Well there is always the scientific approach and a review of the medical literature. Admittedly the literature is a lot drier and less entertaining than Dr. Gupta.
There is also simple arithmetic The American media like to give the impression that violent crime and homicide are at epidemic levels. It is always a shock when people discover that in fact we are at a 30 year low:
The homicide rate has actually declined from 10.2 per 100,000 in 1980 to 5.0 per 100,000 in 2009. What are the odds of that happening if a major new cause of homicide is being added at the same time (namely antidepressants). How does that compare with antidepressant use? A recent study estimated that from 1996 to 2005, the number of Americans older than 6 years of age in surveyed households who received at least one antidepressant in the year studies increased from 5.84% in 1996 to 10.12% in 2005. From the table there was a 24% reduction in the homicide rate during a time that antidepressant use nearly doubled. One in ten Americans received an antidepressant prescription The authors of this study noted this trend was broad based and correlated with a lower percentage of people receiving psychotherapy.
But what does that tell us about the observation that antidepressants cause homicide? Technically there is no current way to demonstrate causality from a negative correlation between homicide rates and the rate of people taking antidepressants. A large scale significant negative correlation between antidepressant use and lethal violence over a 15 year period has already been reported in the Netherlands.
What about the commentator suggesting that the toxicology of homicide perpetrators shows that they can have psychiatric drugs present that explain their homicidal behavior. In fact, a study looking at that issue showed that 2.4% of 127 murder-suicide perpetrators had toxicology that was positive for antidepressants. That is a lower than expected rate of antidepressant use than in the general population. In a study of elderly spousal homicide-suicide perpetrators, depression was seen as an antecedent to this act but none of the perpetrators tested positive for antidepressants.
Given these observations any claim that antidepressant or any psychiatric drug causes homicidal behavior needs to be backed up with some hard data. I don't mean a series of cases reported by somebody to make a point and I don't mean a legal decision where lawyers and judges can pretend that scientific data do not exist and make a decision about what they hear in a court room. I also do not mean listening to somebody claim that we will never know the real relationship until we conduct "prospective double blind placebo controlled studies" of homicidality as a medication side effect. If it isn't obvious, that study would by definition be unethical and would not pass the scrutiny of any human subjects committee.
Anyone with potential homicidal thinking needs close supervision and treatment. They may need inpatient treatment in a unit that specialized in treating homicidal thinking and behavior. Any clinician working in these settings will tell you that the people being treated generally come in with aggressive and violent thoughts and behavior before they take any medication. If they have positive toxicology associated with homicidal thinking it is generally alcohol or an illicit drug like cocaine or methamphetamine. Anyone with this problem also needs close monitoring and management of medication side effects. Antidepressants can cause agitation and restlessness. There are some people who do not benefit from antidepressants. In the case of persons with the potential for aggression and suicide the medication response may need to be determined in a controlled environment before they can be safely treated. Like all medications antidepressants are not perfect medications and they need to be administered by an expert who can provide effective treatment while managing and eliminating any potential drug side effects.
George Dawson, MD, DFAPA
