Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Evaluation and Outcomes of Patients With Palpitations Barbara E. Weber, MD, MPH, Rochester, New York, Wishwa N. Kapoor, MD, MPH, Pittsburgh, Pemsy/vania PURPOSE: To determine: (1) the etiologies of palpitations, (2) the usefulness of diagnostic tests in determining the etiologies of palpitations, and (3) the outcomes of patients with palpitations. PAIIENTS AND METHODS: One hundred and ninety consecutive patients presenting with a complaint of palpitations at a university medical center were enrolled in this prospective cohort study. Patients undement a structured clinical interview and psychiatric screening. The charts were abstracted for results of the physical exam and tests ordered by the primary physician. Assignment of an etiology of palpitations was based on strict adherence to predetermined criteria and achieved by consensus of the two physician investigators. One-year followup was obtained in 96% of the patients. RESULTS An etiology of palpitations was determined in 84% of the patients. The etiology of palpitations was cardiac in 43%, psychiatric in 31%, miscellaneous in lo%, and unknown in 16%. Forty percent of the etiologies could be determined with the history and physical examination, an electrocardiogram, and/or laboratory data. The l-year mortality rate was 1.6% (95% confidence interval [Cl] 0% to 3.4%) and the l-year stroke rate was 1.1% (95% Cl 0% to 2.6%). W&in the first year, 75% of the patients experienced recurrent palpitations. At l-year follow-up, 89% reported that their health was the same or improved compared to that at enrollment, 19% reported that their work performance was impaired, 12% reported that workdays were missed, and 33% repotted accomplishing less than usual work at home. CONCLUSIONS: The etiology of palpitations can often be diagnosed with a simple initial evaluation. Psychiatric illness accounts for the etiology in nearly one third of all patients. The short-term prognosis of patients with palpitations is excellent with low rates of death and stroke at 1 year, but From the Universitv of Prttsburah School of Medicine (WNK),University of Pittsburgh Medical Center, Pi&burgh, Pennsylvania,and the University of Rochester, School of Medicine and Dentistry (BEW),St. Mary’s Hosoltal. Rochester. New York. D;. Wikhwa N. Kapoor is a recipient of a Research Career Development Award from the National Heart, Lung, and Blood Institute (K04L 01899). Requests for reprints should be addressed to Barbara E. Weber, MD, MPH, St. Mary’s Hospital. 89 Genesee Street, Rochester, New York 14611. Manuscript submitted December 29, 1994 and accepted in revised form July 10, 1995. 138 February 1996 The American Journal of Medicine” there is a high rate of recurrence of symptoms and a moderate impact on productivity. P alpitations are one of the most common symptoms in general medical settings, reported by as many as 16% of the patients.’ This symptom may be caused by a variety of disorders, ranging from lifethreatening conditions such as ventricular tachycardia2 to various psychiatric illnesses.3 As a result, patients with palpitations often undergo a wide variety of diagnostic tests and referrals leading to substantial resource utilization.* Currently, clinical experience guides the physician caring for patients with palpitations, since there are no prior studies that describe the spectrum of etiologies or the usefulness of diagnostic tests in the evaluation of palpitations. Furthermore, the outcome of patients with palpitations has not been well described. In the only retrospective study of outcomes in patients with palpitationq5 cases with palpitations and controls without palpitations experienced similar rates of cardiac endpoints. The purpose of this prospective study was to determine (1) the etiologies of palpitations, (2) the usefulness of diagnostic tests in determining the etiologies of palpitations, and (3) the outcomes of patients with palpitations. PATIENTS AND METHODS This was a prospective cohort study of patients presenting with palpitations. Palpitations were defined as one or more of the following patient complaints: fast heart beats, skipped heart beats, irregular heart rate, and heart fluttering, racing, or pounding.” Study Entry Criteria Between January 2 and August 30, 1!391, all patients presenting to the emergency department, admitted to the medical and surgical inpatient service, or attending the medical clinics of the University of Pittsburgh Medical Center were screened for study eligibility. Patients presenting to the psychiatric emergency department or admitted directly to the psychiatric service were not screened far study eligibility. Inclusion criteria were palpitations as a chief complaint for seeking medical care or palpitations as one of the chief complaints during a routine visit to the physician. The symptom must have occurred at least once in the 3 months preceding this index visit. Patients were excluded if their age was less than 18 years, they were known to be aphasic or demented, Volume 100 PATlEriTS WITH PALPITATiONS/WEBER AND KAPOOR or were unable to speak English. Patients transferred from the inpatient service of another hospital, those admitted only for same-day surgery, and patients with palpitations only elicited on review of systems were also excluded. Patient Identification Because palpitations are often not the sole diagnosis for admission or discharge, and patients with palpitations are often labeled with a more specific diagnosis, a comprehensive search strategy was used to capture all patients with palpitations. To identify patients eligible for the study, we performed daily review of the following lists: emergency department discharge diagnoses, hospital admission diagnoses, and outpatient visit discharge diagnoses. If any of the diagnoses noted in Table I was found, the medical chart was reviewed and the patient’s physician was contacted to ascertain if the chief complaint was palpitations. In cases in which this could not be ascertained, the patient was contacted for clarification. Patients who met the entry criteria were asked to participate in the study. Conduct of the study was approved by the institutional review board of the University of Pittsburgh. TABLE I Diagnoses Used to Identify Patients Anemia Transfusion Anxiety Aortic aneurysm Arrhythmia Specific arrhythmias (ie, atrial fibrillation, atrial flutter, bradycardia, multifocal atrial tachycardia, pre-excitation syndrome, sick sinus syndrome, supraventricular tachycardia, ventricular tachycardia, ventricular fibrillation) Atrial myxoma Cardiomyopathy Chest pain Cocaine abuse Congestive heart failure Dizziness Drug toxicity (ie, amphetamines, theophylline) High output failure Hypoglycemia Pacemaker failure Palpitations Panic attack Pheochromocytoma Rule out myocardial infarction Shunt or fistula (peripheral or cardiac) Syncope Thyrotoxicosis Unstable angina Valvular heart disease Patient Evaluation Patients who agreed to participate were interviewed. The structured interview was directed at the following issues: palpitation characteristics, associated symptoms and situations, drug and medication use, and comorbid illness. Medical charts were