- Email: [email protected]
Model to Select On-Therapy vs Off-Therapy Tests for Patients With Refractory Esophageal or Extraesophageal Symptoms
Gastroenterology 2018;155:1729–1740
Model to Select On-Therapy vs Off-Therapy Tests for Patients With Refractory Esophageal or Extraesophageal Symptoms
1
Division of Gastroenterology, Hepatology and Nutrition, Vanderbilt University Medical Center, Nashville, Tennessee; Department of Biostatistics, Vanderbilt University Medical Center, Nashville, Tennessee; 3Vanderbilt Voice Center, Vanderbilt University Medical Center, Nashville, Tennessee; 4Swedish Gastroenterology, Seattle, Washington; 5Division of Otolaryngology and Wisconsin Surgical Outcomes Research Program, University of Wisconsin, Madison, Wisconsin; 6Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota; and 7Division of Gastroenterology, Washington University Medical Center, St Louis, Missouri 2
See Covering the Cover synopsis on page 1661. BACKGROUND & AIMS: It is not clear whether we should test for reflux in patients with refractory heartburn or extraesophageal reflux (EER) symptoms, such as cough, hoarseness, or asthma. Guidelines recommend testing patients by pH monitoring when they are on or off acid-suppressive therapies based on pretest probability of reflux, determined by expert consensus. However, it is not clear what constitutes a low or high pretest probability of reflux in these patients. We aimed to develop a model that clinicians can use at bedside to estimate pretest probability of abnormal reflux. METHODS: We performed a prospective study of 471 adult patients with refractory heartburn (n ¼ 214) or suspected EER symptoms (n ¼ 257) who underwent endoscopy with wireless pH monitoring while they were off acid-suppressive treatment and assigned them to groups based on symptoms at presentation (discovery cohort). Using data from the discovery cohort, we performed proportional odds ordinal logistic regression to select factors (easy to obtain demographic criteria and clinical symptoms such as heartburn, regurgitation, asthma, cough, and hoarseness) associated with esophageal exposure to acid. We validated our findings in a cohort of 118 patients with the same features from 2 separate tertiary care centers (62% women; median age 59 years; 62% with cough as presenting symptom). RESULTS: Abnormal pH (>5.5% of time spent at pH <4) was found in 56% of patients with heartburn and 63% of patients with EER (P ¼ .15). Within EER groups, abnormal pH was detected in a significantly larger proportion (80%) of patients with asthma compared with patients with cough (60%) or hoarseness (51%; P < .01). Factors significantly associated with abnormal pH in patients with heartburn were presence of hiatal hernia and body mass index >25 kg/m2. In patients with EER, the risk of reflux was independently associated with the presence of concomitant heartburn (odds ratio [OR] 2.0; 95% confidence interval [CI] 1.3–3.1), body mass index >25 kg/m2 (OR 2.1; 95% CI 1.5–3.1), asthma (OR 2.0; 95% CI 1.2–3.5), and presence of hiatal hernia (OR 1.9; 95% CI 1.2–3.1). When we used these factors to create a scoring system, we found that a score of 2 excluded patients with moderate to severe reflux, with a negative predictive value of 80% in the discovery cohort and a negative predictive value of 85% in the validation cohort.
CONCLUSION: We developed a clinical model to estimate pretest probability of abnormal pH in patients who were failed by proton pump inhibitor therapy. This system can help guide clinicians at bedside in determining the most appropriate diagnostic test in this challenging group of patients.
Keywords: Body Mass Index; Heartburn, Asthma, and BMI Extraesophageal Reflux (HAs-BEER); Proton Pump Inhibitor; Gastroesophageal Reflux Disease.
A
pproximately 25–75 million people in the United States are affected by gastroesophageal reflux disease (GERD) and 13% of Americans use antireflux medications at least twice weekly.1 GERD is a spectrum of disease classically producing symptoms of heartburn and acid regurgitation, but patients also can present with extraesophageal manifestations such as asthma, cough, hoarseness, and recurrent pneumonia. Because of its variable presentation, GERD remains the leading outpatient physician diagnosis for gastrointestinal disorders in the United States, with a very high economic burden on society.2 The annual direct cost for GERD management is reported to be approximately $9713 per patient, with national expenditures ranging from $9.3 to $12.3 billion.4,5 Patients with symptoms associated with extraesophageal reflux (EER) have even a greater economic burden, with a mean direct cost of $5438 per patient and a national expenditure estimated to be higher than $50 billion.6 This is primarily driven by uncertainty over EER diagnosis because of the lack of a gold standard diagnostic test, which leads to un-
Abbreviations used in this paper: ACG, American College of Gastroenterology; BMI, body mass index; CI, confidence interval; EER, extraesophageal reflux; EGD, esophagogastroduodenoscopy; GERD, gastroesophageal reflux disease; HAs-BEER, Heartburn, Asthma, and BMI Extraesophageal Reflux; MII-pH, multichannel intraluminal impedance with pH; OR, odds ratio; PPI, proton pump inhibitor. Most current article © 2018 by the AGA Institute 0016-5085/$36.00 https://doi.org/10.1053/j.gastro.2018.08.038
CLINICAL AT
Dhyanesh A. Patel,1 Rohit Sharda,4 Yash A. Choksi,1 James C. Slaughter,2 Tina Higginbotham,1 C. Gaelyn Garrett,3 David O. Francis,5 Karthik Ravi,6 Stephen Hasak,7 David Katzka,6 C. Prakash Gyawali,7 and Michael F. Vaezi1
1730 Patel et al
WHAT YOU NEED TO KNOW BACKGROUND AND CONTEXT There are currently no studies that define what constitutes low- or high pre-test probability of reflux in patients with refractory heartburn or extra-esophageal symptoms.
Gastroenterology Vol. 155, No. 6
symptoms (asthma, cough, or hoarseness); (2) develop a simple, predictive model that can be used by clinicians at bedside to determine pretest probability of abnormal pH; and (3) externally validate the predictive model using independent data from a similar population at 2 different institutions.
NEW FINDINGS
CLINICAL AT
The authors developed a prediction model to predict pretest probability of abnormal pH in patients with refractory heartburn and a new scoring system to help determine pre-test probability of abnormal pH in those with suspected extra-esophageal reflux symptoms.
Methods The study was performed in accordance with the Declaration of Helsinki, Good Clinical Practice, and applicable regulatory requirements. The institutional review board approved this clinical trial (number 090872).
LIMITATIONS The recommended percent time with pH < 4 of >12% needs to be prospectively evaluated to determine if this level of reflux consistently results in symptom resolution following fundoplication. IMPACT This externally validated model can be used by clinicians at the bedside employing simple clinical parameters to help determine whether to perform reflux testing OFF or ON therapy.
necessary use of pharmaceuticals (proton pump inhibitor [PPI] therapy), multiple suboptimal diagnostic procedures, and numerous subspecialty referrals.6 The diagnosis of GERD is usually based on symptoms in association with response to acid-suppressive therapies as evidence for the presence or absence of GERD.7 However, heartburn and regurgitation are neither sensitive nor specific for GERD, with sensitivity of 46%–73% and specificity of 45%–58% for patients without esophagitis.8 A therapeutic trial of PPI is similar, with sensitivity and specificity of 78% and 54%, respectively.9 Thus, patients who have refractory esophageal or extraesophageal symptoms despite optimization of PPI therapy and normal endoscopy results (to exclude non–reflux esophageal disorders) often undergo prolonged wireless pH monitoring or multichannel intraluminal impedance with pH (MII-pH) study. The current American College of Gastroenterology (ACG) guidelines recommend that patients with low pretest probability of GERD be tested off medication with pH or MII-pH and that those with high pretest probability of GERD be tested with MII-pH on therapy.10 In the absence of data, expert consensus defined parameters for pretest probability of GERD. Specifically, experts deemed that patients with typical symptoms or partial response to PPI had high pretest probability of reflux, whereas patients with atypical presentations without concomitant typical GERD symptoms had low pretest probability.11 However, there are no studies to date that have evaluated what constitutes low or high pretest probability in this population, creating wide variability in the choice of reflux monitoring and lack of consensus on whether to conduct the test on or off therapy. The aims of this study were to (1) identify clinically significant predictors of abnormal pH (demographic, endoscopic, and pH characteristics) using a large cohort of patients with refractory heartburn or extraesophageal
Study Design and Patient Population The study population consisted of adult patients (18 years of age) with refractory heartburn or suspected EER symptoms (asthma, cough, or hoarseness) referred to the Esophageal Motility Center at the Vanderbilt University Medical Center (Nashville, TN) and had failed > 8 week course of high dose (twice-daily) PPI therapy. Excluded were patients with a history of fundoplication, radiation, or malignancy. The external validation cohort obtained from the Mayo Clinic (Rochester, MN) and the Washington University in St Louis (St Louis, MO) used these same inclusion and exclusion criteria. All patients underwent esophagogastroduodenoscopy (EGD) and 48-hour wireless pH monitoring by a single provider (M.F.V.; described below). Data were prospectively collected and included patient characteristics (age, sex, race, and body mass index [BMI]), presence of secondary GERD symptoms (heartburn with or without regurgitation) in patients with primary extraesophageal symptoms, and current acidsuppressive medication regimen (or no treatment). The presence and size of a hiatal hernia and the presence and severity of esophagitis (graded by the Los Angeles Classification as A, B, C, or D) were determined with endoscopy.12
Wireless pH Monitoring Patients stopped taking PPIs and H2-receptor antagonists at least 7 days before undergoing 48-hour wireless ambulatory pH monitoring (Given Imaging, Duluth, GA). Patients underwent EGD for visual anatomic inspection (for presence and size of hiatal hernia and evaluation of esophageal mucosal integrity for esophagitis) and distance measurements from the incisors to the squamocolumnar junction. Wireless capsules were calibrated by submersion in buffer solutions at pH 7.0 and 1.0 and then activated by magnet removal. Capsules were placed 6 cm above the squamocolumnar junction and attached with vacuum suction using the manufacturer’s delivery system. Placement was confirmed with endoscopy. Subsequently, patients wore wireless pH recorders on their waists or kept them within 3–5 feet at all times. Recording devices received pH data sampling transmitted by the capsule at 433 Hz with 6-second sampling intervals during the 48-hour study. Patients were instructed to perform their normal daily activities and dietary practices.13 After study completion, data were downloaded to dedicated computers. Patient-recorded symptoms were entered into the computer-based record. Measurements of total, upright, and supine percentage of time when esophageal pH was below 4 were determined over days 1 and 2 of the wireless study. Total
Clinical Predictors for Refractory Heartburn and EER 1731
CLINICAL AT
December 2018
Figure 1. Flow diagram of patient enrollment.
acid exposure time (percentage of total time with pH <4) greater than 5.5% (mild reflux) and 12% (moderate to severe reflux) were used as cutoffs based on the average of 2 days.14 A 12% cutoff was used to define moderate to severe reflux based on prior surgical data showing that 12% predicted positive fundoplication outcomes at 1 year in patients with EER symptoms.15 This is the only currently available study in the literature that has evaluated acid exposure time and surgical outcome in this population.
