- Email: [email protected]
Differential reactivity to lactate infusions: The relative role of biological, psychological, and conditioning variables
BIOL PSYCHIATRY 1989;25:469-4g1
469
Differential Reactivity to Lactate Infusions: The Relative Role of Biological, Psychological, and conditioning Variables Thomas A. Aronson, Patricia Whitaker-Azmitia, and Irene Caraseti
Nine patients with panic disorder experienced a lactate-induced panic attack, whereas nine controls did not. Higher preinfusion anxiety levels and heart rates were associated with panic disorder, ana’ high baseline anxiety ratings were associated with atypical, severe lactate-induced panic attacks. Nevertheless, it was difbcult to reconcile patients’ and co#ro~’ reac~v~~ to lactate as entirely se&o~~ to reline di~eren~es. subjects diflered q~l~t#ively in the types of specific symptoms experienced and q~ntitatively in their anxiety and heart rate responses. In most cases, panic began with various central perceptual changes; peripheral cardiovascular and autonomic symptoms followed later. No patient rated a lactate-induced panic attack as identical to a naturally occurring attack. Not only did specific ~mptoms di#er in their severity and order of production, but lactate-inured panic lacked the typicalfears of ping, going crazy, or losing control. The results suggest that though environmental effects, expectancy biases, and baseline psychological states play salient roles in modifying the experience of a lactate-induced panic attack, they do not fully account for lactate sensitivity. The relative role that biological, psychological, and conditioning factors play in lactate-induced panic is discussed.
Introduction Intravenous sodium lactate has been the most extensively studied of the provocative agents used in the psychobiological study of panic disorder-a growing list that includes yo~bine, clondine, 5% and 35% CO, ideation, caffeine, isop~~~nol, m-chlorophenylpi~~ne, and sodium bicarbonate. Since 1967, numerous reports have shown that approximately 70%-80% of patients with panic disorder or panic disorder with agoraphobia experience a lactate-induced panic attack compared to 0%-10% of normal controls (Pitts and McClure 1967; Bonn et al. 1971; Fink et al. 1971; Kelly et al. 1971; Appleby et al. 1981; Rifkin et al. 1981; Liebowitz et al. 1984; Raney et al. 1984; Dager et al. 1987). However, recent evidence suggests that lactate sensitivity is not specific for panic disorder per se, but rather for diagnoses with panic attacks, e.g., agoraphobia with
From the Department of Psychiatry and Behavioral Sciences, State University of New York at Stony Brook, stony Brook, NY. Adams rqxiut reqoests to Dr. T.A. Amnsoo. Depattmnt of Psychiatq and Behavioral Sciences, Health Scieoces CenterTlO, State University of New York at Stony Brook, Stony Brook, NY 11794-8101. Received December 3, 1987; revised Fcbnmy 6, 1988. This work was suppotted in part by a Basic Research Suppt Grant, NE-I grant #RRO5736. 0 1989 Society of Biological Psychiatry
ooo6-3223l89lSO3.50
470
BIOL PSYCHIATRY 1989;25:469-481
[‘.A. Aronson et al
panic attacks (Liebowitz et al. 1984), atypical depression with panic attacks (McGrath et al. 1985), primary major depression with panic attacks (Dager et al. 1987a), posttraumatic stress disorder with panic attacks (Aleem et al. 1987; Rainey et al. 1987), generalized anxiety disorder with panic attacks (Got-man 1987), and panic disorder with infrequent attacks (Cowley et al. 1987a). The mechanism of lactate-induced panic is still poorly understood. From a biological standpoint, the early theories of lowered ionized serum calcium, metabolic alkalosis, and hypersensitive B-adrenergic receptors have been discarded, and the recent theory of intracerebral acidosis is looking unlikely (Dager et al. 1987b). The current dominant theory of locus ceruleus-noradrenergic dysregulation is mainly supported by yohimbine (Charney et al. 1984) and clonidine (Charney and Heningen 1986) studies, whereas lactate effects on plasma norepinephrine, MHPG, epinephrine, and blood pressure have been equivocal at best (Liebowitz et al. 1985, 1986; Carr et al. 1986; Pohl et al. 1987). There is now debate as to whether or not lactate in fact directly provokes panic via a central biological mechanism. A Dutch group (Griez et al. 1987; van den Hout et al. 1987) suggests that lactate provokes panic via a psychologically conditioned response to peripheral bodily sensations (“interoceptive phobic cues”) that lactate nonspecifically produces . The Stanford group (Margraf et al. 1986; Ehlers et al. 1987) argues that expectations about the laboratory procedure and baseline differences in anxiety levels may entirely account for patients’ responses, and questions whether or not lactate is even an appropriate model for panic. Still apparently at issue is whether patients and controls respond differently to lactate and to what extent psychological or conditioning factors may account for the findings of lactate sensitivity, a proposed biological marker for panic attacks. Previous studies have mainly examined global ratings; for example, the overall severity of lactate-induced anxiety between patients and controls or overall similarity between lactate-induced and naturally occurring panic, rather than specific symptomatic differences. There appears to be an increasing polarization in the discussion of biological or psychological factors, rather than a consideration of both. At issue is whether or not lactate-induced panic is a useful model for studying etiological and treatment-related hypotheses concerning panic disorder. A number of these basic issues need reexamination. In this article, we address the following questions: (1) How precisely does lactate recreate a patient’s typical panic attack? (2) Are there differences between patients and controls in their reactivity to lactate that cannot be explained by baseline differences alone? (3) What specific symptoms distinguish between patients’ and controls’ response to lactate, and what nonspecific symptoms does lactate induce in most everyone? (4) What role do psychological factors play in lactate-induced panic, and can patients’ reactivity be explained entirely on a psychogenic or conditioning basis?
MethcIds We analyzed data from lactate infusions performed on 18 subjects at University Hospital at Stony Brook. This was part of neurobiological pilot study of receptor changes in patients with panic disorder. Nine patients met DSM-IIIR (American Psychiatric Association 1987) criteria for panic disorder. They were age- and sex-matched with 9 controls who had no personal or family history of panic disorder, major depression, agoraphobia, drug or alcohol abuse. Patients were self-referred to the Anxiety Disorders Clinic. Controls were recruited from advertisements placed within the hospital and medical school.
Lactate Sensitivity in Biological Variables
BIOL PSYCHIATRY 1989:25:469481
471
AU subjects met the following inclusion criteria: age 18-45, no history of alcohol or drug abuse within the previous 6 months, no major medical problems, and no psychotropic medication for at least 2 weeks prior to the infusion. All subjects were screened with serum electrolytes, liver function tests, fasting blood glucose, blood urea nitrogen, creatmine, thyroxin, complete blood count, electrocardiogram, and, in the case of female subjects, a urine pregnancy test. All signed an informed consent before participating in the study. Subjects were told prior to the procedure that at some point during the infusion they would receive sodium lactate and that there was a possibility that they might experience a panic attack. A single-blind design was used, similar to most recent reports of lactate infusions in the literature. We did not feel that the single-blind design added any significant bias. All subjects stated after the procedure that they could not tell when the lactate was switched on, i.e., the blind was maintained. A physician and nurse remained with each subject throughout the procedure. The subject was not isolated, as is standard in many other centers. Tbe physician and nurse casually talked with each subject, answered questions, and were available if the subject needed support, which varied for each subject. Their overall effect was to allay anxiety. Given the controversy over the effect of baseline differences on lactate-induced panic, we wanted all subjects to be as calm as possible prior to the switch over to sodium lactate, as it was expected that tbe procedure itself would be more anxiety provoking for the patients with panic disorder. Lactate infusions were performed at 8:30 AMafter the subject had fasted after midnight and abstained from tobacco and caffeine for 24 br. After an intravenous line was inserted (time - 30), subjects received a placebo infusion for the first 30 min. Five percent dextrose in half-normal saline DJ12NS) was administered slowly for 28 min, then speeded up to mimic subsequent lactate administration. After 30 min (at time 0), an infusion of 0.5 M sodium oL-lactate was begun under single-blind conditions. Total dosage (10 ml/kg of body weight) was administered over 20 min or to the point of panic, after which the solution was switched back to the DS1/2NS. Both the lactate and DS1/2NS solutions were warmed to body temperature. An electrocardiogram continuously monitored the subject’s heart rate. Three subjective rating scales were used during the protocol. A Spielberger StateTrait Anxiety (STAI) form was completed by each subject just prior to (at time -40) and following the procedure (at time + 40). During the course of the infusion (at times - 25, - 10, +5, +20), subjects were asked to rate their anxiety on a O-10 scale (10 being tbe worst fear that they had ever felt) and to rate each item on a modified Acute Panic Inventory (API) as absent, mild, moderate, or severe by a research assistant blind to the subject’s diagnosis. The API, originally a 17-item scale developed by the Columbia group (Dillon et al. 1987), was expanded to 24 items in this study to measure tbe severity of commonly reported symptoms of panic attacks. Patients were also instructed to report any subjective symptoms during the course of the infusion as soon as they happened; observers paid particular attention to the initial symptomatic changes of a lactate-induced panic attack. A lactate-induced panic attack was defined conservatively as the abrupt onset of fear or anxiety that reached at least moderate levels, with enough physical symptoms to meet DSM-IIIR criteria, and a report by the subject that this was indeed a panic attack. Physical symptoms alone were not sufficient to merit a diagnosis of panic attack, for example, one control subject described a marked “inner shivering,” as if he had just gotten out of
472
BIOL PSYCHIATRY 1989;25:469%481
I‘. A. Aronson et al
a swimming pool, but without anxiety. After the infusion, patients were asked in a semistructured format to compare the similarities and differences between a lactateinduced and naturally occurring panic attack. For statistical analysis, panic disorder patients with and without agoraphobia were combined. In regard to baseline differences, the STAI scores, prelactate mean heart rates, initial anxiety ratings, initial API scores and screening labs were compared between patients and controls using r-tests. In regard to differences in reactivity to lactate, repeated measures Analysis of Variance (ANOVA) was performed across the four time points for total API scores, each individual API item and anxiety self-ratings. A repeated ANOVA was also performed for heart rate at 5-min intervals. Different reactions of patients and controls would be seen as significant interactions between the factors time (changes over the four time points) and group (patients versus controls). Chi-square methods were used in the analysis of dichotomous variables, such as demographic differences between the two groups. This report is preliminary. We report below all significant results for p < 0.05 and near-significant results for p < 0.10, though a Bonferroni correction for the number of tests performed would indicate that a conservative significance level should be less than 0.002. However, given the small sample size and the pilot nature of the study, we felt that assuming the usual criterion for statistical significance was justified for identifying hypotheses for further research, despite the risks of false positive (Type I) errors. All pvalues are two-tailed.
