Address for Correspondence Dr. Mark Fox, Nottingham Digestive Diseases Centre and Biomedical Research Unit, Queen’s Medical Centre, Nottingham, NG7 2UH, UK. Tel: +44 (0)115 9249924; fax: +44 (0)115 9422232; e-mail: email@example.com
Background Catheter-based esophageal pH-monitoring is used to evaluate patients with suspected gastro-esophageal reflux disease (GERD); however false-negative results may occur due to poor tolerance of the catheter with reduced oral intake and activity, or high day-to-day variation in reflux and symptom events. We assessed diagnostic yield and clinical impact of prolonged, wireless pH-monitoring in patients with negative results from 24-h catheter-based studies and ongoing symptoms.
Methods Esophageal acid exposure (percentage time pH <4), Symptom Index, and Symptom Association Probability (SAP) were calculated. Diagnostic yield was assessed using Average (mean) and Worst Day (24-h period with highest acid exposure or symptom load) analyses. Outcome data were assessed 6–36 months (median 24) after initiation of definitive therapy based on physiologic testing.
Key Results Data from prolonged pH-monitoring up to 96-h (median 72-h) were available from 38 patients. Using Average and Worst Day analysis, esophageal acid exposure was pathologic in 37% and 47%, whereas SAP was positive in 34% and 63% of patients, respectively. Overall using Average and Worst Day analyses, 61% and 76% patients were diagnosed with GERD based on either pathologic acid exposure or positive symptom association. Of 12 patients that underwent antireflux surgery, 10(83%) reported a good outcome at a median 24 months follow-up.
Conclusions & Inferences Prolonged, wireless pH-monitoring increases test sensitivity and diagnostic yield in patients with continuing esophageal symptoms despite negative 24-h catheter-based pH-studies. Without a definitive diagnosis, many would not have received effective treatment.
Ambulatory esophageal pH-monitoring is required to document pathological esophageal acid exposure or reflux-symptom association in patients with symptoms suspected gastro-esophageal reflux disease (GERD), especially those with treatment refractory symptoms under consideration for antireflux surgery. Standard catheter-based pH-studies can be uncomfortable and socially embarrassing leading to altered behavior which can reduce reflux events and reflux-associated symptoms.1,2 This can result in a false-negative diagnosis of GERD. In addition, the clinical value and diagnostic yield of 24-h testing is diminished due to the high physiologic day-to-day variability in acid exposure and symptom reporting;2–4 limitations that are more pronounced when reflux events are infrequent and symptoms are intermittent.5
Wireless pH-monitoring is an innovative methodology that may address these limitations; a radiotelemetric capsule is attached to esophageal mucosa and transmits pH data wirelessly over several days.6 Prolonging pH measurement can increase diagnostic reproducibility and sensitivity based on esophageal acid exposure, especially in patients with intermittent symptoms and in the supine period.4 Moreover, the wireless system is better tolerated and is preferred by unselected patients.3,6 Disadvantages include higher costs and, in most centers, the need for endoscopic placement. At present certain national guidelines recommend that ‘wireless esophageal pH-monitoring is suitable (only) for patients who do not tolerate nasal intubation’ of the 24-h pH catheter.7
Another important group that may benefit from prolonged, wireless pH-monitoring is individuals with typical reflux symptoms but a non-diagnostic 24-h catheter-based investigation. Such patients are often considered to have ‘functional’ esophageal symptoms; however, given the limitations of catheter studies, a proportion may have false-negative results and be denied appropriate therapy. This study reports the diagnostic yield of prolonged, up to 96-h, wireless pH-monitoring in patients with negative 24-h catheter-based pH-studies referred for a second opinion with ongoing symptoms suggestive of GERD. The clinical impact of prolonged, wireless pH-monitoring was assessed at clinical follow-up 6–36 months (median 24) after initiation of definitive therapy based on the results of physiologic testing.
