An absolutely fabulous piece in (of all places) the Wall Street Journal called When Drug Labels Make You Sick. It asks the important question, "Is it a good idea to read about all the possible side effects of medications you're taking?"
The answer is definitely "no." The WSJ piece talks not only about the placebo effect (where taking a sugar pill makes you better), but also about the nocebo effect which is essentially thinking a pill you are taking will make you sick, actually will make you sick. Both are powerful forces. The reason we do randomized placebo controlled trials is because in virtually every study, patients taking a placebo get better, by about 25%. The same is also true of side effects of studies, i.e. patients enrolled in studies who don't know whether or not they are getting a real medicine are likely to report side effects. These side effects are ones that people commonly have such as headache, nausea, anxiety, etc.
What the Wall Street Journal does not delve into is how medication side effects are reported. This is critical in understanding both why you shouldn't read all the side effects, as well as which side effects you should actually pay attention to.
Let's use the example of Crestor. I choose this medication, because of the recently published, and highly controversial Jupiter study which I blogged about recently used Crestor 20mg a day.
If you go to Crestor's web site, you first see bulleted side effects of Headache, Muscle pain , Abdominal pain, Weakness and Nausea. Next it says "For a complete list, see the full Prescribing Information for CRESTOR," which is a link to the package insert (more on that in a minute). Finally, it states "If you develop any unexplained muscle pain, tenderness, or weakness at any time during treatment with CRESTOR (especially if you also have a fever or feel ill), call your doctor right away, as these symptoms could be a sign of a rare but serious side effect." If that last part sounds like a TV ad, it's because it is the identical language you see in the their TV commercial.
The package insert is the folded up piece of paper that comes with every prescription. It is designed to tell physicians everything they need to know about the drug: how it works, how well it works, precautions to be taken in prescribing such as interactions with other medications, and of course side effects. The FDA recently changed the format of the PI (as it is known in the industry) with the first page being a summary page. It lists all the major information up front. If you go to the PI, you will note that under "Adverse Reactions" are listed "headache, myalgia, abdominal pain, asthenia, and nausea" which are the same as "Headache, Muscle pain , Abdominal pain, Weakness and Nausea" from the web site. Just above, under Warnigns and Precautions list the severe muscle effects and liver enzyme monitoring. This is the stuff at the end of the TV commercials that the FDA mandates must be disclosed. Thus, if you go to the PI of Viagra, you will find the "erections lasting more than 4 hours" under Warnings and Precautions.
Here's the important thing to know: Side effects or Adverse Effects in the PI are reported effects that happened in 2% or more of the patients REGARDLESS OF WHETHER OR NOT IT WAS MORE THAN THOSE TAKING PLACEBO. If you look at Table 1 in the Crestor PI, you can see that the five side effects listed for Crestor occured at about the same rate for those taking Crestor as those taking placebo. 5.5% of Crestor patients had headache and 5.0% of placebo patients had a headache. Thus, headache is listed as a side effect, but does Crestor really cause headaches? Probably not. This is the nocebo effect that the WSJ piece was talking about. Headache is common, thus you would expect that in a study looking at a new pill, patients would likey report having headache regardless. However, both in what you receive from the pharmacy, what you Ads you see in magazines, and of course those TV commercials, the FDA requires the drug companies to list these as "side effects." The only true side effect caused by Crestor is likely muscle pain, which PI reveals to be 2.8% in Crestor, vs. 1.3% in placebo. In fact, muscle pain is about 3-5% in all statins (Lipitor, Zocor), is usually mild, and usually goes away. What was really interesting about the Jupiter study, which didn't really get picked up in the press, was that there was really no statistically significant difference in side effects of the close to 20,000 patients taking either Crestor 20mg or placebo.
Now the myalgia or rhabdomyolysis is a different story, which is why it is listed under Warnings and Precautions, and why the FDA mandates the be clearly stated. These effects can be very serious, and can even cause death. Fortunately, the occurrence is rare. The true rate of rhabdomyolysis is between 1/100,000 to greater than 1/1,000,000. As carefully documented by the FDA in their response to Dr. Sidney Wolfe who wanted Crestor pulled from the market, stated that rhabdo was rare, and was no different for Crestor than any other statin.
Bottom Line: The WSJ does a great job discussing both the placebo effect and nocebo effect. If you think a medication is going to work, it increases the chances that it will. If you think it is going to cause side effects, you also increase those chances. "Side effects" reported by the drug company are really anything that happened to 2% or more of the patients, regardless whether or not it was different than placebo. When asking about side effects, find out about true side effects (those that occurred much more often then in those taking placebo pills) as well as rare but potentially serious side effects. You will then need to weigh whether or not a small chance of having one of these rare but serious side effects is outweigh by the potential benefits of the drug. Statins decrease heart attacks and strokes, and deaths caused by either. Cardiovascular disease is the single leading cause of death in the US. Compared to a lottery like chance that you could have something seriously go wrong, you can see why these medications are the most commonly prescribed in the country.