Twenty-four-hour esophageal pH monitoring: The most useful test for evaluating noncardiac chest pain

Twenty-Four-Hour Esophageal pH Monitoring: The Most Useful Test for Evaluating Noncardiac Chest Pain EDWARDG. HEWSON,M.D., JANEW. SINCLAIR,P.A.-C., CHRISTINEB. DALTON,P.A.-C., Winston-Salem, North Carolina, JOELE. RICHTER,M.D., Birmingham, Alabama

STUDYOBJECTIV~ To compare the diagnostic capabilities of traditional esophageal tests (manometry and provocative testing with acid and edrophonium) and 24-hour esophageal pH monitoring in identifying an esophageal cause of chest pain. DESIGN: A prospective study of 100 consecutive patients r e f e r r e d by cardiologists to the esophageal laboratory for evaluation of esophageal causes of chest pain. SE'I~ING: Tertlnry-referral university hospital. METHODS:Esophageal mAnometry performed with 10 wet swallows of water. Acid perfusion (0.1 N hydrochloric acid) and edrophonium (80 jzg/kg intravenously) tests were placebo-controlled with a positive study defined as replication of typical chest pain. Esophageal pH monitoring identified (1) abnormn! acid exposure times in the upright, supine, or combined position, and (2) correlation between symptoms and acid reflux, i.e., symptom index. The esophagus was identified as "probably" contributing to chest pain only if the acid or edrophonium test was positive or if there was a positive correlation between symptoms and acid reflux during pH monitoring. RESULTS:Esophageal mAnometry was abnormal in 32 patients (32%), but patients were asymptometic during the study. The acid perfusion test was positive in 18 of 95 patients (19%), and the edrophonium test was positive in 15 of 78 patients (19%). AbnormAl acid exposure times were found in 48 patients (48 %). Of the 83 patients with spontaneous chest pain during 24hour pH testing, 37 patients (46%) had abnormal reflux parameters and 50 patients (60%) had a

From the Gastroenterology Division (EGH, JWS, CBD), Bowman Gray School of Medicine, Winston-Salem, North Carolina, and the Division of Gastroenterology (JER), University of Alabama at Birmingham, Birmingham, Alabama. Requests for reprints should be addressed to Joel E. Richter, M.D., Division of Gastroenterology, University of Alabama at Birmingham, UAB Station, Birmingham, Alabama 35294. Manuscript submitted August 23, 1990, and accepted in revised form February 5, 1991.

576

May 1991 The American Journal of Medicine Volume90

positive symptom index (mean positive score 56%, range 6% to 100%). CONCLUSIONS: Acid reflux is a common and potentially treatable cause of noncardiac chest pain. TraditionAl esophageal tests us, Ally miAs this diagnosis. Twenty-four-hour esophageal pH monitoring with symptom correlation is the single best test for evaluating patients with noncardiac chest pain. with recurrent chest pain who are found p toatients be free of significant coronary artery disease account for 10% to 30% of patients undergoing coronary angiography [1,2].Their symptoms are associated with a low mortality [3,4],but m a n y patients remain physically and emotionally disabled [4]. Most patients stillbelieve that they m a y have heart disease, possibly because they sense their physician'suncertainty and inabilityto identify an alternative cause for their chest pain. The identification of a noncardiac source of pain, therefore, m a y be of considerable value to these patients in allaying their fears. Esophageal disorders are the most c o m m o n cause of noncardiac chest pain, accounting for identifiable abnormalities in 18% to 58% of these individuals [5-8]. Esophageal motility disorders and gastroesophageal reflux (GER) disease are the most frequent findings. However, few studies have assessed the prevalence of both disorders using stateof-the-art esophageal testing. We have studied 100 consecutive patients with angina-like chest pain referred by cardiologists to our esophageal laboratory after extensive evaluations did not identify a cardiac source for their chest pain. Patients underwent traditional esophageal testing (manometry and provocative testing with acid and edrophonium) for noncardiac chest pain [9], as well as outpatient ambulatory esophageal pH monitoring. The latter test is the best technique available for identifying and quantifying the presence of abnormal amounts of acid reflux [10] and can be done in the home or work setting, thereby replicating the environment associated with these patients' chest pains. Diary cards and event mark-

ESOPHAGEAL pH MONITORING / HEWSON ET AL

ers allow accurate recording and correlation of multiple chest pain episodes with the presence or absence of acid reflux for periods of up to 24 hours. The size of our study population permitted us to directly compare the diagnostic capabilities of traditional esophageal tests and pH monitoring for identifying the esophagus as a probable cause of chest pain. Additionally, this large study may suggest common mechanisms for esophageal test abnormalities and a more streamlined diagnostic evaluation for this common problem.

