Jansen Dr Department of Gastroenterology, Academic Hospital Nijmegen St. Radboud, PO Box 9101, 6500 HB Nijmegen,the Netherlands. E-mail: firstname.lastname@example.org
In the treatment of reflux oesophagitis, H2-receptor antagonists are still widely used in spite of the apparent higher efficacy of proton pump inhibitors. In an attempt to compensate for the lower efficacy, H2-receptor antagonists are now increasingly being used at a higher dose.
To assess whether or not standard-dose lansoprazole (30 mg o.d.) is more effective than high-dose ranitidine (300 mg b.d.) in moderately severe reflux oesophagitis (grades II–III).
Lansoprazole or ranitidine was given to 133 patients for 4–8 weeks in a double-blind, randomized, parallel group, multicentre trial.
The percentage of patients with endoscopically-verified healing was significantly higher on lansoprazole than on ranitidine both after 4 weeks (79% vs. 42%) and 8 weeks (91% vs. 66%), though smoking had a negative impact on oesophagitis healing with lansoprazole. Heartburn, retrosternal pain and belching improved significantly better with lansoprazole than with ranitidine, as did the patient-rated overall symptom severity. Relief of heartburn appeared somewhat faster with ranitidine, but was more pronounced with lansoprazole. The number of patients with adverse events was similar in both treatment groups.
Standard-dose lansoprazole is better than high-dose ranitidine in moderately severe reflux oesophagitis.
Gastro-oesophageal reflux is a major cause of dyspepsia in the Western world. The amount and concentration of gastric acid in refluxed material is the principal factor in the development and perpetuation of symptoms and oesophagitis. Many studies have shown that inhibition of both volume and concentration of gastric acid by H2-receptor antagonists results in a significant improvement of reflux symptoms and in healing of the oesophagitis.2 Ranitidine is the most extensively studied H2-receptor antagonist and the majority of studies have been performed at a dose of 150 mg b.d. With the development of proton pump inhibitors in recent years, a new class of highly effective gastric acid inhibitory drugs has become available. Comparative studies have shown that these drugs are superior to ranitidine at its standard dose of 150 mg b.d.3–6 In order to try to compensate for the smaller effect, higher doses of ranitidine are now increasingly being used in daily medical practice. It has been reported that higher ranitidine doses exhibit higher effectiveness.7,8
Lansoprazole is a relatively new proton pump inhibitor that, used at a standard dose of 30 mg o.d., was reported to be safe and not to significantly interfere with the metabolism of other drugs.9 We compared the effectiveness of standard-dose lansoprazole and high-dose ranitidine in the symptomatic improvement and endoscopy-monitored healing of patients with moderately severe reflux oesophagitis (grades II and III, Savary–Miller criteria10).
The trial was designed as a double-blind, randomized, parallel group, multicentre trial and was aimed to compare the effectiveness of standard-dose lansoprazole (30 mg o.d.) and ranitidine at twice the recommended daily dose11 (300 mg b.d.) in patients with endoscopically verified moderate to severe reflux oesophagitis. Following 4 weeks of double-blind treatment with either lansoprazole or ranitidine, patients were endoscoped again. The trial medication was stopped if healing of the oesophagitis had occurred but was continued if complete healing had not occurred. The unhealed patients were endoscoped again at 8 weeks after start.
The endoscopic healing rate after 4 weeks of treatment was chosen as the primary end-point. Healing rates on lansoprazole and ranitidine were anticipated to be 75% and 45%, respectively. Using a significance level of 5%, a power of 80%, and two-sided comparison, at least 41 patients per treatment arm were needed to show a statistically significant difference. To correct for drop-outs and non-evaluable cases the number of patients that were minimally required was set at 120.
Centers and patients
The trial was conducted in 20 hospital centres by 27 internists and gastroenterologists in The Netherlands. Prior to the start of the trial, ethical approval was obtained from a central ethics committee and from the local ethics committees of each of the participating centres. Written or oral witnessed informed consent was obtained from all patients prior to enrolment in the trial. Checks of the Case Report Forms and source data verification were performed by specialized monitoring people of the Medical Departments of Roussel BV, Hoevelaken, and the Department of Clinical Research of Janssen Pharmaceutica BV, Tilburg, the Netherlands. Out-patients with reflux oesophagitis were included between September 1993 and January 1995.
