Central and peripheral chemoreflexes in panic disorder

Psychiatry Research 113 (2002) 181–192

Central and peripheral chemoreflexes in panic disorder Martin A Katzmana,b,*, Lukasz Struzika,c, Nishka Vijaya, Aimee Coonerty-Femianoa, Safraaz Mahamedc, James Duffinc a

Anxiety Disorders Clinic, Centre for Addiction and Mental Health-Clarke Division, 250 College Street, Toronto, Ontario, Canada M5T 1R8 b Department of Psychiatry, Faculty of Medicine, University of Toronto, 1 King’s College Circle, Toronto, Ontario, Canada M5S 1A8 c Department of Physiology, University of Toronto, 1 King’s College Circle, Toronto, Ontario, Canada M5S 1A8 Received 4 March 2002; received in revised form 14 August 2002; accepted 2 October 2002

Abstract Klein (Arch Gen Psychiatry, 50, 1993, 306–317) has suggested that panic disorder patients have a false suffocation alarm that may be associated with a lowered threshold for carbon dioxide detection. We compared the thresholds and sensitivities of the central and peripheral chemoreflexes between panic disorder patients and age- and sex-matched healthy volunteers to test this aspect of the hypothesis. We used a modified version of Read’s rebreathing technique in 11 panic disorder patients and 10 healthy volunteers to examine the peripheral and central chemoreflex characteristics in these two populations. Subjects were examined during three rebreathing tests: training, hyperoxic (central chemoreflex alone) and hypoxic (combined central and peripheral chemoreflex). Panic symptoms were retrospectively assessed between groups using a DSM-IV derived Panic Symptom Scale. Comparisons of panic disorder patients with agoraphobia and healthy volunteers showed no significant differences in sensitivities or thresholds. Klein’s hypothesis is not supported by these data. If a false suffocation alarm exists, its triggering may not be implemented within the respiratory chemoreflexes. 䊚 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Rebreathing; Threshold; Sensitivity; Suffocation alarm; Klein

1. Introduction Panic disorder (PD) is a common and debilitating disease characterized by the presence of spontaneous panic attacks (Wittchen and Essan, 1993; American Psychiatric Association, 1994.) *Corresponding author. Tel.: q416-535-8501; fax: q416979-6853. E-mail address: [email protected] (M.A. Katzman).

Although successful treatments for this condition are now available, the underlying pathophysiology of the illness remains elusive. In laboratory studies, several agents have been used to induce panic, including lactate and bicarbonate infusions, and carbon dioxide (CO2) inhalations. Indeed, the success of CO2 in inducing panic in a number of studies (Lousberg et al., 1988; Papp et al., 1993a) gave rise to the ‘CO2 hypersensitivity theory’ of panic. Upon inhalation of a hypercapnic gas mix-

0165-1781/02/$ - see front matter 䊚 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 1 6 5 - 1 7 8 1 Ž 0 2 . 0 0 2 3 8 - X

182

M.A. Katzman et al. / Psychiatry Research 113 (2002) 181–192

ture (5–35% CO2), PD patients are believed to experience cognitive symptoms (going crazy, losing control, dying) and somatic symptoms (heart palpitations, chest pain, dizziness) and a subjective sense of dyspnea that results in a panic attack. Successful induction of panic with biological agents like CO2 has, in part, been responsible for the conviction that a biological flaw is the cause of panic initiation (e.g. Papp et al., 1993a,b; Gorman et al., 2000). However, others believe that PD patients have a cognitive disposition to panic in the context of aversive sensations (e.g. Ley, 1989, 1992; Clark, 1986, 1988; Rapee, 1986; Barlow, 1988; Salkovskis and Clark, 1990). These models propose that panic is created when benign physical symptoms are catastrophically misinterpreted. Thus, the wide range of biochemical panicprovoking agents, such as CO2, are believed to produce panic through their ability to elicit physical sensations that can be misinterpreted and catastrophised as panic, rather then via some specific biological site. Other popular psychobiosocial theories of panic include anxiety sensitivity, where the fear of anxiety-related symptoms, such as those induced by a panicogen like CO2, arises from beliefs that they have harmful consequences. This sensitivity is thought to amplify anxiety responses, thus facilitating fear conditioning of aversive stimuli, and is believed to be a risk factor for the development of anxiety disorders and PD in particular (Reiss, 1991; Reiss and McNally, 1985; McNally, 1994). In addition, learning theories of panic suggest that conditioned stimuli (interoceptive somatic sensations or exteroceptive cues) in conjunction with panic lead to learning of the conditioned stimuli allowing these to precipitate panic when the stimuli are encountered in any other future setting (see Bouton et al., 2001 for an excellent review). Within this context, CO2-mediated elicitation of panic is the result of panicogeninduced interoceptive cues triggering a learned response via association. Yet, a number of other reports confirm the presence and relevance of respiratory abnormalities in patients with PD (Gorman et al., 1988; Woods et al., 1986; Griez et al., 1987; Carr et al., 1987; Sanderson et al., 1989). Based on this research, Klein (1993) formulated

