TCPR: Dr. Nissen, you have been at the forefront of those who believe that the FDA is not doing a good enough job of warning the public about adverse effects of drugs. What is the problem?

Dr. Nissen: The FDA administers the MedWatch system, which allows physicians to report adverse events in patients they treat with medication. So if a patient has a myocardial infarction soon after starting a new antipsychotic, you can enter the details in the database. The FDA analyzes the data to look for patterns suggesting likely adverse drug effects. But the MedWatch system is voluntary, and studies have shown that only 1%-10% of adverse events are reported. One big problem is evident when an event is common in a population. For example, even if there are a number of reports of strokes in elderly patients after starting a particular drug, it can be difficult to distinguish a possible drug effect from the baseline frequency of strokes in the elderly. It gets a little easier when you look at something like ADHD. Clearly, if you have sudden death when a young person is taking a stimulant drug for ADHD, that is unusual. It is easier to link those events to the drugs that were administered because there are fewer overall occurrences. Nevertheless, when a drug is being given to millions of children, you are going to have some baseline episodes of sudden death. I served on the FDA advisory board that reviewed the cardiovascular effects of the ADHD drugs, and what impressed me was the handful of cases where sudden death or a serious ventricular arrhythmia occurred after the first dose of the drug. If a child took only one dose of a drug and then dropped dead, you tend to think that is probably drug related.

TCPR: In February of 2006, your FDA advisory board voted to recommend that all stimulant makers add a warning (but not a black box) about the cardiovascular risks of these drugs. Do you have any ongoing concerns about their risk?

Dr. Nissen: Yes I do, particularly because one of the most rapidly increasing diagnoses is ADHD and we learned that up to a half a million individuals over the age of 55 are being given these drugs. Well, these stimulants, which are amphetamines or related compounds, raise blood pressure as much as 8 to 10 mmHg. If you look at the cardiovascular literature, that kind of blood pressure elevation has been routinely associated with an increased risk of stroke in older patients. In fact, closely-related compounds such as the oral decongestant phenylpropylamine have been shown in women to increase the risk of stroke.

TCPR: With regard to the second generation antipsychotic drugs, I noticed that in 2003 the FDA required that all of the pharmaceutical companies involved put the same warnings of hyperglycemia and diabetes risks on their labels, even though the risks are clearly different in different members of the class. Why do think they put the whole class on equal footing?

Dr. Nissen: I think that the agency is inappropriately sensitive to commercial interests. They don't want to disadvantage any particular company. We have seen this in other cases. For example, I published a meta-analysis in The New England Journal of Medicine showing that the diabetes drug Avandia increases the risk of adverse cardiovascular outcomes by more than 40 percent. Actos, a drug in the same class with a very similar mechanism, did not have that risk. In fact, a similar meta-analysis indicated that it was actually protective. But the FDA required the same cautionary labeling for all of the drugs in the class. By putting all the drugs on a level playing field, the FDA does not accurately communicate the fact that there are genuine differences in risk within the class.

TCPR: Do you think that there are specific changes that should be made to improve monitoring of adverse events?Dr. Nissen: Yes. The FDA is working on something called the Sentinel System. This is basically an ongoing observational study of a captive population of patients in a large health care delivery organization – the VA system is an example. In this case, you know exactly what percentage of patients are receiving certain drugs, and you can track adverse events and compare rates between those taking the drugs and those not taking them. This is a much more effective and systematic approach to safety monitoring than the MedWatch program.

TCPR: From the standpoint of a practitioner with limited time to scour the databases and the research literature, what are we to do when the FDA comes out with a class warning that doesn’t make sense? How are we to come up with our decisions about the true relative risks of medications?

Dr. Nissen: The FDA is not doing its job in providing more precise guidance for physicians in how to use drugs in an intelligent fashion. I would very much like to see the FDA saying things like “here are the signals we are seeing and here are some of the concerns that we have about specific drugs.” They speak to physicians through drug labels. And physicians simply don’t read drug labels. I could give you innumer- able examples where warnings were put into drug label that had virtually no effect on prescribing practices. For example, you may recall that the statin drug Baychol was withdrawn because it had an 80- to 100-fold higher risk of rhabdomyalysis compared with the other statins. Most of the cases occurred when the drug gemfibrozil was given with Baychol. The FDA warned a year or two earlier that the two drugs should not be combined, but physicians kept right on doing it, because they don’t read drug labels. And you certainly don’t have pharma- ceutical companies placing advertising on television about how not to use a drug.