School Work

Value of extended recording time with wireless pH monitoring in evaluating gastroesophageal reflux disease

Value of extended recording time with wireless pH monitoring in evaluating gastroesophageal reflux disease
of 6
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
  ORIGINAL ARTICLES Value of Extended Recording Time With Wireless pHMonitoring in Evaluating Gastroesophageal Reflux Disease CHANDRA PRAKASH and RAY E. CLOUSE Division of Gastroenterology, Washington University School of Medicine, St Louis, Missouri Background & Aims:  Extending an ambulatory pH studyby using a wireless capsule system could improve diag-nosis of gastroesophageal reflux disease in patients withsymptoms potentially related to acid reflux events. Wedetermined the relative value of a 2-day versus a 1-daystudy in detection of abnormal esophageal acid expo-sure and reflux symptom associations.  Methods:  Sub- jects who had undergone pH monitoring with the wire-less system for symptom evaluation and who had atleast 1 typical or atypical symptom during monitoringwere studied (n    157). Acid exposure time, symptomindex, and a measure of reflux-associated symptomprobability were calculated after 1 day of recording andcompared to final results from analysis of 2 days ofdata.  Results:  Extending recording time increased thenumber of subjects recording symptoms by 6.8% anddoubled the number of symptoms available for associ-ation with acid reflux events. Other advantages weregreater in subjects off antireflux therapy, wherein length-ening recording time increased the proportion of sub- jects with abnormal acid exposure time by 12.4% andsignificantly increased the likelihood of a positive refluxsymptom relationship in subjects with atypical symp-toms. Accuracy of a 1-day study ranged from 0.87–0.88for acid exposure time to as low as 0.72–0.87 for theprobability test, the most sensitive of the symptom as-sociation tests.  Conclusions:  Extending recording time to2 days with the wireless pH monitoring system increasesthe likelihood of detecting reflux disease in patientsundergoing symptom evaluation. Benefits are most evi-dent for subjects with atypical symptoms and for estab-lishing reflux symptom associations by probabilitytesting. A mbulatory pH monitoring is a valuable tool in theclinical evaluation of gastroesophageal reflux disease(GERD). 1,2 The methods help establish the presence of abnormal esophageal acid exposure and determine mean-ingful relationships between symptoms and acid refluxevents. Both are important in determining the relevanceof reflux disease to symptomatic presentations and inselecting appropriate candidates for surgical antirefluxinterventions. 1–5 Wireless pH monitoring by using a radio transmitterattached to the distal esophageal mucosa offers a method of ambulatory monitoring with extended recording capabili-ties over conventional monitoring systems. 6,7 By increasingrecording time from 1 to 2 days, the approach appeared toincrease the yield for detecting abnormal distal esophagealacid exposure in a small group of subjects. 6 A better esti-mate of the relationship between symptoms and acid refluxeventsalsowouldseemlikely,butthishasnotbeenstudied.An increase in number of patients with symptoms as well asan increase in numbers of symptoms for comparison toreflux events would be predicted. Calculation of a statisti-cally significant reflux-symptom relationship by using re-flux-associated symptom probability (RASP) tests is depen-dent on the number of reflux events, symptoms, and reflux-related symptoms, as well as the total recording time. 8,9 Reaching a statistically significant probability that the as-sociation did not occur solely by chance should be enhancedwith the extended recording time of the wireless system;confidence in the result, whether positive or negative, alsoshould be improved with a longer period of observation.In this study, we examined the gains realized from awireless pH monitoring system when extending the record-ing time from 1–2 days in patients with symptoms poten-tially attributable to GERD. The data from the first day of recording were compared to the final outcome, with 2 daysof recording as the reference standard. Besides measuringthe advantages, we were interested in calculating the limi-tations of a single day of data in establishing the presence of GERD in a symptomatic clinical population.  