PATIENTS AND METHODS Patients Over an 18-month period, 100 consecutive patients with noncardiac chest pain were evaluated in our esophageal lal~oratory at the Bowman Gray School of Medicine with traditional esophageal tests (manometry, acid perfusion test, edrophonium test) and ambulatory 24-hour esophageal pH monitoring. This group was comprised of 43 men and 57 women, with a mean age of 50 years (range: 22 to 69 years). All patients had been directly referred to the laboratory by their university cardiologists after extensive work-ups failed to identify a cardiac source for their chest pain. All patients had electrocardiograms during their chest pain episodes, which did not show evidence of ischemic ST segment responses or arrhythmias. Sixty-six patients had undergone cardiac catheterization: norreal coronary arteries were present in 48 patients and nonobstructive disease (less than 50% narrowing of coronary artery diameter) was found in 18 patients. Thirty-four patients did not have cardiac catheterization. Of this latter group, 16 patients had negative stress thallium studies, while the remaining 18 patients were not thought to have cardiac disease primarily based on normal electrocardiograms with pain, negative exercise stress tests, and normal echocardiograms. No patient had mitral valve prolapse. In this group, studies were not explicitly done to exclude microvascular angina since this was considered unlikely by the referring cardiologists based on the absence of ischemic changes during pain and/or normal left ventricular response to exercise during thallium testing [11,12]. Over the time of this study, 583 patients without evidence of prior cardiac disease had undergone cardiac catheterization for the evaluation of anginalike chest pain. From this group, 102 patients (17.4%) were found to have normal coronary,arteries or nonobstructive disease. This was the group from which the 66 patients were referred to our esophageal laboratory. These patients were referred by nine different staff cardiologists, although

76% came from four cardiologists who most commonly in the past had found our esophageal laboratory evaluation helpful. Patients were sent directly to the laboratory without prior consultation with a gastrQenterologist. Whether these patients were screened by the cardiologists for esophageal sympt o m s p r i o r to r e f e r r a l is u n k n o w n to t h e investigators. Detailed information a b o u t gastrointestinal symptoms was obtained from all patients. Patients were specifically questioned about symptoms or habits suggesting esophageal disease, especially GER disease: heartburn (greater than two episodes/ mowth), effortless regurgitation of acidic material, dysphagia for solids or liquids, odynophagia, or a history of antacid/H2 blocker use for peptic symptoms. Seventy-four patients had had prior esophageal evaluations for suspected GER disease: normal endoscopy was reported for 38 patients, unremarkable barium esophagram was documented in 33 patients [13], and three patients had had previous endoscopic evidence of erosive esophagitis. Twenty-six patients had no esophageal evaluation prior to referral to our laboratory for their chest pain. No patient had previous esophageal or ulcer surgery.

Esophageal Tests ESOPHAGEAL MANOMETRY: Studies were done with an 8-lumen polyvinyl catheter (external diameter 4.5 mm; intraluminal diameter 0.8 mm; Arndorfer Specialities, Inc., Greendale, Wisconsin). The distal four openings were 1 cm apart at 90 ° angles, and the four proximal openings were spaced at 5-cm intervals. Each lumen was continuously perfused with distilled water at a rate of 0.5 mL/ minute from a low-compliance, pneumohydraulic capillary infusion system (Arndorfer Specialities Inc.). The catheter was connected to external transducers (model 4-327-C, Beckman Instruments, Inc., Norcross, Georgia) with output to a Beckman recorder (model R-612). Patients were studied in the supine position after an overnight fast. Medications known to affect esophageal contractions or acid secretion or to modulate pain were withheld for at least 24 hours and usually 48 hours before the study. After the catheter was introduced through the nose into the stomach, lower esophageal sphincter (LES) pressure w a s measured by the station pull-through technique. The four proximal catheter openings were next positioned in the esophageal body at 3, 8, 13, and 18 cm above the LES. Esophageal contractions were recorded from these openings in response to wet swallows (5 mL of water) given at 30-second intervals. The basal study consisted of 10 consecutive

