Systematic review of PPI and H2A in GORD


Reflux oesophagitis: quantitative systematic review of the evidence of effectiveness of proton pump inhibitors and histamine antagonists



R A Moore DSc
P Wiffen BPharm
HJ McQuay, DM
Pain Research &
Nuffield Department of Anaesthetics
Oxford radcliffe Hospital
The Churchill
Headington
Oxford

and

C Phillips PhD
Department of Nursing, Midwifery & Healthcare
University of Wales, Swansea
Singleton Park
Swansea


Introduction



This systematic review was prompted by a Bandolier article in late 1994 which examined a comparative trial of omeprazole and ranitidine in reflux oesophagitis healing [1]. This was at a time where the number-needed-to-treat (NNT) was just beginning to be used by Bandolier and others, and the NNT of 3 generated by the data from the Dent paper could still not be put into perspective. Reflux healing and maintenance seemed to be an important area, with good trials, and where a systematic review with economic evaluation was needed.

Two years or so later we have done that. The change in systematic review methods that has occurred during those two years has been significant in the area of treatments where most patients have the expected event. We understand NNTs much better, and we have a better handle on using economic evaluations.

WHAT IS REFLUX OESOPHAGITIS?



Gastro-oesophageal reflux is the process of reflux of stomach contents into the oesophagus. The consequence is a chemical insult from acid and enzymes. Reflux happens commonly but infrequently in many people, and it does not cause major harm because the natural peristalsis of the oesophagus clears the refluxate back into the stomach. In others where acid reflux from the stomach is persistent, the result is damage to the oesophagus causing symptoms or macroscopic oesophageal damage, and here gastro-oesophageal reflux disease (GORD) can be said to be present [2].

Symptoms



GORD produces a characteristic set of symptoms, though significant oesophagitis can be present without symptoms. Heartburn is most common; it is often described as gnawing or burning pain behind the sternum, and it may be severe enough to radiate to the arm or jaw. Usually occurring within an hour or so of a meal it can be made worse by lying down. Heartburn can wake the patient at night, and is most frequent in those with the most severe disease. Heartburn occurs occasionally in many people after a fatty or spicy meal, but in GORD the symptoms occur frequently after any sort of meal. Alcohol and coffee also induce symptoms. Antacids relieve symptoms.

Other symptoms include painful swallowing - when hot liquids or spicy foodstuffs pass over inflamed oesophageal tissue. Regurgitation is common.

Other symptoms are cause for alarm. These include difficulty in swallowing (dysphagia), weight loss, and anaemia. Dysphagia may be due to severe oesophagitis causing a stricture, and these symptoms if severe suggest urgent referral to a gastroenterologist.

Atypical symptoms may also occur. They include oesophageal spasm which causes pain similar to ischaemic heart disease radiating to the arm or jaw, pulmonary symptoms like hoarseness, nocturnal wheezing or coughing (associated with aspiration while supine), or bronchitis, pneumonia and lung abscesses.

Risk factors



Lifestyles which predispose to GORD include smoking, fatty foods, alcohol and coffee. Obesity is a risk factor, as is work or hobbies that require stooping or lying down, like gardening or car maintenance. Hiatus hernia increases the risk of reflux. Use of non-steroidal anti-inflammatory drugs is also associated with oesophageal erosion, or with development stricture (reviewed in [3]).

Oesophageal damage



The lining of the oesophagus is ill-equipped to resist stomach acids. The stomach has cells which produce a bicarbonate-mucus barrier which protects them from stomach acid: oesophagus does not have this protective barrier. The result of refluxed acid stomach acid is to damage the lining of the oesophagus. This damage can be microscopic, but is often macroscopic and seen on endoscopy.

Endoscopic oesophageal damage is graded on a scale of 1-4 with increasing severity of damage.


The correlation between endoscopic grading and symptoms is not good. Severe symptoms can occur with low grade oesophagitis, and conversely severe oesophageal damage can sometimes occur with few symptoms.

Oesophagitis and cancer



Long standing oesophagitis may be associated with the development of Barrett's oesophagitis. A study of 428 patients with chronic gastrointestinal reflux symptoms who received endoscopy as an initial screening procedure showed an incidence of Barrett's oesophagitis of 13% (and of oesophageal cancer of 0.5%) [4].

In Barrett's oesophagitis the normal whitish/grey squamous epithelium of the oesophagus is replaced by red columnar gastric epithelium in the lower oesophagus. Deep ulcers may develop which can cause frank bleeding or perforation. Barrett's oesophagitis is associated with an increased risk of developing oesophageal cancer, which has been estimated as occurring 30-40 times more frequently in Barrett's oesophagitis [2]. More recently a systematic review has shown that Barrett's oesophagitis is also associated with a five-fold increased risk of colon cancer (95% confidence interval 2.4 -12)[5].

Cause of reflux



An abnormally high level of episodic exposure of the oesophagus to gastric contents is due mostly to abnormal neural control of the lower oesophageal sphincter. Most reflux occurs when this sphincter relaxes, though subsequent clearance of gastric contents from the oesophagus can also be impaired in those with established oesophagitis.

PREVALENCE AND NATURAL HISTORY


Prevalence



Indigestion occurs in most people at some time or other. Many individuals will suffer mild or sporadic symptoms, some will have symptoms sufficiently troublesome to consult a general practitioner, and a few will have symptoms sufficiently severe to need specialist consultation. A good review of the epidemiology looking at the numbers in each of these groups has been done by Sontag [3].

Surveys in the USA have indicated that 44% of the adult population has heartburn at least once a month. Six out of ten of these never consult a GP about it. About 13% of the adult population take some type of indigestion aid at least twice a week. There seem to be few reliable figures on the numbers of patients who present to GPs with reflux symptoms, but a health authority with an adult population of 470,000 will have an estimated 7,500 patients seeing GPs with dyspepsia and almost 3,000 having an endoscopy, over half for oesophageal or gastrointestinal problems, including suspected ulcer [6].

The prevalence of endoscopic oesophagitis in the community has been estimated in a number of studies. A review of the Swedish National Register estimated the prevalence of oesophagitis to be less than 5% in the 55 year old group [7]. Two other reports have estimated the prevalence to be of the order of 2% [8][9]. Clearly symptoms of reflux are up to an order of magnitude higher than the prevalence of endoscopic oesophagitis, probably reflecting the proportion who have endoscopy.

How many patients presenting to the GP with symptoms are likely to have endoscopic oesophagitis? Several studies of patients presenting to GPs with symptoms show rates of oesophagitis in the range 40 - 65% [10][4][11]. Open access endoscopy in the UK presents a somewhat different picture. In Gloucestershire the proportion of patient referred by GPs found to have oesophagitis rose from 7% in 1977, rising to 19% by 1987 [12]; it is not clear whether this increasing referral rate is a consequence of reduced ulcer prevalence. A retrospective review of the results of over 8,000 diagnostic endoscopies in Hampshire showed that GORD accounted for 23% of all upper gastrointestinal conditions diagnosed [13].

