Gastroesophageal reflux (GER), or the retrograde flow of gastric content across the esophagogastric junction (EGJ) and the lower esophageal sphincter (LES) , can be physiologic, especially in the postprandial setting. Inherent mechanisms are in place for the LES to relax transiently in response to distension of the fundus of the stomach, resulting in venting of air (belching) [6]. The resting LES tone, inspiratory diaphragmatic crural pinch at the same level as the LES, and the angle between the long axes of the esophagus and the stomach prevent significant retrograde movement of gastric content across the EGJ and LES in the physiologic setting. However, transient LES relaxations (TLESRs) can result in small amounts of gastric content refluxing into the esophagus; in health, esophageal secondary peristalsis is efficient in stripping any refluxed material back into the stomach [7].

GER becomes pathologic (GERD) when associated with symptoms (typically heartburn or regurgitation) or mucosal injury (typically esophagitis or BE) [8, 9]. Symptoms and mucosal injury are not mutually exclusive, and each can occur in the absence of the other. Therefore, subjective symptom analysis, and, indeed, endoscopic inspection of the esophageal mucosa, may not always be indicative of GERD. Symptoms related to GERD can be atypical (noncardiac chest pain, NCCP) or even extra-esophageal (cough, asthma, dental erosion), further complicating the diagnosis of GERD in these settings [1]. Beyond symptom assessment and inspection of the esophageal mucosa at upper endoscopy , the availability of diagnostic tests to quantify reflux and to assess the association of symptoms with reflux episodes affords further insight into the definition of GERD.

The clinical presentation of GERD is predominantly symptom based, as the vast majority of patients present to their physicians with typical symptoms of heartburn and regurgitation . However, there is a significant and growing recognition of atypical symptoms defining GERD, particularly when these atypical symptoms occur in the absence of typical symptoms or endoscopic evidence of mucosal damage. Given the diagnostic challenges associated with the spectrum of clinical symptoms that may be related to GERD with varying definitions across geographic regions, the Montreal classification International Consensus Group was formed to develop a global definition for GERD [1]. Utilizing a modified Delphi process over a 2-year period, this group proposed 50 consensus statements pertaining to GERD definition, published in 2006. At the core, the Montreal group agreed that GERD develops from reflux of stomach contents into the esophagus and proximally, causing troublesome symptoms and/or complications [1].

Symptomatically, the Montreal classification suggested that reflux symptoms must be “troublesome” to meet the definition of GERD. Specifically, this threshold required adverse effects on patient well-being; population-based studies have suggested mild symptoms occurring at least 2 days weekly or moderate-to-severe symptoms occurring at least 1 day weekly may approximate this threshold [10, 11]. Others have suggested that heartburn symptoms occurring more than twice a week negatively impact quality of life [12]. However, in practice, clinicians rely on patients themselves to determine if their reflux symptoms are troublesome, rather than rely on frequency or duration thresholds to meet this definition of GERD. In the absence of esophageal mucosal injury, episodic heartburn not deemed troublesome by the patient does not meet the Montreal criteria for a symptomatic esophageal GERD syndrome [13].

The Montreal classification concluded that heartburn and regurgitation constitute the characteristic symptoms of the typical reflux syndrome, allowing suspicion of GERD based on presence of these symptoms alone, a position adopted by the American Gastroenterological Association in 2008 [14]. However, typical GERD symptoms (heartburn, regurgitation) by themselves are only modestly predictive of GERD. In a large cohort of 33,000 patients undergoing endoscopy for typical GERD symptoms, 27.8 % had erosive esophagitis , 9.1 % had BE , 3.7 % had esophageal strictures , and 44.8 % had a hiatal hernia , leaving 39 % with a normal endoscopy [15]. When compared to endoscopic evidence of GERD, the performance characteristics of these typical symptoms demonstrated sensitivity of only 44 %, but with specificity of 87 %, in another study [16]. When ambulatory reflux monitoring is used as the gold standard, performance characteristics are better. In a selected population of over 300 patients referred for 24-h ambulatory pH monitoring, typical symptoms had 78 % sensitivity and 60 % specificity for GERD [17]. Likewise, in a cohort of 228 patients who had previously undergone laparoscopic anti-reflux surgery, only heartburn significantly correlated with abnormal acid exposure on pH testing, with a positive predictive value of 43 %, negative predictive value of 82 %, and overall accuracy of 78 % [18]. The addition of a further step, the proton pump inhibitor (PPI) test, adds additional confidence in the symptomatic diagnosis of GERD with typical symptoms, as discussed below.

