Skip navigation

Treatment effectiveness and costs in reflux disease

Intermittent treatment
Costs and budgets

Gastro-oesophageal reflux is a common disorder, but only a small proportion of people actually seek help from their doctors. For those who do, there is a bewildering array of possibilities, from advice about lifestyles to endoscopy, depending on age, symptoms, and just where you live. One of the questions Bandolier is often asked is about the relative efficacy of different treatments, particularly histamine antagonists (H2As) and proton pump inhibitors (PPIs). There is one systematic reviews from McMaster [1], which gives some answers, but the date of the last search was 1996.


MEDLINE and manual journal searches were made for English language studies which were randomised single or double blind studies of treatment for gastro-oesophageal reflux. Patients had to have endoscopic grade II to IV oesophagitis, the endpoint was endoscopic healing, and patients had to be adults. The number of healed patients was extracted for up to 12 weeks of treatment.


The analysis in the paper concentrated on the speed of healing. For this analysis, Bandolier extracted information at four and eight weeks for any dose of drug where there were at least two studies and at least 100 patients. What this gives us is the proportion of people healed for each treatment at four and eight weeks (Table 1; Figures 1 and 2). Information was also available on complete relief of heartburn at baseline, and after four and eight weeks (Table 2).

Table 1: Endoscopic healing at 4 and 8 weeks

Endoscopic healing of Grade II-IV oesophagitis with placebo, histamine antagonists and proton pump inhibitors
  4 weeks 8 weeks
Drug / daily dose Healed/Total 95% CI of percent healed Healed/Total 95% CI of percent healed
Placebo 105/629 14 - 20 147/467 27 - 36
Cimetidine 1600 mg 34/147 16 - 30 54/167 25 - 39
Ranitidine 300 mg 258/668 35 - 42 451/831 51 - 58
Ranitidine 600 mg 268/754 32 - 39 422/717 55 - 63
Ranitidine 1200 mg 139/277 44 - 56 184/271 62 - 74
Omeprazole 20 mg 623/1033 57 - 63 171/197 82 - 92
Omeprazole 40 mg 208/320 60 - 70 262/319 78 - 86
Pantoprazole 40 mg 310/422 70 - 78 372/422 85 - 91
Lansoprazole 30 mg 209/255 77 - 87 231/249 90 - 95
From randomised, blind studies, with at least two studies or 100 patients [from Chiba et al, 1997]

Figure 1: Endoscopic healing at 4 weeks. Bars are 95% CI.

Figure 2: Endoscopic healing at 8 weeks. Bars are 95% CI.

Table 2: Symptomatic relief at 4 and 8 weeks

Complete relief of heartburn with placebo, histamine antagonists and proton pump inhibitors
  Baseline 4 weeks 8 weeks
Drug / daily dose Heartburn free /Total 95% CI of % heartburn free Heartburn free /Total 95% CI of % heartburn free Heartburn free /Total 95% CI of % heartburn free
Ranitidine 300 mg 7/305 1 - 4 120/283 37 - 48    
Omeprazole 20 mg 33/860 3 - 5 621/846 70 - 76 323/416 74 - 82
Omeprazole 40 mg 20/279 4 - 10 220/274 76 - 85 162/188 81 - 91
Pantoprazole 40 mg 5/435 0 - 2 362/413 85 - 91    
From randomised, blind studies, with at least two studies or 100 patients [from Chiba et al, 1997]

What the results of the review show is:

  1. Few people get better without treatment. Fewer than 20% are healed with placebo at four weeks, and that increases to about 30% at eight weeks.
  2. Higher healing rates are found after eight weeks than after four weeks. Duration of treatment matters.
  3. Standard doses of histamine antagonists are much less effective than standard doses of proton pump inhibitors. By eight weeks proton pump inhibitors should heal 82-95%: histamine antagonists might heal 25-58%.
  4. Assessment of efficacy by heartburn relief gives similar results to those for endoscopic healing (Table 2).

Symptoms, and especially heartburn, are the critical issue in primary care, and this can be informed by a recent trial looking at treatment in this setting [2]. This randomised double blind study recruited patients with heartburn in primary care. About half were women, the mean age was about 49 years, the mean weight 77 kg, about 40% had hiatus hernia, about half had endoscopic oesophagitis, symptoms were moderate or severe in about 80% of the patients and about half had symptoms every day.

They were randomised between placebo, omeprazole 20 mg daily or cisapride 40 mg daily. After eight weeks, the number with adequate control of heartburn (one day a week or less with no more than mild heartburn) was much higher for omeprazole than other treatments (Figure 3).

Figure 3: Adequate control of heartburn at 8 weeks

What is interesting is that the patients included were different from those in the systematic review [1], as those all had endoscopically proven oesophagitis. Yet the 76% rate of symptom control for omeprazole was in the middle of the confidence interval for symptom control in the review (Table 2).

Intermittent treatment

A continual complaint is that it's not the healing, but the repeat prescriptions that lead to the huge rise in prescribing costs. Another recent randomised trial [3] shows that only half of patients who have their symptoms healed need ongoing treatment.

The study recruited patients from primary and secondary care who had moderate or severe heartburn for more than two days in the previous two weeks. About half were women, the mean age was about 48 years, about a quarter smoked, their mean body mass index was 27, three quarters had had symptoms for more than 12 months and a third had a normal oesophagus on endoscopy.