Abbreviations used in this paper:  AET, acid exposure time; GERD,gastroesophageal reflux disease; GPE, Ghillebert probability estimate;RASP, reflux-associated symptom probability; SAP, symptom associa-tion probability; SI, symptom index. ©  2005 by the American Gastroenterological Association1542-3565/05/$30.00PII: 10.1053/S1542-3565(05)00021-2 CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2005;3:329–334  Methods The subjects for study represent a subset of adultpatients (  18 years old) who had undergone extended ambu-latory pH monitoring with the BRAVO pH capsule system(Medtronic, Inc, Shoreview, MN) at the outpatient facility of Barnes-Jewish Hospital, St Louis, Missouri. Patients who hadbeen referred for pH monitoring to evaluate a predefined,suspected reflux symptom and who recorded this symptom atleast once during the 2-day monitoring period were identifiedretrospectively from the subject database. This included sub-jects undergoing preoperative evaluation, investigation of un-explained symptoms, and evaluation of symptoms persistingafter medical antireflux therapy had been initiated. Conse-quently, subjects with either typical or atypical symptomswere included, as were subjects who were studied either on oroff antireflux therapy. For the purpose of this study, onlyheartburn and regurgitation were considered typical symp-toms. Subjects had discontinued proton pump inhibitors atleast 7 days before the study and histamine 2  receptor antago-nists, prokinetic agents, and antacids at least 3 days beforeplacement of the pH capsule to be considered off therapy.Those who continued treatment during monitoring were con-sidered on therapy. Subjects failing to adhere to this dichoto-mous division were not studied, eg, those with intermittentuse of antisecretory medications. Subjects who had undergoneprior antireflux surgery, failed to maintain a complete diary of activities and symptoms through the length of the pH study,or had   14 hours of good quality tracings on each of the 2study days were excluded from further consideration. Thisstudy was approved by the Human Studies Committee (Insti-tutional Review Board) of Washington University School of Medicine.All potential subjects underwent an upper endoscopic exami-nation immediately before capsule placement; the capsule was notattached if esophagitis was detected. The procedures were per-formed under conscious sedation by using conventional but vari-able combinations of midazolam, meperidine, and fentanyl; somesubjects received propofol. Topical lidocaine for oropharyngealanesthesia was used for subjects who did not receive propofol.Distance from the incisors to the squamocolumnar junction wasmeasured in centimeters at the time of endoscopy. The pHcapsule was attached 6 cm proximal to the junction after oralintroductionbyusingstandardtechnique. 6 Visualconfirmationof attachment uniformly was obtained in subjects on antirefluxtherapy at the time of endoscopy and if alkaline pH recordingswere not observed immediately after placement in patients off therapy. Subjects maintained a daily diary that included symp-toms, activities, and meal periods and returned the receiver after2 days of pH recording.The relative values of a 1-day and 2-day study were exam-ined by measuring outcomes after 1 day of recording andcomparing these to final decisions extracted from the full 2days of recording. This was performed for a quantitative mea-sure of acid exposure and 2 measures of symptom-reflux asso-ciation. Acid exposure time (AET) was used to quantitatereflux, was defined as the percent time the esophageal pHremained below 4, and was calculated separately for day 1 andday 2 of the study. AET  5.33% was used as the threshold forabnormality. 6 Abnormal AET on either of the 2 days was usedas the final determinant of increased acid exposure on thecomplete study. 6 The symptom index (SI) was calculated byusing the method of Wiener et al 10 as the ratio of number of reflux-related symptoms to the total number of recorded symp-toms, expressed as a percentage. This was calculated separatelyfor day 1 and for the total study period. SI   50% wasconsidered indicative of a positive association.The Ghillebert probability estimate (GPE) was used tostudy RASP, 9 and a  P  value    .05 was required for signifi-cance (ie, a positive test result). Like the symptom associationprobability (SAP), this RASP test determines the likelihoodthat a symptom and acid reflux events co-occur solely bychance. The probability is calculated from the sum of partialprobabilities for exact numbers of reflux-associated symptomswithin the context of the total number of symptoms, theproportion of time “at risk” for linking a symptom to a low pHvalue, and the total recording time. Unlike the SAP, the GPEcan be applied post hoc to data extracted from a pH study,correlates highly with RASP tests similar to the SAP, buttends to underestimate statistically significant associations. 11 In part, this relates to the higher “at risk” time estimated fromsumming the total time pH   4 with the 2-minute intervalspreceding reflux events and, thus, the potential for duplicatedcounting of “at risk” time if reflux events are spaced by   2minutes. The GPE was chosen as the RASP test in thisinvestigation because symptom and reflux event informationfor each 2-minute interval was unavailable from all studies,data required for calculation of the SAP and similar tests. 