Dr. Howland is Associate Professor of Psychiatry, University of Pittsburgh School of Medicine, Western Psychiatric Institute and Clinic, Pittsburgh, Pennsylvania.
The author discloses that he has no significant financial interests in any product or class of products discussed directly or indirectly in this activity, including research support.
Address correspondence to Robert H. Howland, MD, Associate Professor of Psychiatry, University of Pittsburgh School of Medicine, Western Psychiatric Institute and Clinic, 3811 O’Hara Street, Pittsburgh, PA 15213; e-mail: HowlandRH@upmc.edu.
On August 24, 2011, the U.S. Food and Drug Administration (FDA) issued a safety announcement that the antidepressant drug citalopram (Celexa®) should not be used at dosages greater than 40 mg per day (or greater than 20 mg per day for patients 60 and older) because it can cause abnormal changes in the electrical activity of the heart. In this article, I will describe and critically evaluate the basis for this warning.
Why did the FDA Issue this Safety Announcement?
Potential cardiotoxicity concerns are not unique to citalopram. An undesirable property of many non-cardiac drugs is their ability to delay cardiac repolarization, an effect that can be measured as a QT interval prolongation on the electrocardiogram (ECG) tracing. Because the QT interval varies according to heart rate, by convention its measurement is adjusted according to the heart rate (referred to as the corrected QT [QTc] interval). An abnormally prolonged QTc interval is associated with an increased risk of developing torsade de pointes (a type of cardiac ventricular tachyarrhythmia associated with sudden death). According to FDA guidelines, drugs should receive an electrocardiographic evaluation beginning early in clinical development (FDA, 2009). In addition to animal studies, this would typically include a clinical trial dedicated to evaluating the effect on cardiac repolarization (referred to as a “thorough QT/QTc study”).
Citalopram is metabolized primarily by the cytochrome P-450 (CYP) hepatic enzymes CYP3A4 and CYP2C19, with a smaller contribution from CYP2D6. The major metabolite is demethylcitalopram (DCT), which is subsequently metabolized by CYP2D6 to the minor metabolite didemethylcitalopram (DDCT). In human beings, unchanged citalopram is the predominant compound in plasma. Concentrations of DCT and DDCT are one half and one tenth, respectively, that of citalopram. In a 1-year toxicology study (described in Forest Pharmaceuticals, Inc., 2011) conducted early in the development of citalopram, 5 of 10 beagle dogs receiving 8 mg/kg per day orally (equivalent to four times the 60-mg-per-day human dose) died suddenly between Weeks 17 and 31. Mortality was not seen in other animal species. A subsequent study determined that in beagle dogs, DDCT caused QT prolongation. This effect occurred at peak DDCT concentrations roughly greater than 250 ng/mL (more than 39 times the mean DDCT concentrations measured in human beings taking 60 mg per day). Concentrations of DDCT (relative to citalopram) are higher in dogs compared with other species and to human beings, suggesting a unique cardiotoxic vulnerability in dogs.
Against this background, the FDA safety notice was based on two clinical concerns. First, the results of a “thorough QT/QTc study” of citalopram became available. Second, since citalopram was approved in 1998, the FDA had received post-marketing reports of QT prolongation and torsade de pointes in some patients taking the drug.
“Thorough QT/QTC Study” of Citalopram
One hundred nineteen healthy non-depressed participants were enrolled in this unpublished randomized, multi-center, double-blind, placebo-controlled, active drug-controlled, crossover study. The antibiotic moxifloxacin (Avelox®) was apparently included as an active control drug because it has been associated with QT interval prolongation. Participants received citalopram 20 mg per day and 60 mg per day. Compared with placebo, the maximum mean QTc interval prolongation was 8.5 milliseconds (msec) for 20 mg per day and 18.5 msec for 60 mg per day. For 40 mg per day, QTc interval prolongation was estimated to be 12.6 msec. No QTc data were reported for moxifloxacin.