A significant advance in defining GERD over the past two decades consists of the distinction between esophageal and extra-esophageal syndromes. In the Montreal classification , esophageal syndromes were further subdivided into symptomatic syndromes (typical reflux syndrome, reflux chest pain syndrome), and syndromes with esophageal injury (reflux esophagitis, reflux stricture, BE, and esophageal adenocarcinoma) [1] . Extra-esophageal syndromes were subdivided into established associations (reflux cough, reflux laryngitis, reflux asthma, and reflux dental erosion syndromes) and proposed associations (pharyngitis, sinusitis, idiopathic pulmonary fibrosis, and recurrent otitis media).

With extra-esophageal reflux symptoms, the diagnostic yield of documentation of GERD on endoscopy and ambulatory reflux monitoring is lower than that established for typical GERD. The accuracy of available diagnostic tests, including laryngoscopy, upper endoscopy , pH-metry, and pH-impedance testing, for the evaluation of suspected extra-esophageal reflux symptoms is suboptimal [19], and contributes substantially to health-care expenditures. In fact, the initial year’s cost for the workup and management of suspected extra-esophageal reflux symptoms may be more than five times than that for typical GERD [20].

At initial presentation, an empiric therapeutic trial of PPI constitutes a commonly employed approach to diagnosis , with symptomatic response to this trial confirming clinical suspicion of GERD. Initial reports of this approach used omeprazole 40 mg before breakfast and 20 mg before dinner for 7 days, and 80 % of GERD patients with heartburn reported symptom improvement, compared to 42 % of patients with heartburn in the absence of GERD [21]. When symptom response to 7 days of twice-daily omeprazole is compared to abnormal acid exposure or erosive esophagitis on endoscopy, the PPI trial has a sensitivity of 75–80 %, but specificity of 55 % [21, 22]. In one study with GERD defined as the presence of erosive esophagitis on endoscopy, a PPI trial had similar sensitivity to acid exposure and symptom index (SI) on 24-h pH monitoring (83 vs. 80 %) [23]. In a meta-analysis incorporating 15 studies investigating the accuracy of empiric PPI trials as a diagnostic strategy for GERD (using ambulatory pH monitoring as the reference standard), the positive likelihood ratio was 1.63–1.87, sensitivity 78 %, and specificity 54 % [24].

Response to PPI trials in non-GERD heartburn has to be interpreted with caution, since there is overlap with other processes that may also improve with antisecretory therapy (such as eosinophilic esophagitis, EoE) or as a placebo effect (such as functional heartburn). Further, antisecretory therapy may not be as effective at improving GERD symptoms in nonerosive disease compared to erosive esophagitis, and PPI nonresponders could still have reflux-triggered symptoms [1, 14, 25]. Nevertheless, lack of response to PPI therapy carries a high negative predictive value for the diagnosis of GERD, and it at least suggests need for further esophageal investigation. Despite the limited specificity of empiric PPI trials, simplicity and limited cost have established their universal utility in the initial evaluation and management of suspected GERD symptoms [26].

The diagnostic yield of empiric PPI therapy for most atypical symptoms, apart from NCCP, is worse than for typical symptoms. Two meta-analyses assessing the accuracy of PPI treatment as a diagnostic test for NCCP (with pH monitoring and/or endoscopy serving as reference standards) found a sensitivity of 80 % and specificity of 74 % [27, 28]. In contrast, the yield of empiric PPI for suspected extra-esophageal symptoms of GERD is abysmal. For example, a Cochrane meta-analysis found no apparent significant differences in symptomatic improvement between 2 and 3 months of PPI therapy and placebo for nonsmokers with chronic cough and normal spirometry [29]. Similarly, in nonsmokers with chronic cough randomized to twice-daily PPI or placebo for 3 months, no differences were found between PPI and placebo in cough-related quality of life or symptoms, even in a subset with positive pH monitoring [30]. These data highlight the fact that extra-esophageal symptoms often have multifactorial etiologies; GERD may represent a cofactor rather than the sole etiology for symptom generation.

Endoscopic definitions of GERD hinge on identification of esophageal mucosal injury visible to the endoscopist. The Montreal classification defined esophageal complications of GERD to include reflux esophagitis, hemorrhage, stricture , BE , and adenocarcinoma. Reflux esophagitis, the most common form of mucosal injury, may be seen as breaks in the distal esophageal mucosa immediately proximal to the squamocolumnar junction on upper endoscopy. Developed by the International Working Group for the Classification of Oesophagitis (IWGCO), the Los Angeles (LA) classification (named for an initial presentation at the 1994 World Congress of Gastroenterology in Los Angeles) is widely used to grade the severity of reflux esophagitis, with its definitive form published in 1999 [31, 32]. The LA classification describes increasing endoscopic grades of severity of esophagitis as follows: grade A, mucosal break(s) < 5 mm in length and not extending between the tops of two mucosal folds; grade B, mucosal break(s) > 5 mm in length, extending across the tops of two mucosal folds; grade C, mucosal break(s) continuous between tops of at least two mucosal folds but not involving > 75 % of esophageal circumference; and grade D, mucosal break(s) involving > 75 % of the esophageal circumference.

There are limited data to suggest that LA grade A esophagitis may rarely be encountered in healthy asymptomatic individuations (e.g., in as many as 8 % of control subjects in one study [33]), but higher grades are rarely seen in the absence of pathologic GERD. The LA grade of esophagitis at presentation has been described to predict healing with PPI therapy, with the highest healing rates described for LA grade A, and lowest for LA grade D. The likelihood of relapse following discontinuation of therapy is highest with LA grade D [34, 35]. The increasing popularity of empiric PPI trials and over-the-counter availability of these agents have further reduced the likelihood of finding esophagitis on endoscopy, limiting the role of endoscopy to the evaluation of treatment failures and complications in the presence of alarm symptoms [14].

While the identification of esophagitis defines erosive GERD (ERD) , a significant proportion of reflux disease is nonerosive (with no mucosal breaks visible at endoscopy), termed nonerosive reflux disease (NERD) . With the increase in popularity of empiric PPI therapy resulting in high likelihood of healing of esophagitis, there has been a diagnostic shift towards NERD in recent decades, since patients on PPI therapy are significantly more likely to be classified as NERD compared to PPI-naïve patients [36]. Population-based estimates suggest only about one third of GERD patients have ERD, with the remaining two thirds falling under the umbrella of the NERD phenotype [9, 37]. While the presence of erosive esophagitis can confirm GERD, the converse is not true: the absence of esophagitis on endoscopy does not rule out GERD, and pH monitoring is necessary to diagnose NERD. In the presence of endoscopically normal mucosa, histologic findings have poor diagnostic yield in GERD [38] (see section “Esophageal Histopathology and Mucosal Integrity”).

However, the finding of intestinal metaplasia on histopathology from suspected esophageal BE segments has a high concordance with abnormal esophageal acid exposure [39], but not necessarily with reflux symptoms [37, 40]. BE develops in patients with presumed genetic predisposition in the setting of prolonged esophageal reflux exposure, as a protective mechanism against corrosive injury and symptoms; therefore, BE segments are less sensitive to acid-triggered symptoms. Population screening suggests BE prevalence of 1.6 % in an adult asymptomatic Swedish population, while the prevalence of BE in diagnosed GERD can be up to 13 % in high-risk groups (chronic GERD, older age, white men) [37]. Although BE is a premalignant condition, risk estimate of development of esophageal adenocarcinoma is approximately 0.5 % per year [41]. Therefore, while targeted screening for BE is recommended in predisposed individuals, population screening is not cost-effective, in terms of both diagnosing reflux disease and esophageal cancer prevention. Nevertheless, the confirmation of BE on histopathology from endoscopic biopsies defines the presence of GERD and establishes the need for therapy of GERD [26] .

Ambulatory pH monitoring assesses and quantifies esophageal acid exposure times, and it helps determine if symptoms co-occur in close proximity to reflux events in assessing symptom–reflux association [42] . Catheter-based ambulatory pH monitoring was introduced in the 1970s for determining esophageal acid exposure over the course of a 24-h period. The most intuitive metric from ambulatory pH monitoring is the acid exposure time (AET, or the fraction of total recording time at pH < 4.0). AET thresholds defining abnormal acid exposure off PPI therapy fall into a narrow range around 4–5 % [42, 43]. While there has been a recent interest in differentiating asleep and awake acid exposure, the analysis of pH monitoring has traditionally been separated by body position—upright or supine. Because acid reflux events occur more frequently in the upright compared to the supine position, in both asymptomatic controls and patients with GERD, acid exposure times are higher in the upright position compared to the supine position [44, 45]. Consequently, the thresholds defining abnormal esophageal acid exposure in the upright position (range of ~ 6–10 %) are much greater than those for the supine position (in a range of ~ 1–6 %) [46–51] .

For patients tested on PPI therapy, a more stringent total distal AET threshold of 1.6 % has been proposed and studied [52, 53]. Wireless pH systems are now available with longer monitoring periods, better patient acceptance, and less restriction of daily activities during the ambulatory study [54]. With these wireless pH systems, recordings of 48–96 h are possible with extended battery life in the portable recording device, but swallowed acidic material cannot be reliably differentiated from acidic reflux events without stringent patient diary recordings of oral intake. With wireless pH monitoring, the 95th percentile for distal esophageal AET for controls over 2-day recordings was 5.3 %, slightly higher than that reported for catheter-based pH systems [55]. Day-to-day variation in AET has been well characterized using wireless pH monitoring, raising questions about the validity of borderline AET elevations on a 24-h study or on any one day of a multiple-day wireless pH study [43, 55]. Nevertheless, abnormal AET is commonly utilized for quantitation of acid exposure in patients with symptoms incompletely responding to antisecretory therapy, or when documentation of acid exposure is needed prior to anti-reflux surgery.

The DeMeester score was developed to quantify esophageal acid exposure as a composite of six measurements extracted from an ambulatory pH study: (1) percentage of total recording time with pH < 4, (2) percentage of upright recording time with pH < 4, (3) percentage of supine recording time with pH < 4, (4) total number of reflux events, (5) number of reflux events > 5 min in duration, and (6) duration of longest reflux event [46]. DeMeester scores of > 14.7–14.9 are commonly considered abnormal [56] .

pH testing off antisecretory therapy is typically utilized for evaluation of patients with a low index of suspicion for GERD or to document reflux in patients being evaluated for endoscopic or surgical anti-reflux therapies. pH testing on therapy does not have as much clinical utility, as pH-impedance testing can provide additional information regarding weakly acidic reflux episodes which may not be detected by pH testing alone. This option is typically utilized to assess patients with known reflux disease with refractory symptoms incompletely responsive to antisecretory therapy, primarily to investigate the presence of persistent reflux parameters despite appropriate antisecretory therapy.

Impedance monitoring is based on recording resistance to flow of tiny electrical currents across pairs of electrodes on an esophageal catheter. Reflux episodes are identified when retrograde decreases of > 50 % in impedance values (corresponding to the presence of refluxate adjacent to the electrodes) are detected across at least three consecutive distal pairs of impedance electrodes [43]. Therefore, the primary advantage of impedance testing over traditional pH testing lies in its ability to detect reflux events regardless of pH, thus detecting weakly acidic reflux and allowing testing on antisecretory therapy.

The first consensus on the use of esophageal multichannel intraluminal impedance (MII) in the evaluation of reflux episodes was published in 2004 [57]. This consensus proposed a distinction between acid (pH < 4), weakly acid (pH 4–7), and nonacid (or weakly alkaline; pH > 7) reflux. Combined MII–pH monitoring thus has greater sensitivity over traditional pH testing alone to detect reflux events. The gain in detection of reflux over pH monitoring is mainly from detection of weakly acid and nonacid reflux episodes, thereby allowing the test to be performed on PPI therapy. Since neutralization of mucosal acidification typically lags behind clearance of refluxate from the esophagus, pH-detected reflux events tend to be longer than impedance-detected events. Hence, bolus contact time with a pair of impedance electrodes in the distal esophagus tends to be significantly shorter than acid exposure times [49]. The impedance correlate of AET is the reflux exposure time (RET), or the fraction of time refluxate is in contact with the distal esophageal impedance electrode 5 cm above the LES (corresponding to the distal esophageal pH sensor). A multicenter examination of healthy controls helped establish a threshold of 1.4 % for an abnormal RET [49]. Despite this development of normative thresholds, RET has not been shown to represent a robust predictor of treatment outcome following reflux therapy [58].

The total numbers of reflux events on ambulatory reflux monitoring have been proposed as a means of defining GERD. Two studies (one American, one European) found very similar 95th percentile values of 73–75 reflux events on 24-h pH-impedance monitoring in healthy volunteers, implying that higher numbers of reflux events suggest the diagnosis of GERD [49, 50]. Recent data suggest that lower thresholds for total reflux events may identify GERD as low as 53 off PPI may be distinctive of GERD [59].

In the setting of antisecretory therapy, acid reflux events decrease while weakly acid reflux events are detected more often. In a landmark study utilizing pH-impedance monitoring before and after omeprazole therapy, acid reflux events significantly decreased, but the numbers of nonacid reflux events almost doubled, despite similar total numbers of reflux events [60]. While heartburn improved following omeprazole therapy, regurgitation events were reported more often. Other reports suggest a reduction in numbers of reflux events with antisecretory therapy in patients with GERD, presumed from reduced volume of gastric secretion [61]. Consequently, the thresholds utilized for numbers of reflux events indicative of GERD are lower when pH-impedance monitoring is performed on PPI therapy. The 95th percentile of normal values for total numbers of reflux events when testing is performed on PPI therapy have ranged from 48 to 57 [49, 59].

Outcome studies with characterization of reflux solely based on numbers of reflux events in the absence of abnormal AET or other reflux parameters are limited. While numbers of reflux events do decrease significantly with anti-reflux surgery in these instances [62], the thresholds alone may not necessarily segregate those with good response to therapy [63]. This may be partly related to the fact that duration of individual reflux events may vary dramatically, and patients with low numbers of reflux events could have significant acid or reflux exposure in the esophagus if prolonged. However, reflux events do have relevance in assessing correlation of symptoms with reflux events.

In addition to quantitation of esophageal acid exposure and reflux events, pH and pH-impedance monitoring can assess correlation of reflux events with esophageal symptoms. The two tests used most often are SI and symptom association probability (SAP). For pH-impedance testing, these symptom–reflux parameters may be calculated for acid-detected reflux events as well as impedance-detected reflux events. The SI is calculated as a simple ratio of the number of reflux-related symptoms to the total number of symptom episodes [64]. Analyses utilizing receiver operating characteristic curves designated a threshold of SI > 50 % as positive for heartburn episodes [65].

Two methods of calculating SAP have been proposed. The Weusten method, used most commonly, involves dividing the 24-h recording time into consecutive 2-min periods [66]. Next, 2 × 2 contingency tables are constructed, depicting the presence or absence of symptoms versus the presence or absence of reflux for each period. Fisher’s exact test is then used to calculate the p value across the contingency table, representing the probability that symptoms and reflux are related by chance alone [66]. The SAP can also be calculated using the Ghillibert probability estimate (GPE), which represents the sum of partial probabilities for the exact numbers of reflux-associated symptoms within the context of the total number of symptoms, taking the total duration of the study and the total exposure time into account [67]. Regardless of how the SAP is calculated, it is considered positive if > 95 %, corresponding to p < 0.05, or a < 5 % chance that the observed association between symptoms and reflux occurred by chance. The Weusten and Ghillibert approaches to SAP can be used virtually interchangeably (with major discordance found in less than 3 % of cases), though the SI may be discordant with SAP, especially in the setting of limited or frequent symptoms [68]. Because of its ability to detect more reflux events, pH-impedance testing increases the yield of detecting a positive symptom–reflux association over traditional pH testing alone [69], especially when performed off pH therapy [58].

Symptom–reflux association is the weakest link in ambulatory pH and pH-impedance monitoring, since it is heavily reliant on patients promptly designating presence of symptoms on their event logger [70]. However, symptom–reflux association has value when positive in particular settings. It contributes to the strength of reflux evidence identified on ambulatory monitoring, especially since patients with strong GERD evidence (both abnormal acid exposure and positive symptom–reflux association) have the best symptomatic outcome with anti-reflux therapy [71, 72]. In the setting of physiologic reflux parameters, positive symptom–reflux association identifies a subgroup of patients with characteristics more akin to functional esophageal disease than GERD. Even though previously classified under the NERD umbrella or termed “acid sensitive,” these patients share psychosomatic and HRQOL characteristics similar to patients with functional heartburn than to true GERD [73]. Reflux hypersensitivity has been used to describe settings where symptom–reflux association is positive on pH-impedance testing that detects all reflux episodes regardless of pH, shifting patients previously diagnosed as functional heartburn with a negative pH study into this category using pH-impedance monitoring [73].

There are several factors that impact the clinical utility of symptom–reflux association. The calculations are highly reliant on symptom episodes, which can vary widely depending on symptom perception, and patient compliance with symptom reporting [74]. Specifically, very high or very low numbers of symptom episodes can significantly influence the calculation of SI [68]. SAP estimates may have better value in these instances because they take into account periods without symptoms (where reflux exposures may also be limited). SI and SAP indices can be over-interpreted, especially in the absence of high rates of reflux [70]. Therefore, a positive symptom–reflux correlation result is much more clinically useful than a negative result in evaluating GERD.

Barium esophagrams are often performed in the setting of esophageal symptoms, but have limited utility in the diagnosis of GERD. Although the overall sensitivity for detection of esophagitis (seen as a reticular or finely nodular pattern) may be around 65 %, the sensitivity decreases for milder grades of esophagitis [20]. Barium radiology without any provocative maneuvers detects one third to one half of patients with GERD [75, 76]; evidence of reflux can be seen with provocative maneuvers in as many as 70 % [77]. The main issue with barium esophagography in GERD is that the most important mechanism of GERD, TLESR, can occur in the normal subject, which can result in reflux of barium from the stomach high into the esophagus in the supine position. On the other hand, if no TLESR is provoked during the study, a patient with reflux disease may have a normal study. Therefore, the sensitivity and specificity of barium studies for diagnosis of GERD make this test inadequate to serve as a screening procedure for GERD [26, 78]. However, barium radiograms provide excellent anatomic detail and are important in assessing complications of GERD (such as a stricture or ring) or evaluating the anatomy of the esophagus prior to intervention [26, 79].