an integrative ‘false suffocation alarm’ theory of panic. Klein (1993) argues that panic patients are hypersensitive to CO2, with blood CO2 levels lower than those of healthy controls, so that when this threshold level is exceeded, ‘the brain’s suffocation monitor erroneously signals a lack of useful air, thereby maladaptively triggering an evolved suffocation alarm system’ that manifests itself as a panic attack (PA). The monitor of the partial pressure of CO2 (PCO2) is thus considered the physiological mechanism for detecting potential suffocation. In a PD patient, this ‘evolutionarily derived set-point has become dysfunctional,’ and the resultant hypersensitivity becomes the instigator of spontaneous panic attacks in a variety of common situations in life where one’s CO2 levels would slightly rise. The chemosensitivity of clinical panickers has been determined using both steady state (MilicEmili and Grunstein, 1976; Bailey et al., 1986; Papp et al., 1989) and rebreathing (Carr et al., 1987; Lousberg et al., 1988; Pain et al., 1988; Papp et al., 1990, 1995) methods. However, these studies have produced mixed results that do not unequivocally support the notion of an enhanced CO2 sensitivity in panickers (Papp and Gorman, 1990). This lack of concordance has been attributed to differences in data analysis and methodology, including the anxiogenic effects of the rebreathing apparatus (Papp et al., 1989, 1995). However, these previous studies were limited for other reasons. First they were confined to the central chemoreflex and have not tested the sensitivity of the peripheral chemoreflex. The peripheral chemoreflex is both a CO2 and oxygen (O2) sensor and as such is an important sensor of suffocation and thus of any ‘suffocation alarm.’ Moreover, none of the techniques employed have been able to measure the chemoreflex threshold, which may be of greater relevance to panic disorder. In this study we use a modified version of the Read (1967) rebreathing technique, a well-known technique in respiratory physiology studies (e.g. Duffin and McAvoy 1988; Baker et al., 1996; Mohan et al., 1999; Vovk et al., 2000) but as yet unemployed in studies of panic etiology, to assess the central and peripheral chemoreflexes in terms

M.A. Katzman et al. / Psychiatry Research 113 (2002) 181–192

of both sensitivity and threshold to CO2. The technique employed herein includes rebreathing against a hyperoxic background, withdrawing peripheral chemoreflex contributions to ventilation (Torrence, 1996) and allowing an independent measure of central chemoreflex responsiveness. Furthermore, prior hypocapnia allows a true measure of the chemoreflex threshold. Thus, this method is the only accurate measure of the chemoreflex threshold as well as the independent central and peripheral chemoreflex contributions to the combined chemoreflex sensitivity response to CO2 (see Mohan and Duffin, 1997). Using a carefully diagnosed group of PD patients and controls, we compare the sensitivities and thresholds of their central and peripheral chemoreflexes in an attempt to identify a possible physiological locale for the ‘suffocation alarm,’ the respiratory sensors of suffocation. These measures attempt to test the validity of Klein’s theory by investigating the likely physiological sensors of suffocation, the only respiratory sensors of CO2mediated suffocation. In addition, we rate the frequency of panic attacks during the procedure. If the chemoreceptors are pathological and responsible for the onset of panic, according to Klein’s theory, we should expect them to display a lower thereshold andyor elevated CO2 sensitivity in PD. 2. Methods 2.1. Subjects Eleven PD patients (5 men, 6 women), with or without agoraphobia (3 non-agoraphobics), participated in the study. They were compared with 10 age- and sex-matched healthy volunteers (5 men, 5 women). All PD patients met criteria for panic disorder, with or without agoraphobia, according to DSM-IV criteria (American Psychiatric Association, 1994). None of the control subjects met criteria for any psychiatric diagnoses, as established by standard psychiatric interviews and confirmed with semi-structured clinical interviews (SCID). As part of the ad-hoc assessment, subjects were asked to complete an anxiety and panic attack diary for 1 week prior to testing.

