Evolution of gastro-oesophageal reflux disease over 5 years under routine medical care the ProGERD study

Transcription

1 Alimentary Pharmacology and Therapeutics Evolution of gastro-oesophageal reflux disease over 5 years under routine medical care the ProGERD study P. Malfertheiner*, M. Nocon, M. Vieth à, M. Stolte, D. Jaspersen, H. R. Koelz**, J. Labenz, A. Leodolter àà, T. Lind, K. Richter & S. N. Willich *Dept of Gastroenterology, Hepatology and Infectious Diseases, Ottovon-Guericke University, Magdeburg, Germany. Institute for Social Medicine, Epidemiology and Health Economics, Charité University Medical Centre, Berlin, Germany. à Institute of Pathology, Klinikum Bayreuth, Bayreuth, Germany. Klinikum Kulmbach, Kulmbach, Germany. Klinikum Fulda, Fulda, Germany. **Division of Gastroenterology, Department of Internal Medicine, Triemli Hospital, Zurich, Switzerland. Department of Internal Medicine, Ev. Jung-Stilling Hospital, Siegen, Germany. àà Klinik für Gastroenterologie, Hepatologie, Stoffwechsel- und Infektions Krankheiten, Remscheid, Germany. AstraZeneca R&D, Mölndal, Sweden. AstraZeneca GmbH, Wedel, Germany. Correspondence to: Prof. Dr med. P. Malfertheiner, Ottovon-Guericke-University, Department of Gastroenterology, Hepatology and Infectious Diseases, Leipziger Str. 44, Magdeburg, Germany. peter.malfertheiner@med.ovgu.de Publication data Submitted 30 June 2011 First decision 17 July 2011 Resubmitted 7 October 2011 Accepted 7 October 2011 EV Pub Online 9 November 2011 SUMMARY Background The evolution of gastro-oesophageal reflux disease (GERD) under current management options remains uncertain. Aim To examine whether, depending on the initial presentation, non-erosive (NERD) and erosive reflux disease (ERD) without Barrett s oesophagus will progress to more severe disease under current routine care following the resolution of the initial condition. Methods Patients with the primary symptom of heartburn were included at baseline, and stratified into non-erosive (NERD) and erosive reflux disease (ERD), LA grades A D (Los Angeles classification). After a 2- to 8-week course with esomeprazole therapy to achieve endoscopic healing in ERD and symptom relief in NERD, patients were treated routinely at the discretion of their physician. We report oesophagitis status and the presence of endoscopic and confirmed Barrett s oesophagus after 5 years. Results A total of 6215 patients were enrolled in the study of whom 2721 patients completed the 5-year follow-up. Progression, regression and stability of GERD severity were followed from baseline to 5 years. Only a few patients with NERD and mild moderate ERD progressed to severe forms of ERD and even Barrett s oesophagus. Most patients remained stable or showed improvement in their oesophagitis; 5.9% of the NERD patients, 12.1% of LA grade A B patients and 19.7% of LA grade C D patients in whom no Barrett s oesophagus was recorded at baseline progressed to endoscopic or confirmed Barrett s oesophagus at 5 years. Conclusion Most GERD patients remain stable or improve over a 5-year observation period under current routine clinical care. Aliment Pharmacol Ther 2012; 35: doi: /j x

2 Gastro-oesophageal reflux disease progression INTRODUCTION Gastro-oesophageal reflux disease (GERD) is a prevalent disease with a tendency to further increase in most populations around the world. 1 3 It has a significant impact on patients quality of life, 4, 5 and in subsets of patients, it leads to complications and may predispose to oesophageal adenocarcinoma development. 6 The recent global definition for GERD (the Montreal definition) addresses the complexity of its clinical spectrum that ranges from typical oesophageal symptoms and lesions to a series of atypical symptoms, including laryngeal and pulmonary symptoms. 7 The current strategies for management of GERD include a range of options. Over the years, patients with sporadic reflux symptoms have been reported to go for self-medication with drugs obtained over the counter. 8 More troublesome and relapsing reflux symptoms are mainly taken care of by primary care medicine. For the initial management of severe GERD, and whenever complex therapies are needed in complicated GERD, specialists usually become involved. 