Discontinuation of proton pump inhibitors in patients on long-term therapy: a double-blind, placebo-controlled trial

Transcription

1 Alimentary Pharmacology & Therapeutics Discontinuation of proton pump inhibitors in patients on long-term therapy: a double-blind, placebo-controlled trial E. BJÖRNSSON, H. ABRAHAMSSON, M. SIMRÉN, N. MATTSSON, C. JENSEN, P. AGERFORZ & A. KILANDER Department of Internal Medicine, Section of Gastroenterology and Hepatology, Sahlgrenska University Hospital, Gothenburg, Sweden Correspondence to: Dr. E. Björnsson, Department of Internal Medicine, Sahlgrenska University Hospital, SE Gothenburg, Sweden. einar.bjornsson@medic. gu.se Publication data Submitted 27 June 26 First decision 6 July 26 Resubmitted 6 July 26 Accepted 6 July 26 SUMMARY Background The proportion of proton pump inhibitor users on long-term therapy who can discontinue proton pump inhibitor (PPI) medication without developing symptoms is unknown. Aim To determine the proportion of patients on long-term PPI therapy who are able to discontinue PPIs without developing symptoms. Methods Patients on long-term PPIs, without a history of peptic ulcer or esophagitis underwent upper endoscopy. Patients were randomized doubleblindly to taper down or continue a constant dosage of omeprazole for three weeks. Thereafter, all patients discontinued PPIs. Results Of the 97 patients enrolled, had used PPIs for 48 months, 78% had GERD. A total of 27% did not use PPIs during the year after discontinuation, 31% of the patients randomized to tapering discontinued PPIs and 22% of those who did not could discontinue therapy (NS). Gastrooesophageal reflux disease (GERD) patients were more prone to continue PPIs than non-gerd patients. Only 16 (21%) of GERD patients were off PPIs vs. 48% of patients without GERD (p <.5). Serum gastrin was higher at baseline in GERD patients who resumed PPIs versus non-resumers (p <.5). GERD and serum gastrin were independent predictors of PPI requirement. Conclusions Discontinuation of PPI was successful in 27% of long-term PPI users. GERD patients had more difficulty discontinuing PPIs than non-gerd patients. Aliment Pharmacol Ther 24, ª 26 The Authors 945 doi:1.1111/j x

2 946 E. BJÖRNSSON et al. INTRODUCTION Proton pump inhibitors (PPIs) are a major economic burden for the healthcare system in many countries. Concerns have been raised about the increasing costs associated with prescription of these drugs as they are often prescribed for minor symptoms and without clear indications. 1 7 Studies from the US, Australia and Europe have demonstrated overuse of PPIs in hospitalized patients 3 5 and in primary care. 2, 7 Very limited data exist on the proportion of PPI users on long-term therapy who could discontinue PPIs without developing symptoms. In a recent uncontrolled study, 15% of gastro-oesophageal reflux disease (GERD) patients on long-term PPI therapy remained asymptomatic without medication after step-down management of PPIs. 8 The proportion of unselected long-term users of PPIs who could discontinue medication is unexplored. Acid rebound hypersecretion following the cessation of PPIs has been demonstrated The clinical importance of this acid rebound following the treatment with acid-suppressive therapy is unclear, but it has been suggested that this may make it more difficult to discontinue PPIs The clinical importance of acid rebound is supported by the fact that discontinuation of ranitidine in healthy subjects was associated with dyspeptic symptoms. 12 No controlled study has previously been undertaken to examine the effects of discontinuation of PPI treatment. The hypothesis that tapering of PPIs makes it easier to discontinue PPIs has not been tested previously. The primary aim of the study was to examine what proportion of long-term PPI users is able to discontinue the medication without developing symptoms and to study the factors that might influence the successful discontinuation. The secondary aim was to study the effect of tapering down the PPIs prior to discontinuation. MATERIAL AND METHODS Patients buying PPIs on prescription at several pharmacies in Gothenburg Sweden were asked to complete a questionnaire on the reason for using PPIs, their symptomatology and the duration of PPI therapy. The patients approached were consecutive patients and there was no selection of patients. Patients with more than 8 weeks of regular daily use of PPIs without a history of peptic ulcer or oesophagitis, and with no significant co-morbidity, were invited to participate. Thus, all patients who did not have exclusion criteria were asked to participate. Patients did not need to be asymptomatic in order to participate but as the aim of the study was to investigate the possibility of discontinuation of PPIs, many patients with troublesome GERD did not choose to participate. Figure 1 shows the patients included in the study and the proportion of patients who chose not to participate. The patients willing to participate underwent endoscopy after an overnight fast, while they were still on PPIs, and the next day they started the study medication. Biopsies for culture of Helicobacter pylori and histology were obtained from the antral and the body Questionnaire at the Pharmacy 593 patients Not suitable patients (37) (use < 8 weeks and/or previous ulcer/oesophagitis) Suitable patients (286) (PPI > 8 weeks & no history of ulcer or oesophagitis) Interested to participate (116) Excluded (19) (Oesophagitis/polyps) Randomized (97) (Normal gastroscopy) Followed for 1 year (96) Figure 1. Patients included in the study.

