NSAID (non-steroidal anti-inflammatory drugs) are used by millions of people worldwide to reduce pain and inflammation. There are two types of drugs:
- The classical ones, such as aspirin and ibuprofen (brand names Advil and Motrin),
- The so-called selective COX-2 inhibitors, such as Celebrex and Vioxx.
The difference between the two types is the claim that the former can cause gastritis (inflammation of the lining of the stomach) and stomach ulcer—and the latter has all the advantages of NSAID without the stomach-related disadvantage.
So if the selective NSAID are so good, how come in 2004, the blockbuster drug rofecoxib (sold commercially as Vioxx) was withdrawn from the U.S. market when a study linked the selective COX-2 inhibitor with a higher rate of heart attack and stroke? The short answer: There was no answer, just a lot of hand-waving. But a paper by Yu and his colleagues, cited in a recent review in Science, ran the definitive experiment and gave us the definitive answer.
How do NSAIDs work?
All tissues have a fatty acid in their membrane, called arachidonic acid. Two enzymes can break it down and release breakdown products. The interesting thing is that they work in opposite directions.
The first enzyme is called COX-1 (cyclooxygenase-1). When arachidonic acid in the membrane of blood platelets is broken down by COX-1, it releases thromboxane A2 (TXA2) which cause vasoconstriction and platelet aggregation.
Makes perfect sense. That’s exactly what we want to stanch the blood flow when an injury occurs. Arachidonic acid in the gastric mucosa (inner lining of the stomach) gets digested by the same enzyme, COX-1, to produce prostaglandins E2 and I2 (PGE2) which protects the stomach from injury by the acid sloshing around.
The second enzyme is cycloxygenase-2, (COX-2). This enzyme attacks arachidonic acid in the membrane of cells lining the joints, and the products, prostaglandins E2 and I2, cause pain and inflammation. It also attacks arachidonic acid in the membranes of endothelial cells, lining the blood vessels. The product, prostacyclin (PGI2) causes vasodilation and inhibition of platelet aggregation, exactly the opposing action of TXA2 released form platelets by COX-1.
You can see now where the confusion came from. We basically have a yin-yang situation, but it’s not a clear-cut division. Each of the enzymes does some good, and some bad. So when the selective COX-2 inhibitors came on the market as the ultimate answer to the ulcer problem of the non-selective NSAID, we neglected to think about all the other things that COX-2 does. Specifically, if we use COX-2 inhibitors, the effect on endothelial cells—vasodilation and inhibition of platelet aggregation—leaves the effect of COX-1— vasoconstriction and promotion of platelet aggregation—without a counter.
The expected clinical result should be hypertension and increased incidence of myocardial infarctions and stroke. Indeed, one of these selective COX-2 inhibitors, Vioxx, was shown conclusively to increase heart attack rates and was withdrawn from the market after a deluge of lawsuits. But is this effect specific to Vioxx? Yu and his colleagues settled the issue.
In a mouse model, they demonstrate that deletion of the gene encoding COX-2 in vascular smooth muscle cells and vascular endothelial cells lowered the urinary excretion of the PGI2 metabolite and predisposed to the development of thrombosis and hypertension. Furthermore, deletion of COX-2 gene in vascular endothelial cells and vascular smooth muscle cells decreased the synthesis and release of nitric oxide (NO), a powerful vasodilator, thereby reducing vascular relaxation.
This was accompanied by a heightened predisposition toward both hypertension and thrombosis. Thus, by genetically engineering mice that lack COX-2 in the vasculature, Yu et al. show that the mechanism of increased cardiovascular risk from COX inhibition results from specifically blocking COX-2 in the blood vessels and not from other tissues in the body.
The implications of the work by Yu et al. are that the increased cardiovascular risk is not an individual drug side effect but is rather a direct pharmacologic consequence of inhibition of COX-2. Furthermore, these risks apply to all COX-2 inhibitors—both selective and nonselective NSAIDs. Naproxen (Naprocyn) is the only NSAID that has not shown increased cardiovascular risk, likely due to its sustained inhibition of COX-1, which provides an antiplatelet effect.
So what’s a person to do?
It is indeed frustratingly confusing. You inhibit COX-1 and you are setting yourself up for a peptic ulcer. You inhibit COX-2 and you are prone to hypertension, myocardial infarction and stroke. The best advice is to avoid the COX-2 inhibitors, such as Celebrex. They’re expensive, to boot.
The non-selective COX-1 and COX-2 inhibitors are cardioprotective, but can give you an ulcer. The way out of this dilemma, at least until new NSAIDs are developed, is to use them at low doses. Those of us taking “baby aspirin”, 81 mg dose, are on safe grounds. Those who take the regular 364 mg pills should not take more than 6-8 pills a day—the less the better.
The NSAID story illustrates something else. The body is maintained in a fine balance, and any disruption is bound to cause some unintended consequences. It is incumbent upon us to re-evaluate the evidence vis-à-vis any theory and clinical practice, and change when the evidence ceases to support them.