In sebastiz/PhysiMineR: Functions to extract, clean, manipulate and analyse upper gastrointestinal physiological parameters
title : "Extended Wireless pH monitoring significantly increases GORD diagnoses in patients with a normal pH impedance study."
shorttitle : "Extended Wireless pH monitoring significantly increases GORD diagnoses in patients with a normal pH impedance study."
author:
- name : "Sebastian S Zeki"
affiliation : "1,3"
corresponding : yes
address : "Gastroenterology Department,Westminster Bridge Road, London SE1 7EH"
email : "sebastiz@hotmail.com"
- name : "Ismail Miah"
affiliation : "1"
- name : "Anna Wolak"
affiliation : "1"
- name : "Minerva daSilva"
affiliation : "1"
- name : "Jason Dunn"
affiliation : "1,2"
- name : "Andrew Davies"
affiliation : "1,2"
- name : "James Gossage"
affiliation : "1,2"
- name : "Abrie Botha"
affiliation : "1,2"
- name : "Guiping Sui"
affiliation : "1"
- name : "Jafar Jafari"
affiliation : "1"
- name : "Terry Wong"
affiliation : "1,2"
affiliation:
- id : "1"
institution : "Centre for Oeosphageal Diseases, Guy's and St. Thomas Hospital, Westminster Bridge Road, London"
- id : "2"
institution : "Karolinska Instituet, Karolinska,Sweden"
- id : "3"
institution : "Bart's Cancer Institute,Charterhouse Square, London"
authornote: |
header-includes:
- \usepackage{setspace}\doublespacing
- \AtBeginEnvironment{tabular}{\singlespacing}
- \AtBeginEnvironment{lltable}{\singlespacing}
- \AtBeginEnvironment{tablenotes}{\doublespacing}
- \captionsetup[table]{font={stretch=1.5}}
- \captionsetup[figure]{font={stretch=1.5}}
abstract: |
Introduction: In cases where a false negative diagnosis of gastro-oesophageal reflux disease (GORD) is suspected, using catheter-based methods such as pH impedance and high resolution manometry (HRM), further testing can be done using a wireless pH monitoring (WPM) device which can assess acidic reflux for 96 hours. The increased yield of GORD-positive diagnoses with WPM is unknown. Further testing is at the clinician’s discretion with no guidance as to which parameters on pH impedance or manometry, when this test is negative, may predict a subsequent positive WPM test. Aim: a) To determine the increased yield and b) determine predictors of GORD positivity in patients with negative pH impedance. Method: The increased GORD-positive diagnostic yield (average day analysis) by with WPM was determined in --NUMBER-- consecutive patients who had also undergone negative pH impedance with HRM. Univariate and multivariate analysis was performed to determine predictors of a positive WPM. Results: Of the 212 patients who underwent WPM after a negative pH impedance study, 56 (29%) were found to have a positive result for GORD. Only the AET at pH impedance and endoscopically visible oesophagitis was significantly associated with a WPM-GORD positive result on multivariate analysis. However the AET was insufficiently accurate to be used as a predictive clinical parameter (AUC: 0.61). --(AET NUMBERS)-- OESOPHAGITIS AND PVALUE NUMBERS. Conclusion: There is a significant increased yield for GORD positive diagnoses with WPM although predicting this from pH impedance and manometry if pH impedance is negative is difficult. --PHRASE MORE EFFICIENTLY-- The increased yield suggests that a more formal study to assess the difference in GORD detection between WPM and impedance may be useful. --PHRASE MORE EFFICIENTLY--
<!-- https://tinyurl.com/ybremelq -->
keywords : "Gastro-oesophageal reflux disease, pH studies, wireless pH monitoring"
wordcount : This document contains r wordcountaddin::word_count() words
bibliography : ["r-references.bib"]
floatsintext : yes
figurelist : no
tablelist : no
footnotelist : no
linenumbers : yes
mask : no
draft : no
documentclass : "apa6"
classoption : "man"
output : papaja::apa6_docx
knitr::opts_chunk$set(cache.extra = knitr::rand_seed, echo = FALSE, comment = FALSE, warning = FALSE, message = FALSE)
set.seed(42)
options(digits=2,xtable.comment = FALSE)
library("papaja")
# Seed for random number generation
library(knitr)
library(here)
library(EndoMineR)
library(PhysiMineR)
library(dplyr)
library(tidyr)
library(kableExtra)
library(ggpubr)
library(readxl)
library(gtsummary)
read_chunk(here("inst","Projects","BRAVOStudies", "NegImpPredictorsOfAllPosBRAVO","munge", "PreProcessing.R"))
Introduction
Wireless pH monitoring (WPM) is a useful method for the prolonged analysis of oesophageal pH. Our previous studies have shown that WPM is a useful subsequent investigation or as an alternative to catheter based studies. REF RAMI's STUDY. A particular advantage of WPM is the accomodation of day to day variation and improved ablility for patients to perform their normal activities of daily living. (REF WONG) False negative catheter studies may occur if reflux is intermittent so that a 24 hour window of assessment is insufficient [@TSENG2005a]. It has been estimated that a significant proportion of patients would have a positive finding of GORD if assessed with extended pH studies rather than a 24 hour pH assessment [@Scarpulla2007a; @Sweis2009] . The increased yield of WPM over standard catheter-based pH impedance specifically however, has never been determined.
When deciding on the need for further pH testing given a negative 24 hour study, a clinican may take into account a number of factors including the nature and frequency of the symptoms, or other tests such as the endoscopy or manometry which may provide circumstantial evidence of GORD [@Gyawali2018]. There is however no guidance as to who should undergo prolonged acid monitoring in the presence of a negative pH impedance study.
The aim of the current study is to establish the diagnostic yield by average day analysis and potential predictors of a wireless pH study demonstrating GORD on a worst day analysis given a negative 24 pH impedance study, using the pH impedance, manometry and endoscopy results.
Methods
Subjects
The results of all consecutive patients who had undergone a 24 hour pH impedance in patients with suspected GORD and a subsequent WPM between r paste(min(lubridate::month(ImpAndBravoWithHRM$MainPtDataDateofAdmission,label=TRUE, abbr=FALSE)),min(lubridate::year(ImpAndBravoWithHRM$MainPtDataDateofAdmission))) and r paste(max(lubridate::month(ImpAndBravoWithHRM$MainPtDataDateofAdmission,label=TRUE, abbr=FALSE)),max(lubridate::year(ImpAndBravoWithHRM$MainPtDataDateofAdmission))) were retrospectively interrogated. Only patients who had undergone an initially negative 24 hour pH impedance result were included. A negative pH impedance test was defined as having more than 40 reflux episodes and an AET for <= 4.2% of a 24 hour period [@Gyawali2018]. Patients with a normal AET but with >80 reflux episodes were also excluded on the basis that they have non-acid reflux. A WPM study was defined as positive if the patient had an AET for >=5.3% of any 24 hour period (worst day analysis-[@Pandolfino2003] and [@Wenner2005]). A worst day analysis was chosen as it is more specific than average day analysis [@Sweis2011].
