The number of patients with proton pump inhibitor (PPI)-refractory reflux symptoms is underestimated since many patients resign after an unsuccessful therapy attempt. Thus, it would be useful having a non-invasive tool that can help identify true gastroesophageal reflux disease (GERD) patients in order to manage them early and properly. The GerdQ is a validated tool developed for this purpose but its applicability in PPI-refractory patients has not yet been investigated. Our aim was to investigate if reflux symptoms per se, the GerdQ and patients characteristics are suitable for non-invasive diagnosis of GERD in patients with PPI refractory reflux symptoms.
A total of 500 patients from a prospectively recorded data base with PPI-refractory reflux symptoms were retrospectively analyzed. All patients received comprehensive diagnostic workup including EGD, pH-impedance measurement and manometry. GERD was diagnosed according to the recent Lyon consensus.
Of all patients enrolled in the study, 280 (56%) finally fulfilled the criteria for objectively verified GERD according to the Lyon consensus. There were no significant differences in age and gender between the patients with and without GERD, whereas the body mass index was significantly higher in the group with verified GERD, but the discriminative value was low (Welch-Test,
Based on our study, neither symptoms and the GerdQ score nor patients’ characteristics are appropriate tools to distinguish between GERD and other causes for reflux symptoms in patients with PPI-refractory reflux symptoms.
Gastroesophageal reflux disease (GERD) affects about 20–25% of the adult population in the Western world (
Typically, patients with reflux symptoms without presenting red flags and risk factors for complications are treated with proton pump inhibitors (PPI) (
The aim of this study was to investigate if the GerdQ score is a suitable tool for symptom-based diagnosis of GERD in patients with PPI-refractory reflux symptoms. For this purpose, an exploratory analysis of prospectively collected data from the Reflux-Center Siegerland was performed.
The Reflux Center Siegerland was founded in 2014 in cooperation between two Siegen hospitals, to help patients with persistent reflux symptoms who are dissatisfied with their previous treatment. Based on comprehensive diagnostics at the Reflux Center the diagnoses are verified and conservative or operative options are discussed with the patients.
Before performing the diagnostic work-up within two working days, the patients needed to stop proton-pump-inhibitor (PPI) therapy, as far as still existing, at least for two weeks. By using specially tailored questionnaires, they were asked about their demographic characteristics, symptom frequency and strength, previous diagnostics and treatment (anamnestic and
In a next step, patients received comprehensive diagnostic workup, including esophagogastroduodenoscopy (EGD) with histological examination from the duodenum, stomach and esophagus, 24-hour pH-impedance measurement and high resolution manometry. Depending on symptoms some patients received additional tests,
Based on the evaluated diagnostic results, the patients can be categorized as follows: Non erosive reflux disease (NERD) with proven acid reflux, hypersensitive esophagus without proven pathologic acid reflux but with significant correlation between physiological acid reflux and symptoms (SAP >95%), functional heartburn without any correlation between reflux events and symptoms, and erosive reflux disease (ERD), divided into mild ERD (Los Angeles A/B) or severe ERD (Los Angeles C/D).
Finally, according to the criteria of the Lyon consensus published in 2018 (
According to the Lyon consensus published in 2018, typical symptoms and response to therapy alone are not sufficient to diagnose GERD. Endoscopy results and pH or pH impedance measurement are required for a definitive diagnosis (
In accordance with the Lyon consensus, conclusive evidence for pathologic reflux are advanced erosive esophagitis (Los-Angeles grades C/D esophagitis), long segment Barrett‘s mucosa, peptic esophageal strictures or an acid exposure time (AET) >6% in the distal esophagus on pH or pH-impedance monitoring. In contrast, an acid exposure time <4% and <40 reflux episodes on pH-impedance measurement argue against pathologic reflux. According to the Lyon criteria, every conclusive finding is strong evidence for the presence of GERD. If the results are inconclusive or borderline, further findings from esophageal examinations (
In our study, the diagnosis of GERD was defined according to the criteria of the Lyon consensus (16) mentioned before. In the synopsis of the results collected from the diagnostic measurement, the patients were divided into those with GERD on the one hand and those without GERD on the other hand. GERD was diagnosed in the presence of erosive esophagitis Los Angeles C&D, an acid exposure time (AET) ≥ 6%, peptic esophageal strictures and/or long-segment Barrett‘s esophagus, whereas an AET <4% was considered as an exclusion criterion for GERD. In case of inconclusive or borderline results, such as an AET between 4–6% or erosive esophagitis Los Angeles A&B, the diagnosis was made by a synopsis of all findings in addition to expert opinion (interdisciplinary reflux board consisting of three experienced gastroenterologists and two visceral surgeons with a special focus on upper GI diseases).