Natural history



The natural history of oesophageal reflux is not well studied, and the frequency of spontaneous remission remains unclear. There is a spectrum of episodic symptoms with periods of remission, through to constant symptoms needing ongoing treatment.

TREATMENT OPTIONS



There are a number of treatment options, which will be reviewed here only briefly. There are useful reviews [3][14][15].

Lifestyle change



Changes in behaviour suggested to patients involve avoiding large meals and provocative food and drink and physical activities. These are probably more useful than stopping smoking, losing weight and raising the bedhead. All of these may be difficult for the patient to follow and are of unproven benefit. Even if the patient is successful in changing lifestyle, only about one in five will benefit [16].

Antacids and alginates



Alginates seem to be more useful than antacids in controlling symptoms, but no placebo-controlled study has shown healing of oesophagitis with either of these classes [15].

Mucoaprotective agents



Preparations such as sucralfate act locally by shielding damaged mucosa or stimulating local defence mechanisms. A problem is that swallowing induces oesophageal peristalsis so reducing mucosal protective time. There seems to be little enthusiasm for this mode of treatment [14].

Motility stimulants



Cisapride is the one stimulant that is well studied, with a principal effect on oesophageal acid clearance. It appears to be effective in grades 1 and 2 oesophagitis but efficacy on more severe disease is not proven [14].

Suppression of acid secretion



Inhibition of gastric acid secretion makes gastric juice less injurious to the oesphageal mucosa, but does not stop reflux. The effectiveness of histamine H2-receptor antagonists (H2A) and proton pump inhibitors (PPI) has been reviewed [14][3], though both include non-randomised studies and those published only in abstract form. Systematic review of randomised comparisons of H2A and PPI is the subject of this review.

Both classes of drug can be effective and doses are flexible to cope with differing severity of disease. However, as many as 15% of patients will fail to heal even on the highest dose of the most effective drug. This suggests that factors other than refluxed acid are important in the development of oesophagitis.

Antireflux surgery



Antireflux surgery involves creating an artificial sphincter by drawing up part of the stomach wall around the lower part of the oesophagus. Surgery is not common, though there is optimism about the potential of laparoscopic techniques [3][2]. The evidence is not yet in for the place of newer surgical methods to be ascertained with certainty. It may be considered for younger patients with proven reflux who seem likely to need lifelong drug therapy [15].

Treatment choices



Most authorities suggest a graded response of treatment to symptom severity, using the lowest dose of any therapy that is likely or proven to be effective. Dent [2], for example suggests the following treatment responses to increasing levels of reflux symptoms:-

Low antacids, non-drug measures

Middle Normal dose of H2A, cisapride or sucralfate

Upper 2-4 times normal dose of H2A, omeprazole 10 mg, lanzoprazole 15 mg

Highest omeprazole 20 mg, lanzoprazole 30 mg, antireflux surgery

Maintenance



Drug treatments are aimed at oesophageal healing, with treatment lasting four to 12 weeks. After complete healing with H2A or PPI, symptomatic recurrence within six months occurs in some 80% of patients [3]. Because symptoms of reflux rarely disappear completely, most patients will continue to take some medication for long periods, at least intermittently. The choice of ongoing treatment will depend on the severity and frequency of symptoms, but will include alginates, H2A and PPI.

Systematic review methods



Reports were sought of comparisons between any proton pump inhibitor (PPI) and any histamine-2 antagonist (H2A) in reflux oesophageal disease with endoscopic healing as the outcome measure. Both short-term healing and long-term maintenance were included.

The tools used in this systematic review include the L'Abbé plot of the data [17], and the calculation of the number-needed-to-treat (NNT) as well as the calculation of relative risk to establish statistical significance. Some words of explanation may aid understanding of the review that follows.

LAbbé plot



This is simply a scattergram of the raw data from all the trials included in a particular comparison. Summary data concerning a particular outcome (endoscopic healing in this instance) for each treatment comparison is plotted for each trial, so that each point is the data from one trial. At 45deg. is drawn the line of equality. The distribution of the points on one side or other of the line of equality gives a simple visual presentation of the benefit (if any) of one treatment over another, together with an indication of how consistent results are between the different trials.

NNT



The NNT is a measure of the treatment-specific benefit of one patient over another using any clinical end-point chosen. The simple formula to calculate it is given in Figure 1. The best NNT for a treatment would be 1, and the lower the NNT the greater the treatment-specific effect. Because few comparisons have treatments which are 100% effective and controls that are 0% effective, NNT values are usually higher than 1.

Low NNTs of the order of 1 can be found for antimicrobial treatments of susceptible pathogens (for antibiotic eradication of Helicobacter pylori , for instance [6]). Effective interventions for treatments (rather than prophylaxis) are usually in the range of 2 - 4.

SEARCHING



MEDLINE was searched electronically (Knowledgeserver vXX) between 1980 and 1996 (date of last search May 1996). A number of sensitive strategies were employed, using MeSH terms "gastroesophageal reflux disease" and treatment", as well as searches for individual therapies - "omeprazole", "lanzoprazole", "cimetidine", "ranitidine", "nizatidine" and "famotidine". Electronic abstracts were read, and hard copies were obtained of papers which may have been randomised comparisons of any PPI with any H2A, or were considered to have any economic analysis data. Independent searches were conducted by the medical information service of Astra Pharmaceuticals Ltd. In addition, reference lists of reviews and reports were examined for additional papers. Only full published articles, not abstracts, were sought. Neither authors nor manufacturers were contacted.

Papers were included in the systematic review of effectiveness if they fulfilled the following criteria:-

DATA ABSTRACTION



Each report which could possibly be described as an RCT was read independently by each of the authors and scored using a three item quality scale [18]. The scale takes into account proper randomisation, double-blinding and reporting of drop-outs and withdrawals. The minimum score of an included RCT was 1, the maximum score 5. There was a pre-hoc agreement that trials without randomisation or with an inadequate randomisation method (without concealment of treatment allocation) would be excluded from further analysis. Consensus was then achieved. Pre-hoc validity criteria were number of patients per treatment group >= 10 [17], pre-treatment endoscopic oesophagitis grade 1 or more, and standardised methods of measuring endoscopic healing.

Information was taken from each report on the number and conditions of the patients, design of the trial, dose of drug and duration of treatment and outcomes. For short-term healing studies the number of patients in each treatment group healed at four and eight weeks was abstracted (defined as oesophagitis grade 1 or less, or reduction in grade to 1), together with any drug-related adverse-event study withdrawals. For long-term studies the outcome chosen was the number of patients in endoscopic remission at 12 months.

Data abstraction was performed so that the analysis could be done on an intention-to-treat basis. Since authors had variable views as to the definition of intention-to-treat analysis we imposed the most conservative definition - that the denominator in any trial was all patients randomised, irrespective of reasons for study withdrawal or lack of follow-up data.

