Article Text
Abstract
Objective To validate Lyon Consensus criteria for diagnosing gastro-oesophageal reflux disease (GORD) by reflux monitoring.
Design Manual review of impedance-pH tracings from patients with proton pump inhibitor (PPI)-dependent heartburn, evaluated off PPI. Acid exposure time (AET) thresholds defined by the Lyon Consensus and impedance parameters were investigated, namely, total refluxes (TRs), postreflux swallow-induced peristaltic wave (PSPW) index and mean nocturnal baseline impedance (MNBI).
Results The study included 488 patients, 178 (36%) with normal (<4%) AET, 89 (18%) with inconclusive (4%–6%) AET and 221 (45%) with abnormal (>6%) AET, alongside with 70 healthy controls. At receiver operating characteristic analysis, area under curve was 0.89, 0.95 and 0.89 for TRs, PSPW index and MNBI, respectively, and threshold values were 40, 50% and 2000 Ω; the 4% physiological AET threshold defined by the Lyon Consensus showed 100% specificity but 63% sensitivity. The thresholds defined for impedance parameters were validated against AET by means of ordered logistic regression, being in concordance with the 4% AET threshold (OR 2.5 for TRs, 18.9 for PSPW index and 5.7 for MNBI). TRs positivity and concordant PSPW index/MNBI positivity were found in 80%–90% of patients in the abnormal AET group, in 73%–74% of cases in the inconclusive AET group and in 28%–40% of cases in the group with normal AET.
Conclusions Our results show the overall validity of the Lyon Consensus approach to GORD diagnosis. Adding evaluation of impedance parameters, namely, TRs, PSPW index and MNBI to AET appraisal, substantially improves the diagnostic yield of reflux monitoring.
- acid-related diseases
- oesophageal ph monitoring
- gastroesophageal reflux disease
Data availability statement
Data supporting the study are available on reasonable request from the corresponding author.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ABOUT THIS SUBJECT?
Reflux monitoring has been regarded as the test of choice to document gastro-oesophageal reflux disease (GORD) in patients with endoscopy-negative heartburn.
A hierarchical approach to reflux monitoring has been proposed based on acid exposure time (AET) and supportive impedance parameters to confirm or exclude GORD.
WHAT ARE THE NEW FINDINGS?
Impedance parameters, namely total refluxes (TRs), post-reflux swallow-induced peristaltic wave (PSPW) index and mean nocturnal baseline impedance (MNBI) showed high efficiency in distinguishing patients with proton pump inhibitor-dependent heartburn from healthy controls.
In the vast majority of patients with inconsistent AET values, GORD diagnosis was confirmed by impedance parameters.
In approximately one-third of patients with normal AET values, GORD diagnosis was confirmed by impedance parameters.
Impedance parameters substantially increase the diagnostic yield of reflux monitoring, showing the high clinical value of impedance-pH monitoring for GORD diagnosis.
HOW MIGHT IT IMPACT ON CLINICAL PRACTICE IN THE FORESEEABLE FUTURE?
Analysis of TRs, PSPW index and MNBI, hopefully aided by refinements of commercial software for automated calculation, can increase the yield of impedance-pH monitoring in the diagnostic workup of GORD.