May 1991 The American Journal of Medicine Volume 90

577

ESOPHAGEALpH MONITORING / HEWSON ET AL

wet swallows over a 5-minute period. Tracings were analyzed for mean distal amplitude and duration at the recording sites 3 and 8 cm above the LES. Contraction sequences were evaluated for the presence of peristalsis as well as simultaneous, nonconducted, retrograde, or triple-peaked contractions. Baseline esophageal motility disorders were classified as nutcracker esophagus, diffuse esophageal spasm, achalasia, hypertensive LES, or nonspecific esophageal motor disorder as defined in our laboratory from prior studies in 95 healthy control subjects [9,14]. PROVOCATIVE TESTING: Immediately after manometry while still in the supine position, patients were administered acid and edrophonium cMoride (Tensilon; Roche Laboratories, Nutley, New Jersey) in an attempt to provoke their typical chest pain. The acid perfusion test was done by first infusing normal saline into the distal esophagus at a rate of 7 mL/minute for 2 minutes [15]. Without the patient's knowledge, the solution was then switched to 0.1 N hydrochloric acid and infused at a similar rate for 10 minutes. Lack of symptom reproduction after acid infusion constituted a negative test. If chest pain was reproduced after acid infusion, normal saline was again infused for 10 minutes or until symptoms eased, at which time the acid infusion was , repeated. A positive test occurred when the patient's typical chest pain was twice reproduced during acid infusion. Although heartburn symptoms were recorded, heartburn was not considered a positive endpoint for the purpose of this study. The edrophonium test was also placebo-controlled. A placebo (1 mL of 0.9% normal saline) and edrophonium chloride (80 ~g/kg) were consecutively administered intravenously by rapid bolus infusion in an order unknown to the patients. Immediately after each injection, 10 wet (5 mL of water) swallows were given and the patients were asked about the similarity of symptoms to their typical chest pain. A positive edrophonium test was defined as the replication of the patient's chest pain only after the edrophonium injection. Motility changes were not required for a positive edrophonium test [16]. Twenty-two patients did not receive edrophonium because they reported histories of bronchospasm (asthma, chronic obstructive lung disease) or cardiac arrhythmias. AMBULATORY 24-HOUR ESOPHAGEAL pH MONITORING: This study was performed on an outpatient

basis, usually i m m e d i a t e l y after the previous esophageal tests. One of two systems was employed: the Sandhill pH system (Sandhill, Littleton, Colorado), which uses an antimony electrode with a 3.0ram outer diameter, or the Synectics pH system 578

May 1991 The American Journal of Medicine Volume90

(Synectics Medical Inc., Irving, Texas), which also uses an antimony electrode with a 2.l-ram outer diameter. Both systems utilize reference electrodes attached to the anterior chest. Before and after each procedure, the pH electrode was calibrated using buffers of pH 2 and 7. The pH electrode was passed nasally and positioned 5 cm above the proximal border of the manometricaUy identified LES. Patients were given a standard diet sheet and instructed to avoid food or drink with a pH less than 5. Alcohol and smoking were allowed. Patients were specifically encouraged to participate in activities that may have been associated with prior chest pain episodes. Meals and chest pain episodes were noted on a written diary card and recorded via an event marker on the pH monitor. All tracings were analyzed by computer and inspected by one of the authors to confirm the calculations and ensure the quality of the tracings. A drop of pH below 4 lasting greater than 20 seconds was considered evidence of GER [17]. The end of a reflux episode was defined as the point at which the pH had risen from below 4 and remained above that pH value for at least 6 seconds. Previous studies in our laboratory have suggested that the most reproducible 24-hour pH parameters are the percentages of upright, recumbent, and total acid exposure times [18]. Therefore, GER disease was deemed present if any of these three parameters exceeded the 95th percentile for normal values obtained by studying 48 healthy volunteers (24 men, 24 women, mean age 32 years, range 19 to 57 years): upright time greater than 6.7%, recumbent time greater than 2.4%, and total time greater than 4.7%. A subgroup of patients has recently been identified who, despite a normal 24-hour pH test, have chest pain episodes temporally associated with periods of acid reflux. A chest pain episode was defined as secondary to acid reflux if the esophageal pH dropped below 4 for longer than 20 seconds and occurred up to 5 minutes before the onset of chest pain. To assess the prevalence of reflux-mediated chestpain episodes, we calculated a "symptom index" in all patients with chest pain during the 24hour pH study [19]. This index used the following formula: number of chest pain episodes occurring when the pH was below 4 divided by the total number of chest pain episodes reported. This quotient was then multiplied by 100 to give the percentage of chest pain episodes associated with reflux. Thus, a negative symptom index implies that no chest pains were acid-mediated, whereas a positive symptom index means at least one chest pain episode was associated with acid reflux. At this time, the appropriate percent cutoff for a positive symptom index is unknown but, for the purposes of this study, any

ESOPHAGEALpH MONITORING/ HEWSONET AL

positive association was considered evidence that GER disease was probably contributing to the patient's complaints of chest pain,

TABLE I

Relationship Between EsophagealTest Abnormalitiesand Esophageal Symptoms

Statistical Analysis

Positive Abnormal Provocative Positive Both Manometry Tests GER SI GER& SI (n=28) (n--22) (n=24) (n=28) Symptoms (n=32)