reviewed and data were abstracted regarding physical examination findings and results of the diagnostic evaluation. This process was completed by the principal investigator (BEW) as soon as possible after the patient presented for medical care. Interviews were performed in person with 42%and on the phone with 58%of patients. The mean time in days between event and evaluation was 1.1, evaluation and interview was 3.4, and event and interview was 4.2. The median time in days between event and evaluation was 0, evaluation and interview was 2, and event and interview was 2. The physical exam and diagnostic evaluation (ie, electrocardiogram [ECG], laboratory tests, arrhyLhmia detection) was determined by the individual physician seeing the patient at the index visit. In cases in which tests for arrhythmia detection were not ordered by the clinician, the investigators made loop monitors available. To screen for generalized anxiety disorder, panic attack, panic disorder, major depression, and somatization disorder, patients were asked to complete two validated self-administered instruments: the General Health Questionnaire (GHQ),7-12and the somatization screening test of Othmer and DeSouza13(SOM). The 30question GHQ was answered using a 4 point Likert scale and scored with 1 point for each response of 3 or 4, allowing for a maximum score of 30. The 7ques tion SOM was scored with 1 point for each positive re sponse. Patients with a GHQ score14J5of 25 or a SOM score16of 23 or in whom these psychiatric illnesses were clinically suspected were further assessedwith the Diagnostic Interview Schedule (DIS)17Jssections for generalized anxiety disorder, panic attack and disorder, major depression, and/or somatization disorder. The DIS was adininistered over the telephonelg by a trained certified registered nurse practitioner. The Diagnostic and Statistical Manual of Mental Disorders, third edition revised, (DSM-III-R>2Ocriteria were used to score the DIS. Generalized anxiety disorder, panic attack and disorder, and somatization disorder were considered present if symptoms were present within the last 6 months. Depression was evaluated as a lifetime disorder. Depression was considered to be C(F morbid and not etiologic, since palpitations are not listed as a criterion symptom in DSM-III-R. Assignment of Etiology of Palpitations Diagnostic criteria for the etiology of palpitations were developed prior to the start of the study after extensive review of the literature (see Appendix). We appraised pertinent articles, case reports, review papers, and cardiology and general medicine text- February 1996 The American Journal of Medicinea Volume 100 139 TABLE Ill TABLE II Patient Characteristics Mean age (yl Range Female White 21 year of college education Site of presentation Emergency department Medical clinic Symptom presentation Chief complaint Complaint during a routine visit Admitted to hospital History of prior palpitations History of Heart disease+ Hypertension Congestive heart failure Diabetes mellitus (N = 190)’ 46 18-87 61 74 57 62 28 87 13 35 77 32 29 13 7 ‘Other than mean age, data shown represent the percentage of patients included. ‘Coronary (ie, history of myocardial infarction, angina, coronary artery bypass grafting, percutaneous transluminal coronary angioplasty), congenital, or valvular heart disease, or cardiomyopathy. books to find disorders associated with the symptom of palpitations. Assignment of an etiology was based on strict adherence to these criteria and achieved by consensus of the two physician investigators. The diagnostic criteria permitted a simple categorization of the etiologies of palpitations. Cardiac etiology included arrhythmias, cardiac and extra-cardiac shunts, regurgitant vahular heart disease, pacemaker, prosthetic heart valve, cardiomegaly, mitral valve prolapse, hyperkinetic heart syndrome, and atrial myxoma. Psychiatric etiology included panic attack, panic disorder, generalized anxiety disorder, and somatization disorder. The category of miscellaneous etiology included medications, habits, metabolic disorders, high output states, dehydration and orthostatic hypotension, and exertion. Given the available information (ie, history, physical, diagnostic evaluation, psychiatric testing), the investigators considered each etiology (see Appendix). Because correlation of symptoms with documented arrhythmias was the most concrete example of causality, whenever this occurred the diagnosis of arrhythmia was assigned. The remaining etiologies were carefully considered for their presence or absence. This hierarchy always designated cardiac arrhythmias as the etiology, although psychiatric or metabolic comorbid conditions may have been present. Similarly, metabolic disorders were considered to be the etiology, although psychiatric conditions may have been present. These diagnoses were considered definite, with only three exceptions. Diagnoses involving medications or 140 February 1996 The American Journal of Medicine* Etiologies of Palpitations No. Cardiac 82 Atrial fibrillation 19 Supraventricular tachycardia 18 Premature ventricular beats 15 Atrial flutter 11 Premature atrial beats 6 Ventricular tachycardia 4 Mitral valve prolapse 2 Sick sinus syndrome 2 Pacemaker failure 2 Aortic insufficiency 2 Atrial myxoma 1 Psychiatric 58 Panic attack or disorder plus anxiety 20 Panic attack alone 17 Panic disorder alone 14 Anxiety alone 6 Panic plus anxiety plus somatization 1 Miscellaneous 19 Medication 5 Thyrotoxicosis 5 Caffeine 3 2 Cocaine 2 Anemia Amphetamine 1 1 Mastocytosis Unknown 31 Percent 43.2 10.0 9.5 7.9 5.8 3.2 2.1 1.1 1.1 1.1 1.1 0.5 30.5 10.5 8.9 7.4 3.2 0.5 10.0 2.6 2.6 1.6 1.1 1.1 0.5 0.5 16.3 habits were considered to be definite, probatble,or possible, adapted from the literature on adverse drug effects.*l Arrhythmias (only supraventricular tihycardia and ventricular tachycardia) and anxiety were considered definite or probable. Only 5.8%of the patients had etiologies that were considered probable or possible. As this is the first study to explore the etiologies for the symptom of palpitations, our focus was on the spectrum of etiologies; therefore, for the analyses presented here, the categories of definite, probable, and possible were combined. These diamostic criteria allowed patients to be assigned to only one etiology category, but could have more than one diagnosis within that category (ie, premature ventricular contraction and premature atrial contraction, cocaine and caffeine, but not supraventricular tachycardia and anxiety). Follow-Up All patients were contacted by phone at 3 and 6 months, and for final follow-up at 9 and/or 12 months. Responses to standard questions regarding recurrences, new cardiovascular events, and mortality were obtained from the patient, the family, or caregiver; interviewers were trained to avoid leading questions and to be specific with respect to morbidity related only to palpitations. All interviews were completed by October 28, 1992; over 90% of inter- Volume 100 PATIENTS WITH PALPITATlONS/WEBER AND KAPOOR TABLE N Selected Significant Variables and Their Relationship to Etiology’ Cardiac Psychiatric Miscellaneous Demographic Mean age (y) Male sex (%I White (%) Comorbidity History of heart disease (%I Symptom characteristics Irregular heart beat (%) Duration >5 minutes (%I Total number of symptoms @I!