3 DISCONTINUATION OF PROTON PUMP INHIBITORS 947 of the stomach. Routine blood tests and serum gastrin were determined. Study protocol A detailed history of GI symptoms was obtained before inclusion, including the occurrence of heartburn, acid regurgitation and/or chest pain rising from the upper abdomen, dyspepsia and whether the most troublesome symptom was upper abdominal pain (ulcer-like dyspepsia), discomfort (dysmotility-like dyspepsia) or both (non-specific dyspepsia) and symptoms of irritable bowel syndrome. Patients with a normal upper endoscopy were randomized double-blindly to either omeprazole 2 mg o.i.d. for 3 weeks (group 1) or 2 mg daily for 1 week, 1-mg omeprazole for 1 week and 1-mg omeprazole every other day for 1 week (group 2), in two identical capsules (in both groups) containing 1 mg of Losec MUPS (AstraZeneca Mölndal, Sweden) with a different dosage in group two. Study medication and randomization were elaborated by the Swedish Pharmacy Production Agency. The patients were instructed to take the study medication every morning at breakfast. All patients also received 3 antacid chewing tablets (Novalucid, AstraZeneca, Sweden) to take on demand. The patients were told that they could cease participation in the study at any time and resume their PPIs but were asked to contact the study nurse by telephone to notify their decision. We told the patients that this was of major importance and the patients could call the study nurse at any time. They were informed about the possibility of acid rebound. After the 3 weeks of treatment all patients discontinued the study medication. Our assumption was that approximately 2% of patients randomized to non-tapering could discontinue therapy whereas approximately 5% of patients tapering their PPIs could discontinue PPI treatment. Thus, based on these assumptions, in order to find significant differences with 8% power between the groups our power calculation demonstrated that we should at least need a total of 78 patients (39 in each group). The lack of data of discontinuation of PPI users in general limits these kind of assumptions and we chose therefore to include approximately 5 patients in each group. One week after discontinuing the study medication, the participants were contacted by the study nurse and interviewed about symptoms. Patients willing to continue in the study who had not resumed their PPIs and who had at least moderate reflux symptoms were asked to undergo 24 h of ph recording. Patients who had abnormal ph findings (total time of 24 h with ph < 4, e 4% of the time and/or positive symptom association probability) were considered to have reached an end-point and were asked to resume PPIs. Questionnaires on quality of life and GI symptoms were completed on inclusion and in those who did not resume PPIs 3, 6 and 12 months after the study start. Patients had a follow-up visit 2 months after the study start and were asked to report changes in their healthrelated quality of life and hospitalization during the 1 year of follow-up. Gastrointestinal symptom and quality of life assessment To assess the presence and severity of gastrointestinal (GI) symptoms, the Gastrointestinal symptom rating scale (GSRS) was used. 13 The questionnaire uses a seven-grade Likert scale and includes 15 items, which are grouped into five domains: reflux, abdominal pain, constipation, indigestion, and diarrhoea. The questions concern symptoms during the week prior to the study. GSRS data are presented as a total score and domain scores. The higher the scores, the more pronounced the symptoms. The modified Glasgow dyspepsia score was used to assess symptom severity, the duration of symptoms and behavioral responses to the symptoms. 12 The Psychological General Well Being Index (PGWB) 14 was used to assess quality of life. The questions deal with symptoms during the week prior to the study. The higher the score, the greater the well-being. Outcome measures Upper endoscopy with biopsies was performed at baseline, while patients were still on PPI therapy and baseline measurements of GI symptoms (GSRS and modified Glasgow dyspepsia score) and quality of life (PGWB). One week after discontinuing the study drug, the participants were contacted and interviewed about symptoms. Patients willing who had not resumed their PPIs with at least moderate reflux symptoms were asked to undergo 24 h of ph recording, which was scheduled within 2 weeks. Follow-up of patients after the period on the study drug was continued for 1 year. The time which they reinstituted PPIs was continu-

4 948 E. BJÖRNSSON et al. ously registered. Questionnaires concerning GI symptoms and quality of life were completed after one, three, 6 and 12 months after being on the 3-week study drug. Statistical analysis The Fisher exact test was used to test differences between groups regarding dichotomous variables; the Mann Whitney test was used for continuous variables. For survival (off PPIs) analysis, Kaplan-Meier estimates were calculated when comparing groups and formally tested using the Log-Rank test. Similarly, Cox Proportional Hazard regression was used for ordered and continuous variables. For multivariate purposes, Stepwise Cox PH-regression was performed. Stepwise logistic regression was performed for multivariate purposes to predict the need for PPIs. All tests are twotailed and were conducted at a 5% significance level. The results are presented as medians and interquartile range (IQR). The study was approved by the Ethics Committee at the University of Gothenburg and all patients signed informed consent. RESULTS Patients A total of 116 patients were willing to participate and underwent upper endoscopy (Figure 1). There was no difference in the age and gender among those who were suitable and approached for participation (data not shown) but a higher proportion of those who were willing to participate had heartburn compared with those who chose not to participate (78% vs. 63%; P <.5). Nineteen were excluded due to endoscopy findings but completed the questionnaires at baseline. Sixteen patients had mild oesophagitis (Los Angeles class A). One patient was excluded due to Barrett s oesophagus and two because of large gastric polyps requiring polypectomy sessions. Ninety-seven patients were randomized to the two treatment arms, 51 to tapering and 46 to non-tapering. The characteristics of the total study group are shown in Table 1. One patient randomized to non-tapering died in a car accident 11 days after the start of the study, leaving 45 patients in the non-tapering arm. Twenty-two (23%) Table 1. The characteristics of the total study group. Results are presented as medians and interquartile range (IQR) or range patients had H. pylori on culture or histology. Twentyfive patients had hypergastrinaemia (>75 pmol/l). Serum gastrin was significantly associated with the duration of PPI therapy prior to inclusion in the study (r ¼.21, P <.5). Indications for PPIs Total Number 96 Age (IQR) 63 (51 7) Gender (F/M) 52/44 Duration of treatment (months), 48 (22 87), 4 18 range Omeprazole (%) 55 (49%) Lanzoprazole (%) 25 (26%) Esomeprazole (%) 16 (17%) Pantoprazole (%) 7 (7%) Rabeprazole (%) 1 (1%) Helicobacter pylori (%) 23 (24%) Hiatal hernia 55 (57%) GERD (%) 75 (78%) Dyspepsia 9 (9.4%) Other indications 12 (12.5%) 24-h ph recording 23 (24%) Previous upper endoscopy 59 (61%) Irritable bowel syndrome (%) 38 (4%) The indication was GERD in the majority of patients (Table 1). The proportion of GERD patients with concomitant H. pylori infection was the same as in non- GERD subjects (24%). Hiatal hernia was present in 5 (67%) GERD patients vs. 12 (57%) in non-gerd patients (P ¼.2). Dyspepsia was the indication for PPIs in nine (9.4%) patients (none had reflux symptoms). Other indications were noted for 12 (12.5%) patients (Table 2). Four patients had PPIs as ulcer prophylaxis. None of these had a history of peptic ulcer, GI bleeding or other risk factors motivating PPIs. Comparison between resumers and nonresumers Twenty-six (27%) patients were off PPIs at the end of the 1-year follow-up whereas 7 (73%) patients had resumed PPIs. Two patients who were off PPIs