The GerdQ is a validated non-invasive tool for diagnosing reflux disease in primary care. This questionnaire includes six items, four positive predictors (heartburn, regurgitation, sleep disturbance due to reflux-symptoms, use of over-the-counter antireflux medication), and two negative predictors (epigastric pain and nausea). Their frequency (0 day/week, 1 day/week, 2–3 days/week, 4–7 days/week) is specified on a Likert Scale (0–3) for the positive predictors and on a reversed Likert Scale (3–0) for the negative predictors (
In cooperation with the Reflux Center Siegerland, the characteristics, symptom manifestations, and diagnostic results of 500 PPI-refractory patients with and without GERD between 2014 and 2020 were collected. Of those, 243 (48.6%) patients were male and 257 (51.4%) are female, the age of the participants ranged from 15 to 89 years with a mean age of 52,84 (SD 15.1). The BMI varied from 16 kg/m2 as minimum to 44 kg/m2 as maximum with a mean BMI of 26,48 kg/m2 (SD 4.8). 362 of the participants (72.4%) were non-smokers, 71 (14.2%) were ex-smokers and 67 (13.4%) were smokers. Of all patients analyzed, 262 (52.4%) did not drink alcohol, whereas 177 (35.4%) consumed alcohol occasionally and 61 (12.2%) regularly.
Of the 500 patients included in our study, 457 (91.4%) participants already received previous diagnostic in form of an EGD before presenting to our Reflux Center the first time; only 8.2% received pH-metry in addition to the EGD. All participants have had an insufficient PPI treatment approach prior to their first visit at the Reflux Center.
At the Reflux Center Siegerland, all of the 500 included patients received EGD and, with only few exceptions, also pH-impedance measurement (
Descriptive statistics with means and standard deviations for continuous variables and with percentages for categorical variables were calculated. Metric variables were tested regarding deviation from the normal distribution with the Shapiro–Wilk test and with tests regarding skewness und kurtosis.
Receiver operating characteristic (ROC) curves with 95% confidence intervals (CI) were used for estimating the optimal cutoff value to differentiate subjects with and without a diagnosis of GERD (
Categorical variables were compared between the two groups using the Chi-square test with effect size Cramér-V. A Cramér-V ≥ .40 signals a large effect (
IBM SPSS 27, R version 4.02 and RStudio 1.2.5042 were used with R packages epiR, pROC (
A total of 500 patients, who were consecutively evaluated in the Reflux Center, were analysed. Of these 500 patients, 280 patients (56%) fulfilled the criteria for objectively verified GERD (please see Methods section for further details). Sample characteristics are displayed in
Variable | GERD | No GERD | Statistics |
---|---|---|---|
Gender | 50% female | 53.2% female | Chi-square |
Age | 53.9 (15.0) | 51.5 (15.1) | Welch-Test, |
BMI | 27.3 (4.8) | 25.5 (4.6) | Welch-Test, |
There were no significant differences regarding gender and age between subjects with and without GERD. Subjects with GERD had a significantly higher body mass index (BMI) than subjects without GERD. Hence, the effect size showed a small effect. The distribution of GerdQ raw scores in subjects with and without diagnosis of GERD is shown in
No GERD | GERD | |||
---|---|---|---|---|
GerdQscore | 1 | 0 | 1 | |
0,0% | 0,4% | |||
2 | 1 | 0 | ||
0,5% | 0,0% | |||
3 | 8 | 7 | ||
3,6% | 2,5% | |||
4 | 3 | 5 | ||
1,4% | 1,8% | |||
5 | 11 | 9 | ||
5,0% | 3,2% | |||
6 | 27 | 27 | ||
12,3% | 9,6% | |||
7 | 11 | 20 | ||
5,0% | 7,1% | |||
8 | 34 | 33 | ||
15,5% | 11,8% | |||
9 | 13 | 26 | ||
5,9% | 9,3% | |||
10 | 17 | 32 | ||
7,7% | 11,4% | |||
11 | 17 | 19 | ||
7,7% | 6,8% | |||
12 | 18 | 26 | ||
8,2% | 9,3% | |||
13 | 11 | 11 | ||
