Imagined Risk of Suffocation as a Trigger for Hyperventilation ILSE VAN DIEST,PHD, STEVEN DE PEUTER, MA, STEPHAN DEVRIESE,PHD, ELKE WELLENS, MA, KAREL P. VAN DE WOESTIJNE, MD, AND OMER VAN DEN BERGH,PHD Objective: Although hyperventilation has been hypothesized to play a role in many pathologies, its critical triggers remain poorly understood. The present experiment aimed to test whether stronger hyperventilation responses occur in response to suggested risk of suffocation compared with other fearful situations in high- and low-trait anxious women. Methods: Fractional end-tidal CO 2 -concentration (FetCO 2 ), respiratory frequency, and inspiratory volume were measured nonintrusively in high- (n = 24) and low- (n = 24) trait anxious women during imagery of 3 fear, 1 tension, 1 depressive, and 3 relaxation scripts. The fear scripts were equal in ratings of unpleasantness and arousal but differed regarding the inclusion of suggested risk of suffocation and entrapment. After each imagery trial, participants rated the emotional dimensions of pleasantness, arousal, and dominance and the vividness of their imagery. Results: Decreases in FetCO 2 occurred in all fear scripts. High-trait anxious women showed a stronger reduction in FetCO 2 compared with low-trait anxious women during the fear script suggesting risk of suffocation but not during the other fear scripts. This effect was unrelated to any of the self-reported fear ratings. Self-reported fear of entrapment was associated with an overall lower FetCO 2 but not with enhanced reactivity to imagined entrapment. Conclusion: High-trait anxiety is associated with stronger respiratory responsivity to imagined risk of suffocation and may constitute a specific vulnerability factor for the development of panic disorder and claustrophobia. Key words: hyperventilation, trait anxiety, suffocation, fear, entrapment. FetCO 2 = fractional concentration of end-tidal carbon dioxide; f r = respiratory frequency; Vi = inspiratory volume; CPS = checklist for psychosomatic symptoms; STAI-T = trait version of the state-trait anxiety inventory; FS = fear of suffocation; FR = fear of restriction. INTRODUCTION H yperventilation is a breathing pattern that exceeds meta- bolic requirements (1): more CO 2 is being expired than produced by the body, leading to a decreased CO 2 -pressure in the blood, and to a lower fractional end-tidal CO 2 concentra- tion (FetCO 2 ). Although hyperventilation has been hypothe- sized to play a role in many pathologies (panic disorder, chronic fatigue syndrome, multiple chemical sensitivity, chronic pain, somatization; 2–7) and can be considered an important cause for stress-related bodily complaints in general (3), its critical triggers remain poorly understood. Previous research from our group has shown that decreases in FetCO 2 can be triggered during imagery of both pleasant and unpleas- ant high-arousal emotional situations, but not when imagining low arousal scenes (8,9). This is in line with a psychophysi- ological perspective that respiratory reactivity reflects the general action tendency underlying an emotion, independently from specific emotional contents (10 –12). In other words, an increase in ventilation during emotional states may be part of the increased arousal associated with emotions preparing the organism for action, irrespective of the direction of that action tendency (approach versus avoidance/escape). However, given the obvious survival value of a satisfactory air supply, it would be adaptive for an organism to be espe- cially reactive to (internal/external) cues signaling potential suffocation. Several observations are consistent with the ex- istence of a primitive alarm system responding with intense fear and hyperventilation to suffocation cues (13–26). If hy- perventilation is part of an unconditioned fear response to suffocation, situations referring to risk of suffocation should trigger hyperventilation more easily than other fearful situa- tions. As far as we know, this simple prediction has never been tested directly. Therefore, the present study aimed to compare responses in respiration and FetCO 2 during imagery of fearful scripts with and without suffocation cues. Whereas we previously showed (27) that there was no overall association between trait anxiety and end-tidal CO 2 after careful control of confounding variables, we did find indications of stronger respiratory reactivity to mental load challenges in high-trait anxious persons. Interestingly, trait anxiety has also been found to be an important vulnerability factor for panic disorder (28). In addition, because there is some evidence that inter-individual differences in reactivity to suffocation stimuli are related to stress-induced modifications in brain functioning (16,17), it would be interesting to see if partic- ularly high-trait anxious participants show elevated reactivity when confronted with suffocation cues. Therefore, the present study also focused on the vulnerability for hyperventilatory re- sponses of high-anxious compared with low-anxious participants. FetCO 2 and respiration were measured in 48 women during imagery of 8 different scripts. Three fear scripts were critical to address our principal research question: 1 suffocation and 2 control fear scripts. We hypothesized that stronger decreases in FetCO 2 would occur for the fear script containing suffoca- tion cues compared with the other fear scripts and that these effects would be more pronounced in high- as opposed to low-trait anxious participants. In order to investigate whether, similarly to previous find- ings (8,9), decreases in FetCO 2 would occur during imagery of high- but not low-arousal scripts, 4 low-arousal scripts were included as well: 1 depressive and 3 relaxation scripts. A tense-anxiety script was additionally used to explore respira- tory behavior under a more attentive defensive set associated with lower action tendencies (29). METHODS Methods were similar to previous studies from our group (8,9). Therefore, a shortened description will be presented in this paper. From the Department of Psychology, University of Leuven, Leuven, Bel- gium (I.V.D., S.D.P., S.D., E.W., O.V.d.B.); Faculty of Medicine, University of Leuven, Leuven, Belgium (P.V.d.W.). Address correspondence and reprint requests to Ilse Van Diest, PhD, University of Florida, CSEA, 2800 SW. Archer Rd., Surge Bldg. 772, Gaines- ville, FL 32608. E-mail: Ilse.Vandiest@psy.kuleuven.ac.be Ilse Van Diest is a postdoctoral researcher of the Fund of Scientific Research, Flanders, Belgium. Received for publication October 13, 2004; revision received April 19, 2005. DOI: 10.1097/01.psy.0000181275.78903.64 813 Psychosomatic Medicine 67:813– 819 (2005) 0033-3174/05/6705-0813 Copyright © 2005 by the American Psychosomatic Society