Patients aged 18 years or above were eligible if they had proven reflux oesophagitis grade II (longitudinal, confluent, non-circumferential erosions) or grade III (longitudinal, confluent, circumferential erosions that bleed easily) according to the classification of Savary and Miller, if they were able to understand and co-operate with the trial requirements and if they had given written or witnessed informed consent.10 Reflux oesophagitis was assessed by endoscopy within 10 days prior to enrolment.
Patients were excluded from participation if they had a bleeding ulcer, Zollinger–Ellison syndrome, a concurrent malignant disease, any uncontrolled significant disease or a history of vagotomy or gastrectomy, if there was any evidence of current alcohol or drug abuse or of hypersensivity to an H2-receptor antagonist or proton pump inhibitor, if they used any other anti-ulcer medication or anticoagulant drug during the trial period, if they had used any investigational drug within the past 4 weeks, or in case of pregnancy or lactation.
During the conduct of the trial the use of concomitant medications was allowed with the exception of anti-ulcer medication such as proton pump inhibitors, H2-receptor antagonists, mucosa protectives, prokinetics or antacids.
All patients had to take two (one white/orange and one yellow) capsules in the morning before breakfast and two (yellow) capsules in the evening at bedtime. Depending on the treatment, white/orange capsules contained 30 mg lansoprazole (Prezal) or 150 mg ranitidine, yellow capsules contained placebo or 150 mg ranitidine. Double-blind trial medication was prepared and properly labelled by the Pharmaceutical Production Unit of the Ziekenhuisapotheek Midden-Brabant, Tilburg, the Netherlands, as follows.
Ranitidine tablets (Zantac300) were commercially obtained, powdered and the resulting material was used to prepare capsules that contained 150 mg ranitidine each. Ranitidine capsules were within 5% of the required weight. Content uniformity and dissolution rate were found to meet the USP requirements for ranitidine tablets. Dissolution profiles showed that equal amounts of ranitidine were released from the commercially obtained tablets and from the locally prepared capsules, but ranitidine dissolution rate was somewhat faster from the capsules than from the tablets. Morning and evening dosing were blistered and separately packed in order to optimize patient compliance. Trial medication sets were numbered and randomized in blocks of four. Per trial centre, trial medications were assigned to patients in ascending number.
Efficacy of the trial medication was assessed at each visit by endoscopy and symptomatic evaluation. The severity of reflux oesophagitis was graded according to the Savary–Miller classification and endoscopic healing was defined as complete re-epithelization of all erosive or ulcerative lesions (grade 0).10 The symptoms heartburn, retrosternal pain, regurgitation, dysphagia, belching, bloating and abdominal distension were each scored on a 4-points scale, ranging from 0 (absent) via 1 (mild) and 2 (moderate) to 3 (severe). In addition, overall symptom severity was rated by the patients on a visual analogue scale (VAS) ranging from 0 mm (no symptoms at all) to 100 mm (symptoms could not be more severe).
Routine laboratory screening was performed in each participating centre for safety reasons. Screening was carried out at the start and after 4 weeks of treatment. If the patient had not healed after 4 weeks, laboratory screen was repeated after 8 weeks. Laboratory screen included haematology (haemoglobin, haematocrit, red blood cells, white blood cells and differential, platelets) and biochemistry (ASAT, ALAT, γ-GT, total bilirubin, potassium, glucose, total protein, uric acid, creatinine and total cholesterol). Separate serum samples were collected for the central assessment of gastrin levels which was carried out in the Laboratory of Gastrointestinal Diseases of the Academic Hospital, Nijmegen St. Radboud, Nijmegen, the Netherlands.
