Review of: Lavigne JE et al, J Gen Intern Med 2019;34(8):1554–1563

Study Type: Observational comparative safety study

Trazodone is one of the most widely prescribed sleep aids in the US despite lacking FDA approval for insomnia. But what if patients taking trazodone attempted suicide at a rate over 1.5 times that of those taking zolpidem (Ambien)? Lavigne and colleagues arrived at this potentially alarming conclusion in their observational safety study.

The authors mined data from the Department of Veterans Affairs (VA) to compare the rate of suicide attempts in veterans treated with various sedative-hypnotics, including trazodone, zolpidem, hydroxyzine, and benzodiazepines. Close to 350,000 patients met study criteria. The authors matched the patients in terms of psychiatric and medical comorbidities, inpatient days, disability ratings, concomitant medications, and several other variables. To control for other confounding variables, they used a “propensity score,” which is an estimate of the likelihood that a given patient would be prescribed zolpidem. This process resulted in four groups of about 75,000 patients divided according to type of sleep medication.

The primary endpoint was suicide attempts as detected in medical records. The rate of attempts was higher only when the trazodone group (n = 57) was compared to zolpidem (n = 38), a 61% increase in attempts. There was no mention of completed suicides.

Before you take your patients off trazodone, remember that observational studies have some glaring limitations. Despite the use of sophisticated statistics, it is impossible to control for all confounding factors. In this study, the authors acknowledge that they did not control for nightmares, and trazodone was a first-line treatment for nightmares in this population. This is an important confound, because the trazodone group may have already been at higher risk of suicide due to greater symptomatic distress. Although the authors controlled for comorbidities, it’s still possible that physicians were more likely to prescribe trazodone than zolpidem to patients with signs of addiction or impulsivity.

The analysis also failed to account for drug interactions. This is important because CYP2D6 inhibitors can increase levels of the trazodone metabolite, m-chlorphenylpiperazine (mCPP). Abrupt increases in this metabolite, which can happen when a CYP2D6 inhibitor is introduced, have been linked to dysphoria and anxiety, as well as to self-harm in adolescents (Shamseddeen W et al, J Child Adolesc Psychopharmacol 2019;29(7):573).

TCPR’s Take
While the apparent association of trazodone with suicidality raises concern, the study does not prove causation and may illustrate that nightmares and possibly drug interactions correlate with suicide attempts in a population of veterans. Prospective controlled trials in varied populations are needed before making recommendations against the use of trazodone.