183

All subjects were in good physical health, as determined by history, laboratory hematology and chemistry, urinalysis, EKG and medical examination. None of the participants were taking any medication and all gave informed consent for the experiments after the Centre for Addiction and Mental Health Research Committee for human experimentation approved the protocol. None of the subjects had been born at high altitudes or were smokers. 2.2. Protocol Subjects were instructed to refrain from food or drink for 9 h, and to avoid caffeine and alcohol for 48 h, prior to testing. On the morning of the test day, the procedure, as outlined in the consent form, was explained again, with a reiteration of the possible physical symptoms that may transiently be experienced as a result of rebreathing. Subjects then underwent a series of three rebreathing tests: a practice test, a hyperoxic test, and a final hypoxic test. Comparison of hypoxic and hyperoxic measurements determined if changes had occurred in peripheral or central chemoreflexes or both, making assumptions discussed in Duffin et al. (2000). Hyperoxic tests generate a net central chemoreflex response wthe peripheral chemoreflex response to CO2 is silenced by hyperoxia; Torrence, 1996)x; hypoxic tests generate a combined central and peripheral chemoreflex response. Differences between experimental groups during the hyperoxic or hypoxic rebreathing tests represent underlying differences in central or peripheral chemoreflex responsiveness to CO2, respectively. The order of the rebreathing tests (hyperoxic, then hypoxic) was applied in a single-blinded fashion. The inter-test interval was 30 min. 2.3. Rebreathing Before rebreathing, subjects voluntarily hyperventilated room air for 5 min, while coached to maintain end-tidal partial pressure of carbon dioxide (PETCO2) between 19 and 25 mmHg, thus avoiding the potential panic symptomatology that accompanies more extreme hypocapnia (Hornsveld et al., 1995). This addition to rebreathing lowered

184

M.A. Katzman et al. / Psychiatry Research 113 (2002) 181–192

subjects’ PETCO2 below chemoreflex threshold of response to CO2, allowing for assessment of this chemoreflex characteristic. Following an expiration, subjects were switched to the rebreathing bag and took three deep breaths to ensure the PCO2 in the bag; lungs and arterial blood quickly equilibrated to the mixed venous PCO2. The latter estimates the PCO2 at the central chemoreceptor. A plateau in PETCO2 at the start of rebreathing evidenced adequate equilibration. The rebreathing bag initially contained a PCO2 of 42 mmHg and a partial pressure of oxygen (Po2) of 150 mmHg for the hyperoxic rebreathing test, or a PO2 of 50 mmHg for the hypoxic rebreathing test. During rebreathing, PO2 was maintained constant (iso-oxic) by the addition of oxygen to the rebreathing bag under computer control. Rebreathing was terminated either when ventilation exceeded 100 lymin. or PETCO2 exceeded 65 mmHg. During each rebreathing test, subjects indicated the onset of any symptoms experienced by an upward hand-swing. Cessation of symptoms was indicated by a downward hand-swing, if it occurred prior to completion of rebreathing. Following each rebreathing test, subjects’ symptoms were retrospectively assessed with the use of a DSM-IV derived Panic Symptom Scale (PSS; Koszycki et al., 1991) with severity of symptoms ranging from 0 (not present) to 5 (extremely severe). A panic attack was defined as the presence of four symptoms with a score of greater than 1 including the ‘anxiety, fear andyor apprehension’ measure, along with the subject’s confirmation (yesyno) that the experienced panic attack was similar to the usual panic attack experience (Koszycki et al., 1991). 2.4. Apparatus The rebreathing apparatus was described previously by Mohan and Duffin (1997). Briefly, subjects wore nose clips throughout the experiment and breathed through a mouthpiece connected to one side of a wide-bore Y valve (Collins P-319, 80-ml dead space) that allowed them to switch from room air to the rebreathing bag. The 5-l rebreathing bag was enclosed in a rigid container with a 50-mm diameter tube connected to a dry

rolling-seal spirometer (Morgan, Spiroflow Model 130). A sample flow of 90 mlymin from the mouthpiece side of the Y valve (gas sample tube UD 5037, Bruel and Kjaer) permitted continuous analysis of carbon dioxide and oxygen. Carbon dioxide and oxygen were analyzed with an anesthetic gas monitor with resolution of 1 and 5 mmHg, respectively (Bruel and Kjaer, type 1304). The sampled gas was returned to the rebreathing bag. Iso-oxia was maintained by a flow of oxygen to the mouthpiece side of the Y valve, under computer control. A 16-bit analog to digital converter (National Instruments, AT-MIO16XE-50) digitized the analog signals for on-line computer analysis using specially written software (National Instruments, LabVIEW). The software calculated tidal volumes, inspiratory and expiratory times, ventilation, PETCO2 and PETO2 on a breath-by-breath basis. The volume changes and PETCO2 and PETO2 were also written on a chart recorder (Graphtec, Lineacorder Mark VII WR 3101) to monitor the initial rebreathing equilibration. This verified proper equilibration of PCO2 between bag, lungs arterial blood and mixed venous PCO2, thus allowing PCO2 monitoring at the central chemoreceptor. The measurement system was calibrated before each experimental session using gases of known concentrations (room air and commercial gas mixture of 5% CO2, 10% O2) and a standardized volume syringe (Hans Rudolph, Model 5540). 2.5. Data analysis Analysis of the breath-by-breath data was accomplished using a spreadsheet (Microsoft, Excel) specially designed for this purpose. Data from the initial equilibration at the start of rebreathing (initial three deep breaths) was excluded from further analysis. Then, plots of ventilation (VE, lymin) tidal volume (VT, ml BTPS) and breathing frequency (f, breathsymin) were plotted against end-tidal PCO2 (mmHg.) These plots were analyzed by fitting a best-fit line composed of a sum of three possible segments to each of ventilation (VE) tidal volume (VT) and frequency of breathing (f) as detailed in Duffin et al. (2000). The first segment representing basal

