Lansoprazole vs. Esomeprazole: A Tale of Two Studies
The different results from these 2 similar studies emphasize the importance of understanding study design.
Two studies were performed to determine whether lansoprazole and esomeprazole -- 2 proton-pump inhibitors (PPIs) -- have differences in clinical healing and symptom resolution in patients with erosive esophagitis. Interestingly, these 2 similar studies reached very different conclusions.
Howden and colleagues randomized 284 patients with erosive esophagitis (grade 2 or higher) to receive lansoprazole (30 mg daily) or esomeprazole (40 mg daily) for 8 weeks. Healing rates at 4 weeks were 77.0% for lansoprazole and 78.3% for esomeprazole; at 8 weeks, corresponding rates were 91.4% and 89.1%. There were no significant differences between agents in symptom resolution. These authors concluded that the 2 agents were equivalent.
Castell and colleagues randomized 5241 patients with Los Angeles grade A or greater esophagitis to receive lansoprazole or esomeprazole daily for 8 weeks at the same standard doses as the other study. Healing rates at 8 weeks were 92.6% for esomeprazole and 88.8% for lansoprazole (P=0.0001). More patients treated with esomeprazole had sustained resolution of heartburn (no episodes for 7 consecutive days). These authors concluded that esomeprazole was superior to lansoprazole for healing and symptom relief.
Comment: Why do these 2 similar studies come to different conclusions? Large sample-size differences partly account for the conflicting results, but the full answer resides in the methods sections.
The Castell group assumed a 3% difference in healing between the 2 agents and used a 95% power calculation with a 0.05 significance level to determine that more than 5000 patients were necessary to detect a significant difference between agents. Indeed, they documented a difference of nearly 4% in healing rates. Conversely, the Howden researchers designed their trial to detect a difference in healing of at least 10% between the 2 PPIs with 80% power at a 0.05 significance level. The sample-size calculation using these criteria indicated that only about 140 patients were necessary in each treatment arm to detect a significant difference between agents. Were these sample-size calculations realistic?
None of the first 4 PPIs released (omeprazole, lansoprazole, rabeprazole, and pantoprazole) has ever been shown to be more effective than another in healing erosive esophagitis. A fifth PPI, esomeprazole, previously was shown to be at most 7% to 10% more effective than omeprazole for healing esophagitis. Thus, a study designed to detect a 10% or greater difference in healing between 2 PPIs is not supported by published data. The Howden study was underpowered and, therefore, unable to detect the true difference between these agents. The Howden investigators also used a power calculation of 80%, resulting in a reduced sample size, and compounding the potential for Type II error. Furthermore, data from previous studies indicate that the greatest difference in healing between esomeprazole and other PPIs occurs in patients with more severe esophagitis. The Howden researchers enrolled only 24 patients with severe esophagitis, versus 327 in the Castell study.
The bottom line is that esomeprazole is a better agent than the other 4 PPIs for healing and symptom relief, especially in patients with advanced esophagitis. The decision to use esomeprazole should be based on the clinician's goals and perception of the importance of the 4% to 10% difference in healing. More important, these 2 papers emphasize the importance of understanding study design to critically and accurately interpret what appear to be conflicting findings.
M. Brian Fennerty, MD
Published in Journal Watch Gastroenterology April 17, 2002
Howden CW et al. Evidence for therapeutic equivalence of lansoprazole 30 mg and esomeprazole 40 mg in the treatment of erosive esophagitis. Clin Drug Invest 2002 Feb; 22:99-109.
Castell DO et al. Esomeprazole (40 mg) compared with lansoprazole (30 mg) in the treatment of erosive esophagitis. Am J Gastroenterol 2002 Mar; 97:575-83.
- Medline abstract (Free)
To ensure that your Reader Remark is not formatted as one long paragraph, precede new paragraphs with either a blank line or an indentation.