Introduction
Gastro-oesophageal reflux disease (GORD) develops when the reflux of gastric contents into the oesophagus leads to troublesome symptoms and/or complications.1 2 Heartburn is the most sensitive and specific GORD symptom with an estimated prevalence of 15%–20% in the general population. Proton pump inhibitor (PPI) therapy represents the mainstay of medical treatment, relieving heartburn in most cases, so that, in the absence of alarm symptoms, heartburn relief by a 4-week PPI trial is regarded as sufficient to confirm GORD diagnosis in clinical practice.1 2
Heartburn quickly recurs after PPI withdrawal, however, and many patients become PPI dependent. Side effects by long-term use of PPI therapy have been claimed in recent years, and strategies for discontinuing long-term PPI therapy have been suggested.3 Thus, many patients are referred to tertiary care centres for objective documentation of GORD in order to justify chronic administration, and several patients ask for surgical GORD management to get rid of PPI dependence despite confirmed evidence of the overall safety of PPIs.4 5
Objective GORD diagnosis is challenging since a definite gold standard has not yet been recognised. In the PPI era, endoscopy reveals erosive reflux oesophagitis in a minority of patients referred for heartburn, then direct reflux monitoring is required in the majority of cases6 to document abnormal reflux burden. Acid exposure time (AET) has long been regarded as the most useful parameter for this purpose, despite detection of normal values in up to 20% of patients with reflux oesophagitis.7 8 Currently, reflux monitoring can be performed with pH monitoring or impedance-pH monitoring, the latter allowing detection of all reflux episodes independent of their acidity.9 In recent years, new impedance parameters have been introduced, namely, postreflux swallow-induced peristaltic wave (PSPW) index,10 representing oesophageal chemical clearance and mean nocturnal baseline impedance (MNBI),11 representing oesophageal mucosal integrity, both reportedly increasing the diagnostic yield of impedance-pH monitoring.12 Following expert meetings, reference threshold AET values have been agreed on13 and recommended by the Lyon Consensus14 : AET >6% and <4% were regarded as definitely abnormal and normal, respectively, while values between 4% and 6% were considered as inconclusive and necessitating supportive evidence by adjunctive parameters, including number of total refluxes (TRs), symptom association probability (SAP) and symptom index (SI) as well as the new metrics PSPW index and MNBI.
In the present study, we report the results of the application of the Lyon Consensus criteria in a large cohort of patients prospectively evaluated at referral centres for PPI-dependent heartburn. Our aim was to define the clinical value of conventional and new impedance parameters as compared with the AET threshold values proposed by the Lyon Consensus.
Methods
Data prospectively collected at Italian referral centres between January 2016 and January 2021 were reviewed. The study was designed and performed in accordance with the Declaration of Helsinki. Since patients were investigated for clinical reasons and were not exposed to any additional intervention for research purposes, according to Italian law, formal medical ethical assessment was not required. A signed informed consent was obtained before any clinical investigation. Heartburn was scored using a validated four-grade Likert-type scale, administered by a senior investigator.15 Briefly, heartburn was scored 0 when absent, 1 when mild/occasional, 2 when moderate/frequent and 3 when severe/constant and was considered troublesome for scores 2 or 3. Previous oesophagogastric surgery, psychiatric disorders, Sjogren syndrome, scleroderma as well as Barrett’s oesophagus and reflux oesophagitis (grades B-C-D Los Angeles score) at upper GI endoscopy carried out within the 6 months period preceding reflux monitoring constituted exclusion criteria. Patients were referred for possible antireflux surgery or for confirmation of the appropriateness of long-term PPI treatment. They were considered to have PPI-dependent heartburn if reporting a history of troublesome heartburn (score 2–3) suppressed (score 0–1) by 4-week standard-dosage PPI therapy, early recurring after PPI wash out and again early suppressed by 4-week standard-dosage PPI. Impedance-pH monitoring was carried out after 2-week PPI withdrawal and was always preceded by conventional or high-resolution oesophageal manometry to exclude major motor disorders.16
