Further comment by Charles Beasley

In a recent (October 2015) report on “Use of selective serotonin re-uptake inhibitors and the heart rate corrected QT interval in a real-life setting”, published in the British Journal of Clinical Pharmacology, Maljuric and associates conclude: “Although no SSRI class effect was observed, use of citalopram was associated with a longer QTc F, even after considering the recommended restrictions. Other SSRIs may not give a clinically relevant QTc F prolongation”. 

The authors sought to collect medical data for the entire population of Ommoord district of Rotterdam, the Netherlands, older than 54 years in 3 waves (1990-1993, 2000-2001, 2006-2008 [latter wave included population older than 44 years], with prospective follow-up every 4-5 years through 2012.  The purpose of the overall study was to investigate various age-related disorders and risk factors.  Across all 3 waves, approximately 69% of the target population participated (n = 14,926).  For this report, persons were included where there was access to ECGs (up to 5) and pharmacy dispensing records.  ECGs were included in these analyses when there was no evidence of atrial fibrillation, a pacemaker rhythm, patients were not receiving any antidepressant, or patients were receiving an SSRI but no other antidepressant (patient number after excluding ECGs based on stated criteria: 12,589).  Several statistical comparisons were made with adjustments for multiple relevant variables (e.g., age, gender, BMI, use of other medications that might affect ventricular repolarization or heart rate, relevant medical disorders).

In a cross-sectional comparison of all ECGs collected that were not associated with an SSRI prescription (n = 26,184) to those collected in association with an SSRI prescription (n = 436), there was a difference in QTcF (QTcFSSRI – QTcFno-SSRI: 2.9 ms [90%CI: 1.3-4.5).  For individual SSRIs studied (fluoxetine, citalopram, paroxetine, sertraline, fluvoxamine, escitalopram) none of these differences were statistically significant except for citalopram (n = 35) where the difference was 12.8 ms (90%CI: 7.3-18.2). While it might be notable that escitalopram prescription was not associated with an increased QTcF (difference: 2.7 ms [90%CI: -11.6-16.9]), there were only 5 ECGs associated with escitalopram.

In a comparison of individual patient ECG sequential pairs when 1 ECG was associated with citalopram prescription and the other with no antidepressant prescription (n = 5), the difference was 28.9 ms (90%CI: 15.3-42.5).  No such pairs were available for escitalopram.

I think these findings are in keeping with my comments posted August 20, 2015, on Edward Shorter’s essay. I would say these data strongly support the assertion that citalopram does prolong ventricular repolarization and more so with real-world clinical use than do other SSRIs (the very limited escitalopram data and lack of pristine random prospective treatment assignment not withstanding).  Although with caveats, as discussed in my "Comment on the Controversy", a mean increase of 12.8 ms with the upper bound of the 90% CI extending to 18.2 ms very probably conveys some excess risk of TDP, although the absolute risk is quite low if it exits.  The TQT study is simply a precise and very sensitive way of detecting the event that can be performed during drug development, not requiring multiple thousands of subjects (subject n = 12,589; ECG n = 26,620) with multiple years of observation (22) to get the right, smaller group of on-treatment ECGs (on SSRI: 436; on citalopram: 35).


Maljuric NM; Noordam R; Aarts N; Niemeijer MN; van den Berg ME; Hofman A; Kors JA; Stricker BH; Visser LE. Use of selective serotonin re-uptake inhibitors and the heart rate corrected QT interval in a real-life setting: the population-based Rotterdam Study. Br J Clin Pharmacol 2015; 698-705.

Charles M. Beasley, Jr
February 25, 2016