Techniques and MethodsThe 35% CO2 inhalation procedure: test–retest reliability
Introduction
Procedures that provoke symptoms of a specific illness have obvious value as tools for the study of that illness and may have important clinical uses as well. Among the mental disorders such procedures have probably been most widely applied in the study of panic disorder. These have included hyperventilation, the infusion of lactate, and the inspiration of carbon dioxide, either at low concentrations in rebreathing apparati, or at higher concentrations in single breaths.
The last of these procedures is likely to come into wider use as a research tool, due in part to its simplicity and, in part, to its apparent sensitivity and specificity to panic disorder. The proportion of individuals who develop brief panic attacks following a single breath of 35% CO2 has robustly separated patients with unmedicated panic disorder from well control subjects van den Hout and van der Molen 1987, Gorman et al 1990, Perna et al 1994a, Caldirola et al 1997 and from patients with major depressive disorder (Perna et al 1995a), obsessive–compulsive disorder Griez et al 1990, Perna et al 1995b, and generalized anxiety disorder (Verburg et al 1995). Moreover, medications effective in the treatment of panic disorder have blocked the panicogenic effects of 35% CO2Pols et al 1991, Sanderson et al 1994, Perna et al 1994b. This effect manifests soon after treatment with tricyclic antidepressants or serotonin reuptake inhibitors begins, well before the antipanic effect of these drugs would be expected Bertani et al 1997, Perna et al 1997.
The subjective response to CO2 inhalation also shows promise as a marker for risk status. Two studies have shown that relatives of panic disorder patients are more likely to experience a panic attack after a single breath of 35% CO2 than are individuals without a family history of panic disorder Perna et al 1995c, Coryell 1997. Another found that panic disorder probands who respond to 35% CO2 inhalation with panic attacks had higher familial loadings for panic disorder than did probands with negative test results (Perna et al 1996). Further evidence that the subjective response to doses of CO2 is genetically influenced derive from a recent twin study in which monozygotic twins were significantly more concordant for a positive test result than were dizygotic twins (Bellodi et al 1998).
The potential utility of this procedure clearly warrants attention to its reliability. To our knowledge, there has been no formal effort to quantify the stability over time of test results. The following does this with additional data from a previously described study of never-ill individuals at high risk for panic disorder (Coryell 1997). Subjects who had a first-degree relative treated for panic disorder, but who lacked a personal history of panic attacks, were substantially more likely to report panic attacks after CO2 inhalation than were either never-ill subjects with a family history of affective disorder, or never-ill subjects with a family history of neither condition. Those subjects underwent the same inhalation procedure after variable intervals, but the results of repeat testing were not described in the earlier report. The following provides these results with a focus on the reproducibility of findings.
Section snippets
Subjects
Advertisements were used to recruit subjects into three groups. Inclusion criteria required that all participants were 18–34 years of age, lacked a personal history of any Research Diagnostic Criteria disorder (Spitzer et al 1978), had never experienced panic attacks, and had never sought professional help for emotional problems. Subjects considered to be at high risk for panic disorder (HR-PD) described at least one first-degree relative who met DSM-III-R criteria for panic disorder and who
Results
The subjects at high risk for affective disorder did not differ from the low-risk control subjects by post-CO2 symptoms scores on either testing occasion. For simplicity these two groups are combined in the following analyses and designated simply as “control subjects.” Twelve HR-PD subjects and 31 control subjects completed the initial inhalation procedure, and none of these failed to return for the repeat procedure. Two control subjects (both at high risk for affective disorder) and 1 HR-P
Discussion
The central findings generated by the first session of testing were clearly reproducible in the second session, yet the within-individual reliability of categorical test results was poor. This instability appeared to arise from variations in sensitivity rather than changes in specificity. Thus, a positive test result with either session was a valid indicator of high-risk status, whereas a negative result in either session was not a strong sign of low-risk status. If replicated, these findings
Acknowledgements
This work was supported by NIMH grant MH 56132.
References (22)
- et al.
The 35% CO2 challenge in panic disorderOptimization by receiver operating characteristic (ROC) analysis
J Psychiatry Res
(1995) - et al.
The 35% CO2 challenge test in patients with social phobia
Psychiatry Res
(1997) Hypersensitivity to carbon dioxide as a disease-specific trait marker
Biol Psychiatry
(1997)- et al.
High-dose carbon dioxide challenge test in anxiety disorder patients
Biol Psychiatry
(1990) - et al.
CO2 inhalation in the treatment of panic attacks
Behav Res Ther
(1986) - et al.
Specific sensitivity of patients with panic attacks to carbon dioxide inhalation
Psychiatry Res
(1990) - et al.
Carbon dioxide/oxygen challenge test in panic disorder
Psychiatry Res
(1994) - et al.
35% CO2 challenge in panic and mood disorders
J Affect Disord
(1995) - et al.
The family history method using diagnostic criteria
Arch Gen Psychiatry
(1977) - et al.
CO2-induced panic attacksA twin study
Am J Psychiatry
(1998)