A l l data are reported as percentages or as mean 4-

SE. X 2 analysis (with Yates' correction for cell sizes less than five when appropriate) was used to compare prevalences between groups. Frequency of chest pain episodes and positive symptom index scores between groups were compared using Student's t-test for unpaired data. An esophageal test was defined as identifying the esophagus as a probable cause of chest pain only if provocative testing replicated the patient's identical chest pain or the patient's spontaneous chest pain correlated with at least one episode of acid reflux (positive symptom index) during 24-hour pH monitoring. This conservative definition was chosen because we believe that a definite diagnosis is only possible in these patients after an esophageal abnormality has been identified and the patients' symptoms have improved with therapy specific for that esophageal disorder [20]. Otherwise, abnormal manometry or acid reflux parameters without associated chest

Heartburn 24(75%)' 23(82%) 16(73%) 20(83%) 19(73%) Regurgitation 25(78%) 21(75%) 14(64%) 16(67%) 20(79%) Dysphagia 19 (59%) 14 (50%) 10 (45%) 16 (67%) 13 (50%) Odynophagia 7(22%) 4(14%) 3(14%) 2(8%) 4(15%) Nosymptoms 3 (10%) 3 (11%) 1 (5%) 2 (8%) 4 (]5%) GER= abnormalacidexposuretimeby24-hourpHtest;positiveSI= positivesymptom indexby24-hourpHtest. • Numberof patientswithsymptoms(percentageof patientswithabnormalesophageal tests).

pain only suggested the esophagus as a possible cause of pain. McNemar's test for matched samples [21] was used to compare the incidence of definite esophageal diagnoses after traditional esophageal tests and 24-hour esophageal pH monitoring.

RESULTS Esophageal Symptoms Esophageal symptoms were common in this consecutive group of patients with noncardiac chest

1OOpatients with NCCP

I

24 hour pH test

I

83 patients CP during test

i orma, , our

/

46 patients

\

I

I nor a, ,,our

/

I

37 patients

\

I NegativeSI I I POsitiveSI I [ PositiveSI I [ NegativeSI [ 22 patients 24 patients 26 patients 11 patients Figure 1. Results of ambulatory 24-hour esophageal pH monitoring in 100 consecutive patients referred by cardiologists for the evaluation of noncardiac chest pain (NCCP). Seve~nteenpatients did not experience chest pain (CP) during the pH test. Of the remaining 83 patients, 55% had normal acid reflux parameters while 45% had abnormal reflux parameters. More than twice as many patients with abnormal reflux parameters had at least one of their chest pain episodes coincident with acid reflux, i.e., positive symptom index (SI). Despite only physiologic amounts of acid reflux in the other group, 24 (52%) patients had a positive syrnptom index. Overall, 50 patients had at least one episode of their chest pain closely associated with gastroesophageal reflux.

May 1991 The American Journal of Medicine Volume 90

579

ESOPHAGEALpH MONITORING/ HEWSONET AL

pain (heartburn [74%], regurgitation [67%], dysphagia [49%], and odynophagia [14%]). Despite the absence of esophageal symptoms in 11 patients, three (27%) had motility abnormalities, three (27%) had positive provocative tests, and seven (63%) had evidence of acid reflux by 24-hour pH testing. As shown in Table I, the abnormalities recorded by esophageal testing were equally distributed among the patients regardless of the presence or type of esophageal symptoms.

index had significantly (p <0.01, unpaired t-test) more chest pain episodes than patients with a negative symptom index (7.6 + 0.9 versus 3.2 + 0.7 episodes, respectively).Table II summarizes the relationship between the symptom index scores and evidence of abnormal acid reflux parameters. The vast majority of patients (42 of 50, 84%) with a positive symptom index had a ratio greater than 25%. As ~ group, these patients were significantly(p <0.01, x 2 analysis) more likely to have abnormal acid exposure times than patients with symptom indexe~ ranging from 0% to 25%. However, the mean s y m p ~ m index was not significantly different (p >0.10, unpaired t-test)between the patients with a positive symptom index alone (58 ± 5%) and those patients with a positive symptom index and abnormal reflux parameters (67 ± 5%).

