Background Ten to 30 percent of patients undergoing cardiac catheterization because of chest pain are found to have normal coronary angiograms. Because these patients may have a visceral pain syndrome unrelated to myocardial ischemia, we investigated whether drugs that are useful in chronic pain syndromes might also be beneficial in such patients.
Methods Sixty consecutive patients underwent cardiac, esophageal, psychiatric, and pain-sensitivity testing and then participated in a randomized, double-blind, placebo-controlled three-week trial of clonidine at a dose of 0.1 mg twice daily (20 patients), imipramine at a dose of 50 mg nightly with a morning placebo (20 patients), or placebo twice daily (20 patients); this treatment phase was compared with an identical period of twice-daily placebo for all patients (placebo phase).
Results Thirteen (22 percent) of the 60 patients had ischemic-appearing electrocardiographic responses to exercise, 22 of the 54 tested (41 percent) had abnormal esophageal motility, 38 of 60 (63 percent) had one or more psychiatric disorders, and 52 of 60 (87 percent) had their characteristic chest pain provoked by right ventricular electrical stimulation or intracoronary infusion of adenosine. During the treatment phase, the imipramine group had a mean (±SD) reduction of 52 ±25 percent in episodes of chest pain, the clonidine group had a reduction of 39 ±51 percent, and the placebo group a reduction of 1 ±86 percent, all as compared with the placebo phase of the trial. Only the improvement with imipramine was statistically significant (P = 0.03). Repeat assessment of sensitivity to cardiac pain while the patients were receiving treatment showed significant improvement only in the imipramine group (P = 0.01). The response to imipramine did not depend on the results of cardiac, esophageal, or psychiatric testing at base line, or on the change in the psychiatric profile during the course of the study, which generally improved in all three study groups.
Conclusions Imipramine improved the symptoms of patients with chest pain and normal coronary angiograms, possibly through a visceral analgesic effect.
Several possible causes of chest pain in such patients have been proposed, including psychiatric illness,10,11,12,13 esophageal dysfunction,14,15,16,17 and coronary microvascular dysfunction18,19,20,21,22; overlap among these conditions has also been reported23,24,25. Recently, several groups of investigators have described exaggerated sensitivity to cardiac26,27,28 and esophageal25,29 pain in patients who have chest pain despite normal coronary angiograms. We conducted a study to categorize the cardiac, gastrointestinal, and psychiatric findings and the results of pain-sensitivity testing in a consecutive series of patients who had chest pain despite normal coronary angiograms. We also assessed the effect of imipramine and clonidine -- drugs that have been used successfully in the management of chronic pain syndromes -- in the same patients in a randomized, double-blind, placebo-controlled trial.
Methods
Patients
We studied 60 consecutive patients who met the following entry criteria: normal coronary angiograms and no epicardial coronary-artery spasm after the intravenous administration of ergonovine (0.15 mg), normal left ventricular function at rest, no evidence of left ventricular hypertrophy or valvular heart disease (including mitral-valve prolapse) on echocardiography, blood pressure no higher than 160/100 mm Hg when they were receiving no medications, and no musculoskeletal sensitivity that accounted for their characteristic chest pain. All the patients had been referred to the National Institutes of Health because of recurrent chest pain; we made no attempt to select patients for the study on the basis of the results of any previous testing. Fifty-seven patients had previously been given antiischemic therapy (calcium-channel blockers, beta-blockers, or nitrates), which had failed to control their symptoms; the remaining three patients were untreated. All the patients discontinued all medications except estrogen replacement (15 patients), thyroid hormone replacement (6 patients), and insulin (2 patients) for at least one week before admission. This study was approved by the National Heart, Lung, and Blood Institute Review Board, and informed consent was obtained from all patients.
