Am Fam Physician. 2007;76(9):1273-1274
Author disclosure: Nothing to disclose.
Each month, three presenters will review an interesting journal article in a conversational manner. These articles will involve “hot topics” that affect family physicians or will “bust” commonly held medical myths. The presenters will give their opinions about the clinical value of the studies discussed. The opinions reflect the views of the presenters, not those of AFP or the AAFP.
This Month's Article
Klok RM, Postma MJ, van Hout BA, Brouwers JR. Meta-analysis: comparing the efficacy of proton pump inhibitors in short-term use. Aliment Pharmacol Ther 2003;17:1237–45.
Is one PPI better than another?
Bob: I hear constantly from sales reps, TV ads, and consultants that one PPI is better than another. This article attempts to review the literature on this topic.
What does this article say?
Bob: It is a meta-analysis of 41 studies comparing various PPIs. The authors identified 20 studies of GERD, nine studies of peptic ulcer disease, and 16 studies of Helicobacter pylori therapy. Let's look at the studies on GERD:
Four studies of pantoprazole (Protonix) 40 mg versus omeprazole (Prilosec) 20 mg found no difference.
Six studies of lansoprazole (Prevacid) 30 mg versus omeprazole 20 mg found no difference.
Two studies of rabeprazole (Aciphex) 20 mg versus omeprazole 20 mg found no difference.
Two studies of esomeprazole (Nexium) 40 mg versus omeprazole 20 mg found a small difference. The “cure rate” of 79.6 versus 67.3 percent results in a number needed to treat (NNT) of 8.
The six remaining studies were a hodge-podge group of trials comparing various PPIs at different doses, none of which demonstrated any superiority.
Andrea: All of the studies, except for the two that compared esomeprazole and omeprazole, found that there were no differences among any of the drugs in the treatment of GERD. As for the two studies that suggested esomeprazole was better than omeprazole, the studies were not fair comparisons. Both studies used higher dosages of esomperazole compared with omeprazole. This cannot be considered a true comparison of effectiveness.
Mark: This is a classic example of a “straw man comparison.” A straw man comparison is when a study intentionally compares drugs of differing doses in order to set up one as a clear “winner” in effectiveness. In the study comparing esomeprazole with omeprazole, esomeprazole did better. Well, that's no surprise—esomeprazole is the active isomer of omeprazole, and the study authors used a dosage four times that of omeprazole. Breaking this down, they used 40 mg of active drug esomeprazole versus 10 mg of active drug omeprazole (the dosage is actually 20 mg, but since omeprazole is a racemic mixture, that is only 10 mg of active drug), and guess what? The drug not going off patent did better! And who sponsored the study? Astra-Zeneca, the manufacturer.
Bob: I knew you would jump on the inappropriate comparison of esomeprazole with omeprazole.
Another problem with this comparison is that a true clinical end point is not being measured. Although an endoscopic result sounds like a good “gold standard” measurement, it is considered a “DOE” or “disease-oriented evidence” and not a clinical outcome. Abdominal pain or gastrointestinal bleeding are clinical end points that are important to patients and would be considered a “POEM” or “patient-oriented evidence that matters.” So, my conclusion regarding GERD is that there are no differences among PPIs.
As for the nine studies on peptic ulcer disease, they found the same results as those found with GERD, except this time, pantoprazole 40 mg barely had better endoscopic healing than omeprazole 20 mg at four weeks (91 versus 86 percent; NNT = 20). But I have the same complaints as I did with the GERD studies regarding the use of unequal dosages and using endoscopic cure rates as the end point.
And do I even have to do the H. pylori results? Sixteen studies with no differences noted among any PPIs.
Should we believe this study?
Bob: Meta-analysis can be tricky. A series of principles needs to be followed for the results to be valid. In this study, the authors appropriately identified a clear study question, performed a fairly comprehensive review of available databases (although they did not include any abstracts from presentations at symposia or obtain unpublished clinical trials from the FDA or the drug companies), and used clear end points to determine cure or failure—so far, so good.
The most obvious faux pas the authors made was not grading the quality of the studies they incorporated in this meta-analysis. They did, however, redeem them-selves slightly by incorporating only randomized prospective trials.
Mark: This faux pas is a problem. How do we know that the studies they included were any good? We don't. I like their conclusions and I am sure they are correct, but they need to grade the papers included in their analysis; other-wise, we have no assurance that they were any good.
What should the family physician do?
Bob: I feel comfortable saying there are no differences in efficacy among any PPIs for the above conditions. If I were to choose a PPI, I would first look at cost (Table 1). It also has been suggested that PPIs are associated with increased rates of community-acquired pneumonia1 and Clostridium difficile colitis.2 In conclusion, PPIs are not cheap and may not be benign.
Andrea: There are no differences among PPIs except cost, so go with the cheapest. However, the cheapest is not what ends up in our sample cabinet. Usually, it is the expensive, nongeneric “purple pills.”
Mark: I agree—save your patient some cash and go for the less expensive drug.
|• All PPIs are equally effective in equipotent dosages.|
|• Prescribe PPIs based on price.|
|• In head-to-head drug trials, watch out for the straw man comparison. Make sure that any article that evaluates treatments uses equipotent dosages of the drugs being compared. As a corollary, a placebo-controlled trial should not typically change your practice; any drug comparison should be against a known effective therapy, if one exists.|
|• Meta-analyses can come to the wrong conclusion for several reasons. First is publication bias—only positive trials are usually published. The second is the “garbage-in, garbage-out” phenomenon, meaning that the authors of a meta-analysis must evaluate the quality of the trials they are using so that only higher-quality studies are included.|
|• Be wary of DOEs; they are surrogate markers of disease and may or may not correlate with important clinical end points (e.g., morbidity, mortality).|