Results The two groups were matched for age and sex. Five from each group were men, four were women. Their average age was 24.2 + 2.9 (range 21-28), p NS. They differed significantly only on marital. status, of doubtful relevance for this study, as all 9 of the controls were single, whereas 4 patients were married; x2 = 5.14, df = 1, p = 0.023. Five of the patients (56%) were diagnosed as having uncomplicated panic disorder, and 4 (44%) had varying degrees of phobia avoidance, i.e., panic disorder with agoraphobia. Patients reported a mean duration of illness of 2.6 L 7.7 years, and an age of onset of panic disorder at 21.9 + 7.7 years of age. They experienced a mean 5.0 -+ 6.6 attacks per week, and rated their usual daily generalized anxiety level at 8 on a O-10 scale (10 being the highest degree of anxiety). They were moderately impaired by their symptoms, as they received a mean GAS score of 56.5 2 8.2. Six of the 9 (67%) had a definite family history for affective or anxiety disorder. Baseline measurements showed that patients consistently rated themselves as more anxious than controls at the beginning of the procedure, but this did not reach statistical significance for all three relevant variables: Spielberger State Anxiety scores (47.0 versus 30.9, t = 4.49, df = 15.4,~ = 0.0005) at time -40 min; total API scores (33.1 versus 30.1, F = 4.01, df = 1, p = 0.063), and anxiety scores (3.0 versus 1.7, F = 1.42, df = 1, p = 0.256) at time - 20 min. Patients did have significantly higher mean prelactate heart rates than controls: F = 5.53, df = 1, p = 0.032. Baseline chemistries, thyroid function, and blood counts did not significantly differ. Nor were there any significant sex differences on anxiety or heart rate measures. All 9 patients experienced a lactate-induced panic attack, whereas none of the 9 controls panicked (x2 = 18.00, df = 1, p = 0.000). The mean time to panic after the lactate was switched on was 6.4 c 2.9 min; there were no “late-panickers” (Dager et al. 1987;
Dillon et al, 1987). The mean duration of panic was 21.1 f 11.7 min. The mean severity of the panic on a O-10 point scale was 6.9 (range 4-10). An unexpected, paradoxical finding was that 8 of the 9 patients spontaneously remarked at the end of the procedure that the laces-indu~ panic was “reassuring;” that is, it meant that “it was not my fault,” there “really was nothing physically wrong with my body,” “I knew I was not going to die,” and “it was not all in my head.” However, though 8 (89%) noted a high degree of similarity between lactate-induced and naturally occurring panic, none of the 9 patients found that the lactate exactly reproduced their typical panic attacks. Lactate-induced panic differed in severity from naturally occurring panic: 6 patients reported that it was milder, 2 more severe. All patients also mentioned one or two dissimilar symptoms that were more idiosyncratic: less globus hystericus (n = l), less faintness (n = 1), more lightheadedness (n = l), less heart pounding (n = 21, heart racing occurred later rather than as an initial symptom (n = 1)) no derealization (n = l), no hyperventilation (n = l), no nausea fn = l), etc. Of note, the two patients who had extremely high baseline anxiety scores rated their lactate-induced panic as more severe than usual and experienced their attack atypically: one patient’s symptoms began at time + 8 and gradually built to peak severity at time +40; another described racing thoughts and “going through the ceiling with tension,” yet noted only a mild degree of similarity with his usual panic attacks; a third, not included in the data analysis because of her ‘day-long” panics aud doubtful diagnosis, was the only subject who panicked on placebo before receiving lactate and had the procedure stopped. Initial symptoms of lactate-induced panic were also noteworthy. The majority (7 patients; 78%) reported various cognitive and CNS changes as the first symptoms of a panic attack: lightheadedness or dizziness (n = 4), inability to concentrate (n = 3), derealization (n = 2), racing thoughts (n = 21, feeling “funny” and confused (n = 2) ~ Less frequently (2 patients, 22%), the peripheral symptoms of muscular tension (n = 2) and inner shakiness or trembling (n = 2) were noted initially. In all cases, patients reported that the classical panic symptoms of palpitations, shortness of breath, sweating, feeling hot, and anxiety followed later. From an observer’s standpoint, it was often difficult to discern when a patient panicked, save for his/her self-report and a heart rate surge on the elec~~iu~ (see beluw); at most, patients appeared more detached and inwardly preoccupied at the beginning of an attack, developing visible tremor, flushing, and distressed vocalizations only later. Two-way repeated measures ANOVA of the individual API items comparing all subjects over the course of the infusion showed significant or near-significant time effects for 11 of the 24 items: difficulty doing one’s job (F = 8.51, df = 3,14, p = 0.002), ~mbl~gori~er shakiness@ = 7.51, df = 3,14,p = O.OtM),dry mouth(F = 7.38, df = 3,14, p = 0.003), tingling (F = 5.03, df = 3,14,p = 0.014), difficulty SF&ing (F = 4.74, df = 3,14, p = 0.018), difficulty catching one’s breath (F = 4.32, df = 3,14, p = 0,024), dizziness (F = 4.31, df = 2,X5, p = 0.033), palpitations (I; = 3.88, df = 2,15, p = 0.044), confusion (F = 3.12, df = 3,14, p = 0.060), hard to concentrate (F = 2.85, df = 3,14, p = O&75), and chest pain or discomfort (F = 2.82, df = 3,14,p = 0.077). When examining group X time effects on repeated measures ANOVA, 6 of the above items were significant or approached significance for the patient group compared to the controis: trembling or inner shakiness (F = 4.29, df = 3,14, p = 0.0241, difficulty catching one’s breath (F = 3.54, df = 3,14, p = 0.043), palpitations (F = 3.88, df = 2,15, p = O.O44), tingling (F = 3.03,
474
BIOL PSYCHIATRY 1989;25:469-481
T.A. Aronson et ai.
Figure 1. Anxiety scores. Controls: . *m* *; patients: ----I--,
df = 3,14, p = 0.065), chest pain or discomfort (F = 2.82, df = 3,14, p = 0.077), and difficulty doing one’s job (F = 2.68, df = 3,14, p = 0.087). In other words, the tendency to become confused, feel dizzy, have difficulty concentrating, trouble speaking, and a dry mouth characterized all subjects, including controls, who received a lactate infusion. These symptoms represented the nonspecific effects that lactate had on all subjects. On the other hand, the tendency to manifest trembling or inner shakiness, difficulty catching one’s breath, palpitations, tingling, chest pain or discomfort, and overall impairment in the ability to perform one’s job characterized the specific effects of lactate in the patient group and accounted for the time effects of the lactate infusion on these API items. These symptoms represented the unique effects of lactate-induced panic. Notably missing were any significant effects of the lactate infusion on control or patient fears of dying, going crazy, or losing control. Lactate-induced panic among our patients lacked the characteristic cognitive concomitants of “panic.” Repeated measures ANOVA for the group X time effects of total API and anxiety scores over the course of the procedure found significant effects for both in the patient groupcomparedtothecontrolgroup(F = 3.55,df = 3,14,p = 0.043andF = 23.36, df = 3,1O,p = 0.0001, respectively). Post-hoc analysis showed that significant differences for both anxiety measures occurred at time points + 5 and + 20 after the lactate was switched on, (see Figures 1 and 2). Repeated measures of ANOVA for heart rate found no group x time effects, despite significant differences in mean prelactate heart rate (63.7 versus 75.2; F = 5.53, df = 1,p = 0.032), postlactate heart rate (77.1 versus 87.3; F = 5.62; df = 1, p = 0.031), and time effects for heart rate for all subjects over the course of the procedure (F = 37.75, df = 1.16, p = 0.0001). In other words, though patients had higher heart rates than controls throughout the infusion, there was no statistically significant differential reactivity in heart rate to the lactate infusion, as
__----
I
5 I*?
1’3
3x.7
415
BIOL PSYCHIATRY 1989;25:469481
Lactate Sensitivity in Biological Variables
-I--
--_-__
___---_.--__--
___=+_________---= _---.,
’
..