Base-Line Studies
All patients underwent symptom-limited treadmill exercise testing using the standard Bruce protocol30 and gated blood-pool scanning with technetium-99m radionuclide angiography both at rest and during exercise31. After an overnight fast and the ingestion of 10 mg of diazepam, all patients underwent cardiac catheterization, during which a 6-French pacing wire was advanced to the apex of the right ventricle. Pacing was then initiated without the patient's awareness at a rate 5 beats per minute faster than the resting heart rate, with the stimulus intensity starting at 1 mA and increasing by 1-mA increments at 10-second intervals to 10 mA. Adenosine (2.2 mg per minute for two minutes) was later infused into the left coronary artery of 57 patients. If chest pain was reported during either test, the patients were asked whether this pain was the same as their characteristic chest pain.
After an overnight fast and anesthesia with topical lidocaine, patients underwent testing of esophageal motility and modified Bernstein testing, as described previously25. To assess esophageal sensitivity, a polyvinyl balloon was inflated 6 to 8 cm proximal to the catheter tip with 1-ml increments of air until the patient had chest pain or the balloon reached a diameter of 30 mm. The test was interpreted as positive if balloon distention precipitated the patient's characteristic chest pain. All patients were interviewed by the same psychiatric nurse using the Schedule for Affective Disorders and Schizophrenia -- Lifetime Version, modified for the study of anxiety disorders32,33. Because chest pain was a criterion for participation in this study, this symptom was not used for the diagnosis of panic disorder. Responses to the diagnostic interview were reviewed by the nurse interviewer and two research psychiatrists, and psychiatric diagnoses were determined according to the criteria of the Diagnostic and Statistical Manual of Mental Disorders, third edition, revised (DSM-III-R)34. Eleven standardized rating scales for psychiatric symptoms and functional impairment were also administered35,36,37,38.
Placebo-Controlled Trial
Placebo Phase
Because we expected that there would be substantial response to placebo in a group of patients with a chronic pain syndrome who had recently undergone an extensive evaluation and who had been reassured about the benign nature of their symptoms, we initially treated all patients with single-blind placebo. The response to subsequent double-blind active treatment could then be compared with the effect of placebo, if any. Accordingly, after the completion of the base-line testing, all patients were discharged with directions to take medication (placebo) twice daily (at 9 a.m. and 9 p.m.) for five weeks. All the patients were issued a symptom diary and were instructed to record each episode of chest pain they experienced. Patients were also instructed to rate the average intensity of chest pain for each day by choosing 1 of 13 terms, ranging from "no pain" to "extremely intense pain." The relative magnitude of pain was quantified on a logarithmic scale and averaged for each week39,40. All the patients returned to the clinical center during the fifth week of the placebo phase, at which time their diaries and ratings of the intensity of chest pain were collected and inspected for compliance. All the patients underwent psychiatric interviews while continuing to take the placebo.
Treatment Phase
All the patients were then randomly assigned in a double-blind fashion to one of three treatment groups: imipramine (25 mg at 9 p.m. for one week, then 50 mg nightly thereafter, always with a matching placebo at 9 a.m.), clonidine (0.05 mg at 9 a.m. and 9 p.m. for one week, then 0.1 mg for both doses thereafter), or placebo at 9 a.m. and 9 p.m. Thus, patients took the full dose of treatment for three weeks (weeks 2 through 4) before returning to the clinical center. All capsules for the single-blind placebo phase and the double-blind treatment phase were identical in appearance. During the fifth week of the treatment phase, all the patients returned and their symptom diaries and ratings of chest-pain intensity were collected. While treatment continued, they underwent testing of sensitivity to cardiac pain, testing of esophageal motility, balloon-distention testing of pain sensitivity, and a psychiatric interview. During this double-blind treatment phase, all tests and analyses were performed by personnel without knowledge of the patients' treatment assignment.