The mean change in the QTc interval was statistically significant (for 60 mg versus placebo), but the magnitude of the change was small (less than 20 msec). The commonly accepted threshold for clinical significance of the QTc interval (i.e., an increased risk of developing torsade de pointes) is a QTc ≥500 msec or a change in the QTc ≥60 msec (change from baseline). The mean change in the QTc interval was less than 20 msec in this study, but the FDA statement does not disclose how many participants taking citalopram (at any dose) exceeded either threshold: an increased QTc ≥60 msec or a QTc ≥500 msec. I assume that the FDA has these data and would have included them in its announcement if any of the participants exceeded either threshold, so I doubt that any participant actually fell into either category.
Additional Citalopram ECG Studies
The results of another unpublished ECG analysis are described in the citalopram package insert (Forest Pharmaceuticals, Inc., 2011), but these findings are not new. They were included in previous editions of the package insert predating the new warning. In this analysis, ECGs from 802 citalopram-treated (dose not reported) and 241 placebo-treated participants were compared with respect to “outliers.” ECG “outliers” were defined as participants with QTc changes greater than 60 msec from baseline or absolute QTc values greater than 500 msec post-dose. Tachycardia “outliers” were defined as participants with heart rate increases to greater than 100 beats per minute (bpm) along with a 25% change from baseline. Bradycardia “outliers” were defined as having heart rate decreases to less than 50 bpm along with a 25% change from baseline. In the citalopram group, 1.9% of participants had a QTc change from baseline greater than 60 msec, compared with 1.2% of those taking placebo. None of the participants taking placebo had a post-dose QTc greater than 500 msec, compared with 0.5% of those taking citalopram. The incidence of tachycardia “outliers” was 0.5% for citalopram and 0.4% for placebo. The incidence of bradycardia “outliers” was 0.9% for citalopram and 0.4% for placebo. There were no observed differences in QT or other ECG intervals. No statistical testing was reported for any of the results.
The effects of citalopram on cardiac conduction have been extensively evaluated in a large number of participants. Rasmussen, Overø, and Tanghøj (1999) reviewed the findings from 40 clinical studies. These investigations included prospective studies conducted in healthy volunteers and prospective studies conducted in adult and older patients (with depression or dementia), using doses of 5 to 60 mg per day. They also conducted a retrospective analysis of all ECG data from all clinical trials conducted from 1978 through 1996. The ECGs were performed at baseline and at study end. The placebo-controlled or active-controlled trials included 966 adult patients (ages 18 to 65) with major depression and 494 participants (ages 65 to 92) with Alzheimer’s disease. The retrospective analysis included 6,517 ECGs recorded from 1,789 citalopram-treated participants and 270 treated with placebo. Rasmussen et al. (1999) found no significant effects of citalopram on PQ, QRS, or QTc intervals. The only notable effect of citalopram was a small reduction in heart rate (less than 9 bpm).
Post-Marketing Reports of Citalopram-Associated Cardiotoxicity
The total number of reports of citalopram-associated cardiotoxicity (QTc prolongation or torsade de pointes) received by the FDA is not known, nor is it clear whether these reports are based on therapeutic doses, unusually high doses, or larger overdoses of citalopram. Only a single case report describes cardiotoxicity in a patient taking a therapeutic dose of citalopram (40 mg per day) for 5 years (de Gregorio, Morabito, Cerrito, Dattilo, & Oreto, 2011). This patient was also taking a diuretic drug. She began showing symptoms after 5 days of acute flu-related diarrhea, was found to have low potassium and magnesium, and had a prolonged QTc. Hypokalemia and hypomagnesemia, which can be caused by diuretic drugs or severe diarrhea, are associated with QT prolongation. Nearly 600 cases of citalopram overdoses (most including other drugs) are described in the published literature (Catalano, Catalano, Epstein, & Tsambiras, 2001; Hayes, Klein-Schwartz, Clark, Muller, & Miloradovich, 2010; Isbister, Bowe, Dawson, & Whyte, 2004; Jimmink, Caminada, Hunfeld, & Touw, 2008; Kelly et al., 2004; Personne, Persson, & Sjöberg, 1997). QTc prolongation or other ECG changes have been described in only about one third of citalopram overdose cases. Six fatalities associated with citalopram overdoses (along with other drugs) were described in one report, but the specific relevance of citalopram to the cause of death has been questioned (Brion, Brion, & Durigon, 1996; Glassman, 1997).