Figure 2. API scores. Controls: **m**; patients: -I-,
_,..--*--.. 1_,, ---.a_
___ ..____*_ _... --
,_..--. .._ ---_*_
___-.-.
60
Lactate
t 50 ’
I
I
1
I
1
-30
-25
-20
-15
-10
-5
0
Infusion
I
I
1
I
I
5
10
15
20
25
Minutes Figure 3. Heart rate over time. Controls: * *n - * ; patients: -I-.
30
476
BIOL PSYCHIATRY 1989;25:469%481
T.A. Aronson
et a~.
lactate appeared to increase heart rate in all subjects (see Figure 3). However, this contradicts our clinical impression that during the beginning stages of a lactate-induced panic attack, patients manifested a 20-40-beat surge in heart rate. Though this is arguably visible in Figure 3, it is not a statistically significant finding.
Discu lssion We found baseline differences between patients and controls for anxiety ratings (though these did not reach statistical significance) and heart rate. This is consistent with most other reports (Freedman et al. 1984; Rainey et al. 1984; Liebowitz et al. 1984, 1985; Ehlers et al. 1986), though not all (Cowley et al. 1987). Higher preinfusion anxiety levels and heart rates were associated with patienthood and, as all patients panicked, with lactate sensitivity as well. Yet it is difficult to reconcile patients’ and controls’ reactions to lactate as entirely secondary to baseline differences, as their reactions differed both qualitatively and quantitatively. We observed differential reactivity to lactate between patients and controls in total API and anxiety scores, as well as in initial heart rate surges, though heat rate increases over the course of the entire procedure did not differ significantly between the two groups. Though some studies have reported a lack of different heart rate reactivity (Freedman et al. 1984; Ehlers et al. 1986), the majority have described sudden increases in heart rate with both lactate-induced panic attacks (Kelly et al. 197 1; Knott et al. 198 1; Liebowitz et al. 1985; Cowley et al. 1987b) and spontaneous panics (Lader and Matthews 1970; Cohen et al. 1985; Taylor et al. 1986). Our results suggest that the negative studies may reflect a too infrequent sampling of heart rate and the tendency of lactate to produce significant increases in heart rate in all subjects over the course of an entire infusion. such that early differences may be obscured statistically. In all subjects, including controls, lactate induced the peripheral symptoms of dry mouth and difficulty swallowing and the central symptoms of dizziness, confusion, and poor concentration. Our findings extend previous reports in the observation that specific API items differentiated between patients’ and controls’ response to lactate. Patients experienced significantly more trembling or inner shakiness, difficulty catching one’s breath, palpitations, tingling, chest pain or discomfort, and impairment in the ability to perform one’s job. These symptoms, as well as the surges in anxiety self-ratings and heart rate 5-10 min after the lactate was switched on, characterized patients’ panic attacks. Central symptoms of dizziness, inability to concentrate, derealization, racing thoughts, feeling “funny” and confused, rather than peripheral cardiovascular or autonomic symptoms, characterized the initial presentation of lactate-induced panic in the majority of patients; the cardiovascular and autonomic symptoms typically followed later. Most patients (8 of 9; 89%) rated the lactate-induced panic as highly similar, but none rated it as identical to a naturally occurring attack, raising questions as to how precisely lactate-induced panic neurobiologically mimics spontaneous panic. Lactate-induced panic differed in severity (usually milder) in one or two specific symptoms and at times in the order of symptom production. interestingly, lactate-induced panic in our laboratory lacked the characteristic cognitive concomitants of panic: fears of dying, going crazy, or losing control. Patients explained later that they felt safe, knew the investigators would not let them die, and so on. Paradoxically, the panic attack was experienced as controlled and reassuring. The supportive environment-at least one physician and nurse who remained
Lactate Sensitivity in biologic
with and freqtentiy
Variables
talked to the patient ~ughout
BmL
PsYcmArnY
477
1989;25:469481
the proced ~ndoub~y
affected
the attack experience. The effect of the environment can also be seen in the patients’ baseline anxiety scores (mean 3.0 on a O-10 scale), which were lower than reported in other studies (Liebowitz et al. 1984; Ehlers et al. 1986). As pointed out by Cowley et al. (1987b), these discrepancies may be due to differences in experimental setting. The Columbia and Stanford groups, for example, place patients alone in a small room, which may be particularly anxiety-provoking for panic disorder patients who often have claustrophobia and fears of being alone. In particular, the inability of Ehlers et al. (1986) to demonstrate differential reactivity in anxiety scores between the two groups may reflect their patients’ and controls’ high baseline anxiety scores (means of 7 and 4, respectively, on a O-10 scale), which may have obscured the significant increases after lactate was switched on. Indeed, our two patients who had high baseline anxiety scores (mean 8) experienced atypical, severe attacks. Another patient with high baseline anxiety scores, who was not included in the data analysis, experienced her attack before the lactate was switched on, an occurrence noted by others (Dager et al. 1987; Yeragani et al. 1987), but rarely commented upon. These prelactate panics suggest that anticipatory anxiety and situational variables play an important role in setting the threshold to panic; this is commonly observed among agoraphobics with predominantly situational panic attacks. They also raise the empirical question of whether or not si~on~ panic attacks differ neurobiologic~ly from spontaneous panic attacks or pharmacologically induced panic attacks. Might prelactate panic be a truer model of the neurobiology of panic attacks than strategies that employ exogenous substances (e.g., lactate, yohimbine, COz, caffeine) that may have confo~~mg, and often con~a~cto~, biological effects? Expectancy biases and baseline psychological states thus appeared to play an important role in our patients’ lactate-induced panic, a position that Margraf et al. (1986) and Ehlers et al. (1986) appear to have taken to an extreme. Our findings suggest that lactate sensitivity as a biological marker may be useful when patients’ baseline anxiety levels are relatively low, but it is unclear in its meaning and often unconvincing when anxiety levels are high. At issue is the relative role that conditioning, psychological, and environmental factors play in ph~a~logic~ly induced panic attacks. Biological theories of hypersensitive Badrenergic receptors, intracerebral acidosis, locus ceruleus activation, etc., all assume that a biological mechanism explains lactate-induced panic. They cannot account for the salient role played by expectancy biases, baseline psychological states, and environmental factors in our and others’ studies. Recently, several groups have elaborated upon the phobic response theory of Ackerman and Sacher (1974) that views lactate-induced panic in terms of an interaction among past experience, environmental, and cognitive factors: “It appears that interoceptive fears of bodily anxiety may, in fact, be sufficient to explain the apparent panicogenic properties of lactate and CO;’ (van den Hout et al. 1987, p 788). “As [panic attacks] reoccur during repeated intense, adversive experiences, the abovementioned adrenergic arousal symptoms themselves become interoceptive conditioned stimuli for further anxiety, contributing to the perpetuation of panic anxiety” (Griez et al. 1987, p 94). “The subject has learned to associate certain bodily sensations with acute anxiety by a process of repeated interoceptive conditioning and subsequent cognitive elaboration . . . . The response.to [lactate-inducedstimuli1depends on the perception of the bodily changes and on the appraisal of env~~~~ and internal cues in terms of perceived un~~~n~,
478
BIOL PSYCHIATRY 1989;25:46%481
f.A. Aronson et al.