Statistical Analysis There was strict control and supervision of the data entry and access for this study and all data were collected and stored at the secure web-based Vanderbilt Digestive Disease Center REDCap (Research Electronic Data Capture; 1 UL1 RR024975 NCRR/NIH). Categorical variables were summarized using percentages and continuous variables were summarized using median and interquartile range (25th–75th percentile). Wilcoxon rank sum and Pearson c2 tests were used to compare continuous and categorical variables between groups, respectively. Separate multivariable proportional odds ordinal logistic regression models were developed for heartburn or EER symptoms to determine factors associated with abnormal pH. The 2 models were constructed by (1) pre-specifying a set of easily obtained demographic and clinical predictors, (2) fitting a multivariable model with these predictors, and (3) using a single backward selection step to remove insignificant predictors (P < .05) to simplify the model. For the heartburn model, the predictors age, BMI, hiatal hernia, and regurgitation were initially inputted. The EER model mirrored the heartburn model except for
the addition of cough or hoarseness and asthma. Age and regurgitation were not significant in either model and were excluded. The model was based on continuous variables and then simplified to a scoring system that could be used at bedside. BMI was modeled flexibly as a continuous variable in the initial predictive model. We subsequently used a cutoff of BMI >25 kg/m2 in the final model because this corresponded to a similar percentage of increase in the probability of abnormal pH as the other binary predictors in the scoring system (heartburn and asthma). All analyses were conducted using the R statistical program using principles for reproducible research.16 We subsequently validated the performance of the prediction model using an external dataset from 2 tertiary care medical centers (Mayo Clinic and Washington University in St Louis).
Results Discovery Cohort Demographics and Endoscopic Findings In total, 471 patients with suspected GERD-related refractory symptoms underwent EGD and 48-hour wireless pH monitoring off PPI therapy. Figure 1 shows the flow diagram for patient enrollment. Of these, 214 of 471 (45%) presented with a chief complaint of refractory heartburn, whereas 257 of 471 (55%) had EER symptoms (60% chronic cough, 18% hoarseness, and 22% asthma). Overall, 30.3% of patients had normal pH and EGD (pH/E) results, 51.4% had abnormal pH but normal EGD (pHþ/E) result, and 18.3% had endoscopic evidence of esophagitis
1732 Patel et al
Gastroenterology Vol. 155, No. 6
Table 1.Baseline Demographics of Patients With Primary Complaint of Heartburn or EER Symptom (Cough, Hoarseness, or Asthma)
CLINICAL AT
Characteristics
Heartburn (n ¼ 214)
EER (n ¼ 257)
P value
Age (y), median (IQR) BMI (kg/m2), median (IQR) Underweight (16–18.5), % Normal (18.5–25), % Overweight (25–30), % Obese (30–60), % Women, % Ethnicity, % White Black Asian Presence of esophagitis, % A or B C or D Presence of hiatal hernia (>2 cm), % Current medical treatment, % No treatment PPI once a day PPI twice a day H2 blocker PPI þ H2 blocker pH parameters Total percentage of time with pH <4, median (IQR) Upright percentage of time with pH <4, median (IQR) Supine percentage of time with pH <4, median (IQR) Total time >5.5% with pH <4, % Total time >12% with pH <4, %
51 (38–59) 27 (24–33) 2 28 32 38 69
55 (46–62) 28 (25–34) 1 25 33 41 74
<.01a .16a
89 9 2 16 80 20 19
93 5 1 20 88 12 22
5 39 42 2 12
9 18 66 2 5
6.8 (2.4–11.6) 7.3 (2.5–13.7) 2.5 (0.3–8.1) 56 23
6.8 (3.9–11.3) 8.8 (4.8–14.7) 2.0 (0.2–6.3) 63 22
.21b .33b
.33b
.41b <.01a
.30a .07a .46a .15b .68b
IQR, interquartile range. a By Wilcoxon test. b By Pearson c2 test
(Eþ; 85% with A or B and 15% with C or D). A hiatal hernia (>2 cm) was present in 21% of the cohort. Table 1 presents baseline covariates stratified by the presenting complaint (heartburn or suspected EER). Most patients (69% and 74%) in the 2 groups were women (P ¼ .21). Patients with suspected EER were older (median age 55 years; interquartile range 46–62) compared with patients with refractory heartburn (P < .01). Median BMI was 27 kg/m2 (interquartile range 24–33) in the heartburn group and 28 kg/m2 (25–34) in the EER group (P ¼ .16). There were no differences between the 2 groups for ethnicity, presence of esophagitis, hiatal hernia, or baseline percentage of time with pH <4 (Table 1). All patients had a previous failed 8-week course of high-dose (twice daily) PPI, but patients with refractory heartburn were more likely to be on once-daily PPI therapy, whereas patients with EER were more likely to be on twice-daily PPI therapy on presentation to our clinic (P < .01).
Predictors of Esophageal Acid Exposure Refractory Heartburn. In the multivariable model, BMI and presence of hiatal hernia were associated with abnormal pH in patients with refractory heartburn (Table 2), whereas age and concomitant presence of
regurgitation were not. The estimated adjusted odds ratio (OR) for an increase in BMI from 25 to 34 kg/m2 was 1.8 (95% confidence interval [CI] 1.2–2.5) and the OR for presence of hiatal hernia was 2.7 (95% CI 1.5–4.8). Extraesophageal Reflux. We found that the type of EER complaint was associated with abnormal pH. Patients with asthma had twice the odds of abnormal pH compared with those with cough or hoarseness (OR 2.0; 95% CI 1.2– 3.5; Table 2). Odds of abnormal pH was associated with the presence of hiatal hernia (OR 1.9; 95% CI 1.2–3.1), concomitant heartburn (OR 2.0; 95% CI 1.3–3.1), and increasing BMI from 25 to 34 kg/m2 (OR 2.1; 95% CI 1.5– 3.1). As in patients with refractory heartburn, regurgitation was not significantly associated with abnormal pH in the EER group (OR 1.0; 95% CI 0.4–2.4) in the multivariable model.
Model to Predict Abnormal Acid Reflux Refractory Heartburn. Probabilities of any abnormal pH (time >5.5% with pH <4) and having moderate to severe reflux (time >12% with pH <4) stratified by BMI and hiatal hernia are shown in Figures 2A and B, respectively. The probability of abnormal pH increased with higher BMI and presence of hiatal hernia (dichotomous). Probability
Clinical Predictors for Refractory Heartburn and EER 1733
increased nonlinearly for BMI, increasing from normal to overweight BMI and plateauing at an obesity level. Being overweight (BMI 25 kg/m2) with heartburn and hiatal hernia conferred a 70% pretest probability of abnormal pH (>5.5%) and a 33% pretest probability indicated this reflux was moderate to severe (>12%). Extraesophageal Reflux. The probability of abnormal acid exposure in those presenting with suspected EER symptoms was dependent on BMI, presence of hiatal hernia, presence of concomitant heartburn, and modified based on the specific presenting symptom (Figure 3). Increased probability was most pronounced in the transition from normal to overweight BMI and plateaued at the obesity level. Our model estimated that patients presenting with cough or hoarseness and no other risk factors with a normal BMI (20–25 kg/m2) had a 25%–40% probability of having an abnormal pH test result, whereas those who were overweight or obese (BMI >25 kg/m2) had a 40%–60% probability of having an abnormal pH using the cutoff of >5.5% (Figure 3A). Presence of concomitant heartburn and hiatal hernia increased this probability in a graded fashion. Patients whose chief complaint was asthma had a higher baseline probability of abnormal pH than those with isolated cough or hoarseness (Figure 3B). Patients with asthma alone and normal BMI had a 38%–60% probability
Probftotalavg 12 j Xg ¼
of having an abnormal pH test result, whereas those with asthma combined with a BMI >25 kg/m2 had a 60%–75% probability. Asthma and concomitant heartburn increased the probability of abnormal pH to 55%–75% in patients with normal BMI and 75%–85% in overweight or obese patients. Addition of hiatal hernia to this group (asthma þ concomitant heartburn þ hiatal hernia) further increased the probability to 70%–85% in patients with normal BMI and to 85%–92% in overweight or obese patients. The baseline probability of moderate to severe acid reflux (time >12% with pH <4) in patients with isolated cough or hoarseness decreased to 4%–8% with normal BMI and to 8%–15% in overweight or obese patients (Figure 3C). Probability was 23%–40% in patients with asthma þ concomitant heartburn þ hiatal hernia who had normal BMI and 40%–58% if they were overweight or obese (Figure 3D).
Scoring System for Moderate to Severe Reflux Refractory Heartburn. Equation 1 presents the clinical model to predict moderate to severe acid reflux (total time >12% with pH <4) using BMI as a continuous variable and hiatal hernia as a dichotomous variable (if present, use 1; otherwise, use 0):
1
; where 1 þ expð6:334065 X b bÞ
Xb b ¼ þ 1:075438 hh þ 0:1820395bmi 0:0006327213ðbmi 21:2Þ3þ þ 0:001041354ðbmi 27:4Þ3þ 0:0004086325ðbmi 37Þ3þ and ðxÞþ ¼ x if x > 0, 0 otherwise
Table 2.Association of Clinical Variables With Odds of Higher Esophageal Acid Exposure in Patients With Heartburn or EER Groups
Clinical variable
Heartburn
Hiatal hernia BMI (25–34) Concomitant regurgitation Age (40–60) Concomitant heartburn BMI (25–34) Asthma Hiatal hernia Concomitant regurgitation Age
EER
OR (95% CI) 2.7 1.8 1.1 1.3 2.0 2.1 2.0 1.9 1.0 0.8
(1.5–4.8) (1.2–2.5) (0.7–1.9) (0.9–1.9) (1.3–3.1) (1.5–3.1) (1.2–3.5) (1.2–3.1) (0.4–2.4) (0.6–1.1)
Given the complexity of the model, the model shown in Figure 2B can be easily used at bedside to determine the probability of abnormal pH at various BMI cutoffs with or without hiatal hernia. Furthermore, as noted in the figures, the relation between BMI and esophageal acid exposure is not linear, thereby making the analysis of a continuous variable only valuable on the left-hand side of the curve, where the predictability is not as high. Because of failed twice-daily PPI therapy in this population, testing should focus on identifying patients with moderate to severe reflux. Thus, based on our model, all patients in this group are considered to have low pretest probability for reflux and should undergo pH testing off therapy. Extraesophageal Reflux. Equation 2 shows the clinical models for prediction of moderate to severe acid reflux
CLINICAL AT
December 2018
1734 Patel et al
Gastroenterology Vol. 155, No. 6
in patients with EER. Equation 2A uses BMI as a continuous variable (without hiatal hernia):
Probftotalavg 12 j Xg ¼
undergo pH or MII-pH off therapy to rule out GERD as the potential contributor of symptoms. We compared the
1
; where 1 þ expð6:552082 X b bÞ
Xb b ¼ þ 0:1751225bmi 0:0006940719ðbmi 22:7Þ3þ þ 0:001058327ðbmi 28:2Þ3þ CLINICAL AT
0:0003642553ðbmi 38:68Þ3þ þ 0:70117 hb þ 0:6509617 asthma and ððxÞþ ¼ xÞifðx > 0Þ; 0 otherwise
Equation 2B shows a model with the addition of hiatal hernia (dichotomous) if that information is available clinically:
Probftotalavg 12 j Xg ¼
simplified HAs-BEER model with the continuous models with BMI and hiatal hernia size and they performed similarly (Supplemental Table 1).