8,12 Like the SI, the GPE was calculated for the first day of pHrecording and compared to a calculation by using the entire2-day study.Data are reported as mean    standard error of the meanunless otherwise indicated. Grouped data were compared byusing two-tailed Student  t   test and Fisher exact test as appro-priate. In all instances, a  P  value    .05 was required forstatistical significance. For determining the value of a 1-daystudy compared with the complete, 2-day study, results fromthe latter were used as the gold standards. Test performancecharacteristics, including sensitivity, specificity, positive pre-dictive value, negative predictive value, and accuracy, werecalculated for the first day of data collection against the goldstandard. Results A total of 157 subjects (age, 47    1 years; 100women) were included in this study. Of these, 87 (55.4%)recorded typical reflux symptoms, whereas the remaining70 subjects (44.6%) had atypical symptoms potentiallyattributable to GERD (chest pain 43, cough 18, othercomplaints 9). The study was performed off antireflux ther-apy in 121 subjects (77.1%; 67 with typical symptoms, 54 330 PRAKASH AND CLOUSE CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 3, No. 4  with atypical symptoms) and on proton pump inhibitors inthe remaining 36 subjects (22.9%; 20 with typical symp-toms, 16 with atypical symptoms). The duration of pHrecordingaveraged43.7  0.2hoursforthecompletestudy(22.0  0.1 hours on day 1, 21.7  0.1 hours on day 2).Sedation and placement procedures appeared to have nosystematic effect on findings, in that, for all subjects, therewere no significant group differences in AET, number of reflux events, number of symptoms, or number of reflux-related symptoms between day 1 and day 2 of study. Targetsymptoms were recorded in the diary by 147 subjects onday 1 and by all 157 subjects by the end of the 2-day study,a 6.8% gain. Subjects Off Antireflux Therapy Of the 121 subjects (age, 46.8    1.2 years; 78women) who underwent pH monitoring off antirefluxtherapy, 67 (55.4%) had typical reflux symptoms, and 54(44.6%) had atypical symptoms. AET was greater in thesubset of subjects with typical reflux symptoms com-pared to those with atypical symptoms for day 1 as wellas the total study period (day 1: 8.6%  0.8% vs 5.3%  0.8%; total: 8.7%  0.8% vs 5.6%  0.7%;  P  .01for each comparison). Increasing the duration of pHrecording to 2 days increased detection of subjects withelevated AET in both subsets; the increment (15 sub-jects, 12.4%) trended toward statistical significance forthe entire group and for those with typical symptoms (  P  .1 for each) (Figure 1). Extending the recording time increased the number of subjects reporting symptoms from 115 on day 1 to 121by the end of the study (5.2% increase). A similar gainoccurred in both symptom subgroups (typical symptoms,4.7%; atypical symptoms, 5.9%). The number of symp-toms available for association with reflux events alsosignificantly increased from day 1 to the entire studyperiod (15.2    1.9 vs 31.5    3.6,  P    .0001) andessentially doubled for both symptom subgroups (ratio of symptom number during the total period to the numberon day 1: 2.1 for subjects with typical symptoms, 2.0 forsubjects with atypical symptoms).The GPE was more likely than the SI to detect sub-jects with reflux-symptom associations, a difference thatreached statistical significance during the total studyperiod (Figure 2). The increase in proportion of subjects with a positive reflux-symptom association was mostpronounced for the GPE in patients with atypical symp-toms (Figure 2). The changes shown in the figure, how- ever, do not reflect completely the potential advantagesof extending recording to 2 days, because some subjectswho were found to have a significant reflux-symptomassociation after 1 day of recording did not retain thisfinding when longer periods of observation and greaternumbers of symptoms and reflux events were taken intoconsideration. Fully 24 subjects (19.8%) either devel-oped (18 subjects) or lost (6 subjects) statistically signif-icant reflux-symptom associations by extending thestudy to 2 days. A smaller number of subjects exhibitedchanges in the outcome for the SI (17 subjects, 14.0%).Four subjects with a positive association by the SI had aninsignificant association by the GPE by the end of thestudy. Considering the AET, the reflux-symptom asso-ciation tests, and number who only reported symptomson day 2 of the study, 40 subjects (33.1%) off antirefluxtherapy gained useful information potentially leading toa GERD diagnosis by the second day. Figure 2.  Increment in the establishment of reflux-symptom associa-tions from 1 day of recording ( solid bars  ) to the entire 2-day period( solid plus hatched bars  ) with a comparison of findings between the 2tests used. The GPE was more likely to identify subjects than the SIby the end of the study (* P   .03 compared to the SI). The incrementin positive findings on the GPE for patients with atypical reflux symp-toms also was statistically significant at the conclusion of the study (** P   .01 compared to the 1-day results.) Figure 1.  