threat/potentialhartn, and availabilityof responsesor coping strategies”(Margrafet al. 1986, P 43). These investigators note that patients habituate to the panicogenic power of COz, which decreases with repeated exposure (van den Hout et al. 1987). Preliminary data suggest that successful cognitive-behavioral treatment of panic disorder also blocks lactate-induced panic (Guttmacher and Nelles 1984; Gorman 1987). The recent findings of behavioral hypersensitivity to CO;! and weak or no evidence for physiological or biochemical hypersensitivity to CO* (Woods et al. 1988) are consistent with this model of interoceptive conditioning in patients with panic disorder. Can purely psychological and conditioning mechanisms entirely account for our findings? It is doubtful. Our patients experienced lactate-induced panic attacks despite the lack of typical cognitive concomitants of panic and with less intense anxiety than normally experienced. Peripheral cardiovascular and autonomic symptoms occurred later in the development of lactate-induced panic, reducing the likelihood that James-Lange or peripheral interoceptive phenomena gave rise to central symptoms. In fact, in the majority of patients, it occurred the other way around. Patients did not simply endorse any API symptom because of heightened anxiety (Margraf et al. 1986)) but rather reported specific panic symptoms consistent with DSM-IIIR nosology. Finally, there was a differential reactivity in anxiety ratings and heart rates that could not be explained by baseline differences alone. The interoceptive cue theory does not explain why patients do not panic with any challenge that disturbs their internal state, such as hypoglycemia (Udhe et al. 1984; Schweitzer et al. 1986) or respiratory alkalosis (Gorman et al. 1984). Nor does it explain what gives rise to initial panic attacks in the first place. It is a theory of a maintenance mechanism, not a causal theory of panic disorder (van den Hout et al. 1987). Similar to environmental, expectancy, appraisal, and mood variables, it may explain some aspects of lactate-induced panic, but it does not completely explain it away. Biological and psychological factors both appear to be important. The latter are more obvious, the former still are poorly understood. In patients with low baseline anxiety. we speculate that biological factors may predominate; with high baseline anxiety, cognitive and conditioning mechanisms may be more relevant for the researcher and clinician. For ’ ” in panic disorder may be a secondary, example, our findings suggest that the “pamc reactive phenomenon; that is, the proper cognitive set or environmental setting may obviate the occurrence of fears of dying, going crazy, or losing control in the face of the interoceptive cues of a panic attack, but a catastrophic or helpless cognitive set may induce a negative feedback loop that may increase both the “panic” and the interoceptive symptoms. A biopsychosocial theory is now needed that can integrate the biological findings of pontine locus ceruleus dysfunction (Chamey et al. 1984, 1986), medullary nucleus reticularis gigantocellularis dysfunction of respiratory control (Gorman 1987), and limbic alterations in hypothalamic (Udhe 1987) and hippocampal function (Reiman et al. 1986) with cortically mediated psychological mechanisms that modulate the course of panic disorder and experience of lactate sensitivity. Implications for the neurobiology of panic disorder and for therapeutic interventions at different levels are obvious. The gate theory of pain suggests one biopsychosocial model: limbic and brainstem centers may mediate the vulnerability to panic attacks, yet can be gated or modulated by higher cortical/psychological influences that alter the threshold and experience of panic. Finally, our study supports other groups’ findings of a paradoxical therapeutic potential
Lactate Sensitivity in Biological VariabLs
BIOL F’SY&IiLKIRY
1989,2546!Usi
479
with lactate i~sions. van den Hout et al.‘s (1987) report of decreased anxiety responses with serial exposure to CO* inhalation, and Bonn et al.‘s (1974) open trial of 6 lactate
infusions in 33 patients suggest that patients can be habituated to many of the interoceptive aspects of panic and anticipatory anxiety. In other words, anxiogenics can be anxiolytic. Whether such desensiti~tion training makes patients panic-free or simply changes the cognitive response to panic attacks, which continue but without the “panic,” is unclear. Further study of this strategy appears warranted. This study has a number of methodological limitations to the generalizations that can be drawn. This was a pilot project with small patient numbers. The patients experienced a relatively high number of weekly panic attacks, and all 9 panicked with lactate, higher than the 70%-75% reported frequency in previous studies with higher patient numbers. The levels of significance reported (p < 0.05) did not include the Bonferroni correction for the number of statistical tests performed. The study perhaps raises more questions than it answers. The results call for confirmation by studies with larger patient numbers.
References Ackerman S, Sachar E (1974): The lactate theory of anxiety: A review and reevaluation. Psychosom Med 36:69-74.
Aleem A, Rainey JM, Grtiz A, et al (1987): Lactate infusion in post-traumatic stress disorder. Presented at the American Psychiatric Association annual convention, Chicago, IL. American Psychiatric Association (1987): Diagnostic and Statistical Manual of Mental Disease, ed 3 (rev). Washington DC: American Psychiatric Association. AppIeby IL, Klein DF, Sachar El, et al (1981): Biochemical indices of lactate-induced panic: A prelii report. In Klein DF, Rabkin J (eds), Anxiety: New Research and Changing Concepts. New York: Raven Press, pp 411-423. Bonn J, Harrison J, Rees W (1971): Lactate-induced anxiety: Therapeutic application. Br J Psychiatry 119~468-470. Can DB, Sheehan DV, Surman OS, et al (1986): Neuroendocrine correlates of lactate-induced anxiety and their response to chronic alprazolam therapy. Am J Psychiatry 143:483-494. Chamey DS , Heninger GR ( 1986): Abnormal regulation of noradrenergic function in panic disorders: Effects of clonidme in healthy subjects and patients with a~omphobia and panic disorder. Arch Gen Psychiatry 4311042-1054.
Chamey DS, Heninger GR, Breier A (1984): Noradrenergic function in panic anxiety: Effects of yohimbine in healthy subjects and patients with agoraphobia and panic disorder. Arch Gen Pact
41:751-763.
Cohen AS, Barlow DH, Blanchard EB (1985): The psychophysiology of relaxation-associated panic attacks. J Abnorm Psycho1 9496-100. Cowley DS, Dager SR, Foster SI, et al (1987a): Clinical characteristics and response to sodium lactate of patients with infrequent panic attacks. Am J ~~chia~ 144:795-798. Cowley DS, Hyde TS, Dager SR, et al. (1987b): Lactate infusions: The role of baseline anxiety. Psychiatry Res 21:169-179. Dager SR, Cowley DS, Dunner DL (1987a): Biological markem in panic states: Lactate-induced panic and mitral valve prolapse. Biol Psychiatry 22339-359. Dager SR, Rainey JM, Dunner DL, et al (1987b): Central effects of lactate infusion in a primate model. Presented at the American College of Neuropsychopharmacology annual meeting, San Juan, Puerto Rico. Dillon DJ, German JM, Liebowitz MR, et al (1987): Measurement of lactate-induced panic and anxiety. Psychiatry Res 20:97-105.
480
BiOJ..PSYCHIATRY 19EWA-B-481
l’.A. Aronson et al.
Ehlers A, Margraf J, Roth WT, et al (1987): Lactate infusions and panic attacks: Do patients and controls respond differently? Prychiutry Res 17:295-308. Fink M, Taylor MA, Volavka J (1970): Anxiety precipitated by lactate. N Engf J Med 281: 1429. Freedman RR, Ianni P, Ettedgui E, et al (1984): Psychophysiological factors in panic disorder. ~~c~upuf~~ugy 17(Suppl I)&%-73 * Gotman JM (1987): Experimental induction of anxiety by lactate and other substances. Presented at the annual meeting of the American College of Neuropsychopharmacology in San Juan, Puerto Rico. German J, Askanazi J, Liebowitz M, et al (1984): Response to hyperventilation in a group of patients with panic disorder. Am J P~c~j~~ 141:857-861. Griez EJL, Lot&erg H, van den Hout MA, et al (1987): CO2 vulnerability in panic disorder. Psychiatry Res 20:87-95. Guttmacher LB, Nelles C (1984): In vivo desensitization alternation of lactate-induced panic: A case study. Behuv Res Ther 15:369-372. Kelly D, Mitchell-Heggs N, Sherman D (1971): Anxiety and the effects of sodium lactate assessed cfinically and physioiogic~ly Er J P~ch~~~ 119:129-141. Knott V, Chat&y R, Lapierre Y (1981): Panic induced by sodium lactate: Elec~ophysiologic~ correlates. Prog Neuropsychophurmacol5:s 1l-514. Lader M, Mathews A (1970): Physiological changes during spontaneous panic attacks. J Psychosom Res l4:377-381. Liebowitz ML, Fyer A, German J, et al (1984): Lactate provocation of panic attacks: I. Clinical and behavioral findings. Arch GermPsychiatry 41:764-770. Liebowitz MR, Gorman JM, Fyer Aj, et al (1985): Lactate provocation of panic attacks. II. Biochemical and physiological findings. Arch Gen Psychiatry 42:709-719. Liebowitz MR, German JM, Fyer A, et al (1986): Possible mechanisms for lactate’s induction of panic. Am J Psychiatry 143:495-502. Margraf J, Ehlers A, Roth W (1986): Sodium lactate infusions and panic attacks: A review and critique. P~c~o~~rn Med 48:23-5 1. McGrath P, Stewart J, Harrison W, et al (1985): Lactate infusion in patients with depression and anxiety. Psychopharmacol Bull 21:555-557. Pitts FN, McClure JN (1967): Lactate metabolism in anxiety neurosis. N Engl J Med 277:13291336.
Pohl R, Ettedgiri E, Bridges M, et al (1987): Plasma MHPG levels in lactate and iosproterenol anxiety states. Biol P~chja~ 22: 1127-l 136. Rainey JM Jr, Pohl RB, Williams M, et al (1984): A comparison of lactate and iosproterenol anxiety states. Psychopathology 17(Suppl 1):74-82. Rainey JM Jr, Aleem A, Ortiz A, et al (1987): A laboratory procedure for the induction of flashbacks. Am J Psychiatry 144:1317-1319.
Reiman EM, Raichle ME, Butler RE, et al (1986): Application of positron emission tomography to study of panic disorder. Am J P~chi~~ 143:469477. Rifkin A, Klein DF, Dillon D, et al (1981): Blockade by imipramine or desipramine of panic induced by sodium lactate. Am J Psychiatry 138:676-677. Schweitzer E, Winokur A, Rickels K (1986): Insulin induced hypoglycemia and panic attacks. Am J Psychiatry 143:654--655.
Taylor CB, She&h J, Agras WS, et al (1986): Ambulatory heart rate changes in patients with panic attacks. Am J Psychiatry 143:478481. Udhe TW (1987): Biology of panic disorder. Presented at Grand Rounds, State University of New York at Stony Brook, Stony Brook, NY, October 6. Udhe TW, Vittone BJ, Post RM (1984): Glucose tolerance testing in panic disorder. Am J Psychiatry 141:1461-1463.
Lactate Sensitivity in Biological Variables
BIOL PSYCHIATRY 1989;25:469-481
481
van den Hout MA, van der Molen GM, Griez E, et al (1987): Reduction of CO&duced anxiety in patients with panic attacks after repeated CO* exposure. Am J Psychiatry 144:788-791. Woods SW, Chamey DS, Goodman WK, et al (1988): Carbon dioxide-induced anxiety. Arch Gen Psychiatry 45143-52.