Isbister et al. (2004) reviewed consecutive overdose admissions for 57 patients taking citalopram, 42 taking fluoxetine (Prozac®), 17 taking fluvoxamine (Luvox®), 78 taking paroxetine (Paxil®), 103 taking sertraline (Zoloft®), and 318 taking non-cardiotoxic drugs. There was a small statistically significant difference in the median QTc interval for citalopram (450 msec) compared with fluoxetine (432 msec), fluvoxamine (433 msec), paroxetine (427 msec), sertraline (429 msec), and non-cardiotoxic drugs (423 msec). The proportion of patients with QTc intervals greater than 440 msec was significantly higher for citalopram (68%) than for fluoxetine (36%), fluvoxamine (24%), paroxetine (40%), and sertraline (40%), but there were no significant differences in the proportion with QTc intervals greater than 500 msec: citalopram (12%), fluoxetine (10%), fluvoxamine (6%), paroxetine (1%), and sertraline (6%). The proportion of patients taking non-cardiotoxic drugs who exceeded these two QTc thresholds was not reported. There were no deaths and no serious cardiovascular sequelae as a result of the overdoses.
Kelly et al. (2004) conducted a 2-year retrospective review of overdose admissions for 88 patients taking citalopram, 96 taking venlafaxine (Effexor®), 29 taking mirtazapine (Remeron®), and 12 taking nefazodone (Serzone®). The mean QTc interval for citalopram (427 msec), venlafaxine (420 msec), mirtazapine (423 msec), and nefazodone (419 msec) were not significantly different. There were no deaths and no serious cardiovascular sequelae.
Escitalopram and Cardiotoxicity
Citalopram is a racemic mixture. The pharmacological activity of citalopram (serotonin reuptake inhibition) is mostly attributable to the S-enantiomer, and the R-enantiomer may competitively interfere with this effect. For this reason, the S-enantiomer escitalopram (Lexapro®) was further developed as an antidepressant drug. Similar to the metabolic pathway of citalopram, escitalopram is converted to its major metabolite S-DCT and subsequently to the potentially cardiotoxic minor metabolite S-DDCT. van Gorp, Whyte, and Isbister (2009) described 78 cases of escitalopram overdoses, and 11 patients (14%) developed QTc prolongation. No patients died, and there were no serious cardiac sequelae. Hayes et al. (2010) conducted a retrospective review of 374 citalopram overdoses and 421 escitalopram overdoses. There was no significant difference in the proportion of citalopram patients having a prolonged QTc (3.7%) compared with escitalopram (1.7%). No patients died, and there were no serious cardiac sequelae.
The statistically significant results from the “thorough QT/QTc study” were small in magnitude, and their clinical significance is questionable. Additional ECG analyses from other studies do not confirm these findings. Nearly 600 cases of citalopram overdoses have been described. Although citalopram overdose is not entirely “cardiac safe,” a close analysis of large datasets of overdose cases demonstrates that only a proportion of patients (one third or less) develop QTc prolongation without serious cardiac sequelae and no deaths. The case report of fatalities associated with citalopram overdoses has been criticized. The inherent methodological limitation of using consecutive admissions to compare overdose outcomes with citalopram and other drugs precludes a definitive assessment of relative cardiotoxicity, but three studies do not demonstrate clinically meaningful differences in cardiotoxic effects. Nurses are likely to field questions and concerns from patients and other health care providers about the cardiac safety of citalopram. They should be able to understand—but critically evaluate—the basis for this safety warning. Next month, I will review additional studies pertinent to evaluating the potential cardiotoxicity of citalopram.
