Review of: Ray WA et al, High-dose citalopram and escitalopram and the risk of out-of-hospital death. J Clin Psychiatry 2016. doi:10.4088/JCP.15m10324

Study type: Retrospective cohort study

Selective serotonin reuptake inhibitors (SSRIs) are considered the first-line treatment for depression; however, our confidence in their safety took a hit when the FDA issued a warning in 2011 about doses of citalopram above 40 mg causing QTc prolongation. The FDA originally said that high-dose citalopram was “contraindicated,” but in 2012 softened its language to “not recommended.” However, even this gentler warning remains controversial, and the study examined here attempts to further pinpoint whether QTc prolongation is any more likely with high-dose citalopram than with other SSRIs.

The authors reviewed the records of Tennessee Medicaid enrollees who had prescriptions for high doses of SSRIs between 1998 and 2011. “High doses” were defined as > 40 mg of citalopram, paroxetine, or fluoxetine; > 20 mg of escitalopram; or > 150 mg of sertraline. The study endpoint was sudden unexpected deaths, which included sudden cardiac deaths, other cardiovascular deaths, and unintentional medication overdose deaths in non-hospitalized patients. The authors believed that these deaths were more likely to be related to cardiac arrhythmias.

Results
There were 54,220 persons with 557,510 prescriptions meeting criteria for high-dose SSRIs. There were 254 deaths during the study period, including 145 sudden unexpected deaths and 100 other deaths. In comparing deaths in patients taking the different SSRIs, the authors adjusted for various comorbidities, such as other risk factors for heart disease. The adjusted risk of sudden unexpected death was not significantly higher in citalopram-treated patients than in patients treated with any other SSRI. Here are the numbers for the hazard ratios for citalopram vs each of the SSRIs examined: citalopram vs escitalopram, 0.84 (95% confidence interval (CI), 0.4–1.75); citalopram vs fluoxetine, 1.24 (95% CI, 0.75–2.05); citalopram vs paroxetine, 0.75 (95% CI, 0.45–1.24); citalopram vs sertraline, 1.53 (95% CI, 0.91–2.55). For those rusty on their statistics, the hazard ratio represents the relative risk. The 95% CI means there’s a 95% chance that the actual risk is somewhere in the cited range. Thus, patients on citalopram were 1.53 times more likely than those on sertraline to have sudden death. That sounds damning for citalopram, but the 95% CI means that the actual risk could be as low as 0.91 (lower than sertraline’s risk) or as high as 2.55. Because the lower range of the CI is below 1, the risk is considered non-significant.

When the authors zeroed in on the subset of patients who had particularly high cardiac risk factors, they still found no significant difference among the medications in the risk of sudden unexpected death.

TCPR’s Take
This was a retrospective chart review, and it’s hard to ascertain causes of death with certainty. However, the study size was quite large, meaning that it is likely to be generalizable to the larger patient population. There is still concern about QTc prolongation with citalopram, but at least in this study, the risk of sudden unexpected death was not significantly different with high-dose citalopram than with escitalopram, fluoxetine, paroxetine, or sertraline users. This finding is aligned with another recent study showing that when VA patients were taken off of citalopram after the FDA warning, they were at higher risk for depression and not at lower risk of arrhythmias (see TCPR, July/August 2016).

Practice Implications
Research continues to chip away at the FDA citalopram warning. While it’s prudent to be aware that there is some concern about citalopram, don’t hesitate to prescribe it in high doses when you and your patient think the benefits outweigh the questionable risks.