Statistical Analysis
Data are presented as means ±SD. Because patients received the full dose of imipramine, clonidine, or placebo during weeks 2 through 4, before returning to the clinical center, we analyzed the frequency of chest pain during these three weeks of the treatment phase and compared the results with the frequency of pain during the corresponding three weeks of the placebo phase. The changes in the frequency of chest pain, ratings of the intensity of chest pain, and responses to psychiatric testing from the placebo phase to the treatment phase (and, for psychiatric testing, from base line to the treatment phase) were compared for the three treatment groups by repeated-measures two-way analysis of variance with assessment of possible interactions between drug and time or drug and diagnosis41. If statistical significance was identified (indicated by a two-tailed P<0.05), intergroup comparisons were performed with Bonferroni's or Dunnett's test to assess the relative effect of the three treatments41. Furthermore, if a statistically significant interaction between drug and diagnosis was found, stratified analyses comparing normal patients and those with abnormal conditions were performed for each treatment group. The results of testing for sensitivity to cardiac and esophageal pain at base line and during the treatment phase were analyzed with an exact binomial test. Categorical data were compared by chi-square analysis.
Results
Of the 60 patients enrolled in the study, 40 were women and 20 were men. The average age was 50 years (range, 29 to 72). The duration of chest-pain symptoms averaged 53 months (range, 3 to 175). Forty-one patients (68 percent) had been hospitalized for chest pain, excluding hospitalizations for the initial cardiac catheterization and the evaluation at the National Institutes of Health, and 30 patients (50 percent) had been hospitalized for chest pain more than once (range, 2 to 25 additional hospitalizations). Twenty-nine patients (48 percent) had been hospitalized for chest pain despite having had normal coronary arteries on angiography during a prior hospitalization (range, 1 to 14 hospitalizations). Eighteen patients (30 percent) had undergone more than one cardiac catheterization before their evaluation at the National Institutes of Health (range, two to nine additional cardiac-catheterization procedures). The clinical characteristics of the patients at randomization are presented in Table 1; there were no substantive differences among the groups.
|
Twenty-three patients (38 percent) had one or more abnormalities on exercise testing (defined as a horizontal or downsloping ST-segment depression of more than 1 mm on the electrocardiogram, bundle-branch-block pattern on the electrocardiogram, decrease in the left ventricular ejection fraction, or wall-motion abnormality). Of these, 13 patients had electrocardiographic responses to exercise associated with chest pain that were similar to that caused by ischemia, commonly designated as syndrome X22.
During cardiac catheterization, 47 patients (78 percent) had chest pain during right ventricular pacing at a rate 5 beats faster than their resting heart rate (median stimulus intensity at onset of pain, 3 mA); 40 of these 47 (85 percent) considered it their characteristic chest pain. During the infusion of adenosine in 57 patients, 54 patients (95 percent) had chest pain, in all cases without electrocardiographic changes characteristic of ischemia; 47 (87 percent) considered it their characteristic chest pain. Overall, the characteristic chest pain was provoked in 52 of 60 patients (87 percent) by either right ventricular pacing or the intracoronary infusion of adenosine.
Fifty-four patients (90 percent) underwent esophageal testing; the remaining six patients were unable to tolerate the motility-testing catheter. Twenty-two of the 54 patients tested (41 percent) met the criteria for a motility disorder (distal high-amplitude peristalsis, 12 patients; nonspecific motility disorder, 10 patients) and 22 (41 percent) had their characteristic chest pain provoked by the administration of edrophonium, the infusion of hydrochloric acid (Bernstein test), intraesophageal balloon distention, or more than one of these tests. Thirty-eight patients (63 percent) had fulfilled DSM-III-R criteria for one or more psychiatric diagnoses at some time during their lives; these were panic disorder (with chest pain excluded as a diagnostic criterion) in 26 patients, a history of major depression in 17, somatization in 11, current major depression in 3, hypochondriasis in 2, alcohol dependence in 2, other anxiety disorders in 5, and other non-anxiety diagnoses in 7. Patients with psychiatric diagnoses at any time during their lives were no more likely to have esophageal dysmotility (41 percent), esophageal pain sensitivity (38 percent), or cardiac pain sensitivity (87 percent) than patients without psychiatric diagnoses (40 percent, 45 percent, and 86 percent, respectively). There were no significant differences in the results of base-line cardiac, esophageal, or psychiatric testing among the treatment groups (Table 2).