Between 29/11/2006 and 10/2/2010, 38 of the 246(15.4%) wireless pH-studies were performed at a tertiary referral center (St Thomas’ Hospital, London) for further investigation of ‘typical’ reflux symptoms (heartburn, regurgitation) in whom 24-h catheter-based pH-studies had not provided a diagnosis of GERD (normal acid exposure and negative or insufficient evidence of reflux-symptom association). Antisecretory medications were discontinued for at least 5 days prior to all pH-studies.8
24-h catheter-based pH-monitoring
Studies were performed by the Digitrapper™ Slimline™ (Medtronic Inc., Shoreview, MN, USA) system using standard technique for preparation, investigation, and analysis.9
Prolonged wireless pH-monitoring
A wireless pH system (Bravo®; Given Imaging, Yoqeam, Israel) was calibrated and a radio-telemetry capsule deployed 6 cm proximal to the squamo-columnar junction after 1 min of suction as described in the literature.6 Capsule attachment and position was confirmed on repeat endoscopy. Median (range) doses for sedation [midazolam 5 mg (3–10 mg) and fentanyl 50 mcg (50–100 mcg)] were higher than for diagnostic endoscopy [midazolam 4 mg (0–5 mg) alone]. After 48 h the patient returned for download of data and a change of batteries. Once the position and function of the pH-sensor was confirmed the patient was released for a further 48-h monitoring. Standardized verbal and written lifestyle advice was provided.
Patient follow-up was performed at least 6 months [median 24 (range 6-36 months)] after initiation of definitive therapy via telephone. Symptoms were rated as good (resolved), moderate or poor (refractory symptoms requiring medical therapy). Any decision for medical or surgical therapy was made by the referring clinician on the basis of the clinical and complete physiologic data.
Both esophageal acid exposure and symptom association data were assessed. Each 24-h pH-recording was analyzed and categorized as normal or pathological on the basis of the percentage time pH <4 and symptom association (described below). Technical success and duration of recording was noted. The effect of day-to-day variability of esophageal acid exposure and symptom association on diagnostic yield based on 24, 48, 72 and 96-h pH-monitoring was calculated. The diagnosis of GERD was based on either pathologic pH or reflux-symptom data. In the case of detachment prior to 96-h, the most complete data were used (last result was carried forward). The clinical impact of pH measurement on management and clinical outcome was assessed following definitive treatment decision at the last documented follow-up on an intention to treat basis (i.e., patients lost to follow-up were assumed not to have responded to treatment).
Esophageal acid exposure
Percentage esophageal acid exposure (pH <4) is the most robust diagnostic marker of GERD and was the primary outcome measurement (total reflux, TR).10 Acid reflux in the upright (UR) and supine (SR) positions were also documented. Measurements from 24-h catheter-based studies were compared to wireless monitoring at 24, 48, 72 and 96 h. ‘Worst Day’ (24-h period with the highest acid exposure) as well as ‘Average’ (mean 24, 48, 72 and 96-h results) are presented for the wireless study. Normal TR values for the two systems differ.6,11 Therefore, for categorical analysis, we applied: (i) a fixed diagnostic cut-off of 4.2% time pH <4 over the monitoring period for both techniques to allow direct comparisons, and (ii) A diagnostic cut-off of 4.2% (catheter-based studies) and 5.3% (wireless studies) to allow technology specific comparisons of TR.6,11
Symptoms were associated with reflux only if they occurred within 2 min following a reflux event. Symptom Index (SI) is the percentage of symptoms associated with reflux episodes (diagnostic cut-off >50%). Symptom Associated Probability (SAP) is a function which calculates the probability that the association between reflux and symptoms is statistically significant (diagnostic cut-off >95%).12 The ‘Worst Day’ measurement (24-h period with the highest reflux-symptom association), as well as the ‘Average’ rolling cumulative assessment in which all symptoms (SI) or all symptoms and reflux events (SAP) during monitoring period were assessed. The number of positive GERD diagnoses based on reflux-symptom association was compared for wireless pH-studies at 24, 48, 72 and 96-h.
Student’s t-test was used for quantitative variables, and Fisher’s exact test for qualitative variables. Wilcoxon test was used for non-parametric analysis of % time pH <4, SI and SAP. Results are reported as median (inter-quartile range). P < 0.05 was considered statistically significant. Statistical analysis was performed using the statistical software spss 16.0 package (SPSS Inc., Chicago, IL, USA) for Windows.