Yeragani VK, Balon R, Pohl R, et al (1987): Do higher preinfusion heart rates predict laboratoryinduced panic attacks? Biol Psychiatry 22:554-555.
469
Differential Reactivity to Lactate Infusions: The Relative Role of Biological, Psychological, and conditioning Variables Thomas A. Aronson, Patricia Whitaker-Azmitia, and Irene Caraseti
Nine patients with panic disorder experienced a lactate-induced panic attack, whereas nine controls did not. Higher preinfusion anxiety levels and heart rates were associated with panic disorder, ana’ high baseline anxiety ratings were associated with atypical, severe lactate-induced panic attacks. Nevertheless, it was difbcult to reconcile patients’ and co#ro~’ reac~v~~ to lactate as entirely se&o~~ to reline di~eren~es. subjects diflered q~l~t#ively in the types of specific symptoms experienced and q~ntitatively in their anxiety and heart rate responses. In most cases, panic began with various central perceptual changes; peripheral cardiovascular and autonomic symptoms followed later. No patient rated a lactate-induced panic attack as identical to a naturally occurring attack. Not only did specific ~mptoms di#er in their severity and order of production, but lactate-inured panic lacked the typicalfears of ping, going crazy, or losing control. The results suggest that though environmental effects, expectancy biases, and baseline psychological states play salient roles in modifying the experience of a lactate-induced panic attack, they do not fully account for lactate sensitivity. The relative role that biological, psychological, and conditioning factors play in lactate-induced panic is discussed.
Introduction Intravenous sodium lactate has been the most extensively studied of the provocative agents used in the psychobiological study of panic disorder-a growing list that includes yo~bine, clondine, 5% and 35% CO, ideation, caffeine, isop~~~nol, m-chlorophenylpi~~ne, and sodium bicarbonate. Since 1967, numerous reports have shown that approximately 70%-80% of patients with panic disorder or panic disorder with agoraphobia experience a lactate-induced panic attack compared to 0%-10% of normal controls (Pitts and McClure 1967; Bonn et al. 1971; Fink et al. 1971; Kelly et al. 1971; Appleby et al. 1981; Rifkin et al. 1981; Liebowitz et al. 1984; Raney et al. 1984; Dager et al. 1987). However, recent evidence suggests that lactate sensitivity is not specific for panic disorder per se, but rather for diagnoses with panic attacks, e.g., agoraphobia with
From the Department of Psychiatry and Behavioral Sciences, State University of New York at Stony Brook, stony Brook, NY. Adams rqxiut reqoests to Dr. T.A. Amnsoo. Depattmnt of Psychiatq and Behavioral Sciences, Health Scieoces CenterTlO, State University of New York at Stony Brook, Stony Brook, NY 11794-8101. Received December 3, 1987; revised Fcbnmy 6, 1988. This work was suppotted in part by a Basic Research Suppt Grant, NE-I grant #RRO5736. 0 1989 Society of Biological Psychiatry
ooo6-3223l89lSO3.50
470
BIOL PSYCHIATRY 1989;25:469-481
[‘.A. Aronson et al
panic attacks (Liebowitz et al. 1984), atypical depression with panic attacks (McGrath et al. 1985), primary major depression with panic attacks (Dager et al. 1987a), posttraumatic stress disorder with panic attacks (Aleem et al. 1987; Rainey et al. 1987), generalized anxiety disorder with panic attacks (Got-man 1987), and panic disorder with infrequent attacks (Cowley et al. 1987a). The mechanism of lactate-induced panic is still poorly understood. From a biological standpoint, the early theories of lowered ionized serum calcium, metabolic alkalosis, and hypersensitive B-adrenergic receptors have been discarded, and the recent theory of intracerebral acidosis is looking unlikely (Dager et al. 1987b). The current dominant theory of locus ceruleus-noradrenergic dysregulation is mainly supported by yohimbine (Charney et al. 1984) and clonidine (Charney and Heningen 1986) studies, whereas lactate effects on plasma norepinephrine, MHPG, epinephrine, and blood pressure have been equivocal at best (Liebowitz et al. 1985, 1986; Carr et al. 1986; Pohl et al. 1987). There is now debate as to whether or not lactate in fact directly provokes panic via a central biological mechanism. A Dutch group (Griez et al. 1987; van den Hout et al. 1987) suggests that lactate provokes panic via a psychologically conditioned response to peripheral bodily sensations (“interoceptive phobic cues”) that lactate nonspecifically produces . The Stanford group (Margraf et al. 1986; Ehlers et al. 1987) argues that expectations about the laboratory procedure and baseline differences in anxiety levels may entirely account for patients’ responses, and questions whether or not lactate is even an appropriate model for panic. Still apparently at issue is whether patients and controls respond differently to lactate and to what extent psychological or conditioning factors may account for the findings of lactate sensitivity, a proposed biological marker for panic attacks. Previous studies have mainly examined global ratings; for example, the overall severity of lactate-induced anxiety between patients and controls or overall similarity between lactate-induced and naturally occurring panic, rather than specific symptomatic differences. There appears to be an increasing polarization in the discussion of biological or psychological factors, rather than a consideration of both. At issue is whether or not lactate-induced panic is a useful model for studying etiological and treatment-related hypotheses concerning panic disorder. A number of these basic issues need reexamination. In this article, we address the following questions: (1) How precisely does lactate recreate a patient’s typical panic attack? (2) Are there differences between patients and controls in their reactivity to lactate that cannot be explained by baseline differences alone? (3) What specific symptoms distinguish between patients’ and controls’ response to lactate, and what nonspecific symptoms does lactate induce in most everyone? (4) What role do psychological factors play in lactate-induced panic, and can patients’ reactivity be explained entirely on a psychogenic or conditioning basis?
MethcIds We analyzed data from lactate infusions performed on 18 subjects at University Hospital at Stony Brook. This was part of neurobiological pilot study of receptor changes in patients with panic disorder. Nine patients met DSM-IIIR (American Psychiatric Association 1987) criteria for panic disorder. They were age- and sex-matched with 9 controls who had no personal or family history of panic disorder, major depression, agoraphobia, drug or alcohol abuse. Patients were self-referred to the Anxiety Disorders Clinic. Controls were recruited from advertisements placed within the hospital and medical school.
Lactate Sensitivity in Biological Variables
BIOL PSYCHIATRY 1989:25:469481
471
AU subjects met the following inclusion criteria: age 18-45, no history of alcohol or drug abuse within the previous 6 months, no major medical problems, and no psychotropic medication for at least 2 weeks prior to the infusion. All subjects were screened with serum electrolytes, liver function tests, fasting blood glucose, blood urea nitrogen, creatmine, thyroxin, complete blood count, electrocardiogram, and, in the case of female subjects, a urine pregnancy test. All signed an informed consent before participating in the study. Subjects were told prior to the procedure that at some point during the infusion they would receive sodium lactate and that there was a possibility that they might experience a panic attack. A single-blind design was used, similar to most recent reports of lactate infusions in the literature. We did not feel that the single-blind design added any significant bias. All subjects stated after the procedure that they could not tell when the lactate was switched on, i.e., the blind was maintained. A physician and nurse remained with each subject throughout the procedure. The subject was not isolated, as is standard in many other centers. Tbe physician and nurse casually talked with each subject, answered questions, and were available if the subject needed support, which varied for each subject. Their overall effect was to allay anxiety. Given the controversy over the effect of baseline differences on lactate-induced panic, we wanted all subjects to be as calm as possible prior to the switch over to sodium lactate, as it was expected that tbe procedure itself would be more anxiety provoking for the patients with panic disorder. Lactate infusions were performed at 8:30 AMafter the subject had fasted after midnight and abstained from tobacco and caffeine for 24 br. After an intravenous line was inserted (time - 30), subjects received a placebo infusion for the first 30 min. Five percent dextrose in half-normal saline DJ12NS) was administered slowly for 28 min, then speeded up to mimic subsequent lactate administration. After 30 min (at time 0), an infusion of 0.5 M sodium oL-lactate was begun under single-blind conditions. Total dosage (10 ml/kg of body weight) was administered over 20 min or to the point of panic, after which the solution was switched back to the DS1/2NS. Both the lactate and DS1/2NS solutions were warmed to body temperature. An electrocardiogram continuously monitored the subject’s heart rate. Three subjective rating scales were used during the protocol. A Spielberger StateTrait Anxiety (STAI) form was completed by each subject just prior to (at time -40) and following the procedure (at time + 40). During the course of the infusion (at times - 25, - 10, +5, +20), subjects were asked to rate their anxiety on a O-10 scale (10 being tbe worst fear that they had ever felt) and to rate each item on a modified Acute Panic Inventory (API) as absent, mild, moderate, or severe by a research assistant blind to the subject’s diagnosis. The API, originally a 17-item scale developed by the Columbia group (Dillon et al. 1987), was expanded to 24 items in this study to measure tbe severity of commonly reported symptoms of panic attacks. Patients were also instructed to report any subjective symptoms during the course of the infusion as soon as they happened; observers paid particular attention to the initial symptomatic changes of a lactate-induced panic attack. A lactate-induced panic attack was defined conservatively as the abrupt onset of fear or anxiety that reached at least moderate levels, with enough physical symptoms to meet DSM-IIIR criteria, and a report by the subject that this was indeed a panic attack. Physical symptoms alone were not sufficient to merit a diagnosis of panic attack, for example, one control subject described a marked “inner shivering,” as if he had just gotten out of
472
BIOL PSYCHIATRY 1989;25:469%481
I‘. A. Aronson et al
a swimming pool, but without anxiety. After the infusion, patients were asked in a semistructured format to compare the similarities and differences between a lactateinduced and naturally occurring panic attack. For statistical analysis, panic disorder patients with and without agoraphobia were combined. In regard to baseline differences, the STAI scores, prelactate mean heart rates, initial anxiety ratings, initial API scores and screening labs were compared between patients and controls using r-tests. In regard to differences in reactivity to lactate, repeated measures Analysis of Variance (ANOVA) was performed across the four time points for total API scores, each individual API item and anxiety self-ratings. A repeated ANOVA was also performed for heart rate at 5-min intervals. Different reactions of patients and controls would be seen as significant interactions between the factors time (changes over the four time points) and group (patients versus controls). Chi-square methods were used in the analysis of dichotomous variables, such as demographic differences between the two groups. This report is preliminary. We report below all significant results for p < 0.05 and near-significant results for p < 0.10, though a Bonferroni correction for the number of tests performed would indicate that a conservative significance level should be less than 0.002. However, given the small sample size and the pilot nature of the study, we felt that assuming the usual criterion for statistical significance was justified for identifying hypotheses for further research, despite the risks of false positive (Type I) errors. All pvalues are two-tailed.
