Dr M. Frazzoni, Medicina Interna e Gastroenterologia, Ospedale S. Agostino, P.zza S. Agostino 228, 41100 Modena, Italy. E-mail: email@example.com
Background : Patients with endoscopy-negative heartburn can be subdivided into non-erosive reflux disease and functional heartburn on the basis of abnormal and normal, respectively, oesophageal acid exposure. Different pathophysiological characteristics could explain the reportedly low efficacy of proton pump inhibitors in functional heartburn.
Aim : To assess if non-erosive reflux disease and functional heartburn are pathophysiologically distinguishable.
Methods : Oesophageal manometry and pH-monitoring were performed in 145 patients with endoscopy-negative heartburn, in 72 patients with erosive reflux disease, in 58 patients with complicated reflux disease, and in 60 controls.
Results : Patients with non-erosive reflux disease (84 cases) and functional heartburn (61 cases) differed with regard to the prevalence of hiatal hernia (49% vs. 31%, P = 0.008), the mean lower oesophageal sphincter tone (18.5 vs. 28.4 mmHg, P < 0.05), and the number of upright diurnal acid refluxes lasting more than 5 min (3.6 vs. 0.37, P < 0.05). The results were very close in thenon-erosive reflux disease, erosive reflux disease and complicated reflux disease groups, whilst patients with functional heartburn were indistinguishable from controls.
Conclusions : Pathophysiological characteristics typical of gastro-oesophageal reflux disease are found in patients with non-erosive reflux disease but not in patients with functional heartburn. This could explain the reportedly low efficacy of proton pump inhibitors in functional heartburn and suggests considering different management strategies.
The term gastro-oesophageal reflux disease (GERD) describes any symptomatic condition or histopathologic alteration resulting from episodes of gastro-oesophageal reflux.1 The dominant mechanism of symptom production in GERD is by contact of the oesophageal mucosa with acid and heartburn is the most common symptom.2 Oesophageal mucosal damage is related to rising levels of oesophageal acidification.3, 4 Up to one-third of patients have endoscopic evidence of erosive reflux oesophagitis and up to one-fifth have complicated reflux disease (CRD) characterized by ulcerative oesophagitis, oesophageal strictures and Barrett's oesophagus.1 The term non-erosive reflux disease (NERD) defines endoscopy-negative patients with reflux symptoms and abnormal oesophageal acid exposure during ambulatory 24-h oesophageal pH-monitoring.1 The term functional heartburn (FH) has been proposed for endoscopy-negative patients with episodic retrosternal burning and normal findings on ambulatory 24-h oesophageal pH-monitoring.5
Heartburn is considered to be most likely due to GERD even in the absence of definite endoscopic oesophagitis, in spite of lack of evidence-based documentation of its positive predictive value.2 The rapid heartburn relief induced by proton pump inhibitors in patients with reflux oesophagitis6 has led to the proposal of using the symptomatic response to these powerful acid-lowering drugs in patients with heartburn as a non-invasive diagnostic test for GERD.7 Interestingly, symptomatic response to a proton pump inhibitor was found to be much lower in heartburn patients with normal oesophageal acid exposure (FH) than in those with abnormal results on oesophageal pH-monitoring (NERD).8 Given the different responses of FH and NERD to acid-suppressing therapy, it could then be argued that heartburn is mainly acid-related in NERD and non-acid related in FH. Consequently, it could be hypothesized that patients with endoscopy-negative heartburn are a heterogeneous population and that FH does not strictly belong to the GERD spectrum.
Gastro-oesophageal reflux disease arises from a failure of the antireflux barrier. Currently, three mechanisms have been identified that allow reflux of the gastric contents across the oesophagogastric junction and into the oesophagus: weakness of the lower oesophageal sphincter (LOS), presence of hiatal hernia and transient LOS relaxations.9, 10 A thorough investigation on the main pathophysiological characteristics of patients with endoscopy-negative heartburn has not yet been performed in a large series and different findings in NERD and FH could further substantiate the hypothesis that patients with endoscopy-negative heartburn are a heterogeneous population. Aim of the present study was to assess if NERD and FH are pathophysiologically distinguishable.