8, 9 Around 70% of patients with GERD may suffer from chronic or relapsing symptoms and they require longterm management by either intermittent, on demand or continuous acid suppressant therapy A proportion of patients require continuous acid suppression with proton pump inhibitors (PPI), and selected cases may also 13, 14 be suitable candidates for anti-reflux surgery. In some countries, including Germany, patients with erosive oesophagitis (ERD) are followed up with upper gastrointestinal endoscopy (GI-endoscopy) for initial healing. In patients with Barrett s oesophagus, upper GI endoscopy is recommended regularly, at 2- to 4-year 15, 16 intervals. Long-term treatment and monitoring of GERD have a significant impact on the health economic burden and pose the question as to whether knowledge on the natural history of GERD under current medical care may influence long-term strategies, depending on whether GERD progresses or not in the individual patient. The issue on the natural history of GERD is controversial, as some advocate significant progression of GERD over time, 17, 18 whereas others state the contrary with patients remaining in the initial GERD stage, and little change among phenotypic expressions of nonerosive reflux disease (NERD), erosive reflux disease (ERD) and Barrett. 19 The issue on progression of GERD is of great clinical importance, given that patients who develop Barrett s oesophagus are at increased risk of developing oesophageal adenocarcinoma, with an estimated annual 6, 20 incidence of 0.5 1%. Our study was designed to address the question of whether, depending on the initial presentation, NERD without Barrett and ERD without Barrett will progress to more severe disease under current routine care following the resolution of the initial condition. 21 This study is unique in that it was a prospective study of the evolution of Barrett s oesophagus and observation of changes in LA grading over 5 years under routine clinical care. METHODS The ProGERD study is a large prospective multicentre open cohort study conducted in Germany (predominantly), Austria and Switzerland with approval from local ethics committees. It comprised an initial treatment phase with esomeprazole for up to 4 weeks in NERD patients (20 mg od) and up to 8 weeks in patients with ERD (40 mg od). Endoscopy was conducted at recruitment and repeated to assess healing at 4 weeks, and again at 8 weeks, if necessary. Attendance of Training seminars for the LA classification and the accurate description of the columnar lined epithelium, including biopsies for the definition of Barretts oesophagus, was an essential requirement for participation of each investigator. During the course of the study, investigators received appropriate updated training. Subsequently, patients were followed up on a long-term basis under routine care, i.e. at the discretion of their physician. Routine care may have involved treatment with PPIs, H 2 -receptor antagonists or antacids on a regular (at least one dose of PPI every third day) or on-demand basis, and endoscopic intervention when considered necessary. Patients were followed up regardless of whether or not they responded to initial treatment and endoscopy with biopsy was repeated at 2 and 5 years, according to the protocol. Two biopsies were taken each from the antrum, corpus and distal oesophagus, according to a location diagram provided to investigators. The GI junction was defined as the proximal end of the gastric folds. Biopsies from the distal oesophagus were to be taken within the 2 cm zone above the cardia. In the presence of endoscopic lesions, biopsies were taken from the adjacent oesophageal mucosa. In addition to the six mandatory biopsies, another two quadrant biopsies from the Z-line were recommended. In the case of endoscopically suspected Barrett s oesophagus, selective biopsies (at least three) were taken from the suspected Barrett s mucosa. All biopsy specimens were sent immediately for histological assessment (including Helicobacter pylori status) at the Aliment Pharmacol Ther 2012; 35:

3 P. Malfertheiner et al. coordinating centre for histology, where two expert pathologists, MS and MV, performed all assessments. Biopsies were histologically graded according to the updated Sydney system (Dixon et al., 1996) and histological changes of the squamous epithelium in the distal oesophagus were graded according to Ismael-Begi and Pope (1974). H. pylori infection was assessed at biopsy using the Warthin-Starry silver stain method for detection of H. pylori. The Intention-to-treat (ITT) population consisted of 6215 patients with either ERD or NERD who had been recruited, in approximately equal numbers, largely from primary care clinics (90%) in Germany, Austria and Switzerland. All patients aged 18 or older with predominant heartburn (with or without erosive oesophagitis) who attended the study sites were enrolled in the study, provided they met the main inclusion criteria. They were not to have been treated continuously with any acid suppressant drug for more than 7 days during the preceding 4 weeks. The results of the initial treatment phase, the epidemiological features of the patients included and the results after a 2-year follow-up have been reported 21, 22 elsewhere. At study entry, patients underwent a physical examination and H. pylori assessment and were asked to complete standardised questionnaires assessing demographic, medical and social characteristics. Upper gastrointestinal endoscopy with biopsy was performed in every patient and they were classified into NERD and ERD groups, the latter according to the Los Angeles classification system. 23 Similar numbers of patients from each category were included from each centre (in blocks of 4; 2 NERD and 2 ERD) to facilitate long-term comparison of the groups. Patients diagnosed with Barrett s oesophagus at endoscopy, either at baseline or after healing treatment, were allowed to enter the ProGERD study. The diagnosis was based either on an endoscopic suspicion (i.e. indications of any columnar lined epithelium in the oesophagus) or on combined endoscopic and histological proof of Barrett s mucosa (i.e. detection of intestinal metaplasia within the columnar epithelium). Patients were followed up annually for a routine check including questions about medication taken for their GERD during the previous 3 months, upper GI symptoms (using the Reflux Disease Questionnaire) and completion of quality of life questionnaires (QoLRAD and SF36), whereas endoscopy and biopsy were performed only after 2 and 5 years. At the 5-year follow-up, a total of 44% of the original ITT population (i.e ) underwent endoscopy. Also, patients were asked whether they had taken medication for GERD during the last 3 months. They were subdivided into 4 groups, according to their endoscopic findings in the oesophagus at baseline i.e. NERD, ERD LA A B, ERD LA C D and Barrett s oesophagus. Disease progression was defined as clinical worsening of the preceding GERD categories and or development of Barrett s oesophagus or oesophageal adenocarcinoma. For study of progression, there was sub-classification of NERD patients at baseline into those who remained unchanged in GERD classification after 5 years, those who had changed to LA grade A B and those who had changed to LA grade C D or Barrett. A similar sub-classification was performed on LA grade A B patients, into those who remained unchanged, those who had regressed to NERD and those who had progressed to LA grade C D or Barrett; and on LA C D patients, into those who remained unchanged, those who had regressed to LA A B or NERD, or those who had progressed to Barrett. The 240 patients who had endoscopic or confirmed (endoscopic + histological) evidence of Barrett s oesophagus at baseline have been excluded from the analysis on progression. Statistical analyses Demographic and clinical characteristics are described according to baseline GERD category (non-erosive, LA grade A B, LA C D or Barrett). The association of various factors with progression to LA A BorLAC D and progression to Barrett was tested in multivariate logistic regression analyses and included the following factors: baseline age, gender, body mass index (BMI), smoking status (smoker or ex-smoker nonsmoker), intake of alcohol(yes no), history of GERD in the family (yes no), duration of GERD ( 5 years <5 years), H. pylori infection (yes no) and also change in GERD symptom score from baseline to year 5, predominant GERD medication (defined as 3 of 5 years) during follow-up (none, regular PPI, on demand PPI or other) and (for the Barrett progression analysis) GERD category at baseline (NERD, LA A B, LA C D). Odds ratios and 95% confidence intervals are reported. Statistical significance was defined as a P < All analyses were performed with SPSS 15.0 (SPSS Inc. Chicago, IL, USA.) RESULTS Patients A total of 6215 patients were enrolled in the original study of whom 2721 attended follow-up at 5 years and 156 Aliment Pharmacol Ther 2012; 35:

4 Gastro-oesophageal reflux disease progression had upper GI endoscopy. The baseline characteristics for the 2721 patients are summarised in Table 1, by GERD category at entry (Appendix S1). They compared well with the characteristics of the dropouts, suggesting that the population sample that completed the study remained consistent. The mean ages for each category were similar, but the proportions of males and of smokers increased with increasing severity of GERD, as did the duration of the disease. Familial history of GERD was slightly more common in those with more severe GERD, and generally, they had a lower prevalence of H. pylori infection. Of the patients, 240 were classified as having either endoscopic (n = 115) or confirmed Barrett s oesophagus (n = 125) at baseline. Their baseline characteristics were most similar to the patient group with LA grade C D oesophagitis, in whom Barrett s oesophagus was most common (Barrett s was present in 2.5% of NERD, 10.8% of LA grade A B and 28.1% of LA grade C D patients). These patients were not included in the analysis of GERD progression. Initial treatment phase Healing rates following initial treatment with esomeprazole are presented in Table 1. Symptomatic healing was reported for 88% of NERD patients, while ERD healing declined from 92% in milder oesophagitis LA grade A B to 84% in LA grade C D patients and to 74% in patients with Barrett s oesophagus at baseline. Progression of GERD severity Progression, regression or stability of GERD categories between grades at both 2 years and 5 years are summarised in Table 2a. Most patients remained stable or showed improvement in their grade of oesophagitis. However, some patients with milder oesophagitis did progress to more severe grades C D and also to Barrett s oesophagus. A multivariate analysis of baseline factors that may be associated with GERD progression to LA grade A BorC D after 5 years (Table 2b) indicated that a family history of GERD was associated with progression, and that remaining unhealed after baseline treatment predisposed patients to progression. Regular intake of PPI reduced the likelihood of progression compared with on-demand PPI or other therapy, although the severity of symptoms at baseline did not seem to be a predictor of progression (Table 2b). Symptoms When reflux symptoms (recorded in the RDQ) were compared at each time point between patients who progressed in GERD severity and those who did not (Figure 1a), there were no significant differences (with the exception of year 1, where symptoms were slightly worse in the group who progressed in severity, P = 0.018). Similarly, no significant differences in epigastric pain levels were noted between patients who progressed in GERD severity and those who did not (Figure 1b). Progression of Barrett s oesophagus At 5 years, 5.9% of the NERD patients, 12.1% of the LA grade A B patients and 19.7% of the LA grade C D patients in whom no Barrett s oesophagus was recorded at baseline, had endoscopic or confirmed Barrett s Table 1 Population characteristics at baseline GERD category NERD LA A B LA C D Barrett (endoscopic confirmed) N Age (mean, s.d.) Gender (male) 45% 57% 72% 69% BMI (mean s.d.) Smoking ex yes 49% 56% 62% 64% GERD 5 years 28% 33% 40% 40% GERD in family 25% 26% 30% 31% GERD sym score (mean s.d.) H. pylori-positive 27% 24% 18% 19% Healing* 88% 92% 84% 74% BMI, body mass index; GERD, gastro-oesophageal reflux disease; NERD, non-erosive reflux disease. *After initial treatment phase: for NERD = symptomatic healing, for ERD = symptomatic + endoscopic healing. Aliment Pharmacol Ther 2012; 35:

5 P. Malfertheiner et al. Table 2 (a) Proportional movement between different grades of oesophagitis or NERD from baseline to 2 years and 2 years to 5 years in patients with data for all three time points. (b) Baseline factors associated with progression to LA A BorLAC D at year 5 (a) Baseline Year 2 Year 5 NERD N = 1041 NERD 73.2% (N = 762) NERD 83.5% (N = 636) LA A B 16.1% (N = 123) LA C D0.4%(N =3) LA A B 25.8% (N = 269) NERD 53.5% (N = 144) LA A B 45.0% (N = 121) LA C D 1.5% (N =4) LA C D 1.0% (N = 10) NERD 40.0% (N =4) LA A B 40.0% (N =4) LA C D 20.0% (N =2) LA A B N = 918 NERD 59.5% (N = 546) NERD 74.0% (N = 404) LA A B 26.0% (N = 142) LA C D0%(N =0) LA A B 38.3% (N = 352) NERD 47.7% (N = 168) LA A B 50.9% (N = 179) LA C D1.4%(N =5) LA C D2.2%(N = 20) NERD 30.0% (N =6) LA A B 50.0% (N = 10) LA C D 20.0% (N =4) LA C D N = 188 NERD 46.3% (N = 87) NERD 66.7% (N = 58) LA A B 27.6% (N =24) LA C D5.7%(N =5) LA A B 42.6% (N = 80) NERD 37.5% (N = 30) LA A B 56.2% (N = 45) LA C D6.3%(N =5) LA C D 11.1% (N = 21) NERD 28.6% (N =6) LA A B 42.9% (N =9) LA C D 28.6% (N =6) NERD, non-erosive reflux disease. (b) Odds ratio 95% CI P-value Age Gender (female) BMI baseline < Smoking baseline Never 1.00 Ex Aliment Pharmacol Ther 2012; 35:

6 Gastro-oesophageal reflux disease progression Table 2 (Continued) (b) Odds ratio 95% CI P-value Yes Alcohol intake baseline (yes) GERD in family (yes) Duration GERD ( 5 years) GERD symptom score baseline Healing (no)* H. pylori baseline (yes) GERD medication** None 1.00 PPI need for regular PPI on demand Others GERD, gastro-oesophageal reflux disease; NERD, non-erosive reflux disease; PPI, proton pump inhibitors. Values in bold denote statistically significant values. * After initial treatment phase: for NERD = symptomatic healing, for ERD = symptomatic + endoscopic healing. ** Predominant (at least 3 of 5 years) GERD medication during follow-up Reflux symptoms (a) No progression Progression Baseline Year 1 Year 2 Year 3 Year 4 Year 5 (b) Epigastric symptoms No progression Progression Baseline Year 1 Year 2 Year 3 Year 4 Year 5 Figure 1 (a) Mean reflux symptom scores throughout the study in patients with and without progression of GERD severity. Difference in year 1 significant (P = 0.018), all other years no significant differences between progression no progression. (b) Mean epigastric pain scores throughout the study in patients with and without progression of GERD severity. No significant differences between progression no progression. Aliment Pharmacol Ther 2012; 35:

7 P. Malfertheiner et al. oesophagus (Table 3a). Of the 241 patients who progressed to endoscopic or confirmed BE at 5 years, data on length of the BE segment were available for 186 cases, and in 73% of them, the length was 2 cm (79% for NERD, 73% for LA grade A B and 66% for LA grade C D). Patients were assigned the diagnosis of confirmed Barrett s oesophagus only if intestinal metaplasia was present in biopsies. In some patients with endoscopic signs, and particularly in short segment cases, only gastric metaplasia was reported. These patients are included under the endoscopic category. Gastric metaplasia was present in 50% of patients with BE length 1 cm and in 29% with BE length >1 <3 cm. In patients with a BE length of 3 cm or more, no gastric metaplasia was observed. Proportions of gastric and intestinal metaplasia recorded for the different lengths of columnar epithelium are presented in Figure 2. The total proportion of patients who progressed from NERD, LA grade A Bor LA grade C D to endoscopic or confirmed Barrett s oesophagus at 5 years was 9.7% (n = 241). The factors selected for assessment of their influence on progression to Barrett s oesophagus are summarised in Table 3b. A multivariate analysis revealed that the presence of oesophagitis at baseline was the most significant factor associated with progression to Barrett s oesophagus. Other significant factors were the intake of alcohol or PPI, especially regular intake. DISCUSSION The findings of this study at 5 years confirm and extend our previous findings at 2 years 22 that, among the different grades of GERD, changes are observed in both directions, with NERD showing progressing to mild ERD (LA A B) and with ERD (LA A B) regressing healing to NERD under routine care. The observed magnitude of progression from NERD to mild moderate degrees of ERD is around 25%, but at the same time, the observed regression of ERD LA A B to NERD is much higher at 63%. However, it is important to note that the 5-year result only gives a snapshot of what happens over 5 years. When results from years 2 and 5 are taken together (Table 2), progression from NERD to mild oesophagitis occurs in more than 40%. It may be that treatment was adjusted by the physician at 2 years in cases where oesophagitis was observed; therefore, fewer patients presented with oesophagitis at 5 years. Whether a patient presents with NERD or ERD LA A B most likely depends on the coincidental or recent consumption of acid suppressant medications. The presence of symptoms is not dependent on whether patients are found with erosive lesions at the time of endoscopy; some NERD patients experience symptoms while some with ERD do not, as we have reported previously from this study. 24 A similar experience was noted recently in a population-based GERD study, where a third of patients with mild moderate oesophageal erosions did not experience symptoms, while a larger group of persons experienced reflux symptoms in the absence of erosions. 25 Data obtained from self-reporting by the patients indicated that 72% of the patients in our cohort took GERD medication in year 5 and 64% took PPI (i.e. 89% of those taking GERD medication). Similar data were reported earlier in the study. 26 There is no comparable experience to date on a similar large and well-defined prospective study of GERD, with initial assessment and healing of lesions and symptoms. The experience from most small observational studies conducted by other investigators, however, is comparable to ours, i.e. the magnitude of progression is quite small, 17, apart from that observed in the study by Pace et al. 18 In that study, all but one patient had developed endoscopic signs of oesophagitis after 5 years, although without complication. In a five-year follow-up study from Japan, 36% of patients with reflux symptoms at entry developed erosive oesophagitis, 30 compared with about 25% in our study. This difference may well be related to different access to acid suppressant therapy in routine medical care, to differences in inclusion criteria or to chance. Although overall changes from baseline were remarkably similar at 2 and 5 years in our study, it is clear that patients move across the groups in both directions. It also needs to be borne in mind that the patient group followed up for up to 5 years is reduced by about 30% compared with the two-year cohort, and by about 56% compared with baseline. Probably, the clinically most important result of our study is that the observed progression of the disease does not appear to increase linearly and the risk of progression from NERD or LA grades A B to LA grade C Dis quite small, although considerable progression was observed from NERD to LA grade A B. Given that LA grade A B was a significant prognostic factor for the development of Barrett s oesophagus, such progression of NERD patients cannot be ignored. However, Barrett s oesophagus in our population did not progress to highgrade dysplasia or cancer in any of the cases. The 6 cases of Barrett s oesophagus cancer, which were detected at the 2-year follow-up, 23 may indicate rather a missed diagnosis of malignancy at study entry rather than true 160 Aliment Pharmacol Ther 2012; 35:

8 Gastro-oesophageal reflux disease progression Table 3 (a) Progression to Barrett s oesophagus (endoscopic confirmed). (b) Factors associated with progression to Barrett s oesophagus (endoscopic confirmed) at year 5 (a) Baseline Year 5 Endoscopic or confirmed Year 5 Endoscopic Year 5 Confirmed NERD N = 1224 LA A B N = 1044 LA C D N = % (N = 72) 1.7% (N = 21) 4.2% (N = 51) 12.1% (N = 127) 4.0% (N = 42) 8.1% (N = 85) 19.7% (N = 42) 9.4% (N = 20) 10.3% (N = 22) Total N =241 N =83 N = 158 (b) Odds ratio 95% CI P-value Age Gender (female) BMI baseline < Smoking baseline Never 1.00 Ex Yes Alcohol intake baseline (yes) GERD in family (yes) H. pylori baseline (yes) Duration GERD ( 5 years) GERD symptom score baseline GERD category baseline NERD 1.00 LA A B <0.001 LA C D <0.001 GERD medication* None 1.00 PPI regular PPI on demand Other GERD, gastro-oesophageal reflux disease; PPI, proton pump inhibitors. Values in bold denote statistically significant values. * Predominant (at least 3 of 5 years) GERD medication during follow-up. progression, as there was no further incidence of malignancy between 2 and 5 years. This observation does not preclude the recognised risk for oesophageal adenocarcinoma development in Barrett s oesophagus patients 6, 20, 32 and so will not impact on our follow-up of patients with Barrett s oesophagus. However, our data Aliment Pharmacol Ther 2012; 35:

9 P. Malfertheiner et al. % Gastric Metaplasia 1cm Intestinal Metaplasia >1cm <3cm 3cm Figure 2 Gastric and intestinal metaplasia (%) found in confirmed Barrett s columnar epithelium. are reassuring as to the progression of NERD and ERD to Barrett s oesophagus, as it occurs in less than 10% of cases during a 5-year follow-up. The progression to Barrett s oesophagus is lowest in patients with NERD, intermediate in LA A B and highest in patients with LA C D. On the other hand, a significant number of patients within all categories of GERD with Barrett s oesophagus at baseline also reversed to normal under routine clinical care. A similar observation of Barrett s oesophagus regression has previously been reported in a study from Germany, in which only 70% of patients with Barrett s oesophagus diagnosis based on the classical confirmed histological definition of specialised intestinal columnar metaplasia, maintained their Barrett s oesophagus diagnosis over time. 33 There are several uncertainties concerning the correct assessment of progression or regression of Barrett s oesophagus and they include underestimation of Barrett s oesophagus at baseline because of persistent confounding inflammation. 34 However, an overestimation of short segment columnar epithelium during endoscopy as Barrett s oesophagus is more likely to occur. Since the Montreal definition 7 included the histological reporting of gastric type epithelium as well as intestinal in the definition of Barrett s oesophagus, even prominent irregular Z-lines may have been incorporated into this group. The influence of inflammation as a confounder is much less likely to have interfered with our assessments, as all patients had been examined for ERD healing after a course of PPI therapy for 4 8 weeks. The issue of short segment Barrett s oesophagus with gastric metaplasia is critical, however, as the endoscopically assessed length of Barrett s oesophagus with gastric metaplasia comprised the majority of patients with newly developed BE. The question of clinical relevance of this short segment Barrett s oesophagus with gastric metaplasia remains unanswered. We would stress in this context that all six cases of Barrett s oesophagus adenocarcinoma that were detected within the 2-year follow-up were Barrett s oesophagus with intestinal type metaplasia, but with metaplastic tissue of varying length. Thus, should only patients with intestinal type Barrett s oesophagus require regular follow-up? For the time being, the safest strategy is probably to follow patients with Barrett s oesophagus extending more than 1 cm, regardless of the type of metaplastic tissue, until more knowledge is available on the malignant potential of gastric type metaplasia in Barrett s oesophagus. The highest frequency of Barrett s oesophagus is found in patients with severe GERD at baseline, so it is to this group that we need to pay particular attention. In a large observational cohort in US, worsening of GERD over a mean follow-up period of 7 years progression occurred in only 11% and complications (i.e. stricture) in 2%. 