Data analysis



Relative risk estimates were calculated with 95% confidence intervals (CI) using a random effects model [19]; the random effects model was chosen because, while data sets were predominantly homogeneous, this produced the most conservative estimate. Number-needed-to-treat and 95% confidence intervals were calculated by the method of Cook & Sackett [20]. A statistically significant difference from control was assumed when the 95% confidence intervals of the relative risk did not include 1. Calculations were performed using Excel v 5.0 on a Macintosh 7100/80.

Systematic review results


SHORT-TERM HEALING



Twenty-three reports with 5,118 patients fulfilled the inclusion criteria and details of the reports are shown in Table 1. One report [21] was had no endoscopic healing data, but did have adverse event information. Of the reports with endoscopic healing, ten (1393 patients) compared omeprazole with ranitidine, two reports (339 patients) omeprazole with cimetidine and three reports (525 patients) lanzoprazole with ranitidine. The following comparisons were found in one trial each: nitazidine and placebo, ranitidine and placebo, omeprazole and placebo, lanzoprazole and omeprazole, omeprazole with ranitidine plus metoclopramide, omeprazole with ranitidine or famotidine and omeprazole at different doses.

Quality scores were high. Four studies had a score of 2, three of 3, eleven of 4 and five of 5. The median score was 4.

The most commonly used doses of drugs used were: omeprazole 20 mg daily (15 studies) and 40 mg (four studies); ranitidine 300 mg daily (usually in a divided dose, 14 studies); cimetidine 1600 mg daily (divided dose); lanzoprazole 30 mg daily.

Omeprazole compared with ranitidine



Ten studies examined endoscopic healing. Omeprazole 20 mg was compared with ranitidine 300 mg in six trials [23][24][26][30][31][33]. Omeprazole 40 mg was compared with ranitidine 300 mg in one study [25]. Omeprazole 40 mg was compared with ranitidine 600 mg by Lundell and colleagues [29]. Green and colleagues [43] used starting doses of 20 mg omeprazole and 300 mg ranitidine, withn doubled doses in thiose who had not healed at 4 weeks. Omeprazole 60 mg was compared with ranitidine 300 mg in one study [22].

There was no simple relationship obvious between dose and outcome (Figure 2) and so these studies were combined. Relative risks for individual trials are shown in Figure 2; nine of the ten trials had a lower 95% confidence interval which did not include 1 and therefore showed omperazole to be statistically superior to ranitidine.

Healing occurred faster and more extensively with omeprazole than with ranitidine, and the extent of the superiority of omeprazole was the same at both times.


Healed with omeprazole
Healed with ranitidine
RR (95%CI)
NNT (95%CI)
4 weeks
410/696 (62%)
206/695 (30%)
2.0 (1.7 - 2.5)
3.5 (2.9 - 4.0)
8 weeks
520/696 (78%)
310/695 (44%)
1.7 (1.5 - 2.0)
3.3 (2.9 - 4.0)

Omeprazole compared with cimetidine



One study [27] compared omeprazole 20 mg with cimetidine 1600 mg and another [28] compared omeprazole 40 mg with cimetidine 1600 mg. Healing occurred faster and more extensively with omeprazole than with cimetidine, and the extent of the superiority of omeprazole was the same at both times.


Healed with omeprazole
Healed with cimetidine
RR (95%CI)
NNT (95%CI)
4 weeks
93/169 (54%)
43/170 (25%)
2.2 (1.6 - 2.9)
3.4 (2.5 - 5.1)
8 weeks
117/169 (68%)
53/170 (27%)
2.3 (1.5 - 3.5)
2.6 (2.1 - 3.5)

Lanzoprazole compared with ranitidine



Two studies [42][41] compared lanzoprazole 30 mg (192 treated patients) with ranitidine 300 mg (204 treated patients) and showed that healing occurred faster and more extensively with lanzoprazole than with ranitidine, and the extent of the superiority of lanzoprazole was the same at both times.


Healed with lanzoprazole
Healed with ranitidine
RR (95%CI)
NNT (95%CI)
4 weeks
157/192 (82%)
93/204 (44%)
1.8 (1.4 - 2.4)
2.8 (2.2 - 3.6)
8 weeks
175/192 (91%)
123/204 (58%)
1.5 (1.4 - 1.7)
3.2 (2.6 - 4.3)

Healing - All PPI vs all H2A



The data from all these studies for healing at eight weeks are combined in Figure 3. Overall eight week healing with PPIs was 78% compared with 44% for H2As. For all PPIs compared with all H2A at eight weeks, the relative risk was 1.7 (1.6 - 1.8) and the NNT 3.2 (2.8 - 3.6).

This means that for every three patients with reflux oesophagitis treated with a proton pump inhibitor, one more will have their oesophagitis healed who would not have done had they been treated with a histamine antagonist.

Adverse events - All PPI vs all H2A



The rate of occurrence of study withdrawals because of drug-related adverse events is shown in Figure 4. For PPI, the rate of adverse event withdrawals was 2.5%, and for H2A it was 4.2%. This tendency for fewer adverse event withdrawals with PPI was significant - relative risk 0.61 (0.41 - 0.91). The NNH was 50 (26 - 251). This means for every fifty patients with reflux oesophagitis treated with a proton pump inhibitor, one will not have a serious treatment-related complication who would have done had they been treated with a H2A.

PPI - other comparisons with H2A



An open study (Marks et al, 1994) compared omeprazole 20 mg with either ranitidine 300 mg or famotidine 40 mg daily in 36 dysphagic patients with peptic stricture and erosive oesophagitis. The outcome was endoscopic healing (grade 0) at 6 weeks. All 18 patients given omeprazole were healed compared with 10 of 18 patients given ranitidine or famotidine.

In another open study, Robinson et al (1993) compared omeprazole 20 mg with ranitidine 300 mg plus 40 mg metoclopramide daily in 184 patients with reflux oesophagitis grade 2-4. The end point was endoscopic healing at grade 0 or 1. At eight weeks 75/92 patients given omeprazole had healed compared with 42/92 patients given ranitidine plus metoclopramide.

PPI comparisons



Bate et al [38] investigated 313 patients with reflux oesophagitis grade 2-4. This parallel-group study was open-label for the first four weeks with omeprazole 20 mg daily. In the second four week block, patients having been randomised at study entry to either 20 mg or 40 mg omeprazole if they were not heales and symptom free at four weeks. At eight weeks, 103 of 158 patients with the lower dose schedule had endoscopic healing grade 0, compared with 114/155 patients given the higher dose schedule.

Hatlebakk et al [36] investigated 229 patients with reflux oesophagitis grade 1 and 2 (75% grade 1) given either omeprazole 20 mg or lanzoprazole 30 mg daily. At eight weeks endoscopic healing occurred in 96/113 patients given omeprazole and 95/116 patients given lanzoprazole.

PPI comparisons with placebo



Sontag et al [35] randomised 230 patients with reflux oesophagitis grade 2 or above to omeprazole 20 or 40 mg daily, or placebo. At eight weeks, endoscopic healing grade 0 or 1 had occurred in 7/46 patients given placebo, 68/93 patients given omeprazole 20 mg and 66/91 patients given omeprazole 40 mg.

