Skip navigation

H pylori eradication for dyspepsia

While eradication of Helicobacter pylori now has an established place in treatment of peptic ulcer, whether it makes any difference in nonulcer dyspepsia remains in question. A review, which looked at a number of early studies, concluded that the evidence was thin [1], though the studies included at that time were rather poor in eradicating Helicobacter.

We now have several randomised trials of Helicobacter pylori eradication in non-ulcer dyspepsia. If you read the titles and abstracts, it seems that some of them say it works, while others say it doesn't. As usual, it's a bit more complicated than that. Though it may be early days, with three solid trials published, it may be time to visit them and see what sort of evidence is emerging.


Three randomised trials have been published in recent months [2-4]. The main details of the trials are shown in the Table, all of which randomised patients between treatments that had an active eradication and placebo arm. They all used roughly the same types of patients, with roughly the same severity of disease, used roughly the same intervention and examined roughly the same type of outcome 12 months afterwards.

Randomised trials of H pylori eradication for dyspepsia
  McColl et al, 1998 Blum et al, 1998 Gilvarry et al, 1997
Included patients Referred to dyspepsia clinic by GP. H pylori positive with clinical symptoms of dyspepsia Patients seeking medical care for persistent dyspepsia present for at least 6 months and no history of peptic ulcer disease Patients attending a dyspepsia clinic who were H pylori positive and had symptoms for at least 3 months
Exclusions Previous peptic ulcer disease, endoscopic oesophagitis, NSAIDs, gastric surgery No ulcers, oesophageal or duodenal erosions, or Barret's oesophagus Documented peptic ulcer, NSAIDs, previous upper GI surgery
Age (years) 17 - 70, mean 42 18 - 79, mean 47 18 - 72, mean 39
Sex distribution about 50:50 41% male 28% male
Number of patients 318 328 100
Duration of dyspepsia 61% > 2 years 81% >1 year at least 3 months
Severity of symptoms 11.5 on scale of 0 - 20 3.3 on scale of 0 - 7 14 on a scale of 0 -20
Main outcome Resolution of symptoms (0 or 1 on range of 0 to 20) Resolution of symptoms (score of 0 or 1 on range 0 - 20 in preceeding week) Symptom score (means, no individual patient data)
Treatment Omeprazole plus antibiotics versus omeprazole plus placebo Omeprazole plus antibiotics versus omeprazole plus placebo Bizmuth plus antibiotics versus bizmuth plus placebo
Time of outcome 12 months 12 months 12 months

Two of the studies [3,4] reported as a primary outcome the number of patients who had complete resolution of symptoms one year after treatment, which is a high hurdle of effectiveness.


McColl et al, 1998 [2]

This study concluded that treatment was effective. One year after treatment 33/154 patients (21%) with eradication therapy were symptom free, compared with 11/154 patients (7%) with placebo antibiotics. This was statistically significant - relative benefit 3.0 (1.6 to 5.7), and NNT of 7 (4.6 to 15).

The paper reported a change in symptom scores from a mean of about 11.5 out of a maximum of 20 at baseline to a mean of 5.4 with omeprazole and antibiotics and 6.2 with omeprazole and placebo. There was a corresponding reduction in antiulcer drug prescribing, from 84% to 43 % of patients in the omeprazole and antibiotic group, and from 80% to 53% of patients in the omeprazole and placebo group.

Blum et al, 1998 [3]

This study concluded that treatment was ineffective. One year after treatment 45/164 patients (27%) with eradication therapy were symptom free, compared with 34/164 patients (21%) with placebo antibiotics. The relative benefit was 1.3 (0.9 to 2.0) and the NNT point estimate 15 (6.3 to no benefit).

The mean symptom score fell from a mean of about 3.3 out of 7 to about 1.7 in both groups. Endoscopically judged healing of gastritis was much better with omeprazole and antibiotics than with omeprazole and placebo. Of antibiotic treated patients, 123/164 (75%) were healed at one year compared with 5/164 (3%). The NNT for one year gastritis healing was 1.4 (1.3 to 1.5).

Gilvarry et al, 1997 [4]

No NNTs from this report, as only mean data were reported. In patients in whom Helicobacter was eradicated, mean symptom scores fell from 14.2 to 9.2. This was a significant change, but with placebo there was no significant improvement in symptoms, which were 12.6 at baseline and 10.0 at one year.

What are we to make of this?

Firstly, patients benefited. In the first two trials about a quarter of patients treated with omeprazole and antibiotics were symptom free one year after treatment, compared with 14% with omeprazole alone. In all the trials the mean symptom score fell, whether treated with eradication therapy or no.

There were other benefits as well, though not quantified. Peptic ulcers occurred in four patients in the placebo group and none in the eradication group in one trial [2], six in the placebo group and one in the eradication group in a second [3], and in seven patients who were noted as eradication failures in the third [4]. This is not conclusive, but has biological plausibility when we know that endoscopic gastritis was so effectively treated by the eradication regimen [3].

Finally there is a cost argument. What scanty evidence we have from one study [2] is that consumption of acid suppressing medicine falls. This can be an expensive business, and an economic analysis might show that treatment benefits patients and saves money too.


Does this add up to a sufficient weight of evidence, or should we be concerned about a negative trial, albeit with some positive aspects. Difficult, isn't it? The combined NNT of 9 (6 to 23) for H pylori eradication in nonulcer dyspepsia for complete symptom relief at one year is on the cusp of action.

We probably have enough information to be sure that Helicobacter eradication is effective (statistically). We need more information to be sure how well it works. There are issues here about testing, endoscopy, and patient age. This is something on which we need some sensible guidelines soon.


  1. SJO Veldhuyzen van Zanten, PM Sherman. Indications for treatment of Helicobacter pylori infection: a systematic overview. Canadian Medical Association Journal 1994150: 189-98.
  2. K McColl et al. Symptomatic benefit from eradicating Helicobacter pylori infection in patients with nonulcer dyspepsia. New England Journal of Medicine 1998 339: 1869-1874.
  3. AL Blum et al. Lack of effect of treating Helicobacter pylori infection in patients with nonulcer dyspepsia. New England Journal of Medicine 1998 339: 1875-1881.
  4. J Gilbvarry et al. Eradication of Helicobacter pylori affects symptoms in no-ulcer dyspepsia. Scand J Gastroenterol 1997 32: 535-540.

previous or next story in this issue