Prevention of ulcer bleeding in high-risk patients: is the enthusiasm for COX-2 selective NSAIDs justified?

Francis K.l. Chan, Lawrence C.t. Hung, Bing Yee Suen, Justin C. Y. Wu, Karla Lee, Vincent K S Leung, Aric J. Hui, Ka‐Fai To, Wai K. Leung, Vincent Wai‐Sun Wong, S. C. S. Chung, Joseph J.�y. SungHong 2007

Introduction

Non-steroidal anti-inflammatory drugs (NSAIDs) are essential for managing arthritis pain but are associated with significant gastrointestinal (GI) toxicity. In the United States, NSAID-related complications contribute to over 100,000 hospitalizations annually. For patients with a history of ulcer bleeding, the risk of recurrence is particularly high. Clinical guidelines typically recommend two strategies for these high-risk individuals: combination therapy (a standard NSAID plus a proton pump inhibitor [PPI]) or selective therapy (a COX-2 selective NSAID). This study evaluates whether COX-2 inhibitors are comparable in efficacy and cost-effectiveness to traditional combination therapy in a high-risk population.

Methods

The researchers conducted a 6-month, prospective, randomized, double-blind trial involving 290 arthritis patients with endoscopically confirmed prior ulcer bleeding. Participants were randomized into two groups:

  1. Celecoxib Group: 200 mg twice daily plus a placebo PPI.
  2. Combination Group: Diclofenac (75 mg twice daily) plus omeprazole (20 mg daily).

The primary endpoint was recurrent ulcer bleeding, confirmed via endoscopy or significant hemoglobin drops. An economic analysis was also performed from the perspective of the Hong Kong Hospital Authority, calculating direct medical costs including medications, diagnostic tests, and hospital visits. The study employed a non-inferiority hypothesis to determine if celecoxib was "no worse" than the combination therapy.

Key Results

The study found that celecoxib was non-inferior to the combination of diclofenac and omeprazole. Recurrent bleeding occurred in 5% of the celecoxib group compared to 6% in the combination group. The absolute difference of -1.5% was not statistically significant, confirming that both strategies offer similar protection against recurrent GI events.

Probability of recurrent ulcer bleeding over 6 months for Celecoxib vs. Diclofenac plus omeprazole.

Discontinuation rates were similar between groups (11-13%), primarily driven by adverse events rather than treatment failure.

Key Insights and Ideas

The Economic Tipping Point

From a public health perspective, celecoxib was generally more cost-effective, with a median total direct cost of HK$9,714 compared to HK$10,915 for the combination group—an 11% reduction. This was largely due to lower "symptom-driven" costs, such as fewer additional diagnostic tests. However, sensitivity analysis revealed a specific threshold: if the price of generic omeprazole falls to HK$3 or less per capsule, the combination therapy becomes the more economical choice.

Sensitivity analysis showing the economic tipping point based on the cost of omeprazole.

Renal Safety and Residual Risk

While COX-2 inhibitors are designed to spare the gastric mucosa, they do not necessarily protect the kidneys. Renal adverse events occurred in over 20% of patients in both treatment arms, highlighting that COX-2 selectivity does not mitigate renal risks in susceptible populations. Furthermore, the 5-6% residual bleeding risk observed in both groups suggests that neither strategy is a complete solution for the highest-risk patients.

Limitations and Future Directions

The study's 6-month duration limits conclusions regarding long-term outcomes in chronic arthritis management. Additionally, the impact of concomitant low-dose aspirin—a common requirement for cardiovascular health that increases ulcer risk—requires further investigation in larger cohorts. Future research might explore "double defense" strategies, such as combining a COX-2 inhibitor with a PPI, to address the persistent residual risk of bleeding in extremely vulnerable patients.

Summary updated

BibTeX
@inproceedings{chan2007prevention,
  author = {Francis K.L. Chan and Lawrence C.T. Hung and Bing Yee Suen and Justin C. Y. Wu and Karla Lee and Vincent K S Leung and Aric J. Hui and Ka‐Fai To and Wai K. Leung and Vincent Wai‐Sun Wong and S. C. S. Chung and Joseph J.�Y. Sung},
  title = {{Prevention of ulcer bleeding in high-risk patients: is the enthusiasm for COX-2 selective NSAIDs justified?}},
  booktitle = {Hong},
  year = {2007},
}
Figure