Healing with both doses was similar. Combining both omeprazole doses and comparing the combined group with placebo produced a relative risk of 4.8 (2.4 - 9.5) and a NNT of 1.7 (1.4 - 2.2).

H2A comparisons with placebo



Roufail et al [34] randomised 342 patients with reflux oesophagitis grade 2-4 to ranitidine 600 mg or 1200 mg daily or placebo. At eight weeks, endoscopic healing to grade 0 or 1 occurred in 68/109 patients given ranitidine 600 mg, 73/120 patients given ranitidine 1200 mg and 30/113 patients given placebo.

Healing with both doses was similar. Combining both ranitidine doses and comparing the combined group with placebo produced a relative risk of 2.4 (2.2 - 2.7) and a NNT of 4.8 (4.3 - 5.4).

LONG-TERM HEALING



Seven reports with 1,635 patients fulfilled the inclusion criteria, and details are shown in Table 2. Four reports (1094 patients) compared omeprazole with ranitidine, one omeprazole with placebo, one lanzoprazole with placebo and one omeprazole and ranitidine alone and in combination with cisapride. Two studies had a quality score of 3, four of 4 and one of 5.

The most commonly used doses were omeprazole 20 mg and ranitidine 300 mg daily.

Omeprazole compared with ranitidine



Data on the long-term effectiveness of daily omeprazole 20 mg (583 treated patients) against ranitidine 300-600 mg (446 treated patients) were obtained from five studies. Remission rates at 1 year were higher with omeprazole (average 70%) than ranitidine (30%), and are shown for individual trials in Figure 5. For all trials combined the relative risk was 2.3 (1.6 - 3.5) and the NNT 2.8 (2.3 - 3.4).

Adverse events sufficiently severe to cause drug withdrawal occurred in 32 of 534 patients given omeprazole (4.2%) and 35 of 357 patients given ranitidine (8.4%). The relative risk was 0.83 (0.28 - 2.49).

Omeprazole compared with placebo



Bate et al [47] randomised 193 patients with healed reflux oesophagitis and who were asymptomatic to omeprazole 10 mg daily, 20 mg daily or placebo. At 12 months, endoscopic remission was apparent in 9 of 62 patients given placebo, 30 of 60 patients given omeprazole 10 mg and 50 of 68 patients given omeprazole 20 mg. Combining the omeprazole patients produced a relative risk of 4.3 (2.3 - 8.0) and a NNT of 2.1 (1.7 - 2.8).

Lanzoprazole compared with placebo



Robinson et al [49], in a similar study but with lanzoprazole, randomised 173 patients with documented healing of erosive oesophagitis to lanzoprazole 15 or 30 mg daily or placebo. At 12 months, endoscopic remission was apparent in 14 of 57 patients given placebo, 47 of 60 patients given lanzoprazole 15 mg and 50 of 56 patients given lanzoprazole 30 mg. Combining the patients produced a relative risk of 3.4 (2.1 - 5.4) and a NNT of 1.7 (1.4 - 2.2).

Systematic review discussion



A notable feature of this systematic review was the methodological quality of the studies included. Khan et al [50] in a review of anti-oestrogens in male infertility have pointed out that, using the same validated quality inclusion scale as in this review, scores of 3 and above are indicative of higher quality and less bias. Only four of the studies included here had a score of 2; the others had scores of 3 and above, and the median score was 4. This means that the results of the review should not be subject to significant bias.

Short-term healing



Healing at four and eight weeks was consistently higher with PPIs than with H2As. There was consistency between the trials whatever the comparison drugs, and whether low or high doses were used. The NNT for PPI compared with H2A was approximately 3 for all comparisons. This means that for every three patients with reflux oesophagitis treated with a proton pump inhibitor, one more will have their oesophagitis healed who would not have done had they been treated with a histamine antagonist.

This NNT of 3 was produced with the most conservative estimate, namely the intention to treat analysis which includes patients randomised but not actually treated. It should be compared with NNTs of a similar order for effective treatments. For example, in acute pain of moderate to severe intensity the best analgesics have an NNT of 3 in producing at least 50% pain relief compared with placebo [51].

The NNT refers to the treatment-specific action of PPIs compared with H2As. H2As themselves produce a significant healing effect. That means that PPIs are very efficient at producing short term healing of reflux oesophagitis. In Figure 3 the 8 week healing rate with PPIs can be seen to be in the region of 60-90% compared with 20-70% with H2As.

Particularly interesting was the adverse event picture that emerged from the systematic review (Figure 4) using drug-related patient withdrawal from a trial as a measure of severity. While there was considerable scatter of results, PPIs never produced a drug-related withdrawal rate of more than 7%, while with H2As it was up to 12%. The increased withdrawal rate with H2As was statistically significant, with a number-needed-to-harm of 50.

Long-term maintenance



Long-term maintenance with omeprazole or ranitidine show rather variable results (Figure 5), but with clear advantages for omeprazole. One year remission with omeprazole was between 50 and 90%, compared with 10 - 50% with ranitidine. The resultant NNT of 2.8 shows omeprazole to be an effective treatment for long-term maintenance in those patients who need it. There were few adverse effects with omeprazole, and again ranitidine had about twice the rate of drug-related adverse-event withdrawal, though this did not achieve statistical significance with this sample.