At each visit, a general physical examination was performed and the patient’s body weight, blood pressure and heart rate were assessed. The occurrence of adverse events was monitored by asking a non-leading question at each follow-up visit. Patients were also given a diary card and were encouraged to use it for the daily registration of severity of heartburn and retrosternal pain, to keep track of proper intake of trial medication and to record any adverse events. Diaries were turned in at each follow-up visit and were inspected by the investigator prior to the registration of adverse events.
The principal measurement of efficacy was the endoscopic healing rate after 4 weeks of treatment. Secondary efficacy parameters were the cumulative endoscopic healing rate after 8 weeks, the symptom resolution rates after 4 and 8 weeks, improvement of the investigator-rated symptom scores and of the patient-rated visual analogue scale (VAS). All data shown at week 8 refer to the cumulative number of patients, so including those healed at week 4 (end-point analysis).
Analysis of laboratory safety data was based on a classification into three categories according to the standards of each local lab, i.e. below, within, or above the normal range, and for each parameter, changes in category were compared between treatment groups. Adverse events were carefully monitored and were reported for all patients for whom follow-up data were available.
Statistical analysis of the results was performed using the statistical package SAS, version 6.10. Binomial proportions were compared using Fisher’s exact probability test while ordinal parameters were analysed with the Wilcoxon two-sample test for intergroup differences. For the intragroup comparisons the Wilcoxon signed rank test was used. All tests were performed two-sided, using a significance level of 5%. Data entry and statistical analysis were carried out at the Department of Clinical Research, Janssen-Cilag B.V., Tilburg.
At total of 133 patients were included; 68 were treated with lansoprazole and 65 with ranitidine according to a previously determined randomization scheme (Table 1). Any difference between the treatment groups at the start must therefore be due to chance. The analysis indicated that, indeed, treatment groups had been comparable at the start for most parameters related to demography, risk factors, and severity of reflux disease. However, the number of smokers was statistically significantly higher (31% vs. 13%, P=0.020) and the dysphagia score just lower (P=0.050) in the ranitidine, compared to the lansoprazole group (Tables 1 and 3.
Table 1. . Patient population and severity of reflux oesophagitis at start (mean ± s.d.)
Table 3. . Mean severity scores of gastro-oesophageal symptoms at the start (week 0) and after 4 and 8 weeks of treatment with either lansoprazole or ranitidine, intention-to-treat analysis. Statistical significance of the difference between treatments is indicated by the P-value, tendencies (0.05 < P < 0.10) are mentioned between brackets
All patients had follow-up endoscopy and could be included in the intention-to-treat analysis. Seven patients were excluded from the per protocol analysis in the lansoprazole group, i.e. five for violation of the selection criteria, one for the use of disallowed co-medication, and one because of poor compliance (less than 70% of medication taken). Two patients were excluded in the ranitidine group for violation of selection criteria. Violation of selection criteria involved four cases without recent endoscopic prove of oesophagitis (endoscopy was not performed within 10 days prior to enrolment), and three cases of alcohol abuse.
The intention-to-treat analysis showed that endoscopic healing occurred in both treatment groups after 4 and 8 weeks (P < 0.001), but that healing was statistically significantly more pronounced in the lansoprazole group than in the ranitidine group (Table 2). The difference between the treatment groups was also apparent in the improvement of Savary–Miller grades ( Figure 1) and was particularly marked in patients with grade-III oesophagitis ( Figure 2).
Table 2. . Endoscopic healing (A) and symptom resolution (B) after 4 and 8 weeks treatment with either lansoprazole (30 mg o.d.) or ranitidine (300 mg b.d.). N = number of patients; P-values indicate the statistical significance of the difference between treatments (intention-to-treat analysis)
Results of the per protocol analysis revealed that endoscopic healing was achieved in 44 out of 58 (76%) of patients in the lansoprazole group vs. 26 out of 63 (41%) of patients in the ranitidine group after 4 weeks (P < 0.001 between groups), and in 55 out of 61 (90%) vs. 40 out of 63 (64%) of patients after 8 weeks, respectively (P < 0.001 between groups). Results of the per protocol analysis were thus very similar to those of the intention-to-treat analysis.