Consecutive patients (N = 38) with reflux symptoms and no diagnosis of GERD on catheter-based pH-monitoring entered the study. Demographic, clinical and endoscopy data are detailed in Table 1. Typical symptoms (heartburn and acid regurgitation) were reported by all patients; however a proportion reported also atypical symptoms including chest pain and cough (Table 1). Wireless pH-studies were performed successfully in all patients and commenced with the intention of acquiring a 96-h pH-recording and symptom log. No patient complained of severe symptoms (e.g., chest pain) during the study and no capsule needed to be removed.
Table 1. Patient characteristics and demographics
Total patients (N)
Male : Female
13 : 25
Median age (range)
Endoscopic findings during Bravo study
1 (5 cm)
All patients completed at least 48 h of the prolonged, wireless pH-study. Capsule detachment occurred between 48 and 72 h in five patients, and between 72 and 96 h in 12 patients. Thus, 96-h recordings were available in 21 of the 38(55%) patients. No demographic or clinical factors (e.g., symptoms) were associated with early capsule detachment. All patients reported normal activities of daily living during the wireless pH-study. No patient experienced more than mild swallowing difficulty or chest discomfort related to the procedure.
Average esophageal acid exposure was higher for wireless than 24-h catheter-based pH-studies at 24, 48, 72 and 96 h. (P < 0.05) (Table S1 Supporting Information).
Prolonged wireless pH-studies provided an objective GERD diagnosis based on acid exposure (TR >5.3%, time <pH4) in 12 of the 38, 12 of the 38, 11 of the 33 and 10 of the 21 patients at 24, 48, 72 and 96 h, respectively. Using the ‘last result carried forward’ approach, such that results from all patients are considered, the average esophageal acid exposure for wireless pH-monitoring was pathological in 14 of the 38(37%) patients at 96-h (Fig. 1). A similar yield was seen (15 of the 38 patients at 96-h) if the same diagnostic cut-off value (TR >4.2%, time <pH4) was used as with catheter-based pH-studies.
If ‘Worst Day’ analysis was applied pathological esophageal acid exposure was present on at least 1 day in 18 of the 38 (Fig. 1) and 22 of the 38 patients using a diagnostic cut-off of 5.3% and 4.2%, respectively.
Similar results were found if pathological acid exposure in the upright and supine positions were considered separately (Table S2 Supporting Information). Esophageal acid exposure was more variable in the supine than the upright positions as evidenced by the large number (50%) of patients with a pathological SR on at least one test day.
A median 7 (IQR 4…15) symptom events were recorded by each patient on each day. Several patients recorded no symptoms on any individual 24-h period, but all patients reported symptoms during the prolonged study. The most common symptoms documented were heartburn (n = 31; HB), regurgitation (n = 27; R) and chest pain (n = 27; CP). Atypical symptoms recorded included cough, sore throat, nausea and dyspnea (n = 7). Most patients recorded more than one type of symptom (Table 2). The number of individual symptoms (HB, R, and CP) and the total number of symptoms recorded each day remained stable over time (P > 0.1; data not shown).
Table 2. Symptoms reported during prolonged wireless pH-monitoring
IQR, inter-quartile range.
Results presented as median (inter-quartile range).
Total # symptom events over 96 h
Daily median # of symptom events
Median number of symptoms (cumulative)
Proportion of patients reporting specific symptoms (n = 38)
Combination of symptoms reported during wireless pH-studies (initial 48 h
Combining all symptoms recorded during the wireless pH-study, 9 of the 38(24%) patients had a positive SI at the end of the study (Fig. 2 and Table 3). ‘Worst Day’ analysis showed a positive SI in 23 of the 38(61%) patients (Fig. 2). Similar results were observed also for individual symptoms (Table S3 Supporting Information). No statistical difference was shown in SI over time (P > 0.1 for all).
Table 3. Symptom Index and Symptom Association Probability calculated for all symptoms combined
24 h (%)
48 h (%)
72 h (%)
96 h (%)
SI, Symptom Index; SAP, Symptom Association Probability; Bravo, wireless pH-monitoring.
The ‘Average’ (rolling cumulative assessment) and ‘Worst Day’ (pathologic SI or SAP for any symptom rated as positive result) analyses are presented using the ‘last result carried forward’ method.