M.A. Katzman et al. / Psychiatry Research 113 (2002) 181–192

respiratory values was fitted with an exponential decline to account for any short-term potentiation secondary to the voluntary hyperventilation prior to rebreathing. The second segment was a straightline segment from the first breakpoint to the second. If a second breakpoint was evident, a third segment was fitted above the second breakpoint. None of the rebreathing responses collected here generated a second breakpoint. The fits were generated using a commercial fitting program (SPSS, Sigmaplot) that simultaneously calculated the breakpoints, slopes, basal values and exponential constants. The first breakpoint (T) measures the chemoreflex threshold, and the slope (S) above the first breakpoint measures the chemoreflex sensitivity. The basal values (B) measure the drives to breathe other than that from the chemoreflexes (i.e. basaly‘wakefulness’ drive to breathe; Fink, 1961). All ventilatory responses and their components (tidal volume and frequency) were assigned these response characteristics (B, T and S) except for those subjects with Type 3 breathing patterns (Bechbache et al., 1979). Despite linear ventilation responses, these subjects’ tidal volume and frequency responses are scattered such that a linear analysis of them is impossible. As a result, not all subjects had their tidal volume and frequency response characteristics reported; only the ventilation response characteristics are reported for all subjects. 2.6. Statistical analyses For each of the three test responses (VE, VT and f), the three response characteristics (T, S and B) were first separated into those resulting from hypoxic or hyperoxic rebreathing tests. Thus, nine respiratory characteristics were generated: basal values for ventilation (VEb,) tidal volume (VTb) and breathing frequency (f b); threshold values for ventilation (VET) tidal volume (VTT) and breathing frequency (f T); and sensitivity values for ventilation (VES,) tidal volume (VTS) and breathing frequency (f S). These characteristics were compared between PD patients and healthy volunteers (CON). Comparisons were made using unpaired t-tests for normally distributed data with equal variance, or Mann-Whitney tests for non-paramet-

185

ric data with the significance level set at 5% (SigmaStat 2.0, SPSS). The ‘Compare Proportions’ (z-test) was used to compare the frequency of panic attacks between groups. Unless otherwise indicated, all results are reported as mean "S.D. 3. Results 3.1. Panic measures Excluding the practice rebreathing tests, two rebreathing tests were done on each of the 11 PD patients and 10 CON to give 22 and 20 tests per, respective group. They resulted in a total of eight panic attacks (36.4%, 6 subjects) in the PD patient group and two panic attacks (10%, 1 subject) in the CON group. There was no significant difference in the number of subjects experiencing panic attacks between the two groups. 3.2. Rebreathing test results Representative rebreathing test responses for a PD patient and a healthy volunteer are shown in Figs. 1 and 2, respectively. Comparisons of chemoreflex characteristics for the different subject groups, separated on the basis of iso-oxia, are made in Tables 1 and 2 (hyperoxic and hypoxic, respectively). There were no significant differences in the chemoreflex thresholds for ventilation (VET) tidal volume (VTT) or breathing frequency (f T) between patients and healthy volunteers for either the hyperoxic tests or the hypoxic tests. Neither was chemoreflex sensitivity (VES, VTS, f S) significantly different between these groups for either the hyperoxic tests or the hypoxic tests. In addition, basal values of ventilation (VEb,) tidal volume (VTb,) and breathing frequency (f b) were not significantly different between these groups for the hyperoxic or hypoxic tests. 4. Discussion To our knowledge, this study is the first to measure central and peripheral chemoreflex threshold differences between PD patients and control subjects. Our finding of no significant differences

186

M.A. Katzman et al. / Psychiatry Research 113 (2002) 181–192

Fig. 1. Representative panic patient’s responses. Points are breath-by-breath values measured during modified rebreathing conditions of slowly rising CO2 concentrations against an iso-oxic background. The plots are best-line fits generated using Sigmaplot employing the Levenberg-Marquardt algorithm of diminishing sum of squared residual to attain convergence.

in the chemoreflexes of PD patients from those of healthy volunteers has important implications for the false suffocation alarm theory. One possible explanation for these findings is a type II error due to the relatively small sample size used in our study. Indeed, the power analysis presented (Tables 1 and 2) does demonstrate this a noteworthy possibility not to be underestimated, although the effect size demonstrated (Cohen’s d) is also negligible, suggesting that any differences

in chemoreflex variables not detected by low power were nonetheless insignificant, particularly in regards to the key defining chemoreflex variables (VET, VES). Furthermore, throughout the data collection process, we observed a trend of diminishing differences in the chemoreflex thresholds and sensitivities between patients and controls as the data accumulated. Also, previous research in the respiratory physiology literature assessing chemoreflex responsiveness (e.g. Berkenbosch et al.,

M.A. Katzman et al. / Psychiatry Research 113 (2002) 181–192

187

Fig. 2. Representative healthy volunteer’s responses. Points are breath-by-breath values measured during modified rebreathing conditions of slowly rising CO2 concentrations against an iso-oxic background. The plots are best-line fits generated using Sigmaplot employing the Levenberg–Marquardt algorithm of diminishing sum of squared residual to attain convergence.