Impedance-pH probes allowed detection of intraluminal impedance at 3, 5, 7, 9, 15 and 17 cm and of pH at 5 cm above the upper border of the manometrically defined lower oesophageal sphincter. Tracings were manually reviewed by senior investigators, aided by the BioView software (Diversatek/Sandhill Scientific, Colorado, USA) and following rigorous rules to identify TRs and PSPWs.17 According to the Lyon Consensus, AET was regarded as definitely abnormal for values >6%, definitely normal for values <4%, and inconsistent for values between 4% and 6%; positive heartburn–reflux association was defined for combined positivity of SAP and SI, that is >95% and >50%, respectively.13 14 A PSPW was defined as an antegrade 50% drop in impedance originating in the proximal oesophagus within 30 s after the end of a reflux event and reaching the distal lumen17; dividing the number of PSPWs by the number of TRs, the PSPW index was obtained.10 MNBI was assessed in the nighttime recumbent period from the most distal impedance channel; three 10 min time intervals (around 1.00 am, 2.00 am and 3.00 am) were selected avoiding swallows, refluxes and pH drops, and the mean was calculated.11 Impedance-pH parameters of patients with PPI-dependent heartburn were categorised according to AET threshold values as proposed by the Lyon Consensus and compared with those of 70 healthy controls evaluated in previous studies.10–12 18
Statistics
Continuous and categorical variables were compared with ANOVA and Χ2 test, respectively. Receiver operating characteristic (ROC) analysis with calculation of the area under the curve (AUC) was used to assess the diagnostic yield of conventional and novel impedance-pH parameters comparing PPI-dependent patients with healthy controls; thresholds of TRs, PSPW index and MNBI were established favouring specificity over sensitivity. Ordered logistic regression analysis was then applied to validate the thresholds of TRs, PSPW index and MNBI against the diagnostic categories as defined by the AET thresholds: for each impedance-pH variable, OR was computed alongside with 95% CI, representing the probability of upgrading from a lower to a higher diagnostic category. STATA statistical software, release V.16 was used. Significance was set at p<0.05, adjusted with Bonferroni correction for multiple comparisons.
Results
The main baseline characteristics of the 488 patients with PPI-dependent heartburn included in the study are reported in table 1 alongside with those of 70 healthy controls. Heartburn was not recorded during impedance-pH testing by 95 of 488 (19%) cases; SAP and SI were both positive in 213 (44%) cases. At ROC analysis, the AUC of AET was 0.92 (95% CI 0.90 to 0.95), quite similar to those of TRs, PSPW index and MNBI (figure 1). The threshold values resulting from maximising the sum of sensitivity and specificity and favouring the latter were 40 for TRs, 50% for PSPW index and 2000Ω for MNBI.
The AET threshold values proposed by the Lyon Consensus14 were used to categorise the 488 PPI-dependent heartburn cases into normal AET (<4%) (178 cases, 36%), inconclusive AET (4%–6%) (89 cases, 18%) and abnormal AET (>6%) (221 cases, 45%) group: the mean values of TRs, PSPW index and MNBI in the three groups are reported in table 2. AET was <4% in all healthy controls; at ROC analysis, the 4% AET physiologic threshold proposed by the Lyon Consensus had a 100% specificity and a 63% sensitivity.
Results of ordered logistic regression are reported in table 3. The impedance parameters, namely, TRs, PSPW index and MNBI were validated against AET: higher values of such parameters were significantly associated to higher values of AET, in concordance with the thresholds defined by the Lyon consensus, the highest OR being displayed by PSPW index.
Table 4 shows positivity of impedance parameters in the 488 PPI-dependent heartburn cases as subdivided according to normal, inconclusive and abnormal AET. Concordant positivity of SAP and SI was quite similar comparing the three groups. Abnormal values of TRs, PSPW index and MNBI were significantly more frequent in the inconclusive and abnormal AET group as compared with the normal AET group. Concordant positivity of PSPW index and MNBI was significantly higher in the abnormal and inconsistent than in the normal AET group (p<0.05) (figure 2).
Table 5 reports positivity of impedance parameters in the 178 PPI-dependent heartburn cases with normal AET, subdivided according to the 2.8% threshold AET value recently proposed for the Diversatek/Sandhill system.19 Concordant positivity of SAP and SI did not differ between the two groups, whereas positivity of TRs, PSPW index and MNBI was more often detected in patients with AET >2.8%.