References



1. Dent, J., Yeomans, N.D., MacKinnon, M., Reed, W., Narielvala, F.M., Hetzel, D.J., Solcia, E., Shearman, D.J.C., Omeprazole V ranitidine for prevention of relapse in reflux oesophagitis. A controlled double blind trial of their efficacy and safety, Gut, 35 (1994) 590-598.
2. Dent, J., Diseases of the oesophagus, Oxford Textbook of Medicine, (Eds DJ Weatherall, JGG Ledingham, DA Warrell), (1996) (Oxford)1865.
3. Sontag, S.J., Rolling review: gastro-oesophageal reflux disease., Aliment Pharmacol Ther, 7 (1993) 293-312.
4. Schnell, T., Sontag, S., Wanner, J., et al, Endoscopic screening for Barrett's oesophagitis, oesophageal adenocarcinoma and other mucosal changes in ambulatory subjects with symptomatic gastroesophageal reflux. Gastroenterol, 1985, (88) (1576)
5. Howden C.W., Hornung, C.A., A systematic review of ther associatioon between Barrett's oesophagitis and colon neoplasms., Am J Gastroenterol, 90 (1995) 1814-1819.
6. Moore, R.A., Helicobacter pylori and peptic ulcer. A systematic review of effectiveness and an overview of the economic benefits of implementing that which is known to be effective, Health Technology Evaluation Association, Oxford, (1995)
7. Norrelund, N., Pedersen, P.A., Prevalence of gastro-oesophageal relux-like dyspepsia., International Congress of Gastroenterology, Rome., (1988) A1031.
8. Wienbeck, M., Barnert, J., Epidemioology of reflux disease and reflux esophagitis., sacnd J Gastroenterol, 24 (Suppl 156) (1989) 7-13.
9. Sonnenburg A., Epidemiologie and Spontanverlauf der refluxkrankheit., In: Refluxtherapie, Gastroesophageale Refluxkrankheit: Konservative and Operative Therapie (Blum AL, Siewert JR (eds): Berlin, Springer Verlag, (1981) 85-106.
10. Johansson, K.E., Ask, P., Boeryd, B., et al, Oesophagitis, signs of reflux and gastric acid secretion in patients with symptoms of gastrooesophageal reflux disease., Scand J Gastroenterol, 21 (1986) 837-847.
11. Behar, J., Biancani, P., Sherahan, D.G., Evaluation of oesophageal tests in the diagnosis of reflux oesophagistis., Gastroenterology, 71 (1976) 9-15.
12. Gear M.W.L., Wilkinson, S.P., Open access upper alimentary endoscopy., Br J Hosp Med, 41 (1989) 438-444.
13. Stokes, D., et al, Reflux oesophagitis., Gastroenterol Int, Suppl 1 (1989) 831.
14. Klinkenberg-Knol, E.C., Festen, H.P.M., Meuwissen, S.G.M., Pharmacological managemant of gastro-oesophageal reflux disease., Drugs, 49 (1995) 695-710.
15. The medical management of gastro-oesophageal reflux., Drug & Therapeutics Bulletin, 34 (1996) 1.1-1.3.
16. Bate, C.M., Reflux oesophagitis, Martin Dunitz, (1992)
17. L'Abbé, K.A., Detsky, A.S., O'Rourke, K., Meta-analysis in clinical research., Ann Intern Med, 107 (1987) 224-233.
18. Jadad, A.R., Moore, R.A., Carroll, D., Jenkinson, C., Reynolds, D.J.M., Gavaghan, D.J., Carroll, D.C., McQuay, H.J., Assessing the quality of reports of randomized clinical trials: is blinding necessary?, Contr Clin Trials, 17 (1996) 1-12.
19. DerSimonian, R., Laird, N., Meta-analysis in clinical trials., Control Clin Trial, 7 (1986) 177-188.
20. Cook, R.J., Sackett, D.L., The number needed to treat: a clinically useful measure of treatment effect., Br. Med. J., 310 (1995) 452-454.
21. Hungin, A.P.S., Gunn, S.D., Bate, C.M., Turbitt, M.L., Wilcock, C., Richardson, P.D.I., A comparison of the efficacy of omeprazole 20 mg once daily with ranitidine 150 mg bd in the relief of symptomatic gastro-oesophageal reflux disease in general practice, British Journal of Clinical Research, 4 (1993) 73-88.
22. Klinkenberg-Knol, E.C., Jansen, J.M.B. J., Festen, H.P.M., Meuwissen, S.G.M., Lamers, C.B.H. W., Dopuble-blind multicentre comparison of omeprazole and ranitidine in the treatment of reflux oesophagitis, Lancet, 1 (1987) 349-351.
23. Sandmark, S., Carlsson, R., Fausa, O., Lundell, L., Omeprazole or ranitidine in the treatment of reflux esophagitis, Scand J Gastroenterol, 23 (1988) 625-632.
24. Havelund, T., Laursen, L.S., Skoubo-Kristensen, E., Andersen, B.N., Pedersen, S.A., Jensen, K.B., Fenger, C., Hanberg-Sorensen, F., Lauritsen, K., Omeprazole and ranitidine in treatment of reflux oesophagitis: Double blind comparitive trial, BMJ, 296 (1988) 89-92.
25. Vantrappen, G., Rutgeerts, L., Schurmans, P., Coenegrachts, J-L, Omeprazole (40mg) is superior to ranitidine in short-term treatment of ulcerative reflux esophagitis, Dig Dis Sci, 33 (1988) 523-529.
26. Zeitoun, P., Rampal, P., Barbier, P., Isal, J-P, Eriksson, S., Carlsson, R., Omeprazole (20 mg o.m.) versus ranitidine (150 mg b.i.d.) in reflux esophagitis. Results of a double-blind randomized trial, Gastroenterol Clin Biol, 13 (1989) 457-462.
27. Bate, C.M., Keeling, P.W.N., O'Morain, C., Wilkinson, S.P., Foster, D.N., Mountford, R.A., Temperley, J.M., Harvey, R.F., Thompson, D.G., Davis, M., Forgacs, I.C., Bassatt, K.S., Richardson, P.D.I., Comparison of omeprazole and cimetidine in reflux oesophagitis: symptomatic, endoscopic, and histological evaluations, Gut, 31 (1990) 968-972.
28. Dehn, T.C.B., Shepherd, H.A., Colin-Jones, D., Kettlewell, M.G.W., Carroll, N.J.H., Dopuble blind comparison of omeprazole (40 mg od) versus cimitidine (400 mg qd) in the treatment of symptomatic erosive reflux oesophagitis, assessed endoscopically, histologially and by 24 h pH monitoring, Gut, 31 (1990) 509-513.
29. Lundell, L., Backman, L., Ekstrom, P., Enander, L-H, Fausa, O., Lind, T., Lonroth, H., Sandmark, S., Sanzen, B., Unge, P., Westin, I.H., Omeprazole or high-dose ranitidine in the treatment of patients with reflux oesophagitis not responding to "standard doses" of H2-receptor antagonists, Aliment Pharmacol Ther, 4 (1990) 145-155.
30. Bate, C.M., Crowe, J.P., Dickinson, R.J., Green, J.R.B., Banerjee, S., Dronfiels, M.W., McFarlane, R.J., Parry-Billings, K.S., Richardson, P.D.I., Reflux oesophagitis resolves more rapidly with omeprazole 20 mg once daily than with ranitidine 150 mg twice daily: omeprazole 40 mg once daily provides further benefit in unresponsive patients, Br J Clin Res, 2 (1991) 133-148.
31. Frame, M.H., Italian Reflux Oesophagitis Study Group, Omeprazole produces significantly greater healing of erosive or ulcerative reflux oesophagitis than ranitidine, Eur J gastroenterol Hepatol, 3 (1991) 511-517.
32. Cloud, M.L., Offen, W.W., Nizatidine versus placebo in gastroesophageal reflux disease. A six-week, multicenter, randomized, double-blind comparison. Nizatidin gastroesophageal reflux disease study group, Dig Dis Sci, 37 (1992) 865-874.
33. Bianchi Porro, G., Pace, F., Peracchia, A., Bonavina, L., Vigneri, S., Scialabba, A., Franceschi, M., Short-term treatment of refractory reflux esophagitis with different doses of omeprazole or ranitidine, J Clin Gastroenterol, 15 (1992) 192-198.
34. Roufail, W., Belsito, A., Robinson, M., Barish, C., Rubin, A., Ranitidine for erosive oesophagitis: a double-blind, placebo-controlled study., Aliment Pharmacol Ther, 6 (1992) 597-607.
35. Sontag, S.J., Hirschowitz, B.I., Holt, S., Robinson, M.G., Behar, J., Berenson, M.M., Mccullough, A., Ippoliti, A.F., Richter, J.E., Ahtaridis, G., McCallum, R.W., Pambianco, D.J., Vlahcevic, Z., Johnson, D.A., Collen, M.J., Lyon, D.T., Humphries, T.J., Cagiola, A., Berman, R.S., Two doses of omeprazole versus placebo an symptomatic erosive esophagitis: The U. S. Multicenter Study, Gastroenterology, 102 (1992) 109-118.
36. Hatlebakk, J.G., Berstad, A., Carling, L., Svedberg, L.-E, Unge, P., Ekstrom, P., Halvorsen, L., Stallemo, A., Hovdenak, N., Trondstad, R., Lansoprazole versus omeprazole in short-term treatment of reflux oesophagitis. Results of a Scandinavian multicentre trial, Scand J Gastroenterol, 28 (1993) 224-228.
37. Robinson, M., Decktor, D.L., Maton, P.N., Sabesin, S., Roufail, W., Kogut, D., Roberts, W., McCullough, A., Pardoll, P., Saco, L., et al, Omeprazole is superior to ranitidine plus metoclopramide in the short-term treatment of erosive oesophagitis, Ailment Pharmacol Ther, 7 (1993) 67-73.
38. Bate, C.M., Booth, S.N., Crowe, J.P., Hepworth-Jones, B., Taylor, M.D., Richardson, P.D.I., Does 40 mg omeprazole daily offer additional benefit over 20 mg daily in patients requiring more than 4 weeks of treatment for symptomatic reflux oesophagitis?, Aliment Pharmacol Ther, 7 (1993) 501-507.
39. Marks, R.D., Richter, J.E., Rizzo, J., Koehler, R.E., Spenney, J.G., Mills, T.P., Champion, G., Omeprazole versus H2-receptor antagonists in treating patients with peptic stricture and esophagitis, Gastroenterology, 106 (1994) 907-915.
40. Bardham, K.D., Hawkey, C.J., Long, R.G., Morgan, A.G., Wormsley, K.G., Moules, I.K., Brocklebank, D., Lansoprazole versus ranitidine for the treatment of reflux oesophagitis. UK lansoprazole clinical research group, Aliment Pharmacol Ther, 9 (1995) 145-151.
41. Robinson, M., Sahba, B., Avner, D., Jhala, N., Greski-Rose, P.A., Jennings, D.E., A comparison of lansoprazole and ranitidine in the treatment of erosive oesophagitis. Multicentre Investigational Group, Aliment Pharmacol Ther, 9 (1995) 25-31.
42. Robinson, M., Cambell, D.R., Sontag, S., Sabesin, S.M., Treatment of erosive reflux esophagitis resistant to H2-receptor antagonist therapy, Dig Dis Sci, 40 (1995) 590-597.
43. Green, J.R.B., Tildesley, G., Theodossi, A.Bate, C. M., Bradby, G.V.H., Axon, A.T.R., Copeman, M.B., Taylor, M.D., Omeprazole 20 mg to 40 mg once daily is more effective than ranitidine 300 mg to 600 mg daily in providing complete symptom relief and endoscopic healing in patients with reflux oesophagitis, Br J Clin Res, 6 (1995) 63-76.
44. Lundell, L., Backman, L., Ekstrom, P., et al, Prevention of relapse of reflux oesophagitis after endoscopic healing: the efficacy and safety of omeprazole compared with ranitidine, Scand J Gastroenterol, 26 (1991) 248-256.
45. Smith, P.M., Kerr, G.D., Cockerel, R., Ross, B.A., Bate, C.M., Brown, P., Dronfiels, M.W., Green, J.R.B., Hislop, W.S., Theodossi, A., McFarland, R.J., Watts, D.A., Taylor, M.D., Richardson, P.D.I., A comparison of omeprazole and ranitidine in the prevention of recurrence of benign esophageal stricture, Gastroenterology, 107 (1994) 1312-1318.
46. Hallerback, B., Unge, P., Carling, L., Edwin, B., Glise, H., Havi, N., Lyrenas, E., Lundberg, K., Omeprazole or ranitidine in long-term treatment of reflux esophagitis, Gastroenterology, 107 (1994) 1305-1311.
47. Bate, C.M., Booth, S.N., Crowe, J.P., Mountford, R.A., Keeling, P.W.N., Hepworth-Jones, B., Taylor, M.D., Richardson, P.D.I., and the Solo Investigator Group, Omeprazole 10 mg or 20 mg once daily in the prevention of reflux oesophagitis, Gut, 36 (1995) (4)492-498.
48. Vigneri, S., Termini, R., Leandro, G., Badalmenti, S., Pantalena, M., Savarino, V., Di Mario, F., Battaglia, G., Mela, G.S., Pilotto, A., Plebani, M., Davi, G., A comparison of five maintenance therapies for reflux esophagitis, N Engl J Med, (1995) 1106-1110.
49. Robinson, M., Lanza, F., Avner, D., Haber, M., Effective maintenance treatment of reflux esophagitis with low-dose lanzoprazole, Annals of Internal Medicine, 124 (1996) 859-867.
50. Khan, K.S., Daya, S., Jadad, A.R., The importance of quality of primary studies in producing unbiased systematic reviews., Arch. Int. Med., 156 (1996) 661-666.
51. Moore, R.A., Collins, S., Carroll, D., McQuay, H.J., Efficacy of oral paracetamol in acute pain states., Pain, in press (1997)
Table 1: Short-term healing