1992; Dahan et al., 1996; Fatemian and Robbins, 1998; Pedersen et al., 2000) suggests that our sample size was to determine significant differences, had any been present. Another possible explanation is that the technique we employed masked the differences—the relatively low percentage of panic attacks (36%) observed may be taken as evidence of this. However, we do not believe this to be the case, rather the opposite. For example, past suggestions were

that hyperoxia in a Read rebreathing test may blunt the ventilatory response to CO2, and the high starting CO2 level in the bag may hide differences between patients and control subjects (Papp et al., 1989, 1993a, 1995). We addressed the former issue by using both hypoxic and hyperoxic iso-oxic rebreathing tests so that both peripheral and central chemoreflexes were measured. The issue of a high starting CO2 level in the bag was addressed by our use of voluntary hyper-

188

M.A. Katzman et al. / Psychiatry Research 113 (2002) 181–192

Table 1 Results of comparisons of respiratory measures for the hyperoxic condition between panic disorder patients and controls PD (ns11) Mean (S.D.)

CON (ns10) Mean (S.D.)

PD vs. CON t-test P-value Cohen’s d CId (upper-lower) Power of test

VET (mmHg)

43.6 (4.9) ns11

43.7 (3.2) ns10

VTT (mmHg)

43.0 (4.9) ns10

42.4 (3.8) ns10

fT (mmHg)

44.1 (5.3) ns7

45.8 (3.1) ns7

VES (lyminymmHg)

1.6 (0.6) ns11

1.7 (0.7) ns10

VTS (mlymmHg)

73.6 (27.1) ns10

75.6 (34.2) ns10

fS (byminymmHg) VEb (lymin)

0.74a ns7 9.3 (3.7)

0.64a ns7 8.9 (5.9)

VTb (ml) fb (bymin)

622.9a ns10 8.5 (6.2) ns10

555.0a ns10 12.2 (6.5) ns8

Ps0.972 y0.02 0.14–1.46 0.05 Ps0.785 0.14 y0.75–1.01 0.06 Ps0.467 y0.39 y1.42–0.69 0.10 Ps0.808 y0.15 y1.01–0.71 0.06 Ps0.888 y0.06 y0.94–0.81 0.05 Ps1.0a NyA Ps0.858 0.08 y0.78–0.94 0.05 Ps0.571a NyA Ps0.234a y0.58 y1.5–0.39 0.21

a Reported values are medians obtained using the Mann–Whitney test for non-parametric data. Cohen’s dsCohen’s d coefficient for effect size; CId sConfidence intervals for Cohen’s d; PDspanic disorder; CONscontrol subjects; VET sventilation threshold; VTTstidal volume threshold; f Tsfrequency threshold; VES sventilation sensitivity; VTS stidal volume sensitivity; f Ssfrequency sensitivity; VEbsbasal ventilation; VTbsbasal tidal volume; f bsbasal frequency.

ventilation prior to rebreathing to lower the starting CO2 level in the bag below that used in a normal Read rebreathing test. This modification of the Read rebreathing test has the added advantage of allowing assessment of the true chemoreflex threshold of response to CO2. Although previous studies have examined the sensitivity of the response to CO2 in PD patients (Woods et al., 1986; Lousberg et al., 1988; Zandbergen et al.,

1991), they have not directly measured the threshold; extrapolating the ventilatory slope to the PCO2 axis does not give a measure of the chemoreflex threshold of response to CO2 because the intercept is also dependent on the sensitivity. We also point out that the voluntary hyperventilation subjects used prior to rebreathing is not itself panicogenic (Rapee et al., 1986; Gorman et al., 1988; Griez et al., 1988; Bass et al., 1989;

M.A. Katzman et al. / Psychiatry Research 113 (2002) 181–192

189

Table 2 Results of comparisons of respiratory measures for the hypoxic condition between panic disorder patients and controls PD (ns11) Mean (S.D.)

CON (ns10) Mean (S.D.)