Results The two groups were matched for age and sex. Five from each group were men, four were women. Their average age was 24.2 + 2.9 (range 21-28), p NS. They differed significantly only on marital. status, of doubtful relevance for this study, as all 9 of the controls were single, whereas 4 patients were married; x2 = 5.14, df = 1, p = 0.023. Five of the patients (56%) were diagnosed as having uncomplicated panic disorder, and 4 (44%) had varying degrees of phobia avoidance, i.e., panic disorder with agoraphobia. Patients reported a mean duration of illness of 2.6 L 7.7 years, and an age of onset of panic disorder at 21.9 + 7.7 years of age. They experienced a mean 5.0 -+ 6.6 attacks per week, and rated their usual daily generalized anxiety level at 8 on a O-10 scale (10 being the highest degree of anxiety). They were moderately impaired by their symptoms, as they received a mean GAS score of 56.5 2 8.2. Six of the 9 (67%) had a definite family history for affective or anxiety disorder. Baseline measurements showed that patients consistently rated themselves as more anxious than controls at the beginning of the procedure, but this did not reach statistical significance for all three relevant variables: Spielberger State Anxiety scores (47.0 versus 30.9, t = 4.49, df = 15.4,~ = 0.0005) at time -40 min; total API scores (33.1 versus 30.1, F = 4.01, df = 1, p = 0.063), and anxiety scores (3.0 versus 1.7, F = 1.42, df = 1, p = 0.256) at time - 20 min. Patients did have significantly higher mean prelactate heart rates than controls: F = 5.53, df = 1, p = 0.032. Baseline chemistries, thyroid function, and blood counts did not significantly differ. Nor were there any significant sex differences on anxiety or heart rate measures. All 9 patients experienced a lactate-induced panic attack, whereas none of the 9 controls panicked (x2 = 18.00, df = 1, p = 0.000). The mean time to panic after the lactate was switched on was 6.4 c 2.9 min; there were no “late-panickers” (Dager et al. 1987;
Dillon et al, 1987). The mean duration of panic was 21.1 f 11.7 min. The mean severity of the panic on a O-10 point scale was 6.9 (range 4-10). An unexpected, paradoxical finding was that 8 of the 9 patients spontaneously remarked at the end of the procedure that the laces-indu~ panic was “reassuring;” that is, it meant that “it was not my fault,” there “really was nothing physically wrong with my body,” “I knew I was not going to die,” and “it was not all in my head.” However, though 8 (89%) noted a high degree of similarity between lactate-induced and naturally occurring panic, none of the 9 patients found that the lactate exactly reproduced their typical panic attacks. Lactate-induced panic differed in severity from naturally occurring panic: 6 patients reported that it was milder, 2 more severe. All patients also mentioned one or two dissimilar symptoms that were more idiosyncratic: less globus hystericus (n = l), less faintness (n = 1), more lightheadedness (n = l), less heart pounding (n = 21, heart racing occurred later rather than as an initial symptom (n = 1)) no derealization (n = l), no hyperventilation (n = l), no nausea fn = l), etc. Of note, the two patients who had extremely high baseline anxiety scores rated their lactate-induced panic as more severe than usual and experienced their attack atypically: one patient’s symptoms began at time + 8 and gradually built to peak severity at time +40; another described racing thoughts and “going through the ceiling with tension,” yet noted only a mild degree of similarity with his usual panic attacks; a third, not included in the data analysis because of her ‘day-long” panics aud doubtful diagnosis, was the only subject who panicked on placebo before receiving lactate and had the procedure stopped. Initial symptoms of lactate-induced panic were also noteworthy. The majority (7 patients; 78%) reported various cognitive and CNS changes as the first symptoms of a panic attack: lightheadedness or dizziness (n = 4), inability to concentrate (n = 3), derealization (n = 2), racing thoughts (n = 21, feeling “funny” and confused (n = 2) ~ Less frequently (2 patients, 22%), the peripheral symptoms of muscular tension (n = 2) and inner shakiness or trembling (n = 2) were noted initially. In all cases, patients reported that the classical panic symptoms of palpitations, shortness of breath, sweating, feeling hot, and anxiety followed later. From an observer’s standpoint, it was often difficult to discern when a patient panicked, save for his/her self-report and a heart rate surge on the elec~~iu~ (see beluw); at most, patients appeared more detached and inwardly preoccupied at the beginning of an attack, developing visible tremor, flushing, and distressed vocalizations only later. Two-way repeated measures ANOVA of the individual API items comparing all subjects over the course of the infusion showed significant or near-significant time effects for 11 of the 24 items: difficulty doing one’s job (F = 8.51, df = 3,14, p = 0.002), ~mbl~gori~er shakiness@ = 7.51, df = 3,14,p = O.OtM),dry mouth(F = 7.38, df = 3,14, p = 0.003), tingling (F = 5.03, df = 3,14,p = 0.014), difficulty SF&ing (F = 4.74, df = 3,14, p = 0.018), difficulty catching one’s breath (F = 4.32, df = 3,14, p = 0,024), dizziness (F = 4.31, df = 2,X5, p = 0.033), palpitations (I; = 3.88, df = 2,15, p = 0.044), confusion (F = 3.12, df = 3,14, p = 0.060), hard to concentrate (F = 2.85, df = 3,14, p = O&75), and chest pain or discomfort (F = 2.82, df = 3,14,p = 0.077). When examining group X time effects on repeated measures ANOVA, 6 of the above items were significant or approached significance for the patient group compared to the controis: trembling or inner shakiness (F = 4.29, df = 3,14, p = 0.0241, difficulty catching one’s breath (F = 3.54, df = 3,14, p = 0.043), palpitations (F = 3.88, df = 2,15, p = O.O44), tingling (F = 3.03,
474
BIOL PSYCHIATRY 1989;25:469-481
T.A. Aronson et ai.
Figure 1. Anxiety scores. Controls: . *m* *; patients: ----I--,
df = 3,14, p = 0.065), chest pain or discomfort (F = 2.82, df = 3,14, p = 0.077), and difficulty doing one’s job (F = 2.68, df = 3,14, p = 0.087). In other words, the tendency to become confused, feel dizzy, have difficulty concentrating, trouble speaking, and a dry mouth characterized all subjects, including controls, who received a lactate infusion. These symptoms represented the nonspecific effects that lactate had on all subjects. On the other hand, the tendency to manifest trembling or inner shakiness, difficulty catching one’s breath, palpitations, tingling, chest pain or discomfort, and overall impairment in the ability to perform one’s job characterized the specific effects of lactate in the patient group and accounted for the time effects of the lactate infusion on these API items. These symptoms represented the unique effects of lactate-induced panic. Notably missing were any significant effects of the lactate infusion on control or patient fears of dying, going crazy, or losing control. Lactate-induced panic among our patients lacked the characteristic cognitive concomitants of “panic.” Repeated measures ANOVA for the group X time effects of total API and anxiety scores over the course of the procedure found significant effects for both in the patient groupcomparedtothecontrolgroup(F = 3.55,df = 3,14,p = 0.043andF = 23.36, df = 3,1O,p = 0.0001, respectively). Post-hoc analysis showed that significant differences for both anxiety measures occurred at time points + 5 and + 20 after the lactate was switched on, (see Figures 1 and 2). Repeated measures of ANOVA for heart rate found no group x time effects, despite significant differences in mean prelactate heart rate (63.7 versus 75.2; F = 5.53, df = 1,p = 0.032), postlactate heart rate (77.1 versus 87.3; F = 5.62; df = 1, p = 0.031), and time effects for heart rate for all subjects over the course of the procedure (F = 37.75, df = 1.16, p = 0.0001). In other words, though patients had higher heart rates than controls throughout the infusion, there was no statistically significant differential reactivity in heart rate to the lactate infusion, as
__----
I
5 I*?
1’3
3x.7
415
BIOL PSYCHIATRY 1989;25:469481
Lactate Sensitivity in Biological Variables
-I--
--_-__
___---_.--__--
___=+_________---= _---.,
’
..
