Prompt endoscopy was more cost-effective than H2-blockers in dyspepsia
ACP J Club. 1994 Sept-Oct;121:31. doi:10.7326/ACPJC-1994-121-2-031
Bytzer P, Hansen JM, Schaffalitzky de Muckadell OB. Empirical H2-blocker therapy or prompt endoscopy in management of dyspepsia. Lancet. 1994 Apr 2;343:811-6.
To compare empirical histamine-2 (H2)-blocker treatment with prompt endoscopy before drug treatment for symptoms, quality of life, and health care costs in patients with dyspepsia.
Randomized controlled trial with 1-year follow-up.
77 primary care practices in a city (population 170 000) in Denmark.
414 patients with dyspepsia (epigastric pain or discomfort of sufficient severity to justify empirical treatment with an H2-blocker) and with no history of peptic ulcer or esophagitis. Exclusion criteria were age < 18 years, treatment with ulcer-healing drugs (except antacids) within the previous 2 months, anemia or jaundice, suspected upper gastrointestinal bleeding, contraindications to endoscopy, previous surgery on the upper gastrointestinal tract, pregnancy, suspected serious or fatal condition, or lack of cooperation. Follow-up at 4 weeks and 1 year was 97% and 90%, respectively.
208 patients were randomized to prompt endoscopy within 1 week of referral. Only patients with structural abnormalities (duodenal or gastric ulcers or reflux esophagitis) received H2-blocker therapy. 206 patients were randomized to immediate treatment with H2-blockers. In these patients endoscopy was done only in case of treatment failure or relapse.
Main outcome measures
Dyspepsia symptoms; consumption of antacids, ulcer-healing drugs, and other medication for dyspepsia; sick-leave days taken; satisfaction with medical care; number of patient visits to a physician; total endoscopies; quality of life; and total health care costs.
The groups did not differ for symptoms (Table), quality of life, or additional diagnostic procedures. Patients who had immediate endoscopies, compared with patients who had drug treatment, had more satisfaction with diagnostic and management efforts (P < 0.001) (Table), fewer general practitioner visits (47 vs 114 visits), and fewer dyspepsia-related sick-leave days (542 vs 1096 d) and sick-leave days (2323 vs 3436 d) for other reasons, and lower total costs (mostly related to drug costs and sick-leave days). Case selection for endoscopy was not improved by the empirical H2-blocker strategy because 40% of ulcers remained undiagnosed.
Prompt endoscopy was more cost-effective and had fewer associated sick-leave days than did immediate treatment with H2-blockers in patients with dyspepsia. Prompt endoscopy also led to greater satisfaction with medical care.
Sources of funding: Odense University, Denmark, and Glaxo Pharmaceuticals (ranitidine).
For article reprint: Dr. P. Bytzer, Department of Gastroenterology, Glostrup Hospital, University of Copenhagen, DK-2600 Glostrup, Denmark. FAX 45-4323-3950.
Table. Prompt endoscopy vs immediate treatment with H2-blockers in dyspepsia*
|Outcome at 1 year||Prompt endoscopy||H2-blockers||RBR (95% CI)||NNH|
|No symptoms||21%||22%||3% (-42 to 33)||Not significant|
|Symptoms worse||3%||2%||49% (-54 to 386)||Not significant|
|RBI (CI)||NNT (CI)|
|Symptoms improved||18%||17%||3% (-34 to 59)||Not significant|
|Satisfied or very satisfied with medical care||50%||28%||76% (35 to 132)||5 ( 4 to 9)|
|Dissatisfied or very dissatisfied with medical care||3%||25%||87% (72 to 94)||5 (4 to 7)|
*RBR = relative benefit reduction. Other abbreviations defined in Glossary; RBR, RRI, NNH, NNT, and CI calculated from data in article.
Dyspepsia (pain or discomfort centered in the upper abdomen) affects up to one third of the population. Patients with uninvestigated dyspepsia who consult physicians may have a serious underlying disease (e.g., gastric cancer) or may occasionally develop complications (e.g., bleeding from an ulcer), but a large proportion have no structural explanation for their symptoms and are diagnosed as having nonulcer (or functional) dyspepsia (NUD) (1). Traditionally, management of the patient with recent-onset dyspepsia has involved prescribing empirical therapy, such as an H2-blocker, or arranging for appropriate investigations, such as endoscopy, to rule out structural disease. The American College of Physicians (2) issued guidelines in 1985 recommending that in patients without alarm features, initial empirical therapy was desirable and cost-effective. The excellent randomized controlled trial by Bytzer and colleagues calls this policy into question.
This is the first study to objectively test the empirical therapy strategy. An empirical trial presupposes that the need for long-term investigation is eliminated in patients who respond, but the study by Bytzer and colleagues indicates that this assumption is not valid. Moreover, with empirical therapy, true responders cannot be differentiated from placebo responders. The conclusion that the empirical approach is not a cost-saving strategy has been confirmed by recent decision analyses (3).
What are the advantages of early investigation? Therapy can be appropriately targeted at peptic ulcer (e.g., eradicate Helicobacter pylori or eliminate nonsteroidal anti-inflammatory drugs), gastric cancer, and reflux esophagitis. In such cases, empirical drug treatment may mask the condition and delay diagnosis. In patients with NUD, firm reassurance about the benign but recurrent nature of the disorder can be given; this may improve patient satisfaction and reduce the need to prescribe medications that have uncertain efficacy.
Nicholas J. Talley, MD, PhD
University of SydneySydney, New South Wales, Australia