Traditional Esophageal Tests Traditional esophageal tests were abnormal and/ or reproduced the chest pain in 52 patients.Esophageal manometry was abnormal in 32 of 100 patients (32%): nonspeciflc esophageal motor disorder (21),nutcracker esophagus (six),diffuse esophageal spasm (four), and hypertensive L E S (one). However,"no patient with these motility abnormalities Overall Diagnostic Yield had spontaneous chest pain. The acid perfusion test Traditional esophageal tests identified the was positive in 18 of 95 patients (19%) and the edro- esophagus as probably contributing to chest pain in phonium test was positive in 15 of 78 patients 28% of patients: positive acid perfusion test, 13%; (19%). positive edrophonium test,10%; both tests positive, Evidence of GER disease (abnormal acid expo- 5%. Additionally, 24% of patients had abnormal masure and/or positive symptom index) by the 24- nometry resultsonly, suggesting the esophagus as a hour pH test was no more frequent in the patients possible cause of their chest pain. In contrast, a with abnormal manometry results (22 of 32 pa- positive symptom index during the 24-hour p H test tients, 69%) than in those with normal baseline ma- identified G E R as probably contributing to chest nometry (48 of 68 patients, 71%). Patients in whom pain in 50% of patients. Another 22% of patients (11 the acid perfusion test was positive more commonly without chest pain, 11 with a negative symptom had evidence of reflux disease than did patients index) had abnormal acid exposure times, suggestwith negative acid perfusion results (18 of 18 pa- ing that G E R may be a possible cause of their chest tients [100%] versus 48 of 77 patients [62%]; pain. These differences are particularly striking p <0.01, X2 analysis). A similar trend was seen with when the test results in individual patients are exthe edrophonium test: in 13 of 15 (87%) patients, amined. Twenty-four-hour p H testing permitted a the edrophonium test was positive, whereas 39 of 63 probable esophageal diagnosis in 31 patients when (52%) patients had a negative edrophonium test traditional tests were negative or inconclusive, but (p = 0.07, X 2 analysis). the converse was true in only nine patients. Therefore, 24-hour p H testing with symptom index was Ambulatory 24-Hour Esophageal pH Test significantly superior (p <0.001; McNemar's test) Abnormal parameters of acid exposure time were to traditional esophageal tests in identifying the found in 48 patients (upright, 12; recumbent, 18; esophagus as the probable source of chest pain both, 18). Seventeen patients did not have chest (Figure 2). This superiority (p <0.025) persisted pain during the 24-hour p H test, but 11 of these even when the eight patients with a positive symppatients had abnormal acid exposure times. Of the tom index between 1% and 25% (Table If) were remaining 83 patients with chest pain during the reclassifiedas having negative or inconclusive evi24-hour p H test, 37 patients (46%) had abnormal dence of GER. reflux parameters and 50 patients (60%) had a positive symptom index (Figure 1). As might be ex- COMMENTS pected, more than twice as many patients with abW e report the largeststudy to date investigating normal reflux parameters had a positive symptom patients with noncardiac chest pain with state-ofindex while patients with normal 24-hour acid pa- the art esophageal testing including manometry, rameters had an equal ratioof positiveand negative provocative tests,and 24-hour p H monitoring. All symptom indexes. patients were directly referred by their cardioloFor the 50 patients with a positive symptom in- gists;therefore,the frequency of esophageal abnordex, the mean score was 56%, median 50%, and malities found was not a selection bias introduced range 6% to 100%. Patients with a positive symptom by the gastrointestinalinvestigators.The 52% prev580

May 1991 The American Journal of Medicine Volume 90

ESOPHAGEALpH MONITORING/ HEWSONET AL

(p <0.001) to traditional esophageal tests in identifying the esophagus as the probable source of chest pain. Previous smaller studies using prolonged pH monitoring have identified a remarkably similar prevalence of GER disease in patients with noncardiac chest pain. DeMeester et al [22] noted abnormal GER during 24-hour pH monitoring in 23 (46%) of 50 patients who had angina-type chest pain and normal coronary angiograms. Of the 13 reflux patients having chest pain during the study, 12 had the onset of pain coinciding with the reflux episode on the pH record. DeCaestecker and co-workers [7] found abnormal reflux in 14 (47%) of 30 patients who had unexplained chest pain. Schofield et al [23] identified GER in 23 (44%) of 52 patients with chest pain and normal coronary angiograms; however, exercise testing was often necessary to elicit their reflux symptoms. Eleven patients had abnormal 24hour esophageal pH scores. Nine of these patients, as well as an additional 13 patients with normal baseline reflux parameters, had their typical chest pain occur coincident with acid reflux during exercise testing. Two additional studies [24,25] using prototype combined esophageal pH/pressure systems found GER in 30% and 39% of patients with noncardiac chest pain. These studies have also

TABLE II Relationship Between Symptom Index and Abnormal Acid Reflux Parameters SymptomIndex

