Vioxx and heart attacks

Just two days ago, we looked at how aspirin prevents heart attacks. There is an interesting offshoot to this idea - the Vioxx saga. This was a drug, developed by Merck & Co., that initially promised to revolutionise the NSAID (non-steroidal anti-inflammatory) group of analgesics.

What I didn't mention in the above post was that cyclooxygenase (COX) comes in two flavours. COX-1 is constitutionally expressed in most of our cells, and has several routine functions. The most important, clinically, is that the prostaglandins it helps to manufacture induce the stomach to produce more bicarbonate. This bicarbonate sits on the stomach surface and protects the stomach's cells from being damaged by the harsh acidic conditions there (the pH can go down to 0.4!).

COX-2, however, is expressed only during inflammation, and, amongst other things, this allows the endothelial cells to make their prostacyclin. This has an antithrombotic effect.

One of the biggest dangers of using any of the NSAIDs for long is that although they are excellent at reducing pain and inflammation, they tend to cause stomach ulcers due to the depletion of the stomach's bicarbonate. Why us doctors were so enthusiastic about Vioxx is that it seemed to get around this problem by selectively inhibiting only the COX-2 enzyme. In theory, this meant that we could use the drug for much longer periods without having to worry about its gastrointestinal side-effects.

In theory. But when they were used, a dark picture began to emerge - it seemed as though patients who took Vioxx were placed at an increased risk of developing cardiovascular complications like heart attacks and strokes. This was precisely the opposite effect that its conventional NSAID cousins, like aspirin, had. Although the data is still somewhat controversial, Merck eventually voluntarily withdrew the product due to these concerns.

But why did this happen? Here it is very convenient to use the wisdom of hindsight, so I must stress that the picture wasn't (and to some extent isn't) as clear as I portray it here. Nonetheless, it seems that the underappreciated fact was that COX-1 (not COX-2) is the form of the enzyme used by platelets to make their thromboxane A₂. This substance has a pro-thrombotic effect. Since, unlike aspirin, COX-1 was unmolested by Vioxx, this net COX-1 'clot forming' effect was therefore left intact. What Vioxx did inhibit though, was the 'anti-clotting' effects of COX-2. And so patients on Vioxx developed slightly more heart attacks than average.

The whole saga, with both sides of the story, can be read here.