29 The widespread use of adequate acid suppressant therapy is likely to be the main reason for the rather small subsets of patients with progression. On the other side, progression of patients to Barrett s oesophagus is observed more frequently alongside a need for regular PPI intake, probably because this is a more severe patient group. Also in a recent study, 40 patients followed up over a 20-year period had a higher use of acid suppressants concomitantly with progression to Barrett s oesophagus, and progression was also related to severity of GERD, which in itself would most probably increase PPI use. 35 Whether this points to an inadequate dosing of PPI remains speculative. When it comes to risk factors for Barrett s oesophagus development, as well as progression to more severe forms of ERD, these are male gender and alcohol, but seemingly not the presence of H. pylori at baseline. The implication of the 56% drop out of patients on the robustness of the patterns in the results and for the study conclusions is probably not of relevance, as the baseline characteristics of the group who dropped out were similar to those of the total study population. In conclusion, the 5-year follow-up of patients with NERD and ERD revealed several important aspects: (i) With regard to LA grading, there was movement in both directions across all grades of the disease. For NERD LA-A and B, a cycling between these categories was common, possibly influenced by the use or non-use of PPIs. (ii) Under routine medical care, progression to severe forms of GERD is uncommon, indicating that current therapeutic management is usually adequate. Patients 162 Aliment Pharmacol Ther 2012; 35:

10 Gastro-oesophageal reflux disease progression who remained unhealed after initial treatment were predisposed to GERD progression. (iii) Patients with a more severe form of ERD (LA C D) at initial endoscopy have the highest risk for progression to Barrett s oesophagus and they probably need more effective therapeutic management. (iv) Less than 10% of patients with GERD are likely to progress to a diagnosis of Barrett s oesophagus at 5 years, a significant number with a histological diagnosis of gastric metaplasia. Although of uncertain clinical relevance, it is our belief that gastric metaplasia probably does not pose a significant risk to the patient for cancer development. ACKNOWLEDGEMENTS We thank Dr Madeline Frame for assistance with the manuscript preparation, sponsored by AstraZeneca R&D, Mölndal, Sweden. Declaration of personal interests: Peter Malfertheiner accepts full responsibility for the conduct of the study and had access to the data. Tore Lind and K. Richter are employees of AstraZeneca. Declaration of funding interests: The study was funded by AstraZeneca, Wedel, Germany, but run independently by the steering committee, with data analyses performed at an academic institution. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article: Appendix S1. Baseline characteristics of patients participating in year 5 endoscopy and those lost to followup. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by authors. Any queries (other than missing material) should be directed to the corresponding author for the article. REFERENCES 1. Dent J, el-serag HB, Wallander MA, Johansson S. Epidemiology of gastroesophageal reflux disease: a systematic review. Gut 2005; 54: El-Serag HB. Time trends of gastresophageal reflux disease: a systematic review. Clin Gastroenterol Hepatol 2007; 5: Wong BC, Kinoshita Y. Systematic review on epidemiology of gastresophageal reflux disease in Asia. Clin Gastroenterol Hepatol 2006; 4: Wiklund I. Review of the quality of life and burden of illness in gastresophageal reflux disease. Dig Dis 2004; 22: Kulig M, Leodolter A, Vieth M, et al. Quality of life in relation to symptoms in patients with gastro-esophageal reflux disease an analysis based on the Pro- GERD initiative. Aliment Pharmacol Ther 2003; 18: Shaheen NJ, Crosby MA, Bozymski EM, Sandler RS. Is there publication bias in the reporting of cancer risk in Barrett s esophagus? Gastroenterology 2000; 119: Vakil N, van Zanten SV, Kahrilas P, Dent J, Jones R. The Montreal definition and classification of gastresophageal reflux disease: a global evidence-based consensus. Am J Gastroenterol 2006; 101: Tytgat GN, McColl K, Tack J, et al. New algorithm for the treatment of gastroesophageal reflux disease. Aliment Pharmacol Ther 2008; 27: Fass R. Proton pump inhibitor failurewhat are the therapeutic options? Am J Gastroenterol 2009; 104(Suppl. 2): S Carlsson R, Galmiche JP, Dent J, Lundell L, Frison L. Prognostic factors influencing relapse of esophagitis during maintenance therapy with antisecretory drugs: a meta-analysis of long-term omeprazole trials. Aliment Pharmacol Ther 1997; 11: Freston JW, Triadafilopoulos G. Review article: approaches to the long-term management of adults with GERD-proton pump inhibitor therapy, laparoscopic fundoplication or endoscopic therapy? Aliment Pharmacol Ther 2004; 19(Suppl. 1): Lauritsen K, Deviere J, Bigard MA, et al. Esomeprazole 20 mg and lansoprazole 15 mg in maintaining healed reflux esophagitis: Metropole study results. Aliment Pharmacol Ther 2003; 17: Lundell L, Attwood S, Ell C, et al. Comparing laparoscopic antireflux surgery with esomeprazole in the management of patients with chronic gastro-esophageal reflux disease: a 3-year interim analysis of the LOTUS trial. Gut 2008; 57: Lundell L. Surgery of gastresophageal reflux disease: a competitive or complementary procedure? Dig Dis 2004; 22: Wang K, Sampliner RE. Updated guidelines 2008 for the diagnosis, surveillance, and therapy of Barrett s esophagus. Am J Gastroenterol 2008; 103: Koop H, Schepp W, Muller-Lissner S, et al. Consensus conference of the DGVS on gastresophageal reflux. Z Gastroenterol 2005; 43: Pace F, Bianchi PG. Gastresophageal reflux disease: a typical spectrum disease (a new conceptual framework is not needed). Am J Gastroenterol 2004; 99: Pace F, Bollani S, Molteni P, Bianchi PG. Natural history of gastro-esophageal reflux disease without esophagitis (NERD) a reappraisal 10 years on. Dig Liver Dis 2004; 36: Fass R. Distinct phenotypic presentations of gastresophageal reflux disease: a new view of the natural history. Dig Dis 2004; 22: Jankowski JA, Provenzale D, Moayyedi P. Esophageal adenocarcinoma arising from Barrett s metaplasia has regional variations in the west. Gastroenterology 2002; 122: Malfertheiner P, Lind T, Willich S, et al. Prognostic influence of Barrett s esophagus and Helicobacter pylori infection on healing of erosive gastro-esophageal reflux disease (GERD) and symptom resolution in non-erosive GERD: report from the ProGERD study. Gut 2005; 54: Labenz J, Nocon M, Lind T, et al. Prospective follow-up data from the Pro- GERD study suggest that GERD is not a categorial disease. Am J Gastroenterol 2006; 101: Aliment Pharmacol Ther 2012; 35:

11 P. Malfertheiner et al. 23. Lundell L, Dent J, Bennett JR, et al. Endoscopic assessment of esophagitis: clinical and functional correlates and further validation of the Los Angeles classification. Gut 1999; 45: Kulig M, Nocon M, Vieth M, et al. Risk factors of gastresophageal reflux disease: methodology and first epidemiological results of the ProGERD study. J Clin Epidemiol 2004; 57: Ronkainen J, Aro P, Storskrubb T, et al. High prevalence of gastresophageal reflux symptoms and esophagitis with or without symptoms in the general adult Swedish population: a Kalixanda study report. Scand J Gastroenterol 2005; 40: Nocon M, Labenz J, Jaspersen D, et al. Long-term treatment of patients with gastresophageal reflux disease in routine clinical care-results from the ProGERD study. Aliment Pharmacol Ther 2007; 25: Hartmann D, Kudis V, Jakobs R, Riemann JF, Schilling D. Course of erosive gastresophageal reflux disease (ERD) a prospective examination under therapeutic daily life conditions with a mean follow-up of 6.5 years. Z Gastroenterol 2006; 44: Manabe N, Yoshihara M, Sasaki A, Tanaka S, Haruma K, Chayama K. Clinical characteristics and natural history of patients with low-grade reflux esophagitis. J Gastroenterol Hepatol 2002; 17: Sontag SJ, Sonnenberg A, Schnell TG, Leya J, Metz A. The long-term natural history of gastresophageal reflux disease. J Clin Gastroenterol 2006; 40: Kawanishi M. Will symptomatic gastresophageal reflux disease develop into reflux esophagitis? J Gastroenterol 2006; 41: Isolauri J, Luostarinen M, Isolauri E, Reinikainen P, Viljakka M, Keyrilainen O. Natural course of gastresophageal reflux disease: year follow-up of 60 patients. Am J Gastroenterol 1997; 92: Falk GW. Barrett s esophagus. Gastroenterology 2002; 122: Meining A, Ott R, Becker I, et al. The Munich Barrett follow up study: suspicion of Barrett s esophagus based on either endoscopy or histology only- what is the clinical significance? Gut 2004; 53: Hanna S, Rastogi A, Weston AP, et al. Detection of Barrett s esophagus after endoscopic healing of erosive esophagitis. Am J Gastroenterol 2006; 101: Falkenback D, }Oberg S, Johnsson F, Johansson J. Is the course of gastresophageal reflux disease progressive? A 21-year follow-up. Scan J Gastroenterol 2009; 44: Aliment Pharmacol Ther 2012; 35: