Concerns about the critical appraisal of the role of H pylori in dyspepsia.

1732 Canadian Family Physician•Le Médecin de famille canadien❖VOL 46: SEPTEMBER • SEPTEMBRE 2000

Letters ^❖ Correspondance Letters ^❖ Correspondance

Concerns about the critical appraisal of the role of H pylori in dyspepsia

A

family practice (CCFP) colleague sent us a critical appraisal article written by Drs Hammet and Evans¹ that reviewed the Talley study of Helicobacter pylorieradication in func- tional dyspepsia.²The article was sent to us, as we are both gastroenterolo- gists with research interest in dyspep- sia and H pylori.

This placebo-controlled study showed that eradication of H pylori did not improve dyspeptic symptoms in investigated patients diagnosed with functional dyspepsia. This finding is not in dispute. However, the review of this article¹and its interpretation con- tained factual errors that could have led the authors to miss some of the subtle nuances of the study.

We have specific comments about this appraisal.

1. In the section “relevance to family physicians,” the authors state that 15% to 20% of patients have dyspepsia and later state 5% to 10% develop dys- peptic symptoms in a given year.

Which is it? They state that 50% do not have symptoms the following year, but do not give a reference.

This prevalence of symptoms is inconsistent with our knowledge of the literature or personal experience.

2. Patients were screened for H pylori with a whole blood test or a rapid urease test from biopsies taken during diagnostic endoscopy.

Patients with positive test results were enrolled in a 1-week run-in period and patients had to have at

least moderate symptoms before they could be randomized in the study (ie, they were not patients with just trivial symptoms at base- line). At this point, the urea breath test was done and patients were considered H pylori—positive only if two tests had positive results.

The method of H pylori documenta- tion reported by the authors was not technically correct.

3. The authors write that patients who missed more than two of the 14 doses were excluded. This is incor- rect, but those that missed more than two doses were considered noncompliant. They were all included in the ITT analysis.

4. The authors repor ted that two patients were subsequently exclud- ed for lack of follow-up data. Why one patient in the placebo group was excluded is unclear. Also, Table 1 contains several errors.

The placebo ar m included 142 patients, not 143. Results in the treatment group of those with mini- mal or no dyspepsia should be reported for 32/133 (not 31/133) and in the placebo group for 31/142 (not 28/143). We are uncer- tain where they derived the num- bers they report.

5. In the “Results” and in the “Bottom line” sections, the authors report that patients with mild gastritis and functional dyspepsia are more like- ly to have symptom improvement than those with severe gastritis.

This is a misinterpretation. The data they report pertains to after treatment in which those who had no or only mild gastritis at study end had better symptom relief than those with persistent, severe gastri- tis. This is consistent with the hypothesis that H pylori eradication heals gastritis slowly over time and could result in better symptom relief. These data need to be inter- preted with caution, as analysis was post-hoc. Indeed, these results were not confirmed in a similar, subsequent study.³

6.The ORCHID study² studied patients with functional dyspepsia.

The diagnosis was made only after a patient under went an investigation and had no abnor- mality identified that accounted for the symptoms. This critical appraisal does not specify that, unless their patients all are investi- gated (endoscopy or barium x-ray),

Make your views known!

Contact us by e-mail at letters.editor@cfpc.ca on the College’s website at www.cfpc.ca

by fax to the Scientific Editor at (905) 629-0893 or by mail to Canadian Family Physician College of Family Physicians of Canada

2630 Skymark Ave Mississauga, ON L4W 5A4

…

Faites-vous entendre!

Communiquez avec nous par courier électronique:

letters.editor@cfpc.ca au site web du Collège:

www.cfpc.ca par télécopieur au Rédacteur scientifique (905) 629-0893 ou par la poste

Le Médecin de famille canadien Collège des médecins de famille du Canada

2630 avenue Skymark Mississauga, ON L4W 5A4

VOL 46: SEPTEMBER • SEPTEMBRE 2000❖Canadian Family Physician•Le Médecin de famille canadien 1733

Letters

^❖

Correspondance

the results of the ORCHID study are not generalizable to patients that present with undiagnosed and uninvestigated dyspepsia. This rep- resents a conundrum in dyspepsia management. Studies such as this one do not replace clinical judg- ment, and physicians are remind- ed that a positive noninvasive H pyloritest (eg, serology or breath test) is not a surrogate for diag- nosing ulcer disease. Studies in a primar y care setting are needed that address whether noninvasive- ly testing for H pylori and treating a positive result would benefit uninvestigated dyspeptic patients.

Recently a Canadian study addr essing this cr ucial issue showed a significant benefit of the

“test for H pylori by 13C-UBT and eradicate if positive” strategy.⁴ This primar y care study did not include initial endoscopy; thus, it probably included some patients with ulcers in addition to patients with functional dyspepsia.

7. The authors suggest that it is advisable to test for H pylori only when there is proven ulcer dis- ease or MALT lymphoma.

Patients with these conditions can be diagnosed only through inves- tigation. Surely the authors did not intend to convey the message that ever y dyspeptic patient should be investigated befor e treatment.

8. The authors cite a reference by Dr Blaser,⁵which discusses possi- ble negative consequences of treating H pylori–positive dyspep- sia patients. This could have led them to state that eradicating H pylori can be mor e har mful than beneficial. This har m includes precipitation of gastroe- sophageal reflux disease (GERD);

however, most studies to date do not suggest that curing H pylori infection causes GERD. Some data suggest that intragastric pH is less well contr olled after H pylori eradication.6 The clinical

significance of this obser vation is currently unclear. Some suggest that H pylori infection helps to protect against development of esophageal cancer. However, it is impor tant to realize that, while the incidence of distal esophageal adenocarcinoma is rising, it is still less than that of gastric cancer. As gastric cancer is linked to H pylori infection, it makes sense to eradi- cate H pylori. Indeed, we believe that, if a patient is known to be infected with H pylori, treatment must be of fered.7 Whether all patients with dyspepsia should be tested for H pylori is a dif ferent matter.

This Critical Appraisal series has merit. In future, we suggest that these appraisals are done by both family physicians and exper ts, so that the reviews can be put into proper clinical perspective.

We hope that our comments will be helpful.

— Naoki Chiba, MD, FRCPC

Associate Clinical Professor of Medicine McMaster University, Hamilton, Ont Surrey GI Clinic/Research, Guelph, Ont

— Sander JO Veldhuyzen van Zanten,

MD, PHD, FRCPC

Associate Professor of Medicine Dalhousie University, Halifax, NS by mail

References

1. Hammett DC, Evans MF. Functional (non- ulcer) dyspepsia and Helicobacter pylori infection. To treat or not to treat? Can Fam Physician1999;45:2323-6.

2. Talley NJ, Janssens J, Lauritsen K, Racz I, Bolling-Sternevald E. Eradication of Helicobacter pyloriin functional dyspep- sia: randomized double blind placebo controlled trial with 12 months’ follow up. The Optimal Regimen Cures HelicobacterInduced Dyspepsia (ORCHID) Study Group. BMJ 1999;318(7187):833-7.

3. Talley NJ, Vakil N, Ballard ED, Fennerty MB.

Absence of benefit of eradicating Helicobacter pylori in patients with nonulcer dyspepsia. N Engl J Med1999;341(15):1106-11.

4. Chiba N, Veldhuyzen van Zanten SJO, Sinclair P, Ferguson RA, Escobedo S, CADET-Hp Study Group. Beneficial effect of H pylorieradication therapy on long term symptom relief in primary care patients with uninvestigated dyspepsia: the CADET-Hp study [Abstract]. Can J Gastroenterol 2000;14(Suppl A)17A.

5. Blaser MJ. In a world of black and white, Helicobacter pylori is gray. Ann Intern Med 1999;130(8):695-7.

6. Labenz J, Tillenburg B, Peitz U, Idström JP, Verdú EF, Stolte M, et al. Helicobacter pylori augments the pH-increasing effect of omeprazole in patients with duodenal ulcer.

Gastroenterology1996;110(3):725-32.

7. Hunt RH, Fallone CA, Thomson AB.

Canadian Helicobacter pylori Consensus Conference update: infections in adults.

Canadian Helicobacter Study Group. Can J Gastroenterol1999;13(3):213-7.

1734 Canadian Family Physician•Le Médecin de famille canadien❖VOL 46: SEPTEMBER • SEPTEMBRE 2000

Letters

^❖

Correspondance

Response

T

hank you. We appreciate your input and the time you took to comment on our Critical Appraisal arti- cle. The errors in the table appear to have been typographical; we appreci- ate that you made note of them. Also, we were incorrect to say that candi- dates who failed the r un-in period were excluded. The study was an intention-to-treat analysis, so all patients were included regardless of compliance. For tunately, these two errors made little dif ference to the interpretation of the study.

We have reviewed our description of the methods of the study and find it to be a correct representation of the article. If your knowledge of the meth- ods is related to personal participation in the study, then the article itself is incorrect.

While we appreciate that knowl- edge in the area of Helicobacter pylori continues to evolve, we did not have access to any of the literature you describe that was published after the article we reviewed. Certainly we did not have access to your abstract, which was published months after we submitted our review. Thank you for taking the time to provide us with this new information.

While this new literature does ask us to reconsider some of the subtle nuances of the study, it does not change the “bottom line” of the study we reviewed for family physicians. You interpreted our writing to convey to readers that ever y dyspeptic patient should be investigated before treat- ment. The message we meant to con- vey was that every dyspeptic patient a family physician considers treating with eradication therapy be investigat- ed. If an investigated patient has an ulcer (and is H pylori—positive), then treat; if not, eradication therapy will probably make little dif ference to symptoms.

Also, in our review we addressed literature that suggests there might be

“good” H pylori and that eradicating

them would be “bad” because it might be harmful to patients. This literature is not definitive, however, where patients could be harmed when a treat- ment has been shown not to improve morbidity (in this case H pylori eradi- cation in a dyspeptic patient without ulcer). In such cases, it is the usual standard of practice not to of fer the treatment. It is probably reasonable not to investigate dyspeptic patients without ulcer for H pylori because testing for something inherently leads to treating it, even though treatment might not be beneficial. Thank you for apprising our readers of the con- nection between H pylori and gastric cancer and the rising incidence of dis- tal esophageal adenocar cinoma.

Obviously, eradication for cancer pre- vention is an important topic, and we look for ward to increasing evidence that this strategy makes sense in pri- mary care.

Incidence numbers given at the beginning of the ar ticle were taken from our first reference.¹The numbers are different because there is a differ- ence in prevalence between dyspepsia and upper abdominal pain (as stated in the article). As you know, these rates var y somewhat depending on the source and, of course, the clinic setting.

We appreciate your reiteration of our main points and your correction of the data. Your suggestion regard- ing specialty contributors to the criti- cal appraisal section is interesting and something we have wr estled with. Clearly, you know more about dyspepsia than we do, and this could be an advantage for our readers. This needs to be balanced with the fact that some of the subtle nuances impor tant to you as gastroenterolo- gists are not as impor tant to us as family physicians. As well, given selection bias, prevalence, and so on, the strategies that specialists derive from the literature might differ from strategies derived by family physi- cians. We are more likely to write critical appraisals that address the diagnostic and therapeutic concerns

other family physicians face in their offices each day.

As a result, we think that family physicians are the best primar y reviewers. Also, many of the articles we review are not written by special- ists. We do think many specialists are becoming more sensitive to the reali- ties of primary care, and we did send this critical appraisal to the primar y author of the article we were apprais- ing for feedback before publication. As well, a gastroenterologist with much expertise in the area of H pylori and non-ulcer dyspepsia gave us feedback on the review prior to publication.

Thank you again for your interest and useful comments on our article.

References

1. Locke GR III. Prevalence, incidence and natural history of dyspepsia and functional dyspepsia. Baillieres Clin Gastroenterol 1998;12(3):435-42.

—Michael Evans, MD, CCFP

Toronto, Ont

—Clayton Hammett, MD

Toronto, Ont

Be careful about drawing conclusions

O

ur group of family medicine resi- dents at McMaster University in Hamilton, Ont, have the following response to Graham Worrall’s article,¹

“One hundred earaches. Family prac- tice series.”

In the article, Dr Worrall reviews a family practice case series of 100 ear- aches. Of the 100 patients assessed, only four received antibiotics: two on the first visit and two in follow-up vis- its. The author subsequently con- cludes that “most people who present to primar y care physicians with ear- ache do not need antibiotics.”

We are concerned about the general- ity of this claim. This author’s work only demonstrates that most patients seen with earache in his clinic did not return to his clinic for earache. It does not show that antibiotics are unnecessary