Materials and methods
We studied patients referred for ambulatory 24-h oesophageal pH-monitoring to our open-access gastro-intestinal service from February 1999 to February 2004. The reasons for referral were diagnosis of NERD in patients with endoscopy-negative heartburn, evaluation of atypical manifestations possibly related to GERD (asthma, chronic laryngitis, chronic cough, non-cardiac chest pain, sleep apnoea syndrome), evaluation for possible antireflux surgery, and the tailoring of medical management in CRD (Barrett's oesophagus of any length with intestinal metaplasia documented at histology; peptic oesophageal strictures; deep peptic oesophageal ulcers).
Patients were included in the present study only if they had been submitted to upper gastrointestinal endoscopy at our endoscopy service, where all physicians train together and adopt standardized criteria to record oesophageal abnormalities. Patients were considered to have CRD when: (i) deep oesophageal ulcers (with or without strictures) were detected (ulcerative oesophagitis), or (ii) a change of oesophageal epithelium of any length was recognized and was confirmed to have intestinal metaplasia by biopsy of the tubular oesophagus (intestinal metaplasia of the cardia excluded) (Barrett's oesophagus).11 Erosive reflux disease (ERD) was considered to be present when at least one mucosal break was detected.3 Hiatal hernia was considered to be present when the gastro-oesophageal junction was detected 3 cm or more above the diaphragmatic impression. Multiple biopsies of the distal oesophagus were taken in all cases when deep oesophageal ulcers were present and when a change of the oesophageal epithelium of any length raised the suspicion of Barrett's oesophagus.11 Exclusion criteria included both treatment with PPIs in the 4 weeks before endoscopy, except for patients in whom a final diagnosis of Barrett's oesophagus was made, and endoscopy performed outside our endoscopy service and/or more than 12 months before oesophageal pH-monitoring.
After written informed consent had been obtained and an adequate wash-out period of any antireflux therapy (at least 20 days) had been completed, patients were evaluated by means of a symptom score, oesophageal manometry and ambulatory 24-h oesophageal pH-monitoring, all performed by the same physician (MF). Patients who were not able to undergo an oesophageal manometric evaluation before pH-metric assessment were excluded from the study. During the 20-day wash-out period, patients were permitted to use an antacid containing a balanced formulation of magnesium and aluminium hydroxides as needed for the relief of heartburn.
Symptoms were assessed using a scoring system validated in a previous study.12 Briefly, they were graded as follows: 0 = none; 1 = mild, symptom could be ignored and was recalled only after specific inquiring; 2 = moderate, symptom could not be ignored and was spontaneously reported, but neither daily activities nor sleep were influenced; 3 = severe, symptom influenced daily activities and/or sleep. Specific (heartburn, acid regurgitation) and aspecific symptoms of GERD (wheezing, hoarseness, cough, chest pain) were assessed as well as symptoms of dyspepsia (chronic or recurrent pain or discomfort centred in the upper abdomen).13
Oesophageal manometry was performed in every patient before pH-monitoring studies according to current standard guidelines.14 The station pull-through method was used to locate the LOS. Briefly, an 8-channel, water perfused manometry catheter was placed nasally into the stomach and slowly withdrawn. The LOS was identified, and then, at each 0.5 cm station the basal tone was measured at the end of expiration and the relaxation was assessed with wet swallows. Oesophageal body peristalsis was assessed and pressures measured with at least 10 wet swallows. The upper oesophageal sphincter was finally identified and the catheter removed. All manometric data were processed using the Polygram for Windows module (Medtronic Functional Diagnostics A/S, Tonsbakken, Denmark) in order to obtain the LOS tone and the mean distal oesophageal amplitude.
For the ambulatory pH-monitoring study, a probe was passed nasally such that an antimony pH electrode was placed 5 cm above the upper border of the LOS (defined by previous manometry). Data were recorded by a Digitrapper Mark III Gold (Medtronic Functional Diagnostics A/S). The patients consumed three meals during the 24-h study period and were instructed to stay in bed only between 22:00 and 07:00 hours. All pH-metric data were processed using the Polygram for Windows module (Medtronic Functional Diagnostics A/S). Acid reflux was considered to take place when pH dropped below pH = 4.0 for at least 4 s. Patients were instructed to report in a diary the time spent in the recumbent position between 22:00 and 07:00 hours and this was considered to be the supine nocturnal period. All the remaining time was considered to be the upright diurnal period. The fraction of time at pH < 4, i.e. the percentage acid reflux time, was computed for the total study period (total percentage acid reflux time) and for the upright diurnal and supine nocturnal periods (upright diurnal percentage acid reflux time and supine nocturnal percentage acid reflux time). Values of at least 5.8% for the total percentage acid reflux time and of at least 3% for the supine nocturnal percentage acid reflux time were considered to be abnormal.4 Acid refluxes lasting for more than 5 min for the upright diurnal and supine nocturnal periods were also computed.
Patients evaluated for atypical manifestations possibly related to GERD and with normal findings at endoscopy (except for the presence of hiatal hernia), a heartburn score <2, a dyspepsia score ranging from 0 to 3, and normal pH-metric results served as controls. Patients with a heartburn score <2 and abnormal pH-metric results were excluded from the study, even if a firm endoscopic diagnosis of CRD or ERD had been made.
For all the continuous variables, the analysis of variance (anova) and the Student–Newman–Keuls test for multiple comparisons were adopted to analyse the differences between the groups. For all the categorical variables, the chi-squared test was used to analyse the differences between the groups and the Bonferroni's correction was applied for multiple comparisons.
P < 0.05 was considered to be significant.
Between February 1999 and February 2004, 335 patients met the enrolment criteria and were included in the study. One-hundred and forty-five had endoscopy-negative heartburn, 130 had endoscopy-positive heartburn (72 with ERD and 58 with CRD), and 60 were considered to be controls. On the basis of ambulatory 24-h oesophageal pH-monitoring results (total and supine nocturnal percentage acid reflux time), 84 patients (58%) with endoscopy-negative heartburn were classified as NERD and 61 (42%) were classified as FH. The baseline characteristics of the five groups of patients so obtained are shown in Table 1. The median interval between endoscopy and oesophageal manometric and pH-metric assessment was 2 months in all five groups of patients. Patients with CRD were older than all the other groups (P < 0.05). Male sex was more prevalent in CRD, ERD and NERD patients than in controls and FH cases (P = 0.000, P = 0.000 and P = 0.04, respectively). The heartburn score was higher than the dyspepsia score in 64 of 145 patients with endoscopy-negative heartburn (44%) (Table 1).
Table 1. Baseline characteristics of patients
Controls (n = 60)
FH (n = 61)
NERD (n = 84)
ERD (n = 72)
CRD (n = 58)
Age: CRD vs. ERD P < 0.05; CRD vs. NERD P < 0.05; CRD vs. FH P < 0.05; CRD vs. controls P < 0.05.
Male sex: FH vs. controls P NS; NERD vs. FH and controls P = 0.04; ERD vs. FH and controls P = 0.000; CRD vs. FH and controls P = 0.000.
Heartburn score > dyspepsia score: FH vs. all other groups P = 0.000.
Table 2 shows that the total percentage acid reflux time was significantly greater in CRD than in ERD and in NERD (P < 0.05). Moreover, the total percentage acid reflux time was significantly greater in ERD than in NERD (P < 0.05). There was no significant difference in the upright diurnal percentage acid reflux time between these three groups of patients, but the supine nocturnal percentage acid reflux time was significantly greater in patients with CRD than in those with ERD or NERD, and in patients with ERD vs. those with NERD (P < 0.05). Normal values for the total percentage acid reflux time were found in eight of the 72 patients with ERD (six had abnormal supine nocturnal percentage acid reflux times) and in two of the 58 patients with CRD (one with an abnormal supine nocturnal percentage acid reflux time). Therefore, the total percentage acid reflux time gave abnormal values in 120 of the 130 GERD patients (92%) with endoscopically-detectable mucosal lesions, but, when the supine nocturnal percentage acid reflux time was also taken into account, abnormal pH-metric values were found in 127 of the 130 cases (97%).
The mean number of refluxes lasting more than 5 min did not differ between the FH patients and controls for both the upright diurnal period (0.37 and 0.41, respectively) and the supine nocturnal period (0.09 and 0.06, respectively) (Figure 1), but it was significantly higher in the NERD group (3.6 in the upright diurnal period and 1.8 in the supine nocturnal period) (P < 0.05). Interestingly, in the NERD and in the ERD groups similar results were found with regard to the upright diurnal period (3.6 and 3.8, respectively), but in the ERD group values significantly higher than in the NERD group were observed for the supine nocturnal period (3.1 vs. 1.8) (P < 0.05).
The prevalence of hiatal hernia did not significantly differ between the FH patients and controls (31 and 25%, respectively), but it was found to be significantly higher in the NERD, ERD and CRD groups (49, 68 and 64%, respectively) (P = 0.008, P = 0.000 and P = 0.000, respectively) (Figure 2).
No difference was observed between the FH patients and controls with regard to the mean LOS tone (28.4 and 26.9 mmHg, respectively) which, conversely, was found to be significantly lower in the NERD, ERD and CRD groups (18.5, 17 and 16 mmHg, respectively) (P < 0.05) (Figure 3).
The mean distal oesophageal amplitude did not differ between the NERD group, the FH group and controls (87.2, 101.5 and 99.3 mmHg, respectively), whilst significantly lower values were observed in patients with ERD and CRD (69.7 and 65 mmHg, respectively) (P < 0.05) (Figure 4).
In the present study, we have shown that patients with endoscopy-negative heartburn are a pathophysiologically heterogeneous population. After dividing them into NERD and FH on the basis of the total and supine nocturnal percentage acid reflux time as assessed by oesophageal pH-monitoring, we found that they differed with regard to the prevalence of hiatal hernia, the mean LOS tone and the number of upright diurnal refluxes lasting more than 5 min: in the NERD group the results were very close to the ERD and CRD groups, whilst in the FH group the results were indistinguishable from those found in controls.
Gastro-oesophageal reflux occurs when the pressure gradient between the oesophagogastric junction and the stomach is lost. Both hiatal hernia and weakness of the LOS favour gastro-oesophageal reflux by reducing the pressure gradient between the oesophagogastric junction and the stomach.9, 10, 15, 16 In the present study we found that the prevalence of hiatal hernia did not significantly differ between the FH patients and controls, but it was significantly higher in the NERD, ERD and CRD groups (Figure 2); moreover, no difference was observed between the FH patients and controls with regard to the mean LOS tone which, conversely, was found to be significantly lower in the NERD, ERD and CRD groups (Figure 3). Therefore, pathophysiological characteristics favouring gastro-oesophageal reflux were found in patients with NERD but not in patients with FH.
Transient LOS relaxations are defined as prolonged relaxations of the LOS which are not associated with a swallow.10, 15 They constitute a major mechanism of reflux only in patients without hiatal hernia.17 Transient LOS relaxations are no more frequent in patients with GERD than in asymptomatic volunteers but only in the former they are followed by acid refluxes.18 Accordingly, the total rate of gastro-oesophageal reflux episodes is similar in patients with GERD and controls but in the former a higher proportion and rate of acid reflux than in the latter has been shown.19 In our study transient LOS relaxations were not directly assessed owing to the high inconvenience to the patient of prolonged ambulatory oesophageal manometric monitoring. As the perception of reflux symptoms depends on the duration of acid-exposure episodes 20 and refluxes because of transient LOS relaxations occur in the daytime,16 we evaluated upright diurnal acid refluxes lasting more than 5 min, as well as supine nocturnal acid refluxes lasting more than 5 min. Acid refluxes lasting more than 5 min may be considered to reflect long-lasting, i.e. transient, LOS relaxations in the upright diurnal period (drainage by gravity, primary peristalsis and salivation all present) and defective secondary oesophageal peristalsis in the supine nocturnal period (drainage by gravity, primary peristalsis and salivation all absent). They are not strictly dependent from the total and supine nocturnal percentage acid reflux time, the pH-metric parameters that we adopted to separate NERD from FH patients, and this is confirmed by somehow different results for the two sets of parameters in our series (Table 2, Figure 1). In particular, the upright diurnal acid refluxes lasting more than 5 min did not differ between the NERD and the ERD group whilst, again, no difference was found between the FH group and controls. Accordingly, it is unlikely that transient LOS relaxations followed by acid refluxes occur with comparable frequencies in FH and NERD, further confirming that FH and NERD are two pathophysiologically distinct clinical entities.
We found that the mean distal oesophageal amplitude was lower in patients with ERD and CRD than in patients with NERD, and that the supine nocturnal acid refluxes lasting more than 5 min were more frequent in the ERD and the CRD groups when compared with the NERD group, confirming that ineffective oesophageal motility is associated with prolonged oesophageal clearance in the supine position.21
We did not attempt to subdivide the FH group on the basis of a positive symptom-reflux association. Although the clinical gain of studying symptom association is clear, there is lack of agreement on how to correlate reflux episodes and symptoms, the cut-off points defining a positive score are arbitrary, the most appropriate computer software is not widely available and the methodology may still not be developed sufficiently.2, 22 Moreover, the average percentage of perceived acid reflux events during ambulatory oesophageal pH-monitoring is reportedly very low (<4%)23 and this casts further doubts on the reliability of symptom-reflux scores.
In this study we confirm that ambulatory oesophageal pH-monitoring has a high sensitivity in GERD patients with endoscopically-detectable mucosal lesions if the total percentage acid reflux time is combined with the supine nocturnal percentage acid reflux time (97% in the present series).4 In patients with endoscopy-negative heartburn we found abnormal pH-metric values in 58% of cases, a percentage higher than that reported by others (45%).23 These values do not reflect the sensitivity of oesophageal pH-monitoring in patients with endoscopy-negative heartburn, however, because an independent standard of accuracy such as endoscopy is lacking in this subset of patients with GERD. Heartburn cannot be regarded as the gold standard for diagnosing GERD in endoscopy-negative patients: in the landmark study concerning the symptoms of GERD,24 a high specificity for heartburn in endoscopy-negative patients was found only when heartburn dominated the patient's complaints but when it was a co-dominant symptom its specificity dropped to about 50%. Dyspeptic symptoms are reportedly present in one half of endoscopy-negative heartburn cases and frequently co-dominate the clinical picture.25 In our series, we found that the heartburn score was higher than the dyspepsia score in less than one half of patients with endoscopy-negative heartburn: therefore, a reliable diagnosis of GERD could not have been made on the basis of symptoms for the majority of these patients. Therefore, ambulatory oesophageal pH-monitoring should still be regarded as the most reliable test for the diagnosis of GERD.26
Patients with endoscopy-negative heartburn are less responsive to proton pump inhibitors than patients with ERD.25, 27 Symptomatic response to a proton pump inhibitor was found to be much lower in heartburn patients with normal oesophageal acid exposure (FH) than in those with abnormal results on oesophageal pH-monitoring (NERD).8 Our data further support the concept that patients with endoscopy-negative heartburn are a heterogeneous population:27 they can be pH-metrically divided into two distinct clinical entities, namely NERD and FH, the former only with pathophysiological characteristics typical of GERD whilst the latter pathophysiologically indistinguishable from controls. It appears then that heartburn is acid-reflux related in NERD and non-acid reflux related in FH and that FH does not strictly belong to the GERD spectrum.
According to authoritative suggestions, in patients with endoscopy-negative heartburn management strategies should be tailored to address the specific mechanism for symptom generation.25, 27 Pain modulators such as tricyclic antidepressants, trazodone, and selective serotonin reuptake inhibitors could have a role in patients in whom heartburn is not acid-related.28 In keeping with our data, by means of ambulatory oesophageal pH-monitoring a sub-group of patients can be reliably identified, namely FH, with pathophysiological characteristics not typical of GERD and in whom acid reflux has a low probability of being implicated in the genesis of heartburn. This could explain the reportedly low efficacy of powerful acid-lowering drugs such as proton pump inhibitors in patients with FH and suggests considering different management strategies, such as pain modulators.
There was no financial support for the present study.