R eference
Number/ Condition
Design/treatment
Outcomes
Results
Adverse events/ withdrawals
[22]

QS: 1+2+1=4
51 outpatients with erosive gastritis grade 2 or 3

Parallel group with symptoms at 2, 4 and 8 weeks. Endoscopy at 4 and 8 weeks
25 O 60 mg
26 R 300 mg
Healing defined as change to grade 0 or 1.
At 4 weeks:
O 60 19/25
R 300 7/26
At 8 weeks:
O 60 22/25
R 300 10/26
AE withdrawal
O 60 1/25
R 300 1/26
[23]

QS: 2+2+1=5
152 patients with erosive or ulcerative oesophagitis grade 2-4
Parallel group with endoscopy at 4 and 8 weeks
75 O 20 mg
77 R 300 mg
Macroscopic healing of lesions.
At 4 weeks:
O 20 46/75
R 300 23/77
At 8 weeks:
O 20 56/75
R 300 33/77
AE withdrawal
O 20 2/75
R 300 2/77
[24]

QS: 2+2+1=5
162 patientswith endoscopically proven reflux oesophagitis grade 1-3
Parallel group with endoscopy at 4, 8 and 12 weeks
80 O 20 mg
82 R 300 mg
Endoscopic healing (grade 0)
At 4 weeks:
O 20 59/80
R 300 30/82
At 8 weeks:
O 20 69/80
R 300 49/82
AE withdrawals:
O 20 3/80
R 300 1/82
[25]