AbnormalRefluxParameters

0% 1%-25% 26%-50% 51%-75% >75%

11/33 patients(33%) 2/8 patients (25%) 10/18 patients(56%) 5/I0 patients(50%) 10/14 patients(71%)

alence rate of abnormal traditional esophageal tests is similar to previous studies from our laboratory [9] and others [6,7]. However, manometry and provocative testing only identified the esophagus as probably contributing to chest pain in 28% of patients: positive acid perfusion tests, 13%; positive edrophonium test, 10%; both tests positive, 5%. In contrast, outpatient 24-hour esophageal pH monitoring identified GER as probably contributing to chest pain in 50% of patients based upon their symptoms occurring coincident with acid reflux, i.e., a "positive symptom index." An additional 22% had abnormal reflux parameters that could not be associated with their chest pain symptoms. Overall, 24-hour pH monitoring was significantly superior

Traditional EsophagealTests ~ ~

iiiiiiil

( A l l tests

negative

~+APT 8 %~ +E /-

24 Hour pH Study

Possible

Possible

220/o

Iiiii!i !~iiii ii~!l

/

Probable

28%

"

Probable

50%

5%

Figure 2. Comparison between the diagnostic yield of traditional esophageal tests (manometry and provocative testing with acid and edrophonium) and 24-hour esophageal pH monitoring in 100 consecutive patients with noncardiac chest pain. An esophageal test was defined as identifying the esophagus as a probable cause of chest pain only if provocative testing replicated the patient's identical chest pain or the patient's spontaneous chest pain correlated with at least one episode of acid reflux (positive symptom index) during 24-hour pH monitoring. Otherwise, abnormal manometry or acid reflux parameters without associated chest pain only suggested the esophagus as a possible cause of chest pain. Esophageal pH monitoring with symptom index was significantly superior (p <0.001, McNemar's test) to traditional esophageal tests in identifying the esophagus as probably contributing to these patients' chest pain. +APT = positive acid perfusion test; +E -- positive edrophonium test; +GER = abnormal acid reflux parameters; +SI = positive symptom index. May 1991 The American Journal of Medicine Volume 90

581

ESOPHAGEALpH MONITORING/ HEWSONET AL

shown that GER is two to three times more common a cause of chest pain than esophageal motility disorders. Our study confirms the high prevalence of GER disease in a population with noncardiac chest pain. Esophageal symptoms were common in our study but did not distinguish patients with GER disease from those with motility disorders or positive provocative tests. This may represent the common interaction between acid reflux and motility disturbances as well as the limited number of responses that the esophagus has to noxious stimuli. For example, spastic motility disorders may promote poor acid clearance, thereby producing symptoms of heartburn and regurgitation, whereas acid reflux can produce complaints of dysphagia from esophageal inflammation or secondary motility disturbances. However, the lack of esophageal symptoms does not preclude the presence of esophageal test abnormalities, especially GER. It appears that approximately 10% of patients will have no esophageal symptoms at the time they present with severe chest pain. GER disease may be the common link between most of the esophageal test abnormalities observed in patients with noncardiac chest pain. From a treatment standpoint, this would be very important since drug therapy for spastic motility disorders (nitrates, anticholinergics, calcium channel blockers) could exacerbate acid reflux [8]. The resuits of this study and others are tantalizing but inconclusive. Schofield et al [26] found that GER was significantly more common (p <0.01) in patients with esophageal motility disorders (19 of 26, 73%) than in patients with normal manometry (five of 26, 19%). We could not confirm this observation. Abnormal 24-hour acid exposure times and/or positive symptom indexes were no more frequent in our patients with esophageal motility disorders (22 of 32, 69%) than in those patients with normal manometry (48 of 68, 71%). However, we did note a trend that nearly reached significance (p = 0.07) for GER to be more common in patients with a positive edrophonium test (13 of 15, 87%) than in those patients with a negative test (39 of 63, 52%). A similar overlap between positive edrophonium testing and GER has been reported by Schofield et al [26] and Vantrappen et al [27]. The patients with a positive symptom index, who account for 50% of the abnormalities identified by 24-hour pH testing, are an interesting group. Nearly half of these patients (49%) had normal acid reflux parameters, but their mean symptom index (58 ± 5%) was not significantly different from that of patients with a positive symptom index and abnormal reflux parameters (67 + 5%). In agreement with our experience, Schofield and colleagues [23] found 582