All patients selected were adults over the age of 18. Ethics was approved (IRAS 18/NW/0120). Patients were excluded for the following reasons: 1. Previous oesophageal surgery or intervention such as endoscopic mucosal resection or radiofrequency ablation. 2. Less than 24 hours of recording on pH impedance. 3. Less than 48 hours of recorded data from the patient's WPM study.
High resolution manometry protocol
Following local analgesia of the nares the catheter was introduced trans-nasally and the patient was instructed to drink water through a straw whilst the HRM catheter was advanced to the stomach. The high resolution manometry (HRM) catheter depth was adjusted to ensure manometric visual of the upper oesophageal sphincter (UOS), the gastro-oesophageal junction (GOJ) and gastric pressures. 10 single swallows of 5ml were performed with each being 20 seconds apart. Each 5ml water swallow was then assessed in accordance to Chicago classification (version 3)[@Kahrilas2015] using Manoview software (version 3) (Sierra Scientific Instruments).
pH impedance protocol:
Patients underwent reflux monitoring using Sandhill Scientific multichannel impedance pH catheters (ZAN-BG-44) which were inserted trans-nasally after applying local anaesthesia (xylocaine). The dual pH sensors of the catheter were positioned 5cm below and above the manometric LOS. The impedance sensors were positioned above the LOS by 3cm, 5cm, 9cm, 15cm and 19cm. Acid reflux was considered when a retrograde impedance flow was observed and the oesophageal pH sensor was detecting a pH value of <4. The data was captured by ZepHrTM recording device and data was analysed using the BioVIEW Analysis software (5.7.1.0).
96 hour WPM protocol:
The 96 hour WPM procedure was performed after a 6 hour fasting period and discontinuation of acid suppressant The patient was asked to complete a diary of symptoms and meal/drink times (except for still water), using the clock on the WPM receiver, during the 96-hour pH recording. 3 symptoms of reflux were decided by the patient and symbols allocated for each symptom. The patient was requested to record these symptoms by pressing the appropriate symptom markers (symbols) on the receiver, when the symptom occurred.
After the WPM capsule was calibrated, the patient was sedated according to local practice. Following a complete endoscopic examination of the upper GI tract a pH sensor located in a capsule (WPM pH capsule: 25mm x 6mm x 5.5 mm) was pinned temporarily to the wall of the Oesophagus 6 cm proximal to the Z line.
The following parameters were obtained from the analysis and compared against normal reference values 1. Number of reflux episodes per day, 2. Total percent of time spent in reflux, 3. Percent of time spent in reflux in the upright position, 4. Percent of time spent in reflux in the supine position, 5. Percent of time spent in reflux post prandial, 6. De Meester Score
Oesophagitis
Endoscopy records were retrospectively consulted for patients who had been submitted for a BRAVO with a negative impedance catheter study. The degree of oesophagitis recorded by the endoscopist at the time of BRAVO insertion was recorded for the patient. All endoscopists inserting BRAVO catheters used the Los Angeles oesophagitis grading system. Los Angeles grade A to D was included as classifying the patient as having oesophagitis.
Statistical analyses:
WPM, HRM and pH impedance results were ordered according to the patient identifier and date of the study. To ensure that there was no duplicated data for patients who had undergone more than two of any test, only tests chronologically closest to the test to be merged were chosen. To ensure relevance, tests were merged only if there was less than one year's difference between the two. Variables for analysis were chosen from impedance and HRM data on the following basis: Lack of multicolinearity and an adequate number of data points to remove skew from missing data. All quantitative data were presented as mean±SD. Fisher's exact test or Wilcoxon rank sum test was used to compare non-parametric numerical data. Student's t-test was used to compare parametric numerical data. Categorical data was compared using a chi-squared test of independence. The following variables were chosen for analysis: Age (years), basal respiratory minimum (mmHg),mean residual pressure (mmHg),distal contractile integral (mean mmHg/cm/s), contractile front velocity (m/s),distal latency, percentage failed swallows, percentage panoesophageal pressurization,percentage large breaks, number of small breaks, number of acid episodes, acid exposure time, mean acid clearance time, the longest acid episode and the presence of oesophagitis at endoscopy. The symptom association probability (SAP) at pH impedance was also used. Analysis was performed on individual SAP symptoms or symptoms grouped as typical (heartburn,regurgitation,chest pain), atypical (belching, cough, throat symptoms, abdominal pain, vomiting, nausea) or both (mixed).
Statistical comparison was carried out on all recorded parameters and those with p<0.05 were selected for evaluation in a multivariate model with a stepwise (forward selection/backward elimination) method. p<0.05 was taken as the threshold of significance for the multivariate model and the strength of association was expressed as odds ratio (OR) with 95% confidence interval (CI).
Results
Baseline demographic data
Of the r nrow(NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal2) included, r nrow(NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal2[NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal2$AcidRefluxBRAVOAv=="Positive",]) studies were positive for GORD (Male:Female r resOR$Gender$descriptive[3,2]: r resOR$Gender$descriptive[3,1], average age of r inline_text(so, variable = "ageInYears", column = "Positive") years and
r nrow(NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal2[NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal2$AcidRefluxBRAVOAv=="Negative",]) negative for GORD (Male:Female r resOR$Gender$descriptive[2,2]: r resOR$Gender$descriptive[2,1] average age of r inline_text(so, variable = "ageInYears", column = "Negative") years.
Using a worst day analysis, r nrow(NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal[NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal$WorstDayAnalysisGORDPositive==1,]) were positive (r (nrow(NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal[NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal$WorstDayAnalysisGORDPositive==1,])/nrow(NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal))*100)%.