-131 Unknown 50 51 85 41 29 67 46 26 74 47 32 58 45 19 16 32 34 21 67 4.3 50 50 2.9 :: 51 3.5 5.3 'P do.05 (analysis of variance for age and number of symptoms and Fisher’s exact test for categorical variables). One or more of the etiology categories is different from the others. views were performed by the same physician’s assis- gender from those who refused. Selected features of tant. Whenever follow-up events occurred, attempts the 190 enrolled patients are shown in Ttible Il. were made to obtain further details from the primary physician and medical chart. The cause of death was Etiologies assigned on the basis of information obtained from An etiology for palpitations was assigned in 159 the patient’s family and the medical chart. [84%) patients. The specific etiologies are listed in Table III. Overall, 43%of the patients had1a cardiac Statistical Analyses etiology, 31%had a psychiatric etiology, 10%had misStandardized forms for entry of clinical, laboratory, cellaneous etiologies, and 16%were unknown. For the and outcome data were utilized. Data management and subgroup of patients presenting to the medical clinic, analyses were conducted with the use of RBase the etiology was cardiac in Zl%, psychiatric in 45%, (Microrim Inc, Seattle, Washington),22 BMDP (UC miscellaneous in 696,and 28% unhewn. For the subPress, Berkeley, California),23 StatXact (Cytel Corp, group of patients presenting to the emergency deCambridge, Massachusetts),24and ROCFIT (University partment, the etiology was cardiac in 47%,psychiatric of Chicago, Chicago, Illin~is)~~ software packages. in 27%, miscellaneous in 13%, and 13% unknown. Statistical tests included the &i-square, Fisher’s exact, There was a significant difference in etiology by site and analysis of variance tests to evaluate differences of presentation (P <0.002). The distribution of etibetween groups. Logistic regression analysis (BMDP ologies was not statistically different when those with LRa) was used to test for independence among pre- prior palpitations were compared to those without dictors of a cardiac etiology of palpitations. Our intent prior palpitations (P = 0.24). was to create a clinical prediction model that incorTwenty-four patients were assigned more than one porated variables readily available in the initial history. etiology: 21 were coexisting psychiatxic illnesses (eg, All variables tested were dichotomous, including the generalized anxiety disorder and panic disordler), 2 had composite variable ‘history of heart disease’, which both symptomatic premature atrial and ventricular was considered present if any of the following was pre- beats, and 1 had coexisting cocaine and ctieine use. Of the 159 patients for whom an etiology could be sent: histow of angina, myocardial infarction, cardiac surgery, percutaneous lxansluminal coronary angio- determined, 148were definite, 10 were probable, and plasty, pulmonary hypertension, congestive heart fail- 1 was possible. In the miscellaneous category, there ure, and va,lvular or congenital disease.To evaluate this were 4 probable medication and habit etiologies and 1 possible medication/thyrotoxicosis etiology. In the prediction model, a receiver-operatingxharacteristic (ROC) curve25was constructed based on the number psychiatric category, there were 3 probable anxiety of multivariate predictors of cardiac etiology present. etiologies. In the cardiac category, there were 3 probOne-year mortality and stroke rates were calculated able arrhythmias. (See Methods section for specific definitions). using the Kaplan-Meier method.26 Table IV lists selected variables (ie, demographic, RESULTS historical, and symptom characteristic) and their reEntry criteria were met by 229 patients. Thirty-nine lationship with the four categories of etiology. The patients were not enrolled due to patient or physician cardiac group had the highest mean age and the greatrefusal. Therefore, 83% agreed to participate. Those est percentage of males, and its patients were more agreeing to participate did not differ in age, race, or likely to describe an irregular heart beat and report February 1996 The American Journal of Medicine’ Volume X00 141 TABLE V Multivariate Predictors of Cardiac Etiology Odds 95% Confidence Ratio Interval Male sex Description of an irregular History of heart disease >5minute duration of palpitation event beat 2.6 3.2 3.5 5.7 1.2 to 5.4 1.5 to 6.8 1.6 to 7.8 2.4 to 13.7 TABLE VI Evaluation Methods and the Number of Patients for Whom Each Test Revealed the Etiology of Palpitations Initial database History and physical alone Electrocardiogram’+ Laboratory alone History and physical plus laboratory Diagnosis suggested by history and physical Diagnostic Interview Schedule Monitoring’ Telemetry Halter monitor Loop monitor 13 43 5 3 4 55 12 8 3 2 Telemetry plus Holter Dthert 3 Electrophysiologic study 1 Echoaardiogram 1 Aortogram 1 Pacemaker evaluation 5 Previous diagnosis or evaluation 159s Total ‘The numberof tests performed for arrhythmiadetectionwas as follows: 166 electrocardiograms, 79 radioelectrocardiograms(telemetty), 53 Hotter mom itors, 10loop monrtors,10electrophysiologtcstudies,and 2 pacemakerevaluations. ‘Pacemaker evaluation plus electrocardiogram was diagnostic in 1 patient. $The number of selected other tests performed was as foltows: 93 chest roentgenograms, 48 echocardiograms, 16 exercise treadmill tests, 9 cardiac catheterizations, and 6 muitigated angiograms. §Anetiology could not be determined in 31 patients. the duration of their palpitation event as greater than 5 minutes. The highest mean number of associated symptoms was reported by patients in the psychiatric group. One variable that did not distinguish the etiologies was a history of prior palpitations. After the noncardiac etiologies were combined and compared to the cardiac etiology group, we found six clinically meaningful variables that were significant univariate predictors of a cardiac etiology of palpitations (P ~0.05). These variables were older age (continuous), male sex, description of an irregular heart beat, history of heart disease, duration of palpitation event >5 minutes, and fewer number of associated symptoms (continuous variable). When these variables were entered into a multivariate logistic regression model, male sex, description of an irregular heart beat, history of heart disease, and event duration of >5 minutes were found to be independent predictors 142 of a cardiac etiology (Table V). Although~ univa,ria~~ analysis revealed that presentation to the emergency department compared with the medical clinic ‘was associated with a cardiac etiology of palpitations (P <0.002), this was not a significant multivariate predictor of cardiac etiology (odds ratio 2.03, 95% confidence interval 0.76 to 5.4). None of the 17 patients with 0 predictors had a cardiac etiology, 13 (26%) of the 50 pa,tients with 1 predictor had a cardiac etiology, 28 (48%)) of t.he 58 patients with 2 predictors had a cardiac etiology, 22 (71%) of the 31. patients with 3 predictors had a cardiac etiology, and 9 (90%) of the 10 patients with all 4 predictors had a cardiac etiology. Twenty fcmr patients with missing data were excluded from this analysis. The area under the ROC curve for this model was 0.79 (standard deviation of the area = 0.04). Diagnostic Testing Table VI describes the evaluation of Ipatients in terms of the types of tests ordered, as well as the number of patients for whom a diagnostic test led to the etiology of palpitations. The basic patient evaluation consisted of a history and physical (completed in loo%?),psychiatric screening with the GHQ (completed in 76%), and arrhythmia detection (ECG completed in 87%, prolonged electrocardiographic monit80ring completed in 64%). The performance of history and physical examination, ECG, and prolonged electrocardio graphic monitoring was not significantly different (P >0.2) between patients who had a known etiology of palpitations and those who had unknown etiology. Of the 159 patients for whom a diagnosis could be made, 64 (40%) were accomplished with either the history and physical examination, an ECG, and/or laboratory data The laboratory data that was diagnostic in 8 patients included thyroid function studies and serum theophylline level and hematocrit determinations. Cardiac arrhythmia etiologies were detected by ECG in 43 patients and by prolonged electrocardiographic monitors in 25 patients; of the remlaining 122 patients, 110 (90%) underwent electrocardiography or prolonged electrocardiographic monitoring, however, none of these tests revealed the etiology of palpitations. There was no evidence for gender, age, or racial bias in the performance of ECG or monitoring. There was no evidence for bias in the performance of monitoring based on site of presentation. An ECG was more likely to be performed if the patient presented to the emergency department (94:0/o),rather than any other site (78%); conversely, an ECG was less likely to have been performed if the patient presented to the medical clinic (70?), rather than any other site (95%) (P ~0.01). Psychiatric screening instruments were completed by 144 (76%) patients. A GHQ score 25 and/or SOM score 23 was found in 73 individuals. The DIS was ad- February 1996 The American Journal of Medicine” Volume 100 PATIENTS WITH PALPITATiONS/VVEBER ministered to these 73 patients and to an additional 23 patients with clinical features suggestive of a possible psychiatric illness. Of the 96 patients assessed by the DIS, one or more psychiatric illness was diagnosed in 78 patients (for a total of 154 diagnoses). Additionally, 4 other patients were felt to have clinically significant psychiatric illness (3 with anxiety, 1 with depression and suicidal ideation). The diagnoses made by DIS were panic attack and/or disorder (n = 69), generalized anxiety disorder (n = 39), somatization disorder (n = l), and depression (n = 45). The majority of patients (52 of 78) had 2 or more coexisting psychiatric illnesses; most frequently, depression accompanied one of the other disorders (n = 41). There were significantly more DIS performed in nonwhite (65%) than white (45%) patients (P ~0.02). Ultimately, 58 of these 82 patients (55 with a positive DIS, 3 with a clinical diagnosis) were assigned to the psychiatric etiology group (Table Ill); 5% (29/55) of these patients were also depressed, however depression was considered to be comorbid and not etiologic. There were 19patients with a positive DIS in whom the psychiatric illness was considered to be comorbid and not etiologic. The assigned etiologies in these patients were: arrhythmias in 14, medications in 2, thyrotoxicosis in 1, and unknown in 2. These 2 patients were categorized as unknown etiology instead of psychiatric etiology because one had coexisting alcoholism and the second had a remote history of the psychiatric disorder. Follow-Up By October 28, 1992, follow-up was completed in 98%of the patients. At least l-year follow-up (365 days + 14days) was available for 96%of the patients. Table VII details the outcomes at final follow-up. All reports of stroke were confirmed by the medical record. Mortality was documented in 3 patients; none of the deaths was sudden. One death was due to a subarachnoid hemorrhage in a patient taking warfarin following aortic valve replacement for aortic insufficiency. The second death was due to congestive heart failure and renal failure in a patient following a stroke. The third death was due to severe congestive heart failure in a patient with cell&is and soft-tissue abscess. There were 2 patients with new arrhythmias documented in follow-up; it is probable that these were responsible for the original palpitations. One patient, originally in the unknown category, had symptomatic correlation of palpitations with premature ventricular beats upon repeat presentation. Another patient with increasing symptoms, originally in the psychiatric category, had documentation of symptomatic correlation of palpitations with supraventicular tachycardia. The percent with recurrent palpitations varied by etiology; however, this difference was only significant AND KAPOOR TABLE VII Outcomes at Final Follow-Up Mortalitv -. 1%). Stroke 1%) - Cardiac (n = 82) 1 (1.2) 1 (1.2) Psychiatric (n = 58) l(1.7) 0 Miscellaneous (n = 19) 1 (5.3) 1 (5.3) Unknown (n = 31) 0 0 Total’ (n = 190) 3 (1.6) 2 (1.1) ‘One-year mortakty and stroke rates were determined by the Kaplan-Meier method. The 95% confidence interval for the total mortality rate was 0% to 3.4% and for the total stroke rate was 0% to 2.6%. Mean and median follow-up times for both stroke and death exceeded 365 days. at the first follow-up when recurrent symptoms were experienced by 61% of the psychiatric group, 53%of the cardiac group, 17% of the miscellaneous group, and 48% of the group with an unknown etiology (P ~0.02). By the last follow-up, 75%of all patients reported recurrent symptoms. This rate of recurrent symptoms varied by history of prior palpitations; 79% of the patients with and 61% of the patients without a prior history of palpitations had recurrent events within the follow-up period (P ~0.03). At 1 year, 95% of the patients were satisfied with the care they received for their palpitations. Eightynine percent reported that their health was the same or improved compared to 1 year before. Of those working outside of the home (52%of the cohort), work performance was impaired in 19%,and workdays were missed by 1%. Because of palpitations, 33%reported accomplishing less than usual work at home. DISCUSSION Although palpitations are a common symptom, there are very limited descriptive, etiologic, or prognostic data for nonreferred patients on this subject. This is the first study that describes the etiologies of palpitations and the outcomes of patients in a cohort of patients presenting for care at a university-based medical center. We found that (1) an etiology could be dletermined in 84%of the cohort, (2) 40%of those etiologies could be determined with the history and physical examination, an ECG, and/or laboratory data, 113)psychiatric illnesses were common causes of palpitations, and (4) the prognosis was excellent except that most patients continued to have recurrent symptoms. Our finding that only 16%of patients had no clear etiology for their palpitations differs from research on other common symptoms in primary care. In a 3year incidence study analyzing the probable etiology of 14 common symptoms in 1,000 internal medicine outpatients, Kroenke and Mangelsdorf?!7 reported that the percent of each symptom with an unknown etiology ranged from 47% for insomnia to 100%for constipation. February 1996 The American Journal of Medicinea Volume 100 143 PATIENTS WITH PALPITATIONS/WEBER AND KAPOOR A wide variety of etiologies were diagnosed in this cardiac disease, mortality was documented in only 3 population of patients with palpitations. Previous re- women over the age of 70. None of the (deaths was ports of the etiology of palpitations are limited to stud- sudden or directly related to the original etiology of ies of patients referred for ambulatory electrocardiopalpitations. In the only other study of outcomes in graphic monitoring. 