5 DISCONTINUATION OF PROTON PUMP INHIBITORS 949 Table 2. Other indications than gastro-oesophageal reflux disease (GERD) or dyspepsia Reasons for taking proton pump inhibitors Withdrawal possible Ulcer prophylaxis 4: corticosteroids ¼ 2 non-steroidal anti-inflammatory drug ¼ 1, low-dose aspirin ¼ 1 Non-cardiac chest pain 3 3 Non-specific abdominal pain 2 2 Non-acid reflux to airways at night 1 Dysphagia 1 Nausea/retchning 1 2 (corticosteroids 1 and low-dose aspirin ¼ 1) had used H2 receptor antagonists occasionally on demand during the last 6 months of follow-up. The others did not use any acid-suppressive therapy, apart from antacids occasionally. One patient who had been on omeprazole 2 mg daily prior to the study patient reinstituted PPIs after just 4 days of the study medication. One patient on corticosteroids and warfarin used PPI as ulcer prophylaxis therapy, was also put on low-dose aspirin and then PPIs were reinstituted. There was no difference in age, gender, occurrence of hiatal hernia, H. pylori, irritable bowel syndrome or duration of PPI therapy prior to enrollment between those who resumed PPIs vs. those who were off PPIs (Table 3). There was a correlation between GERD as an indication for PPIs and the need to resume PPI treatment. Reinstitution of PPIs was observed in 79% of GERD patients, 67% in dyspeptics and 42% in other patients (Table 3 and Figure 2). Comparison of the time to rein- Table 3. Comparison between those who resumed proton pump inhibitors (PPI) and those who did not resume PPI therapy Resumed PPI n ¼ 7 Not resumed PPI n ¼ 26 P-value Age 64 (52 72), (45 69), 21 8 NS Gender (F/M) 38/32 14/12 NS Duration of treatment 48 (3 92), (18 6), NS GERD (%) 59/75 (79%) 16/75 (21%).4 Non-GERD (%) 11/21 (52%) 1/21 (58%) NS Dyspepsia 6/9 (67%) 3/9 (33%) NS Other indication 5/12 (42%) 7/12 (52%) NS Helicobacter pylori pos. (%) 18/7 (26%) 5/26 (19%) NS 24-h ph study abnormal/normal 16/3 /4 Previous gastroscopy 43 (61%) 16 (62%) NS Irritable bowel syndrome (%) 25 (36%) 13 (5%) NS PPI mg 21.6 (.7) 21.5 (1.1) NS Omeprazol % 36 (51%) 11 (42%) NS Hiatal hernia 41/55 (75%) 29/41 (7%) NS Glasgow dyspepsia score d 4.5 ( 8) 4.5 (1 7) NS Glasgow dyspepsia score d 28 6 ( 9) 4.5 ( 7) NS Gastrointestinal symptom rating scale (GSRS) total d 42 (29 51) 34 (25 43) NS GSRS total d (28 51) 35 (2 44).2 GSRS reflux score d 5 (3 7) 3 (2 4).2 GSRS reflux score d 28 6 (4 9) 4 (2 5).6 GSRS abdominal pain d 8 (6 11) 6 (4 9).3 GSRS abdominal pain 28 8 (6 12) 6 (3 9).1 Psychological general well-being (PGWB) total score d 11 (84 19) 95 (8 11) NS PGWB total score d (87 111) 15 (95 113) NS

6 95 E. BJÖRNSSON et al. Proportion off PPls Logrank P = Time (days) Dyspepsia GERD Others Figure 2. Comparison between patients who were off Proton pump inhibitors at the end of the 1 year follow-up with gastro-oesophageal reflux disease, dyspepsia and other indications (Others). stitution of PPIs in patients with GERD vs. those without GERD revealed a highly significant difference in the median time for reinstitution of PPIs in the two groups (Figure 3). GERD patients resumed PPIs 18 days after the 3-week study medication period (4 34; range 4 193) whereas the non-gerd patients resumed therapy 95 days after (17 114; range 22 23) (P <.1). Sixteen patients with an abnormal ph recording were told to resume PPIs. Gastrointestinal symptom rating scale total and Glasgow dyspepsia score were similar at baseline in those who resumed PPIs vs. non-resumers (Table 3). The GSRS reflux score and the GSRS abdominal pain score were higher at baseline in those who reinstituted PPIs. GI symptoms were not significantly different at baseline in GERD patients who resumed PPIs vs. those who did not (data not shown). Patients who had oesophagitis had reflux scores similar to the score for GERD patients with normal endoscopy (5.2 vs. 4.9; P ¼ NS). The only significant difference between GERD patients off PPIs after 1 year vs. those who resumed PPIs was higher serum gastrin among those who reinstituted PPIs [59 (36 81) vs. 38 (25 57); P ¼.3]. All 15 GERD patients with serum gastrin levels higher than 1 pmol/l reinstituted PPIs. The area under the curve (AUC) for the model with GERD and serum gastrin had a value of.758. Moreover, the discontinuation of PPIs was associated with a significantly better quality of life as measured by the PGWB index (Table 4). In stepwise logistic multivariate regression analysis, GERD (P ¼.14, odds ratio 8.5, CI ) and s-gastrin (P ¼.35, odds ratio 1.18, CI ) contributed to the strongest model as independent predictors for the need for reinstitution of PPIs. Possible predictors in the stepwise model were selected from the univariate analysis, all variables with P <.1 were included. Proportion off PPls GERD Non-GERD Comparison between tapering and non-tapering Sixteen (31%) of 51 patients who were randomized to tapering discontinued PPIs whereas 1 (22%) of 45 who did not taper down discontinued PPIs (NS) (Figure 4) (Table 5). The only difference between patients randomized to tapering vs. those who did not taper down was a higher proportion of hiatal hernia in those who tapered down their PPIs (Table 5). Logrank P = Time (days) Figure 3. Comparison between patients who were off proton pump inhibitors at the end of the 1 year follow-up among the gastro-oesophageal reflux disease (GERD) and the non-gerd patients. DISCUSSION During the last two decades, peptic ulcer as an indication for PPIs has decreased whereas GERD has become a major indication although dyspepsia and prophylactic therapy are also common indications. 1 7 In a recent study, approximately 6% of patients on long-term PPIs had GERD. 15 This is very similar to our original cohort of PPI users participating in our symptom

7 DISCONTINUATION OF PROTON PUMP INHIBITORS 951 Table 4. Comparison of gastrointestinal symptom rating scale (GSRS) total score, GSRS reflux score, GSRS abdominal pain score and Glasgow dyspepsia score as well as the psychological general well-being (PGWB) score during the course of the study and in those who had a successful PPI discontinuation At baseline 28 days 3 months 6 months 12 months GSRS total 34 (25 43) 35 (2 44) 23 (2 3) 28 (2 47) 28 (21 39) GSRS reflux 3 (2 4) 4 (2 5) 3 (2 4) 3 (2 4) 3.5 (2 5) GSRS Abdominal pain 6 (4 9) 6.5 (3 9) 4 (3 6) 4 (3 7)* 5 (4 8) Glasgow dyspepsia score 4.5 ( 8) 4.5 ( 7) ( 8)** ( )** ( 5) PGWB Total score 95 (8 11) 15* (95 113) 112* (15 116) 17* (93 114) 14* (91 116) * P <.5, ** P <.1 compared with baseline. Proportion off PPls Non-tapering Tapering Logrank P = Time (days) Figure 4. Comparison between patients who were off Proton pump inhibitors at the end of the 1 year follow-up among those who were randomized to tapering vs. those who were randomized to non-tapering. survey. 16 However, 78% of patients willing to participate in the current study had GERD. We observed a positive association between GERD as an indication for PPIs and the need to resume PPIs. Consequently, the proportion of patients resuming therapy seems to depend to a great extent on the indication. At the end of follow-up 27% of patients were off PPIs in the current study but this proportion would probably have been larger if more non-gerd patients had participated. Our results do not permit firm conclusions to be drawn on the proportion of long-term PPI users in general who could do well without medication. However, it seems unlikely that the patients who were willing to participate were in less need of PPIs than those who did not, as GERD was more common in the participants than in those who chose not to participate. Very limited data exist on the need for long-term acid-suppressive therapy in PPI users in general. Cooper et al. 17 carried out an audit of PPI prescribing in general practice in the UK and 12% of patients had their acid-suppressive therapy discontinued completely. However, the proportion of PPI users with a confirmed diagnosis of GERD was unclear, follow-up was short and no quality of life assessments was made. 17 In the current study, we found that 79% of GERD patients had resumed PPIs at the end of follow-up whereas 21% were off PPIs, although two patients used H2-receptor antagonists occasionally. Our results with 19% of GERD patients being asymptomatic without medication are similar to the 15% of GERD patients not needing medical therapy in a study of step-down management of PPIs. 8 However, no upper endoscopy was performed and their method of tapering down the PPIs for successful discontinuation was not evaluated. 8 In the current study 14% of the patients who had endoscopy had oesophagitis and had a very similar reflux score at baseline compared with those without oesophagitis. It would therefore appear to be reasonable to perform an upper endoscopy before a decision can be made in patients with GERD to discontinue PPIs completely as patients who are already on long-term PPIs but still have erosive disease should not have their PPIs discontinued. The GERD patients who reinstituted PPIs shortly after discontinuation seem to have a truly acid-related disorder whereas the non-gerd patients resumed PPIs after a median of 95 days. One patient resumed PPIs just 4 days after the start of the study medication, choosing the same treatment that she was on previously. Psychological factors in these patients cannot therefore be ignored. 18

8 952 E. BJÖRNSSON et al. Table 5. Comparison between those who were randomized to tapering of their proton pump inhibitors (PPI) and those who did not have tapering of their PPI therapy Tapering PPI n ¼ 51 No tapering of PPI n ¼ 45 P-value Age 64 (56 73) 6 (49 69) NS Gender (F/M) 2/31 21/24 NS Duration of treatment 6 (26 18) 36 (18 68) NS Gastro-oesophageal reflux disease (GERD) (%) 37 (73%) 38 (84%) NS Non-GERD (%) 14 (27%) 7 (16%) NS Dyspepsia 8 (16%) 1 (2%) NS Other indication 6 (12%) 6 (13%) NS Helicobacter pylori pos. (%) 13 (25%) 1 (22%) NS Previous gastroscopy 34 (67%) 25 (56%) NS Irritable bowel syndrome (%) 19 (37%) 19 (42%) NS PPI mg 22.6 (.7) 2.5 (.9) NS Omeprazol % 29 (56%) 18 (4%) NS Hiatal hernia 34 (67%) 21 (47%).3 Glasgow dyspepsia score d 3. ( 7) 5. (2 8) NS Glasgow dyspepsia score d ( 8) 5.5 ( 8) NS GSRS total d 36 (23 49) 41 (33 51) NS GSRS total d (28 47) 38 (27 53) NS GSRS reflux score d 4 (2 6) 4 (2 6) NS GSRS reflux score d 28 5 (3 8) 6 (3 9) NS GSRS abdominal pain d 8 (5 1) 8 (6 11) NS GSRS abdominal pain 28 7 (5 1) 8 (5 11) NS Psychological general well-being total score d 99 (85 11) 11 (81 19) NS PGWB total score d (94 111) 14 (86 11) NS Recent studies have shown that many GERD patients 1, take PPIs on demand. In the current study, patients taking PPIs on demand were excluded. Most patients with GERD were willing to continue with PPIs on demand to control their symptoms but 5 6% of patients continued with placebo on demand The reasons for the different results of these trials compared with our results are unclear. Placebo treatment is not equivalent to absence of therapy. The recruitment of patients was also different and prior longterm therapy was not a prerequisite for inclusion but endoscopy-negative GERD as defined by symptoms. 22 As our patients were strictly those who used PPIs every day it seems logical that it might be more difficult for these patients to discontinue PPIs completely. We observed a significant improvement in the quality of life as measured by the PGWB index at all intervals in those who were off PPIs. The improved quality of life is probably not due to the absence of PPIs but rather associated with being in a clinical study with endoscopy, lab tests, being cared for by a study nurse and evaluation by a gastroenterologist with reassurance. Those who were off PPIs also had unchanged or fewer GI symptoms. Thus, most patients who were off PPIs at the end of follow-up had been prescribed PPIs for symptoms unlikely to be due to acid production. Acid rebound hypersecretion after discontinuation of acid-suppressive therapy has been well documented with acid output measurements The clinical implication of the acid rebound hypersecretion after discontinuation of PPIs is still unclear. It is conceivable that this might lead to exacerbation of reflux symptoms and explain the clinical experience of many doctors that discontinuation of PPIs is difficult due to symptom recurrence. The short period of tapering did not seem to make it easier for patients to be off PPIs compared with patients who discontinued therapy promptly. It is possible that a more prolonged tapering would have been more effective in order to step down from a PPI to no therapy. Very limited data exists on symptoms associated with discontinuation of PPIs. A pilot study with a short-term course of omeprazole 2 mg in healthy volunteers induced significant rebound heartburn symptoms. 23 The current study is the first to test the hypothesis that tapering might be helpful in stepping down from a PPI to no therapy.

9 DISCONTINUATION OF PROTON PUMP INHIBITORS 953 Withdrawal of ranitidine induced dyspeptic symptoms that lasted for only a few days 12 and in a previous uncontrolled study tapering was performed in 2 weeks. 8 However, the results of the current study put the question of tapering into a new perspective. Gastrin at baseline was significantly higher in GERD patients who resumed PPIs vs. GERD patients who were off PPIs and was independently associated with PPI requirement. Long-term hypergastrinaemia is associated with increased gastric acid secretion. 24 Furthermore, Gillen et al. 1 demonstrated that the degree of increase in maximal acid output after discontinuation of omeprazole was related to fasting gastrin levels. It thus seems likely that GERD patients with high fasting gastrin might be more prone to develop symptomatic acid rebound and would therefore experience more difficulty in discontinuing PPIs. It is conceivable that only patients with high gastrin levels would need tapering in order to decrease the risk of a clinically significant acid rebound. We found gastrin levels above the upper limit of normal in a minority of our 1, patients, which is in line with other studies. In conclusion, discontinuation of PPI was successful in 27% of long-term PPI users. GERD patients had more difficulty discontinuing PPI than non-gerd patients. Hypergastrinaemia seems to be an important predictor of PPI requirement in GERD patients. Discontinuation of these drugs might be possible in some PPI users without a strong indication for the drug, especially if they do not have symptoms of GERD, although recommendations on the mode of discontinuation can not be made at the present time. ACKNOWLEDGEMENTS The authors would like to express their sincere gratitude for generous support from the Federation of County Councils in Sweden (Landstingsförbundet) and Professor Sten Iwarson (The Gothenburg Committee for Drugs and Pharmaceuticals). This study was also supported by the Faculty of Medicine, Göteborg University. No support from any pharmaceutical company was received for this study. REFERENCES 1 Raghunath AS, O Morain C, McLoughlin RC. Review article: the long-term use of proton-pump inhibitors. Aliment Pharmacol Ther 25; 22 (Suppl 1): Bashford JNR, Norwood J, Chapman SR. Why are patients prescribed proton pump inhibitors? Retrospective analysis of link between morbidity and prescribing in the General Practice Research Database. BMJ 1998; 317: Nardino RJ, Vender RJ, Herbert PN. Overuse of acid suppressive therapy in hospitalised patients. Am J Gastroenterol 2; 95: Pillans PI, Kubler PA, Radford JM, et al. Concordance between use pf proton pump inhibitors and prescribing guidelines. Med J Aust 2; 172: Strid H, Simrén M, Björnsson ES. Overuse of acid suppressant drugs in patients with chronic renal failure. Nephrol Dial Transplant 23; 18: Niklasson A, Bajor A, Strid H, et al. Overuse of acid suppressant drugs in patients with pulmonary diseases. Respir Med 23; 97: Ryder SD, O Reilly S, Miller RJ, et al. Long term acid suppressing treatment in general practice. BMJ 1994; 38: Inadomi JM, Jamal R, Murata GH, et al. Step-down management of gastroesophageal reflux disease. Gastroenterology 21; 121: Waldum HL, Arnestad JS, Brenna E, et al. Marked increase in gastric acid secretory capacity after omeprazole treatment. Gut 1996; 39: Gillen D, Wirz AA, Ardill JE, et al. Rebound hypersecretion after omeprazole and its relation to on-treatment acid suppression and Helicobacter pylori status. Gastroenterology 1999; 116: Gillen D, Wirz AA, Neithercut WD, et al. Helicobacter pylori infection potentiates the inhibition of gastric acid secretion by omeprazole. Gut 1999; 44: Smith AD, Gillen D, Cochran KM, et al. Dyspepsia on withdrawal of ranitidine in previously asymptomatic volunteers. Am J Gastroenterol 1999; 94: Revicki DA, Wood M, Wiklund I, et al. Reliability and validity of the Gastrointestinal Symptom Rating Scale in patients with gastroesophageal reflux disease. Qual Life Res. 1998; 7: Dimenäs E, Glise H, Hallerbäck B, et al. Quality of life in patients with upper gastrointestinal symptoms. An improved evaluation of treatment regimens? Scand J Gastroenterol 1993; 28: Jacobson BC, Ferris TG, Shea TL, et al. Who is using chronic acid suppression therapy and why? Am J Gastroenterol. 23; 98: Moss UD, Carlsten A, Björnsson E. High gastrointestinal symptom burden in patients using proton pump inhibitors. Gastroenterology 24; 126: Suppl. (abstract T 1144). 17 Cooper AL, Langworthy H, Porter S. Cost-effective prescribing of proton pump inhibitor therapy: an audit in general practice. Int J Clin Pract 2; 54: Johnston Bt, Gunning J, Lewis SA. Health care seeking by heartburn sufferers is associated with psychosocial factors. Am J Gastroenterol 1996; 12: Lee TJ, Fennerty MB, Howden CW. Systematic review: is there excessive use of proton pump inhibitors in gastro-oesophageal reflux disease? Aliment Pharmacol Ther 24; 2:

10 954 E. BJÖRNSSON et al. 2 Bardhan KD, Müller-Lissner S, Bigard MA, et al. Symptomatic gastro-oesophageal reflux disease: double blind controlled study of intermittent treatment with omeprazole or ranitidine. The European Study Group. BMJ 1999; 318: Lind T, Havelund T, Lundell L, et al. On demand therapy with omeprazole for the long-term management of patients with heartburn without oesophagitis a placebo-controlled randomized trial. Aliment Pharmacol Ther 1999; 13: Talley NJ, Lauritsen K, Tunturi-Hihnala H, et al. Esomeprazole 2 mg maintains symptom control in endoscopy-negative gastro-oesophageal reflux disease: a controlled trial of on-demand therapy for 6 months. Aliment Pharmacol Ther 21; 15: Lindberg G, Vieri J. Dyspepsia after withdrawal of omeprazole in healthy volunteers. Gastroenterology 23; 124: Suppl (abstract S 158). 24 Pisegna JR, Norton JA, Slimak GG, et al. Effects of curative gastrinoma resection on gastric secretory function and antisecretory drug requirement in the Zollinger Ellison syndrome. Gastroenterology 1992; 12: Sanduleanu S, Stridsberg M, Jonkers D, et al. Serum gastrin and chromogranin A during medium- and long-term acid suppressive therapy: a case control study. Aliment Pharmacol Ther 1999; 13: Reilly JP. Safety profiles of the proton pump inhibitors. Am J Health Syst Pharm 1999; 56: S11 7.