The fraction of time with pH<4 (acid exposure time, AET), measured at WPM, is shown in Figure 1. Worst day analysis demonstrated a positive skew towards a GORD diagnosis on the first day. r myNumDayspositiveTable$Freq[2] (r round(myNumDayspositiveTable$Percentage[1],digits=2))% of patients classified as WPM-GORD-positive were positive for only 1 day. Of those diagnoses, r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD[NegImp_FromImpWithBRavoAndHRMOnlyGORD$NumDaysBravoPositive==1,]$worstDaypH)[1] patients were positive on the first day. For patients who only had one day of GORD, the extra number of GORD diagnoses on day 2, 3 and 4 of WPM was r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD[NegImp_FromImpWithBRavoAndHRMOnlyGORD$NumDaysBravoPositive==1,]$worstDaypH)[2],r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD[NegImp_FromImpWithBRavoAndHRMOnlyGORD$NumDaysBravoPositive==1,]$worstDaypH)[3] and r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD[NegImp_FromImpWithBRavoAndHRMOnlyGORD$NumDaysBravoPositive==1,]$worstDaypH)[4] patients respectively. The number of studies that were negative for GORD in the first 24 hours and 48 hours of the WPM study was r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day1Pos,NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day2Pos)[1] (r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day1Pos,NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day2Pos)[1]/(nrow(NegImp_FromImpWithBRavoAndHRMOnlyGORD))*100 % and r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day1Pos,NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day2Pos)[2] (r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day1Pos,NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day2Pos)[2]/(nrow(NegImp_FromImpWithBRavoAndHRMOnlyGORD))*100%) respectively. The worst day symptom association probability (SAP) was positive in r SAPWDA%. The majority of the worst SAP days occurred in the first 24 hours for both positive and negative WPM studies.
Table 1: Table of demographics for GORD-positive ('Positive') WPM and GORD-negative ('Negative') WPM studies
High resolution manometry
WPM-GORD-positive patients had no significant difference in DCI when compared with WPM-GORD-negative patients (WPM-GORD-positive patients (mmHg): r inline_text(so, variable = "DistalcontractileintegralmeanmmHgcms",column="Positive") vs. WPM-GORD-negative patients (mmHg): r inline_text(so, variable = "DistalcontractileintegralmeanmmHgcms", column = "Negative") r inline_text(so, variable = "DistalcontractileintegralmeanmmHgcms", column = "p.value")). There was no significant difference in the distal latency between the two groups (WPM-GORD-positive patients (s): r inline_text(so, variable = "Distallatency", column = "Positive") vs. WPM-GORD-negative patients (s): r inline_text(so, variable = "Distallatency", column = "Negative") r inline_text(so, variable = "Distallatency", column = "p.value").
There was also no significant difference in the basal respiratory minimum (mmHg) between the two groups (WPM-GORD-positive (mmHg): r inline_text(so, variable = "BasalrespiratoryminmmHg", column = "Positive") vs WPM-GORD-negative (mmHg): r inline_text(so, variable = "BasalrespiratoryminmmHg", column = "Negative") r inline_text(so, variable = "BasalrespiratoryminmmHg", column = "p.value") ) Furthermore no significant difference was noted for any other manometric finding.
pH Impedance measurements
WPM-GORD-positive patients had a longer AET at pH impedance when compared with WPM-GORD-negative patients (WPM-GORD-positive patients (%): r inline_text(so, variable = "MainAcidExpTotalClearanceChannelPercentTime", column = "Positive") vs. WPM-GORD-negative patients (%): r inline_text(so, variable = "MainAcidExpTotalClearanceChannelPercentTime", column = "Negative") r inline_text(so, variable = "MainAcidExpTotalClearanceChannelPercentTime", column = "p.value")).
The AET when measured in an upright position was significantly increased in WPM-GORD-positive patients when compared with the WPM-GORD-negative group (p=0.001). There was no difference in supine reflux (data not shown). The longest time spent in reflux did not show any significant difference although the number of refluxes was significantly increased in the WPM-GORD-positive group r inline_text(so, variable = "MainAcidExpTotalClearanceChannelNumberofAcidEpisodes", column = "Positive") vs. r inline_text(so, variable = "MainAcidExpTotalClearanceChannelNumberofAcidEpisodes", column = "Negative") r inline_text(so, variable = "MainAcidExpTotalClearanceChannelNumberofAcidEpisodes", column = "p.value").
Symptoms and endoscopic findings
The SAP did not differ for any of the individual or grouped symptoms between the two groups. Oesophagitis was more common in the WPM-GORD-Positive group r round((resOR$Oesophagitis$descriptive[3,2]/(resOR$Oesophagitis$descriptive[3,1]+resOR$Oesophagitis$descriptive[3,2])),digits=3)*100% vs. r round(resOR$Oesophagitis$descriptive[2,2]/(resOR$Oesophagitis$descriptive[2,1]+resOR$Oesophagitis$descriptive[2,2]),digits=3)*100%. Of this group, the patients were graded as Los Angeles grade A,B,C or D in 4.2%, 2.1%, 1.4% and 0.7% respectively.
Univariate analysis highlighted r paste0(Hmisc::label(chosen[2]),collapse=" ,"), r paste0(Hmisc::label(chosen[3]),collapse=" ,") and r paste0(Hmisc::label(chosen[4]),collapse=" ,") for further investigation with a multivariate analysis
\pagebreak
\pagebreak
\pagebreak
Multivariate analysis:
Table 2: Table of multivariate analysis to assess pH impedance and manometry variables from the univariate analysis.
A stepwise multiple logistic regression analysis was performed where the independent variables were those being analysed and the dependent variable was the presence of pathological acid reflux on a WPM study (Table 2). Variables chosen were those found to be significantly different in the univariate analysis. Only oesophagitis at the point of BRAVO insertion was also a significant predictor of GORD on mutivariate analysis OR: r mytable[4,6]
\newpage
Discussion
The current study represents the largest cohort of pH impedance GORD-negative patients who subsequently underwent WPM testing at the clinician's discretion. This cohort is unique in that the experience of increased positive diagnosis yield made with wireless pH studies is replicated for the first time with pH impedance as opposed to standard pH studies. This allows patients with predominant non-acid reflux to also be excluded.
1. Discussion of increased yield
1b. Discuss the difference between WDA and average
2. Discussion of metrics if BRAVO ie day1 vs day 2 etc.
It is notable that the WPM studies have shown a distribution of positivity similar to previous studies: the diagnosis of GORD is most often made on the basis of one day of positivity and of those patients who are positive for GORD, the worst day of GORD is predominantly on the first day. [@Sweis2011]. Furthermore a significant number of WPM studies were positive after a negative impedance study. This has been noted on previous pH-only catheter studies [@Sweis2011] and may be related to the extended monitoring time [@Pandolfino2003a].