Increment in the detection of elevated AET comparing 1 day of recording to the final results from the 2-day study. An elevation inAET on either recording day was considered abnormal. The differenceapproached statistical significance when all subjects or those withtypical symptoms were considered (* P   0.1). April 2005 WIRELESS pH MONITORING IN GERD 331  Subjects on Antireflux Therapy A total of 36 subjects (age, 48    2 years; 22women), 20 (55.6%) with typical symptoms and 16(44.4%) with atypical symptoms, underwent pH monitor-ing while on antireflux therapy. AETs were equally sup-pressed in subjects with typical and atypical symptoms (day1: 1.6%  0.6% vs 2.5%  0.8%; total: 2.6%  1.1% vs1.8%    0.5%;  P    .3 for each comparison). Only 7(19.4%) subjects had elevated AET on either day of study.Four subjects reported symptoms solely on day 2 of study,a 12.5% gain in symptomatic subjects during monitoring(17.6% for those with typical symptoms, 6.7% gain forthose with atypical symptoms). Number of recorded symp-toms again increased significantly from 1–2 days of record-ing (11.8  2.4 vs 26  5.3,  P  .0001), with nearly adoubling of symptoms in both subject groups (ratio of symptom number during the total period to the number onday 1: 2.5 for subjects with typical symptoms, 1.9 forsubjects with atypical symptoms).The GPE identified 6 subjects with a significant symp-tom-reflux association on day 1 and 9 subjects across theentire study; 4 subjects had a positive SI on day 1 and 5subjects during the total study period. Although a greaterproportion of subjects had a positive GPE than a positive SIat each recording length, no differences were significant. Asfor the subjects off antireflux therapy, subjects both gainedandlostmeaningfulreflux-symptomassociationsbyextend-ing the recording period. The GPE changed in 5 subjects(13.9%) and the SI in 3 (8.3%); most changes were gains inassociation. The numbers were too small to examine sepa-rately subjects with typical or atypical reflux symptoms.Considering the AET, the reflux-symptom association tests,and number who only reported symptoms on day 2 of thestudy, 8 subjects (22.2%) on antireflux therapy gaineduseful information potentially leading to a GERD diagnosisby the second day. Performance Characteristics of a One-DayStudy To take into consideration all changes that occurredin detecting abnormal acid exposure and reflux-symptomassociations between the first day and the final decision,sensitivities, specificities, accuracies, and predictive valueswere calculated for the day 1 results by using the outcomefrom 2 days of recording as the gold standard (Table 1). The number of abnormalities in subjects who remained on an-tireflux therapy for the study was small, and only thosesubjects who were off therapy were used for these calcula-tions. Because an elevation in AET on either day of studywas considered indicative of an abnormal result, the speci-ficity of an abnormal AET on day 1 is 1.0, by definition.However, the sensitivity of only 1 day of data was only 0.80for all subjects; the value of a limited study was mostpronounced in the subjects with atypical symptoms (sensi-tivity, 0.73).The limitations of 1 day of recording compared with 2daysfromthewirelesssystemalsoareapparentbyreviewingthe findings from the symptom association tests in Table 1.Specificities on day 1 were higher than sensitivities for boththe SI and the GPE; consequently, the positive predictivevalues are better than the negative predictive values from 1day of data. Because the SI was less affected than the GPEby extending the recording time, a single day of data moreaccurately predicted the final outcome. However, as de-scribed above, the GPE was more likely to detect a reflux-symptom association (  P  .03). This RASP test was par-ticularly influenced by extending recording time, an effectthatwasmostpronouncedinthosewithatypicalsymptoms.The GPE after 1 day was sensitive to only 48% of thesesubjects; the resultant negative predictive value in thispatient mix was 0.7, and the overall accuracy of the 1-dayrecording was only 0.72. Table 1.  Value of 1 Day of pH Recording in Predicting the Final Results in Subjects Off Antireflux Therapy at Time of Study  Sensitivity Specificity Accuracy Positive predictive value Negative predictive valueElevated AETAll subjects 0.80 1 0.88 1 0.76Subjects with typical symptoms 0.83 1 0.88 1 0.70Subjects with atypical symptoms 0.73 1 0.87 1 0.80Positive SIAll subjects 0.77 0.93 0.88 0.83 0.89Subjects with typical symptoms 0.81 0.88 0.85 0.81 0.88Subjects with atypical symptoms 0.69 0.98 0.91 0.9 0.91Positive GPEAll subjects 0.68 0.91 0.8 0.87 0.76Subjects with typical symptoms 0.82 0.91 0.87 0.9 0.83Subjects with atypical symptoms 0.48 a  0.9 0.72 0.79 0.7 a  P   .01 compared to subjects with typical symptoms. 