May 1991 The American Journal of Medicine Volume 90

that 13 of 22 (59%) patients with a positive symptom index had normal 24-hour acid reflux parameters. However, two other reports [7,22] infrequently documented acid-related symptoms without associated abnormal reflux parameters. It is not known whether patients with a positive symptom index alone represent a group with hypersensitivity to physiologic episodes of acid reflux or patients in whom abnormal quantities of acid reflux were missed by the pH probe [28]. We also do not know if these acid-sensitive patients are as responsive to antireflux therapy as are patients with a positive symptom index and concomitant abnormal reflux parameters. The frequency of acid reflux-induced symptoms in patients with noncardiac chest pain is an important finding since excellent medical and surgical therapies are available for GER disease. However, does acid-induced chest pain respond as well as heartburn to vigorous antireflux measures? Only the study by DeMeester et al [22] addresses this issue, and it is not a controlled trial. Of the 12 patients they identified with documented acid reflux during chest pain episodes, eight had symptom relief following Nissen fundoplication and four improved with medical therapy (postural changes, antacids, cimetidine). The remaining 11 patients with abnormal reflux parameters but no correlation of their chest pain with reflux or no chest pain during monitoring did not achieve such good results. Two of three patients responded to antireflux surgery, but only three of eight medically treated patients were pain-free. This excellent response to antire flux surgery suggests that near total elimination of acid reflux may be required for relief of chest pain. Recommended doses of H2 blockers may not accomplish this goal [29,30], consistent with our observations that these patients frequently need higher doses (ranitidine 300 mg twice a day or famotidine 40 mg twice a day) for initial pain control. The substituted benzimidazole, omeprazole, which markedly suppresses gastric acid secretion by inhibiting the H-K-ATPase found on the parietal cell secretory surface, offers another exciting therapeutic option. Additionally, omeprazole may permit investigators to characterize the treatability of the various abnormal groups identified by esophageal pH testing. Unfortunately, the sensitivity and specificity of these tests for identifying an esophageal source of chest pain are not known because no "gold standard" exists for the diagnosis of esophageal chest pain. It is known that these tests are rarely positive in healthy individuals [10,12,16,18,unpublished data]. Furthermore, esophageal motility disorders have only been reported in 10% of patients with angina pectoris and significant coronary artery dis-

ESOPHAGEAL pH MONITORING / HEWSON ET AL

ease [26], but the prevalence of positive provocative tests and acid reflux has not been studied. In addition, the symptom index needs further refinement to exclude the possibility that some of the refluxassociated events may occur randomly. Nevertheless, similar limitations in sensitivity and specificity exist for the tests diagnosing microvascular angina [11] and panic disorders [31], which are the "new" candidate diseases popularized as common causes of noncardiac chest pain. In the end, the only means to resolve these issues is the development of effective therapies for these pain problems. At this time, only acid reflux disease has effectivemedical and surgical therapy, which reinforcesthe need not to miss this diagnosis. Our resultssuggest that 24-hour esophageal p H monitoring is the singlebest testfor evaluatingpatients with noncardiac chest pain. The 50% prevalence of acid-induced chest pain is important since these patientshave a potentiallytreatableproblem. Esophageal p H monitoring can be done on an outpatient basis with excellentdiagnosticyieldmerely by encouraging activitiespreviouslyknown to exacerbate chest pain. Unfortunately, current accurate placement of the p H probe requires the L E S to be located by esophageal manometry [32].Future p H systems will need to include a pressure transducer in the p H probe to simplify this task. The evaluation of noncardiac chest pain could then be streamlined, with manometry and provocative testingreserved for patients with negative p H tests or those patients with positivetests who do not respond to vigorous antireflux regimens.

REFERENCES 1. Proufit WL, Shirley EK, Jones FM. Selective cine coronary arteriography: correlation with clinical findings in 1000 patients. Circulation 1966; 33: 901-10. 2. Kemp HG, Vokonas PS, Cohn PF, Gorlin R. The anginal syndrome associated with normal coronary arteriograms: report of a 6-year experience. Am J Med 1973; 54: 735-42. 3. Dart AM, Alban Davies H, Dalai J, Rutley M, Henderson AM. "Angina" and normal coronary arteriograms: a follow-up study. Eur Heart J 1980; 1: 97-8. 4. Ockene IS, Shay MJ, Alpert JS, Weiner BH, Dalen JE. Unexplained chest pain in patients with normal coronary arteriograms: a follow-up of functional status. N Engl J Med 1980; 303: 1249-52. 5. Brand DC, Martin D, Pope CE. Esophageal manometrics in patients with angina-like chest pain. Dig Dis Sci 1977; 22: 300-4. 6. Davies HA, Jones DB, Rhodes J. Esophagealangina as the cause of chest pain. JAMA 1982; 248: 2274-8. 7. deCaestecker JS, Blackwell JN, Brown J, Heading RC. The oesophagus as a cause of recurrent chest pain: which patient should be investigated and which tests should be used? Lancet 1985; 2: 1143-6. 8. Richter JE, Bradley LA, Castell DO. Esophagealchest pain: current controversies in pathogenesis, diagnosis and therapy. Ann Intern Med 1989; 110: 66-78. 9. Katz PO, Dalton CB, Richter JE, Wu WC, Castell DO. Esophageal'testing in

patients with non-cardiac chest pain or dysphagia: results of three years' experience with 1161 patients. Ann Intern Med 1987; 106: 593-7. 10. Johnson LF, DeMeester TR. Twenty-four hour pH monitoring of the distal esophagus: a quantitative measure of gastroesophageal reflux. Am J Gastroenterol 1974; 62: 325-32. 11. Cannon RO, Watson RM, Rosing DR, Epstein SE. Angina caused by reduced vasodilator reserve of the small coronary arteries. J Am Coil Cardiol 1983; 1: 1359-73. 12. Cannon RO. The gastroenterologist and microvascular angina. Gastroenterology 1990; 98: 1103-5. 13. Ott DJ, Chen YM, Gelfand DW, Wu WC. Analysis of the multiphasic radiographic examination for detecting reflux esophagitis. Gastrointest Radio11986; 11: 1-6. 14. Richter JE, Wu WC, Johns DN, eta/. Esophageal manometry in 95 healthy adults: variability of pressures with age and frequency of "abnormal" contractions. Dig Dis Sci 1987; 32: 583-92. 15. Richter J E, Johns DN, Wu WC, Castell DO. Are esophageal motility abnormalities produced during the intraesophageal acid perfusion test? JAMA 1985; 253: 1914-7. 16. Richter JE, Hackshaw BT, Wu WC, Castell DO. Edrophonium: a useful provocative test for esophageal chest pain. Ann Intern Med 1985; 103: 14-21. 17. Shaker R, Kahrilas PJ, Dodds WJ, Hogan WJ. Esophageal clearance of small amounts of acid [Abstract]. Gastroenterology 1986; 90: 1628A. 18. Wiener GJ, Morgan TM, Copper JB, eta/. Ambulatory 24-hour esophageal monitoring. Reproducibility and variability of pH parameters. Dig Dis Sci 1988; 33: 1127-33. 19. Wiener GJ, Richter JE, Copper JB, Wu WC, Castell DO. The symptom index: a clinically important parameter of ambulatory 24-hour esophageal pH monitoring. Am J Gastroenterol 1988; 83: 358--61, 20. Richter JE. Selected summary: esophageal test for all seasons and all reasons but.., how do they help the patient? ? Gastroenterology 1990; 99: 18468. 21. McNemar Q. Psychological statistics. New York: Wiley, 1955. 22. DeMeester TR, O'Sullivan GC, Bermudez G, MideU AI, Cimochowiski GE, O'Drobinak JO. Esophagealfunction in patients with angina-type chest pain and normal coronary angiograms. Ann Surg 1982; 196: 488-98. 23. Schofield PM, Bennett DH, Whorwell PJ, et al. Exertional gastro-oesophageal reflux: a mechanism for symptoms in patients with angina pectoris and normal coronary angiograms. Br Med J 1987; 294: 1459-61. 24. Janssen J, Vantrappen G, Ghillebert G. 24-hour recording of esophageal pressure and pH in patients with non-cardiac chest pain. Gastroenterology 1986; 90: 1978-84. 25. Peters LJ, Maas LC, Petty D, eta/. Spontaneous non-cardiac chest pain: evaluation by 24-hour ambulatory esophageal motility and pH monitoring. Gastroenterology 1988; 94: 878-86. 26. Schofield PM, Whorwell PJ, Brooks NH, Bennett DH, Jones PE. Oesophageal function in patients with angina pectoris: a comparison of patients with normal coronary angiograms and patients with coronary artery disease. Digestion 1989; 42: 70-8. 27. Vantrappen G, Janssen J, Ghillibert G. The irritable oesophagus: a frequent cause of angina-like chest pain. Lancet 1987; 1: 1232-4. 28. Murphy DW, Yuan T, Castell DO. Does the intraesophageal pH probe accurately detect acid reflux? Dig Dis Sci 1989; 34: 649-56. 29. Orr WC, Robinson MG, Humphries TJ, Antonella J, Caglalio A. Dose response effect of famotidine on gastroesophageal reflux [Abstract]. Gastroenterology 1987; 92: 1562A. 30. RussellJ, Orr WC, King JF, Finn AL. The effects of high doses of ranitidine on esophageal reflux and severity [Abstract]. Am J Gastroentero11988; 83: 1025A. 31. Beitman BD, Mukerji V, Lamberti JW, eta/. Panic disorders in patients with chest pain and angiographically normal coronary arteries. Am J Cardio11989; 63: 1399-1403. 32. Walther B, DeMeester TR. Placement of the esophageal pH electrode for 24hour esophageal pH monitoring, In: DeMeester TR, Skinner DB, eds. Esophageal disorders: pathophysiology and therapy. New York: Raven Press, 1985: 53941.

May 1991

The American Journal of Medicine

Volume 90

583