Values in parenthesis are expressed in percentage.
SI all symptoms (cumulative)
# patients reporting symptoms
SI all symptoms (Worst Day)
# patients reporting symptoms
SAP all symptoms (cumulative)
# patients reporting symptoms
SAP all symptoms (Worst Day)
# patients reporting symptoms
Symptom Association Probability
Combining all symptoms recorded during the wireless pH-study, 13 of the 38(34.2%) patients had a positive SAP (Fig. 2 and Table 3). Worst Day analysis showed a positive SAP in 24 of the 38(63.2%) patients (Fig. 2). Similar results were observed also for individual symptoms (Table S4 Supporting Information). No statistical difference was shown in SAP over time (P > 0.1 for all).
Agreement of reflux-symptom association assessments
As expected, GERD diagnosis based on Worst Day analysis provided a higher diagnostic yield than that based on the rolling cumulative analysis (19 vs 5 patients, respectively). Diagnostic agreement between a positive SAP and SI was only fair (Fig. S1 Supporting Information).
GERD diagnosis based on acid exposure and reflux-symptom association
Whereas 8 of the 38(21%) patients had a GERD diagnosis based on pathologic average esophageal acid exposure as well as positive symptom association (cumulative SAP or SI); 23 of the 38(61%) patients had a GERD diagnosis based on either average esophageal acid exposure or positive symptom association. (Fig. 3A)
Applying the Worst Day analysis, 17 of the 38(45%) patients had a GERD diagnosis based on esophageal acid exposure and positive symptom association on any day, and 29 of the 38(76%) patients had a GERD diagnosis based on either average esophageal acid exposure or positive symptom association on any day. (Fig. 3B)
Treatment decisions were based on a GERD diagnosis from prolonged wireless pH-monitoring using a ‘Worst Day’ analysis for acid exposure or Symptom Association Probability as described above. Outcome at least 6 months after initiation of definitive therapy was available for 33 of the 38(87%) patients.
Medical therapy (e.g., omeprazole 20–40 mg BID) or conservative advice (lifestyle modification and antacids) was pursued by the referring physician in 26 of the 38(68%) patients with outcome assessed at median 24 months (N = 21/26, range 12–36 months). Nine patients had a moderate or good improvement in symptoms and 17 had a poor outcome or were lost to follow-up. Of these 9 of the 9(100%) and 9 of the 17(53%) had a GERD diagnosis on the basis of pathologic acid exposure or positive symptom association, respectively (P = 0.361).
Antireflux surgery was performed in 12 of the 38(32%) with outcome for all assessed at median 24 months [range 6 (2 patients)–36 months]. Of these patients, 9 of the 12 had a positive diagnosis based on pathologic acid exposure, 10 of the 12 based on positive reflux-symptom association (or if both criteria were considered) and two were operated on despite a negative studies. One of these had pathological supine reflux and a large hiatus hernia on endoscopy. The other had ongoing symptoms of non-acid food and fluid regurgitation with normal endoscopy and manometry with a recording of 119 symptoms over 96 h most directly after meals almost certainly due to rumination. At follow-up, 10 patients had a good outcome and 2 patients had persistent reflux symptoms (awaiting repeat pH-studies). Of these 8 of the10 and 2 of the 2 had a positive GERD diagnosis, respectively (P = 1.0). Similarly, 7 of the 10 patients with good outcome and both patients with poor outcome had at least a partial (50%) response to PPI prior to the operation. A further patient with a good initial result had new-onset dysphagia after 6 months that was shown to be due to a trans-diaphragmatic herniation of the wrap (awaiting repair).
This prospective study reports an increased sensitivity and a high diagnostic yield of prolonged, wireless pH-monitoring in patients with reflux symptoms in whom initial 24-h catheter-based pH-studies were negative. Patients diagnosed on the basis of this investigation and referred for definitive, surgical management had good outcomes.