Worst day versus average day analysis:
The univariate results here are presented as both average and worst day analysis for comparison, although the multiple logistic regression has been performed using average day analyis; this is justified on the basis that this is most commonly used analysis clinically. Worst day analysis (WDA) is also often criticised for being over sensitive *** COMPARISON BETWEEN AV AND WDA * . A higher proportion of patients were GORD positive with WDA (Average day analyis: r (nrow(so$inputs$data[so$inputs$data$AcidRefluxBRAVOAv=='Positive',])/(nrow(so$inputs$data)))*100 % ),WDA: r (nrow(soWDA$inputs$data[soWDA$inputs$data$AcidRefluxBRAVOAv=='Positive',])/(nrow(soWDA$inputs$data)))*100%. This is similar to the increased yield with average day analysis when compared to standard pH catheter studies [@Sweis2011]. Although the overall yield may be different with the two different analyses, there is no difference between the metrics that distinguish between GORD positive and negative patients. This indicates that from an academic perspective whether WDA or average day analysis is unlikely to be relevant. READ VAEZI PAPER ***
Yield versus pH impedance:
The increased yield of wireless pH as compared with standard pH studies has been demonstrated previously [@Scarpulla2007a]. The reasons for the increased yield may be multifactorial including increased patient tolerability [@Wong2005a] and better detection of intermittent reflux episodes. The fact that there is an increased yield over and above pH impedance is interesting for two reasons. Firstly, pH impedance is often promoted as a tool to detect further episodes of GORD in standard pH negative patients as it is able to detect non-acid reflux episodes. This study suggests that in fact there are further acid related reflux episodes that pH impedance is not detecting and that rather than look for non-acid related causes of a patient's symptoms, a longer assessment for acid reflux may in fact be more appropriate as the next line of investigation. Secondly, pH impedance is time consuming when compared to BRAVO analysis and standard pH analysis; the analysis of much of the two latter studies can be automated in a way that is not possible with pH impedance. Given the increased patient comfort and tolerability associated with WPM studies [@Wong2005a], automation of the much of the analysis, the ability to perform an endoscopic examination during capsule placement and the increased positive diagnosis yield, there may be a strong case for further clinical studies assessing the use of WPM as a first line investigation instead of standard 24 hour pH or 24 hour pH impedance monitoring.
Dissection of days in BRAVO wireless capsule assessment:
In this current study, of the patients who turn out to be GORD positive, the number of patients who were GORD negative in the first 24 hours was r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day1Pos,NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day2Pos)[2]/(nrow(NegImp_FromImpWithBRavoAndHRMOnlyGORD))*100%. A further r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day1Pos,NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day2Pos)[1]/(nrow(NegImp_FromImpWithBRavoAndHRMOnlyGORD))*100% were still negative after 48 hours indicating that even a 48 hour pH study may be insufficient. The number of patients who were GORD positive on only one day also argues for monitoring for longer than 24 hours as it suggests that a significant number of GORD positive patients in this cohort may have intermittent symptoms.
Predictors of GORD positive study:
The fact that most motility variables were not associated with WPM GORD-positivity is in keeping with the established literature that there is no specific pattern at HRM that predicts the presence of GORD. There are findings however that are more likely in GORD. Distal contractile integral (DCI), a measure of contractile vigour, has been shown to be lower in GORD patients [@VanHoeij2015] although in mild ineffective oesophageal motility the relationship to GORD is less well proven ([@FORNARI2007]) Univariate analysis has only shown a tendency towards significance for the DCI between the two groups although multivariate analysis did not demonstrate this to be a significant discriminator. A reduced basal respiratory minimum pressure is also known to be associated with increasing severity of GORD ([@Savarino2011]) as has been found in this study. Again this was not found to be an important factor on the multivariate analysis. The fact that this is the case for both DCI and the basal respiratory minimum may be related to the study population. r round(myNumDayspositiveTable$Percentage[1],digits = 2)% of patients positive for GORD by WPM were only positive on one day indicating the intermittent nature of reflux in a proportion of this cohort. Furthermore most of the patients did not have erosive oesophagitis. The DCI is well documented to be more severely affected in patients with oesophagitis and motility disorders are more commonly seen as patients progress from non-erosive reflux disease to increasing severity of GORD to Barrett's oesophagus [@Savarino2011]. The study population therefore represents a GORD population less likely to have motility dysfunction. That the presence of oesophagitis is a predictor of a positive WPM is not surprising. Arguably, the presence of oesophagitis at endoscopy should preclude the need for further pH testing (although in this study the oesophagitis was diagnosed at the point of WPM insertion) so that this should not be used as a predictor of a subsequent WPM study.
Change this as AET is not relevant any more- discuss oesophagitis instead ####
The significant difference in the acid exposure time at pH impedance, between those who have GORD at WPM and those who do not is interesting. Some false negative pH studies may be related to patients altering their behaviour because of the presence of a nasal catheter [@Fass1999] therefore reducing the AET into the normal range. It is also possible that the cohort in this study represent more borderline GORD diagnoses making the ability to use predictors more difficult. This significant difference does not translate in to a value with sufficient accuracy to allow physicians to determine who should go on to have a WPM study. Furthermore, no other measurement taken at impedance, including parameters of non-acid reflux, are relevant albeit in this selected group.
Discussion re symptom changes:
The decision to submit a patient to a WPM study following a negative pH impedance study may be more likely in those who have typical and convincing symptoms. Typical GORD symptoms such as regurgitation and heartburn are well documented to be better predictors of oesophagitis than atypical symptoms (https://doi.org/10.1016/j.gtc.2013.11.002). It is therefore a surprise that symptoms do not predict the presence of GORD. This may be because the symptom was user-defined as it was recorded by the patient. Patient-defined features do not always correlate well with questionnaire defined symptoms.
Limitations
The study is limited insofar as it is retrospective and the selection criteria for those who were selected for WPM were at the physician's discretion. Also, as noted above, patient defined symptoms may be difficult to interpret. However the patient cohort does represent a real-world selection at a high throughput centre.
In conclusion this study demonstrates that there is a significant increase in positive GORD diagnosis yield in patients with a negative pH impedance in a cohort selected on clinical suspicion for further testing. It is difficult to determine which patients may benefit from prolonged oesophageal pH monitoring in the context of a negative 24 hour pH impedance test. There is no single factor that reliably distinguishes those who will and will not have GORD on WPM with sufficient sensitivity and specificity. The presence of oesophagitis at endoscopy should indicate that a further period of monitoring is not necessary rather than act as an indicator for further testing. The combination of increased diagnostic yield and the inability to accurately predict who will have a positive extended WPM could strengthen the argument for a more formal assessment of WPM as the first line investigation in patients who require a positive diagnosis of GORD.
Data analysis
We used r cite_r("r-references.bib") for all our analyses.
\newpage
Author contribution
Sebastian Zeki,Jafar Jafari,Guiping Sui,Ismail Miah,Minerva daSilva,Anna Wolak: acquisition of data, Andrew Davies, Abrie Botha, Terry Wong, Jafar Jafari: data analysis and interpretation, Sebastian Zeki: statistical analysis, drafting of the manuscript; Jafar Jafari, Guiping Sui, Ismail Miah, Minerva daSilva, Anna Wolak, Jason Dunn, Andrew Davies, James Gossage, Abrie Botha, Terry Wong: critical revision of the manuscript for important intellectual content.
