Gastro-oesophageal reflux disease: symptoms, erosions, and Barrett's--what is the interplay?

T he outcomes of patients with erosive oesophagitis, treated with acid suppression therapy (proton pump inhibitors), has been dictated by the baseline severity of erosive oesophagitis, presence of hiatus hernia, duration of therapy and, in some studies, by the Helicobacter pylori status of the patients. 2 It has been shown that higher grades of erosive oesophagitis (Los Angeles grades C and D) have significantly lower healing rates as opposed to those with lower grades of erosive oesophagitis (grades A and B). Moreover, the majority of the oesophagitis trials have evaluated healing at four and eight weeks, showing a higher proportion of patients with all grades of erosive oesophagitis healed at week 8 compared with week 4. 4 Similar data on healing at .8 weeks are not consistently available in the literature. Not only do patients with severe grades of erosive oesophagitis have a higher degree of oesophageal acid exposure compared with those with either no oesophagitis or low grades of oesophagitis, but they also have low amplitude of oesophageal contractions and the presence of large hiatus hernias. Therefore, it is not surprising that the poor pathophysiology associated with severe erosive oesophagitis leads to poor healing rates. Although a few studies have correlated H pylori status with oesophagitis healing, with H pylori positivity associated with improved healing rates, this has not been consistently documented. This may be a phenomenon related not just to the presence or absence of H pylori infection but rather to the pattern of gastritis, presence of hiatus hernia, acid output states, etc. Although patients with Barrett’s oesophagus also have abnormal pathophysiology, very similar to patients with severe grades of erosive oesophagitis, the impact of the presence of Barrett’s oesophagus in patients with erosive oesophagitis has not been systematically evaluated. In fact, previous trials of erosive oesophagitis have excluded patients with Barrett’s oesophagus and therefore the effect of healing of erosive oesophagitis in the presence of Barrett’s oesophagus is not known. In this issue of Gut, Malfertheiner and colleagues report results from the Progression of gastro-oesophageal reflux disease (ProGORD) trial, a large, multicentre, prospective, follow up study of 6215 patients with reflux disease treated with esomeprazole (open label) (see page 746). Results for heartburn resolution in patients with erosive oesophagitis and non-erosive reflux disease (NERD) were presented for the last visit and the prognostic influence of the baseline grade of erosive oesophagitis, presence of Barrett’s oesophagus, age, sex, body mass index, and H pylori infection was studied on the healing of erosive oesophagitis and, for NERD patients, on complete resolution of heartburn. Barrett’s oesophagus was detected in 14% of patients with erosive oesophagitis and in 2.3% of NERD patients. The overall healing rates of erosive oesophagitis at eight weeks in all patients (with and without Barrett’s oesophagus) was 77.5%; 79.3% in grades A and B compared with 69.9% in grades C and D (p,0.0001). In patients without Barrett’s oesophagus, the healing rate of oesophagitis was 79.3% compared with 66.7% in those with Barrett’s (p,0.0001). These eight week healing rates in patients with Barrett’s oesophagus were also directly related to baseline oesophagitis severity (78.6% in grades A and B; 63% in grades C and D). Healing rates were lower in those with ‘‘confirmed Barrett’s oesophagus’’ (with histological documentation of intestinal metaplasia) and also those with endoscopic Barrett’s oesophagus (that is, oesophageal columnar segment). Whereas the presence of severe grades of erosive oesophagitis (that is, C and D) have been shown to influence healing of erosive oesophagitis, this is one of the initial reports to show the presence of Barrett’s oesophagus as having a negative impact on healing of erosive oesophagitis. Systematic biopsies were not obtained from the oesophageal columnar segment; the number of biopsies and endoscopic measurement of the length of Barrett’s oesophagus were also not standardised between participating centres. Although all endoscopists were trained on the LA classification system for erosive oesophagitis, the diagnosis of Barrett’s oesophagus was performed without any predetermined criteria. Furthermore, obtaining biopsies from the oesophagus were left up to the discretion of the endoscopists; additional biopsies were requested but were not mandatory from the endoscopists. It is well known that there is large interobserver variability in the endoscopic recognition of the oesophageal columnar segment and that detection of intestinal metaplasia is directly related to the endoscopy/biopsy technique and number of biopsies obtained. 9 Moreover, it is possible that patients with higher grades of erosive oesophagitis (grades C and D) may be more likely to have been included in the ‘‘Barrett’s group’’ as inflammatory lesions might have been mistaken as columnar areas in the distal oesophagus. Complete symptom resolution, as determined by a validated reflux disease questionnaire, was 58.5% at two weeks and 64.8% at the last visit in the NERD group compared with 61.1% and 70.4%, respectively, in the oesophagitis group. Thus the absolute difference in patients with heartburn resolution between the oesophagitis and NERD groups at the last visit was 5.6%, suggesting that these are relatively similar patient groups in terms of both pathophysiology and treatment response. These data however do not reflect the same point in time in each group and although the comparison is not ideal, this highlights the fact that complete symptom resolution is difficult to achieve. Symptom resolution (measured by validated questionnaires) can be achieved in approximately 60–75% of GORD patients treated with proton pump inhibitor therapy and although the numbers may be numerically higher in patients with erosive oesophagitis, they are still nowhere closer to healing rates, suggesting that symptoms are more resistant to acid suppression than mucosal breaks (that is, erosions). On the other hand, it is not clear if patients actually seek complete symptom resolution and maybe goals such as complete resolution of symptoms as evaluated in this and other COMMENTARIES 739