QS: 1+2+1=4
61 patients with ulcerative reflux oesophagitis grade 2 or 3
Parallel group with endoscopy at 4 and 8 weeks
31 O 40 mg
30 R 300 mg
Endoscopic healing
At 4 weeks:
O 40 23/31
R 300 13/30
At 8 weeks:
O 40 25/31
R 300 15/30
AE withdrawals:
O 40 1/31
R 300 0/30
[26]

QS: 2+2+1=5
156 patients with refux oesophagitis grade 2-4
Parallel group with endoscopy at 4 and 8 weeks
76 O 20 mg
80 R 300 mg
Endoscopic healing grade 0 or 1
At 4 weeks:
O 20 56/76
R 300 33/80
At 8 weeks:
O 20 66/76
R 300 45/80
AE withdrawals:
O 20 1/76
R 300 2/80
[27]

QS: 1+2+1=4
272 patients with reflux oesophagitis grade 1-4
Parallel group with endoscopy at 4 and 8 weeks
138 O 20 mg
134 C 1600 mg
Endoscopic healing grade 0
At 4 weeks:
O 20 77/138
C 1600 34/134
At 8 weeks:
O 20 97/138
C 1600 46/134
AE withdrawals:
O 20 9/138
C 1600 11/134
[28]

QS: 1+2+1=4
67 outpatients with reflux oesophagitis grade 1-4
Parallel group with endoscopy at 4 and 8 weeks
31 O 20 mg
36 C 1600 mg
Endoscopic healing grade 0
At 4 weeks:
O 20 16/31
C 1600 9/36
At 8 weeks:
O 20 20/31
C 1600 7/36
AE withdrawals:
O 20 1/31
C 1600 3/36
[29]

QS: 2+2+0=4
98 patients with erosive/ulcerative oesophagitis despite cimetidine or ranitidine treatment at least grade 2
Parallel group with endoscopy at 4, 8 and 12 weeks
51 O 40 mg
47 R 600 mg
Endoscopic healing grade 0
At 4 weeks:
O 40 32/51
R 600 8/47
At 8 weeks:
O 40 44/51
R 600 18/47
Minor adverse events similar in both groups
[30]

QS: 1+2+1=4
283 patients endoscopically diagnosed reflux oesophagitis grade 1-4
Parallel group with endoscopy at 4 and 8 weeks
143 O 20 mg
140 R 300 mg
Endoscopic healing grade 0
At 4 weeks:
O 20 81/143
R 300 42/140
At 8 weeks:
O 20 98/143
R 300 53/140
AE withdrawals:
O 20 5/143
R 300 14/140
[31]

QS: 1+2+1=4
172 patients with reflux oesophagitis grade 2 or 3
Parallel group with endoscopy at 4 and 8 weeks
86 O 20 mg
86 R 300 mg
Endoscopic healing grade 0 or 1
At 4 weeks:
O 20 44/86
R 300 37/86
At 8 weeks:
O 20 63/86
R 300 53/86
AE withdrawals:
O 20 2/86
R 300 0/86
[32]

QS: 2+2+1=5
515 patients with gastroesophageal reflux grade 1-4
Parallel group with endoscopy at 3 and 6 weeks
168 N 300 mg
169 N 600 mg
178 placebo
Endoscopic healing grade 0
At 3 weeks:
N 300 29/168
N 600 33/169
placebo 21/178
At 6 weeks:
N 300 65/168
N 600 70/169
placebo 46/178
AE withdrawals:
N 300 8/168
N 600 5/169
placebo 10/178
[33]

QS: 1+2+1=4
60 patients with erosive/ulcerative reflux oesophagitis grade 2-4
Parallel group with endoscopy at 4 and 8 weeks
30 O 20 mg
30 R 300 mg
Endoscopic healing grade 0
At 4 weeks:
O 20 15/30
R 300 6/30
At 8 weeks:
O 20 23/29
R 300 10/30
AE withdrawals:
O 20 0/30
R 300 0/30
[34]

QS: 1+1+1=3
342 patients with erosive oesophagitis grade 2-4
Parallel group with endoscopy at 4, 8 and 12 weeks
109 R 600 mg
120 R 1200 mg
113 placebo
Endoscopic healing grade 0 or 1
At 4 weeks:
R 600 47/109
R 1200 50/120
placebo 21/113
At 8 weeks:
R 600 68/109
R 1200 73/120
placebo 30/113
AE withdrawals:
R 600 3/109
R 1200 3/120
placebo1/113
[35]

QS: 1+2+1=4
230 patients with reflux oesophagitis grade 2 or above
Parallel group with endoscopy at 4 and 8 weeks
93 O 20 mg
91 O 40 mg
46 placebo
Endoscopic healing grade 0 or 1
At 4 weeks:
O 20 36/93
O 40 40/91
placebo 3/46
At 8 weeks:
O 20 68/93
O 40 66/91
placebo 7/46
AE withdrawals:
O 20 4/93
O 40 5/91
placebo 2/46
[21]

QS: 1+0+1=2
424 patients with gastroesophageal reflux in 111 general practices, endoscopically confirmed within 2 years
Parallel group, open study for four weeks
214 O 20 mg
208 R 300 mg
Symptom relief at 4 weeks
Heartburn
O 20 129/214
R 300 80/208
AEwithdrawals:
O 20 2/214
R 300 8/208
[36]

QS: 1+2+1=4
229 patients with reflux oesophagitis grade 1 and 2 (75% grade 1)
Parallel group with endoscopy at 4 and 8 weeks
116 L 30 mg
113 O 20 mg
Endoscopic healing
At 4 weeks:
L 30 71/116
O 20 73/113
At 8 weeks:
L 30 95/116
O 20 96/113
AEwithdrawals:
L 30 0/116
O 20 1/113
[37]

QS: 1+0+1=2
184 patients with reflux oesophagitis grade 2-4
Parallel group, open, with endoscopy at 4 and 8 weeks
92 O 20 mg
92 R 300 mg plus 40 mg metoclopramide
Endoscopic healing grade 0 or 1
At 4 weeks:
O 20 63/92
R 300+M 28/92
At 8 weeks:
O 20 75/92
R 300+M 42/92
AEwithdrawals:
O 20 0/92
R 300+M 10/92
[38]

QS: 1+1+1=3
313 patients with reflux oesophagitis grade 2-4
Parallel group, open 4 weeks blind 4 weeks with endoscopy at 4 and 8 weeks
158 O 20/20 mg
155 O 20/40 mg
Endoscopic healing grade 0
At 4 weeks:
O 20 154/313
At 8 weeks:
O 20/20 103/158
O 20/40 114/155
AEwithdrawals:
At 4 weeks:
O 20 7/313
At 8 weeks:
O 20/20 0/158
O 20/40 0/155
[39]
QS: 1+0+1=2
34 dysphagic patients with peptic stricture and erosive oesophagitis
Parallel group with endoscopy at 3 and 6 weeks
18 O 20 mg
18 R 300 mg or F 40 mg
Endoscopic healing grade 0
At 3 weeks:
O 20 11/18
R/F 300/40 8/18
At 6 weeks:
O 20 18/18
R/F 300/40 10/18
AEwithdrawals:
O 20 0/18
R/F 300/40 0/18
[40]