|
During the three weeks of full-dose treatment, the average change in the frequency of chest pain from the placebo phase was a reduction of 52 ±25 percent for the patients in the imipramine group (from 88 ±108 episodes during the preceding three weeks of the placebo phase); all 20 patients in this group reported a reduction in the frequency of chest pain. There was a reduction of 39 ±51 percent in the clonidine group (from 63 ±62 episodes during the placebo phase), with 15 of the 20 patients in this group reporting a reduction in the frequency of chest pain. In contrast, the average change in the frequency of chest pain in the placebo group, as compared with the corresponding three weeks of the placebo phase, was a reduction of 1 ±86 percent (from 54 ±45 episodes), with 14 of 20 patients reporting a reduction. The reduction in the frequency of chest pain in the imipramine group was significantly greater than the response in the placebo group (P = 0.03).
Although there were more women than men in the imipramine group (16 vs. 4), the proportion of women was not significantly different from that in the clonidine and placebo groups, and the response of women and men to imipramine was similar, with women reporting a 53 ±25 percent reduction and men reporting a 48 ±24 percent reduction in the frequency of pain. Ratings of the intensity of chest pain were significantly lower during the administration of imipramine (P = 0.001) and clonidine (P = 0.002), as compared with the placebo phase, but the difference between the treatment phase and the placebo phase in the placebo group was not significant (P = 0.31). However, the magnitude of the effects of imipramine and clonidine on the intensity of chest pain was not significantly different from that of the effects in the placebo group (P = 0.23 by analysis of variance).
Relation of Treatment Response to Results of Base-Line Testing
In all three treatment groups, there were no differences in the degree of change in the frequency of chest pain from the placebo phase to the treatment phase between patients with normal results on noninvasive cardiac or esophageal testing at base line and those with abnormal results (Figure 1). Repeat assessment of right ventricular sensitivity during treatment showed a significant reduction in the prevalence of chest pain provoked by right ventricular electrical stimulation in the imipramine group (P = 0.01), but not the clonidine or placebo group (Figure 2).
|
|
|
|
During the treatment phase, side effects considered to be a consequence of treatment were reported by 15 patients in the imipramine group, 14 patients in the clonidine group, and 11 patients in the placebo group. In the imipramine group, electrocardiograms obtained at week 5 of the treatment phase showed slight but significant prolongation of the corrected QT (QTc) interval, as compared with the placebo phase, from 0.42 ±0.04 to 0.43 ±0.03 second (range of prolongation, -0.01 to 0.04 second; P = 0.02). In the three patients with bundle-branch-block patterns in this group, imipramine increased the QTc interval by 0.02, 0.01, and 0.01 second. At the relatively low dose used in this study, no patient in the imipramine group reported alterations in consciousness during treatment, which might suggest a proarrhythmic effect of the drug.
At the end of the study, 58 patients were offered long-term open-label imipramine treatment (2 patients preferred to continue to receive clonidine). After an average follow-up of 21 months (range, 9 to 33), we were able to contact 55 of these 58 patients to ascertain their current treatment regimen; 40 of 55 (73 percent) were still receiving imipramine (dose range, 50 to 150 mg per day; mode, 50 mg per day). No patient reported symptoms suggesting a proarrhythmic effect of the drug.
Discussion
Our major finding is that imipramine reduced the frequency of chest pain in patients with the syndrome of chest pain and normal coronary angiograms by approximately 50 percent. The benefit of imipramine was independent of the results of extensive cardiac, esophageal, and psychiatric testing and was observed in men as well as women. In the majority of patients in this study (87 percent), the characteristic chest pain was provoked by right ventricular electrical stimulation or infusion of adenosine; only in the imipramine group was there a significant reduction in the incidence of chest pain provoked by right ventricular electrical stimulation during the treatment phase, as compared with the base-line response. The symptomatic benefit of imipramine in our study is consistent with the results of previous placebo-controlled trials of antidepressants in the treatment of chronic pain syndromes such as painful diabetic neuropathy,42,43,44 postherpetic neuralgia,40,45 migraine headache,46,47 presumed esophageal pain,48 and other chronic pain syndromes49. Blockade of norepinephrine reuptake by drugs such as imipramine may enhance the inhibitory action of pain-modulating neurons that descend from the brain stem to the spinal cord50,51. Despite the benefit of the 
2-adrenoreceptor agonist clonidine in other pain syndromes,52,53 the effect of clonidine on symptoms in our study was not significantly different from that of placebo.