Figure 2. API scores. Controls: **m**; patients: -I-,
_,..--*--.. 1_,, ---.a_
___ ..____*_ _... --
,_..--. .._ ---_*_
___-.-.
60
Lactate
t 50 ’
I
I
1
I
1
-30
-25
-20
-15
-10
-5
0
Infusion
I
I
1
I
I
5
10
15
20
25
Minutes Figure 3. Heart rate over time. Controls: * *n - * ; patients: -I-.
30
476
BIOL PSYCHIATRY 1989;25:469%481
T.A. Aronson
et a~.
lactate appeared to increase heart rate in all subjects (see Figure 3). However, this contradicts our clinical impression that during the beginning stages of a lactate-induced panic attack, patients manifested a 20-40-beat surge in heart rate. Though this is arguably visible in Figure 3, it is not a statistically significant finding.
Discu lssion We found baseline differences between patients and controls for anxiety ratings (though these did not reach statistical significance) and heart rate. This is consistent with most other reports (Freedman et al. 1984; Rainey et al. 1984; Liebowitz et al. 1984, 1985; Ehlers et al. 1986), though not all (Cowley et al. 1987). Higher preinfusion anxiety levels and heart rates were associated with patienthood and, as all patients panicked, with lactate sensitivity as well. Yet it is difficult to reconcile patients’ and controls’ reactions to lactate as entirely secondary to baseline differences, as their reactions differed both qualitatively and quantitatively. We observed differential reactivity to lactate between patients and controls in total API and anxiety scores, as well as in initial heart rate surges, though heat rate increases over the course of the entire procedure did not differ significantly between the two groups. Though some studies have reported a lack of different heart rate reactivity (Freedman et al. 1984; Ehlers et al. 1986), the majority have described sudden increases in heart rate with both lactate-induced panic attacks (Kelly et al. 197 1; Knott et al. 198 1; Liebowitz et al. 1985; Cowley et al. 1987b) and spontaneous panics (Lader and Matthews 1970; Cohen et al. 1985; Taylor et al. 1986). Our results suggest that the negative studies may reflect a too infrequent sampling of heart rate and the tendency of lactate to produce significant increases in heart rate in all subjects over the course of an entire infusion. such that early differences may be obscured statistically. In all subjects, including controls, lactate induced the peripheral symptoms of dry mouth and difficulty swallowing and the central symptoms of dizziness, confusion, and poor concentration. Our findings extend previous reports in the observation that specific API items differentiated between patients’ and controls’ response to lactate. Patients experienced significantly more trembling or inner shakiness, difficulty catching one’s breath, palpitations, tingling, chest pain or discomfort, and impairment in the ability to perform one’s job. These symptoms, as well as the surges in anxiety self-ratings and heart rate 5-10 min after the lactate was switched on, characterized patients’ panic attacks. Central symptoms of dizziness, inability to concentrate, derealization, racing thoughts, feeling “funny” and confused, rather than peripheral cardiovascular or autonomic symptoms, characterized the initial presentation of lactate-induced panic in the majority of patients; the cardiovascular and autonomic symptoms typically followed later. Most patients (8 of 9; 89%) rated the lactate-induced panic as highly similar, but none rated it as identical to a naturally occurring attack, raising questions as to how precisely lactate-induced panic neurobiologically mimics spontaneous panic. Lactate-induced panic differed in severity (usually milder) in one or two specific symptoms and at times in the order of symptom production. interestingly, lactate-induced panic in our laboratory lacked the characteristic cognitive concomitants of panic: fears of dying, going crazy, or losing control. Patients explained later that they felt safe, knew the investigators would not let them die, and so on. Paradoxically, the panic attack was experienced as controlled and reassuring. The supportive environment-at least one physician and nurse who remained
Lactate Sensitivity in biologic
with and freqtentiy
Variables
talked to the patient ~ughout
BmL
PsYcmArnY
477
1989;25:469481
the proced ~ndoub~y
affected
the attack experience. The effect of the environment can also be seen in the patients’ baseline anxiety scores (mean 3.0 on a O-10 scale), which were lower than reported in other studies (Liebowitz et al. 1984; Ehlers et al. 1986). As pointed out by Cowley et al. (1987b), these discrepancies may be due to differences in experimental setting. The Columbia and Stanford groups, for example, place patients alone in a small room, which may be particularly anxiety-provoking for panic disorder patients who often have claustrophobia and fears of being alone. In particular, the inability of Ehlers et al. (1986) to demonstrate differential reactivity in anxiety scores between the two groups may reflect their patients’ and controls’ high baseline anxiety scores (means of 7 and 4, respectively, on a O-10 scale), which may have obscured the significant increases after lactate was switched on. Indeed, our two patients who had high baseline anxiety scores (mean 8) experienced atypical, severe attacks. Another patient with high baseline anxiety scores, who was not included in the data analysis, experienced her attack before the lactate was switched on, an occurrence noted by others (Dager et al. 1987; Yeragani et al. 1987), but rarely commented upon. These prelactate panics suggest that anticipatory anxiety and situational variables play an important role in setting the threshold to panic; this is commonly observed among agoraphobics with predominantly situational panic attacks. They also raise the empirical question of whether or not si~on~ panic attacks differ neurobiologic~ly from spontaneous panic attacks or pharmacologically induced panic attacks. Might prelactate panic be a truer model of the neurobiology of panic attacks than strategies that employ exogenous substances (e.g., lactate, yohimbine, COz, caffeine) that may have confo~~mg, and often con~a~cto~, biological effects? Expectancy biases and baseline psychological states thus appeared to play an important role in our patients’ lactate-induced panic, a position that Margraf et al. (1986) and Ehlers et al. (1986) appear to have taken to an extreme. Our findings suggest that lactate sensitivity as a biological marker may be useful when patients’ baseline anxiety levels are relatively low, but it is unclear in its meaning and often unconvincing when anxiety levels are high. At issue is the relative role that conditioning, psychological, and environmental factors play in ph~a~logic~ly induced panic attacks. Biological theories of hypersensitive Badrenergic receptors, intracerebral acidosis, locus ceruleus activation, etc., all assume that a biological mechanism explains lactate-induced panic. They cannot account for the salient role played by expectancy biases, baseline psychological states, and environmental factors in our and others’ studies. Recently, several groups have elaborated upon the phobic response theory of Ackerman and Sacher (1974) that views lactate-induced panic in terms of an interaction among past experience, environmental, and cognitive factors: “It appears that interoceptive fears of bodily anxiety may, in fact, be sufficient to explain the apparent panicogenic properties of lactate and CO;’ (van den Hout et al. 1987, p 788). “As [panic attacks] reoccur during repeated intense, adversive experiences, the abovementioned adrenergic arousal symptoms themselves become interoceptive conditioned stimuli for further anxiety, contributing to the perpetuation of panic anxiety” (Griez et al. 1987, p 94). “The subject has learned to associate certain bodily sensations with acute anxiety by a process of repeated interoceptive conditioning and subsequent cognitive elaboration . . . . The response.to [lactate-inducedstimuli1depends on the perception of the bodily changes and on the appraisal of env~~~~ and internal cues in terms of perceived un~~~n~,
478
BIOL PSYCHIATRY 1989;25:46%481
f.A. Aronson et al.
threat/potentialhartn, and availabilityof responsesor coping strategies”(Margrafet al. 1986, P 43). These investigators note that patients habituate to the panicogenic power of COz, which decreases with repeated exposure (van den Hout et al. 1987). Preliminary data suggest that successful cognitive-behavioral treatment of panic disorder also blocks lactate-induced panic (Guttmacher and Nelles 1984; Gorman 1987). The recent findings of behavioral hypersensitivity to CO;! and weak or no evidence for physiological or biochemical hypersensitivity to CO* (Woods et al. 1988) are consistent with this model of interoceptive conditioning in patients with panic disorder. Can purely psychological and conditioning mechanisms entirely account for our findings? It is doubtful. Our patients experienced lactate-induced panic attacks despite the lack of typical cognitive concomitants of panic and with less intense anxiety than normally experienced. Peripheral cardiovascular and autonomic symptoms occurred later in the development of lactate-induced panic, reducing the likelihood that James-Lange or peripheral interoceptive phenomena gave rise to central symptoms. In fact, in the majority of patients, it occurred the other way around. Patients did not simply endorse any API symptom because of heightened anxiety (Margraf et al. 1986)) but rather reported specific panic symptoms consistent with DSM-IIIR nosology. Finally, there was a differential reactivity in anxiety ratings and heart rates that could not be explained by baseline differences alone. The interoceptive cue theory does not explain why patients do not panic with any challenge that disturbs their internal state, such as hypoglycemia (Udhe et al. 1984; Schweitzer et al. 1986) or respiratory alkalosis (Gorman et al. 1984). Nor does it explain what gives rise to initial panic attacks in the first place. It is a theory of a maintenance mechanism, not a causal theory of panic disorder (van den Hout et al. 1987). Similar to environmental, expectancy, appraisal, and mood variables, it may explain some aspects of lactate-induced panic, but it does not completely explain it away. Biological and psychological factors both appear to be important. The latter are more obvious, the former still are poorly understood. In patients with low baseline anxiety. we speculate that biological factors may predominate; with high baseline anxiety, cognitive and conditioning mechanisms may be more relevant for the researcher and clinician. For ’ ” in panic disorder may be a secondary, example, our findings suggest that the “pamc reactive phenomenon; that is, the proper cognitive set or environmental setting may obviate the occurrence of fears of dying, going crazy, or losing control in the face of the interoceptive cues of a panic attack, but a catastrophic or helpless cognitive set may induce a negative feedback loop that may increase both the “panic” and the interoceptive symptoms. A biopsychosocial theory is now needed that can integrate the biological findings of pontine locus ceruleus dysfunction (Chamey et al. 1984, 1986), medullary nucleus reticularis gigantocellularis dysfunction of respiratory control (Gorman 1987), and limbic alterations in hypothalamic (Udhe 1987) and hippocampal function (Reiman et al. 1986) with cortically mediated psychological mechanisms that modulate the course of panic disorder and experience of lactate sensitivity. Implications for the neurobiology of panic disorder and for therapeutic interventions at different levels are obvious. The gate theory of pain suggests one biopsychosocial model: limbic and brainstem centers may mediate the vulnerability to panic attacks, yet can be gated or modulated by higher cortical/psychological influences that alter the threshold and experience of panic. Finally, our study supports other groups’ findings of a paradoxical therapeutic potential
Lactate Sensitivity in Biological VariabLs
BIOL F’SY&IiLKIRY
1989,2546!Usi
479
with lactate i~sions. van den Hout et al.‘s (1987) report of decreased anxiety responses with serial exposure to CO* inhalation, and Bonn et al.‘s (1974) open trial of 6 lactate
infusions in 33 patients suggest that patients can be habituated to many of the interoceptive aspects of panic and anticipatory anxiety. In other words, anxiogenics can be anxiolytic. Whether such desensiti~tion training makes patients panic-free or simply changes the cognitive response to panic attacks, which continue but without the “panic,” is unclear. Further study of this strategy appears warranted. This study has a number of methodological limitations to the generalizations that can be drawn. This was a pilot project with small patient numbers. The patients experienced a relatively high number of weekly panic attacks, and all 9 panicked with lactate, higher than the 70%-75% reported frequency in previous studies with higher patient numbers. The levels of significance reported (p < 0.05) did not include the Bonferroni correction for the number of statistical tests performed. The study perhaps raises more questions than it answers. The results call for confirmation by studies with larger patient numbers.
References Ackerman S, Sachar E (1974): The lactate theory of anxiety: A review and reevaluation. Psychosom Med 36:69-74.
Aleem A, Rainey JM, Grtiz A, et al (1987): Lactate infusion in post-traumatic stress disorder. Presented at the American Psychiatric Association annual convention, Chicago, IL. American Psychiatric Association (1987): Diagnostic and Statistical Manual of Mental Disease, ed 3 (rev). Washington DC: American Psychiatric Association. AppIeby IL, Klein DF, Sachar El, et al (1981): Biochemical indices of lactate-induced panic: A prelii report. In Klein DF, Rabkin J (eds), Anxiety: New Research and Changing Concepts. New York: Raven Press, pp 411-423. Bonn J, Harrison J, Rees W (1971): Lactate-induced anxiety: Therapeutic application. Br J Psychiatry 119~468-470. Can DB, Sheehan DV, Surman OS, et al (1986): Neuroendocrine correlates of lactate-induced anxiety and their response to chronic alprazolam therapy. Am J Psychiatry 143:483-494. Chamey DS , Heninger GR ( 1986): Abnormal regulation of noradrenergic function in panic disorders: Effects of clonidme in healthy subjects and patients with a~omphobia and panic disorder. Arch Gen Psychiatry 4311042-1054.
Chamey DS, Heninger GR, Breier A (1984): Noradrenergic function in panic anxiety: Effects of yohimbine in healthy subjects and patients with agoraphobia and panic disorder. Arch Gen Pact
41:751-763.
Cohen AS, Barlow DH, Blanchard EB (1985): The psychophysiology of relaxation-associated panic attacks. J Abnorm Psycho1 9496-100. Cowley DS, Dager SR, Foster SI, et al (1987a): Clinical characteristics and response to sodium lactate of patients with infrequent panic attacks. Am J ~~chia~ 144:795-798. Cowley DS, Hyde TS, Dager SR, et al. (1987b): Lactate infusions: The role of baseline anxiety. Psychiatry Res 21:169-179. Dager SR, Cowley DS, Dunner DL (1987a): Biological markem in panic states: Lactate-induced panic and mitral valve prolapse. Biol Psychiatry 22339-359. Dager SR, Rainey JM, Dunner DL, et al (1987b): Central effects of lactate infusion in a primate model. Presented at the American College of Neuropsychopharmacology annual meeting, San Juan, Puerto Rico. Dillon DJ, German JM, Liebowitz MR, et al (1987): Measurement of lactate-induced panic and anxiety. Psychiatry Res 20:97-105.
480
BiOJ..PSYCHIATRY 19EWA-B-481
l’.A. Aronson et al.
Ehlers A, Margraf J, Roth WT, et al (1987): Lactate infusions and panic attacks: Do patients and controls respond differently? Prychiutry Res 17:295-308. Fink M, Taylor MA, Volavka J (1970): Anxiety precipitated by lactate. N Engf J Med 281: 1429. Freedman RR, Ianni P, Ettedgui E, et al (1984): Psychophysiological factors in panic disorder. ~~c~upuf~~ugy 17(Suppl I)&%-73 * Gotman JM (1987): Experimental induction of anxiety by lactate and other substances. Presented at the annual meeting of the American College of Neuropsychopharmacology in San Juan, Puerto Rico. German J, Askanazi J, Liebowitz M, et al (1984): Response to hyperventilation in a group of patients with panic disorder. Am J P~c~j~~ 141:857-861. Griez EJL, Lot&erg H, van den Hout MA, et al (1987): CO2 vulnerability in panic disorder. Psychiatry Res 20:87-95. Guttmacher LB, Nelles C (1984): In vivo desensitization alternation of lactate-induced panic: A case study. Behuv Res Ther 15:369-372. Kelly D, Mitchell-Heggs N, Sherman D (1971): Anxiety and the effects of sodium lactate assessed cfinically and physioiogic~ly Er J P~ch~~~ 119:129-141. Knott V, Chat&y R, Lapierre Y (1981): Panic induced by sodium lactate: Elec~ophysiologic~ correlates. Prog Neuropsychophurmacol5:s 1l-514. Lader M, Mathews A (1970): Physiological changes during spontaneous panic attacks. J Psychosom Res l4:377-381. Liebowitz ML, Fyer A, German J, et al (1984): Lactate provocation of panic attacks: I. Clinical and behavioral findings. Arch GermPsychiatry 41:764-770. Liebowitz MR, Gorman JM, Fyer Aj, et al (1985): Lactate provocation of panic attacks. II. Biochemical and physiological findings. Arch Gen Psychiatry 42:709-719. Liebowitz MR, German JM, Fyer A, et al (1986): Possible mechanisms for lactate’s induction of panic. Am J Psychiatry 143:495-502. Margraf J, Ehlers A, Roth W (1986): Sodium lactate infusions and panic attacks: A review and critique. P~c~o~~rn Med 48:23-5 1. McGrath P, Stewart J, Harrison W, et al (1985): Lactate infusion in patients with depression and anxiety. Psychopharmacol Bull 21:555-557. Pitts FN, McClure JN (1967): Lactate metabolism in anxiety neurosis. N Engl J Med 277:13291336.
Pohl R, Ettedgiri E, Bridges M, et al (1987): Plasma MHPG levels in lactate and iosproterenol anxiety states. Biol P~chja~ 22: 1127-l 136. Rainey JM Jr, Pohl RB, Williams M, et al (1984): A comparison of lactate and iosproterenol anxiety states. Psychopathology 17(Suppl 1):74-82. Rainey JM Jr, Aleem A, Ortiz A, et al (1987): A laboratory procedure for the induction of flashbacks. Am J Psychiatry 144:1317-1319.
Reiman EM, Raichle ME, Butler RE, et al (1986): Application of positron emission tomography to study of panic disorder. Am J P~chi~~ 143:469477. Rifkin A, Klein DF, Dillon D, et al (1981): Blockade by imipramine or desipramine of panic induced by sodium lactate. Am J Psychiatry 138:676-677. Schweitzer E, Winokur A, Rickels K (1986): Insulin induced hypoglycemia and panic attacks. Am J Psychiatry 143:654--655.
Taylor CB, She&h J, Agras WS, et al (1986): Ambulatory heart rate changes in patients with panic attacks. Am J Psychiatry 143:478481. Udhe TW (1987): Biology of panic disorder. Presented at Grand Rounds, State University of New York at Stony Brook, Stony Brook, NY, October 6. Udhe TW, Vittone BJ, Post RM (1984): Glucose tolerance testing in panic disorder. Am J Psychiatry 141:1461-1463.
Lactate Sensitivity in Biological Variables
BIOL PSYCHIATRY 1989;25:469-481
481
van den Hout MA, van der Molen GM, Griez E, et al (1987): Reduction of CO&duced anxiety in patients with panic attacks after repeated CO* exposure. Am J Psychiatry 144:788-791. Woods SW, Chamey DS, Goodman WK, et al (1988): Carbon dioxide-induced anxiety. Arch Gen Psychiatry 45143-52.
Yeragani VK, Balon R, Pohl R, et al (1987): Do higher preinfusion heart rates predict laboratoryinduced panic attacks? Biol Psychiatry 22:554-555.