In August 2011, the US Food and Drug Administration issued a warning that doses of the selective serotonin reuptake inhibitor (SSRI) antidepressant citalopram (Celexa) that exceeded 40mg/day could prolong the QT interval, a measure of heart rate used to diagnose abnormal heart rhythms. A study of records from the Veterans Health Administration showed that 35,848 veterans whose dose of citalopram was reduced from an average of 64mg/day to under 40mg/day faced increased deaths, hospitalizations for any cause, and hospitalizations for depression specifically after the reductions.
The FDA warning meant to prevent heart problems had the unintended consequence of increasing hospitalizations and deaths among the veterans affected. These findings by Thomas S. Rector and colleagues were published in the American Journal of Psychiatry in 2016.
Editor’s Note: There are some similarities between this case and findings by researchers Andrew Nierenberg and Andrew Stoll, who noticed that patients taking 40mg/day of fluoxetine (Prozac) had better long-term outcomes than those taking 20mg/day, even though those taking 40mg were more ill and more likely to relapse at the start of the study.
Researchers Ellen Frank and David Kupfer found that 90% of unipolar depressed patients relapsed when their antidepressant doses were halved, even though they had been stable for 5 years before the change.
These and the findings from Rector and colleagues lead this editor to believe that reducing the dosage of effective treatments should not be done without reason—that is, in the absence of side effects, or simply to achieve the minimal effective dose. Dose reductions without cause not only may increase the risk of relapse, but may also put the patient at increased risk of developing tolerance to the medication, for example hastening the onset of ‘Prozac poop-out.’
When long-term maintenance drug therapy is going well, it may be best to be conservative and stay the course. Conversely, in the absence of a good long-term response, be as active and creative as possible to achieve mood stabilization.
A 30-year observational study published by Andrew Leon and colleagues in the American Journal of Psychiatry has found that anticonvulsants used in epilepsy and for bipolar depression (carbamazepine, lamotrigine, and valproate) do not increase suicidal behavior in bipolar patients.
Editor’s Note: The FDA gave a warning in 2009 that these anticonvulsants were associated with suicidal ideation. This was based on studies of a mixed group of psychiatry and neurological patients in acute placebo-controlled studies, where suicidal ideation is typically a reason for exclusion from the study. Leon et al. used more powerful longitudinal methods to compare the risk of suicidal ideation in individuals taking and not taking anticonvulsants and found no such increase in suicidal behavior.
This is like the FDA warning for antidepressants and suicide, which was based on data from placebo-controlled clinical trials in acute depression (where suicidal patients are excluded). When investigators used the same longitudinal methods as Leon et al. in the anticonvulsant study, they found that antidepressants actually reduced suicidal behavior by 30%.
The bottom line is that the use of anticonvulsants for bipolar disorder should not be discouraged based on the FDA warning about suicidal ideation in mixed neurological and psychiatric patients. In bipolar patients, anticonvulsants do not increase the risk of suicidal behaviors, i.e. suicidal acts or completed suicides.
In 2007, the FDA began labeling antidepressants with a warning that patients aged 18-24 were at risk for increased suicidality during the first weeks of treatment. New evidence shows antidepressants actually have beneficial effects on suicide risk in adults. A study of all published and unpublished data on the SSRI fluoxetine (Prozac) and the SNRI venlafaxine (Effexor) published in 2012 by Gibbons et al. in the Archives of General Psychiatry showed that these antidepressants substantially reduced suicidal thoughts and behavior in adults and produced no increase in suicidal thoughts or behavior in children and adolescents.
The protective effect on suicidality in adults was mediated by mood, i.e. the patients’ mood improved and they became less suicidal. Children’s mood also improved on the antidepressants, but their risk of suicidal ideation did not change.
Editor’s Note: These are important findings. When the FDA box warning on antidepressants and suicidal ideation appeared, antidepressant treatment of youth decreased without an accompanying increase in psychotherapy, and the actual suicide rate in youth increased.
We now know that childhood-onset depression carries a bigger risk for a poor outcome in adulthood than adult-onset illness. In parallel, greater numbers of depressions are associated with more impairment, disability, cognitive dysfunction, medical comorbidities, treatment resistances, and neurobiological abnormalities.
It is important to treat illness in young people in order to prevent these difficulties, and the suicide warning should not deter the use of antidepressants. Patients should be careful about suicidal ideation in the first several months after starting an antidepressant, as other data suggest that this is a time of slightly increased risk of suicidal thoughts in children and adolescents.