- Brion, F., Brion, N. & Durigon, M. (1996). Fatal overdose with citalopram?Lancet, 348, 1380. doi:10.1016/S0140-6736(05)65441-4 [CrossRef]
- Catalano, G., Catalano, M.C., Epstein, M.A. & Tsambiras, P.E. (2001). QTc interval prolongation associated with citalopram overdose: A case report and literature review. Clinical Neuropharmacology, 24, 158–162. doi:10.1097/00002826-200105000-00007 [CrossRef]
- de Gregorio, C., Morabito, G., Cerrito, M., Dattilo, G. & Oreto, G. (2011). Citalopram-induced long QT syndrome and torsade de pointes: Role for concomitant therapy and disease. International Journal of Cardiology, 148, 226–228. doi:10.1016/j.ijcard.2009.05.060 [CrossRef]
- Forest Pharmaceuticals, Inc. (2011). Celexa® (citalopram hydrobromide) tablets/oral solution. Retrieved from http://www.accessdata.fda.gov/drugsatfda_docs/label/2011/020822s038s040,021046s016s017lbl.pdf
- Glassman, A.H. (1997). Citalopram toxicity. Lancet, 350, 818. doi:10.1016/S0140-6736(05)62620-7 [CrossRef]
- Hayes, B.D., Klein-Schwartz, W., Clark, R.F., Muller, A.A. & Miloradovich, J.E. (2010). Comparison of toxicity of acute overdoses with citalopram and escitalopram. Journal of Emergency Medicine, 39, 44–48. doi:10.1016/j.jemermed.2008.06.030 [CrossRef]
- Isbister, G.K., Bowe, S.J., Dawson, A. & Whyte, I.M. (2004). Relative toxicity of selective serotonin reuptake inhibitors (SSRIs) in overdose. Journal of Toxicology, 42, 277–285.
- Jimmink, A., Caminada, K., Hunfeld, N.G. & Touw, D.J. (2008). Clinical toxicology of citalopram after acute intoxication with the sole drug or in combination with other drugs: Overview of 26 cases. Therapeutic Drug Monitoring, 30, 365–371. doi:10.1097/FJC.0b013e3181379ef6 [CrossRef]
- Kelly, C.A., Dhaun, N., Laing, W.J., Strachan, F.E., Good, A.M. & Bateman, D.N. (2004). Comparative toxicity of citalopram and the newer antidepressants after overdose. Journal of Toxicology, 42, 67–71.
- Personne, M., Persson, H. & Sjöberg, E. (1997). Citalopram toxicity. Lancet, 350, 518–519. doi:10.1016/S0140-6736(05)63109-1 [CrossRef]
- Rasmussen, S.L., Overø, K.F. & Tanghøj, P. (1999). Cardiac safety of citalopram: Prospective trials and retrospective analyses. Journal of Clinical Psychopharmacology, 19, 407–415. doi:10.1097/00004714-199910000-00004 [CrossRef]
- U.S. Food and Drug Administration. (2009). Clinical evaluation of QT/QTc interval prolongation and proarrhythmic potential for non-antiarrhythmic drugs. Retrieved from http://www.fda.gov/RegulatoryInformation/Guidances/ucm129335.htm
- U.S. Food and Drug Administration. (2011). FDA drug safety communication: Abnormal heart rhythms associated with high doses of Celexa (citalopram hydrobromide). Retrieved from http://www.fda.gov/Drugs/DrugSafety/ucm269086.htm
- van Gorp, F., Whyte, I.M. & Isbister, G.K. (2009). Clinical and ECG effects of escitalopram overdose. Annals of Emergency Medicine, 54, 404–408. doi:10.1016/j.annemergmed.2009.04.016 [CrossRef]