Discussion
In this large multicentre cohort of patients with PPI-dependent heartburn, we defined threshold values of TRs, PSPW index and MNBI through ROC analysis. All three metrics efficiently separated patients with normal AET from those with inconclusive and abnormal AET as defined by the Lyon Consensus,14 and all the three metrics were validated against AET by ordered logistic regression.
In the present series, concordant SAP/SI positivity was detected in less than half of our patients and did not differ among the various AET groups, indicating the lack of relationship with acid reflux burden. SAP and SI are considered complementary, and concordant positivity is regarded to show the best link between symptoms and reflux.13 14 However, SAP and SI are overly dependent on patient’s recording accuracy, and symptoms are not recorded during reflux monitoring in many cases: then, assessment of more objective and applicable parameters is warranted for a firm diagnosis of GORD.
The diagnostic value of TRs has not yet been thoroughly investigated in patients complaining of heartburn. On PPI, TRs positivity was detected in more than half of patients with PPI-refractory heartburn who underwent successful antireflux surgery.15 Recently, it has been suggested that TRs may represent a predictor of GORD treatment outcome.20 Notably, the threshold value of TRs (40), we defined by means of ROC analysis, is the same proposed as the physiological threshold by the Lyon Consensus.14 In the present study, TRs positivity was detected in much more cases in the inconsistent and abnormal AET than in the normal AET group, showing that this parameter can be useful in the diagnostic work-up for patients with the typical GORD syndrome.
The high diagnostic yield of combined assessment of PSPW index and MNBI has been shown by several studies.11 12 15 18 21–27 PSPW index represents oesophageal chemical clearance, a major defence mechanism against reflux elicited by the oesophago-salivary reflex,28 in turn directly related to the harmfulness of refluxate,29 while MNBI represents mucosal integrity, lower values of both parameters reflecting increasing GORD severity.12 PSPW index does not improve significantly after effective antireflux surgery, in turn confirmed by subtotal abolition of TRs, whereas normalisation of mucosal integrity as measured with MNBI occurs in nearly all cases15: these findings suggest that impairment of chemical clearance is clinically relevant when associated with detectable mucosal damage documented by low MNBI values and confirm that concordant positivity of both the two metrics is more consistent for diagnostic purposes.12 18 22 24 25 The diagnostic efficiency of PSPW index and MNBI was confirmed in the present series, with concordant positivity detected in much more cases in the inconclusive and in the abnormal than in the normal AET group.
Nearly half of our patients had AET >6%, the threshold defined by the Lyon Consensus as conclusive evidence of pathologic reflux14 ; TRs and concordant PSPW index/MNBI positivity was found in 80% and 90% of these cases, respectively, confirming the validity of the >6% AET threshold, which can then stand alone for confirming GORD diagnosis in patients with endoscopy-negative heartburn.
The inconsistent AET group comprised less than one-fifth of our patients. GORD diagnosis was confirmed by TRs positivity and concordant PSPW index/MNBI positivity in three-fourth of cases, showing the high usefulness of these impedance metrics in such cases. Notably, mean PSPW index and MNBI were significantly lower in the abnormal AET group as compared with the inconsistent AET group, whereas mean number of TRs was similar. Likewise, the rate of concordant positivity of PSPW index and MNBI was significantly higher in the abnormal AET group as compared with the inconsistent AET group, whereas the rate of positivity of TRs was similar. These results indicate that reflux burden is comparable in patients with abnormal and inconsistent AET, while the difference lies in the response to reflux, that is, chemical clearance as measured with the PSPW index, and its consequences, that is, acid exposure time as measured with AET and mucosal integrity as measured with MNBI.