1 1 þ expð6:484742 X b bÞ
; where
Xb b ¼ þ 0:6555836 hh þ 0:1617812bmi 0:000626957ðbmi 22:7Þ3þ þ 0:0009559897ðbmi 28:2Þ3þ 0:0003290328ðbmi 38:68Þ3þ þ 0:6860861 hb þ 0:7113895 asthma and ððxÞþ ¼ xÞifðx > 0Þ; 0 otherwise
Because of the complexity of the clinical models, we developed a new simple scoring system using probability data (HAs-BEER) to help clinicians estimate pretest probability of abnormal pH at bedside based on clinical history in patients with EER. The acronym HAs-BEER stands for Heartburn, Asthma, and BMI (>25) in EER. One point is earned for the presence of each of the following clinical parameters: concomitant heartburn, asthma, and BMI >25 kg/m2. No points are earned for cough and hoarseness without the other clinical parameters. BMI >25 kg/m2 was used as a cutoff because this corresponded to a similar percentage of increase in the probability of abnormal pH as the other binary predictors in the scoring system (heartburn and asthma; Table 2). A score of 3 had a positive predictive value of 92% (sensitivity 97.8%) for abnormal pH defined using the traditional cutoff of 5.5%. Given the high-pretest probability of reflux in this group, these patients should undergo MII-pH on therapy according to ACG guidelines. We made our model more specific by using a pH cutoff >12% with a HAs-BEER score 2 having 80% negative predictive value for moderate to severe reflux. Thus, these patients should
External Validation. Validation of the proposed model was performed in an external cohort from 2 different tertiary care medical centers (62% women; median age 59 years). The derivation cohort had a higher median total percentage of time with pH <4 (6.8 vs 3.9; P < .01) and therefore more patients (63% vs 42%; P < .01) with abnormal pH than the validation cohort (Table 3). There were no differences in the number of patients with moderate to severe acid reflux between cohorts. The validation cohort included patients with cough (62%), hoarseness (12%), and asthma (17%). Despite the noted baseline demographic differences between the derivation and validation cohorts (Table 3), the HAs-BEER score 2 had a negative predictive value for abnormal pH (defined as >12%) of 85% in the external cohort.
Discussion Reflux monitoring in patients with refractory heartburn or suspected EER-associated symptoms is widely variable because of the lack of consensus on definition, gold standard test, and therapeutic implications of a positive result.
Figure 2. Probability of abnormal pH in patients with refractory heartburn stratified by BMI and hiatal hernia with cutoff of abnormal pH defined as total time (A) >5.5% and (B) >12% with pH <4. HH, hiatal hernia.
Uncertainty has resulted in significant and rising management costs with serious implications for the patient with persistent reflux symptoms on PPI therapy.17 Current ACG guidelines recommend differential testing depending on the pretest probability of reflux, the criteria for which were left nebulous by the expert consensus panel. To date, guidance is lacking as to what constitutes low or high pretest probability in those patients with failed PPI trials and normal EGD findings.10 This study aimed to fill this important knowledge gap by identifying easily accessible clinical variables that clinicians can use to estimate the pretest probability of reflux in those with refractory heartburn or EER-associated symptoms using traditional pH or moderate to severe reflux cutoffs. Moderate to severe reflux was chosen as an additional target based on the only available prior surgical outcome data, which have been reported to have a more reliable treatment response to antireflux surgeries.15 We evaluated a large cohort of patients with refractory heartburn or suspected EER-associated symptoms (cough,
Clinical Predictors for Refractory Heartburn and EER 1735
hoarseness, or asthma) to determine independent predictors of abnormal pH and created a model to estimate pretest probability of abnormal pH. We found that BMI and presence of hiatal hernia in patients with refractory heartburn were independently associated with abnormal pH. Similar to previous studies, our study highlights the impact of obesity in esophageal acid exposure and as an independent risk factor for GERD symptoms and erosive esophagitis.18–21 As presented in Figure 2A, our model shows that the pretest probability of abnormal esophageal acid exposure using the traditional cutoff of 5.5% is too high in this group of patients in whom PPI therapy has failed. Thus, the primary objective in patients with refractory heartburn is to evaluate for moderate to severe reflux (cutoff 12%). Being overweight (BMI 25 kg/m2) with heartburn and hiatal hernia confers a 70% pretest probability of abnormal pH (>5.5%) and a 33% pretest probability that this reflux is moderate to severe (>12%). As presented in Figure 2B, the pretest probabilities of moderate to severe reflux are significantly lower. Thus, based on our model, we consider all patients with refractory heartburn and failed twice-daily PPI therapy as having low pretest probability for moderate to severe reflux and recommend pH testing off therapy. If they do not have moderate to severe reflux, then it is unlikely that reflux is the primary etiology of their ongoing symptoms (refractory heartburn) given prior failure of twice-daily PPI therapy and focus should be shifted toward identifying other etiologies (such as hypersensitive esophagus). Furthermore, we extend these observations to show that in patients with suspected EER symptoms, esophageal acid exposure is dependent on their chief complaint or presenting symptoms in addition to other clinical factors. We found that abnormal esophageal acid exposure (total time >5.5% with pH <4) was present in 56% of patients with heartburn and 63% of patients with EER (P ¼ .15). Although this difference was nonsignificant between the 2 groups, the higher percentage of abnormal acid exposure in patients with EER was due primarily to those patients presenting with asthma. Within EER groups, abnormal pH was significantly more prevalent (80%) in patients with asthma compared with cough (60%) or hoarseness (51%; P < .01). The derivation cohort also had a significantly higher prevalence of abnormal esophageal acid exposure (total time >5.5% with pH <4) at 63% compared with the external validation cohort (42%; P < .01), but this is likely due to the larger percentage of obese (BMI 30–60 kg/m2) patients in our derivation cohort (41% vs 32%). The prevalence of abnormal esophageal acid exposure in our patients with asthma is similar to previous studies showing a large percentage of patients with asthma having abnormal acid exposure ranging from 62% to 80%.22–26 Furthermore, we found that probability of abnormal esophageal acid exposure in patients with suspected EER is not uniform across presenting symptoms. In this group, presence of asthma, hiatal hernia, concomitant heartburn, and BMI increased the probability of abnormal pH by 2-fold
CLINICAL AT
December 2018
1736 Patel et al
Gastroenterology Vol. 155, No. 6
CLINICAL AT
Figure 3. Probability of abnormal pH in patients with EER stratified by presenting symptom and additive contributions of hiatal hernia and heartburn at various BMIs with cutoff of abnormal pH defined as total time >5.5% with pH <4 (A, cough or hoarseness; B, asthma) and total time >12% with pH <4 (C, cough or hoarseness; D, asthma). HB, heartburn; HH, hiatal hernia.
Table 3.Comparison of Baseline Demographics Between Derivation and Validation (External) Cohorts for EER (Cough, Hoarseness, or Asthma) Characteristics
Derivation cohort (n ¼ 257)
Validation cohort (n ¼ 118)
P value
Age (y), median (IQR) BMI (kg/m2), median (IQR) Underweight (16–18.5), % Normal (18.5–25), % Overweight (25–30), % Obese (30–60), % Women, % Chief complaint, % Asthma Cough Hoarseness None Secondary heartburn, % pH parameters Total percentage of time with pH <4, median (IQR) Upright percentage of time with pH <4, median (IQR) Supine percentage of time with pH <4, median (IQR) Total time >5.5% with pH <4, % Total time >12% with pH <4, % HAs-BEER score, % 0 1 2 3
55 (46–62) 28 (25–34) 1 25 33 41 74
59 (49–68) 28 (24–31) 1 29 38 32 62
<.01a .39a
21 60 9 0 45
17 62 12 9 26
<.01b
6.8 (3.9–11.3) 8.8 (4.8–14.7) 2.1 (0.2–6.3) 63 22
3.9 (1.3–8.2) 4.4 (1.9–10.5) 0.8 (0–5.1) 42 14
<.01a <.01a .02a <.01b .06b
13 42 35 10
20 56 18 7
<.01b
IQR, interquartile range. a By Wilcoxon test. b By Pearson c2 test.
.01b <.01b
Clinical Predictors for Refractory Heartburn and EER 1737
CLINICAL AT
December 2018
Figure 4. Proposed algorithm for testing patients with refractory heartburn or EER (asthma, cough, and hoarseness) for moderate to severe reflux (total time >12% with pH <4) based on the clinical model.
compared with those with isolated cough or hoarseness. In other words, the likelihood of a positive reflux test result is lower for a patient presenting with cough and/or hoarseness without concomitant heartburn than for a patient presenting with asthma. We also confirmed prior findings showing that being overweight and obese significantly contributed to the likelihood of increased esophageal acid exposure, with a substantial increase from normal BMI to overweight than from overweight to obese.27 There are several important and clinically relevant observations from our study. (1) There is no such thing as low probability of reflux as outlined by the ACG guidelines. Patients with the lowest probability, those with refractory
heartburn (Figure 2A) or cough or hoarseness (Figure 3A) and an average BMI (27 kg/m2 in this study cohort), had 50% baseline likelihood of reflux. Probability of positive test results increased in the presence of other clinical factors such as hiatal hernia, concomitant heartburn, and presence of asthma in the EER group (Figure 3A and B). (2) Not all EER presenting symptoms had the same pretest probability for abnormal esophageal acid exposure; that is, patients with asthma had higher odds of having abnormal esophageal acid exposure compared with those presenting with cough and/or hoarseness. In fact, most patients with asthma had abnormal esophageal acid exposure (abnormal ¼ time 5.5% with pH <4). Thus, in this population, does it make
1738 Patel et al
CLINICAL AT
sense to order a test that will shows a positive result in most patients? (3) We suggest using a different metric of abnormality (time >12% with pH <4 or moderate or severe reflux) for assessing degree of reflux in patients with refractory heartburn or EER symptoms. This is based on data indicating that time >12% with pH <4 was an independent predictor of EER symptom response to surgical fundoplication.15 As such, we developed a scoring system (HAsBEER) to determine which patients with refractory EER symptoms should undergo testing off or on PPI therapy based on probability of baseline moderate to severe esophageal acid exposure. Our model for refractory heartburn (Figure 2) or EER (Figure 3) is a useful bedside tool to determine pretest probability of abnormal pH based on readily available clinical parameters during a clinic visit. In this novel scoring system (HAs-BEER) based on clinical history, patients earn 1 point for each clinical parameter present (heartburn, asthma, or BMI >25 kg/m2). A score of 3 indicates high pretest probability of having moderate to severe reflux, which suggests these patients should undergo MII-pH testing on therapy to evaluate for non–acid reflux as a cause of persistent symptoms. Based on our cohort, this would indicate MII-pH testing on therapy in 9.7% of patients. In contrast, those with a score 2 have low pretest probability and should undergo pH or MII-pH testing off therapy to rule out GERD as the potential contributor of symptoms. Furthermore, we externally validated this model using a cohort of patients with EER from 2 separate institutions with a HAs-BEER score 2 having 85% negative predictive value for abnormal reflux (defined as >12%). This external cohort consisted of patients with cough or hoarseness and asthma. Thus, the present study provides a roadmap to define low and high probability, which contrasts with the current clinical approach. These findings further suggest that the traditional cutoff of >5.5% for total percentage of time with pH <4 is too low to exclude any patient groups from pH testing. Based on our model, a HAs-BEER score of 3 had 92% positive predictive value for abnormal pH using the traditional cutoff (>5.5%). Because PPI use has failed in these patients, the therapeutic implications of a positive pH study result based on traditional parameters is unclear and could result in referral for surgical fundoplication. Figure 4 shows our proposed algorithm for testing patients with refractory heartburn or EER (asthma, cough, and hoarseness) for moderate to severe reflux (total time >12% with pH <4) based on the clinical models. In addition, it should be noted that despite the significant baseline demographic and pH monitoring differences between the derivation and validation cohorts, our model performed very well in the external validation. This study has some limitations. (1) Validated instruments were not used to determine symptom severity at presentation. However, we aimed to make a model that can be easily used clinically; because symptom questionnaires are not used in routine daily clinical practice, we used the presenting chief complaint to develop our model and externally validated it to demonstrate its predictive
Gastroenterology Vol. 155, No. 6
accuracy. Our model also was developed with a robust sample size and fills a critical gap in the current ACG guidelines. Specifically, it provides a clinical algorithm to estimate the pretest probability of positive pH test results in patients with refractory heartburn or EER. (2) The recommended time >12% with pH <4 needs to be prospectively evaluated to see whether this level of reflux consistently results in symptom resolution after fundoplication. However, we use this cutoff for moderate to severe acid reflux based on the only surgical outcome data study currently available.15 Based on these data, we believe that this higher physiologic cutoff ensures that clinicians judiciously select patients with refractory symptoms who might benefit from surgical fundoplication. (3) We acknowledge that our cohort is based on a singlecenter experience with potential risk of referral bias, but the large number of patients included in this cohort minimizes this bias, as does external validation from 2 different tertiary care medical centers. Furthermore, results from this study will serve as the important backbone to pursue further multicenter prospective validation studies. (4) It is important to recognize that our findings are limited to patients in the United States because the BMI value might not be applicable for the Asian or European population. (5) Although we recognize that hiatal hernia is an important predictor of abnormal esophageal acid exposure in patients with suspected EER (as shown in Figure 3), we did not include it in our HAs-BEER score because we aimed to develop a predictive score that can be used at bedside by clinicians and presence or size of hiatal hernia is not usually available to clinicians at bedside. We addressed this by providing Equations 1 and 2A and 2B that show clinical models with hiatal hernia (if that information is available clinically). We also included Supplemental Table 1, which shows that predicted probabilities of moderate to severe reflux (total time >12% with pH <4) from the simplified HAs-BEER model (using BMI cutoff >25), the HAs-BEER model with BMI as a continuous variable (model 2), the HAs-BEER model with hiatal hernia (model 3), and the model with BMI and hiatal hernia as continuous variables (model 4) perform similarly. In summary, choice of reflux testing in patients with refractory heartburn and EER-associated symptoms (cough, hoarseness, and asthma) is currently fraught with confusion, which primarily derives from a lack of understanding of what constitutes high or low pretest probability of abnormal pH. We found abnormal esophageal acid exposure (defined as total time >5.5% with pH <4) in 56% of patients with refractory heartburn and 63% of patients with EER. The present data suggest the traditional cutoff value is too sensitive and not specific enough, especially in those with EER. Our novel bedside prediction tool can help determine pretest probability of abnormal pH in patients with failed PPI therapy and can help guide clinicians to determine the best diagnostic approach in this challenging group of patients and decrease unnecessary testing.
Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at https://doi.org/10.1053/j. gastro.2018.08.038.
References 1. Sontaj SJ. The medical management of reflux esophagitis. Role of antacids and acid inhibition. Gastroenterol Clin North Am 1990;19:683–712. 2. Dent J, El-Serag HB, Wallander MA, et al. Epidemiology of gastro-oesophageal reflux disease: a systematic review. Gut 2005;54:710–717. 3. Fenter TC, Naslund MJ, Shah MB, et al. The cost of treating the 10 most prevalent diseases in men 50 years of age or older. Am J Manag Care 2006;12:S90–S98. 4. Sandler RS, Everhart JE, Donowitz M, et al. The burden of selected digestive diseases in the United States. Gastroenterology 2002;122:1500–1511. 5. Everhart JE, Ruhl CE. Burden of digestive diseases in the United States part I: overall and upper gastrointestinal diseases. Gastroenterology 2009;136:376–386. 6. Francis DO, Rymer JA, Slaughter JC, et al. High economic burden of caring for patients with suspected extraesophageal reflux. Am J Gastroenterol 2013; 108:905–911. 7. Vakil N, van Zanten SV, Kahrilas P, et al. The Montreal definition and classification of gastroesophageal reflux disease: a global evidence-based consensus. Am J Gastroenterol 2006;101:1900–1920; quiz 1943. 8. Klauser AG, Schindlbeck NE, Muller-Lissner SA. Symptoms in gastro-oesophageal reflux disease. Lancet 1990; 335:205–208. 9. Numans ME, Lau J, de Wit NJ, et al. Short-term treatment with proton-pump inhibitors as a test for gastroesophageal reflux disease: a meta-analysis of diagnostic test characteristics. Ann Intern Med 2004; 140:518–527. 10. Katz PO, Gerson LB, Vela MF. Guidelines for the diagnosis and management of gastroesophageal reflux disease. Am J Gastroenterol 2013;108:308–328; quiz 329. 11. Roman S, Gyawali CP, Savarino E, et al. Ambulatory reflux monitoring for diagnosis of gastro-esophageal reflux disease: update of the Porto consensus and recommendations from an international consensus group. Neurogastroenterol Motil 2017;29:1–15. 12. Lundell LR, Dent J, Bennett JR, et al. Endoscopic assessment of oesophagitis: clinical and functional correlates and further validation of the Los Angeles classification. Gut 1999;45:172–180. 13. Kavitt RT, Yuksel ES, Slaughter JC, et al. The role of impedance monitoring in patients with extraesophageal symptoms. Laryngoscope 2013;123: 2463–2468. 14. Pandolfino JE, Richter JE, Ours T, et al. Ambulatory esophageal pH monitoring using a wireless system. Am J Gastroenterol 2003;98:740–749.
Clinical Predictors for Refractory Heartburn and EER 1739 15. Francis DO, Goutte M, Slaughter JC, et al. Traditional reflux parameters and not impedance monitoring predict outcome after fundoplication in extraesophageal reflux. Laryngoscope 2011;121:1902–1909. 16. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria;(2014). Available at: http:// www.R-project.org/. Accessed November 13, 2008. 17. Becher A, El-Serag H. Systematic review: the association between symptomatic response to proton pump inhibitors and health-related quality of life in patients with gastro-oesophageal reflux disease. Aliment Pharmacol Ther 2011;34:618–627. 18. El-Serag HB, Ergun GA, Pandolfino J, et al. Obesity increases oesophageal acid exposure. Gut 2007;56:749– 755. 19. Crowell MD, Bradley A, Hansel S, et al. Obesity is associated with increased 48-h esophageal acid exposure in patients with symptomatic gastroesophageal reflux. Am J Gastroenterol 2009;104:553–559. 20. El-Serag HB, Graham DY, Satia JA, et al. Obesity is an independent risk factor for GERD symptoms and erosive esophagitis. Am J Gastroenterol 2005;100: 1243–1250. 21. Hampel H, Abraham NS, El-Serag HB. Meta-analysis: obesity and the risk for gastroesophageal reflux disease and its complications. Ann Intern Med 2005;143: 199–211. 22. Harding SM, Guzzo MR, Richter JE. The prevalence of gastroesophageal reflux in asthma patients without reflux symptoms. Am J Respir Crit Care Med 2000;162:34–39. 23. Sontag SJ, O’Connell S, Khandelwal S, et al. Most asthmatics have gastroesophageal reflux with or without bronchodilator therapy. Gastroenterology 1990;99:613– 620. 24. Bucknall C, Stanton A, Miller G, et al. The impact of normalization of esophageal acid profile by incremental protein pump inhibitors dosing in difficult asthma patients with proven gastro-esophageal acid reflux. J Asthma 2009;46:506–511. 25. DeMeester TR, Bonavina L, Iascone C, et al. Chronic respiratory symptoms and occult gastroesophageal reflux. A prospective clinical study and results of surgical therapy. Ann Surg 1990;211:337–345. 26. Alhabib KF, Vedal S, Champion P, et al. The utility of ambulatory pH monitoring in patients presenting with chronic cough and asthma. Can J Gastroenterol 2007; 21:159–163. 27. Aslam M, Slaughter JC, Goutte M, et al. Nonlinear relationship between body mass index and esophageal acid exposure in the extraesophageal manifestations of reflux. Clin Gastroenterol Hepatol 2012; 10:874–878.
Received January 9, 2018. Accepted August 22, 2018. Reprint requests Address requests for reprints to: Michael F. Vaezi, MD, PhD, MSc (Epi), FACG, Professor of Medicine, Clinical Director, Division of Gastroenterology, Hepatology, and Nutrition, Vanderbilt University Medical Center, Nashville, Tennessee 37232. e-mail: [email protected].
CLINICAL AT
December 2018
1740 Patel et al Acknowledgments Author contributions: Dhyanesh A. Patel, C. Gaelyn Garrett, David O. Francis, and Michael F. Vaezi conceived the study. Dhyanesh A. Patel and Michael F. Vaezi designed the study. Dhyanesh A. Patel, Rohit Sharda, Tina Higginbotham, and Michael F. Vaezi acquired the data. Karthik Ravi, Steven Hasak, David Katzka, and C. Prakash Gyawali acquired external data. Dhyanesh A. Patel, Rohit Sharda, Yash A. Choksi, James C. Slaughter, and
Gastroenterology Vol. 155, No. 6 Michael F. Vaezi analyzed and interpreted the data. Dhyanesh A. Patel and James C. Slaughter drafted the manuscript. Dhyanesh A. Patel, James C. Slaughter, and David O. Francis revised the manuscript. Michael F. Vaezi critically reviewed the manuscript and supervised the study. Conflicts of interest The authors disclose on conflicts.
CLINICAL AT
December 2018
Clinical Predictors for Refractory Heartburn and EER 1740.e1
Supplemental Table 1.Predicted Probabilities of Moderate to Severe Reflux (Total % Time pH <4 of >12%) From Simplified HAs-BEER Model (BMI Cutoff >25), HAs-BEER Model With Continuous BMI (Model 2), HAs-BEER Model With Hiatal Hernia (Model 3), and Model With Continuous BMI and Hiatal Hernia (Model 4)a BMI
Heartburn
Hiatal hernia size
Asthma
HAs-BEER score
HAs-BEER probability
Model 2
Model 3
Model 4
25 25 25 25 25 25 25 25 25 30 30 30 25 25 25 30 30 30 30 30 30 30 30 30
0 0 0 0 0 0 1 1 1 0 0 0 1 1 1 0 0 0 1 1 1 1 1 1
0 2 4 0 2 4 0 2 4 0 2 4 0 2 4 0 2 4 0 2 4 0 2 4
0 0 0 1 1 1 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 1 1 1
0 0 0 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 3 3 3
8% 8% 8% 15% 15% 15% 15% 15% 15% 15% 15% 15% 29% 29% 29% 29% 29% 29% 29% 29% 29% 47% 47% 47%
10% 10% 10% 18% 18% 18% 19% 19% 19% 17% 17% 17% 30% 30% 30% 29% 29% 29% 30% 30% 30% 45% 45% 45%
6% 11% 10% 12% 23% 20% 12% 23% 20% 12% 23% 20% 23% 40% 37% 23% 40% 37% 23% 40% 37% 40% 60% 56%
8% 15% 14% 16% 28% 26% 14% 26% 24% 13% 24% 22% 27% 44% 41% 25% 42% 39% 23% 39% 36% 40% 59% 56%
a
Dichotomous variables were classified as 1 (yes) or 0 (no) based on presence of the symptom. Probabilities are given for representative values of BMI, heartburn presence (or absence), hernia size, and asthma.
Model to Select On-Therapy vs Off-Therapy Tests for Patients With Refractory Esophageal or Extraesophageal Symptoms
1
Division of Gastroenterology, Hepatology and Nutrition, Vanderbilt University Medical Center, Nashville, Tennessee; Department of Biostatistics, Vanderbilt University Medical Center, Nashville, Tennessee; 3Vanderbilt Voice Center, Vanderbilt University Medical Center, Nashville, Tennessee; 4Swedish Gastroenterology, Seattle, Washington; 5Division of Otolaryngology and Wisconsin Surgical Outcomes Research Program, University of Wisconsin, Madison, Wisconsin; 6Division of Gastroenterology and Hepatology, Mayo Clinic, Rochester, Minnesota; and 7Division of Gastroenterology, Washington University Medical Center, St Louis, Missouri 2
See Covering the Cover synopsis on page 1661. BACKGROUND & AIMS: It is not clear whether we should test for reflux in patients with refractory heartburn or extraesophageal reflux (EER) symptoms, such as cough, hoarseness, or asthma. Guidelines recommend testing patients by pH monitoring when they are on or off acid-suppressive therapies based on pretest probability of reflux, determined by expert consensus. However, it is not clear what constitutes a low or high pretest probability of reflux in these patients. We aimed to develop a model that clinicians can use at bedside to estimate pretest probability of abnormal reflux. METHODS: We performed a prospective study of 471 adult patients with refractory heartburn (n ¼ 214) or suspected EER symptoms (n ¼ 257) who underwent endoscopy with wireless pH monitoring while they were off acid-suppressive treatment and assigned them to groups based on symptoms at presentation (discovery cohort). Using data from the discovery cohort, we performed proportional odds ordinal logistic regression to select factors (easy to obtain demographic criteria and clinical symptoms such as heartburn, regurgitation, asthma, cough, and hoarseness) associated with esophageal exposure to acid. We validated our findings in a cohort of 118 patients with the same features from 2 separate tertiary care centers (62% women; median age 59 years; 62% with cough as presenting symptom). RESULTS: Abnormal pH (>5.5% of time spent at pH <4) was found in 56% of patients with heartburn and 63% of patients with EER (P ¼ .15). Within EER groups, abnormal pH was detected in a significantly larger proportion (80%) of patients with asthma compared with patients with cough (60%) or hoarseness (51%; P < .01). Factors significantly associated with abnormal pH in patients with heartburn were presence of hiatal hernia and body mass index >25 kg/m2. In patients with EER, the risk of reflux was independently associated with the presence of concomitant heartburn (odds ratio [OR] 2.0; 95% confidence interval [CI] 1.3–3.1), body mass index >25 kg/m2 (OR 2.1; 95% CI 1.5–3.1), asthma (OR 2.0; 95% CI 1.2–3.5), and presence of hiatal hernia (OR 1.9; 95% CI 1.2–3.1). When we used these factors to create a scoring system, we found that a score of 2 excluded patients with moderate to severe reflux, with a negative predictive value of 80% in the discovery cohort and a negative predictive value of 85% in the validation cohort.
CONCLUSION: We developed a clinical model to estimate pretest probability of abnormal pH in patients who were failed by proton pump inhibitor therapy. This system can help guide clinicians at bedside in determining the most appropriate diagnostic test in this challenging group of patients.
Keywords: Body Mass Index; Heartburn, Asthma, and BMI Extraesophageal Reflux (HAs-BEER); Proton Pump Inhibitor; Gastroesophageal Reflux Disease.
A
pproximately 25–75 million people in the United States are affected by gastroesophageal reflux disease (GERD) and 13% of Americans use antireflux medications at least twice weekly.1 GERD is a spectrum of disease classically producing symptoms of heartburn and acid regurgitation, but patients also can present with extraesophageal manifestations such as asthma, cough, hoarseness, and recurrent pneumonia. Because of its variable presentation, GERD remains the leading outpatient physician diagnosis for gastrointestinal disorders in the United States, with a very high economic burden on society.2 The annual direct cost for GERD management is reported to be approximately $9713 per patient, with national expenditures ranging from $9.3 to $12.3 billion.4,5 Patients with symptoms associated with extraesophageal reflux (EER) have even a greater economic burden, with a mean direct cost of $5438 per patient and a national expenditure estimated to be higher than $50 billion.6 This is primarily driven by uncertainty over EER diagnosis because of the lack of a gold standard diagnostic test, which leads to un-
Abbreviations used in this paper: ACG, American College of Gastroenterology; BMI, body mass index; CI, confidence interval; EER, extraesophageal reflux; EGD, esophagogastroduodenoscopy; GERD, gastroesophageal reflux disease; HAs-BEER, Heartburn, Asthma, and BMI Extraesophageal Reflux; MII-pH, multichannel intraluminal impedance with pH; OR, odds ratio; PPI, proton pump inhibitor. Most current article © 2018 by the AGA Institute 0016-5085/$36.00 https://doi.org/10.1053/j.gastro.2018.08.038
CLINICAL AT
Dhyanesh A. Patel,1 Rohit Sharda,4 Yash A. Choksi,1 James C. Slaughter,2 Tina Higginbotham,1 C. Gaelyn Garrett,3 David O. Francis,5 Karthik Ravi,6 Stephen Hasak,7 David Katzka,6 C. Prakash Gyawali,7 and Michael F. Vaezi1
1730 Patel et al
WHAT YOU NEED TO KNOW BACKGROUND AND CONTEXT There are currently no studies that define what constitutes low- or high pre-test probability of reflux in patients with refractory heartburn or extra-esophageal symptoms.
Gastroenterology Vol. 155, No. 6
symptoms (asthma, cough, or hoarseness); (2) develop a simple, predictive model that can be used by clinicians at bedside to determine pretest probability of abnormal pH; and (3) externally validate the predictive model using independent data from a similar population at 2 different institutions.
NEW FINDINGS
CLINICAL AT
The authors developed a prediction model to predict pretest probability of abnormal pH in patients with refractory heartburn and a new scoring system to help determine pre-test probability of abnormal pH in those with suspected extra-esophageal reflux symptoms.
Methods The study was performed in accordance with the Declaration of Helsinki, Good Clinical Practice, and applicable regulatory requirements. The institutional review board approved this clinical trial (number 090872).
LIMITATIONS The recommended percent time with pH < 4 of >12% needs to be prospectively evaluated to determine if this level of reflux consistently results in symptom resolution following fundoplication. IMPACT This externally validated model can be used by clinicians at the bedside employing simple clinical parameters to help determine whether to perform reflux testing OFF or ON therapy.
necessary use of pharmaceuticals (proton pump inhibitor [PPI] therapy), multiple suboptimal diagnostic procedures, and numerous subspecialty referrals.6 The diagnosis of GERD is usually based on symptoms in association with response to acid-suppressive therapies as evidence for the presence or absence of GERD.7 However, heartburn and regurgitation are neither sensitive nor specific for GERD, with sensitivity of 46%–73% and specificity of 45%–58% for patients without esophagitis.8 A therapeutic trial of PPI is similar, with sensitivity and specificity of 78% and 54%, respectively.9 Thus, patients who have refractory esophageal or extraesophageal symptoms despite optimization of PPI therapy and normal endoscopy results (to exclude non–reflux esophageal disorders) often undergo prolonged wireless pH monitoring or multichannel intraluminal impedance with pH (MII-pH) study. The current American College of Gastroenterology (ACG) guidelines recommend that patients with low pretest probability of GERD be tested off medication with pH or MII-pH and that those with high pretest probability of GERD be tested with MII-pH on therapy.10 In the absence of data, expert consensus defined parameters for pretest probability of GERD. Specifically, experts deemed that patients with typical symptoms or partial response to PPI had high pretest probability of reflux, whereas patients with atypical presentations without concomitant typical GERD symptoms had low pretest probability.11 However, there are no studies to date that have evaluated what constitutes low or high pretest probability in this population, creating wide variability in the choice of reflux monitoring and lack of consensus on whether to conduct the test on or off therapy. The aims of this study were to (1) identify clinically significant predictors of abnormal pH (demographic, endoscopic, and pH characteristics) using a large cohort of patients with refractory heartburn or extraesophageal
Study Design and Patient Population The study population consisted of adult patients (18 years of age) with refractory heartburn or suspected EER symptoms (asthma, cough, or hoarseness) referred to the Esophageal Motility Center at the Vanderbilt University Medical Center (Nashville, TN) and had failed > 8 week course of high dose (twice-daily) PPI therapy. Excluded were patients with a history of fundoplication, radiation, or malignancy. The external validation cohort obtained from the Mayo Clinic (Rochester, MN) and the Washington University in St Louis (St Louis, MO) used these same inclusion and exclusion criteria. All patients underwent esophagogastroduodenoscopy (EGD) and 48-hour wireless pH monitoring by a single provider (M.F.V.; described below). Data were prospectively collected and included patient characteristics (age, sex, race, and body mass index [BMI]), presence of secondary GERD symptoms (heartburn with or without regurgitation) in patients with primary extraesophageal symptoms, and current acidsuppressive medication regimen (or no treatment). The presence and size of a hiatal hernia and the presence and severity of esophagitis (graded by the Los Angeles Classification as A, B, C, or D) were determined with endoscopy.12
Wireless pH Monitoring Patients stopped taking PPIs and H2-receptor antagonists at least 7 days before undergoing 48-hour wireless ambulatory pH monitoring (Given Imaging, Duluth, GA). Patients underwent EGD for visual anatomic inspection (for presence and size of hiatal hernia and evaluation of esophageal mucosal integrity for esophagitis) and distance measurements from the incisors to the squamocolumnar junction. Wireless capsules were calibrated by submersion in buffer solutions at pH 7.0 and 1.0 and then activated by magnet removal. Capsules were placed 6 cm above the squamocolumnar junction and attached with vacuum suction using the manufacturer’s delivery system. Placement was confirmed with endoscopy. Subsequently, patients wore wireless pH recorders on their waists or kept them within 3–5 feet at all times. Recording devices received pH data sampling transmitted by the capsule at 433 Hz with 6-second sampling intervals during the 48-hour study. Patients were instructed to perform their normal daily activities and dietary practices.13 After study completion, data were downloaded to dedicated computers. Patient-recorded symptoms were entered into the computer-based record. Measurements of total, upright, and supine percentage of time when esophageal pH was below 4 were determined over days 1 and 2 of the wireless study. Total
Clinical Predictors for Refractory Heartburn and EER 1731
CLINICAL AT
December 2018
Figure 1. Flow diagram of patient enrollment.
acid exposure time (percentage of total time with pH <4) greater than 5.5% (mild reflux) and 12% (moderate to severe reflux) were used as cutoffs based on the average of 2 days.14 A 12% cutoff was used to define moderate to severe reflux based on prior surgical data showing that 12% predicted positive fundoplication outcomes at 1 year in patients with EER symptoms.15 This is the only currently available study in the literature that has evaluated acid exposure time and surgical outcome in this population.
Statistical Analysis There was strict control and supervision of the data entry and access for this study and all data were collected and stored at the secure web-based Vanderbilt Digestive Disease Center REDCap (Research Electronic Data Capture; 1 UL1 RR024975 NCRR/NIH). Categorical variables were summarized using percentages and continuous variables were summarized using median and interquartile range (25th–75th percentile). Wilcoxon rank sum and Pearson c2 tests were used to compare continuous and categorical variables between groups, respectively. Separate multivariable proportional odds ordinal logistic regression models were developed for heartburn or EER symptoms to determine factors associated with abnormal pH. The 2 models were constructed by (1) pre-specifying a set of easily obtained demographic and clinical predictors, (2) fitting a multivariable model with these predictors, and (3) using a single backward selection step to remove insignificant predictors (P < .05) to simplify the model. For the heartburn model, the predictors age, BMI, hiatal hernia, and regurgitation were initially inputted. The EER model mirrored the heartburn model except for
the addition of cough or hoarseness and asthma. Age and regurgitation were not significant in either model and were excluded. The model was based on continuous variables and then simplified to a scoring system that could be used at bedside. BMI was modeled flexibly as a continuous variable in the initial predictive model. We subsequently used a cutoff of BMI >25 kg/m2 in the final model because this corresponded to a similar percentage of increase in the probability of abnormal pH as the other binary predictors in the scoring system (heartburn and asthma). All analyses were conducted using the R statistical program using principles for reproducible research.16 We subsequently validated the performance of the prediction model using an external dataset from 2 tertiary care medical centers (Mayo Clinic and Washington University in St Louis).
Results Discovery Cohort Demographics and Endoscopic Findings In total, 471 patients with suspected GERD-related refractory symptoms underwent EGD and 48-hour wireless pH monitoring off PPI therapy. Figure 1 shows the flow diagram for patient enrollment. Of these, 214 of 471 (45%) presented with a chief complaint of refractory heartburn, whereas 257 of 471 (55%) had EER symptoms (60% chronic cough, 18% hoarseness, and 22% asthma). Overall, 30.3% of patients had normal pH and EGD (pH/E) results, 51.4% had abnormal pH but normal EGD (pHþ/E) result, and 18.3% had endoscopic evidence of esophagitis
1732 Patel et al
Gastroenterology Vol. 155, No. 6
Table 1.Baseline Demographics of Patients With Primary Complaint of Heartburn or EER Symptom (Cough, Hoarseness, or Asthma)
CLINICAL AT
Characteristics
Heartburn (n ¼ 214)
EER (n ¼ 257)
P value
Age (y), median (IQR) BMI (kg/m2), median (IQR) Underweight (16–18.5), % Normal (18.5–25), % Overweight (25–30), % Obese (30–60), % Women, % Ethnicity, % White Black Asian Presence of esophagitis, % A or B C or D Presence of hiatal hernia (>2 cm), % Current medical treatment, % No treatment PPI once a day PPI twice a day H2 blocker PPI þ H2 blocker pH parameters Total percentage of time with pH <4, median (IQR) Upright percentage of time with pH <4, median (IQR) Supine percentage of time with pH <4, median (IQR) Total time >5.5% with pH <4, % Total time >12% with pH <4, %
51 (38–59) 27 (24–33) 2 28 32 38 69
55 (46–62) 28 (25–34) 1 25 33 41 74
<.01a .16a
89 9 2 16 80 20 19
93 5 1 20 88 12 22
5 39 42 2 12
9 18 66 2 5
6.8 (2.4–11.6) 7.3 (2.5–13.7) 2.5 (0.3–8.1) 56 23
6.8 (3.9–11.3) 8.8 (4.8–14.7) 2.0 (0.2–6.3) 63 22
.21b .33b
.33b
.41b <.01a
.30a .07a .46a .15b .68b
IQR, interquartile range. a By Wilcoxon test. b By Pearson c2 test
(Eþ; 85% with A or B and 15% with C or D). A hiatal hernia (>2 cm) was present in 21% of the cohort. Table 1 presents baseline covariates stratified by the presenting complaint (heartburn or suspected EER). Most patients (69% and 74%) in the 2 groups were women (P ¼ .21). Patients with suspected EER were older (median age 55 years; interquartile range 46–62) compared with patients with refractory heartburn (P < .01). Median BMI was 27 kg/m2 (interquartile range 24–33) in the heartburn group and 28 kg/m2 (25–34) in the EER group (P ¼ .16). There were no differences between the 2 groups for ethnicity, presence of esophagitis, hiatal hernia, or baseline percentage of time with pH <4 (Table 1). All patients had a previous failed 8-week course of high-dose (twice daily) PPI, but patients with refractory heartburn were more likely to be on once-daily PPI therapy, whereas patients with EER were more likely to be on twice-daily PPI therapy on presentation to our clinic (P < .01).
Predictors of Esophageal Acid Exposure Refractory Heartburn. In the multivariable model, BMI and presence of hiatal hernia were associated with abnormal pH in patients with refractory heartburn (Table 2), whereas age and concomitant presence of
regurgitation were not. The estimated adjusted odds ratio (OR) for an increase in BMI from 25 to 34 kg/m2 was 1.8 (95% confidence interval [CI] 1.2–2.5) and the OR for presence of hiatal hernia was 2.7 (95% CI 1.5–4.8). Extraesophageal Reflux. We found that the type of EER complaint was associated with abnormal pH. Patients with asthma had twice the odds of abnormal pH compared with those with cough or hoarseness (OR 2.0; 95% CI 1.2– 3.5; Table 2). Odds of abnormal pH was associated with the presence of hiatal hernia (OR 1.9; 95% CI 1.2–3.1), concomitant heartburn (OR 2.0; 95% CI 1.3–3.1), and increasing BMI from 25 to 34 kg/m2 (OR 2.1; 95% CI 1.5– 3.1). As in patients with refractory heartburn, regurgitation was not significantly associated with abnormal pH in the EER group (OR 1.0; 95% CI 0.4–2.4) in the multivariable model.
Model to Predict Abnormal Acid Reflux Refractory Heartburn. Probabilities of any abnormal pH (time >5.5% with pH <4) and having moderate to severe reflux (time >12% with pH <4) stratified by BMI and hiatal hernia are shown in Figures 2A and B, respectively. The probability of abnormal pH increased with higher BMI and presence of hiatal hernia (dichotomous). Probability
Clinical Predictors for Refractory Heartburn and EER 1733
increased nonlinearly for BMI, increasing from normal to overweight BMI and plateauing at an obesity level. Being overweight (BMI 25 kg/m2) with heartburn and hiatal hernia conferred a 70% pretest probability of abnormal pH (>5.5%) and a 33% pretest probability indicated this reflux was moderate to severe (>12%). Extraesophageal Reflux. The probability of abnormal acid exposure in those presenting with suspected EER symptoms was dependent on BMI, presence of hiatal hernia, presence of concomitant heartburn, and modified based on the specific presenting symptom (Figure 3). Increased probability was most pronounced in the transition from normal to overweight BMI and plateaued at the obesity level. Our model estimated that patients presenting with cough or hoarseness and no other risk factors with a normal BMI (20–25 kg/m2) had a 25%–40% probability of having an abnormal pH test result, whereas those who were overweight or obese (BMI >25 kg/m2) had a 40%–60% probability of having an abnormal pH using the cutoff of >5.5% (Figure 3A). Presence of concomitant heartburn and hiatal hernia increased this probability in a graded fashion. Patients whose chief complaint was asthma had a higher baseline probability of abnormal pH than those with isolated cough or hoarseness (Figure 3B). Patients with asthma alone and normal BMI had a 38%–60% probability
Probftotalavg 12 j Xg ¼
of having an abnormal pH test result, whereas those with asthma combined with a BMI >25 kg/m2 had a 60%–75% probability. Asthma and concomitant heartburn increased the probability of abnormal pH to 55%–75% in patients with normal BMI and 75%–85% in overweight or obese patients. Addition of hiatal hernia to this group (asthma þ concomitant heartburn þ hiatal hernia) further increased the probability to 70%–85% in patients with normal BMI and to 85%–92% in overweight or obese patients. The baseline probability of moderate to severe acid reflux (time >12% with pH <4) in patients with isolated cough or hoarseness decreased to 4%–8% with normal BMI and to 8%–15% in overweight or obese patients (Figure 3C). Probability was 23%–40% in patients with asthma þ concomitant heartburn þ hiatal hernia who had normal BMI and 40%–58% if they were overweight or obese (Figure 3D).
Scoring System for Moderate to Severe Reflux Refractory Heartburn. Equation 1 presents the clinical model to predict moderate to severe acid reflux (total time >12% with pH <4) using BMI as a continuous variable and hiatal hernia as a dichotomous variable (if present, use 1; otherwise, use 0):
1
; where 1 þ expð6:334065 X b bÞ
Xb b ¼ þ 1:075438 hh þ 0:1820395bmi 0:0006327213ðbmi 21:2Þ3þ þ 0:001041354ðbmi 27:4Þ3þ 0:0004086325ðbmi 37Þ3þ and ðxÞþ ¼ x if x > 0, 0 otherwise
Table 2.Association of Clinical Variables With Odds of Higher Esophageal Acid Exposure in Patients With Heartburn or EER Groups
Clinical variable
Heartburn
Hiatal hernia BMI (25–34) Concomitant regurgitation Age (40–60) Concomitant heartburn BMI (25–34) Asthma Hiatal hernia Concomitant regurgitation Age
EER
OR (95% CI) 2.7 1.8 1.1 1.3 2.0 2.1 2.0 1.9 1.0 0.8
(1.5–4.8) (1.2–2.5) (0.7–1.9) (0.9–1.9) (1.3–3.1) (1.5–3.1) (1.2–3.5) (1.2–3.1) (0.4–2.4) (0.6–1.1)
Given the complexity of the model, the model shown in Figure 2B can be easily used at bedside to determine the probability of abnormal pH at various BMI cutoffs with or without hiatal hernia. Furthermore, as noted in the figures, the relation between BMI and esophageal acid exposure is not linear, thereby making the analysis of a continuous variable only valuable on the left-hand side of the curve, where the predictability is not as high. Because of failed twice-daily PPI therapy in this population, testing should focus on identifying patients with moderate to severe reflux. Thus, based on our model, all patients in this group are considered to have low pretest probability for reflux and should undergo pH testing off therapy. Extraesophageal Reflux. Equation 2 shows the clinical models for prediction of moderate to severe acid reflux
CLINICAL AT
December 2018
1734 Patel et al
Gastroenterology Vol. 155, No. 6
in patients with EER. Equation 2A uses BMI as a continuous variable (without hiatal hernia):
Probftotalavg 12 j Xg ¼
undergo pH or MII-pH off therapy to rule out GERD as the potential contributor of symptoms. We compared the
1
; where 1 þ expð6:552082 X b bÞ
Xb b ¼ þ 0:1751225bmi 0:0006940719ðbmi 22:7Þ3þ þ 0:001058327ðbmi 28:2Þ3þ CLINICAL AT
0:0003642553ðbmi 38:68Þ3þ þ 0:70117 hb þ 0:6509617 asthma and ððxÞþ ¼ xÞifðx > 0Þ; 0 otherwise
Equation 2B shows a model with the addition of hiatal hernia (dichotomous) if that information is available clinically:
Probftotalavg 12 j Xg ¼
simplified HAs-BEER model with the continuous models with BMI and hiatal hernia size and they performed similarly (Supplemental Table 1).
1 1 þ expð6:484742 X b bÞ
; where
Xb b ¼ þ 0:6555836 hh þ 0:1617812bmi 0:000626957ðbmi 22:7Þ3þ þ 0:0009559897ðbmi 28:2Þ3þ 0:0003290328ðbmi 38:68Þ3þ þ 0:6860861 hb þ 0:7113895 asthma and ððxÞþ ¼ xÞifðx > 0Þ; 0 otherwise
Because of the complexity of the clinical models, we developed a new simple scoring system using probability data (HAs-BEER) to help clinicians estimate pretest probability of abnormal pH at bedside based on clinical history in patients with EER. The acronym HAs-BEER stands for Heartburn, Asthma, and BMI (>25) in EER. One point is earned for the presence of each of the following clinical parameters: concomitant heartburn, asthma, and BMI >25 kg/m2. No points are earned for cough and hoarseness without the other clinical parameters. BMI >25 kg/m2 was used as a cutoff because this corresponded to a similar percentage of increase in the probability of abnormal pH as the other binary predictors in the scoring system (heartburn and asthma; Table 2). A score of 3 had a positive predictive value of 92% (sensitivity 97.8%) for abnormal pH defined using the traditional cutoff of 5.5%. Given the high-pretest probability of reflux in this group, these patients should undergo MII-pH on therapy according to ACG guidelines. We made our model more specific by using a pH cutoff >12% with a HAs-BEER score 2 having 80% negative predictive value for moderate to severe reflux. Thus, these patients should
External Validation. Validation of the proposed model was performed in an external cohort from 2 different tertiary care medical centers (62% women; median age 59 years). The derivation cohort had a higher median total percentage of time with pH <4 (6.8 vs 3.9; P < .01) and therefore more patients (63% vs 42%; P < .01) with abnormal pH than the validation cohort (Table 3). There were no differences in the number of patients with moderate to severe acid reflux between cohorts. The validation cohort included patients with cough (62%), hoarseness (12%), and asthma (17%). Despite the noted baseline demographic differences between the derivation and validation cohorts (Table 3), the HAs-BEER score 2 had a negative predictive value for abnormal pH (defined as >12%) of 85% in the external cohort.
Discussion Reflux monitoring in patients with refractory heartburn or suspected EER-associated symptoms is widely variable because of the lack of consensus on definition, gold standard test, and therapeutic implications of a positive result.
Figure 2. Probability of abnormal pH in patients with refractory heartburn stratified by BMI and hiatal hernia with cutoff of abnormal pH defined as total time (A) >5.5% and (B) >12% with pH <4. HH, hiatal hernia.
Uncertainty has resulted in significant and rising management costs with serious implications for the patient with persistent reflux symptoms on PPI therapy.17 Current ACG guidelines recommend differential testing depending on the pretest probability of reflux, the criteria for which were left nebulous by the expert consensus panel. To date, guidance is lacking as to what constitutes low or high pretest probability in those patients with failed PPI trials and normal EGD findings.10 This study aimed to fill this important knowledge gap by identifying easily accessible clinical variables that clinicians can use to estimate the pretest probability of reflux in those with refractory heartburn or EER-associated symptoms using traditional pH or moderate to severe reflux cutoffs. Moderate to severe reflux was chosen as an additional target based on the only available prior surgical outcome data, which have been reported to have a more reliable treatment response to antireflux surgeries.15 We evaluated a large cohort of patients with refractory heartburn or suspected EER-associated symptoms (cough,
Clinical Predictors for Refractory Heartburn and EER 1735
hoarseness, or asthma) to determine independent predictors of abnormal pH and created a model to estimate pretest probability of abnormal pH. We found that BMI and presence of hiatal hernia in patients with refractory heartburn were independently associated with abnormal pH. Similar to previous studies, our study highlights the impact of obesity in esophageal acid exposure and as an independent risk factor for GERD symptoms and erosive esophagitis.18–21 As presented in Figure 2A, our model shows that the pretest probability of abnormal esophageal acid exposure using the traditional cutoff of 5.5% is too high in this group of patients in whom PPI therapy has failed. Thus, the primary objective in patients with refractory heartburn is to evaluate for moderate to severe reflux (cutoff 12%). Being overweight (BMI 25 kg/m2) with heartburn and hiatal hernia confers a 70% pretest probability of abnormal pH (>5.5%) and a 33% pretest probability that this reflux is moderate to severe (>12%). As presented in Figure 2B, the pretest probabilities of moderate to severe reflux are significantly lower. Thus, based on our model, we consider all patients with refractory heartburn and failed twice-daily PPI therapy as having low pretest probability for moderate to severe reflux and recommend pH testing off therapy. If they do not have moderate to severe reflux, then it is unlikely that reflux is the primary etiology of their ongoing symptoms (refractory heartburn) given prior failure of twice-daily PPI therapy and focus should be shifted toward identifying other etiologies (such as hypersensitive esophagus). Furthermore, we extend these observations to show that in patients with suspected EER symptoms, esophageal acid exposure is dependent on their chief complaint or presenting symptoms in addition to other clinical factors. We found that abnormal esophageal acid exposure (total time >5.5% with pH <4) was present in 56% of patients with heartburn and 63% of patients with EER (P ¼ .15). Although this difference was nonsignificant between the 2 groups, the higher percentage of abnormal acid exposure in patients with EER was due primarily to those patients presenting with asthma. Within EER groups, abnormal pH was significantly more prevalent (80%) in patients with asthma compared with cough (60%) or hoarseness (51%; P < .01). The derivation cohort also had a significantly higher prevalence of abnormal esophageal acid exposure (total time >5.5% with pH <4) at 63% compared with the external validation cohort (42%; P < .01), but this is likely due to the larger percentage of obese (BMI 30–60 kg/m2) patients in our derivation cohort (41% vs 32%). The prevalence of abnormal esophageal acid exposure in our patients with asthma is similar to previous studies showing a large percentage of patients with asthma having abnormal acid exposure ranging from 62% to 80%.22–26 Furthermore, we found that probability of abnormal esophageal acid exposure in patients with suspected EER is not uniform across presenting symptoms. In this group, presence of asthma, hiatal hernia, concomitant heartburn, and BMI increased the probability of abnormal pH by 2-fold
CLINICAL AT
December 2018
1736 Patel et al
Gastroenterology Vol. 155, No. 6
CLINICAL AT
Figure 3. Probability of abnormal pH in patients with EER stratified by presenting symptom and additive contributions of hiatal hernia and heartburn at various BMIs with cutoff of abnormal pH defined as total time >5.5% with pH <4 (A, cough or hoarseness; B, asthma) and total time >12% with pH <4 (C, cough or hoarseness; D, asthma). HB, heartburn; HH, hiatal hernia.
Table 3.Comparison of Baseline Demographics Between Derivation and Validation (External) Cohorts for EER (Cough, Hoarseness, or Asthma) Characteristics
Derivation cohort (n ¼ 257)
Validation cohort (n ¼ 118)
P value
Age (y), median (IQR) BMI (kg/m2), median (IQR) Underweight (16–18.5), % Normal (18.5–25), % Overweight (25–30), % Obese (30–60), % Women, % Chief complaint, % Asthma Cough Hoarseness None Secondary heartburn, % pH parameters Total percentage of time with pH <4, median (IQR) Upright percentage of time with pH <4, median (IQR) Supine percentage of time with pH <4, median (IQR) Total time >5.5% with pH <4, % Total time >12% with pH <4, % HAs-BEER score, % 0 1 2 3
55 (46–62) 28 (25–34) 1 25 33 41 74
59 (49–68) 28 (24–31) 1 29 38 32 62
<.01a .39a
21 60 9 0 45
17 62 12 9 26
<.01b
6.8 (3.9–11.3) 8.8 (4.8–14.7) 2.1 (0.2–6.3) 63 22
3.9 (1.3–8.2) 4.4 (1.9–10.5) 0.8 (0–5.1) 42 14
<.01a <.01a .02a <.01b .06b
13 42 35 10
20 56 18 7
<.01b
IQR, interquartile range. a By Wilcoxon test. b By Pearson c2 test.
.01b <.01b
Clinical Predictors for Refractory Heartburn and EER 1737
CLINICAL AT
December 2018
Figure 4. Proposed algorithm for testing patients with refractory heartburn or EER (asthma, cough, and hoarseness) for moderate to severe reflux (total time >12% with pH <4) based on the clinical model.
compared with those with isolated cough or hoarseness. In other words, the likelihood of a positive reflux test result is lower for a patient presenting with cough and/or hoarseness without concomitant heartburn than for a patient presenting with asthma. We also confirmed prior findings showing that being overweight and obese significantly contributed to the likelihood of increased esophageal acid exposure, with a substantial increase from normal BMI to overweight than from overweight to obese.27 There are several important and clinically relevant observations from our study. (1) There is no such thing as low probability of reflux as outlined by the ACG guidelines. Patients with the lowest probability, those with refractory
heartburn (Figure 2A) or cough or hoarseness (Figure 3A) and an average BMI (27 kg/m2 in this study cohort), had 50% baseline likelihood of reflux. Probability of positive test results increased in the presence of other clinical factors such as hiatal hernia, concomitant heartburn, and presence of asthma in the EER group (Figure 3A and B). (2) Not all EER presenting symptoms had the same pretest probability for abnormal esophageal acid exposure; that is, patients with asthma had higher odds of having abnormal esophageal acid exposure compared with those presenting with cough and/or hoarseness. In fact, most patients with asthma had abnormal esophageal acid exposure (abnormal ¼ time 5.5% with pH <4). Thus, in this population, does it make
1738 Patel et al
CLINICAL AT
sense to order a test that will shows a positive result in most patients? (3) We suggest using a different metric of abnormality (time >12% with pH <4 or moderate or severe reflux) for assessing degree of reflux in patients with refractory heartburn or EER symptoms. This is based on data indicating that time >12% with pH <4 was an independent predictor of EER symptom response to surgical fundoplication.15 As such, we developed a scoring system (HAsBEER) to determine which patients with refractory EER symptoms should undergo testing off or on PPI therapy based on probability of baseline moderate to severe esophageal acid exposure. Our model for refractory heartburn (Figure 2) or EER (Figure 3) is a useful bedside tool to determine pretest probability of abnormal pH based on readily available clinical parameters during a clinic visit. In this novel scoring system (HAs-BEER) based on clinical history, patients earn 1 point for each clinical parameter present (heartburn, asthma, or BMI >25 kg/m2). A score of 3 indicates high pretest probability of having moderate to severe reflux, which suggests these patients should undergo MII-pH testing on therapy to evaluate for non–acid reflux as a cause of persistent symptoms. Based on our cohort, this would indicate MII-pH testing on therapy in 9.7% of patients. In contrast, those with a score 2 have low pretest probability and should undergo pH or MII-pH testing off therapy to rule out GERD as the potential contributor of symptoms. Furthermore, we externally validated this model using a cohort of patients with EER from 2 separate institutions with a HAs-BEER score 2 having 85% negative predictive value for abnormal reflux (defined as >12%). This external cohort consisted of patients with cough or hoarseness and asthma. Thus, the present study provides a roadmap to define low and high probability, which contrasts with the current clinical approach. These findings further suggest that the traditional cutoff of >5.5% for total percentage of time with pH <4 is too low to exclude any patient groups from pH testing. Based on our model, a HAs-BEER score of 3 had 92% positive predictive value for abnormal pH using the traditional cutoff (>5.5%). Because PPI use has failed in these patients, the therapeutic implications of a positive pH study result based on traditional parameters is unclear and could result in referral for surgical fundoplication. Figure 4 shows our proposed algorithm for testing patients with refractory heartburn or EER (asthma, cough, and hoarseness) for moderate to severe reflux (total time >12% with pH <4) based on the clinical models. In addition, it should be noted that despite the significant baseline demographic and pH monitoring differences between the derivation and validation cohorts, our model performed very well in the external validation. This study has some limitations. (1) Validated instruments were not used to determine symptom severity at presentation. However, we aimed to make a model that can be easily used clinically; because symptom questionnaires are not used in routine daily clinical practice, we used the presenting chief complaint to develop our model and externally validated it to demonstrate its predictive
Gastroenterology Vol. 155, No. 6
accuracy. Our model also was developed with a robust sample size and fills a critical gap in the current ACG guidelines. Specifically, it provides a clinical algorithm to estimate the pretest probability of positive pH test results in patients with refractory heartburn or EER. (2) The recommended time >12% with pH <4 needs to be prospectively evaluated to see whether this level of reflux consistently results in symptom resolution after fundoplication. However, we use this cutoff for moderate to severe acid reflux based on the only surgical outcome data study currently available.15 Based on these data, we believe that this higher physiologic cutoff ensures that clinicians judiciously select patients with refractory symptoms who might benefit from surgical fundoplication. (3) We acknowledge that our cohort is based on a singlecenter experience with potential risk of referral bias, but the large number of patients included in this cohort minimizes this bias, as does external validation from 2 different tertiary care medical centers. Furthermore, results from this study will serve as the important backbone to pursue further multicenter prospective validation studies. (4) It is important to recognize that our findings are limited to patients in the United States because the BMI value might not be applicable for the Asian or European population. (5) Although we recognize that hiatal hernia is an important predictor of abnormal esophageal acid exposure in patients with suspected EER (as shown in Figure 3), we did not include it in our HAs-BEER score because we aimed to develop a predictive score that can be used at bedside by clinicians and presence or size of hiatal hernia is not usually available to clinicians at bedside. We addressed this by providing Equations 1 and 2A and 2B that show clinical models with hiatal hernia (if that information is available clinically). We also included Supplemental Table 1, which shows that predicted probabilities of moderate to severe reflux (total time >12% with pH <4) from the simplified HAs-BEER model (using BMI cutoff >25), the HAs-BEER model with BMI as a continuous variable (model 2), the HAs-BEER model with hiatal hernia (model 3), and the model with BMI and hiatal hernia as continuous variables (model 4) perform similarly. In summary, choice of reflux testing in patients with refractory heartburn and EER-associated symptoms (cough, hoarseness, and asthma) is currently fraught with confusion, which primarily derives from a lack of understanding of what constitutes high or low pretest probability of abnormal pH. We found abnormal esophageal acid exposure (defined as total time >5.5% with pH <4) in 56% of patients with refractory heartburn and 63% of patients with EER. The present data suggest the traditional cutoff value is too sensitive and not specific enough, especially in those with EER. Our novel bedside prediction tool can help determine pretest probability of abnormal pH in patients with failed PPI therapy and can help guide clinicians to determine the best diagnostic approach in this challenging group of patients and decrease unnecessary testing.
Supplementary Material Note: To access the supplementary material accompanying this article, visit the online version of Gastroenterology at www.gastrojournal.org, and at https://doi.org/10.1053/j. gastro.2018.08.038.
References 1. Sontaj SJ. The medical management of reflux esophagitis. Role of antacids and acid inhibition. Gastroenterol Clin North Am 1990;19:683–712. 2. Dent J, El-Serag HB, Wallander MA, et al. Epidemiology of gastro-oesophageal reflux disease: a systematic review. Gut 2005;54:710–717. 3. Fenter TC, Naslund MJ, Shah MB, et al. The cost of treating the 10 most prevalent diseases in men 50 years of age or older. Am J Manag Care 2006;12:S90–S98. 4. Sandler RS, Everhart JE, Donowitz M, et al. The burden of selected digestive diseases in the United States. Gastroenterology 2002;122:1500–1511. 5. Everhart JE, Ruhl CE. Burden of digestive diseases in the United States part I: overall and upper gastrointestinal diseases. Gastroenterology 2009;136:376–386. 6. Francis DO, Rymer JA, Slaughter JC, et al. High economic burden of caring for patients with suspected extraesophageal reflux. Am J Gastroenterol 2013; 108:905–911. 7. Vakil N, van Zanten SV, Kahrilas P, et al. The Montreal definition and classification of gastroesophageal reflux disease: a global evidence-based consensus. Am J Gastroenterol 2006;101:1900–1920; quiz 1943. 8. Klauser AG, Schindlbeck NE, Muller-Lissner SA. Symptoms in gastro-oesophageal reflux disease. Lancet 1990; 335:205–208. 9. Numans ME, Lau J, de Wit NJ, et al. Short-term treatment with proton-pump inhibitors as a test for gastroesophageal reflux disease: a meta-analysis of diagnostic test characteristics. Ann Intern Med 2004; 140:518–527. 10. Katz PO, Gerson LB, Vela MF. Guidelines for the diagnosis and management of gastroesophageal reflux disease. Am J Gastroenterol 2013;108:308–328; quiz 329. 11. Roman S, Gyawali CP, Savarino E, et al. Ambulatory reflux monitoring for diagnosis of gastro-esophageal reflux disease: update of the Porto consensus and recommendations from an international consensus group. Neurogastroenterol Motil 2017;29:1–15. 12. Lundell LR, Dent J, Bennett JR, et al. Endoscopic assessment of oesophagitis: clinical and functional correlates and further validation of the Los Angeles classification. Gut 1999;45:172–180. 13. Kavitt RT, Yuksel ES, Slaughter JC, et al. The role of impedance monitoring in patients with extraesophageal symptoms. Laryngoscope 2013;123: 2463–2468. 14. Pandolfino JE, Richter JE, Ours T, et al. Ambulatory esophageal pH monitoring using a wireless system. Am J Gastroenterol 2003;98:740–749.
Clinical Predictors for Refractory Heartburn and EER 1739 15. Francis DO, Goutte M, Slaughter JC, et al. Traditional reflux parameters and not impedance monitoring predict outcome after fundoplication in extraesophageal reflux. Laryngoscope 2011;121:1902–1909. 16. R Core Team. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria;(2014). Available at: http:// www.R-project.org/. Accessed November 13, 2008. 17. Becher A, El-Serag H. Systematic review: the association between symptomatic response to proton pump inhibitors and health-related quality of life in patients with gastro-oesophageal reflux disease. Aliment Pharmacol Ther 2011;34:618–627. 18. El-Serag HB, Ergun GA, Pandolfino J, et al. Obesity increases oesophageal acid exposure. Gut 2007;56:749– 755. 19. Crowell MD, Bradley A, Hansel S, et al. Obesity is associated with increased 48-h esophageal acid exposure in patients with symptomatic gastroesophageal reflux. Am J Gastroenterol 2009;104:553–559. 20. El-Serag HB, Graham DY, Satia JA, et al. Obesity is an independent risk factor for GERD symptoms and erosive esophagitis. Am J Gastroenterol 2005;100: 1243–1250. 21. Hampel H, Abraham NS, El-Serag HB. Meta-analysis: obesity and the risk for gastroesophageal reflux disease and its complications. Ann Intern Med 2005;143: 199–211. 22. Harding SM, Guzzo MR, Richter JE. The prevalence of gastroesophageal reflux in asthma patients without reflux symptoms. Am J Respir Crit Care Med 2000;162:34–39. 23. Sontag SJ, O’Connell S, Khandelwal S, et al. Most asthmatics have gastroesophageal reflux with or without bronchodilator therapy. Gastroenterology 1990;99:613– 620. 24. Bucknall C, Stanton A, Miller G, et al. The impact of normalization of esophageal acid profile by incremental protein pump inhibitors dosing in difficult asthma patients with proven gastro-esophageal acid reflux. J Asthma 2009;46:506–511. 25. DeMeester TR, Bonavina L, Iascone C, et al. Chronic respiratory symptoms and occult gastroesophageal reflux. A prospective clinical study and results of surgical therapy. Ann Surg 1990;211:337–345. 26. Alhabib KF, Vedal S, Champion P, et al. The utility of ambulatory pH monitoring in patients presenting with chronic cough and asthma. Can J Gastroenterol 2007; 21:159–163. 27. Aslam M, Slaughter JC, Goutte M, et al. Nonlinear relationship between body mass index and esophageal acid exposure in the extraesophageal manifestations of reflux. Clin Gastroenterol Hepatol 2012; 10:874–878.
Received January 9, 2018. Accepted August 22, 2018. Reprint requests Address requests for reprints to: Michael F. Vaezi, MD, PhD, MSc (Epi), FACG, Professor of Medicine, Clinical Director, Division of Gastroenterology, Hepatology, and Nutrition, Vanderbilt University Medical Center, Nashville, Tennessee 37232. e-mail: [email protected].
CLINICAL AT
December 2018
1740 Patel et al Acknowledgments Author contributions: Dhyanesh A. Patel, C. Gaelyn Garrett, David O. Francis, and Michael F. Vaezi conceived the study. Dhyanesh A. Patel and Michael F. Vaezi designed the study. Dhyanesh A. Patel, Rohit Sharda, Tina Higginbotham, and Michael F. Vaezi acquired the data. Karthik Ravi, Steven Hasak, David Katzka, and C. Prakash Gyawali acquired external data. Dhyanesh A. Patel, Rohit Sharda, Yash A. Choksi, James C. Slaughter, and
Gastroenterology Vol. 155, No. 6 Michael F. Vaezi analyzed and interpreted the data. Dhyanesh A. Patel and James C. Slaughter drafted the manuscript. Dhyanesh A. Patel, James C. Slaughter, and David O. Francis revised the manuscript. Michael F. Vaezi critically reviewed the manuscript and supervised the study. Conflicts of interest The authors disclose on conflicts.
CLINICAL AT
December 2018
Clinical Predictors for Refractory Heartburn and EER 1740.e1
Supplemental Table 1.Predicted Probabilities of Moderate to Severe Reflux (Total % Time pH <4 of >12%) From Simplified HAs-BEER Model (BMI Cutoff >25), HAs-BEER Model With Continuous BMI (Model 2), HAs-BEER Model With Hiatal Hernia (Model 3), and Model With Continuous BMI and Hiatal Hernia (Model 4)a BMI
Heartburn
Hiatal hernia size
Asthma
HAs-BEER score
HAs-BEER probability
Model 2
Model 3
Model 4
25 25 25 25 25 25 25 25 25 30 30 30 25 25 25 30 30 30 30 30 30 30 30 30
0 0 0 0 0 0 1 1 1 0 0 0 1 1 1 0 0 0 1 1 1 1 1 1
0 2 4 0 2 4 0 2 4 0 2 4 0 2 4 0 2 4 0 2 4 0 2 4
0 0 0 1 1 1 0 0 0 0 0 0 1 1 1 1 1 1 0 0 0 1 1 1
0 0 0 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 3 3 3
8% 8% 8% 15% 15% 15% 15% 15% 15% 15% 15% 15% 29% 29% 29% 29% 29% 29% 29% 29% 29% 47% 47% 47%
10% 10% 10% 18% 18% 18% 19% 19% 19% 17% 17% 17% 30% 30% 30% 29% 29% 29% 30% 30% 30% 45% 45% 45%
6% 11% 10% 12% 23% 20% 12% 23% 20% 12% 23% 20% 23% 40% 37% 23% 40% 37% 23% 40% 37% 40% 60% 56%
8% 15% 14% 16% 28% 26% 14% 26% 24% 13% 24% 22% 27% 44% 41% 25% 42% 39% 23% 39% 36% 40% 59% 56%
a
Dichotomous variables were classified as 1 (yes) or 0 (no) based on presence of the symptom. Probabilities are given for representative values of BMI, heartburn presence (or absence), hernia size, and asthma.