for all the ventilatory variables were significant (P < 0.001) indicating that the CO, manipulations were effective in altering breathing. Heart rate also showed a significant univariate F for Phases, F(3,32) = P < 0.001. Post -hoc analyses revealed significant differences with phase comparisons made with the 5% CO, inhalation phase (P < 0.001). The multivariate F for Groups was nonsignificant which indicated that the panic disorder patients and controls showed no evidence of physiological differences.

Agoraphobic avoidance and panic frequency

In order to determine if agoraphobic avoidance and panic frequency behavior differentially predicted psychological and/or physiological responses within panic sufferers during the exper- iment, the pre-experimental measures of agoraphobic avoidance (MIA and FQ-AG), agoraphobic cognitions and sensations (ACQ, BSQ) and frequency of panic attacks, as well as the blood-injury (FQ-BI), and social phobias (FQ-S) and depression (BDI) were examined in a series of simple and multiple stepwise regression analyses. A separate analysis was conducted for baseline to determine if pre-experimental S characteristics predicted anticipatory reactions prior to actual provocation.

The simple regression coefficients for the baseline and 5% CO, inhalation phases are presented in Tables 3 and 4, respectively. Inspection of these tables indicates that agoraphobic avoidance and panic frequency were predictive of subjective reactions during both baseline and following the 5% CO, inhalation. The blood-injury subscale (FQ-BI) also predicted subjective reactions across the phases. The FQ-S was a nonsignificant predictor during both phases while the BDI predicted ACQ and BSQ scores during 5% CO, inhalation. The pre-experimental assessment of ACQ and BSQ yielded only one significant regression coefficient: ACQ(pre) predicted ACQ scores following 5% CO, provocation.

The stepwise multiple regression analysis found agoraphobic avoidance, measured by the MIA, to be predictive of Anxiety scores during baseline [F(1,18) = 5.55, P < 0.051, BSQ scores during baseline [F( 1,18) = 11.08, P < 0.011, and ACQ scores during baseline [F( 1,lS) = 16.58, P < 0.001 J. Number of panic attacks entered as a second step for BSQ at baseline. The stepwise predictors changed for responses to 5% CO,. Panic frequency predicted Anxiety, F(1,18) = 8.19, P < 0.01, and BSQ, F( 1,18) = 8.75, P < 0.01, following 5% CO, inhalation. ACQ scores following inhalation were predicted by FQ-BI, F( 1,18) = 11.14, P < 0.01.

The pre-session measures of agoraphobic avoidance, frightening cognitions and sensations, panic attacks, and depression failed to account for significant variance associated with the ventilatory

Table 3. Pre-session psychological predictors of anxiety, BSQ and ACQ scores during baseline

(simple reeression)

Dependent variables

Anxiety BSQ ACQ - Regression coefficient

MIA Panic Freq. FQ-AG FQ-BI FQ-S BDI ACQ (PN BSQ (ore)

0.02’ 0.38** 0.19** 0.27’ 4.66’* 1.12 1.21” 1.21* 0.63** 0.10* 1.41* 0.72’

-0.02 0.15 0.21 0.02 0.44 0.23 0.06 0.11 0.27 0.26 0.29 0.17

‘P < 0.05; **P < 0.01

Table 4. Pre-session psychological predictors of anxiety, BSQ and ACQ scores following 5% CO,

inhalation (simple regression)

Dependent variables

Anxiety BSQ ACQ Reeression coefficient

MIA Panic Freq. FQ-AG FQ-BI FQ-S BDI ACQ (pre) BSQ (PW

0.03* 0.44* 0.24*’ 0.46” 1.4s’ 3.59** 0.11’ I .36 0.75’ 0.16** I .98* 1.3s** 0.03 1.00 0.72 0.03 I .53* 0.78; 0.21 0.23 0.39’ 0.24 0.23 0.31

Predictors of laboratory induced panic reactions 595

measures during baseline and during 5% CO, inhalation. However, panic frequency was related to heart rate levels during baseline, F( 1,18) = 9.57, P < 0.01 and during 5% CO, inhalation, F(l,l8) = 9.79, P < 0.01.

A further examination of the relationship between agoraphobic avoidance and patient reactivity was suggested by the work of Mavissakalian (1986) using the Agoraphobia subscale of the Fear Questionnaire (FQ-AG). Mavissakalian proposed that a score of > 30 on this subscale is indicative of severe agoraphobic avoidance. The Panic Disorder patients were divided into those who had a FQ-AG score > 30 (n = 4) and those who had a score < 30. The controls all had a scale score ~6. A two way (Groups, Phases) MANOVA performed on the total scores of these rating scales revealed a significant main effects for Groups, F(3,32) = P < 0.002 but not for Phases. The univariate comparisons for the > 30 FQ-AG Ss were significantly different from the < 30 FQ-AG panic disorder Ss and controls on the Anxiety scale (P < 0.001) and the BSQ (P < 0.001). The two panic disorder groups did not differ statistically in terms of their scores on the ACQ. The <30 FQ-AG panic group did not differ from the controls, although their scores were consistently greater than those observed for the controls. A MANOVA analysis of the physiologic data failed to produce a significant Groups effect.

DISCUSSION

This study found evidence for the importance of agoraphobic avoidance and panic frequency as predictors of panic anxiety, cognitions and sensations during a standard laboratory provocation. The clinical measures predicted levels of subjective reactions during both resting baseline and during 5% COz provocation, suggesting that their influence was pervasive and to some extent independent of the manipulations in use.

Agoraphobic avoidance seemed to enhance anticipatory anxiety during baseline in some patients, in spite of repeated reassurances that this was not a study of panic attacks. Others have observed that anticipatory anxiety often accompanies clinical agoraphobic avoidance. Cox, Swinson, Norton and Kuch (1991) observed that scores on the FQ-AG subscale were highly correlated with perceived anticipation of panic attacks in specific situations.

Although both agoraphobic avoidance and panic frequency predicted subjective reactions of panic sufferers across the baseline and 5% CO, inhalation, there was a shift in their dominance with agoraphobic avoidance predicting greater variance of the subjective measures during baseline and panic frequency accounting for more variance following 5% CO, inhalation. The reliability of the differential predictions remains uncertain given that panic frequency was assessed with one item which asked Ss to report the number of panic attacks experienced during the past 7 days. Agoraphobic avoidance and panic frequency may represent different aspects of panic disorder as the correlations of the panic frequency measure with the two agoraphobic avoidance measures were quite low. Craske et al. (1987) also observed nonsignificant relationships between panic frequency and agoraphobic avoidance and suggested that avoidance behavior was more closely related to anticipation of panic rather than panic frequency.

The observed baseline differences in panic-related thoughts and sensations raise the question of whether it is possible to test severe panic patients in a laboratory without provoking considerable anxiety prior to the provocation. The cardinal characteristic of agoraphobic avoidance is “the fear of being in places or situations from which escape might be difficult (or embarrassing) or in which help might not be available in the event of a panic attack” (American Psychiatric Association, 1987). Entering a lab and having transducers attached made a rapid or unnoticed escape impossible.

The physiologic data from the present study provided little or no support for ventilatory differences between panic sufferers and controls. The use of the 5-min inhalation period may not have been sufficient for Ss to have reached asymptotic levels of minute ventilation. Klein and his group have used inhalation periods of 20 min and found that panic attacks occur well after 5 min (D. F. Klein, pers. commun., 21 August 1991). The present results suggest that provocation paradigms need to consider the possibility that observed panic is the result of agoraphobic-based anticipatory fears and panic severity in interaction with situational cues and demands which accompany the panicogenic agent.

596 PATRICK LYNCH et al,

REFERENCES

American Psychiatric Association (1987). Diagnostic and statistical manual ofmental disorders (3rd Edn, rev.). Washington, D.C.: American Psychiatric Association,

Beck, A. T., Ward, C. H., Mendelson, M., Mock, J. & Erbaugh, J. (1961). An inventory for measuring depression. Archives of General Psychiatry, 4, 561-571.

Chambless, D. L., Caputo, G. C., Bright, P. & Gallagher, R. (1984). The assessment of fear of fear in agoraphobics: The body sensations questionnaire and the agoraphobic cognitions questionnaire. Journal of Consulting and Clinical Psychology, 5.2, 1090-1097.

Chambless, D. L., Caputo, G. C., Jasin, S. E., Gracely, E. J. & Williams, C. (1985). The Mobility Inventory for Agoraphobia. Behaviour Research and Therapy, 23, 35-44.

Clark, D. M. (1986). A cognitive approach to panic. Behaviour Research and Therapy, 24, 461470. Cox, B. J., Swinson, R. P., Norton, G. R. & Kuch, K. (1991). Anticipatory anxiety and avoidance in panic disorder with

agoraphobia. Behaviour Research and Therapy, 29, 363-365. Craske, M. G., Sanderson, W. C. & Barlow, D. H. (1987). The relationships among panic, fear, and avoidance. Journal

of Anxiety Disorders, I, 153S160. Goldstein, A. J. & Chambless, D. L. (1978). A reanalysis of agoraphobia. Behavior Therapy, 9, 4749. Hout, M. A. van den (1988). The explanation of experimental panic. In S. Rachman & J. D. Maser (Eds), Panic:

Psychological perspectives. Hillsdale, N.J.: Erlbaum. Lynch, P. B., Bakal, D. A. & Whitelaw, W. A. (1990). Carbon dioxide recovery in panic disorder patients and controls,

Psychosomatic Medicine, 52, 226 (Abstract). Margraf, J., Ehlers, A. & Roth, W. T. (1986). Sodium lactate infusion and panic attacks: A review and critique.

Psychosomatic Medicine, 48, 23-5 I. Marks, I. M. & Mathews, A. M. (1979). Brief standard self-rating scale for phobic patients. Behaviour Research and Therapy,

17, 263-267. Mavissakalian, M. (1986). The Fear Questionnaire: A validity study. Behaviour Research and Therapy, 24, 83-85. McNally, R. J. (1990). Psychological approaches to panic disorder: A review. Psychological Bulletin, 108, 403419.