2a42 These studies report the patients with palpitations5 the proportion experiencarrhythmias recorded by ambulatory electrocardioing a cardiac endpoint (ie, myocardial infarction, vengraphic monitoring, however not all studies indicate tricular tachycardia, ventricular fibrillation, cardiac the relationship between symptoms reported and ar- arrest, or death) in 5 years was similar between cases rhythmias detected. The reported yield of ambulatory with palpitations (6.4%) and clinic-based controls electrocardiographic monitoring in the detection of (7.2%). In only 4 patients was ventricular t,achycardia symptomatic arrhythmias ranges from 13%to 69%..“541,@ responsible for the symptom of palpitations. The lHowever, referral bias and the inclusion of patients year mortality rate of 1.6% in patients with palpitawith symptoms other than palpitations in most of tions is in striking contrast to our experience with these studies raise concerns about the generalizabilsyncope, where there is a 28% mortality ;md 15% inity of these findings. The prevalence of the various eti- cidence of sudden death at 1 year in patients with carologies of palpitations may also be misrepresented by diac etiologies.47 case series of patients with specific diseases associIn contrast, the morbidity from palpitations was ated with the symptom of palpitations.44 substantial. Although 77% had prior palpitations, it Psychiatric illness was frequently diagnosed and was surprising that the majority of patients with and was associated with the highest rate of recurrent without a history of prior palpitations had recurrent symptoms at 3 months. Palpitations are one of the symptoms. This is in contrast to the 35% of syncope DSM-III-R diagnostic criteria for panic attack, gener- patients with recurrent symptoms at 5 years.“7 The realized anxiety disorder, and somatization disorder.20 currences appeared to have substantial effect on qualAlthough depression may be a comorbid illness, we ity of life since at least one third of the patients redid not consider this to be an etiology of palpitations. ported accomplishing less than usual work in the Previously published data from our own general med- home and a smaller fraction had impaired work perical clini~,~~@ documented a 9.2% prevalence of major formance or missed work. These data suggest that depressive disorder and 1.7% prevalence of panic dis- this common symptom has characteristics similar to order in primary care populations. Depressed patients a chronic disease with exacerbation and remission of had more physical illness, somatic symptoms, and dis- symptoms over time. ability than nondepressed patients. It is likely that deLimitations of this study should be acknowledged. pression may result in palpitations, but 41 of 45 of our First, we did not assemble an inception co‘hort by limpatients with depression had other coexisting psychiiting our patient enrollment to those with new onset atric illness. Thus, we were unable to clarify the role of palpitations because patients often could not define of depression in leading to palpitations. As suggested the first onset of this symptom. Despite this limitation, our findings are relevant to the patients seen with palby Bar~ky,~ further studies of the relationship between pitations since the vast majority of patients present common symptoms such as palpitations and psychiatric illness are needed, since there is evidence for sig- with chronic symptomatology. Second, not all patients nificant unrecognized psychiatric morbidity in ambu- underwent all diagnostic tests. Specifically, the DIS was performed less frequently in patients for whom an latory care patients with common medical symptoms. etiology could not be determined. We believe this is Based on our findings, assessment for generalized anxiety, panic, and somatization disorders and de- not a major limitation since, surprisingly, an etiology pression should become an important focus of evalu- could be determined in the majority of the patients. Based on this study, we suggest the following stratation of patients with palpitations. egy for a practical evaluation of patients with palpiWe identified four variables that were independent predictors of a cardiac etiology of palpitations. This tations. A careful history, physical examination, and ECG along with selective use of laboratory tests will model performed better than chance, as demonstrated by the area under the ROC curve. However, the pre- identify the etiology of palpitations in close to half of the patients. In the remaining patients, a major focus diction of a cardiac etiology must be interpreted with should be screening (with the GHQ) for psychiatric caution because (1) assignment to the cardiac etiology group does not necessarily imply a higher mor- disorders to detect generalized anxiety, panic, and detality, (2) further testing beyond the initial database pression. In those without a diagnosis who have heart may not be required to make this diagnosis, and (3) disease, palpitations lasting longer than 5 minutes, or irregular beats, prolonged cardiac monitoring may diverse etiologies were included in the cardiac group. show an etiology of palpitations. Since symptomatic Palpitations were associated with low mortality and cardiac morbidity. Despite the high prevalence of correlation is critical in determining whether an ar144 February 1996 The American Journal of Medicine@ Volume 100 PATIENTS WITH PALPITATlONS/WEBER AND KAPOOR rhythmia is the etiology of palpitations, event monitoring appears to be well suited for diagnostic evaluation of this symptom.31 Electrophysiologic testing should be reserved for specific high-risk groups such as patients with accessory pathways or when therapy of documented arrhythmias is needed.@ In conclusion, the etiology of palpitations can often be diagnosed with a simple initial evaluation. Psychiatric illness accounts for the etiology in nearly one third of all patients with palpitations. The shortterm prognosis of patients with palpitations is excellent; however, the recurrence rate is high. Future studies are needed to develop interventions that may decrease recurrent symptoms in these patients and improve their quality of life. ACKNOWLEDGMENT We are Indebtedto Nancy Brant Miller, CRNP,for the administration of the DIS, Karen Brich, PA-C,for completing patient follow-up, Barbara Hanusa, PhD, for statistical assrstance, Terry Sefcik, MSc, for data managementguidance, Lisa Joseph for data entry, and Roberta Eckmanfor manuscriptpreparation REFERENCES 1. Kroenke K. Arrington ME, Mangelsdorff AD. The prevalence of symptoms in medical outpatients and the adequacy of therapy. Arch intern Med. 1990;150:168~1689. 2. Braunwald E. The history. In: Braunwald E, ed. Heart Disease: a textbook of cardiovascular medrcine.4th ed. Philadelphia:WB Saunders; 1992:1-12. 3. Barsky AL Palpitabons,cardiac awareness, and panic disorder. Am J Med. 1992;92:31s-34s. 4. Jones GE, Dinoff BL, Jones KR, Leonberger FT. Survey of cardiac awareness in rehabilitatedcardiac patients. Psychophysiology.1983;20:450-451. 5. Knudson MP. The natural history of palpitations in a family practice. J Fam Pratt. 1987;24:357-360. 6. Degowrn CL, Degowin RL. Bedside diagnostic exam. 4th ed. New York: MacmillanPublrshrngCo., 1981. 7. Goldberg DP. The detecbon of psychiatric illness by questionnarre: a technique for the identification and assessment of non-psychotic psychiatric illness. London: Oxford University Press, 1972. 8. Berwick DM, BudmanS, DamicoWhiteJ, et al. Assessment of psychological morbidity in primary care: explorations with the General Health Queshonnaire.J Chron Dis. 1987;4Ofsuppl1):71s-79s. 9. Huppert FA, Walters DE, Day NE, Elliott BJ. The factor structure of the General Health Quesbonnaire(GHQ30). A reliability study on 6317 communrty residents. Br J Psychiatry. 1989;155:178-185. 10. Vieweg BW, Hellund JL. The General Health Questionnatre (GHQ): A comprehensive revrew. 1 OperationalPsychiatry. 1983;14:74-81. 11. vonAmmonCavanaughS, WeftsternRM. Emotionaland cognitkredysfunction assocratedwith medicaldisorders. J PsychosomRes. 1989;33:505-514. 12. Finlay-JonesRA, Murphy E. Severity of psychiatric disorder and the 30iitem General Health Quesbonnaire.Br J Psychiatry. 1979;134:609-616. 13. Othmer E, DeSouza C. A screening test for somatizatron disorder. Am J Psych. 1985;142:1146-1149. 14. Newman SC, Bland RC, Om H. A comparison of the methods of scoring the General Health Questionnaire.Compr Psychiatry. 1988;29:402-408. 15. Goldberg DP, Rrckels K, Downing R, Hesbacher P. A comparison of two psychiatric screening tests. Br J Psychiatry. 1976;129:61-67. 16. Smith GR, Brown F. Screening indexes in DSM-IIIRsomatrzation disorder. GenHasp Psychiatry. 1990;12:148-152. 17. Helzer JE, Robins LN. The dragnostic interview schedule: its development, evolution, and use. Sot PsychratrEpidemiol. 1988;23:6-16. 18. Robins LN, Helzer JE, CroughanJ, Ratcllff KS. National lnstttute of Mental Health Diagnostic Interview Schedule: its history, characterishcs, and validity. Arch Gen Psychratry. 1981;38:381-389. 19. Paulsen AS, Crowe RR, Noyes R, Pfohl B. Rehabrlity of the telephone intervrew in diagnosing anxiety disorders. Arch ‘C;en Psychiatry. 1988;45:62-63. 20. American Psychiatric Association, Committee on nomenclature and statistics. Diagnosbc and Statistical Manual of Mental Disorders, 3rd ed, revised. Washington,DC: AmerrcanPsychiatric Association, 1987. 21. Naranjo CA, Busto U, Sellers EM, et al. A method for estimating the probability of adverse drug reacbons. Bn PharmacolJher. 1’381;30:239-245. 22. Rbase data managementsoftware. Microrim Inc., 1987. 23. BMDPstatistical software manual. Volumes 1 & 2. Berkeley: University of California Press, 1988. 24. Statxact, Version 2. Cambndge,MA: CMEL Corp., 1991. 25. Metz CE, Shen J, Wang P, Kronman HB. ROCFIT,Chicago: University of Chicago, 1989. 26. Kaplan EL, Meier P. Nonparametric estimation form incomplete observations. JAm Stat Assoc. 1958;53:457-481. 27. Kroenke K, Mangelsdorff AD. Common symptoms in ambulatory care: incidence, evaluation,therapy, and outcome. Am J Med. 1989;86:262-266. 28. Gooch AS, Rahim A, McKerthen R. Exercise testing alnd portable ECG recording in arrhythmias-pronepatients. Angiofogy. 1976;27:133-137. 29. RingqvrstI, Jonason T, Nilsson G, Khan AR. Dragnosbcvalue of longterm ambulatory ECGin patients with syncope, dizziness or palpitatrons.Ch Physiol. 1989;9:47-55. 30. Tabatznik 8. Halter recording stakes out three clinical areas. C/in Trend Cardiol. 1976;6:6-7. 31. Brown AP, Dawkrns KD, Davies JG. Detecbon of arrhythmias: use of a patient-activated ambulatory electrocardrogram devrce with a solid state memory loop. Br Heart J. 1987;58:251-253. 32. Camm AJ, Ward DE, Spurrell RAJ. Arrhythmias in ambulatory persons. Biotelemetry and PabentMonitoring. 1978;5:167-181. 33. Judson P, Holmes DR, Baker WP. Evaluation of outpatIent arrhythmias using transtelephonic monitoring. Am Heart J. 1979;97(6):759-761. 34. LrpskrJ, Cohen L, EsprnosaJ, et al. Value of bolter monitoring in assessing cardiac arrhythmias in symptomatic patients. Am J Cardiol. 1976;37:102-107, 35. DiamondTH, Smith R, Myburgh DP. Halter monitoring-a necessity for the evaluationof palpitations, S Afr MedJ. 1983;63:5-7. 36. Goldberg AD, Raftery EB, CashmanPMM.Ambulatory elec:rocardrographic records in patients with transrent cerebral attacks or palpitabon. BMJ. 1975;4:569-571. 37. Grodman RS, Capone RJ. Most AS. Arrhythmia surverllance by transtelephonic monitoring: comparison w&h bolter monitoring n symptomatic ambulatory pabents. Am Heart J. 1979;98:459-464. 38. Hasin Y, David D, Rogel S. Dragnosbc and therapeutic assessment by telephone electrocardiographrc monrtoring of ambulatory patients. BMJ. 1976;2:609-612. 39. Kunz G, Raeder E, Bruckhardt D. What does the symptom “palpitahon” mean?-Correlation between symptom and the presence of cardiac arrhythmias in the ambulatory ECG.ZKardiol. 1977;66:138-141. 40. Thomas LE, Shapiro LM, Perkins EJ, Fox KM. Detection of arrhythmia: limited usefulness of patient actrvated recording devices. BMJ. 1984;289: 1106-1107. 41. Zeldis SM, Levine BJ, Mrchelson EL, Morganroth J. Cardiovascular complaints: correlation wrth cardiac arrhythmias on 24.hour electrocardiographic monrtonng.Chest. 1980;78(3):456461. 42. Barsky AJ, Cleary PD, Coeytaux RR, Ruskin JN. Psychratric disorders in medical outpatients complaining of palpitations. J Gen Intern Med. 1994;9: 306-313. 43. DrMarco JP, Philbrick JT. Use of ambulatory electrocardiographic (Halter) monitoring. Ann Int Med. 1990;113:53-68. 44. Bravo EL, Gifford RW. Pheochromocytoma: Diagnosis, localization and management.NEJM.1984;311f20):1298-1303. 45. SchulbergHC, Saul M, McClellandM, et al. Assessingdepres,sionin primary medical and psychiatric pracbces. Arch GenPsychiatry. 1985;42 1164-l 170. 46. CoulehanJL, Schulberg HC, Block MR, et al. Medical comorbrdity of major depressive disorder In a primary medical practice. Arch Intern Med. 1990;150(111:2363-2367. February 1996 The American Journal of Mediiine” Volume 100 145 47. Kapoor WN. Evaluation and outcome of patients with syncope. Medicine. 1990;69:160-175. 48. Reid P. Indications for intracardiac electrophysiologic studies in patients with unexplained palpitations. Circulation. 1987;75(4 pt 2):lll 154-160. APPENDIX Diagnostic Criteria Used for the Categorization of the Etiologies of Palpitations Arrhythmias. Any new deviation from normal sinus rhythm or a significant change in the rate of a stable arrhythmia (eg, atrial fibrillation) can cause the symptom of palpitations. Definite: Symptomatic correlation of palpitations with the arrhythmia recorded on electrocardiographic monitoring. &oba&: Greater than 5 beats of supraventricular tachycardia or ventricular tachycardia in the absence of symptomatic correlation. Increased Stroke Volume. Stroke volume may be increased with cardiac and extracardiac shunts13 or regurgitant valvular heart disease.4 Palpitations may be noted at different stages, so general guidelines were used to determine causality. 0 Cardiac and extra-cardiac shunts. A cardiac shunt was considered causal if there was either echocardiographic or cardiac catheterization documentation of moderate to severe shunt flow in a patient with palpitations. Any extra-cardiac arteriovenous fistula in a patient with palpitations was considered causal if no other etiology was present. 0 Regurgitant valvular heart disease. Valvular heart disease such as mitral regurgitation and aortic insufficiency was considered causal if there was either echocardiographic or cardiac catheterization documentation of moderate or severe regurgitant flow in a patient with constant palpitations and no concurrent arrhythmia. Pacemaker. Paced beats or intercostal muscle and diaphragmatic flutter may be sensed by the patient with a pacemaker. Pacemaker syndrome may also be associated with palpitations5 To be considered causal, the sensed beats must have been correlated with the paced beats. Prosthetic Heart Valve. Each heart beat in a patient with a prosthetic heart valve may be sensed and reported as constant palpitations.4 To be considered causal, the sensed beats must have been correlated with the normal heart rhythm. Cardiac Disease. Various cardiac diseases have been associated with palpitations in the absence of arrhythmias or other causes for palpitations. These reports were interpreted to define causality. 0 Cardiomegaly. The enlarged cardiac silhouette can cause palpitations, probably on the basis of increased cardiac output or contractility. This was considered causal if there was chest radiographic evi146 February dence of at least moderate cardiomegaly in a patient with palpitations. l Mitral Valve Prolapse (MVP). Although the literature reports series of patients with MVP in whom there is symptomatic correlation of arrhythmias on ambulatory monitoring, there are also patients with symptoms in the absence of arrhythmias.fi-8 For the patient with clinical (ie, classic murmur or click) or echocardiographic evidence of mitral valve prolapse and palpitations, MVP was considered causal in the absence of symptomatic arrhythmias. Patients in whom symptomatic arrhythmias were documented were classified as an arrhythmia etiology. l Hyperkinetic heart syndrome. Classification into this category was limited to young males -with a systolic murmur and a hyperdynamic precordium and pulse, in the absence of any other etiology for a hyperadrenergic state.gJO l Atrial myxoma Although this entity is rare, there have been reports of atrial myxoma presenting with palpitations. l1 For the patient with echocardiographic evidence of atrial myxoma and palpitations, the myxoma was considered causal in the absence of symptomatic arrhythmias. Psychiatric Disease. Palpitations commonly occur in patients with panic attack and disorder, generalized anxiety disorder, and somatization disorder and are included in their Diagnostic and Statistical Manual of Mental Disorders, Third edition, Revised (DSM-IILR) deftitions.12 Definite: A diagnosis of one or more of’ these disorders was made if the patient met DSM-III-R criteria as diagnosed with Diagnostic Interview Schedule @IS).‘3J4 and there was no other signiticant medical comorbidity or etiology for palpitations. Probable: When the DIS was not administered but there was strong clinical evidence, the investigators considered a category of probable anxiety to fit the DSM-III-R category of generalized anxiety disorder that was not otherwise specified. Cmorbid: In cases in which any other etiology for palpitations existed, the psychiatric disorder was not considered as the etiology. Medications. Palpitations occurring with a temporal relationship to the use of medications such as sympathomimetic agents, vasodilatom, anticholinergics, or during withdrawal from S-blockers are well recognized.15 Each potential etiologic medication must have been reported to cause palpitations as a possible adverse effect. Criteria were devised to justify causality, similar to previous reports on adverse drug reactions. l6 Definite: (1) Palpitations following a temporal sequence after the medication was introduced or reached an abnormal level, (2) resolving after withdrawal of the medication or normalization of the 1996 The American Journal of Medicine@ Volume 100 PATIENTS WITH PALPITATIONS/WEBER AND KAPOOR blood concentration, and (3) having no other apparent etiologic factor. Probable: Two of the three factors listed above. Possible: One of the three factors listed above OR two of the three factors listed above in the presence of another etiologic factor. Habits. Palpitations occurring with a temporal relationship to the use of cocaine17-1g and amphetamines,15 caffeine,2@23 and nicotine15 have been reported. Criteria have been developed to define this association. 0 Cocaine or amphetamines Definite: (1) Palpitations temporally related to cocaine or amphetamine use, (2) with resolution after discontinuation of the drug, and (3) no other apparent etiology. Probable: Palpitations temporally related to cocaine or amphetamine use and one of the other two factors listed above. Possible: Palpitations temporally related to cocaine or amphetamine use, but other possible etiologies exist. 0 Caffeine Definite: (1) Palpitations temporally related to caffeine intake of greater than 4 cups of coffee (or equivalent) per day, (2) with resolution after discontinuation of caffeine, and (3) no other apparent etiology. Probable: Palpitations temporally related to caffeine intake of greater than 2 cups of coffee (or equivalent) per day and one of the other two factors listed above. Possible: Palpitations temporally related to any caffeine intake, but other possible etiologies exist. 0 Nicotine Definite: (1) Palpitations temporally related to nicotine product use, (2) with resolution after cessation, and (3) no other apparent etiology. Probable: Palpitations temporally related to nicotine product use and one of the other two factors listed above. Possible: Palpitations temporalIy related to nicotine product use, but other possible etiologies exist. Metabolic Disorders. Several metabolic disorders have been reported to be associated with palpitations and are often accompanied by a sinus tachycardia. These abnormal metabolic states are often recognized by other characteristic associated signs and symptoms. 0 Hypoglycemia. 24Hypoglycemia was documented in association with palpitations that resolved with restoration of normal glucose levels. 0 Thyrotoxicosisz5 Palpitations were associated with the clinical presence of thyrotoxicosis, confirmed by laboratory tests. 0 Pheochromocytoma. 26 Palpitations were associated with the clinical presence of pheochromocytoma, confiied by laboratory tests. l Mastocytosis. 27Palpitations were associated with the clinical presence of mastocytosis, confirmed by laboratory tests. l Scombroid food poison&g.2x Palpitations following ingestion of scombroid fish. 0 Idiopathic flushing. 2g The clinical scenario was consistent with idiopaQic flushing and other disorders were excluded (ie, diagnosis by exclusion). High Output States. Increased cardiac output can be responsible for the symptom of palpitations. The conditions listed below are known causes of a high output state. The following definitions were used to relate the palpitations to the high output state: 0 Anemia1n30 There is no clear level of hemoglobin at which the cardiac output rises. Symptoms vary with the rate of onset of anemia and the underlying comorbidity of the patient. For the purpose of this study, anemia was considered causal if palpitations occurred in a patient with a hemoglobin < 100 g/L and resolved after correction of the anemia. 0 Pregnancy.31 Peak cardiac output occurs between the 20th and 24th week of gestation. Palpitations in a pregnant woman after the 20th we(3k of gestation were attributed to the high output state of pregnancy when no other etiology was present. 0 Paget’s disease.1B32Cardiac output rises .when more than 15% of the skeleton is involved with active Paget’s disease. For the purpose of this study, Paget’s disease was considered causal in a patient with palpitations if there was active disease of at least two skeletal locations and no other etiology was present. 0 Fever.33 There is a linear relationship between rise in temperature above normal and the heart rate. As baseline heart rate varies among individuals, so will the appearance of tachycardia Fever was considered causal when palpitations were temporally related to a new temperature of 238” C, with resolution after defervescence, if no other etiology was present. Dehydration and Orthostatic Hypotension. Palpitations occurring in the patient with dehydration or orthostatic hypotension are generally believed to be related to a physiologic (compensatory) sinus tachycardia. These were considered present in the appropriate clinical setting if clinical or laboratory signs of dehydration or symptomatic orthostatic hypotension existed.% Exertion. Exertional palpitations are related to an increased stroke volume. Exertion was considered causal if historical data confirmed physical exertion for that $iividual, in the absence of another etiology of palpitations. REFERENCES 1. BraunwaldE, GrossmanW. Clinical aspects of heart failure. III: BraunwaldE, ed. Heart Disease: A Textbook of CardiovascularMedicine.4th ed. Philadelphia: WB Saunders;1992:45%462. February 1996 The American Journal of Medicine@ VOlUml? 100 147 2. Dedichen H, Thalow E, Naess A. Traumatic arteriovenous fistula. Case Report. Acta Chir Stand. 1989;155:297-300. 3. Holman E. Abnormal arteriovenous communrcations. Great variability of effects with particular reference to delayed development of cardiac failure. Circulation. 1966;32:1001-1008. 4. Braunwald E. Valvular heart disease. In: Braunwald E, ed. Heart Disease: A Textbook of Cardiovascular Medicine. 4th ed. Philadelphia: WB Saunders; 1992:1007-1065. 5. Ausubel K, Furman S. The pacemaker syndrome. Ann fnt Med. 1985;103:420-429. 6. Kramer HM, Kligfield P, Devereux RB, et al. Arrhythmias in mitral valve prolapse: effect of selection bias. Arch Intern Med. 1984;144:2360-2364. 7. Devereux RB, Kramer-Fox R, Kligfield P. Mitral valve prolapse: causes, clinical manifestations,and management.Ann Int Med. 1989;111:305-317. 8. DeMaria AN, Amsterdam EA, Vismara LA, et al. Arrhythmias in the mitral valve prolapse syndrome. Ann Int Med. 1976;84:656-660. 9. Gillum RF, Teichjolz, Herman MV, Gorlin R. The idiopathic hyperkinetic heart syndrome: clinical course and long-term prognosis. Am Heart J. 1981;102: 728-734. 10. Gorlin R. The hyperkinetic heart syndrome. JAMA.1962;182:823-829. 11. Awang Y, SallehuddinA. Surgical experience with cardiac tumours at the General Hospital. Med J Malaysia. 1991;46:28-34. 12. American Psychiatric Association, Committee on nomenclature and statistics: Diagnostic and Statistical Manual of Mental Disorders. 3rd ed, revised. Washington,DC: American Psychiatric Association; 1987. 13. Helzer JE, Robins LN. The diagnostic interview schedule: its development, evolution, and use. Sot Psychiatr Epidemiol. 1988;23:6-16. 14. Robins LN, Helzer JE, Croughan J, Ratcliff KS. National Institute of Mental Health Diagnostic Inter-viewSchedule: its history, characteristics, and validity. Arch Gen Psychiatry. 1981;38:381-389. 15. Physician’s Desk Reference. 46th ed. Montvale, NJ: Medical Economics Data; 1992. 16. Naranjo CA, Busto U, Sellers EM, et al. A method for estimating the probability of adverse drug reactions. Clin Pharmacol Ther. 1981;30:239-245. 17. Petronis KR, Anthony JC. An epidemiologic investigation of marijuana and cocaine-relatedpalpitations. Drug Alcohol Depend. 1989;23:219-226. 18. Cregler LL, Mark H. Medical complications of cocaine abuse. NEJM. 1986:315:1495-1500. 148 February 19. Rich JA. Singer DE. Cocaine-relatedsymptoms in patients presenting to an urban emergency department. Ann Emerg Med. 1991;20:616-621. 20. Greden KF. Anxiety or caffernism: a diagnostic dilemma. .AmJ Psychiatry. 1974;131:1089-1092. 21. Myers MG. Caffeine and cardiac arrhythmias. Ann ht Med. 1991;114: 147-150. 22. Prineas RJ, Jacobs DR, Crow RS, Blackburn H. Coffee, tea and VPB. J Chronic Dis. 1980;33:67-72. 23. Shirlow MJ, Mathers CD. A study of caffeine consumption and symptoms: indigestion, palpitations, tremor, headache and insomnia. Int J Epidemiol. 1985;14:239-248. 24. Cryer PE, Gerich JE. Glucose counterregulation, hypoglycemia, and intensive insulintherapy in diabetes mellitus. NEJM. 1985;313:232-241. 25. Klein I. Thyroid hormone and the cardiovascular system. Am J Med. 1990;88:631-637. 26. Bravo EL, Gifford RW. Pheochromocytoma: diagnosis, localization and management.NE&f. 1984;311:1298-1303. 27. Tharp MD. The spectrum of mastocytosis. Am J Med Sci. 1985;289: 117-132. 28. Merson MH, Baine WB, Gangarosa EJ, et al. Scombroicl fish poisoning: outbreak traced to commercially canned tuna fish. JAMA. 1974;228: 1268-1269. 29. Aldrich LB, Moattari AR, Vinik Al. Distinguishing features of idiopathic flushing and carcinoid syndrome. Arch Intern Med. 1988;148:2614-2618. 30. RosenthalDS, Braunwald E. Hematological-oncologicaldisorders and heart disease. In: Braunwald E, ed. Heart Disease: A Textbook of Cardiovascular Medicine.4th ed. Philadelphia:WB Saunders; 1992:1742-1744. 31. Elkayam U. Pregnancy and cardiovascular disease. In: Braunwald E, ed. Heart Disease: A Textbookof Cardiovascular Medicine. 4th red. Philadelphia: WB Saunders; 1992:1790-l 793. 32. Arnalich F, Plaza I, Sobrino JA, et al. Cardiac size and function in Paget’s disease of bone. Int J Cardiol. 1984;5:491-505. 33. Dinardello CA, Wolff SM. Fever. In: Mandell GL, Douglas RG, Bennett JE, ed. Principles and Practices of Infectious Diseases. 3rd ed. New York: Churchill Liiingstone; 1990:464-467. 34. Thomas JE, Schirger A, Fealey RD, Sheps SG. Orthostatic hypotension. Mayo C/inProc. 1981;56:117-125. 1996 The American Journal of Medicine@ Volume 100