Is the reflux disease questionnaire useful for identifying GERD according to the Montreal definition?

In Western societies, 30% [1] to 60% [2] of persons suffer from heartburn or regurgitation, yet not all those who report these symptoms suffer from gastro-oesophageal reflux disease (GERD). Indeed, the differentiation between frequent and occasional symptoms is unsatisfactory from a practical standpoint, in view of the fact that a considerable proportion of patients with occasional symptoms consult their physician [3] and suffer from a deterioration in their quality of life [4]. Consequently, for every patient that seeks advice, it is the physician who has to decide whether the symptoms are irrelevant or circumstantial (and, by extension, not constitutive of disease), or alternatively, whether the patient is really suffering from GERD.

The Montreal Consensus defines GERD as "a condition which develops when the reflux of stomach contents causes troublesome symptoms and/or complications" [5], the conclusion being that, in clinical practice, it is the patient who defines whether or not the symptoms are troublesome, without the use of arbitrary cut-points of frequency and severity. Any given person's perception of a symptom as troublesome goes beyond its frequency, its severity or even its impact on quality of life [6], since it implies individual cognitive assessment which may vary according to the circumstances [7].

When it comes to the severity of GERD -an aspect of crucial importance to the physician because it is precisely on this that his attitude to diagnosis and therapy will so often depend [8, 9]- the Montreal Consensus does not specify how GERD severity is to be assessed in clinical practice, thus assuming the degree of severity to be that reported by the patient [10].

Yet, this commitment to patient-centred clinical practice comes up against the reality of the fact that, as compared to the opinion held by their patients, physicians tend to underestimate the impact of gastro-oesophageal reflux symptoms [11, 12]. To try and minimise the discrepancies between patient-based information and physicians' subjective evaluation, useful instruments have been developed to collect information directly from the patient as uniformly as possible and help physicians diagnose GERD and assess its severity, e.g., the Reflux Disease Questionnaire (RDQ) [13‐15]. This questionnaire was developed using definitions of GERD based on symptom frequency and severity, and its ability to correctly identify GERD was recently validated using 24-hour pH monitoring as a gold standard [16]. However, GERD is usually diagnosed without invasive testing, and the neither of the following aspects is known: the extent to which the RDQ is useful for differentiating troublesome from non-troublesome reflux symptoms, a key factor for diagnosing GERD within the new conceptual framework proposed by the Montreal Consensus; and, the RDQ's capacity to distinguish symptom severity.

Hence, the primary objective of this study was to assess the utility of the RDQ in standard clinical practice for identifying reflux symptoms defined as troublesome by the patient. As a secondary objective, we proposed to evaluate its capacity for classifying the severity of GERD symptoms.

On the basis of the screening criteria, the study initially included a total of 5332 eligible patients attending 926 centres distributed nationwide (48 out of 51 provinces): 758 patients were excluded for not fulfilling the inclusion and exclusion criteria, due to the lack of some item of basic information (e.g., age). Accordingly, 4574 (85.8%) patients were available for analysis; of these, 4410 patients reported heartburn and 4230 acid regurgitation in the screening questions.

Our study's principal contribution is that patients' perception of reflux symptoms as "troublesome" depends, above all, on the characteristics of the symptoms themselves, and that the RDQ is a relatively useful instrument for identifying symptoms regarded as troublesome by the patient. Accordingly, the RDQ would be useful for differentiating troublesome from non-troublesome reflux symptoms (the basis of the Montreal Consensus), thereby adding value to its recently shown ability to make a symptom-based diagnosis of GERD among patients with troublesome upper GI symptoms [16].

The Montreal Consensus [5] lays down a disease definition based on a situation where the symptoms are perceived as troublesome by the patient and the definition's application to clinical practice is feasible [23]. Its implementation depends basically on the physician judging the impact of the symptoms relayed to him by the patient as being or not being troublesome. It is reasonable to surmise that questions about symptom frequency, severity and impact on the patient would have been incorporated into most physicians' anamnesis years ago, and yet there continues to be discordance between physicians' and patients' respective subjective impressions, with disease severity being generally underestimated [11, 12, 24]. The use of patient-administered questionnaires can thus help physicians ascertain the patients' symptoms more objectively and decide whether such symptoms are indeed troublesome.

The principal aim of this study was to assess the capacity of a simple scale that is widely used in clinical practice and clinical studies -the RDQ- to discern typically troublesome symptoms. In our study, the RDQ displayed good sensitivity and specificity when it came to distinguishing troublesome from non-troublesome symptoms. This is so in the two principal clinical situations provoked by patients with typical GERD symptoms, namely, the need to decide whether the symptoms are constitutive of disease (GERD diagnosis); and the persistence of symptoms with treatment (adequate or inadequate response). The negative predictive values are low because they depend on prevalence, and in this case, the prevalence of troublesome symptoms is so high that it increases the amount of false negatives.

Furthermore, the cut-points for identifying troublesome symptoms in the RDQ are homogeneous across all patients, with or without PPI treatment. Patients' perception of TS appears to be fairly uniform and homogeneous over the disease course, something that simplifies clinical application considerably, since the same result is obtained in two populations of patients under different circumstances. It seems reasonable to propose that, as GERD is defined on the basis of the presence of troublesome symptoms, inadequate response to treatment should be defined in like manner.

Moreover, other recent studies have shown that RDQ is useful for diagnosis of GERD in real-world practice [25].

The factors that influence the perception of symptoms as troublesome are not well known, though there are undoubtedly subjective elements personal to each patient, and it has been reported that in the general population such perception can depend on factors extraneous to the symptoms themselves [26]. While the aim of our study was not specifically to study this aspect, our data suggest that, among patients who consult primary care physicians, apart from the frequency of typical symptoms, impact on sleep is a relevant characteristic in such symptoms being viewed as troublesome. In this regard, recent studies conducted mainly in Western countries have shown a two-way relationship between GERD and sleep, in cases where night-time reflux leads to sleep deprivation and sleep deprivation per se can exacerbate GERD by enhancing perception of intra-oesophageal stimuli [27‐29]. Other factors of a socio-cultural nature may have an influence: in our study, we covered all areas country-wide to try and minimise the potential impact that these factors might have. We did not conduct a comprehensive study of all the potential factors that might exert an influence, psychosocial factors in particular. This was not the primary goal of the study and would have required increasing the number of questions to be completed by patients, thereby endangering compliance with the protocol and, by extension, putting the primary goal at risk. Nevertheless, the fact that depression and anxiety disorder diagnosed by the physician may not represent a related factor, suggests that psychosocial factors play a secondary role in the primary care setting. It would seem that neither other comorbidities nor the use of drugs play a relevant role, though it has to be said that patients with severe diseases which might potentially have had a substantial impact on health-related quality of life were excluded. It should be noted that hypercholesterolemia was associated with reporting TS. Although there is no evident explanation, this does not appear to be a random result, given the great difference in prevalence in both groups. The fact that it was not associated with TS among patients on treatment suggests that, in reality, hypercholesterolemia is associated with GERD rather than with troublesome symptoms. In this respect, this association has already been reported by Ponce et al. [30] in this country and was recently observed in a population-based study by Eslick and Talley [31]. It has been pointed out that this association could be mediated by obesity but in our study this relationship persisted after adjusting for BMI. At all events, frequency of symptoms plus their impact on sleep emerge as the principal factors for determining TS.

There are no universally accepted severity scales for GERD, and indeed the use of the GSRS was the only Spanish-language standard available. Obviously, "severity", like "troublesome", is a subjective measure and difficult to define, seeing as many factors can be responsible for it. Nonetheless, the RDQ displayed a great capacity for discerning these groups when the GSRS was used as the standard. This tells us two things: first, that having cut-points enables patients to be stratified, which is undeniably useful in clinical practice and clinical research alike; and second, that the scales indirectly indicate a relationship with severity and so express severity quantitatively, which enables changes in severity to be assessed. It will be for future studies to demonstrate this.

Our results are important because they mean that a therapeutic attitude can be uniformly implemented and treatment escalations can be modulated, and because they provide a single cut-point for cases where PPIs are or are not being taken, which is vital when it comes to defining therapeutic failure for every individual patient.

Our study has a number of strengths. The first is its large sample size. The inclusion of over 4,500 patients allowed us, not only to ensure a sufficient number of persons with NTS, but also to provide statistical power to detect small yet potentially relevant effects. Furthermore, the study population represents clinical reality, in that persons with troublesome and non-troublesome symptoms who consulted physicians were studied, free of any interference in the form of persons with symptoms not seeking medical consultation. Secondly, all the instruments used to collect data were validated, with the exception of the question referring to TS (which solely seeks to represent the clinical question homogeneously), and the clinical information was drawn directly from the data source (the physician). Although the degree to which a given symptom is troublesome can be evaluated numerically, as in the case of a recent population-based study [26], we nevertheless feel that a yes/no question obliges patients to define themselves, prior to any other question on the impact of symptoms, and so conforms to the Montreal philosophy, i.e., that it is for the patient to define what is and what is not troublesome. Thirdly, patients from over 900 health centres were included, thereby avoiding any single-researcher bias vis-à-vis the patient profile included. Fourthly, the inclusion criteria were as lax as possible, so as to enable a truly representative sample of patients with GERD symptoms seeking medical advice to be included. The exclusions were only those required by protocol, as well as that of patients with diseases which might have an important impact on quality of life.

It is also necessary to acknowledge some limitations. Our study was observational and thus the validity of the cut-points was not confirmed in other populations. Even so, it must be stressed that the same cut-points were reproduced in both subsamples, which indirectly indicates the validity of these instruments. No sub-analysis was performed to try and identify the questionnaire items which were most relevant or even whether the combination of the two questionnaires or of the most relevant items from both might have had greater efficacy. The standard used was strictly based on symptoms and specifically on the simple specification by the patient that he/she had suffered from heartburn/acid regurgitation and that this was troublesome. To what extent this is representative of the diagnosis obtained using invasive physiological tests is uncertain. Despite the fact that a recent study reported a low sensitivity for clinical diagnosis based on the Montreal Consensus in obese patients [32], the use of the RDQ for symptom-based diagnosis of GERD in the primary-care setting has been clinically validated [16], and all clinical guidelines support symptom-based diagnosis of GERD.

In brief, it can be concluded that patients' perception of reflux symptoms as troublesome depends, above all, on the characteristics of the symptoms themselves and that the RDQ in a useful instrument for helping the primary care physician differentiate relevant from irrelevant reflux symptoms, which is fundamental for diagnosis of GERD based on the Montreal Consensus.