Skip navigation
Link to Back issues listing | Back Issue Listing with content Index | Subject Index

PPIs for reflux oesophagitis

Systematic review
Results
Comment

When estimating relative efficacy of different treatments in meta-analyses of randomised trials, the usual situation is that we have many comparisons with, typically, placebo, but few direct comparisons between treatments. As a consequence of this we resort to indirect efficacy. It is a bit like testing every athlete for how long it takes them to run 100 metres individually (indirect comparison = world record) as opposed to who is fastest in a single race (direct comparison = Olympic champion).

It is unusual to have a feast of large, good quality, direct comparisons, but that is the situation in a meta-analysis of proton pump inhibitors (PPI) for healing of reflux oesophagitis [1]. This sort of data can help us generate information on relative efficacy in order to help formulate cost-effective strategies.

Systematic review


The systematic review built on an earlier one, with wide searching up to early 2005 for randomised trials comparing PPIs with esomeprazole. Trials chosen were those of European licensed standard doses of a PPI with esomeprazole 40 mg.

The outcome of interest was endoscopic healing data at four and eight weeks, in patients with comparable grades of oesophagitis (Los Angeles A-D or equivalent). Where necessary data from trials was recalculated with the number of patients randomised, to ensure a consistent intention to treat approach.

Results


Eight trials were identified, with 14,800 patients. Of these about 7,400 used esomeprazole 40 mg, 3,300 lanzoprazole 30 mg, 2,400 omeprazole 20 mg, and 1,700 pantoprazole 40 mg. No trials were identified with rabeprazole. Trials generally examined patients with grades A-D oesophagitis, though two limited patients to grades B and C or C and D.

The main results calculated from data in the paper are shown in Table 1. Esomeprazole 40 mg was significantly better than other PPIs used in these trials, with higher healing rates at four and eight weeks (Figure 1).


Table 1: Comparison of esomeprazole 40 mg daily with other PPIs in endoscopic healing of reflux oesophagitis after four and eight weeks of treatment



Number of
Percent healed with
Comparator
Trials
Patients
Esomeprazole
40 mg
Comparator
Relative benefit
(95% CI)
NNT
(95% CI)
At 4 weeks
Lanzoprazole 30 mg
3
6526
73
68
1.07 (1.04 to 1.10)
22 (15 to 42)
Omeprazole 20 mg
3
4877
74
65
1.14 (1.10 to 1.18)
11 (8.6 to 15)
Pantoprazole 40 mg
2
3397
77
71
1.09 (1.04 to 1.13)
16 (11 to 32)
At 8 weeks
Lanzoprazole 30 mg
3
6526
86
83
1.04 (1.02 to 1.06)
30 (20 to 65)
Omeprazole 20 mg
3
4877
89
82
1.08 (1.06 to 1.11)
16 (12 to 23)
Pantoprazole 40 mg
2
3397
90
88
1.02 (1.00 to 1.04)
49 (24 to infinity)




Figure 1: Percentage of patients with endoscopic healing of reflux oesophagitis for four common PPI doses






Analysis by baseline Los Angeles classification showed that, at eight and four weeks, healing rates tended to be lower at higher initial grade. Thus four week healing rates for esomeprazole 40 mg ranged from about 82% for grade A to about 50% for grade D. Eight week healing rates for esomeprazole 40 mg ranged from about 92% for grade A to about 77% for grade D. Similar but lower results were reported for the other PPIs combined.

Comment


The first thing to bear in mind is that two of the three authors of the meta-analysis were employees of the manufacturers of esomeprazole. That is not necessarily a bad thing, but the thrust of the analysis, with esomeprazole 40 mg as the common comparator which had to be in an included trial, would tend to exclude other trials and limit the evidence we have to look at.

A different approach, which might be interesting, would be to compare relative efficacy using placebo, and using esomeprazole or other common comparators to see if they give the same order of efficacy. Such an approach might also include non-standard or non-licensed doses, further broadening the available evidence if there were sufficiently large amounts of data in properly conducted trials with the same outcomes and conducted in patients with similar initial disease severity. A case for an extended systematic review, probably.

A second observation from looking at the individual trials is how consistent the results were. Figure 2 shows the eight week healing rates in the esomeprazole arms of the eight trials. With high event rates and large numbers of patients, the result of each trial is close to the overall average of 88%. This is quite unlike the situation of small numbers and low event rates.


Figure 2: Eight week healing rates for esomeprazole 40 mg in individual trials






A third moment for reflection is for the economic consequences of small differences between healing rates. The immediate thought on costs would be to leap to the lowest acquisition cost, in this case generic omeprazole 20 mg, at about £13 for four weeks treatment, rather than somewhat more effective, but expensive, branded PPIs that cost up to twice as much.

It all depends on the cost of someone not healed. As that increases, the economics change, so a good health economic analysis would help in decision-making.

Reference:

  1. 1 SJ Edwards et al. Systematic review: proton pump inhibitors (PPIs) for the healing of reflux oesophagitis – a comparisons of esomeprazole and other PPIs. Alimentary Pharmacology & Therapeutics 2006 24: 743-750.

previous or next story