Last month ISMP Canada did a safety bulletin on a suspected preventable death due to a known drug-drug interaction (ISMP Canada 2014). The case involved an elderly woman admitted with fever and presumed pneumonia. Prior to admission she was on citalopram 40 mg. daily, an anticoagulant, an antihypertensive, and nonprescription supplements. Initial antibiotics were ampicillin and gentamycin but these were later switched to azithromycin and ceftriaxone. She subsequently suffered a series of syncopal spells and a cardiac arrest. She had a markedly prolonged QT interval and was also hypokalemic. The citalopram and azithromycin were discontinued. She ultimately died. It was felt she had a prolonged QT interval syndrome secondary to citalopram and azithromycin that contributed to her death.
The article discusses the literature on the QT interval-prolonging effect of both citalopram and azithromycin and notes the effects may be additive. They note the warnings on both drugs that have been disseminated by both Health Canada and the US FDA. They note 2 potential points for intervention: (1) when the pharmacy processed the order for azithromycin and (2) when the prolonged QT interval was first noted.
In several of our prior columns (see list at the end of today’s column) we’ve discussed the risks of developing ventricular tachycardia and, specifically, Torsade de Pointes in hospitalized patients with prolonged QT intervals. There are a number of reasons why this syndrome is more likely to both occur and result in death in hospitalized patients. Hospitalized patients have a whole host of other factors that may help precipitate malignant arrhythmias in vulnerable patients. They tend to have underlying heart disease, electrolyte abnormalities (eg. hypokalemia, hypomagnesemia, hypocalcemia), renal or hepatic impairment, and bradycardia, all of which may be precipitating factors. More importantly they may have the sorts of conditions for which we prescribe the drugs that are primarily responsible for prolonging the QT interval (eg. haloperidol, antibiotics, antiarrhythmic agents, etc.). And many of those drugs are given intravenously and in high doses in the hospital as compared to the outpatient arena. Rapid intravenous infusion of such drugs may be more likely to precipitate Torsade de Pointes than slow infusion.
In our previous columns we provided a number of steps your organization should be taking to minimize the risks in such cases. We’ve recommended use of clinical decision support systems (CDSS) and computerized alerts as an important step. No one could possibly remember all the drugs that can prolong the QT interval. For a full list of drugs that commonly cause prolongation of the QT interval and may lead to Torsade de Pointes, go to the CredibleMeds™ website. That site also has a list of drugs that prolong the QT interval and might possibly cause Torsade de Pointes and another list of drugs that have conditional risk (eg. only when combined with other drugs). Hence, the need for other tools to help alert the clinician that a drug he or she is about to order might prolong the QT interval. In our April 9, 2013 Patient Safety Tip of the Week “Mayo Clinic System Alerts for QT Interval Prolongation” we discussed one such CDSS tool that had been implemented at the Mayo Clinic (Haugaa 2013).
Now another new study has demonstrated that use of CDSS and computerized alerts can reduce the risk of QT interval prolongation (Tisdale 2014). One of the most important considerations is developing a system in which the risk of alert fatigue is minimized. We know from multiple studies done in the past that physicians override over 90% of computer alerts during CPOE (computerized physician order entry). To minimize the risk of alert fatigue and still accomplish your goal of reducing the risk to patients it is important to (1) deliver the alert to the right person (2) deliver alerts only for the most potentially serious events and (3) provide alternative options for the physician’s response.
They system developed and implemented by Tisdale and colleagues did all three. First, the alerts first went to the pharmacist, who would then evaluate the situation and decide whether discussion with and recommendations for the physician were appropriate. Second, the thresholds to trigger the alerts were set at levels expected to minimize alert fatigue. And, third, the pharmacist responding to the alert would present the physician with some options for actions.
After reviewing extensive data on prior patients in their CCU’s and reviewing the literature, the researchers opted to trigger the alert when the QTc interval was >500 ms or there was an increase in QTc of ≥60 ms from baseline. Their system also identified through the electronic medical record multiple other conditions or laboratory results that identified patients at higher risk for QT interval prolongation.
After implementation of the CDSS system the found a significant reduction in the risk of QT prolongation (odds ratio 0.65). In addition, they found a significant reduction in the prescription of non-cardiac drugs known to prolong the QT interval (especially fluoroquinolone antibiotics and intravenous haloperidol).
Overall, 82% of alerts were overridden. That still compares favorably to the frequency with which other alerts are overridden. Most of the overrides were for cardiac drugs (eg. amiodarone or other anti-arrythmic drugs). The authors point out that overriding the computer alert did not mean that nothing was done. For example, even though the order for the drug may have been overridden the pharmacist and physician may have modified some other risk factor (eg. corrected an electrolyte disturbance or stopped another medication) or increased the frequency of QTc surveillance.
We recommend you read our previous columns listed below for suggestions on what your organization should be doing. Since many of the risk factors in hospitalized patients are potentially modifiable or avoidable it is imperative that we put into place systems that will help early identification of patients at risk.
Some of our prior columns on QT interval prolongation and Torsade de Pointes:
June 29, 2010 “Torsade de Pointes: Are Your Patients At Risk?”
February 5, 2013 “Antidepressants and QT Interval Prolongation”
April 9, 2013 “Mayo Clinic System Alerts for QT Interval Prolongation”
ISMP Canada. Preventable Death Highlights the Need for Improved Management of Known Drug Interactions. ISMP Canada Safety Bulletin 2014; 14 (5): 1-7 May
Haugaa KH, Bos JM, Tarrell RF, et al. Institution-Wide QT Alert System Identifies Patients With a High Risk of Mortality. Mayo Clin Proc 2013; 88(4): 315-325
Tisdale JE, Jaynes HA, Kingery J, et al. Effectiveness of a Clinical Decision Support System for Reducing the Risk of QT Interval Prolongation in Hospitalized Patients. Circulation: Cardiovascular Quality and Outcomes 2014; published online before print May 6, 2014
Print “PDF version”