332 PRAKASH AND CLOUSE CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 3, No. 4  Discussion In this report, we found an increased yield in theidentification of GERD by extending pH recording timewith a wireless pH monitoring system. Overall, additionalinformation that could incriminate GERD in a symptom-atic patient (reporting of symptoms, detection of elevatedAET, positive SI, and significant reflux-symptom associa-tion) was obtained in 48 subjects (30.1%) compared to a1-day study. The benefit was most evident when symptom-atic subjects were studied off acid suppressive therapy,wherein an additional 12.4% of subjects exhibited elevatedAET,andmorethan1of5eitherdevelopedorlostevidenceofreflux-symptomassociationsbyextendingthedurationof pH recordings. Subjects with atypical symptoms comparedto those with heartburn or regurgitation had the greatestgains.The wireless pH monitoring system is considered a sig-nificant advance in pH monitoring for its practical advan-tages. Technical aspects of capsule deployment are easy tomaster, and the procedure is better tolerated compared withcatheter-based pH testing. 6,7,13 Patients report less impair-ment in activities of daily living and little need for dietarymodification, allowing pH monitoring during a patient’snormal daily routine. 13 The pH capsule typically falls off spontaneously within 2 weeks. 14 Discomfort from the cap-sule is reported by up to 34% of patients 6 but requiresendoscopic capsule dislodgement in a very small percentageof cases. 15 Data adequate for pH evaluation were obtainedin 89%–96% of patients during a 36- to 48-hour pHmonitoring period. 6,7 Early studies with simultaneous wireless and catheter-based pH monitoring demonstrated concordance of tech-niques in quantifying esophageal acid exposure. 16,17 Subse-quently, wireless monitoring was shown to distinguishaccurately patients with GERD from control subjects witha sensitivity of 84% and a specificity of 85% 6 by using anupper limit of normal for total esophageal acid exposurethat is comparable to or slightly higher than that derivedfrom catheter-based pH monitoring (5.3% vs3.4%–5.8%). 6,18–21 Some investigators have suggestedthat measurements obtained from the second day of record-ing might have greater accuracy in distinguishing patientswith GERD from control subjects, because normal dailyactivities and symptom reporting might be suppressed byconscious sedation used during capsule placement on thefirst day of study. 22 We could not document a systematiceffect of sedation and the procedure on symptom reportingin our subjects, however. Preliminary data also have sug-gested a role for the extended study in evaluating atypicalsymptoms, such as chest pain. 23 Thus, results are accumu-lating to support an important role of the wireless moni-toring system in GERD evaluations.This study has quantitated the diagnostic advantagerealizedbyextendingpHrecordingtimewithawirelesspHsystem when evaluating symptomatic patients for the pres-ence of GERD. Identifying additional patients with ele-vated AET forms part of the advantage, but greater gainsare obtained in establishing or refuting significant reflux-symptom associations, particularly with the use of RASPtests. Extending the recording time for ambulatory pHmonitoring beyond 1 day increases both the likelihood of having symptoms during the study and the number of symptoms available for association with reflux events.Longer monitoring enhances the likelihood of finding ahigh proportion of reflux-related symptoms and of detect-ing patients in whom symptoms and reflux events arestatistically unlikely to have occurred by chance. Theseadvantages are important, because both presence of elevatedAET and presence of meaningful reflux-symptom associa-tions are helpful in selecting candidates for antireflux sur-gery and predict favorable response. 1,5,12 The RASP test used in this investigation was the mostaffected measurement parameter and influenced final deci-sions of reflux-symptom associations in 1 in 5 subjects. TheGPE was proposed as a method of calculating an associationprobability post hoc from ambulatory pH studies by usingcommonlyextracteddata,specificallynumberofsymptoms,number of reflux-associated symptoms, number of refluxevents, AET, and total recording time. 9 This method canunderestimate significant associations in approximately14% of subjects, yet rarely overstates the case. 11 The chanceof error is greater in cases having large numbers of symp-toms and high AET, and, thus, some subjects losing sig-nificance with extended recording might not have done so if the SAP had been used. However, it remains unknownwhether either method is superior in predicting outcome byusing an external standard, such as response to medical orsurgical therapy. Preliminary data indicate that most dis-crepant cases between the SAP and GPE demonstrate atrend toward significance by the GPE, 11 and a small num-ber of our RASP-negative cases did fall into this level of statistical significance by the end of the study (8%). 12 Thus,it is possible that a larger number of subjects would havegained evidence supporting a GERD diagnosis by usingother RASP test methods. 8,12 A more important question relates to the importance of this type of reflux-symptom association testing in clinicaldecision making. Probability testing provides the statisticallikelihood that the symptom and acid reflux events areassociated purely by chance. It offers supportive evidencethat a symptom and acid reflux are linked and likely should April 2005 WIRELESS pH MONITORING IN GERD 333
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks