Isotretinoin and suicide: Epidemiological study shows no association

London, Ontario, Canada - Suicidal behavior among patients taking isotretinoin represents an uncommon, idiosyncratic phenomenon, according to the author of a recent study of nationally representative patient data spanning a decade.

"As a clinician, I've seen patients who have had emotional reactions to isotretinoin," says Madhulika A. Gupta, M.D., professor of psychiatry (with a long-standing interest in psychodermatology), at the University of Western Ontario; London, Ontario, Canada.

Anecdotally and in scientific literature, she says, isotretinoin has been associated with a range of psychiatric reactions, including suicide. However, she says, such reports primarily involve small case series ’ the link doesn't hold up in larger studies.

Gathering data

To determine how frequently suicide and self-injury occur in patients taking isotretinoin, Dr. Gupta investigated the comorbidity between isotretinoin and suicide in data collected between 1993 and 2003 from the National Ambulatory Medical Care Survey (NAMCS) and the National Hospital Ambulatory Medical Care Survey (NHAMCS).

Both are nationally representative samples of patient visits for all purposes occurring in the United States, she says.

"This data is in the public domain, available on the Centers for Disease Control and Prevention (CDC) website," Dr. Gupta tells Dermatology Times.

For the study, she merged the databases from 1993 to 2003. Once the data were properly formatted, Dr. Gupta then searched for ICD codes representing suicide and the drug codes for isotretinoin.

To test the study’s validity, she and her colleagues also searched for links between suicide and two other drugs: the antidepressant fluoxetine and Narcan (naloxone hydrochloride, Endo Pharmaceuticals), the latter of which is used almost exclusively to treat narcotic overdoses, she says.

Findings

In total, Dr. Gupta found 521 patient visits involving isotretinoin use. When she applied an epidemiologic weighting variable supplied by the CDC, she says, "Those 521 patient visits translate to 9.6 million visits throughout the United States over 11 years."

Somewhat similarly, her database included 1,176 patient visits associated with "suicide and self-inflicted injury," and these visits represent more than 4.2 million weighted visits.

Additionally, her analysis uncovered 5,939 visits involving fluoxetine (more than 79 million weighted visits) and 360 visits involving Narcan (more than 1.2 million weighted visits).

Looking at other conditions

"I wanted to analyze other conditions where I know for sure that there is an association with suicide," Dr. Gupta says.

Although the cross-sectional study could not determine that taking fluoxetine causes suicidality, she says, "One would expect a strong association because a person who's taking antidepressants is invariably more suicidal than one who isn't."

Indeed, researchers found a significant association between fluoxetine and suicide (OR = 2.5; 95 percent CI = 1.36-4.7) in their sample.

The link between Narcan and suicide was even stronger (OR=417.9; 95 percent CI = 257.8-677.4).

"One would expect that," Dr. Gupta says, "because Narcan is used almost exclusively in situations where the patient is unconscious and one suspects a narcotic overdose, intentional or otherwise. The strength of this association, and of the fluoxetine-suicide association, establishes the concurrent validity of the statistical analysis."

-No association-

In contrast, Dr. Gupta says researchers found no association between isotretinoin and the suicide variable.

"Over the years," she says, "I have seen a few patients who are on isotretinoin who become emotional, irritable and sometimes suicidal. So, I have no doubt that some people do become suicidal, even some who had no previous psychiatric problems."

Nevertheless, she says her study proves such reactions are unpredictable and idiosyncratic.

"I thought I might have seen one or two cases linking isotretinoin and suicide, even though the results would have been not statistically significant. But I didn't see any," Dr. Gupta says.

Therefore, suicidality among patients taking isotretinoin is very uncommon, she says, "Because otherwise we would have seen something in these samples."

Study strengths

The study-s strengths include the fact that the NAMCS and NHAMCS databases are highly representative of the U.S. population as a whole, Dr. Gupta says.

She says that when she examined suicide in the community at large, for example, "The numbers one gets from these databases are almost identical to rates published in textbooks.

"That is amazing. This data has not been gathered with any sort of a priori agenda," which could have introduced a sampling bias.

Conversely, she says that, as with all cross-sectional studies, even the strongest statistical association does not establish causation.

"It-s unlikely that a lot of the visits are just the same people being counted more than once," Dr. Gupta says, "but it-s definitely possible."

At press time, she was analyzing data from 2004 through 2006 that's recently been added to the databases. DT

Disclosure: Dr. Gupta reports no relevant financial interests.

For more information:

www.aad.org

www.uwo.ca