Electrocardiography during exercise stress testing and radionuclide angiography did not identify patients who were more or less likely to respond to imipramine. Indeed, six patients in the imipramine group fulfilled the descriptive criteria for syndrome X, in that they had an apparently ischemic electrocardiographic response to exercise in association with chest pain22. The effect of imipramine on their symptoms and their sensitivity to cardiac pain was similar to that in the other 14 patients in this treatment group.
Esophageal dysmotility is reportedly common in patients with chest pain and normal coronary angiograms14,15,17,25. However, esophageal-motility testing in our study did not identify patients who were more or less likely to respond to imipramine. Furthermore, there were no significant differences among the three treatment groups with regard to the prevalence of esophageal dysmotility or pain in response to balloon distention between base-line testing and repeat testing during the treatment phase.
A high prevalence of psychiatric disorders has been reported in several studies of patients who have chest pain despite normal coronary angiograms,10,11,12,13 and panic disorder is a particularly common finding11,13. However, the patients in our study who had psychiatric disorders at base line were no more likely to have exaggerated sensitivity to cardiac or esophageal pain than patients without psychiatric disorders. Furthermore, patients with psychiatric disorders at base line who were treated with imipramine or clonidine were no more likely to have a reduction in the frequency of chest pain than patients without psychiatric disorders. Serial assessment with psychiatric rating scales during the study generally showed psychiatric improvement over time, regardless of treatment.
In our study, the reduction in sensitivity to cardiac pain among patients who received imipramine suggests that this drug may have a visceral analgesic effect, at the dose used in this study, that is unrelated to changes in esophageal motility or to the psychiatric profile. Although imipramine in doses used to treat depression has been associated with electrophysiologic effects similar to those of class I antiarrhythmic drugs,54 at the relatively low dose used in this study the effect of imipramine on the repolarization pattern of the electrocardiogram was minimal. Moreover, the absence of side effects suggestive of a proarrhythmic effect of imipramine during the study and an average follow-up period of almost two years indicate that the drug is safe at this dose in patients with structurally normal hearts. Because the symptomatic benefit of imipramine was independent of the results of cardiac, esophageal, or psychiatric testing before treatment, this drug may have wide applicability in the management of chest pain in patients with normal coronary angiograms who continue to have symptoms despite reassurance that their prognosis is good.
Source Information
From the Cardiology Branch (R.O.C., A.A.Q., R.M., A.M.S.) and the Biostatistics Research Branch (M.A.W.), National Heart, Lung, and Blood Institute; the Neurobiology and Anesthesiology Branch, National Institute of Dental Research (R.H.G, W.B.S.); and the Section on Anxiety and Affective Disorders, National Institute of Mental Health (M.F.G., B.C.B., T.W.U.) -- all at the National Institutes of Health, Bethesda, Md.; and the Division of Gastroenterology, Georgetown University Hospital, Washington, D.C. (K.M., S.B.B.). Presented in part at the 57th Scientific Meeting, American College of Gastroenterology, Miami Beach, Fla., October 26, 1992, and at the 42nd Scientific Session, American College of Cardiology, Anaheim, Calif., March 18, 1993.
Address reprint requests to Dr. Cannon at Bldg. 10, Rm. 7B-15, National Institutes of Health, Bethesda, MD 20892.
References
The New England Journal of Medicine is owned, published, and copyrighted © 2004 Massachusetts Medical Society. All rights reserved.
 |
 |
 |
 |