The demographic and clinical characteristics of patients recruited (Table 1) was typical of the larger group referred for pH-studies in our institution.3,13 No important adverse effects occurred and all patients reported normal oral intake and daily activities. At least 48-h of continuous pH-recording was obtained in all studies and 21 of the 38 completed 96-h (median 72-h). More than half of patients (61%) received a positive diagnosis of GERD based on well-validated criteria and a conservative data analysis (average esophageal acid exposure and cumulative reflux-symptom association). The yield increased further (76%) if diagnosis was based on ‘Worst Day’ analysis as is most often applied in the literature6 and in our clinical practice (Fig. 3). The study included 10 patients with mild erosive disease; however this does not obviate the need for investigation as esophagitis LA grade A and endoscopy negative reflux disease have similar clinical and physiologic characteristics, and longitudinal studies show frequent cross-over between these two groups on repeated endoscopy.14 In line with published guidelines it is our policy to proceed with antireflux surgery only with the results of physiologic testing.8,15
The data provides insight into how prolonged wireless pH-studies increase diagnostic yield compared with 24-h catheter-based studies. In this study, 12 of the 38 patients had pathological acid exposure in the first 24 and 48-h after capsule placement and eight of these had pathological acid reflux on at least three of the 4 days. From 48 to 96-h there was a further, modest increase in diagnostic yield based on pH measurement with two additional cases using ‘Average’ and six using ‘Worst Day’ analysis (Fig. 1). For patients with consistently positive results on prolonged wireless pH-studies the negative finding on 24-h catheter studies was likely due to discomfort and altered behavior which are known to reduce reflux events and symptom-assocaition.1,3,16 The remainder represents the ‘true increase’ of diagnostic yield on prolonged pH-monitoring. The latter effect was marginal when diagnosis was based on average esophageal acid exposure as this measurement is stable over time; however there was a steady increase in yield when diagnosis is based on Worst Day analysis due to day-to-day variation in esophageal acid exposure (Fig. 1).
Heartburn was the most common symptom recorded. The majority of patients complained of at least two different symptoms per day with a median of 7 (IQR 4…15) on any one day and 30 (14…59) symptom events over the entire study (Table 2). Different patients report different symptoms in response to reflux events; therefore, to increase power, typical symptoms such as heartburn, regurgitation and chest pain were combined to assess the overall reflux-symptom association. Twenty-four percent and 34% patients had a positive SI and SAP at the end of the study respectively using average cumulative measurements (Fig. 2 and Table 3). Results for individual symptoms were similar (Tables S3 and S4 Supporting Information).
No patient remained symptom free during the prolonged monitoring period. A diagnosis of GERD based on positive SI or SAP was present in 45% patients using the Average analysis and in 74% patients using Worst Day analysis (Fig. S1 Supporting Information). Agreement between diagnoses based on SI and SAP were far from perfect and it remains uncertain which method provides the best and most clinically relevant summary of this data;17 however, regardless of which method is used, the large data set from prolonged monitoring greatly facilitates a statistically robust assessment of reflux-symptom association.18
The optimal duration of monitoring and the most appropriate analysis of prolonged pH-studies have not been determined;4 however this study shows that any increase in diagnostic yield from 4-day studies depends on whether data analysis is based on the ‘Average’ or a ‘Worst Day’ approach. In general, for both acid exposure and reflux-symptom association, results based on ‘Average’ or ‘cumulative’ measurements are statistically sound and stable over time; however ‘Worst Day’ analysis might be more appropriate in patients with intermittent symptoms as it maximizes diagnostic sensitivity and yield.4 Notwithstanding this possible advantage, there must be a limit to the ‘Worst Day’ approach. For example, it would be wrong to select one abnormal day out of 20 days monitoring as this would be expected to occur in a healthy person by chance alone. A simple approach that can be applied in clinical practice is to make a decision at 48 h whether to prolong the Bravo study. Detailed analysis suggests that if the first 2 days provide consistent results (i.e., both positive or both negative) then it is unlikely that the subsequent 2 days will alter the diagnosis and further prolongation is not required. Conversely, if the results of the first 2 days are inconsistent prolonging the study provides more data on which to base a definitive diagnosis. Ongoing studies will examine these important issues in greater detail using the association of results with clinical outcome to determine the best approach.