Other parameters further supported the better efficacy of lansoprazole, such as the height of the erosive oesophageal segment which fell from an average of 38.0 mm at the start to 3.8 mm and 1.5 mm in the lansoprazole group after 4 and 8 weeks, respectively, and from 39.7 mm at the start to 19.2 mm and 10.4 mm in the ranitidine group (intention-to-treat analysis, results not shown). Both after 4 and after 8 weeks, the height of the erosive segment was statistically significantly lower in the lansoprazole group than in the ranitidine group (P < 0.001).
Results on overall symptom resolution (Table 2) were very similar to those on endoscopic healing (Table 2). Severity scores per symptom were analysed for patients who indicated a score> 0 at any time during the trial. Table 3 shows that heartburn and retrosternal pain were the most common symptoms, dysphagia was the least common symptom (but happened to be somewhat more frequent in the lansoprazole group than in the ranitidine group, only at the start). After 4 and 8 weeks, heartburn, retrosternal pain, regurgitation, dysphagia, belching, bloating and abdominal distension had all statistically significantly improved in both treatment groups, but symptom scores were statistically significantly lower in the lansoprazole group than in the ranitidine group for heartburn (after 4 and 8 weeks) and for retrosternal pain and belching (after 4 weeks, with just a tendency [0.05 < P < 0.10] after 8 weeks).
Patient-rated VAS scores indicated a statistically significant improvement of symptoms in both treatment groups, but the improvement was more pronounced in the lansoprazole group than in the ranitidine group both after 4 and 8 weeks ( Figure 3).
Daily registration of the severity of heartburn and retrosternal pain in the patients’ diaries allowed analysis of the improvement of these symptoms over the days. As can be seen for heartburn ( Figure 4), improvement tended to be better on ranitidine during the first 2 days, but was more pronounced on lansoprazole after the 3rd day. The pattern for the rest of the month persistently indicated better improvement of heartburn with lansoprazole, although statistical significance was reached at just a few days ( Figure 4).
Adverse events were mentioned by a similar number of patients in each group, i.e. by 34 out of 68 (50%) in the lansoprazole group and by 30 out of 65 (46%) in the ranitidine group. Only a minority of the adverse events was indicated as possibly or probably drug related, i.e. 20% of the adverse events in the lansoprazole group and 27% in the ranitidine group. Adverse events that were mentioned by at least three patients in one of the treatment groups are listed in Table 4. Headache, diarrhoea, common cold and influenza appeared to be more frequent in the lansoprazole group, while a sore throat was mentioned more often in the ranitidine group, but differences between treatments were not of statistical significance. No significant changes were found in any of the routine laboratory parameters either in the lansoprazole group or in the ranitidine group nor between groups. Systolic and diastolic blood pressure, heart rate and body weight did not change over the 8 weeks treatment period.
Table 4. . Most frequent adverse events
Fasting serum gastrin levels were similar in both treatment groups at the start, but rose statistically significantly during the first 4 weeks of the double-blind treatment phase in both groups, i.e. from 37.8 to 66.6 pmol/litre in the lansoprazole group and from 35.4 to 48.8 pmol/litre in the ranitidine group (results not shown). The rise in gastrin was statistically significantly greater in the lansoprazole group than in the ranitidine group (P < 0.01).
Since the number of smokers had been larger in the ranitidine group than in the lansoprazole group, separate analyses were performed of smokers and non-smokers. First, characteristics of smokers and non-smokers were similar in percentage of men/women, mean height and weight, and the use of alcohol, but smokers were younger than non-smokers (mean age was 45 vs. 56 years, P < 0.05) and smokers had reflux for a shorter time than non-smokers (first reflux episode was a mean of 14 and 25 months ago, respectively; P < 0.05). Second, it appeared that smokers in the lansoprazole group and smokers in the ranitidine group used a similar number of cigarettes or cigars, indicating that any difference between the treatment groups was not related to the extent of smoking.