PD vs. CON t-test P-value Cohen’s d CId (upperlower) power of test

VET (mmHg)

41.1 (3.6) ns10

40.0 (3.5) ns10

VTT (mmHg)

39.6 (2.5) ns10

39.1 (3.6) ns10

fT (mmHg)

41.1 (3.4) ns9

41.0 (2.6) ns7

VES (lyminymmHg)

2.9 (1.0) ns10

2.8 (1.6) ns10

VTS

106.4

92.4

(mlymmHg) fS (byminymmHg)

(28.4) ns10 0.85 (0.56) ns9

(22.7) ns10 0.80 (0.53) ns8

VEb (lymin)

8.9 (5.7) ns11

8.3 (2.4) ns9

VTb (ml)

715.4 (503.7) ns11

574.3 (281.3) ns9

fb (bymin)

9.5 (6.0) ns11

12.4 (7.9) ns9

Ps0.508 0.31 y0.58–1.18 0.10 Ps0.705 0.16 y0.72–1.03 0.06 Ps0.909 0.03 y0.96–1.02 0.05 Ps0.896 0.07 y0.81–0.95 0.05 Ps0.233 0.54 y0.37–1.41 0.21 Ps0.849 0.09 y0.87–1.04 0.05 Ps0.769 0.14 y0.76–1.01 0.06 Ps0.464 0.41 y0.56–1.21 0.19 Ps0.357 y0.41 y1.29–0.49 0.14

Reported values are medians obtained using the Mann–Whitney test for non-parametric data. Cohen’s dsCohen’s d coefficient for effect size; CIdsConfidence intervals for Cohen’s d; PDspanic disorder; CONscontrol subjects; VETsventilation threshold; VTTstidal volume threshold; f Tsfrequency threshold; VES sventilation sensitivity; VTS stidal volume sensitivity; f Ssfrequency sensitivity; VEbsbasal ventilation; VTbsbasal tidal volume; f bsbasal frequency.

Zandbergen et al., 1990), and we employed a practice rebreathing test to relieve any anticipatory anxiety associated with the complete test procedure. Though we cannot be certain that anticipatory anxiety was completely eliminated, previous

work has shown that 10 min is sufficient time for laboratory adaptation with PD patients; in this study subjects had a minimum of a half-hour acclimatization (Beck and Berisford, 1992; Beck and Shipherd, 1997). Indeed, we suggest that the

190

M.A. Katzman et al. / Psychiatry Research 113 (2002) 181–192

relatively low percentage of panic attacks (36%) in our study could be the result of the gradual rise of CO2 employed with our method, with rebreathing begun below resting CO2 levels. Furthermore, the level of hypocapnia maintained during the hyperventilation phase (19–25 mmHg) is below that believed to induce significant symptomatology (though not panic; Hornsveld et al., 1995). Previous studies have examined the response to CO2 against a hyperoxic background, which minimizes the peripheral chemoreflex response (Torrence, 1996). Since hyperoxia markedly reduces the peripheral chemoreflex contribution to ventilation, our hyperoxic rebreathing tests measured the threshold and sensitivity to CO2 mediated by the central chemoreflex. Neither differed between PD and CON groups. The threshold has not been measured before, but others have measured sensitivity with conflicting findings. Some have found that it did not differ significantly between PD and CON (Woods et al., 1986; Pain et al., 1988; Papp et al., 1995), while others found heightened sensitivity (Carr et al., 1987; Lousberg et al., 1988; Zandbergen et al., 1991). In contrast to hyperoxia, the hypoxic rebreathing tests measured ventilatory responses mediated by a combination of central and peripheral chemoreflexes. Any differences between the hypoxic and hyperoxic responses are therefore due to the peripheral chemoreflex. Again we found no differences between PD and CON groups for either thresholds or sensitivities. Neither has been previously examined. In summary, although the PD patients we studied experienced a panic attack in 36% of the rebreathing tests we used to measure their respiratory chemoreflexes, their chemoreflex responsiveness was normal. We therefore conclude that the central and peripheral chemoreceptors, the sole respiratory sensors of suffocation, are not likely to be hypersensitive to CO2 in PD patients. Our preliminary findings do not support the ‘false suffocation alarm hypothesis’ of heightened chemoreflex responsiveness as an explanation for panic attack initiation. We suggest that if a false suffocation alarm exists, it may not lie within the respiratory chemoreflexes; while the respiratory chemoreflexes may partici-

pate in the genesis of a panic attack, the triggering event may lie elsewhere. Acknowledgments We thank Ms Monica Vermani for her assistance in the final stages of study completion. We also thank the Centre for Addiction and Mental Health for funding this research. References American Psychiatric Association, 1994. Diagnostic and Statistical Manual of Mental Disorders, 4th edition. American Psychiatric Press, Washington, DC. Bailey, P.L., Andriano, K.P., Goldman, M., Stanley, T.H., Pace, N.L., 1986. Variability of the respiratory response to diazepam. Anesthesiology 64, 460–465. Baker, J.F., Goode, R.C., Duffin, J., 1996. The effect of a rise in body temperature on the central-chemoreflex ventilatory response to carbon dioxide. European Journal of Applied Physiology and Occupational Physiology 72, 537–541. Barlow, D.H., 1988. Current models of panic disorder and a view from emotion theory. In: Frances, A.J., Hales, R.E. (Eds.), American Psychiatric Press Review of Psychiatry, Vol. 7. American Psychiatric Press, Washington, DC. Bass, C., Lelliott, P., Marks, I., 1989. Fear talk vs. voluntary hyperventilation in agoraphobics and normals: a controlled study. Psychological Medicine 19, 669–676. Bechbache, R.R., Chow, H.H.K., Duffin, J., Orsini, E.C., 1979. The effects of hypercapnia, hypoxia, exercise and anxiety on the pattern of breathing in man. Journal of Physiology 293, 285–300. Beck, J.G., Berisford, M.A., 1992. The effects of caffeine on panic patients: response components of anxiety. Behaviour Therapy 23, 405–422. Beck, J.G., Shipherd, J., 1997. Repeated exposure to interoceptive cues: does habituation of fear occur in panic disorder patients? Behaviour Research and Therapy 35, 551–557. Berkenbosch, A., Dahan, A., DeGoede, J., Olievier, I.C., 1992. The ventilatory response to CO2 of the peripheral and central chemoreflex loop before and after sustained hypoxia in man. Journal of Physiology (London) 456, 71–83. Bouton, M.E., Mineka, S., Barlow, D.H., 2001. A modern learning theory perspective on the etiology of panic disorder. Psychological Review 108, 4–32. Carr, D.B., Fishman, S.M., Systrom, D., Beckett, A., Sheehan, D.V., Rosenbaum, J.F., 1987. CO2 in panic disorder and premenstrual normals. Paper presented at the 140th annual meeting of the American Psychiatric Association. Chicago. Clark, D.M., 1986. A cognitive approach to panic. Behaviour Research and Therapy 24, 461–471. Clark, D.M., 1988. A cognitive model of panic attacks. In: Rachman, S., Maser, J. (Eds.), Panic: Psychological Perspectives. Erlbaum, Hillside, NJ.

M.A. Katzman et al. / Psychiatry Research 113 (2002) 181–192 Dahan, A., van Kleef, J., van den Elsen, M., Valk, R., Berkenbosch, A., 1996. Slow ventilatory dynamics after isocapnic hypoxia and voluntary hyperventilation in humans: effects of isoflurane. British Journal of Anaesthesia 76, 374–381. Duffin, J., McAvoy, G.V., 1988. The peripheral-chemoreceptor threshold to carbon dioxide in man. Journal of Physiology 406, 15–26. Duffin, J., Mohan, R., Vasiliou, P., Stephenson, R., Mahamed, S., 2000. A model of the chemoreflex control of breathing in humans: model parameter measurement. Respiration Physiology 120, 13–26. Fatemian, M., Robbins, P.A., 1998. Human ventilatory response to CO2 after 8 h of isocapnic or poikilocapnic hypoxia. Journal of Applied Physiology 85, 1922–1928. Fink, B.R., 1961. Influence of cerebral activity in wakefulness on regulation of breathing. Journal of Applied Physiology 16, 15–20. Gorman, J.M., Fyer, M.R., Goetz, R., Askanazi, J., Liebowitz, M.R., Fyer, A.J., Kinney, J., Klein, D.F., 1988. Ventilatory physiology of patients with panic disorder. Archives of General Psychiatry 45, 31–39. Gorman, J.M., Kent, J.M., Sullivan, G.M., Coplan, J.D., 2000. Neuroanatomical hypothesis of panic disorder, revised. American Journal of Psychiatry 157, 493–505. Griez, E., Lousberg, H., van den Hout, M.A., van der Molen, M., 1987. CO2 vulnerability in panic disorder. Psychiatry Research 20, 87–95. Griez, E., Zandbergen, J., Lousberg, H., van den Hout, M., 1988. Effects of low pulmonary CO2 on panic anxiety. Comprehensive Psychiatry 29, 490–497. Hornsveld, H., Garssen, B., van Spiegel, P., 1995. Voluntary hyperventilation: the influence of duration and depth on the development of symptoms. Biological Psychology 40, 299–312. Ley, R., 1989. Dyspneic fears and catastrophic cognitions in hyperventilatory panic attacks. Behaviour Research and Therapy 27, 549–554. Ley, R., 1992. The many faces of panic: psychological and physiological differences among three types of panic. Behaviour Research and Therapy 30, 347–357. Klein, D.F., 1993. False suffocation alarms, spontaneous panics, and related conditions. Archives of General Psychiatry 50, 306–317. Koszycki, D., Bradwejn, J., Bourin, M., 1991. Comparison of the effects of cholecystokinin-tetrapeptide and carbon dioxide in health volunteers. European Neuropsychopharmacology 1, 137–141. Lousberg, H., Griez, E., van den Hout, M.A., 1988. CO2 chemosensitivity in panic disorder. Acta Psychiatrica Scandinavica 77, 214–218. McNally, R.J., 1994. Panic Disorder: A Critical Analysis. Guilford, New York. Milic-Emili, J., Grunstein, M.M., 1976. Drive and timing components of ventilation. Chest 70, 131–133.

191

Mohan, R., Duffin, J., 1997. The effect of hypoxia on the ventilatory response to carbon dioxide in man. Respiration Physiology 108, 101–115. Mohan, R.M., Amara, C.E., Cunningham, D.A., Duffin, J., 1999. Measuring central-chemoreflex sensitivity in man: rebreathing and steady-state methods compared. Respiration Physiology 115, 23–33. Pain, M.C.F., Biddle, N., Tiller, J.W.G., 1988. Panic disorder, ventilatory response to CO2’s respiratory variables. Psychosomatic Medicine 50, 541–548. Papp, L.A., Goetz, R., Cole, R., Liebowitz, M.R., Fyer, A.J., Klein, D.F., Jordan, F., Liebowitz, M.R., Fyer, A.J., Hollander, E., Gorman, J.M., 1989. Hypersensitivity to CO2 in panic disorder. American Journal of Psychiatry 146, 779–782. Papp, L.A., Gorman, J.M., 1990. Respiratory neurobiology of panic. Paper presented at the annual meeting of the American College of Neuropsychopharmacology, San Juan, P.R. Papp, L.A., Klein, D.F., Gorman, J., 1993b. Carbon dioxide hypersensitivity, hyperventilation and panic disorder. American Journal of Psychiatry 150, 1149–1157. Papp, L.A., Klein, D.F., Martinez, J., Schneier, F., Cole, R., Liebowitz, M.R., Hollander, E., Fyer, A.J., Jordan, F., Gorman, J.M., 1993a. Diagnostic and substance specificity of carbon-dioxide induced panic. American Journal of Psychiatry 150, 250–257. Papp, L.A., Martinez, J.M., Klein, D.F., Coplan, J.D., Gorman, J.M., 1995. Rebreathing tests in panic disorder. Biological Psychiatry 38, 240–245. Pedersen, M.E., Robach, P., Richalet, J.P., Robbins, P.A., 2000. Peripheral chemoreflex function in hyperoxia following ventilatory acclimatization to altitude. Journal of Applied Physiology 89, 291–296. Rapee, R., 1986. Differential response to hyperventilation in panic disorder and generalized anxiety disorder. Journal of Abnormal Psychology 95, 24–28. Rapee, R., Mattick, R., Murrell, E., 1986. Cognitive mediation in the affective component of spontaneous panic attacks. Journal of Behavioral Therapy and Experimental Psychiatry 17, 245–253. Read, D.J.C., 1967. A clinical method for assessing the ventilatory response to CO2. Australian Annals of Medicine 16, 20–32. Reiss, S., 1991. The expectancy model of fear, anxiety and panic. Clinical Psychology Review 11, 141–153. Reiss, S., McNally, R.J., 1985. Expectancy model of fear. In: Reiss, S., Bootzin, R.R. (Eds.), Theoretical Issues in Behavior Therapy. Academic Press, San Diego, CA, pp. 107–121. Salkovskis, P.M., Clark, D.M., 1990. Affective responses to hyperventilation: a test of the cognitive model of panic. Behaviour Research and Therapy 28, 51–61. Sanderson, W.C., Rapee, R.M., Barlow, D.H., 1989. The influence of an illusion of control on panic attacks induced via inhalation of 5.5% carbon dioxide enriched air. Archives of General Psychiatry 46, 157–162. Torrence, R.W., 1996. Chemoreception upstream of transmitters. In: Zapata, P., Eyzaguirre, C., Torrence, R.W. (Eds.),

192

M.A. Katzman et al. / Psychiatry Research 113 (2002) 181–192

Frontiers of Arterial Chemoreception. New York, Plenum Press, pp. 13–38. Vovk, A., Duffin, J., Kowalchuk, J.M., Paterson, D.H., Cunningham, D.A., 2000. Changes in chemoreflex characteristics following acute carbonic anhydrase inhibition in humans at rest. Experimental Physiology 85, 847–856. Wittchen, H.U., Essan, C.A., 1993. Epidemiology of panic disorder: progress and unresolved issues. Journal of Psychiatric Research 27, 47–68.

Woods, S.W., Charney, D.S., Loke, W.K., Goodman, G.R., Redmond, D.E., Heninger, G.R., 1986. Carbon dioxide sensitivity in panic anxiety. Archives of General Psychiatry 43, 900–909. Zandbergen, J., Lousberg, H., Pols, H., de Loof, C., Griez, E.J.L., 1990. Hypercarbia vs. hypocarbia in panic disorder. Journal of Affective Disorders 18, 75–81. Zandbergen, J., Pols, H., de Loof, C., Griez, E.J.L., 1991. Ventilatory response to CO2 in panic disorder. Psychiatry Research 39, 13–19.