The clinical utility of any investigation and diagnosis is based on its ability to guide effective management. Pathologic esophageal acid exposure or reflux-symptom association on pH testing predicts a positive outcome after antireflux surgery 19–21 and guidelines recommend against surgical management in the absence of a definitive GERD diagnosis.8 Although prolonged wireless pH-studies increase test sensitivity, specificity may decrease with repeated testing,22 especially if Worst Day analysis is the chosen method of interpretation. However, a false negative result on pH testing can also have important consequences, and it may be more prudent to ‘rule out’ than ‘rule in’ a diagnosis of GERD in this patient group.
Consistent with this statement, this study reports excellent control of reflux symptoms following fundoplication in 10 of the 12 patients with a negative catheter-based pH-study. In contrast, only 9 of the 26(34%) patients reported a good response to optimal conservative therapy (all had GERD); whereas 17(66%) had a poor response (nine had GERD). Three patients with objective evidence of GERD await surgery and one with evidence also of delayed gastric emptying is being treated with prokinetics. The remaining five patients with poor response to PPI had a GERD diagnosis based on positive reflux-symptom association to only a few reflux events. While such individuals may profit from surgery,20,21 the primary problem may be visceral hypersensitivity rather than reflux per se18 and, similar to the patients without GERD, treatment with low-dose tricyclic antidepressants directed at reducing chronic visceral pain was preferred.
This study has limitations. This is a case series in a highly selected group of patients referred by physicians and surgeons due to the clinical suspicion of a ‘false-negative’ diagnosis on initial catheter-based pH-study. Thus, the findings cannot be generalized to all patients with negative catheter-based pH-studies many of whom have atypical reflux symptoms that are much less likely to be caused by reflux events.23 Furthermore, the catheter-based pH-study performed in our institution does not routinely include impedance. Inclusion of this modality increases test sensitivity ‘off PPI treatment’ by 10–20%.23 Although this could have reduced the pool of patients entering the study, the principle limitations of 24-h catheter-based studies remain and it is unlikely that this technology would remove the need for prolonged wireless pH-studies in some patients. More critically, this study did not have a control group. Inclusion of patients with a positive catheter-based 24-h pH-study would have provided information about the concordance of GERD diagnosis between methods. Thus, the results are biased ‘in favor’ of prolonged wireless pH-monitoring in that the diagnosis could only change from negative to positive for GERD and not vice versa. As a true ‘gold standard’ does not exist, outcome data is critical. The high proportion of patients with a negative catheter-based test that then had a positive GERD diagnosis on prolonged wireless testing and a good outcome post-fundoplication surgery provides an independent validation of the clinical utility of this method. Moreover, study procedures reflect ‘real life’ practice in which most patients are referred for pH-studies only if reflux symptoms fail to respond to acid suppressant treatment and/or to confirm the diagnosis prior to antireflux surgery.
In conclusion, prolonged wireless pH-monitoring had an important clinical impact on diagnosis and management of patients with typical reflux symptoms but negative catheter-based studies. In most centers at least one in three patients that undergo catheter-based pH-studies have a negative test result.13 This study suggests that if a false-negative diagnosis is suspected, and especially if there is a suggestion of modified behavior due to catheter discomfort, then prolonged wireless pH-studies should be considered. Indeed, for patients with intermittent symptoms and those troubled by nasal catheterization, prolonged wireless pH-studies are likely to be the optimal investigation for GERD diagnosis.
Rami Sweis: acquisition of data, data analysis and interpretation, statistical analysis, drafting of the manuscript; Mark Fox: study concept, study supervision, data analysis and interpretation, critical revision of the manuscript for important intellectual content; Angela Anggiansah: study concept, administrative and material support, acquisition of data; Terry Wong: clinical lead, administrative and material support, critical revision of the manuscript for important intellectual content.
We are grateful to Dr Maxwell Asante, Consultant Gastroenterologist at Princess Royal University Hospital, Orpington, UK, for his help with funding research at the esophageal lab for 1 year through the Maxwell Asante Foundation; Dr Jeremy Sanderson, Consultant Gastroenterologist at Guy’s and St Thomas’ Hospitals, London, UK and Senior Clinical Research Fellow at King’s College London for his support and advice; and the Manoukian Foundation for its support of the St Thomas’ Hospital Oesophageal Laboratory.