They were randomised between ranitidine 300 mg a day, or omeprazole 10 mg or 20 mg. After two weeks patients with no symptoms over the previous week entered a one-year follow up. For those who were not symptom free, doses were increased or continued for a further two weeks. After four weeks all patients, whether healed or with mild symptoms entered the follow up period.

Seventy-two percent of patients had no relapse or only a single relapse needing intermittent treatment over the year. Forty-seven percent reached the end of the study using an intermittent treatment and without recourse to maintenance treatment. The only feature of patients that indicated a shorter time to final treatment failure was smoking.

Costs and budgets

Bandolier has been told that evidence-based medicine plays second fiddle to budget-based medicine, and disaggregated budgets at that. Medicine acquisition costs rule, OK! We don't look at costs in health services as a whole.

With reflux disease the argument seems to be between two different approaches. Step-up treatment involves beginning with lifestyle advice (eating, smoking, drinking), raising the bedhead, and then using alginates, perhaps H2As and then maybe an endoscopy before using a PPI. Step-down treatments involves starting with an effective medicine like a PPI, and when the patient is cured advising about smoking cessation, weight reduction, and reflux control with alginates and H2As from the pharmacist.

A study from 1995 [4] suggests that step up is better, because it assessed costs over six months for initial treatment with cisapride, ranitidine and omeprazole from a computer database. The overall six-month costs were (in 1996 £) £136 for cisapride, £177 for ranitidine and £189 for omeprazole per patient.

This was based on information from only 250 patients. Problems were that it was audit of clinical practice during 1995 rather than what is known to be best practice, that patients with more severe disease were excluded, that patient outcomes are not mentioned, nor the quality of the database. Many other studies conclude that step-down treatment is more cost-effective (and at least it is effective). One such [5] is a cost-effectiveness analysis from the randomised trial [3] comparing initial ranitidine and initial omeprazole. There were no significant savings from initial use of ranitidine, which was significantly less effective [3]. The trouble is that few of these health economic analyses are truly independent, which is a problem.

Bandolier was struck by the representations of Figures 1 and 2. They show not only the expected proportion of patients cured, but also those not cured. Patients who are not cured presumably have other things happen to them - more visits to the GP or outpatients, more drugs, more diagnostics, more endoscopy. That all consumes resources. Perhaps we should think about the balance between the simple costs of treatment, usually the acquisition costs of the medicines, and the consequences of not being cured.

So we assumed that the consequences of not being cured for moderate or severe gastro-oesophageal reflux disease was £100. It could easily be this, given the cost of a GP visit as £16, an outpatient visit at £65, an endoscopy at £200 and a course of drugs from about £20-£70 per month. If you do a few sums, it is clear that PPIs are cheaper than H2As and doing nothing, both in cost per patient healed at eight weeks (Figure 4) and in total cost (Figure 5). A new study from Stamford confirms this view of the world [6] demonstrated for the UK previously [7], and concludes that step down treatments are likely to be the most cost-effective when success rates with PPIs are above 59%, which they clearly are at four or eight weeks (Table 1). Following the evidence could save considerable resources in the NHS ( ImpAct July 1999 )

Figure 4: Cost for each patient healed at 8 weeks

Figure 5: Total cost of treatment


Bandolier has learned several lessons from this. Firstly that the world does not stand still, and that more good information keeps coming our way. Decisions made today may be right on what we know. The trouble is what we choose to know then becomes rigid while the truth moves on.

Another lesson is to beware stereotypes. We are told that reflux disease is a lifestyle problem brought about by eating curries, drinking lager, smoking and being overweight. Yet in the two randomised trials mentioned above the patients were predominantly in their 40s or older, few smoked, they weren't obese and had suffered for a long time before seeking help. Not exactly typical lager louts.

Perhaps finally it brings home something important. That budget based medicine is not good medicine. If budget based medicine leads to using less effective treatments, then we are in tricky territory. It's all a matter of balance, of course, and for those who have to manage budgets in isolation, life must be difficult. Bandolier sympathises, and would like to think that a more holistic approach would make it easier for all of us. In the meantime the old adage that the most expensive medicine is the one that doesn't work seems to be provable.


  1. N Chiba et al. Speed of healing and symptom relief in grade II to IV gastroesophageal reflux disease: a meta-analysis. Gastroenterology 1997 112: 1798-1810.
  2. JG Hatlebakk et al. Heartburn treatment in primary care: randomised, double- blind study for 8 weeks. BMJ 1999 319: 550-553.
  3. KD Bardhan et al. Symptomatic gastro-oesophageal reflux disease: double blind controlled study of intermittent treatment with omeprazole or ranitidine. BMJ 1999 318: 502-507.
  4. A Eggleston et al. Cost effectiveness of treatment for gastro-oesophageal reflux disease in clinical practice: a clinical database analysis. Gut 1998 42: 13-16.
  5. NO Stalhammer et al. Cost effectiveness of omeprazole and ranitidine in intermittent treatment of symptomatic gastro-oesophageal reflux disease. Pharmacoeconomics 1999 16: 483-497.
  6. LB Gerson et al. A cost-effectiveness analysis of prescribing strategies in the management of gastroesophageal reflux disease. American Journal of Gastroenterology 2000 95: 395-407.
  7. C Phillips, A Moore. Trial and error - an expensive luxury: Economic analysis of effectiveness of proton pump inhibitors and histamine antagonists in treating reflux disease. British Journal of Medical Economics 1997 11: 55-63.
previous or next story in this issue