References
#r_refs(file = "r-references.bib")
\begingroup
\setlength{\parindent}{-0.5in}
\setlength{\leftskip}{0.5in}
\endgroup
sebastiz/PhysiMineR documentation built on Oct. 3, 2023, 3:46 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
title : "Extended Wireless pH monitoring significantly increases GORD diagnoses in patients with a normal pH impedance study." shorttitle : "Extended Wireless pH monitoring significantly increases GORD diagnoses in patients with a normal pH impedance study."
author:
- name : "Sebastian S Zeki"
affiliation : "1,3"
corresponding : yes
address : "Gastroenterology Department,Westminster Bridge Road, London SE1 7EH"
email : "sebastiz@hotmail.com"
- name : "Ismail Miah"
affiliation : "1"
- name : "Anna Wolak"
affiliation : "1"
- name : "Minerva daSilva"
affiliation : "1"
- name : "Jason Dunn"
affiliation : "1,2"
- name : "Andrew Davies"
affiliation : "1,2"
- name : "James Gossage"
affiliation : "1,2"
- name : "Abrie Botha"
affiliation : "1,2"
- name : "Guiping Sui"
affiliation : "1"
- name : "Jafar Jafari"
affiliation : "1"
- name : "Terry Wong"
affiliation : "1,2"
affiliation: - id : "1" institution : "Centre for Oeosphageal Diseases, Guy's and St. Thomas Hospital, Westminster Bridge Road, London" - id : "2" institution : "Karolinska Instituet, Karolinska,Sweden" - id : "3" institution : "Bart's Cancer Institute,Charterhouse Square, London"
authornote: | header-includes: - \usepackage{setspace}\doublespacing - \AtBeginEnvironment{tabular}{\singlespacing} - \AtBeginEnvironment{lltable}{\singlespacing} - \AtBeginEnvironment{tablenotes}{\doublespacing} - \captionsetup[table]{font={stretch=1.5}} - \captionsetup[figure]{font={stretch=1.5}}
abstract: | Introduction: In cases where a false negative diagnosis of gastro-oesophageal reflux disease (GORD) is suspected, using catheter-based methods such as pH impedance and high resolution manometry (HRM), further testing can be done using a wireless pH monitoring (WPM) device which can assess acidic reflux for 96 hours. The increased yield of GORD-positive diagnoses with WPM is unknown. Further testing is at the clinician’s discretion with no guidance as to which parameters on pH impedance or manometry, when this test is negative, may predict a subsequent positive WPM test. Aim: a) To determine the increased yield and b) determine predictors of GORD positivity in patients with negative pH impedance. Method: The increased GORD-positive diagnostic yield (average day analysis) by with WPM was determined in --NUMBER-- consecutive patients who had also undergone negative pH impedance with HRM. Univariate and multivariate analysis was performed to determine predictors of a positive WPM. Results: Of the 212 patients who underwent WPM after a negative pH impedance study, 56 (29%) were found to have a positive result for GORD. Only the AET at pH impedance and endoscopically visible oesophagitis was significantly associated with a WPM-GORD positive result on multivariate analysis. However the AET was insufficiently accurate to be used as a predictive clinical parameter (AUC: 0.61). --(AET NUMBERS)-- OESOPHAGITIS AND PVALUE NUMBERS. Conclusion: There is a significant increased yield for GORD positive diagnoses with WPM although predicting this from pH impedance and manometry if pH impedance is negative is difficult. --PHRASE MORE EFFICIENTLY-- The increased yield suggests that a more formal study to assess the difference in GORD detection between WPM and impedance may be useful. --PHRASE MORE EFFICIENTLY--
<!-- https://tinyurl.com/ybremelq -->
keywords : "Gastro-oesophageal reflux disease, pH studies, wireless pH monitoring"
wordcount : This document contains r wordcountaddin::word_count() words
bibliography : ["r-references.bib"]
floatsintext : yes figurelist : no tablelist : no footnotelist : no linenumbers : yes mask : no draft : no
documentclass : "apa6" classoption : "man" output : papaja::apa6_docx
knitr::opts_chunk$set(cache.extra = knitr::rand_seed, echo = FALSE, comment = FALSE, warning = FALSE, message = FALSE) set.seed(42) options(digits=2,xtable.comment = FALSE) library("papaja")
# Seed for random number generation library(knitr) library(here) library(EndoMineR) library(PhysiMineR) library(dplyr) library(tidyr) library(kableExtra) library(ggpubr) library(readxl) library(gtsummary)
read_chunk(here("inst","Projects","BRAVOStudies", "NegImpPredictorsOfAllPosBRAVO","munge", "PreProcessing.R"))
Introduction
Wireless pH monitoring (WPM) is a useful method for the prolonged analysis of oesophageal pH. Our previous studies have shown that WPM is a useful subsequent investigation or as an alternative to catheter based studies. REF RAMI's STUDY. A particular advantage of WPM is the accomodation of day to day variation and improved ablility for patients to perform their normal activities of daily living. (REF WONG) False negative catheter studies may occur if reflux is intermittent so that a 24 hour window of assessment is insufficient [@TSENG2005a]. It has been estimated that a significant proportion of patients would have a positive finding of GORD if assessed with extended pH studies rather than a 24 hour pH assessment [@Scarpulla2007a; @Sweis2009] . The increased yield of WPM over standard catheter-based pH impedance specifically however, has never been determined.
When deciding on the need for further pH testing given a negative 24 hour study, a clinican may take into account a number of factors including the nature and frequency of the symptoms, or other tests such as the endoscopy or manometry which may provide circumstantial evidence of GORD [@Gyawali2018]. There is however no guidance as to who should undergo prolonged acid monitoring in the presence of a negative pH impedance study.
The aim of the current study is to establish the diagnostic yield by average day analysis and potential predictors of a wireless pH study demonstrating GORD on a worst day analysis given a negative 24 pH impedance study, using the pH impedance, manometry and endoscopy results.
Methods
Subjects
The results of all consecutive patients who had undergone a 24 hour pH impedance in patients with suspected GORD and a subsequent WPM between r paste(min(lubridate::month(ImpAndBravoWithHRM$MainPtDataDateofAdmission,label=TRUE, abbr=FALSE)),min(lubridate::year(ImpAndBravoWithHRM$MainPtDataDateofAdmission))) and r paste(max(lubridate::month(ImpAndBravoWithHRM$MainPtDataDateofAdmission,label=TRUE, abbr=FALSE)),max(lubridate::year(ImpAndBravoWithHRM$MainPtDataDateofAdmission))) were retrospectively interrogated. Only patients who had undergone an initially negative 24 hour pH impedance result were included. A negative pH impedance test was defined as having more than 40 reflux episodes and an AET for <= 4.2% of a 24 hour period [@Gyawali2018]. Patients with a normal AET but with >80 reflux episodes were also excluded on the basis that they have non-acid reflux. A WPM study was defined as positive if the patient had an AET for >=5.3% of any 24 hour period (worst day analysis-[@Pandolfino2003] and [@Wenner2005]). A worst day analysis was chosen as it is more specific than average day analysis [@Sweis2011].
All patients selected were adults over the age of 18. Ethics was approved (IRAS 18/NW/0120). Patients were excluded for the following reasons: 1. Previous oesophageal surgery or intervention such as endoscopic mucosal resection or radiofrequency ablation. 2. Less than 24 hours of recording on pH impedance. 3. Less than 48 hours of recorded data from the patient's WPM study.
High resolution manometry protocol
Following local analgesia of the nares the catheter was introduced trans-nasally and the patient was instructed to drink water through a straw whilst the HRM catheter was advanced to the stomach. The high resolution manometry (HRM) catheter depth was adjusted to ensure manometric visual of the upper oesophageal sphincter (UOS), the gastro-oesophageal junction (GOJ) and gastric pressures. 10 single swallows of 5ml were performed with each being 20 seconds apart. Each 5ml water swallow was then assessed in accordance to Chicago classification (version 3)[@Kahrilas2015] using Manoview software (version 3) (Sierra Scientific Instruments).
pH impedance protocol:
Patients underwent reflux monitoring using Sandhill Scientific multichannel impedance pH catheters (ZAN-BG-44) which were inserted trans-nasally after applying local anaesthesia (xylocaine). The dual pH sensors of the catheter were positioned 5cm below and above the manometric LOS. The impedance sensors were positioned above the LOS by 3cm, 5cm, 9cm, 15cm and 19cm. Acid reflux was considered when a retrograde impedance flow was observed and the oesophageal pH sensor was detecting a pH value of <4. The data was captured by ZepHrTM recording device and data was analysed using the BioVIEW Analysis software (5.7.1.0).
96 hour WPM protocol:
The 96 hour WPM procedure was performed after a 6 hour fasting period and discontinuation of acid suppressant The patient was asked to complete a diary of symptoms and meal/drink times (except for still water), using the clock on the WPM receiver, during the 96-hour pH recording. 3 symptoms of reflux were decided by the patient and symbols allocated for each symptom. The patient was requested to record these symptoms by pressing the appropriate symptom markers (symbols) on the receiver, when the symptom occurred.
After the WPM capsule was calibrated, the patient was sedated according to local practice. Following a complete endoscopic examination of the upper GI tract a pH sensor located in a capsule (WPM pH capsule: 25mm x 6mm x 5.5 mm) was pinned temporarily to the wall of the Oesophagus 6 cm proximal to the Z line. The following parameters were obtained from the analysis and compared against normal reference values 1. Number of reflux episodes per day, 2. Total percent of time spent in reflux, 3. Percent of time spent in reflux in the upright position, 4. Percent of time spent in reflux in the supine position, 5. Percent of time spent in reflux post prandial, 6. De Meester Score
Oesophagitis
Endoscopy records were retrospectively consulted for patients who had been submitted for a BRAVO with a negative impedance catheter study. The degree of oesophagitis recorded by the endoscopist at the time of BRAVO insertion was recorded for the patient. All endoscopists inserting BRAVO catheters used the Los Angeles oesophagitis grading system. Los Angeles grade A to D was included as classifying the patient as having oesophagitis.
Statistical analyses:
WPM, HRM and pH impedance results were ordered according to the patient identifier and date of the study. To ensure that there was no duplicated data for patients who had undergone more than two of any test, only tests chronologically closest to the test to be merged were chosen. To ensure relevance, tests were merged only if there was less than one year's difference between the two. Variables for analysis were chosen from impedance and HRM data on the following basis: Lack of multicolinearity and an adequate number of data points to remove skew from missing data. All quantitative data were presented as mean±SD. Fisher's exact test or Wilcoxon rank sum test was used to compare non-parametric numerical data. Student's t-test was used to compare parametric numerical data. Categorical data was compared using a chi-squared test of independence. The following variables were chosen for analysis: Age (years), basal respiratory minimum (mmHg),mean residual pressure (mmHg),distal contractile integral (mean mmHg/cm/s), contractile front velocity (m/s),distal latency, percentage failed swallows, percentage panoesophageal pressurization,percentage large breaks, number of small breaks, number of acid episodes, acid exposure time, mean acid clearance time, the longest acid episode and the presence of oesophagitis at endoscopy. The symptom association probability (SAP) at pH impedance was also used. Analysis was performed on individual SAP symptoms or symptoms grouped as typical (heartburn,regurgitation,chest pain), atypical (belching, cough, throat symptoms, abdominal pain, vomiting, nausea) or both (mixed). Statistical comparison was carried out on all recorded parameters and those with p<0.05 were selected for evaluation in a multivariate model with a stepwise (forward selection/backward elimination) method. p<0.05 was taken as the threshold of significance for the multivariate model and the strength of association was expressed as odds ratio (OR) with 95% confidence interval (CI).
Results
Baseline demographic data
Of the r nrow(NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal2) included, r nrow(NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal2[NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal2$AcidRefluxBRAVOAv=="Positive",]) studies were positive for GORD (Male:Female r resOR$Gender$descriptive[3,2]: r resOR$Gender$descriptive[3,1], average age of r inline_text(so, variable = "ageInYears", column = "Positive") years and
r nrow(NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal2[NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal2$AcidRefluxBRAVOAv=="Negative",]) negative for GORD (Male:Female r resOR$Gender$descriptive[2,2]: r resOR$Gender$descriptive[2,1] average age of r inline_text(so, variable = "ageInYears", column = "Negative") years.
Using a worst day analysis, r nrow(NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal[NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal$WorstDayAnalysisGORDPositive==1,]) were positive (r (nrow(NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal[NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal$WorstDayAnalysisGORDPositive==1,])/nrow(NegImp_FromImpWithBRavoAndHRMMinSetQualityFinal))*100)%.
The fraction of time with pH<4 (acid exposure time, AET), measured at WPM, is shown in Figure 1. Worst day analysis demonstrated a positive skew towards a GORD diagnosis on the first day. r myNumDayspositiveTable$Freq[2] (r round(myNumDayspositiveTable$Percentage[1],digits=2))% of patients classified as WPM-GORD-positive were positive for only 1 day. Of those diagnoses, r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD[NegImp_FromImpWithBRavoAndHRMOnlyGORD$NumDaysBravoPositive==1,]$worstDaypH)[1] patients were positive on the first day. For patients who only had one day of GORD, the extra number of GORD diagnoses on day 2, 3 and 4 of WPM was r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD[NegImp_FromImpWithBRavoAndHRMOnlyGORD$NumDaysBravoPositive==1,]$worstDaypH)[2],r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD[NegImp_FromImpWithBRavoAndHRMOnlyGORD$NumDaysBravoPositive==1,]$worstDaypH)[3] and r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD[NegImp_FromImpWithBRavoAndHRMOnlyGORD$NumDaysBravoPositive==1,]$worstDaypH)[4] patients respectively. The number of studies that were negative for GORD in the first 24 hours and 48 hours of the WPM study was r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day1Pos,NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day2Pos)[1] (r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day1Pos,NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day2Pos)[1]/(nrow(NegImp_FromImpWithBRavoAndHRMOnlyGORD))*100 % and r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day1Pos,NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day2Pos)[2] (r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day1Pos,NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day2Pos)[2]/(nrow(NegImp_FromImpWithBRavoAndHRMOnlyGORD))*100%) respectively. The worst day symptom association probability (SAP) was positive in r SAPWDA%. The majority of the worst SAP days occurred in the first 24 hours for both positive and negative WPM studies.
Table 1: Table of demographics for GORD-positive ('Positive') WPM and GORD-negative ('Negative') WPM studies
High resolution manometry
WPM-GORD-positive patients had no significant difference in DCI when compared with WPM-GORD-negative patients (WPM-GORD-positive patients (mmHg): r inline_text(so, variable = "DistalcontractileintegralmeanmmHgcms",column="Positive") vs. WPM-GORD-negative patients (mmHg): r inline_text(so, variable = "DistalcontractileintegralmeanmmHgcms", column = "Negative") r inline_text(so, variable = "DistalcontractileintegralmeanmmHgcms", column = "p.value")). There was no significant difference in the distal latency between the two groups (WPM-GORD-positive patients (s): r inline_text(so, variable = "Distallatency", column = "Positive") vs. WPM-GORD-negative patients (s): r inline_text(so, variable = "Distallatency", column = "Negative") r inline_text(so, variable = "Distallatency", column = "p.value").
There was also no significant difference in the basal respiratory minimum (mmHg) between the two groups (WPM-GORD-positive (mmHg): r inline_text(so, variable = "BasalrespiratoryminmmHg", column = "Positive") vs WPM-GORD-negative (mmHg): r inline_text(so, variable = "BasalrespiratoryminmmHg", column = "Negative") r inline_text(so, variable = "BasalrespiratoryminmmHg", column = "p.value") ) Furthermore no significant difference was noted for any other manometric finding.
pH Impedance measurements
WPM-GORD-positive patients had a longer AET at pH impedance when compared with WPM-GORD-negative patients (WPM-GORD-positive patients (%): r inline_text(so, variable = "MainAcidExpTotalClearanceChannelPercentTime", column = "Positive") vs. WPM-GORD-negative patients (%): r inline_text(so, variable = "MainAcidExpTotalClearanceChannelPercentTime", column = "Negative") r inline_text(so, variable = "MainAcidExpTotalClearanceChannelPercentTime", column = "p.value")).
The AET when measured in an upright position was significantly increased in WPM-GORD-positive patients when compared with the WPM-GORD-negative group (p=0.001). There was no difference in supine reflux (data not shown). The longest time spent in reflux did not show any significant difference although the number of refluxes was significantly increased in the WPM-GORD-positive group r inline_text(so, variable = "MainAcidExpTotalClearanceChannelNumberofAcidEpisodes", column = "Positive") vs. r inline_text(so, variable = "MainAcidExpTotalClearanceChannelNumberofAcidEpisodes", column = "Negative") r inline_text(so, variable = "MainAcidExpTotalClearanceChannelNumberofAcidEpisodes", column = "p.value").
Symptoms and endoscopic findings
The SAP did not differ for any of the individual or grouped symptoms between the two groups. Oesophagitis was more common in the WPM-GORD-Positive group r round((resOR$Oesophagitis$descriptive[3,2]/(resOR$Oesophagitis$descriptive[3,1]+resOR$Oesophagitis$descriptive[3,2])),digits=3)*100% vs. r round(resOR$Oesophagitis$descriptive[2,2]/(resOR$Oesophagitis$descriptive[2,1]+resOR$Oesophagitis$descriptive[2,2]),digits=3)*100%. Of this group, the patients were graded as Los Angeles grade A,B,C or D in 4.2%, 2.1%, 1.4% and 0.7% respectively.
Univariate analysis highlighted r paste0(Hmisc::label(chosen[2]),collapse=" ,"), r paste0(Hmisc::label(chosen[3]),collapse=" ,") and r paste0(Hmisc::label(chosen[4]),collapse=" ,") for further investigation with a multivariate analysis
\pagebreak
\pagebreak
\pagebreak
Multivariate analysis:
Table 2: Table of multivariate analysis to assess pH impedance and manometry variables from the univariate analysis.
A stepwise multiple logistic regression analysis was performed where the independent variables were those being analysed and the dependent variable was the presence of pathological acid reflux on a WPM study (Table 2). Variables chosen were those found to be significantly different in the univariate analysis. Only oesophagitis at the point of BRAVO insertion was also a significant predictor of GORD on mutivariate analysis OR: r mytable[4,6]
\newpage
Discussion
The current study represents the largest cohort of pH impedance GORD-negative patients who subsequently underwent WPM testing at the clinician's discretion. This cohort is unique in that the experience of increased positive diagnosis yield made with wireless pH studies is replicated for the first time with pH impedance as opposed to standard pH studies. This allows patients with predominant non-acid reflux to also be excluded.
1. Discussion of increased yield
1b. Discuss the difference between WDA and average
2. Discussion of metrics if BRAVO ie day1 vs day 2 etc.
It is notable that the WPM studies have shown a distribution of positivity similar to previous studies: the diagnosis of GORD is most often made on the basis of one day of positivity and of those patients who are positive for GORD, the worst day of GORD is predominantly on the first day. [@Sweis2011]. Furthermore a significant number of WPM studies were positive after a negative impedance study. This has been noted on previous pH-only catheter studies [@Sweis2011] and may be related to the extended monitoring time [@Pandolfino2003a].
Worst day versus average day analysis:
The univariate results here are presented as both average and worst day analysis for comparison, although the multiple logistic regression has been performed using average day analyis; this is justified on the basis that this is most commonly used analysis clinically. Worst day analysis (WDA) is also often criticised for being over sensitive *** COMPARISON BETWEEN AV AND WDA * . A higher proportion of patients were GORD positive with WDA (Average day analyis: r (nrow(so$inputs$data[so$inputs$data$AcidRefluxBRAVOAv=='Positive',])/(nrow(so$inputs$data)))*100 % ),WDA: r (nrow(soWDA$inputs$data[soWDA$inputs$data$AcidRefluxBRAVOAv=='Positive',])/(nrow(soWDA$inputs$data)))*100%. This is similar to the increased yield with average day analysis when compared to standard pH catheter studies [@Sweis2011]. Although the overall yield may be different with the two different analyses, there is no difference between the metrics that distinguish between GORD positive and negative patients. This indicates that from an academic perspective whether WDA or average day analysis is unlikely to be relevant. READ VAEZI PAPER ***
Yield versus pH impedance:
The increased yield of wireless pH as compared with standard pH studies has been demonstrated previously [@Scarpulla2007a]. The reasons for the increased yield may be multifactorial including increased patient tolerability [@Wong2005a] and better detection of intermittent reflux episodes. The fact that there is an increased yield over and above pH impedance is interesting for two reasons. Firstly, pH impedance is often promoted as a tool to detect further episodes of GORD in standard pH negative patients as it is able to detect non-acid reflux episodes. This study suggests that in fact there are further acid related reflux episodes that pH impedance is not detecting and that rather than look for non-acid related causes of a patient's symptoms, a longer assessment for acid reflux may in fact be more appropriate as the next line of investigation. Secondly, pH impedance is time consuming when compared to BRAVO analysis and standard pH analysis; the analysis of much of the two latter studies can be automated in a way that is not possible with pH impedance. Given the increased patient comfort and tolerability associated with WPM studies [@Wong2005a], automation of the much of the analysis, the ability to perform an endoscopic examination during capsule placement and the increased positive diagnosis yield, there may be a strong case for further clinical studies assessing the use of WPM as a first line investigation instead of standard 24 hour pH or 24 hour pH impedance monitoring.
Dissection of days in BRAVO wireless capsule assessment:
In this current study, of the patients who turn out to be GORD positive, the number of patients who were GORD negative in the first 24 hours was r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day1Pos,NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day2Pos)[2]/(nrow(NegImp_FromImpWithBRavoAndHRMOnlyGORD))*100%. A further r table(NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day1Pos,NegImp_FromImpWithBRavoAndHRMOnlyGORD$Day2Pos)[1]/(nrow(NegImp_FromImpWithBRavoAndHRMOnlyGORD))*100% were still negative after 48 hours indicating that even a 48 hour pH study may be insufficient. The number of patients who were GORD positive on only one day also argues for monitoring for longer than 24 hours as it suggests that a significant number of GORD positive patients in this cohort may have intermittent symptoms.
Predictors of GORD positive study:
The fact that most motility variables were not associated with WPM GORD-positivity is in keeping with the established literature that there is no specific pattern at HRM that predicts the presence of GORD. There are findings however that are more likely in GORD. Distal contractile integral (DCI), a measure of contractile vigour, has been shown to be lower in GORD patients [@VanHoeij2015] although in mild ineffective oesophageal motility the relationship to GORD is less well proven ([@FORNARI2007]) Univariate analysis has only shown a tendency towards significance for the DCI between the two groups although multivariate analysis did not demonstrate this to be a significant discriminator. A reduced basal respiratory minimum pressure is also known to be associated with increasing severity of GORD ([@Savarino2011]) as has been found in this study. Again this was not found to be an important factor on the multivariate analysis. The fact that this is the case for both DCI and the basal respiratory minimum may be related to the study population. r round(myNumDayspositiveTable$Percentage[1],digits = 2)% of patients positive for GORD by WPM were only positive on one day indicating the intermittent nature of reflux in a proportion of this cohort. Furthermore most of the patients did not have erosive oesophagitis. The DCI is well documented to be more severely affected in patients with oesophagitis and motility disorders are more commonly seen as patients progress from non-erosive reflux disease to increasing severity of GORD to Barrett's oesophagus [@Savarino2011]. The study population therefore represents a GORD population less likely to have motility dysfunction. That the presence of oesophagitis is a predictor of a positive WPM is not surprising. Arguably, the presence of oesophagitis at endoscopy should preclude the need for further pH testing (although in this study the oesophagitis was diagnosed at the point of WPM insertion) so that this should not be used as a predictor of a subsequent WPM study.
Change this as AET is not relevant any more- discuss oesophagitis instead ####
The significant difference in the acid exposure time at pH impedance, between those who have GORD at WPM and those who do not is interesting. Some false negative pH studies may be related to patients altering their behaviour because of the presence of a nasal catheter [@Fass1999] therefore reducing the AET into the normal range. It is also possible that the cohort in this study represent more borderline GORD diagnoses making the ability to use predictors more difficult. This significant difference does not translate in to a value with sufficient accuracy to allow physicians to determine who should go on to have a WPM study. Furthermore, no other measurement taken at impedance, including parameters of non-acid reflux, are relevant albeit in this selected group.
Discussion re symptom changes:
The decision to submit a patient to a WPM study following a negative pH impedance study may be more likely in those who have typical and convincing symptoms. Typical GORD symptoms such as regurgitation and heartburn are well documented to be better predictors of oesophagitis than atypical symptoms (https://doi.org/10.1016/j.gtc.2013.11.002). It is therefore a surprise that symptoms do not predict the presence of GORD. This may be because the symptom was user-defined as it was recorded by the patient. Patient-defined features do not always correlate well with questionnaire defined symptoms.
Limitations
The study is limited insofar as it is retrospective and the selection criteria for those who were selected for WPM were at the physician's discretion. Also, as noted above, patient defined symptoms may be difficult to interpret. However the patient cohort does represent a real-world selection at a high throughput centre.
In conclusion this study demonstrates that there is a significant increase in positive GORD diagnosis yield in patients with a negative pH impedance in a cohort selected on clinical suspicion for further testing. It is difficult to determine which patients may benefit from prolonged oesophageal pH monitoring in the context of a negative 24 hour pH impedance test. There is no single factor that reliably distinguishes those who will and will not have GORD on WPM with sufficient sensitivity and specificity. The presence of oesophagitis at endoscopy should indicate that a further period of monitoring is not necessary rather than act as an indicator for further testing. The combination of increased diagnostic yield and the inability to accurately predict who will have a positive extended WPM could strengthen the argument for a more formal assessment of WPM as the first line investigation in patients who require a positive diagnosis of GORD.
Data analysis
We used r cite_r("r-references.bib") for all our analyses.
\newpage
Author contribution
Sebastian Zeki,Jafar Jafari,Guiping Sui,Ismail Miah,Minerva daSilva,Anna Wolak: acquisition of data, Andrew Davies, Abrie Botha, Terry Wong, Jafar Jafari: data analysis and interpretation, Sebastian Zeki: statistical analysis, drafting of the manuscript; Jafar Jafari, Guiping Sui, Ismail Miah, Minerva daSilva, Anna Wolak, Jason Dunn, Andrew Davies, James Gossage, Abrie Botha, Terry Wong: critical revision of the manuscript for important intellectual content.
References
#r_refs(file = "r-references.bib")
\begingroup \setlength{\parindent}{-0.5in} \setlength{\leftskip}{0.5in}
\endgroup
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.