After 4 weeks treatment, the overall endoscopic healing rate was significantly higher in non-smokers (n=104) than in smokers (n=29), i.e. 66% vs. 41% (P=0.019). This effect was rather outspoken in the lansoprazole group, i.e. 49 out of 59 (83%) of non-smokers vs. 5 out of 9 (56%) of smokers were healed (difference between smokers and non-smokers did not reach statistical significance, P=0.079), but was less apparent in the ranitidine group, i.e. 20 out of 45 (44%) of non-smokers vs. 7 out of 20 (35%) of smokers were healed. After 8 weeks treatment, the healing rate on lansoprazole was higher in non-smokers than in smokers, i.e. 95% vs. 67% (P < 0.001), but the healing rate on ranitidine was similar in non-smokers and smokers, i.e. 67% vs. 65% (N.S.).
The results of the present study indicate that lansoprazole (30 mg in the morning) is superior to ranitidine (300 mg in the morning plus 300 mg in the evening) in the healing of reflux oesophagitis. The data are particularly striking in view of the fact that the majority of the patients had only modest (grade II) oesophagitis and that ranitidine was used at a twice the higher dose than is recommended for this type of oesophagitis.11,12 Overexposure to gastric acid is known to be the principal factor in the development of erosive lesions of the oesophagus and the extent of the lesions is an important negative factor in the healing process.2 This was supported once more by the presently found lower healing rates in more severe (grade III) as compared to less severe (grade II) oesophagitis patients.
Both ranitidine and lansoprazole have been demonstrated to dose-dependently inhibit gastric acid secretion,13,14 but standard-dose lansoprazole appeared to be superior to standard-dose ranitidine in the treatment of duodenal ulcers,15 dyspeptic complaints16 and reflux oesophagitis.17 The results presented here indicate that even a daily dose of 600 mg ranitidine is markedly less effective than standard-dose lansoprazole. Raising the ranitidine dose further, even up to 1200 mg per day, does not appear to yield any increase in effectiveness.18 Our study therefore provides additional evidence that mild to moderate forms of reflux oesophagitis are better treated with standard-dose proton pump inhibitor than with high-dose H2-receptor antagonist.19,20
Not only the endoscopically verified lesions, but also the subjective symptoms heartburn and, to a lesser extent, retrosternal pain and belching improved significantly better on lansoprazole than on ranitidine. No significantly better improvement of the symptoms regurgitation, dysphagia, abdominal distension and bloating was found. These observations are in line with previous reports that indicate heartburn as the most commonly and consistently reported symptom of gastro-oesophageal reflux disease.20,21 Diaries revealed that the initial improvement in heartburn occurred earlier in patients treated with ranitidine than with lansoprazole, while after 3 days of treatment lansoprazole was persistently more effective than ranitidine in providing relief of heartburn. The reason for this may well lie in the balance between the fading effect of H2-receptor antagonists, often indicated as ‘tolerance phenomenon’ or ‘exaggerated first dose response’,22,23 and the rising effectiveness of the proton pump inhibitors during the first 3–5 days of treatment.24 The latter effect is probably the result of a gradual increase in the percentage of proton pumps that are knocked out by subsequent doses of the irreversible proton pump inhibitor.22,25 The initial lag phase in the action of lansoprazole is, however, less prolonged than with omeprazole, which may be due to the higher bioavailability of lansoprazole compared to omeprazole.24,26 So a faster relief of heartburn can be achieved when lansoprazole is selected as the proton pump inhibitor to treat reflux symtoms26–28. Although the difference in the onset of action of H2-receptor antagonists and proton pump inhibitors is known from pH-metry studies,29 the data presented here indicate that this is also reflected in the disappearance rate of the primary acid reflux symptom, heartburn.
Previous studies have shown that the incidence of adverse events in patients treated with ranitidine or lansoprazole was similar to that of placebo.30 In the present study, adverse events were mild and the incidence was similar in both treatment groups. Patients treated with lansoprazole reported headache, diarrhoea, influenza and common cold more frequently, but a sore throat less frequently, than patients treated with ranitidine. It is, however, unclear whether this bears any clinical relevance since the difference between the treatment groups was not of statistical significance. From literature we know that headache and diarrhoea are relatively frequently reported with proton pump inhibitors.27,30 The origin of the headache is still unclear, but profound inhibition of gastric acid secretion has been found to result in small bowel bacterial overgrowth, leading to higher defecation frequency, looser stool consistency and, in a number of patients, diarrhoea.31–33 A more extensive suppression of gastric acid secretion by proton pump inhibitors, compared to H2-receptor antagonists appears to also underlie the higher rise in serum gastrin levels following treatment with lansoprazole.17,19,34
Another interesting finding of the present study was the influence of smoking on oesophagitis healing. It is known that smoking negatively interferes with the healing of reflux oesophagitis.12,19 In accordance with another report21 we did not find a difference in healing rate between smokers and non-smokers when treated with ranitidine, but non-smokers responded significantly better to lansoprazole than did smokers. At randomization, patients had not been stratified for smoking habit, and the final analysis revealed that smokers were over-represented in the ranitidine group relative to the lansoprazole group. A potentially negative impact of the high number of smokers on the overall effectiveness of ranitidine is, however, ruled out since ranitidine was found equally effective in smokers and non-smokers. Reduced effectiveness in smokers has earlier been reported for proton pump inhibitors in the treatment of reflux oesophagitis34 and of duodenal ulcers irrespective of antibiotic co-treatment to eradicate Helicobacter pylori.35–37 The reason why proton pump inhibitors are less effective in smokers has not yet been elucidated and deserves further investigation.
In summary, (1) lansoprazole 30 mg o.d. is clearly superior to ranitidine 300 mg b.d. in the healing of grade II and III reflux oesophagitis. (2) Smoking appears to reduce the healing of oesophagitis with lansoprazole. (3) Reflux symptoms improve better when treated with lansoprazole than with high-dose ranitidine, but the onset of symptom relief occurs somewhat earlier with ranitidine. (4) Adverse events are mild and occur with similar frequency during both treatments.
We wish to thank the following persons for special contributions to the trial: Mrs KA Simons, RPh, Ziekenhuisapotheek Midden-Brabant, Tilburg, for the preparation of the trial medication. A Tangerman, PhD, Laboratorium voor Maag-, Darm en Leverziekten, Nijmegen, for the assessment of serum gastrin levels. R de Graaf, Janssen-Cilag BV, Tilburg, Mrs PGM van Oosterbosch, Mrs TH Hendriks-Luiten, M Gerretsen, PhD, and Mrs YJB van Megen, PhD, Hoechst Marion Roussel BV, Hoevelaken, for their contributions in the organization and proper conduct of the trial. Mrs RMAB Jansen and Mrs MJ Lahaye-Ghering, Janssen-Cilag BV, Tilburg, for developing the case report form and for performing the statistical analysis, respectively.
Financial support provided by Janssen-Cilag B.V., Tilburg, and Hoechst Marion Roussel B.V., Hoevelaken, The Netherlands, is gratefully acknowledged.
The Dutch Lansoprazole Study Group consisted of the following specialists: E Balk, Bennekom; PHE Berghuis, Helmond; R Beukers, Dordrecht; GH de Groot, Beverwijk; SJ Graafsma, Tilburg; PWE Haeck, Stadskanaal; C Harlaar, Winterswijk; BP Hazenberg, Dordrecht; SS Hofstra, Leeuwarden; JBMJ Jansen, Nijmegen; HJA Jebbink, Leeuwarden; FJM Klessens-Godfroy, Rotterdam; PHT Koch, Helmond; FAT Lustermans, Heerlen; WW Meijer, Den Helder; LA Noach, Amstelveen; RE Nikkels, Heerlen; PR Oosting, Purmerend; P Snel, Amsterdam; JM Spijker, Terneuzen; JAJ Staal, Hengelo; TG Tan, Hengelo; RJ Timmerman, Den Helder; JJ van As, Bennekom; CTBM van Deursen, Brunssum; HPC van Roermund, Roosendaal; JR Vermeijden, Haarlem; AMH Wetzels, Stadskanaal.