5,0% | 3,9% | |||
14 | 13 | 15 | ||
5,9% | 5,4% | |||
15 | 20 | 17 | ||
9,1% | 6,1% | |||
16 | 10 | 17 | ||
4,5% | 6,1% | |||
17 | 1 | 6 | ||
0,5% | 2,1% | |||
18 | 5 | 9 | ||
2,3% | 3,2% | |||
Total | 220 | 280 | ||
100,0% | 100,0% |
Descriptive statistics of both groups regarding the GerdQ score are displayed in
Mean | SD | Min | Max | Shapiro–Wilk-Test | Skewness | Kurtosis | |
---|---|---|---|---|---|---|---|
GERD | 10.26 | 3.76 | 1 | 18 | <.0001 | ||
No GERD | 9.94 | 3.79 | 2 | 18 | <.0001 |
Both distributions significantly deviate from a normal distribution (Shapiro–Wilk-Test,
The Welch
ROC analyses revealed that the GerdQ was not able to differentiate between the two groups. The area under the curve (AUC) was .525 (95% CI [.474–.576]), p =.34 which is slightly above chance level. The Youden Index (−.003, 95% CI [−.13–.12]) revealed a GerdQ score of 9 as an optimal cut-off value for the separation of subjects with and without diagnosis of GERD (
Setting this threshold leads to the classification in
In
Our study showed for the first time that the GerdQ is not an appropriate tool to distinguish between GERD and other causes in patients with PPI-refractory reflux symptoms. Moreover, we could demonstrate in a large cohort with prospectively gathered data and an elaborated diagnostic work-up that about half of the patients complaining PPI-refractory reflux symptoms do not have true GERD based on the Lyon consensus (
Some studies in different countries provided evidence on the general applicability of the GerdQ, for example in Norway (
In contrast to previous studies we examined the GerdQ in patients with refractory reflux symptoms in the setting of a center. Furthermore, all of our study participants discontinued their PPI therapy at least two weeks before the study and the diagnosis of GERD was made according to the Lyon consensus criteria (
In addition, an important strength of our study is the large number of patients (
GERD | No GERD | ||
---|---|---|---|
GerdQ score ≥9 | 120 | 95 | 215 |
GerdQ score <9 | 160 | 125 | 285 |
280 | 220 |
GERD | 95% CI | |
---|---|---|
Sensitivity | 43% | 37–49% |
Specificity | 57% | 50–63% |
Positive predictive value | 56% | 49–63% |
Negative predictive value | 44% | 38–50% |
Positive likelihood ratio | 0.99 | 0.81–1.22 |
Negative likelihood ratio | 1.01 | 0.86–1.17 |
The data derived from our reflux center clearly showed that many patients suffering from so called reflux symptoms that are refractory to PPI therapy do not have reflux disease defined as troublesome symptoms as a result of reflux of stomach content into the esophagus (
In order to improve care of patients with persistent reflux symptoms despite PPI-therapy in the near future, the development of a stepwise approach is desirable to better distinguish between GERD and other diseases in this important patient population.
acid exposure time
body mass index
confidence intervals
esophagogastroduodenoscopy
erosive reflux disease
Gastroesophageal reflux disease questionnaire
Gastro-esophageal reflux disease Impact Scale
Gastroesophageal reflux disease
Gastrointestinal Symptom Rating Scale
high resolution manometry
positive likelihood ratios
negative likelihood ratios
Non erosive reflux disease
negative predictive values
proton pump inhibitor
positive predictive values
Reflux Disease Questionnaire
Receiver operating characteristic
symptom association probability
The authors declare there are no competing interests.
The following information was supplied relating to ethical approvals (i.e., approving body and any reference numbers):
The Ethics Committee of the University of Essen Medical School approved the study (16-7125-BO).
The following information was supplied regarding data availability:
The raw data are available as a