More than one-third of our patients had AET <4%, to be considered as evidence against pathologic reflux according to the Lyon Consensus.14 In a recent study on healthy subjects from around the world,19 the 95th percentile of AET threshold detected with the Diversatek/Sandhill system, the same used in our study, was 2.8%, considerably lower than 4%. In our series of 488 PPI-dependent heartburn cases and 70 healthy controls, the AET 4% threshold showed a 100% specificity but a 63% sensitivity. Indeed, limitations in diagnostic sensitivity of various AET thresholds have long been recognised,7 8 in part due to day-to-day variability of acid reflux burden as shown by prolonged wireless pH monitoring.30 We found TRs positivity and concordant PSPW index/MNBI positivity in approximately one-third of our patients with AET <4% and in approximately half of cases with AET between 2.9% and 4%, showing that different AET cut-offs directly affect positivity rate of impedance parameters. This confirms the soundness of the impedance parameters thresholds we detected with ROC analysis and further highlights the diagnostic yield of impedance parameters for confirming GORD diagnosis in patients with PPI-dependent heartburn and variably defined AET thresholds.
Reflux monitoring has been recommended off PPI when GORD is unproven, and the choice between catheter-based/wireless pH-monitoring and pH-impedance monitoring has been suggested on the basis of cost and availability.14 The Lyon Consensus proposed a hierarchical approach to GORD diagnosis based on defined AET thresholds, while SAP/SI positivity, abnormal number of TRs and low values of PSPW index and MNBI were considered as supportive evidence for GORD diagnosis in case of inconclusive AET values.14 Actually, among the parameters supporting inconclusive AET, only SAP and SI for acid refluxes can be assessed with catheter-based/wireless pH-monitoring while TRs, PSPW index and MNBI can be assessed only with impedance-pH monitoring. Therefore, only impedance-pH monitoring provides a comprehensive evaluation of total reflux burden (TRs), of the reaction against reflux (PSPW index) and of the consequences of reflux (MNBI). Since our results show the high diagnostic yield of these parameters, it can be inferred that impedance-pH monitoring is the test of choice for reflux monitoring in order to diagnose GORD.
A few limitations of our study should be acknowledged. Threshold values for TRs, PSPW index and MNBI found in the present series are at minor variance with those previously suggested,12 a result probably due to the smaller size of that series and to the subsequent development of consensus criteria about definition of reflux episodes,17 in turn directly influencing calculation of PSPW index. Refinements of software automated analysis, possibly aided by artificial intelligence, are needed to simplify analysis and interpretation of impedance-pH monitoring. A potential bias of our study relates to retrospective review of patient records; however, data were prospectively collected at tertiary care centres adopting standardised protocols, and the patient population was decidedly homogeneous, being constituted by PPI-dependent heartburn cases. We focused on heartburn since it is the cardinal symptom of the typical GORD syndrome,1 often becoming dependent on long-term PPI treatment: it remains to be established the applicability of our findings in patients with non-cardiac chest pain and extraoesophageal reflux symptoms as well as the usefulness of impedance metrics in heartburn patients failing PPI therapy.
In conclusion, our results show the overall validity of a hierarchical approach as suggested by the Lyon Consensus for GORD diagnosis, based on AET and impedance parameters when AET does not afford conclusive evidence of pathologic reflux. Indeed, the diagnostic yield of reflux monitoring is substantially improved by assessment of impedance parameters, namely, TRs, PSPW index and MNBI, indicating that impedance-pH monitoring represents the most informative and useful reflux-monitoring test for GORD diagnosis.
Data availability statement
Data supporting the study are available on reasonable request from the corresponding author.
Ethics statements
Patient consent for publication
Ethics approval
Patients were investigated for clinical reasons and were not exposed to any additional intervention for the study purposes. According to Italian law, formal medical ethical assessment was not required.
Footnotes
Contributors LF: study concept and design; analysis and interpretation of data; drafting of the manuscript. MF: study concept and design; collection, analysis and interpretation of data; drafting of the manuscript. NDB, MR, ES: study concept; collection, analysis and interpretation of data; critical revision of the manuscript. ST, SR, RLC: collection, analysis and interpretation of data; critical revision of the manuscript. RP, LF, RMZ: study concept; analysis and interpretation of data; critical revision of the manuscript.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.
Provenance and peer review Not commissioned; externally peer reviewed.