QS: 2+2+1=5
229 patients with reflux oesophagitis grade 1-3
Parallel group with endoscopy at 4 and 8 weeks
77 L 30 mg
75 L 60 mg
77 R 300 mg
Endoscopic healing grade 0
At 4 weeks:
L 30 63/77
L 60 49/75
R 300 29/77
At 8 weeks:
L 30 70/77
L 60 59/75
R 300 37/77
AEwithdrawals:
L 30 0/77
L 60 1/75
R 300 0/77
[41]

QS: 1+2+1=4
242 patients with reflux oesophagitis grade 2 or more
Parallel group with endoscopy at 4 and 8 weeks
115 L 30 mg
127 R 300 mg
Endoscopic healing grade 0 or 1
At 4 weeks:
L 30 94/115
R 300 64/127
At 8 weeks:
L 30 105/115
R 300 86/127
AEwithdrawals:
L 30 7/115
R 300 5/127
[42]

QS: 1+1+1=3
54 patients with therapy resitant erosive reflux oesophagitis grade 1-4
Parallel group with endoscopy at 2, 4, 6, 8 and 12 weeks
23 L 30 mg
27 R 300 mg
Endoscopic healing grade 0
At 4 weeks:
L 30 20/23
L 60 20/27
At 8 weeks:
L 30 22/23
L 60 24/27
AEwithdrawals:
L 30 0/23
L 60 0/27
[43]

QS: 1+0+1=2
198 patients with reflux oesophagitis grade 1-4
Parallel group with endoscopy at 4 and 8 weeks
99 O 20 mg to 40 mg @ 4 weeks
97 R 300 mg to 600 mg @ 4 weeks
Endoscopic healing grade 0 plus no symptoms for 7 days
At 4 weeks:
O 20 35/99
R 300 7/97
At 8 weeks:
O 20 54/99
R 300 24/97
AEwithdrawals:
O 20 3/99
R 300 11/97

Table 2: Long-term maintenance

R eference
Number/ Condition
Design/treatment
Outcomes
Results
Adverse events/ withdrawals
[41]

QS: 2+2+1=5
98 patients with grade 2 oesophagitis in healing phase.


12 week healing stage, parallel group:
51 patients O40 followed by 020 maintenance 1 year in 46.
47 patients R600 followed by R300 maintenance 1 year in 22.
Endoscopy 3/12 Gross apprearance & biopsy
Symptom evaluation
Intention to treat analysis.
Remission at 12/12:
O 23/46
R 2/22
Endoscopic relapse:
O 11/46
R 14/22
Withdrawn:
O 2/46
R 2/22
Minor adverse events, same frequency
[42]

QS: 1+1+1=3
366 patients with benign oesophageal stricture secondary to acid reflux.
1 year parallel group:
180 O20
185 R300
after endoscopic dilatation to 12-18 mm diameter
Primary end point was complete relief of symptoms - absence of symptoms, normal diet, no use of antacids.
3/13 visits
Free of reflux symptoms at 12/12:
O 115/139
R 84/122
Redilation:
O 43/143
R 66/143
Stricture at 12/12:
O 56/137
R 73/121
Normal mucosa at 12/12:
O 83/136
R 26/121
260 AE in 115 of 181 on O,
255 AE in 115 of 185 on R.

AE withdrawals 24 O, 29 R
[1]

QS: 2+1+1=4
204 patients with erosive reflux oesophagitsis for 1 year after a healing phase.
1 year parallel group: 159 randomised to
53 O20 daily
55 O20 3 days each week
51 R300 daily
3/12 visits for endoscopy, symptoms.
Relapse defined as endoscopically verified oesophagitis of at least grade 2.
Remission at 12/12
O20 daily 47/53
O20 3 days each week 18/55
R300 daily 13/51
AE Withdrawal:
O20 daily 0/53
O20 3 days/week 3/55
R300 1/51
[46]

QS: 1+2+1=4
426 patients with reflux oesophagitis grade 2 or more healed with Omeprazole.
392 entered maintenance study
1 year parallel group: 392 randomised to
131 O20 (single dose)
133 O20 (split dose)
128 R300
3/12 visits.
symptomatic relapse plus endoscopy.
endoscopy at 12/12
Endoscopic remission at 12/12
O20 (single dose) 94/131
O20 (split dose) 82/133
R300 58/128
AE Withdrawal:
O20 (single dose) 2/131
O20 (split dose) 6/133
R300 1/128
[47]

QS: 1+2+1=4
193 patients with healed reflux oesophagitis and asymptomatic
1 year parallel group: 193 randomised to
60 O10
68 O20
62 placebo
3/12 clinic visits for symptoms
endoscopic relapse recurrence grade 2 or more
symptomatic relapse, moderate or severe symptoms
Endoscopic remission at 12/12
O10 30/60
O20 50/68
placebo 9/62
Symptomatic remission at 12/12
O10 46/60
O20 56/68
placebo 21/62
AEs
33 in 19 of 61 with O10
42 in 25 of 69 with O20
16 in 13 of 63 with placebo
AE withdrawals:
O10 2/61
O20 4/69
placebo 3/63
[48]

QS: 2+0+1=3
175 patients with endoscopically confirmed reflux oesphagitis healed with Omeprazole
1 year parallel group: 175 randomised (35/group) to
Cisapride 30
R 450
O20
R + Cisapride
O + Cisapride
endoscopy at 6 and 12 months
Endoscopic healiong and symptoms
Remission - absence of symtoms or endoscopic signs
Remission at 12/12:
Cisapride 30 19/35
R 450 17/35
O20 28/35
R + Cisapride 23/35
O + Cisapride 31/35
AE withdrawals:
Cisapride 30 2/35
R 450 2/35
O20 0/35
R + Cisapride 1/35
O + Cisapride 0/35
[49]

QS: 1+2+1=4
173 patients with documented healing of erosive oesophagitis with Lanzoprazole
1 year parallel group: 173 randomised to
60 L15
56 L30
57 placebo
3/12 visits
Recurrence of erosive oesophagitis (grade 2 or higher), either symptomatic or asymptomatic
Remission at 12/12:
L15 47/60
L30 50/56
placebo 14/57
Asymptomatic at 12/12:
L15 43/60
L30 38/56
placebo 20/57
AE withdrawals:
L15 1/60
L30 3/56
placebo 2/57



FIGURE 1: NNT CALCULATION




where:

IMP act = number of patients given active treatment achieving the target
TOT act = total number of patients given the active treatment
IMP con = number of patients given a control treatment achieving the target
TOT con = total number of patients given the control treatment

FIGURE 2:



FIGURE 3:



FIGURE 4:



FIGURE 5: